Encyclopedia of Energy Research and Policy

ENCYCLOPEDIA OF ENERGY RESEARCH AND POLICY

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

ENCYCLOPEDIA OF ENERGY RESEARCH AND POLICY

A. L. ZENFORA EDITOR

Nova Science Publishers, Inc. New York

Copyright © 2010 by Nova Science Publishers, Inc. All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works. Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.

Library of Congress Cataloging-in-Publication Data Available upon request

ISBN 978-1-61324-544-6 (eBook)

Published by Nova Science Publishers, Inc. Ô New York

CONTENTS Preface

xi

Research and Review Studies

1

Chapter 1

Energy Markets United States Government Accountability Office

3

Chapter 2

Three-Dimensional Simulation of Base Carrier Transport Effects in Back Side Point Contact Silicon Solar Cells K. Kotsovos and K. Misiakos

53

Multiple Effect Distillation of Seawater Water Using Solar Energy – The Case of Abu Dhabi Solar Desalination Plant Ali M. El-Nashar

85

Solid State Organic Photoelectrochemical Solar Energy Conversion Based on Conjugated Substituted Polythiophenes Teketel Yohannes

159

Chapter 3

Chapter 4

Chapter 5

A New Approach to Hybrid Systems of Renewable Energy Utilization Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley, P. Horley and L. Bulat

201

Chapter 6

Dye-Sensitized Nano SnO2:TiO2 Solar Cells Weon-Pil Tai

Chapter 7

Strategies for Reducing Carbon Dioxide Emissions - The Case of Botswana Rural Communities C. Ketlogetswe and T.H. Mothudi

231

The Applying of Coatings and Surface Thermal Treatment of Materials in Solar Furnaces: Theory and Experiments V.V. Pasichny and B.A. Uryukov

245

Chapter 8

219

vi Chapter 9

Chapter 10

Contents Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides for Solar Cells Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley, V. Khomyak, S. Bilichuk, V. Grechko and P. Horley Dynamic Impedance Characterization of Solar Cells and PV Modules Based on Frequency and Time Domain Analyses D. Chenvidhya, K. Kirtikara and C. Jivacate

Chapter 11

Wind Energy Technology Overview United States Department of the Interior, Bureau of Land Management

Chapter 12

Federal and State Regulatory Requirements Potentially Applicable to Wind Energy Projects United States Department of the Interior, Bureau of Land Management

277

301 327

367

Chapter 13

Commercial Wind Energy Projects United States Department of the Interior, Bureau of Land Management

Chapter 14

Biomass And Bioenergy Research In Tropical Africa: State Of The Art, Challenges And Future Directions Jonathan C. Onyekwelu and Shadrach O. Akindele

385

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material Bojana Klasnja, Sasa Orlovic, Zoran Galic and Milan Drekic

419

Biobased Polymers by Chemical Valorization of Biomass Components B. Kamm, M. Kamm, I. Scherze, G. Muschiolik and U. Bindrich

449

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets J.C. Morán, J.L. Míguez, E. Granada and J. Porteiro

481

Chapter 15

Chapter 16

Chapter 17

Chapter 18

Chapter 19

Chapter 20

Negative Emission Biomass Technologies in an Uncertain Climate Future Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný and Michael Obersteiner A Review of the Socio-Economic and Environmental Benefits of Biomass Gasification Based Power Plant: Lessons from India Kakali Mukhopadhyay The Energy Balance and Fuel Properties of Biodiesel Mustafa Acaroglu and Mahmut Ünaldı

377

501

549 589

Contents Chapter 21

An Experimental Study on Performance and Exhaust Emissions of a Diesel Engine Fuelled with Various Biodiesels Nazim Usta

Chapter 22

New Materials from Lignin Carlo Bonini and Maurizio D’Auria

Chapter 23

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic Treatment of Swine Waste: Pilot Studies and FullScale Design James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

Chapter 24

Nuclear Dynamics Modelling by Recurrent Neural Networks F. Cadini, E. Zio and N. Pedroni 675

Chapter 25

Primary Cosmic Ray Studies Based on Atmospheric Cherenkov Light Technique at High-Mountain Altitude A.L. Mishev, S. Cht. Mavrodiev and J.N. Stamenov

Chapter 26

Development of Subchannel Analysis Code for CANDU-SCWR Yu Jiyang, Wang Songtao, Jia Baoshan

Chapter 27

Application of Best Estimate Computational Tools for Safety Accident Analysis in Nuclear Plants Anis Bousbia Salah, Tewfik Hamidouche and Francesco D’Auria

vii

603 625

653 675

731 779

811

Chapter 28

Advanced Fuel Fusion Reactors: Towards a Zero-waste Option Massimo Zucchetti

829

Chapter 29

Solar Thermal Power Generation on Mars Viorel Badescu

843

Chapter 30

Equilibrium Phases in Zirconium Alloys of Concern to the Nuclear Industry: Isothermal Sections of the Zr-Cr-Sn and Zr-Cr-Ti Systems S.F. Aricó, R.O. González and L.M. Gribaudo

893

Nuclear Nonproliferation: IAEA Safeguards and Other Measures to Halt the Spread of Nuclear Weapons and Material Gene Aloise

915

Radial-Bias-Combustion and Central-Fuel-Rich Swirl Pulverized Coal Burners for Wall-Fired Boilers Zhengqi Li

939

Chapter 31

Chapter 32

Chapter 33

Chapter 34

Fuel Cell Combined Cycle Power Generation System Installed into Micro-Grid Shin’ya Obara

1061

Electricity from Renewable Energy Sources: A Multi-Criteria Evaluation Framework of Technologies Fausto Cavallaro

1107

viii

Contents

Chapter 35

Gas Turbines and Electric Distribution System Francisco Jurado

1139

Chapter 36

Micro CCHP: Future Residential Energy Center R. Z. Wang and D. W. Wu

1173

Chapter 37

Sensitivity Calculation in Real Time Transmission Network and Energy Markets Jizhong Zhu

1199

Wide-Area Monitoring and Analysis of Inter-Area Oscillations Using the Hilbert-Huang Transform A. R. Messina, M. A. Andrade and E. Barocio

1219

Chapter 38

Chapter 39

Unconventional Problems in Power Systems Protection Mahmoud Gilany and Mohamed A. Mahmoud

Chapter 40

BME-Generated Temperature Maps of the Nea Kessani Geothermal Field Konstantinos Modis, Hwa-Lung Yu, George Christakos, Robert Stewart and George Papantonopoulos

Chapter 41

Chapter 42

Chapter 43

Chapter 44

Chapter 45

1251

1265

Advances in Studies of Thermal-Fluid Geochemistry and Hydrothermal Resources in China Jianguo Du, Youlian Zhang and Heping Li

1281

A Comparative Analysis of the Geothermal Fields of Larderello and MT Amiata, Italy Giovanni Gianelli

1321

Sedimentary Characteristics of Coal Beds in Intramontane Basins (Massif Central, France) Wang Hua and Xiao Jun

1349

Coupling of Thermal and Chemical Simulations in a 3-D Integrated Magma Chamber-Reservoir Model: A New Geothermal Energy Research Frontier Surendra P. Verma and Jorge Andaverde Determination of the Damage Effect in Geothermal Wells Using Inflow Type Curves A. A. Aragón, S. L. Moya and A. M. C. Suárez

1361

1403

Expert Commentaries

1443

Commentary A Innovative Techniques for the Simulation and Control of Nuclear Power Plants Antonio Cammi and Lelio Luzzi

1445

Contents

ix

Commentary B Analysis and Characterization of Complex Inter-Area Oscillations from Measured Data: A Time-Frequency Perspective A. R. Messina, E. Barocio and M. A. Andrade

1449

Index

1453

PREFACE Chapter 1 - To better understand how changes in domestic and international petroleum products markets have affected prices, this book evaluates trends in (1) the international trade of petroleum products, (2) refining capacity and intensity of refining capacity use internationally and in the United States, (3) international and domestic crude oil and petroleum product inventories, and (4) domestic petroleum supply infrastructure. This is an excerpted and indexed version. Chapter 2 – This work presents a theoretical investigation of rear junction point contact silicon solar cells through three-dimensional numerical simulation based on the solution of minority and majority carrier transport equations in the base of the cell. The device series resistance is evaluated through the simulated current-voltage (IV) curves under AM1.5 illumination conditions and its dependence on back contact geometry is examined. Results are presented which show the influence of the majority carrier transport in the base to the solar cell performance. A comparison is also performed with two other similar types of point contact solar cells, one with the emitter located on the front surface and the other on both surfaces, as well as with a conventional solar cell structure. Chapter 3 - This report describes the solar desalination test plant in Abu Dhabi, UAE and gives a summary of its first year performance and economics. The plant has been operating successfully for 18 years supplying fresh water to the City of Abu Dhabi. The plant was commissioned in September 1984 and was running until the year 2002 when it was dismantled after fulfilling its objectives. The aim of the plant is to investigate the technical and economic feasibility of using solar desalination of seawater in providing fresh water to remote communities in the Middle East and to obtain long-term performance and reliability data on the operation of the plant. The plant has proved its technical feasibility and proved to be reliable in operation with few minor maintenance problems that required slight plant modification. Maintenance routines were established to maintain high plant performance. The economic feasibility of the plant was established by comparing the cost of water from a solar MED plant with a conventional MED plant using fossil fuel for plant capacity ranging from 100 m3/day to 1000 m3/day. It was found that the cost of water from solar MED plants is competitive with that from a conventional MED plant if the cost fuel continues to rise. Chapter 4 - The utilization of organic materials for photovoltaic devices has been investigated intensely during the last couple of decades. Earlier studies concentrated on molecules that had high optical absorption in the visible region of the electromagnetic spectrum. Recent discovery of conjugated polymers having semiconductor-like behavior has

xii

A. L. Zenfora

started to stir excitement because such materials are not only able to function in a similar manner to the inorganic semiconductors but also have important advantages such as: low cost, light weight, ease of fabrication and the possibility of large area coatings. Their use as photoactive electrodes is of increasing interest, as the processing possibilities of conjugated polymer materials have become more developed. Furthermore, the high absorption coefficients of these materials and the possibility of varying the band gap by molecular engineering have opened up new options for solar energy conversion. Among the conjugated conducting polymers, neutral, substituted polythiophenes exhibit interesting properties as semiconducting photoactive materials and are used for conversion of optical energy into electrical energy. Investigation of the photoelectrochemistry of conducting polymers was mainly focused on their use as protective films against photocorrosion and as photoactive electrodes in liquid junction photoelectrochemical cells (PECs). Photocorrosion and side reactions involving the electrolyte solution and the difficulty of packaging limit the working life of liquid electrolyte PECs. Solid-state PECs with the use of solid polymer electrolytes provide a means to eliminate this problem since they can easily be processed into thin films over large areas and are easier to encapsulate. The solvent-free ion conducting polymer electrolytes eliminated handling, portability, and packaging problems encountered in liquid junction photoelectrochemical cells. Basically, the photoelectrochemical properties occurring in these systems are the same as those occurring in systems based on semiconductor photoelectrodes in contact with liquid electrolytes. In this chapter an overview of the studies made on solid-state photoelectrochemical solar energy conversion devices using standard photoelectrochemical and photoelectrical characterization techniques is presented. The photoelectrochemical cells contain a thin film of semiconducting conjugated substituted polythiophenes as a light-harvesting unit, a redox couple complexed with an ion conducting polymer electrolyte, and a counter electrode. Chapter 5 - A general analysis is given of hybrid systems consisting of different combinations of 4 devices frequently employed for renewable energy utilization: Photovoltaic Solar Panel (PV), Solar Thermal Plane Collector (ST), Wind Generator (WG) and Heat-toElectric/Mechanic Energy Convertor (HE); some of the combinations include radiation energy flux concentrators of different degrees. The main result of the consideration made is that the hybrid systems are more efficient than the sum of the constituents and more stable in relation to spontaneous variations of the renewable energy source potential (like wind velocity, insolation, etc.). However, to realize the possibilities mentioned, all the elements of a given hybrid system have to be especially designed and made for this specific system. For example, the PV panel for the hybrid PV/Thermal system ought to have a substrate with high thermal conductivity, to allow for heat extraction from the panel by the adjacent Solar Thermal Plane Collector, and practically no commercial panels with these characteristics are available. Besides, the PV panel as a part of the hybrid system will demand a special choice of semiconductor material and surface treatment which could be different from those for conventional panels. The limiting efficencies for some hybrid systems are estimated; these efficiencies exceed the efficiencies of separate use of the devices discussed. The most promising hybrid system is the PV panel made as spectrum splitter in combination with HE converter, of which total efficency could be around 50 %. Chapter 6 – The nanostructured SnO2:TiO2 bilayered and composite solar cells sensitized by eosin Y and RuL2(NCS)2 dyes are prepared and the photoelectrochemical

Preface

xiii

properties of the cells are investigated. The semiconductor films possess the grain size of nanometer order and have nanoporous structure. The bilayered cell shows higher IPCE (incident photon- to-current conversion efficiency) value than the single and composite cells. A maximum IPCE value of 88.1% was reached at 540 nm wavelength in the bilayered cell with 3.5μm-thick SnO2 and 7μm-thick TiO2 sensitized by RuL2(NCS)2 dye. The higher IPCE value in the bilayered cell is attributed to the promotion of the charge separation by fast electron transfer process from the excited dye to SnO2 in the SnO2/TiO2/dye system with different conduction band edge energy positions. Chapter 7 - The International Community’s pre-occupation with the ever-escalating dangers posed by gaseous pollutants need not be overemphasized. Suffice to mention, however that the magnitude of the dire negativity of pollutants is reflected in the numerous international charters that were promulgated with a common objective to sensitise the world about the need to move toward setting up minimum permissible levels of emission for activities whose execution result in atmospheric pollution. In addition, authorities have also gone so far as to offer incentives / motivation as a means of assuaging nations towards implementing various strategies for minimising atmospheric pollutions. This paper explicates efforts taken by. The government of Botswana in an effort to strive for compliance with international protocols and standards to safeguard against deterioration of the planet. Focus will specifically be paid to examining any concrete measures taken with the view to curb the negative impacts of carbon dioxide gas. The suitability and sustenance or, otherwise, of government projects envisaged for reducing carbon dioxide emission levels generated during the combustion of fuelwood and other related energy sources used by rural communities in Botswana will also be discussed. Chapter 8 - Solar furnaces make it possible to obtain a temperature of heating equivalent to 3500 K and above it an oxidizing air medium and without any outside contamination. They are used for investigation of materials in the Institute for Problems of Materials Science (IPMS) of theNational Academy of Science of Ukraine (NASU) for the past 40 years. The created experimental base consists of 14 different solar installations of power from 0.1 up to 10 kW. They are included in the two laboratories located in Kyiv and on the Black Sea coast. Some optical furnaces on Xe arc lamps which are the simulators of solar furnaces are added to the experimental base. In the given chapter the works of the last few years are concentrated. They are dedicated to surface heating of materials intended for obtaining coatings and improving their protective, decorative and other operational characteristics. The specialists of various fields of engineering and production are engaged in the development of these energy-intensive processes with the use of traditional energy sources. Their substitution for renewable solar radiation if it is possible can cause not only saving on utilities saving but in some cases the improvement of coatings quality due to chemical purity of the heating source. Some theoretical and experimental results of the investigation in the given field fulfilled in the IPMS are represented in the proposed work. Using an approximate integral method for solving heat conduction equation the problem is solved for the determination of the rate of thermal treatment of a surface by partial melting in a solar furnace when the sample is stationary and moves relative to the focal spot depending upon the given thickness of fused layer. Taking into account the absence hitherto of industrial (commercial) production of solar furnaces the theoretical and practical foundations have been developed in the IPMS for the creation of solar radiation concentrators on the basis of metallic antennae with plane mirror facets. As it is described in the given work the energy characteristics of these

xiv

A. L. Zenfora

concentrators fully come up to the standards which are necessary for the realization of the greatest part of the investigated processes. Chapter 9 - Transparent conductive oxides SnO2, In2O3-SnO2 (ITO) and CdO are widely used for different optoelectronic devices, including photovoltaic applications. Depending on technological conditions, oxide films can be either high- or low-resistive. This paper presents the results of complex investigation of technological parameters influence (such as chamber pressure, substrate temperature, magnetron cathode power, and duration of isothermal annealing in the air) on specific resistance and transmission coefficient of oxide thin films, grown by reactive magnetron sputtering. Ar-O2 mixture was used as a carrier gas for direct current sputtering; high-frequency sputtering was performed in pure Ar atmosphere. Substrates for the films were made of quartz glass and silicon. Significant attention was paid to the transformation of defect subsystems after isothermal annealing in the air. The authors determined optimal technological regimes allowing to obtain reproducible high-quality thin films of tin, indium and cadmium oxides with the following electrical and optical parameters: SnO2 – specific resistivity ρ = 6 – 15.10-4 Ω⋅cm, optical transmission T = 90 – 95% in transparency region; ITO – ρ = 4 – 6.10-4 Ω⋅cm, T = 90 – 95%; CdO – ρ = 5 – 20.10-4 Ω⋅cm, T = 80 – 90%. Chapter 10 - This article describes new methods to derive dynamic impedance of solar cells and PV modules from time and frequency domain analyses. Initially, the authors propose a new method, based on the frequency domain analysis, to measure dynamic impedance of x-Si solar cells and PV modules in the dark using basic instruments and FFT analysis. The dynamic parameters in the AC equivalent circuit, in addition to the DC model, consists of dynamic resistance, diffusion capacitance and transition capacitance. Loci of impedance in the complex plane can be obtained by inputting a small signal square wave, superimposing on either forward bias or reverse bias, to cells or modules. Such technique is compared with sinusoidal inputting. All of these parameters can be obtained from impedance loci in the complex plane. The impedance of a cell or a module can be derived in a closed form equation in terms of frequency dependent and voltage dependent resistance and capacitance under the dark condition with reverse bias. The relationship between the dynamic and static characteristics is compared for solar cell modules having low and high fill factors. Another new analytical method determining solar cell and module dynamic impedance is demonstrated using the same measuring techniques. Determination of dynamic parameters, previously outlined, and time constant of solar cells and modules, based on a time domain response, can be simultaneously obtained at each bias condition. The merits of this second characterization method using square wave inputs are reduction in measuring steps and yielding of dynamic parameters and time constants in a single measurement. Experiments on polycrystalline and amorphous silicon cells and modules are also conducted and their results will be separately revealed at a later date. Knowledge of dynamic impedance characterization of solar cells and modules will lead to better understanding of behaviors of PV grid-connected systems and improvement of power quality from such distributed power generation systems. Chapter 11 - Modern wind energy technologies rely heavily on the very complex scientific discipline of fluid dynamics (which includes the study of the atmosphere) and the equally complex engineering discipline of aerodynamics. A comprehensive treatment of either of these disciplines is well beyond the scope of this programmatic environmental

Preface

xv

impact statement (PEIS). The discussions that follow are intended only to establish a basic understanding of wind technology and the factors that control its evolution. References are provided for those who wish to have a more detailed understanding of wind technology. This appendix provides an overview of the fundamentals of wind energy and wind energy technologies, describes the major components of modern wind turbines, and introduces terms that are unique to the field of electric power generation using wind energy. Important site characteristics and critical engineering aspects of wind energy technologies are presented, and their respective influences on future development decisions are discussed.[1] An overview of the current state of wind energy technology and ongoing research and development (R&D) is provided. Descriptions of a typical wind energy project and the major actions associated with each phase of development — site monitoring and testing, construction, operation, and decommissioning — are presented in part 3 of this PEIS. Chapter 12 - The tables that follow list the major federal and state laws, Executive Orders, and other compliance instruments that establish permits, approvals, or consultations that may apply to the construction and operation of a wind energy project on Bureau of Land Management (BLM)-administered lands. The general application of these federal and state authorities and other regulatory considerations associated with such construction and operation are discussed in Chapter 3. The tables are divided into general environmental impact categories. The citations in the tables are those of the general statutory authority that governs the indicated category of activities to be undertaken under the proposed action and alternatives. Under such statutory authority, the lead federal or state agency may have promulgated implementing regulations that set forth the detailed procedures for permitting and compliance. Definitions of abbreviations used in the tables are provided here. Chapter 13 - Data on commercial wind energy projects in the western states that are within the scope of this programmatic environmental impact statement (PEIS) are displayed in the tables below. The American Wind Energy Association (AWEA) compiles and maintains all of the data displayed below. All data presented are current as of January 14, 2004. All data are accessible electronically from the AWEA Web site at http://www.awea.org/projects/index.html. Data presented in the tables below are updated quarterly by the AWEA. The Bureau of Land Management (BLM) cannot guarantee the completeness or accuracy of these listings. Submission by wind farm developers or operators of project information to AWEA for inclusion in these listings is voluntary. Chapter 14 - Forest biomass and bioenergy production currently play a very important role in energy generation in tropical African countries, especially in the rural areas where between 75 and 95% of the populace depend on fuelwood as the primary energy source. Given the current high population growth, the low rates of switching to non-carboniferous household energy sources as well as the inefficiency of other energy sources, the importance of biomass and bioenergy in household energy generation in tropical African countries is expected to increase in the future. This paper examines the sources and extent of biomass production in tropical African countries as well as their current contribution to bioenergy supply and possible future trend. The current status and prospects of bioenergy technologies, the state of biomass and bioenergy research in the sub-region as well as the methodologies used in obtaining local and national biomass estimates were reviewed. The paper also discussed the challenges facing biomass and bioenergy research in tropical Africa, and

xvi

A. L. Zenfora

stressed the need for more collaboration with the developed countries to be able to tackle the challenges. The paper finally examines the likely future research directions and makes recommendations towards a more efficient and environmental-friendly utilization of biomass and bioenergy in the sub-region. Chapter 15 - Fast-growing broadleaf species (poplars, willows and black locust), raised in dense, short-rotation plantations, very often on the soils unsuitable for agricultural crops, produce a high yield of biomass. A significant amount of thermal energy can be obtained by direct combustion of young plant biomass (aged from one to three years) converted into chips by chipping the whole trees together with bark and branches. In this aim, the Institute carried out systematic multiannual research on the improvement of several poplar clones in order to increase the yield of biomass. Also for this purpose, the selection focused on the clones which are best adapted to the conditions of very dense planting, which is the main condition required from the foresters in the establishment of energy plantations. Based on the calorific value of wood and bark of the study poplar clones, it is assessed the quantity of energy which could be produced by the combustion of the chipped biomass of one-year, i.e. two-year-old plants. The higher heating value of wood and bark was determined for several poplar clones (Populus spp.) of different ages and plants, as well as the trees from mature plantings (aged from 8 to 14 years). By FVI (Fuel Value Index) which takes into account ash content, wood basic density, as well as moisture content, it was determined that poplar wood can be significant energy raw material, primarily because of its short production cycle and very high volume increment. The plantations are established in two variants, by planting the cuttings of the selected poplar clones, with two planting spaces, i.e. with 38,461 plant/ha, and 83,333 plant/ha, on the previously selected and prepared soil. To define the produced biomass of individual clones, the increment elements were measured after the cycles of one and two years. Average dry matter biomass yield reached 21 t ha-1 year-1 (38,461 plant/ha), and 12 t ha1 year-1 (83,333 plant/ha). Based on calorific values of oven dry wood and bark of each clone, average energy potential of researched poplar clones was estimated up to 395 GJ ha-1 year-1, and for denser plantations up to 222 GJ ha-1 year-1. Chapter 16 - Plants represent a natural chemical and polymer factory and food plant. Biorefineries combines necessary technologies between biogenic raw material and intermediates and final products. The paper present two strategies for producing of polymeric materials, firstly the utilization of the pre-determined natural macromolecular structure and secondly the using of biogenic building blocks. The first step is the fractionation technology from green biomass for producing of fiber-rich press cake and a nutrient rich-green juice. The main focus is directed on products, such as proteins, polylactic acid, cellulose and levulinic acid- sequence products and their application as well as their market. Chapter 17 - In this chapter a set joint of experimental techniques for assessing biomass combustion devices is presented. Small scale energy converters such as chimneys, boilers, stoves, etc, producing heat and/or hot water by combustion of biomass (wood, pellets, briquettes, etc.) are especially suited to domestic purposes. However, in regular commercial combustion conditions, this kind of use still has some disadvantages: besides the fact that some emissions (volatile organic carbons, carbon monoxide or NOx) may still be high, it is difficult to compare the quality and performance of equipment working in very different combustion conditions.

Preface

xvii

Due to their relatively low cost and the complexity of combustion in such devices, modelling by numerical analysis is seldom attempted. Controlling operational factors are usually designed and regulated based on the manufacturer’s experience or on handbook values. In order to protect customers, and to assure compliance with minimum requirements for energy performance and maximum limits on pollutant emissions, several national and international regulations have been developed in recent years. Experimental analysis of these devices is a key technique for control and improvement. Chapter 18 - Mitigation of and adaptation to climate change belong to the most pressing global challenges for the 21st century. Major mitigation options include improved energy efficiency, shifting towards less carbon-intensive fossil fuels, increased use of energy sources with near-zero emissions, such as renewables and nuclear, CO2 capture and permanent storage (CCS), and carbon sequestration by protection and enhancement of biological absorption capacity in forests and soils. Bioenergy is one of several energy sources which could provide society with energy services with near-zero emissions. Bioenergy has a unique feature, however, which distinguishes it from other low-emitting energy supply options, such as solar, wind, nuclear, and clean fossil energy technologies. Bioenergy conversion could be integrated with a process which separates carbon. If the biomass feedstock is sustainably produced and the separated carbon is subsequently isolated from the atmosphere for a very long time the entire process becomes a continuous carbon sink – in other words such technologies yield negative CO2 emissions. Negative emission biomass technologies can be centralised or distributed; Centralised negative emission biomass technologies, biomass energy with CO2 capture and storage (BECS), build on the conversion of biomass into energy carriers in centralised conversion plants integrated with CO2 capture. The captured CO2 is subsequently transported and stored in geological formations. Distributed negative emission biomass technologies are based on the production of long-term carbon-sequestering charcoal soil amendment, with or without co-production of biofuels. In this chapter a BECS implementation scenario study is presented. The study analyses investments in BECS in a pulp and paper mill environment. The investment analysis is carried out within a real options framework taking into account the potential revenue from trading generated emission allowances on a carbon market. Uncertainty is considered in the economic modelling through the use of stochastically correlated price processes of one input price (biomass) and two output prices (electricity and CO2 emission permits) that are consistent with shadow price trajectories of a large-scale global energy model. The results suggest that BECS can be economically feasible within approximately 40 years. The chapter also discusses Research and Development needs for better understanding of the future overall potential of negative emission biomass technology implementation. Chapter 19 - There is a steady and continuing interest in biomass gasification in both the developed countries and developing countries. While the advanced countries are interested primarily from considerations of reduced emissions and waste utilisation, the developing countries look at biomass gasification as a means to augment commercial energy like electricity, diesel, fuel oil etc. India, a tropical country with a vast geographical area is richly endowed with renewable energy sources like solar, wind, biomass which can play a crucial role in meeting end use energy needs in a decentralised manner. One of the major goals of the ninth and tenth five year plan is strengthening of infrastructure (energy, transport, communication, irrigation) in

xviii

A. L. Zenfora

order to support the growth process on a sustainable basis. It is usually the tendency of the developing countries to equate development with economic growth and to further equate economic growth with energy consumption especially electricity. India being a developing country has also given due emphasis on strengthening its energy position accordingly. Moreover threat from Green House Gasses (GHG) also has caused worldwide concern. In India electric power generation is the largest source of GHG emissions. It accounts for 48% of carbon emitted. These concerns point towards more rational energy use strategies. The renewable and recycling process makes biomass possible to generate power without adding to air emissions. Biomass (firewood, agricultural residue, and dung) is one of the main fuels in India, particularly in the energy-starved rural sector. The biomass power potential in India was 16,000 MW (excluding co-generation), but the achievement in this respect is negligible (Installed capacity - 630 MW Project under implementation - 630 MW, as on March 2005). It brings out the fact that much of the potential of biomass gasification is still unexplored. Globally, India is in the fourth position in generating power through biomass and with a huge potential, is poised to become a world leader in utilization of biomass. According to the Planning Commission of India, in its Tenth Five Year Plan, announced that 26.10 per cent of the Indian populations are below the poverty line and mostly belongs to rural areas. The inequitable distribution has been evident from the fact that although 70% of India’s population lives in the rural areas, only 29% of rural households have electricity supply as against 92% of urban households. Of the half a million or so villages in India, about 3, 10,000 villages have been declared to be electrified and 80,000 more villages remain completely un-electrified. There are a number of constraints to supply power to remote rural area such as small human settlements, geographically dispersed villages, seasonally of loads etc. In the absence of adequate network and hence supply of power to remote rural areas the household depend largely on primary energy sources like kerosene and diesel for lighting. No commercial investments in micro enterprises can therefore be made by either individuals or companies without installing diesel generators which have a very high generating cost. Biomass gasifier is a leading option in that respect. Besides, the supply of power to remote rural areas from the centralised grid is not competitive than a modern biomass gasification based decentralised power plant. Estimate from an Indian village shows that modest 50 kW of installed capacity per village will lead to total saving of 52000 million Rs (Rs 5200 Crore / 1100 million US $) in power plant investments. In energy terms, the saving in TandD losses will release a generation capacity of 800 MW for profitable sale. Reduced pollution and reduction of CO2 emissions will be the other advantages of a decentralised renewable energy based system for the rural areas. The purpose of the present paper is to evaluate the rural electrification programme in India undertaken by the Ministry of Non Conventional Energy Sources (MNES), Government of India, through biomass gasifier power plant. It explores the eradication of poverty that has been made possible by introducing biomass gasification based power plant in remote rural areas in India. Creation of jobs in the power stations, small-scale business, commerce and industries and also improvement in the quality of life is assessed. The paper concludes with policy options relevance for the other developing countries. Chapter 20 - In this study energy balance and fuel properties of biodiesel has been calculated. Accordingly, the cost of 1 liter of oil is calculated 0.32 € after the income from the seed meal is deduced. Finally, the cost of per unit of biodiesel (1 liter) was calculated as 0.55

Preface

xix

€, after deduction of the income provided by the sales of glycerin for use in soap and cosmetic industry. The energy equivalent of total output was calculated 147605.50 MJ per hectare. The net energy gain (refined oil) was found as 15105.63 MJ per hectare (The net energy ratio 11.031) according to yield and inputs values. The viscosity values of vegetable oils vary between 27.2 and 53.6 mm2/s whereas those of vegetable oil methyl esters between 3.59 and 4.63 mm2/s. The flash point values of vegetable oil methyl esters are highly lower than those of vegetable oils. The flash point values of vegetable oil methyl esters are highly lower than those of vegetable oils. An increase in density from 860 to 885 kg/m3 for vegetable oil methyl esters or biodiesel increases the viscosity from 3.59 to 4.63 mm2/s and the increases are highly regular. There is high regression between density and viscosity values vegetable oil methyl esters. The relationships between viscosity and flash point for vegetable oil methyl esters are irregular. An increase in density from 860 to 885 kg/m3 for vegetable oil methyl esters increases the flash point from 401 to 453 K and the increases are slightly regular. The LHV values of vegetable oils methyl ester vary between 35.74 and 39.16 MJ/kg. Chapter 21 - Instability and increases in prices of petroleum-based fuels, gradual depletion of world petroleum reserves and increases in environmental pollution caused by exhaust emissions speed up research on renewable alternative fuels. Vegetable oils have been considered as renewable alternative fuels in compression ignition engines for a long time. However, they have not been widely used as fuels in the engines due to some technical and economical drawbacks. Some properties of vegetable oils such as high viscosity, lower volatility and lower heat content result in technical problems in direct using of vegetable oils in short and long term applications. From economical point of view, the main problem is that vegetable oils have been more expensive than petroleum Diesel fuel. There are various ongoing studies on solving these problems to be able to use vegetable oils in Diesel engines. Different methods such as preheating oils, blending or dilution with other fuels, thermal cracking/pyrolysis and transesterification have been developed. Among these techniques, transesterification appears to be the most promising one. It is a chemical process converting vegetable oils to alcohol ester of oil named as biodiesel. In general, biodiesel-Diesel fuel No.2 blend can be used as a fuel in Diesel engines without modification. Specifications of biodiesel mainly depend on oil, transesterification process, type and amount of alcohol, type and amount of catalysis, reaction time and temperature. Biodiesel can be produced from different kinds of vegetable oils. Since prices of edible vegetable oils are higher than that of Diesel fuel No. 2, waste vegetable oils and non-edible crude vegetable oils are mostly preferred as potential low priced biodiesel sources. It is also possible to use soapstock, a by-product of edible oil production, for cheap biodiesel production. In this study, various biodiesels were produced from raw vegetable oils (rapeseed oil, soybean oil, cotton seed oil, palm oil and tobacco seed oil), waste sunflower vegetable oils and hazelnut oil soap stock-waste sunflower vegetable oil, and their specifications were compared with each other. The biodiesel (20% in volume) - Diesel fuel No.2 (80% in volume) blends were tested in a four cycle, four cylinder turbocharged indirect injection Diesel engine. The effects of biodiesel addition to Diesel fuel No.2 on the performance and emissions of the engine were investigated at full load. Experimental results showed that the

xx

A. L. Zenfora

biodiesels can be partially substituted for Diesel fuel No.2 at most operating conditions in terms of performance parameters and emissions without any engine modification and preheating of the blends. Chapter 22 - Lignin, obtained through steam explosion from straw, was completely characterized via elemental analysis, gel permeation chromatography, ultraviolet and infrared spectroscopy, 13C and 1H nuclear magnetic resonance spectrometry. Lignin powder was used for the preparation of blends with low-density polyethylene (LDPE), linear low-density polyethylene (LLDPE), high-density polyethylene (HDPE) and atactic polystyrene (PS). The obtained blends are processable through the conventional techniques used for thermoplastics; the modulus slightly increases for most lignin-polymer blends, while the tensile stress and elongation reduce. Moreover, lignin acts as a stabilizer against the UV radiation for PS, LDPE and LLDPE. Polyurethanes were obtained treating steam exploded lignin from straw with 4,4’methylenebis(phenylisocyanate), 4,4’-methylenebis(phenylisocyanate) – ethandiol, and poly(1,4-butandiol)tolylene-2,4-diisocyanate terminated. The obtained materials were characterized by using gel permeation chromatography, infrared spectroscopy and scanning electron microscopy. Differential scanning calorimetry analysis showed a Tg at -6 °C, assigned to the glass transition of the poly(1,4-butandiol) chains. The presence of ethylene glycol reduced the yields of the polyurethanes. The use of the prepolymer gave the best results in polyurethanes formation. Steam exploded lignin was used as starting material in the synthesis of polyesters. Lignin was treated with dodecanoyl dichloride. The products were characterized by using gel permeation chromatography, infrared spectroscopy, 13C and 1H nuclear magnetic resonance spectrometry, and scanning electron microscopy. Chapter 23 - Pilot plant experiments with both a 3.79 m3 batch and semi-continuous reactor have been performed with whole, fresh swine manure and the production of biochemical energy as heat has been both measured and calculated. The reactor operates at near atmospheric pressure and about 55˚ C. The systems were equipped with a patented offgas recycle process that may be shown to increase the amount of recoverable and useful energy from the reactor compared with a once-through aeration system. The batch study, a statistically-designed series of experiments, was held to investigate the relationships of initial or feed total solids concentration, fresh air fed, and offgas recycle rate to the total biochemical energy produced in the system. A linear model was developed to determine the importance of these factors in design. The model indicates optimism for improved operation over pilot plant work performed. The recycle concept is most useful when a reactor design is desired with a relatively shallow depth (e.g., 3 m aeration submergence), as may be found in in-ground concrete tanks. Good results may be achieved in deep aeration submergence reactors with no offgas recycle, owing to the higher oxygen transfer efficiencies in tall tanks. A combination of tall tanks and offgas recycle is synergistic with improved results. These results will be presented and discussed in the context of a full-scale farm application. When compared to once-through aeration systems, offgas recycle also leads to major reductions of emitted offgas, and aiding odor and pollutant reductions. Other potential applications will also be discussed. Chapter 24 - The design, operation and control of highly risky industrial systems, such as in the nuclear, chemical and aerospace, entail the capability of accurately modelling the

Preface

xxi

nonlinear dynamics of the underlying processes. In this respect, Artificial Neural Networks (ANNs) have gained popularity as valid alternatives to the lengthy and burdensome analytical approaches to reconstructing complex nonlinear and multivariate dynamic mappings. In particular, Recurrent Neural Networks (RNNs) are attracting significant attention, because of their intrinsic potentials in temporal processing, e.g., time series prediction, system identification and control, temporal pattern recognition and classification, whereas classical feedforward neural networks are in general capable of representing only static input/output mappings. The aim of this chapter is to present two kinds of recurrent neural networks and show their capabilities of approximating the temporal evolution of complex dynamical systems. First, the Elman’s recurrent network is considered, in which external feedback connections feed the output of the hidden nodes back to a set of additional nodes placed in the input layer. The network’s modelling capabilities are demonstrated on a case study concerning the prediction of the behaviour of a steam generator in a nuclear power plant. A more advanced type of recurrent architecture is then presented: the Infinite Impulse Response-Locally Recurrent Neural Network (IIR-LRNN), characterized by nodes which contain local, internal feedback paths realized by means of IIR synaptic filters providing the network with the necessary system state memory. The effectiveness and criticalities of this type of recurrent neural network are tested on two highly nonlinear dynamic systems of literature, the discrete-time Back-Tsoi model and the continuous-time Chernick model describing the evolution of the neutron flux in a nuclear reactor. Chapter 25 - A new method for primary cosmic ray investigations based only on atmospheric Cherenkov light flux analysis is presented. The method is applied for the solution of two of the main problems in astroparticle physics: ground based gamma ray astronomy, selection of events initiated by primary gamma quanta and the energy and mass composition estimation of primary cosmic ray in the region around the “knee”. The lateral distribution of atmospheric Cherenkov light flux in extensive air showers initiated by primary proton, Helium, Oxygen and Iron nuclei with energies in the range from 1013 eV to 1017 eV were obtained with the help of the CORSIKA 5.62 code, using VENUS and GHEISHA hadronic interaction models for the Chacaltaya observation level of 536 g/sm2. The lateral distribution of Cherenkov light flux in extensive air showers is approximated using a nonlinear fit such as Breit-Wigner. A detailed study of the energy dependence of the proposed model function parameters is carried out and the fit of model parameters as a function of the energy is obtained as well. On the basis of the difference between the model parameters, precisely their behavior as a function of the energy, the strong nonlinearity of the model, the authors propose a method, which permits the making of the distinction between a primary gamma quanta from a primary nuclei. The efficiency of the method is estimated and studied. An additional analysis for primary nuclei is carried out, towards the development of a similar method for simultaneous energy and mass composition estimation of simplified cosmic ray spectra of protons, iron, helium and oxygen. Different detector displacements are analyzed using the simulation of simplified primary mass composition. The detector response is simulated taking into account the physical fluctuations of the processes, the statistical and possible systematic errors. The simulated and reconstructed events are compared and the accuracy in energy and primary mass estimations

xxii

A. L. Zenfora

is obtained. Moreover, the accuracy in shower axis localization is studied and the corresponding criteria are proposed. On the basis of the obtained approximation of the lateral distribution of Cherenkov light, a fast Monte Carlo simulation of the response of a different detector displacement is carried out. The possible triggers for two different detector arrays are studied and the registration efficiency is estimated. Chapter 26 - The paper presents the development of a sub-channel thermal hydraulic analysis code named SUBCHAN. The code was originally developed to analyze a super critical CANDU type reactor which has such characteristics as horizontal fuel channels, heavy water moderated, super critical light cooled water, and any type of fuel bundle with or without thorium rods. Thermal-hydraulic model of SUBCHAN is based on four partial differential equations that describe the conservation of mass, energy and momentum vector in axial and lateral directions for the water liquid/vapor mixture. The heat transfer correlations and pressure drop correlations used in the SUBCHAN code are presented in this paper. The water properties package of the code is based on the Industrial Formulation 1997 for the Thermodynamic Properties of Water and Steam. The heat transfer correlation of super critical region is based on the experimental investigation of Xi'an Jiaotong University. By calculating the TACR case, which is operating at 12.5MPa pressure, compared with the results of ASSERT-PV code, the paper arrives at the conclusion that the development of the SUBCHAN code with super critical water property package is successful. Then the paper uses the SUBCHAN code to analyze CANDU-SCWR operating at 25.0 MPa pressure. The paper draws the conclusion that the SUBCHAN code can be used to analyze sub-channel thermal hydraulic analysis of CANDU-SCWR fuel channel. Chapter 27 - Computer codes are widely used for Nuclear Power Plants (NPP) safety analysis within a wide set of purposes including licensing issues, safety improvement programs of existing NPPs, better utilization of nuclear fuel, and higher operational flexibility, for justification of lifetime extensions, development of new emergency operating procedures, analysis of operational events, and development of accident management programmes. A safety key parameter of the evaluation and assessment of NPPs is closely related to the code ability in determining the time-space thermal-hydraulic conditions throughout the reactor coolant system and especially in the core region. In the beginning, the code development took place between the sixties and seventies during which sets of conservative models were used. Furthermore, the latter were also limited due mainly to the restricted computer memory, Central Process Unit (CPU) time, and performances. However, in light of the sustained development in computer technology and computational methods, the potential of computational features has been enlarged accordingly. Nowadays, it has become possible to switch to a new generation of computational tools consisting of coupling advanced computer codes and getting better realistic simulations of complex phenomena and transients that could occur in NPP. These packages include mainly a thermal-hydraulic system and reactor kinetics codes, as well as specific codes for the containment thermal-hydraulics, structural mechanics codes, and more sophisticated Computational Fluid Dynamics (CFD) codes. However, notwithstanding the complexity of these codes and the level of the present scientific knowledge, a computer code cannot be expected to accurately model phenomena that are not yet fully understood by the scientific community. In general, the results of code predictions, specifically when compared with experimental data, often reveal some

Preface

xxiii

discrepancies. These discrepancies could be attributed to several reasons as model deficiencies, approximations in the numerical solution, nodalization effects, imperfect knowledge of boundary and initial conditions. Therefore, it is necessary to investigate the uncertainty of the results and the sensitivity effect of the most effective parameters. The purpose of the present paper is to characterize the present situation as far as the code assessment and uncertainty predictions are concerned. This is achieved through a reevaluation of some typical activities carried out at the University of PISA. These examples concern mainly application of Best Estimate tools for PWR, BWR, VVER1000 and Research nuclear reactors accident analysis. On this basis, requirements and future needs in the field of Best Estimate tools are outlined. Chapter 28 - Most of the studies and experiments on nuclear fusion are currently devoted to the Deuterium-Tritium (DT) fuel cycle, the easiest way to reach ignition. Some of the main technological questions of future DT fusion reactors have been identified previously. Among those, in particular, the radioactive inventory in such reactors is due, besides tritium, to the neutron-induced radioactivity in the reactor structures. The recent stress on safety by the world community has stimulated research on fuel cycles other than the DT cycle, based on ‘advanced’ reactions, such as Deuterium-Helium-3 (DHe). Several studies have addressed the design of DHe reactors: concerning small-size near-term experiments, to begin to explore the possibilities of DHe plasmas, a DT burning plasma experiment at high magnetic field and high plasma densities is particularly compelling. Ignitor is a proposed compact high magnetic field tokamak, aimed at reaching ignition in DT plasmas and at studying them for periods of a few seconds. A design evolution of Ignitor in the direction of a reactor using a DHe fuel cycle has been proposed: a feasibility study of a high-field DHe experiment of larger dimensions and higher fusion power than Ignitor, still based on the core Ignitor technologies, has led to the proposal of the Candor fusion experiment. This paper deals with the radioactive waste issue for fusion reactors, proposing an innovative solution (the “zero-waste” option), which is a clear advantage of fusion power versus fission, in view of its ultimate safety and public acceptance. Even if feasible in theory, a zero-waste option for fusion reactors using the DT fuel cycle will be difficult to obtain. As a further step towards the zero-waste option, the features of fusion reactors based on alternative advanced fuel cycles have been examined, to assess whether that goal could be reached for such devices. Fusion reactors with advanced DHe fuel cycle turn out to have quite outstanding environmental advantages. Activation behaviour of materials after service in a DHe advanced fuel fusion experiment has been investigated. EUROFER, SiC/SiC and V-Cr-Ti materials have shown the possibility of being declassified to non-radioactive material (clearance) after their irradiation in the reactor plasma chamber wall, if a sufficient interim cooling time is allotted. AISI 316L, on the contrary, suffers the presence of Ni and N (alloying elements) and Nb and Mo (impurities). Chapter 29 - A "dynamic" solar power plant (which consists of a solar collector - thermal engine combination) is proposed as an alternative for the more usual photovoltaic cells. Upper bounds for the efficiency of solar thermal power plants operating in the Martian environment are first evaluated. A general thermodynamic approach, first presented here, clearly shows which of the three theories usually quoted in literature gives the exergy of thermal radiation. Recent works reporting accurate upper bounds for the efficiency of thermal radiation energy conversion into work are subsequently used in this chapter. The results refer to thermal

xxiv

A. L. Zenfora

engines powered by direct or diffuse solar radiation on Mars. Diffuse solar radiation is modeled as diluted or multiply scattered thermal radiation. A more elaborated model uses an endoreversible Carnot cycle to describe solar engine operation. Two strategies to collect solar radiation are analyzed: a solar horizontal collector and a solar collector whose tilt and orientation are continuously adjusted to keep the receiving surface perpendicular on Sun rays. Meteorological data measured at Viking Landers (VL) sites are used in computations. Results show that generally the influence of latitude on performance is important. In some situations the meteorological effects compensate the latitudinal effects and the output power is quite similar at both VL1 and VL2 sites. During a winter dust-storm day the maximum output power is much smaller than during autumn. High efficiency thermal engines should be used in combination with solar collectors kept perpendicular to the Sun’s rays. When a horizontal solar collector is considered, the dependence of the maximum output power on optimum solar efficiency seems to be quadratic at both VL1 and VL2 sites. When a collector perpendicular to the Sun’s rays is considered, this dependence is more complicated, but keeps the quadratic feature. No obvious difference exists between power plant performances in the two years of VL2 operation. A solar Stirling engine based on a horizontal selective flat-plate converter is analyzed in the last part of this chapter. All the computations were performed for a solar collection area similar in size with that of Mars Pathfinder’s Sojourner. The solar efficiency at noon is as high as 18 %. The power provided by the engine is as high as 16 W during autumn and winter. These results suggest that under the Martian environment the performance of properly designed solar Stirling engines is comparable with that of PV cell power systems. Chapter 30 - Zirconium has a low neutron capture cross-section and it is used in alloys for internal components of nuclear reactors, the currently named Zircaloy, Zr-Nb, ZIRLO, etc. In Zircaloy-2 and Zircaloy-4, chromium is an important component in order to assure good corrosion performance, and tin is one of the strengthening elements. On the other hand, titanium, in spite of its poor neutron transparency, has sometimes been considered an element, which could substitute zirconium in this kind of alloy. The present experimental study concerns two ternary systems Zr-Cr-X (being the X component Sn or Ti). Published data on phase equilibriums of these systems are very scarce and found only in Russian works. Many contributions to the knowledge of phase equilibriums in ternary and quaternary systems involving zirconium as the principal component were assessed by Ivanov O.S. et al. and published by the Metallurgical Institute of Moscow in the monograph Zirconium Alloys Structures in 1973. Stability domains of phases at different temperatures of those two ternaries were presented, especially as isothermal sections of the equilibrium diagram. The knowledge of transformations through equilibrium diagrams is essential in order to design or improve technological applications, especially in the temperature range where the Zr rich hcp/bcc solid solution reaction is possible. Alloys were prepared by melting the metal components in a non-consumable tungsten electrode arc furnace with a copper crucible under a high purity argon atmosphere. Phase characterizations and determination of their compositions were carried out by metallographic observations and electron microprobe analysis. X-ray diffraction was performed on some samples. The study of the Zr-Cr-Sn system involves alloys with compositions between 0 and 15 at. % Cr and 0 to 15 at. % Sn and heat treatments at temperatures of 860, 900, 960 and 980 ºC.

Preface

xxv

Three alloys of the Zr-Cr-Ti system with 40 at. % Cr and different Zr/Ti ratios and one more, richer in Cr, were elaborated. Specimens were heat treated at 900 and 1100 ºC respectively. Results of equilibrium between the solid solutions and the intermetallic compounds are presented as tie lines and isothermal sections where the phase boundaries are also sketched. Chapter 31 - The International Atomic Energy Agency’s (IAEA) safeguards system has been a cornerstone of U.S. efforts to prevent nuclear weapons proliferation since the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) was adopted in 1970. Safeguards allow IAEA to verify countries’ compliance with the NPT. Since the discovery in 1991 of a clandestine nuclear weapons program in Iraq, IAEA has strengthened its safeguards system. In addition to IAEA’s strengthened safeguards program, there are other U.S. and international efforts that have helped stem the spread of nuclear materials and technology that could be used for nuclear weapons programs. This testimony is based on the U.S. Government Accountability Office’s (GAO’s) report on IAEA safeguards issued in October 2005 (Nuclear Nonproliferation: IAEA Has Strengthened Its Safeguards and Nuclear Security Programs, but Weaknesses Need to Be Addressed, GAO-06-93 [Washington, D.C.: Oct. 7, 2005]). This testimony is also based on previous GAO work related to the Nuclear Suppliers Group—a group of more than 40 countries that have pledged to limit trade in nuclear materials, equipment, and technology to only countries that are engaged in peaceful nuclear activities— and U.S. assistance to Russia and other countries of the former Soviet Union for the destruction, protection, and detection of nuclear material and weapons. Chapter 32 - The kind of swirl coal burners is given. Radial-biased-combustion and centrally-fuel-rich swirl coal combustion technology was developed. In the air and the airparticle test facilities, the single sensor hot-film and the anemometers were used to measure air and air-particle flows in the near-burner region of different swirl burners. Both cold air flow and reacting flow experiments were performed in the industrial 50, 220, 410, 670 and 1025 ton per hour boilers. On an air-particle test facility, the characteristics of the pulverizedcoal concentrator with cone vanes were investigated. The influence of structure parameters, such as run parameters such as swirling vane angle and burner cone angle and length, and run parameters, such as non-swirl secondary air, central air and air supply, and primary air flow type on divergent angles, diameter and length of the central recirculation zone, mixing characteristic of the primary air and the secondary air, in-situ gas temperature and NOx formation near the burner zone, carbon in ash and NOx emission of boilers was determined with the radial-biased-combustion burner. The difference characteristics of gas/particle flow and coal combustion of the centrally-fuel-rich and dual register burners were obtained. The experimental results show that the two new burners simultaneously have the ability of high combustion efficiency, flame stability, low NOx emission and resistance to slagging and high temperature corrosion. The air-surrounding-fuel combustion theory was put forward. Chapter 33 - The introduction to urban areas of the micro-grid system has the following characteristics. (a) The distance between the heat-supply side and the heat-demand side is short, and effective utilization of exhaust heat is possible. (b) It is linked with the load leveling of the existing large-sized electric power facilities. (c) Since a facility suitable for the energy-demand characteristics of a region can be installed, energy efficiency may increase and facility costs may decrease. The micro-grid using a proton exchange membrane type fuel cell (PEM-FC) may greatly reduce environmental impact. However, when connecting an energy system to the micro-grid of a city area and operating, partial load operation occurs

xxvi

A. L. Zenfora

frequently and power generation efficiency falls. And, the electrode material (especially the catalyst material and the proton exchange membrane) of PEM-FC is expensive, and its system is complex. Consequently, it is necessary to connect two or more power generation systems to the micro-grid, and to design optimization of an operation plan for the purpose of maximization of power generation efficiency. Therefore, the methods of an improvement of the efficiency of the power generation system connected to the micro-grid installed into a city area are described. In this chapter, it consists of subjects of three studies on the micro grid. In these studies described in this chapter, the improvement of the subject of the micro grid is tried by combining fuel cell and other power equipment. Section 1 describes "Operation Plan of Micro Grid Using PEM-FC/Diesel Engine Generator Combined System." Section 2 describes "Carbon-Dioxide Emission Characteristic of Micro Grid Using PEMFC/Hydrogenation City gas-Engine Combined System." Section 3 describes "Dynamic Characteristics of Micro Grid Using PEM-FC/Woody Biomass Engine Combined System." Chapter 34 - The energy policy of many Western governments aims to diversify supply and reduce dependence on foreign sources and thus to maximise benefits from internal resources. Undoubtedly, the main strategy underlying this is one that seeks to optimise the use of renewable energy sources (RES). The development of these sources, as well as their market penetration, depends however not only on political will but also on sound management of energy demand in order to rationalise and stabilise energy consumption. In addition to fortifying the guaranteed energy supply, RES represent a potential that cannot be overlooked. This lies in their ability to reduce greenhouse gas emissions and thus to stem the growing trend of global warming, one which has accelerated particularly in recent years and which is due mainly to the use of fossil fuels for producing electricity. The use of RES for the production of electric power brings huge benefits both in terms of environmental protection as well as savings in non-renewable resources. Nevertheless, the very nature of RES raise technical and economic problems that create a considerable gap between their potential capacity and ways to feasibly exploit them. Their many different forms and the ways in which they may be used have to be carefully examined in order to evaluate the costs and other technical and environmental factors involved. The planning and appraisal of sustainable energy projects involve rather complex tasks. This is due to the fact that the decision making process is the closing link in the process of analysing and handling different types of information: environmental, technical, economic and social. Such information can play a strategic role in steering the decision maker towards one choice instead of another. Some of these variables (technical and economic) can be handled fairly easily by numerical models whilst others, particularly ones relating to environmental impacts, may only be adjudicated qualitatively. In many cases therefore, traditional evaluation methods and the chief economic and financial indicators are unable to deal with all the components involved in an environmentally valid energy project. Multicriteria methods provide a flexible tool that is able to handle and bring together a wide range of variables appraised in different ways and thus offer valid assistance to the decision maker in mapping out the problem. Chapter 35 - Lately, the use of gas turbines following the deregulation of the electricity supply industry has become greater quickly. The motivation for modeling the gas turbines and their controllers is determinant to the interpreting of their impacts on distribution systems. The model predictive control (MPC) is used to damp the oscillation when the power distribution system is subjected to a disturbance. MPC is selected because it can explicitly

Preface

xxvii

handle the nonlinearities, and constraints of many variables in a single control formulation. The IEEE 13 node power distribution system is employed to demonstrate the effectiveness of MPC to damp the oscillations of gas turbines. Among fossil fuels, gas is the most quickest, with a growth rate nearly double that of coal and oil. The electricity generation field is the leading market for gas. The natural gas business has a great interaction with the electricity market in terms of fuel consumption and energy conversion. On the other hand, the transmission and distribution activities are very similar with the natural gas transportation through pipelines. The power losses in gas and electric systems are compared. It is also demonstrated that the electricity system results more convenient for longer distances of gas wells from electricity consumption area. Chapter 36 - Combined cooling, heating and power (CCHP) system, as a distributed energy system, can work all the year and provide cooling/hot-water/power in summer, heating/hot-water/power in winter and hot-water/power in other seasons. In CCHP systems, the total energy efficiency increases to over 85%, while the average energy efficiency of conventional fossil fuel fired electricity generation systems is around 40%. The energy efficiency promotion of CCHP systems results in emission reduction compared to the conventional methods of generating heat and electricity separately. And as a distributed energy resource, CCHP systems also increase in the reliability of the energy supply. With the overall development of CCHP systems and related technologies, the utilization of micro CCHP systems in the residential sector is emerged as a growing potential. The article focuses on the micro CCHP systems for single-family applications (around 10 kW) and multi-family or residential district applications (under 200 kW). The status quo of micro CCHP systems is briefly presented and diverse combinations of technologies existing in applications or experimental units are listed through comprehensive literature review. Various technologies available or under development are introduced, such as reciprocating internal combustion engine, micro-turbine, fuel cell, Stirling engine, absorption chiller, adsorption chiller and so on. Afterward, the tendency and issues of micro CCHP systems are discussed. The review shows that micro-CCHP applications are entering into average families as a nextgeneration residential energy supply center. Chapter 37 - The calculations of the several sensitivities such as loss sensitivity, voltage sensitivity, generator constraint shift factor, and area based constraint shift factor become very important in energy management system (EMS) and energy markets. This chapter focuses on the analysis and implementation details of the above-mentioned sensitivities calculations in the practical transmission network and energy markets. The power operator uses them to study and monitor market and system behavior and detect possible problems in the operation. These sensitivities calculations are also used to determine whether the on-line capacity as indicated in the resource plan is located in the right place on the network to serve the forecasted demand. If the congestion or violation exists, the generation scheduling based on the sensitivities calculations can determine whether or not a different allocation of the available resources could resolve the congestion or violation problem. This chapter also comprehensively discusses how to compute and use the sensitivities under the different references such as the market-based reference, and the energy management system based reference. The calculation results of the several sensitivities are illustrated using the IEEE 14 bus system and AREVA T & D 60-bus system. Chapter 38 - Many transient processes in power systems involve phenomena that vary in time and space in complicated ways. Comprehensive monitoring of large-scale power systems

xxviii

A. L. Zenfora

by means of properly placed time-synchronized phasor measurement units (PMUs) provides the opportunity to analyze and characterize complex inter-area swing dynamics involving all or most of the power system. Wide-area real-time monitoring may prove invaluable in power system dynamic studies by giving a quick assessment of the damping and frequency content of dominant system modes after critical contingencies. Measured data, however, may exhibit quite different dynamics at each system location or exhibit abrupt changes, dynamic irregularities, or be complicated by nonlinear trends or noise. Traditional Fourier and Prony methods for system identification are unable to resolve the localized nature of these processes and hence provide little useful information concerning the nature of noisy, time-varying oscillatory processes. In this Chapter, a new method for analyzing the temporal dynamics of nonlinear and nonstationary inter-area oscillations using a local empirical mode decomposition (EMD) method and the Hilbert transform is presented. Two novel algorithms are developed to address nonlinear and non-stationary issues. The first method is a local implementation of the empirical mode decomposition technique. The second is an algorithm to compute the Hilbert transform using finite impulse response (FIR) filters. By combining these approaches, the method can be used to analyze complex signals for which the conventional assumptions of linearity and stationarity may not apply and can be implemented for on-line estimation of modal damping and frequency using synchronized wide-area measurement systems. The physical mechanism underlying nonlinear time-varying inter-area oscillations is investigated and methods to characterize the observed oscillatory phenomena in terms of physically meaningful modal components are proposed. Emphasis is placed on identifying modal content in the presence of noise and nonlinear trends. Issues concerning the implementation of the method and numerical considerations are also discussed. As specific applications, data obtained from PMU measurements from a real event in the northern systems of the Mexican interconnected system are used to examine the potential usefulness of nonlinear time series analysis techniques to characterize the spatio-temporal characteristics of the observed oscillations and to determine the nature and propagation of the system disturbance. The efficiency and accuracy of the method is demonstrated by comparison to other approaches. Chapter 39 - Of the numerous electric power faults an Electric Engineer comes across in a life time, only a few of these faults are memorable- the rest being routine ones. In this chapter, some of those unconventional faults, which are mainly related to power system protection, are presented. The chapter presents five case studies of actual field incidents rather than hypothetical scenarios. The objective of the chapter is to present a typical approach for analyzing the faults in power systems. Chapter 40 - Temperature profiles have been empirically investigated in the underground geological formations of the Nea Kessani (Greece) geothermal system by the Greek Institute of Geology and Mineral Exploration using measurements in a set of vertical drill holes. In this work, we used the BME method to derive spatial temperature estimates in the Nea Kessani region in a mathematically rigorous and scientifically meaningful manner. The proposed analysis involves the solution of a stochastic partial differential equation representing the geothermal field and is conditioned on site-specific information (random boundary conditions reflecting in-situ uncertainties etc.). Temperature probability distributions were generated at the nodes of a dense spatial grid, which can provide a detailed understanding of the geothermal situation by means of various temperature maps (most probable, error minimizing

Preface

xxix

etc.), depending on the objectives of the study. The BME solutions are more informative than the direct (analytical and numerical) solutions of the geothermal model obtained in a formal mathematical sense. Chapter 41 - This chapter introduces briefly distribution of hydrothermal resources, potential of hydrothermal energy, geochemistry of geothermal fluids and correlation between geothermal areas and seismic zones in China. More than 3,200 hydrothermal manifestations have been found in China. About 2,240 drilled wells reveal that 275 high temperature sites of hydrothermal energy, which are expected to supply a need of electric generators with total annual output of 5,800 MW. More than 2,900 sites of low and intermediate temperature geothermal systems have been found, which can be utilized for heating, medicine treating, bathing, farming, etc. Most geothermal waters in China are (Na, Ca)-HCO3 type, and some are (Ca, Na)-SO4 and Na-Cl types. Stable isotopic compositions of oxygen and hydrogen indicate the geothermal waters are derived from meteoric water, with small amount of magmatic volatile. Reservoir temperatures calculated with chemical geothermometers range from about 100 °C to 350 °C. Geochemical variations of geothermal fluids with time are found, which are correlated to hydrothermal eruption, earthquakes and exploitation. Main gaseous components of geothermal systems in China are CO2, N2, O2, and trace amount of H2S, H2, CH4, NH3, CO, C2H6, C3H8 as well as noble gases (Rn, He, Ar, Ne, Kr, Xe). The gaseous concentrations of geothermal systems are correlated to the temperatures of geothermal systems and seismic faults. The gases have a multiple origins of crust, mantle and atmosphere. The chapter emphasizes on both the spatial correlation between the geothermal areas/zones and the seismic zones and the energetic relationship between geothermal-fluid geochemistry and seismic activity. The more amounts of mantle gases the geothermal systems contain, the higher temperature of geothermal systems and the more active the seismic zones. The deep earth fluids provide both matter and energy for geothermal fields and earthquake generation, and carry the messages of geothermal reservoir and earthquake. Chapter 42 - The Larderello and Mt Amiata geothermal fields in Tuscany are large active thermal systems. Both likely overlie young plutonic rocks that serve as the principal sources of heat. The features of the two geothermal systems are similar. 1 Structural setting. The geothermal fields of Larderello and Mt. Amiata are located in the inner part of the Northern Apennines, characterized asthenosphere uplift and delamination of the crustal lithosphere or underplating. 2 The heat source both at Larderello and Mt. Amiata can be ascribed to the presence of shallow igneous intrusions. 3 The heat flow data for the area surrounding both the Larderello and Mt. Amiata geothermal fields show a comparable areal extension and similar values (up to 200-300 mW/m2 ). 4 Cap rocks and reservoirs. Both the Larderello and the Mt. Amiata fields have shallow vapor-dominated sedimentary, and deep metamorphic reservoirs. At Larderello super-heated steam is present in both reservoirs, to depth of more than 3.5 km, whereas the deep reservoir of the Mt. Amiata geothermal fields is likely water-dominated. In both fields the upper reservoir is present below the flysch units forming the cap rocks. 5 Permeability is due to rock fracturing, even at depths of about 4 km and temperatures as high as 350°C. Pressure greater than hydrostatic and a supercritical fluid can occur in the deepest part of the geothermal fields. 6 Hydrothermal alteration and contact metamorphism. At Larderello and Mt. Amiata there is evidence of an early contact metamorphism related with

xxx

A. L. Zenfora

the intrusion of the granites. 6 Recharge. The water stable isotope values of the steam discharged by the geothermal wells at Larderello indicate a meteoric origin. A geochemical regional study on the thermal waters and gases of the Mt. Amiata area indicates that the geothermal reservoirs originated from a meteoric fluid, mainly stored in a regional Mesozoic dolomite-anhydrite unit, and evolved in a Na-Cl, CO2 gas-reach reservoir by interaction with calcite-bearing metamorphic rocks. The high temperatures existing in correspondence of a deep seismic reflector suggest the occurrence of a deep-seated unconventional geothermal resource (UGR), which can be possibly exploited. The heat could be mined from silica-rich rocks close to a plastic state, but where fracturing can be induced by fluid overpressure and abrupt high strain rates. This geothermal resource is very important, requires a re-assessment of the geothermal resources in Italy, considering the possibility of the exploitation of the new reservoirs. Chapter 43 - High-temperature geothermal reservoir is under consideration, consisting of two high-permeability layers, which are separated by a low-permeability stratum. The thermodynamic conditions are assumed to imply that the upper and lower high-permeability layers are filled in by water or by vapor, respectively. The stable stationary regimes of vertical phase flow between water and vapor layers in the low-permeability stratum may exist. The authors give possible types of transition to instability of the vertical flows in such a system under the condition of smallness of the advective heat transfer in comparison with the conductive one. It is found that in the generic case there exist three different scenarios of the instability onset of the stationary vertical flows. They are accompanied by the bifurcations of solutions describing the destabilizing vertical flows. The possible scenarios of the evolution of the system over the threshold of instability are discussed. Chapter 44 - Abundant genetic and sedimentary indicators has been found in the thick coal beds from three fault-controlled coal basins on the Central Massif France. A new formation model for thick continental (intra-mountainous) lacustrine peat swap is proposed. In the new coal accumulation mechanism, thick coal beds were associated with various gravity-influenced breccia and sandstone interlayer sediments and the subaquatic gravitary current transported the organic (peat) and inorganic clasts formed in lakeshore swamp were formed in active clastic environment, and were associated with various gravity-influenced mudstone and sandstone interlayers. The presence of a great number of gravity-flow sediments such as detrital flow, diluted slurry flow or turbidity-current sediments in the coal seams, and that of the contemporaneous gravity slump and deformation structures in the coal seams both indicate that the accumulation of the thick coal beds was characterized by the relatively deep water environment and allochthonous sedimentation. This new model interprets reasonably the accumulation mechanism of the thick coal beds developed in the fault basins in the Central Massif (France) and provides a completely new idea with respect to the traditional coal formation models. Chapter 45 - As an innovation, the authors propose that a new frontier in geothermal research should be explored that involves the coupling of thermal and chemical simulations in an integrated “magma chamber-reservoir” model. To achieve this innovation in geothermal research, the authors have written a new computer program (in Fortran 90) in modular structure that runs on a PC under the dynamic memory concept and simulates heat transfer conductive and convective processes both in a magma chamber and the overlying geothermal reservoir as well as computes in-situ major element chemistry of magmas that evolved in the magma chamber as a result of processes of assimilation, crystallization (liquid line of

Preface

xxxi

descent), magma mixing, recharge, and eruption. This combined task is accomplished in three dimensions (3-D) – a substantial improvement as compared to the current practice of obtaining thermal solutions in 1-D or 2-D and of modeling chemical data obtained from the analysis of surface outcrops without reference to the actual location within the Earth where the magmas were stored prior to eruption. In fact, if temperature estimates in drill wells and chemical data for surface rocks were available, this information can be used to constraint the model the authors are proposing as a new research frontier. The practice of “direct” modeling can be replaced in the future by inverse modeling when greater computing and storage capacities of personal computers will be available. This chapter briefly reviews the current state of thermal modeling of geothermal areas and presents the salient features of our new research approach, including a brief description of our computer program. An application example of a Mexican geothermal field (Los Humeros, Puebla), currently under exploitation for electricity production, will highlight the use of our software. This particular geothermal field was chosen for illustration purposes because of the availability of required thermal and chemical data to test the 3-D simulation model. The authors have successfully reproduced some of the major element chemical characteristics for the most voluminous caldera-forming eruption at about 0.46 Ma and the present-day thermal regime inferred from static formation temperatures using a quadratic regression of the actually measured bottom hole temperature data. Expert Commentary A - In the last decade, nuclear energy has gained a widespread renewal of interest as an important contributor to energy security, supply and sustainability. A number of new designs of nuclear power plants (NPP) has emerged recently, in attempts to achieve advances in the following areas: sustainability, competitive economics, safety and reliability, proliferation-resistance and physical protection. Actually, in the framework of the Generation IV International Forum (GIF), a task force has announced in 2002 the selection of six reactor technologies, which would represent the future shape of nuclear fission energy: these reactors operate at higher temperatures than today's reactors, allowing new and attractive applications, such as the thermo-chemical production of hydrogen. In addition to these six concepts for deployment between 2010 and 2030, the GIF has recognised a number of International Near-Term Deployment advanced NPPs available before 2015. Moreover, several international research projects are ongoing, which concern subcritical AcceleratorDriven Systems for radioactive wastes incineration, in conjunction with Partitioning and Transmutation technologies. Expert Commentary B - Large sparse power systems form an extremely complex dynamical system which usually possess many degrees of freedom and poses a challenge for simulation and analysis. Forced complex oscillations triggered by the loss of major system resources may manifest highly complex spatial and temporal dynamics and involve a large number of machines and take place over a great range of time and time scales. Proper understanding of the underlying dynamics causing these oscillations requires investigation of the various types of temporal nonlinear interactions involving the fundamental modes of the system. Such features may be obscured or distorted in the normal spectral analysis approach. The analysis of spatio-temporal dynamic patterns is important for many reasons. Nonlinearity causes the fundamental waves or temporal modes to interact, leading to frequency and amplitude modulation and to a phase relationship known as quadratic phase coupling between the frequency components involved. Mounting evidence suggest that these interactions can have a significant impact on system performance such as the modal content

xxxii

A. L. Zenfora

of the observed oscillations and may the design of controllers. Further, it is also possible that nonlinearity contributes to non-stationary behavior in the record.

RESEARCH AND REVIEW STUDIES

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 3-51

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 1

ENERGY MARKETS* United States Government Accountability Office WHY GAO DID THIS STUDY To better understand how changes in domestic and international petroleum products markets have affected prices, GAO was asked to evaluate trends in (1) the international trade of petroleum products, (2) refining capacity and intensity of refining capacity use internationally and in the United States, (3) international and domestic crude oil and petroleum product inventories, and (4) domestic petroleum supply infrastructure. To address these objectives, we reviewed numerous studies, evaluated data, and spoke to many industry officials and experts and agency officials.

WHAT GAO RECOMMENDS GAO is making recommendations aimed at improving the functioning of petroleum product markets, including that the Secretaries of Transportation and Energy coordinate with other agencies to (1) encourage more uniform biofuel and petroleum product blending practices, (2) conduct a study of infrastructure system adequacy, and (3) evaluate the assignment of a lead agency to coordinate permitting of infrastructure construction. In commenting on the report, the Federal Energy Regulatory Commission generally agreed with the report’s findings and recommendations, while the Departments of Energy and Transportation neither fully agreed nor disagreed. To view the full product, including the scope and methodology, click on GAO-08-14. For more information, contact Mark Gaffigan at (202) 512-3841, [email protected].

*

This is an edited, reformatted and augmented version of GAO Report GAO-08-14, dated December 2007 publication.

4

United States Government Accountability Office

WHAT GAO FOUND International trade in petroleum products has expanded over the past two decades, making markets for gasoline and other petroleum products increasingly global in nature. Recent plans and mandates in the United States and other countries to greatly expand the use of biofuels blended with petroleum products—for example, ethanol blended with gasoline and biodiesel blended with petroleum diesel—may have the unintended effect of reducing opportunities for trade because blending different levels of biofuels with petroleum blending stocks will require changes to these blending stocks and thereby reduce their fungibility. For most of the past 25 years, there has been excess refining capacity globally, but this excess has shrunk considerably in recent years as demand has increased faster than capacity growth, causing refineries to run closer to their production capacity, and contributing to recent increases in petroleum product prices, price volatility, and refining profits. However, experts say it is unclear whether or for how long the current market tightness will continue, in part because of uncertainties about how much additional refining capacity will actually be built in the face of rising construction costs and initiatives that may reduce future demand for petroleum products such as through the blending of large volumes of biofuels into the transportation fuels markets. When measured as average days of consumption, inventories of petroleum products and crude oil in the United States indicate a general decline over the past 20 years. A number of factors have contributed to this decrease in the United States, including reductions in crude oil production and the number of refineries as well as efforts to reduce inventory holding costs by applying advances in technology. Lower operating costs associated with lower inventories may have translated into lower consumer prices during normal periods. However, lower than normal inventories can lead to higher or more volatile prices in the event of supply disruptions or surges in demand. The nation’s petroleum product supply infrastructure is constrained in key areas and is likely to become increasingly constrained, unless timely investments are made. A constrained supply infrastructure can exacerbate price effects and price volatility due to a supply disruption. However, no central source of data tracks system bottlenecks. While there is widespread recognition that a study is needed to fully identify the extent of infrastructure inadequacy and the impact on prices, to date, no such analysis has been undertaken, though such a study was mandated by Congress in 2006 with a June 2008 deadline. Significant infrastructure expansion plans in the private sector could alleviate the stresses. However, a complex permitting and siting process involving as many as 11 federal agencies and numerous state and local stakeholders has slowed or impeded the expansion and construction of new pipelines. Unlike in the case of natural gas pipelines, no central federal agency acts to coordinate this permitting process.

ABBREVIATIONS BP DOE DOT EIA

British Petroleum Department of Energy Department of Transportation Energy Information Administration

Energy Markets EPA FERC FTC IEA MARAD NYMEX OECD SPR

5

Environmental Protection Agency Federal Energy Regulatory Commission Federal Trade Commission International Energy Agency U.S. Maritime Administration New York Mercantile Exchange Organisation for Economic Co-operation and Development U.S. Strategic Petroleum Reserve

December 20, 2007 The Honorable Daniel Inouye Chairman, Committee on Commerce, Science, and Transportation United States Senate The Honorable Maria Cantwell United States Senate In 2003, the price of West Texas Intermediate crude oil, a widely watched benchmark crude oil price, averaged about $31 per barrel. By 2006, the average was about $66 per barrel, and in mid-November, 2007 the price rose to over $90 per barrel. Wholesale and retail prices of petroleum products refined from crude oil, including gasoline, diesel, and jet fuel, which normally rise and fall with crude oil prices, also generally rose over the period. For example, U.S. retail regular gasoline prices—equivalent to wholesale prices plus taxes, marketing costs, and retail profit margins—averaged $1.52 per gallon in 2003, but by August 2006, they had almost doubled to $3.00 per gallon, and as of July 2007, remained relatively high at $2.85 per gallon. Such large and sustained increases in gasoline prices have not been seen in the United States since the late 1970s and early 1980s, when the start of the Iran-Iraq war pushed prices up—even higher than today’s prices when adjusted for inflation—causing severe economic hardship for many Americans and contributing to a global economic recession. While this more recent increase in petroleum product prices does not appear to have had such far-reaching economic effects, consumers want to know the reasons for the large and relatively sudden price increases. Figure 1 shows retail regular gasoline prices in the United States, in both nominal and inflation-adjusted terms during the past 30 years. In addition to crude oil prices, a number of factors affect the price of petroleum products. As we recently testified before Congress, these factors include domestic capacity to refine crude oil into petroleum products; inventories of these products; the proliferation of special blends of gasoline; the capacity and functioning of the crude oil and petroleum product supply infrastructure, which is composed of pipelines, barges, tanker vessels, marine terminals, rail, trucking and storage tanks; and mergers in the oil industry.[1] In addition, because the United States imports and exports petroleum products, events outside the United States can affect domestic petroleum product prices. Imports to or exports from the United States typically enter or leave through port facilities on tankers or across national borders via pipeline. Our imports of petroleum products come from all over the world into ports in the Gulf of Mexico and the east and west coasts, and by pipeline from Canada. Refineries process crude oil into petroleum products through a variety of complicated processes, and a single barrel of crude oil produces a varying amount of gasoline, diesel, jet fuel, and other products depending on the configuration––or complexity––of the refinery as well as the type of crude oil being refined. Refineries can be optimized—or “upgraded”––to

6

United States Government Accountability Office

process different grades of crude oil through the addition of specialized refining equipment. U.S. refineries are generally optimized to produce large proportions of gasoline to meet domestic transportation demand. Cleaner-burning fuels have proliferated in response to legislation including the Clean Air Act Amendments of 1990, leading to additional investments in the refining equipment needed to produce the new fuels.

Source: GAO analysis of EIA data. Figure 1. U.S. Retail Regular Unleaded Gasoline Prices, Annual Average, 1976 – 2006.

More recently, a number of European countries, the U.S. federal government, and a number of individual states and localities have proposed or mandated the use of biofuels— such as ethanol made from corn or biodiesel made from soybeans or other crops—partly in an effort to reduce greenhouse gas emissions and reduce consumption of petroleum products. These mandates call for biofuels to be blended in varying proportions with traditional gasoline or diesel. For example, U.S. federal biofuel standards call for a minimum proportion and volume of biofuels to be sold each year but do not specify how that proportion is met. In addition, a number of states and at least one city have requirements or plans to require use of biofuels in varying proportions, blended with gasoline and diesel. For example: •

•

•

•

Hawaii, Minnesota, and the city of Portland, Oregon, all currently require ethanol to be blended at a 10 percent by volume rate with gasoline, although Hawaii only requires this for 85 percent of the gasoline sold in the state. Minnesota and Portland, Oregon require 2 and 5 percent biodiesel, respectively, to be blended with diesel fuel. Minnesota also requires the expansion of ethanol blending to 20 percent by volume by 2013. Four other states—Missouri, Montana, New Mexico, and Oregon—have biofuel mandates that will require 10 percent ethanol blended into gasoline and/or varying blends of biodiesel: Missouri and Montana have no mandated plans for biodiesel; New Mexico calls for 5 percent biodiesel blending and Oregon for 2 percent. Other states have “flexible standards.” For example, Iowa provides tax credits if at least 10 percent of the fuels used by 2009 are renewable, with the threshold rising to 23 percent in 2018. Yet this can be achieved in a flexible way, using a blend consisting of 85 percent ethanol and 15 percent gasoline, while other gasoline would

Energy Markets

7

be blended with less or no ethanol at all. Louisiana will require both ethanol and biodiesel to be blended at 2 percent, but only when state production reaches certain levels and prices of ethanol and biodiesel are sufficiently low. Finally, Washington will require that at least 2 percent of diesel sold be biodiesel by November 30, 2008, or when a determination is made that state biodiesel production can meet the 2 percent requirement. Automakers and refiners told us that these varying biofuel blends will require changes to the gasoline and diesel blendstocks––the fuels that will be mixed with the ethanol or biodiesel—to maintain engine performance and emissions requirements. The production of these new blends may also require further refinery changes as well as changes to automobile engines. Automakers also told us that in addition to increasing the costs of production, changing engines to be able to meet performance and emissions standards using a wide mix of biofuel blends would also entail potential losses in fuel efficiency. From refineries, petroleum products are distributed through an extensive supply infrastructure composed of pipelines, barges, tanker vessels, marine terminals, rail roads, trucks, and storage tanks. Pipelines are generally the cheapest domestic mode for transporting crude oil and petroleum products. Crude oil and petroleum products are transported in separate pipelines, and while different types and specifications of petroleum products are shipped in the same pipelines, they must be kept separate during transport and storage in order to maintain the specific desirable performance and emissions characteristics of these different fuels. Crude oil pipelines connect several large refining centers to crude oil sources, and petroleum product pipelines connect these refineries to population centers all over the country. Trucks and rail have generally distributed only a small fraction of petroleum products to wholesale terminals. However, they are being increasingly utilized to move ethanol to locations near final demand centers where the ethanol is blended with gasoline. This is because existing pipelines cannot currently accommodate ethanol due to an insufficient collector pipeline network linking ethanol refineries with major pipelines, and because ethanol has corrosive and other properties that complicate its transport in pipelines that also carry petroleum products. Refiners, distributors, and marketers of petroleum products maintain inventories of crude oil and petroleum products to facilitate smooth supply operations and mitigate the effects of supply disruptions. Crude oil and petroleum product inventories consist of three levels. Primary inventories comprise the crude oil or petroleum products held at production sites, refineries, and storage terminals, and in pipelines, tankers, barges, and other transportation centers. Secondary inventories consist of retail outlets and small storage facilities—those with less than 50,000 barrels of total capacity––that exist between the primary distribution system and the end user. Tertiary inventories are the petroleum products in the hands of end users, for example, in drivers’ gasoline tanks. The federal government also maintains strategic stocks of crude oil and, in the Northeast, heating oil to be released in the event of a major supply shortage. The Energy Information Administration (EIA) collects inventory data for the primary system. Information about changes in inventory levels can inform market participants about underlying demand or supply conditions that will influence prices. A number of federal agencies have programs and activities related to the oversight or monitoring of the refining, distribution, or importing of petroleum and petroleum products. For example, the Department of Transportation (DOT) oversees crude oil and petroleum

8

United States Government Accountability Office

product pipelines and monitors their operations to ensure public safety. The Federal Energy Regulatory Commission (FERC) regulates the “tariffs”—or rates and conditions—under which interstate crude oil and petroleum product pipelines operate, while individual states have oversight over intrastate pipelines within their borders. EIA collects data from refiners and others about shipments of crude oil and petroleum products by pipeline and barge between regions of the United States. In addition, a number of federal and state agencies and other local and private entities become involved in approving new supply infrastructure projects. For example, the approval to build or repair a pipeline could involve DOT’s Pipeline and Hazardous Materials Safety Administration, the Environmental Protection Agency, Bureau of Land Management (if pipelines cross federal lands), Army Corps of Engineers, U.S. Fish and Wildlife Service, as well as other federal agencies, and state and local stakeholders. Legislation in 2002 mandated the formation of an interagency committee to help expedite pipeline review and permitting processes for pipeline repairs. That committee is composed of 11 federal entities. For construction of interstate natural gas pipelines, the Federal Energy Regulatory Commission takes a lead role in coordinating the permitting process across the relevant federal agencies and can convey the right of eminent domain to builders of natural gas pipelines to resolve disputes with owners of land needed to build a pipeline.[2] However, no such federal coordinating authority or power of eminent domain exists for construction and expansion of new interstate petroleum product or crude oil pipelines. To better understand changes in domestic and international markets for petroleum products and the implications of these changes for recent price increases, you asked us to evaluate trends and effects on petroleum product prices in (1) international trade of petroleum products; (2) refining capacity and intensity of refining capacity use internationally and in the United States; (3) international and domestic crude oil and petroleum product inventories; and (4) domestic crude oil and petroleum product supply infrastructure, particularly pipelines and marine transportation. To evaluate trends in the international trade of crude oil and petroleum products, we analyzed data from EIA and the International Energy Association (IEA) and spoke with numerous government agency and oil company officials and industry experts. To assess trends in refining capacity, we evaluated IEA, EIA, and Oil and Gas Journal data, and spoke with numerous industry experts. To evaluate trends in inventories of crude oil and petroleum products, we reviewed data from EIA and IEA on inventories and demand to analyze international, U.S. national, and regional inventories. We analyzed New York Mercantile Exchange (NYMEX) and other futures market data, as well as EIA data, and asked experts about the effects of these futures prices for crude oil on inventory holding decisions. To evaluate trends in supply infrastructure for crude oil and petroleum products, we collected and analyzed available data on the pipeline and marine infrastructure system, capacity, throughputs, and constraints. We examined reports and data from supply disruption case studies to examine those cases’ impact on prices and price volatility. We spoke with numerous government agency and pipeline company officials and industry experts. This report focuses on long-term trends in the industry, rather than recent events that have influenced prices of gasoline and other petroleum products. GAO currently has ongoing work looking at such recent trends as refinery outages through the spring of 2007 and mergers in the industry since 2000.

Energy Markets

9

This report uses data from domestic and international wholesale petroleum products and crude oil markets and domestic retail markets. In contrast to retail prices, wholesale prices do not include taxes, distribution and marketing expenses, and profits. In every case for the data used in this report, we assessed and determined that the data were sufficiently reliable for our purposes. A more detailed description of the scope and methodology of our review is presented in appendix I. We performed our work from August 2006 through September 2007 in accordance with generally accepted government auditing standards.

RESULTS IN BRIEF International trade in petroleum products has expanded significantly over the past two decades, making markets for gasoline and other petroleum products increasingly global in nature. This trend has been particularly important for the United States; while in 1970 the United States was largely self-sufficient in gasoline, we now import over 10 percent of our annual gasoline consumption. Having access to more sources of supply can benefit the United States in the event of domestic supply disruptions. For example, the benefit of such flexibility in sources of supply helped U.S. marketers and retail sellers obtain gasoline and other petroleum products in the aftermath of Hurricanes Katrina and Rita, when imports of gasoline to the United States increased to fill the void left by damaged or shut-down domestic refineries and pipelines. However, the fact that petroleum product markets are international means that supply disruptions or unexpected increases in demand anywhere in the world can influence U.S. prices. Our analysis of wholesale prices in the United States, Europe, and Asia shows that prices in geographically dispersed markets rose significantly following Hurricanes Katrina and Rita, indicating that prices in these markets are linked to some extent. We also evaluated petroleum product import data and found that products came from a wider range of countries during this period, again indicating that products move in response to price signals globally. Recent plans and mandates in the United States and other countries to greatly expand the use of biofuels blended with petroleum products—for example, ethanol blended with gasoline and biodiesel blended with petroleum diesel—may have the unintended effect of reducing opportunities for trade because blending different levels of biofuels with petroleum blending stocks will require changes to these blending stocks and thereby reduce their fungibility. For example, if European countries adopt widely different blending levels biofuels in gasoline and diesel products as current plans call for, the refineries serving these countries will have to alter petroleum blending stocks for those blending levels, and this could make the blending stocks themselves less tradable across countries. For most of the past 25 years, there has been excess refining capacity globally, but this excess capacity has shrunk considerably in recent years as demand has increased faster than capacity growth, causing refineries to run closer to their production capacity, and, along with rising crude oil prices, contributing to recent increases in petroleum product prices and price volatility. Demand for petroleum products has grown more quickly than has refinery capacity for much of the past 25 years, in large part because excess refining capacity historically caused profitability of the refining sector to be low compared to that of many other industries. More recently, this tightening of the balance between supply and demand for petroleum products has, along with higher crude oil prices and other factors, contributed to increased petroleum product prices and higher oil industry profits, and could contribute to greater price

10

United States Government Accountability Office

volatility. Recently high petroleum product prices and increased profits over those seen during the 1990s in the refining industry have spurred new refinery capacity investments in the United States and internationally. However, experts say it is unclear whether or for how long the current market tightness will continue, in part because of uncertainties about how much additional refining capacity will actually be built in the face of rising construction costs, and initiatives that may reduce future demand for petroleum products such as through the blending of large volumes of biofuels into the transportation fuels markets in many countries. The absence of national standards for blending biofuels with gasoline and diesel could also increase the number of gasoline and diesel blending stocks refiners have to make, which could require additional refining investment to make those blends that could crowd out investment in refining capacity expansions. When measured as average days of consumption, long-term trends in inventories of petroleum products and crude oil in the United States indicate a general decline over the past 20 years. Similarly, gasoline and crude oil inventories in the Organisation for Economic Cooperation and Development (OECD) countries, excluding the United States, have also generally fallen over the same period.[3] Inventories, as measured by EIA IEA, and others, have some limitations as a measure of what is available to meet demand in the event of a supply shortfall, in part because the United States has imported an increasing share of its gasoline over the period during which inventories have fallen, and as such, the domestic inventory data do not account for large volumes of these products on the water or in tankers from foreign sources that are destined for the U.S. market or in storage terminals at foreign ports serving this trade in gasoline. A number of factors have contributed to the long-term decrease in inventory holdings in the United States, including reductions in both domestic crude oil production and the number of refineries. Advances in technology and changes in management processes also may have contributed to reduced inventories by enabling refiners to more closely time the production of supplies to meet expected demand. Lower operating costs associated with lower inventories may have translated into lower consumer prices during normal periods. However, in the short term, because inventories provide a smoothing effect against temporary demand and supply fluctuations, lower than normal inventories can lead to higher or more volatile prices in the event of supply disruptions or surges in demand. The nation’s petroleum product supply infrastructure is constrained in key areas and is likely to become increasingly constrained, unless timely investments are made. Industry and federal agency officials report a systemic lack of pipeline capacity in the supply infrastructure system in key states including Arizona, California, Colorado, and Nevada, and note the existing supply infrastructure is insufficient to carry the commensurate volumes of petroleum products and crude oil needed to meet growing demand there. A constrained supply infrastructure can exacerbate price effects and price volatility due to a supply disruption. For example, during a pipeline outage in 2003 that affected pipeline supplies to Arizona, retail prices of gasoline rose by about 45 cents per gallon. However, we were unable to assess the full extent of supply infrastructure constraints or the impacts of these constraints on prices and price volatility, in large part because there is no central source of data that tracks system bottlenecks. In 2006, DOT put forth a legislative proposal and Congress passed legislation that mandated the Secretaries of Energy and Transportation to conduct periodic analyses of (1) where unplanned petroleum product pipeline outages or insufficient pipeline capacity increase prices and (2) whether or not regulation is adequate to minimize the potential for unplanned losses of pipeline capacity. While there is widespread recognition that such a study

Energy Markets

11

is needed to fully identify the extent of infrastructure inadequacy and the impact on prices, to date, no such analysis has been undertaken. DOT and Department of Energy (DOE) officials told us that they were not allocated funds specifically to do the mandated analyses and that the agencies have not re-allocated other funds for this, although DOE told us it has met with DOT to discuss how this work could be approached. However, given that the study has not begun, it seems highly unlikely that agencies will be able to meet their June 2008 deadline for reporting to Congress. There are many private sector plans to expand the supply infrastructure, and if implemented in a timely fashion, these plans could significantly alleviate the stresses on the system. However, a complex permitting and siting process involving as many as 11 federal agencies and numerous state and local stakeholders has slowed or impeded the expansion and construction of new pipelines. The permitting process for building natural gas pipeline infrastructure has been made easier by the designation of FERC as a lead federal agency to streamline permitting for interstate natural gas pipeline expansion, but no such lead federal agency exists to facilitate permitting of crude oil or petroleum product pipeline construction or upgrading. GAO is making recommendations aimed at improving the functioning of petroleum product markets, including that the Secretaries of Transportation and Energy coordinate with other relevant agencies to (1) encourage uniform biofuel and petroleum product blending practices, (2) conduct a study of infrastructure system adequacy, and (3) evaluate the feasibility of assigning a lead federal agency to coordinate the permitting of infrastructure construction. In commenting on the report, the Federal Energy Regulatory Commission generally agreed with the report’s findings and recommendations, while the Departments of Energy and Transportation neither fully agreed nor disagreed.

BACKGROUND The United States is the largest consumer of crude oil and petroleum products of all nations, though demand for crude oil is growing faster globally, led by growth in developing countries such as China and India. When processed, crude oil is refined to produce petroleum products such as gasoline, diesel, and jet fuel, which have been instrumental in providing the nation with affordable fuel for automobiles, trucks, airplanes, and other forms of transportation and heating in some parts of the country. The petroleum industry consists of three main segments: the exploration and production segment (upstream); the refining and marketing segment (downstream); and a third segment typically referred to as the midstream, which consists of the infrastructure used to transport crude oil and petroleum products. Several U.S. agencies regulate and monitor the downstream and midstream oil industry and petroleum product markets.

12

United States Government Accountability Office

THE UNITED STATES IS THE LARGEST CONSUMER OF CRUDE OIL AND PETROLEUM PRODUCTS, BUT GLOBAL DEMAND HAS GROWN SIGNIFICANTLY IN RECENT YEARS While the United States is the largest consumer of crude oil and petroleum products, global demand for crude oil and petroleum products is growing at a faster pace than U.S. demand, driven by growing consumption of crude oil and certain petroleum products in developing countries such as China and India. In 2006, the United States’ share of world oil consumption was approximately 25 percent. The EIA projects in its reference, or “baseline,” scenario that world oil consumption will continue to grow and will reach 118 million barrels per day in 2030. About 43 percent of this growth will come from non-OECD countries, particularly China and India, but the United States will remain the world’s largest consumer. Under the assumptions of EIA’s reference case scenario, U.S. demand for oil is projected to increase by 30 percent between 2005 and 2030—from about 21 million barrels per day in 2005 to about 27 million barrels per day in 2030—compared to 39 percent for the entire world. Meanwhile, domestic production of oil has generally been in decline for decades, leading to greater reliance on imported oil. In 2006, the United States imported about 66 percent of its crude oil.

Source: GAO analysis of BP Statistical Review of World Energy June 2007. Notes: Other light distillates consists of aviation gasoline and light distillate feedstock (LDF) Other middle distillates consists of jet and heating kerosene, and gas and diesel oils (including marine bunkers) Fuel oil includes marine bunkers and crude oil used directly as fuel Others consists of refinery gas, liquid petroleum gas, solvents, petroleum coke, lubricants, bitumen, wax, and other petroleum products and refinery fuel and loss

Figure 2. U.S. Consumption of Petroleum Products, 1965-2006.

Energy Markets

13

When processed, crude oil produces petroleum products such as gasoline, diesel, and jet fuel, which have been instrumental in providing the nation with affordable fuel for automobiles, trucks, airplanes, and other forms of transportation, and—in some parts of the country—heating. Overall, demand for petroleum products in the United States has generally increased over the last 25 years, as shown in figure 2.

KEY ASPECTS OF THE PETROLEUM PRODUCT MARKETS: REFINING, INVENTORIES, AND INFRASTRUCTURE The petroleum industry consists of three main segments: the exploration and production segment (upstream); the refining and marketing segment (downstream); and a third segment typically referred to as the midstream, which consists of the infrastructure used to transport crude oil and petroleum products. This report is mainly concerned with certain aspects of the downstream and midstream segments, namely refining, inventories, and the pipeline and marine supply infrastructure.

Refining Refineries change crude oil into petroleum products primarily through a distillation process that separates the crude oil into different fractions based on boiling point ranges. One barrel of crude oil produces a varying amount of gasoline, diesel, jet fuel, and other petroleum products depending on the configuration–or complexity–of the refinery and the type of crude oil that is being refined. Through the addition of specialized equipment, refineries can be optimized—or “upgraded”—to produce greater proportions of specific types of products or to use different grades of crude oil. For example, hydrocracking units enable refiners to increase the production of lighter fuels including gasoline, diesel fuel, and jet fuel; catalytic cracking units increase the production of gasoline; and hydrotreating units enable refiners to produce lower-sulfur fuels required by the European Union, United States, and many other countries. Changes in product specifications, shifts in demand, and environmental regulations all have important implications for refineries. For example, the regulated shift to unleaded gasoline that began in 1974 caused refineries to install equipment to produce high-octane components to replace the lost lead. Similarly, in response to environmental regulations such as limits on the emissions of certain air pollutants refineries have invested in equipment and processes to control such emissions. The proliferation of some special gasoline blends, or “boutique fuels,” has made it more complicated to supply gasoline and raised costs, significantly affecting operations at refineries.[4] Last, to the extent that varying amounts of biofuels blended with gasoline and diesel require changes to the gasoline and fuel blendstocks, further refinery changes may be required to accommodate these blends. Shifting demand for petroleum products, such as Europe’s declining demand for gasoline and growing consumption of diesel, can also cause refiners to invest in different processes to produce the mix of products desired by the market. In general, the United States’ refineries are among the most sophisticated in the world, and domestic refineries have generally been optimized to produce large proportions of cleaner-burning gasoline to meet the huge transportation demand subject to various

14

United States Government Accountability Office

environmental constraints. Historically, U.S. and international refining capacity has broadly grown and fallen in response to shifts in demand for petroleum products. For example, U.S. and international refining capacity fell sharply during the early 1980s in response to falling demand for petroleum products, caused in part by high prices of these products and worldwide recession. By 1983, demand had fallen so much that almost 30 percent of U.S. refinery capacity was not being used. Many refineries were shut down or idled and refining capacity thus fell. Demand began growing again in the United States and internationally around 1982.

Inventories Inventories of petroleum products and crude oil are maintained by refiners, distributors, marketers, and others to mitigate the effects of disruptions, and to ensure a continuity of supply to their customers. Companies build inventories in preparation for planned maintenance and production, refining, and logistical systems. The primary inventory system comprises the crude oil or petroleum products held at production sites, refineries, and storage sites, and in pipelines, tankers, barges, and other transportation centers. Secondary inventories exist between the primary distribution system and the end user, and consist of retail outlets and small tank farms, which have less than 50,000 barrels of total capacity. Tertiary inventories are inventories held by consumers, for example, in automobile tanks. EIA collects inventory data for the primary system. EIA collects inventory data for crude oil and petroleum products held in storage at refineries, pipelines, and tank farms, and bulk terminals that can store at least 50,000 barrels of petroleum products. EIA also collects inventory data for Alaskan crude oil in transit by tanker from the terminus of the Alaskan pipeline in Valdez, Alaska, to other U.S. ports, as well as oil in the Strategic Petroleum Reserve.[5]

Pipeline and Marine Supply Infrastructure System The supply infrastructure is composed of petroleum product and crude oil pipelines, barges, vessels, marine terminals, and storage tanks. Trucks and rail also distribute a small fraction of the products––about 6 and 4 percent respectively––but are being increasingly utilized with the rise of biofuels such as ethanol, which existing pipelines cannot currently accommodate.[6] As displayed in figure 3, about 90 percent of all petroleum products currently travel by either pipeline or marine transport. Because of these two modes dominance, our analysis of the nation’s supply infrastructure system will be limited to the pipeline and the marine transport system. Pipelines are generally the least expensive mode for transporting oil and most petroleum products.[7] Most of the United States pipeline infrastructure—approximately 166,000 miles of crude oil and petroleum product pipeline—was constructed in the 1950s, 1960s and 1970s to accommodate the needs of the refining sector and demand centers at that time. These main pipelines were built to transport petroleum products from the Gulf Coast and Midwestern oil fields—where many of the nation’s refineries were—to the East Coast, the United States’ largest consuming region. The first large transmission pipelines for petroleum products were

Energy Markets

15

constructed during World War II, and ran primarily from the Gulf Coast to the Mid-Atlantic states; the network expanded significantly until the 1970s. Pipelines feed refinery centers and market hubs because the regions with the most supply are not the regions with the most demand. Most pipelines are common carriers, offering transportation services to anyone who wants them, but subject to some regulations. While crude oil and petroleum products generally do not travel on the same pipelines, numerous different petroleum products are shipped back to back in batches through the same pipelines. During this process, some blending of any two adjacent batches of petroleum products occurs where the two batches interface. This blended material may be simply mixed with the lower-valued product—for example, the mix of high- and low-octane gasoline at the interface between batches of these commodities would be downgraded, or mixed with the low-octane fuel—or, if the blended material is incompatible with either of the two petroleum products that interfaced, it must be removed and reprocessed into something that can be used. To access space on a pipeline, a shipper must ask for the right to use capacity by nominating amounts of liquid for service to be received, delivered or stored by the pipeline company. Different shippers’ nominations of common products are often combined by the pipeline in order to reduce the number of batches and therefore the amount of downgrading or reprocessing of blended products.

Source: GAO analysis of Bureau of Transportation Statistics data. Figure 3. Transport Mode of Petroleum Products in the United States, 2004.

Marine transportation of crude oil and petroleum products accounts for nearly one-third of domestic shipments. The marine transport system consists primarily of waterways; ports and vessels, including crude oil tankers; and product tankers and tank barges. Built to accommodate smaller vessels, many of the major ports have had to expand in response to increasing marine transport and trade and to accommodate larger tanker vessels.

16

United States Government Accountability Office

SEVERAL U.S. AGENCIES REGULATE AND MONITOR THE DOWNSTREAM AND MIDSTREAM OIL INDUSTRY AND PETROLEUM PRODUCT MARKETS Several U.S. agencies have jurisdiction over or monitor the U.S. downstream oil industry and petroleum product markets: •

•

•

•

•

Within DOE, EIA collects and analyzes data on the supply, consumption, and prices of crude oil and petroleum products, including inventory levels, refining capacity and utilization rates, and product movements into and within the United States. DOE’s Office of Fossil Energy manages the U.S. Strategic Petroleum Reserve (SPR), which is a federally maintained stockpile of about 700 million barrels of crude for use in the case of a major disruption of oil supplies, as well as the Northeast Home Heating Oil Reserve, a component of the SPR that has 2 million barrels of emergency fuel oil for homes and businesses in the Northeast that could be released during heating oil supply interruptions or high periods of demand caused by severe winter weather. FERC is an independent agency that regulates the transmission of oil through interstate pipelines by setting and enforcing pipeline “tariffs”—the prices and terms under which shippers send their products through the pipelines and the rules governing access to these pipelines.[8] The Federal Trade Commission (FTC) enforces antitrust and consumer protection statutes. For example, in the petroleum industry, the FTC generally reviews proposed mergers and approves such mergers only if they are deemed not to have anticompetitive effects. DOT’s Pipeline and Hazardous Materials Safety Administration regulates safety for oil pipelines that transport oil and petroleum products. Among other things, it oversees oil pipelines’ design, maintenance, and operating procedures. DOT’s Maritime Administration (MARAD) reports to Congress on the status of public ports’ supply infrastructure needs. The Environmental Protection Agency (EPA) develops and enforces regulations that implement environmental laws including the Clean Air Act, the Clean Water Act, and the Oil Pollution Act, which aim to control the discharge of pollutants into the environment by refiners and other industries. EPA also administers the National Environmental Policy Act, which requires federal agencies to consider environmental impacts of proposed actions.

In addition, individual foreign countries play regulatory roles and can affect trade conditions for products through their individual or collective actions. The IEA is an organization established by treaty of 26 mainly net oil-importing OECD countries to cope with oil supply disruptions and coordinate an international response in the case of a disruption to the global oil supply. Member countries agree to keep significant strategic stocks of crude oil and/or petroleum products to be available in the event of a severe supply disruption. IEA also maintains a database that provides information on IEA member crude oil and petroleum product inventory levels, refining capacity, and utilization rates.

Energy Markets

17

PETROLEUM PRODUCTS MARKETS HAVE BECOME INCREASINGLY GLOBAL WITH GREATER TRADE AND PRICES INCREASINGLY LINKED ACROSS COUNTRIES International trade in petroleum products has expanded significantly over the past 20 years, making the markets for gasoline, diesel, and jet fuel increasingly global in nature, and providing additional gasoline supply options for the United States. This trend has been particularly important for the United States, which has seen large increases in the volume of imported gasoline. A key impetus for global trade in petroleum products has been a structural surplus in production of gasoline and deficit in production of diesel in Europe as a result of a systematic switch in European countries to diesel-burning automobiles. While many experts we spoke with believe that growth in international trade of petroleum products will likely continue, they identified several factors that may limit or change the patterns of trade, including plans and mandates to introduce significant volumes of biofuels and the potential expansion of differing fuel specifications that a proliferation of biofuel blends would entail.

International Trade in Petroleum Products has Expanded Significantly International trade in petroleum products has expanded significantly over the past 20 years, making the markets for gasoline, diesel, and jet fuel increasingly global in nature. Specifically, our analysis of IEA data shows that OECD imports of gasoline, diesel, and jet fuel more than doubled between 1984 and 2006, from about 80 million barrels per month to over 160 million barrels per month. Similarly, OECD exports increased from about 55 million to over 140 million barrels per month over the same time period. While OECD exports and imports in these products have more than doubled, OECD demand for these products rose by less than 40 percent during the same time period. Figure 4 shows the increase in OECD imports and exports of gasoline, kerosene-type jet fuel, and diesel fuel.[9] Trade in gasoline has been particularly important for the United States, which has seen large increases in the volume of imported gasoline. While in 1970 the United States was largely self-sufficient in gasoline, we now import over 10 percent of our annual consumption of gasoline and smaller percentages of jet fuel and some other products.[10] U.S. imports of gasoline and gasoline blending components, which accounted for about 31 percent of our imports of refined petroleum products in 2005, averaged about 1.1 million barrels per day, or more than 10 percent of U.S. daily consumption. According to DOE, imports have supplied about half of U.S. gasoline demand growth from 1993 to 2005.[11] These “total gasoline” imports include finished gasoline, which can be sold directly to retail markets, as well as gasoline blending components that are combined in the United States to make finished gasoline to serve various markets, some of which use special, cleaner-burning gasoline blends as part of their plans to meet federal air quality requirements.

18

United States Government Accountability Office

Source: GAO analysis of IEA data. Figure 4. Imports and Exports of Gasoline, Kerosene-type Jet Fuel, and Diesel Fuel for All OECD Countries, 1984 – 2007.

Source: GAO analysis of EIA data. Figure 5. U.S. Imports of European Gasoline and Gasoline Blendstocks, 1993 – 2007.

A key impetus for global trade in petroleum products has been a structural surplus in production of gasoline and a deficit in production of diesel in Europe. This surplus of gasoline is largely the result of a systematic switch in European countries toward automobiles with diesel-powered engines, which are more efficient than gasoline-powered engines. European regulators promoted diesel fuel use in Europe by taxing diesel at a lower rate, and European demand for diesel fuel-powered vehicles rose. The European refining and marketing sector responded to this change in demand by importing increasing amounts of diesel, and exporting a growing surplus of gasoline to the United States. The United States has purchased

Energy Markets

19

increasing amounts of gasoline, including gasoline blendstocks, from Europe in recent years, as shown in figure 5. These imports have generally had a strong seasonal component, with higher levels of imports during the peak summer driving months and lower imports during the fall and winter. The major exception to this seasonality came in the months October 2005 through January 2006, when imports surged in response to U.S. shortfalls as a result of damage to and shutdowns of refineries and pipelines following Hurricanes Katrina and Rita in August and September 2005, respectively. Experts and company representatives told us they believe this structural imbalance within the European Union will continue for the foreseeable future, and perhaps widen, resulting in more exports of European gasoline and blending components to the United States. Specifically, company representatives and industry experts told us that European refiners are unlikely to significantly expand their refining capacity in the near future or reconfigure to produce less gasoline for a number of reasons, including the following: •

•

•

The profitability of the U.S. gasoline market acts as a draw for surplus gasoline worldwide. Many company representatives told us that the United States’ continued appetite for gasoline—combined with many countries’ declining demand—has resulted in most surplus gasoline being exported to the United States, and that this trend would likely continue in the future. For example, some refining interests in Europe told us they had configured their refinery operations to be essentially a U.S. “gasoline machine.” Construction costs have increased significantly, raising the cost of investments in refining capacity or upgrades. For example, some refining interests in Europe and elsewhere told us that some planned conversion and upgrading of refinery capacity in Europe was on hold, because of increased construction costs worldwide. Some of these upgrade plans called for enhanced diesel fuel production mainly for the European market, as well as surplus gasoline exported to the United States. European refiners told us that they are reluctant to make large investments necessary to produce significantly more diesel because doing so will increase their greenhouse gas emissions. Their concern is that as greenhouse gas emissions caps are lowered, companies will be required to pay to reduce emissions or buy costly emissions credits.

EIA and other experts have stated that, at times, imports from Europe could be provided more competitively than gasoline from the U.S. Gulf Coast and other domestic refineries. In addition, more sources of supply can also benefit the United States in the event of domestic supply disruptions. For example, the flexibility in sources of supply helped U.S. marketers and retail sellers obtain gasoline and other petroleum products in the aftermath of Hurricanes Katrina and Rita in August and September of 2005, respectively, when a large fraction of the nation’s refineries and pipelines were temporarily shut down. During the 3 months following the hurricanes, imports of gasoline to the United States increased by about 30 percent compared to what they were during the same months in the previous year, and imports came from a number of countries that do not typically sell to the U.S, market. Imports of other petroleum products into different regions of the country also rose. As illustrated by figure 6, U.S. imports of petroleum products surged in response to Hurricanes Katrina and Rita compared to levels during the same months of the previous year.[12]

20

United States Government Accountability Office

In addition to gasoline, kerosene-type jet fuel imports into the Gulf Coast surged to about 3.3 million barrels in October of 2005, compared to just 20,000 barrels in October 2004. Some countries that did not export significant quantities of this fuel in 2004 exported significant quantities following the hurricanes to the United States in 2005. For example, France exported 580,000 barrels of kerosene-type jet fuel to the United States in October 2005, but nothing in October 2004 or October 2006.

Source: GAO analysis of EIA data. Figure 6. Petroleum Product Imports into the United States, 2004-2006.

Our analysis of wholesale prices in the United States, Europe, and Asia shows that prices in geographically dispersed markets rose significantly following Hurricanes Katrina and Rita, indicating that prices in these markets are linked to some extent. Because imports surged from many countries in response to the resulting supply disruptions in the United States, gasoline prices around the world rose along with prices in the United States before prices eventually returned to pre-hurricane levels. Figure 7 illustrates the price spikes that occurred in late August and late September 2005 as a result of the severe damage to oil and gas production facilities in the Gulf of Mexico and to refineries and pipelines onshore from Hurricanes Katrina and Rita. The figure clearly shows that European and, to a lesser extent, Asian spot gasoline prices—wholesale prices for gasoline traded on a daily basis at major market centers—responded to the resulting petroleum product supply disruptions in the United States. The additional supplies to U.S. markets from Europe and elsewhere reduced prices in the United States, and spot prices everywhere declined to pre-hurricane levels before the middle of October.

Energy Markets

21

Source: GAO analysis of EIA data. Note: Breaks in lines represent days for which no data were recorded. Figure 7. Wholesale Gasoline Prices during the 2005 Hurricanes, 2005 – 2006.

While experts have stated that the availability of additional sources of petroleum product supplies has benefited the United States through lower and less volatile prices, and foreign gasoline supplies clearly helped reduce prices following Hurricanes Katrina and Rita, the fact that petroleum product markets are international means the United States will be exposed to supply disruptions or unexpected increases in demand anywhere in the world. Further, because some foreign suppliers are further away from the U.S. demand centers they serve than the relevant domestic supply center, the length of time it takes to get additional product to a demand center experiencing a supply shortfall may be longer than had the United States had more refining capacity. For example, imports of gasoline to the West Coast may come from as far away as Asia or the Middle East, and the transport time and therefore cost is greater. To the extent that imported gasoline or other petroleum products come from far away, the lengthening of the supply chain has implications for the ability to respond rapidly to domestic supply shortfalls. Specifically, if supplies to relieve a domestic regional supply shortfall must come from further away, the price increases associated with such shortfalls may be greater and/or last longer. In this sense, the West Coast is more vulnerable to price increases or volatility than is the Northeast, which can receive shipments of gasoline into New York Harbor or elsewhere in the U.S. Northeast from Europe, often on voyages of less than a week.

22

United States Government Accountability Office

Growth in International Trade of Petroleum Products is Expected to Continue but Growth in Biofuel use May Limit or Change the Patterns of Trade With demand for petroleum products growing globally, experts we spoke with believe the trade in petroleum products will continue to increase for a number of reasons. For example, global trends toward lower-sulfur fuels have resulted in more uniform sulfur specifications, creating more trade opportunities. Strong global demand for certain petroleum products— especially distillates such as diesel and jet fuel—will increase competition for, and facilitate global trade of, these petroleum products. For example, since 2005, diesel wholesale prices have generally been at a premium compared to the price of gasoline, in response to sharp consumer demand, and in the United States, diesel demand grew 6.9 percent in 2005, compared to 2.5 percent for gasoline. Demand for jet fuel is growing with the increase in air transportation, and given that jet fuel has uniform global specifications, jet fuel will continue to trade relatively freely based on global price signals. While many experts we spoke with believe that growth in international trade of petroleum products will likely continue, the planned expansion of the use of biofuels, such as ethanol made from corn or other crops, and biodiesel made from soybeans or other crops, in the United States and many other countries could reduce the growth of demand for petroleum products and thereby reduce the opportunity for trade. At the U.S. federal level, the EPA administers the Renewable Fuel Standard Program, which went into effect in 2007 and requires most U.S. gasoline refiners, importers and blenders to sell a minimum portion of biofuels each year. Refiners can meet the standard by blending biofuels with conventional gasoline or diesel in various proportions. Plans and mandates in a number of countries and regions, including the United States to introduce larger volumes of biofuels, primarily as additives to gasoline or diesel, could displace demand for and trade in petroleum products. In addition and as discussed previously in this report, some of these biofuel policies mandate that all gasoline or diesel sold in an area be blended in specific proportions with biofuels, with differences across states in the timing and level of such blending. Still other states specify a certain proportion of biofuels to be blended but allow flexibility in how they are blended, thereby creating the potential for widely different biofuel blended fuels within even a single state—for example, the mandated biofuel requirement could be met by a uniform proportion of ethanol blended into every gallon of gasoline used in the state, or by using a small amount of E85 (fuel composed of 85 percent ethanol and 15 percent gasoline components) with ethanol blended into the rest of the gasoline, and any combination of blends and volumes that meets the overall requirements would also satisfy the mandate. States and localities have pursued such policies for a variety of reasons, including viewing biofuels as a means to boost farm economies by increasing demand for feedstock crops while also contributing to a cleaner environment.[13] However, the current absence of uniform standards for biofuels and varying plans by various countries and regions to blend different volumes of biofuels with petroleum-based gasoline and diesel could reduce the fungibility of these fuels and thereby reduce opportunities for trade.

Energy Markets

23

GLOBAL AND DOMESTIC REFINING CAPACITY HAVE NOT KEPT PACE WITH DEMAND, LEADING TO TIGHT DEMAND AND SUPPLY BALANCE AND RECENTLY CONTRIBUTING TO HIGHER PETROLEUM PRODUCT PRICES For most of the past 25 years, there has been excess refining capacity globally, but this excess capacity has been reduced over time as demand has increased faster than capacity has grown. Capacity growth has lagged behind demand for a number of reasons, including low profitability in the refining sector and demands on industry to meet changing fuel specifications and reduce emissions of environmental pollutants. More recently, unexpectedly rapid growth in demand for petroleum products caused refineries to run closer to their production capacity. Current market tightness has contributed to higher and more volatile prices and increased profits in the refining industry. While these higher profits have encouraged increased investments in refining capacity, it is unclear whether or for how long the current market tightness will continue. This uncertainty is, in part, because rising construction costs and uncertain future demand make it difficult to estimate how many of the planned refining projects will actually be completed and because biofuel initiatives in many countries could reduce demand for petroleum products while potentially requiring further refining investment to make and keep separate different gasoline and diesel specifications to be blended with ethanol and biodiesel.

Demand for Petroleum Products Has Grown More Quickly than has Refinery Capacity, Tightening the Supply and Demand Balance Worldwide For much of the past 25 years, demand for petroleum products in the United States and internationally has outpaced growth in refining capacity. Demand for petroleum products fell dramatically from 1978 to 1982, creating significant excess capacity—by 1983, almost 30 percent of U.S. and world refining capacity was idle. Demand for petroleum products began growing again around 1982, and this demand growth, along with the shutting down of some idle refining capacity, began to narrow the gap between capacity and demand. Since that time, growth in demand for petroleum products has generally exceeded growth in refinery capacity, causing refineries to run more intensively to meet demand. Figure 8 shows how refinery utilization in the United States and internationally, with a few exceptions, including the countries of the former Soviet Union, has increased significantly since the early 1980s. Refining capacity in the United States has been growing since 1994 through expansions at existing refineries. The last major complex refinery on a new, or “green field” site in the United States was built in the 1970s, and many, mostly smaller, refineries were shut down starting in the early 1980s. However, as figure 9 shows, even as the number of refineries in the United States fell since 1981, refiners have since 1994 generally expanded total capacity at remaining facilities. Capacity expanded by an annual average of 192,000 barrels per day between 1994 and 2006—more than the average-sized refinery in 2006, which had a capacity of 116,000 barrels per day. For example, ExxonMobil’s Baytown refinery grew by about 166,000 barrels per day in capacity between 1994 and 2006, more than equivalent to adding a new refinery. In this sense, it is potentially misleading to say that no new refineries have been

24

United States Government Accountability Office

built in the United States since the 1970s. Instead, experts have said that expansion in the United States has centered at existing facilities because such expansion is less expensive than building an entirely new refinery at a new, “green field” site because of lower construction, permitting, and land acquisition costs. Some industry officials we spoke with said that construction at a green field site can be about two to three times more expensive than expanding capacity at existing sites on a per barrel basis.

Source: GAO analysis of BP Statistical Review of World Energy June 2007. Figure 8. Refinery Utilization in the World and Selected Countries, 1980-2006.

Source: GAO analysis of EIA data. Figure 9. U.S. Crude Oil and Petroleum Product Consumption and Number and Capacity of Operable Refineries, 1973 – 2006.

Energy Markets

25

Figure 10 shows how much of the recent growth in refining capacity in the United States has been concentrated in the Gulf Coast. This growth in capacity in the Gulf Coast is consistent with the view of many industry experts we spoke with that the Gulf Coast provides one of the most competitive environments for U.S. refiners. Experts cited several factors, including ready access to imported crude oil supplies, numerous options for shipping product to the rest of the United States by pipeline and waterways, and a concentration of highly skilled workers. U.S. refineries also have invested in equipment to upgrade their refineries to be able to produce more high-value products from a wider variety of raw inputs. For example, hydrocracking equipment enables refiners to adjust the yields of various products, and coking capacity allows refiners to process heavier crude oils. Figure 11 shows how the capacity of such downstream units, particularly hydrocracking and coking, has grown faster than distillation capacity overall. The addition of such downstream units does not increase the distillation capacity of refineries—the traditional measure of capacity—but enables refineries to produce a greater portion of products in high demand (such as gasoline, diesel, and jet fuel) and also to process more heavy and sour crude inputs. In fact, the proportion of gasoline, diesel fuel and jet fuel produced per barrel of crude input in the United States has increased from 77 percent in 1993 to 81 percent in 2005 even as the quality of crude oil inputs used has deteriorated.

Source: GAO analysis of EIA data. Figure 10. Refinery Capacity by U.S. Region, 1985-2006.

26

United States Government Accountability Office

Source: GAO analysis of EIA data. Figure 11. U.S. Refinery Distillation Capacity and Capacity of Selected Downstream Units, 1982-2006.

Industry officials and experts we spoke with said that several factors have caused refinery capacity to grow more slowly than demand in the United States. First, industry officials and experts said that refining has been a low-, even negative-return business for much of the past two decades, with profits too low to encourage significant expansion. Companies in the oil industry overall, which includes upstream oil exploration and production activities as well as downstream refining and retail marketing, have in general performed better than some industries and worse than others. However, according to an analysis by Deutsche Bank, cash returns on investment for oil companies in the Standard and Poor’s 500 index were less than the cost of capital from 1986 to 2000. In other words, it cost companies more to raise the money to invest than those investments earned. Within the oil industry, the refining segment has been less profitable than other lines of business in the petroleum industry, according to EIA data, as illustrated in figure 12. Except for a few years since 1977, returns for U.S. refining and marketing operations have been lower than returns in foreign refining and marketing and lower than exploration and production. Specifically, during the entire period 1977-2005, average return on investment for the U.S. refining industry was 7 percent, compared to 9 percent for foreign refining and slightly over 10 percent for all other lines of business. For the integrated oil companies that still control a major portion of the nation’s refining capacity, U.S. refining must compete with foreign refining and upstream investment options for capital. The lower returns for U.S. refining can make it harder for companies to justify expanding U.S. refining capacity.

Energy Markets

27

Source: GAO analysis of EIA data. Note: This graphic is based on the performance of major energy producers covered by EIA’s Financial Reporting System (FRS). These companies represented about 81 percent of U.S. refining capacity in 2005. Figure 12. Return on Investment in U.S. and Foreign Refining Compared to Other Lines of Business for Major Energy Producers, 1997 – 2005.

Another indication that the refinery industry has long had low expectations of profitability is that existing refinery capacity has sold very cheaply. U.S. refineries have often sold for significantly less than what it would cost to build a new similar refinery. According to an analysis by the National Petroleum Council of the value of existing refinery purchases between 1998 and 2004, refineries sold for about one-fourth to one-third the cost of equivalent new construction. The cost of buying an existing refinery was also less than the general cost of expanding capacity on an existing refining site, which experts indicated could be less than one-half the cost of new construction. This suggests that refiners have had low expectations of future returns in the U.S. refining market. This also indicates that until recently, a refiner looking to expand capacity in the United States may have been able to do so more affordably by purchasing an existing refinery. This would add to that refiner’s capacity, but would not expand domestic refinery capacity overall. A second reason experts cited for slow domestic refinery capacity growth is that more rigorous product specifications; the proliferation of special gasoline blends, or “boutique fuels” around the country; and environmental controls have all required refineries to invest in additional processes in order to meet the specifications and regulations, and these investments did not typically add to base capacity. Officials we spoke with said that the large investments required to reduce harmful air emissions at refineries and meet more stringent product specifications drew from the capital that may otherwise have been available to invest in expanding capacity. A third reason for slow domestic refinery growth, according to some industry representatives, is that permitting difficulties have discouraged refinery expansions. Refineries are required to obtain permits from relevant state and local authorities in order to build or expand refinery capacity. These are often difficult to obtain owing to regulatory

28

United States Government Accountability Office

hurdles and public opposition. Other experts suggested that permitting adds to the difficulty of expanding capacity but has been a less important factor than overall low expectations of returns. DOE officials told us that resistance to refinery expansions by nearby communities could be a more important factor in discouraging new domestic refinery construction or expansions. Finally, EIA officials and others pointed to the surplus production of gasoline in Europe as a major, more recent, reason domestic refinery capacity has not kept up with domestic demand. They stated that Europe could at times provide gasoline to the Northeast more competitively than some Gulf Coast refiners, and therefore gasoline imported from Europe has displaced domestic supplies and discouraged domestic refinery expansions. More recently, global demand for gasoline, diesel, and jet fuel grew particularly quickly around 2004, partly as a result of rapid growth in China, where demand surged by over 15 percent. In order to increase production and meet this recent surge in demand for petroleum products, refiners have had to run their refineries even more intensively—capacity cannot be added quickly because of the long lead times involved in designing and constructing a refinery or an expansion project. Since 2004, world refinery utilization rates have risen to around 86 percent, the highest levels since 1980, when data first became available. Experts told us that many refineries have been running near their production capacity in order to meet demand. This is particularly true in the United States and Europe, where refineries have been running at or near 90 percent utilization since the 1990s, even though spare capacity still existed worldwide, particularly in the countries of the former Soviet Union and to a lesser extent in the Asia Pacific region.

Current Market Tightness has Contributed to Higher Petroleum Product Prices, Higher Price Volatility, and Higher Industry Profits The recent tightening of the balance between supply and demand for petroleum products has, along with higher crude oil prices and other factors, contributed to increased petroleum product prices and higher industry profits, and has contributed to greater price volatility. In addition, a tight demand and supply balance means less flexibility in industry’s response to unanticipated events. For example, at times of excess capacity, if a particular refinery were to unexpectedly shut down for emergency maintenance, capacity that wasn’t being used could be brought on line to meet demand. However, when refineries are generally running near capacity, there is less excess capacity to call on, and what available capacity there is tends to be located farther away from demand because the lower-cost and nearer refining capacity tends to be used up first. An analysis by the FTC illustrated the effects of tight refining capacity at the regional level. This analysis compared the impact on gasoline prices of two refinery outages in the upper Midwest in the spring and summer of 2001 with a major refinery outage in Oklahoma in July 2003. Each of the Midwest refinery failures was associated with wholesale gasoline price increases of between 30 and 40 cents.[14] By contrast, the Oklahoma refinery failure was found to have little effect on gasoline prices in that state. The FTC attributed the difference in the price responses to the fact that the upper Midwest region lacks sufficient refinery capacity to meet the region’s demand, while Oklahoma produces significantly more petroleum products than the state needs. Therefore, when a major disruption occurs, the upper

Energy Markets

29

Midwest must rely on supplies from distant refineries, while Oklahoma simply exports fewer petroleum products to other states. Further, as we have previously reported, the West Coast of the United States generally has higher gasoline prices than much of the rest of the country.[15] Among the reasons for these consistently higher gasoline prices are a tight supply and demand balance for gasoline, the fact that the region is isolated from other major domestic and foreign refining regions, and the adoption in California of a unique blend of gasoline that is more costly to make than many other blends and that is not routinely produced by many refineries outside the West Coast. Prices in the West Coast can rise rapidly in response to supply disruptions as a result of these factors. Profits in the refining industry have increased significantly since 2002, in part as a result of sustained market tightness and, in the United States, owing to wide price differentials between heavy and light crude oils. For example, the difference between crude oil input prices and petroleum product prices, a strong indicator of refining profits, has increased worldwide, though particularly in the United States. In the United States, these wide price spreads have caused returns on investments in the refining and marketing segment of the petroleum industry to reach record levels in 2004 and 2005, the latest data we were able to obtain. In the United States, these higher margins are, in part, the result of the ability of U.S. refineries to take advantage of low-price, low-quality crude oil inputs. Sophisticated U.S. refineries are able to convert large quantities of low-quality crude oil inputs into highervalued products, while refineries in the rest of the world do not have such capacity to the same extent. Shifts in global crude oil production and demand have contributed to a glut of such low-quality oils, lowering their price relative to higher-quality crude oils and improving the position of U.S. refineries relative to that of their international competitors.

Increased Profit Margins have Led to More Investment, but Future Market Tightness Will Depend on Several Factors Currently high petroleum product prices and high profits in the refining industry have spurred new refinery capacity investments in the United States and internationally. Global investment in refining has increased in recent years. According to IEA data, capital spending grew from $34 billion in 2000 to $51 billion in 2005 and is expected to average $60 billion per year between 2006 and 2010. Analysis by IEA of plans and projects currently underway worldwide suggest that almost 10.5 million barrels per day of capacity may be added through 2011. This rate of refinery capacity growth is somewhat higher than expected demand growth, which is projected to grow by about 9.2 million barrels per day by 2011. The majority of this capacity expansion is expected to take place overseas, especially in China, India, and the Middle East. In the United States, EIA officials have said that announced investments through 2011 could add an additional 1 million barrels per day to domestic refinery distillation capacity, along with significant additions to downstream processing capacity. If realized, these domestic and international investments could help alleviate the tight balance between refining supply and demand. However, since tight refining capacity is just one of a number of factors affecting prices, the biggest factor being crude oil prices, even a less tight refining market may not bring much price relief at the gasoline pump.

30

United States Government Accountability Office

While recent profits and prices have renewed interest in expanding refining capacity, experts said it is unclear whether or for how long current refining market tightness will continue. Future refining market tightness depends on changes in refining capacity and on changes in the demand for petroleum products. Industry officials and experts we spoke with said future conditions are highly uncertain for several reasons: •

•

•

•

It has become much more costly to expand refinery capacity in recent years due to rapidly rising construction costs. Various construction materials such as iron, steel, and concrete are important in energy projects, and their costs have increased significantly. For example, while prices for iron and steel fell in the decades prior to 2002, prices increased by 9 percent annually between 2002 and 2004, and by 31 percent from 2004 to 2005. Similarly, industry officials said that many decades of low investment levels have led to a small pool of qualified project engineers to design and oversee construction and expansion projects, causing labor prices to soar. Moreover, the Nelson-Farrar refinery construction cost index, which tracks prices for materials such as iron and steel, equipment and skilled and unskilled labor, shows that costs for refinery investment rose by 17 percent from 2002 to 2005 in real terms. Industry officials indicated that these cost estimates did not capture the full extent to which refinery expansion costs have increased. Officials also said that the waiting lists to purchase key refinery equipment are getting longer. In the United States and in Europe, some planned refining expansions have been delayed or canceled, in part because of these rising costs and delays in acquiring equipment and skilled labor. Uncertainty about future demand makes refinery investments risky and may inhibit investments. The United States is considered a mature market, with demand for motor gasoline forecast by EIA to grow by 1.2 percent annually between 2005 and 2030. Meanwhile, refinery capacity has on average expanded by almost 1 percent annually since 1999. Some industry officials we spoke with believe that U.S. demand for petroleum products will peak in the next decade and then begin to decline, implying only a temporary need for additional refining capacity. Company representatives told us that various proposals under consideration dampen the U.S. refining investment climate. For example, the Administration has proposed to reduce U.S. petroleum gasoline consumption by 20 percent by 2017 through increased use of biofuels and more stringent automobile fuel economy standards. If achieved, this could turn the United States from a gasoline importer to a net exporter within 10 years; and current refining capacity could meet future demand even without expansions that are currently planned. Similar initiatives to blend large volumes of biofuels into the transportation fuels markets in other countries could similarly displace demand for petroleum products. It is unclear whether such initiatives could ease the demand and supply tightness that currently exists. On the one hand, reducing demand growth can reduce pressure on refinery capacity. On the other hand, reduced expectations of future demand can alter the attractiveness of refinery investments, and some refiners may respond by cutting back refinery expansion plans. New initiatives to blend biofuels in varying proportions into transportation fuels could potentially add to the need for further refining investments both to refine and to keep separate new blending stocks, possibly absorbing resources that could have

Energy Markets

31

been used to expand capacity. Automobile industry experts we spoke with agreed that each different ethanol blend requires a specific gasoline or diesel blend stock in order for the resulting blended fuel to meet performance and emissions standards. In other words, the gasoline that is blended with ethanol to make E10 (10 percent ethanol) is different than the gasoline used to make E85 (85 percent ethanol). The absence of national standards for blending biofuels with gasoline and diesel could also increase the number of gasoline and diesel blending stocks refiners have to make, and could lead to a proliferation of new blendstocks. Further, to the extent that new equipment is needed at refineries in order to produce, handle, or keep separate these various blendstocks, refineries will need to invest in this equipment in order to meet various federal, state, and local biofuel mandates and standards. These added investments could crowd out resources that could otherwise have gone to expanding refinery capacity.

DOMESTIC AND OECD INVENTORIES OF PETROLEUM PRODUCTS AND CRUDE OIL HAVE DECLINED RELATIVE TO DEMAND, WITH MIXED EFFECTS ON PRICES AND PRICE VOLATILITY When measured as average days of consumption, long-term trends in inventories of petroleum products and crude oil in the United States indicate a general decline over the past 20 years. Similarly, gasoline and crude oil inventories in OECD countries excluding the United States have also generally fallen over the same period. However, there are limitations to inventory data as measured by EIA and IEA, in part because these data do not fully match stocks with their intended markets; in general petroleum product exporting regions will typically have large stocks of these products relative to that region’s demand, while inventories held in net importing regions will typically be lower relative to demand. For example, petroleum products stocks of gasoline in Canada, Europe, and the Caribbean that are destined for the United States are counted as inventories in those countries but not as inventories in the United States. A number of factors have contributed to the long-term decrease in inventory holdings in the United States, including reductions in domestic crude oil production and in the number of refineries as well as advances in technology and management processes that allowed for reduced inventories and a concomitant reduction in operating costs. Lower operating costs associated with lower inventories may have translated into lower consumer prices during normal periods. However, in the short term, inventory levels tend to fluctuate within a “normal” range, and—since inventories provide a smoothing effect against temporary demand and supply fluctuations—lower than normal inventories can lead to higher prices in the event of supply disruptions or surges in demand.

Inventory Levels of Petroleum Products in the United States and Other OECD Countries have Generally Fallen over the Past Two Decades Privately held inventories of petroleum products and crude oil in the United States have generally fallen over the past two decades, in terms of average number of days worth of

32

United States Government Accountability Office

supply, or “days forward cover.”[16] Specifically, as illustrated in figure 13, days forward cover for gasoline in the United States averaged about 30 days in 1984 but fell to an average of about 12 days for the first 5 months of 2007. Similarly, crude oil days forward cover fell from about 29 to about 22 days, and jet fuel and diesel fuel days forward cover also fell over the same period.

Source: GAO analysis of EIA data (annual average of monthly data). Figure 13. Crude Oil (Excluding Strategic Reserves), Finished Motor Gasoline, Kerosene-type Jet Fuel and Diesel Fuel Days Forward Cover in the United States, 1984 – 2007.

Other OECD countries have also generally seen a reduction in days forward cover for motor gasoline. Figure 14 shows the general downward trends in gasoline days forward cover for the OECD regions of Europe, Pacific, and North America excluding the United States. Specifically, European stocks declined from about 50 days in 1984 to about 40 days in 2001, before increasing to almost 46 days on average for the first 5 months of 2007. The much larger inventory figure for Europe compared to that for the United States reflects the fact that the inventory data include strategic stocks of gasoline held by some private companies.[17] The recent increase in European stocks of gasoline coincides with a period in which demand for gasoline fell in Europe relative to supply, and exports of gasoline to the United States increased a great deal. Figure 14 also shows large reductions in gasoline stocks in North America, excluding the United States, from about 44 to 18 days forward cover over the same time period, while stocks in Pacific OECD countries fell more modestly from 20 to almost 16 days. Crude oil stocks in two of the three other OECD regions, Europe and Pacific, decreased over the period, while stocks in North America excluding the United States rose significantly, driven in part by increases in Canadian oil sands production and the storage and delivery infrastructure associated with this increased production. Figure 15 illustrates these changes in crude oil inventories in days forward cover. We do not have inventory data for non-OECD countries. However, as with petroleum products, net crude oil exporting countries would be expected to have much higher levels of days forward cover for crude oil than net importing countries, especially if strategic stocks are excluded.

Energy Markets

33

Source: GAO analysis of IEA data (annual average of quarterly and monthly data). Figure 14. Motor Gasoline Days Forward Cover, by OECD Region, 1984-2007.

Source: GAO analysis of IEA data (annual average of quarterly and monthly data). Figure 15. Crude Oil Days Forward Cover, by OECD Region, 1984 – 2007.

Inventories, as measured by IEA, EIA, and others, have some limitations as a measure of what is available to meet demand in the event of a supply shortfall. For example, as discussed above, the United States has imported an increasing share of its gasoline over the period during which inventories have fallen, and as such, the domestic inventory data do not account for large volumes of these products on the water in tankers from foreign sources that are destined for the U.S. market or in storage terminals at foreign ports serving this trade in gasoline. Our analysis indicates that about 16 million barrels of gasoline and gasolineblending components were en route to the United States on the average day during 2006,

34

United States Government Accountability Office

representing about an additional 2 days of forward cover, and an unknown additional amount is held in storage terminals that would be available for shipment in the event of a supply shortfall in the United States. Data on U.S. gasoline inventories may further under-represent available inventories if we ignore the part of inventories held at foreign refineries that are intended to serve the U.S. petroleum products market. The inventories represented in these refineries’ storage systems and in the tanker and pipeline system supporting the flow of products to the United States, or at least some portion, could be considered part of U.S. inventories for the purpose of evaluating our days forward cover of products in the event of a supply disruption. However, it would be difficult to estimate the precise volumes of these foreign inventories as we do not collect such data from offshore suppliers and because many of these inventories are at varying distances from U.S. markets and would have to be evaluated differently, depending on how long they would take to reach the United States in the event of a domestic supply shortfall.[18] Another limitation in interpreting inventory data arises because much of the measured volumes of petroleum products in pipelines cannot be effectively removed from the pipelines in the event of a supply shortfall because they are needed to keep the pressure in the pipelines at operable levels. Similarly, some inventories are in so-called “tank bottoms,” or the part of storage tanks that cannot effectively be retrieved in the event of a supply shortfall. As a result of these and other limitations, we do not have an accurate measure of precisely how much is in the full supply chain to the United States, or the actual number of days’ worth of usable supply we could rely on in the event of a supply disruption. Looking forward, the refining expansions discussed previously in this report may lead to increases in the days forward cover measure as pipelines and storage facilities associated with the new refining capacity add to inventory holdings. However, any increase in days forward cover is likely to be modest overall because demand is also projected to grow, and companies continue to strive to develop more efficient inventory holding practices and reduce costs associated with holding any excess inventory.

A Number of Factors Have Contributed to the Long-Term Decrease in Inventory Levels Since 1980 A number of factors have contributed to the long-term decrease in U.S. days forward cover. These factors include (1) a reduction in the number of refineries and falling domestic crude oil production, (2) the fact that demand has been rising faster than refining capacity for much of the past 20 years, (3) gains in technological and management efficiency that have allowed companies to reduce the level of operating inventories, and (4) the rise of futures markets for crude oil and gasoline that have enabled oil companies and others to reduce exposure to market risk by holding financial as well as physical stocks of these commodities.[19] In the United States the decline in U.S. crude oil production resulted in decreased inventory in gathering pipelines and storage infrastructure, as pipelines and storage tanks were decommissioned. This decline in production-related inventories could be quite significant, although we do not have data to measure it directly. Oil production in the United States peaked in 1970 at around 10 million barrels per day, but by 2005 had fallen to less than 6 million barrels per day. This decline in production has left a number of abandoned crude oil

Energy Markets

35

pipelines and therefore represents a reduction in measured inventories. Similarly, the closure of many small refineries and the decommissioning of these refineries’ storage and pipeline connections to the greater supply infrastructure also reduced inventories held at these facilities and in the pipeline connections. As discussed previously in this report, this reduction in the number of refineries was significant. For example, in 1980, there were well over 300 refineries in the United States, while in 2006 the number was about 150. This sheer drop in numbers probably overstates the drop in associated inventories because, while the number of refineries fell, the average size of refineries rose, both because it was smaller refineries that tended to be shut down and because many of the remaining refineries expanded their capacity significantly. Nonetheless, EIA has stated that refinery closures had an important impact on petroleum stocks prior to 1995. A related cause of the reduction in days forward cover for petroleum products has been the fact that for much of the past 20 years, demand for these products has risen faster than domestic refining capacity. Because days forward cover is measured as the number of days’ worth of demand that is in the domestic supply chain, any increase in demand that is not met by a commensurate increase in domestic supply will lead to a reduction in days forward cover. However, as discussed previously in this report, this is potentially misleading because the supply chain between foreign refiners and the United States is relevant for measuring actual days forward cover and the U.S. imports of gasoline have been increasing dramatically over the past decade, and imports of crude oil have been increasing for much longer than that. According to company representatives and industry experts we spoke with, as well as the National Petroleum Council, delivery system efficiency improvements have also resulted in reduced crude oil inventory levels. Company representatives told us improved information technology has given managers better tools needed to optimize stock levels, and that this was mirrored in many other industries over this same period, as improved logistics and management practices enabled companies to more closely track production and delivery. By reducing inventories, refiners were able to reduce their operating costs, providing incentives to invest in efficiency-improving measures. Some officials told us that low refining profit margins were a major driver in getting companies to reduce their inventory holdings. Finally, the relationship between the future price of crude oil and petroleum products and the amount of inventory stored has, at times, contributed to changes in overall inventory levels. For example, according to a 1997 EIA report, during the period between 1995 and 1996, the prevalence of instances where the current trading price of crude oil or petroleum products was higher than the future expected price appeared to be an important factor behind the reduction in stocks.[20] In such a price environment, oil companies and others could sell currently held inventories and buy futures contracts to meet their future expected needs. In recent years, the future price of crude oil has most frequently been higher than the current price, and industry officials told us that this has been a factor in explaining why absolute inventory levels increased between December 2004 and early 2007. For example, total U.S. crude oil stocks, excluding strategic stocks, were about 355 million barrels in June 2007, or about 31 percent higher than in January 2004, according to EIA data.

36

United States Government Accountability Office

Long-Term Inventory Cost Reductions have Likely Reduced Prices of Gasoline and Other Petroleum Products, but, in the Short Term, Reductions in Inventory Levels below Normal Ranges Can Lead to Higher Prices during Supply Shortfalls To the extent that improved technology and inventory management over the long term have resulted in lower operating costs, some of the savings may have been passed on to consumers in the form of lower prices. We found no consensus among industry experts about the extent of such price reductions, nor any empirical analyses that would quantify the savings to consumers from lower inventory holding costs. However, because refineries compete with one another to sell their products, they would likely be forced to pass on some savings in operating costs in order to remain in operation, especially during that portion of the last three decades in which there was unused refining capacity as well as during recent years when surplus gasoline production in Europe has increasingly found its way to U.S. markets. However, because inventories provide a smoothing effect against temporary demand and supply fluctuations, lower than normal inventories can signal underlying changes in supply and demand conditions that will cause prices to rise. For example, if a large refinery in the United States were to suffer an unexpected outage, the resulting reduction in domestic supplies would likely result in a drawdown of that refinery’s inventories to meet its demand, and if that is insufficient, the refinery would buy from other refiners. If inventories were on the high end of the normal range, such a disruption would likely have little effect on petroleum product prices, all else remaining the same. On the other hand, if inventories were on the low side of or below the normal range—the result of other supply shortfalls or unexpectedly high demand—the additional refinery outage would be more likely to cause significant price increases. The size of the supply disruption relative to available inventories, as well as to the size of the refining sector, can also influence how prices respond. For example, if a large refinery outage were to occur in the Gulf Coast refining region, the large volume of inventories and the large number and capacity of other refiners relative to that refinery’s production would likely mean that the effect on prices of petroleum products would be small. Similarly, the availability of large stocks of gasoline in Europe, often less than 1 week away by tanker to the U.S. East Coast market, probably insulates the latter market from extreme price fluctuations in the event of a domestic supply shortfall. On the other hand, if the refinery outage were to occur in the West Coast, where one refinery’s production would be significantly larger relative to available inventories and total sector capacity, a significant price response is more likely. For these reasons, lower than normal inventories are frequently cited as a factor in price run-ups of petroleum products. For example, the Northeast diesel price spike of January 2000 was preceded by lower than normal inventory buildup, and the California Energy Commission cited relatively low inventory levels as one of several contributing factors to the gasoline price spike in the spring of 2006.

Energy Markets

37

U.S. SUPPLY INFRASTRUCTURE IS CONSTRAINED IN KEY AREAS AND LIKELY TO BECOME INCREASINGLY CONSTRAINED, THEREBY INCREASING PRICES AND PRICE VOLATILITY UNLESS TIMELY INVESTMENTS ARE MADE The nation’s crude oil and petroleum product supply infrastructure is constrained in key areas and may become inadequate to handle future volumes of petroleum products and biofuels unless sufficient investment is undertaken. Inadequate supply infrastructure can lead to higher prices and price volatility during supply disruptions or unexpected increases in demand because the supply infrastructure cannot handle the changed or increased delivery of fuels. However, the extent of the problem and the prospect for the future of the supply infrastructure is uncertain, in part because there has been no comprehensive study done to assess infrastructure adequacy. There are many planned infrastructure expansions that could alleviate the stress on the system to some extent. However, a complex approval process— involving numerous federal, state, and private entities—and other factors increase the time and cost of building and maintaining infrastructure.

The Nation’s Supply Infrastructure is Constrained in Key Areas and Likely to Become More Constrained Industry and agency officials report that key crude and petroleum product pipelines are constrained and operating at or near capacity. As the Secretary of Energy noted in a December 4, 2007 discussion with industry media, the U.S. energy infrastructure system— including oil pipelines—is “pressed,” and it is important that pipeline and other energy infrastructure owners maintain their assets effectively, to maintain adequate supplies. Both DOT and industry officials report a systemic lack of pipeline capacity in the supply infrastructure system in key states including Arizona, California, Colorado, and Nevada, and note existing pipeline supply infrastructure is insufficient to carry the commensurate volumes of petroleum products and crude oil needed to meet growing demand. Industry officials told us that pipelines in the Southwestern region, such as Arizona and Las Vegas, have reached maximum utilization, or become “constrained.” For example, industry experts told us that a new petroleum product pipeline from the Gulf of Mexico to El Paso is already approaching full capacity. Denver’s petroleum product pipelines have also become generally constrained and unable to meet increased gasoline demand for summer travel. This raises the cost of delivering petroleum products to Denver; in instances when pipelines are full, shippers must make alternate shipping arrangements by more costly rail or truck. Further, a key petroleum product line from the U.S. Gulf Coast to North Carolina is reportedly constrained, thereby increasing delivery costs for petroleum products in that region, and key petroleum product pipelines radiating outward from the major refining center of Houston are also reportedly constrained. Finally, in certain areas, pipeline infrastructure to support certain demand or production centers’ needs does not exist. For example, there are no petroleum product pipelines into Florida. Additionally, despite strong demand in California, the existing petroleum product pipelines support the flow of product from California to other Southwestern states, but no petroleum pipelines flow into California from other regions.

38

United States Government Accountability Office

Industry representatives and federal studies also report that many of the nation’s port facilities are operating at or near capacity. For example, one-fourth of the ports in a U.S. Maritime Administration (MARAD) survey described their infrastructure impediments as “severe.” Officials from the interagency U. S. Committee on the Maritime Transportation System, which includes MARAD, the National Oceanic and Atmospheric Administration, and the U.S. Army Corps of Engineers told us that U.S. ports and waterways are constrained in capacity and utilization, and anticipate marine supply infrastructure will become more constrained in the future. The Ports of Los Angeles, Long Beach, Oakland, Houston, Savannah, and Charleston reported congestion and emphasized in a 2005 report that they are experiencing higher than projected growth levels. The capacity of the supply infrastructure not keeping pace with increasing demand in certain areas has raised concerns about the adequacy of the infrastructure to accommodate expected increasing volumes of crude oil and petroleum products. Population increases in the West and South are expected to increase the need for pipelines, marine transportation, and capacity utilization there. DOT reports that already high pipeline capacity utilization levels may not meet growing demands unless significant expansion occurs. The situation is similar for the U.S. marine infrastructure. In a 2005 report, MARAD evaluated the status of U.S. ports and waterways and concluded that domestic marine transport supply infrastructure will become more constrained in the future. As imports of petroleum products are projected to increase by over 80 percent by volume between 2004 and 2030, according to EIA, this anticipated demand growth will challenge a marine transport system that is already operating, in some instances, at the limits of its capacity. The introduction of biofuels will also increase the strain on the existing supply infrastructure. For example, ethanol-producing plants tend to be relatively small near the sources of biofeedstocks—currently mostly corn—used to make ethanol. At present, the ethanol produced by these plants, unless they are located next to rail facilities, are typically trucked to central rail loading facilities and then shipped to demand regions on unit trains— trains whose cars are entirely made up of a single product and typically going to a single destination. Trucking biofuels to these central locations is costly and also uses petroleum products, thereby reducing the volumes of these latter fuels the ethanol can displace. Experts we spoke with generally agreed that eventually a more efficient collection system will likely be built—probably consisting of feeder pipelines—to connect the relatively small ethanol plants to major rail or supply and demand centers. Nonetheless, according to DOE, the existing petroleum product pipelines are currently not configured to transport ethanol from regions where it is currently produced to regions where it is consumed. Because pipelines are ultimately the cheapest form of domestic shipment of petroleum products and crude oil, it may make sense to ultimately ship ethanol through the pipeline system, and existing or new petroleum pipelines could be used in certain areas to transport ethanol if ongoing efforts by operators to identify ways to modify their systems to make them compatible with ethanol or ethanol-blended gasoline are successful. In addition, as discussed previously in this report, a proliferation of biofuel blends in this country will require additional variations in the blends of petroleum products that are mixed with these biofuels. Pipeline companies report that varying fuel specifications complicate petroleum product delivery and supply infrastructure systems by requiring separate storage and increasing the complexity of the distribution system. Also, pipeline operators told us that sending more and smaller batches of these special blends has slowed the flow of fuels through

Energy Markets

39

pipelines because pulling off more and smaller batches of fuels requires a slower speed to not miss significant parts of these batches. However, when we asked, these pipeline operators did not offer any quantification of the extent to which effective tank capacity reduction or pipeline slowness has occurred.

Infrastructure Disruptions Lead to Increases in Prices and Price Volatility and Constraints in Supply Infrastructure Could Exacerbate Price Effects A constrained supply infrastructure can be a major factor influencing prices of petroleum products during supply disruptions. For example, during the rupture in the Kinder Morgan pipeline in Arizona in August 2003, Arizona’s gasoline prices rose by about 45 cents during the 3-week period ending on August 25, 2003. Due to the connectivity of the pipeline network among California, Arizona, and Nevada, the disruption not only caused prices to spike in Arizona itself, but the extra burden from Arizona’s demand also contributed to higher prices elsewhere in the West; during the disruption, California’s prices rose by 40 cents to peak at $2.10, and Washington, Nevada and Oregon all experienced price increases of over 30 cents per gallon.[21] Any constraint in the supply infrastructure can reduce supply reliability by making it more difficult to reallocate supplies in response to even relatively minor disruptions in the supply and distribution system. In this way, a constrained supply infrastructure could increase price volatility and exacerbate price effects due to disruptions. When certain localities are inadequately served by pipelines or reasonably priced marine supply infrastructure, alternative transport modes tend to be more costly, leading to higher prices for consumers. For example, since relatively few pipelines connect the West Coast with other regions, some supplies of petroleum products and crude oil must be shipped by truck or barge from other domestic regions or by tanker from foreign countries; such modes of transport are slower and more costly than via pipelines. For example, it can take around 2 weeks for a vessel to travel from the Gulf Coast to Los Angeles port—including transit time through the Panama Canal. This can increase recovery time from an unplanned refinery outage, other supply disruption, or an unanticipated surge in demand, thereby leading to higher or longer-lasting price spikes. Federal agency officials and industry experts told us that the slow permitting process and corresponding delays in infrastructure development could lead to higher and more volatile petroleum product prices in the future. For example, while the recent expansion of pipeline capacity from the Gulf Coast to El Paso, following the opening of the Longhorn pipeline in June 2004, has been expected to ease the infrastructure constraint on Arizona’s petroleum product supplies, permitting impediments continue to perpetuate the lag between the growth of demand for petroleum products on the West Coast on the one hand and the growth of the pipeline capacity to move products to the region on the other. The California Energy Commission has recently stated that similar constraints on marine infrastructure expansions to accommodate future growth in demand for imports of petroleum products will be a major challenge for the West Coast. Such failure of the region’s supply infrastructure to handle the requisite volumes of petroleum products to meet rising demand will continue to contribute to the persistence of higher and more volatile prices in the West Coast compared to other regions.

40

United States Government Accountability Office

We were unable to assess the extent of supply infrastructure constraints or the impacts of these constraints on prices and price volatility, in large part because there is no central source of data that tracks system bottlenecks. Information that would indicate whether a pipeline is operating at or near capacity is also not collected in a central location by federal agencies or industry trade groups. These data would include pipeline throughputs, measured by the amount of product flowing into a pipeline and the volume of output received at key market locations. Companies are not required to report such information. By contrast, FERC requires natural gas pipelines to report, via their web sites, throughput information that allows regulatory, public, and private entities to track bottlenecks and identify where shortages in supply, or system constraints, affect regional prices. A number of studies and analyses of constraints in natural gas pipelines have quantified the effects on natural gas prices. For example, EIA routinely uses natural gas pipeline capacity and outages in making projections about natural gas prices. These data on natural gas pipeline capacity and the flow of natural gas are collected and evaluated to determine the reliability of the infrastructure to meet demand, and it is well understood that constrained pipelines lead to higher natural gas prices and can even lead to disruptions of service in severely constrained cases.[22] We recognize there are differences between the natural gas industry on one hand and the petroleum industry on the other, particularly because of the fact that the former industry evolved under a rate-regulated utility framework, while the petroleum industry did not. Specifically, under rate regulation, the former requirement that utilities meet all demand at their regulated prices at any point in time necessitated the monitoring of supply and infrastructure constraints that could cause a failure of service. By contrast, petroleum product prices have largely not been regulated, and prices have generally been allowed to adjust to equilibrate supply with demand at any point in time. Further, we are not suggesting in this report that petroleum product markets should be regulated like natural gas or any other markets. However, these historical regulatory differences notwithstanding, we believe that it is important to understand the extent to which constraints on the current petroleum product supply infrastructure affect prices as well as the adequacy of the infrastructure to meet growing demand. Federal agencies, industry experts, and Congress have all recognized this as a priority. For example, industry consultants and agency officials have acknowledged the importance of a system-wide study of pipeline capacity constraints and regulatory impediments to future investment. In addition, DOT officials have stated that the extent of capacity restrictions in the nation’s pipeline infrastructure is becoming more apparent, that the current regulatory mechanisms may not lead to appropriate reinvestment in the industry. In June 2006, DOT put forth a proposal and in December 2006 Congress passed legislation that mandated the Secretaries of Energy and Transportation to conduct periodic analyses of the adequacy of the nation’s pipeline supply infrastructure. The first report to Congress of the results of such an analysis is required by June 2008.[23] The language for the mandate stated that “such analyses should identify areas of the United States where unplanned loss of individual pipeline facilities may cause shortages of petroleum products or price disruptions and where shortages of pipeline capacity and reliability concerns may have or are anticipated to contribute to shortages of petroleum products or price disruptions. Upon identifying such areas, the Secretaries may determine if the current level of regulation is sufficient to minimize the potential for unplanned losses of pipeline capacity.” Despite widespread recognition that such a study is needed to fully identify the extent of infrastructure inadequacy and the impact on prices, to date, no such analysis has been undertaken. DOT and DOE officials told us that

Energy Markets

41

they were not appropriated funds specifically to do the mandated analyses and that the agencies have not re-allocated other funds for this, although DOE told us in its comments that DOE and DOT staff have met to discuss how this work could be approached. Given that the study has not begun, it seems highly unlikely that the agencies will be able to meet their June 2008 deadline for reporting to Congress.

Expansions in Supply Infrastructure Are Planned, but High Construction Costs, Investment Risk, and a Complex Regulatory Environment Can Deter or Delay These Needed Infrastructure Investments There are many private sector plans to expand the supply infrastructure, and if implemented in timely fashion, these plans could significantly alleviate the stresses on the system. For example, there is a long-anticipated project for a 500-mile petroleum product pipeline expansion from Louisiana to Georgia, several plans for new crude pipelines to accommodate the expected increased flows of Canadian oil sands, as well as other crude and refined product pipeline plans to meet more localized needs. However, many such plans are in a conceptual stage and/or subject to permitting approval and other possible complications. Thus, industry representatives told us, it is difficult to determine how many of the industry plans for new construction or expansion of existing pipelines will be realized. However, the high cost of construction, uncertain investment climate, and complex regulatory environment increase the time it takes to build this supply infrastructure and raises risk and investment costs. With regard to construction costs, a shortage of skilled labor and specialized equipment to perform the work, and high prices of steel and concrete have increased construction costs and the time it takes to expand the nation’s supply infrastructure system. For example, pipeline companies and other industry experts we spoke with said that major pipeline expansion and construction projects take anywhere from 2 to 15 years to complete and currently cost about $1 million per mile to build. With regard to the uncertain investment climate, pipeline companies and industry experts told us that uncertainty about petroleum product demand, biofuel development and shipping, and future changes to fuel specifications complicate the decisions about where and when to build new or expand existing infrastructure. Regulations governing pipeline and other infrastructure expansions, including regulations governing water and air pollution, endangered species protection, and public safety, have evolved to protect the environment and ensure public safety. However, there can be tension between these goals and the goals of ensuring adequate energy supplies and keeping prices down. For example, in order to build a new pipeline or significantly expand capacity or upgrade an existing pipeline, companies must first navigate a mixed and sometimes complex jurisdiction of federal, state, and local regulators, as well as secure right of way approval from the necessary landowners whose lands will be crossed by the pipeline. At the federal level alone, as many as 11 agencies may be involved in granting approval to build new pipeline projects. In addition, industry experts told us that some potential market entrants have had difficulty meeting permitting requirements and are often unable or unwilling to wait out lengthy delays in obtaining permits, such as when two companies in southern California reportedly recently backed out of plans to build storage terminals there after trying to complete the federal, state and local approval processes. A study conducted for Association of Oil Pipe Lines, an FTC report on gasoline prices, and industry officials told us

42

United States Government Accountability Office

that building or expanding pipelines has become increasingly difficult in certain situations. For example, a major pipeline operator encountered federal and local legal and regulatory issues that delayed for 10 years the development of a key pipeline from the Gulf Coast to El Paso, Texas. As a result of such delays and impediments to investment, regional demand that could support new pipeline capacity must be served by more costly transportation modes for years, as has been the case in parts of the Rocky Mountains and Southwest and West Coast regions. Finally, an uneven balance of costs and benefits of expansion for various entities can also contribute to declining investment in supply infrastructure by certain entities. For example, DOT reports that common carrier pipelines achieve only modest returns from relieving constrained pipeline capacity. However, it reports consumers would benefit proportionately greater through the enhanced competition resulting from the increased capacity of new pipeline investments. Pipeline companies, on the other hand, report they will expand when sufficient demand is secured, particularly through the “presale” of capacity in the proposed pipeline. Ideally, the permitting and approval process should be streamlined without sacrificing the important protections provided by regulatory oversight. Industry and federal agency officials have pointed out that a federal model exists for this in the permitting process for interstate natural gas pipelines. Specifically, FERC facilitates expansions and construction of natural gas pipelines by serving as the lead agency to process company permit applications, conduct the required environmental impact study, and coordinate the timing of other necessary permits that fall under the purview of various federal agencies. In addition, FERC authorizations convey the right of eminent domain to pipeline builders to resolve specific right of way issues in the event an agreement cannot be reached between a landowner and a project sponsor. FERC officials told us that although its authorizations convey the right of eminent domain, pipeline companies rarely have to exercise it because its existence is usually sufficient to get landowners to negotiate a solution with pipeline builders. Streamlining the federal regulatory process with regard to crude oil and petroleum product pipeline repairs has already begun in response to a federal statute passed in 2002 to coordinate environmental reviews and permitting needed for pipeline repairs and more clearly define federal roles in the pipeline repair process.[24] However, this streamlined federal process has not been applied to constructing new crude oil or petroleum product pipelines or significantly upgrading or increasing capacity of existing pipelines.[25]

CONCLUSION The choices the United States and other countries make about how to ensure sufficient supplies and stable prices of petroleum products and other fuels such as ethanol and biofuels will greatly influence energy prices in the United States. For biofuels in particular, cost and availability will depend in part on how well international, federal, state, and local governments coordinate their biofuel standards and methods of integrating them with petroleum products. Harmonizing fuel specifications worldwide, while continuing to allow for regional differences in fuels specifications that are there to meet specific environmental or vehicle performance goals, would make it easier to refine and transport common blends, streamline delivery, increase opportunities for trade, provide additional sources of supply, and potentially reduce prices and price volatility. However, if the world and the United States end

Energy Markets

43

up with numerous different biofuel blends—as appears to be happening under existing plans and mandates—this could expand the array of incompatible gasoline and diesel blending stocks and final blended products that cannot be interchanged at the retail level, reducing opportunities for trade. In addition, these products will have to be segregated during shipment, further straining the supply infrastructure. Unless the supply infrastructure catches up and keeps up with these changes, the domestic energy supply will be less secure and prices will tend to be higher or more volatile. Even without these changes, rising demand for crude oil and petroleum products over the last 25 years has challenged the supply infrastructure for these commodities in certain areas, leading to higher prices during supply disruptions or during periods when pipelines or ports lack sufficient capacity to transport the products suppliers wish to ship. As noted by the Secretary of Energy in December 2007, the U.S. energy infrastructure system—including oil pipelines—is “pressed,” and it is important that pipeline and other energy infrastructure owners maintain their assets effectively, in order to maintain adequate supplies of energy. In the absence of a comprehensive analysis of the likely weaknesses in our infrastructure, policy makers and regulatory agencies involved in overseeing the safety and adequacy of supply infrastructure remain in the dark about the extent of these problems and their effects on prices of petroleum products. Further, as demand for petroleum products and biofuels grows, the existing system may become increasingly constrained and need to be upgraded and expanded to handle greater and different product flows. Because federal and state agencies and other entities will be involved in approving such upgrades and expansions, it is essential that they be well informed as to the current state of the supply infrastructure and the areas in most critical need of further investment. Furthermore, the lack of a lead agency to streamline the complex and costly permitting process for U.S. supply infrastructure construction or expansion projects and the lack of ability of federal agencies to convey the power of eminent domain in cases where conflicts over infrastructure placement cannot be resolved may deter potential market entrants from investing in much-needed upgrades in a timely fashion. As a result, we could end up with less security of supply and higher and more volatile prices in the future.

RECOMMENDATIONS FOR EXECUTIVE ACTION To better monitor and evaluate the development of our nation’s supply infrastructure systems, as well as to facilitate the continued tradability of products across domestic and global markets and to ensure that gasoline supplies from Europe and elsewhere remain compatible with U.S. gasoline specifications, we are making a number of recommendations that, if adopted, should improve prospects for the future security of petroleum product supplies and price stability. •

To avoid additional proliferation of differing fuel specifications that would further burden the existing supply infrastructure and create impediments to trade, we recommend that the Secretary of Energy coordinate with EPA and other relevant federal agencies, states, IEA, the European Union, and other foreign entities to encourage development of biofuels and petroleum products standards and blending practices that maximize the fungibility of these fuels and minimize the spread of

44

United States Government Accountability Office

•

•

differing fuel types that would further strain the supply infrastructure, while recognizing that some fuel differences to reflect local environmental requirements, engine performance, or other factors are likely beneficial. To comprehensively analyze the U.S. supply infrastructure’s capacity to accept, handle, and transport the increasing volumes and types of petroleum products and biofuels expected to traverse its system, we recommend the Secretaries of Energy and Transportation undertake the comprehensive study of existing and projected increases to the infrastructure system—including terminal capacity and pipeline throughputs—to evaluate whether future demand is likely to be met by existing infrastructure and planned increases as mandated by Congress in 2006. To the extent that the data to comprehensively conduct such analyses may at present not be collected, the Secretaries should consider evaluating the merits of enhancing the reporting of utilization and throughputs, perhaps using natural gas pipeline and storage reporting requirements as a model. In conjunction with the completion of the first comprehensive study of the supply infrastructure, we recommend the Secretary of Transportation work with DOE, FERC, EPA, and other federal agencies to evaluate the feasibility and desirability of designating a lead federal agency, with authority to convey the power of eminent domain, to coordinate across agencies and streamline the permitting and siting process for crude oil and petroleum product interstate pipeline expansions, upgrades, and new construction, using FERC’s role with natural gas pipelines as a model. If this is found to be feasible and desirable, we recommend the aforementioned agencies work together to determine which agency should take the lead role and to prepare a legislative proposal for Congress to provide any additional authority needed to implement this recommendation.

APPENDIX I: SCOPE AND METHODOLOGY The Chairman and a member of the Senate Commerce, Science, and Transportation Committee asked GAO to evaluate trends and effects on petroleum product prices in (1) international trade of petroleum products; (2) refining capacity and intensity of refining capacity use internationally and in the United States; (3) international and domestic crude oil and petroleum product inventories; and (4) domestic crude oil and petroleum product supply infrastructure, particularly pipelines and marine transportation. To address the first objective, we examined data from the Department of Energy’s (DOE) Energy Information Administration (EIA) and the International Energy Association (IEA) to evaluate trends in the international trade flows for crude oil and petroleum products and their price correlations over time at international trading hubs. In addition, IEA data were used to calculate total global imports and exports of crude oil and petroleum products as well as for key global regions including Europe, Asia and the United States. We met with more than 20 oil industry companies—including refiners and pipeline companies—a number of financial and investment corporations, more than 25 industry groups, and more than 15 domestic and international government agencies to corroborate trend analyses, reports, and data. We conducted audit work in various locations in Texas, California, New York, and Washington, D.C., as well as Belgium, France, Germany, and the United Kingdom to obtain industry’s

Energy Markets

45

perspective on recent trends in the international trade of petroleum product as well as prospective trends going forward. In addition, we analyzed EIA and New York Mercantile Exchange, (NYMEX) data on historical spot and futures prices for crude oil and petroleum products at international and domestic trading hubs to see how price volatility has changed over time. To address the second objective, we assessed trends in refining capacity, refining capacity additions, utilization, complexity, and planned investments using IEA, EIA, and Oil and Gas Journal data, and determined the data were sufficiently reliable for our purposes. We met with more than 20 oil-industry companies—including refiners and pipeline companies—a number of financial and investment corporations, more than 25 industry groups, and numerous staff and officials of more than 15 domestic and international government agencies in California, Texas, New York, Washington D.C. ,as well as Belgium, France, Germany and the United Kingdom to corroborate trend analyses, reports and data. We also reviewed and analyzed trends in refinery investment, operating costs, and profitability in the U.S. and internationally, using literature and data on U.S. and international refining practices, trends and forecasts, and interviewed experts on these trends. To address the third objective, we used data from EIA and IEA on crude oil and petroleum product inventories and projected demand to conduct international, U.S. total domestic, and U.S. Petroleum Administration for Defense District (PADD) inventory trend analysis on inventories in absolute terms and in “days forward cover” terms. We analyzed NYMEX and other futures market data, as well as EIA data, to observe the effects of the expected future price for crude oil on inventory holding decisions. To collect these data, we conducted a site visit to meet with industry and government representatives in Belgium, France, Germany, and the United Kingdom to gain information about the European Union’s policy of maintaining strategic petroleum product reserves and their effects on price levels and price volatility. To address the fourth objective, we interviewed federal and state agencies that oversee the economic, safety, and environmental impacts of pipelines and marine transportation on current and future utilization capacity of the petroleum product infrastructure. Where possible, we collected and analyzed data on the age of the pipeline and marine infrastructure system, capacity, throughputs, and constraints. We compared data reporting requirements for petroleum products with reporting requirements for liquefied natural gas, and identified differences in such reporting requirements. We spoke with common carrier pipeline operators, port authorities, government entities, and trade association and consumer advocate groups to gain their perspectives on supply infrastructure investment, capacity utilization levels, and potential system constraints. We also reviewed previous relevant GAO reports and testimonies, and Department of Energy and Department of Transportation reports. In addition, we examined reports and data from supply disruption case studies to examine those cases’ impact on infrastructure, prices, and price volatility. During our audit work we consulted with the following entities: •

We met with the following oil industry companies, including refiners, supply infrastructure and oil service companies: BP; Buckeye Partners; Chevron Corporation; ConocoPhillips Company; ExxonMobil Corporation; Fluor Corporation; Frontier Oil Corporation; Hess Corporation; Holly Corporation; Kinder Morgan Energy Partners; Longhorn Partners Pipeline; Magellan Midstream Partners;

46

United States Government Accountability Office

•

•

•

Marathon Oil Company; Mid-continent Express Pipeline; Oiltanking GmbH; Paramount Petroleum Corporation; Plains All American Pipeline L.P.; RaceTrac Petroleum, Inc.; Sunoco, Inc.; TEPPCO Partners L.P.; Tesoro Corporation; UOP LLC; Valero Energy Corporation. We met with the following financial organizations: Deutsche Bank; Goldman, Sachs and Co; JP Morgan Chase Bank; Morgan Stanley; New York Mercantile Exchange, Inc. (NYMEX). We met with the following industry groups and expert institutions: Allegro Energy Consulting; American Association of Port Authorities; American Petroleum Institute (API); Association of Oil Pipe Lines (AOPL); Conservation of Clean Air and Water in Europe (CONCAWE); Consumer Federation of America; Energy Analysts International, Inc.; European Petroleum Industry Association (EUROPIA); Global Insight, Inc.; Institut Francais du Petrole (IFP); Muse Stancil and Co.; National Association of Regulatory Utility Commissioners; National Petrochemical and Refiners Association; Oil and Gas Journal; Petroleum Marketers Association of America; Pipeline Safety Trust; PIRA Energy Group; Purvin and Gertz, Inc.; Stillwater Associates LLC; Turner, Mason and Company; the Rabinow Consortium, LLC; UK Petroleum Industry Association; Union of European Petroleum Independents (UPEI); University of California Energy Institute; Western States Petroleum Association; Wood Mackenzie Research and Consulting. With regard to government and agency sources, we met with the following U.S. agencies and governmental institutions: Department of Defense, including the Army Corps of Engineers; Department of Energy, including the Energy Information Administration; Department of State; Department of Transportation, including Pipeline and Hazardous Materials Safety Administration (PHMSA); Department of Homeland Security; Federal Energy Regulatory Commission; Federal Trade Commission; Interagency Committee on Marine Transportation; Oak Ridge National Laboratory. We met with the following state and local governmental agencies: California Energy Commission (CEC); California Environmental Protection Agency Air Resources Board (CARB); Hawaii Energy Planning and Policy Branch; Port of Houston Authority. We met with the following international government and multilateral organizations: European Commission Directorate-General for Energy and Transport; EBV (German Stockholding Agency); French General Directorate for Energy and Raw Materials; International Energy Agency (IEA); International Monetary Fund (IMF).

The report primarily uses data from the domestic and international wholesale petroleum product and crude oil markets. In contrast to retail markets, wholesale prices do not generally include extra costs such as federal and state taxes, distribution and marketing expenses and profits. In every case for the data used in this report, we assessed and determined that the data were sufficiently reliable for our purposes. We performed our work from August 2006 through September 2007 in accordance with generally accepted government auditing standards.

Energy Markets

APPENDIX II: COMMENTS FROM THE DEPARTMENT OF ENERGY

47

48

United States Government Accountability Office

APPENDIX III: COMMENTS FROM THE FEDERAL ENERGY REGULATORY COMMISSION

Energy Markets

49

REFERENCES [1]

GAO, Energy Markets: Mergers and Other Factors That Influence Gasoline Prices, GAO-07-894T (Washington, D.C.: May 23, 2007); GAO, Energy Markets: Factors Contributing to Higher Gasoline Prices, GAO-06-412T (Washington, D.C.: Feb. 1, 2006); GAO, Energy Markets: Gasoline Price Trends, GAO-05-1047T (Washington, D.C.: Sept. 21, 2005); GAO, Motor Fuels: Understanding the Factors That Influence the Retail Price of Gasoline, GAO-05-525SP (Washington, D.C.: May 2005). [2] FERC also serves as the lead agency in coordinating the permitting process across federal agencies and can similarly convey the right of eminent domain for electricity transmission lines. [3] The Organisation for Economic Co-operation and Development is a group of 30 countries committed to democracy and the market economy to support sustainable economic growth, maintain financial stability, and assist other countries’ economic development. These countries are Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Spain, Sweden, Switzerland, Turkey, United Kingdom, and the United States. However, we consider the United States separately for the purposes of this report. [4] GAO, Gasoline Markets: Special Gasoline Blends Reduce Emissions and Improve Air Quality, but Complicate Supply and Contribute to Higher Prices, GAO-05-421 (Washington, D.C.: June 2005). [5] The Strategic Petroleum Reserve (SPR) is a federally maintained stockpile of about 700 million barrels of light crude oil for use in the case of a major disruption of oil supplies. [6] Access to the rail market is limited and tanker trucks’ expenses depend on distances traveled. [7] By petroleum products, we refer to primarily gasoline, diesel, jet fuel, heating oil. Most petroleum products and crude oil are shipped primarily by pipeline within the United States. Imports of petroleum products and crude oil, however, travel to the United States mainly over sea by vessel. [8] FERC also collects administrative, financial, and operational information on crude oil and petroleum product pipeline companies. [9] Figures represent trade originating or ending in OECD member nations, including trade between OECD nations, from OECD nations to non-OECD nations, and from nonOECD nations to OECD nations. Because figures include some trade from OECD nations to other OECD nations, such trade is counted as both an import and an export and therefore includes some duplication of counting. Furthermore, figures do not account for trade between non-OECD nations and therefore understate the total global trade of these products. [10] Total gasoline includes both finished motor gasoline and motor gasoline blending components. [11] Imports of distillate fuels and jet fuel have also risen in the last 20 years, while imports of residual fuel oil have declined.

50

United States Government Accountability Office

[12] The graphic shows that imports remained significantly higher than in the same months during the previous year at least through January 2006. This was likely the result of lasting damage to U.S. refining production caused by the hurricanes. [13] States are subject to some federal requirements in setting biofuel policies. Currently, EPA has determined that only blends of up to 10 percent ethanol are allowed in conventional gasoline vehicles and blends of up to 85 percent ethanol are permitted in flexible fuel vehicles. However, the state of Minnesota and the Renewable Fuels Association, as well as DOE are developing research and tests to gather the data required to facilitate EPA certification of fuel blends up to E15 or E20. [14] The price increases were measured in gasoline prices in Chicago relative to Houston prices. [15] See, for example, GAO’s Motor Fuels: Gasoline Prices in the West Coast Market, GAO-01-608T, (Washington, D.C.: Apr. 25, 2001). [16] In the United States, inventory data reported in this report refer only to privately held stocks, not the federally held crude oil and heating oil strategic reserves. As will be discussed later in this report, this is not true of some other OECD member inventory data. [17] Unlike in the United States, where the federal government holds strategic stocks of primarily crude oil—but also a relatively small stock of fuel oil in the U.S. Northeast Home Heating Oil Reserve—European countries hold a large fraction of their strategic stocks in petroleum products, including gasoline and certain distillate fuels. Some European countries require private companies to maintain these stocks. [18] It is also not clear that the benefits of collecting and maintaining such data outweigh the costs. Evaluating these trade-offs was beyond the scope of this report, but such an evaluation would have to be made before making a decision to collect a broader range of inventory data. [19] Assessing the relative importance of these factors with any precision would be very difficult and we did not undertake this task in this report, so the list of factors should not be seen as a ranking of those factors in any way. [20] Energy Information Administration, Petroleum 1996: Issues and Trends, (Washington D.C., September 1997). [21] Note that there may have been additional factors influencing prices during this period, so we are not asserting that the pipeline outage was responsible for the entire change in prices. [22] It should be noted that whether or not the benefits of collecting and maintaining such data outweigh the costs is unknown. Evaluating these trade-offs was beyond the scope of this report, but such an evaluation would have to be made before making any decision to collect a broader range of pipeline or other infrastructure data. [23] Pub. L. No. 109-468, §8. [24] Pub. L. No. 107-355, § 16 [25] In 2006, DOT identified the need for additional Congressional authority to reduce the regulatory burden on companies trying to construct new pipelines or repair existing ones. Specifically, DOT proposed legislation that, according to DOT, would among other things provide “minimal authority” to assist pipeline operators in overcoming state and local-level impediments to constructing new pipelines and would further streamline the permitting process for pipeline repairs. At this time, Congress has not

Energy Markets

51

provided this additional authorization. DOT’s proposal did not call for a federal agency to have the authority to convey the power of eminent domain in cases where conflicts over infrastructure placement cannot be resolved but it would have authorized the Secretary of DOT to “designate an ombudsman to assist resolving disagreements between Federal, State, and local agencies and pipeline operators arising during agency review of pipeline repairs and hazardous liquids pipeline construction projects…”

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 53-83

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 2

THREE-DIMENSIONAL SIMULATION OF BASE CARRIER TRANSPORT EFFECTS IN BACK SIDE POINT CONTACT SILICON SOLAR CELLS* K. Kotsovos† and K. Misiakos Institute of Microelectronics, NCSR Demokritos, Attiki, Greece

ABSTRACT This work presents a theoretical investigation of rear junction point contact silicon solar cells through three-dimensional numerical simulation based on the solution of minority and majority carrier transport equations in the base of the cell. The device series resistance is evaluated through the simulated current-voltage (IV) curves under AM1.5 illumination conditions and its dependence on back contact geometry is examined. Results are presented which show the influence of the majority carrier transport in the base to the solar cell performance. A comparison is also performed with two other similar types of point contact solar cells, one with the emitter located on the front surface and the other on both surfaces, as well as with a conventional solar cell structure.

I. INTRODUCTION Rear point contact (locally diffused) silicon solar cells with backside p/n junctions are structures which have already shown their promising potential in solar energy production, reaching very high conversion efficiency (27.5%) under concentrated illumination [1]. Although these devices were ideal for concentrator applications due to low series resistance and surface recombination losses, they have some additional interesting advantages, *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Institute of Microelectronics, NCSR Demokritos, P. O Box 60228 153 10 Aghia Paraskevi, Attiki, Greece Tel: (+30)2106503113, Fax: (+30)2106511723, E-mail: [email protected]

54

K. Kotsovos and K. Misiakos

compared to typical solar cells designed for one-sun operating conditions. Specifically, since the metallization grid lies entirely on the back surface, there is no shading loss on the illuminated surface of the solar cell, while the interconnection of individual cells into modules is more easily implemented. However, this optimized solar cell design was considered to be too complex for use at low concentrations, so a simplified structure was proposed by Sinton et al. [2], suitable for cost-effective production. Therefore, SunPower Corporation has developed a process for that purpose, providing solar cells fabricated on high quality FZ substrates with efficiencies greater than 20% under normal sunlight [3]. The choice of high quality material is necessary for this type of solar cells, since the photogenerated carriers need to reach the back surface in order to be collected. Results of theoretical simulations regarding the back contact structure have already been published in the literature. A 3D model based on the solution of semiconductor transport equations using a variational approach has been developed by Swanson [4-5], which was applied in order to optimize the back point contact solar cell design under concentrated illumination. An optimization of the interdigitated back contact cell was performed by Chin et al. [6], while the simulated efficiency limit of this cell was calculated by Ohtsuka et al. [7] by 3D simulations. Epitaxial layer transfer has also been proposed as an alternative way to produce back contact solar cells [8], where this method is used to create thin silicon films on foreign substrates and a two-dimensional model was applied for this case. The purpose of this work is a theoretical investigation of back junction point contact solar cells by means of numerical three-dimensional simulation based on the solution of minority and majority carrier transport equations in the base of the cell. The method is based on the transformation in x and y dimensions of the basic partial differential equations through 2D Fast Fourier Transform (FFT). In Fourier space these equations become algebraic in Kx and Ky (the transformed x, y variables), thus reducing to ordinary differential equations with respect to z, that can be solved in analytical form. The basic assumption for such a problem reformulation is planar geometry and low injection. The solution of the transport equations under illumination conditions provides the device IV characteristics and solar cell’s series resistance is extracted. This model was previously used [9] to simulate a structure similar to the PERL [10-11] solar cell, a device that is consisted of an emitter covering the front illuminated surface and point contacts in the back surface. The same method was later applied for the simulation of the double junction solar cell [12], a device with an additional emitter in the back surface. Since the back junction point contact solar cell and the previous two types of solar cells, are high efficiency structures, a direct comparison among them is performed. The influence of back contact size and spacing in solar cell performance is discussed in detail. The following section presents a description of the mathematical model of our method as applied on the point contact structures under consideration. The third section includes our simulation results and discussion.

Three-Dimensional Simulation of Base Carrier Transport Effects…

55

II. MATHEMATICAL MODEL- SIMULATION ALGORITHM

II.1. Assumptions –Device Geometry y

x z

w

a

d l

3l/4

Oxide

3l/4 Base

l/4

b

Contact Emitter

l/4

c

Oxide

Contact

d

w

d

Emitter

l Emitter

Base

Contact Oxide

Emitter

e

d Contact Emitter

Figure 1. a) Three dimensional back surface geometry of the simulated front junction, back junction and double junction devices b) Back junction structure, c) Back point junction structure (the locations of the diffused regions are shown in the insert) d) front junction structure e) Double junction structure. This pattern is repeated periodically in x and y directions with a period length l.

The base of the solar cells is considered to be under low injection conditions and assumed as homogeneous with thickness w, while the junctions are infinitesimally shallow. Photogeneration in the emitter regions is considered negligible, while their ohmic losses are neglected. We set as x, y the directions parallel to the junction while z is the perpendicular

56

K. Kotsovos and K. Misiakos

one. The geometry of the simulated devices is shown in figure 1. The base contacts on the back surface are assumed as squares with side length d, while the period length, or field length, of the repeated pattern as shown in figure 1(a) is l. Figure 1(b) illustrates the structure of the back junction solar cell, figure 1(c), 1(d) the corresponding point back junction and single front emitter devices and figure 1(e) the double junction solar cell. The dimensions of the back point junction and base contact of the cell as shown in figure 1(c) are the same. In addition, zero front surface reflectance is assumed, while light trapping is similar to pyramidal texture scheme.

II. 2. Minority Carrier Continuity Equation and Boundary Conditions The minority carrier continuity equation for a p-type base under low-level injection and steady state is given by

∇ 2 n( x, y , z ) =

n ( x , y , z ) G ( x, y , z ) − L2n Dn

(1)

where n(x,y,z) is the minority carrier concentration, Ln is the corresponding diffusion length, Dn the diffusion constant and G(x,y,z) the local generation rate.

II. 2.1. Boundary Conditions (Back Junction Structure) The following relation describes the boundary condition at the front of the oxide passivated solar cell’s surface as shown on figure 1(b)

J n ( x, y, w) = eDn

dn( x, y, z ) dz

z=w

= eS1 n( x, y,0)

(2)

where Jn(x,y,w) is the minority carrier diffusion current in the back surface and S1 is the recombination velocity in the area covered by the oxide. At the back surface, the boundary condition at the diffused contacts is also expressed by the minority carrier diffusion current, which depends on the surface recombination velocity in that area:

J n ( x, y, w) = eDn

dn( x, y, z ) dz

z=w

= −eS 2 n( x, y, w)

(3)

where Jn(x,y,w) is the minority carrier diffusion current in the back surface and S2 is the recombination velocity in the diffused contact areas, which is assumed as constant and given by the following expression

S2 =

J 0C N A eni2

(4)

Three-Dimensional Simulation of Base Carrier Transport Effects…

57

where J0C is the saturation current density in the diffused contacts, ni the intrinsic carrier concentration of the semiconductor and NA the base doping. The rest of the back surface area is covered by the junction, and the boundary condition is:

n( x, y, w) =

ni2 NA

⎛ ⎛ e(V + Vdrop ( x, y, w) ) ⎞ ⎞ ⎜ exp⎜ B ⎟⎟ − 1⎟ ⎜ ⎜ ⎟ KT ⎝ ⎠ ⎠ ⎝

(5)

where VB is the junction bias voltage and Vdrop is the voltage drop caused by the majority carrier flow through the base series resistance. This voltage drop is initially set to zero.

II. 2.2. Boundary Conditions (Back Point Junction Structure) The front surface of this structure, shown on figure 1(c) is covered by oxide, so relation (2) gives the expression of the boundary condition in that region. Expressions (3) and (5) define the boundary conditions in the back diffused contacts and the junction area respectively. The rest of the back surface is oxide passivated, so the boundary condition is defined by the minority carrier diffusion current

J n ( x, y, w) = eDn

dn( x, y, z ) dz

z=w

= −eS 3 n( x, y, w)

(6)

where S3 is the recombination velocity in the area covered by the back oxide.

II. 2. 3. Boundary Conditions (Front Junction Structure) The emitter of the front junction solar cell, which is illustrated in figure 1(d) covers the whole illuminated surface, so the boundary condition inside the junction is

ni2 n( x, y,0) = NA

⎛ ⎛ e(V + Vdrop ( x, y,0) ) ⎞ ⎞ ⎜ exp⎜ B ⎟⎟ − 1⎟ ⎜ ⎟ ⎜ KT ⎝ ⎠ ⎠ ⎝

(7)

At the back surface, the minority carrier diffusion current is determined by the surface recombination velocity, where in the diffused base contact regions is defined by relation (4), while in the oxide passivated surface has a constant value (S1). Therefore, the general form of the boundary condition at the back surface may be written as

J n ( x, y, w) = eDn

dn( x, y, z ) dz

z=w

= −eS ( x, y )n( x, y, w)

(8)

II. 2. 4. Boundary Conditions (Double Junction Structure) The double junction structure (figure 1(e)) is consisted of an emitter covering the whole front surface (as in the front junction device), so the boundary condition in that area is given by (7). In a similar way, the conditions in the back surface are expressed by relations (3) and

58

K. Kotsovos and K. Misiakos

(5) of section II.2.1. The front and back emitters of this device are biased with same voltage VB.

II.3. Majority Carrier Voltage Drop Equation The solution of continuity equation (1) may be used to obtain the voltage drop caused by the majority carrier flow. We begin from the current density relation for the majority carriers

J p = eμ p pE − eD p∇p

(9)

Charge neutrality in the semiconductor is assumed, so it follows that δp(x,y,z)=δn(x,y,z). Since the cell is operated under low injection, p≈NA, where NA is the base doping. Using these assumptions and with the aid of (1), we differentiate (9) ∇J p = eμ p N A ∇E − eD p ∇ 2 p ⇒ ∇J p = eμ p N A ∇E − eD p ∇ 2 n ⎛ n G ⎞ ⎟⎟ ⇒ ∇J p = eμ p N A ∇E − eD p ⎜⎜ 2 − ⎝ Ln D n ⎠ Dp ⎛ n ⎞ n ⎜⎜ − G ( z ) ⎟⎟ ⇒ − G = − μ p N A ∇E + Dn ⎝ τ n τn ⎠ D p − Dn ⎛ n G ⎞ ⎜⎜ 2 − ⎟ ∇E = μ p N A ⎝ Ln Dn ⎟⎠

The comparison of this equation with (1), gives a more compact expression

∇ 2V =

1 ( Dn − D p )∇ 2 n μpNA

(10)

where Dn, Dp are the diffusion constants for electron and holes respectively and μp is the hole mobility. The solution of this equation provides the voltage drop due to majority carrier flow and is subjected to the boundary conditions given in the next subsection.

II. 3.1. Boundary Conditions (Back Junction Structure) Since there is no total current flow in the oxide covering the whole front surface of the back junction solar cell, the majority carrier current value is exactly the opposite of the minority carrier equivalent:

J p ( x, y,0) = −eS1 n( x, y,0)

(11)

At the back surface, the diffused back contact areas are considered as the ground terminal, so the majority carrier voltage drop is zero:

Three-Dimensional Simulation of Base Carrier Transport Effects…

Vdrop ( x , y , w ) = 0

59 (12)

At the rest of the back surface area, covered by the rear junction, the majority carrier current is given by

⎛ ⎛ e(VB + Vdrop ( x, y, w)) ⎞ ⎞ ⎟⎟ − 1⎟ J p ( x, y, w) = J 0 ⎜⎜ exp⎜⎜ ⎟ KT ⎝ ⎠ ⎠ ⎝

(13)

where J0 is the emitter saturation current density. Expressions (11) and (13) can be converted as boundary conditions for the electric field E if we make use of (9) with the following way:

J p = eμ p pE − eDp∇p ⇒ J p = eμ p N A E − eDp∇n ⇒ J p = eμ p N A E −

Dp Dn

Jp + Jn ⇒ E =

Dp

Dn eμ p N A

Jn

(14)

The minority carrier current density Jn, which is required in expression (14) is obtained from the solution of the continuity equation described in section II. 2.

II. 3.2. Boundary Conditions (Back Point Junction Structure) The front surface of this structure is covered by oxide as in the case of the back junction structure, so expression (11) defines the boundary condition in that region. Relations (12), (13) also describe the boundary conditions inside the back diffused contact and junction areas respectively, while the rest of the back surface is covered by oxide, so the following condition holds

J p ( x, y, w) = eS 3 n( x, y, w)

(15)

As reported on the previous subsection, the majority carrier expressions may be converted to electric field boundary conditions by making use of (14).

II. 3.3. Boundary Conditions (Front Junction Structure) The boundary condition for the majority current at the front surface and inside the junction is written as:

⎛ ⎛ e(VB + Vdrop ( x, y,0)) ⎞ ⎞ ⎟⎟ − 1⎟ J p ( x, y,0) = − J 0 ⎜⎜ exp⎜⎜ ⎟ KT ⎝ ⎠ ⎠ ⎝

(16)

where J0 is the saturation current density in the emitter. The expression of the boundary condition in the oxide-covered part of the back surface is also given by

60

K. Kotsovos and K. Misiakos

J p ( x, y, w) = eS1 n( x, y, w)

(17)

The diffused back contact areas are considered as the ground terminal, where the majority carrier voltage drop is given by (12).

II. 3.3. Boundary Conditions (Double Junction Structure) Since the front emitter of this structure covers the whole surface, relation (16) of the previous subsection defines the boundary condition in that area. In a similar way, back surface boundary conditions are expressed by relations (12) and (13) of section II.3.1.

II. 4. Algorithm Description In this section we will provide a description of the algorithm, which is used to obtain a numerical solution of the problem formulated in the previous subsections. The derived expressions from the solutions of the minority carrier diffusion equation and the majority carrier voltage drop equation are given in appendices B and C respectively.

II. 4. 1. Diffusion Equation Solution Algorithm (1) The algorithm starts with an initial guess for the minority carrier concentrations in front and back surface areas not covered by the emitters, while in the junction regions the corresponding boundary conditions (depending on the investigated structure) that define the minority carrier concentrations are applied. Majority carrier voltage drop is initially set to zero. (2) A two-dimensional Fast Fourier Transform (FFT) with respect to x, y is performed to both surface concentrations and the minority carrier current density in Fourier space is calculated by differentiating the general solution of the diffusion equation with respect to z (appendix B). (3) An inverse FFT is then applied to each of the transformed current densities to obtain the real current densities at the areas not covered by the junctions, while at the regions covered by oxide or the back contacts the current densities are acquired from the boundary conditions. (4) Subsequent Fast Fourier Transforms are used in order to calculate the new carrier

~

concentrations as functions of the transformed current densities J n ( k x , k y ,0) ,

~ J n (k x , k y , w) (Appendix B).

(5) Inverse Fast Fourier Transforms are performed to the previously obtained carrier concentrations to calculate the new estimated ones in real space. (6) The solution is set as a mixture of the previously calculated and the newly obtained minority carrier distributions with a defined percentage. If this solution fulfills the convergence condition, the results are stored in order to proceed with the voltage drop equation, else calculations are repeated from step 2.

Three-Dimensional Simulation of Base Carrier Transport Effects…

61

II. 4. 2. Voltage Drop Equation Solution Algorithm The results of the solution of the minority carrier diffusion equation are required in order to obtain the majority carrier voltage drop. Therefore, the corresponding procedure for the case of equation (10), which follows the one referred to the previous section, is described through the following steps: (1) An initial guess for the majority carrier voltage drop (Vdrop) on both surfaces is used. This is considered as zero. This estimate also fulfills the boundary condition at the back-diffused contacts. (2) Two-dimensional Fast Fourier Transforms with respect to x, y are performed to the voltage distributions and the electric field distributions in Fourier space

~ ~ E (k x , k y ,0) , E (k x , k y , w) are calculated by differentiating the general solution of

equation (10) with respect to z (appendix C). (3) An inverse FFT is then applied to each of the transformed electric field distributions to obtain the corresponding values in real space, while at the regions covered by oxide or the junctions the current densities are acquired from the boundary conditions. (4) In this step, Fast Fourier Transforms are used in order to obtain the electric field distributions in Fourier space and the new transformed voltage

~

~

distributions V (k x , k y ,0) , V (k x , k y , w) as

~ E (k x , k y , w) are calculated (appendix C).

functions

of

~ E (k x , k y ,0) ,

(5) Inverse Fast Fourier Transforms are performed to the previously obtained voltage distributions to calculate the new estimated equivalents in real space. (6) The solution is set as a mixture of the previously calculated and the newly obtained one in a similar way as that referred to in the previous subsection. If this solution fulfills the convergence condition, the results are stored in order to be inserted in the boundary conditions of the minority carrier diffusion equation, else calculations are repeated from step 2.

II. 4. 3. Solution of the Final Coupled Problem The separate solutions of the differential equations (1) and (10) are not necessarily selfconsistent for all working conditions, since the boundary conditions that define the minority carrier diffusion equation depend on the calculated majority carrier voltage distribution in a non-linear way. Therefore self-consistency should be achieved by following a proper iterative procedure described as following: (1) Calculation of the solution of equation (1) by following the steps described in section II. 4.1. The calculations are performed under the assumption of zero voltage drop. (2) Numerical solution of equation (10) using the minority carrier currents and density distributions as obtained previously. (3) New solution of the minority carrier diffusion equation taking into account the voltage distributions calculated from the previous step. (4) The procedure continues between steps 2 and 3 until self-consistency is achieved.

62

K. Kotsovos and K. Misiakos

III. SIMULATION RESULTS The substrate of the simulated solar cells is considered as monocrystalline silicon, with base doping density NA=1016cm-3. We assume that the oxide-covered surfaces have ideal passivation properties, so there are no recombination losses at these areas. Therefore, the recombination velocity at these surfaces is zero while in the diffused contact regions is calculated from the relation (4), where we assume that the recombination current in these regions is J0c=10-12 A/cm2. The emitter saturation current value of all devices is the same (1013 A/cm2), while for all acquired results of the following sections III.1-III.4, a base diffusion length Ln of 800μm is assumed. The mobilities for minority and majority carriers are taken from Klaasen [13]. The simulated illumination is considered as the global AM1.5 sun spectrum [14] normalized to 100mW/cm2, where light trapping similar to the pyramidal textured scheme is assumed (Appendix A). The back surface contact has reflective characteristics with reflectivity R=95%. The simulation program, which is based on the algorithm of section II. 4 is used to calculate the IV characteristic of the cell and from that curve the maximum power, the short circuit current, the open circuit voltage and the series resistance of the cell are obtained. The series resistance of the cell is calculated for each point of the curve using the following relation

Voc − V − Rs =

I KT ln sc e I sc − I I

(18)

Voltage Drop (m

V)

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 400

0.0

320

80

240

160

x (μ m)

160

240

80

320 400

) μm ( y

Figure 2. Majority carrier voltage drop at the back surface of a back point junction structure with period length 400 μm and 80 μm back diffused junction sidelength. The cell is operated in the maximum power point (576 mV). Base thickness w is 400μm and Ln=800μm.

Three-Dimensional Simulation of Base Carrier Transport Effects…

63

The value of Rs which is obtained by (18) is caused by the current crowding effect at the back point contacts, since inside these regions current density values are large [9]. Such an effect induces a voltage drop which rises fast near the base contact edges. This effect is more intense in the back point junction structure, where near the back point emitter as shown in figure 2 an additional voltage drop is induced. The maximum voltage drop value is reached inside the emitter area, where it remains constant. In this structure d/l equals 0.2, while the period of the repeated pattern and the base thickness is 400μm.

III. 1. Short Circuit Current (Jsc) Figure 3 illustrates the short circuit current of the back junction structure as a function of the back base contact size and its spacing as parameter, assuming base thickness of 200μm. As expected, the reduction of the back contact area results to an increased photocurrent since the surface recombination velocity in this area (S2) is high. Jsc is also improved when the back contact spacing is smaller since in this case current crowding is reduced and carriers are collected more efficiently.

40.6

2

Jsc (mA/cm )

40.4 40.2 40.0 39.8

d=400μm d=200μm d=50μm

39.6 39.4

0.1

0.2

0.3

0.4

0.5

d/l Figure 3. Short circuit current (Jsc) of the back junction structure as a function of back contact size for different contact spacing l. Base thickness is 200μm and Ln=800μm.

The short circuit current of the back point junction structure is shown on figure 4. In this case, the increase of the back contact size is beneficial to the device photocurrent in contrast to the previously analyzed structure. A dramatic reduction in Jsc is also observed for the largest contact spacing (400μm), since in this case the required path for the collection of photogenerated carriers is significantly increased, thus the base minority carrier diffusion length should be higher for a more efficient current collection.

64

K. Kotsovos and K. Misiakos

40

36

2

Jsc(mA/cm )

38

34 32

d=400μm d=200μm d=50μm

30 28 26

0.1

0.2

0.3

0.4

0.5

d/l

42

1.03

40

0.98

38

0.93

36

0.88

34

0.83

32

0.79 front junction device double junction device back junction device point junction device

30 28 26

0.1

0.2

0.3

0.4

0.5

0.74

3D to 1D ratio

2

Jsc(mA/cm )

Figure 4. Short circuit current versus back contact size of the back point junction solar cell. The other parameters are the same as of figure 3.

0.69 0.64

d/l Figure 5. Short circuit current of all different point contact structures versus back contact size for a given contact spacing of 400μm, device thickness 200μm and Ln=800μm. The right axis is the current normalized to the corresponding value of the conventional 1D structure.

42.0

1.03

41.6

1.02

41.2

1.01

40.8

1.00

40.4

0.99

40.0

0.98

39.6

0.97

39.2

front junction device double junction device back junction device point junction device

38.8 38.4 38.0

0.1

0.2

0.3

0.4

0.5

0.96 0.95

65

3D to 1D ratio

2

Jsc(mA/cm )

Three-Dimensional Simulation of Base Carrier Transport Effects…

0.94 0.93

d/l Figure 6. Short circuit current of all different point contact structures versus back contact size for a given contact spacing of 50μm. The other parameters are the same as of the previous figure.

The next figure shows a comparison of the Jsc of the four different structures of figure 1 when back contact spacing is 400μm. The right hand axis is the current normalized to the corresponding typical solar cell structure where the back contact covers the whole back surface (1D case). Figure 6 is the same plot calculated for the smallest back contact spacing of 50μm. Since all devices are illuminated on the front side, most carriers are generated near the front surface, thus the single, double junction as a well as the typical solar cell device show an improved Jsc compared to both back junction structures. The reduction of the back contact spacing to 50μm results to a Jsc increase of these devices, especially in the point junction one, while the corresponding short circuit current of single and double junction devices remains almost unaffected from that change in l [9, 12]. A greater diffusion length would significantly improve the carrier collection ability of back contact devices, as will be discussed later.

III. 2. Open Circuit Voltage (Voc) Figure 7 demonstrates the open circuit voltage of all different structures as a function of the back contact size. The graph is referred to a specific contact size of 50μm, but it is also valid for the other ones since our simulations have shown that the influence of back contact spacing l on Voc is negligible, while base thickness is 200μm, as in the previous section. We observe that in contrast to the short circuit current, the back point junction device has the highest open circuit voltage compared to the other solar cell structures.

K. Kotsovos and K. Misiakos 684

front junction dev. double junction dev. back junction dev. point junction dev.

680 676

Voc(mV)

672

1.078 1.071 1.065 1.059

3D to 1D ratio

66

668

1.052

664

1.046

660

1.040

656

1.034

652

1.027

648

1.021

644

0.1

0.2

0.3

0.4

0.5

1.015

d/l Figure 7. Open circuit voltage (Voc) of all different point contact structures versus back contact size. The right axis is the current normalized to the corresponding value of the conventional 1D structure. Base thickness is 200μm and Ln=800μm.

The improved open circuit voltage may be attributed to the reduced surface recombination of the point junction structure, since the minimization of the area of the diffused regions is required to maximize the voltage [5], so Voc is improved at a faster rate compared to the other devices when the d/l ratio is reduced. The Voc of the front junction solar cell follow the corresponding point junction equivalent due to the low back surface recombination, while the back junction structure Voc values are slightly lower compared to the front junction cell. Finally, the double junction device has the lowest open circuit voltage of all point contact structures as expected, due to recombination in both emitters.

III. 3. Base Series Resistance (Rs) Figure 8 shows the base series resistance of the back junction structure near the cell maximum power point as a function of the back base contact size and its spacing as parameter when base thickness is 200μm. This graph shows that decreasing contact size leads to greater series resistance that grows dramatically for the smallest back contact area coverage fraction due to the current crowding at the back contact, as already reported for the front [9, 15,16] and double junction rear point contact solar cells [12]. In addition, the reduction of the back contact spacing limits the series resistance considerably. The current crowding effect is more evident in figure 9, where the series resistance of the back point junction structure is shown. It can be observed that there is an almost ten-fold increase to the Rs value for the smallest d/l ratio when the back contact spacing is changed from 50μm to 400μm. A rapid reduction of Rs is also observed when the back-diffused coverage ratio is increased.

Three-Dimensional Simulation of Base Carrier Transport Effects…

67

Figure 10 shows the series resistance dependence on the d/l ratio of the four different structures when the base contact spacing is 400μm. The simulations are performed for devices with different thickness, 200μm and 400μm respectively. 0.25

l=400μm l=200μm l=50μm

2

Rs (Ohm.cm )

0.20 0.15 0.10 0.05 0.00

0.1

0.2

0.3

0.4

0.5

d/l Figure 8. Base series resistance of the back junction structure near the cell maximum power point as a function of the back base contact size and its spacing as parameter. Base thickness is 200μm and Ln=800μm.

0.35

l=400μm l=200μm l=50μm

0.30

2

Rs (Ohm.cm )

0.25 0.20 0.15 0.10 0.05 0.00

0.1

0.2

0.3

0.4

0.5

d/l Figure 9. Base series resistance of the point back junction structure near the cell maximum power point as a function of the back base contact size and its spacing as parameter. Base thickness is 200μm and Ln=800μm.

68

K. Kotsovos and K. Misiakos 0.35

22.88

2

Rs (Ohm.cm )

0.30 0.25

19.61 16.34

0.20

13.07

0.15

9.80

0.10

6.54

0.05

3.27

0.00

0.1

0.2

0.3

0.4

0.5

3D to 1D ratio

front junction device double junction device back junction device point junction device

0.00

d/l A 10.50

front junction device double junction device back junction device point junction device

0.295

2

Rs (Ohm.cm )

0.246

9.00 7.50

0.197

6.00

0.148

4.50

0.098

3.00

0.049

1.50

0.000

0.1

0.2

0.3

0.4

0.5

3D to 1D ratio

0.344

0.00

d/l B Figure 10. Series resistance of the four different structures versus d/l ratio when the base contact spacing is 400μm, Ln=800μm and different base thickness: (A) 200μm and (B) 400μm. The right hand axis is the series resistance normalized to the corresponding value of the typical solar cell structure (1D case).

The right hand axis is the series resistance normalized to the corresponding value of the typical solar cell structure (1D case). In this case the back junction device exhibits the lowest Rs, while the corresponding series resistance of the double junction structure is slightly higher. On the contrary, the single front and point junction devices exhibit the highest Rs values. The series resistance of the front junction structure is significantly influenced from the base thickness in contrast to the rest of the point contact cells, except for the case of the smallest contact coverage fraction. Therefore, it can be concluded that the series resistance of the front junction cell is significantly influenced by the majority carrier flow in the vertical (z) direction. In most cases the conventional structure (1D) has the lowest series resistance, except for the largest d/l ratio, where the back and double junction structure Rs values are smaller. The situation is different in figure 11, which is the same plot as of figure 10 where the back contact spacing is reduced to 50μm.

Three-Dimensional Simulation of Base Carrier Transport Effects…

front junction device double junction device back junction device point junction device

0.045

2

Rs (Ohm.cm )

0.040 0.035

3.27 2.94 2.61 2.29

0.030

1.96

0.025

1.63

0.020

1.31

0.015

0.98

0.010

0.65

0.005

0.33

0.000

0.1

0.2

0.3

0.4

0.5

3D to 1D ratio

0.050

69

0.00

d/l A 3.05

front junction device double junction device back junction device point junction device

0.09

2

Rs (Ohm.cm )

0.08 0.07

2.74 2.44 2.13

0.06

1.83

0.05

1.52

0.04

1.22

0.03

0.91

0.02

0.61

0.01

0.30

0.00

0.1

0.2

0.3

0.4

0.5

3D to 1D ratio

0.10

0.00

d/l B Figure 11. Series resistance of the four different structures versus d/l ratio when the base contact spacing is 50μm, Ln=800μm and different base thickness: (A) 200μm and (B) 400μm. The right hand axis is the series resistance normalized to the corresponding value of the typical solar cell structure (1D case).

In this case, the back and point junction structures show the smallest Rs values, with the back junction one having the lowest. The Rs of the double junction solar cell is slightly higher in almost all cases compared to the previous structures and it is significantly reduced when the base thickness is changed from 400μm το 200μm. This is an indication that this device is also influenced by the majority carrier flow in the vertical direction as already reported. On the contrary, the series resistance of the front junction structure is by far the highest of all, approaching the limit of the conventional 1D device for large back contact coverage fractions. This may be attributed to the fact that in the front junction and the conventional solar cell devices the emitter and back contact are located on different surfaces and minority and majority carriers move to opposite directions, thus minority carrier flow opposes majority carrier movement, while in the back and point junction devices, the diffused regions lie in the same surface, so both minority and majority carriers flow towards the back surface. Therefore in this case, the reduced series resistance values of the back and point junction structures, is a

70

K. Kotsovos and K. Misiakos

clear advantage for concentrator applications, where ohmic losses need to be minimized, due to the high current generated by the solar cell. The double junction structure is also a good potential candidate for such applications since its series resistance is also low. It must be additionally noted, that all structures except from the single front junction one, may reach much lower Rs values compared to the conventional 1D device.

III. 4. Maximum Efficiency (η) Figure 12 shows the dependence of the back contact size on the back junction cell’s maximum efficiency with the contact spacing as a parameter. The plots demonstrate that there is a significant increase of the efficiency when the contact size decreases, provided that the d/l ratio is not less than 0.2, while with a further decrease of the back contact spacing shifts the efficiency to a smaller d/l ratio. According to the discussion of the previous sections, the reduction of the back contact size results to lower back contact recombination, so the open circuit voltage and the short circuit current are improved. On the other hand, a minimization of the back contact coverage fraction results to an intense current crowding effect, where the series resistance is significantly increased (figure 8), thus limiting the efficiency. The reduction of the back contact spacing limits this effect, so efficiency is improved. 22.8

efficiency (%)

22.6 22.4 22.2 22.0

d=400μm d=200μm d=50μm

21.8 21.6 0.1

0.2

0.3

0.4

0.5

d/l Figure 12. Maximum conversion efficiency of the back junction structure near the cell maximum power point as a function of the back base contact size and its spacing as parameter. Base thickness is 200μm and Ln=800μm.

Figure 13 illustrates the corresponding efficiency plots of the point back junction solar cell. In contrast with the previous figure, the efficiency of this structure is improved by increasing the back-diffused regions coverage, since in this case the photocurrent is enhanced as shown on figure 4, while series resistance is reduced. However, when back contact spacing is 50μm the device efficiency is improved when the d/l ratio is reduced from 0.5 to 0.2. In addition, when the back contact spacing is large (400μm) and the d/l ratio is 0.1 the efficiency is greatly reduced. This is due to the current crowding effect, which not only increases the

Three-Dimensional Simulation of Base Carrier Transport Effects…

71

series resistance but also requires a significantly larger minority carrier diffusion length for efficient carrier collection, as already discussed [5]. 23 22

efficiency (%)

21 20 19 18

d=400μm d=200μm d=50μm

17 16 15

0.1

0.2

0.3

0.4

0.5

d/l

23

1.07

22

1.02

21

0.97

20

0.93

19

0.88

18

front junction device double junction device back junction device point junction device

17 16 15

0.1

0.2

0.3

0.4

0.5

0.83 0.79

3D to 1D ratio

efficiency (%)

Figure 13. Maximum conversion efficiency of the back point junction structure near the cell maximum power point as a function of the back base contact size and its spacing as parameter. Base thickness is 200μm and Ln=800μm.

0.74 0.70

d/l Figure 14. Efficiency of the four different structures versus d/l ratio when the base contact spacing is 400μm. The right hand axis is the series resistance normalized to the corresponding value of the typical solar cell structure (1D case). Base thickness is 200μm and Ln=800μm.

Figure 14 compares the efficiency dependence on the d/l ratio of the four different structures, when the base contact spacing is 400μm. The right hand axis is the efficiency normalized to the corresponding value of the typical solar cell structure (1D case). The plots show that the efficiency of the back and point junction solar cell devices is significantly lower due to less efficient carrier collection, since the emitter of both structures

72

K. Kotsovos and K. Misiakos

is located on the back surface. As already discussed, the efficiency of the point junction solar cell is severely limited by current crowding. The highest performing structures are the single front and double junction solar cells, which have almost the same efficiency. Figure 15 is the same plot as of figure 10 calculated for reduced the back contact spacing to 50μm. As expected, the efficiency of all structures (except from the conventional solar cell) is improved and especially the corresponding point junction one. The best performing structures are still the single front and double junction solar cells, where the single front junction cell edges out the corresponding double junction one for small contact sizes. The efficiency of the back and point junction solar cell devices is still lower, where the back junction structure has the best efficiency of these two.

23.4

efficiency (%)

23.2

1.09 1.08

23.0

1.07

22.8

1.06

22.6

1.05

22.4

1.04

22.2

1.03

22.0

1.02

3D to 1D ratio

front junction device double junction device back junction device point junction device

1.01

21.8 0.1

0.2

0.3

0.4

0.5

d/l Figure 15. Efficiency of the four different structures versus d/l ratio when the base contact spacing is 50μm. The right hand axis is the series resistance normalized to the corresponding value of the typical solar cell structure (1D case). Base thickness is 200μm and Ln=800μm.

It should be pointed out, however that no front surface reflection is assumed for all devices and in reality all devices except from the back and point junction ones have additional losses due to front grid shadowing, which limits significantly their efficiency. On the other hand, larger diffusion lengths would greatly improve the performance of back and point junction solar cells; therefore the influence of this parameter is investigated in the next section.

III. 5. Minority Carrier Diffusion Length Influence on Device Short Circuit Current, Open Circuit Voltage and Efficiency Figure 16 shows the short circuit current of all different structures, including the conventional 1D solar cell, versus minority carrier base diffusion length (Ln). The calculations are performed for device thickness 200μm and 400μm.

Three-Dimensional Simulation of Base Carrier Transport Effects…

73

42 41

39

2

Jsc (mA/cm )

40

38 37

Front junction device Double junction device Back junction device Point junction device Conventional device

36 35 34 33

400

600

800

1000

1200

1400

1600

Ln (μm) A 42

2

Jsc (mA/cm )

40 38 36 34 32

front junction device double junction device dack junction device point junction device conventional device

30 28 26 24

400

600

800

1000

1200

1400

1600

Ln(μm) B Figure 16. Short circuit current of all different back point contact structures, including the conventional 1D solar cell versus minority carrier base diffusion length (Ln) and different base thickness: (A) 200μm and (B) 400μm. The simulations are performed for the smallest contact spacing of 50μm for improved efficiency, while point contact side length in (A) and (B) is set as 16μm and 10μm respectively.

The simulations are performed for the smallest contact spacing of 50μm for improved efficiency, while point contact side length is considered as 16μm and 10μm for figure 16(A) and 16(B) respectively, as a good balance between back surface recombination and series resistance. As expected the short circuit current of the double junction solar cell is the highest, due to the enhanced carrier collection of the back emitter [12,17], which remains almost constant for the considered diffusion lengths. The Jsc of the single front junction solar cell is near to the levels of the previous structure when Ln is greater or equal than 800μm, followed by the conventional 1D structure, where there is negligible photocurrent improvement when Ln is increased. In contrast to the former structures, the back and point junction solar cells benefit a lot from the diffusion length increase, since in this case the minimization of bulk recombination results to greatly improved carrier collection. This is more evident in the case of the smaller base thickness (200μm).

74

K. Kotsovos and K. Misiakos

Voc(mV)

Figure 17 shows the open circuit voltage of all different structures versus minority carrier base diffusion length (Ln) in the same manner as of previous figure.

705 700 695 690 685 680 675 670 665 660 655 650 645 640 635 630 625

Front junction device Double junction device Back junction device Point junction device Conventional device

400

600

800

1000

1200

1400

1600

Ln (μm) A 680 670

Voc (mV)

660 650 640 front junction dev. double junction dev. back junction dev. point junction dev. conventional dev.

630 620 610

400

600

800

1000

Ln(μm)

1200

1400

1600

B Figure 17. Open circuit voltage of all different back point contact structures, including the conventional 1D solar cell versus minority carrier base diffusion length (Ln). Other simulation parameters are the same as of figure 16.

As already discussed in section III.2, the limited front and back surface recombination of the point junction structure, is the cause of the improved open circuit voltage compared to the other solar cells. However, when Ln and base thickness is 400μm, bulk recombination limits the Voc of the former as well as of the back junction structure to values lower than the other structures. On the contrary, for the greatest Ln value of figure 17 the open circuit voltage gain of the point junction solar cell compared with the corresponding front junction equivalent is almost 10mV when w=400μm and exceeds 20mV when w=200μm. The Voc of the back junction structure almost equals or exceeds the corresponding front junction one for diffusion

Three-Dimensional Simulation of Base Carrier Transport Effects…

75

efficiency (%)

lengths greater than 800μm and w=200μm, while the voltage of the double junction structure is significantly lower compared to the three previously referred solar cells due to front and back emitter recombination. The conventional structure shows the most limited open circuit voltage, which is more than 35mV lower than the corresponding back point junction one when Ln is 1600μm and w=400μm, while this difference is increased to 60mV when w=200μm.

24.5 24.0 23.5 23.0 22.5 22.0 21.5 21.0 20.5 20.0 19.5 19.0 18.5 18.0 17.5

Front junction (shading loss) Double junction (shading loss) Conventional (shading loss)

Front junction device Double junction device Back junction device Point junction device Conventional device

400

600

800

1000

1200

1400

1600

Ln (μm) A Front junction (shading loss) Double junction (shading loss) Conventional (shading loss)

24

efficiency (%)

22 20 18 Single junction device Double junction device Back junction device Point junction device Conventional device

16 14 12

400

600

800

1000

1200

Ln (μm)

1400

1600

B Figure 18. Conversion efficiency of all different back point contact structures, including the conventional 1D solar cell versus minority carrier base diffusion length (Ln). Other simulation parameters are the same as of figure 16 and 17. The open-symbol colored plots refer to the front, double and the conventional solar cell structures when a 4% front grid shading loss is taken into account.

Finally, figure 18 shows the conversion efficiency of all different structures as a function of the minority carrier base diffusion length (Ln) in the same manner as of figures 16 and 17.

76

K. Kotsovos and K. Misiakos

As expected the point and back junction structures benefit the most from the diffusion length increase due to more efficient carrier collection, where for diffusion lengths greater than 800μm their efficiencies are almost equal, while the efficiency of the point junction solar cell exceeds the corresponding back junction one by a small margin when Ln=1600μm and w=400μm. In addition when w=400μm, these structures show superior performance compared to the conventional solar cell, when Ln=1200μm or larger, while for the largest diffusion length value of the graph, their efficiencies approach the levels of the single front and double junction ones. This small efficiency premium (approximately 0.5% absolute for the case of the single front junction cell and 0.34% for the double junction equivalent) is eliminated if front surface grid shadowing is taken into account, as shown on the graphs where a 4% shading loss is assumed. In this case, the back and point junction solar cells exhibit the highest efficiencies when Ln is greater than 1200μm (Ln/w>3). If the device thickness is reduced to 200μm and base diffusion length is greater than 1200μm (Ln/w>6), the point junction cell shows the highest efficiency of all structures, neglecting shadowing losses. When grid shadowing is set to 4%, the back and point junction structures reach higher efficiencies compared to the single and double junction solar cells for diffusion lengths greater than 800μm (Ln/w>4). Therefore, the choice of thin, high quality silicon wafers is absolutely necessary for the fabrication of the back and point junction solar cells. Topsil produces such FZ wafers with minority carrier lifetimes greater than 1ms for use in the PV industry [18], while wafers grown under the MCZ method (magnetically confined Czochralski) that are already used for the fabrication of high efficiency PERL structures [19] are good candidates as a starting material and they cost less than electronic quality FZ wafers.

IV. CONCLUSION In this work back junction, point contact (locally diffused) solar cells have been investigated through 3D simulations and compared with corresponding single front junction, double junction as well as conventional (1D) solar cell devices. It was shown that the simulated base series resistance of the back junction structure reached significantly lower values compared to the single front and double junction devices, especially for small back contact spacing. The back point junction solar cell reached the highest open circuit voltage due to reduced surface recombination, although current-crowding effects would severely affect its efficiency by reducing the solar cell’s photocurrent and increasing the base series resistance if the diffused areas are too small or too remotely spaced. A proper choice of back diffused contact spacing and size, would result to low Rs, close to the values of the back junction structure. Therefore, these cells are preferable for concentrator applications, since their efficiency would be significantly less affected from resistive losses compared to the single front junction back point contact solar cell and the conventional device. However, high quality starting material and relatively thin substrates (Ln/w>4) are required so that these devices reach efficiencies significantly higher than the conventional 1D device and close to or higher than the single front or double junction structure. On the other hand, the double junction solar cell could also be proposed as a very good choice for all applications, since it performs marginally lower compared to the corresponding front junction one on high quality substrates, and it has the best efficiency on low quality ones. In addition its simulated base series resistance reached values near to those of the back junction solar cells.

Three-Dimensional Simulation of Base Carrier Transport Effects…

77

REFERENCES [1] [2] [3]

[4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

R. A.Sinton, Y. Kwark, J. Y. Gan, R. M. Swanson, IEEE Electron Device Lett., Vol. 7 (10), p.1855, 1986. R. A. Sinton and R. M. Swanson, IEEE Trans. on Electron Devices, Vol. 37 (10), p.348, 1990. W. Mullikan, D. Rose, M. J. Cudzinovic, D. M. De Ceuster, K. R. McIntosh, David D. Smith, and R. M. Swanson, Proceedings of the 19 EPVSEC, Paris, France, p. 387, 2004. R. M. Swanson, EPRI Rep., AP-2859, 1983. R. M. Swanson, Solar Cells, Vol. 17, p. 85, 1986. D. J. Chin and Navon D. H., Solid State Electron., Vol. 24, p.109, 1981. H. Ohtsuka, Y. Ohkura, T. Uematsu and T. Warabisako, Prog. Photovoltaics: Res. Appl., Vol. 2, p. 275, 1994. Nichiporuk O., Kaminski A., Lemiti M., Fave A. and Skryshevski V, Sol. Energy Mat. and Sol. Cells, 86, p. 517, 2005. K. Kotsovos and K. Misiakos, J. Appl. Phys., Vol. 89, p. 2491, 2001. J. Zhao, A. Wang, P. Altermatt and M. A. Green, Appl. Phys. Lett., Vol. 66 , p. 3646, 1995. Zhao J., Wang A., και Green M. A., Progr. In Photovoltaics: Res. Appl., Vol. 7, p. 471, 1999. K. Kotsovos and K. Misiakos, Sol. En. Mat. and Sol. Cells, Vol. 77, p. 209, 2003. D.B.M. Klaassen, Solid-State Electronics, Vol. 35, p. 953, 1992. R. Hulstrom R. Bird and C. Riordan, Solar Cells, Vol. 15, p. 365, 1985. Zhao J., Wang A., and Green M. A., Sol. Energy Mat. and Sol. Cells,Vol. 32, p. 89, 1994. Catchpole K. R. and Blakers A. W., Sol. Energy Mat. and Sol. Cells, 73, p. 189, 2002. E. Van Kerschaver, C. Zechner, and J. Dicker, IEEE Trans. El. Devices, Vol. 47 (4) , p. 711, 2000. Vedde J., Jensen L., Larsen T. and Klausen T., Proceedings of the 19 EPVSEC, Paris, France, p. 1075, 2004. Zhao J., Wang A. and Green M. A., Progr. In Photovoltaics: Res. Appl., Vol. 8, p. 549, 2000.

APPENDIX A. LIGHT GENERATION PROFILE MODEL The surface of the simulated devices is textured as shown in figure A.1. The back surface is assumed reflective with a constant reflection coefficient Rb. This light-trapping scheme improves the absorbing properties of the investigated material, since incoming rays enter the cell with an angle of incidence, which is different than normal, so the material absorption coefficient αi is increased according to the following relation.

78

K. Kotsovos and K. Misiakos

θ Figure A.1. Assumed light trapping scheme used for the model calculations.

α ieff =

αi sin θ

(A.1)

where αieff is the effective material absorption coefficient under a given wavelength, while θ is the angle shown in figure A.1. Assuming that photon flux decays exponentially with increasing depth and that light is coupled out after performing a double pass across the cell, then the photon generation rate which is independent from x and y directions for all wavelengths of the considered AM1.5 spectrum, is written as N −1

G ( z ) = ∑ (α ieff g i e

−α ieff z

i =1

+ Rb e

−α ieff w

α ieff g i e

−α ieff ( w − z )

N −1

) Δλ i = ∑ Gi ( z )

(A.2)

i =1

where gi is the number of generated electron-hole pairs for the given wavelength i. By substituting in (A.2), the generation rate in front and back surface are obtained N −1

G (0) = ∑α ieff g i (1 + Rb e

− 2α ieff w

i =1

N −1

G ( w) = ∑ α ieff g i (e i =1

−α ieff w

N −1

)Δλi = ∑ Gi (0)

(A.3)

i =1

+ Rb e

−α ieff w

N −1

) Δλi = ∑ Gi ( w) i =1

(A.4)

The differentiation of (A.2) provides the following expression N −1 dG ( z ) N −1 −α z −α w −α ( w − z ) 2 = ∑ − α ieff g i (e ieff − Rb e ieff e ieff ) Δλ i = ∑ ΔGi ( z ) dz i =1 i =1

(A.5)

while the corresponding values for both surfaces are

dG ( z ) dz

N −1

z =0

2 = ∑ − α ieff g i (1 − Rb e i =1

− 2α ieff w

N −1

)Δλi = ∑ ΔGi (0) i =1

(A.6)

Three-Dimensional Simulation of Base Carrier Transport Effects…

dG ( z ) dz

N −1

z=w

2 = ∑ − α ieff g i (e i =1

−α ieff w

− Rb e

−α ieff w

N −1

)Δλ i = ∑ ΔGi ( w)

79 (A.7)

i =1

The expressions (A.2)-(A.7) will be used in the following sections for the solution of the transport equations.

APPENDIX B. SOLUTION OF THE MINORITY CARRIER CONTINUITY EQUATION Performing a two-dimensional Fourier Transform on equation (1) the following expression is obtained:

d 2n~(k x , k y , z ) dz 2 where

= (k x2 + k y2 +

~ G( z) 1 ~ n k k z ) ( , , ) − x y L2n Dn

(B.1)

~ n~ (k x , k y , z ), G ( z ) are the Fourier transforms with respect to x, y of

n( x, y, z ), G ( z ) respectively. This is an ordinary differential equation with independent variable z, which has the following general solution:

~ G( z) R1 z − R1 z ~ n (k x , k y , z ) = A(k x , k y )e + B (k x , k y )e + k1 Dn where R1 =

k x2 + k y2 + 1

L2n

(B.2)

and k1 is determined by differentiating (B.2) with respect to z

twice and equating the result with the right part of (B.1). By performing the necessary operations and with the use of (A.2) we get N −1

n~ (k x , k y , z ) = A(k x , k y )e − R1z + B(k x , k y )e R1z + ∑ i =1

~ Gi ( z ) 2 Dn ( R12 − α ieff )

(B.3)

The solution defined by (B.3) incorporates the constants A and B, which should be determined through the boundary conditions. These constants may be expressed as functions

~ (k , k ,0), n~ (k , k , w) in the following way: of n x y x y

A( k x , k y ) =

~ N −1 ~ G ( w) − Gi (0)e R1w n~ (k x , k y ,0)e R1w − n~ (k x , k y , w) + ∑ i 2 Dn ( R12 − α ieff ) i =1 e R1w − e − R1w

(B.4)

80

K. Kotsovos and K. Misiakos

B (k x , k y ) =

~ N −1 ~ Gi (0)e − R1w − Gi ( w) − R1w ~ ~ n ( k x , k y , w) − n ( k x , k y ,0)e +∑ 2 ) Dn ( R12 − aieff i =1

(B.5)

e R1w − e − R1w

~

~

where G (0), G ( w) are the carrier generation rates in the front and back surface respectively and are given by (A.3) and (A.4). The minority carrier diffusion current can be obtained by differentiating (B.3) with respect to z

~ J n (k x , k y , z ) eDn

=

dn~ (k x , k y , z ) dz

(

=

R1 B (k x , k y )e R1z − A(k x , k y )e − R1z

)

(B.6)

~ ΔGi ( z ) +∑ 2 2 i =1 D n ( R1 − α ieff ) N −1

After substituting (B.4) and (B.5) in (A.6) we get the final expressions for the current in both surfaces: ~ J n (kx , k y ,0) eDn

(

= R1

)

~ N −1 ~ G (0) eR1w + e−R1w − 2Gi (w) 2n~(kx , k y , w) − n~(kx , ky ,0)(eR1w + e−R1w ) + ∑ i 2 Dn (R12 −αieff ) i =1 eR1w − e−R1w

~ ΔGi (0) +∑ 2 2 i =1 Dn (R1 −αieff )

(B.7)

N−1

~ Jn (kx , ky , w) eDn

(

= R1

eR1w −e−R1w

~ ΔGi (w) +∑ 2 2 i=1 Dn (R1 −αieff) N−1

~

)

~ N−1 ~ 2Gi (0) − eR1w +e−R1w Gi (w) R1w −R1w ~ ~ n(kx , ky , w)(e +e ) −2n(kx , ky ,0) +∑ 2 Dn (R12 −αieff ) i=1 (B.8)

~

where ΔGi (0), ΔGi ( w) are defined in (A.6) and (A.7). The expressions (B.7) and (B.8) may be used to calculate the minority carrier diffusion currents in Fourier space as a function of the corresponding surface concentrations. The opposite procedure could be performed by solving the system of (B.7) and (B.8) to obtain the transformed surface minority carrier concentrations as a function of the corresponding diffusion currents

Three-Dimensional Simulation of Base Carrier Transport Effects…

n~(k x , k y ,0) =

N −1 ~ − J n (k x , k y ,0) + ∑ i =1

~ ΔGi (0) 2 Dn ( R12 − α ieff )

R1

N −1 ~ J n (k x , k y , w) − ∑

coth(R1 w) +

~ ΔGi ( w) 2 Dn ( R12 − α ieff )

~ N −1 Gi (0) 1 +∑ 2 R1 sinh(R1 w) i =1 Dn ( R12 − α ieff ) ~ N −1 ΔGi (0) ~ − J n (k x , k y ,0) + ∑ 2 2 i =1 Dn ( R1 − α ieff ) 1 n~(k x , k y , w) = + R1 sinh(R1 w) ~ N −1 ΔGi (w) ~ J n (k x , k y , w) − ∑ ~ 2 2 N −1 Gi (w) i =1 Dn ( R1 − α ieff ) coth(R1 w) + ∑ 2 2 R1 i =1 Dn ( R1 − α ieff ) i =1

81

(B.9)

(B.10)

The expressions (B.7)-(B.10) are used to solve the diffusion equation by application of the algorithm described in section II.4.1.

APPENDIX C. SOLUTION OF THE MAJORITY CARRIER VOLTAGE DROP EQUATION A similar analysis is used for the solution of equation (9), so by performing a twodimensional Fourier Transform in (9) and using (B.3) we get ~ d 2V ( k x , k y , z ) dz

2

~ = R 2V (k x , k y , z ) +

~ ⎞⎞ Dn − D p ⎛⎜ A(k x , k y )e − R1z + B ( k x , k y )e R1z N −1 Gi ( z ) ⎛ 1 ⎜ ⎟⎟ + − 1 ∑ 2 ⎜ 2 2 ⎟⎟ μ p N A ⎜⎝ L2n i =1 D n ⎝ Ln ( R1 − α ieff ) ⎠⎠

where R =

(C.11)

k x2 + k y2 .This ordinary differential equation has the following general solution

when R≠0 ~ V (k x , k y , z) = A1 (k x , k y )e − Rz + B1 (k x , k y )e Rz + ~ ⎞⎞ ⎛ Dn − D p ⎛⎜ c1 A(k x , k y )e − R1z + c2 B(k x , k y )e R1z N −1 c3 G 1 i ( z) ⎜ + − 1⎟ ⎟ ∑ 2 2 2 2 ⎜ ⎜ μpNA ⎝ Dn ⎝ Ln ( R1 − α ieff ) ⎟⎠ ⎟⎠ Ln i =1

(C.12)

where c1, c2, c3 are constants, which can be calculated by differentiating (C.12) with respect to z twice and equating the result with the right part of (C.11). Therefore, by completing these operations the general solution may written in the following form

82

K. Kotsovos and K. Misiakos

D p − Dn ~ ~ V (k x , k y , z ) = A1 (k x , k y )e − Rz + B1 ( k x , k y )e Rz + n (k x , k y , z ) μpNA

(C.13)

The constants A1 and B1 may be expressed as a function of the transformed surface voltage distributions in a similar manner as that of previous section

A1 (k x , k y ) =

B1 (k x , k y ) =

~ ~ V (k x , k y ,0)e Rw − V (k x , k y , w) + k1 (n~(k x , k y ,0)e Rw − n~(k x , k y , w)) e Rw − e −Rw ~ ~ V (k x , k y , w) − V (k x , k y ,0)e− Rw + k1 (n~(k x , k y , w) − n~(k x , k y ,0)e− Rw )

(C.14)

e Rw − e−Rw (C.15)

where k1 =

Dn − D p

μpNA

. The electric field can be calculated through differentiation of (C.13)

with respect to z, as following

~ dV (k x , k y , z ) dn~ ( k x , k y , z ) ~ = R A1 (k x , k y )e − Rz − B1 (k x , k y )e Rz + k1 E (k x , k y , z ) = − dz dz ~ J n (k x , k y , z) = R A1 (k x , k y )e − Rz − B1 (k x , k y )e Rz + k1 eDn

(

(

)

)

(C.16) The substitution of (C.14) and (C.15) in (C.16) leads to the following expressions for the electric field on both surfaces as a function of the corresponding voltage distributions

~ ⎛ ~ V (kx , k y , w) + k1n~(kx , k y , w) ⎞ ~ ~ ⎟+ ⎜ E (kx , k y ,0) = R V (kx , k y ,0) + k1n (kx , k y ,0) coth(Rw) − ⎟ ⎜ sinh ( ) Rw ⎠ ⎝ ~ J (k , k ,0) k1 n x y eDn

(

)

~ ⎞ ⎛ V (k x , k y ,0) + k1n~(k x , k y ,0) ~ ~ E(k x , k y , w) = R⎜ − V (k x , k y , w) + k1n~(k x , k y , w) coth(Rw) ⎟ + ⎟ ⎜ sinh(Rw) ⎠ ⎝ ~ J n (k x , k y , w) k1 eDn

(

(C.17)

)

(C.18)

Conversely, the surface voltage distributions may be related to the corresponding electric fields by using (C.17) and (C.18)

Three-Dimensional Simulation of Base Carrier Transport Effects…

~ J n (kx , k y ,0) ~ E(k x , k y ,0) − k1 eDn ~ coth(Rw) − V (kx , k y ,0) = R ~ J n (kx , k y , w) ~ E(kx , k y , w) − k1 eDn 1 − k1n~(kx , k y ,0) R sinh(Rw) ~ J n (kx , k y ,0) ~ E(kx , k y ,0) − k1 eDn 1 ~ V (k x , k y , w) = − R sinh(Rw) ~ J n (k x , k y , w) ~ E(kx , k y , w) − k1 eDn coth(Rw) − k1n~(kx , k y , w) R

83

(C.19)

(C.20)

Expressions (C.17)-(C.20) are valid when R≠0 and may be used to solve equation (9) by application of the algorithm described in section II.4.2. If R=0 the general solution of (9) is reduced to the following simple form

~ V (0,0, z ) = A1 (0,0) z + B1 (0,0) − k1n~ (0,0, z )

(C.21)

so the previously described procedure may be used to find the required relations for the electric field on both surfaces depending on the voltage distributions and conversely.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 85-158

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 3

MULTIPLE EFFECT DISTILLATION OF SEAWATER WATER USING SOLAR ENERGY – THE CASE OF ABU DHABI SOLAR DESALINATION PLANT* Ali M. El-Nashar ADWEA Research Center, UAE

ABSTRACT This report describes the solar desalination test plant in Abu Dhabi, UAE and gives a summary of its first year performance and economics. The plant has been operating successfully for 18 years supplying fresh water to the City of Abu Dhabi. The plant was commissioned in September 1984 and was running until the year 2002 when it was dismantled after fulfilling its objectives. The aim of the plant is to investigate the technical and economic feasibility of using solar desalination of seawater in providing fresh water to remote communities in the Middle East and to obtain long-term performance and reliability data on the operation of the plant. The plant has proved its technical feasibility and proved to be reliable in operation with few minor maintenance problems that required slight plant modification. Maintenance routines were established to maintain high plant performance. The economic feasibility of the plant was established by comparing the cost of water from a solar MED plant with a conventional MED plant using fossil fuel for plant capacity ranging from 100 m3/day to 1000 m3/day. It was found that the cost of water from solar MED plants is competitive with that from a conventional MED plant if the cost fuel continues to rise.

Keywords: desalination, solar energy, solar desalination, economic feasibility, operating performance, solar distillation.

*

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

86

Ali M. El-Nashar

1. INTRODUCTION Many remote areas of the world such as coastal desert areas in the Middle East or some Mediterranean and Caribbean islands are suffering from acute shortage of drinking water particularly during the summer season. Drinking water for these locations are normally hauled in by tankers or barges or produced by small desalination units using the available saline water. The transportation of water by tankers or barges involves a lot of expense and is fraught with logistical problems which can make fresh water not only very expensive when available but also its supply being very susceptible to frequent interruptions. The use of small conventional desalination units using a fossil fuel such as diesel oil as the energy supply can suffer from the same procurement problems that are encountered with transporting fresh water, namely transportation expenses and supply reliability. Some of the remote areas are blessed with abundant solar radiation which can be used as an energy source for small desalination units to provide a reliable drinking water supply for the inhabitants of the remote areas. Recently, considerable attention has been given to the use of solar energy as an energy source for desalination because of the high cost of fossil fuel in remote areas, difficulties in obtaining it, interest in reducing air pollution and the lack of electrical power source in remote areas. Desalination of seawater and brackish water is one of the ways for meeting future fresh water demand. Conventional desalination technology is fairly well established, and some of the processes may be considered quite mature although there is still considerable scope for improvement and innovation. Conventional desalination processes are energy intensive, and one of the major cost items in operating expenses of any conventional desalination plant is the energy cost. Thus, one of the major concerns about using desalination as a means of supplying fresh water to remote communities is the cost of energy. Apart from energy cost implications, there are environmental concerns with regard to the effects of using conventional energy sources. In recent years it has become clear that environmental pollution caused by the release of green house gases resulting from burning fossil fuels is responsible for ozone depletion and atmospheric warming. The need to control atmospheric emissions of greenhouse and other gases and substances will increasingly need to be based on growing reliance on renewable sources of energy.

Figure 1. Picture of Abu Dhabi solar desalination plant.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

87

A solar-assisted desalination plant was designed, constructed and put into operation on September 1984 as part of a cooperative research program between Japan and the United Arab Emirates (UAE) to test the technical and economic feasibility of using solar energy for desalination of seawater[1,2,3,4]. The plant (see figure 1) has been in operation in a Umm Al Nar near Abu Dhabi City until the year 2002 when it was dismantled. This report describes the main features of the first year of operation and compares its economics with conventional systems using the same desalination technology.

2. HISTORY OF ABU DHABI SOLAR DESALINATION PLANT In July 1979, when Mr. Ezaki, the then Japanese Minister of International Trade and Industry, visited the United Arab Emirates (UAE) and discussed the utilization of solar energy utilization in the UAE with Dr. Mana Saeed Al-Otaiba, the UAE Minister of Petroleum and Mineral Resources, they agreed on a joint project between the two countries to develop solar energy utilization for desalination of seawater. Under this agreement, several discussions were held at various levels. On January 22, 1983, the Record of Discussion (ROD) was finally signed for the joint implementation of a Research and Development Cooperation on Solar Energy Desalination Project by the New Energy Development Organization (NEDO) in Japan, and the Water and Electricity Department in the Abu Dhabi Emirate of the UAE[5]. An outline of the ROD is as follows: • • • • •

Execusion period of the project is 3 years starting January 22, 1983 Location of the project is Umm Al Nar in the suburbs of Abu Dhabi City Product water capacity of the test plant has a yearly average value of 80 m3/day Research operation period of the test plant is one year Japanese project executor: Engineering Advancement Association of Japan (ENAA)

The design, procurement and fabrication of the test plant started in February 1983 and the test plant was completed in October 1984. For the following year, research operation on the test plant was jointly conducted by ENAA and WED and was concluded in October 1985. Upon completion of the cooperative research project, the test plant was put in operation and was used as a research tool for a number of research projects carried out by WED. The plant was decommissioned in June 2002 after successfully operating for 18 years producing fresh water to Abu Dhabi City.

3. DESCRIPTION OF ABU DHABI SOLAR DESALINATION PLANT The solar desalination plant is designed for an expected yearly average fresh water production of 80 m3/day. A simplified schematic of the plant is shown in Figure 2.A bank of evacuated tube solar collectors, whose orientation with respect to the sun has been optimized to collect the maximum amount of solar radiation, is used to heat the collector fluid to a maximum temperature of about 99oC. The effective collector area of this bank is 1862 m2.

88

Ali M. El-Nashar

The heat collecting water leaving the collector bank flows into the top of the heat accumulator which has a total capacity of 300 m3. The heat accumulator is of the thermally stratified liquid type where, by virtue of density variation between the top and bottom layers, the higher temperature water is located in the upper region of the accumulator tank while the lower temperature water occupies the lower region. The lower temperature water is drawn from the tank bottom and pumped through the collectors by the heat collecting pump which has a capacity of 80m3/hr at 26m discharge head. The heat collecting water is drawn from the top of the accumulator tank by the heating water circulating pump and is forced to flow into the heating tubes of the first effect of the MED evaporator. This evaporator is designed for a maximum distillate production of 120m3. By transferring heat to the cooler brine flowing on the outside of the tubes, the heating water is cooled down and is then discharged into the accumulator.

Figure 2. A simplified schematic of the solar desalination plant.

The MED evaporator has 18 effects stacked one on top of the other with the highest temperature effect (No. 1) located at the top of the stack and the lowest temperature effect (No. 18) located at the bottom. The 18 effects are actually arranged in a double-stack configuration where effects 1, 3, 5,….17 are in one stack and effects 2, 4, 6…18 in the second. The double-stack arrangement is incorporated into one evaporator vessel as will be shown in detail later. In addition to the 18 effects, the evaporator has a final condenser designed to condense the vapor generated in the bottom (last) stage (No. 18). Heat input supplied to the first effect by the heating water is repeatedly used by evaporating a portion of the brine flowing into each effect. The evaporator operates under vacuum that is effected by a positive displacement pump connected to the final condenser. The absolute pressure to be maintained in the final condenser is designed to be 50 mmHg. The pressure to be maintained in each effect varies from slightly below atmospheric in the first effect to about 50 mmHg in the 18th effect. Seawater is used to condense the vapor generated in the 18th effect. Part of the discharged warm seawater leaving the final condenser returns to the sea, while the other part constitutes the evaporator feedwater. The feedwater flow rate amounts to 17.3 m3/hr; it flows through 17

Multiple Effect Distillation of Seawater Water Using Solar Energy …

89

preheaters before reaching the first effect, one preheater for each effect except the 18th effect. These preheaters are designed to raise the feedwater temperature incrementally by flowing from the bottom effect (No. 18) to the top effect (No.1).

3.1. Plant Description 3.1.1. The Solar Heat Collector Subsystem The solar energy collecting system (SECS) has the function of collecting the solar energy when it is available during the day using the collector bank and storing this energy in the heat accumulator which supplies thermal energy to the evaporator with minimum fluctuations in the supply temperature. This is desirable since steady state operation of the evaporator near its optimum operating condition is highly recommendable. The basic unit in the collector bank is the Sanyo evacuated tube solar collector which is shown in isometric in figure 3. This is a flat plate-type collector that employs selective coating absorber plates enclosed in glass tubes maintained under high vacuum of 10-4 mm Hg. Ten glass tubes with their absorber plates are incorporated in each collector. Along the centerline of each glass tube is located a single copper tube which is attached to the middle of the absorber plate. The heat collecting water flows through this center pipe and absorbs the solar energy collected.

Figure 3. Isometric view of a collector.

The ends of each glass tube are sealed to a special stainless steel end cap using a ceramic glass material having a coefficient of thermal expansion approximately the same as that of the glass tube. The difference in the thermal expansion between the copper tube and the glass tube is taken up by bellows installed between the end cap and the copper tube. Each collector consists of 10 individual tubes arranged in parallel. The heat collecting water moves inside the center tubes in a parallel/series arrangement whereby in five of the tubes the flow is in one direction and in the other five it is in the opposite direction. Attached to one end of the center tubes is a header tube with an orifice located in the middle of the header tube. The other ends are connected to return bends which are used to connect pairs of center tubes in series. Several collectors (14 in number) are connected in series by coupling the different header tubes. Each collector has an absorber area of 1.75 m2 and is coated with a black selective coating having an absorptivity, α ≥ 0.91 and an emittance,

90

Ali M. El-Nashar

ε ≤ 0.12. The specifications of a single collector as provided by the manufacturer are shown in Table 1. Table 1. Specifications of a single collector Item Selective coating Absorber area External dimensions Net weight Flow rate Max. operating pressure

Specification Absorptivity α ≥ 0.91 Emissivity ε ≤ 0.12 1.75 m2 2860 mm x 985 mm x 115 mm 64 kg 700 – 1,800 lit/hr 6 bar

Figure 4. One array pair.

The collector bank consists of 1064 collector units making up a total collector area of 1064 × 1.75 = 1862 m2. 28 collectors are combined to form a single array pair of collectors with its own support structure as shown in Figure 4. Each array pair consists of two parallel stacks of collector with each stack consisting of 14 collectors in series. The array pair is 14.5 meters long and 6.0 meters wide and is oriented in the north/south direction at a slope of 1/50. Water is supplied from the main pipe on the south side and passes through the 14 collectors connected in series and exit into the main pipe on the north side. 76 array pairs are arranged in a U-shape to form the whole collector bank. All array pairs are connected in parallel and each is provided with two isolating valves- at inlet and exit-, a drain valve, and an air vent. The bank is divided into six blocks designated A, B, C, D, E and F. Blocks A and F consists of 12 array pairs while the other blocks each consists of 13 array pairs. Figure 5 is a block diagram of the collector field.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

91

from MED evaporator

Tank #3

Block F

Block A Tank #2

Block E

Block B Block C

Tank #1

Block D

to MED evaporator

Figure 5. Collector bank consisting of six blocks A, B, C, D, E and F.

3.1.2. The Heat Accumulator Subsystem The heat accumulator subsystem (see Figure 6) is designed to provide thermal energy to the evaporator during its 24 hours per day operation. It consists of three carbon steel tanks having a total capacity of 300 m3 and contains hot water at a temperature ranging from 74oC to 99oC and at atmospheric pressure. The tanks are insulated with a 100 mm layer of fiberglass to minimize heat loss to the ambient air. All three cylindrical tanks have the same internal diameter (3.8m) and wall thickness (9mm). However, the tank heights are not identical with tank No. 1 having an effective height of 10m while tanks No. 2 and 3 having an effective height of 7.6m. The heat collecting water from the collector bank is introduced at the top of tank No. 1. The heat collecting water to the collector bank is taken from the bottom of tank No. 3. Heating water to the evaporator is drawn from the top of tank No. 1 and returns to the bottom of tank No. 3. The water is therefore stratified in such a way that the top water layers of tank No. 1 are always at the highest temperature and the bottom layer of tank No. 3 at the lowest temperature. The heat accumulator tanks have enough capacity to be able to provide the required thermal load (for the evaporator) for about 16 hours after sunset provided that the tanks are fully charged just before sunset. This feature makes it possible to operate the desalination unit during night time. Only during extended overcast or hazy days when sandstorms prevail we expect plant shut-down to occur due to insufficient energy collection.

92

Ali M. El-Nashar

Figure 6. Heat accumulator subsystem.

3.1.3. MED Evaporator Subsystem A horizontal tube, thin film multi-effect distiller (MED) is used for desalination of seawater. The distiller is manufactured by Sasakura Engineering Co., Ltd. It was chosen because of its capability to accommodate large load fluctuations and its small consumption of electrical power. The maximum capacity of the distiller is 120 m3/day. A flow diagram of the MED desalination process is shown in Figure 7. Preheated feedwater is sprayed into the top of the first effect and descends down the evaporator stack, flowing as a thin film over the tube bundle in each effect. The feedwater flashes and thereby cooled by several degrees as it passes from one effect to the next. It is rejected at the bottom of the plant as cool, concentrated brine. In the top effect, heating water from the accumulator is used to partially evaporate the thin seawater film on the outside of the tubes. The generated vapor passes through demisters to the inside of the tubes in the second effect where it condenses to form part of the product. It simultaneously causes further evaporation from the external seawater film and the process is repeated from effect to effect down the plant. The heat input from the accumulator is thus used over and over again in successive evaporation/condensation heat exchangers in each effect to produce more product and new vapor, thereby obtaining a maximum quantity of fresh water with minimum heat input. The vapor generated in the last effect (18th) is condensed in a seawater-cooled condenser and part of the seawater is used as feedwater to the stack. The remaining seawater is rejected to the sea and carries most of the heat away from the process.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

93

Figure 7. The MED evaporator.

3.2. Design Features Table 2 shows the design conditions and specifications at the time of the original planning of the test plant. At the planning stage, no detailed solar radiation data was available for Abu Dhabi city and the only data available was that for nearby Kuwait. Therefore the data for Kuwait was used with the annual mean daily solar radiation on horizontal surface taken as 5000 kcal/m2 day. Based on the measurements at the test plant made during 1985 and subsequent years, the annual average daily values were found to be slightly higher than this value (5270 kcal/m2 day for 1985).

94

Ali M. El-Nashar Table 2. Plant design conditions assumed for design of test plant

Design parameter Solar radiation Ambient temperature Rainfall Wind speed Relative humidity Seawater temperature System capacity Solar collector type Heat accumulator Evaporator

Assumed value/range 5,000 kcal/m2 day (annual mean value on horizontal surface) 30oC (daytime mean temperature) 18.1 – 390.1 mm/year 5 m/s (for collector design) 30 m/s (for structures) Max. 100%, min. 10%, normal 25-90% 55,000 ppm TDS (design base) 80 m3/day as expected yearly average Evacuated glass tube collector 1862 m2 (effective absorbing area) Thermally stratified vertical cylinder Capacity 300 m3 Horizontal-tube, multiple effect stack type Evaporator, capacity 120 m3/day, specific heat consumption 43.8 kcal/kg- product water

4. MEASUREMENTS AND DATA ACQUISITION SYSTEM Table 3 lists the plant parameters measured every 15 minutes and the measured data is sent to the data acquisition system (DAS). The DAS is shown in Figure 8 and consists of two separate subsystems: one is the on-line control room subsystem and the other is data analysis subsystem. The data analysis subsystem consists of a Thermodac 32 data logger manufactured by Eto Denki, Co., PC (model PC-8001 mkII manufactured by NEC company) and a PC printer (model PC-8023C manufactured by NEC company). The data analysis subsystem consists of a data logger (model Thermodac 3 manufactured by Eto Denki, Co.), PC (model PC-8801mk II by NEC company) and PC printer (model PC-8024 by NEC company).

Figure 8. The data acquisition system.

Multiple Effect Distillation of Seawater Water Using Solar Energy … The control room subsystem has the following functions: • • • •

Sampling of data at 15 minute intervals Calculate hourly average values once per hour Record data on CD at even hours (i.e. 8:00, 10:00, 12:00,…) Print a summary report every 12 hours.

The data analysis subsystem has the following functions: • • • •

Make daily, weekly and monthly reports. Format new data disks Copy data disks for backup Edit hourly or daily data on data disks. Table 3. Items of data acquired every 15 minutes from data loggers Thermodac 32 and Thermodac 3 Channel # Measuring survice THERMODAC 32 1 Ambient temp. 2 Collector field outlet temp. 3 Accumulator inlet temp. 4 Accumulator outlet temp. 5 Heating water inlet temp. 6 Heating water outlet temp. 7 No. 1 effect temp. 8 Preheater No. 1 outlet temp. 9 No. 18 effect temp. 10 Seawater temp. 11 Empty collector temp. #1 12 Empty collector temp. #2 13 Relative humidity 14 Heat collecting water flow 15 Heating water flow 16 Product water flow 17 Solar radiation 18 Heat collected from field 19 Heat used by evaporator 20 Heat collected by block F 21 Heat collected by block A 22 Pump P-101 running hours (heat collecting pump) 23 Pump P-205 running hours (product water pump) 24 Electrical energy consumption THERMODAC 3 1 No. 1 effect temp. 2 No. 4 effect temp. 3 No. 7 effect temp. 4 No. 10 effect temp. 5 No. 13 effect temp. 6 No. 16 effect temp. 7 No. 18 effect temp.

Tag #

Signal output

Unit

TE-111 TE-102-1 TE-102-2 TE-102-3 TE-104 TE-105 TE-206-1 TE-203 TE-206-2 TE-202 TE-301 TE-302 HUE-111 FIT-101 FIT-105 FQ-205 SQ-111 CAQ-102 CAQ-105 CAQ-101-1 CAQ-101-2 N/A

DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V DC 1-5 V Thermocouple Thermocouple DC 1-5 V DC 1-5 V DC 1-5 V DC 24V pulse DC 24V pulse DC 24V pulse DC 24V pulse DC 24V pulse DC 24V pulse On/off pulse

o

C C o C o C o C o C o C o C o C o C mV mV % m3/hr m3/hr m3 kcal/hr m2 kcal kcal kcal Kcal hr

N/A

On/off pulse

hr

N/A

DC 24V pulse

kWh

TEW-206-1 TEW-206-4 TEW-206-7 TEW-206-10 TEW-206-13 TEW-206-16 TEW-206-18

Pt 100 Pt 100 Pt 100 Pt 100 Pt 100 Pt 100 Pt 100

o

o

C C o C o C o C o C o C o

95

96

Ali M. El-Nashar

The items of data shown in table 3 are transferred to the computer through an RS-232C interface to the control room computer to be processed. The results of the mean of four 15minute data values are displayed every hour on the CRT in the control room. The following hourly values are displayed on the CRT: • • • • • • • • • • • • • • • • • • • • • • • • •

Climatic temperature (oC) Solar radiation (kcal/m2hr) Heat consumption (kcal/kg dist.) OHTC of heater (first effect) (kcal/m2hroC) Seawater flow (m3/hr) Seawater TDS (ppm) Seawater inlet temp. (oC) Seawater outlet temp. (oC) Heat collected by block F (kcal/hr) Heat collected by block A (kcal/hr) Heat collecting pump flow rate (m3/hr) Heat accumulator inlet temp. (oC) Heat accumulator outlet temp. (oC) Heating water inlet temp. to evaporator (oC) Heating water return temp. from evaporator (oC) Heat supplied to accumulator (kcal/hr) Heat supplied to evaporator (kcal/hr) Heat supplied to evaporator (kcal/hr) Evaporator feedwater flow rate (m3/hr) Preheater #1 outlet temp. (oC) First effect temp. (oC) 18th effect temp. (oC) Product water flow rate (m3/hr) Absorber plate temp. of empty collector (oC) Header tube temp. of empty collector (oC)

4.1. Measuring the Heat Collected in Block F Measuring the heat loss in the piping system in such a large collector bank is obviously laborious and will require the accurate measurement of collector fluid temperature at many locations within the field. This was deemed impractical and was ruled out from the outset. The solution which was found practical is to isolate a single block of collectors and use it for test measurements in order to find the piping heat loss in this block, then estimate the heat loss in the piping system of the whole collector bank based on the results obtained from the measurements carried out on the selected block. Block F was selected for this purpose since this block was already provided with resistance temperature detectors at inlet and outlet of the block, as well as a vortex flowmeter for measuring the water flow rate through the block. Figure 9 is a schematic diagram of block F showing the location of the temperature measuring probes. Two RTDs are attached at block inlet (location E) and block outlet (location F). The RTDs, which are three-wire sheathed platinum resistance elements, are

Multiple Effect Distillation of Seawater Water Using Solar Energy …

97

model RN33-AMAS and are manufactured by Yokogawa Hokushin Corporation. They are connected to Thermodac 32 via resistance-to-voltage converters which produce 1-5 volt DC signal outputs that are fed to temperature recorders. The two output signals from the resistance-to-voltage converters which are attached to the RTDs at locations E and F are connected to a programmable computing unit (PCU), model SPLR-100A manufactured by Yokogawa (see Yokogawa Instruction Manual for Model YF100 vortex flowmeters (1993)), in which the 4-20 mA signal from the vortex flow meter is also connected to the PCU. A variety of arithmetical computational functions can be performed by the PCU. Programs can be developed and written to ROM (Read Only Memory) using a dedicated programming language connecting the PCU to an SPRG Programmer. The pulse output signal from the PCU represents the heat collected between E and F and was measured by subtracting the two input temperature signals, TF - TE, and multiplying by the flow rate signal and the specific heat of the heat collecting fluid (water) at the operating temperature to obtain the heat collected between E and F (block heat collected).

Figure 9. Location of measuring sensors in block F.

98

Ali M. El-Nashar

The analog signals from the resistance-to-voltage converters, as well as the pulse output signal from the PCU were connected to the DAS for recording and printing and for transfer to a PC computer system via an RS-232 interface cable. The pulse output signal was integrated over a period of one hour and scanning of the data was made every 15 minutes. The inlet and outlet temperature measurements (TA and TB) of each of the 12 arrays making up block F were measured using copper-constantan thermocouples which were attached to the array header pipes as shown in Figure 9. The assembly of these thermocouples into the array supply and return pipes, as well as the locations of the 24 (2 x 12 arrays) thermocouples used are also shown in this figure. The mV output signals from these thermocouples were recorded on an hourly basis. The heat collected by the 12 arrays was calculated by multiplying the temperature difference, TA – TB, by the flow rate and the specific heat in a manner similar to that used for estimating the block heat collected. A remote converter-type vortex flowmeter (model YF105, Yewflo by Yokogawa) is used with a model YFA11 vortex flow converter are used for flow measurement. This vortex flow meter measures the flow rates and converts the measurements to a 4 to 20 mA DC output signal. The accuracy of the instrument is ± 1.0% of reading plus ± 0.1% of full scale. A solar radiation sensor (pyranometer)- model H 201 manufactured by Nakaasa Instrument Co.- is used to measure the global insolation on the collector absorber plate. The sensor has a measuring range of 0 to 2 kW/m2 and has an accuracy of ± 0.5% of full scale. The mV output signal from the sensor is first amplified before being converted into a pulse signal for connection to Thermodac 32 data logger. This signal is integrated over hourly intervals in order to obtain the hourly values of the solar radiation. The ambient temperature was measured by a three-wire RTD connected to the DAS. Figure 10 shows a block diagram of the data acquisition and analysis for the heat collection system. The estimated percentage error in the heat measurement was estimated as 1.5 – 3% while the error in the collector efficiency measurement 2.0 – 3.5%.

Figure 10. Block diagram of data acquisition system for heat collection.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

99

5. DATA ANALYSIS 5.1. Calculating the Solar Radiation on Absorber Plate The global solar radiation intensity on a tilted surface having a tilt angle α to the ground, and an azimuth angle β can be expressed by Eq. 1.

It = Id + I s

(1)

where Id is the direct component of solar radiation and Is is the diffuse component. According to [5], the components can be expressed as:

Id = Io × P

1 sin( h )

× cos( θ )

(2) 1

1 + cos(α ) 1 1 − P sin( h ) × I s = × I o × sin( h ) × 2 1 − 1.4 ln( P ) 2

(3)

where P is the atmospheric transmittance, h is the solar altitude, θ is the incidence angle on the tilted surface. The solar angles are shown in Figure 11. The transmittance P is defined as the ratio between the normal solar radiation at the ground and the corresponding value at the outer limit of the atmosphere:

P=

IN I oN

(4)

where IN is direct normal radiation at the ground and IoN is the corresponding value at the edge of the atmosphere. Since the hourly global radiation is measured at a tilted surface having the same tilt angle as the collector absorber plates, It, it is possible to solve the following equation for the hourly values of P:

It = Io × P

1 sin( h )

× cos(θ ) +

1 sin( h )

1 1− P 1 + cos(α ) × I o × sin( h) × × 2 1 − 1.4 ln( P) 2

(5)

With the hourly values of P estimated, it is possible to calculate the direct and diffuse components from Eqs. 2 and 3.

100

Ali M. El-Nashar

Figure 11. Solar angles.

5.2. Calculating the Performance of the Collector Bank The heat collection amount, Qc, was calculated from the measured collector water temperatures and flow rate

Qc = mc C p (Tc 2 − Tc1 )

(6)

where mc is the water flow rate through the collector bank (or block) and Tc1 and Tc2 are the inlet and outlet water temperatures, respectively. The heat collection efficiency is expressed by the following polynomial equation

ηc ≡

Qc = a + b x + c x2 Ac I t

(7)

a, b, and c are constants and x is a parameter defined as

Tc1 + Tc 2 − Ta 2 x= It

(8)

Multiple Effect Distillation of Seawater Water Using Solar Energy …

101

where Ta is the ambient air temperature and It is the solar radiation on a tilted surface. Susbstituting the expression for Qc from Eq. 6 and the expression for x from Eq. 8 into Eq. 7 we get a relationship for the collector outlet temperature, Tc2 in terms of the collector inlet temperature, ambient temperature, solar radiation and collector absorber area

C1Tc22 + 2C 2Tc 2 + C 3 = 0

(9)

where C1, C2 and C3 are given by the following expressions

C1 = γ Ac C 2 = γ Ac (Tc1 − 2Ta ) + β Ac I t − 2mc I t C 3 = γ Ac (Tc1 − 2Ta ) 2 + 2βAc I t (Tc1 − 2Ta ) + 4αAc I t2 + 4mc I t Tc1 with the above equations, the hourly values of the outlet water temperature Tc2 can be derived if the hourly inlet water temperature, Tc1, is given.

5. 3. Calculating the Performance of the Evaporator The performance of the evaporator consists of estimating the overall heat transfer coefficients (OHTC) of the first effect (heater), the other evaporator effects (2nd – 18th effects), the 17 preheaters and the condenser as well as evaluating the economy (or specific heat consumption) of the evaporator. The list of measurements carried out for the evaporator are shown in Table 4. Table 4. Measurements made at evaporator Measurement

Location

Symbol

Unit

Flow rate

Heating water flow

mhw

m3/h

Seawater flow

msw

m3/h

Feedwater flow

mfw

m3/h

Product water flow

md

m3/h

Heating water inlet

Thw1

o

C

Thw2

o

C

First effect

Tev(1)

o

C

18th effect

Tev(18)

o

C

Tpr(1)

o

C

Seawater inlet to condenser

Tcon1

o

C

Seawater outlet from condenser

Tcon2

o

C

Heat amount

Heat supplied by heating water

Qhw

kcal/h

Total dissolved solids

Seawater

Cb(0)

Kg salt/kg water

Temperture

Heating water outlet

st

1 preheater outlet

102

Ali M. El-Nashar

5.3.1. Calculating the Brine Concentration at Each Effect The brine total dissolved solids (TDS) at each effect is estimated from a mass balance for each effect considering the fact that the heater (first effect) receives all the feedwater while all the other effects receive only half of the amount of brine leaving the first effect. This stems from the fact that the evaporator consists of one effect at the top followed by two stacks in parallel with all odd-numbered effects in one stack and even-numbered ones in the other. The brine concentration in the first effect is estimated from:

Cb (1) =

mfw × C b ( 0) ( mfw − md / 18)

(10)

where mfw is the feedwater flow rate, md is the distillate flow rate and Cb(0) is the seawater concentration. The concentration of the even-numbered effects is obtained from

C b (2 N ) =

0.5 × ( mfw − md / 18) × C b (1) [0.5 × ( mfw − md / 18) − N × md / 18]

(11)

where N varies from 1 (2nd effect) to 9 (18th effect) The concentration of the odd-numbered effects is obtained from Cb(2N + 1) = Cb(2N) Where N varies from 1 (3rd effect) to 8 (17th effect).

(12)

5.3.2. OHTC of Heater (First Effect) In the heater (first effect) hot water from the accumulator flows through the tubes in a horizontal tube bundle while a relatively cold seawater is flowing as a thin boiling film on the outside of the tubes. Heat is therefore transmitted from the hot water inside the tubes to seawater on the outside. The overall heat transfer coefficient for the heater can be expressed as

Uh =

Qh Ah (ΔT ) h

(13)

where Qh is the rate of heat transfer (kcal/h), Ah is the heat transfer area (Ah = 24.9 m2) and (ΔT)h is the log-mean-temperature difference obtained from the equation:

(ΔT ) h =

{Thw1 − Tev (1)} − {Thw 2 − T pr (1)} {T − Tev (1)} ln hw1 {Thw 2 − T pr (1)}

The rate of heat transfer can be estimated from

(14)

Multiple Effect Distillation of Seawater Water Using Solar Energy …

Qh = mhw C p (Thw1 − Thw 2 )

103 (15)

The specific heat Cp is calculated at the average temperature (Thw1 + Thw2)/2.

5.3.3. Average OHTC of other Evaporator Effects In the 2nd –18th effects the mechanism of heat transfer is different from that in the first effect. In these effects, boiling takes place on the outside of the horizontal tubes while condensation occurs on the tube inside. The average overall heat transfer coefficient for the 17 evaporators was estimated from the equation

U ev =

Qev Aev (ΔT ) ev

(16)

where Qev is the average heat transfer rate for each evaporator effect, (ΔT)ev is the average log-mean-temperature difference for each evaporator and Aev is the heat transfer area of each evaporator (Aev = 63.1 m2). The heat transfer rate Qev is obtained from the fact that the distillate production consists of two components: production by the 17 (2nd –18th ) effects and production by the 17 (1st – 17th) preheaters:

md = (md ) ev + (md ) pr

(17)

Where ( md )ev is the distillate production by the 17 evaporator effects and ( md )pr is the production by the 17 preheaters. ( md )pr is calculated from a heat balance equation over all the preheaters:

(md ) pr =

mfw C p (Tpr (1) − Tcon2 ) Lav

(18)

The latent heat of vaporization Lav is estimated at an average temperature of [Tev(1) + Tev(18)]/2. The average heat transfer by each evaporator effect is obtained from:

Qev =

mev × Lav 17 md Lav − mfw C p (Tpr (1) − Tcon2 ) = 17

(19)

The average log-mean-temperature difference (ΔT)ev is calculated as the average temperature difference between the heating steam inside the evaporator tube bundles and the boiling brine on the outside. Noting that the heating steam temperature in a particular effect i is slightly lower than the brine temperature in the preceding effect (i.e., effect i -1) due to:

104

Ali M. El-Nashar • •

the boiling point elevation (BPE), the temperature drop across the demister inside (Δp)demis,

(ΔT)ev was estimated from the following relation

(ΔT ) ev =

18

17

i =1

i =1

Tev (1) − Tev (18) − ∑ BPE − ∑ (Δp) demis 17

(20)

The values of the boiling point elevation (BPE) and the demister pressure drop (Δp)demis are calculated using the correlations given in the appendix.

5.3.4. Average OHTC of Preheaters In the 17 preheaters seawater flowing inside the tube bundles of heat exchangers is heated up by steam condensing on the tube outside. For any preheater, the average OHTC (kcal/hr m2oC) can be calculated using the equation:

U ph=

Q ph A ph × (ΔT ) ph × 17

(21)

where Qph is the heat transfer rate for all preheaters , kcal/hr and Aph is the heat transfer area of each preheater, m2 (Aph = 19.5 m2). Qph is calculated from the measured feedwater flow rate and the temperature difference between the outlet of preheater #1 (top preheater) and the outlet of the condenser:

Q ph = M fw × C p × (T ph (1) − Tcond 2 )

(22)

The log-mean temperature difference (ΔT)ph is assumed to be identical for each preheater.

5.3.5. OHTC of Condenser The OHTC of the condenser is calculated from the measured seawater flow rate and condenser inlet and outlet temperatures according to the equation:

U cond =

Qcond Acond × (ΔT ) cond

(23)

where Qcond is condenser heat flow rate, kcal/hr, Acond is the condenser heat transfer area, m2 (Acond = 35.3 m2). The log-mean temperature difference for the condenser is calculated from the equation:

(ΔT ) cond =

(Tev (18) − Tcond 1 ) − (Tev (18) − Tcond 2 ) (T (18) − Tcond 1 ) ln ev (Tev (18) − Tcond 2 )

(24)

Multiple Effect Distillation of Seawater Water Using Solar Energy …

105

5.3.6. Evaporator Economy The performance ratio (PR) of the evaporator is calculated as the ratio between the product flow rate, Md (kg/hr) and the equivalent amount of heating steam to the heater (first effect):

PR =

md (Qh / L)

(25)

where Qh is amount of heat supplied to the heater (first effect) and L is the latent heat of vaporization (kcal/kg).

6. WEATHER CONDITION IN ABU DHABI For a solar desalination plant, the solar radiation and ambient temperature have a great effect on the heat collection efficiency of the collectors while the temperature and salinity of the seawater influence the capacity of the distiller. Consequently, these data which have been yielded by the research operation constitute valuable basic material for analyzing the performance of the operation of the plant.The total solar radiation on a horizontal surface for 1985 ( see Table 5) came to 1,923,000 kcal/m2 year which gave a mean daily value of 5,270 kcal/m2 day. When the monthly average total solar radiation an a horizontal surface is considered, the maximum value was achieved in June which was 1.95 times as high as the minimum value recorded in December. Table 6 shows the atmospheric transmittance and the figures in this table were determined by calculations from the total solar radiation values on a tilted surface using the Bouger’s and Berlage’s formulae[5]. Generally, there is a tendency on any given clear day for the atmospheric transmittance at dawn and dusk when the solar altitude is low to work out on the high side and for it to be on the low side around midday when the solar altitude is high. Table 6 gives the values for the atmospheric transmittance at noon everyday and consequently the air mass (air mass = sin(h) where h is the solar altitude) is ranging from 1.0 to 1.5. Since values yielded for cloudy days are assumed to be meaningless, the same table gives the mean atmospheric transmittance values for the above five days in each month as the monthly averages. This indicates that the closer the atmospheric transmittance is to 1.0, the purer is the air, and seasonal changes are apparent in that it is high during winter and low during summer. The monthly average daily mean, maximum and minimum ambient temperatures are shown in Table 7. Table 8 shows the monthly average daily maximum relative humidity in Abu Dhabi and Table 9 shows the seawater salinity. The results of the TDS analysis of seawater are indicated in table 9. Variations are visible in the measured values with the salinity being slightly higher in summer months compared to winter months.

Table 5. Daily total radiation on tilted surface at 21o 09’ (data of 1985) Month Solar radiation Kcal/m2 day Solar radiation kWh/m2 day

Jan. 3700

Feb. 5010

Mar. 5270

Apr. 6210

May 6350

Jun. 6770

Jul. 5810

Aug. 5830

Sep. 5740

Oct. 4970

Nov. 4130

Dec. 3470

4.3

5.8

6.1

7.2

7.4

7.9

6.7

6.8

6.7

5.8

4.8

4.0

Table 6. Atmospheric transmittance at noon Month Transmittance

Jan. 0.74

Feb. 0.76

Mar. 0.69

Apr. 0.70

May 0.65

Jun. 0.66

Jul. 0.61

Aug. 0.61

Sep. 0.64

Oct. 0.67

Nov. 0.70

Table 7. Monthly average daily mean, max. and min. ambient temperatures (1985) Month Mean Maximum Minimum

Jan. 20.2 24.8 16.7

Feb. 19.4 23.9 15.7

Mar. 22.8 28.2 18.7

Apr. 26.1 31.8 21.3

May 30.5 36.3 25.9

Jun. 32.1 37.8 27.2

Jul. 34.0 39.9 30.0

Aug. 35.2 42.6 30.7

Sep. 32.3 39.0 27.3

Oct. 29.4 34.5 24.8

Nov. 25.5 29.9 21.5

Dec. 21.0 25.6 17.1

Table 8. Monthly average daily maximum relative humidity in Abu Dhabi (1985) Month Humidity

Jan. 90.9

Feb. 82.6

Mar. 82.0

Apr. 81.5

May 77.6

Jun. 80.4

Jul. 78.0

Aug. 77.6

Sep. 87.8

Oct. 85.2

Nov. 87.3

Dec. 83.6

Table 9. Seawater salinity Sampling date TDS (ppm)

25 Sep. 84 52,100

2 Jan. 85 51,200

2 Mar. 85 51,900

17 Jul. 85 53,500

22 Oct. 85 53,000

Dec. 0.72

Multiple Effect Distillation of Seawater Water Using Solar Energy …

107

7. OPERATING CHARACTERISTICS In this section the results of plant characteristics during the first year of operation is presented. Typical measured performance for each of the main plant subsystems, e.g. heat collecting subsystem, heat accumulator subsystem and evaporator subsystem, are presented. The major performance parameters for the whole plant are also shown.

7.1. Heat Collecting Subsystem 7.1.1. Heat Collector Efficiency Instantaneous Heat Collection Efficiency Figure 12 shows the measured collector efficiency of the whole collector bank for a typical month (June 1985) and the ideal efficiency of a single collector measured under controlled conditions at the manufacturer’s laboratory. The ideal efficiency can be correlated to the x-parameter by the following polynomial equation:

η c0 = 0.913 − 2.46 x − 1.92 x 2

(26)

The measured collector efficiency is seen to be lower than the ideal efficiency due to heat losses from the piping system as well as losses due to attenuation of solar radiation received by the absorber plates because of dust deposition on the glass tubes of the collectors. In order to exclude the data during the warm-up and cool down periods in the early morning and before sunset, only the data for the period 10:00 am to 5:00 pm were plotted. The amount of heat collected was estimated from the measured inlet and outlet water temperature to the collector field and the flow rate of water. Therefore, all the heat loss from the internal and external piping system was included in the instantaneous efficiency shown in Figure 12. Figure 13 shows the instantaneous efficiency of the collector bank at mid-day during the months of January and June 1985. In a clear day, the efficiency at mid-day (12:00 noon) is usually close to the highest value for that day. It can be seen that, for the month of January where some days are usually overcast, the mid-day efficiency drops for those overcast days. June is normally a sunny month with rare overcast periods, the mid-day efficiency fluctuates only slightly. The mid-day efficiency can drop slightly during periods of sand storms where the air is laden with small dust particles that reduce the solar radiation falling on the absorber plates of the collectors.

108

Ali M. El-Nashar Measured efficiency

Ideal efficiency

collector efficiency

1

y = 0.913 - 2.46 x - 1.92

0,8

2

0,6 0,4 0,2

Data of June 1985

0 0

0,02

0,04

0,06 0,08 x-param eter, hm 2oC/kcal

0,1

0,12

0,14

Figure 12. Measured efficiency of collector bank and the ideal efficiency of a single collector for a typical day.

Collectir mid-day efficiency

Jan. 1985

Jun. 1985

0,8 0,6 0,4 0,2 0 0

5

10

15

20

25

30

35

Day number

Figure 13. Daily instantaneous efficiency at mid-day during January and June 1985.

Daily efficiency

Jan. 85

июн.85

1 0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 1

6

11

16

21

Day num ber

Figure 14. Daily heat collection efficiency for January and June 1985.

26

31

Multiple Effect Distillation of Seawater Water Using Solar Energy …

109

7.1.2. Daily Heat Collection Efficiency The daily heat collection efficiency is defined as the ratio of the amount of heat produced by the collector bank divided by the amount of solar radiation falling on the absorber plates. The daily amount of heat produced and the daily amount of incident solar radiation are estimated from the summation of their hourly values during a day which are measured and recorded by the data acquisition system. Figure 14 shows the daily heat collection efficiency for the months of January and June 1985. It can be seen that the daily efficiency for sunny days is normally above 50% except for days with prolonged overcast periods, such as in January, during which the daily efficiency can drop below 40%.

7.2. Heat Accumulator System 7.2.1. Heat Loss from the Heat Accumulator Table 10 gives the monthly heat loss from the heat accumulator as a percentage of the incident solar radiation on the collector field for several months during 1985. As can be seen, the heat loss varies from 4.6% (for June) to 6.6% (for December). The percentage heat loss is increased during winter months compared to summer months and also increases during the month where the plant experiences several shutdowns either emergency shutdown or automatic shutdown due to insufficient accumulator charge. Months where the plant has been in emergency shutdown for long periods of pump maintenance, for example, can have accumulator loss exceeding 8%. Table 10. Monthly heat loss from the heat accumulator as a percentage of incident solar radiation (data of 1985) Month Heat loss %

Jan. 5.6

Mar. 6.3

May 5.0

Jun. 4.6

Sep. 5.6

Nov. 5.3

Dec. 6.6

7.2.2. Thermal Stratification Ratio The thermal stratification ratio is the ratio of the mass of the strata of water in the heat accumulator where a temperature gradient exists to the total mass of water inside the accumulator. The total mass of water inside the 3 accumulator tanks is essentially constant at 300 m3. The thickness of the temperature gradient strata is measured using the temperature sensor (RTD) located in the middle of tank # 2 (mid-temperature tank). The heat accumulator operates in two main modes: the simultaneous heat collection and discharge mode during day periods and the discharge mode during night periods. Table 11 shows the percentage thermal stratification ratio for each of these two modes for a number of days. During the heat discharge mode (at night), the temperature gradient strata averages 19.5 m3 which corresponds to a stratification ratio of 6.5%. During the simultaneous heat collection and discharge mode (during day time) this strata averages 21.9 m3 which corresponds to a stratification ratio of 7.3%.

110

Ali M. El-Nashar Table 11. Thermal stratification ratio of heat accumulator (data of 1985) Date

Heat discharge

Jan. 14 Jan. 15 Jan. 19 Feb. 2 Feb. 13 Feb. 20 Mar. 10 Mar. 20 Mar. 29 Apr. 4 Apr. 10 Apr. 20 May 3 May 11 May 20 Average

4.6% 3.8% 4.6% 6.7% 6.6% 6.1% 10.6% 8.8% 6.5% 9.4% 7.2% 1.7% 6.4% 5.6% 8.3% 6.5%

Simultaneous heat collection and discharge 6.0% 7.9% 4.8% 8.1 6.1% 8.3% 6.3% 3.9% 5.8% 5.7% 7.6% 11.5% 8.9% 9.2% 9.3% 7.3%

7.3. Evaporating System 7.3.1. Evaporator Performance The performance ratio (PR) is defined here as the amount of product water produced by the evaporator per 526 kcal of heat supplied by the heating water. Table 12 shows the average PR values and average specific heat consumption for several months during the first year of the test plant operation. The effect of the product water flow rate on the PR is shown in Figure 15 which is based on actual tests carried out during plant commissioning in November 1984. Table 12. Performance ratio of the evaporator Month Jan. 1985 Feb. 1985 Mar. 1985 Apr. 1985 May 1985 Jun. 1985 Sep. 1985

Product water flow m3/hr 4.6 5.3 5.0 5.2 5.0 5.1 4.9

Specific heat consumption kcal/kg 40.71 39.04 39.31 38.83 39.04 39.98 40.31

Performance Ratio 12.9 13.5 13.4 13.5 13.5 13.2 13.0

Multiple Effect Distillation of Seawater Water Using Solar Energy …

111

Performance Ratio, PR

14 13 12 11 10 0

1

2

3 4 Product w ater flow , m 3/hr

5

6

7

Figure 15. Performance Ratio versus product water flow.

OHTC, kcal/hr m2 oC

Overall Heat Transfer Coefficients The overall heat transfer coefficients (OHTC) shown in Figure 16 are estimated from the measured temperatures and flow rates. Shown in this figure are the average HTC for the evaporators, the average OHTC of the preheaters and the OHTC of the heater (first effect) and condenser. The evaporators are heat exchangers in which vapor is condensed inside tubes while seawater brine boils on the outside of the tubes. These heat exchangers have the highest OHTC compared with the other heat exchangers as shown in the figure. The data shown in this figure represent typical values of the OHTC,s obtained during the first year of plant operation. Some deterioration in the OHTC’s has occurred during the subsequent years which necessitated acid cleaning to remove scale deposited on the heat exchanger tubes.

3000 2500

Heater

2000 1500 1000

Evaporators Preheaters 1

2

3

4

5

6

7

8

9 10 11 12 13 14

Condenser

Measurem ent #

Figure 16. Measured OHTC during the period June 1-7, 1985.

7.4. Performance of the Plant Figure 17 and Figure 18 show pie chart plots depicting the January and June split of the incident solar radiation falling on the collector bank among collector loss (including piping loss and loss due to dust effect), accumulator heat loss, heat loss by the evaporator and the heat going for desalination. It can be seen that for January, 37% of the incident solar radiation is converted into thermal energy for desalination while for June it is 47%. This is due to the fact that for winter months heat losses from the collector bank, the accumulator tanks and the evaporator are larger than during the summer period because ambient winter temperatures are substantially lower than summer temperatures.

112

Ali M. El-Nashar

January 1985 37% 54%

3% 6%

Collector loss

Accumulator loss

Evaporator loss

Heat for distillation

Figure 17. Split of the incident solar radiation for January. June 1985

47%

47%

1% 5%

Collector loss

Accumulator loss

Evaporator loss

Heat for distillation

Figure 18. Split of the incident solar radiation for June.

60 40 20 0 Jan.

Feb.

Mar.

Specific w ater production (kg/m 2day)

Apr.

May

Jun.

Specific heat consum ption (kcal/kg dist.)

Specific pow er consum ption (kWh/m 3)

Figure 19. Plant performance parameters for the first six months of 1985.

Figure 19 shows the monthly values of three major plant performance parameters: the specific water production in kg/m2 day, the specific heat consumption in kcal/kg dist. and the specific power consumption in kWh/m3. The specific water production is defined as the rate of water production per unit collector area. It varies between 43.9 kg/m2 day and 78.1 kg/m2 day with the lower values for the winter months and the higher values for the summer months. The specific heat consumptions is defined as the rate of heat supplied to the

Multiple Effect Distillation of Seawater Water Using Solar Energy …

113

evaporator per unit of water production. It is essentially constant at about 39 kcal/m3 of product water. The specific power consumption, defined as the electrical energy required per cubic meter of product water, varies between 7.08 kWh/m3 and 8.15 kWh/m3 with the higher values typical for winter months and lower values for summer months.

8. PLANT MAINTENANCE AND MODIFICATIONS In this section, the major maintenance and plant modification activities will be described. Among the major maintenance activities that had to be carried out regularly are the following: • • • • • • •

Cleaning the collectors to remove dust and dirt deposited. Inspection for corrosion replacing corroded components Inspect vacuum inside collector glass tubes Monitor scale formation on evaporator heat exchanger tubes and carry out acid cleaning if necessary Monitor level of water in heat accumulator and add makeup water and anticorrosive chemical as required Evaporator pump maintenance Monitor pressure difference across seawater intake filters and clean when necessary

Plant modifications were necessary in order to avoid the harmful effects of emergency plant shutdown due to power failure.

8.1. Heat Collecting System 8.1.1. Cleaning the Solar Collector Field The performance of the solar collector field is affected by the extent of dust deposition on the glass tube which influences the transmittance of the glass tube to solar radiation. It is therefore important to clean the collector field at regular intervals to maintain good performance. The cleaning was carried out using a high-pressure water jet spray device. Since fresh water is used for cleaning, it is important to economize on the use of fresh water for collector cleaning without adversely affecting the performance of the collectors. The amount of water required for each cleaning session depends on the extent of dust deposition on the collectors; more water is required when more dust has accumulated on the collectors. Since sandstorms are seasonal in character, the amount of dust deposited in a particular period depends on the month of the year. Several tests were carried out to determine the required amount of fresh water needed for each collector block. Figure 20 shows the measured water quantity used for each cleaning session for one block for different months. As can be seen, there is wide variation in the quantity of cleaning water required which may be attributed to variation in the personal skills of the different cleaners as well as variation in the amount of dust deposition on the glass tubes. The average quantity of water for each cleaning session is about 1000 liters.

114

Ali M. El-Nashar

2500 2000 Water 1500 quantity 1000 (liters) 500 0 January

M ay

Se ptem be r

M onth

Figure 20. Water quantity required for cleaning a single block for different months.

8.1.2. Corrosion of the Collector Air Vent Valves A total of 76 air vent valves were installed on the arrays of solar collectors to vent the air out during the water filling up process. Water started to leak from these valves after few months of operation and several valves have to be replaced with spare ones. Because of a limited supply of vent valves, few of the leaking valves have to be plugged as a makeshift measure until new valves are procured, a procedure usually taking several months due to administrative delays. The problem with the leaky vent valves occurred soon after a power failure happened during a sunny day causing the heat collecting pump to shutdown while the water is still inside the collectors which caused partial evaporation of the water and frequent opening and closing of the vent valves. To prevent further damage to the vent valves, it was decided that plant re-start after the restoration of power following a power failure is made either early in the morning or at dusk. 8.1.3. Vacuum Loss Inside Glass Tubes As previously stated, the glass tubes operate under a high vacuum of 10-4 mmHg. High collector efficiency depends on maintaining this high vacuum level inside the collector. The vacuum level is monitored every year by checking the condition of the getter. At the end of the first year of operation, the condition of the vacuum on almost all glass tube was essentially as new. 8.1.4. Scale Prevention The most important item for the evaporator is scale prevention. There are two kinds of scale that could form in seawater distillation: a hard scale and a soft scale. The hard scale consists mainly of calcium sulfate (CaSO4) and very difficult to remove from the heat transfer surfaces once formed. The only way to avoid the formation of this scale is to operate the evaporator within the solubility limits of Ca+2 and SO42- ions in the brine. The evaporator was designed in such a way that in normal operation, the concentration of these ions are not allowed to reach saturation. On the other hand, soft scale which consists mainly of calcium carbonate (Ca CO3) and magnesium hydroxide (Mg(OH)2) can be avoided by injecting a chemical inhibitor such as Belgard EV into seawater feed. In our evaporator 10 ppm of Belgard EV is injected and was found satisfactory with the high-salinity seawater of the Gulf (52,000 ppm). Scale formation was monitored on an hourly basis during plant operation by

Multiple Effect Distillation of Seawater Water Using Solar Energy …

115

observing the heat transfer coefficient of the heater (first effect) since it has the most likelihood of scale formation since it operates at the highest temperature. For an emergency, an acid injection system is used to carry out an acid cleaning procedure to remove any soft scale formed.

8.1.5. Anti-Corrosion Chemical for Use in the Heat Collecting Water An anticorrosive chemical must be used in the heat collecting water, which circulates through a closed system, to protect the equipment from corrosion. “High Clean CL-100” was selected in view of the heat collector tube material being of copper and the heat accumulator and piping material being of iron. “High Clean CL-100” was found to be effective in protecting both materials from corrosion. This chemical is a solution of alkanol-amino salt of nitrogen-containing condensate having the chemical symbol: [R-SO2NH(CN2)n COOH] where R = C6H5 and n = 1 to 3. The solution is a light brown transparent liquid having a pH of 8 – 8.5. A concentration of 5,000 ppm was recommended by the manufacturer. However, the concentration of the chemical as poured into the makeup water tank after draining and leakage was gradually reduced as the stock of the chemical decreased because the heat collecting water was forced to be drained more frequently than was planned due to unexpected power failures and pump troubles. The average concentration of the chemical in the collector water during the first year of operation was approximately 2,8000 ppm. Table 13 shows the results of analysis of the heat collecting water of four samples taken during the first year of operation. It can be seen that there was virtually no change in the total Fe content indicating an adequate anticorrosive effect for Fe. On the other hand, the Cu content tended to increase gradually as the concentration of the anticorrosive chemical drops. The thickness of the copper tubes corroded in the solar collector field was estimated at 0.001 mm/year[5]. Table 13. Chemical analysis of heat collecting water Sampling data

31 Jan. 85

27 Mar. 85

30 Jul. 85

21 Oct. 85

pH

8.6

8.5

7.9

8.2

Conductivity (μS/cm)

209

210

373

242

Temperature (oC)

-

20

28

18

M Alkalinity (ppm as CaCO3)

526

486

468

452

Total hardness (ppm as C(ppm)aCO3)

22

8

20

4

Chloride (ppm as Cl-)

37

21

44

33

Silica (ppm as SiO2)

1.3

2.0

1.9

2.8

Total iron (ppm as Fe)

1.6

1.82

1.1

1.4

Copper (ppm as Cu )

0.55

0.62

1.8

2.83

Total Nitrogen (ppm as N)

217

177

156

161

High-Clean CL-100

5,167

4,214

3,714

3,843

++

8.1.6. Measures Against Power Failure The original plant design was such that in the case of a power failure during the day, the heat collecting pump stops and the water in the collectors, subject to solar radiation, evaporates resulting in dangerous water hammer in the header pipes. In order to protect

116

Ali M. El-Nashar

against such power failure, the solar field piping system was modified by installing three motorized valves (normally closed) on three 25 mmφ lines, one motorized valve (normally open) on the 125 mmφ on the outlet piping of the heat collecting pump and a check valve on the 125 mmφ on the return header pipe from the collector field to the accumulator. The modifications are shown in Figure 21. In the modified system, following a power failure, the three 25 mmφ motorized values will be open thus draining all the water in the collector field and the 125 mmφ motorized valve in the pump outlet will close thus preventing the collector water from flowing by gravity from the heat accumulator to the collector field. The modified system proved effective in protecting against the hazard of evaporation in the collector field. The collector field so not automatically restart following the restoration of power and the valves has to be rest manually by the operator.

Figure 21. Modifications to the solar collector field to protect against power failure.

8.2. Evaporating System 8.2.1. Evaporator Pump Maintenance Pump problems were one of the main causes of plant downtime. The plant had 12 pumps: seawater intake pump, seawater pump, seawater feed pump, brine blow-down pump, product water pump, vacuum pump, drain pump, heat collecting pump, heating water pump, drinking water pump, plant water pump and priming vacuum pump. All these pumps are motoroperated centrifugal pumps pumping seawater, brine or product water except the vacuum pump which is an oil ring-seal vacuum pump drawing in a mixture of vapor and noncondensable gas from the evaporator. Most of the pump problems were started by a highpitch noise and vibration which gets worse as time goes on. The pumps dealing with seawater

Multiple Effect Distillation of Seawater Water Using Solar Energy …

117

or its brine were found to be more amenable to failure than the other pumps due to seawater corrosion problems that affect the pump bearings. Continuous operation of the evaporator had to be interrupted by pump trouble several times. The pumps which gave most of the problems were the drain pump, the seawater intake pump. Most of the problems were resolved by replacing the pump/motor bearings or replacing the mechanical seal (or packing material). Many emergency plant shutdowns happened because of “intake pit level low” signal due to air leakage through the shaft packing of the seawater intake pump because of a worn out packing material. The suction line of this pump operates at a vacuum due to the fact that the seawater suction point is below the pump level. The air leakage into the pump would break the vacuum and drastically reduces the pump flow. The drain pump gave a lot of trouble. All the drains from the evaporator was collected into a pit in which the pump suction line is immersed. A level switch is used to start and stop the operation of the pump. This pump was found to be easily clogged with debris which accumulates inside the pit and caused a relay trip due to motor over-current. To solve this problem, a fine screen was installed on the tip of the suction line inside the pit to prevent the trash from entering the pump. The screen has to be cleaned regularly and trash removed from it. The product water pump was another source of problems and was responsible for plant shutdown particularly in the first year of plant operation. Faulty bearings and a pump shaft were replaced.

8.2.2. Inspection of the Evaporator An overhaul was carried out at the end of one year of plant operation which consisted of: •

•

•

•

The first effect, which is subject to high temperature and hence high possibility of scaling was inspected after the first year of operation. An accumulation of silt (fine sand) approximately 1 cm thick was found on the bottom of the 1st effect and was removed. Blackened spots were seen on the outer surface of the first effect tube bundle but they remained the same as those at the start of plant operation and should cause no problems. Scale marks were seen on parts of the tube bundle of the first effect near the water chambers at the heater inlet and outlet, but they did not represent any advanced state. A re-inspection few months later revealed no advanced state. No abnormal conditions were seen on the 18th effect.

8.2.3. Change in Operating Sequence A modification was carried out in a part of the operating sequence so that the test plant could be safely operated. This change brought about the following improvements in plant operation: • •

If the intake strainer is clogged with foreign matter like seaweed, the seawater intake pump will be stopped temporarily. If the concentration of the discharged brine becomes too high, an alarm will be triggered and the plant shut down. This is to prevent scale formation on the heat

118

Ali M. El-Nashar

•

•

transfer tubes if the flow of heating water increases more than necessary or if the feedwater quantity is reduced. If the product water quantity becomes too small, an alarm will be triggered and the plant shutdown. This is to prevent the product water pump fro running at no load if the temperature of the heating water drops enough to make water production too low. The manner of emergency plant shutdown will be either of the two types given below depending on the degree of emergency: o Immediate plant shutdown for motor relay tripping, too low seawater flow, or the like. o Gradual plant shutdown for too high first effect evaporator temperature, too high brine concentration, or the like.

8.2.4. Modification of the System for Injecting Anti-Scale Chemical The original plan called for a single system for injecting the scale preventing chemical “Belgard EV” into seawater feed. However, because the injection pump has failed during unmanned operation of the plant and it was feared that the plant could have been subjected to scaling, it was decided to modify the injection system to include the standby system as a secondary injection system. In case the primary injection system fails, the secondary system automatically turns on. 8.2.5. Modification of the Method of Feeding Sealing Water to the Priming Vacuum Pump Because the seawater level is lower than the level of the seawater intake pump, when the plant is not in operation, the seawater in the suction line of this pump drains back by gravity to the sea and the line is full of air. The priming vacuum pump, which is a water ring sealed vacuum pump, is used to prime (start) the seawater intake pump by insuring that this pump and its suction line is full of water. Seawater was initially used for the sealing water, but since the primary vacuum pump is operated at plant startup only, seawater inside the pump and sealing tank has caused considerable amount of rust and made the water brown. It was therefore decided that fresh water be used for the sealing water instead of seawater.

9. SIMULATION PROGRAM AND ITS VALIDATION 9.1. Simulation Program 9.1.1. Outline The performance of a solar desalination plant that uses solar energy as the heating source for a seawater evaporator is affected by the amount of solar radiation, ambient temperature and other climatic conditions. These conditions are extremely unstable and are subject to constant change. This means that computer-based simulation is indispensable if, to cope with these changes, the system operation states are to be grasped and the amount of product water annually is to be forecast. As shown with the basic design, computer simulation was used to investigate the operating characteristics even at the planning stage of the test plant. However, it was not possible to grasp with any accuracy the climatic conditions when design was commenced and

Multiple Effect Distillation of Seawater Water Using Solar Energy …

119

forecasted values based on experience to date were used even for the performance of the various components. Subsequently, research operation was begun, actual climatic data were accumulated and the characteristics of the equipment started to become clear through the winter and summer component tests and the continuous operation tests. Thus, a new look was taken at the simulation program. As a result of comparing and examining the calculated values and th actually measured values, it was concluded that the program must be revised in order that the operation of the test plant was reproduced as faithfully as possible and more accurate forecasts were made to cope with different operating conditions and plants of different scales. The program was revised and manuals were prepared to use the program between October 1985 and January 1986.

9.1.2. Flow Chart of the SOLDES Program Evacuated glass tube-type of collectors are used and the absorber area can be varied from 500 m2 to 20,000 m2. The heat collecting system uses a bypass circuit. When the temperature of the heat collecting system drops below the set value, the bypass operation is performed, and once the temperature rises above the set value, operation is switched over to the accumulator side. Solar cell type control is excercised to start and stop the heat collecting water pump. The accumulator is treated as a thermal stratified type, and the temperature distribution inside the tank is determined while bearing in mind the inflow and outflow of the heat collecting water and heating water. The number of collectors used and the heating water flow rate are varied so that the maximum operating capacity of the evaporator is not exceeded. However, for the simulation, consideration was given to maximize the effective use of the collectors and a bypass circuit was installed between the accumulator and the evaporator. By this means, a system of control was adopted where some of the heating water returning from the evaporator is bypassed and forwarded to the evaporator so that the temperature of the heating water entering the evaporator is kept below the rating. This is particularly useful when the accumulator water temperature is excessively high. The evaporator capacity can be varied over the range of 100 to 2,000 m3/day, the maximum brine temperature can be varied from 60 – 80oC and the number of effects of the evaporator can be varied fro 13 to 32. The simulation program takes into consideration the influence of the heating water temperature, heating water flow rate, seawater temperature and seawater flow rate. Figure 21 is an abbreviated flow chart of the simulation program. The equipment specifications, calculation conditions and other data are input to the program and are divided into two groups: System data No. 1 and System data No. 2. The data input with the System data No. 1 group serve to output error messages and suspend the execusion of the calculations in cases where there are errors in the input data or where allowable ranges have been exceeded. Any calculation period ranging from one day to one year can be designated and a balance in the system is determined every 30 minutes in relation to the calculation loop. The calculation results can be output either by the day or by the month. The “SOLDES” simulation program for the solar desalination plant is composed of 22 subroutine programs, two sets of system data and four types of climatic data.

120

Ali M. El-Nashar

Figure 22. Flow chart of Solar Desalination Plant computer simulation program.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

121

9.1.3. Program Input and Output Data In conducting a computer simulation of a solar desalination plant, input data need to be revised according to changes in plant specifications such as absorber area, accumulator capacity, or evaporator capacity. The items of data input that can be revised using program “SYDT1” are as follows: • • • • • • • • • • • • • • • • • • • • • • •

Title of simulation run Starting and finishing date of simulation Specifications of daily data print out for each month Correction coefficient of solar radiation by month Correction coefficient of ambient temperature by month Collector absorption area of field Azimuth angle of collector Collector support angle Tilt angle of absorber plate Heat collecting water flow rate Heat collection pump rated power consumption Heat accumulator capacity Heat accumulator initial temperature distribution Evaporator capacity Maximum brine temperature Number of effects of evaporator Heating water flow rate Seawater flow rate Power consumption of vacuum pump and other evaporator pumps By-pass valve open/close temperatures Evaporator start/stop temperatures Correction coefficient of dust influence Specification of collector cleaning days Table 14. Types of meteorological data which can be used in "SOLDES"

Data name

Solar radiation data

Ambient temperature data

MEDT1

Hourly ambient temperature Input of data obtained at the Abu Dhabi solar desalination plant Instantaneous ambient temperature by hour

MEDT3

Hourly total solar radiation on tilt surface Input of data obtained at Abu Dhabi solar desalination plant Instantaneous total solar radiation on horizontal surface by hour Daily total solar radiation on tilt surface

MEDT4

Daily total solar radiation on horizontal surface

Daily mean, daily maximum and daily minimum ambient temperature

MEDT2

Daily mean, daily maximum and daily minimum ambient temperature

Four types of meteorological data can be accommodated in the program (see table 14). The type MEDT1 is used in the solar desalination plant where the hourly global radiation and hourly ambient temperature were measured. The other types of meteorological data can accommodate hourly or daily global solar radiation on horizontal surface instead of tilt surface as well as hourly temperature or mean-max-min daily temperatures.

122

Ali M. El-Nashar

9.1.4. Mathematical Models Mathematical models were developed for the different components in preparation of the simulation program “SOLDES”. The following models were developed : 1. Numerical Method to Estimate Solar Radiation on a Tilted Surface This model calculates the beam and diffuse components of solar ration on a tilted surface having the same angle as the absorber plate of the solar collectors given as input the following information: • • • •

Measured hourly solar radiation on a tilted surface and measured hourly ambient temperature (MEDT1) as the case in the solar desalination plant. Measured hourly solar radiation on a horizontal surface and measured hourly ambient temperature (MEDT2), Estimated daily total solar radiation on tilted surface and estimated mean, maximum and minimum daily ambient temperatures (MEDT3) Estimated daily total solar radiation on horizontal surface and estimated mean, maximum and minimum daily ambient temperatures (MEDT4)

The solar radiation on tilted surface, It is estimated from the following equation:

It = Ib + Id = I0 × P

1 sin( h )

1

1 + cos(α 0 ) 1 − P sin( h ) 1 × × cos(θ 0 ) + × I 0 × sin( h) × 1 − 1.4 × ln( P ) 2 2

(27)

If the atmospheric transmittance, P, is given as known data, hourly solar radiation on a tilted surface can be calculated from the above equation. However, the only data available is the hourly global radiation or the daily global radiation. Therefore, to proceed with the computer simulation, it is necessary to convert global radiation to hourly bean and diffuse components. This is achieved by estimating the hourly P values from Eq. 23 knowing the global radiation, then calculating the beam and diffuse components from the following the following equations:

Ib = I0 × P

Id =

1 sin( h )

× cos(θ 0 )

(28) 1 sin( h )

1 + cos(α 0 ) 1 1− P × 0 × sin(h) × × 2 1 − 1.4 × ln( P ) 2

(29)

When daily global radiation is available instead of the measured hourly values, the hourly solar radiation is first estimated using an iterative procedure then the beam and diffuse components are estimated from Eqs. 28 and 29[5].

Multiple Effect Distillation of Seawater Water Using Solar Energy …

123

2. Effect of Shade on Solar Radiation on Absorber Plate The amount of solar radiation on the absorber plate is less than the amount falling on a tilted surface having the same angle as the absorber plate but situated in the open. This is due to the shading cast on the absorber plate by several collector components: • • •

Shade by the neighboring absorber plates (shade length = s1) Shade by the neighboring glass tubes (shade length = s2) Shade by the header box of the collector (shade length = s3)

Figure 23 shows the shadow length due to adjacent absorber plate (s1) and adjacent glass tube (s2). Figure 24 shows the shadow length due to collector header box (s3). The three shadow areas are shown in a plan view in Figure 25. Note that the area marked #1 represe

Figure 23. Shadow effect due to adjacent plate and adjacent glass tube.

These shadow effects are particularly evident in early morning and late afternoon but have a minimal effect throughout the rest of the time. Based on the solar angles (solar altitude and solar azimuth), absorber plate dimensions and tilt angle, pitch of absorber plates, glass tube diameter and collector header box height, a model was developed to estimate the hourly value of the length of the shade cast by each of the above three effects. We present here the results of this model.

l × [cos(α c ) × tan(hN' ) + sin(α c )] − L × tan(hN' ) Morning: s1 = cos(α c ) × tan(hN' ) + sin(α c )

(30)

124

Ali M. El-Nashar

Figure 24. Shadow effect due collector header box.

Figure 25. Different shadow areas in a plan view.

Afternoon: s1 =

l × [cos(α c ) × tan(hN' ) − sin(α c )] + L × tan(hN' ) cos(α c ) × tan(hN' ) − sin(α c )

(31)

Multiple Effect Distillation of Seawater Water Using Solar Energy …

l l γ × sin(α c ) + − tan(hN' ) × [ L − × cos(α c )] ' 2 2 cos(hN ) Morning: s 2 = ' cos(α c ) × tan(hN ) + sin(α c ) l l γ − × sin(α c ) + − tan(hN' ) × [( L − × cos(α c )] ' 2 2 cos(hN ) Afternoon: s 2 = ' cos(α c ) × tan(hN ) − sin(α c )

s3 =

H × sin(γ N ) tan(hN )

125

(32)

(33)

(34)

With reference to Figure 25 we can write: Total area of absorber plate…………A = LT × l

(35)

Shadow area #1……………………...A1 = s3 × l

(36)

Shadow area #2………………………A2 = s1 × (LT – s3)

(37)

Shadow area #3……………………….A3 = (s2 – s1) × (LT – s3)

(38)

The area exposed to direct solar radiation is therefore the difference between the total area of the absorber plate, A, and the three shadow area, i.e. A4 = A – (A1 + A2 + A3)

(39)

It is convenient to introduce the ratios R1, R2, R3 and R4 such that

s3 × l A s × ( LT − s3 ) R2 = 1 A ( s 2 − s1 ) × ( LT − s3 ) R3 = A (l − s 2 ) × ( LT − s3 ) R4 = A R1 =

(40)

It is to be noted that areas #1 and #2 are completely shadowed by solid obstacles which does not transmit any radiation whereas area #3 (shadow due to adjacent glass tube) has an attenuated solar radiation due to the transmittance of three laters of glass through which each

126

Ali M. El-Nashar

solar ray has to travel (see Figure 23). It is assumed that the daily average transmittance of the three glass layers is equal to 0.7. This assumption is suggested by the manufacturer (Sanyo) and is based on detailed hourly computer simulation. Consequently, the net beam radiation on the absorber plate can be expressed by the equation:

I b' = I b × ( R4 + 0.7 × R3 )

(41)

The net total radiation on the absorber plate is obtained by adding the diffuse component to the beam component

I t' = I b' + I d

(42)

Ambient temperature model When hourly ambient temperature data is not available, hourly data can obtained from knowledge of the daily mean, maximum and minimum values using a model developed for this purpose. 3. Effect of Dust on Transmittance of Glass Tubes The effect of dust on the transmittance of the glass tubes varies with the season. The heat collection amount drops sharply especially when there is a sandstorm and dust accumulates rapidly. This causes the transmittance of the glass tubes to drop. When it rains, on the other hand, the transmittance is restored because the rainfall washes away the dust. Climatic conditions in the UAE are such that there is little rainfall, but nevertheless, the cleaning effect of rainfall appears in the collector measurements. In the UAE, moreover, the temperature differences between day and night are relatively large; in the mornings, dew accumulates on glass tube surface and they become damp, so mucu so that water droplets sometimes fall from them. If the level of dew accumulation is slight, dust can easily adhere to the tube surface, but if the extent of dew is great enough, it will serve to wash away dust from the tube. The relationship between the cumulative level of dust affecting the transmittance of the glass tube and the cleansing effect of various natural climatic conditions is extremely complex. In order to develop a model of the effect of dust accumulation on the transmittance of the glass tube, it was assumed that the clean glass tube transmittance of 98% is restored after each tube cleaning and that following a tube cleaning the transmittance drops exponentially due to dust effect according to the equation (see Sayigh et al. [..]):

τ −τ m = exp(−0.055 × N ) τ 0 −τ m

(43)

Where τ is the transmittance after N days has elapsed since cleaning, τm is the transmittance after one month has elapsed since cleaning, τ0 is the transmittance immediately measured after cleaning (= 98%). The monthly drop in transmittance (τ0–τm) is dependent on the month with summer months experiencing larger drops than winter months due to the sandstorm prevailing mainly in summer months.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

127

4. Control of Heat Collecting Operation The solar collector field is operated by the solar controller which uses a solar cell to switch the heat collecting pump on and off depending on the solar radiation. The solar radiation condition for pump startup, Ion (kcal/hm2), and shutdown, Ioff (kcal/hm2), are determined from the measured values of the heat accumulator outlet temperature, Tw1 (oC) and the ambient temperature, Ta (oC) according to the relations:

I on ≥ 5 × (Tw1 − Ta ) − 25 I off < 5 × (Tw1 − Ta ) − 10

(44)

5. Heat Collection Amount from Solar Collector Field The heat collection efficiency is defined as the amount of heat collected divided by the amount of solar radiation falling on the absorber plates of the solar collector field and is expressed by the following formula:

η c = α + β .x + γ . x 2

(45)

where ηc is the collector efficiency, α,β,γ are constant parameters and x is a variable defined

1 − (Tc1 + Tc 2 ) − Ta , Tc1 and Tc2 are the collector field inlet and outlet water as: x = 2 I t' temperature and I’t is the solar radiation on the absorber plate. The rate of heat collected, Qc, from the solar field is obtained from the equation:

Qc = Ac .η c .I t' = mc .C p .(Tc 2 − Tc1 ) (46) Using Eqs. 34 and 35 the collector outlet temperature can be obtained as follows:

Tc 2 =

− C 2 − C 22 − C1C 3 C1

(47)

Where C1, C2 and C3 are given by the following equations:

C1 = γ . Ac C 2 = γ . Ac (Tc1 − 2Ta ) + β . Ac .I t' − 2mc .C p .I t' C 3 = γ . Ac .(Tc1 − 2Ta ) 2 + 2 βAc .I t' (Tc1 − 2Ta ) + 4α . Ac .I t'2 + 4mc .C p .I t' .Tc1

(48)

128

Ali M. El-Nashar

With this equation the collector outlet water temperature can be obtained at each hour of the day given the inlet water temperature, the solar radiation, water flow rate, and the collector parameters αi,β and γ.

6. Evaporator Performance A simple model is used for predicting the performance of the MED evaporator at part load given the evaporator’s capacity, maximum brine temperature, number of effects, heating water flow rate, heating water temperature and seawater temperature. The aim is to calculate the hourly distillate production, the hourly heating water outlet temperature and the hourly pumping power requirement. The performance ratio and specific heat consumption of the evaporator depends on the number of effect and was estimated from the following equations provided by the manufacturer (Sasakura): Number of effects N ≤ 13:

PR = −1.875 × 10 −2 × N 2 + 1.15 × N − 1.625

(49)

Number of effects N > 13:

PR = −2.500 × 10 −3 × N 2 + 0.625 × N + 2.56

(50)

Specific heat consumption, SPC (kcal/kg):

SHC =

526 PR

(51)

The effective temperature difference per single effect δT at the design condition was estimated from the equation:

δT =

(T2 d − T1d − BPE av × N ) N

(52)

where T2d is the design brine maximum temperature in the first effect (top effect), T1d is the design brine temperature in the last effect (bottom effect), BPEav is the average boiling point elevation in the N effects. For the evaporator in the solar desalination plant the following values are used: T2d = 68oC, T1d = 40.7oC, BPEav = 0.71oC. The average heat transfer coefficient for the N effects, U (kcal/h m2oC) was estimated from:

U = (1888 × L + 1313) ×

C (T1 ) + C (T2 ) 2

(53)

where C is a correction coefficient which is dependent on the brine temperature, L is the load. The design value is that when L = 1.0, i.,.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

U d = (1888 + 1313) ×

C (T1 ) + C (T2 ) 2

129

(54)

The correction coefficient can be expressed by the equation:

C (Tb ) = −0.4678 + 0.050 × Tb − 0.0005 × Tb2 + 0.17 × 10 −5 × Tb3

(55)

The operating condition of the evaporator depends on the load parameter L =(Md/Md100%). The operating temperatures for each load are evaluated as follows: Calculate the brine temperature in the last effect from knowledge of the seawater temperature and the condenser load:

T2 = Tc 2 + 1.2 = Tc1 +

Qc + 1.2 m sw × C p

= Tc1 +

0.9Qh + 1.2 m sw × C p

(56)

In this equation the assumption is made that the condenser load, Qc, is 90% of the heating load, Qh due to heat loss to the environment. It is also assumed that the last effect brine temperature is smaller than the condenser outlet temperature by 1.2oC. These assumptions are verified by actual measurements at the evaporator of the solar plant. Calculate an approximate value for the first effect temperature, T1,

T1 = T2 + ( L × δT + BPE ) × N

(57)

Calculate an average overall heat transfer coefficient for the N effects,

U = (1888 × L + 1313) × •

(58)

Calculate an improved value for the first effect temperature, T1,

T1 = T2 + ( L × δT × •

C (T1 ) + C (T2 ) 2

U + BPE ) × N Ud

Iterate to get improved values for T1.

The heating water outlet temperature was assumed to be equal to the first effect temperature: Th2 = T1 and the heating water inlet temperature is calculated from a heat balance equation:

130

Ali M. El-Nashar

Th1 = Th 2 +

Qh mh × C p

(59)

The power consumption (in kW) of the evaporator was assumed to depend mainly on its capacity and was estimated from the following equations provided by the manufacturer (Sasakura): Evaporator capacity md ≤ 500 m3/day:

P = −1.25 × 10 −5 × md2 + 9.0 × 10 −2 × md + 5.125 + Z

(60)

Evaporator capacity md ≥ 1000 m3/day:

P = −6.00 × 10 −6 × md2 + 5.1 × 10 −2 × md + 32.0 + Z

(61)

Evaporator capacity 500 m3/day < md < 1000 m3/day

P = −4.48 × 10 −5 × md2 + 0.1272 × md − 5.4 + Z

(62)

where Z is the pumping power (kW) of the vacuum pump and is given by:

Z = 0.050535 × md + 0.93045

(63)

9.2. Comparison of Simulation and Actually Measured Values In order to determine the accuracy of the simulation program, a comparison between the simulation results and the results from actual plant operation was conducted for January and June 1985. The operating condition of the test plant underwent various changes over the test period and are not constant. The plant had also experienced several power failures that caused complete shutdown and the plant had to be restarted manually after the restoration of power. The power failures that occurred during the test period (January and June 1985) are listed in table 15. Table 15. Power failures causing emergency shutdown Month January 1985

June 1985

Date 21 28 29 16 22 24

Time 11:10 16:30 14:30 0:41 9:35 13:15

Multiple Effect Distillation of Seawater Water Using Solar Energy …

131

In order to account for the variation in the operating parameters occurring during the month as well as the emergency shutdowns that occurs during each month, each monthly period is divided into a number of smaller periods with the operating parameters maintained constant during each small period. The computer program is run for each small period using the prevailing operating condition of the test plant. The operating parameters which were input to the program are: • • • • • • •

Collector absorber area in service (January: 1543.5 m2 – 1837.5 m2, June: 1543.5 m2) Simulation start and end dates (January: 1st – 20th, June: 1st – 15th) Temperature of heat accumulator at the start of simulation period (bottom 64.0oC, top 76.5oC) Heating water flow rate (16.5 m3/hr) Monthly bypass valve open and close temperatures (open: 70.0oC, close: 73.0oC) Monthly evaporator start and stop temperatures (start: 72.0oC, stop: 66.0oC) Specification of collector cleaning days (January: 1st, 30th, June: 8th)

Figure 26 shows the measured and simulation results of the daily net amount of heat collected by the solar collector field and delivered to the heat accumulator during the month of January 1985. This amount is equivalent to the total heat collected minus the heat loss due to the collector piping system as well the heat loss due to dust effect on the glass tubes of the collectors. January is usually characterized by severe variations in the daily solar radiation and this variation is reflected on the daily amount of heat collected as shown in the figure. Figure 27 shows the measured and simulation results of the daily water production for January 1985. The simulation results appear to be in reasonable agreement with the measured values with the exception of January 6 for which there is a large discrepancy between measured and simulation values. This is because, while the operation of the evaporator was stopped at 11:30 due a drop in the temperature of the heating water in the actual plant, operation of the evaporator was continued in the simulation until 19:30 when it was stopped. This, in turn, was due to the fact that in the simulation the temperature of the heating water was slightly higher than the set level at which operation was to stop. Daily Heat Collected for January 1985

Daily heat collected (10 6 kcal/day)

6 5 4 3 2 1

Measured

0

Simulation 0

5

Pow er failure 10

15 Day

Figure 26. Daily heat collected for January 1985.

20

25

30

132

Ali M. El-Nashar

Water production (m3/day)

Daily Wate r Production (m 3 /day) for Januar y 1985 140 120 100 80 60 40 20 0

Pow er 0

10

20

30 M e as ur e d

Day

Sim ulation

Figure 27. Daily water production for January 1985.

Figure 28 shows the simulation and measured values of the net amount of heat collected for June. The agreement seems to be quite good. Figure 29 shows the results of the daily water production for June. It is observed that the simulation result for the water production on June 1 is low. It appears that this happened because the initial temperature setting of the accumulator at the start of the calculations was lower than the actual temperatures. On June 24, the actual operation data are missing starting from 11:30 because of a power failure. To remove the dust on the glass tubes in June, blocks A and B were cleaned on June 4 and blocks D and E on June 8 while block F was cleaned regularly every 4 or 5 days. In the simulation, on the other hand, all blocks were presumed to have been cleaned on June 8. In June, the effect of the dust was significant and it appears that an error was produced due to the differences between the dust model and the actual dust accumulation. Nevertheless, the value for the amount of heat collected from June 10 through June 20 match relatively well since by this data the dust has been removed. As indicated here, there were areas where the simulation did not match the actual operation status exactly but, overall, the simulation was quite accurate even in cases where the evaporator was frequently started and stopped, as it was in January.

Daily heat collected (10 6 kcal/day)

Daily Heat Collected (106 kcal/day) for June 1985

7 6 5 4 3 2 1 0

Pow er

Measured Sim ulation 0

5

10

15 Day

Figure 28. Daily heat collected for June 1985.

20

25

30

Multiple Effect Distillation of Seawater Water Using Solar Energy …

133

Daily production (m3/day)

Daily Water Production (m 3/day) for June 1985 150 100 Measured

50

Pow er

Simulation

0 0

5

10

15

20

25

30

Day

Figure 29. Daily water production for June 1985.

Table 16 shows a comparison between the actually measured values and the calculation values for January and June. Although the operation of the evaporator and the heat collection operation are greatly influenced by subtle changes in the set temperature conditions for operation and the ever-changing climatic conditions, it can be seen that the two sets of figures match relatively well. Consequently, it is assumed that this program will be quite serviceable for studying the optimum operating conditions of the test plant and also for designing a similar plant under different conditions. Table 16. Comparison between measured and simulation results for January and June 1985 Heat collection amount (kcal/month)

Product water amount (m3/month)

Simulation Measured Error (%) Simulation Measured Error (%)

Jan. 1985 120,100,000 115,100,000 4.3 2,290 2,340 -2.1

Jun. 1985 152,400,000 153,100,000 -0.45 3,340 3,430 -2.6

10. EVALUATION OF THE TEST PLANT 10.1. Optimum Operating Conditions The simulation program of the test plant was run after inputting the values for the following: • • • •

Performance of individual equipment of the plant obtained from the research operation (for example, the collector efficiency y= 0.84 – 2.46 x – 1.9 x2, and the maximum capacity of the evaporator = 6.0 m3/h). Weather conditions for January through December 1985 (average solar radiation over inclined surface = 5,589 kcal/m2day, average daily atmospheric temperature = 27.4 o C, and average seawater temperature = 27.5oC). Standard operating conditions of the test plant as shown in table 17. Heating water quantity (variables) = 12, 13.5, 15, 16.5, 18, 19.5 and 21 m3/h.

134

Ali M. El-Nashar Table 17. Operating condition of test plant for simulation calculations Item Collector absorber area Heat collection water flow Frequency of solar collector cleaning Maximum brine temperature Seawater flow rate Feedwater flow rate Evaporator startup temperature Evaporator shutdown temp.

Value 1862 m2 83.6 m3/h Once a month 68oC 36.7 m3/h 17.5 m3/h Heating water temp. corresponding to 80% load Heating water temp. corresponding to 60% load

A summary of the results of the simulation is shown in table 18. As seen from this table, a heating water quantity of 16.5 m3/h results in the maximum water production: 43,400 m3/year or 118.9 m3/day. However, the plant should not be operated such that TME110 (i.e., the number of hours the plant is operated at above 110% load) exceeds 5% of any months total operating hours. The data in table 18 show the following in terms of TME110: • • •

TME110 for April is 7.6% even if the heating water flow rate is reduced to 12 m3/h. TME110 for the other months can be below 5% when the heating water flow rate is 16.5 m3/h or below. By increasing the heating water flow to 18.0 or 19.5 m3/h without allowing TME110 to exceed 5%, the water production for the month concerned can be increased over that for the heating water flow of 16.5 m3/h.

Table 18. Summary of simulation results for different heating water flow rates Month Jan.

Feb.

Mar.

H/W flow m3/h 21 19.5 18 16.5 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12

Production m3/month 2360 2530 27600 2890 2830 2740 3680 3710 3700 3680 3640 3600 3560 3680 3860 3870 3850 3820 3780 3740

TME110 hrs 11 10 0 0 0 0 267 207 109 68 46 23 5 291 233 190 126 85 41 14

Startups 28 20 11 5 4 7 1 0 0 0 0 0 0 8 2 1 1 1 1 1

Multiple Effect Distillation of Seawater Water Using Solar Energy … Table 18. (Continued). Month Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

H/W flow m3/h 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12

Production m3/month 4030 4050 4040 4010 3990 3950 3900 3940 3980 3950 3940 3910 3850 3830 3940 3930 3900 3870 3840 3800 3750 3350 3440 3420 3400 3350 3340 3310 3850 3860 3860 3830 3790 3750 3700 4050 4050 4020 3990 3960 3920 3870 3800 4010 4000 3970 3930 3890 3850

TME110 hrs 428 385 355 317 227 140 55 224 89 53 21 2 0 0 212 163 72 33 18 4 0 16 10 4 2 0 0 0 70 60 19 8 2 0 0 350 291 131 74 42 12 0 149 85 58 30 11 4 1

Startups 1 0 0 0 0 0 0 3 1 1 0 0 1 0 0 0 0 0 0 0 0 6 1 1 1 2 1 1 3 1 1 1 1 1 1 0 0 0 0 0 0 0 9 0 0 0 0 0 0

135

136

Ali M. El-Nashar Table 18. (Continued). Month

H/W flow m3/h 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12 21 19.5 18 16.5 15 13.5 12

Nov.

Dec.

Total

Production m3/month 2830 3180 3300 3340 3320 3280 3190 2150 2170 2370 2600 2650 2580 2420 41700 42800 43200 43400 43000 42600 41900

TME110 hrs 3 0 0 0 0 0 0 0 0 0 0 0 0 0 2020 1532 989 677 421 223 74

Startups 24 9 4 1 1 1 1 28 27 18 7 4 4 12 111 61 37 16 10 13 23

Table 19. Optimum operating conditions for the Abu Dhabi solar desalination plant Month

Absorber area (m2)

H/C flow (m3/h)

Jan. Feb. Mar. Apr. May. Jun. Jul. Aug. Sep. Oct. Nov. Dec.

1862 1862 1862 1764 1862 1862 1862 1862 1862 1862 1862 1862

83.6 83.6 83.6 79.2 83.6 83.6 83.6 83.6 83.6 83.6 83.6 83.6

Bypass valve Open (oC) 63 65 68 69 70 71 70 72 74 69 67 65

Bypass valve Close (oC) 67 69 72 69 74 75 74 76 78 73 71 69

H/W flow (m3/h)

F/W flow (m3/h)

Evaporator Start (oC)

Evaporator Stop (oC)

16.5 13.5 12.0 13.5 16.5 16.5 19.5 18.0 13.5 16.5 16.5 15.0

17.5 17.5 17.5 17.5 17.5 17.5 17.5 17.5 17.5 17.5 17.5 17.5

67 70 73 73 73 75 74 76 78 73 71 69

62 64 67 68 69 70 69 71 73 68 66 64

Since it was found that the quantity of heat collected in April exceeds the allowable capacity of the evaporator, the collector absorber area for April needs to be reduced in order to protect the evaporator from scaling. In order to find the number of arrays that need to be taken out of service during April, the simulation program was run for different number of arrays in operation. It was found that 4 arrays of the 76 arrays available need to be drained making the available absorber area for this month 1,764 m2. The other operating conditions are as given in table 17. These results translate into the optimum operating conditions shown in table 19.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

137

10.2. Simulation Results Simulated operation of the test plant under the optimum operating conditions noted above was carried out using the solar radiation data of 1985. The results are shown in table 20. The effective annual water production is 42,900 m3, the annual operating rate of the evaporator is 97.6% (number of evaporator starts and stops = 13), and the specific power consumption is 5.1 kWh/m3 of product water. Table 20. Simulation results at optimum operating condition Item Total radiation on tilt surface Drop in solar radiation due to dust Heat collection amount Amount of heat supplied to heat accumulator Heat supplied to evaporator Heat effectively used by evaporator

Simulation output 3,780,000,000 kcal/year 195,000,000 kcal/year 2,212,000,000 kcal/year 1,950,000,000 kcal/year 1,808,000,000 kcal/year 1,796,000,000 kcal/year

Product water Collector cleaning water Effective product water Heat collecting pump running hours Evaporator operating hours Frequency of operator startup Power consumption per m3 Anti-scalant consumption (Belgard EV) Sodium hypochlorite consumption (NaClO)

43,000 m3/year 76 m3/year 42,900 m3/year 3,655 hr/year (10.0 hr/day) 8,546 hr/year (yearly average rate 97.6%) 13 times/year 5.1 kWh/m3 1,496 kg/year (0.035 kg/m3 product) 5,577 kg/year (0.13 kg/m3 product)

Ratio 1.0 0.052 0.585 0.516 0.478 0.475

10.3. Evaluation of the Solar Plant In this section we compare the planned values used in the design of the solar plant, the experimental results of the research operation obtained during the first year of plant operation and the values introduced in the simulation program which are based on the measured values. The results are summarized in table 21. Column one of this table shows the design values; column two the results of actual measured performance data and column three shows the introduced in the simulation program.

A. The Solar Collector Field The heat collection efficiency of the solar collectors was about 9% lower than that for the basic design under the same weather conditions. The efficiency of a clean single collector was measured at the manufacturer’s laboratory under ideal operating conditions and was reported as the catalog efficiency, ηoc, which can be represented by the following equation:

ηoc = 0.913 – 2.46 x – 1.92 x2

(64)

where x is a parameter defined as : x = {(To + Ti)/2 – Ta}/It and the unit of oC hr m2/kcal. When a number of collectors are connected together to form a block in the solar collector

138

Ali M. El-Nashar

field, the efficiency of such a clean block was measured and found to fit the following equation:

ηm = 0.84 – 2.46 x – 1.92 x2

(65)

The measured efficiency is therefore lower than the catalog efficiency by about 8% for x = 0 and 9.3% for x = 0.05 oC hr m2/kcal. The drop in collector efficiency of a group of interconnected collectors can be attributed to the heat loss by internal piping and connectors. In the basic design, the influence of dust deposition on the solar collectors was assumed to cause a 10% loss of the incoming solar radiation. Measurements of the heat loss due to dust effect carried out during plant operation showed that the influence of dust deposition has a seasonal character with the loss in solar radiation varying from 4% during winter months to as much as 20% in the summer. To account for the monthly variation of the dust effect in the simulation program, a mathematical model was developed to estimate the dust effect from month to month. Heat loss from the collector piping system is another cause of heat loss that has to be accounted f-cor. The piping system consists of insulated pipes varying in diameter from 30 mm to 125 mm as well as valves, pipe supports, expansion joints and safety valves. In the basic design, heat loss from the piping system was estimated based only on heat loss from all the pipes with heat loss from valves, supports, etc. neglected. Measurements were made to estimate the total piping loss which takes into consideration all components of the piping system. For a single block, the measured heat loss from the piping system was correlated by the following formula:

Qloss-m = 66.6 (Tw – Ta)1.3 kcal/hr

(66)

The theoretical (calculated) value of the piping heat loss excluding valves, supports, etc. was estimated as:

Qloss-c = 93.5 (Tw – Ta) kcal/hr

(67)

For a (Tw – Ta) = 40oC , the measured heat loss is 2.15 times the theoretical value.

B. The Heat Accumulator The heat accumulator (HA) in the solar plant consists of 3 series-connected thermally stratified vertical cylinders. Heat loss from the accumulator tanks was estimated in the basic design as 1.05oC per day for a water temperature of 99oC and an ambient temperature of 30oC. For a storage capacity of 300 m3 of water, the estimated daily heat loss is 315,000 kcal/day. Heat loss measurements from plant heat balance indicates that daily temperature drop varies from month to month with the average yearly value of 2.61oC per day which is more than double the value used in basic design. This value is also based on a water temperature of 99oC and ambient temperature of 30oC.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

139

C. The MED Evaporator As seen from table 21, the evaporator is the only piece of equipment in which the performance in the basic design was upgraded in the simulation program: 20% up in water production capacity, and 0.2 to 7.5% less in specific heat consumption. In the basic design, the maximum evaporator capacity was 120 m3/day (5 m3/hr) and specific heat consumption is 43.8 kcal/kg- product at 35oC seawater temperature and 55,000 ppm salt concentration. This design maximum production was assumed to be achieved at a heating water (HW) temperature of 99oC and a HW flow rate of 18.4 m3/hr. Measured values of maximum production was 130 m3/day (5.42 m3/hr) at a corresponding specific heat consumption of 42.4 kcal/kg- product. Bas ic De s ign He at Balance incide nt s olar r adiation = 100%

us e ful he at to e vapor ator 36%

colle ctor he at los s 43%

e vapor ator los s 3% accum ulator colle ctor los s piping los s 3% 5%

los s by dus t de pos ition 10%

(a) Measured Heat Balance incident solar radiation = 100%

collector heat loss 36%

useful heat to evaporator 48%

loss by dust deposition 5% evaporator loss 0.3%

accum ulator loss 4%

collector piping loss 7%

(b) Figure 30. Total plant heat balance - comparison between design and simulation values.

140

Ali M. El-Nashar

The design and measured annual plant heat balance are shown in pie chart form in figure 30. In this figure it is assumed that the incident solar radiation represent 100% and the percentage of this energy going to each of the different losses are specified along with the net amount of useful energy for desalination. The effective heat input to the evaporator increased a remarkable 11.5% from 36.0% for the basic design to 47.5% in terms of the ratio to the total solar radiation quantity on the collector tilted surfaces. This sharp increase combined with the improved heat efficiency of the evaporator to produce a synergistic effect: The water production increased from an annual average of 80 m3/day for the basic design to 117.8 m3/day, and the effective water production (after subtracting the quantity of water used for cleaning the collectors) was 117.5 m3/day, a dramatic 46.9% increase. In addition to using performance data from actual plant measurement, the simulation program incorporates various improvements, which include: • •

Adopting a heating water temperature control system based on a three-way proportional control valve for efficient use of the energy of the collected heat. Reducing power consumption by adopting pumps with appropriate capacity (several pumps used in the solar plant are oversized).

Table 21. Comparison between performance values used in basic design, measured values and simulation program Item Heat collection efficiency

Basic Design Catalog efficiency: η = 0.913 – 2.46 x – 1.92 x2 x={ (Ti + To)/2-Ta}/ It o C hr m2/kcal

Heat loss from piping

Dust influence

Heat collecting system

Calculated value ignores valves and piping supports. For a single block: Qloss = 93.5 (Tw – Ta) kcal/hr Tw = water temp., Ta=air temp. It is assumed that the solar radiation is reduced by 10% due to dust effect throughout the year Bypass operation is not considered. Water enters the accumulator immediately after pump start up.

Measured value Value of η when x=0 is: Winter = 0.83 ~ 0.84 Summer = 0.80 ~ 0.81 Average = 0.82 Slope of curve is: Winter = large Summer = small Efficiency of block F includes heat loss from internal piping and dust effect Value measured for a single block is: Qloss = 66.6 (Tw – Ta)1.3 kcal/hr

Dust effect varies seasonally. In winter about 4% and in summer reaches more than 20%. A heat collecting system using bypass operation and control valves are used.

Simulation * Efficiency used for simulation is: η = 0.84 – 2.46 –1.92 x2 * For completely clean condition, the value of η for x=0 has been increased by 0.2 to 0.84 because of the fact that dust isn’t completely removed after cleaning. Value used is: Qloss = 66.6 (Tw – Ta)1.3 kcal/hr

Based on measured values, a model was developed to estimate the monthly dust influence. Heat collection amount takes into consideration bypass operation.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

141

Table 21. (Continued). Item Heart accumulator

Basic Design * When calculated heat loss from accumulator is converted to ΔT = Tw – Ta = 69oC, the heat loss is Qloss = 312123 kcal/day. * Heat loss is based on complete mixing in the tank and no stratification considered

Measured value * Based on measured values the heat loss at ΔT = 69oC is Qloss = 765000 kcal/day. * Thermal stratification ratio was measured as 7.5% during daytime and 6.5% at night

Simulation * After heat loss is calculated, compensation is performed using (Qloss)measured/(Qloss)calc.=2.45 * Heat accumulator is divided into 1000 equal parts, mixing takes place at the top and bottom of tank. Mixing at top = 7.5% and mixing at bottom = 5.0%.

Operation of evaporator when excessive heat is collected

The evaporator can be operated up to a boiling temperature of 100oC.

Adjust the heat collecting area or the heating water flow rate so that the load of the evaporator does not exceed 110%.

Evaporator water production Evaporator operating flow rates

Max. capacity = 5 m3/hr

Max. capacity = 6.46 m3/hr

Install a 3-way proportional control valve between the heat accumulator and the evaporator in order to control the heating water temperature so that the load of the evaporator does not exceed 110%. Max. capacity = 6.0 m3/hr

* Heating water flow rate = 18.4 m3/hr * Feedwater to evaporator = 10.5 – 17.5 m3/hr * Seawater to evaporator = 39.4 m3/hr

* Heating water flow rate = 12 – 17 m3/hr * Feedwater flow rate = 17 – 21 m3/hr * Seawater flow rate = 36 –40 m3/hr

* Heating water flow rate = 11.2 – 22.4 m3/hr * Feedwater flow rate = 17.5 m3/hr * Seawater flow rate = 20.2 –40.3 m3/hr

Operating temp. at 5 m3/day load: H/W inlet temp. = 99oC H/W outlet temp. = 87oC 1st effect temp. = 83oC 18th effect temp. = 43oC SW temp.(design value) = 35oC SW concentration = 55,000 ppm

Operating temp. at 5 m3/day : Jan. Jul. 71oiC 79oC 58oC 66oC 66oC 58oC 28oC 39oC o 21.8 C 32.4oC

Operating temp. at 5 m3/day load 78.5oC 67.5oC 67.5oC 40.8oC 35oC

51,200

52,000

Evaporator operating temperatures

53,500

11. ECONOMIC CONSIDERATIONS AND COMPARISON WITH CONVENTIONAL MED PLANTS On the basis of data obtained from the research operation of the test plant, an economic study was made to estimate the cost of water from solar MED plants with a capacity in the range 100 – 1000 m3/day and compare the results with conventional MED plant. The pumping power of the solar and conventional MED plant is to be provided by a diesel generator of appropriate capacity. Steam for the conventional MED plant is to be provided by a steam generator. The product water costs of these practical plants were calculated and are

142

Ali M. El-Nashar

shown below. Note that these trial designs incorporate improvements such as the inclusion of the three-way proportional control valve.

11.1. Basic Economic Parameters The product water cost was calculated using the following economic parameters: •

•

Expected life of plant components o Evaporator…………………..20 years o Heat accumulator……………20 years o Solar collectors………………20 years o Diesel generator……………..10 years Interest rate……………………………8% per annum

11.2. Capital Equipment Cost 11.2.1. Capital Cost of MED Evaporator The cost of MED evaporator is based on budget offers from different manufacturers and on cost information available in the open literature. The capital cost depends on the design capacity, number of effects and maximum brine temperature according to the correlation given by Fosselard et al.[6]:

C ev = 5,005,000( where Cev md PR Tbmax

md 0.7 PR 0.5 70 0.47 ) ( ) ( ) 2,500 8 Tb max

100 ≤ md ≤ 1000

(68)

= evaporator capital cost, $ = rated (design) capacity, m3 day–1 = performance ratio = design maximum brine temperature, °C.

The performance ratio (PR) of the evaporator (defined as the distillate output in kg per 1055 kJ of heat input) is related to the number of effects by the following equation[1]:

PR = −0.809 + 0.932N − 0.0091N 2

(69)

11.2.2. Capital Cost of Solar Thermal Collectors Solar collectors used for this application should be capable of producing hot water at a temperature ranging between 70 and 90°C. Evacuated tube collectors and high-efficiency flat plate collectors can be used to produce hot water at a temperature in excess of 80°C. The specific cost of the solar collectors are assumed to range 200 - 600 $ m–2 (flat plate and evacuated tube collectors). The cost is assumed to include both the solar collector proper as well as the support structure, piping, valves, etc.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

143

11.2.3. Capital Cost of Heat Accumulator The heat accumulator is assumed to be a vertical cylindrical tank made of mild steel with a thick layer of fiberglass insulation to reduce heat loss. The tank is designed to operate at atmospheric pressure and is provided with a pressure relief valve as a safety measure. Hot water from the collector field is supplied to the tank at the top via a special water distribution grid that ensures that hot water diffuses slowly through the surrounding water with causing too much turbulence in order to enhance thermal stratification through the tank. The capital cost of the heat accumulator as obtained from manufacturer’s data is obtained from the following relation

C st = 7803.9 × m st0.525 100 ≤ mst ≤ 600 where, Cst mst

(70)

= cost of heat accumulator, $ = storage capacity, m3

11.2.4. Capital Cost of Steam Generator for Conventional MED Systems Low pressure and low capacity steam generators are required to supply the MES evaporator with the low-temperature thermal energy necessary to drive the unit. The capacity of the steam generator depends on the capacity of the MES unit as well as its performance ratio. For a unit having a PR = 13 and having a capacity 200 m3 day–1 at design conditions, requires an estimated 0.6 ton h–1 of low-pressure steam. A fire-tube packaged steam generator producing steam at 10 bar and having an efficiency (LHV) of 86 per cent is considered appropriate. The capital cost of such unit, Cb ($) is obtained from [7] and adjusted to the current cost level using the Marshall and Swift Equipment Cost Index. The resulting correlation is shown below:

C b = 115,700 + 18,200 × where, Cb ms

ms 12

0.15 ≤ ms ≤ 12

(71)

= capital cost of steam generator, $ = steam generating capacity, ton h–1

11.2.5. Capital Cost of Diesel Generator A diesel generator whose capacity will obviously depend on the capacity of the plant itself can supply the electrical demand of the desalination plant. The capital cost of such diesel generator is obtained from the following relation:

C dg = 50 800 (

P 0 .5494 ) 40

where Cdg is the cost in $ and P is the rated power in kW.

(72)

144

Ali M. El-Nashar

11.3. Operation and Maintenance Expenses 11.3.1. Consumable Chemical Expenses The cost of consumable items is as follows: •

Cost of chemicals o Scale preventive (Belgard EV) o Anti-corrosive (Nalco 2000) o Seawater disinfection (NaClO) o Caustic soda o Sodiun bisulfite o Coagulant o Calcium chloride

3.42 $/kg 9 $/kg 0.40 $/kg 1.78 $/kg 0.97 $/kg 3.4 $/kg 0.81 $/kg

Chemical consumption for the solar MED plant were estimated from the following assumptions: o o o o o o

Belgard EV to be added to the feed water at 10 ppm Nalco 2000 to be added at 7,000 ppm to the makeup water, amounting to 30% of the accumulator capacity per year. Sodium hypochlorite (NaClO) to be added to intake seawater at 18 ppm Sodium hypochlorite (NaClO) to be added to product water at 2 ppm Sodium bicarbonate (NaHCO3) to be added to product water at 23 ppm Calcium chloride (CaCl2) to be added to product water at 18 ppm

11.3.2. Electrical Energy Consumption Electrical energy consumption by the solar-MED and conventional MED systems are provided by a diesel-generator of appropriate capacity. • • • • • • •

Fuel consumption of diesel generator Cost of one liter of diesel oil (cf) One barrel of oil = 167 liter Cost of one barrel of oil Cost of electricity Electricity consumption (solar MED) Electricity consumption (conv. MED)

3.0 kWh per liter of diesel oil cbarrel/167 $/liter 50 –120 $/barrel = cf/3.0 $/kWh kWh/m3 product water kWh/m3 product water

The electrical power required by the solar MED plant consists of the following components: •

Pumping power for the evaporator which is given by:

Pev = ( 4 . 805 + 0 . 094 × m d − 2 . 1 × 10

−5

m d2 ) × ( 2 . 1 + 0 . 06 × PR ) / 2 . 88

(73)

Multiple Effect Distillation of Seawater Water Using Solar Energy … •

Power of the vacuum pump

Pvac = −1.866 + 0.057 × md − 2.7 × 10 −6 × md2 •

(74)

Power of the heat collecting pump

Pc = [83.6 × ( •

145

Ac 1.13 )] × 2.691 × 10 − 2 1862

(75)

Power of the heating water pump

Phw = 1.5 × (

md ) 130

(76)

Thus, the total pumping power, kW, can be expressed as

P1 = Pev + Pvac + Pc + Phw

(77)

The electrical power required by the conventional MED plant consists of the following components: • •

Pumping power of the evaporator as before Pumping power for the steam generator

The pumping power of the steam generator was estimated from the relation:

Pb = 65.0 ×

ms 15.0

(78)

where ms is the design capacity of the steam generator, ton/hr. The steam generator is capable of providing low-pressure heating steam for the MED evaporator and medium pressure steam to a steam ejector to create the operating vacuum inside the evaporator. The total electrical power required by the conventional MED plant is:

P2 = Pev + Pb

(79)

11.3.3. Spare Parts Cost An amount of 2% of the direct capital investment has been estimated as yearly cost of spare parts.

146

Ali M. El-Nashar

11.3.4. Personnel Cost The staff required for operation and preventive maintenance (for 3 shift operation) are assumed as follows: Plant capacity

Supervisor

Mechanic

Electrician

Chemist

helper

100 – 300 m3/day

1

2

1

0.5

2

300 – 1000 m3/day

1

4

2

1

4

Monthly salary $

1000

600

600

600

400

The above estimates results in a cost of water due to personnel of 1.5 $/m3 for a 100 m /day plant and 0.27 $/m3 for a 1000 m3/day plant. 3

11.4. Estimating the Cost of Water Produced The estimates of the cost of water that are given below are based on the life-cycle cost analysis of the plant which includes capital, OandM and fuel costs. The total life-cycle cost, TLC, is the sum total of the capitals cost plus the present value of all future OandM annual expenses:

TLC = Ctot + PW (F) + PW (OM)

(80)

where, Ctot = total capital cost including engineering, installation and management costs, PW (F) = present worth of all annual fuel costs incurred throughout the lifetime of the plant (for conventional plants), PW (OM) = present worth of all annual OandM expenses incurred throughout plant lifetime. The present worth of the annual fuel and OandM expenses are calculated from the following expressions [5]

PW ( F ) = F0 (

1+ g f k − gf

)[1 − (

1+ g f 1+ k

)N ]

1 + g om 1 + g om N )[1 − ( ) ] PW (OM ) = OM 0 ( 1+ k k − g om where F0 OMo gf gom k

= fuel cost in the first year of operation, $ = OandM cost in the first year of operation, $ = annual fuel escalation rate (assumed 0.03) = annual OandM cost escalation rate (assumed 0.03) = interest rate (assumed 0.08)

(81)

Multiple Effect Distillation of Seawater Water Using Solar Energy … N

147

= plant lifetime, years (N = 20 years)

The cost of water, cw, ($/m3) was calculated as follows

cw =

where, md PF

TLC md (365) N ( PF )

(82)

= desalination plant rated capacity, m3/day = plant factor (assumed 0.85)

It should be noted that all water costs given in this section do not include seawater intake and outfall costs or cost of land. These additional costs are very much site dependent and has to be added to the cost estimates given here.

12. RESULTS OF THE ECONOMIC STUDY Figure 31 shows how the cost of water varies with the number of effects and cost of collector (in $/m2) for a solar MED plant having a capacity of 130 m3/day which is identical to that of the test plant. The design maximum temperature Tb = 90oC and the fuel cost is assumed to be cb = 50 $/barrel. The cost of water is seen to be quite sensitive to the number of effects and to a lesser degree on the cost of collector with the cost of water varying between 8 $/m3 and 4 $/m3. The water cost tends to decrease with increasing the number of effects and reducing the cost of collector. The increase in the number of effects results in an increase in the performance ratio and thus leads to a reduction in the heat demand of the evaporator for a given water production. Since this heat demand is produced by a field of solar collectors, the reduction in this demand is expected to cause a similar reduction in the area of the collector field and thus in its corresponding capital cost. On the other hand, the increase in the number of effects results also in an increase of the capital cost of the evaporator due to the increased structural complexity of the evaporator but this increase in capital cost is usually small compared to the benefits of larger number of effects. Figure 32 shows the influence of the number of effects and cost of fuel (expressed in $/barrel) on the resulting cost of water for a solar MED plant having a capacity of 130 m3/day. The evaporator is assumed to have a maximum brine temperature, Tb = 90oC and the collector cost is assumed to be ccol = 300 $/m2. It can be seen that increasing the cost of fuel to the diesel generator results in an increase in the water cost due to the increase of the cost of electricity produced by the diesel generator. Figure 33 demonstrates how the plant capacity and fuel cost affect the cost of water. As expected, higher plant capacity results in a lower water cost and higher fuel cost results in a higher water cost. Figure 34 shows the effect of the cost of collector on the resulting water cost for solar MED plant capacity ranging from 100 m3/day to 1000 m3/day. It can be seen that the collector cost has a vital contribution to the cost of water. For a plant capacity of 1000 m3/day, the cost

148

Ali M. El-Nashar

of water is 2.24 $/m3 for a collector cost of 200 $/m2 and is 3.35 $/m3 for a collector cost of 600 $/m2. The effect of the cost of fuel on the water cost is displayed in Figure 35.

Figure 31. Effect of the number of effects and the collector cost on the resulting water cost – fuel cost, cbarrel = 50 $/barrel, max. brine temp. , Tb = 90oC (solar MED plant capacity = 130 m3/day).

Figure 32. Effect of the number of effects and the fuel cost on the resulting cost of water – collector cost, ccol = 300 $/m2, max. brine temp., Tb = 90oC (solar MED plant capacity = 130 m3/day).

Multiple Effect Distillation of Seawater Water Using Solar Energy …

149

Figure 33. Cost of water as a function of plant capacity and fuel cost, Neff = 25, Tb = 70oC, ccol =300 $/m2 (solar MED plant).

Figure 34. Cost of water as a function of plant capacity and collector cost, Neff = 25, Tb = 90oC, cbarrel = 50 $/barrel (solar MED plant).

Figure 35. for different conventional MED plants of different capacities. For a plant capacity of 1000 m3/day and a fuel cost of 60 $/barrel (close to the current oil price), the cost of water is 2.68 $/m3

150

Ali M. El-Nashar

Figure 36. Cost of water as a function of plant capacity and cost of fuel, Neff = 25, Tb = 70oC (conventional MED plant).

Figure 37 is 3-dimentional plot of the cost of water versus the number of effects and cost of fuel for a conventional MED plant having a capacity of 130 m3/day. A conventional MED plant with this capacity produces water at a cost of 3.92 $/m3 assuming an oil price of 60 $/barrel. As the oil price doubles to 120 $/barrel, the cost of water is expected to reach 4.99 $/m3. The corresponding water cost from a solar MED plant is 4.58 $/m3 at a collector cost of 300 $/m2.

Figure 37. Effect of the number of effects and the cost of fuel on the resulting cost of water – max. brine temp., Tb = 90oC (conventional MED plant capacity = 130 m3/day).

Multiple Effect Distillation of Seawater Water Using Solar Energy …

151

The difference between the cost of water of a solar MED plant (cw1) and the corresponding cost of a conventional MED plant (cw2) is shown in Figure 38 for a capacity of 130 m3/day and for different collector and fuel costs. It can be seen that for higher fuel costs and lower collector costs the difference (cw1 – cw2) is negative indicating that the cost of water from a solar MED plant is cheaper than that of a conventional MED plant having the same capacity.

Figure 38. Difference in cost of water between solar MED plant, cw1, and conventional MED plant, cw2 , (cw1 - cw2) ,$/m3.

13. CONCLUSION The research operation of the test plant, conducted jointly between ENAA and WED, ended successfully at the end of October 1985, as scheduled, giving many useful results that are summarized below.

1. System Reliability The reliability of the automated continuous operation of the test plant for its first year of operation has been successfully demonstrated. However, the plant has suffered from pump trouble, etc. associated with plants of the same type in the early stage of development. The plant’s heat collecting subsystem, the heat accumulator subsystem and evaporator subsystem have also proved to have no particular problems:

152

Ali M. El-Nashar • •

•

Corrosion of the heat collectors and heat accumulators was prevented by the use of corrosion prevention chemical additive; Scaling in the evaporator was prevented by using a scale inhibitor which maintained a good evaporator performance throughout the first year without any need for an acid cleaning; Experimental results showed that there was no problem with the vacuum in the evacuated glass tube collectors, despite early worries that the vacuum might deteriorate if the collectors were left at no load (drained condition) without operation under conditions of high solar radiation typical of the Middle East. The collector joints were found to pose no problems in the first year of operation.

The above results have led to estimating the life expectancy of the solar collectors, accumulators and evaporator at 20 years. The reliability of the plant has been further demonstrated during its succeeding years of operation.

2. Response to Varying Weather Condition Weather conditions, including solar radiation, ambient temperature, seawater temperature and humidity varies widely according to the time of day and season. For example, the solar radiation on a horizontal surface varied from 2,150 kcal/m2 day to 7,080 kcal/m2 and the daily average temperature varied between 16.7oC and 38.0oC for 1985 in Abu Dhabi. Despite such wide fluctuations, the results of the simulation program call for the evaporator of the test plant to be shutdown only 13 times a year for lack of thermal charge in the accumulator. Based on the simulation results also, the plant’s annual operating time is 8,546 hours which translate into an availability of 97.6%. Thus, the test plant has a very good adaptability to weather conditions.

3. Establishment of a Method for Cleaning of Solar Collectors One of the early worries about the test plant was whether there was an easy and economical method for removing fouling material (dust) due to sandstorms and polluted air. In an effort to solve this problem, an investigation was made of solar installations in the Middle East, and based on the results of this investigation, a preliminary test was made for collector cleaning. As a result, high-pressure water spray gun was adopted as the method of cleaning the solar collectors. This method worked very well. A cleaning once a month reduced the heat loss from the solar collectors from the planned 10% to 5.2%. The annual amount of water used for this purpose was about 0.2% of the total water production. The applicability of this cleaning method to practical plants was therefore demonstrated.

4. Comparison of the Test Plant Results with that of the Basic Design A. Improved Effective Use of Solar Radiation It was difficult to make an accurate comparison of the effective use of solar radiation between the test plant and the basic design because the weather conditions were different.

Multiple Effect Distillation of Seawater Water Using Solar Energy …

153

Nevertheless, the table below shows that the ratio of solar radiation used for water production to the total solar radiation obtained using the simulation program (47.5%) increased markedly over the value that was predicted by the basic design (36.0%). Table 22. Plant energy balance Energy quantity Total solar radiation on tilted surface Heat loss due to fouling Heat loss from solar collectors Quantity of Heat loss from collector piping collected heat Heat loss from heat accumulators Heat loss from evaporator Heat quantity used for desalination

Basic design 100% 10% 42.7% 5.3% 2.4% 2.6% 36.0%

Simulation 100% 5.2% 36.3% 6.9% 3.8% 0.3% 47.5%

The reasons for this marked increase are: •

• • •

The heat loss due to dust deposition was reduced by almost half, although the heat loss from the solar collectors was actually larger than was planned in the basic design. The quantity of collected heat was larger than was planned in the basic design because it was possible to collect heat at lower temperature. The number of evaporator starts and stops was considerably reduced over the planned number, which in turn reduced the heat loss from the evaporator. Improvements made on the test plant: Motor-operated valves, check valves, etc. were installed at the inlet and outlet of the solar collectors to make provision for power failure during the day. If a power failure occurs, this installation automatically drains out the solar collector water, preventing water hammer due to overheated collector water.

For the test plant, there was no choice but to reduce the collector absorber area when dealing with peak radiation intensity during summer time. If all the collectors installed are to be fully used, however, it is suggested that a three-way proportional control valve is to be added between the evaporator and the heat accumulators and to mix some of the return heating water from the evaporator with the hot water from the accumulator to prevent the evaporator from overloading.

B. Development of Computer Simulation Program A simulation program was prepared based on the results of the research operation of the test plant. It is a substantially improved version of the simulation program over that used in the basic design. The results of running this new program showed water productions of 102% and 97.4% of the test plant’s actual performance for January and June 1985, respectively, and there was good agreement. The simulation program is used to calculate the water production for data input into it, and involves the weather conditions at the plant concerned, such as solar radiation, ambient temperature and seawater temperature; and the specifications and capacities of the major individual plant components – capacity (100 to 2,000 m3/day), maximum brine temperature

154

Ali M. El-Nashar

(60 to 80oC), number of effects (13 to 32) for the evaporator, for example; and the operating conditions of the plant (eg. Flow rate and temperature of the heating water). If this simulation calculation is repeated for various sets of data, the optimum combination for the geographical area concerned, including the specifications and capacities of solar collectors, heat accumulator and evaporator, can be determined. Other data can also be obtained that may be useful in the selection of pump capacities, piping lengths, angles of absorber plates of solar collectors, etc. Therefore, the simulation program can be used to facilitate precise conceptual design of solar plants of similar design.

C. Water Production Costs • The collector cost has a vital contribution to the cost of water. For a plant capacity of 1000 m3/day, the cost of water is 2.24 $/m3 for a collector cost of 200 $/m2 and is 3.35 $/m3 for a collector cost of 600 $/m2. • For a plant capacity of 130 m3/day which is identical to that of the test plant, the cost of water is quite sensitive to the number of effects and to a lesser degree on the cost of collector with the cost of water varying between 8 $/m3 and 4 $/m3. • The difference between the cost of water from a solar MED plant and the corresponding cost from a conventional MED plant can be in favor for high fuel cost and low collector cost situation indicating that the cost of water from a solar MED plant can be cheaper than that of a conventional MED plant under these conditions.

14. NOMENCLATURE Ac BPE Cb cbarrel ccol Cdg Cev Cst Ctot cf cw1 cw2 di Fo gf gom H h hN I I0 Ib

collector absorber area (m2) boiling point elevation, oC capital cost of steam generator, $ cost of 1 barrel of oil, $ specific collector cost, $/m2 capital cost of diesel generator, $ capital cost of evaporator, $ capital cost of heat storage tank, $ total capital cost of plant, $ cost of 1 liter of fuel, $ cost of water from a solar MED plant cost of water from a conventional MED plant insulation thickness of pipe i fuel cost in the first year of plant operation, $ fuel cost annual escalation rate annual escalation rate for OandM expenses height of collector header box (m) solar altitude (rad) solar altitude measured with respect to collector plane (rad) solar radiation on horizontal surface (kcal/h m2) solar radiation at the outer limit of the atmosphere (kcal/h m2) beam component of solar radiation (kcal/h m2)

Multiple Effect Distillation of Seawater Water Using Solar Energy … Id It k L l li m md mst ms N OMo P Pev Pvac Pc Phw Pb P1 P2 PF PW(Fo) PW(OMo) Qa Qc Qev Qph ri s1 s2 s3 T Tb Tbmax TLC U Va x

diffuse component of solar radiation (kcal/h m2) solar radiation on tilted surface (kcal/h m2) interest rate pitch of absorber plate (m), also latent heat of vaporization (kcal/kg) width of absorber plate (m) length of pipe i, m mass flow rate (kg /s) product flow rate ((kg/ s) heat storage capacity, m3 steam flow rate, ton/hr No. of effects; also lifetime in years OandM cost in the first year of plant operation, $ pump power (kW), atmospheric transmittance pumping power of MED evaporator, kW pumping power of vacuum pump, kW power of heat collecting pump, kW power of heating water pump, kW power of steam generator pump, kW total power of solar MED plant, kW total power of conventional MED plant, kW plant factor present worth of fuel cost, $ present worth of OandM cost, $ rate of heat supplied to accumulator, kcal/ h rate of heat collected by solar field and evaporator condenser, kcal/ h rate of heat supplied to evaporator, kcal/ h rate of heat transfer in preheaters, kcal/ h radius of pipe i, m shadow length by neighboring absorber plate, m shadow length of neighboring glass tube, m shadow length of collector header box, m temperature, oC brine temperature, oC maximum brine temperature, oC total life cycle cost, $ overall heat transfer coefficient, kcal/ h m2 oC wind speed, m/s collector parameter, oC h m2/kcal

Greek Symbols λ θ γ α

heat conductivity of insulation material (kcal/h moC) incidence angle on collector plane (rad) solar azimuth angle (rad) support angle of collector (rad)

155

156

Ali M. El-Nashar τ α0 ηc αc γN

transmittance of glass tube heat trasnsfer coefficient at air side, (kcal/h m2oC) collector efficiency tilt angle of absorber plate with respect to collector (rad) solar azimuth angle measured from tilted collector plane (rad)

Subscripts a av c c1 c2 F in m out w

ambient average collector or condenser inlet to condenser outlet from condenser block F inlet measured outlet water

ACKNOWLEDGMENT The author would like to thank Dr. Darwish Al Qubaisi for his continued support and encouragement throughout this project.

REFERENCES [1] [2]

[3] [4] [5] [6]

El-Nashar, A.M., “Economics of small solar-assisted multiple-effect stack distillation plants”, Desalination. 130 (2000) 201-215 El-Nashar, A.M., “Predicting part-load performance of small MED evaporators- a simple simulation program and its experimental verification”, Desalination. 130 (2000) 217-234 El-Nashar, A.M., “The economic feasibility of small solar MED seawater desalination plants for remote arid areas”, Desalination. 134 (2001) 173-186 El-Nashar, A.M., “Validating the performance simulation program “SOLDES” using data from an operating solar desalination plant”, Desalination. 130 (2000) 235-253 ENAA and WED, Research and development cooperation on a solar energy desalination plant, Final Report, 1986. Fosselard, G. and Wangnick, K. “Comprehensive study on capital and operational expenditures for different types of seawater desalting plants (RO, MVC, ME, METVC, MSF) rated between 200 m3/day and 3000 m3/day”, Proceeding Fourth World Congress on Desalination and Water Reuse, Vol. IV, Kuwait 4-8, 1989

Multiple Effect Distillation of Seawater Water Using Solar Energy … [7]

[8]

[9]

[10]

[11]

[12]

[13] [14]

[15]

157

Garcia-Rodriguez and Gomez-Camacho, C., “Design parameter selection for a distillation system coupled to a solar parabolic trough collector”, Desalination. 122 (1999) 195-204 Garcia-Rodriguez, L. and Gomez-Camacho, C., “Conditions for economical benefirs of the use of solar energy in multi-stage flash distillation”, Desalination. 125 (1999) 133-138 Garcia-Rodriguez, L. and Gomez-Camacho, C., “Thermo-economic analysis of a solar multi-effect plant installed at the Platforma Solare de Almeria (Spain)”, Desalination. 122 (1999) 205-214 Goosen, M. F.A.; Sablani, S.S.; Shayya, W.H.; Paton, C. and Al-Hinai, H., “Thermodynamic and economic considerations in solar desalination”, Desalination. 129 (2000) 63-89 Milow, B. and Zarza, E., “Advanced MED solar desalination plants. Configurations, costs, future – seven years of experience at the Platforma Solare de Almeria (Spain)”, Desalination. 108 (1996) 51-58 Sayigh, A. et al., “Dust effect on solar flat surface devices in Kuwait”, Proceedings of the international symposium on Thermal Applications of Solar Energy, April 7-10 (1985), 95. Tsilingiris, P.T., “The analysis and performance of large-scale stand-alone solar desalination plants”, Desalination. 100 (1995) 249-255 Voivontas, D.; Misirlis, K.; Manoli, E.; Arampatzis, G. and Assimacopoulos, D., “ A tool for the design of desalination plants powered by renewable energies”, Desalination. 133 (2001) 175-198 Voivontas, D.; Yannopoulos, K.; Rados, K.; Zervos, A. and Assimacopoulos, D., “Market potential of renewable energy powered desalination systems in Greece”, Desalination. 121 (1999) 159-172

APPENDIX: PHYSICAL PROPERTIES OF SEAWATER Density (kg/m3)

ρ (T , C ) = 1002.28 − 0.18302 × T + 703.13 × C − 0.32954 × C × T Boiling point elevation (oC)

BPE (T , C ) = −0.05 +

(0.9576 + 0.8189 × 10 −2 × T + 0.1647 × 10 −4 × T 2 ) × C 0.114

Latent heat of vaporization (kcal/kg)

L(T ) = 0.5976 × 10 3 − 0.565 × T + 0.7828 × 10 −4 × T 2 − 0.2859 × 10 −5 × T 3 Specific heat at constant pressure (kcal/kgoC)

158

Ali M. El-Nashar

C p = 1.0022 − 0.6405 × 10 −4 × T − 1.14185 × C + 6.0118 × 10 −3 × T × C + 2.109 × 10 −6 × T 2 + 2.1753 × C 2 + 1.5937 × 10 − 2 × T × C 2 − 0.135 × 10 − 2 × T × C The temperature T is in oC and the salt concentration C is in kg /kg water.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 159-200

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 4

SOLID STATE ORGANIC PHOTOELECTROCHEMICAL SOLAR ENERGY CONVERSION BASED ON CONJUGATED SUBSTITUTED POLYTHIOPHENES* Teketel Yohannes† Chemistry Department, Addis Ababa University, Addis Ababa, Ethiopia

ABSTRACT The utilization of organic materials for photovoltaic devices has been investigated intensely during the last couple of decades. Earlier studies concentrated on molecules that had high optical absorption in the visible region of the electromagnetic spectrum. Recent discovery of conjugated polymers having semiconductor-like behavior has started to stir excitement because such materials are not only able to function in a similar manner to the inorganic semiconductors but also have important advantages such as: low cost, light weight, ease of fabrication and the possibility of large area coatings. Their use as photoactive electrodes is of increasing interest, as the processing possibilities of conjugated polymer materials have become more developed. Furthermore, the high absorption coefficients of these materials and the possibility of varying the band gap by molecular engineering have opened up new options for solar energy conversion. Among the conjugated conducting polymers, neutral, substituted polythiophenes exhibit interesting properties as semiconducting photoactive materials and are used for conversion of optical energy into electrical energy. Investigation of the photoelectrochemistry of conducting polymers was mainly focused on their use as protective films against photocorrosion and as photoactive electrodes in liquid junction photoelectrochemical cells (PECs). Photocorrosion and side *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Chemistry Department, Addis Ababa University, P. O. Box, 1176, Addis Ababa, Ethiopia E-mail: [email protected] or [email protected], Tel: 251-91-1408839

160

Teketel Yohannes reactions involving the electrolyte solution and the difficulty of packaging limit the working life of liquid electrolyte PECs. Solid-state PECs with the use of solid polymer electrolytes provide a means to eliminate this problem since they can easily be processed into thin films over large areas and are easier to encapsulate. The solvent-free ion conducting polymer electrolytes eliminated handling, portability, and packaging problems encountered in liquid junction photoelectrochemical cells. Basically, the photoelectrochemical properties occurring in these systems are the same as those occurring in systems based on semiconductor photoelectrodes in contact with liquid electrolytes. In this chapter an overview of the studies made on solid-state photoelectrochemical solar energy conversion devices using standard photoelectrochemical and photoelectrical characterization techniques is presented. The photoelectrochemical cells contain a thin film of semiconducting conjugated substituted polythiophenes as a light-harvesting unit, a redox couple complexed with an ion conducting polymer electrolyte, and a counter electrode.

1. INTRODUCTION An enormous amount of radiant energy is received from the sun. The solar energy falling on the Earth's surface in a fortnight is estimated to be equivalent to the energy contained in the world's supply of fossil fuels [1, 2]. About 23% of the incident solar energy is consumed in the evaporation, convection, and precipitation of water in the hydrological cycle, about 47% goes to heat the atmosphere, the land surface, and the oceans, and about 30% is reflected and scattered back into space. Approximately 0.03% of the solar input is stored as chemical energy by the photosynthetic growth of green plants, yet this process provides all our food and has generated the fossil fuels. Solar energy is one of the most promising renewable energy sources for our future energy needs when the supply of the conventional energy sources, such as coal, petroleum, and natural gas, gets depleted. Energy from the sun is not only available in plentiful supply, but also introduces no direct contamination of the environment. As a result considerable research work has been aimed at harnessing solar energy. The first experiment on conversion of solar energy to electrical energy was performed in 1839 by the French physicist Edmond Becquerel [3], who demonstrated that a photovoltage and a photocurrent are produced when a silver chloride electrode in an electrolyte solution is illuminated. The modern era of photovoltaic solar energy conversion began in 1954 at Bell Laboratories (USA) when Chapin and his co-workers reported a solar energy conversion efficiency of 6% for a single-crystal silicon cell [4]. With improved technology, the silicon cell efficiency under terrestrial sunlight had reached 14% by 1958. The first satellite powered with silicon solar cells was launched in the same year. A large amount of continued research effort has improved the efficiency to 25% at the moment [5, 6]. Although these devices are used for various optoelectronic applications, the high manufacturing cost of the cells has created a stumbling block for large-scale use. An alternative technology that might offer comparable efficiency and reduced cost is photoelectrochemical solar energy conversion. Photoelectrochemical cells (PECs) are based on the junction between a semiconductor photoelectrode and an electrolyte containing a redox couple. The semiconductor is responsible for the absorption of the incident light, while the

Solid State Organic Photoelectrochemical Solar Energy…

161

interface between the semiconductor and the electrolyte is the key factor in the subsequent chemical steps that lead to energy conversion. Less expensive polycrystalline or nanocrystalline semiconductors are used in PECs, which have low efficiency when used in photovoltaic devices. PECs combine photosynthetic and photovoltaic aspects in that they can be constructed either to produce and store chemical fuel, or to produce electricity, or both. Such devices based on dye sensitized nanocrystalline semiconductors have demonstrated solar energy conversion efficiencies exceeding 15% [7-10]. These cells are attractive because they (a) are less expensive, (b) can be used to store energy in the form of conventional fuels in addition to converting light to electrical energy, and, (c) can be fabricated with considerable ease since there are no solid-solid junctions unlike the case of photovoltaic devices. In addition, (d) the band-bending characteristics of the semiconductor can be conveniently varied by suitable choice of electrolyte, (e) they do not have problems associated with different thermal expansion of solid-solid junctions, and, (f) no anti-reflection coating is required. The potential application of PECs for storage of chemical fuels appeared for the first time in Japan in 1972 when Fujishima and Honda [11] studied the photoelectrolysis of water to oxygen and hydrogen at illuminated semiconducting titanium dioxide electrodes. They suggested that such a system would be applicable to the problem of using sunlight to photoelectrolyze water, a process that results in the conversion of sunlight to stored chemical energy. The first follow-up work did not appear until 1975 [12-15]. Since 1975 the main research efforts, which have been pursued by an increasing number of researchers, have been to study smaller band gap semiconductors (silicon, gallium arsenide, cadmium sulphide, etc.), which have acceptable solar energy conversion efficiency due to their good match to the solar spectrum. The main challenge is that these materials tend to be unstable when illuminated because of unwanted chemical reactions between the illuminated semiconductor and some components of the electrolyte solvent. On the other hand, the use of large band gap semiconductors is hampered by the fact that these materials utilize only a small portion of the solar spectrum for efficient energy conversion. Attempts to achieve good visible light response with these stable materials suffer principally from low efficiency. While inorganic semiconductor photovoltaic cells have high efficiency and inorganic semiconductor PECs have the advantage of a storable chemical fuel, both suffer from high cost. One possibility for circumventing this problem is to use inexpensive organic molecules. Earlier studies concentrated on molecules that had high optical absorption in the visible region of the electromagnetic spectrum, such as chlorophyll, cyanine, merocyanines, phthalocyanines, porphyrins, and tetracenes [16, 17]. Recent discovery of conjugated polymers having semiconductor-like behavior have started to stir excitement because they are able to function in a similar manner to the inorganic semiconductors but also have important advantages such as: low cost, light weight, ease of fabrication, and the possibility of large area coatings. Studies on the photoelectrochemistry of conducting polymers was mainly focused on their use as protective films to stabilize photocorrosion of small band gap inorganic semiconductors in liquid-junction PECs [18-23] and as photoactive electrodes in liquid junction PECs [24-43]. Their use as photoactive electrodes is of increasing interest, as the processing possibilities of conjugated polymer materials becomes more developed. Furthermore, the high absorption coefficients of these materials and the possibility of varying

162

Teketel Yohannes

the band gap by molecular engineering have opened up new options for photoelectrochemical solar energy conversion. Although conducting polymers could be used as photoelectrodes in PECs, there are only relatively few reports in the literature, probably due to their low overall conversion efficiencies. Apart from the low conversion efficiency, liquid-junction PECs suffer from problems in handling, portability, and packaging. Such problems might be eliminated by the use of polymer electrolytes, which are solvent-free ion conducting polymers. Polymer electrolytes can easily be processed into thin films over large areas, and with these materials the rectifying contacts to semiconductors, be they inorganic or organic, can easily be made. Basically, the photoelectrochemical processes occurring in these systems are the same as those occurring in systems based on semiconductor photoelectrodes in contact with liquid electrolytes. The chapter is divided into four sections. Section 1 is a general introduction. Section 2 and 3 presents the general properties of electrically conducting conjugated polymers and the ion conducting polymer electrolytes, respectively. Section 4 discusses the fundamental principles of photoelectrochemical solar energy conversion. In section 5 the experimental methods used to characterize and the most important findings of the solid-state photoelectrochemical cells based on substituted polythiophenes will be described. The last section summarizes the chapter.

2. ELECTRONICALLY CONDUCTING POLYMERS 2.1. Introduction Polymers are macromolecules produced by the union of many (102-106) small repeating units called monomers. Many polymers are strong, elastic, plastic, tough, friction-resistant, and insulating. We encounter them in our day-to-day life in a wide range of products from most consumer goods to highly specialized applications. Starting in the late 1970s new materials having high electrical conductivity became available. In 1974 a graduate student in Shirakawa's laboratory at the Tokyo Institute of Technology was trying to make a black powder of polyacetylene following a modified Natta route [44] when he accidentally prepared a shiny, free-standing polyacetylene film [45]. Looking back over his chemical recipe, the student saw that he had mistakenly added 103 times more catalyst than the instructions called for. In 1977 Shirakawa, together with MacDiarmid and Heeger [46, 47], discovered that the conductivity of that insulating polyacetylene, which had a conductivity lower than 10-5 S cm-1, could be increased up to 103 S cm-1 at room temperature by chemical oxidation with iodine or AsF5. Further improvements in polymerization methods produced polymers containing fewer defects in their structure and having conductivity as high as 1.5 x 105 S cm-1 [48]. Another significant breakthrough occurred in 1980 with the discovery that polyparaphenylene could also be doped and have conductivity levels comparable to those of polyacetylene [49]. Thereafter the number of conjugated polymers has been enlarged to include a variety of aromatic and heterocyclic polymers as well as many derivatives of these parent materials. The chemical structures of some common conjugated polymers are shown in

Solid State Organic Photoelectrochemical Solar Energy…

163

figure 1. All have an extended conjugated structure (alternating single and double bonds) along the main chain with the exception of polyaniline, which has instead an extra electron pair on the nitrogen atoms that gives it the behavior of a conjugated polymer. The main focus of the early research activity on conducting polymers was on their electrical conductivity. Neutral conjugated polymers are insulators or semiconductors. To make these polymers electrically conducting, one should introduce species, which are electron accepting (oxidizing agents) or electron donating (reducing agents). The oxidation and reduction reactions which induce high conductivity are termed, using the language of semiconductor physics, p-doping and n-doping, respectively. However, the process of doping in conducting polymers is different from that of semiconductors. In semiconductors doping involves replacing some of the atoms with atoms that have either more or less electrons while in conducting polymers the dopant molecules never replace any of the atoms of the polymer; rather they simply act as associates that accept or donate electrons. The doping process of conducting polymers can be represented by the following general scheme: For an oxidation process

Polymer + X

⇄

n+

n¯

+X

(Polymer)

and for a reduction process

Polymer + M

⇄

n¯

n+

(Polymer) + M n¯

n+

where X is the oxidizing agent and M is the reducing agent. X and M are the dopant counter anion and cation, respectively. By appropriately adjusting the doping level, conductivity anywhere between that of the undoped (insulating or semiconducting) and that of the fully doped (metallic) forms of the conjugated polymer may be obtained. The doping process may be achieved either electrochemically via application of a potential or chemically by using an oxidizing or a reducing agent. Oxidizing agents such as I2, Br2, Cl2, AsF5, FeCl3, and NOPF6, and reducing agents like Li, Na, and K, have been used. p-type doping is more common and gives higher conductivity and a better stability. Significant problems were apparent at an early stage: the materials were often unstable in air and were intractable, infusible, insoluble films or powders that, once synthesized, could not be further manipulated into forms with more ordered and controlled structures. A great deal of research has been directed to overcome these problems. Approaches include the use of a nonconducting solution processable precursor polymer that can be transformed by a simple thermal elimination reaction into an insoluble conjugated polymer [50-53]. Another is the substitution of flexible side chains onto the rigid conjugated main polymer chain to confer melt and solution processability [54-58].

164

Teketel Yohannes H N

n

N H

polypyrrole

n poly(p-phenylene)

S

n

n

polythiophene

n poly(p-phenylenevinylene)

polyaniline

n polyfluorene

Figure 1. Chemical structures of some conjugated polymers.

Because they had the important optical and electronic properties of semiconductors and metals, conducting polymers became potential materials for technological applications. Conducting polymers became so popular and they are now the main topic in international conferences and meetings on synthetic metals. Researchers from various disciplines that include chemistry, physics, electrical engineering, and material science are involved in studying them. Several reviews have been published covering many different aspects of these materials [59-66].

2.2. Synthesis of Substituted Polythiophenes In the following section the two main routes, i.e., chemical and electrochemical, for the synthesis of substituted polythiophenes will be presented in brief.

2.2.1.Chemical Synthesis Polythiophene is not melt or solution processable. It is obtained in the form of a powder or film. The delocalized electronic structure of π-conjugated polymers tends to yield relatively stiff chains and strong interchain attractions, which make them insoluble and nonprocessable. Modification of the chemical structure by addition of side groups to the thiophene ring has resulted in polymers with different degrees of stability, conductivity, solubility, and band gap. For instance, substitution onto the thiophene ring an alkyl group at the 3-position renders both solution and melt processable conjugated polymers [54-58]. Water-soluble derivatives of polythiophene have also been prepared by placing carboxylic, sulfonic, or amino acid groups attached to the alkyl chains [67-69]. Furthermore, substitution at the 3- and 4-positions by an electron donating group stabilizes the positive charges generated in the doped polymer and lower the relative band gap [70-72]. When monosubstituted thiophenes are polymerized three different couplings are possible: head-to-tail (HT), head-to-head (HH), or tail-to-tail (TT) (figure 2.). The standard synthesis of all substituted polythiophenes gives some proportion of HT linkages. The dihedral angle between two HT-coupled rings is smaller than between two HH-coupled rings. This difference in the dihedral angle affects the optical properties of the conjugated polymers. Generally, when the dihedral angle between the rings increases, the conjugation decreases and the absorption occurs at a shorter wavelength. Large substituents lead to a large dihedral angle between the rings and short conjugation along the polymer backbone. The co-planarity

Solid State Organic Photoelectrochemical Solar Energy…

165

of the rings in the main chain also affects the optical absorption. If all the thiophene rings are co-planar, the conjugation along the main chain is extended and the absorption occurs at a longer wavelength. Among the substituted thiophenes, the highest electron mobility, highest conductivity, and narrowest band gap is found in the perfect HT linked isomers [73]. R

R

R

S

S

S

S

S R

HT

HH

S R

R

TT

Figure 2. Three different coupling modes for monosubstituted thiophenes.

The most commonly used method of preparing 3-substituted monomers of thiophene is the procedure of Tamao and co-workers [74], a nickel-catalyzed Grignard coupling (figure 3 (a)). The polymer is produced by diiodinating the substituted thiophene monomer [75] and Grignard coupling of the 2,5-diiodo-3-alkylthiophenes. The scheme for the polymerization is shown in figure 3 (b). This method gives a polymer with approximately 50-60% HT coupling [76], which is random. Br

R 1. RMgBr 2. Ni(dppp)Cl2, Et2O

S

S

(a) R

R I2 S

HNO3

R 1. Mg, THF

I

S

I

2. Ni(dppp)Cl2

S

n

(b) Figure 3. (a) Monomer synthesis and (b) polymerisation of 3-alkylthiophenes by nickel-catalyzed Grignard coupling. [Ni(dppp)Cl2 = [1,3-bis(diphenylphosphino)propane]nickel(II)chloride and THF = tetrahydrofuran].

Yoshino and his co-workers [77] used another polymerization method, which involved chemical oxidation of the monomers with transition metal halides such as FeCl3, MoCl5, and RuCl3. The most widely used method now is that with FeCl3 (figure 4.). FeCl3 acts as an oxidizing agent and is reduced to FeCl2 during the polymerization reaction. The polymerization gives irregular polymers with approximately 80% HT coupling [78].

166

Teketel Yohannes

R

R FeCl3 CHCl3

S

S

n

Figure 4. Polymerization of 3-alkylthiophenes with FeCl3.

Recently, two new polymerization reactions have been developed. The one mostly used is a Grignard type of reaction (figure 5.) [79]. The regularity is high with 93-96% HT coupling. R LDA Br

S

R

R

THF

MgBr2-Et2O Li

S

BrMg

Br

Br

S

Ni(dppp)Cl2 R

S

n

Figure 5. Grignard type polymerization of 3-alkyl-2-bromothiophenes. [LDA = lithium diisopropylamide].

An alternative method of preparing 3-alkylthiophene uses zinc and a special catalyst [1,3with 3-alkyl-2,5bis(diphenylphosphino)ethane]nickel(II)chloride (Ni(dppe)Cl2), dibromothiophene (figure 6.) [80]. The HT content in this polymer is claimed to be 98.5 ± 1.5%. R

R

R

Zn Br

S

Br

THF

BrZn

S

Br

+

Br

S

ZnBr

Ni(dppe)Cl2 R

S

n

Figure 6. Polymerization of 3-alkyl-2,5-dibromothiophenes with Zn.

The stability of doped conducting polymers is important for applications. Unsubstituted, doped polythiophenes are relatively stable [81, 82], but when the thiophene rings are substituted the stability of the doped polymer severely decreases [83, 84]. Conducting polymers that have fewer, more regularly distributed side chains have better stability. The

Solid State Organic Photoelectrochemical Solar Energy…

167

number of side chains along the polymer backbone can be reduced by polymerizing bithienyls or terthienyls with only one side chain [58]. Use of a phenyl ring between the flexible alkyl side chain and the polymer backbone also increases the stability of the doped polymer [85]. The stability therefore seems dependent on the rigidity of the polymer main chain and the availability of space for the counter ion. Synthesis of one such substituted thiophene is depicted in figure 7. The 3-alkylthiophenes are selectively brominated, using Nbromosuccinimide (NBS) in N, N-dimethylformamide (DMF) [86], and can then be coupled with 2-thiopheneboronic acid according to the method described by Gronowitz [87]. (The 2thiopheneboronic acid is prepared by treating lithiated thiophene with trimethylborate [88].) The polymerization is performed using FeCl3 in chloroform (figure 7.).

R

R NBS DMF

S

Br

S

(a) R R

Pd0(PPh3)4

HO Br

+ S

B

NaHCO3, DMF

S

HO

S S

(b) Figure 7. (a) Selective bromination of 3-alkylthiophene and (b) coupling of two different substituted thiophenes. [Pd0(PPh3)4 = tetrakis(triphenylphosphine)palladium(0)].

2.2.2. Electrochemical Synthesis Electrochemical polymerization has been used to synthesize substituted polythiophenes from their monomers using both cathodic and anodic routes. The anodic route is the most convenient and widely used. The electrochemical polymerization may be carried out with a classical three-electrode electrochemical cell, consisting of a working electrode, a reference electrode, and a counter electrode, in a solution containing the monomer and a supporting electrolyte. The nature of the working electrode is critical for the preparation of these films and depends on the type of polymer to be synthesized and on the electrolytic medium. Since an oxidative or reductive process produces the films, it is important that the electrode does not oxidize or reduce concurrently with the monomer. Working electrodes such as gold, platinum, and transparent indium doped tin-oxide (ITO) coated glass can be used. The ITO-coated glass electrode is particularly suitable for spectroscopic studies. For the synthesis, potentiostatic (fixed polymerization potential), galvanostatic (fixed current), or potentiodynamic (varying potential) methods can be used. A preliminary study is necessary in order to find a solvent in which the monomer is soluble and to determine the

168

Teketel Yohannes

potential at which the polymerization may be performed. Very often, cyclic voltammetry is a good tool to determine the best polymerization conditions. The polymerization potential should not be too low, since this causes the polymerization process to be very slow and to form soluble oligomers, nor too high in order to avoid a material having lower conductivity due to overoxidation. In general, the polymerization potential must be chosen not too far (± 0.1 V) from the corresponding oxidation peak potential of the monomer. The electrodeposition potential is specific for any given electropolymerization process. As the potential needed for monomer oxidation is always higher than the charging of the existing polymer, both polymerization and doping processes may be driven by a single electrochemical operation which, starting from the monomer, first forms the polymeric chain and then induces its oxidation and deposition on the working electrode. The polymeric film coating the working electrode can be studied afterwards in a monomer free solution. The general mechanism of the electrochemical polymerization reaction is as shown in figure 8. The first step consists of the irreversible electrochemical oxidation of the neutral monomer to form delocalized radical cations. The radical cations are unstable and reactive. Since the electron transfer reaction is much faster than the diffusion of the monomer from the bulk solution, a high concentration of radicals is continuously maintained at the electrode/solution interface. The next step in the polymerization sequence involves the radical cation-radical cation coupling with dimer formation followed by a deprotonation and rearomatization. As the dimer is more easily oxidized than the monomer, under the given experimental conditions it is immediately reoxidized to the cation. Chain growth proceeds between the radical cations of the monomer and those of the continuously forming oligomers. This in turn is followed by another proton elimination and oxidation of the propagated oligomeric unit to a cation. For the polymerization to continue, radical cations must be present in the vicinity of the working electrode and therefore, the electrode potential should be kept at the oxidation potential of the monomer. In the terminology of electrochemical reaction mechanisms, the electrochemical polymerization proceeds through successive electrochemical and chemical steps according to a general E(CE)n scheme, until the oligomer becomes insoluble in the electrolytic medium and precipitates onto the electrode surface. Once deposited as conductive films, the polymers can be repeatedly cycled from the undoped to the doped forms (and vice versa) in an electrochemical cell similar to that used for the electropolymerization reactions in the absence of the monomer. The doping-undoping process involves transport of the counter ions from the electrolyte solution into the polymer matrix (between chains) during charging and out of the polymer matrix during discharging to neutralize the electronic charge on the polymer chain. Although electrochemical synthesis gives rise to cross-linked, insoluble materials that are not dense or of high quality, it is widely used because of the following advantages over chemical synthesis. (a) The polymeric material, directly grafted onto the electrode surface is produced in one step. (b) There is no need for a catalyst; therefore, the electrodeposited polymer is pure. (c) By controlling the amount of charge supplied, the thickness of the polymer film may be controlled from a few Angstroms to many micrometers. (d) By changing the nature of the counter ions in solution, the electrical and physico-chemical properties of the polymer may be changed for a particular purpose. (e) There is the possibility of performing in situ characterization of the growth process or of the polymer by electrochemical and/or spectroscopic techniques.

Solid State Organic Photoelectrochemical Solar Energy…

169

Electrochemistry is an excellent tool for conducting polymer synthesis and can also be used to study the polymerization and the doping-undoping processes. Furthermore, electrochemistry can be considered as excellent means to study processes involving the doping-undoping processes that occur in batteries, electrochromic devices, sensors, and electromechanical devices [60, 89, 90]. - eS

S

Monomer oxidation H

S

S

-2H+

.

2

.

S

S

S

H

Radical cation-radical cation coupling -e-

S S

.

S

S

S

n

S n

. S

S

+

S

2H+

S n+1

Chain propagation Figure 8. Mechanism of electropolymerization of thiophenes.

2.3. Electronic Properties Polyacetylene is the most extensively studied conducting polymer. It has served as a model for understanding the electronic and physical properties of conducting polymers. The electronic band structure of polyacetylene developed from the smallest unsaturated hydrocarbon ethene. The carbon atoms in the ethene molecule are sp2 hybridized and are connected to each other and to hydrogen atoms by σ-bonds. The remaining carbon pz orbitals interact with each other to produce two orbitals of π symmetry, one of which is bonding (π) and the other, antibonding (π*). Each of the pz orbitals contributes one electron, and in the ground state, two electrons occupy the bonding π orbital. A discrete optical transition exists between the two molecular orbitals (MOs), the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO).

170

Teketel Yohannes

As the basic ethene units are coupled together to form a larger molecule, the π-bonding will become more delocalized and more atomic orbitals must be included in the overall MO description. In the simplest conjugated hydrocarbon 1,3-butadiene, the four valence p orbitals will produce four π-type MOs. Two of these orbitals are bonding in character and are occupied by electron pairs, while the other two are antibonding and are empty. Addition of another ethene unit to 1,3-butadiene creates 1,3,5-hexatriene, a molecule with three fully occupied π-bonding orbitals and three empty π-antibonding orbitals. As this process of adding is continued, longer conjugated systems will be created until an infinite one-dimensional polyacetylene chain is formed. This chain will contain an infinite number of bonding and antibonding MOs. In this infinite chain, the orbitals will tend to cluster together into tightly packed groups. Because the fully developed polyacetylene has an extremely large number of MOs, a description of the π-bonding in this molecule can be simplified by considering groups of orbitals together as sets. Even though each of the MOs consists of a large number of orbitals that are packed tightly together into a finite energy interval, for most purposes we can ignore the energy spacing between the individual orbitals. We can therefore consider the orbitals as forming continuous bands of energy levels clustered together. The cluster of fully occupied bonding orbitals (π-band) is referred to as the valence band (VB), and the cluster of vacant antibonding orbitals (π*-band) is called the conduction band (CB) of the polymer, and the difference between the bands is called the band gap or forbidden gap. Partially filled energy bands characterize metallic conductors. On the other hand, semiconductors and insulators are characterized by the presence of completely full VB and completely empty CB separated by an energy gap. Semiconductors have reasonably low band gaps, whereas the gap for insulators is rather large. Conjugated polymers have the electronic profile of insulators or semiconductors. The existence of the alternating single and double bonds allows polyacetylene to have a number of possible structures, as shown in figure 9. As can be seen from this figure, it has two trans and two cis structures. The two trans structures are energetically degenerate (the single and double bonds can be interchanged with no cost in energy) and are thermodynamically stable. On the other hand the two cis structures are not energetically equivalent and therefore have a non-degenerate ground state: the quinoid cis structure is of higher energy than the aromatic cis structure. As a consequence only the latter is thermodynamically stable.

Trans degenerate structures

Aromatic Cis non-degenerate structures Figure 9. Trans- and cis-polyacetylene.

Quinoid

Solid State Organic Photoelectrochemical Solar Energy…

171

If all bond lengths along the polyacetylene backbone were equal, with each bond having a partial double bond character, then the polymer would have a half-filled band and behaves as a quasi-one-dimensional metal having good conductive properties. This is not the case, however. Analysis of the physics of one-dimensional metals has led to the conclusion that this type of configuration is unstable (Peierls theorem) [91], and so the one-dimensional system will undergo lattice distortion by alternating short (partial double) and long (partial single) bonds of the linear chain. The Peierls theorem states that a one-dimensional metal will be unstable against a metal-to-semiconductor transition and an energy gap will form due to the occurrence of the lattice distortion so that the material becomes either a semiconductor or an insulator [91]. Elastic energy is used during lattice distortion, which is compensated by a lowering in the electronic energy of the occupied states and the generation of a band gap. Hence in trans-polyacetylene there will be a periodic alternation of the carbon-carbon bond length along the polymer chain resulting in a stable structure of low energy. If the two degenerate structures of trans-polyacetylene are on the same chain, they will be separated by a defect in the bond alternation as shown in figure 10. The carbon atom between the conjugated segments will be sp3-hybridized and contains one unpaired electron although the overall charge remains zero. As a result, a new localized electronic state is created at the middle of the forbidden gap i.e., the unpaired electron resides in a non-bonding orbital. These conformational defects are called solitons and have a spin of 1/2 [92]. The soliton is often depicted as being localized at a certain position on the chain. However, this is not the case as noted from theoretical calculations; the defect is delocalized over 15 carbon atoms [93]. Figure 10. shows schematic band diagrams and the three classes of solitons. The soliton is positively charged with spin zero when the electron is removed and negatively charged with spin zero when electron is added. Furthermore, when charge is removed or added to the polymer chain it will be located in the midgap states.

+ Positive Soliton

Neutral Soliton

CB

CB

CB

VB

VB

VB

Charge = 0 Spin = 1/2

Negative soliton

Charge = +e Spin = 0

Charge = -e Spin = 0

Figure 10. Neutral, positive, and negative solitons and their corresponding energy band diagrams with the allowed interband (dashed arrow) and subgap (normal arrow) transitions. Dark circles indicate electrons.

All other conjugated polymers, including cis-polyacetylene, possess non-degenerate ground states and the formation of single solitons, as a result of inherent defects or doping, is energetically unfavorable. In such polymers, there are two different possible chemical structures, which are non-degenerate and are referred to as aromatic and quinoid (figure 11).

172

Teketel Yohannes

The energy of the aromatic form is lower than the energy of the quinoid form and the ground state of a non-degenerate conjugated polymer corresponds to the aromatic structure. One therefore cannot utilize the concept of soliton transport since the two regions separated by a topological defect are not energetically degenerate.

S

S S

S S n

Aromatic

S

S S

S S n

Quinoid

Figure 11. The aromatic and quinoid forms of polythiophene.

If an electron is removed, a polaron (singly charged radical cation) will be generated (figure 12) accompanied by a geometrical relaxation (lattice distortion) from the aromatic structural geometry towards the quinoid form. Further removal of an electron from the already oxidized polymer containing the polaron results in the generation of a doubly charged state termed as bipolaron (figure 12). The formation of bipolaron is also supported by calculations which show that the formation of one bipolaron is thermodynamically more stable than that of two separated polarons, despite the coulombic repulsion between the two charges confined in the same site [94]. The increase of energy due to electrostatic interaction between the two charge units is compensated by the fact that the aromatic form surrounds the bipolaron. The polarons and bipolarons are self-localized to minimize the energy of the main chain and are assumed to extend over four or five rings along the chain.

S

S S

S S n

Polaron S

S S

S S

Bipolaron Figure 12. Polaron and bipolaron in polythiophene.

n

Solid State Organic Photoelectrochemical Solar Energy…

173

The neutral non-degenerate conjugated polymer has full valence and empty conduction bands separated by a band gap. Generation of a polaron and a bipolaron creates two localized electronic levels which are energetically separated in the energy gap, a bonding polaron or bipolaron state above the valence band edge and an empty antibonding polaron or bipolaron state below the conduction band edge, unlike the single midgap state of soliton. The energy band picture of the polaron and bipolaron states is depicted in figure 13. For polarons the discrete level within the gap is singly charged while for bipolarons it is all empty (charge +2e) or all full (charge -2e). Bipolarons are spinless in contrast to polarons and neutral solitons. For polythiophenes at very high p-doping levels the transport properties become those of a metal [95]. Theoretically, this can be understood by considering that the broadening of the bipolaron states in the gap upon increasing the dopant concentration eventually leads to the merging of the lower and upper bipolaron bands with the valence band and conduction band respectively forming a bipolaronic band, thus approaching the metallic regime [96]. Optical absorption spectroscopy provides experimental evidence for the existence of the electronic states in the conjugated polymers [97]. CB

CB

VB

CB

VB Charge = +e Spin = 1/2

VB Charge = -e Spin = 1/2

Polarons

CB

VB Charge = +2e Spin = 0

Charge = -2e Spin = 0

Bipolarons

Figure 13. Energy band diagrams for polarons and bipolarons and possible allowed interband (dashed arrow) and subgap (normal arrow) transitions. Dark circles indicate electrons.

3. IONICALLY CONDUCTING POLYMERS 3.1. Introduction Electrolytes are materials that have high ionic but negligible electronic conductivity. Most electrolytes are liquids, either a molten salt or a salt dissolved in a liquid solvent. There are a number of advantages to be expected if the electrolyte could instead be a solid. The solids so far investigated have primarily been inorganic materials, most commonly β-alumina and silver salts. Recent discoveries in ion conducting polymer electrolytes have broadened the advantages of the solid electrolytes. Polymer electrolytes are complexes of metal salts with high molecular weight polymers containing electron donor atoms or groups of atoms that co-ordinate with the metal ion in the salt. To be successful as a host, a polymer should posses (a) electron donor atoms or groups of atoms (such as -O- ether, -S- sulphide, -N- amine, -P- phosphine, C=O carbonyl, and C≡N cyano) that form co-ordinate bonds with the cations, (b) low barriers to bond rotation for atoms in the main chain so that high flexibility and hence the segmental motion of the

174

Teketel Yohannes

polymer chain can take place readily, and (c) a suitable distance between co-ordinating centres which ensures adequate jumping distance for charge carriers. Wright first reported this type of polymer in 1973 [98], and Armand proposed the potential of these materials for the development of electrochemical devices in 1978 [99]. Since then, a rapid growth in research and development appears to be devoted to fundamental understanding of the ionic conduction and to their technological applications. The flexibility, processability, ease of handling, and relatively low impact on the environment that polymers inherently possess make these materials suitable for use as solid electrolytes in batteries, electrochromic devices, photoelectrochemical cells, fuel cells, sensors, and media for electrochemical reactions in solid state voltammetry. The rapid progress in this field has led to publication of a number of reviews [100-110]. Most of the early work on polymer electrolytes concentrated on the simple linear homopolymer poly(ethylene oxide), PEO, as the host for a number of salt species. PEO is readily available and is a useful solvent for a wide range of salts while exhibiting acceptable chemical and electrochemical stability. Pure PEO is a semicrystalline polymer consisting of CH2CH2O repeating units; it possesses both an amorphous and a crystalline phase at room temperature. The ionic conductivity was initially thought to occur in the regular crystalline lattice [111]; however, it is now known that such phases are electrical insulators and significant ionic transport occurs only within the amorphous phase [102, 112]. The crystalline nature of PEO is a great hindrance to the ionic transport, and at room temperature the ionic conductivity is only in the order of 10-6 S cm-1 [113]. Although it exhibits excellent ion conductivity at temperatures greater than 60oC [106], the mechanical properties of the polymer at such high temperatures are significantly poorer. This loss of mechanical stability is largely the result of melting of the crystalline phases, giving the material too small a resistance to shear stress for practical applications. The linear homopolymers poly(methylene oxide) and poly(trimethylene oxide) do not act as polymer hosts [111] despite their similarity to PEO. On the other hand, low molecular weight poly(tetramethylene glycol) dissolves salts, but the resulting electrolytes are very poor ion conductors [114, 115]. These results imply that the repeating unit in PEO provides just the right spacing between co-ordinating ether oxygens for solvation of the cations. Similarly, the homopolymer poly(propylene oxide) (PPO), having a repeat unit CH2CH(CH3)O, is less effective in dissolving salts than PEO as a result of the steric hindrance of the methyl groups [104] and has lower conductivity [116]. Linear homopolymers are prone to crystallization and formation of crystalline phases with salts, both of which reduce the conductivity of the system. As the possibility of using polymer electrolytes as components in practical devices became clear, demands were made for materials with better mechanical stability and higher ionic conductivity. A wide variety of polymer structures have been designed and produced in an attempt to optimize these properties by chemical modification of the polymer structures based on formation of linear copolymers (random or block), branched (comb-shaped) polymers, or cross-linked network polymers. Examples of some polymer host structures are given in table 1. Random copolymers with fully amorphous morphology are obtained by interspersing ethylene oxide with methylene oxide units [117-121]. The methylene oxide units break up the regular helical pattern of PEO, and in doing so suppresses crystallization. Both the host polymers and the electrolytes derived from them are amorphous; sometimes they are referred to as amorphous PEO (aPEO). Electrolytes based on aPEO generally have conductivities

Solid State Organic Photoelectrochemical Solar Energy…

175

around 10-4 S cm-1 [122, 123] at ambient temperature. Similarly, dimethyl siloxy units have been introduced between medium length PEO units to produce an amorphous polymer, dimethyl siloxy linked PEO [124]. Comb-branched polymers, with short-chain polyethers attached to a polyphosphazene [125-127] or polysiloxane backbone [128-129] have also been found to be excellent hosts with high conductivity (see examples in table 1). Unfortunately the mechanical strength of these materials is poor and it is necessary to develop more complex structures in order to optimize both mechanical and electrical properties. Polymer materials, which are mechanically stable and have reduced crystallinity have been produced by the formation of cross-linked network structures. Cross-linking can be introduced in a polymer or polymer-salt complex chemically or by exposure to intense gamma radiation. Many polymer electrolytes have been developed from network polymers [130-133], and example is given in table 1. Table 1. Chemical structures of some polymer hosts Linear Polymers Poly (ethylene oxide) (PEO)

CH2CH2O

Poly (propylene oxide) (PPO)

CH2CH(CH3)O n

Oxymethylene-linked poly (oxyethylene)

(CH2CH2O)m

n

CH2O

n

CH3

Poly(dimethyl siloxane-co-ethylene oxide) (DMS-n EO)

SiO (CH2CH2O)m CH3

n

Comb Polymers Poly[bis (methoxyethoxyethoxy) phosphazene] (MEEP)

N

(CH2CH2O)mCH3

O P

n (CH2CH2O)mCH3

O

CH3 Poly{[ω-methoxyoligo(oxyethylene)ethoxy]methyl siloxane} (PMMS-m)

SiO O

n (CH2CH2O)mCH3

CH3 Poly{[ω-methoxyoligo(oxyethylene)propyl]methyl siloxane} (PAGS-m)

SiO

n (CH2)3

O

Networks Cross-linked Siloxane

CH3

CH3 PEO

(CH2)3

Si

O

(CH2)3

Si CH3

(CH2)3

PEO

(CH2)3

PEO

O

O PEO

Si

O

Si CH3

(CH2CH2O)mCH3

176

Teketel Yohannes

A wide variety of salts based on alkali, alkaline earth, transition metals, and lanthanides can be complexed with the polymers. The conductivity of multivalent cation salts in polymer electrolytes are generally lower than those based on monovalent salts. This might be due to the greater strength of the ion-ion interactions in the former, which play a significant role in lowering the conductivity. In general, the salts that most readily complex with the polymer host contain large, singly charged anions that have low lattice energies. The most common anions used are ClO4-, CF3SO3-, (CF3SO2)2N-, (CF3SO2)3C-, BPh4-, AsF6-, PF6-, and SCN-. Salts containing monatomic anions are also soluble, provided they are large and polarizable. Hence, iodide and bromide based salts dissolve, but only few chlorides are soluble and fluorides are insoluble.

3.2. Mechanism of Ion Transport The mechanism for ion transport in polymer electrolytes is distinct from the processes occurring in conventional liquid or solid electrolytes. In conventional liquid electrolytes, ions are solvated by low molecular weight polar molecules. The ions move with their solvent sheath intact, and transport is related to the macroscopic viscosity of the electrolyte. On the other hand, solvation does not arise in inorganic solid electrolytes. The ions can be visualized as hopping between fixed sites, the positions of which do not change significantly with time. During early investigations, cations were thought to move by hopping through a rigid polymer framework along channels within the polymer helices [111]. Such models were introduced for several reasons, partly because the earlier materials (such as PEO) were in fact partially crystalline. However, various experimental techniques have shown the key role of a dynamic polymer environment for ion transport. In polymer electrolytes the solvent molecules, which are part of the polymer chain surround each cations, which are also covalently linked to each other. Transport in polymer electrolytes is considered to take place by a combination of ion motion coupled to the local motion of polymer segments. The segmental motions are thought to promote ion motion by making and breaking the co-ordination sphere of the solvated ion and by providing space into which the ion may diffuse under the influence of the electrical field. In other words, the motion of ions appears to occur by a liquid-like mechanism in which the movement through the polymer matrix is assisted by the segmental motion of the polymer backbone. The liquidlike behavior of the polymer electrolytes dominate the transport process, although an ionhopping mechanism, characteristic of solids, may also contribute to ion transport. Polymer electrolytes could be considered to have macroscopic mechanical properties similar to those of conventional solids owing to chain entanglement and cross-linking of various types, but at a microscopic level they have properties similar to those of liquids. This combination of solid-like and liquid-like properties is what makes them so fascinating scientifically and technologically. A simplified schematic representation of cation movements in polymer electrolytes is given in figure 14. An individual cation is linked to several heteroatoms at any instant of time, old co-ordination links breaking and new links being formed as the ion moves along the polymer chains in its progress through a polymer electrolyte material. The heteroatoms to which the cations are co-ordinated may well belong to more than one polymer chain. Such chains are therefore temporarily linked together by their common bonding to an individual

Solid State Organic Photoelectrochemical Solar Energy…

177

cation. Ion transport is therefore a dissociative process in which cations hop between ion coordinating sites on different polymer molecules or different parts of the same molecule. In contrast with the situation in conventional crystalline or vitreous solid electrolytes, these sites are not fixed in time or space but are created or destroyed on a continuous basis as a result of a segmental motion of the polymer chains [134]. Unlike classical solid electrolytes, both the anions and the cations are mobile in polymer electrolytes. A great deal of evidence indicates that anions are far more mobile than cations in these systems. The strong cation bonding to the polymer chain liberates anions from cation association and thus results in higher anion mobility and conductivity. Cation mobility depends on the strength of the cation-polymer interactions; if these are strong, cation transport is suppressed.

O

O M

O

M

+

O

O

O

O

O

O

O

+

O

O O

O

Figure 14. Schematic representation of a cation (M+) motion in a polymer.

4. PHOTOELECTROCHEMICAL SOLAR ENERGY CONVERSION 4.1. Introduction Photoelectrochemical solar energy conversion is based on the junction between a semiconductor and an electrolyte. A typical photoelectrochemical cell (PEC) is shown in figure 15. It consists of a semiconductor electrode, a counter electrode, and an electrolyte containing a redox couple. The PECs that convert light into electricity are termed "electrochemical photovoltaic" or "regenerative cells" and those that generate chemical fuels are "photoelectrosynthetic" or "non-regenerative cells". In regenerative PECs only one redox couple predominates in the electrolyte, so the electrochemical reaction at the semiconductor is exactly reversed at the counter electrode. Since no net chemical change occurs in the electrolyte, the light-induced current will result in electrical power in the circuit that can be used to drive a load. In non-regenerative PECs the electrolyte should contain two different redox reactions that proceed in opposite directions at both electrodes. The respective oxidation and reduction reactions occurring at the anode and cathode are different, leading to a net chemical change. Hence, the incident optical energy is converted into chemical energy by the light-induced current flow.

178

Teketel Yohannes

Figure 15. Photoelectrochemical cell.

The semiconductors that most efficiently convert solar energy to electrical energy are those with small band gaps because they capture a large fraction of the incident energy. However, they also exhibit decomposition or passivation processes which compete with the desired energy conversion reactions [138-140]. This is a serious problem for liquid-junction PECs. Therefore, for small band gap n-type semiconductors to be used for the sustained conversion of light, the photoanodic decomposition of the electrode must be suppressed. Various strategies have been pursued to chemically control the properties of semiconductor/liquid junctions. These include (a) surface modifications by electrodeposition of an electrically conducting polymer films [18, 141-150], (b) use of redox reagents that can compete kinetically with photocorrosion and photodecomposition processes [151-154], (c) application of solvent-free ion conducting polymers as electrolytes [155-165], and (d) use of non-aqueous electrolyte solutions [166-169]. On the other hand, large band gap semiconductors are stable under illumination. However, they utilize only a small portion of the solar spectrum. Good visible light response was achieved by coating a dye onto the semiconductor [170-177]. By doing so, the location of the light absorption could be transferred to the dye, and a much broader range of the solar spectrum could be used while retaining the chemical stability of the semiconductor. PEC having photoelectrochemical parameters competitive with commercial solid-state photovoltaic devices were attained with semiconductor films consisting of nanometer-sized titanium dioxide particles sensitized with a dye [173]. In the following section, an overview of the fundamental principles of photoelectrochemical solar energy conversion based on inorganic semiconductors will be presented. Knowledge of these provides a starting point for understanding the behavior of conjugated polymer based solid-state PECs. Several detailed reviews covering the science and applications of inorganic semiconductor/liquid electrolyte junction PECs are available [178184].

Solid State Organic Photoelectrochemical Solar Energy…

179

4.2. The Semiconductor/Electrolyte Interface Before examining the properties of the semiconductor/electrolyte interface, it is necessary to make a connection between the conventional way of representing the energy in a semiconductor and the electrochemical potential in an electrolyte. For semiconductors, the electrochemical potential of electrons is given by the Fermi level. Changes in the electrode potential corresponds to changes in the position of the Fermi level with respect to a reference energy, which in solids is the energy of an electron in a vacuum. For electrolytes containing a redox couple, the electrochemical potential (Eredox) of an electron is determined by the redox potential. For a reversible redox system (ox + ne- → red), this is given by the Nernst Equation [185] Eredox = Eoredox + RT/nF ln(aox/ared)

(1)

in which Eoredox is the standard electrochemical potential of the redox couple, aox and ared are the activities of the oxidized and reduced species of the redox system, and n is the number of electrons that are exchanged during the reaction. Usually, concentration instead of activity is employed given by a = f c, where f is the activity coefficient. The electrochemical potential of a redox system is given with respect to a reference, usually the normal hydrogen electrode (NHE). To treat the process occurring in PECs quantitatively, the Fermi level of the semiconductor and that of the electrolyte must be placed on a common energy scale. Using an absolute energy scale, the energy of a redox couple (EF, redox) is given by EF, redox = Eref - eEredox

(2)

in which Eredox is the redox potential versus NHE and Eref is the energy of the reference electrode versus the vacuum level. The usual value of Eref taken for the NHE is -4.5 eV, although measurements range from -4.5 to -4.7 eV [186-189]. Equation 2 can then be rewritten as EF, redox = -4.5 eV - eEredox

(3)

with respect to the vacuum level. The relationship between the energy in a semiconductor and the electrochemical potential in an electrolyte, using the solid-state scale and that of the electrochemical scale of a redox couple, are shown in figure 16. Most of the PECs are quite analogous to Schottky barrier solar cells, with the metal layer being replaced by an electrolyte containing a redox couple. When an n-type semiconductor is brought into contact with an electrolyte containing a redox species, rapid exchange of electrons between the redox species and the electrode occurs because of the difference in electrochemical potentials. If the initial Fermi level (or electrochemical potential) in a semiconductor is above the initial Fermi level in the electrolyte, then equilibration of the two Fermi levels occurs by transfer of electrons from the semiconductor to the electrolyte. The charge transfer process stops when the electrochemical potentials of both phases are equal, that is, when equilibrium has been reached. This produces a positive space charge layer in the semiconductor (also called a depletion layer since the region is depleted of majority charge

180

Teketel Yohannes

carriers) and a negative charge in the electrolyte. As a result, the conduction and valence band edges are bent upward such that a potential barrier is established against further electron transfer into the electrolyte (figure 17). Like the situation in the Schottky barrier, the magnitude of the potential barrier is determined by the bulk properties of the semiconductor, the redox levels in the electrolyte, and by the interface properties of the junction.

Figure 16. Diagram showing the relationship between the energy in a semiconductor and the electrochemical potential in an electrolyte. EF is the energy Fermi level, EC the conduction band edge, EV the valence band edge, and Eg the energy band gap.

The inverse but analogous situation occurs with p-type semiconductors having an initial Fermi level below that of the electrolyte. A negative space charge or depletion layer is formed in the semiconductor, with the valence and conduction bands bending downward to produce a potential barrier against further hole transfer into the electrolyte. A charged layer, known as the Helmholtz layer, also exists in the electrolyte adjacent to the interface with the solid electrode. This layer consists of charged ions from the electrolyte adsorbed onto the solid electrode surface. These ions are of opposite sign to the charge induced on the solid electrode. The width of the Helmholtz layer is generally on the order of a few Angstroms. The potential drop across the Helmholtz layer depends upon the specific ionic equilibrium at the surface. A very important consequence of the presence of the Helmholtz layer is that it markedly affects the band bending that develops in the semiconductor when it equilibrates with the electrolyte. Without the Helmholtz layer, the band bending would simply be expected to be equal to the difference in initial Fermi levels between the two phases (i.e., the difference between their respective work functions). Because of the high charge density and small width of the Helmholtz layer, the potential drop across the Helmholtz layer does not vary with applied electrode potential or charge transfer across the semiconductor/electrolyte interface. Instead, all of any externally applied voltage appears across the depletion layer in the semiconductor. Consequently, at a given electrolyte composition the band edges of the semiconductor at the surface are fixed with respect to the redox potential of the electrolyte, and independent of an applied voltage across the semiconductor/electrolyte interface.

Solid State Organic Photoelectrochemical Solar Energy…

181

Figure 17. Energy level diagram for the junction between an n-type semiconductor and an electrolyte in the dark. (a) Before contact. (b) After contact and equilibrium condition. φSC is the work function and χ the electron affinity.

4.3. Photoinduced Charge Transfer at the Interface To convert solar energy into electrical and/or chemical energy, current must flow across the semiconductor/electrolyte junction. When an n-type semiconductor/electrolyte junction is illuminated with light, photons having energies greater than the semiconductor band gap are absorbed. In the dark, no current flows in the cell. But when it is illuminated, electrons are freed in the valence band and move into the conduction band. The free charges can then be separated under the influence of the electric field present in the space charge region. The electric field at the space charge does not require a constant energy input from an external source; rather, it occurs spontaneously whenever two phases with different electrochemical potentials are brought into contact. Electron-hole pairs produced by absorption of photons beyond the depletion layer will separate if the minority carriers can diffuse to the depletion layer before recombining with majority carriers. If they do not disappear by recombination, either by direct coulombic interaction or by collision with other carriers in their path through the space charge layer, the minority carriers in the semiconductor are swept to the surface where they are subsequently injected into the electrolyte to drive a redox reaction. On the other hand, the majority carriers are swept towards the semiconductor bulk, where they subsequently leave the semiconductor via an ohmic contact and are then ejected at the counter electrode to drive a redox reaction opposite to that occurring at the semiconductor electrode. Since the electrons and holes travel in opposite directions, a continuous current will flow as long as the cell is illuminated and connected to an external load. The type of the redox species used is governed by the type of the semiconductor and the position of the energy bands. For n-type semiconductors, minority holes are injected to produce an oxidation reaction, while for p-type semiconductors, minority electrons are injected to produce a reduction reaction.

182

Teketel Yohannes

Oxidations for n-type semiconductors will occur from holes in the valence band if the solution species Fermi level lies above the surface valence band level and reductions for ptype semiconductors will occur from electrons in the conduction band if the solution species Fermi level lies below the surface conduction band level. The photogeneration causes the Fermi level in the semiconductor to return towards its original position, before the semiconductor/electrolyte junction was established (see figure 18). Under open circuit conditions between an illuminated semiconductor electrode and a metal counter electrode, a photovoltage is produced. The photovoltage produced between the electrodes is equal to the difference between the Fermi level in the semiconductor and the redox potential of the electrolyte. Under closed circuit conditions, the Fermi level in the system is equalized and no photovoltages exist between the two electrodes.

Figure 18. Energy level diagram of a junction between a semiconductor and an electrolyte under illumination.

5. SOLID STATE PHOTOELECTROCHEMICAL SOLAR ENERGY CONVERSION BASED ON SUBSTITUTED POLYTHIOPHENES 5.1. Introduction As discussed in Section 2, conjugated polymers exhibit a range of interesting properties as semiconducting materials and have a band gap between 1 eV and 4 eV. In certain respects, their electrical properties are similar to those of inorganic semiconductors, which allow them to be used as photoelectrodes for solar energy conversion. They have advantages over conventional inorganic semiconductors such as low cost, lightweight, ease of fabrication, processability, and the possibilities of large area coating, and material design through molecular engineering. Most of the earlier studies on solar energy conversion using conjugated polymers focused on Schottky junction photovoltaic devices, in which the polymers were sandwiched between

Solid State Organic Photoelectrochemical Solar Energy…

183

low work function metals such as aluminum or indium and high work function metals such as gold or silver. The optical absorption of conjugated polymers in the visible region of the electromagnetic spectrum was high, but the poor charge transport across the materials and the collection at the contacts kept the conversion efficiencies low. Current researches focus mainly on understanding the mechanism of operation in order to find a way to improve the performance of these devices. Understanding the mechanism of junction formation, dependence of photoactivity on molecular structure and device composition, and the sequence of events for carrier photogeneration, i.e., photon absorption, exciton generation, diffusion, recombination, and exciton dissociation, is necessary to increase the efficiency. Photoelectrochemical studies of conjugated polymers have mainly focused on their use as protective films against photocorrosion [18-23] and as photoactive electrodes in liquid junction PECs [24-43]. Liquid junction PECs suffers from handling, portability, and packaging problems. Solid-state PECs that use solvent-free ion conducting polymer electrolytes provide a means for eliminating these problems. As discussed in section 3, the ion conducting polymer electrolytes can easily be processed into thin films over large areas and are easier to encapsulate. The photoelectrochemical properties occurring in these systems are basically the same as those occurring in systems based on liquid junction PECs. Various studies on the use of polymer electrolytes in PECs have appeared in the literature in combination with inorganic semiconductors [155-165]. The properties that a solid polymer electrolyte should possess when used in PECs are: (a) it must be capable of dissolving the redox species, (b) it should have a high ionic conductivity, (c) it should be stable over a considerable period of time, (d) it should not allow unnecessary reactions at the interface between the polymer film and the semiconductor, (e) it should have a low optical absorption to prevent losses in conversion, and (f) its mechanical and electronic properties should be such that a semiconductor/electrolyte interface with the desired electronic properties can be formed. In the following section the materials, their properties, and methods used to investigate the physics and chemistry occurring in regenerative solid-state PECs based on substituted polythiophene photoactive electrode/polymer electrolyte junction will be discussed.

5.2. Materials, Device Structure, and Experimental Set-Up Among the conjugated polymers, substituted polythiophenes have attracted attention because of their chemical and environmental stability, reliable mechanical adhesion to different electrodes, solubility in common organic solvents, high conductivity, and variable band gap. They have been observed to be fairly stable under irradiation and to have high optical absorption in the visible range. Therefore, they are used as a photoelectrodes and the results obtained are presented in this section. The chemical structures of the conjugated polymers used in the construction of the solid state PECs studied in this work are shown in figure 19. Except for poly(3-methylthiophene) and poly(3,4-ethylenedioxythiophene) which were synthesized electrochemically, they were synthesized chemically and obtained in the neutral, semiconductive state. After dissolving polymers in a solvent, films were coated on ITO either by spin coating or solvent casting. The thickness of the film was controlled by the concentration of the polymer solution and/or the rotation speed. Electrochemical polymerization was done from a solution of the monomer dissolved in a solvent having an

184

Teketel Yohannes

appropriate supporting electrolyte in a three-electrode one-compartment electrochemical cell. The film was obtained in its oxidized state directly on ITO and was then reduced to the semiconductive state in a monomer-free solution. The thickness of the film was controlled by the amount of charge supplied during polymerization. The thickness of the conjugated polymer films ranged between 10 nm and 2 μm. H3 C

C6H13

S

n

C8H17

S

(I)

n

(II)

O

O

S

S

n

n

(IV)

(III)

C8H17

C8H17

S S (V)

n

S

n (VI)

Figure 19. Chemical structures of substituted polythiophenes used in these studies. (I) poly[3methylthiophene], P3MT, [190] (II) poly(3-hexylthiophene), P3HT, [191] (III) poly[3-octylthiophene], P3OT, [192] (IV) poly[3, 4-ethylenedioxythiophene], PEDOT, [193] (V) poly[3-(4-octylphenyl) thiophene], POPT, [193] (VI) poly[3-(4-octylphenyl)-2,2'-bithiophene], PTOPT [194].

The polymer electrolyte was amorphous poly(ethylene oxide), from the poly[oxymethylene-oligo(oxyethylene)] family [117-121], with a repeating unit of CH2O(CH2CH2O)9, (POMOE). It has a melting point below room temperature and a glass transition temperature of 209 K. At room temperature it will not crystallize or form crystalline polymer-salt complexes with moderate salt concentrations [121]. Amounts with the desired stoichiometry of the redox couple and the polymer were dissolved separately or together in an appropriate solvent. After thorough mixing of the two solutions, the polymer electrolyte thin films were produced by solvent casting, in which the solvent is slowly evaporated from a homogeneous solution. The thickness of the polymer electrolyte films was in the range of 0.1 μm to 6 μm. The counter electrode was platinum, deposited by vacuum evaporation onto ITO-coated glass, or oxidized poly(3,4-ethylenedioxythiophene) coated electrochemically on ITO. Platinum or oxidized poly(3,4-ethylenedioxythiophene) were required on ITO because they improve charge transfer between ITO and the iodide/triiodide redox couple: it is known that bare ITO is irreversible for the iodide/triiodide oxidation/reduction reaction [193, 195]. The basic structure of the solid-state PECs used in this study is shown in figure 20. It contains electrically conducting conjugated polymer, substituted polythiophenes, an ionically conducting polymer electrolyte, POMOE complexed with iodide/triiodide redox couple, and a counter electrode, Pt or PEDOT coated on ITO.

Solid State Organic Photoelectrochemical Solar Energy…

185

Figure 20. The basic structure of the solid-state PECs.

The general experimental set-up used for the photoelectrochemical measurements is as shown in figure 21. It contains a power supply, a lamp housing, a monochromator, a sample holder, and an output-measuring instrument. The cell is assembled and mounted inside a sample holder with light entrance window. The measurements to be taken to characterize the solar cells include current-voltage, stability towards light, spectral response, short-circuit current and open-circuit voltage dependence on incident light intensity, variation of opencircuit voltage with redox couple concentration, etc., using standard electrical, optical, photoelectrical and photoelectrochemical techniques.

Figure 21. General experimental set-up used for the photoelectrochemical measurements. (a) Power supply, (b) lamp housing, (c) monochromator, (d) sample holder, and (e) an output measuring instrument.

5.3. The Conjugated Polymer/Electrolyte Interface The property of a conjugated polymer/electrolyte interface is similar to that of an inorganic semiconductor/electrolyte interface. If a conjugated polymer is brought into contact with an electrolyte containing a redox couple with a different electrochemical potential, charge will flow across the interface and equilibrate by building up a space charge layer at the

186

Teketel Yohannes

interface. The bands bend at the interface depending on the relative positions of the electrochemical potentials of the two media. This band bending results in the formation of a potential barrier, i.e., Schottky barrier, at the interface. The energy band diagram when they are in contact with an electrolyte is depicted in figure 22.

Figure 22. Energy band diagram for a conjugated polymer/electrolyte junction.

There is a difference between light absorption in conventional semiconductor and in conjugated polymeric semiconductors. Conventional semiconductors have a fairly rigid crystalline lattice. Polymeric semiconductors have a soft one-dimensional lattice and have inter- and intra-molecular interactions, local structural disorders, amorphous and crystalline regions, and chemical impurities. If we break one bond, the lattice will be severely distorted because in a one-dimensional lattice the atoms are kept in position by the two neighboring bonds, whereas in a three-dimensional lattice they are bound to the neighbors in all three directions. In conventional inorganic semiconductor, absorption of a photon produces electron and hole charge carriers, but in conjugated polymer-based PECs light absorption results in the generation of excitons. These excitons diffuse as uncharged particles until they dissociate into free charge carriers for the photocurrent generation. If they dissociate at the conjugated polymer/electrolyte interface, the majority carriers (holes) move into the conjugated polymer and the corresponding electrons travel towards the electrolyte. The Fermi levels shift due to this charging effect and a photovoltage is produced.

5.4. Steady State and Transient Properties The parameters that are used to describe solar cells are the short circuit current (Isc) and the open circuit voltage (Voc). The photovoltage developed during illumination is measured with a voltmeter having a very high internal resistance so that negligible current can flow through the cell. The photocurrent is measured using a very small load resistance. These parameters determine the efficiency and are the key parameters in experimental investigation of solar cells. Steady state and transient measurements of Isc and Voc, established during long and short period of irradiation, are used to characterize the stability of the PECs towards illumination. As an example, the transient photocurrent in the study of P3HT-based solid-state PECs is

Solid State Organic Photoelectrochemical Solar Energy…

187

shown in figure 23. [191]. The transient photocurrent of the PEC is characterized by a rise to a steady state value when the light is switched on and decay at approximately the same rate when the light is switched off. The photocurrent obtained with a longer period of irradiation indicated that the stability of the P3HT towards light illumination is fairly good. Similar results were obtained for the other substituted polythiophenes based solid state PECs studied. Table 2 summarizes Isc and Voc of the various solid-state PECs studied.

Photocurrent [µ A/cm2]

0.6 on

off

on

off

200

300

400

500

0.4

0.2

0.0

Time [s] Figure 23. Transient photocurrent of P3HT-based solid-state PEC [191].

Table 2. Isc and Voc of a solid state PECs illuminated with light intensity of 100 mW/cm2 from the front side PEC

Isc (μA/cm2)

Voc (mV)

Ref.

ITO/P3MT/POMOE, I3/I¯/Pt/ITO

0.35

140

[190]

ITO/P3HT/POMOE, I3/I¯/PEDOT/ITO

0.47

130

[191]

ITO/P3OT/POMOE, I3/I¯/Pt/ITO

0.04

250

[192]

ITO/POPT/POMOE, I3/I¯/PEDOT/ITO

0.20

166

[193]

ITO/PTOPT/POMOE, I3/I¯/PEDOT/ITO

0.40

240

[194]

5.5. Current-Voltage Characteristics If an external field is applied with forward bias (negative voltage at the counter electrode relative to the working electrode), it acts to diminish the effects of the internal barrier field. Carriers can acquire sufficient energy to cross the barrier, and at high enough external voltages a large current will flow. On the other hand, if the bias is reversed, the external field enhances the barrier potential and only a small current flows. Thus, the junction acts as a rectifier since the current flowing for a given positive external voltage is quite different from

188

Teketel Yohannes

the current flowing at the same negative voltage. Such a current-voltage characteristic is described mathematically as [196] I = Io [exp(qV/nkT) - 1]

(4)

where Io is the saturation current, q the electron charge, k the Boltzmann constant, T the absolute temperature, V the applied voltage, and n the ideality factor. The rectification characteristic is typical of diodes, and equations that have the form of Equation 4 are generally called diode equations. The conjugated polymer/electrolyte interface also obeys the diode equation. As example, the current-voltage characteristic of a junction between P3HT and a polymer electrolyte in the dark is shown in figure 24. [191]. The forward current corresponds to a positive bias where the barrier height is lowered, whereas in reverse polarization the increased barrier prevents the passage of a current. 3

Current [mA/cm2 ]

2

1

0

-1

-2 -0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

Voltage [V]

Figure 24. Current versus voltage characteristics in dark (open circles) and under white light illumination (solid squares) from the front side of the P3HT based solid-state PEC with a light intensity of 100 mW/cm2 [191].

Under illumination, absorption of photons creates excitons, and later both the majority and the minority carriers. The concentration of photogenerated majority carriers are usually small. This implies that illumination does not significantly perturb the majority carrier behavior either in the semiconductor or at the semiconductor/electrolyte interface. Because the majority carrier concentrations are essentially unchanged, the majority charge flow is also unchanged. Majority carriers should thus exhibit the I-V characteristic that is well described by the diode equation, regardless of whether it is in the dark or under illumination. The current under illumination can generally be described by adding the current from photogenerated carriers to the dark current. From Equation 4 the I-V characteristic under illumination is given by

Solid State Organic Photoelectrochemical Solar Energy… I = Iph - Io [exp(qV/nkT) - 1]

189 (5)

where Iph is the component of the current that has been generated by illumination. The I-V curve obtained under illumination for a neutral P3HT-based solid-state PEC is depicted in figure 25. [191]. A cathodic photocurrent was observed at cathodic potentials, indicating that P3HT behaves as p-type semiconductor. The cathodic photocurrent is due to photoinduced minority carrier injection from the P3HT electrode into the electrolyte, where they react with the electron acceptor. The majority carriers go to the back contact and react with the electron donor at the counter electrode. The process is cyclic, and therefore there is no net chemical reaction and the PEC converts light to electricity in a regenerative mode. The possible operation mechanism is depicted in figure 25.

Figure 25. Operation principle of the solid-state PECs.

One of the new results from these studies is that the redox species may be diffusing inside the polymer photoelectrode. If this is the case, which is probable as conjugated polymers are mixed electronic-ionic conductors, novel possibilities arise. We do not necessarily have to move the excitons to the electrode/electrolyte boundary; we can position the redox species in many, many different geometries inside the polymer solid; we can collect excitons deep inside the polymer film. The utilization of these phenomena to bring about higher photocurrents is non-trivial, but some important avenues are already visible. First, we may use redox species that are effective acceptor (or donor) molecules. There are reports of very fast electron transfer from polymers onto fullerenes, which could be an inspiring example [197]. There is further the possibility of dynamic geometry with diffusing redox species, which could help in bringing about an effective charge separation. The possibility of movement by diffusion might also be of some importance, when it comes to preventing back electron transfer. And thirdly, the possibility of making mixed electronic-ionic conductors with enhanced ionic transport, such as in the main chain conjugated polymers decorated with side chains of oligo(ethylene oxide), is showing promise [198]. The thin films used in these devices will allow an effective absorption of light, but the photocurrents will be limited by the

190

Teketel Yohannes

ionic transport. It has been shown that even with present day polymer electrolytes, the ionic transport can be sufficient for photovoltaic purposes [192].

5.6. Spectral Response Most substituted polythiophenes are coloured and absorb in the visible range of the electromagnetic spectrum. The transition between the π- and π *-orbitals can be seen in the UV-VIS optical absorption spectra. In general, the absorbance (A) and transmittance (T) of a material can be expressed by Beer's law A = ln(Ii/I) = εcl

and

T = I/Ii = exp(-εcl)

(6)

where I is the transmitted light intensity, Ii is the incident light intensity, ε is the molar extinction coefficient, l is the optical path length, and c is the concentration of the absorbing material. For solids, the concentration of the absorbing material is constant and Equation 6 can be rewritten as A = ln(Ii/I) = αl

and

T = I/Ii = exp(-αl)

(7)

where α = εc is the absorption coefficient of the material. A and α are functions of the wavelength at which the light absorption is measured. This latter equation is the form that is usually used for expressing light absorption in thin films of conjugated polymers. The photocurrent collected at different wavelength relative to the number of photons incident on the surface at that wavelength determines the spectral response of the device (sometimes known as the external quantum efficiency or collection efficiency at each wavelength). Light of different wavelengths is absorbed at different depths in the conjugated polymer film. The ability of a solar cell to generate photocurrent at a given wavelength of the incident light is measured by the incident monochromatic photon to current conversion efficiency (IPCE), defined as the number of electrons generated per number of incident photons. It can be obtained from the photocurrents by means of the following equation [175] IPCE% = (1240 Isc)/(λ Ii)

(8)

where Isc is the short circuit current (μA cm-2), λ the excitation wavelength (nm) and Ii the photon flux (W m-2). The quantum efficiencies obtained for the solid-state PECs studied was less than 1%. In general the quantum efficiency of the solid-state PECs when illuminated from the front side is greater than for backside illumination. This is due to the optical filtering effect of the conjugated polymer films. When light is illuminated from the backside only a small fraction of the excitons produced by light absorption reach the interface to dissociate into carriers. In addition, the presence of a high density of traps in the film reduces the number of carriers for the photocurrent generation. The greater the distance from the surface, the smaller is the probability for an exciton to reach the interface and dissociate into carriers. In conventional inorganic semiconductor solar cells, the absorption with a photon energy greater than the band gap leads to a direct generation of an electron-hole pair that is separated by the built-in electric field; the charges are transported to opposite electrodes and produce a

Solid State Organic Photoelectrochemical Solar Energy…

191

photocurrent. However, in molecular semiconductors the absorption of a photon creates an exciton rather than free charge carriers. To generate photocurrent, these excitons must dissociate into free carriers either in the bulk or at the interface. Based on experimental results obtained from many molecular semiconductor solar cells [16, 199-203], only those excitons that reach the active junction produced free charge carriers. The excitons reach this interface by diffusion, then dissociate into carriers that are transported to opposite electrodes for collection. Several researchers working with different molecular semiconductors [16, 199203] have noticed a difference in the action spectra depending on which side of the cell is illuminated: front (rectifying contact) or back (ohmic contact). For front side illumination, the strongly absorbed light creates excited states in the barrier region. Under these conditions, the absorption spectrum is usually well matched to the action spectrum. For backside illumination the organic material itself acts as an optical filter for the strongly absorbed light in the fieldfree bulk region, and only weakly absorbed photons penetrate into the depletion region and make a major contribution to the photocurrent. The spectral response then does not match the optical absorption spectra. A comparison of the optical absorption spectrum and the spectral photoresponse can be used to identify the active junction responsible for the photoelectrochemical phenomena. If illumination through the front side of the PEC produces a spectral response, which corresponds to the absorption spectrum of the conjugated polymer, then the conjugated polymer/electrolyte junction is responsible; if illumination from the backside produces a matching spectrum, then it is the conjugated polymer/ITO junction, which is active. The photocurrent action spectra plotted in terms of IPCE versus wavelength for front side and backside illumination of the solid-state PEC with P3HT photoactive electrode, together with the optical absorption of the P3HT film coated on ITO, are shown in figure 26. [191]. Normalization was done to the peak values in order to facilitate the comparison between the photocurrent action spectra and the optical absorption spectrum.

Normalized to peak value

1.0

0.8

0.6

0.4

0.2

0.0 300

400

500

600

700

800

Wavelength [nm]

Figure 26. Optical absorption spectrum of P3HT coated on ITO (open triangles) and normalized photocurrent action spectrum of P3HT based solid-state PEC from the front side illumination (solid squares) and from the backside illumination (open circles) [191].

192

Teketel Yohannes

The optical absorption spectrum and the action spectrum for front side illumination match. This indicates that the active junction responsible for the photoelectrochemical properties of the PEC is that between the conjugated polymer and the polymer electrolyte. Since the generation of charge carriers occurs close to the active interface rather than in the bulk of the polymer film, the kinetics of the interfacial electron transfer processes may determine the photoeffects. This hypothesis can thus explain why excitons, which are produced predominantly in the interior, yield a weaker photocurrent than those that are produced at the interface.

5.7. Dependence of Photovoltage on Redox Couple Concentration The formation of junction between the conjugated polymer and the electrolyte gives some means of varying the redox potentials of the electrons and holes in the conjugated polymer relative to the redox systems in solution. By changing the solution redox potential one can manipulate the properties of the semiconductor/electrolyte interface. There are two different methods of changing the solution redox potential: varying the concentration ratio of the reduced and oxidized species or holding the redox concentrations constant while varying the molecular species. One can therefore vary the ratio of the redox couple concentration or the molecular species to vary the photovoltage of the PEC. For p-type semiconductors, negative redox potentials produce highly rectifying contacts, while positive redox potentials produce poorly rectifying contacts. Therefore, high photovoltages are generated with comparatively large negative redox potential for p-type semiconductors. For a given photoactive electrode, the magnitude of the photovoltage is dependent on the concentration of the redox couple. Table 3 shows the effect of redox couple concentration on the open circuit voltage for a P3OT-based solid-state PEC [192]. As the concentration ratio between the oxidized and reduced form increases, the open circuit voltage decreases. On the other hand, as the concentration ratio increases, there will be a point at which the PEC does not produce a photovoltage at all. Chemical doping of the conjugated polymer, which changes the state of the conjugated polymer from semiconductive to the conductive, causes this. The effect of increasing the concentration of the redox couples causing doping of the conjugated polymer can be seen from the optical absorption of a P3OT-based solid-state PEC (curve a figure 27.) [192]. Two new optical transitions below the band gap transition appeared, together with a decrease of the band gap transition, indicating doping. Table 3. Variation of open circuit voltage with concentration ratio of I3- to I- [192] [I3-]/[I-] 10 1.0 0.1 0.01

Voc (mV) 223.0 250.0 523.3

Isc (nA) 53.3 31.0 32.3

Solid State Organic Photoelectrochemical Solar Energy…

193

Figure 27. Optical absorption spectrum for P3OT-based solid-state PEC at different concentration ratios of I3- to I-: (a) 10, (b) 0.01, (c) 0.1, and (d) 1 [192].

5.8. Dependence of Photocurrent and Photovoltage on Light Intensity For organic and some inorganic solar cells, the short circuit current increases with increasing light intensity (Ii) and is proportional to Iiα [204]. Thus, a plot of log Isc versus log Ii yields a straight line whose slope is characteristic of the photoactive material. The Isc dependence on the incident light intensity for all solid-state PECs gave a linear plot with a slope less than unity. Such sublinear dependence implies bimolecular recombination of excitons and/or the presence of a high density of traps in the film [204-207]. The trapping could arise from structural defects on the materials that promote charge recombination and reduce the average lifetime of the charge carriers. For Schottky junction solar cells under open circuit, no net current will flow through the junction and Equation 5 can be rearranged to yield the following relationship: Voc = nkT/q ln[(Iph/Io) + 1] = nkT/q ln(Iph/Io)

for Iph >> Io

(9)

As can be seen from Equation 9, Voc increases logarithmically with the light intensity because Iph is linearly proportional to the incident light intensity. All of the solid-state PECs studied show an open circuit voltage which is logarithmically dependent on the light intensity, in agreement with the projected behavior of Schottky barrier solar cells.

194

Teketel Yohannes

6. CONCLUSION Organic solar cells may provide a unique alternative to inorganic photovoltaics. They function in a similar manner to the inorganic semiconductors but also have important advantages such as: high absorption coefficient, low cost, light weight, ease of fabrication and the possibility of large area coatings. One such type of cells is solid-state PECs. In this chapter it is shown that solid-state PECs convert solar energy to electrical energy. The photoactive electrodes were electron conducting, neutral, substituted polythiophenes and the ion conducting polymer used as an electrolyte, was amorphous poly(ethylene oxide) complexed with the iodide/triiodide redox couple. In order to understand the physics and chemistry of the cell studies such as I-V characteristics, steady and transient properties, spectral response, dependence of photocurrent and photovoltage on light intensities, and dependence of photovoltage on the redox species concentration ratios were made.

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

M.D. Archer, J. Appl. Electrochem., 5 (1975) 17. M.K. Hubbard, Scientific American, 224 (1971) 61. E. Becquerel, C.R. Acad. Sci. Paris, 9 (1839) 561. D.M. Chapin, C.S. Fuller, and G.L. Pearson, J. Appl. Phys., 25 (1954) 676. P.E. Gruenbaum, J.Y. Gan, and R.M. Swanson, Appl. Phys. Lett., 58 (1991) 945. J.M. Li, M. Chong, J.C. Zhu, Y.J. Li, J.D. Xu, P.D. Wang, Z.Q. Shang, Z.K. Yang, R.H. Zhu, and X. Cao, Appl Phys. Lett., 60 (1992) 2240. E. Aharon-Shalom and A. Heller, J. Electrochem. Soc., 129 (1982) 2865. J.F. Gibbons, G.W. Cogan, C.M. Gronet, and N.S. Lewis, Appl. Phys. Lett., 45 (1984) 1095. B.J. Tufts, I.L. Abrahams, P.G. Santangelo, G.N. Ryba, L.G. Casagrande, and N.S. Lewis, Nature, 326 (1987) 861. S. Licht and D. Peramunage, Nature, 345 (1990) 330. A. Fujishima and K. Honda, Nature, 238 (1972) 37. R. Gomer, Electrochim. Acta, 20 (1975) 13. W.J. Albery and M.D. Archer, Electrochim. Acta, 21 (1976) 1155. T. Sakata, Y. Suda, J. Tanaka, and H. Tsubomura, J. Phys. Chem., 81 (1977) 537. A. Albery and M.D. Archer, Nature, 270 (1977) 399. G.A. Chamberlain, Solar Cells, 8 (1983) 47. D. Wöhrle and D. Meissner, Adv. Mater., 3 (1991) 129. R. Noufi, D. Tench, and L.F. Warren, J. Electrochem. Soc., 128 (1981) 2596. A.J. Frank, Mol. Cryst. Liq. Cryst., 83 (1982) 341. A.J. Frank and K. Honda, J. Phys. Chem., 86 (1982) 1933. R. Noufi and A.J. Nozik, J. Electrochem. Soc., 129 (1982) 2261. F. Garnier, M. Lemaire, G. Horowitz, and J. Roncali, J. Ber. Bunsen-Ges. Phys. Chem., 92 (1988) 1261. A.J. Frank, A.J. Nelson, and S. Glenis, J. Phys. Chem., 93 (1989) 3818.

Solid State Organic Photoelectrochemical Solar Energy…

195

[24] S.N. Chen, A.J. Heeger, Z. Kiss, A.G. MacDiarmid, S.C. Gau, and D.L. Peebles, Appl. Phys. Lett., 36 (1980) 96. [25] H. Shirakawa, S. Ikeda, M. Aizawa, I. Yoshitake, and S. Suzuki, Synth. Met., 4 (1981) 43. [26] T. Yamase, H. Harada, T. Ikawa, S. Ikeda, and H. Shirakawa, Bull. Chem. Soc. Jpn., 54 (1981) 2817. [27] Y. Nosaka, K. Norimatsu, and H. Miyama, Chem. Lett. Jpn., (1983) 801. [28] T. Inoue and T. Yamase, Bull. Chem. Soc. Jpn., 56 (1983) 985. [29] S. Glenis, G. Tourillon, and F. Garnier, Thin Solid Films, 122 (1984) 9. [30] M. Kaneko, K. Okuzumi, and A. Yamada, J. Electroanal. Chem., 183 (1985) 407. [31] T. Kenmochi, E. Tsuchida, M. Kaneko, and A. Yamada, Electrochim. Acta, 30 (1985) 1405. [32] M. Kaneko and H. Nakamura, J. Chem. Soc. Chem. Commun., (1985) 346. [33] E.M. Genies and M. Lapkowski, Synth. Met., 24 (1988) 69. [34] J.A. De Saja and K. Tanaka, Phys. Stat. Sol., 108 (1988) K109. [35] E.M. Genies, P. Hany, and M. Lapkowski, Synth. Met., 25 (1988) 29. [36] P.K. Shen and Z.Q. Tian, Electrochim. Acta, 34 (1989) 1611. [37] D. Yaohua and M. Shaolin, Electrochim. Acta, 36 (1991) 2015. [38] Z. Li and S. Dong, Electrochim. Acta, 37 (1992) 1003. [39] M. Shaolin and X. Caifen, Synth. Met., 59 (1993) 243. [40] Y. Yang and Z.G. Lin, Synth. Met., 64 (1994) 43. [41] F.L.C. Miquelino, M.A.D. Paoli, and E.M. Genies, Synth. Met., 68 (1994) 91. [42] O.A. Semenikhin, E.V. Ovsyannikova, N.M. Alpatova, Z.A. Rotenberg, and V.E. Kazarinov, Russian J. Electrochem., 30 (1994) 666. [43] J. Lukkari, M. Alanko, V. Pitkanen, K. Kleemola, and J. Kankare, J. Phys. Chem., 98 (1994) 8525. [44] G. Natta, G. Mazzanati, and P. Corrandini, Att. Acad. Naz. Lincei. A. Sci. Fis. Mat. Nat. Rend., 25 (1958) 2. [45] T. Ito, H. Shirakawa, and S. Ikeda, J. Polym. Sci. Polym. Chem. Ed., 12 (1974) 11. [46] H. Shirakawa, E.J. Louis, A.G. MacDiarmid, C.K. Chiang, and A.J. Heeger, J. Chem. Soc. Chem. Commun., (1977) 578. [47] C.K. Chiang, C.R. Fincher, Y.W. Park, A.J. Heeger, H. Shirakawa, E.J. Louis, and A.G. MacDiarmid, Phys. Rev. Lett., 39 (1977) 1098. [48] N. Basescu, Z.X. Liu, D. Moses, A.J. Heeger, H. Naarman, and N. Theophilou, Nature, 327 (1987) 403. [49] L.W. Shacklette, H. Eckhardt, R.R. Chance, and R.H. Baughman, J. Chem. Soc. Chem. Commun., (1980) 854. [50] J.H. Edwards and W.J. Feast, Polym. Commun., 21 (1980) 595. [51] J.H. Edwards, W.J. Feast, and D.C. Bott, Polymer, 25 (1984) 395. [52] I. Murase, T. Ohnishi, T. Noguchi, and M. Hirooka, Polym. Commun., 25 (1984) 327. [53] F.E. Karasz, J.D. Capistran, D.R. Gagnon, and R.W. Lenz, Mol. Cryst. Liq. Cryst., 118 (1985) 327. [54] R.L. Elsenbaumer, K.Y. Jen, and R. Oboodi, Synth. Met., 15 (1986) 169. [55] M. Sato, S. Tanaka, and K. Kaeriyama, J. Chem. Soc. Chem. Commun., (1986) 873. [56] K. Yatsumi, S. Nakajima, M. Fuji, and R. Sugimoto, Polym. Commun., 28 (1987) 309.

196

Teketel Yohannes

[57] S. Hotta, S.D.D.V. Rughooputh, A.J. Heeger, and F. Wudl, Macromolecules, 20 (1987) 212. [58] M.R. Andersson, Q. Pei, T. Hjertberg, O. Inganäs, O. Wennerström, and J.-E. Österholm, Synth. Met., 55-57 (1993) 1227. [59] T.A. Skotheim (Ed), Handbook of Conducting Polymers, Vol. 1 and 2, Marcel Dekker, New York (1986). [60] B. Scrosati, Prog. Solid State Chem., 18 (1988) 1. [61] A.O. Patil, A.J. Heeger, and F. Wudl, Chem. Rev., 88 (1988) 183. [62] R.H. Friend, J. Mol. Electron., 4 (1988) 37. [63] M.G. Kanatzidis, Chem. Eng. News, 68 (1990) 36. [64] G.P. Evans, In Advances in Electrochemical Science and Engineering, Vol. 1, H. Gerisher and C.W. Tobias (Eds), VCH, Weinheim (1990) pp. 1-74. [65] H. Stubb, E. Punkka, and J. Paloheimo, Mater. Sci. Eng., 10 (1993) 85. [66] A.K. Bakhshi, Bull. Mater. Sci., 18 (1995) 469. [67] A.O. Patil, Y. Ikenoue, N. Basescu, N. Colaneri, J. Chen, F. Wudl, and A.J. Heeger, Synth. Met., 20 (1987) 151. [68] A.O. Patil, Y. Ikenoue, F. Wudl, and A.J. Heeger, J. Am. Chem. Soc., 109 (1987) 1858. [69] M.R. Andersson, P.O. Ekeblad, T. Hjertberg, O. Wennerström, and O. Inganäs, Polym. Commun., 32 (1991) 546. [70] G. Heywang and F. Jonas, Adv. Mater., 4 (1992) 116. [71] Q. Pei, G. Zuccarello, M. Ahlskog, and O. Inganäs, Polymer, 35 (1994) 1347. [72] M. Dietrich, J. Heinze, G. Heywang, and F. Jonas, J. Electroanal. Chem., 369 (1994) 87. [73] R.D. McCullough, S. Tristram-Magle, S.P. Williams, R.D. Lowe, and M. Jayaraman, J. Am. Chem. Soc., 115 (1993) 4910. [74] K. Tamao, S. Losama, I. Nakajima, M. Kumada, A. Minato, and K. Suzuki, Tetrahedron, 38 (1982) 3347. [75] J. Barker, P.R. Huddleston, and M.L. Wood, Syn. Commun., 5 (1975) 59. [76] H. Mao, B. Xu, and S. Holdcroft, Macromolecules, 26 (1993) 1163. [77] R. Sugimoto, S. Takeda, H.B. Gu, and K. Yoshino, Chemistry Express, 1 (1986) 635. [78] M. Leclerc, F.M. Diaz, and G. Wegner, Makromol. Chem., 190 (1989) 3105. [79] R.D. McCullough, R.D. Lowe, M. Jayaraman, and D.L. Anderson, J. Org. Chem., 58 (1993) 904. [80] T.-A. Chen and R.D. Rieke, J. Am. Chem. Soc., 117 (1995) 233. [81] G. Tourillon and F. Garnier, J. Electrochem. Soc., 130 (1983) 2042. [82] J.-E. Österholm, P. Passinemi, H. Isotalo, and H. Stubb, Synth. Met., 18 (1987) 213. [83] M.T. Loponen, T. Taka, J. Laakso, K. Väkiparta, K. Suuronen, P. Valkeinen, and J.-E. Österholm, Synth. Met., 41 (1991) 479. [84] Y. Wang and M.F. Rubner, Synth. Met., 41 (1991) 1103. [85] M.R. Andersson, D. Selse, H. Järvinen, T. Hjertberg, O. Inganäs, O. Wennerström, and J.E. Österholm, Macromolecules, 27 (1994) 6503. [86] R.M. Mitchell, Y.-H. Lai, and R.V. Williams, J. Org. Chem., 44 (1979) 4733. [87] S. Gronowitz and D. Peters, Heterocycles, 30 (1990) 645. [88] S.-O. Lawesson, Arkiv Kemi, 11 (1957) 387. [89] J. Heinze, Topics in Current Chemistry, 152 (1990) 1. [90] J. Simonet, J. Rault-Berthelot, Prog. Solid State Chem., 21 (1991) 1.

Solid State Organic Photoelectrochemical Solar Energy…

197

[91] R. Peierls, Quantum Theory of Solids, Oxford University Press, London (1955). [92] A.J. Heeger, S. Kivelson, J.R. Schrieffer, and W.-P. Su, Rev. Mod. Phys., 60 (1988) 781. [93] J.L. Bredas, R.R. Chance, and R. Silbey, Mol. Crys. Liq. Cryst., 77 (1981) 319. [94] J.L. Bredas and G.B. Street, Acc. Chem. Res., 18 (1985) 309. [95] T.C. Chung, J.H. Kaufman, A.J. Heeger, and F. Wudl, Phys. Rev. B, 30 (1984) 702. [96] J.L. Bredas, B. Themans, J.G. Fripiat, J.M. Andre, and R.R. Chance, Phys. Rev. B, 29 (1984) 6761. [97] M. Mastragostino, C. Arbizzani, A. Bongini, G. Barbarella, and M. Zambianchi, Electrochim. Acta, 38 (1993) 135. [98] D.E. Fenton, J.M. Parker, and P.V. Wright, Polymer, 14 (1973) 589. [99] M.B. Armand, J.M. Chabagno, and M. Duclot, Second International Conference on Solid Electrolytes, St Andrews (1978). [100] M.B. Armand, Solid State Ionics, 9 and 10 (1983) 745. [101] M.B. Armand, Annu. Rev. Mater. Sci., 4 (1986) 245. [102] C.A. Vincent, Prog. Solid State Chem., 17 (1987) 145. [103] M.A. Ratner and D.F. Shriver, Chem. Rev., 88 (1988) 109. [104] J.M.G. Cowie and S.H. Cree, Ann. Rev. Phys. Chem., 40 (1989) 85. [105] M.B. Armand, Adv. Mater., 2 (1990) 278. [106] F.M. Gray, Solid Polymer Electrolytes, Fundamentals and Technological Applications, VCH, New York (1991). [107] B. Scrosati, Mater. Sci. Eng., B12 (1992) 369. [108] P.G. Bruce and C.A. Vincent, J. Chem. Soc. Faraday Transactions, 89 (1993) 3187. [109] P.G. Bruce, Solid State Electrochemistry, Cambridge University Press, Cambridge (1993) pp. 95-160. [110] B. Scrosati, Applications of Electroactive Polymers, Chapman and Hall, New York (1993) pp.1-28, 75-112. [111] M.B. Armand, J.M. Chabagno, and M.J. Duclot, In Fast Ion Transport in Solids, P.Vashishta, J.M. Mundy, and G.K. Shenoy (Eds), Elsevier, New York (1979) p. 131. [112] C. Berthier, W. Gorecki, M. Minier, M.B. Armand, J. Chabagno, and P. Rigaud, Solid State Ionics, 11 (1983) 91. [113] M.B. Armand, In Polymer Electrolyte Reviews-1, J.R. MacCallum and C.A. Vincent (Eds), Elsevier Applied Science, London (1987) p. 1. [114] M.S. Mendolia and G.C. Farrington, Electrochim. Acta, 37 (1992) 1695. [115] G.G. Cameron, M.K. Ingram, and K. Sarmouk, Eur. Polym. J., 26 (1990) 1097. [116] M. Watanabe and N. Ogata, In Polymer Electrolyte Reviews-1, J.R. MacCallum and C.A. Vincent (Eds), Elsevier Applied Science, London (1987) p. 39. [117] J.R. Craven, R.H. Mobbs, C. Booth, and J.R.M. Giles, Makromol. Chem. Rapid Commun., 7 (1986) 81. [118] J.R. Craven, C.V. Nicholas, R. Webster, D.J. Wilson, R.H. Mobbs, G.A. Morris, F. Heatley, C. Booth, and J.R.M. Giles, Br. Polym. J., 19 (1987) 509. [119] C.V. Nicholas, D.J. Wilson, C. Booth, and J.R.M. Giles, Br. Polym. J., 20 (1988) 289. [120] E. Linden and J.R. Owen, Br. Polym. J., 20 (1988) 237. [121] M. Nekoomanesh, H.S. Nagae, C. Booth, and J.R. Owen, J. Electrochem. Soc., 139 (1992) 3046. [122] F.M. Gray, Solid State Ionics, 40/41 (1990) 637.

198

Teketel Yohannes

[123] J.P. Lemmon and M.M. Lerner, Macromolecules, 25 (1992) 2907. [124] K. Nagoka, H. Naruse, I. Shinohara, and M. Watanabe, J. Polym. Sci. Polym. Lett., 22 (1984) 659. [125] P.M. Blonsky, D.F. Shriver, F. Austin, and H.R. Allcock, J. Am. Chem. Soc., 106 (1984) 6854. [126] P.M. Blonsky, D.F. Shriver, P. Austin, and H.R. Allcock, Solid State Ionics, 18/19 (1986) 258. [127] S.G. Greenbaum, K.J. Adamic, Y.S. Pak, M.C. Wintersgill, and J.J. Fontanella, Solid State Ionics, 28-30 (1988) 258. [128] D.W. Xia, K. Soltz, and J. Smid, Solid State Ionics, 14 (1984) 221. [129] P.G. Hall, G.R. Davies, J.E. McIntyre, I.M. Ward, D.J. Bannister, and K.M.F. Le Brocq, Polym. Commun., 27 (1986) 98. [130] D. Fish, I.M. Khan, and J. Smid, Makromol. Chem. Rapid Commun., 7 (1986) 115. [131] D. Fish, I.M. Khan, E. Wu, and J. Smid, Br. Polym. J., 20 (1988) 281. [132] H. Tsutsumi, M. Yamamoto, M. Morita, Y. Matsuda, T. Nakamura, and H. Asai, Electrochim. Acta, 37 (1992) 1183. [133] J.R.M. Giles and M.P. Greenhall, Polym. Commun., 27 (1986) 360. [134] H. Cheradame and J.F. Le Nest, In Polymer Electrolyte Reviews-1, J.R. MacCallum and C.A. Vincent (Eds), Elsevier Applied Science, London (1987) p. 103. [135] J.F. Le Nest, A. Gandini, and H. Cheradame, Br. Polym. J., 20 (1988) 253. [136] J.F. Le Nest, S. Callens, A. Gandini, and M.B. Armand, Electrochim. Acta, 37 (1992) 1585. [137] S.D. Druger, A. Nitzan, and M.A. Ratner, J. Phys. Chem., 79 (1983) 3133. [138] H. Gerischer, J. Electroanal. Chem., 82 (1977) 133. [139] A.J. Bard and M.S. Wrighton, J. Electrochem. Soc., 124 (1977) 1706. [140] S.M. Park and S.E. Barber, J. Electroanal. Chem., 99 (1977) 67. [141] R. Noufi, D. Tench, and L.F. Warren, J. Electrochem. Soc., 127 (1980) 2310. [142] R. Noufi, A.J. Frank, and A.J. Nozik, J. Am. Chem. Soc., 103 (1981) 1849. [143] F.R.F. Fan, B. Wheeler, A.J. Bard, and R. Noufi, J. Electrochem. Soc., 128 (1981) 2042. [144] T.A. Skotheim, I. Lundström, and J. Prejza, J. Electrochem. Soc., 128 (1981) 1625. [145] R. Noufi, J. Electrochem. Soc., 130 (1983) 2126. [146] A.J. Frank and K. Honda, J. Electroanal. Chem., 150 (1983) 673. [147] G. Horowiz and F. Garnier, J. Electrochem. Soc., 132 (1985) 634. [148] K. Rajeshwar, M. Kaneko, A. Yamada, and R.N. Noufi, J. Phys. Chem., 89 (1985) 806. [149] J.M. Bolts, A.B. Bocarsly, M.C. Plazzotto, E.G. Walton, N.S. Lewis, and M.S. Wrighton, J. Am. Chem. Soc., 101 (1979) 1378. [150] C.L. Clement, S. Arvamuthan, and K.S.V. Santhanam, J. Electroanal. Chem., 248 (1988) 233. [151] A.B. Ellis, S.W. Kaiser, and M.S. Wrighton, J. Am. Chem. Soc., 98 (1976) 1635. [152] G. Hodes, J. Manassen, and D. Cohen, Nature, 261 (1976) 403. [153] B. Miller and A. Heller, Nature, 262 (1976) 680. [154] A. Heller, K.C. Chang, and B. Miller, J. Electrochem. Soc., 124 (1977) 697. [155] T.A. Skotheim, App. Phys. Lett., 38 (1981) 712. [156] T.A. Skotheim and I. Lundström, J. Electrochem. Soc., 129 (1982) 894. [157] T.A. Skotheim and O. Inganäs, J. Electrochem. Soc., 132 (1985) 2116.

Solid State Organic Photoelectrochemical Solar Energy…

199

[158] O. Inganäs, T.A. Skotheim, and S.W. Feldberg, Solid State Ionics, 18 and 19 (1986) 332. [159] T.A. Skotheim and O. Inganäs, Mol. Cryst. Liq. Cryst., 121 (1985) 285. [160] A.F. Sammels and P.G.P. Ang, J. Electrochem. Soc., 131 (1984) 617. [161] A.F. Sammels and S.K. Schmidt, J. Electrochem. Soc., 132 (1985) 520. [162] R.L. Cook and A.F. Sammels, J. Electrochem. Soc., 132 (1985) 2429. [163] A.K. Vijh and B. Marsan, Bull. Electrochem., 5 (1989) 456. [164] B. Marsan, D. Fauteux, and A.K. Vijh, J. Electrochem. Soc., 134 (1987) 2508. [165] B. Marsan, D. Fauteux, and A.K. Vijh, Solid State Ionics, 28 and 29 (1988) 1058. [166] D. Laser and A.J. Bard, J. Phys. Chem., 80 (1976) 459. [167] P.A. Kohl and A.J. Bard, J. Electrochem. Soc., 126 (1979) 59. [168] M.E. Langmuir, P. Hoenig, and R.D. Rauh, J. Electrochem. Soc., 128 (1981) 2357. [169] M.E. Langmuir, M.A. Parker, and R.D. Rauh, J. Electrochem. Soc., 129 (1982) 1705. [170] H. Gerischer and H. Tributch, Ber. Bunsenges. Phys. Chem., 72 (1968) 437. [171] M. Matsumura, Y. Nomura, and H. Tsubomura, Bull. Chem. Soc. Jpn., 50 (1977) 2533. [172] M. Matsumura, S. Matsudaira, H. Tsubomura, M. Takata, and H. Yanagida, I and E Prod. Res. Devel., 19 (1980) 415. [173] B. O'Regan and M. Grätzel, Nature, 353 (1991) 737. [174] J. Desilvestro, M. Grätzel, L. Kavan, and J. Moser, J. Am. Chem. Soc., 107 (1985) 2988. [175] N. Vlachopoulos, P. Liska, J. Augustynski, and M. Grätzel, J. Am. Chem. Soc., 110 (1988) 1216. [176] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Muller, P. Liska, N. Vlachopoulos, and M. Grätzel, J. Am. Chem. Soc., 115 (1993) 6382. [177] A. Hagfeldt, B. Didriksson, T. Palmqvist, H. Lindström, S. Södergren, H. Rensmo, and Ster-Eric Lindquist, Sol. Energy Mater. Sol. Cells, 31 (1994) 481. [178] H. Gerischer, Advances in Electrochemical Engineering, Interscience, New York (1961) Vol. 1. [179] H. Gerischer, Physical Chemistry: An Advanced Treatise, Academic Press, New York (1967) Vol. 9A. [180] A.J. Nozik, Annu. Rev. Phys. Chem., 29 (1978) 189. [181] H. Gerischer, Electrochim. Acta, 35 (1990) 1677. [182] C.A. Koval and J.N. Howard, Chem. Rev., 92 (1992) 411. [183] M.X. Tan, S.T. Laibinis, J.M. Nguyen, C.E. Stanton, and N.S. Lewis, Prog. Inorg. Chem., 41 (1994) 21. [184] A.J. Nozik and R. Memming, J. Phys. Chem., 100 (1996) 13061. [185] A.J. Bard and L.R. Faulkner, Electrochemical Methods Fundamentals and Applications, John Wiley and Sons, Inc., New York (1980). [186] R.M. Noyes, J. Am. Chem. Soc., 84 (1962) 513. [187] F. Lohmann, S. Naturforsch., 22A (1967) 843. [188] S. Trasatti, J. Electroanal. Chem., 52 (1974) 313. [189] R. Gomer and G. Tryson, J. Chem. Phys., 66 (1977) 4413. [190] Teketel Yohannes, Theodros Solomon, O. Inganäs, Synth. Met., 82(1996)215. [191] Mosissa Adi, Teketel Yohannes, Theodros Solomon, Solar Materials and Solar Cells, 83 (2004) 301. [192] Teketel Yohannes, O. Inganäs, J. Electrochem. Soc., 143 (1996) 2310.

200

Teketel Yohannes

[193] Teketel Yohannes, O. Inganäs, Solar Materials and Solar Cells, 51(1998) 193. [194] Teketel Yohannes, O. Inganäs, Synth. Met., 107 (1999) 97. [195] T.A. Skotheim, S.W. Feldberg, and M.B. Armand, J. Phys. Paris Colloq., C3 (1983) 615. [196] S.M. Sze, Physics of Semiconductor Devices, 2nd ed., Wiley, New York (1981). [197] N.S. Sariciftci, Prog. Quant. Electr., 19 (1995) 131. [198] Q. Pei and Y. Yang, J. Am. Chem. Soc., 118 (1996) 7416. [199] A.K. Ghosh and T. Feng, J. Appl. Phys., 49 (1978) 5982. [200] G.A. Chamberlain, J. Appl. Phys., 53 (1982) 6262. [201] F-R. Fan and L.R. Faulkner, J. Chem. Phys., 69 (1978) 3341. [202] S-An Chen and Y. Fang, Synth. Met., 60 (1993) 215. [203] Y. Fang and S-An Chen, Synth. Met., 52 (1992) 261. [204] H. Meier, Organic Semiconductors, Verlag Chemie, Berline (1974) pp. 318-321. [205] S. Glenis, G. Horowitz, G. Tourillon, and F. Garnier, Thin Solid Films, 111 (1984) 93. [206] S. Glenis, G. Horowitz, G. Tourillon, and F. Garnier, Thin Solid Films, 139 (1986) 221. [207] R.O. Loutfy, J.H. Sharp, C.K. Hsiao, and R. Ho, J. Appl Phys., 52 (1981) 5218.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 201-218

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 5

A NEW APPROACH TO HYBRID SYSTEMS OF RENEWABLE ENERGY UTILIZATION* Yu.V. Vorobiev1, J. Gonzalez-Hernandez2†, P. Gorley3, P. Horley3 and L. Bulat3 1

CINVESTAV, Unidad Queretaro, Queretaro 76230, Mexico 2 CIMAV, Chihuahua 31109, Mexico Department of Electronic and Energy Engineering, Chernivtsi National University, Chernivtsi 58013, Ukraine 3 St. Petersburg State University of Refrigeration and Food Eng., St. Petersburg 191002, Russia

ABSTRACT A general analysis is given of hybrid systems consisting of different combinations of 4 devices frequently employed for renewable energy utilization: Photovoltaic Solar Panel (PV), Solar Thermal Plane Collector (ST), Wind Generator (WG) and Heat-toElectric/Mechanic Energy Convertor (HE); some of the combinations include radiation energy flux concentrators of different degrees. The main result of the consideration made is that the hybrid systems are more efficient than the sum of the constituents and more stable in relation to spontaneous variations of the renewable energy source potential (like wind velocity, insolation, etc.). However, to realize the possibilities mentioned, all the elements of a given hybrid system have to be especially designed and made for this specific system. For example, the PV panel for the hybrid PV/Thermal system ought to have a substrate with high thermal conductivity, to allow for heat extraction from the panel by the adjacent Solar Thermal Plane Collector, and practically no commercial panels with these characteristics are available. Besides, the PV panel as a part of the hybrid system will demand a special choice of semiconductor material and surface *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † On sabbatical leave from CINVESTAV-Queretaro

202

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al. treatment which could be different from those for conventional panels. The limiting efficencies for some hybrid systems are estimated; these efficiencies exceed the efficiencies of separate use of the devices discussed. The most promising hybrid system is the PV panel made as spectrum splitter in combination with HE converter, of which total efficency could be around 50 %.

INTRODUCTION Among many devices employed for the renewable energy generation, a special place is occupied by photovoltaic solar cells, for many reasons (like low maintenance, long lifetime, etc.). The key problem in solar cells science and engineering is the restricted efficiency of Solar Energy conversion resulting from the impossibility of efficient utilization of a wide solar spectrum with one semiconductor material. The possibilities of using more than one material (multijunction cells, spectral splitting with many cells [1-6]) or hypothetical materials with very specific parameters [7,8] do not give a simple solution to the problem because of technological problems and very high cost. On the other hand, the hybrid systems of a different kind recently became very popular (for example, [9-14]); some of them, like PV/Thermal, are developed up to commercial stage. It is evident that these hybrid systems a priori could provide a higher efficiency of utilization of renewable energy source resulting from the smaller total area than that of the sum of separate parts of a system, having at the same time a relatively low cost. In addition the hybrid systems have advantages compared to their elements, since each one of the separate devices forming the system has its own working conditions which quite often are contradictory (for example, the windy weather is profitable for wind generators, but it reduces efficiency of solar thermal collectors), therefore the hybrid system can be made almost independent upon the variation of these conditions, and thus more stable and reliable. The common way to build a hybrid system is to use a combination of existing devices, which is definitely the simplest and the cheapest approach. However, this way may not be the most efficient one. The present paper intends to show that to optimize the construction and performance of any hybrid system, one has to make a special analysis of the coupling conditions of the devices used, and, in the majority of cases, to design each device accordingly to the specific system’s demands. The result will be the highest efficiency and stability of the system.

THEORETICAL ANALYSIS Separate Performance of the System’s Elements Here we give the basic information necessary to discuss the different combinations of the elements (for the 4 elements, there are 10 possible combinations; we shall discuss only those which we consider are the most important). Photovoltaic Solar Panel (PV) with the sensitive area APV, efficiency ηPV is exposed to solar radiation of power W during Δt1 hours per day (the last parameter is the averaged

A New Approach to Hybrid Systems of Renewable Energy Utilization

203

considering the hourly changes in the solar position). Thus, the total energy generated by a PV per day is EPV = ηPV W APV Δt1

(1)

For a typical Si solar photovoltaic module with an efficiency of 12 % and area of 1 m2, under solar illumination corresponding to AM1.5 conditions (844 W/m2), the power generated will be approximately 100 W, and the energy corresponding to the average 5.5 sun hours per day (data taken for the state of Queretaro, Mexico, no Sun tracking is assumed), in agreement with (1) is 550 Wh. Solar Thermal Plane Collector (ST) is converting the sun radiation into heat stored in water or air circulating within the collector. Its efficiency ηΤ is usually 50 – 70 % ([9-11]), and is very dependent upon the working conditions like the temperature difference between input and output, cooling conditions, etc., so thermal energy generated by ST with area AT during the sun hours will be ET = ηT W AT Δt1

(2)

and for each square meter of area, the daily thermal energy produced is around 3 kWh. Wind Generator (WG) utilizes a secondary product of solar radiation – the wind, and its performance is determined, first of all, by the wind velocity v. For a WG “controlling” the area AA of air flux, the power WW converted to electricity is the part ηW (WG efficiency) of the corresponding air flux kinetic energy per unit time which is equal to 0.5 ρVv2, ρ being the air density and V – volume per second of the air flux controlled. For the latter we write V = AA v, and thus obtain WW = 0.5 ηE ρ ΑΑ v3,

(3)

so the total energy EW produced during the working time interval Δt2 will be EW = 0.5 ηE ρ ΑΑ v3 Δt2

(4)

Here the corresponding time interval can be 24 hours per day, although the wind velocity could vary considerably during the day. Taking the flux controlled area equal again to 1 m2, air density ρ = 1.29 kg/m3, wind velocity 8 m/s which is considered as the minimum for a good WG performance, and the effectivity 0.25 (some average value), we get from (3) the power 82.6 W; for 24 hours per day performance, according to (4) it gives approximately 2 kWh of energy. The estimations made show that each of the devices considered produces considerable amount of energy per day (from 0.55 to 3 kWh), which could be utilized for domestic or other applications. The analysis below will show the advantages in combining these elements in hybrid systems, and the specific demands related with the array. Heat-to-Electric/Mechanic Energy Convertor (HE). In this part, some heat engines could be considered (like Stirling Engine frequently used in the field discussed, with extensive literature). Here we present information about thermoelectric generators (TEG) which is less

204

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

familiar to the general audience. The thermoelectric method is a direct method of energy conversion similar to the photovoltaic one, so it is natural to combine these two methods in a hybrid system. It is also important that thermoelectric devices (again, like the photovoltaic ones) are highly reliable and could work 10 – 30 years practically without any technical service (see [15-19]) which unfortunately could not be said in relation to heat engines. The TEG efficiency. The efficiency of a thermoelectric generator TEG is determined by so called thermoelectric figure of merit

Z=

α2 , ρκ

(5)

where ρ is the electric resistivity,

κ - the thermal conductivity and α - the thermo-emf

coefficient of a thermoelectric (semiconductor) used. A modern thermoelectric module (a unit thermoelectric converter) is a battery – a number of alternate n- and p-type semiconductor branches. They are connected electrically in series with metallic connection strips, sandwiched between two electrically insulating and thermally conducting ceramic plates to form a module. So, any TEG contains two different materials (corresponding indexes “1” and “2”), and the optimized figure of merit for a module can be presented in the form

Z0 =

(α1 − α 2 )2

( κ 1 ρ1 +

κ 2ρ2

)2

(6)

The efficiency of TEG can be characterized in two ways: [1]

Efficiency corresponding to the classic condition of the maximum power output R = r (the external electric resistance R equal to the TEG internal resistance r )

⎛ 4 T − Tc +2− h η1 = η 0 ⎜⎜ 2Th ⎝ Z 0Th [2]

⎞ ⎟⎟ ⎠

−1

(7)

Since the thermoelectric devices (the same as photovoltaic ones) are not characterized by a constant internal resistance, there exist another condition for the maximum TEG efficiency corresponding to the relation R = M = 1 + Z 0 T M , where TM is the average temperature r opt

( )

TM = 0.5 (Th + Tc). In this case

η max = η 0

M −1 . Tc M+ Th

(8)

A New Approach to Hybrid Systems of Renewable Energy Utilization

The coefficient

η0 =

205

Th − Tc = ΔT/Th in (7) and (8) is the ideal thermodynamic Th

efficiency of a heat engine (the Carnot efficiency). Usually ηmax is a little higher than η1; as a rule, the difference between the values of the TEG efficiency given by the two expressions above does not exceed 4% [15]. Our purpose is to estimate the highest efficiency of the hybrid system, therefore we use for calculations the expression (8). Thermoelectric materials. To proceed, we have to specify value of the figure of merit Z . It is necessary to point out that the figure of merit of semiconductors depends on temperature, and different kinds of semiconductor materials should be selected for different operating temperatures. The industrial thermoelectric materials can be divided in dependence of operating temperature into the following groups [15-18]: -

for temperatures up to 500K - the solid solutions based on bismuth telluride (Bi2Te3, Bi2Te3-Bi2Se3) are used; for temperatures up to 800K - the PbTe; for space applications (T>900K) - the solid solutions based on Ge-Si.

The main problem of thermoelectric material research is how to increase the figure of merit Z . Despite the efforts of many research groups in different countries, during 1950 2000 years the increase achieved in the dimensionless thermoelectric coefficient ZT was only from 0.75 to 1.0 (the data refer to room temperature). But an essential progress in the field was made during the last few years. The remarkable successes have been achieved on different directions: 1. Usage of the new physical ideas in nano-scale microstructures. A high quality 2D quantum superlattice with nano-scale films based on p-Bi2Te3/Sb2Te3 having ZT = 2.4 at room temperature was obtained [19-21]. A nanomaterial with quantum dots (1D structure) based on PbSeTe has the value of ZT = 2.0 at room temperature [22]. The increase of the figure of merit was also obtained in a special structure with cold points (like contacts between a plate and edge of a cone). The measured figure of merit of the last structure based on p-Bi0.5Sb1.5Te3 and n- Bi2Te2.9Se0.1 corresponds to ZT = 1.7 at room temperature [23,24]. 2. Alongside with semiconductor TEG there exist thermionic devices that also can be used for direct generation of electricity. However, the traditional thermionic converters work only at very high temperatures, above 800 К, necessary for electrons to overcome the potential barrier. But recently a theory has been offered [25] of a thermionic converter with a nanometers barrier thickness; here electrons can overcome the potential barrier by quantum tunneling. Apparently such thermotunnel converter has been realized [26]. The parameter equivalent to thermoelectric figure of merit of this converter is ZT = 4. 3. Another option is the application of some new bulk materials as thermoelectrics – so called skutterrudites and clathrates. It was shown that the thermal conductivity can be reduced in such materials [27], therefore the figure of merit will increase. The examples of skutterrudites are CoSb3 (n-type) and Zn4Sb3 (p-type) and of the

206

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al. clathrates - CoFe4Sb12. These materials have good figure of merit in a wide temperature range and therefore can be used for TEGs.

Thus, for calculation of the efficiency of hybrid system in the temperature range 300 – 600 K we can use the following thermoelectric parameters: ZT = 1, value that have modern industrial thermoelectrics; ZT = 2, this value has thermoelectrics received in laboratories (nano-scale microstructures), and ZT = 4, value for thermionic converter with quantum tunneling. Using the expression (8) to estimate the efficiency and taking the temperature difference between hot and cold TEG sides equal to 300 K, we obtained for the three cases mentioned the corresponding efficiency of 10, 15 and 20 %.

Hybrid Systems (Different Combinations of Elemental Devices) Thermal-Wind (ST+WG). This complex may be useful for thermic applications (for example, to stabilize the temperature regime of a greenhouse or small living house in case of great daily variations of temperature, which is typical for Mexico), the energy generated during the day by ST is collected by hot water which is stored in thermally insulated water tank, to be used at night for heating of a building. The electricity generated by the WG will be used also for heating (through an electric heater with an efficiency practically of 100 %, or in a heat pump system where the efficiency could be almost doubled). The efficiency of the ST decreases almost linearly with the wind velocity; this decrease ought to be compensated by the effect of wind generator WG: its power is never linear with the wind velocity, so the compensation will be within some limits; actually, for a strong wing, there could be overcompensation. We can introduce an average wind velocity va and find the conditions of energy stabilization in relation to relatively small deviations (± Δv) of this velocity, taking |Δv| < va. Thus, for the ST efficiency we write ηT = ηTo (1 − α v) = ηTo [1 − α( va ± αΔv)], so that ET = C [1 − α( va ± αΔv)], where C = W AT Δt1

(2A)

and for the WG EW = C* (va ± Δv)3 ≈ C* (va3 ± 3 va2Δv),

(4A)

C* = 0.5 ηE ρ ΑΑ Δt2. Then for the total energy produced by the system per day we have Etot = [C(1 − α va) + C* va3] ± (3 C* va2 − Cα) Δv

(9)

A New Approach to Hybrid Systems of Renewable Energy Utilization

207

Now we have the condition of stability of the system energy as (3 C* va2 − Cα) = (1.5ηE ρ ΑΑ Δt2 va2 − W AT α Δt1) = 0

(10)

We shall see below how this condition could be applied to a real system. However, one general remark could be made now. The wind generators most frequently used (HA, or Horisontal Axis devices) demand very high wind velocity for normal performance (8 m/s or more). At these conditions, the thermal collectors have very low efficiency, and their use is not practical; the supportable wind velocity for their application is 3 – 5 m/s. It means that for the hybrid system of the kind discussed, the HA WG is not a good choice; at relatively low wind velocities, the Vertical Axis (VA) devices are more efficient. Photovoltaic-Thermal (PV + ST) System. This system (so-called “combi-panel”) has an evident advantage compared to the performance of the separated devices. In that system the ST absorbs the excessive heat of the PV thus cooling it and therefore enhances the efficiency of the PV panel. It is also clear that the total amount of energy produced per unit area increases by this combination. The detailed analysis of thermal balance in this system was published in our previous paper [28]; the results of the present analysis support the main conclusion. All the previous papers on the subject have taken for granted that the systems elements have to be of the same area. However, it is evident that the presence of the PV panel above the solar collector reduces the heat flux to the collector and thus its efficiency; on the other hand, the collector ability to extract heat from the PV panel is reduced while the water (air) inside it is heated. Therefore, the optimal case would be to make PV panel of smaller area than that of the heat collector, and to place it above the initial collector’s part (i.e. that corresponding to the entrance of the cold water). Another point is that the conventional PV usually could not provide a good thermal contact with the ST heat collector, so, this hybrid system demands a special construction of the PV with a high thermal conductivity of the substrate used; one example is given in the paper mentioned [28]. Wind-Photovoltaic (WG + PV) System. These systems are already in use; they are unique among the hybrids discussed in a sense that there are no specific demands to the devices combined. The system generates and stores electric energy day and night; since the WG has more working hours per day than the PV, low-power WG could match the higher power PV. To guarantee the normal performance of the system at different conditions of weather, the excessive battery bank ought to be used. This system is more effective in places with very high wind velocity: in addition to the WG driving, the wind stabilizes the PV temperature and therefore the efficiency of its performance. Wind-Thermal-Photovoltaic (WG+ST+PV) System. Here we consider the PV/Thermal system discussed above in combination with the wind generator WG, the system provides electricity generated by PV panel (at daytime) and by the wind generator (day and night), as well as the hot water from the ST collector (ought to be stored in thermally insulated vessel to be used in absence of illumination by sun). The total energy (electric + thermal) produced by the system per day will be Etot = (ηPV APV + ηT AT)WΔt1 + 0.5 ηE ρ ΑΑ v3 Δt2.

(11)

208

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

The specific demands to the PV based on the necessity of a good heat exchange between the PV and the ST were discussed above. Photovoltaic-Heat-to-Electric (PV+HE) system. The scheme of a hybrid system discussed is presented in figure 1, parts A and B refer to the two different options (see below). It is evident from the figure that the two versions are very similar and contain the same basic elements (the concentrator “CONC”, photovoltaic cell “PVC”, High Temperature Stage “HTS” which is the HE converter, and the 2-axis Solar Tracking System “STS2”), although the type and construction of the cell (and the concentrator) could be different in the two cases. Below we discuss the working principles and possible parameters of the system in each version. A. System with separation of “thermal solar radiation” and low temperature operation of solar cell. (figure 1 A). This system needs a solar cell which at the same time acts as a spectral filter (splitter) neither absorbing nor dissipating the solar radiation part with quantum energy hν < Eg (“thermal solar radiation”) where Eg is the cell absorbing material band gap. It is shown below that this part could be quite large, especially for semiconductor materials with relatively wide band gap. Thus, in case of a semiconductor with Eg = 1.75 eV, approximately 50 % of solar radiation energy corresponds to the condition hν > Eg and is suitable for photovoltaic conversion, and another 50 %, with hν < Eg, could be used as thermal energy. In our system, this “thermal solar radiation” should be concentrated on the hot side of the HTS (thus providing its high temperature), and is converted by it into electric energy when HTS is Thermoelectric Generator (TEG), or to mechanic energy by heat engine as the HTS (like Stirling Engine [29] which has efficiency very close to that of the Carnot cycle). It is known that mechanic energy could be converted into electricity with efficiency above 90 % which allows to have rather high total energy conversion coefficient of the hybrid system.

Figure 1. Schemes of hybrid system with High Temperature Stage.

To calculate the percentage of “thermal solar radiation” as a function of semiconductor band gap, we used the simple graphical analysis of solar radiation spectrum introduced by Henry [2] and included now in textbooks [30], with a numerical integration to determine the function nph (Eg), the solar flux absorbed by a semiconductor (as in Henry paper, we assume that the semiconductor is opaque for photon energies greater than Eg and transparent for energies less than Eg). This flux is given by

A New Approach to Hybrid Systems of Renewable Energy Utilization ∞

n ph ( E g ) = ∫ E

g

dn ph dhω dhω

209

(12)

The function obtained is shown in figure 2: the lower curve (squares) for AM1.5 spectrum, and upper one (circles) – for AM0. Following the graphical analysis procedure [2,30]), on the basis of these curves we found the percentage ξ of “thermal solar radiation”: in figure 2 it is shown by triangles (upper curve for AM0, lower for AM1.5; the corresponding values of the figure 2 ought to be multiplied by 10 to get ξ in %). One can see that ξ varies between 10 – 12 % for Eg = 0.8 eV and 78 – 80 % for Eg = 2.3 eV. The ξ vaues obtained for AM1.5 spectrum are also included in figure 3 (circles). This thermal part of solar radiation energy has to drive the HTS (heat engine or TEG); assuming that HTS efficiency is proportional to that of Carnot cycle, with a coefficient K < 1 (the difference between K and 1 shows how close the HTS is to the ideal engine; in general, K could be temperature dependent, in particular, in the TEG case). Thus, this part of solar energy conversion is characterized by the efficiency ηtherm = ξ⋅K ΔT/Th

(13)

nph (1017 cm-2 sec-1 eV-1), 0.1 ξtherm

where Th is the temperature of the hot side of the HTS, and ΔT the temperature difference between cold and hot sides.

8 6 4 2 0

0

1

2

3

ENERGY(eV)

Figure 2. Graphical analysis of the efficiency of an ideal solar cell, for AM1.5 and AM0 solar spectra, with the calculated percentage of “thermal” radiation.

The solar cells for hybrid system of this kind have yet to be designed and made, but we do not see any principal obstacles to that. To estimate the cell efficiency ηCA as a function of band gap, we took 0.75 of the corresponding cell’s ideal efficiency at non-concentrated radiation ηid [2,30]; since for all well developed cells (based on Ge, GaAs, CuInSe2) the best efficiency obtained is well above this value, we consider this approximation reasonable. The corresponding values (ηCA = 0.75 ηid) for the AM1.5 irradiation are shown in figure 3 by

210

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

squares (lower curve). We may add that if a variety of these cells-filters will be available, the cells with decreasing band gaps could be connected optically in series to increase the total efficiency, as an alternative to the multijuncion cells. An additional note is that the cell for this system should not have any surface texturizing which is typical for conventinal solar cells. Using for the HTS efficiency the expression (13), we get the total system efficiency

Efficiency; "Thermal" radiation (%)

ηtotA = ηCA + ηtherm = ηCA + ξ⋅K ΔT/Th

(14)

80

60

40

20

0

1,0 1,2 1,4 1,6 1,8 2,0 2,2 2,4

Eg, eV

Figure 3. Percentage of “thermal” radiation in the AM1.5 spectrum, PV cell estimated efficiency and the total system efficiency as function of Eg..

The curves calculated according to (14) for K = 0.8, ΔT = 500 and 1000 K, are shown in figure 3 by up- and down-sided triangles. In this case, the total efficiency could exceed 50 55 %, without an employment of expensive high technologies. We must note that the temperature differences used for our calculations do not exceed those practically employed in Stirling engines (for example, Bernd Kammerich 1 kW Engine, Germany). B. System without the solar spectrum division, a PV cell operates at high temperature. This is more straitforward system (figure 1B) than the one discussed previously; the conditions of the PV cell performance are not quite favorable here, but the system, in principle, could be constructed on the basis of the elements which already exist (although the lifetime of the PV cell in question needs special study for this working regime). The cell is subjected to concentrated sunlight, which usually enhances its efficiency; the thermal flux through the cell is transfered into the HTS by direct thermal contact, thus the cell working temperature is equal to the Th parameter of the HTS.

A New Approach to Hybrid Systems of Renewable Energy Utilization

211

Efficiency, %

40 30 20 10 0 300

350

400

450

500

550

600

Th , K

Figure 4. Calculated temperature dependencies of the PV cells efficiency (squares, triangles) and of the total hybrid system efficiency, in case of the heat engine as HTS (circles, diamonds) and the TEG (three intermediate curves, ZT = 1, 2, 4).

For calculation of the expected system efficency, two types of cells with relatively high (but not record) efficiency are considered: a GaAs single junction cell [31] with room temperature efficiency (ηο) of 24 %, and multijunction GaAs-based cell with corresponding efficiency of 30 % [3]. Both values refer to radiation concentration of approximately 50 Suns which is sufficient to achieve cell temperature higher than 450 K. The temperature dependence of cell efficiency η(T) is assumed linear with the coefficient β = (dη/dT)/η equal to – 2.7X10-3 K-1 [32]; it gives for the temperature 450 K the efficiency values η∗ = 14.3 % for single junction (SJ) cell, and 17.8 % for multijunction (MJ) cell. Thus, practically 80 % (i.e. 1 – η∗) of solar radiation energy will be transformed to heat within a cell, and may be used for a heat-to-electric/mechanic energy conversion by the second stage of a hybrid system – a HTS. The total conversion efficiency of hybrid system could be written as ηtotB = η∗ + (1 – η∗) ηHTS

(15)

Here we have η∗ = ηο (1 − β ΔΤ), ΔΤ = Τh − Τroom ≈ Τh − Τc, (Τc is the temperature of the cold size of the HTs which is approximately equal to the ambient temperature Τroom), and for the HTS we take, as in part “A”, that it is proportional to the Carnot engine efficiency: ηHTS = K ΔT/Th, which gives for the total conversion efficiency ηtotB = ηο (1 − β ΔΤ) + [1 − ηο (1 − β ΔΤ)] K ΔT/Th.

(16)

Figure 4 shows the calculated results for the hybrid system with the two cells mentioned above; the temperature dependence of the cells efficiency is shown by squares (SJ) and triangles (MJ). The total efficiency with the heat engine acting as HTS is represented by the growing curves, (circles for single junction cell, diamonds for multijunction cell, K = 0.8).

212

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

One could see that the second stage gives significant increase in the total efficency for relatively modest values of ΔT, and the total efficency could be as high as 35 – 40 %. For the system “B” having the thermoelectric generator TEG as the HTS, we present in figure 4 the results of calculations of the total efficiency for three ZT values given above, taking as a basis the equation (15) and using ηmax (8) instead of ηHTS. The PV cell is supposed to be of MJ type, with the room temperature efficiency of 30 %; the corresponding results for ZT = 1, 2 and 4 are shown in figure 4 by three descending curves starting from efficiency 30 % at 300 K (the larger ZT, the higher curve). It is seen that the TEG has considerable effect on the efficiency in all cases, although the total efficiency of the system with TEG is lower than with heat engine. Taking into account the extraordinary recent progress in the TEG field, we can expect further increase of the TEG efficency, and that of the hybrid system.

EXPERIMENTAL The basic elements employed in experimental study were the following. The solar heat collector ST used was of the model Powermat, with water as a heat collecting agent, having a surface area of 4 m2; approximately 90 % of the surface area is controlled by the internal tube system providing an efficient heat interchange. The black PVC absorber covering the collector is resistant to UV solar radiation (the guaranteed lifetime is 20 years) and provides a small weight of the panel (about 3 kg in empty state, and 7 kg when filled with water). The maximum heating efficiency at the absence of the wind estimated using the Hottel-Whillier model [11-13] was around 60 – 70 %, which corresponds to the best collectors known. This result is illustrated in figure 5, which gives the dependence of the ST efficiency upon the temperature difference between the water entering the collector and the ambient temperature Ta. 1.00

Efficiency (%)

0.75

0.50

0.25

0.00

0.008

0.016

0.024 o

0.032

2

(Ti-Ta)/G ( Cm /W) Figure 5. The temperature dependence of the thermal efficiency of the ST studied.

A New Approach to Hybrid Systems of Renewable Energy Utilization

213

The dependence is practically linear, in good agreement with the Hottel-Whillier model, and the maximal value is close to 70 % indicating a good performance of the collector. The expression used to find the ST efficiency from experimental data was: ηT = Mt Cc (To – Ti) / GS

(17)

where Mt is the water flux (mass per second), G – solar radiation intensity, Cc - the water specific heat capacity, To is the output water temperature, and Ti - its input temperature. The PV panels investigated and used in PV/Thermal systems were of crystalline Si type (c-Si; in particular, we used the panel made by Russian plant OKB “Krasnoe Znamya” OKBKZ M100/12 with an area of about 1 m2 and the power generated under AM1.5 solar radiation of 100 W, and the other one assembled in our laboratory from 85×85 mm2 cells made by the same company). Besides, in the experiments designed to achieve a good thermal contact with the collector, we used the panels made of amorphous Si (α-Si, ECD Company, Troy, Michigan, USA) and CuInSe2 (CIS) commercial panel made by Siemens, USA. The electric efficiency of all the PV panels used was between 10 and 18 %. The prototype panel which we constracted to achieve the best thermal contact was made of Al substrate (thickness 2 mm) covered with 1 μm thick insulating film of PMMA/silica composite dielectric coating, with OKBKZ-made six c-Si solar cells (85X85 mm2) area, and a glass cover. One option to increase the efficiency of conventional solar panels is the use of bifacial panels [33,34]; to take full advantage of the rear photosensitive face, the panel ought to be placed at some distance from the white diffuse reflecting surfaces [34], or small degree radiation concentrators might be used. We employed the latter option (see photo in figure 6), the concentrator made of 5 stainless steel plates collected solar radiation from an area twice as big as the panel. It resulted in considerable heating of the panel; to overcome this effect, we plan to develop the water-running plane heat collector transparent in visible (not to affect much the rear face photovoltaic performance) and put it in thermal contact with the panel rear face. This system will produce electric energy and hot water, as any PV/Thermal system; we expect it to be useful for future domestic applications.

Figure 6. Bifacial solar cell with radiation concentrator for rear face.

214

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

The wind generator WG was developed and built in our laboratory (figure 7). The vertical axis (VA) model of Savonius type, 3 m high, with an active area ΑΑ = 1.5 m2, having an automobile alternator with a home made gear-box as generating unit, under conditions of moderate wind, typical for the region of application (state of Querétaro, Mexico, average wind velocity about 5 m/s) provides approximately 40 W of electric energy operating at 120 rpm (see figure 8).

Figure 7. The constructed Vertical Axis Wind Generator . 60

Power (W)

50

40

30

20

40

60

80

100

Rotation velocity (rpm) Figure 8. Characteristics of the experimental wind generator.

120

A New Approach to Hybrid Systems of Renewable Energy Utilization

215

The PV+ST+WG system made according to the discussion above, is schematically shown in figure 9 (construction scheme – left part, electric scheme – right one). The electric part of the system includes the battery bank (4 sealed lead-acid batteries “Prism” 12 V, capacity of 105 AH each), Solar Charge Controller of the model Steca (Germany) and the DC-AC Inverter Proam (China). The arrangement made to investigate the regime of the solar thermal collector ST included two home-made electrical digital thermometers (1, 2 in figure 9) based on the temperature sensors LM335, the water flux meter CICASA Delaunet MD-15 (3 in figure 9), all three devices mentioned were connected to the computerized data acquisition system. The electrical connections scheme is made according to the generally accepted rules and notations, and does not demand comments.

Figure 9. Construction and electric shemes of the PV/ST/WG hybrid system (see text).

The devices used for study of hybrid systems performance were capable of production per day at normal working conditions discussed above EPV = 0.55 kWh (PV panel M100/12) and EW = 0.96 kWh (home made WG) of electric energy, plus ET = 12 kWh of thermal energy (ST Powermatt; the value given refers to the absence of wind). From (3) we find that our WG generating 40 kW at wind velocity va = 5 m/s has an efficiency ηE = 0.3 which could be considered as a good parameter. At the wind velocity mentioned, the ST will produce, according to (2A), 10.1 kWh of thermal energy per day (to find this value, we determined experimentally the coefficient α in (2A), taking measurements of the ST efficiency depending on v. Thus we obtained α = 0.032 s/m). In Thermal-Wind Hybrid System (ST+WG) designed for building heating purposes, according to discussion above, at average wind velocity of 5 m/s (expression (9)) it will provide approximately 11 – 12 kWh of heat energy, depending on the way of transforming of electric-to-heat energy. To discuss the heat production stability of the system in relation to small variations of wind velocity, we have to calculate the variations in the energy generation by the system’s elements according to (10). For the parameters given, we obtain the first part of (10) describing the dependence of WG productivity upon the changes of v (1.5ηE ρ ΑΑ Δt2 va2) equal to 0.576 kWhs/m, and the second part (similar dependence of the ST productivity, − W AT α Δt1) equal to − 0.594 kWhs/m. Thus we see that their sum is reasonably close to zero, i.e. our system is well balanced. According to the values found, the daily variations in

216

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

energy production caused by changes of wind velocity Δv = ± 3 m/s will be around 0.05 kWh, which is less than 0.5 % of the total energy generated per day. In PV/Thermal Hybrid System (PV+ST) discussed above we obtain, in first approximation, the above mentioned amount EPV of electric energy and ET of thermal energy, with the difference that now the whole energy is provided by the area AT, not the sum of areas AT + APV. To get an additional advantage caused by the PV cooling in thermal contact with the ST, a special PV construction is needed; we have shown [28] that the actual increase of the PVM efficiency could be as high as 10 %. Our Wind-Photovoltaic Hybrid System (WG+PV) produces daily approximately 1.5 kWh of electric energy at average conditions of state of Queretaro, Mexico (we should stress that our model of WG is good for moderate wind velocities; for more windy places, the HA models will be better). For actual utilization of this energy with conentional electrical devices, an energy storage, control and conversion appliances are necessary, as shown in figure 8. The Wind-Thermal-Photovoltaic System (WG+ST+PV) provides at average conditions of the state around 1.5 kWh of electric energy and 10 kWh of thermal energy. Depending on the necesities and the actual climatic conditions, any part of electric energy produced can be used for stabilization of thermal regime; on the other hand, this amount of energy is sufficient to operate one remote classroom equipped with receptor of educational satellite programs and TV-video set during 6 hours a day, which was proved experimentally in one of the rural schools of the state of Queretaro, Mexico (the classroom is in constant use since August 2002). In relation to hybrid systems with High Temperature Stage (PV+HE), the preliminary investigations were made of the performance of different solar cells (those of α-Si, CuInSe2, c-Si with p-n junction and with Schottky barrier) at high temperatures, using concentrated and non-concentrated solar radiation. For the former type of experiments, the concentrator with 2axial Sun tracking system was made (figure 10), the radiation concentration degree around 30, providing the temperature at focal plane up to 200 oC. The corresponding theory for high temperature cell performance was also developed. Figure 11 shows theoretical and experimental dependencies of parameters of the c-Si solar cells with p-n junction at the temperature interval 60 – 170 oC (the detailed description of theoretical model and experimental equipment is given in [35]). The results obtained are related to type “B” of the system discussed, and show the definite possibility of solar cell to be used at high temperatures. For the realization of all the advantages of the system (both “A” and “B” types), the new cells are necessary, and their design and construction are in process now.

A New Approach to Hybrid Systems of Renewable Energy Utilization

217

Figure 10. Sun-tracking radiation concentrator with c-Si solar cell in focal plane.

1,0

Uoc

P

0,8

2

Voc [V], Isc [A/cm ], P [W]

Isc

0,6

0,4

0,2 60

80

100

120

140

160

180

0

Temperature [ C]

Figure 11. Temperature dependence of c-Si solar cell parameters:theory (solid symbols) and experiment (hollow symbols).

CONCLUSION The recent development in solar photovoltaic cell technologies brought great progress to the field, but unfortunately, there are no clear prospects today to develop economic cells with efficiency higher than 20 %. Hybrid systems of different kinds promise much higher efficiency of utilization of solar (renewable) energy, which could be in total around 40 – 60 %, with a reasonable cost and some other advantages. However, hybrid systems, as a rule, are not efficiently working when made by a mechanical combination of the existing elemental devices. Some of them, if not all, have to be designed and constructed especially for a particlular hybrid system. Thus, a photovoltaic panel for applications in a PV/Thermal system has to be designed in an optimized way to provide a good thermal exchange with the heat collector, which needs specific materials and arrangements. The ideal for hybrid systems would be a photovoltaic panel transparent for part of the solar spectrum (infrared one) which is not absorbed by its semiconductor material, and having a good heat exchange with the heat collector. On the other hand, the solar heat collector could be made transparent in the visible region, to be able to use with bifacial panels and not affect greatly their performance. The experimental systems designed, built and studied confirm the conclusions made.

REFERENCES [1] [2] [3] [4]

Karam, N.H.; King, R.R., et al., Sol. En.Mat. Sol. Cells., 2001, vol. 66, pp. 453-466. Henry, C.H., J. Appl. Phys., 1980, vol. 51, pp. 4494-4500. Yamaguchi, M., Sol. En. Mat. Sol. Cells., 2003, vol. 75, pp. 261-269. Tobias, I; Luque, A., Progr. in Photovolt., 2002, vol. 10, pp. 323-329.

218 [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

[27] [28] [29] [30] [31] [32] [33] [34] [35]

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al. Andreev, V.M.; Grilikhes, V.A., et al., Sol. En. Mat. Sol. Cells., 2004, vol. 84, pp. 3-17. Imenes, A.G.; Mills, D.R., Sol. En.Mat. Sol. Cells., 2004, vol.84, pp. 19-69. Luque, A.; Marti, A., Phys. Rev. Lett., 1997, vol. 78, pp. 5014-5017. Luque, A.; Marti, A., Progr. in Photovolt., 2001, vol. 9, pp. 73-86. Kern Jr., E.C.; Russel, M.C., Proc. 13 ISES PVSC, Washington D.C., USA, 1978, pp. 1153-1157. Huerta, J., et al., Proc. ANES/ISES, Mexico, 2000, pp. 529-533. Raghuraman, P., J. Sol. En. Eng., 1981, vol. 103, pp. 291-298. Takashima T., et al., Sol. En., 1994, vol.52, p.241-245. Florschuetz, L.W., Sol. En., 1979, vol. 22, pp. 361-366. Zondag, H.A., et al., Sol. En., 2002, vol. 72, pp. 113-128. Rowe, D.M., Ed.;CRC Handbook of Thermoelectrics, CRC Press, London, N.Y., 1995, 702 p. Anatychuk, L. I. Thermoelements and Thermoelectric Devices (Handbook), Naukova Dumka, Kiev, 1979, 768 p. Bulat, L.P., et al., Thermoelectric Cooling. St. Petersburg, SPbSURandFE, 2002, 47 p. L.P.Bulat, L.P., Kholodilnaya Tekhnika, 2004, No. 8, 7 p. Fleurial, J.-P., et al., Proc. of 23rd International Conference on Thermoelectrics, Book of Abstracts, Adelaide, Australia, 2004, p.70. Venkatasubramantan, R., et al., Nature, 2001, vol. 413, pp. 597-608. Venkatasubramantan, R., US Patent No.: 6,300,150,B1. Oct. 9, 2001. Harman, T.C., et al., Science, 2002 September 27; 297: 2229. Ghoshal, U., Ghoshal, S., McDowell, C., Shi, L., Appl. Phys. Letters, 2002, vol.80. pp. 3006-3008. Ghoshal, U., Proc. XXI International Conf. on Thermoelectrics, 2002. IEEE, p.540. Hishinuma Y., Geballe, T.H., Moyzhes, B.Y, Kenny T.W., Appl. Phys. Lett., 2001, vol.78, p.2572. Tavkhelidze, A., Skhiladze, G., Bibilashvili, A., Tsakadze, L., Jangadze, L., Taliashvili, Z., Cox, I., Berishvili, Z., Proc. XXI International Conf. on Thermoelectrics, 2002. IEEE. P.435. Caillat, T. et al., Proc. XX International Conf. on Thermoelectrics, 2001. IEEE, p.282. Zakharchenko, R., et al., Sol. En. Mater. Sol. Cells., 2004, vol. 82, pp. 253-261. Organ, A.J., The Regenerator and the Stirling Engine, Harvill Press, 1997, pp. 35-88. Sze, S.M., Physics of Semiconductor Devices, 2nd ed., John Wiley and Sons, N.Y., 1981, pp. 830-835. Algora, C., et al., IEEE Trans. Electron Devices, 2001, vol. 48, pp. 840-844. Nann, S.; Emery, K., Sol. En.Mat. Sol. Cells, 1992, vol. 27, pp. 189-216. Luque, A., et al., Sol. Cells, 1980, vol. 2. pp. 151-166. Luque, A., et al., Sol. Cells, 1984-85, vol. 13, pp. 277-292. Meneses-Rodriguez, D., et al., Sol. En. J., 2005, vol. 78, pp. 243-250.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 219-229

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 6

DYE-SENSITIZED NANO SNO2:TIO2 SOLAR CELLS* Weon-Pil Tai Fine Chemical Industry Support Center, Ulsan Industry Promotion Techno Park; 2F Small and Medium Business Center, 758-2, Yeonamdong, Buggu, Ulsan 683-804, South Korea

ABSTRACT The nanostructured SnO2:TiO2 bilayered and composite solar cells sensitized by eosin Y and RuL2(NCS)2 dyes are prepared and the photoelectrochemical properties of the cells are investigated. The semiconductor films possess the grain size of nanometer order and have nanoporous structure. The bilayered cell shows higher IPCE (incident photon- to-current conversion efficiency) value than the single and composite cells. A maximum IPCE value of 88.1% was reached at 540 nm wavelength in the bilayered cell with 3.5μm-thick SnO2 and 7μm-thick TiO2 sensitized by RuL2(NCS)2 dye. The higher IPCE value in the bilayered cell is attributed to the promotion of the charge separation by fast electron transfer process from the excited dye to SnO2 in the SnO2/TiO2/dye system with different conduction band edge energy positions.

1. INTRODUCTION Nanostructured materials are, recently, an intensive research area with many potential applications. Dye-sensitized nanostructured TiO2 solar cells having a wide-band gap offer an alternative method for the fabrication of low-cost solar cells [1-5]. The interaction of carboxylic acid groups with a porous TiO2 film surface results in effective monolayer adsorption of the dye molecules onto porous TiO2 films. A monolayer of dye adsorbed onto the porous TiO2 surface is sufficient to collect a large part of the solar spectrum. Such solar cells possess the large surface area and high porosity of the nanostructured TiO2 film, *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

220

Weon-Pil Tai

enhancing the light-harvesting capability of the dye adsorbed onto the film surface. The chemical and physical processes involved in the operation of these solar cells take place in a two-phase system consisting of a nanostructured porous TiO2 film interpenetrated by a I3-/Iredox electrolyte. Charge injection from the photoexcited dye and regeneration of the dye by electron transfer from I- lead to transport of electrons in the TiO2 as well as transport of I3and I- ions in the electrolyte. Electron transfer from I- to the oxidized dye and regeneration from I3- to I- at the counter electrode link the two transport processes. The nanostructured-bilayered (coupled) solar cells have shown an improved IPCE (incident photon-to-current conversion efficiency) value compared with single layered solar cells [6-9]. The charge recombination between the electrons injected in the conduction band of the semiconductor and the oxidized sensitizer could be suppressed by using two semiconductors with different energy levels, i.e., different conduction band edge energy positions. A better charge separation in the bilayered film electrode can be achieved by using two different oxide semiconductors e.g. SnO2 and TiO2 having different energy levels. This suppresses the charge recombination in the bilayered film. Namely, the electrons injected from the excited RuL2(NCS)2 dye molecule (E°= -0.66V vs NHE, in acetonitrile) [10] into the conduction band of TiO2 (ECB = -0.5V vs NHE) [11] could be transferred quickly into the lower lying conduction band of SnO2 (ECB = 0V vs NHE) [7]. Furthermore, Tennakone et al. [12] reported the suppression of charge recombination for the composite of SnO2 with a nanocrystalline size of 15 nm and ZnO with microcrystalline size of 2μm. Kiesewetter et al. [13] also tried to improve the photocurrent by embedding the large microcrystalline WSe2 and MoS2 (absorbers) into the nanocrystalline TiO2 without dye. In this paper, nanostructured SnO2:TiO2 thin films were prepared by the sol-gel process and the photoelectrochemical properties of SnO2/TiO2 bilayered solar cells sensitized by eosin Y and RuL2(NCS)2 dyes were investigated. The photoelectrochemical properties of RuL2(NCS)2 dye-sensitized TiO2-SnO2 composite cells were also investigated for the direct comparison in IPCE.

2. EXPERIMENTAL DETAILS 2.1. Synthesis TiO2 colloidal solution was prepared by the following procedure. 150 ml of Titanium tetra isopropoxide (Ti(C3H7O)4, Wako pure chemical, 95%) was rapidly added to 270 ml of deionized water. The resulting precipitate was washed with deionized water. The precipitate cake was transferred into a well-sealed autoclave vessel containing 0.5 M tetramethylammonium hydroxide ((CH3)4NOH), Aldrich). Peptization occurred after heating at 1100 for 6 h. The suspension which resulted from peptization was treated hydrothermally in the autoclave at 1900 for 6 h. SnO2 colloidal solution was prepared by the following procedure. Tin chloride pentahydrate (SnCl4·5H2O, Wako pure chemical) was dissolved in ethanol (0.4 M). The solution was well stirred and then refluxed for 3 h. An 5.5 M aqueous ammonia solution (Wako pure chemical, 25~27.9%) was added dropwise to a refluxed solution and the resulting

Dye-Sensitized Nano SnO2:TiO2 Solar Cells

221

precipitate was washed thoroughly with deionized water to remove NH4Cl. Finally, SnO2 colloidal solution was prepared by adding deionized water.

2.2. Preparation of ITO/SnO2:TiO2 Electrodes The colloidal TiO2 dispersions were prepared as follows. TiO2 colloidal solution of 6 g and commercial TiO2 (Nippon Aerosil, P25) of 0.08 g was ground in a mortar. A detergent (Triton X-100) of 20μL was added to facilitate the spreading of the colloid on the substrate. Polyethylene glycol (PEG 20000) of 0.35 g was added to facilitate the adsorption of dye molecules onto porous film. Finally, the colloid was diluted by the addition of ethanol (3 mL). The colloidal SnO2 dispersions were prepared by the same method as the colloidal TiO2 dispersions except the addition of commercial TiO2. The TiO2-SnO2 composite solution, a viscous dispersion of SnO2 and TiO2 mixed nanocrystallites, was prepared by adding SnO2 colloidal solution into TiO2 colloidal solution with different ratios. There were two types of electrodes, SnO2/TiO2 (a upper TiO2 layer on a lower SnO2 layer) bilayered electrode and TiO2-SnO2 composite electrode. In the bilayered electrode, nanostructured SnO2 film was prepared first by spreading a viscous dispersion of nanocrystalline SnO2 onto a conducting ITO (indium tin oxide)-coated glass substrate(14 ohm/□, Sanyo vacuum industries), and then coated by spreading a viscous dispersion of nanocrystalline TiO2 onto the nanostructured SnO2 film. In the composite electrode, nanostructured TiO2-SnO2 composite film was prepared by spreading a viscous composite solution onto a conducting ITO-coated glass. After air drying, the film electrodes were fired at 500°C for 30 min in air using an electric furnace, at an increasing rate of 50/min. The resulting film thickness was about 7~14.5μm. A dye molecule, cis-di(thiocyanato)-N,N'-bis(2,2'-bipyridyl-4,4'-dicarboxylic acid)ruthenium(II) dihydrate, referred to as a RuL2(NCS)2 dye, was synthesized in accordance with published procedures [2, 14]. IR spectra exhibited 2117, 1720, 1612, 1548, 1408/cm, similar to the infrared peaks of commercial RuL2(NCS)2 dye (Solaronix SA, Ruthenium 535). Also, commercial RuL2(NCS)2 (Ruthenium 535) and eosin Y dyes (Aldrich) were used. The adsorption of the RuL2(NCS)2 dye was done immediately after high-temperature firing in order to avoid rehydration of the porous thin films. The ITO/SnO2:TiO2 film electrodes were dipped into a 3 ×10-4 M RuL2(NCS)2 dye solution while it was 800 and the immersed electrode was refluxed at 800 for 1 h. After the adsorption treatment, the film electrode was dried by a stream of argon. The film electrodes were also dipped into a 3.2 ×104 M ethanol solution of eosin Y dye and the immersed electrode was refluxed at 800 for 30min.

2.3. Photoelectrochemical Measurements and Analyses Three electrode cells were used, comprising of a platinum wire counter electrode, an Ag/AgCl reference electrode, and a RuL2(NCS)2 dye-adsorbed nanocrystalline SnO2:TiO2 working electrode. A solution of 0.03 M I2 and 0.3 M LiI in acetonitrile was used as an electrolyte. A potentiostat (Toho Technical Research, 2020) and a programmed function generator were used to measure the photoelectrochemical response of the solar cell. The eosin Y and RuL2(NCS)2 dyes-adsorbed semiconductor film electrode was illuminated through the

222

Weon-Pil Tai

conductive glass support using a 500W xenon lamp served as a light source. The intensity of the illumination source was measured using a power meter. UV-visible absorption spectra of the eosin Y and RuL2(NCS)2 dyes onto semiconductor film were recorded by using spectrophotometer (V-550, Jasco) before and after the adsorption of the dye by refluxing at 800 for 30 min or 1 h. Surface area was measured by the nitrogen adsorption method using a Autosorb-1 (Qunta Chrome). The microstructure of the films was observed by scanning electron microscope (SEM, Hitachi S-4200). The thickness of the films was evaluated from SEM images of the cross section of the films.

3. RESULTS AND DISCUSSION 3.1. Characterization of Film Electrodes Figure 1 shows the SEM micrographs of the surfaces in the SnO2/TiO2 bilayered film and TiO2-5wt% SnO2 composite film. The films possess nanocrystalline and nanoporous structure, which is composed of interconnected nano-sized grains. It indicates that such nanostructured films can adsorb easily the dyes. The P25 TiO2 with larger nano-sized grain in the SnO2:TiO2 films was added 5 wt% to TiO2 colloidal solution to supply the photon scattering center of light. The phases formed in the SnO2/TiO2 bilayered films were anatase TiO2, containing a small amount of SnO2. The phases in the TiO2-SnO2 composite films were anatase TiO2 and SnO2. Further, no second phases formed in the TiO2:SnO2 films except TiO2 and SnO2.

Figure 1. SEM micrographs for samples : (a) SnO2/TiO2 bilayered film and (b) TiO2-5wt% SnO2 composite film.

Dye-Sensitized Nano SnO2:TiO2 Solar Cells

223

Figure 2 shows the absorption spectra of SnO2/TiO2 bilayered film electrodes. The film electrode without the dye absorbs UV below 400 nm (figure 2(a)). The ITO/3.5μm SnO2/7μm TiO2 bilayered electrode adsorbed with an eosin Y exhibits absorption over the wide range including the visible light (figure 2(b)). The absorption peak of eosin Y dye on SnO2/TiO2 (525 nm) was shifted towards blue region compared to that of the dye in the ethanol (534nm, figure 2(c)). Redmond et al. [15] reported that the blue shift in the ZnO adsorbed with a Ru dye was due to a change in the dielectric constant at the semiconductor electrode-electrolyte solution interface, dye molecules adsorbed as aggregates or oligomers, and dye chemisorption onto specific sites of the ZnO surface. In this study, the blue shift is attributed to chemisorption of eosin Y dye onto specific sites of the nanoporous SnO2/TiO2 film surface.

Figure 2. Absorption spectra of (a) ITO/SnO2/TiO2 bilayered electrode, (b) ITO/3.5μm SnO2/7μm TiO2/eosin Y bilayered electrode, and (c) eosin Y in ethanol.

Figure 3 shows the UV-visible absorption spectra of the SnO2/TiO2 bilayered film electrodes sensitized by RuL2(NCS)2 dye. The film electrode without the dye absorbs UV below 400 nm (figure 3(a)). The ITO/SnO2/TiO2/RuL2(NCS)2 electrodes exhibit absorption over the wide range including the visible light (figure 3(b and c)). It insures efficient photon capture by a RuL2(NCS)2 dye in the visible spectral range. The thicker TiO2 film on the SnO2 film (figure 3(c)) results in higher absorption peak. Figure 3(d) shows the absorption spectrum of the 3.5μm SnO2/7μm TiO2 bilayered electrode by a commercial RuL2(NCS)2 dye. It has wider absorption range in the visible range than homemade RuL2(NCS)2 dyesensitized SnO2/TiO2 bilayered electrode. The absorption spectra of the TiO2-SnO2 composite film electrodes are shown in figure 4. The film electrode without the dye exhibits absorption of UV light below 400 nm (figure 4(a)). The increase of absorption near 700 nm wavelength is due to the transparent and porous TiO2-SnO2 composite films even though the thickness of the film is about 10μm. The ITO/TiO2-SnO2/RuL2(NCS)2 electrodes exhibit absorption in the wide range including the visible light range (figure 4(b and c)). The composite film electrode with 95 wt% TiO2 content, 5 wt% SnO2 addition (figure 4(c)) exhibits higher absorption peak than the film electrode with 80 wt% TiO2 content, 20 wt% SnO2 addition (figure 4(b)). The intensity of absorption peak for the composite film electrode with 60 wt% TiO2 content is similar to that for the film electrode with 80 wt% TiO2 content.

224

Weon-Pil Tai

Figure 3. Absorption spectra of (a) ITO/3.5μm SnO2/7μm TiO2, (b) ITO/3.5μm SnO2/3.5μm TiO2/RuL2(NCS)2, (c) ITO/3.5μm SnO2/7μm TiO2/RuL2 (NCS)2, and (d) ITO/3.5μm SnO2/7μm TiO2/commercial RuL2(NCS)2 bilayered electrodes.

3.2. Photoelectrochemical Properties of Dye-Sensitized Solar Cells Figure 5 shows the incident monochromatic photon-to-current conversion efficiency (IPCE) of ITO/SnO2/TiO2/eosin Y bilayered cells as function of excitation wavelength and TiO2 film thickness at a constant SnO2 film thickness. Figure 5(a) exhibits SnO2 of 3.5 μm and TiO2 of 3.5 μm in film thickness. Figure 5(b) exhibits SnO2 of 3.5 μm and TiO2 of 7 μm in film thickness. Moreover, figure 5(c) shows SnO2 of 3.5 μm and TiO2 of 11 μm. The IPCE is defined as the number of electrons generated by light in the external circuit divided by the number of incident photons, as follows:

Figure 4. Absorption spectra of (a) ITO/TiO2-20wt% SnO2, (b) ITO/TiO2-20wt% SnO2 /RuL2(NCS)2, and (c) ITO/TiO2-5wt% SnO2/RuL2(NCS)2 composite electrodes.

Dye-Sensitized Nano SnO2:TiO2 Solar Cells

225

70 60

(b )

IPCE / %

50 (a )

40 30

(c )

20 10 0 400

450

500

550

600

650

700

W a v e le n g t h / n m

Figure 5. Photocurrent action spectra of (a) ITO/3.5μm SnO2/3.5μm TiO2/eosin Y, (b) ITO/3.5μm SnO2/7μm TiO2/eosin Y, and (c) ITO/3.5μm SnO2/11μm TiO2/eosin Y bilayered cells.

The photocurrent density was obtained at short circuit where the SnO2/TiO2 bilayered film electrode is poised to a potential of 0.2 V measured against Ag/AgCl. The electrolyte used was a solution of 0.03 M I2 and 0.3 M LiI in acetonitrile. The SnO2/TiO2 bilayered cell was illuminated from the back face through the ITO glass support. The maximum IPCE value attains 63% at 525 nm wavelength in the 3.5 μm-thick SnO2 and 7 μm-thick TiO2 bilayered cell sensitized by an eosin Y dye, as shown in figure 5(b). The IPCE decreases with thicker and thinner TiO2 thickness at a constant SnO2 thickness. Hagfeldt et al. [16] explained an optimal colloidal film thickness. The quantum efficiency for back-face illumination exhibited a maximum value in the cell with 4μm-thick TiO2 film. The optimum film thickness in the bilayered cell sensitized by a mercurochrome dye was SnO2 of 3.5μm and TiO2 of 6 7μm [17]. RuL2(NCS)2 dye-sensitized Nb2O5 single cell showed the highest photocurrent in a 8μm thickness [10]. At enough large film thickness, the hole transfer to the electrolyte is limited, due to too slow diffusion of the hole acceptor to the inner part of the film. Figure 6 shows the IPCE of ITO/SnO2/TiO2/RuL2(NCS)2 bilayered cell as a function of wavelength. Figure 6(a) shows the IPCE of the bilayered cell with SnO2 of 3.5μm and TiO2 of 3.5 μm in the thickness of the film. Figure 6(b) exhibits SnO2 of 3.5μm and TiO2 of 7μm in the film thickness. The RuL2(NCS)2 dye of figure 6(a) and (b) was synthesized by the author. figure 6(c) shows the IPCE of the SnO2/TiO2 bilayered cell sensitized by a commercial RuL2(NCS)2 dye with the film thickness of a 3.5μm SnO2 and 7μm TiO2. A maximum IPCE value in the 3.5μm-thick SnO2 and 7μm-thick TiO2 bilayered cell sensitized by a commercial RuL2(NCS)2 dye attained 88.1% at 540 nm wavelength. The difference of the IPCE values in the cells sensitized by homemade and commercial RuL2(NCS)2 dyes in figure 6(b) and (c) seems to be due to the dye purification. The increase of TiO2 film thickness up to 7μm at a constant SnO2 thickness resulted in an increase in the value of IPCE. The shape of the photocurrent action spectra was similar to that of the absorption spectra. The SnO2 film exhibits higher transparency and conductivity than the TiO2 film, which results in a decreased electron-loss at an interface between the SnO2 and ITO layers in the ITO/SnO2/TiO2 bilayered cell. Lagemaat et al. [18] reported that the electrical potential drop at a transparent conducting oxide, TCO/TiO2 interface, occurred over a narrow region.

226

Weon-Pil Tai

100 (c) 80

IPCE / %

(b) 60

(a)

40 20 0 400

450

500

550

600

650

700

Wavelength / nm Figure 6. Photocurrent action spectra of (a) ITO/3.5μm SnO2/3.5μm TiO2/RuL2(NCS)2, (b) ITO/3.5μm SnO2/7μm TiO2/RuL2(NCS)2, and (C) ITO/3.5μm SnO2/ 7μm TiO2/commercial RuL2(NCS)2 bilayered cells.

The IPCEs of ITO/TiO2-SnO2/RuL2(NCS)2 composite cells as a function of wavelength are shown in figure 7. The IPCE value of TiO2-5 wt% SnO2 composite cell is higher than those of TiO2-20 wt% SnO2 and TiO2-40 wt% SnO2 composite cells in the visible range. It means that RuL2(NCS)2 dye molecules adsorb well onto the TiO2 particle surface of the composite film with higher TiO2 content, and charge recombination occurs in the composite cell of higher SnO2 content. The photocurrent flow in the TiO2-SnO2 composite cell was unstable at the cell with higher SnO2 content. The IPCE below 450 nm exhibits relatively high value due to weak photon flux despite the low photocurrent. The SnO2/TiO2 bilayered system shows higher IPCE values than the TiO2-SnO2 composite system in using two oxide semiconductors with different energy levels, different conduction band edge energy positions. In the composite system, the charge recombination between the electron injected into the conduction band of the semiconductor and the oxidized sensitizer is not suppressed because the electrons injected from the excited RuL2(NCS)2 dye (E°= -0.66V vs NHE, in acetonitrile) [10] into the conduction band of SnO2 (ECB = 0V vs NHE) [7] can not migrate to the higher lying conduction band of TiO2 (ECB = -0.5,V vs NHE) [11]. In the bilayered system, however, the charge recombination is suppressed by fast electron transfer process between two oxide semiconductors with different energy levels. In other words, the electrons injected from the excited RuL2(NCS)2 dye molecule into the conduction band of TiO2 are transferred quickly into the lower lying conduction band of SnO2. It means that a better charge separation in the SnO2/TiO2/RuL2(NCS)2 bilayered system is enhanced by fast electron transfer process from excited RuL2(NCS)2 dye to SnO2. The charge recombination is suppressed more in the bilayered system than in the composite system.

Dye-Sensitized Nano SnO2:TiO2 Solar Cells

227

60 (a)

50 (c)

IPCE / %

40

(b)

30 20 10 0 400

450

500

550

600

650

700

Wavelength / nm

Figure 7. Photocurrent action spectra of (a) ITO/TiO2-5wt% SnO2/RuL2(NCS)2, (b) ITO/ TiO2-20wt% SnO2/RuL2(NCS)2, and (c) ITO/TiO2-40wt% SnO2/RuL2(NCS)2 composite cells.

Nasr et al. [8] reported that the rate of back electron transfer in the SnO2/TiO2 bilayered (coupled) system decreased 3-5 times compared to those in single SnO2 and TiO2 systems, which supports an improved charge separation in the bilayered system. In another study, it has also been reported that the maximum IPCE value of 78.3% at 530 nm was attained in the single-layered TiO2 cell sensitized by RuL2(NCS)2 dye [9], which is slightly lower than that of SnO2/TiO2 bilayered cell. The SnO2 layer between ITO and TiO2 in the bilayered system provides a driving force to pull electrons away from TiO2 layer and away from holes to improve charge separation. Thus, charge separation in the SnO2/TiO2 bilayered cell resulted in lower charge recombination losses and leading to an increased IPCE. An increase in the photocurrent was achieved by reflux treatment of the semiconductor film in a dye solution [19], which was attributed to the increase in the ester-like linkage between the RuL2(NCS)2 dye and TiO2 film. It was also noted that the increase in the photocurrent is due to an increase in the ratio of ester-like linkage to chelating linkage. The binding state between the RuL2(NCS)2 dye and different types of semiconductors was studied by Sayama et al. [10] using FTIR spectroscopy. The absorption at about 2100 cm-1 is the SCN ligand stretching modes [19]. The absorption at about 1730 cm-1 is related to the carboxyl group C=O stretching band, indicating an ester-like linkage between the dye molecules and the semiconductor surface; whereas the absorption at about 1605cm-1 is assigned to the carboxyl group O-C-O asymmetrical stretching, indicating the interaction through chelating or bridging modes to the semiconductor surface [17, 20]. The RuL2(NCS)2 dye adheres to semiconductor surface through an ester linkage, and electrons are transferred mainly through the conjugated orbitals of the ester linkage and the semiconductor conduction band. The TiO2 conduction band consists of a Ti3d orbital and that of SnO2 consists of a Sn5s orbital. The orbital overlap between the conduction band and the ester linkage means that electron transfer from the RuL2(NCS)2 dye to the TiO2 d-orbital is more effective than that from the

228

Weon-Pil Tai

RuL2(NCS)2 dye to the SnO2 s-orbital due to the orbital shape [10]. The one cause of higher IPCE value in the SnO2/TiO2 bilayered cell compared with the TiO2-SnO2 composite cell is the stronger ester-like linkage between the TiO2 surface layer and the RuL2(NCS)2 dye molecules by reflux treatment of the SnO2/TiO2 film in the RuL2(NCS)2 dye solution.

4. CONCLUSION Nanostructured SnO2:TiO2 thin films were prepared by the sol-gel process and the photoelectrochemical properties of the SnO2:TiO2 bilayered and composite solar cells sensitized by eosin Y and RuL2(NCS)2 dyes were studied. The SnO2/TiO2 bilayered cell showed higher IPCE values than the TiO2-SnO2 composite cell in using two semiconductors with different energy levels. A maximum IPCE value of 88.1% was reached at 540 nm in the 3.5μm-thick SnO2 and 7μm-thick TiO2 bilayered cell, which has been attributed to a better charge separation by fast electron transfer process between the constituent layers. The dyesensitized SnO2/TiO2 bilayered cell could be utilized to fabricate a low-cost solar cell.

REFERENCES [1] [2]

[3] [4]

[5]

[6]

[7]

[8]

O'Regan, B., and Gratzel, M. (1991) A Low-Cost, High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO2 Films, Nature, 353, 737-739. Nazeeruddin, M.K., Kay, A., Rodicio, I., Humphry-Baker, R., Muller, E., Liska, P., Vlachopoulos, N., and Gratzel, M. (1993) Conversion of Light to Electricity by cisX2Bis(2,2'-bipyridyl-4,4'-dicarboxylate) ruthenium(II) Charge-Transfer Sensitizers (X=Cl-, Br-, I-, CN-, and SCN-) on Nanocrystalline TiO2 Electrodes, J. Am. Chem. Soc., 115, 6382-6390. Hagfeldt, A., and Gratzel, M. (1995) Light-Induced Redox Reactions in Nanocrystalline Systems, Chem.Rev., 95, 49-68. Burnside, S.D., Brooks, K., McEvoy, A.J., and Gratzel, M. (1998) Molecular Photovoltaics and Nanocrystalline Junctions, Advanced Materials in Switzerland, Chimia. 52, 557-560. Lindstrom, H., Magnusson, E., Holmberg, A., Sodergren, S., Lindquist, S.-E., and Hagfeldt, A. (2002) A New Method for Manufacturing Nanostructured Electrodes on Glass Substrates, Solar Energy Materials and Soalr Cells, 73, 91-101. Hotchandani, S., and Kamat, P.V. (1992) Charge-Transfer Processes in Coupled Semiconductor Systems. Photochemistry and Photoelectrochemistry of the Colloidal CdS-ZnO system, J. Phys. Chem., 96, 6834-6839. Nasr, C., Hotchandani, S., Kim, W.Y., Schmehl, R.H., and Kamat, P.V. (1997) Photoelectrochemistry of Composite Semiconductor Thin Films. Photosensitization of SnO2/CdS Coupled Nanocrystallites with a Ruthenium Polypridyl Complex, J. Phys. Chem. B, 101, 7480-7487. Nasr, C., Kamat, P.V., and Hotchandani, S. (1998) Photoelectrochemistry of Composite Semiconductor Thin Films. Photosensitization of the SnO2/TiO2 Coupled System with a Ruthenium Polypyridyl Complex, J. Phys. Chem. B, 102, 10047-10056.

Dye-Sensitized Nano SnO2:TiO2 Solar Cells [9] [10]

[11]

[12]

[13] [14]

[15]

[16]

[17] [18]

[19]

[20]

229

Tai, W.-P., Inoue, K., and Oh, J.-H. (2002) Ruthenium Dye-Sensitized SnO2/TiO2 Coupled Solar Cell, Solar Energy Materials and Soalr Cells, 71, 553-557. Sayama, K., Sugihara, H., and Arakawa, H. (1998) Photoelectrochemical Properties of a Porous Nb2O5 Electrode Sensitized by a Ruthenium Dye, Chem. Mater., 10, 38253832. Fessenden, R.W. and Kamat, P.V. (1995) Rate Constants for Charge Injection from Excited Sensitizer into SnO2, ZnO, and TiO2 Semiconductor Nanostallites, J. Phys. Chem., 99, 12902-12906. Tennakone, K., Kumara, G.R.R.A., Kottegoda, I.R.M., and Perera, V.P.S. (1999) An Efficient Dye-Sensitized Photoelectrochemical Solar Cell made from Oxides of Tin and Zinc, Chem. Commun., 15-16. Kiesewetter, T., Tomm, Y., Turrion, M., and Tributsch, H. (1999) Composite Materials for Photovoltaics: A Realistic Aim?, Solar Energy Mat. and Solar cells, 59, 309-323. Liska, P., Vlachopoulos, N., Nazeeruddin, M.K., Comte, P., and Gratzel, M. (1988) cisDiaquabis(2,2'-bipyridyl-4,4'-dicarboxylate)-ruthenium(II) Sensitizes Wide Band Gap Oxide Semiconductors Very Efficiently over a Broad Spectral Range in the Visible, J. Am. Chemi. Soc., 110, 3686-3687. Redmond, G., Fitzmaurice, D., and Graetzel, M. (1994) Visible Light Sensitization by cis-Bis(thiocyanato) bis(2,2'-bipyridyl-4,4'-dicarboxylato) Ruthenium(II) of a Transparent Nanocrystalline ZnO Film Prepared by Sol-Gel Technique, Chem. Mater., 6, 686-691. Hagfeldt, A., Bjorksten, U., and Lindquist, S.-E. (1992) Photoelectrochemical Studies of Colloidal TiO2-Films: the Charge Separation Process Studied by Means of Action Spectra un the UV Region, Solar Energy Mat. and Solar Cells, 27, 293-304. Tai, W.-P. (2001) Photoelectrochemical Properties of SnO2/TiO2 Coupled Electrode Sensitized by a Mercurochrome Dye, Materials Lett., 51, 451-454. van de Lagemaat, J., Park, N.-G. and Frank, A.J. (2000) Influence of Electrical Potential Distribution, Charge Transport, and Recombination on the Photopontial and Photocurrent Conversion Efficiency of Dye-Sensitized Nanocrystalline TiO2 Solar Cells: A Study by Electrical Impedance and Optical Modulation Techniques, J. Phys. Chem. B 104, 2044-2052. Murakoshi, K., Kano, G., Wada, Y., Yanagida, S., Miyazaki, H., Matsumoto, M., and Murasawa, S. (1995) Importance of Binding States between Photosensitizing Molecules and the TiO2 Surface for Efficiency in a Dye-Sensitized Solar Cell, J. Electroanal. Chem., 396, 27-34. Meyer, T.J., Meyer, G.J., Pfenning, B.W., Schoonover, J.R. Timpson, C.J., Wall, J.F., Kobusch, C., Chen, X., Peek, B.M., Wall, C.G., Ou, Erickson, B.W. and Bignozzi, C.A. Molecular-Level Electron Transfer and Excited State Assemblies on Surfaces of Metal Oxides and Glass, Inorg. Chem., 33, 3952.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 231-244

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 7

STRATEGIES FOR REDUCING CARBON DIOXIDE EMISSIONS - THE CASE OF BOTSWANA RURAL COMMUNITIES* C. Ketlogetswe a and T.H. Mothudi b a

Department of Mechanical Engineering, University of Botswana, Private Bag 0061 Communication and Study Skill Unit, University of Botswana, Private Bag 0061

b

ABSTRACT The International Community’s pre-occupation with the ever-escalating dangers posed by gaseous pollutants need not be overemphasized. Suffice to mention, however that the magnitude of the dire negativity of pollutants is reflected in the numerous international charters that were promulgated with a common objective to sensitise the world about the need to move toward setting up minimum permissible levels of emission for activities whose execution result in atmospheric pollution. In addition, authorities have also gone so far as to offer incentives / motivation as a means of assuaging nations towards implementing various strategies for minimising atmospheric pollutions. This paper explicates efforts taken by. The government of Botswana in an effort to strive for compliance with international protocols and standards to safeguard against deterioration of the planet. Focus will specifically be paid to examining any concrete measures taken with the view to curb the negative impacts of carbon dioxide gas. The suitability and sustenance or, otherwise, of government projects envisaged for reducing carbon dioxide emission levels generated during the combustion of fuelwood and other related energy sources used by rural communities in Botswana will also be discussed.

*

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

232

C. Ketlogetswe and T.H. Mothudi

1. INTRODUCTION The adverse impact of climate change is a matter of grave concern, the world-over. As a result, increased pressure from the United Nations Framework Convention on Climate Change (UN FCCC) to reduce the world’s energy-related greenhouse emission levels particularly carbon dioxide gas is expected to stimulate rapid development of renewable energy facilities in many countries. The authors also believe that pressure to preserve environmental purity is poised to exert immense challenges to energy resources conservation strategies in many developing countries, Botswana included. As testimony to the global nature of this concern, the UK government predicts that by 2010, 10% of the country’s electricity sales should be generated from resources which are eligible for the renewable obligation including waste-to-energy. This figure compares favourably with the entire European Community’s target of 15% by the same period (http:defra March 2003). The Kyoto Protocol to the UN FCCC requires that both the Organisation for Economic Cooperation and Development (OECD) and Non-OECD countries reduce the emissions of greenhouse gas especially carbon dioxide to permissible minimums (The Kyoto Protocol and Beyond (1999) and Porteous (2001)). The emphasis on carbon dioxide is driven by the fact that combustion of hydrocarbons particularly fossil fuels, which are a major global energy resource, has hugely increased the total load of carbon dioxide in the atmosphere. The longterm consequences of this includes alteration of global climate as greenhouse gases trap heat at the earth’s surface. The Kyoto Protocol is designed to ensure that communities across the world begin a long-term shift to a future with lower emissions and more efficient energy sources. The protocol also grants countries opportunities to work out their modalities for reducing emissions. However, it specifies limits for the carbon dioxide emission levels produced by a country or a group of countries. As a result, the target reductions vary from nation to nation. For example, the Protocol requires developed nations to achieve a combined minimum reduction of 5% of the combined collective emissions from the 1990 baseline levels. There are several techniques and measures used world-over to address the problem posed by the release of carbon dioxide gas into the atmosphere. One such technique is geological storage, which is rapidly receiving widespread recognition among countries of the world. This method employs three storage options, namely gas and oil reservoirs which are no longer productive, coal seams which cannot be further mined and deep aquifers. Available evidence indicates that when carbon dioxide is injected in gas and oil reservoirs which are nearing depletion, the injection process enhances the production process by enabling the natural gas trapped in these structures to be released. As a result, the process considerably reduces the capture and storage costs of carbon dioxide. This factor leads to the conclusion that there will be a corresponding increase in the economic case for geological storage of carbon dioxide gas. Research on this approach is on-going. Some of the countries involved in the existing research efforts are those that subscribe to a more technology-oriented approach in their national policy on climate changes. These countries include, the United States, Canada, Japan, Norway, the Netherlands, Australia and some European Union Member States. The clamour for low-emissions, more efficient energy resources and the use of the geological techniques is a step towards sustainable development.

Strategies for Reducing Carbon Dioxide Emissions

233

This paper, as mentioned earlier, discusses efforts taken by the government of Botswana in a bid to comply with international protocols and requirements to reduce carbon dioxide emission levels to be emitted into the atmosphere. Most importantly this paper seeks to examine the present strategies designed for rural communities with the view to consider their feasibility and sustainability as described in sections 3, 4 and 5. Prior to engaging into a comprehensive review of the proposed strategies for rural communities, it is pertinent to consider the uses to which energy resources are placed in Botswana particularly in the context of rural communities. A detailed exploration of this matter is undertaken in Section 2.

2. RURAL COMMUNITIES AND HOUSEHOLDS EXPENDITURE ON ENERGY RESOURCES Rural communities account for more than 70% of Botswana’s population (Government of Botswana Statistics 2001). The majority of these communities are heavily dependent on traditional agriculture for their livelihood. This situation appears to be a common feature of most developing countries particularly those situated in the Southern African Development Countries (SADC) region. For example, estimates indicate that 67% of Namibia’s population live in a rural area, mostly in the northern regions of the country where rainfall is higher (Government of Namibian Statistics 1996). In Botswana, estimates indicate that there are approximately 160 000 rural household populations with an average monthly income below Botswana Pula (BWP) 530 per month, which is equivalent to US$ 106 (Botswana government statistics 2001). The income generated by these communities is used to cover a household’s expenditure including energy supply which is of particular interest to the present study. A study by the Japan International Cooperation Agency (Japan International Cooperation Agency report, 2003) on the households expenditure by rural communities in Botswana found that approximately 85% of homeowners in rural communities spend up to a maximum of BPW 50 (US$ 10) per mouth on energy provision (Japan International Cooperation Agency report, 2003). Figure 1 illustrates the overall expenditure levels on energy resources for the rural communities in Botswana. The data in figure 1 shows that less than 10% of the rural communities in Botswana spend approximately BPW 100 per month on energy resources, which is equivalent to US$ 20 per month. The percentage falls below 3% in respect of those house owners spending BPW 150 or more per month on energy resources. Overall, the data suggests that affordability window for energy resources for the majority of rural communities in Botswana range from PBW 0 to 50 (US$ 10). On the basis of the above observations, it is obvious that an investigation on sustainability of the proposed projects for rural dwellers aimed at reducing energy-related carbon dioxide emission levels generated by the rural community in Botswana is of considerable practical interest. It should be explained that the data presented in figure 1 was collected from nine rural villages, located in various parts of the country. Consequently, therefore, it is assumed that such data reflects the practical situation found in the rest of the rural communities in Botswana.

234

C. Ketlogetswe and T.H. Mothudi 90 80

% households

70 60 50 40 30 20 10 0 0

50

100

150

200

250

300

350

Expenditure (BPW / month)

Source: Japanese International Cooperation Agency Report (2003). Figure 1. Typical rural community expenditure levels on energy resources in Botswana.

Regarding climate changes, Botswana contributes approximately 7% of Africa’s total greenhouse gas emissions while the entire continent of Africa is estimated to contribute 5% of the global total emission (Government of Botswana framework on climate change report 2001). Although this study places more emphasis on proposed strategies for reducing energyrelated carbon dioxide emission levels generated by rural community in Botswana, it is pertinent also to highlight energy supply and demand in various sectors. This is considered a vital means of linking the supply and demand to the generation of gaseous emissions. In line with this understanding, figures 2 (a) and (b) show net energy supply and demand in the country on sectoral basis. The results in figures 2 (a) and (b) should be viewed in parallel. The data in figure 2 (a) shows that fuelwood is the second major source of energy supply after petroleum. Regarding demand, the data in figure 2 (b) shows that the residential sector is the major energy consumer in Botswana. The high rate of fuelwood supply as shown in figure 2 (a) indicates that fuelwood is the major energy source for rural communities in Botswana and perhaps the least expensive. In many cases, it is harvested at no monetary cost as a common property. Figure 3 shows an increased supply with time of two major energy sources used by rural communities in Botswana. With respect to the reduction of energy-related gaseous emission levels generated by rural communities, the data in figures 2 and 3 suggest that emphasis ought to be placed on the reduction in consumption rate of fuelwood which is mainly used for cooking in open fire, space and water heating. It should be noted that Botswana with ‘estimated population of 1.7 million recorded woodfuel consumption of about 1.5 million tonnes in 1999’ (http:worldenergy;statistical Bulletin, 2001). Based on this observation, and particularly on the results in figure 3, it can be assumed that the carbon dioxide emission levels from the combustion of woodfuel are relatively high.

Strategies for Reducing Carbon Dioxide Emissions

235

Since the proposed strategies for reducing energy-related carbon dioxide gas generated by rural communities do not include reduction in the use of fuelwood, the utilisation of this approach is likely to conjure several questions especially those relating to the techniques for using proposed strategies and their ability to significantly reduce gaseous emissions generated by these communities. Section 3 briefly discusses some of the government strategies aimed at addressing the problem of climate change.

45

% contribution

40 35 30 25 20 15 10 5 0 Petrolium

Electricity

fuelwood

Coal

Sources of energy supply

Ot he r

Ag ric ul tur e

Tr an sp or t

In du st ry

45 40 35 30 25 20 15 10 5 0 Re si d en ti a l

% energy demand

a.

Energy demand by sector b. Source: Botswana energy statistics report 2000. Figures 2 (a) and (b). Botswana’s sources of energy supply and demand on sectorial basis.

236

C. Ketlogetswe and T.H. Mothudi 25000 22500

woodfuel paraffin & LPG gas

20000 Energy (Terajoules)

17500 15000 12500 10000 7500 5000 2500 0 1981

1983

1985

1987

1989

1991

1993

1995

1997

Time (years)

Source: Government of Botswana (2001) Report No: TB 10/1/9/99/2000. Figure 3. Uses of various energy sources in a typical rural household

3. CENTRALISED PV BATTERY CHARGING STATION IN BOTSWANA There is scope for action to reduce energy-related carbon dioxide emission levels generated by the rural communities in Botswana. This opportunity is exemplified by the introduction of a number of solar energy projects including centralised a photovoltaic battery charging station at Lorolwana village, hitherto, the country’s sole centralised photovoltaic battery station. The village is located approximately 200 km South of Gaborone City and comprises of a population size and households number of 952 and 180 respectively (Government of Botswana statistics 2001). The facility was constructed in 2003, through funding by the Global Environmental Facility (Energy Affairs Division annual report 2004). The main goals of the project was to provide lighting and also to operate radios. Although adding more solar panels could easily increase the size of such system, the present design is such that its capacity is approximately 1.1 kWp. Figure 4 presents a photograph of the facility. The batteries demonstrated in figure 4 have capacities ranging from 25 to 50Ah and are supplied with an electronic unit to prevent undercharging which is one of the most common causes of plate buckling, due to the plate strain caused by the lead sulphate. In addition to two 12V DC sockets designed for lighting, these batteries incorporate 9V DC sockets for appliances such as radio and black and white television sets.

Strategies for Reducing Carbon Dioxide Emissions

237

Figure 4. Exterior and interior views of the battery charging station at Lorolwana village.

In order to assess the socio-economic merits of the facility shown in figure 4, 42 households were reviewed as part of a pilot study of the technology. The total number of households reviewed as part of the pilot project was determined taking into consideration that the station was going to be supplied with 84 batteries, and also that each household was to be allocated two batteries to ensure availability of a permanent standby battery whilst another is being used. Further, the same was promised on the understanding that in the event socioeconomic factors indicated possible sustenance of the project, the facility will be expanded to provide the entire community of Lorolwana. Furthermore, it was envisaged that upon confirmation of the sustenance of the project, it will ultimately be handed over to the community to take full charge of its running. The overall aim was that the concept could be replicated in other parts of the country and the data collected during the implementation phase would serve as the basis for formulating policies on renewable energy in Botswana particularly among rural communities. Although there are some indications for the successful introduction and utilisation of solar technologies in rural communities, the present study revealed that the facility demonstrated in figure 4 is currently dysfunctional. To pave way for an ensuing discussion, it is pertinent to point out that participants were expected to pay for service charges which included fixed and battery charging coupons which expire during the last day of each month. The overall charge was BPW pula 20 which is equivalent to US$ 4. Based on the results in figure 1, it is clear that the overall charge was within the affordable window as discussed in Section 2. Notwithstanding these conclusions, however, evidence from interviews with facility operator at community level revealed that majority of the selected house owners defaulted from paying the fixed charges and burying battery-charging coupons at the introductory stage of the project alleging high service costs and inflexibility in the payment method. It is perhaps; appropriate to observe that charging coupons are an unsuitable method of payment for most rural communities in Botswana. This confirmation is largely predicated on the fact that majority of this population sector are generally dependent on traditional agriculture and pastoralism as discussed earlier in Section 2. Consequently, the majority of households in rural communities have three settlements, at the same time, thus lands, cattle post and village status, [further making the payment system of charging coupons with a fixed expiring date unsuitable]. This is largely due to movement of participants between these three

238

C. Ketlogetswe and T.H. Mothudi

places. It should also be noted that although the overall charge may appear to be within affordable range as discussed earlier in the same section, lack of economic activities in most rural communities appears to contribute to low development of such projects at rural communities levels. Second, the study had also revealed that most components particularly batteries needed rapid replacement because they were unable to store photovoltaic generated electrical energy. Lack of periodic maintenance checks on the overall facility by the management to ensure efficient functioning of equipment as well as provide further user guidance appear to have been the major causes for the present condition in which the facility is non-functioning. This point is reinforced by observations from interviewees (mainly those at management level from Energy Affairs Division) who revealed that management is seeking funds for replacing old batteries at Lorolwana centralised battery charging station (Mr. Sethare)1. Although there are no periodic maintenance checks as mentioned above, it should also be noted that Botswana Power Corporation (the only power generation and distribution company within the country) was sanctioned to monitor the design and installation of the project. It is believed that Botswana Power Corporation was bestowed this mandate because it is the sole government owned company in the specialist area. However, the contract in respect of the Lorolwana project appear not to have included periodic maintenance service, as well. On the basis that the facility demonstrated in figure 4 is currently not functioning, this invites several questions primarily relating to its viability as a mechanism for the reduction of energy-related carbon dioxide emission levels generated by rural communities in Botswana. The current situation at Lorolwana leads to the conclusion that government ambition to replicate the same project in other parts of the country aimed to strive for compliance with international protocols on global climate change is till a far cry. It should be stressed that problems relating to failure to put in place periodic maintenance checks as mentioned in the previous page is likely to bring to surface the existing public perception about poor quality of solar equipment in Botswana, further making the rapid development of solar industry in rural communities difficult. The centralised battery charging system is not the only renewable technology, which is facing resistance from the public, among the major ones being those, highlighted in Section 4, below.

4. MANYANA PHOTOVOLTAIC PILOT PROJECT This section considers the status of photovoltaic facilities at Manyana village. The project at Manyana was selected to constitute the present study on consideration of its status of being among the pioneer project that the Energy Affairs Division had embarked upon during its 8 years of existence. The establishment of Energy Affair Division was to effectively provide policy direction on issues pertaining to energy in order to promote rapid and effective development of National Electricity Grid infrastructure network particularly among in rural communities in Botswana. The findings from the present study would be used to make a case as to whether renewable technologies including photovoltaic techniques could significantly reduce gaseous emission levels generated by rural communities in the country. The village is located 50 km 1

D.Sethare, Renewable Energy Officer

Strategies for Reducing Carbon Dioxide Emissions

239

from Gaborone city. The Manyana photovoltaic project started in 1992. Like the project mentioned in Section 3, the major objective of this project was to assess viability and sustainability of solar energy technology as an alternative source of energy especially for rural based communities. Depending on the findings, such project was also intended to be replicated in other parts of the country. It is pertinent to explain that it is government procedure that after two to three years of project inception, the project changes from pilot to commercial status. As a direct result, the Manyana photovoltaic project went through the above-mentioned processes in 1995 under the auspices of Rural Industrial Innovation Center. At its launch, 42 households were selected to form part of the piloting phase of the technology. To facilitate the implementation processes, government offered to the 42 selected households to purchase photovoltaic systems. The loan scheme was payable over a 4 years period. These loans ranged between BWP 4000 to 20,000 (Porter 1994 in BOT/00/G41/A/1G/99), which is equivalent to between US$ 800 to 4000. The actually costs of the system is influenced by the size of the system. The present study revealed that an evaluation report on the viability and sustainability of Manyana project was carried out in 1994, thus approximately 2 years after the implementation phase (Porter 1994 in BOT/00/G41/A/1G/99). This study has found out that all 42 selected households paid off their loans and were satisfied with the performance of their systems. In contrast to the above findings by Porter (1994), the present study revealed that 83% of the selected households disconnected photovoltaic systems from their houses. From interviews with the majority of the 42 Clients, it had been alleged that the exorbitant costs for replacing components, especially the battery proved to be the main rational for disconnections. The study confirmed that the cost in question is approximately BPW 650 which is equivalent to US$ 150. Regarding the high percentage of disconnecting photovoltaic system by household at Manyana, it should be noted that the present study also confirmed the observation made in Section 3 that donor projects often offer these technologies at heavily subsidised consumer price at the period of project inception. As a direct result, most of such projects become unsustainable upon handover of control to local communities or private enterprise. On the basis of the finding by Porter (1994 in BOT/00/G41/A/1G/99), and the fact that approximately 83% of households at Manyana have disconnected the system, it becomes clear that the evaluation process within 2 years of project launch is too soon to indicate reliable data on viability and sustainability of such projects. On the basis of this evidence, as well as any others such as increasing use of fuelwood consumption as shown in figure 3, it becomes apparent that the use of renewable technologies in rural communities in Botswana does not offer significant potential to contribute to the reduction of energy-related gaseous emission levels. This observation is based on the fact that efforts at reducing energy-related carbon dioxide emission levels have placed more emphasis on the provision of alternative energy system than on intervention techniques aimed at reducing the use of woodfuel from 1.5 million tonnes as discussed in Section 2. As discussed earlier in the same section woodfuel in rural communities in Botswana is used for cooking in open fires, it should be expected that the combustion of 1.5 million tonnes of woodfuel would generate significant gaseous emissions. The major challenge facing government is to put in place new and more effective measures or augment the existing main measures aimed at reducing the use of woodfuel in rural communities. The utilisation of this approach is likely to bring closer to reality the

240

C. Ketlogetswe and T.H. Mothudi

Botswana government’s effort to strive for compliance with international protocols and requirement to safeguard deterioration of the planet. Prior to the discussion on the proposed approach on the reduction of combustion emission levels in rural areas in Botswana, it is pertinent to briefly discuss the status of photovoltaic mini-grid system in Motshegaletau village as described in Section 5, below. The findings would be used in the present study to make a case for whether renewable technologies including photovoltaic min-grid systems could play a major role in reducing gaseous emission levels generated by the rural communities in Botswana.

5. CURRENT STATUS OF MOTSHEGALETAU PV MINI-GRID SYSTEM Todate, Botswana possesses only one centralised photovoltaic power station, located at Motshegaletau village, approximately 50km from the nearest national electricity grid line. The village has a population of approximately 440 with approximately 88 households (Government of Botswana Statistics 2001). The system, with a capacity of 5.5kWp, started commercial operation in August 1998. The present study revealed that only 11 households have access to electricity generation from the facility. It is believed that this low access rate to electricity in Motshegaletau is explained by a number of factors including high tariffs. In fact electric tariff at Motshegaletau is 25 thebe/kWh which is the same tariff rate as charge by the Botswana Power Corporation, the only power generation and distribution company within the country. The problem of service costs as discussed in Section 3 appears to surface again here. On the basis of the above observations, and the limitation of photovoltaic electricity generation, it is believed that these could adversely affect government efforts in the battle against global warming and climate change.

6. CURRENT ACTION BY THE GOVERNMENT Preparations to replicate photovoltaic technology packages in small villages in Botswana are at advanced stages at national government and local authorities levels. The project is receiving financial assistance from the United Nation Development Programme Global Environmental Funds (UNDP-GEF). It is estimated that the project should reduce carbon dioxide emission levels by approximately 52 000 tonnes over a 20 year period (Government of Botswana report BOT/00/G41/A/1G/99). This is in the expectation that the use of photovoltaic technology will displace liquid paraffin that would otherwise have to be burned to provide light. The report estimated that on an average saving of 10 litres of liquid paraffin per month per household for selected 88 villages would be achieved. The first observation which cab be made based on the estimated saving to the value of 52 000 tonnes is that the report failed to give similar estimated value of carbon dioxide emission levels generated during the combustion of 1.5 million tonnes of woodfuel over the same period for the purpose of making comparisons. The value of 52 000 tonnes appears to an overestimation. As a subsequent discussion, it would be noted that the cost of 10 litres of paraffin in the city of Gaborone is BPW 32 which is equivalent to US$ 6.4. It can also be noted that this value is 35% more than the overall charge that residents of Lorolwana village

Strategies for Reducing Carbon Dioxide Emissions

241

(discussed in Section 2) were expected to pay as service charge. It is pertinent to mention that the price of liquid paraffin is expected to increase significantly as the location of rural communities increases relative to urban areas. On the bases of the above observations, it is of interest to conclude that the government report (Government of Botswana report BOT/00/G$1/A/1G/99) overestimated the average amount of liquid paraffin used by rural communities in Botswana. Although there is no available data on emission levels of carbon dioxide generated during the combustion of woodfuel used by rural communities to compare with the value for liquid paraffin, the authors believe that combustion of woodfuel generate relatively high gaseous emissions. On these bases, the only attractive option should be the one considered to possess potential to cause significant reduction in consumption rate of fuelwood by the rural communities in Botswana. A discussion on this matter is provided in Section 7, below.

7. PROPOSED APPROACH FOR RURAL COMMUNITIES IN BOTSWANA It is noted that between 1995 and 2002, serious educational campaign was undertaken aimed at discouraging the use of fuelwood particularly in public institutions, such as schools and prisons, and/or rehabilitation centres. The action was largely due to unsustainable harvesting of woodfuel and cutting of live trees for the purpose of drying and using them at a later stage (Energy Affairs Division annual report 2002). As a direct result, the majority of public schools particularly in cities, towns and in big villages are now using liquid petroleum gas which is supplied by local authorities. However, such opportunity is not readily available to most rural communities including areas such as Lorolwana due to lack of infrastructure. On this basis, fuelwood still remains the major source of energy for the majority of the rural communities in Botswana. The challenge to engineers and government is to provide rural communities with a facility which should reduce the consumption rate of woodfuel resource, at a cost that is relatively affordable to rural communities in Botswana. One of the facilities earmarked to promote this objective is demonstrated in figure 5. Figure 5 presents a woodstove designed by Morupule Colliery, a local coal mining company and tested by the University of Botswana. Tests conducted included the determination of thermal characteristics and the emission levels for different woodfuel species commonly used in rural communities in Botswana. The facility is in the form of two rectangular boxes (inner and outer) with detachable top cover. The inner and outer boxes are separated by an air gap of approximately 26 mm. The detachable top cover which forms part of the effective heating zone, has three circular holes of 200 mm, 150 mm and 110 mm diameter and are shown herewith as dark regions. These are closed using three circular plates consisting of recesses to ensure positive location and also to prevent the smoke from the top part of the stove from leaking. The wood stove is made of a 3 mm thick mild steel plate. The section below the fuel grate is open to the atmosphere to ensure that the combustion air enters the combustion zone under natural convection and also to facilitate the removal of ash after use. The cross sectional area of the combustion chamber is approximately 0.066 m2. The stove is internally lined with a 25 mm thick refractory material in an effort to decrease heat loss and thermal distortion of the combustion chamber particularly the grate which is made of perforated thin, mild steel plate.

242

C. Ketlogetswe and T.H. Mothudi

Removable top plate

Exhaust system

250 mm

150 mm

330 mm

1500 mm

700 mm

560 mm

300 mm

Position of smoke level meter Fuel grate

Combustion chamber

Positions of thermocouples

Figure 5. Schematic arrangement of proposed wood stove.

An experimental study was conducted to determine the thermal characteristics of the combustion process of woodfuel particles (Sickle bush) using the proposed facility as demonstrated in figure 5. The maximum charge was fixed at 1.5 Kilograms. The description of the instrumentation and operational procedures needed to determine thermal characteristics as mentioned above are described fully in Ketlogetswe (2004). For simplicity, only a sample of the results obtained in Ketlogetswe (2004) are presented in figure 6 and discussed. It should be explained that during the experimental study in Ketlogetswe (2004), an aluminium pot filled with approximately 2.5 litres of tap water was put on top of a 200 mm diameter circular plate. The time needed for the pot to boil was recorded. This was done with the knowledge that a typical rural household family may require the same quantity of water for preparing a single meal. The average boiling time was found to be 45 minutes from the start of the combustion process.

Bed fuel temperature (oC)

Strategies for Reducing Carbon Dioxide Emissions

243

800 700 600 500 400 300 200 100 0 0

5 10 15 20 25 30 35 40 45 50 55 60 65 Combustion time (min)

Figure 6. Temperature profile for the combustion process of 1.5 kg Sickle bush.

As a basis for subsequent discussions, it should be explained that the present study revealed that a typical rural household using open fire consumes approximately 4.1 tonnes/year. This compares with a value of approximately 1.6 tonnes/year required when a stove demonstrated in figure 5 is used. The difference in the rate of consumption suggests that the use of proposed wood stove by the rural communities would significantly reduce energyrelated gaseous emissions levels by approximately 61%. Based on the above observations, it is clear that one of the most attractive options with the potential to cause significant reductions in energy-related gaseous emission levels would be the use of the facility as demonstrated in figure 5. It should also be noted that the approach will address the issue of conservation of such energy resource.

8. CONCLUSION This paper has examined the status of renewable energy including woodfuel in rural communities of Botswana. In particular, the study revealed that woodfuel is the major energy resource used in the country’s rural communities. It can be concluded further that lack of network solar service centers in Botswana is the major factor hampering the development of renewable technology in the country. The implementation of the proposed project involving 88 villages does not appear to adequately address the current problem of over harvesting of woodfuel. The project under review appears to have been concerned primarily with promoting rural electrification rather than being clearly biased towards reducing energyrelated gaseous emission levels in the country. The following measures are therefore, proposed as likely to go a long way in reducing energy-related gaseous emission levels generated by rural communities in Botswana: i.

Simultaneous with encouraging the development of renewable energy-systems in rural communities, authorities should also encourage the use of a woodstove demonstrated in figure 5. Further households should be assisted financially to purchase such stoves. To stimulate public interest in the facility, demonstrations should be mounted to best highlight the advantages of using the stove under review

244

ii.

C. Ketlogetswe and T.H. Mothudi as compared to open fires. This argument is based on the fact that the use of renewable energy by rural communities will not replace the use of woodfuel for cooking. To assist rapid development and also maintain public confidence on renewable technologies particularly among rural populations, it is vital for authorities to ensure that there is framework on a network of solar service centers and service technical support. Otherwise the development of the solar industry in Botswana will never prosper.

REFERENCES [1] [2] [3]

[4] [5]

[6]

[7] [8] [9] [10] [11]

[12] [13] [14]

Botswana energy statistics. 2000. Energy Affairs Division. Ministry of minerals, energy and water resources. Energy Affairs Division annual report 2004. Ministry of minerals, energy and water resources Energy Affairs Division annual report, 2002. Fuelwood depletion. Cutting down trees for firewood, leads to environmental degradation, 2002. Ministry of minerals, energy and water resources Government of Botswana Report No [TB 10/1/9/99/2000]. Rural energy needs and requirements in Botswana, 2001. Ministry of minerals, energy and water resources Government of Botswana, report [BOT/00/G41/A/1G/99], Identifying and overcoming barriers to widespread adoption of renewable energy-based rural electrification in Botswana. Ministry of minerals, energy and water resources Government of Botswana. Initial National Communication to the United Nation Framework Convention on Climate Change, 2001. Ministry of works, transport and communication. Government of Botswana. Statistical bulletin, population of towns, village and associated localities, 2001. Government of Namibia. Statistical bulletin, household income and expenditure survey. Living conditions in Namibia, Namibia planning commission, 1996. http://www.defra.gov.uk. March 2003. http:worldenergy. Statistical Bulletin , 2001 Japan International Cooperation Agency Report. Master plan study on photovoltaic rural electrification in Botswana, 2003. Ministry of minerals, energy and water resources. Ketlogetswe. C. Thermal performance of wood particles on household stove. Eastern Africa Journal of Rural Development. 2004, 20, 72 - 80. Porteous, A. Energy from waste incineration – a state of the art emission review with an emphasis on public acceptability. Journal of Applied Energy, 2001, 70, 157 – 167. The Kyoto protocol and Beyond. Action against climate change, OECD Coll, 1999, 7 – 83.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 245-276

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 8

THE APPLYING OF COATINGS AND SURFACE THERMAL TREATMENT OF MATERIALS IN SOLAR FURNACES: THEORY AND EXPERIMENTS* V.V. Pasichny and B.A. Uryukov Institute for Problems of Material Sciences of National Academy of Sciences of Ukraine, Krzhizhanovsky str. 3, Kyiv 03142, Ukraine

ABSTRACT Solar furnaces make it possible to obtain a temperature of heating equivalent to 3500 K and above it an oxidizing air medium and without any outside contamination. They are used for investigation of materials in the Institute for Problems of Materials Science (IPMS) of theNational Academy of Science of Ukraine (NASU) for the past 40 years. The created experimental base consists of 14 different solar installations of power from 0.1 up to 10 kW. They are included in the two laboratories located in Kyiv and on the Black Sea coast. Some optical furnaces on Xe arc lamps which are the simulators of solar furnaces are added to the experimental base. In the given chapter the works of the last few years are concentrated. They are dedicated to surface heating of materials intended for obtaining coatings and improving their protective, decorative and other operational characteristics. The specialists of various fields of engineering and production are engaged in the development of these energy-intensive processes with the use of traditional energy sources. Their substitution for renewable solar radiation if it is possible can cause not only saving on utilities saving but in some cases the improvement of coatings quality due to chemical purity of the heating source. Some theoretical and experimental results of the investigation in the given field fulfilled in the IPMS are represented in the proposed work. Using an approximate integral method for solving heat conduction equation the problem is solved for the determination of the rate of thermal treatment of a surface by partial melting in a solar furnace when the sample is stationary *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

246

V.V. Pasichny and B.A. Uryukov and moves relative to the focal spot depending upon the given thickness of fused layer. Taking into account the absence hitherto of industrial (commercial) production of solar furnaces the theoretical and practical foundations have been developed in the IPMS for the creation of solar radiation concentrators on the basis of metallic antennae with plane mirror facets. As it is described in the given work the energy characteristics of these concentrators fully come up to the standards which are necessary for the realization of the greatest part of the investigated processes.

INTRODUCTION Solar energy is considered all over the world to be one of the most promising renewable sources, the intensive mastering of which can be, according to the calculations of the specialists of some countries, an alternative for traditional kinds of energy. So, the representatives of Germany at the 'Euro Sun 2004' International conference (Freiburg Germany, June 2004) have declared that in 2050 the share of renewable sources (Solar energy and wind mainly) in the total energy balance of the country will exceed 50%. Among varied ways of solar energy application the high degree of concentration of solar radiation in solar furnaces (SF) ranks high. Without going into details of the history of this orientation development which is described in many papers and monographs we shall note that the works on creating this kind of energy installation started after World War II. It was caused by intensive development of atomic power engineering, rocket and space technology and other new industries and the advent of corresponding demand in new high temperature and refractory materials and coatings. The synthesis of chemical components and also the investigation of physicochemical and technical properties and characteristics of new materials demand high temperature and specific conditions of its realization. Solar furnaces in spite of some shortcomings (dependence upon weather, time of day, season, geographical coordinates and other factors) turned out to be a very useful instrument for the investigation in the field of new materials. It is explained first of all by the opportunity to obtain high temperature of heating (up to 3800 K and above it) within an oxidizing air medium by complete absence of any contaminating impurities caused by the source of heating. Besides, practically instantaneous heat supply to an object of heating and inertialess control of heating intensity are ensured in solar furnaces. It is very important in some cases. Solar radiation is the only renewable source which permits the obtainment of high temperature of heating without intermediate conversion into electric energy. One-sided heating and strictly directed action of radiant flux can be both a shortcoming and advantage of heating in solar furnaces. In the end we shall note that the potential of solar energy depends strongly upon the location of consumer. The presence occurring everywhere in the world and even in the near-earth orbit is also and advantage of this source. The advent of plasma generators, lasers, arc Xe-lamps of high pressure and other hightemperature installations in the second half of the last-century has reduced the interest for SF because of relative high cost of optical systems, the dependence upon environment and other causes. The world energy crisis in 70-th increased the interest for renewable sources including the systems concentrating solar radiation. Though this attention concerns to a large measure the installations of energy purpose (making electric energy, natural gases conversion and others) the new SF have been created in Spain, Switzerland. Israel, Uzbekistan and other

The Applying of Coatings and Surface Thermal Treatment…

247

countries for the investigation of materials and technological processes in addition to the installations existing mainly in France, Japan, USA and republics of the former USSR. Due to the qualities described above SF are widely adopted for the synthesis and investigation of the properties of high-temperature oxide compounds (France, Japan, USA, Ukraine, Uzbekistan and other) and also for the investigation of thermal and temperature stability of materials and coatings by simulating extreme conditions of their operation (USA, France, Ukraine, Spain and other). In the last case the one-sided directional nature of radiant flux supply to the surface of the sample being heated is of great importance. This quality of solar furnaces as the other ones can be very useful for realizing technological processes of the kind of surface thermal treatment and applying the coatings of various functional purposes. But at present there are very few known works of the kind. The purpose of this work is to demonstrate the opportunities of solar furnaces utilization for obtaining glazed decorative coatings on building materials, wear resistant and protective coatings on metals and in other similar processes using the investigation carried out in the Institute for the Problems of Materials Science of National Academy of Sciences in Ukraine (IPMS NASU). Taking into account the absence of industrial production of solar furnaces and high cost of their individual manufacturing, special attention has been paid to the creation of the concentrators of solar radiation based on obsolete metallic antennae. The technology of their manufacturing has been developed in the IPMS NASU. The theoretical models and analytical methods for the calculation of optimum structure of facet mirror coating for solar concentrators have been developed. The problem has been solved for the determination of the rate of surface thermal treatment depending upon the properties of a material and the nature of power supply. Separately it is necessary to note that in that work not only the solar installations were used. Optical furnaces based on arc Xe-lamps of high pressure, were used also. By their technological potentialities they can be of completely independent importance. Like solar furnaces they possess almost all the advantages of radiant heating. Their functioning does not depend upon the Sun. But they possess their own shortcomings such as rather high expenditure of energy when efficiency is small, ozone and heat liberation to the environment, UV radiation etc. But if optical furnaces are compared with other technological equipment more or less of the same type for obtaining coatings (laser, for example) we can distinguish the following extra advantages: considerably greater area of heating when their dimensions and expenditure of energy are lesser (the question is not about the cases of the use of the most power light fluxes to obtain the hole in super hard materials, when lasers are out of competition).

1. DEVELOPMENT OF SOLAR FURNACES EXPERIMENTAL BASE ON IPMS NASU Taking into account the variety of the problems defined for the scientists in the field of materials science within high temperature region the SF constructions are also highly variable. The simplest SF is a helioinstallation with parabolic concentrator with solar radiation (dish) of direct tracking of the sun. Often they are created on the basis of military searchlights. They permit to obtain the highest heating temperatures. But they are

248

V.V. Pasichny and B.A. Uryukov

characterized by the essential disadvantages. They are the following: object the of heating moves constantly in a space; heating is directed upwards which causes the fall of high temperature particles onto the mirror in the case of sample melting or destruction if certain precautions are not taken. Besides, such installations operate in the open and they require special cover being out of operation state. The installations realized by the 'concentrator + heliostat' scheme are the most widespread.

1.1. Solar and Optical Furnaces of IPMS NASU The solar energy concentration systems developed at the IPMS NASU are divided into high−temperature (2500-3500 K), medium temperature (1500-2000 K) and low-temperature (500-800 K). Table 1-1 gives the main characteristics of the solar furnaces. All high−temperature systems were developed on the basis military searchlight equipment. The systems with the medium temperature level (SGU-6 and SGU-7) used as concentrators metallic radio antennae with bonded flat glass facets. The concentrator of low-temperature equipment SGU-11 was developed on the basis of 90 plane facets. It contains hinged sections for adjusting forming the concentrator facets, a sun tracking system and other elements. SGU9 equipment contains a photoelectric battery for separate feed of tracking system drives. SGU-8 horizontal-axial equipment has two adjacent concentrators: the first one in the form of a parabolic single mirror 2 m of diameter, the second in the form of a set of spherical facets situated on the parabolic surface with an extended focal distance and a correspondingly larger focal spot (the facets of this concentrator was developed by the Moscow Prozhector plant. All installations were certificated for the thermal parameters and other characteristics. Figure 1-1 shows the principle diagrams of a 'Cascade' two-position solar furnace being completed at the IPMS NASU. The optical system of equipment consists of two parabolic mirrors of the searchlight type 2 m in diameter and two flat heliostats. The first one tracks the sun and reflects the radiant flux onto the horizontal-axial concentrator and the other one been set at an angle of 45 degrees radiates the vertical-axial mirror. This design of equipment, regardless of certain losses of energy in the heliostats, is universal and should ensure the most suitable conditions for precision experiments with heating stationary object located on a vertical or horizontal plane. Figure 1-2 shows the general view of the IPMS solar furnaces laboratory in Kyiv. The most serious disadvantage of the solar energy system is that, as is well known, their efficiency depends on weather and seasonal conditions. The developed base of solar furnace simulators – optical furnaces based on high-pressure xenon arc lamps – makes it possible not only to eliminate this shortcoming but also widen the experimental possibilities of apparatus. Table 1-2 gives the main characteristics of optical furnaces developed at the IPMS NASU.

The Applying of Coatings and Surface Thermal Treatment…

249

Table 1-1. Main characteristics of the solar furnaces Equipment SGU-1 SGU-2

Optical system (mirror dimensions) Monoparaboloid (d = 1.5 m) The same

d min, mm

q max,

N max’ kW

Tmax’K

1.2

3800

6.0

1500

1.2

3800

6.0

1500

W/cm2

SGU-3

Monoparaboloid (d = 2.0 m)

1.8

3600

8.4

1200

SGU-4

The same

1.8

3600

8.4

1200

SGU-5

The same

1.8

3600

8.4

1200

SGU-6

Metallic antennae ( d=2.8 m), facets in the form of an equilateral triangle (a=50mm)

3.5

1800

36

80

SGU-7

Metallic antenna (d= 5 m), facets are the same as in SGU-6

8.5

1800

70

80

SGU-8 (ATON-1)

Heliostat 3.5 · 3.0 m, concentrator 2.0 ·2.0 m, 64 spherical facets 0.25 ·0.25 m

1.0

2300

25

100

SGU-9

Monoparaboloid (d = 1.0 m)

0.7

2800

2.2

600

SGU-10

Monoparaboloid (d =2.0 m)

1.8

3500

8.4

1000

SGU-11

Concentrator

1.6

600

150

4

'Cascade'

Heliostat (3.0 2.75 m), auxiliary heliostat (2.0 ·3.0 m) , 2 monoparaboloid (d=2 m)

1.0

2800

8.4

560

Additional description. Application Production of amorphos film, testing of light guides Loading mechanism. Thermal and mechanical tests Supersonic jet generator. Gas dynamic and thermal tests of materials. Loading mechanism, thermomechanical, physical and physico-chemical investigations. Vacuum system. Welding, brazing, coating processes. Thermomechanical test, heat treatment of constructional materials and biological objects, water freshening. Television equipment. Thermal tests of materials, thermo-chemical transformation of solar energy. Equipment for light pulse irradiation of materials and biological object. Photoelectric accumulator for feeding tracking systems. Cutting of cloth, heat treatment. Physicochemical investigations of oxides, heat treatment. Combustion of materials, irradiation of biological objects. Thermal surface treatment of materials and technological processes.

Comments: N max - power, Tmax temperature, dmin - diameter of the focal spot, q max - irradiation in the focal spot.

Table 1-2. Main characteristics of the optical furnaces Furnace Uran USS-1 'Crystal' 'Orbit'

Optical elements ( number, dimension, angle of beam convergence) Monoelipsoid, one piece; d=0.6 m; α=35ο Monoelipsoid, one piece ; d=0.35 m; α=66ο Monoelipsoid, three pieces ;d= 0.6 m; α=35ο Parabolic, six pieces, d=0.5 m; concentrating parabolic (one piece; d=1.5 m); α=120ο

Type and number of lamps, required power DKsShRB 10000-1, one piece; 10 kW DKsShRB 150 A-1 (one piece; 3 kW DKsShRB 10000-1, three pieces; 30 kW DKsShRB 10000-1, six pieces; 60 kW

Diameter of focal spot, mm

Heat flux, W/cm2

Temperature, K

10-12

600

2800

2.0

2000

3500

10

1500

3500

10

1300

3400

Application Heat treatment, thermal tests, irradiation of biological objects Brazing and welding cutting of cloth Examination of oxides, surface treatment and coating Thermal and gas dynamic test of materials

Comment. Orbita furnace is originally equipped with a vacuum chamber and a plasma generator of 300 kW power. Number of systems: Uran –3 systems, rest –1 system.

The Applying of Coatings and Surface Thermal Treatment…

251

Figure 1-1. Basic diagram of bivariant helioinstallation: 1 - movable heliostat; 2 - optical sensor of the tracking system; 3 - solar radiation; 4 - vertical - axial concentrator; 5 -, 7- sample with a coordinate device; 6 - horizontal -axial concentrator; 8 - conditionally immovable heliostat; 9 - flat facets.

Figure 1-2.

252

V.V. Pasichny and B.A. Uryukov

1.2. Development of the Analytic Method for Computation of the Facet Mirror Coating for Applying onto Paraboloid Antennae Now there are many technologies for manufacturing of solar concentrators. They are of single-unit production mainly. But there are no economically justified solutions of the problem and it hampers the use of the concentrators. This work is based on the idea we have been testing during 30 years. From 1974 in the Institute for Problems of Materials Science (IPMS) two concentrators of solar energy are operated without noticeable deterioration of technical data. They have been manufactured on the base of metal antennae ∅2.8 and 5.0 m. The mirror facets have been glued on the surface of the antennae [1]. The technological method used in the IPMS have been found at one time after relatively short period of search by the way of tests and errors and it has demonstrated its value in practice completely. Of course, it cannot be considered universal but it could be widely adopted if some problems are solved. These problems have not been considered in general or have been solved in insufficient volume.

1.2.1. Characteristics of Solar Radiation Concentrators (SRC) with Square Facets It has been demonstrated earlier [2] that when the facets are being placed onto SRC surface along the main rings (concentric circles) the facets configuration must be trapezium form to obtain a focal spot of the given diameter. The facets size changes in going from one main ring to another. Precise manufacturing the optimum facets calls for considerable financial and time expenditures especially when their required number is estimated in thousands. The facets of the identical size are used in practice. Usually they are square and placed along rectilinear secants of a surface as it is shown in the figure 1-3. It demonstrates naturally only qualitative pattern of the facets arrangement because it does not represent the spatialness of an object.

Figure 1-3. Pattern of facets arrangement on SRC.

The Applying of Coatings and Surface Thermal Treatment…

253

Figure 1-4. Radiant flux reflection onto surface. focal plane.

1.2.2. Focal Spot Size by Square Facets Let us determine the focal spot diameter for the SRC the surface of which is formed by square facets. The facet configuration in focal plane is deformed in comparison with initial one. The figure 1-4 explains the cause of this. Linear segment image in a plane being perpendicular to the figure plane remains to be unchanged if we neglect the small change connected with the radiation source finite size and the image of the linear segment laying in the figure plane changes: li = Kl ; K =

cos(θ / 2) cos θ

(1.1)

where l and li the length of the segment and its image, correspondingly; θ is an angle at which the segment center is visible from the focal spot center. So, in the figure central part the image is demonstrated for one of the facets arranged along the radius intersecting the facets along a diagonal (shaded band 1 in the figure 1-3. Depending upon the facets arrangement relatively the planes passing through the concentrator axis their images can take the form of rhomb (band 1), rectangle (for example for the facets lying within the shaded band 2) or nonequal-side parallelogram (band 3). The image deformation causes the change of the zone area in the focal plane center where all the rays reflected by the facets come together. This zone is a focal spot. So, the focal spot radius determined by the least (among all the facets images in focal plane) distance from the center to the image boundaries. Let us analyze the image configuration for the facet located at arbitrary place of the SRC reflecting surface (figure 1-5). Facet boundaries are denoted by thin line of a facet is projected onto focal plane without configuration change. This image is turned by an angle ϕ relatively the images symmetry axis being an axis perpendicular to the figure 1-4 plane and passing though focal spot center. A heavier line demonstrates the boundaries of the facet all the linear sizes of which in the y-axis direction have changes with the same proportionality coefficient.

254

V.V. Pasichny and B.A. Uryukov

Figure 1-5. Modification of facet configuration in focal plane.

Geometric analysis demonstrates that minimum distance from the center O to the image boundary is equal to: lmin K = a cos ϕ K 2 + tg 2ϕ

(1.2)

where a is the half of a square side; K is the coefficient of linear magnification by reflection (1.1). Analyzing the eq. (1.2), we can find that the minimum value of the lmin takes place when ϕ = 0 i.e. lmin min = a. Hence, the focal spot diameter will be equal to square facet side (by perfect conditions of facets attaching to SRC bearing surface when the ray being reflected from a facet center gets to focal spot center). When the SRC is great enough and the facets are small it is necessary to take into account the change of the spot diameter due to finite size of radiation source being the Sun. The focal spot radius reduction for this reason is equal to:

ΔrF =

sin α S ⎛ sin(θ / 2) ⎞ ⎜R+a ⎟ cos(θ − α S ) ⎝ cos θ ⎠

where aS is an angular radius of the Sun being equal to 0,004654 Rad, R is an radius drawn from the focal spot center to a facet center (figure 1-4) witch is determined by the following formula for SRC parabolic frame:

R=

2f 1 + cos θ

where f is a focal distance.

The Applying of Coatings and Surface Thermal Treatment…

255

The maximum value of the ΔrF takes place when the θ value is maximum i.e. when θ = θm (the half of SRC aperture). Since under usual conditions as << θm and a << R., we can use the approximate formula: ΔrF =

2α S f αS D = cos θ m (1 + cos θ m ) sin 2θ m

(1.3)

where D is a SRC diameter being equal to:

D=

4 f sin θ m 1 + cos θ m

For example, when D = 1.8 m, θm = 600 the focal spot radius reduction will be almost 1 cm approximately.

1.2.3. Number of Square Facets Covering All the SRC Area The facets number which is necessary for covering all the SRC bearing frame is calculated by the simplest method: the frame area is divided by a separate facet area. Let us estimate how such calculation raises too high the result. The error arises first of all because there is a technological gap between separate facets and secondly because a plane facet 'straightens' the section of the frame curved surface above which it is arranged. Let the average technological gap is equal to ε. Its half fits to each neighboring side i.e. the total area of the gap fitting to each square facet with the side of 2a is equal to 4aε and its relation to facet area is equal ε /a. So, when the gap is 0.3 mm and square side is 25 mm some more than 2% of SRC area is used for the gap ensuring. To take the second factor into account it is necessary to calculate the area of ‘curved square’ of the frame surface, which is substituted for a plane facet. Let us arrange the origin of Cartesian coordinates at the SRC center. Then the equation for parabolic frame will be in the following form: z=

x2 + y2 4f

(1.4)

Curved surface area is calculated by the formula:

S=

x2 y 2

∫∫

x1 y1

2

2

⎛ ∂z ⎞ ⎛ ∂z ⎞ 1 + ⎜ ⎟ + ⎜⎜ ⎟⎟ dx dy ⎝ ∂x ⎠ ⎝ ∂y ⎠

Substituting the (1.4) here we shall get:

256

V.V. Pasichny and B.A. Uryukov S=

x2 y 2

∫∫

x1 y1

1+

x2 + y2 dx dy 4f 2

(1.5)

To determine the range of integration let us consider the simplest case when facet center is arranged in the plane y = 0 a facet itself oriented as it is demonstrated in the figure 1-6.

Figure 1-6. Layout of the facet being investigated.

The facet side edges lie in the planes being at a distance y = -a and y = +a from the plane y = 0. Obviously, they are the range of integration with respect to the y coordinate. To determine the range of integration with respect to the x coordinate it is necessary to take into account that a facet 'replaces' the part of the frame surface lying under it. Therefore the range of integration (x1 and x2) does not agree with the abscissa x01 and x02 of the facet edges. The problem of their determination is solved in the following way. Let some point x0 is given in the region of which a facet must be arranged. In this case the point x0 assignment of the angle θ, at which it is seen from the focus also. In order that the ray reflected from the facet center O would get at the SRC focus the center must lay on the ray reflected from the point x0 if the frame surface is reflecting. As a result the facet location is determined unambiguously and its corners touch the frame on the radii ρ1 and ρ2 as demonstrated in the first figure 4.

θ

θ

x1o, 2 = x0o m a cos ; z1o.2 = z0o m a sin 2 2 Let us calculate the coordinates of facet sides being perpendicular to the second figure 4 plane. It follows from the first figure 1-6 that:

x1o, 2 = ρ12, 2 − a 2 At the same time the facet center lies on the ray directed at the focus i.e.:

The Applying of Coatings and Surface Thermal Treatment…

257

z0o = f − x0o ctgθ where f is a SRC focal distance. Overlapping these formulae makes it possible to calculate the ρ1 and ρ2 radii: 2

⎛ a ⎛ ρ1.2 ⎞ ⎜⎜ ⎟⎟ = ⎜⎜ ⎝2f ⎝2f ⎠

2 ⎡ ⎛ a ⎞ 1 ⎟⎟ + ⎢ − ⎜⎜ 2 ⎢ sin θ ⎝ 2 f ⎠ ⎣

2 ⎤ ⎞ a ⎟⎟ m − ctg θ ⎥ ⎥ 2 f sin( θ /2) ⎠ ⎦

2

(1 . 6 )

and other coordinates of facet points with their help. The sought x1 and x2 points are determined on the basis of the fact that they are the points of the intersection of the normals emanating from the facet corresponding ends with the frame contour. As a result we shall get: 2

θ θ ⎛ ⎞ x1, 2 = ⎜ 2 f ctg + ρ12, 2 − a 2 ⎟ + a 2 − 2 f ctg 2 2 ⎝ ⎠

(1.7)

Double integral (1.5) can be calculated in general view. But the formulae being obtained are inconvenient for estimating the relation between the areas of plane and curved squares because of its complexity. Simpler result is obtained if we assume that there is a small parameter in this problem i.e. a / f << 1. Let us denote:

b=

a x y S ; ξ= ; η= ; s= 2f 2f 2f 4f 2

Then the formulae (1.5) will change in the following way:

s=

+ bξ 2

∫ ξ∫

−b

1 + ξ 2 + η 2 dξ dη

1

Let us expand the s into a series in the environ of b = 0:

b ⎛ ∂s ⎞ b2 ⎛ ∂ 2s ⎞ b3 ⎛ ∂ 3s ⎞ b4 ⎛ ∂ 4s ⎞ s = ⎜ ⎟ + ⎜⎜ 2 ⎟⎟ + ⎜⎜ 3 ⎟⎟ + ⎜⎜ 4 ⎟⎟ + ... 1! ⎝ ∂b ⎠b =0 2! ⎝ ∂b ⎠b =0 3! ⎝ ∂b ⎠b =0 4! ⎝ ∂b ⎠b =0 a put down some derivatives entering in the expansion taking into account that when b → 0 the integrals being within the limits from – b to + b and from ξ1 to ξ2 will tend to zero (ξ1, ξ2 → ξ0):

258

V.V. Pasichny and B.A. Uryukov

⎛ ∂s ⎞ ⎜ ⎟ = 0; ⎝ ∂b ⎠ b = 0

⎛ ∂2s ⎞ ⎛ ∂ξ ∂ξ ⎞ ⎜⎜ 2 ⎟⎟ = 4 1 + ξ 02 ⎜ 2 − 1 ⎟ ; ⎝ ∂b ∂b ⎠ b =0 ⎝ ∂b ⎠ b =0

⎛ ∂3s ⎞ ⎛ ∂ 2ξ ∂ 2ξ ⎞ ⎜⎜ 3 ⎟⎟ = 6 1 + ξ 02 ⎜⎜ 22 − 21 ⎟⎟ ; ∂b ⎠ b = 0 ⎝ ∂b ⎠ b = 0 ⎝ ∂b

⎛ ∂4s ⎞ ⎛ ∂ 3ξ ∂ 3ξ ⎞ ⎜⎜ 4 ⎟⎟ = 8 1 + ξ 02 ⎜⎜ 32 − 31 ⎟⎟ ∂b ⎠b =0 ⎝ ∂b ⎠ b =0 ⎝ ∂b

Let us put down the expression for determining the ξ1,2 in the form being convenient to seek the derivatives:

(

)

2

(ξ1, 2 + γ ) 2 = α + α 2 m βb − b 2 + b 2 ; α =

1 1 ;β = ; γ = ctg (θ / 2) sin θ sin(θ / 2)

As a result of all the calculations we shall get: 3 S θ⎞ ⎛ 1 = 1 + b 2 sin 2 θ ⎜1 + cos 2 ⎟ 2 4a 8 2⎠ ⎝ 3

(1.8)

The maximum correction to one takes place when the θ value is maximum i.e. when θ = θm. It is ~ 0.35 b2 for θm = 600. So, when 2a = 2,5 cm, D = 1,8 m it is smaller than 1⋅10-4 i.e. the difference between the areas of bearing contour and the facets attached to it can be not taken into account.

1.2.4. Calculation of Characteristics of SRC with Square Facets The flux of solar radiation being incident onto a facet of the area Sn is equal (figure 1-3) to qS Sn cos(θn/2). This flux is incident on focal plain reflecting from the mirror and distributed over the area Snf = SnKn in accordance with the law of linear size change by reflection in connection with eq. (1.1). Thus, radiant flux being incident onto focal plane differs from solar radiation flux: q fn = qS cos θ n

(1.9)

Illuminance coefficient for the focal spot is equal to the sum of the qfn /qS relations over all the facets: N

q fn

n =1

qS

q =∑

N

=∑ cos θ n

(1.10)

n =1

Thus, to find the focal spot total illuminance it is enough to determine angular arrangement of each separate facet center relatively to the focus. Taking into account the SRC frame surface is curved (paraboloid of revolution) it is very difficult to do this because the boundaries of the bands where the facets are arranged are the lines of intersection of a paraboloid and some curved surfaces being unknown beforehand. The property characterizing

The Applying of Coatings and Surface Thermal Treatment…

259

them is a fact that they are at a distance 2a one from another in the direction of a tangent to a paraboloid in the middle of this distance. Seeking the equations of such surfaces is a very complex problem of analytic geometry. We shall use the different method. In accordance with the result of the previous part the frame total area does not differ practically from total area of facets arranged on it. Relying on this fact we shall map paraboloid surface on to a plane in such way that its area should be equal to the area of formed circle similarly to the figure 1-3. In this case all the paraboloid surfaces bounded by the angle θ = const, i.e. by the circles which lie on its surface with the centers on the axis of SRC are mapped into plane circles with the same areas. The area of every such circle is equal: Sθ =

⎞ πD 2 ⎞ 8π f 2 ⎛ 1 1 ⎛ θ ⎞⎛ ⎜⎜ − 1⎟⎟ = ctg 2 ⎜ m ⎟⎜⎜ − 1⎟⎟ 3 3 3 ⎝ cos (θ / 2) ⎠ 6 ⎝ 2 ⎠⎝ cos (θ / 2) ⎠

(1.11)

The total area of SRC is calculated by the (1.11) when θ = θm. If θm = 60o S = 0,2698 πD . It differs only by 8% from the area of a circle of the same diameter. Radius of a circle of mapping is equal: 2

r1θ =

Sθ

(1.12)

π

Write r1 for mapping radius of the entire SRC surface, which is determined through the area S. Let us fill the plane mapping with square facets. It is obvious that their total number will be equal to number of facets arranged on a paraboloid. We shall number them in the following way (first quadrant being in the figure 1-3): the number j in horizontal band is counted off from 'zero' band through the middle of which coordinate axis runs along. Accordingly, the number i in vertical band counted off. The explanation of the way of numbering is given in the figure 1-3. The distance from circle center to the center of the facet being itemized under number i, j is equal to: ri , j = 2a i 2 + j 2

(1.13)

The area of the circle of this radius is equal to the area of the circular surface of paraboloid which is visible at angle θi,j from the focus. Putting the (1.11) equal to the (1.12) we shall get the expression for the given facet angular coordinate on bearing paraboloid:

θi, j

⎡ a2 ⎤ θ = ⎢24 2 tg 2 m i 2 + j 2 + 1⎥ cos 2 ⎣ D 2 ⎦

(

)

−1 / 3

(1.14)

Thus, the increase coefficient of focal spot illuminance will be equal to: im

jm ( i )

i =0

j =1

q = 1 + 4∑ ∑ cos θ i , j

(1.15)

260

V.V. Pasichny and B.A. Uryukov

The first term (one) is in agreement with the central facet; the multiplier 4 came out of the fact that only the facets arranged in single one quadrant were dealt with. Different initial numbers in the sums make it possible to avoid the doubling of the number of facets arranged in the bands along which the coordinate axes run. The maximum number along the vertical is equal to the number of the facets, which can be arranged within the band of the length r1: ⎛ r ⎞ im = INT⎜ 1 ⎟ ⎝ 2a ⎠

(1.16)

where the symbol INT stands for an ‘integral number’. The facets number within horizontal band depends upon its number along the vertical: ⎛ r2 ⎞ jm = INT⎜ 1 2 − i 2 ⎟ ⎜ 4a ⎟ ⎝ ⎠

(1.17)

1.2.5. Comparison of Characteristics of SRC with Optimum and Square Facets The developed method for calculating the characteristics of SRC with square facets makes it possible to compare its properties with SRC with the facets of optimum configuration. The characteristics of the existing concentrator SGU-12 of 1.8 m diameter with square facets with a side 2a = 2,5 cm have been calculated. The results of the 1-st section demonstrate that the diameter of this concentrator focal spot is equal to 0,5 cm. The facets total number is N = 4400. The coefficient of the focal spot illuminance turned out to be equal to ⎯q = 3170. For the practical applications such a characteristic of SRC energy efficiency as the temperature of an object being heated by solar radiation in a focal spot is more clear and important. In the ideal case it is equal to radiation - equilibrium temperature, which is calculated by the formula: 1/ 4

⎛q q ⎞ TR = ⎜⎜ S ⎟⎟ ⎝ σS ⎠

where σs is a Stefan-Boltzmann constant. Solar radiation flux was assumed to be equal to 800 W/m2. The comparison of the characteristics was made for the three versions in which one of the main parameters was assumed to be the same as the initial SRC one: 1. 2. 3. 4.

The same focal spot diameter df ; N,⎯q were compared. The same number of facets N; df , q were compared. The same coefficients of illuminance increase; df , N were compared. The results are given in the table 1-3.

The Applying of Coatings and Surface Thermal Treatment…

261

Table 1-3. The comparison of the characteristics of SRC with optimal and square facets Parameters SGU-3 Version 1 Version 2 Version 3

df, cm 0,5 0,5 1,8 1,75

N 4400 14530 4400 4520

⎯q 3170 10640 3140 3170

TR K 2590 35002580 2590

The comparison in the version 1 demonstrates that the optimization of facets configuration makes it possible to improve substantially SRC characteristics in comparison with square facets application due to the facets number increase. The version 2 demonstrates that when the number of facets is the same the SRC with the facets of optimum configuration makes it possible to increase noticeably the focal spot size without the change of irradiancy. At last when the irradiancy in a focal spot is the same the optimization of facets configuration permits also to increase noticeably the size of a spot by small increase of facets number. Thus, the method is developed for the calculation of the characteristics of the SRC with square facets, which are arranged over the surface of parabolic frame. The comparison of the characteristics of the SRC with square facets and facets of optimum configuration [2] demonstrates that the optimization of facets configuration makes it possible to increase essentially the focal spot size and improve considerably energy properties of SRC.

1.3. Selection of Materials and Development of the Process Flow Diagram for Applying of the Mirror Coating onto Metal Base 1.3.1. Grounds of the Technology Applying of the mirror coating onto metal surface is a very significant element of the technology for the creation of high-powered solar installation on the base of antenna. The method of mirror coating glueing onto antenna surface has been developed and tested in the IPMS. Two technologies have been worked through: 1. The glueing of polyethylenetheryphthalate (lavsan) film; 2. The glueing of glass mirror facets. Accordingly to the developed technology the mirror coating on the base of metallized lavsan film was applied onto the parabolic concentrator (∅2.8 m). On the grounds of service test the conclusion have been made, that the quality of the obtained mirror coating is up to the requirements and the technology of applying is simple, the coating is light and not expensive. But the service test has revealed some short-comings of the coating: the lavsan film ages under the influence of solar radiation, atmospheric precipitation and wind, metal coating of the film gets washed off grows dim. It leads to the uselessness of the coating and to the necessity of the coating renewal every year. As the result of these short-comings the first technology has not been widely adopted and has been declined.

262

V.V. Pasichny and B.A. Uryukov

The second technology consists of glueing the glass facets onto the metal surface. The mirror coating is applied onto the inner surface of the facets regarding the incident light flux. During working through the technology the sampling of the glueing compound was the most complicated problem. Rather serious requirements are raised to the service characteristics of the glue. These characteristics are caused by the specificity of the concentrator construction and its operating conditions. The concentrators of solar radiation are operated in the open. They stand the influence of atmospheric precipitation, of temperature gradient of the seasons and climatic conditions, of vibration and wind load. The lack of construction rigidity also influences upon the operational reliability of the concentrators. Thus, the glueing compound should be elastic and at the same time rather strong to be deformed without destruction. It should be possessed of high adhesion to the construction material. The glue should be characterized by the wet strength, the heat- and cold resistance. It is necessary to protect the butt joints of the glass facets from moisture penetration and the destruction of the mirror coating. The glueing compound must harden without shrinkage by room temperature, it should be technological and should guarantee the operating of the construction for a long time. We have got some glueing compounds which could solve the problem for a while. There is, for example, such glue as elastic compound UP6-104 on the base of epoxyaliphatic resin modified with rubber PDI-3АК. In the Institute for Plastic the elastic epoxide compound UP6-123 and the epoxide compound UP-5-244 have been developed. However the newest scientific-analytical publications [3] assert that the sealant U30MES5 selected 25 years ago is one of the best even to-day proceeding from the whole spectrum of its properties. Operating of the solar concentrators ∅2.8 and 5 m with the mirror coating made out of glass facets for all these years has shown exceptionally high qualities of the sealant used for glueing the facets to the metal base by the reflecting layer. On these grounds the quite logical conclusion can be made about the expediency of the sealant U30MES5 further use. The polysulphide thiocol sealant U30МES5 is the thermoreactive material which hardens by the room temperature and under the influence of special vulcanizer. The base of this sealant is the liquid bifunctional polysulphide rubber (liquid thiocol) with the molecular mass 2⋅10-3 - 6⋅10-3 and the viscosity from 20 up to 50 Pa by the temperature 25 -30oC. This sealant is spattled and it could be applied by the layer of 4-5 mm. It makes possible to carry out the adjusting of the facets. The glue is three-components. The components are mixed immediately before using. The sealant U30MES5 has the properties of rubber and glue as well. Just as rubber it has elasticity and tensile strength not less than 1.5 MPa and relative elongation not less than 200%. As glue it has certain viscosity and adhesion relatively different materials: ferrous and non-ferrous metals, silicate and organic glass, wood, concrete, plastic and rubber products. The presence of sulphur atoms in the main chain of thiocol molecule guarantees the high stability of the sealant to many aggressive mediums (oil, petroleum, fuels) and to the influence of ozone, light and radiation. The absence of the double bond in the molecular chain guarantees the high stability to the influence of oxygen and ozone, to the aging in the open and to the temperature rise.

The Applying of Coatings and Surface Thermal Treatment…

263

Thus, after hardening the sealant U30MES5 has high elasticity and adhesiveness in the temperature range from -60 up to +130 oC, and necessary mechanical characteristics. The sealant does not cause corrosion of the materials being in contact, has high dielectric properties. It does not age for years. The technology of glass facets glueing with the sealant U30MES5 is simple. The sealant is not toxic. Table 1-4. Main characteristics of the sealant

Trade Mark

Work temperature O C

Vitality, hours

U30МES5

from -60 up to +130

2-9

Color

Density Kg/m3

Shore hardness, nominal units

Nominal tensile strength MPa

Relative tensile elongation, %

black

1450

40-60

1.5-2.5

200-400

1.3.2. Selection of the Concentrator and Analysis of the Surface Quality The second-hand radio-antenna or the specially manufactured parabolic dish could be used for the manufacturing of solar furnaces concentrators. Dimensions and configuration of the concentrator and of the facets should be matched according to the purpose. The quality of the surface, presence of roughness, hollows and extraneous impregnations, reproduction of the assigned profile should be controlled by sight or with the help of a template. Small unevenness and hollows should be removed by puttying (the putty should be selected accordingly to the metal of the concentrator). The concentrators with considerable roughness or profile deviation should be rejected. Clearing of extraneous impregnations, projections and getting the roughness of the concentrator surface could be realized by its stripping with coarse-grained emery and by following cleaning with vacuum cleaner. If the dimension of the concentrator is above 5-10 m in diameter it is expediently to make the facets with different dimension in the form of an equilateral trapezium. In this and other possible cases it is necessary to make preliminary marking of the concentrator surface considering the chinks between the sides of the facets being about 0.3 mm. Immediately before applying of the coating the concentrator should be degreased by cleaning with soft cotton waste wetted in sweet gasoline. For drying the concentrator should be kept by the room temperature during 30 min or blown round by warm and dry air from an air heater. The glass for the facets manufacturing should be sampled depending upon the dimension and the purpose of the concentrator. It is necessary to take into account the glass optical density and the flatness of its surface. The thickness of the glass can vary from 2 up to 4-6 mm depending upon acceptable dampness load. The glass gets cut to pieces of optimum dimensions taking into account technological specificity of reflecting layer applying and following cutting out into facets. It is expected to use the industrial technologies taking into account the operating requirements to the material and the characteristic of the reflecting coating (Al, Ag etc.). The recommended glue-sealant is considered to be chemically neutral regarding the absolute majority of materials. Taking into account the great number of new modern substances which can be used to fix sprayed metallized layer it is expediently to carry out

264

V.V. Pasichny and B.A. Uryukov

preliminary examination if there is negative chemical interaction between the glue sealant and outer layer of used substance. It could be done by glueing of two or three facet samples and examination of the joint quality in 30-36 hours. Regarding the concentrator dimension the facets could be of the same dimension or of different configuration and dimension. The quality of the glass and of the mirror coating are controlled by sight. The facets are sorted according to marking of the concentrator surface. Immediately before using the glass facets with the applied mirror-coating get them degreased by dipping into the vessel with sweet gasoline. The degreased facets are kept under room temperature not less than 40 min. Thorough mixing of the components is the necessary condition to make the glue. Special mixers for viscous substance can be used. One can mix by hand with a spattle during not less than 15 min. The ratio of the pastes U30MES5 and N9 is 100:5 accordingly. This mixture can be made in advance and be preserved during 10 days. Immediately before using the vulcanizer-hardener DFT should be added to the mixture in number of 0.5 - 1 g per 100 g of the sealant U30MES5. The glue is being mixed during 15 min. The glue-sealant U30MES5 can be used during 3 hours . The layer of the sealant (the thickness is 2-4 mm) is applied onto the inner surface of the glass facet with a spattle. The facet is laid on the surface of the concentrator, is pressed for compact joining and is adjusted if necessary. The following facets are laid accordingly to the marking of the concentrator surface compactly one to another. To apply the coating onto the concentrator of greater diameter it is necessary to have auxiliary equipment such as suspended seats. After glueing all the facets the concentrator is kept by room-temperature during 24 hours till complete hardening of the sealant. When applying the mirror-coating getting of moisture on the concentrator surface and on the facets is not permitted till complete hardening of the sealant. The optimum conditions of making and using of the sealant are the following: the temperature is 18-25 oC and the relative air humidity is 50-75%. The unnecessary sealant squeezed out of the spaces between the facets is cut off with a knife. If it is necessary the drops and marks of glueing substance on the outer surface of the facets are removed with sweet gasoline.

1.3.3. The Manufacturing of the Concentrator Demonstrative Sample The concentrator demonstrative sample based on ∅1.8 m metal antenna was manufactured to verify the developed technology with up-to-date components of sealant. The antenna produced by Krasyliv Aggregate Plant (Ukraine) was manufactured out of duralumin and aimed for the reception of satellite TV signals. Unfortunately, the antenna technical certificate does not contain the information concerning the paraboloid geometric accuracy and characteristics because the prognostication of power parameters and the selection of the glass facets dimensions have been complicated. Using the computationanalytic method above given, it has been determined the facets dimensions and their number taking into account the necessity to obtain the focal spot of 20-25 mm diameter with even enough distribution of the irradiancy. Accordingly to [2] a trapezium is an optimum configuration for the facets. The dimensions of the trapezium depends upon its location on the antenna surface. Taking into account the great number of the facets necessary to cover the antenna (about 4000), their small dimension, relatively small dimension of the antenna itself and insignificance power advantages in this case it was used the quadratic form of facets of the same dimension 25x25 mm. It has enabled to change considerably the time and

The Applying of Coatings and Surface Thermal Treatment…

265

expenditures for the manufacturing and gluing of the mirrors. The polished glass of high quality of 2.0 mm thickness was used for the facets manufacturing. The metallization of the glass by aluminium and its cutting have been realized by the Kyiv Plant of Glass Products. The sealant U30MES5 has been used as a gluing substance. It is necessary to take into consideration the significant viscosity of the sealant when the facets dimensions are being determined. It is especially important when a concentrator being operated in wintertime. Over-diminution of the facet dimensions can cause the errors of the accuracy of their gluing to the base due to the glue remains and insufficient enough contact of the facet ends with metal. It can cause some decrease of the concentrator specific power parameters due to extra light scattering of solar radiation near by the centre of the focal spot. The concentrator total power should not decrease in this case.

1.3.4. Plant Flow Diagram of Mirror Concentrator Manufacturing

Note: There are not design and calculations works in the technological plant flow diagram.

266

V.V. Pasichny and B.A. Uryukov

1.3.5. Comparative Analysis of the New Sealant State and Its State after 25 Years of Operation The comparative analysis of physical state has been carried out by the method of IR spectroscopy for the sealant manufactured in 1973 and 1998. It has been done for the U30MES5 used for glueing the mirror facets onto antenna metal base. The samples of the sealant have been taken from the installation SGU-6 and the concentrator of ∅1.8 m. The analysis has been carried out by the specialists of the Institute for Chemistry of HighMolecular Compositions of NASU. The traditional research methods (transmission of a thin film, KBr tablets) have turned out to be useless due to the specificity of physical properties of the sealant U30MES5. But in this case also there have been considerable complications because the sealant contains great amount of soot which is characterized by continuous spectrum and strong absorption in IR. region. As a result of the search the conditions was found out when the faint spectrum of vulcanized thiocol was recorded. That is why the relatively new method RBTIR (repeatedly broken total internal reflection) was used. The V-prism was an element of RBTIR, the number of reflections was N-4, the incidence angle was 450. Thiocol is characterized by the typical I-R absorption spectrum in the region 900-15000 cm-1 with the intensive bands of about 1000 cm-1 wave length. All the bands of thiocol spectrum are observed in the spectra of the samples both 1973 and 1998. It follows from this that the chemical structure of the samples is the same. That is to say it is not destroyed and does not change in time. The spectrum of the sample of 1973 differs by greater diffusivity. It can be caused by some loss of elasticity due to the “oozing” of additions - plasticizer. In such case the optical contact of the sample with working face of the RBTIR prism becomes worse. It is accompanied by the worsening of spectrum contrast. Some drop of relative intensity for the absorption bands near 1200 and 1600 cm-1 (oscillations of benzene rings of dibenzthiazolylsulfide or diphenylguanidine for example) can be connected with the “oozing” of the plasticizers. The widening of the band 1730 cm-1 in the spectrum of the sample 1973 indicates the oxidation or the absorption of gaseous products with the group C = 0. As a result of the carried out analysis the important conclusion are drawn that the considerable changes of the chemical structure of polymer base for the sample 1973 is not found out in comparison with the sample 1998. Some changes of the sample 1973 spectrum perhaps are stipulated by the oxidation or the absorption of the compounds with the group C = 0 and also by the decrease of additions - plasticizers due to the “oozing” which is the nature of rubbers. It can cause some loss of elasticity but not of strength. Thus, it has obtained the confirmation of high quality of selected glue-sealant and conservation of its operational characteristics for a long time.

CONCLUSION FOR PART 1. 1.

2.

The analytic methods for computation of number, dimensions and optimum configuration of the plane facets have been developed for forming the mirror coating under the given clearances of the parabolic solar energy concentrator and focal spot dimensions. The plant flow diagram has been developed for applying the mirror coating by

The Applying of Coatings and Surface Thermal Treatment…

3. 4.

267

glueing the plane mirror facets onto the metal antennae. The use of the sealant U30MES5 as a glueing substance has been grounded. The demonstrative sample of the solar energy concentrator has been manufactured on the base of the satellite television antenna of ∅1.8 m by the proposed technology. The fulfilled complex of methodic-analytical and technological research and elaboration will permit the grounded and purposeful use of obsolete and up-to-date parabolic antenna of different purpose as the base of solar energy concentrators.

2. THEORETICAL MODEL OF SURFACE THERMAL TREATMENT OF NONMETALS IN SOLAR FURNACES Introduction The investigations of carried out in the IPMS NASU has demonstrated the processes for surface thermal treatment and materials fusing with the formation of coatings of various purposes such as corrosion resistance, conductivity, wear resistance, decorativeness and so on [5] are promising. In this case physicochemical transformations take place in surface layers of the wares being treated both for ware material and the substances forming the coatings. Both physicochemical and thermomechanical properties of base material and coating substance and radiant flux characteristics and the conditions of its supplying influence upon the quality of the coating being formed. Among them are the focal spot size and the irradiancy, instant or gradual running into the given conditions of heating, stationary heating all the surface or its treatment by scanning and so on Experimental determination of optimum conditions of thermal treatment is very difficult often and not always permits to answer all the questions. Preliminary calculations and theoretical analysis can promote substantially the process development. The purpose of this paper is a presentation of relatively simple methods for the calculation of radiant flux influence on the wares being irradiated and which take into account the features of heating in solar furnaces. High-temperature heating the wares in solar furnaces is nonlinear from the point of view of the theory of heat conduction even by constant thermal properties of the materials being treated due to substantial reradiation of the incident onto a surface flux. In the same time all the most developed methods for analytical solving the problems of heat conduction are linear. Solving nonlinear problems is realized mainly by numerical methods. The generalization of their results requires a great number of computer experiments. But it is a good idea always to have an opportunity to evaluate quickly and simple the results being expected which cannot be ensured by computer experiments. Some nonlinear problems including the radiation from a surface, melting or ablation can be solved with the use of integral approximate methods [6– 10]. The two such methods are known in the publications: the Goodman method [6] and the Tsyan θ-moment method [8]. The Goodman method is based on the approximate solution of the equation obtained by formal integrating the initial differential heat conduction equation over the region under consideration with the following approximation of temperature profile by some function of time and coordinates and the determination of the approximation parameters with the help of this equation. Side by side with the mentioned equation in the

268

V.V. Pasichny and B.A. Uryukov

Tsyan method the equation is used also which is obtained by integrating the equation of heat conduction being multiplied by temperature. This method being more tedious though gives the results being more accurate substantially.

2.1. The Heating of Stationary Plate Problem Equations and Solution Method In the case when a surface of a ware (it is a plate usually) is arranged in focal spot completely where radiant flux is uniform more or less we can consider the problem of onedimensional heating. The equation of one-dimensional heat propagation in a material with constant thermal properties being used in the Goodman method is in the following form: x

λ ⎡ ∂T d 2 dx dx ∂T ⎤ T dx = T ( x2 ) 2 − T ( x1 ) 1 + ( x2 ) − ( x1 )⎥ ∫ ⎢ dt x1 dt dt ρc ⎣ ∂x ∂x ⎦ 2 x x2 1 d 2 2 1 2 dx2 1 2 dx1 λ ⎡ ∂T ∂T ⎛ ∂T ⎞ ⎤ T dx T ( x ) T ( x ) T ( x ) T ( x ) = − + − − ⎢ 2 1 2 1 ∫ ⎜ ∂x ⎟⎠ dx ⎥⎥ 2 dt x∫1 2 dt 2 dt ρc ⎢⎣ ∂x ∂x x1 ⎝ ⎦

(2.1)

(2.2)

where х1 and х2 are arbitrary coordinates (х1 < х2). The equation being used in the Tsyan method in addition The problem statements are the following: radiant flux is incident on a surface x = 0. Another boundary of the plate is for enough from the first one so it can be considered that it is infinitely for where the temperature is equal to the initial one Та, and thermal flow is equal to zero. When the surface is partially melted the boundary between the melt and solid body is under х = х0(t), its rate Vm = dx0/dt. Accordingly to these conditions the equations (2.1) and (2.2) are in the following form: ∞

d λ ∂T (T − Ta ) dx = −[T ( x0 ) − Ta ]Vт − ( x0 ) ∫ dt x0 ρ c ∂x 2 ∞ ∞ 1 d 1 λ ⎡ 1 ∂T 2 ⎛ ∂T ⎞ ⎤ 2 2 ( T T ) dx [ T ( x ) T ] V ( x ) − = − − − + ⎢ a a т 0 0 ∫ ⎜ ∂x ⎟⎠ dx⎥⎥ 2 dt x∫0 2 ρ c ⎢⎣ 2 ∂x x0 ⎝ ⎦

(2.3)

(2.4)

The condition of heat input on the plate outer surface taking in account the radiation from the surface: 4

qi = εσ T0 − λ

∂T ( 0) ∂x

where Т0 is the temperature of the plate surface by the heating and the surface of the melt film by the partial melting; λ is the coefficient of thermal conductivity of the plate during the

The Applying of Coatings and Surface Thermal Treatment…

269

period of heating and λ = λт (coefficient of thermal conductivity of the melt) during the period of melting. When the film thickness is small the temperature distribution within it can be considered to be linear, i.e. thermal flow getting into it on the outer boundary reaches the boundary with solid body without any change. Energy expenditure are necessary for realizing phase transformation so thermal flow penetrating into a solid material decreases. Thus, the condition of heat input on the boundary of the melt film and solid body is in the following form:

∂T ( х0 ) + ρ Vт L ∂x

qi = εσ T0 − λ 4

(2.5)

where L is the specific melting heat. On account of the linearity of thermal flow within the film the temperature of outer surface of the melt is equal to: T0 = Tm +

x0

λm

[q − εσ T ] 4

i

(2.6)

0

The success of soling depends in great degree upon the accuracy of the selection of the approximating function Let as take the simple enough function answering the obvious physical conditions of the problem: T − Та= [T(x0) − Ta](1+bξ) exp(−βξ) (2.7) where: ξ = (x −х0)qi/λ(Tm − Ta) is the dimensionless coordinate being counted off moving boundary (by melting); β and b are the form-parameters of approximating profile which are time functions being subject to the determination. During the process of heating (until melting) Т(х0) = Т0 (surface temperature). When melting takes place the temperature on the boundary of solid body and the melt is equal to melting temperature, i.e. Т(х0) = Тт; in this case the temperature of the outer surface of the melt film Т0 > Тт. Let us take the complex: τ = tqi2/ρλc/(Tm−Та)2, and θ = (T – Та)/(Tm−Та) as a dimensionless temperature. Then the problem and the boundary conditions (2.3) – (2.6) will be in the form: ∞

d ∂θ θ dξ = −v − (0) dτ ∫0 ∂ξ

(2.8)

2

∞ ∞ ⎛ ∂θ ⎞ ∂θ 2 d 2 ⎜⎜ ⎟⎟ dξ = − − − θ d ξ v (0) 2 ∫ ∫ ∂ξ ∂ξ ⎠ dτ 0 0⎝

where v=Vтρ c(Tm−Та) /qi is dimensionless rate of surface failure; ∂θ 1 = C (θ 0 + ω a ) 4 − (0) + Bv ∂ξ

(2.9)

(2.10)

270

V.V. Pasichny and B.A. Uryukov

θ0 = 1 +

λ [1 − C (θ 0 + ω a ) 4 ]η 0 λm

(2.11)

where С=εσ (Т−Та)4/qi; B = L/c(Tm−Та); ωа = Та/(Tm−Та) are the physical parameters of the given problem; η0=х0qi/λ(T−Та) is the dimensionless coordinate of the boundary of the melt film and solid body and also

v=

dη 0 dτ

(2.12)

The approximation of temperature profile: θ =θ0 (1+bξ) exp(−βξ)

(2.13)

During the partial melting period θ = (1+bξ) exp(−βξ), because on the boundary with the melt the temperature is equal to melting temperature. In the partial melting stage b(τ), β(τ), θ0(τ) v(τ) and η0(τ), are the functions being determined. For their determination there are 5 equations (2.8)–(2.12). The initial conditions for solving the problem of melting are values of these functions at the moment when the plate surface temperature attains melting temperature. For their determination it is necessary to solve the problem of surface heating. In this period it is obviously that v=0, η0 = 0. The equations (2.8)–(2.10) are the equations of the given problem in which we must assume v = 0. In its turn the initial conditions for the given problem are the values of the functions b(τ) and β(τ), at the moment τ = 0, when θ0 = 0. The analysis has demonstrated that when τ → 0 functions b(τ) and β(τ) → ∞, as τ−1/2 and θ0 ~ τ1/2. Therefore the solving was found at the surface temperature being close to the initial temperature. In this case thermal flow penetrating into a body can be considered to be constant and the mentioned equations can be solved analytically:

β = 1.644 τ -1/2; b = 0.756 τ -1/2; θ0 = 1.126(1–С θа4) τ ½

(2.14)

For constant thermal flow on a surface q = q0 the exact solution is known:

T − Та x = exp(−u 2 ) − π u[1 − erf (u )]; u = T0 − Т а 4at

(2.15)

where а = λ/ρ c is the coefficient of temperature conductivity: The dependence of surface temperature upon time:

T0 − Т а =

2 q0 π λ

at

(2.16)

Surface temperature calculated by the last formula (2.14), in which 1–С θа4 ≈ 1, almost dose not differ from the accurate one (the error is 0.2%).

The Applying of Coatings and Surface Thermal Treatment…

271

The time from the beginning of heating till the melting temperature is determined from the equation (16): τ = τf = π/4. At this moment the variables и and ξ are connected by the relation: ξ = π1/2 и, and the approximating profile (2.13) can be represented in the form:

θ = (1 + π b и ) exp(− π β и )

(2.17)

It follows from the eq. (2.14), that b = 0,851, β = 1,851 (it is necessary to assume θ0 = 1). The comparison of exact and approximating profiles under constant thermal flow on a surface is given in the figure 2-1. It is obvious that noticeable distinctions of the profiles take place only on the distribution ‘tail’. In connection with the mentioned feature of the unknown functions behavior by small time for decreasing calculation error they were represented in the form: b = b0 τ−1/2, β = β0 τ−1/2, θ0 = θ00τ1/2, were the multipliers b0, β0, θ00 are time function generally speaking.

Figure 2-1. Temperature profile in semi-infinite body by constant thermal flow on a boundary.1 – dependence (2.15); 2 – approximation (2.13) by b = 0.851, β = 1.851.

Calculation Example As an example the calculation of heating and partial melting has been carried out for the sample of ceramics characterized by the following thermal properties: ρ = 1500 kg/m3; λ = λт = 0.6 W/m K; ε= 0.8; с=1200 J/kg K; L = 1.5 105 J/kg; Tm = 1900 K. For melting the ware the irradiancy in solar furnace must be above some minimum level which is defined by the condition qi min =εσTm4. For the given conditions qimin =0.59 MW/m2. As the qi decreases and approaches the qmin heating time increases exponentially. The data concerning the time of heating up to melting temperature t = tf. are tabulated in the table 2-1.

272

V.V. Pasichny and B.A. Uryukov Table 2-1. Time of heating stationary plate tf up to melting temperature and time of partial melting moving plate tm up to film thickness being equal to 0.5 and 1.0 mm

qi, MW/m2 tf, s tm(0.5), s tm(1.0), s

0.7 85.8 336 740

0.8 25.2 106 242

0.9 12.4 52.9 126

1.0 7.50 33.2 80.2

1.2 3.72 17.3 44.2

1.4 2.25 11.3 29.6

1.6 1.53 8.10 21.6

1.8 1.10 6.20 17.1

2.0 0.84 5.02 13.8

Figure 2-2. Dynamics of thickness of melt film under different irradiancies. 1 – qi = 0.7 MW/m2; 2 – qi = 0.8 MW/m2; 3 – qi = 0.9 MW/m2; 4 – qi = 1.0 MW/m2; 5 – qi = 2.0 MW/m2;

Figure 2-3. Dynamics of rate of melt film growth under different irradiancies. 1 – qi = 0.8 MW/m2; 2 – qi = 0.9 MW/m2; 3 – qi = 1.0 MW/m2; 4 – qi = 2.0 MW/m2;

The time dependences of the melt thickness are given in the figure 2-2 for different values of the irradiancy.

The Applying of Coatings and Surface Thermal Treatment…

273

The dynamics of the rate of the melt film growth is represented in the figure 2-3 for different irradiances. The rate of liquid film growth is maximal in the beginning of the melting process and then it decreases going into almost constant level.

2.2. Partial Melting of Moving Plate In the case when focal spot area is less than the area of the surface being treated a ware (plate) must be transported relatively the spot. The problem resides in the determination of the velocity of travel being sufficient for necessary level of treatment (necessary thickness of melt film). The elemental evaluation of this velocity V can be carried out on the basis of the results of the previous part. Let us assume that the process of heating for each section of the surface runs in the same way as the process of heating stationary plate by uniform radiant flux. In this case for being melted to the given depth x0 it must be under the action of radiant flux during the time tr being equal to the time of heating up to melting temperature tf and the time of partial melting tm: tr = tf + tm, The time tm is determined from integral equation because melting rate depends upon time: tm

∫V

m

dt = х0

0

The values of tm are given in the table 2-1 for х0 = 0.5 and 1.0 mm depending upon irradiancy. If focal spot diameter is equal to d for the points arranged along the diameter being parallel to the direction of motion: V = d/(tf + tm). The dependence V(qi) is demonstrated in the figure 2-4 for d = 1cm and х0 = 0.5 and 1.0 mm.

Figure 2-4. Dependence of velocity of ware movement relatively focal spot. 1 – х0 = 0.5 mm; 2 – х0 = 1.0 mm.

It is evident that at the periphery of the zone at which focal spot acts (in the direction being perpendicular to the direction of motion) the time of irradiation will be lesser than in the middle. Close to the spot boundary the surface sections will not be heated up to melting

274

V.V. Pasichny and B.A. Uryukov

temperature at all. Therefore it is advisable to determine ‘effective’ region of treatment. It can be limited by the condition that the surface section being treated must be under radiant flux action during the time which is not lesser than the time of heating up to melting temperature. It is not difficult to determine that the bandwidth of such ‘effective’ treatment hef will be:

⎛ tf ⎞ ⎟ hef = d 1 − ⎜ ⎜t +t ⎟ f m ⎝ ⎠

2

In the given particular case the value he/d held has turned out to be close to one (>0.98) because of strong distinction between tf and tm.

2.3. Energy Efficiency of Surface Treatment by Melting By energy efficiency of an ware treatment by partial melting of its surface we mean the relation of the energy expended directly for melting a material to total expended during some period of time. In this case the energy expended for heating up to melting temperature should be considered to be useful: tf ⎤ 1 ⎡ η ef = ⎢ ρ x0 L + ∫ [qi − ε σ T0 4 ]dt ⎥ qi t ⎢⎣ ⎥⎦ 0

In dimensionless variables: τf ⎤ 1⎡ η ef = ⎢ Вη 0 + ∫ [1 − С (θ 0 + θ a ) 4 ]dτ ⎥ τ ⎢⎣ ⎥⎦ 0

The change of the ηef in time is demonstrated in the figure 2-5 for different values of the irradiancy. The beginning of the each curve corresponds to the beginning of melting. As we can see the efficiency is maximum in the very beginning of the process and then decreases quickly especially under great irradiancies. Insufficiently high efficiency of a ware treatment by partial melting is stipulated by the fact that the main amount of thermal energy is expended for reradiation and a considerable part of the rest energy goes to solid mass. So directly only the small part of the energy being supplied is expended for melting. It is a result of the fact that specific expenditures of thermal energy under radiant heating are 3–4 times less than in industrial furnaces running on natural gas [5]. The efficiency of the process of materials treatment in solar furnaces under high temperatures can be increased essentially by reradiated energy returned onto the treated surface, at least partly, especially as the technical solutions of such a problem are possible in some cases in principal.

The Applying of Coatings and Surface Thermal Treatment…

275

Figure 2-5. Efficiency of thermal treatment under different irradiancies. 1 – qi = 0.7 MW/m2; 2 – qi = 0.8 MW/m2; 3 – qi = 0.9 MW/m2; 4 – qi = 1.0 MW/m2; 5 – qi = 2.0 MW/m2.

2.4. Conclusion 1. The approximate integral method of Tsyan for solving heat equation permits to solve the problems of dynamics of heating stationary and moving plane wares by radiant flux in solar furnaces including the process of a surface partial melting. 2. The velocity of the movement of ceramic ware surface being treated relatively focal spot, time and other characteristics of the process under the given level of thermal treatment (the thickness of melted layer and other) are determined. 3. Energy efficiency of surface thermal treatment of wares by melting decreases with the increase of treatment time.

DESIGNATIONS а – coefficient of temperature diffusivity (m2/s); b − form-parameter of approximating temperature profile; b0 − multiplier in the dependence of the b upon time; В − physical parameter of the problem; с – thermal capacity (J/kg K); d – focal spot diameter; hef – width of treatment bond under ware movement relatively focal spot; L – specific melting heat (J/kg); q, q0, qi – thermal flux, constant thermal flux, irradiancy (W/m2); t, tf, tm – time, time of heating up to melting temperature, time which is necessary for the partial melting of the surface up to the given film thickness (s); Т, Т0, Тт, Тa – temperature, temperature of plate surface or outer surface of melt film, melting temperature, ‘normal’ temperature (K); и – variable in the formula (15); v – dimensionless rate of melt film growth ;

276

V.V. Pasichny and B.A. Uryukov V – velocity of ware movement relatively focal spot (mm/s); Vm – rate of melt film growth (mm/s); х, x0, х1, х2 – coordinate, melt film thickness; arbitrary points; β − form-parameter of temperature profile approximation; β0 − multiplier in the dependence of the β upon time; ε – blackness degree of a surface; η0 – dimensionless thickness of melt film; ηef – energy efficiency of the process of ware surface thermal treatment by partial melting; λ, λт – coefficient of thermal conductivity of solid material, coefficient of thermal conductivity of melt (W/m K); θ, θ0, θ00 – dimensionless temperature, dimensionless temperature of plate surface or outer surface of melt film, multiplier in the dependence of the θ0 upon time; ωа − dimensionless ‘normal’ temperature; ρ – density (kg/m3); σ − Stefan-Boltzmann constant (W/m2K4); τ, τf,τm – dimensionless time, dimensionless time of heating up to melting temperature, dimensionless time which is necessary for the partial melting of the surface up to the given film thickness; ξ – dimensionless coordinate. Indexes: ef – efficient (эффективный); f – front (of melting); m – melting; i – irradiation.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 277-300

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 9

TRANSPARENT CONDUCTIVE LAYERS OF TIN, INDIUM, AND CADMIUM OXIDES FOR SOLAR CELLS* Yu.V. Vorobiev1, J. Gonzalez-Hernandez2†, P. Gorley3, V. Khomyak3, S. Bilichuk3, V. Grechko3 and P. Horley3 1

CINVESTAV, Unidad Queretaro, Queretaro 76230, Mexico 2 CIMAV, Chihuahua 31109, Mexico 3 Department of Electronics and Energy Engineering, Chernivtsi National University, Chernivtsi, Ukraine

ABSTRACT Transparent conductive oxides SnO2, In2O3-SnO2 (ITO) and CdO are widely used for different optoelectronic devices, including photovoltaic applications. Depending on technological conditions, oxide films can be either high- or low-resistive. This paper presents the results of complex investigation of technological parameters influence (such as chamber pressure, substrate temperature, magnetron cathode power, and duration of isothermal annealing in the air) on specific resistance and transmission coefficient of oxide thin films, grown by reactive magnetron sputtering. Ar-O2 mixture was used as a carrier gas for direct current sputtering; high-frequency sputtering was performed in pure Ar atmosphere. Substrates for the films were made of quartz glass and silicon. Significant attention was paid to the transformation of defect subsystems after isothermal annealing in the air. The authors determined optimal technological regimes allowing to obtain reproducible high-quality thin films of tin, indium and cadmium oxides with the following electrical and optical parameters: SnO2 – specific resistivity ρ = 6 – 15.10-4 Ω⋅cm, optical transmission T = 90 – 95% in transparency region; ITO – ρ = 4 – 6.10-4 Ω⋅cm, T = 90 – 95%; CdO – ρ = 5 – 20.10-4 Ω⋅cm, T = 80 – 90%.

*

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † On sabbatical leave from CINVESTAV-Queretaro

278

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

1. INTRODUCTION Recently, significant attention was paid to the technology and investigation of transparent conductive films of metal oxides, such as SnO2, ITO, and CdO, caused by wide application perspectives of these materials for optoelectronic devices: photodiodes [1-4], liquid-crystal displays [5], image sensors and switches [6,7], resistive ecological monitoring sensors measuring the concentration of toxic and explosion-hazardous gases in the atmosphere [8-10], devices with charge coupling [11], fiber optics coating [12], conducting transparent layers (windows) for solar cells, acting as effective anti-reflection coatings [13-16]. SnO2 and ITO oxide compounds are wide-band semiconductors (Eg = 3.5-4.0 eV [17]) with high conductivity and transparency in the visible spectral range. CdO has a somewhat narrower band-gap (Eg ≈ 2.4-2.5 eV [18]), but it can be successfully used along with SnO2 and ITO for solar cells applications (especially those based on CdTe) because of high transparency and low specific resistivity. Transmission coefficient T of all described SnO2, ITO, and CdO layers is rather high (T = 0.75 - 0.95); unfortunately, their electric parameters are characterized by significant discrepancies, even for the films grown with the same method but under different technological conditions [13, 18-31]. To our opinion, one of the possible explanations of this fact is connected with polycrystalline nature of the films. It is worth noting that polycrystalline grain size D (ranging from several nanometers (amorphous films) to dozens of nanometers [17, 18, 29, 32-37]) depends on both deposition and the post-processing techniques used. To obtain transparent films of conductive oxides, several methods can be successfully applied: ion sputtering in glowing direct-current discharge [18, 38-40], high-frequency ion sputtering [41-43], canon-ray sputtering [44, 45], pulverization with further pyrolysis [17, 21, 46-48], chemical deposition from the vapor [23, 49], reactive thermal evaporation [50, 51], laser pulse sputtering [52], direct current magnetron sputtering (DCMS) or radio-frequency magnetron sputtering (RFMS) [25-28, 53-59]. Among all the methods mentioned, vacuum reactive magnetron sputtering yields the best results [25-28, 56, 57, 59]. During sputtering process, magnetron plasma is kept in the target area at low pressure (around 5⋅10-2 Pa) with an additional magnetic field, forcing free electrons to move along the spiral trajectory, colliding with argon atoms more frequently. This increases discharge intensity and, as a consequence, leads to higher film deposition rates comparing to ion sputtering systems. As the substrate is situated beyond the area of active plasma influence, substrate heating and plasma particles bombardment are minimal, which leads to more homogeneous and less resistive films. At the same time, even for magnetron-sputtered layers, additional investigations are still necessary to find out how to obtain highly transparent and low-resistive metal oxide films. One of the possible solutions, indicated by previous investigations, lays in technology optimization of film deposition and further thermal processing. In this paper we are analyzing the published data and our own results to outline proper technological regimes for reproducible high-quality thin films of tin, indium and cadmium oxides, usable for different optoelectronic devices, including photovoltaics. The paper consists of three parts: the first one presents the analysis of the results obtained in previous investigations of thin metallic oxide films, mainly discussing their electrical and optical properties. The second part is devoted to the description of reactive magnetron sputtering

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

279

technology to obtain SnO2, ITO, and CdO. Our own experimental studies were performed for film specimens, deposited using specially modernized industrial set-up VUP-5, which allows to perform magnetron sputtering of high-quality films. In the third part, the authors analyze the influence of technological conditions during film deposition and further annealing on the properties of resulting SnO2, ITO and CdO films. In particular, it was shown that it is possible to control specific resistivity ρ and transmission T of the films under proper annealing regime, which result in the following film properties: SnO2: ITO: CdO:

ρ = 6 – 15.10-4 Ω⋅cm, ρ = 4 – 6.10-4 Ω⋅cm, ρ = 5 – 20.10-4 Ω⋅cm,

T = 90 – 95%; T = 90 – 95%; T = 80 – 90%.

The conclusion summarizes the discussion over main causes influencing transparency and specific resistivity of investigated layers of tin, indium, and cadmium oxides.

2. ELECTRICAL AND OPTICAL PROPERTIES OF METAL OXIDES (AN OVERVIEW) Analysis of numerous literature sources shows that in the majority of cases oxide layers were deposited by pulverization of organic solutions. Results of electrical and optical investigations of obtained films and discussion on their use in photovoltaics can be found in [17, 21, 46-48, 60-64]. Further on, several new methods were developed using different physical processes. In particular, Stjerna and Grangvist [19] successfully used reactive magnetron sputtering in atmospheres of О2 and Ar to obtain SnOx films with minimal resistivity ρ = 10-2 Ω⋅cm and high optical transparency (up to 75 %). They found an optimal proportion of carrier gas flow О2/Ar (Г*) and deposition rate (3 nm/s), which yielded the best resitivity and transmission values for SnOx films. It was shown that increasing of magnetron power P results in higher film deposition rate. For Г<Г* the films were yellowish, proving the formation of low-conductive SnO, while for Г>Г* the composition of the films was closer to SnO2, characterized with higher transmission. Authors of [20] obtained fluorine-alloyed SnO2 by the convection-sputtering method; the resulted films had optical transmission T ≈ 75 % and specific resistivity ρ ≈ 10-3 Ω⋅cm for temperatures 523-623 K and impurity concentration of 2 atomic percents. Kim and Chun [21] obtained SnO2 films by chemical evaporation using controlled oxidation of SnCl4. They mentioned the decrease of film specific resistivity for higher deposition temperatures; in particular, the films deposited at 773K have specific resistivity of 1.0-2.0⋅10-2 Ω⋅cm. Visual observation of film surface shown that films deposited at the temperatures lower than 773K have better transparency than those deposited at 873K and higher. In the paper [52] presented the results of microstructural and electrooptical investigations of the films, obtained by laser pulse deposition in vacuum and atmosphere of О2 at pressures 20.0 Pa for different substrate temperatures Тs. Depositing SnO2 films in О2 atmosphere with increase of Тs from 293 to 623 K gave specific resistivity of the material ranged from 200 Ω⋅cm to 2.0 Ω⋅cm, while the crystalline grain size was about 4-12 nm. Film transmission

280

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

was quite high (T = 75–90%). Band gap of film SnO2 determined from optical absorption spectra was about Eg= 2.6-4.0 eV depending on the value of Тs and carrier gas type. Authors of [59] obtained thin ITO films by reactive magnetron sputtering using both direct current (DC) and radio-frequency (RF) sputtering; they investigated influence of carrier gas proportion O2/Ar on resistivity and transparency of the films. For DC-sputtered films obtained under 1% O2/Ar ratio minimal specific resistivity was 2.5⋅10-3 Ω⋅cm, while their average transmission was T = 83%. For RF-sputtered films (O2/Ar proportion 4-5 %) corresponding parameters were ρ = 2⋅10-3 Ω⋅cm and T = 93 %. Eze [30] obtained CdO films with minimal specific resistivity ρ = 9.1⋅ 10-4 Ω⋅cm and maximum transmission T = 80 % using modified reactive vacuum evaporation in oxygen atmosphere with pressure 4.0⋅10-2 Pa. Authors of [18] fabricated CdO films with activated reactive sputtering in glowing-discharge plasma under different substrate temperatures in oxygen atmosphere. At certain substrate temperatures (Тs = 473K) films featured minimal resistivity ρ = 1.5⋅10-3 Ω⋅cm and maximum transmission of 85 %. Many investigators (e.g. [32-37, 65-67]) paid significant attention to investigation of film microstructure and its influence on optical and electrical parameters of the metal oxides obtained, especially for SnOx compounds. They found out that the increase of annealing temperature leads to grain size increase [32-34] and overall film resistivity decrease [34, 35]. According to diffusion-reaction model [65] it was assumed that the conductivity changes observed for SnO2 nano-crystals subjected to the action of air are caused by fast diffusion of atmospheric oxygen through the pores of film microcrystals with its further chemisorbtion at the surface of SnO2 grains. In the paper [37] it was suggested that film resistivity decrease observed after 5-hour constant temperature annealing was caused by increase of grain size D. As it was shown in the paper [36] according to the data of electron microscopy analysis, grains of polycrystalline SnO2 films solder together during annealing process, forming conductive chains lowering overall film resistivity. Authors of [21] basing on X-ray diffraction and scanning electron microscopy studies shown that SnO2 films deposited at Ts > 673 K are polycrystalline; at higher substrate temperatures (Ts ≈ 773K) obtained metal oxide layers are low-resistive and highly transparent. To achieve minimal specific resistivity of tin oxide and ITO films, special alloying with indium, boron or fluorine impurity is necessary [17,66,67]. Therefore, several different technological methods are used to obtain conductive and transparent thin metal oxide films (including various deposition methods, alloying of the initial materials with proper impurities, special post-annealing, etc.) Not all the methods applied are studied deep enough to make exact judgments regarding their optimization possibility to obtain better electrical and optical properties of the resulted films, essential for their photovoltaic applications. Our investigations [22, 28, 68-71] were focused on analysis of technology influence on magnetron-sputtered high-conductive and transparent thin unalloyed SnO2, ITO, and CdO films.

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

281

3. TECHNOLOGY FOR THIN SNO2, ITO AND CDO FILMS Thin films of SnO2, ITO and CdO were obtained using DC and RF magnetron sputtering (DCMS, RFMS), the former one in the atmosphere of Ar and O2. The latter method was used to deposit ITO films on the base of specially modernized industrial setup VUP-5. The magnetron itself was mounted in the window of the vacuum chamber (figure 1). Advanced external magnetron power source allowed controllable settings of direct cathode voltage and current in the ranges of 50-1000 V and 10-500 mA, respectively. For RFMS we used additional HF generator coupled with magnetron power source, operating with frequency 13.56 MHz and maximum power output of 50 W. The main advantage of our film deposition setup comparing to industrial magnetron VUP-5M was current stabilization of magnetron power source, allowing to avoid un-controlled cathode (target) discharge and keeping necessary current during deposition process. The authors also designed a special table for correct mounting and pre-heating of the substrate; we also performed general device optimization for most appropriate pressure control of carrier gas mixture, which allowed more flexible control over film deposition process. 4 1

5 6

3 2

7

50 – 1000В

A

V 9

8

Figure 1. Magnetron sputtering setup: 1– cooling agent (water); 2– magnetron; 3– vacuum chamber; 4– screen; 5– plasma; 6– substrate; 7– sputtered particles; 8– main body of vacuum setup VUP-5; 9– external magnetron power supply.

Total gas pressure in the chamber Pg was kept within the range of 0.1-10 Pa, using 5Npurity argon in carrier gas mixture. DCMS cathode was made of vacuum-synthesized 5Npurity tin, cadmium or 9:1 indium-tin alloy, forming the disk 60 mm in diameter. For RFMS the cathode of the same form and dimensions was made of pressed ITO ceramics. We used both glass and silicon substrates to deposit metal oxide films; the former to study their optical, and the latter to investigate their electrical properties and measure film thickness. Prior to deposition, glass substrates were cleaned in ethanol, silicon ones – in acetone. The substrate was located in some 5-7 cm away from the target. If necessary, substrates were preheated before the deposition to certain temperatures Ts (in the range of 300 - 473 K), kept constant till the end of the sputtering process. RF sputtering required pure argon plasma, formed when the pressure of carrier gas overcame some minimal threshold value. To perform film deposition at different pressures,

282

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

the chamber was first pumped out to gain the limit pressure to glow up the plasma, and then Pg was adjusted to the required value, as plasma was self-maintaining in all the pressure ranges studied. Optimal cathode power for DCMS was determined experimentally to be within the range of 7.5-38 W (for the voltages 150-250 V and currents 50-250 mA). For the case of RFMS, cathode power was kept constant and equal to 15 W. Our investigations have shown that higher values of cathode power can result in local melting of target material, changing stoichiometric composition of the film. In general, film deposition rate was about 0.8 - 60 nm/min depending on target material and technological conditions. As we know [72], there exists critical cathode power for reactive sputtering, overcoming which it is impossible to ensure complete metal oxidation even in the atmosphere of pure oxygen. That is why it is possible (depending on partial oxygen pressure and discharge power) to deposit oxide mixtures with different content of Cd, In, Sn and O2, deviated in composition from stochiometric CdO, SnO2, and ITO. For example, increase of oxygen content in the carrier gas improves film transparency, but makes electrical properties of SnO2 and ІТО worse. Investigations have shown that DC-sputtered ITO films have better parameters if carrier gas is composed of 79% Ar and 21% of O2. Film thickness was estimated with interference Linnick microscope MII – 4, yielding different values depending on parameters of film deposition technology (target material, carrier gas pressure, cathode power, substrate temperature, deposition time, etc.). Surface morphology of the films was investigated using X-ray Diffraction (XRD), Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM) at Queretaro Branch of research centre CINVESTAV (Mexico). Investigations of their electrical properties included four-probe surface resistivity measurements. Optical transparency (transmission) of SnO2, ІТО, and CdO films was investigated with spectral photometer SF-20 at room temperatures in wavelength range 250 –1000 nm. To determine the influence of the temperature on electrical properties of the films, they were isothermally annealed for 10 minutes at 323 – 773 K in the air. For the sake of comparison, the samples were divided into two groups consisting of the samples obtained under the same conditions. First group was subjected to multiple 10-minutes long annealing stages, increasing temperature to 773 K in 50 K steps and measuring the resistivity of the sample after each stage. The second group was annealed only once during 10 minutes under constant temperature with further resistivity measurement. As it follows from the equal values of specific resistivity for both sample groups (within the measurement precision), the annealing time considered was sufficient to establish all the stationary states in the material.

4. RESULTS AND DISCISSION 4.1. Thin SnO2 films Thin SnOx films obtained were slightly yellowish, very clear and transparent, with smooth mirror-like surface characteristic to tin dioxide [73]. The technological conditions during film deposition are summarized in table 1. Film deposition rate was 1.0–13.0 nm/minute. All the films obtained had electron conductivity.

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

283

As one can see from the table 1 and figure 2, dependence of deposition rate on chamber pressure has a maximum at certain optimal pressure (samples 3 and 9). At lowest and highest pressures investigated, deposition rate were the smallest (e.g., samples 1, 2 and 6). It also diminishes with temperature increase (samples 3, 5, 10, 11), but can be increased if the magnetron current becomes greater (samples 3, 4, 9, 10). Film thickness depends on the deposition time if the other conditions are kept invariable (samples 1, 2). With carrier gas pressure increase the adhesion of the film to non-pre-heated substrate becomes worse and deposition rate shorter (samples 6, 7). Under low chamber pressure (0.13–0.4 Pa) adhesion of the films are quite high even if the substrates were not pre-heated, but the magnetron power has to be reduced (samples 5, 11). In the opposite case the surface of the substrate gets the metallic speckle because of deposition of tin atoms that escaped oxidation. Table 1. Technology of SnO2 film growing by means of reactive magnetron sputtering Sample No.

1 2 3 4 5 6 7 8 9 10 11

Deposition time min

Substrate temperature Ts, K 300 300 300 300 300 300 473 473 473 473 473

Magnetron current Im, mА 100 100 100 70 70 100 160 160 160 100 70

30 15 30 15 60 30 30 30 30 30 60

Chamber pressure Pg, Pa 0.67 0.67 1.46 1.46 0.13 7.98 7.98 0.67 1.46 1.46 0.13

Film thickness, Nm 150 80 400 140 130 100 130 180 270 200 70

Deposition rate Vd, nm/min

14 Ts = 300K, Im = 100mA Ts = 473K, Im = 160mA

12 10 8 6 4 2

0

1

2

3

4

5

6

7

8

Carrier gas pressure, Pg

Figure 2. SnO2 film deposition rate depending on Ar/O2 carrier gas pressure.

9

Deposition rate Vd, nm/min 5.0 5.0 13.0 9.0 2.2 3.3 4.0 6.0 9.0 6.7 1.2

284

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

Measurement of specific resistivity of the films immediately after formation yields the values ρ = 1–200 Ω·cm, failing quality requirements for conductive covers. The subsequent isothermal annealing of the films leads to significant resistivity drop. After annealing at 773 K specific resistivity decreases in about four-five orders of magnitude, reaching the values 1.5·10-3–6·10-4 Ω·cm. SnO2 films featured high parameter reproduction; after heating and cooling in the temperature ranges investigated film properties were stable and remained constant after a year of storage under normal conditions. As it was shown, sputtering rate and surface adhesion of SnO2 films depend significantly on integral pressure in the chamber (see table I.) Lower deposition rates at higher pressure (samples 1, 3, 6-9) can be attributed to increase of multiple collisions of sputtered atoms with those of carrier gas, which scatter the metal particles and decrease the total flux of the atoms moving towards the substrate. Moreover, the atoms show lower adhesion upon reaching the substrate if they lose their kinetic energy even partially. Decrease of deposition rate under lower pressures is caused by the lack of the positively charged ions of the carrier gas. Metallic layer of tin atoms forming at low pressure is caused by insufficient amount of oxygen atoms, leaving a portion of tin atoms without oxidation. The composition of SnOx film in this case will deviate from stoichiometric value for SnO2 towards the oxides of lower valence. This assumption is proved by dark yellow hue characteristic to SnO [73,74] if the films are grown at low chamber pressure (0.13 Pa) and magnetron current 250 mA. Figure 3 shows an example of annealing effect on film resistivity under various temperatures Ta for two films deposited on substrates with different substrate temperature (samples 1 and 10, according to table I). All the samples studied featured similar curves for ρ = f (103/Ta) with slight variations depending on substrate temperature while film deposition.

103

35eV EA=0.1

Resistivity ρ, Ω.cm

102 101 E

100 10-1

= EA

10-2 10-3

V 0e 5 . 0

eV 70 . =0 A EA=0.075eV

V 7e 4 . =0

EA

V 30e =0.

EA

10-4 1.0

0. E A=

V 30e

1.5

2.0

103/T

2.5 a,

K-1

Figure 3. Specific resistivity of SnO2 versus annealing temperature.

Sample 1 Sample 10 3.0

3.5

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

285

As it follows from the figure 3, all the interval of temperatures studied could be divided into several parts, in which the dependence ρ (Тa) has activation character and could be described by the formula [75]: ρ (Тa) = 1 / (e n μeff) ≈ f (Ta, m, r).exp( EA / kTa ),

(1)

Here n and μeff are concentration and effective mobility of the carriers, f (Ta, m, r) is a function depending on annealing temperature Ta, material parameters m and dominant carrier scattering mechanism r; EA is the activation energy; other designations are common. As it is seen from the figure 3, films deposited on the substrate without pre-heating (sample 1) under low annealing temperatures (300 K < Ta < 330 K) feature activation energy EA = 0.135 eV. Good correlation of the latter value with formation energy for oxygen vacancy in SnO2 (EV = 0.130 eV [76]) makes sound the assumption that namely oxygen vacancy ionization leads to comparably small lowering of specific resistivity of the SnO2 films in this case. For annealing temperature 350 K < Ta < 470 K activation mechanism changes, which is reflected in significantly greater activation energy EA = 0.69 ± 0.01 eV, suggesting that oxygen divacancies are formed in the film or that these vacancies (or divacancies) become trapped by impurity atoms, resulting into complex associates. Hence, activation energy will have significant value, being a sum of vacancy (or divacancy) migration energy and associate formation energy. The associates appeared in the material are instable formations, dissociating under further temperature increase (470 K < Ta < 700 K), which leads to appearance of new defects with EA = 0.50 ± 0.01 eV. It is peculiar, that for Ta > 730K another defect subsystem with EA = 0.30 ± 0.06 eV becomes dominant. Kinetics of annealing processes for the films deposited on pre-heated substrate (473K, sample 10, figure 3) differs from the previous case. First, for 300 K < Ta < 370 K activation energy is significantly lower and is equal to EA = 0.075 ± 0.005 eV. It worth noting that this value is quite close to the height of inter-crystallite barriers Vd in the films of other semiconductor materials (e.g. for CdS the value of Vd is about 0.06-0.2 eV [75]). It makes it possible to assume that decrease of specific resistivity of SnO2 films, deposited over preheated substrates is caused by the crystallite grains forming in the material studied. It is important that there are no traces of complex associates formation in the films deposited over pre-heated substrate, which lowers defect level of the material and significantly reduces its specific resistivity. It is also curious that activation energies EA = 0.47 ± 0.03 eV and EA = 0.30 ± 0.04 eV, determined for Ta > 470 K, show no dependence on the substrate temperature during film deposition, at least within experiment precision. Further annealing time increase and multiple annealing does not change SnO2 resistivity in any significant way. Analysis of transmission spectra of SnO2 films (figure 4) shows that isothermal annealing at 473K for ten minutes improves their transparency to 90 – 95% (for example, figure 4 curves 1 and 1') and makes steeper the slope of intrinsic absorption edge. The latter becomes steeper also if the film deposition had carried out at higher substrate temperatures (samples 8 and 10). Such a behavior of transmission coefficient could be explained by specific transformations of defect subsystem during SnO2 film deposition on pre-heated (T = 473 K) substrate. Band gap estimations made from spectral dependence of transmission coefficient give the values of 3.54 – 3.75 eV, which corresponds quite well with SnO2 data, published in different sources [17, 77, 78].

286

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

1.0

Transmission, d.u.

10

8 1' 1

0.5

0.0 200

400

600

800

1000

Wavelength, nm Figure 4. Optical transmission of SnO2 film (curves 1, 8, 10 – data for samples1, 8 and 10 (see table I), curve 1' – data for sample 1 after annealing at 473 K for 10 min).

4.2. Thin ITO Films Analysis of the experimental results allowed us to determine the optimal technology to grow high conductive and transparent ITO films with good adhesion to the substrate. Table 2 summarizes main technological film deposition regimes. All the films obtained had electron conductivity. As it follows from the table 2, film deposition rate Vd depends on the temperature of the substrate Ts, carrier gas pressure Pg and cathode power Pc. In the figure 5 deposition rate Vd is plotted versus Pc for different Ts and constant Pg. As one can see, for all the values Pc and Ts studied this dependence is linear in the terms of measurement precision: Vd = C . Pc,

(2)

with coefficient C decreasing linearly with substrate temperature growth (inset to the figure 5), which could be explained by increase of back-evaporation of the material from the heated substrate [72]. On the other hand, deposition rate Vd = f(Pg) at constant Pc and Ts has a form of the curve with a maxima (figure 6) characteristic to dynamic systems with damping, which obeys the expression Vd = C1 . Pg . Py / ( (P02 – Pg2 )2 + Pg2 Py2 ),

(3)

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

287

Table 2. Technological parameters for ITO films obtained by magnetron sputtering Sample No.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Cathode power Pc, W

Substrate temperature Ts, K

300 300 300 300 300 300 300 300 300 373 373 373 373 373 373 473 473 473 473 473 473 473

7.5 7.5 7.5 7.5 15.0 15.0 15.0 15.0 38.0 7.5 7.5 7.5 7.5 15.0 38.0 7.5 7.5 7.5 15.0 38.0 38.0 15.0

Film thickness, nm

0.13 0.5 0.8 1.5 0.13 0.5 0.8 1.5 0.8 0.13 0.5 0.8 1.5 0.8 0.8 0.13 0.8 1.5 0.8 0.8 7.8 0.8

400 490 540 620 450 520 600 710 980 300 320 400 400 650 1000 250 250 250 600 1000 100 550

Deposition rate Vd, nm/min

15

40 44 46 52 35 39 41 47 26 20 21 23 27 20 11 8 9 10 7.5 5.5 9.5 4.8

0.6

1

0.4

2

0.2

10

300

400

TS, K

Specific resistance ρ, 10-2Ω⋅cm

1.5 3.5 4.4 3.7 2.3 6.1 9.0 7.2 20.0 1.1 2.3 3.0 2.6 6.0 16.5 0.8 1.8 1.5 5.3 11.5 1.2 11.0

C, nm min-1 W -1

Deposition rate Vd, nm/min

20

Chamber pressure Pg, Pa

500

Surface resistance ρ, Ω/□

10000 8980 8519 8387 7778 7500 6833 6620 2653 6667 6562 5775 6750 3077 1100 3200 3600 4000 1250 916 9500 850

3

5

0

0

10

20

30

40

Power Pc, W Figure 5. Dependence of ITO film deposition rate Vd for different substrate temperatures Ts (1 – 300 K; 2 – 373 K; 3 – 473 K) on the cathode power Pc for direct current reactive magnetron sputtering and pressure of carrier gas Ar/O2 = 0.8 Pa. Inset: dependence of Vd on Ts..

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

Deposition rate Vd, nm/min

20

1

16

2

12

3

C1, nm min-1 Pa2

288

300

200 300

8

400

500

TS,K

4 0

0

2

4

6

Carrier gas pressure Pg, Pa

8

Figure 6. Dependence of ITO film deposition rate Vd for different substrate temperatures Ts (1 - 300K; 2 – 373K; 3 – 473K) on the pressure of carrier gas Ar/O2 in the chamber, magnetron cathode power Pc = 38 W. Inset: plot of constant C1 from (3) versus Ts.

with damping coefficient Py = 10.17 Pa and parameter P0 = 3.23 Pa, analogous to intrinsic gas pressure. The damping in the system considered may appear due to scattering processes taking place for the sputtering atoms over carrier gas atoms or other sputtering atoms. Coefficient С1, corresponding to the peak value of Vd plot, appears decreasing linearly with increase of Ts (inset to figure 6). Physical mechanisms leading to damping could be connected with increase of multiple collisions and scattering of the sputtered atoms over carrier gas atoms in the chamber, and also by lowering of general atomic flow towards the substrate. It is worth noting that under pressure increase Pg > 1.1 Pa film adhesion to room-temperature substrates also becomes lower, which could be addressed to partial decrease of kinetic energy of sputtering atoms after the scattering. Under the lower pressures for Pg < Pgmax (where Pgmax is the pressure corresponding to a maximum of Vd(Pg) curve), one can observe deposition rate decrease, caused by the lack of positive charged ions of the gas necessary to keep with phase equilibrium. Under low pressures film adhesion is rather high even for room-temperature substrates, but in this case one has to decrease cathode power to avoid deposition of nonoxidized In and Sn atoms due to the lack of oxygen, causing metallic shine of the film and significant deviations from stoichometry. Analysis of experimental data from the table 2 shows that the films grown under the constant cathode power have rather high resistance (ρ ≥ 0.06 Ω.cm), which increases slightly with carrier gas pressure Pg and depends inversely on substrate temperature Ts (figure 7). The similar dependence ρ = f(Pg) was observed also in [79]. Significant improvement of film electric properties upon isothermal annealing was reported in [80, 81]. Figure 8 shows characteristic curves of specific resistivity ρ versus annealing temperature Та for the samples 7, 14 and 19 (table 2), obtained under the same Pg and Pc, but different Ts. Our investigations proved that 10-minute annealing under constant temperature is completely enough for establishing equilibrium state in the films, judging from equal values of specific resistance obtained for both groups of the samples.

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

289

Specific resistivity ρ, Ω⋅cm

100

1

2 10-1

0

2

4

6

8

Ar/O2 carrier gas pressure Pg, Pa Figure 7. Dependence of ITO specific resistance ρ on carrier gas pressure Pg for different substrate temperatures (1 – 373K, 2 – 473K) and cathodepower Pc = 15W.

100

14

19

10-1

10-2

10-3

10-4 200

Activation energy EA, 10-2eV

Specific resistivity ρ, Ω⋅cm

7

18

1

15

2

12 9

300 400 500 TS, K

300

400

500

600

700

800

Temperature of annealing TA, K Figure 8. Specific resistance of ITO films versus annealing temperature (curve numbers correspond to the sample number in the table 2). Inset: constants Ea (1) and E1a (2) versus Ts.

As it follows from the figure 8, the character of ρ=f(Ta) is significantly different for given Ta, making it possible to divide all the temperature range investigated into three parts. Lowtemperature annealing (300 K < Ta < 473 K) is characterized with slight changes of ρ with increase of Ta. This dependence becomes even weaker for higher substrate temperatures. Moderate-temperature annealing (473 K < Ta < 600 K) leads to fast decrease of resistance by several orders of magnitude. Upon reaching the minimal value (4 – 9)⋅10-4 Ω⋅cm at nearly 530

290

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

K, ρ starts to grow and at Ta=600 K it becomes equal to (1–2)⋅10-3 Ω⋅cm. High-temperature annealing (600 K < Ta) features slight resistance increase with Ta. To explain possible mechanisms causing these changes of resistance behavior, one have to keep in mind that ITO films obtained were amorphous according to the XRD and AFM surface structure investigations (figure 9). In this case, as it was discussed in [82], changes of specific resistance of the film in wide temperature ranges could be described by

ρ = ( A exp( -∆E / kT ) + A1exp( -∆E1/kT ) + A2exp( -∆E2/kT ) )-1,

(4)

with constants A, A1 and A2 determined by technological parameters during film deposition but independent on the temperature, and different activation energies ∆Е, ∆Е1 and ∆Е2. Three items in right hand side of (4) describe partial contributions to current transport by different groups of carriers: those at delocalized states, excited carriers at the edge of allowed energy band (i.e. close to the energy of donor and acceptor levels), and by electrons contributing to hopping conductivity between localized states close to Fermi level. As it is generally accepted, third mechanism is dominating only for low temperatures, while two others – for higher ones [83]. Nevertheless, results of experimental investigations for amorphous silicon and germanium proved the hopping conductivity to be significant in the wide temperature range of 40-400K. Annealing could lead to the changes of dominating mechanism of conductivity (for the case of amorphous germanium, one can observe hopping conductivity prevailed by band conductivity [83]), resulting in changes of specific resistance. Moreover, analyzing influence of the annealing in the air over ρ of ITO films obtained, one may also have in mind presence of stoichiometric and molecular oxygen. The former already exist in crystalline lattice and forms complex associates upon annealing [84], while the latter could be absorbed by the surface of the film and undergo further chemisorption by trapping of conductivity band electrons, which will result in formation of depleted layers at the surface [66]. It is important, that with increase of annealing temperature, initially amorphous ITO could be gradually turned into monocrystalline material with grain size depending on Ta. This effect triggers formation of carrier-depleted, inverted or enriched areas at grain boundaries resulting in different kinds of potential barriers [83], depending on the nature of subsurface states and energy distribution of surface levels. Such a complex mechanism determining specific resistance could be properly considered only with corresponding complicated physical model, which will require separate research work. Here we will present only qualitative discussion of the data presented in figure 8. The low-temperature annealing changes of specific resistance ρ with temperature are caused mainly by activation mechanism with activation energy Еа, obeying to [83]

ρ = 1 / σ = 1 / ( e n μeff ) = A0 ( Та / 300 )3/2 exp( Ea / kTa ),

(5)

where n is carrier concentration and μeff - effective mobility, e - absolute value of elementary charge, and independent on the temperature normalization constant A0. In (5) it was taken into account that for the temperature range investigated the most probable carrier scattering mechanism involve acoustic phonons. Calculation results obtained with (5) correlate well with experimental data (dashed lines in figure 8, plotted for three different substrate

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

291

temperatures Ts). Activation energy Ea linearly decreases with increase of Ts (curve 1 at inset to the figure 8), which could be connected with different concentration of oxygen vacancy acting as donor. Obtained values 0.10 ≤ Ea ≤ 0.18 eV, correlating well with formation energy of oxygen impurity EF = 0.13eV in SnO2 [76] could be considered as qualitative proof of this assumption.

Figure 9. Surface morphology AFM data for ITO film deposited on room-temperature substrate.

For high-temperature annealing ( 600 K < Ta ) experimental results fit well the formula

ρ = A exp ( -E1a / kTa ),

(6)

with normalization constant A and parameter E1a, which most probably describes the difference of energies corresponding to formation of different structural defects, such as vacancies, interstitial atoms, associates, etc. Dash and dot lines in figure8 represent calculation data obtained according to (6). Comparing (5) and (6), one may notice that physical processes taking place in low- and high-temperature annealing modes are significantly different. It is also proved by the fact that the value of E1a increases with Ts (curve 2, the inset to figure 8) on the contrary to the behavior of energy Ea for lowtemperature annealing. As formula (6) has no coefficient depending on temperature by the exponent, one can assume that carrier scattering is connected mainly with neutral impurities or their clusters for high-temperature annealing. One of the possible explanations of ρ = f(Ta) behavior after annealing at 600K < Ta could be either crystallization processes leading to increase of grain size or qualitative changes in defect sub-system. Our experiments have shown that specific resistance of the films in moderate-temperature annealing mode (473 K < Ta < 600 K) could not be described as superposition of (5) and (6). It means that to determine the nature of physical and chemical processes taking place at these annealing temperatures it will be necessary to apply methodology developed for chemical kinetics and non-equilibrium thermodynamics [83]. Construction of appropriate physical model will allow finding functional dependence of ρ on technological conditions during film deposition, opening ways to obtain ITO films of maximal possible conductivity. At the moment we have only determined optimal technological regimes for magnetron-sputtered

292

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

low-resistive ITO on the base of analysis of experimental data. Our films featured reproducible parameters and were characterized with specific resistance ρ = 4 - 6⋅10-4 Ω⋅cm, being among the best results published [83, 86, 88]. Investigation of optical properties of obtained ITO films proved their high transparency (visible spectra transmission for the best samples 90-95%, figure 10). For the sake of comparison with ITO films annealed in the vacuum figure 10 features experimental data from [87]; as one can see, our best films has a little higher and homogeneous transmission for visible light. Insignificant deviation of transmission slope at the intrinsic absorption edge between our films and data of [87] could be caused by difference of their thickness. Experimental results also shown a correlation between electric and optical parameters of ITO films obtained – samples with smaller minimal specific resistance (figure 8) has higher slope of transmission curves at intrinsic absorption edge (figure 10). At the same time films deposited on room-temperature substrates (i.e. sample No. 7) are characterized with less steep absorption edge than those deposited on pre-heated substrates (samples No. 14, 19, and 21). Band gap of ITO films estimated from figure 10 was about 3.7eV, which agrees well with previously published data [17, 72, 89-91]. 100

Optical transmission, d.u.

19 80

14 60

22

7

40 20 0

400

600

800

1000

Wavelength λ, nm Figure 10. Optical transmission of ITO films (curve numbers 7, 14, and 19 corresponds to the sample numbers from the table 1; sample 14 was additionally annealed at 523K, 19 – at 773K, 22 – high frequency magnetron sputtering, squares corresponds to the data from [87]).

4.3. Thin CdO Films Thin CdO films are usually obtained by pulverization with further pyrolysis or either by growing from the melt [17, 92, 93]. However, these methods do not allow to obtain films with high transparency and conductivity or to ensure their stoichiometric homogeneity, causing moderate or low efficiency of solar cells with CdO films grown with either of these two methods. Results of our investigations prove that magnetron sputtering is a promising method to obtain high-conductive transparent CdO layers, suitable for photovoltaic applications.

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

293

Deposition rate Vd, nm/min

Main technological data and properties of DC magnetron sputtered CdO films are presented in the table 3 and figures 11-13. Our CdO films were transparent and continuous, with smooth mirror-like surface and good film adhesion to the substrate. Thermal probe measurements proved electron conductivity type of all the films obtained. As it follows from the figure 11 and table 3, film deposition rate decreases with substrate temperature increase; this effect is most pronounced for maximum constant magnetron current and maximum pressure of carrier gas (Pg = 1 – 2 Pa), for which deposition rate Vd is the highest (curves 1-3 and 4-6). 60 50

1

40

2

30

4

20

5 3

10

6 300

350

400

450

500

Substrate temperature Ts, K

Deposition rate Vd, nm/min

Figure 11. CdO film deposition rate depending on substrate temperature for different carrier gas pressures Pg and magnetron currents (curves 1, 2, 3 – Рg = 1.5 Ра, Іm = 130, 100, 50 mA; curves 4, 5, 6 – Pg = 6.0 Ра, Іm = 130, 100, 50 mA).

60

1

50

2

40

3 4 5 6

30 20 10 60

80

100

120

140

Magnetron current Im, mA Figure 12. CdO film deposition rate versus magnetron current (curves 1, 3, 4 – Тs = 300 K, Pg = 1.5, 0.13, 6.0 Pa; curves 2, 5, 6 – Тs = 473 K, Pg = 1.5, 0.13, 6.0 Pa).

294

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

Table 3. Technology and parameters of DC-reactive magnetron sputtered CdO films Sample No.

Substrate temperature Тs, K

Magnetron current Im, mA

Carrier gas pressure Pg, Pa

Film thickness, nm

Deposition rate, nm/min

Specific resistivity ρ×10 2 Ω⋅cm

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

300 300 300 300 300 300 300 300 300 300 300 373 373 373 373 373 373 473 473 473 473 473 473 473 473 473 473 473

50 50 50 50 50 50 50 100 100 100 130 50 50 50 100 130 100 50 50 50 50 100 100 100 130 130 130 130

0.13 0.5 0.8 1.5 2.5 3.5 6.0 0.13 1.5 6.0 1.5 0.13 0.8 1.5 1.5 1.5 6.0 0.13 0.5 1.5 6.0 0.8 1.5 2.5 0.8 1.5 3.5 6.0

550 750 880 950 875 800 1300 1175 1500 1075 1800 520 725 850 760 750 925 475 550 575 425 440 500 525 450 550 500 575

14 19 22 24 22 20 13 29 47 27 60 13 18 21 42 54 23 12 14 17 10 34 38 35 39 48 38 27

6.12 8.21 9.40 12.10 13.78 13.48 14.60 4.75 5.80 5.95 5.52 0.70 0.75 0.80 0.63 0.59 0.68 0.17 0.18 0.22 0.24 0.19 0.21 0.22 0.20 0.25 0.24 0.27

Higher magnetron power (or current) and lower deposition temperature also increase film deposition rate (figure 12). The latter also depends on pressure of carrier gas (figure 13), but, as it could be seen comparing figures 13, 15, table 3 and data presented in figures 2, 5, 6 and tables 1, 2, Vd(CdO) is significantly higher than Vd(SnO2, ІТО) under the same technological conditions. The peak of maximum deposition rate for CdO is blurred and shifted towards higher pressures (Pg = 1 – 2 Pa). Investigated technological regimes made it possible to change film deposition rate in wide ranges of 10-60 nm/min.

Deposition rate Vd, nm/min

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

295

60 50 40 1 30

2

20

3 4 5

10 0

0

1

2

3

4

5

6

Carrier gas pressure Pg, Pa Figure 13. CdO deposition rate depending on carrier gas pressure for different substrate temperatures Тs and magnetron currents Іm (curves 1, 4 – Тs = 300 K, Іm = 130 i 50 mA; curves 2, 3, 5 – Тs = 473 K, Іm = 130, 100, 50 mA).

Specific resistivity ρ, Ω⋅cm

10-1

10-2

9

15 23

10-3

300

400

500

600

700

800

Temperature annealing Ta, K Figure 14. Specific resistivity of CdO films obtained for different substrate temperatures Ts as a function of annealing temperature TA (curve numbers correspond to sample numbers in the table 3).

Typical specific resistivity curves for CdO films are presented in figure 14 for the samples 9, 15, and 23, grown under different Ts (see table 3), with subsequent annealing in the air at Ta = 323 – 773 K. This dependence of resistivity on annealing temperature was observed for all the films obtained. For Ts= 300 – 473 K, resistivity of the films remains within the ranges of 5.8⋅10-2 – 2.1⋅10-3 Ω⋅cm. With increase of annealing temperature specific resistivity decreases at first in the temperature range Тa = 300 – 573 K, reaching the minimum ρ = 5⋅10-4 Ω⋅cm. The samples, deposited on the non-pre-heated substrate (curve 9) show higher resistivity deviation comparing to the samples deposited on pre-heated substrates (curves 15, 23). With further increase of annealing temperature Тa > 573 K specific resistivity of the resulted films shows no particular dependence on the latter. Average temperature

296

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

Optical transmission, d.u.

coefficient of resistivity dρ / dTa for the samples 9, 15, 23 are 2.1⋅10-4, 0.2⋅10-4, and 0.07⋅104 Ω⋅cm/K, respectively. Almost the same values of dρ / dTs coefficient are characteristic for non-annealed samples (e.g. curves 9, 15, and 23 for ТА =300 K and Тs = 300 – 473 K). Resistivity values obtained correlate well with other published data [88, 89]. Resistivity decrease with higher substrate and annealing temperatures can be most probably connected with film structure improvement due to grain size increase, and therefore decrease of the area of inter-barrier crystalline phases. These factors diminish the concentration of traps and interbarrier scattering over grain boundaries which, in turn, increase carrier mobility and conductivity. The similar behavior of resistivity is characteristic to transparent conductive SnO2 films (see 4.1). Transmission spectra of CdO films obtained for different substrate temperatures or after annealing are presented in figure 15. Films, which were grown under room temperatures (curve 9) feature rather low transmission (Т ≈ 65 %) and small intrinsic absorption slope, which is caused by their low structural perfection and possible presence of un-oxidized cadmium in the films, which was proved on the base of X-ray investigations in the paper [18]. 1,0 0,8 0,6

23 a 15′

15

9

9′

0,4 0,2 0,0 400

500

600

700

800

900

Wavelength λ, nm

1000

Figure 15. Optical transmission of CdO films (curves 9, 15, 23 – corresponds to sample numbers from the table 3; curves 9′, 15′ – data for annealed samples 9, 15 (TA= 573 K); curve а – transmission data from [18]).

Films deposited at temperatures 373 and 473 K are characterized with higher transmission (Т = 80 – 85 %) with abrupt absorption edge, insignificantly shifted to the shorter wavelengths. For comparison, transmission data for CdO films according from [18] for Тs = 523 K are plotted in figure 15 (curve a), showing good correlation with the data obtained by the authors. Film annealing (at temperature Тa characteristic to minimum film resistivity) leads to increase of transmission coefficient and steeper transmission curve (compare figure 15, curves 9, 15 and 9′, 15′). For the sample 15′ one can observe the interference pattern, not presented in the figure 15 to avoid visual overloading. Transmission of the samples, deposited on room-temperature substrates, can be significantly improved by annealing (from 60-65 to

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides…

297

75-80 %), while for the samples grown with Тs = 473 K the slope of transmission curve and transmission coefficient itself do not feature significant changes (Т = 80-90 %), while their interference peaks become sharper (e.g., for the film 23 transmission prior and after annealing are practically the same). Improvement of optical properties of CdO films deposited on preheated substrates or after isothermal annealing in the air can find reasonable explanation by improvement of their crystalline structure, proved by the character of the curves ρ = f(Ts) and ρ = f(TA), as well as by the presence of sharp interference pattern for the samples, obtained for Ts > 373 K, as well as those annealed at Тa = 573K. Presence of well-defined interference peaks made it possible to estimate film thickness using the following formula [95] d = λ1λ2 / [4n (λ1 – λ2)],

(7)

with CdO refraction coefficient n = 2.49 [96] and position of two subsequent maximums or minimums λ1 and λ2. Calculated film thickness agrees well with the thickness measured on MII-4 setup. Band gap value, estimated from the optical transmission of the films grown under different technological conditions was approximately 2.2 – 2.45eV. For higher substrate temperatures (figure 15, curves 9, 15) or after annealing (curves 9, 9′ and 15, 15′) the band gap decreases, which can be caused by the Burstein-Moss effect [17].

5. CONCLUSION Results of current investigations prove the possibility to obtain high-transparent and conductive thin metal oxide films with magnetron sputtering and further controlled annealing. Direct-current and radio-frequency magnetron-sputtered films were obtained with specially modernized industrial setup VUP-5 allowing more flexible control over magnetron power, carrier gas pressure, substrate temperature, etc. Optimal film deposition and annealing regimes were determined empirically, resulting in thin films with the following parameters: SnO2 – specific resistivity ρ = 6 – 15.10-4 Ω⋅cm, optical transmission T = 90 – 95%; ITO – ρ = 4 – 6.10-4 Ω⋅cm, T = 90 – 95%; CdO – ρ = 5 – 20.10-4 Ω⋅cm, T = 80 – 90%. Further improvement of these parameters (i.e. resistivity decrease keeping high transparency), to our opinion, will be possible when the physical processes taking place during film deposition and annealing will be better understood, which will take place upon development of proper theoretical models describing metal oxide layer formation. This model has to consider the decrease of inter-crystalline boundaries caused by grain size increase during film deposition and further annealing, which may lead to decrease of their influence on electrical and optical properties of the films. To our point of view, it is impossible to obtain better high-conductive and transparent films of SnO2, In2O3-SnO2 (ITO) and CdO films without additional investigations (including ellipsometric data), which will allow for the determination of the influence of grain boundaries, and hence, to enhance technology optimization to obtain thin metal oxide layers with better electrical and optical parameters.

298

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

[29]

S.M. Tadayyon, K. Griffiths, P.R. Norton, C. Tripp, Z. Popovic, J. Vac. Sci. Technol. A 17 (1999) 1773. M. Powell, C. Classe, I. French, A. Franklin, J. Hughes, J. Curran, Mater. Res. Soc. Symp. Proc. 467 (1997) 863. J.S. Kim, M. Granstrom, R.H. Friend, N. Johansson, W.R. Salaneck, R. Daik, W.J. Feast, F. Cacialli, J. Appl. Phys. 84 (1998) 6859. J.S. Kim, R.H. Friend, F. Cacialli, Appl. Phys. Lett. 74 (1999) 3084. J. Lan, J. Kanicki, Mater. Res. Soc. Symp. Proc. 424 (1997) 347. T.Hirayashi, Sensors Technology 6 (1986) 78. W. Gopel, Sensors and Actuators B 16 (1989) 167. O.V. Safonova, G. Delabouglisi, B. Chenevier, et al. Mater. Sci. Eng. C 21 (2002) 105. M.N. Rumyantseva, M. Labeau, J.P. Senateur, et al. Mater. Sci. Eng. B 41 (1996) 228. B.A. Akimov, A.V. Albul, A.M. Galkov, et al. Physics and Technics of Semiconductors 31 (1997) 400. N. Harada, Amorphous Semiconductor Technologies and Devices (1984) JARECT Vol. 6, p.283. D. Fujita, T. Ohgi, W. Deng, H. Nejo, T. Okamoto, S. Yokoyama, K. Kamidado, S. Mashiko, Surf. Sci. 454 – 456 (2000) 1021. J.C. Manifacier, Thin Solid Films 90 (1982) 297. M.L. Olvera, A. Maldonado, R. Asomoza, M. Konagai, M. Asomoza, Thin Solid Films 229 (1993) 196. A.J. Varkey, A.F. Fort, Thin Solid Films 235 (1993) 47. M. Ruckh, D. Hariskos, U. Ruhle, H.W. Schock, R. Menner, B. Dimmler, 25th IEEE Photov. Spec. Conf., Washington, USA, 1996, p.825. K.I. Chopra, S.R. Das Thin film solar cells, Mir, Moscow, 1986, pp. 435. K.T. Ramakrishna Reddy, C. Sravani, R.W. Miles, J. of Crystal Growth 184/185 (1998) 1031. B. Stjerna, C.G. Grangvist, Thin Solid Films 193:194 (1990) 704. M. Fantini, T. Torriani, Thin Solid Films 138 (1986) 255. K.H. Kim, J.S. Chun, Thin Solid Films 141 (1986) 287. P.M. Gorley, V.V. Khomyak, S.V. Bilichuk, I.G. Orletsky, P.P. Horley, V.O. Grechko, Material Science and Engineering B 118 (2005) 160. T. Maruyama, K. Fukui, Thin Solid Films 203 (1991) 297. T. Karlsson, A. Roos, C. Ribbing, Sol. Energy Mater. 11 (1985) 469. R. Latz, K. Michael, M. Scherer, Jpn. J. Appl. Phys. 30 (1991) L 149. M. Muta, S. Ohgushi, Y.Matsuda, H. Fujiyama, Thin Solid Films 341 (1999) 225. R.N. Joshi, V.P. Singh, J.C. Me Clure, Thin Solid Films 257 (1995) 32. S.V. Bilichuk, P.M.Gorley, V.V. Khomyak, Zs.J. Horvath Solid State Surfaces and Interfaces IV (SSSI IV) Abstract book, Castle Smolenice, Slovak Republic, November 8-12, 2004, p.17. A.W. Metz, J.R. Ireland, J-G. Zheng, et al., J.of the American Chemical Society 126 (2004) 8477.

Transparent Conductive Layers of Tin, Indium, and Cadmium Oxides… [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68]

299

F.C. Eze, Materials Chemistry and Physics 89 (2005) 205. R. Javakrishnan, G. Hodes, Thin Solid Films 440 (2003) 19. S.M. Kudryavtseva, A.A. Vertegel, S.V. Kalinin, et al. J.Mater. Chem. 7 (1997) 2269. O. Safonova, I. Bezverkhy, P. Fabritchny, et al. J.Mater. Chem. 12 ( 2002) 1174. Chaonan Xu, Jun Tamaki, Norio Miura, et al. Sensors and Actuators B, 3 (1991) 147. A.H. Hamdi, R.C.O. Laugal, A.B. Catalan, et al. Thin Solid Films 198 (1991) 9. S.I. Rembeza, T.V. Svistova, E.S. Rembeza, O.I. Borsyakova, Physics and Technics of Semiconductors 35 (2001) 796. A.S. Bakin, M.V. Bestaev, D.Tz. Dimitrov, V.A. Moshnikov, Yu.M. Tairov, Thin Solid Films 296 (1997) 168. D.B. Fraser, H.D. Cook, J. Electrochem. Soc. 119 (1972) 1368. E. Leja, T. Pisarkiewicr, A. Kolodrieg, Thin Solid Films 67 (1980) 45. G. Kienel, Thin Solid Films 77 ( 1981) 213. G. Haacke, W.E. Mealmaker, L.A. Siegel, Thin Solid Films 55 (1978) 67. N. Miyata, K. Miyaka, K. Koga, T. Fukushima, J. Ekectrochem. Soc. 127 (1980) 918. W.G. Haines, R.H. Bube, J. Appl. Phys. 49 (1978) 304. A.K. Chosh, C. Fishman, T, Feng, J. Appl. Phys. 49 (1978) 3490. R. Banerjee, D. Das, S.Ray, A. Batabyal, Sol. Energy Mater. 13 (1985) 11. K.L. Chopra, S. Major, D.K. Pandya, Thin Solid Films 102 (1983) 1. E. Shanthi, A. Banerjee, V. Dutta, K.L. Chopra, J. Appl. Phys. 53 (1982) 1615. M. Lebeau, B. Gautheron, F. Celier, et al., J. Sol. St. Chem. 102 (1993) 379. Y.S. Hsu, S.K. Gandhi, J. Electochem. Soc. 126 (1979) 1434. M. Mizuhashi, Thin Solid Films 70 (1980) 91. P. Nath, R.F. Bunshah, Thin Solid Films 69 (1980) 63. R. Dolbec, M.A. El Khakan, A.M. Serventi et al. Thin Solid Films 419 (2002) 230. T.L. Chu, J.C. Lien, H.C. Mollenkopf, et al. Solar Energy 17 (1975) 229. T.L. Chu, K.N. Singh, Solid – State Electronics 19 (1976) 837. Nodo Kazuhiro, Sato Hirotoshi, Jpn. J. Appl. Phys. Pt 1, 42 (2003) 217. B.L. Gehman, S. Jonsson, T. Rudolph, et al., Thin Solid Films 220 (1992) 333. S.H. Shin, J.H. Shin, K.L. Park, T. Ishida, O. Tabata, H.H. Kim, Thin Solid Films 341 (1999) 225. K.Suzuki, M.Mizuhasi, J. Vac. Soc. Jpn. 21 (1978) 158. W. Deng, T. Ohgi, H. Nejo, D. Fujita, Appl. Phys. A 72 (2001). .K. Davis, Tele-Tech, 11 (1952) 55. R.E. Aitchison, Australian J. Appl. Sci. 5 (1954) 10. R.H.W. Burkett, J. Brit. IRE 21 (1961) 301. D.F.A. MacLachlan, L.S. Philips, K.R. Honick, G.V. Planer, J. Brit. IRE 21 (1961) 221. J.D. Hayden, Brit. Elecrical Eng. 1962 (1962) 804. Julian W. Gardner, Semicond. Sci. Technol. 4 (1989) 345. A. Ortiz, S. Lopez, C. Falcony, et al., Sol. Energi Mater. 20 (1990) 381. S.A. Nasser, Thin Solid Films 342 (1999) 47. V.V. Khomyak, V.O. Grechko, S.V. Bilichuk, L.M. Mazur. Physycs and Technology of Thin Films Materials of IX International Conference (Ivano-Frankivsk, Ukraine), May 19-24, 2003 p.61

300

Yu.V. Vorobiev, J. Gonzalez-Hernandez, P. Gorley et al.

[69] V.V. Khomyak, V.O. Grechko, S.V. Bilichuk, Ya.P.Boyiko, V.V.Pylypko. Physycs and Technology of Thin Films Materials of IX International Conference (Ivano-Frankivsk, Ukraine), May 19-24, 2003 pp. 62-63. [70] P.M. Gorley, V.V. Khomyak, S.V. Bilichuk, I.G. Orletsky, P.P. Horley, V.O. Grechko. E-MRS 2004 Spring meeting (Strasbourg, France), May 24-28, 2004. D/PI.12. [71] V.V. Khomyak, I.G. Orletsky, S.V. Bilichuk, V.O. Grechko. ІІ Ukrainian Scientific Conference on Semiconductor Physics (Chernivtsi-Vizhnitsa, Ukraine): Ruta, September 20-24, 2004 Vol.2 p.478. [72] L. Maisell, R. Glend, Thin film technology handbook, New York, 1970. [73] R.M. Voshchilova, D.P. Dimitrov, N.I. Dolotov, et al. Physics and Technics of Semiconductors, Vol. 29 (1995) 1987. [74] K.Suzuki, M. Mizuhashi, Thin Solid Films 97 (1982) 119. [75] A. Fahrenbruck, R. Bube, Fundamentals of Solar Cells, Academic Press, New York, 1983. [76] C.G. Fonstad, R.H. Rediker, J. Appl. Phys. 42 (1971) 2911. [77] E.E. Kohnke, J. Phys. Chem. Sol. 23 (1962) 1557. [78] V.K. Miloslavskij, Phys. Solid State 7 (1965) 1550. [79] V. Craciun, D. Craciun, Z. Chen, J. Hwang, R.K. Singh, Mat. Res. Soc. Symp. 617 (2000) J3.13.1. [80] J.Y. Seto, J. Appl. Phys. 46 (1975) 5247. [81] N.F. Mott, E.A. Davis, Electronic properties in non-crystalline materials, Clarendon Press, Oxford, 1971. [82] B.I. Shklovski, A.L. Efros, Electronic properties of doped semiconductors, Nauka, Moscow, 1979. [83] V.L. Bonch-Bruevich et al., Electronic theory of disordered semiconductors, Nauka, Moscow, 1981. [84] F. Kröger, The Chemistry of imperfect crystals, North-Holland Publishing Complany, Amsterdam, 1964. [85] Huai-Wu Zhang, Wu-yi Xu, Vacuum 43 (1992) 835. [86] William G. Haines, Richard H. Bube, J. Appl. Phys. 49 (1978) 304. [87] H. Kimura, H. Watanabe, S. Ishihara, Y. Suzuki, T. Jto, J. Vac. Soc. Jpn. 30 (1987) 546. [88] Sh. Bashar, Study of Indium Thin Oxide (ITO) for Novel Optoelectronic Devices, Dissertation to obtain Ph.D.degree, King’s College, London, University of London, 1998. [89] J. Scryrbowski, A. Dietrich, H. Hoffman, Phys. Stat Sol. (a) 78 (1983) 243. [90] A. Dietrich, K. Schmalzbauer, H. Hoffman, J. Scryrbowski, Thin solid Films 122 (1984) 19. [91] N. Balasubramanian, A. Subrahmanyan, J. of Phys D: Appl Phys. 22 (1989) 206. [92] C. Sravani, P. Sreedhara Reddy, P. Jayarama Reddy, Mater Lett 12 (1992) 406. [93] A.J. Varkey, A.F. Fort, Thin Solid Films 239 (1994) 211. [94] K. Kurumurugan, D. Mangalraj, S.K. Narayandass, Mater. Letts. 28 (1996) 307. [95] Yu.I. Uhanov. Optical properties of semiconductors, Nauka, Moscow, 1977. [96] Physico-chemical properties of semiconductor materials, Handbook edited by A.V. Novoselova et al., Nauka, Moscow, 1979.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 301-326

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 10

DYNAMIC IMPEDANCE CHARACTERIZATION OF SOLAR CELLS AND PV MODULES BASED ON FREQUENCY AND TIME DOMAIN ANALYSES* D. Chenvidhya†, K. Kirtikara and C. Jivacate Clean Energy Systems Group (CES), King Mongkut’s University of Technology Thonburi (KMUTT), Bangkok, Thailand

ABSTRACT This article describes new methods to derive dynamic impedance of solar cells and PV modules from time and frequency domain analyses. Initially, we propose a new method, based on the frequency domain analysis, to measure dynamic impedance of x-Si solar cells and PV modules in the dark using basic instruments and FFT analysis. The dynamic parameters in the AC equivalent circuit, in addition to the DC model, consists of dynamic resistance, diffusion capacitance and transition capacitance. Loci of impedance in the complex plane can be obtained by inputting a small signal square wave, superimposing on either forward bias or reverse bias, to cells or modules. Such technique is compared with sinusoidal inputting. All of these parameters can be obtained from impedance loci in the complex plane. The impedance of a cell or a module can be derived in a closed form equation in terms of frequency dependent and voltage dependent resistance and capacitance under the dark condition with reverse bias. The relationship between the dynamic and static characteristics is compared for solar cell modules having low and high fill factors. Another new analytical method determining solar cell and module dynamic impedance is demonstrated using the same measuring techniques. Determination of dynamic parameters, previously outlined, and time constant of solar cells and modules, based on a time domain response, can be simultaneously obtained at each bias condition. *

A version of this chapter was also published in Leading Edge Research in Solar Energy edited by P. N. Rivers published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Email: [email protected] or [email protected]

302

D. Chenvidhya, K. Kirtikara and C. Jivacate The merits of this second characterization method using square wave inputs are reduction in measuring steps and yielding of dynamic parameters and time constants in a single measurement. Experiments on polycrystalline and amorphous silicon cells and modules are also conducted and their results will be separately revealed at a later date. Knowledge of dynamic impedance characterization of solar cells and modules will lead to better understanding of behaviors of PV grid-connected systems and improvement of power quality from such distributed power generation systems.

1. INTRODUCTION Understanding of solar cell characteristics is very important for studying solar cells performances. In practice, this involves understanding the fundamentals and development of devices, performance evaluation and various methods of measurement. Characteristics measurement would yield information to identify and the improvement of material properties in cell production, to identify solar cell grades for module production, to verify suitable models, and to evaluate cell or module performances. Solar cells parameters are basically determined in static characteristics. The solar cell characteristics consist of open circuit voltage (Voc), short circuit current (Isc), voltage, current and power at the maximum point (Vmp, Imp and Pmp), the fill factor (FF), series resistance (Rs) and shunt resistance (Rsh). These parameters can normally be characterized by various static characterization methods. In static characteristic determination, measurements can be done either under illuminated or under dark conditions. Under illumination, the measurements of IV characteristics are used in most standard measurements. However, the characteristics in the dark can often be found in various research works. The static characteristics or I-V characteristics of solar cells, both illuminated and dark characteristics, are often described in most of solar cell textbooks. These include the dc equivalent circuit model (lumped and distributed model), equations, operating conditions, curves, etc. Therefore, it will not be repeated here. In the past, most of solar cell applications were for stand-alone purposes, with battery storages in most cases. Knowledges of static parameters are adequate as any fluctuation in output from solar cell modules due to light intensity will have no impact on loads. Either constant dc voltages can be drawn from the battery or uniform ac voltages are available from inverters connected between the module and loads. However, it is clear from all situation reports that the direction of PV applications is a rapid increase and trend in grid-connected applications. Cloud movements will cause fluctuating module output. The module output, in turn, becomes the inverter input whose outputs are sensitive to the inverter operating condition (e.g. input voltage, input power). As the loads draw powers from both the inverter and the line, the electrical quality of the loads, the inverter and the line are related. To understand the dynamics of module and grid connection under varying cloud conditions, the solar cell, module and array ac or dynamic parameters are essential. The solar cell dynamic parameters, such as minority carrier lifetime, time constants, diffusion length, diffusion and transition capacitances (CD and CT) can be calculated by measurements when solar cell operates under dynamic or transient conditions.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

303

For modeling of a solar cell or a PV module, a dc model or dc equivalent circuit is adequate for describing solar array behavior in general. A dc circuit model consists of a series resistance (Rs), a shunt resistance (Rsh) and a diode having a non-ideal diode factor. The solar cells dynamic characteristics and impedance measurements had been described and reported in some specific applications, such as satellite applications [5, 6], the Mars Pathfinder and measurement of impedance of GaAs/Ge solar cells [3] and the Si Back Surface Field Reflector (Si-BSFR) solar cells [3]. The results are used to design efficient, reliable, high power devices with stability. In all cases, the impedances are determined by special equipment, such as an electrochemical interface (ECI), a frequency response analyzer (FRA) using an impedance spectroscopy technique. In previous studies, they highlight solar cell impedances in terms of material properties, and measured solar cells under dark conditions with either forward bias or reverse bias conditions. In our studies, we aim that understanding of dynamic impedances of solar cells and arrays are essential in determining the dynamic performance of arrays when connected to electricity distribution networks. It will become increasingly important when more PV systems are connected to the networks. This article describes new methods to measure and to derive dynamic impedances of a solar cell and a PV module from time and frequency domain analyses. The outline of this chapter consists of theoretical background for an ac equivalent circuit and Impedance Spectroscopy, a new method for solar cell impedance characterization, dynamic impedance in frequency domain analysis, PV module diagnostic method with dynamic impedance, and dynamic response in time domain analysis. The detail of each section is following.

2. THEORETICAL BACKGROUD

2.1. AC Equivalent Circuit An ac model of a solar cell, shown in figure 1, contains the illumination controlled current source (IL), series resistance (Rs), shunt resistance (Rsh), diode resistance (Rd) and in additional to equivalent components of a dc model, the transition capacitance (CT) and diffusion capacitance (CD). Rd represents non-ideal diode characteristic. CT is the junction or space-charge capacitance of the depletion region. CD is the capacitace due to minority carrier oscillation in response to the ac signal. It is recognized that an ac equivalent circuit of a module or and array takes similar forms to that of a solar cell. The particular forms depend on characteristics of individual solar cells and how solar cells are connected in modules and connection of modules into arrays.

304

D. Chenvidhya, K. Kirtikara and C. Jivacate

Figure 1. Ac equivalent circuit of a solar cell connecting to a load ZL under illumination.

The values of Rs and Rsh, are both voltage independent. The CT and Rd are voltage dependent, and a CD is voltage and frequency dependent. The values of CD, CT and Rd vary with the level of incident light, the cell temperature, the cell operating voltage, the solar cell material constant and processing parameters. In the high frequency range, the value of CD is approaching zero with increasing frequency. This is due to the minority carrier storage effects. Therefore, the CT is the only capacitance remaining at the higher frequencies. The value of CT depends upon dc voltage. From the static I-V characteristic of a solar cell under steady state illumination, the incremental diode dc resistance (Rdc) at any point on the output is defined as

R dc = dVdc dI dc

(1)

Rdc is related to Rd as follows

R dc =

R d R sh + RS R d + R sh

(2)

Variations of Rdc can be illustrated as a function of the dc voltage, Vdc. The values of Rdc of a solar cell vary in the range of 103 to 10-1 Ω. Near the short circuit condition, Rd becomes very large, and Rsh dominates the cell impedance. But near the open circuit condition, Rd becomes small, Rs exerts a large influence on the impedance.

2.2. Impedance Spectroscopy Technique In previous works, the impedance measurement technique mostly uses the method of impedance spectroscopy. The concept of the method can simply be described in figure 2.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

305

Figure 2. Block diagram describing the simply concept of Impedance Measurement using sinusoidal signal inputs.

The impedances of a solar cell are measured in the dark using sinusoidal signal inputs with varying frequencies and superimposing on the dc bias. It is measured in both forward bias and reverse bias conditions. The impedance loci are plotted and interpreted in a complex plane at any bias levels.

3. A NEW METHOD FOR SOLAR CELL IMPEDANCE CHARACTERIZATION A new method developed under this research work has been introduced by the author. To characterize the dynamic impedance of solar cell or module, we use square wave inputs superimposed on a dc biasing voltage in either forward or reverse condition. A square wave signal, being periodic, consists of infinite numbers of sinusoidal signal. Thus, inputting one square wave to a solar cell is, in theory, equivalent to simultaneously inputting infinite sinusoidal signals. In principle, any periodic inputting signal can be used. A square wave is used because it is readily available from most signal generator equipment. In this method, the responses are calculated by signal processing input and output signals (voltages and currents) using FFT technique to obtain various harmonic contents of input and output. A block diagram describing the concept of measurement is shown in figure 3. At the same time, this method permits analyzing data regarding the transient behavior of solar cell under study so that response time, rise and fall time, or time constant can be determined.

Figure 3. Block diagram of the method of dynamic impedance measurement using square wave inputs.

306

D. Chenvidhya, K. Kirtikara and C. Jivacate

In this method, the measuring results can be obtained by basic instrument such as oscilloscope and signal generator. At each square wave input of a certain frequency, phases and amplitudes of up to 10 corresponding harmonics can be determined with accurate measurement. It means that one whole impedance locus can be measured by using 2-3 square wave inputs at different frequencies. For previous studies, a large number of sinusoidal test frequencies covering 4-5 decades are required, hence, more measuring steps and time. Therefore, the advantages of this method, compared with the sinusoidal input method, are less expensive equipment setup, simpler measurement, and reduction in measuring steps, yielding comparable results. The experiment is setup to ensure the results obtained from this characterization method for solar cell dynamic impedance using square wave inputs. Both are done in the dark with forward bias conditions. The experiment is measured and calculated impedance of a solar cell using square wave inputs and using sinusoidal signal inputs. The measurements are using the same basic instrument, such as digital oscilloscope, signal generator and audio amplifier. The measuring data are analyzed by the FFT technique on MATLAB. The measured solar cell impedances are plotted in a complex plane. In this experiment, the impedance loci obtained by the new method using square wave inputs, shown in figure 4, are compared with the impedance loci obtained by the previous method using sinusoidal, shown in figure 5. The impedance loci, at the same bias level 0.2 V, obtained from the two methods are compared in figure 6. It is found that the trends of the impedance loci determined by both methods are the same in semicircular shape.

Figure 4. Impedance loci of the 10 cm x 10 cm solar cell measured by this new method using square wave inputs at forward bias levels of 0.2, 0.3 and 0.4 V.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

307

Figure 5. Impedance loci of the 10 cm x 10 cm solar cell using sinusoidal inputs at forward bias levels of 0.2 V, 0.3 V and 0.4 V.

Figure 6. Comparison of impedance loci obtained by square wave inputs and sinusoidal inputs at forward biasing level of 0.2 V.

From figure 1, the ac equivalent circuit can be derived and modeled with a simplified equivalent circuit at the frequency ω consisting of one resistive component and one reactive component in series in the form of RPV+jXPV by the following equation:

308

D. Chenvidhya, K. Kirtikara and C. Jivacate Z PV = R PV + jX PV 2 ⎤ ⎤ ⎡ ⎡ [(R sh + R d )R sh R d ] ω(R sh R d ) (C D + C T ) j − = RS + ⎢ ⎢ 2 2 ⎥ 2 2 ⎥ ⎣⎢ [ωR sh R d (C D + C T )] + [R sh + R d ] ⎦⎥ ⎣⎢ [ωR sh R d (C D + C T )] + [R sh + R d ] ⎦⎥

(3)

The series resistance of solar cell, Rs is constant at any bias level, and is given by the high frequency end interception of the impedance locus on the real axis. Due to the reactive component of solar cell impedance is close to zero at low frequency (ω→0) and consists of only the series resistance at high frequency (ω→∞). By the comparison results in figure 6, the values of Rs obtained by the square wave inputs method is 0.352 Ω, and Rs obtained by the sinusoidal inputs is 0.356 Ω. It is noted that the results determined by the two methods, using the same basic instrument, are not different. In addition, the other experiment is setup to compare the results obtained by the method with the results measured by the special instrument, impedance gain-phase analyzerSolartron Analytical or Frequency response analyzer (FRA).

4. DYNAMIC IMPEDANCE IN FREQUENCY DOMAIN ANALYSIS 4.1. Derivation of an AC Equivalent Circuit The ac equivalent circuit of a solar cell is shown in figure 7. In figure 7 (a), it shows the normal operating condition of cell. The figure 7 (b) shows the cell, without light generated current and load connecting, which the dynamic impedance are measured and derived in this condition. It is noted that Rd, CT and CD are voltage dependent.

(a) AC equivalent circuit of a solar cell connecting to a load ZL under illumination.

(b) AC equivalent circuit of a solar cell under dark condition.

Figure 7. AC equivalent circuit of a solar cell.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

309

Under dark condition when looking into the solar cell output port, normally connecting to a load, the solar cell impedance Zpv (ω) at a frequency ω can be shown to consist of one resistive component and one reactive component in series in the form of Rpv+jXpv. At each bias voltage V, the impedance Zpv(ω) can be expressed as the equation Z PV ( V , ω) = R PV ( V , ω) + jX PV ( V , ω) ⎡ ⎤ [{R sh + R d (V )}R sh R d (V ) ] = RS + ⎢ 2 2 ⎥ ⎢⎣ [ωR sh R d ( V ){C D ( V , ω) + C T ( V )}] + [R sh + R d ( V ) ] ⎥⎦ 2 ⎡ ⎤ ω{R sh R d ( V )} {C D ( V , ω) + C T ( V )} − j⎢ 2 2 ⎥ ⎣⎢ [ωR sh R d ( V ){C D ( V , ω) + C T ( V )}] + [R sh + R d ( V )] ⎦⎥

(4)

where RsRsh Rd(V) CD(V,ω) CT(V) ω

series resistance -shunt resistance dynamic resistance of diode diffusion capacitance transition capacitance -signal frequency

For simplicity, we drop V and ω in the equation (4) and write Z PV = R PV + jX PV ⎤ ⎡ [(R sh + R d )R sh R d ] − = RS + ⎢ 2 2 ⎥ ⎣⎢ [ωR sh R d (C D + C T )] + [R sh + R d ] ⎦⎥

2 ⎤ ⎡ ω(R sh R d ) (C D + C T ) j⎢ 2 2 ⎥ ⎣⎢ [ωR sh R d (C D + C T )] + [R sh + R d ] ⎦⎥

(5) Further, in the equation (5) we replace Rsh in parallel with Rd by Rp, and CT in parallel with CD as Cp. The simplified dynamic impedance equation can be rewritten as the equation

⎤ ⎡ ωR 2P C P ⎤ ⎡ RP Z PV = ⎢R S + ⎥ ⎥ − j⎢ 2 2 + 1 ω ( R C ) P P ⎦ ⎣ (ωR P C P ) + 1⎦ (6) ⎣ 4.2. Interpretations of Impedance Loci In the dark with forward bias, (1)

(2)

Rsh is much greater than Rd because diode resistance normally decreases with increasing conduction or biasing voltage, but Rsh is quite constant. Therefore, Rp is close to Rd. CD is more dominant than CT.

310

D. Chenvidhya, K. Kirtikara and C. Jivacate (3)

(4)

The reactive component of the solar cell dynamic impedance is close to zero at low frequency (ω → 0). Moreover, the impedance consists of only the series resistance at high frequency (ω → ∞). The impedance locus on the complex plane is close to semi-circular.

The two intercepts of the impedance locus on the real axis are approximately Rs and Rs+Rp. Rp is close to Rd in this case. In the dark with reverse bias, (1) (2) (3)

Rd is much greater than Rsh (in case of low level of biasing). Therefore, Rp is close to Rsh, CT is more dominant than CD. The reactive component of solar cell dynamic impedance is close to zero at low frequency (ω → 0). The impedance consists of only the series resistance at high frequency (ω → ∞). The impedance locus on the complex plane is close to semicircular.

The two intercepts of the impedance locus on the real axis are approximately Rs and Rs+Rp. Rp is close to Rsh in this case.

Figure 8. Impedance loci of a PV cell or module at low (V1), intermediate (V2) and high (V3) biasing voltages.

It can be shown that the equation describing each semi-circular impedance locus is of the form ⎡ ⎡ R R − R S (R sh + R d )⎤ ⎛ R sh R d + R S (R sh + R d ) ⎞⎤ ⎟⎟⎥ + X 2PV = ⎢ sh d ⎢R PV − ⎜⎜ ⎥ 2(R sh + R d ) 2(R sh + R d ) ⎣ ⎦ ⎝ ⎠⎦ ⎣ 2

At any biasing voltages, the equation (7) can be simplified as

2

(7)

Dynamic Impedance Characterization of Solar Cells and PV Modules… 2

⎡ ⎡R − RS ⎤ ⎛ R P + R S ⎞⎤ ⎟⎥ + X 2PV = ⎢ P ⎢R PV − ⎜ ⎥ 2 2 ⎦ ⎣ ⎠⎦ ⎝ ⎣

311

2

(8)

The equation (8) when expressed in terms of the magnitude |Zpv| and the argument θ of the impedance Zpv can be written as

Z PV − (R P + R S ) Z PV cos θ + R P R S = 0 2

(9)

In case of, ω = 0 or dc resistance of solar cell, therefore, equation (6) turns into

Z PV = R S + R p Zpv is corresponding to Rdc in static model of solar cell in equation (2).

Figure 9. Impedance loci of the module under reverse bias, plotted in the complex Zpv plane.

By the derivations of an ac equivalent circuit of solar cell and the interpretations of impedance loci in a complex plane, it is found that the impedance loci in the reverse bias conditions can yield the parameters of a cell or a PV module, such as the series resistance, the shunt resistance. Therefore, the measurement on the dynamic impedance of a 20 Wp crystalline silicon PV module is setup to confirm the above principle. The measurement is done in the dark with any reverse bias levels, and the impedance loci are plotted in complex plane as figure 9. It is the first time that a report on commercial PV module dynamic impedance is made. At low reverse biasing (0.3 and 0.5 V/cell), the two impedance loci are closely located implying that the impedances are nearly identical. We know that the intercept is equaled to Rd

312

D. Chenvidhya, K. Kirtikara and C. Jivacate

in parallel with Rsh, Rd is voltage dependent and Rsh is voltage independent, therefore, the two coincident loci imply that Rd is much greater than Rsh and the intercept on the real axis of Zpv is basically Rsh. The high frequency end intercept is approximately Rs. For this tested module Rs is 0.79Ω. When we plot the real axis intercepts at low frequency ends of the loci and the biasing voltage, we obtain the relationship between Rp (of Zpv at ω → 0) and the voltage. The plot is shown in figure 10. 160 140 120

Rp (k Ω)

100 80 60 40 20 0 0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

V (Volt/cell)

Figure 10.The relationship of Rp (of Zpv at ω → 0) versus biasing level.

4.3.Voltage and Frequency Dependencies of RPV and XPV Voltage and Frequency Dependencies of a Cell From the loci of impedances Zpv, we can determine the following: (a) Rpv, Xpv, |Zpv| and argument of Zpv. The equations (4), (5) and (6) give the relationship between the above quantities with the equivalent components and the frequency. (b) Rs from the high frequency intercepts and Rsh from the low frequency intercepts.

Forward Bias

Frequency Response Plots Based on the results shown in figure 4, we calculate the Rpv, Xpv, |Zpv| and argument of Zpv of the cell. The results are shown as figure 11, 12, 13 and 14.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

313

20 18 16

Rpv (Ω )

14 12 10 8 6 4 2

0.2V

.

0.3V

1.E+05

1.E+04

f (Hz)

1.E+03

1.E+02

1.E+01

1.E+00

0

0.4V

Figure 11. Rpv of the cell as a function of frequency.

0 -1 -2

Xpv (Ω )

-3 -4 -5 -6 -7 -8 -9

0.2V

Figure 12. Xpv of the cell as a function of frequency.

0.3V

0.4V

1.E+05

1.E+04

f (Hz)

1.E+03

1.E+02

1.E+01

1.E+00

It is seen that the complex impedance plot yields information on Rs and Rd. However, the Rd vs. frequency plot gives an additional information on the frequency and the voltage dependent nature of Rd. At low forward bias voltage of 0.2 V Rd is about 15.53Ω and remains constant and dominant until a frequency of 100 Hz. However, at higher bias voltage Rd decreases. Moreover, Rd remains constant and dominant upto a higher frequency range, about 1 kHz in our cell.

314

D. Chenvidhya, K. Kirtikara and C. Jivacate 20 18 16

|Zpv| (Ω )

14 12 10 8 6 4 2

0.2V

0.3V

1.E+05

1.E+04

f (Hz)

1.E+03

1.E+02

1.E+01

1.E+00

0

0.4V

Figure 13. |Zpv| of the cell as a function of frequency.

0 -10

Arg. Zpv (deg.)

-20 -30 -40 -50 -60 -70 -80

0.2V

0.3V

Figure 14. Arg. of Zpv of the cell as a function of frequency.

0.4V

1.E+05

1.E+04

f (Hz)

1.E+03

1.E+02

1.E+01

1.E+00

Frequency Dependence of Capacitance From the Rpv frequency plot (11), the Xpv frequency plot (12) and the equations (4) and (5), Cp (CD in parallel with CT) has been determined. The calculated Cp at the bias voltage of 0.2 V are 3.19, 5.45 and 5.85 μF at 100, 500 Hz and 1 kHz. It may be recalled that the transition capacitance CT is voltage dependent whereas

Dynamic Impedance Characterization of Solar Cells and PV Modules…

315

the diffusion capacitance CD is both voltage and frequency dependent. Moreover, CD increases with frequency. From the above results, it can be seen that Cp, and hence CD, increases with frequency as expected. CD and CT can separately be determined. By selecting an impedance locus at a fixed bias voltage, therefore CT is fixed, CD can be calculated.

Reverse Bias Based on the results of the cell in reverse bias condition, we calculate the Rpv, Xpv, |Zpv| and argument of Zpv of the cell. The results are shown as figure 15, 16, 17 and 18. 200

Rpv (Ω)

160

120

80

40

1.E+04

1.E+05

1.E+04

1.E+05

1.E+03

1.E+02

1.E+01

1.E+00

0

f (Hz) 0.3V

0.5V

0.9V

1.5V

1.E+03

1.E+02

1.E+01

1.E+00

Figure 15. Rpv of the cell as a function of frequency. f (Hz)

0

Xpv (Ω)

-20

-40

-60

-80

-100

0.3V

0.5V

Figure 16. Xpv of the cell as a function of frequency.

0.9V

1.5V

316

D. Chenvidhya, K. Kirtikara and C. Jivacate 200

|Zpv| (Ω)

160

120

80

40

1.E+05

1.E+04

1.E+03

1.E+02

1.E+01

1.E+00

0

f (Hz) 0.3V

0.5V

0.9V 1.5V

Figure 17. |Zpv| of the cell as a function of frequency.

0

Arg. Zpv (deg.)

-2 0

-4 0

-6 0

-8 0

-1 0 0

0 .3 V

0 .5 V

0 .9 V

Figure 18. Arg. of Zpv of the cell as a function of frequency.

1 .5 V

1.E+05

f (H z )

1.E+04

1.E+03

1.E+02

1.E+01

1.E+00

Voltage and Frequency Dependence of Rpv and Xpv of A PV Module Based on similar procedures in calculating the Rpv, Xpv, |Zpv| and Arg. of Zpv of the cell, we undertake similar calculation of those of the module. We use the results shown in figure 9 (impedance loci of the module under reverse bias of 0.3, 0.5, 0.9, 1.2 and 1.5 V/cell). Figure 19 to 22 show the frequency response plots. From the frequency plot of Rpv, figure 19, we see that the shunt resistance of the module (Rpv at low frequency and the low bias voltage of 0.3 V/cell) is about 143 kΩ. This gives Rsh of each cell of 4.33 kΩ.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

317

1.6E+05 1.4E+05 1.2E+05

Rpv (Ω )

1.0E+05 8.0E+04 6.0E+04 4.0E+04 2.0E+04 0.0E+00 1.E+00

1.E+01

1.E+02

0.3V

0.5V

1.E+03

f (Hz) 0.9V

1.2V

1.E+04

1.E+05

1.5V/cell

Figure 19. Rpv of the module as a function of frequency.

However, when we consider the frequency dependent nature of Rpv of the selected test cell and that of the module, we see that Rpv of the module drops more rapidly with frequency. This implies that, per cell basis, Cp (CD in parallel with CT) of the module is larger than that of the cell. f (Hz) 1.E+00 0.0E+00

1.E+01

1.E+02

1.E+03

-1.0E+04 -2.0E+04

Xpv (Ω )

-3.0E+04 -4.0E+04 -5.0E+04 -6.0E+04 -7.0E+04 -8.0E+04

0.3V

0.5V

0.9V

Figure 20. Xpv of the module as a function of frequency.

1.2V

1.5V/cell

1.E+04

1.E+05

318

D. Chenvidhya, K. Kirtikara and C. Jivacate 1.6E +05 1.4E +05 1.2E +05

|Zpv| (Ω)

1.0E +05 8.0E +04 6.0E +04 4.0E +04 2.0E +04 0.0E +00 1.E +00

1.E+01

1.E +02

0.3V

0.5V

1.E +03

f (H z) 0.9V

1.2V

1.E+04

1.E +05

1.5V/cell

Figure 21. |Zpv| of the module as a function of frequency.

In this section, it is shown that the frequency range of which the solar cells and module impedances rapidly change is in the region of few tens to few hundreds Hz. However, changes due to cloud movements are slow, over few seconds or few minutes period. So we are of the view that the solar cells impedances variation would largely be caused by voltage changes arising: -

in some part due to cloud and radiation fluctuation and in other part due to electrical nature of power conditioning units, loads and distribution network. We cannot yet be certain about the relative importance of these electrical entities connected to the PV array. f (H z )

Arg. Zpv (deg.)

1 .E + 0 0 0 .0 E + 0 0

1 .E + 0 1

1 .E + 0 2

1 .E + 0 3

-3 .0 E + 0 1

-6 .0 E + 0 1

-9 .0 E + 0 1

0 .3 V

0 .5 V

0 .9 V

1 .2 V

Figure 22. Arg. of Zpv of the module as a function of frequency.

1 .5 V /c e ll

1 .E + 0 4

1 .E + 0 5

Dynamic Impedance Characterization of Solar Cells and PV Modules…

319

5. PV MODULE DIAGNOSTIC METHOD USING DYNAMIC IMPEDANCE From the static characteristic of solar cell, the series resistance, shunt resistance, fill factor and efficiency can be calculated from the illuminated current-voltage curve. These parameters determine performance of a solar cell or a module. Cells or modules having low Rs, high Rsh and high FF are recognized as having good quality. The current-voltage characteristic of a cell having a high fill factor is rather rectangular-like while that of a cell with a low fill factor is triangular-like. Based on the derivation outlined in the Section 4, in the complex plane, impedance loci of solar cells with low Rs would have the high frequency intercepts lying close to the origin. On the other hand, impedance loci of solar cells having large Rsh would have the low frequency intercepts lying towards the right hand side of the real axis. Thus, based only on measuring the dynamic impedance of solar cells and plot them in the complex plane, without having to determine their static characteristics, one can quickly distinguish between solar cells having good quality (low Rs and high Rsh) and those having low quality (high Rs and low Rsh). Comparatively speaking, the impe-dance loci of good solar cells would be semi-circularly large.

(a) Schematic picture of static characteristics of two PV modules, M1 being of good quality and M2 being of low quality.

(b) Schematic picture of impedance loci of two PV modules, M1 and M2,

Figure 23. Comparative of schematic pictures of the static characteristics and impedance loci of two PV modules being of good quality and low quality.

320

D. Chenvidhya, K. Kirtikara and C. Jivacate

Figure 23 shows a schematic picture of current-voltage static characteristics and dynamic characteristics of solar cells of good and low quality. Two modules, one having a high fill factor of about 0.7 and the other low fill factor of about 0.5, having been selected from a number of PV modules. Illuminated static I-V characteristics of the modules are determined. Static parameters obtained for each module are the open circuit voltage (Voc), the short circuit current (Isc), the module efficiency (η) and the fill factor (FF). The I-V characteristics of the two modules are compared in figure 24. 3.00

2.50

Current (A)

2.00

1.50

Module1 Module1: Pm=28.15W (13.8V,2.04A) Isc=2.20A Voc=18.0V FF=71.08%

1.00

Module2 Module2: Pm=22.40W (10.0V,2.24A) Isc=2.52A Voc=18.2V FF=48.84%

0.50 Test Conditions: Radiation of 780 W/m2 and Module Temperature of 43C

0.00 0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

Voltage (V)

Figure 24. Illuminated Static I-V Characteristics of Two Modules.

The dynamic impedances of the modules under dark condition with reverse biasing are measured from their frequency responses. Measurement of dynamic impedances of each module yields its Rs, Rsh, dynamic resistance (Rd), diffusion capacitance (CD) and transition capacitance (CT). The dynamic impedance loci, under a biasing voltage of 0.3 V/cell, are plotted in figure 25. Rs of the Module 1 and the Module 2 are 0.93 and 3.13 Ω, respectively. On the other hand, Rsh of the Module 1 and the Module 2 are 52.6 and 13.95 kΩ, respectively. From the comparison, we see the well established correspondence between the FF, Rs, and Rsh of solar cell modules, i.e. low Rs and high Rsh results in high FF modules. We can say that dynamic characterization can yield qualitative information on comparative fill factors of modules, but not quantitative values. However, other essential electrical parameters of modules required to predict the module–grid interaction can only be obtained from the dynamic characterization. This becomes essential as the trend in solar cell applications are in the grid interactive applications. If we are mainly interested in investigating the module-grid interaction, information from the dynamic characterization of modules is necessary and sufficient. We note that the same information on the module dynamic characteristics provide an adequate qualitative nature of fill factors as well.

Dynamic Impedance Characterization of Solar Cells and PV Modules… 0.000 0.000

0.002

321

0.004

-0.020

Rpv (k ohm)

-0.040

0 -0.060

0

-0.080

-5

10

20

30

40

50

60

ω

Xpv (k ohm)

-10

-15

-20

-25

Module1 Module2 Module 1: Rs=0.93ohm, Rsh=52.6 k ohm Module 2: Rs=3.13ohm,Rsh=13.95kohm

-30

Figure 25. Dynamic Impedance Loci of Two Modules (Dark Reverse Bias of 0.3 V/Cell).

6. DYNAMIC RESPONSE IN TIME DOMAIN ANALYSIS An alternative method can also analyze dynamic characteristics of a solar cell from time domain responses. In previous sections of this chapter, the method of measurements and derivations of dynamic impedances of a solar cell or a PV module are done in frequency domain analysis. A solar cell or module are measured by impedance spectroscopy method using sinusoidal signal inputs in general, and a new proposed method measures by using square wave inputs with FFT technique to analyze harmonic contents of the output response. Both measuring results are analyzed in frequency domain analysis, and plotted the magnitudes and the arguments of impedances of solar cell in a complex plane. In the same measurement on the new characterization method for dynamic impedances using square wave inputs in section 3, the time constants of a solar cell or module can be simultaneously obtained, in the time domain responses, at each bias condition. Time responses can be directly observed and analyzed in the time domain whereas the frequency domain analysis can be done by using FFT to determine harmonics contents. Under dark conditions at each bias voltage, the dynamic impedance of a solar cell, ZPV(ω) can be modeled with one resistive component and one reactive component in series in form of RPV+jXPV. The impedance can be expressed in function of input signal frequency ω

322

D. Chenvidhya, K. Kirtikara and C. Jivacate

as the equation (4), and ZPV can be rewritten in simplicity form in equation (5) and (6) detailed in section 4. The impedance loci of a solar cell or PV module can be described in frequency domain analysis as frequency response plots or impedance plotted in a complex plane at any bias levels, as shown in figure 8. In the time domain analysis with the same measurement using square wave inputs, a response time or time constant of a solar cell, at any bias voltage V, can be derived and written as equation (10) Time Constant =

(10)

R s R sh R d (V)[(CT (V)) + C D (V, ω)] R d (V)R sh + R s [R d (V) + R sh ]

The time constant can be yielded from the slope of output response curve at the transition point of a square wave. It is noted that some components in an equivalent circuit model are voltage and frequency dependencies as mentioned in section 4. At any fixed bias level, CD is only one component that is voltage dependence. It is observed that the time constants are also varied with frequency. From the same measured results using square wave inputs, the time constants can be analyzed and plotted as a function of frequency, at any bias voltages either forward and reverse conditions. These are shown in figure 26 and 27. 250

200

0.2 Volt

Time Constant (µS)

0.3 Volt 0.4 Volt

150

100

50

0 10

100

1000

10000

100000

Frequency (Hz)

Figure 26. Relation between frequency and time constant on forward bias.

As shown in figure 26 and 27, it is observed that time constants of a solar cell are also voltage and frequency dependencies. Time constants in reverse bias conditions rapidly drop in lower frequency than in forward bias conditions. Time constants under reverse bias conditions are larger than under forward bias conditions. This is due to the dominance of the junction capacitance in reverse bias, especially at low level. However, the effect of the junction capacitance should prominent in cells with good junction quality. The cell that we use has low junction quality reflecting in low Rsh reported prviously.

Dynamic Impedance Characterization of Solar Cells and PV Modules…

323

1000

0.3 Volt

Time Constant (µS)

750

0.5 Volt 0.7 Volt

500

250

0 10

100

1000

10000

100000

Frequency (Hz)

Figure 27. Relation between frequency and time constant on reverse bias.

From the results, the relationship of bias voltage and the time constant can be plotted in Figure 28 and 29 to illustrate that at the same frequency, the time constants decrease with increasing voltages. We note that the voltage dependent components are Rd, CT and CD. We do not determine the relative dominance between them. In principle, we can use this method to determine all the equivalent components. This will be reported separately. 200

30 Hz

Time Constant (µS)

150

500 Hz 4 kHz 10 kHz

100

50

0 0.15

0.2

0.25

0.3

0.35

0.4

Bias Voltage (V)

Figure 28. Relation between bias voltage and the time constant on forward bias.

0.45

324

D. Chenvidhya, K. Kirtikara and C. Jivacate 1000

50 Hz 200 Hz 700 Hz Time Constant (µS)

750

3 kHz

500

250

0 0.2

0.3

0.4

0.5

0.6

0.7

0.8

Bias Voltage (V)

Figure 29. Relation between bias voltage and the time constant on reverse bias.

As we mention earlier, from one single set of measurements this method allows simultaneous analyses in both the frequency domain and the time domain. The relationship between the time constant and the bias voltage and frequency can be used to determine the AC equivalent circuit components. If the frequency domain analysis is carried out with consistency of the equivalent components, then comparison of results, as obtained from both domains, can be compared.

7. CONCLUSION In this chapter, dynamic impedances of a solar cell or a PV module can be characterized and derived in both frequency domain analysis and time domain analysis. A new characterization method for solar cell dynamic impedance is proposed, based on frequency domain analysis. By this method, a solar cell is measured in the dark by using a basic instrument, such as a digital oscilloscope, signal generator and audio amplifier. The new simple method for measuring solar cell impedances using square wave inputs instead of sinusoidal inputs reported earlier. The impedance loci are calculated from the output responses by FFT technique to analyze harmonics contents, and then plotted in a complex plane. Comparison is made with results obtained from the measurement, based on the same principle, using sinusoidal inputs. It is confirmed the comparable results. The dynamic impedance of a solar cell or a PV module in the dark can be derived in a closed form equation in terms of frequency dependent and voltage dependent resistance and

Dynamic Impedance Characterization of Solar Cells and PV Modules…

325

capacitance in reverse and forward bias conditions. The dynamic parameters in an AC equivalent circuit, in addition to the DC model, consisting of dynamic resistance, diffusion capacitance and transition capacitance are also derived in the frequency domain analysis. Impedance loci of solar cells are separately interpreted in forward bias and reverse bias conditions. Intercepts of each impedance locus yield series, shunt and dynamic resistance of a cell or a module. Moreover, the dynamic characteristics in the dark can also diagnose the quality of PV modules. Comparison of the static I-V characteristics and the dynamic impedance of PV modules having low and high fill factors are made. It is noted that the dynamic characteristics of modules can provide an adequate qualitative nature of respective fill factors as well. Another new analytical method is made in the same measurement of dynamic impedance of a solar cell using the square wave inputs method, both in the frequency domain analysis and in time domain analysis. Dynamic impedance parameters are analyzed in the frequency domain, and the time constants are analyzed in time domain. Dynamic impedance parameters and time constants of solar cells or PV modules can be simultaneously obtained at each bias condition. The advantages of this simple measuring method are reduction in measuring steps and yielding of dynamic parameters and time constants in single measurement. However, the dynamic impedance characterization method is not only done on crystalline silicon solar cells, but also done on polycrystalline and amorphous silicon cells and modules, and their results will be separately revealed at a later date. Knowledge of dynamic impedance characterization of solar cells and modules will lead to better understanding of characteristics of PV grid-connected systems in distribution networks, and be used to diagnose the quality of PV modules.

ACKNOWLEDGMENTS The authors are grateful to the Clean Energy Systems Group (CES), the Pilot Plant Development and Training Institute (PDTI), the Department of Electrical Engineering Department, the School of Energy and Materials of the University for equipment and facilities. We would like to thank Dr. Veerapol Monyakul, Mr. Nattavut Chayavanich, Mrs. Tasanee Chayavanich, Mr. Jutturit Thongpron and Mr. Chamnan Limsakul and the staff of CES for their valuable discussion and comments.

REFERENCES [1] [2] [3] [4]

Rauschenbach, H.S. Solar Cell Array Design Handbook; Van Nostrand Reinhold; New York; 1980. Pierret, R.F.; Semiconductor Device Fundamentals; Addison-Wesley, Reading, M.A.; 1996. Mueller, R.L.; Wallance, M.T.; Illes, P. Scaling nominal solar cell impedances for array design, WCPEC-1, December 5-9, 1994. 2034-2037. Suresh, M.S. Measurement of solar cell parameters using impedance spectroscopy, Sol Energy Mater Sol. Cells. 1995, vol. 43, 21-28.

326 [5]

D. Chenvidhya, K. Kirtikara and C. Jivacate

Kumar, R.A.; Suresh, M.S.; Nagaraju, J. Measurement of AC parameters for Gallium Arsenide (GaAs/Ge) solar cell by impedance spectroscopy, IEEE Trans Electron Devices. 2001, vol.48, 2177-2179. [6] Chenvidhya, D. PV Grid-Connected Systems for Residential Distribution System: Dynamic Impedance Characterization of Solar Cells and PV modules; Doctor of Engineering Thesis, King Mongkut’s University of Technology Thonburi; Bangkok, Thailand; 2002. [7] Chenvidhya, D.; Kirtikara, K.; Jivacate, C. A new characterization method for solar cell dynamic impedance; Sol. Energy Mater Sol. Cells. 2003, vol. 80, 459-464. [8] Chenvidhya, D.; Kirtikara, K.; Jivacate, C. PV module dynamic impedance and its voltage and frequency dependencies; Sol. Energy Mater Sol. Cells. 2005, vol. 86, 243251. [9] Chenvidhya, D.; Kirtikara, K.; Jivacate, C. On dynamic and static I-V characteristics of solar cell modules having low and high fill factors, WCPEC-3, May 12-16, 2003. [10] Wongyao, N.; Kirtikara, K.; Jivacate, C.; Chenvidhya, D. Time Responses of a Crystalline Silicon Solar Cell to Varying Light Inputs: Equivalent Components Determination, PVSEC-14, January 26-30, 2004. [11] Chenvidhya, D.; Limsakul, C.; Thongpron, J.; Kirtikara, K.; Jivacate, C. Determination of Solar Cell Dynamic Parameters from Time Domain Responses, PVSEC-14, January 26-30, 2004.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 327-365

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 11

WIND ENERGY TECHNOLOGY OVERVIEW* United States Department of the Interior, Bureau of Land Management SUMMARY Modern wind energy technologies rely heavily on the very complex scientific discipline of fluid dynamics (which includes the study of the atmosphere) and the equally complex engineering discipline of aerodynamics. A comprehensive treatment of either of these disciplines is well beyond the scope of this programmatic environmental impact statement (PEIS). The discussions that follow are intended only to establish a basic understanding of wind technology and the factors that control its evolution. References are provided for those who wish to have a more detailed understanding of wind technology. This appendix provides an overview of the fundamentals of wind energy and wind energy technologies, describes the major components of modern wind turbines, and introduces terms that are unique to the field of electric power generation using wind energy. Important site characteristics and critical engineering aspects of wind energy technologies are presented, and their respective influences on future development decisions are discussed.[1] An overview of the current state of wind energy technology and ongoing research and development (R&D) is provided. Descriptions of a typical wind energy project and the major actions associated with each phase of development — site monitoring and testing, construction, operation, and decommissioning — are presented in Chapter 3 of this PEIS.

*

A version of this chapter was also published in Wind Energy: Technology, Commercial Projects and Laws edited by Marco A. Telles published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

328

United States Department of the Interior, Bureau of Land Management

1. IMPORTANT TERMS AND CONVENTIONS Discussions in the following sections introduce important terms and conventions, some of which are unique to the wind energy industry. The terms and conventions are described in the text where they are first introduced. Additional details are provided in the glossary of this PEIS (Chapter 10).

2. WIND ENERGY Wind represents the kinetic energy of the atmosphere. In simplest terms, wind is the movement of air in the earth’s atmosphere relative to a fixed point on the earth’s surface. The major initiator of that movement is the uneven heating of the earth’s surface by solar radiation. The materials that compose the patchwork of the earth’s surface (e.g., vegetation, exposed rock, snow/ice cover, and water) react differently to solar radiation, absorbing heat energy and reflecting some of that energy back into the atmosphere at different rates. The result is a nonequilibrium condition in which adjacent air masses have different heat energies and, as a result of adiabatic expansion or compression, different barometric pressures. Wind is one result of the atmosphere’s attempt to normalize those differences and return to the lowest possible equilibrium state. The rotation of the earth around its axis initially causes a generally uniform global flow of air from west to east; however, many other factors add complexity to the dynamics of the earth’s atmosphere. The text box on the next page has additional information on atmospheric motion.

3. EXTRACTING THE POWER OF THE WIND The kinetic energy of wind is related to its velocity. This relationship is represented mathematically by the following equation: P=½×ρ×A×V3,

(1)

where P = wind power (W), ρ = air density (typically 2.70 lb/m3 [1.225 kg/m3] at sea level and 59 F [15 C]), A = cross-sectional area of the wind being measured (m2), and V = mean velocity of the wind within the measured cross section (m/s). A careful examination of this power equation reveals the following important fundamental truths about wind energy. Both the air’s density and the cross-sectional area of the wind being intercepted have a direct relationship to wind power. The air’s density varies with temperature, elevation, and humidity, but, in all instances, the density remains relatively low. Thus, any changes to air density have a minimal effect on the wind’s inherent power. Doubling the cross-sectional area of a wind front leads to a doubling of the intrinsic power. Most important to wind farmers is the fact that the wind’s power is proportional to the cube of

Wind Energy Technology Overview

329

its average velocity. Thus, a doubling of the average or mean wind speed results in an eightfold increase in its power. As a practical matter, wind energy technologists focus on the wind’s “power density” or power per unit area of wind being intercepted, expressed in W/m2. Simple manipulation of the above power equation allows power density to be calculated by using the following expression: Power density = P/A = ½ × ρ × V 3.

(2)

The height of the wind above the earth’s surface also affects the average wind speed. Frictional drag and obstructions near the surface of the earth generally retard wind speed and induce a phenomenon known as wind shear (the change in a wind’s speed with elevation). The rate at which wind speed increases with height varies on the basis of local conditions of the topography, terrain, and climate, with the greatest rates of increase observed over the roughest terrain. Unique local conditions notwithstanding, a reliable approximation is that wind speed increases approximately 10% with each doubling of height (Gipe 1995). Understanding Atmospheric Motion

Wind represents the earth’s atmosphere in motion. Understanding the development and progression of wind involves understanding the complex array of forces that constantly act upon the earth’s atmosphere and cause its continuous motion. The velocity, direction, and variability of wind are products of those collective forces. The major forces at play include basic laws of thermodynamics, the force of the earth’s gravity, frictional forces and obstructions imposed by the topography of the earth’s surface, and the Coriolis effect caused by the earth’s rotation. Thermodynamics governs the ways in which a given air mass behaves as it exchanges heat energy with its surroundings. Although the atmosphere’s density is quite low, the gravitational forces of the earth nevertheless exert a constant downward force on the atmosphere that continuously affects its behavior. It can be intuitively understood that the surface of the earth over which wind passes can also have some influence on wind, especially in the planetary boundary layer (the portion of the atmosphere immediately above the earth’s surface). Topography can either increase or decrease wind speed in localized areas. Topography can also contribute to or induce wind shear (the rapid change of direction of wind with altitude). When other overriding forces are absent, topographic obstructions and friction at the earth’s surface generally result in higher wind speeds at higher altitudes, with the highest wind speeds being achieved when all surface influences disappear. This wind is called the geostropic wind. The height or thickness of the planetary boundary layer varies over the surface of the earth (and actually changes slightly over the course of the day as a result of solar heating), reaching to thousands of feet in some locations. For the practical purpose of harvesting wind energy, the wind regime of greatest interest

330

United States Department of the Interior, Bureau of Land Management is contained completely within the boundary layer and, ideally, is composed largely of geostrophic wind. The force commonly referred to as the Coriolis effect is more difficult to comprehend. Although it is easy to understand wind as being the motion of the atmosphere relative to one’s point of observation on the surface of the earth, it is also important to recognize that one’s point of observation, while it is fixed on the earth’s surface, is not fixed in space, and it is itself moving as the result of both the earth’s rotation and its orbit around the sun. The Coriolis effect is most easily defined as that apparent force on the wind that would not have otherwise occurred except for the earth’s rotation and movement through space. It is manifested as a bending or redirection of the wind into circular patterns as air masses move from high-pressure to lowpressure areas. The magnitude of the Coriolis effect is a function of latitude. Winds directly above the earth’s equator and moving in a direction parallel to the earth’s axis of rotation experience very little in the way of a Coriolis effect. Winds occurring at other latitudes experience a Coriolis effect that is roughly proportional to the distance of that latitude from the equator. This fact can be easily understood by recognizing that any given point on the earth’s surface along its equator is traveling at roughly 373 mph (600 km/h) around the earth’s axis of rotation, while both the north and south poles have virtually no angular momentum. Other characteristics of atmospheric motion that are of great practical significance to wind energy development are those factors that contribute to its variability over both time and geographic location. These factors include topography-induced variations, annual and seasonal wind speed variability, synoptic variations (resulting from or influenced by broad-area weather patterns and storm fronts), diurnal variations (reflecting changes in levels of solar radiation over a 24-hour cycle), turbulence (the uneven, chaotic motion of air), wind gusts, and extreme wind speeds. All such factors are critical to identifying ideal wind regimes and to designing wind turbines that can capture wind energy with the greatest efficiency while still withstanding the forces to which they will be exposed over their lifetimes. Since most of these forces exhibit their greatest influence on atmospheric motion in the planetary boundary layer (the portion of the atmosphere in which wind turbines normally operate), their influence on siting decisions and turbine design is substantial. While many of these variability factors can be intuitively understood, many others cannot. This uncertainty leads directly to the difficulties that now exist in accurately predicting weather. This uncertainty also greatly increases the complexity involved in selecting and developing the ideal wind farm.

Because wind flows not only more quickly but also more uniformly as the elevation from the earth’s surface increases, the power contained in the wind is both greater and more easily extractable at higher elevations. Because turbulence decreases as the distance from surface obstructions increases, power actually increases faster with height than the relationship of

Wind Energy Technology Overview

331

power to the cube of the wind’s speed would indicate. Thus, for example, a fivefold increase in height results in nearly a doubling of available wind power. To take advantage of this relationship, wind turbine developers pursue designs that not only allow the capture of the greatest cross-sectional area of wind but also allow the capture of wind at the highest practical elevation possible. There are trade-offs, however. Higher turbine elevations require more substantial support systems (both towers and their foundations) and substantially greater initial investments. Higher altitudes also subject the rotor and the nacelle, as well as the tower itself, to greater aerodynamic forces, which can require extensive design modifications and can shorten the expected operating lives of the tower and its components. Finally, operation and maintenance (O&M) activities can also be more complicated and costly with increases in the elevation of the rotor.

3.1. Characterizing Candidate Sites and Site Selection The wind energy industry has adopted a convention by which annual average wind power densities and speeds are divided into seven power classes. It is also common practice to represent wind speed at a specified elevation above the land surface to allow comparative evaluations of sites within a given class to be made. To facilitate the identification of ideal wind regimes, the U.S. Department of Energy’s (DOE’s) National Renewable Energy Laboratory (NREL) has developed comprehensive wind maps for the United States that show the spatial distributions of these power classes. These maps were derived from meteorological data collected at thousands of locations. Figure 1 shows the wind resource distribution map for the contiguous 48 states. (Power density maps have also been developed for Alaska and Hawaii. However, since lands administered by the Bureau of Land Management [BLM] in those states are outside the scope of this PEIS, maps for those two states are not displayed here.) A more detailed discussion on the distribution of ideal wind regimes and more detailed maps showing ideal wind regimes on BLM-administered lands and their locations relative to existing electric power transmission lines are provided in Appendix B. Developers using currently available wind turbine technologies have found that sites with wind power densities at Class 4 or higher represent economically viable sites for a wind farm. These wind maps serve only as a preliminary screening tool for site selection. Developers must still investigate the properties of the wind regime at any candidate site in much greater detail before assigning a practical value to the site and deciding on a course of development. The principal limitation to the wind power distribution map displayed here is that it shows only the annualized average wind speeds and power densities. Two sites with identical annual average wind speeds and power densities may have arrived at those average values by entirely different paths. Sites whose average speeds and power densities are the product of widely varying instantaneous wind speeds over time are much less attractive than sites displaying lesser wind speed variations over time with few or no instances of excessive, potentially damaging wind speeds.

332

United States Department of the Interior, Bureau of Land Management

Figure 1. Wind Resource Distribution Map for the 48 Contiguous United States (Source: EERE 2004b)

The developer must understand the time variability of the instantaneous wind speed. The ideal wind regime is one at which the instantaneous wind speed is near the upper limit of the operating range of commercially available wind turbines for the greatest percentage of time over the course of the year, thus maximizing annual energy production. (See Section 5.3 for additional discussion on turbine operating ranges.) Therefore, the first step in any future wind farm development involves the collection of meteorological data (primarily wind speed and direction) at a potential candidate site for at least 1 year. For candidate sites in complex terrain or in areas with weather extremes, as many as 3 years of meteorological data may be necessary to support site development decisions. To realize their fullest value, the data must be collected at various locations within the site to support “micrositing” decisions (e.g., selecting the precise positioning of a wind turbine) and at various elevations to validate wind turbine decisions (e.g., selecting a turbine model and tower in which the rotor hub can be positioned at or near the elevation of maximum wind speed within its operating range and at a sufficiently high elevation so as to be above the chaotic and potentially damaging wind turbulence at or near the ground surface).[2] When the wind regime is precisely mapped, wind farms can consist of a variety of turbine models operating at different hub elevations to reach maximum sitewide efficiency. However, this type of composition complicates site development, construction, operation, and maintenance and may also complicate the collection and conditioning of the electric power that is generated. The use of various turbine models is unlikely; however, placing turbines at different hub elevations is technically feasible.

Wind Energy Technology Overview

333

3.2. Other Factors in Site Selection Site selection primarily involves matching wind regimes to turbine performance characteristics. The wind’s elevation profiles and variability over time and location, as well as the range of extant wind speeds, must be matched to turbine designs (and vice versa). All such efforts to find the perfect match are conducted with the intention of maximizing the capacity factor of each turbine. This capacity factor is the ratio of expected energy output to the turbine’s maximum rated power capacity, expressed as an annualized percentage (see additional discussion on capacity factors in Section 5.3). A wind farm’s expected capacity factor is the single greatest influence on the farm’s return on investment (ROI). Obviously, selecting a location with the highest average wind speed within the operating range of the proposed wind turbine for the greatest percentage of time is a principal site selection objective. In practice, many other circumstantial factors, such as transmission access and road access, substantially affect the costs of site development and O&M; therefore, they also play a key role in site selection.

4. WIND TURBINE TECHNOLOGIES The centuries-old history of efforts to harvest wind energy is fascinating, and an extensive discussion is beyond the scope of this PEIS. However, many excellent sources exist, including Gipe (1995), Hau (2000), Burton et al. (2001), Manwell et al. (2002), and Wilson (1994) and the references therein, as well as Web sites maintained by the DOE Office of Energy Efficiency and Renewable Energy (EERE 2004a), NREL (2004a), Sandia National Laboratories (2004a), the National Wind Coordinating Committee (NWCC 2004), and the American Wind Energy Association (AWEA 2004c). Sailing ships probably represent the earliest attempt to harness the wind. Windmills, the most familiar wind technology, have been used for myriad applications, most commonly to grind grain and pump water and crude oil. There is speculation that the earliest windmills went into service more than 3,000 years ago. More reliable historical documentation dates the earliest use of windmills to 200 B.C. in Persia (now Iraq) (Sandia National Laboratories 2004a). There is also evidence that windmills may have been used much earlier in China to drain rice fields, but the earliest dates of service are unclear. The use of windmills to generate electricity began in the late 19th century to provide electric power in rural areas, before the advent of far-ranging power transmission and distribution systems. Many windmills used in rural areas of Europe and the United States to pump water were converted for the production of electricity. Windmills such as the one shown in Figure 2 were used to generate small amounts of electricity, normally to satisfy the demand for electric power in the immediate vicinity. Windmills are the progenitors of the modern wind turbine.[3] In fact, they share a common fundamental function: converting the kinetic energy of the wind into the mechanical energy of a rotating shaft. Throughout the development and evolution of the windmill, a variety of designs have been explored. The evolution of wind turbine design has followed a similar path. The earliest windmills had their axis of rotation oriented vertically, and verticalaxis wind turbines (VAWTs) were also developed. Later-model windmills have their axis of

334

United States Department of the Interior, Bureau of Land Management

rotation in the horizontal position, and the analogous horizontal-axis wind turbines (HAWTs) also evolved. Although the orientation of the rotational axis defines the two primary design categories of wind turbines, many variations exist within each category. Early sailing ships and the earliest windmills utilized the principle of “aerodynamic drag” to capture wind energy. Applying this principle involves installing an obstruction in the path of the wind. Depending on how this obstruction is oriented and what it is connected to, the force of the wind striking it can cause work to be performed (e.g., propelling a square-rigged sailing ship through the water). The common instrument for measuring wind speed, the cup anemometer, is an example of a present technology that still utilizes aerodynamic drag. Machines utilizing aerodynamic drag are easy to construct, and they make few design or operational demands. However, despite the relative simplicity of aerodynamic drag machines, their overall efficiency is generally low.

Figure 2. Great Plains Windmill (Source: EERE 2004a)

No modern wind turbine operates on the principle of aerodynamic drag; instead, “aerodynamic lift” is utilized. When this principle is employed, the wind turbine’s blades do not obstruct the wind; rather, they direct its flow. The cross-sectional shape of all modern wind turbine blades is that of an “airfoil.” These blades are similar in shape and purpose to an airplane wing. Wind flowing around an airfoil creates two different regions of pressure: a low-pressure region on the convex or “suction” side of the airfoil, and a higher-pressure region on its concave or “pressure” side. The atmosphere’s attempt to return to pressure equilibrium creates the phenomenon of aerodynamic lift. However, whereas an airplane’s

Wind Energy Technology Overview

335

airfoils are oriented in such a way that aerodynamic lift helps the plane defy the laws of gravity (i.e., air pressure is lower above the wing than below it, causing the wing to “lift”), the orientation of a wind turbine’s blades relative to incident wind converts aerodynamic lifting forces into the rotation of the blades around an axis parallel to the direction of the wind.[4] Wind turbines utilizing aerodynamic lift can have power efficiencies up to 50 times greater than the efficiencies of turbines operating on aerodynamic drag (Wilson 1994). As noted previously, wind turbines have been developed with their axis of rotation in both the vertical orientation and the horizontal orientation. The VAWT traces its ancestry farther back in time than does the HAWT, to as early as 200 B.C. (Sandia National Laboratories 2004b). Modern VAWTs are variations of a design first introduced by French scientist Georges Darrieus around 1920. Figure 3 shows examples of a commercial VAWT in California and an experimental VAWT currently operating at a DOE test facility in Texas. In theory, both VAWTs and HAWTs should be able to capture the wind’s energy by means of the principle of aerodynamic lift. However, VAWTs have a number of practical advantages. Because their blades are always perpendicular to the prevailing wind, they do not need to be reorientated when the wind direction changes in order to operate at their maximum efficiency. Thus, both their design and the complexity of their required operational controls are simplified. They are generally easier to erect than HAWTs and can have serviceable components located at or near ground level, thereby greatly simplifying their O&M. However, some of those same design characteristics contribute to the VAWT’s intrinsic limitations. Many VAWT designs are not “free-wheeling” and must use an external energy source to start their rotation. Many also have limited wind speed operating ranges. VAWTs also have certain design limitations with respect to their maximum practical height. Most important to their commercial application, however, is blade reliability and working life. VAWT blades must pass through the “wind shadow” or wake of their rotational axis, which also serves as the machine’s primary support. This region typically exhibits a good deal of turbulence, which not only reduces power capture efficiencies but also subjects the blades to forces that are different and opposite to those that they experience when they are upwind of the center support; thus, significant engineering issues, such as fatigue, are introduced. Considerable research continues even today on how to overcome the intrinsic shortcomings of VAWTs, and VAWTs are being used as test platforms to generally advance the understanding of wind turbine technology. DOE’s Sandia National Laboratories play a key role in this effort. However, only a few commercial wind farms that utilize VAWTs have ever been developed, and none are anticipated in the foreseeable future. Wind farms at Tehachapi Pass in California; Pincher Creek in Alberta, Canada; and Cap-Chat in Quebec, Canada, utilize or have utilized VAWTs. The leading manufacturer of commercial VAWTs, FloWind Corporation, is no longer in business. No VAWTs have ever gone into commercial service in Europe (Gipe 1995). Therefore, it is likely that HAWTs will continue to dominate the commercial market in the foreseeable future. Additional discussion of VAWT technology is therefore unnecessary for purposes of this PEIS.

336

United States Department of the Interior, Bureau of Land Management

Figure 3. Examples of VAWTs (Left: FloWind Corporation VAWT at Tehachapi, California. Photo credit: R. Thresher. Source: Photo #04688, NREL 2004b. Right: Darrieus-design VAWT operated as a wind energy technology test bed by Sandia National Laboratories at the U.S. Department of Agriculture research station at Bushland, Texas; 138 ft (42 m) high, 112 ft (34 m) in diameter. Photo credit: Sandia National Laboratories. Source: Photo #01671, NREL 2004b.)

In recent years, HAWTs have become the predominant technology used in commercial wind farms; thus, they are the focus of discussion in this PEIS. Figure 4 shows an example of a typical front-facing HAWT. Within this category, Manwell et al. (2002) identified the following significant design variants: front-facing or rear-facing rotors and blades, rigid or teetering hubs, rotor rotation controlled by pitch or stall, number of blades (usually two or three), and free or controlled yaw motion. The majority of these design characteristics influence the overall performance of a turbine, but most have little or no influence on the environmental impacts of an operating turbine and thus are not discussed in further detail.

Figure 4. Typical Front-Facing or Upwind HAWT (GE’s 3.6-MW prototype wind turbine is an example of a front-facing HAWT. It is one of the largest HAWTs in existence, with a rotor diameter of 341 ft [104 m], giving a swept area of the blades of 91,432 ft2 [8,495 m2]. Rotor speed is variable between 8.5 and 15.3 rpm. The tower is constructed of concrete [lower portion] and tubular steel. Here, the turbine faces into the wind, which enters from the left.Sources: Photo adapted from EERE 2004c. Turbine specifications available from GE 2004.)

Wind Energy Technology Overview

337

5. IMPORTANT CONCEPTS OF MODERN HAWT OPERATION Figure 5 shows the major components of a HAWT. As noted previously, many factors influence the design and performance of modern wind turbines. This section focuses on the aspects of wind turbine design and operation that can have direct and/or cumulative environmental impacts. Also discussed here is the spatial arrangement of wind turbines on a wind farm, which can also result in environmental impacts.

5.1. Power Coefficients Intercepting the greatest practical cross-sectional area of wind creates the opportunity for capturing the greatest amount of energy; therefore, the primary design focus is on the rotor, which is the part of the turbine that actually extracts the wind’s energy. No mechanical device, including the wind turbine, is 100% efficient. The practical efficiency of a wind turbine is usually represented as its power coefficient, Cp , defined as that fraction of the wind power that may be captured by the turbine and converted to mechanical work (and, subsequently, electricity). The power coefficient of a wind turbine is almost entirely a function of the rotor’s efficiency.The power coefficient is represented by the following expression: P= ½ × Cp ×ρ × A × V 3,

(3)

where P = power output of the turbine, Cp = power coefficient of the rotor,

Figure 5. Major Components of a Modern HAWT (Source: EERE 2004c)

Anemometer: Measures the wind speed and transmits wind speed data to the controller. Blades: Most turbines have either two or three blades. Wind blowing over the blades causes the blades to “lift” and rotate. Front-facing turbines normally have three blades. Brake: A disc brake, which can be applied mechanically, electrically, or hydraulically to stop the rotor in emergencies.

338

United States Department of the Interior, Bureau of Land Management

Controller: The controller starts the machine at wind speeds of about 8 to 16 mph (13 to 26 km/h) and shuts off the machine at about 65 mph (105 km/h). Turbines cannot operate at wind speeds above about 65 mph (105 km/h) because their generators could overheat. Gear box: Gears connect the low-speed shaft to the high-speed shaft and increase the rotational speeds from about 30 to 60 rotations per minute (rpm) to about 1,200 to 1,500 rpm, the rotational speed required by most generators to produce electricity. The gear box is a costly (and heavy) part of the wind turbine, so engineers are exploring “direct-drive” generators that operate at lower rotational speeds and do not need gear boxes. Generator: Usually an off-the-shelf induction generator that produces 60-cycle alternating current (ac) electricity. High-speed shaft: Drives the generator. Low-speed shaft: The rotor turns the low-speed shaft at about 30 to 60 rpm. Nacelle: The rotor attaches to the nacelle, which sits atop the tower and includes the gear box, low-speed and high-speed shafts, generator, controller, and brake. A cover protects the components inside the nacelle. Some nacelles are large enough for a technician to stand inside while working. Pitch: Blades are turned, or pitched, out of the wind to keep the rotor from turning in winds that are too high or too low to produce electricity. Rotor: The blades and the hub together are called the rotor. Tower: Towers are made from tubular steel (shown here) or steel lattice. Some taller towers may incorporate concrete over the lower portions of their height. Because wind speed increases with height, taller towers enable turbines to capture more energy and generate more electricity. Wind direction: This is an “upwind” turbine, so-called because it operates facing into the wind. Other turbines are designed to run “downwind,” facing away from the wind. Wind vane: Measures wind direction and communicates with the yaw drive to orient the turbine properly with respect to the wind. Yaw drive: Upwind turbines face into the wind; the yaw drive is used to keep the rotor facing into the wind as the wind direction changes. Downwind turbines do not require a yaw drive, since the wind blows the rotor downwind. Yaw motor: Powers the yaw drive.

ρ = air density (typically 2.70 lb/m3 [1.225 kg/m3] at sea level and 59 F [15 C]), A = rotor-swept area, and V3 = cube of the incident wind speed. The power coefficient of the rotor has a theoretical maximum value of 0.593, called the Betz limit or Lancaster-Betz limit. This value is based upon the physical reality that even the most aerodynamically efficient turbine blade disrupts the airflow of incident wind, even before the wind front reaches the rotating blade. In actuality, the air molecules within the cross-sectional area swept by the rotor slow down as they approach rotating turbine blades and thus lose kinetic energy proportional to the cube of that velocity loss.[5] The power coefficient of the rotor can be thought of as a correction factor, introduced into the above power equation to reflect the reality that the rotor’s power-capturing efficiency is less than perfect. To calculate the power coefficient of the entire wind turbine, one simply

Wind Energy Technology Overview

339

has to introduce additional correction factors to represent the mechanical inefficiencies of the entire turbine drivetrain. However, for the purpose of this discussion, the power coefficient of the rotor is the source of greatest turbine inefficiency to the extent that drivetrain inefficiencies need not be discussed in detail. A comparison of the turbine efficiency equation above with the equation presented in Section 3, which represents the power inherent in the wind, leads one to fully appreciate how energy is produced by wind turbines. The Betz limit actually reflects the impossibility of extracting all the energy from the wind. Because the theoretical limit of rotor efficiency is always considerably less than 100%, the power produced by a wind turbine is always less than the power contained in the wind cross section that the turbine is intercepting. And because the rotor’s efficiency is the major contributor to the overall turbine efficiency, rotor design considerations are of paramount importance.

5.2. Turbine Power Curves The graphical representation of a turbine’s electric power output as a function of incident wind speed is known as the turbine’s power curve. At a fixed rotor speed, the power production of a wind turbine is defined by the following equation: Pel

= cp × ρ /2 × (vw)3 × A ,

(4)

where Pel = electric power (expressed in W, kW, or MW), cp = power coefficient of the turbine, ρ = air density (kg/m3), vw = wind speed (m/s), and A = swept area of the rotor (m2). Turbine manufacturers routinely use the power curve as a representation of their wind turbine’s official certificate of performance. Certain design features can have minor influences on the exact shape of the power curve; however, these influences notwithstanding, the power curves of virtually all commercial wind turbines are strikingly similar. As incident wind speed increases from zero to the “cut-in velocity,” the net power extracted from the wind becomes greater than that which is necessary to overcome the mechanical drag of the turbine’s drivetrain, and the excess power is used to begin producing usable electric power. With increasing wind speed, power production increases rapidly until the “rated velocity” is reached. At this wind speed, the turbine has reached its maximum electric power production capability. Power production continues at this maximum level with further increases in wind speed until the “cut-out velocity” is reached. At the cut-out velocity, the wind’s energy is so great that it can cause mechanical damage to major turbine components. To prevent such damage, designers introduce various controls (such as pitch and stall control on the rotor, mechanical braking of the rotor shaft, and clutching mechanisms on the rotor shaft) that can decouple the rotor from the remainder of the turbine drivetrain.[6] With the application of such controls, the electric power production drops precipitously to zero, and the turbine effectively becomes nonfunctional as a power

340

United States Department of the Interior, Bureau of Land Management

source. The range of wind velocities over which the turbine can produce electricity is referred to as its operating range; however, the maximum electric power production (i.e., the turbine’s nameplate rating) is achieved only at the upper end of the operating range. At incident wind speeds between the cut-in velocity and the rated velocity, power production is well below the nameplate rating. In general, commercial wind turbines have operating ranges between 2.5 and 25 m/s. (Table 2 in Section 6, which provides commercial wind industry profiles, has examples of operating ranges.) A turbine’s power output can be derived solely from engineering calculations. However, because the power curve represents the manufacturer’s guarantee of a turbine’s performance, theoretical calculations are also carefully validated with real-world measurements. To overcome myriad real-world variables that can affect power production, such empirical verifications of power output are based on the statistical evaluation of a large number of measurements. Hau indicates that measurements averaged over a minimum of 10 minutes are usually sufficient to account for the time variability of operating conditions (Hau 2000).

5.3. Capacity Factors Although the power curve is an accurate measure of the turbine’s ability to generate electricity from incident wind, it does not adequately describe expectations of real-world power production. Overlaying the relevant characteristics of a given wind regime (most importantly, the percentage of time the incident wind is at the uppermost portion of the operating range) and introducing additional correction factors that reflect the turbine’s technical availability (i.e., periods when the turbine is fully functional and not down for maintenance or repairs)[7] yield the capacity factor, the most realistic and reliable prediction of the energy yield for a given candidate site. Capacity factors are dimensionless, expressed as a ratio in which the turbine’s annual predicted energy production is divided by the energy it would produce if it operated at its nameplate rating continuously. Capacity factors are normally represented as annualized values to account for seasonal variations in wind regimes. In practice, the most efficient wind farms exhibit individual turbine capacity factors of 30 to 35% (EPRI 2001; DOE/TVA/EPRI 2003; Robichaud 2004). However, capacity factors as high as 45% have been observed (Manwell et al. 2002; EPRI 2001; McGowan and Conners 2000). Capacity factors of at least 25% are considered minimally necessary for a site to be considered economically viable (McGowan and Conners 2000). Because it is rooted in the real world, the capacity factor becomes a much more valuable tool for supporting decisions about wind farm development than the turbine’s power curve alone. The ideal site from a power production perspective is one that yields the highest capacity factor for each of the turbines. That being said, however, it is important to also recognize that power-producing potential, although important, is not the exclusive basis for site development decisions. Many other factors, including ease of site access, access to transmission lines, site development costs, the absence of sensitive ecosystems, and market price for energy, are always also considered in site selection decisions. Thus, it is often the case that the sites with the ideal wind regimes yielding the highest predicted capacity factors are not necessarily assigned the highest priority for development.

Wind Energy Technology Overview

341

5.4. Rotor Tip Speed and Tip Speed Ratio The rotor tip speed is the tangential velocity of the very end of the blade of a rotating rotor (i.e., the speed at which the tip of the blade moves around the circumference of the swept area of the rotor). Early wind turbine designs sought to match the rotor speed with the rotational speed requirements of the electric generator’s rotor.[8] However, modern designs utilizing more sophisticated and more reliable transmissions (Figure D-5) can adequately maintain the rotational speed of the electric generator’s central shaft at much lower rates of rotor rotation. This results in substantial additional benefits, including reductions in the bending moments on the blades and reductions in the forces on the turbine drivetrain, by minimizing the effective weight of the rotor. Wind turbine designers concern themselves not with the blade’s tip speed but rather with the tip speed ratio, which is defined as the ratio of the angular velocity of the blade tip to the mean velocity of the wind entering the rotor. For a given mean wind velocity and a rotor with a given number of blades, the design objective is to select a tip speed ratio that maximizes the opportunity for the incident wind to interact with the turbine blades and impart aerodynamic lift while simultaneously minimizing the disruptions of airflow ahead of the rotor blades. A rotor spinning too fast will present a greater obstruction to incident wind. Conversely, a rotor revolution that is too slow will allow large amounts of air to pass through the rotor’s plane without ever interacting with a turbine blade and imparting aerodynamic lift. At a given mean wind speed, the power coefficient of a turbine initially increases with an increasing tip speed ratio until a maximum is reached; beyond this point, performance actually decreases with further increases in the tip speed ratio. A more detailed discussion of this relationship and the influence of the Betz limit on turbine performance is provided by Burton et al. (2001). The ideal tip speed ratio is empirically derived and is inversely related to the number of blades. Because the rotor’s (and the turbine’s) power coefficient is directly related to the tip speed, controlling that ratio is a desirable objective. For a specific rotor operating in a given wind regime, the tip speed ratio at which maximum performance is achieved becomes the controlling design basis value. In addition to the basic performance relationship between the blade’s tip speed and the turbine’s power coefficient, two impacting factors are directly related to rotor rotation and tip speed: aerodynamic noise and shadow flicker. Both can influence turbine design decisions. The aerodynamic noise generated by a wind turbine is proportional to the fifth power of the tip speed.[9] Thus, small variations in tip speed can dramatically affect the noise profile of a wind turbine. Empirical data have led turbine designers to limit the tip speed to no more than 213 ft/s (65 m/s). Limiting the tip speed (which is proportional to the rotor’s rate of rotation and based on the swept area of the rotor) and limiting the distance to the nearest habitation to at least 1,312 ft (400 m) are expected to result in a turbine noise level at or near ambient levels (Burton et al. 2001). However, other factors, such as the height of the rotor and the topography of the site, can significantly influence the propagation of sound energy. In addition to the mathematical and geometric relationships between the rotor’s rate of revolution and the tip speed and the relationships between the tip speed ratio and the power coefficients, rotor revolution can also cause a visual phenomenon unique to wind turbines known as shadow flicker. Shadow flicker refers to the shadows that a wind turbine casts over structures and observers at times of the day when the sun is directly behind the turbine rotor from an observer’s position. Shadow flicker is most pronounced in northern latitudes during

342

United States Department of the Interior, Bureau of Land Management

winter months because of the lower angle of the sun in the winter sky. However, it is possible to encounter shadow flicker anywhere for brief periods after sunset and before sunrise. Empirical data suggest that shadow flicker can have a disorienting effect on a small segment of the general population. Empirical data also suggest that limiting the frequency of rotor rotation to below 2.5 Hz can mitigate the deleterious effects of shadow flicker.[10] Burton et al. (2001) indicates that limiting a (three-bladed) rotor revolution to 35 rpm will result in a blade passing frequency of 1.75 Hz (i.e., where the passing is between the sun and the observer). Increasing the spacing between a turbine rotor and the nearest observer to at least 10 rotor diameters also dramatically mitigates shadow flicker effects. Finally, another closely related phenomenon is “blade glint,” which is the reflection of sunlight off the surfaces of rotating blades. Such glint can also have a disruptive effect on some observers. However, as discussed elsewhere, the trend in the industry is toward longer blades. To control the resulting weight (and provide better aerodynamic properties), modern blades are now constructed almost exclusively of carbon composites or plastics, the natural surfaces of which are quite dull, especially relative to the steel and aluminum blades of the past. In the majority of cases, this technological development has made blade glint a relatively moot point with regard to modern turbines.

5.5. Blade Length and Tower Height Because the speed of the incoming wind cannot be controlled, attaining and maintaining the ideal tip speed ratio involves controlling the tip speed. There are two paths to this objective: changing the rate of rotor rotation or increasing the blade length. Increasing the blade length is often the preferred option for a number of engineering reasons. However, the law of diminishing returns is also at play here. Larger rotor diameters result in additional bending moments on the blades that must be accounted for. Longer blades mean additional rotor weight and increased strain on the mechanical drivetrain components. Research on alternative materials and fabrication procedures is being conducted by turbine manufacturers and under government sponsorship. (See Section 7 for more details on blade research.) Preliminary DOE-sponsored research on the technological impediments to scaling up current blade designs has identified the need to modify construction materials and processes (Griffin 2002) and the need to take a fundamentally different approach to airfoil design for extremely long blades (TPI Composites, Inc. 2002). To accommodate longer blade lengths, the turbine support towers have to be taller and more substantial. Irrespective of blade length, taller towers allow the rotor to operate in geostrophic wind regimes above the interferences introduced by surface topography. Principal performance factors affecting tower height selection include the wind profiles of the candidate site and the blade length of the turbine model selected. Costs of fabrication and erection are balanced against the performance advantages. Other factors related to site conditions can also influence tower height selection. These include access to the site by the larger equipment needed to transport towers (or tower segments), longer blades, and lifting/erection equipment; temporary amendment of site surface conditions to accommodate erection activities; and subsurface conditions that could affect the difficulty and the cost of constructing sufficient foundations for larger towers.[11] Installation costs, site access, and transportation logistics are important limiting factors with regard to tower height, and all

Wind Energy Technology Overview

343

factors must be considered in calculating improved performance with height. Developers are not likely to erect towers any taller than necessary to achieve economic power production (Steinhower 2004). The principal impacting factors that directly relate to a rotor’s geometry and the elevation at which it operates are listed below: • • •

• •

Larger rotors require higher, more formidable towers that are more expensive to fabricate and erect. Higher towers, in turn, are visible from greater distances, increasing the size of the impacted viewshed. Larger rotors allow for the economical capture of wind energy at slower rotor revolutions, which could lessen or completely eliminate the adverse viewshed impacts and bird-strike hazards. Larger rotors can rotate at frequencies less than the frequencies that induce shadow flicker. Larger rotors operating at fewer rotations per minute produce less aerodynamic noise than their smaller counterparts, which must rotate faster to capture the same amount of wind energy.

5.6. Grid Interconnection Issues The distance to an existing transmission line of suitable voltage and with reserve powercarrying capacity is a critical factor to consider with regard to future wind energy development projects, because the wind farm developer is expected to absorb the cost of establishing the physical link from the wind farm to the nearest existing transmission grid.[12] However, connecting to the grid is not necessarily a straightforward process. In reality, many factors related to grid interconnectivity can influence site development costs, design selection, initial installation and subsequent operating costs, and ROI schedules. To prevent disrupting the grid, the electric power generated at the wind farm must first be conditioned. This requires installing various power management and conditioning devices. Other devices are required to automatically isolate a wind farm from the grid during certain disruptive events. Sophisticated supervisory control and data acquisition (SCADA) systems are also required to ensure that the operating conditions of both the individual turbines and the overall wind farm and any rapid changes to grid interconnections are adequately controlled, in order to prevent the effects of potentially damaging disruptive events at the wind farm from cascading onto the grid. Although power management and control devices and SCADA systems certainly affect site development costs and the ability of the wind farm to interconnect to the grid, they represent only an incremental change to the footprint of the wind farm, and most have little or no direct or cumulative environmental impacts.[13] There are two notable exceptions, however: “voltage flicker” and lightning protection. If not adequately conditioned and controlled, wind farm power introduced onto the grid can result in voltage flicker. Voltage flicker occurs when changes to the network voltage occur faster than steady-state voltage changes that exist within the transmission system. Voltage flicker can cause perceptible changes to the brightness of incandescent lights that

344

United States Department of the Interior, Bureau of Land Management

draw power from the grid. Such changes, in turn, can have a disorienting effect on certain individuals. Transmission grid operators can be expected to require wind farm operators to establish power management systems capable of eliminating conditions leading to voltage flicker. Lightning protection is also required for wind farm components to prevent catastrophic impacts to the grid. Each individual turbine tower on the wind farm, as well as the electrical substation, must be protected, and control systems must be capable of isolating the wind farm from the grid during upset conditions caused by lightning. Although lightning protection technologies are available, their application in some wind farm settings may appreciably increase site development costs. Conventional lightning control involves providing a lowimpedance path for the lightning’s electrical energy to pass to the ground.[14] To establish adequate lightning protection for wind farms developed on rocky ground where there is no soil mantle, it may be necessary to drill one or more wells into which a current-conducting metal rod is inserted to extend the grounding path to the nearest aquifer. Moreover, the aquifer must be continuous over a large area rather than perched to provide reliable protection. In some western states within the study area, the nearest appropriate aquifer may be thousands of feet below a candidate wind site. Installation of such grounding wells will increase costs ↓ not only costs directly related to well installation, but also costs to support the hydrogeologic studies that may be required to identify appropriate aquifers.[15]

5.7. Variable versus Fixed Rotor Rotation Wind turbines can be designed to operate at both fixed and variable rotor rotation speeds. Of the two systems, variable-speed systems are preferred for a number of reasons related to overall wind turbine performance. However, while variable-speed machines can take fuller advantage of variations in the incident wind speed, the alternating current (ac) electricity they produce has a variable frequency that cannot be safely delivered to existing power transmission grids without conditioning. Variable-speed wind turbines are routinely connected “indirectly” to the grid to allow this power conditioning to occur at the wind farm. The majority of modern turbines include transmissions, clutches, and rotor shaft braking systems or aerodynamic stall features that act on the rotor blades to maintain the variations in a rotor shaft’s rotation within prescribed design limits. Such turbines are also equipped with SCADA systems that can adjust operating conditions (e.g., aerodynamic stall and blade pitch) to changing wind conditions. Variable-speed capability allows the turbine to operate at ideal tip speed ratios over a larger range of wind speeds. The most dramatic increase in performance is realized at lower wind speeds. Wind turbines with either a fixed or variable rotor rotation speed can be outfitted with either synchronous or asynchronous electric power generators.[16] In general, initial installation costs for asynchronous generators are lower, and the generators are generally very reliable. More important, asynchronous generators have mechanical properties that make them very suitable for wind turbine applications, including good overload capabilities and a relatively small generator slip.[17] Asynchronous generators can easily accommodate changes in the torque applied by the wind turbine’s rotor shaft (through the transmission), thus reducing overall mechanical wear and tear over the generator’s operating life. Because of the relatively constant operating conditions of asynchronous generators, turbines equipped

Wind Energy Technology Overview

345

with such generators are normally directly connected to the grid with little additional conditioning. The use of synchronous electric generators rather than induction generators improves the wind turbine’s overall power-generating performance and reduces the likelihood that the turbine will be a source of harmonic electric currents that can be disruptive to the power grid. However, initial installation costs are higher, and the power produced by synchronous generators must first be conditioned before delivery to the grid, further increasing installation and operational costs. As rotor diameters increase, the turbine’s rated power increases proportionally to the square of the rotor diameter. The amount of torque produced by the rotor shaft also increases markedly, placing significant operating demands on transmissions and generators. Industry and government researchers are now exploring the use of multiple generators or the use of multipole generators as a way of distributing torque and reducing its damaging effects on mechanical systems (Cotrell 2002). The use of multiple generators operating at different shaft speeds is also being investigated as a means of producing optimal levels of power at more widely varying rotor rotational speeds. Regardless of turbine and generator design choices, the attendant power-conditioning prerequisites do not themselves have additional environmental impacts of any significance. Operation at variable rotor speeds increases the complexity of the initial turbine design as well as the SCADA system required. However, it also promises to increase the overall longevity of major system components and to reduce O&M costs. Thus, turbines with variable-speed rotors can be expected to have less of an environmental impact over their operating lives than would their fixed-speed counterparts. Wind farms could consist of a mixture of fixed-speed and variable-speed turbines. Although the development costs of such a wind farm would be incremental, the increased sophistication of power management systems and SCADA systems and the expected greater O&M costs of such a configuration make such a wind farm unlikely. Wind farms consisting of identical turbines operating at different rotor elevations in order to take the fullest advantage of existing wind profiles are still a conceivable option, however. The following impacting factors relate to rotor operation at a variable rotation speed: •

•

•

Reducing the dynamic forces on the turbine drivetrain, extending the operating lives of major components, extending the maintenance intervals, and reducing the incidence of breakdowns, all of which would result in a smaller environmental impact over the life of the wind farm; Allowing the turbine to be “elastic” with respect to its interaction with the grid, thereby reducing the generation of power harmonics that can be disruptive to the grid; and Allowing the turbine to efficiently generate power at lower wind speeds, thus reducing the aerodynamic noise signal of the blades.

346

United States Department of the Interior, Bureau of Land Management

5.8. Micrositing and Site Development Once a candidate site has been selected and more detailed meteorological data have been gathered for a minimum of 1 year, site developers have the data necessary to make micrositing decisions (i.e., determine the precise location on the site at which the wind turbines will be located). The natural turbulence at the site due to the surface topography and obstructions and the induced turbulence of each wind turbine tower are the primary factors that govern turbine micrositing. Empirically derived nomographs[18] exist that indicate the necessary minimum distances for turbine placement from natural obstructions; however, they are often imprecise. Improving the methods for characterizing site-specific turbulence and understanding the influence of turbulence on site development make up a major ongoing R&D initiative (Section 7). It is possible that site developers may find it appropriate to remove some natural obstructions (e.g., trees) to mitigate turbulence caused by natural obstructions.[19] It is also reasonable to conclude, however, that the extent to which natural features of the site will be altered to improve the wind regime will be limited by site development costs. Thus, while tree removal is a feasible step associated with site development, major alterations of the existing grade over a large scale are not. It is also reasonable to expect that a site developer will seek to take advantage of economies of scale and develop a candidate site to its fullest potential. Thus, multiple turbines will likely be erected, and turbulence considerations will again be the primary factor governing their number and interspatial relationships.[20] Empirical nomographs that describe the induced turbulence of a wind turbine and its tower and that indicate the minimum distance of separation needed to avoid such interferences will likely be used to support micrositing decisions. (Research is ongoing to develop more precise modeling tools for characterizing the wind regimes on a site; see Section 7.) Avoiding the wind shadow of turbines will probably be a first priority in siting multiple turbines, and access to the indicated micrositing location will be of secondary importance. Pursuing economies of scale in site development will amortize site characterization and site development costs. However, the extent to which a site will be developed can have additive effects on many of its impacting factors. Primary impacting factors related to site development and micrositing include the following: •

• • • • •

Potential for ancillary activities, such as tree and vegetation removal, that will result in surface scarring and additional impacts to the viewshed beyond the impact of turbine visibility itself; Increased potential for fugitive dust, proportional to the area of disturbed ground surface; Potential for invasive species being established in disturbed areas before indigenous vegetation can be reestablished; Potential for bird strikes, generally proportional to the number of turbines installed; Increased time required for construction, with proportional increases in both the magnitude and duration of impacts related to construction; Potentially additive impacts from individual turbines, including noise and viewshed impacts; and

Wind Energy Technology Overview •

347

Proportional increases in O&M costs, including costs to deal with wastes associated with system maintenance and repair.

6. COMMERCIAL WIND ENERGY INDUSTRY PROFILES This section provides an overview of the existing commercial wind energy industry within the study area. The AWEA compiles and maintains data on commercial wind farms.[21] The review and analysis of these data provide a reasonable basis from which to anticipate the characteristics of future wind farms. Industrywide reviews of the commercial utility-scale wind energy industry have identified the following important trends, each of which will greatly influence future wind farms. •

•

•

• • •

•

In general, average individual wind turbine power-generating capacities have steadily increased in North America, from 500− 750 kW in the late 1990s to megawattcapacity turbine installations beginning in 1999, resulting in typical wind farm generating capacities of 50 MW or larger (Kaygusuz 2004). The (worldwide) average growth rate of the cumulative installed wind energy powergenerating capacity over the period 1998 to 2004 has been about 30% per year (Kaygusuz 2004). As the understanding of aerodynamics has been increasing and as designs have been defined, wind turbine efficiencies have been increasing, especially for turbines with larger rotor-swept areas. Average annual yields per unit of rotor-swept area (RSA) have increased by more than 50% as rotor diameters have increased from 66 to 262 ft (20 to 80 m) (Milborrow 2002). Wind turbines now have power-generating capacities of as much as 600 W/m2 of RSA. Three-bladed, upwind turbines dominate the commercial utility-scale market (Milborrow 2002). The majority of wind turbines run at fixed rotor speeds and utilize induction generators. However, newer models equipped with sophisticated electric power conditioning controls have rotors that run at a variable rotational speed (Milborrow 2002). Newer-model turbines tend to run at slower rotor rotational speeds but have relatively high energy capture/conversion efficiencies (Milborrow 2002). About the AWEA

The American Wind Energy Association (AWEA) is a national trade association that represents wind power plant developers, wind turbine manufacturers, utilities, consultants, insurers, financiers, researchers, and others involved or interested in the wind energy industry. The AWEA provides up-to-date information on wind energy projects operating worldwide and projects under development, and it conducts technology and policy development activities related to wind energy.

348

United States Department of the Interior, Bureau of Land Management The AWEA compiles and regularly updates relevant domestic and worldwide statistics on the wind energy industry and makes them available to industry participants, the interested general public, and the news media. These data are available at the association’s Web site at http://www.awea.org. Also available on the AWEA Web site is access to various wind-energy-related information resources, including wind energy fact sheets and a catalogue of related publications. The AWEA also publishes a weekly newsletter devoted to wind energy news and hosts an annual national conference, WINDPOWER. Detailed information on AWEA activities and services can be obtained by visiting the Web site. Information developed by the AWEA has been incorporated into this PEIS without independent verification. The BLM does not endorse the AWEA and does not make any warranty regarding the accuracy or completeness of the data it provides.

Although the commercial wind energy market in the United States has existed for some time, it has only recently (since 1999) begun to experience substantial growth, with calendar years 2001 and 2003 witnessing the two largest single-year’s growth. Figure 6 graphically depicts the rise in wind energy capacity (nameplate ratings in megawatts of electricity; the bars in the foreground represent capacities added annually; the bars in the background represent cumulative power capacity) over the period from 1981 through 2003. Data published by the AWEA indicate that the total installed capacity for all domestic commercial wind energy as of December 2003 was 6,374 MW, with 1,687 MW coming on line in 2003, which was a 36% increase from the capacity at the previous year’s end (AWEA 2004d). Calendar year 2003 compared favorably with the previous year, showing a worldwide increase in capacity of 6,868 MW to reach a total of 31,128 MW and a U.S. increase of 410 MW to reach a year-end total of 4,685 MW, which represents 15% of the world’s market (AWEA 2003a). Of the current total domestic capacity of 6,374 MW, 2,999.7 MW (or 47%) is being produced in the 11-state study area of this PEIS. The increase in overall generating capacity has been accompanied by a steady increase in individual turbine proportions and capacities. In the late 1980s, average turbine power outputs averaged 450 kW. Outputs increased to an average of 600 to 750 kW by the late 1990s. Now, individual turbines with ratings greater than 2 MW (2,000 kW) are commonplace (McGowan and Connors 2000). Figure 7 shows the distribution of wind energy power-generating capacity across the United States. The numbers represent power capacities of utility-scale wind farms only, all of which deliver power directly to the electric power transmission grid. Additional power capacities from distributed energy systems are not included. The power capacities represent nameplate ratings and are rarely realized in practice. (See the discussion on typical capacity factors in Section 5.2.) Within the 11-state study area for the PEIS, the total installed wind energy capacity is 2,999.7 MW.

Wind Energy Technology Overview

349

Figure 6. U.S. Installed Capacity (MW) for 1981 through 2003 (Source: AWEA 2004d. Reprinted with permission. Courtesy of the AWEA.)

Figure 7. Wind Energy Projects in the United States (Source: Adapted from AWEA 2004a. Reproduced with permission. Courtesy of the AWEA.)

Table 1 lists the commercial wind energy projects completed in 2003. Projects completed within the 11-state study area are in bold type. The projects listed in the table represent new wind farms and phased expansions, or “repowering” of existing wind farms (i.e., replacing existing turbines with ones of newer design). Facility expansions and repowering activities are not expected to have the same array and magnitude of impacting factors as would a completely new facility. By definition, such site modifications are outside the scope of this PEIS. In general, the number of manufacturers of wind turbines has greatly decreased from earlier years. In fact, a number of manufacturers have gone out of business. However, also represented in this decline are a number of mergers among manufacturers.

350

United States Department of the Interior, Bureau of Land Management Table 1. Wind Energy Projects Installed in 2003a

State Alaska Arkansas California California California California

Project Name Selawik Wind Project Bitworks High Winds Mountain View III

Illinois

CalWind II CECrepower Whitewater expansion Karen Avenue II Colorado Green Lewandoski wind farm Mendota Hills

Iowa Iowa

Flying Cloud Henry Hills

Iowa Iowa Iowa

Lenox Wall Lake Sibley Hills

Minnesota Minnesota

Chanarambie Moraine Wind Power Project Farmers’ cooperative corporations McNeilus

California California Colorado Idaho

Minnesota Minnesota Minnesota Minnesota Minnesota Minnesota Minnesota Minnesota Minnesota Minnesota

McNeilus Viking McNeilus Fairmont Missouri River Energy Systems Shaokatan Power Partners McNeilus Don Sieve Wind Farm

Minnesota Pipestone School District New Mexico New Mexico Wind Energy Center New Mexico Llano Estacado Wind Ranch at Texico North Dakota North Dakota Ohio Oklahoma Blue Canyon Wind Power

Location Selawik Prairie Grove Industrial Park, Washington County Solano San Gorgonio Sacramento Tehachapi

Combine Hills

Turbine Manufacturer AOC

No. of Wind Turbines 4

NEG Micon

1

Vestas Vestas Vestas Vestas

90 34 15 13

4.5

FPL Energy `PPM Energy SMUD CalWind Resources, Inc. Cannon Power Corp.

GE Wind

3

San Gorgonio Near Lamar

4.5 162 0.216

San Gorgonio Farms GE Wind Bob Lewandoski

GE Wind GE Wind

3 108 2

Lee County, near Mendota Near Spirit Lake Osceola County, near Sibley Lenox Wall Lake Near Sibley

50.4

Navitas Energy

Gamesa Eolica

63

43.5 3.6

GE Wind Gamesa Eolica

29 2

NEG Micon Vestas Vestas

1 1 1

85.5 51

PPM Energy Northern Alternative Energy Lenox Municipal Wall Lake Municipal Northern Alternative Energy enXco PPM Energy

GE Wind GE Wind

57 34

22.8

DanMar & Associates

Suzlon Energy

24

22.8

Garwin McNeilus

NEG Micon

24

16.5 12 6 1.9 1.9

NEG Micon NEG Micon NEG Micon NEG Micon

11 8 4 2 2

Gamesa Eolica

2

NEG Micon NEG Micon

1 1

NEG Micon

1

NEG Micon

1

204

Garwin McNeilus Project Resources Garwin McNeilus SMMPA Missouri River Energy Systems Northern Alternative Energy Garwin McNeilus Diversified Energy Solutions Diversified Energy Solutions Pipestone School District FPL Energy

1.32

Cielo Wind Power

Vestas

2

40.5

FPL Energy

GE Wind

27

21

FPL Energy

GE Wind

14

Bowling Green Zilkha Renewable Energy & Kirmart Corp. FPL Energy FPL Energy Eurus

Vestas NEG Micon

2 45

GE Wind GE Wind Mitsubishi

34 34 41

Murray County Pipestone & Murray Counties

Near Minn. Highway 56 Murray County Fairmont Worthington Lincoln County, near Hendricks

162 22.44 9.9 8.58

0.75 0.66 0.66

1.6

Lincoln Co.

1.65 0.95

Lincoln Co.

0.9

Minnesota

Oklahoma Oklahoma Oregon

Capacity Developer (MW) 0.2 Kotzebue Electric Association 0.1 Bitworks, Inc

0.75 Quay, DeBaca Counties

Near Edgeley Near Kulm Bowling Green North of Lawton

3.6 74.25

Near Woodland Near Woodland

51 51 41

GE Wind

136

Wind Energy Technology Overview State

Project Name

Pennsylvania Waymart Pennsylvania South Dakota South Dakota Texas Texas Texas Texas Washington Wyoming a

Meyersdale Highmore

Location Clinton & Canaan Township Somerset Near Highmore

Capacity Developer (MW) 64.5 FPL Energy

351 Turbine Manufacturer GE Wind

No. of Wind Turbines 43

30 40.5

FPL Energy FPL Energy

NEG Micon GE Wind

20 27

Rosebud Sioux

0.75

DisGen

NEG Micon

1

Brazos Wind Ranch 90 miles south of Lubbock Sweetwater Sweetwater Hansford County, Texas Indian Mesa Nine Canyon, Phase Benton County II Evanston Evanston

160 37.5 3

Cielo Wind Mitsubishi Power/Orion Energy DKR/Babcock-Brown GE Wind FPL Energy Vestas

3 15.6

Energy Northwest

Vestas Bonus

1 12

144

FPL Energy

Vestas

80

160 25 1

Bold type indicates projects within the 11-state study area. Source: Adapted from AWEA (2003b). Reprinted by permission. Courtesy of the AWEA.

Table 1 lists the manufacturers of commercial wind turbines whose products were installed in U.S. wind farm projects in 2003. Although there are many other manufacturers, those listed in Table 1 nevertheless represent a cross section of vendors. One should therefore take a more careful look at the turbine models offered by these vendors. Table 2 lists the ranges of values for critical parameters of wind turbines installed in 2003. Although it is assumed that installations in 2003 constitute a reasonable representation of the most current facility installations and expansions, there is still a possibility that future wind farms will utilize turbines from other manufacturers. Nevertheless, it is reasonable to assume that the turbines installed in 2003 met the technical requirements of the sites at which they were installed. It is therefore also reasonable to assume that future developments at sites with similar wind regimes may also utilize turbines with these approximate specifications. It is not the BLM’s intention to endorse any specific equipment manufacturer.[22] Consequently, rather than present the specifications of individual turbines, the table displays a range of values for each parameter that is addressed. Only design specifications that were readily available from manufacturers’ Web sites are included in the range calculations. Not always accurately reflected in the range value displayed, but nevertheless important for anticipating future wind farm characteristics, is the fact that many manufacturers offer modules rather than complete turbines, providing a number of options for each major component. Thus, the developer can custom build a turbine that is precisely suited to a particular site’s wind conditions and to the stipulations of a particular interconnection agreement with the transmission line operator. For the reader’s convenience, the Web sites for the manufacturers whose turbines are represented in the range of values displayed are provided as footnotes to Table 2. The data displayed in Table 1 appear to support the following conclusions about the characteristics of future wind farms. Notwithstanding the fact that calendar year 2003 was an exceptional year for industry growth, a reasonable assumption is that the projects that went on line in 2003 reflect the state of the technology with respect to commercially available wind turbines. Another reasonable assumption is that the wind turbine models installed in 2003 offered operating parameters that matched well with the specific conditions at the sites at which they were installed. A further assumption is that future sites with wind characteristics

352

United States Department of the Interior, Bureau of Land Management

similar to those at sites developed in 2003 will utilize turbines with operating parameters similar to those displayed in Table 2. Table 2. Specifications for Wind Turbines Installed in 2003a Parameterb Power (nameplate rating)d Turbine type Cut-in speed (m/s) Nominal wind speed (m/s) Cut-out speed (m/s) Rotor diameter (m) Rotor-swept area (m2) Rotor speed (rpm) Rotor hub height (m)e Tower construction material

a

Tower weight (kg)f Nacelle weight (excluding rotor) (kg)e,f Rotor weight (kg)g Total weight (kg)h

<9,070–30,839 <37,188–158,300

Data presented in this table represent the range of options offered by the manufacturers listed in Table 1 for which data were readily available. No attempt was made to identify the specific turbine models used in the 2003 projects. Instead, all available models of the manufacturers listed were used to compute the ranges. Additional information on individual turbine models is available at that turbine manufacturer’s Web site. Web sites are listed here as follows: Atlantic Orient Corp. Bonus Energy Products Gamesa Eolica GE Energy Mitsubishi Electric NEG-Micon Suzlon Energy Vestas Wind Systems A/S

b

Ranges for Available Optionsc 200 kW–3.6 MW Upwind HAWT 2.5–4.0 11–16 25 30–104 706–8495 8–46 30–120 Cylindrical or tubular steel, hot-dip galvanized lattice steel, combination concrete and tubular steel <30,500–216,780 <19,954–55,329

http//www.aocwind.net/specs.htm http//www.bonus.dk/uk/produkter/ http//www.gamesa.es/ingles/nucleos_negocio/ga mesa_eolica/presentacion/ presentacion.htm http//www.gepower.com/businesses/ge_wind_en ergy/en/products.htm http//www.global.mitsubishielectric.com/bu/wind power/index2_b.html http//www.neg-micon.com (Only limited data are available; data are not included in ranges presented in the table.) http//www.suzlon.com/technical_data http//www/vestas.com/produkter/

By industry convention, all specifications are presented in metric units. c Range does not include data from AOC Model 15/50 turbine, the use of which has been confined to distributed energy systems in remote locations. d Range represents individual turbine nameplate ratings. Additional specifications for power generation and management devices are available at the manufacturers’ Web sites. However, since these devices have little or no influence on the environmental impacts of an operating wind turbine, they are not represented here. e Rotor hub height is considered to be approximately equivalent to tower height, measured from ground elevation. f All weights are approximate; the weight range is based on models manufactured by Vestas Wind Systems A/S and Bonus Energy Products only. The weight of the smallest tower option was not available. g Rotor weight includes active pitch control equipment, if present. h Nacelle weights may differ as a result of drivetrain component selection. Source: Derived from AWEA (2003b).

Wind Energy Technology Overview

353

Following a strategy of extracting the maximum potential wind energy from a given site will minimize the overall environmental impacts. However, phased site development can cause changes to some impacting factors related to site development and operation. Some of the impacts in phased development will simply be additive over time. For example, the noise levels from individual turbines will be logarithmically additive for each turbine installed; however, because of the expected distances between turbines in a typical wind farm, the addition of each turbine will increase the area potentially impacted by noise, but it will not substantially increase the average or maximum noise levels throughout that area. Site topographic features can also greatly influence noise levels at a given distance from a noise source. See Section 4.5 of the PEIS for a detailed discussion on noise generation and propagation and Section 5.5 for a discussion on potential noise impacts from wind farms. Impacting factors associated with turbine foundations and erections will also be additive within a given phase of development and then reoccur during subsequent development phases, although not necessarily at the same magnitude or for the same duration. Other impacts related to initial site development may not reoccur at all during subsequent site expansions. For example, if it is assumed that the initial site development plan accounts for all future site expansions, a single main site access road can be selected and constructed as part of initial site development, and it can continue to serve as the site access road for subsequent phases of development. In such a scenario, only the expansions of on-site roads would be impacting factors in later development phases.

7. WIND ENERGY TECHNOLOGY RESEARCH AND DEVELOPMENT A review of the current state of the commercial wind turbine market can provide a basis for predicting the types of turbines that are likely to be installed at future sites. However, it is also reasonable to predict that future site developers will avail themselves of technological advances and improved performance models. Therefore, a brief review of wind energy industry R&D activities is warranted. Although much of the R&D effort has been undertaken by the equipment manufacturers, the federal government also provides support. The discussions below are confined to R&D activities unique to the commercial wind energy industry. Note that R&D efforts to improve the design and performance of many of the major components of a wind turbine, such as transmissions and electrical generators, are also ongoing within the respective industry sectors. Likewise, R&D efforts in the general area of monitoring and control systems continue as well. Although these R&D efforts are not discussed here, it is assumed that wind farm developers and/or equipment manufacturers will incorporate technological advances from these other sectors into their wind farms and turbines at appropriate times.

7.1. Industry-Sponsored Research and Development Leading equipment manufacturers are already engaged in R&D on many aspects of their products. Their primary objective is to maintain or improve their competitive positions in the markets in which they operate. R&D can also help them conform to quality standards (Section 8).

354

United States Department of the Interior, Bureau of Land Management

Industry research focuses on improving the reliability of major components, improving overall efficiency, reducing manufacturing costs, and mitigating the adverse aspects of individual products. For example, manufacturers who hope to participate in the European wind energy market are exploring ways to mitigate the noise signals of their equipment. Because most wind farms in Europe are located close to inhabited areas, controlling noise is critical to maintaining market position. In its overview of worldwide wind energy industry trends, Shikha et al. (2003) found that continuous improvements were being made to applied technologies in the expanding wind energy industry. They found that energy output capacities of individual turbines increased 100-fold in the 15 years ending in 2003, while the overall weight of turbines was halved in the 5 years ending in 2003, and the noise emitted was halved over the 3-year period ending in 2003. Steady gains were attributed to a number of factors, including improved aerodynamics, improved structural dynamics, and improved micrometeorology, which resulted in precise turbine siting at the most ideal location. Additional improvements were attributed to the increase in rotor size and improved blade performance. Together with the benefits derived from reduced rotor weight, overall improvements in the drivetrain design and the reliability of individual components also resulted in a reduction in O&M costs. It is estimated that O&M costs constitute as much as 10 to 15% of the unit energy costs of a new wind farm; however, O&M costs increase to 20 to 30% near the end of the farm’s design life (McGowan and Connors 2000). However, O&M costs are also expected to rise slightly over the design life of the turbine. Steady improvements in drivetrain design and efficiency are expected to reduce O&M costs from a U.S. average of $0.01/kWh in 1997 to $0.005/kWh by 2005 (McGowan and Connors 2000).

Source: Photo # 12449, NREL 2004b.) Figure 8. Lattice-Type Wind Turbine Tower in South Dakota (A Vestas Model V17 wind turbine mounted on a lattice-type tower in Gary, South Dakota. Photo credit: Energy Maintenance Service, Inc., Sept. 1, 2002.

Wind Energy Technology Overview

355

Manufacturers are also adopting modular design strategies that allow the replacement of individual turbine drivetrain components, thereby reducing downtime and costs. Often such strategies are further enhanced by equipping towers with internal lifting devices that allow the replacement of individual components without the necessity of bringing heavy-duty lifting devices to the site to remove the rotor assembly and/or the entire nacelle. Although the majority of industry R&D initiatives focus on improving the design and efficiency of rotors and turbine drivetrain components, some innovative tower designs and materials can also affect future wind farms. Early wind farms utilized lattice-type towers (Figure 8). However, smooth-skinned, tapered steel towers now dominate the commercial utility-scale market. The size and weight of the steel towers required for larger turbines increase installation costs and createsignificant problems related to the transportation of both the tower segments and the cranes required for their erection. A number of innovative tower designs and erection methodologies have been developed to overcome these impediments. Towers that can be erected by using mobile, temporary elevators have been developed, obviating the need for independent cranes and thus greatly simplifying erection costs and reducing transportation logistics (e.g., see Valmont 2004). A government-sponsored study completed in May 2001 identified a number of unique tower erection strategies and evaluated each against its impact on the overall cost of energy produced (Global Energy Concepts, LLC 2001). Two technologies were evaluated in depth and compared with conventional crane technologies. The study concluded that one of the two alternative erection methods compared favorably to conventional cranes for 1.5-MW and larger turbines, but it was more expensive than conventional cranes for smaller turbines. The study further postulated that alternative erection methodologies might be favored over conventional cranes for sites with complex terrain or difficult access, but they could be at a disadvantage at sites with significant wind shear. Other developments include constructing towers of tubular carbon composites in an integrated pyramidal shape, resulting in stronger and substantially lighter towers (e.g., IsoTruss Structures, Inc. 2004). Again, such lighter towers can substantially reduce transportation logistics and reduce site development costs.

7.2. Government-Sponsored Research and Development Government-sponsored research and government-industry partnerships also account for a major portion of ongoing R&D efforts. DOE/EERE is the principal funding agency for government-sponsored research. Government participation also includes the personnel and facilities of NREL in Boulder, Colorado, and Sandia National Laboratories in Albuquerque, New Mexico. Government-industry partnerships proceed under the auspices of DOE’s Cooperative Research and Development Agreement (CRADA) program. Under CRADA programs, government and industry collaborate to identify and better understand the fundamental science and engineering issues critical to technology advancement. Government personnel also conduct tests on prototypes and develop software that aids designers. Industries then have access to the published reports on CRADA research and use their contents to shape their own additional technology R&D. The government-industry partnership in DOE’s Wind Energy Program is known as the Wind Partnerships for Advanced Component Technologies (WindPACT).[23]

356

United States Department of the Interior, Bureau of Land Management

DOE’s R&D objectives and strategies are outlined in Wind and Hydropower Technologies Program; Wind Energy Program Multi Year Technical Plan for 2004–2010 (EERE 2003). The overall strategic objective is to protect the nation’s energy security by fostering the development of technologies that utilize a diverse supply of affordable and environmentally sound energy. Specific research objectives are defined in terms of reducing the ultimate costs of electricity generated by wind energy. Individual research initiatives, or technology improvement opportunities (TIOs), are distributed throughout all segments of the wind energy industry. The research initiatives of greatest importance to the utility-scale sector of the industry include improving the viability of low-wind-speed technology and facilitating the application of technologies and technological advances by engaging in fundamental research, developing quality standards and certification programs, conducting field verification tests, and analyzing and addressing technological and market impediments. Researchers have identified a number of TIOs, including the following: • • • •

Advanced drivetrain designs that use rare-earth permanent magnets for excitation, reduced gear box stages, and low- and medium-speed generators; Advanced power electronics that allow variable-speed operation while improving overall power capture/conversion efficiencies; Advanced rotors that use adaptive blades; and Advanced tower designs and materials that either reduce erection costs and simplify transportation logistics or are fabricated completely on site.

Research critical to the advancement of utility-scale turbines, especially in lower wind power classes,[24] includes the development of (1) advanced rotors; (2) a more complete understanding of a site’s atmospheric dynamics; (3) improved generator, drivetrain, and power management subsystems; and (4) better integrated operational controls. Turbines harvesting wind at lower wind classes are expected to need larger RSAs and operate at higher hub elevations. Rotor development focuses on the development of blades that are stiffer and stronger but also more slender, lighter, and more flexible (i.e., more adaptive to the dynamic forces they will encounter during operation). These apparently mutually exclusive characteristics hold the key to the successful advancement of large turbines. Although blade technology has already advanced significantly, it is thought that new materials and fabrication methods, as well as new design philosophies and criteria, will be necessary to support further substantial technological advances. Prototype blades made of long-fiber carbon composites are being tested for durability, and manufacturing processes are being refined.[25] If successful, this research will lead to turbines with greater RSAs and power-capturing efficiencies. There are, nevertheless, technical and economic limits to blade length. Rotor weight increases by the cube of its swept area, while the rated power efficiency increases by the square of the swept area. Consequently, there are some diminishing ROIs in the development of extremely long blades. Furthermore, with regard to extremely long blades, gravitational forces and torsional forces on the hub and the rotor shaft will become controlling forces in turbine design. Finally, as noted earlier, the torque produced by the rotor shaft increases with the square of the rotor diameter, thus significantly increasing the demand on transmissions and generators to withstand such increased torque moments. Some anticipate that the point at which these adverse forces will preempt rotor size expansions will be reached

Wind Energy Technology Overview

357

at rotor diameters of 256 ft (200 m), although the introduction of lightweight composites, such as fiber-reinforced plastics, may extend the practical rotor diameter to even greater values (Milborrow 2002). Other possible dividends from increased blade length include lower operating costs and less aerodynamic noise. However, another real-world consequence of the use of very long blades is significant transportation logistics. Research conducted by Sandia and its contractor has explored the possibility of manufacturing turbine blades at the wind farm location (TPI Composites, Inc. 2003). The research concluded that on-site manufacturing was fraught with significant quality control issues and not feasible at this time. However, fabrication of the blades at nearby manufacturing sites (i.e., sites specifically constructed to support blade fabrication for use at a particular wind farm) was still considered feasible, since such a strategy would significantly reduce transportation distances and, if located judiciously, would significantly simplify transportation logistics. Other scaling and related logistics issues associated with transportation and erection also accompany any consideration for significantly enlarging wind turbines. WindPACT research initiatives will identify these obstacles and evaluate ways to overcome them. Up to this point of development, rotor aerodynamic design criteria have borrowed heavily from aerodynamic codes[26] developed in the aircraft industry. However, these codes do not reflect the aerodynamic conditions in which a wind turbine operates to a sufficiently high level of precision. New code development efforts are necessary to better understand the aerodynamic forces affecting both the performance and reliability of turbine rotor blades. Newly developed and validated codes will expedite the development of design criteria for longer, lighter, and more slender adaptive blades that can withstand dynamic forces and also impart minimum loads on the turbine drivetrain. A more complete understanding of aerodynamic forces impinging on turbine blades will also allow designers to mitigate aerodynamic noise impacts. Another facet of research is the development of a semiempirical noise prediction code to be used by rotor and blade designers to ensure that new rotor systems have acceptable noise signatures. As turbines become larger and operate at higher rotor hub heights, additional information about the atmospheric dynamics at these higher altitudes will be necessary to support design and micrositing decisions. It has already been established that the tallest turbines may be influenced by jet stream turbulence, especially by what are known as nocturnal jets (DOE 2002). Such turbulence is routinely present in low wind power classes, especially in the Great Plains regions. Successful advancement of wind turbines in such areas, especially in lower wind power classes, requires a much more complete understanding of jet stream turbulence and candidate site aerodynamics. Other research initiatives on improving the power generation and management performance of the electric generator will have a direct impact on the interconnectivity of turbine power into the electrical grid but are expected to have little impact on environmental factors. Nevertheless, such improvements in overall turbine performance efficiency can be expected to reduce the mechanical noise emanating from the turbine blades and drivetrain components, as well as to reduce the number of breakdowns and maintenance shutdowns. Finally, research on the advancement of integrated systems and controls attempts to enhance the precision with which turbines are monitored and controlled, promising better control of yaw and blade pitch to maximize performance. Such research pays its greatest dividends by improving the interconnection opportunities for wind farms. However,

358

United States Department of the Interior, Bureau of Land Management

maintaining the turbine’s operation at the highest performance level is also expected to improve overall reliability and reduce unwanted impacts that are manifestations of inefficiency (such as aerodynamic noise).

8. TESTING AND VERIFICATION PROGRAMS DOE sponsorship of wind energy R&D also extends to field testing and verification programs. NREL and Sandia personnel, in collaboration with representatives of the Electric Power Research Institute (EPRI), other wind energy industry participants, and individual wind farm operators, conduct evaluations of wind project development experiences and conduct field verifications of critical aspects of operational wind farms. The verification efforts help to identify issues related to site development, as well as design and operation, and provide the empirical basis for additional research on how to address or eliminate those issues. Published reports provide the opportunity for transferring lessons learned to other interested parties. Additional details about these verification programs and the published reports are available on the NREL and Sandia Web sites (NREL 2004c; Sandia National Laboratories 2004d).

9. STANDARDS AND CERTIFICATIONS One clear indication of the maturation of the wind energy industry is the development and application of quality standards. International standards are already largely in place. Analogous U.S. standards are under development. Standards related to wind energy turbines promulgated by the International Electrotechnical Commission (IEC) are listed in Table 3. The AWEA is the U.S. industry representative to this international standard-setting body. Many turbine manufacturers voluntarily conform to these standards to maintain their competitive position in the marketplace and to better guarantee the connectivity of windgenerated electric power to transmission grids. Conformance with international standards is a requirement for some wind farms in Europe. U.S. wind energy industry consensus standards have been under development since 1974. The AWEA is the lead organization in domestic standard development. The development process involves the participation of various industry organizations, including the American Society of Mechanical Engineers (ASME), American Society for Testing and Materials (ASTM), American National Standards Institute (ANSI), National Fire Protection Association (NFPA), American Gear Manufacturers Association (AGMA), and Institute of Electrical and Electronics Engineers (IEEE). Personnel from NREL and Sandia also participate in standards development. Domestic standards are expected to parallel and be compatible with IEC standards in order to ensure that American manufacturers maintain their access to European markets.

Wind Energy Technology Overview

359

Table 3. International Wind Turbine Standards Standard No.

IEC 61400-1 IEC 61400-1 Ed 2 IEC 61400-2 IEC 61400-12 IEC 61400-11 IEC 61400-13 IEC 61400-22 IEC 61400-23 IEC 61400-21

Title

Wind Turbine Safety and Design Wind Turbine Safety and Design Revision Small Wind Turbine Safety Power Performance Noise Measurement Mechanical Load Measurements Wind Turbine Certification Blade Structural Testing Power Quality

In addition to quality standards for the design and construction of major turbine components, international standards are in place for the certification of turbines and ancillary systems by independent third-party auditors. Leading equipment manufacturers routinely submit their products and systems to such certifications so that they have evidence that their quality and performance goals have been met. Personnel from NREL are working in collaboration with Underwriters Laboratories, Inc. (UL) to develop analogous domestic certification standards and processes. Until those are in place, U.S. manufacturers are submitting their products and systems to certification against the international standards. As the wind energy industry continues to mature, it is reasonable to expect that future wind farm developers and their equipment vendors will conform to applicable quality standards and submit their products and systems to third-party certifications. Conformance to quality standards and certifications provides a better guarantee of safe design and construction and generally increases both the reliability and performance of major wind turbine components. Given the levels of participation that already exist, it is reasonable to conclude that proposals for future wind farms and the equipment represented in those proposals will involve a commitment to conform to all applicable quality standards and to submit to all relevant third-party certifications.

10. IMPACTING FACTORS RELATED TO REASONABLY FORESEEABLE SITE DEVELOPMENT ACTIVITIES The data in Tables 1 and 2 provide a reasonable representation of commercially available turbines and allow a reasonable prediction of the types of turbines that will be used in future sites. They are less adequate, however, in supporting further conclusions regarding site development. Nevertheless, past project experiences, together with the current state of wind energy technology and the advances expected from ongoing R&D activities, lend support to the following likely future site development scenarios.

360

United States Department of the Interior, Bureau of Land Management •

•

•

• •

• •

•

•

•

•

Business plans for future sites will involve developing candidate sites to their fullest wind energy potential as a means of quickly amortizing initial site development costs. The majority of large or extensive wind farms will probably be developed in phases, with the schedule of development being based largely on available development capital, as well as on myriad electric power market conditions. It is less likely that development will be speculative (i.e., built in advance of electric power sale agreements with transmission line operators) (Osborne 2004).[27] Sites developed in phases will not necessarily consist of the same turbine model throughout the site, and portions of the site may be owned and operated by more than one business entity.[28] Future sites are likely to take advantage of state-of-the-art wind turbine technology, leading to larger and taller but fewer turbines at a given site. It is possible that existing sites will expand into less-ideal areas that cannot, at this time, be economically farmed for wind energy by state-of-the-art turbine technologies. Sites may be repowered by replacing original turbines with technologically advanced models.[29] Modular construction of turbines will allow for their customization to address sitespecific characteristics. Modular construction, together with sophisticated SCADA systems, now make it technically feasible for future farms to consist of various models of turbines operating at different elevations on the basis of site-specific wind regime characteristics. Site development strategies will take fullest advantage of economies of scale. Activities will be grouped by type (e.g., foundations for all planned turbines will be installed over the same period), thereby simplifying logistics. Although the majority of wind turbine construction will still occur at the manufacturer’s facility, larger turbines, longer and more slender adaptive blades, and taller towers will impose unique problems related to the transportation of those components and may result in additional subassembly work being conducted on site during site construction. The use of innovative, self-erecting towers constructed of lightweight composite materials may dramatically minimize problems related to transportation logistics and site development times and costs. Reduced transportation requirements may expand the array of candidate sites to some that were previously excluded because of access difficulties. Equipment manufacturers can be expected to conform to international quality standards for manufacturing and operation (and to analogous U.S. standards as they are promulgated) as a way of maintaining market competitiveness. This conformance to standards will, in turn, lead to higher quality and greater reliability of major turbine components. Maintenance intervals are expected to increase as maintenance procedures become more regimented and are based on empirically derived isochronal factors rather than elapsed time.

Wind Energy Technology Overview •

•

•

361

Sophisticated SCADA systems will allow wind turbines at a given site to operate independently of one another, enabling the economical development of sites with different wind regimes throughout. It will become increasingly feasible for wind farms to include ancillary technologies, such as battery charging and elevated water storage, which will allow for the delayed delivery of wind-generated electricity to the transmission grid. The expanded capabilities and operating ranges of turbines will allow economical harvesting of wind energy at sites with Class 3 wind regimes.

REFERENCES AWEA (American Wind Energy Association), 2003a, Record Growth for Global Wind Power in 2002, news release, March 3. Available at http://www.awea.org/news/ news030303.gbl.html. Accessed March 2004. AWEA, 2003b, Projects Completed in 2003 (Preliminary). Available at http://www.awea.org/ projects/summaries/2003projects.pdf. Accessed March 2004. AWEA, 2004a, Wind Energy Projects throughout the United States of America, Jan. 22. Available at http://www.awea.org/projects/. Accessed March 2004. AWEA, 2004b, Online Membership Directory. Available at http://mcweb01 .memberclicks.com/ mc/prelogin.do?hidOrgID=awea. AWEA, 2004c, home page. Available at http://www. awea.org. AWEA, 2004d, Boom: 2003 Close to Best Year Ever for New Wind Installations; Bust: Expiration for Key Incentive Lowers Hopes for 2004, new release, Jan. 22. Available at http://www.awea.org/ news040122r03. Accessed July 9, 2004. Burton, T., et al., 2001, Wind Energy Handbook, John Wiley & Sons, Ltd., Chichester, United Kingdom. Cotrell, J., 2002, “A Preliminary Evaluation of a Multiple-Generator Drivetrain Configuration for Wind Turbines,” preprint, presented at the 21st American Society of Mechanical Engineers Wind Energy Symposium, Reno, Nev., Jan. 14–17. Available at http://www.nrel.gov/ docs/ fy02osti/31178.pdf. DOE (U.S. Department of Energy), 2002, Wind Power Today, 2002 Wind Energy Research Highlights. Available at http://www.nrel.gov/ docs/fy03osti/33149.pdf. Accessed March 2004. DOE/TVA/EPRI (DOE, Tennessee Valley Authority, and Electric Power Research Institute), 2003, Tennessee Valley Authority Buffalo Mountain Wind Power Project, First- and Second-Year Operating Experience: 2001–2003, DOE-EPRI Wind Turbine Verification Program, Dec. DWIA (Danish Wind Industry Association), 2004, Wind Turbine Generators. Available at http://www.windpower.org./en/tour/wtrb/ electric.htm. EERE (Office of Energy Efficiency and Renewable Energy), 2003, Wind and Hydropower Technologies Program; Wind Energy Program Multi Year Technical Plan for 2004– 2010, U.S. Department of Energy, Nov. Available at http://www.nrel.gov/wind_meetings/ 2003_imp_meeting /pdfs/mytp_nov_2003.pdf.

362

United States Department of the Interior, Bureau of Land Management

EERE, 2004a, Wind & Hydropower Technologies Program, U.S. Department of Energy. Available at http://www.eere.energy.gov/ windandhydro/wind_history.html?print. Accessed March 2004. EERE, 2004b, Wind Energy Resource Potential, U.S. Department of Energy. Available at http://www.eere.energy.gov/windandhydro/wind_potential. html. EERE, 2004c, How Wind Turbines Work, U.S. Department of Energy. Available at http://www.eere.energy.gov/windandhydro/wind_how.html. Accessed March 2004. EPRI (Electric Power Research Institute), 2001, Big Spring Wind Power Project, SecondYear Operating Experience: 2000–2001, Final Report, DOE-EPRI Wind Turbine Verification Program, Dec. GE (General Electric), 2004, 3.6 MW Wind Turbine Technical Specifications. Available at http://www.gepower.com/prod_serv/ products/wind_turbines/en/36mw/36mw_specs.htm. Accessed March 2004. Gipe, P., 1995, Wind Energy Comes of Age, John Wiley & Sons, Inc., New York, N.Y. Global Energy Concepts, LLC, 2001, WindPACT Turbine Design Scaling Studies Technical Area 3 — Self-Erecting Tower and Nacelle Feasibility March 2000–March 2001, NREL/SR-500-29493, prepared by Global Energy Concepts, LLC, Kirkland, Wash., for National Renewable Energy Laboratory, Golden, Colo., May. Griffin, D., 2002, Blade System Design Studies Volume I: Composite Technologies for Large Wind Turbine Blades, SAND2002-1879, prepared by Global Energy Concepts, LLC, Kirkland, Wash., for Sandia National Laboratories, July. Hau, E., 2000, Windturbines: Fundamentals, Technologies, Application, Economics, Springer-Verlag, Berlin, Germany. IsoTruss Structures, Inc., 2004, IsoTruss Structures, Inc., home page. Available at http://www.isotruss.com. Kaygusuz, K., 2004, “Wind Energy: Progress and Potential,” Energy Sources 26:95–105. Manwell, J.F., et al., 2002, Wind Energy Explained: Theory, Design, and Application, John Wiley & Sons Ltd., Chichester, United Kingdom. McGowan, J.G., and S. Connors, 2000, “Windpower: A Turn of the Century Review,” Annual Review of Energy and the Environment 25:147–197. Milborrow, D., 2002, “Wind Energy Technology — The State of the Art,” in Proceedings of the Institution of Mechanical Engineers, Part A, Journal of Power and Energy 216:23– 30. Momentum Technologies, LLC, 2004, The Source for Renewable Energy, source guide. Available at http://energy.sourceguides.com.index.shtml. NREL (National Renewable Energy Laboratory), 2004a, home page. Available at http://nrel.gov/. Accessed March 2004. NREL, 2004b, NREL Photo Archive, Photo #04688, #01671, and #12449. Available at http://www.nrel.gov/data/pix/. NREL, 2004c, National Wind Technology Center. Available at http://www.nrel.gov/wind/. NWCC (National Wind Coordinating Committee), 2004, home page. Available at http://www.nationalwind.org/. Osborn, D., 2004, personal communication from Osborn (Distributed Generation Systems, Inc., Lakewood, Colo.) to R. Kolpa (Argonne National Laboratory, Argonne, Ill.), March 15.

Wind Energy Technology Overview

363

Robichaud, R., 2004, personal communication from Robichaud (National Renewable Energy Laboratory, Golden, Colo.) to R. Kolpa (Argonne National Laboratory, Argonne, Ill.), March 11. Sandia National Laboratories, 2004a, home page. Available at http://www.sandia.gov/. Sandia National Laboratories, 2004b, What Is It? Available at http://www.sandia.gov/Renewable_Energy/wind_energy/abstracts/brochure.pdf. Sandia National Laboratories, 2004c, Online Abstracts and Reports. Available at http://www.sandia.gov/Renewable_Energy/wind_energy/ topical.htm. Sandia National Laboratories, 2004d, Wind Energy Technology. Available at http://www.sandia.gov/wind. Shikha [no first initial given], et al., 2003, “Aspects of Technological Development of Wind Turbines,” Journal of Energy Engineering 129(3):81–95, Dec. Steinhower, S., 2004, personal communication from Steinhower (SeaWest, Inc., Oakland, Calif.) to R. Kolpa (Argonne National Laboratory, Argonne, Ill.), March 19. TPI Composites, Inc., 2002, Parametric Study for Large Wind Turbine Blades, WindPACT Blade System Design Studies, SAND2002-2519, prepared by TPI Composites, Warren, R.I., for Sandia National Laboratories, Aug. TPI Composites, Inc., 2003, Blade Manufacturing Improvements: Remote Blade Manufacturing Demonstration, SAND2003-0719, prepared by TPI Composites, Warren, R.I., for Sandia National Laboratories, May. Available at http://www.sandia.gov/Renewable_Energy/wind_energy/ other/030719.pdf. Valmont, 2004, Valmont, home page. Available at http://www.valmont.com. Wilson, R.E., 1994, “Aerodynamic Behavior of Wind Turbines,” pp. 215–282, in Wind Turbine Technology, Fundamental Concepts of Wind Turbine Engineering, D. Spera (editor), American Society of Mechanical Engineers Press, New York, N.Y.

ENDNOTES [1]

[2]

[3]

[4] [5]

Wind farm developers and their investment capitalists must select among myriad options related to turbine design and site development and operation. Only those factors that have direct relationships to direct or cumulative impacting factors that are analyzed in this PEIS are discussed here. Although actual measurements of wind profiles at candidate sites are preferred, statistical methods can be utilized to extrapolate wind data from one site to nearby sites. An exhaustive discussion of these statistical methods is beyond the scope of this PEIS; additional information can be obtained from appropriate engineering texts (e.g., Burton et al. 2001; Manwell et al. 2002). For this discussion, a wind turbine is defined as any device operated expressly for generating electricity, regardless of whether that electricity is utilized locally or introduced into power transmission and/or distribution systems. Empirical studies have shown that the greatest turbine efficiencies are realized when the turbine rotor’s axis of rotation is tilted slightly from the horizontal. The Betz limit is named after Albert Betz, the German dynamicist who first identified and defined the phenomenon. A more detailed discussion of the influence of turbine

364

[6]

[7] [8] [9] [10] [11] [12]

[13]

[14]

[15] [16]

[17]

[18]

[19]

[20]

United States Department of the Interior, Bureau of Land Management blades on airflow and the derivation of the Betz limit is provided in Burton et al. (2001). In practice, such controls can be applied at any point throughout the operating range of the turbine to maintain the quality of electric power being produced and to overcome the real-world variability in incident wind energy over time. Hau (2000) cites studies from Denmark and Germany that support the claim that annualized availabilities of modern-day wind turbines can approach 98%. The center shaft, or rotor, of a typical induction generator rotates at 1,500 to 2,000 rotations per minute (rpm). The angle at which the airfoil of a rotor blade faces the wind, sometimes known as the angle of attack, can also influence the production of aerodynamic noise. One hertz, or one cycle per second, is equal to 1/60th rpm. However, innovative tower designs can dramatically influence erection costs and simplify transportation logistics. See Section D.7.1 for additional discussion. Detailed discussions on the development of interconnecting links to existing transmission lines are provided in the cumulative impacts portion of this PEIS. Nevertheless, the development of power links between any wind farm and existing power transmission lines will receive separate National Environmental Policy Act (NEPA) evaluations, which are outside the scope of this PEIS. Although many issues associated with power management and control and interconnection to the grid are outside the scope of this PEIS, they are, nevertheless, expected to be stipulations to any agreement between a power transmission company and a wind farm operator regulating grid interconnection. Where the soil mantle provides adequate grounding capacity, lightning protection systems routinely involve one or more grounding rods. For electrical substations, this grounding path is often enhanced by the installation of a grounding grid of wire located below the entire footprint of the substation and at some depth below the ground surface. Properly designed and installed “grounding wells” have no potential to adversely impact groundwater quality. Asynchronous generators are also commonly called induction generators. Expanded discussions on electric generators are available in appropriate engineering textbooks. A simplified discussion regarding generators used in wind turbines can be found in DWIA (2004). The difference in rotational speeds of the generator at idle and at peak load is called the generator slip, expressed as a percentage of the synchronous speed. Thus, the rotational speed of the generator’s center shaft (called the stator), which is turned by the action of the turbine rotor, varies little over the entire operating range of the generator. A nomograph is any chart representing numerical relationships. In this case, the relationship is between the degree of turbulence and the distance from a wind turbine to any natural or human-made wind obstruction, including other turbines. However, for wind turbines operating on very tall towers with their rotors largely within the geostropic wind regime, even mature trees represent relatively inconsequential ground-level obstructions to winds at the turbine hub’s elevation. The rotation of both a turbine rotor and the support tower induce turbulence in the downwind direction. Spacing of wind turbines to avoid turbulence effects is usually

Wind Energy Technology Overview

[21] [22] [23] [24] [25] [26] [27] [28]

[29]

365

represented by rotor diameters. Normally, a distance of 10 rotor diameters is considered to be the minimum downwind distance for spacing turbines in the downwind direction. The text box on the next page describes the AWEA and information compiled by the AWEA regarding the wind energy industry. For a comprehensive list of turbine manufacturers, consult AWEA (2004b) or commercial business source guides such as Momentum Technologies, LLC (2004). Many of the WindPACT technical reports may be accessed electronically at the NREL and Sandia Web sites; see NREL (2004a) and Sandia National Laboratories (2004d). Within the context of the WindPACT program, DOE defines lower wind classes as Class 4 and below (5.8 m/s [13 mpg] at a height of 10 m [33 ft]). See Sandia National Laboratories (2004c) for access to published reports of blade research being conducted by Sandia. Aerodynamic codes are an industry convention that describe the geometries of differently shaped airfoils. Nevertheless, speculative construction (sometimes referred to as a merchant plant) in advance of electric market agreements has occurred in the past. The Foote Creek Rim site, located near Arlington, Wyoming, is an example of one possible wind farm development scenario. This project, which was initiated on BLMadministered land and has subsequently been expanded to adjacent non-BLMadministered lands, represents one of the most ideal wind regimes in existence, with average wind speeds in excess of 23 mph (37 km/h). Four separate wind farms have been developed by two separate developers, delivering electric power to three separate utilities. The first farm, completed in April 1999, involved the erection of sixty-nine 600-kW turbines built by Mitsubishi (Model 600) and distributed over a land area of 2,156 acres (872 ha). The footprints of the turbines, control buildings, and other structures make up less than 1% of the land area in the parcel. A second farm completed in June 1999 added an additional three Mitsubishi turbines and 1.8 MW of generating capacity. A third farm, also completed in June 1999, added 33 NEG Micon turbines, representing a capacity of 24.8 MW. A final phase of development, completed in October 2000, involved an additional 16.8 MW of capacity from an additional 28 Mitsubishi Model 600 turbines. The remainder of the parcel continues to be used for ranching, as was the case before the wind farm was constructed. Repowering is already occurring. Many of the wind farms constructed in California in the early 1980s have been repowered.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 367-376

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 12

FEDERAL AND STATE REGULATORY REQUIREMENTS POTENTIALLY APPLICABLE TO WIND ENERGY PROJECTS* United States Department of the Interior, Bureau of Land Management SUMMARY The tables that follow list the major federal and state laws, Executive Orders, and other compliance instruments that establish permits, approvals, or consultations that may apply to the construction and operation of a wind energy project on Bureau of Land Management (BLM)-administered lands. The general application of these federal and state authorities and other regulatory considerations associated with such construction and operation are discussed in Chapter 3. The tables are divided into general environmental impact categories. The citations in the tables are those of the general statutory authority that governs the indicated category of activities to be undertaken under the proposed action and alternatives. Under such statutory authority, the lead federal or state agency may have promulgated implementing regulations that set forth the detailed procedures for permitting and compliance. Definitions of abbreviations used in the tables are provided here. ARS CRS CFR IC MCA NMSA *

Arizona Revised Statutes Colorado Revised Statues Code of Federal Regulations Idaho Code Montana Code Annotated New Mexico Statutes Annotated

A version of this chapter was also published in Wind Energy: Technology, Commercial Projects and Laws edited by Marco A. Telles published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

368

United States Department of the Interior, Bureau of Land Management NRS ORS RCW UCA USC WS

Nevada Revised Statutes Oregon Revised Statutes Revised Code of Washingtonaa Utah Code Annotated United States Code Wyoming Statutes

Table 1. Wind Energy Project Siting Authority Federal

Citation No primary statutory authority

Arizona

No primary statutory authority

California

• • • •

Public Resources Code, Division 13, § 21000 et seq. Public Resources Code, Division 15, Chapter 6, Power Facility and Site Certification, § 25500 et seq Public Resources Code, Division 15, Chapter 8, § 25743, Development of new in-state renewable electricity generation facilities Government Code, Division 1, Chapter 4, Article 2.11, Wind Energy, § 65892.13 et seq.

Colorado

Local government regulation — location, construction, or improvement of major electrical or natural gas facilities — legislative declaration (CRS 29-20-108)

Idaho

No primary statutory authority

Montana

• • •

Nevada

Construction of Utility Facilities; Utility Environmental Protection Act(NRS 704.820 et seq.)

New Mexico

No primary statutory authority

Oregon

Regulation of Energy Facilities, Energy Facility Siting Council (ORS 469.300-469.520)

Utah

Electrical Facility Review Board Act (UCA 54-14-10 et seq.)

Washington

• State Environmental Policy Act (RCW 43.21 C.010 et seq.) • Washington Energy Facility Site Evaluation Council (RCW 80.50.010 et seq.)

Wyoming

Industrial Development Information and Siting Act (WS 35-12-101 et seq.)

Montana Environmental Policy Act (MCA 75-1-101 et seq.) Major Facility Siting (MCA 75-20-101 et seq.) Wind Energy Easement (MCA 70-17-303)

Federal and State Regulatory Requirements ... Table 2. Land Use Authority Federal

•

Citation Federal Land Policy and Management Act of 1976 (43 USC 1701 et seq.) BLM Right-of-Way Regulation (43 CFR 2800) Coastal Zone Management Act (16 USC 1451 et seq.) Coastal Zone Act Reauthorization Amendments of 1990 (16 USC 1456 (c)(3)(A)) Wild and Scenic Rivers Act (16 USC 1271 et seq.) Farmland Protection and Policy Act (7 USC 4201 et seq.) Soil and Water Conservation Act of 1977 (16 USC 2001 et seq.) Structures Interfering with Air Commerce (49 USC 44718) Objects Affecting Navigable Airspace (14 CFR 77) Federal Aviation Administration, Advisory Circular 70/74602K, March 1, 2000 Oregon and California Grant Lands Act of 1937 (43 USC 1181 a, b, d-f) The Northwest Forest Plan

Arizona

•

No primary statutory authority

California

•

Public Resources Code, Division 5, Wild and Scenic Rivers Act, § 5093.50-5093.70 Public Resources Code, Division 25, Coastal Resources and Energy Assistance, § 35000 et seq

• • • • • • • • • • •

•

Colorado

Areas and Activities of State Interest (CRS 24-65.101 et seq.; CRS 24-65.1-101)

Idaho

Local Land Use Planning Act (IC 67-6501 et seq.)

Montana

• • • •

Nevada

Conservation; Regulations for Use of Land (NRS 548.410 et seq.)

New Mexico

No primary statutory authority

Oregon

Comprehensive Land Use Planning Coordination, Statewide Planning Goals and Guidelines (including Oregon Ocean-Coastal Management Program; Agricultural Lands; Open Spaces, Scenic and Historic Areas, and Natural Resources; and Air, Water and Land Resources) (ORS 197.005 et seq.)

Utah

No primary statutory authority

Washington

• •

Wyoming

Wyoming Environmental Quality Act (WS 35-11-101 et seq.)

Land Use Regulations (MCA 76-15-701 et seq.) Wild and Scenic Resources (MCA 76-12-101 et seq.) Timber Resources (MCA 76-13-101 et seq.) Rangeland Resources (MCA 76-14-101 et seq.)

Shoreline Management Act of 1971 (RCW 90.58.010 et seq.) Wetland Mitigation Banking (RCW 90.84.005 et seq.)

369

370

United States Department of the Interior, Bureau of Land Management Table 3. Floodplains and Wetlands Authority Federal

Citation Clean Water Act (33 USC 1344) Rivers and Harbors Act of 1899 (33 USC 401 et seq.) Executive Order 11988, “Floodplain Management,” May 21, 1977 Executive Order 11990, “Protection of Wetlands,” May 24, 1977

Arizona

Floodplain delineation; regulation of use (ARS 48-3609)

California

Public Resources Code, Chapter 7, Wetlands Preservation (Keene-Nejedly California Wetlands Preservation Act), § 5810 et seq. Water Code, Division 5, Cobey-Alquist Flood Plain Management Act, § 8400 et seq.

Colorado

Areas and Activities of State Interest (CRS 24-65.1-101 et seq.; CRS 24-65.1-202) Local Government Land Use Control Enabling Act (CRS 29-20-104)

Idaho

Local governments may adopt floodplain zoning ordinances (IC 46-1022)

Montana

Aquatic Ecosystem Protections (MCA 75-7-101 et seq.) Flood Plain and Floodway Management (MCA 76-5-101 et seq.)

Nevada

Planning and Zoning, Contents of Regional Plans (NRS 278.0274)

New Mexico Powers of Municipalities, Additional County and Municipal Powers; Flood and Mudslides Hazard Areas; Floodplain Permits; Land Use Control; Jurisdiction; Agreement(3-18-7(C) NMSA 1978) Oregon

Wetlands Conservation (ORS 196.600 et seq.) Removal of Material and Fill (ORS 196.795)

Utah

Quality Growth Act of 1999 (UCA 11-38-101 et seq.)

Washington Wetlands Mitigation Banking (RCW 90.84.005 et seq.) Floodplain Management (Chapter 86.16, RCW) Wyoming

Water Quality (WS 35-11-301 et seq.)

Table E-4. Water Bodies and Wastewater Authority Federal

Citation Clean Water Act (33 USC 1251 et seq.)

Arizona

Water Quality Control (ARS 49-201 et seq.)

California

Water Code, Division 7, Water Quality, § 13000 et seq.

Colorado

Water Quality Control (CRS 25-8-101 et seq.)

Idaho

Water Quality (IC 39-3601 et seq.)

Montana

Water Quality (MCA 75-5-101 et seq.)

Nevada

Water Controls; Water Pollution Controls (NRS 445A.300 et seq.)

New Mexico

Water Quality (74-6-1 NMSA 1978 et seq.)

Oregon Utah

• Water Quality, Water Pollution Control (ORS 468B.005 et seq.) • Sewage Treatment and Disposal Systems (ORS 454.010 et seq.) Water Quality Act (UCA 19-5-101 et seq.)

Washington

•

Wyoming

Domestic Wastewater Treatment Plants — Operators (RCW 70.95B010 et seq.) • On-site Sewage Disposal Systems (RCW 70.118.010 et seq.) • Water Pollution Control (RCW 90.48.010 et seq.) Water Quality (WS 35-11-301 et seq.)

Federal and State Regulatory Requirements ...

371

Table 5. Groundwater, Drinking Water, and Water Rights Authority Federal

Citation Safe Drinking Water Act (42 USC 300(f) et seq.)

Arizona

Water Quality Control (ARS 49-201 et seq., 49-241 et seq., 49-255 et seq., 49-351 et seq., 45151 et seq.)

California

• •

Health and Safety Code, Division 104, California Safe Drinking Water Act, § 116270 Water Code, Division 2, Water, § 1000 et seq.

Colorado

• • • • • •

Water Quality Control (CRS 25-8-101 et seq.) Water Rights and Irrigation (CRS 37-92-501 et seq.) Groundwater Recharge (IC 42-4201) Irrigation and Drainage ↓ Water Rights and Reclamation (IC 42-101, et seq.) Domestic Water and Ice (IC 37-2102) State Policy on Environmental Protection (IC 39-102)

• • • •

Water Use (MCA 85-2-101 et seq.) Public Water Supplies, Distribution, and Treatment (MCA 75-6-101 et seq.) Underground Water and Wells (NRS 534.010 et seq.) Water Controls; Public Water Systems (NRS 445A.800 et seq.)

Idaho

Montana Nevada New Mexico

Compliance with the Federal Safe Drinking Water Act (74-1-12 NMSA 1978)

Oregon

• • •

Oregon Drinking Water Quality Act of 1981 (ORS 448.115-448.990) Water Quality, Pollution Prevention Control (Groundwater) (ORS 468B.150 et. seq.) Water Resources Administration (ORS 536.220 et seq.)

Utah

• • • •

Safe Drinking Water Act (UCA 19-4-101 et seq.) Water & Irrigation (UCA 73-1-1 et seq.) Water Code (RCW 90.03.005 et seq.) Regulation of Public Ground Water (RCW 90.44.020 et seq.)

• •

Water Rights; Administration and Control (WS 41-3-101 et seq.) Water Quality (WS 35-11-301 et seq.)

Washington Wyoming

Table 6. Source Water Protection Authority Federal

Citation Safe Drinking Water Act, Protection of Underground Sources of Drinking Water (42 USC 300h-7)

Arizona

Aquifer Protection Permits (ARS 49-241 et seq.)

California

• • •

Colorado

Water Quality Control (CRS 25-8-101 et seq.)

Idaho

Groundwater Recharge (IC 42-4201)

Water Supply, Division 104, § 116975 et seq. Water Code, Division 7, § 13751 et seq. Areas and Activities of State Interest (CRS 24-65.1-204)

Montana

Montana Wellhead Protection Program (MCA 75-6-120)

Nevada

Underground Water and Wells (NRS 534.010 et seq.)

New Mexico

No primary statutory authority

Oregon

Water Pollution Control (Groundwater) (ORS 468B.167 et seq.)

Utah

Utah Water Quality Act (UCA 19-5-101 et seq.)

Washington

No primary statutory authority

Wyoming

Protection of Public Water Supply (WS 35-4-201 et seq.)

372

United States Department of the Interior, Bureau of Land Management Table 7. Cultural Resources

Authority Federal

Citation • Executive Order 13007, “Indian Sacred Sites,” May 24, 1996 • Executive Order 13175, “Consultation and Coordination with Indian Tribal Governments,” Nov. 9, 2000 • Native American Graves Protection and Repatriation Act (25 USC 3001) • American Indian Religious Freedom Act (42 USC 1996) • Archeological Resources Protection Act (16 USC 470(aa) et seq.) • Archaeological and Historic Preservation Act (16 USC 469a et seq.) • Antiquities Act (16 USC 431 et seq.) • National Historic Preservation Act (16 USC 470 et seq.) • Theft of Government Property (62 Stat. 764; 19 USC 1361) • Executive Order 11593, “Protection and Enhancement of the Cultural Environment,” May 15, 1971

Arizona

• Duties, board; partnership fund; state historic preservation officer (ARS 41-511.04) • Arizona Historical Society; powers; officers; duties of board of directors (ARS 41-821 et seq.) • Historic Preservation (ARS 41-861 et seq.) • Archeological Discoveries (ARS 41-841 et seq.)

California

Public Resources Code, Division 5, Historical Resources, § 5020 et seq.

Colorado

Montana

• Historical, Prehistorical, and Archeological Resources (CRS 24-80-401 et seq.) • Unmarked Human Graves (CRS 24-80-1302 et seq.) • Idaho Archaeological Survey (IC 33-3901 et seq.) • Protection of Graves (IC 27-501 et seq.) • Preservation of Historic Sites (IC 67-4601 et seq.) Antiquities (MCA 22-3-101)

Nevada

Historic Preservation and Archeology (NRS 383.011 et seq.)

New Mexico Oregon

Cultural Properties Act (18-6-3 NMSA 1978) • Historical Properties (ORS 358.475 et seq.) • Indian Graves and Protected Objects (ORS 97.740 et seq.)

Utah

• History Development (UCA 9-8-102 et seq.) • Native American Grave Protection and Repatriation Act (UCA 9-9-401 et seq.)

Washington

• Archaeological Sites and Resources (Chapter 27.53, RCW) • Indian Graves and Records (Chapter 27.44, RCW) • State Historical Societies — Historic Preservation (Chapter 27.34, RCW)

Wyoming

Antiquities Act (WS 36-1-114 through 36-1-116)

Idaho

Table 8. Wildlife Authority Federal

• • • • • • •

Citation Bald and Golden Eagle Protection Act (16 USC 668) Migratory Bird Treaty Act (16 USC § 703) Endangered Species Act (16 USC 1531 et seq.) Wild Free-Roaming Horse and Burro Act of 1971 (Public Law 92-195) Executive Order 13112, “Invasive Species,” February 3, 1999 Executive Order 13186, “Responsibilities of Federal Agencies to Protect Migratory Birds,” February 10, 2001

Arizona

• Powers and Duties (ARS 17-231 et seq.) • Taking and Handling of Wildlife (ARS 17-301 et seq.) • Wildlife Habitat Protection (ARS 17-451)

California

• Fish and Game Code, Division 3, Chapter 1.5, Endangered Species, § 2050 et seq. • Fish and Game Code, Division 5, Protected Reptiles and Amphibians, § 5000 et seq.

Federal and State Regulatory Requirements ...

373

Table 8. Wildlife (Continued) Authority Citation Colorado • Non-game, Endangered, or Threatened Species Conservation Act (CRS 33-2-101 et seq.) • Migratory Birds — Possession of Raptors — Reciprocal Agreements — Reports to General Assembly (CRS 33-1-115) • Colorado Natural Areas (CRS 33-33-101 et seq.) • Protection of Fishing Streams (CRS 33-5-101 et seq.) Idaho

• Species Conservation (IC 36-2401 et seq.)

Montana

• Wildlife Protection (MCA 87-5-101 et seq.)

Nevada

• Wildlife (NRS 501.002 et seq.)

New Mexico • Wildlife Conservation Act (17-2-46 NMSA 1978 • Endangered Plant Species (75-6-1 NMSA 1978) • Habitat Protection (17-6-1 NMSA 1978 et seq.) Oregon

• Threatened and Endangered Wildlife Species (ORS 496.171 et seq.) • Wildflowers; Threatened and Endangered Plants (ORS 564.010 et seq.) • General Protective Regulations, Commercial Fishing and Fisheries, Fish Passage; Fishways; Screening Devices; Hatcheries near Dams (ORS 509.580 et seq.) • Hunting, Angling, and Trapping Regulations; Wildlife Protective Provisions • (ORS 498.002 et seq.)

Utah

• Wildlife Resources Code of Utah (UCA 23-13-1 et seq.)

Washington

• Protection of Bald Eagles and Their Habitats (RCW 77.12.650-655)

Wyoming

• Bird and Animal Provisions (WS 23-3-101 et seq.) • Predatory Animals; Control Generally (WS 11-6-101 et seq.)

Table 9. Air Quality Authority Federal Arizona

Citation Clean Air Act (42 USC 7401 et seq.) Air Quality (ARS 49-401 et seq.)

California

Health and Safety Code, Division 26, Air Resources, § 39000 et seq.

Colorado

Air Quality Control (CRS 25-7-101 et seq.)

Idaho

• Registration of Persons Engaged in Operations or Construction Where Air Pollution Is a Factor (IC 39-110) • Pollution Source Permits (IC 39-115) • Relationship to Federal Law (IC 39-118B)

Montana

Air Quality (MCA 75-2-101 et seq.)

Nevada

Air Pollution (NRS 445B.100 et seq.)

New Mexico Oregon

• Environmental Improvement Act (74-1-1 NMSA 1978 et seq.) • Air Quality Control Act (74-1-1 NMSA 1978 et seq.) Air Quality (ORS 468A.005 et seq.)

Utah

Air Conservation Act (UCA 19-2-101 et seq.)

Washington

Washington Clean Air Act (Chapter 70.94, RCW

Wyoming

Air Quality (WS 35-11-201 et seq.)

374

United States Department of the Interior, Bureau of Land Management Table 10. Noise Authority Federal Arizona California Colorado Idaho Montana Nevada New Mexico Oregon Utah Washington Wyoming

Citation Noise Control Act, as amended by the Quiet Communities Act (42 USC 4901 et seq.) No primary statutory authority Health and Safety Code, Division 28, Noise Control Act, § 46000 et seq. Noise Abatement (CRS 25-12-101 et seq.) No primary statutory authority No primary statutory authority Prevention of Excessive Noise (NRS 244.363) No primary statutory authority Noise Control (ORS 467.010 et seq.) No primary statutory authority Noise Control (RCW 70.07.010 et seq.) No primary statutory authority

Table 11. Hazardous Materials Authority Federal

Citation • Hazardous Materials Transportation Law (49 USC 5101-5127) • Emergency Planning and Community Right-to-Know Act of 1986, as extended to federal facilities by Executive Order 12856, August 3, 1993 • Oil Pollution Control Act (33 USC 2701 et seq.) • Pollution Prevention Act of 1990 (42 USC 13101 et seq.)

Arizona

• Emergency Planning and Community Right-to-Know Act (ARS 26-341 et seq.)

California

• Health and Safety Code, Division 20, Chapter 6.11, Unified Hazardous Waste and Hazardous Materials Management and Regulatory Program, § 25404 et seq. • Health and Safety Code, Division 20, Chapter 6.6, Safe Drinking Water and Toxics Enforcement Act of 1986 (Proposition 65), § 25249.5 et seq. • Health and Safety Code, Division 20, Chapter 6.95, Hazardous Materials Release Response Plans and Inventory, § 25500 et seq.

Colorado

• Implementation of Title III of Superfund Act (CRS 24-32-2601 et seq.) • Hazardous Substances (CRS 25-5-501 et seq.)

Idaho

Hazardous Substances Emergency Response Act (IC 39-7101 et seq.)

Montana

Montana Response to Hazardous Material Incidents Act (MCA 10-3-1201 et seq.)

Nevada

• Hazardous Materials; Regulation of Highly Hazardous Substances and Explosives • (NRS 459.380 et seq.)

New Mexico

• Hazardous Chemicals Information Act (74-4E-1 NMSA 1978 et seq.) • Hazardous Material Transportation (74-4F-1 NMSA 1978 et seq.)

Oregon

Hazardous Substances (ORS 453.001-453.527)

Utah

No primary statutory authority

Washington

• Oil and Hazardous Substance Spill Prevention and Response (RCW 90.56.005 et seq.)

Wyoming

• Water Pollution from Underground Storage Tanks Corrective Action Act of 1990 • (WS 35-11-1414 et seq.)

Federal and State Regulatory Requirements ... Table 12. Pesticides and Noxious Weeds Authority Federal

Citation • Federal Insecticide, Fungicide, and Rodenticide Act (7 USC 136 et seq.) • Noxious Weed Act of 1974 (7 USC 2801-2813), as amended by Section 15, Management of Undesirable Plants on Federal Lands 1990

Arizona

• Pesticide Contamination Prevention (ARS 49-301) • Pesticides (ARS 3-341 et seq.) • Pesticide Control (ARS 3-361 et seq.)

California

• Food and Agriculture Code, Division 7, Agricultural Chemicals, Livestock Remedies, and Commercial Feeds, § 12500 et seq. • Food and Agriculture Code, Division 4, Weeds, § 7201 et seq.

Colorado

Pesticide Act (CRS 35-9-101 et seq.)

Idaho

• Application of Fertilizers and Pesticides (IC 39-127) • Pesticides and Chemigation (IC 22-3401 et seq.) • Noxious Weeds (IC 22-2401 et seq.)

Montana

• Pesticides (MCA 80-8-101 et seq.) • Weed Control (MCA 80-7-701 et seq.)

Nevada

Control of Insects, Pests, and Noxious Weeds (NRS 555.005 et seq.)

New Mexico

Pesticide Control Act (76-4-1 NMSA 1978 et seq.)

Oregon

Pesticide Control (ORS 634.005 et seq.)

Utah

Utah Pesticide Control Act (UCA 4-14-1 et seq.)

Washington

Washington Pesticide Application Act (RCW 17.21.010 et seq.)

Wyoming

Wyoming Weed and Pest Control Act of 1973 (WS 11-5-1001 et seq.)

Table 13. Solid Waste Authority Federal

Citation Solid Waste Disposal Act (SWDA) (42 USC 6901 et seq.)

Arizona

Solid Waste Management (ARS 49-701 et seq.)

California

Public Resources Code, Division 30, Waste Management, § 40000 et seq.

Colorado

Solid Waste Disposal Sites and Facilities (CRS 30-20-100.5 et seq.)

Idaho

Idaho Solid Waste Facilities Act (IC 39-7401 et seq.)

Montana

Montana Solid Waste Management Act (MCA 75-10-201 et seq.)

Nevada

Sanitation; Collection and Disposal of Solid Waste (NRS 444.440 et seq.)

New Mexico

Solid Waste Act (74-9-1 NMSA 1978 et seq.)

Oregon

Solid Waste Management (ORS 459.005 et seq.)

Utah

Solid Waste Management Act (UCA 19-6-501 et seq.)

Washington

Solid Waste Management — Reduction and Recycling (RCW 70.95.010 et seq.)

Wyoming

Solid Waste Management (WS 35-11-501 et seq.)

375

376

United States Department of the Interior, Bureau of Land Management Table 14. Hazardous Waste and Polychlorinated Biphenyls (PCBs) Authority Federal

Arizona

Citation • Toxic Substances Control Act (15 USC 2605(e)) • Solid Waste Disposal Act, as amended by the Resource Conservation and Recovery Act (42 USC 6901 et seq.) and the Hazardous Solid Waste Amendments of 1984 Hazardous Waste Disposal (ARS 49-901 et seq.)

California

Health and Safety Code, Division 20, Hazardous Waste Control, § 25100 et seq.

Colorado

Hazardous Waste (CRS 25-15-101 et seq.)

Idaho

• Hazardous Waste Management (IC 39-4401 et seq.) • PCB Waste Disposal (IC 39-6201 et seq.)

Montana

Montana Hazardous Waste Act (MCA 75-10-401 et seq.)

Nevada

Hazardous Material; Disposal of Hazardous Waste (NRS 459.400 et seq.)

New Mexico

Hazardous Waste Act (74-4-1 through 74-4-14 NMSA 1978)

Oregon

• Hazardous Waste and Hazardous Materials I (ORS 465.005 et seq.) • Hazardous Waste and Hazardous Materials II (ORS 466.005 et seq.)

Utah

Solid and Hazardous Waste Act (UCA 19-6-101 et seq.)

Washington

Hazardous Waste Management (RCW 70.105.005 et seq.)

Wyoming

Solid Waste Management (WS 35-11-503 and 35-11-516)

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 377-384

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 13

COMMERCIAL WIND ENERGY PROJECTS* United States Department of the Interior, Bureau of Land Management SUMMARY Data on commercial wind energy projects in the western states that are within the scope of this programmatic environmental impact statement (PEIS) are displayed in the tables below. The American Wind Energy Association (AWEA) compiles and maintains all of the data displayed below. All data presented are current as of January 14, 2004. All data are accessible electronically from the AWEA Web site at http://www.awea.org/projects/ index.html. Data presented in the tables below are updated quarterly by the AWEA. The Bureau of Land Management (BLM) cannot guarantee the completeness or accuracy of these listings. Submission by wind farm developers or operators of project information to AWEA for inclusion in these listings is voluntary.

*

A version of this chapter was also published in Wind Energy: Technology, Commercial Projects and Laws edited by Marco A. Telles published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

378

United States Department of the Interior, Bureau of Land Management

Commercial Wind Energy Projects Wind Energy Projects in California (Continued)

379

380

United States Department of the Interior, Bureau of Land Management Wind Energy Projects in California (Continued)

Commercial Wind Energy Projects

381

382

United States Department of the Interior, Bureau of Land Management

Commercial Wind Energy Projects

383

384

United States Department of the Interior, Bureau of Land Management

Washington State Wind Energy Development

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 385-417

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 14

BIOMASS AND BIOENERGY RESEARCH IN TROPICAL AFRICA: STATE OF THE ART, CHALLENGES AND FUTURE DIRECTIONS *

Jonathan C. Onyekwelu1,2 and Shadrach O. Akindele2,3 Technische Universität, München, Am Hochanger 13, D-85354, Freising, Germany Department of Forest Resources Management, University of British Columbia, 2424 Main Mall, Vancouver, B.C., Canada V6T 1Z4

ABSTRACT Forest biomass and bioenergy production currently play a very important role in energy generation in tropical African countries, especially in the rural areas where between 75 and 95% of the populace depend on fuelwood as the primary energy source. Given the current high population growth, the low rates of switching to noncarboniferous household energy sources as well as the inefficiency of other energy sources, the importance of biomass and bioenergy in household energy generation in tropical African countries is expected to increase in the future. This paper examines the sources and extent of biomass production in tropical African countries as well as their current contribution to bioenergy supply and possible future trend. The current status and *

A version of this chapter was also published in Biomass and Bioenergy: New Research published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. 1 Corresponding author: Current address: Chair of Silviculture and Forest Management. Technische Universität, München, Am Hochanger 13, D-85354, Freising Germany. Tel: ++49-8161-714687; Fax: ++49-8161-714616. Email: [email protected] 2 Permanent address: Department of Forestry and Wood Technology, Federal University of Technology, P.M.B. 704, Akure Ondo State, Nigeria. Tel:++234-8034721633 3 Current address: Department of Forest Resources Management, University of British Columbia, 2424 Main Mall, Vancouver, BC., Canada. V6T 1Z4. Tel: +1 604-822-5689; Fax: +1 604-822-9106. Email: [email protected]

386

Jonathan C. Onyekwelu and Shadrach O. Akindele prospects of bioenergy technologies, the state of biomass and bioenergy research in the sub-region as well as the methodologies used in obtaining local and national biomass estimates were reviewed. The paper also discussed the challenges facing biomass and bioenergy research in tropical Africa, and stressed the need for more collaboration with the developed countries to be able to tackle the challenges. The paper finally examines the likely future research directions and makes recommendations towards a more efficient and environmental-friendly utilization of biomass and bioenergy in the subregion.

Keywords: Biomass and bioenergy production, tropical Africa, poverty, energy, deforestation, natural forests, forest plantations.

INTRODUCTION Geographical Extent of Tropical Africa The geographical tropic4 is the geographic region of the earth centred on the equator and lying between the two tropics: the tropics of Cancer and Capricorn in the northern and southern hemispheres, respectively. The tropic of Cancer lies at approximately 23.5°N latitude while the tropic of Capricorn lies at approximately 23.5°S latitude. In the tropics, the sun reaches a point directly overhead at least once during the solar year (north of the tropic of Cancer and south of the tropic of Capricorn the sun never reaches an azimuth of 90° or directly overhead). Consequently, tropical Africa includes the whole geographical zone of the African continent lying between the tropic of Cancer and the tropic of Capricorn (Fig. 1). It encompasses a large and rather heterogeneous assemblage of countries and covers about 22.7 million km2 (about 73.9%) of Africa’s total land mass of about 30.7 million km2 (Parry, 1955; Brown and Gaston, 1996). Morocco, Tunisia, Lesotho and Swaziland are the only countries completely outside the tropics in Africa, while Western Sahara, Algeria, Libya, Egypt and South Africa have a significant portion of their land mass outside the tropics and are thus not considered as part of tropical Africa for the purpose of this study. All other countries in the continent (46) lie either predominantly or totally within the tropics (Fig. 1). Basically, the consistency and uniformity of climate is used to distinguish the tropics from other parts of the globe (Evans and Turnbull, 2004). Unlike in temperate lands, where seasons are differentiated based on temperature variation, seasons in the tropics are delineated based mainly on rainfall variation (amount and distribution). Two broad seasons (rainy and dry seasons), defined basically on the presence and absence of rainfall, are recognised in tropical Africa. While the rainy (wet) season is marked by consistent rainfall, the dry season (also called harmattan in some parts of Africa, e.g. Nigeria) is characterised by the absence of rainfall. However, there could be little rainfall during the dry months, especially in the very humid parts of tropical Africa where the rainy season could last for 10 months in a year and where annual rainfall could be as high as 3000 – 4000 mm. The duration and distribution of

Biomass And Bioenergy Research In Tropical Africa…

387

rainfall significantly impacts agriculture and forestry in tropical Africa, with two crop growing seasons possible in wetter parts of the continent. On the other hand, only one cropping season is practiced in the driers parts of tropical Africa with about 4 months of rainfall and mean annual rainfall of less than 1000 mm. Forest tree growth in the humid tropics is continuous because of the long duration of rainy season and because precipitation exceeds or equals potential evapotranspiration and thus the soil is continuously moist and there is hardly any season in which the soil dries out (Nwaoboshi, 1982; Evans and Turnbull, 2004). Though not as important as rainfall, temperature is also considered in defining the tropics. Thus, in addition to the amount and distribution of rainfall, it has become more meaningful and acceptable to define the tropics as where mean monthly temperature variation between the three coldest and the three warmest months is less than 50C (Evans and Turnbull, 2004). Based primarily on rainfall and temperature, Parry (1955) categorised the climate of tropical Africa into seven groups, which are: (i) wet tropical lowland, (ii) moist tropical lowland, (iii) wet montane, (iv) dry montane, (v) dry plateau, (vi) dry lowland and (vii) semidesert. This classification is broader than that of Evans and Turnbull (2004).

Forest Types of Tropical Africa and Their Characteristics The total forest cover in Africa is about 650 million ha and accounts for 21.8% of the continent’s land area and 16.8% of global forest cover (FAO, 2001a). About 81% (528 million ha) of Africa’s forest cover is found within the tropical region (Singh, 1993). The distribution of forests in Africa is uneven, with North and West Africa being the least forested sub-regions due largely to their extremely arid conditions. Central Africa, with about 44% of its land under forest cover, accounts for 37% of Africa’s forest ecosystems (FAO, 2003a). The forest resources in Africa are currently threatened by high rates of deforestation, which resulted in a loss of about 53 million ha of forest cover between 1990 and 2000 and accounted for 56% of global forest cover decrease, with tropical countries like Sudan, Zambia and the Democratic Republic of the Congo (DR Congo) accounting for almost 44% of the deforestation in Africa within this period (FAO, 2001b; 2003a). The forest resources in tropical African countries are presented in Table 1. About 74% (2.69 million km2) of the forests are located in Central Africa, 19% (680,000 km2) in West Africa and 7% (250,000 km2) in East Africa. The five countries with the highest tropical forest resources are the DR Congo, Angola, Sudan, Tanzania and Zambia, which accounts for about 53% of the tropical forest resources in Africa. Africa’s tropical forests are classified either based on their structure or their ecological differences. Based on structure, they are broadly classified either as closed forests or open forests. While the closed forests are defined as land covered by trees with a canopy cover greater than 40% and a height exceeding 5 m, the open forests (also known as fragmented forests and includes woodland and wooded savannah) are defined as land covered by trees with canopy cover of between 10 and 40 percent and a height exceeding 5 m or mosaics of forest land (FAO, 2001a). Both forest types include natural and plantation forests. Based on this classification, much of the forest resources in tropical Africa are open forests as indicated by the analyses conducted by Brown 4

The term “geographical tropic” is used to emphasis the distinction between geographical and ecological tropics, which is usually made when defining “tropics”. The ecological tropics refer to climatic conditions, with the main defining characteristic being the absence of any month with an average temperature below 00 C.

388

Jonathan C. Onyekwelu and Shadrach O. Akindele

and Gaston (1996), which revealed that 18% of tropical Africa are classified as closed forests while 36% are classified as open forests (Fig. 2). The closed forests are predominantly found in the Congo basin5 (Fig. 2). In West Africa, the closed forests exist as a narrow strip along the western coasts of Nigeria, Ghana, Liberia and Sierra Leone (Fig. 2). Based on ecological differences, tropical forests are broadly classified as either moist or dry forests (Fig. 3a). The tropical moist forests are mainly found in the Congo basin6 of Central Africa and a narrow band along the coast in West Africa (FAO, 2001a; FAO 2003a; Were, 2001). The moist forests are found where the climate is hot and humid throughout the year, with about 2000 mm of annual precipitation (Lauer, 1989). More important however, is that rainfall is distributed throughout the year, thus moist forest occur in potions of Africa where annual precipitation is sometimes about 1600 mm but falls throughout the year (Hendrick, 2001). The moist forests constitute Africa’s most diverse terrestrial ecosystem, with many rare, endemic and endangered plant and animal species, rich vegetation and tall and closely set trees that often form a continuous multi-layered canopy and emergent trees reaching a height of 50 to 60 m (Jones, 1955; Nwoboshi, 1982; Sayer et al., 1992; Gonder et al., 1997; Hilton-Taylor, 2000; Were, 2001; Boukongou, 2005). The moist forests of the Congo basin and the humid zone in West Africa are inherently more productive (FAO, 2003a). However, the moist forests of West Africa are less diverse, with relatively lower endemism than those of Central Africa (IUCN, 1996 cited in CIFOR, 2005). The Congo basin contains the largest remaining contiguous expanse of moist tropical forest in Africa and is second largest in the world after the Amazon (FAO, 2003b; CIFOR, 2005). The moist forest contains many resources for local subsistence and of commercial importance, such as timber, fuelwood, rattan, fruits, nuts, medicinal plants and rubber and is also home to a large number of indigenous peoples. Tropical dry forests occur in regions with pronounced dry periods of between 4 to 8 months and annual precipitation of about 1000 mm or less (Hendrick, 2001). They are most extensive in eastern and southern Africa, where they stretch over large areas (FAO, 2001a). Unlike the moist forests, the dry forests have vegetations that are typically made up of deciduous trees of 10 to 20 m tall, with a grass understorey. Though this forest type consists mainly of steppe vegetation, thorny bushes and open savannah woodlands with grass and shrubs dominating, they are productive and contribute significantly to the livelihoods of rural people (CIFOR, 2005). The dry forest also includes subtropical dry forest or dry sclerophyll forest (Fig. 3a), with natural vegetation similar to the Mediterranean climate type (mild humid winters and dry summers). Though a large proportion of the dry forest types has been cleared and is currently dominated by shrubs, it still harbours a particularly rich flora, including many endemic species. The African tropical dry forests and savannah are major habitats for wildlife and also provide local people with valuable products and services such as fuelwood, honey, timber, bushmeat, medicines and grazing ground for cattle (FAO, 2001a).

5 6

The Congo basin covers such countries as Cameroon, Congo, DR Congo, Central African Republic, Equatorial Guinea, Gabon, Burundi, Rwanda, Angola and Chad (Boukongou, 2005). The Congo basin contains a vast moist forest area covering approximately 2.3 million km2 or 26 percent of the world’s rainforests (Boukongou, 2005). However, the moist forests cuts across only six of the ten countries within the basin, which includes: Cameroon, Central African Republic, Congo, DR Congo, Equatorial Guinea and Gabon (WRI, 2000)

Biomass And Bioenergy Research In Tropical Africa…

389

DEFINITION AND MEANING OF BIOMASS AND BIOENERGY Biomass is a very broad term covering a wide variety of plant and animal materials that can be seen as energy resources, thus its definition usually depends on the context in which it is used. For the purpose of this study, biomass is defined as the total amount of organic matter present in plant (woody and non-woody) and animal residues (e.g. dung) expressed in dryweight (oven-dry) basis. When expressed as mass per unit area (e.g. t ha-1), it is referred to as biomass density and is a useful way of quantifying the amount of resource available for all traditional uses. Most biomass estimations and definitions (especially in forestry) involve only the aboveground components of the tree or plant, mainly because these components generally account for the greatest fraction of total living biomass and does not pose too much logistical problems in estimation (Brown, 1997). For most forests, biomass is only estimated for trees with diameters equal or greater than 10 cm, though minimum diameter could be smaller than 5 cm for forests of smaller stature such as those in the dry tropical zones or highly degraded forests (FAO, 1993). Forest undergrowth and forest floor fine litter are not usually included in biomass estimation, except where they are locally important. This is because they are rarely used as biofuel and when they are so used, it is usually in isolated cases and out of necessity and lack of sufficient alternatives. Ryan and Openshaw (1991) classified biomass broadly into (i) woody biomass, (ii) nonwoody biomass and (iii) animal residue (dung). Woody biomass comes from ligneous plants such as trees or bamboos while non-woody biomass includes leaves, herbaceous plants and crop residues and animal residues may come from the excrement of animals, though cattle dung is the most common form (Ryan and Openshaw, 1991). Because they come from organic matter, biomass resources are renewable. For example, many biomass resources are replenished through the cultivation of fast-growing trees and grasses. Unlike fossil fuels which take millions of years to create, biomass can be replaced relatively quickly without permanently depleting the earth's natural resources. Biomass is processed to create bioenergy in form of electricity, heat, steam, and fuels (e.g. bio-diesel, ethanol, etc). Consequently, bioenergy is defined as the energy stored in biomass. Organic matter may either be used directly as a fuel processed into liquids and gases or be a residual of processing and conversion. It is very important to note that the burning of biomass (bioenergy), like fossil fuels, can produce carbon dioxide (CO2). However, because its source is renewable, bioenergy creates what has been termed “net gain of zero” by experts in bioenergy research. This is because the small amount of emissions (mainly CO2) put into the atmosphere by burning biomass is offset by the amount of CO2 that was absorbed by the biomass while it was growing. So, while the overall amount of emissions in the atmosphere is not reduced, they are not increased either and the net emission of CO2 from bioenergy will be zero as long as plants continue to be replenished. Bioenergy is experiencing a surge in interest in many parts of the world due to a greater recognition of its current role and future potential contribution as modern fuel, its availability, versatility and sustainability, a better understanding of its global and local environmental benefits, perceived potential role in climate stabilization, the existing and potential development and entrepreneurial opportunities. Consequently, a World Bank report concluded that energy policies will need to be as concerned about the supply and use of biofuels as they are about modern fuels (World Bank, 1996).

390

Jonathan C. Onyekwelu and Shadrach O. Akindele

SOURCES OF BIOMASS AND THEIR CONTRIBUTIONS TO BIOENERGY SUPPLY Though the consumption pattern of biomass is noted to be declining over much of Asia and almost static in Latin America, it has continued to rise in Africa. An estimated 583 million people in Africa relied on fuelwood, charcoal and other biomass for cooking and heating in year 2000, which is projected to increase to 820 million people within the next three decades (Arnold et al., 2003). The volume of woodfuel consumption in Africa increased from about 399.5 million m3 in 1980 to 635.1 million m3 in 2000 (an increase of about 59%) and is expected to increase to about 850.2 million m3 in 2020 (FAO, 2003c). Bioenergy is the dominant energy source in tropical African countries; especially in rural households where between 75 and 95% of energy needs are met with biomass as the primary energy source (Fuwape and Onyekwelu, 1995; Marufu et al., 1997; IEA, 2003; World Bank, 2003; Karekezi et al., 2004). Also a significant number of small-scale rural and urban industries such as agricultural processing industries (tea, tobacco, rice parboiling, cassava production) brick and tile industries, alcoholic beverage production, wood processing, bakeries, restaurants/canteens, etc, rely on biomass as their major source of energy. For example, biomass (firewood, charcoal and agricultural waste) and petroleum products accounted for 78% and 19%, respectively of the total energy consumption in Kenya while 91.6%, 6.8% and 1.6% of the total energy consumed in Tanzania were sourced from biomass, petroleum products and electricity, respectively (IEA, 2003; Karekezi et al., 2004). This heavy reliance on bioenergy in tropical Africa is dictated by the unavailability of conventional energy sources (e.g. electricity, gas, kerosene, coal, etc) in most rural areas, the characteristic inefficiency (epileptic services) of the conventional energy sources (i.e. where they are available) as well as the inability of the rural dwellers to pay for the conventional energy sources. The most critical issue relating to biomass and bioenergy consumption in African has to do with the current state of the sources of supply and whether these sources are sustainable. In tropical Africa, woody biomass is obtained from various sources depending on the dominant vegetation type as well as the distance, accessibility and availability of woody biomass stocks in the ecosystem. Woody biomass is obtained in large quantities from closed forests, open forests, woodlands and wooded grasslands (Ryan and Openshaw, 1991). In addition, a significant quantity of woody biomass are sourced from trees outside the forest (i.e. non-forest sources) such as community woodlots, trees in home gardens, trees scattered in farmlands (agroforestry trees), trees and shrubs used for life fence, trees used for hedges or wind breaks as well as trees along the roads and waterways (Ryan and Openshaw, 1991; FAO, 1993). Furthermore, palms and bamboos are increasingly being included as sources of biomass, though only where they comprise an important component of the forest and are important for local uses (FAO, 1993). Residues from logging, wood industries, wood recovered from construction waste are also becoming important additional sources of woody biomass in tropical Africa. For example, the petroleum products crises in Nigeria in the 1990s (especially between 1992 and 1995) and the consequent scarcity and soar in prices of conventional energy resulted in a situation where lots of people in the rural and urban centres were collecting sawmill wastes for use in cooking (Fuwape and Onyekwelu, 1995). This practice has continued till date.

Biomass And Bioenergy Research In Tropical Africa…

391

Non-woody biomass is usually sourced from crop harvest residues such as maize, sorghum and millet stalks, coconut husk and kernels, maize-hub, etc, while the common animal residue used for bioenergy is animal dung (especially cow dung). In rural Malawi, women use twigs to fry foods and Ethiopian women use maize stalks to make griddled flatbread while in South Africa animal dung is considered suitable for beer brewing, which requires a long-lasting low heat (van Horen and Eberhard 1995; Brouwer et al., 1996). Though a potential huge source of bioenergy, non-woody and animal residue biomasses are currently used as alternative bioenergy sources (Marufu et al., 1997) in rural areas when woody biomass is scarce or simply too extensive. Compared to other forms of biomass, woody biomass is by far the dominant contributor to bioenergy consumption in tropical Africa. Consequently further discussion on this subject will concentrate on woody biomass. A high percentage (about 60%) of woody biomass consumed in tropical Africa is sourced from closed forest or moist forests, where vegetation is richer, trees are more abundant and the forest is more productive. Currently, the humid zones do not have any major problem with regard to bioenergy supply (mainly fuelwood) and will most likely not have any shortage in the near future due mainly to the high biomass productivity of humid forests as well as preponderance of tree crops on homesteads and in other areas (Fuwape and Onyekwelu, 1995; FAO, 2003c). However, this will not be the case in the dry tropical forest zones where biomass productivity is low, because a large portion of the forests has been cleared, the tree poverty of the remaining forests as well as their low productivity. Consequently, the inhabitants of the dry zone are dependent on biomass transported from distant sources (mostly from the humid zones), which is increasingly making traditional bioenergy (mainly fuelwood) too expensive for the rural poor. The current high rate of deforestation in natural forests (both humid and dry forests) in tropical Africa, which was estimated to amount to a total forest loss of over 1 million ha between 1990 and 2000 in the West Africa sub-region alone (FAO, 2003c), implies that future biomass supply from Africa’s tropical natural forests is under serious threat and as a result, the sustainability of supply cannot be guaranteed. With the exception of Gambia and Cape Verde, other tropical African countries witnessed a decrease of between 0.2 and 9.0% in their total natural forest estates in just one year (between 1999 and 2000 (Table 1)). Contrary to this, forest plantation areas in tropical African countries is increasing, which is an indication that they will play a major and significant role in the future supply of biomass and thus bioenergy consumption in the continent. Africa’s tropical forest plantation estates increased from about 2.7 million ha in 1980 to 3.8 million ha in 1990 and 4.6 million ha in 2000 (Evans and Turnbull, 2004), which represented an increase of 68% within two decades. The main purpose of such plantations is either for industrial wood production or domestic use as building poles, fuelwood, wood for charcoal and fodder (Pandey, 1995; Evans, 1998). Many fuelwood plantations with Gmelina arborea, Acacia spp, Eucalyptus spp are reported to exist in Northern Nigeria (Onyekwelu, 2001). There are indications that current contribution of forest plantations to bioenergy supply in Africa is low, probably due to the fact the plantations established in the last three decades are just maturing. However, when compared to other geographical regions of the world, Africa has the highest proportion of forest plantations used for bioenergy production, with countries like Sudan, Ethiopia, Madagascar and Rwanda having between 52 to 88% of their total forest plantation areas (predominantly Eucalyptus spp and Acacia spp) as bioenergy plantations (FAO, 2000; FAO, 2001b). The proportion of biomass sourced from forest plantation in tropical Africa is

392

Jonathan C. Onyekwelu and Shadrach O. Akindele

expected to increase in the future. The estimations of Marrison and Larson (1996) revealed that about 39.4% (4.9 EJ yr-1) of the total biomass supply for energy in Africa in 2025 will be sourced from plantation forests which is expected to increase to about 60.4% (14.2 EJ yr-1) in 2050. For this target to be realised, however, it is expected that in addition to state funded large-scale forest plantation programmes, bioenergy will be sourced from smaller-scale biomass production systems like community woodlots, farm forestry systems, home gardens, etc.

BIOMASS PRODUCTION OF THE MAJOR FOREST TYPES Biomass Production from Natural Forests Extreme variability in growing conditions exist between the moist and dry natural tropical forests, thus differences in biomass productivity exists across the African continent. For example, while the mean woody biomass production of the forests in the Sahelian zone is as low as 4 t ha-1, it is more than 200 t ha-1 in the humid tropical rain forests of Central Africa (FAO, 2003c). This trend in biomass production between the dry and moist tropical forests was also confirmed by the investigation of Brown and Gaston (1995), who reported actual biomass production of 20 Mg ha-1 and 299 Mg ha-1 for dry lowland and moist lowland forests of tropical Africa, respectively. Brown and Gaston (1995) also reported biomass production of 37 and 105 Mg ha-1 for African montane seasonal and montane moist forests, respectively. Future biomass production in the various forest types of tropical Africa is expected to follow this trend as indicated by the maximum biomass density potential in tropical Africa (see Brown and Gaston, 1996). It is therefore apparent that the moist forests, especially the humid tropical forests at the Congo basin and West Africa, are inherently more productive than the other forest types in Africa. Due to the high demand for bioenergy in the dry tropical regions, extraction from the dry forests often far exceed natural productivity. This implies heavy pressure on the forests and potential acute scarcity of woody biomass in the dry regions given the low biomass productivity of the dry forests. High variability in biomass production also exits between the African countries, with countries within the humid forest zones having significantly higher biomass production than those within the dry forest zones. Actual biomass production is reported to be as low as 10 to 15 Mg ha-1 in Botswana, Niger, Somalia, and Zimbabwe and as high as 305 to 344 Mg ha-1 in Congo, Equatorial Guinea, Gabon, and Liberia (Brown and Gaston, 1995). Similar biomass productivity variation between the moist and dry forests also exists between the closed and open forests, with the closed forests having significantly higher biomass productivity per unit area than open forests. Brown and Gaston (1996) reported mean aboveground biomass values of 208.7 Mg ha-1 and 67.1 Mg ha-1 for closed and open forests in tropical Africa, respectively. Biomass production per unit area from closed forests is higher than that from open forests in many tropical African countries (Table 2).

Biomass And Bioenergy Research In Tropical Africa…

393

Biomass Production of Major Tropical Forest Plantation Species The high productivity associated with many tropical forest plantation species, makes them important in meeting the ever growing demand for biomass in Africa. Among the factors responsible for the increasing trend of forest plantation estate, their ability to produce high amount of biomass within a relatively short period of time and their fast-growth rate are the most notable. Forest plantations possess the capacity of producing between 3 to 10 times greater aboveground woody biomass per ha than natural forests (Pandey, 1995; Evans, 1999; Evans and Turnbull, 2004). For instance, while the maximum mean annual volume increment (MAI) in a natural tropical rain forest in Nigeria is 5 m3ha-1yr-1, that of an adjacent Nauclea diderrichii and Gmelina arborea plantations are 16 and 51 m3ha-1yr-1, respectively (Lowe, 1997; Onyekwelu, 2001). The desire for high productivity from forest plantations informed the use of fast growing trees species. However, biomass productivity in forest plantations depends on several factors such as tree species, soil quality, amount and distribution of rainfall, silvicultural practice, planting space, etc. In tropical Africa, over 30 different exotic and indigenous tree species have been established in plantations, with exotic species dominating (Pandey, 1995). This is particularly true of Nigeria where exotic species plantations account for over 80% of total forest plantations in the country (Onyekwelu, 2001). The dominance by exotic species can be attributed to their faster growth rate as well as shorter rotation length. Due largely to climatic and ecological differences, the dominant plantation species in tropical Africa varies from region to region. Pandey (1995) reported that Eucalyptus spp (E. camaldulensis, E. microtheca), Acacia species (A. albida, A. nilotica) Tectona grandis, Gmelina arborea, Azadirachta indica, Cassia simaea, and Prosopis spp are the dominant exotic species in West Sahel, Sudan and West Africa while the dominating indigenous species in these regions are Terminalia spp (T. superba, T. ivorensis), Nauclea diderrichii, Triplochiton scleroxylon, Khaya ivorensis and Mansonia altissima. In tropical Southern Africa and Kenya more than 50% of plantations are reported to be conifers like Cupressus lusitanica, Pinus spp (P. caribaea, P. elliottii, P. patula, and P. kesiya) while the rest are broadleaved like Eucalyptus spp (E. globulus, E. robusta, E. saligna), Gmelina arborea, Acacia mearnsii, Cassia simaeaa and Casuarina equisetifolia (Pandey, 1995). Biomass productivity is generally higher in forest plantations in moist forest zones than in the dry zones. For example, Eucalyptus species plantations accumulated 64.7 t ha-1 of biomass in 4 years in moist forest zone of Congo but produced only 43.2 t ha-1 in 7 years in the dry tropics of Cameroon (Table 3). Other species in the dry forest zones (e.g. Acacia species) also yields small amount of biomass. Other exotic species with high biomass productivity in the moist forest zones of Africa are Gmelina arborea and Tectona grandis. Pinus caribaea produced relatively lower biomass when compared with other exotic species plantations in moist forest zones. The indigenous species plantations (Nauclea diderrichii and Terminalia spp) in the moist forest zone generally produced lower biomass than their exotic counterpart in the same zone but relatively more biomass than the plantations in the dry zones (Table 3). Mean annual biomass increment followed the same trend as stand biomass production, with forest plantations in moist forest zones generally having higher mean annual biomass increment than those in dry forest zone. Depending on the species and plantation age, annual biomass increment varied from 5.6 to 27.2 t ha-1yr-1 in the moist forest zone and from 1.0 to 11.2 t ha-1yr-1 in the dry forest zone (Table 3). The exotic species have higher biomass

394

Jonathan C. Onyekwelu and Shadrach O. Akindele

increment (between 12.2 and 27.2 t ha-1yr-1) than the indigenous species (between 4.6 and 14.9 t ha-1yr-1). Given proper planning, good management and application of tree breeding, much higher biomass yield is possible from forest plantations as has been demonstrated by the results of some research works. For example, it is generally accepted that mean annual volume increment (MAI) of fast growing Eucalyptus spp is between 30 and 40 m3ha-1yr-1, with maximum value of about 50 m3ha-1yr-1. With good provenance selection and management, the MAI of Eucalyptus urophylla increased from average of 20 to 30 m3ha-1yr-1 to 83 m3ha-1yr-1 at age 8 years in Mangombe, Cameroon (Eldridge et al., 1993; FAO, 2001c). Similarly, the use of genetically improved Eucalyptus grandis significantly improved the MAI of its plantations in Brazil from an average of about 25 m3ha-1yr-1 to between 70 and 89.5 m3ha-1yr-1 (Betancourt 1987 cited in FAO, 2001c; Pandey, 1995). Similar results have also been reported for plantations of other species e.g. Leuceana leucocephala (Lamprecht, 1990).

Future Biomass Production from Forest Plantations The increasing trend of forest plantation area in tropical Africa (Evans and Turnbull, 2004) is an indication of its greater contribution to future biomass and bioenergy supply in the continent. This is in consonance with the estimations of Marrison and Larson (1996), who projected that the yearly contribution of forest plantations to total bioenergy supply in Africa will increase to 39.4% and 60.4% in year 2025 and 2050, respectively. The past three decades of forest plantation establishment in Africa could generally be described as a learning period, as the period was largely characterised by species selection, provenance trial, site selection, matching of species with site, determination of appropriate silvicultural and management techniques. It is therefore expected that the application of the experiences acquired during last three decades will results to improved productivity of future plantations, as is being done in some tropical African countries. For instance, Eucalyptus spp plantations established in the savannah region of Congo in 1978 was generally unproductive with an increment of about 12 m3ha-1yr-1 but research results indicated that productivity of these plantations could be greatly improved to 25-30 m3ha-1yr-1 through the use of more productive clones (FAO, 2003b). Consequently, the old and relatively unproductive plantations are currently being replaced with plantations of the productive clone at the rate of about 500 ha per month since 2001, which has resulted to a total of about 42,000 ha of plantations (FAO, 2003b). Future forest plantations for biomass production (energy plantations) are likely to be established with tree species of high productivity planted either on good quality agricultural land or on marginal land. Also, the emerging and growing markets for forest plantation products (Akindele and Fuwape, 1998; Onyekwelu, 2002) will encourage further establishment of forest plantations and thus increased biomass production from forest plantations. However, given the current move towards a private sector driven economy by governments of several African states, government minimal investment in public enterprises as well as the paucity of funds, the prospects of governments of African countries sustaining their investment in large-scale forest plantation establishment in the future appear weak. In some African countries (e.g. Nigeria), government has already commenced the process of contracting out the management of existing forest plantations as well as further establishment of forest plantations to private sector. Consequently, future biomass production from forest

Biomass And Bioenergy Research In Tropical Africa…

395

plantations in Africa will mostly likely be private sector driven. This development, therefore, calls for a change in the notion that contiguous, large-scale plantations are required to make a significant contribution to national or global bioenergy supply as well as to take advantage of economies of scale that can make bioenergy competitive. This has become imperative given the fact that most private sectors do not have the financial muscle that large-scale forest plantation establishment requires. An alternative is therefore to target biomass production from small-scale forest plantation programmes such as village/community or individual woodlots (farm forestry system), home gardens, etc. A woodlot is any area of farmland with trees, with the primary purpose of providing bioenergy (e.g. fuelwood) and the secondary purpose of providing shades, shelter, soil improvement, animal fodder, etc. However, woodlots are not necessarily restricted to farmlands but include small plantations or group of trees on farms, around villages and on waste lands (Evans and Turnbull, 2004). The primary importance of woodlots is the provision of bioenergy, though they can produce small roundwood such as posts and poles. Bioenergy need was one of the major motivating forces for the development of woodlots in social forestry programmes in the late 1970s, especially in Africa and Asia (Pandey, 1995). It is however pertinent to note that the current policy of woodlots programmes in tropical African needs some modifications if it must play a significant role in future biomass and bioenergy production in the continent and beyond. Currently, most woodlot programmes are individually based and are mostly centred on meeting the fuel needs of immediate families. The essential motivation to produce biomass for local, regional and national markets is usually lacking. The motivation can be provided by the joint efforts of government, private companies and bioenergy researchers. Where this motivation has been provided, biomass yields from woodlots were not much lower than those reported for large-scale plantations as was the case in Brazil (Larson and Williams, 1995) and yields were expected to increase as farmers apply improved methods and approaches from bioenergy researchers. The Brazilian experience is briefly discussed for emphasis and with the hope that African countries can learn from it. The rapid growth of farm forestry systems (woodlots) in Brazil is attributed to the encouragement from the private sector, the federal, state and local governments as well as the growing motivation of farmers. Several hundred thousand hectares of forest plantations with predominantly Eucalyptus spp have been established in Brazil through this programme (Larson and Williams, 1995). Small-scale farm forestry programmes in Brazil is a symbiotic relationship between farmers on the one hand and forest companies on the other. In a typical Brazilian farm forestry programme, the material inputs (e.g. seedlings, fertilisers, herbicides, pesticides, etc) and technical know-how for establishing trees on a farmer's land are usually provided by a forestry company, which contracts with the farmer to buy some or all the products from the first harvest for an agreed price that incorporates repayment for the initial inputs and services (Larson and Williams, 1995). In this way, the farmers were motivated as they were assured of the necessary inputs and services for establishing the woodlots as well as a ready market for the products. On the other hands, the forest companies have a continuous flow of biomass at a reasonable price. Through this initiative, farmers-owned plantations in Brazil have risen to account for about 20% of forestry companies' total plantation area, with a plan of increasing this proportion to about 50% or more (Larson and Williams, 1995). Some tropical African countries have begun to undertake similar initiative. In DR Congo for example, an NGO (Action Centre for Integrated and Sustainable Development in the

396

Jonathan C. Onyekwelu and Shadrach O. Akindele

Community) in conjunction with a local community in the Sud-Kivu Province, have established 22 ha of agroforestry plantations for primarily fuelwood production (FAO, 2003c). The plantation is jointly managed by the NGO and a village management committee. In Cameroon, a company (SOCAPALM) has indicated interest in involving local people in rubber and palm production by promoting small-scale plantations with technical support from the company and with the plan to buy the produce from the plantations (FAO, 2003c). For farm forestry and small-scale energy plantation programmes to contribute significantly to future biomass production in Africa, some incentives to encourage private sector participation in forestry must be put in place by governments of the various tropical African countries. Some of these measures are briefly discussed below. -

-

-

-

Financial incentives: Although there is growing interest by the private sector (individuals, communities, corporate bodies) in establishing forest plantations in Africa, this interest is being militated against by some government policies as well as the capital intensive nature of forest plantation establishment and management, which a high percentage of the populace in Africa cannot afford. Thus there is the need to create soft loans and easy access to credit, with very low interest rates. The loans should have a long repayment period, probably as long as the rotation period of the intended energy plantation. Tax incentives: Tax exemption on at least the products of the first rotation should be granted while tax on the products of subsequent rotations should be low. Also royalties should not be charged for exploiting trees from small-scale forest plantations. These measures will contribute to enhanced biomass production by encouraging private individuals and communities to establish forest plantations on marginal cropland and degraded non-crop land. Ready market: People will be encouraged to establish small-scale plantations if they are sure of the market for the products. Measures should be taken towards developing both local and international markets for products from small-scale plantations. Promotional initiatives: An agency of government should be charged with the responsibility of promoting the establishment of bioenergy plantation. The agency should educate the people on the economical, social and biological importance of small-scale plantations, inform them of available financial and tax incentives as well as available market prospects for the products. The agency should interact with the private sector on a regular basis.

STATUS AND PROSPECTS OF BIOENERGY TECHNOLOGIES IN TROPICAL AFRICA Recent advances in bioenergy research have shown that biomass can be converted to different types of final energy such as heat, liquid fuels, gaseous fuels and electricity, with generation of electricity being the most popular and advanced. Biomass power has the potential of being the most important renewable energy option within the next 25 years (Lal and Singh, 2000). Modern bioenergy technologies can be set up in virtually any location

Biomass And Bioenergy Research In Tropical Africa…

397

(urban as well as rural) where bioenergy plantations exist or can be established. In addition bioenergy has the advantage of being relatively cheaper than most fossil fuel options and produces fewer emissions than petroleum fuels. A number of biomass fuelled electricity generating technologies, such as modern wood fired plant, whole tree energy system (WTE), biomass co-firing, biomass integrated gasification combined cycle (BIGCC) and Stirling engine, etc., are currently available (Bhattacharya et al., 2003). Bio-based refineries that produce biofuels for transportation (e.g. biodiesel, ethanol) as well as bio-based chemical are also available (USDE, 2005). However, most of these technologies are only available in some developed countries. In the U.S., for example, hundreds of power plants use biomass resources to generate about 65 billion kilowatt-hours of electricity each year but in China, biomass electricity generation technology is still on its testing stage (Junfeng and Runqing, 2003; USDE, 2005). In some parts of Germany, biodiesel is already available to the public in some fuel stations. In the U.S., biodiesel is not yet widely available to the general public but some federal, state, and transit fleets, as well as tourist boats and launches use blended biodiesel or pure biodiesel and more than 300 billion pounds of bio-based products (plastics, cleaning products, natural fibres, natural structural materials, and industrial chemicals from biomass) are produced each year (USDE, 2005). The biodiesel association of Australia, which aims at building viable and sustainable biodiesel industry in Australia, has been successful in promoting biodiesel and lobbying for its acceptance by Local, State and Federal Governments (BAA, 2005). Contrary to the above developments, there is persistence reliance on small-scale traditional biomass conversion technologies and limited use of modern ones in Africa. Over the centuries, the dominant bioenergy technology in Africa has remained charcoal production, which has continued to be a major source of employment to the rural poor (Karekezi and Afrepren, 2003). The charcoal production system in Africa relies on the traditional and rudimentary earth kiln technology, which is an inefficient biomass conversion technology as it results in very high loss of energy during the kilning process (Arnold et al., 2003; Karekezi and Afrepren, 2003). Most earth kilns in Africa have an energy conversion efficiency of between 20 and 25% (Arnold et al., 2003), which implies that large quantities of wood are transformed into relatively small quantities of charcoal. Efforts to improve and modernise small-scale biomass energy constitute an important component of national energy strategies in many African countries, especially the sub-Saharan countries, and could potentially yield major benefits to both the urban and rural poor. The use of newer metal kilns has resulted to an average conversion efficiency of about 30, with the potential of increasing it to 50% if more efficient kilns used in other parts of the globe are adopted. There have also been recent attempts at using briquetting technology for biomass conversion in Africa. Given the enormous quantity of wastes from wood processing industries7 in Africa, briquette industry can be said to have a bright prospects. Despite this prospect, briquette production in Africa is still in small-scale. However, there is the need to adopt more modern, efficient and sustainable bioenergy production, conversion and end-use technologies as this could secure and enhance the renewable energy base of not only rural Africa but also the urban centres.

7

Due the current wood processing technology, log recovery ratio in many wood processing industries is between 40 and 60% (Akindele and Fuwape, 1998; Onyekwelu, 2002). This implies that about 50% of the logs taken to the wood processing industries end up as wastes, which are mostly disposed.

398

Jonathan C. Onyekwelu and Shadrach O. Akindele

There has been recent interest in modern bioenergy technology in Africa. This interest is driven by the high prospects for the wide-scale development and distribution of bioenergy, the increasing trend of crude oil price8 and the consequent increase in the cost of conventional energy sources (e.g. petrol, diesel, electricity, kerosene, etc), which is becoming increasing difficult to afford in Africa as well as the recurrent crises faced by most power utilities in Africa. For example, in year 2000 alone, Ethiopia Kenya, Malawi, Nigeria and Tanzania faced unprecedented power rationing, which adversely affected their economies (Karekezi and Afrepren, 2003). Despite the prospects and interest in bioenergy technologies, there is a general lack of experience in emerging modern biomass conversion technologies in Africa. In addition, technical and financial barriers have largely contributed to the inability to adequately harness the prospects of modern bioenergy conversion technologies in Africa. In a most cases, biomass conversion and utilisation technologies are still at research and development (RandD) stage. For example, a GTZ sponsored projects on “Regional biomass energy conservation programme for Southern African countries" covering such countries as Lesotho, Malawi, Mozambique, Namibia, South Africa, Zimbabwe, has among its objectives: to develop new and adapt existing energy-saving technologies for households and small-scale industries; support production, marketing and installation of energy-saving technologies, elaborate strategies to promote rational biomass use measures and exchange of experience among partners (Anon, 2005). A carbonization and pyrolysis thermal reactor, which is effective in converting wood to fuel and chemical products was developed Titiladunayo and Fapetu (2003) but is yet to be commercialized for use in cottage industries. Brew-Hammond (2001) reported that Ghana’s “vision 2020 can and should serve as a framework for knowledge-based, private-sector-led and government-supported renewable energy resource development and capacity building not only for Ghana, but also for the West African SubRegion as a whole”. Although not yet developed, modern bioenergy technologies appear to hold great promise for the future in Africa. The evolution of modern bioenergy technologies and the consequent increased use of bioenergy will benefit farmers and rural communities. It is expected that each new bioenergy industries will increase the income of their host communities, provide employment and make energy readily available, thus increasing the standard of living of rural and urban populace alike. But to harness the potentials of bioenergy technologies, key challenges must be overcome. Firstly, the growth of an integrated bioenergy industry that links resources with the production of a variety of energy and material products must be facilitated. Secondly, an aggressive campaign to promote the use of bioenergy should be undertaken by both the government and the private sector. Thirdly, the cost of bioenergy must be kept considerably lower than those of competing conventional fuels. If it costs less or even the same to make electricity, transportation fuels and products from fossil fuels than it does to make bioenergy, people will neither be interested in investing in bioenergy technology nor in its consumption. Lastly, it must be ensured that increasing use of bioenergy will not adversely affect our environment.

8

Crude oil price recently reached an unprecedented record high of over US$70 a barrel.

Biomass And Bioenergy Research In Tropical Africa…

399

STATE OF BIOMASS AND BIOENERGY RESEARCH In sub-Saharan Africa, an average of 52% of the population lives on less than US$1.00 per day (ECA, 2001). Over the past decade, opportunities for employment and household income generation have diminished thereby limiting family savings and ability to invest in modern energy systems (Kituyi, 2002). Consequently, at least 80% of the African population continues to depend on traditional biomass fuels (charcoal and fuelwood) for their energy needs. This heavy dependence on bioenergy resources has resulted in large scale deforestation and forest degradation in many countries in tropical Africa. To promote sustainability of the resources, series of studies have been conducted in various parts of the world. These research activities cover the entire spectrum of material flows from production and extraction through processing to utilization and disposal.

Studies on Assessment of Biomass and Bioenergy Resources Assessment of forest biomass in tropical Africa has been carried out at local and regional levels with the aim of providing reliable estimates on the state of the resource. The methods adopted can be broadly classified as: (a) Direct estimation of forest biomass from field data (e.g. Stomgaard, 1985; Fuwape and Akindele, 1997; Akindele, 2001; Fuwape et al., 2001; Onyekwelu, 2004). (b) Estimation from existing forest inventory data (e.g. Brown and Lugo, 1984; Brown et al., 1989). (c) Estimation using remote sensing and GIS techniques (e.g. Millington et al., 1989; Millington et al., 1994). The first method involves harvesting sample trees and taking weighed samples from their components (mainly stem, branches, and foliage) for oven-drying. The ratio of fresh to dry weight is then used to calculate the oven-dry weight of the sample trees. The results are then projected for the entire population from where the sample trees were selected. A common feature of this method is the development of regression equations for computing biomass estimates from measured tree dimensions. The most popular equation form is the allometric regression equation, expressed as a logarithmic function. This method of biomass estimation provides the most accurate estimates but it is often used for relatively small areas. In tropical Africa, this method has been used for many plantation species such as Pinus caribaea (Kadeba, 1991), Gmelina arborea (Nwoboshi, 1985; 1994; Fuwape and Akindele, 1997; Akindele, 2001; Fuwape et al., 2001; Onyekwelu, 2004), Nauclea diderrichii (Fuwape et al., 2001; Onyekwelu, 2005 (in press)), Eucalyptus species (Laclau et al., 2000; Deans et al., 2003; Nzila et al., 2004), Terminalia species (Ola-Adams, 1993; Deans et al., 1996), Triplochiton scleroxylon (Deans et al., 1996), Leucaena leucocephala (Oyerinde and Akindele, 2004), Acacia species (Okello et al., 2001; Deans et al., 2003; Harmand et al., 2004), Prosopis species (Deans et al., 2003), and Azadirachta indica (Deans et al., 2003). The high level of species diversity in the natural forests has hindered the use of this method for computing their biomass estimates.

400

Jonathan C. Onyekwelu and Shadrach O. Akindele

The second method that is used in the assessment of biomass estimates in tropical Africa is biomass estimation from existing forest inventory data. Since volume estimates are available for larger areas, this method is very useful in obtaining biomass estimates on regional or national basis. To obtain the biomass estimates, stand volume is multiplied by volume-weighted average wood density and biomass expansion factor (Brown, 1997). The biomass expansion factor is the ratio of aboveground oven-dry biomass of trees to oven-dry biomass of inventoried volume. This method can be applied to both natural forests and plantations using the average values recommended by Brown (1997) for various forest types and tree species. However, the method has a major setback arising from the fact that most forest inventory programmes concentrate on forests with commercial value and tree species above a certain minimum diameter limit. Biomass components such as branches, twigs, dead wood, woody undergrowth, shrubs and non-commercial species are usually excluded under this method. Areas with relatively sparse vegetation such as woodlands and shrub lands are also excluded, whereas they are very significant bioenergy resource base. The third method, which is usually applied to very large areas, involves the use of remote sensing and GIS techniques in biomass estimation. The use of this method requires the acquisition and review of existing satellite imageries. The low-spatial-resolution imageries are suitable for large country or multi-country region, while the high-spatial-resolution imageries are suitable for small country or portion of a country. These imageries are used to produce land cover maps. Low intensity biomass inventory is then carried out in each of the delineated land cover types. This will involve destructive sampling to establish biomass regression equations for land cover types where such equations are not available. The equations are then used to estimate the biomass of each land cover type. The results can be aggregated over large areas to reflect regional estimates. Remote sensing technology allows for acquisition of repeated imageries for any particular scene. This makes this method very suitable for monitoring biomass resource. The major setback to the widespread use of this method in tropical Africa is the high cost of acquiring the satellite imageries and the level of technical skill required in interpretation of the imageries. Since most of the countries in the region are the poorest in the world, little or no funds are set aside by their various Governments for biomass and bioenergy research involving the use of remote sensing technology. However, the need for research on assessment of the bioenergy resource base in tropical Africa was emphasized at a Household Energy Seminar organized by the Energy Sector Management Assistance Programme (ESMAP) in Harare, Zimbabwe, in 1988. ESMAP is a joint program of the World Bank and the United Nations Development Programme. Participants at the seminar concluded that there was the need to address the lack of data on energy resources for most African countries. As a follow-up to the seminar, funding was provided by the Government of the Netherlands for mapping of Africa’s land cover and assessment of its woody biomass. It was the first time a study of such magnitude was carried out for the continent. Using the Normalized Difference Vegetation Index (NDVI) with satellite data acquired by Advanced Very High Resolution Radiometer (AVHRR), land cover maps and descriptions were generated for forty-three distinct classes. For each class, estimates of woody biomass were computed. Although the accuracy of the assessment was limited by the resolution of the satellite imageries and by the paucity of existing data on woody biomass in the various vegetation types, the study resulted in a very useful mapping reference (Millington et al., 1994).

Biomass And Bioenergy Research In Tropical Africa…

401

Taking advantage of geographic information systems (GIS) technology, a different method of estimating forest biomass density for tropical Africa has been used by Brown and Gaston (1995). The method was a multi-stage approach based on the assumption that the present distribution of forest biomass density depends on the potential amount that the landscape can support under prevailing environmental conditions, and the cumulative impact of human activities on forests that reduce the biomass density. Several spatial data layers such as climatic index map, FAO soil map, Forest Resources Assessment (FRA) 1990 vegetation map, soil map, etc. were incorporated into the GIS and overlaid to produce an index of potential biomass density. The resulting digital maps were then calibrated and validated using existing information from past forest inventory and ecological studies. The influence of human activities that result in reduction of forest biomass was estimated as degradation ratio and then incorporated to obtain the final estimates of potential and actual aboveground biomass density of all forest types in tropical Africa. Detailed description of the procedure has been given by Brown (1997), Brown and Gaston (1995) and Brown et al. (1996).

Studies on Improved Utilization of Biomass for Energy Traditional biomass energy use in Africa is characterized by indoor air pollution (Muchiri and Gitonga, 2000) and heat loss (Openshaw, 1980) due to inefficient systems used for combustion. In general, women and children spend much of their time near biomass-based cooking fires, and are therefore more adversely affected by particle emissions from biofuels smoke. A recent study in a rural area in Kenya found that women, who undertook most of the cooking at the household level, were exposed to twice as much particulate emission as their male counterparts, and were on the average twice as likely to suffer from respiratory infections (Schirnding, 2001). The key challenge facing Africa is not to increase energy consumption per se, but to ensure access to cleaner energy services preferably through energy efficiency and renewable energy thus promoting sustainable consumption (Karekezi et al., 2004). Consequently, some research efforts in tropical Africa have been directed towards development and dissemination of improved biomass energy technologies. An example is the development of a thermal reactor for carbonization and pyrolysis in Nigeria (Titiladunayo and Fapetu, 2003). The thermal reactor was very effective for converting wood to fuel and chemical products, and with more laboratory and household trials, it could be commercialized and used in cottage industries to boost the micro-economy of rural areas in tropical Africa. Other specific examples include the development of shielded fire stoves (Kabuleta, 2004) and household rocket stove (Scott, 2005) in Uganda, as well as the Kenya Ceramic Jiko stove (KENGO, 1991) and the Maendeleo/Upesi Improved Woodfuel Stove (Muriithi, 1995) in Kenya. These improved stoves are designed to reduce heat loss, increase combustion efficiency and attain a higher heat transfer. Since they produce less smoke than the traditional biofuel stoves, they can reduce respiratory health problems associated with smoke emission. In addition, the use of the improved biomass technologies will help to alleviate the burden placed on women in fuel collection, freeing up more time for them to engage in other activities, especially income generating activities (Karekezi and Kithyoma, 2002). Other aspects of bioenergy research in tropical Africa include assessment of the combustion characteristics of some tropical fuelwood species (Lucas and Fuwape, 1984; Abbot et al., 1997), determination of the energy

402

Jonathan C. Onyekwelu and Shadrach O. Akindele

value of short-rotation species (Fuwape, 1993; Fuwape and Akindele, 1997), evaluation of energy requirements for household cooking (Openshaw, 1980), assessment of the effects of carbonisation temperature on charcoal production (Fuwape, 1996), and household energy consumption pattern (Marufu et al., 1997; Karekezi et al., 2004).

CHALLENGES FACING BIOMASS AND BIOENERGY RESEARCH IN TROPICAL AFRICA Biomass and bioenergy research has not received enough attention from various Governments in tropical Africa in terms of funding. Energy supply paradigms have tended to be built on lucrative energy resources (e.g. oil) that supply energy to the industrial complex and urban centres. Access to modern energy in particular is restricted by the low incomes of the majority of Africans, hence limiting affordability (Zhou, 2003). Also, the low access indirectly limits the reliability of supply and energy distribution networks in rural areas. Since the economies of the various countries are dominated by a few multinational corporations that are often profit oriented, African Governments tend to owe them more allegiance at the detriment of the poor. While the lucrative energy resources are being exploited by foreign oil companies for their markets in the developed world, the bioenergy resources on which majority of the people depend are largely neglected. African Heads of Government have proposed the New Partnership for Africa’s Development (NEPAD) initiative as the pathway towards sustainable development of the continent. The initiative aims to foster cooperation and cohesion among Africans, and realistically address the problems of poverty, economic retardation and marginalization of women, children and the disabled. According to Mochebelele (2004), energy projects in the NEPAD’s short-term action plan are: • • • • • • • • • • • • • • •

Mepanda Unkua Hydropower Ethiopia-Sudan Interconnection West Africa Power Pool (WAPP) Algeria-Morocco-Spain Interconnection (strengthening) Algeria-Spain Interconnection and Algeria Gas-Fired Power Station Mozambique-Malawi Interconnection Kenya-Uganda Oil Pipeline West Africa Gas Pipeline Grand Inga Integrator DRC-Angola-Namibia Connection Nigeria-Algeria Gas Pipeline Sub-Regional Interconnections (East, West, South, Central) Feasibility studies Requirements to facilitate AFREC Operationalisation REC Capacity Building

Biomass And Bioenergy Research In Tropical Africa…

403

It is clear from these projects that NEPAD places no emphasis on bioenergy development. It appears that African leaders are yet to learn from the failure of past similar initiatives that were largely driven by donors and foreign experts, with short term interests in the outcomes of the initiatives. The funding of these projects will require assistance from development agencies, most of who often provide funds with strings attached. Even if the new projects succeed and energy is also channeled to rural areas, the level of poverty in these rural areas will make it unaffordable. If energy is to make significant impact in reducing poverty, the focus should be to target groups with potentials to create sustainable employment and improve incomes in the rural and peri-urban areas. Since these are the areas where most of the populations reside, energy development in these areas will make the largest impacts. Most of the studies on biomass in tropical Africa are small isolated studies, covering very few tree species in specific localities. Sometimes, different methodologies are used thereby making comparison difficult. Where large scale studies have been carried out, the results often incorporate many assumptions for validity due to the paucity of data on which they are based. In addition, such studies are usually conducted by foreign experts without much attention on building local capacity for subsequent studies. Where local experts are available, some of the modern research techniques cannot be used due to lack of facilities. Unlike in the developed countries where various agencies provide research grants, African scientists have frequently had abandon novel research ideas due to lack of funds. The conclusions about future availability of forest wood as well as residues from forest industries and agriculture vary substantially among previous studies. While many concluded that forest wood will continue to play a leading role as a potentially major source of biomass for energy in the future, others presented a less prominent role of forest wood in supplying biomass for energy. The divergence can be explained by different approaches to estimating the bioenergy potential of forest wood: the lower end estimates restrict the bioenergy potential to certain shares of the wood flows in the forest sector (and thus to the future forest product demand), while the higher end estimates do not make such restrictions (Berndes et al., 2003). In computing regional estimates, there is always the problem of having to reconcile different values for the same area.

FUTURE DIRECTIONS During the combustion of plant biomass, CO2 is released into the atmosphere, along with CO, CH4, NO, N20, aerosol particles and a wide range of organic compounds, which are collectively called non-methane hydrocarbons or volatile organic compounds (Crutzen and Andreae, 1990). Due to the scarcity of, and inconsistencies in the data on biofuel consumption, it is not well established, and there is little agreement on, just how much domestic biomass burning contributes to the emission of these trace gases into the atmosphere (Marufu et al., 1997). There is the need for further studies to obtain reliable data on household biofuel consumption, especially in Africa where dependency on bioenergy in on the increase. The methodology for carrying out such studies should be standardized so that the results will be compatible to allow for aggregation of the data in order to compute regional estimates. The plant species diversity in tropical Africa is very high. Only relatively few of them have been studied in terms investigating their suitability as feedstocks in biomass energy

404

Jonathan C. Onyekwelu and Shadrach O. Akindele

production. There is the need for further research into the combustion characteristics of many tropical plant species which can serve as potential feedstock in the bioenergy industry. Such research has the potential of helping us to identify those species we should focus on due to their desirable properties. In addition, researchers should take advantage of the advances in biotechnology to conduct plant improvement research aimed at producing fast-growing genetically engineered plant species with high energy values. This will increase both the level of usable biomass resources from crops per unit of input and the productivity per unit area. The increase in biomass yields as well as the higher levels of valuable constituents in plants grown for bioenergy products will help to reduce the pressure on the tropical rainforest and savannah ecosystems. Some progress has been made in the area of developing improved biofuel stoves that are more environmentally friendly than the traditional stoves. Research is needed to advance the technologies for these improved stoves and also develop new technologies. In particular, researchers across the region should work together on pilot plant facilities that focus on evaluating and developing processing technologies for bioenergy products using a variety of raw material resources. Relevant Government agencies and the industry should also be involved so that together they can work out how best to commercialize these technologies without making them too expensive for the rural poor. In the area of biomass thermochemical conversion research, future directions should be on how to improve existing biomass gasification technologies to enable the conversion of a wide range of feedstocks, including forest and agricultural residues. There will also be need for analytical studies on performance and costs of the new technologies with a view to making them affordable when commercialized. Over the years, there has been a decline in the quality of training programmes in many institutions in tropical Africa. The major reason for this is poor funding, which has resulted in massive exodus of the cream of African scholars to Europe and North America. The facilities in many Universities and research centres in tropical Africa need to be completely overhauled in order to be able to perform meaningful research. Attention should also be focused on capacity building in the region with the aim of improving the skills of local scientists in all the areas of biomass and bioenergy research. Specifically, training in remote sensing and GIS analyses, field assessment techniques, and modern bioenergy technology should be increased. Energy policies for many countries in Africa tend to put more emphasis on conventional (fossil) energy, thus denying biomass energy the comprehensive treatment it deserves (Kituyi, 2002). There is the need to change this perspective since majority of people in tropical Africa depend on biomass resources for their energy needs. The future direction would be to formulate a coherent bioenergy policy to be coordinated by the relevant government agency in each country. Such agencies across various countries should work with the African Energy Policy Research Network (AFREPREN) to develop networks between biomass energy researchers. This will help to facilitate data sharing and continuous monitoring of biomass resources in tropical Africa. Since AFREPREN is a collective regional response to the widespread concern over the weak link between energy research and the formulation and implementation of energy policy in Africa, it can complement the activities of relevant government agencies to strengthen and harness local research capacity in formulating bioenergy policy.

Biomass And Bioenergy Research In Tropical Africa…

Figure 1: Map of African showing the geographical extent of tropic in Africa

405

406

Jonathan C. Onyekwelu and Shadrach O. Akindele

Figure 2: Land use in tropical Africa Source: Brown and Gaston (1996)

Biomass And Bioenergy Research In Tropical Africa…

407

Figure 3a: Forest types of Africa based on ecological differences Adapted from WCMC (1997) 2750 2500

Moist Forests

Dry Forests

2250 2000

2

Area (000 km )

1750 1500 1250 1000 750 500 250

Forest types

Figure 3b: Extent of forest types in tropical Africa based on ecological differences

fo re st dr y

er op hy llo us Sc l

re es /p

ar kl an

d

fo re st Sp ar se t

Th or n

le a br oa d

De c

id uo us /se m i-d

ec id uo us

ea f ne ed lel ix ed M

ff

ff le a /b ro ad

dl e oi st br oa nm

or es t

or es t

nf or es t af ra i

ra i af br oa dl e Se m i-e

ve rg re en de lan

Lo w

ve rg re e

m on ta er Lo w

nf or es t

ne fo re st

ef or es t m on ta n

U pp er

Fr es h

w

ate

rs w

M

am

pf

an gr ov e

or es t

0

Table 1: Forest resources of tropical African countries Forest area (2000) Countrya Angola Benin Botswana Burkina Faso Burundi Cameroon Cape Verde Central African Republic Chad Comoros Congo Côte d'Ivoire Dem. Rep. of the Congo Djibouti Equatorial Guinea Eritrea Ethiopia Gabon Gambia Ghana Guinea Guinea-Bissau Kenya Liberia Madagascar b Malawi Malib

Land Area (000 ha) 124 670 11 063 56 673 27 360 2 568 46 540 403 62 297 125 920 186 34 150 31 800 226 705 2 317 2 805 11 759 110 430 25 767 1 000 22 754 24 572 3 612 56 915 11 137 58 154 9 409 122 019

Natural Forest 69 615 2 538 12 426 7 023 21 23 778 0 22 903 12 678 6 21 977 6 933 135 110 6 1 752 1 563 4 377 21 790 479 6 259 6 904 2 186 16 865 3 363 11 378 2 450 13 172

Forest plantation 141 112 1 67 73 80 85 4 14 2 83 184 97 22 216 36 2 76 25 2 232 119 350 112 15

Total Forest (000 ha) 69 756 2 650 12 427 7 089 94 23 858 85 22 907 12 692 8 22 060 7 117 135 207 6 1 752 1 585 4 593 21 826 481 6 335 6 929 2 187 17 096 3 481 11 727 2 562 13 186

(%) 56.0 24.0 21.9 25.9 3.7 51.3 21.1 36.8 10.1 4.3 64.6 22.4 59.6 0.3 62.5 13.5 4.2 84.7 48.1 27.8 28.2 60.5 30.0 31.3 20.2 27.2 10.8

Area Change (total forest: 1999 -2000) (000 ha) (%) -124 -0.2 -70 -2.3 -118 -0.9 -15 -0.2 -15 -9.0 -222 -0.9 5 9.3 -30 -0.1 -82 -0.6 n.s. -4.3 -17 -0.1 -265 -3.1 -532 -0.4 n.s. n.s. -11 -0.6 -5 -0.3 -40 -0.8 -10 n.s. 4 1.0 -120 -1.7 -35 -0.5 -22 -0.9 -93 -0.5 -76 -2.0 -117 -0.9 -71 -2.4 -99 -0.7

Above ground biomass (t/ha) 54 195 63 16 187 131 127 113 16 65 213 130 225 46 158 32 79 137 22 88 114 20 48 196 194 143 31

Table 1: (Continued) Forest area (2000) Countrya Mauritaniab Mauritius Mozambiqueb Namibiab Niger Nigeria Réunion Rwanda Sao Tome and Principe Senegal Seychelles Sierra Leone Somalia Sudan Tanzania Togo Uganda Zambia Zimbabwe

Land Area (000 ha) 102 522 202 78 409 82 329 126 670 91 077 250 2 466 95 19 252 45 7 162 62 734 237 600 88 359 5 439 19 964 74 339 38 685

Natural Forest 293 3 30 551 8 040 1 256 12 824 68 46 27 5 942 25 1 049 7 512 60 986 38 676 472 4 147 31 171 18 899

Forest plantation 25 13 50 0 73 693 3 261 263 5 6 3 641 135 38 43 75 141

Total Forest (000 ha) 317 16 30 601 8 040 1 328 13 517 71 307 27 6 205 30 1 055 7 515 61 627 38 811 510 4 190 31 246 19 040

(%) 0.3 7.9 39.0 9.8 1.0 14.8 28.4 12.4 28.3 32.2 66.7 14.7 12.0 25.9 43.9 9.4 21.0 42.0 49.2

Area Change (total forest: 1999 -2000) (000 ha) (%) -10 -2.7 n.s. -0.6 -64 -0.2 -73 -0.9 -62 -3.7 -398 -2.6 -1 -0.8 -15 -3.9 n.s. n.s. -45 -0.7 n.s. n.s. -36 -2.9 -77 -1.0 -959 -1.4 -91 -0.2 -21 -3.4 -91 -2.0 -851 -2.4 -320 -1.5

Adapted from (FAO, 2001b) a b

Countries with significant portion of their land mass outside the two tropics (Cancer and Capricorn) are not included Allowances were not made for the small portion of these countries outside the tropics.

Above ground biomass (t/ha) 6 95 55 12 4 184 160 187 116 30 49 139 26 12 60 155 163 104 56

Table 2: Mean biomass production in closed and open forests in tropical African countries

Angola Benin Botswana Burkina Faso Burundi Cameroon Central African Republic Chad Congo DR Congo Côte d'Ivoire Djibouti Equatorial Guinea Ethiopia Gabon Gambia Ghana Guinea Guinea-Bissau Kenya Liberia Madagascar Malawi Mali Mauritania Mozambique Namibia Niger Nigeria

Forest type (Area) Closed forest (km2) 66750 4325 0 0 2575 221750 105425 0 247675 1860300 102250 0 24725 375450 228450 5150 74200 26975 11225 38000 79825 132975 4650 0 0 97950 0 0 185900

2

Open forest (km ) 1063775 103675 101800 97375 6900 182500 515850 316825 375 378175 164875 0 0 325400 0 875 125725 199450 11750 315250 6550 49475 83825 138575 125 639650 141175 48075 490600

Mean biomass production (Mg ha-1) Closed forest Open forest 109.4 69.7 75.6 56.4 0.0 13.1 0.0 34.0 57.1 40.3 297.5 120.0 262.9 185.7 0.0 38.0 344.9 36.4 223.9 118.2 211.8 133.4 0.0 0.0 318.0 0.0 65.3 35.6 342.5 0.0 30.2 15.7 121.5 68.3 170.7 132.7 98.2 70.2 64.5 35.6 309.2 244.5 196.3 198.1 83.9 38.7 0.0 44.4 0.0 5.6 46.6 58.6 0.0 10.7 0.0 8.4 100.3 27.8

Table 2: (Continued) Forest type (Area) Closed forest (km2) Rwanda 7300 Senegal 13500 Sierra Leone 30875 Somalia 350 Sudan 5425 Tanzania 5550 Togo 725 Uganda 12075 Zambia 47250 Zimbabwe 5800 Mean biomass production in tropical Africa

Source: Adapted from Brown and Gaston (1996)

2

Open forest (km ) 1525 63975 28400 165675 737150 529800 49575 131825 579650 360375

Mean biomass production (Mg ha-1) Closed forest Open forest 31.5 46.3 52.8 25.8 220.6 174.9 11.4 12.4 134.0 63.0 73.5 48.6 40.9 71.7 124.0 101.7 45.9 52.3 37.5 13.3 208.7 67.1

412

Jonathan C. Onyekwelu and Shadrach O. Akindele

Table 3: Biomass yield and biomass increment of some major forest plantation species in tropical Africa Location

Eucalyptus spp

Age (years) 4

Biomass yield (t ha-1) 64.7

Biomass increment (t ha-1 yr-1) 13.2 – 21.6

Gmelina arborea

21

394.9

12.2 – 27.2

Tectona grandis Pinus caribaea

25 10

352.5 158.0

13.7 – 19.8 12.0 – 15.9

Nauclea diderrichii

30

287.7

5.6 – 9.6

Nzila et al., 2004, Laclau et al., 2000 Nwoboshi, 1994; Fuwape and Akindele, 1997; Onyekwelu, 2004; Ola-Adams, 1993 Kadeba and Aduayi, 1986 Fuwape et al., 2001; Onyekwelu (2005, in press)

Terminalia spp

13

193.7

10 – 14.9

Ola-Adams, 1993; Norgrove and Hauser, 2002

Eucalyptus spp

7

43.2

5.3 – 12.0

Acacia species

14

76.7

1.0 – 10.2

Senna siamea,

6

23.91

-

Dry tropical zone

Moist tropical zone

Tree Species

References

Deans et al., 2003; FAO, 2003b; Harmand et al., 2004 Okello et al., 2001; Deans et al., 2003; Forrester et al., 2004; Harmand et al., 2004 Harmand et al., 2004

REFERENCES Abbot, P., J. Loworet, C. Khofit and M, Werren, 1997. Defining firewood quality: a comparison of quantitative and rapid appraisal techniques to evaluate firewood species from a southern African savanna. Biomass and Bioenergy 12 (6): 429-437. Akindele, S.O and J.A Fuwape, 1998. Wood-based industrial sector review. A consultancy report prepared as part of the national Forest Resources Study, Nigeria, 1998. 74 pp. Akindele, S. O., 2001. Biomass table for Gmelina arborea pulpwood plantation at Jebba, Kwara State. Nigerian Journal of Forestry. Vol. 31 (1and2): 63 - 70. Anon, 2005. B7-6201/97-21/VIII/FOR "Regional biomass energy conservation programme for Southern Africa" Tropical forestry projects information systems. http://www.odi.org.uk/tropics/projects/2013.htm Arnold, M., Köhlin, G., Persson, R. and Shepherd, G., 2003. Fuelwood Revisited: What has changed in the last decade? CIFOR occasional paper No. 39, Bogor, Indonesia. BAA (Biodiesel Association of Australia), 2005. http://www.biodiesel.org.au/ Berndes, G., M. Hoogwijk and R. van den Broek, 2003. The contribution of biomass in the future global energy supply: a review of 17 studies. Biomass and Bioenergy 25: 1–28. Bhattacharya, S.C., Salam, P.A., Pham, H.L. and Ravindranath, N.H., 2003. Sustainable biomass production for energy in selected Asian countries. Biomass and Bioenergy 25: 471 – 482

Biomass And Bioenergy Research In Tropical Africa…

413

Boukongou, J.D., 2005. The protection of the Congo basin: a multilateral challenge. African Geopolitics, http://www.african-geopolitics.org/show.aspx?ArticleId=3836. Brew-Hammond, A. 2001. Renewable Energy in Ghana – Challenges for the 21st century. First workshop on knowledge networks for sustainable energy in Africa, Accra Ghana 29 – 30 June 2001. Brouwer, I.D., den Hartog, A.P., Kamwendo, M.O.K. and Heldens, M.W.O. 1996. Wood quality and wood preferences in relation to food preparation and diet composition in central Malawi. Ecology of Food and Nutrition, 35: 1-13. Brown, S. and A. E. Lugo, 1984. Biomass of tropical forests: A new estimate based on forest volumes. Science 223:1290-1293. Brown, S., A. J. R. Gillespie, and A. E. Lugo, 1989. Biomass estimation methods for tropical forests with applications to forest inventory data. Forest Science 35:881-902. Brown, S. and G. Gaston, 1995. Use of Forest Inventories and Geographic Information Systems to Estimate Biomass Density of Tropical Forests: Applications to Tropical Africa. Environmental Monitoring and Assessment 38:157-168. Brown, S, and Greg Gaston. 1996. Tropical Africa: Land Use, Biomass, and Carbon Estimates for 1980. NDP-055, Carbon Dioxide Information Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee. Brown, S., G. Gaston and R. Daniels, 1996. Tropical Africa: Land Use, Biomass, and Carbon Estimates for 1980. ORNL/CDIAC-92, NDP-055. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee. Brown, S., 1997. Estimating Biomass and Biomass Change of Tropical Forests: a Primer. For the Food and Agriculture Organization of the United Nations. FAO Forestry Paper 134, 55pp. CIFOR (Centre for International Forestry Research), 2005. Contributing to African development through forests. CIFOR’s strategy for engagement in sub-Saharan Africa, 34pp Crutzen, P. J. and M. O. Andreae, 1990. Biomass burning in the tropics: impact on atmospheric chemistry and biogeochemical cycles. Science 250: 1669-1678. Deans, J.D., Moran, J and Grace, J., 1996. Biomass relationships for tree species in regenerating semi-deciduous tropical moist forest in Cameroon. Forest Ecology and Management, 88: 215-225 Deans, J.D., Diagne, O., Nizinski, J., Lindley, D.K., Seck, M., Ingleby, K. and Munro, R.C., 2003. Competitive growth, biomass production, nutrient use and soil amelioration by nitrogen-fixing tree species in semi-arid Senegal. Forest Ecology and Management, 176: 253 – 264. ECA, 2001. Transforming Africa’s Economies. Economic Report on Africa 2000, Economic Commission for Africa, Addis Ababa. 85pp. Eldridge, K.; Davidson, J.; Hardwood, C. and van Wyk, G. 1993. Eucalypt Domestication and Breeding. Oxford Science Publications. USA, 288 p. Evans, J., 1998. The suitability of wood production in plantation forestry. Unasylva, 192 (49): 47-52 Evans, J. 1999. Sustainability of Forest Plantations: a review of evidence and future prospects. International Forestry review, 1(3): 153 – 162. Evans, J. and Turnbull, J.W., 2004. Plantation forestry in the tropics (3rd edt.). Oxford University press, 467 pp.

414

Jonathan C. Onyekwelu and Shadrach O. Akindele

FAO, 1993. Forest Resource Assessment 1990: tropical countries, FAO Forestry Paper, 112, Rome FAO, 2000. The global outlook for future wood supply from forest plantations: Global Forest Products Outlook Study. Working Paper No: GFPOS/WP/03, compiled by Christopher Brown. FAO, Rome FAO, 2001a. State of the world’s forests (SOFO), 2001. FAO Forestry Paper, FAO, Rome 181pp FAO. 2001b. Global Forest Resource Assessment 2000. Main report. Rome. FAO, 2001c. Mean annual volume increment of selected industrial forest plantation species compiled by L. Ugalde and O. Pérez. Forest Plantation Thematic Papers, Working Paper FP/1. FAO, Rome, 27pp FAO, 2003a: Forestry outlook study for Africa: regional report - opportunities and challenges towards 2020. FAO Forestry Paper 141. FAO, Rome, 66 pp FAO, 2003b. Sustainable management of tropical forests in Central Africa: in search of excellence. FAO forestry papers- 143, FAO Rome, 126pp. FAO. 2003c. Forestry Outlook Study for Africa: Sub-regional Report, West Africa, Rome, 66 pp. Forrester, D.I., Bauhus, J. and Khanna, P.K., 2004. Growth dynamics in a mixed-species plantation of Eucalyptus globulus and Acacia mearnsii. Forest Ecology and Management 193: 81–95 Fuwape, J. A., 1993. Charcoal and fuel value of agroforestry tree crops. Agroforestry Systems, 22, 55-59. Fuwape, J. A., 1996. Effects of carbonisation temperature on charcoal from some tropical trees. Bioresource Technology 57(1): 91-94. Fuwape, J. A. and S. O. Akindele, 1997. Biomass yield and energy value of some fastgrowing multipurpose trees in Nigeria. Biomass and Bioenergy 12(2): 101-106. Fuwape, J.A and Onyekwelu, J.C., 1995. The effect of uncontrolled fuelwood collection on the environment. Proceedings of UNESCO-MAB regional training workshop, 23 - 26th June 1995, pp 240-245. Fuwape, J.A., Onyekwelu, J.C. and Adekunle, V.A.T., 2001. Biomass equations and estimation for Gmelina arborea and Nauclea diderrichii stands in Akure forest reserve. Biomass and Bioenergy (21): 401 – 405 Gonder, K. M., Oates, J. F. Disotell, T. R. Forstner, M. R. J. Morales J. C., and Melnick D. J., 1997. A new West African chimpanzee subspecies? Nature 388: 337. Harmand, J.M., Njiti, C.F., Bernhard-Reversat, F and Puig, H., 2004. Aboveground and belowground biomass, productivity and nutrient accumulation in tree improved fallows in the dry tropics of Cameroon. Forest Ecology and Management 188: 249–265 Hendrick, L.D., 2001. Forest types and classification. In: Evans, J (ed). Forestry Handbook Volume 1, Blackwell Science, London, pp 23 – 60. Hilton-Taylor, C. 2000. 1998. The IUCN 2000 Red List of Threatened Species. IUCN, Gland, Switzerland and Cambridge, United Kingdom. IEA (International Energy Agency), 2003. Energy Balances of Non-OECD Countries, 20002001. Paris, France. Jones, E. W. 1955. Ecological studies of the rain forest of southern Nigeria IV. The plateau forest of the Okomu Forest Reserve, Part 1. The environment, the vegetation types of the forest, and the horizontal distribution of species. Journal of Ecology 43: 564-594.

Biomass And Bioenergy Research In Tropical Africa…

415

Junfeng, L. and Runqing, H., 2003. Sustainable biomass production for energy in China. Biomass and Bioenergy 25: 483 – 499 Kabuleta, R. K. (Editor), 2004. How to build the improved household stoves: a construction manual for the Rocket-Lorena and shielded fire stoves. Ministry of Energy and Mineral Development, Energy Advisory Project, Kampala, Uganda, 28p. Kadeba, O. and Aduayi, E.A., 1986. Dry matter production in Pinus caribaea stand planted in a subhumid tropical savannah site. Oikos, 46: 237 – 242. Kadeba, O., 1991. Above-ground biomass production and nutrient accumulation in an age sequence of Pinus caribaea stands. Forest Ecology and Management, 41: 237-248 Karekezi, S. and Kithyoma, W., 2002. Renewable energy strategies for rural Africa: is PV led renewable energy strategy the right approach for providing modern energy to the rural poor of Sub-Saharan Africa? Energy Policy 30 (11-12) 1071-1086. Karekezi, S. and Afrepren, W.K., 2003. Renewable Energy in Africa: Prospects and Limits. In: Renewable energy development. Workshop for African energy experts on operationalizing the NEPAD energy initiative, 2- 4 June, 2003, Dakar, Senegal, 30pp. Karekezi, S., J. Wangeci and E. Manyara, 2004. Sustainable Energy Consumption in Africa. UN-DESA Final Report. African Energy Policy Research Network, Nairobi, Kenya. 54pp. KENGO, 1991. How to Make and Use the KCJ, KENGO/ Regional wood energy programme for Africa (RWEPA), Nairobi. Kituyi, E., 2002. Towards Sustainable Charcoal Production and Use: A Systems Approach. In: Proceedings of a Regional Workshop on Woodfuel Policy and Legislation in Eastern and Southern Africa (F. Mugo, ed.) RELMA, Nairobi, 4-6 March 2002. pp 46-52. Laclau, J.P, Bouillet, J.P. and Ranger, J., 2000. Dynamics of biomass and nutrient accumulation in a clonal plantation of Eucalyptus in Congo. Forest Ecology and Management 128:181-196 Lal, M. and Singh, R., 2000. Sustainable forestry as a source of bioenergy for fossil fuel substitution. In: Innes, et al. (eds), Advances in global change research (vol 3) - Biomass burning and its inter-relationships with the climate system. Kluwer Academic publishers, Boston, pp 281-298. Lamprecht, H. 1990. Silviculture in the Tropics: tropical forest ecosystems and their tree species-possibilities and methods for their long-term utilization. GTZ, Eschborn. Germany. 296 pp. Larson, E.D. and Williams, R.H., 1995. Biomass plantation energy systems and sustainable development. In: Goldemberg, J. and Johansson, T.B. (eds). Energy as an Instrument for Socio-Economic Development. UNDP, New York, http://www.undp.org/ seed/energy/policy/ Lauer, W., 1989. Climate and weather. In: Lieth, H. and Werger, M.J.A. (eds). Tropical rainforest ecosystems. Ecosystems of the world Vol 14B, pp 7 – 53, Elsevier, Amsterdam. Lowe, R.G. 1997. Volume increment of natural moist tropical forest in Nigeria. Commonwealth Forestry Review, 76 (2): 109-113. Lucas, E. B. and Fuwape, J. A., 1984. Burning characteristics of forty-two Nigerian fuelwood species. Nigerian Journal Forestry, 14(1/2), 45-52.

416

Jonathan C. Onyekwelu and Shadrach O. Akindele

Marrison, C.I. and Larson, D.E., 1996 A preliminary analysis of the biomass energy production potential in Africa in 2025 considering projected land needs for food production. Biomass and energy 10(5/6): 337-351. Marufu, L., J. Ludwig, M. Andreae, F. Meixner and G. Helas, 1997. Domestic biomass burning in rural and urban Zimbabwe - Part A. Biomass and Bioenergy 12 (1): 53-68. Millington, A. C., J. R. G. Townshend, P. A. Kennedy, R. Saull, S. D. Prince and R. Madams, 1989. Biomass Assessment. Woody biomass in the SADCC region. Earthscan, London, U.K. Millington, A. C., R. W. Critchley, T. D. Douglas and P. Ryan, 1994. Estimating woody biomass in sub-Saharan Africa. The World Bank, Washington, D.C., USA. 191pp. Mochebelele, R., 2004. The New Partnership for Africa’s Development: Implementation of NEPAD. Presentation to the African Energy Ministerial Meeting, Nairobi, Kenya on 7th May, 2004. 24pp. Muriithi, J., 1995. Women and Energy Project – Kenya an Impact Study. Boiling Point, No. 35, ITDG/ GTZ, January pp 7-8. Muchiri L. and Gitonga S., 2000. Gender and Household Energy Technology in East Africa. Energia News, Vol. 3 Issue 4. Energia Secretariat, The Netherlands Norgrove, L. and Hauser, S., 2002. Measured growth and tree biomass estimates of Terminalia ivorensis in the 3 years after thinning to different stand densities in an agrisilvicultural system in southern Cameroon. Forest Ecology and Management, 166: 261 - 270 Nwoboshi, L.C., 1982. Tropical Silviculture: Principles and Techniques. Ibadan University press 333pp. Nwoboshi, L.C., 1985. Biomass and nutrient uptake and distribution in a Gmelina pulpwood plantation age-series in Nigeria. Journal of Tropical Forest Resources, 1 (1): 53-62. Nwoboshi, L.C., 1994. Development of Gmelina arborea under the Subri Conversion Technique: first three years. Ghana Journal of Forestry, 1: 12-18. Nzila, J.D., Deleporte, P., Bouillet, J.P., Laclau, J.P. and Ranger, J., 2004. Effect of slash management on tree growth and nutrient cycling in second-rotation Eucalyptus replanted sites in Congo. In: Nambiar, et al., (eds.), site management and productivity in tropical plantation forests. Proceedings of Workshops in Congo July 2001 and China in February 2003. CIFOR, 2004, PP 15-30. Okello, B.D., O'Connor, T.G. and Young, T.P., 2001. Growth, biomass estimates, and charcoal production of Acacia drepanolobium in Laikipia, Kenya. Forest Ecology and Management, 142:143-153. Ola-Adams, B.A., 1993. Effects of spacing on biomass distribution and nutrient content of Tectona grandis Linn. F. (teak) and Terminalia superba Engl. and Diels. (afara) in southwestern Nigeria. Forest Ecology and Management, 58: 299 – 319. Onyekwelu, J.C., 2001. Growth characteristics and management scenarios for plantationgrown Gmelina arborea and Nauclea diderrichii in south-western Nigeria. Hieronymus Verlag, Munich, 196pp. Onyekwelu, J.C., 2002. Prospects of plantation grown indigenous species in wood processing industries in Nigeria. Applied Tropical Agriculture, 7: 62 – 69. Onyekwelu, J.C., 2004. Above-ground biomass production and biomass equations for evenaged Gmelina arborea (Roxb) plantations in south-western Nigeria. Biomass and Bioenergy, 26(1): 39 – 46

Biomass And Bioenergy Research In Tropical Africa…

417

Onyekwelu, J.C., 2005 (in press). Biomass yield and biomass functions for plantation-grown Nauclea diderrichii in humid tropical rainforest zone of Nigeria. Ghana Journal of Forestry. Openshaw, K., 1980. Energy requirements for household cooking in Africa with existing and improved cooking stoves. Proc. Bio-Energy 80 World Congress and Exposition, pp. 256258. Oyerinde, O. V. and Akindele, S.O., 2004. Spacing effects on the productivity of young Leucaena leucocephala woodlots. Journal of Sustainable Tropical Agricultural Research, 9: 1 – 4. Pandey, D., 1995. Forest resources assessment 1990. Tropical forest plantation resources. FAO forestry paper 128. 81pp. Parry, M. S., 1955. Planting in tropical Africa. Unasylva - Vol. 9, No. 1. Ryan, P. and Openshaw, K., 1991. Assessment of Biomass Energy Resources: A Discussion on its Need and Methodology. Industry and energy department working paper, Energy series paper no. 48, 76 pp. Sayer, J. A., C. S. Harcourt, and N. M. Collins. 1992. The Conservation Atlas of Tropical Forests: Africa. IUCN and Simon and Schuster, Cambridge. Schirnding, Y., 2001. Gender Differences in the Impact of Biomass Fuel on Health. In: Energia News, Vol. 4, No. 4. ENERGIA, the Netherlands. Scott, P., 2005. New Rocket Stove Design from Uganda. Online information. http://www.repp.org/discussiongroups/resources/stoves/Scott/ugandarocket/Uganda_Hou sehold_Rocket.pdf Singh, K.D., 1993. The 1990 tropical forest resources assessment. Unasylva - No. 174 Forest resources assessment, 1993 Stomgaard, P., 1985. Biomass estimation equations for Miombo woodland, Zambia. Agroforestry Systems 3(1): 3-13. Titiladunayo, I. F. and P. O. Fapetu, 2003. Development of a thermal reactor for the conversion of woody biomass to fuel and chemicals. Nigerian Journal of Forestry, 33(1): 32-39. USDE (U.S. Department of Energy), 2005. Bioenergy: an overview - energy efficiency and renewable energy. U.S. Department of Energy. http://www.eere.energy.gov/ consumerinfo/factsheets/nb2.html Van Horen, C. and Eberhard, A. 1995. Energy, environment and the rural poor in South Africa. Development Southern Africa 12: 197-211. WCMC (World Conservation Monitoring Centre), 1997. Forest type and protected area distribution: Africa. Were, J. L. R., 2001. Nigerian lowland forests (AT0123). http://www.worldwildlife.org/ wildworld/profiles/terrestrial/at/at0123_full.html World Bank, 1996. Rural energy and development: Improving energy supplies for 2 billion World Bank, 2003. African Development Indicators 2003. World Bank Industry and Energy Department Report No 1512 GLB. Washington, D.C. USA. Zhou, P. P., 2003. Taking the NEPAD energy initiative forward: regional and sub-regional perspectives. Paper presented at the Workshop for African Energy Experts on operationalizing the NEPAD Energy Initiative. Dakar, Senegal. 2-4 June 2003. 19pp.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 419-443

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 15

POPLAR BIOMASS OF SHORT ROTATION PLANTATIONS AS RENEWABLE ENERGY RAW MATERIAL* Bojana Klasnja, Sasa Orlovic, Zoran Galic and Milan Drekic Faculty of Agriculture, Institute of Lowland Forestry and Environment Novi Sad, Serbia and Montenegro

ABSTRACT Fast-growing broadleaf species (poplars, willows and black locust), raised in dense, short-rotation plantations, very often on the soils unsuitable for agricultural crops, produce a high yield of biomass. A significant amount of thermal energy can be obtained by direct combustion of young plant biomass (aged from one to three years) converted into chips by chipping the whole trees together with bark and branches. In this aim, the Institute carried out systematic multiannual research on the improvement of several poplar clones in order to increase the yield of biomass. Also for this purpose, the selection focused on the clones which are best adapted to the conditions of very dense planting, which is the main condition required from the foresters in the establishment of energy plantations. Based on the calorific value of wood and bark of the study poplar clones, it is assessed the quantity of energy which could be produced by the combustion of the chipped biomass of one-year, i.e. two-year-old plants. The higher heating value of wood and bark was determined for several poplar clones (Populus spp.) of different ages and plants, as well as the trees from mature plantings (aged from 8 to 14 years). By FVI (Fuel Value Index) which takes into account ash content, wood basic density, as well as moisture content, it was determined that poplar wood can be significant energy raw material, primarily because of its short production cycle and very high volume increment.

*

A version of this chapter was also published in Biomass and Bioenergy: New Research published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

420

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al. The plantations are established in two variants, by planting the cuttings of the selected poplar clones, with two planting spaces, i.e. with 38,461 plant/ha, and 83,333 plant/ha, on the previously selected and prepared soil. To define the produced biomass of individual clones, the increment elements were measured after the cycles of one and two years. Average dry matter biomass yield reached 21 t ha-1 year-1 (38,461 plant/ha), and 12 t -1 ha year-1 (83,333 plant/ha). Based on calorific values of oven dry wood and bark of each clone, average energy potential of researched poplar clones was estimated up to 395 GJ ha-1 year-1, and for denser plantations up to 222 GJ ha-1 year-1.

1. INTRODUCTION Biomass has a large energy potential. A comparison between the available potential with the current use shows that, on a worldwide level, about two-fifths of the existing biomass energy potential is used. In most areas of the world the current biomass use is clearly below the available potential. Only for Asia does the current use exceed the available potential, i.e. non-sustainable biomass use. Therefore, increased biomass use, e.g. for upgrading is possible in most countries. Biomass currently represents approximately 14% of world's final energy consumption[1]. About 25% of the usage is in industrialized countries, where a significant level of investment in environmental protection has been made to meet emissions standards, especially air emissions [2]. The other 75% of primary energy use of biomass is in heat production for developing country household energy needs and in process heat production for biomass-based industries through the use of their generated residues [2]. The world's total above-ground biomass in forests amounts to 420 (109) tones, of which more than 40% is located in South America [3]. Estimates by FAO [4] show that global production and use of wood fuel and roundwood reached about 3300 (106) m3 in 1999. About 55% is used directly as fuel, e.g. as split firewood, and about 90% of this is produced and consumed in the developing countries. The remaining 45% is used as industrial raw material, but about 40% of this is used as primary or secondary process residues, suitable only for energy production. According to the European Commission's White Paper [5], the overall aim is to double the share of renewable energy from 6 to 12% of the total energy consumption in the European Union by 2010. According to the White Paper, the major part of this renewable energy could come from woody biomass. This means that, additionally, over 160 million m3 of woody biomass per year would be used for energy in Europe [6]. However, if 50% of available agricultural residues in Europe (such as straw) were used for bioenergy purposes, then additional biomass needs are reduced to 101 million m3 of roundwood equivalents per annum [7]. Most projections of global energy use predict that biomass will be an important component of primary energy souces in the coming decades, and that SRWC will be a primary source of biomass [8]. In addition to combustion and gasification conversion pathways for power and heat production, SRWC represent a uniform, locally available feedstock for the production of bioproducts - liquid fuels, chemicals and advanced materials curently made from petroleum products.

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 421 In northern temperate areas [9], SRWC development has focused on willow clones (Salix spp.) and hybrid poplar (Populus spp.), while eucalyptus (Eucalyptus spp.) has been a model species in warmer climates. Poplars, which are the focus of this paper, have several characteristics that make them ideal for SRWC system including high yields that can be obtained in a few years; case of vegetative propagation; a broad genetic base; a short breeding cycle; ability to resprout after multiple harvests; and feedstock uniformity. While they have many characteristics in common, growth habits, life history, and resistance to pests and diseases vary considerably. This diversity is important in the successful development of SRWC. Plantations help ease shortage of forestry wood. In 1995 the industrial plantation area was estimated to be 103 million ha and the non –industrial plantation area to be 20 million ha [10]. Over 50% of the plantations are assessed to be less than 15 years and 25% are less than five years [11,12]. The establishment of new plantations is assumed to increase between 160 and 235 million ha in year 2050 [12]. Around 2030 the industrial wood supply from plantations is estimated to be 45% of the total consumption of industrial wood compared to 22% in 1997. Thus, the above-identified regional and global shortages of wood supply would be much worse without the establishment of plantations. Under short rotation intensive management the end product is generally woody biomass (feedstock) and as such, tree size and form are not characteristics of particular importance. Management objectives center on maximizing annual woody biomass yield per unit area. The success of the biomass production concept depends, in part, on the efficient production systems. Agricultural management practices (plant spacing, high density, use of herbicides, short rotation and regular harvests) are applied to fast growing tree species. Poplar appears to be a model species and prototype for such tree biomass plantations. The idea of producing large amounts of wooden biomass by cultivation of fast growing tree species with different rotation periods is a well known approach [13,14,15,16,17,18,19,20,21, 22,23,24,25,26]. Among the main conditions of energy forestry at a wider level, the decisive role is that of the available land area. The basic limiting factor of a wide introduction of energy plantations is the fact that proportionally large land areas are to meet the proportionally small part of energy demands of the country. For this reason, in many cases, energy plantations compete with agricultural lands or other land uses. To eliminate the competition and its undesirable impacts between energy and agricultural production, numerous analysts propose the reduction of energy production to degraded sites in developing countries. This will intensify the trend of using marginal lands, abandoned agricultural lands, semi-harvested or unharvested coppice forests of poor quality, land along the roads and water courses, etc. In the establishment of energy plantations, the greatest technical challenge is to determine the rotation period, in order to define the favourable period of crop rotation between two production cycles. The aim of crop rotation is to improve soil properties, the contents of organic and mineral nutrients, moisture and other properties, which is one of the conditions of keeping the yield at a desired level. Other focal problems concern all the developing regions, e.g. land ownership, lack of roads and other means for biomass transport to the plants, as well as the fact that in undeveloped regions, the owners cannot wait for 3-8 years needed for the first economic results. Although most authors advocate the advantages of establishing uniform plantations on large areas, the experience in South America shows the success of the program of forest farms

422

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

in which the companies supply the material and training to the owners of 1-50 ha holdings, with a subsequent purchase of a part of the production. The correct choice of species for the establishment of energy plantations is conditioned by many factors, the most important of which are the following: -

species potential to fulfil the management goal; fast growth in juvenile stage, especially height; harmony of site conditions and bioecological properties of the species; tolerance and resistance to harmful impacts of abiotic and biotic factors; costs of plantation establishment; rate of sale and collection on the market.

The previous experience in plantations for biomass production points to bioecological advantages of some species for the establishment of energy plantations on degraded sites, peat land and spoil banks of opencast mines and thermoelectric power stations. The key criterion for the selection of species, along with bioecological characteristics, is energy value of the produced material. Each new location requires a specific previous study, to prevent the failures resulting from some local specificities or insufficient knowledge of the planting stock origin. It is especially significant that planting stock should be produced from seeds or cuttings of known origin in the aim of adequate use of their genetic potential. In Serbia to date, the best results of biomass production in specialised plantations have been achieved with the clones and (or) varieties in the genera Salix, Populus and Robinia. Detailed study of the production potential of the above species raised in plantations for biomass production, by improving the technology of establishment and use, will reduce the pressure on natural populations as the sources of energy raw materials, the use of fossil fuels and electric energy. Site characteristics are important factors determining the yield and to some extent also the quality of biomass from SRIC. Planting SRIC on unsuitable sites may be one of the most important reasons of low yield. The basic demands of poplar cultivars used in traditional poplar plantations are: • • • • • • • • •

Vegetation periods of at least 150 days and absence of extreme early and late frost An average temperature between June and September of at least 14°C Soils with good water holding capacity, but without stagnant water Capillary moisture of soil profiles Uniform textural composition of the soil per profile depth Soils that allow deep root penetration Profile depth minimum 100 cm Good nutrient supply from the soil and pH of at least 5.5, but maximum pH 8.5 Along the large alluvial rivers in Vojvodina, the best systematic units of soil for plantation establishment are sandy loam and loamy forms of fluvisol and humofluvisol.

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 423 The main barrier for a more common use of biomass from SRC as a renewable energy source is the economic background. As an energy source, wood from SRC has to compete with fossil fuels, residues from agriculture (e.g. straw) and forestry as well as with other renewable energy sources and in most cases is inferior under the given economic and political frame conditions. Consequently, in order to promote the use of biomass from SRC as an energy source, costs for its production and use have to be decreased. This can, among others, be attained by increasing the biomass yield and optimizing fuel quality. The latter mainly influences the processing and combustion costs for biomass from SRC. The future supply of wood from plantations is based on three scenarios from the FAO [27] that differ on the net rate of plantation establishment and yields (including the effect of the time delay between plantation establishment and harvest).Two types of plantations are included: industrial plantations (established to produce industrial roundwood) and nonindustrial plantations (established for fuelwood production, water or soil protection, recreation or similar non-productive purposes). -

The low scenario of the FAO assumes no growth in the plantation area (6,9 EJ industrial roundwood in 2050 and 2,0 EJ fuelwood from non-industrial plantation). The medium scenario assumes a fixed plantation establishment rate of 1% of the 1995 plantation area, resulting in 9,7 EJ and 2,8 EJ in 2050. The high scenario assumes a gradual reduction from current establishment rates. The FAO states this scenario seems to be achievable in physical terms and represents the uper boundary of new planting rates.

Large-scale production of energy crops should not compete for land needed for food and fibre production. There have been careful calculations made that there is enough suitable land available to provide the world,s population with all its needs for food, fibre and energy throughout this century [27] althoough equitable distribution of these basic necessities is another issue yet to be resolved. In some regions the availability of water will be the constraining factor to growing energy crops rather than available land. Initial analysis has shown that sufficient arable land is available to meet all world population needs for food, fibre and energy until 2050.By 2100, the global land requirement for food and fibre production is estimated to reach about 1,7 Gha, with a further 0,69-1,35 Gha needed to support future biomass energy requirements in order to meet a high-growth energy scenario. This exceeds the 2.495 Gha total cropping land available so land-use conflicts could then arise. More detailed analysis should be undertaken to include possible constraints of local water availability, proximity to markets, export trading of biomass and social factors [28]. In the industrialised countries there is still a large area potential comprising agricultural set-aside areas, agricultural marginal land and reclamation sites of postmining landscapes [29]. More recently production of woody biomass in agroforestry systems such as alley cropping has come into focus in order to integrate crop and bioenergy production [30,31]. The objective of this study is to present the results of multiannual research of the creation and improvement of the selected poplar clones intended for the establishment of the so-called energy plantations, i.e. plantations with a very high number of plants per hectare and short rotation cycles of one to three years. The results on the biomass yield in multiannual plantations with different numbers of production cycles were analysed, as well as the effect of

424

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

rotation period on biomass yield, i.e. potential amount of energy. This study presents the results of measurement of higher heating values of wood and bark of one-year and two-yearold rooted cuttings of the selected clones, as well as of adult plants. The values of fuel value index were determined, which includes bulk density and ash content, which can affect significantly the heating value of wood. Also, based on dendrometric measurement of mean trees in two plantations with high numbers of plants (38,461plants/ha, and 83,333 plants/ha), oven dry biomass weights per hectare were calculated after one-year cycle, and the potential energy obtained by burning the total (dry) biomass.

2. MATERIALS AND METHODS 2.1. Experimental Set-Up and Plant Material In this study, the wood of Populus x euramericana cl.I-214, aged one, two and 12, Populus deltoides cl.PE 19/66, aged one, two and eight, Salix alba L. cl.378 aged one, two and 14, and Robinia pseudoacacia cl.R-54, aged one, two and ten was examined. The mature trees originated from the corresponding plantations, and one-year and two-year plants are from rooted cuttings. Mature trees in both poplar clones are from plantations with 400 trees per hectare. After the selection of characteristic sample trees (three trees in each species and age) measured parameters of growth elements were determined and the trees were felled. Sample trees were chosen as average plants based on average diameter and height on the experimental plot. An experimental field plantations were established in experimental estate “Kacka suma”. In the field trial 6 poplar clones (P.x euramericana cl. Ostia, P.nigra cl.53/86, P deltoides cl. PE 19/66, P.x .euramericana cl.I-214, P.x euramericana cl. S6-7, P.x euramericana cv. Robusta), with two different plant densities (38,461 plants ha−1, and 83,333 plants ha−1,) are being tested. All clones were planted as 25-cm-long dormant, unrooted cuttings obtained from the Institute of Lowland Forestry and Environment Novi Sad, Serbia and Montenegro. The cuttings were stored at 40 and than soaked in water for 24 hours prior to planting.. Cuttings were planted manually to a depth of 22-23cm, leaving one or two buds above the soil surface. Above ground biomass was harvested at the and of the first growing season.

2.2 Site Characteristics Short rotation coppice plantation was established on systematic units of soil fluvisol with fossil soil, with morphogenetic formula of soil Ap – I – IIGso - Ab. The soil was characterized with a great variability of properties in profile, especially of the soil textural class, and consequently soil water and air regimes. The textural class was sandy in layer I, sandy loam in Ap horizon, loamy in layer IIGso and heavy loam in horizon Ab. The pH was high (8.05 – 8.22). The content of CaCO3 was from 3.40 % in horizon Ab to 14.42 % in horizon Ap. The upper soil horizon contained between 0.98 to 1.34 % of organic matter. The nutrient reserves were low (nitrogen 0.001 to 0.046%, pottasium from 2.4 to 6.2, phosphorus 3.2 to 7.8%).

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 425

2.3. Laboratory Procedures After the selection of characteristic sample trees (three plants in each species, clones and age) measured parameters of growth elements were determined and the trees were felled. Sample trees were chosen as average plants based on average diameter and height on the experimental plot. The weight of each tree was measured, separately for wood and bark and the proportion of bark calculated. Immediately after felling, samples discs (discs cut at breast height - 130cm) were taken to assess moisture content and wood and bark densities. After natural seasoning of samples for one month at room temperature, wood was ground into wood flour suitable for pellet pressing. When a moisture content of about 10% was achieved, ash content was determined and the exact moisture content according to standard methodology [32]. The density was determined on the basis of oven-dry weight per green volume of an individual disk segment. Green volumes were obtained by soaking disk segments for 10 days in water until constant volume was achieved. Excess moisture was removed from the surface of the sample, and each sample's water displacement (volume) was measured. The sample then was oven-dried to constant weight at 1040C and weighted to determine the dry weight. For the determination of moisture content wood and bark samples were oven dried at 1040C to a constant weight. The ash content was determined by burning 5g of oven-dried and ground sample in a platinum crucible in a muffle furnace at 5500C±250C. All analyses were done in duplicate and the results were expressed on a dry weight basis. The calorific value was determined for ground air-dried samples. Pellets were made by a special device, producing pellets ranging from 0.35 to 0.64g. Samples were combusted in a Parr 1341 adiabatic calorimeter. Correction factors for the formation of acids were not included in the gross heat of combustion (higher heating value) calculations [33]. However, calorific values were corrected for moisture regained during storage. There were three replications for each sample. Also FVI (Fuel Value Index) was determined by the formula [34]: FVI =

Calorific value(kJ / g) x Density(g / cm3 ) Ash content(g / g) x Moist ure content(g / g)

3. RESULTS AND DISCUSSION 3.1 Selection of Black Poplar Clones for Biomass Production In the scope of a wider selection program, the Institute tends to create new fast growing cultivars of poplars which can be used in plantations for biomass production. In this sense, one of the main objectives is the speeding up of selection procedure, i.e. the shortening of the period between the creation of the clone and the beginning of production. The program promotes the 3 – 10-year rotation, which requires the improvement of fast growing poplar varieties, nursery production, technology of intensive plantation establishment, as well as tending operations adapted to short rotation. This led to intensified research and creation of new breeding methods, especially in the selection of the varieties - poplar clones of vigorous

426

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

growth. Taking into account that the main elements of growth in forestry (diameter, height, biomass) result from the activities of complex interactions between the plant genetic constitution and environmental factors, the study methods had to be multidisciplinary, to include all the complexity of the plant organism. For this reason, most of the poplar breeding programs are directed towards the highest possible yield of their genetic potential and adaptation values [35]. In this sense, the tendency is to create the varieties (clones) characterised by superior (fast) growth and resistance to pests and leaf, stem and bark diseases. The main problem faced by any breeder is to recognize the genotypes with desirable properties as early as possible. Shortened periods of selection, along with time-saving, also reduce the expenses required for the research. This type of research is complex because growth vigour is monitored by growth parameters (diameter, height and biomass), which are, just as the majority of other quantitative properties, controlled by several genes. For this reason, numerous anatomic properties, as well as physiological and biochemical processes related to growth, are being investigated in order to evaluate genotype potentials [35]. The Institute of Lowland Forestry and Environment in Novi Sad has established a program of long-term research of poplar anatomic properties and physiological processes [23] to accelerate the procedure of producing fast-growing poplars and to conduct selection in the earliest possible ontogenetic phases. In the first stage, the research points out the rate of variability and the potential relationship with growth elements. It is partly performed on ontogenetically younger selected material of one-year-old rooted cuttings (ramets) of the clones that are known to be characterized by fast growth in their adult stage. Along with the variability of anatomic properties and physiological processes, the results of the research will indicate the directions of further work on the creation of fast-growing poplar clones. This part of paper presents the study results of the number of net photosynthesis stomata, leaf area, stomata number, as well as diameter, height and biomass production of one-year-old rooted cuttings of poplar clones of the following taxonomy: Populus nigra L., Populus x euramericana (Dode) Guinier and Populus deltoides Bartr. A polyclonal experiment in randomised block design was established for research purposes. The greatest number of stomata per mm2 on the adaxial and abaxial was found in the clone PE 19/66 (P. deltoides) and lowest in the clone 53/86 (Populus nigra). The coefficient of variation (V) for this parameter ranging from 0.99 to 6.11 (Table 1). The coefficient of variation for this parameter ranging from 1.21 to 10.76. Table1. Stomata number per mm2 Clone 53/86 Ostia Robusta S6-7 PE 19/66 F values

Adaxial X average 67.06 122.54 112.61 134.97 172.23 83.67 ***

V, % 2.33 1.69 3.24 0.99 1.15

Abaxial X average 152.34 170.57 178.03 153.18 211.15 0.77ns

V, % 2.56 5.14 2.65 5.21 3.66

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 427 Photosynthesis intensity (net photosynthesis) was measured three times during the vegetation growth period (Table 2). Table 2. Net photosynthesis and leaf area Clone 53/86 Ostia Robusta S6-7 PE 19/66 F values

Net photosynthesis, (μmol m-2s-1) X average V, % 7.98 4.86 12.08 7.65 14.04 4.13 18.60 1.15 22.91 5.40 26.25***

Leaf area, (cm2) X average 3215.70 7298.30 10707.30 11339.50 19346.30 139.69***

V, % 1.70 1.30 2.18 1.30 3.25

The samples was taken at the end of vegetative growing period and ramets were selected randomly. Leaf area was measured by the apparatus LI 3000 (Leaf portable areameter). The results show (Table 2) that the highest average net photosynthesis and leaf area was reached by the clones of eastern cottonwood (PE 19/66 and S6-7), and the lowest - by the clone of the European black poplar (53/86). The clones of Euramerican poplar were intermediary in this respect, too. The statistically highly significant differences in this parameter were calculated among the study clones. Diameter and height were measured on 20 ramets per clone after the first growing season. The ramets were selected randomly. Diameters were measured at the height of 10 cm above the root collar. The rooted cuttings of the clones PE 19/66 and S6-7 attained the best values of diameters, heights and biomass, the clones Ostia and Robusta had the lower values, and the clone 53/86 had the lowest value (Table 3). Table 3. Stem diameter and height Clone 53/86 Ostia Robusta S6-7 PE 19/66 F values

Diameter (mm) X average 13.63 14.06 13.76 18.24 22.23 18.19***

V, % 2.88 2.54 5.55 6.22 5.64

Height (cm) X average 169.00 190.60 192.30 208.50 300.00 125.10 ***

V, % 3.62 6.25 5.21 4.25 3.65

Dry weight biomass was measured by drying the entire ramet. The samples was taken at the end of the first growing season and ramets were selected randomly. The plants were dried at 105 Co to constant weight. (Table 4). The data were subjected to various statistical analysis including: means, calculation of coefficient of variation, analysis of variance (ANOVA) and cluster analysis. The analysis of variance determined the statistically highly significant differences among the clones regarding growth elements and biomass.

428

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

Highly positive correlation between the growth elements and biomass yield and the number of stomata on the leaf adaxial and abaxial sides, net photosynthesis and leaf area was confirmed (Table 5). Table 4. Biomass of poplar clones Clone 53/86 Ostia Robusta S6-7 PE 19/66 F values

Fresh biomass, g X average 150.00 402.15 469.55 582.03 1261.92 837.53 ***

Dry biomass, g X average 45.18 108.55 179.14 181.12 302.11 30.72***

V, % 12.15 8.21 7.52 6.51 7.85

V, % 2.8 6.11 5.12 5.88 6.14

Table 5. Values of coefficient of correlation Clone Diameter Height Fresh biomass Dry biomass

Stomata number per mm2 adaxial Abaxial 0.72 0.88 0.92 0.90 0.93 0.89 0.93 0.86

Net photosynthesis

Leaf area

0.83 0.95 0.98 0.90

0.83 0.95 0.98 0.99

The cluster analysis (Fig.1) of the study physiological parameters, the growth elements and biomass production shows a tendency of clone grouping in two groups. The first group is the clone 53/86 and Ostia, and the second group is Robusta and S6-7, while the clones Ostia and PE 19/66 remained separate. 0.000

DISTANCES

5000.00

53/86

Ostia Robusta S6-7 PE19/66 Figure 1. Cluster analysis of investigation parameters

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 429

3.1.1. Discussion and Conclusion The study results show a very strong genetic variability of the analysed parameters. The analysed parameters show a statistically highly significant difference among the clones, which leads to a conclusions that they are under a high degree of genetic control. The highest number of stomata was measured in the eastern cottonwood clones, which is consistent with the studies reported by [36,37,38].This is in harmony with the leaf anatomy of these genotypes, because Orlović et al., [38], determined that they had very thick photosynthetically active tissues and fine cells with numerous intercellulars. Simultaneously, eastern cottonwood clones also had the highest photosynthesis intensity and leaf area. The results show that all the study parameters (number of stomata, net photosynthesis and leaf area) were highly correlated with the growth elements (diameter and height) and biomass production. This is consistent with the results of a series of authors [40,41,42,43], who determined that net photosynthesis was positively correlated with the yield. This practically means that these physiological parameters can be applied in the early selection of poplar genotypes for fast growth. Also, the results of cluster analysis indicate that all the study parameters are specific for the species, i.e. that they relate to the genotypes of eastern cottonwood and Euramerican poplar, which means that further research of genotype selection for biomass production should be carried out within these species.

3.2. Produced Poplar Biomass Depending on Planting Density and Rotation Period The study of energy plantations with several poplar clones was performed. In this aim, a series of experiments was established with several poplar clones in dense plant spacing of 1.00 x 0.25 m to 2.00 x 2.00 m, i.e. from 2,500 to 40,000 plants per hectare. Production cycles were one to five years, after which the coppicing ability was applied. 2-3 shoots were left on the stump and the smaller shoots were removed during the first vegetation. In this way, the initial density in the successive production cycles increased to 5,000-100,000 shoots per hectare depending on the selected plant spacing. This part of paper presents the results of the research of potential biomass production of plantations with high density which can be used for the production of bioenergy. The possibility of biomass production was researched in one-year, two-year, three-year, four-year and five-year rotations, and plantations established by poplar rooted cuttings, roots and seedlings, regenerated by coppice vigor after felling. In this way, the production process during 8-10 years and from two to nine rotations lasting from one to five years, produces annually on the average between 40.9 and 53.9 m3, i.e. 14.8 to 19.8 tons (of oven-dry mass) of wood and bark per ha, which can provide (combustion of whole tree chips) energy from 216 to 285 GJ. The paper presents the main elements of the technology of plantation establishment, tending and protection, with the main characteristics regarding the number and size of average trees, as well as the structural percentage of wod and bark in the total produced biomass.

430

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

3.2.1. Results and Discussion The results of the research mainly refer to Populus x euramericana Guinier (Dode) cl. I214, and the spacing in the plantaiton from 1.00 x .025 m (density 40,000 plants ha-1), to 2.0 x 2.0 m (i.e. 2,500 plants ha-1). All the experimental plantations were established on alluvial (fluvisol) and humofluvisol soils of the Middle Danube Basin [44]. Five experimental plantations were selected for investigations consisted of the above poplar clone - cultivar, on the soils of approximate productive capacity, with planting density and spacing which enable relatively short rotations (one to five years) and plantation renewal by coppice vigor, and with maximum wood and bark biomass productivity. The main characteristics of all the technological forms of production are as follows: Within the same production process of 8 to 10 years, the reduction of the number of rotations is determined by the development of a smaller number of trees per unit area. Thus in the plantations with one-year rotations, with plant densities from 40,000 plants ha-1 to 100,000 plants ha-1, and in the plantations with two rotations, with 2,500 plants ha-1 to 7,500 plants ha1 (Tables 6-10). Varying densities and the number of trees per unit area determine the main characteristics of sizes, quantity and structural composition (wood and bark) of the yielded biomass. Thus in the plantations with one-year rotations (Table 6), average diameter amounts to 2.0 to 2.8 cm, height 3.0 to 3.5 m. In the plantations with two rotations (Tables 9,10), average diameters are 8.6 to 12.8 cm; height 12.3 to 14.4 m. This is the main reason why volume percentage of bark in the total biomass in the plantations with one-year rotations is about 47%, and in the plantations with two rotations the share of bark is only about 16%. In the aim to assesed the produced energy of the obtained biomass in different poplar plantations, the mass of oven-dry wood and oven-dry bark (nominal densities) were previously calculated per unit area. Assesment of produced biomass weight on the basis: wood density 300 kg m -3, i.e. 450 kg m -3 for bark [45]. Table 6: One-year rotation in nine-year production process Plant spacing, m

Rotation number

I II III IV 1.00 x V 0.25 VI VII VIII IX Total for 9 years Annual average % Energy (GJ)

Plant number per ha, N 40,000 80,000 100,000 100,000 100,000 80,000 60,000 50,000 40,000

Dimensions ds Hs cm m 2.0 3.0 2.5 3.5 2.8 3.5 2.8 3.5 2.8 3.5 2.8 3.5 2.8 3.5 2.8 3.5 2.8 3.5

Volume Total Wood m3/ha 15 7 50 25 80 45 80 45 80 45 60 35 50 25 40 20 30 15 485 262 53.9 29.1 100.0 54.0 for 9 years Annual average %

Bark 8 25 35 35 35 25 25 20 15 223 24.8 46.0

Mass (o.d.) Wood Bark t/ha 2.1 3.6 7.5 11.25 13.5 15.75 13.5 15.75 13.5 15.75 10.5 11.25 7.5 11.25 6.0 9.00 4.5 6.75 78.6 100.3 8.7 11.2 43.9 56.1 1354 1216 150 135 52.7 47.3

Total 5.7 18.75 29.25 29.25 29.25 21.75 18.75 15.00 11.25 178.9 19.95 100.0 2570 285 100.0

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 431 Table 7: Two-year rotation in eight-year production process Plant spacing m

Rotation number

I II III IV Total for 8 years Annual average % 1.20 x 0.50

Plant number per ha, N 16.667 48.000 48.000 18.000

Dimensions ds Hs cm m 4.2 5.8 4.0 5.0 4.0 5.0 4.3 5.9

Energy (GJ)

Volume Total Wood m3/ha 48 28 114 66 114 66 51 28 327 188 41 24 100.0 57.5 for 8 years Annual average %

Bark 20 48 48 23 139 17 42.5

Mass (o.d.) Wood Bark t/ha 8.4 9.0 19.8 21.6 19.8 21.6 8.4 10.3 56.4 62.5 7.0 7.8 47.4 52.6 972 758 121.5 94.8 56.2 43.8

Total 17.4 41.4 41.4 18.7 118.9 14.8 100.0 1730 256.3 100.0

Table 8: Three-year rotation in nine-year production process Plant spacing m

Rotation number

I II III Total for 9 years Annual average % 1.80 x 0.80

Plant number per ha, N 6,944 13,400 6,500

Dimensions ds Hs cm m 7.4 11.0 6.2 9.6 7.0 9.0

Energy (GJ)

Volume Total Wood m3/ha 156 117 167 117 116 86 439 320 49 36 100 72.9 for 9 years Annual average %

Bark 39 50 30 119 13 27.1

Mass (o.d.) Wood t/ha 35.1 35.1 25.8 96.0 10.7 64.1 1654 184 71.9

Bark 17.6 22.5 13.5 53.6 6.0 35.9 650 72 28.1

Total 52.7 57.6 39.3 149.6 16.7 100 2304 256 100

Table 9: Four-year rotation in eight-year production process Plant spacing m

Rotation number

2.0 x I 2.0 II Total for 8 years Annual average % Energy (GJ)

Plant number per ha, N 2,500 4,850

Dimensions ds Hs cm m 12.8 14.4 8.6 11.7

Volume Total Wood m3/ha 192 169 189 155 381 324 47.6 40.5 100.0 85.0 for 8 years Annual average %

Bark 23 34 57 7.1 15.0

Mass (o.d.) Wood Bark t/ha 50.7 10.3 46.5 15.3 97.2 25.6 12.1 3.2 79.1 20.9 1675 310 209 39 84.3 15.7

Total 61.0 61.8 122.8 15.3 100.0 1985 248 100,0

432

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al. Table 10: Five-year rotation in ten-year production process

Plant spacing m

Rotation number

1,70 x I 1.40 II Total for 10 years Annual average % Energy (GJ)

Plant number per ha, N 4,170 7,506

Dimensions ds Hs cm m 10.3 14.0 8.7 12.3

Volume Total Wood m3/ha 236 204 246 202 482 406 48.2 40.6 100.0 84.2 for 10 years Annual average %

Bark 32 44 76 7.6 15.2

Mass (o.d.) Wood Bark t/ha 61.2 14.4 60.6 19.8 121.8 34.2 12.2 3.4 78.1 21.9 2099 414 209.9 41.4 83.5 16.5

Total 75.6 80.4 156.0 15.6 100.0 2513 251.3 100.0

Under planting density of 40,000 plants per hectare (Table 6) production cycle is one year, attaining small dimensions: diameter 2-3 cm and height 3.0-3.5 m, i.e. volume 15-80 m3/ha (average 55 m3/ha) depending on production cycle. Such plantations produce high percentage of bark, almost 50% of the volume and above 50% of the mass, so that it is possible to produce averagely 285 GJ per hectare annually. In the plantations established with density 1.20 x 0.50 m, i.e. 16,670 plants per hectare (Table 7) production process (production cycle) was 2 years, and in the second and third production cycles, the number of trees increased up to 50,000 trees per hectare. Such plantations attain somewhat larger sizes; diameters 4.0-4.5 cm, heights 5.0-6.0 m, volume 2555m3/ha (average 40 m3/ha) annually. In such plantations, bark percentage is somewhat lower than in the plantations of the maximal density, and the average 216 GJ can be produced per hectare annually. Plantations established with plant spacing 1.80 x 0.80 m to 2.0 x 2.0 m, i.e. 2,500-7,000 plants per hectare (Tables 8-10) produce somewhat greater tree sizes in the production cycles of 3-4 years. In a 3-year rotation, mean diameters reach 6.0-7.5 cm, heights 9-11 m, volume 40-60 m3/ha (average 50 m3/ha) annually. In a 4-year rotation, mean diameters attain 8.5-13.0 cm, heights 12.0-14.5 m, and volume 50 m3/ha annually. Energy production in these plantations is very similar and amounts to about 250 GJ per hectare annually. The greatest production is achieved in the plantations with one-year rotations, with 2,570 GJ per ha in 9 years, and the lowest production is in the plantations with four rotations, where for 8 years 1,730 GJ per ha is produced (Table 11). Regarding yield, the shorter the rotation cycle, the more important the stand density becomes. For very short rotations (3–4 years), high stand densities are required. With increased rotation length, competition in dense stands reduces growth vigour and the number of shoots decreases because self thinning occurs. If rotations are not too short (at least 5 years), spacing generally has only a slight effect on total yield but an important effect on yield composition. Average simulated yield is highest at a rotation cycle of 3 or 4 years. Increasing the rotation cycle to 5 or 6 years also reduces average yield. Similar results have been found in field studies [52]. Relative growth rate of trees is highest before canopy closure. Of course, if the spacing is adapted (5,000 or less plants ha-1) longer rotation cycles become relatively more productive, since canopy closure is postponed. The absolute yield of the systems

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 433 however decreases. These results confirm the practice of 3–4 year coppice culture result in the highest yield [52]. Table 11: Poplar biomass and energy production Rotation Product ion Dura dur., Numbe tion, years r year s Total production 1 IX 9 2 IV 8 3 III 9 4 II 8 5 II 10 Average production 1 IX 9 2 IV 8 3 III 9 4 II 8 5 II 10 Share of wood and bark (%) 1 IX 9 2 IV 8 3 III 9 4 II 8 5 II 10

Biomass production per ha Volume (m3) Wood Bark Total

Mass (o.d.) t Wood Bark

Total

Energy (GJ) Wood Bark

Total

262 188 320 324 406

223 139 119 57 76

485 327 439 381 482

78.6 56.4 96.0 97.2 121.8

100.3 62.5 53.6 25.6 34.2

178.9 118.9 149.6 122.8 156.0

1354 972 1654 1675 2099

1216 758 650 310 414

2570 1730 2304 1985 2513

29.1 23.5 35.6 40.5 40.6

24.8 17.4 13.2 7.1 7.6

53.9 40.9 48.8 47.6 48.2

8.7 7.0 10.7 12.1 12.2

11.1 7.8 6.0 3.2 3.4

19.8 14.8 16.7 15.3 15.6

150 121 184 209 210

135 95 72 39 41

285 216 256 248 251

54 57 73 85 84

46 43 27 15 16

100 100 100 100 100

44 47 64 79 78

56 53 36 21 22

100 100 100 100 100

53 56 72 84 84

47 44 28 16 16

100 100 100 100 100

One effect of the vigorous juvenile growth of poplars is that, in Central Europe, yields of about 100 m3 ha−1 can be attained within four years with hybrid poplars plantations. Within the genus Populus the black poplars (Aigeiros-section) have the most pronounced pioneer tree character and are therefore not adapted to dense stands and this makes them unsuitable for SRIC. Balsam poplars (Tacamahaca-section) show a peak main annual increment at the age between 4 and 10 years [26]. Maximization of the yield therefore requires rotation cycles of 5–7 years for balsam poplars and hybrids of balsam and black poplars. Some authors even recommend a minimum rotation length of 10 years for Populus in general [26]. The yields given in the literature for poplars in short rotation plantations differ considerably. While reported maximum yields lie between 20 and 35 o.d.tha−1 yr−1 mean annual increment [46,47], other publications report that it was in the range of 2–3 o.d.tha−1yr−1 [48,49,50]. These differences partly reflect the type of trials. Based on this information it can be estimated that average harvestable yields of poplars from SRIC in temperate regions of Central Europe and North America range between 10 and 12 o.d.tha−1 yr−1 [25,51,52,26,66]. Nevertheless, the wide range of reported yields indicates the potential to optimize productivity of short rotation coppices. Considering the fact that the greatest production was achieved in experimental plantaitons with one-year rotations in the production process of 9 years, as well as the fact that plantaion establishment and tending, felling, manipulation and preparation for combustion is far simpler and economical than the biomass resulting form other forms of

434

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

production, it can be concluded that this form of poplar plantation is the most profitable for the production of biomass intended for thermal energy.

3.3.Wood and Bark of Some Fast Growing Broadleaf Tree Species as the Source of Renewable Energy Raw Material In this part of study, the wood of Populus x euramericana cl.I-214, aged one, two and 12, Populus deltoides cl.PE 19/66, aged one, two and eight, Salix alba L. cl.378 aged one, two and 14, and Robinia pseudoacacia cl.R-54, aged one, two and ten was examined. After the selection of characteristic sample trees (three trees in each species and age) measured parameters of growth elements were determined and the trees were felled. Sample trees were chosen as average plants based on average diameter and height on the experimental plot.The weight of each tree was measured, separately for wood and bark and the proportion of bark calculated. Immediately after felling, samples discs (discs cut at breast height - 130cm) were taken to assess moisture content and wood and bark densities. The density was determined on the basis of oven dry weight per green volume of an individual disk segment.

3.3.1. Results and Discussion Based on growth elements of the selected trees, their oven dry mass was determined per age (Table 12). Mature trees in both clones are from plantations with 400 plants ha-1, while the plantation density of one-year old trees is 10,000 plants ha-1, and two-year old trees 4,444 plants ha-1. With age, tree size increases and the share of bark decreases and also average densities of wood and bark decrease. The differences in taxation elements of study clones are significant, so that the amounts of oven dry mass of one-year and two-year trees for the clone PE 19/66 are twice as high as those for the clone I-214 (Table 12). Oven dry mass of black locust stem is min. In older trees these differences are still higher, which is also contributed by the higher densities of the clone PE 19/66. Moisture content of almost all prepared samples is rather uniform, ranging between 8 and 12%. Ash content in the wood is significantly lower (0.47 to 1.16%) compared to bark (5.32 to 7.34%), as shown in Table 12. It can be inferred that the obtained values cannot affect to a higher degree the differences in thermal characteristics of the researched clones. Higher heating values of wood and bark and the calculated values for the whole tree based on the share of bark have been presented in Fig. 2,3. The calorific values of the bark of examined species (Fig.2) have the same tendency as calorific values of the wood, and the maximum values were for two year cuttings: clone I-214 22.076 MJ/kg, and black locust 20.114 MJ/kg. Calorific value of bark of 12 year old poplar clone I-214 (17.473 MJ/kg) is somewhat higher when compared with the value in Danon et al., 16.065 MJ/kg for the same clone, 15 year old [54], and similar with value of black poplar bark (P. nigra) which is 17.260 MJ/kg [55]. Lignin content of the bark of the clone I-214 is about 24% [56] and this could indicate the higher calorific value, bearing in mind that higher heating value of lignin is about 25.000 MJ/kg [56].

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 435 Table 12: Taxation elements and stem mass by species

Species

Age year

Height m 2.6 5.6 22.4

Bark share % 19.0 18.0 12.0

Wood density kg/m3 338 336 320

Stem mass o.d.kg 0.31 1.26 86.75

Ash content, % wood bark 1.16 0.73 0.82

6.84 5.56 5.32

Poplar I-214

1 2 12

DBH Cm 1.9 3.8 12.2

Poplar PE 19/66

1 2 8

2.5 4.4 26.2

3.3 5.6 23.7

18.6 18.2 11.0

403 402 388

0.62 2.53 206.26

1.13 0.73 0.47

5.95 5.95 6.26

Willow cl.378

1 2 14

1.0 2.4 18.8

3.4 3.9 23.2

26.7 16.7 15.2

402 381 377

0.11 0.75 122.11

0.67 0.89 0.52

4.77 4.92 5.94

Black Locust cl.R-54

1 2 10

0.7 1.8 8.4

1.5 2.6 14.4

38.5 20.0 16.8

580 578 576

0.035 0.38 45.97

0.88 0.70 0.52

7.34 6.64 5.94

older trees

two year

one year

0

5

10

15

20

Higher heating v alue, MJ/kg

clone I-214

PE 19/66

willow

black locust

Figure 2. Higher heating values of bark of examined clones

The higher heating values calculated for whole stem (with corresponding share of bark) were lower than for wood . The highest calorific value of whole stem (Fig.3), referred from two year old trees (24.275 MJ/kg for cl.I-214; 23.392 MJ/kg for black locust; 22.572 MJ/kg for willow; and 20.817 MJ/kg for cl. PE 19/66). This is due to the higher proportion of bark and juvenile wood with high lignin content. The minimum values were measured for poplar clone I-214 (one year), and for willow mature wood (14 year), 15.787 MJ/kg and 16.169 MJ/kg respectively. Ciria et al., [57] reported similar heating values of 3-5 years old SRIC poplar wood (stem and branches) 18.1 – 18.3 MJ/kg, and Ugrenovic [58] for willow wood 17.849 MJ/kg. Benetka et al., [64] for 1-3 years old poplar clones (wood at breast height and

436

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

basal part, and branches) reported heating values from 18.60 MJ/kg to 19.27 MJ/kg. Higher heating values for black locust (20.784 MJ/kg to 23.392 MJ/kg) are some higher than cited in references: 19.578 MJ/kg [59]; 19.766 MJ/kg [60]; 18.858 MJ/kg [61] for seven year old trees.

older trees two year one year 0

5

10

15

20

25

Higher heating value, MJ/kg

clone I-214

PE 19/66

willow

black locust

Figure 3. Higher heating values of whole stem of examined clones

Although the wood of studied clones characterized by different ash contents, basic density, and share of bark, the ranging was changed when FVI indexes were calculated (Fig.4). Calculated FVI for the bark is influenced by high ash content and basic density. All FVI values are similar and low, from 1,217 to 2,832 (Fig.4). The higher FVI index of whole stem was for black locust (16,832), than for cl.PE 19/66 eight year old (16,047). Two year old cutting of the same clone (11,615), two year old wood of the cl.I-214 (10,430), mature willow wood (10,882), and one year old black locust (10,822) have the similar values. The minimum value of FVI index was for one year cutting of cl. I-214 (3,794), which characterized by higher values of ash content (12.16%), and moisture content (10.30%), and lower basic density (338 kg m-3). High ash content of wood is less desirable for fuelwood as it noncombustible and reduces the heat of combustion. The results of calculated FVI indexes indicated that higher value of wood density can contribute to the total heating value of wood combustion. Average basic wood density of twoyear old wood of cl. PE19/66 was 402 kg m-3, which is about 20% higher than for cl. I-214 (336 kg m-3). Wood density can be used as a useful parameter to fix harvest rotation cycles, particularly for short rotation plantations. However, decisions would be specific for each tree species on a given site. Both biomass production and quality are important considerations to fix the optimum age for fast growing energy plantations (poplars and willows).These results indicate the possibility of the energy production from whole very young poplar and willow trees from short rotation plantations, by chipping together with branches and bark.

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 437

3.3.2.Conclusion Poplars and willows, as the most represented species grown very successfully in short rotation plantations, as well as black locust, can be a significant source of thermal energy, being a relatively quickly renewable energy raw material. Higher heating value of wood was researched on the clones Populus x euramericana (cl.I-214) one, two and 12 years; Populus deltoides (cl. PE 19/66) one, two and 8 years; Salix alba (cl. 378) one, two and 14 years; Robinia pseudoacacia (cl. R-54) one, two and 10 years old respectively. As the share of bark depends on the age of wood, calorific values were determined separately for bark and for wood. Based on the share of bark, calorific value was assessed for individual trees of the analyzed clones. Average higher heating value for poplar stem is about 19.60 MJ/kg (cl.I214) and 18.90 MJ/kg (PE 19/66);

stem older tree

stem two year

stem one year

bark older tree

bark two year

bark one year

0

2500

5000

7500

10000

12500

15000

17500

FVI index

clone I-214

PE 19/66

willow

black locust

Figure 4: Values of FVI of bark and whole stem of examined clones

for calorific value of whole stem were for two year old trees (24.28 MJ/kg for cl.I-214, 23.39 MJ/kg for black locust, 22.57 MJ/kg for willow, and 20.82 MJ/kg for cl. PE 19/66). This is due to the higher proportion of bark and juvenile wood with high lignin content The bark of all species has a lower calorific value than wood. Primarily, it should be noted that wood of greater density has a higher calorific value. If we consider a tree as a whole, these differences are lower due to lower deviations of bark density values compared to wood, depending on the clone. It should be pointed out that the calorific value of wood is more favorable than that of bark. Consequently, higher densities of wood and bark, as well as lower moisture and ash contents, have a positive effect on heating value. Density is primarily characterized by the species of wood, then by site, climatic conditions and increment, as well as by planting density.

438

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

3.3.Short Rotation and High Plant Density Poplar Plantations for Energy Production 3.4.1 Results and Discussion Experimental field plantations were established in experiment estate “Kacka suma”. In the field trial 6 poplar clones (P.x euramericana cl. Ostia, P.nigra cl.53/86, P deltoides cl. PE 19/66, P.x .euramericana cl.I-214, P.x euramericana cl. S6-7, P.x euramericana cv. Robusta), with two different plant densities (38,461 plants ha-1 and 83,333 plants ha-1) were tested. Trees were planted using 25cm long hardwood cuttings obtained from the Institute of Lowland Forestry and Environment collection. After first growing season diameter at breast height and plant height were measured on the selected samples in field tests. Immediately after felling, the mass of freshly cut trees was measured and the bark was measured after barking in the green state. The specimens were taken for moisture content measurement. After the biomass drying in the laboratory, it was kiln dried and its oven-dry weight was measured . The results (Table 13) show significant differences in tree diameters and heights, depending on planting density. The consequences of denser plantings are significantly lower diameters, especially cl. Ostia (drop for about 64%) and Robusta (for about 48%). Other clones range in the interval between 16% for the cl. I-214 (min) and 37% (cl. 53/86 and S67). The changes of plant height are not so prominent, and the maximal values are attained by cl. S6-7 and Robusta. The changed tree sizes, which are the consequence of significantly greater planting density, result also in a significantly lower biomass yield. Based on the weights of measured plants, biomass ranges up to 85% (cl. Ostia and Robusta), i.e. more than 70% for cl. 53/86 and S6-7. The minimal value (cl. I-214) amounts to only a half of the biomass weight reached in the lower-density plantation. Table 13: Average tree parameters and oven dry weights of stem and bark Stem dimensions Diameter,cm Plant density 38,461 plants ha-1 Ostia 2.8 53/86 1.9 PE19/66 2.5 I-214 1.9 S6-7 2.5 Robusta 2.1 Plant density 83,333 plants ha-1 Ostia 1.0 53/86 1.2 PE19/66 1.8 I-214 1.6 S6-7 1.6 Robusta 1.1 Clone

Height, m

Average weight, o.d. kg Stem with bark Bark

Wood

2.95 3.30 3.30 2.60 3.65 3.50

0.590 0.433 0.620 0.310 0.748 0.625

0.049 0.097 0.120 0.054 0.129 0.088

0.541 0.336 0.500 0.256 0.619 0.537

2.40 2.90 3.27 2.70 2.50 2.30

0.083 0.105 0.260 0.156 0.157 0.113

0.028 0.031 0.068 0.048 0.043 0.036

0.055 0.074 0.192 0.108 0.114 0.077

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 439 Biomass yield per unit area depending on the plantation density was calculated based on the weight of test trees and the number of plants. The quantity of bark per hectare of plantation (Table 14) was calculated based on the bark percentage. Due to the fact that this study deals with the biomass of very young trees, practically one-year-old seedlings in which bark percentage is very high, and because of great differences in diameters of the study mean trees, bark weight per unit area is presented separately, disregarding the fact that the bark is not removed from so young plants, i.e. the trees are not barked before chipping. However, as the bark has a relatively high calorific value, it is significant to present the percentage of bark in the total energy released by biomass combustion. The study results of biomass yield after the first year show (Fig.5) that the increase of planting density has not the same effect on all the study clones. Namely, cl. I-214 shows the rise of biomass yield for about 8%, i.e. if only the bark yield is taken into account, it is the increase of more than 90%. Biomass yield of the clone PE 19/66 has a downward tendency for about 9% (higher yield of bark by about 23%). The clones Ostia and Robusta are significantly behind, because their yield is lower by 60%. Table 14. Biomass production after first growing season Biomass yield, o.d. t ha-1 Wood Bark Plant density 38,461 plants ha-1 Ostia 20.807 1.885 53/86 12.923 3.731 PE 19/66 19.231 4.615 I-214 9.846 2.077 S6-7 23.808 4.961 Robusta 20.653 3.385 Plant density 83,333 plants ha-1 Ostia 4.583 2.333 53/86 6.167 2.583 PE 19/66 15.999 5.667 I-214 8.999 3.999 S6-7 9.450 3.583 Robusta 6.417 2.999 Clone

Stem with bark 22.692 16.654 23.846 11.923 28.769 24.038 6.917 8.750 21.667 12.999 13.083 9.417

Maximal values of biomass yield in the plantations with 38,461 plants ha-1 were attained by the clones S6-7 (28.769 t ha-1 year-1) and PE 19/66 (23.846 t ha-1 year-1). It should be noted that PE 19/66 had the maximal yield also in a denser plantation (21.667 t ha-1 year-1). Clone S6-7, with 13.083 t ha-1 year-1, is the second by the yield in a denser plantation, although this is only cca 55% of its yield attained in the thinner plantation. In spite of the higher plant density, the biomass production figure is generally in accordance with other studies reporting biomass production of 10 to 12 o.d.t ha-1 year-1 [16]; the one-year-old shoots of willow clones (52,500 plants ha-1) also produced about 12 o.d.t ha-1 year-1 wood [10]. Jiranek and Weger [62] report that natural clones grow slower than the hybrids, and in good natural conditions annual yield of best poplar clones is expected to be over 15 t ha-1 year-1 of dry biomass. The yield after first year (18,000 plants ha-1) ranges from 2.2 to 3.6 o.d.t ha-1 for poplar clones and 2 to 2.5 o.d. t ha-1 year-1 for willow clones [13].

440

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

After the first of four years rotation cycle in medium density poplar plantations (10,000 plants ha-1) mean annual increment was 10 to 12 o.d.t ha-1 [63]. Riddel-Black et al., [17] report that yield of six poplar clones (16,500 plants ha-1) after first growing season was 4.88 to 9.54 o.d. t ha-1.

Biomass production, o.d. t/ha

35 30 25 20 15 10 5 0 stem,38461 plants ostia

53/86

bark,38461 plants 19/66

stem,83333 plants I-214

bark,83333 plants

S6-7

robusta

Figure 5. Biomass production after the first growing season

The greatest production of 11.25 o.d. t ha-1 annually can be achieved in experimental plantations with one-year rotations in the production process of 9 years with 40,000 plants ha1 [64], as well as the fact that plantation establishment and tending, felling, manipulation and preparation for combustion is far simpler and economical than biomass resulting from other forms of production. To be able to assess the amount of energy obtained from the unit area in two study planting densities, by plantation clear cutting after one-year rotation, the calorific values – higher heating values, were determined for wood and bark specimens of the study clones, clones I-214 and PE 19/66 [44, 64] were used. The study values are presented in Table 15. Table 15: Average calorific values of wood and bark Clone Ostia 53/86 PE 19/66 I-214 S6-7 Robusta

Higher heating value, MJ/kg Wood Bark 17.131 19.808 18.747 16.757 17.420 15.539 15.680 16.245 21.145 17.685 19.698 19.084

Stem with bark 17.583 18.293 17.070 15.787 20.505 19.559

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 441 The analysis of the study poplar wood and bark calorific values shows that the calorific value ranges within the interval from 15.68 MJ/kg (min) for the clone I-214, to 21.14 MJ/kg (max) for the clone S6-7. The bark calorific values have a narrower range, between 15.54 MJ/kg and 19.81 MJ/kg and they also have both positive and negative deviations from the respective wood calorific values. The values calculated for unbarked wood show that cl. I-214 (15.79 MJ/kg) has the min value and that the max value was recorded for clone S6-7 – 20.50 MJ/kg. The amount of energy that could be produced by the combustion of wood of the study clones was assessed based on the number of trees per unit area and the mass of mean trees of each individual clone, separately for wood and bark, and for the whole tree, based on bark percentage (Table 16). Table 16: Calculated amounts of energy per unit area of plantation Energy, GJ/ha Wood Plant density 38,461 plants ha-1 Ostia 360.169 53/86 242.267 19/66 335.004 I-214 154.385 S6-7 503.420 Robusta 406.823 Plant density 83,333 plants ha-1 Ostia 78.511 53/86 115.613 19/66 278.703 I-214 141.104 S6-7 199.820 Robusta 126.402 Clone

Bark

Stem with bark

37.338 62.520 71.712 33.741 87.735 64.599

398.993 315.144 407.051 188.728 589.908 470.159

46.212 43.283 94.962 64.964 63.365 57.233

121.622 160.064 396.354 205.215 268.267 184.187

The calculated amounts of energy show that there is a similar tendency as in the calculation of biomass yield. Max value is recorded for the clone S6-7 (589.908 GJ/ha) in the plantation with 38,641 plants ha-1. The minimal amount of energy is produced by cl. I-214 (188.728 GJ/ha). Robusta and PE19/66 have the advantage over the other study clones in the plantations with a lower number of trees. It is interesting that the energy obtained by the biomass combustion of clone PE19/66 is similar also in a denser plantation (396.354 GJ/ha) and that the drop is only 3%, which is minor compared to the drop of almost 70% for Ostia or 61% for Robusta. Clone I-214 showed a slight increase (cca 8%) in the denser planting, which is explained by insignificant changes of biomass yield.

3.4.2 Conclusion The analysis of results obtained by measuring and computing the yield of biomass (and energy) in two field tests with different planting densities, i.e. in the tests with a great number of plants per unit area, shows that in such studies it is necessary to know the characteristics of individual clones. The reaction of the clones to increased planting density is very different. Evidently, the clones I-214 and PE 19/66 are the least susceptible to higher planting density.

442

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

Of course, due to its significantly higher basic density, clone PE 19/66 is much more interesting, because its biomass yield is in both cases significantly higher.

4. CONCLUSION The Institute of Lowland Forestry and Environment in Novi Sad has established a program of long-term research of poplar anatomic properties and physiological processes to accelerate the procedure of producing fast-growing poplars and to conduct selection in the earliest possible phases. The study results show a very strong genetic variability of the analysed parameters, and statistically highly significant difference among the clones, which leads to a conclusions that they are under a high degree of genetic control. The highest number of stomata was measured in the Eastern cottonwood clones, which is in harmony with the leaf anatomy of these genotypes, because they had very thick photosynthetically active tissues and fine cells with numerous intercellulars. Simultaneously, Eastern cottonwood clones also had the highest photosynthesis intensity and leaf area. The results show that all the study parameters (number of stomata, net photosynthesis and leaf area) were highly correlated with the growth elements (diameter and height) and biomass production. The possibility of biomass production was researched in one-year, two-year, three-year, four-year and five-year rotations, in plantations established by poplar rooted cuttings, roots and seedlings, regenerated by coppice vigor after felling. In this way, the production process during 8-10 years and from two to nine rotations lasting from one to five years, produces annually on the average between 40.9 and 53.9 m3, i.e. 14.8 to 19.8 tons (of oven-dry mass) of wood and bark per ha, which can provide (combustion of whole tree chips) energy from 216 to 285 GJ ha-1 year-1. Regarding yield, the shorter a rotation cycle, the more important the stand density becomes. For very short rotations (3–4 years), high stand densities are required. With increased rotation length, competition in dense stands reduces growth vigour and the number of shoots decreases because self thinning occurs. If rotations are not too short (at least 5 years), spacing generally has only a slight effect on total yield but an important effect on yield composition. Considering the fact that the greatest production was achieved in experimental plantations with one-year rotations in the production process of 9 years, as well as the fact that plantation establishment and tending, felling, manipulation and preparation for combustion is far simpler and economical than the biomass resulting form other forms of production, it can be concluded that this form of poplar plantation is the most profitable for the production of biomass intended for thermal energy. Higher heating value of wood was researched on the clones Populus x euramericana (cl.I-214) one, two and 12 years; Populus deltoides (cl. PE 19/66) one, two and 8 years; Salix alba (cl. 378) one, two and 14 years; Robinia pseudoacacia (cl. R-54) one, two and 10 years old respectively. Average higher heating value for poplar stem is about 19.6 MJ/kg (cl.I-214) and 18.9 MJ/kg (PE 19/66); for willow about 19 MJ/kg, and for black locust about 21.7 kJ/kg. It is interesting that the highest calorific value of whole stem were for two year old trees (24.28 MJ/kg for cl.I-214, 23.39 MJ/kg for black locust, 22.57 MJ/kg for willow, and 20.82 MJ/kg for cl. PE 19/66). This is due to the higher proportion of bark and juvenile wood with high lignin content The bark of all species has a lower calorific value than wood.

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 443 Consequently, higher densities of wood and bark, as well as lower moisture and ash contents, have a positive effect on heating value. Density is primarily characterized by the species of wood, then by site, climatic conditions and increment, as well as by planting density. The analysis of results obtained by measuring of the yield of biomass (and energy) in two field tests with different planting densities, i.e. in the tests with a great number of plants per unit area, shows that in such studies it is necessary to know the characteristics of individual clones. The study results of biomass yield after the first year show that the increase of planting density has not the same effect on all the study clones. Namely, cl. I-214 shows the rise of biomass yield for about 8%, i.e. if only the bark yield is taken into account, it is the increase of more than 90%. Biomass yield of the clone PE 19/66 has a downward tendency for about 9% (higher yield of bark for about 23%). The clones Ostia and Robusta are significantly behind, because their yield is lower for 60%. Maximal values of biomass yield in the plantations with 38,461 plants ha-1 were attained by the clones S6-7 (28.769 o.d.t ha-1 year-1) and PE 19/66 (23.846 o.d.t ha-1 year-1). Evidently, the clones I-214 and PE 19/66 are the least susceptible to higher planting density. Of course, due to its significantly higher basic density, clone PE 19/66 is much more interesting, because its biomass yield is in both cases significantly higher. The calculated amounts of energy show that there is a similar tendency as in the calculation of biomass yield. Max value is recorded for the Eastern cottonwood clone S6-7 (589.91 GJ/ha) in the plantation with 38,641 plants ha-1. The minimal amount of energy is produced by Euaramerican poplar clone cl. I-214 (188.73 GJ/ha). Robusta and PE19/66 have the advantage over the other study clones in the plantations with a lower number of trees. It is interesting that the energy obtained by the biomass combustion of clone PE19/66 is similar also in a denser plantation (396.35 GJ/ha) and that the drop is only 3%, which is minor compared to the drop of almost 70% for Ostia, or 61% for Robusta. Clone I-214 showed a slight increase (cca 8%) in the denser planting, which is explained by insignificant changes of biomass yield.

REFERENCES [1] [2]

[3] [4] [5]

[6] [7]

International Energy Agency (2002): Renewables in global energy supply. Paris, France, IEA Overend,R.P. (2002): Bioenergy production and environmental protection In: A Sayigh, (Edit.), Workshop Proceedings, World Renewable Energy Congress, Cologne, Germany. FAO. FAOSTAT-database 2002, http://ww.fao.org, 2002. FAO. State of the World's Forests—2001, www.fao.org, 2001. European Commission. Energy for the future, renewable sources of energy.(1997): White Paper for a Community Strategy and Action Plan Communication from the commission, FOR/086/97, Brussels. Parikka, M.(2004): Global biomass fuel resources. Biomass and Bioenergy 27(6):613620. Kuiper,L.C., Sikkema,R., Stolp,J.A.N. (1998): Establishment needs for short rotation forestry in the EU to meet the goals of the commission’s White Paper on renewable energy (November 1997). Biomass and Bioenergy 15 (6): 451-456.

444 [8]

[9]

[10] [11] [12]

[13] [14]

[15]

[16]

[17]

[18] [19]

[20]

[21] [22]

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al. Berndes, G., Hoogwijk, M., Van den Broek, R.(2003): The contribution of biomass in the future global energy system: A review of 17 studies. Biomass and Bioenergy 25: 128. Volk,T.A., Verwijst, T., Tharakan, P. J., Abrahamson, L. P., White, E. H. (2004): Are short-rotation woody crops sustainable? In: W.P.M. Van Swaaij, T. Fjallstrom, P. Helm, A. Grassi (Eds.) Second World Biomass Conference, Rome, Italy. ETA Florance and WIP Munich, vol.I: 34-39. Nilsson, S. (1996): Do we have enough forests? IUFRO Occasional Paper No.5. Brown, C. (1999): Global Forest Products Outlook Study, Thematic study of plantation. FAO Rome Italy. Whityeman, A., Brown, C. (1999): The potential role of forest plantations in meeting future demands for industrial wood products. International Forestry Review, vol.1, No.3. Hanson, E.A. (1991): Poplar woody biomass yields; a look to the future. Biomass and Bioenergy :1-7. Markovic, J., Roncevic, S., Pudar, Z.(1996): Effect of poplar planting density on short rotation production of biomass. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.1: 745-750. Markovic, J., Roncevic, S., Pudar, Z.(1996): Possibility of poplar biomass production as raw material for bioenergy production. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.1: 739-744. Pitscher, K.(1996): The development of a gasification plant utilizing short rotation coppice: Project ARBRE. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.2: 1096-1101. Riddel-Black, D.M., Rowlands, C., Snelson, A.(1996): Short rotation forest productivity using sewage sludge as a nutrient. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.1:103-108. Klasnja, B., Kopitovic, S., Orlovic, S. (2001):Wood of some poplar and willow clones as raw material for kraft pulp production. Drevarsky Vyskum 46 (4): 1-8 Klasnja, B., Orlovic. S., Galic,Z., Pilipovic, A., Markovic,M.(2002): Short rotation and high plant density poplar plantations for energy production in: W. Palz, J. Spitzer, K. Maniatis, K. Kwant, P. Helm, A. Grassi (Eds.) Biomass for Energy, Industry and Climate Protection, Amsterdam, The Netherlands, ETA Florance: 223-226. Interdisciplinary Regional Research – Hungary, Romania, Yugoslavia, Novi Sad, Yugoslavia, Klasnja, B., Roncevic, S., Andrasev, S., Galic,Z.(2002): Short rotation plantations of fast growing broadleaf tree species as the source of renewable energy raw material. 6th International Symposium on CD. Klasnja, B., Kopitovic, S., Orlovic, S.(2002): Wood and bark of some poplar and willow clones as fuelwood Biomass and Bioenergy 23: 427-432. Orlovic, S., Klasnja, B., Kovacevic,B., Pekec,S., Pilipovic,A.(2003): Selection of black poplars (Section Aigeiros Duby) for special purpose production of wood for pulp and paper.In: G. Baudin, J. Felegi, G. Gellerstedt, S. Katuscak, I. Pikuluk, J. Paris (Eds.).

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 445

[23]

[24]

[25]

[26]

[27]

[28] [29]

[30]

[31]

[32] [33] [34] [35]

[36]

Chemical Technology of Wood, Pulp and Paper. Bratislava, Slovak Republic, Slovak University of Technology, Bratislava: 309-314. Klasnja,B., Orlovic,S., Drekic,M., Markovic,M.(2003): Energy production from short rotation poplar plantations.7th International Symposium on Interdisciplinary Regional Research – Hungary, SerbiaandMontenegro, Romania, Hunedoara, Romania, CD: 353358. Orlovic,S., Klasnja,B., Ivanisević,P., Galic,Z., Radosavljevic,N.(2004): Selection of black poplar clones for biomass production. In: W.P.M. Van Swaaij, T. Fjallstrom, P. Helm, A. Grassi (Eds.) Second World Biomass Conference, Rome, Italy. ETA Florance and WIP Munich, vol.I: 434-437. Fischer,G., Prieler,S., Van Veltuizen,H. (2005): Biomass potentials of miscanthus, willow and poplar: results and policy implications for Eastern Europe, Northern and Central Asia. Biomass and Bioenergy 28 (2): 119-132. Kauter,D., Lewandowski I., Claupein,W. (2003): Quantity and quality of harvestable biomass from Populus short rotation coppice for solid fuel use—a review of the physiological basis and management influences . Biomass and Bioenergy 24 (6): 411427. FAO, The global outlook for future wood supply from forest plantations, 2000, Food and Agricultural Organisation, Forestry Policy and Planning Division. Working Paper GFPOS/WP/03: Rome Italy. Hall, D.O., Scrase, J.I. (1998): Will biomass be the environmentally friendly fuel of the future; Biomass and Bioenergy 15: 357-367. Sims, R.E.H. (2004): Bioenergy – its global potential in future decades. In: W.P.M. Van Swaaij, T. Fjallstrom, P. Helm, A. Grassi (Eds.) Second World Biomass Conference, Rome, Italy. ETA Florance and WIP Munich, vol.I: 96- 102. Bungart, R., Huettl,R.F.(2000): Biomass Production, Water Budget and Nutrition of Fast-Growing Tree Species on a Mine Spoil in the Lusatian Mining Region, in: W. Palz, J. Spitzer, K. Maniatis, K. Kwant, P. Helm, A. Grassi (Eds.) Biomass for Energy, Industry and Climate Protection, Amsterdam, The Netherlands, ETA Florance:: 170172. Holger Gruenewald, B., Brandt,V., Bens,O., Uwe Schneider, B., Huettl,R.F.(2004): Production of biomass for energy transformation purposes in agroforestry systems. In: W.P.M. Van Swaaij, T. Fjallstrom, P. Helm, A. Grassi (Eds.) Second World Biomass Conference, Rome, Italy. ETA Florance and WIP Munich, vol.I:254-257. TAPPI. TAPPI test method. (1999): The American Paper and Pulp industries Institute (TAPPI) Press, Norcross, GA, USA, 1999. Murphey, W.K., Cutter, B.E.(1974): Gross heat of combustion of five hardwood species at differing moisture contents. Forest Products Journal 24 (2): 44-45. Goel, V., Behl, H.M.(1996): Fuelwood quality of promising tree species for alkaline soil sites in relation to tree age. Biomass and Bioenergy 10 (1): 57-61. Orlović, S., Guzina, V., Krstić, B., Merkulov, LJ. (1998): Genetic variability in anatomical, physiological and growth characteristics of hybrid poplar (Populus x euramericana Dode (Guinier)) and eastern cottonwood (Populus deltoides Bartr.) clones. Silvae Genetica 47 (4): 183-190 Ceulemans, R.,Impens,I., Imler,R. (1988): Stomatal conductance and stomatal behavior in Populus clones and hybrids. Canadian Journal of Botany 66: 404-414.

446

Bojana Klasnja, Sasa Orlovic, Zoran Galic et al.

[37] Orlovic,S., Djokovic,R. (1991): Variabilty of number and dimensioms of stomata of some poplar clones of section Aigeiros Duby. Works, Poplar Research Institute 23: 4552 (in Serbian). [38] Orlovic,S., Guzina,V. (1994):Variability of stomata number on the leaves of Black poplars and their hybrids. Matica srpska - Proceedings for natural sciences 86: 63-68. [39] Orlovic,S., Merkulov,Lj., Guzina,V. (1994): Variability of elements of poplar leaf anatomic structure. Matica srpska - Proceedings for natural sciences 87: 65-72. [40] Ceulemans,R.,Impens,I. (1983): Photosynthetic, morphological, and biochemical gas exchange characteristics in relation to growth of young cuttings of Populus clones. Advances in photosynthesis research.In: Prooceedings of the VI th International Congress on Photosynthesis, Brussels, Belgium, vol. IV: 141-144. [41] Isebrands, J.G., Ceulemans,R.,Wiard,B. (1988): Genetic variation in photosynthetic traits among Populus clones in relation to yield. Plant Physiology and Biochemistry 26 (4): 427-437. [42] Ceulemans,R., Hinckley, T.M.,Impens,I. (1989): Stomatal response of hybrid poplar to incident light, sudden darkening leaf excision. Physiologia Plantarum 75: 174-182. [43] Orlovic,S., Pajevic,S., Krstic,B. (1995): Possibility of utilizing some morphological and physiological parameters in poplar breeding. IUFRO XX World Congress, Poster abstracts: 70. [44] Markovic, J., Roncevic, S., Pudar, Z. (1996): Possibility of poplar biomass production as raw material for bioenergy production. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol. I: 739-744. [45] Markovic, J. (1975): Wood fiber dimensions and wood basic densities depending of breedeng technology, age and stem height of cl. I-214. Topola (Poplar)103/106: 135150 (in Serbian). [46] Beale,C.V., Heywood, M.J. (1997): Productivity of commercial crops of short rotation coppice at six sites in Southern England. Aspects of Applied Biology 49: 181–390. [47] Harrington, C.A., Radwan, M.A., DeBell, D.S. (1997): Leaf characteristics reflect growth rates of 2-year-old Populus trees. Canadian Journal of Forest Research 27: 1321–1325. [48] Pontailler,J.Y., Ceulemans,R., Guittet,C., Mau,P. (1997): Linear and non-linear functions of volume index to estimate woody biomass in high density young poplar stands. Annales de Science Forestiere 54: 335–345. [49] Hartmann,H. (2001): Brennstoffzusammensetzung und eigenschaften. In: M. Kaltschmitt and H. Hartmann, Editors, Energie aus Biomasse—Grundlagen, Techniken und Verfahren, Springer, Berlin : 248–272. [50] Scarascia-Mugnozza, G.E., Ceulemans,R., Heilman, P.H., Isebrands, J.G., Stettler, R.F., Hinckley, T.M. (1997): Production physiology and morphology of Populus species and their hybrids grown under short rotation. II. Biomass components and harvest index of hybrid and parental species clones. Canadian Journal of Forest Research 27: 285–294. [51] Deckmyn,G., Laureysens,I., Garcia,J., Muys,B., Ceulemans,R. (2004): Poplar growth and yield in short rotation coppice: model simulations using the process model SECRETS.Biomass and Bioenergy 26 (3):221-227. [52] Herve,C., Ceulemans,R.(1996): Short-rotation coppiced vs. non-coppiced poplar: a comparative study at two different field sites. Biomass and Bioenergy 11: 139–150.

Poplar Biomass of Short Rotation Plantations as Renewable Energy Raw Material 447 [53] Senelwa,K., Sims, R.E.H.(1999): Fuel characteristics of short rotation forest biomass. Biomass and Bioenergy 17: 127–140. [54] Danon, G., Stevanovic Janezic, T., Bujanovic, B., Stanojevic, G.(1996): Short-rotation poplar bark utilization for the production of light briquettes. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.2: 942-947. [55] Ninic, N., Oka, S.(1992): Using biomass for energy production, Jugoslovensko drustvo termicara, Naucna knjiga Beograd,Yugoslavia (in Serbian). [56] Stevanovic Janezic, T., Kolin, B., Jaic, M., Danon, G.(1995): Enhancement of wood technologies in correlation with properties of wood chemical constituents, Faculty of forestry, Belgrade, Yugoslavia (in Serbian). [57] Ciria, M.P., Gonzales, E., Mazon, P., Carrasco, J.(1996): Influence of the rotation age and plant density on the composition and quality of poplar biomass. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.2: 968-973. [58] Ugrenovic,I. (1950): Wood technology, 254, Nakladni zavod Hrvatske, Zagreb, Yugoslavia (in Croatien). [59] Stringer,J.W. (1992): Wood properties of black locust (Robinia pseudoacacaia): Physical, mechanical, and quantitative chemical variability on black locust. Proceedings, Int.Conf.: 277. [60] Panshin, A.J., Zeeuw, C.(1970): Textbook of wood technology, 3rd ed. Mc Graw Hill Inc New York, 722. [61] Geyer,W.A., Walawender,W.P.(1994): Biomass properties and gasification behavior of young black locust. Wood and Fiber Science 26 (3): 354-359. [62] Jiranek, J., Weger, J.(1998): The potential and utilisation of biomass in the Czech Republic.In: H. Kopetz, T. Weber, W. Palz, P. Chartier, G.L. Ferrero (Eds.), Biomass for Energy and Industry, Wurzburg, Germany, C.A.R.M.E.N. 1002-1005. [63] Kuiper, L.C., Kolster, H.W.(1996): Twenty years of research on poplar biomass production in the Netherlands. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.1: 96-102. [64] Klasnja, B., Kopitovic, S.(1996): Basic thermal characteristics of poplar wood in direct combustion process. In: P.Chartier, G.L. Ferrero, U.M. Henius, S. Hultberg, J.Sachau, M.Winblad (Eds.). Biomass for Energy and the Environment, Copenhagen, Denmark. Pergamon, vol.2: 974-979. [65] Benetka,V., Bartakova,I., Mottl,J. (2002): Productivity of Populus nigra L., ssp.nigra under short-rotation culture in marginal areas. Biomass and Bioenergy 23 (5):327-336. [66] Scholz,V., Ellerbrock,R. (2002): The growth productivity, and environmental impact of the cultivation of energy crops on sandy soil in Germany. Biomass and Bioenergy 23 (2):81-92.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 449-480

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 16

BIOBASED POLYMERS BY CHEMICAL VALORIZATION OF BIOMASS COMPONENTS* B. Kamm†,1, M. Kamm2, I. Scherze3, G. Muschiolik3 and U. Bindrich4 1

Research Institute of Bioactive Polymer Systems e.V. Research Center Teltow-Seehof Kantstraße 55 D-14513 Teltow, Germany 2 Biorefinery.de GmbH, Potsdam, Germany 3 FS-University Jena, Department of Food Technology 4. DIL (Deutsches Institut für Lebensmitteltechnik) e.V., Quakenbrück

ABSTRACT Plants represent a natural chemical and polymer factory and food plant. Biorefineries combines necessary technologies between biogenic raw material and intermediates and final products. The paper present two strategies for producing of polymeric materials, firstly the utilization of the pre-determined natural macromolecular structure and secondly the using of biogenic building blocks. The first step is the fractionation technology from green biomass for producing of fiber-rich press cake and a nutrient richgreen juice. The main focus is directed on products, such as proteins, polylactic acid, cellulose and levulinic acid- sequence products and their application as well as their market.

Keywords: green biomass, biorefinery, proteins, poly(lactic acid), cellulose, levulinic acid

*

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Corresponding author: Research Institute of Bioactive Polymer Systems e.V. Research Center Teltow-Seehof Kantstraße 55 D-14513 Teltow, Germany; e-mail: [email protected]; Tel.:0049-3328-332210; Fax: 0049-3328332211.

450

B. Kamm, M. Kamm, I. Scherze et al.

ABBREVIATIONS BR BJ DM dt e.g. GBR GJ GLNC ha kg LNC LP LPC PC pound (eng) SJ SPC t t/y

Biorefinery Brown Juice Dry matter deciton, decimal metric tonne (1dt = 100 kg) (for example) Green biorefinery Green Juice Green leaf nutrient concentrate hectare (10.000 m2 =100m x 100m) Kilogramme (1kg = 1000 gramme) Leaf nutrient concentrate‘ (LNC) Leaf protein Leaf proteine concentrate Press cake 1 pound (germ) = 1.1023 pound (eng) Silage Juice Silage press cake ton, decimal metric tons (1t = 1000 kg) ton(s) per year

INTRODUCTION Sustainable economical growth requires safe resources of raw materials for the industrial production. Today’s most frequently used industrial raw material, petroleum, is neither sustainable, because limited, nor environmentally friendly. While the economy of energy can be based on various alternative raw materials, such as wind, sun, water, biomass, as well as nuclear fission and fusion, the economy of substances is fundamentally depending on biomass, in particular biomass of plants. Special requirements are placed to both, the substantial converting industry as well as research and development regarding the efficiency of the product line as well as sustainability. “The development of biorefineries represents the key for the access to an integrated production of food, feed, chemicals, materials, goods, and fuels of the future” (National Research Council, 2000). Many of the currently used industrially made biobased products are results of a directly physical or chemical treatment and processing of biomass, such as cellulose, starch, oil, protein, lignin and terpenes. By biotechnological processes and methods feedstock chemicals are produced such as ethanol, butanol, acetone, lactic acid and itaconic acid as well as amino acids, e.g. glutamic acid, lysine, tryptophane. On the other side, currently only 6 billion tons of the yearly by photosynthesis produced 170 billion tons biomass are used; in addition, only 3 percent of these in the non-food area, such as chemistry (Zoebelin, 2001). The today’s product lines in the chemical industry produce a few basic chemicals from petrochemical raw material which represent the basis for the synthesis of a wide product palette for nearly all life areas.

Biobased Polymers by Chemical Valorization of Biomass Components

451

The development of comparable biorefineries – however not in the sense of copy – is necessary to produce a broad variety of biobased products in an efficient construction set system. Each biorefinery refines and converts its corresponding biological raw materials into a multitude of valuable products. The product palette of a biorefinery includes not only such products also producable in a petroleum refinery, but in particular such products, which are not accessable in petroleum refineries. Therefore, it is necessary to develop new biorefinery basis technologies, such as (1) (LCF)-Lignocellulosic Feddstock Biorefinery, (LCF)-pretreatment and effective separation into lignin, cellulose and hemicellulose, (2) further development of thermal, chemical and mechanical processes, such as extractive methods, gasification (syngas) and liquefaction of biomass, (3) further development of biological processes (biosynthesis, bacteria for degradation of starch and cellulose, etc, (4) combination of substantial conversions, such as biotechnological and chemical processes; (5) cornbiorefinery-concepts, (6) green biorefinery-concepts (7) promotion of research and development into phase III-biorefinery: feedstock-mix + process-mix ⇐ product-mix, (Kamm and Kamm, 2004). (Figure 1). Therefore well-known technologies and methods have to be applied in a combinatory way. From today’s point of view there are two principle ways for the utilization of the synthesis power of the nature for the application area of degradable polymers based on biogenic raw materials:

Feedstock(s) biological raw material various, mixed

• Food and Feed Grains, • Ligno-cellulosic Biomass, (e.g. late grass, reed, bush, harvest rest)

• Forest Biomass, (e.g. wood, underwood, waste wood-processing)

• Municipal Solid Waste (MSW), (e.g. paper/cardboard, town-cleaning, hospitals)

Processing-Technologies various, combined

Products Substances and Energy various, multi product systems

• Bioprocesses ( bacterial, enzymatic a.o.), • Chemical Processes, • Thermo-chemical Processes, • Thermal Processes, • Physical Processes,

• Fuels, • Chemicals, • Materials ( e.g. Polymers ) • Specialities, • Commodities, Goods

Figure 1. Basic principles of a biorefinery (Type III biorefinery).

452

B. Kamm, M. Kamm, I. Scherze et al.

Strategy I. Polymeric Materials from Biogenic Macromolecular Structures First way is the utilization of the pre-determined natural macromolecular structure under extensive maintenance of this specific structure, if necessary also of parts of the complex plant morphology and direct modification of properties for particular applications. Included are the following classes of substances: (1) nucleic acids, (2) proteins, (3) polysaccharides, (a) poly(α)-glucoses ⎯ starch and dextrins, (b) poly(ß)-glucoses ⎯ cellulose, lignocellulose, (c) xylane ⎯ hemicellulose, (d) poly(galactoses) ⎯ pectins, (e) poly(mannoses) ⎯ alginates, (f) poly(ß-glucosamines) ⎯ chitins and (4) poly(hydroxy-fatty acids). These are hetero chain polymers in form of ester -, amid-(peptid-) and/ or glycosidstructures, which are hydrolytically degradable, that means acid- or base-catalyzed as well as enzymatically (Ebert, 1993). Penetration of water and thus degradability of polymers can be influenced by means of changing the physical structure, as (plastic) shaping and/or modification of the chemical structure, as increasing or decreasing of hydrophobicity as well as hydrophilicity of the corresponding polymers. In particular for functional applications of renewable raw materials, as fibre composite, starch-resultant products etc. this way is followed (Sixth Symposium on Renewable Resources and Fourth European Symposium on Industrial Crops and Products, 1999). For applications as basic chemical building blocks these polymers have limits due to their non-uniform structures depending on the respective quality of the nature-built batch.

Strategy II. Polymeric Materials from Biogenic Building Blocks Main requirement in the construction system of chemistry are uniformly structured compounds, which are converted via tailor-made syntheses into highly processed degradable structures (Verband der Chemischen Industrie, 1994). This is the guidance of the second principle of the synthesis of degradable structures: Biogenic raw materials can be degraded to well-defined uniform monomer structures by means of biotechnological or chemical methods. These building blocks can then be used for the synthesis of the target compounds. Ideally, the breakdown and build-up of the polymer structures are then combined. By means of chemical degradation of hexoses- as well as pentoses-containing raw materials well-defined structures such as levulinic acid (γ-oxocarbonic acid), hydroxymethylfurfural (HMF) or furfural are available. Currently applied monomers biotechnologically produced from hexosenic raw materials are (1) α-hydroxycarbonic acid, as lactic acid, malic acid, (2) olefinic carbonic acid, as fumaric acid, itaconic acid (3) polyvalent alcohols, as 2,3-butanediol, 1,3-propanediol, dihydroxyacetone, (4) α-aminocarbonic acid, as glutamic acid and lysine as well as (5) subsequent products such as carnitine (ßhydroxybetaine) (Kamm, 2004).

Biobased Polymers by Chemical Valorization of Biomass Components

453

2. THE GREEN BIOREFINERY 2.1. Principles Green biorefinery represents a complex system of ecological technologies for the comprehensive (holistic) substantial and energetic utilization of renewable raw and natural materials in form of green and waste biomass from a targeted sustainable regional land utilization. Such green biomass are for example grass from cultivation of permanent grass land, closure fields, nature preserves or green corps, such as lucerne, clover, immature cereals from extensive land cultivation. Thus, green plants represent a natural chemical factory and food plant. The careful wet fractionation technology is used as first step (primary refinery) to isolate the content-substances in their natural form. Thus, the green crop goods (or humid organic waste goods) are separated into a fiber-rich press cake (PC) and a nutrient-rich green juice (GJ). Beside cellulose and starch, the press cake contains valuable dyes and pigments, crude drugs and other organic substances. The green juice contains proteins, free amino acids, organic acids, dyes, enzymes, hormones, minerals, high-quality crude drugs and other organic substances. By the help of the bio-technology, the eco-technology, the ‘soft’ and ‘green’ chemistry, these valuable materials can be isolated in their natural form, or via mild conversion carefully be devoted to an economical utilization (Kamm et al., 2000). Green Crop Drying Plant Wet Fractionation

Energy

Green (Wet) Raw Material Grass, Lucerne, Alfalfa, Herb a.o.

Press

Power Station Heat, Electricity

Press Juice

Press Cake drying to Pellets + Bales Valuable Products

Biogas

Separation Fermentation Fermenter

Carbohydrate Sources Pre-Treatment

Enzymes

Decanter Lactic Acid + Derivatives Amino Acids Proteins

Flavourings Dyes Carbohydrates

Green Pellets for Fodder Pellets or Bales for Solid Fuel Raw Material for Syngas Raw Material for Hydrocarbons Raw Material for Biogas

Proteins

Enzymes Whole Crops, Straw, Seeds, Starch, Hydrolyzate, Molasse, a.o.

Fields

Organic Acids Ethanol

Figure 2. A System Green Biorefinery combined with a green crop drying plant.

Raw Material for Fibres + Fleece Raw Material for Chemicals e.g. Levulinic Acid

454

B. Kamm, M. Kamm, I. Scherze et al.

Concept of Havelland-Biorefinery, Selbelang, State of Brandenburg, Germany The activities in the field green biorefinery system grown more and more and developed into an independent line within the large field of biomass technologies. Beside raw material and technology aspects, this system is particularly characterized by the approach of consideration and attention of sustainability criteria (Hector et al.) and incorporation of technologies in regional living and business spaces (sustainable economy, sustainable agriculture, sustainable regional development). The term ‘Green Biorefinery’ is on the one hand-side used for model procedures, but on the other hand-side also for a entire program. To ‘refine’ is originally French (raffiner) and means ‘something to improve, to purify’. A refinery is per definition a technical facility for the purification, separation and refinement of materials and products. ‘Green’ in the field of plants means the simultaneousness of high concentration of chlorophyll, nutrients and water, ‘Bio’ is Greek (bios) and means ‘live’, something biological and natural. Programmatically, The Green Biorefinery stands for a technology (refinery), formed by the nature imitated (biologically) with the target to be careful, sustainable and ecologic. The Green Biorefinery pursues the following approaches. The Green Biorefinery represents • • •

a complex system of ecological technologies a model for the study of ecological process management finally, an economically self-consisting enterprise and economic entity, respectively.

Therefore, the green biorefinery can be defined as follows. The green biorefinery is:

•

•

•

a complex system of ecological technologies for the comprehensive material and energetically use and utilization of renewable raw and natural materials in form of green and waste biomass from a targeted sustainable regional land utilization. a model for the study of ecological process management, that means for the environmental friendly reorientation of the production and energy supply under the premise of sustainability. This model includes the following fields: − the supply of raw materials from sustainable, that means environmental and social friendly land utilization, − regional sustainable economic procedures based on modern stock-flow management , − the development of a value-material oriented agriculture − the step-wise replacement of material and energy management of fossil raw materials by technology transfer, − the introduction of ecological technologies and products into market and practice. an economic self-consisting business of complex technical facilities for the purification, separation and refination of renewable raw materials in form of green and waste biomass together with a self-supporting energy supply on the basis of

Biobased Polymers by Chemical Valorization of Biomass Components

455

renewable raw materials, e.g. bioenergy via biogas, an in-plant material circulation, in particular water and an ecological water and waste treatment.

2.2. Raw Materials A main raw material is the ‘green biomass’. This includes the large group of green plant materials: green grasses (meadows, willows, natural resources, extensive willow management), the wild fruit and crops lucerne (table 1) and clover as well as immature cereals and plant shoots. The green plant material contains complex natural and value materials in form of carbohydrates, proteins, fibres, fragrances, dyes, fats, hormones, amino acids, enzymes and others (Pirie, 1971; Carlsson 1989, Carlsson 1997). By primary production of photosynthesis in green plants more than 20 tons of dry matter and 3 tons of protein per ha in temperate climates can be obtained per year (Carlsson 1985). Table 1. Crude components of Medica sativa L. (lucerne alfalfa) Medica sativa L. alfalfa, lucerne, green plant ingredients Nitrogen free Crude fibre Crude Crude fat Crude ashes extractives** proteins *** components in 6-14.4 3.5-13.4 2.8-7.3 0.5-1.0 1.8 wt % yield /ha* 0.55-1.32 0.32-1.23 0.27-0.67 0.046-0.092 0.165 in tons [t] * Lucerne yield 9,3 t/ha/harvest (DM), DM-dry matter, harvest in july (Robowsky 1998), **Nitrogen free extractives: crude drugs, sugar, dyes and pigments, enzymes, vitamines, free acids a.o., ***Ashes: Ca, P, K, Mn, trace elements Fe, Zn, Cu, (Hagers handbook of pharmaceutical practice, 1972-78)

Table 2. Yields of sugar beet herbage/foliage per year in the district Havelland area under cultivation* yields of sugar beet herbage/foliage** 35-40 t/ha

1225 ha (1998) 42.875,0 -49.000,0 t/y 16,5% 7.074,0-8.085,0 t/y

dry matter substance (DM)** 15-18 ( 16,5) %/ha yields of sugar beet herbage/foliage 5,77-6,6 t/ha (DM) (DM) *district Havelland, Germany, (State of Brandenburg, Brandenburg, (2003), ** (Fechner and Hertwig, 2003) t/y = tons per year

456

B. Kamm, M. Kamm, I. Scherze et al.

•

•

•

The second large raw material source are the green harvesting residue materials from agricultural cultivated crops. In particular such vegetables, that are harvested with green foliages. This includes e. g. not insignificant amounts sugar beet leaves (sugar beet for sugar industry) (table 2), hemp scrapes and leaves (hemp for fibre production), residue from flax processing, residue of the fresh vegetable production. Further potential refinery raw materials are the less normed (standardized) juice-rich waste biomass. These should contain moisture and on the top listed natural and value- materials or also conversion grade. According to coupling effects of material and energetic use, the constitution can strongly vary. Such waste biomass are not standardized goods, but renewable natural waste good that has mainly to be waste managed. This can be residuals of plant production (mixed and ripe harvest residuals), potato juices, hydroxycarboxylic acid-rich wastes, as silage seepage, juices of the canned foods industry, remainder of the sugar industry, remainder of the animal production. The 4th large group are less normed (standardized) dried biomass and waste biomass. These often contain a high amount of plant cellulose and will therefore be supplied as raw material to press-cake-using production lines. This can be residual straw, hay and all kinds of dried foliage (e.g. maize hay). But also residuals of inplant waste paper and wood, e.g. for energy production or cardboard production. This group also includes modern concepts of dry crop fractionation, as immature cereals (Coombs and Hall, 1997).

It should be mentioned, that the transitions between raw material types will and should be fluid.

2.3. Primary Technologies The special feature of the green biorefinery is the wet fractionation or watery-fractionation of green biomass (Figure 3: way A). This is also called first fractionation step or primary refinery step. (This includes for example the harvest, fractionation, conservation and storage of the primary fraction). Here, fresh harvest and waste goods are treated. Thus, the plant compounds are mostly unadulterated; however, the green good should in any case immediately be re-worked. This process step, in generally by technical press produced a faser-rich quantity of water-unsoluble solid material, Press cake (PC) and a nutrients-rich Green juice (GJ) or Brown juice (BJ). The wet fractionation based on the soft separation of water soluble and water-unsoluble components of the green biomass.

Biobased Polymers by Chemical Valorization of Biomass Components

457

Green plant, Green crops grass, lucerne, beet herbage, green biomass etc.

Harvest, Green crops cutting, comminution plant

A

B

Wet fractionation • press

juice fraction Green juice (GJ), Brown juice (BJ)

solid matter fraction Press cake (PC)

C

Decomposition process • enzymatic (saccharification) • hydrolytic • enzymatic/fermentative • thermal (steam/ pressure)

Fractionation • solid matter separation

liquid fraction

Silage production • silage (before-fermentation)

Wet fractionation • press

juice fraction

solid matter fraction

solid matter fraction

Silage juice (SJ); [GJ, BJ)]

Silage press cake (S-PC)

biotechnological and chemical conversion • non-food- and food-products (chemistry, pharmacy, cosmetics a.o.)

practical uses and functional applications • food, fibres, proteines, raw paper a.o.

thermal conversion and fermentation gas • energy (heat, cold), energy products (current ), bio- and synthesis gas, pyrolyzate a.o.

Figure 3. Green Biorefinery ⎯ Primary refinery. Methods for fractionation of green crops.

The silage wet-fractionation is a form of the primary refinery technology (Figure 3, way C). The green goods are conserved by organic acids or fermentation processes before treating them in the following procedure. The treatment of silage green goods have any advantages (decentral raw material preparation, simple and low price conservation and storage, reasonable whole year operation of the processing-step and other more (Kromus et al. 2004). The use of the end products of silage is restricted because silage-chemicals attacks the cell walls and can modify substances. The so-called breaking up methods are the 3rdcategory of the primary refinery technology (Figure 3, way B). Breaking up methods consider mainly the humid or dry whole plant. The use procedures are working by enzymatic, fermentative, hydrolytic, chemical, thermal or thermal in combination with press methods. The strength (deepness of operation) of the breaking up methods is differently, and ranges from a low (enzymatic, fermentative) to a high level (chemical, hydrolytic). For every step a classification is needed to check if it belongs to the green biorefinery technology. A high single yield of products can be achieved if the complete plant breaking up methods takes place at the primary refinery step (e.g. by saccharification, increases the total amounts of sugar of the raw charge. But these procedure decrease the level of the utilization and product diversity. Nevertheless breaking up methods have been consider to be usable from technological and economic point of view. This is also

458

B. Kamm, M. Kamm, I. Scherze et al.

the case for the secondary product lines. Green biorefinery-system contradiction can be solved by a further development of new biotechnological breaking up methods. All primary fractionations following the secondary refinery steps contain processes for substantial and energetical utilization of fractionation products. The kind and number of the secondary fractionation steps are determinate by composition and energy potential of the input green biomass and waste biomass, the status of the technology as well as the market ability of potential products of refinery.

3. PRODUCTS FROM THE PRESS CAKE The kind and number of products of a green biorefinery is nearly unlimited, if the fractal character of the biosynthesis and biochemistry of the green plant materials is considered (Peitgen and Richter, 1986). The characterization of the plant by a new analysis method usually discovers in addition to the main product innumarable new products. Not all ingredients have been discovered and technologically gained from natural products even from plants with large trade importance as e.g. lucerne (Starke et al. 2000). For the green biorefinery the main products, beside products as well as charge-impurities are interesting. However the ecotechnology and the biotechnology are determinate by the cost-use-efficiency. The economic aspect reduces the diversity of products, also soft technologies are used to reduce the complex molecules of nature materials. Nevertheless the scientific branch of ecotechnology tries to develop new methoda, preferring a reducing of strengths (deepness of operation). This can be done by using e.g. biodiversity before molecule modification or by applying low injure degree methods etc. (Moser, 1997). Following products and group products are possible (technological; after fractionation according to variant A): From the solid matter fraction (press cake, PC) [see also Fig 2]

• • •

The application of the PC as feeding stuff (silage, bale press food, green pellets (Fechner, 1998) The utilization of the PC as source of energy (burning, fermentation gas) (Hertwig and Scholz, 1998) The extraction of the dyes of plant (chlorophyll, carotene, xanthophyll) (Schertz, 1938) and application such in the food and candle industry (Judah, 1954) and environmental analytical or after pure refining in the cosmetic, medicine, biochemistry (Wantanabe 1983), electronic, as nematic liquid crystals (Leblanc et al., 1984) and photovoltaic, as organic dyes (Meissner,1997).

Due to structural similarities of chlorophyll and blood hemoglobine one can expect interesting developments in the field of plant dyes and colorants. The resulting fraction will materially and thermally be treated analogue to PC. The suitability (and applicability) as feed depends mainly on the corresponding extraction compounds and has to be tested. The press cake fraction can be separated analogue to wood raw materials into its main components (Figure 4). On the one side, this green plant press cake fractionation seems from an economical point of view to make not much sense today (due to wood competition). On the other side, there are interesting applications for special vegetable celluloses,

Biobased Polymers by Chemical Valorization of Biomass Components

459

hemicelluloses, lignins and monosaccharides. However, short-term possibilities could be the partial fractionation in combination with other applications (paper, cardboard, moulded articles).

Grass, Lucerne a.o. press cake, hay, late hay crop Neutral solvent and/or water, and/ or steam

soluble or volatile

Extractive free Grass (a.o.) Mild oxidation and extraction

Inorganics degraded, soluble

Lignin Mannose, Xylose, Galactose, Arabinose, Uronic acids a.o.

Polysaccharidefraction

Dilute aqueous alkali

Extractives dyes, proteines a.o.

soluble

Acid hydrolysis

Hemicelluloses ‚Grass‘-cellulose

Acid hydrolysis

Glucose + traces of other carbohydrates and impurities

Figure 4. Classification of the major components of grass press cake, hay and late hay crop (in analogy to Janes R L, 1969).

The ‘green plant’ polyoses (hemicelluloses, pseudocelluloses, polysaccharides) are nutrient-physiologically valuable (Authors group,1978). Furthermore, they can be used (similar to plant rubber) as protecting colloids, emulsifiers in cosmetics, thickeners in food industry (Aspinall, 1983a), adhesives, additives in pulp and paper industry, stabilizers for environmental friendly inks and dyes (Aspinall,1983b), or as thickeners for crude oil drillings Davidson,1980). Lignin is one of the components of lignocellulose. Isolated lignin can be used as dispersant in food industry, as stabilizer for foams and bitumen or as environmental friendly adhesive (ACS,1989; Perl,1967; Crawford,1981).

460

B. Kamm, M. Kamm, I. Scherze et al.

3.1. Cellulose - High Valuable Products Highly pure plant celluloses isolated according to the process shown in Figure 4 can be used as support for immobilization and chromatographic purification of proteins (immune bodies, antigens, enzymes, lectines) (Boeden, 1991; Baeseler, 1992) or as cellulose derivatives for polyelectrolyte complex microcapsules of biological materials (proteins, enzymes, cells) (Dautzenberg et al., 1996, Dautzenberg et al., 1999, Ulrich et al. 2001). Recent development of cellulose sulfate (polyanionic component) and polycarnitine (polycationic component) based polyelectrolyte complex microcapsules are for special interests for application in medicine, pharmacy and cosmetics (Kamm et al., 2001, Kamm et al., 2005).

3.2. Cellulose - Low Valuable Products •

One of the short-term applications could be the use of the press-cake-fraction for the production of rough paper, wallpaper, cardboard and moulded articles for protective materials, (egg packings etc.) (Holm-Christensen, 1989; Fechner and Hertwig 1994). Furthermore, additives for composite materials (fiber-enhanced polymers) or conventional cellulose industry. Paper from press cake of lucerne has been manufactured in Denmark (Carlsson, 1993).

Two interesting developments should be mentioned: First, the studies to produce paper out of Reed canary-grass (Phalaris arundinacea) and cock's-foot (Dactylis gromerata) according to a conventional CTMP-process (Chemo-Thermo-Mechanical Pulp). In this respect, the Reed canary-grass (Phalaris arundinacea) from eastern German lowland moor‘s shows particularly good paper properties (Figure 5). Second, the prenacellR-process regarding the production of raw paper or tech paper from late harvested grass from nature preserve area and waste lay from filament garn industry. Using the so-called yellow lye (17% waste NaOH) under water-steam pressure the fiber is breaked to the grass pulp and following is pressed to boards by filter pressing. It could be shown, that grass card boards have the same or even a better quality (for the paper re-working industry) than analogue waste papers and in addition, they are less expensive (price for waste paper: 100-140 Euro/tons) (Hille Ch, 1999). The prenacell-pulp can also be added to polymeric foam-forming compounds, which can be foamed and cured via micro-wave technology to produce insulation sheets for construction industry (based on grass).

•

A further product group produced from press cake are oxygen-containing chemicals, which are bio-technologically synthesized via fermentation or from pentose/hexose. For this purpose, PC-carbohydrates (e.g. cellulose and hemicellulose) will be degraded via fermentation or chemically to monosaccharides (saccharification). This can be done stepwise or even in one step. In this way, basic chemicals such as lactic acid, ethanol, glycerine, acrylic acid and 1,3-propanediol are bio-technologically accessible in which technological access varied strongly (Danner et al. 1997).

Biobased Polymers by Chemical Valorization of Biomass Components

461

1• Wood, 2• Waste (old) paper, 3• Waste (old) paper, 4•Phalaris arundinacea - Reed canary-grass, 5• Wild mixed grass, 6• Dactylis gromerata - Cock’s-foot Figure 5. Paper made from grass, CTMP-process.

3.3. Levulinic Acid and Sequence Products A major role as high-valuable chemical compound obtained from biomass is levulinic acid, which can be obtained from hexoses and pentoses. During the conversion, 5hydroxymetylfurfural (HMF) is formed via acid catalyzed dehydration, which can be split via dehydration into levulinic acid and methanoic acid (formic acid). Even with raw materials that strongly vary in their quality from batch to batch, the yields that can be obtained are very high. The formic acid can be removed via distillation for further use. Levulinic acid is a versatile chemical intermediate (Dahlmann, 1968, Kuster, 1990, Olson, 2001) (Figure 6). To decrease the waste problems, the State New York (U.S.A.) has built two levulinic acid pilot plants (1 t per day). In these pilot plants, different possibilities shell be tested for the exploitation of the whole variety of carbohydrate-rich and humid waste materials (waste paper, sewage sludge) for levulinic acid production. Thermo-chemical processes tolerate fluctuations in feedstock compositions. In the future, decentred facilities are planned with volumes of 50 to 1000 t/day and more. The high prices of levulinic acid has inhibited largescale use. It currently has a world wide market of about one million pounds per year at a price of $4 to $6 per pound. The New York “biofine-process” is projected to be capable of producing levulinic acid at $0.04 to $0.32 per pound, depending on the scale of operation NYSERDA, 1998, Fitzpatrick, 1999.

462

B. Kamm, M. Kamm, I. Scherze et al. OH

H3C

Levulinicacid—Productionand acid—ProductionandUse Use Levulinic

O O

Levulinic acid

Feedstocks Feedstocks

LevulinicacidacidLevulinic derivates derivates

plant hexoses (press cake a.o.)

organic acids

papermill sludge

solvents

scrap paper

biopolymers

wood waste

fuels

forestry residuals

misc. organics

Consumer Consumer products products • coatings • dyes • herbicides • adhesives • cosmetics • fuel -additives • plasticsubstitutes • food-additives • nylon • pesticides • pharmaceuticals • PVC-plastics • textiles • resins(auto & electrical parts)

Figure 6. Levulinic acid⎯production and use.

•

The PC has also been used as media for growing mushrooms, mulch/green crop enhancer and fertiliser .

Actually, the PC can be used as a renewable sources of carbohydrates for multipurpose uses in the same way as crude mineral oil.

4. PRODUCTS FROM THE JUICE FRACTION (GJ-GREEN JUICE, BJ-BROWN JUICE) In the (esp. fresh pressed) GJ we can find proteins, lipids, glycoproteins, lectins, sugars, free amino acids, dyes (carotenes), hormones, enzymes, minerals and others, so especially crude drugs. The GJ can be fractionated by heat, organic and inorganic acids, acid anaerobic fermentation, centrifugation and gel filtration into a leaf nutrient concentrate (LNC) and a brown juice (BJ). The LNC consists of a mixture of chloroplastic and other organic membrane plus denaturized earlier soluble plant cell proteins. The composition of a LNC is: true protein (60-70%); lipid: [esp. palmitic acid, linoleic acid, lionolenic acid ] (20-30%), starch (5-10%), ash (1-10%), carotenoide/polyene dyes: ß-carotene (1-2 g/kg) and xanthophyll (Pirie, 1975, Schwenke, 1985). The LNC is mainly used for non-ruminant feed to enhance the colour (trough red ß-carotene) of chicken skin or egg yolk. Its also produces tender meat in chickens, ducks and pigs. Pigs fed LNC give pork with increased contents of healthy oleic and linoleic fatty acids in the fat (Carlsson, 1997).

•

Lectins are special proteins, which are able to specifically recognise and complex polysaccharides even in lipid- or protein-bonded form. The old name Phytohämaglutinine / phytaglutinine shows, that the first lecitins have been found in plants (e.g. wheat germs, potato’s). However, not very much is known about the detailed

Biobased Polymers by Chemical Valorization of Biomass Components

•

•

•

•

•

463

mechanism of lectins in plants. However, there are already several applications in human and veterinary medicine (e.g. cancer diagnostics) (Berenzin et al.,1995; Pusztai and Bardocz, 1995) and plant protection (Payene et al. 1993); feed and food control (Schröder et al.,1993). Plant lecitines are available via preparative isolation and subsequent purification by means preparative chromatography. Sugars (monosaccharides, disaccharides and derivatives): GJ and BJ contain valuable special sugars and have highly valuable and sometimes expensive applications (Kirk, Othmer, 1994). Before also these sugars are fermented, they are studied regarding to their potential characterization, and isolation (Starke et al., 2000). Dyes: The GLNC enriched in ß-carotene may have anti-cancer effects. Apart from the use of beta-carotene as pro-vitamin A, ß-carotene and other carotenoid (xanthophyll) are used in cosmetic drugs and as food-, textiles- as well as toys colouring substance (see also chlorophyll (Schertz, 1938; Shearon and Gee, 1949; Judah et al., 1954). Fatty acids: The GLNC is also rich in oleic and linoleic fatty acids, especially palmitic acid, linoleic acid, ,linolenic acid. The lipids have with a good health values. The lipids can be extracted with hydrocarbons. They are interesting for the cosmetic industry (Schwenke, 1985). Crude drugs/ingredients: As the BJ contain specific secondary plant substances, such as, saponins and nicotine, these can be separated from the juice for pharmacological or pesticide purposes (for the isolation see Hagers handbook of pharmaceutical practice, 1972-78). Fertilizer: GJ or BJ used as a bio-fertiliser (soil bioactivators) to return to the soil the macro and micro mineral nutrients, which were removed by harvesting the green crop (Carlsson, 1997).

4.1. Proteins The study of green leaf proteins (LP) goes back more than two hundred years to 1773 when Hilare Rouelle published the first known report on the subject (Rouelle, 1773). The pioneering work of Pirie (1987) since World War II focused on bulk extraction of leaf protein and possibilities for its incorporation into human diets. In the 1980’s leaf protein was the main subject of three international conferences (1982 in India, 1985 in Japan and 1989 in Italy). The most intensive research on leaf proteins has been conducted with alfalpha and tobacco. A comprehensive and critical review of the plant sources, chemistry and nutrition of leaf protein concentrates as well as their preparation is given in the book “Leaf protein concentrates” (Telek and Graham, 1983). Recently research has been focused on Rubisco, the main soluble protein of the leaf (Barbeau and Kinsella, 1988; Douillard and de Mathan, 1994). The present paper is not intended to be a definitive summary of LP research. With protein structure as the starting point, the interrelationship between extraction method, functional properties and recent and future industrial applications in the food and non-food fields will be addressed. Furthermore, experiments with other plant proteins demonstrating the potential of leaf proteins in non-food applications will be discussed.

464

B. Kamm, M. Kamm, I. Scherze et al.

4.1.1. Composition, Fractions and Structure of Leaf Protein From Green Biomass On a wet weight basis the protein content of green leaves varies considerably between 1.2% and 8.2% (lucerne ~ 4%) depending upon plant species, age and growing conditions. The proteins present in leaves possess diverse structure and functions in the living plant including lipoproteins (membranes), structural proteins, photoactive pigment-bound proteins (chloroplasts), and enzymes. Leaf proteins can be divided into two classes: (i) insoluble chloroplastic leaf proteins consisting primarily of lipoproteins and pigment-bound proteins, and (ii) the water-soluble cytoplasmic fraction. The content of soluble proteins is relatively high in the leaves of C3 plants (33 to 48%, e.g. wheat, lucerne, tobacco) compared to C4 grasses (26 to 30%, e.g. maize or sorghum) (Carlsson, 1995). The main soluble protein in leaves is the enzyme ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco EC 4.1.1.39) also known as fraction-I protein. In lucerne leaves, it accounts for 30 to 70% of total nitrogen, depending on the physiological stage or genotype (Douillard and de Mathan, 1994). Rubisco has, irrespective of the plant source, a molecular weight (MW) of between 500,000 and 600,000 daltons and is composed of eight large and eight small subunits with MWs of about 55,000 and 12,500, respectively. The sedimentation coefficient is close to 18.5S. Rubisco possesses a compact, tightly folded threedimensional structure typical of globular proteins. Due to its amino acid composition Rubisco is mildly acidic and carries a negative charge at neutral pH (isoelectric point pH 4.4-4.7) (Bahr et al., 1977). It also has a relatively high average hydrophobicity value of 1275 cal/g residue calculated according to Bigelow (1967). Native Rubisco from alfalfa contains 90 sulfhydryl groups, of which eight are “free” (one per protomer), 36 one exposed after denaturation by SDS and 46 are involved in the formation of disulfide bonds within the Rubisco subunits (Hood et al., 1981). The denaturation temperature of lucerne Rubisco varied between 70°C (pH 7.5) and 61°C (pH 10.3) (Burova et al., 1989). For more details about Rubisco, the reader is referred to the reviews of Barbeau and Kinsella (1988) and Douillard and de Mathan (1994). Table 3. Comparison of the amino acid composition of the different protein fractions of Lucerne (Hatch and Bruce, 1968, Douillard, 1985) Protein Rubisco White protein Green protein Soluble protein Oligomeric soluble protein Membranous protein

Amino acids (in parts per thousand) Hydrophilic (H) Charged Apolar (A) 414 289 285 421 303 272 432 286 268 491 333 239 451 310 275 427 288 299

Small 180 183 180 143 166 169

H/A 1.45 1.55 1.61 2.10 1.70 1.40

Hydrophilic: Asp + Glu + Ser + Thr + Arg + Lys + His; Charged: Asp + Glu + Arg + Lys; Apolar: Val + Ile + Leu + Phe + Met; Small: Gly + Ala

Fraction-II protein consists of a mixture of proteins originating from the chloroplasts and cytoplasm with molecular weights from 10,000 to 300,000 daltons and has a sedimentation constant of from 4S to 10S (Jones and Mangan, 1976).

Biobased Polymers by Chemical Valorization of Biomass Components

465

On the basis of the amino acid composition, Rubisco as well as the green and white fraction of leaf protein are considered hydrophobic (Table 3).

4.1.2. Extraction Methods of Leaf Protein A review of the various methods to prepare leaf protein concentrates (LPC) has been published by Hernandez-Garcia and Martinez-Para (1988). Depending on the methods used, it is possible to obtain (i) an unfractionated leaf protein concentrate (whole LPC), (ii) a decolorized whole LPC, (iii) the green chloroplastic fraction or (iv) the white cytoplasmic fraction. The unfractionated leaf protein concentrate (whole LPC) is defined as the complex precipitated from the green juice without further fractionation, containing both the chloroplastic and cytoplasmic fractions. Such a complex is typically dark green and contains from 50 to 65 wt% protein, 15 to 30 wt% lipids, chlorophylls, carotenoids (xanthophylls), and approximately 5 wt% ash. These concentrates can be easily prepared by heat coagulation at 85-90°C using direct steam injection followed by centrifugation to separate the curde and drying (ProXan process, Knuckles et al., 1972). Due to the similar isoelectric point of both fractions, acid precipitation or anaerobic fermentation of the juice leads to whole LPC, as well (Ohshima et al., 1997). Hernandez et al. (1988) described the preparation of unfractionated LPC by freezing alfalpha juice at –25°C. The subsequent thawing at room temperature resulted in a freezing curde which can be separated by centrifugation and contains 50% of the dried matter and 60 wt% of the nitrogen present in the original juice. Ultrafiltration of the whole herbage juice using membranes with a 1 to 2 x 104 MW cut-off was suitable for obtaining a whole LPC with good solubility, but the process was time- and energy-consuming (Ostrowski-Meissner, 1983a). In order to remove the green color and increase the storage stability of the LPC, extraction with polar solvents has been described (Bray et al., 1978). Although effective, very large amounts of solvents were necessary making this process less economical. Fractionation of chloroplastic and cytoplasmic proteins can be achieved by exploiting their different physico-chemical properties. A well-known method (the ProXan II process) is based on the differential heating of green juice (Edwards et al., 1975). In the first step the green chloroplastic fraction containing 50 wt% proteins, 14 wt% lipids, carotenes and xanthophylls was extracted by steam injection at 60°C and centrifugation. After removing traces of green particles by filtration, the white proteins can be recovered from the clarified juice by heat (80°C, Edwards et al., 1975), acid precipitation (pH 4, Miller et al., 1975), ultrafiltration, or gel filtration (Knuckles et al., 1975). These procedures influenced considerably the functionality of the resulting LPC. A two-stage ultrafiltration procedure using membranes with a 6.5 x 104 MW cut-off (which retained approximately 75-80 wt% of the total recoverable proteins) and a 6 x 10³ MW cut-off led to a fractionation similar to differential heating (Ostrowski-Meissner, 1983a). The most advantageous method of protein separation appeared to be steam coagulation for chloroplastic fraction recovery followed by membrane filtration to separate the cytoplasmic fraction. The application of high molecular polyelectrolytes (cationic or anionic flocculants) allowed for the separation of the pigmented chloroplastic fraction at room temperature (Knuckles et al., 1980a, Baraniak et al., 1989). Antonov and Tolstoguzow (1990) showed that

466

B. Kamm, M. Kamm, I. Scherze et al.

linear polysaccharides (e.g. 0.55% pectin and methylcellulose) were most effective in precipitating chloroplast particles which can then be easily separated by centrifugation. After a short alkali treatment (pH 11.5 for 1 min) a completely water-soluble white protein concentrate was obtained by subsequent acid coagulation and washing. Organic solvents with limited water solubility (e.g. butanol, ethylacetate) were most effective at a concentration of 3 wt% whereas highly polar solvents (e.g. acetone, isopropanol) should be used at 10% in order to remove the chloroplastic fraction (Bray and Humphries, 1978). Optimal solvent extraction was achieved at pH 6.0 and 25°C. In Hungary the Vepex-(vegetable protein extract) process was developed which involves the preparation of a protein concentrate from plant juices, fermentation of the remaining juice to obtain a yeast for feed, and preparation of a feed meal from the plant tissue residue (Hollo and Koch, 1978). The fractionation process, which aimed to obtain the largest amount of white protein fraction possible, was carried out by separating the chloroplast fraction at lower temperature (40-50°C) in the presence of polyvinylpyrrolidone or Ca++, Al+++, and/or Fe++ flocculants, or at 55°C with the addition of surfactants, to delay the separation of the white protein fraction (Koch, 1983). Tobacco is the only species from which crystalline Rubisco has been so far extracted in significant amounts (Pedone et al., 1995). Recently a process leading to a protein isolate rich in Rubisco was described (Levesque and Rambourg, 2001). The green juice of alfalfa was separated into (i) a brown juice rich in Rubisco and (ii) a fraction rich in proteins, pigments, vitamins, insolubles and trace elements. Fraction (ii) was subsequently dried to from a food supplement. Rubisco in fraction (i) was acid-precipitated, concentrated, purified by microfiltration and diafiltration, and dried. Alternatively, purification of the brown juice was accomplished with a resin adsorber.

4.1.3. Optimization of the Protein Extraction Process A comprehensive review of the factors limiting protein recovery and their overall effect on the yield is given by Ostrowski-Meissner (1983b). The factors were grouped into four categories: (i) ecology and agronomy, (ii) plant harvesting, (iii) plant processing and (iv) juices processing. Some important factors will be mentioned. The maximum yield of leaf protein per hectare resulted from cuts of regrown plants, where each regrowth of the plant has reached the physiological stage of prebloom (Carlsson, 1995). With respect to the processing, the pressure conditions were important. To get maximum yields of LP it is favorable to apply pressure slowly so as to extract as much juice as possible with gentle pressure. Any delay in processing should be avoided. Delays can cause an increase in degradation by proteolytic enzymes and an increase in the binding of phenolic compounds to proteins (Pirie, 1994). Metabisulphite or sulfite salts are commonly used in leaf protein isolation to inhibit polyphenoloxidases, improve retention of methionine and available lysine, and increase leaf protein yields. 4.1.4. Applications of the Proteins A prerequisite for exploiting proteins in special applications is the tailored modification of their functional properties. In the case of food, these properties have been defined as the “intrinsic physicochemical characteristics which affect the behavior of proteins in food

Biobased Polymers by Chemical Valorization of Biomass Components

467

systems during processing, storage, preparation and consumption” (Kinsella, 1979). Functional properties can be described as the ability of the protein to emulsify, to adsorb water, to form gels, to provide viscosity and elasticity, or to foam. In addition to the intrinsic attributes of the proteins (i.e. composition, amino acid sequence, conformation and structure), the functionality of proteins is affected by interactions with other components present (i.e. carbohydrates, lipids, salts, surfactants) and the process parameters (i.e. temperature, pH, ionic strength, reducing agents, storage conditions) (Kinsella, 1979). Several technofunctional properties of proteins, their required physico-chemical and molecular parameters and the consequences for different industrial applications are summarized in Table 4.

4.1.5. Nutritional and Functional Properties of Leaf Protein To date, the described utilization of leaf protein concentrates has been limited to feed and food applications. The whole leaf concentrates or chloroplastic fraction are primarily used for feeding nonruminant animals and poultry. The high pigment content (e.g. xanthophylls) of LPC distinguishes it from the products based on soybeans. Broiler and layer feeding trials have shown that the xanthophylls in Pro-Xan are utilized more efficiently than the xanthophylls in dehydrated alfalfa (Kuzmicky et a.,1977). LPC included in the diet of laying hens increased their egg production slightly, and also increased the intensity of the yolk color of the eggs. The aim of a recent Austrian project was to compare the use of fresh green protein and silage protein as a specialty feed (milk re-placer) for calves (Koschuh et al., 2003). Whereas high quality protein products were obtained from fresh biomass, the yield of silage protein was too low due to a high degree of proteolysis. On the basis of amino acid composition, LPC showed high nutritive values. The limiting amino acids in leaf proteins are the sulfur-containing amino acids methionine and cysteine. Depending on the production method, the digestibility of whole LPC is 80-90%, while for cytoplasmic LPC it is usually higher than 95%. The nutritive value is affected adversely by the presence of antinutritional factors (e.g. amino transferase activity, trypsin inhibitory activity, tannins, saponins) which differed dependent on the plants and production process (Hussein et al., 1999). A review of the investigated functional properties of leaf proteins in food applications has been written by Barbeau (1990). It is obvious that the extraction method has considerable impact on the properties and suitability of the leaf protein preparations in food systems. Good solubility plays a key role in the quality of LPC because it is a prerequisite for emulsification, foaming or gelation. The frequently used method of heat coagulation results in irreversible denaturation and loss of solubility. Using the differential heating method, both green chloroplastic and white cytoplasmic alfalfa protein concentrates showed poor solubility between pH = 2 and 10. On the other hand, acid-precipitated leaf protein was soluble above pH = 6 and below pH = 3 (Betschart and Kinsella, 1973). Removal of phenolic compounds by ultrafiltration resulted in increased solubility of alfalfa protein concentrate (Knuckles et al., 1980b). The emulsifying activities and emulsion stabilities of acid-precipitated alfalfa leaf protein concentrates were better than those of soybean concentrate (Wang and Kinsella, 1976a). Extraction of the lipids with acetone led to a slight decrease in solubility and a 50% reduction in water and fat adsorption capacities compared with controls. Both foam formation and stability, however, were markedly improved by extraction of the lipids (Wang and Kinsella, 1976b). The addition of sucrose and NaCl reduced the foaming of the leaf proteins.

Table 4. Relationship between physico-chemical and techno-functional properties and the consequences for different industrial applications Techno-functional properties Emulsification

Foaming

Gelation

Water adsorption and holding Viscosity, thickening power Fat adsorption Adhesion; Cohesive strength Metal-binding

Physico-chemical properties Molecular and structural prerequisites Amphiphilic character, high solubility, ease of unfolding, high molecular flexibility and hydrophobicity High solubility, surface hydrophobicity, ease of unfolding at an interface, reorientation at the interface without aggregation or coagulation; decreased molecular weight; enhancement of protein-protein-interaction (electrostatic) to decrease lipid-sensitivity Long, coiled polypeptides capable of loosing and unfolding by heat; deposition of charged groups; ordered reassociation via hydrophobic association, hydrogen bonding, ionic interactions and/or disulphide linkages Reduced solubility for physical water binding

Increased hydrodynamic properties (molecular volume, size, axial ratio, shape of the molecule) caused by denaturation without gelation High surface hydrophobicity Unfolding (flat conformation, less structured), high polar affinity to the surface, exposure of specific groups, Unfolding, exposure of specific groups

Kinsella (1979), Feeney (1987), Muschiolik (1991), De Graaf (2000) * Damadoran & Hwang, 1995

Suitable modifications (selection) Partial hydrolysis (DH<10, MW>10kDa); acetylation, succinylation

Industrial applications (examples) Preparation of stable oil-in-water emulsions

Partial hydrolysis (DH<10, MW>10kDa); physical protein degradation (mechanolysis)

Preparation of aerated solutions, products and foams

Acetylation, alkali and heat treatment, carbonyl-amine reaction, reaction with bivalent ions

Gel formation (food gels, emulsion gels, surface coatings, film formation, membranes stable to organic solvents)

Chemical modification by acylation and crosslinking of the acyl-modified protein with dialdehyde (*) Alkali treatment (> pH 9)

Water binding powders, textiles, absorber

Heat denaturation and drying Cross-linking to enhance cohesion and reduce water sensitivity Chemical modification by acylation and crosslinking of the acyl-modified protein with dialdehyde (*)

Binding of nonpolar liquid contaminations Adhesive, coatings

Thickener, stabilizer (dispersed systems), regulation of flow behavior of coatings or building materials

Industrial absorbents, waste water reclamation, heavy metal sequestration

Biobased Polymers by Chemical Valorization of Biomass Components

469

The edible protein concentrate from alfalfa prepared by ultrafiltration and spray-drying at 85°C showed excellent functional properties (Knuckles and Kohler, 1982). Emulsions containing 2 wt% alfalfa protein and oil were stable and their consistency was similar to mayonnaise. Firm gels were obtained by heating solutions of 3 wt% alfalfa protein to 72°C and cooling. Foam volume and stability were similar to those of egg white. Since the 1980’s much research has focused on the purification and technological evaluation of Rubisco, the main soluble protein of leaves. Due to the excellent functional and nutritional properties of this tasteless and odorless white powder, purified Rubisco is regarded as having great potential as an ingredient in animal and/or human foods. The reviews of Barbeau and Kinsella (1988) and Douillard and de Mathan (1994) demonstrated that Rubisco was able to form stable foams and emulsions at low temperature, to produce firm gels when heated and to give high fat-binding capacity when the protein has not been modified by heat treatment.

4.1.6. Effects of Enzymatic and Chemical Modification of Leaf Proteins The modification of proteins via enzymatic, chemical or physical procedures is a common way to tailor the functionality. It is generally accepted that limited hydrolysis enhances the surface properties, whereas extensive hydrolysis has a detrimental effect on the stabilization properties. It was shown that the initial enzymatic degradation of native crystalline tobacco fraction-1 protein led to the removal of oligopeptides from the large subunits without a loss of small subunits and destruction of the quaternary structure (Sheen and Sheen, 1987). Insoluble alfalfa protein concentrate was solubilized by the proteolytic enzyme Delvolase and a peptide isolate suitable as a protein supplement in human diets was obtained by coupling the process with ultrafiltration using a 10 kDa MW cut-off membrane (Prevot-D’Alvise et al., 2003). According to Yang et al. (2004) the pepsin digest of leaf protein with a high Rubisco content could potentially be useful in the prevention and/or treatment of high blood pressure (hypertension). So far, the influence of the degree of hydrolysis on the functional properties of leaf protein samples has not been investigated. Chemical modification can be subdivided into (i) hydrophilization (incorporation of polar groups, e.g. –COOH, -NH2, -OH), (ii) hydrophobization (incorporation of apolar groups, e.g. alkyl or aromatic) and (iii) cross-linking (covalent linking of protein molecules). Only a few studies exist regarding the chemical modification of leaf proteins. The succinylation of 84% of the ε-amino groups of lysine resulted in improved flavor, increased solubility (>10-fold), enhanced emulsifying activity (32%) and increased foaming capacity (3-fold) of protein isolated from alfalfa leaves (Franzen and Kinsella, 1976). Acetylation of leaf protein improved the solubility and foaming capacity also but to a much lesser extent. Studies of the soluble tobacco leaf proteins (F-1-p and F-2-p) confirmed an improvement in functionality by succinylation (Sheen, 1991). 4.1.7. Future Trends and Necessary Investigation In order to force the valorization of biomass components, it is important to develop a new market for leaf proteins. Besides the food and feed field, technical applications of proteins are a promising area. Such non-food applications cover a very heterogeneous market ranging from surfactants (emulsifiers, detergents, wetting agents), coatings (e.g. in paint, ink, paper, packaging), and adhesives to biodegradable plastics or materials (e.g. disposable dishes or cutlery). The market potential of renewable materials depends primarily on their superior

470

B. Kamm, M. Kamm, I. Scherze et al.

performance, including functionality and price. The suitability of leaf proteins for such nonfood applications has been insufficiently tested. In most cases, the presence of antinutritive components, dark color or unpleasant grassy smell is not prohibitive to the use of leaf proteins in non-food products. The excellent interfacial and gelling properties of a properly extracted cytoplasmic fraction and Rubisco make them suitable for use as surfactants or film-forming agents in personal care products or biodegradable packaging films. The extreme purity of crystallized leaf protein (e.g. Rubisco from tobacco) may support its use in the field of medicine and pharmaceutics. Due to its large charge/mass ratio, Rubisco has a high tendency to bind ionic substances (Barbeau and Kinsella, 1988). This may allow it to be used as a sequestering agent for environmental pollution. Although for food applications it is important that the proteins are in the native (not denatured) state in order to exert the desired functionality, for technical applications this is less relevant (de Graaf, 2000). The basis for the formation of biomaterials is the reactivity of the macromolecules. The ability of water-insoluble plant proteins to produce biodegradable materials with antistatic and flame resistant properties (useful in electrical devices) will be described in order to demonstrate the potential of leaf protein in such applications. During a simultaneous treatment of pressure (1.2 MPa) and temperature (120 – 170°C) the raw materials (e.g. plant protein with 25% water or a mixture of plant proteins, native potato starch and/or dialdehyde starch with 25% water) were plasticized, fluidized and cooled to create materials with special mechanical properties (Bindrich, 2004). Different plant protein sources when used to make “bioplastics” demonstrate differences in mechanical behavior which cannot be predicted solely from the molecular status of the protein samples (Figure 7). Under certain conditions, the mechanical properties (bending elastic modulus) of composite materials made with wheat were superior to plexiglass (bending elastic modulus 1000 MPa) (Figure 8). At temperatures above 120°C it is so far impossible to predict changes within the protein molecules, i.e. breaking or forming of bonds, and their effects on the mechanical properties of biodegradable materials. Only experimental tests can provide the desired information. Hence, leaf protein concentrates (single or combined with other plant proteins or starch) have potential in the production of biodegradable materials. To date, nearly all information on protein denaturation has been obtained from research on solutions with water content > 90 wt%, and little research has been done in systems with water content lower than 30 wt%. The denaturation temperature of proteins strongly depends on the water content, and can increases to 120-200°C at water contents < 20 wt% (de Graaf, 2000). By using green leaf proteins in the production of packaging materials, the existing pigments (e.g. chlorophyll) could be used to protect light-sensitive substances from oxidation.

Biobased Polymers by Chemical Valorization of Biomass Components

B endin g elastic mo dulus (M Pa)

1400

471

P P-nS P-DAS P-nS-DAS

1200 1000 800 600 400 200 0 Pea

Maize Wheat Potato Rape

Plant Protein Figure 7. Mechanical properties of biomaterials made from protein (P) or a mixture of proteins, native starch (nS) and/or dialdehyde starch (DAS).

B ending elastic m odulus (MPa)

4000 3500

W W-nS W-nS-DAS

3000 2500 2000 1500 1000 500 0 120

130

140

145

150

155

160

170

Structuring temperature (°C)

Figure 8. Effect of structuring temperature on the mechanical properties of biomaterials made from wheat protein II (W) or a mixture of proteins, native starch (nS) and dialdehyde starch (DAS).

The tailoring of protein functionality by physical, chemical and enzymatic methods is a promising method for developing new non-food applications and should be utilized more extensively for leaf proteins. Due to numerous potential glutaminyl and lysyl sites, Rubisco (mostly its large sub-unit) is a good substrate for transglutaminase-directed modifications (Margosiak et al., 1990). This offers new opportunities for solvent stable membranes, biodegradable films and “bioplastics” with different mechanical firmness.

472

B. Kamm, M. Kamm, I. Scherze et al.

4 2. Green Juice and Brown Juice as an Fermentation Media and the Products The GJ and BJ are an excellent fermentation media, as such, and mixed with cereal grain flours, molasses or other hexosen-riches substrates (Thomsen et al., 2004). As fermentation medium, BJ has been used for production of ethanol, lactic acid and other organic acids, plus feed proteins, lysine, soil bioactivators. Especially the fermentation products ethanol, lactic acid and other organic acids, amino acids such as lysine are high-values chemical Biorefinery products. Together with the PC levulinic acid these chemicals parent products and substances can be able to establish a biomass industry and a modern green chemistry.

4.2.1. Lactic Acid - Sequence Product Poly(Lactic Acid) Lactic acid has the potential to become a commodity chemical. High growth rates are expected: an annual volume of 1.36 to 1.8 million tons for lactic acid sequence products alone for the U.S.- market (Gruber, 2001).Chemical products of sequence are propylene glycol, propylene oxide, and epoxides. Propylene oxide is a starting material for the production of polyester, polycarbonates, polyurethanes. Further products are acrylic acid as monomer for polyacrylic acid and resins as well as alkyllactates for application as „green solvents“. Furthermore enantiomeric forms of lactic acid are applied in drugs, pharmaceuticals and agrochemicals. Classical areas of application are such as so-called culinary delight lactic acid, e.g. in food industry and technical processes such as tanneries, textile industry, chemical industry (Kamm et al, 2001). Enormously growing amounts are expected for polymeric materials of lactic acid, poly(lactic) acid. Polylactide is a versatile thermoplastic, which can be processed in manifold ways: e.g. spinning fibres, melting spinning fibres, extrusion foils, injection moulding, thermoforming sheets, extrusion coating for paper and board and many other applications. It is fully biodegradable and compostable and does not disturb the normal process of biodegradation in compost. Especially for the market segments of food packaging and food service, e.g. disposable articles) and performance-products for agriculture as well as fibres for textiles PLA is an very interesting material. In 1993, a market volume of 140.000 – 900.000 tons per year for biodegradable polymers on basis of lactic acid was estimated for the U.S.A. (Cargill,1993). Since then efforts were increased to build major industrial capacities. The company Cargill Dow LLC has built a commercial production facility for polylactide (PLA) in Blair Nebraska, U.S.A.. The Blair facility started its operations in late 2002 and has a maximum capacity of 140,000 metric tons of PLA per year (Hovey, 2002). The establishment of further capacities of the company shall follow within the next 10 years up to a capacity of 450,000 metric tons in Asia and Europe. So it is expected that the price will decrease within the next years to the level of petrochemical based thermoplastics (Cargill Dow, 2001). The Polylactid technology starts with the classical variant of lactic acid fermentation (primary production of inorganic salts of lactic acid) via further steps of production of lactic acid and oligomeric lactic acid, subsequently synthesis of cyclic diesters of lactic acid, from which the polylactide is accessible by ring opening reaction. In this technology principle the two processing steps (biotechnological production of lactic acid and the subsequent chemical steps) are clearly separated (Kamm and Kamm, 1998).

Biobased Polymers by Chemical Valorization of Biomass Components

473

A new technology principle is patented and briefly drafted in Figure 4 (Kamm et al. 1996, Kamm et al. 1997). Aminium lactates act as alternative coupling reagents to connect biotechnological and chemical conversion (synthesis of dilactide). Requirements on amium lactates are low melting points, good crystallinity and good thermal or hydrolytic dissociation ability. Requirements on amines are good water solubility, a pH-value within basic area, sufficient thermal, acid and catalyst stability for the recycling according to the cyclization process, ecological harmlessness and economical availibility. Piperazine dilactate is for example a well investigated model substance (Kamm et al, 2001), (Figure 9).

CH2OH O H H

H HO HO

H

OH H

OH

+

H N

-2 H2O - (H2O)

piperazine

glucose

H N

H

H N

H3C

CH COO

OH piperazinium-dilactate

N H H2SO4 + 2 H3C CH COOH

piperaziniummonosulfate H N

N H

( H2O ) fermentation (lactobacillus)

2

N H

H

+ H2SO4

OH lactic acid + 2 NaOH

N H + 2 H2O + Na2SO4

piperazine (or piperazinium hydrate)

cat. / temp.

H3C O

+ 2 H2O

O

O

O

CH3

dilactide

elektrodialysis H2SO4 + 2 NaOH

Figure 9. Principle of procedure of extremely pure lactic acid and of dilactide based on aminium lactates.

REFERENCES ACS (1989) (American chemical society). ACS Symp. Ser. No. 397. Antonov YuA, Tolstoguzov VB (1990). Food protein from green plant leaves. Nahrung 34:125-134. ASPINALL (1983), (ed.),The polysaccharides. Vol. 2 [Academic Press, New York a) pp. 98193, b) pp. 411-490. Authors group (1978), Deutsches Ärzteblatt, 75, 2735-2741. Baeseler M, Boeden H F, Koelsch R, Lasch J (1992), Cellulose beads: A weak leaking affinity support. J. Chromatogr. 589, 93-100. Bahr JT, Bourque DP, Smith HJ (1977). Solubility properties of fraction I proteins of maize, cotton, spinach, and tobacco. J Agr Food Chem 25:783-789.

474

B. Kamm, M. Kamm, I. Scherze et al.

Baraniak B, Baraniak A, Bubicz M (1989). Protein concentrate from alfalfa and cocksfoot by polyelectrolyte precipitation. Nahrung 33:491-495. Barbeau WE (1990). Functional properties of leaf proteins: Criteria required in food application. Ital J Food Sci 2:213-225. Barbeau WE, Kinsella JE (1988). Ribulose bisphosphate carboxylase/oxygenase (Rubisco) from green leaves – potential as a food protein. Food Rev Int41:93-127. Berenzin et. al. (1995); The Sorbents for Lectin Fraction into Isolectins. Appl. Biochem. Microbiol., 3 254-258. Betschart A, Kinsella JE (1973). Extractability and solubility of leaf protein. J Agr Food Chem 21:60-65. Bigelow CC (1967). On the average hydrophobicity of proteins and the relation between it and protein structure. J Theor Biol 16:187-211. Bindrich U (2004). Werkstoffe aus Protein-Kohlenhydrat-Verbunden. Proceedings 10th International Conference for Renewable Resources and Plant Biotechnology, Magdeburg, Germany. Boeden H F, Pommering K, Becker M, Rupprich C, Holtzhauer M, Loth F, Müller R, Bertram, D (1991), Bead cellulose derivatives as support for immobilization and chromatographic purification of proteins. J. Chromatogr., 552 389-414 Bray WJ, Humphries C (1978). Solvent fractionation of leaf juice to prepare green and white protein products. J Sci Food Agr 29:839-846. Bray WJ, Humphries C, Ineritei MS (1978). The use of solvents to decolorize leaf protein concentrate. J Sci Food Agr 29:165-171. Burova TV, Soshinskii AA, Danilenko AN, Antonov YuA, Grinberg Vya, Tolstoguzov VB (1989). Conformation stability of ribulose diphosphate carboxylase of alfalfa green leaves according to the data of differential scanning microcalorimetry. Biofizika 34:545549. Cargill-Dow (2001) Cargill-Dow to Up PLA Capacity to 450,000 t/y in 10 Yrs. Japan Chemical Week, 11.10.2001. Carlsson R (1989) Green Biomass of Native Plants and new Cultivated Crops for Multiple Use: Food, Fodder, Fuel, Fibre for Industry, Photochemical Products and Medicine. In: New Crops for Food and Industry [Wickens et al., Chapmann and Hall, London]. Carlsson R (1993) Pressed crop for possible production of paper crop. Proc. 4th int. Conf. Leaf Protein Res., New Zealand-Australia, 69-74. Carlsson R (1995). Sustainable primary production - green crop fractionation: Effects of species, growth conditions, and physiological development on fractionation products. In Handbook of plant and crop physiology (Ed. Pessarakli M), Marcel Dekker Inc., N.Y., USA, 941-963. Carlsson R (1997) Food and Non-food uses of immature cereals. In: Cereals ⎯ Novel Uses and processes [G.M. Campbell, C. Webb & S.L. McKee (eds.), Plenum Publ.Corp., New York] 159-67. Carlsson, R (1985) An ecolocically better adapted agriculture. Wet fractionation of biomass as green crops. Macro alga and tuber crops. [Proc. 2nd int. Conf. Leaf Protein Res., Naoya, Japan]. Carlsson, R (1998). Status quo of the utilization of green biomass In: The Green Biorefinery, concept of technology. (First international symposium on green biorefinery) Neuruppin,

Biobased Polymers by Chemical Valorization of Biomass Components

475

Society of Ecological Technology and System Analysis, Berlin, ISBN 3-929672-06-5 ] pp. 39-45. Coombs J, Hall K (1997) The potential of cereals as industrial raw materials: Legal technical, commercial considerations. In: Cereals - Novel Uses and processes [G.M. Campbell, C. Webb & S.L. McKee (eds.), Plenum Publ.Corp., New York,] 1-12. Crawford (1981) (ed.), Lignin Biodegradation and Transformation. [John Wiley & Sons, New York]. Dahlmann J (1968); Notiz über die Darstellung von Lävulinsäure. Chem Ber, 101, 4251-53. Damodaran S, Hwang D-C (1995). Carboxyl-modified superabsorbent protein hydrogel. US 5,847,089. Danner H, Boller V, Neureiter M, Sandhofer M Braun R (1997) Neue biotechnologische Verfahren zur Produktion marktrelevanter C2-, C3- und C4-Verbindungen aus Biomasse. In: Chemie nachwachsender Rohstoffe. Proc. of the conf,. Wien, [BMUJF (ed.), Wien, Austria, ISBN: 3-901 305-71-8]. Dautzenberg et al. (1999) Development of Cellulose Sulfate-based Polyelektrolyte Complex Microcapsules for medical applications, In: Bioartifical Organs II, Technology, Medicine, and Materials, Annals of the New York Academy of Sciences, Vol. 875, pp. 4663. Dautzenberg H, Arnold G, Tiersch B, Lukanoff B, Eckert U (1996) Polyelektrolyte complex formation at the interface of solutions, Progr Colloid Polym Sci 101, 149-156. Davidson (1980), (ed.); Handbook of Water soluble Gums and Resins. [McGraw-Hill, New York]. De Graaf LA (2000). Denaturation of proteins from a non-food perspective. J Biotechnol 79:299-306. Douillard R, de Mathan O (1994). Leaf protein for food use: potential of Rubisco. In New and developing sources of food proteins (Ed. Hudson BJF), Publisher Chapman & Hall, London, UK, 307-342. Douillard, R (1985). Biochemical and physicochemical properties of leaf proteins. In Proteines Veg, Publisher Tech. Doc. Lavoisier, Paris, France, 211-244. Ebert G. (1993) Biopolymere, Struktur und Eigenschaften. B. G. Teubner Stuttgart, (ISBN: 3519-03516-2). Edwards RH, Miller RE, de Fremery D, Knuckles BE, Bickoff EM, Kohler GO (1975). Pilot plant production of an edible white fraction leaf protein concentrate from alfalfa. J Agr Food Chem 23:620-626. Estimates from Battelle, SRI, Cargill (1993) announcement. Fechner M (1998) Grassland economy in Brandenburg, In: The Green Biorefinery, concept of technology. (First international symposium on green biorefinery) Neuruppin, Society of Ecological Technology and System Analysis, Berlin, ISBN 3-929672-06-5 ] pp.47-52. Fechner M, Hertwig F (1994); Paper made of grass very much in the future. Neue Landwirtschaft, 11 29-31 (germ.) Fechner M, Hertwig, F (2003) State institute of Agriculture Brandenburg, Paulinenaue, personal information. Feeney RE (1987). Chemical modification of proteins: comments and perspectives. Int J Pept Res 29:145-161.

476

B. Kamm, M. Kamm, I. Scherze et al.

Fitzpatrick S (1999) (biofine inc.); Process for Conversion of Cellulosic Biomass to Chemicals. In: Green Chemistry and Engineering. Proc. 3rd Ann. Conf. Wahington, D.C. USA, june 1999 [ACS (American Chemical Society), Washington D.C.]. Franzen KL, Kinsella JE (1976). Functional properties of succinylated and acetylated leaf protein. J Agr Food Chem 24:914-919. Gruber P(2001), Nature Works: New Products, Markets, and Sustainability, Presentation 19.12.2001, Industrial Investment Council, Berlin. Hagers Handbuch der pharmazeutischen Chemie (1972-78) (Hagers handbook of phamaceutical practice) Vol. 1-5, chemicals and drugs, [W. Kern et al. (eds), Springer Verlag, Berlin, Heidelberg, New York, Vol. 1, pp.732-36 Hatch FT, Bruce AL (1968). Amino acid composition of soluble and membranous lipoproteins. Nature 218:1166-1168. Hector A, et al. (1999) Plant Diversity and Productivity Experiments in European Grasslands, Science, 286, 11231127. Hernandez A, Martinez C, Gonzalez G (1988). Effects of freezing and pH of alfalfa leaf juice upon the recovery of chloroplastic protein concentrates. J Agr Food Chem 36:139-143. Hernandez-Garcia A, Martinez-Para MC (1988). Leaf protein concentrates. A review. Anales de Bromatologia 40:311-325. Hertwig F, Scholz, V (1998), Thermal conversion from grass. In: The Green Biorefinery, concept of technology. (First international symposium on green biorefinery) Neuruppin, Society of Ecological Technology and System Analysis, Berlin, ISBN 3-929672-06-5 ] pp. 93-100. Hille Ch (1999) Prenacell-process, UVER – Umweltverfahrenstechnik GmbH, Premnitz, Germany. Hollo J, Koch L (1978). VEPEX, das ungarische Blattprotein-Verfahren. Ernaehrungsumschau 25:210-216. Holm-Christensen (1989); The dehydration plant as producer for the cellulose industry. Proc. Dry-Crops 89, 4th Int. Green Crop Drying Congr., Cambridge [Agra Europe Ltd., London, UK] pp. 91-94 Hood LL, Cheng SG, Koch U, Brunner JR (1981). Alfalfa proteins: Isolation and partial characterization of the major component - fraction I protein. J Food Sci 46:1843-1850. Hovey A (2002), Cargill delivers $300M message with new plant. Lincoln Journal Star Lincoln, Nebrasca, 03.04.2002. Hussein L, El-Fouly MM, El-Baz FK, Ghanem SA (1999). Nutritional quality and presence of antinutritional factors in leaf protein concentrates (LPC). Int J Food Sci Nutr 50:333343. Janes R L (1969); In: The pulping of wood. 2nd edn, vol 1 [R.G. Mac Donald (ed), McGrawHill, New York,] p. 34. Jones WT, Mangan JL (1976). Large-scale isolation of fraction 1 leaf protein (18S) from lucerne (Medicago sativa L). J Agr Sci 86:495-501. Judah M A, Burdick E M, Carroll, R G, (1954) Chlorophyll by solvent extraction. Industrial and Engineering Chemistry, 46, 2262-2271. Kamm B (2004) Neue Ansätze in der Organischen Synthesechemie – Verknüpfung von biologischer und chemischer Stoffwandlung am Beispiel der Bioraffinerie-Grundprodulte Milchsäure und Carnitin, Habilitationschrift, Universität Potsdam.

Biobased Polymers by Chemical Valorization of Biomass Components

477

Kamm B et al (2000) Green biorefinery Brandenburg, article to development of products and of technologies and assessment. Brandenburg Umweltber 8, 260-269. Kamm B, Kamm M (1998)Milchsäuren als grüne Biochemikalien. In: The Green Biorefinery, concept of technology. (First international symposium on green biorefinery) Neuruppin, Society of Ecological Technology and System Analysis, Berlin, ISBN 3-929672-06-5 ] pp.89-92 Kamm B, Kamm M (2004) Mini Review, Principles of biorefineries, Appl Microbiol Biotechnol 64, 137-145. Kamm B, Kamm M, Fischbach M, Allwohn J, Birkel S, Krause Th (2001) Polyesters and oligoesters of cationic hydroxy acids, method for production them and use thereof, PCT/WO 01/94441 A1, 11.05.2001/13.12.2001. Kamm B, Kamm M, Kiener A, Meyer H P (2005) MINI-REVIEW, Polycarnitine – a new biomaterial, Appl Microbiol Biotechnol, in press. Kamm B, Kamm M, Richter K (2001) Biobased industrial products. Heterocyclic aminium lactates. New fermentation products and feedstocks for multiple use, Agro-Food-Industry Hi- Tech, 12 (3), 15-19. Kamm B, Kamm M, Richter K, et al. (1996) Verfahren zur Herstellung von 1,4-Dioxan-2,5dionen, neue Piperazoniumsalze von α-Hydroxycarbonsäuren und Fermentationsverfahren zur Herstellung von organischen Aminiumlactaten. Offenlegungsschrift DE 196 28 519 A1, Cl. C07 D 319/12 (04.07.1996/ 14.08.1997) 18 S. Kamm B, Kamm M, Richter K, et al. (1997) Verfahren zur Herstellung von organischen Aminiumlactaten und deren Verwendung zur Herstellung von Dilactid. Europäische Patentanmeldung EP 0 789 080 A2, Cl. C12P 7/56, (07.02.1997/ 13.08.1997). Kamm B, Kamm M, Richter K, Reimann W, Siebert A (2000) Formation of Aminium Lactates in Lactic Acid Fermentation. Fermentative production of 1,4-Piperazinium(L,L)-dilactate and its use as starting material for the synthesis of dilactide (Part 2), Acta Biotechn., 20, (3-4), 289-304. Kinsella JE (1979). Functional properties of soy proteins. J Am Oil Chem Soc 56:242-258. Kirk R E, Othmer D F(1994) (ed.).; Encyclopedia of chemical technology [John Wiley & Sons, New York]. Knuckles BE, Bickoff EM, Kohler GO (1972). PRO-XAN process: Methods for increasing protein recovery from alfalfa. J Agr Food Chem 20:1055-1057. Knuckles BE, de Fremery D, Bickoff EM, Kohler GO (1975). Soluble protein from alfalfa juice by membrane filtration. J Agr Food Chem 23:209-212. Knuckles BE, Edwards RH, Kohler GO, Whitney LF (1980a). Flocculants in the separation of green and soluble white protein fractions from alfalfa. J Agr Food Chem 28:32-36. Knuckles BE, Edwards RH, Miller RE, Kohler GO (1980b). Pilot scale ultrafiltration of clarified alfalfa juice. J Food Sci 45:730-734. Knuckles BE, Kohler GO (1982). Functional properties of edible protein concentrate from alfalfa. J Agr Food Chem 30:748-752. Koch L (1983). The Vepex process. In Leaf protein concentrates (Eds. Telek L, Graham HD) AVI Publishing: Westport, CT, USA, 601-631. Koschuh W, Kromus S, Krotscheck C (2003). Grüne Bioraffinerie – Gewinnung von Proteinen aus Grassäften. Berichte aus Energie- und Umweltforschung 19/2003, BMVIT, Wien. http://www.nachhaltigwirtschaften.at/nw_pdf/0319_proteine.pdf

478

B. Kamm, M. Kamm, I. Scherze et al.

Kromus et al. (2004) The Green Biorefinery Austria – Development of an Integrated System for Green Biomass Utilization, Chem Biochem Eng Q 18 (1) 13-19 Kuster B F M (1990); 5-Hydroxymethylfurfural (HMF). A Review Focussing on ist Manufacture. starch/Stärke, 42, 314-321 Kuzmicky DD, Livingston AL, Knowles RE, Kohler GO, Guenthner E, Olson OE, Carlson CW (1977). Xanthophyll availability of alfalfa leaf protein concentrate (Pro-Xan) for broilers and laying hens. Poultry Sci 56:1504-1509. Leblanc R M, et al.; (1984) Photobiochem. Photobiophys. 7, 41. Levesque D, Rambourg J-C (2001). Procede pour traiter le jus vert issu du pressage d’une matiere foliaire riche en proteins telle que la luzerne. FR 2819685. Margosiak SA, Dharma A, Bruce-Carver, MR, Gonzales AP, Louie D, Kuehn GD (1990). Identification of the large subunit of ribulose 1,5-bisphosphate carboxylase/oxygenase as a substrate for transglutaminase in Medicago sativa L. (alfalfa). Plant Physiol 92:88-96. Meissner D (1997) Farbstoffzellen-Solarenergieumwandlung mit organischen Farbstoffen. Photon, 2, 24-27. Miller RE, de Fremery D, Bickoff EM, Kohler GO (1975). Soluble protein concentrate from alfalfa by low-temperature acid precipitation. J Agr Food Chem 23:1177-1179. Moser A (ed.) (1997], Eco-tech The technical Development, Verlag TU Graz, Austria Muschiolik G (1991). Proteinprodukte als funktionelle Lebensmittelkomponenten (I). Lebensmitteltechnik 11/91: 630-636. National Research Council, USA (2000) Biobased Industrial Products, Priorities for Research and Commercialization, National Academic Press, Washington D.C. NYSERDA (1998) (New York State Energy Research and Development Authority); Commerzialing Biomass Technologies in New York State. Producing a High-Value Chemical from Biomass (Levulinic acid). Project paper. [NYSERDA, Albany, New York] Ohshima M, Proydak NI, Nishino N (1997). Effects of addition of lactic acid bacteria or previously fermented juice on the yield and nutritive value of alfalfa leaf protein concentrate by anaerobic fermentation. Anim Sci Tech 68:820-826. Olson E S, Micelle R K, Schlag A J; Sharma R K (2001) Levulinate Esters from Biomass Wastes, ACS, Symposium Series 794, 51-63. Ostrowski-Meissner HT (1983a). Protein concentrates from pasture herbage and their fractionation into feed- and food-grade products. In Leaf protein concentrates (Eds. Telek L, Graham HD). AVI Publishing: Westport, CT, USA, 437-489. Ostrowski-Meissner, HT (1983b). Optimization of the protein extraction process from grasslands in temperate and subtropical regions. In Leaf protein concentrates (Eds. Telek L, Graham HD). AVI Publishing: Westport, CT, USA, 562-600. Payene M J et al. (1993); Lectin-magnetic separation can enhance methods for the detection of Staphylococus aureus, Salmonelle enterididis and Listera monocytogenes. Food Microbiology, 10 75-83. Pedone S, Selvaggini R, Fantozzi P (1995). Leaf protein availability in food: significance of the binding of phenolic compounds to ribulose-1,5-diphosphate carboxylase. Food Sci Technol-Leb 28:625-634. Peitgen H O, Richter P H (1986) The beauty of fractals, Berlin. Perl (1967) (ed.), The Chemistry of Lignin [Dekker, New York ]. Pirie N W (1975) Leaf protein: a beneficiary of tribulation. Nature, 253 239-241.

Biobased Polymers by Chemical Valorization of Biomass Components

479

Pirie NW (1987). Leaf protein and its by-products in human and animal nutrition. 2nd ed. Publisher Cambridge Univ. Press, Cambridge, Engl. Pirie NW (1994). The bulk extraction and quality of leaf protein. In Modern methods of plant analysis, Vol. 16 Vegetables and vegetables products (Eds. Linskens HF, Jackson JF), Springer Verlag Berlin Heidelberg, 1-22. Pirie N W, 1971 Leaf protein: In: Agronomy, Preparation, Ouality and Use. In: IPB Handbook 20 [Blackwell Scientific Publications, Oxford,]. Prevot-D'Alvise N, Lesueur-Lambert C, Fertin-Bazus A, Fertin B, Dhulster P (2003). Development of a pilot process for the production of alfalfa peptide isolate. J Chem Technol Biotechnol 78:518-528. Pusztai A, Bardocz S (1995) Physiological Role(s) of lectins in Plants and the Effects of their Inclusion in the Diet on the GUT and Metabolism of Mammals. In: Phytochemical and Health. [D.L. Gustine; H.E. Flores (ed.) American Society of Plant Physiologists]. Robowsky K D (1998) The potential of Amino acids from green plants. In: The Green Biorefinery, concept of technology. (First international symposium on green biorefinery) Neuruppin, Society of Ecological Technology and System Analysis, Berlin, ISBN 3929672-06-5 ] pp.89-92. Rouelle HM (1773). Sur les fécules ou parties vertes des plantes & sur la matière glutineuse ou végéto-animale. J Méd Chir Pharm 40:59. Schertz F M (1938) Isolation of Chlorophyll, Carotene and Xanthophyll by Improved Methods. Industrial and Engineering Chemistry, 30,1073-75. Schröder, M.R. et al (1993); Colosiation of Barley Lectin and Sporamin in Vacuolos Transgenic Tobaco Plants. Plant Phys., 101 451-458. Schwenke K D (1985) Eiweisquellen der Zukunft [Aulis-Verlag Deubner, Köln, 1985, ISBN 3-7614-0858-7] pp. 82. Shearon W H, Gee O F (1949); Carotene and Chlorophyll. Industrial and Engineering Chemistry, 41, 220-26. Sheen SJ (1991). Effect of succinylation on molecular and functional properties of soluble tobacco leaf protein J Agr Food Chem 39:1070-1074. Sheen SJ, Sheen VL (1987). Characteristics of fraction-1-protein degradation by chemical and enzymatic treatments. J Agr Food Chem 35:948-952. Sixth Symposium on Renewable Resources and Fourth European Symposium on Industrial Crops & Products (1999), Schriftenreihe Nachwachsende Rohstoffe, Vol. 14, Landwirtschaftsverlag, Münster (a) S. Girardeau, J. Aburto, C. Vaca-Garcia, I. Alric, E. Berredon, A performing method of transesterfication of cellulose and amylose; 832-836; (b) G. A. van Ingen, Plastics based on proteins, 837-839, (c) L. Heier, M. Heintges, K.H. Kromer, Short flaxfibre –production and quality, 842-849. Starke I, Holzberger A, Kamm B, Kleinpeter E. (2000) Qualitative and quantitative analysis of carbohydrates in green juices (wild mix grass and alfalfa) from a green biorefinery by gas chromatography/ mass spectrometry, Fresenius J. Anal. Chem., 367, 65-72. State of Brandenburg (2003), Agriculture and food economy report (germ.) [MELF Brandenburg (ed.), Potsdam ]. Telek L, Graham HD (Eds.) (1983). Leaf protein concentrates, AIV Publishing: Westport, CT, USA.

480

B. Kamm, M. Kamm, I. Scherze et al.

Thomsen M H, Bech D, Kiel P (2004);Manufacturing of Stabilised Brown Juice for L-lysine production – from University Lab Scale over Pilot Scale to Industrial Production, Chem Biochem Eng Q 18 (1) 37-46. Ulrich F, Tullius S G, Gärtner, M, Kamm B, Müller C, Neuhaus P, Steinmüller Th (2001) Mikroenkapsulierung von humanem Nebenschilddrüsengewebe mit Natriumcellulosesulfat und PolyDADMAC, Acta Chir. Austriaca, Vol. 33 (179) 15. Verband der Chemischen Industrie e.V., (1994) Fachvereinigung Organische Chemie, Konzeption des Fachausschusses Nachwachsende Rohstoffe. Wang JC, Kinsella JE (1976a). Functional properties of novel proteins: alfalfa leaf protein. J Food Sci 41:286-292. Wang JC, Kinsella JE (1976b). Functional properties of alfalfa leaf protein: foaming. J Food Sci 41:498-501. Wantanabe T, Fujishima A, Honda K, (1983) Dye-sensitive electrodes. In: Energy Resour. Photochem. Cat. [M.Grätzel (ed.), Academic Press] 359 Yang Y, Marczak ED, Usui H, Kawamura Y, Yoshikawa, M (2004). Antihypertensive properties of spinach leaf protein digests. J Agr Food Chem 52:2223-2225. Zoebelin H (2001) Dictionary of Renewable Resources, WILEY-VCH, Weinheim

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 481-500

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 17

EXPERIMENTAL ANALYSIS OF SMALL COMBUSTION THERMAL SYSTEMS BASED ON PELLETS *

J.C. Morán†1, J.L. Míguez, E. Granada and J. Porteiro Universidad de Vigo - E.T.S. Ingenieros Industriales, Lagoas-Marcosende, s/n. 36200 -Vigo (Pontevedra), Spain

ABSTRACT In this chapter a set joint of experimental techniques for assessing biomass combustion devices is presented. Small scale energy converters such as chimneys, boilers, stoves, etc, producing heat and/or hot water by combustion of biomass (wood, pellets, briquettes, etc.) are especially suited to domestic purposes. However, in regular commercial combustion conditions, this kind of use still has some disadvantages: besides the fact that some emissions (volatile organic carbons, carbon monoxide or NOx) may still be high, it is difficult to compare the quality and performance of equipment working in very different combustion conditions. Due to their relatively low cost and the complexity of combustion in such devices, modelling by numerical analysis is seldom attempted. Controlling operational factors are usually designed and regulated based on the manufacturer’s experience or on handbook values. In order to protect customers, and to assure compliance with minimum requirements for energy performance and maximum limits on pollutant emissions, several national and international regulations have been developed in recent years. Experimental analysis of these devices is a key technique for control and improvement.

1. INTRODUCTION There are numerous discussions in the literature about the uncertainties in the analysis and design of thermal systems, taking into account their stochastic nature. This chapter *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Universidad de Vigo - E.T.S. Ingenieros Industriales. Lagoas-Marcosende, s/n. 36200 -Vigo (Pontevedra). Spain. e-mail: (1) [email protected]

482

J.C. Morán, J.L. Míguez, E. Granada et al.

presents a description of the problem in terms of experimental design and subsequent analysis. Although there are also several practical methods for performing relative analysis the authors present here a still relatively unknown technique based on Grey Theory. Particularly, a method based on this theory called the grey relative analysis is a very simple, reliable analysis tool for time sequence and dispersed series. The use of grey relational analysis combined and compared with classical statistical experimental design and analysis applied to the optimising of a pellet combustion device is described as an application of the combination of the techniques. The use of biomass, a renewable energy source, instead of fossil fuels in heat and power generation is increasing, and is a way of reducing global CO2 and sulphur emissions and saving on diminishing fossil fuels. The use of wood pellets in small-scale systems is an increasing alternative for producing heat and hot water in the household sector. In Spain, just over half energy from renewable sources used involves biomass, which is limited almost solely to heat applications. With regard to thermal applications of biomass, the Renewable Energy Promotion Plan in Spain 1999-2010 [1] developed under the recommendations of the European White Paper [2] highlights the need for improvement in domestic equipment (typical residential equipment in a power range of 8 to 30 kW). The regional distribution of consumption is heavily related to the presence of the paper and pulp manufacturing sector, timber and food industries and the household sector. According to estimates from the 3rd National Forest Inventory, the amount of biomass in Galicia (northwest of Spain) exceeds 600,000 tm a year [3]. Residues derived directly from forest waste represented about 40% of the total (of which: Pine 38%, Eucalyptus Branches 22%, Crust of Eucalyptus 28%, Others 12%). The energy use of residual wood is limited because of its low density. Waste from primary and secondary timber processing accounts for 20-30% in volume of all the raw material used. Biomass pelletising is a densification process that improves its handling characteristics, enhances its calorific value per volume, reduces transportation costs and produces a uniform fuel. Therefore, wood pellets are an alternative not only to non-treated wood but also to the traditional fossil fuels used in small-scale house heating systems [4, 5]. Typical characteristics of lignocellulosic pellets are low moisture content (<10% Wet Basis (WB)), high density (>700 kg⋅m-3) and a heating value around 17000 kJ/kg with a diameter between 5 and 12 mm [6]. The previous mentioned specifications make pellets very interesting for ordinary central heating systems. Moreover, they are an alternative to fossil fuels in small and average size boilers, and also in district heating plants and small scale house heating systems [7]. However, this kind of use still has some drawbacks, as emissions of pollutants such as VOC (Volatile Organic Carbons), CO (carbon monoxide) and NOx still tend to be high (Table 1) [8]. Consequently, small-scale biomass pellet combustion technologies need further research to help to develop new equipment to meet current and future demand [9,10, 11]. Pellets burn cleaner than wood because the feeding rate is regulated and coupled with the accurate amount of air in order to obtain an optimum burn rate. Pellet fuel is made mainly of sawdust and shavings, or fine branches and leaves left over after processing trees for lumber and other wood products. Wood pellet handling is well-known and its aptness for boilers is one of its main advantages, especially in small-scale residence heating systems. In such pellet combustion devices, it is important to select appropriate operational parameters to achieve optimal performance. The desired parameters for solid fuels are usually determined on the

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 483 basis of manufacturer experience or on handbook values [12]. But even with optimal operation parameters, it is difficult to compare the quality and performance of equipment working in such rather different combustion conditions [13, 14, 15]. The main reasons for high pollutant emissions are the relatively low combustion temperatures and the insufficient mixing rate between air and combustible gases resulting from pyrolysis and heterogeneous combustion of solid particles [16]. Table 1. Typical stove emissions Emissions (mg/kWh) CO SO2 NOx Particles NMVOC

Fuel

Gas

Pellet

10 350 350 20 5

150 20 150 0 2

250 20 350 150 10

2. THEORY One important issue would be to control the operational factors to achieve an overall optimal response, or at least some sort of indicator to quantify the gap related to the most possible favourable value. Thus, this kind of problem requires innovative management approaches, such as those that will ensure joint consideration of economic, ecological and social issues. Another growing concern is consumer interest in environmental labels and certificates. Several methods and certification systems exist in Europe [17, 18, 19]. One of the most popular is the Swedish method called P-marking or labelling. From this point of view, integrating and qualifying all performance aspects of a device in a 0 to 1 scale is a very simple and intuitive labelling possibility. Grey relational analysis is part of grey system theory, which is suitable for solving the complicated interrelationships between multiple factors and variables; in this case, how efficiency, emission variables and fuel prices clutch on to control variables. A grey system is defined as a system containing unknown information presented by grey numbers and grey variables [20]. In grey theory, black represents a system devoid of information, while white stands for complete information. Thus, the information that is either incomplete or undetermined is called grey. System theory was initiated by Deng in 1982, and has been widely applied in the last decade in China in several research fields [21, 22]. In many cases the improvement of one performance characteristic may be to the detriment of another. In this research, the successively combination of experimental design classical statistic analysis and the use of grey relational analysis applied to the energetic and economic assessment and optimisation of combustion of wood pellets has been studied [23, 24, 25, 26, 27].

484

J.C. Morán, J.L. Míguez, E. Granada et al.

2.1. Experimental Design In many situations it is desirable to learn something about the association between two attributes of an individual, material, a product, or for instance a complex process like pellet combustion. The problem then consists of determining the nature and degree of the relation. In regression analysis a regression equation is fitted to an observed point cloud of measured variables [28]. The effect caused by a simple variable when many influencing variables change at the same time is not easily to determine [29, 30]. Experimental design and subsequent statistic analysis is useful to find these correlations with minimum test runs and avoiding lost of generality in conclusions [31, 32, 33]. Statistics experimental design demands a minimum number of experiments avoiding a lost of generality in conclusions. These techniques provide valid conclusions and coupled with Anova define which are the most influential factors. This work is focused on the analysis of the response of certain variables as an empirical function of one or more quantitative factors. A general form of this type of response function is y = f ( x 1 , x 2 , x 3 ......)

(1)

where y variable is the response, and x1, x2, x3 are quantitative levels of the factors of interest. Knowledge of the form of the function f, will be found by fitting accurate correlations models to data obtained from designed experiments. When the response is a function of one or two factors, the fitted model is referred as a response surface because it can be graphed in one or two dimensions respectively, and can be explored to determine important characteristics such as optimum operating conditions or relevant tradeoffs. Complete factorial designs require a lot of experiences depending on number of factors, k, and their corresponding number of levels, number of tests points in which the range of a selected variable will be measured. test runs = n = p k

(2)

In order to avoid an unaffordable number of experiences, three level factorial experiments are often conducted in order to fit such response surfaces. Two level factorial would require less tests, but could not detect curvature on the surface (maximums or minimums). As the number of factors increases, even three levels complete factorials become inefficient and impractical. Further, these design do not give equal precision for fitted responses at points (factor level combinations) that are at equal distances from the centre of the factor space, so that the exploration of the response surface would be dependent on the orientation of the design. In this study every experiment set has at least three levels in order to study possible second grade tendencies. A design that has the property of equal precision at equal distances is termed a rotatable design. Rotability is a desirable property for response surfaces models because prior to the collection of data and the fitting of the response surface, the orientation of the design with respect to the surface is unknown. In general rotatable designs can be constructed from equally spaced points on circles or spheres. For more than two factors, design points should

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 485 lie on a sphere, or a hyper sphere in four or more dimensions. The design points must also form a regular geometric figure. All 2k complete factorials are rotatable but 3k are not. Table 2. Experimental layout examples in coded variables

1

3 factors Central composite design Factor Factor Factor a b c -1 -1 -1

2 factors Complete factorial design Factor Factor a b -1 -1

2

-1

-1

+1

-1

0

3

-1

+1

-1

-1

+1

4

-1

+1

+1

0

-1

5

+1

-1

-1

0

0

6

+1

-1

+1

0

+1

7

+1

+1

-1

+1

-1

8

-1.68

+1

+1

+1

0

9

+1.68

0

0

+1

+1

10

0

0

0

11

0

-1.68

0

12

0

+1.68

0

13

0

0

-1.68

14

0

0

+1.68

15

0

0

0

16

0

0

0

17

0

0

0

Run Test

There are classes of designs for two or more factors that can be used in place of 3k factorials for fitting second order polynomials to response surfaces. One design that makes more efficient use of the test runs is the central composite design, and an extra advantage is, that this design can be made rotatable. The total number of test runs in a central composite design, is based on a complete 2k factorial design, but they only require enough extra observations to estimate the second order effects of the response surface. In this case n = 2 k + 2k + m

(3)

which is less than 3k, so that fewer observations are required than in a 3k factorial design. “m” is the number of repeated observations at the design centre necessary to ignore interaction effects and “a” is the absolute extreme value of the desired studied interval. The central composite design can be made to be rotatable by choosing a = F1/4, where F is the number of factorial points. (F = 2k) In this case five levels of each factor has to be considered, coded (+a, +1,0,-1, -a)

486

J.C. Morán, J.L. Míguez, E. Granada et al.

2.2. Contrast of Equations Many times it is difficult to decide which regression or correlation equation is appropriate. In the natural sciences every equation connects only precise quantities and the question of which variable is independent is often irrelevant. But in our case the aim is to find cause, find out which are the actual factors that influence each and every variable. A measure for how well the regression line explains the observed values is the so called coefficient of determination R2. How ever the factual interpretation of any statistical relations, and their testing for possible causal relation, lies outside the scope of statistical methodology. Even when stochastic dependence appears certain, one must bear in mind that the existence of a functional relation says nothing about a causal relation. The recognition of a causal correlation thus follows from the exclusion of other possibilities. Consequently the size of the correlation coefficient will not make any difference in this context. In order to contrast a correlation (mathematical model of the process) an evaluation process is necessary. One common procedure is by the analysis of variance (ANOVA). As result correlations are evaluated and contrasted by ANOVA and validated considering phenomenon like multicollinearity (variance inflation factors, eigenvalues, …) influence factors (Cook distance…), auto-correlation (Durbin Watson…), heterocedasticity, etc.

2.3. Grey Theory The grey system theory, first proposed by Professor Julong Deng in 1982 avoids he inherent defects of conventional statistical methods and only requires a limited amount of data to estimate the behaviour of an uncertain system. The major advantage of Grey theory is that it can handle both incomplete information and unclear problems very precisely. It serves as an analysis tool especially in cases when there is insufficient data [34]. It was recognized that the Grey relational analysis in Grey theory had been largely applied to project selection, prediction analysis, performance evaluation, and factor effect evaluation due to grey relational analysis software development. There are numerous literature discussions regarding the uncertain in the analysis and design of the thermal system taking into account the stochastic nature. There are also several practical methods to perform relative analysis: parameter analytical method in mathematical statistics, orthogonal test analytical method and grey relative analytical method. Of these, the method of grey relative analysis is best suited to the relative analysis of time sequence and dispersed series [35, 36]. Considering a grey system where a set of variables yi depend on certain independent factors xj, all the possible combination of normalized experimental results xij can be expressed if the performance characteristic is the higher the better as: y ij -min y ij x ij =

j

max y ij - min y ij j

j

(4)

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 487 where yij is the ith experimental result in the jth experiment. Or if the performance characteristic is the lower the better pattern, as: y ij -max y ij x ij =

j

min y ij - max y ij j

(5)

j

The grey relational coefficient, ξij, is calculated to express the relationship between the ideal (best) and the actual normalized experimental results

ξij =

min min x io - x ij + ζ max max x io - x ij i j i j x io - x ij + ζ max max x io - x ij i j

(6)

where xi0 is the ideal standardised result for the ith performance characteristic and ζ is the socalled distinguishing coefficient (0≤ζ ≤1). The overall evaluation of the multiple variables is based on the grey relational grade, which is computed by averaging the grey relational coefficient corresponding to each performance characteristic,

γj =

1 m ∑ξ m i =1 ij

(7)

where γj is the grey relational coefficient for the jth experiment and m is the number of performance characteristics (variables).

2.4. Procedure for the Experiment To reduce the influence of any factors other than those studied here, a strict measurement procedure is applied. A stove of this kind never attains complete steady-state operation, mainly due to its batch feeding system, which in our case is controlled “pellet by pellet”. If the measuring system is fast enough, an appreciable fluctuation may be detected after each pellet falls into the bed, as a result of the stir produced in the particularly thin bed that exists in this kind of fixed bed combustion system (this fluctuation is normally only detected by the CO analyser). For that reason we chose to follow the remaining variables (20 temperature readings, 6 pressure readings, 5 flows and the concentration of the rest of the gas species) until no variation was observed (which in several cases took more than 2 hours). Once the steady state was achieved, data recording was started. This took 20 minutes initially, but was subsequently extended to 60 minutes [37, 38].

488

J.C. Morán, J.L. Míguez, E. Granada et al.

3. EXPERIMENTAL The University of Vigo (Spain) has developed a pilot plant to study pellet combustion in its different configurations and load conditions, taking into consideration both the energy and environmental viewpoints. The core of the plant is a commercial stove which has been modified to allow multiple controlled combustion conditions [39]. As with every complex system, in pellet combustion the effect caused by a single variable is not easily determined when many influential variables change at the same time. Experimental design techniques and subsequent statistical analysis of their results are useful to determine these correlations with a minimum number of test runs with no loss of generality in the conclusions. Many papers on this methodology have been published in many research fields since Box and Wilson developed the idea, some of which have dealt with combustion systems, but none has been reported that applies to pellet combustion in small systems. Previous tests at our experimental plant revealed that the use of flue gases to preheat primary air improves the system’s overall performance. Consequently the use of secondary air and/or gas recirculation, which has been successfully applied in bigger systems, may also be interesting [40]. The planning of the experimental part demands that the number of experiments should be minimised without loss of generality in conclusions. Due to the heterogeneity of the pellet, a minimum number of experiments is needed [41]. The most appropriate type of experiment and also the number of experiments are determined. These experiment design techniques provide valid conclusions and define which the most influential factors are. They also show the predicted tendencies in order to achieve a better efficiency of the general process [42].

3.1. Plant Description An experimental plant has been used for this research (Figure 1). The main parts of this plant are the feeding system and the combustion unit.

Figure 1. Pilot Plant General View.

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 489 In order to study the combustion of different mixtures of pellets and even potentially cheap forest residues, a new feeding hopper for domestic biomass boilers has been implemented in the plant. This new hopper consists of a 200 mm wide (100 mm per product) band that drags both materials, pellet or forest residues, at the same time. The feeding rate can be regulated by means of a floodgate (0-40 mm height) or by changing the conveyor speed.

Figure 2. Combustion chamber scheme.

The combustion unit can be described as a pellet stove surrounded by a water jacket and vacuum operated, where pellets drop from the hopper into the combustion grate while combustion air is supplied upwards, through the grate. The stove has also a window door for ornamental purposes, maintenance and radiation heating purposes, which is closed for normal heat production. The combustion chamber has been adapted to allow multiple combinations of gas supply (primary and secondary air or already burnt gases, as desired) and also several access ports have been installed in the sidewalls and in all its gas-lines to facilitate the measurement of pressure, temperature, mass flow and gas concentration. The whole combustion device is isolated to calculate energy balance and efficiency.

Control and Analyses Equipment • Pressure: is measured by SIEMENS transducers (Sitrans – P) • Temperature: measured through thermocouples and RTD’s (Pt-100). • Combustion chamber exchangers, air and gas tubes have been isolated with polystyrene. • Pellet flow rate control: moved by a controlled motor that determines the amount of pellet introduced in the combustion chamber. Flow is determinate measuring hopper weigh changes with an extensiometric gauge. • Air flow rate: is controlled pressure drop by means three speed regulated fans. • Water jacket temperature: is fixed with a regulated secondary water flow circuit • Gas composition: gas analyses were carried out using a Horiba PG250 and TESTO-350 to measure gas concentrations of the main components involved: NOx/SO2/CO/CO2/O2

490

J.C. Morán, J.L. Míguez, E. Granada et al.

3.2. Material Typical characteristics of lignocelluloses pellets are low moisture content (<10% Wet Basis (WB)), high density (>700 kg⋅m-3) and a heating value around 17000 kJ/kg with a diameter between 5 and 12 mm, and prices in Galicia are between 25 to 30 €/MWh. The previous mentioned specifications make pellets very interesting for ordinary central heating systems, and especially for single-family houses, which are very common in Galicia. In this work, two relatively different kinds of pellet are tested. Table 3. Pellets properties

Diameter (mm) Length (mm) Density1 (kg⋅m-3) Moisture (% WB) Proximate Analysis (wet % dry basis) Ash (550 ºC) Ultimate Analysis (wet % dry basis Ash free) Carbon Hydrogen Oxygen Nitrogen Sulphur 1 Geometric method.

Pellet 1 7 7-21 1166 8.5

Pellet 2 6.6 10-15 943.4 9.5

0.68

0.90

51.7 6.7 40.7 0.17 < 0.05

52.0 5.3 42.0 0.92 < 0.05

4. RESULTS AND DISCUSSION Previous test analysis shows the dominance of just two main factors mp and n. To study secondary air influence an amount M2 (air flow rate in l/s) of the total air introduced into the stove, MT A, (air flow rate in l/s), is forced through the lateral pipes. The ratio sa = M2 / MT A is selected as the factor for Set SA. The study covers a range from 0 to 0.30 of this factor. Analogously, to study the effect of smoke recirculation part of the flue gas, MR (gas flow rate in l/s), of the combustion chamber MTS (gas flow rate in l/s) is again introduced into the stove, sr = MR / MT S is the third factor selected for Set SR. The study covers a range from 0 to 0.30. Changes in the factors have an effect on variables but there is no simple direct relation in these changes, as each variable also depends on other factors. Therefore, correlations between factors and variables needed a statistical approach.

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 491

4.1. Variables and Factors In accordance with previous tests, two different experimental designs are used to determine the controlling factors for pellet combustion in such stoves. The variables studied for different combustion conditions are heat in water Qw , heat in flue gases Qs , stove efficiency ηst , combustion efficiency ηc and exhaust gas composition. Statistical experimental design coupled with ANOVA defines which the most influential factors are. Table 4. Factors and variables Factors Stoichiometric ratio Pellet supply Secondary air Recirculation fraction Variables QW (kW) Heat in water QS (kW) Heat in flue gas QT (kW) = QW + QS Liberated heat 1 Supplied heat QA (kW) = mp × LHV Stove efficiency ηst = QW / QA Combustion efficiency ηc = QT / QA CO (g/kWh Pellet) CO emissions (gases) NO (g/kWh Pellet) NO emissions (gases) n mp (gr/s) sa sr

1

LHV = low heating value wet basis

Table 5. SET SA and SET SR factor levels before randomising

Set SA Coded levels mp(gr/s) +1.68 2.0 +1 1.8 0 1.5 -1 1.2 -1.68 1.0

n 2.00 1.84 1.60 1.36 1.20

Set SR sa (%) mp (gr/s) 30 2.0 24 1.8 15 1.5 6 1.2 0 1.0

n 2.00 1.84 1.60 1.36 1.20

sr (%) 30 24 15 6 0

4.1.1. Secondary Air (Set Sa) The heat transferred to water depends on the amount of pellet supplied (mp), and, to a lesser extent, on n. The amount of secondary air used has no significant influence. Heat in smoke increases linearly with the contribution of both mp and n. Secondary air reduces CO emissions, with n still being the critical factor, but the amount of secondary air is also significant even for combustion conditions where CO emissions are at their lowest (n approximately 1.65).

492

J.C. Morán, J.L. Míguez, E. Granada et al. Table 6. Experimental results. Set SA Mp

n

sa

ηst

ηc

Qw

CO

NO

g/s 2.00

1.60

0.15

0.73

0.87

23.25

g/kWh 17.26

g/ kWh 0.63

1.50

1.60

0.30

0.75

0.86

19.97

2.72

0.68

1.80

1.80

0.24

0.72

0.87

21.92

2.92

0.74

1.20

1.40

0.06

0.74

0.81

15.31

40.56

0.45

1.20

1.40

0.24

0.74

0.81

15.15

21.29

0.51

1.20

1.80

0.06

0.77

0.87

15.84

9.90

0.61

1.20

1.80

0.24

0.78

0.89

16.30

3.67

0.64

1.50

1.60

0.15

0.71

0.82

18.66

23.71

0.53

1.50

1.20

0.15

0.62

0.69

16.21

40.45

0.43

1.50

2.00

0.15

0.74

0.88

19.07

4.87

0.67

1.50

1.60

0.15

0.73

0.84

18.65

15.20

0.47

1.00

1.60

0.15

0.80

0.88

12.99

11.12

0.52

1.50

1.60

0.00

0.74

0.85

18.99

17.89

0.53

1.80

1.40

0.24

0.70

0.81

21.21

18.25

0.49

1.80

1.80

0.06

0.68

0.83

21.56

3.55

0.85

1.80

1.40

0.06

0.70

0.81

20.9

23.49

0.45

1.50

1.60

0.15

0.74

0.85

18.6

12.32

0.48

Lowest value

0.62

0.69

12.99

2.72

0.43

Highest value

0.80

0.89

23.25

40.56

0.85

Test

SA -1 SA - 2 SA - 3 SA - 4 SA - 5 SA - 6 SA - 7 SA - 8 SA - 9 SA - 10 SA - 11 SA - 12 SA - 13 SA - 14 SA - 15 SA - 16 SA - 17

4.1.2. Smoke Recirculation (Set SR) The heat transferred to water depends mainly on the two factors mp and n, with mp being the more significant of the two. The amount of recirculation has only a small influence. Heat in smoke (QS) increases linearly with the contribution of both mp and n. Emissions, however, are affected by recirculation. CO can be reduced significantly by ensuring a proper air-fuel rate and with a relatively low degree of recirculation of smoke.

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 493 Table 7. Experimental results. Set SR Test

Mp

n

sr

ηst

ηc

Qw

g/s

CO

NO

g/kWh

g/ kWh

SR -1

1.2

1.36

0.06

0.81

0.89

17.46

23.84

0.25

SR - 2

1.5

1.20

0.15

0.70

0.79

19.45

27.86

0.21

SR - 3

1.5

1.60

0.15

0.80

0.93

21.40

7.04

0.33

SR - 4

1.2

1.84

0.06

0.82

0.94

17.85

3.45

0.34

SR - 5

1.5

1.60

0.30

0.84

0.99

20.11

2.41

0.40

SR - 6

1.8

1.36

0.06

0.80

0.92

23.81

35.31

0.30

SR - 7

1.5

2.00

0.15

0.73

0.88

20.39

2.90

0.42

SR - 8

1.8

1.84

0.06

0.79

0.95

24.26

4.80

0.42

SR - 9

1.0

1.60

0.15

0.87

0.95

15.49

3.02

0.27

SR - 10

1.2

1.36

0.24

0.83

0.91

17.33

7.05

0.25

SR - 11

2.0

1.60

0.15

0.76

0.91

25.02

3.04

0.39

SR - 12

1.5

1.60

0.15

0.78

0.91

20.47

5.76

0.33

SR - 13

1.8

1.36

0.24

0.72

0.84

23.08

2.37

0.32

SR - 14

1.8

1.84

0.24

0.77

0.94

22.39

1.86

0.42

SR - 15

1.5

1.60

0.00

0.87

0.99

21.50

20.99

0.33

SR - 16

1.5

1.60

0.15

0.83

0.95

21.72

7.21

0.32

SR - 17

1.2

1.84

0.24

0.80

0.92

17.39

1.25

0.35

Lowest value

0.70

0.79

15.49

1.25

0.21

Highest value

0.87

0.99

25.02

35.31

0.42

4.2. Statistical Analysis An extensive statistical analysis of each variable may obtained contrasted correlations for both sets of measures. Relatively high correlation index R2 is obtained in the regressions for both energy and emissions variables. Concerning energy efficiency neither recirculation of air nor smoke has major advantages, at least directly. Best efficiency is obtained with low loads and high air fuel ratio (n = 2). However if CO is an important question recirculation of smoke reduces remarkably the emissions. Greatest reduction is obtained with low air fuel ratios (n < 1.5). In addition, as NO emission increase with the air fuel ratio, recirculation of smoke can help to get an overall optimum combustion as it make possible to work with lower n, low NO emissions, without a big loss in efficiency (only 5%) and lower CO emission as combustion without recirculation and an air fuel ratio of n = 2.

494

J.C. Morán, J.L. Míguez, E. Granada et al. Table 8. Statistical results R2C SET SA QW(kW) = 9.83⋅mp + 2.35⋅n

99.9%

QS(kW) = - 5.62 + 3.39⋅mp + 2.14⋅n

98.6%

CO gr/kWhPellet = 62.68/n - 62.43⋅(sa)

84.1%

NO gr/kWhPellet = 0.135 ⋅ n2 + 0.145⋅mp

98.3%

2

ηC = 0.78 + 0.26⋅n⋅mp-1 - 0.33⋅mp-1 ηst = 0.58 + 0.14⋅(n⋅mp-1)

SET SR QW(kW) = 10.9 ⋅ mp + 2.58⋅n

99.8%

QS(kW)= -5.48 + 3.51⋅mp + 2.17⋅n

98.6%

CO gr/kWhPellet = 51.33/n2 - 76.7⋅ (sr)

81.9%

NO gr/kWhPellet = 0.07 ⋅ n2 + 0.10⋅mp

99.5%

ηC (%) = 82.84 + 27.3⋅n⋅mp-1 - 31.5⋅mp-1 ηst (%) = 62.67 + 14.8⋅(n⋅mp-1)

The optimisation of pellet combustion as a whole means achieving that combination of independent process parameters (factors) that leads to the highest possible values of heat in water, with the best achievable efficiencies, and the lowest attainable emissions of NO and CO. As a result, an improvement of one variable may be in detriment to another variable that depends on the same controlling factor (example SET SA in Table 9). Table 9. Performance characteristics. SET SA Factors

Correlated Variable

Goal

mp

n

QW

↑

↑

↑

ηst

↑

↓

↑

Not coupled

ηc

↑

↑

↑

Not coupled

CO

↓

Not coupled

↑

NO

↓

↓

↓

↓ Not coupled

Sa Not coupled

Optimisation of the multiple variables cannot be as straight-forward as the optimisation of a single variable. Consequently the problem is not only to detect which is the best configuration but also to define the optimal combustion that would involve simultaneous maximal stove and combustion efficiency (highest value) and minimum emissions (lowest value). Thus, the optimal operation conditions cannot be easily established and need the definition of the grey relational grade (GRG). Through this new variable, tests can not only be

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 495 compared, but also the optimisation studying the correlation of the GRG with the factors means the optimisation of the whole combustion process.

4.3. Grey Relational Grade In the grey relational analysis method, experimental data (combustion efficiency, stove efficiency, CO emissions, NO emissions) are first normalised in the range between zero and one, which is also called the grey relational generation. Subsequently, the grey relational coefficient is calculated from the normalised experimental data to express the relationship between the desired and actual experimental data. The grey relational grade is then computed by averaging the grey relational coefficient corresponding to each process response. The overall evaluation of the multiple process responses is based on the grey relational grade. Experimental results of selected performance characteristics (stove and combustion efficiency and CO and NO emissions) were standardised to calculate the grey relational coefficient of each variable. The grey relational grade (GRG) was calculated by averaging the grey relational coefficient. Certain experimental results are given in Table 10. The influence of each factor using the averaged grey relational grade for each level was analysed. A higher value of the grey relational grade represents a stronger relational degree between the reference sequence x0(k) and the given sequence xi(k). As mentioned before, the reference sequence x0(k) is the best process response in the experimental layout. In other words, optimisation of the complicated multiple process responses can be converted into optimisation of a single grey relational grade. Depending on which group of performance characteristics are included in order to calculate the Grey Relational Grade, different scales are obtained. As the grey relational coefficient, ξij, express the relationship between the ideal (best) and the actual normalized experimental results, so best operational conditions for ηst performance were obtained in tests coded SR-9 and SR-15, those for ηc performance were obtained in tests coded SR-5 and SR-15, those for CO performance in test number SR-17, and those for NO performance was obtained in test number SR-2. Just as classical statistical analysis show, but straightforwardly, the grey relational grade including all tests (SET SA+ SET SR) illustrate that SET SR is better (Figure 3). The influence of each factor on grey relational grade can be studied as each factor is an independent variable. According to this factor analysis, optimum is obtained around mp = 1.00, n = 1.60, sr = 30%. This optimum was verified through a confirmation experiment and recalculation of new relational coefficients and grey relational grades. The inclusion of confirmation test achieved a new highest GRGN on this new scale (SET SA + SET SR + OPTIMUM), higher than previous best value test (former value 0.89, in new scale 0.87). This optimum means that although not all performance characteristics achieved the best values individually, this combination of factors produced the best possible overall averaged values.

496

J.C. Morán, J.L. Míguez, E. Granada et al. Table 10. Grey relational grade SET SA + SET SR Test

mp

n

sa

sr

g/s

ηst

ηc

CO

NO

ξij

ξij

ξij

ξij

GRG

SA-1

2.0

1.6

0.15

-

0.48 0.56 0.55 0.43

0.50

SA-2

1.5

1.6

0.30

-

0.51 0.53 0.93 0.40

0.60

SA-3

1.8

1.8

0.24

-

0.45 0.55 0.92 0.37

0.57

SA-4

1.2

1.4

0.06

-

0.49 0.45 0.33 0.57

0.46

SA-5

1.2

1.4

0.24

-

0.49 0.46 0.50 0.52

0.49

SA-6

1.2

1.8

0.06

-

0.56 0.56 0.69 0.45

0.56

1.2

1.8

0.24

-

0.59 0.60 0.89 0.43

0.63

SA-8

1.5

1.6

0.15

-

0.45 0.47 0.47 0.50

0.47

SA-9

1.5

1.2

0.15

-

0.33 0.33 0.33 0.60

0.40

SA-10

1.5

2.0

0.15

-

0.49 0.58 0.84 0.41

0.58

SA-11

1.5

1.6

0.15

-

0.47 0.50 0.58 0.55

0.53

SA-12

1.0

1.6

0.15

-

0.63 0.57 0.67 0.50

0.59

SA-13

1.5

1.6

0.00

-

0.50 0.52 0.54 0.50

0.51

SA-14

1.8

1.4

0.24

-

0.42 0.46 0.54 0.54

0.49

SA-15

1.8

1.8

0.06

-

0.40 0.48 0.90 0.33

0.53

SA-16

1.8

1.4

0.06

-

0.43 0.45 0.47 0.57

0.48

SA-17

1.5

1.6

0.15

-

0.49 0.51 0.64 0.54

0.55

SR -1

1.2

1.36

-

0.06

0.68 0.60 0.47 0.89

0.66

SR -2

1.5

1.20

-

0.15

0.43 0.43 0.42

0.57

SR -3

1.5

1.60

-

0.15

0.65 0.71 0.77 0.73

0.71

SR -4

1.2

1.84

-

0.06

0.72 0.75 0.90 0.71

0.77

SR -5

1.5

1.60

-

0.30

0.81

0.94 0.62

0.84

SR -6

1.8

1.36

-

0.06

0.65 0.68 0.37 0.79

0.62

SR -7

1.5

2.00

-

0.15

0.48 0.58 0.92 0.60

0.64

SR -8

1.8

1.84

-

0.06

0.61 0.79 0.85 0.60

0.71

SR -9

1.0

1.60

-

0.15

0.79 0.92 0.84

0.89

SR -10

1.2

1.36

-

0.24

0.76 0.65 0.77 0.91

0.77

SR -11

2.0

1.60

-

0.15

0.54 0.65 0.92 0.64

0.69

SR -12

1.5

1.60

-

0.15

0.59 0.65 0.81 0.73

0.70

SR -13

1.8

1.36

-

0.24

0.46 0.50 0.95 0.75

0.66

SR -14

1.8

1.84

-

0.24

0.56 0.75 0.97 0.60

0.72

SR -15

1.5

1.60

-

0.00

0.50 0.73

0.81

SR -16

1.5

1.60

-

0.15

0.76 0.79 0.77 0.74

0.77

SR -17

1.2

1.84

-

0.24

0.65 0.68

0.76

1

1

1

1

1

1

0.70

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 497 SA GRG (SA+SR)

SR GRG (SA+SR)

1,00 0,90 0,80 0,70

GRG

0,60 0,50 0,40 0,30 0,20 0,10 0,00 1,15

1,25

1,35

1,45

1,55

n

1,65

1,75

1,85

1,95

Figure 3. Grey relational grade. Configuration analysis.

SR GRG (SA+RS) 1,00 0,90 0,80 0,70

GRG

0,60 0,50 0,40 0,30 0,20 0,10 0,00 0,85

1,05

1,25

1,45

1,65 mp

Figure 4. Grey relational grade. Factor dependence analysis.

1,85

2,05

2,05

498

J.C. Morán, J.L. Míguez, E. Granada et al. Table 11. GRGN - SET SA + SET SR + Optimum Test SR -9 Optimum

mp g/s 1.0 2.0

n

sa

sr

ηst

ηc

1.60 1.6

-

0.15 -

0.87 0.88

0.95 0.97

CO g/kWh 3.02 1.23

NO g/ kWh 0.27 0.36

GRGN 0.87 0.89

CONCLUSION This work carried out relative analyses on the multivariable relationships of pellet combustion in an improved domestic pellet boiler by means of the grey relative analytical method. Grey relational analysis is part of a Grey theory and is here applied to evaluate the combustion process. In addition to classical statistic techniques and by means of the grey rational grade, this methodology permits qualifying all the aspects of combustion with just a single variable. Consequently, the optimisation of complicated multiple responses can be converted into the optimisation of a normalized single grey relational grade. Indeed, the confirmation experiment shows that factor dependence analysis of the grey relational grade led to an “optimal combustion”.

REFERENCES [1] [2] [3]

[4] [5]

[6]

[7]

[8]

Instituto para la Diversificación y Ahorro de la Energía, IDAE. Plan de Fomento de las Energías Renovables en España. 1999 The European Commission. 97- 599 Energy for the Future: Renewable Sources of Energy. White Paper for a Community Strategy and Action Plan. 1997 Hernandez, C . Biomass in Spain: Activities, policies and strategies. The renewable energy promotion plan. 1st World Conference on Biomass for Energy and Industry. Sevilla. London: James and James Science Publishers Ltd. 2000. pp 1213-1216. OPET Sweden Framework. Refined biomass - a source for climate change and business opportunities Work package: Market development of refined biomass. EM-BM-1. 2001 Gustavsson, L; Tullin, C. and Wrande I. Small-scale biomass combustion in Sweden research towards a sustainable society. Proceedings of the 1st World Conference on Biomass for Energy and Industry. Sevilla. London: James and James Science Publishers Ltd. 2000 pp 1553 - 1555. Rösch, C.; Kaltschmitt, M. and Limbrick A. Standardisation of Solid Biofuels in Europe. Proceedings of the 1st World Conference on Biomass for Energy and Industry. Sevilla. London: James and James Science Publishers Ltd. 2000. pp 609-612. Obernberger, I. and Thek G. The Current State of Austrian Pellet Boiler Technology. BIOS Bioenergiesysteme GmbH. First World Conference on Pellets. Stockholm. Sweden. 2002. pp 45-48. Instituto para la Diversificación y Ahorro de la Energía, IDEA. Biocombustibles para edificios de viviendas.Ministerio de Economía. España M-49974-2002.

Experimental Analysis of Small Combustion Thermal Systems Based on Pellets 499 [9] [10]

[11] [12] [13]

[14]

[15] [16] [17] [18]

[19] [20] [21] [22]

[23] [24]

[25]

[26]

Cotton, R.A and Giffard, A. Introducing wood pellet fuel to the UK. Renewable Heat and Powder Ltd. DTI Sustainable Energy Programmes – Altener Programme. 2001 Frandsen, S. Biomass Survey in Europe. Country Report of Denmark. Danish Technological Institute. Danish Energy Authority. European Bioenergy Networks. 2002. Egger, Ch. and Öhlinger, Ch. Strategies and Methods to Create a New Market. First World Conference on Pellets. Stockholm. Sweden. 2002. pp 35-37 Lasselsberger, L. Quality Marking and Environmental Testing of Small-Scale Biomass Boilers in Austria. Landtechnische Forschung (2000). Skreiberg, O. and Saanum, O. Comparison of Emission Levels of Different Air Pollution Components from Various Biomass Combustion Installations in the IEA Countries. Supplemental Report to the IEA-Project -Emissions from Biomass Combustion. The Norwegian Institute of Technology. Institute of Thermal Energy and Hydropower. 1995. Fiedler, F. The state of the art of small-scale pellet-based heating systems and relevant regulations in Sweden, Austria and Germany. Renewable and Sustainable Energy Reviews. 2004 vol, 8 pp 201 - 221. Dias, J:; Costa, M. and Azevedo, J.L.T. Test of small domestic boiler using different pellets. First World Conference on Pellets. Stockholm. Sweden. 2002. pp 137-141 Sjöstrom E. Wood Chemistry: Fundamentals and Applications. Academic Press. 1993. Oravainen, H. Testing methods and emission requirements for small boilers (<300kW) in Europe. VTT Energy. Helsinki - Finland.2000. EN 303-5 Heating boilers. Part 5: Heating boilers for solid fuels, hand and automatically stocked, nominal heat output of up to 300 kW. Terminology, requirements, testing and marking. EN-12809. 2001 Residential independent boilers fired by solid fuel - Nominal heat output up to 50 kW - Requirements and test methods Wu, H. The Introduction of Grey Analysis. Gauli Publishing Co. Taipei. China. 1996. Deng, J. Grey System:Society and Economy. Publishing House of National Defense Industry. Wuhan. China 1985. Deng, J. Introduction to Grey System Theory. The Journal of Grey Systems. Sci-Tech Information Services, England and China Petroleum Industry Press, P. R. China. 1989. vol 1. Coetzer, R.L.J. and Keyser, M.J. Experimental design and statistical evaluation of a full scale gasification project. Fuel Processing Technology. 2003. vol 80 pp 263-278 Lin, J.L. and Lin, C.L. The Use of the Orthogonal Array with Grey Relational Analysis to Optimize the Electrical Discharge Machining Process with Multiple Performance Characteristics. International Journal of Machine Tools and Manufacture. London. GB. 2002 vol 42 Lin, C.L. and Lin, J.L. Optimisation of the EDM Process Based on the Orthogonal Array with Fuzzy Logic and Grey Relational Analysis Method. The International Journal of Advanced Manufacturing. London. GB. 2002 vol 42 Huang, J.T and Liao, Y.S. Optimization of Machining Parameters of Wire-EDM Based on Grey Relational and Statistical Analyses. International Journal of Production Research. 2003. vol. 41- 8 pp 1707-1720.

500

J.C. Morán, J.L. Míguez, E. Granada et al.

[27] Huang, J.T and Liao, Y.S. Application of Grey Relational Analysis to Machining Parameters Determination of Wire Electrical Discharge Machining. Proceedings International Symposium for Electromachining, (ISEM XIII). Fundación Tekniker. Bilbao. Spain .2001 [28] Sánchez de Rivera, P. Modelos lineales y series temporales. Alianza Universidad Textos. Spain.1989 [29] Mason, R.L.; Gunst, R.F. and Hess, J.L. Statistical design and analysis of experiments.Willey and Sons.1989 [30] Lothar Sachs. Applied Statistics. A Handbook of Techniques. Second Edition. Spinger Verlag. 1984 [31] Nordin, A.; Eriksson L. and Öhman M. NO reduction in a fluidized bed combustor with primary measures and selective non-catalytic reduction: A screening study using statistical experimental designs Fuel. 1995. vol 74 pp 128-135. [32] Box, G.E.P. and Wilson, K.B. On the experimental attainment of optimun conditions. J. R. Stat. Soc. B. 1951 vol 13 pp 1-45 [33] Yunyan G. Design Method of Orthogonal and Regression Test. Publishing House of Shanghai Traffic University. Shanghai. China. 1980. [34] Kun-Li, W. Grey Systems. Modelling and Prediction. Yang's Scientific Press. 2004 [35] Xueming, W. and Jianjun, L. Simple Course of Grey System Method. Publishing House of Chengdu Technology College. Chengdu. China. 1993. [36] Lin, Y.; Chen, M-y. and Liu, S. Theory of grey systems: capturing uncertainties of grey information. Kybernetes. Emerald Group Publishing Limited. 2004. vol 33-2 pp. 196218. [37] Míguez, J.L.; Porteiro, J.; Morán, J.C.; Granada, E. and López, L.M. Description of a pilot lignocellulosic pellets stove plant. Sixth European Conference on Industral Furnaces and Boilers. Estoril. Portugal. 2002 [38] AENOR.UNE 9-044-086. Spain. 1986 [39] Granada, E.; Morán, J.C., Porteiro, J.; Miguez, J.L. and Lareo, G.. Pilot lignocellulosic pellet stove plant. First World Conference on Pellets. Stockholm. Sweden. 2002. pp 147-151 [40] Houck, J. E.; Scott A. T.; Purvis, C.R. Kariher, P.H.; Crouch, J. and Van Buren, M.J. Low emission and high efficiency residential pellet-fired heaters. Proceedings of the Ninth Biennial Bioenergy Conference, Buffalo, NY -USA. 2000 [41] Morán, J.C.; Granada E.; Porteiro,J. and Míguez, J.L. Pellet Combustion in Stove: Performance and Emissions Statistical Approach. International Conference on Renewable Energy and Power Quality. Vigo. Spain. 2003. pp 122-126 [42] Morán, J.C.; Granada E.; Porteiro,J. and Míguez, J.L. Experimental modelling of a pilot lignocellulosic pellets stove plant. Biomass and Bioenergy. 2004. vol 27- 6 pp 577-583.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 501-547

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 18

NEGATIVE EMISSION BIOMASS TECHNOLOGIES IN * AN UNCERTAIN CLIMATE FUTURE Kenneth Möllersten1, Zuzana Chladná1,2, Miroslav Chladný3 and Michael Obersteiner1 1

International Institute for Applied Systems Analysis, Schlossplatz 1, A-2361 Laxenburg, Austria 2 Department of Applied Mathematics and Statistcs, Comenius University, 84248 Bratislava, Slovakia 3 Department of Computer Science, Comenius University, Bratislava, Slovakia

ABSTRACT Mitigation of and adaptation to climate change belong to the most pressing global challenges for the 21st century. Major mitigation options include improved energy efficiency, shifting towards less carbon-intensive fossil fuels, increased use of energy sources with near-zero emissions, such as renewables and nuclear, CO2 capture and permanent storage (CCS), and carbon sequestration by protection and enhancement of biological absorption capacity in forests and soils. Bioenergy is one of several energy sources which could provide society with energy services with near-zero emissions. Bioenergy has a unique feature, however, which distinguishes it from other low-emitting energy supply options, such as solar, wind, nuclear, and clean fossil energy technologies. Bioenergy conversion could be integrated with a process which separates carbon. If the biomass feedstock is sustainably produced and the separated carbon is subsequently isolated from the atmosphere for a very long time the entire process becomes a continuous carbon sink – in other words such technologies yield negative CO2 emissions. Negative emission biomass technologies can be centralised or distributed; Centralised negative emission biomass technologies, *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

502

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al. biomass energy with CO2 capture and storage (BECS), build on the conversion of biomass into energy carriers in centralised conversion plants integrated with CO2 capture. The captured CO2 is subsequently transported and stored in geological formations. Distributed negative emission biomass technologies are based on the production of longterm carbon-sequestering charcoal soil amendment, with or without co-production of biofuels. In this chapter a BECS implementation scenario study is presented. The study analyses investments in BECS in a pulp and paper mill environment. The investment analysis is carried out within a real options framework taking into account the potential revenue from trading generated emission allowances on a carbon market. Uncertainty is considered in the economic modelling through the use of stochastically correlated price processes of one input price (biomass) and two output prices (electricity and CO2 emission permits) that are consistent with shadow price trajectories of a large-scale global energy model. The results suggest that BECS can be economically feasible within approximately 40 years. The chapter also discusses Research and Development needs for better understanding of the future overall potential of negative emission biomass technology implementation.

LIST OF ACRONYMS ADt BIG BIG/CC BECS BLG BLG/CC CCS CHP COE GHG HRSG IPCC IPPM LBM MPM MWe ppm UNFCCC

Air-dry tonne Biomass integrated gasification Biomass integrated gasifier with combined cycle Biomass energy with CO2 capture and storage Black liquor gasification Black liquor integrated gasifier with combined cycle CO2 capture and storage Combined heat and power production Cost of electricity Greenhouse gas Heat recovery steam generator Intergovernmental panel on climate change Integrated pulp and paper mill Limited biomass model Market pulp mill Megawatt electric Parts per million United Nations framework convention on climate change

INTRODUCTION Evidence is mounting that human-induced increase in greenhouse gas (GHG) levels in the atmosphere is creating one of the direst environmental changes in human history. Since the pre-industrial era, the concentration of CO2 in the atmosphere has increased from 280 to

Negative Emission Biomass Technologies in an Uncertain Climate Future

503

over 370 parts per million by volume (ppmv), mainly as a result of the burning of fossil fuels and land clearing. The third assessment report of the Intergovernmental Panel on Climate Change (IPCC) states that most of the observed global warming over the last 50 years is likely to have been due to the increase in greenhouse gas (GHG) concentrations in the atmosphere [IPCC, 2001a]. The IPCC further concludes that the stabilisation of the atmospheric CO2 concentration requires CO2 emissions to eventually drop well below current levels. The required reduction in net emissions to the atmosphere can be realised by means of [IPCC, 2001b]: • • • • •

Demand reductions and/or effiency improvements Substitution among fossil fuels Switching to renewables or nuclear energy CO2 capture and storage (CCS) Carbon sink enhancement.

The IPCC [2001b] concludes that none of these measures alone can achieve the emission reductions required. Thus, a portfolio of technologies needs to be developed and adopted. The choice of a particular stabilization level from any given baseline significantly affects the technology portfolio required for achieving the necessary emissions mitigation. Generally, a wider range of technological measures and their widespread diffusion and more intensive use is required for stabilizing at 450 ppmv compared with stabilization at higher levels [Nakicenovic et al., 2002]. Technology plays a crucial role in addressing the challenge of long-term GHG abatement. One way of categorising technologies is to arrange them in the following three groups: Technologies that give rise to substantial net CO2 emissions and thus contribute to a build-up of GHGs in the atmosphere; Conventional energy technologies based on fossil fuels is the most obvious example of technologies that belong to this category. Technologies with near-zero net CO2 emissions, i.e. technologies that can be regarded as more or less climate-neutral; Renewables-based technologies such as solar and geothermal energy, hydro power, wind power, and sustainable bioenergy1 along with nuclear energy are well known examples of this category. Among possible energy technologies with near-zero emissions, fossil fuel utilisation in combination with so-called CO2 capture and storage (CCS) is receiving increasing recognition as a mitigation option.2 CCS is the separation of CO2 from anthropogenic sources, transport to a storage location, and isolation from the atmosphere, see Fig. 1. Fossil fuel utilisation with CCS is seen by many as a key future technological option to reduce CO2 emissions that can be attributed with a large global CO2 reduction potential. Technologies for CO2 capture and transportation are technically proven and commercially available, although there is a large potential for improvements both regarding efficiency and 1

Conventional bioenergy can be carbon-neutral or -positive depending on how land use is affected by the biomass source. If biomass extracted for energy purposes from an existing biomass stock is subsequently replanted, the same amount of carbon that is releases from combustion is extracted from the atmosphere through photosynthesis and the bioenergy system is generally CO2 neutral (for a detailed discussion see, for example, Schlamadinger et al., 1997). 2 See IPCC [2005] for a comprehensive overview.

504

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

costs. It remains to be established, however, criteria and standards for safe CO2 storage over very long time periods. Moreover estimates of regional and global storage potentials are still very uncertain (see, for example, Bradshaw et al., 2006). With respect to CCS, scientists, policy makers and industry have largely focussed on fossil fuel-based systems. However, it is important to note that capturing CO2 from an emission source and storing it away from the atmosphere has the same climate change-mitigating effect regardless if the CO2 capture is applied to emissions from fossil fuels or from biomass.

Figure 1. Schematic diagram of possible CCS systems (Source: IPCC, 2005).

Negative emission technologies, i.e. technologies that remove CO2 from the atmosphere on a net basis; Biomass conversion systems, with or without the generation of heat, power and/or fuels, could be combined with a process which separates carbon. If the biomass is produced in a sustainable way and the separated carbon is subsequently stored away from the atmosphere for a very long period of time, the entire process becomes a continuous carbon sink, i.e. negative emissions are yielded. Net lifecycle emissions would be strongly negative because the carbon in the biomass was originally extracted from the atmosphere via photosynthesis and carbon will be extracted again when new biomass replaces what has been harvested. This feature distinguishes biomass from other carbon-lean energy sources, such as solar, wind, geothermal or nuclear energy. In this chapter these technologies are referred to as “negative emission biomass technologies”. An other approach to yielding negative emissions would be air CO2 capture, whereby ambient CO2 is removed by passing a natural air flow

Negative Emission Biomass Technologies in an Uncertain Climate Future

505

over absorber surfaces [Lackner, 2003]. The CO2 concentration in ambient air is around a factor of 100 or more lower than in flue gas. The IPCC [2005] conclude that capturing CO2 from air by the growth of biomass and its use in industrial plants with CO2 capture is more cost effective and based on foreseeable technologies. The capture of CO2 from ambient air is not discussed further in this chapter. This chapter deals with negative emission biomass technologies, which can be divided into two sub groups; centralised or distributed: 1. Centralised negative emission biomass technologies build on the conversion of biomass into CO2-free or CO2-lean energy carriers in centralised conversion plants. CO2 is captured in the conversion plant and subsequently transported and stored in geological formations, see Fig.2. This is referred to as biomass energy with CO2 capture and storage (BECS). For economic reasons the CO2 capture, transportation and storage must be stationary and large-scale. 2. Distributed negative emission biomass technologies build on the conversion of biomass into long-term carbon-sequestering charcoal. The conversion may be combined with the production of biofuels. Carbon can be sequestered by utilizing the wood charcoal as a soil amendment in forests and arable lands to improve soil productivity.3

Figure 2. The principle of bioenergy with CO2 capture and storage (BECS).

3

In addition, the option of “remote sequestration” exists, i.e. biomass may be harvested and separately sequestered, for example by burying the trees (see Keith, 2001, for further discussion concerning this alternative).

506

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Early work on negative emission biomass technologies was presented by Seifritz [1993] and Williams [1998]. Seifritz [1993] suggested that massive implementation of charcoal sequestration in developing countries could be implemented to off-set fossil emissions from industrialised countries. It is argued in the paper that the strategy outlined would offer industrialised countries a relatively cheap way to pay the external costs of their GHG emissions. At the same time well-needed capital flows from industrialised to developing countries would be created. Williams [1998] analysed production of hydrogen from biomass combined with CCS. The author proposes that BECS could open up possibilities for achieving deep emission reductions globally even if certain countries are unable (or unwilling) to significantly reduce emissions. The solution, then, would be that negative emissions generated by one party could be used to permanently off-set emissions generated by another. Obersteiner et al. [2001] brought a new dimension to the discussion surrounding negative emissions by arguing that removal of CO2 from the atmosphere through massive deployment of BECS would make it possible for mankind to bring about a downwards movement of atmospheric CO2 levels. The negative emission feature could thus be used to off-set emissions from the past. Following the observations by Obersteiner et al. [2001], modelling exercises simulating the global energy system have been carried out in order to assess the potential of BECS on a global scale [Azar et al., 2006; Smith et al., 2006, Rao and Riahi; van Vuuren et al.]. The importance of considering BECS when judging the feasibility of ambitious GHG stabilisation targets was illuminated in a study by Azar et al. [2006] which concluded that the option of CCS applied to fossil fuels and bioenergy could reduce the cost of meeting a 350 ppm stabilisation targets by 75 % compared to a case where these technologies are unavailable. According to the study, the corresponding number when CCS was applied exclusively to fossil fuels was 45 %. A rapid downwards movement of atmospheric CO2 levels could prove to be a necessity of a mitigation strategy in conformity with the United Nations Framework Convention on Climate Change (UNFCCC) [Obersteiner et al., 2004]. The ultimate objective of the UNFCCC [1992] is a "Stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system”. This level should be achieved within a time frame sufficient to, inter alia, allow ecosystems to adapt naturally to climate change. According to Thomas et al. [2004], gradual climate change can be expected to commit 24 % of terrestrial species to extinction by 2050 due to rising temperatures (on the basis of mid-range climate-warming scenarios). Thomas et al. [2004] argue that returning to near pre-industrial global temperatures as quickly as possible could prevent much of the projected climate-related (and irreversible) extinction from being realized. In other words, negative emission biomass technologies allowing for a downward movement in atmospheric CO2 levels could be instrumental in achieving the ultimate objective of the UNFCCC. Negative emission biomass technologies also have the potential to bring new perspectives to environmental product branding. Möllersten and Yan [2001] identified clearly negative CO2 balances for BECS-based electricity production. Williams [1998] and Möllersten et al. [2003a] reported distinctly negative net life cycle CO2 emissions for biofuels (ethanol, 4

Net power output/lower heating value of fuel input. The lower heating value (LHV) is used as the basis for the calculations and numbers presented throughout this paper.

Negative Emission Biomass Technologies in an Uncertain Climate Future

507

hydrogen and methanol). One could thus consider negative emission biomass technologies to be a technological possibility enabling the generation of energy carriers and other products that are “greener than green”. Möllersten et al. [2006] and Day et al. [2005] reported clearly negative carbon balances with respect to the production of pulp/paper and hydrogen, respectively, when negative emission biomass technology is integrated in the production. Consequently, manufacturers could potentially co-generate specific products and atmospheric CO2 removal, thus adding a new dimension to environmentally friendly products. Day et al. [2005] put forward the novel idea that when a product has a negative carbon budget on a net basis, consumerism “becomes an agent of climate mitigation”. Negative emission biomass technologies have a potential role to play as a component of a carbon management technology portfolio – a role that could not easily be replaced by any other technology. However, the assessment of the individual technologies, economic analysis has mostly been restricted to estimates concerning key numbers such as the cost of electricity (COE) or the cost of CO2 captured. To our knowledge, no study has so far been carried out that assesses the commercial potential of an individual negative emission biomass technology taking into account the long-term development of energy and carbon markets. Furthermore, only few of many perceivable negative emission biomass technology system configurations (from biomass production to carbon storage) have been studied in any detail (for example, Williams, 1998; Möllersten et al., 2006; Okimori et al., 2003, Larson et al., 2006). One should note that negative emission biomass technologies form a technology cluster with a large number of perceivable system configurations. Only a few examples of these can be found scattered throughout the literature. This brings us to the aim and scope of this chapter; Firstly, a BECS implementation scenario study is presented. The study analyses investments in BECS under multiple uncertainties in a pulp and paper mill environment. An introduction to pulp and paper production and related energy issues of interest to the study is followed by a section where the mill environments (market pulp mills and integrated pulp and paper mills) and the selected integrated BECS systems are defined. The results from an analysis of the energetic performance of the selected BECS systems are presented as well as an analysis of overall CO2 balances. Uncertainty is considered in the economic modelling through the use of stochastically correlated price processes of one input price (biomass) and two output prices (electricity and CO2 emission allowances) that are consistent with shadow price trajectories of a large-scale global energy model. The investment analysis is carried out within a real options framework (see, for example, Dixit and Pindyck, 1994). A major advantage of this approach is the ability to analyze the value of flexibility and uncertainty. The aim of this study is to present a comprehensive modelling effort targeted at evaluating the commercial potential of a negative emission biomass technology. Secondly, having analysed the economic performance of BECS in a pulp and paper mill setting in some detail, we turn our focus to a broader discussion concerning negative emission biomass technologies. The aim is to identify important knowledge gaps and Research and Development challenges ahead of us.

508

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

EVALUATION OF CO2 CAPTURE AND STORAGE IN PULP AND PAPER MILLS UNDER ECONOMIC AND TECHNICAL UNCERTAINTY Combined Heat and Power Production in Pulp and Paper Mills Pulp is used as a raw material to produce paper and board. Pulp production starts with a fibre source, the prime source being trees. Wood pulp is made by a mechanical or a chemical pulping process, or a combination of these two pulping processes called semi-chemical pulping. Paper and board production can be integrated with pulp production in integrated pulp and paper mills. In many cases, however, the pulp is produced in market pulp mills and then transported to another site where the paper or board is produced. In chemical pulping wood chips are cooked in a solution of chemicals whereby the wood cellulose fibres, which are used for pulp production, are separated as the chemicals dissolve the lignin (delignification). A chemical pulping process called Kraft pulping accounts for around 70 % of pulp production worldwide [FAO, 2005]. The Kraft process generates a byproduct from fibre extraction known as black liquor, which is a mixture of lignin and inorganic chemicals. The dissolved lignin in the black liquor represents just over half of the biomass entering a Kraft pulp mill. In modern Kraft pulp mills the black liquor is burned in recovery boilers (RB), which recover important pulping chemicals and produce steam (by utilising the energy value of the dissolved biomass) that is fed to the mill combined heat and power (CHP) system. The efficient utilisation of the black liquor energy content can reduce the Kraft pulp and paper industry's reliance on fossil fuels. In the most energy efficient existing Kraft market pulp mills the fuel requirement for the CHP system, which satisfies the mill’s requirement for medium-pressure (MP) and low-pressure (LP) steam, is typically covered through black liquor and internally generated bark. In contrast, integrated pulp and paper mills and paper mills need to import fuels to satisfy the process steam demand. In nearly all Kraft pulp production fossil fuels are still used in lime kilns, although a limited number of kilns were converted to biofuels in the 1980’s [Siro, 1984]. Most pulp mills and all integrated mills rely on electricity import to cover the part of their electricity demand that is not covered by power generated with internal CHP. In existing Kraft pulp mills with modern CHP systems based on recovery boilers and biomass boilers with steam turbines, electrical efficiencies are fairly low (up to 15 %4) [Larson et al., 2000]. Improved overall energy efficiency and increased electrical efficiency could be accomplished through the introduction of black liquor integrated gasification combined cycle (BLG/CC) [Berglin et al., 1999; Larson et al., 1999, Larson et al., 2000; Maunsbach et al., 2001], which is a promising technology that is however not commercially available today. Larson et al. [1999] modelled the performance of black liquor and biomass integrated gasification combined cycle in a typical present-day North American pulp mill. The results predict electrical efficiencies around 28–29 %. Today, there is a trend in the pulp and paper industry towards a further “closing” of the process. This means minimising the amount of effluents together with reducing the need for additional raw materials and energy. Generally this will reduce the heat demand through improved heat integration. The higher power-to-heat ratio of BLG/CC compared to recovery

Negative Emission Biomass Technologies in an Uncertain Climate Future

509

boilers with steam turbines makes this development favourable for BLG/CC. Several studies have shown that CHP systems based on black liquor gasification (BLG) in mills with low energy requirements could turn the mills into net exporters of electricity (see, for example, Berglin et al., 1999; STFI, 2003).

Technologies for Reducing CO2 Emissions in Chemical Pulp and Paper Mills The awareness of carbon management is growing as the issue of anthropogenically induced climate change is gaining importance. Industry is looking for strategies, policies, and measures that could be adopted to address and reduce its GHG emissions and the pulp and paper industry is no exception, see, for example, Browne [2003], Bruce [2000] and Miner and Lucier [2004]. A large share of CO2 emissions that can be allocated to pulp and paper production are closely related to the energy utilisation in pulp and paper mills. Opportunities for CO2 reductions in Kraft pulp and paper mills can be organised in the following categories: • • •

Decreased specific energy utilization Fuel switching (to less carbon-intensive fossil fuels or biomass fuels) CO2 capture and storage (CCS)

Possible projects range from highly process-integrated, as in the case of energy conservation through the optimisation of black liquor evaporation to pure energy projects, such as switching from fossil fuels to bark for steam generation (see, for example, Martin et al., 2000; Upton and Mannisto, 2001; Möllersten, 2002). When fossil fuels have been phased out, the main remaining mitigation option is BECS. BECS would lead to avoided on-site emissions thus potentially generating eligible tradable emission permits for the mill owner.5 Several studies analysing the potential for implementing BECS in pulp and paper mills have been published [Ekström, 1997; Möllersten and Yan, 2001; Möllersten, 2002; Möllersten et al., 2003a-b; Möllersten et al., 2006]. Before the results from these studies are summarised a brief background on CCS is provided.

A Background on CO2 Capture and Storage (CCS) The principle of CCS is to prevent CO2 from escaping to the atmosphere by implementing technologies that separate, or “capture”, the CO2 from fuel conversion and store CO2 or carbon in some form for long periods of time. CO2 capture, transportation and storage technologies are feasible and technically proven. There is considerable experience accumulated in the chemical and petroleum industries for operating chemical reactors and absorption units used for the capture of CO2 as well as for CO2 transportation systems. CO2 is routinely separated today at some large industrial plants such as natural gas processing and ammonia production facilities, although these plants separate CO2 to meet process demands and not for storage [IPCC, 2005]. In its third assessment report, the IPCC [2001b] predicts that CCS technologies could give major contributions to CO2 abatement by 2020 provided that the integrity of storage can be guaranteed. 5

An accounting system must be in place that enables crediting of avoided emissions through BECS (see section XX for further discussion about issues related to the accounting of captured and stored biotic CO2).

510

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

CO2 Capture CO2 capture applied to emissions from biomass can be divided into process groups with respect to the capture technology that is used. The purpose of CO2 capture is to produce a concentrated stream which can readily be transported to a CO2 storage location. Here, we choose to focus the description on three main process groups, see Fig. 3. “Post-combustion capture”, the capture the CO2 from combustion products in flue gas, is the most conventional approach for CO2 capture. CO2 capture from process streams is an established concept that has achieved widespread industrial application. These applications have, however, focused on gas separation from high purity, high pressure streams where the energy penalties and cost for capture are moderately low. Post-combustion capture for CO2 abatement involves capturing CO2 from gas with low CO2 concentration. This means that a large amount of inert gas has to be treated which leads to a significant cost and efficiency penalty because of the size of any downstream scrubbing and heat recovery equipment, etc. A main challenge associated with post-combustion capture for abatement of GHG emissions is reducing costs and the amount of energy required for capture. “Pre-combustion capture” systems share a common objective: to produce a fuel stream that contains little, or none, of the carbon contained in a carbonaceous feedstock fuel. This approach necessarily involves the separation of CO2 at some point in the conversion process. The resulting H2-rich fuel can be fed to a hydrogen consuming process such as production of synthetic liquid fuels, oxidised in a fuel cell, or burned in the combustion chamber of a gas turbine to produce electricity. For solid fuels like biomass (or coal), the first step in a precombustion system is always a gasification process, by which the solid fuel is reacted with steam and/or oxygen to produce a fuel gas that contains large quantities of carbon monoxide (CO) and hydrogen (H2). The synthesis gas is cleaned and the CO would be reacted with steam (“CO/water-shift reaction”) to produce more H2 and CO2, thereby increasing the amount of CO2 available for capture.6 The separation of these two gases can be achieved with well known, commercial absorption-desorption methods, producing the CO2 stream suitable for storage. These technologies have a high strategic importance because their capability to deliver a suitable mix of electricity, hydrogen and lower carbon-containing fuels. Conversion of carbonaceous fuels to synthesis gas, CO/water-shift conversion and CO2 separation are well-known processes which could be applied in power stations. “Oxygen combustion” (often referred to as “oxy-fuel”) is an approach that builds on the combustion of a fuel using pure oxygen instead of air. The dilution of CO2 with nitrogen in the flue gases is thus avoided. Consequently, the flue gases will consist of CO2 and steam, which in turn enables the CO2 to be separated by condensing the steam. The oxygen combustion approach will require a considerable amount of energy for the production of pure oxygen. CO2 Transportation Because of the large volumes involved in CCS, pipelines are suitable for the transportation of CO2 to a storage location once it has been captured. Transport of CO2 can best be done at high pressure in the range of 80 to 140 bars. Compression and pipeline transport of CO2 is feasible and technically proven. Several million tons of CO2 are transported annually, mainly in the USA, over long distances on-shore in pipelines for use in the enhanced oil recovery 6

CO + H2Ovap → CO2 + H2 +44.5 MJ/Molco

Negative Emission Biomass Technologies in an Uncertain Climate Future

511

industry [IEA, 2002]. Economies-of-scale effects are significant on the cost of CO2 transportation by pipeline. This is illustrated in Fig. 4 which illustrates the costs of transportation and on-shore storage as a function of distance for a set of CO2 flow rates.7 The cost data presented in the diagram were generated using two independent cost models [IEA, 2002; Ogden, 2002]. This is a major reason why CCS is only feasible at larger scales.

Figure. 3. Schematic representation of capture systems.

Other means of transportation that can be used are motor carriers, railway and water carriers. Experiences from these means of transportation are mainly found in the food and brewery industry, and the amounts transported are in the range of some 100,000 tons of CO2 annually, which is much smaller than the amounts associated with CCS [Svensson et al., 2004]. Svensson et al. [2004] conclude that pipeline and water carriers and a combination of these are the only economically feasible alternatives.

CO2 Storage A key issue is where CO2 should be stored. The discussion on CO2 storage covers the injection of supercritical-state CO2 into underground geological formations or the deep oceans and technologies for converstion to stable carbonates or bicarbonates. Much further work is 7

CO2 injection is assumed to take place in CO2-retaining deep saline aquifers and the depth of the injection wells is 1000 m. Capital costs were annualised using an interest rate of 10 % and a plant life of 25 years. A capacity utilization of 90 % was assumed.

512

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Transportation cost [USD/t CO2]

required to investigate the permanent storage of CO2. Deep underground storage is regarded as the most mature storage option today [IPCC, 2005]. Candidate underground storage locations are exhausted natural gas and oil fields, not exhausted oil fields, unminable coal formations, and deep, saline water bearing formations, see Fig. 5. Several commercial projects involving the injection of CO2 into reservoirs where it displaces and mobilises oil (so-called enhanced oil recovery) are in commercial operation [IPCC, 2005]. Underground storage is characterised by minimum interference with other ecological systems and provision of storage for very long time periods whereas ocean storage has considerable uncertainties regarding potential environmental damage, especially effects on marine life due to increased acidity, and the long-term isolation of the CO2 [Falkowski et al., 2000; IPCC, 2005]. International monitoring of current storage projects will help to define criteria and standards for safe geological CO2 storage.

60 10 [kg/s]

50

20

40

40 60

30

10

20

20 40

10

60

0 0

200

400

600

800

1000

Distance [km] Figure 4. CO2 pipeline transportation costs according to two independent cost models by IEA [2002] (dotted line) and Ogden [2002] (full line).

CO2 could also be reacted with certain minerals and subsequently be stored as carbonates (stable products that are common in nature) [Lackner, 2003; DOE, 1999]. The mineralization option is, however, flawed by high costs and energy requirements with current best approaches [Baciocchi et al., 2006]. Improved methods for accelerating carbonation are needed. One major concern with underground storage is the possibility of leakage of the stored CO2. Leakage rates have to be very small for carbon capture and storage to play a large and meaningful role in global efforts to meet stringent climate targets.8 For more details on the leakage risk, see, for example, Ha-Duong and Keith [2003].

8

If CCS would become a major option for the abatement of CO2 emissions very large quantities of CO2 would be stored towards the end of this century. If, for example, 1 % of 600 GtC (in line with the quantity stored globally the year 2100 according to some modelling exercises) would leak from storage every year, total emissions from

Negative Emission Biomass Technologies in an Uncertain Climate Future

513

Figure 5. Overview of geological storage options (Source: IPCC, 2005).

Several estimates of the global underground storage potential have been carried out. In Table 1, the comprehensive assessment by the IPCC [2005] is reproduced. Table 1. Potential for geological carbon storage options Reservoir type

Lower estimate of storage capacity (GtCO2)

Upper estimate of storage Capacity (GtCO2)

Deep saline formations Oil and gas fields Unminable coal seams

1000 675 3 - 15

Uncertain, but possibly 10 000 900 200

Source: IPCC [2005] CO2 could also be reacted with certain minerals and subsequently be stored as carbonates (stable products that are common in nature) [Lackner, 2003; DOE, 1999]. The mineralization option is, however, flawed by high costs and energy requirements with current best approaches [Baciocchi et al., 2006]. Improved methods for accelerating carbonation are needed. One major concern with underground storage is the possibility of leakage of the stored CO2. Leakage rates have to be very small for carbon capture and storage to play a large and

leakage would amount to as much as 6 GtC/year, which is roughly equal to current total global CO2 emissions from fossil fuels.

514

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

meaningful role in global efforts to meet stringent climate targets.9 For more details on the leakage risk, see, for example, Ha-Duong and Keith [2003].

CO2 Capture and Storage in Pulp and Paper Mills The performance of BECS in pulp and paper mills has been the focus of some published works. Here, we first provide a summary of studies based on CHP systems of existingstandard market Kraft pulp mills followed by results concerning future-standard mills. Ekström et al. [1997] analysed post-combustion capture and oxygen combustion in pulp mill CHP systems based on recovery boilers and steam turbine technology. The investment cost required for the entire CHP system was estimated to double due to the introduction of CO2 capture. Möllersten and Yan [2001] analysed advanced BLG-based systems with CCS for coproduction of biomass-based transportation fuels, power, and heat. Clearly negative CO2 balances were identified. Furthermore, the systems’ rates of return on investment (ROR) were compared in order to determine the price of CO2 emission permits that would justify the extra costs required for the more advanced and costly systems compared to a reference case based on recovery boiler and steam turbine technology (without CCS). The analysis considered potential incomes for both energy products delivered and from trading of CO2 emission permits. It was found that, assuming a 700 km CO2 transportation requirement from the pulp mill to the injection site, a price of approximately 60 $/tCO2 justified the introduction of the advanced systems with CCS.10 Möllersten et al. [2003] evaluated CO2 abatement potentials of CCS in CHP systems based on recovery boilers and pressurised BLG/CC, respectively. The largest abatement potential found was for post-combustion CO2 capture by chemical absorption from recovery boiler and bark boiler flue gases. For the analysed BLG/CC systems with pre-combustion capture, the analysis was restricted to partial CO2 capture without a CO/water-shift reaction prior to the CO2 separation by chemical absorption. According to the study BLG/CC with partial pre-combustion CO2 capture features a higher electrical efficiency compared to boilers with post-combustion capture but a significantly lower overall CO2 abatement potential. One BECS study has been published that takes into consideration the potential of energy efficiency improvements in pulp and paper mills. Möllersten et al. [2006] investigated the integration of CHP systems with CCS in predicted future market pulp mill and integrated pulp and paper mill environments (“reference mills”) with considerably lower process steam demand than existing-standard mills. The studies considered three main types of CHP systems with CO2 capture: post-combustion capture from recovery boiler flue gases, pre-combustion capture in BLG/CC systems without CO/water-shift enhancement, and precombustion in BLG/CC systems with CO/water-shift enhancement. A few conclusions can be drawn from a comparison of the analysed systems; Considerably less CO2 is captured in BLG/CC systems without a CO/water-shift reaction than in the other two systems considered in the study. Meanwhile, the CO2 capture level in BLG/CC systems with CO-water shift 9

10

If CCS would become a major option for the abatement of CO2 emissions very large quantities of CO2 would be stored towards the end of this century. If, for example, 1 % of 600 GtC (in line with the quantity stored globally the year 2100 according to some modelling exercises) would leak from storage every year, total emissions from leakage would amount to as much as 6 GtC/year, which is roughly equal to current total global CO2 emissions from fossil fuels. Note that the method used for the evaluation is not consistent with the design of CO2 accounting principles in emerging emission trading systems, e.g. the EU Greenhouse Gas Emission Trading Scheme (EU-ETS).

Negative Emission Biomass Technologies in an Uncertain Climate Future

515

reaction and post-combustion capture from recovery boiler flue gases were approximately the same. The studies further showed that CO2 capture systems based on pre-combustion capture have advantages compared to post-combustion capture systems in terms of higher electrical efficiency and lower biomass fuel consumption. Furthermore, estimated costs of CO2 capture can be reported from the studies mentioned above. Möllersten et al. [2003a] estimated the cost of CO2 capture to 34 USD/tCO2 for the case of post-combustion capture. For pre-combustion capture the costs in the range 22 and 34 USD/tCO2 were estimated [Möllersten et al, 2003; Möllersten et al., 2006].

Economic Evaluation Under Uncertainty Investment strategies in capital- and energy-intensive industries, like the pulp and paper industry, are driven by long-run price signals and their respective uncertainties. The implementation of climate policies is a major source of uncertainty for these industries. In the economic evaluation presented here, BLG-based11 CHP systems with a CCS option in predicted future-standard Kraft pulp and paper mills are evaluated against market conditions predicted by large scale global energy models.12 The economic feasibility of BECS is evaluated given correlated uncertainties of the biomass fuel, electricity and CO2 emission permit prices.13 Uncertainties in connection with climate change, and society’s response to the threat of climate change, along with the irreversible characteristics of many investments in mitigation technologies, create the conditions for decision-makers to value delaying investment decisions until more information is available. For the valuation of the investment decision we will build on principles of the real option theory (see, for example, Dixit and Pindyck, 1994; Mun, 2002). In the next section the evaluated CHP systems are defined. Based on simple process simulations, the performance of the evaluated systems are then presented, emphasising the rates of electricity production, CO2 capture and biomass consumption. Some effort is made to estimate and discuss overall CO2 balances of the evaluated systems, taking into account primary (“on-site”) and secondary (“off-site”) impacts on the emissions.14 The capital costs of the evaluated CHP systems are estimated in the following section. Next, the modelling 11

This case study does not address the issue whether recovery boilers or BLG can be expected to be the preferred technology in the future, but simply assumes the BLG as the baseline technology. As mentioned earlier, the currently applied process for recovery of chemicals and energy from black liquor is based on recovery boilers with Rankine steam cycle. If and when BLG will be introduced in reality is dependent on a number of factors. The recovery boiler with Rankine steam cycle has some drawbacks; the thermal efficiency and power-to-heat ratio are low, capital cost is high, maintenance under corrosive conditions is complicated, and there is a risk of smelt/water explosions. BLG is addressing the majority of these drawbacks. In addition, BLG opens up the opportunities for tailoring the delignification process towards higher yields and/or improved pulp physical properties [Stigsson and Berglin, 1999]. The gasification technology development has taken place during the last two decades. A number of BLG technologies have been developed, e.g., the Chemrec, MTCI, and ABB processes. In an on-going 10 M€ project the Chemrec BLG technology is demonstrated in Sweden. 12 Clearly, the results have to be interpreted with consideration to implicit assumptions made thereby (the conditions underlying the scenario family chosen etc.). 13 Consider that it is probable that more stringent CO2 restrictions will not only lead to an increase in the CO2 price, but also lead to elevated electricity and biomass prices. One realises that as the rate of CO2 capture, electricity production, and biomass fuel consumption are correlated, this has implications for the economic feasibility of the technologies and their relative competitiveness. 14 However, it is only the direct impacts that matter to the economic evaluation, as the mill owner is only accounts for the “own” emissions within the framework of an emission trading system.

516

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

framework, that optimises the pulp mill or integrated pulp and paper mill owner's decisions, is described. Finally, the results of the economic evaluation are presented.

Definition of the Evaluated CHP Systems In this section, the mill environments and CHP system configurations evaluated in this study are defined. The systems’ energy performance and overall CO2 balances are presented.

Mill Environment The modelling of CHP systems in this study is carried out in two different mill environments: a market pulp mill (MPM) and an integrated pulp and paper mill (IPPM). The MPM is identical with the model MPM defined by the Swedish Ecocyclic Pulp Mill research programme [STFI, 2003]. In the MPM of the Ecocyclic Pulp Mill Programme, which is assumed to employ the best technologies available in commercial use in the late 1990’s in all departments of the mill, the required process steam is 11 GJ/ADt pulp (Air-Dry tonne pulp) a reduction by 24 % compared to the 1994 Swedish average. The MPM has a capacity to produce 1550 ADt pulp/d. The IPPM, defined by Berglin et al. [1999], is an extension of the Ecocyclic Pulp Mill programme MPM. The IPMM steam consumption is approximately 5 % lower than the average Swedish 1994 fine paper mill. The IPPM has a capacity to produce 1860 tonnes of paper/d. The process steam and electricity requirements of the MPM and IPPM are presented in Table 2. Table 2. Energy requirements of the considered mills Energy requirement (GJ/ADt end product) Electricity Medium pressure steam (12 bar) Low pressure steam (4 bar) a

Market pulp mill 2.5 4.3 5.7

Integrated pulp and paper milla 4.8 7.5 8.3

1.2 tonnes paper are produced for every air-dry tonne (ADt) pulp produced in the IPPM.

CHP System Configuration Table 3 summarizes the alternative CHP system configurations considered. All cases are based on a pressurised high-temperature, oxygen-blown black liquor gasifier. In all evaluated systems the synthesis gas is cooled in a quenching bath using the weak wash as coolant whereby the weak wash is evaporated using the sensible heat of the synthesis gas. The quenching adjusts the fraction of steam in the synthesis gas to ensure an adequate amount of water for a CO/water-shift reaction to proceed in a downstream CO/water-shift reactor. Prior to shift-conversion, further gas cleaning would be required in order to protect the shift reactor from contamination. Gas cleaning was not modelled in this study.

Negative Emission Biomass Technologies in an Uncertain Climate Future

517

Table 3. Summary of analysed CHP system configurations

a

Case

Black liquor conversion Gasifier

Biomass conversiona Gasifier

MPM/BLG1 MPM/BLG2 MPM/BLG3 IPPM/BLG1 IPPM/BLG2 IPPM/BLG3

X X X X X X

X X X X X X

CO2 caputreb None

No CO/watershift

CO/water-shift

X X X X X X

Defines the technology used when fuel in addition to black liquor is required to meet process steam demands. b CO2 capture from both black liquor and bark/woody biomass is considered when applicable.

In the MPM/BLG3 and IPPM/BLG3 cases the cooled synthesis gas is sent to a shift unit to adjust the CO/H2 ratio via a CO/water-shift reaction. The shift unit is divided into two stages in series: The first (high-temperature) stage the temperature ranges from 225○C to 470○C. Most of the shift reaction is accomplished in this reactor. After being cooled down to 225○C, the synthesis gas is sent into the second (low-temperature) stage. The heat released in the shift unit is recovered through generating MP steam, which can be made useful for pulp and paper production. Sensible heat (above 70○C) of the synthesis gas leaving the shift unit is used to generate LP steam and some feed water for the heat recovery steam generator (HRSG). Subsequently the synthesis gas is cooled to 33○C before entering the cleanup unit. In the MPM/BLG2, MPM/BLG3, IPPM/BLG2, and IPPM/BLG3 cases, CO2 separation is carried out in physical absorption units (meant to approximate the Selexol process) upstream from the gas turbine combustion chamber. The captured CO2 is subsequently compressed to 80 bar in a multi-stage intercooled compressor. After the clean-up section the synthesis gas is used to fuel a gas turbine for power generation.15 The exhaust gas from the gas turbine is recovered in a HRSG and the generated steam is used for process steam needs in the mills, either directly or via a back-pressure steam turbine which generates additional electricity. When additional fuel is required to satisfy the process steam demand a supplemental BIG/CC is considered (see Table 3). In the same way as stated above, the biomass gasifier is modelled as a “black box” The main assumptions of the CHP systems are given in Table 4. Table 5 presents the main characteristics of the CO2-lean fuel gas which is fed to the gas turbine after the physical CO2 absorption.

15

THERE ARE a few areas relating to the use of gas turbines in the studied systems where technical development is less advanced. BLG/CC without CO2 capture requires that burners be modified for low and medium heating value gases. Fortunately, BLG/CC can draw on burner development for other gasification applications (e.g. coal and biomass) [Stigsson and Berglin, 1999]. The use of decarbonised fuels in gas turbine systems presents new technical challenges. Further development of gas turbines that can operate at very high inlet temperatures will be necessary to enable the firing with hydrogen-rich gas.

518

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al. Table 4. Main assumptions for the evaluated CHP systems Gasifiers Cold gas efficiency (%) Synthesis gas properties Temperature (ºC) Pressure (bar) Composition (mol%) N2 CO CO2 H2O H2 H2S CH4

Black liquor 77

Biomass 77

Raw gas 950 32

After quench 211 25

Raw gas 900 27

After quench 209 25

0.2 29.5 14.6 22.0 31.1 1.5 1.1

0.1 13.5 6.7 64.3 14.2 0.7 0.5

0.2 30.0 24.2 15.9 24.1 0.0 5.6

0.1 13.0 10.4 63.7 10.4 0.0 2.4

Gas turbinea Turbine inlet temperature (ºC) Pressure ratio Mechanical efficiency (%) Isentropic efficiency, expander (%) Isentropic efficiency, compressor (%)

1250 17 98 92 87

Steam cycle Turbine inlet temperature (ºC) Turbine inlet pressure (bar) Mechanical efficiency (%) Isentropic efficiency, expander (%) (High pressure / Medium pressure) Pinch temperature difference of HRSG (ºC) Feed water temperature (ºC)

440 66 98 85 / 87 15 120

a

Commercial gas turbines come in well-defined sizes that cannot be changed. Here, however, we refer to a “generic” gas turbine with the characteristics in Tab. 3. Table 5. Characteristics of CO2-lean fuel gas to the gas turbinea Temperature (C) Pressure (bar) Composition (mol%) N2 CO CO2 H2O H2 H2S CH4 a

110 20 0.4 0.4 0.7 0.0 96.8 0.0 1.8

The composition is based on MPM/BLG2 and would vary slightly when gasified woody biomass is added as in IPPM3.

Negative Emission Biomass Technologies in an Uncertain Climate Future

519

CHP System Performance Simple simulations of CHP systems were carried out by using the ASPEN PLUS process simulator [AspenTech, 2003]. The entire energy system modelling and optimisation as well as the cost analysis could be done in much greater detail. We regarded, however, the detail of the energy system assessment in this case study to be sufficient for our level of systems analysis.16 The performance of the analysed CHP systems is summarised in Tables 6 (MPM cases) and 7 (IPPM cases). Table 6. Performance of the MPM CHP systems (Pulp production 1550 ADt/d) MPM/BLG1 Black liquor (MW) Bark and woody biomass (MW)

MPM/BLG2

MPM/BLG3

338 0

0

0

CO2 recovery (%)

0

31

90

CO2 capture rate (kg CO2/s)

0

10

27

MP steam to mill (12 bar-t/h) LP steam to mill (4.5 bar-t/h)

101 137

Power consumption for CO2 absorption (MW)

N.A.

2

4

Heat consumption for CO2 separation (MW)

N.A.

N.A

N.A.

CO2 compressor (MW)

N.A.

4

13

Air separation unit (ASU) (MW)

5

5

5

Others (MW)

10

10

10

GT output (MW)

100

99

93

ST output (MW)

21

16

10

Net electricity output (MW)

106

94

71

Mill electricity consumption (MW)

39

Electricity surplus (MW)

67

55

32

Electricity surplus (MWh/ADt pulp)

1.0

1.0

0.5

Electrical efficiency (%)

31 76

28 72

21 65

Internal power consumption

Total efficiency (%)

16

The gasification itself is not simulated since the purpose of this study is to investigate the energy system not the gasification. The black liquor gasifier is treated as a “black box” in the simulation. However, the energy and material balances in the gasification have been considered in the simulation and the composition of the synthesis gas generated is correct. Only the parts that interact with the rest of the system are included such as compressors, heat exchangers and steam generators where process steam is produced. A large fraction of the sulphur in the black liquor is converted to hydrogen sufide (H2S) and is recovered and turned into useful sulphur for economic reasons. The H2S is removed from the gas in an acid gas removal system downstream from the gasifier. Data on the gasifier were taken from Berglin et al. [1999]. The composition of the raw synthesis gas from the gasifier was obtained from Lindblom [2001]. The high-temperature stage of the shift reactor was modelled as adiabatic reactor and the second stage as a constant temperature reactor. In the same way as stated above, the capture unit is modelled as a “black box”. The work consumed for the physical absorption depends on the partial pressure of the CO2 in the gas mixture. The work required for operating the absorption plant was set to 0.14 MJ/kgCO2 captured.

520

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

The tables show the mill-integrated systems' performance with regard to fuel requirement, CO2 capture rate (when applicable), electricity production and overall energy efficiency. Note that in all cases the mills' process steam demand is satisfied precisely. All MPM CHP systems generate a net electricity surplus which allows for power export to the grid, although the electricity surplus drops when CO2 is captured. Moreover, we can see that process steam requirements are satisfied through the black liquor-based CHP system in the MPM cases. Also in the IPPM cases an electricity surplus is generated for all cases. In contrast to the MPM cases, however, the process steam requirements are not satisfied by the black liquor alone in any of the IPPM cases.17 Table 7. Performance of the IPPM CHP systems (Paper production 1860 ADt/d)a IPPM/BLG1

IPPM/BLG2

IPPM/BLG3

Black liquor (MW)

338

Bark and woody biomass (MW)

114

114

184

CO2 recovery (%)

0

33

90

CO2 capture rate (kg CO2/s)

0

14

45

MP steam to mill (12 bar-t/h)

176

LP steam to mill (4.5 bar-t/h)

200

Power consumption for CO2 absorption (MW)

N.A.

3

6

Heat consumption for CO2 absorption (MW)

N.A.

N.A.

N.A.

CO2 compressor (MW)

N.A.

6

20

Air separation unit (ASU) (MW)

6

6

7

Others (MW)

14

14

16

GT output (MW)

135

135

146

ST output (MW)

0

0

16

Net electricity output (MW)

115

107

113

Mill electricity consumption (MW)

74

Electricity surplus (MW)

42

33

39

Electricity surplus (MWh/ADt paper)

0.5

0.5

0.5

Electrical efficiency (%)

25

24

22

Total efficiency (%)

78

76

68

Internal power consumption

The results obtained from our simulation of BLG/CC are consistent with other studies evaluating BLG/CC (without CO2 capture) in the same mill environment. For example, Berglin et al. (1999) reported electrical efficiencies in the range 29-31 % and 25-27 % for market pulp mills and integrated pulp and paper mills, respectively. We are not aware of 17

We can also observe that CCS leads to a larger drop in electrical efficiency for the MPM than for the IPPM. This can be regarded as an effect of the system configurations chosen and that the systems were optimised with respect to the mills’ steam requirements rather than with respect to the total electrical efficiency. Note that an optimisation of a total energy system with the objective to maximise the electrical efficiency may lead to different results.

Negative Emission Biomass Technologies in an Uncertain Climate Future

521

published studies analysing CCS in similar mill environments that could be used for the sake of comparison.

Impact of CCS on Overall CO2 Emissions It is only the direct emissions from a mill, and the impact that CCS would have on these, that would be relevant in relation to an emission trading system. However, the overall impact on the emissions is relevant with respect to the environmental credibility of a mitigation option.18 In order to calculate the overall CO2 reduction impact of introducing CCS in pulp and paper mills we consider changes in on-site and off-site net emissions compared to the reference cases MPM1 and IPPM1 following the principles outlined below.19 The MPM/BLG3 and IPPM/BLG3 cases were selected for this analysis. The systems’ overall CO2 impact compared to the reference cases was assessed based on the performance of the systems presented in Tables 6 and 7. The only change in on-site emissions compared to the reference cases is the CO2 which is captured and put into long-term storage and thereby not allowed to reach the atmosphere. Regarding off-site emissions the following approach was applied: •

•

18

The mills’ electricity production from CHP is affected by introduction of CO2 capture. Reduced or increased electricity production will have an impact on the electricity balance between the mill and the grid. We assume that the marginal electricity supply which has to compensate for a change in the electricity balance comes from either natural gas-fired combined cycle (NGCC) power plants with a 60 % electrical efficiency20 or coal-fired condensing power plants with 40 % electrical efficiency. Extraction of biomass requires energy, which leads to net emissions of CO2 if fossil fuels are used. The internally generated black liquor available at the mills is a byproduct of the pulp production which has no alternative use. Consequently, the emissions due to the extraction of the black liquor biomass fraction should be allocated to the production of pulp and paper. Therefore, only emissions from the extraction of biomass required in excess of the available black liquor are allocated to the CHP systems in our analysis. It is reasonable to consider two alternative levels of CO2 emissions for biomass extraction. As a lower value, we considered data for unrefined forestry residues. Börjesson and Gustavsson [1996] estimate that 2.9 kg CO2 is emitted per GJ forestry residues extracted. The figure, based on Swedish conditions, includes 50 km transportation of the fuel. As a value on the high end we used data for dedicated biomass plantations. The result of a comprehensive environmental life cycle assessment of fuel supply from dedicated eucalyptus plantations shows that 21 kg CO2 is emitted per GJ biomass extracted [Dowaki et al., 2002].

Doubts are often expressed whether large-scale biomass systems can really mitigate climate change if life cycle emissions are taken into account, such as CO2 emitted during biomass cultivation, harvesting, transportation and processing. Similar doubts have been expressed concerning BECS. 19 Note that all CHP systems are analysed in an environment of predicted future pulp and paper mills with considerably lower process steam demand than today’s existing mills. Thus the results presented do not reflect the advantages of the predicted pulp and paper mills that are per se energetically superior to existing mills. 20 Representing the best available technology for natural gas-fired combined cycle power plants today and in the near future.

522 •

•

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al. CO2 transportation by pipeline requires work for pressurisation. The initial pressurisation is considered in our analysis in that compression penalises the net power output of the mill CHP systems. The impact of emissions due to work required for booster compressors along the pipelines is regarded as negligible. CCS storage requires additional infrastructure such as pipelines. It is important to ensure that emissions are not moved from the tailpipe to the construction process to any significant extent. Comprehensive life cycle assessments of large-scale hydrogen production with CCS show only negligible CO2 emissions due to the construction of additional infrastructure [Strømman, 2003]. The impact of these emissions was regarded as negligible in this study. Table 8. Impact on CO2 emissions compared to reference case (marginal electricity from NGCC) Case

MPM2b IPPM3c

On-site emissions compared to reference (tCO2/ADt) CO2 capture and storage -1.51 -2.15

Off-site emissions compared to reference (tCO2/ADt) Marginal electricity production 0.18 0.02

Overall emissions compared to reference (tCO2/ADt)a

Limited biofuel model (tCO2/ADt)

-1.37/-1.37 -2.12/-2.05

-1.37d -1.75

Biomass fuel extractiona 0d 0.01/0.08

a

First value based on higher biomass extraction CO2 emission level/Second value based on lower biomass extraction CO2 emission level b MPM1 is reference case. CO2 emissions are per ton pulp. c IPPM1 is reference case. CO2 emissions are per ton paper. d No additional biomass fuel is extracted

Table 9. Impact on CO2 emissions compared to reference case (marginal electricity from coal-fired power plants) Case

MPM2b IPPM3c

On-site emissions compared to reference (tCO2/ADt) CO2 capture and storage -1.51 -2.15

Off-site emissions compared to reference (tCO2/ADt) Marginal electricity production 0.45 -0.05

CO2 emissions compared to reference (tCO2/ADt)a

Limited biofuel model (tCO2/ADt)

-1.06/-1.06 -2.19/-2.12

-1.06d -1.82

Biomass fuel extractiona 0d 0.01/0.08

a

First value based on higher biomass extraction CO2 emission level/Second value based on lower biomass extraction CO2 emission level. b MPM1 is reference case. CO2 emissions are per ton pulp. c IPPM1 is reference case. CO2 emissions are per ton paper. d No additional biomass fuel is extracted

The resulting overall CO2 impact of introducing CCS based on these assumptions is presented in Tables 8 and 9 (marginal electricity from NGCC and coal-fired power plants, respectively) as tonnes CO2 saved per air-dry tonne of final product (pulp or paper). In Tables

Negative Emission Biomass Technologies in an Uncertain Climate Future

523

8 and 9 one can also see the relative importance of the different emission sources. The results show that the CO2 penalty, which derives from off-site emissions is quite small. So far in this analysis the net emissions of CO2 from biomass fuel without CO2 capture have been considered to be zero. The real net emissions depend upon the carbon storage in soil and trees, the time span considered, the alternative use of the biofuel, and other factors [Schlamadinger et al., 1997]. Karlsson [2003] and Grönkvist et al. [2003] argue that the “wood-fuel efficiency” should be included into the analysis of biomass energy utilisation. If biofuel becomes a scarce resource, the limited amount of biofuel available would already be used in full to replace fossil fuels. Additional use of biofuels in, for example, a CHP plant would then have to be taken from another application. Subsequently, this other application has to resort to some other primary energy source, most probably a fossil fuel. A new biofuelled energy project causing additional biofuel consumption “on the margin” may thus lead to an increase in the use of fossil fuels somewhere else. To use the convention that biofuels emit no CO2 could be regarded as misleading in such a situation, because it would imply that it does not matter how efficiently the biomass fuel is utilised. Grönkvist et al. [2003] defined “the Limited Biofuel Model” (LBM) where scarcity of biomass fuel is explicitly considered by debiting biomass fuel use with the same CO2 emissions as coal. The direct comparison with coal comes naturally because coal can be replaced by biofuel in many applications without any extensive technical modifications. We applied the LBM, thus debiting additional biofuel consumed with 341 kgCO2/MWh fuel. The effect of using this alternative approach is presented in the last columns of Tables 8 and 9. The results show that using the biomass in mills with CCS adds leverage to the CO2 reduction potential of biomass, even when the limited biomass model is applied. Note that in this case emissions from the extraction of biomass were not accounted for because according to the LBM the biomass would be used for another application if it were not used for mill CHP requirements. This is an example of how scarcity of biomass could be included in an analysis based on the LBM. Region-specific conditions decide exactly at what rate biomass consumption should be debited with CO2 emissions. Debiting could be based on another fossil fuel than coal, or a mixture of different fuels. In conclusion, the results of Tables 6 - 9 show that substantial amounts of CO2 could be captured in market pulp and integrated pulp and paper mills while biomass-based electricity could simultaneously be delivered to the grid on a net basis.21 21

As a final note to the discussion concerning CO2 balances we would like to add a comment on the postcombustion capture options. As previously mentioned an alternative option for CO2 capture in pulp mills would be recovery boilers with post-combustion capture. For reference, applying the same methodology as for BLG-based systems, we made a simple estimation of CO2 balances of CCS integrated with recovery boilers with back-pressure steam cycle. Steam data representing the most advanced recovery boilers in use today were assumed. CO2 separation by chemical absorption using the chemical solvent Mono-Ethanolamine (MEA) was assumed. Energy required to regenerate the solvent and for stripping is typically 2.7-3.2 MJ/tCO2 for state-of-the-art, high efficiency coal-fired power stations. As the CO2 concentration of the flue gases from black-liquor and biomass-fired boilers are similar to that of a coal-fired boiler, steam consumption for regeneration was assumed to be 2,9 MJ/tCO2 (MP steam). If additional biomass was regarded as CO2-neutral, the results indicate that 10-40 % larger CO2 reductions could be achieved with post-combustion option. This effect can be attributed to the inferior energy efficiency of systems based on recovery boilers in the following way; less efficient post-combustion technology leads to more additional biomass needed to satisfy the mill steam demand. As a consequence more CO2 can be captured thus generating additional negative emissions. In this case it is illustrative to apply the limited biofuel model. If that is done, systems based on BLG are on par with, or slightly better than, the boiler based systems. This is explained by the superior energy efficiency of BECS based on BLG compared to BECS based on recovery boilers.

524

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Cost Analysis

CHP System Capital Costs Capital cost for the system components were first estimated by Möllersten et al. [2006]. The original cost data derives from several literature sources [Larson et al., 2000; STFI, 2000; Warnqvist, 2000; Brandberg et al., 2000; Williams, 2002; IEA, 2002; David and Herzog, 2000; Freund and Davison, 2002]. Tables 10 and 11 present the estimated capital costs. A scaling factor of 0.7 was used to adjust capital costs for size. An estimated initial accuracy of the source cost data is approximately 30%. Table 10. Estimated capital costs of MPM/BLG CHP systems [MUSD] Component Black liquor gasification island Biomass gasification island Shift reactor CO2 absorbtion Gas turbine HRSG Steam turbine CO2 compressor Total

MPM/BLG1 74

MPM/BLG2 74

MPM/BLG3 74

-

-

-

38 13 7 132

7 38 13 6 4 142

14 14 42a 13 4 10 167

a

The turbine is fuelled with predominantly H2. A 10% increase of the specific capital cost was assumed for the H2-fuelled gas turbine.

Table 11. Estimated capital costs of IPPM/BLG CHP systems [MUSD] Component Black liquor gasification island Biomass gasification island Shift reactor CO2 absorbtion Gas turbine HRSG Steam turbine CO2 compressor Total

IPPM/BLG1 74

IPPM/BLG2 74

IPPM/BLG3 74

53

53

75

47 16 190

8 47 16 5 203

20 20 57a 18 6 13 283

a

The turbine is fuelled with predominantly H2. A 10% increase of the specific capital cost was assumed for the H2-fuelled gas turbine.

The Cost of CO2 Transportation and Storage The cost of CO2 transportation and storage was determined using a model issued by the GHG Research and Development Programme of the International Energy Agency [IEA, 2002] (see model description earlier in this chapter). The assumed transportation cost as a function of transportation rate and distance is illustrated in Fig. 4. CO2 injection was assumed to take place in CO2-retaining deep saline formations with negligible seepage back to the

Negative Emission Biomass Technologies in an Uncertain Climate Future

525

atmosphere. 22 Pulp and paper mills are situated at sea or river ports enabling tanker transportation. Therefore, a cost ceiling of 20 USD/tCO2 for transportation and storage was assumed representing the cost for long-distance tanker transportation (over 2000 km) [IPCC, 2005]. Longer transportation distances were regarded as unrealistic.

Modelling Framework This section develops a model for evaluating the complex capital budgeting problem, which is analysed in this case study.23 The model applied for optimising the pulp mill owner's decisions considers two real options. The first belongs to the category of capital options: an option to invest in building a new module (system). There are two possibilities concerning how the new module can be built. The owner can either build an entirely new module or, if at least one system has already been built, invest only in the components that are necessary to add to an already built module. The second option deals with the mill's operating strategy. Once a certain module has been built, we assume that the pulp mill owner has the flexibility to activate and deactivate it again if this behaviour is found to be profitable. We call this kind of option a “switch option”. We will discuss these options in detail in the forthcoming paragraphs.

Actions In the search for the optimal strategy of the pulp mill operation we consider a time horizon T = 50 years. In our model the pulp mill owner has an option to change the operation plan each year (that is 50 times altogether). The equipment used in the pulp or pulp and paper mill for the combined CHP/chemical recovery process (BLG island and supporting elements) reaches the end of its economic life time after Tretire = 25 years, that is there are at least two 25-year periods (called CHP plant periods) to be taken into account. However, if the pulp mill owner finds it optimal, the retirement of the equipment can be advanced, thus leading to a scenario with more than two CHP plant periods within the time horizon T.24

22

Note that our cost assessment only considers dedicated single pipelines for each project. If a CO2 grid with trunk pipelines would become a reality (similar to the case of natural gas) thus allowing numerous CO2-emitting point sources to be connected to a CO2 transport network, the average scale of the transportation system would increase thereby reducing the average cost per ton CO2 transported. Under such circumstances, transportation by pipeline could be done over much longer distances before tanker transportation became the economically preferred option. It might be mentioned that the IEA (2002) estimated the cost of transporting CO2 5000 km in large-diameter pipelines at 25 USD/tCO2 (IEA, 2002), which gives some idea of the cost level that could be expected with a large-scale transportation system. 23 For a comprehensive discussion on the model setup, please refer to Chladna et al. [2004]. 24 To abandon the current equipment before the actual retirement time elapses could be a reasonable decision if, e.g., the prices that the pulp mill owner takes into consideration change suddenly and the equipment is near the end of its economic life time. Then the owner faces the following three options in our model: 1. To ignore the sudden change in prices, wait until the equipment reaches the end of its economic life time and then invest in new equipment; 2. To add to the existing equipment immediately; such investments must, however, be fully renewed in a few years when the original equipment reaches the end of its economic life time; 3. The existing equipment can be retired prematurely and investment made immediately in equipment upgrade (that is, the owner has an option to enforce the start of a new CHP plant period). When compared to option (2), such investment need not be renewed in a few years (which is an advantage), albeit there are losses due to premature retirement of the old equipment (which is a disadvantage). It is reasonable that, in certain cases, the third option becomes superior which justifies our decision to consider it.

526

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

There are three types of BLG-based modules (denoted by BLG1, BLG2, and BLG3), which we are interested in. Since exactly one module must be running in each year, the pulp mill owner deals with the following options every year: •

•

•

To invest in and build a new module, which has not yet been built in the current CHP plant period. This option will be called a capital investment action. Note that this action is a necessary choice in the first year of the respective period; To switch to using a module that has been built earlier in the current CHP plant period, but then (as perhaps temporarily economically inefficient) deactivated. This, in fact, means that an existing module will be re-activated. This option will be called a switch action; and To stay with the module that has been active so far. This option will be called a stay action.25

We assume that each BLG module is actually organized into components, which can be reused. For example, if the pulp mill exploited the module BLG1 so far and the owner now decides to upgrade to the BLG3 module, the components that the modules BLG1 and BLG3 share in common have not to be built: the respective components from BLG1 can be reused within BLG3. However, if the owner terminates the current period (either because the retirement time Tretire elapses or because she just finds it optimal), no reuse is possible anymore; all components must be built again from the beginning (and can be used or reused for the next Tretire years).

Costs Naturally, taking each of the above actions induces costs that we must consider. Therefore, we assume that there is a cost function c(.), set up as an input parameter of the model that prescribes the necessary investment for the respective actions:

•

c(m1| M): the cost of a capital investment action. With this action we build a new module m1 assuming that each module in the set M has already been built earlier in the current CHP plant period. Therefore, the components of modules in M can now be reused. The module m1 requires components of two kinds: the units that are not present in any module in M (i.e., the components that must be constructed for the first time in the current CHP plant period) and the components that are contained in some module of M (i.e., the units that will be reused). The cost c(m1| M) (also known as capital investment cost) comprises not only the investment in the former units but also the costs associated with repetitive implementation of latter ones; c(m1 → m2 ) : the cost of a switch action, when switching from using module m1 to

•

using the existing module m2;26 c(stay) = 0: the cost of the stay action, as it requires no investments.

•

25 26

Performing two actions in the same year is not allowed however. The switching costs were defined to be 15% of the capital cost for components that are switched on from an offstate and 10% of the capital costs for components that are switched off into off-state.

Negative Emission Biomass Technologies in an Uncertain Climate Future

527

As there are three BLG-based types of modules, only 6 values must be specified to cover all possible switch actions. Similarly, only 12 values must be specified to cover all possible capital investment actions. Note that these costs are not the only ones associated with the module operation (see, for example, operational costs, transportation costs, etc., which we introduce later). We assume that the structure of function c is “reasonable”. In particular, we assume that: •

•

Switching to the usage of an existing module cannot be more expensive than building that module for the first time, under otherwise identical circumstances, in the same year (that is, re-activation of a module is a reasonable choice); Capital investment is not made into a module if the same module was built earlier in the same CHP plant period (that is, the reactivation of a module component is always cheaper than its construction “from scratch”).

Learning In our model we consider technological learning by introducing the learning rate R. This means that any capital investment cost or switch cost, or any other technology-related cost has to be decreased by the factor 1/(1 + R)t, if associated with year t. Note, that there is a significant difference between the effect of learning and the effect of discounting: Learning means real reduction in costs, while discounting is just projection of the same amount of money in time. The time 0 cost of investment c performed in the year t is therefore:

c

(1 + r ) (1 + R )t t

that is, both effects apply.

Price Processes So far we have considered the module set up. However, once the module is active it operates and perhaps produces profit (i.e., operational profit). In our model, the operational e profit is governed by price processes varying in time: the electricity price pt , the biomass b

c

price p t , and the CO2 price pt . This introduces uncertainty into the model as the above prices are generated/simulated (see the subsequent section on generating price processes for details of the price processes simulation).

Operational Profit When the module m operates, it produces a fixed amount qc(m) of captured CO2 and surplus electric power, qe(m), on an annual basis. It is assumed that the captured CO2 is sold at the mill fence to an external undertaker that carries out the transporting and storage activities and, furthermore, that the mill owner and the CO2 undertaker share equally the revenue which is the income from selling CO2 emission permits reduced by the cost for transportation and storage. Depending on prices, the income for the pulp mill owner in the year t is calculated as follows:

528

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

0.5 × ptc × q c (m) +

pte × q e (m) .

(1)

The associated production costs in the year t are:

ptb × qb (m) +

c oper (m)

(1 + R )t

+

0.5 × c trans (m, d ) × qc (m)

(1 + R )t

,

(2)

where qb denotes the biomass requirement (additional to black liquor) for the CHP process and coper is the annual operational cost related to production. Furthermore, we assume the fixed annual transportation and storage costs (ctrans) for each unit of CO2 captured given the transportation distance d. The difference of equations (1) and (2) will be called operational oper

profit pt

( m) .

The values qc, qe, qb, coper, ctrans, and d are input constants of the model. We will assume that the annual costs for operation of module m amounts to 4% of its full initial capital investment costs. Note that the last period may be subject to different treatment as is outlined below.

Special Case: The Last CHP Plant Period The last CHP plant period requires special treatment in computations, as it may be artificially and thus prematurely terminated when the simulation reaches the time horizon T. In such a case the capital investment originally meant for period of Tretire years is applied in an artificially shortened period, thus the corresponding action could be found to be suboptimal. Instead of approximating the future profit that one would obtain if operating the module behind the time horizon T (which is hard to estimate), we reduce the capital investment costs in the last CHP plant period by a factor relative to the length of the last period with respect to the full Tretire-long period. (For example, if the last CHP plant period is artificially terminated by a time horizon T after five years, all capital investment costs realized within this period are reduced by factor 5/25 = 20%.) Solution Techniques The presented study is an application of the real option theory,27 which is nowadays a widely accepted technique in the analysis of investment problems under uncertainty. In our analysis we assume that the uncertainty arises only from the fluctuation of the prices in time. Therefore we assume that the stochastic part of the operational profit corresponds to the underlying source of randomness (St). Further, we assume that it can be modeled as the geometric Brownian motion process28, which is a classical assumption in the real option analysis. Actually, we adopt the real option framework in discrete time and determine the optimal actions by building a decision tree. A first step in the real options analysis requires us to construct a binomial lattice, which represents the time development of the underlying process. A basic idea of the binomial 27 28

See, e.g., Dixit and Pindyck, (1994) - an excellent overview of the real options techniques e.g., we assume that the underlying process can be modelled as: dSt = μtdSt + σtdWt , where increment of the Wiener process. For more details see Dixit and Pindyck, (1994), page 68.

dW

is an

Negative Emission Biomass Technologies in an Uncertain Climate Future

529

lattice construction is to structure the stochastic process in the simple manner: it is assumed that the underlying process can either tick up or tick down at each time point. Further, the discretization is proposed in such a way that the number of time steps is sufficiently large in order to achieve the same results as in the continuous case. Once the binomial lattice has been constructed the optimal decisions can be determined for each decision node. We use the standard backward moving algorithm to solve the problem. At each node for which a feasible decision can be adopted, the optimal decision is given as a solution of the following optimization problem:

V (ω , t ) = max π (S t , a t ) + e − rΔt E (V ( f (ω , a t ), t + Δt )) , at

(3)

i.e., the maximization procedure selects that feasible action, which maximizes the sum of the current profit flow π plus the expected discounted value of the continuation profit. In Equation (3) St represents the underlying process, at the feasible action, and the symbol E stays for the expected value operator. For the real option analysis we assume that a sufficiently wide set of financial instrument exists, such that the standard replication argument known from the option theory can be applied. Under such a consideration expected value in the formula (3) is discounted using the risk free rate r. This also means, that the expected value is determined using the risk neutral probabilities with respect to the underlying binomial lattice. Finally, note that due to the structure of the problem space, the states ω, that is the nodes of our binomial lattice, are actually multidimensional, therefore optimization described by (3) is multidimensional as well: taking the action at corresponds hence to the change of state described by the transition function f.

Generating the Price Processes Price information has been taken from Riahi et al. [2003] and Nakicenovic and Riahi [2002]. Riahi et al. [2003] computed electricity and carbon prices as shadow prices in GHG stabilisation runs including CO2 capture technologies aiming at an atmospheric CO2 concentration of about 550 ppmv. Two stabilisation scenarios for each baseline were developed ― one assuming constant costs for capture technologies (A2-550s, B2-550s), and one including learning for capture technologies (A2-550t, B2-550t). All four stabilisation scenarios are based on iterated runs of the global optimisation framework MESSAGEMACRO [Messner and Schrattenholzer, 2000]. Global emissions peak at about 9 to 12 GtC around 2050 and then proceed to decline to slightly less than the 1990 emissions level (6 GtC) by 2100. These emissions profiles are similar to other emissions trajectories for 550 ppmv stabilisation cases found in the literature (for example, Wigley et al., 1996; Riahi and Roehrl, 2000).

CO2 Price The carbon value is an endogenous output calculated by the MESSAGE model. It can be interpreted either as a carbon tax or value of an emission permit that has to be introduced in a GHG-constrained world in order to meet the stabilisation target. In the stabilisation scenarios,

530

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

CO2 prices grow steadily from about 5 US$/tCO2 in 2020, to about 7–17 US$/tCO2 in 2050, to about 33–70 US$/tCO2 in 2070, and to about 110–136 US$/tCO2 in 2100. The sharp increase at the end of the century is partly due to discounting with a 5% annual rate. In order to constrain the range within which the prices are allowed to fluctuate we selected the A2550t price trajectory as an approximation of the upper boundary and the B2-550t price trajectory as an approximation of the lower boundary. Mainly due to discounting, but also due to population and GDP growth the CO2 price process is modelled to be time dependent. More precisely, the CO2 price trajectories are generated as follows:

= e c1 + c2t + ε tc ,

ptc

(4)

where ε t ~ N(0, (σ t ) 2). The parameters ci, i = 1, 2 were estimated to fit the carbon shadow c

c

prices by Riahi et al. [2003] as described above.

σ tc were conjectured based on subjective

judgment. Figure 6 shows the CO2 price trajectories produced by Equation (4) for 10 simulations. 70

60

50

40

30

20

10

0 0

5

10

15

20

25

30

35

40

45

50

Figure 6. CO2 price trajectories produced by 10 simulations.

Electricity Price Riahi et al. (2003) compute an electricity price increase of about 100% in the coming century due to the stabilisation constraint. Based on the expectation of the price increase we fitted the electricity price trajectory to the following equation:

pte

= e e1 + e2t + ε te

(5)

Negative Emission Biomass Technologies in an Uncertain Climate Future

531

where ε t ~ N(0, σ e ). The electricity price is modelled as a time dependent process. In fact, e

2

we use the same structure as in case of the CO2 price. In order to obtain the underlying data for the parameter estimation of the process (5) we altered the electricity shadow prices in OECD countries calculated by Riahi et al. (2004) such that the particular information from the Nordpool energy market is taken into the account. We see such an adjustment of the global OECD scenario as a necessary modification in order to predict the prices in particular countries. This is due to the fact that electricity cannot be freely traded among the countries (for example, because of limitations in transmission capacities.) More precisely, as the main underlying information for the parameters estimation we consider the historical evolution of the yearly electricity spot price averages in Sweden modified by trends of the electricity price in the OECD countries predicted by the MESSAGE model.

σ te and the correlation between

the CO2 and the electricity price, ρ(pc, pe), were conjectured based on subjective judgement. Figure 7 shows the electricity price trajectories produced by Equation (5) for 10 simulations. 30

28

26

24

22

20

18

16 0

5

10

15

20

25

30

35

40

45

50

Figure 7. Electricity price trajectories produced by 10 simulations.

Biomass Price The modelling of the biomass price follows on the results presented by Ådahl and Harvey [2004]. In this study the authors propose a price-setting model for biofuel that assumes a constant price ratio for biomass and electricity in the Nordic countries. Empirical evidence shows that the value of the ratio is close to 2.9. However, so far, biomass fuel markets have not faced the impact of regional and global biomass fuel trading. Biomass markets might become increasingly international with global trade, main importers of biomass being the OECD countries.

532

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Therefore, we suggest that the development of the biomass price in Sweden will be correlated to the average of Sweden and the OECD marginal electricity prices. The biomass prices are generated as:

= e b1 + b2t + ε tb ,

ptb

(6)

where ε t ~ N(0, (σ t ) 2). Similarly to the electricity price and the CO2 price, the biomass b

b

price is hence modelled as a time dependent process. The parameters bi, i = 1, 2 were estimated to fit the biomass prices generated by the procedure proposed above. σ t and the b

correlation between the CO2 price and the biomass price, ρ(pc, pb), were conjectured based on subjective judgment. Figure 8 shows the biomass price trajectories produced by Equation (6) for 10 simulations. 19 18 17 16 15 14 13 12 11 10 0

5

10

15

20

25

30

35

40

45

50

Figure 8. Biomass price trajectories produced by 10 simulations.

Results of the Economic Evaluation In order to analyze the potential of commercial use of the described MPM/BLG and IPPM/BLG modules, respectively, we have performed several numerical experiments. We have simulated the price trajectories and used the decision tree (see Section “Solution Techniques”) to find the optimal sequence of a pulp mill owner’s actions. We have run 1000

Negative Emission Biomass Technologies in an Uncertain Climate Future

533

simulations and considered the total time period to be 50 years (T = 50) starting in 202029, the retirement time of the equipment to be 25 years (Tretire = 25) and the risk free rate (r) to be 5%. Since our main interest is to find the expected optimal time to enter the carbon market, i.e., to start capturing CO2, the presentation of the results will be biased towards CO2 capture technology. The results will be presented so that the sensitivity of optimal commitment to technological learning assumptions and various transportation distances for CO2 can be assessed. We consider three different learning rates: 5%, 10% and 15%, respectively and three different CO2 transportation distances: 100 km, 400 km and 1000 km, respectively. It means that altogether we present the results of nine numerical experiments in each MPM/BLG and IPPM/BLG case.

Market Pulp Mill When dealing with MPM/BLG modules, the numerical experiments show that BLG1 is consistently chosen as the optimal module to be built in the first year. Module BLG3 is built in all nine cases, too. The building time of BLG3 is concentrated within the first eleven years of the second simulation period (i.e., within years 26-36). However, it is observed that BLG2 is not a competitive technology: in most cases we observe a direct switch from BLG1 to BLG3. Only in some cases BLG2 acts as a transitory technology. This phenomenon is significant only under the high learning rate. Simulations show that under the high learning rate in approximately 53-71% of simulations BLG2 is built in year 26 that is in the first year of the second simulation period. Another interesting aspect is the sensitivity analysis of the expected commitment time for BLG3 technology. Figure 9 presents the resulting frequency distributions with the emphasis on the effect of the assumed learning rate. Not surprisingly, irrespective of the length of the CO2 transportation distance, the higher the learning rate, the sooner the first introduction of BLG3 technology. MPM/BLG3, CO2 transportation distance 100 km learning 5% learning 10% learning 15%

600 500 400 300 200 100 0 no build

4

8

12

16

20

24

28

32

36

40

44

48

Figure 9. Continues.

29

The starting year 2020 has been chosen because the IPCC [2001b] predicts that CCS could give major contributions to CO2 abatement by 2020. It may, therefore, be reasonable to assume the existence of a readiness (technical and institutional) to implement CCS by that year.

534

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

MPM/BLG 3 , CO 2 transportation distance 400 km learning 5% learning 10% learning 15%

600 500 400 300 200 100 0 no build

4

8

12

16

20

24

28

32

36

40

44

48

MPM/BLG3, CO2 transportation distance 1000km learning 5% learning 10% learning 15%

600 500 400 300 200 100 0 no build

3

6

9

12

15

18

21

24

27

30

33

36

39

42

45

48

Figure 9. Frequency distribution of commitment time for MPM/BLG3 technology.

The spread of BLG3 commitment times caused by change in the learning rate varies from about 3.6 years for the case of short transportation distances (100 km) up to 5.9 years in the case of long transportation distance (1000 km). Sensitivity analysis with respect to CO2 transportation distance shows overall less variability. The results in form of expected commitment time are depicted in Figure 10. We observe that the transportation distance does not influence the expected commitment time significantly: the maximum30 spread induced by the increase of the CO2 transportation distance (from 100km to 1000km) is 2.4 years.

30

Reached under the learning rate (5%).

Negative Emission Biomass Technologies in an Uncertain Climate Future

535

MPM/BLG3 40 35 30 25

distance 100km

20

distance 400km

15

distance 1000km

10 5 0 0.05

0.1

0.15

lear ning r at e

Figure 10. The expected commitment time for MPM/BLG3 technologies under different transportation distances subject to learning rate.

Integrated Pulp and Paper Mill Similarly to the MPM/BLG case, the module BLG1 is always built in the first year of the first investment period. However, in the IPPM case, BLG2 acts as an active transitory technology: no direct transition from BLG1 to BLG3 has been observed. Actually, in all simulations the modules have been built sequentially in the following order: first BLG1, then BLG2 and only afterwards BLG3. The exact timing of the occasion that BLG2 and BLG3 are built, respectively, depends on the learning rate and transportation distance setup. Figure 11 shows that the expected time of the first commitment of BLG2 varies between years 18 and 28. Not surprisingly, the earliest commitment is observed under the fastest learning (15%) and the shortest CO2 transportation distance (100 km). For higher learning rates (10% and 15%) the technology switch occurs always prior to first technical retirement and, most interestingly, earlier than in the MPM case. IPPM/BLG2 30 25 20 distance 100km 15

distance 400km distance 1000km

10 5 0 0.05

0.1

0.15

learning rate

Figure 11. The expected commitment time for IPPM/BLG2 technologies under different transportation distances subject to learning rate.

536

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

In contrast to this, for the slow learning assumption (5%), irrespective of the length of the transportation distance, the expected building time of BLG2 is concentrated within the first three years of the second investment period (26-28). Furthermore, we observe that for the slow learning rate (5%) and short transportation distance (100 km), the timing to build BLG2 is mostly dominated by the technical retirement restriction (i.e. lifetime of 25 years): the first build of BLG2 is postponed to the second investment period rather than being switched to just before the end of the first period. Figure 12 presents the frequency distribution of the first commitment time of BLG3 module. IPPM/BLG3, CO2 transportation distance 100km

learning 5% learning 10% learning 15%

500 400 300 200 100 0 no build

4

8

12

16

20

24

28

32

36

40

44

48

IP P M / BLG 3 , CO 2 transportation distance 400km

learning 5% learning 10% learning 15%

500 400 300 200 100 0 no build

5

10

15

20

25

30

35

40

45

50

IPPM/BLG3, CO2 transportation distance 1000km

learning 5% learning 10% learning 15%

400 350 300 250 200 150 100 50 0 no build

3

6

9

12

15

18

21

24

27

30

33

36

39

42

45

48

Figure 12. Frequency distribution of commitment time for IPPM/BLG3 technology

Negative Emission Biomass Technologies in an Uncertain Climate Future

537

The numerical results show that BLG3 is always committed after the first technical retirement: for all setup combinations of learning rate and transportation distance the optimal building time falls within the range of years 27-35. As we already mentioned, under all nine studied setups BLG3 is built subsequent to the BLG2 module. We calculated the transit period to be approximately 6-9 years. A variance of the first commitment time of BLG3 is sensitive neither to the transportation distance nor to the learning rate. However, the variance is greater than in the MPM/BLG3 case. Moreover, Figure 11 and Figure 13 group the expected time of the first commitment of the BLG2 and BLG3 so that the sensitivity with respect to the CO2 transportation distance can be easily derived. The results show that in both cases commitment times are rather insensitive to the transportation distance of CO2 ― despite the rather high assumption on the transportation price of about 20 €/tCO2 for longer distances. The maximum spread of the expected commitment times caused by the change in the transportation distance is only 1.65 years.31 On the other hand, the learning assumptions make a notable difference: in general, the higher the learning rate, the sooner the introduction of CCS. The spread caused by the increase of the learning rate varies from 8 up to 8.6 in case of BLG2 and from 5.1 up to 7.2 years in BLG3 case, respectively. Overall, the sensitivity analysis of the MPM/BLG and IPPM/BLG results feature similar characteristics. IPPM/BLG3 40 35 30 25

distance 100km

20

distance 400km

15

distance 1000km

10 5 0

0.05

0.1

0.15

learning rate

Figure 13. The expected commitment time for IPPM/BLG3 technologies under different transportation distances subject to learning rate.

31

Considering the positive economies-of-scale in CO2 transportation, the results' insensitive character (for the MPM and IPPM cases) to the distance from the point of capture to the point of storage comes with some surprise since the analysed biomass-based energy conversion systems are of modest scale with associated modest rates of CO2 generated.

538

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

DISCUSSION AND OUTLOOK ON NEGATIVE EMISSION BIOMASS TECHNOLOGIES In this chapter we have evaluated investments in BECS in a pulp and paper mill context. The analysis considered biomass-based systems with a fuel input between approximately 300 to 500 MW. Our analysis shows that in the range of scales considered, BECS can be economically feasible within approximately 40 years given carbon, electricity and biomass prices predicted by one of the leading IPCC global energy scenario models.32 However, it has to be noted that market imperfections may yield much higher prices than those predicted in an idealized model world, which would suggest that price signals triggering BECS could emerge earlier. As mentioned in the introduction to this chapter, an overwhelming majority of all work done related to CCS has focused on applications to emissions from fossil fuels. One likely explanation is the issue of scale. Sizes of coal-fired power plants in the range from 500 to over 1000 MWe are currently common in the world and natural gas-fired power plants with unit sizes of several hundred MWe are abundant. Pulp and paper mills, along with sugar canebased ethanol mills, are among the few industries where large-scale bioenergy conversion takes place in the world today, thus providing potential niche markets for BECS implementation [Möllersten et al., 2003a].33 The fact remains, however, that most of the biomass in the world is currently used in traditional ways at small scales not compatible with CCS for economic reasons. Nevertheless, the IPCC [2005] states that “it is perfectly conceivable that these technologies [BECS, authors’ remark] might play a significant role by 2050 and produce negative emissions across the full technology chain”. This would require a significant increase in large-scale biomass energy conversion on a global scale compared to the present situation. In the future, one might expect small-scale uses of biomass to continue to remain significant, but larger-sized operations compatible with CCS is likely to become increasingly important with increasing interest in bioenergy implementation. In order to justify this statement we refer to Table 12 that summarises the results from several studies concerning economically optimal scales of bioenergy conversion. Given the rather large optimal scales, a growth in commercial bioenergy implementation could very well bring about an increase in large-scale bioenergy conversion, which would imply improved opportunities for BECS. The potential for converting biomass into long-term carbon-sequestering charcoal (see the introduction to this chapter) adds to the total theoretical potential of negative emission biomass technologies. Whereas BECS is a centralised technology, charcoal carbon sequestration is feasible for small-scale distributed biomass based energy generation and carbon sequestration. In this sense the two technology groups are complementary. It is also worth noting that while BECS requires advanced technologies and technological know-how,

32

In the pulp and paper mills investigated in this chapter approximately half of the biomass delivered to the mill ends up as fuel and the other half as fibre. Moreover, large chemical pulp mills of today produce in the range 1500–3000 tons of pulp per day, i.e. up to twice as much as the mills studied here. Thus, it has already been proven that it is logistically feasible to operate facilities that receive biomass at very high rates - in the order of 1000 MW. 33 In addition, in some countries (e.g. Sweden and Finland) large heating and CHP plants connected to district heating networks are fired with biomass.

Negative Emission Biomass Technologies in an Uncertain Climate Future

539

charcoal carbon sequestration can be based on relatively simple and well-known technologies and practices. Table 12. Estimates of optimum scales of bioenergy conversiona Type of plant/mode(s) of operation BIG/CC/Power production

BIG/CC/Power production

BIG/CC/Power production

Fuel supply

Optimum scale [MW]

Author(s)

Dedicated energy crop (switchgrass) or plantation wood Dedicated energy crop (switchgrass) or plantation wood Dedicated energy crops (williow)/Municipal RDF

230-320 MWe

Larson and Marrison [1997]

110-142 MWe

Larson and Marrison [1997]

COE declined with increasing scale within the studied capacity range from 51-215 MWe Up-scaling decreases costs per unit primary energy saved within the studied range 10-200 MW thermal input Up-scaling decreases costs per unit primary energy saved within the studied range 10-300 MW thermal input 137/450/900 MWe

Faij et al. [1998]

Fluidised bed combustion and grate firing/ Power production and CHP

Forestry residues/industrial waste wood

Fluidised bed gasification/ Power production and CHP

Forestry residues/industrial waste wood

Steam turbine power plants/ Power production Bioethanol plants

Forestry residues/ agricultural residues/whole forest Cane sugar/Cane sugar plus sorghum Dedicated energy crops

Electricity production/ Synthetic fuel production Methanol, ethanol, hydrogen and FischerTropsch diesel plants

Woody biomass

140-200 MW thermal output Approximately 1000 Mw thermal input Strong economies-of-scale effects between 80 and 400 MW thermal input. Production costs continue to decrease between 400 and 2000 MW thermal input albeit at a slower rate.

Dornburg and Faaij [2001]

Dornburg and Faaij [2001]

Kumar et al. [2002] Nguyen and Prince [1996] Greene [2004] Hamelinck [2004]

a

All studies presented in the table consider scale effects in the cost of biomass feedstock supply. All authors except Hamenlinck [2004] assumed the biomass to be produced near the conversion facility. Hamenlinck [2004] considered long-distance international biomass transportation from biomass production areas to energy import regions. The analysis, therefore, assumes a flat biomass feedstock cost. The investigation shows that transcontinental bioenergy trade is possible while maintaining competitive biomass prices and modest energy losses. Hamenlinck [2004] concludes that the energy requirement to deliver solid biomass from South America to Europe is 1.2 – 1.3 MJprimary/MJdelivered. As a comparison, the corresponding energy requirement for coal is approximately 1.1 MJ/MJ.

Taking these considerations into account, it is perceivable that negative emission biomass technologies could turn out to be an important option in a wider portfolio of GHG abatement technologies allowing for long-term energy security and sustainable climate management. The uncertainties surrounding the competitiveness of negative emission biomass technologies is however larger than that of conventional mitigation technologies Therefore, we will conclude with a discussion concerning areas where Research and Development could fill important knowledge gaps. Note that the discussion that follows focuses on issues specific to

540

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

negative emission biomass technologies. Research and Development in the general area of CCS and energy technology, such as methods for biomass production, technologies for transportation and storage of CO2, issues surrounding the integrity of CO2 storage options, CO2 separation technologies, thermochemical conversion of solid fuels and advanced bioenergy conversion will be beneficial also for the development of negative emission biomass technologies.

Important Areas of Research Technologies and Systems

BECS Although a broad variety BECS “technology chains” can be envisaged, linking different biomass feedstock sources, biomass pre-treatment and transportation systems, bioenergy conversion and CCS options, only few BECS “technology chains” have been studied in any detail. Regional solutions that encompass biomass production, bioenergy conversion and CCS in one region are one possibility, but another option would be to integrate long-distance transportation of solid biomass or refined fuels at various stages along the technology chains. For example, BECS could involve large-scale production of transport fuel with CCS in biomass-producing regions and subsequent export of the fuels produced to energy-importing regions. BECS could also comprise biofuel export from biomass-producing regions to largescale energy conversion facilities in energy-importing regions where CCS would then take place. One further option would be to combine the latter two examples. A number of competing bioenergy technologies will likely be available. Today’s biomass to electricity or CHP capacity is based on mature, direct-combustion boiler/steam turbine technology. Direct-fired combustion technologies for electricity or CHP production can be integrated with post-combustion or oxy-fuel CO2 capture. An important near-term low-cost option for the extended use of biomass is co-firing with coal in existing boilers, i.e. the practice of introducing biomass as a supplementary fuel in high efficiency boilers (see, for example, Veijonen et al, 2003). Due to scale advantages, co-firing is likely to be a costeffective BECS option in the more near-term [IPCC, 2005]. In addition, the flexibility to choose between fossil and a carbon neutral feedstock allows to hedge against uncertainties of future carbon market uncertainties. Another potentially attractive biomass to power or CHP option is based on gasification. Biomass integrated gasification/combined cycle (BIG/CC) systems would be expected to have thermal efficiencies nearly double those of directcombustion systems. Biomass gasification is, as mentioned earlier in this chapter, compatible with pre-combustion CO2 capture technology. Advanced biomass power systems based on gasification benefit from the substantial investments made in coal-based gasification combined cycle systems. Biomass gasification systems with CO2 capture will also be appropriate to provide fuel to fuel cell and hybrid fuel-cell/gas-turbine systems. A near-term opportunity for BIG/CC technology is in the forest products industry (more details are presented in the main section of this chapter). The few published studies concerning BECS in the power sector include Audus and Freund [2004], Hochenauer et al. [2004], Larson et al. [2005] and Rhodes and Keith [2005].

Negative Emission Biomass Technologies in an Uncertain Climate Future

541

Audus and Freund [2004] identified high costs for CCS with a 30 MWe BIG/CC power plant. CCS doubled the COE compared to a case without CO2 capture. The study clearly illustrates the high cost of CCS in small-scale applications. With respect to CHP several studies have been carried out analysing the energetic performance, carbon balances and costs of BECS in pulp and paper mills, considering several system configurations including post-combustion capture, pre-combustion capture and oxygen combustion. The main section of this chapter summarises studies in this area. Biomass-based liquid and gaseous fuels (“biofuels”) can be produced either via biochemical processes or gasification. Gas emissions from fermentation, the biochemical process that converts sugars into ethanol, are concentrated CO2 and capture of CO2 is essentially an activity consisting of condensing-out water and compression. Methanol, hydrogen and so-called Fischer-Tropsch diesel are examples of biofuels that can be produced from biomass via gasification. Several production routes are possible. CO2 separation is part of all processes, which reduces the additional cost for employment of CCS. With respect to liquid fuels and hydrogen, Walsh [1993], Williams [1998] and Möllersten and Yan [2001], Larson et al. [2005] presented studies investigating BECS-based production of methanol, hydrogen, and Fischer-Tropsch fuels. In addition to the publications mentioned above, Azar et al. [2006] estimate key performance figures of different BECS alternatives calculated on a consistent basis. Capital costs and energetic performance of biomass-based power, CHP, liquid fuel and hydrogen production with CCS are estimated, albeit on a highly aggregated level. It is evident that these BECS studies provide very limited information regarding the potential of BECS in the power sector, for CHP outside the pulp and paper sector as well as in relation to the full range of biofuel production options. The efficiency and economic performance of BECS needs to be studied over the full range of fuel supply options, bioenergy conversion technologies and CO2 capture methods. The potential improvement through integrated process configurations and the development of new technologies needs to be studied further. Furthermore, a geographically explicit analysis of the relationship between possible sources and storage locations would add valuable information. Optimal scales for BECS have not been analysed. Note that the estimated optimum facility sizes for bioenergy conversion quoted in Table 12 are for applications without CCS. Because of positive economies-of-scale in CCS, combining bioenergy conversion with CO2 capture could be expected to lead to elevated optimum sizes. This issue would need to be further analysed.

Charcoal Carbon Sequestration Charcoal carbon sequestration technologies are based on the carbonisation of biomass materials to make charcoal which is subsequently put into repository or utilised for carbonsequestering purposes such as soil amendment in agriculture and forestry or even advanced building material. In its least advanced form, the simple transformation of organic carbon in biomass into inorganic carbon in the form of wood charcoal in some common type of furnace, the carbonisation process involves technologies that are readily available. With modern technologies hydrogen can be co-produced with charcoal using reforming technologies. The

542

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

hydrogen produced can be converted to synthetic liquid fuels or used in, for example, fuel cells..The charcoal can be further processed for its use as a soil amendment. Okimori et al. [2003] presented a case study based on carbonization of biomass residue and waste from tree plantations and pulp mills. Ogawa et al. [2005] analysed three cases to evaluate the potential of the charcoal carbon sequestration finding fixed carbon recovery ratios35 in the range 21 to 32 %. In a study by Day et al. [2005] a process has been proposed which co-produces a nitrogen-rich, slow-release charcoal fertiliser and hydrogen. Furthermore, the combination of the nitrogen-enriched charcoal production process with a chemical process that can directly convert CO2, SOx and NOx from fossil fuel plants to valuable fertilizers is proposed. The number of studies analysing process solutions is quite limited, which calls for futher Research and Development. Further work is also needed to reduce the uncertainty surrounding the stability of charcoal in soil. Ogawa et al. [2005] point out that there are only a few studies that clarify oxidation and degradation processes of carbon in charcoal.

Monitoring, Inventories and Accounting The requirement for explicit treatment of BECS with respect to accounting issues has been noted in some publications [IEA, 2004; IPCC, 2005; Grönkvist et al., 2006]. Grönkvist et al. [2006] propose desirable characteristics for an accounting approach for BECS and suggest that the desirable criteria can be achieved with an accounting approach that would allow for the assignment of “removal units” in a biotic CCS carbon pool for biomass CO2 collected. Ogawa et al. [2005] identify essential details of a method for accurately monitoring the course of charcoal carbon sequestration. In brief, the method recommends the monitoring of (i) the source, quantity and quality of biomass consumed in the carbonisation, (ii) carbonization method, charcoal yield, carbon composition in charcoal, external heating requirement, and (iii) the identification of non-fuel use of charcoal and related carbon stability aspects. Further work on monitoring, inventories and accounting is necessary if negative emission biomass technologies shall be incorporated in the accounting systems of international climate regimes. Global Potentials Some modelling using global energy system models has been performed with consideration taken to the possibilities of BECS (see the introduction to this chapter for some details and references). In relation to charcoal carbon sequestration, Lehman et al. [2005] estimated the total technical sequestration potential in the range 5.5 to 9.5 Pg C/year. Further modelling exercises including both BECS and charcoal sequestration technology are called for. Finally, it is important to point out that there are still a number of issues open with respect to the environmental and social sustainability of large-scale global applications of negative emission biomass technologies which go beyond the scope of this paper. Given that land markets are rather sticky, large infrastructure investments are necessary and diffusion 35

“Fixed carbon" indicates carbon in biochar. Fixed carbon (%) = 100 % - (moisture % + ash % + volatile %). Recovery of fixed carbon indicates the percentage of (Fixed carbon)/(Carbon from fuelwood).

Negative Emission Biomass Technologies in an Uncertain Climate Future

543

times will span several decades. In particular, it has been observed that large scale biomass plantations have faced political and social objection both in the developed and developing world. It can be concluded that as long as the high option value, the high environmental and potentially social values of negative emission biomass technologies are not recognized, it will fail to be developed adequately as a competitive mitigation and energy technology cluster. Furthermore, as already noted, technological spillovers can be expected from the implementation of CCS technologies with fossil fuels.

CONCLUSION This chapter has presented a biomass energy with CO2 capture and storage (BECS) implementation scenario study. Investments in BECS in a pulp and paper mill environment were analysed within a real options framework. Uncertainty was considered in the economic modelling through the use of stochastically correlated price processes of one input price (biomass) and two output prices (electricity and CO2 emission permits) that are consistent with shadow price trajectories of a large-scale global energy model. The analysis suggests that (in the range of scales considered), BECS can be economically feasible within approximately 40 years. Combined with a number of economic factors, the uncertainties on the competitiveness of the negative emission biomass technology cluster is larger than that of more conventional mitigation technologies. Therefore we provided a discussion concerning Research and Development efforts that would allow increasing certainty and buying down costs.

ACKNOWLEDGMENTS The authors gratefully acknowledge financial support from the Kempe foundations, the Swedish Association of Graduate Engineers, the IIASA Forestry program and the EU projectIntegrated Sink Enhancement Assessment (INSEA), contract number: SSPI-CT-2003/503614. The authors are also grateful to Andre Faaij and Yasuyuki Okimori for providing helpful comments on early drafts.

REFERENCES Ådahl, A.; Harvey, S. In: Process industry energy projects in a climate change conscious economy; Ådahl, A; PhD thesis, Chalmers University of Technology, Goteborg, 2004. AspenTech (2005). http://www.aspentech.com Audus, H.; Freund, P. In Proceedings 7th International Conference on Greenhouse Gas Control Technologies; Rubin, E. S.; Keith, D. W.; Gilboy, C. F.; Eds.; IEA Greenhouse Gas Programme: Cheltenham, 2004.

544

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Azar, Ch.; Lindgren, K.; Larson, E. D.; Möllersten, K. Climatic Change. 2006, 74, 47–79. Baciocchi, R.; Storti, G.; Mazzotti, M. Proceedings 8th Greenhouse Gas Control Technologies Conference, Trondheim, 2006. Berglin, N.; Eriksson, T.; Berntsson, T. Proceedings 2nd Johan Gullichsen Colloquium, 1999, 55-67. Börjesson, P.; Gustavsson, L. Energy. 1996, 21, 747-764. Bradshaw, J.; Bachu, S.; Bonijoly, D.; Burruss, R.; Holloway, S.; Christensen, N. P.; Mathiassen, O. M. Proceedings 8th Greenhouse Gas Control Technologies Conference, Trondheim, 2006. Brandberg, L.; Ekbom, T.; Hjerpe, C.-J.; Hjortsberg, H.; Landälv, I.; Sävbark, B. BioMeet: Planning of Biomass-Based Methanol Energy Combine - Trollhättan Region; Report no XVII/4.1030/Z/98/368; Ecotraffic Research and Development AB and Nykomb Synergetics AB: Stockholm, 2000. Browne T. Pulp Pap. 2003, 104, 27-31. Bruce D. M. Pulp Pap. 2000, 101, 35-38. Ceila, M.A; Bachu, S. In Gale, J.; Kaya, Y.; Eds.; Proceedings 6th International Conference on Greenhouse Gas Control Technologies, Elsevier Science: Oxford, 2003. . Chladná, Z.; Chladný, M.; Möllersten, K.; Obersteiner, M. Investment under Multiple Uncertainties: The Case of Future Pulp and Paper Mills; IR-04-077; International Institute for Applied Systems Analysis (IIASA): Laxenburg, 2004. David, J.; Herzog, H. J. In Proceedings: 5th International Conference on Greenhouse Gas Control Technologies; Williams, D. J.; Durie, R. A.; McMullan, P.; Paulson, C. A. J.; Smith, A. Y.; Eeds.; CSIRO Publishing: Collingwood, 2000, pp 985-990. Day, D.; Evans, R. J.; Lee, J. W.; Reicosky, D. Energy. 2005, 30, 2558-2579. Dixit, A.; Pindyck, R. S. Investment under Uncertainty; Princeton University Press: Princeton, NJ, 1994. DOE (U.S. Department of Energy). Carbon Sequestration, Research and Development. U.S. Department of Energy:Washington, DC, 1999. Dornburg, V.; Faaij, A. Biomass and Bioenergy. 2001, 21, 91-108. Dowaki, K.; Ishitani, H.; Matsuhashi, R.; Sam, N. Technology. 2002, 8, 193-204. Ekström, C.; Blumer, M.; Cavani, A.; Hedberg, M.; Hinderson, A.; Svensson, C. G.; Westermark, M.; Erlström, M.; Hagenfeldt, S. Technologies and costs in Sweden for capture and storage of CO2 from combustion of fossil fuels in production of power, heat, and/or transportation fuels; Vattenfall Utveckling AB: Stockholm, 1997. (In Swedish). FAO (United Nations Food and Agriculture Organization) (2005). Statistical Databases. http://www.fao.org. Faaij, A.; Meuleman, B.; Van Ree R. Long term perspectives of biomass integrated gasification/combined cycle (BIG/CC) technology; performance and costs; EWAB 9840: NOVEM, 1998. Falkowski, P.; Scholes, R.J.; Boyle, E.; Canadell, J.; Canfield, D.; Elser, J.; Gruber, N.; Hibbard, K.; Högberg, P.; Linder, S.; Mackenzie, F.T.; Moore, B.; Pedersen, T.; Rosenthal, Y.; Seitzinger, S.; Smetacek, V.; Steffen, W. Science. 2000, 290, 291-296. Freund, P.; Davison, J. In Proceedings IPPC workshop on carbon dioxide capture and storage; Regina, 2002.

Negative Emission Biomass Technologies in an Uncertain Climate Future

545

Greene, N.. Growing energy: how biofuels can help end America’s growing oil dependence; National Resources Defence Council, New York, NY, 2004. Grönkvist, S.; Sjödin, J.; Westermark, M. Int. J. Energ. Res. 2003, 27, 601-613. Grönkvist, S.; Möllersten, K.; Pingoud, K. Mitigation and adaptation strategies for global change. Published online March 2006. Ha-Duong, M.; Keith, D. W. Clean Technologies and Environmental Policy. 2003, 5, 181189. Hamelinck, C. N. Outlook for advanced biofuels; Ph.D. thesis. Utrecht University; Utrecht, 2004. Hochenauer, C.; Hohenwarter, U.; Sanz, W.; Schlamadinger, B. Proceedings ASME Turbo Expo. 2004, Vienna, 2004. IEA (International Energy Agency). Transmission of CO2 and energy; Report no PH4/6; International Energy Agency Greenhouse Gas Research and Develoment Programme: Stoke Orchard, 2002. IEA (International Energy Agency). Carbon dioxide capture and storage issues – Accounting and baselines under the United Nations framework convention on climate change (UNFCCC); International Energy Agency: Paris, 2004. IPCC (Intergovernmental Panel on Climate Change). Climate Change 2001: The scientific basis; Cambridge University Press: Cambridge, 2001a. IPCC (Intergovernmental Panel on Climate Change). Climate Change 2001: Mitigation; Cambridge University Press: Cambridge, 2001b. IPCC (Intergovernmental Panel on Climate Change). Special Report on Carbon Capture and Storage; Cambridge University Press: Cambridge, United Kingdom and New York, NY, USA, 2005. Karlsson, Å. Comparative assessment of fuel-based systems for space heating. Doctoral thesis. Lund, Sweden: Lund University, Department of Technology and Society, 2003. Keith, D. W. Climatic Change. 2001, 49, 1-10. Keith D. W.; Rhodes J. S. Climatic Change. 2002, 54, 375-377. Kumar, A.; Cameron, J. B.; Flynn, P. C. Biomass Bioenergy. 2002, 24, 445-464. Lackner, K. S. Science. 2003, 300, 1677-1678. Larson, E. D.; Marrison, C. I. J. Eng. Gas Turbines Power. 1997, 119, 285–290. Larson, E. D.; Kreutz, T. G; Consonni, S. J. Eng. Gas Turbines Power. 1999,121, 394-400. Larson, E.; Consonni, S.; Kreutz, T. J. Eng. Gas Turbines Power. 2000, 122, 255-261. Larson, E. D.; Jin, H.; Celik, F. E. Gasification-based fuels and electricity production from biomass, without and with carbon capture and storage. Princeton, NJ: PEI, Princeton Univ; 2005. Lehmann, J.; Gaunt, J.; Rondon M. Mitigation and Adaptation Strategies for Global Change (forthcoming). Lindblom, M. Chemrec AB. Personal communication. 2001. Martin, N.; Anglani, N.; Einstein, D.; Khrushch, M.; Worrell, E.; Price L. K. Opportunities to improve energy efficiency and reduce greenhouse gas emissions in the US pulp and paper industry; Report no. LBNL-46141; Lawrence Berkeley National Laboratory: Berkeley, CA, 2000. Maunsbach, K.; Isaksson, A.; Yan, J.; Svedberg, G.; Eidensten, L. J. J. Eng. Gas Turbines Power. 2001, 123, 734-740.

546

Kenneth Möllersten, Zuzana Chladná, Miroslav Chladný et al.

Messner, S.; Schrattenholzer, L. Energy. 2000, 25, 267-282. Miner, R.; Lucier, A. A Value Chain Assessment of Climate Change and Energy Issues Affecting the Global Forest-Based Industry; National Council for Air and Stream Improvement (NCASI): Research Triangle Park, NC, 2004. Möllersten, K.; Yan, J. World Resour. Rev. 2001, 13, 509-525. Möllersten, K. Opportunities for CO2 reductions and CO2-lean energy systems in pulp and paper mills; TRITA-KET R161; Royal Institute of Technology: Stockholm, 2002. Möllersten, K.; Yan, J.; Moreira, J.R. Biomass Bioenergy. 2003a, 25, 273-285. Möllersten, K.; Gao, L.; Yan, J. Mitigation and Adaptations Strategies for Global Change. Published online August 2006. Mun, J. Real Options Analysis: Tools and Techniques for Valuing Strategic Investments and Decisions. Wiley finance series, USA, 2002. Nakicenovic, N.; Riahi, K. An Assessment of Technological Change across Selected Energy Scenarios; RR-02-005 (Reprint from a WEC report); International Institute for Applied Systems Analysis: Laxenburg, 2002. National Research Council. Abrupt Climate Change – Inevitable Surprises. National Academy Press, Washington, DC, 2002. Nguyen, M. H.; Prince, R. G. H. Biomass Bioenergy. 1996, 10, 361-365. Obersteiner, M.; Azar, Ch.; Kauppi, P.; Möllersten, K.; Moreira, J.; Nilsson, S.; Read, P.; Riahi, K.; Schlamadinger, B.; Yamagata, Y.; Yan, J. van Ypersele J-P. Science. 2001, 294, 786-787. Obersteiner, M.; Möllersten, K.; Azar, Ch.; Bohlin, B. Climate risk management – Are we ignoring the obvious? EuroScience Open Forum, Stockholm, 2004. Ogawa, M.; Okimori, Y.; Takahashi F. Mitigation and adaptation strategies for global change. 2005 (forthcoming). Ogden, J. M. In Gale, J.; Kaya, Y.; Eds.; Proceedings 6th International Conference on Greenhouse Gas Control Technologies, Elsevier Science: Oxford, 2003.. Okimori, Y.; Ogawa, M.; Takahashi, F. Mitigation and adaptation strategies for global change. 2003, 8, 261-280. Rao, S.; Riahi, K The Energy Journal (in press). Rhodes, J.; Keith, D. Biomass Bioenergy, 2005, 29, 440-450. Riahi, K.; Roehrl, R. A. Technol. Forecast. Soc. Change. 2000, 63, 175-205. Riahi, K.; Rubin, E. S.; Schrattenholzer, L. In Gale, J.; Kaya, Y.; Eds.; Proceedings 6th International Conference on Greenhouse Gas Control Technologies, Elsevier Science: Oxford, 2003. pp. 1095-1100. Schlamadinger, B.; Apps, M.; Bohlin, F.; Gustavsson, L.; Jungmeier, G.; Marland, G.; Pingoud, K.; Savolainen, I. Biomass Bioenergy. 1997, 13, 359-375. Seifritz, W. Int. J. Hydrogen Energy. 1993, 18, 405-407. Siro, M. First Finnish Kraft pulp and paper mill independent of oil as an energy source. Proceedings: 1984 TAPPI Pulping Conference, 1984, 591-595. Smith, S. J.; Brenkert, A.; Edmonds, J. Proceedings 8th Greenhouse Gas Control Technologies Conference, Trondheim, 2006. STFI (Swedish Pulp and Paper Research Institute). The Ecocyclic Pulp Mill, phase 1. Final report KAM1 1996-1999; Swedish Pulp and Paper Research Institute: Stockholm, 2000.

Negative Emission Biomass Technologies in an Uncertain Climate Future

547

Swedish Pulp and Paper Research Institute (STFI). Ecocyclic Pulp Mill – “KAM”, Final report 1996-2002; KAM A100; Swedish Pulp and Paper Research Institute: Stockholm 2003. Stigsson, L.; Berglin, N. In Proceedings 6th International Conference on New Available Technologies at SPCI99; Stockholm, 1999. Strømman, A. Norweigan University of Science and Technology, Trondheim, Norway. Personal communication, 2003. Svensson, R.; Odenberger, M.; Johnsson, F.; Strömberg, L. Energy Convers. Manag. 2004, 45, 2343-2353. Thomas, C. D.; Cameron, A.; Green, R. E.; Bakkenes, M.; Beaumont, L. J.; Collingham, Y. C.; Erasmus, B. F. N.; Ferreira De Siqueira, M.; Grainger, A.; Hannah, L.; Huges, L.; Huntley, B.; Van Jaarsveld, A. S.; Midgley, G. F.; Miles, L.; Ortega-Huerta, M. A.; Townsend Peterson, A.; Phillips, O. L.; Williams, S. E. Nature. 2004, 427, 145-148. UNFCCC (United Nations Framework Convention on Climate Change) (1992). http://www.unfccc.int. Upton, B. H.; Mannisto, H. Technologies for reducing carbon dioxide emissions: A resource manual for pulp, paper, and wood products manufacturers; Report no. 01-05. National Council for Air and Stream Improvement: Research Triangle Park, NC, 2001. Van Vuuren, D.; den Elzen, M.; Lucas, P.; Eickhout, B.; Strengers, B.; van Ruijven, B.; Wonink, S; van Houdt, R. Climatic Change (in press). Veijonen, K., Vainikka, P. Järvinen, T., Alakangas, E. Biomass co-firing – an efficient way to reduce greenhouse gas emissions; VTT Processes: Jyväskylä, 2003. Walsh, J. H. Energy Convers. Manag. 1993, 34, 1031. Warnqvist, B. Techno-economic evaluation of black liquor gasification processes; Report no 698; Värmeforsk: Stockholm, 2000 (in Swedish). Wigley, T. R.; Richels, R.; Edmonds, J. Nature. 1996, 359, 240-243. Williams, R. H. In Eco-Restructuring: Implications for Sustainable Development; Ayres, R. Ed; United Nations University Press: Tokyo, 1998, 180–222. Williams, R. H. In Proceedings VIII Biennial Asilomar Conference on Transportation, Energy, and Environmental Policy: Managing Transitions; Kurani, K. S.; Sperling D.; Eds., Transportation Research Board: Washington, DC, 2002.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 549-588

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 19

A REVIEW OF THE SOCIO-ECONOMIC AND ENVIRONMENTAL BENEFITS OF BIOMASS GASIFICATION BASED POWER PLANT: LESSONS FROM INDIA* Kakali Mukhopadhyay† UNEP-NISD; Centre for Development and Environment Policy, Indian Institute of Management Calcutta, India; Joka, D.H.Road, Calcutta-700104, India

ABSTRACT There is a steady and continuing interest in biomass gasification in both the developed countries and developing countries. While the advanced countries are interested primarily from considerations of reduced emissions and waste utilisation, the developing countries look at biomass gasification as a means to augment commercial energy like electricity, diesel, fuel oil etc. India, a tropical country with a vast geographical area is richly endowed with renewable energy sources like solar, wind, biomass which can play a crucial role in meeting end use energy needs in a decentralised manner. One of the major goals of the ninth and tenth five year plan is strengthening of infrastructure (energy, transport, communication, irrigation) in order to support the growth process on a sustainable basis. It is usually the tendency of the developing countries to equate development with economic growth and to further equate economic growth with energy consumption especially electricity. India being a developing country has also given due emphasis on strengthening its energy position accordingly. Moreover threat from Green House Gasses *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Consultant UNEP-NISD; Centre for Development and Environment Policy, Indian Institute of Management Calcutta, India; Joka, D.H.Road, Calcutta-700104, INDIA; [email protected]; kakali_mukhopadhyay @yahoo.co.in; Tel: 913324678300-04,; Fax: 913324678062

550

Kakali Mukhopadhyay (GHG) also has caused worldwide concern. In India electric power generation is the largest source of GHG emissions. It accounts for 48% of carbon emitted. These concerns point towards more rational energy use strategies. The renewable and recycling process makes biomass possible to generate power without adding to air emissions. Biomass (firewood, agricultural residue, and dung) is one of the main fuels in India, particularly in the energy-starved rural sector. The biomass power potential in India was 16,000 MW (excluding co-generation), but the achievement in this respect is negligible (Installed capacity - 630 MW Project under implementation - 630 MW, as on March 2005). It brings out the fact that much of the potential of biomass gasification is still unexplored. Globally, India is in the fourth position in generating power through biomass and with a huge potential, is poised to become a world leader in utilization of biomass. According to the Planning Commission of India, in its Tenth Five Year Plan, announced that 26.10 per cent of the Indian populations are below the poverty line and mostly belongs to rural areas. The inequitable distribution has been evident from the fact that although 70% of India’s population lives in the rural areas, only 29% of rural households have electricity supply as against 92% of urban households. Of the half a million or so villages in India, about 3, 10,000 villages have been declared to be electrified and 80,000 more villages remain completely un-electrified. There are a number of constraints to supply power to remote rural area such as small human settlements, geographically dispersed villages, seasonally of loads etc. In the absence of adequate network and hence supply of power to remote rural areas the household depend largely on primary energy sources like kerosene and diesel for lighting. No commercial investments in micro enterprises can therefore be made by either individuals or companies without installing diesel generators which have a very high generating cost. Biomass gasifier is a leading option in that respect. Besides, the supply of power to remote rural areas from the centralised grid is not competitive than a modern biomass gasification based decentralised power plant. Estimate from an Indian village shows that modest 50 kW of installed capacity per village will lead to total saving of 52000 million Rs (Rs 5200 Crore / 1100 million US $) in power plant investments. In energy terms, the saving in TandD losses will release a generation capacity of 800 MW for profitable sale. Reduced pollution and reduction of CO2 emissions will be the other advantages of a decentralised renewable energy based system for the rural areas. The purpose of the present paper is to evaluate the rural electrification programme in India undertaken by the Ministry of Non Conventional Energy Sources (MNES), Government of India, through biomass gasifier power plant. It explores the eradication of poverty that has been made possible by introducing biomass gasification based power plant in remote rural areas in India. Creation of jobs in the power stations, small-scale business, commerce and industries and also improvement in the quality of life is assessed. The paper concludes with policy options relevance for the other developing countries.

1. INTRODUCTION Electricity is the key factor to the economic and social development of any country. During the last five decades, the demand for electricity has increased manifold in India, primarily due to the rapid rate of urbanization and industrialization. In India demand for electricity is growing at 7% annually, particularly in rural areas. However, its annual per

A Review of the Socio-Economic and Environmental Benefits…

551

capita consumption is less than 400 kWh, as compared to a world average of over 2,000 kWh (MNES, 2002). Over 70% of the population of India live in villages. It is a matter of shame for all of us even 58 years after Independence, 63 per cent of all rural households in India do not have electricity and use kerosene for lighting. Even for those rural areas, which are electrified, there is a tremendous shortage of power supply. Thus it is not uncommon for these areas to have 10-15 hours of blackouts and brownouts every day. There is a shortfall of about 15,00020,000 mw of electricity in the country and we require about 140,000 mw of additional capacity by 2010 with an estimated outlay of Rs 5, 50,000 crores (MNES, 2004). Moreover, with any problems in the national grid, rural areas are mostly affected because the state electricity boards provide urban areas with electricity on priority basis. Further, the electricity supply to low-load rural areas is characterized by high transmission and distribution costs and losses and subsidized pricing. So the power generation and supply situation is grim, with shortages in installed capacity and peak power supply. Because of acute shortage of electricity, industrial growth and general life in the country is also affected seriously. Educational facilities for higher studies do not exist in these villages and sophisticated hospitals or industries are also absent due to lack of electricity supply. Furthermore, most of the villages in India is situated in remote rural areas and hence cannot be connected to the normal conventional grid. This is a vital problem. Thus, the sole option is to consider non-conventional sources of energy instead of considering fossil fuel based electricity. India, a tropical country with a vast geographical area is richly endowed with renewable energy sources like solar, wind, biomass which can play a crucial role in meeting end use energy needs in a decentralised manner. India's need for power is growing at a remarkable rate and this requirement is being met by both commercial and renewable energy sources. India, today, has a total installed capacity of about 3400 MW of power from renewables, which is over 3% of the total power generation capacity in the country, still leaving a large capacity untapped. Annual electricity generation and consumption have nearly doubled since 1990. Electricity generation has grown from 275.5 billion kilowatt-hours (kWh) to 547.2 kWh, while consumption has grown from 257.1 kWh to 510.1 kWh. The country's projected increase in electricity consumption, of between 2.6 per cent and 4.5 per cent up to 2020, is the highest for any major country (UNDP, 2005). An examination of India’s primary energy balance shows that renewable account for about 33% of the primary energy consumption in India. Of this, the major contributor is traditional bio-mass that is used for electricity generation from gasifier plants. Thus renewable energy development programme is gaining momentum in India. It has emerged as a viable option to achieve the goal of sustainable development (Thomas, 2002). As we know that coal is the major feeding item for the conventional electricity generation. Coal-based thermal power accounts for 70% of the installed capacity. (MNES report, 2004). Coal-based power generation is characterized by local and regional environmental degradation as well as greenhouse gas emissions, leading to climate change. Coal mining, storage, transportation, and combustion lead to environmental degradation. Coal combustion for power generation is the dominant source of greenhouse gases, both in India and globally. With global concern over greenhouse warming, India, like other developed and developing countries needs to consider this aspect as one of its guiding principles in selecting energy options in the medium and long term. Coal combustion also leads to the emission of oxides of

552

Kakali Mukhopadhyay

sulphur and nitrogen, and production of large amounts of ash. This results in environmental pollution for the population residing around the power plant as well as further a field. Currently more crucial for developing countries are local air pollution, land degradation, and waste disposal problems arising from coal use. Thus, the search is on for an environmentally sound alternative for meeting the power needs, particularly of rural areas in developing countries. Adverse local environment impacts (SOx, NOx, SPM) and global environmental impacts (green house gas emissions mainly due to carbon dioxide) associated with fossil fuel use have resulted in an increased emphasis on renewables. India has now the world’s largest programme for deployment of renewable energy products and systems (Thomas, 2002). The spread of various renewable energy technologies in the country has been supported by variety of incentives and policy measures. Table 1.1 shows a listing of some of the commonly used renewable options. Renewables can be used for ace heating, cooling, water pumping, cooking and for almost any endues that is presently met by fossil fuels. Table 1.1. Renewable Energy Options

*Modern Renewables Source: Planning Commission, 2005

Several studies, surveys and documents show that biomass is the most convenient option for decentralized power generation in rural India compared to other renewables. The arguments are discussed below.

Why Biomass India being an agrarian economy there is easy availability of agricultural based mass which can be used to generate energy. The term ‘biomass’ refers to organic matter, which can be converted to energy. Biomass is available all round the year. It is cheap, widely available,

A Review of the Socio-Economic and Environmental Benefits…

553

easy to transport, store, and has no environmental hazards. It can be obtained from plantation of land having no competitive use. Biomass-based power generation systems, linked to plantations on wasteland, simultaneously address the vital issues of wastelands development, environmental restoration, rural employment generation, and generation of power with no distribution losses. It can be combined with production of other useful products, making it an attractive by product. Some of the most common biomass fuels are wood, agricultural residues, and crops grown specifically for energy. In addition, it is possible to convert municipal waste, manure or agricultural products into valuable fuels for transportation, industry, and even residential use. Burning this biomass is the easiest and oldest method of generating energy. Among the various renewable energy sources, biomass conversion technologies appear to be one of the best suited for conversion to shaft power/electricity through gasifier. The gasifier is essentially a chemical reactor, where several thermo-chemical processes such as pyrolysis, combustion and reduction of biomass take place under controlled conditions. (Rehman, 2002). India has done some pioneering work in this area for many years and built open-top gasifiers and integrated them with internal combustion engines or boilers. Biomass exists in rural areas and needs to be tapped to provide not only electricity but also water to irrigate and cultivate fields to further increase production of biomass (either as a main product or as a by-product). The various applications of biomass energy include thermal or heat, mechanical water pumping for irrigation and power generation including village electrification. The availability of waste biomass from the biomass gasifier plant to be used as fertilizer is an added advantage. As a renewable fuel, biomass is used in nearly every corner of the developing world as a source of heat, particularly in the domestic sector. Unlike other renewables, biomass is a versatile source of energy, particularly attractive for decentralized applications which can be converted to ‘modern’ forms such as liquid and gaseous fuels, electricity, and process heat. Bioenergy also permits operation at varying scales. For example, small-scale (5–10 kW), medium-scale (1–10 MW) and large-sale (about 50 MW) electricity generation systems or biogas plants of a few cubic metres (Indian and Chinese family plants for cooking) to several thousand cubic metres (Danish systems for heat and electricity). This variety of scales is useful for power generation for decentralized applications at the village level as well as for supply to the national grids. Unlike wind, solar or micro-hydroelectric systems, modern biomass energy systems could be set up in virtually any location where plants can be grown. (WEC, 1993) Renewables such as solar, wind, and micro-hydro require ‘spare’ or additional capacity to produce adequate energy when the conditions are favourable, such as water flow or wind speed. This intermittent feature of such renewable energy sources necessitates electricity storage facilities, especially with small and local systems, if sustained demand requirements are to be met (WEC1993). Bioenergy sources such as producer gas systems do not require electricity storage. This is an important advantage. Currently, biomass contributes 15 per cent of the total energy supply worldwide and 40 per cent of this energy is consumed in developing countries, mostly in the rural and traditional sectors of the economy. According to WEC (1993), under the minimum case scenario, modern biomass is the most important of the renewables and is projected to account for 45 per cent of the new contribution by renewables to world energy by 2020. In the maximum scenario projections,

554

Kakali Mukhopadhyay

modern biomass will account for 42 per cent of the total renewable energy contribution by 2020. The US Department of Energy suggests that biomass power will be the most important renewable energy option for the next quarter of a century and has projected a possible 25 GW in installed power in the US alone by 2010 (USDOE 1993). This clearly demonstrates a dominant role for biomass energy among all the renewable. Most common source of biomass is wood waste and agricultural wastes. In India development of biomass gasification has received serious attention with establishment of biomass research centers and gasifier action research centers at various locations spread all over the country. These institutions have played a key role in upgradation and adoption of suitable technologies, testing, monitoring and development of biomass gasification systems. Studies reveal that the low grade of land suitable only for scrub vegetation can be turned to advantage and form an excellent source of biomass – fast growing trees and shrubs. India has actively promoted research and development programmes for efficient utilization of biomass and agro wastes and further efforts are on. In short, biomass energy systems offer an opportunity for sustainable (as biomass can be grown sustainable), self-reliant (biomass is available in all countries and can be converted to gaseous or liquid fuels and electricity, leading to a decrease in imports of oil), and equitable development (between and within countries due to universal availability of biomass and the fact that decentralized bio-energy systems lead to local control and employment). As far as biomass-based power generation is concerned, the commissioned biomass power capacity reached 290 MW (52 projects) and the commissioned co-generation capacity installed mainly at sugar mills reached 437 MW (57 projects) by the end of year 2004. In the area of small-scale biomass gasification, significant developments in technology have made India a world leader. A total capacity of 55 MW (1817 projects) of biomass gasifier system has so far been installed in India, mainly for stand-alone applications. Biomass gasifiers capable of producing power from a few kilowatts up to 500 kW have been successfully developed indigenously and are also now being exported to the developing countries of Asia and Latin America, and also Europe and USA (MNES, 2005). In India more than 2000 gasifiers are estimated to have been established with a capacity in excess of 22 MW and a number of villages have been electrified with biomass gasifier based generators. These gasifiers mainly consumed, fuel wood (200-300 million tonnes), animal waste (80- 100 million tonnes) and crop residues (100-120 million tonnes) annually as the main biomass fuels. Fuel wood contributes nearly 60 per cent of the total biomass energy in India estimated between 200-300 million tonnes. There are immense scope and potential to be acquired from biomass gasification i.e., water pumping, electricity generation - 3 to 1 MW power plants, heat generation: for cooking gas – smokeless environment and rural electrification ensuring better healthcare, better education and improved quality of life (MNES, 2005). There is no doubt that the environment friendly electricity generation is only possible through renewables. Since the central focus of this paper is on biomass gasification and corresponding generation, its eco- friendly role also receives proper attention. When biomass is used to produce power, the carbon dioxide released at the power plant is recycled back into the re-growth of new biomass. This renewable and recycling process makes it possible to generate power without adding to air emissions. Due to the non-availability of the sufficient resources and a considerable amount of emission of pollutants from commercial energy, it is now being felt that renewable energy has to be utilised to a greater extent. In India electric

A Review of the Socio-Economic and Environmental Benefits…

555

power generation is the largest source of GHG emissions and accounts for 48% of carbon emitted. These concerns point towards more rational energy use strategies. The statistics shows that estimated biomass power potential is always high but its installed capacity is low, though its commercial viability is quite well known. This fact is quite apparent in tables 1.2 and 1.3. Biomass collection logistics and competing uses of biomass are the most serious constraints in the widespread utilization of this resource for power generation. For the grid interactive (table 1.2) case the total cumulative installed capacity is 7097.54 MW as on 31December 2005 but the share of biomass power is nominal. As against the 16,000MW estimated potential of biomass power, a cumulative installed capacity of 376 MW has been commissioned till 31.12.05. Maximum capacity is under implementation in Tamil Nadu, Andhra Pradesh, Karnataka, UP, Maharashtra, Punjab and Haryana. The cumulative installed capacity is further low for decentralized energy systems (table 1.3). Table 1.2. Grid-interactive renewable power Source/ system

Estimated potential(MW)

Wind power Biomass power Bagasse cogeneration Small hydro (up to 25 MW) Waste-to-energy Solar photovoltaic

45 000 16 000 3500 15 000 2700 20 MW per km2 Total

Cumulative installed capacity(as on 31 December 2005) (MW) 4434.00 376.00 491.00 1747.98 45.76 2.80 7097.54

Source: Planning Commission, 2005

Table 1.3. Decentralized renewable energy systems Source/system Estimated potential Cumulative installed capacity (as on 31 December 2005) 1. Family-size biogas plants 120 lakh 38.00 lakh 2. Community/institutional/night soil biogas plant — 3952 nos 3. Improved chulha 12 crore 3.52 crore 4. Solar photovoltaic systems 20 MW/km2 5. Solar water heating systems 140 million m2 collector area 1.5 million m2 collector area 7. Solar PV pumps — 6818 nos 8. Wind pumps — 1087 nos 9. Hybrid systems — 410 kW 10. Biomass gasifiers — 69 MW

Source: Planning Commission, 2005

Biomass for power generation has been recognized as an important component of the renewable energy programme in India and this is reflected in the priority by the MNES. Recently, the Prime Minister has set up the Rural Electricity Supply Technology (REST) mission in the Union ministry of power. The State Governments have been directed to take up the electrification of 62,000 villages through the Electricity Boards under the traditional rural electrification programmes by 2007 under the Pradhan Mantri Gramodhaya Yojna.. The Government of India has also directed MNES to take up renewable energy based

556

Kakali Mukhopadhyay

electrification of 18,000 villages in remote and inaccessible parts of the country by 2012. These villages would be electrified through decentralised plants based on biomass, gasification of biomass, hydel power, solar thermal power etc. The action plans so far implemented or to be undertaken has been portrayed in section 2. The above background simply states that Biomass potential is enormous in India and it can be considered as a feasible option in all respect among other renewable. Though the potential is high in comparison to other renewable, its application is not up to the desired level, however. The purpose of the paper is to evaluate the feasibility of the biomass gasifier based power plants for rural electrification. The feasibility mentioned here basically encompasses the benefits acquired by the villagers due to the power plant. The evaluation will also help the policy makers and researchers to get a comprehensive overview of rural electrification in India through decentralized power plant. Further this will encourage those unelectrified villages going to undertake biomass gasifier power plant as an option. The rest of the paper is structured as follows. Section 2 reviews implementation of the rural electrification programme, actions and strategies formulated by the Government of India and the various programmes undertaken by the Ministry of Power. Section 3 describes the village level case studies mainly in respect of rural electrification. It evaluates the socioeconomic and environmental benefits of the installed BGBPP. Overall assessment of the case studies are explained in section 4 and section 5 concludes the paper with policy options.

2. RURAL ELECTRIFICATION PROGRAMME THROUGH BIOMASS GASIFICATION BASED POWER PLANT Rural electrification programme is a common agenda in every five year plans in India. Its acute necessity has also been reflected in various plans. But the proper implementation of the programme is a fundamental problem. In this section the paper presents the various schemes towards rural electrification undertaken by MNES, Government of India so far.

Rural Electrification Initiatives Rural electrification programmes began in the 1950s as a social amenity, but they gathered importance in the mid-sixties as a source of energy for pumping irrigation water. The Rural Electrification Corporation was formed in 1969 with the task of electrifying all villages in India. Significant success in this effort has been achieved and 84 per cent of the villages were electrified by October 1993 (CMIE 1994). Many major states, such as Andhra Pradesh, Gujarat, Karnataka, Maharashtra, and Punjab, have the distinction of electrifying nearly 100 per cent of villages. However, in West Bengal on 73 per cent of the villages are electrified, while the lowest level of electrification is in Bihar where only 70 per cent of villages are electrified. Thus the achievement of the rural electrification programme has been excellent; for example between 1980 and 1991 the number of villages newly electrified was 212634. However, there are many critics of the rural electrification programme, who argue that there are many villages (no data are available) where electricity transmission lines pass through the village and yet none of the households or farms has benefited from electrification.

A Review of the Socio-Economic and Environmental Benefits…

557

Such villages are, however, classified as ‘electrified’. There could be some truth in these misgivings since the data from some states, such as Bihar and Uttar Pradesh show that only 4.4 and 9.3 per cent, respectively, or rural homes are actually electrified (NSS 1992). What is important is that at least 70 000 villages are yet to be connected to the grid, and even more important is that these villages could be remote, requiring very large investments to connect them to the grid. This assumption is based on the logic that the state electricity utilities, wanting to show progress and achieve physical targets, start with villages near existing grid transmission lines and leave all the remote and difficult villages to the last. These villages provide the first opportunity of alternative power generation systems. During the year 2002-2003, 3056 inhabited villages were electrified as on 31.12.2002 and 213618 pumpsets /tubewells energized as on 30.11.2002. Cumulatively 509678 villages have been electrified and 13355909 electric irrigation pumpsets have been energized as on 30.11.2002(Ministry of Power, 2003). Table 2.1 shows the status of rural electrification recently. The major percentage of unelectrified villages and corresponding household is concentrated in Jharkhand, Bihar and Uttar Pradesh. The number of remote villages in India is large mostly concentrated in the eastern part of Indian peninsula. The number of the remote village electrification is also growing (table 2.2). Towards this end it is better to check the state wise total number of biomass gasifier installation (table2. 3). A total capacity of 55.105 MW has so far been installed, mainly for stand-alone applications. The most worth mentioning plants in this respect are: 1) A 5 x 100 KW biomass gasifier installation on Gosaba Island in Sunderbans area of West Bengal is being successfully run on a commercial basis to provide electricity to the inhabitants of the Island through a local grid. 2) A 4X250 kW (1.00 MW) Biomass Gasifier based project has alsdo been commissioned at Khtrichera, Tripura for village electrification. Table 2.1. Status of Rural Electrification - Selected States State Jharkhand

Villages to be electrified 22,920

%village unelectrified 78%

Households to be electrified 3,422,425

%households unelectrified 90%

Bihar

20,449

53%

12,010,504

95%

Uttar Pradesh Assam Orissa West Bengal

40,389 5,640 9,682 7,694

42% 23% 21% 20%

15,505,786 3,522,331 6,651,135 8,899,353

80% 84% 81% 80%

Source: Ministry of Power, 2003

The purpose of the present study is to evaluate the rural electrification programme in India undertaken by the Ministry of Non Conventional Energy Sources (MNES), Government of India, through biomass gasifier power plant. Recently, MNES provides power to 15 villages in Karnataka by the utilization of biomass. The project aims at sustainable transformation of energy in rural areas of Karnataka, it is providing power to 15 villages in the Tumkur district of the state. The energy is produced through biomass gasifiers for standalone applications and supplying to the villages. The stand-alone gasifiers are located in every village within the capacity of 60 to 100 kW (The Business Standard, 25 February 2000).

558

Kakali Mukhopadhyay Table 2.2. Remote village electrification Remote villages/hamlets electrified through RE (as on 31 December 2005) 2195 remote villages 594 remote hamlet

Item Remote village Electrification

Source: MNES, 2005

Table 2.3. State-wise Cumulative Total Number and Capacity of Biomass Gasifiers Installed upto 30th June, 2003 S.No.

State

Cumulative No. of systems

Capacity (in kW)

1.

Andhra Pradesh

231

15384

2.

Arunachal Pradesh

3

180

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

Assam Bihar Chhatisgarh Goa Gujarat Haryana Himachal Pradesh J and K Karnataka Kerala Madhya Pradesh Maharashtra Mizoram Orissa Punjab Rajasthan Tamilnadu Tripura Uttar Pradesh West Bengal AandN Island Delhi Others Total

6 2 1 3 237 25 2 4 476 13 144 316 2 16 27 21 83 4 50 27 17 16 91 1817

123 20 500 22 11961 964 7 120 4499 725 4529 3823 200 72 700 218 2653 1000 2746 4100 167 74 318 55105

Source: MNES, 2004

In pursuance of the national agriculture policy, a National Biomass Resource Atlas is being prepared with the specific intention of boosting power generation from biomass. The national agriculture policy had called for increasing power generation from renewable sources for meeting the needs of agriculture. The National Biomass Resource Assessment Programme has been assigned this task. 150 taluka level studies have been completed and another 175 are being taken up (The Financial Express, 4 September 2000). According to a recent initial assessment made by the MNES about 500 million tonnes of biomass is generated every year from crop residues, bagasse, agro residues, and forest sources. The newspaper also says that out of this, only 170 million tonnes is used every year for power generation.

A Review of the Socio-Economic and Environmental Benefits…

559

The action and strategies for rural electrification taken by the Government of India through Ministry of Power and MNES under various planning period is explained below: Rajiv Gandhi Gram Vidyutikaran Yojana (RGGVY) ---- This initiative is to provide electricity access to all households in five years covering the entire country. This programme provides for ninety per cent capital subsidy for rural electrification projects covering to electrify the 1, 25,000 unelectrified villages. It will connect all the estimated 2.34 crore unelectrified households below the poverty line (BPL) with 90 percent subsidy on connecting costs and augment the backbone network in all the already electrified 4.62 lakh house holds by 2010. The 5.46 crore households above the poverty line, which are currently unelectrified, are expected to get electricity connection on their own without any subsidy. Such expansion of connectivity will require a corresponding expansion of supply capability. Given the present widespread and acute shortage of power in many states, special action is needed to facilitate and encourage decentralized distributed generation (DG) system so that community can take their destiny in their own hands instead of waiting for utility companies to supply electricity reliably. As per the NSS 55th Round Survey in 1999-2000, among the households in rural areas who had electricity, the households who belonged to the poorest 5 percent of all rural households, spent more than Rs.300 per year for electricity. Thus a charge of Rs.1.0 per kWhr for the first 30 units per month should be within the capacity and willingness of even the poorest 5 percent households. Above 30 units per month the normal charge should be levied. For this electricity would have to be metered. An effective way of targeting the subsidy to BPL households will have to be found. Increasing the level of rural household electrification from 44 per cent (2001) to the targeted 100 per cent by 2012 would lead to a huge growth in both demand and consumption. The Rs. 16,000 crores (nearly US$ 3.5 billion) outlay for the scheme, also opens up big opportunities for electrical equipment manufacturers.

Pradhan Mantri Gramodaya Yojana (PMGY) Rural Electrification was included under Pradhan Mantri Gramodaya Yojana (PMGY) from 2001-02 to achieve human development at the village level. The six components of PMGY now are: Primary Health, Primary Education, Rural Drinking Water, Rural Shelter, Nutrition and Rural Electrification. During 2002-03, the PMGY is being administered by the Planning Commission. Under the revised guidelines, the States would have flexibility to decide their inter-se allocation of ACA among the six PMGY sectors as per their own plan priorities and discretion. The funds for village electrification are available as Additional Central Assistance with 90% grant and 10% loan for the special category States, and 30% grant and 70% loan for other States. Government has released Rs.36066.35 lakhs to various States as first installment (50%) under PMGY for 2002-03. The Centre proposes to electrify 62,000 villages through grid power, during the 10th Five-Year Plan (2002-07) under the Pradhan Mantri Gramodhaya Yojna. The Centre hopes to achieve 100 per cent village electrification in the current Plan. Another 18,000 remote villages will be electrified through non-conventional energy as grid power could prove uneconomical. These villages would be electrified through decentralised plants based on biomass, gasification of biomass, hydel power, solar thermal power etc.

560

Kakali Mukhopadhyay

Minimum Needs Programme (MNP) The revised criteria for the MNP components of rural electrification adopted since the beginning of 7th Plan are as under: (a) all North-Eastern hilly States; (b) all States with less than 65% electrification and in these States those districts will be taken up which has less than 65% electrification provided that districts having least percentage coverage will be given priority over the others; and (c) all areas including in the tribal sub plan. During 2002-03 Rs.600 crores have been allocated to the eligible States under MNP. The break up is follows. Table 2.4. Minimum Needs Programme allocated outlay (Rs. in lakhs) S.No.

States

Amount

1.

Arunachal Pradesh

1200

2.

Assam

6000

3.

Bihar

6800

4.

Chattisgarh

800

5.

Himachal Pradesh

200

6.

Jharkhand

6800

7.

Madhya Pradesh

800

8.

Manipur

270

9.

Meghalaya

3000

10.

Nagaland

130

11.

Orissa

6000

12.

Uttar Pradesh

15000

13.

Uttaranchal

7000

14.

West Bengal

6000

Total

60000

Source: Ministry of Power, 2003

Rural Electricity Supply Technology Mission (REST) Distributed Generation has been identified as one of the technologies for ensuring supply of power in rural areas by way of setting up of small generating units based on a variety of local funds along with localized distribution. The electricity distribution in the rural areas is characterized by low density, high cost of delivery, poor availability of supply of power and commercially unviable on account of high faxed cost and high variable cost. In order to utilize technology in proving for an affordable solution in making available electricity in rural areas, it has been decided to constitute the Rural Electricity Supply Technology Mission (REST) under the auspices of Minister of Power. The Mission would evolve a strategy based on technology which could provide for low cost power generation and low cost of delivery in the rural areas which can be managed by local institutions like Village Panchayats of Non-Government Organizations and to identify feasible size of generating units for different fuels, which are locally available and for mini and micro hydel projects. It is hoped through this mission to electrify all villages by 2010. Of the half a million or so

A Review of the Socio-Economic and Environmental Benefits…

561

villages in India, about 3, 10,000 villages have been declared to have been already electrified. According to government statistics, 80,000 more villages remain to be electrified. The State Governments have been directed to take up the electrification of 62,000 villages through the Electricity Boards under the traditional rural electrification programmes by 2007. The Government of India has also directed MNES to take up renewable energy based electrification of 18,000 villages in remote and inaccessible parts of the country by 2012. According to Ministry of Power officials, funds of about Rs 10,000-15,000 crore will be made available to the rural power utilities at 2-2.5 per cent per annum interest rate. With the new Electricity Act and this type of funding it becomes very attractive for micro utilities to come up in rural areas. Thus it is envisaged that a small rural power cooperative can be set up to produce 200-500 KWe of power and supply all the electricity demands of one or two villages. Again this utility can lease the existing SEB power line infrastructure for its purposes.

Accelerated Rural Electrification Programme (AREP) Government of India in the Budget for 2002-03, has announced the introduction of a new Interest Subsidy Scheme called Accelerated Rural Electrification Programme. With the Interest Subsidy Scheme, States should be able to give this programme the requisite momentum. An outlay of Rs.164 crores has been provided for this Scheme during 2002-03. The interest subsidy will be at 4% and would be provided to the States for the loans to be taken for rural electrification of un-electrified villages including Dalit Basti.

Kutir Jyoti Scheme The Government of India in 1988-89 launched a programme called Kutir Jyoti for extending single point light connections to the households of rural families below the poverty line including Harijan and Adivasi families to improve the quality of life of such poor families. Under this programme, one time cost of internal wiring and service connection charges is provided by way of 100% grant to the State Governments/State Electricity Boards through REC The money provided for release of Kutir jyoti connections covers the service line from the pole, the fuse unit, switch, the meter and board. It also covers the cost of single point internal wiring and the cost of bulb. Keeping in view the current cost of material and labour charges, the Government has now approved to revise the cost from the present Rs.l ,000/- to Rs.1800/- per connection in respect of special category States and Rs.1500/ - per connection in other States. Government has also decided that under this Programme, only metered connections should be given. During 200203, it was proposed to provide 6, 53,007 connections. The programme, by and large, has been successfully implemented in the States barring a few States like Assam, Goa, J and K, Manipur, Orissa, and Uttar Pradesh and West Bengal. These States have not been able to implement the programme at the desired pace based on performance review upto March ending 2004.

562

Kakali Mukhopadhyay

Exploitation of the abundant biomass energy resources available in our country is being accorded a high priority by the MNES. The implementation of projects is being facilitated through comprehensive programmes by the Ministry. The ministry aims to create a favourable policy environment, encourage technology upgradation and ensure market for the power generated. In January 2000, the MITCON (Maharashtra Industrial and Technical Consultancy Organization) was appointed the lead programme partner of the MNES, for the promotion of biomass and bagasse cogeneration on an all-India basis. The mandate under this programme for the next 15 months was to financially close projects equivalent to 200 MW of exportable surplus, including 160 MW from sugar mill cogeneration and 40 MW from other biomass materials. This effort was expected to cover nine major sugar-producing states in the country and also other states with biomass availability. MNES has also taken up few strategies to improve this sector. 1) Biomass gasifiers capable of producing power from a few KW upto 1 MW capacity have been successfully developed indigenously. 2) Indigenously developed small biomass gasifiers have successfully undergone stringent testing abroad. 3) Biomass Gasifiers are now being exported not only to developing countries of Asia and Latin America, but also to Europe and USA. 4)A large number of installations for providing power to small scale industries and for electrification of a village or group of villages have been undertaken. 5) The Biomass Gasifier Programme has been recasted to bring about better quality and cost effectiveness. 6) The programmes on biomass briquetting and biomass production are being reviewed and a new programme on power production linked to energy plantations on waste lands is proposed to be developed. The following steps have already undertaken in the 10th five year plan to facilitate the rural electrification programme (Government of India, 2002). 1. To identifying remote areas where power supply from the conventional grid will be prohibitively expensive and make it a priority to provide off-grid supply from renewables for these areas. Create provisions for integrated generation and distribution of off-grid energy supply. 2. To encourage private sector investments in renewable energy sources by promoting a bidding process for available subsidies. Award contacts to private entrepreneurs who provide maximum benefit with the lowest amount of subsidies. 3. To promote local private sector management of both generation and distribution for offgrid supply from renewable sources. 4. To optimize energy plantation by raising plants on degraded forest and community land. The above plans, programmes, schemes and initiatives taken at the government level no doubt bear some fruitful results in respect of village electrification. But being a developing country its implementation is not at the desirable space. In the next section, the paper highlights a few successful case studies in respect of rural electrification through biomass gasifier.

A Review of the Socio-Economic and Environmental Benefits…

563

3. CASE STUDIES OF BIOMASS GASIFICATION BASED POWER PLANT IN INDIA A country like India where most of the people is living in the village and normal grid electricity is difficult to access in the remote village then decentralized power generation based on renewable is an attractive option to meet village energy needs. A list of biomass gasifier installations in India is given below: 3.1. Illustrative list of application-wise biomass gasifier installations* power applications Type of application

Supplied Designed by

Place of installation

Capacity

Forest Development Corporation, West Bengal

1 x 30 kVA

2 Textile units in Gujarat Ankur 4 cold storage units in Uttar Pradesh Captive power

IISc

Ankur

Rural electrification IISc

1 x 120 kW 1x 60 kW 3 units with 1 x 100 kW capacity each 1 unit with Ix 200 kW capacity

Biomass feedstock Woody biomass

Rice husk

Manufacturer of electrical insulation boards, filter grade paper and allied products, Karnataka

500 kW

solid bioresidue such as mulberry 'stalk

Navodaya Vidyalaya, Karnataka

100 kW

N.A.

Gosaba island, Sunderbans, West Bengal

5 x 100 kW

Chhotomollakhali, Sunderbans, West Bengal

4 x 125 kW

Khtrichera, Tripura

4 x 250 kW

Hoshalli village, Karnataka

20kW

Ungra village, Karnataka

20kW

Hanumanthanagara, Karnataka

20kW

Port Blair, AandN islands

100 kW

Karavatti, Lakshwadeep island (oronosed)

250 kW

Woody biomass

Woody biomass N.A. Woody biomass Woody biomass N.A.

564

Kakali Mukhopadhyay 3.1. Continued.

Type of application Other electrification projects

Supplied Designed by DA/DESI Power/ NET PRO

Place of installation

Capacity

Biomass feedstock

Desi Power Orchha(P) Ltd, MP

100 kW

Woody biomass

lIT Delhi, Mechanical Engg

12kW

Desi power,bodhadhara Dewan estate,karnataka Desi power,mahanadi,orissa Devpower corp. tamilnadu GB Engineering enterprises, Tamilnadu Desi power,KOSI Ltd. Bihar Desi power, baharbari Bihar

100kw 50kw 120KW 120KW

SKET, Phase I, Kamataka

120 kW

120KW

N.A.

50kw 50kw

SKET, Phase 11, Kamataka

120 kW

MVIT, Phase I, Bangalore

120 kW

MVIT, Phase 11, Bangalore

120 kW

Varlakonda, Kamataka

50kW

GB Food Oils, Tamil Nadu V.LT Vellore Tamil Nadu

120 kW 120 kW

Woody biomass N.A.

Thermal applications Supplied designed by

Place of installation

Capacity

CO2 manufacturing, Guiarat

I x 150 kW

6 ceramic firms in Gujarat

One firm with I x 300 kW and I x 500 kW units; One firm with I x 300 kW unit; Four firms with 2 x 300 kW units

8 firms in Gujarat

One firm with I x 10 kW unit; One firm with I x 20 kW unit; One firm with I x 40 kW unit; One firm with I x 60 kW unit; One firm with I x 100 kW unit; One firm with I x 300 kW unit; One firm with 2 x 300 kW unit; One firm with I x 500 kW unit;

Industrial abrasives manufacturing unit, Tamil Nadu

I x 500 kW

Ankur

Ankur

IISc

Biomass feedstock

Woody biomass

Firm, Maharashtra

I x 500 kW

Firm, West Bengal Mis Agro Biochem, Kamataka for marigold flower drying Thermal Central Building Research Institute, Roorkee, Uttar Pradesh

I x 100 kW

Rice husk

IMW

Woody biomass

Tea drying, Bangalore

800 kW 1.5MW

N.A.

A Review of the Socio-Economic and Environmental Benefits…

565

Thermal applications (continued) Supplied designed by

Place of installation

Capacity

Biomass feedstock

CO2 manufacturing, Junagarh, Guiarat

2 x 150 kW

Fire wood

2 x 150 kW

Firewood

13 x 20 kW

"

Manufacture of magnesium chloride, Kharagodha, Guiarat Silk dyeing, Bangalore Green brick drying, Palghat, Kerala

TERI (through several manufacturers)

I x 20 kW

"

Rubber drying, Kerala and Tamil Nadu

6xlOOkW

Rubberwood, coconut shells, cashew nut shells

Silk reeling, Kamataka and Tamilnadu

30xlOkW

Firewood

Cardamom curing, Sikkim, Bhutan

-150 x 20 kW

"

Institutional cooking, Beas Satsang, Punjab

I x 100 kW

Cooking in tribal school, Gram Vikas, Berhampur, Orissa

I x 10 kW

"

Rice mill, Orissa

"

I x 20 kW

" " "

Crematorium, Nagarik Sewa MandaI, Ambemath, Maharashtra.

I x 100 kW

"

Puffed rice making, Dharwar Kamataka

I x 10 kW

"

Khoya making, Raiastan

I x 10 kW

"

Steel re-rolling, Haryana

1x100kw

"

Cooking in hostels, jails (AP. and T.N.)

N.A.

Rubber drying, Kerala

I x 100 kW

Cosmo

Steel re-rolling, Raipur

Several

Radhe Industries

Ceramic firms, Gujarat

Several

Harris

Rubber drying Kerala

6 x 100 kW

*

"

Bamboo mat factory, Gram Vikas, Orissa I x 20 kW Drying of mushroom and mahua flowers, I x 10 kW Pradan, MP Food processing (Tooty-frooty) I x 20 kW factory, Bangalore Melting of Lead in a battery reclamation I x 20 kW factory, Bangalore

AEW

"

N.A.

Sources: (1) Internal documents and in-house publications from institutions such as Ankur, NETPRO, DESI Power, TERI, and lISC, (2) Web-based information: http://ankurscientific.com http://www.desiDower.co,in http://cgpl.iisc.ernet.in, http://www.teriin.org

From the above table it is quite clear that application on rural electrification is very few compared to other applications. Though the potential of biomass in India is enormous still its application is hardly few. It is concentrated in a few states such as Karnataka and West Bengal. However, it is worth mentioning that several biomass projects which are installed and found successful in many respect. Now we shall present some of these successful projects.

566

Kakali Mukhopadhyay

1) Hosahalli and Hanumanthanagara Project The biomass gasifier-based decentralized power generation systems were implemented in Hosahalli and Hanumanthanagara villages of Tumkur district in Karnataka (Somasekhar et al.2000). Hosahalli was a non-electrified village and Hanumanthanagara was electrified (gridconnected); however, only 30% of the houses were electrified, at the time of project initiation. The number of households in Hosahalli is 35 and Hanumanthanagara 58 (Table 3.2). Hosahalli did not have any pumps or a flourmill. Kerosene-based traditional wick lamps were used for lighting. Women carried water from a polluted open water tank nearly 1 km away from the village. Farmers depended on rainfed agriculture which is subjected to vagaries of monsoon, with low crop yields and occasionally hired diesel engines to pump water for irrigation or partial supply from irrigation tank. The bio-energy project was planned and implemented by CST (The Centre for Sustainable Technologies). Energy forests were established in both the villages; Hosahalli during 1988 and Hanumanthanagara during 1996. The gasifier-based power generation system was installed in Hosahalli during 1988 and in Hanumanthanagara during 1996 and the end-use systems were installed in phases. Local youth were trained to operate and undertake minor maintenance of the systems. CST obtained funds for the implementation of gasifier power system in both the villages. Village committees managed the systems, taking decisions on operation, supervision of the operator, protection of the forest and ensuring payment for the services provided. Features and Performance: The installed capacity and load for different services are given in Table 3.2. The capacity of biomass gasifier system installed is 20 kW in both the villages. Even though the installed end-use capacity is higher (30 and 37 kW), the load is distributed such that the irrigation load is scheduled for day hours and other load activities are planned for evening hours (6 to 11 p.m.). In Hosahalli, initially a 3.7 kW power generation system was commissioned in 1988. The 20 kW biomass gasifier system was commissioned in 1997 at Hosahalli and 1996 at Hanumanthanagara. Table 3.2 provides details regarding the performance of the system and services provided to the village over the last six years. It is important to note that the power generation system was operational in Hosahalli for over 90% of the days during 1998–2003. Table 3.2. Features of biomass gasifier systems Description Year of establishment Size of village (number of households) Population Energy plantation (ha) raised Installed capacity(kWe) Installed end-use capacity(load) Lighting Drinking water Flour mill Irrigation pump Total installed end-use capacity

Hosahalli 1988* 35 218 4 20

Hanumanthanagara 1996 58 319 8 20

4.0 2.6 5.6 18.5 30.7

4.0 2.6 5.6 25.5 37.7

The installed capacity in 1988 was 3.7 kW and was expanded to 20 kW in 1997. Source: Ravindranath et al., 2004

A Review of the Socio-Economic and Environmental Benefits…

567

The power generation system functioned for over 70% of the days on dual fuel mode, indicating the operation of the gasifier system for power generation using biomass. The system was operated in diesel-only mode for power generation in about 25% of days, due to non availability of processed wood fuel or problems with the gasifier system and subelements or non-availability of trained operators. The total electricity generated varied from 12 to 22 MWh per year. Table 3.3. Hosahalli: System operation and provision of utility services System operation and services provided

2003

2002

2001

2000

1999

Days operated during the year

355

358

347

298

343

1998 349

Days on dual fuel mode

287

272

250

162

257

269

Days on diesel mode

68

86

97

136

86

80

Days services provided for lighting during the year

355

349

347

287

310

300

Days services provided for drinking water

353

344

339

293

338

295

Days services provided for flour milling

162

97

155

92

180

125

Days services provided for irrigation water

79

88

39

-

-

-

Source: Ravindranath et al., 2004

The general philosophy adopted in this project implementation was to provide the basic services like piped water and lighting at home and streets in a reliable manner. To achieve this objective, occasionally services are provided using diesel system, if gasifiers cannot be operated. In Hosahalli, lighting and piped drinking water services were provided for over 85% of the days during most years. The flour mill was operated twice or thrice a week depending on the demand for milling of grains. The irrigation system was operated depending on the crops grown, area irrigated, cropping season and demand from farmers. The basic services such as home and street lighting and piped water supply were provided on most days. This is a unique achievement for a village in India. Table 3.4 provides data on electricity generation and biomass and diesel consumption per kWh during 1998–2003 at Hosahalli. Table 3.4. Hosahalli: Total annual electricity generation and fuel consumption Description Electricity generated kWh/yr in dual fuel mode Electricity generated kWh/yr in diesel-only mode Total electricity generated kWh/yr Average wood consumption rate kg/kWh/ in dual fuel mode Diesel use in duel fuel mode 1/kwh Diesel use in diesel-only mode 1/kWh Diesel substitution in per cent under dual fuel mode

2003 18651 3326 21977 1.8

2002 17185 3992 21557 1.64

2001 12775 3476 16251 2.07

2000 7238 5251 12489 1.28

1999 9617 3267 12884 1.27

1998 9300 2723 12023 1.32

0.063 0.567 85.55

0.077 0.76 87.02

0.086 0.779 80.69

0.109 0.564 80.67

0.173 0.379 54.35

0.182 0.432 58.33

Source: Ravindranath et al., 2004

Cost analysis is carried out using only the variable or recurring costs, namely the mean monthly fuel, OandM costs. The fuel, OandM costs at different loads are given in Table 3.5.

568

Kakali Mukhopadhyay

It is observed that the OandM cost per unit declines from Rs 5.85/kWh at a load of 5 kW to Rs 3.34/kWh at a full load of nearly 20 kW on selected days (2003 cost data). The biomass-based decentralized power generation system implemented in Hosahalli village has provided multiple social, economic and environmental benefits; some measurable and others not. The potential benefits of large scale spread of decentralized biomass power systems are as follows. Table 3.5. Fuel consumption, units of electricity generated and O and M costs at different loads Maintenance Load (kW) Diesel cost Biomass (Rs/h) cost (Rs/h) 6.0 16.4 9

Engine (Rs/h) Gasifier (Rs/h) Labour cost Total cost (Rs/h) (Rs/h) 5.42 0.98 6.25 38.05

Cost/kWh(Rs/k Wh) 5.85

7.0

21.1

10.5

5.42

0.98

6.25

44.25

4.92

8.5

18.74

10.5

5.42

0.98

6.25

41.81

4.65

11.5

22.26

15

5.42

0.98

6.25

49.91

3.56

15

25.77

18

5.42

0.98

6.25

56.42

3.52

20

42.17

25.5

5.42

0.98

6.25

80.32

3.34

Cost: Wood = Rs 0.75/kg; Diesel+transport = Rs23.45/1; Engine maintenance = Rs 5.42/h; Gasifier maintenance = Rs 0.98/h, Operator wage = Rs 6.25/h. Source: Ravindranath et al., 2004

Provision of reliable and safe water supply for households is made on most days near the door-step, from a deep borewell. This reduces the drudgery involved in lifting and carrying water from an open water pond nearly a kilometer away. Women in this region spend about 2.6 h/day in collecting water. Further, the quantity of water consumed, an indicator of quality of life, which was low earlier (26 l/capita/day) has gone up based on field observations, due to nearly 2 h of water supply near the door-step of the houses. Women reported improvement in their own health as well as that of their children due to safe water supply and increased use of water and better hygiene. Electricity for lighting in all houses has helped school going children in their studies and women in their household chores. Earlier, women used to walk miles to the neighbouring village for milling grains at least once a week. Now a flourmill has been installed in Hosahalli itself. The unique feature of the project in Hosahalli is equitable sharing of benefits by all the households and reliable provision of services on most days in a year, contributing to improved quality of life for all and counted as social benefit. The economic benefits include employment and income generation and increased crop production. Establishment of energy forest, harvesting, transportation, wood-fuel chips, preparation and operation of the decentralized power generation system has created employment for two persons on most days and many more during different seasons. In Hosahalli 17 farmers irrigated 20 acres, growing labour-intensive cash crops such as vegetables and mulberry, thus creating employment and generating income. Environmental impacts are also considered. Raising multi-species energy forest has led to soil and water conservation in the lands subjected to degradation. If mixed species forestry, as done in Hosahalli, is adopted, it will contribute to biodiversity conservation in degraded or wastelands. Biomass is low in ash compared to coal, leading to no or insignificant ash production. Finally, biomass combustion leads to insignificant sulphur emission. In the

A Review of the Socio-Economic and Environmental Benefits…

569

absence of biomass gasifier-based power supply, Hosahalli village would have used kerosene for lighting and diesel engine for pumping water for irrigation or would have been connected to the centralized grid, where nearly 70% of electricity is generated from coal power plants. Diesel, kerosene or coal combustion leads to emission of CO2 and other greenhouse gases. In Hosahalli, if all the services were to be provided by diesel-based decentralized generation system, the total diesel consumption for generating 18,900 kWh of electricity annually (average of 2001 and 2002) would be 12,995l. The CO2 emission from diesel use avoided would be 35 t CO2/yr. The Hosahalli case study shows the potential of sustainable biomassbased decentralized power generation system for mitigation of greenhouse gases in small villages (< 500 population), which dominate rural India. Another two successful projects in West Bengal are explained below. West Bengal Renewable Energy Development Agency (WBREDA), Calcutta India has awarded the contract for Electrification of Chhotomollakhali Islands in the Sunderbans using biomass gasifier to our associates in India. This is the second such project with a power plant rating of 500 kW.

2) Chottomollakhali Project Chottomollakhali Island in Sunderbans is situated in the district of South 24 Parganas is about 130 km away from Kolkata. It is difficult to extend grid electricity to Chhottomollakhali Island due to prohibitive cost involved in crossing of various rivers and creeks. In the absence of electricity, the economic activities of the island were suffering. The Biomass Gasifier based Power Plant (four modules of 125 KW each) on 29th June 2001 has been set up set up by West Bengal Renewable Energy Development Authority (WBREDA) and it is catering to electricity needs of domestic, commercial and industrial users. Four villages (Chhottomollakhali, Taranagar, Kalidaspur and Bodo Mollakhali) of the island are benefited with electricity from the power plant (Mukhopadhyay 2004). Table 3.6. Power plant in detail Lifetime of the plant Plant capacity Average power generation Internal loss Line loss final selling average No. of consumers Hours of operation Fuel consumption pattern under full load condition Length of distribution line Substations Name of the manufacturer of gasifier

Source: Mukhopadhyay, 2004

15 years 4x125 kW 400 kW 100%(40kW) 5%(20kW) 340kW 225 (industry:1, commercial:74, household:150 5PM to 11PM (a) Biomass (80%) (b) Diesel (20%) HT line-5kms and LT Line-7kms 4 M/S Ankur Scientific Energy Technologies Ltd.(ASCENT) Baroda, India

570

Kakali Mukhopadhyay

The power plant detail is given in Table 3.6. Average generated power is 400 kW. The generated capacity depends on the local demand factor. It shows that the demand peaks on Monday (450 kW), which is a ‘Hat baar’ and falls to 350 kW on Sunday. To run the gasifier plant, woody biomass is being bought from local markets with an average price of Rs. USD 0.02/kg. This BGBPP needs on an average 1-kg woody biomass per unit of power generation (per kWh). The survey period requirement of wood is 400 kgs/day (drywood), which is met by the purchase from the local market. Since the woody biomass content moisture (10–15%) the amount of purchase is higher than 400 kg. To ensure steady supply of woody biomass for the project and the gasifier programme to be environmentally viable needs to be integrated to the plantation/ afforestation programme to guarantee the source of raw material without a threat of deforestation. It will also help to maintain eco-balance through replantation. WBREDA has initiated the programme and planted quick growing plants like eucalyptus, sirish, tetul and babla (energy plants) in the riverine wasteland. In future, the wood from these trees will be used as one of the fuels for this plant. This program has also created employment opportunities and villagers are normally employed according to the requirements of the power plant. WBREDA has set up a 10 ha energy plantation in 2000–2002. The average cost to set up this plantation is USD 10235.41. Rotation period is estimated to be 5 years and the average production is 5 t /ha/ year as estimated. Three types of tariff exist for three categories of consumers, i.e. USD 0.10/kW for industry, USD 0.09/kW for commercial and household are paying USD 0.08/kW in consultation with the village community. Revenue collection is administered by WBREDA. Total average monthly revenue generated varies from USD 511.77–USD 614.12. This amount is utilised in plant operation and maintenan « (OandM) purpose. The deficit expenditure of plant’s operation and maintenance is met by Government of West Bengal. It has been revealed that the gasifier has influenced the economic activities and quality of life of the villagers in a number of positive ways. The benefits generated through the programme over the year are also estimated. Before the introduction of gasifier power for lighting, the commercial units used to depend on power supply from privately owned a few 10 kW-diesel power generator set. The household sector was completely depended on kerosene and in very rare cases on solar power. Use of kerosene is very inefficient use of fuel in India. It has already a high scarcity value and is very expensive. The additional scarcity value of diesel in the study area arises from the typical geographic location of it by way of high cost and problems in transportation of the fuel in a boat from the nearest mainland canning. The high scarcity value of diesel gets reflected in the high cost of unit of power generated by this privately owned generator supply. The inefficiency of diesel use per unit of power generation was USD 0.49. But the cost of per unit of gasifier-generated power is USD 0.09. Excess payment is an indicator of economic inefficiency. Not only the fuel efficiency of the diesel power generating sets are low they also have many environmental effects in terms of noise and air pollution. The efficiency derived in case of fuel use pattern for cooking and lighting has also been estimated. The market demand for fuelwood is gradually declining by the domestic users and will benefit a BGBPP in the long run. This reduction in cost of power supply can be assessed from the savings in kerosene, diesel and by pulling out generator set that has been made possible

A Review of the Socio-Economic and Environmental Benefits…

571

by the gasifier power. In the post-gasifier period inefficient use of diesel generation set has been completely eliminated for the commercial units (Table 3.7). As none of the commercial units are continuing with the decentralised diesel set using appliance very few of household and commercial sectors are still continuing with their use of kerosene (which is as an input to light the lamp power) as a source to supplement gasifier power to meet the unmet part of power demand from gasifiers. For the domestic sector the supplementary sources are kerosene and solar power. The cost of diesel use per unit of power generation is USD 0.49 but the introduction of the gasifier has reduced it to USD 0.09 per unit. Thus during the post-gasifier period commercial units as well as households are benefited. In terms of simple economic cost it can be said that holding all other components of cost of power generation constant the use of new technology has been successful in bringing down the fuel cost of power from USD 0.49 to USD 0.09 per unit of power generated. Table 3.7. Fuel uses pattern for lighting Commercial Pre-gasifier None 100 None None

Kerosene Diesel Solar Gasifier

Post-gasifier None None None 100

Household Pre-gasifier 83 None 17 None

Post- gasifier None None None 100

Source: Mukhopadhyay 2004

This is an indicator of benefit from the installation of gasifier power. The gain in efficiency has been reflected through gain in consumer’s welfare. The beneficiaries have expressed several kinds of benefits, tangible and intangible. The benefits which can be quantified and which have influenced the quality of life is discussed below (Table 3.8 and 3.9). Table 3.8. Benefits of commercial units Type of benefit Savings in monthly exp(USD) 0.02-1.02 1.02-2.40 2.40-3.07 3.04-4.09 4.09 and above Increase in business hours Nil <1 1-2 >2 Connected points Pre-gasifier 1-4 45 5-10 5 >10

Source: Mukhopadhyay, 2004

Number of units reporting

% in sample

9 10 15 9 7

18 20 30 18 14

1 6 32 11 Post-gasifier 12 35 3

2 12 64 22 Post- gasifier 24 70 6

Pre-gasifier 90 10

572

Kakali Mukhopadhyay Table 3.9. Benefits of household Type of benefit

Number of units reporting

% in sample

1.02-2.40

16

32

2.04-3.07

18

36

3.07-4.09

11

22

4.09 and above

5

10

Savings in monthly exp.(Rs)

Source: Mukhopadhyay, 2004

The quantifiable benefits are those, which emerged due to gain in efficiency in generation, and hence reduction in price per unit of power supply set at the consumer end. The price paid by the consumers in the pre-gasifier period to the diesel generators was much higher and arbitrary. A policy with correct economic incentive will be one, which does not lead to overuse/wasteful use/overstress on existing supply set. Essentially an appliance type uniquely represents a particular level of power consumption at certain instants of time. Assuming identical supply quantity at the distributor end a consumer with 100-W bulb consumes 100-W/h of power, while one with a tube consumes 40 W/h of power. But under pre-gasifier regime since the different consumer categories were supposed to pay the uniform price per day both the categories as mentioned above had to pay the same price per day despite different levels of consumption. It is noteworthy that given the lump sum nature of the price charged there was no incentive to conserve power at the consumer end through use of efficient appliances. The pricing structure was unable to distinguish between efficient users and inefficient users giving rise to inefficient system of pricing and appliance choice. Inefficiency of pricing has been removed in the gasifier regime. Per unit price of gasifier power is less than that of diesel power. An economically efficient tariff structure in post-gasifier regime has been successful in encouraging consumers to choose efficient appliances. Although in post-gasifier regime their total units of power consumption has gone up due to more connected points. Still compared to the pre-gasifier their total cost burden is much less. The table above indicates the range of actual monthly savings for beneficiaries. These results from efficient appliance choice by consumers as well as effective reduction in unit cost for gasifier power compared to diesel generated based power. All commercial units connected to the gasifier supply generate a collective illuminationgiving rise to external benefits for the villagers. This has, in fact, given rise to business hours beyond 6:30 PM or even 7:00 PM, which was usually the standard closing time for the pregasifier period. Out of the total surveyed commercial units 70% have reported that increased business hour has been productive to them by way of increase in the total sales and hence a more profitable turn over. This qualitative improvement can be quantified in terms of gain in total turn over. Availability of gasifiers power at a cheaper rate has in turn given rise to power consumption through purchase and use of larger number of lighting appliances reflected through their demand for a larger number of connected points compared to pre-gasifier scenario. It shows a significant saving in monthly expenditure of electric bill with an average of USD 2.67 for each commercial unit. At the same time, there has been a considerable increase in business hours and number of connected points leading to an average increase of

A Review of the Socio-Economic and Environmental Benefits…

573

monthly business turnover of USD 20.47. So on the whole post-gasifier period has brought in a propitious time for the commercial sector of this village island. Direct qualitative benefits gained by households can be seen from the savings in the monthly expenditure on power. Some of the households could now afford to buy durable consumer goods like radio, T.V, electric iron etc with the savings generated by the installation of BGBPP. They are aware of children’s comfort and enhancement of study hours during night. The study concluded that the consumers are saving on an average USD 2.67/month in post-gasifier period. From the field survey it has also been noted that the household consumers are now using on average eight plug points. Prior to gasifier power to save kerosene or diesel power, nighttime activities of the villagers were almost non-existent. Now with gasifier power available upto 11:00 PM., us ël household activities like watching T.V, socialisation, etc. have increased and thus improved the quality of life. The households are aware of the benefits and are demanding longer hours of service from the BGBPP. In this connection we have to mention that the benefit (or monthly saving) incurred from the BGBPP by the household will help them to switch over from traditional chullah to kerosene/LPG type. More specifically, beneficiaries are now willing to spend the extra amount they have saved from electricity on the efficient type of fuel for cooking. Overall this benefit will indirectly save time and health costs for households. A huge potential demand for BGBPP power has been observed. Moreover problems also have been identified due to limited hours of power supply. All the surveyed beneficiaries and the non-beneficiaries have expressed their willingness to pay the existing rate of tariff despite their awareness that it is 2.5 times higher than conventional grid power tariff of West Bengal. Fifty percent of the surveyed non-beneficiaries have shown interest to take up the BGBPP lines but the lines have not yet been extended due to transmission and distribution constraints. A huge demand from potential consumers has also been experienced. The further extension of transmission and distribution would be justifiable given the density of settlements in the neighbouring areas. It was assessed that with increasing expansion of the TandD network unmet demand would be minimized in the process. It is found that all households and commercial beneficiaries would like to receive power increased from 6 to 24 h. This increased power supply will lead to more profitable turn over for the commercial sector. Though the average income of the household is not high still they are willing to pay more because it is expected that increased power supply will help the students to study for more hours comfortably and use of more electric appliances leading to further improvement of quality of life for the villagers. The findings of the study (Mukhopadhyay, 2004) indicate that the BGBPP has made a very positive impact on the life of the villagers of Chottomollakhali Island. This has led to increased economic activities and more profitable turn over for the commercial consumers and improved quality of life for the household sector. The consumers now are gaining in terms of lesser electric bill expenditure. All of them have shown a willingness to pay a higher price to get 24 h of power supply. From the BCA(benefit cost analysis) it is evident that the plant can recover its full investment after 7 years. It has also been observed that the BCR is greater than 1 (1.68). Moreover the IRR is 19%, which is more than the cost of capital (8%). Assessing above criteria we may conclude that the project is economically viable. Though the environmental benefits have not been measured in the BCA but the study finds that the BGBPP has a positive environmental impact. But the environmental awareness is very poor among the villagers.

574

Kakali Mukhopadhyay

3) Gosaba Project The first rural electrification demonstration project in the country considered to be a success is at Gosaba, an island of about 156 sq. kms. area in the Sunderbans area in the state of West Bengal in 1997 (Akshay Urja, 2005). Gosaba was selected as a site for rural electrification based on decentralized supply sources, as this was the only option for this region. In addition, the area also has an abundance of biomass resources. The project was implemented by WBREDA in association with MNES, Sunderban Development Department (the local development body), Forest Department and South 24 Parganas Zilla Parishad (local administration). MNES subsidized 75 percent of the project cost and state government gave the remaining. The state electricity board has set up the distribution network, with financing from WBREDA. The total electricity generation capacity is 500 kW, with five individual gasifier-based units of 100 kW capacity each. The gasifiers are closed-top downdraught systems based on woody biomass, supplied by Ankur. The plant has two dual-fuel engines that are synchronized with the system and can be operated in parallel. The entire project cost was Rs. 10 million, including setting up of TandD network. The investment for the distribution network in Gosaba amounted to Rs. 1.8 million. The capital cost for the gasifier installation approximates Rs. 25 million per MW. The transmission and distribution line spans over a length of 6.25 kms of high-tension lines and 13.67 kms of low-tension lines, with a cost of around Rs. 175,000 per km. The electricity generation in the plant is at 400V. Within the TandD network, around 45 to 50 consumers are connected every kilometer. The average TandD losses are only 4 percent. At present around 900 consumers are being provided with power 16 hours a day. In the initial stages of the project, a single 100 kW gasifier unit was installed as the existing load at that time was just around 10 to 20 kW. The villagers were reluctant to participate and there were only around 25 consumers of power. It took some time to convince a local people of the potential benefits and the load growth took around a year. The operating load of the system is 300 kW. Therefore, a maximum of three gasifier units are operating at one point of time and the other two units are kept as standby. The average daily generation is 950 units over the period of 16 operating hours. The tariff for domestic consumers is at Rs. 5/kWh, for commercial shops and establishments it is Rs. 5.50/kWh and for industrial consumers Rs. 6/kWh. The average household consumption is in the range of 1 to 3 units per day. The households with electricity supply connection have to pay a fixed charge of Rs. 75 per month in addition to the variable charges for the units consumed. The monthly revenue generation at Gosaba is around Rs. 160,000. There have been no defaulters in payment of electricity bills and no electricity thefts are reported. The charges are affordable to the users and they are willing to pay the price for reliable electricity supply. Before the Gosaba project, people in this area were paying around Rs. 9/kWh for diesel-based generation. The state nodal agency’s assessment is that demand for electricity among villagers is increasing steadily as more local industries are coming up. The plant is run by a local co-operative society, which receives funds from WBREDA. This cooperative, which is responsible for ensuring biomass supply, daily plant operation and maintenance, and financial record keeping. For undertaking renovation, repair and maintenance of plants, around 75 percent of the financing comes from the co-operative and

A Review of the Socio-Economic and Environmental Benefits…

575

the rest from MNES. The success in running this rural energy co-operative is partly attributed to the history of success in co-operative movements in that area. Ankur undertook turnkey operation for the project and at present intervenes in major maintenance and retrofitting functions. It has trained local people in plant operation and maintenance. It also periodically reviews plant operation. One of its service engineers based in eastern India supervises these activities. The state nodal agency, WBREDA, functions as the Technical Backup Unit for the project. It provides both technical and non-technical support for running the system, monitors the operation of the plant and performance of the plant personnel. It periodically conducts tests for the plant operators to monitor their performance. The socio-economic development of Gosaba after electrification has been immense. The availability of electricity has allowed students to study at night. Small-scale industries ( for example lathe machine units, boat-repairing works, and grill welding , domestic iron implements sharpening machines and machines to grind spices like chilli and turmeric, using automated electricity-operated machinery) have been established in the region. Establishment of Xerox machines in shops has facilitated photocopy of costly study books and to make copies of important documents like deeds. An opera÷Áon theatre has been made functional in the Government Health Centre in the Island. With the availability of refrigerators it has become possible to store life-saving vaccines or medicines. Water from ponds is being drawn with the help of electricity- operated pump motors for the purpose of irrigation and cultivation of crops Availability of electricity has given way to entertainment. People are able to watch sports and other programmes on cable television, which were not thought to be possible earlier. Cinema shows are being organized by the local video parlours. A computertraining institute has started in Gosaba (Akshay Urja, 2005). Nowadays people need not go to Canning for any printing press related jobs, or sizing the lens of spectacles. Time is being saved, and the expenditure and risk of travelling has reduced. Electric sewing machines are being used to manufacture fishing nets for the fishermen. Tailoring shops are using electric press. The local shopkeepers and traders have been able to decorate their shops and establishments through strategic illuminations to attract customers. Presently, the income levels have also increased due to increase of sales. Electronic goods shops have proliferated in the region where all kinds of electrical goods are available. Many telephone booths have been established. The scope of work has also increased for people of the region, and there has been a huge improvement in the communications system. The use of wood-generated electricity has upgraded manpower resources, and led to an overall development of the region. The biomass gasifier power plant has been a boon for the people of Gosaba. It has brought about an overall development in their lives. The villagers have really benefited from the above three power plants located in southern and eastern part of India. In this context the paper also evaluates the two power plants located in northern part of India.

4) Producer-Gas Electricity for Six Villages in North India A village electrification programme using biomass gasification has been implemented in a cluster of six villages near Kotdwara (Garhwal region, North India), by the Himalayan Environmental Studies and Conservation Organization and the Indian Institute of Technology (IIT), Delhi. A brief review of the biomass energy project in a cluster of villages around

576

Kakali Mukhopadhyay

Kumbichar village based on (Ravindranath et al., 1995). A 3.5 kW gasifier was installed in Kumbichar village, which uses mainly Lantana camara, a wild shrub growing in the village, as feedstock. It is a small village and all 18 houses are provided with electricity for lighting form a 3.5 kW system. During the day, the system is used to operate a flour mill, spice grinding, and a leaf plate making machine. The system is operated by trained village youth. Each household is to pay Rs.20/month for lighting. The operational cost of producer-gas electricity generated was Rs0.93/kWh compared to the grid electricity rate of Rs1.38/kWh (1992 price). A diesel substitution of 70 per cent has been achieved. The system has been expanded to five more villages. An interesting aspect of the project is the use of the relatively low-density shrub Lantana camara. Lantana grows wild in many parts of India and thus using it for power generation in this demonstration project provides an excellent example of using such a hardy, quickgrowing, and copping weed for economic gain. It is also relevant that the households collect and cut Lantana sticks for gasification each day, thus avoiding the growing and collection of wood. This contribution of the village community demonstrates their commitment to the programme and reduces the investment and operational costs.

5) Nari Project Work done at the Nimbkar Agricultural Research Institute in Phaltan, Maharashtra, has shown that each taluka in the Phaltan district produces enough agricultural residues so that all its electricity demands can be met by using them in 10-20 mw biomass-based power plants (Rajvanshi,1995). The NARI study also showed that besides providing power, the taluka energy self-sufficiency plan could also create 30,000 jobs/year. With the new Electricity Act, taluka energy self-sufficiency can become a reality since the utility can produce and supply power to its customers without the need to go through SEBs. The taluka utility company can also lease the existing transmission and distribution infrastructure of SEBs so that it need not invest in developing its own. This will also help the SEBs to get regular income from their infrastructure. The NARI study also showed that the taluka energy programme could produce Rs 100 crore/year wealth for its inhabitants in terms of biomass production and setting up of new electricity-based industries. With about 3,500 talukas in the country it is therefore possible to produce about Rs 3, 50,000 crore/year extra wealth through the taluka programme. NARI has recently suggested this concept to the Maharashtra Electricity Regulatory Commission. It is also envisaged that electric cooperatives may function on the lines of TV cable operators in rural areas. However, for small power packs of 500 KWe and less to function smoothly in rural areas it is necessary that they be powered by fuel from locally available resources. Thus there is a need to do sophisticated RandD in producing biofuels from renewable energy sources. These biofuels can easily power the existing diesel gensets. Development of liquid fuels like ethanol and bio-diesel from multipurpose crops should be done so that the issue of food and fuel from the same piece of land is taken care of. This will help in creating fuel supply network for small rural based utilities. Besides it will create wealth in the rural areas by producing value-added items like liquid fuel from agricultural residues. Apart from the rural electrification there are several other successful stories of biomass gasification which can also be mentioned in this respect.

A Review of the Socio-Economic and Environmental Benefits…

577

a. The Desi Power Unit at Orchha The first DESI Power plant was installed in 1996 at Orchha, near Jhansi, a five hour train ride south of Delhi. This plant has now been operating for more than three years (SEI, 1999). It is located at and supplies power to TARAgram, a campus where TARA carries out research, demonstration, training and production activities using appropriate technology. Although TARAgram has acquired a grid connection, the bulk of its electricity requirements are still met by the DESI Power plant. As for all DESI Power stations, a careful site analysis and feasibility report was prepared for the Orchha plant, including an assessment of electricity demand, availability of renewable fuels and existence of interested partners for setting up an IRPP. On the basis of the study regarding current energy sources purchasing power of households small business opportunities in the region, accessibility of local investment capital and availability of skilled labour and availability of local biomass and 80 kW power station was ordered in December 1995. The DESI Power Orchha unit, located at TARAgram, went into operation in April 1996. The capital cost of the station was Rs. 22 Lakhs. Initially, the sole client was TARA. Ipomea was used as the fuel. An 80 kW plant needs almost one tonne of biomass fuel every day. Managing biomass resources properly is crucial for reliable plant operation. More than ten local families have created new livelihoods through harvesting, chopping and transporting the weed and delivering it to the power station, where it is sun-dried before being fed into the gasifier. However, it is estimated that the cost of fuel is adding some 20% more than it should to the final cost of electricity produced. The current indications are that the quantity of biomass accessible to the plant in its first year should be at least 50% greater than the long-run annual requirements. The plant is satisfactorily operated and maintained by local personnel. Local people, with fairly minimal education is being trained and employed for the tasks. Further employment of over 100 persons was created through the factories that purchase electricity from the Orchha plant. Over the past three years, the gasifier has logged about 5,000 hours of operation, running nearly 10-12 hours per day during normal operation. The percentage of diesel replaced by gas depends on the plant load factor. It has been recorded as high as 85%, with the average at about 75 per cent. Competitive pricing of electricity will require DESI Power stations to maintain an average of well above 80%. Plant load factor is a critical parameter to which the economics of power generation is extremely sensitive. The plant load factor is important not only for spreading out the fixed costs, but for raising the diesel replacement rate, reducing the use of costly diesel fuel. The breakeven plant load factor for the Orchha plant is between 50% and 60%. Even with 40% load factor and high biomass cost, the cost of electricity has been Rs.4.0 to Rs 4.5 per kWH, which is still competitive with electricity from the grid. The Orchha plant has shown that the cost of electricity can be highly competitive with conventional systems, provided that investment capital is obtained on terms similar to those available to larger Independent Power Producers. The two most sensitive determinants of production cost are the cost of biomass fuel and the plant load factor. Efforts are being made to raise the Plant Load Factor at Orchha by adding new clients and industries with energyintensive applications, reducing the high fluctuations in some of the loads staggering certain activities.

578

Kakali Mukhopadhyay

b. Valli Chlorate Company at Kovilpatti, Tamil Nadu Valli Chlorate Company has been chosen as a successful case study by installing 100% producer gas based gasifier. The company has also increased its profitability by installing the gasifier by reducing cost of energy purchased from TNEB(ITCOT,2005). The company sells Potassium Chlorate at a rate of Rs. 36 per kg. The production cost with EB Power Supply was around Rs. 34 to 35 per kg results in a meager profit margin. Due to gasifier installation, the production cost of Potassium chlorate has come down to Rs. 28 to Rs. 30. This has helped the plant in improving their profits. The operation of the unit is for 8 hours a day and extends some times to 10 hours a day and six days a week. c. M. Vishveswaraiah Institute of Technology Installation at Bangalore, Karnataka M. Vishveswaraiah Institute of Technology (MVIT) has sought the assistance of Netpro Renewable Energy Private Limited, Bangalore to install a Dual Fuel gasifier-based power plant in their college premises and to supply power to the college. In the educational institutions category, MVIT is the oldest gasifier unit operating successfully for the last three years. The unit also operates for longer period in a year and has generated more units when compared to other units. Netpro is presently successful in providing captive power to the college to the greatest satisfaction. d. Bagavathi Bio Energy Limited Installation at Coimbatore, Tamil Nadu In the industry sector, the unit that is successful in implementing the gasifier program is Bagavathi Bio Energy installation at Mettupalayam. The company has successfully been running as a separate entity to serve the energy demand of a Textile bleaching unit by installing the system in their premises without any major problem (ITCOT,2005). Though other units covered by the study were also operating the gasifiers, the data available from the units other than the above four is scanty and further, units were also not willing to share the real performance of the gasifiers. Some units found the operation not economical on dual mode due to high cost of diesel, and hence, gasifiers are not run. e. PSG College of Technology, Coimbatore through Ministry of Non-Conventional Energy Sources (MNES) New Delhi implemented a 1x 100kW biomass gasifier on electrical mode at the PSG foundry campus, Neelambur to partially meet the foundry electrical energy requirement and also to promote gasifier system based academic research projects at the college. The plant is commissioned in 30th July 2004(ITCOT, 2005). f. Jagat Alloys Private Limited (JAPL), Tamil Nadu (JAPL) has set up only 1 x 500 kW gasifier to meet the load of Ferro Alloy Plant. The plant is commissioned on July 4, 2001 at Nellithurai, Coimbatore district. The project is commissioned to meet the energy requirement of the raw water pumping station at Nellithurai Municipality and the lighting of the surrounding area. The total power requirement is 3 HP and 2 HP for water pumps and a lighting load of 3 kW(ITCOT,2005).

A Review of the Socio-Economic and Environmental Benefits…

579

g. Vellore Institute of Technology (VIT), Tamil Nadu has installed a 90 kWe Gasifier for generating power for captive consumption in the hostels and to carryout R and D activities in the field of Renewable Energy. VIT conveyed that overall performance of the gasifier was satisfactory (ITCOT, 2005). h. Periyar Maniammai College of Technology for Women (PMCTW), Tamil Nadu has installed 100 kW Dual Fuel Mode Gasifier system during 2001 and later installed 200 kW 100% producer gas based Gasifier system during 2004. The college has installed the gasifiers for captive consumption and for carrying out RandD activities in the field of renewable energy. The dual fuel based gasifier plant was installed during 2001 and the 100% producer gas based plant was installed on 24th June 2004(ITCOT,2005). From the above case studies it is reflected how far the village people get benefited from the rural electrification programme. It also gives us a good picture about the pros and cons of biomass gasifier power plant. The attempts so far initiated to install BGBPP in India are mostly with other purposes without implementing rural electrification programme intensively. Moreover, according to the statistics most unelectrified villages are in Bihar, Rajasthan and Madhya Pradesh.

4. ASSESSMENT OF THE CASE STUDIES The case studies discussed above predict the feasibility of the biomass gasifier based power plant in rural remote villages in India. Several benefits have been derived from the proper implementation of the gasifier plant. The Hosahalli and Hanumanthanagara, case study has demonstrated the technical and operational feasibility of a decentralized biomass gasifierbased power generation system to meet all the rural electricity needs in an environmentally sustainable manner. It is one of the oldest applications of rural electrification programme. Another feasible option is Chottomollakhali situated in eastern part of India indicates that BGBPP has made a very positive impact on the life of the villagers of Island. This has led to increased economic activities and more profitable turnover for the commercial consumers and improves quality of life for the household sector. Further, Gosaba Island installation has also been a successful case study for the reason that the energy generated from gasifier has turned the socio economic status of the Island along with community development. Even when the cost of energy generation is high, the energy generated from gasifier has played a pivotal role in improving the life style of the rural masses of Sunderbans, where grid penetration is impossible. Several other feasibility studies(BERI, ITCOT,2005) on biomass gasifier based power plant not exactly for rural electrification but dealt with the objectives like water pumps system, cooling machine, rubber drying, cardamom drying, silk reeling machine for convenience of the village activities. The above application of the biomass power plant also helps to generate employment opportunities and reduce poverty in the village. But some of these villages are not yet electrified, though there are biomass gasifier activities running for other industrial purposes. In our opinion these villages should be first in the targeted list for electrification. It will be easier also to introduce the biomass plant in that area.

580

Kakali Mukhopadhyay

Whatever the programmes so far taken by MNES and other state governments regarding electrification through gasifier , mostly concentrates on southern part of India and }¶w in eastern part. So the regional variation is a dominant factor in this regard. Moreover the attempts of biomass gasifier plant in southern part of India mostly applied on small and medium scale industries. So there are rooms for trading off between rural electrification and Small and Medium Enterprises facility. These vital issues need to be addressed by the Government of India according to priority basis. Further, forestry in degraded lands is a well-known climate mitigation option to sequester carbon in soil and standing trees. Biomass power based on sustainable biomass supply is a climate mitigation option for substituting fossil fuels (coal used in power stations, kerosene and diesel), leading to reduction in the emission of greenhouse gases. This is a very positive contribution from the environmental point of view. However the most important issue is cost related. There is a general perception that the cost of electricity generated by renewable energy technologies is always higher than electricity generated by fossil fuel sources. While this is true in many places, it is no longer valid for the rural areas of l India. In most of the Indian villages diesel generators are often the only source of power but power from biomass gasifier based plants are considerably cheaper where ever biomass is available. Even for dual fuel operation where 20 % diesel is used, the generation costs are lower, especially with high running hours and loads. The savings are high when pure gas engines are used. Even when grid power is available, the actual cost of power at the point of consumption is very high largely due to line losses in transmission and distribution. High subsidies and financial losses keep the power price low for agricultural pumps but now that industrial and commercial consumers pay the actual cost of power, the biomass gasification based electricity can easily compete when pure gas engines are used. Table 4.1 explains the capital cost differences among the renewables. It clearly shows the cost advantages in most cases. Financial Advantages of Decentralised Biomass Power Plants here are assessed by the DESI power in a separate manner. As the experience of DESI Power’s EmPower partnership programme shows, grid supply to remote areas is not competitive with electricity supply from modern decentralised renewable energy power plants (table 4.2). Table 4.1. Capital Costs and the Typical Cost of Generated Electricity from the Renewable Options Sl No. Source

1. Small Hydro-Power 2. Wind Power 3. Bio-mass Power 4. Bagasse Cogeneration 5. Bio-mass Gasifier 6. Solar Photovoltaic 7. Energy from Waste

Capital cost (Crores of Rs/MW) 5.00-6.00 4.00-5.00 4.00 3.5 1.94 26.5 2.50-10.0

Source: Planning Commission, 2005

Estimated cost of Generation per Unit (Rs./kWh) 1.50-2.50 2.00-3.00 2.50-3.50 2.50-3.00 2.50-3.50 15.00-20.00 2.50-7.50

Total installed Capacity(MW) 1601.62 2483.00 234.43 379.00 60.20 2.54 41.43

A Review of the Socio-Economic and Environmental Benefits…

581

Table 4.2. Cost of Supplying Power to a Village Generation MW TandD Losses End Use Energy MW

Cost Rs

MW

Cost Rs

MW

Cost Rs/MW

Centralised Grid Supply

1

35 million

0.3

5million

0.7

57 million

Decentralised biomass power plant (gasification) SAVINGS: Decentralised Vs. Centralised

1

35 million

0.1

5 million

0.9

44 million

Power Generation/End 0.2MW Use CO2 emissions

Power Saving

Avoided Cost

Saving in Cost

22%

13million Rs/MW

29.5%

Amount

5500t/y per MW

Source: Desi Power (2004)

Though the above explanation gives us a fair idea about the cost differences between decentralized and centralized power system and also several benefits which can be derived from the Biomass plant still the negative side cannot be ignored. The advantages and problems exclusively from the BGBPP are outlined below. 1) In India biomass technology is mature with several designs and manufacturers who undertake planning and commissioning of small-scale biomass power systems and who also provide performance guarantee. 2) Biomass gasifiers are available in different capacities for decentralized applications from 5, 20, 100 to 500 kW in India. 3) Biomass gasifier-based system can be operated from 1 to 24 h a day, depending on the load. The system can be operated 365 days in a year, if needed. Woody biomass feedstock can be transported over shorter distances and stored. 4) Biomass gasifiers can be installed and operated in any village where biomass is available or can be grown, except probably in desert areas. Such flexibility does not exist for other renewables such as solar, wind, micro-hydro and biogas systems.5) Biomass gasifier technology is indigenously developed and transferred to manufacturers. Maintenance, spare-part supply and servicing facility and infrastructure are available or can be organized. 6) The economic viability is yet to be proven for renewables in India based on monitoring of field-based systems. Preliminary assessments available show that biomass gasifiers are economically feasible and have lower cost per kilowatt hour compared to other energy technologies. 7) Biomass gasifier-based power generation systems create jobs and skills in rural areas in biomass feedstock production, transportation and processing, and in operation and maintenance of the gasifier–engine–genset systems as well as end-use systems. 8) India has vast degraded or wastelands (over 60 million ha), which urgently require revegetation to prevent further degradation. Biomass production, as feedstock for power generation, provides economic incentive to re-vegetate wastelands with energy forests.9) Appropriate guidelines to discourage monoculture plantations and incentives to promote mixed species forestry, with appropriate density will promote biodiversity in degraded lands. Use of forestlands to supply biomass or conversion of natural forest to energy plantations is not desirable and banned in India. 10) In India, coal-based power plants account for 70% of electricity generated. Thus, every kilowatt hour of electricity generated based on sustainable biomass supply, leads to reduction in CO2 emission by 1.0 kg. Biomass power is recognized

582

Kakali Mukhopadhyay

as a prime option with high potential for reduction in carbon emission and climate change mitigation. Despite the above advantages the rate of spread of biomass- based power generation technology in India is low, due to a number of policies, institutional and financial barriers. There is need to address such barriers to promote biomass power in India.

Problems Encountered from the BGBPP A decentralized power generation system, implemented in a village could face a number of technical, social, political and financial problems over the period of its life. Here, some of the broad problems encountered particularly focusing on the performance of the plant are presented.

Technical Problems With respect to the technology package, the two critical components are the gasifier and diesel engine. These include the failure of material in the top shell (twice), filter replacement and grate repairs. The gasifier maintenance-related problems which related to the operational problems, use of moist fuel, wrong size fuel, etc. affecting the quality of gas. Engine-related problems which are related partly to gasifier operation and also to the failure of the radiator. Electrical system of the alternator failed twice. These give an indication about the reliability of different components of the system. Some of the gasifier-related problems have been addressed by the RandD group. The problem related to the reactor material has now been resolved using high temperature ceramics and also the filter failure using fabric filters. Input Supply The utility does not have adequate storage facility. Thus during rainy seasons the cutting and drying of wood is a problem. The labour availability for harvesting wood in the energy forest and then chopping to the size needed is limited, particularly during the peak crop season. The power generation system was operated on diesel mode during such days, when cut and dried wood chips were not available. The nearest diesel source is about 30 km away and often diesel is not available or accessible to the operator. The system could be operated on days when diesel was not available. The mitigation measures for sustained biomass feedstock supply include: creation of storage capacity, mechanical system for cutting wood to the desired size and drying of wood using the exhaust gas of the engine. The ultimate solution to the diesel problem is the adoption of a ‘gas engine’, which is at an advanced field testing stage. Non Availability of Operator Due to financial constaints fully trained operators necessary for the running of the plants could not be employed in most cases. Social Problems The social problems include some disagreement among the members of the management committee or political rivalry. Social problems-related disruption was infrequent; this is a tribute to the most of the village community. The other social problems like unauthorized removal of trees from the forest occasionally, grazing of livestock in the forest, attempted

A Review of the Socio-Economic and Environmental Benefits…

583

encroachment of the forest land and opposition to the operator from one section of the village community. It is no surprise that the village communities in India are divided over political affiliations, castes, land ownership, etc. But what is surprising is that despite the social and political problems and lack of their understanding by the most of the project team, the decentralized system has functioned uninterruptedly.

Subsidy and Government Policy Related Problems While the government program has relied mainly on subsidies and an orientation towards target fulfillment, it has had little emphasis on performance. Furthermore, the implementation of the government effort has not been driven by need assessment and performance evaluation. In fact, there has been little emphasis on systematic review of the program. There are also distortions in subsidy policies in terms of the structure and nature of subsidies. Some of these arise due to subsidies being applicable even for commercially viable applications of gasifiers such as productive uses in industries and region-wise (higher subsidy to locations in the north-eastern regions) and category-wise (higher level of subsidy offered to consumer categories belonging to certain socio-economic classes) classification of subsidies. Frequently changing government policy guidelines with respect to subsidies also results in awareness problems among users. The installations of systems are often driven by subsidy motives and draw little commitment from users. There has been no shift from capital subsidy to performance-based incentives such as soft loans and tax credits. Bureaucracy The procedures for government subsidy approval and disbursement are lengthy and cumbersome that deters potential beneficiaries (although to be fair, this is not a particular problem for the renewables areas only). A bottom-up structure exists related to project development and implementation that leads to high cost and time overruns due to factors such as procedural bottlenecks and approval needs from multiple agencies. Finally, there are also adverse impacts due to uneven support to RandD institutions and sudden withdrawals in government support without alternate support in place. For example, after the dismantling of government support for GARPs, there remains no agency for testing and certification of gasifiers. There are uncertainties with respect to resumption of these activities that adversely affects dissemination. Thus,, an overall assessment lead us to state that the Gasifier based Power Generating systems are technically feasible and financially more viable, when run on 100% producer gas mode and not on dual fuel mode. The project would be more successful if the availability and cost of biomass used in the system are kept at constant check. Energy Plantation would be the most feasible and reliable option for adequate and cheaper biomass availability. Constant Load operation and increased Plant Load Factor (PLF) would further improve the financial viability of the Project.

584

Kakali Mukhopadhyay

5. CONCLUSION AND POLICY OPTIONS The objective of the present paper is to evaluate the rural electrification programme in India undertaken by the Ministry of Non Conventional Energy Sources (MNES), Government of India, through biomass gasifier power plant. Among the renewable energy technologies, biomass gasifier-based decentralized power generation holds great promise for meeting rural energy needs. A number of case studies have been presented and these case studies more or less establish the feasibility of the gasifier from various fronts. On the one hand it provides commercial, social and environmental benefit, on the other hand it helps to achieve better quality of life through creation of jobs in the power stations, small-scale business, commerce and industries. We have discussed that the Government of India had already undertaken a number of strategies and policies for rural electrification programme. Due to more feasible option, biomass as considered, Government has taken one step further initiatives in this regard. Among the various programme, the Rajiv Gandhi Gram Vidyutikaran Yojana (RGGVY), Pradhan Mantra Gramodaya Yojana, Distributive Generation, Kutir Jyoti Programme deserve mention. The biggest challenge in this regard is to electrify 62,000 villages through grid power, during the 10th Five-Year Plan (2002-07) under the Pradhan Mantri Gramodhaya Yojna. The Centre hopes to achieve 100 per cent village electrification in the current Plan. Another 18,000 remote villages will be electrified through decentralised plants based on biomass, gasification of biomass, hydel power, solar thermal power etc. We have also discussed the problems related to biomass gasifier rather than rural electrification. Unless we introduce proper policies and implement them, our objective will not be fulfilled. Towards this end the paper is suggesting a few policy options.

Suggested Policies The paper suggests three steps policy suggestions. First the broad policies; secondly a few subsidy oriented policies and finally some specific policies have been suggested.

Broad Policies Renewable energy including biomass may need special policies to encourage them. This should be done for a well-defined period or up to a well-defined limit and should be done in a way that encourages outcomes and not just outlays. The programme can be accelerated if the support of the government and the budgeted public funds were used to leverage local, private and corporate sector investments in these rural projects. A policy framework can be established for utilising sanctioned funds earmarked for renewable energy based rural electrification as well as for other rural development programmes (e.g., for kerosene, small scale industries, job creation schemes, etc,) in a more focused and integrated manner. An annual renewable energy report should be published providing details of actual performance of different renewable technologies at the state and national level. This would

A Review of the Socio-Economic and Environmental Benefits…

585

include actual energy supplied from different renewable options, availability, actual costs, operating and maintenance problems. The monitoring should also encompass other parameters like user profile (in order to ensure that government support is indeed going to poor households), livelihood outcomes such as increased income, improved food security and gender impacts. In addition to monitoring the performance of devices the assessment should critically review the programme objective and the strategy adopted and suggest course corrections as required. Information on any system that is receiving government support should be made publicly available. It is essential to ensure that independent assessment of performance is done for all renewable projects receiving Government funding. This help in tracking programmes, repeating mistakes and providing mid-course correction. Apart from the facilities of existing Technology Business Incubators being set up by the DST energy entrepreneurs for renewable energy, energy efficiency, or rural energy also need finance. Financial institutions should be encouraged to set-up venture capital funds for energy entrepreneurs.

Subsidy Oriented Policies It is well known that capital subsidy plays a very important role for rural electrification in India. A critical issue in distributed generation for rural electrification is the cost recovery and the implementation mechanism. Different policy experiments for implementation of DG appropriate technology option (biogas, bio-mass gasification) for their village. For isolated systems it is beneficial to link the DG system to an industrial load (cold storage, oil mill etc.) to improve its load factor and hence it’s economic viability. The capital subsidy should be based on the annual generation. This is preferably in the form of an annualised subsidy to be provided on actual generation. These pilot projects can be set by panchayats, independent power producer or renewable energy service companies. A mechanism of bidding can be used to obtain the annualised subsidy level. For example, if it is decided to electrify a village using a dedicated producer gas engine and bio-mass gasifier, bids may be obtained from suppliers for the support required annually for the first three years per kWh of actual generation. The project would be given to the lowest bidder. This would require actual tracking of annual generation. This is feasible using existing technologies of remote monitoring and would add only incrementally to the system cost. A premium on feed-in tariff may not benefit a stand alone plant in a remote area. For such a plant, capital subsidy may be required. Such capital subsidy can be linked however, to the amount of power actually generated, if it is given in the form of Tradable Tax Rebate Certificates (TTRC). The rebate claim becomes payable when electricity is generated and linked to the amount of electricity generated. This will also encourage exploitation of better wind sites earlier. The need to keep the certificates tradable arises from the possibility that small generators may not have adequate taxable income to benefit fully from tax rebate. In areas where there is no electricity grid, there should be minimum clearances/permissions required for setting up a Distributed Generation (DG) system. Supply companies/ entrepreneurs should be free to set-up micro-grids and recover revenues from customers. This is already provided for in the electricity act. Each state should clearly define guidelines to facilitate this.

586

Kakali Mukhopadhyay

Price subsidy for renewables especially for biomass may be justified for several grounds. A renewable energy source may be environmentally benign. It may be locally available making it possible to supply energy earlier than when a centralised system can do. It may provide employment and livelihood to the poor. The environmental subsidy for renewables should be financed by a cess on renewables and fuels causing environmental damage. Price subsidy should be linked to outcomes. For grid connected renewables like biomass Regulatory Commissions (RC’s) should provide feed laws to permit renewables to supply electricity to the grid. RC’s should ensure that renewables are given a tariff equal to the avoided cost of generation.

Specific Policies The following specific policies are to promote rural electrification through BGBPP. Fuelwood Plantation: Cooperatives should be encouraged and facilitated to grow tree plantations in villages. Cooperatives which are open to all members of the community and which are non-discriminatory should be given government land on long-term lease. Women should be encouraged to set-up and manage such plantations so that the time they now spend in gathering fuel can be spent productively in a way that empowers them. They should also be provided finance. If organised and managed properly, such plantations are economic and successful as shown by the experience of National Tree Growers Cooperatives Federation, Parikh et al (1997). Field based NGOs could also be involved in this activity. To encourage large-scale plantations, contract farming should be facilitated. Electricity from Wood Gasification: This can provide electricity based on gasification of wood and can be very useful especially in remote villages. Some institutional arrangements to promote renewable energy are needed: Finally, for the rural electricity supply mission to succeed it is necessary that close cooperation between corporate sector, government and NGOs is needed. The corporate sector can provide the necessary technological and managerial support, NGOs can create the necessary trust in such utilities and Government of India can help provide soft financing through its many rural development programmes. An energy self-sufficient and hence prosperous rural India will be the first step in making us a developed nation. Thus the paper gives us a glimpse of hope for the future development of the rural sector in India through sustainable biomass power. There is need for large-scale demonstration in different parts of India along with policy, institutional and financial support for large-scale spread of environmentally sustainable biomass power systems in India and other developing countries. The State as well as central govt. should be more pro active in that respect to install Biomass gaisifier power plant in those un-electrified villages with the help of MNES. These successful efforts in India briefed in section 3 are really lessons for other countries of this tropical region like Cambodia, Laos, Burma, Vietnam, Thailand, China, Malaysia, Philippine and Indonesia. As we know that the interest in biomass energy is growing in most Asian countries because of scarcity of the conventional fuel. Most of the countries have greater number of rural population than its urban counterpart. The above discussed rural electrification programme in India through biomass gasifier will help to show a guideline for the countries like Cambodia, Laos, Burma and Vietnam especially. With the growing awareness of the importance of renewable energy sources and their potential role in

A Review of the Socio-Economic and Environmental Benefits…

587

decentralised energy generation along with their eco-friendly nature, the rate of growth of biomass energy systems is expected to accelerate in the future. Biomass energy technologies still face a number of barriers in Asian countries, including subsidy for fossil fuels. The pace of commercialisation and deployment of these can be further facilitated through removal of a host of barriers these technologies face at present (Bhattacharya, 2002). By and large, research and development efforts of the Asian developing countries have not yet achieved any remarkable success. The relative successes of India appear to suggest that resource constraints have hindered RandD efforts in the smaller countries. A regional network would be useful in overcoming resource constraints faced by small countries in carrying out research and development efforts. Climate change-related developments are expected to create interest in and awareness of renewable energy sources, including biomass. Thus, renewable energy projects executed in developing countries can generate revenues though certified emissions reduction under the Clean Development Mechanism (CDM) of the Kyoto Protocol; this will make Renewable Energy projects more attractive. To meet the challenges of rising global concern regarding climate change Biomass gasifier power generation should not be an individual attempt rather a regional effort.

ACKNOWLEDGMENT Author gratefully acknowledges Professor Debesh Chakraborty (Dept. of Economics, Jadavpur University, Calcutta, India) for his valuable comments and suggestions on the work.

REFERENCES Akshay Urja (2005) Renewable Energy, 1 (6), Nov-Dec Bhattacharya S.C(2002) Biomass energy in Asia: a review of status, technologies and policies in Asia, Energy for Sustainable Development l Volume VI(3l) September CMIE(1994) Centre for Monitoring Indian Economy Report Desi Power (2004) Advantages of decentralised biomass based power plants, Power to the People Government of India, 2002: The Tenth Five Year Plan (2002-7),Vol II, Sectoral Policies And Programmes, Planning Commission, New Delhi ITCOT (2005) “India: Biomass Energy for Rural India (PMU - BERI), BERI Mukhopadhyay, K.(2004), An p5sessment of a Biomass Gasification based Power Plant in the Sunderbans, Biomass and Bioenergy, 27 (2004) :253 – 264 Ministry of Power, Rural Electrification Programme, Chapter 4, Annual report 2002-3, Government of India, New Delhi. MNES (2004) Ministry of non conventional Energy Sources, Renewable Energy in India, Business opportunities, February. MNES (2002) Renewable energy options in India MNES (2005) Ministry of non conventional Energy Sources, Renewable Energy in India, February. NSS (1992) National Sample Survey, Village level data on rural electrification.

588

Kakali Mukhopadhyay

Planning Commission (2005), Draft report of the expert committee on integrated energy policy, Government of India, New Delhi. Rajvanshi A.k. (1995) Energy Self Sufficient Talukas - A Solution to National Energy Crisis, Economic and Political Weekly (EPW), Vol. XXX, pp. 3315-3319, 1995 Ravindranath, N. H. and Hall, D. O. (1995), Biomass, Energy, and Environment: A Developing Country Perspective from India, Oxford University Press. Ravindranath (2004) Sustainable biomass power for rural India: Case study of biomass gasifier for village electrification, Current Science, Vol. 87, No. 7, 10 October 2004 Rehman, I.H. (2002) Non conventional energy and rural reconstruction, Yojana, January: 3033 SEI (1999)--Stockholm environment Institute-- Renewable Energy for Development, September 1999, Vol. 12, No. 2/3 ---News letter of the Energy Program Somashekhar, H.I. S. Dasappa and N.H. Ravindranath (2000), Rural bioenergy centres based on biomass gasifiers for decentralized power generation: case study of two villages in southern India, Energy for Sustainable Development, IV (3) October: 55-63 Thomas E.C (2002) Renewable Energy in India, Yojana, October,: 47-50 UNDP (2005) ---Equator Initiatives—Plant Power India (Series 5) WEC (1993)—World Energy Council--Energy Information Centre Report

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 589-602

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 20

THE ENERGY BALANCE AND FUEL PROPERTIES OF BIODIESEL* Mustafa Acaroglu and Mahmut Ünaldı Selçuk University, Technical Education Faculty Kampüs Konya, Turkey

ABSTRACT In this study energy balance and fuel properties of biodiesel has been calculated. Accordingly, the cost of 1 liter of oil is calculated 0.32 € after the income from the seed meal is deduced. Finally, the cost of per unit of biodiesel (1 liter) was calculated as 0.55 €, after deduction of the income provided by the sales of glycerin for use in soap and cosmetic industry. The energy equivalent of total output was calculated 147605.50 MJ per hectare. The net energy gain (refined oil) was found as 15105.63 MJ per hectare (The net energy ratio 11.031) according to yield and inputs values. The viscosity values of vegetable oils vary between 27.2 and 53.6 mm2/s whereas those of vegetable oil methyl esters between 3.59 and 4.63 mm2/s. The flash point values of vegetable oil methyl esters are highly lower than those of vegetable oils. The flash point values of vegetable oil methyl esters are highly lower than those of vegetable oils. An increase in density from 860 to 885 kg/m3 for vegetable oil methyl esters or biodiesel increases the viscosity from 3.59 to 4.63 mm2/s and the increases are highly regular. There is high regression between density and viscosity values vegetable oil methyl esters. The relationships between viscosity and flash point for vegetable oil methyl esters are irregular. An increase in density from 860 to 885 kg/m3 for vegetable oil methyl esters increases the flash point from 401 to 453 K and the increases are slightly regular. The LHV values of vegetable oils methyl ester vary between 35.74 and 39.16 MJ/kg.

Keywords: biodiesel, fuel, energy balance *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

590

Mustafa Acaroglu and Mahmut Ünaldı

1. INTRODUCTION Biomass is a renewable resource so far as its production is continued in a sustainable way. Biomass conversion is a process to convert photosynthetic material into a more useful form. Fuels from biomass can take various forms such as solid, gas, and liquid. Liquid biomass (Vegetable oils) from rapeseed, safflower, soybean, palm oil, sunflower, and others can be used for diesel engine. Vegetable oils are used for food, as industrial raw materials, and for the generation of energy. Vegetable oils are a blend of free fatty acids (FFA’s); monoglycerides, diglycerides, and triglycerides; phosphatides, lipoproteins, and glycolipids; waxes; terpenes, gums, and other less important compounds. Today’s diesel engines require a clean-burning, stable fuel that performs well under a variety of operating conditions. Biodiesel is the only alternative fuel that can be used directly in any existing, unmodified diesel engine. Because it has similar properties to petroleum diesel fuel, biodiesel can be blended in any ratio with petroleum diesel fuel [1-3, 13]. Biodiesel is the name for a variety of ester-based oxygenated fuels made from vegetable oils or animal fats. Biodiesel is the name for a variety of ester-based oxygenated fuels made from vegetable oils or animal fats. The concept of using vegetable oil as a fuel dates back to 1895 when Dr. Rudolf Diesel developed the first diesel engine to run on vegetable oil. Biodiesel can be produced by several processes. Micro emulsification, pyrolysis transesterification and super critical method are the four different techniques used to production biodiesel. Vegetable oils or fats can be converted to fatty acids, which in turn are converted to esters. Oils or fats can also be converted to methyl or ethyl esters directly, using an acid or base to accelerate (catalyze) the transesterification reaction. Biodiesel is produced through a process known as transesterification, as shown in the equation below, where R1, R2, and R3 are long hydrocarbon chains, sometimes called fatty acid chains. There are only five chains that are most common in soybean oil and animal fats (others are present in small amounts) (Figure I) [4, 6, 8-11, 16-30].

Figure I. Transesterification flow chart [1, 16].

The Energy Balance and Fuel Properties of Biodiesel

591

Base catalyzation is preferred, because the reaction is quick and thorough. It also occurs at lower temperature and pressure than other processes, resulting in lower capital and operating costs for the biodiesel plant. The most common method of producing biodiesel is to reaction animal fat or vegetable oil with methanol in the presence of sodium hydroxide (a base, known as lye or caustic soda). This reaction is a base-catalyzed transesterification that produces methyl esters and glycerin. Methanol is preferred, because it is less expensive than ethanol. The majority of the alkyl esters produced today is done with the base catalyzed reaction because it is the most economic for several reasons: • • •

Low temperature (65 0C) and pressure (1.37*105 Pascal) processing. High conversion (98%) with minimal side reactions and reaction time. Direct conversion to methyl ester with no intermediate steps.

The use of biodiesel in a conventional diesel engine results in substantial reduction of unburned hydrocarbons, carbon monoxide, and particulate matter. Emissions of nitrogen oxides are either slightly increased depending on the duty cycle and testing methods. The use of biodiesel decreases the solid carbon fraction of particulate matter (since the oxygen in biodiesel enables more complete combustion to CO2), eliminates the sulphate fraction (as there is no sulphur in the fuel), while the soluble, or hydrocarbon, fraction stays the same or is increased. Therefore, biodiesel works well with new technologies such as catalysts (which reduces the soluble fraction of diesel particulate but not the solid carbon fraction), particulate traps, and exhaust gas recirculation (potentially longer engine life due to less carbon). Biodiesel is less toxic than petroleum diesel and biodegrades as fast as dextrose. In addition, biodiesel has a flash point of over 125˚C which makes it safer to store and handle than petroleum diesel fuel [15-28, 30, 31].

2. FUEL PROPERTIES OF BIODIESEL Vegetable oils and their derivatives (especially methyl esters), commonly referred to as “biodiesel,” are prominent candidates as alternative diesel fuels. Biodiesel is generally made of methyl esters of fatty acids produced by the transesterification reaction of triglycerides with methanol in the presence alkali as a catalyst [8]. Among the alcohols that can be used in the transesterification reaction are methanol, ethanol, propanol, butanol and amyl alcohol. Methanol and ethanol are used most frequently, ethanol is a preferred alcohol in the transesterification process compared to methanol because it is derived from agricultural products and is renewable and biologically less objectionable in the environment, however methanol because of its low cost and its physical and chemical advantages. The transesterification reaction can be catalyzed by alkalis [9, 10, 40, 61], acids [15], or enzymes [11, 18, 44, 46, 54]. The transesterification of triglycerides by supercritical methanol, ethanol, propanol and butanol, has proved to be the most promising process [1, 36, 37]. Supercritical methanol has a high potential for both transesterification of triglycerides and methyl esterification of free fatty acids to methyl esters for diesel fuel substitute. In the supercritical methanol transesterification method, the yield of conversion raises 95% for 10 minutes.

592

Mustafa Acaroglu and Mahmut Ünaldı

The catalyst (NaOH) is dissolved into methanol by vigorous stirring in a small reactor. The oil is transferred into the biodiesel reactor and then the catalyst/alcohol mixture is pumped into the oil. The final mixture is stirred vigorously for 1 hour at 338 K in ambient pressure. A successful transesterification reaction produces two liquid phases: ester and crude glycerin. Crude glycerin, the heavier liquid, will collect at the bottom after several hours of settling. Phase separation can be observed within 10 minutes and can be complete within 2 hours of settling. Complete settling can take as long as 20 hours. After settling is complete, water is added at the rate of 5.5 percent by volume of the methyl ester of oil and then stirred for 5 minutes and the glycerin is allowed to settle again. Washing the ester is a two-step process, which is carried out with extreme care. A water wash solution at the rate of 28 percent by volume of oil and 1 gram of tannic acid per liter of water is added to the ester and gently agitated. Air is carefully introduced into the aqueous layer while simultaneously stirring very gently. This process is continued until the ester layer becomes clear. After settling, the aqueous solution is drained and water alone is added at 28 percent by volume of oil for the final washing [1, 35, 36, 39, 41-43, 52-54, 57-61]. All the runs of supercritical methanol transesterification were performed in a 100-mL cylindrical. The sample was loaded from the bolt-hole into the autoclave, and the hole was plugged with a screw bolt after each run. In a typical run, the autoclave was charged with a given amount of vegetable oil (20-30 g) and liquid methanol (30-50 g) with changed molar ratios. The autoclave was supplied with heat from an external heater, and power was adjusted to give Zv approximate heating time of 15 min. The temperature of the reaction vessel was measured with an iron-constantan thermocouple and controlled at ±5 K for 30 min. Transesterification occurred during the heating period. Eight different samples of biodiesel were used for viscosity, flash point and density measurements. A Redwood No. 1 viscosity meter with a measuring cup and a thermostat was used to measure the viscosity of all samples. The viscosity measurements were carried out at 313 K temperature. The temperatures were checked with a digital thermometer within the thermostat and the viscosity meter. At the beginning of each measurement a volume of 50 ml of the sample was filled into the measuring cup. We had to adjust shear rates for the different kinds of samples, because the viscosities are quite different and the viscosity meter has to be used within the correct measuring range. Flash point measurements were carried out using a Koehler mark apparatus. Table 1. Fatty acid compositions of vegetable oil samples [1, 10] Sample Cottonseed Rapeseed Safflowerseed Safflowerseed Palm Soybean Hazelnut kernel

16:0 28.7 3.5 7.3 6.4 42.6 13.9 4.9

16:1 0 0 0 0.1 0.3 0.3 0.2

18:0 0.9 0.9 1.9 2.9 4.4 2.1 2.6

18:1 13.0 64.1 13.6 17.7 40.5 23.2 83.6

18:2 57.4 22.3 77.2 72.9 10.1 56.2 8.5

18:3 0 8.2 0 0 0.2 4.3 0.2

Others 0 0 0 0 1.1 0 0

The Energy Balance and Fuel Properties of Biodiesel Table 2. Viscosity, density and flash point measurements of ten vegetable oils [1, 17] Oil Source Corn Cottonseed Crambe Linseed Peanut Rapeseed Safflower Sesame Soybean Sunflower

Viscosity mm2/s (at 311 K) 34.9 33.5 53.6 27.2 39.6 37.0 31.3 35.5 32.6 33.9

Density kg/m3 909.5 914.8 904.4 923.6 902.6 911.5 914.4 913.3 913.8 916.1

Flash Point K 550 509 447 514 544 519 533 533 527 447

Table 3. Some fuel properties of six methyl ester biodiesels Source Sunflower Soybean Palm Penaut Babassu Tallow

Viscosity cSt at 313.2 K 4.6 4.1 5.7 4.9 3.6 4.1

Density g/mL at 288.7 K 0.880 0.884 0.880 0.876 -0.877

Cetane Number 49 46 62 54 63 58

Reference [49] [53] [49] [57] [57] [4]

Table 4. Viscosity, density and flash point measurements of eight oil methyl esters [3] Methyl ester Cottonseed oil Hazelnut kernel oil Mustard oil Palm oil Rapeseed oil Safflower oil

Viscosity mm2/s (at 313 K) 3.69 3.59 4.10 3.70 4.63 4.03

Density kg/m3 (at 288 K) 880 860 881 870 885 880

Flash point K 437 401 446 443 428 453

Soybean oil Sunflower oil

4.08 4.22

885 880

447 443

Table 5. LHV (Low Heating Value) properties of methyl ester biodiesel* Methyl ester Cottonseed oil Hazelnut kernel oil Mustard oil Palm oil Rapeseed oil Safflower oil Sunflower oil

* Analysis time July 2005 (Acaroğlu)

LHV (MJ/kg) 39.16 38.70 39.04 35.74 38.24 36.98 35.92

Ash (wt %)

Moisture (wt %)

0.013 -0.003 0.035 0.004 0.002

0.75 -6.30 1.30 0.61 0.90

593

594

Mustafa Acaroglu and Mahmut Ünaldı

3. BIODIESEL EMISSIONS COMPARED TO CONVENTIONAL DIESEL Biodiesel is the first and only alternative fuel to have a complete evaluation of emission results and potential health effects submitted to the U.S. Environmental Protection Agency (EPA) under the Clean Air Act Section 211 (b). These programs include the most stringent emissions testing protocols ever required by EPA for certification of fuels or fuel additives in the US. The data gathered through these tests complete the most thorough inventory of the environmental and human health effects attributes that current technology will allow. The overall ozone (smog) forming potential of biodiesel is less than diesel fuel. The ozone forming potential of the speciated hydrocarbon emissions was nearly 50 percent less than that measured for diesel fuel. Sulfur emissions are essentially eliminated with pure biodiesel. The exhaust emissions of sulfur oxides and sulfates (major components of acid rain) from biodiesel were essentially eliminated compared to sulfur oxides and sulfates from diesel. Criteria pollutants are reduced with biodiesel use. The use of biodiesel in an unmodified Cummins N14 diesel engine resulted in substantial reductions of unburned hydrocarbons, carbon monoxide, and particulate matter. Emissions of nitrogen oxides were slightly increased. Carbon Monoxide - The exhaust emissions of carbon monoxide (a poisonous gas) from biodiesel were 47 percent lower than carbon monoxide emissions from diesel. CO2 emission index is defined as the CO2 emission (%) divided by corresponding fuel consumption rate (in unit of g/h), the CO emission index is defined as CO emission (ppm) divided by the corresponding fuel consumption rate (in unit of g/h) and the NOx emission index is defined as NOx emission (ppm) divided by the corresponding fuel consumption rate. The catalytic converter reduced CO and HC emissions [1]. Benzene emissions increased with the amount of RME (rapeseed oil methylester) either with catalytic converter or without catalytic converter. This is remarkable because benzene is absent in RME. This indicates that the main source of benzene emissions may be a synthesis that occurs during combustion, rather than unburned fuel. However, the catalytic converter reduced the emissions by one third. Particulate Matter - Breathing particulate has been shown to be a human health hazard. The exhaust emissions of particulate matter from biodiesel were 47 percent lower than overall particulate matter emissions from diesel, Hydrocarbons - The exhaust emissions of total hydrocarbons (a contributing factor in the localized formation of smog and ozone) were 67 percent lower for biodiesel than diesel fuel. Nitrogen Oxides(NOx) emissions from biodiesel increase or decrease depending on the engine family and testing procedures. NOx emissions (a contributing factor in the localized formation of smog and ozone) from pure (100%) biodiesel increased in this test by 10 percent. However, biodiesel's lack of sulfur allows the use of NOx control technologies that cannot be used with conventional diesel. So, biodiesel NOx emissions can be effectively managed and efficiently eliminated as a concern of the fuel's use [14, 26, 31, 33, 38, 49, 50, 59]. Biodiesel reduces the health risks associated with petroleum diesel. Biodiesel emissions showed decreased levels of PAH and nitrited PAH compounds which have been identified as potential cancer causing compounds. In the recent testing, PAH compounds were reduced by

The Energy Balance and Fuel Properties of Biodiesel

595

75 to 85 percent, with the exception of benzo(a)anthracene, which was reduced by roughly 50 percent. Targeted nPAH compounds were also reduced dramatically with biodiesel fuel, with 2-nitrofluorene and 1 -nitropyrene reduced by 90 percent, and the rest of the nPAH compounds reduced to only trace levels.

Figure 2. Average emission impacts of biodiesel fuels in CI engines [59].

4. THE ENERGY BALANCE OF SAFFLOWER OIL Energy analysis, along with economic and environmental analyses, is an important tool to define the behavior of agricultural system. Energy analysis started as a relevant subject in agricultural production in the 1970’s as a results of the dramatic increase of oil derivative prices. In an energy analysis of production systems it is necessary to consider the following steps ([1, 32, 47]: • • • • •

Set a limit in the process or system to be analyzed in such a way that all inputs and outputs, which pass that limit in a certain time interval, are evaluated. Assign energy requirements to all inputs Identify and quantify all outputs, establishing criteria for energy embodied in the main products and that corresponding to by-products. Relate output energy to total sequestered energy to obtain the energy ratio and the energy productivity. Apply energy analysis results

Safflower, Carthamus tinctorius, is among the oldest crops known to man. The Safflower (Carthamus tinctorius L.) is probably native to an area bounded by the Eastern Mediterranean

596

Mustafa Acaroglu and Mahmut Ünaldı

and Persian Gulf. Believed to have originated in southern Asia and is known to have been cultivated in China, India, Persia and Egypt almost from prehistoric times. During middle Ages it was cultivated in Italy, France, and Spain, and soon after discovery of America, the Spanish took it to Mexico and then to Venezuela and Colombia. It was introduced into United States in 1925 from the Mediterranean region and is now grown in all parts west of 100th meridian. Safflower is commercially cultivated to a large extent in India, Mexico and the USA (Table 6, Table 7) [5, 12, 34, 48]. Table 6. Safflower in World (Country, Sown area, production and yield) [29] Country Indian USA Ethiopia Mexico Australia Turkey World

Sown Area (ha) 404100 79320 72000 52758 35000 30 756055

Production (ton) 226000 135160 38000 52855 26000 15 577555

Yield (kg/ha) 559 1704 528 1002 743 500 764

Table 7. Safflower in Turkey (year, sown area, production and yield) (1995-2002), [29] Years 1995 1996 1997 1998 1999 2000 2001 2002 Konya (2003-2004)

Sown area (ha) 134 81 74 75 50 30 35 30 35

Production (ton) 125 74 65 72 50 18 25 15 67.375

Yield (kg/ha) 930 910 880 960 1000 600 714 500 1925

Safflower is cultivated for the edible oil obtained from the seed. It contains a higher percentage of essential unsaturated fatty acids and a lower percentage of saturated fatty acids than other edible vegetable seed oils. Safflower oil lowers blood cholesterol levels and is used to treat heart diseases. The flowers have been the source of yellow and red dyes, largely replaced by synthetics, but still used in rouge. Seeds used for tumors, especially inflammatory tumors of the liver [19]. Flowers considered diaphoretic, emmenagogue, laxative, sedative, stimulant, in large doses laxative; used as a substitute or adulterant for saffron in treating measles, scarlatina, and other exanthematous diseases. Charred safflower oil used for rheumatism and sores; seeds, diuretic and tonic [7]. In China, prescribed as uterine astringent in dysmenorrhea. In Iran, the oil is used as a salve for sprains and rheumatism. Safflower is propagated by seed. The seed is sown 2-3 cm deep, with planting distances 10 cm in the row and 30-60 cm between the rows. The desired crop density is 40-50 plants per m2, and up to 70 plants per m2 on light soil. The growing period is 200-250 days or 110140 days, respectively [5, 48, 51].

The Energy Balance and Fuel Properties of Biodiesel

597

No recommendations can be made for herbicide use, as this aspect of weed control is still being tested, but herbicides appropriate for sunflower crops are believed to be appropriate for safflower too [12]. The yields of seed can reach 1.1 – 1.7 t/ha/year [29]. With irrigation and good fertilization, yields 2.8-4.5 t/ha/year can be achieved, but the world average yield is 0.5 t/ha/year. The cultivars have 30-48 % oil in the fruit. The oil contains 73-79 % linoleic acid, depending on the cultivar [29, 45, 55, 56, 60]. In Turkey, today there are only three Safflower kinds. This study was used Dinçer variety (Table 8). Middle Anatolia region was selected and 35 ha Safflower was sowed in Konya. Sowing time was the first week of April (Table 9). Table 8. The properties of Safflower variety crops [5] SORT (VARIETY) YENICE DINÇER 5-154

THORNY NOT THORNY NOT THORNY THORNY

FLOWER COLOR

PLANT HEIGHT (CM)

COLOR

OIL CONTENT (%)

WEIGHT (GR/1000 SEED)

RED ORANGE YELLOW

100-120 90-110 60-80

WHITE WHITE WHITE

24-25 25-28 35-40

38-40 45-49 46-50

SEED

Table 9. The properties on sowing operations Operation Row distance Row upper Sowing deep Sowing (seeding) norm Crop density Fertilizer (N) Fertilizer (P2O5) Fertilizer (K)

Properties 15 cm 10 cm 2.5-4.0 cm 20 kg/ha 66 – 67 plant per m2 120 kg/ha per year 50 kg/ha per year Not used

Harvesting is making with Combine harvester.

5. RESULTS AND DISCUSSION In this study energy balance in Safflower production has been determined and calculated (Table 10 and Table 11). In making the energy balance, input and output values used in calculation are measured in field conditions and calculated using literature. Table 5 list the cultivation operations carried out, with the corresponding work hours and technical means and used with their respective consumption.

598

Mustafa Acaroglu and Mahmut Ünaldı Table 10. Technical characteristic of machines and energy inputs Safflower (MJ/ha)

Agricultural Operation

(h/ha)

Tractors Energy

Equipment Energy

Labor Energy

TOTAL

14.35

Fuellubricant Energy 604.53

Primary Tillage Plough

2.14

73.36

4.00

696.24

Secondary tillage (Disk harrow) Seeding

0.84

28.91

10.16

238.33

1.58

278.97

0.68

23.34

49.94

192.47

1.27

267.03

Fertilizer

0.22

7.55

1.96

62.18

0.41

72.11

Harvesting (Combine)

0.94

121.20

390.74

1.76

513.70

197.62

1488.25

9.01

1828.04

Total

133.16

Table 11. Fertilizer and Seed Energy Inputs Fertilizer Energy N P

kg/ha 120 40

Seed Energy

20

MJ/kg 49.1 17.78 TOTAL 14

MJ/ha 5892 711.2 6603.2 280

In harvesting, the Safflower seed yield was obtained as 1925 kg per hectare on an average and Safflower stalk was obtained as 6650 kg DM per hectare in the moisture of 15 % (18.08 MJ/kg). The average seed oil content was found 36 % (693 kg per hectare oil yield), (39.50 MJ/kg). (Table 12, Table 13). Table 12. Total Energy Inputs of Safflower Energy Input

MJ/ha

%

Fuel-oil Tractor Energy Machinery Energy Labour Energy Fertilizer Energy Seed Energy Industrial process (extraction) Industrial process (refining) TOTAL

1488.25 133.16 197.62 9.01 6603.20 280.00 2406.25 1150.38 12267.87

12.13 1.09 1.61 0.07 53.83 2.28 19,61 9.38 100.00

Table 13. Total Energy Outputs of Safflower Energy Output

MJ/kg

MJ/ha

Safflower oil (Refined)

39.5

27373.5

Safflower stalk

18.08

120232

Safflower product residues

18.09

28871.6

Total

147605.5

The Energy Balance and Fuel Properties of Biodiesel

599

Cost inputs in safflower oil biodiesel production has been calculated (table 14). In making the cost analysis, input and output values used in calculation are measured in process conditions and calculated using literature. Table 14. Input and cost safflower biodiesel Raw material

Cost per liter (€)

Conversion %

Quantity / liter biodiesel

Cost / liter biodiesel (€)

Oil

0.32

90

1.11

0.36

Methanol

0.250

-

0.20

0.05

NaOH

1.200

-

0.035

0.04

Various

-

-

-

0.01

Energy

0.100

-

-

0.10

Biodiesel cost

-

-

-

0.55

The energy equivalent of total output was calculated 147605.50 MJ per hectare. The net energy gain (refined oil) was found as 15105.63 MJ/ha (The net energy ratio 11.031) according to yield and inputs values. Net energy gain of Safflower seed is higher then other oil seeds in the Turkey. In addition to these, Safflower crops will have a great importance in Turkey as oil resource and biomass energy source as it solves environmental problems, its inputs are low, and it can be grown in arid zones.

REFERENCES [1] [2]

[3] [4] [5] [6] [7] [8] [9]

Acaroğlu, M. 2003. Renewable Energy Sources, Atlas Yayın and Dağıtım, Istanbul, (Turkish Book) Acaroğlu, M., Gezer, I., 2004, Renewable Energy Sources and Future of Biomass Energy in Turkey, pp. 413-416, 2nd World Conference on Biomass for Energy, Industry and Climate Protection, 10-14 May 2004, Rome, Italy Acaroğlu, M., Demirbaş, A., 2005, Relationships between Viscosity and Density Measurements of Biodiesel Fuels, Energy Sources (in press). Ali Y, Hanna MA, Cuppett SL. 1995. Fuel properties of tallow and soybean oil esters. JAOCS 1995; 72:1557-1564. Babaoğlu, M., 2004 Dünya’da ve Türkiye’de Aspir Bitkisinin Tarihçesi, Kullanım Alanları ve Önemi (Turkish), Trakya Tarımsal Araştırma Enstitüsü Edirne Bala, B. K. 2005. Studies on biodiesels from transformation of vegetable oils for diesel engines, Energy Edu. Sci. Technol. 15:1-43. C.S.I.R. (Council of Scientific and Industrial Research), 1948-1976, The wealth of India. 11 vols, New Delhi Clark,S. J., L. Wagner, M. D. Schrock, and P. G. Pinnaar. 1984. Methyl and ethyl esters as renewable fuels for diesel engines. JAOCS 61:1632–1638. Demirbas A. Biodiesel from vegetable oils via transesterification in supercritical methanol, Energy Convers. Mgmt. 2002; 43:2349–56.

600

Mustafa Acaroglu and Mahmut Ünaldı

[10] Demirbas A. Biodiesel fuels from vegetable oils via catalytic and non-catalytic supercritical alcohol transesterifications and other methods: a survey. Energy Convers Management 2003;44:2093-2109. [11] Du W, Xu Y, Liu D, Zeng J. Comparative study on lipase-catalyzed transformation of soybean oil for biodiesel production with different acyl acceptors, J. Molecular Catal. B: Enzymatic. 2004;30:125–129 [12] Duke, J.A., 1983. Handbook of Energy Crops (unpublished). [13] El Bassam, N., 1998. Energy Plant Species, Their use and impact on environment, JamesandJames (Science Publishers), 320 p., ISBN 1 873936 75 3 [14] EPA, 2002, A Comprehensive Analysis of Biodiesel Impacts on Exhaust Emissions, Draft Technical Report, EPA420-P-02-001 October 2002 [15] Furuta S, Matsuhashi H, Arata K. Biodiesel fuel production with solid superacid catalysis in fixed bed reactor under atmospheric pressure. Catal. Commun. 2004; 5:721–723. [16] Gerpen, V., 2005, Biodiesel processing and production, Fuel Processing Technology 86 (2005) 1097– 1107 [17] Goering, E., W. Schwab, J. Daugherty, H. Pryde, and J. Heakin, 1982, Fuel properties of eleven vegetable oils. Trans ASAE 25: 1472–1483. [18] Hama S, Yamaji H, Kaieda M, Oda M, Kondo A, Fukuda H. Effect of fatty acid membrane composition on whole-cell biocatalysts for biodiesel-fuel production. Biochem. Eng. J. 2004;21:155–60 [19] Hartwell, J.L. 1967-1971. Plants used against cancer. A survey. Lloydia 30-34. [20] http://journeytoforever.org/biodiesel.html [21] http://ww2.green-trust.org:8383/biodiesel2.htm [22] http://www.biodiesel.org [23] http://www.biodiesel.org/pdf_files/fuelfactsheets/emissions.pdf [24] http://www.canentec.com/whatisbiodiesel.html [25] http://www.distributiondrive.com/Article16.html [26] http://www.epa.gov/otaq/models/biodsl.htm [27] http://www.fact-index.com/b/bi/biodiesel.html#History [28] http://www.fact-index.com/m/ma/main_page.html [29] http://www.fao.org [30] http://www.liquid-biofuels.com/pub.htm [31] http://www.soypower.net/BiodieselPDF/BiodieselEmissions.pdf [32] Intosh, C. S., Withers, R. V., Smith, S. M., 1982. The Economics of On-Farm Production and Use of Vegetable Oils for Fuel, Paper No.177, Proceedings of the International Conference on Plant and Vegetable Oils as Fuels, Holiday Inn Fargo North Dakota. [33] Kaltschmitt, M., Reinhardt, G.A., 1997. Nachwachsende Energieträger - Grundlagen, Verfahren, ökologische Bilanzierung. Vieweg-Verlag. Braunschweig. [34] Keys, J.D. 1976, Chinese herbs, their botany, chemistry, and pharmacodynamics, Chas. E. Tuttle Co., Tokyo [35] Krawczyk, T., “Biodiesel – Alternative fuel makes inroads but hurdles remain,” INFORM vol. 7, no. 8(1996), pp.801-815, American Oil Chemist Society [36] Kusdiana D, Saka S. Effects of water on biodiesel fuel production by supercritical methanol treatment. Biores.Technol. 2004; 91: 289-295.

The Energy Balance and Fuel Properties of Biodiesel

601

[37] Kusdiana D, Saka S. Kinetics of transesterification in rapeseed oil to biodiesel fuels as treated in supercritical methanol. Fuel 2001; 80: 693–698. [38] Lin, C.Y., Lin, H.A, 2005, Diesel engine performance and emission characteristics of biodiesel produced by the peroxidation process, Fuel xx (2005) 1-8. [39] Lyell, K., 2003. Design of a Methanol Extraction Process for Bio-Diesel Production, Fall 2003, Austin Bio-Fuels, LLC, Texas. [40] Ma F, Hanna MA. Biodiesel production: a review. Biores Technol 1999, 70:1-15. [41] Ma, F., “Biodiesel fuel: The transesterification of beef tallow.” PhD dissertation, Biological Systems Engineering, University of Nebraska-Lincoln (1998) [42] Ma, F., Clements, L.D., Hanna, M.A, 1998. Biodiesel fuel form animal fat. Ancillary studies on transesterification of beef tallow, Ind. Eng. Chem. Res. vol.37 (1998b), pp. 3768-3771. [43] Ma, F., Hanna, M.A., Biodiesel production: a review, Biosource Technology vol.7 (1999), pp.1-15. [44] Noureddini H, Gao X, Philkana RS. Immobilized Pseudomonas cepacia lipase for biodiesel fuel production from soybean oil, Biores. Technol. 2005; 96:769–777. [45] O’Brien, R., 1998, Fats and Oils, Formulating and Processing for Applications, 667 p., Technomic Publishing AG, Basel, Switzerland [46] Oda M, Kaieda M, Hama S, Yamaji H, Kondo A, Izumoto E, Fukuda H. Facilitatory effect of immobilized lipase-producing Rhizopus oryzae cells on acyl migration in biodiesel-fuel production. Biochem. Eng. J. 2004; 23:45–51. [47] Ortiz-Canavate, J., Hernanz, J.L., 1999. Energy for Biological Systems, Energy Analysis and Saving, CIGR Handbook of Agricultural Engineering, Energy and Biomass Engineering, pp.13-42, Published by ASAE, USA. [48] Ozturk, O, 2004. Aspir Tarımının Önemi ve Orta Anadolu Şartlarında Yetiştirme İmkanları, Konya Ticaret Borsası, April 2004, Year: 7, N. 17, pp. 54-60 (Turkish), Konya, Turkey. [49] Pischinger GM, Falcon AM, Siekmann RW, Fernandes FR. Methylesters of plant oils as diesels fuels, either straight or in blends. Vegetable Oil Fuels, ASAE Publication 482, Amer. Soc. Agric. Engrs. St. Joseph, MI, USA, 1982. [50] Pryor, R. W., M. A. Hanna, J. L. Schinstock, and L. L. Bashford. 1982. Soybean oil fuel in a small diesel engine. Trans ASAE 26:333-338. [51] Raghu, J.S. and Sharma, S.R. 1978, Response to irrigation and fertility levels of safflower. Indian J. Agron. 23(2):93-97. [52] Riva, G., Sissot, F., 1999. Vegetable Oils and Their Esters (biodiesel), CIGR Handbook of Agricultural Engineering, Energy and Biomass Engineering, pp. 164-201, ASAE, USA. [53] Schwab AW, Bagby MO, Freedman B. Preparation and properties of diesel fuels from vegetable oils. Fuel 1987;66:1372–1378. [54] Shieh C-J, Liao H-F, Lee C-C. Optimization of lipase-catalyzed biodiesel by response surface methodology. Bioresource Technology 2003;88:103–106. [55] Smith, J.R., 1985. “Safflower: Due for a Rebound”, J. Am. Oil Chem. Soc., 62(9):12861291. [56] Sonntag, N.O.V., 1979. “Composition and Characteristics of Individual Fats and Oils” in Bailey’s Industrial Oil and Fat Products, Vol. 1, 4th Edition, D. Swern, ed. New York, NY: A Wiley- Interscience Publication, pp. 398-403.

602

Mustafa Acaroglu and Mahmut Ünaldı

[57] Srivastava A, Prasad R. Triglycerides-based diesel fuels. Renew. Sustain. Energy Rev. 2000;4:111–133 [58] Tickell, J., 2000.From the Fryer to the Fuel Tank the Complete Guide to Using Vegetable Oil as an Alternative Fuel. Tickell Energy Consulting, USA [59] Tyson, K.S., 2004, Biodiesel Handling and Use Guidelines, DOE/GO 102004-1999 Revised November 2004; U.S. Department of Energy. [60] Weiss, T.J., 1983. “Commercial Oil Sources”, in Food Oil and Their Uses, Second Edition [61] Zhang Y, Dub MA, McLean DD, Kates M. Biodiesel production from waste cooking oil: 2. Economic assessment and sensitivity analysis. Biores. Technol. 2003; 90:229– 240.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 603-623

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 21

AN EXPERIMENTAL STUDY ON PERFORMANCE AND EXHAUST EMISSIONS OF A DIESEL ENGINE FUELLED WITH VARIOUS BIODIESELS* Nazim Usta† Pamukkale University, Mechanical Engineering Department Camlık 20017 Denizli, Turkey

ABSTRACT Instability and increases in prices of petroleum-based fuels, gradual depletion of world petroleum reserves and increases in environmental pollution caused by exhaust emissions speed up research on renewable alternative fuels. Vegetable oils have been considered as renewable alternative fuels in compression ignition engines for a long time. However, they have not been widely used as fuels in the engines due to some technical and economical drawbacks. Some properties of vegetable oils such as high viscosity, lower volatility and lower heat content result in technical problems in direct using of vegetable oils in short and long term applications. From economical point of view, the main problem is that vegetable oils have been more expensive than petroleum Diesel fuel. There are various ongoing studies on solving these problems to be able to use vegetable oils in Diesel engines. Different methods such as preheating oils, blending or dilution with other fuels, thermal cracking/pyrolysis and transesterification have been developed. Among these techniques, transesterification appears to be the most promising one. It is a chemical process converting vegetable oils to alcohol ester of oil named as biodiesel. In general, biodiesel-Diesel fuel No.2 blend can be used as a fuel in Diesel engines without modification. Specifications of biodiesel mainly depend on oil, *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Nazim Usta; Pamukkale University; Mechanical Engineering Department; Camlık 20017 Denizli; Turkey; Tel: +90 258 2125532; Fax: +90 258 2125538; E-mail: [email protected] ; [email protected]

604

Nazim Usta transesterification process, type and amount of alcohol, type and amount of catalysis, reaction time and temperature. Biodiesel can be produced from different kinds of vegetable oils. Since prices of edible vegetable oils are higher than that of Diesel fuel No. 2, waste vegetable oils and non-edible crude vegetable oils are mostly preferred as potential low priced biodiesel sources. It is also possible to use soapstock, a by-product of edible oil production, for cheap biodiesel production. In this study, various biodiesels were produced from raw vegetable oils (rapeseed oil, soybean oil, cotton seed oil, palm oil and tobacco seed oil), waste sunflower vegetable oils and hazelnut oil soap stock-waste sunflower vegetable oil, and their specifications were compared with each other. The biodiesel (20% in volume) - Diesel fuel No.2 (80% in volume) blends were tested in a four cycle, four cylinder turbocharged indirect injection Diesel engine. The effects of biodiesel addition to Diesel fuel No.2 on the performance and emissions of the engine were investigated at full load. Experimental results showed that the biodiesels can be partially substituted for Diesel fuel No.2 at most operating conditions in terms of performance parameters and emissions without any engine modification and preheating of the blends.

Keywords: Biodiesel, diesel engine, performance, emission.

NOMENCLATURE C D P R S SOW T W

cotton seed oil methyl ester Diesel fuel No.2 palm oil methyl ester rape seed oil methyl ester soybean oil methyl ester mixture (waste sunflower oil in 50%–hazelnut kernel soap stock in 50%) methyl ester tobacco seed oil methyl ester waste sunflower oil methyl ester

1. INTRODUCTION Increases in prices of petroleum-based fuels, the environmental pollution due to exhaust emissions from these fuels, the gradual depletion of world petroleum reserves and economical and political instabilities in countries exporting petroleum have encouraged studies to search for alternative renewable fuels. Vegetable oils have been considered as alternative renewable fuels for compression ignition (CI) engines which are mainly used in transport sector. Vegetable oils are non-toxic, biodegradable, and have low emission profiles (Williamson and Badr, 1998; Ma and Hanna, 1999; Srivastava and Prasad, 2000; Kalam et al. 2003; Kalligeros et al. 2003). Use of vegetable oils in CI engines is known since the invention of CI engine by Rudolph Diesel. He used peanut oil in the engine. However, vegetable oils have not been widely and effectively

An Experimental Study on Performance and Exhaust Emissions…

605

used in CI engines due to some drawbacks such as higher viscosity, lower volatility and lower heat content (Ergeneman et al., 1997a; Altin et al., 2001; Nwafor, 2003). Especially the higher viscosity which causes poor atomization and results in incomplete combustion is an important disadvantage of vegetable oils. It may also lead to formation of injector deposits, ring sticking, development of gumming, as well as incompatibility with lubricating oils in long term operations (Williamson and Badr, 1998; Karaosmanoğlu et al. 2000). In addition, prices of vegetable oils have been higher than that of diesel fuel No.2 and this is an important economical drawback related to use of vegetable oil in CI engines. Researchers have been tried to develop different techniques such as preheating oils, blending or dilution with other fuels, thermal cracking/pyrolysis and transesterification to use vegetable oils in CI engines effectively and efficiently (Williamson and Badr, 1998; Ma and Hanna, 1999; Karaosmanoglu, 1999; Agarwal and Das, 2001; Demirbaş, 2003). Among these techniques, transesterification which is a chemical process of converting vegetable oil into methyl or ethyl ester of vegetable oil has been widely preferred and used successfully (Kalligeros et al. 2003; Agarwal and Das, 2001; Demirbaş, 2002; Kumar et al., 2003). In general, the methyl or ethyl ester of vegetables is called as biodiesel. The transesterification process has been modified depending on type and condition of the vegetable oil. The detailed reviews about the process are available in the literature (Ma and Hanna, 1999; Demirbaş, 2002; Agarwal and Das, 2001; Karaosmanoglu, 1999; Demirbaş, 2003; Ozaktas et al. 1997). There are different kinds of sources for biodiesel production. Although virgin edible vegetable oils are mainly used in food sectors, rapeseed oil, soybean and palm oil have been mainly used for biodiesel production (Crabbe et al., 2001; Kumar et al. 2003). There are some studies related to non-edible vegetable oils such as tobacco seed oil (Usta, 2005). To the author’s best knowledge, tobacco seed oil was used first time by the author for biodiesel production and usage as a fuel in a diesel engine. There are also some other studies on nonedible vegetable oils (Srivastava and Prasad, 2000; Agarwal and Das, 2001; Pramanik, 2003). The extensive usage of biodiesel may results in the finding of new sources. Waste cooking oil is an important cheap source for biodiesel production (Gonzalez Gomez et al. 2000; Dorado et al., 2003; Ozaktas, 2000; Alcantara et al., 2000; Al-Widyan and Al-Shyoukh, 2002; Tomasevic and Siler-Marinkovic, 2003). In addition, soapstocks have been considered to use in biodiesel production (Haas et al., 2000; Haas et al., 2001; Haas and Foglia, 2002; Haas et al., 2003; Graboski et al., 2003; Usta et al., 2005). Soapstocks are also cheap sources for biodiesel production. Biodiesel can be used in different proportions in CI engines without any modification. In the US, the most commonly used blend is 20 % (in volume) biodiesel in diesel fuel No.2, which is called as a B20 blend. However, in Europe the most commonly used blend is 5 % (in volume) in the diesel fuel (Piazza and Fogila, 2001). There is an important compositional difference between biodiesels and the diesel fuel. Biodiesels contain approximately 10–12% oxygen in weight basis. This leads to reduction in the energy content of the fuel resulting in lower engine torque and power (Altin et al., 2001; Nwafor et al., 2000; Bari et al. 2002; Antolín et al., 2002). However, the oxygen in the fuel helps to reduce exhaust emissions such as smoke, CO and HC mainly due to the effect of complete combustion (Kalam et al., 2003; Kalligeros et al., 2003; Gonzalez Gomez et al. 2000; Ozaktas et al., 1997; Ergeneman et al., 1997b; Marshall et al., 1995; Chang and Van Gerpen, 1997; Monyem and Van Gerpen, 2001; Graboski and McCormick, 1998; Curran et al., 2001; Kitamura et al., 2001a; Kitamura et al., 2001b).

606

Nazim Usta

Since vegetable oils includes very little sulphur compared to the diesel fuel No.2, some reduction in SO2 emission is obtained depending on the proportion of biodiesel in the fuel (Dorado et al., 2003). The main disadvantage of biodiesel on emissions is related to NOx. Although NOx emissions mainly depend on the engine fuelling system, engine type and engine loading, in general biodiesel usage increases NOx emissions due to oxygen content of the fuel and higher temperatures of combustion chamber (Gonzalez Gomez et al., 2000). In this study, biodiesels produced from seven different sources (rapeseed oil, soybean oil, cotton seed oil, palm oil, tobacco seed oil, waste vegetable oils and hazelnut kernel oil soap stock-waste vegetable oil mixture) were blended with diesel fuel No.2 in 20 % (in volume). The blends were tested in a four cycle, four cylinder turbocharged indirect injection diesel engine. The effects of biodiesel addition in 20% to Diesel No. 2 on the performance and emissions of the engine were investigated at full load. All tests were performed without any modification on the engine.

2. MATERIALS AND METHODS 2.1. Raw Materials In this section, the seven different sources, which are important sources for biodiesel productions worldwide and used in this study, are described briefly. Rapeseed is one of the most important sources of vegetable oil in the world. Natural rapeseed oil contains erucic acid which is a kind of toxic material to humans in large doses. An edible form of rapeseed (canola) which has low erucic acid content was developed in Canada and then the rapeseed production has increased. In Europe, low-erucic rapeseed oil is the major source for biodiesel production (Piazza and Fogila, 2001). Soybean oil is the world's most widely used edible oil and it is a very healthy food ingredient. It is not only used in food products but is also used to produce some non-food products like biodiesel, inks, plasticizers, crayons, paints and soy candles. It is the primary oil of biodiesel production in the United States (Piazza and Fogila, 2001). Cottonseed oil is edible oil and it is mainly used in food sector. However, it is less favored than sunflower, corn and soybean oils in foods. Meanwhile the cotton seed oil which is not used in the food sector may be utilized oil in biodiesel production, especially in certain countries where cotton is widely produced (Karaosmanoğlu et al., 1999). Palm oil is an edible vegetable oil which is extracted from the fruit of oil palm tree. In addition it can be used in non-food products. It is also likely to be used in biodiesel production. The high oil yield per hectare area has made it the main source of vegetable oil for many tropical countries. Main palm oil producers and exporters in the world are Malaysia and Indonesia. Tobacco plant is a beautiful plant with large oval leaves, pink flowers and green capsules containing numerous very small seeds (U.S. Department of Agriculture, 2004). The seed has a strong shell, therefore it can resistant to high humidity and can be stored in dry conditions at ordinary temperatures. The oil content of seeds changes from 36% to 41% by weight (Giannelos et al., 2002). The rest of the seed consists of protein, carbohydrate, inorganic material and crude fiber. The plant is grown in 119 countries in the world mainly for leaves

An Experimental Study on Performance and Exhaust Emissions…

607

which are commercial products and used in the production of cigarettes in the tobacco processing industries. Since the oil which can be extracted from tobacco seeds is a non-edible oil and not used in other applications, small amount of tobacco seeds is collected from fields for next year production. But most of them are left unused in fields. Tobacco harvesting area and leaves production of different countries are available in (U.S. Department of Agriculture, 2004). However, there is no statistical information on potential tobacco seeds in the literature to the best knowledge of the author. Meanwhile, the seed and oil potential may be estimated using the harvesting area information. Tobacco seed oil was used first time by the author for the biodiesel production. Since the tobacco seed oil is not available in markets it was required to extract the oil from the seeds. The detailed information can be found in a paper of the author (Usta, 2005a). Waste cooking oil is a cheap source for biodiesel production. However the operational cost of biodiesel production from this oil is higher than that from virgin oil due to the requirement of filtration and water removal. In addition, the properties of waste cooking oils vary depending on the raw oil and the cooking application. One of the important properties of waste cooking oil is the free fatty acid content. If the free fatty acid content of the oil is higher than %2, it is required to use acid catalyser or acid/base catalysers in biodiesel production. Soapstock is a by-product of edible vegetable oil production and also a cheap source for biodiesel production. In general it is used in soap industry. There are limited numbers of studies on soapstock as a biodiesel source in the literature as mentioned above. Since soapstock contains large amounts of free fatty acids (45–50%), it cannot be effectively converted to biodiesel using only an alkaline catalyst. It is required to reduce the free fatty acids of the feedstock using an acid catalysed pre-treatment to esterify the free acids before transesterifying the triglycerides with an alkaline catalyst to complete the reaction. The fatty acid composition of vegetable oils is the one of the important parameter which affects specification of their methyl esters. Therefore the fatty acid compositions of the oils are given in Table 1. Table 1. Fatty Acid Compositions of the Vegetable Oils (% by weight) Fatty acid

Tobacco seed *

Soybean **

Rapeseed**

Palm**

3.5

Cotton seed** 28.7

Palmitic (16.0) Palmitoleic (16:1) Stearic (18:0) Oleic (18:1) Linoleic (18:2) Linolenic (18:3) Others

10.96

13.9

0.2

0.3

0.0

3.34

2.1

14.54

42.6

Sunflower seed** 6.4

Hazelnut kernel** 4.9

0.0

0.3

0.1

0.2

0.9

0.9

4.4

2.9

2.6

23.2

64.1

13.0

40.5

17.7

83.6

69.49

56.2

22.3

57.4

10.1

72.9

8.5

0.69

4.3

8.2

0.0

0.2

0.0

0.2

0.78

0.0

1.0

0.0

1.9

0.0

0.0

* Giannelos et al. (2002) **Demirbaş (2003)

608

Nazim Usta

2.2. Biodiesel Production There are different kinds of biodiesel production techniques. Researchers try to increase yield and quality of biodiesels. The quality of a biodiesel is checked with standards, mainly ASTM 6751 and EN14214. The general procedure applied in this study for the virgin oils (rapeseed, tobacco seed, soybean and palm oils) can be summarised as follows: the oils were converted into methyl esters by a transesterification process in which the triglycerides of the oils react with methyl alcohol in the presence of a base catalyst to produce glycerol and fatty acid esters. 6/1 molar ratio of alcohol to triglycerides, twice of stoichiometric ratio, was used to get the largest ester yield (Al-Widyan and Al-Shyoukh, 2002; Tomasevic and Siler-Marinkovic, 2003; Darnoko and Cheryan, 2000). NaOH as a catalyst was used due to its high activity (Vicente et al. 1998) and the amount of NaOH was determined using a titration process as advised in (Tickell and Tickell, 1999). Since methyl alcohol was used in the process, 55°C which is below the boiling temperature of methanol was chosen for the reaction temperature. NaOH is dissolved in methanol to produce the methoxide and then it is poured into the oil which is heated to 55°C previously. The solution was stirred for 1.5 h holding the temperature at 55°C, then the heater is turned off and the stirring is continued for 1.5h without heating. The mixture is allowed to form two layers overnight. The upper layer was the ester while the bottom layer was glycerine. The glycerine was taken out. The remaining ester was washed with pure water three times. The water is settled at bottom and it was removed. At the end of the process the ester was heated to 100°C to remove unused methanol and water from the oil left in the ester. The waste cooking oil also was transesterified as the process explained as above. Because the free fatty acid content of the oil is lower than 2%. However, it is required to be careful, if the free fatty acid content of the oil is higher than 2%, the alkaly catalyser is not advised to prevent soap production. In that case, it is advised to use acid/base catalyser process as explained below. Soap stocks have very high free fatty acid (FFA) (45–50%). When a base catalyser is directly applied to the soap stocks, the high free fatty acid content causes high soap formation (Canakci and Van Gerpen, 2001). Therefore they can not be esterified with alkali catalysers. It is required acid or acid/base catalyser. Haas et al. (2000) reported that fatty acid methyl esters (FAME) from soybean soapstock can be produced in a two stage process that involves alkaline hydrolysis of all lipid linked fatty acid ester bonds and acid catalysed esterification of the resulting fatty acid sodium salts. In addition Haas et al. (2003) reported that FAME can be produced from soapstock using only acid catalysed esterification. They found the maximum esterification occurred at 65°C and 26 h reaction at a molar ratio of total fatty acid (FA)/methanol/sulphuric acid of 1:15:1.5. In this study hazelnut soapstock and waste sunflower oil mixture was used as a source in approximately equal volume proportions. Although the waste sunflower oil reduced the FFA content of the mixture, the mixture was still not suitable for base catalyser. In this case the free fatty acid content of the mixture was reduced using an acid catalysed pre-treatment at 35°C to esterify the free acids before transesterifying the triglycerides with an alkaline catalyst to complete the reaction at 55° C. The acid/base catalyser process is summarised as follows: After the removing particles and water from the mixture, methanol (8% of the mixture in volume) was added to the mixture which is at 35°C, and it was stirred for 5 min. Then one millilitre of 95% sulphuric acid was added to the mixture-methanol. The stirring was continued for 1 h with heating

An Experimental Study on Performance and Exhaust Emissions…

609

keeping the temperature constant at 35°C and for 1 h without any heating. The mixture was left overnight. In the second stage, 3.5 g NaOH per litre of the mixture was dissolved in methanol (12% of the hazelnut soapstock-waste sunflower oil mixture) to produce methoxide. Initially half of the methoxide was poured into the mixture and mixed for 5 minutes. Then the mixture was heated to 55°C, and the rest of the methoxide was added to the heated mixture. The stirring was continued for 90 minutes. The mixture was allowed to form two layers overnight, sometimes it may take more time. The bottom layer was glycerine, while the upper layer was the ester. The rest of the process is similar to the process mentioned as above.

2.3. Experimental Apparatus for Engine Tests and Exhaust Emissions The fuel testing system consists of a Cussons-P8651 type engine test bed, a four cylinder four stroke turbocharged indirect injection diesel engine, a gas analyser and a smokemeter. The schematic of the system is shown in Fig. 1. The specifications of the diesel engine are given in Table 2. The engine test bed consists of a hydraulic dynamometer, measurement instruments and a control panel. The water cooled hydraulic dynamometer is rated for 112 kW power absorption at 9000 rpm. A strain gauge load sensor which was calibrated by using standard weights just before the experiments was used to measure the load on the dynamometer. The speed of the engine was measured using inductive pickup speed sensor calibrated by an optical tachometer. The air flowrate was measured by means of an air box, a venturi meter and a manometer. The fuel flowrate was measured with a burette with 50 and 100 ml volumes and a stopwatch. A mechanical actuator was used to adjust different loads. A data logger with K type thermocouples was used to measure air inlet, fuel, engine coolant inlet-outlet, lubricating oil, exhaust gas temperatures. A second fuel tank was fixed to the system for alternative fuels.

Figure 1. Experimental rig (1-Engine chassis, 2- Hydrokinetic dynamometer, 3-Engine tank, 4-Engine cooling unit, 5-Air tank 6- Control unit 7-Main fuel tank, 8-Alternative fuel tank, 9-Biodiesel control valve, 10-Diesel control valve, 11-Exhaust gas analyser and smokemeter).

610

Nazim Usta Table 2. Specifications of the diesel engine Type

Ford XLD 418T Turbocharged 4 Stroke, Water Cooled, IDI

Number of Cylinder

4

Stroke

82.0 mm

Bore

82.5 mm

Compression Ratio

21.5:1

Displacement

1.753 litres

Maximum Torque

152 Nm at 2200rpm

Maximum Power

55kW at 4500 rpm

Type of Injection Pump

Rotary Distributor

A Testo 350 M/XL gas analyzer was used to measure CO, NO, NO2, H2S, O2 and SO2 emissions. The smoke was measured using a Bosch BEA 170 smokemeter. Table 3 shows the accuracies of the measurements and the uncertainties in the calculated results. Table 3. The accuracies of the measurements and the uncertainities in the calculated results Measurements Accuracy Load Speed Time Temperatures CO K NOx SO2 Dynamic Viscosity Specific gravity Calculated Results Uncertainty Kinematic Viscosity Power bsfc Thermal Efficiency

m 2N m 2 rpm m 0.5 % m 1 °C m 20 ppm m 0.1 % m 20 ppm m 20 ppm m 1% m 1% m 1.4 % m 2% m 2.3 % m 2.5 %

Diesel fuel No.2 and blends containing 20% the biodiesels by volume were tested in the engine at full load. The engine was run approximately 30 min to warm up and then the engine speed was increased to 3000 rpm. The measurements were done at five different engine speeds, namely 3000, 2500, 2200, 2000 and 1500 rpm. At each speed, the engine was run approximately four minutes and then the measurement parameters were recorded at fifth minute. For biodiesel blends experiments, the diesel fuel valve was shut down and the blend valve was opened to run the engine with the biodiesel blends.

An Experimental Study on Performance and Exhaust Emissions…

611

In this study, abbreviations for diesel fuel No.2 and the biodiesels are given in the nomenclature. In addition numbers just after the abbreviations show the percentage of the fuel in the blends. For example, the legend D100 represents 100% the diesel fuel No.2, while R20 indicates a blend containing 80% the diesel fuel No.2 and 20% rape seed oil methyl ester.

3. EXPERIMENTAL RESULTS AND DISCUSSIONS 3.1. Comparison of Biodiesel’s Specifications In general, it is known that specifications of a biodiesel depend on the oil and the transesterification process. After production of a biodiesel, the specifications should be determined and checked whether they are within the limits of standards, mainly ASTM D6751 and EN14214. Some specifications such as density, viscosity and cetane number mainly depend on the fatty acid composition of the oil. Therefore if these specifications are out of the limits, it is required to use some additional improvers before using the biodiesel. However, some other specifications such as water content, monoglyceride, diglyceride, triglyceride content, free glycerol and total glycerol mainly depend on the transesterification process. These specifications can be improved by modifications in the process. Different biodiesels can be produced from the same oil using different processes. In most cases, the specifications of biodiesels produced and examined in this study are within the limits of EN14214 standard, except SOW100. Since the biodiesels were tested in a diesel engine as a blend (20% in volume), the study was focused on the performance and emission tests rather than the details of specifications. However, the densities and viscosities of the biodiesels are compared with D100. The changes of viscosities with temperature are also presented here. The densities (at 15°C) and kinematic viscosities (at 40°C) of the biodiesels are compared in Figs. 2 and 3, respectively. The densities of virgin oil biodiesels (R100, S100, C100, T100, and P100) and waste cooking oil methyl ester (W100) are close each other and approximately 5% higher than that of D100. However, the difference between D100 and SOW100 reached to 9.5 %. The higher density results in higher mass flowrate of the fuel in the engine. The kinematic viscosities of biodiesel from virgin oils (R100, S100, C100, T100, P100) are close each other and below 5mm2/s. However, the viscosity of W100 is slightly higher than 5mm2/s. SOW100 has the highest viscosity which is approximately seven times higher than viscosity of diesel fuel No. 2. Therefore, it can not be advised to use SOW100 alone in a diesel engine. In addition the changes of dynamic viscosities of the fuels with temperature range from 5°C to 40°C are compared in Fig. 4. It is shown in the figure that the differences between the viscosities of biodiesels and the viscosity of D100 increase as the temperature decreases. Again, SOW100 shows fairly high viscosity at low temperatures.

612

Nazim Usta

940 920

3

Density (kg/m )

900 880 860 840 820 800 D100

R100

S100

C100

P100

T100

W100 SOW100

Figure 2. Density values of fuels used in the tests.

30

2

Kinematic Viscosity (mm /s)

25 20 15 10 5 0 D100

R100

S100

C100

P100

Figure 3. Kinematic viscosity values of fuels used in the tests.

T100

W100

SOW100

An Experimental Study on Performance and Exhaust Emissions… 70

D100 P100

60

Dynamic Viscosity (mPas)

613

T100 W100

50

C100 40

R100 S100

30

SOW20 20 10 0 0

5

10

15

20 25 Temperature (°C)

30

35

40

45

Figure 4. Change of dynamic viscosity values of the fuels with temperature.

3.2. The Engine Performance and Emission The effects of different biodiesel additions on the performance and emission of the engine running at full load are introduced in this section. Although it is not shown here, the power variation depending on the biodiesel content in the blend was also examined using different biodiesels and it was found that the optimum power was obtained with the blends containing approximately 20 % (in volume) biodiesel. It should be pointed out that the biodiesel content in the blends is very important. After a certain amount of biodiesel content, the power reduction occurs due to the lower heating value and the higher viscosity. Although some biodiesels’ viscosities are fairly close to the diesel fuel’s viscosity, the reduction in power and efficiency with higher content of biodiesel in the blend is inevitable due to lower heating value (Kalam et al., 2003). If the biodiesel content is continued to increase, the power and efficiency of the blend will be lower than those of diesel fuel. The detailed information about this concept can be found in Usta et al. (2005) and Usta (2005b). This was more important for W100 and SOW100, because their viscosities are higher than that of diesel fuel and higher viscosity causes the bad atomization. Therefore, the seven different biodiesels were blended with diesel fuel No.2 in 20% (in volume). These blends and Diesel fuel No.2 were tested in the engine. Although the measurements were recorded at 1500, 2000, 2200, 2500 and 3000 rpm, only the measurements at 3000rpm are compared here for the sake of clarity. In addition it is better to say that the maximum torque was occurred at 2200 rpm for both diesel fuel No.2 and the blends. The effects of the seven different biodiesel additions (20% in volume) on the engine power at 3000 rpm engine speed are shown in Fig. 5. As it is seen on the figure, although the heating values of the blends are 10-11% less than that of diesel fuel, no any significant difference was determined in the engine power. The change of the power is in m 1% for

614

Nazim Usta

different blends. This can be explained with three main reasons. Firstly, since the fuel is pumped to the engine cylinders on volumetric basis, slightly higher density of the blends causes larger mass flow rate for the same fuel volume. Secondly, since the viscosities of biodiesels are higher than that of diesel fuel, more viscous blends provide less internal leakage in the fuel pump (Lang et al., 2001; Wagner et al., 1984; Al-Widyan et al., 2002). Thirdly, the biodiesels contain approximately 10-12 % (in weight) oxygen, this oxygen helps more complete combustion, thereby increasing the power (Kalam et al., 2003; Agarwal and Das, 2001). 3

Change in Power (%)

2 1 0 R20

S20

C20

P20

T20

P20

T20

W20

SOW20

-1 -2 -3

Figure 5. Change in power with respect to D100. 3

Change in bsfc (%)

2 1 0 R20

S20

C20

-1 -2 -3

Figure 6. Change in bsfc with respect to D100.

W20

SOW20

An Experimental Study on Performance and Exhaust Emissions…

615

In general, brake specific fuel consumption (bsfc) of a diesel engine depends on fuel specific gravity, viscosity, heating value and volumetric fuel injection system The fuels with lower heat content and higher density results in slightly higher bsfc with respect to the diesel fuel (Kalam et al., 2003). In the present study, bsfc values of the blends increased in the range of 0.2%-2.6% as it is shown in Fig. 6 as in (Bari et al., 2002; Kalam et al., 2003; Nwafor, 2003; de Almeida et al., 2002; Machacon et al., 2001). Highest increase in bsfc appeared with S20, meanwhile the lowest increase occurred with W20. The thermal efficiency of a diesel engine is inversely proportional to bsfc and heating value of fuel. The bsfc values of the blends were slightly higher than diesel fuel No.2, while the heating values of the blends are lower than that of diesel fuel No.2. The thermal efficiencies obtained with the blends are compared in Fig. 7. Although the thermal efficiencies of C20, P20, T20, and W20 are slightly higher than that of diesel fuels, the efficiencies of R20, S20 and SOW20 are slightly lower than that of diesel fuel. It can be said that 20 % biodiesel addition also did not make any significant change in the thermal efficiency. Cetane number of a biodiesel is very important specification. A biodiesel having a slightly lower cetane number results in longer ignition delay and slower burning rate (Nwafor at al. 2000) and longer ignition delay causes higher exhaust gas temperature. Late combustion in expansion stroke causes higher exhaust and lubrication oil temperatures. Also, there is another concept related to the biodiesel usage. Some biodiesels may contain some constituents which have higher boiling points are not sufficiently evaporated during the main combustion phase. They continue to burn in the late combustion phase resulting lower thermal efficiency and higher exhaust temperature (Yu et al., 2002). 3

Change in thermal efficiency (%)

2

1

0

R20

S20

C20

P20

T20

W20

SOW20

-1

-2

-3

Figure 7. Change in thermal efficiency with respect to D100.

The exhaust gas temperatures of the fuels are shown in Fig. 8. It can be said that the change of the temperature is not significant. The temperature of the diesel fuel No.2 was measured as 560°C at 3000 rpm. Although R20, S20 and P20 resulted in slightly higher temperatures, the other blends caused the lower temperatures. The lowest temperature was

616

Nazim Usta

obtained with T20 as 531°C. The results may imply that the addition of the biodiesel did not change the ignition delay. Figure 9 shows the lubrication oil temperatures of the blends which are in general lower than that of diesel fuel in the range of 0 - 2°C. There was no significant difference in the lubrication oil temperatures with 20% biodiesel additions. This was supported with the measurements of cooling water inlet and exit temperatures of the engine. It was found that the difference between water inlet and outlet temperatures of the engine cooling system was not affected with the biodiesel addition.

Exhaust Gas Temperature (°C)

600

580

560

540

520

500 D100

R20

S20

C20

P20

T20

W20

SOW20

Figure 8. Comparison of the exhaust gas temperatures.

Lubrication Oil Temperature (°C)

100

95

90

85

80 D100

R20

S20

C20

Figure 9. Comparison of the lubrication oil temperatures.

P20

T20

W20

SOW20

An Experimental Study on Performance and Exhaust Emissions…

617

Although 20% biodiesel addition did not change the performance significantly, the emissions were affected with the biodiesel additions. One of the important emissions from CI engines is smoke. It was determined that biodiesel addition resulted in reduction in smoke emission due to complete combustion (Fig. 10) (Kalam et al.,2003; AlWidyan et al., 2002; Monyem et al., 2001). The smoke reduction reached to 32% with S20 and C20. Meanwhile the reductions in CO emission are shown in Fig. 11. The maximum reduction was determined as 35% with W20. These reductions may be explained with extra fuel oxygen. Soot and CO compete for the available oxygen in the rich combustion regime at the full engine load. In addition although the details are not shown here, at partial loads, there is no appreciable difference between the fuels due to the dominant premixed lean combustion with excess oxygen (de Almeida et al., 2002; Monyem et al., 2001). 0 R20

S20

C20

P20

T 20

W20

SOW20

Change in K (%)

-10

-20

-30

-40

-50

Figure 10. Change in smoke emission.

Since the biodiesels contain fairly low sulphur compared to the diesel fuel (Chang et al., 1996; Giannelos at al., 2002), it was expected to decrease in SO2 emission. Figure 12 shows the change in SO2 emission due to the biodiesel addition. The reduction in SO2 emissions was higher than the expected value. Since the blend includes 20% biodiesel, SO2 reduction was expected around 20 %. However, the measured reduction reached up to 45% at full load, similar to the results found by Dorado et al. (2003). Although smoke, CO and SO2 emissions decreased with the addition of the biodiesels, NOx emission was slightly increased as it is shown in Fig. 13. It is known that three factors mainly affect NOx emission. These are oxygen concentration, combustion temperature and time. Especially, at full load, the higher temperatures of combustion chamber and the presence of fuel oxygen causes higher NOx emission (Gonzalez Gomez et al., 2000; Dorado et al., 2003; Yu et al., 2002). The increases in NOx emission are in 1-6%. The maximum increase in NOx was obtained with T20 as 5.6 %. It was also determined that the increase was fairly negligible at partial loads. Increasing of NOx exhaust emissions is an important problem

618

Nazim Usta

facing biodiesel usage. The biodiesels which are higher cetane numbers are correlated with reduced NOx emissions. However, this may not always hold for all types of engine technologies (Knothe et al., 2003). 0 R20

S20

C20

P20

T 20

W20

SOW20

Change in CO (%)

-10

-20

-30

-40

-50

Figure 11. Change in CO emission. 0 R20

S20

C20

Changee in SO 2 (%)

-10

-20

-30

-40

-50

Figure 12. Change in SO2 emission at 3000 rpm speed.

P20

T 20

W20

SOW20

An Experimental Study on Performance and Exhaust Emissions…

619

10

Change in NOx (%)

8

6

4

2

0 R20

S20

C20

P20

T 20

W20

SOW20

Figure 13. Change in NOx emission at 3000 rpm speed.

4. CONCLUSIONS In this study, seven different biodiesels were produced, and tested in a diesel engine as alternative renewable fuels. The experimental results are described as follows: 1. In addition to biodiesel production from virgin oils and waste cooking oils, it is possible to produce biodiesels from the soap stocks which have high free fatty acid content. However, the viscosity of the biodiesel produced from soapstock is fairly higher than that of diesel fuel, especially at lower temperatures. 2. The power, brake specific fuel consumption, thermal efficiency, exhaust gas temperature and lubrication oil temperature as performance parameters were examined with the addition of 20 % (in volume) seven different biodiesels. It was determined that although the heating values of biodiesels are lower than that of the diesel fuel, the biodiesel addition (20% in volume) did not cause any significant variation in the engine performance. 3. The addition of biodiesel decreased smoke and CO emissions due to the fact that biodiesels contain approximately 11- 12 % oxygen by weight, and this fuel borne oxygen helps to oxidize the combustion products in the cylinder, especially in rich region. There was a significant SO2 reduction with blends due to lower sulphur content of the biodiesel. NOx emissions slightly increased due to the presence of fuel oxygen and higher combustion temperature with the blends at full load. However, the increase is lower than 6%. 4. In these short term experiments, no obvious wear or effect on the diesel engine components was observed. However, the effects of biodiesel usage on the engine

620

Nazim Usta

components and the lubrication oil are the subject of an ongoing project in the university. As a result of all of the findings mentioned above, it can be concluded that biodiesels can be partially (20%) substituted for diesel fuel No.2 at most operating conditions in terms of performance parameters and emissions without any engine modification and preheating of the blends.

REFERENCES Agarwal AK, Das LM. Biodiesel development and characterization for use as a fuel in compression ignition engines. Journal of Engineering for Gas Turbine and Power, Transactions of ASME 2001;123:440-447. Alcantara R, Amores J, Canoira L, Fidalgo E, Franco MJ, Navarro A. Catalytic production of biodiesel from soybean oil, used frying oil and tallow. Biomass and Bioenergy 2000; 18(6):515-527. Altin R, Çetinkaya S, Yücesu HS. The potential of using vegetable oil fuels as fuel for diesel engines. Energy Conversion and Management 2001;42(5):529-538. Al-Widyan MI, Al-Shyoukh AO. Experimental evaluation of the transesterification of waste palm oil into biodiesel. Bioresource Technology 2002;85(3):253-256. Al-Widyan MI, Tashtoush G, Abu-Qudais M. Utilization of ethyl ester of waste vegetable oils as fuel in diesel engines. Fuel Processing Technology 2002;76:91-103. Antolín G, Tinaut FV, Briceño Y, Castaño V, Pérez C, Ramírez AI. Optimisation of biodiesel production by sunflower oil transesterification. Bioresource Technology 2002; 83(2):111114. Bari S, Lim TH, Yu CW. Effects of preheating of crude palm oil (CPO) on injection system, performance and emission of a diesel engine. Renewable Energy 2002;27:339–351. Canakci M, Van Gerpen J. Biodiesel production from oils and fats with high free fatty acids. Transaction of the ASAE 2001; 44(6): 1429-1436. Chang DYZ, Van Gerpen JH, Lee I, Johnson LA, Hammond EG, Marley SJ. Fuel properties and emissions of soybean oil esters as diesel fuel. J. Am. Oil Chem. Soc. 1996; 73:1549– 55. Chang DYZ, Van Gerpen JH. Fuel properties and engine performance for biodiesel prepared from modified feedstocks. Society of Automotive Engineers Paper No. 971684, SAE, Warrendale, PA, 1997. Crabbe E, Nolasco-Hipolito C, Kobayashi G, Sonomoto KIA. Biodiesel production from crude palm oil and evaluation of butanol extraction and fuel properties. Process Biochemistry 2001;37(1):65-71. Curran HJ, Fisher EM, Galude PA, Marinov NM, Pitz WJ, Westbrook C.K, Layton DW, Flynn PF, Durrett RP, Zur Loye AO, Akinyemi OC, Dryer FL. Detailed chemical kinetic modeling of diesel combustion with oxygenated fuels. Society of Automotive Engineers Paper No. 2001-01-0653971684, SAE, Warrendale, PA, 2001. Darnoko D, Cheryan M. Kinetics of palm oil transesterification in a batch reactor. JAOCS 2000;77(12):1263-1267. de Almeida SCA, Belchiora CR, Nascimentob MVG, Vieirab LDSR, Fleuryb G. Performance of a diesel generator fuelled with palm oil. Fuel 2002;81:2097-2102.

An Experimental Study on Performance and Exhaust Emissions…

621

Demirbaş A. Biodiesel from vegetable oils via transesterification in supercritical methanol. Energy Conversion and Management 2002;43:2349–2356. Demirbaş A. Biodiesel fuels from vegetable oils via catalytic and non-catalytic supercritical alcohol transesterifications and other methods: A survey. Energy Conversion and Management 2003;44(13):2093-2109. Dorado MP, Ballesteros E, Arnal JM, Gómez J, López FJ. Exhaust emissions from a diesel engine fueled with transesterified waste olive oil. Fuel 2003;82(11):1311-1315. Ergeneman M, Ozaktas T, Karaosmanoglu F, Arslan HE. Ignition delay characteristics of some Turkish vegetable oil-diesel fuel blends. Petroleum Science and Technology 1997a; 15(7-8):667-683. Ergeneman M, Ozaktas T, Cigizoglu KB, Karaosmanoglu F, Arslan E. Effect of some Turkish vegetable oil-diesel fuel blends on exhaust emissions. Energy Sources 1997b;19(8):879-885. Giannelos PN, Zannikos F, Stournas S, Lois E, Anastopoulos G. Tobacco seed oil as an alternative diesel fuel: Physical and chemical properties. Industrial Crops and Products 2002;16:1–9. Gonzalez Gomez ME, Howard-Hildige R, Leahy JJ, O’reilly T, Supple B, Malone M. Emission and performance characteristics of a 2 Litre Toyota diesel van operating on esterified waste cooking oil and mineral diesel fuel. Environmental Monitoring and Assessment 2000;65:13–20. Graboski MS, McCormick RL, Alleman TL, Herring AM. The Effect of Biodiesel Composition on Engine Emissions from a DDC Series 60 Diesel Engine, Final Report February 2003, NREL/SR-510-31461, National Renewable Energy Laboratory, Golden, Colorado, USA. Graboski MS, McCormick RL. Combustion of fat and vegetable oil derived fuels in diesel engines. Prog. Energy Combust. Sci. 1998;24:125-164. Haas MJ, Bloomer S, Scott K. Simple, high-efficiency synthesis of fatty acid methyl esters from soapstock. JAOCS 2000; 77(4): 373-379. Haas MJ, Scott KM, Alleman TL, McCormick RL. Engine performance of biodiesel fuel prepared from soybean soapstock: A high quality renewable fuel produced from a waste feedstock. Energy and Fuels 2001; 15(5): 1207-1212. Haas MJ, Foglia TA. Cheaper feedstocks for biodiesel. Industrial Bioprocessing 2002; 24(5): 4-5. Haas MJ, Michalski PJ, Runyon S, Nunez A, Scott KM. Production of FAME from acid oil, a by-product of vegetable oil refining. JAOCS 2003; 80(1): 97-102. Kalam MA, Husnawan M, Masjuki HH. Exhaust emission and combustion evaluation of coconut oil-powered indirect injection diesel engine. Renewable Energy 2003;28:24052415. Kalligeros S, Zannikos F, Stournas S, Lois E, Anastopoulos G, Teas CH and Sakellaropoulos F. An investigation of using biodiesel/marine diesel blends on the performance of a stationary diesel engine. Biomass and Bioenergy 2003;24(2):141-149. Karaosmanoglu F. Vegetable oil fuels: A review. Energy Sources 1999;21:221-231. Karaosmanoğlu F, Kurt G, Özaktas T. Long term CI engine test of sunflower oil. Renewable Energy 2000;19:219-221. Karaosmanoglu F., Tuter M., Gollu E., Yanmaz S., Altıntıg E., Fuel Properties of Cotton seed Oil. Energy Sources 1999; 21:821- 828,

622

Nazim Usta

Kitamura T, Ito T, Senda J, Fujimoto H. Extraction of the suppression effects of oxygenated fuels on soot formation using a detailed chemical kinetic model. JSAE Review 2001a;22:139-145. Kitamura T, Ito T, Senda J, Fujimoto H. Detailed chemical kinetic modeling of diesel spray combustion with oxygenated fuels. Society of Automotive Engineers Paper No. 2001-011262, SAE, Warrendale, PA, 2001b. Knothe G., Matheaus A.C., Ryan T.W. Cetane numbers of branched and straight-chain fatty esters determined in an ignition quality tester. Fuel 2003;82: 971–975. Kumar MS, Ramesh A, Nagalingam B. An experimental comparison of methods to use methanol and Jatropha oil in a compression ignition engine. Biomass and Bioenergy 2003;25:309–318. Lang X, Dalai AK, Bakhsi NN, Reaney MJ, Hertz PB. Preparation and characterization of bio-diesels from various bio-oils. Bioresource Technology 2001;80:53-62. Ma F, Hanna MF. Biodiesel production: A review. Bioresource Technology 1999;70:1-15. Machacon HTC, Shiga S, Karasawa T, Nakamura H. Performance and emission characteristics of a diesel engine fueled with coconut oil-diesel fuel blend. Biomass and Bioenergy 2001; 20: 63-69. Marshall W, Schumacher LG, Howell S. Engine exhaust emissions evaluation of a cummins l10e when fueled with a biodiesel blend. Society of Automotive Engineers Paper No. 952363, SAE, Warrendale, PA, 1995. Monyem A, Van Gerpen JH. The effect of biodiesel oxidation on engine performance and emissions. Biomass and Bioenergy 2001;20:317-325. Monyem A, Van Gerpen JH, Canakci M. The effect of timing and oxidation on emissions from biodiesel-fueled engines. Transactions of the ASAE 2001;44(1):35-42. Nwafor OMI. The effect of elevated fuel inlet temperature on performance of diesel engine running on neat vegetable oil at constant speed conditions. Renewable Energy 2003;28: 171-181. Nwafor OMI, Rice G, Ogbonna AI. Effect of advanced injection timing on the performance of rapeseed oil in diesel engines. Renewable Energy 2000;21:433-444. Ozaktas T. Compression ignition engine fuel properties of a used sunflower oil-diesel fuel blend. Energy SourA s 2000;22(4):377-382. Ozaktas T, Cigizoglu KB; Karaosmanoglu F. Alternative diesel fuel study on four different types of vegetable oils of Turkish origin. Energy Sources 1997;19(2):173-181. Piazza G.J. and Foglia T.A., Rapeseed oil for oleochemical usage Eur. J. Lipid Sci. Technol. 2001; 103: 450–454. Pramanik K. Properties and use of jatropha curcas oil and diesel fuel blends in compression ignition engine. Renewable Energy 2003;28(2):239-248. Srivastava A, Prasad R. Triglycerides-based diesel fuels. Renewable and Sustainable Energy Reviews 2000;4:111-133. Tickell J, Tickell K. From the fryer to the fuel tank: The complete guide to using vegetable oil as an alternative fuel. Green Teach Pub., Sarasota, FL, 2nd ed. 1999, 66-67. Tomasevic AV, Siler-Marinkovic SS. Methanolysis of used frying oil. Fuel Processing Technology 2003;81(1):1-6. Usta N., Öztürk E., Can Ö., Conkur E.S., Nas S., Çon A.H., Can A.Ç., Topcu M. Combustion of biodiesel fuel produced from hazelnut soapstock/waste sunflower oil mixture in a diesel engine. Energy Conversion and Management 2005; 46: 741-755.

An Experimental Study on Performance and Exhaust Emissions…

623

Usta N. Use of tobacco seed oil methyl ester in a turbocharged indirect injection diesel engine, Biomass and Bioenergy 2005a; 28: 77-86. Usta N. An experimental study on performance and exhaust emissions of a diesel engine fuelled with tobacco seed oil methyl ester, Energy Conversion and Management 2005b; 46: 2373–2386. U.S. Department of Agriculture, National Agricultural Statistics Service. Statistics of Cotton, Tobacco, Sugar Crops and Honey, Chapter II. http://www.usda.gov/nass/pubs/agr03/03_ch2.pdf, 2004. Vicente G, Coteron A, Martinez M, Aracil J. Application of the factorial design of experiments and response surface methodology to optimize biodiesel production. Industrial Crops and Products 1998;8(1):29-35. Wagner L, Clark S, Schrock M. Effects of soybean oil esters on the performance, lubricating oil, and wear of diesel engines. Society of Automotive Engineers Paper No. 841385, SAE, Warrendale, PA, 1984. Williamson AM, Badr O. Assessing the viability of using rape methyl ester (RME) as an alternative to mineral diesel fuel for powering road vehicles in the UK. Applied Energy 1998; 59(2-3):187-214. Yu CW, Bari S, Ameen A. A comparison of combustion characteristics of waste cooking oil with diesel as fuel in a direct injection diesel engine. Proc. Instn. Mech. Engrs. Part D: J. Automobile Engineering 2002;216:237-243.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 625-652

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 22

NEW MATERIALS FROM LIGNIN* Carlo Bonini and Maurizio D’Auria Dipartimento di Chimica, Universita’ della Basilicata, Via N. Sauro 85, 85100 Potenza, Italy

ABSTRACT Lignin, obtained through steam explosion from straw, was completely characterized via elemental analysis, gel permeation chromatography, ultraviolet and infrared spectroscopy, 13C and 1H nuclear magnetic resonance spectrometry. Lignin powder was used for the preparation of blends with low-density polyethylene (LDPE), linear low-density polyethylene (LLDPE), high-density polyethylene (HDPE) and atactic polystyrene (PS). The obtained blends are processable through the conventional techniques used for thermoplastics; the modulus slightly increases for most lignin-polymer blends, while the tensile stress and elongation reduce. Moreover, lignin acts as a stabilizer against the UV radiation for PS, LDPE and LLDPE. Polyurethanes were obtained treating steam exploded lignin from straw with 4,4’methylenebis(phenylisocyanate), 4,4’-methylenebis(phenylisocyanate) – ethandiol, and poly(1,4-butandiol)tolylene-2,4-diisocyanate terminated. The obtained materials were characterized by using gel permeation chromatography, infrared spectroscopy and scanning electron microscopy. Differential scanning calorimetry analysis showed a Tg at 6 °C, assigned to the glass transition of the poly(1,4-butandiol) chains. The presence of ethylene glycol reduced the yields of the polyurethanes. The use of the prepolymer gave the best results in polyurethanes formation. Steam exploded lignin was used as starting material in the synthesis of polyesters. Lignin was treated with dodecanoyl dichloride. The products were characterized by using gel permeation chromatography, infrared spectroscopy, 13C and 1H nuclear magnetic resonance spectrometry, and scanning electron microscopy. *

A version of this chapter was also published in Progress in Biomass and Bioenergy Research edited by Steven F. Warnmer published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research.

626

Carlo Bonini and Maurizio D’Auria

INTRODUCTION Lignin is a natural amorphous polymer, which, together with cellulose and hemicellulose, is one of the main constituents of wood. It is generally obtained, as a by-product in the paper production, through the separation from the cellulose fibers. Its structure depends of the kind of process used for delignification [1]. Due to its phenolic nature, many chemical modifications have been studied. For example, it has been used as a main chain on which other synthetic polymer chains can be grafted [2]. Moreover, due to the presence of the phenolic groups it is expected that it can increase the oxidation, thermal and light stability of polymeric materials. It is a low-density, low abrasive and low-cost material, features that could be interesting in its use as filler instead of inorganic fillers [3]. With certain polymers, in suitable formulations, it can give partially or completely biodegradable composites [4]. All these features are attractive from the industrial point of view, nevertheless rarely lignin has been used to obtain new materials; in fact only in the last years some trials have been made to use it as a thermoplastic. In the literature some papers have been published concerning the use of lignin as a stabilizer for plastics and rubbers, where it acts as antioxidant or modifier of the mechanical properties [4,5]. The main papers concern blends of lignin with polyvinylacetate [6], where it improves some mechanical properties, with polyethylene and polypropylene, where it acts as a stabilizer against degradation reactions [3,4,7-9], with polyvinylchloride [10], in which it increases the yielding stress, with biodegradable polyesters and polyamides, where, in the first case it improves the impact resistance, and in the second the rigidity [11]. In these works wood lignin is always used. In fact, a wide quantity of lignin is obtained as a byproduct of the paper industry and traditionally has been used as a fuel to produce energy. On the other hand, non-used lignin constitutes a major environmental problem, therefore it would be important to find for it new applications. Several uses of lignin in the synthesis of new materials have been reported. In particular, lignin has been used as raw material in the preparation of polyurethanes [12-20] and in the synthesis of graft copolymers [21-32]. Furthermore, several examples of the synthesis of polyesters from lignin have been reported [33-37]. Polyesters can be used in the formulation of polyurethane coatings. The separation of lignin from cellulose is made by means of the pulping process in the paper industry, which presents environmental problems and give altered lignin.. Nevertheless, another important source of lignin exists, straw, a very diffuse and very low cost agricultural residue and nowadays new delignification processes exist that have been largely developed mainly due to the environmental problems related to the pulping process and that allow to obtain a less altered lignin, in particular the Steam Explosion Process. Steam explosion is a technology useful for the treatment of every lignocellulosic material. In the steam explosion saturated vapour at high pressure is used to rapidly warm the biomass in a digester. The biomass is maintained at the desired temperature (130 – 180 °C) for a short time: during this period the hemicellulose is hydrolysed and dissolved. At the end of this period, an explosive decompression gives rise to a loss of water from the cells (due to the immediate evaporation of water) and the cleavage of cellular structures. In our work we use mainly straw lignin, largely present in our region, due to the large abundance of agricultural crops, separated through the Steam Explosion process, developed at the ENEA of Trisaia (Matera, Italy) to obtain blends with synthetic polymers (polyethylenes

New Materials from Lignin

627

and polystyrene), which have been chosen for their large commercial diffusion, low cost and processing temperature range close to that of lignin. The aim of our work is the obtainment of low cost materials, of potential interest in the field of packaging, health care products, agricultural films, disposable objects, to find new applications for lignin, a natural abundant and low cost polymer, for which, nowadays, only a small market exists.

CHARACTERIZATION OF LIGNIN To perform our experiments we used a steam-exploded lignin from straw. The results of the elemental analysis were C: 62.13, H: 5.88; N: 1.26; S: 0.00; O: 30.73%. We analyzed the presence of carbon and hydrogen in order to characterize the lignin, but also the presence of both nitrogen, as a marker of the presence of proteins in the lignin, and sulphur, as a marker of the presence of sulphate lignin. The presence of sulphur in our sample was not detected. The elemental analysis allowed us to give the molecular weight of the lignin expressed in phenylpropanoid (C9) units. In our case, the molecular formula was C9H10.22O3.34 with a molecular weight of 172. Elemental analysis showed that this sample was highly oxidised with a large amount of oxygen in the molecular formula. The distribution of the molecular weights of acetylated lignin was obtained by using GPC: it gave Mn = 3509, Mw = 15096, and Mz = 40966. These data confirmed the evidence that the steam explosion process induced a strong destructuration in the lignin structure giving samples with relative low molecular weight. The UV spectrum of the lignin from straw was recorded in DMF. It showed absorption at 231 nm (D = 16.8 L g-1 cm-1). We recorded the differential spectrum carrying out the spectrumin1 M NaOH vs.the standard solution in DMF. These data allowed us to give the amount in mEq g-1 of some structural features in the lignin sample. We could give the amount of syringyl and guaiacyl phenols (Type I), the amount of phenols containing conjugated double bonds (i.e. HO-Ar-CH=CH-CH2OH, Type II), and the amount of stilbenic phenols (Type IV). We found in our sample Type I (0.43mEq g-1) and Type IV (0.12 mEq g-1). The infrared spectrum showed absorptions at 1702 (carbonyl stretching), 1655 (C=O stretching in aryl ketones), 1605 and 1513 (aromatic stretching), 1459 (C-H bending in methyl and methylenic groups), 1424 (aromatic vibration coupled with C-H bending in plane), 1330 (C-H bending in plane in syringyl and guaiacyl rings substituted on C-5), 1220 (C-C, C-O, and C=O stretching), 1123 (C-H bending in syringyl units and C-O stretching in secondary alcohols), 1030 (C-H bending in plane in guaiacyl units and C-O stretching in primary alcohols), and 840 cm-1 (aromatic C-H bending out of plane). The 13C NMR spectrum of lignin from straw gave signals at δ 173 (C=O), 153 (C-3/C3’in 5-5’etherified units), 148 (C-4 in etherified guaiacyl units), 145 (C-4 in β-O-4 non etherified guaiacyl units), 138 (C-1 in β-O-4 etherified syringyl units), 135 (C-4 in β-O-4 etherified and non etherified syringyl units), 133 (C-1 in β-O-4 non etherified guaiacyl units), 130 (C-2/C-6 in benzoate), 120 (C-6 in etherified and non etherified guaiacyl units), 115 (C-5 in etherified and non etherified guaiacyl units), 112 (C-2 in guaiacyl units), 111 (C-2 in guaiacyl-guaiacyl stilbenes), 105 (C-2/C-6 in syringyl units), 87 (C-β in β-O-4 threo syringyl units), 72 (C-α in β-O-4 erythro guaiacyl and syringyl units), 60 (C-γ in β-O-4 erithro and threo syringyl and guaiacyl units), 56 (methoxy groups), and 34-20 ppm (CH3 and CH2 in

628

Carlo Bonini and Maurizio D’Auria

saturated chains). The 13C NMR spectrum was compatible with the presence of both guaiacyl and syringyl units. Furthermore, the 1H NMR spectrum on acetylated lignin showed signals at δ 1.9-2.0 (aliphatic acetates), 2.18 and 2.30 (aromatic acetates), 2.6 (benzyl protons in 3-aryl1-propanol units), 3.8 (methoxy groups), 6.6 (aromatic protons in syringyl units), 6.9 (aromatic protons in guaiacyl units), and 7.6 ppm (aromatic protons ortho to carbonyl groups). 1H NMR spectrum showed the presence of signals due to the presence of aromatic acetates, in agreement with the differential UV spectrum showing the presence of large amount of phenolic hydroxy groups in the structure (0.55 mEq g-1). The Figure 1 represents the ESEM image of the lignin from straw.

Figure 1. ESEM of lignin from straw.

PREPARATION OF BLENDS The blends have been prepared through the extrusion of the synthetic polymers and the lignin, in a nitrogen atmosphere, in a single screw and single chamber Brabender extruder with the following parameters: a. extruder diameter: 30 mm, b. L/D ratio: 25, c. compression ratio 1/3.5. In the choice of the temperatures (in the range from 160 °C to 190 °C), the degradation of lignin, observed trhough TGA, see below, has to be taken into account. In Table 1 all the samples with the legends used to identify them are indicated. The blend composition is expressed as weight % of lignin The DSC curve of lignin from 60 °C to 200 °C at 20 °C/min is reported in Figure 2. The scan shows a large endotherm near 100 °C, indicating the presence of water, then a gradual increase of the specific heat at about 161 °C, that can be easily attributed to the glass transition of lignin.

New Materials from Lignin

629

Table 1. Samples examined in the preparation of blends SAMPLE Lignin Flexirene (LLDPE) Blend containing 10% by weight of lignin and 90% by weight of Flexirene Blend containing 20% by weight of lignin and 80% by weight of Flexirene Riblene (LDPE) Blend containing 10% by weight of lignin and 90% by weight of Riblene Blend containing 20% by weight of lignin and 80% by weight of Riblene Lupolen (HDPE) Blend containing 10% by weight of lignin and 90% by weight of Lupolen Blend containing 20% by weight of lignin and 80% by weight of Lupolen Polystyrene Aldrich Blend containing 10% by weight of lignin and 90% by weight of Polystyrene Aldrich Blend containing 20% by weight of lignin and 80% by weight of Polystyrene Aldrich Polystyrene Dow Blend containing 20% by weight of lignin and 80% by weight of Polystyrene Dow Blend containing 30% by weight of lignin and 80% by weight of Polystyrene Dow

Figure 2. DSC scan of lignin from 60 °C to 200° C at 20 °C/min.

LEGEND Lignin Flexirene LF10 LF20 Riblene LR10 LR20 Lupolen LLU10 LLU20 PSA LPSA10 LPSA20 PSD LPSD20 LPSD30

630

Carlo Bonini and Maurizio D’Auria

120

Weight (%)

100 80 60

B

40 20

A

0 50

150

250

350

450

550

650

750

TEMPERATURE (°C) Figure 3. TGA scan of lignin in air (A) and in nitrogen (B) at 20 °C/min.

The subsequent heating of lignin, after cooling from 200 °C to 50 °C at the same scanning rate, does not show any change in the specific heat. Such behavior is similar to what reported in the literature for various wood lignins, that, after thermal treatment, do not show any glass transition [6]. This can be interpreted by admitting that the thermal treatment determines an increase in the glass transition temperature (Tg) related to a decrease in the chain flexibility, likely due, in turn to a chemical or physical (e.g. hydrogen bond formation) crosslinking process. The TGA scans in air and in a nitrogen atmosphere, reported in Figure 3, reveal a weight loss of about 5% at 100 °C, due to the adsorbed water, followed by the complete degradation of lignin at about 210 °C in air and 230 °C in nitrogen.

WEIGHT LOSS (%

100 99 98 97 96 95 94 93 92

0

5

10

15

TIME (min)

Figure 4. TGA isotherm of lignin at 190 °C for 30 min.

20

25

30

New Materials from Lignin

631

In order to careful evaluate the degradation of lignin, which has to be taken into account to define the processing conditions, various isotherms in a nitrogen atmosphere have been performed. As an example Figure 4 shows the isotherm obtained after taking the sample at 190°C, keeping it for 30 min at this temperature. A weight loss of 6.62% takes place. The TGA measurements allowed to establish 200 °C as the maximum temperature to be used for the extrusion, as at higher temperatures a too large degradation of lignin occurs (the complete degradation takes place at 230 °C in a nitrogen atmosphere), taking into account the weight loss of 6.62% already at 190 °C after 30 min and of the further heat developing in the extruder because of the friction. The Karl-Fisher titration revealed a water content of 4.3% in weight, confirming the result obtained through TGA. The Melt Flow Index (MFI) of the neat polymers and of the blends has been evaluated, in order to have a qualitative indication of the effect of the addition of lignin on the viscosity, and hence on the processing properties, of the materials. Figure 5 shows the trend of MFI as a function of the lignin content in the blend.

MELT FLOW INDEX (g/10min)

30 Flexirene

25 PS D

20 15

PS A

10

Riblene

5 Lupolen 0 -5

0

5

10

15 LIGNIN (wt%)

20

25

30

35

Figure 5. Dependence of MFI as function of the lignin content in the blends with Flexirene, Riblene, Lupolen, PSA and PSD.

In all cases a decrease of the MFI is obtained and, as a consequence, an increase of the viscosity, as expected, on increasing the lignin content in the blends. Tensile tests have been performed to study the effect of lignin on the basic mechanical properties of the polyethylenes and polystyrenes. In Table 2 for all the samples are reported the values of the maximum tensile stress and of elongation at breaking evaluated in the tensile tests. The maximum tensile stress decreases for all samples on increasing the lignin content. In particular for Flexirene a strong decrease occurs (about 68%) passing from the neat polymer to the blend LF10, then a slight further decrease takes place when the content of lignin is increased to 20%.

632

Carlo Bonini and Maurizio D’Auria

Passing from neat Riblene to the blend LR10 a very strong reduction of the maximum tensile stress (about 75%) is observed, value that remains constant after a further addition of lignin in LR20. Table 2. Maximum tensile stress (σmax) and elongation (εmax) at breaking Sample Flexirene LF10 LF20 Riblene LR10 LR20 Lupolen LLU10 LLU20 PSA LPSA10 LPSA20 PSD LPSD20 LPSD30

σmax (MPa) 23.0 6.1 4.4 20.8 5.2 5.5 34.5 18.0 18.8 29.6 12.0 4.6 28.8 8.5 5.9

εmax (%) 2000 112.5 104.2 1391.8 54.2 58.3 1962.5 38.5 41.8 6.6 11.8 5.6 2.5 -

For Lupolen the decrease of the maximum tensile stress is about half than in the previous cases for the blend containing 10% of lignin (about 32%). Nevertheless, on increasing the content of lignin up to 20% a further decrease of 16% occurs. For PSA and PSD a decrease of the maximum tensile stress is observed upon increasing the lignin content. Also the elongation at breaking decrease upon the addition of lignin. This decrement is particularly high for the sample containing polyethylene, while is lower for those containing polystyrene. The stress-strain curves for some samples are reported in Figure 6. In Figure 6A the curve of Flexirene is reported as an exemplum of polyethylene behavior. Similar curves are recorded for Riblene ans Lupolen, which therefore are not reported. The usual curve observed in the tensile mode at a constant deformation rate of plastic material is obtained. In the case of lignin-modified samples (FL10 is reported as an exemplum in Figure 6B), the curves appear strongly modified. The fracture thakes place at much lower stress values and the elongation at breaking is much lower. The shape of the curve is different with respect to that of neat polyethylene. It shows a ductile-stable trend, with a first linearly increasing trend, then a deviation from linearity occurs before the strain reaches its maximum value. Then, there is a region of plastin deformation with constancy of stress on elongation, then the failure takes place up to the breaking of the specimen.

New Materials from Lignin

633

Figure 6. Stress-strain curves for neat Flexirene (A), LF10 (B), neat PSA (C) and LPSA 10 (D).

In Figure 6C the stress-strain curve for PSA is reported as an example of poly(styrene)containing samples. A fragile fracture, typical of glassy polymers is observed. In particular, the fracture is fragile-unstable, in that it propagates in an unstable manner through the specimen until the elastic energy storerd in the sample itself begins not sufficient for the further growth and the fracture stops. This phenomenon repeats itself more times and it is possible to evidence on the curve the points of the crake initiation and stop up to the fracture of the specimen. For the modified sample LPSA10 (Figure 6D) the fracture is fragile too, but occurs at much lower value of the stress. The thermal propertied of the samples have been evaluated by DSC. In Table 3 the glass transition temperatures (Tg), the melting temperatures (Tm) and the heat of fusion, obtained by DSC for the neat polymers and for the blends are reported. In Figure 7 the DSC scan of Lupolen (a) and for the blends LLU10 (b) and LLU20 (c) are reported, as exempla of poly(ethylene) containing samples. From Table 3 and Figure 7 it can be observed that the presence of lignin scarcely influences the thermal behavior of neat Lupolen. In fact, the melting peak, centered at 133 °C for the neat polymer (a), slightly decreases to 129 °C for both blends (b and c). The normalized heat of fusion decreases on increasing the lignin content in the blends is observed. This beahavior may point to a lower nucleation density, then to the formation of poorer and less crystals. For the other PE-containing samples a very similar behavior, with even smaller effects, is observed.

634

Carlo Bonini and Maurizio D’Auria Table 3. Melting temperatures (Tm) and heats of fusion (ΔHm) normalized with respect to the synthetic polymer content in the blends, glass transition temperatures (Tg), evaluated by DSC Sample Flexirene LF10 LF20 Riblene LR10 LR20 Lupolen LLU10 LLU20 PSA LPSA10 LPSA20 PSD LPSD20 LPSD30 Lignin

Tm(°C) 111.0/119.8 113.7/120.0 113.7/119.9 111.5 112.5 112.5 132.7 129.5 129.4

ΔHm(J/g) 63.7 62.6 42 65.5 66.5 56.3 148.5 117.4 93.6

Tg (°C)

99.9 100.6 97.9 95.6 95.9 94.6 161

a

HEAT FLOW (endo up)

c b

60

70

80

90

100

110

120

130

140

150

TEMPERATURE (°C)

Figure 7. DSC scan from 60 °C to 150 °C at 20 °C/min for Lupolen (a), LLU10 (b), LLU20 (c).

For PSD, in Figure 8 and Table 3, we observe the glass transition at a Tg around 95 °C, scarcely influenced in the blends by the presence of lignin. A similar behavior is obtained with PSA. The dynamic mechanical properties, namely the storage component of the modulus and the loss factor have been evaluated for the neat polymers and for the blends. In table 4 the glass transition temperatures of all samples are reported together with the α-transition ones, in the case of PE-containing samples, evaluated by DMTA.

New Materials from Lignin

635

HEAT FLOW (endo up)

b

c a

80

85

90

95

100

105

TEMPERATURE (°C)

Figure 8. DSC scan from 50°C to 200 °C at 20 °C/min for PSD (a), LPSD20 (b), LPSD30 (c).

In all cases in tanδ curve two peaks are detected, the first one at lower temperature, due to the glass transition of polyethylene, the other at temperatures lower than the melting point, which corresponds to the α-transition of polyethylene and has been attributes to a crystalline relaxation due to both intralammellar and intralamellar-c schear [38]. Table 4. α-Transition temperatures (Tα) and glass transition temperatures (Tg) evaluated through DMTA Sample Flexirene LF10 LF20 Riblene LR10 LR20 Lupolen LLU10 LLU20 PSA LPSA10 LPSA20 PSD LPSD20 LPSD30

Tα(°C) 50.7 71.6 52 56.8 56.8 51.6 103.8 97.8 95.8

Tg(°C) -113.2 -111 -111 -121.6 -118 -118 -110.5 -110.5 -110.5 115 117.5 117.5 105.4 105.5 107.1

Figure 9 shows the trend of the storage component of the modulus (log E’) and of the loss factor (tanδ) as a function of the temperature, for Lupolen (a), LLU10 (b) and LLU20 (c), as exempla of the PE-containing samples.

636

Carlo Bonini and Maurizio D’Auria

Figure 9. Storage component of the modulus (logE’) and loss factor (tanδ) as a function of temperature for Lupolen (a), LLU10 (b), LLU20 (c) evaluated by DMTA.

Figure 10. Storage component of the modulus (logE’) and loss factor (tanδ) as a function of temperature for PSD (a), LPSD20 (b), LPSD30 (c) evaluated by DMTA.

In general, on passing from the neat polymer to the blends, a slight increase in Tg and a decrease in Tα are observed (except for Lupolen for which Tg is nearly constant at about – 110.5 °C).

New Materials from Lignin

637

In Figure 10 the trend of the storage component of the modulus (log E’) and of the loss factor (tanδ) as a function of temperature are reported for PSD (a), LPSD20 (b) and LPSD30 (c), as exampla of the polystyrene-containing samples. tan δ curve shows for such samples the presence of a single peak, attributable to the polystyrene glass transition. As far as the modulus is concerned, it generally increases, after the addition of lignin. Morphological investigations have been carried out, by scanning electronic microscopy, examining the surface of the extruded samples, fractured in liquid nitrogen. In Figures 11-18 the ESEM micrographs of the neat polymers and of the blends containing 10% by weight of lignin and 20% in the case of PSD are reported. The ESEM micrograph of neat flexirene (Figure 11) shows a rough, irregular surface, with evidente yielding signs, as typical for a polyethylene sample. In the presence of lignin , sample LF10 (Figure 12), the surface appears smoother, showing a more fragile fracture. Lignin particles are clearly distributed nonuniformly within the matrix and with a large distribution of their dimensions. Such particles are characterized by an internal structure and protrude out of the matrix, revealing a poor adhesion. The fracture surface of Riblene (Figure 13) shows a fibrillar appearance, where yielding signs can be noted. The sample LR10 presents a rough surface (Figure 14), nevertheless the lignin particles, inhomogeneously dispersed within the matrix are completely separated, evidencing the complete absence of interfacial adhesion.

Figure 11. ESEM micrograph of the fracture surface of Flexirene.

638

Carlo Bonini and Maurizio D’Auria

In the ESEM micrograph of neat Lupolen (Figure 15) the surface, although quite smooth, shows clear signs of yielding.

Figure 12. ESEM micrograph of the fracture surface of LF10

Figure 13. ESEM micrograph of fracture surface of Riblene.

New Materials from Lignin

639

The surface of the lignin-modified polymer appears rough (Figure 16). Also here some plastic deformation is present, nevertheless the lignin particles, smaller than in the other cases, are, separated from the matrix. In the case of PSD the fracture surface, shown in Figure 17 is mirror-like, revealing an essentially fragile fracture, typical of glassy polymers.

Figure 14. ESEM micrograph of the fracture surface of LR10

Figure 15. ESEM micrograph of the fracture of Lupolen.

640

Carlo Bonini and Maurizio D’Auria

Figure 16. ESEM micrograph of the fracture surface of LLU10.

Figure 17. ESEM micrograph of the fracture surface of PSD.

In the case of the blend containing 20% by weight of lignin, holes within the matrix are observed (Figure 18), likely due to the ablation of the lignin particles, confirming the very poor adhesion between lignin and matrix. The sample have been tested for the UV stability determination of the effect of lignin presence and content. The results of the photodegradation are reported in Figures 19-22. As far as the neat polymers are concerned, we observe that under irradiation PSD and Flexirene undergo some degradation, Riblene remains unvaried, while for Lupolen an increase of the molecular weight occurs, that may be due to crosslinking reactions among the chains induced by the UV radiation.

New Materials from Lignin

641

Figure 18. ESEM micrograph of the fracture surface of LPSD20.

Figure 19. Percent of the weight average molecular weight hold after irradiation of blends LigninFlexirene with a 15W UV lamp for 48 hours.

The addition of lignin to PSD, Flexirene and Riblene causes an increase in the average molecular weight after irradiation, on increasing the content of lignin, suggesting a protecting action against photoossidation exerted by the filler. In contrast, the addition of lignin causes a neat degradation of Lupolen which clearly decomposes. The addition of lignin to polyethylene and polystyrene does not hinder their processability. In fact, although MFI increases upon addition of lignin, the samples are always well processable. During processing, the molecular wight of polymers and their distribution can be modified, in particular the molecules undergo mechanical stress that can result in the rupture of the macromolecules themselves, with the consequent decrease of the molecular weight. In contrast, in the present case, a decrease in MFI is obtaines. It is worthy noting that such a result is in contrast with that obtained in ref. 4 for blends of polyethylene and

642

Carlo Bonini and Maurizio D’Auria

polypropylene with wood lignin. In fact, in this case, always an increase in MFI was observed upon addition of lignin to the thermoplastic polymers. Therefore, our results appear encouraging in suggesting the use of straw-lignin as a processing stabilizer. Nevertheless, in order to understand if lignin can effectively act as a processing stabilizer for the above polymers, instead of the conventional inorganic stabilizers, which are more expensive, toxic and abrasive, the trend of MFI as a function of the number of subsequent extrusions should be performed.

Figure 20. Percent of the weight average molecular weight hold after irradiation of blends LigninRiblene with a 15W UV lamp for 48 hours. 400

% Mw after irradiation

350 300 250 200 150 100 50 0 0

10

20

Lignin (wt%) Figure 21. Percent of the weight average molecular weight hold after irradiation of blends LigninLupolen with a 15W UV lamp for 48 hours.

New Materials from Lignin

643

Figure 22. Percent of the weight average molecular weight hold after irradiation of blends Lignin-PSD with a 15W UV lamp for 48 hours.

As far as the mechanical properties are concerned, in all cases a decrement of the maximum tensile stress and of the elongation at breaking occurs in the blends with respect to the neat polymers. Similar results are reported in literature for non-polar thermoplastics and lignin blends [3,4] and, more generally, for polymer composites, in which a poor compatibility takes place between the two components. Therefore, in order to evaluate the compatibility between lignin and synthetic polymers used in the blends, we have performed a series of investigations. Firstly, DSC investigations have been performed on the neat polymers and on the blends. For both the kind of blends, those containing semicrystalline polymers (polyethylenes) and those containing glassy one (polystyrenes) the DSC results point to the absence of miscibility with lignin. In fact, in the first case melting of polyethylene is scarcely affected by the addition of lignin, except for a very small lowering of the melting temperatures. For the samples containing polystyrene, the constancy of the polystyrene glass transition temperature in the blend also indicates the absence of miscibility with lignin. The same indications are given by the DMTA analysis. In fact, from the loss factor curve as a function of temperature it can be observed that the typical transition of the neat polymer (glass and α transition for polyethylenes, glass transition for polystyrenes) appear unchanged in the blends, as far as their temperature and shape are concerned. Therefore, the decrement of the stress and the elongation at breaking of the considered blends have to be ascribed to the poor miscibility between polar lignin and non polar synthetic polymers. In these systems rigidity is ordinarilt improved, but streght, elongation and toughness sacrified. In fact, also in our case, as far as the modulus is concerned, it generally increases, after the addition of lignin. This result would suggest that lignin may be used as a reinforcing agent, i.e. as a filler suitable to increase such mechanical features of the material, in general rigidity, but even the dimensional stability and the shrinkage. In general, inorganic fillers are

644

Carlo Bonini and Maurizio D’Auria

used to this target, but the lignin low abrasivity, low cost and the absence of toxicity could render it competitive with respect to these ones. In order to further investigate the interactions between lignin and the thermoplastic polymers, we have performed morphological investigations, by ESEM, on the neat polymers and on the blends. In all cases the lignin particles are distributed in a inhomogeneous way within the polymer matrix, moreover they appear separated from the matrix itself. Often they have been ablated leaving holes within the matrix. Therefore, the mechanical behavior can been ascribed to the poor adhesion between the nonpolar polymer and the more polar lignin used as a filler, that gives a poor stress transfer between the matrix and the filler and yielding of the lignin-polymer composites at lower stress values than for the unblended polymers. The reduction of the tensile stress can also be ascribed to a poor dispersion of the filler into the matrix. In all cases, therefore, the decrease of the mechanical properties is clearly attributable to the poor adhesion and dispersion of the lignin particles in the apolar polymeric matrix, which create defects that act as stress concentrators and make the specimen fragile. Nevertheless, in our opinion, the observed mechanical properties constitute a lower limit, since the distribution of the lignin particles in the matrix could be improved by using a more efficient mixing technology and the adhesion by using a compatibilizing agent. The photodegradation behavior of PSD and Flexirene clearly shows that the neat polymers are strongly degraded by the UV radiation, while Riblene remains more stable. In the lignin-containing blends we observe a much lower degradation than in the unblended polymer, while for Riblene and PSD not only a constancy of the molecular weight, which would indicate an antioxidant action by lignin, but even an increase. This could indicate the occurrence of coupling reactions between the radicals present on the polymer chains and on lignin, giving formation of branched chains. Lupolen shows a completely different behavior, with an increase of the molecular weight of the neat polymer after irradiation, likely due to branching and crosslinking. The addition of lignin induces, in contrast, high degradation. The UV stabilization performed by lignin on low-density polyethylenes (Flexirene and Riblene) and polystyrene is due to the presence of phenolic groups, which make lignin to act as a radical scavenger, inhibiting or slowering the radicalic processes of degradation. The different behavior observed in Lupolen blends may be explained taking into account that lignin (as well as other fillers, in general) distributes inhomogeneously in the polymer, in particular it concentrates in the material amorphous regions. This could explain the strong UV stabilization induced on polystyrene (amorphous), on Riblene (LDPE) and Flexirene (LLDPE), which show low crystallinity degrees, evaluated by DSC, (0.22 for both). The negative effect detected in Lupolen could lie in the much higher crystallinity (0.51) of such polymer (HDPE). Therefore lignin is poorly mixed with it, in contrast it can act as initiator in its degradation, as reported in the literature for polypropylene [4].

SYNTHESIS OF POLYURETHANES Lignin and 4,4’-methylenebis(phenylisocyanate) were suspended in THF for 5 h in the presence of catalytic amount of stannous octoate. After solvent evaporation, the product was maintained in an oven at 72 °C overnight (Table 5).

New Materials from Lignin

645

Table 5. Polyurethanes from steam-exploded lignin Entry

Lignin [g]

1 2 3 4 5 6 7 8 9 10 11 12

1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.0

Diisocyanatea [g] A 0.138 0.200 0.270 0.138 0.138 0.200 0.200 0.270 0.270

mEq

Ethandiol [g]

Yield [g]

B

0.891 1.510 2.103

2.04 3.06 4.08 2.04 2.04 3.06 3.06 4.08 4.08 1.02 1.53 2.04

0.070 0.100 0.070 0.100 0.070 0.100

0.826 0.806 0.830 0.752 0.710 0.650 0.810 0.830 0.802 1.784 2.451 3.063

A: 4,4’-methylenebis(phenylisocyanate); B: poly(1,4-butandiol)tolylene-2,4-diisocyanate terminated.

The same procedure was performed using ethandiol as reagent (Table 5). In this case lignin, 4,4’-methylenebis(phenylisocyanate), and ethandiol were suspended in THF for 5 h in the presence of catalytic amount of stannous octoate. After evaporation of the solvent the residue was treated in an oven at 72 °C overnight. The reaction was carried out using poly(1,4-butandiol)tolylene-2,4-diisocyanate terminated (Table 5). In this case, lignin and the prepolymer were suspended in THF for 5 h in the presence of catalytic amount of stannous octoate. After solvent evaporation, the residue was maintained at 72 °C in an oven overnight (Table 5). The synthesis of polyurethanes and polyesters involves the use of lignin as a source of hydroxyl groups. The total hydroxy groups in lignin was determined as described by Mansson [39] We obtained a value of 1.02 mEq g-1. In the synthesis of polyurethanes we used two, three, and four equivalents of diisocyanate, respectively. The best result was obtained using two equivalents of diisocyanate. The treatment of lignin with different amounts of 4,4’-methylenebis(phenylisocianate) gave the results reported in Table 6 (entries 1-3). After the treatment with the diisocyanate the obtained material was maintained at 72 °C in an oven to obtain crosslinking. The FTIR spectrum of this material showed the presence of a peak at 1730 cm-1 (urethane) and peaks at 1643, 1551, and 1237 cm-1 that could be attributed to the stretching of the carbonyl group in urea and to the N-H bending. The ESEM of the same material (Figure 23) showed that there was no relation between the morphology of native lignin and the obtained polymer. Polyurethanes were obtained also by using ethylene glycol as co-reagent (Table 5, entries 4-9). The presence of ethylene glycol reduced the yields of the polyurethanes. The best result was obtained when 0.270 g of the diisocyanate and 0.070 g of the glycol were used (Table 5, entry 8). FTIR spectrum showed the same type of absorption described before (1730, 1643, 1547, and 1236 cm-1). ESEM (Figure 24) showed that the morphological aspect of the new material was very different from that of lignin: while lignin showed the presence of grains of ca. 50 μm diameter, the polyurethane appeared as grains with dimension higher than 200 μm.

646

Carlo Bonini and Maurizio D’Auria

Figure 23. ESEM of the polyurethane obtained from lignin and 4,4’-methylenebis(phenylisocyanate).

Figure 24. ESEM of the polyurethane from lignin and 4,4’-methylenebis(phenylisocyanate) in the presence of ethandiol.

We used also a poly(1,4-butandiol) terminated with tolylene-2,4-diisocynate prepolymer (Aldrich, Mn = 1600). The results of our experiments are reported in Table 5 (entries 10-12). We used 1, 1.5, and 2 equivalents of the prepolymer in relation to the hydroxy content of lignin. The use of this prepolymer gave the best results in the formation of polyurethanes. While using 4,4’-methylenebis(phenylisocyanate) the obtained materials were not soluble in common solvents used for GPC analysis, the material obtained by using the butanediol prepolymer allowed this type of analysis. In Figure 25 we report the molecular weight distribution observed. In Table 6 we have collected the average molecular weights for all the polyurethanes thus obtained. We observed a large increase of the molecular weight in comparison with the lignin used for the experiments.

New Materials from Lignin

647

Figure 25. Molecular weight distribution in the polyurethane obtained from steam exploded lignin and poly(1,4-butandiol) terminated with tolylene-2,4-diisocyanate.

Table 6. Average molecular weights for polyurethanes obtained from steam exploded lignin from straw and poly(1,4-butandiol) terminated with tolylene-2,4-disocyanate Entry 10 11 12

Mn 766 1301 1371

Mw 5517 7819 6830

Mz 35268 37927 28857

Mp 939 3268 4047

FTIR spectrum showed absorption at 1730, 1548, and 1223 cm-1, in agreement with the formation of a polyurethanes (Table 7). The 13C NMR spectra of these materials showed peaks at δ 155, 154 (C=O), 148, 147 (C4 in etherified guaiacyl units), 138.5, 138 (C-1 in etherified β-O-4 syringyl units), 137 (C-1 in syringyl units), 131 (C-2/C-6 in p-hydroxyphenyl units), 130, 126 (aromatic carbons), 116 (C-5 in etherified and non etherified guaiacyl units), 112 (C-2 in guaiacyl units), 107 (aromatic carbons), 104.5 (C-2/C-6 in syringyl units), 70, 64.5, 62.5, 61 (ethereal and alcoholic carbons), 56.5 (methoxy group), 33.5, 31.5, 30, 26, 24.5, 23, 18, 16, and 14.5 ppm. These data were in agreement with the presence of signals of both lignin and aliphatic ethereal prepolymer. The same conclusion could be obtained from 1H NMR spectrum (Table 7).It showed signals due to the presence of lignin and prepolymer: in particular, the peaks at δ 1.3, 1.7, 2.0, 3.3 and 5.1 could be attributed to the presence of polybutandiol moiety. The ESEM image of the obtained materials showed the presence of very large grains (Figure 26). DSC analysis showed a Tg at -6 °C, assigned to the glass transition of the poly(1,4-butandiol) chains. In the temperature range used in DSC analysis no thermal decomposition was observed.

648

Carlo Bonini and Maurizio D’Auria

Table 7. FTIR and 1H NMR data for polyurethane obtained from lignin and poly(1,3butandiol)tolylene-2,4-diisocyanate terminated and for a polyester Polyurethane IR absorptions [cm-1] 3357 2941 2857 1730 1602 1542 1450 1370 1227 1111

H NMR peaks [δ, ppm]

1

8.71 7.54 7.22 5.45 5.04 4.28 3.80 3.32 2.36 2.08 1.60

Polyester IR absorption [cm-1] 3449 2921 2854 1811 1736 1473 1412 1335 1269 1176 1075

Figure 26. ESEM of the polyurethane obtained from steam exploded lignin and poly(1,4-butandiol) terminated with tolylene-2,4-diisocyanate.

SYNTHESIS OF POLYESTERS We tested the possible use of this type of lignin in the preparation of polyesters. Lignin (1 g) and dodecandioyl dichloride were dissolved in THF (25 mL) in the presence of a stoichiometric amount (referred to the acyl chloride) of triethylamine. After 94 h, the mixture was extracted with ethyl acetate and dried over Na2SO4. We used lignin from steam explosion as substrate in reactions with dodecandioyl dichloride. We used different lignin/dodecandioyl dichloride ratios on the basis of hydroxy content of lignin (Table 8).

New Materials from Lignin

649

Table 8. Polyesters from lignin Entry

Lignin [g]

1 2 3

1.0 1.0 1.0

R(COCl)2 [g] 1.0 0.7 0.5

mEq

Yield [g]

Mn

Mw

Mz

3.7 2.6 1.9

0.8 0.6 0.4

1915 2312 6382

18863 20153 29069

49371 355141 76416

Figure 27. ESEM of the polymer between lignin and dodecandioyl dichloride.

In Table 8 we collected the average molecular weights of the obtained polymers. The polyester with the highest values of Mn, Mw, and Mz, and, then, with a low amount of molecules with high molecular weight, was obtained by using the conditions reported in entry 2. FTIR spectra were in agreement with the formation of polyesters (Table 7). The absorptions at 1736 (C=O stretching) and 1074 cm-1 (C-O stretching) were diagnostic for the presence of the ester function. The ESEM image of the obtained polyester showed that the new material had a homogeneous structure (Figure 27). 13 C and 1H NMR spectra of the polyesters showed signals in agreement with the presence of ester function . 13C NMR spectrum showed signals at δ 179, 173.9, 169.6, 67.6, 67.4, 63.3, 44.4, 29.2, 29.1, 29.0, and 24.1 ppm. The 1H NMR spectrum showed signals at δ 7.0, 4.1, 3.6, 2.4, 2.3, 1.8, 1.6, and 1.3 ppm.

CONCLUSION Biomass has an enormous potential as a source of new interesting polymeric materials in a wide range of applications. Nevertheless, up to now only a small number of research groups is involved in such an activity. In the Basilicata region (Italy) agricultural residues are largely present, for example straw, which represents a particularly low-cost lignin source. Moreover, in the same region, Enea uses a particularly efficient and low environmental impact method (the Steam Explosion) of separation of lignin from cellulose and hemicellulose.

650

Carlo Bonini and Maurizio D’Auria

In this work we have used straw lignin, produced by Enea, blended with different commercial polymers, as polyethylenes and polystyrenes. The obtained blends are processable with the techniques in use for thermoplastics. Lignin stabilizes polystyrene and low-density polyethylene against the UV radiation. As far as the mechanical properties are concerned, the addition of lignin increases the modulus of blends, with respect to the neat polymers. Nevertheless, the presence of lignin decreases the tensile strength and the elongation at breaking. These results have been related with the poor compatibility between lignin and the synthetic apolar polymers and to the nonuniform distribution and poor adhesion of the lignin particles to the matrix. These problems may be overcome by using more efficient mixing techniques and a compatibilizing agent On the basis of the results described above we have shown that steam explosion lignin from straw can be used as partner in wood adhesive: in fact it reacts thermally with diisocyanates to give polyurethanes. This result is particularly significant considering that in our experimental conditions only a gentle thermal treatment (very different from 90-170 °C used in the farm) was used. The presence of lignin as additive in wood adhesive could be important in order to have a more stable material [40,41] In conclusion, we showed that lignin from straw obtained through steam explosion process can be used as starting material for the preparation of polyurethanes potentially useful in the formulation of wood adhesive and in the synthesis of polyesters, that can be used in the formulation of polyurethane coatings. Results on the possible use of lignin in the formulation of wood adhesive will be presented in the near future.

REFERENCES [1] [2] [3]

[4] [5] [6] [7] [8]

[9]

Lin, S. Y.; Lebo, S. E. Jr.. In: Encyclopedia of Chemical Technology, Kirk-Othmer, New York, Wiley – Interscience 1995 (volume 4). Glasser, W. G.; Sarkanen, S. Lignin: Properties and Materials. ACS Symposium series 398, Washington DC 1989, American Chemical Society. Sanchez, C. G.; Exposito Alvarez L. A. Michromrchanics of lignin/polypropylene composites suitable for industrial applications, Angew. Makromol. Chem. 1999; 272 (1), 65-70 Alexy, P.; Kosikova, B.; Podstranka, G. The effect of blending lignin with polyethylene and polypropylene on physical properties. Polymer 2000, 41 (13), 4901-4908 Glasse W, Lignin. In: Pulp and Paper, 3rd Ed., Casey JP editor, New York, Interscience Publisher, 1981, p. 39 Li, Y.; Mlynar, J.; Sarkanen, S. The first 85% kraft lignin-based thermoplastics. J. Polym. Sci. Part B: Polym. Phys. 1997, 35 (12), 1899-1910. Kosikova, B.; Demianova, V.; Kacurakova, M. Sulfur-free lignins as composites of polypropylene films. J. Appl. Polym. Sci.. 1993, 47 (6), 1065-1073 Kosikova, B.; Revajova, A.; Demianova, V. The effect of adding lignin with polyethylene and polypropylene on physical properties. Eur. Polym. J. 1995, 31 (10), 953-956 Kharade, A. Y.; Kale, D. D. Lignin-filled polyolefins. J. Appl. Polym. Sci. 1999, 72(10), 1321-1326.

New Materials from Lignin

651

[10] Feldman, D.; Banu, D. Contribution to the study of rigid PVC polyblends with different lignins. J. Appl. Polym. Sci. 1997, 66(9), 1731-1744 [11] Nitz, H.; Semke, H.; Mulhaupt, R. Influence of lignin type on the mechanical properties of lignin based compounds. Macromol. Mater. Eng. 2001, 286 (12), 737-743 [12] Cheradame, H.; Detoisien, M.; Gandini, A.; Pla, F.; Roux, G. Polyurethane from kraft lignin. Br. Polym. J. 1989, 21, 269-275. [13] Kelley, S. S.; Glasser, W. G.; Ward, T. C. Multiphase materials with lignin. 9. effect of lignin content on interpenetrating polymer network properties. Polymer 1989, 30, 22652268. [14] Yoshida, H.; Morck, R.; Kringstad, K. P.; Hatakeyama, H. Kraft lignin in polyurethanes. II, effectd of the molecular weight of kraft lignin on the properties of polyurethanes from a kraft lignin-polyethertriol-polymeric MDI system. J. Appl. Polym. Sci. 1990, 40, 1819-1832. [15] Reimann, A.; Morck, R.; Yoshida, H.; Hatakeyama, H.; Kringstad, K. P. Kraft lignin in polyurethanes. III. Effects of molecular weight of PEG on the properties of polyurethanes from kraft lignin-PEG-MDI system. J. Appl. Polym. Sci. 1990, 41, 39-50. [16] Kelley, S. S.; Ward. T. C.; Glasser, W. G. Multiphase materials with lignin. VIII. Interpenetrating polymer networks from polyurethanes and poly(methyl methacrylate). J. Appl. Polym. Sci. 1990, 41, 2813-2828. [17] Thring, R. W.; Vanderlaan, M. N.; Griffin, S. L. Polyurethanes from Alcell lignin. Biomass Bioenergy 1997, 13, 125-132. [18] Vanderlaan, M. N.; Thring, R. W. Polyurethanes from Alcell lignin fractions obtained by sequential solvent extraction. Biomass Bioenergy 1998, 14, 525-531. [19] Evtuguin, D. V.; Andreolety, J. P.; Gandini, A. Polyurethanes based on oxygenorganosolv lignin. Eur. Polym. J. 1998, 34, 1163-1169. [20] Sarkar, S.; Adhikari, B. Synthesis and characterization of lignin-HTPB copolyurethane. Eur. Polym. J. 2001, 37, 1391-1401. [21] Kundu, S. K.; Ray, P. K.; Day, A.; Sen, S. K. Infrared spectra of acrylonitrile-grafted jute fibers. J. Appl. Polym. Sci. 1989, 38, 1951-1955. [22] Lathia, A.; Chang, F. F.; Meister, J. J. Effect of class, order, family, genus, species, and recovery method of lignin on product properties of grafted lignin. Polym. Prep. 1990, 31, 648-649. [23] Meister, J. J.; Li, C. T. Synthesis of cationic graft copolymers of lignin. Polym. Prep. 1990, 31, 653-654. [24] Lathia, A.; Meister. J. J. Formation of lignin-alkoxy polyols from yellow poplar lignin. Polym. Prep. 1990, 31, 660-661. [25] Meister, J. J.; Lathia, A.; Chang, F. F. Solvent effects, species and extraction method effects, and coinitiator effects in the grafting of lignin. J. Polym. Sci., Part. A: Polym. Chem. 1991, 29, 1465-1473. [26] Meister, J. J.; Li, C. T. Graft 1-phenylethylene copolymers of lignin.1. Synthesis and proof of copolymerization. Macromolecules 1992, 25, 611-616. [27] Gunnels, D. W.; Gardner, D. J.; Chen, M. J.; Meister, J. J. Alteration of surface energy of wood using thermoplastic graft copolymers. Polym. Mater. Sci. Eng. 1992, 67, 227. [28] Meister, J. J.; Zhao, Z. Lignin graft copolymers containing a methyl methacrylate graft side chain. Polym. Mater. Sci. Eng. 1992, 67, 228-229.

652

Carlo Bonini and Maurizio D’Auria

[29] Meister, J. J.; Chen, M. J. J Graft copolymers of wood pulp and 1-phenylethylene. I. Generality of synthesis and proof of copolymerization. Appl. Polym. Sci. 1993, 49, 935951. [30] Chen, M. J.; Meister, J. J.; Gunnels, D. W.; Gardner, D. J. Binding to a hydrophobic surface by altering of the surface energy of wood using graft copolymers. Polym. Mater. Sci. Eng. 1993, 68, 243-244. [31] Meister, J. J.; Zhao, Z. Physical characterization of lignin graft copolymers with poly(methyl methacrylate) sidechains. Polym. Prep. 1993, 34, 606-607. [32] Meister, J. J.; Aranha, A.; Wang, A. Poly(3-hydroxybutyrate)-3-(hydroxyvalerate)lignin graft copolymer blends. Polym. Prep. 1993, 34, 608-609. [33] Gandini, A.; Naceur, B. M.; Guo, Z. X.; Montanari, S. In Chemical modification, properties and usage of lignin, Hu, T. Q., Ed.; Kluwer Academic/Plenum Publisher: New York, 2002, p. 57. [34] Glasser, W. G.; Jain, R. K. In Chemicals and materials from renewable resources, Bozell, J. J., Ed.; American Chemical Society: Wasshington DC, 2001, p 191. [35] Evtugin, D. V.; Gandini, A. Polyesters based on oxygen-organosolv lignin Acta Polym. 1996, 47, 344-350. [36] Guo, Z. X.; Gandini, A.; Pla, F. Polyesters from lignin. 1. The reaction of kraft lignin with dicarboxylic acid chlorides. Polym. Int. 1992, 27, 17-22. [37] Guo, Z. X.; Gandini, A. Polyesters from lignin. 2. The copolyesterification of kraft lignin and polyethylene glycols with dicarboxylic acid chlorides. Eur. Polym. J. 1991, 27, 1177-1180. [38] Mc Crum, N.C,; Read, B. E.; Williams, G. Anelastic and Dielectric Effects in Polymeric Solids, New York, Wiley-Interscience 1967. [39] Chen, C. –L. In Methods in lignin chemistry, Lin, S. Y.; Dence, C. W., Eds; SpringerVerlag: Berlin, 1992, Chap 7.1. [40] Ferri, R. Uses of lignin for an environmental sustanaible development. EPA Newsletter 2004, 77, 50-53. [41] Pucciariello, R.; Villani, V.; Bonini, C.; D’Auria, M.; Vetere, T. Physical properties of straw lignin-based polymer blends. Polymer 2004, 45, 4159-4169.

In: Encyclopedia of Energy Research and Policy Editor: A. L. Zenfora, pp. 653-674

ISBN: 978-1-60692-161-6 © 2010 Nova Science Publishers, Inc.

Chapter 23

OFFGAS RECYCLE FOR INCREASED HEAT PRODUCTION FROM AEROBIC THERMOPHILIC TREATMENT OF SWINE WASTE: PILOT STUDIES AND FULL-SCALE DESIGN *

James W. Blackburn1,2,†, Zhe Wang1,2 and Mahesh Mudragaddam1,3 Departments of Mechanical Engineering and Energy Processes1, Civil and Environmental Engineering2, and Materials Technology Center3, Southern Illinois University, USA

ABSTRACT Pilot plant experiments with both a 3.79 m3 batch and semi-continuous reactor have been performed with whole, fresh swine manure and the production of biochemical energy as heat has been both measured and calculated. The reactor operates at near atmospheric pressure and about 55° C. The systems were equipped with a patented offgas recycle process that may be shown to increase the amount of recoverable and useful energy from the reactor compared with a once-through aeration system. The batch study, a statistically-designed series of experiments, was held to investigate the relationships of initial or feed total solids concentration, fresh air fed, and offgas recycle rate to the total biochemical energy produced in the system. A linear model was developed to determine the importance of these factors in design. The model indicates optimism for improved

*

A version of this chapter was also published in Recycling: New Research edited by Christian S. Gallo and Lorenzo F. Rossi published by Nova Science Publishers, Inc. It was submitted for appropriate modifications in an effort to encourage wider dissemination of research. † Address correspondence to: James W. Blackburn, Departments of Mechanical Engineering and Energy Processes, Southern Illinois University, Carbondale, IL 62901.

654

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam operation over pilot plant work performed. The recycle concept is most useful when a reactor design is desired with a relatively shallow depth (e.g., 3 m aeration submergence), as may be found in in-ground concrete tanks. Good results may be achieved in deep aeration submergence reactors with no offgas recycle, owing to the higher oxygen transfer efficiencies in tall tanks. A combination of tall tanks and offgas recycle is synergistic with improved results. These results will be presented and discussed in the context of a full-scale farm application. When compared to once-through aeration systems, offgas recycle also leads to major reductions of emitted offgas, and aiding odor and pollutant reductions. Other potential applications will also be discussed.

INTRODUCTION This chapter discusses the aerobic thermophilic process applied to swine waste with special emphasis on batch and semi-continuous pilot plant operations. A main emphasis is on the novel idea of offgas recycle for increased heat recovery and diminished fresh air feed and emitted offgas feed. (Zhe Wang, 2001, 2003) [1,2]. The approach is shown to make more heat available for recovery without significant new power costs. The use of offgas recycle in this process has received a U.S. patent (No. 6,730,224).

STUDIES ON THE AEROBIC THERMOPHILIC TREATMENT OF SWINE WASTE Aerobic treatment of livestock slurries or other organic wastes is a natural biological degradation and stabilization process. Biodegradation is accelerated by both optimizing the supply of oxygen for microorganisms within the slurry and increasing the temperature of treated slurry arising from the exothermic (heat-evolving) treatment process. Elevation of temperature, 40 º to over 60 ºC, can be achieved in insulated reactors or lagoons. Recovered heat is significant enough for application to other on-farm functions. Mudragaddam (2002) and Mudragaddam and Blackburn (2003) [3,4] reported that heat use for building heating and for drying the swine solids after a liquid-solid separation to make a value-added fertilizer product are feasible and economical. This technology is also applicable to other agricultural and industrial aqueous slurry streams. In earlier studies, Popel and Ohnmacht (1972) [5] conducted thermophilic aerobic treatment on highly concentrated substrates including animal manure. It was found that under insulated conditions, the amount of heat due to the exothermic reaction can be utilized for heating the substrates up to 65-70ºC, while accelerating the rate of degradation of the organic matter and pasteurizing the substrates. Surucu and Chian (1976) [6] studied the effect of aerobic thermophilic treatment of high strength wastewater. They simulated the wastewater by using glucose as the main carbon and -3

energy source at a concentration of 2 kg COD m . The system could be self-sustained at 56.4°C, and over 90% removal of soluble COD (chemical oxygen demand) in this liquid feed system was observed with an RT (residence time) of 2 days. Woods and Ginnivan, 1978 [7] studied the effects of temperature on the treatment of swine slurry using a bench-scale apparatus, but with daily feeding and stirring. All of the

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

655

studies were conducted at residence times varying from 1 to 10 days, and the solid feed level -3

-3

was 45 kg m to 65 kg m . The reduction of total solids, COD and BOD5 (biochemical oxygen demand at 5 days) were greater than corresponding results at a temperature of 15°C, particularly with short residence times. Approximately 90% of the COD was removed in 3 days and 75% of the BOD5 (5-day biochemical oxygen demand) reduced in two days. It was also found that the dry matter reduction rate in all reactions at 55°C were of the same order as COD removal. Heat production from aerobic metabolism should be proportional to oxygen consumption. In the oxidation of swine waste to CO2, two major processes may occur, the breakdown of organic substrates by a variety of heterotrophs and the oxidation of ammonia by the nitrifying bacteria. At thermophilic temperatures, nitrifying bacteria were largely inhibited; therefore, the heterotrophic activity was dominant. Temperature stability played an important role in thermophilic aerobic treatment and good process control was necessary (Evans and Baines, 1980) [8]. It was also reported that the inoculation of fresh waste with previously aerated waste had a beneficial effect for waste stabilization. The waste stabilization after a period of incubation time was proportional to the initial concentration of organic matter. However, the stabilization was generally achieved in less than two days. Thermophilic aerobic treatment also had distinct effects on effluent particle size distribution with a higher percentage of small particles (74%<75µm) compared to untreated slurry (56%<75µm) (Jewell and Kabrik, 1980) [9]. With long settling times, the solid was observed to settle into 3 different zones. It is reported that the true protein was concentrated in the upper layer with smaller particles, which represents a possible means of separation. The increased ease of liquid–solid separation was observed and was presumed to be related to the decreased viscosity of the waste at thermophilic temperatures. It was well established that the net rate of bacteria growth increased with temperature and growth rate was approximately linear with the reciprocal of the absolute temperature. Above a certain maximum temperature, growth rate fell rapidly. For thermophilic conditions, the maximum temperature was 60ºC. In 1982, Evans and Svoboda [10] conducted a heat balance experiment on aerobic treatment using piggy slurry (swine waste). They reported a substantial increase of carbonaceous matter decomposition at 50ºC compared with 24º-45ºC. However, it was also found that the recovery of the heat produced in aerobic thermophilic systems was somewhat difficult. Counting the potentially recoverable heat as a percentage of the total heat released by microbial activity, the potentially recoverable heat at 50ºC was similar to that at 40ºC. In the experiment when treatment temperature was increased, heat loss at short residence times exceeded total heat production. At longer residence times, there was a net surplus of heat. The heat released from the nitrification process was obtained at 15ºC, 25ºC, and 40ºC. There was no significant nitrification at 50ºC. Evans and Svoboda also studied the effect of temperature and residence time on aerobic treatment of piggy slurry. In their opinion, the major advantage of the aerobic thermophilic process was the greater extent of degradation and higher rate of reaction. They obtained 10% more COD removal at 50ºC with residence times greater than one day. The residence time also had an effect on the thermophilic process. The BOD5 and COD removal rates received the most rapid increase within 12 hours. However, the degradation of solids was significantly more at 50ºC than that at 25ºC to 40ºC when residence times were in excess of 4 days. Although there was little need for oxygen for nitrification, the

656

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

experiments showed the oxygen requirement was higher than that at lower temperatures, presumably because of higher COD removal rates. Tjernshaugen (1982) [11] described three reactors on Norwegian pig farms designed for pollution control with heat recovery. All of these were batch cultures with the maximum temperature controlled at about 40ºC. The heat energy recovery was about 3 to 4 kwh per kwh supplied for aeration. Initially the microorganisms in slurry removed the dissolved components. These components (organic acids, phenols, indoles, sulphur compounds and low molecular weight proteins) are reflected in high BOD5 of the liquid slurry phase (supernatant) with concentrations reaching 10 kg m-3. With aerobic treatment this BOD5 can be decreased to approximately 0.1 kg m-3. A continuous treatment system removed offensive odor from hog slurry within three days was demonstrated by Evans and Thacker (1987) [12]. The suspended solids were degraded much more slowly and the slurry total BOD5, even after the treatment, remained relatively high. Slurry treated in such a way, therefore has to be applied onto land and not discharged into surface waters. In 1990, Beaudet and Gagnon [13] used swine waste to study the microbiological aspect of aerobic thermophilic treatment. Twenty-two thermophilic bacteria strains originating from different sources were found and isolated in swine waste with thermophilic treatment. Twenty-one could grow at 60ºC, 9 could grow at 65ºC, and none of the strains grew at 70ºC. All strains except five could grow at pH 9.5. Cheng and Blackburn studied the aerobic thermophilic treatment of swine waste with heat generation (Cheng, 2000, Blackburn and Cheng, 2005). [14,15] The experiments were carried out in a bench-top reactor containing 1.2 l of swine waste. The agitation was of a turbine type and the speed was 300 rpm. The default airflow was 590 ml/min that was adjusted for comparative experiments. The temperature of the reactor was controlled between 55 ± 2 °C. A series of experiments with differing initial total solids and COD strengths obtained by dilution of a common initial sample led to the conclusions that the COD removal was a good predictor of the total amount of heat generated in this aerobic thermophilic system. While a linear relationship between operating factors was developed, it was cautioned to be limited to bench-scale units with similar aeration processes. Biological heat production ranged from 0.9 to 2.5 W, depending on the operating conditions.

HEAT PRODUCTION IN THE AEROBIC THERMOPHILIC BIOLOGICAL PROCESSES Bioenergetics has as it roots, early studies on the heats of combustion of organic -1

compounds. Thornton, 1917 [16] reported the release of 13.4 Mj (kg of oxygen consumed ) for complete combustion of many types of organics, with some exceptions. This work was extended (Kharasch and Sher, 1929) [17] to suggest that most organic compounds would –1 release 13.7 Mj (kg oxygen consumed ) in complete combustion. Several kinds of microorganisms produce heat when degrading one gram of carbon presented as different substrates (Cooney et al., 1968) [18]. For instance E. coli and B. subtilis produce from 11.41 to 12.63 kcal (g carbon) –1 when supplied as glucose, molasses or soybean meal. McCarty

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

657

(1965) [19] determined the heat evolution of mixtures of microrganisms acting on specific substrates and using COD (chemical oxygen demand) as the measure for oxygen uptake or utilization. Popel and Ohnmacht, 1972 [5] found the heat evolved by heterotrophs supplied with a variety of organic substrates averaged 14.5 Mj (kg of oxygen consumed-1 ) for organic matter degradation. Roels (1983) [20] studied many cases of mixed and single substrates with axenic or mixed cultures. There was a constancy of heat production for microbial aerobic systems relative to oxygen consumption and this was equal to 14.4 Mj (kg of oxygen consumed-1). Our results (Cheng, 2000, Blackburn and Cheng, 2005) [14,15] of swine waste -1

thermophilic biooxidation was 13.9 Mj (kg of oxygen consumed ) and was typical for all of our experiments with swine waste and an indigenous inoculum. Sneath, 1978 [21] discovered a correlation between COD removal rate, which is directly related to oxygen consumption, and heat evolution. With suitable insulation, it was theoretically possible to conserve heat released in continuous culture so that the reaction would proceed at any pre-determined value up to an extent that inhibits microbial activity. When heat production is low compared to the heat losses, the need for additional energy input has been considered the main disadvantage of thermophilic aerobic treatment. The additional heat input results in excessive costs. Work by Popel and Ohnmacht, 1972 [5] show, the energy produced by the oxidation of concentrated waste can be sufficient to maintain thermophilic temperature at the high ranges (60-70ºC) for sustaining the reactions. Much of the work in the past has focused on mesophilic temperatures or low thermophilic temperatures and dilute agricultural, municipal or industrial wastewaters. Some applications exist with higher levels of COD in the feed and, as long as landfarming the residue remains acceptable in agriculture, there is no need for a high percentage of solids removal from the effluent. Effluent containing low suspended solids is possible, but requires more extensive and costly processing. These applications fit well into the objective of recovering heat from the aerobic thermophilic process. As noted, Mudgragaddam and Blackburn (2003) [4] have investigated four options for utilizing the 50-55 °C hot water produced for heating, ventilation and drying applications on the farm. Two options—space heating and producing a dry valueadded product fertilizer—were found technically and economically feasible.

PILOT PLANT RESULTS Batch Pilot Plant Studies Early work was performed using a 2-l bench-scale reactor (Blackburn and Cheng, 2005) [15]. Successful results at this scale supported the development of first, a batch pilot-scale reactor and second, a semi-continuous pilot scale reactor. A drawing of the batch pilot-scale system is shown in Figure 1. The 3.79-m3 digester tank was loaded with 3.02 m3 of fresh swine waste from a growerfinisher building with scraper for manure collection and an outside pit. With the experimental objective of taking heat and chemical measurements with the reactor near a steady temperature of 55°C, and the reality that the initial charge was over 30°C cooler, it was decided to supplement the reactor heat production with an outside water heater and internal separate heat exchanger. Thus, the system was typically agitated and heated for about two

658

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

days without aeration until the reactor temperature achieved 55°C, and then the aeration system was started, with the starting sample designated “Elapsed Time Day Zero.” The water heater remained on during the run while measuring the flow and temperatures in and out of the reactor. The heat added was measured and subtracted from the total biological heat production. Thus the heat added was not included as heat produced by the reaction. The heat load (a water-air finned heat exchanger) was turned on and off as a controller sensed the reactor temperature when exceeding the 55°C set point temperature and maintained the reactor temperature within limits of about ±2˚ C. Flow rates and temperatures were recorded to enable calculation of heat leaving the reactor through the cooling system and these values were added to the heat production of the reactor. These measured heat balance results are reported, but a calculated result based on measured changes in the COD concentrations were used for all calculations. As noted in the earlier discussion of thermodynamics in this area, calculation of biochemical heat evolution from COD removal data have been well founded.

Figure 1. Schematic diagram of the batch pilot plant.

Air flow to the variable-speed lobe blower, internally-plated with nickel and chromium, was comprised of both fresh ambient air from the building and recycle air from the offgas scrubber. After an initial charge with tap water, the scrubber did not require addition of recirculating fluid since the temperature of the scrubber was lower than the reactor offgas temperature saturated with water. Condensation from the offgas provided scrubber fluid and an overflow (not shown in Figure 1). The scrubber helped to prevent reactor foam from being recycled to the blower intake. Turbine-type air flow meters and manual valves permitted control of the fresh air flow and offgas recycle ratio. Periodically the air intakes were flushed with water spray and calibrated and reset. Readings tended to be reliable for long periods of time.

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

659

A heat loss through the reactor and piping surfaces was estimated in a series of unsteadystate high temperature pure water runs at different building temperatures. In the batch studies, the building was closed from the outside and had some heating in the winter, but building temperatures remained above freezing in the winter. In the summer, the building had no air conditioning and was often higher than the outside temperature as the temperatures also varied. In the semi-continuous runs, a new building was used with heating to avoid freezing in the winter, but had no air conditioning in the summer. Estimates of system heat loss to the surroundings were made based on the unsteady-state studies done with tap water at the tested conditions. As shown in Table 1, the measured heat balances to calculate the total biochemical reactor heat production rate generally trend in the same direction and are of the same magnitude as the heat balances calculated from COD reduction. The average biochemical heat produced by a COD removal calculation is given in equation 1. Q& B io c h e m ic a l ( k W ) = C O D i ·V · η C O D | 6 d a y s ·1 3 ,9 0 0 COD

Eqn. 1

R T |6 d a y s

& Biochemical is the biochemical heat production rate in kW, COD is the initial or Where, Q i COD

feed concentration of COD in kg COD m-3, V is the wetted reactor volume in m3, ηCOD |6 days is the fractional removal of COD over a 6-day residence time, 13,900 are the kJ of heat equivalent for utilization of 1 kg COD, and RT |6days is the reactor mean residence time for 6 days days in sec. The value of V was set for the wetted reactor volume, 3.02 m3 to calculate the average biochemical heat produced each experiment, by COD (kW), and was set to 1 m3 to calculate average volumetric biochemical heat production (kW m-3). Equation 2 presents the calculation for the COD removal rate, later shown also to be a pseudo-zero-order rate constant.

COD Removal Rate = CODi · ηCOD| 6 days

Eqn. 2

RT |6days

COD Removal Rate is in kg COD removed (m-3 day-1), and the residence time at 6 days is expressed in days. Biochemical heat produced during the experiments based on COD removal will be used -1 in this analysis, along with the measured value of 13.9 Mj/(kg oxygen consumed ), (Blackburn and Cheng, 2005) [15], a value similar to several reported in literature. The statistical design selected for the operating conditions was a fractional factorial design with the specific fresh air flow (SFAF) and the offgas recycle rate (ORR) set at high, medium and low values. A third parameter was the initial total solids (ITS) of the waste fed. This material was taken from a grower-finisher building with external pits. Because we were unable to agitate the pits, solids accumulated at the bottom. The solids content of the feed

660

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

material depended on how long it was held in the pit before our sampling. The ITS in our experiments ranged independently from 1.04 to 4.59 dry wt. percent. This provided a third major parameter to investigate relative to the biochemical heat production rate and COD removal rate. Experimental results (Table 1) yielded data of good quality with which to investigate the effects of SFAF, ORR, and ITS on the heat production calculated based on COD removal. Experiments were based on removals from the beginning of aeration until 6 days had elapsed. Other conditions and results from the batch pilot experiments are also presented in Table 1. It is noted that in our experiments, the ratio of Initial COD/Initial Total Solids (converted to kg m-3 using a specific gravity of 1), was an average of 1.36. This value will be used in calculations presented later in this chapter. Two simple linear models demonstrate the significance of the major operating variables responsible for COD reduction rates and heat production. The first linear model predicts the average biochemical heat production (ABHP) as calculated by COD from the parameters, SFAF (sec-1), the ORR (fraction), and the ITS (%). It’s square regression coefficient is 0.908. The relationship is given:

& Biochemical (kW ) = -1.887 = 822.7 ⋅ SFAF=2.002 ⋅ ORR+0.05243*ITS Q COD

Eqn. 3

& Biochemical , is the ABHP. The SFAF is based on the wetted reactor volume. Where, Q COD

-3

The second linear relationship predicts the COD Removal Rate in kg COD removed m -1 day . This relationship also had a square correlation coefficient of 0.906.

COD Removal Rate (kgCODm-3 day-1)= - 3.872 +1688.5 · SFAF + 4.117 ·ORR +1.074 · ITS

Eqn. 4.

This equation may be used to estimate what percent conversion of COD is possible given a particular initial COD concentration and residence time. Experimental COD removal was strikingly linear over time for the number of samples and replicates analyzed. We should remember that the analysis was based on the whole samples--solids-plus-liquids. Only in perhaps two of the eleven experiments did we find evidence of non-linearity in the COD vs. time relationship. We therefore utilized a pseudozero-order rate model to describe and compare the removal and heat production performance. In this model, − dCCOD / dt = k pseudo−zero. The integrated form for an ideal batch,

kpseudo-zero = CODinit - COD|RT completely-mixed, constant volume reactor is RT the residence time in days. k pseudo−zero carries units of kg COD m-3 day-1.

where RT is

Table 1. Operational conditions, thermal output and COD removal kinetics. Batch Experiment

1 2 3 4 5 6 7 8 9 10 11

SFAF, Specific Fresh Air Flow (sec-1)

0.00188 0.00249 0.00249 0.00156 0.00156 0.00147 0.00118 0.00150 0.00118 0.000968 0.00165

ORR, Offgas Recycle Rate (fraction)

0 0 0 0 0.4 0.4 0.6 0.4 0.6 0.6 0.4

ITS, Initial Total Solids (%)

4.27 1.58 2.70 4.59 1.04 3.46 3.70 3.31 2.52 1.54 1.40

CODi, Initial COD (kg m-3)

51.4 23.0 33.9 64.0 11.5 46.0 61.3 27.0 41.7 22.0 23.2

ηCOD|6 days COD Removed in 6 days (fraction)

0.473 0.527 0.611 0.338 0.786 0.520 0.500 0.697 0.430 0.600 0.516

COD Removal Rate (kg m-3 day-1)

4.05 2.02 3.45 3.61 1.51 3.99 5.11 3.14 2.99 2.20 2.00

QCOD _

Biochem

Average Biochemical Heat Produced, by COD (kW) 1.98 0.99 1.68 1.76 0.73 1.94 2.49 1.53 1.46 1.07 0.97

Q

Biochem measured _ Average

Biochemical Heat Produced, Measured (kW) 1.84 1.15 1.38 2.77 0 1.43 2.92 0 3.62 3.70 0.46

Average Volumetric Biochemical Heat Production, by COD (kW m-3)

0.654 0.327 0.555 0.581 0.241 0.641 0.823 0.505 0.482 0.353 0.320

662

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

Staying within the experimental limits tested, an estimate of best conditions was possible. Table 2 presents these estimates with the highest experimental values of the parameters used, within the experimental ranges of parameters measured. Table 2. Potentially best conditions for batch thermophilic aerobic heat production from swine waste. COD Removal SFAF (maximum ORR (maximum ITS (maximum ABHP Average value value Rate (kg COD value Biochemical experimentally experimentally m-3 day-1) experimentally Heat Production used) used) used) (kW) 3.77

7.73

0.00249

0.6

4.59

Since linear relationships are assumed, these conditions should be experimentally explored to certify that the higher performance is achievable. Higher performance might translate into residence times shorter than 6 days.

SEMI-CONTINUOUS PILOT PLANT The operation of the semi-continuous pilot plant (Figure 2) was similar to that for the batch pilot plant, except the system was intermittently filled from a feed tank and effluent was sent to an effluent tank. At typical reactor residence times of 6 days, the feed and effluent tanks held approximately two weeks worth of liquid. In practice, the feed tank was filled each 10-15 days. Once a day, one-sixth of the reactor liquid was removed to the effluent tank and 3

-1

refilled from the feed tank. This amounted to 0.50 m day . All three tanks were agitated.

Figure 2. Diagram of the semi-continuous pilot plant.

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

663

Figure 3 presents a photograph of the semi-continuous pilot plant. Figure 4 presents the results from a semi-continuous run with a residence time of 6 days with SFAF of 0.0012, ORR=0.48 and ITS (estimated fed to the reactor) =2.27 %. Also the feed for COD was measured at 30.9 kg m-3. Normally in biological degradation system studies, the total experimental elapsed time should be at least three- or four-times as long as the residence time. Since the feed tank only held 14 days worth of feed material and different samples of feed ranged widely in ITS composition, the experimental elapsed time was only 2.3-times the residence time. However this experiment yields a useful estimate of the COD removal rate at operating conditions.

Figure 3. Semi-continuous aerobic thermophilic pilot plant.

Figure 4. Results from a pilot semi-continuous run.

664

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

The reactor was filled with 3.02 m3 of initial swine waste and allowed to come to temperature at times before elapsed time zero. It was agitated during this time. When at temperature, aeration was begun and 0.5 m3 day-1 of liquid was removed and replaced with material from the feed tank at 30.9 kg COD m-3. For an ideal continuous, constant volume, mixed reactor, the rate of removal is expressed as -dCCOD/dt=k pseudo-zero, and the integrated form is k·RT=CODfed-CODeffluent. With the CODfed=30.9 kg COD m-3, CODeffluent= 12.4 kg COD m-3, and residence time, RT = 6 days, k=3.1 kg COD m-3 day-1. The CODeffluent was calculated from an average of the final four measurements. While this experiment was semi-continuous and the equations only approximately apply, the rate constant from this run can be compared with that predicted by batch runs in Equation 5. For SFAF=0.0012, ORR=0.48 and ITS= 2.27, Equation 4 provides an estimate of the kpseudo-zero of 2.6 kg COD m-3 day-1, in fairly good agreement with the average experimental values from the semi-continuous run of 3.1. It is possible that the semicontinuous design is more effective than a batch design, but conclusions must await further semi-continuous data sets.

IMPORTANCE OF OFFGAS RECYCLE IN HEAT RECOVERY Recently the concept of reducing operating heat losses by recycling the offgas air stream has been proposed and patented. It is reported to enable this process to increase the level of recoverable or extractable energy (Blackburn, 2000; Blackburn, 2001) [22,23]. This concept has been more recently confirmed in subsequent batch and semi-continuous pilot scale runs in a 3.79 m3 reaction system (Wang and Blackburn, 2003) [24]. An unpublished full-scale design model suggests about 60% of the total biochemical heat produced may be recovered as useable hot water in the 50-55 ºC temperature range It is noted that if regenerative heat exchange is used, or if heated feedstock comes to the reactor, heat recoveries of 70 or even 80% are possible. Some reactor heat losses in batch runs include heat related to raising the reactor temperatures to the operating temperature, to increasing temperature in the off-gas and the higher levels of moisture escaping at these conditions, and to heat lost through the vessel or piping walls. In continuous or semi-continuous designs, heat is also lost with the hot effluent from the reactor. For one design from an unpublished program by Blackburn for a full-scale hog farm design, the heat expected to be lost from a semi-continuous reactor included 21% in warming semi-continuous makeup feed, 1-3% through reactor and equipment walls, and about 30% in reactor offgas humidified air. The heat lost in humidified once-through offgas is the largest loss in the system we designed and any approach to lessen this sink should increase the amount of recoverable heat. High aeration rates, or at least high oxygen transfer efficiencies are also required to maximize the biological oxidation and heat production process, while minimizing the offgas flow and losses of humidified air. This is not a new problem in the design of biological aeration systems and many studies have focused on improving the oxygen transfer efficiency. A higher oxygen transfer efficiency leads to several benefits. With all other parameters being the same, higher efficiencies reduce the overall fresh airflow requirements, leading to less expensive blowers and lower energy consumption. The lower fresh air flow rates, while being humidified in the reactor to near saturation, evaporate less water than systems with lower oxygen transfer

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

665

efficiencies depending simply on mass action. There have been many efforts to improve oxygen transfer efficiency including reduction of bubble diameter and increasing the bubble’s residence time in the reactor. These may be combined and result in systems with fine bubble diffusion and a large aeration submergence. An example of this approach is presented in Figure 5 for an application to a swine farm with 2000 finishing animals and 1200 nursery animals. The waste is used for its biochemical heat content, and the blower is a once-through system with the energy of humidification minimized by using fine bubble cylindrical diffusers at high density and a 6 m aeration submergence for a high air bubble residence time. In this case, oxygen transfer efficiencies were estimated from both the pilot plant oxygen transfer efficiency data and the available literature (Newbry, 1998) [25]. Of course the increase in height increases the pressure the blower must operate against and the related power requirements for the blower. Actual designs must be optimized for these and other important factors affecting performance and economics.

Figure 5. Isometric design of a once-through air, 6 m aeration submergence aerobic thermophilic treatment system for a pork farm with 2000 finishing hogs and 1200 nursery hogs. Heat recovered as hot water is used in radiators to heat the nursery.

The total oxygen transfer efficiency turns out to be a sensitive parameter in design calculations. Actual oxygen transfer efficiency (AOTE) with published information on the total oxygen transfer efficiency (OTE) as a function of aeration submergence distance for tap water may be used to develop an estimate of the AOTE expected vs. aeration submergence in a full-scale design. First we must say that direct correlations for the strength of feed, and the temperatures and high solids conditions operated for this application seems scarce in the literature. Newbry, 1998 [25] provides real OTE data from a fine pore diffuser system with tap water at 35°C (continuous perforated membrane diffusers at various submergence depths).

666

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

We chose these data to best approximate the tube-type membrane diffusers we used in our pilot work and designs. A linear regression of Newbry’s data leads to the following linear relationship with tap water.

OTE(%) =15.3006 + 4.46914 · aerator submergence(m)

Eqn. 5.

We measured the offgas oxygen concentration in our 6-day residence time semicontinuous pilot plant run, with a 1.2 m aeration submergence and a feed Total Solids Concentration of about 2.27%. We used a paramagnetic continuous oxygen analyzer with a dehumidifier that read out in vol. % O2 concentrations in the offgas. Considering that the aeration feed had an average 48% offgas recycle at a measured concentration as low as 17%, a corrected oxygen concentration for the aeration feed gas can be estimated.

Oxygen in the Aeration Feed Gas (vol.%) = (ORR) · measured offgas concentration +(1- ORR)· 21%

Eqn. 6 .

With a measured offgas oxygen percentage at 17%, and ORR=0.48, the oxygen concentration in the combined aeration feed gas was about 19%. The AOTE was therefore 100-(17*100/19)=10.5 % Actual Oxygen Transfer Efficiency. We can use the linearized relationship from Newbry and calculate an α correction factor (AOTE/OTE) of 0.53, and project actual OTE’s for the design comparison at 1.2, 3, 5 and 10 m aeration submergence as 10.5, 15.1, 19.8, and 31.5 %, respectively. An estimation of the effects of the higher reactor aeration submergence depth and ORR may be seen in some calculations for reactors at various aeration submergence depths, with and without offgas recycle. For simplification, we based the calculation on one kg day-1 of total COD fed, a 60% maximum removal of COD in the aerobic thermophilic reactor, an assumption that the oxygen transfer rate in the reactor is equal to the COD removed (0.6 kg day-1), and both a fractional ORR for once-through systems of zero and a 0.5 fractional ORR, for the recycle offgas cases. S.T.P. conditions will be used. We will neglect actual conditions in temperature and pressure. Figure 6 presents a schematic of the simplified aeration and offgas flow.

Figure 6. Schematic of an offgas recycle system.

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

667

The ORR, is defined as the volume flow of offgas recycled, Frecycle , divided by the sum of the volume flow of fresh air, Ffresh, and the Frecycle. Rigorous calculation would require the consideration of actual pressure, temperature, and humidity conditions, but for this example, standard m3 hr-1at zero humidity will be used. The example is idealized with no additional pressure drops taken in the vessel or piping. While blowers are not very efficient and generate heat during the compression process and bubbles expanding through diffusers do work in the process, 30% of the blower power will be added to the biochemical heat production as the heat equivalent of work. Also in this example, no additional mixing is added to the reactor, other than the mixing related to rising bubbles. The reactor diameter is expected to change depending on the aeration depth to yield a constant wetted reactor volume in each case. A real reactor would have an additional 1-2 m of vapor space added to the total volume for foam management. Therefore for the same amount of COD fed, the oxygen uptake, COD removal and biochemical heat generation is constant between cases. The aeration flow rate for a constant oxygen transfer of 0.6 kg/day may be calculated with the following equation:

Faeration = Ffresh + Frecycle =

COD

max

fed

·η COD ·ν

M O2 · yO2 · AOTE

Eqn. 7. 3

-1

-1

Where, Ffresh+Frecycle are volumetric flow rates in m hr at S.T.P., CODfed (kg day ), ηCODmax, is the fraction of COD fed capable of biochemical conversion, υ is the gas molar volume, (22.4 -1 m3 kg-mole-1 at S.T.P.), MO2 is the molecular weight of oxygen =32 kg kg-mole , yO2 is the mole fraction of oxygen in air (0.21for R=0% and 0.19 for R=50%), and AOTE is a function of submergence as noted above. The Blower Power Requirement, BPR (kW) may be estimated by the following equation assuming a 60% blower efficiency:

BPR(kW ) = 0.000061· BPD · Aeration Flowrate(m3 hr -1 )

Eqn. 8.

Where BPR is the blower power requirement assuming 60% efficiency, BPD is the blower pressure drop in cm of water, and Aeration Flow rate is the volumetric flow rate exiting the blower at S.T.P. The recoverable biochemical heat production (RBHP) is calculated by calculating the total biochemical heat production, (Eqn.1), subtracting the 1) estimated losses due to the reactor and piping heat losses and the hot effluent heat loss (24% of total biological heat produced), 2) the offgas heat loss (OHL) and 3) adding a fraction of the mechanical equivalent of heat from the blower (30% of BPR) to the RBHP. Table 3 offers useful unit ratios of flow rates and heat and power parameters based on 1 kg total COD day -1 fed. A useful comparison may be made by assuming a full-scale case of 2400 finishing animals producing waste per Midwest Plan Guidelines [26]. The solids load may be estimated from the guidelines, but the COD will be based on the 1.36 factor of COD/TS determined earlier in this chapter. This leads to a COD daily feed of 1483 kg COD day-1. Concentrations depend on the amount of water used in production and wash water, a site-specific value. Total solids concentrations of 5% and COD concentrations of 69 kg m-3 are assumed.

668

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam Table 3. Unit Ratios of aeration air, offgas air and BPHP/BPR and blower power consumption

Case

A, no recycle 1.2 m submergence B, 50% recycle 1.2 m submergence C, no recycle 3 m submergence D, w/ recycle 3m submergence E, no recycle, 5m submergence F, 50% recycle, 5 m submergence G, no recycle, 10 m submergence H, 50% recycle, 10 m submergence

Blower Output (m3 hr1 daily kg COD fed-1 , STP)

BPD Blower Pressure Drop (cm H2O)

BPR Blower Power Requirement (KW kg daily COD fed-1)

Offgas to atmosphere, saturated with water vapor at 52°C, dry gas, STP basis, m3 hr1 daily kg COD fed-1

OHL Heat loss in offgas (297 kJ/kg dry gas), kW daily kg COD fed-1

RBHP Recoverable Biochemical Heat Production (kW daily kg COD fed–1)

Ratio of RBHP/ BPR** (kW/kW)

0.764

120

0.006

0.764

0.081

0

0

0.841

120

0.006

0.382

0.041

0.034

5.59

0.552

300

0.010

0.552

0.059

0.018

1.74

0.607

300

0.011

0.304

0.032

0.044

4.00

0.421

500

0.013

0.421

0.045

0.032

2.51

0.464

500

0.014

0.232

0.025

0.053

3.74

0.265

1000

0.016

0.265

0.028

0.050

3.09

0.291

1000

0.018

0.146

0.016

0.063

3.55

*Calculations based on one kg COD day-1 fed, and 60% biochemically degradable. **Does not include additional mixing power (if needed), pumping power and power required for measurement and process control. Some of each electrical-driven mechanical input would also increase the amount of recoverable heat because of the heat equivalent of work input.

It can be seen in Figure 7 that the BPR is largely a function of submergence, with deep submergence reactors having the highest BPR. The OHL is very high for shallow, no recycle systems but becomes comparable to the recycle systems for deeper submergence reactors. The percent of recoverable heat is much lower for the no recycle systems and increases significantly for the systems with recycle. It also increases with reactor aeration submergence.

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

1.2 m no recycle

3 m no recycle

5 m no recycle

10 m no recycle

12 m w/ recycle

3m w/ recycle

5m w/ recycle

669

10m w/ recycle

Figure 7. Projections of BPR, OHL and the percent of RBHP which is recoverable for the various submergences, with and without recycle.

It can be seen in Figure 7 that the BPR is largely a function of submergence, with deep submergence reactors having the highest BPR. The OHL is very high for shallow, no recycle systems but becomes comparable to the recycle systems for deeper submergence reactors. The percent of recoverable heat is much lower for the no recycle systems and increases significantly for the systems with recycle. It also increases with reactor aeration submergence. A common calculation to compare the energy efficiency of aerobic thermophilic systems is the RBHP/BPR ratio. Low values indicate a system that may not produce as much heat as the BPR. High values of the ratio are related to systems that produce much more heat than the electric power input (the major energy input to the process). Figure 8 clearly shows the importance of offgas recycle relative to no recycle systems. Further, if the relative cost of electric power divided by the value of recovered heat is, for instance, 2, the region of positive cost savings lies above 2 RBHP/BPR ratio. This assumes that the value of heat produced is equivalent to the value of the fuel it is replacing. The fuel in this case is propane. Since this is a linear function and is constant with the size of the plant, and the plant’s capital costs rise with size to the 0.6 exponent the plant economics and profitability will improve with scale.

670

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

Figure 8. RBHP/BPR Ratio for various submergences and with and without recycle.

Finally, there is a major difference in the offgas emitted with no recycle systems being much higher than recycle systems as seen in Figure 9. Recycle systems have 40-50% less offgas emitted than the no recycle systems (calculated at S.T.P.). Since the gas can be expected to be saturated with volatile organic compounds (VOC) present in the reactor liquid, a major reduction in VOC emissions and related odor emissions may also be expected. 1200

Offgas Flow (sm3 hr-1)

1000

800

Offgas Flow, no recycle (sm3 hr-1)

600

Offgas Flow, 50% recycle (sm3 hr-1)

400

200

0 0

2

4

6

8

10

12

Aeration Submergence (m) 3

Figure 9. Reductions in offgas emissions based on 50% offgas recycle. (Note that in English units, sm 1 3 1 hr =1.7 sft min .)

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic…

671

FURTHER CONSIDERATIONS Using offgas recycle in the aerobic thermophilic system requires additional considerations. The aerobic thermophilic process applied to swine waste sometimes leads to significant foam levels at the top of the reactor. These seem to be most prevalent when the reactor is anoxic or anaerobic for a time before aeration begins. When these conditions occur, the possibility of foam being recycled to the blower exists. Earlier it was mentioned that a 1-2 m head space be used in the reactor to give extra time for the foam to coalesce. As a second defense, either chemical anti-foams must be added or devices installed in the reactor head space to mechanically cut the foam. These devices may also be used on the recycle line with a liquid trap to break the recycled foam. The economics of each alternative should be considered for a recycle system. Finally, the blower must be chosen with the potential of coping with foam/biological solids entering the blower. Many blowers can handle a certain amount of contamination, while others require a clean gas stream. This requires selection of an appropriate blower design for the application. The foam is corrosive and alloy blowers or alloy-plated blowers must be specified. It is also appropriate for the blower intake to automatically and periodically be washed with tap water droplets to minimize any accumulation of solids in the blower.

OTHER APPLICATIONS This process may be considered for a number of agricultural or industrial aqueous slurries containing COD. One such possibility is the application of aerobic thermophilic processing to the dry-grind corn-to-ethanol process. In this case, a large fraction of the corn fed to the process leaves the system as a dry, distiller’s grain product, currently sold as an animal feed supplement. With an appropriate inoculum, the aerobic thermophilic process can convert some of the by-product’s organic constituents to heat. In an ethanol plant, the energy balance is critical. Large amounts of fossil fuels are used to provide processing heat as boiler steam. Hot water produced in the aerobic thermophilic process can be used to heat the make-up boiler feed water, reducing the amount of fossil fuel required. Another application in the ethanol process for the hot water produced is to supply energy to a vacuum distillation unit used to recover ethanol from the fermentation broth. This might be the most important single heat requirement of an ethanol process. The processing energy in the ethanol plant can be lowered significantly with this strategy. Residuals from the aerobic thermophilic system can still be sold as an animal feed, likely with a higher protein content. Also the stream coming to the aerobic thermophilic system may have temperatures well above ground water temperatures. This reduces the amount of energy lost in heating the feed to reactor conditions and increases the energy recovery efficiency based on the heat equivalent of biochemicallydegradable COD fed to 70-80%.

672

James W. Blackburn, Zhe Wang and Mahesh Mudragaddam

CONCLUSION Batch pilot plant experiments at the scale of 3.79 m3 confirm the energy-producing characteristics of the process as applied to swine waste. Batch pilot-scale system results yield linear multiple regression models that permit estimation of both heat production and the rate of COD removal as functions of the specific volume of fresh air added, the offgas recycle rate and the initial total solids concentration. Results from these models indicate that there may exist more optimal conditions with related increases in heat production and COD removal rate. If so, the time of the process could be reduced from the current 6-day residence time used in the experiments and full-scale design. The semi-continuous pilot experiment indicates some agreement in applying the batch linear regressions to a semi-continuous process, however the performance of the experimental run was superior to that predicted by model. Until further pilot explorations are performed, no conclusions can be drawn from the increased performance of the semi-continuous experiment. A design concept for an on-farm treatment system was developed and an isometric drawing of a possible full-scale system was presented. To a point, successful results may be enjoyed using a deep aeration submergence reactor without recycle, but recycle can be effective in attaining superior energy production potential in reactors with lower aeration submergence. Offgas recycle reduces the emitted offgas volume to 40-50% lower values than no recycle systems. This should lead to a significant reduction in both VOC and odor emissions. Energy operating cost savings are a linear function with capacity while capital cost increases with a 0.6 exponent relative to system capacity. A large economy-of-scale appears to exist. Aerobic thermophilic treatment is relatively generic across applications and should be considered for use on other agricultural or industrial aqueous/organic slurries.

ACKNOWLEDGMENTS This work was performed through grants from the Swine Odor and Waste Management project of the Illinois Council on Food and Agricultural Research. It was also supported by Illinois DCCA and a supplementary grant from US DOE. Southern Illinois University at Carbondale matched funds through the SIUC Materials Technology Center and other University organizations.

REFERENCES [1]

[2]

Wang, Zhe (2001) Energy production from swine waste using pilot-scale thermophilic treatment with offgas recovery. MS thesis. Graduate School, Southern Illinois University at Carbondale. University Microfilms, Ann Arbor MI. Wang, Zhe (2003) Energy production from swine waste using continuous pilot-scale thermophilic treatment with offgas recovery. MS thesis. Graduate School, Southern Illinois University at Carbondale. University Microfilms, Ann Arbor MI.

Offgas Recycle for Increased Heat Production from Aerobic Thermophilic… [3]

[4]

[5] [6] [7]

[8]

[9] [10] [11]

[12] [13] [14]

[15]

[16] [17]

[18]

673

Mudragaddam, M. (2002) Hot water usage from an aerobic-thermophilic swine waste reactor during the hot months. MS thesis. Graduate School, Southern Illinois University at Carbondale. University Microfilms, Ann Arbor MI. Mudgragaddam, M. and Blackburn, J.W. (2003) Aerobic thermophilic swine waste reactor—use of heat production during warm weather months. Report to the Illinois Dept. of Commerce and Community Affairs, Springfield IL. Pöpel, F. and Ohnmacht, C. (1972) Thermophilic bacterial oxidation of highly concentrated substrates. Water Res., 6(7): 80. Surucu, G. A., Chian, E. S. K. (1976) Aerobic thermophilic treatment of high strength waste waters. J. Wat. Pol. Cont. Fed. 48(4): 665-610. Woods, J. L. and Ginnivan, M. J. (1978) Thermophilic treatment of piggy slurry. Engineering problems with effluents from livestock. EEC Seminar, Cambridge. (pp. 415427). Evans, M. R. and Baines, S. (1980) Factors in the design of aerobic treatment for th animal wastes. Proceedings of the 4 International Symposium on Livestock Wastes. ASAE. St. Joseph, MI. Jewell, W. & Kabrik, R. D. (1980) Autoheated aerobic thermophilic digestion with aeration. J. Wat. Pol. Cont. Fed., 52: 512-513. Evans, M. R. and Svoboda, I. F. (1982) Heat from aerobic treatment of piggy slurry. J. Agr. Engr. Res. 27:45-50. Tjernshaugen, O. (1982) Methods for recovery of heat