INTERFACE SCIENCE AND TECHNOLOGY - VOLUME 1
Clay Surfaces Fundamentals and Applications Edited by
Fernando Wypych and Kestur Gundappa Satyanarayana* Universidade Federal do Paraná Curitiba, Brazil * Visiting Professor
2004
ELSEVIER ACADEMIC PRESS
Amsterdam - Boston - Heidelberg - London - New York - Oxford - Paris San Diego - San Francisco - Singapore - Sydney - Tokyo
Clay Surfaces Fundamentals and Applications
INTERFACE SCIENCE AND TECHNOLOGY Series Editor: ARTHUR HUBBARD In this series: Vol. 1: Clay Surfaces: Fundamentals and Applications Edited by F. Wypych and K.G. Satyanarayana
ELSEVIER B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam The Netherlands
ELSEVIER Inc. ELSEVIER Ltd 525 B Street, Suite 1900 The Boulevard, Langford Lane San Diego, CA 92101-4495 Kidlington, Oxford OX5 1GB USA UK
ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK
© 2004 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier Ltd, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier's Rights Department in Oxford, UK: phone (+44) 1865 843830, fax (+44) 1865 853333, e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http ://www .elsevier .com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 207631 5555; fax: (+44) 207631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier's Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made.
First edition 2004
Library of Congress Cataloging in Publication Data A catalog record is available from the Library of Congress. British Library Cataloguing in Publication Data A catalogue record is available from the British Library.
ISBN: 0-12-088439-9 ISSN (series): 1573-4285 The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.
Preface The purpose of this book is to introduce the reader to the fascinating world of the chemistry and physics of the clay mineral surfaces/interfaces and to demonstrate that, in spite of lot of information that has been accumulated over the decades, a vast field still exists to be explored. This book will not only be useful for the specialists directly involved with clay science and related areas but also as a text or supplementary reading for undergraduate and graduate students. When a crystalline or amorphous solid is brought in contact with a liquid, a gas or another solid, the possible reactions and interactions take place via their surfaces. The study of the chemistry and physics of the surfaces plays a fundamental role in an attempt to understand those phenomena in a general sense. Apart from the structure, the composition is usually very different from the bulk of the crystal, mainly due to the sorption of ions and/or neutral molecules on the surface of the crystal. Another fundamental point in the reactivity of materials is related to the defects on the crystals' surface and surface reconstruction, the first originating in the preparation procedure and the second in a mechanism of compensation of the unsaturated bonds when a surface is generated by growth or fracture. All these phenomena are of fundamental importance for several branches of science. However, many of them remain slightly misunderstood, mainly based on the fact that only in the last few decades has appropriate equipment been developed and/or improved for this kind of studies. Usually, the information generated by the atoms of the surface is infinitely weaker than that by the atoms of the bulk, which makes it difficult or hinders the studies of the interfaces of the crystals. Considering the clay minerals and the possible interactions with multi components of the soil (that can be obtained by the degradation of organic matter of vegetable or animal origin, deliberately introduced through agricultural activities or via the decomposition cycle of rocks and minerals, apart from the activity of microorganisms), it is easy to imagine the extreme complexity of the phenomena on the surfaces involved in those systems. In this book, we have made an attempt to describe the clay surface/interfaces phenomena with the main purpose to try to understand the interrelations among the minerals, living organisms and human activities on our planet. The above aspects of clays are presented in this volume. This book contains 17 chapters, which have been classified under two main headings, viz., Natural clays: theoretical aspects and applications and Synthetic clays: Synthesis and applications with an introductory chapter by Prof. Fernando Wypych. This chapter defines layered materials and deals with derivatives of simple hydroxides, hydroxysalts, synthetic clays (Layered Double Hydroxide (LDH)) and natural clays. Herein various reactions such as adsorption, solvation, grafting, exfoliation, thermal, mechanochemical modifications and intercalation have been explained in respect of the mentioned derivatives. The Section under Natural clays: theoretical aspects and applications includes 10 chapters. Chapters 2—5 deal with the theoretical aspects of clay surfaces. Knowing that the electrokinetic properties of a substance are used to explain the mechanism of dispersion and agglomeration in a liquid phase and identification of the adsorption mechanisms of ions or molecules at a solid-liquid interface, the second chapter presents an overview of the electrokinetic properties of clay minerals, elucidating the electrokinetic behavior of clay surfaces and the mechanism of particle-particle interactions in aqueous systems. On the other hand, considering various forces that operate at the interface where the two surfaces interact and the importance of such
vi
Preface
studies on colloidal materials such as clays, the third chapter presents the recent thermodynamic study of clay surfaces, both swelling and non-swelling type, and relates it to the data obtained by measurements of surface tension, immersion enthalpy and sorption. The properties of any material near surfaces or interfaces are different to the properties of the same material in the bulk because of the different coordination environment of surface atoms than that of those in the bulk. This special character of surfaces has some thermodynamic implications, especially in terms of interaction with molecules of other substances in close contact. This aspect is described in the fourth chapter. Taking into account the vital role played by the adsorption/desorption processes in determining the efficacy and environmental behavior of pollutants or nutrients in soil, and most of the adsorption in natural systems occurs in the dispersed phase that consists predominantly of inorganic colloids such as clays, the fifth chapter presents a model for these processes and also other mechanisms and reactions for predicting the behavior of reactive solutes in complex systems in both single and multi-component systems with several interacting species. Characterization of clay surfaces, including modification of kaolinite surfaces, intercalation by mechanochemical method as studied by Raman and IR spectroscopic techniques are described in Chapter 6. It is well known that NMR studies provide structural and dynamic information at a local and sometimes mesoscopic scale. Hence a brief outline of the relevant NMR theory, starting with the solid state and simplification brought about by the mobility increase in the liquid state and studies of natural, synthetic, or modified clay suspensions along with some NMR results on solid systems are summarized in Chapter 7. Some of the important applications of clays include ceramics, paper, paint, plastics, drilling fluids, foundry bondants, chemical carriers, liquid barriers, decolorization, and catalysis. Further, pesticides, toxic organic chemicals, greenhouse gases, heavy metals, undesirable inorganic substances are as much targeted molecules to be controlled for the preservation of flora and fauna and safety of the earth. Accordingly, Chapters 8-10 deal with various application areas of natural clays mentioned, including removal of a wide range of contaminants from industrial effluents or wastewater by anion exchange and adsorption processes or catalytic remediation, using LDHs, modified LDHs or calcined LDHs, pharmaceutical & cosmetics, environmental remediation, waste treatment including nuclear wastes, heavy and toxic metals removal, biological applications, etc. Removal of metals, particularly of chromium in wastewater, by natural and modified clays is described in Chapter 10. The catalytic and adsorption properties of modified clays is given in Chapter 11 covering some general aspects of structure of clays, cation exchange capacity and swelling capacity of clays, clay-organic cation- interaction and acid-activated organo clays with a view to understand better the relationship between the clays and their catalytic uses. The section on Synthetic clays: Synthesis and applications has six chapters. Over the last two decades, interest has been growing in the availability for the intercalation of various organic anions having flexible or rigid molecular frameworks into LDH, not only from the scientific but also from industrial viewpoints. This includes synthetic anionic clay materials. Accordingly, the state of the art for the materials composed of the assembly between Layered Double Hydroxides (LDH) and Polymer are described in these chapters and new trends in term of applications are identified. It is pointed out here that the polymer/LDH assembly, not yet extensively studied, constitutes an appealing new class of nanocomposites in numerous topical applications. Chapter 12 is related to the synthetic methods of layered double hydroxides and mechanism of decomposition studied by in-situ techniques. Chapter 13 concentrates
Preface
vii
on the newest research progress in the compositions, structures, synthesis methods and photocatalytic and oxidation-catalytic functions of the POM-LDHs complexes. The preparation and applications of a new emerging class of bio-clay hybrids are presented and discussed in Chapter 14. A wide range of contaminants can be removed from industrial effluents or wastewater by anion exchange and adsorption processes or catalytic remediation, using LDHs, modified LDHs or calcined LDHs. With this background, Chapter 15 describes the investigations on the potential uses of LDHs for decontamination of environmental sites or prevention of pollutant dispersion in Nature. Chapter 16, dealing with Layered Double Hydroxide/Polymer nanocomposites, gives a description of LDH materials including their natural occurrence, chemical composition and aspects of the stacking sequence, the building of inorganic-organic assemblies including synthetic pathways for the LDH/polymer assembly along with the layer charge density and the colloidal and exfoliation properties. Chapter 17 describes a broad spectrum of catalytic applications of layered double hydroxides and possibilities of designing catalysts tailored for specific reactions and/or substrates. With the accumulation of the above knowledge, we expect to understand in the future the mechanisms of the interactions of the soils when in contact with the wastes originated from human activities. We also hope to understand what action should be taken in the case of an accidental spilling of chemicals, how to improve the production of food without affecting the ecological balance significantly, how to increase the fixation of carbon in the soil to increase the production of cereal grains and reduce carbon dioxide emissions to the atmosphere, and safely predict the effects of a certain activity on the soil, waters and atmosphere of our fragile planet. We hope this book will be good reading material for all concerned with all aspects of mother Earth. It is not an exaggeration to say that this volume represents the expertise, time, efforts and opinions of 27 specialists in their subject areas. The statements, views and recommendations made by each of the contributors are their own and should be considered to be made with appropriate responsibility. The editors express their sincere and heartfelt gratitude to all these contributors for their devotion and providing the chapters, for carrying out alterations/modifications as and when called for to suit the overall uniformity of this volume. They also express their gratitude to Louise Morris and Derek Coleman of Elsevier for their guidance and help from time to time and last but not the least, to Dr. Arthur Hubbard for the invitation, encouragement and the first table of contents of the book. Fernando Wypych Kestur Gundappa Satyanarayana
Contributors ALBERTO LOPEZ-GALINDO *' and CESAR VISERAS 2 1 Instituto Andaluz de Ciencias de la Tierra (CSIC-UGR). Facultad de Ciencias, Campus Fuentenueva. 18071 - Granada - SPAIN. 2 Departamento de Farmacia y Tecnologia Farmaceutica, Facultad de Farmacia. Universidad de Granada. 18071, Granada - SPAIN. E-mail:
[email protected] * E-mail:
[email protected] ALEXANDER MORONTA Centro de Superficies y Catalisis, Facultad de Ingenieria, Universidad del Zulia, Maracaibo 4003-A - VENEZUELA. E-mail:
[email protected] B.S. JAI PRAKASH Department of Chemistry, Bangalore Institute of Technology, k.r. road, Bangalore 560 004, INDIA. E-mail:
[email protected] CHANGWEN HU '* and DANFENG LI 2 1 Department of Chemistry, Beijing Institute of Technology, Beijing, P.R. CHINA, 100081 2 Institute of Polyoxometalate Chemistry, Northeast Normal University, Changchun, P.R. CHINA, 130024 * E-mail:
[email protected] CLAUDE FORANO Laboratoire des Materiaux Inorganiques, UMRCNRS 6002, Universite Blaise Pascal, 63177, Aubiere Cedex - FRANCE E-mail:
[email protected] CRISTINA VOLZONE Centro de Tecnologia de Recursos Minerales y Ceramica - CETMIC (CIC-CONICET-UNLP) - CC 49, Cno. Centenario y 506, (1897) M.B. Gonnet Provincia de Buenos Aires - ARGENTINA E-mail:
[email protected] /
[email protected] EIJI KANEZAKI Department of Chemical Science and Technology, Faculty of Engineering, The University of Tokushima, 2-1 Minamijosanjima, Tokushima 770-8506 - JAPAN E-mail:
[email protected] FABRICE LEROUX * and JEAN-PIERRE BESSE Laboratoire des Materiaux Inorganiques, UMR 6002-CNRS, Universite Blaise Pascal, 24 av. des Landais, 63177 Aubiere cedex, FRANCE. * E-mail:
[email protected] FERNANDO WYPYCH Centro de Pesquisas em Quimica Aplicada - CEPESQ. Universidade Federal do Parana - UFPR - Departamento de Quimica CP 19081 - Centro Politecnico - 81531-990 - Curitiba - PR - BRAZIL. E-mail:
[email protected]
Contributors
ix
GIORA RYTWO School of Environmental Sciences and Technology - Tel Hai Academic College, Upper Galilee 12210, ISRAEL. MIGAL, Galilee Technological Center, Kiryat Shmona, ISRAEL. E-mail:
[email protected] JEAN GRANDJEAN Universite de Liege - Institut de Chimie B6a - COSM Sart Tilman - B-4000 Liege BELGIUM E-mail.:
[email protected] JEAN MARC DOUILLARD* and FABRICE SALLES University of Sciences, L.A.M.M.I., CC015, Universite Montpellier 2. Sciences et Techniques du Languedoc. PL Eugene Bataillon, Montpellier Cedex 05 - FRANCE * E-mail:
[email protected] JIN-HO CHOY* and MAN PARK National Nanohybrid Materials Laboratory (NNML) -School of Chemistry and Molecular Engineering - Seoul National University, Seoul, 151-747 - KOREA * E-mail:
[email protected] JUAN CORNEJO*, RAFAEL CELIS, LUCIA COX and M. CARMEN HERMOSIN Instituto de Recursos Naturales y Agrobiologia de Sevilla, CSIC. P.O. Box 1052. 41080 Sevilla - SPAIN. * E-mail:
[email protected] MEHMET SABRI CELIK Istanbul Technical University - Mining Engineering Dept, Mineral Processing Section Ayazaga 34469 Istanbul - TURKEY E-mail:
[email protected] RAY L. FROST * ] and JANOS KRISTOF 2 1 Inorganic Materials Research Program, Queensland University of Technology, GPO Box 2434, Brisbane Queensland 4001 - AUSTRALIA. 2 Department of Analytical Chemistry, University of Veszprem, H8201 Veszprem, PO Box 158-HUNGARY. * E-mail:
[email protected] SIMONE ALBERTAZZI, FRANCESCO BASILE and ANGELO VACCARI* Dipartimento di Chimica Industriale e dei Materiali, Alma Mater Studiorum - Universita di Bologna, INSTM-UdR di Bologna, Viale del Risorgimento 4, 40136 Bologna ITALY. * E-mail:
[email protected]
This page is intentionally left blank
Contents Preface List of contributors Introduction 1 - Chemical Modification of Clay Surfaces Fernando Wypych
v viii
1
Section I - Natural clays: theoretical aspects and applications 2 - Electrokinetic Behavior of Clay Surfaces 57 Mehmet Sabri Qelik 3 - Surface Thermodynamics of Clays 90 B.S. Jai Prakash 4 - Phenomenology of Water Adsorption at Clay Surfaces 118 Jean Marc Douillard and Fabrice Salles 5 - A Worksheet Model for Adsorption/desorption of Ions on Clay Surfaces 153 Giora Rytwo 6 - Raman and Infrared Spectroscopic Studies of kaoUnite Surfaces Modified by Intercalation 184 Ray L. Frost andJanos Kristof I - Nuclear Magnetic Resonance Spectroscopy of Molecules and Ions at Clay Surfaces Jean Grandjean 216 8 - Pesticide-clay interactions and formulations 247 Juan Cornejo, Rafael Celis, Lucia Cox andM.Carmen Hermosin 9 - Pharmaceutical and Cosmetic Applications of Clays 267 Alberto Lopez- Galindo and Cesar Viseras 10 - Removal of Metals by Natural and Modified Clays 290 Cristina Volzone I1 - Catalytic and Adsorption Properties of Modified Clay Surfaces 321 Alexander Moronta Section II - Synthetic clays: Synthesis and applications 12 - Preparation of Layered Double Hydroxides Eiji Kanezaki 13 - Polyoxometalate Complexes of Layered Double Hydroxides Changwen Hu and Danfeng Li 14 - Cationic and anionic clays for biological applications Jin-Ho Choy and Man Park 15 - Environmental Remediation Involving Layered Double Hydroxides Claude Forano 16 - Layered double hydroxide/polymer nanocomposites Fabrice Leroux and Jean-Pierre Besse 17 - Catalytic properties of Hydrotalcite-type Anionic Clays Simone Albertazzi, Francesco Basile andAngelo Vaccari Index
345 374 403 425 459 496
547
This page is intentionally left blank
CHEMICAL MODIFICATION OF CLAY SURFACES FERNANDO WYPYCH Centra de Pesquisas em Quimica Aplicada - CEPESQ. Universidade Federal do Parana - UFPR - Departamento de Quimica CP 19081 - Centra Politecnico - 81531 -990 - Curitiba - PR - BRAZIL. E-mail:
[email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
2
F.Wypych
1 - Introduction Layered materials belong to a special class of compounds in which the crystals are built by the stacking of "two-dimensional" units known as layers that are bond to each other through weak forces [1,2]. Intercalation reactions occur by the topotactic insertion of mobile guest species (neutral molecules, anhydrous or solvated ions) into the accessible crystallographic defined vacant sites located between the layers (interlayer spacings) in the layered host structure. In this intercalation compounds, strong covalent bonds occur in the layers and weak interactions, between host lattice and guest species or co-intercalated solvents. Ionic and solvent exchange reactions are related to the replacement of solvated guest species (cations and anions) located into the interlayer spacings. In this case, only the solvent, the cations or the solvated cation can be replaced, depending on the reaction conditions. Grafting reactions occur by establishing covalent bonds between the reactive groups of the layer and an adequate reactant molecule, which ensures higher chemical, structural and thermal stability for the compound. These reactions can be restricted to the crystal surface (the basal spacing remains unchanged) or layer surface (in this case an interlayer expansion occurs). These compounds can be collectively defined as hybrid materials, or more specifically, surface-modified inorganic layered materials. One of the simplest families in the compounds with layered structures belongs to the alkaline earth or transition metal hydroxides. Common examples involve the structure of brucite (Mg(OH)2) [3,4], gibbsite, bayerite, nordstrandite and doyleite (polymorphic modifications of A1(OH)3), among others. Brucite has the most representative structure that is adopted by several simple hydroxides. In brucite structure [3], atoms of magnesium are octahedrically bonded to six hydroxyl groups, being these units linked to each other through the edges, producing charge neutral two-dimensional layers. Both sides of the layers are covered in hydroxyl groups, being potentially susceptible to be grafted by adequate organic/inorganic molecules, producing grafted or pillared derivatives. Brucite layers are the fundamental building units of a great variety of geologically important hydrous phyllosilicates, including micas, clays and layered double hydroxides (LDH). Figure l(a) displays the structure of brucite, Figure l(b), the schematic representation of a single layer and Figure l(c), the layer top view
Figure 1 - (a) Structure of brucite, (b) the schematic representation of a single layer and (c), the layer top view [5].
3
Chemical Modification of Clay Surfaces
Another class of compounds with a little more complex structures, is the layered hydroxysalts, being typical structures of hydrozincite (Zn5(OH)6(CO3)2) [6], zinc hydroxide nitrate (Zn5(OH)g(NO3)2.2H2O) [7] or copper (II) hydroxide acetate (Cu2(OH)3CH3COOH.H2O) [8]. In these compounds, part of the hydroxyl anions is exchanged by other anions. Water molecules may also be incorporated in the interlayer region for stability. The typical formulation can be written as: Mx+(OH)x_yBn"y/n.zH2O (Mx+ = metal cation and Bn"= anion). Normally, the substituting ion neither needs to have similar hydroxyl ion chemical characteristics nor the same size. In this case, an interlayer expansion is expected in order to accommodate the solvated ion. In hydroxysalts, the process of grafting of specific molecules to the hydroxylated side of the layer (as in the case of the simple hydroxides) and processes of anionic exchange, are also possible. The synthesis of the copper(II) hydroxide acetate is presented in [Eq. 1]. 2 Cu+2 + 3 OH" + CH3COOH" -» Cu2(OH)3CH3eOO.H2O
(Eq. 1)
Controlling the conditions, copper(II) hydroxide or copper(II) hydroxide acetate can be precipitated. Figure 2(a) presents the structure of the zinc hydroxide nitrate, Figure 2(b), the layer top view and Figure l(c), the schematic representation of a single layer.
Figure 2 - (a) Structure of the zinc hydroxide nitrate, (b) the layer top view and (c), the schematic representation of a single layer [5,8]. Other most typical examples of minerals belonging to clay minerals class are hydrotalcite (Mg6Al2(OH)16CO3.4H2O) [9] and pyroaurite (Mg6Fe2(OH)16CO3.4.5H2O) [10], which have the layered double hydroxides (LDH) structure, also known as anionic clays [11-14]. In these materials of variable compositions and mainly of synthetic origin, the layered structure is intimately related to the structure of brucite, where hydroxyl ions are hexagonally close packed and magnesium (or aluminum) ions occupy octahedral sites. Both sides of the layer are covered in hydroxyl groups. LDHs have a generic formulation [M+21.xM+3x(OH)2]x+(Am")x/m.nH2O, where +3 M and M+2 represent metal ions in octahedral sites and Am" represents the interlayer
4
F.Wypych
anion. In these compounds, the trivalent metal substitutes isomorphically a metal in the divalent state of oxidation of the hydroxide structure, generating charges that are compensated by the intercalation of hydrated anions (Eq. 2). LDHs can also be prepared with a single metal in two different oxidation states (Ex.: [Fe+21.xFe+3x(OH)2]x+(Am" )x/m.nH2O) producing the "green rusts". 0,67Mg+2 + 0,33 Al+3 + 2OH" + CO3"2 ^
-> Mg0 67Al0 33(OH)2(C03)o i65.nH2O
(Eq. 2) When another interlayer ion is needed instead of carbonate, the reactions should be performed under protective gas and using boiled and degassed solutions. In the case of using alkaline metal hydroxides as precipitation agent, the excess carbonate should be removed. The salts should preferably contain the anions to be intercalated or an excess of the anion should also be added to the reaction media to compensate the undesirable species (Eq. 2). These hydrated anions are free to move, as they are located in the interlayer spacings. They are exchangeable, being attributed to these compounds the characteristic anionic exchange capacity (AEC). The anions class to be exchanged is wide, going from organic [15], inorganic [16] and even complexes [14,17], with varied oxidation states. Figure 3 (a) presents the structure of the layered double hydroxides and Figure 3(b), the schematic representation of a single layer.
Figure 3 — (a) Structure of the layered double hydroxides and (b) the schematic representation of a single layer [5]. Another class of similar compounds and of natural or synthetic origin involves the clay minerals of the phyllosilicate group [18,19]. In these compounds, generically two structural units are normally involved in the construction of their crystalline lattices. The first is constituted of octahedrons of oxygen atoms and hydroxyl groups at the corners with an aluminum atom in the center or of the gibbsite type. The other is constituted of tetrahedrons with oxygen atoms at the corners and with a silicon atom in the center or of the silica (tridymite) type. If we consider the connection of those isolated units along the plane (octahedrons or tetrahedrons) we will have superposed atomic planes that constitute a sheet. The superposition of the sheets build the layers, the ones which were stacked and separated by the interlayer spacings constitute the structural units [19]. Clay minerals are essentially hydrated crystalline aluminosilicates
Chemical Modification of Clay Surfaces
5
containing several main elements as iron, alkaline and alkaline earth metals. The crystalline lattices are classified as three-dimensional, two-dimensional (layered) or mono-dimensional (fibrous structures) [18,19]. The hydrated aluminosilicates can be neutral or ionic exchangers and the groups treated in this work involve basically the phyllosilicates of the smectite group (type 1:2- derived from the idealized formula of pyrophyllite (Al2Si4Oi0(OH)2 dioctahedral = only 2/3 of the octahedral sites are occupied) and kaolin group (type 1:1). More specifically montmorillonite (typical formulation: 0.33M+(Al167Mgo33)Si4Oio (OH)2) and kaolinite (ideal formulation: Al2Si2O5(OH)4). The composition of the montmorillonites is variable and depends on its own genesis, which is attributed to the characteristic of different cationic exchange capacity. However, in general, it is very superior to that of kaolinite. The low cationic exchange capacity of kaolinite is justified by the low isomorphic structural substitution (usually Al+3 for Fe+2 or Fe+3). In other words, its composition can be considered fixed. The nomenclature 1:2 and 1:1 refers to the construction of the layers: in the case of the smectites, the layer is built by two sheets of silicon atoms tetrahedrically bonded to oxygen atoms that involve a sheet of aluminum atoms octahedrically bonded to oxygen atoms and hydroxyl groups. Both sides of the layer expose planes of oxygen atoms (siloxane surface), having distorted hexagonal cavity formed by six-corner sharing silica tetrahedron (Fig. 4(c)). The low interaction between the layers provides the material easy cleavage and consequently, anisotropic properties. In nature, aluminum atoms are isomorphically replaced by atoms of lower oxidation states, thus producing negatively charged layers, in higher or lower degree. The negative layers are charge balanced through the process of intercalation of hydrated cations in interlayer vacant sites, which is being attributed to these clay minerals for their high cationic exchange capacity characteristics (measured in meq/lOOg). Figure 4(a) presents the structure of montmorillonite, Figure 4(b), the schematic representation of a single layer and Figure 4(c), the silicate sheet top view.
Figure 4 — (a) Structure of the montmorillonite, (b) the schematic representation of a single layer and (c), the silicate sheet top view [5].
6
F.Wypych
In kaolinite, the layers are built of only a sheet of silicon, tetrahedrically bonded to oxygen atoms and a sheet of aluminum octahedrically bonded to oxygen atoms and hydroxyl groups. As a consequence, the aluminum side of the layer is covered in hydroxyl groups (aluminol) and one third of the octahedrons are vacant in order to maintain a neutral sheet (dioctahedral) (Fig. 5(c)). The silicon side is covered in oxygen atoms (siloxane surface) and the same distorted hexagonal cavity present in the 1:2 group, is observed (Fig. 4(c)). These characteristics turn kaolinite structure into an unique matrix in which the confined molecule will be subjected to an asymmetric chemical environment, producing materials with interesting properties suitable for applications in the non-linear optics as recently reported [20,21]. Adjacent layers of kaolinite are linked to each other by hydrogen bonds, which involve the aluminol and siloxane groups. A high cohesion between the layers results from this type of bonding that hinders intercalation, grafting and exfoliation reactions. Although kaolinite has a fixed composition, depending on its genesis, different degrees of crystallinities can be obtained (low, medium and high) with decisive factors in its chemical reactivity. Figure 5(a) presents the kaolinite structure, Figure 5(b), the schematic representation of a single layer and Figure 5(c), the hydroxide sheet top view.
Figure 5 — (a) Structure of kaolinite, (b) the schematic representation of a single layer and (c), the hydroxide sheet top view [5]. The surface reactions and the corresponding surface complexes play a fundamental role in the behavior and properties of the materials with layered structures. A specific example can be found in the acid catalytic properties of the clay minerals, which can promote processes of polymerization of organic matter residues of the animal/vegetable origin, contributing to the humification processes [22]. Being clay minerals the main components of the colloidal fraction of soils, the above mechanism promotes the formation of organo-mineral compounds, which have a large capacity of ionic exchange and the ability of complexing metals and nutrients. Thus, this system becomes important for the agricultural activities, mainly in tropical countries [23]. Another interesting characteristic of the clays in soils is the possible modification (or destruction) of pesticides through oxidation/reduction reactions, being isomorphic substituting iron the key factor for these mechanisms.
Chemical Modification of Clay Surfaces
7
One of the most important examples of a layered material application involves the structure of molybdenum disulfide doped with metals (normally Ni or Co), as an hydrotreatment catalyst. Although these materials have been used for many decades, one of the main doubts have been persisting for years is related to the positioning of the dopant in the structure of the sulfide. Only recently, results have demonstrated that the active phase in this catalyst is generated when the dopant is bonded to the layered crystal edges producing a phase of the type Ni(or Co)-Mo-S [24,25]. As this phase is preferably generated on the edges of the crystals, it is difficult to characterize and still lead to divergent views among the researchers. Being predominantly surface processes, the understanding of these phenomena in compounds with layered structures is of fundamental importance for several branches of science, mainly those devoted to soils science; environment pollution control, special materials, catalysts design are among others. Due to the similar structural characteristics of the five groups of layered compounds described above, the intercalation and grafting processes are similar, differing basically in the particularities of the layers' surfaces in each system and accessibility of the reactants to the interlayer spacings. We start in the following sections, the description of generic and specific reactions involved for each group of layered materials mentioned above. 2 - Simple hydroxides derivatives The reactions with hydroxides are quite limited, mainly based on the fact that a great difficulty exists in penetrating in the interlayer spacings. Although the layers are linked to each other through weak forces, only small molecules can potentially be inserted and grafted to the layers. The five potentially possible reactions in simple hydroxides are related with the process of surface adsorption (basal and edge planes), substitution of the hydroxyl groups for other anions transforming the materials in hydroxysalts and processes of grafting of the interlayer (and surface) hydroxyl groups as well as the processes of interlayer hydroxyl groups' solvation and intralayer metal oxidation/reduction reactions. Only a few examples are reported in the literature about the reactions described above [26-29]. In the case of gibbsite, which has octahedral vacant sites in the layer (Figure 5(c)), reactions with lithium salts producing compounds similar to the layered double hydroxides were also reported [30,31]. This reaction is very rare as both cation and anion are simultaneously intercalated into the layered structure, being the layer octahedral vacant site directly involved in the process. 2.1 - Surface adsorption Based on the fact that the unit cell is electrically neutral, the simplest reactions for the hydroxides involve the processes of reversible surface adsorption, through neutral or charged organic (org.) and inorganic (inorg.) molecules (Eq. 3) Mg(OH)2 + x inorg/org <-> Mg(OH)2(inorg/org)x
(Eq. 3)
As the adsorbed species are located mainly in the unsaturated bonds present on the edges and basal defects of the crystallites as single (or multiple) layers, and does not develop a three dimensional structure, the structure of the original hydroxide is retained. Depending on the type and energy of interaction involved, chemical or physical processes can be considered and the surface complexes can acquire important characteristics depending on the involved constituents. When a single layer of a specific
8
F.Wypych
dye can be adsorbed preferably on the edges or basal surface of the crystals, a special procedure for the surface area determination can be developed. 2.2 - Exchange of hydroxyl ions Potentially the hydroxyl anions bonded to the layer can be exchanged for other anions of the same charge, where non-stoichiometric compounds similar to the hydroxysalts can be produced (Eq. 4). If the exchanged anions are of different charges, compositions similar to the layered double hydroxides (although being cationic exchangers) can also potentially be obtained. (Eq.5) Mg(OH)2 + x CH3COO" -> Mg(OH)2_x(CH3COO)x + x OH" (Eq. 4) Mg(OH)2 + x CO3-2 H> [Mg(OH)2.x(CO3)x]-1 + x OH" (Eq. 5) These types of compounds are obtained easily when a nickel salt is precipitated with sodium hydroxide. If the conditions are not very well established, besides Ni(OH)2 (a phase) the non-stoichiometric (3 phases can also be observed (Ni(OH)2.x(An" X/n.yH2O; An"= NO 3 \ Cl", SO4"2, CO3"2) [32]. Obviously, if the anion charge is larger than 1, the layers are negatively charged and cations must be intercalated between the layers (See Section 3). 2.3 - Interlayer (or surface) grafting The process occurs through the grafting of the layer surfaces with organic or inorganic molecules, in the same way as in glass, silica or other materials with modified surfaces, used mainly in chromatographic purposes (Eq. 6). The compounds display stronger interactions (through covalent bonds) which ensure higher chemical, thermal and structural stability. When both sides of the layer can be grafted and an appropriate molecule is used (with two reactive terminal functions), organic pillared compounds can be prepared. (Eq. 7). Mg(OH)2 + x CH3-OH -> Mg(OH)2.x(O-CH3)x + x H2O
(Eq. 6)
Mg(OH)2 + x/2 HO-(CH2)2-OH -> Mg(OH)2.x(O-(CH2)2-O)x/2 + x H2O (Eq. 7) In this case it is important to emphasize that only small molecules of appropriate acidity that can have access to the interlayer spacings and consequently, to the hydroxyl ions, can be grafted. The resultant non-stoichiometric compounds have organic functions in an inorganic layer matrix. This type of compound has unique characteristics since the grafting process will preserve the structure of the layered compound, being obtained only through this kind of procedure. By controlling the size, the distribution, and the nature of the organic pillars, interesting materials with special properties can be engineered and synthesized. 2.3.1 - Grafting of ethylene glycol and glycerol into brucite [33] In this case the "esterification" or "acid/base" reaction with ethylene glycol is presented (Eq. 7). The reaction concentrates on the insertion of ethylene glycol molecules between the layers and the reaction with the interlayer hydroxyl groups. In
Chemical Modification of Clay Surfaces
9
the case of the grafting reaction we should imagine that similar processes could happen in unsaturated bonds on the edges or basal surface defects of the crystals. The X-ray powder diffraction patterns of brucite and the reacted composites are shown in Figure 6. Brucite has a basal spacing of 4.78A, in perfect agreement with the literature value (Fig. 6(a)) [3]. After the reaction with ethylene glycol, a new phase was obtained with a basal spacing of 8.3A (Mg-EG), as shown in Figure 6(b). Only one broad basal diffraction peak was observed, showing that this phase has low crystallinity with low stacking order. The increase of the basal spacing was of 3.5 A compared to that of the brucite. Small diffraction peaks of brucite were still observed even after reaction times of 72 hours. In the reaction of brucite with ethylene glycol for 24 hours (Fig. 6(b)), the concentration of brucite is higher. This can be observed in Figure 6(c) where the reaction time was increased to 72 hours.
Figure 6 - X-ray powder diffraction patterns (a) of brucite and (b) the ethylene glycol derivative obtained after reaction for 24 hours (Mg-EG), (c) 72 hours and (d) the glycerol derivative (Mg-GL). Powder of silicon was used as internal standard (*). [Reprinted by kind permission ofJ. Coll. Interface Sci., (253, 180, 2002)] [33]. The TG/DSC curves can be observed in Figure 7. In brucite (Fig. 7(a)), an initial mass loss of 3.1% is associated with the elimination of absorbed/adsorbed water molecules (weak endothermic peak centered at 92°C) followed by a large endothermic peak centered at 393°C, which is attributed to the dehydroxylation of the layered structure, producing MgO. The mass loss of 31.5% after water evaporation due to dehydroxylation is very close to the theoretically expected value. The final product was additionally characterized by XRD (not shown) and consists of MgO crystals with the periclase structure, as expected [34]. The Mg-EG sample (Fig. 7(b)) shows two endothermic peaks at 67 and 139°C, attributed to elimination of water molecules, which amounts to a 19.7% mass loss between room temperature and 160°C. The exothermic peaks at 204, 371 and 403°C are collectively attributed to the combustion of organic matter. The 50.9% mass loss between 160 and 1000°C leads us to the following empirical formula from the experimental data: Mg(C2H402)o,95(OH)o,o5.1,12H20 (mixture of 95% Mg(C2H4O2) and 5% of Mg(OH)2). The Mg-GL phase (Fig. 7(c)) presented a slightly different profile,
10
F.Wypych
where only one endothermic peak at 70°C was observed. A 9.4% mass loss was found up to 160°C, attributed to water removal. Three exothermic peaks centered at 272, 364 and 420°C, are also clearly associated to the combustion of organic material combustion. Here we find an empirical formula of Mg(C3H6O3).0,66H2O considering the 67.2% experimental mass loss above 160°C for the Mg-GL original sample.
Figure 7 - TG/DSC measurements of (a) brucite, (b) Mg-EG and (c) Mg-GL grafted phases. [Reprinted with kind permission of J. Coll. Interface ScL, (253, 180, 2002)] [33]. Figure 8 present the FTIR spectra of (a) brucite, (b) Mg-EG and (c) neat ethylene glycol. Brucite displays the typical hydroxyl stretch bands at 3698 and 3643cm"1 and an extended band centered at 3428cm"1 relative to hydroxyl stretching of water molecules in various states [35]. We note a small shoulder at 3275cm"1, which corresponds to strongly bonded water. Small bands were also observed at 1637cm"1 (surface adsorbed/co-intercalated water), 1425 and 1118cm"1. The Mg-EG sample shows a small brucite contamination
Chemical Modification of Clay Surfaces
11
(band at 3698cm" ) as already detected in the X-ray powder diffraction patterns. The 3700 and 2700cm"1 region of the FTIR spectra provides interesting information about the structure of the brucite grafted derivatives. Two important bands were found within these spectral regions: the out-of-plane stretching vibrations of C-H bonds at 31002700cm"1 and the stretching vibrations of O-H groups at 3700-3100cm"1. After the covalent grafting of ethylene glycol into brucite, two C-H stretching bands of ethylene glycol, originally centered at 2946cm"1 (antisymmetric) and 2879 (symmetric) were either displaced or converted to at least four new absorption bands at 2954, 2919, 2852 and 2704cm'1.
Figure 8 - FT1R spectra of (a) brucite, (b) Mg-EG and (c) neat ethylene glycol. [Reprinted by kind permission ofJ. Coll. Interface Sci., (253, 180, 2002)] [33], This FTIR observation confirms the successful insertion of ethylene glycol into the layer structure of the host matrix, because their C-H groups are now vibrating in a distinct chemical environment. Two small bands observed at 1325cm"1 and 1359cm"1, which may suggest that oxyethylene units (O-CH2-CH2-O) are in trans conformation [36,37]. The ISOO-nOOcm"1 spectral region of the grafted material reveals also a series of peaks with low intensity, collectively attributed to CH2 twisting, wagging and scissoring vibrations, in ethylene glycol [38]. The absorption bands at 1030-1100cm"1, typically attributed to Al-O-C and CC-0 bonds in kaolinite and 1100cm"1 in barium aluminate glycolate [39], have been observed at 1040, 1076, 1109 and 113 lcm"1 for Mg-EG. Similar bands were observed in 1043, 1072 and 1081cm"1 in the compound obtained by the grafting of ethylene glycol in layered double hydroxide [40]. Rocking vibrations of the CH2 groups and C-C stretching vibrations, generally centered at 865 and 882cm"1 for ethylene glycol, were observed at 858 and 880cm"1. A similar effect over the CH2 and C-C vibrations was also observed when boehmite was used as the host matrix, where the original bands were displaced to 868 e 900cm"1 after the monodentade grafting of ethylene glycol [41]. Therefore, larger shifts to higher frequencies are expected when ethylene glycol is grafted into the host matrix under a bidentade conformation [41]. The absorption bands at 447 and 558cm"1, which can be assigned to Mg-0 lattice vibrations in brucite now,
12
F.Wypych
appear enlarged at 561, 608 and 655cm"1. In the case of the functionalization of brucite with glycerol (Fig. 9), the spectra are very similar mainly based on the similarity of the structures of ethylene glycol and glycerol.
Figure 9 - FTIR spectra of (a) brucite, (b) Mg-GL and (c) neat glycerol. [Reprinted by kind permission ofJ. Coll. Interface Sci., (253, 180, 2002)] [33]. 2.4 - Solvation reaction of the interlayer hydroxyl groups Potentially, after breaking the weak forces that maintain the layers together, the interlayer hydroxyl groups can be solvated with small polar molecules. In this case experimental evidences of this reaction type do not exist. However, similar results were accomplished in the process of interlayer groups' solvation of the kaolinite structure [42-46] (described in section 5.1.) (Eq. 8). Mg(OH)2 + x solv. <-> Mg(OH)2(solv.)x
(Eq.8)
2.5 - Oxidation-reduction reactions If we consider that layered double hydroxides can be obtained with the same metal in two oxidation states, this type of compounds can be synthesized starting from the oxidation of iron(II) hydroxides (Eq. 9). These processes are potentially reversible. This reaction (or alternatively for the co-precipitation of Fe+2/Fe+3 salts in alkaline pH conditions) originates the family of the green rusts [47-50]. These rusts are very important materials in the solubility and transport of iron in soils and underground water, inactivation of pesticides and maybe in pollution control due to its oxidation/reduction potential [50]. The reduction reaction of a +3 metal can potentially generate cationic exchangers, being those processes also reversible (Eq. 10). It is important to emphasize that this kind of reaction is only possible if the metals in the two oxidation states have compatible sizes and can occupy sites of identical geometry and coordination numbers. The diameter of Fe+2 being larger than Fe+3, the replacement of Fe+2 for Fe+3 is expected ( Eq. 9). However, the opposite reaction is quite unlikely (Eq. 10). Fe(OH)2 - xe + A" « [Fe+3xFe+2,.x(OH)2]x+(A"n)x/n.yH2O
(Eq.9)
Chemical Modification of Clay Surfaces
Fe(OH) 3 + xe" + B + " » [Fe +2 x Fe +3 ,. x (OH) 2 ] x -(B +n ) x/n .yH 2 O
13
(Eq. 10)
2.6 - Reaction of Gibbsite with lithium chloride In the structure of gibbsite only 2/3 of the layer octahedrons are occupied by Al+3 ions, being the structure classified as dioctahedral (Fig. 5(c)). As the third position of the aluminum ion in the layer is empty, it can be filled out with a alkaline metal cation with reduced dimensions leaving the layer positively charged, which is balanced with a hydrated anion intercalation (Eq. 11) [30,31]. 2 A1(OH)3 + LiCl -> [LiAl2(OH)6]Cl.nH2O
(Eq. 11)
In this case the intercalated ions can potentially be exchanged as described in the session 3.1 for the hydroxysalts and 4.2 for the layered double hydroxides. This reaction type is rare and, as far as we know, it has been described only for the derivatives of aluminum hydroxide with lithium salts. The size and the stereochemistry of the cavity in the gibbsite layer can explain this limitation. This kind of reaction can also potentially be applied to the kaolinite structure, where a hexagonal siloxane cavity and empty octahedral sites can be found in both sites of the building layers. The product of the reaction similar to layered double hydroxides can be obtained with well-ordered structure. Computer calculations can be performed to find the correct position of the intercalated molecules or ion. This procedure is important when good quality single crystals are not available and the powder X-ray powder diffraction patterns are of poor quality. 3 - Hydroxysalts derivatives The hydroxysalts reactions supply new alternatives, mainly for the previous expansion of the layers for ions of larger dimensions, located in the crystallographic positions of the hydroxyl ions [7,8,51-53]. The reactions can be classified as surface adsorption [similar to the case of the simple hydroxides (Eq. 3), ionic exchange reactions, processes of replacement of the hydroxyl ions (Eq. 04, 05) and probable simultaneous ionic exchange and grafting reactions (Eq. 6). Exfoliation reactions (also this kind of reaction can be potentially applied to other layered structure compounds) can be performed and new classes of materials can be obtained as the nanocomposites involving polymers. In this class of compounds, only the reactions of ionic exchange have been described in the literature. However, a large research field still remains to be explored, which may produce very interesting materials. 3.1 - Ionic and solvent exchange reactions The ionic exchange reactions take place, when the solvated interlayer ions are replaced (Eq. 12). Cu2(OH)3CH3COO.H2O + NO3" - Cu2(OH)3NO3.H2O + CH3COC>- (Eq. 12) Depending on the used solvent, not only the replacement of the cation but also of the solvent can take place (Eq. 13,14). The solvent can be any polar molecule, including polymers of synthetic or natural origin. The mechanism involved in the exchange reactions occur with the shift of the basal spacing to higher values, being regulated by the size of the ion (in anhydrous conditions) or by the solvation energy of
14
F.Wypych
the ion. Normally the size of the solvent is bigger than the one of the ion and determines the value of the basal spacings. Cu2(OH)3CH3COO.H2O + CH3-OH -> Cu2(OH)3CH3COO.CH3-OH + H2O (Eq. 13) Cu2(OH)3CH3COO.H2O + NO3" + CH3-OH -> Cu2(OH)3NO3.CH3-OH + CH3COO" + H2O (Eq. 14) Depending on the time and the concentration of the ions in the solution, an equilibrium can be established where cations and solvents mixtures will be present in the final composition. The ionic exchange capacity of a compound is measured in meq/lOOg of the material. 3.2 - Hydroxyl ions substitution + ionic exchange reactions Depending on the reaction conditions, not only the ionic exchange as well as reactions of hydroxyl replacement for other ions can take place simultaneously. This reaction can also include the replacement of solvents (not shown, however similar to the one described in the ionic exchange reactions) (Eq. 15). These possibilities have shown that exchange reaction can be very complex, effects that are rarely described in the specialized literature. Cu2(OH)3CH3COO + 1+y NO3" - Cu2(OH)3.y(NO3)yNO3 + yOH" + CH3COO" (Eq. 15) 3.3 - Grafting reaction [54-57] Although the grafting reaction with simple hydroxides is difficult to happen, the same kind of reactions with the hydroxysalts is very much facilitated, mainly based on the fact that the layers had already been moved away previously by the presence of the exchangeable ion (Eq. 16). The possibility that the exchangeable ion can be simultaneously replaced during the grafting reaction should also be considered (Eq. 17). Cu2(OH)3CH3COO + y CH3-OH -> Cu2(OH)3_y(CH3-O)yCH3COO + y H2O (Eq. 16) Cu2(OH)3CH3COO + 2 CH3-OH -+ Cu2(OH)2(CH3-O)CH3O + H2O + CH3COOH (Eq. 17) 3.3.1 - Grafting of copper(II) hydroxide acetate with benzoic acid [54] The hydrated copper(ll) hydroxide acetate (Cu2(OH)3CH3COO.H2O) was reacted with benzoic acid in water, under magnetic stirrer agitation at 60° C for 36 hours. Figure 10 shows the X-ray powder diffraction patterns for the described samples. Pure copper (II) hydroxide acetate (Fig. 10(a)) is clearly layered with basal plane diffraction peaks and their multiples corresponding to basal spacing of 9.3 A. The reacted material, independent from the reacted amount shows a basal spacing of 15.6A. The line width of the X-ray powder diffraction patterns shows that the crystallinity of the reacted compounds reduces as more benzoic acid is incorporated between the layers. The stoichiometric 1/0.75 ratio (Fig. 10(c)) produced the highest quality crystals.
Chemical Modification of Clay Surfaces
15
Figure 10 — X-ray powder diffraction patterns (a) for copper (II) hydroxide acetate and CuOHAc reacted with benzole acid in the proportions: (b) 1/0.5; (c) 1/0.75; (d) 1/1 and (e) 1/4. *= Si internal standard. Reprinted by kind permission of [J. Coll. Interface Set, (240, 245, 2001)] [54]. Figure 11 shows the TG/DSC measurements for copper (II) hydroxide acetate (Fig. 1 l(a)) and the reacted materials ((b) 1/0.75 and (c) 1/4). In pure hydroxide acetate, two endothermic peaks observed at 128 and 188°C can be attributed to water elimination and fragmentation followed by oxidation, respectively. The 7.2% mass loss up to 142°C due to water removal and the 33% mass loss up to 1000°C leads to the Cu2(OH)3(CH3CO2).(H2O)1>03 stoichiometry. The theoretical values for the Cu2(OH)3(CH3CO2).(H2O)ii(, formula would be of 7.06% water content and 32.92%, considering CuO as the final oxidation product of this experiment. Figure ll(b) shows the TG/DSC results for the 1/0.75 stoichiometry reacted material with a more complex profile. The 0.5% moisture is eliminated up to 100°C. Three endothermic peaks are observed at 180, 208 and 220°C. These peaks are attributed to fragmentation and oxidation steps respectively. The three exothermic peaks at 275, 295 and 315°C are attributed to organic matter oxidation. At 322°C part of the material is ejected from the crucible. Considering the mass loss of 54.9% up to 322°C we obtain a stoichiometry for the reacted material of Cu2(OH)2j4(C6H5CO2)i,6 while the predicted theoretical stoichiometry, considering the reaction proportions, would be Cu2(OH)3 25(C6H5C02)o,75. Considering the % stoichiometry material, whose thermal decomposition and reaction behavior is displayed in Figure ll(c), we observe a broad endothermic peak centered at 232°C followed by three exothermic peaks at 254, 299 and 313°C respectively. The mass losses are of 0.3% up to 100°C and 65.8% from 100 to 325CC. The 2% mass gain from 325 to 1000°C is related to copper oxidation. The calculated stoichiometry of this material would be Cu2(OH)! 6(C6H5CO2)2i4, while the theoretical expected stoichiometry for this reacted material should be Cu(C6H5C02)2,oThe general findings of these thermal experiments are that the ideal organic functionalization occurs for the 1/4 proportion of reaction, where clearly an excess of benzoic acid is present while the maximum incorporation never exceeds 60%, even after
16
F.Wypych
36h of reaction time. We synthesized copper benzoate in hydrated and anhydrous form and none of these compounds could be identified in the grafted samples.
Figure 11 - TG/DSC measurements for (a) copper (II) hydroxide acetate and the reacted materials ((b) 1/0.75 and (c) 1/4). [Reprinted by kind permission of J. Coll. Interface ScL, (240, 245, 2001)] [54]. Figure 12 shows the FTIR spectra for the different reacted materials in comparison to (a) pure copper (II) hydroxide acetate and (f) pure benzoic acid. For comparison we also show FTIR spectra for the 1/4 proportion of sample (Fig. 12(e)) as well as for benzoic acid (Fig. 12(f)), obtained by pressing the material with KBr powder (circa 1% of material relative to KBr). Since the spectra for the reacted materials are very similar, we concentrate our discussion on the 1/2 proportion of reaction (Fig. 12(d)). The region between 3000 and 3600cm'1, which is attributed to hydroxyl groups' vibrations, demonstrates that water molecules are absent in the grafted materials and that hydroxyl groups are still present in the samples even after functionalization by excess benzoic acid. This interpretation is
Chemical Modification of Clay Surfaces
17
in accordance with the proposed stoichiometries and partial grafting reaction. Low intensity bands observed at 3609, 3576, 3520, 3449, 3403 and 3263cm"1 attributed to water and matrix hydroxyl groups have been substituted by very well defined absorption bands at 3613, 3601 and 3586cm"1. In the grafted phase, the bands can collectively be attributed to free hydroxide vibrations. Absorption peaks at 3089, 3066, 3056, 3025 and 2932cm"1 demonstrates that the organic molecule has been added to the host due to the characteristic C-H vibrations.
Figure 12 - FTIR spectra for (a) copper (II) hydroxide acetate, (f) benzoic acid and their reacted products in the reaction proportions: (b) 1/0.75; (c) 1/1; (d) 1/2 and (e) 1/4. [Reprinted with kind permission ofJ. Coll. Interface Sci., (240, 245, 2001)] [54]. The absorption bands in the 3000-3100cm"1 region grow in intensity in correlation with benzoic acid proportion. These bands are practically absent in CuOHAc. The large benzoic acid band in this region does not contribute to the grafted samples. The IR absorption in the 1200 and 400 cm"1 region of the functionalized samples differs markedly from pure CuOHAc. In this region benzoic acid group absorption is relevant. The most important bands in the functionalized material occur at 683, 703, 871, 928, 1029, 1066 and 1310cm"1 (667, 685, 709, 936, 1006, 1026, 1073, 1101, 1127, 1182cm"1 in benzoic acid) and 651, 799, 947, 1022 and 1047cm"1 in copper (II) hydroxide acetate. The most interesting characteristic of the IR spectra in this region is the maintenance of absorption bands of carboxyl groups. Peaks observed at 1409cm"1 attributed to symmetric C=O vibrations and 1549cm"1 attributed to asymmetric C=O in CuOHAc are seen at 1404, 1429, 1551 and 1595cm"1. Since the two bands in the original compound transform to four bands after functionalization, it is not clear which type of grafting reaction occurs for the carboxylate groups to the layers [54,57]. The unidentate grafting probably dominates the process, but simple intercalation and bidentate grafting cannot be excluded [58]. The absorption band relative to the asymmetric C=O vibration is more affected by the chemical environment and to ligand fields than the symmetric band, an effect clearly confirmed here [59]. But the absorption band of the C=O symmetric vibration also shows a splitting into two new bands. The band at 1429cm"1 grows proportionally to the benzoic acid content in the
18
F.Wypych
reacting medium. We observe that, while the use of KBr pellets is a common procedure to prepare samples for IR analyses, that this preparation in our case has led to moisture absorption (and other unwanted reactions), masking important features of the spectra in the range from 3000 to 3500cm"1 (Fig. 12(e)). The spectral region between 2000 and 500cm"1 show only small effects besides slight dislocations and intensity shifts. Figure 13 shows the SEM micrographs of the samples starting with pure CuOHAc.
(d)
Figure 13 - Scanning electron microscopy micrographs for (a) copper (II) hydroxide acetate and CuOHAc reacted with benzoic acid in the proportions: (b) 1/0.75, (c) 1/4 and (d) 1/2 reacted for 96h. [Reprinted with kind permission ofJ. Coll. Interface Sci., (240, 245, 2001)] [54]. Bar = 2^im. We observe that the original hydroxide acetate (Fig. 13 (a)), is composed of platelets of approximately 5um diameter. Upon grafting with benzoic acid (1/0.75 proportion), the highly crystalline platelets are degraded (Fig. 13(b)), while the higher addition of benzoic acid (1/4 proportion - Fig. 13(c)) leads to a fibrous compound. For the 1/2 proportion (Fig. 13 (d)) we have also tried a longer reaction time of 96 h, and observed a complete separation of the sample into submicrometric fibers of about 5000A diameter. This longer reaction time has shown otherwise no effects on all the physicochemical characterizations. Several visual evidences in the micrographs point to the fact that the pellets are fragmented according to preferential easy directions, originating from the fibers.
Chemical Modification of Clay Surfaces
19
3.4 - Exfoliation reactions Exfoliation reactions refer to the process of separation of individual layers in an appropriate solvent [60-66]. Literally, it is a process of rupture of a layered crystal in such a way that stacked single layers are removed of the crystal and taken to suspension (Eq. 18). Cu2(OH)3CH3COO -> Cu2(OH)3CH3COO (single layers)
(Eq. 18)
Potentially, the process of separation of individual layers can be used for reactions of direct functionalization since in those conditions, the bonds do not exist among the layers that hinder the access of the reactants to the interlayer spacings (Eq. 6,16,17). Similarly, this can be used in the transition metal dichalcogenides coprecipitation of polymers (charged or not) or other species to the surfaces of the layers followed by the rearrangement of the layers in a crystal [67-70]. That can lead to a range of interesting materials, from a technological point of view, including lightemitting diodes and chemical sensors (Eq. 19,20). Cu2(OH)3CH3COO(single layers) Cu2(OH)3CH3COO(polymer)x (Eq. 19)
+
x
polymer
-»
Cu2(OH)3CH3COO(single layers) + x polymer" -> Cu2(OH)3(CH3COO)!.x (polymer)x + x CH3COO" (Eq. 20) During the process of restacking of the layers, they are free to rearrange in a way to minimize the energy (being influenced by the adsorbed molecules). This can produce new phases, which are normally impossible to be prepared for the conventional intercalation reactions. The nanocomposites are singular material, in which, at least one of the phases is in the nanometer scale range. In the case of layered materials as only one dimension of the layer is in the nanometer range (perpendicular to the layer), the materials are denominated as polymer-layered crystal nanocomposites [71]. Depending on the chosen system, each component can contribute to the properties in a synergistic way, being produced materials with special properties. Although in this system the matrix is a hydroxysalt, after the intercalation process, the nanocomposite can be incorporated in a polymer. The nanocomposite can facilitate the interaction and the compatibilization of the filler in the polymer, giving the material any desired property. In that specific case, we can mention the possibility of incorporation single layers of clay minerals as fillers in the reinforcement of polymers (from synthetic or natural origin), resins or rubbers of special uses [71-76]. Alternatively, those suspensions of single layers (or the nanocomposites) can be used in the production of thin films with similar characteristics to those of the Langmuir-Blodgett films and with potential applications such as: sensors, electric devices, materials with enhanced mechanical properties, resistant to fire, etc. Special attention should be focused on the production of interstratified thin films obtained by restacking mixtures of single layers with opposite charges (or even with the same charge). An interesting example is the mineral lithiophorite (Al2LiMn+2o.5Mn+42.506(OH)6 = [Al2Li(OH)6][Mn+20.5Mn+425O6], obtained by alternate positive layers of aluminum/lithium hydroxide and negative manganese oxide layers [19] (fig. 14) (ideal formulation Al 2 Mn 3 0 9 .3H 2 0 = 2Al(OH)3.3MnO2).
20
F.Wypych
Figure 14 - (a) Structure of lithiophorite and (b) the schematic representation of the corresponding single layers [5]. 4 - Layered double hydroxide derivatives Layered double hydroxides react similarly to the hydroxysalts (Eq. 12-20). However, two metals with different oxidation states can produce interesting materials in several branches of chemistry, physics and engineering [11-17]. If only one metal with two different oxidation states is involved, similar reactions proposed for the simple hydroxides can be observed (Eq. 9,10). In these systems, surface adsorption, ionic exchange reactions (involving the exchange of solvents or not), replacement of hydroxyls simultaneously with ionic exchange reactions, grafting and exfoliation reactions (nanocomposites obtaining) are perfectly possible. Another reaction, still not explored, refers to the process of solvation of the hydroxyl groups, as reported in the case of kaolinite (section 5.1). 4.1 — Surface adsorption reactions Based on the fact that the unit cell is positively charged, the processes of surface adsorption of charged species involve the ionic interactions as well as the charge residues in the crystal edges, similarly to those described in the case of the simple hydroxides (Eq. 3) and hydroxysalts [51-53]. These extensively studied processes, involve the first layer of adsorption quite organized (usually constituted of hydrated anions), followed by other less organized layers until the interaction with the external aqueous solution happens. 4.2 - Exchange + grafting reactions 4.2.1 - Grafting of layered double hydroxide (LDH) with ethylene glycol [40] The layered double hydroxide with the nominal composition Zno.66Alo.34(OH)2(C03)o.i7.nH20 was dispersed in 15cm3 of ethylene glycol (Merck) or glycerol, in a 50 cm3 flat bottomed reaction flask connected to a reflux condenser. The reaction mixture was heated up to 80°C and kept under magnetic stirrer agitation for 5 days. The X-ray powder diffraction patterns of both phases, (a) Zn-Al-CO3 and (b) Zn-
Chemical Modification of Clay Surfaces
21
Al-EG, are shown in Figure 14. The Zn-Al-CO3 compound showed a basal spacing of 7.78A and this data was in perfect agreement with that found in the literature. This typical basal spacing was readily attributed to the summation of the basal lattice parameters of brucite (4.78A) [3] and the diameter of the intercalated anion (CO3"2, 3A)[11]. After reaction of the Zn-Al-CO3 phase with ethylene glycol, a new pure phase was generated with a basal lattice parameter of 9.78A (Zn-Al-EG). The main reflections of these phases, labeled as "Hn" and "En" in Figure 14, were respectively associated with the sequence of basal reflections of pure Zn-Al-CO3 and Zn-Al-EG, where "n" is integral number. The "n" values could not be used to represent normal indexing because structural transitions of the original hydroxide could not be clearly excluded after intercalation. This was due to the tendency of the layered crystallites in getting organized on the surface of the glass sample-holder, therefore intensifying the basal reflections from the background (in contrast to other existing signals). The basal spacing, calculated in relation to the basal reflection of highest order, was normally 6 or 7 depending on the type of material under analysis. Variations in basal spacing were obtained by subtracting the basal spacing of the intercalated LDH derivative (9.78A) from that of both Zn-Al-CO3 (7.78A) and pure brucite (4.78A). Compared to brucite, the variation of 5.0A in the basal spacing of the Zn-Al-EG compound was consistent with the establishment of either two EG loops on the surface of each of the adjacent layers (resulting in two opposite bidentade forms in cis) or one EG bridge between two adjacent layers (bidentade form in trans linking both layers). As both opposite layer surfaces are susceptible to grafting and the grafted molecule contains two carbon atoms, grafting between two layers forming a bridge provides an equivalent basal expansion as the independent double-grafting of each layer (EG loop).
Figure 14 - X-ray powder diffraction patterns of (a) the Zn-Al-CO3 phase before and (b) after reaction with ethylene glycol. Powdered silicon, identified by an asterisk (*), was used as the internal standard. [Reprinted with kind permission ofj. Coll. Interface Sci., (227, 445, 2000)] [40]. Theoretically, the latter case {trans conformation) seems to be preferable over the former case (cis conformation), since the independent grafting of each layer would require a perfect and conserved orientation and/or spacing of the two hydroxyl groups
22
F.Wypych
involved in grafting. Of course, being a topotactic reaction, it is perfectly conceivable that both linkage types are present in the resulting grafted Zn-Al-EG derivative. When kaolinite was grafted with ethylene glycol, two distinct basal expansions of 2.2A and 3.6A were observed [77,78]. Since these variations are considerably smaller than 4.2A, which is the nominal diameter of the isolated ethylene glycol molecule [41,79], the former seems to be related to the direct grafting of ethylene glycol (Ac= 2.2A), whereas the latter would indicate its simple intercalation within the layer structure (Ac= 3.6A)[77,78]. Variations lower than the molecular diameter of ethylene glycol are justified by the interpenetration or keying of the grafted molecules into the siloxane hexagonal cavity (Fig. 4(c)). As reported in the literature, grafting of boehmite with ethylene glycol generated a single basal expansion of 5.5A [41,80,81] of the starting material. This variation was attributed to the double-layer monodentade grafting of ethylene glycol, in which the remaining unreacted hydroxyl groups contributed to the larger expansion in the basal spacing [79,80], compared to the isolated molecule (4.2A in size). The TG/DSC measurements of both (a) Zn-Al-CO3 and (b) Zn-Al-EG phases are shown in Figure 15. The temperatures indicated there were determined at every minimum and maximum of the DSC profiles. In the Zn-Al-CO3 phase (Fig. 15(a)), a large endothermic peak centered at 216°C was attributed to both the removal of water (6% in mass) and dehydroxylation of the layered structure. This peak was followed by an exothermic band centered at 814°C, which was attributed to the combustion of the residual organic matter or partial crystallization of the oxide mixture. Considering that the residue was composed of a 0.17:0.66 ratio of A12O3 and ZnO, a mass loss of 31.5% was observed until temperatures of 550°C were reached and this experimental observation was in perfect agreement with the theoretical prediction. Between 550°C and 950°C, there was an additional mass loss of 3.5% and this was presumably attributable to complementary reactions involving oxides. The empiric formulae derived from the experimental data was calculated as Zn0.66Alo.34(OH)2(C03)o i?.0.4H2O, which is in perfect correlation with the Al:Zn ratio (1:2) used for sample preparation (Table 1).
Figure J5 - TG/DSC measurements of both (a) original Zn-Al-CO} and (b) the chemically modified Zn-Al-EG phase. Reprinted by kind permission of [J. Coll. Interface Sci., (227, 445, 2000)] [40].
Chemical Modification of Clay Surfaces
23
The Zn-Al-EG phase (Fig. 15(b)) presented a much more complex TG profile in which several simultaneous (therefore superimposed) processes were observed within the experimental temperature range (room temperature to 950°C). The corresponding DSC curve was characterized by a small endothermic peak at 60°C, which could be readily attributed to the removal of water from the crystal structure (nearly 6% in weight until 100°C). This peak was followed by the elimination and burning of ethylene glycol from the sample, characterized by an intense exothermic peak at 218°C. Smaller exothermic peaks at 340 and 399°C were attributed to the elimination of the remaining organic matter present within the sample. A 44% mass loss was observed between room temperature and 500°C. Between 500 and 950°C, an additional 3% mass loss was observed and this was again attributed to complementary reactions involving oxides. An exothermic peak, attributed to the partial crystallization of oxides, was also observed and centered at 752°C. Considering that the reaction procedure did not impair any changes to both Zn and Al contents, a theoretical mass loss of 43.3% could be predicted from the [Zno66Al034(OCH2CH20)](OH)o34.0.4H20 empirical formulae. Therefore, a slight deviation was observed between the theoretical and experimental data and this was probably a result of the residual ethylene glycol still present within the grafted material. Nevertheless, this small contamination was not enough to impair changes in the C:H ratio of the Zn-Al-EG phase (see Table 1 and FTIR data). Table 1 shows the chemical characterization of both Zn-Al-CO3 and Zn-Al-EG phases in relation to their wet mass for a typical moisture content of 6%. These results are in close agreement with the grafting of all hydroxyl groups present within the structure of the original LDH phase, giving the following empirical formulae, [Zno.66Alo.34(0-(CH2)2-0)](X"n)o.34/n.0.4H20, where OH" is the probable resident counterion. Table 1 - Chemical composition of both AI-Z11-CO3 and Al-Zn-EG phases, as determined by elemental analysis (C.H.N) and atomic absorption spectroscopy (Al,Zn) (wet basis for an average moisture content of 6%). Element Al-Zn-CO3 Al-Zn-EG T. (%, m/m) Exp. Exp. T. Carbon 19.2 2.0 2.7 19.7 Hydrogen 2.7 4.1 3.4 2.7 Nitrogen ND ND Aluminum 7.3 7.2 8.8 8.9 34.4 Zinc 41.6 34.1 41.7 ND = not detected. T. = theoretical; Exp. = experimental [Reprinted with kind permission from J. Coll. Interface Set, (227, 445, 2000)] [40]. To investigate whether the experimental procedure could cause any loss of Al and Zn by leaching, the original matrix was also subjected to the same experimental conditions. As a result, there were no observable changes in the Al/Zn ratio of the matrix. Therefore, this experimental control clearly demonstrated that no leaching of Al and Zn had occurred during sample preparation and that the assumptions made for the calculation of the empirical formulae were correct. Higher C contents in both samples were probably due to the occurrence of some residual ethylene glycol and contaminating solvent, even though this was not corroborated by the corresponding change in H content.
24
F.Wypych
The possibility of having carbonate as the counter-ion of the intercalated ZnAl-EG phase was eliminated by the complete absence of any FTIR band that could be attributed to its presence. However, the exchange of OH" for CO3"2 does not result in a significant variation in C, H, Al and Zn contents. Hence, the presence of CO3"2 could not be eliminated by elemental analysis and the theoretical yields of the empirical formulae, containing this counter-ion [Zn0.66Alo.34(0-(CH2)2-0)](C03)o n.0.4H2O, lie perfectly within the acceptable range depicted in the experimental data of Table 1 (C= 20.0%; H= 3.7%, Zn= 33.6%, Al= 7.0%). The FTIR spectra of (a ) Zn-Al-CO3,(b) Zn-Al-EG and (c) pure ethylene glycol are shown in Figure 16 with two distinct spectral ranges. The 2700 and 3700cm"1 region of the FTIR spectra provided interesting information about the structure of the LDH grafted composites. Two important bands were found within this spectral region: the out-of-plane stretching vibrations of C-H bonds at 2700-3100cm"1 and the stretching vibrations of O-H groups at 3100-3700cm"'. After the covalent grafting of ethylene glycol into the Zn-Al-CO3 phase, it was observed that two C-H stretching of ethylene glycol originally centered at 2879 (symmetric) and 2946cm"1 (antisymmetric) were either displaced or converted to at least three new absorption bands at 2856, 2896 and 2923cm"1. This observation confirmed the successful grafting of ethylene glycol within the layer structure of the host matrix because their C-H groups were then vibrating in a distinct chemical environment. Band displacements such as these can be used to characterize the higher rigidity of the grafted composite since the out-of-plane stretching of C-H bonds were naturally shifted to higher wavenumbers. However, other chemical interactions such as those with residual water and/or unreacted hydroxyl groups might have also contributed to the band displacements discussed above.
Figure 16 - FTIR spectra of (a) Zn-Al-COs, (b) Zn-Al-EG and (c) pure ethylene glycol (EG). [Reprinted with kind permission ofJ. Coll. Interface Sci., (227, 445, 2000)] [40]. Similar effects were also observed when both kaolinite and gibbsite were successfully grafted with ethylene glycol. The out-of-plane C-H stretching vibrations in kaolinite were shifted to 2969, 2945 and 2895cm'1 after grafting, whereas these same bands were displaced to 2920 and 2870cm"1 when gibbsite was used as the host matrix.
Chemical Modification of Clay Surfaces
25
For the simple intercalation of ethylene glycol into the kaolinite layer structure, there was no considerable change in the absorption profile at this spectral range and the observed bands at 2890 and 2945cm"1 were similar to those present in the FTIR spectra of the pure EG (2879 and 2946cm"1). Vibration frequencies other than those strictly related to the grafting of ethylene glycol into Zn-Al-CO3 were also observed in the FTIR spectra of the Zn-Al-EG phase. Hence, the occurrence of a number of relatively weak absorption bands, such as those centered at 2705, 2713, 3026, 3061 and 3082cm"1, were probably associated with minor contamination that might have been incorporated within the structure of the covalently grafted material. Sample preparation for FTIR was carried out after drying at 50°C to avoid exposure of the grafted composite to exceedingly higher temperatures. Therefore, complete removal of water could not be easily achieved and this was detrimental to the interpretation of absorption bands occurring around 3431cm"1 (water O-H stretching vibrations). The occurrence of a band at 1635cm"1 was the strongest evidence that some adsorbed water had remained within the sample. Likewise, this spectral region (1630cm" 1 and 1650cm"1) has also been used to detect water in other similar compounds such as kaolinite grafted with ethylene glycol [77,78]. The 1500-1200cm"1 spectral region of the grafted material revealed a series of peaks with low intensity, collectively attributed to CH2 stretching vibrations. This was an additional evidence for the strong rigidification of the ethylene glycol backbone after grafting [77]. The broad absorption band attributed to the carbonate counter-ion (1365cm1) (see the FTIR spectra of Zn-Al-CO3 in Fig. 16(a)) was completely removed from the LDH after grafting. In fact, this band was replaced by a sharp peak of low intensity, probably attributed to CH2 deformations of the ethylene glycol backbone. Therefore, as stated above, the occurrence of carbonate as the counter-ion for the Zn-AlEG phase was completely discarded and OH" was considered the actual counter-ion that was intercalated within the grafted material. Ethylene glycol is a very hygroscopic compound and any small amount of water present within the reaction mixture may trigger the gradual displacement of carbonate from the layered structure. This proposed exchange of counter-ions can also partly explain the broad association band (at 343 lcm"1) found in the FTIR spectra of the grafted material. Even though the absence of an absorption band at 1325cm"1 may be used to suggest that the conformations of oxyethylene units (O-CH2-CH2-O) are not in a trans conformation, it is possible that this same absorption band was slightly displaced to 1362cm"1 in the LDH-EG compound, thus characterizing a shift that has been already observed in other systems. For instance, absorption bands at 1030-1100cm"1, typically attributed to Al-O-C and CC-0 bonds in kaolinite, have been observed at 1043, 1072, 1081 and 1124cm"1 for ZnAl-EG. Rocking vibrations of the CH2 groups, generally centered at 864 and 882cm"1 for ethylene glycol, were almost completely absent from the FTIR spectra of the LDHEG compound. In fact, after grafting, these bands were converted into three new bands at the higher wavenumbers of 903, 911 and 919cm"1. A similar effect over the CH2 rocking vibrations was also observed when boehmite was used as the host matrix, where the original bands at 865 and 885cm"1 were displaced to 868 and 900cm"1 after the monodentade grafting of ethylene glycol [79]. Therefore, larger shifts of these rocking vibrations are expected when ethylene glycol is grafted into the host matrix under a bidentade conformation [79]. The relatively weak bands observed at 865, 882, 1041 and 1085cm"1 may be an additional evidence for the existence of small amounts of adsorbed
26
F.Wypych
ethylene glycol within the LDH-EG matrix. Nevertheless, the same observation led to the conclusion that this organic compound was successfully grafted onto the matrix from both ends (bidentade conformation), since the FTIR spectra brought little evidence for the persistence of any considerable amount of free ethylene glycol hydroxyl groups after grafting. 4.3 - Thermal reactions When hydrated ions intercalated compounds are heated to higher temperature, the solvent is released and a contraction of the basal spacing occurs. In most of the cases only the reversible dehydration/rehydration process occur (Eq. 21). [AlxMgl.x(OH)2]Clx.nH2O <-> AlxMgl.x(OH)2Clx + nH2O
(Eq. 21)
In some cases, the structure of the layered double hydroxide is preserved and intercalated ions can undergo chemical transformations that allow grafting to the layers (Eq. 22) [82,83]. Potentially this kind of reactions can also be applied to the hydroxysalts. [AlxMg1.x(OH)2](A")x(H2O)y + x/n B"n -> grafted derivatives
(Eq. 22)
By increasing the temperature, a mixture of amorphous oxides or ternary compounds (spinel like) can be obtained. One of the most interesting features of LDHs is the memory effect or reconstruction of the structure. This reconstruction is totally reversible (Eq. 23) when moderate temperature calcination temperatures are employed (ca. 300 - 500°C depending on the metals of the structure) and the amorphous material is put in contact with a solution or water vapor. The amorphous basic mixed oxides with a high surface area, high porosity, homogeneous dispersion of metallic particles, have many practical applications like catalysts, catalysts supports, ions exchangers, stabilizers in polymers, adsorbents, etc. [AlxMgl.x(OH)2]Clx.nH2O -> x/2 A12O3 + 1-x MgO o
[Al x M gl . x (0H) 2 ]-
(Eq. 23) In the case of ion intercalation containing a metallic atom, the thermal treatment under controlled conditions (inert or hydrogen atmosphere or vacuum), metallic particles can be obtained in mixtures of oxides matrix (Eq. 24) or still metal alloys [84,85,86]. The metallic particles will eventually be able to be used in the production of devices of the most varied species, catalysts, etc. [AlxMg1.x(OH)2](Fe(CN)6)x/3.nH2O -> x/2 A12O3 + 1-x MgO + x/3 Fe° (Eq. 24) 4.3.1 - Iron nanoparticles embedded in Al2O3-ZnO matrix [86]. Nitrate ions from a layered double hydroxide were exchange by hexacyano Fe(III) complex. The decomposition of the hexacyano Fe(III) complex and subsequent dehydroxilation of the LDH matrix was achieved by thermal treatment in high vacuum at 450°C during 2 hours, generating nanoparticles of Fe in a A12O3 and ZnO matrix. Figure 17 presents the FTIR spectra of the (a) original LDH and (b) after the exchange reaction with the hexacyano Fe(III) complex. The original LDH presents characteristic bands at 619, 1111, 1175, and 1384cm"1 attributed to nitrate and sulfate ion bands, respectively.
Chemical Modification of Clay Surfaces
27
After the hexacyano Fe(III) complex exchange reaction, bands attributed to carbonate ion bands were observed at 1357 cm"1 [40], nitrate bands at 1384 cm"1, sulfate bands at 1109 cm"1 and cyanide bands at 2097 cm"1 and 2110 cm"1 [14], respectively. The position of the cyanide bands depends on the composition of the LDH and possible processes of the iron oxidation/reduction of the hexacyano Fe(III) complex ion [14]. The cyanide bands are very strong, demonstrating that in spite of the presence of the other ions, the hexacyano Fe(III) complex ion is in a large concentration in the sample. The broad band centered at 3450 cm*1 and a narrow band in the 1630 cm"1 region, are attributed to adsorbed/absorbed/coordinated water molecules. Figure 18 shows the X-ray powder diffraction patterns of the (a) LDH as prepared, (b) after the hexacyano Fe(III) complex exchange reaction and (c) after thermal treatment of the LDH-FeCN at 450 °C under vacuum. Firstly, a compound of low crystal quality with a basal spacing of 10.7A, is observed. Although sulfate, nitrate and chloride salts have been used in the synthesis, it is expected that the LDH present a basal spacing of the larger diameter ion, i.e., sulfate. The basal spacing of 10.7A is usually observed when besides the sulfate ion, a neutral salt is also co-intercalated.
Figure 17 - FTIR spectra of (a) the original LDH and (b) after the exchange reaction with the hexacyano Fe(IH) complex ion. [Reprinted with kind permission ofJ. Phys. D: Appl. Phys., (36, 428, 2003)] [86]. The compounds intercalated exclusively with carbonate and nitrate ions would have the basal spacings of 7.8A [40] and 8.8A, respectively. After the exchange reaction, a compound with basal spacing of 10.9A was obtained. This basal spacing corresponds to the intercalation of the hexacyano Fe(III) complex ion, with the three-fold axis perpendicular to the LDH layers [17]. Obviously, in this case, the exchange reaction does not necessarily processed entirely. However, the basal expansion corresponds to those of the ion with larger diameter or hexacyano Fe(III) complex. After thermal treatment, a polycrystalline material with preponderant double-oxide Al2O3-ZnO composition was obtained. Some Bragg reflection peaks are identified in the powder X-ray powder diffraction pattern, whose detailed investigation was performed using transmission electron microcopy.
28
F.Wypych
Figure 18 - X-ray powder diffraction patterns of the (a) original LDH, (b) LDH-FeCN, and (c) after thermal treatment of the LDH-FeCN at 450°C under vacuum. The asterisk denotes the internal standard peak of Si. [Reprinted with kind permission ofJ. Phys. D: Appl. Phys., (36, 428, 2003)] [86].
Figure 19 - (a) Bright-field image obtained of the sample after thermal annealing under vacuum obtained with a transmission electron microscope operating at 120 kV, (b) SAED pattern of a large sample area showing diffraction rings associated with A12O3, ZnO and Fe, and (c) particle size distribution obtained by computational method. [Reprinted with kind permission ofJ. Phys. D: Appl. Phys., (36, 428, 2003)] [86].
Chemical Modification of Clay Surfaces
29
Figure 19(a) shows the bright-field image obtained with transmission electron microscope of the sample after thermal annealing under vacuum. The dark regions in the image are associated with Fe-rich particles. The Fe particles have rounded shapes with clear boundaries and small connectivity. Figure 19(b) shows the selected-area electron diffraction (SAED) pattern of a 30 um-diameter area of the same sample. The polycrystalline character of the double-oxide Al2O3-ZnO matrix is clearly observed from the diffraction rings. Since several interplanar spacings of the A12O3 bulk are superimposed to those of the ZnO bulk, an unambiguous identification of the diffraction rings could not be obtained. The presence of diffraction rings related to strained or unstrained Fe oxides cannot also be discarded. SAED patterns obtained from the electron beam focalization on the largest dark-regions reveal a predominance of the diffraction rings, which are unambiguously associated with metallic Fe. The doubleoxide diffraction pattern of the matrix appears rather uniform among distinct analyzed regions in the sample. Figure 19(c) exhibits a particle size distribution with a predominant maximum around 1.5 nm2 and several secondary peaks at 13, 29, 40, and 51 nm2. A dedicated software was used to obtain particle size area from the image computation. This procedure described the production of metallic Fe nanoparticles in a double-oxide of aluminum and zinc. The route is very attractive technologically since it involves simple chemical procedures with low cost resources. In the case of the doubleAl-Zn-hydroxide intercalate with the Fe(CN)6"3 complex anion, the annealing and subsequent dehydroxylation in vacuum leads predominantly to 7A-radius metallic Fe particles embedded in a double Al2O3/ZnO oxide host. 4.4 - Exfoliation reactions and preparation of nanocomposites Although the process of separation of individual layers can be potentially applied to any layered compound, transition metal dichalcogenides of the V and VIB groups were the first ones to be subjected to this kind of reaction [60-66]. The literature reports also two specific examples for the layered double hydroxides exfoliation [87,88] and probably none in simple hydroxides or hydroxysalts. 4.4.1 — Exfoliation of a layered double hydroxide and reaction with poly(ethylene oxide) [89] Sulfate ions from the layered double hydroxide were exchanged with dodecylsulfate, using sodium dodecylsufate (SDS). The supernatant suspension containing an excess of SDS was separated and reacted with an aqueous solution of poly(ethylene oxide) in 50ml of water. Figure 20 shows the X-ray powder diffraction results for the entire sample preparation sequence. The sequence of diffraction results shows layered compounds with increasing crystal quality and increasing basal plane separation. Figure 20(a) shows the typical powder X-ray diffraction pattern of the sulfate intercalated layered double hydroxide (LDH-SO4) with a 11.1 A layer separation in accordance with the literature results [90]. The replacement of SO4"2 for the anion dodecylsulfate leads to the LDH-DDS sample shown in Figure 20(b). This sample has a basal spacing of 26.2A in good agreement with the literature [91,92]. The nanocomposite formed after the LDH-DDS reaction with PEO, which we call LDHDDS-PEO, has an expanded basal spacing of 35.9A, as shown in Figure 20(c). This separation grows to 38.2A for the 110°C heat-treated nanocomposite, as shown in Figure 20(d). The X-ray diffraction results of the LDH-DDS-PEO composite show clear modulation of the high order diffraction peaks. In one case (Fig. 20(d)), these peaks go up to 15 orders. The high diffraction orders indicate very well crystallized samples with
30
F.Wypych
long order correlation lengths, while the modulation reveals ordered substructures within layers. The transmission electron microscopy results were obtained by using a JEOL 1200 EX-II instrument operating at 40 kV. The measurements were extremely difficult, since the interaction of the electron beam with the sample caused rapid deterioration of the latter. The selected area electron diffraction (SAED) pattern shows some welldefined spots together with a strong diffuse ring for the LDH-DDS-PEO nanocomposite (not shown). The ring is due to the already mentioned deterioration processes. In contrast, the well-defined spots reveal a hexagonal structure with (4.7J6 0.248)A and (2.715 + 0.082 )A for the interplanar spacing of the [ 1 100] and [2110] directions, respectively. No distortion in the basal plane projection could be observed, indicating that the lattice distortion occurs only in the "c " axis direction. The composite dried at 110°C could not be observed due to the above mentioned beam-sample interactions. The TG results are shown in Figure 21 for (a) LDH-SO4, (b) LDH-DDS and (c) LDH-DDSPEO. DSC measurement for LDH-DDS-PEO was also included (Fig. 18(d)). Clearly, all results are distinct in each case, indicating the presence of different layered materials. In the case of LDH-SO4 the TG data permit us to calculate the formula for the compound: Al0,33Mgo,67(OH)2(S04)o,i7.0,61H20. After the exchange of the sulfate ions by DDS ions, the TG data are consistent with the following formula: Alo,33Mgo,67(OH)2(Ci2H25S04)o,33.0,64H20. Here, a mass loss of 64,5% was observed between 150 and 1000°C. In the HDL-DDS-PEO case, an additional mass loss was observed, that is consistent with PEO incorporation into the sample. We observed a 71.5% sample mass loss between 150 and 1000°C.
Figure 20 ~ X-ray powder diffraction patterns (a) for LDH-SO4, (b) LDH-DDS, (c) LDH-DDS-PEO at room temperature and (d) heated at JOO°C for 1 hour. The asterisk indicates the silicon standard (111) peak. Copyright - Langmuir, (18, 5967, 2002) [89]. Also distinct are the Fourier Transform Infrared (FTIR) results for the samples as shown in Figure 22. Here again we find that the intercalation always lead to new features and characteristic absorption bands. With the exception of HDL-SO4, all other samples display the characteristic absorption bands of DDS at 2850, 2919 and 2957cm"1 [91,92]. Finally, some bands related to PEO [93] are superimposed on the strong DDS
Chemical Modification of Clay Surfaces
31
bands in the same region. Additionally, we observe that the sulfate ion in HDL-DDSPEO is coordinated in a different form as compared to HDL-SO4 (449, 619, 991, 1115 and 1190cm"1) [90,94] and SDS (1221 and 1247cm"1), since the bands have been shifted to 1214, 1249 and 1270cm"1. Absorption bands relative to KNO3 (695, 828 e 1370cm"1) [94] or nitrate ions (1380cm"1) [94,95] as well as carbonate ions (1365cm"1) were not observed [11]. The intercalation of layered double hydroxides with long chain surfactants yields hydrophobic surface properties. These systems are potential adsorbents for the removal of charged and neutral organic molecules from aqueous systems, being important from environmental pollution control [96-99].
Figure 21 - Thermogravimetry (TG) curves for (a) LDH-SO 4> (b) LDH-DDS and (c) LDH-DDS-PEO. (d) Differential scanning calorimetry (DSC) curve for LDH-DDSPEO. Copyright - Langmuir, (18, 5967, 2002) [89].
Figure 22 - Fourier transform infrared spectra (FTIR) for (a) LDH-DDS, (b) LDH-SO4 and(c) LDH-DDS-PEO. Copyright - Langmuir, (18, 5967, 2002) [89].
32
F.Wypych
5 - Kaolinite derivatives In kaolinite, due to presence of hydroxyl groups on the aluminum side of the layer (Fig. 5), the process of solvation of those groups may occur (including polymers) [100-104], as well as the grafting reactions [77,78,105,106]. An important characteristic is the possibility to bond covalently specific molecules to the layer (through interlayer aluminol groups) or to modify them after the grafting process, with an intention of attributing their own characteristics to the matrix. Through this procedure interesting materials can be obtained, such as those obtained by the intercalation of organic or inorganic compounds (dyes or pigments, catalysts, precursors of catalysts, ionic exchangers, materials with controlled surface area, with controlled micro and macroporosity, etc.). Apart from these possibilities, the confinement (through grafting or intercalation) of molecules in an asymmetrical chemical nanoenvironment can generate materials with differentiated physical properties of from those that are observed with the free molecules or in the crystalline state [20,21]. The intercalated molecules can also be located into the hexagonal siloxane cavity (Fig. 4(c)) or occupy the octahedral aluminum vacant site (Fig. 5(c)). Despite such infinite possibilities, there have been very few attempts made for the kaolinite as host matrix for such reactions. The most important reactions involve the solvation of the hydroxyl groups (intercalation) and aluminol groups' functionalization (grafting). However, reactions of surface adsorption, exfoliation and synthesis of nanocomposites should also be considered [107,108]. 5.1 - Direct solvation (intercalation) Considering that appropriate polar molecules can have access to the interlayer hydroxyl ions linked to aluminum atoms (aluminol), a solvation process is perfectly possible [42-46]. Apart from organic molecules, simple salts as potassium acetate can be intercalated directly by the simple milling process with pure kaolinite [109-111], or its contact with a solution containing the molecule to be intercalated or the simple contact with the solvent to be intercalated. Usually the intercalation processes are carried out at temperatures slightly higher than the room temperature. After several days of reaction, the solid material is separated by centrifugation and washed with an appropriate solvent or dried at a controlled temperature, to avoid the removal of the intercalated molecule (when the intercalated molecule is sufficiently volatile) (Eq. 25). Al2Si2O5(OH)4 + x (CH3)2SO <-> Al2Si2O5(OH)4((CH3)2SO)x
(Eq. 25)
5.1.1 - Intercalation of Dimethylsulfoxide (DMSO) 9.0g of kaolinite were dispersed in a mixture composed of 60mL of (DMSO) and 5.5mL of distilled water. The reaction was carried out at room temperature for a period of 10 days in a 50mL plane-bottom glass flask equipped with magnetic stirrer agitation. The resulting material was centrifuged at 4000 rpm and dried at 50°C for 24 h, to eliminate the excess of DMSO. A light brow expanded kaolinite-DMSO complex was obtained with an intercalation ratio of about 85% and the basal spacing of 11.21 A, which represent an expansion of 4.04A in relation to the basal spacing of the raw kaolinite (7.16A) (Fig. 23). The basal spacings were obtained from the powder X-ray powder diffraction patterns, by using the reflection of a higher possible order (normally 5). The results of the TG/DSC analysis are presented in Figure 24. For pure kaolinite (results not shown) two endothermic peaks could be observed. The first one, was related
Chemical Modification of Clay Surfaces
33
to elimination of adsorbed/absorbed water, centered at 51°C while the second centered at 529°C, was attributed to the dehydroxylation process to metakaolinite.
Figure 23 -X-ray powder diffraction of (a) kaolinite and (b) the K-DMSO (*= Si).
Figure 24 - TG/DSC of kaolinite reacted with DMSO. The K(DMSO)X phase showed two endothermic peaks, one centered at 175°C, which could be attributed to DMSO elimination, and other centered at 509°C that corresponded to the kaolinite dehydroxylation process. Considering the concentration of non-reacted kaolinite (near of 16.5%) and the moisture of the material (0.7%), the measured loss of organic matter (8.8%) was in accordance with the theoretical value obtained from the proposed formula (8.8%). The concentration of the final residue (79%) was also in accordance with the decomposition of non-reacted kaolinite and the intercalated material (theoretical value: 78.5%). Due to the relatively low temperature of DMSO elimination (175°C), it could be taken for granted that the process involved integral elimination of the molecule instead of its burning. This fact has been
34
F.Wypych
corroborated by the absence of any exothermic peak that could be attributed to DMSO combustion, even when the experiments were carried out in an air atmosphere. Considering the results of the thermal analysis it has been possible to estimate that the stoichiometry is K(DMSO)0 5.2 - Intercalation by displacement [112-117] Some specific molecules cannot be intercalated directly into kaolinite. However, they can displace molecules previously intercalated (Eq. 26), mainly based on the fact that hydrogen bonds that maintain the structure have been already partially broken. Al2Si2O5(OH)4((CH3)2SO)x + y C6H5CONH2 -> Al2Si2O5(OH)4(C6H5CONH2)y + x(CH 3 ) 2 SO (Eq.26) Sometimes the exchange process is incomplete and compounds containing mixtures of intercalated molecules in varied proportions can be obtained. Several are the replacement procedures of previously intercalated molecules. If the exchanging process involves a solvent, it is enough to keep the intercalation compound in contact with the new solvent. In the case of intercalation of a solid, a solution in a non-reactive solvent is used or through the fusion of the solid to be inserted [112]. In the case of polymers, the procedure involves the intercalation of a monomer and subsequent polymerization by chemical, thermal treatment [101,102] or fusion of the polymer in contact with the previously intercalated compound [100]. 5.2.1 -Displacement of DMSO by benzamide [112] The apparent intercalation ratio (I.R.) of both K-DMSO and K-BZ was determined in their X-ray powder diffraction patterns using the Eq. 27 [117]. I.R. = Intensity (first peak) Intercalate / Intensity (first peak) Intercalate + Intensity (first peak) kaolinite (Eq. 27) The DMSO-intercalated kaolinite (K-DMSO fraction) has shown to be paleyellow powder with an intercalation ratio (I.R.) of 81.5% and basal spacing of 11.21 A. The benzamide-intercalated kaolinite (K-BZ fraction), obtained from K-DMSO, was also shown to be pale-yellow powder with an I.R. of 73% and a basal spacing of 14.29A, which represents an expansion of 7.14A in relation to raw kaolinite. In this case, the I.R. could not be directly calculated from the X-ray powder diffraction pattern because the second reflection of the K-BZ fraction (d=7.14A) and the first reflection of the raw kaolinite (d=7.16A) were almost perfectly superimposed. Therefore, measurements for I.R. calculations were only performed in K-BZ after its normalized X-ray powder diffraction pattern was subtracted from the K-DMSO normalized X-ray diffraction background. For the experimental control in which raw kaolinite was used in the absence of DMSO, there was no evidence that benzamide could be intercalated within the kaolinite host, suggesting that pre-treatment with DMSO (displacement method) was a requirement for the successful inclusion of this intercalation compound. The X-ray powder diffraction patterns of (a) pure benzamide, (b) raw kaolinite, (c) both K-DMSO and (d) K-BZ composites are collectively shown in Figure 22, where the diffraction pattern of the internal standard (powdered silicon) is labeled with an asterisk
Chemical Modification of Clay Surfaces
35
(*). Figure 25 clearly indicates that there is no evidence for crystalline benzamide in the K-BZ X-ray powder diffraction pattern, suggesting that benzamide substitution has been fully accomplished and that the only crystalline materials found within the matrix are kaolinite and the resulting K-BZ intercalation compound. To facilitate interpretation of Figure 22, the basal reflections of pure kaolinite, K-DMSO and K-BZ have been respectively labeled as Kn, Dn and Bn. However the "n" values observed in this figure do not represent normal indexations because it was not possible to determine if kaolinite has undergone any structural transition after intercalation. As stated above, there was no evidence that any DMSO had remained co-intercalated within the K-BZ matrix, unless that occurs at a relatively low level, which would not interfere with its basal spacing. Pure benzamide was also apparently absent from the X-ray powder diffraction pattern of K-BZ. Variations in basal spacing were determined by subtracting the basal spacing of the intercalated kaolinite from the basal spacing of raw kaolinite (7.16A). The molecular diameter of benzamide, measured between the /^-substituted aromatic hydrogen and the oxygen atom of the carbonyl group, was determined as 7.7A by using a modeling software [118]. Therefore, based on the 7.14A variation in the kaolinite basal spacing after intercalation, it seems that only one type of hydroxyl group is interacting directly with a single intercalated benzamide molecule and that each intercalated molecule is displaced at an angle of 68° in relation to the plane of the kaolinite layer. In fact, this assumption is consistent with the 50 to 75° orientation angle that is normally assumed by hydroxyl groups on the surface of the kaolinite layer [119,120].
Figure 25 - X-ray powder diffraction patterns of (a) pure benzamide, (b) raw kaolinite, (c) K-DMSO and (d) K-BZ composites. [Reprinted with permission from J. Coll. Interface Set, (221, 284, 2000)] [112]. Figure 26 shows the TG/DSC/DTG measurements made on (a) raw kaolinite and (b) on the K-BZ intercalation compound. For raw kaolinite, the 0.7% mass loss observed at temperatures below 250°C was attributed to the loss of moisture. After that, the dehydroxylation of kaolinite into metakaolinite was observed as an endothermic peak centered at 532°C. This process generated a mass loss of 14.1% and this was in good agreement with the 13.96% value predicted from the theoretical formula of pure kaolinite (Al2Si2O5(OH)4). Therefore, the kaolinite sample used in this study was of a
36
F.Wypych
very good quality, a property that could be even improved by appropriate chemical treatments to remove iron (e.g., the dithionite/citrate/bicarbonate method). As the same amount of impurities was detected after several attempts to further purify the untreated kaolinite, these elements appear to be present within the kaolinite matrix as isomorphic/non-isomorphic substitutions. The exothermic peak centered at 985°C was attributed to the crystallization of both Si and Al oxides. The K-BZ phase showed one small endothermic peak at 60°C, readily attributed to the loss of adsorbed water (0.5% mass loss), and one broad endothermic band centered at 225°C, followed by two endothermic peaks with their average intensities centered at 314 and 341°C, respectively. These measurements were collectively attributed to the loss of organic matter from the kaolinite host. As both processes were characteristically endothermic, benzamide molecules appeared to be displaced from their interlayer spacing without being burnt. The endothermic peak centered at 514°C was associated with dehydroxylation of the lattice matrix, whereas crystallization was observed at 988°C as the last exothermic event of the DSC profile.
Figure 26 - TG/DSC measurements made on (a) raw kaolinite and (b) on the K-BZ intercalation compound. Reprinted with permission, from J. Coll. Interface Sci., (221, 284,2000)] [112]. These two processes were also observed in raw kaolinite at 532 and 985°C, respectively. The complete absence of benzamide melting peaks (130°C) in the K-BZ thermal curves demonstrated that there was no excess of this compound within the intercalated derivative. Assuming that K-BZ has nearly 9.6% organic matter in its chemical composition (dry basis), the overall mass loss for a K(BZ)o 2 stoichiometry up to 350°C would theoretically correspond to approximately 9.57%. Indeed, there was a perfect agreement between this theoretical value and the data determined experimentally. Likewise, the amount of residues recovered after the experiment at 1000°C (77.6%) also revealed a perfect agreement with the expected theoretical value of 77.8% in relation to the K-BZ dry weight. After 300°C, no more organic matter was expected to be present in K-BZ and assuming that the K-BZ has an I.R. of 73%, the experimental mass loss of 13.9% was in perfect agreement with both theoretical and
Chemical Modification of Clay Surfaces
37
experimental values obtained previously for raw kaolinite (13.96% and 14.1%, respectively). Considering that at temperatures beyond 35O°C, all of the benzamide molecules had been completely removed from the kaolinite matrix, it is possible to conclude from the TG/DSC measurements that both intercalation derivatives have very similar stoichiometries of K(BZ)o 2 and K(DMSO)0! - This observation suggests that the intercalation of benzamide molecules was dependent upon the substitution of DMSO molecules from the host matrix. As the I.R. ratio is not precisely known, it is possible that the actual K-BZ stoichiometry is slightly different from that proposed above. The TG/DSC measurements of the K-BZ phase also suggested that DMSO was completely substituted by benzamide during intercalation of the K-DMSO derivative. This was supported by the complete absence of an endothermic peak at 189°C, which corresponds to the elimination of DMSO from the kaolinite matrix. Figure 27 shows the FTIR spectra of (a) raw kaolinite, (b) the K-BZ phase and (c) pure benzamide.
Figure 27 - FTIR spectra from (a) raw kaolinite, (b) the K-BZ phase and (c) pure benzamide. Reprinted with permission, from J. Coll. Interface Sci., (221, 284, 2000)] [112]. A tentative interpretation of the FTIR spectra is given in Table 2, on the basis of FTIR data available in the literature for raw kaolinite, K-DMSO, benzamide and others [121-123]. The FTIR spectrum of the K-BZ derivative contained all the major FTIR bands attributed to kaolinite and benzamide. However, there was no evidence for bands associated with DMSO and this confirmed the absence of any co-intercalated DMSO within the benzamide-intercalated kaolinite. Compared to kaolinite, the FTIR spectrum of the K-BZ derivative showed variations within the region characteristically attributed to O-H axial deformations (3400-3800cm"'). There was a considerable increase in the absorption intensities at the 3647cm"1 region with the concomitant appearance of a shoulder at higher wavenumbers while both 3666cm"1 and 3696cm"1 bands remained relatively constant or even decreased in their relative intensities. This observation led to the hypothesis that, out of the two (or three) distinct hydroxyl groups found on the surface of the kaolinite layered structure [123], only one contributed the most to the hydrogen bonding established directly with the intercalation compound. The
38
F.Wypych
FTIR absorption band at 3619cm"1 has been previously attributed to hydroxyl groups that are embedded within the kaolinite matrix [124,125]. Figure 27 shows that the intensity of this band was not influenced by the intercalation process and this was apparent from the FTIR spectra of both K-DMSO and K-BZ derivatives. Therefore, it seemed that hydroxyl groups that are distributed internally are not directly affected by intercalation of either DMSO or benzamide into kaolinite. The intercalation process of K-DMSO with benzamide also generated three new absorption bands in the FTIR spectra of the resulting K-BZ intercalation compound. These bands, located at 3598cm" ', 3549cm"1 and 3472cm"1, were tentatively attributed to the axial deformations of hydroxyl groups that are involved in hydrogen bonding with the carbonyl group of benzamide. Considering that FTIR could be used to characterize the nature and strength of hydrogen bonds and that the weaker the hydrogen bonding, the lower the wavenumber in which the associated O-H stretching occurs, it seemed that the new absorption band at 3549cm"1 could indicate the occurrence of hydrogen bonding between the carbonyl group of benzamide and co-intercalated water molecules. Nevertheless, the amount of co-intercalated water molecules must be very low because there was no evidence for considerable mass loss in the TG/DSC experiments. .data of K-BZ. Table 2 Attribution Peak Wavenumber 1 K:3694 - O-H surface 3696 2 K:3666 - O-H surface 3670 3 3647 K:3650 - O-H surface 3619 4 K:3619 - O-H inner 3598 5 K:O-H O=C 6 K:O-H O=C or H-O-H 3549 7 K:O-H O=C 3472 8 3391 B:3370-N-H 9 3372 B:3370-N-H 3180 10 B:3176-N-H 11 B:1625 - H-O-H; B:1660 - C O 1638 B:1603, 1617 - N-H and/or H-O-H 12 1606 B:1578-N-H 13 1574 B:1449 14 1447 B:1404 15 1407 B:1298 16 1300 K:1107-Si-O-Si 17 1108 B:1073 18 1083 B:1073 1057 19 B: 1026;K:1033-Si-O 1034 20 B: 1001;K:1006-Si-O 21 1007 K:936 - O-H inner 23 938 24 914 B: 919; K:913 - O-H surface 879 25 K:877 790 26 B:792; K:791 - Si-O-Si 27 K:752 - Si-O-Si 754 28 K:697 - Si-O-Si 692 29 548 K:538 - Al-O-Si 30 472 K:467 - Si-O 431 31 K:431-Si-O B: 414; K:411-Si-0 32 411 [Reprinted by permission from J. Coll. Interface Sci., (221, 284, 2000)] [112].
Chemical Modification of Clay Surfaces
39
The external hydroxyl groups are also responsible for an absorption band at 936cm"1 in the FTIR spectrum of kaolinite, whereas internal hydroxyl groups contribute with a band located at 913cm"1. Figure 27 shows that the intensity of the former band (internal hydroxyl groups) in K-BZ decreases in relation to the latter (external hydroxyl groups) after intercalation. Therefore, this was another strong evidence that the intercalated molecule was directly associated with the kaolinite matrix through hydrogen bonding. The FTIR spectrum of pure benzamide showed a sharp, single absorption band at 919cm"1. Therefore, if any free benzamide were present in K-BZ, this band would have partially contributed to the relative intensity of the broader K-BZ band at 913cm"1 (internal hydroxyl groups). However, no other spectral evidence for free benzamide was found in K-BZ, particularly within the 1000-4000cm"' region, suggesting that benzamide was indeed absent from the K-BZ composite. Additional variations in the FTIR spectra were observed within the 1500-1700cm"1, which corresponds to N-H and C=O deformation modes in amides. Based on the benzamide FTIR spectrum, the C=O stretching at 1660cm"1 was shifted to a band centered at 1638cm"1. Likewise, both N-H deformation modes located at 1578 and 1625cm"1 were detected as a single peak at 1574cm"1 with a shoulder at a slightly higher wavenumber. This is an additional evidence that, besides the C=O bond, the N-H bond in benzamide is also affected by the intercalation process. 5.3 - Direct grafting Apart from the processes of simple intercalation, processes of direct grafting can be achieved with kaolinite (Eq. 28) [105,106]. Al 2 Si 2 0 5 (0H) 4 + x C6H5PO(OH)2 -+ Al2Si2O5(OH)4.x(C6H5PO3H) + x H2O (Eq. 28) Usually only part of the interlayer hydroxyl groups are reacted, in which a mixed composition containing both Al-O-H and Al-O-C chemical bonds (for the specific case of kaolinite reaction with an alcohol) are obtained. In this kind of reaction, kaolinite is suspended in a non-reactive solvent containing the molecule to de grafted and kept in refluxing conditions by several days or in a pressurized container. After reaction, the solid material is separated by centrifugation and washed with an appropriate solvent or dried in a controlled temperature, to avoid the removal of the grafted molecule. Usually, those compounds are more stable than those obtained by simple intercalation, but in some cases the grafted molecules can be removed by hydrolysis. The range of possible reactions for this case involves all those available in the classic organic chemistry or others, characteristic of the system. 5.3.1 — Reaction with phenylphosphonic acid [105] The reaction of phenylphosphonic acid with kaolinite in aqueous/acetone solution was carried out at C for a period of 20 days. The X-ray powder diffraction patterns (Fig. 28) showed that all the reflections of the raw kaolinite could be indexed, which implied that no crystalline impurities were present. A basal spacing of 15.02A (KPPA15) was observed on the final compound, which represented an interlayer expansion of 7.88A. This value is coherent with the grafting of phenylphosphonate groups into the layer of kaolinite. For a better observation of the non-basal reflections, a ten-fold expanded X-ray powder diffraction pattern of the final product is shown in Figure 29. In
40
F.Wypych
this sequence, a single number denotes the basal reflections of the modified material. K denotes the signal of the residual kaolinite, while an asterisk (*) represents silicon as an internal standard. One spurious phase was identified as a splitting of the first basal reflection of the phase K-PPA-15, occasionally observed during the synthesis process. With the assumption that the basal spacing of this compound was of 16.45A, it was denominated K-PPA-16. Considering the possibility of formation of a hydrated species, the K-PPA-15 phase was subjected to a hydration process. For this purpose, 0.5g of the dry material was reacted with 50 mL of distilled water for a period of 48 hours. In this experiment (results not shown), a partial decomposition of K-PPA-15 phase to kaolinite was observed (about 20%). The hydrated form was observed as a small shoulder on the left side of the first basal reflection of K-PPA-15.
Figure 28 - X-ray powder diffraction pattern of the (a) kaolinite and product of the reaction (K-PPA) with different reaction times. [Reprinted by permission from J. Coll. Interface Set, (206, 281, 1998)] [105].
Figure 29 - X-ray powder diffraction patterns of the final product with the expansion of the powder diffraction pattern (lOx) to better observe the non-basal reflections. Reprinted, by permission, from [J. Coll. Interface Sci., (206, 281, 1998)] [105].
Chemical Modification of Clay Surfaces
41
Although the material was in contact with water for several days, the isolation of a pure K-PPA-16 phase was not possible. If we consider that the observed interlayer expansion is larger than those observed for water molecules intercalated in zinc and cobalt phenylphosphonate and methylphosphonate, it is possible to assume that the KPPA-16 phase corresponds to a product of partial hydrolyses, instead of a hydrated form. 5.4 - Grafting by displacement Apart from the substitution of the intercalated molecules, processes of functionalization can be also accomplished (Eq. 29).
+ y H2O
Al 2 Si 2 O 5 (OH)4(DMSO) x + y CH 3 -OH -> Al2Si2O5(OH)4_y(O-CH3)y + x DMSO (Eq. 29)
Depending on the organic molecule, bonds through one or more bridges can be established. In the case of one bridge, reactions between kaolinite and primary alcohols can be used as examples, being the resulting material consisted of a composition similar to an ester, releasing a water molecule. In the case of two bridges, reactions with ethylene glycol or glycerol can be used as examples. The grafted compounds can be chosen in such a way to produce very interesting compounds for potential industrial applications. Examples of such applications could be obtained with molecules containing ionic exchange groups (cationic or anionic) that are positioned between the layers, reactions that produce colored materials (having some specific physical properties) or even those containing catalysts or precursors of catalysts. The process of grafting of long linear molecules can produce materials with high porosity with potential applications in filtration, chromatographic or environmental pollution control [96-99]. 5.4.1 - Grafting of ethylene glycol by the displacing of dimethylsulphoxide
Figure 30 - X-ray powder diffraction patterns of (a) raw kaolinite, (b) K-EG and (c) KDMSO.
42
F.Wypych
Figure 30 shows the powder X-ray powder diffraction patterns of (a) raw kaolinite, (b) kaolinite grafted ethylene glycol (K-EG) and (c) kaolinite intercalated with dimethylsulphoxide (K-DMSO). The main reflections of these phases, labeled as "Dn", "En" and "Kn" in Figure 30, were respectively associated with the sequence of basal reflections of K-DMSO, K-EG and K, where "n" is an integral number. In the K-EG phase, a basal spacing of 9.5 A was observed, which is consistent with the grafting of an ethylene glycol single layer to the aluminol side of the kaolinite layer. The basal expansion of 2.3A is similar to those described previously for the same compound [77,78]. Variations lower than the molecular diameter of ethylene glycol are justified by the interpenetration of the grafted molecules into the hexagonal siloxane cavity of the silicate sheet (Fig. 4(c)). 5.5 — Mechanochemical modifications [109-111] Intercalation reactions, chemical or morphological modifications of layered crystals, can be obtained by milling specific chemicals with the layered materials. Using kaolinite structure as an example, basically the process consists of dry milling kaolinite with appropriate chemicals (urea, potassium acetate, etc.). Apart from the intercalation reaction as described in Eq. 25, the crystals can change the morphology from layered to cylindrical or tubular, being this change is more pronounced when pressurized vessels are employed after the mechanochemical activation. The real positioning of the intercalated molecules is still under discussion. However, as described previously, part of them (the cation or the anions) can be inserted into the hexagonal cavity of the siloxane surface sheet (Fig. 4(c)). 5.5.1 - Intercalation of urea and preparation of hydrated kaolinite [111] The X-ray powder diffraction patterns of the resulting material from the milling of kaolinite with urea, in the proportion of 20% in mass are shown in Figure 3 l(b).
Figure 31 - X-ray powder diffraction patterns of (a) neat urea, (b) kaolinite intercalated urea in the proportion of 20%, (c) 30% and (d) raw kaolinite. [Reprinted from Quimica Nova, (24, 6, 761, 2001), with Permission from Sociedade Brasileira de Quimica] [111].
Chemical Modification of Clay Surfaces
43
It can be observed that practically all kaolinite was intercalated, generating a compound with basal spacing of 10.76A (expansion of 3.6A in relation to kaolinite; d = 7.16A [109,110]. In the case of the intercalation with 30% of urea (Fig. 31(c)), all kaolinite was reacted, however, through an intense X-ray diffraction reflection, urea was detected in the region of 26=26°(Fig. 31 (a)). The second basal reflection of kaolinite (28 =28.9°) coincides with the third basal reflection of the urea-intercalated compound. A priori this could indicate a compound that did not undergo reaction with the urea. However, that is not the case as evidenced below. It is observed that after washing the intercalated sample with 20% of urea (the same happens with the 30% sample), urea is totally eliminated from the interlayer spacings. While the sample is still wet (Fig. 32(b)) few X-ray diffraction peaks were observed, which demonstrate that the coherency of stacking of layers is poor. Apart from raw kaolinite at 7.2A, two small reflections with basal spacings around 20.1 A and 8.4A were identified.
Figure 32 — X-ray powder diffraction patterns of the intercalated sample (a) with 20% of urea after washing at 90°C in an ultrasound bath and air drying, (b) still wet and (c) after the milling of the air dried product (hydrated kaolinite) with 30% of urea as described for the pure kaolinite and (d) hydrated kaolinite produced through the methanol washing of the dimethylsulfoxide intercalated kaolinite. [Reprinted from Quimica Nova, (24, 6, 761, 2001), with Permission from Sociedade Brasileira de Quimica] [111]. The first refection could be attributed to an intermediary hydrated phase or formation of ordered heterostructures and the second, to the stable hydrated phase. After the air drying process (Fig. 32(a)), hydrated kaolinite was observed, which was characterized by a broad reflection with an basal spacing of 8.4A [126], apart from a small concentration of raw kaolinite. After milling of the dry hydrated kaolinite with 30% of urea, the urea intercalated kaolinite phase has reappeared, with an excess of urea (Fig. 32(c)), demonstrating that the hydrated kaolinite was chemically similar to raw kaolinite. Keeping in mind that water plays an important role in breaking the hydrogen bonds that hold the layers together, it is expected, that with a bigger basal spacing, this phase is
44
F.Wypych
more appropriate for the intercalation of molecules, which is normally not directly intercalated with raw kaolinite. TG/DSC measurements were performed with raw kaolinite, pure urea, urea intercalated kaolinite in the proportion of 30% and 20% and hydrated kaolinite, as can be seen in Figure 33. TG curve of raw kaolinite (Fig. 33(a)) presents a mass loss of 1.05% up to 200°C attributed to moisture, followed by a process of dehydroxylation of the matrix (endothermic peak in the DSC curve at 527°C) and the crystallization of the oxides (exothermic peak in the DSC curve at 987°C). The mass loss up to 1000°C (in dry matter basis) of 14.1% is in a very close agreement with the expected value of 13.96% for the ideal composition proposed for the kaolinite (Al2Si2O5(OH)4) and formation of oxides at the end of the thermal treatment: A12O3 and SiO2. Pure urea (Fig. 33(b)) presents a complex decomposition profile under air, beginning with the fusion process associated with an endothermic peak centered at 142°C observed in the DSC curve. An endothermic peak is observed at 213°C following by endothermic peaks at 252, 343, 374 and 396°C. At that sweeping speed, at least 4 decomposition steps were observed till the complete elimination of the sample from temperatures up to 420°C. The decomposition process of the 30% urea intercalated kaolinite (Fig. 33(c)) presents a typical profile of a mixture of urea and urea intercalated kaolinite. In the TG curve up to 120°C, the elimination of the sample moisture was observed (1.36%), associated with an endothermic peak at 54°C in the DSC curve. At 141°C in the DSC curve, the fusion of the excess urea was observed along with a complex decomposition process of the residual urea and the destruction of the intercalation compound (endothermic peaks at 189, 199, 219, 236 and 308°C). The matrix dehydroxylation was observed at a slightly lower temperature (513°C) than in raw kaolinite (527°C), demonstrating that the intercalation process produced crystal delamination (an effect that facilitates the structure dehydroxylation process). The exothermic peak relative to the crystallization of the oxides was observed at 991°C in the DSC curve. The TG/DSC curves of the intercalated sample with a composition of 20% of urea (Fig. 33(d)) present a quite simplified profile in relation to the 30% proportion. One endothermic peak was observed at 59°C, being associated with the dehydration of the sample (mass loss of 2.48% between room temperature and 120°C) followed by a step of organic matter removal (endothermic peak at 231°C in the DSC curve and a mass loss of 15.87% between 120°C and 380°C). The dehydroxylation of the matrix was observed in 517°C (loss of mass of 12.37% between 370°C and 1000°C) apart from one exothermic peak at 988°C. An interesting characteristic in this system is associated with the elimination of the whole organic matter of the sample without the destruction of the matrix. Apart from the simplification of the urea decomposition process that takes place in only one step, the absence of urea fusion peak in the DSC curve, demonstrates that the intercalated phase with 20% does not present any nonintercalated urea as also evidenced by powder X-ray diffraction. Another interesting fact is the stabilization of the decomposition of the urea after the intercalation process that takes place at 231°C in comparison with 213°C in the pure urea. The generated stoichiometry starting from the obtained data, Al2Si205(OH)4(N2H4CO)o,84 is quite close to the predicted stoichiometry starting from the mixture of the chemicals (Al2Si20;5(OH)4(N2H4CO)o 86 ). The hydrated phase (Fig. 33(e)) presents a quite different decomposition profile. The process of moisture elimination between room temperature and 100cC in the TG curve (1.03%) is accompanied by two endothermic peaks at 40°C and 59°C in the DSC curve.
Chemical Modification of Clay Surfaces
45
Figure 33 — TG/DSC curves of (a) raw kaolinite, (b) pure urea, (c) urea intercalated kaolinite in the proportion of 30%, (d) 20% and (e) hydrated kaolinite. [Reprinted from Quimica Nova, (24, 6, 761, 2001), with Permission from Sociedade Brasileira de Quimica] [111].
46
F.Wypych
The process of intercalated water removal between 100 and 370°C (3.27%) is accompanied by two endothermic peaks of low intensity centered at 113°C and 192°C, in the DSC curve. Based on these data, the stoichiometry can be obtained (Al2Si205(OH)4(H20)o.64), this being identical to the phase observed through water washing of the ethylene glycol intercalated kaolinite and through methanol washing of the dimethylsulfoxide intercalated kaolinite. Between the temperatures 370 to 1000°C, a mass loss of 13.72% is observed in the TG curve to which one endothermic peak related do the dehydroxylation process at 524°C and one characteristic exothermic peak at 989°C, were observed. The total residue of 82.8% is close to the expected value of 83.2% for the proposed stoichiometry. As the organic molecules of the described phases are eliminated before the beginning of the process of the matrix decomposition, being heated until a certain temperature, kaolinite can be fully recovered, although it looses crystallinity as a consequence of the intercalation reaction.
Figure 34 - FTIR measurements of (a) the raw kaolinite, (b) urea intercalated kaolinite in the proportion of 20%, (c) hydrated kaolinite and (d) pure urea. [Reprinted from Quimica Nova, (24, 6, 761, 2001), with Permission from Sociedade Brasileira de Quimica] [111]. After the intercalation reaction (Fig. 34(b)), the region related to the hydroxyl groups in the FTIR spectra (between 3200 and 3800cm"1) was well affected with the disappearance of the external hydroxyl groups' bands at 3651 and 3669cm"1 and a decrease of the intensity of the band at 3698cm"1. This was slightly moved to larger wavenumbers in relation to the raw kaolinite. Even then, the relative bands for internal hydroxyl groups were maintained at 3619cm"1 and the appearance of new bands at 3504, 3411 and 3388cm*1. Those new bands are associated with the bonding of the urea molecule to the external hydroxyl groups of the kaolinite layer. This was another evidence for the urea interaction with the interlayer hydroxyl groups. The bands observed at 1467, 1617 and 1686cm"1 in the pure urea have moved up to 1475, 1590, 1622 and 1673cm"1 in the intercalated compound. The band attributed to the surface hydroxyl groups has moved from 914cm"1 in the kaolinite to 903cm"1 in the urea
Chemical Modification of Clay Surfaces
47
intercalated sample. The other bands did not undergo significant changes. In the hydrated phase (Fig. 34(c)) a characteristic profile was observed with the maintenance of the relative band for internal hydroxyl groups (3618cm"1), a shoulder at 3649 and a band at 3692cm"1 (relative to the external hydroxyl groups) apart from the appearance of new bands at 3600 and 3556cm"1. The small band at 1655cm"1 is attributed to the deformation band of the water molecule. The band attributed to the surface hydroxyl groups returned practically to the original position at 912cm"1. However, a new band was observed at 965cm"1, which demonstrated the interaction between the intercalated water molecules and the external hydroxyl groups. The possibility that some product from the decomposition of the urea could remain between the layers cannot be excluded. However, after washing the sample with hot water, the hydrated kaolinite appeared as a final product. This phase was identical to the previously reported one, as confirmed by powder X-ray diffraction, FTIR and thermal analysis. The hydrated kaolinite should not be confused with halloysite, since the two differ in some aspects, although some coincidences can be cited. The most outstanding difference is related with the habit of the crystals. Hydrated kaolinite presents itself in the form of layered crystals while halloysite (with two molecules of hydration water also known as Endellite and anhydrous) is tubular. In spite of this difference, halloysite can show an uncoiling of the crystals after dehydration, turning into layered structure. The stoichiometries between halloysite (Al2Si205(OH)4(H20)2,o and hydrated kaolinite (Al2Si2O5(OH)4(H2O)0)64) are very different, although variable amounts of water can be observed during the dehydration process of halloysite to anhydrous halloysite. The same happens to the basal spacings that are very different [(halloysite=10.lA [127] and hydrated kaolinite (unstable 10.lA and stable 8.4A)] [126,128-130], although intermediate spacings between the two extremes can be identified by "in-situ" X-ray diffraction measurements during the process of dehydration of halloysite. The hydrated kaolinite and halloysite regenerate the kaolinite (or anhydrous tubular halloysite) after total dehydration. The observed bands in the FTIR spectra in the region between 3200 and 3800cm"1 differ only in the intensity, although that characteristic can be explained by the variable amount of interlayer water (halloysite with two molecules of hydration 3695, 3620, 3602 and 3550cm"1 and hydrated kaolinite - 3692, 3649, 3619, 3600 and 3556cm"1). In the region between 2000 and 400cm"1, although the differences are very subtle, the spectra are closer to kaolinite than to halloysite. The bands in the FTIR spectrum of raw kaolinite (Fig. 34(a)) attributed to the external hydroxyl groups were observed at 3694, 3669 and 3651cm"1 and those attributed to the internal hydroxyl groups at 3619cm"1 [131-133]. Although hydrated kaolinite was never reported to occur in nature, this appearance could be predicted mainly in mixed deposits of halloysite/kaolinite. Hydrated kaolinite can be used to prepare long chain amine derivatives [115] and aminoacids [134] that are similar to hydrophobic organo layered double hydroxides [96-99]. These compounds are very easy to prepare and can be an interesting alternative for the environmental cleanup and remediation of contaminated soils, groundwater and industrial effluents that are resistant to biological degradation. 6 - Smectite derivatives In the smectites, the reactions that are more different from kaolinite are related with the isomorphic substitution, which attribute to smectite a higher capacity of cationic exchange. Due to presence of planes of oxygen atoms on both sides of the layer
48
F.Wypych
(siloxane surface), structural modifications through grafting reaction that are quite important for kaolinite do not play an important role for the preparation of new smectite derivatives (structure type 1:2). On the other hand, similar materials can be obtained through the thermal treatment of intercalated smectites with some specific compounds producing pillared clays, with several applications as catalysts and adsorbent materials. Apart from the reactions of ionic exchange, many of the reactions described for the other systems can be applied to the smectites as well. 6.1 - Cationic exchange reactions The importance of the cationic exchange reactions will not be described in this work in full detail, but it can be mentioned that those processes are extremely important sources of essential cations for the growth of plants, apart from many important industrial applications [23,135]. The reactions proceed through the unsaturated bonds on the crystal edges (silanol and aluminol bonds), replacement of the cations in the crystalline lattice, replacement of the hydrogen ions of the hydroxyl groups with other cations or through the replacement of the interlayer cations. Although the crystallite size can influence the capacity of ionic exchange attributed to the edges of the crystals, the process of interlayer cations exchange can be considered predominant. Apart from the process of ionic exchange, other solvents can replace the layers of solvation of the intercalated ions in a similar way to those described for the hydroxysalts (Eq. 13,14), including natural and synthetic polymers (Eq. 30). When the hybrid materials are used as fillers in polymers [136-144], those phases allow a larger interaction among the phases and in some cases, a total exfoliation inside the matrix of the polymer can take place. Those processes allow a great improvement of the mechanical properties of the reinforced polymers and help save polymer consumption and reduce costs as well. Clay(A+)x(H2O)y + polymer —> nanocomposites
(Eq. 30)
6.2 - Thermal reactions Depending on the thermal treatment, the dehydration processes or a reversible removal of solvents of the intercalated cations can take place (Eq. 31,32). After increasing the temperature, the dehydration is followed by a process of dehydroxylation of the clay mineral matrix, producing a mixture of amorphous or crystalline oxides, depending on the involved temperature (Eq. 33). Clay(A+)x(H2O)y <-> Clay(A+)x + y H2O (-» heating and <- presence of water) (Eq.31) Clay(A+)x(solv)z <-> Clay(A+)x + z solv ( -> heating and <— presence of solvent) (Eq. 32) Clay(A+)x(H2O)y -> oxides mixtures (high temperatures)
(Eq. 33)
Adopting an appropriate strategy of the intercalation of key cationic complexes (Ex.: salts of Al, Ti, etc) and depending on the involved thermal treatment, a process of decomposition of the intercalated compound can take place, leaving the structure of the
Chemical Modification of Clay Surfaces
49
clay mineral almost intact. A reaction between the clay mineral layer with the product of decomposition of the complex material occur, which produces pillars between the layers, generating a class of extremely important catalyst known as "pillared clays" (Eq. 34) [145,146]. Clay(A+)x(H2O)y -> pillared clays (presence of adequate cationic complexes and controlled temperatures) (Eq. 34). 6.4 — Grafting reactions Clays from the smectite groups are used as fillers in polymers [136-144]. As the interaction between the hydrophilic clay and hydrophobic polymer is poor, it is difficult to disperse and the reinforcing property is not maximized. In order to improve the interaction between the phases, hydrophilic phyllosilicate surfaces can be modified through grafting of organic groups, silanes being the most widely employed material [147-150]. The reaction is similar to kaolinite surface (Section 5.3), but restricted to crystal edges and basal defects, as both sides of the smectite layers are covered in oxygen atoms. In some cases the clays are subjected to acidic activation that leaches octahedral aluminum from the structure, producing more available sites to be reacted with the silane [147]. The surface modification with a silane coupling agent is shown in Eq. 35 (R denotes an organic group that can interact (or react) with the polymer and improve the interactions between the phases. Clay/-3OH + RSi(OH)3 -> Clay/-O3Si-R + 3 H2O
(Eq. 35).
Modifying the terminal group R, this kind of surface modified material can be used to immobilize several interesting compounds going from enzymes, catalysts (oxidation catalysts based on metalloporphyrins is one example), pigments, surfactants, etc. 6.3 - Exfoliation and preparation of nanocomposites One of the most important characteristics of the smectites is that, depending on the energy of hydration of the intercalated cations, those can be intercalated with one, two, three or more layers of hydration. The distribution of the layers of water in the smectites can be regular or a randomic mixture of different hydrate forms and depends on the relative humidity that the sample was exposed to and the solvation energy of the cation. As the interaction between layers are through the intercalated cations, the tendency of those materials are to be presented in low ordinate structures which hinders the processes of characterization especially through X-ray diffraction. The structure determination is normally obtained by X-ray powder diffraction patterns using specific programs, when single crystals are usually not available. When a colloidal particle of the mineral clay of the smectite group is put in contact with water (through a dispersing agent or otherwise), that is associated to the hydrated exchangeable cations which interact strongly with water molecules and counter-ions of the solution. The ions are not directly bound to the surfaces but build a diffuse layer of ions around the colloidal particles. In special conditions (usually in the presence of ions of phosphate, silicate, hydroxide, etc.), a sol or gel can be reversibly stabilized. As attractive forces act between the particles, the colloidal dispersion can be destabilized specially by the influence of salts or pH. The coagulated particles can
50
F.Wypych
interact in different ways (surface-surface, surface-edge or edge-edge), producing morphologies that depend on the method employed in the drying procedure. The smectites derived organophyllic gels have wide application as components of drilling muds for the perforation of oil wells and in cosmetics, toiletries, lubricants, adhesives, paint and other related industries (Eq. 35) [151-157]. Clay(A+)x(H2O)y -+ sols or gels
(Eq. 35)
Although the single-layer suspensions can potentially be used for the nanocomposites preparation or dispersion in polymers, few reports are described in the specialized literature [158,159]. Most of the examples that involve clay gels and polymers are centered around rheological behavior and studies related with the interactions of the soluble polymer molecules and the clay single layers [160-165]. Recently, hectorite [76] and other layered compounds [158] were used as fillers in order to improve the mechanical properties of glycerol plasticized starch films. The films were characterized by several techniques. Dynamic mechanical analyses have shown that the composite films present three relaxation processes, attributed to glass transition of glycerol rich phase; water loss including the interlayer water molecules from the clay structure and the starch rich phase. The film with 30% in mass of hectorite showed an increase of more than 70% in Young's modulus compared to non-reinforced plasticized starch. Both X-ray diffraction and infrared spectroscopy have shown that glycerol can be intercalated into the clay galleries and there is a possible conformational change of starch molecule in the plasticized starch/clay composite films. In the unplasticized clay/starch mixtures, the clay is almost totally exfoliated in the starch matrix. These data show that the glycerol/starch interactions, hinders the exfoliation process of the clay. 7 - Concluding remarks Although clay minerals represent a small fraction in the layered compounds family, their chemistry and physics are very rich and fascinating. Despite the efforts that have been made during the last few decades, many aspects related to their reactivity and properties are still obscure. Studies centered about those compounds, be they of synthetic or natural origin have been exciting to the researchers due to their applications flexibility. It can be safely predicted with certainty that clay minerals will reveal in the near future their tremendous potential, especially in the formulation of new and exciting compounds having bearing on important industrial applications. Apart from clay minerals, mono dimensional materials like natural phyllosilicates from the serpentine sub-group (serpentine-kaolin group) will be used to prepare nanocomposites with unthinkable properties. These materials will achieve improved properties specially when functionalized surfaces will be employed to help in the compatibility of the polymeric matrix with the nanoparticles surfaces and also to increase the strength of the interface. Oriented nanotubes within the matrix will produce anisotropic mechanical properties and researches will resurrect banned materials like chrysotile, as ideal material for this purpose. Aknowledgements The author gratefully acknowledges CNPq, PRONEX/MCT, Fundacao Araucaria (Brazilian agencies), his undergraduate and graduate students as well as research co-workers.
Chemical Modification of Clay Surfaces
51
8 - References [I] R. Schollhorn, in: Progress in Intercalation Research, ed. by W. Miiller-Warmuth. and R. Schollhorn, Kluver Academic Publishers, Dordrecht, 1994. [2] R. Schollhorn, in: Intercalation Chemistry, ed. by M.S. Whittingham and AJ. Jacobson, Academic Press, New York, 1982. [3] K. Nakamura, S.I. Hirano and S. Somiya, J. Am. Ceram. Soc, 58, 7 (1975) 349. [4] Y. Ding, G.T. Zhang, H. Wu, B. Hai, L. Wang and Y Qian, Chem. Mat., 13, 2 (2001)435. [5] http://sdpd.univ-lemans.fr/cod/ and http://www.ccdc.cam.ac.uk/prods (Mercury Crystal Structure Visualisation Available for free download) [6] S. Ghose, Acta Cryst., 17, 8 (1964) 1051. [7] W. Stahlin and H.R. Oswald, Acta Cryst. B., 26 (1970) 860. [8] N. Masciocchi, E. Corradi, A. Sironi, G. Moretti, G. Minelli and P. Porta, J. Solid State Chem., 131 (1997)252. [9] W.T. Reichle, Solid State Ionics, 22, 1 (1986) 135. [10] R. Allmann, Acta Crystall. B, 7 (1968) 972. [II] A.S. Prakash, P. Vishnu Kamath and M.S. Hegde, Mat. Res. Bull, 35, 13 (2000) 2189. [12] E. Kanezaki, Solid State Ionics, 106, 3-4 (1998) 279. [13] A.J. Marchi, C.R. Apesteguia, Appl. Clay Sci., 13, 1 (1998) 35. [14] I. Crespo, C. Barriga, V. Rives and M.A. Ulibarri, Solid State Ionics, 101-103 (1997) 729. [15] M.Z. Bin Hussein, Z. Zainal, A.H. Yahaya and A.A. Aziz, Mat. Sci. Eng. B, 88, 1 (2002) 98. [16] M. Badreddine, A. Legrouri, A. Barroug, A. De Roy and J.P. Besse, Mat. Lett., 38, 6(1999)391. [17] J.X. He, K. Kobayashi, M. Takahashi, G. Villemure and A. Yamagishi, Thin Solid Films, 397,1-2(2001)255. [18] R.E. Grim, Clay Mineralogy, McGraw-Hill Book Company, Inc., New York, 1953. [19] G.W. Brindley and G. Brown, in: Crystal Structure of Clay Minerals and their XRay Identification, Mineralogical Society, London, 1980. [20] K. Kuroda, K. Hiraguri, Y. Komori, Y. Sugahara, H. Mouri and Y. Uesu, Chem. Commun., 22 (1999) 2253. [21] R. Takenawa, Y. Komori, S. Hayashi, J. Kawamata and K. Kuroda, Chem Mat., 13, 10 (2001)3741. [22] W. Zech, N. Senesi, G. Guggenberger, K. Kaiser, J. Lehmann, T.M. Miano, A. Miltner and G. Schroth, Geoderma, 79, 1-4 (1997) 117. [23] E.H. Novotny, W.E.H. Blum, M.H. Gerzabek and A.S. Mangrich, Geoderma, 92, 1-2 (1999) 87. [24] R. Prins, V.H.J. De Beer, and G.A. Somorjai, Cat. Rev. Sci. Eng., 31 (1-2) (1989) 1. [25] R. Cattaneo, F. Rota and R. Prins, J. Catal, 199,2 (2001) 318. [26] M. Inoue, K. Kitamura, H. Tanino, H. Nagayama and T. Inui, Clays Clay. Miner., 37(1989)71. [27] S. Ogata, I. Miyazaki, Y. Tasaka, H. Tagaya, J. Kadokawa and K. Chiba, J. Mat. Chem., 8, 12(1998)2813. [28] S. Ogata, H. Tagaya, M. Karasu and J. Kadokawa, J. Mat. Chem., 10, 2 (2000) 321.
52
F.Wypych
[29] S. Ogata, Y. Tasaka, H. Tagaya, J. Kadokawa and K. Chiba, Chem. Lett., 3 (1998) 237. [30] A.V. Besserguenev, A.M. Fogg, R.J. Francis, S.J. Price, D. O'Hare, V.P. Isupov and B.P. Tolochko, Chem. Mat., 9, 1 (1997) 241. [31] K.R. Poeppelmeier and S.J. Hwu, Inorg. Chem., 26 (1987) 3297. [32] M. Rajamathi, P.V. Kamath and R. Seshadri, Mat. Res. Bull., 35, 2 (2000) 271. [33] F. Wypych, W.H. Schreiner, R. Marangoni, J. Colloid Interface Sci., 253 (2002) 180. [34] Joint Committee on Powder Diffraction Standards (JCPDS) file number 45-946. [35] R.L. Frost and J.T. Kloprogge, Spectroch. Acta A. 55 (1999) 2195. [36] P. Aranda and E. Ruiz-Hitzky, App. Clay Sci., 15, 1-2 (1999) 119. [37] B.L. Papke, M.A. Ratner and D.F. Shriver, J. Phys. Chem. Solids, 42 (1981) 493. [38] K. Fukushima and H. Matsuura, J. Molec. Struct. 350, 3 (1995) 215. [39] B. Herreros, T.L. Barr and J. Klinowski, J. Phys. Chem., 98, 3 (1994) 738. [40] J.L. Guimaraes, R. Marangoni, L.P. Ramos and F. Wypych, J. Colloid Interface Sci., 227(2000)445. [41] M. Inoue, Y. Kondo and T. Inui, Inorg. Chem., 27, 2 (1988) 215. [42] J.G. Thompson and C. Cuff, Clays Clay Miner., 33, 6 (1985) 490. [43] R.L. Frost, D.A. Lack, G.N. Paroz and T.H.T. Tran, Clays Clay Miner., 47, 3(1999)297. [44] R.L. Frost, J. Kristof, G.N. Paro and J.T. Kloprogge, Phys. Chem. Miner., 26, 3 (1999)257. [45] S. Olejnik, A.M. Posner and J.P. Quirk, Clays Clay Miner., 19, 2 (1971) 83. [46] R.L. Frost, J. Kristof, E. Horvath and J.T. Kloprogge, J. Colloid Interface Sci., 214, 1 (1999) 109. [47] A. Gehin, C. Ruby, M. Abdelmoula, O. Benali, J. Ghanbaja, P. Refait and J.R. Genin Solid State Ionics, 4, 1 (2002) 61. [48] L. Legrand, M. Abdelmoula, A. Gehin, A. Chausse and J.M.R. Genin, Electroch. Acta, 46, 12(2001)1815. [49] J.R. Gancedo, M.L. Martinez and J.M. Oton, An. Quim. A - Fis.Tec, 79, 3(1983) 470. [50] I.R. Mcgill, B. Mcenaney and D.C. Smith, Nature, 259 (1976) 200. [51] D. Louer, J. Rius J, P. Benard-Rocherulle and M. Louer, Powder Diffraction, 16, 2 (2001) 86. [52] M. Rajamathi and P. Vishnu Kamath, Int. J. Inorg. Mat., 3 (2001) 901. [53] H. Tagaya, N. Sasaki, H. Morioka, et al, Polyedron, 19, 3 (2000) 259. [54] R. Marangoni, G.A. Bubniak, M.P. Cantao, M. Abatte, W.H. Schreiner and F. Wypych, J. Colloid Interface Sci., 240 (2001) 245. [55] V. Laget, C. Hornick, P. Rabu, and M. Drillon, J. Mat. Chem., 9, 1 (1999) 169. [56] V. Laget, M. Drillon, C. Hornick, P. Rabu, F. Romero, P. Turek and R. Ziessel, J. Alloy Compds., 262 (1997) 423. [57] W. Fujita and K. Awaga, Inorg. Chem., 35 (1996) 1915. [58] W. Fujita, K. Awaga and T. Yokoyama, Appl. Clay Sci., 15 (1999) 281. [59] W. Fujita, K. Awaga and T. Yokoyama, T., Inorg. Chem., 36 (1997) 196. [60] W.M.R. Divigalpitiya, R.F. Frindt and S.R. Morrison, Science, 246 (1989) 369. [61] M.A. Gee, R.F. Frindt, P. Joensen and S.R. Morrison, Mat. Res. Bull, 21 (1986) 543. [62] P. Joensen, R.F. Frindt and S.R. Morrison, Mat. Res. Bull., 21 (1986) 457.
Chemical Modification of Clay Surfaces
53
[63] M.G. Kanatzidis, R. Bissessur, D.C. DeGroot, J.L. Schindler and C.R. Kannewurf, Chem. Mat, 5(1993)595. [64] R. Bissessur, M.G. Kanatzidis, J.L. Schindler, et al, J. Chem. Soc. Chem. Commun., 20 (1993) 1582. [65] R. Bissessur, J. Heising, W. Hirpo and M.G. Kanatzidis, Chem. Mat., 8 (1996) 318. [66] J. Heising, F. Bonhomme, M.G. Kanatzidis, J. Solid State Chem., 139 (1998) 22. [67] F. Wypych, L.B. Adad and M.C. Grothe, Quim. Nova, 21, 6 (1998) 687 (in Portuguese). [68] F. Wypych, N. Seefeld and I. Denicolo, Quim. Nova, 20, 4 (1997) 356 (in Portuguese). [69] S. Nakagaki, A.S. Mangrich, F. Wypych, Inorg. Chim. Acta, 254 (1997) 213. [70] R. Bissessur, J.L. Schindler, C.R. Kannewurf, C.R., et. al., Mol. Cryst. Liq. Cryst. A., 244(1994)249. [71] M. Alexandre and P. Dubois, Mat. Sci. Eng., 28 (2000) 1. [72] T.L. Porter, D. Thompson, M. Bradley, M.P. Eastman, M.E. Hagerman, J.L. Attuso, A.E. Votava and E.D. Bain, J. Vac. Sci. Technol. A, 15 (1997) 500. [73] G. Lagaly, Appl. Clay Sci., 15 (1999) 1. [74] P.C. Lebaron, Z. Wang and T.P. Pinnavaia, Appl. Clay Sci., 15 (1999) 11. [75] P. Aranda, E. Ruiz-Hitzky, Chem. Mat., 4 (1992) 1395. [76] H.M. Wilhelm, M.R. Sierakowski, G.P. Souza and F. Wypych, Carbohyd. Polym., 52(2003)101. [77] J.J. Tunney and C. Detellier, Chem. Mat., 5 (1993) 747. [78] J.J. Tunney and C. Detellier, Clays Clay Miner., 42 (1994) 552. [79] M. Inoue, Y. Kondo and T. Inui, Chem. Lett., 9 (1986) 1421. [80] M. Inoue, H. Tanino, Y. Kondo and T. Inui, Clays Clay Miner., 39 (1991) 151. [81] M. Inoue, H. Kominami, Y. Kondo and T. Inui, Chem. Mat. 9, 7 (1997) 1614. [82] H. Nijs, A. Clearfield and E.F. Vansant, Microp. Mesop. Mat., 23, 1-2 (1998) 97. [83] S. Carlino, Solid State Ionics, 98, 1-2 (1997) 73. [84] K.A. Tarasov, V.P. Isupov, B.B. Bokhonov, Y.A. Gaponov, B.P. Tolochko, M.R. Sharafutdinov and S.S. Shatskaya, J. Mat. Synt. Proces., 8 (2000) 21. [85] I. Dekany, L. Turi, A. Szucs and Z. Kiraly, Coll. Surf. A, Phys. Eng. Asp., 141 (1998)405. [86] O. Santini, D.H. Mosca, W.H. Schreiner, R. Marangoni, J.L. GuimarSes, F. Wypych and A.J A. de Oliveira, J. Phys. D: Appl. Phys, 36 (2003) 428. [87] M. Adachi-Pagano, C. Forano and J.P. Besse, J. Chem. Commun., 1 (2000) 91. [88] T. Hibino and W. Jones, J. Mat. Chem., 11, 5 (2001) 1321. [89] G.A. Bubniak, W.H. Schreiner, N. Mattoso and F. Wypych, Langmuir, 18 (2002) 5967. [90] M. Khaldi, A. De Roy, M. Chaouch and J.P. Besse, Ann. Chim. Sci. Mat., 23, 1-2 (1998)337. [91] M.Z.B. Hussein, Z. Zainal and C.Y. Ming, J. Mat. Sci. Lett., 19, 10 (2000) 879. [92] A. Clearfield, M. Kieke, J. Kwan, J.L. Colon and R.C. Wang, J. Incl. Phen. Mol. Recog., 11,4(1991)361. [93] T. Yoshihara, S. Murahashi and H. Tadokoro, J. Phys. Chem., 41, 9 (1964) 2902. [94] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, John Wiley & Sons, New York, 1986. [95] J. Olanrewaju, B.L. Newalkar, C. Mancino and S. Komarneni, Mat. Lett., 45, 6 (2000) 307.
54
F.Wypych
[96] Y. Youwen, H. Zhao, G.F. Vance, Colloid Surf. A: Phys. Engen. Aspects, 205 (2002)161. [97] K. Esumi, S. Yamamoto, Colloid Surf. A, 137, 1-3 (1998) 385. [98] M.V. Villa, M.J. Sanchez-Martin, M. Sanchez-Camazano, J. Environ. Sci. Heal. B, 34, 3 (1999) 509. [99] R. Celis, W.C. Koskinen, A.M. Cecchi, et al, J. Environ. Sci. Heal. B, 34, 6 (1999) 929. [100] J.E. Gardolinski, L.C.M. Carrera, M.P. Cantao and F. Wypych, J. Mat. Sci., 35 (2000)3113. [101] Y. Sugahara, S. Satokawa, K. Kuroda and C. Kato, Clays Clay Miner., 38 (1990) 137. [102] Y. Sugahara, S. Kitano, S. Satokawa, K. Kuroda and C. Kato, Bull. Chem. Soc. Jpn., 59(1986)2607. [103] Y. Komori, Y. Sugahara and K. Kuroda, Chem. Mat., 11 (1999) 3. [104] J.J. Tunney and C. Detellier, Chem. Mat., 8 (1996) 927. [105] J.L. Guimaraes, P. Peralta-Zamora and F. Wypych, J. Colloid Interface Sci., 206 (1998)281. [106] C. Breen, N. D'Mello and J. Yarwood, J. Mater. Chem., 12,2 (2002) 273. [107] E. Tombacz, M. Szekeres, L. Baranyi and E. Micheli, Colloids and Surf.,A., 141 (1998)379. [108] D. Ghosh and K.G. Bhattacharyya, Appl. Clay Sci., 20 (2002) 295. [109] R.L. Frost, J. Kristof, G.N. Paroz, T.H. Tran and J.T. Kloprogge, J. Colloid Interface Sci., 204 (1998) 227. [110] K. Tsunematsu and H. Tateyama, J. Am. Ceram. Soc, 82 (1999) 1589. [111] J.E. Gardolinski, M.P. Cantao and F. Wypych, Quim. Nova, 24, 6 (2001) 761 (in Portuguese). [112] J.E. Gardolinski, L.P. Ramos and F. Wypych, J. Colloid Interface Sci., 221 (2000) 284. [113] Y. Sugahara, S. Satokawa, K. Yoshioka, K. Kuroda and C. Kato, Clays Clay Miner., 37 (1989) 143. [114] Y. Komori, Y. Sugahara and K. Kuroda, J. Mat. Res., 13 (1998) 930. [115] Y. Komori, Y. Sugahara and K. Kuroda, Appl. Clay Sci., 15 (1999) 241. [116] J.E. Gardolinski, P. Peralta-Zamora and F. Wypych, J. Colloid Interface Sci., 211 (1999) 137. [117] T. Itagaki, Y. Komori, Y. Sugahara and K. Kuroda, J. Mat. Chem., 11 (2001) 3291. [118] Hyperchem version 4.5. Hypercube Inc., http://www.hyper.com [119] S. Olejnik, L.A.G. Aylmore, A.M. Posner and J.P. Quirk, J. Phys. Chem., 72, 1 (1968)241. [120] E. Akiba, H. Hayakawa, S. Hayashi, R. Miyawaki, S. Tomura, Y. Shibasaki, F. Izumi, H. Asano and T. Kamiyama, Clays Clay Miner., 45 (1997) 781. [121] A. Weiss, W. Thielepape, H. Schafew and G. Z. Anorg. Allg. Chem., 320 (1963) 183. [122] R.L. Frost and U. Johansson, Clay Clay Miner., 46, 4 (1998) 466. [123] J. Kristof, R.L. Frost, A. Felinger, et al, J. Mol. Struct., 410 (1997) 119. [124] R.L. Frost and A.M. Vassallo, Clays Clay Miner., 44, 5 (1996) 635. [125] Q. Hao, X.J. Yang, Y. Wang, L. Lu and X. Wang, Spec. Spectra Anal, 20, 3 (2000) 302.
Chemical Modification of Clay Surfaces
55
[126] J.J. Tunney and C. Detellier, Clays Clay Miner., 42 (1994) 473. [127] I. Bobos, J. Duplay, J. Rocha, and C. Gomes, Portugal, Clays Clay Miner., 49, 6 (2001) 596. [128] P.M. Costanzo, R.F. Giese and C.V. Clemency, Clays Clay Miner., 32, 1(1984) 29. [129] R. Raythatha and M. Lipsicas, Clays Clay Min., 33 (1985) 333. [130] P.M. Costanzo, C.V. Clemency and R.F. Giese Jr., Clays Clay Miner., 28 (1980) 155. [131] S. Shoval, S. Yariv, K.H. Michaelian, I. Lapides, M. Boudeuille and G. Panczer, J. Colloid Interface Sci., 212 (1999) 523. [132] R.L. Frost, Clays Clay Miner., 46, 3 (1998) 280. [133] R.L. Frost and A.M. Vassallo, Clays Clay Miner., 44, 5 (1996) 635. [134] M. Sato, Clays Clay Miner., 47, 6 (1999) 793. [135] P.W. Faguy, W. Ma, J.A. Lowe, W. Pan, and T. Brown, J. Mat. Chem., 4 (1994) 771. [136] K.A. Carrado and L. Xu, Microp. Mesop. Mat., 27 (1999) 87. [137] D.C. Lee and L.W. Jang, J. Appl. Pol. Sci., 68 (1998) 1997. [138] M.S. Wang and T.J. Pinnavaia, Chem. Mat., 6 (1994) 468. [139] T. Lan and T.J. Pinnavaia, Chem. Mat., 6 (1994) 2216. [140] T. Lan, P.D. Kaviratna and T.J. Pinnavaia, Chem. Mat., 7 (1995) 2144. [141] K.A. Carrado and L. Xu, Chem. Mat., 10, 5 (1998) 1440. [142] A. Okada and A. Usuki, Mat. Sci. Eng. C3 (1995) 109. [143] Y. Kurokawa, H. Yasuda, M. Kashiwagi and A. Oyo, J. Mat. Sci. Lett., 16 (1997) 1670. [144] M. Biswas and S.S. Ray, Adv. Polym. Sci., 155 (2001) 167. [145] D.E.W. Vaughan, Cat. Today, 2 (1988) 187. [146] K. Sapag and S. Mendioroz, Coll. Surf., A, 187 (2001) 141. [147] J. C. Dai and J. T. Huang, Applied Clay Sci., 15 (1999) 51. [148] P.K. Pal and S.K. DE, Rubber Chem. Tech., 56, 4 (1983) 737. [149] T. Seckin, A. Gultek, M.G. Icduygu and Y. Onal, J. Appl. Polym. Sci., 84, 1 (2002) 164. [150] K. Song and G. Sandi, Clays Clay Miner., 49, 2 (2001) 119. [151] L. Bailey, M. Keall, A. Audibert and J. Lecourtier, Langmuir, 10, 5 (1994) 1544. [152] B.J. Briscoe, P.F. Luckham and S.R. Ren, Phil. Trans. Royal Soc. London, AMat. Phys. Eng. Sci., 348, 1686 (1994) 179. [153] S. Abend and G. Lagaly, Appl. Clay Sci., 16, 3-4 (2000) 201. [154] P.F. Luckham and S. Rossi, Adv. Coll. Interf, 82, 1-3 (1999) 43. [155] P.K. Singh and V.P. Sharma, Energ. Source, 13, 3 (1991) 369. [156] V.P. Sharma, S. Laik and S. Srinivasan, Res. Ind., 31, 3 (1986) 230. [157] B. Bloys, N. Davis, B. Smolen, et al, Oilfield Ver., 6, 2 (1994) 33. [158] H.M. Wilhelm, M.R. Sierakowski, G.P. Souza and F. Wypych, Polym. Int., 52 (2003) 1035. [159] K.A. Carrado, P. Thiyagarajan and K. Song, Clay Miner., 32, 1 (1997) 29. [160] J. Lai and L. Auvray, Mol. Cryst. Liq. Cryst, 356 (2001) 503. [161] J. Lai and L. Auvray, J. Appl. Cryst., 33, 1 (2000) 673. [162] O.I. Ece, N. Gungor and A. Alemdar, J. Incl. Phen. Macr. Chem., 33, 2 (1999) 155.
56
F.Wypych
[163] D.Y. Gao, R.B. Heimann, M.C. Williams, L.T. Wardhaugh and M. Muhammad, J. Mat. Sci., 34, 7(1999)1543. [164] J. Dau and G. Lagaly, Croat. Chem. Acta, 71,4 (1998) 983. [165] M.V. Smalley, H. Jinnai, T. Hashimoto and S. Koizumi, Clays Clay Miner., 45, 5 (1997)745.
ELECTROKINETIC BEHAVIOR OF CLAY SURFACES MEHMET SABRI CELIK Istanbul Technical University Mining Engineering Dept., Mineral Processing Section Ayazaga 34469 Istanbul - TURKEY E-mail:
[email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
58
M.S. Çelik
1 - Introduction Clay minerals are hydrous silicates or alumina silicates and constitute a major proportion of soils, sediments, rocks and waters [1]. Suspended particulate matter, an important component of natural waters and ore dewatering systems, is largely composed of various clay particles which often control the kinetics and deposition of different organic and inorganic species. Interfacial phenomena at clay/water interface are usually controlled by electrokinetics properties including zeta potential (zp), the structure of electrical double layer (EDL), surface potential, and isoelectric point (iep). The electrokinetic properties of a substance, inorganic or organic, are used to explain the mechanism of dispersion and agglomeration in a liquid phase and to identify the adsorption mechanisms of ions or molecules at a solid-liquid interface. They, therefore, play an important role in a spectrum of applications including ceramics, mining, paper, medicine, water and wastewater treatment and emulsions. A larger quantity of literature on electrokinetics of mineral particles is available [2-6]. This chapter presents an overview of the electrokinetic properties of clay minerals, to elucidate the electrokinetic behavior of clay surfaces and the mechanism of particle-particle interactions in aqueous systems. 2 - Electrokinetic properties 2.1 - Zeta Potential Zeta (Q potential is an intrinsic property of a mineral particle in a liquid. It determines the strength of the EDL repulsive forces between particles and identifies the stability of a colloidal system. The zeta potential (zp) is known as the measurable surface potential of a particle viz., the potential at the shear plane. There is no direct experimental method for determining both the surface potential ((p0) and Stern layer potential (q>5) [7]. So far the exact position of the shear plane within the diffuse layer of the EDL could not be determined, but it is assumed that the position of the shear plane is very close to the outer Helmholtz plane (OHP) [8]. The C, potential is fairly close to the Stern potential, cp5, in magnitude, and definitely less than the potential at the surface, cp0. The conventional position of the shear plane is usually thought to be two to three water molecule diameters, or about 5 A from the surface of particle. However, a recent study by Li et al [9] particularly for swelling type clay minerals has shown that the shear plane is closer to the Gouy plane as illustrated in Figure 1. Accordingly, the distance between the shear plane and Stern plane may be about 200-300 A in the presence of 10"4 M of 1:1 electrolyte. Application of this concept using the Gouy-Chapman theory provides a reasonable prediction to the swelling of clay minerals. More importantly, calculation of the Stern potential (cp6) from known values of crs (the surface charge density) using the Gouy-Chapman theory revealed that the zeta potential cannot be approximated to the surface potential [9]. jlRTecx, , ,
CT5
fZFnX
= V —^— smhV^^7^.J
0)
where c0 is the concentration of the electrolyte in bulk solution, s is the dielectric constant of the medium, Z and F are the valency and Faraday constant, respectively. Li et al [9] have also shown that the swelling pressure (P) of the electrostatic repulsion in the diffuse double layer between two adjacent particles can be calculated from the following equation:
Electrokinetic Behavior of Clay Surfaces
F = 2*r,{co sh (^fi)-i]
59
(2)
where cpd is the potential at the overlapping point of the diffuse double layer. The relation between swelling pressure and interlayer distance is presented in Fig. 2
Figure 1 - A diagram illustrating the position of the shear plane in a diffuse layer [9].
Figure 2 - The relationship between the swelling pressure (P) and the interlayer distance (X) as determined experimentally and by double layer theory using different values of(q>^: (—) theoretical curve with cps = - 270 mV; (0) theoretical curve with cps ~ £, = - 55.6 mV; (%) experimental data [9J.
60
M.S. Çelik
2.2 - Potential Determining Ions The interface between the solid and solution may be treated as a semipermeable membrane which allows only the charged species common to both the solid and the solution to pass through. These species are called potential determining ions (pdi). They are the major ions responsible for the establishment of the surface charge of particle [7]. Their activities in the liquid play a crucial role in the generation of potential difference across a solid-liquid interface. They are also able to reverse the sign of zp of the solid. As a simple recipe, for a cation to be the pdi, it must make the surface more positive upon increasing the cation concentration. Similarly, for an anion to be the pdi it must impart the surface more negative charges with increasing the anion concentration. The pdi for ionic solids such as Agl, BaSO4, and CaCO3 etc. are the lattice constituent ions, i.e. Ag+, I", Ba2+, SO42", Ca2+, CO32; whereas H+ and OH" ions are for metal oxides and hydroxides, silicate or clay minerals, some hydrophobic minerals (e.g., coal) and some synthetic polymers with sulfate groups [2,10]. 2.3 - Surface Potential Surface potential is theoretically the potential at zero distance from the surface. It is the highest potential and exhibits an exponential decay with distance. For oxides and silicates, for example, the surface potential is determined by the activities of pdis and expressed by Nernst-like equation q)0 = (RT/F)ln(a H+ /a 0 H+ )
(3) +
where F is the Faraday constant, aH+ is the activitiy of H and a°H+ refers to the point of zero charge (pzc). The operational formula for aqueous solutions at 25 °C is (p0 = 0.059 (pHc - pH) volts where pH° refers to the pzc [11]. Although the isoelectric point (iep) and the point of zero charge (pzc) are identical by definition, there are some differences between them. While the pzc denotes the state in which the net surface charge of the solid is zero, the iep describes the condition at which the potential at the shear plane, i.e., the zp obtained from electrokinetic measurements is zero; the iep and pzc are the same in the absence of the specific adsorption. But the pzc of a mineral need not coincide with the iep in most cases [3]. 2.4 - Origin of Surface Charge Each mineral particle in a liquid whether colloidal (< 1 um) or nanoparticle (<100 run) carries electrokinetic charges depending on the properties of liquid phase such as pH and ionic strength [12]. The surface charge of clay minerals can originate due to a number of mechanisms discussed below. (i) Dissociation of Surface Groups In most solid minerals dissociable functional surface groups such as carboxyl (-COOH) and hydroxyl (-OH) are present. These groups may be ionizable depending on the solution pH; a surface is charged either negatively at high pHs or positively at low pHs. For metal oxides and hydroxides:
Electrokinetic Behavior of Clay Surfaces
61
where M represents a metal cation at the surface. Consequently, for simple metal oxides and hydroxides, e.g. SiO2, A12O3, Fe2O3, and A1OOH, complex metal oxides including clay minerals [10] and some hydrophobic minerals, e.g. coal [13], H+ and OH" ions are considered as the potential determining ions (pdi). (ii) Preferential Adsorption of Ions from Solution This mechanism is observed most commonly and results from the differences in the affinity of two phases for some ions. Some specific ions such as pdis can adsorb strongly on a solid surface and charge the particle or vice versa, a charged particle can become non-charged through such adsorption. For example, adsorption of H+ and OH" on oxide minerals (Eqs. 5 and 6), Ag+ and I" adsorption on silver iodide particles, and Al3+ and HDTMA+ (Hexadecyltrimethylammonium) adsorption on clinoptilolite [14]. It should be noted that especially for H+ and OH" ions, it is difficult to distinguish whether the charging of a particle is generated from the adsorption or dissociation of these ions (12).
In clay minerals the anionic centers are represented by the polar silanol groups (Si-OH) or relatively non-polar siloxane groups (Si-O-Si). Stumm [15,16] has outlined a number of reactions between the hydroxyl groups at the edges and metal ions in aqueous solution including surface metal binding (Eq. 7), ligand exchange (Eq. 8) and their combination (Eq.9): S-OH + MZ+
«
S-OM(Z"" + H+
(7)
S-OH + L"
«
S-L + OH"
(8)
S-OH + Mz+
+ L" o
S-L-Mz+ + OH"
(9)
The proportion of the edge surfaces and respective reactive hydroxyl groups for kaolinite is reported as 12 % [17] and 14 % [18] and for smectite minerals less thanl % [19]. (iii) Isomorphic substitutions Most clay and zeolite minerals are generally characterized by aluminum silicates and exhibit negative charges in water which results from the substitutions within the crystal lattice of Al3+ for Si4+ or Mg2+ for Al3+. Consequently, negative charges are developed in the lattice to compensate the so called exchangeable cations i.e., Na+, K+, and Ca2+ entering the crystal structure. When such minerals come in contact with water, some of these cations can easily dissociate leading to negatively charged surfaces. When a mineral particle is immersed in a liquid, charged species are transferred across the solid/liquid interface through one of the mechanisms discussed above. In equilibrium this condition is characterized by a surface potential (cp0) and a surface charge density (crs). Let us imagine a negatively charged solid particle in an
62
M.S. Çelik
electrolyte solution; while the oppositely charged counterions will congregate in the vicinity of the particle, coions which have the same sign with that of the particle will be repelled from the surface due to electrostatic interactions. Thus, a charged surface layer (layer 1) and an ionic layer (layer 2) all the way to the bulk water constitute the EDL with a thickness usually ranging from a few nanometer to a few hundred angstroms. 3 - Electrokinetics of clay minerals 3.1 - Layer Charge density of clay minerals Clay layers are electrically charged and thus undergo swelling upon the uptake of water into the interlayer space. Swelling occurs when the clay mineral expands beyond the original limit of about 0.95 nm. Talc and pyrophyllite never swell in an aqueous medium because of their zero layer charge, as illustrated in Table 1 [20]. Table 1 - Layer Charge density and crystalline swelling of clay minerals [20]. Interlamellar Mineral type Charge per Interlayer spacing in dilute unit cell clay suspensions (nm) cation Talc 0 0.93 Pyrophyllite 0 0.91 Illite 1.3 K 1.00 Vermiculite 1.3 Li >4.00 1.4-1.5 Na 1.16 K 1.4-1.5 Ca >4.00 Montmorillonite 0.67 Li >4.00 Na 1.55 and >4.00 K 1.91 Ca >4.00 Baidellite 0.25-0.6 Li 1.52 Na 1.27 K 1.54vw, 1.89s Ca >4.00 Saponite 0.25-0.6 Li 1.52 Na 1.26 K 1.54s, 1.87vw Ca 3.2 - Structure of clay minerals Classification of clay minerals is not within the context of this paper. However, in order to present the data available in the literature in some order, it is necessary to follow one of the accepted classifications. Grim [21] has grouped crystalline clay minerals based on their structural arrangement of layers as two-layer, three-layer, mixed layer and chain structure types. Since there is very little literature on the mixed layer, for the sake of simplicity, the electrokinetic properties of clay minerals will be presented in the order of the most common names, i.e. kaolinite, smectite and palygorskite. Literature data on other specific minerals will be also separately elaborated. The most important types of clay minerals are: kaolinites, smectites, attapulgites, illites and chlorites. Hikes and kaolinites exhibit plate-like particles without expanding lattice due to strong interlayer bonding and strong hydrogen bonding, respectively [22]. Chlorites show positive charge on one layer balanced by an
Electrokinetic Behavior of Clay Surfaces
63
additional negative charge practically with no interlayer water. The composition of clay mineral and the structural arrangement of octahedral and tetrahedral sheets and minerals account for the differences in the electrokinetics properties. Structural arrangements of tetrahedral and octahedral layers and their combinations constitute the layer charge in clay minerals. Some of the important properties of clay minerals pertinent to electrokinetics are presented in Table 2 [23]. Table 2 - Important properties of clay minerals related to electrokinetics Kaolin 1:1 layer Little substitution
Smectite 2:1 layer Octahedral and tetrahedral substitution Minimal layer charge High layer charge Low exchange capacity High exchange capacity capacity
Palygorskite 2:1 layer inverted Octahedral substitution Moderate layer charge Moderate exchange
3.3 - Kaolin Kaolin with an ideal formula of Al4Si4O10(OH)8 exhibits very little substitution in the structural lattice resulting in a minimal layer charge and low base exchange capacity. Linkage of a tetrahedral siloxane layer to one dioctahedral (gibbsite) or trioctahedral layer (brucite) forms the structure of 1:1 layer silicate. While in trioctahedral clay minerals all the three sites are occupied by magnesium ions to achieve charge balance, in dioctahedral layer two out of three positions are occupied by trivalent aluminum ions leading to practically uncharged layers which are held together through van der Waals' forces and partly by hydrogen bonding; kaolinite is a dioctahedral non-expendable clay mineral [4]. Published data on the electrokinetic properties of kaolin minerals are relatively few [24-30]. The iep of kaolinite is reported by several researchers [24-27,31]. Smith and Narimatsu [24] have used both microelectrophoresis and streaming potential techniques and found an iep of 2.2 for the former and no iep for the latter technique (Fig. 3). Cases et al [25] reported an iep of kaolinite at pH 3 and found that the sample to spontaneously flocculate at pH < 3.5 and stabilized at pH >5. The surface charge of kaolinite at pH= 3 did agree with the zeta potential vs. pH profile due to the electronegative character of the (001) faces characterized by OH" or O2" ions; this induced the particles to move under an applied field even at zero surface charge. Five commercial deposits from Georgia yielded iep values in the range of 1.5 to 3.5. The variation in zeta potential values were ascribed to the differences in ionic composition of the clay samples particularly those of exchangeable or soluble calcium ions [27]. Similarly, among the three commercial ball clays (kaolinite rich clay, smectitic kaolinite rich clay and illitic kaolinite rich clay) only kaolinite rich clay displayed a variation as a function of pH whereas the other two clays showed practically constant zp profiles against pH. Zeta potential curves versus pH with 10 mM Mg(NO3)2 and Ca(NO3)2 showed negative values in the entire pH range, the lower zp being about - 25 mV in the neutral and alkaline pH region. Dependence of zp on pH for ripidolite and kaolinite is presented in Fig. 4. Both minerals are characterized by two regions of distinctly different slopes, the steeper slope being below pH= 5.
64
M.S. Çelik
Figure 3 - Zeta potential ofkaolinite and a-Al2O3 as a function of pH; sp-streaming potential, mep-microelectrophoretic mobility [24].
Figure 4 - Zeta potential ofripidolite (U) and kaolinite (M) in 10'3 MNaCl solution as a function ofpH[19].
Electrokinetic Behavior of Clay Surfaces
65
When an alumino- or magnesium silicate layer is disrupted, the valences of the exposed crystal atoms are not completely compensated as they are in the interior of the crystal. These surfaces are called broken-bond surfaces or edge surfaces. The exposed functional groups are very active and may act as electron pair donors or acceptors. The free charges may be balanced by the uptake of cations or anions through either chemisorption or electrostatic attraction [20]. The zeta potential of clay particles inferred from the electrophoretic mobility using the Schmoluchowski equation has been criticized because of the heterogeneous nature of the particle charge. Kaolinite particles are not spherical but exist as hexagonal platelets. Calculation of zp and surface charge densities for non-spherical kaolinite assuming an equivalent sphere may result in quite misleading values. The zp calculated from such mobilities does not reflect the potential at the shear plane because of the screening effect of positive charges on the edges relative to those of negative charges at the faces resulting in a lower negative mobility [32]. Williams and Williams [26] have developed a method of calculating the zp of edge surfaces assuming a linear combination of the zeta potentials of quartz and alumina (Fig. 5).
Figure 5 - Estimation of edge potential by linear combination of quartz and a-alumina [26].
66
M.S. Çelik
They also investigated the effect of preparation techniques, pH and NaCl concentration on the electrophoretic mobility of sodium kaolinite. It is shown that above pH 4 the kaolinite platelets carry a net negative charge at different salt concentrations. A method based on the edge zero point charge is proposed to calculate face zp values. The Gouy-Stern-Graham model was shown to reasonably fit experimental data after making some unrealistic assumption of model parameters. A number of evidences suggest that the edge is composed of silica and alumina layers with a net positive charge at low pH and negative at high pH; the iep value of the edge varies in the range of 5-8. The face, on the other hand, is negatively charged at all pH values due to isomorphous substitutions within the lattice [25]. Delgado at al. [33] using the concept of Williams and Williams [26] together with the formulation of O'Brien and White [34], which considers double layer polarization and surface conductance, predicted the zp of sodium montmorillonite particles reasonably well. The electrokinetic properties of montmorillonite suspensions investigated by Callaghan and Ottewill [35] and Rioche and Siffert [36] indicate a strong pHindependent negative charge on the clay surfaces and a much weaker positively charged double layer on the edges. Kaolinite is composed of pH-independent (permanent) and pH-dependent surface charges [1,37-40]. The kaolinite basal surfaces carry a negative permanent charge ascribed to replacement of Si4+ by Al3+ in the tetrahedral layer of kaolinite and to isomorphic substitution. Because kaolinite has very small CEC (1-5 mequiv/100 g) [41-43], the permanent negative charge on the tetrahedral basal plane is often neglected in model calculations [37]. The pH-dependent surface charge of kaolinite may stem from the following [44]: (i) protonation and deprotonation of aluminol (> A1OH) on the edge; (ii) deprotonation of silanol (> SiOH) on the edge; and (iii) protonation and deprotonation of basal plane hydroxyl groups that are coordinated to two underlying aluminum atoms (> A12OH). The last site which protonates at lower pH is less reactive than the aluminol at the edge. Surface charge models presented on kaolinite tend not to include contributions from basal plane [37]. As most of the basal plane aluminum surface sites (> A12OH) undergo neither protonation nor deprotonation [45], the pHdependent surface charge of kaolinite in the pH range of 3 to 9 is dominated by the surface charge of edges which is equivalent to the difference of the amount of protonated Al edges (> A1OH+) and that of deprotonated Si edges (> SiO"). A plausible evidence for the existence of edge-face interactions in kaolinite suspensions has been observed [46] owing to increased edge surface area of kaolinite platelets relative to their lateral extent [21,47]. The ratio of total aluminol to total silanol edge sites is governed by the kaolinite structure and the electrostatic valence principle of Pauling and appears constant in various kaolinite samples. Ganor et al [44] have shown that the absolute PHPZNPO which is reported in the range of 3 to 7.5 in different kaolinites [1], is about constant. Since at the pHpZNPC only the edge contributes to the charge, changes in the total edge surface area should not appreciably change the pHPZNPC. The discrepancies observed in the 6 titration curves were attributed to the differences in the calculation of PHPZNPC- The authors further emphasized that incorrect assumption of PHPZNPC will introduce error in the calculation of surface potential and of the electrostatic correction factor [44]. Van Olphen [48] put forward that the edge potential is a combination of the potential on the silica and alumina surfaces. A simplified model of the edge utilizing a linear combination of quartz and alumina in the presence of electrolyte predicted a value of 7.2 close to that of actual zpc value. The potentials at low pH were found to
Electrokinetic Behavior of Clay Surfaces
67
agree well with those of Johansen and Buchanan for aluminum [49] silicate composed of about equal proportions of silica and alumina. Several studies have considered the isolated contribution of SiO2 and various aluminum hydroxyl species [1,19,50-51]. The absolute value of negative charge generally depends upon the Si4+/Al3+ substitution and adsorption of charged species onto the clay surfaces.
Figure 6 - The electroakustic zeta potential behavior of kaolinite particles as a function ofAlCU andpH in 10'3 MNaCl with kaolinite volume fraction of 0.02 for wt % AlCl3 levels of(9) 0, (O) 2.5xlO'3, (a) 2.5xlO'2, (A) 0.25, (v) 2.5 [28].
Figure 7 - Zeta potential of ripidollite in Iff2 MNaCl versus pH - (a) natural sample, (M) - sample milled for 6 min. [47]. Electroacustic zp measurements in nondilute kaolinite suspensions against pH and Al3+ concentrations revealed remarkable features for different faces [28].
68
M.S. Çelik
Increasing Al3+ concentration shifted the iep values to higher pH and the zp values became more positive (Fig. 6) At low Al3+ levels, the zp values were represented by the silica-like kaolinite face, pH dependent face, and edge interactions. Sondi and Pravdic [4] have demonstrated a major change in the zp-pH profile of ripidolite and beidellite minerals upon grinding. While no iep was found for both minerals in the absence of grinding, an iep of ripidolite at pH 6 (Figure 7) and that of beidellite at pH 3 [47,52]. These findings were attributed to the increase of fraction of edge surfaces upon mechanical disintegration of clay particles and the corresponding increase in both surface areas and cation exchange capacities and in turn their favorable effect on electrokinetic properties. Indeed, exposure of reactive hydroxyl sites upon milling of ripidolite by a planetary ball mill for 6 min was shown to induce a 12.3-fold increase in specific surface areas and 3-fold increase in cation exchange capacities [47]. Kaolinites exhibit surface acidity due to terminated bonds and structural coordination across the edge faces. The character of acid sites are usually ascribed to Bronsted acidity arising from broken -O-Si-O-Si networks or to Lewis acidity from strained gibbsite layer [43]. Bronsted acid generation may derive either from silanol protonation or aluminol protonation both of which will equally contribute to the generation of acid sites and also to electrokinetic charge. The formation of various aluminum species particularly those of A1(OH)2+ and colloidal A1(OH)3 and their subsequent adsorption onto the solid surface are important in the charging of kaolinite. The effect of a series of metal chlorides of different valency on electrophoretic mobility of kaolinite as a function of their concentration is presented in Figure 8. It is evident that while the zp profiles in NaCl and MgCl2 suspensions exhibit negative charges in the entire concentration range, lanthanum and particularly aluminum ions are capable of reversing the sign of zp. Although both La and Al have the same trivalent charge, Al appears to reverse the charge at much lower concentrations than lanthanum. The presence of highly charged polyhydroxy aluminum species such as A14(OH)2O4+ particularly at pH 6 is plausible based on the species diagram of Aluminum and also the work of Matijevic et al [53]. A change in the valency of anion for potassium salts showed a much smaller effect compared to that of cations, with the effect becoming relatively small as the size of the anion increases from chloride to iodide [54]. In another work of Buchanan and Oppenheim [55]. Comparison of raw and leached kaolinite at pH 2.25 and 6 have yielded different zp-pH profiles, which were attributed to preferential removal of aluminum leaving a dealuminated surface layer. The solution behavior of multivalent metal cations (Mn+) has been extensively covered by various researchers [56,57,7]. The hydrolyzed species will undergo the following modifications in aqueous solution [7]: Mn+ + x O H " » M(OH)x(n-x)+
(10)
M(0H)x(n"x)+ + O H " o M(OH)x+1(nx+1)+
(11)
M(OH)X
+ +H+ <» M(0H) x .i (nx+l)+ + H2O
(12)
The above reactions indicate that the solution pH is of utmost importance in the formation of hydrolyzed products. The adsorption of Co2+, La3+ and Th4+ onto silica and titania was studied by James and Healy [56]. Charge reversals due to the adsorption of hydrolysable metal ions and their subsequent precipitation were correlated in terms of electrophoretic mobility.
Electrokinetic Behavior of Clay Surfaces
69
Figure 8 - Effect of cation valence on electrophoretic mobility of kaolinite particles at pH 6. Ofor a series of metal chloride solutions [54].
Figure 9 - Dynamic mobility of kaolinite as a function of pH with and without Co2' (Co (NO3)2: (O) nil; (open diamond) Iff3 M; (u) 2xlO~3 M; (M) 4xlO'3 M; (x) represents solution conductivity at 4x10'3 M; A marks the pH at which hydroxide precipitation is expected [58]. They developed a model involving three charge reversals (CR) in zeta potential with pH. The first (CR1) represents the point of zero charge (pzc) of the solid and the second is ascribed to the specific adsorption of partially hydrolyzed metal ions inducing a charge reversal (CR2). At a critical pH corresponding to the onset of metal hydroxide precipitation, the adsorbed hydroxyl ions are converted to the respective
70
M.S. Çelik
metal hydroxide. Above this pH value the zp is reversed from positive to negative at (CR3). A similar study was performed for kaolinite and hydrolysable metal ions, i.e. Co2+, Cu2+ and Cd2+, as illustrated in Figure 9 [58]. At about 2xlO"3 M of Co2+ addition, the CR2 occurs at around 7.5; CR3 takes place at a pH of about 9 that is much lower pH than the pzc of 10.7, the theoretically known value for Co(OH)2. These values correspond to the concentration of 100 times more than that required by James and Healy [56]. The most interesting conclusion out of this study is that the unhydrolyzed form (Mn+) does not strongly adsorb to a metal surface despite its large negative charge but it is the hydrolyzed product (M(0H) x (nx)+ that strongly adsorbs onto a metal surface. The adsorption of metal hydroxides continues with increasing pH until a mono layer is attained. Similar results are also reported in a recent publication [59]. 3.4 - Smectites Linkage of two tetrahedral layers to one central octahedral layer creates a 2:1 layer. Smectite, vermiculate and mica are the most important groups in this layer structure. Sodium and calcium montmorillonites are the most common clay minerals in the smectite group. Montmorillonite has an idealized structural formula of My+nH2O(Al2yMgy)Si4O10 (OH)2 [22]. The montmorillonite structure is classified as dioctahedral with two thirds of the octahedral sites taken up by trivalent cations. Depending on the dominant exchangeable cation, the mineral is named either sodium or calcium. The substitution of alumina for silica in the tetrahedral sheet and that of iron and magnesium for iron and magnesium creates a charge imbalance in the 2:1 layer. The charge imbalance in Smectite is about 0.66 per unit cell [21]. This net positive charge deficiency is balanced by exchangeable cations adsorbed between the unit layers on the edges [23]. In aqueous suspension, both anions and cations may exchange with ions in bulk solution; these are called exchangeable ions. The total amount of cations adsorbed on the clay, expressed in miliequivalents per hundred grams of dry clay, is called the cation exchange capacity (CEC) [60]. Montmorillonite is typically reported to have CEC values of 81-124 meq/100 g [61]. The zp-pH curves for a series of smectite minerals are given in Fig. 10. Evidently, except for beidellite no dependency of zp on pH is noted.
Figure 10 - Zeta potential of smectite clay minerals in Iff3 MNaCl solution against pH (O) saponites; (U) beidellite; (n) Otay montmorillonite [19].
Electrokinetic Behavior of Clay Surfaces
71
Various data reported on the electrokinetic properties of smectites ranging from montmorillonite to mica reveal that the zp is negative and no iep is observed in the pH range of 2-12 [19,33,62-64]. Sodium montmorillonite particles are laminar in shape with their surface charges inhomogeneously distributed. While face surfaces bear a negative charge generated by substitution of lattice ions and thus basically independent of the aqueous composition, edge surfaces show a pH-dependent charge. Zeta potential of montmorillonite particles as a function of pH at constant ionic strength of 10'2 M NaCl is shown in Figure 11 [65]. Zeta potential is negative throughout pH region and basically independent of pH due to the dominant role of constant negative charge of faces; this is in accord with literature which suggests that edges induce a negligible contribution in the electrokinetics of smectites [19,35,66]; this inference is logical considering that only about 1 % of the total surface area is attributed to edges [19,67]. Heath and Tadros [66] proposed that the iep of edges must be close to that of kaolinite, viz. pHiep 7. Based on experimental data on bentonite coagulation, Permien and Lagaly [68,69] have inferred an iep value of about 6. Sondi et. al. [19] related the zp of edges to that of weighed average of silica and alumina using zedge=l/3(^SiO2 +2^A12O3). Accordingly the zp of edges varied with pH with an iep of about 7, as shown in Figure 11; this is in agreement with some finding in the literature [65,66].
Figure 11 - Zeta potential of montmorillonite (NaMt), A12O3, and SiO2 powders as a function of pH in the presence of Iff2 M NaCl. The curve labeled NaMt edges was calculated as 1/3 {Qstoi + 2Q,2O3}[65J.
72
M.S. Çelik
A number of researchers disagree with the concept of edge-face interactions being dominant in clay suspensions [70-73]. The non-rigidity of montmorillonite platelets which exhibit high lateral extent is the primary reason for discrediting the edge-face mode of interaction. A majority of findings supporting the edge-face interactions were reported above pH 8 [74]; however, this suspension pH must be lower than the iep of the edges for the establishment of face-edge interactions. Various studies indicate that the estimated iep of the edge of montmorillonite platelets vary in the range of 7 to 8 based on non-crystalline A1(OH)3 and SiO4 compounds [2,30,75]. Montmorillonite platelets, contrary to kaolinite ones, exhibit high lateral extent relative to their thickness and thus face double layer would extend out sideways such that it will screen any charge developed by the edges [35,76]. Controversial edge charge screening appears to be unlikely in distilled water and at low electrolyte concentrations [48]. Zeta potential values for a montmorillonite samples at 0.8 and 2.7 % solids concentration yielded slightly less negative trend particularly at low salinities. Increasing NaCl concentrations produced a profile typical of double layer compression effect [76]. Montmorillonite particles carry two kinds of electrical charges: a pH dependent charge arising from proton adsorption/desorption reactions on surface hydroxyl groups located at the edges and a structural negative charge resulting from isomorphous substitutions at the faces of platelets [75]. Protonation-deprotonation behavior of metal hydroxides and materials with structural charges such as clay minerals exhibit considerable differences. While the proton adsorption curves in the presence of different supporting electrolytes in the former intersect at the point of zero charge, in the latter case H+ curves against pH at different NaCl concentrations for montmorillonite [77], sodium illite and several soils [78] and for sodium-attapulgite [79] did not cross each other. Avena and De Pauli [75] utilized potentiometric titrations to obtain proton adsorption vs. pH curves at different NaCl concentrations and mass titrations to determine the dependence of the point of zero charge (PZNPC) with the ionic strength (Fig. 12). The PZNPC was found to decrease with the ionic strength without and crossing points for several titration curves conducted at different NaCl concentrations. A model assuming the presence of structural negative charges into the clay particle and variable charge sites and cation exchange sites at the particle sites were put forward. The model predicted that proton adsorption at high pH takes place mainly on variable charge sites and become positively charged at pH values lower than the PZNPC. Acidbase potentiometric titration results and model predictions at three different ionic strengths are presented in Figure 13. As apparent, the model slightly overestimates PZNPC at high ionic strength probably due to the assumption that all the surface sites experience the same potential rather than smeared out. However, much better prediction was obtained with an illite sample. The authors [75] suggested that a more sophisticated model developed by Chang and Sposito [80] that considers edges and face separately and the spill over effect from the basal plane on the edges could improve the prediction at the expense of loss in simplicity. The most up-to-date model has been proposed by Lero and Revil [81]. An electrochemical triple-layer model (Fig. 14) involving a speciation model of the active crystallographic surface sites and EDL comprising the Stern and diffuse layers. The model computes both the zp and surface conductivity. For the latter, the model consists of two contributions; the Stern layer and the diffuse Gouy-Chapman layer which shelters the excess of counter ions. Although both contributions are significant the
Electrokinetic Behavior of Clay Surfaces
73
Stern layer contribution dominates. The model is shown to be applicable to both 1:1 (e.g. kaolinite) and 2:1 (e.g smectite) under a variety of thermodynamic conditions such as pH and salinity. The surface properties taken as a base in the model is illustrated in Fig. 14 for both kaolinite and smectite. Surface complexation reactions which considers aluminol and silanol surface groups together with isomorphic substitution on the {010} and {110} planes of clay minerals are assumed and corresponding surface site densities incorporated in the model [81].
Figure 12 - PZNPC values obtained from (a) acid-base potentiometric titrations and (U) mass titrations. The line represents the model prediction [75]. The basal charge in clays stems from a partition of the counterions between the Stern layer and the diffuse layer and remains constant irrespective of structural charge of clay and the monovalent electrolyte The effect of layer charge of smectites on their electrophoretic mobilities using electrophoresis technique has been reported by Thomas et al [61]. Two clay series covering Cu-montmorillonite (0-0.7) and synthetic saponites (0.7-2) were selected to represent the charge domain from 0 to 2 charges per unit structural cell. At neutral and alkaline pH, the EM does not significantly change with layer charge. Such lack of dependency has been also observed by Low (82) and Miller and Low [83]. A schematic model illustrating the effect layer charge and ionic strength on the electrophoretic mobility of swelling clays is shown in Figure 15. It is the layer charge that dictates the aggregation/dispersion behavior of clays. Thomas et al [61] have shown that low and high charge clays are non-expandable and not dispersed in aqueous suspension which displays variable electrophoretic mobility. Conversely, medium charge clays are highly dispersed and exhibit constant electrophoretic mobility. There is a lack of dependency between cation exchange and zp of smectites from various sources [61]. However, they found an excellent
74
M.S. Çelik
correlation between theoretical layer charge per unit cell and CEC, particularly for layer charges < 1.5, as shown in Figure 16.
Figure 13 - Acid-base potentiometric titration results (M) I = 0.006 M; (O) I = 0.0014; (a) I = 0.088 M. Lines are model predictions [75]. Homoionic smectites with bivalent interlayer cations display lower absolute values of zps than monovalent exchanged smectites. Cation exchange by trivalent were shown to result even in charge reversal [33]. The structure of mica is basically the same as that of smectite except a small amount of excess negative charge is balanced by potassium ions resulting in a nonswelling structure. Illite is the most common hydrous mica group found in soils and sediments [4]. Chlorites are a group of nonexpanding clay minerals with low CEC and surface charge densities [84]. Since negative charges generated by isomorphous subsititution are balanced by Mg2+ ions of brucite or gibbsite, unlike other clay minerals, it exhibits an iep at about pH 5 [63,85]. Chlorite displays a relatively lower level of zp values in both mono- (NaCl) and bi-valent (MgCl2 and CaCl2) electrolytes (Fig. 17) a charge reversal occurring only at around 10"3 M CaCl2 concentration [63]. Illite is a nonexpanding clay mineral with low CEC. The negative charge generated by isomorphous substitution is compensated by a layer of potassium ions [84]. Addition of mono- (NaCl) and bi-valent (CaCl2 and MgCl2) ions basically compresses the EDL in an expected manner. Omitting the last dubious points in Fig 18, no charge reversal is observed at all electrolyte concentrations. Illite seems to exhibit higher negative potentials than chlorite indicating a lower level of isomorphous substitution in the case of chlorite.
Electrokinetic Behavior of Clay Surfaces
75
Figure 14 - Active surface sites at the edge of (a) 1:1 clays (e.g. kaolinite and (b) 2:1 clays (e.g. smectite); note the difference in surface sites densities and types on the edge of the mineral {110} and {010} planes [81]. Unlike most other minerals, electrokinetic measurements on some montmorillonites exhibit an unusual behavior. While the addition of monovalent salts such as NaCl and KCl is expected to compress the EDL, an opposite trend is observed. A set of systematic experiments with monovalent and multivalent ions have been carried out to understand the mechanism of this process [86]. The monovalent cations were found to increase the negative charge of smectite in the order of Li > K > Na. Above about 10 M salt concentration for almost all these ions, there occured a minimum peak followed by a decrease in the absolute potential; this turning point usually indicates the slowing down of the ion exchange process and the onset of normally observed double layer compression. The ion exchange data shown in Table 3 vividly shows that increasing the concentration of monovalent cations leads to an increase in the released calcium concentration. There are indications that the ion exchange mechanism is related to the cation diameter [42]. Although the order of effectiveness follows K + > Li + > Na + , the hydrated ion diameter follows that of Li + > Na + > K + . The decrease obtained in zp values upon addition of monovalent salts again confirms the release of Ca^ + type cations.
76
M.S. Çelik
Figure 15 - Schematic model explaining the combined effect of layer charge and ionic strength on the electrophoretic mobility of swelling clays at acidic pH [61].
Figure 16 - Variation ofCEC with layer charge per unit cell [61].
Electrokinetic Behavior of Clay Surfaces
11
Figure 17 - Zeta potential of chlorite in various electrolyte solutions: (A) NaCl; (%) CaCl2; (a) MgCl2 atpH6.5 0.2 [63].
Figure 18 - Zeta potential ofillite in various electrolyte solutions (A) NaCl; (%) CaCl2; (a) MgCl2 atpH6.5 0.2 [63]. Similar experiments were conducted with a well-known model bentonite, the Wyoming bentonite. Zeta potential of Wyoming bentonite as a function of NaCl concentration is presented in Figure 19. The concentration of Ca2+ ion in solution upon NaCl addition has been simultaneously measured by a calcium ion selective electrode. Although both an increase in the zp and Ca2+ concentration are at lower levels, nevertheless, the results clearly illustrate that the same phenomenon may occur with all cation exchangable clays. Indeed, sepiolite exhibits a similar behavior but at much lower scale as discussed in the subsequent section.
M.S. Çelik
78
Table 3 - Release of calcium ions upon addition of monovalent salts to 1 % by weight smectite suspensions [86]. Monovalent ion Cone, M
Calcium ions released, M Na
10-3
io-
2
5xlO-
2
+
K+
Li +
1.9xlO'5
2.1xlO"4
1.5xlO'4
2.8xl0" 4
3.2xlO' 4 -4 3.7x10
2.3xlO" 4 -4 3.5x10
3.5xlO' 4
Figure 19 - Effect of NaCl concentration on zeta potential (zp) of Wyoming bentonite and the corresponding release of calcium ions (Ca) from it [86]. 3.5 - Palygorskite and Sepiolite Palygorskite and attapulgite are synonymous terms for the same hydrated magnesium aluminum silicate mineral. Sepiolite ((Si^XMgg^oCOH^ (OH2)4. 8H2O [87] is almost structurally and chemically identical to palygorskite except it has a slightly larger unit cell [23]. However, aluminum in sepiolite has been considerably substituted by magnesium in the octahedral layer such that it gives moderately high layer charge. Structurally, it is formed by alternation of blocks and tunnels that grow up in the fiber direction (see Fig. 20). Each structural block is composed of two tetrahedral silica sheets sandwiching a central sheet of magnesium oxide-hydroxide [88]. Its unique fibrous structure with interior channels (3.6x10.6 A) that allows incorporation of organic and inorganic ions into the structure of sepiolite.
Electrokinetic Behavior of Clay Surfaces
79
Sepiolite undergoes acid-base interactions in the vicinity of pH 8.5 and thus exhibits a strong buffering capacity, particularly in the acidic pH. It takes less than a minute for a sepiolite suspension adjusted to pH 3 to attain its natural pH of 8.5. Since Mg ions located in the octahedral sheet are conducive to ion exchange, they are released into solution, as the pH is made more acidic. It should be noted that while the Mg concentration at natural pH is about 20 ppm, it is about 400 ppm at pH 3. Zeta potential measurements conducted as a function of solids concentration revealed significant differences. The zp-pH profile of sepiolite at two different solids concentration is presented in Fig. 21. It is seen that the isoelectric point (iep) of sepiolite at 0.2 % solids concentration yields 3.2 and that at 5 % gives 6.3. Such difference can be explained on the basis of increased Mg concentration at high solids concentration. Increasing the solids concentration from 0.2 to 5 % proportionally shifts the zp values from negative to more positive values; this behavior was also found in the case of colemanite (Ca2B5On.5H2O), which yielded an iep of 8.5 at 0.1 % solids concentration and 10.5 at 1 % solids concentration [89]. The literature on Palygorskite is practically none. However, unpublished zp measurements in our laboratories reveal that white sepiolite containing carbonaceous impurities, yield an iep of about 4 whereas those made with brown or beige Sepiolite exhibit an iep around 5. Variation of zp with solids concentration and more importantly its strong buffer feature of sepiolite require an additional care to be exercised. The zeta potential behavior of sepiolite in the presence of monovalent ions is shown in Fig. 22. The zp curves are characterized by two regions with different slopes. The first region is dominated by ion exchange reaction between the added monovalent ion and magnesium ions in the octahedral layer. The slope remains virtually horizontal indicating that the electroneutrality is maintained. The second region is represented by double layer compression indicative of adsorption of monovalent ions in the EDL.
Figure 20 -A schematic model representing the sepiolite structure [88].
80
M.S. Çelik
Figure 21 - Variation ofzp ofsepiolite with pH at different solids concentrations
Figure 22 - Effect ofmonovalent ions on zeta potential ofsepiolite [91].
Electrokinetic Behavior of Clay Surfaces
81
Examination of Fig. 22 reveals that there is an order of effectiveness particularly in the second region; this is found to follow the order of electronegativity of the ions in the order of H>Li >Na>K>Cs [90]. Irrespective of the ionic size, the ability of ions to be incorporated in the EDL and their ion exchange favors the ions of lower electronegativity. Accordingly, the exchange of Cs+ ions with Mg+2 in the octahedral layer is attained until 10"4 M ion concentration. Similarly, Li+ ion with its highest electronegativity exhibits ion exchange over a wide range of ion concentration and remains negatively charged at the entire ion concentration. In order to test the validity of the above assertion, a series of systematic adsorption tests was conducted in sepiolite/monovalent ion system and the results are given in Fig. 23. As a measurement approach, the released Mg ion concentration which identifies the magnitude of ion exchange was analyzed. The earlier suggested order is again prevalent in adsorption measurements; this indicates the influence of electronegativity of ions in this particular system. The largest ion (Cs+) with the lowest electronegativity undergoes ion exchange with Mg+2 in a rapid manner whereas the smallest Li+ ion requires a wider range of concentration. Research on adsorption kinetics is required in order to identify if factors other than electronegativity plays a role in this phenomenon. Zeta potential behavior of minerals in the presence of added monovalent salt such as NaCl can undergo three possible modes: (a) reduction of positive charges upon electrolyte addition through adsorption of ions in the EDL and its consequent compression, (b) flat type zeta potential profile indicating exchange of monovalent ions with those in the solid to maintain electrical neutrality until ion exchange ceases, (c) increase of negative charges on addition of electrolyte due to exchange of monovalent ions and resultant release of higher valency ions in excess of electrical neutrality leaving a negatively charged deficit surface. When a nonionic polymer is adsorbed on a particle surface, a displacement of the shear plane occurs compared to the position in the absence of adsorbed polymer [92]. This displacement depends upon the thickness of the adsorbed layer [93]. The adsorbed polymer is shown to induce no change on zpc of oxides in the presence of indifferent electrolytes. The presence of polymer is assumed to induce no effect on the surface charge density, specific adsorption of ions in the Stern plane and the charge distribution in the diffuse layer. The observed decrease in zp was related to the shift in the Stern plane which corresponds to the hydrodynamic thickness of the adsorbed layer [94]. Rossi et al [92] calculated the adsorbed polymer layer thickness for a flat double layer as follows: ^(x) = tanh (zeyd/4kT) exp [-K(S - A)
(13)
where *¥(x) is the potential distribution in the diffuse layer against distance x, z is the valence of the counterion, e is the electronic charge, A id the thickness of the Stern layer, 1/K is the thickness of the EDL. Assuming the potential measured in the absence of polymer (vj/d) is equal to the zp in the absence of adsorbed polymer, they calculated the thickness of the adsorbed layer (8). The adsorbed layer thickness was found to increase with the number of ethylene oxide units of the surfactant as shown in Figure 24; this indicates that the length of the poly ethylene oxide chain determines the length of the adsorbed layer thickness and the shift of the Stern layer.
82
M.S. Çelik
Figure 23 - Released Mg+2 ion concentration sepiolite matrix upon addition of different monovalent salts [91]. 4 - DLVO Theory 4.1 - Interaction Energy Curves DLVO (Derjaguin and Landau [95] Verwey and Overbeek [96]) theory explains the stability of colloidal systems considering the total potential energy of interaction between colloidal particles depending on the distance between them. The total or net interaction energy is equal to the summation of the EDL interaction energy (VER) and the van der Waals interaction energy (VVA) and given as VT = V E R + V V A
(14)
The EDL interaction energy between two colloidal particles [97,98] in an electrolyte solution results from the overlapping of their diffuse layers, that is, it results from an osmotic pressure of counterions in repulsive character when the two particles have the same sign of charge. But, when they are opposite in charge, the DL interaction energy becomes attractive in character. The van der Waals interaction energy (V V A) between the particles arises from the London-van der Waals forces. If the two particles are identical, van der Waals interaction is always negative (attractive) but in the case of different particles, this may change depending upon the Hamaker constant of the particles and the medium. For two spherical colloidal particles of equal size which usually appear in most colloidal systems, the total interaction energy using the DLVO theory is described provided that a » H as follows [98]: VT = 32 7i s 80 a (kT/ze)Y exp (-KH) - (Aa / 12H)
(15)
Where a is the radius of colloidal particles H is the shortest distance between
83
Electrokinetic Behavior of Clay Surfaces
them, A is the effective Hamaker constant depending on the Hamaker constants of the particles and the medium, K is the Debye-Huckel parameter and % is given by X = tanh (zel, /4kT)
(16)
In Eq. (15), the first term denotes the EDL repulsion energy (VER) and the second term the van der Waals attraction energy (V V A)- Note that here i; is used as an effective surface potential of the particles. If VT, VER and V V A are plotted as a function of the distance (H), the characteristic curves like in Fig. 25 are obtained. Here the VT value at each distance is obtained by the summation of the V E R and V V A value that is the smaller energy is subtracted from the larger. If repulsive, the net value is plotted above, if attractive below, and then the VT curve is formed. As seen in Fig. 25, both repulsive and attractive interactions become weaker as the separation distance becomes larger. At sufficiently large distances the particles exert no influence on each other.
Figure 24 - Adsorbed layer thickness as a function of the number of ethylene oxide units for the surfactants: 27.5; 48; 79.4 [92]. If the colloidal particles are very close, the van der Waals attractive forces take over with a resultant negative energy of interaction leading to the coagulation of particles. At contact state, the total interaction energy is known as the primary minimum. There is also negative attraction energy usually beyond 3 nm known as the aggregation region or the secondary minimum [97]. But the coagulation in this region is not stable and reversible with respect to the case in the primary minimum [99,100]. Rheological properties such as thixotropy are closely related to coagulation at the secondary minimum. If the particles are further away, van der Waals attraction forces decrease sharply because of the large exponent of inverse distance, and the EDL repulsion forces take over with an energy barrier occurring between the particles. If aggregation is required, the height of energy barrier shown in Fig. 25 should be lowered or disappeared. Conversely, for a good dispersion the height of energy barrier must be enlarged. These two cases can be realized by changing the EDL repulsive forces, as it is perhaps impossible to change the van der Waals forces. The EDL repulsive forces can be altered by changing the zp of particles through changing parameters such as the type and concentration of electrolyte and
84
M.S. Çelik
solution pH. For example, for negatively charged colloidal silica particles the zp values distilled water, in KC1 and in A1C13 are - 30, -14 and 0 mV, respectively [7]. The total interaction energy (VT) curves for zp value of-30 mV exhibit dispersion and the height of the energy barrier is considerably high. When KC1 is added into water the zp of silica particles comes down to the -14 mV due to the double layer compression with a resultant decrease in repulsive energy and in turn in the height of the energy barrier. Sometimes the van der Waals attractive forces may become dominant depending on the kinetic conditions and/or the existence of the non-DLVO forces such as hydration, hydrophobic, and steric. When a trivalent electrolyte, A1C13, is added into system, the zp comes down to zero, that is called the isoelectric point owing to the charge neutralization on the silica surface; the height of the energy barrier disappears and the van der Waals attractive forces become dominant in the system [101]. Accordingly, the colloidal particles come in contact and coagulate. At the point, where the energy barrier just disappears: dVT/dH = 0
and
VT = 0
(18)
Applying these conditions to Eq. (15) results in an expression for the critical coagulation concentration, Cc (ccc), for a symmetrical (z+=-z., z is the valence number of electrolyte, such as NaCl) electrolyte as follows: Cc = K (P4/A2 z6)
(19)
Figure 25 - Repulsive and attractive forces as a function of distance of separation
Electrokinetic Behavior of Clay Surfaces
85
Where K is a constant which depends only on the properties of the dispersion medium and A is the effective Hamaker constant. When the zp is very high, the term P approaches unity and the critical coagulation concentration (ccc) becomes inversely proportional to the sixth power of the valency, z. This dependence of ccc on 1/z6 known as the Schultz-Hardy rule is consistent with the DLVO theory. For instance, if coagulation occurs at 1 M with a 1:1 electrolyte, it will occur at 1/26 (= 0.016) M with a 2:2 electrolyte, and at 1/36 (= 0.0014) M with a 3:3 electrolyte. Although the ccc is proportional to 1/26, the surfaces were assumed to have a very high potential which is contrary to common observations. The use of linear approximation under low potential conditions predicts ccc a z2 [97]. 4.2 - Application of DLVO Theory to Clay particles The knowledge of the EDL surrounding clay particles in an aqueous media is of great interest in various diverse applications. Since clay particles are colloidal and non-uniform in size and shape, the treatment of EDL becomes more challenging. Duran et al [65] used an extended DLVO model including electrostatic, van der Waals and, polar acid-base contributions to the total energy demonstrated that while face-to-face interactions are practically independent of pH, edge-to-edge interactions are most attractive at the iep of edges.
Figure 26 - Potential energy of interaction (per unit surface area) between montmorillonite platelets as a function of their separation (H): Acid-base (AB), Lifshitz-van der Waals (LW), and electrostatic (EL) contributions (at different pH values) for edge to face interaction [65].
86
M.S. Çelik
The surface free energy was calculated using the thin layer wicking technique developed by Van Oss [102]; the results showed that both faces and edges are almost completely monopolar and electron donor in nature. The zeta potentials and surface free energies of faces and edges were then used in the determination of the potential energy of interactions face to face (F-F), face to edge (F-E), and edge to edge (E-E). Figure 26 show repulsive short-range interaction, attractive van der Waals interaction and EL contribution with pH. Figure 27 summarizes the minimum values of Vi3j against pH. It is vividly shown that f-F interactions are independent of pH, whereas E-E interactions are most attractive at the iep. Interestingly, the most significant variations are observed with E-F potential energy which is comparable to E-E and F-F but at pH values below 7 a strong attraction leading to house-of card gel structure is proposed in concentrated clay suspensions and at ionic strengths higher than 5x10"'M [35,66,103,104]. Calculations made by de Kretser et al [105] at different NaCl concentrations identify a secondary minimum begin to occur at 0.05 M NaCl concentration corresponding to a separation distance of 90-140 A, They also indicated that while below 0.1-0.2 M NaCl level the edge-edge and face-face interactions are likely, at higher salinities face-face interactions are dominant in the region of primary minimum.
Figure 27 - Minimum value of the potential energy of interaction as a function ofpH [54].
Electrokinetic Behavior of Clay Surfaces
87
5 - References [I] B.K. Schroth and G. Sposito, Clays Clay Miner., 45 (1997) 85. [2] G.A. Parks, Chem. Rev., 65 (1965) 177. [3] K.K. Das, Interfacial Electrokinetics and Electrophoresis, Surfactant Sci. Series, vol. 106, Ed. A.V. Delgado, Dekker, New York, 2002. [4] I.J. Sondi, and V. Pravdic, Interfacial Electrokinetics and Electrophoresis, Surfactant Science Series vol. 106, Ed. A.V. Delgado, Dekker, New York, 2002. [5] M.S. Celik and B. Ersoy, in Encyclopedia of Nanoscience and Nanotechnology, Eds. J. A. Schwarz, C. Contescu, and K. Putyera. In Press, 2003. [6] S.L. Swartzen-Allen and E. Matijevic, Chemical Rev., 74 (1974) 385. [7] J.R. Hunter, Zeta Potential in Colloid Science, Principles and Applications. Third Printing, Academic Press, San Diego, 1988. [8] J. Leja., Surface Chemistry of Froth Flotation, Plenum Press, New York, 1983. [9] H. Li, S. Wei, C. Qing and J. Yang, J. Colloid Interface Sci., 258 (2003) 40. [10] P. Somasundaran, Advances in Interfacial Phenomena, AIChE Symposium Series, vol.71, no. 150, 1975. [II] S. Usui, Electrical Phenomena at Interfaces, Fundamentals, Measurements and Applications, Ed. A. Watanable, A. Dekker, New York, 1984. [12] J. Lyklema, Colloidal Dispersions. Ed. J.W. Goodwin, Dorset Press, Amsterdam, 1982. [13] J.S. Laskowski, Developments in Mineral Processing, Advisory Ed. D.W. Fuerstenau, Elsevier, Amsterdam, 2001. [14] B. Ersoy, and M.S. Celik, Microp. and Mesop. Mat., 55 (2003) 305. [15] W. Stumm, The Chemistry of Solid-Water Interfaces, New York, Wiley, 1992. [16] W. Stumm, Colloids Surfaces, 73 (1993) 1. [17] A.P. Ferris and W.B. Jepson, J. Colloid Interface Sci., 51 (1975) 245. [18] A.E. James and D.J.A. Williams, Adv. Colloid Interface Sci., 17 (1982) 219. [19] I. Sondi, O Milat and V. Pravdic, J. Colloid Interface Sci., 189 (1997) 66. [20] S. Yariv, Modern approaches to Wettability: Theory and Applications, Eds. M.E. Schrader and G.I. Loeb, Plenum Press, 1992. [21] R.E. Grim, Clay Mineralogy, McGraw Hill, New York, 1968. [22] G.W. Brindley and G. Brown, Crystal Structures of Clay Minerals and their XRay Identification, Miner. Soc, London, 1980. [23] H.H. Murray, Applied Clay Sci., 17 (2000) 207. [24] R.W. Smith and Y. Narimatsu, Minerals Engin., 6 (1993) 753. [25] J.M. Cases, C. Touret-Poinsignon, and D. Vestier, Acad Sci Ser. C, 272 (1971) 728. [26] D.J.A. Williams and K.P. Williams, J Colloid Interface Sci., 65 (1978) 79. [27] J. Yuan and R.J. Pruet, Miner. Metall. Process., 15 (1998) 50. [28] S.B. Johnson, D.R. Dixon and P.J. Scales, Colloids Surfaces A, 146 (1999) 281. [29] A.P. Ferris and W.B. Jepson J. Colloid Interface Sci., 51 (1975) 245. [30] S.K. Nicole and R.J. Hunter., Aust. J. Chem., 23 (1970) 2177. [31] A.C. Pierre and K. Ma, J. Mat. Sci., 32 (1997) 2937. [32] P.F. Luckham and S. Rossi, Adv. Colloid Interface Sci., 82 (1999) 43. [33] A. Delgado, F. Gonzalez-Caballero and J.M. Bruque, J. Colloid Interface Sci., 113 (1986)2003. [34] R.W. O'Brien and L.R. White, J. Chem. Soc. Faraday Trans., 2, 74 (1978) 1607. [35] I.C. Callaghan and R.H. Ottewill, J. Chem. Soc. Faraday Disc, 57 (1974)110. [36] D. Rioche and B. Siffert, Proceed. IV Int. Clay Conference, 465 (1978).
88
M.S. Çelik
[37] P.V. Brady, R.T. Cygan and K.L. Nagy, J. Colloid Interface Sci, 183 (1996) 356. [38] D.B. Ward and P.V. Brady, Clays Clay Miner., 46 (1998) 453. [39] M.D.A. Bolland, A.M. Posner and J.P. Quirk, Clays Clay Miner., 28 (1980) 412. [40] Z. Zhou and W.D. Gunter, Clays Clay Miner., 40 (1992) 365. [41] R.K. Schofield and H.R. Samson, Disc Farad. Soc, 18 (1954)135 [42] A.S. Michaels and J.C. Bolger, Ind. Eng. Chem. Fund., 1 (1962) 153. [43] R.F. Conley and A.C. Althoff, J. Colloid Interface Sci., 37 (19719 186. [44] J. Ganor, J. Cama and V. Metz, J. Colloid Interface Sci., 264 (2003) 67. [45] J.F. Huertas, L. Chou and R. Wollast, Geochim. Cosmochim. Acta, 62 (1998) 417. [46] B. Rand and I.E. Melton, J. Colloid Interface Sci., 60 (1977) 308. [47] I. Sondi, M. Stubicar and V. Pravdic, Colloids Surfaces, 127 (1997) 141. [48] H. Van Olphen, Characterization of powder surfaces, Eds. G.D. Parfitt and K.S.W. Sing, Academic Press, New York, 1976. [49] G. Johansen and A.S. Buchanan, Aust. J. Chem., 10 (1957) 398. [50] W.N. Rowlands and R.W. O'Brien, J. Colloid Interface Sci., 175 (1995) 190. [51] G. Tari, J. Bobos, C.S.F. Gomes and J.M.F. Ferreira, J. Colloid Interface Sci., 210 (1999) 360. [52] I. Sondi, and V. Pravdic, J. Colloid Interface Sci., 181 (1996) 463. [53] E. Matijevic, J. Phys. Chem., 65 (1961) 826. [54] A.S. Buchanan and R.C. Oppenheim, Aust. J. Chem., 25 (1972) 1857. [55] A.S. Buchanan and R.C. Oppenheim, Aust. J. Chem., 21 (1968) 2367. [56] R.O. James and T.W. Healy, Parts I, II and III, J. Colloid Interface Sci, 40 (1972) 42, 53, 65. [57] C.F. Baes and R.E. Mesmer, The hydrolysis of cations, R.E. Krieger Publishing, Malabar, 1986. [58] RJ. Hunter and M. James, Clays Clay Miner, 40 (1992) 644. [59] Y. Yukselen and A. Kaya, Water Air Soil Poll, 145 (2003) 155. [60] RJ. Hunter, Adv. Colloid Interface Sci, 17 (1982) 197. [61] F. Thomas, L.J. Michot, D. Vantelon, E. Montarges, B. Prelot, M. Cruchaudet and J.F. Delon, Colloids Surfaces, 159 (1999) 351. [62] R.M. Pashley, Clays Clay Miner, 33 (1985) 193. [63] I. Sondi, J. Biscan and V. Pravdic, J. Colloid Interface Sci, 178 (1996) 514. [64] I. Sondi, and V. Pravdic, J. Colloid Interface Sci, 181 (1996) 463. [65] D.G. Duran, M.M. Ramos-Tejada, F.J. Arroyo and F. Gonzalez-Caballero, J Colloid Interface Sci, 229 (2000) 107. [66] O. Heath, Th.F. Tadros, J. Colloid Interface Sci, 93 (1983) 307. [67] M. Benna, N. K-bir-Ariguib, A. Magnin, F.J. Bergaga, J. Colloid Interface Sci, 218(1999)442. [68] T. Permien and G. Lagaly, Clay Miner, 29 (1994) 761. [69] T. Permien and G. Lagaly, Clay Miner, 29 (1994) 751. [70] R. Keren, I. Shainberg and E. Klein, Soil. Sci. Soc. Am. J , 52 (1988) 76. [71] R. Keren Soil Sci. Soc. Am. J , 53 (1989) 25. [72] B. Rand, E. Peckenc J.W. Goodwin and R.W. Smith, J. Chem. Soc. Faraday Trans, 1,76(1980)225. [73] M. Morvan, D. Espinat, J. Lambard, and Th. Zemb, Colloids Surfaces A, 82 (1994) 193. [74] F. Miano and M.R. Rabaioli, Colloids Surfaces A, 84 (1994) 229. [75] M.J. Avena and C.P. De Pauli, J. Colloid Interface Sci, 202 (1998) 195. [76] R.G. de Kretser, P.J. Scales and D.V. Boger, Colloids Surfaces A, 137 (1998) 307. [77] L. Madrid and E. Diaz-Barrientos, J. Soil Sci, 39 (1988) 215.
Electrokinetic Behavior of Clay Surfaces
89
[78] W.H. Hendershot and L.M. Lavkulich, Soil Sci. Soc. Am., 47 (1983) 1252 [79] E. Cao, R. Bryant and D.J.A. Williams, J Colloid Interface Sci., 179 (1996) 143. [80] F.C. Chang and G. Sposito, J Colloid Inteface Sci., 178 (1996) 555. [81] P. Leroy and A. Revil, J. Colloid Interface Sci., In press (2003). [82] P.F. Low, Soil Sci. Soc. Am. J., 45 (1981) 1074. [83] S.E. Miller and P.F. Low, Langmuir, 6 (1990) 572. [84] C D . Newman and G. Brown, Chemistry of Clays and Clay Minerals, Ed. A.C.D. Newman, Wiley, New York, 1987. [85] Y. Horikawa, R.S. Murray, and J.P. Quirk, Colloids Surfaces, 32 (1988) 181. [86] M.S. Celik, Y. Akin, M. Hancer, Proceedings of Society of Mining Engineers, SME Preprint 96-91, SME Annual Meeting, Phoenix, 1996. [87] K. Brauner and A. Preisinger, Miner. Petr. Mitt., 6 (1956) 120. [88] E. Ruiz-Hitzky, J. Mater. Chem., 11 (2001) 86. [89] M.S. Celik,and E. Yasar, J. Colloid Interface Sci., 173 (1995) 181. [90] J.E. Brady and G.E. Hauston, General Chemistry; Principles and Structure, John Wiley and Sons Inc. New.York, 1975. [91] U. Mart, M. Cinar, B. Ersoy, and M.S. Celik, Proceedings 10th National Clay Semp., Izmir, Turkey, 2003. [92] S. Rossi, P.F. Luckham and Th.F. Tadros, Colloids Surfaces, 201 (2002) 85. [93] J.J. Spitzer, C.A. Midgley, H.S.G. Slooten and K.P. Lok, Colloids Surfaces, 39 (1989) 273. [94] A.M'Pandou and B. Siffert, Colloids Surfaces, 24 (1987) 159. [95] B.V. Derjaguin and L.D. Landau, Acta Physicochim. URSS, 14 (1941) 622. [96] E.J.W. Verwey and J.Th.G. Overbeek, Theory of the stability of lyophobic colloids, Elsevier, Amsterdam, 1948, 168. [97] J.N. Israelechvili, Intermolecular and Surface Forces, Second Edition; Academic Press, San Diego, 1995. [98] J. Gregory, Critical Reviews in Environmental Control., 13 (1989) 185. [99] P. Sennet and J.P. Olivier, Colloidal dispersions, Industrial and Eng. Chem. The Interface Symposium, vol. 57, 1965. [100] K. Furusawa and M. Matsumoto, Electrical Phenomena at Interfaces, Fundamentals, Measurements and Applications, Ed. A. Watanable, Dekker , New York, 1984. [101] J.A. Schwarz, C.T. Driscoll, and A.K. Bhanot, J. Colloid Interface Sci., 97 (1984)55. [102] C.J. Van Oss, Intefacial Forces in Aqueous Media, Dekker, New York, 1994. [103] G. Lagaly, Coagulation and Flocculation, Surfactant Science Series 47, Ed. B. Dobias, Dekker, 1993. [104] B. Ravid, E. Pekenc, J.W. Goodwin and R.W. Smith, J. Chem. Soc. Faraday Trans., 76(1980)225. [105] R.G. de Kretser, P.J. Scales and D.V. Boger, Colloids Surfaces, 137 (1998) 307.
SURFACE THERMODYNAMICS OF CLAYS B.S. JAI PRAKASH Department of Chemistry, Bangalore Institute of Technology, K.R. Road. Bangalore 560 004, INDIA. E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Surface Thermodynamics of Clays
91
1 - Thermodynamics of clay surfaces by surface tension. The effect of surface tension of a solid on the thermodynamic properties is so small that it may be neglected unless the subdivision is exceedingly fine. The behaviour of nanosized particles, although not new to clay scientists [1], is drawing the attention of nanotechnologists, particularly after the advent of techniques such as atomic force microscopy and scanning tunneling microscopy. In spite of the excellent techniques that are available, accurate determination of interface thickness remains a challenge [2]. It was not until the monumental work by the school of van Oss et al [3] on the determination of interfacial tension by judicious selection of interacting liquids, that the apolar and polar contribution of clays could be quantified. An interface is characterized by a surface tension y expressed in terms of force per unit length with units mj m"2 or k T urn"2 where k is Boltzmann constant and T is the temperature in Kelvin. To increase the area of an interface, work must be expended by the system and is given by Equation 1 and 2.
or
dW = -y dA y = -dW/dA
(Eq. 1) (Eq. 2)
The surface tension or surface work y is the reversible work needed to create a surface of unit area under conditions of constant temperature and volume. The term surface tension is somewhat confusing when one talks about solids since it implies the existence of a stress. The term tension is in common usage for liquid surfaces where the process involves stretching the surface, whereas for solids y is related to work spent in forming the surface [4]. The increase in interatomic separation due to stretching seen in the case of liquids, is not observed in the case of solids where the work is spent in forming the surface. In solids, this results in an interface with the medium. At constant volume and temperature, the excess surface energy 9GXS, is related to surface tension y = (3Gx79A)T,v
(Eq. 3)
The surface tension of a solid surface i, given by term y,, is equal to one half the free energy of cohesion [3,65] yi = - I / 2 AG ii
(Eq.4)
Unlike for liquids, where AGi; refers to their free energy of cohesion, the AG;J in solids is the free energy available for interaction with liquids [5]. The surface tension component (STC) theory considers that interfacial tension of a surface can be deduced from a combination of interactions such as apolar and polar interactions called components [6]. Thus, for a surface i, AG can be given as AG = AGapolar + AGpolar
(Eq. 5)
Substituting for AG from Eq. (4) in Eq (5) Yi = yr'" r + yf"
(Eq- 6)
92
B.S. Jai Prakash
Apolar (electrodynamic) interactions have been recognized to be made of three types of van der Waals forces which include orientation, induction and dispersion. Based on the Lifshitz theory of molecular attracting forces between condensed bodies, Chaudhury [7] showed that the three forces could be combined as Lifshitz- van der Waals (LW) forces and the corresponding surface tension term y™. Similarly, the dominant polar interaction has been recognized to be acid-base in nature. Consequently, the surface tension term yj"1" is written as y*B so that y.= y ' w
+
(Eq- 7 )
Y™
1.1 - LW interactions For a surface of clay in contact with a liquid L, the interfacial tension y™ for LW interaction is given by,
Ycr=(V^-V^) 2
(Eq-8)
where yc is the surface tension of the clay surface and ji is that of the pure liquid. LW _
LW ,
Yci. — Te
LW o
^ Yi.
z
( /
LW
tyYc
/
LW )
VYL )
/p
Q^
v^-q- y J
And the corresponding AG is given by the Dupre equation [8], AG^=y--y--y L w
(Eq. 10)
For clay particles in contact with a liquid L, the relation is rewritten as
AG ™ = -2 y - = -2 (^f - Vy^) 2
(Eq. 11)
1.2 - AB (polar) interactions All polar forces on the surface could be regarded as arising from Acid-base (AB) interactions, the acid A being a proton donor (Bronsted acid) or a species that can accept a share in a pair of electrons (Lewis acid); the base B being a proton acceptor (Bronsted base) or a species that can donate a share in a pair of electrons (Lewis base). All types of AB interactions, including hydrogen bonding, may be represented by Lewis electron donating and accepting properties; the corresponding surface tensions are denoted as y ' for electron donating and y+ for electron accepting groups. The free energy of interaction between a clay surface in contact with a liquid is given by the relation AG™
= -2 ( VYTYL+ + 2 VY^YT )
(Eq. 12)
where the electron donor and acceptor parameters of the clay surface and the liquid are designated respectively as y~, y^ and y*, y^. For example, the oxygen atom on the surface donates a share in a pair electron to a molecule of water (Lewis basicity) and a proton in the hydroxyl group on the surface accepting a share of pair of electron (Lewis acidity) from the oxygen of a water molecule.
Surface Thermodynamics of Clays
93
The free energies of interaction arising out of polar and apolar interactions are additive [3,65]. AG = AG LW +AG AB
(Eq. 13)
According to the van Oss - Good - Chaudhury model [9], one can write for the surface tension of a clay surface in contact with a liquid, Yc, = JL: + Y£
(Eq. 14)
2+
+
= (v^-v^] 2(y^i ^+j^i
+^
]
From Dupre equation, the interaction energy between two clay particles immersed in a liquid, given as AGCLC> is AGCLC=-2yCL
LW = -7 ClvYCYC + y- T+ VYLTL L + v' z [/v |y I c - Vl-j^y TL J - H4 W
(Eq. 15)
T
+ V 'CTL L + v"T +V TL/ VT+CT/-I 1
Of the many analytical techniques that are known to survey the surface of solids, the contact angle is known to measure the tension at the precise surface of solids [3]. For determining the different components of surface tension, yLW, Y+ and y ", it is desirable to find the contact angles of liquids on solid surfaces which are known to be apolar and polar. For example, a non polar liquid such as octane will give only the LW interactions on a non polar solid such as PTFE. 1.3 - Determination of contact angle Contact angle measurement is a simple method to adopt that can be used to calculate the surface tension using the famous Young equation (3) YLCOSO = y s - YSL
(Eq. 16)
where G is the contact angle which is obtained by drawing a tangent line from the drop shape starting at the triple point, solid - liquid - air and measuring the angle between the tangent line and the solid surface (Figure la). ySL represents the interfacial tension between the liquid and the solid and y s is the solid surface tension or the free surface energy. The surface, which gets wetted easily with water (hydrophilic), will have lower 9 values (less than 65°) whereas hydrophobic surfaces exhibit higher 9 values. A typical contact angle report chart is shown in Figure 2 (b). Mica shows a large hydrophilic character whereas talc is very hydrophobic. Most of the clay minerals are hydrophilic showing contact angles in the range of 30° to 40°.
94
B.S. Jai Prakash
Figure I - (a) Contact angles made by the solid surface at the air-liquid interface. Spreading liquids which wet the surface show low contact angles. Liquids which do not wet easily show higher values, (b) Typical contact angle chart relating the hydrophilic and hydrophobic character of some solids. Clays show a hydrophilic character while talc is hydrophobic.
1.4 - Surface tension components - The Young- Dupre equation The Dupre equation for a clay surface is given by Eq. 10 as A G C L = YCL-YC-YL
where AGCL is the free energy of the clay surface in contact with liquid L. Combining this with the Young equation (Eq. 15), we get the Young - Dupre equation, as -AGCL=(l+cose)yL
(Eq. 17)
Combining Eqs. 11, 12 and 13 with Eq. 17, we get,
(l+cose)y L =2(Vy c LW Yr + V Y T Y T - V T O O
(Eq. 18)
Determination of surface tension components requires a judicious selection of liquids which can specifically interact through van der Waals forces to measure yLW and through acid-base interactions to measure y*8. Highly apolar solvents which are known to interact through LW interactions include organic solvents such as hydrocarbon liquids (liquid alkanes). Surface tensions of such liquids have been determined by finding the contact angles on apolar surfaces such as teflon. However, for polar liquids, determination of yL+ and yL" poses some problems. Known monopolar solids (which show either acid (y+) or base (y~) character) are used. For example, a solid like polymethylmethacrylate (PMMA) can donate a share in lone pair of electrons of the oxygen atom in the carbonyl group acting like monopolar Lewis acid. Thus, the electron
Surface Thermodynamics of Clays
95
accepting property (y+) of a liquid drop on PMMA could be measured from the contact angle formed. Many other useful methods have been suggested by Giese and van Oss [3] for measuring the surface tension components of liquids. Similarly, for water, having a total surface tension yw of 72.8.mJ m"2, y™ can be measured on a low energy apolar solid such as Teflon. This is found to be 21.8 mJ m"2 and thus the yAB component of water is obtained by the difference 72.8 - 21.8 = 51.0 mJ m"2 [7,10], Since there is no method by which the y w + and yw~ can be separately determined and due to the tendency of the water molecule to easily accept a pair of electrons, unlike many other liquids, it is convenient and generally accepted to set y^, = y^ = 25.5 mJ m"2 [9-13]. Y™ = Yw" + lw + Tw = 21.8+25.0+ 25.0 = 71.8 mJm"2 (Eq. 19) To determine the three surface tension components of a solid, y1"*, y+ and y", it is necessary that at least 3 liquids, whose surface tension values are known, be used for contact angle measurements. The yLW component could be determined using an apolar solvent. The other two liquids should be polar, preferably one with a large y ' value and the other with a large y + value. Water is a bipolar liquid with high electron donating ability (y " = 25.5) as well as a high electron accepting capacity (y+ = 25.5). Most solids when dried predominantly exhibit Lewis base property (y") and a negligible Lewis acid property (y+). When both the tendencies are strong, normally it is due to the wetness or water of hydration present on the solid [14]. The clay surface should therefore be dry before any measurements are made. The school of van Oss et al [15] advocates contact angle measurements with at least 2 polar liquids in addition to water to get an accurate picture of y/^ of a solid surface. The choice of liquids further get restricted because the yL should be greater than ys to get a finite, measurable contact angle. Thus liquid alkanes (C6 -C 16 ) having low y values in the range of 20 to 30 form no contact angle on most high energy solids and hence are not suitable for contact angle measurements of solids like clays, a- bromonaphthalene (y = 44.4, y+= 0,y" = 0) and diiodomethane (y =50.8, y+= 0,y" = 0) are used for determining yLW. In addition to water, formamide (y =58, yLW=39, y+= 2.28,y " = 39.6) and glycerol (y = 64, y LW=34, y+= 3.92,y" = 57.4) are recommended [15]. A clay surface for contact angle measurements could be prepared simply by transferring an aqueous suspension of appropriate concentration of the clay on to a clean microscope slide through a pipette and allowing it to dry overnight. The plate is heated to 110°C, cooled in a vacuum desiccator and equilibrated with the atmosphere before use. Liquid drop is placed on the slide and the contact angle is measured, usually by studying the shape of the drop by image capture. Nowadays high speed cameras are used for image capture and analyzed with computer software. The contact angles measured with the liquids are fitted into the Young- Dupre equation in the form
(1+ C OS9) Y, 2
=
VTTVF+
VK^-V^r
(Eq.20)
The contact angles with apolar liquids are first used to calculate the surface tension due to LW interaction since yL and yL~ would be zero. Therefore,
96
B.S. Jai Prakash
L
= V ^
(Eq.21)
The average value obtained with the two apolar liquids is used to calculate the -y/Yc" . According to the method suggested by van Oss et al [15], the y+ and y~ are calculated by fitting the contact angles obtained with polar liquids say, Lj L2 and L3 in the Young - Dupre equation expressed in the linear form, a = bxi + cx2
a
bx!
cx2
The Lewis basic groups (y "), being the predominant species present on a dry solid surface, require a liquid that can quantitatively measure them. Water, with its very high basicity (y w = 25.5mj m'2) is the liquid of choice. Therefore, out of the three liquids, one is always water, the other two polar liquids could be glycerol (y = 64) and formamide (y = 58). Three simultaneous equations could thus be written with L] L2 and water. The equations could be solved for X] and x2 (i.e.,y c ' and yc+ ) but always including the equation with water contact angle. The results are unreliable if the contact angles for polar liquids are considered omitting that for water [3]. The equations that are used for the calculations are
- VYY] 2
AG"C = -2 y™ = -2 \fif it
= Ycw + Yc™ = Yc* + 2 V Y I Y I , AG"FWC = A G " C + A G ^ C = -2 yCw z
Yew
^ l_VYcYc
T
V'wTw
VYcYw ^ VYwYc J
Giese and van Oss [3] have listed the thermodynamic parameters for clay dispersed in water. Table 1 gives the results of free energy values measured for a montmorillonite sample using the three liquids. The average values for the various thermodynamic parameters are: The Lifshitz van der Waals surface free energy, y Lw = 43.0 mJ m"2, The Lewis acid parameter, yc+ =1.6 mJ m"2, and The Lewis base parameter yc = 36.9 mJ m~2. The higher value of y^ indicates the predominance of the oxygen of the surface hydroxyl group on clays in donating a pair of electrons to the surrounding water molecules. The magnitude of y LW is also indicative of the large tendency of the surface to hold through van der Waals forces. The interfacial energy AGJ.FWC obtained by summing up AG™C and AG*^C are mostly + ve with an average value of 10.6 for clays. The value changes to -ve (or low + ve) when smectite clays are exchanged with Li+, Na+, K+ and Cs+ The free energy values indicate that the majority of the clays have AG >0 and therefore form stable suspensions with water (hydrophilic character). Some silicates such as talc and pyrophylite, on the other hand, show -ve values of AG. The hydrophobic character of talc is also indicated by high contact angles (75°) and low yAB values (4mJ m"2), both y+ and y" values being low.
Surface Thermodynamics of Clays
97
OTable 1 - Thermodynamic parameters for Wyoming montmorillonite saturated with a specific metal cation and organic cations with different number of carbon atoms and tetraalkyl ammonium ions. Reprinted by kind permission of [Colloid and Surface Properties of Clays and Related Minerals (231-243,105, 2002)]. Swy I cec = 68 meq/lOOg
AG
AG
AG
LW
K
na
NH4
Cs
li
Mg
Ca
Ba
Sr
Nat
-6.6
-7.,1
-6.4
-6.:5
-6.7
-7.1
-6.9
-5.6
-6.2
-6.1
7.6
15 .4
16.2
2.7
7.2
25.1
21.5
-8.1
7.7
11.0
1.02
8.:50
9.80
-4. 14
0.48
18.00
14.65
-13.74
1.43
4.83
iWi
AB
iWi
IF
iWi
« 0 9.9
AG
6 -38 .5
7 -40.3
NUMBER OF CARBON ATOMS" 8 11 12 9 10 -47.4 -52.1 -42.0 -44.0 -25.1
13 -71.3
14 -53.8!
15 -89.6
IF
iWi
<: IF
AG iWi
ORGANIC CATIONS TMA -25.6
HDTMA -42.0
TMPA -31.6
TEA -15.9
TMA = tetramethyl ammonium, HDTMA = hexadecyl trimethyl ammonium, TMPA = trimethylphenyl ammonium, TEA = tetra ethyl ammonium The higher value of yc indicates the predominance of the oxygen of the surface hydroxyl group on clays in donating a pair of electrons to the surrounding water molecules. The magnitude of y LW Is also indicative of the large tendency of the surface to hold through van der Waals forces. The interfacial energy AG[,FWC obtained by summing up AG™C and AG*®C are mostly +ve with an average value of 10.6 for clays. The value changes to -ve (or low +ve) when smectite clays are exchanged with Li+, Na+, K+ and Cs+ The free energy values indicate that the majority of the clays have AG >0 and therefore form stable suspensions with water (hydrophilic character). Some silicates such as talc and pyrophylite, on the other hand, show -ve values of AG. The hydrophobic character of talc is also indicated by high contact angles (75°) and low y*6 values (4mJ m"2), both y* and y' values being low. Interaction of clays with organic substances, both by adsorption on the surface (e.g., amine) and by ion- exchange (e.g., quaternary ammonium cation) renders the clay mineral hydrophobic (AGIF = -ve). This takes place by the neutralization of Lewis basic sites (O atoms on the surface) by the organic cation. Consequently, there will be a reduction of the term YL+YC in the equations given above causing a reversal of sign on the AG jfwc term resulting in the organo clay becoming hydrophobic (Table 1). 1.5 - Hydrophobic and hydrophilic nature of clays - relation to interparticle free energy. The behaviour of clay minerals to form stable aqueous suspension is also attributed to the layer charge. In the case of silicate minerals the charge at the edge sites
98
B.S. Jai Prakash
and that arising due to isomorphous replacement necessitates the addition of another term AGEL in the equation of total interaction energy between the clay particles immersed in water. AG <£ = AG ™c + AG - c + AG fwc
(Eq. 23)
where AGEL expresses the electrostatic contribution arising due to charge. A positive AG ""c would show a hydrophilic nature indicating the tendency of the clay particles to stay in a suspension whereas a negative value would show hydrophobic character resulting in flocculation. The total interaction free energy depends on the following: i) interparticle distance - the interaction energy varies with the variation in the distance between the particles, ii) geometry of the particles.
Figure 2 - (a) A classical DLVO plot of the free energy of interaction (in KT) between ljjm spherical particles ofhectorite as a function of the interparticle distance (in nm) immersed in water of different NaCl concentrations. These concentrations (in M) are shown on each curve. For all concentrations, at contact, a substantial repulsion between particles would be expected, (b). An XDLVO plot of the free energy of interaction (in kT) between 1 jMn spherical particles of hectorite as a function of the interparticle distance (in nm) immersed in water of different NaCl concentrations. These concentrations (in M) are shown for each curve. For all concentrations, at contact, there is a substantial attraction between particles. A suspension in lMNaCl is predicted to flocculate, the 0.1M suspension has an energy barrier at close approach with a secondary minimum beyond 5 nm, while for all other concentrations there is a substantial energy barrier as the particles approach each other. Reprinted by kind permission of [Colloid and Surface Properties of Clays and Related Minerals (205-206, 105, 2002) J.
Surface Thermodynamics of Clays
99
The contribution of electrostatic interaction AGEL could be determined by studying the particle potential fo, the C, - potential and the thickness of the diffuse ionic double layer by electrokinetic methods that have been developed over recent years. Some of these include microelectrophoresis, electroosmosis, streaming potential and sedimentation potential. The AG ^ c values for clay particles may range from 2 to 3 kT (1 kT = 4.045 xio 2 1 J at 25°C) per nm2 of surface at contact point. This value is dependent on the electrolyte concentration and varies with the distance between the particles and their shape. In general, the interaction energy is proportional to the particle radius. For irregularly shaped particles, such as in clays, the repulsion between particles is much smaller compared to spherical bodies. The lower the ionic strength, the greater the stability [5]. This is attributed to the relatively large thickness of the diffuse double layer, which results in maximum EL repulsion. The XDLVO analysis (a plot of AG ^ which is the sum total of all the three types of interaction energies expressed in kT against L, the distance between the particles in nm) gives a clear picture of the stability of clay suspensions at different ionic concentrations, van Oss et al [16] have studied the hydrophobic nature of a hectorite sample by XDLVO which correctly predicts the stability of hectorite as a function of ionic strength. However, the classical DLVO analysis (a plot of AG ™c + AGjiwc against particle distance at different ionic strength) failed to predict the behaviour of hectorite. Figure 2 shows the results obtained by van Oss et.al. [3] on a sample of hectorite. Ignoring AB contribution in studying the stability of colloidal particles in polar solvents like water would lead to erroneous conclusions. The hydrophilicity of smectites are attributed to the strong Lewis basicity ( y" ~ 40 mj m"2) of the oxygen atoms on their surface. 1.6 - Hydrophobicity of talc and pyrophylite Talc and pyrophylite have low yAB values (y+ » y" « 1.7 to 6.5 mJ m"2 and consequently have weak Lewis acioTbase interactions with water molecules resulting in a highly hydrophobic material. These two materials do not form films suitable for contact angle measurements. A method suggested by Giese et al[17] referred to as thin layer wicking involves measuring capillary flow rate of liquid through a thin uniform layer of a powdered material deposited on a smooth glass plate. The capillary flow rate is related to the contact angle by the Washburn equation [18]. h2 = t Reff yLcos 6 / 2n
(Eq. 24)
where h is the height of the capillary rise of the liquid at time t, yL and r| respectively are the surface tension and viscosity of the liquid. R is the average pore radius and is obtained by measuring h / 1 for a low energy wetting liquid like alkane whose contact angle is assumed to be zero (cos 6 = 1 ) . The thin layer wicking method is not recommended for swelling clays. Talc and pyrophylite exhibit hydrophobic character because of their weak tendency for electron donicity (y" « 7 mJ m"2) showing that the oxygens on the surface of these minerals, perhaps by the absence of layer charge, are not influenced to share their lone pair of electrons easily. The basal oxygens thus are electrostatically saturated and incapable of accepting hydrogen bonds [5,17]. Micas and smectites, on the other hand, have layer charges and also exhibit a high degree of electron donicity (y"« 40-60 mJ m"2). The negative charge on the surface of smectites will cause electron repulsion on the oxygen atoms directing the lone pair away from the
100
B.S.JaiPrakash
001 plane enabling their donation to form hydrogen bonds. They can then interact with water and organic molecules and such interactions are known to depend on the layer charge and the type of exchangeable cations as established by many catalytic reactions on their surfaces involving Lewis acidity and basicity [19, and references therein]. A systematic study of the surface tension arising due to electron donicity on the surfaces (001 plane) and electron accepting property on the lateral surfaces of smectites would thus throw light on the mechanism of catalytic reactions on clay surfaces. Surface modification with cationic surfactants are known to hydrophobize the clay surfaces. Such a surface may be considered to be a mosaic surface of clay (1) and the surfactant (2). Then, from Casie's equation [20], the contact angle of a liquid on the mosaic surface Gi2 is related to the contact angles of the liquid on the clay surface (60 and on the surfactant modified surface (02) cose 12 = fcos9i + (l-f)cose 2
(Eq. 25)
where f is the fraction of the surface covered, van Oss and Giese [15] have shown that in the case of octadecylamine coated talc particles, the interfacial free energy AG jTwc of the coated talc particle became more negative (hydrophobic) with the increase in fraction of the surface covered. Van Oss and Giese have also studied the influence of particle size on the hydrophilic nature of clays and related minerals. A decrease in y" (electron donating property) was noticed on grinding which is attributed to the increase in the electron acceptor sites ( y ) which neutralizes the adjacent (y") sites resulting in a material with high AG jj,; (hydrophilic) to change over to low AG *m (hydrophobic). 2 - Immersion enthalpy studies on clays Medout - Marere et al [21] studied the thermodynamic aspects of the immersion of an Algerian montmorillonite. This analysis was to interpret the high value of the enthalpy of immersion in water obtained with swelling clays. This is attributed to interactions between exchange cations and interfacial water, which furnishes some heat during swelling. A microcalorimeter was used to measure the wetting enthalpy AWH (when the solid is fully covered by the vapour of the liquid before immersion). AHimm is given by H C L - Hc° where H C L is the interfacial enthalpy when the clay is in contact with the liquid, and the enthalpy when the clay is kept in vacuum is H c °. Based on the approach by van Oss, Chaudhury and Good (VCG model), the authors related the interfacial surface enthalpy A ^ H to various surface enthalpy parameters - apolar, polar, chemical and electrochemical. A ^ f H = A ^ " r H + A^" r H + A^e™oalb<)ndH
+A^'™"kffectH
(Eq.26)
In physisorption, the third and the fourth terms on the right hand side of the equation are zero. Medout - Marere et al [21] found out the immersion enthalpy in heptane, benzene, formamide and water for three clay samples- kaolinite, illite and montmorillonite. The average values of immersion enthalpies per m2 of the surface were comparable for kaolinite and illite. The authors deduced that montmorillonite, a swelling clay, having composition close to that of kaolinite and illite, would also give values of the same order. (H° 650 mJ m"2 and AH imm in heptane 100 mj m"2). This suggested a possibility of getting an idea of the specific surface area from the
Surface Thermodynamics of Clays
101
experimental immersion enthalpy per unit weight of montmorillonite. The wetting enthalpy Aw H could be related to the specific surface area, Asp and the solid - vapour surface enthalpy H s v and the solid - liquid surface enthalpy HSL [22,23] by the relation (Eq.27)
-AWH = (HSV-HSL)ASP
Combining this equation with Gibbs - Helmholtz relationship and then with Young equation, one obtains -A W H=
f T (dyLcos9) ASP YL cosG- -j—
^1
(Eq. 28)
This implies that in cases of perfect wetting, i.e., when contact angle is zero (See [22]) -A W H = A SP H LV
(Eq. 29)
where HLV is the surface enthalpy of the liquid which in the case of water at ambient temperature is 119 mj m"2. Using the above equation, the authors calculated the specific surface area of montmorillonite clay for different adsorbents at different partial pressures. The values obtained are as follows: A*1 (N2) = 44m2/g , A ° ' = 99m2/g ; A"
(H2O) =110 m2/g and A " (H2O) = 235 m2/g where the superscripts 0.1 and 0.7
are taken as mean values for the partial pressure range for the corresponding adsorbent. The increase in the surface area in the case of adsorption of water is linked to the variation of the interlayer distance (from 9.8 A to 11.9 A) measured by XRD. The continuous changing of surface topology is attributed by the authors to the driving force derived from the chemical potential of the adsorbent molecule. At ambient temperature, for low pressures of interacting solvent and for any pressure of apolar solvents, the surface area has a value of 110 m2/g. At relatively high water pressure the montmorillonite expands to a constant value of 235 m2/g which, very near to the saturation vapour pressure, attains a value of'277 m2/g. By fixing a mean surface area value of 180 m2 g"1, Medout- Marere et al computed the enthalpy values but found some disagreement between computed and experimental immersion enthalpies. They have proposed the introduction of another energetic term, swelling enthalpy, HsWE which is interlinked to the interlayer cohesion energy. It is however, to be noted that the results are not in agreement with a finite contact angle between water and montmorillonite. Douillard [6] has concluded that the combined measurements of AadsG ^" and AadsH *f"°° for different probe liquids (apolar, monopolar and bipolar) will allow surface entropy phenomena to be understood and a correct theory of surface enthalpy components could then be constructed. Only this approach can overcome the approximations of the Good -van Oss - Chaudury theory, so useful even in the approximated form [21] 2.1 - Adsorption enthalpy and entropy studies on smectite surfaces Adsorption of cationic species, both inorganic and organic, within the interlayer of smectites open up the surface resulting in swelling. Study of such new surfaces presents certain problems. A complete understanding of the fresh surface could be obtained by studying the thermodynamic changes that occur during the swelling process. The changes that accompany could be estimated by simple methods of
102
B.S.JaiPrakash
measuring heat changes normally adopted in adsorption studies. The adsorption isotherm obtained reveal a great deal of information about sorption characteristics of the active sites of adsorption at different concentrations. Different approaches adopted using various equations are considered below. It is known that the much higher affinities of organic cations (than inorganic cations) for the clay interlayer is partly due to the lateral interactions of the adsorbed organic cations (cooperative adsorption). The most extensively studied organic substituents are quaternary ammonium ions of the type [RN(CH3)3]+ [24] and various n - primary aliphatic amine salts. Cowan and White [25] made a systematic study on the cation- exchange reaction occurring between Na - montmorillonite and various n primary aliphatic amine salts. Based on the free energy of adsorption obtained for the various amines (from 2 carbon atoms to 10 carbon atoms), they found that in the case of the lower amines up to 6 carbon atoms, the exchange was minimum but showed an increase in tendency for exchange with the increase in the number of carbon atoms . Thus, salts of n - heptyl amine and n - octyl amine exchanged completely while n -decyl amine adsorbed higher than the CEC. The adsorption depended on the amine size and the adsorptive effect due to van der Waals force was found to be the operative factor. Vansant and Peeters [26] investigated the exchange of various alkyl ammonium cations from aqueous solutions by sodium laponite. The affinity of the clay for these organic cations was linearly related to the molecular weight and molecular size or chain length of the alkyl ammonium ions. The overall equilibrium constant, Ka for the exchange reaction Na - clay + Alk+ ^ Alk - clay + Na+ was estimated by the equation given by Gaines and Thomas [27]
lnKa= j l n K f dN^ o where N ^ is the equivalent fraction of the alkyl ammonium ions on the clay and K ^ is given as T^Alk K
Na
Aik-mN,
=
N
~
~
(Eq-31)
N
, , ,
Na-mA,k
where N and m are the equivalent fractions of the ions on the clay and in the liquid phase, respectively. The variations of -RTlnKa (or AGa), a non standard free energy of exchange, showed that the affinity of alkyl ammonium ions for the clay decreases in the order R3NH+ > R2NH2+ > RiNH 3 + . The average increment of AGa per -CH 2 group for the straight chain monoalkyl ammonium ions was 1.35 kj mol"1. The increments per CH3 and -C2H5 going from monomethyl to trimethyl ammonium and monoethyl to triethyl ammonium, respectively were 1.86 kJ mol"1 and 3.69 kJ mol"1 This was attributed to the increased contribution of van der Waals forces to the adsorption energy. This, the authors expected, would be for a flat orientation, as van der Waals forces are additive and hence increase as the size of the adsorbed cation becomes larger. X-ray
Surface Thermodynamics of Clays
103
analysis indicated a single layer of organic ions in the interlamellar space possibly with their shorter axis perpendicular to the clay surface. A comparison in exchange behaviour of the monoalkyl ammonium ions for the montmorillonite and the laponite clay minerals reveal that the affinity of the organic cations for the Na - montmorillonite is larger compared to the Na - laponite. This was attributed by Vansant and Peeters [26] to the higher surface charge density of the Na - montmorillonite compared to the laponite clay. 2.1.1 - DKR equation. Many publications have appeared recently focusing mainly on the adsorption sites in the interlamellar region which open up on swelling. From the thermodynamic point of view, the enthalpic changes caused by the interaction of molecules responsible for swelling involve both the free energy changes contributed by the LW and AB interactions of the surface and entropic changes of the hydration sphere of the exchangeable cations. The thermodynamics of such surfaces could be studied from the energy changes accompanying the swelling considering the uptake of the the adsorbate molecules as a function of concentration. One such study is provided by the DR equation. Dubinin - Radushkevitch (DR) equation measures the surface characteristics from the low and medium pressure parts of the adsorption isotherm [28]. The D-R coverage is the greatest coverage when the lateral adsorbate-adsorbate interactions can be neglected [29]. This is based on the Polanyi's theory of adsorption [30] which is related to the equilibrium concentration as 6 = RTln(l/C E )
(Eq. 32)
where s is the Polanyi's potential which is, according to Gregg and Singh [28], clearly equal to -AG, the differential free energy of adsorption, R is the gas constant, T is the temperature in kelvin and CE is the equilibrium concentration of the adsorbate. According to Polanyi's treatment, the adsorption space in the vicinity of a solid surface, such as in microporous solids, is characterized by a series of equipotential surfaces [31]. Polanyi pictured the adsorbate as in intimate contact with the solid leading to the micropore filling. The adsorption potential arising due to the dispersion and polar forces between the solid and the adsorbate molecules was assumed to be of Gaussian distribution. Kaganer in 1959 [32] modified the model of micropore filling by introducing the concept of surface coverage and evaluating the surface area from equilibrium uptake curves. The new equation, often called the Dubinin - Kaganer Radushkevitch (DKR) equation thus relates the surface coverage with respect to monolayer capacity at low and intermediate coverages. lnCads = In Cm - (3s2
(Eq. 33)
where Cm is DKR monolayer capacity which is always less than that measured by the BET method [29], Cads is the amount adsorbed in the DKR region of the isotherm. When lnCads is plotted against s2, a straight line is obtained. The slope of the curve gives the value of p (mol2 J"2) and the intercept yields the value of sorption capacity, Cm(mmol kg" '). The value of p is related to sorption energy E via the following relationship given by Hobson [29]
104
B.S. Jai Prakash
s
=
-1 ^ =
(Eq. 34)
V-2(3
2.1.2 - Adsorption of long chain quaternary amine cations Clays, which are inherently hydrophilic due to hydration of metal ions show a hydrophobic character when they are interacted with surfactants having long chain quaternary amine cations. Cationic surfactants are known to swell montmorillonite; the gallery expansion being attributed to the grafting of the surfactant molecule to silanol groups within the interlayer where the silica framework is in contact with the clay layers. However, the sorption lowers the BET surface area, apparently due to the constriction of the pore channels as a result of attachment of the surfactant moieties to the internal framework surfaces. The quaternary amine organic cations can replace the metal ions by ion exchange. Among the quaternary ammonium cations studied, hexadecyltrimethyl ammonium ion (HDTMA) is one of the most effective in modifying the sorptive capabilities of clays. The organic clays thus prepared have been extensively studied for adsorption and isolation of organic contaminants in aqueous medium [33,34,35]. Boyd et al [36] have modified a smectite clay with hexadecyl trimethyl ammonium (HDTMA - clay) and have studied the adsorption of trichloroethylene and benzene. Lawrence et al [37] have examined the sorption of phenols and chlorinated phenols from aqueous suspensions by tetramethyl ammonium (TMA) and tetramethylphosphonium (TMP) modified smectite clays. Some papers have been published recently on the thermodynamic studies of the uptake of inorganic anions by surfactant modified swelling clays. For swelling clays, the specific surface area is not constant and depends on the opening of sheets. This makes the evaluation of thermodynamic parameters in the case of swelling clays less easy. A few papers that have mainly focused on the evaluation of thermodynamic parameters from equilibrium uptake curves are discussed below. Lee et al [38] have reported that soil samples exchanged with organic cations of the form [(CH3)3 NR]+, where R is a C9 - Ci6 hydrocarbon, displayed high sorptive uptake of common ground water contaminants. Grim et al [39] have investigated the sorption of n - butyl amine, n- dodecyl amine and ethyl dimethyl octadecanyl amine on clays. Adsorption of the smaller molecules of n - butyl amine did not occur beyond the cec of the clays even when the amine cation was present in large excess in the solution. On the other hand, adsorption of larger amines did not stop at cec, but went on beyond twice the value of cec. The excess amount adsorbed beyond the cec of the clay was ascribed to van der Waals interactions. Similar reports have been made by many workers including Zhang et al [40 and the references therein]. In general, it is observed that the affinity of clays for the organic cations increased with the chain length and van der Waals forces contributed more to the adsorption energy [26,41]. Theng et al [42] have noted that the relationship between free energy change and chain length is of general applicability to the adsorption of organic compounds on montmorillonite owing to the increased contribution of van der Waals forces to the adsorption energy. For the exchange reaction,
Surface Thermodynamics of Clays
105
Clay-M + amine+ ** Clay-amine + M+ attempts have been made to evaluate the molar free energy change that would be associated with the increase in chain length.[25,42]. The selectivity coefficient Km for the reaction is given by [40] Km = (CM+/ N M + ) x (Namine / Camme)
(Eq. 35)
where M+ denotes monovalent metal ions such as Na+ and K+ , C represents the concentration of the appropriate ion in solution and N is the mole fraction of the specified ions exchanged in the solid phase. Cowan and White [25] used the Km value to calculate the molar free energy change AGm for the cation exchange reaction. Adsorption of amylamine to decylamine showed a linear relationship between AGm and the number of carbon atoms in the alkyl chain. The increment of AGm per - CH2 group was found to be 1.67 kJ mol"1. Sullivan et al [43] have studied the thermodynamics of cationic surfactant sorption on to a natural clinoptilolite, a zeolite having two dimensional channel system that allows the mineral to act as a molecular sieve. Large surfactant molecules such as HDTMA sorb only to the external surface of the zeolite. In such cases the external cation exchange capacity (ECEC) characterizes the exchange capacity of the mineral surface for HDTMA. Further, the enthalpic changes due to swelling noticed in the case of montmorillonite will be absent and therefore this study by Sullivan et al, could be comparable to non-swelling aluminosilicate clays such as kaolinite. Sorption enthalpies of HDTMA as monomers and micelles and tetraethyl ammonium bromide (TEA) have been studied. TEA, which does not form micelles, was used to approximate the sorptive behaviour of an amine head group, to further elucidate the effect of the HDTMA tail group on the sorption process. The sorption mechanism of HDTMA was investigated using sorption isotherms and calorimetry. The heat of sorption was measured for a range of surfactant coverages to elucidate the surface - surfactant bond strength and the data was used by Sullivan et al, to develop a thermodynamic description of the sorption process as given below. For sorbed quantities of surfactant less than ECEC, sorption was described as occurring via cation exchange and above ECEC through van der Waals forces. The respective ideal equilibrium constants for a divalent exchangeable cation were given by
and
Kideai(<ECEC) = (X+/ [S+] x ([M2+]°5 / XM2+ °'5)
(Eq. 36)
Kldeal(>ECEc) = N s + / [S+]
(Eq. 37)
where Xs+ is the mole fraction of monovalent surfactant cation S+ on the zeolite surface and XM2+ is the mole fraction of exchangeable metal ion M2+on the zeolite surface given by the equations
and
X s + =N s + /(n s + + nM2+)
(Eq. 38)
XM2+ = nM2+/ (N s+ + NM2+)
(Eq. 39)
106
B.S. Jai Prakash
where n refers to the amount of the material sorbed in mmol kg"1 of clinoptilolite. [S+] and [M2+] are respectively the concentrations of the monovalent surfactant ions and the exchangeable cations in solution. For the case where there are no data for solution or surface activities, deviation from ideal behaviour in the system was expressed by the equation lnK=lnK ideal + f(ns+)
(Eq. 40)
where f(ns+) is a function of the amount of surfactant sorbed ns+. Isotherm data was regressed in terms of lnKj
(Eq.41)
The enthalpy of sorption was measured directly to estimate AH^ and the entropy of sorption was calculated from the relation -(AG^, -AH° ) AS0
=
(Eq. 42) T
Both the monomer and the micelle forms of HDTMA showed similar enthalpies of sorption. Below the ECEC, the enthalpy data from the micelle system did not reflect an additional contribution from demicellization. The measured enthalpies of sorption of HDTMA below ECEC was around -11.98 2.85 kJ mol"1 which was close to the enthalpy of micellization (»-9.7kJ mol"1). This made the authors to suggest that the dominant enthalpic effect is the transfer of HDTMA from the water phase to the bound phase, regardless of the HDTMA structure on the surface or in solution. In the bound phase the micelle appears to rearrange to a monolayer slowly. Slightly endothermic value of the sorption of TEA (+3.03 \.19kJmo\~l) obtained by Tamura et al [44] was attributed to the disruption of the interfacial water structure and release of water molecules from the surface. Enthalpy of sorption of TEA (which cannot form a micelle) would approximately give the enthalpy of sorption of the head group in HDTMA. Thus the difference between the two enthalpies (»7kJ mol"1) would give an estimate of the van der Waals, tail - tail or tail - surface interactions. AG ° values were found to become more negative with increasing tail length. According to Sullivan et al [43], AG °m can be broken into electric (AGeiec) and specific (AGspec) additive components. The polar head group dominates AGspec, while the chain chain and chain - surface interactions usually dominate AGspec. Thus AGspec is approximately equal to the AG °m for TEA as the latter lacks a tail group. One can obtain AGspec by subtracting AG "m for TEA from the total AG °m . The value of AG ° for TEA is small and thus the contribution to AG^ is mostly from the AGspec. The latter is the driving force for the sorption of HDTMA and explains why HDTMA monomers sorb as admicelles of bilayer, on the aluminosilicate surface studied, even when the solution concentration is less than CMC. Based on the free energy values of micellization and the hydrophobic energy associated with the transfer of surfactant chains from water into
Surface Thermodynamics of Clays
107
the bulk phase reported by earlier workers [45,46,47] and the values obtained in this work, Sullivan et al concluded that fewer -CH 2 groups actively participated in the hydrophobic bonding. Unlike in the case of small molecules, entropy changes accompanying sorption of surfactants are governed by many factors which include, mobility of long chain molecules, rigidity of the head group, random movement of the flexible tail, tail - tail interactions, the hydration sphere, the cohesive energy of the medium, the hydration sphere of the substrate itself and its cec. High affinity for the surface (-AG) drives the hydrophobic part of the surfactant molecules to get rid of their hydration sphere and to overcome the high cohesion energy density of water. Sorption, in general, leads to an arrangement of surfactant on the surface and generally there is a net loss in entropy (disorder) upon adsorption. At the same time the surfactant may have to overcome the well-arranged hydration sphere by water destructuring resulting in the water phase gaining entropy. Sullivan et al [43] have made a relative comparison of the sorption characteristics between the HDTMA systems and the less hydrophobic and less mobile TEA molecule. They have illustrated the weak to strong water destructuring in the case of monomer sorption and both the moderate water structuring and the weak to moderate water destructuring in the case of micelle sorption (Fig.3). Sorption of TEA produced a moderately small positive value of TASm° (+4.07 kJ mo,r') whereas sorption of HDTMA at less than ECEC (low loading values) showed a small net loss of entropy (l.lOkJmol" 1 ). Lee and Kim [48] have studied HDTMA adsorption by smectite at different loadings of HDTMA. They observed that swelling took place at different stages depending on the HDTMA loading. The d(001) spacing was 14A at HDTMA < 0.5 CEC suggesting a monolayer arrangement. Increasing the added HDTMA from 0.5 to 1.5 CEC produced a bilayer arrangement (17 - 18 A) or a pseudotrimolecular layer (~22A). A spacing of 40A was also observed at higher HDTMA concentrations. Surfactant orientation would thus depend on the quantity of surfactant in the interlayer. Interestingly, the HDTMA adsorbed at higher concentrations could not be removed easily by washing with water showing a greater binding energy characteristic of ion dipole and ion - ion interactions. The isotherm had a low HDTMA sorption initially but increased dramatically to a maximum around 0.7 cec. According to Lee and Kim [48] this indicates that the cation exchange of HDTMA for Na in interlayers is not easy at an early stage due to the difficulty in prising open the layers by very large HDTMA cations. The interlayers once opened, may be very suitable to adsorb further HDTMA. There was a large variation in the d(001) spacings depending on the relative humidity (%RH), surfactant loading and its orientation. The surfactant orientation and its structure depend also on the charge distribution and chemical composition of the smectite in addition to the quantity of surfactant. However, very few attempts have been made so far to study the thermodynamics of adsorption of surfactants with the changing topology as a result of swelling in smectites.
108
B.S. Jai Prakash
Figure 3 - Conceptual model ofentropic changes due to water destructuring caused by sorption of TEA and HDTMA to natural clinoptilolite. Sorption of TEA: (A) electrostatic interactions and weak water destructuring. Sorption of HDTMA monomers: (B) electrostatic and weak water destructuring; (C) electrostatic and moderate water destructuring; (D) electrostatic and strong water destructuring; (E) electrostatic and weak water destructuring. Sorption of HDTMA micelles: (F) electrostatic and moderate water structuring; (G) electrostatic and weak to moderate water destructuring. Reprinted by kind permission of [J. Coll. Interface Sci. (206, 369, 1998)] 2.2 - Thcrmodv nainic studies of adsorption on the modified clay surfaces. Krishna et al [49,50] have studied the thermodynamics of the adsorption of two nonionic surfactants. Tween -80 (polyoxyethylene sorbitan monooleate) and PEG -300 (polyethylene glycol-300) on an Indian montmorillonite sample. In order to determine the surface heterogeneity, Scatchard equation was used. The adsorption data were fitted to the Scatchard equation in its simplest and most utilized linear form [51] X — =k s n s -k s X Ce
(Eq. 43)
Surface Thermodynamics of Clays
109
where X is the amount of adsorbate adsorbed per gram of the adsorbent, Ce is the equilibrium concentration (umol ml"1), and ks (slope) and ns (intercept) are Scatchard constants. Scatchard plot gives a measure of the fraction of the adsorbate retained on the solid at different adsorbate concentrations. A steep decrease in the fraction adsorbed (a higher slope of the isotherm line) would indicate a more active site and conversely a line with lower gradient would indicate a lesser active site. Thus, breaks in the Scatchard plot (X/Ce > < X) show a non - homogeneous surface with different stability constant. Adsorption of tween - 80 on the clay surface at 298 K and 318 K showed a clear break in the curve mentioned as region I in Figure 4(a) indicating stronger sites of adsorption (ks = 16.2 at 283K).
Figure 4 - (a) Scatchard plot of tween-80 adsorbent on modified clay (b) Scatchard plot ofPEG300 adsorption on unmodified clay. Reprinted by kind permission of [Bull. Mater. Sci. (21, 359, 1998)].
The AH0 reported for the tween -80 adsorption is -86.8 kj mol"1 indicating chemisorption. The Scatchard plot of adsorption of PEG -300 on clay surface shows two regions similar to tween-80 adsorption. The sites, however, are much weaker (ks=2.78 x 10"2 at 283 K) and the adsorption is physical as indicated by the heat of adsorption (AH = -8.5kJ mol"1). Krishna et al [49] carried out studies on the sorption of iodine on montmorillonite chemically modified with tween - 80 and PEG -300. The Scatchard plot at 283 K showed two breaks with 3 regions of adsorption (Figure 5). The structure of tween -80, polyoxyethylene sorbitan monooleate, consists of three moieties - hydrophilic ethylene oxide chains, sorbitan and a hydrophobic group. The breaks in the curves are attributed to the iodine adsorption on these three groups. However, at 318 K (Figure 5) there was only one break in the curve with two regions; one of the regions, perhaps the one that showed a very weak adsorption at 283 K, was absent at 318 K. The authors attributed strong adsorption at region I to be due to oleic group (ks = 2.6 x 10"1), very weak adsorption at region II to be due to ethylene oxide groups (ks = 7.4 x 10"2 at 283 K) and region III was attributed to adsorbate - adsorbate interaction that takes place once the surface is saturated and is stronger (ks = 4 x 10"1 at 283 K) than adsorption over polyoxyethylene sites. PEG-300, however, showed no breaks in Scatchard plots
110
B.S. Jai Prakash
showing a homogeneous surface consisting of only ethylene oxide groups. The ks value at 283 K (2.3 x 10"2), was comparable with that at region II at 283 K for tween-80. The thermodynamic parameters for the sorption of the surfactants on the clay surface and that for iodine sorption are given in Table 2 . Meckamer and Assaad [52] treated a Mg2+- exchanged vermiculite sample with varied amounts of Sr2+ ions and studied the thermodynamics of the release of Mg2+ in the presence and absence of polyvinyl alcohol (PVA). The Sr - Mg exchange reaction represented as CMg + C S r ->Cs r +C M g
(Eq.44)
where C and C represent the concentrations of the ions in the solution phase and on the clay surface respectively. To study the effect of PVA (M.Wt. = 72,000) on the interlayer magnesium release, samples of Mg - clay were suspended in PVA solution and varied amounts of SrCl2 were added and shaken. The amount of Mg2+released was determined.
Figure5 - Scatchard Plot of iodine adsorption on tween-80 modified clay. Reprinted kind permission of [Bull. Mater. Sci. (21, 359,1998)].
by
Table 2 - Scatchard constants for surfactant and iodine adsorption on unmodified clays (Reprinted by kind permission of [Clay Res. ( 1 8 , 1 , 1 9 9 ) ] .
-AG° kJ/mol"1 -AS0 J/mol" i K -i
-AH0 kJ/mol"1
T (K)
Tween -80 adsorption on montmorillonite
Iodine adsorption on tween-80 modified clay
PEG-300 adsorption on montmorillonite
278 318 278 318
43.4 39.5 156 156 86.8
30.4 33.0 -66.4 -66.4 11.9
30.0 34.1 -103 -103 1.47
Iodine adsorption on PEG modified clay 12.0 12.9 -2.25 -2.25 5.50
Surface Thermodynamics of Clays
111
The experiments were also conducted in the absence of PVA. The Sr - Mg isotherms at 20°C indicated that the affinity decreased as the unhydrated ionic radius increased, i.e., Mg was much more selective than Sr ion. This was reverse of what was found in organic resins and most other clay minerals [53,54,55] in which the radii of the fully hydrated ions determine the distance of the approach of the cations to the negatively charged surface. This unusual behaviour was attributed to the strong attraction forces between the vermiculite sheets which prevent the full hydration of the interlayer ions. The thermodynamics of these interactions were studied by the following equation 1
lnK= JlnK c dX S r o
(Eq. 45)
where K is the thermodynamic equilibrium constant, K c is the corrected selectivity coefficient for the preferences of the solution phase and XSr is equal to Csr / Co. The standard free energy of the exchange reaction was calculated from AG° = -RTln K and the standard enthalpy using the vant Hoff relationship In [K40 / K20] = AH0 / R (1/ T40 - I/T20)
(Eq. 46)
and the standard entropy AS0 was calculated from the relationship AG° = AHC - TAS° The value of the enthalpy, AH0 (+22.57 cal/ mol) indicates that the exchange reaction is endothermic in the direction of forming Sr - vermiculite. This was attributed by Mekhamer and Assaad to the expansion of the interlamellar space to accommodate the larger Sr2+ ion. They obtained a value of+6.79 cal/degree mol which is the algebraic sum of all entropy changes in the system of reaction CMg + C Sr -» CSr + C Mg which may be given as AS0 = AS ^
AS L i .
(Eq.47)
Substituting a value for AS0 as -18.6 cal / degree - mol, obtained by Home [56], AS°ciay becomes even more positive. This indicated that the entropy increase due to the clay component in the reaction was more than that experimentally obtained for the overall reaction. The term AS°ciay actually represents in this case (SSl - SMg )ciay. According to the authors, the increase in entropy points to a more orderly structure with Mg ions on the surface than that with Sr ions. This could be attributed to the greater proximity of the magnesium ions to the surface. This, according to Mekhamer and Assaad, restricts their freedom in such a way as to decrease the number of arrangements which Mg ions may assume causing entropy loss for Mg -vermiculite and entropy gain for Sr - vermiculite. The c- spacings obtained by XRD also supported this. The c spacing of Mg-PVA vermiculite has the same value as for Mg-vermiculite without PVA. By increasing the Sr ion concentration in solution, the c -spacing becomes larger for the treated PVA -vermiculite than the untreated. Krishna et al [49,57] carried out batch experiments on Cr(VI) species sorption from aqueous solutions by montmorillonite clays modified with HDTMA. Under the conditions of the experiment, the amount of HDTMA sorbed was equivalent to the cec
112
B.S.JaiPrakash
of the clay (0.8 mmol g"1). The amount of Cr(VI) species sorbed by the HDTMA in the clay interlayer was found to be maximum at pH « 1. The attack of protons on the structure of the clay was found to be diminished possibly because of hydrogen bonding with oxyanion of chromate bonded to the surfactant. At pH a 1, the species present are HCr2O7" ions. The HCr2C>7~ anion displaced the surfactant counter ion from the exchange sites on the clays forming clay - (HDTMA) - HCr2O7. The exchange sites on the clays are balanced by the protons present in the medium. Mont - HDTMA+ + H+ - HCr2O7"_-> Mont - H+ + HDTMA+ - HCr2O7". Beyond pH = 1 and up to pH = 6, the species present are the divalent anions of Cr2O72" with small quantities of HCrO4" [58]. For the same reason, the amount of the Cr(VI) species sorbed between pH 2 and 6 is half the quantity adsorbed at pH<2. Mont - HDTMA+ + V2 [ H2Cr207] - Mont - H+ + (HDTMA)+ - !4 (Cr2O7)2' Table 3 - DKR Parameters -Adsorbent Capacity (Cm), Sorption Energy Constant (P) for Sorption of HDTMA on Montmorillonite and Cr(VI) Anionic species by Modified Montmorillonite Clay. Reprinted by kind permission of [J. Coll. Interface Sci. (220.230.2000)1 DKR constants Cn.molkg 1 Pmol 2 J 2 R correlation factor Sorption energy E, kJ mol"1
sorption of HDTMA on clay 345 -1.62xlO'8 0.986 5.55
sorption of Cr(VI) species by modified clay 340 -1.026x10"" 0.996 7.00
The Dubinin - Kaganer - Radushkevich (DKR) isotherm was assessed for the sorption of HDTMA on the bare clay surface and Cr (VI) species on the HDTMA modified clay surface. The DKR parameters - adsorbent capacity (Cm) and the sorption energy constant (p) for the sorption are given in Table 3. The sorption energy values (E) are found to be 5.55 kJmol"1 for HDTMA sorption on montmorillonite and 7.0 kJ mol"1 for Cr (VI) sorption by modified montmorillonite clay. These values are of the order expected of an ion exchange mechanism [59]. The adsorption capacity Cm at the DKR region is calculated to be 345 mmol kg"1 for the sorption of HDTMA and 340 mmol kg"1 for the sorption of Cr (VI) species. These values are less than the adsorption capacity observed at the Langmuir region. The DKR equation apparently is obeyed at a lower concentration range of 2.5 to 8.5 x 10"3 M for HDTMA and 0.5 to 2.5 x 10"3 M for Cr (VI) sorption respectively. The thermodynamic parameters for concentrations higher than this are shown in Table 4. The AG° value for the surfactant sorption below the CMC value of the HDTMA (<0.9mM) is lower than that obtained for above CMC. The change of enthalpy (AH0) values for the sorption of surfactant is negative showing that the sorption is exothermic in nature. The magnitude shows a stronger adsorbent - adsorbate interaction above the CMC value. The higher value of change in enthalpy obtained here compared to that on the surface of the clinoptilolite by Sullivan et al [43] shows that the dominant effect was to transfer the species from the water phase to diffuse into the interlayer. Washing did not remove the surfactant molecule from the interlamellar region.
Surface Thermodynamics of Clays
113
Similarly, the change in enthalpy values for Cr (VI) species sorption by modified clay shows that the anionic species are bound sufficiently strongly on the surface of the HDTMA - clay. The entropy changes, however, were slightly positive (Table 4) in agreement with the report by earlier workers. HDTMA being a large molecule with a long flexible tail group brings about a variation in entropy changes upon adsorption that mainly depends on its CMC values and its ability to destructure the water molecules surrounding them and also present on the clay surface [43]. According to Krishna et al [57], below the CMC value, surfactant molecules are not arranged in an ordered manner in solution phase; hence there is a loss in entropy upon adsorption. Above the CMC value, the surfactant molecules are arranged in a more ordered manner by the formation of a micelle in the solution phase. Once the clay adsorbs the surfactant molecules, they get randomly distributed to satisfy the scattered exchange sites within the interlamellar region. There will thus be a gain in entropy. Krishna et al attributed the slightly positive entropy change, for the adsorption of Cr (VI) species on the modified clay, to the fixation of anionic species on to the exchange sites of the randomly distributed surfactant species. Table 4 - Thermodynamic Parameters for the sorption of HDTMA on Montmorillonite Above and Below cmc and Cr(VI) Species on Modified Montmorillonite. Reprinted by kind permission of [J. Coll. Interface Sci. (220, 230, 2000)].
-AG^mol" 1 -AH°Kjmol"' -AS0 JK-'mol"1
Sorption of HDTMA on montmorillonite Below cmc value 22.3 16.1 1.02
Sorption of HDTMA on montmorillonite above cmc value
Sorption of Cr(VI) species bymodified clay
68.2 49.9 -3.08
39.2 19.1 -1.65
Juang et al [60] have studied the thermodynamics of sorption of phenol, m nitro phenol and o - cresol from water on to montmorillonite modified with HDTMA. the sorption capacity decreased in the order phenol > o - cresol > m - nitro phenol. The Langmuir, dual mode sorption and Redlich - Peterson models were tested to fit the sorption isotherms of single solute systems, whereas the Langmuir competitive model was used to describe bisolute sorption equilibria. Thermodynamic parameters (AH0 and AS0) and the mean free energy (E) for the sorption were determined from the temperature dependence of the distribution constant and the DKR equation, respectively. The value of sorption free energy obtained was indicative of ion exchange. The AH0 values for the sorption of phenol were in the range -3.6 to - 6.7 kj mol"1 and AG values were in the range - 9.7 to -12.8 kJ mol"1. The adsorption showed a gain in entropy. Removal of Cu2+ and Zn2+ from aqueous solutions by sorption on the montmorillonite modified with sodium dodecyl sulphate (SOS) was investigated by Lin and Juang [61]. The DKR parameters were determined. The sorption energy (E) values were -12.6 kj mol"1 for Zn2+ sorption and - 13.8 kJ mol"1 for Cu2+ sorption on the modified clay. This lies in the range of 8 - 1 6 kJ mol"1, indicative of ion-exchange
114
B.S. Jai Prakash
reaction. The sorption capacity in the DKR equation is found to be 31.9 mmol kg"1 for Zn2+ and 83.0 mmol kg*1 for Cu2+ for the clay having a cec of 37.2 mmol / lOOg. The values of AHC and AS0 are 7.39 kJ mol"1 and 6.39 J mol"1 K"1 respectively for Zn2+ and 7.05 kJ mol"1 and 9.09 J mol"1 K"1 respectively for Cu2+. A slightly positive entropy change was attributed by the authors to the fixation of ions on the exchangeable sites of the randomly distributed surfactant species. Doula et al [62] studied the thermodynamics of copper adsorption - desorption by Ca - kaolinite. Both the Freundlich and Gouy - Chapman models were found to describe the adsorption successfully. The adsorption enthalpy ranged from -2.71 to 4.78 kcal mol"1 indicating a physical process. The Cu2+ desorption was endothermic, since an increase in desorption was observed with increasing temperature. It was, however, observed by the authors that the calculated AH0 values include the enthalpies of hydration, mixing and exchange. Because of uncertainties about the energy of hydration of the adsorbed -desorbed ions, most of these terms cannot be calculated with a high degree of accuracy. Similar observations have been made on K -Ca and Mg - Ca exchange on the surface of kaolinite soil clay by Udo [63] and in calcium-bentonite clay byDoulaetal[64]. A study has been made recently to adsorb lead ions on to surfactant immobilized interlayered species bonded to clays (SUS-Clay) by Mahadevaiah et al [67]. Chromate immobilized by HDTMA surfactant was found to easily adsorb lead ions. The magnitude of DKR sorption energy (E) of lead ions on SHCr- montmorillonite was found to be 51.5 kJ mol"1. This value is greater than the order expected of an ion exchange mechanism [59] showing that the adsorption may be due to precipitation. The adsorption capacity Cm at the DKR region is calculated to be 119 mmol, kg"1 for the adsorption of lead ions which is less than the adsorption observed at the Langmuir region [67] 3 - Conclusions Probing the surface of a solid by measuring contact angle at the solid-liquid-air interface using the famous Young-Dupre equation which relates the surface tension to thermodynamic parameters has been extended to the study of surface of clays and related minerals. The surface tension components theory of van Oss - Good -Chaudhury has been successful in interpreting the contact angles and distinguishing the Lifshitz-van der Waals ) and the acid-base forces (y+ and y") on the surface. The average value of LW y , y" and y" for clay surfaces in mJ m"2 are, respectively 43, 1.6 and 36.9. The higher value of y indicates the predominance of the oxygen of the surface hydroxyl group in donating a pair of electrons. The magnitude of y+ is also indicative of the high tendency of the surface to hold through van der Waals forces. The sum total of the non covalent interactions between surfaces of particles immersed in water are designated as interfacial (IF) interactions. The interfacial free energy between clay particles immersed in water (AG c w c = AGLW + AGAB) are mostly positive with an average value of +10.6 mJ m"2 indicating hydrophilic character of clays. The value changes to negative when the clay surfaces are modified with organic molecules. Negative values of interfacial free energy is a quantitative measure of the degree of hydrophobicity of the clay. Thus organo clays are hydrophobic indicating their tendency to flocculate. A value of y" = 27.9 mJ m"2 is fixed as the boundary between hydrophobic (<27.9 mJ m"2) and hydrophilic (>27.9 mJ m"2). The contribution of electrostatic interaction arising due to charge, AGEL could be determined by electrical
Surface Thermodynamics of Clays
115
methods. Thus AGtotal = AGLW + A G ^ + AGEL. The classical DLVO analysis ( a plot of AGLW + AGEL as a function of particle distance in nm at different ionic strengths) fails to predict the stability of clay suspension in polar solvents like water. However, the XDLVO analysis (a plot of AGtotal against the distance between the particles) gives a clear picture of the stability of clay suspensions at different ionic concentrations. Ignoring the AB interactions between the clay particles suspended in water would lead to incorrect conclusions. For highly hydrophobic clays such as talc and pyrophylite, which do not form films suitable for contact angle measurements, thin layer wicking method is recommended. These materials, having zero layer charge, exhibit hydrophobic character because of their weak tendency for electron donicity. A decrease in the value of y~ was noticed on grinding the clay minerals possibly because of the increase in the electron acceptor sites (y+) which neutralizes the adjacent y" sites resulting in a material with high AG cwc (hydrophilic) to change over to a low AG J.FWC (hydrophobic). The average values of the immersion enthalpies and the interfacial surface enthalpies per m2 of the surface, in contact with water, were similar for kaolinite and illite. With swelling clays, however, a high value of immersion enthalpy has been observed. The extra heat liberated is attributed to the interaction between exchange cations and interfacial water during swelling. It is realized that the introduction of another term called swelling enthalpy may be in order to interpret the data obtained by enthalpic measurements in the case of swelling clays. The continuous changing of sorption topology is attributed to the driving force derived from the chemical potential of the adsorbate molecules. Combined measurementrs of AadsG j ^ " * and AadsH j ^ " * for different probe liquids (apolar, monopolar and bipolar) would thus allow surface entropy phenomena to be understood and would help to evolve a proper theory of surface thermodynamics of clays and related minerals. Accordingly, an attempt has been made in the latter part of the chapter to compile the thermodynamic data obtained in the recent studies on adsorption of organic cations, used as tracers of surface heterogeneity, by swelling and non-swelling clays. The main purpose is to present and try to relate the thermodynamic data derived by different techniques such as surface tension, immersion enthalpy and sorption. One might conclude, in the end, that it is just a recognition of several functioning variables operative in a dynamic clay system. Acknowledgements A part of the work compiled in this chapter was carried out under a project grant received from the Department of Atomic Energy, Government of India. The author expresses his grateful thanks to the Department. The author also wishes to thank the management of Vokkaligara Sangha for the encouragement. 4 - References [1] J. Lloyd, J. Amer. Pharm. Assn., 5 (1916) 381. [2] J.M. Douillard, J. Colloid Interface Sci., 182 (1996) 308. [3] R.F. Giese and CJ. Van Oss, Colloid and Interface Properties of Clays and Related Minerals, Marcel Dekker Inc, New York, vol. 105: Surfactant Science Series (2002). [4] R.A. Swalin, Thermodynmics of Solids, John Wiley & Sons Inc. Singapore 221, (1991). [5] R.F. Giese, W. Wu and CJ. van Oss, J. Disp. Science and Tech., 17 (1996) 527.
116
B.S. Jai Prakash
[6] J. M. Douillard, J. Colloid and Interface Sci., 188 (1997) 511 [7] M.K.Chaudhury, Short Range and Long Range Forces in Colloidal and Macroscopic Systems, Ph.D. Thesis, SUNY at Buffalo, (1984) 215pp. [8] Dupre A. Theorie Mecanique de la Chaleur: Gauthier Villars, Paris (1869) 484pp [9] C.J. van Oss, R.J. Good , M.K. Chaudhury, Langmuir, 4 (1988) 884. [10]F.M. Fowkes, Chemistry and Physics of Interfaces: Predicting Attractive Forces at Interfaces. Analogy to Solubility Parameter, Ed. S. Ross, (1965) 153. [11] C.J. van Oss, R.J. Good , M.K. Chaudhury, Ad. Colloid Interface Sci., 28 (1987a) 35. [12] C.J. van Oss , R.J. Good , M.K. Chaudhury, Sep.Sci. Technol., 22 (1987b) 1. [13] C.J. van Oss , R.J. Good , M.K. Chaudhury, Colloids Surfaces, 23 (1987c) 369. [14] C.J. van Oss, R.F. Giese and W. Wu, J. Adhesion Sci. Technol., 63 (1997) 71. [15] C.J. van Oss and R.F. Giese, J. Disp. Sci. Technol., 24 (2003) 363 [16] C.J. van Oss, R.F. Giese and P.M. Costanzo, Clays Clay Miner., 38 (1990a) [17] R.F. Giese, P.M. Costanzo and C.J. van Oss, Phys. Chem. Miner., 17(1991) 611. [18] E.W. Washburn, Phys. Rev., 17 (1921) 273. [19] J.P. Rupert, W.P. Granquist and T.J. Pinnavaia, Chemistry of Clays Clay Miner. monograph no. 6: Catalytic Properties of Clay Minerals, Ed. A.C.D. Newman, Longman Scientific and Technical, England, (1987). [20] A.B.D. Casie, Discuss. Faraday Soc, 3 (1948) 11. [21] Medout - Marere, H. Belarbi, P. Thomas, F. Morato, J.C. Giuntini and J.M. Douillard, J. Colloid and Interface Sci., 202 (1998)139. [22] J.G. Jura, and W.D. Harkins, J.Am. Chem. Soc, 66, (1944) 511. [23] S. Partyka, S. Rouquerol and J. Roquerol, J. Colloid and Interface. Sci., 68 (1970) 30. [24] S. Y. Lee and S.J. Kim, Clays Clay Miner., 50 (2002) 435. [25] C.T. Cowan and D. White, Trans. Faraday Soc, 39 (1958) 691. [26] E.E. Vansant and G.Peeters, Clays Clay Miner., 26 (1978) 279. [27] G.L. Gaines and H.C.Thomas, J.Chem. Phys., 21(1953) 14. [28] S.J.Gregg and K.S.W.Singh Adsorption, Surface Area and Porosity, 2nd Edition, Academic Press, London(1962). [29] J.P. Hobson, J. Phys. Chem., 73 (1969) 2720. [30] M. Polanyi, Verb. Deutch Physik. Ges., 16 (1914) 1012. [31] J.M. Thomas and W.J. Thomas, Principle and Practice of Heterogeneous Catalysis, VCH Publishers. Inc. NY (1997). [32] M.G. Kaganer , Zhur Fiz. Khim., 33, (1959) 2202. [33] S. Lin and R.S. Juang, J. Hazardous Mater., 92 (2002) 315. [34] B.S. Krishna, D.S. R. Murthy, B.S. J. Prakash, J. Colloid and Interface Sci., 229 (2000) 230. [35] J.L. Huh, D.L. Song and Y.W. Jeon, Sep. Sci. Technol., 35 (2000) 243 [36] S.A. Boyd, S. Shaobai, J.F. Lee .and M.M. Mortland, Clays Clay Miner., 36, (1988) 125. [37] M.A.. Lawrence, R.K. Kukkadappu and Boyd S.A. Appl. Clay Sci., 13 (1998) 3. [38] J.F. Lee , J.R. Crum, S.A. Boyd, Environ. Sci. Technol., 23 (1989) 1365. [39] R.E. Grim, W.H. Allaway, F.L. Cuthbert, J.Am.Chem. Soc, 30 (1947) 137. [40] Z.Z. Zhang, D.L. Sparks and N.C. Scrivner, Environ. Sci. Technol., 27 (1993) 1625. [41] E.F. Vansant and B. Uytterhoeven, Clays Clay Miner., 20 (1972) 47.
Surface Thermodynamics of Clays
117
[42] B.K.G. Theng, D.J. Greenland and J.P Quirk, Clay Miner., 7 (1967) 1. [43] E.J.Sullivan, J.N.Carey and R.S. Bowman, J. Colloid and Interface Sci., 206 (1998) 369. [44] H. Tamura, K. Noriaki and F. Ryusaburo, Environ. Sci. & Tech., 30 (1996) 1198. [45] J.N. Israelachvili, Intermolecular and Surface Forces 2nd Ed, AP, San Diego, CA (1991). [46] J.Blandamor, P.M. Cullis, L.G. Soldi, J.B.F.N.Engbrets, P.A.Dooreman., A. Kasperska, K.C. Rao, and M.C.S. Subha, J. Chem. Soc. Faraday Trans., 91 (1995) 1229. [47] J.T. Bashford. and E.M. Woolley, J.Phys. Chem., 89 (1985) 3173. [48] Y.Lee and S.J.Kim, Clays Clay Miner., 50 (2002) 435. [49] B.S. Krishna, D.S.R. Murty and B.S. Jai Prakash, Bull. Mater. Sci., 21 (1998) 355. [50] B.S. Krishna, D.S.R. Murty and B.S. Jai Prakash, Clay Res., 18 (1999) 1. [51] R. Narine and R.D. Guy, Soil Sci., 133 (1982) 356. [52] W.K. Mekhamer and F.F. Assaad, Thermochimica Acta, 334 (1999) 33. [53] G.L. Gaines and H.C.Thomas, J.Chem. Phys., 21 (1953) 714. [54] G.L. Gaines and H. Thomas, J. Chem. Phys., 23 (1955) 2322. [55] R.G. Gaust, R. van Bladel, R.B.Deshpande, Soil Sci, Soc. Amer. Proc, 33 (1969) 661. [56] R.A. Home, Water and Aqueous Solutions, Structure, Thermodynamics and Transport, Wiley Interscience, NY, (1972). [57] B.S. Krishna, D.S.R. Murty and B.S. Jai Prakash, J. Colloid and Interface Sci., 229 (2000) [58] N.N. Greenwood and A. Earnshaw Chemistry of the Elements, Pergamon, NY, (1988) [59] F. Helfferich F, Ion Exchange McGraw - Hill, NY, (1962) [60] R.S. Juang, S.H. Lin and K.H. Tsao, J. Colloid and Interface Sci., 254 (2002) 234. [61] S.H. Lin and R.S. Juang [52], J. of Hazardous Mater., 92( 2002) 315. [62] M. Doula, A. Ioannou and A. Dimirkou, Adsorption, 6 (2000) 325. [63] E.J .Udo, Soil Sci. Soc. Amer. Proc, 42 (1978) 556. [64] M. Doula, A. Ioannou and A. Dimirkou, Commun. Soil Sci. Plant Anal., 26 (1995) 1535.
[65] R.J. Good, SCI Monographs, 25(1967) 328. [66] F.M. Fowkes, Acid-base interactions in Polymer Adhesion:Physicochemical Aspects of Polymer Surfaces, Ed. K.L. Mittal,, Plenum Press, New York (1983). [67] N. Mahadevaiah, B.S.Krishna and B.S.Jai Prakash (unpublished results).
PHENOMENOLOGY OF WATER ADSORPTION AT CLAY SURFACES JEAN MARC DOUILLARD* and FABRICE SALLES University of Sciences, Montpellier, FRANCE (L. A.M.M.I., CCO15, Universite Montpellier 2. Sciences et Techniques du Languedoc. PL Eugene Bataillon, Montpellier Cedex 05, FRANCE) * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych andK.G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Phenomenology of Water Adsorption at Clay Surfaces
119
1 - Introduction Clays are some of the most important minerals in many diverse uses. Indeed, clays and clay minerals are important in myths, art, medicine, agriculture, construction, engineering, process industries, and environmental applications. Some of the more important applications include ceramics, paper, paint, plastics, drilling fluids, foundry bondants, chemical carriers, liquid barriers, decolorization, and catalysis. In many of these cases, clays are in contact with water. Research activities and improved processing and mining produce continually innovative clays of higher purity, or modified surface chemistry. However, in basic, clays can be considered as part of a few great crystallographic families. In the practice, clays are powders produced from the crushing of rocks. Then, the particle size, shape, and distribution are important physical properties, which are intimately related to the applications of the clay minerals. Some other important properties are surface chemistry, surface area, and surface charge, which are dependent on the former properties, but can be studied as fundamental parameters. The properties of any material near surfaces or interfaces are different than the properties of the same material in the bulk. This is expected because the surface atoms coordination environment is different when it is in the bulk. This special character of surfaces has some thermodynamic implications, especially in terms of interaction with molecules of other substances in close contact. Clays are lamellar aluminosilicates showing a wide variety of chemical effects, when put in contact with water; for instance dissolution, hydroxylation, fracture, delamination, and electrification. Many physicochemical phenomena limited to surfaces are also observed: adsorption, swelling (adsorption can be considered as a step in the swelling process), ionization, ionic exchange and surface acidity. This complex landscape of clay-water interactions is due to a general structure giving many possibilities. Basically, clays are composed of sandwich layers of aluminum/magnesium octahedrons (see Figure 1) or silica tetrahedrons (see Figure 2). Different combinations of sandwiches are possible. The composition of the layers is very variable. The cations present in the layers can be of multiple types: Al, Si, Mg, Fe, etc.
Figure 1 - Schematic representation of a Kaolinite where the octahedral coordination oftheAl atom is shown This chapter focuses on the case of adsorption of water, i.e. the interaction of low energy between surface atoms of solids and vapor molecules, which modifies surface properties but does not modify bulk properties, and which is reversible. Adsorption occurs when there is a certain activity (i.e. a significant chemical potential)
120
J.M. Douillard and F. Salles
of the vapor, when it is close to the saturation pressure corresponding to the chosen temperature. Water can adsorb at ambient temperature and pressure. In these conditions, the vapor pressure of water is about 3.2 kPa.
Figure 2 - Schematic representation of a Kaolinite where the tetrahedral coordination of the Si atom is shown If the scope is thus limited to surfaces, the microscopic interaction with water deals with the nature of the bulk crystallographic structure of the clay, with the presence and nature of compensation cations, and with the location of active atoms, either on edges or in the plane of sheets of clay. The interaction between water and a solid surface can be viewed as a chemical reaction characterized by certain energy, which is the product of an extensive term and an intensive term. In surface thermodynamics, the extensive term is the surface area, and the intensive term is then the energy per unit of surface area. The understanding of the interaction deals mainly with the analysis of the energy and free energy changes per unit area associated with the contact of water and clay minerals. Unfortunately, the thermodynamic definitions are dependent on the volume of the spatial zone defined as the "surface" when analyzing the problem. Despite of this, the key macroscopic parameter allowing the understanding of this phenomenon is the surface tension of the solid, i.e. a derivative of the Gibbs free energy per unit surface area of the solid. 2 - Nomenclature In very general terms [1-3], "clay" refers to a fine-grain mineral, which is plastic at certain water contents and will harden when dried. This is a broad definition: clay materials usually contain phyllosilicates but can contain other materials, which differ, chemically or crystallographically, such as quartz. From a more practical viewpoint, clay minerals are aluminosilicates that predominate in clay fractions of earth materials at intermediate stages of weathering. From the mineralogists' basic viewpoint [4-11], clay minerals are sandwiches of tetrahedral and octahedral sheet structures. We will consider here that they belong to three major groups: Kaolinite (containing Lizardites), Montmorillonite/Smectite (which contains "clay-mica" products such as Illite) and Chlorite. The differences between the three groups are due to the chemical formula and to the stacking of silicate sheets ((Si2O5)2") denoted s (see Figure 3) bonded to gibbsite layers (A12(OH)4) denoted g (see Figure 4) or to brucite layers (Mg2(OH)4); denoted b (see Figure 5) [12].
Phenomenology of Water Adsorption at Clay Surfaces
Figure 3 - Schematic view from the top of a Siloxane layer.
Figure 4 - Schematic representation of the side of a Gibbsite layer.
Figure 5 - Schematic representation of a side of a Brucite layer.
121
122
J.M. DouillardandF. Salles
Brucite and Gibbsite are octahedrally coordinated and silicate sheets are tetrahedrally coordinated. A stacking sequence corresponds to a group. Kaolinites are sg (see Figure 6) and Lizardites are s-b. Montmorillonites are s-g-s (see Figure 7). Illites and Mica-clays are also s-g-s (but with more water inside the structure than Montmorillonite). Interlayer ions can be present in the structure, modifying behaviors and interlayer distances.
Figure 6 - Schematic representation of a kaolinite.
Figure 7 - Schematic representation of a montmorillonite Chlorites are formed of ((s-b-s) b) sequences (see Figure 8) [13]. Talc is a very special clay, with a stacking sequence: s-b-s and can be placed in different groups (see Figure 9). The classification of clays can also use layer types. These one distinguished by the number of tetrahedral (denoted T) and octahedral (denoted O) sheets combined. For example, the 2:1 layer type has two tetrahedral sheets that sandwich an octahedral sheet. This can be also written T-O-T. This classification gives less chemical details than the preceding one.
Phenomenology of Water Adsorption at Clay Surfaces
123
Figure 8 - Schematic representation of a chlorite.
Figure 9 - Schematic representation of a talc. 3 - Surface Sites All these materials present differences in their chemical structure. The paramount question, from the viewpoint of adsorption, is often considered to be the nature of active sites at the surface. In many theories [14-15], adsorption is described in terms of a set of complex formation reactions between the fluid molecules and surface functional groups. The surface and the density of functional groups control the adsorption capacity. The materials can be classified in terms of the nature of the surface functional groups that they possess. Hydrous oxide minerals possess proton bearing or proton accepting surface functional groups [16]. Furthermore, aluminosilicate minerals are divided into those that possess a permanent structural charge in the layer and those that do not. Substitution of divalent cations for trivalent cations in the octahedral layer and trivalent cations for Si(IV) in the tetrahedral layers gives rise to the permanent structural charge on the composite layer, that is compensated by the complexing of mono or divalent cations located between the sheets near the ditrigonal cavities. (The layer charge can be calculated from the total element analysis. For the calculation, it is often considered that only Al atoms substitute for Si in the tetrahedral layer, even
124
J.M. Douillard and F. Salles
though other elements such as Fe, Mn or Mg are present. They are generally considered only as structural elements of the clay [17]). Thus, minerals with permanent structural charge have ion-bearing exchange sites in addition to proton bearing or accepting surface functional groups. Therefore the interface between clay surfaces and water can be viewed in certain conditions as an interface between a charged surface and water, namely an electrical double layer [18-19]. The case of phyllosilicates is even more complex. In addition to surface hydroxyl groups, these minerals present rings of siloxane groups, which occur on the basal planes, and interlayer regions that dominate the surface area of these minerals. In many cases, these groups are not hydroxylated because the coordination environments of these bridging oxygens are satisfied by the two Si(IV) ions with which they are coordinated. These oxygens are obviously strong electron donors and can interact with water, which is itself particularly reactive [20] and can interact through its hydrogen atoms. The importance of the siloxane rings as surface functional groups is enhanced by the magnitude of the permanent charge of the crystal lattice of clay minerals, which is due to isomorphic cationic substitutions. Moreover, linkage of tetrahedra to form the tetrahedral layer gives rise to hexagonal cavities surrounded by siloxane groups. Joining of the tetrahedral and octahedral layers causes distortion of these cavities from hexagonal to ditrigonal symmetry [21]. Then a clay can be characterized by a large number of adsorption sites (oxygens, aluminols, magnesols, silanols, siloxane rings, cations, etc.), some of them being in very specific configurations. Some of adsorption sites (such as aluminol) are both proton donor and proton acceptor. Some of them are only proton donor or proton acceptor. When the structural charge and water quantity are important, the interface can be considered as an electrical double layer. The detailed characteristics of this layer, such as the thickness are given by solutions of a Poisson-Boltzmann equation. More details can be found in specialist books on this subject [18, 22]. In conclusion, the global behavior of the clay-water interface can be viewed as a combination between two processes: physical adsorption (onto n surface site types) and progressive building of the electric double-layer. We will now focus on the physical adsorption step only, because the "adsorbed phase" is strictly speaking a mixture of gas phase and solid phase, when the electric layer is a solid-liquid interface. 4 - Particle size, shape, and surface area In the special case of solids, a surface is the exterior boundary of the solid phase in a vacuum or a gaseous environment. An interface is the boundary between two condensed phases e.g. a solid-liquid interface or a solid-solid grain boundary. However, the term "surface" can be taken in a generic sense, the differences between surfaces and interfaces being not so important. In practice, clays are powders. The surface area of a divided solid increases relative to volume as grain size decreases and also with increasing porosity either intern or between grains. The area and volume of a grain are related to some characteristic length (such as a diameter d) through shape factors (such as the number n). Then, the surface area A and the volume V of the grain [which can be determined by some optical measurements (for instance [23])] are related. It is possible to state some specific surface areas [24], which can be defined per unit volume, or mass (m), or mole. The most convenient specific surface area is the surface area per unit mass, because it can be determined by some standardized experimental methods.
Phenomenology of Water Adsorption at Clay Surfaces
As=
125
(Eql)
miv
The most common is the BET method [25], named after Brunauer, Emmett, and Teller, who derived it. It is an analysis of the adsorbed quantities of a certain gas retained by a known mass of the solid, at certain partial pressures of the gas. The key parameter of a BET determination is the average area occupied by a molecule of adsorbate in the completed monolayer, a,,,. The value used in the calculation is set, and depends on different assumptions made by the various schools of thought. The main point is to give the value of am used when reporting a specific surface area. An average enthalpy of adsorption of the first layer can be computed from the BET determination. It is expressed by a parameter c. Clays either natural or synthetic have generally high specific surface area. This is due to their special shape of sheets. In the literature, specific surface areas of materials range from
Y=
(Eq2)
fel,P,n
The subscripts T, P, and n specify that pressure temperature and composition (in the surface phase) remain constant during the process. The total free energy of the surface can then be written G^^njUj+Ay i
(Eq. 3)
Depending on the different processes studied, it is possible to distinguish between molecules of the solid and of the fluid and between tensions of different crystal faces, even though this distinction has a poor meaning from the experimental viewpoint. For instance
G s = X n^ olk V-° lid + X n J u i V f M + Ay i
(Eq. 4)
j
or G s = £ n j m +Abasalybasal+Aedgeyedge i
( E q 5)
126
J.M. Douillard and F. Salles
The surface tension of solids has to be defined more precisely. Indeed, when considering a solid, the stretching of the surface and the forming of a new surface are different processes [28]. The surface tension (ys) is the derivative of the free energy associated with the formation of a new surface. There is an equivalent term associated to the stretching of a surface, which is called the surface stress T. Generally whatever, adsorption phenomena are studied in absence of strains applied to the solid. Therefore, it is the surface tension, which is relevant. Generally, the superscript ° is added to indicate that the surface tension is defined in absence of strain (y2). 5.2 - Adsorption Adsorption onto a solid surface can be viewed as a chemical reaction between gas molecules and some solid molecules, forming a new surface: n baresolid
+n?as ^ ( r i j +n] )surface
(Eq. 6)
The energy of the reactants can be written as follows:
E r =-PV r +TS r +£n solid i4 olid +]F]n? a! Vf as +y°A solid i
(Eq. 7)
j
Similarly, the energy of the surface phase is given by:
E s ^ P V ^ ^ T S ^ ^ E n r " ^ i d + E n f °^>f +7sv A surface (Eq. 8) i
j
The energy change for the reaction corresponding to adsorption at constant temperature and pressure of a pure gas is: AE^-PCVSU-^+TCS^-S^^ni^fO-^^+^nX^-^HySLAsu-YsA 1 3 5 0 i j
( Ec l- 9 )
with the notations r for reactants, bso for the bare solid, su for the surface phase, ag for the adsorbed gas and g for the gas. In many cases, the thermodynamic surface area does not vary during the process and can be substituted by the specific surface area. The heat q corresponding to the process is linked to the energy change by: q = AE + P ( V s u - V r )
(Eq. 10)
Therefore, a heat change (dq) at constant pressure can be expressed as: dq = T d S a d s + ^ A n j i o l i d dn j + RT(A In P / P°)dn j + (y S L - y ° )dA s i
(Eq. 11)
Phenomenology of Water Adsorption at Clay Surfaces
127
where P° is the saturation pressure of the vapor chosen and dSads the entropy change due to adsorption. 5.3 - Surface excesses: the problem of dividing surfaces The other thermodynamic properties of interfaces need an analysis of the volume considered as a "surface phase" i.e. the spatial zone where the thermodynamic properties are different from the equivalent properties taken in the adjacent bulk phases. It is clear, starting alternatively from experiments [29] or from both crystallographic and chemical bonds theories [30-31], that an interface "composed" by a solid and water molecules can be very complex. Its analysis has to take into account several molecular layers inside the solid and several layers inside the liquid [32]. It is therefore clear that the interfacial thermodynamics must be in agreement with the thickness of the claywater interface. However, in reality, it is very difficult to estimate this thickness. From the thermodynamical viewpoint, it is necessary to consider that the surface phase is taken inside two Gibbs dividing surfaces (GDS) [33-39] This analysis, which can be called the "multilayer model", has been applied to interfacial electrochemistry and to liquidvapor interfaces [40-42], using different estimations of the thickness. In such a case, the thermodynamical analysis is in agreement with experimental results coming from very different techniques such as ellipsometry, X-Ray reflectivity, STEM or classical adsorption methods. Equations are identical to the Gibbs model, but in such an analysis, adsorbed quantities cannot be negative. Surface excesses (i.e. the numbers of molecules adsorbed) are actual concentrations in the surface considered as a phase. The molar properties (such as the volumes) of the species have the same physical meanings as in the classical bulk thermodynamics. In this multilayer model a choice remains. This choice is the surface area value. In the case of solid-fluid surfaces, it is convenient to take the specific surface area As, i.e. a value per weight corresponding to the solid, because this value is constant in the majority of the processes. Swelling modifies the specific surface area, and this special case will be discussed below. Therefore, dividing the surface Gibbs free energy by A s gives
§ s = Z r i^i + Y
(Eq-12)
i where, gs is the Gibbs free energy per unit surface area and Tj the surface excess of the species i. We obtain at equilibrium a relationship, which is called the "Gibbs equation":
Y{=-\^-\
(Eq.13)
Considering the gas phase, the dujterm is a function of the partial pressure. Therefore, adsorption corresponds to the introduction of a certain number of molecules in the surface phase, accompanied by a decrease of the surface tension. The corresponding thermal effect can be defined from the preceding equations by using a Gibbs-Helmoltz relationship:
128
J.M. Douillard and F. Salles
AH=AG-T^P
(Eq. 14)
The equation (13) gives a method of determination of the quantities of molecules in a fluid-fluid surface using surface tension measurements. However, in the case of a solid-fluid interface, this quantity is directly determined by the modification of weight of the solid during the experiment or by the modification of concentrations in the fluid bulk phase. For solid-solid interfaces, it is an integral energy, such as adhesion energy, which is determined. Furthermore, the equation (13) requires that a decrease in surface free energy implying a positive adsorption at the interface. The surface tension of the solid yS, which is strongly dependent on the bulk energy, can be considered as the "attractive potential" of the adsorption phenomena. A large surface tension of the solid implies large adsorption energy of water molecules, so a low value of the tension of the resulting surface phase. 5.4 - Adsorption Isotherms Interpretation of adsorption is rigorously the analysis of energy and free energy changes associated with the contact of water and solids. The interaction between water and clays, and more generally with solids can be viewed as a study of the changes of adsorbed water quantities versus the partial pressure of water. In the literature adsorbed quantities are expressed as surface excess F (in moles or milligrams of gas per gram of solid, or in adsorbed volume (cm3 S.T.P.) per gram of solid) versus the partial pressure. These curves are called adsorption isotherms even though they do not rigorously correspond to the thermodynamic definition of the chemical potential, which must use a Log function. In literature, adsorption isotherms are often considered as "theoretical equations", more than experimental data. Then, mostly used relations are the Langmuir isotherm and the Freundlich isotherm [14-15]. However, it is well known that data correspond to processes more complex than the concepts underlying these simple relationships. The rigorous procedure has been widely exposed [43], and can be summarized by the following phenomenological concepts. - The adsorption isotherm belongs to a certain type of standard isotherms (Figure 10), numbered as six by the I.U.P.A.C [25]. - The left hand side of the isotherm corresponds to the direct interaction of the adsorbing sites of the solid with the vapor molecules. - The right hand side of the isotherm corresponds to the creation of a quasiliquid phase on the solid surface. This creation is strongly dependant on the similarities between the liquid network and the solid network downer. - Type I is a type II isotherm, determined for a microporous solid; types II and IV corresponds to "high surface energy solids"; type III and V corresponds to "low surface energy solids"; types IV and V corresponds to solids with a macroporosity. - If the solid possesses some mesoporosity, adsorption cannot be reversible because it is necessary to give a supplementary energy to extract the liquid from pores. This energy is due to the Kelvin effect. Some pores in clays have diameters of Angstroms to tens of Angstroms; thus the pressure of water inside these materials can be greater than outside. This phenomenon creates a hysteresis when the water vapor is desorbed from the solid surface (isotherms of types IV and V). The hysteresis loop can be used to analyze the mesoporosity of the solid, by using equations corresponding to varying expected shapes of pores [14,24-25]. This phenomenon depends more on the topology of the solid surface than on its energy. Generally, it is observed with clays of
Phenomenology of Water Adsorption at Clay Surfaces
129
type II and type IV isotherms, allowing the specific surface area to be determined correctly by BET method (Figure 11 shows a typical experimental isotherm). However, generally, smectites do not give correct Type II isotherms (see the corresponding following Paragraph).
Figure 10 - The six types of adsorption isotherms, I to VI, following classification.
IUPAC
Figure 11 - Adsorption isotherm of water on a Na-Kaolinite at ambient temperature. This isotherm is typical of the majority of isotherms of water on non-swelling clays. It is a type II isotherm.
130
J.M. Douillard and F. Salles
5.5. - Adsorption Enthalpies The quantity of heat corresponding to the adsorption phenomenon can be defined from alternative procedures. Using "immersion calorimeters", it is possible to obtain an integral enthalpy of adsorption, which corresponds to the transition between the pure solid under vacuum and the adsorbed phase. The corresponding equation is [44-45]:
V
/
Some differential enthalpies of adsorption can be measured using calorimeters. The exact definition depends on the type of experimental arrangement employed. In an isothermal volume calorimeter, the reversible adsorption of an amount of gas, dn, is accompanied by an exchange of heat dq. The differential heat (dq/dn) is related to a differential enthalpy, conventionally called the isosteric enthalpy of adsorption AaHThe usual method of calculating this term is by experimental isotherms, plotting ln(P/P°) for a given r as a function of 1/T. Rigorously speaking, this term is clear when extrapolated at zero coverage, e.g. at nil pressure of the adsorbed gas. It corresponds to the association of molecules of water with bare solid surface. When taken at finite value of the partial pressure, it corresponds to an unclear derivative of the enthalpy with some terms such as the chemical potential of adsorbing sites taken as constant. The exact definitions of the differential enthalpies of adsorption, with respect to the number of molecules or with respect to the temperature, are very complex. They depend on many assumptions such as the solid considered as inert, or as the volume of the adsorbent being negligible in comparison with that of the gas, etc. More details can be found in the following references [14-15,22,24,27]. Basically, it is possible to consider that the extrapolation of differential heats at zero coverage is equivalent to the integral enthalpy of adsorption but is expressed in joule per mole. Obviously, the integral enthalpy of adsorption is a key parameter, because it can be analyzed as the energy corresponding to a reaction between two pure species: water and clay. Clay is assumed as a "pure substance". It is possible if clay is considered homogeneous at the scale of a mole. The adsorption energy can now be viewed as a combination of the energies of two pure species interacting [27,44-47] and can be used to determine the energy of the solid. 5.6 - Wettability Shapes of drops on solid surface or wetting behaviors of powder immersed in liquids are governed by a mechanical equilibrium known as the Young equation, related to the 2D picture of a drop on a plate surface:
y l v cose = y s v - y s l
(Eq. 16)
The interfacial tensions are denoted y. The subscripts s, 1, v correspond to the solid, liquid and vapor phase respectively. 0 is the "contact angle". The main practical difficulty when using this equation is the fact that it is generally impossible to measure solid-fluid surface tensions. It is however possible to overcome this problem. Firstly, it is necessary to consider that the actual solid sample,
Phenomenology of Water Adsorption at Clay Surfaces
131
which in case of clay is generally a powder, can be represented by a "mean" surface tension, averaged over the whole crystal and grains faces. The reference state is the solid surface tension taken in vacuum and in the absence of mechanical stress (y°). From a semi-theoretical analysis of immersion experiments [45], it is known that the range of magnitude of this tension for clays goes from 100 mN/m (Talcs) to 500 mN/m (Chlorites). When the clay is put in presence of a water vapor phase, few molecular layers are adsorbed. The tension of the formed interface, which is the relevant solid surface tension when observing drops of water on clays, is then denoted ysv. The surface tension drop nt (i.e. y§ -ysv), between this surface state and the preceding, can be found by integration of the adsorption isotherm. The values are from 100 mN/m to 300 mN/m [48] for the same type of solids. Therefore, the solid-vapor surface tension can be higher or lower than the water liquid-vapor surface tension (about 72 mN/m at 25°C). In the first case, water spreads. In the second case, drops are observed (typically in the case of talcs). It is convenient to characterize the solid by an empirical surface tension, which is called critical surface tension (yc) after Zisman [49]. It is the maximal surface tension of the different liquids, which wet the solid. This empirical term, which is obtained experimentally by observation of wetting by liquids of different surface tensions has been shown [50] to be linked to yg. At a primary level of approximation, and for clays, the following assumption can be used yc«0.5yj?
(Eq. 17)
The change of tension due to adsorption and the contact angle can be related by a simple equation of the following type: jvo avo COS0=-(l+b)--^-+2 p _ Y lv F i v
(Eq. 18)
where a and b, which take into account the adsorption effects, are dependant on the experimental case considered. In the reality, the physical links between a, b and y° (i.e. the attracting force) imply that the numerical form of this equation does not greatly vary. Therefore, it is possible to use in practice an empirical and numerical form of this equation such as the Neumann's equation of state [51], which is written
cose = - i + 2 p ^ r p ^ - ^ 2 VYlv
(i9)
with (3=0.0001247 (m2/mj)2. Indeed, this equation is a numerical approximation, which takes into account the different processes acting at a triple line of contact: adsorption of vapor and formation of the liquid-solid surface. Obviously, the greater the surface tension of the solid, the better the wettability. From this set of approximated equations, it is clear that it is possible to use the theoretical (see below) or semi-theoretical estimations of y£ in
132
J.M. Douillard and F. Salles
order to estimate yc and the contact angle or more generally wetting behaviors. The order of magnitude of the surface tensions given below is a rough indication of the wetting behavior of clays versus water. Moreover, using this route, it has been observed that the differences between behaviors of plates and powders are not important [52]. It can see below that some aspects of the wetting behavior of the solid versus water can also be linked to chemical hardness. 6 - Experimental methods Many experimental methods used in order to study adsorption onto surfaces, include ellipsometry [53], NMR [54], ESR [55] or X-Ray spectroscopy, neutron diffraction [56], etc. However, few of them are relevant when studying divided solids such as clays, and few of them are directly interpretable in terms of surface energy or adsorption energy. We will focus on some techniques which give directly an energy term or which have been already qualitatively already linked to adsorption energy. 6.1 - Adsorption isotherms The "adsorption isotherm" corresponds to quantities of vapor adsorbed by a known mass of the solid, versus the corresponding partial pressures of the vapor in equilibrium with the solid in a certain volume. Various methods can be used. For instance, the clay samples can be kept in dessicators over salt solutions of different r.h. until equilibrium has been maintained. The amount of adsorbed water is obtained by weighting. More sophisticated methods are currently used [57]. A balance of high precision maintained in a controlled atmosphere can be used. Volumetric methods where a known quantity of vapor is introduced in a known volume are often used. A comparison is made between the actual pressure and a computed pressure corresponding to the number of molecules introduced. The difference corresponds to the molecules adsorbed. 6.2 - Adsorption calorimetry Calorimetry is a measurement of heat. Different types of calorimeters can be used: adiabatic, isothermal and heat-flow calorimeters. Heat-flow calorimeters use thermocouples insuring thermal conduction between the calorimetric cell and its surrounding to measure the heat flow. This type of apparatus possesses high sensitivity, which is particularly suitable for interfacial calorimetry. Two different experiments can be performed. The first one corresponds to an "integral adsorption" and is called [45] "immersion calorimetry" or "wetting calorimetry". The second one corresponds to a "differential adsorption" and is called "adsorption calorimetry" [58]. In the first case (see Figure 12 a typical set used for these experiments), the process is the contact of two phases: a pure solid, dry or wetted and a pure liquid. In the second case, small known quantities of the adsorbing vapor are put in contact with the solid, and the process is iterated. In this case, the heat evolved is attributed to a certain number of moles of vapor, but to an unknown surface of solid. In the first case, the heat evolved corresponds to a known quantity of solid but to an unknown quantity of molecules in the surface phase. The enthalpy change corresponds to the transition from a starting point, which is the energy of the solid, either dry or wet. The end point is the establishment of a solid-liquid interface. The integral heat corresponding to "immersion" is A
i m m Q = A s ( H s r H s ° ) (J/g of solid)
(Eq. 20)
133
Phenomenology of Water Adsorption at Clay Surfaces
Figure 12 - Sketch of a typical immersion set used as part of a differential calorimeter The integral heat corresponding to "wetting" is: AwQ=As(Hsl-Hsv)
(J/g of solid)
(Eq.21)
By definition, the heat of adsorption is the difference between immersion and wetting. The heat measured during a differential process is:
d
ads
q= ads
^
(J/mole of vapor)
(Eq. 22)
an
v
JAS
All these processes are exothermic because adsorption is spontaneous and the entropy of the vapor molecules decreases. Corresponding enthalpies can be defined, and the most widely used are enthalpies per unit surface area or per mole of water, which are both intensive parameters. Using enthalpies per mole of water is not always appropriate because the number of moles of clay is not well defined, either chemically or numerically. 6.3 - Gas chromatography Inverse gas chromatography (IGC) under infinite dilution conditions, i.e. when injecting small amounts of gas allows the detection of some surface properties, and then allows a characterization of solid-water interactions [59-61]. A chromatographic column is filled with the solid of interest. The retention time and the retention volume of the carrier gas necessary to evacuate the probe through the column are determined as in a current chromatographic experiment. The retention volume Vn is linked to a free energy of adsorption at a certain pressure:
134
J.M. Douillard and F. Salles
A ads G = -RT l n V n + C
(Eq.23)
where R is the ideal gas constant, T the temperature, Vn the retention volume and C a constant depending both on the total solid surface area in the experiment and on the value set for the water molecule in the temperature and pressure conditions considered. Different procedures have been proposed to extract a thermodynamic value from experimental data. The most often used is a surface tension y sv . This tension is depending on y°. The main question however, is the actual pressure where these energetic terms have to be attributed. 6.4 - Surface infrared spectroscopy Infrared spectroscopy is a valuable tool to identify the types of interactions between water and surface sites [62]. This is because O-H bonds give rise to a stretching frequency in a particular region of the infrared spectrum (typically between 3200 and 4000 cm"1). However, i) intensities generally cannot be used in the analysis, because they are very sensitive to the experimental settings, ii) the exact link between adsorption energy and band frequency is not currently established, iii) the exact position of the band depends upon the underlying structure of the surface layer, and the bands are easy to detect but complex to assign in detail. For example, Aluminol groups are believed to give rise to distinct bands depending on whether the aluminium atom is tetrahedrally or octahedrally coordinated. It is useful to consider the crystal structure to predict the number of OH bands that will occur in the infrared spectra in consideration. The number of surface hydroxyl species for a given mineral is related to the coordination number of oxygen in the bulk structure. Obviously, the occurrence of particular surface planes has also an influence and the problem could become quite complex. The main problem could be experimental in that sense that it is often difficult to distinguish between isolated, vicinal and geminal groups in infrared spectra. Furthermore, some OH group can be present in the bulk structure and then they contribute to the signal. Some relations between the frequencies and the energy of adsorption have been proposed [63], but a general and clear analysis is still lacking. 6.5 - TSDC The Thermally Stimulated Depolarization Currents (TSDC) technique [64] consists of determining the current created by a dielectric, which has previously been polarized at low temperature, returning to equilibrium state. Its performance can be significantly increased when used in a fractional analytical mode (Relaxation Map Analysis, RMA). This technique consists of polarizing only in a small fragment of the full spectrum to isolate a single relaxation process. By varying the values of the temperature of polarization Tp and depolarization Td (Tp-Td ~ 5°C) and repeating the process over the entire range, some elementary modes can be isolated one by one, allowing the elementary relaxations to be determined. Currents observed, concerning water adsorption onto clay, can be attributed to the exchangeable ions moving inside the surface phase during the adsorption process [65-66]. The evolution of the TSDC signal with different relative water vapor pressures can be followed by submitting samples to successive heating, and testing them at each step. 7 - Theoretical methods Different approaches are possible when investigating the adsorption of vapors onto solids. The exact approach would be a calculation performed in a same box with an
Phenomenology of Water Adsorption at Clay Surfaces
135
infinite crystal and an infinite number of vapor molecules, representing the dilute or condensed vapor. However, the number of atoms necessary in such a calculation exceeds the capability of computers. Furthermore, the crystallographic methods allowing division of the number of solid atoms are not valid for the vapor phase. Therefore, it is necessary to use approximated methods [67-72]. Two approaches are thus possible, which are similar in one point. In the both approaches, it is necessary to build the infinite crystal or clusters or layers of the crystal. This is done using lattice modeling techniques. Generally, crystallographic data is used. For clays, it is necessary to use neutron scattering data in order to know the location of hydrogen atoms. At this point, some assume that the solid is rigid and that the geometry of the sites does not vary during the adsorption process. However, some others use chemical modeling methods to build a "model crystal" of minimal energy. Therefore, in this case, the positions of surface atoms do not necessarily correspond to the positions deduced from crystallographic data. This point is a very controversial. Even though it is obvious that bulk solid atoms are fixed in geometry in agreement with the constraints imposed by the solid state, it seems possible that the surface groups, such as hydroxyls, should move in the adsorption process. Therefore, a possible move of surface groups and a rigidity of the sub-layer solid atoms should be the better model. It is very difficult to perform calculations using such a model and to relate such a model to experimental information. The point of the link between actual thickness of the interface and available information is still present. Both approaches are thus similarly limited by this problem. The difference between the two approaches is as follows. On the one hand, it is possible to study the solid independently of the vapor. It is then possible, to compute some terms corresponding to the infinite crystal or to semiinfinite layers. These terms are basically the energy, surface energy, electronegativity and chemical hardness. Generally, it is difficult to interpret such results at the microscopic scale, even though some local reactivity indexes have been proposed [73]. It is possible to compare related characters, computed independently for the solid and the vapor, such as the HOMO-LUMO combination or some chemical reactivity indexes. The surface enthalpy can also be used in order to estimate the surface tension of the solid [45]. Some results are listed in following paragraphs. On the other hand, it is possible to build clusters or layers representative of the solid and to perform calculations mixing solid clusters and vapor molecules in the same box. Computational chemistry methods, such as Force Fields (FF) [74], Monte Carlo (MC), molecular dynamics (MD), and quantum methods or density functional theory (DFT) can be used to address the problem. Each method uses a similar strategy by attempting to determine the most stable, or lowest energy, conformational state for the structure by examining the differences in potential energy caused by random changes in location and orientation of component molecules. Some assumptions have to be made in order to represent the different atoms necessary to the calculation [75]. The underlying approach in MC and MD simulations is to construct potential functions that model all of the known interactions in a system of ions, atoms, and molecules, then devise a strategy for sampling the phase space of the interacting system in order to compute its properties. In a MC simulation, the configuration space of the system is sampled randomly under the guidance of an algorithm following statistical mechanics. In a MD simulation, the phase space of the system is sampled through numerical integration of the equations of motion for each molecular species. MD can then give indications about the diffusion of vapor molecules in the surface phase. Quantum calculations are approximations, but these are of an extremely general kind. The calculations yield the total energy of assemblies of atoms, and forces
136
J.M. Douillard and F. Salles
on the atoms. This knowledge enables the calculation of a vast range of properties, which can be known at the microscopic scale. Unfortunately, for complex systems such as crystals of clays, some terms have to be set, such as the charges in order to get neutrality. This point has to be verified when trying to interpret results. The limit of these approaches is obviously the respective numbers of vapor and solid atoms used in the computation. Moreover, with such sophisticated methods, the comparison between representative clay structures is still at the beginning. The most studied sample in these approaches is montmorillonite [76-80]. 8 - Trends For sake of simplicity, chemists can consider that clays belong to three great families, namely Kaolinite, Smectite-Illite, and Chlorite. Some average values can be compiled from literature data and used as guidelines. They describe the behavior patterns of the adsorption of water upon clays. 8.1 - Surface Energy Surface enthalpy can be considered as the macroscopic force, which drives adsorption. In the case of crystals of clays, which are composed of layers, it is possible to compute the surface energy by subtracting the energy of a semi-infinite layer from the energy of the infinite crystal. This difference is related to energy of extraction and then to the surface energy corresponding to a precise surface area. This is a well established approach for ionic crystals [81-83]. Obviously, this approach is valid if electrostatic forces, which depend on the dimensions of the solid considered, are predominant. In the case of solids such as clays, which are not strictly speaking ionic, such a method could be considered as non-valid. However, in such solids it is possible to use the concept of atomic charges in molecules [84] and then to compute realistic electrostatic forces acting, although atomic charges are not quantum-mechanical observables. For this reason, the charges can be derived by different models and it is possible to get non-unique values. However, keeping the same atomic charges in the layer and in the crystal allows the computation of a sound difference between the effective attractive electrostatic energies of the crystal and the layer. Thus, as recently published, it is possible to estimate a surface energy (without taking into account repulsive forces) for non-ionic crystals following different routes [85-86]. The Table 1 presents such results, using one of the methods [85] (the precise data and crystal descriptions will be detailed and published elsewhere in the near future). Results reported here correspond to crystals representative of the different groups composing clays, as detailed in the nomenclature part of this paper. Of course actual clays are not as simple as the crystallographic files used to establish these results. Moreover, as we will see below, it seems that the siloxane surface energy depends very strongly on the environment. Potassium kanemite therefore, even though it is a very simple network of siloxane, is perhaps not representative of the actual type of siloxane constitutive of clay structures. These results are in excellent agreement with a semitheoretical analysis of immersion data [45] (see figure 13) and hence are supposed sufficiently accurate to build a general view of the energy of clay surfaces. Indeed, if the repulsive forces are a small fraction of the total forces, then it is possible to assume that the energy discussed here is overestimated by about 10 %. An order of energy appears and two points are clear. In terms of surface energy, brucite, silica and gibbsite external layers cannot be considered as elemental bricks composing clay. The bonds between a silica layer and a gibbsite layer for instance, strongly modify the energy of the "pure" silica layer. This is clear when looking at the difference of energies between gibbsite
Phenomenology of Water Adsorption at Clay Surfaces
137
and brucite and the near energy between serpentine (a lizardite) and kaolinite. Similarly, it is not possible to consider that the surface energy of a chlorite is the sum of the surface energy of a brucite sheet and of a talc sheet. The second point is the influence of the sheet effective charge. When the charge of the interlayer cation increases, the surface energy of the clay increases. This trend can be observed inside the group of montmorillonites. The substitutions in illites or chlorites act in the same way: the surface energy increases with the effective charge of the layer. A scheme summarizing the trends can be drawn (see Figure 14). The surface energy increases with the complexity of the structure and with the effective charge (see also [96]). Obviously, the structure influences the charge, because the number of substitutions and combinations able to increase the charge increases with the sandwiching.
Figure 13 - Comparison with values of surface energy (mJ/m2) obtained for actual clays from heat of immersion data [45].
Figure 14 - Schematic evolution of the surface energy with the structure of model clays.
138
J.M. Douillard and F. Salles
Table 1 - Surface energies (enthalpies) computed from the model described in reference [85] using crystallographic data [77, 87-95). Basic layei
Sandwich
Model Solid
Surface Energy (ffJmJ/m2 245 738 2270 480 565 470 390 425
Brucite Brucitea Gibbsite Gibbsiteb Silica ring Kanemitec Silica-Brucite 2 layers Serpentine6 Silica-Gibbsite Kaolinitef Silica-Brucite-Silica 3 layers Talcd Silica-Brucite-Silica Li-Montmorillonite8 Silica-Brucite-Silica Na-Montmorillonite8 Complex 3 (Silica-Brucite-Silica)840 Cs-Montmorillonite8 strong charge layers (Silica-Brucite-Silica)strong charge Illiteh 650 (Silica-Brucite-Silica)Brucite ideal Chlorite without charge1 930 (Silica-Brucite-Silica)Chlorite containing Fe III, 1465 Brucite with charge on sheets1 a Ref[87], b Ref[88], c Ref[89J; d Ref [90]; e Ref[91], fRef[92J; g Ref[93] and Ref [94]; h Ref [77]; i Ref [95]. 8.2 - Chemical reactivity indexes Some chemical properties of solids or molecules are considered useful in order to understand chemical reactivity, namely the electronegativity, the hardness and the Fukuy indices [97]. Some of them can be easily calculated using computers and lattice modeling techniques associated with crystallographic data. There are many definitions of electronegativity, the capability of a molecule to attract electrons. This debate is not essential here, because it is possible to pass from one scale to another. The definition according to Mulliken, however, is very useful, namely the arithmetic average of ionization energy I and electron affinity, EA: X=
(£,.24)
In this definition, elctronegativity appears to be energy. Furthermore, it is also possible to relate an "exact" electronegativity to the negative of the chemical potential, defined inside the DFT theory (fJ,——/)- The derivative terms of this chemical potential with respect to the number of electrons have a physical meaning, namely the chemical hardness (or simply hardness) r| and its inverse, the softness S:
139
Phenomenology of Water Adsorption at Clay Surfaces
In the same DFT theoretical description, it is possible to define three Fukuy functions, as derivative of the electron density. These three functions are respectively indexes of the nucleophilic attacks, electrophilic attacks and radical attack. Due to the physical meaning of electronegativity as expressed by equation (24), it is obvious that these reactivity indices are related, at least qualitatively, to the value and the shape of the Frontier Orbitals, namely the HOMO and the LUMO of the global system. It has been shown that the HOMO density gives the same indication about electrophilic attack as the corresponding Fukuy index. The LUMO density description gives indication about nucleophilic attack. In Table 2 the electronegativities computed for the model clays selected are reported. The Table also presents the maximal partial charges observed for oxygen atoms of the structure. In order to obtain these two terms, the Winpacha [98] software and the model of electronegativity both due to Henry, which combines the electronegativity scale of Allen [99-101] and the Henry's hardness scale have been used. Chem 3D software (semi-empirical AMI computation) has been used in order to qualitatively determine the HOM O and LUMO locations. This has been done for neutral layers built using the crystallographic sub-software of Winpacha and same crystallographic files as preceding issued from Neutron scattering data. Table 2 - Electronegativity and hardness values of the layers corresponding to the solids described in Table 1. Allen and Henry scales have been used.
Model Solid Brucite" Gibbsite" Kanemite (siloxane) 0 Serpentine e Kaolinite' Talc11 Li-Montmorillonite g Na-Montmorillonite 8 Cs-Montmorillonite 8 Illite" ideal Chlorite without charge 1 Chlorite containing Fe III, with charge on sheets'
Electronegativity (eV)
Hardness (eV)
8.9
11 9.9 9.6
3.09 6.21 6.50 2.46 1.92 3.03 1.53 1.52 1.53 1.38 2.84
Maximal Atomic Charge on Oxygen 0.649 0.667 0.560 0.696 0.654 0.682 0.582 0.581 0.519 0.596 0.763
9.8
2.87
0.746
10.7 9.55 10,5 11.5
11 11.5 11.33
The full results (indeed, electronegativity and hardness are related to HOMO and LUMO levels, see equation (24)) are in quantitative agreement and have been compared satisfactorily with more precise calculations performed using the Gaussian 98 software for some selected samples. The pictures of the orbitals of chlorite are reported in Figures 15 and 16 as an example. The trends are the same for all the systems. The results are in qualitative agreement. The differences in electronegativity between the different clay models are small. The electronegativities of the different solids lie inside values imposed by the electronegativity of the constituting atoms, i.e.
J.M. Douillard and F. Salles
140
about 10 eV (see Table 3). The maximal charge is always located on the oxygen atoms, which do not coordinate protons. Most of them are situated on the siloxane planes. The HOMO is located on these atoms. However, the LUMO is more diffuse. Therefore, if these oxygens can be considered as well-defined sites for electron-donor action, there are no precisely located electron acceptor sites. Even though, it is frequently argued in the literature that OH groups are reactive groups [102-104], this type of calculation (quantitatively speaking approximated, but qualitatively correct) is not in agreement. Table 3 - Electronegativity and Hardness values of the constituting Atoms, in Allen and Henry scales. Atom 0 Si Al Mg
Electronegativity (Allen's scale) eV 21.36 11,33 9.54 7.65
Hardness (Henry's scale) eV 31.999 13.483 10.975 11.259
Figure 15 - Schematic representation of an isocontour map of the HOMO of a "layer" of chlorite (about 1000 atoms). The hydrogen atoms are located at the upper layer of the structure (see Figure 8). Neither the electropositivity, nor the LUMO density is in agreement with the expected special reactivity of some precise electron-acceptor sites on the surface of clays. Moreover, the differences of reactivity between the different clays do not appear important, even though there is a trend, which is in agreement with experiment. It is clear that talc and chlorite are harder than the other samples. (It has been argued that the chemical hardness corresponds to a physical hardness [47], even though the necessary link between the microscopic and macroscopic scales has not been established. It is clear that this point is not confirmed here: talc is the softer crystalline material known in the Mohs hardness physical scale).
Phenomenology of Water Adsorption at Clay Surfaces
141
This result has to be compared with the inertia of these types of talcite minerals versus water. Therefore, in Table 2, the low hardness values indicate a possibility of exchange reactions with water of the ions constituting the solid.
Figure 16 - Schematic representation of an isocontour map of the LUMO of a layer of chlorite (about 1000 atoms) 8.3 - Adsorption enthalpies The integral adsorption enthalpies reported below, which are in agreement with compiled data, are taken from immersion results obtained in the laboratory for welldefined members of each family, saturated where possible with Na+. The adsorption enthalpies at zero coverage have been deduced from these results taking the molecular area of the adsorbing water as 18 A2 which seems a reasonable value at low coverage. With this value, the results obtained are in agreement with average extrapolated values at zero coverage generally reported in the literature for a family. The film pressures have been obtained by integration of adsorption isotherms obtained by volumetric methods in the laboratory. The film pressures are in agreement with classical literature data. The number of layers is estimated in agreement with the usual multilayer models. The surface tension versus vacuum is obtained by using an empirical model [45]. It is clear, looking at the result obtained for illite, that this estimation is approximate. However the results deduced from this surface tension, namely the contact angle assuming that ySL is nil (following classical literature) are surprisingly good in the case of talc [52,105], and correct for kaolinite (see below). The enthalpies of adsorption obtained have to be compared to the surface energies reported above. There is a strong link between these two types of results, indicating that an estimation of the adsorption effect using the crystallographic data is possible (see Figure 17 for the correlation). Linearity is not expected, even though we believe that surface energy is the major influence on the adsorption heat. At a primary level of interpretation, the surface tension is 40% of the surface energy.
142
J.M. Douillard and F. Salles
Table 4 - Characteristics of the samples and values of the different adsorption energies measured or computed for the different clay families at ambient temperature and pressure. Chlorite Family Talc Kaolinite Mite Montmorillo nite Pyrenees Origin Pyrenees Charentes Vosges Algeria 30 2.5 Specific Surface Area of the 50 1.9 110 (P/P°<0.3) sample 277 (m2/g) (P/P°>0.9) 230 260 75 130 89 Film Pressure ne (mN/m) 750 304 325 203 Integral Enthalpy of 222 Adsorption (mJ/m2) 22 35 24 Differential 33 81 Enthalpy of Adsorption (kJ/mol) 376 1052 465 Surface Enthalpy from 216 650 immersion analysis (mJ/m2) 8 Final number of layers in the 8 1.7 4.5 5 multilayer 506 97 181 Estimated Surface Tension 226 310 Ts (mN/m) Estimated contact angle (°)
71
45
0
?
0
Figure 17 - Comparison of the surface energies computed using crystallographie files of model solids and of representative adsorption enthalpies on the same type of solid. The line is a guide for the eyes.
Phenomenology of Water Adsorption at Clay Surfaces
143
9 - Difficulties of Interpretation 9.1 - Estimation of the quantitative surface energy relative to the cations The influence of cations in such energies could be estimated, by exchanging a cation present in the structure by another cation and then by comparing clays saturated by different ions. Experiments have been performed and the observed trends give indications about the influence of the cation on water adsorption. Generally the heat evolved during hydration increases when substituting Ca2+ and Ba2+ to Li+ or Na+ [106]. The heat increase is about 8 kJ/mole with a free energy term about 4 kJ/mole [29, 107]. Monovalent ions have also an influence on the observed heat of hydration, in the order Cs+, Rb+
Mgr+ Ba^ K+ A1 J+
Contact Angle (°) 16 15 23 24 26 28
Adsorption of water on faces or edges of sheets can be viewed approximatively as a direct physical interaction, i.e. interaction between a single molecule and a group of atoms in the absence of a chemical reaction. The influence of entropic terms is then easy to comprehend in the framework of the structuration of the surface phase. The structure of the adsorbed phase composed of solid molecules and of "frozen" vapor molecules has an entropy value between solid and vapor. However, the interaction between cations and water is different. It is well known that cations move inside the clay structure, and the cation-water interaction can be viewed as a change of the solid microscopic structure or even as an extraction from the solid structure, that is a dissolution effect. Essentially, energies of dissolution of solids in water are the difference between the energy of the solid network and the energies of hydration [84]. If one assumes that the solid can be separated into two ions, namely the cation and a global anion corresponding to the sheet or to the network, all the energies depend on the radii of the considered couple of ions, with two varying laws. The total energy, which is a difference, depends on one hand on the radius of the cation, but on the other hand, on the "radius" of the cavity inside the clay network. Therefore, it is not possible to compare different clays having the same cation-form, without introducing a systematic error. This is due to the structure of the cavity, which is unfortunately not known precisely. It is a particular characteristic of clay, relating to its history and not simply to its mineralogical family. (Evidently, the same discussion is possible for mobile anionclay forms). Moreover, two other effects have to be taken into account. Firstly, it is difficult to interpret values obtained per mole of water, because a cation interacts with more than one molecule of water. This number depends on the radius and the charge of
144
J.M. Douillard and F. Salles
the cation [110]. The number of water molecules interacting at short distance has been estimated in some papers, and varies between 3 (Li) and 7 (Ca2+). Secondly, the hydration energy of the individual cation can vary from -262 kJ/mol of cation for Cs+ to -515 kJ/mol of cation for Li+. The ratio between entropic effects and energy depends, however, both on radius and charge of the cation and varies greatly (from positive values (Cs+) to negative values (small ions or multivalent ions: Al3+, Mg2+)). Moreover, little is known about this ratio in interfaces. A complete interpretation is thus quite impossible and could only been achieved by comparison with theoretical estimations. It has however been possible, through TSDC measurements [66], to estimate in some clay structures the energy corresponding to the influence of hydration in the displacement of Na+ ions due to adsorption of water. This is about 30 kJ/mole, and compares with the gas phase binding enthalpy for Na+/H2O: 100 kJ/mol. It seems reasonable to consider that this effect is linked more or less directly to the quantity of ions in the structure (about 10"4 mol/g of clay, depending obviously on the clay studied). An order of magnitude of the enthalpy due to the hydration of ions Na+ can then be given. For the samples studied, which belong to the Smectite-Montmorillonite-IUite group [66], this influence varies between 21 % and 66 %, a large range. To the best of our knowledge, such work has not been undertaken for other ions. 9.2 - Estimation of the enthalpy of adsorption onto one face of a sheet Starting from these results, it is possible to estimate the energy corresponding approximately to one type of face for talcites. This is because talc and chlorite are very similar and talc has the stacking sequence: s-b-s. From this point of view, chlorite could be considered as a "composite" compound with one s face and one b face (the reasoning is not as general as that in section H. 1). The value obtained for the enthalpy of adsorption of pure talc (203 mJ/m2) indicates that one virtual face of a sheet composed of talc siloxanes gives an enthalpy of adsorption of 101.5 mJ/m2. It is then possible to estimate the value due to the b face. Immersion results obtained with controlled mixtures of talc and chlorite are in agreement with this rough analysis [111-113]. Table 6 - Empirical Estimation of the enthalpy of adsorption onto the different faces of talcites. Face
Brucite Silica
Enthalpy of adsorption of water computed for one virtual face of talc or chlorite (mJ/m2) 650 101
Enthalpy of adsorption of water on a virtual half-sheet (kJ/mol) 70 11
These estimations, however, can only be used as guidelines for mixtures of talc and chlorite, which are frequently used in industry. The reasoning cannot be extended to other clays, where siloxanes, as discussed in the section H. 1, appear very energetic and seem to depend strongly on their environment. These results are, however, in good agreement with some analysis of water adsorption on clays, aluminosilicates and silicates [114]. In these approaches, the silica faces or layers appear as "low-energy surfaces" [14, 115-116]. However, it is frequently argued that the presence of hydroxides enhances the adsorption energy. Two points complicate the problem but can also explain this discrepancy. The first is the influence of the cation on the charge of the clay network. The charge increases the attractive
Phenomenology of Water Adsorption at Clay Surfaces
145
forces. This point has been discussed above, but other arguments can be added. For instance, generally clays with charged layers have more wetting ability than uncharged clays. It is clear when comparing contact angles for a TOT uncharged structure (talc for instance: 60-70° with water) with a TOT charged structure (vermiculite: 15° with water) [108]. Therefore, a term due to the surface charge has to be added to the values listed in Table 6, although it is not yet quantitatively estimated. The second point is due to an electrostatic force related to the length of the lamellae. It is well known that the dimension of a plane has an influence on the strength of its attracting field. Some results obtained with well-characterized samples have shown that the apolar part of the adsorption energy is enhanced by the lamellarity of the samples [117]. 9.3 - Estimation of the ratio between enthalpy and entropy It is often not clear why adsorption occurs when considering on the one hand the enthalpy of adsorption on some sites and on the other hand the enthalpy of association of water in gas phase. These both being of the same order of magnitude. The actual driving force of such a pseudo-reaction is the free energy. The loss of entropy due to adsorption is of paramount importance. Considering the movements of the water molecules, adsorption corresponds to the replacement of energy of translation by energy of vibration in the cluster water/solid. It is often argued that the fact that clays are materials which are different from others concerning water adsorption, is due to the charge of the surface. Even though clays are very complex materials concerning the interpretation of adsorption, starting from analysis of literature data and results presented here, they are not so different from other materials, for instance silicates. It is easy to estimate the entropy of adsorption from available data of heat and isotherms for representative solids. In the case of silicates, the entropy variation related to the adsorption of water is of the order of 8 J/mol.K (hydrophobic silica) up to 48 J/mole.K (quartz). For very hard materials such as silicon carbide, the entropy drop is largest, at about 131 J/mol.K. The values obtained for clays are in this range (from 46 J/mol.K for talc, to 177 J/mol.K for chlorite). This is obviously due to the combined presence of many sites of high potential energy: siloxanes, cations, and charges of sheets. This influence of a strong enthalpy due to chemical attraction, however, is limited by the fact that these chemical effects are not acting in the same way concerning the structure of the adsorbed phase. Firstly, concerning non-electric terms (i.e. dispersive), the entropic term decreases when the enthalpic term increases (compared to the enthalpy of adsorption, the tension drop is proportionally less important for chlorite (ratio = 0.35) than for kaolinite (ratio = 0.42), see Table 4). Physical adsorption gives rise to short-range forces, which structure up to five layers of water. When the attractive field is strong on the first layer, this creates a distortion of the hydrogen bonds between water molecules, and a loss of energy is due to the rearrangements necessary to recover the bulk water structure. The ease of rearrangement depends on the distance between the adsorbing sites and then on the surface crystallography of the solid. Secondly, forces due to electrostatic are more longranged. This creates a distortion of the hydrogen bonds of the non-electrostatic adsorbing layer. Therefore, if the two enthalpy terms (electric and non-electric) are additive, the two free-energy terms, i.e. tensions are not. The result is that the global adsorption behavior of water onto clays is not greatly different from those observed with other materials not bearing electric charges such as silicate [118] or carbides [119120]. An attempt has been made to attribute the entropy drop only to movements of the
146
J.M. Douillard and F. Salles
water molecules and not to the molecules of the solid surface layer. The results were not conclusive, even though the appropriate orders of magnitude were obtained [121]. 9.4 - Position and distance of the water molecules These points can be deduced from chemical modeling calculations. Programs allow minimal energy configurations to be computed. It is generally considered that a minimal energy corresponds to an equilibrium configuration. However, the estimations of the literature are in poor agreement, depending greatly on the chosen route. Some researchers have studied the influence of the position of the atoms of hydrogen, pointing towards or away from the clay surface. The majority of them have focused on the distance between some selected planes of molecules onto the surface of the clay and the center of the atoms composing water. The reliability of such results is obviously related to the number of atoms involved in the calculation. (This is not the case for calculation of chemistry reactivity indices, which are less expensive in time and less dependant on the number of molecules studied). Moreover, in many cases, Monte-Carlo simulations are used and the quality of such calculations is driven by the force field chosen. Unfortunately, the forces actually acting in the surface phase are not precisely known if complete quantum mechanics calculations are not achieved. It seems that the location of hydrogen atoms in water pointing towards the clay surface is favored. However, there is a poor agreement between the different calculations giving the distance between the solid and the water molecules. The distance between Si and hydrogen atoms is often given about 4 A [122-123] even though some very short distances have been obtained, about 2 A [124]. Both values seem realistic when compared to the few results of neutron diffraction, giving the distance of water towards zeolites, which are comparable samples [125-126]. Supplementary neutron data is now necessary in order to understand more fully the situation. 9.5 - Swelling Adsorption of water onto clays from the montmorillonite group leads to a very special effect called "swelling", which defines the "smectite" clays. This is mainly due to the fact that the interlamellar counterions are hydrated in the interlamellar space. The effect occurs because the electrostatic charge of the montmorillonite layers is very low and does not imply a massive presence of positive charges in order to be equilibrated. The reason behind this is the major contribution of the substitutions in octahedral sites. When water is adsorbed, the beginning of the process for montmorillonites is similar to other clays. However, at a certain value of water inside the interlamellar space, the chemical potential of water is sufficient to break the low energy bonds between layers, and the interlamellar distance increases. The attraction between the layers is then insured by "liquid network" composed of water and ions. The surface area available to adsorption then varies during this process, and it is necessary to interpret adsorption isotherms taking into account this variation of the extensive term of energy (Figure 18). It is possible to assume a discontinuous variation of the surface area or a continuous variation. The surface area can increase by more than two hundred percent during the process. Many simulations have been done to interpret the behavior of this very special liquid, which fills the interlamellar space. This problem is similar to the problem of attractive forces between plates immersed in solvents [127]. From the viewpoint of adsorption, the montmorillonite group is not a special case, if taking into account the change of surface area during the process [29]. Indeed, the heats observed are comparable to the heat evolved during immersion of kaolinite, for instance. The only difference is the force of the interaction between layers, which is mainly due to a charge
Phenomenology of Water Adsorption at Clay Surfaces
147
effect. When swelling, the energy between layers is decreased, and the driving force of adsorption is the migration of the cation from the solid surface to the interlayer structure (extraction and hydration). This energy has been estimated [66] to be about -176 kj/mol of exchange cation. In conclusion, montmorillonites are characterized by interactions of the same energy between layers and between water and faces of layers. This is not the case of kaolinites (strong interaction between layers) or of talcs (low interaction between layers but low adsorption energies).
Figure 18 - Adsorption isotherm of water onto a Na-Montmorillonite at ambient temperature. It is possible to distinguish some steps corresponding to a modification of the thickness between layers. The relative importance of the steps depends on the nature of the exchangeable cation. The specific surface area increases since P/P°>0.3. 1 0 - Summary At the microscopic level, three main interactions can be considered. Firstly; the dispersive interaction, which can be attributed to a "network", i.e. to the clay considered as a whole. The complexity of the sandwiching enhances the attraction force. The lamellarity also has an influence here, due to an increase of dispersive force with increasing of the dimensions of the sheet. Secondly, the donor-acceptor interaction, which occurs between polar groups and water. The active surface polar groups of clays are mainly silica rings. Associated with this siloxane surface is a roughly hexagonal cavity, formed by the bases (triads of oxygen ions) of six corner-sharing silica tetrahedra. The reactivity of the siloxane surface depends on the structure where it is involved and on the value of the charge in the clay mineral layer. If cation substitutions are absent from the underlying layer structure, a siloxane surface will function as a weak electron donor. If substitution of Al3+ by Mg2+ occurs in the clay layer, the resulting excess of negative charge distributed over several surface oxygens increases the electron donor effect. If isomorphic substitution of Si4+ by Al 3+ occurs in the tetrahedral sheets, the excess negative charge is greater and the electron donor effect is stronger. If the hydrogen of water is pointed towards the clay surface, the interaction is easy between the water LUMO and the clay HOMO localized on oxygens of siloxanes. Therefore, it is probably the favored
148
J.M. Douillard and F. Salles
situation. Concerning electron acceptor effect, it is dubious to state that they will be localized upon some surface groups. Something looking like an hydrogen bond can probably be formed between the oxygen of the incoming water and the LUMO of the clay surface. When cations are present, the electrostatic forces are predominant. This explains why in theoretical simulations, water molecules are grouped around interlayer cations. Energies of the different surface hydroxides have been widely investigated [128-132] in terms of acid-base strengths or hydrophilic character. However, it appears that the hydrophilicity of clays cannot be related to the number of hydroxyls located on the surface. The poor hydrophilic character of talcs presenting siloxane surfaces and where the silanol groups are absent casts a shadow on this reasoning. However, the reactivity of the siloxane is greatly influenced by its environment. Therefore, the main reason for the poor hydrophilic character of talc is more its hardness than the absence of hydroxyl groups. The third major type of interaction is due to cations. Water adsorption onto cations located inside sites creates two competitive charge transfers. The first is between the electrons of an oxygen atom from the solid, and a hydrogen atom from a water molecule. The second is between the cation and the oxygen atoms of water molecule. These two effects compete with the initial transfer between cation and oxygen atoms of the solids, and facilitate the extraction of the ion. The low level of extraction energies measured by TSDC shows this. Adsorption at high partial pressure is due to direct interactions between water vapor molecules and adsorbed molecules. The energies are the order of liquefaction energies and the limiting effect is the structure of the first layer, which is imposed by the heterogeneity of the solid surface. Some samples with "middle" energy surfaces can adsorb high quantities of water probably because they present on a same plane electron donor sites at appropriate distances compared to the liquid water structure. Approximate but realistic energies of surface and of adsorption can be computed from theoretical models using crystallographic neutron data. The main theoretical problem remains the estimation of the energy of the individual surface groups on real clay samples. 11 -References [1] R.C. Mackenzie, De Natura Lutorum, in Proceedings of the Eleventh National Conference on Clays and Clay Minerals, Pergamon Press, pp. 11-28, 1963. [2] S. Guggenheim and M. T. Martin, Clays and Clay Miner., 43 (1995) 255. [3] P.A. Schroeder, Web site: http://www.gly.uga.edu/schroeder/geol6550 /reservebook.html. [4] S.W. Bailey, Summary of Recommendations of AIPEA Nomenclature Committee, Clays and Clay Miner., 28 (1980) 73 [5] S.W. Bailey Ed., Reviews in Mineralogy, vol. 13: Micas, Series Editor P. H. Ribbes, Mineralogical Society of America Publications, 1984. [6] W.A. Deer, R.A. Howie and J. Zussman, Rock Forming Minerals, 2nd ed., Longman, 1986. [7] S.W. Bailey Ed., Reviews in Mineralogy: vol. 19: Hydrous Phyllosilicates (exclusive of Micas), Series Ed. Paul H. Ribbes, Mineralogical Society of America Publications, 1988. [8] M.J. Singer and D.N. Munns, Soils an introduction, MacMillan Publishing Company, 1987. [9] L. Yan, C.B. Roth, and P.F. Low, Langmuir 12 (1996) 4421.
Phenomenology of Water Adsorption at Clay Surfaces
149
[10] Web site: http://mineral.galleries.com/minerals/silicate/class.htm. [11] S.A. Nelson, Web site: http://www.tulane.edu/~sanelson/geol211 /clayminerals.pdf. [12] Web site: http://www.man.ac.uk/Geology/MineralWeb/MineralWeb.html. [13] G.W. Brindley, G. Brown, Crystal structures of Clay Minerals and Their X-Ray Identification, Mineralogical Society, London, 1980. [14] A.W. Adamson and A.P. Gast, Physical Chemistry of Surfaces, Wiley, New York, 6th ed., 1997. [15] W. Rudzinski. and D.H. Everett, Adsorption of Gases on Heterogeneous Surfaces, Academic Press, London, 1992. [16] R.K. Her, The Chemistry of Silica, John Wiley and Sons, New York, 1979. [17] Z. Zhou and W.D. Gunter, Clays and Clay Miner., 40 (1992) 365. [18] M.J. Sparnaay, The Electrical Double Layer, Pergamon Press, Oxford, 1972. [19] J.A. Greathouse, S.E. Feller, and D.A. McQuarrie, Langmuir, 10 (1994) 2125. [20] P. Hobza, J. Sauer, C. Morgeneyer, J. Hurych and R.J. Zahradnic, J. Phys. Chem., 85(1981)4061. [21] W.F. Bleam, Clays and Clay Miner. 38 (1990) 522. [22] J. Lyklema, Fundamentals of Interface and Colloid Science, Academic Press, New York, 1995. [23] C.S. Johnson, Jr. and D.A. Gabriel, Laser Light Scattering, Dover, New York, 1981. [24] S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, London, 1982. [25] J. Rouquerol, D. Avnir, C.W. Fairbridge, D.H. Everett, J.M. Haynes, N. Peraicone, J.D. Ramsay, K.S.W. Sing and K.K. Hunger, Pure Appl. Chem., 66 (1994) 1739. [26] D.H. Everett, Basic Principles of Colloid Science, Royal Society of Chemistry, 1988. [27] R. Defay, I. Prigogine, A. Bellemans and D.H. Everett, Surface tension and adsorption, Longmans, London, 1966. [28] R.G. Linford, Chem. Rev., 78 (1978) 81. [29] V. Medout-Marere, H. Belarbi, P. Thomas, F. Morato, J.C. Giuntini and J.M. Douillard, J. Colloid Interface Sci., 202 (1998) 139. [30] C.H. Bridgeman, A.D. Buckingham, N.T. Skipper and M.C. Payne, Mol. Phys., 89 (1996) 879. [31] B.J. Teppen, K. Rasmussen, P.M. Bertsch, D.M. Miller and L. Shafer, J. Phys. Chem. B, 101,(1997)1579. [32] A. Delville and M. Letellier, Langmuir, 11 (1995) 1361. [33] J.W. Gibbs, The Collected Works of Josiah Williard Gibbs, Green and Co, London, 1928. [34] J.D.V. Van der Waals und P. Kohnstamm, Lehrbuch der Thermostatik, Leipzig, 1927. [35] J.S. Rowlinson and B. Widom, Molecular Theory of Capillarity, Clarendon Press, Oxford, 1982. [36] J.W. Cahn and J.E. Hilliard, J. Chem. Phys., 28 (1958) 258 [37] E.A. Guggenheim, Thermodynamics, 3 ed., North-Holland Publishing Company, Amsterdam, 1957. [38] J.C. Ericksson, Arkiv. fur Kemi, 25 (1965) 331; 25 (1965) 342; 26 (1966) 49. [39] A.I. Rusanov, Progress in Surface and Membrane Science, Academic Press, New York, 1971.
150
J.M. Douillard and F. Salles
[40] E. Tronel-Peyroz, J.M. Douillard, L. Tenebre, R. Bennes and M. Privat, Langmuir, 3 (1987) 1027. [41] M. Privat, R. Bennes, E. Tronel-Peyroz and J.M. Douillard, J. Colloid Interface Sci., 121 (1988) 198. [42] E. Tronel-Peyroz, J.M. Douillard, R. Bennes and M. Privat, Langmuir, 5 (1989) 54. [43] J. Fraissard and C.W. Conner, Eds., Physical Adsorption: Experiment, Theory and Applications, Nato ASI Series, Kluwer AC, Dordrecht, 1997. [44] T.L. Hill, J. Chem. Phys. 17 (1949) 520, J.A.C.S. 74 (1952) 1598. [45] J.M. Douillard, Immersion Phenomena, Interfacial Dynamics, (Surfactant Science Series/88), Ed. N. Kallay, Marcel Dekker, New York, 1999. [46] J.M. Douillard, J. Colloid Interface Sci., 188 (1997) 511. [47] R.G. Pearson, Chemical Hardness, Wiley-VCH, New York, 1997. [48] C.S. Brooks, J. Phys. Chem., 64 (1960) 532. [49] W.A. Zisman, Adv. Chem. Ser., vol.43: Contact Angle, Wettability, and Adhesion, Ed. R.F. Gould, p. 1, A.C.S., Washington DC, 1964. [50] J.M. Douillard and V. Medout-Marere, J. Colloid Interface Sci., 223 (2000) 255. [51] D.Y. Kwok, D. Li, and A.W. Neumann, Langmuir, 10 (1994) 1323. [52] J.M. Douillard, J. Zajac, H. Malandrini and F. Clauss, J. Colloid Interface Sci., 255 (2002)341. [53] D. Aspnes, Optical Properties Of Solids - New developments, Chap. 15: Spectroscopic Ellipsometry of Solids, Ed. B. Seraphin, North-Holland, 1976. [54] V. Laperche, J.F. Lambert, R. Prost and J.J. Fripiat, J. Phys. Chem., 94 (1990) 8821. [55] M.B. Me Bride, T.J. Pinnavaia and M.M. Mortland, J. Phys. Chem., 79 (1975) 2430. [56] N.T. Skipper, M.V. Smalley and G.D. Williams, A.K. Soper, C.H. Thompson, J. Phys. Chem., 99 (1995) 14201. [57] J. Rouquerol, K.S. Sing, F. Rouquerol, Adsorption by Powders and Porous Solids, Elsevier, 1999. [58] B. Fubini, Thermochimica Acta 135 (1988) 19 [59] A. Saada, E. Papirer, H. Balard and B. Siffert, J. Colloid Interface Sci., 175 (1995) 212. [60] H. Balard, O. Aouadj, E. Papirer, Langmuir, 13 (1997) 1251. [61] H. Balard, Langmuir, 13 (1997) 1260. [62] J.T. Yates Jr. and T.E Madey, Eds., Vibrational Spectroscopy of Molecules on Surfaces, Plenum Press, New York, 1987. [63] A. Auroux and A. Gervasini, J. Phys. Chem., 94 (1990) 6371. [64] J. Vanderschueren and J. Gasiot, Thermally Stimulated Relaxation in Solids, Chap.: Field-Induced Thermally Stimulated Currents, Ed. P. Braunlich, pp. 135-223, Springer-Verlag, New York, 1979. [65] S. Devautour, J.C. Giuntini, F. Henn, J.M. Douillard, J.V. Zanchetta and J. Vandeschueren, J. Phys. Chem. B, 103 (1999) 3275. [66] V. Medout-Marere, H. Belarbi, A. Haouzi, J.C. Giuntini, J.M. Douillard, J.V. Zanchetta and J. Vanderschueren, J. Colloid Interface Sci., 223, (2000) 61. [67] P. Comba, T.W. Hambley, Molecular Modeling of Inorganic Compounds, WilerVCH, Weinheim,2001. [68] J.B. Foresman, A. Frisch, Exploring Chemistry with Electronic Structures Methods: A Guide to Using Gaussian, Gaussian Inc. Pub., 1993. [69] G.H. Grant and W. Graham Richards, Computational Chemistry, Oxford Science Publications, Oxford, 1995.
Phenomenology of Water Adsorption at Clay Surfaces
151
[70] F. Frank Jensen, Introduction to Computational Chemistry, John Wiley & Sons, 1999. [71] A.R. Leach, Molecular Modelling: Principles and Applications, 2nd Edition, Prentice Hall, 2001. [72] R.T. Cygan and J.D. Kubicki, Eds., Reviews in Mineralogy and Geochemistry, Vol. 42: Molecular Modeling Theory: Applications in the Geosciences, Mineralogical Soc. America Ed., 2001. [73] R.G. Parr, W. Yang, International Series of Monographs on Chemistry, n° 16: Density-Functional Theory of Atoms and Molecules, Oxford University Press, N.Y., 1994. [74] S. Hwang, M. Blanco, E. Demiralp, T. Cagin, and W.A. Goddard III, J. Phys. Chem. B, 105(2001)4122. [75] B.J Teppen, D.M. Miller, S.Q. Newton, L. Schafer, J. Phys. Chem., 98 (1994) 12545. [76] N.T. Skipper, F.R Chou Chang, and G. Sposito, Clays and Clay Miner., 43 (1995) 285. [77] S.H. Park and G. Sposito, J. Phys. Chem. B, 104 (2000) 4642. [78] J.A. Greathouse, K. Refson, and G. Sposito, J.A.C.S., 122 (2000) 11459. [79] R. Sutton, G. Sposito, J. Colloid Interface Sci. 237 (2001) 174. [80] S.H. Park and G. Sposito, Phys. Rev. Lett., 89 (2002) 85501. [81] M.L. Huggins and J.E. Mayer, J. Chem. Phys., 1 (1933) 643. [82] R. Car, M. Parinello, Phys. Rev. Lett., 55 (1985) 2471. [83] CM. Bertoni, A.I. Shkrebtii, R. Di Felice and F. Finocchi, Progr. Surf. Sci., 42 (1993)319. [84] J.E. Huheey, E.A. Keiter, R.L. Keiter, Inorganic Chemistry: Principles of Structure and Reactivity, Harper Collins, 1993. [85] J.M. Douillard and M. Henry, J. Colloid Interface Sci., 263 (2003) 554. [86] M. Henry, Solid State Sci., 5 (2003) 1201. [87] F. Zigan, R. Rothbauer, Neues Jahrbuch fuer Mineralogie, 4 (1967) 137. [88] H. Saalfeld, M. Wedde, Zeitschrift fuer Kristallographie, 139 (1974) 129. [89] M.T. le Bihan, A. Kalt, R. Wey, Bull. Soc. Franc. Min. Crist., 94 (1971) 15. [90] B. Perdikatsis, H. Burzlaff, Zeitschrift fuer Kristallographie, 156 (1981) 177. [91] M. Mellini and P. F. Zanazzi, Am. Min., 72 (1987) 943. [92] R.A. Young, A.W. Hewat, Clays and Clay Miner., 36 (1988) 225. [93] S.I. Tsipursky and V.A. Drits, Clay Miner. 19 (1984) 177. [94] R. Calvet, Thesis, Hydratation de la Montmorillonite et Diffusion des Cations Compensateurs, Paris 6 University, 1972. [95] D.O. Nelson, S. Guggenheim, Am. Min., 78 (1993) 1197. [96] J.M. Douillard, V. Medout-Marere, Acid-Base Interactions: Relevance to Adhesion Science and Technology, vol. 2, Surface Energy and Acid-Base Properties of Solids studied by Immersion Calorimetry, pp. 317-347, Ed. K. Mittal, VSP, The Netherlands, 2000. [97] H. Chermette, J. Comp. Chem., 20 (1999) 129. [98] M. Henry, in "Modelling of Minerals and Silicated Materials", Eds. B. Silvi and P. d'Arco, pp. 273-334, Kluwer Ac. Pub., The Netherlands, 1997; Chem. Phys. Chem., 3 (2002) 561; Chem. Phys. Chem., 3 (2002) 607. [99] L.C. Allen, J.A.C.S., 111 (1989) 9003. [100] J.B. Mann, T.L. Meek, and L.C. Allen, J.A.C.S., 122 (2000) 2780. [101] J.B. Mann, T.L. Meek, E.T. Knight, J.F. Capitani, and L.C. Allen, J.A.C.S., 122 (2000)5132.
152
J.M. Douillard and F. Salles
[102] W. Hertl and M.L. Hair, J. Phys. Chem., 72 (1968) 4666. [103] T. Morimoto, M. Nagao and J. Imai, Bull. Chem. Soc. Jpn., 44 (1971) 1282. [104] C. Contescu, J. Jagiello and J.A. Schwarz, Langmuir, 9 (1993) 1754. [105] R. Chi, Z. Xu, T. Difeo, J.A. Finch, J.L. Yordan, J. Pulp Pap. Sci., 27 (2001) 152. [106] R. Keren and I. Shainberg, Clays and Clay Miner., 28 (1980) 204. [107] I. Berend, J. M. Cases, M. Francois, J.P. Uriot, L. Michot, A. Masion and F. Thomas, Clays and Clay Miner., 43 (1995) 324. [108] B. Janczuk, E. Chibowski, M. Hajnos, T. Bialopiotrowicz and J. Stawinski, Clays and Clay Miner., 37 (1989) 269. [109] M. E. Schrader and S. Yariv, J. Colloid Interface Sci., 136 (1990) 85. [110] J. Burgess, Ions in Solutions, Ellis Horwood, 1988. [111] H. Malandrini, Ph. D. Thesis, Montpellier (France), University of Sciences, 1995. [112] J. M. Douillard, H. Malandrini, T. Zoungrana, F. Clauss, S. Partyka, J. Therm. Anal., 41 (1994)1205. [113] J. Zajac, H. Malandrini, Pol. J. Chem., (1997) 686. [114] A. Delville, J. Phys. Chem., 99 (1995) 2033. [115] J. Sauer, P. Hobza and R. Zahradnik, J. Phys. Chem., 84 (1980) 3318. [116] N. T. Skipper, K. Refson and J. D. C. Me Connel, Clay Miner., 24 (1989) 411. [117] H. Malandrini, F. Clauss, S. Partyka and J.M. Douillard, J. Colloid Interface Sci., 194(1997)183. [118] H. Malandrini, R. Sarraf, B. Faucompre, S. Partyka and J.M. Douillard, Langmuir, 13 (1997) 1337. [119] V. Medout-Marere, A. el Ghzaoui, C. Charnay, J.M. Douillard, G. Chauveteau and S. Partyka, J Colloid Interface Sci., 223 (2000) 205. [120] V. Medout-Marere, S. Partyka, R. Dutartre, G. Chauveteau+ and J.M. Douillard, J. Colloid Interface Sci., 262 (2003) 309. [121] J.J. Fripiat, A. Jelli, G. Poncelet and J. Andr<5, J. Phys. Chem., 69 (1965) 2185. [122] A. Chaterjee, T. Iwasaki, T. Ebinaz, H. Hayashi, Applied Surface Sci., 121-122 (1997) 167. [123] A. Delville, S. Sokolowski, J. Phys. Chem., 97 (1993) 6261. [124] S. Berhouet and H. Toulhoat, Langmuir, 10 (1994) 1832. [125] F. Pechar, W. Schaefer, G. Will, Zeitschrift fuer Kristallographie, 84 (1933) 373. [126] K. Shiokawa, M. Ito, K. Itabashi, ZEOLD, 9 (1989) 170. [127] J.N. Israelachvili and R.M. Pashley, Nature, 306 (1983) 249. [128] E. Garrone and P. Uglingo, in Structure and Reactivity of Surfaces, Eds. C. Morterra, A. Zecchina and G. Costa, Elsevier, Amsterdam, The Netherlands, 1989. [129] J. Sauer, J. Phys. Chem., 91 (1987) 2315. [130] W.J. Mortier, J. Sauer, J.A. Lercher and H. Noller, J. Phys. Chem., 88 (1984) 905. [131] L.J. Michot, F. Villieras, M. Francois, J. Yvon, R. Le Dred and J. M. Cases, Langmuir, 10(1994)3765. [132] M.L. Hair and W. Hertl, J. Phys. Chem., 74 (1970) 91.
A WORKSHEET MODEL FOR ADSORPTION/DESORPTION OF IONS ON CLAY SURFACES GIORA RYTWO School of Environmental Sciences and Technology, Tel Hai Academic College, Upper Galilee 12210, Israel, MIGAL, Galilee Technological Center, Kiryat Shmona, ISRAEL. E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
154
G. Rytwo
1 - Introduction Adsorption/desorption processes play a vital role in determining the efficacy and environmental behavior of pollutants or nutrients in soil. Adsorption is defined as the accumulation of matter at the solid-liquid interface. Most of the adsorption in natural systems occurs in the dispersed phase that consists predominantly of inorganic colloids (clays, metal oxides, metal hydroxides and metal carbonates) and organic colloidal matter [1]. Adsorption processes are considered to be very important, since they determine the amounts of nutrients, metals, pesticides, and other organic chemicals retained on the sorbent [2]. By doing that, the amounts of nutrients available for the plants, or pollutants free to migrate to undesired zones are determined. Other mechanisms and reactions as precipitation, dissolution, catalysis and redox processes on the matrix depend on adsorption processes: by influencing the distribution of substances between the aqueous and the solid phase, adsorption also affects the electrostatic properties of suspended particles and colloids, making impact on their tendency to coagulate or settle, and the reactivity of the surfaces [1]. Thus, the determination of the amounts adsorbed to sorbents is crucial for prediction of the environmental fate of pollutants or nutrients. Predicting the behavior of reactive solutes in complex systems requires an accurate description of the sorption, and relatively complicated experiments to confirm the predictions. In the case of multi-component systems with several interacting species, the processes can only be described with models accounting for the chemical reactions between the matrix and the species in the solution [3]. However, in the case of a single component adsorbed by clay minerals, the retention can be estimated from relatively simple experiments. The use of adsorption/desorption models to predict scenarios, is essential to evaluate critical concentrations of chemicals before they became an environmental or agricultural hazard. There is a wide array of adsorption models based on equilibrium reached in the system [2]. Modeling is usually carried out by two different approaches: (i) empirical models, and (ii) mechanistic models [4]. The empirical models give a simple description of the experimental data with no particular theoretical basis. An example is the frequent employment of the Freundlich isotherm. Some models arise from theoretical considerations. The Langmuir adsorption isotherm, for example, is developed from the mass law [1] following four basic assumptions (adapted from [2]): (1) Adsorption occurs on a fixed number of identical sites placed on planar layers, and only 1:1 adsorption is allowed. (2) Desorption might occur- thus, adsorption is reversible. (3) There is no lateral movement of molecules. (4) The adsorption energy is the same for all sites. Thus, adsorption is independent of surface coverage. The Frumkin equation evolved from the Langmuir isotherm model by considering lateral effects. More mechanistic models, or surface complexation models, make reference to thermodynamic concepts such as specific reactions described by mass action laws and material balance equations, and/or electrostatic equations. Several different mechanistic models use different descriptions of the electric double layer, around the charged particles. These models simplify the ion distribution next to the surface, and require that the ions be located on some main adsorption planes. The models differ in the number of specific planes for location of ions, in such a way that the surface reaction is modeled, and how the electrostatic potential is assumed to change between the planes [5]. In capacitance models, for example, there is a linear relationship between surface charge and surface potential, based on the assumption that all surface complexes are inner sphere, and anions are adsorbed by ligand exchange [2, 6]. All the ions are considered specifically adsorbed on one plane, and thus they will
Adsorption/Desorption of Ions on Clay Surfaces
155
experience the same surface potential, with no explicit consideration of the effect of the background electrolyte. In diffuse layer models [7], the equations that relate the surface potential to the charge are based on the Gouy-Chapman theory. Triple layer models are in general an extension of the Stern theory, in which two constant capacitance planes and one diffuse layer plane between them are considered [8,9]. One of the major problems with many diffuse double layer and surface complexation models is the large array of adjustable parameters that are employed to fit the experimental data to the calculation [2]. This led Westall and Hohl [10] to the conclusion that one model is about as good as another, as long that the adjustable parameters are correctly set. The essential set of equations, and the initial computerized version of the model presented in this study were developed in 1978 by S. Nir [11] for predicting the adsorption of mono- and divalent inorganic cations by synthetic membranes. The model was extended afterwards to deal with batch experiments in closed systems [12] and adapted to sorption on clay minerals [13]. The main concept is the solution of the electrostatic Gouy-Chapman equations, combined with calculations of the adsorbed amounts of the cations as the sum of those residing in the double-layer region (Gouy layer), and the cations that are chemically bound to the surface (Stern layer). All the calculations are performed in a closed system, caring constantly for the mass balance of each ion, thus, the adsorption causes a decrease in the equilibrium concentrations of cations in solution. During the years the model was tested, extended and adapted for a variety of systems. The first study on clay minerals tested the adsorption of mono and divalent inorganic cations by SWy-1 montmorillonite [14]. In 1988 adsorption of monovalent organic cations by SWy-1 montmorillonite was successfully predicted by introducing the formation of charged complexes consisting of two cations and one surface site [15], as presented in Section 2.3. The model could successfully predict the unexpected increase in the adsorption of organic monovalent cations upon the increase in the concentration of inorganic cations by more that one order of magnitude (see Section 3.2). Hirsch et al [16] introduced complexation of Cd in solution, and its adsorption as divalent (Cd2+) and as monovalent (CdCl+) species. Yermiyahu et al [17] used the model to evaluate competitive adsorption of mono- and divalent inorganic cations to melon plasma membranes. Until 1995, the model solved the Gouy-Chapman equations analytically. This limited the calculations to mono- and di-valent cations and only one type of anions. Rytwo et al [18] introduced an algorithm that solved the electrostatic equations numerically. That allowed expanding the model to additional ions and species. Adsorption of trivalent cations, such as La3+ and Al3+ by wheat plasma membranes was evaluated, while elucidating the mechanism of plant resistance to Al toxicity [19]. Neutral sorption sites, e.g. silanol groups, were introduced [20], in order to account for adsorption of organic cations to sepiolite. This version was tested later for adsorption of the herbicides diquat and paraquat, which are divalent organic cations [21]. These studies employed binding coefficients that were determined in previous studies for adsorption of those cations to montmorillonite [22], without any further adaptation or calibration. Undabeytia et al [23] tested simultaneous binding of Cu to edge and planar charged sites, on the same clay mineral, considering the double layers formed by both sites while clay platelets are surrounded by the same equilibrium solution. Model calculations also accounted for changes in pH and their influence on the binding of H+, the speciation of Cu, and their binding as Cu2+ and as CuOH+. Other papers applied the mentioned adsorption model as it, and gave a very good correspondence between measured and calculated adsorbed amounts, after calibrating the model for the chemical tested, with only two adjustable binding coefficients for each cation [24-29]. Recent
156
G. Rytwo
studies [30] tested the applicability of the mentioned model for the adsorption of a monovalent and a divalent organic cation on several soils, considering the cation exchange capacity and the specific surface area of the soils. As all Gouy-Chapman-Stern (GCS) models, this one observes what occurs to charged particles in suspension. It should be recalled that irrespective of the origin of the surface charge, electrical neutrality requires that an equal amount of charge of opposite sign must accumulate in the liquid near the charged surface. Several models can be developed to describe such behavior in suspensions. The Gouy-Chapman equation was widely used for soil systems, but such models had also several biological applications as evaluating electrokinetic surface charge [31] or electro-osmotic models [32]. The main equation arises from the combination of Poisson's Law and Boltzmann's distribution, and the main assumptions involved are [33]: 1. The charge is uniformly spread on the surface. 2. The charge in the solution is built up by unequal distribution of point charges. 3. The solvent is considered as a continuous medium, with invariant properties (dielectric coefficient, density, temperature). 4. Ions and surface are only involved in electrostatic interactions. 5. The surfaces are flat and infinite, and the distance between surfaces is also infinite. The GCS model used for this study solves numerically the electrostatic GouyChapman equation for the relevant suspension in each experiment or sample. All cations in the suspension, with their respective valencies, are considered, and their spatial distribution as a function of the distance from the adsorbing surfaces is accounted. This allows evaluation of the surplus of each cation and the depletion of each anion in the double layer region. Changes in the actual surface charge density of the adsorbing surfaces that occur due to the specific cation binding, are iteratively considered while solving the equations. The interpretation of the results calculated by electrostatic/mechanistic models should not be overestimated: Even the simplest natural system contains several types of clay minerals, organic matter and other colloidal particles that would influence the sorption behavior. As a partial list of the severe limitations in the use of any model we may quote three different studies: 1. When modeling natural systems, it is imperative to understand rates of chemical processes to accurately predict the fate and transport of ions and organic compounds in soil and water natural environments [34]. In other words- equilibrium based model are limited, and kinetic rates might influence considerably. 2. Utilization of bulk mineralogical data of clay minerals as cation exchange capacity (CEC) or specific surface are (SSA) to represent predominant phases in natural systems often has failed to predict reliably solute and contaminant behavior [35]. 3. Electrostatic models may not be defensible for complex system as soil, or natural environments, due to the complexity of the sorbents, that do not conform to the essential assumptions of double layer models [36], summarized above. As the matter of fact- even a simple planar surface of a smectite platelet does not fulfill all the limitations, since the charges (for example) are not uniformly spread, neither homogenous: some of them arise from octahedral isomorphic exchanges, others from tetrahedral exchanges. Under those severe limitations, the model presented here still manages to deliver initial predictions that fit very well with measured results. This is achieved by using only two adjustable parameters for each cation in the system: (1) the coefficient for the formation of a neutral complex, and (2), the coefficient of the formation of a charged
Adsorption/Desorption of Ions on Clay Surfaces
157
complex. Those two parameters remain unchanged even upon changes in temperature, sorbent concentration, ionic strength, etc. The advantage of the model presented here is its availability and flexibility. As it can be obtained from the internet as a worksheet version, it can be easily adapted for use by any researcher. 2 - Description of the model A wide and complete description of the model at different stages of its development was presented in several papers [4,11,15,16]. In this work, we will attach to the computational procedure of the worksheet version of the model, presented in Fig. 1. By following it we will explain the equations and assumptions needed for each stage. 2.1-The sorbent The first stage is to present details of the sorbent in case. Since the ions adsorb on the charged sites of the clay mineral, we need to know the amount of charged sites per unit volume of suspension; in other words, the concentration of the charged sites, denoted as PT. The concentration of charges is a function of two other, more common parameters: the cation exchange capacity (CEC) of the clay in case, and the mass of clay per volume of suspension. By knowing these details we may obtain PT by means of Eq. (1):
PT ^
j = CEC p ^ l (clay concentration)^]
(Eq. D
Since double layer calculations depend on the surface charge density, the area of each charge is needed. An idealization of uniformity and equality for all charges is assumed here. Such idealization might be easily justified in cases were the charge arise from isomorphic exchanges in the octahedral layer, thus the negative charge is spread on the surface over a relatively broad area. Such is a case generally for montmorillonite clays. For clays with large amount of tetrahedral exchange the approximation of homogenous spread charge assumed in most double layer model might not be valid. For homogenous spread charges the area of each charge equals the specific surface area (SSA) of the clay divided by the amount of charges, which depends on the CEC and on Avogadro number (NAvogadro) that defines the number of atoms (or in this case, charges) per mole:
A [ m ;]
CEC I""10'6" I N
fatomsl
L kg J — LmoleJ
(Eq. 2)
SSAM
LkgJ This value can be transformed to any convenient unit of area. For example, by multiplying by (1010)2 = 1020, we might calculate A in units of A2. Additional parameters that should be incorporated at an early stage of the calculation are the thermodynamic temperature (T [K°]), the dielectric coefficient of the solvent (s), and several physical constants such as the gas constant (/?), the charge of an electron (e), Boltzmann's coefficient (k), although the latter might be evaluated using k=R/NAvogadro since it represents the gas constant for a single atom or molecule.
158
G. Rytwo
Figure 1 - Schematic diagram of the adsorption model, and the computational procedure. Numbers in brackets refer to sections in the text. In the worksheet version of the. model, all the initial details and sorbent parameters are introduced in "general input" worksheet. Calculation of PT and A is performed after introducing the concentration of the sorbent (mass per volume), its CEC and SSA. For example, a suspension of 1 g L"1 Wyoming montmorillonite, with a CEC of 0.796 molec kg"1, and a SSA of 8xlO5 m2 kg"1 will yield PT=1.59xlO"3 M and A=
159
Adsorption/Desorption of Ions on Clay Surfaces
1.67xlO"18 m 2 = 167 A2. The intrinsic initial surface charge density, aini, can be calculated, by dividing the absolute charge of an electron (e) by the area per single charge. In the sample above, cr/ra=95.9xlO~3 C m"2. Table 1 lists all the parameters that are incorporated in the model and do not depend on the ions concentration. 2.2 - Cations in suspension Table 1 - General parameters used bv the model Parameter
Symbol1 Units
area per site
A
Avogadro number Boltzmann's constant Cation exchange capacity concentration of adsorbing sites concentration of empty sites concentration of sorbent in suspension dielectric coefficient electron charge gas constant intrinsic initial surface charge density specific surface area y(x)- function of potential at distance x from clay y(0)- y(x) at the surface of the clay
N
K
m2 atoms mole"1 JK"1
Default value
Remarks calculated
6.02xl0
23
1.38xlO"23
Relevant equations (2.1-2.5) 2
constant
2
constant
9,10 1,2
CEC
molec kg"1
external input
PT
M
calculated
1
P"
M
calculated
3, 4, 5, 6, 8,
external
kgL OA
oU
s e R
C electron'1 J mole"1 K"1
-1.6xlO"19 8.314
Cm"2 SSA
m 2 kg-'
input external input constant constant calculated from e and A external input calculated
Y(x)
1
9, 10
2
15, 16, 1S
1O
Initial value inputted by user. Calc. final value
Y(0)
calculated
8,11
external 9,10 input a - cations in the equilibrium solution, b - cations in the Stern layer, c - surplus of cations in the diffuse Gouy layer. This "surplus" might be negative (actually, deficit) in cases where the surface is positively charged due to adsorption above the CEC of the sorbent. temperature
T
K°
298
160
G. Rytwo
As described in the introduction, several models might yield very good predictions when the concentrations of all species, except the ion in case, are fixed. In order to account for changing conditions in all species in suspension, all the ions must be introduced to the model. Usually an increase in the amount of one cation might lower the amount adsorbed of other ions. One of the important features of the model presented is that it can calculate the influence of the amounts adsorbed on the total concentration, thus, it does not consider the equilibrium concentration as constant. Due to that, the total concentration (Cro/) of each ion should be introduced, that means that all the contributions of each ion should be considered: added salts, ions that arrive as adsorbed to the clay, and background concentration in the solvent. In the general case, the total concentration of each cation is divided to three different sinks: Even though it is obvious than some anions might change considerably the adsorbed amounts by forming complexes in solution, the worksheet version of the model is relatively simple, and considers only adsorption of cations to negatively charged sites in the clay. More specific versions of the model accounted for complexes of cations with CT [16,26], OH" [19,23] or even initial evaluations of complexation with organic acids [29]. Such kinds of complexes might be easily added to the worksheet version, if needed, upon considering the speciation coefficients, and the binding coefficients of the charged species formed. 2.2.1 - Cations bound to the surface of the clay (Stern layer) When a negatively charged particle, for example, a clay platelet, is introduced into a solution containing cations and anions, migration of the ions occurs. Cations in the suspension move toward the surface of the particle, being attracted by electrostatic forces. Fractions of those cations are specifically and chemically bound to the surface of the sorbent (cations in the Stern layer). For each cation in the suspension two general types of complexes are considered: neutral and charged. A neutral complex is formed when a +Zi valent cation binds to a -Zi charged site on the surface of the clay. Figure 2 gives a schematic representation of a neutral complex formed between a monovalent cation and a monovalent site (Fig. 2a), a divalent cation and a divalent site (Fig. 2b) or a bivalent cation and a trivalent site (Fig. 2c). Since, for example, a divalent charged site on the surface of the clay is formed by two adjacent monovalent sites, the general description of the formation of a neutral complex PCi is: [Ci(O)Zl+]+[PZl ~] = [ C i ( 0 ) Z l + ] + ^ o [ P C i ]
(Eq. 3)
where [Ci(O)Zl+] represents the concentration of the +Zi valent cation i near the surface of the clay (at zero distance of the site), [P2''] is the concentration of -Zi valent sites, that equals the concentration of monovalent sites [P~] divided by the valency of the ion Zi. An interesting issue not sufficiently elaborated is whether activity coefficients in the double layer are not considerably lower than unity. Kinraide [37] considered activity coefficients when modeling the binding of ions to membranes. Although this issue will be further elaborated upon in Section 2.8, Njus [38] denotes the similarity between the Debye-Huckel theory for activity coefficients around fixed point-charges, and the GouyChapman theory that describes ion distribution as a function of the distance from a planar surface. The model assumes activity coefficients in double layer are close to unity, leading to activities that equal the concentrations of all ions. Thus, the binding coefficient of such reaction, Ki, is defined by the mass law:
Adsorption/Desorption of Ions on Clay Surfaces
161
Figure 2 - Different types of complexes considered by the adsorption model. The left side shows formation of neutral complexes between a monovalent cation and a monovalent site (a), a divalent cation and a divalent site (b) and a trivalent cation with a trivalent site (c). The right figure shows charged complexes between two monovalent cations and one monovalent site (d), a divalent cation and a monovalent site (e) and a trivalent cation with a monovalent site (/). Ki
[PCi] _ [PCi] -[pz-][ci(o)-]-[n[c.(O)Z1+]
(Eq.4)
Zi The assumption [PZl~] = " y . is based on the fact that since all the sites are placed on the same clay platelet, each two sites might react as a divalent site, and each three sites might act as a trivalent site (see Fig. 2). This definition does not come instead thermodynamic or kinetic considerations in general definitions of equilibrium coefficients, where concentration of each participant in the reaction is raised to the power of the stoichiometric coefficient. The combination of those two considerations might lead to a theoretical expression for the rate of formation of a divalent or trivalent site, as described in Section 2.9. Until now we assumed the formation of neutral sites only. For cations of higher valency we have to consider the possibility of charged complexes that arise by the binding of a monovalent site with a di-, tri- or four-valent cation. A schematic representation of the formation of charged complexes appears in Fig.2. Divalent and trivalent cations form charged complexes following the same mechanism: a monovalent site binds to a divalent cation (Fig.2e) or to a trivalent cation (Fig.2f):
162
G. Rytwo
p-+Ci(0) Z l + oPC p i ( z - 1 ) +
(Eq. 5)
where PCpi represents the concentration of a charged complex between cation CiZl+ and monovalent sites. As in Eq.(4), a general definition for the binding coefficient in such reaction might be written: —
[PC i < Z W ) + l
Ki = J i V
_L
(Eq.6)
[p][Ci(0)Zl+]
It is clearly seen that monovalent organic cations can't yield to a charged complex with this mechanism. However, it was observed that large monovalent organic cations might adsorb to clays beyond the CEC [39] forming positively charged organoclays. Such effects were ascribed to a combination of both coulombic interactions and van der Waals attractive forces [40]. Margulies et al [15] introduced this interaction into the adsorption model, by assuming a non-coulombic reaction between a second monovalent organic cation, and a neutral complex formed by one monovalent cation and one monovalent site on the surface (Fig.2d): PCi+Ci(0)+<S>P(Ci)2+
(Eq. 7)
since this reaction differs from Eq.(5) it demands a different definition for the binding coefficient for the formation of a charged complex of a monovalent organic cations. Introducing into it the concentration of a neutral complex as calculated using (Eq.4), thus: Ki =
[P( C l ) 2 J [PCi][Ci(0n
[P(CQ2 ] Ki[P][Ci + fy(0) 2
(monovalent organic cations only) (Eq. 8)
Apparently, the same model should have been used for divalent organic cations, thus: PCi+Ci(0) + + » P ( C i ) 2 + +
(Eq. 7b)
However, Rytwo et al [22] had shown that charged complexes for adsorbed divalent organic cations as diquat and paraquat, arise from the binding of one divalent molecule to one monovalent site, as shown in Eq. (5) and Fig. (2e). It should be emphasized that in Eqs (3)-(8) the concentrations of the cations should be obtained near the surface of the clay platelet. This is not the concentration at the equilibrium solution. Since the distribution of the cation as a function of their distance from the surface, depends on the electric potential, we may use Boltzmann's equation:
[Ci(x)] = [Ci(oo)]e
kT
(Eq.9)
were e is the absolute magnitude of an electronic charge, q>0 is the potential at the surface of the sorbent, k is Boltzmann's factor, and 7 is the absolute temperature. For the sake of
Adsorption/Desorption of Ions on Clay Surfaces
163
simplicity we may define a variable named y(x), were x denotes the distance from the surface as:
y(x) = e
kT
(Eq. 10)
and Eq.(9) at the surface of the platelet (x=0) may became: [Ci(0)] = [Ci(a>)].y(0)Zl
(Eq. 11)
where y(0) denotes the value ofy at the surface of the sorbent. Thus, for each ion its concentration near the surface may be calculated upon knowing its concentration in the equilibrium solution, Ci(<x>), and the potential of the surface which yields a specific value of y(0). The value obtained for Ci(0) might be used afterwards to evaluate all complexes of the cation in case, using Eqs. (3)-(8). In summary, for each cation in the system the following details are needed: a. The valency of the cation, Zi b. Total concentration of the cation in the whole system, Cilol c. Binding coefficient of the cation for the formation of a neutral site, Kt d.
Binding coefficient of the cation for the formation of a charged site, Kt
In the worksheet version of the model, all those parameters are introduced in "ions input" worksheet. For each cation in the system there is a line in the worksheet named "output", that contains all the details about it: those that were introduced as input (name, valency, total concentration and binding coefficients), those that represent the different complexes formed (neutral and charged) and the surplus of each cation in the double layer, which is described in the next section. More specific versions of the model had considered also the formation of mixed complexes between two different monovalent organic cations, aggregation of organic cations in solution [18], adsorption to neutral sites [20] and even adsorption on two different type of charged sites, with different surface potential, PT and A [41, 23]. Each different complex might be added to the model, upon considering its influence on the mass balance of the cations in case, and the changes in the surface potential. 2.3 - Cations in the diffuse layer In the equilibrium solution there is equality between the amount of negative and positive charges. Near the negatively charged surface, we will usually find a surplus of cations neutralizing the surface charges. The assumptions in Gouy-Chapman theory lead to the conclusion that all ions with the same valency behave exactly the same in the double layer. This assumption is merely an approximation, as Cs+ will neutralize sites more effectively than Na+, due to it less tightly attached hydration shell. However, considering the basic assumption as correct, the proportion between all the Zi valent ions in the equilibrium solution will be exactly the same as in the double layer. For example: If Na+ is 50% of the monovalent cations in the bulk solution, it will still remain 50% of the monovalent cations in the diffuse layer, although the exact amount will obviously not be the same, due to the amplification caused by the negative potential of the charged surface. The same arguments can be used for anions, and the only difference is that their concentration near the surface will be lower than at infinity. If we define as Qz the surplus of the z-valent ions, then we might define Di, the excess
164
G. Rytwo
concentration of the Z valent ion / in the double layer region above the equilibrium concentration, were Z can be any integer number between (-2) and (+4) except 0, (thus can be either a mono or divalent anion or a mono-, di-, tri or tetra-valent cation) as: [Di] = Q z ^ C i ( c ° ) ]
(Eq-12)
were XCi(z>(°o) will be the sum the concentrations of all the Z valent ions in the equilibrium solution. The quantities Qz are a result of an integration of the excess of cations of valency Z over the double layer region. When the system includes only mono and divalent cations, and one type of anions, they can be calculated analytically [11]. For the general case, that will be described here, these values can be obtained numerically [42]. We define as Csj the excess of ions of type ;' in the double layer region above the equilibrium concentration. As said before, this excess would be the sum over the whole width of the double layer, denoted as d, of the concentration of ions / at the segment dx width, (Cj(x)) minus the concentration of those ions at the segment at infinite distance of the surface (Cj(oo)). Reducing dx to an infinitesimal value, would transform the summation into an integration. Introducing the relationship between the concentration as a function of the potential (Boltzmann distribution, Eq.(9)) and the definition of y(x)(Eq.(lO)) we may write: J
d
CS1 = J(C,(x)-C,(*))dx=C,(<x>) J(y(x)Zl -l)dx 0
(Eq- 13)
0
Thus, if we will be able to obtain a value for y(x) for any distance from the surface x, the integration in Eq.(13) can be made numerically. If a function y(x) is given, then evaluations of its value for any x can be made by developing the function to a Taylor series:
y(x+ h) = y(x) + hg + ^-§ + .... dx
(Eq.14)
2! dx
we may begin at the surface of the sorbant, and advance each time and use y(x) to calculate y(x+h), providing we know the value of y(0) and at least the first and second derivative of y(x). The full development of the first derivative is presented in Section 2.10. The first derivative ofy(x) is: ^tyCx^Cl^C.MiyW21-!)
(Eq. 15)
the sign + is essential due to the square root operation. Cl is a constant that depends on the units in use. When non-rationalized electrostatic units are used, C12=—— To make EkT the second derivation easier we will rewrite after assigning:
Adsorption/Desorption of Ions on Clay Surfaces
S y =£c,(»)(y(x) Zi -l)
165
(Eq. 16)
Eq. (15) then becomes: y" y
dx
2
(Eq. 15b)
and the second derivative ofy can be easily calculated from that, yielding:
2
dx
s " 2 ^ + y(x).S - - - I . ^ ^ l [
y
dx
yy
'
y
(Eq. 17)
2 dy dx\
and from the definition of Sj, (Eq. 16), we can calculate: ^=yZi-Ci(oo)-y(x)Zi-1 dy *—'
(Eq. 18)
and infroduce the value in (17). By knowing the first and the second derivative of y we can evaluate y(x) at any value of x using Eq.(14), if an initial value of y(0) is known. Although the calculations are only approximate, the use of the second term of the Taylor series, and a small length step (h), minimizes the error. The value ofy(x) at any x allows to perform the numerical integration in Eq. (13). The values of Qz for the different valencies can then be calculated, by summing the results of Eq.(13) for all ions with valence z. In the worksheet version of the model, double layer calculations are numerically performed in "double layer" worksheet. The length of the step h is limited to a maximum value of 0.2 nm, but close to the platelet it calculated as a function of the first derivative of y(x) in order to account for sharp changes when the slope is steep. The numerical integration is performed up to a distance of approximately 80 nm from the surface. Such distance is long enough even for very low ionic strengths that yield relatively high surface potentials. Closure of the system was achieved even at values of y(0)=300 (equivalent to cpo=-15O mV). In any case, width of the double layer may easily be increased by adding rows in the relevant worksheet, although that increases slightly the time of each calculation. After obtaining values of Qz for all valencies, the surplus in double layer of each ion can be calculated using Eq.(12), if the equilibrium concentrations of all ions are known. 2.4 - Mass balance of ion /'. Most ion adsorption models calculate the adsorbed amounts by taking the equilibrium concentration as known. However, for most applications we are interested in evaluating sorbed amounts as a function of the total amount of nutrient or pollutant spilled over the system.
166
G. Rytwo
Figure 3 - Illustrative representation of the different contributions to the total amount of a monovalent ion. Table 2 -Relevant parameters for each ion in the model Relevant equations (2.1-2.5) 3,4,9, 11, 13, 15, 16, 18, 19b 9, 11, 12, 13, 15, 16, 18, 19, 19b 3,4,5,6, 11,
Parameter
Symbol
Units
Remarks
Valence of the ion
Zi
absolute integer
external input
Concentration of the ion at equilibrium solution
Ci(«)
M
calculated
Ci(0)
M
calculated
Ci,o,
M
external input
19, 19b
Ki
M"1
external input
4,19b
Ki
M"1
external input
6, 8, 19b
Pci
M
PC i
M
P(Ci) 2 +
M
Di
M
calculated
12, 19
Qz
M
calculated
12, 19b
calculated
12, 19b
Concentration of the ion near surface total concentration of the ion binding coefficient for neutral complex binding coefficient for charged complex neutral complex Charged complex of di or trivalent Charged complex of monovalent surplus of ion at diffuse layer Proportion factor of Z valency ions surplus in diffuse layer sum of al Z valent ions in the equilibrium solution
calculated using Ci(0), K, P" and Zi calculated using Ci(0), K, P" calculated using Ci(0), K, PCi
3,4,6, 19 5, 19 7 8
Adsorption/Desorption of Ions on Clay Surfaces
167
Clearly, if the amount of total cation is constant, sorption processes will reduce the equilibrium concentration, thus- the latter is not known a-priori in most cases. The model presented here considers the system as closed- thus a mass conservation balance for each ion in the system is constantly calculated, and adsorption of a molecule causes a reduction of its concentration in the equilibrium solution, and vice-versa. The total amount of each cation ,Cilot, includes (see Fig. 3): a. b. c. d.
the sum of the ions in the equilibrium solution- Ci(ao), the amount bound as neutral complexes- PCi, (Eq.(4)) the amount bound as charged complexes- PCpi, (Eq.(6) or (8), depending in the valency of the cation) the surplus of the specific cation in the diffuse layer around the sorbent- Di (Eq. (12)).
If the total amount of a cation is known, the concentration at equilibrium can be calculated by evaluating the mass balance of the cation: [Ci tot HCi(»)]+[PCi]+[PC p i]+[Di]
(Eq. 19)
In previous versions of the model, Ci(ao)v/as specifically calculated. For example, for the case of a divalent cation, Eq.(19) would become: [Ci tot ] = [Ci(»)]+Ki^[Ci(»)]y 0 Z l +Ki[p-][Ci(«)]y o zi
+ Q z
=i^L(Eq.l9b)
As can be seen, if the amount of empty sites F, the potential at the surface (and y(0), that is related to it by Eq.(lO)), Qz and the concentration of all other ions is known the equilibrium concentration can be isolated and calculated from a quadratic equation based on Eq.(19b). In the worksheet version of the model, the calculation of Ci(<x>) is numerically performed, based on the "goal seek" function present in worksheet programs. This procedure is fast and accurate, and allows fast convergence of the model. Table 2 concentrates all the relevant parameters for each ion. Additional species (i.e., complexes between ions, dimers, aggregates, binding to another type of sites, etc.) might be introduced to the system, as long as they are also considered in the mass balance for each relevant ion. 2.5 - Evaluating the potential at the surface and y(0). As can be understood from previous sections, the potential at the surface is essential in order to evaluate all other parameters: without it, it would be impossible to calculate the concentration at the surface needed for the Stern layer, neither the distribution in the double layer needed for Di. Without all those, the mass balance would not be calculated. This specific model is based on the Gouy-Chapman equation. Details on the steps needed for the development of the equation can be obtained in Singh and Uehara [33]. In the worksheet model, the parameter evaluated is y(0) which is a function of physical constants and the potential at the surface, q>(0) (see Eq.(lO)). Although there are several studies and models based on the same equation, different notations might be found for the Gouy-Chapman equation. In order to clarify the apparent discrepancies, we will quote three different examples:
168
G. Rytwo
g2=2sRTC,
7i
(oo) r . ^ ZiF(p(0) ^ V 2RT )
(
-ZiF(p(0)
basgd
md
Uehara
[33]
(Eq 20a)
N
RT _ j
e
on Singh
a2=—^ni(oo)(y0Zl-l)
based on Obi etal [43]
(Eq. 20b)
based on Nir etal [4]
(Eq. 20c)
In all the three versions,
(Eq. 20)
This equation must include all species in solution, including anions and neutral complexes. If we adopt a general notation Sj as the sum of all J-valent ions at the equilibrium solution where 4<J<-2 (tetravalent cations up to divalent anions) S^fC^oo)];
S 3 =X[C 1 ( 3 + ) (°o)];
S2=X[C'<2+)(<X)^
S^^JC.^K)];
[C <2 )(M)];
S-2= X i "
S
[C ( )(a))];
-'= Z ' "
S o = S 4 +S 3 + S 2 +S, +S_! +S_ 2 Eq.(20) yields:
(Eq.21)
Adsorption/Desorption of Ions on Clay Surfaces
G2=2ss0RT S 4 y 0 4 +S 3 y 0 3 +S 2 y 0 3 + S i y o +S0 + ^ + %
L
y
169
(Eq. 22)
» yo J
Even if all the concentrations are known there are two unknown variables in Eq.(22): y(0) and a. In order to close the system, we must also take a look at what happens at the surface of the sorbent. As described in Section 2.1, the initial amount of charged sites, and the area per site are properties of the sorbent, and depend on the CEC and the SSA of the clay. The intrinsic surface charge density of the sorbent (<7M) is given by dividing the amount of charges by the area per site. If only neutral complexes are present, the actual surface charge density,CT,will be a function of the concentration of free sites in the sorbent, since the occupied sites are uncharged. However, if charged complexes are possible, then the positive complexes make also influence to the actual value of a. The ratio between a and aini equals the ratio between the negative charges and the total concentration of sites, denoted [PT\. Since each free site contributes to the amount of negative charge, whereas each charged complex causes the negative charge decrease, then:
« ^-IPCP+
(Eq.23)
[PT] where \^ PC p + is the sum of all charges that arise from charged complexes. Thus, a charged complex formed by a trivalent cation contributes twice its concentration. By squaring Eq.(23) and combining it with Eq(22) we get the core equation of the model:
c2 = E J l 2 Z - I C J 2 i n i
=288oRT£[Cl(°o)](y(0)z. -l)
(Eq. 24)
which can be solved numerically, for y(0) using the "goal seek" function of the worksheet software. Figure 4 shows Eq.(24) and all the parameters that influence the calculation of it. 2.6 - Computational procedure From the mass balance equation for each cation, we may isolate the concentration of each ion in the equilibrium solution, Ci(co), as shown in Section 2.5, as long as we know y(0), P~, Qz, and all other ions concentrations. We may calculate all the species of each ion using Sections 2.3 and 2.4, as long as we know equilibrium concentrations (C/'fooj), empty sites (P~), y(0) and Qz. If all species are known, the amount of empty sites P~ might be calculated. The potential can be calculated using Eq.(24), as long as all other parameters are known. However, the potential is needed for y(0), which is needed for the first stage of calculating concentrations. From this short description it can be seen that the entire system can be solved only iteratively. In previous versions we followed a linear algorithm of calculating each parameter once, and going back to the beginning.
170
G. Rytwo
charges that arise from charged complexes f
[ee charged
\\ U
equilibrium
intrinsic surface charge density of the sorbent
*AJ « / ^
concentration of all species in suspension
\
^a ini =2E80RT£[Ci(«>)](y(0)Zi - l ] 2
IP II initial amount of charged sites in suspension
physical constants and unit transformation
a function of the potential at the surface
Figure 4 - Parameters influencing the "core function" of the model. For the worksheet version we saw that it converges faster if some steps are performed twice before going on to the following stage. Regardless of the exact procedure of computation, the system converges to one set of values that is the result for all the equations. The convergence is checked upon consecutive values of y(0) and can be brought to any desirable level. The full computational path used by the worksheet version of the model is presented in Fig. 1. The solution of the complete set of equations yields the surface potential of the sorbents, and the distribution of each of the ions among its various species and residence sites (neutral and charged complex, double layer, equilibrium solution, etc.). In the worksheet version any desired output can be designed. Some examples are presented in Section 3. 2.7 - Activity coefficients in the diffuse layer 2.7.1 - Debye length Theoretical expressions for activity coefficients are based on Debye-Huckel limiting law [1]. In a neutral solution the coulombic potential at a distance r from an isolated ion of charge Zi times an elementary electron charge e (1.609xl0"19 C), is cp,=-^ 47ter
(Eq.25)
B is the permittivity of the medium, and for water at room temperature it equals 78.5x8.85xlO"12 C2J"'m"'. The effect of the ionic atmosphere in a real solution around this ion is to cause the potential to decay with the distance more sharply than this expression implies [45]. We define then the shielded Coulomb potential as:
(Eq.26)
Adsorption/Desorption of Ions on Clay Surfaces
rD=P^
171
(El-27)
v p. and arises from the derivation of Poisson's equation, that gives the relationship between the volume charge density (pj) and the electrostatic potential. When rrj is large, the shielded potential is virtually the same as the unshielded potential, but for small values of the Debye length, the shielded potential can be considerably smaller, leading to the differences in the behavior of electrolyte solutions, that are attributed to activity coefficients y
Pi
=X e z J " J =X F z i C j(°) = Z F z i C J ( o o ) e ~ i 7 1 = Z F z J C J ( r o ) e ~ ^
(Eq28)
where F and R were already defines as the Faraday constant (96500 C mole"'), and the gas constant (8.314 J mole"1 K"1), respectively, the right wing can be developed to an exponential series that, for cases where the value of the exponent is very small, P I =
£FZJCJ(»)-]>>Z/CJ(^
(Eq.29)
The first term in the right wing is the sum of all the charges far from the cation. Since a neutral solution, far from the cation should be in equilibrium, with equality between negative and positive charges- this term is zero by definition. Introducing to (29) the definition of the ionic strength / a s [1]:
I =- Y c i Z j 2 1*-'
(Eq.30)
' '
will lead to a specific expression for Debye length:
r
°=f¥i
(Eq3l)
as a simple example we may calculate the Debye length for a solution containing 0.005 M CaCl2. As a first step, we will calculate the ionic strength using Eq.(30): I = -YCz2=-(o.OO5-22+O.OM2)=O.O15—
(Eq. 32)
172
G. Rytwo
and then using Eq.(31), and introducing the transformation from L to m3: C2 T 298K 8.85xlO~12 — - 8 . 3 1 — rD= p *™!5 = 2.5e-9m = 2.5nm 1U.964002-^-0.015—-1000-^ I mole L m I
33
)
The result is that in such solution, at a distance of 2.5 nm of the cation, the shielded potential will be almost one third (1/e) of the value evaluated by the Coulomb potential, reducing considerably the electrostatic interactions. This will of course lead to an activity (or "effective concentration") lower than the concentration itself, thus- y
(Eq.34)
where 5p has a small positive value, since the surplus is of cations. We will recall that the value of p7- is negative, since in the general case, the atmosphere that cation i forms causes in its close surrounding a surplus of anions. Thus, Eq. (27) became:
r
= I
E—cp
(Eq. 35)
Due to the differences in the signs, |p;+5p| will always be smaller than |pj|. Dividing to a smaller number will definitively lead to larger numbers of rD. Large values
Adsorption/Desorption of Ions on Clay Surfaces
173
of Debye length mean less influence of the surrounding on cation / chemical behavior, thus making activity coefficient closer to unity. A simpler way to visualize this effect, would be by recalling that very close to the surface, at the place where the concentration and the activity coefficients are needed, there is only a very low concentration of anions. Thus, the shielding effect will be minimized, leading to the conclusion that for the binding reactions, we may use the concentrations and rely on the fact that for all ions (and even for the sites, following the same logic) y=l. 2.8 - Rate of formation of a Z-valent site In Eqs. (3) and (4) we presented the relationship between the monovalent sites and the Zi charged sites, [Pzi~] = ™ vL. . Kinetic consideration might lead to interesting results for the rate of formation of such multiple valency sites. It is clear that in order to form such site, Zi monovalent sites are needed: Zi[P]o[Pzi"]
(Eq.36)
where ka and ka' will denote the forward and reverse rate constants. The Z-valent site would afterwards participate in the reaction presented in Eq. (3), with forward and reverse rate constants of kb and kb'. Combining reactions (3) and (36), we can write down the net rates of changes in the concentrations of the products:
^ P = -MP"f+ka[P-] ^
l dt
from
= -k b '[P z -][Ci(O) Zl+ ] + k b [PCi]
Eq(36)
^
^
from Eq(3)
At equilibrium, the net rates of formation equal to zero, leading to:
K
A
=
K_=P^ k,' [ p - f
-
K
'
K
[PCi] ~ z k b ' [P -][Ci(0) Zl+ ]
(Eq-38)
where KA represents the equilibrium coefficient for the formation of a multiple-charge site. Combining its definition with the fact that [pzi~ ] = ™ y
K -
1
A
[p]Z,-,zi
forZi = l ^ K A =
rr = l [P-] M 1
forZi = 2 - > K , =
\— = —— [p-]2-'2 2[P"]
A
forZi = 3 ^ K . =
l
-r—=——.T [P"] 31 3 3[P"] 2
yields:
i-c ir>\ (Eq. 39)
174
G. Rytwo
Since the concentration of empty sites will be in all the relevant cases considerably less than 1 M, thus [P"]«l, that means that KA>1. Evaluating the free energy of such reaction, presented in Eq.(36), and using AG = -RTlnK, will lead to the conclusion that the formation of multiple charged sites is thermodynamically favored, since AG<0. 2.9 -First derivative ofy(x). For the sake of consistency with previous publications [4,11,15,23], this auxiliary section will be presented in un-rationalized electrostatic units. The first derivative oiy(x) is needed in order to evaluate the distribution of ions in the diffuse double layer, using Taylor's approximation and the detailed described at Section 2.4. To get Eq.(15) we need to go back to the basic Maxwell's equations. Unrationalized notation of Poissons' law, which defines the rate of change of the potential with the distance from a charged infinite surface, is: a 2 (p_47i P (x)
dx2
(Eq
40)
s
where s is the permittivity of the matrix, and p(x) stands for the charge density per unit volume at distance x from the charged surface. This charge density depends on the amount of charges per unit volume in the solution, i.e.: p(x) = ^ e n , ( x ) - Z i
(Eq.41)
where H/xjrepresents the amount of ions / at distance x from the surface. This summation has to be made over all the cations and the anions. Using Boltzmann's distribution and Eq.(41) into Eq.(40), we obtain: 3 ^ = i 5 V eZin (oo)y(x)Zl dx e '
(Eq. 42)
differentiating Eq.(9) for x will yield to: M0=_^.e-^T.^E=_^y(x).^P Sx kT dx kT dx
(Eq.43)
we can multiply Eq.(42) by 2dq>/dx, and recall that
scprs^l^recpf
s ^ y
,
(Eq 42b)
isolating 9(p/9x from Eq.(43), and substituting in the right wing of Eq.(42b) will yield:
![M=-^!ZeZin,(oo)y(x)dxldxj
se dx*—1
(Eq.44)
Adsorption/Desorption of Ions on Clay Surfaces
175
we may integrate once, from x up to ao. The left wing becomes simply (9(p/3x)2. p o r t jj e right side it will be useful to recall from the integration rules that
/dx while the minus sign inverts the order of the integration, and knowing that the boundary conditions imply that the potential and the derivative of the potential zeroes at infinite distance, thus y(°o)=l and <5y/5x(oo)=0 , then
2
21 1
[f] —!".^^) - )
(Eq45)
-
By taking the square root of Eq.(45), and substituting dcp/dx using dy/dx and Eq.(43) we receive an explicit expression for dy/dx, and transforming from amount of ions to concentration by means of the Avogadro number, we obtain Eq. (15). 3 - Examples In order to demonstrate the features of the model, two simple examples are presented: 1. Ca2+/Na+ exchange on a SWy-1 montmorillonite, compared with the simple model of Gapon coefficient. 2. Increased adsorption of an organic monovalent cation at high ionic strength. 3.1 -Na+/Ca++ exchange over a montmorillonite clay The simplest model used in soils for the cation exchange reaction between monovalents as Na+ or K+ and divalent cations as Ca4^ or Mg*"1" is based on the Gapon equation, that states [46]: K
G
r
= ^ SAR
(Eq.46)
where KG is the Gapon coefficient, SAR is the sodium adsorption ration, defined as {Na+}/{Ca++}05, where the curled brackets indicate concentrations in mM units, and ESR is the exchangeable sodium ratio, and is defined as the amount of sites bound to Na+ divided by the amount of sites bound to Ca**. This model is widely used to express cation selectivity, and it indicates the irrigation water quality, and the danger of "sodification" (increasing the amount of sodium in the exchange suite of the soil) when irrigating for a long period using water with high values of SAR. For the exchange sodium-calcium, the average value of KGNa/Ca= 0.0145 was determined for a wide range of California soils, with a relative amounts of montmorillonite. We tested the prediction of the model for a suspension containing 0.5 g L"1 of SWy-1 montmorillonite. This montmorillonite was widely studied, and we took the relevant parameters and the binding coefficients from previous studies [47]: CEC=0.796 molec kg"1, SSA=8xlO5 m2 kg"1, KNa=l M'1, and Kca=4, whereas the assumption was that neither Na nor Ca form charged complexes on the surface of the platelet (thus,
176
G. Rytwo
K.Na = Kca = 0 ) We performed calculations for several cases in two systems: (a)
(b)
A total normality of 0.01 N of cations and anions, ranging from total concentrations of 9.6 mM Na+(C1O4)" and 0.2 mM Ca++(C1O4)"2, and up to 0.4 mM Na+(C1O4)" and 4.8 mM Ca++(C1O4)"2. A total normality of 0.05 N of cations and anions, ranging from total concentrations of 9.6 mM Na+(C1O4)' and 0.2 mM Ca++(C1O4)"2, and up to 0.4 mM Na+(C1O4)" and 4.8 mM Ca^.
We used perchlorate anion, (C1O4)~, since previous studies [48] show that this anion does not tend to form complexes with cations. Table 3 concentrates the results of the calculations. It can be seen that the values obtained for K o by evaluating SAR and ESR from the model calculations, are for most cases close to the value previously presented from the literature. The average value, for a total normality of 10 mN, if we exclude the first case, is 0.0148+0.0086. For the higher normality, the average KG=0.0139+0.0038. Table 3 - Model calculations for the ion exchange reaction between NaVCa"1"1". The calculations were performed for a SWy-1 montmorillonite suspension, at 0.5 g clay L1. Ca
Na Total
C(»)
PCi
total
C(co)
PCi
(mM)
(mM)
(-mV)
Width of double layer (nm) (nm)
Ko
9.6
6.179
1.543
0.2
0.003
0.161
122.65
25.18
0.0406
9.0
6.162
1.194
0.5
0.010
0.389
117.49
24.67
0.0254
8.0
6.080
0708
1.0
0.049
0.718
106.67
23.53
0.0179
7.0
5.811
0.389
1.5
0.179
0.942
94.33
19.69
0.0151
6.0
5.255
0.229
2.0
0.454
1.057
84.41
16.55
0.0139
5.0
4.509
0.148
2.5
0.826
1.116
77.73
14.42
0.0134
4.0
3.672
0.099
3.0
1.245
1.152
73.05
12.63
0.0131
3.0
2.786
0.065
3.5
1.688
1.177
69.54
11.49
0.0129
2.0
1.873
0.039
4.0
2.146
1.197
66.77
10.33
0.0127
1.0
0.943
0.018
4.5
2.612
1.212
64.48
9.85
0.0126
0.2
0.189
0.003
4.9
2.990
1.223
62.91
9.40
0.0125
48.0
45.391
1.246
1.0
0.504
0.388
67.88
8.15
0.0251
45.0
43.155
0.835
2.5
1.613
0.664
60.80
6.80
0.0185
40.0
38.770
0.541
5.0
3.799
0.866
54.02
5.63
0.0157
35.0
34.110
0.390
7.5
6.126
0.972
49.81
4.93
0.0146
30.0
29.340
0.291
10.0
8.510
1.041
46.78
4.63
0.0139
25.0
24.510
0.218
12.5
10.925
1.092
44.43
4.11
0.0134
20.0
19.645
0.159
15.0
13.359
1.133
42.50
3.86
0.0131
15.0
14.755
0.111
17.5
15.804
1.167
40.88
3.72
0.0128
10.0
9.848
0.069
20.0
18.259
1.196
39.47
3.51
0.0126
5.0
4.929
0.033
22.5
20.719
1.221
38.24
3.26
0.0124
1.0
0.986
0.006
24.5
22.691
1.239
37.34
3.10
0.0123
Adsorption/Desorption of Ions on Clay Surfaces
111
The use of the Gapon approach has been criticized as potentially in error due to effects of ionic strength and complexation on activity of the ions [49] (Sposito and Mattigod, 1977). Levy et al [50] studied exchange in kaolinitic soils, and reported a wide range of values for KG. However, it is interesting to note the very good fit obtained by the model with the average value reported for California smectitic soils. Figure 5 shows the usual representation for such kind of exchange experiments: the fraction of sites bound to Ca++, is presented as a function of the fraction of Ca++ in the equilibrium solution. The figure compares the "ideal" behavior, assuming KQ=0.0145, with the model calculations. It can be seen that the fit between calculated and "ideal" values is very good (R2=0.994). It should be emphasized that these results were obtained without making any calibration whatsoever, and only from values and parameters obtained from the literature.
Figure 5 - The equivalent fraction of Ca++ bound to the surface, as a function of the equivalent fraction in solution, in a virtual exchange experiment between Ca and Na. The lines represent "ideal" behavior (Ka=0.0145), and the points represent model calculations. The total normality is a sum of the total cations concentration. Values of the negative potential at the surface, and thickness of the double layer are presented in Table 3. It is obvious that the potential at the surface is negative, since it arises from isomorphic substitution in the lattice, forming a negatively charged clay surface. It can be seen from both sample studies, that when the same normality is composed mainly by divalent cations, the absolute potential decreases to almost half the value of the calculated when almost all cations are monovalent: At a normality of 10
178
G. Rytwo
mN, the potential when most of the cations are monovalent is |-122| mV, decreasing to about |-63| mV when most of the cations are divalent. At a 50 mN normality, values change from |-68| to |-37| mV. Thickness of the double layer is taken as the distance up to the point where the potential zeroes, thus the amount of negative charges in solution equals the amount of positive charges. It can be seen that for the 10 mN normality the double layer width decreases from 25 to 9 nm. In the 50 mN case, the decrease is from 8 to 3 nm.
Figure 6 - Changes in the absolute value of the potential (left y-axis) and the concentration (right y-axis) as a function of the distance from the surface. The lines were calculated by the model for a system containing 5 g L~' SWy-1 clay, 5 mM total NaClO4, and2.5mMCa(ClO4)2. Such change can explain the flocculation effects observed at the same total ion concentration, when divalent cations are used instead of monovalent. Such effect emphasizes the potential hazard to agricultural soils, when irrigating with water containing high amounts of monovalent ions (high SAR). A broad double layer, influenced by Na+ cations, might cause dispersion of the soil particles and destruction of the structure, leading to formation of impermeable crust and increased runoff. Additional information that can be obtained from the model, is the distribution of the ions as a function of the distance. Figure 6 shows the distribution of the ions and the potential as a function of the distance from the surface. The specific figure shows the distribution for equal normal total concentrations of Na+ and Ca++ ({Na+}={Ca++}=5 mN), and 5 g L"1 of SWy-1 clay. It can be observed (and confirmed in Table 3) that the potential at the surface of the clay is about |78 mV|. Molar concentrations of divalent cations at the surface are almost three fold larger than monovalent cations. Since the concentration relevant for the binding is near the surface (Eq.(3)), it is obvious that even at similar binding coefficients- there will be a preferential adsorption of divalent cations. In the worksheet version of the model, a graph like Fig.6, visualizing the potential and the concentration of the different ions as a function of the distance is simultaneously
Adsorption/Desorption of Ions on Clay Surfaces
179
prepared, giving a picture of the situation in the double layer and near the surface. Such picture can be helpful to describe other colloidal effects, for example, flocculation at higher ionic strengths, or increased adsorbed amounts of higher valence cations. 3.2 - Adsorption of an organic monovalent dye, upon increasing ionic strength. It is known that some monovalent organic cations adsorb to clays at very large amounts [39]. Monovalent organic dyes as methylene blue, crystal violet, acriflavin, etc., were reported to adsorb almost irreversibly in a wide range of pHs [51], and cannot be desorbed even at large concentrations of inorganic cations. In several cases, adsorption of the organic cations even increased upon adding high concentrations of inorganic salts [15]. Such increase was in some cases explained as precipitation of the organic dye due to high ionic strengths. However, such precipitation was not observed when high ionic strengths were added to similar concentrations of dyes, without clay being presented in the system. Furthermore- in studies using pillared clays [52] part of the adsorbed dye molecules were released when the system was diluted with distilled water, but were not released from the complex upon diluting the organoclay on large volumes of high ionic strength solutions, indicating that the effect is electrostatic, and governed by the ionic strength and not by the concentration of the dye directly. These results can not be explained by simple exchange considerations: whatever the binding coefficients are, it is obvious at first sight that increasing 100 fold the concentration might increase the competition to the binding, and by that some decrease in the adsorption of the dye, and increase in the adsorption of the competing cation should be observed. Margulies et al [15] showed that an adsorption model can account for such effect, upon allowing noncoulombic interactions of large organic molecules (Eq.(7)). The explanation to the effect lies in the fact that when the organic cation is adsorbed at amounts larger than the CEC of the clay, charge reversal occurs. Thus, the surface of the clay, with the organic cations adsorbed to it in the Stern layer, becomes positively charged. At this stage, cations are repelled from the surface, and the double layer contains a surplus of anions and a deficit in cations when compared with the equilibrium solution. Table 4 - Model calculations of the surface potential, the amounts of organic cation (OC) adsorbed, and its concentration in the equilibrium solution. The system contained 0.5 g L-l SWy-1 clay, 0.2 mM total Ca(ClO4)2, 1 mM of a monovalent organic cation, and concentrations of NaClO4 ranged from 1 mM to 250 mM „ . .. „ Concentration of T r wi Na[mM] 1 5 25 50 100 250
n (On L m vJ Vo
15.2 13.5 10.9 9.5 8.1 6.3
A i mn Amount of OC , , , adsorbed ro/o0fCECl J
Residual OC ... concentration at .... . equilibrium solution [mM]
146.4% 162.1% 179.7% 185.9% 190.8% 195.1%
0.417 0.355 0.285 0.260 0.241 0.224
Table 4 shows the surface potential, equilibrium concentration of organic cation, and amounts of organic cation adsorbed for a simulated experiment performed with 0.5 g L"1 SWy-1 clay, 0.2 mM Ca(ClO 4 ) 2 , 1 mM organic monovalent cation (OC),
180
G. Rytwo
and Na(ClO4) concentrations ranging from 1 to 250 mM. A single monovalent anion (C1O4) was considered, the binding coefficients of inorganic cations were taken as in the previous section, and for the organic cation we took values similar to those reported in the literature for acriflavin or methylene blue (K oc =10 7 M"1, K~~ =106 M"1).
Figure 7 - Concentration of the organic cation as a function of the distance from the surface. The lines were calculated by the model for a system containing 0.5 g L'1 SWy-1 clay, 0.2 mM total Ca(ClO4)2, 1 mM of a monovalent organic cation Concentrations of Na(ClO4) are (a) 1 mM, (b) 5 mM, (c) 25 mM, (d) 50 mM, (e) 100 mM and (f) 250 mM. Inset II is a magnification of the situation very close to the platelet. The potential is positive, since as it can be seen in the table, the amounts adsorbed are higher than the CEC. It can be seen that the increase in the background concentration of Na+ makes the surface potential less positive. Due to that, cations will be repelled less effectively (Eq.(ll)). This feature is demonstrated in Fig. 7, that shows the concentrations of the organic cation as a function of the distance. It can be seen that the lower the concentration of Na+, the higher the concentration of OC in the equilibrium solution. This indicates clearly that with low amounts of Na+ less OC is adsorbed (as it appears in Table 4). The important feature to be seen is in inset II in Fig.7, which represents the concentrations very close to the surface. It can be observed that at the higher Na+ concentration, the amounts of OC near the surface are almost ten fold the amounts at the low background Na+ concentration. Since the binding reactions (Eqs.(3), (5) and (7)) show that the concentration of the complexes depend in any case on the concentration near the surface, it is understandable that at high ionic strengths more complexes will be formed, and more OC will be adsorbed, due to the increase in the concentration near the surface. From Fig.7 we can also understand about the influence of the background concentration of Na+ on the thickness of the Gouy diffuse layer. The diffuse layer ends where the concentration reaches a plateau. It can be seen that at [Na+]= lmM, even at a
Adsorption/Desorption of Ions on Clay Surfaces
181
distance of 30 nm a plateau is not achieved yet (line a). When [Na+]= 5 mM, (line b) a plateau is reached at about 20 nm. At very high ionic strengths, e.g. [Na+]= 250 mM, (line/) the thickness of the Gouy layer is only about 2.5 nm. This is more or less the size of some large organic cations. Since at that point, local effects not considered by such macroscopic model may influence, we may deduce that at very high ionic strengths (>500 mM), the calculations of the model might be erroneous. 4 - Summary This study intends to present an adsorption/desorption model that is on one hand relatively easy to manipulate by any researcher, even without any programming knowledge, but on the other hand comprehensive enough to consider a wide range of changes in the system. The model can consider effectively changes in the concentrations of any of the adsorbing cations, changes in the amount of sorption sites (CEC or clay concentration), changes in temperature or solvent properties, and even visualize the distribution of the different cations around the clay platelets. The model is available for free download at the author's homepage, or can be mailed upon request. 5 - References [I] W. Stumm, J.J. Morgan. The solid-solution interface in: Aquatic chemistry: chemical equilibrium and rates in natural waters, 3rd. ed. John Wiley & Sons, Inc. New York, 1996. [2] D. Sparks. Sorption phenomena on soils, In: Environmental Soil Chemistry, Academic Press, San Diego, CA; 1995. [3] A. Voegelin, V.M. Vulava, F. Kuhnen, R. Kretzschmar. Multicomponent transport of major cations predicted from binary adsorption experiments. Journal of Contaminant Hydrology, 46 (2000) 319. [4] S. Nir, G. Rytwo, T. Undabeytia, T. Pulobesova. Adsorption of organic cations to clays: Experimental results and modeling in: Organo-clay complexes and interactions. S. Yariv and H. Cross (eds), Marcel Dekker Publ. 193-222. (2002). [5] G. Sposito. The surface chemistry of soils. Oxford Univ Pr. New York, 1984. [6] M. Stadler and P.W. Schindler, Clays Clay Miner., 41 (1993) 288. [7] C.P. Huang and W. Stumm, J. Colloid Interface Sci., 43 (1973) 409. [8] J.A. Davis and J.O. Leckie, J. Colloid Interface Sci. 67 (1978) 90. [9] K.F. Hayes and J.O. Leckie, J. Colloid Interface Sci., 115 (1987) 564. [10] J. Westall and H. Hohl, Adv. Coll. Interf. Sci. 12 (1980) 265. II1] S. Nir, C. Newton and D. Papahadjopoulos, Bioenergetics, 5 (1978) 116. [12] S. Nir, J. Colloid Interface Sci. 102 (1984) 313. [13] S. Nir, Soil Sci. Soc. Am. J. 50 (1986) 52. [14] S. Nir, D. Hirsch, J. Navrot and A. Banin, Soil Sci. Soc. Am. J. 50 (1986) 40. [15] L. Margulies, H. Rozen and S. Nir, Clays Clay Miner., 36 (1988) 270. [16] D. Hirsch, S. Nir and A. Banin, Soil Sci. Soc. Am. J. 53 (1989) 716. [17] U. Yermiyahu, S. Nir, G. Ben-Hayyim and U. Kafkafi, J. Membrane Biol., 138 (1994) 55. [18] G. Rytwo, S. Nir and L. Margulies, Soil. Sci. Soc. Am. J. 59 (1995) 554. [19] U. Yermiyahu, G. Rytwo, D. Brauer and T. Kinraide, J. Membrane Biol. 159 (1997) 239.
182
G. Rytwo
[20] G. Rytwo, S. Nir, L. Margulies, B. Casal, J. Merino, E. Ruiz-Hitzky and J.M. Serratosa, Clays Clay Miner. 46 (1998) 340. [21] G. Rytwo, C. Serban and D. Tropp, Appl. Clay Sci. 20/6 (2002) 273. [22] G. Rytwo, S. Nir and L. Margulies, J. Colloid Interface Sci. 181 (1996) 551. [23] T. Undabeytia, S. Nir, G. Rytwo, E. Morillo and C. Maqueda, Environ. Sci. & Technol., 36 (2002) 2677. [24] T. Pulobesova, G. Rytwo, S. Nir, C. Serban and L. Margulies, Clays Clay Miner. 45 (1997) 834. [25] T.B. Kinraide, U. Yermiyahu and G. Rytwo, Plant Physiology, 118 (1998) 505. [26] T. Undabeytia, S. Nir, T. Polubesova, G. Rytwo, E. Morillo and C. Maqueda, Environ. Sci. & Technol., 33 (1999) 864. [27] S. Nir, T. Undabeytia, D. Marcovich, Y. El-Nahhal, T. Polubesova, C. Serban, G. Rytwo, G. Lagaly and B. Rubin, Environ. Sci. & Technol., 34 (2000) 1269. [28] Y.G. Mishael, T. Undabeytia, G. Rytwo, B. Papahadjopoulos-Sternberg, B. Rubin and S. Nir, J. Agric. Food Chem. 50 (2002) 2856. [29] R. Vulkan, U. Yermiyahu, T.B. Kinraide, U. Mingelgrin and G. Rytwo, 7th International Conference on the Biogeochemistry of Trace Elements. Uppsala (eds. G.R. Gobran and N. Lepp), 2 (2003) 142. [30] G. Rytwo, Applied Clay Science, in press (2003). [31] M. Filek, M. Zembala and M. Scezhynska-Hebda, Z. Naturforsch 57c (2002) 696. [32] S. Genet, R. Costalat and J. Burger, Biophys. J. 81 (2001) 2442. [33] U. Singh, G. Uehara,. Electrochemistry of the Double Layer: Principles and applications to soils. In: D. Sparks (Ed), Soil Physical Chemistry, 2nd Edition. CRC Press p. 1-46, 1998. [34] D.L. Sparks, Geoderma 100 (2001)303. [35] P.M. Berstch and J.C. Seaman, Proc. Natl. Acad. Sci. 96 (1999) 3350. [36] J.M. Zachara and J.C. Westall, Chemical modeling of ion adsorption to soils. In: D. Sparks (Ed.), Soil Physical Chemistry, 2nd Edition, CRC Press p. 47-96 1998. [37] T.B. Kinraide, Plant Physiol. 106 (1994) 1583. [38] D. Njus. Membrane electrostatics, in Fundamental Principles in Membrane Biophysics, Department of Biological Sciences Wayne State University, http://sun.science.wayne.edu/~bio669/Chap04.pdf, p. 40 (2000). [39] R.E. Grim, W.H. Allaway and F.L. Cuthbert, J. Am. Chem. Soc, 30 (1947) 137. [40] B.K.G. Theng. Clay Organic Interactions. Adam Hilger Ltd., London UK, 1974. [41] G. Rytwo, T. Undabeytia and S. Nir, 36th annual meeting of the Clay Mineral Society, Purdue University, Lafayette IN, (1999) 88. [42] G. Rytwo. Interactions between organic cations and montmorillonite: adsorption and structural changes. Ph.D. Thesis, Hebrew University of Jerusalem (1994). [43] I. Obi, Y. Ichikawa, T. Kakutani and M. Senda, Plant Cell Physiol. 30 (1989) 439. [44] N.E. Hill. Theoretical treatment of permittivity and loss, in Dielectric Properties and Molecular Behavior, T.M. Sudgen (Ed.) , Van Nostrand Reinhold Co. London, 1968. [45] P.W. Atkins. The Debye-Huckel theory in Physical Chemistry (Fifth edition). Oxford University Press, Oxford, UK, p. A8-A10, 1995. [46] M.E. Sumner, W.P. Miller. Cation exchange capacity and exchange coefficients. In: D. Sparks (Ed), Methods of Soil Analysis, Part 3: Chemical Methods, SSSA Press, Madison, WI p. 1201-1229 (1996). [47] G. Rytwo, A. Banin and S. Nir, Clays Clay Miner. 44 (1996) 276.
Adsorption/Desorption of Ions on Clay Surfaces
183
[48] G. Sposito, K.M. Holtzclaw, L. Charlet, C. Jouany and A.L. Page, Soil Sci. Soc. Am. J., 47(1983)51. [49] G. Sposito and S.V. Mattigod, Soil Sci. Soc. Am. J. 41 (1977) 324. [50] G.J. Levy, H.v.H. van der Watt, I. Shainberg and H.M. du Plessis, Soil Sci. Soc. Am. J. 52 (1988) 1259. [51] D.R. Narine and R.D. Guy, Clays Clay Miner. 29 (1981) 205. [52] Y.G. Mishael, G. Rytwo, S. Nir, M. Crespin, F. Annabi- Bergaya and H. Van Damme, J. Colloid Interface. Sci. 209 (1999) 123.
RAMAN AND INFRARED SPECTROSCOPIC STUDIES OF KAOLINITE SURFACES MODIFIED BY INTERCALATION RAY L. FROST *' and JANOS KRISTOF 2 1
Inorganic Materials Research Program, Queensland University of Technology, GPO Box 2434, Brisbane Queensland 4001 - AUSTRALIA. 2 Department of Analytical Chemistry, University of Veszprem, H8201 Veszprem, PO Box 158-HUNGARY. * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Raman and Infrared Spectroscopic Studies
185
1 - Introduction Kaolinite is described as a 1:1 clay mineral consisting of two layers joined through an apical oxygen. One layer is known as the siloxane layer and consists of silicon tetrahedra joined in an hexagonal array. This layer is coupled to a gibbsite-like layer consisting of octahedral aluminium bonded to four OH units and two oxygen atoms. These layers are joined as sheets to other layers and may form a large set of clay layers known as a kaolinite book. Figure 1 illustrates these kaolinite stacks. In this example there are many kaolinite layers. The layers have a low aspect ratio (thickness/length ratio)
Figure 1 - Scanning electron micrograph of an Australian kaolinite illustrating kaolinite stacks and kaolinite books Individual crystals may be selected. Figure 2 displays SEM images of a low defect kaolinite showing crystals with a high aspect ratio. The kaolinite is from Kiralyhegy, Hungary. Kaolinites may be classified according to their order. A highly ordered kaolinite is a kaolinite with very few defect structures and the stacking is perfect. In Xray crystallography various techniques have been employed to provide some measure of this order. One such technique known as the Hinckley index is based upon the ratio of intensity of peaks in the XRD pattern. Kaolinites with many defects in the stacking are known as disordered kaolinites. In order to understand modification of the kaolinite surfaces through a number of processes, it is important to understand the structure of kaolinite. The kaolinite consists of the siloxane tetrahedra marked in dark black, which are bonded to the hydroxylated aluminium layer (marked in black - big sphere) through the
186
R.L. Frost and J. Kristof
apical oxygen marked in black (small sphere). Oxygens are marked in light gray. Four hydroxyl groups marked in white are divided into two groups; firstly the inner hydroxyl marked as OHl and secondly what is known as the inner surface hydroxyl groups marked as OH2, OH3, OH4 (Fig. 3). The inner hydroxyl points towards the ditrigonal cavity of the siloxane layer. The inner surface hydroxyls point away from the surface and hydrogen bond to the next adjacent siloxane surface.
Figure 2 - Scanning electron micrograph of a low defect kaolinite from Kiralyhegy, Hungary
Figure 3 - Computer generated model of the kaolinite unit cell
Raman and Infrared Spectroscopic Studies
187
2 - Modification of kaolinite surfaces Kaolinite surfaces may be modified in a number of ways. These include (a) mechanochemical activation of the kaolinite (b) modification of the surfaces through intercalation (c) adsorption of molecules on the external surfaces (d) acid or base reaction with the kaolinite layers (e) plasma treatment of the kaolinite (f) thermal treatment of the kaolinite Modification causes disruption of the kaolinite layer stacking resulting in increased disorder and increased defect structures. Chemical reaction with acids or base results in the loss of aluminium from the octahedral layer. This may result in a porous material with catalytic properties. 3 - Vibrational spectroscopy of kaolinites The vibrational spectroscopy of kaolinite may be divided into spectral regions depending on which vibrational modes are being measured. Such spectral regions are: (a) the hydroxyl stretching region [1] (b) the SiO stretching region [2,3] (c) the hydroxyl deformation region [4] (d) the hydroxyl translation region (e) the low wavenumber region [5] (f) the overtone and combination region [6] The vibrational spectra of kaolinite and its polytypes is complex especially to the uninitiated. Farmer has reviewed the structure and vibrational spectroscopy of kaolinite and other layer silicates [7]. Four distinct bands are observed in the infrared spectrum of kaolinite at 3697, 3669, 3652 and 3620 cm"1. The three higher frequency bands (designated v b v2 and v3) are assigned to OH stretching modes of the three inner surface hydroxyl groups. The band at 3620 cm"1 is designated v5 and is assigned to the stretching mode of an inner hydroxyl group. The Raman spectrum in the OH stretching region of kaolinite contains five features. In addition to the four frequencies noted above, an infrared inactive feature is observed at 3685cm"1. This is designated v4 and is observed as a component of an unresolved doublet of medium intensity at 3695/3685 cm"1. The relative intensities of the components of this doublet depend on the orientation of the kaolinite crystal. These two bands are described as the transverse and longitudinal optic modes. In Raman spectra observed with 90° scattering geometry, the incident radiation interacts with crystal vibrations of similar wavelength. If a unit cell vibration develops an oscillating polarisability tensor, two corresponding long wavelength crystal modes exist (a) a longitudinal mode (LO) and a transverse mode (TO). In Raman spectroscopy both the TO and LO modes are active, the LO mode exists at a higher frequency because of induced dipoles. Therefore, the Raman active infrared inactive band at 3685 cm"1 is ascribed to the transverse optic mode. The Raman band at 3692 cm"1 is assigned to the longitudinal optic mode and is of low intensity in the Raman spectrum but is strongly infrared active. The 3669 and 3652 cm"1 bands (v2 and v3) are weak and are described as the out-of-phase vibrational modes of the in-phase vibration observed at 3695 cm"1. The
188
R.L. Frost and J. Kristof
3620 cm"1 band (v5) is strong and sharp. The intensity of the OH stretching bands depends on the defect structure of the kaolinite. In highly ordered kaolinites v 4 has been detected in the infrared spectrum at 3685 cm"1 when band resolution was carried out. A comparison may be made between the vibrations of kaolinite hydroxyls and ammonia. Ammonia has a C3v symmetry and therefore one in phase and two out of phase vibrations. Kaolinite has a pseudo C3v symmetry and therefore also has one in phase and two out of phase vibrations. The kaolinite polytype unit cells contain only four hydroxyl groups, one (OH1) of which lies in the ab plane and the other three (OH2OH4) lie at angles between 65 and 73° to the ab plane [8,9]. The inner surface hydroxyl groups give rise to infrared bands at 3695 cm"1, a band with the transition moment lying perpendicular to the ab plane, and two weaker absorptions at 3650 and 3670 cm"1 with transition moments in the ab plane . Recently theoretical predictions of the vibrational spectra of kaolinite have been attempted [9]. Teppen et al, developed empirical force field models for aluminous clay minerals [10,11]. Additionally, computer simulations using force field calculations to predict the infrared and Raman spectra from the dynamic and structural characteristics of kaolinites have been undertaken. However, whilst these theoretical calculations show a good correlation between the experimental infrared and Raman spectra in the low frequency region, the correlation is not as good in the hydroxylstretching region. It could also be expected that this correlation would be improved by obtaining spectra at liquid nitrogen temperatures. This is not the case. The reason may be attributed to (a) the coupling of the hydroxyl-stretching vibrations of the inner surface hydroxyls and (b) the existence of the transverse-longitudinal splitting. 4 - Modification of kaolinite surfaces through mechanochemical activation Kaolinite surfaces may be altered through the grinding of the mineral [12-14]. Both wet and dry grinding are used. This grinding is often referred to as mechanochemical treatment or mechanochemical activation. Such treatment may result in the reduction of the particle size to submicron size and to even into the nanometre scale.studies of both the wet and dry grinding of kaolinite by Hiroshi Takahashi, were undertaken in the late fifties [15-19]. It was found that two types of material were produced; an amorphous type of alumino-silicate and an aggregated material some of which showed zeolytic properties. Such materials have potential as catalysts for specific reactions. X-ray diffraction, thermal analysis and electron microscopy were the main techniques use to study these mechamochemically activated materials. Importantly the effect of dry grinding produced a material with high surface areas [20-22]. It is this high surface area which offers the potential use of mechanochemically activated kaolinite as a catalyst. The effect of mechanochemical action on kaolinite has been measured by infrared absorption spectroscopy and the results compared with those obtained by X-ray diffraction [23,24]. Indeed mechanochemical activation of kaolinite has been used to explore the disordering of kaolinites [25]. The grinding of kaolinite results in the formation of particles of small size which may be on the nano-scale but which may agglomerate into larger particles [26,27]. Dry grinding is a method of intercalating kaolinite with molecules which otherwise would not readily insert between the kaolinite layers [28,29]. Infrared spectroscopy has been used to follow the changes in the structure of kaolinite with intercalation with group 1 chlorides [30]. Destruction of the kaolinite structure and
Raman and Infrared Spectroscopic Studies
189
formation of amorphous material, solid-state diffusion of atoms, particularly protons, leading to the formation of lattice defects, delamination of the tactoid of kaolinite and enhanced hydration were observed. It has been proposed that water inserts between the kaolinite layers with mechanochemical action and the water bonds to the siloxane surface [31-33]. Deuteration resulted in the exchange of the inner surface hydroxyls of kaolinite [31]. Whilst there have been infrared studies of the intercalation of kaolinite by dry grinding with group 1 halides, few studies of the mechanochemical activation of kaolinite have been forthcoming [34].
Figure 4 - DRIFT spectra of the hydroxyl stretching region of kaolinite ground with equal amounts of quartz for a range of times as shown
Figure 5 - Variation in relative intensity of the kaolinite hydroxyls stretching modes and the water OH stretching modes
190
R.L. Frost and J. Kristof
The XRD patterns of the mechanically ground kaolinite/quartz mixture show the rapid changes in the kaolinite structure during the grinding process. The effect of grinding causes the diminution of the d(001) spacing and after 2 hours of grinding no intensity remains in this peak. The significance of the loss of intensity of the d(001) peak means the stacking between the layers is disrupted and lost. The disorder of the kaolinite has been increased. The mechanochemical treatment has broken the hydrogen bonding between adjacent kaolinite layers. Thus, the kaolinite has been completely delaminated through the mechanical grinding process. After 4 hours of grinding no XRD pattern of the kaolinite is present. This means that the long-range ordering in the layers is disturbed so that there is no regular pattern of atoms, which can cause the diffraction. The mechanochemical treatment not only changes the morphology of the kaolinite particles but also causes a reduction in particle size. This modification of the kaolinite surfaces is reflected in the DRIFT spectra. The destruction of the molecular structure of the ground kaolinite/quartz mixtures as determined by DRIFT spectroscopy is illustrated in Figure 4. This figure shows the loss in intensity of the hydroxyl stretching vibrations as a function of grinding time and the concomitant increase in intensity of OH stretching vibrations, attributed to water OH stretching vibrations in the 3200 to 3550 cm"1 region. This variation in intensity is shown in Figure 5. The dry grinding results in point heating and the mechanochemical activation of kaolinite causes dehydroxylation through local heating. The bands at 3695 (v^ and 3685 (v4) cm"1 are attributed to the longitudinal and transverse optic vibrations. This latter band is intense in Raman spectra of low defect kaolinites but is of low intensity in the infrared spectra and is only determined as a component in the overall band profile. Interestingly the (v4) mode shows a decrease in intensity as the mechanochemical treatment of the kaolinite is taking place. Such a transverse vibration depends on the aspect ratio of the kaolinite crystals i.e. the thickness of the crystals (see Figure 2). Thus, as the size of the kaolinite crystals is reduced, this vibration apparently shifts to lower wavenumbers. The bands observed at 3668 (v2) and 3652 (v3) cm"1 also show a decrease in band position with the length of grinding, and after 1 hour of mechanochemical treatment no intensity is observed in these bands. These bands result from the out-of phase vibrations of the inner surface hydroxyls corresponding to the in-phase vibrations observed at 3695 and 3685 cm"1. This means that the inner surface hydroxyls are no longer behaving in a cooperative vibrational pattern. The mechanochemical treatment causes significant changes in the surface structure at the molecular level. The position of the band attributed to the stretching vibration of the inner hydroxyl at 3619 cm"1 (v5) is not effected by the grinding process even though the intensity decreases. An increase in bandwidth is noted. This may result from the intense localised heating of the kaolinite surfaces during mechanochemical treatment. Figure 6 shows the hydroxyl deformation modes observed at 937 and 914 cm"1 attributed to the inner surface and inner hydroxyls, respectively. In harmony with the decrease in intensity of the hydroxyl stretching vibrations, the decrease in intensity of the kaolinite hydroxyl deformation vibrations is linear with grinding time. The grinding process destroys the hydrogen bonding between the adjacent kaolinite layers, as is observed from the loss of the hydroxyl vibrations. At the same time, an increase in intensity of bands ascribed to the hydroxyl stretching vibrations of water are observed. Figure 6 clearly shows the loss of intensity of the 937 cm"1 band at a grater rate than the
Raman and Infrared Spectroscopic Studies
191
914 cm"1 band. This result is significant as it means that the grinding process results in the loss of the inner surface hydroxyls before the inner hydroxyls. It is proposed that the band at around 3200 cm"1 is coordinated water, which is bonded to the aluminium. What this means is that the hydroxyls on the gibbsite like surface of the kaolinite have been replaced with water coordinated to the surface. Mechanochemical treatment causes intense local heating and sufficient energy is supplied to break the hydroxyl bonds. Three low intensity bands are observed at 3738, 3742, and 3748 cm"1. Such bands are attributable to the hydroxyl stretching vibrations of SiOH, confirming the effects of intense local heating on the kaolinite surfaces by the breaking of the SiO bonds and the replacement with SiOH. Changes in the SiO stretching vibrations are shown in Figure 7.
Figure 6 - DRIFT spectra of the OH deformation modes of mechanochemically activated kaolinite. The bands at 1056 and 1034 cm"1 are attributed to the SiO stretching vibrations of the siloxane layer of the kaolinite. The bands observed at 1196, 1159 and 1103 cm"1 are associated with the SiO stretching vibrations of quartz. The relative intensities remain constant upon grinding. The intensity in the spectral profile at 1113 cnrtncreases upon the mechanochemical treatment. It is proposed that this band is associated with the new surface phase produced upon grinding. The observation of new SiO stretching and bending vibrations at 1113 and 520 cm"1 suggests that the new material has a different surface structure from that of the starting material. The concomitant decrease in the SiO stretching and bending
192
R.L. Frost and J. Kristof
modes of kaolinite support the concept of the synthesis of a new material which has a very different surface structure from that of the untreated kaolinite. Upon mechanochemical treatment of the kaolinite with quartz, significant structural alteration occurred rapidly to form a new material with a significantly modified kaolinite surface of reduced crystal size with a somewhat higher surface area. Two processes occur upon mechanochemical activation of kaolinite: firstly the reduction in particle size which results in the maximization of surface area after two hours of grinding and secondly the agglomeration of the small particles. If this aggregation could be prevented then the potential to produce very high surface areas equivalent to say fuming silica exists.
Figure 7 - DRIFT spectra of the SiO stretching modes of mechanochemically activated kaolinite 5 - Kaolinite adsorbed and intercalated formamide - application of the controlled rate thermal analysis experiment. One of the problems associated with the spectroscopic analysis of the intercalated kaolinites is the uncertainty of the nature of the intercalate. The formamide-intercalated kaolinite may contain both adsorbed water as well as adsorbed formamide. Thus the infrared spectra of the intercalated kaolinites probably determines more than one phase i.e. the formamide-intercalated kaolinite and the formamide adsorbed kaolinite. The technique of CRTA thermal analysis enables the separation of
Raman and Infrared Spectroscopic Studies
193
the adsorbed from the intercalated kaolinite. This is achieved through stopping the experiment between the desorption step and the deintercalation step. Thermal analysis is normally carried out as a dynamic experiment with a constant and continuous heating rate. Such experimentation is not able to determine phase changes, which occur at close temperature intervals. New thermoanalytical techniques, which can separate thermal processes have been developed [35]. Thermal analysis is normally carried out under dynamic conditions with constant heating rate. Under these conditions, however, closely overlapping mass loss stages cannot always be recognized and separated. Under controlled rate conditions, i.e. at a pre-set, low rate of decomposition, sufficient time is provided for slow heat and mass transfer processes to occur. With this technique, some "hidden" reactions, transformations, etc, can be recognized, which are not normally seen with linear-even slow-heating. The method is known as constant rate thermal analysis (CRTA) and depends on the rate of mass loss, such that no heating occurs when the phase change occurs. Such thermoanalytical experiments are known as isothermal TGA or quasi-isothermal TGA. Raman spectroscopy is particularly useful for studying intercalated kaolinites. One of the difficulties in studying both the Raman and infrared spectra of the intercalated kaolinites is the uncertainty of what is actually being measured, since the intercalated kaolinite may also contain formamide adsorbed on the kaolinite surfaces. Thus three phases could be present (a) kaolinite which has not been intercalated and has no adsorbed formamide (b) kaolinite with adsorbed formamide (c) kaolinite which has been intercalated with formamide and which also has adsorbed formamide. It is unlikely that the phase with intercalated formamide and no adsorbed formamide would be observed. The application of CRTA to formamide-intercalated kaolinites helps eliminate the phase in (b). The formamide-intercalated kaolinites are removed from the thermal analysis instrument after the desorption step and before the de-intercalation step. Figure 8 displays the Raman spectrum of formamide-intercalated kaolinite before and after CRTA treatment. No bands are observed in the 1630 cm"1 region and although the water bending modes in the Raman spectrum are weak, no bands are observed in this position in either the infrared and Raman spectrum. This proves that the CRTA technology has removed the adsorbed and intercalated water. The band observed at 1670 cm"1 is attributed to the C=O stretching vibration of the intercalated formamide molecule. The band at 1594 cm"1 is assigned to the NH deformation vibration of a primary amide. Two bands are observed at 1394 and 1319 cm"1, which are intense in the Raman spectrum compared with the infrared spectrum. These bands are assigned to the CH deformation and CH rocking vibrations. An unknown broad band is also observed at 1247 cm"1. A major difference between the infrared and Raman spectra of the CRTA-treated formamide-intercalated kaolinites is in the 950 to 1150 cm"1 region attributed to the SiO stretching vibrations. In the Raman spectrum, the band at 1159 cm"1 is assigned to the stretching vibration of quartz impurity. The two bands observed at 1084 and 1055 cm"1 are the SiO stretching vibrations of kaolinite. Figure 9 shows the Raman spectra of the formamide-intercalated and waterformamide intercalated kaolinites before the CRTA treatment. The spectra show a single intense band (v6) at 3629 cm"1 superimposed upon a broad background. For the formamide-intercalated kaolinite, almost no intensity remains in the bands attributed to the inner surface hydroxyls. Further the band (v5) assigned to the inner hydroxyl and
194
R.L. Frost and J. Kristof
normally at 3620 cm"1 is not observed in the spectrum. The band may be simply a part of the 3629 cm"1 band profile. This band is attributed to the inner surface hydroxyls hydrogen bonded to the formamide molecule.
Figure 8 - Raman spectra of the 950 to 1750 cm'1 region of formamide-intercalated kaolinite.
Figure 9 - Raman spectra of the OH stretching region of formamide-intercalated kaolinites after CRTA treatment.
Raman and Infrared Spectroscopic Studies
195
A broad band is observed at around 3603 cm"1 and may be attributed to the hydroxyl-stretching vibration of water, which is in the interlayer spaces. The difference between intercalating with 100% formamide or with 50%formamide/50%water is minimal. In Figure 9, a small amount of unreacted inner surface hydroxyls are observed as intensity in the (v,) mode. A broad band is observed in the Raman spectra at around 3639 cm"1. This band is not to be confused with the out-of-phase vibration of the inner surface hydroxyls normally observed at 3652 cm"1. This band is ascribed to the hydroxyl stretching vibration of the inner surface hydroxyls which are hydrogen bonded to adsorbed formamide molecules. The bandwidth of the band is ~50 cm"1 is indicative of a wide range of hydrogen bonding of the adsorbed molecules. The Raman spectrum of the formamide-water intercalated kaolinite shows some intensity in the hydroxyl stretching vibration of the inner hydroxyl (v^. A small shoulder on the lower wavenumber side of the (v6) band may be that of the (v5) inner hydroxyl stretching vibration. Figure 10 displays the Raman and infrared spectra of the NH stretching region of the CRTA treated 100% formamide-intercalated kaolinite. Bands are observed at (a) 3475 and 3460 cm"1 (b) 3370 and 3337 cm"1 and are attributed to the NH stretching vibrations of the formamide intercalated to the inner surface hydroxyls. These bands are assigned to the anti-symmetric and symmetric NH stretching modes and make up 84% of the total intensity. Bands are also observed at 3247, 3165 and 3069 cm"1 and are assigned to formamide molecules in the interlamellar space.
Figure 10 - Raman and infrared spectra of the NH stretching region of CRTA treated formamide intercalated kaolinite. 6 - CRTA treated hydrazine intercalated kaolinite The interaction of kaolinite with hydrazine has been known for considerable time. This interaction between the kaolinite surfaces and hydrazine is used to distinguish different types of kaolinite polytypes including kaolinite and halloysite through the different rates of expansion.
196
R.L. Frost and J. Kristof
Figure 11 - DRIFT spectra of the hydroxyl stretching region of CRT A treated hydrazine-intercalated kaolinite treated under dry nitrogen or CRTA treated and heated to 50, 75, 85 °C. Halloysite reacts more slowly with hydrazine and may take up to four hours before expansion occurs. This distinction is related to the different surface morphology of the kaolinite, which is flat, and the halloysite, which may take on different morphologies, but the common one is rolled. Part of this distinction may be attributed to the different ways in which the hydrazine molecule reacts with the kaolinite surfaces. Kaolinite has several surfaces, which may bond with the hydrazine in different mechanisms. Firstly kaolinite has both internal and external surfaces. Hydrazine chiefly adsorbs on these external surfaces. The internal surfaces are composed of a siloxane layer consisting of linked silica tetrahedra and a gibbsite like layer with a hydroxyl surface. The hydrazine may interact with the siloxane layer through hydrogen bonding of the hydrazine NH2 groups to the surface. Hydrazine may also bond to the gibbsite-like surface through the lone pair of electrons on the nitrogen of the hydrazine molecule. One of the difficulties of using hydrazine is its highly hygroscopic nature and the fact that is obtained as a hydrazine hydrate. Previous studies have shown that intercalation is kinetically faster if water is present [36]. This is the case of the intercalation of formamide into kaolinite. Intercalation of hydrazine-hydrate into kaolinite increases the complexity of the molecular interaction between the kaolinite surfaces and the hydrazine [37]. Ledoux and White first reported the expansion of kaolinite from 7.2 A to 10.4 A upon intercalating hydrazine into the kaolinite structure
Raman and Infrared Spectroscopic Studies
197
[36]. Mild heating resulted in deintercalation accompanied with the partial collapse of the structure to 9.4A. Johnston and Stone showed the effect of evacuation on the kaolinite-hydrazine complex with the subsequent collapse of the structure from 10.4 A to 9.6 A [38,39]. Frost et al proposed a new model for hydrazine intercalation based on the insertion of a hydrazine-water unit [37]. The thermal behaviour of hydrazineintercalated kaolinite shows a close similarity to that of the formamide-intercalated complex. In addition to the involvement of water in the intercalation process and in the structure of the complex, hydrazine is also liberated from the intercalated clay in two overlapping stages. The innovative technique of controlled rate thermal analysis allows the possibility to separate the adsorbed and intercalated molecules [40,41].
Figure 12 - Raman spectra of the hydroxyl stretching region of CRTA treated hydrazine-intercalated kaolinite treated under dry nitrogen or heated to 50, 75, 85 °C. The DRIFT and Raman spectra of the hydroxyl stretching region are shown in Figures 11 and 12, respectively. The Raman spectra of the NH stretching region are shown in Figure 13. The figures shows a comparison of the spectra from (a) hydrazineintercalated kaolinite treated with dry nitrogen for 1 hour (b) CRTA treated hydrazineintercalated kaolinite at 50°C (c) CRTA treated hydrazine-intercalated kaolinite at 70°C (d) CRTA treated hydrazine-intercalated kaolinite at 85°C. The DRIFT spectra of untreated kaolinite displays four bands in the hydroxyl-stretching region at 3695 (vi), 3668 (v2) 3652 (v3) and 3619 (v5) cm"1.
198
R.L. Frost and J. Kristof
Figure 13 - Raman spectra of the NH stretching region of CRTA treated hydrazineintercalated kaolinite treated under dry nitrogen or heated to 50, 75, 85 °C. In addition, a band may be resolved at 3684 (v4) cm"1 for the low defect kaolinite. Upon intercalation of the low defect kaolinite with hydrazine, the V] band is diminished in intensity and the v4 and v2 modes are not observed in spectra a to c. In spectrum d, the v 2 band is observed. The significance of this observation means that the cooperative in-phase and out-of-phase vibrations of the hydroxyl stretching modes is no longer present. In addition a new band at 3626 cm"1 attributed to the inner surface hydroxyls of kaolinite hydrogen bonded to the hydrazine is observed. Further for the spectra a to c, a band at 3604 cm'1 is observed. This band is ascribed to the hydroxyl stretching of interlayer water. The band at 3695 cm'1 is of low intensity and represents the inner surface hydroxyls, which are not hydrogen bonded to the hydrazine. The band centred on 3657 cm'1 is a broad background profile. The band at 3626 cm'1 has a band width of 11.9 cm"1 when the hydrazine-intercalated kaolinite is treated under dry nitrogen. The CRTA treatment of the hydrazine-intercalated kaolinite at both 51 and 70°C causes the band to be narrower with a bandwidth of 9.3 cm"1. Heating the sample with CRTA treatment at 85°C also causes a broadening of the band to 11.9 cm"1. This means that the CRTA treatment of the hydrazine-intercalated kaolinite is the equivalent of placing the sample under dry nitrogen for 1 hour. In previous work, we therefore proposed a model based on the formation of psrH2-NH3]+[OH]" units as hydrazine functions as a weak monoacid base forming a monohydrate [37]. The interaction of the hydrazine complex occurs between the
Raman and Infrared Spectroscopic Studies
199
negative charge on the OH group and the inner-surface hydroxyls. This suggests that the band occurs at 3626 cm"1 because the interaction of the [NH2-NH3]+[OH]" unit and the inner-surface hydroxyls of the kaolinite is weak. A second interaction can occur between the hydrogen atoms of the hydrazine and the siloxane layer. Whereas the hydrated part of the hydrazine molecule bonds to the inner-surface hydroxyls, the opposing end of the molecule bounds to the siloxane surface between the amine hydrogen atoms and the oxygen atoms of the siloxane surface. Water is essential to the intercalation of the kaolinite and is intimately involved in the intercalation process. Based on this model, there are two types of NH groups and hence, two sets of bands are observed in the DRIFT spectra. Two sharp bands are observed in the spectra of b and c at 3514 and 3481 cm"1 which are absent in spectra a and d. These bands are absent when the hydrazineintercalated kaolinite is either CRTA treated at 85CC or exposed to dry nitrogen for 1 hour. These two bands are assigned to the hydroxyl stretching vibrations of water molecules involved in the hydrazine-intercalation complex. When the hydrazineintercalated complex is CRTA treated at 70°C, the hydrazine intercalation complex collapses to the 9.6A phase. In the DRIFT spectra of this complex two additional bands are observed at 2885 and 2921 cm"1. These bands are assigned to N-H vibrations of hydrazine involved in strong hydrogen bonding with the kaolinite surface. These bands are not observed in the CRTA treated hydrazine-intercalated complex at 85°C. Some low intensity broad bands are observed in about these positions for the nitrogen dried hydrazine-intercalated complex. In the spectra a to c, we also observe a broad band at 2976cm"1 for the dry nitrogen treated hydrazine-intercalated kaolinite and at 2959 and 2957 cm"1 for the 51 and 70°C CRTA treated hydrazine-intercalated kaolinites. The band is of low intensity and is not observed for the 85°C treated hydrazine-intercalated kaolinite. Thus the results reported in this paper are in agreement with the results of Johnston et al These workers suggest that upon partial collapse of the hydrazine-intercalation complex to 9.5A, that the broad band at 2975 cm"1 is no longer observed but is replaced by a band at 3270 cm"1. We also observe a broad band at 3270 cm*1 for the dry nitrogen treated hydrazine-intercalation complex. The band appears to shift to lower wavenumbers upon CRTA treatment and is observed at 3244, 3266 and 3260 cm"1. However, we observe this broad band for all the CRTA treated intercalation complexes. The 70cC CRTA treated hydrazine intercalated complex results in the collapse of the complex to 9.6A, and at the same time two new bands are observed at 2885 and 2921 cm"1. These two bands are assigned to the NH vibrations of very strongly hydrogen bonded hydrazine molecules. Such increased hydrogen bonding strength would be brought about through the reduction in interlamellar space by 0.80A. Two bands are observed at 3362 and 3356 cm"1 and are attributed to the antisymmetric NH stretching vibrations. The observation of two bands means that two different types of NH2 units are observed. The slight differences in wavenumbers means that the two NH2 units are involved with slightly different hydrogen bond strengths. Several models are possible: one likely model is that one NH2 unit hydrogen bonds to the oxygen of the siloxane layer through the hydrogen and the second NH2 unit bonds to the inner surface hydroxyls of the kaolinite through the lone pair of electrons on the nitrogen. If the assumption is made that the 3362 cm"1 band is due to
200
R.L. Frost and J. Kristof
weaker hydrogen bonding, then this may be attributed to the bonding between the lone pair of electrons on the nitrogen to the inner surface hydroxyl.
Figure 14 - DRIFT spectra of the hydroxyl stretching region of CRTA treated hydrazine-intercalated kaolinite treated under dry nitrogen or CRTA treated and heated to 50, 75, 85 °C. Not only is the hydroxyl-stretching region sensitive to the modification of the kaolinite surface through bonding with hydrazine, but the hydroxyl-deformation region centered around 914 cm"1 is also sensitive to this bonding. Figure 14 displays the DRIFT spectra for the suite of hydrazine-intercalated kaolinites treated under a range of conditions. Untreated kaolinite displays two infrared bands at 930 and 914 cm"1, which are attributed to the hydroxyl deformation modes of the inner surface hydroxyls and the inner hydroxyls. Upon intercalation with hydrazine, two bands are observed at 895 and 914 cm"1. These bands are assigned to the inner surface hydroxyls hydrogen bonded to the hydrazine and to the inner hydroxyls. No intensity was observed in the 930 cm"1 position. This loss of intensity is in harmony with the loss of intensity of the inner surface hydroxyl stretching modes. It is interesting to compare the relative intensities of the two bands at 914 and 895 cm"1. If the relative intensity of the band at 914 cm"1 is normalised to unity
Raman and Infrared Spectroscopic Studies
201
then the relative intensity ratio of the 895 to 914 cm'1 bands becomes 1.44, 0.285, 0.173 and 0.35 for the four hydrazine intercalation systems studied in this research. For the 85°C CRTA treated hydrazine intercalated kaolinite, some intensity in the 930 cm"1 band is observed. Fundamentally the relative intensity of the 895 cm"1 band decreases with increasing CRTA treatment temperature. The ratio is 1.44 for the dry nitrogen treated hydrazine-intercalated kaolinite. This value decreases to 0.285 for the 51°C CRTA treated hydrazine intercalated kaolinite and to 0.173 for the 70°C CRTA treated sample. These results imply that the maximum degree of bonding between the hydrazine and the inner surface hydroxyls occurs for the dry-nitrogen treated sample. The results are in harmony with the conclusions reached from the NH stretching vibrations at 3356 and 3362 cm"1. It was concluded that by CRTA treatment of the hydrazine intercalated kaolinites at 50 and 70°C that increased bonding of the hydrazine to the siloxane surface was observed. A low intensity band is observed at around 954 cm"1. This band is attributed to a out-of-plane HNH wag. Figure 14 also shows the SiO stretching region and also the N-N stretching region. The N-N stretching symmetric stretching band is found at 1126 cm"1 in the infrared spectrum of solid hydrazine, 1098 cm"1 for liquid hydrazine and at 1112 cm" 1 in the Raman spectrum of liquid hydrazine. This band is observed at 1126 cm"1 for the hydrazine intercalated kaolinites and is asymmetric on the high wavenumber side. Two bands may be resolved at 1124 and 1034 cm"1 with bandwidths of 8.9 and 16.4 cm"1 respectively. The observation of two N-N vibrations implies that two different types of hydrazine molecules are present in the intercalation complex. A number of models are envisaged: (a) hydrazine molecule bonded to the siloxane surface and a second molecule bonded to the hydroxyl surface with one end of the hydrazine molecules free from bonding to a surface (b) hydrazine molecule bonded to both surfaces through the two NH2 units and a second hydrazine bonded to the intercalating hydrazine molecule (c) hydrazine interacting with the kaolinite surface through a water molecule and a hydrazine molecule interacting without water. For the dry nitrogen treated hydrazine intercalated kaolinite, the relative intensity of the two bands at 1124 and 1134 cm"1 is 4:1. When the hydrazine-intercalated kaolinite is CRTA treated at 85°C, the ratio is 7:1. Thus in harmony with other results, the treatment with dry nitrogen is similar to that of the CRTA treatment at 85°C. The ratio for the 51°C and 70°C CRTA treated hydrazine-intercalated kaolinites are 3:2. These two intercalation systems have predominantly different phases namely the 10.3A phase for the 51°C CRTA system and 9.5A for the 70°C system. It is concluded that the N-N symmetric stretching vibration does not appear to be influenced by the interlamellar spacing. Controlled rate thermal analysis allows the separation of adsorbed and intercalated hydrazine. CRTA displays the presence of three different types of hydrogen-bonded hydrazine in the intercalation complex (a) adsorbed loosely bonded in the kaolinite structure fully expanded by hydrazine-hydrate. The first type of adsorbed hydrazine (some 0.20 mol hydrazine-hydrate/mol inner surface OH) is liberated between approx. 50 and 70°C. (b) The second type hydrazine is lost between approx. 70 and 85CC, corresponding to a quantity of 0.12-0.15 mol hydrazinehydrate/mol inner surface OH. (c) The third type of hydrazine molecule is lost in the 85-130°C range. The quantity of this reagent is fairly constant (0.29-0.32 mol hydrazine/mol inner surface OH), independently of the conditions of sample
202
R.L. Frost and J. Kristof
preparation (drying). CRTA at 70 °C enables the removal of some hydrazine-water and results in the partial collapse of the hydrazine-intercalated kaolinite structure. The occurrence of the 9.6A band in the partially decomposed complex is due to the presence of hydrazine hydrogen-bonded directly to the inner surface OH groups. The presence of the 10.3 A reflection is explained by the connection of hydrazine molecules to the OH groups through water (i.e. in the form of hydrazine-hydrate). The effects of various CRTA treatments on the hydrazine-intercalation complex can be followed by the changes in the spectra of the hydroxyl stretching region of kaolinite or through the changes in the vibrational modes of the inserting hydrazine and water molecules. The change in the quantity of loosely bonded (surface bonded) hydrazine-hydrate cannot be seen in the FT-IR spectra of the OH stretching range. In contrast, the liberation of the strongly bonded (hydrogen-bonded) reagent can be followed by changes in intensity of the OH stretching bands. In particular the intense band observed at 3628 cm"1 is attributed to the hydrogen bonding of the inner surface hydroxyls to the hydrazine. CRTA treatment at 85°C or through the application of dry nitrogen results in the removal of water from the intercalation system, resulting in a water free hydrazine-intercalation complex. Concomitantly increases in the inner surface hydroxyls at 3695 cm"1 are observed. 7 - Intercalation of mechanochemically activated kaolinite intercalated with potassium acetate In this work kaolinite is mechanochemically treated for 1, 2, 6, 10 hours, intercalated the kaolinite with potassium acetate, dried the modified kaolinite in a desiccator and then exposed the samples to moist air for 0, 1 and 24 hours. The DRIFT spectra of these three experiments are shown in Figures 15 and 16. The DRIFT spectrum of the hydroxyl stretching region of kaolinite shows five important features: Bands are observed at (a) (v^ 3695 cm"1 which is assigned to the hydroxyl stretching of the inner surface hydroxyl (b) (v4) 3684 cm"1, attributed to the transverse optic vibration and is only observed in kaolinite crystals with a high aspect ratio (c) bands (v2) at 3684 cm"1 attributed to the out-of-phase vibration of the inner surface hydroxyls (d) bands (v3) at 3653 cm"1 attributed to the second out-of-phase vibration of the inner surface hydroxyls and (e) bands (v5) at 3619 cm"1 assigned to the inner hydroxyls. The v4 band is not normally observed in infrared spectra but contributes significant intensity in Raman spectra. Upon modification of the kaolinite through mechanochemical treatment such as dry grinding changes to the DRIFT spectra occur. Also if the kaolinite is intercalated with potassium acetate, additional bands (v6) are observed at around 3602 cm"1 which are assigned to the hydroxyl stretching of the inner surface hydroxyls of kaolinite which have been hydrogen bonded to the acetate. The figures clearly show a number of important results: (a) the intensity of the hydroxyl stretching bands decreases in intensity with the length of the grinding time, (b) the phase behaviour of the hydroxyl vibrations disappears with grinding (c) the intensity of the vibrations of the hydroxyl stretching of water increases with time of grinding (d) an additional band is observed at around 3602 cm"1 and is assigned to the kaolinite hydroxyl hydrogen bonded to the acetate. Bands are also observed in the 3625 to 3630 cm"1 region. Such bands may arise from the distortion of the kaolinite lattice. Bands are observed in this position in the spectrum of halloysite. The type of lattice, which is found in halloysite, is the more stable structure, since the aluminium
Raman and Infrared Spectroscopic Studies
203
octahedron is slightly less in size than the silica tetrahedra. It is proposed that the mechanochemical treatment has caused the curvature of the kaolinite layers similar to that which is observed for halloysite.
Figure 15 - DRIFT spectra of the mechanochemically activated kaolinite for 1, 2, 3, 10 hrs and intercalated with potassium acetate and not exposed to air. The decrease in the intensity of the V] and Vj bands is illustrated in Figures 15 and 15. There is an apparent exponential decrease in the intensity of the hydroxyl stretching bands with grinding time. The variation in relative intensity of the 3695, 3553 and 3619 cm"1 bands is shown in Figures 17 and 18. For the inner surface hydroxyl the percentage relative intensity increases from 1 to 2 hours of grinding and then shows a decrease up to ten hours. The effect of exposure to air for 1 and 24 hours appears to increase the relative intensity of the band. Exposure to air caused an increase in intensity of the band assigned to the stretching vibration of the inner surface hydroxyl. This increase may be attributed to two factors (a) moisture has caused the de-intercalation of the kaolinite resulting in the increase in intensity or (b) moisture enabled some of the hydroxyls removed through the mechanochemical treatment to be restored. It is apparent that after 10 hours of grinding and intercalation with potassium acetate, the relative % area of the three sets of data all end finish at the same point. The effect of exposure of the modified kaolinites to moist air for 24 hours is more dramatic. The initial intensity drops from 5.2 to almost zero with the additional one hour of grinding. The uptake of moisture has caused the in-phase out-of-phase behaviour of the inner surface hydroxyls to be
204
R.L. Frost and J. Kristof
removed. Water molecules must apparently react with the kaolinite surface and cause loss of short range order.
Figure 16 - DRIFT spectra of the mechanochemically activated kaolinite for 1, 2, 3, 10 hrs and intercalated with potassium acetate and exposed to air for 1 hour. The variation in the relative intensity of the band of the inner hydroxyl at 3619 cm"1 is shown in Figure 18. For the modified kaolinite not exposed to air the relative intensity shows an increase from 1 to 2 hours of grinding and then deceases to a minimum value at 10 hours. Upon exposure to air for 1 hour, the intensity is in a similar position as for the zero exposure, and then decreases exponentially with grinding time. Exposure to air for 24 hours results in a decrease in intensity with grinding time, which is below that of the other two curves. Just as for the intensity of the inner surface hydroxyl, the intensity of the inner hydroxyl stretching vibration ends up at the same point after 10 hours of grinding. The implication is that the kaolinite surface structure has been so changed that the exposure to moist air has no effect on the molecular structure. An additional band at 3605 cm"1 is observed and is assigned to the hydroxyl stretching vibration of the inner surface hydroxyl hydrogen bonded to the acetate ion. This bonding in moist conditions takes place through a water molecule. This results in an expansion of the kaolinite layers to 13.9A. In the absence of water the expansion occurs to a lesser value. In this case the position of the hydroxyl-stretching vibration of the inner surface hydroxyl is found at higher wavenumbers around 3611 cm"1. This is
Raman and Infrared Spectroscopic Studies
205
due to a weaker hydrogen bond formed between the kaolinite inner surface hydroxyl and the acetate ion.
Figure 17 - Variation in the relative intensity of the 3695 cm band of the mechanochemically activated kaolinite, the kaolinite mechanochemically activated with quartz and the mechanochemically activated kaolinite intercalated with potassium acetate.
Figure 18 - Variation in the relative intensity of the 3619 cm'1 band of the mechanochemically activated kaolinite, the kaolinite mechanochemically activated with quartz and the mechanochemically activated kaolinite intercalated with potassium acetate.
206
R.L. Frost and J. Kristof
When intercalation approaches completeness no intensity remains in the inner surface hydroxyl stretching vibrations at 3695, 3668 and 3653 cm"1. A number of conclusions can be made: (a) the amount of intercalation as measured by the % relative intensity of the 3602 cm"1 band is low (b) the amount of intercalation is highest for the intercalated kaolinite that has not been exposed to air (c) the amount of intercalation decreases significantly with length of grinding time (d) the amount of intercalation decreases remarkably with exposure to moist air. 8 - Deintercalation of hydrazine intercalated kaolinite One excellent method of studying the changes in surfaces properties of kaolinite is through deintercalation. Figure 19 shows the XRD patterns of kaolinite intercalated with hydrazine hydrate after exposure to air.
Figure 19 -X-ray diffraction patterns for the deintercalation of hydrazine-intercalated kaolinite. The kaolinite has been fully intercalated at ambient temperatures, and 100% intercalation is required to make spectroscopic interpretations less complicated. Hydrazine adsorbed on the 1:1 layer surfaces show characteristic 'intercalation' type spectra, although the XRD data may not show a fully expanded phase. The cf(001)value for fully intercalated kaolinite is 10.39 A, which is consistent with 10.4 A observed by others [36,38,42]. Upon exposure to air, the kaolinite deintercalates as the
Raman and Infrared Spectroscopic Studies
207
hydrazine is lost spontaneously. The kaolinite loses stacking order from 50 to 120 h, as indicated by the loss of intensity of the d(00[) reflections. After the 120-h period (Figure 19), the 10.39-A was no longer observed. The intensity of the 7.16-A peak increased only after 200 h. The decrease in intensity of the 10.39-A peak without the consequent formation of the 7.16-A peak indicates that the stacking order is restored only after deintercalation. The intensity loss of the 10.39 A peak is exponential with time suggesting that the deintercalation reaction follows first-order kinetics. These results are consistent with the reverse reaction reported by Johnston and Stone (1990) [38]. In our study, the 7.16-A peak after deintercalation is different from the untreated kaolinite. The peak is broad and asymmetric on the low-angle side with portions extending to 7.39 A. This suggests that some layers remain slightly expanded. This slight expansion may result from the incorporation of H2O between the layers.
Figure 20 - DRIFT spectra of the hydroxyl stretching region of the deintercalation of hydrazine-intercalated kaolinite as a function of time. The DRIFT spectra of the deintercalation of the hydrazine-intercalated kaolinite are shown in Figure 20. The OH-stretching region shows the complete absence of the vi, v2 and v3 bands owing to the breaking of the H-bonding of the inner surface hydroxyls consistent with expansion as indicated by XRD. If kaolinite is not fully intercalated then significant intensity remains in the bands caused by the innersurface hydroxyl groups. An additional band is observed upon intercalation with
208
R.L. Frost and J. Kristof
hydrazine at 3626 cm"1 and this band is assigned to the inner-surface hydroxyls Hbonded to the hydrazine-water complex. The band attributed to the inner hydroxyls at 3620 cm ~l is superposed by the 3626 cm"1 band and appears as a shoulder. Note that the 3626 cm"1 band occurs at 1 bar and 25°C. Upon exposure to air, the occurrence of the vi, v2, and v3 bands show a spectral shift towards their original, lower frequencies upon decomposition. The frequency shift in the partially decomposed complex is due to the formation of an intermediate structure where the inner-surface hydroxyls can temporarily move freely, i.e. without forming H-bonding to the siloxane layer. This agrees with the XRD data where a loss of stacking order is observed. The intensity decrease of the 3626 cm"1 band as a function of time indicates the gradual decomposition of the complex. With the intensity decrease of the 3626-cm'1 band, however, another band occurs at 3599 cm"1, which is attributed to intercalated H2O. Thermal analysis (not shown) implicates that H2O is present in the intercalate. The simultaneous loss of water and hydrazine occurs at 140cC. DRIFT spectroscopy of the kaolinite heated to 200°C for 1 h showed that the band at 3599 cm"1 no longer occurs. Upon heating the deintercalated kaolinite, the DRIFT spectrum resembles that of untreated kaolinite. Flat, broad bands centered around 3550 cm"1 owing to loosely bonded water in the decomposing intercalate are observed. Thus, hydrazine is replaced by H2O upon deintercalation. For kaolinite intercalated with potassium acetate, the H-bonded surface OH groups produce a peak at 3605 cm"1 [43]. Because this H-bonded inner-surface OH band appears at 3626 cm"1 for hydrazine-intercalated kaolinite, rather than at 3605 cm"1 as is observed, for example, in acetate intercalated kaolinites, this indicates a weaker hydrogen bonding interaction. Therefore, we propose that the intercalating hydrazine (in fact, hydrazine hydrate) is hydrogen bonded to the inner-surface hydroxyls via H2O molecules. At zero time, peaks are observed at 3626 and 3620 cm"1 with 70.0 and 28.3% of the total band intensity. Such intensities are near the theoretical, i.e., 25.0% for the inner hydroxyl and 75.0% for the inner-surface hydroxyls. At 3 h, the intercalate decomposes and the kaolinite interlayer hydrogen bonding is reforming, as indicated by the appearance of a band at 3695 cm"1, the loss in intensity of the 3626cm"1 band, and the appearance of another band at 3599 cm"1. Bands are observed at 3699 and 3653 cm"1 with 3.5 and 18.5% intensity. The 3626 cm"1 band intensity decreased to 33.0% and the 3620-cm"1 band remains nearly constant at 27.7%. After 4.5 h, the 3626-cm"1 band decreased in intensity to 13.5% with significant intensity in bands at 3696, 3685, 3673, and 3654 cm"1. Most changes in the DRIFT spectra have taken place by this time. After 22 h and 30 days, the observed changes are smaller. As the 3626-cm"1 band decreases in intensity, the 3599-cm"1 band and the bands relating to the inner-surface hydroxyls increase. Thus, the 3599-cm"1 band is due to H2O bonding to the 1:1 layer surface. Deintercalation of kaolinite as a mechanism for hydrating the 1:1 layer surfaces is possible [44,45]. The incorporation of H2O between the 1:1 layers as intercalated H2O is consistent with the infrared band observed at 3599 cm"1 and the XRD results of a broad, low intensity 001 peak. The Raman spectrum of the hydrazine-intercalated kaolinite consists of one band at 3620 cm"1 with a bandwidth of 5.7 cm^attributed to the inner-hydroxyl group (Figure 21). No band corresponding to the DRIFT band at 3626 cm"1 is observed in the Raman spectrum indicating a Raman inactive and infrared active band. Such a band
Raman and Infrared Spectroscopic Studies
209
occurs when there is a large change in dipole moment and no change in the polarizability of the OH bond. Such bands often occur when H2O is involved. These observations help determine the assignment of the 3626 cm"1 band, because the band is attributed to the hydrogen bonding of the hydrazine-water unit and the inner surface hydroxyls.
Figure 21 - Raman spectra of the hydroxyl stretching region of the deintercalation of hydrazine-intercalated kaolinite as a function of time.
The DRIFT spectra of the NH region of the hydrazine and water OHstretching regions are shown in Figure 22. The band observed at 3301 cm"1 is attributed to symmetric stretching, and the bands at 3356 and 3362 cm"1 to the antisymmetric stretching of the NH vibrations. (Durig et al, [46]). The symmetric stretching band is weak in intensity with 8% relative intensity and a bandwidth of 8.0 cm"1. The bandwidths of the 3362 and the 3356-cm"1 bands are 8.4 and 9.1 cm"1, respectively. The bandwidth of the 3301 cm"1 band is 8.0 cm"1. Bands are observed in the Raman spectra of the amine-stretching region at 3346 cm"1 for pure hydrazine and at 3363 and 3338 cm"1 in the hydrazine intercalated kaolinite (Figure 23). The 3362 cm"1 band is both infrared and Raman active. However, the 3356 cm"1 band is infrared active only. Thus this band is attributed to the antisymmetric stretching frequency of the amine NH vibration of the [-NH3]+unit.
210
R.L. Frost and J. Kristof
Figure 22 - DRIFT spectra of the NH stretching region of the deintercalation of hydrazine-intercalated kaolinite as a function of time.
Two bands observed at slightly different frequencies in the antisymmetric stretching region suggests that there are two types of interaction between the hydrazine and the 1:1 layer surfaces. One interaction may occur between both the lone pairs of electrons of the hydrazine nitrogen coordinated to water. This complex ion then interacts with the inner-surface hydroxyls. The band component analysis reveals four broad bands at 3346, 3285, 3260, and 3199 cm"1 with 16.1, 46.5, 9.5, and 27.9% relative intensity. Upon intercalation of the kaolinite with hydrazine, bands are observed 3363, 3342, 3312, 3304, 3283, 3263, and 3206 cm"1. At time zero and at 30 min, NH stretching bands are observed at -3363, 3342, and 3304 cm"1. At 60 min, an additional band at 3312 cm"1 is only observed after deintercalation. The three bands at 3340, 3286, and 3209 cm"1 are assigned to the normal vibrations of adsorbed hydrazine. The band at 3363 cm"1 is assigned to the antisymmetric stretching vibration. Note that only one band is observed in the Raman spectrum whereas two bands were observed in the DRIFT spectra. Thus, the 3356-cm"1
Raman and Infrared Spectroscopic Studies
211
band in the DRIFT spectrum is infrared active, but Raman inactive. Such a band results from large changes in dipole moment with no or minor changes in polarizability.
Figure 23 - Raman spectra of the NH stretching region of the deintercalation of hydrazine-intercalated kaolinite as a function of time. The band at 3363 cm"1 is common to both the DRIFT and Raman spectra. Bands at 3312 and 3302 cnf'are observed in the Raman spectrum of the symmetric stretching region of the amine, whereas in the DRIFT spectra only one band isobserved. The 3302-cm"1 band is common to both the DRIFT and Raman spectra, whereas the 3312-cm'1 band is only Raman active. The hydrazine bands of the intercalated kaolinite are at higher frequencies than for the hydrazine liquid. A broad band centered on 3200 cm'1 is attributed to H2O associated with the hydrazine, although H2O is difficult to determine in the Raman spectra. The spectra clearly show a decrease in intensity of both H2O and hydrazine bands as progresses. Hydrazine functions as a weak monoacid base forming a monohydrate. We therefore propose a model based on the formation of [NH2-NH3]+[OH]" units. The interaction of the hydrazine complex occurs between the negative charge on the OH group and the inner-surface hydroxyls. This suggests that the band occurs at 3626 cm"1 because the interaction of the [NH2-NH3]+[OH]" unit and the inner-surface hydroxyls of the kaolinite is weak. A second interaction can occur between the hydrogen atoms of the hydrazine and the siloxane layer. Whereas the hydrated part of the hydrazine
212
R.L. Frost and J. Kristof
molecule bonds to the inner-surface hydroxyls, the opposing end of the molecule bounds to the siloxane surface between the amine hydrogen atoms and the oxygen atoms of the siloxane surface. Water is essential to the intercalation of the kaolinite and is intimately involved in the intercalation process. Based on this model, there are two types of NH groups and hence, two sets of bands are observed in the DRIFT spectra. An additional band is observed at 3206 cm"1, relating to strongly bonded or coordinated H2O. Thus the 3206 cm"1 band is attributed to the OH bonds of the H2O, which are hydrogen bonded to the hydrazine in the intercalated complex. Spectroscopic evidence supports the concept of the hydrazine having two different NH 2 groups in the intercalated kaolinite. In the infrared spectra, two antisymmetric vibrations are observed at 3362 and 3356 cm"1 with only one infrared symmetric vibration at 3302 cm"1. In the Raman spectra only one antisymmetric vibration is observed at 3363 cm"1. However two symmetric vibrations are observed at 3312 and 3302 cm'1. The 3302 cm"1 band is common in both the DRIFT and Raman spectra, while the 3312 cm"1 band is only Raman active. The 3312 cm"1 band therefore represents a highly symmetric vibration. It is proposed that the 3312 cm"1 band arises from amine NH stretching of the hydrazine [-NH2] hydrogen bonded to the siloxane layer. The hydrazine molecule is rigid in structure and the hydrogens are not freely rotating. Therefore if the NH 2 forms a symmetric linkage with the siloxane layer, then the other half of the molecule is asymmetric to the gibbsite-like layer. The hydroxyl deformation region shows broad bands around 940 and 970 cm" ', which are attributed to the inner surface hydroxyl deformation modes, in addition to the inner-hydroxyl deformation band at 913 cm"1 (Figure 24). In the DRIFT spectrum of the hydroxyl deformation region of the hydrazine-intercalated kaolinite at zero time, the 913-cm"1 band attributed to the inner-hydroxyl group contains 59.6% of the total band intensity. The remaining intensity in the 895 cm"1 band is attributed to the hydroxyl deformations of weakly hydrogen bonded inner-surface hydroxyls. Additional minor intensity occurs in the 925 and 953-cm"1 bands. Hydroxyl deformation vibrations show little change at 0.5 h, but at 3 h, however, a significant decrease in intensity of the 895-cm"1 band occurs (27.2%). A weak band also occurs at 904 cm"1. At 4.5 h, the intensity of the 895-cm"1 band is 11.6% and the 905-cm"1 band is 6.7%. Significant intensity is found in other hydroxyl deformation modes at 923 and 937 cm"1 with 12.9 and 22.6% intensity. In the spectra at 22, 50, 168 h and 30 days, the intensity of the 923-cm"1 band remains constant within experimental error, whereas the band at -937 cm"1 increases in intensity. The spectrum of the de-intercalated kaolinite after 30 days resembles closely that of the untreated kaolinite. The hydroxyl-stretching region of the fully intercalated kaolinite at zero time showed bands at 3620 and 3626 cm"1. The hydroxyl deformation region also shows bands at 913 and 895 cm"1. Upon deintercalation of the hydrazine-intercalated kaolinite to 3 h, the intensity of the bands at 913 and 895 cm"1 remains essentially constant. At 4.5 h, the inner-surface hydroxyl bands at 3653 and 3696 cm"1 show significant intensity. At this stage, the deformation bands at 923 and 937 cm"1 show increased intensity. At 50 h, hardly any to no intensity remains in either the 895 cm"1 and the 3626 cm"1 band. Thus, the 895-cm"1 band is the hydroxyl deformation band corresponding to the 3626-cm"1 hydroxyl-stretching band. After 50 h the band at 905 cm"1 has 13.9% intensity. Note that there is no intensity in the 962-cm"1 band after 50 h. As further deintercalation occurs, the results after 120 h are similar to the 50-h
Raman and Infrared Spectroscopic Studies
213
spectrum, but at 168 h significant intensity occurs in the 966-cm"1 band. This result corresponds to the XRD patterns where there is no 13.9-A peak remaining, and the 7.2A peak occurs. The 966-cm'1 peak is related to the hydroxyl deformation of the innersurface hydroxyls, which are strongly hydrogen bonded to the hydrazine-water complex.
Figure 24 — DRIFT spectra of the hydroxyl deformation region during the deintercalation ofhydrazine intercalated kaolinite.
This hydroxyl deformation band also corresponds with the increased intensity of the 3599-cm"1 band. Thus the formation of the hydrated kaolinite through deintercalation seems to indicate that the 966-cm"1 band is associated with the interaction of the intercalated water and the Al-OH surface. The observations made in the analysis of the hydroxyl-stretching region are consistent with the results of the hydroxyl deformation region. With deintercalation, bands associated with weakly hydrogen-bonded inner-surface hydroxyl groups decrease in intensity and bands relating to the inner-surface hydroxyl groups hydrogen bonded to the adjacent siloxane layer increase in intensity. The spectrum of the hydroxyl deformation region after 7 days and 30 days closely resembles that of the untreated kaolinite. The only difference
214
R.L. Frost and J. Kristof
is the band observed at 966 cm"1. The band at 895 cm"1 is due to the NH wag and the disappearance of this band at higher temperatures or longer periods shows the loss of hydrazine from the intercalate. 9 - Conclusions Intercalation of kaolinite is a mechanism for the synthesis of new and novel phases. Intercalation followed by deintercalation can reduce the particle size and invent materials with varying porosities. For example intercalation of kaolinite with potassium acetate followed by deintercalation results in the formation of a mesophase-like material. Such materials can be used for cation size selection and are useful for the removal of for example heavy metals from water. Kaolinite surfaces modified by intercalation offers potential materials for catalysis. When kaolinite is mechanochemically treated i.e. ground for varying periods of time, the resultant material also has very different properties to that of the starting material. Such mechanochemically modified kaolinite has a high surface area and also has the potential to act as a catalyst and as a catalyst support. Such a material is also suitable for incorporating into composite materials. Much of the potential applications of intercalated, deintercalated and mechanochemically activated kaolinite remain to be explored. 10 - References [1] U. Johansson, R.L. Frost, W. Forsling and J.T. Kloprogge, in 3rd Australian Conference on Vibrational Spectroscopy, University of Melbourne, Australia, 1998, pp. 168. [2] R. L. Frost and S. J. Van Der Gaast, Clay Miner. 32 (1997) 471. [3] R. L. Frost, T. H. Tran and T. Le, Mikrochim. Acta, Suppl. 14 (1997) 747. [4]. R. L. Frost, Clays Clay Miner. 46 (1998) 280. [5].R. L. Frost and J. T. Kloprogge, Appl. Spectrosc. 53 (1999) 1610. [6].R. L. Frost and U. Johansson, Clays and Clay Minerals 46 (1998) 466. [7].V.C. Fanner, Spectrochimica Acta, Part A: Molecular and Biomolecular Spectroscopy 56A (2000) 927. [8].A.C. Hess and V.R. Saunders, Journal of Physical Chemistry 96 (1992) 4367. [9].D. Bougeard, K.S. Smirnov and E. Geidel, Journal of Physical Chemistry B 104 (2000)9210. [10].B.J. Teppen, PhD Thesis - Quantum mechanical development of potential energy parameters for molecular dynamics simulations of the clay/solution interface (silicates, phyllosilicates) - Univ. of Arkansas, Fayetteville, AR, USA, 1994. [11].B.J. Teppen, K. Rasmussen, P.M. Bertsch, D.M. Miller and L. Schaefer, Journal of Physical Chemistry B 101 (1997) 1579. [12].R.L. Frost, J. Kristof, J.T. Kloprogge and E. Horvath, Langmuir 17 (2001) 4067. [13].R. Groulx, R.L. Frost and Y. Tremblay, in Brit. UK Pat. Appl, (Mitel Corporation, Can.). Gb, 2001, p. 8 pp. [14].E. Mako, R.L. Frost, J. Kristof and E. Horvath, Journal of Colloid and Interface Science 244 (2001) 359. [15].H. Takahashi, Bull. Chem. Soc. Japan 32 (1959) 381. [16].H. Takahashi, Clays, Clay Minerals. Proc. Natl. Conf. Clays, Clay Minerals, 6th, Berkeley (1959) 279.
Raman and Infrared Spectroscopic Studies
215
[17].H. Takahashi, Bull. Chem. Soc. Japan 32 (1959) 235. [18].H. Takahashi, Bull. Chem. Soc. Japan 32 (1959) 245. [19].H. Takahashi, Bull. Chem. Soc. Japan 32 (1959) 252. [20J.S.J. Gregg, T.W. Parker and M.J. Stephens, J. Appl. Chem. (London) 4 (1954) 666. [21].S.J. Gregg, T.W. Parker and M.J. Stephens, Clay Minerals Bull. 2 (1953) 34. [22].S.J. Gregg, K.J. Hill and T.W. Parker, J. Appl. Chem. (London) 4 (1954) 631. [23].E.F. Aglietti, J.M. Porto Lopez and E. Pereira, Int. J. Miner. Process. 16 (1986) 135. [24J.E.F. Aglietti, J.M. Porto Lopez and E. Pereira, Int. J. Miner. Process. 16 (1986) 125. [25].S. Deluca and M. Slaughter, Am. Mineral. 70 (1985) 149. [26].F. Gonzalez Garcia, M.T. Ruiz Abrio and M. Gonzalez Rodriguez, Clay Miner. 26 (1991) 549. [27].F. Gonzalez Garcia, M. Gonzalez Rodriguez, C. Gonzalez Vilchez and M. Raigon Pichardo, Bol. Soc. Esp. Ceram. Vidrio 31 (1992) 297. [28].I.D.R. Mackinnon, P.J.R. Uwins, A.J.E. Yago and J.G. Thompson, Clays Controlling Environ., Proc. Int. Clay Conf, 10th (1995) 196. [29J.K.H. Michaelian, I. Lapides, N. Lahav, S. Yariv and I. Brodsky, J. Colloid Interface Sci. 204 (1998) 389. [30].S. Yariv, Powder Technol. 12 (1975) 131. [31].S. Yariv, Int. J. Trop. Agric. 4 (1986) 310. [32].S. Yariv and S. Shoval, Clays Clay Miner. 24 (1976) 253. [33].S. Yariv, J. Chem. Soc, Faraday Trans. 171 (1975) 674. [34].A.K. Bandopadhyay, D.G. Bharathi, S. Maitra, S.H. Ansari, S. Mitra and R. Sen, Fuel Sci. Technol. 16 (1997) 115. [35].J. Kristof, E. Horvath, R.L. Frost and J.T. Kloprogge, J. Therm. Anal. Calorim. 63 (2001)279. [36].R.L. Ledoux and J.L. White, J. Colloid Interface Sci. 21 (1966) 127. [37J.R.L. Frost, J.T. Kloprogge, J. Kristof and E. Horvath, Clays Clay Miner. 47 (1999) 732. [38].C.T. Johnston and D.A. Stone, Clays Clay Miner. 38 (1990) 121. [39].C.T. Johnston, ACS Symp. Ser. 415 (1989) 432. [40].R.L. Frost, J. Kristof, E. Horvath, W.N. Martens and J.T. Kloprogge, Journal of Colloid and Interface Science 251 (2002) 350. [41].J. Kristof, R.L. Frost, J.T. Kloprogge, E. Horvath and M. Gabor, J. of Thermal Analysis and Calorimetry 56 (1999) 885. [42].C.T. Johnston, D.L. Bish, J. Eckert and L.A. Brown, J. Phys. Chem. B 104 (2000) 8080. [43].R.L. Frost, T.H. Tran and J. Kristof, Clay Miner. 32 (1997) 587. [44].P.M. Costanzo and R.F. Giese, Jr., Clays Clay Miner. 38 (1990) 160. [45].P.M. Costanzo, R.F. Giese, Jr. and C.V. Clemency, Clays Clay Miner. 32 (1984) 29. [46] Durig Jr., S.F. Fush and E.E. Mercer, J. Chem. Phys., 44, 11 (1966) 4238
NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY OF MOLECULES AND IONS AT CLAY SURFACES JEAN GRANDJEAN Universite de Liege - Institut de Chimie B6a - COSM Sart Tilman - B-4000 Liege BELGIUM E-mail.: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych andK.G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Nuclear Magnetic Resonance Spectroscopy
217
1 - Introduction Clays and particularly swelling clays are added to many emulsions designed for industrial applications, such as painting [1] or cosmetic industry [2]. Adsorption of ionic and neutral species onto clays is also of interest for environmental issues [3,4] or enhanced oil recovery [5,6]. Nuclei of solvent, solute molecules, and clay counterions are appropriate probes to characterize the behavior of interfacial species. Molecules and ions adsorbed or intercalated in clays are also studied in the solid state. In the last decade, study of organoammonium-exchanged layered silicates has been boosted by new possible applications [7]. For instance, preparation of new hybrid materials such as polymer/clay nanocomposites, sharing new properties compared to the pure organic compound, constitutes a promising area of research [8,9]. Thus, ions, neutral or charged (macro)molecules can be adsorbed on clay particles, intercalated between clay platelets or blended in clayey minerals. Therefore, nuclear magnetic resonance (NMR) of such diverse systems covers a broad range of NMR techniques. Dispersion of clay materials in solvent is investigated by pulse sequences characteristic of isotropic liquids. Decreasing the amount of the fluid phase leads to preferential orientation of the involved species, and NMR theory is similar to that of crystal liquids. On another hand, the techniques of solid-state NMR are appropriate for dry samples or pastes. Whatever the system may be, NMR studies provide structural and dynamic information at a local and sometimes mesoscopic scale. The next section gives a brief outline of the relevant NMR theory, starting with the solid state and showing simplification brought by mobility increase in the liquid state. The third section deals with studies of natural, synthetic, or modified clay suspensions. NMR results on solid systems are summarized in the last section. 2 - Theory A brief survey of NMR theory is only described here. More detailed aspects, necessary to understand studies on clay systems, are reported in the next sections. The interested reader can also find a more complete description in published books on this matter (for instance [10,11]). 2.1 - Nuclear spin interactions If the mass number and the atomic number of a nucleus are even, the spin quantum number I equals to zero. Such nuclei are not observable by NMR. The other nuclear isotopes have non-zero nuclear spin and possess a magnetic moment proportional to the magnetogyric ratio (also called gyromagnetic ratio). Table 1 shows the properties of some nuclei considered in this chapter. The spectrometer usually supplies two kinds of external field: a strong static magnetic field B z and a radiofrequency field Bi perpendicular to it. The interaction of the spin system and the static field creates a distribution of the spin nuclei among 2 1 + 1 energy levels, according to the Maxwell-Boltzmann law. The rotation of the nucleus changes the orientation of the magnetic moment with respect to the surrounding field, varying the energy of the nucleus. The radio-frequency pulse generates transitions among these energy levels. Internal fields are also produced by the spin system itself. The external magnetic field induces currents in the electron clouds of the molecules forming a small induced field. The nuclear spin senses the sum of the external and induced fields giving rise to measurable shifts in the spin precession frequencies (chemical shift, often an anisotropic interaction CSA). On another hand, each nuclear spin generates a magnetic
218
J. Grandjean
field. Surrounding nuclear spins interact with it. This interaction which is called through-space or direct dipole-dipole coupling (DD) decreases with the inverse cube of the internuclear distance. An indirect dipole-dipole coupling (J-coupling) is mediated through the bonding electrons between the interacting nuclei. Table 1- A selection of nuclear isotopes and their properties Isotope I Abundance
H 2
H 6 Li 7 Li 13 C 14 N 15 N 17 O 19 F 23
Na Mg 27 A1 29 Si 25
31p 39
K
113
Cd 133 Cs
Natural
rads"'T "'xlO 6 99.989% 1/2 0.012% 1 1 7.59% 92.41% 3/2 1/2 1.01% 1 99.63 1/2 0.368% 5/2 0.038% 1/2 100% 3/2 100% 5/2 10% 5/2 100% 1/2 4.683% 1/2 100% 93.258% 3/2 1/2 12.2% 7/2 100%
Magnetogyric ratio moment (fm2) 267.522 41.066 39.371 103.977 67.283 19.338 -27.126 -36.281 251.623 70.808 -16.388 69.763 -53.190 108.394 12.501 -59.610 35.333
Quadrupole
0.262 -0.0808 -4.01 2.044 -2.56 10.4 19.9 14.7
5.85 0.343
For spins-1/2 nuclei, the electric charge distribution is spherical and the magnetic effects are only to be considered. As the charge distribution of higher-spin nuclei does not show such a symmetry, the electric energy of the nucleus depends on its orientation with respect to the rest of the molecule. The most important term results from the interaction of the nuclear quadrupole charge distribution which is expressed by the (electric) quadrupole moment eQ of the nucleus and the electric field gradient eq generated by the environment at the nucleus. Nuclei of spin > 1/2 are termed quadrupolar. For such nuclei, the quadrupolar (Q) interaction is usually dominant. Thus, the NMR of quadrupolar nuclei is a more complicated and richer field than that of spins-1/2 because there are electric as well as magnetic influences on the reorientation of the nuclei. Each internal interaction depends on the orientation of the atomic framework with respect to the external magnetic field. Such interaction is proportional to 3cos29 - 1 in most situations where 6 is the angle between the principal axis of the interaction tensor and the magnetic field B z . A powder consists of very many crystals, all with different orientations. Then, the NMR spectrum is very broad, because the interaction in each crystallite is different. In contrast to the other interactions, anisotropy of the J coupling is very small. In liquids, the interaction terms fluctuate as a result of fast molecular motion giving a motionally averaged value which is zero for the direct dipole-dipole and quadrupolar interactions. On another hand, the indirect dipole-dipole and chemical shift
Nuclear Magnetic Resonance Spectroscopy
219
interactions are reduced to their isotropic values: the J coupling and the (isotropic) chemical shift, respectively. Liquid crystals provide an intermediate situation in which mobility shortens the interaction strength. Table 2 - Summary of the nuclear spin interactions in different media Solid Chemical shift (CSA) Direct dipolar (DD) J-coupling (small) Quadrupolar(I>l/2)
Oriented medium reduced CSA reduced DD. reduced J reduced Q
Liquid isotropic C.S. cancelled isotropic J cancelled
In the solid state, all the possible internal interactions occur simultaneously, providing a broad useless signal except with very small molecules. Rotation of the sample around an axis forming an angle 0 of 54.7° (magic angle: <3cos20 -1> =0) leads to a narrow isotropic signal when the spinning rate is fast compared to the magnitude of the interaction. Partial or total decoupling of the interactions can be monitored by appropriate radio-frequency pulse sequences. Pulse sequences are also used to recover selectively one interaction or to obtain multidimensional spectra. More details on these technical aspects can be found in a recent review [12]. 2.2 - Relaxation Radio-frequency pulses disturb the equilibrium of the spin system. Relaxation is the process by which the equilibrium is regained, through fluctuating interaction of the spin system with the thermal molecular environment. For quadrupolar nuclei, the relative importance of the relaxation mechanisms is usually: Quadrupole » dipole-dipole > CSA. The relaxation process can be followed either from the time evolution of the magnetization along the static field direction (longitudinal or spin-lattice relaxation rate Ri) or from that in the perpendicular plane (transverse or spin-spin relaxation rate R2). The relaxation theory is detailed in several NMR textbooks [for instance, 13,14]. 2.3 -Self-diffusion Self-diffusion is the random translational motion of molecules and ions driven by internal kinetic energy. The self-diffusion coefficient is closely related to the molecular size as expressed by the Stokes-Einstein equation. This parameter can be measured by NMR spectrometers equipped with pulsed magnetic field gradient accessories. This methodology is described in recent reviews [15,16]. 3 - NMR studies of aqueous clay suspensions This section deals mainly with suspensions of 2:1 phyllosilicates. Clay platelets are negatively charged as a result of cation isomorphous substitution either in the octahedral layer [Li(I) for Mg(II) and Mg(II) for Al(III) substitution for hectorite and montmorillonite, respectively] or in the tetrahedral layer [Al(III) for Si(IV)] for saponite or beidellite. This section is focussed only on NMR results although these data are often complemented by other techniques. Appropriate references can be found in two recent reviews [17,18] or in the cited articles. These NMR studies have been arranged according to the mainly used NMR parameter.
220
J. Grandjean
3.1 - Quadrupolar splitting Pure water molecules have no preferential orientation and a single line is detected in its proton, deuteron and 17O NMR spectra. In anisotropic media, if no processes are fast enough to cancel the dipolar (protons) or quadrupolar interactions (deuterons and 17O), the constrained molecular motion induces line splitting. For instance, fast exchange of protons between ordered water molecules reduces the water proton resonance to one single line [19]. Water molecules near the surface of a solid are preferentially oriented as indicated by the splitting (multiplicity = 21) in the 2H NMR spectrum of aqueous clay suspensions. The orientation of the principal axes of the quadrupolar interaction (electric field gradient tensor, more exactly) with respect to the applied magnetic induction B z determines the splitting value. The resulting splitting A is given by [20]: A = 3x<3cos20LD - 1>A/(4I(2I-1)
(Eq. 1)
where A is the residual anisotropy with a value between +1 and - 1 , 9LD is the angle between the z axes of the coordinated systems associated with the director (clay platelets) and with Bz, and % ' s the quadrupolar coupling constant (= e2qQ/h). The residual quadrupolar anisotropy A of water deuterons is [20]: A(2H) = 0.05(3cos2eDM - l)/2 - 0.4W6(sin29DM cos 2<|)DM)/2 (Eq. 2) where 0DM and <|>DM are the polar angles between the director and the molecular frames. As the origin of the quadrupolar interaction is intramolecular, the orientation of the principal axis system defining the spatial extent of the interaction with respect to the molecular frame is unaffected by molecular motion. Fast exchange of water molecules between different environments results in a weighted average splitting: A = Spi A;
(Eq. 3)
where p ; denotes the molar fraction of the water site i (Ep; = 1). For instance, in the twostate model of fast exchange between "free" bulk water and "bound" water molecules at the clay interface, one may write A = Pf A f + P b A b
(Eq.4)
Molecular ordering upon application of a magnetic field is a characteristic property of liquid crystals. Such an effect has been also reported in clay gels formed by suspending various montmorillonites and vermiculites (ca. lg.) in heavy water (1 mL). The splitting of the 2H resonance was detected in a static field B z of 0.587 T[19]. Details on the orientation of interfacial water molecules have been obtained later on more dilute clay suspensions by using an orienting field stronger by one order of magnitude. A linear relationship between water deuteron (or 17O) splitting and the amount of montmorillonite (Wyoming) is observed [21]. Fast exchange of water molecules between the bulk (no splitting) and the clay interface (4) can explain such an observation. In the used aqueous suspensions (a few tens of mg. per mL), the water deuteron and oxygen-17 splittings are ca. 3-4 orders of magnitude lower than the
Nuclear Magnetic Resonance Spectroscopy
221
quadrupolar coupling constant measured in ice [22,23]. Thus, a minute fraction of water molecules (one in 10000, typically) is influenced by the solid. The two experimental plots yield to deuteron and 17O quadrupolar couplings of "bound" water significantly smaller than for ice [21]. Among different explanations, such a reduction can be ascribed to water molecules exchanging between anionic and cationic sites having anisotropy of the quadrupolar interaction (A in (2)) with opposite signs. This is a likely explanation since interfacial water molecules could coordinate exchangeable cations (cationic director) and be associated with the negatively charged clay (anionic director). Accordingly, the nature of interlamellar cations should change the splitting value. Indeed, first studies with alkali-exchanged montmorillonites indicate a decreasing value from K+ to Li+ [24]. The most drastic effect is observed by increasing the Ca2+/Na+ ratio of interlamellar cations: the water deuteron splitting first decreases in magnitude until it vanishes, then increases again in magnitude, but with a sign reversal [25] (figure 1). The two components of the doublet exhibit a differential line broadening, which ensures validity of the observed sign change. An explanation of this differential line width has been published [26]. No sign inversion was detected for clays with a single cation isomorphous substitution either in the octahedral layer or in the tetrahedral layer [27-29]. As montmorillonite from Wyoming shares cation replacement both in the octahedral (70%) and in the tetrahedral layers, the anisotropies A (2) of opposite signs may be associated to the two isomorphous substitutions, respectively. This is supported by 2H NMR spectra of aqueous suspensions of clay mixtures (50:50) which indicate that the "cationic" director prevails with tetrahedral cation replacement, and with montmorillonite from Wyoming when the calcium content is larger than -10% [27,29].
Figure 1 - 2H NMR quadrupolar splitting versus Ca2*/Na+ molar ratio of montmorillonite dispersed in water With rather viscous systems, ordering of clay particles in the magnetic field is a slow process which has been followed from the increase of the splitting versus the residence time in the spectrometer [30]. Water deuteron splitting has also been used to detect the sol-gel transition in clay suspensions as a quadrupolar splitting is observed in the gel state only [31]. Similar experiments have been performed on clay suspensions in aqueous binary mixtures. Cosolvent molecules such as acetonitrile, acetone, dimethylsulfoxide
222
J. Grandjean
or methanol are also oriented by the solid surface. Again, the 2H NMR splitting of the deuterated cosolvent has an opposite sign for the tetrahedrally and octahedrally substituted smectites [29,30], and two plausible orientations of acetonitrile molecules are inferred from these data [30]. However, no splitting has been observed in Laponite suspensions, a commercial synthetic hectorite. The small size (diameter: 30 run, thickness: 1 nm) of isolated particles does not produce sufficient ordering in the magnetic field. A complete understanding of the behavior of interfacial species is provided by suspensions of Li-exchanged saponite from Ballarat in water-acetonitrile-Jj mixtures. Indeed, water deuterons and oxygen-17, acetonitrile deuterons and nitrogen-14, and lithium-7 nuclei exhibit quadrupolar splittings. Interfacial water orientation remains basically unchanged within the 0-50% (v/v) concentration range of the organic solvent since the splitting of water deuterons and oxygen-17 is not significantly affected. There is a strong water ordering in the interfacial region which remains unaffected by the organic solvent [32]. In the same concentration range, the deuteron and nitrogen-14 splittings of acetonitrile-c/j molecules decrease, vanish, and reappear with the opposite sign. Isomorphous cation substitution occurs only in the tetrahedral layer of saponites, and one mean interfacial site is expected from previous studies [27]. Thus, change of the acetonitrile splittings is due to the change of the (3cos2GDM - 1) factor (2) varying through the magic angle (54.7°). Simultaneously, the lithium-7 splitting decreases progressively, and acetonitrile molecules penetrate the water structure with lithium ions moving away from the clay surface (migration from the Stern layer to the Gouy Chapman diffuse layer) [32]. A similar behavior is observed for acetone molecules when the same Li-saponite is suspended in water-acetone-rf<s mixtures. By contrast, in water-methanol-J., mixtures, the 2H splittings of both solvent molecules vary similarly in suspensions of this lithiated saponite. Splitting data and self-diffusion measurements attest to the increasing liquid structuring at the clay surface compared to that of the solution in the absence of clay [33]. The layer charge is expected to influence the behaviour of interfacial species. To verify this, similar studies have been performed with synthetic saponites of variable interlayer charge. As observed with the natural saponite [32], water orientation at the surface of the highly charged clays (0.5, 0.6 and 0.75 charge per half-unit cell) is not perturbed by the concentration of acetonitrile, within the 0-50% (v/v) range. With a less charged clay (0.3 unit charge), water molecules do not interact strongly enough with the solid surface, and their mean orientation varies, leading to a sign reversal of the water deuterons splitting with the increase of the cosolvent content [34]. To summarize, the determination of the quadrupolar splitting provides a quite sensitive probe to follow ordering of species at a clay surface. In particular, a sign reversal can occur when the population of two fastly exchanged interfacial sites of opposite anisotropies A (2) varies or when the mean orientation of the interfacial molecules changes progressively, spanning the zero value of (3cos29DM - 1) (2). Both situations have been observed with the studies in aqueous clay suspensions. In anisotropic medium, the NMR spectrum of spins-3/2 nuclei (Table 1) exhibits a triplet pattern (21 transitions) with the central band (m = 1/2—»-1/2 transition) and each satellite (m = 2 transitions) accounting for 40% and 30% of the total signal intensity, respectively. Such behavior is shown for the 7Li NMR spectrum of suspensions of the Li-saponite (Ballarat) discussed previously [32]. By contrast, the 7Li NMR spectrum of water-rich Laponite suspensions (clay amount < 8% w/w) exhibits a
Nuclear Magnetic Resonance Spectroscopy
223
single super-Lorentzian line. Such a lineshape indicates motional restriction (see relaxation results) but the medium remains isotropic. This different behavior may originate from the more diffuse negative charge brought by octahedral substitution, decreasing lithium ion interaction with the clay surface, and/or from the size of the clay platelets, reducing ordering of clay particles in the high magnetic field. The size of the tetrahedrally substituted synthetic saponites is larger than that of Laponite (typically similar to discs with a diameter of 50 nm instead of 30 nm). Accordingly, small quadrupolar splittings are observed in suspensions of Li-saponites [35]. Local ordering increases with the clay content, as supported by the increase of the 2H and 17O water nuclei and 7Li counterion splittings. The opposite effect was observed for the 7Li splitting of the high charge saponite (0.75). The high charge density can change the mean cation orientation, resulting in splittings of opposite signs (2) [35]. Rather surprisingly, the 23Na NMR spectra of such laponite and saponite suspensions show one single Lorentzian band. The 23Na nucleus has indeed a quadrupolar moment greater than 7Li (Table 1), and sodium cations are known to interact more strongly with the clay surface as a result of their lower hydration energy. Both effects point to greater quadrupolar interaction. Competition experiments between alkali counterions and tetraalkylammonium cations for interaction with Laponite surface have been monitored by 7Li and 23Na NMR. The organic cations release progressively lithium or sodium ions from the clay surface into the bulk. The intensity of the 7Li signal remains constant in the whole concentration range of the organic salt, and the line recovers progressively a single Lorentzian shape as lithium ions are expelled from the clay surface. By contrast, addition of the organic salt increases the 23Na line intensity of Na-laponite suspensions. Assuming the central band (40%) is only detected in the absence of organic cations, the total NMR visibility (100%) is progressively reached by increasing the organic salt content [36]. Likewise, the central transition is only observed in the 23Na NMR spectra of other investigated clay suspensions. Similarly, the low hydration power of Cs+ ions leads to a strong quadrupolar interaction and a loss of NMR visibility of ca. 80%, as found in the I33Cs NMR spectrum of Cs-bentonite suspended in water [37]. Loss of NMR visibility arises from the rather slow recording procedure of "liquid-type" NMR spectrometers. Furthermore, the available frequency range is limited to ca. 100 kHz, and satellites may be outside the observable domain. Using a fast recording spectrometer, all the three transitions have been detected [38]. Thus, local ordering is detected in most water-rich clay suspensions, but macroscopic alignment might be anticipated in dense clay suspensions. The nematic ordering of dense aqueous suspensions (29-52% w/w) of Laponite has been revealed by analysing 7Li and 23Na quadrupolar splittings [39,40]. The 23Na residual splitting results from the anisotropy of the first hydration shell of the sodium counterions polarized by the electric charges of the clay particles. Because of the smaller quadrupole moment of the 7Li nucleus (Table 1), the 23Na splitting is larger than that of 7Li. However, the ratio of their quadrupole moment {ca. 2.6) cannot explain the ratio of the quadrupolar splittings {ca. 7.2) [39]. Clearly, the first hydration sphere of the lithium counterion is less polarised as a result of its higher hydration power. 3.2 - Self-diffusion Water translational motion has been studied by NMR in aqueous suspensions of clays. In synthetic hectorite suspensions ranging from 1.83 to 51.2 wt.% clay, the water self-diffusion coefficient linearly decreases with the increase of the clay content.
224
J. Grandjean
Water molecules might be exchanged fast between the bulk and the clay surface. The observed self-diffusion coefficient Dobs is an averaged value expressed by the equation: Dobs = Pf D f +PbD b
(Eq.5)
where the subscripts have the same meaning as before (Eq. 4). Such a two-sites model can account for this linear relationship. Alternatively, obstruction effect of the clay platelets increases with the solid content, decreasing the water self-diffusion coefficient. Varying the diffusion time shows restricted diffusion applies [41]. Similar data obtained with other clay suspensions have been analysed in terms of only obstruction effects [42,43]. The random walk of water molecules in the gel is disrupted by clay particles. The average pore size is typically smaller than 6 um A later more exhaustive analysis on water-rich clay suspensions suggest that the thickness of the bound water layers in the gels depends on the phyllosilicate nature, varying between 8 and 1.2 nm. Thus, this study confirms that the obstruction effects of bound water layers near the clay surfaces are important in restricting self-diffusion of the unbound water molecules in water-rich smectite gels [44]. The self-diffusion coefficient of water protons has been measured as a function of the Laponite content in the range between 0 and 56.4% w/w [45]. This parameter is not sensitive to the sol-gel transition as found with the Japanese hectorite [41], and the normalized self-diffusion coefficient of water is not linearly related to the solid volume fraction (<&). More interestingly, with <& higher than 0.065, diffusion becomes orientation-dependent, and an anisotropic phase is formed. The diffusion data are explained by considering obstruction of water diffusion by the Laponite particles or by assuming mobility decrease of water molecules interacting (bound) with the clay surface. Both models similarly fit the experimental data [45]. The anisotropy of water self-diffusion is used as a new procedure to detect nematic ordering [46], confirming the analysis of the 23Na quadrupolar data [40]. Self-diffusion measurement has been also used to describe the pore distribution in Fe(III)-doped kaolin clay [47]. In contrast to polymer gels, the investigated clay gels are characterized by much smaller pores, which remain essentially interconnected by the liquid phase [42,43,45,47]. In lithium battery research, polymer gels electrolytes, in which the liquid polymer electrolyte is entrapped, are used to increase their energy density and their safety. Flurohectorite or a synthetic fluoromica has been incorporated into gel electrolytes in order to preserve a porous structure that maximizes the absorption of the liquid electrolyte. Self-diffusion has been used to monitor mobility of cations (7Li), anions (19F) of LiCF3SO3, or solvent molecules ('H). Compared with the free-filler sample, the self-diffusion coefficients are slightly reduced in the presence of fluorohectorite. An increasing concentration of the smectite content results in a decrease of the anion self-diffusion coefficient and lower the activation energy values obtained from Arrhenius plots. Strong and undesirable ion interactions with the mica-filler lead to significant decrease of ion diffusion [48]. 3.3 - Relaxation When relaxation is dynamically averaged out between the different environments, a single magnetisation decay is observed [14]. This is the fast exchange limit and an equation similar to (4) and (5) may be written for a two-sites process:
Nuclear Magnetic Resonance Spectroscopy
Ri = p f R if + pBRib
225
(Eq. 6)
where R (= 1/T) is the relaxation rate and i indicates longitudinal (i =1) or transverse (i = 2) relaxation. The relaxation rates in one site are expressed as [13,14]: R, = T, 1 = K f(coTc)
(Eq. 7)
R2 = T/ 1 = K' f (COTC)
(Eq. 8)
where K and K' are constants dependent on the interaction, and the functions f and f are sums of spectral densities J(mco). For an isotropic diffusional rotation, the spectral density functions are given by: J(mco) = TC/(1 + (mco)2Tc2)
(Eq. 9)
where co is the Larmor frequency of the observed nucleus and TC the correlation time associated with the variation of the interaction tensor. For longitudinal or transverse relaxation, m is equal to 1 and 2, or 0, 1 and 2, respectively. The longitudinal and transverse relaxation rates are equal and proportional to the correlation time under the extreme narrowing condition (fast motion: coxc « 1). Increasing the molecular motion leads to a continuous increase of the transverse relaxation rate, whereas the longitudinal relaxation rate reaches a maximum at COTC « 1 (figure 2).
Figure 2 - Typical plot of f (solid line) and f correlation time (t = TC).
(dashed line) as a function of the
For free solvent molecules, relaxation rates are small, while at the solid surface their isotropic motion becomes constrained and the rates change according to the Figure 2. There have been very few studies up to the present date regarding NMR solvent relaxation in clay suspensions. Sur et al [49] demonstrated that the water-proton
226
J. Grandjean
longitudinal relaxation rates are dominated by magnetic interactions with paramagnetic (Fe(III)) centres of montmorillonite, and not by dynamics of solvent molecules. Thus, hectorite or synthetic clays, which are poorly or non-paramagnetic, have been used to investigate molecular dynamics in clay suspensions. Transverse R2 and longitudinal Ri water relaxation data as a function of temperature and angular frequency (magnetic field) indicate that the dominant relaxation mechanism results from the water protonproton dipolar interaction modulated not by molecular rotation but diffusion of adsorbed water at the clay surface. A strong liquid-solid interaction is responsible for the 2-D diffusion of the adsorbed molecules. At times longer than the residence time, larger than 10~6s, water molecules start to desorb and diffuse in a 3-D space [50]. Variation of the water proton longitudinal relaxation rates in synthetic saponite pastes as a function of the magnetic field (relaxation dispersion curve) shows motion characterized by correlation times greater than 10"5s. The low-frequency longitudinal relaxation dispersion is governed by reorientation mediated by translational displacements. The Levy-walk statistics account for these data when strongly adsorbed molecules perform many desorption/adsorption cycles before escaping to more remote regions [51]. Although some quantitative differences may result from the clay structure (isomorphous substitution in the octahedral (hectorite) or tetrahedral layer (saponite)), these data converge towards a similar overall picture [50,51]. Water proton relaxation data have been also used to complement water self-diffusion measurements reported in the previous section [43,45]. Addition of any solute into the clay dispersion may perturb dynamics of water. Polyethers and non-ionic polymers (polyethylene oxide; polypropylene oxide, and their copolymers) can interact with the clay surface through two types of mechanisms. The first involves ion-dipole interactions between the clay counterions and the polar groups of the solute. The second type of interaction includes hydrogen bonding, either by the direct interaction between the adsorbed polymer and the oxygens (or hydroxyls) of the clay surface, or through a "water bridge" between clay surface and polymer. When a non-ionic polymer is introduced into the dispersion, the polymer segments adsorbed at the interface change the number and mobility of the adsorbed solvent molecules. Since Rib (6) is dependent upon both the water mobility at the interface and the mean residence time near its vicinity, the polymer adsorption alters its value. The transverse relaxation rate R2 of water protons has been measured as a function of the polymer concentration in Na-montmorillonite. The observed enhancement at low polymer content is due to an increase of the bound solvent molecules close to the clay surface. Among the few investigated polymers, the relaxation rate increases with the ethylene oxide chain length [52]. As the adsorbed polymer amount on the montmorillonite surface increases, the micelles start adopting a less extended conformation on the clay particles. It reduces the interaction between the ethylene oxide chain and the clay surface, hence a decrease of the water proton relaxation rate is observed [52]. Instead of following such interaction from the relaxation of solvent molecules, the solute nuclei may be studied too. Thus, the 13C NMR relaxation rates have been used to monitor the interaction of polyethers (non ionic surfactants: H-(OCH2CH2)n-OR) with synthetic Na-saponites dispersed in water. Fast exchange occurs between surfactant molecules adsorbed at the clay interface and in the bulk. Interaction of the solute molecules with the clay surface decreases their motion compared to that in solution, perturbing the observed parameter (6). Here, the extreme narrowing condition prevails (figure 2: left part), and an increase of the longitudinal relaxation rate
Nuclear Magnetic Resonance Spectroscopy
227
corresponds to a reduced mobility of the organic molecules. When the saponite interlayer charge increases, the observed relaxation rate R] decreases and no significant effect is observed with the highest charge clay (0.75 per half-unit cell). Thus, the surfactant interaction raises with the decrease of the clay charge. These data have been supported by the determination of the water deuteron splittings and water self-diffusion coefficients [53]. The number of ethylene oxide units governs the intensity of the interaction [35]. Polyethers are known to complex sodium cations in solution but such interaction can also occur through hydrogen bonds with the clay surface. With the high charge saponite (0.75), interfacial water molecules are strongly bound and polyether molecules do not penetrate the interfacial region. Weaker interaction between the clay and the highly hydrated lithium cation is known. Accordingly, higher 13C longitudinal relaxation rates of dodecyldimethylamine oxide suspensions have been measured with the Li-saponites than with the corresponding Na-saponites [54]. 23Na and 7Li NMR studies have been performed to complement the above 13C data and to verify whether clay counterions are involved in the surfactant interaction [35,53,54]. Relaxation theory of quadrupolar nuclei depends on the spin quantum number which intervenes in the proportionality constant (7,8). Similarly to dipolar nuclei, the longitudinal and the transverse relaxation rates are equal under the extreme narrowing condition. However, the relaxation process is related to the spin quantum number when COTC > 1, and for half-integer quadrupolar nuclei, the relaxation rate is characterized by I + Vi exponentials [55-57]. Thus, the longitudinal and transverse relaxation decays are described by a biexponential law although this behavior is more easily detected for the transverse component. The signal after Fourier transformation shows a bi-Lorentzian lineshape formed by a narrow peak superposing a broad band, accounting for 40% and 60% of the signal intensity, respectively. In anisotropic system, species are preferentially oriented, and the residual quadrupolar interaction gives rise to a triplet pattern, as noted previously. When the splitting remains unresolved, the signal may also exhibit a bi-Lorentzian lineshape. To distinguish between these two situations, a fourpulse sequence which selects double quantum coherences (m = -3/2 -> +1/2; -1/2 -> +3/2) has been introduced [58]. When the two last pulses correspond to a nutation angle of the magnetization equals to 54.7° the isotropic contribution is cancelled whereas an anisotropic environment is revealed by an asymmetric doublet (Figure 3). Interaction of counterions with the clay surface reduces their mobility compared to cations in solution. The 7Li NMR spectrum of Li-Laponite suspensions exhibits one single signal with a bi-Lorentzian lineshape. The double-quantum filtered spectrum shows no signal, indicating no ordering of the ion surroundings [54]. By contrast, the corresponding spectrum of Li-saponite suspensions is characterized by an asymmetric doublet (figure 3) [35]. The stronger interaction with saponites reduces not only lithium ion mobility but also orients preferentially lithium counterions with respect to the clay surface. As noted previously, the 23Na NMR spectrum of Na-clay suspensions recorded with a "liquid state" spectrometer shows usually one single Lorentzian line corresponding to the m -1/2 ->l/2 transition (40% of the total line intensity). The satellite peaks are outside the observed frequency range. Competition experiments between lithium and sodium cations for interacting with the saponite surface result in a drastic reduction of the splitting, and the 23Na NMR spectrum is characterized by a biLorentzian lineshape. After double quantum filtering, the spectrum shows the
228
J. Grandjean
asymmetric doublet (figure 3) characteristic of ordering of the counterion surroundings near the clay surface [59].
Figure 3 - Typical 7Li (23Na) NMR spectrum showing a bi-Lorentzian lineshape (bottom) and their double-quantum filtered spectra in isotropic (middle) and anisotropic surroundings (top) Thus, local ordering of the Li+ (Na+) environment has been shown in saponite suspensions [35,59]. Such an effect has been indirectly detected by inspection of the variation of the 23Na nuclear relaxation rates of Laponite suspensions over a broad range of frequencies (magnetic fields). Indeed, to explain the unexpected effect of the temperature on the relaxation rates, the modulation of the quadrupolar coupling probed by sodium counterions originates from microstructural rather than dynamic effects. These microdomains have approximately the same size as the particle diameter (300 A) [38,40]. The 23Na relaxation rates Ri and R2 increases with the layer charge of saponites suspended in water. In the non-extreme narrowing condition which applies here, the variations of Ri and R> versus the correlation time are opposite (figure 2). Therefore, the enhancement of both relaxation rates is not determined by cation dynamics but results from the increase of the electric field gradient (eq) generated by surroundings at the nucleus site, giving rise to stronger quadrupolar interaction (K = e2qQ/h in (7) and (8)). A stronger interaction of hydrated sodium ions with the clay surface leads to the deformation of the cation environment. The less symmetric ion surroundings enhance the value of the electric field gradient (eq). The interaction of a few non ionic [35,53,54], one zwitterionic and one anionic [59] surfactants with Laponite and saponites of variable interlayer charge has been investigated by relaxation measurements. Although the interaction of the non ionic surfactants through hydrogen bonds of the silicate surface remains possible, the increase of the 23Na relaxation rate in the presence of such detergent molecules indicates the change of the cation surroundings. This supports the complexation of sodium (lithium) cations by the
Nuclear Magnetic Resonance Spectroscopy
229
polyether molecules [35,53,54]. In the zwitterionic and anionic surfactant suspensions, the clay charge modulates the sodium exchange between the interfacial sites and bulk water [59]. In all cases, the intensity of the surfactant interaction with the clay surface decreases with the increase of the saponite charge and the greatest interaction is observed with Laponite. These results corroborate data of the previous section.. 4 - NMR studies of solid clayey materials The first part of this section deals with NMR studies of cations from original or ion-exchanged clays. In the last part of this chapter, intercalation of (macro)molecules between the clay platelets is considered. 4.1 - Interlayer inorganic cations In dried solids or in clay pastes, counterion mobility is reduced, and the corresponding transverse relaxation time is very short. Therefore, their observation requires a solid-state (fast recording) NMR spectrometer. As noted in the introduction, high resolution spectra of solids are obtained by using appropriate techniques to cancel or reduce strong interactions [12]. Dipolar counterions such as 113Cd2+ are rather isolated from other NMR active nuclei and the main interaction is chemical shift anisotropy (CSA). Analysis of the asymmetric resonance provides a CSA interaction of ca. 100 ppm for the dioctahedral montmorillonite compared to ca. 40 ppm for the trioctahedral vermiculite and hectorite. Different orientations of the O-H group at the bottom of the ditrigonal cavity formed by the silicate layers accounts for this observation. Addition of water gives rise to two cationic sites, associated to Cd2+ - water and Cd2+ - clay interactions, respectively. A two- and three-water layers are clearly shown [60]. The CSA interaction can be cancelled by fast spinning of the sample at the magic angle (MAS). Using highly concentrated Cd solutions, the U3Cd MAS NMR spectrum of the resulting Cd-montmorillonite indicate adsorption of hydrated Cd2+ and CdCl+ ions in the interlayer and partially hydrated Cd2+ on the external surface [61]. These studies have been complemented later with Ca-hectorite and Ca-montmorillonite partially exchanged with cadmium cations. In the former smectite with no tetrahedral charge, the hydrated cadmium ions occupy one single site in the centre of the interlayer along the c axis but in montmorillonite, a second site is present in which these ions are displaced closer to the negatively charged basal oxygens. 'H MAS NMR experiments indicate that cadmium cations interact with hydroxyl groups through one solvating water molecules in the first coordination sphere [62]. The hydroxyl NMR bands show a value at ca. 0.5 ppm for the trioctahedral samples and of ca. 2.0 ppm for the dioctahedral ones. In trioctahedral minerals, the OH bond, perpendicular to the layers, points toward the pseudohexagonal cavity, while this bond is nearly parallel to the layer in dioctahedral clays. This latter position of the OH group allows the formation of a hydrogen bond with the apical oxygen of the tetrahedral sheet, shifting the resonance to lower field. As dioctahedral minerals share an average of different hydroxyl orientations, their 'H MAS spectra show broader lines compared to trioctahedral 2:1 phyllosilicates. In contrast to the hydroxyl proton resonance which is not affected by the charge of the interlamellar cation, the water proton signal in the 4-5 ppm range moves downfield as the proton acidity increases with the counterion charge [63]. 23Na and 113 Cd NMR studies have been performed on pastes of Na-and Cd-vermiculites, varying the water content. A downfield (left) shift means a increasing number (0,1 and 2) of water layers between the clay platelets. These results also point to preferential
230
J. Grandjean
orientation of water molecules with respect to sodium or cadmium cation [64]. Similarly, 39K NMR allows discrimination between the mobile highly hydrated and the "fixed" weakly hydrated forms of the alkali ion in several K-silicates [65]. Spectra of hectorite slurries in CsCl solution depend on the temperature. The high and low field 133 Cs signals occurring at temperatures lower than -60°C are assigned to tightly and loosely bound cations, respectively. The first peak (-29 ppm) was associated to cations in the Stern layer and the second one (-8 ppm) to cations in the Gouy diffuse layer surrounded by their hydration shells. Increasing temperature leads first to the signal overlapping, and at higher temperatures, the signals of the two Cs-species undergoing rapid exchange are motionally averaged. Upon heating at 450°C, the spectra of the dehydrated smectites (hectorite, saponite, and montmorillonite) show two peaks, one narrow at ca. -110 ppm and another broad at 30-50 ppm, that were ascribed to caesium cations occupying sites with 12-fold and 9-fold oxygen coordination, respectively. These two sites can be envisaged to be formed from two basal oxygen planes of adjacent layers by the superposition of two ditrigonal holes or from one hexagonal hole and one oxygen triad, respectively [66,67]. Because of their greater hydration energy, Na+ ions exist primarly as outer-sphere complexes. They successfully compete with Cs+ for outer-sphere sites [68]. In the case of Li-montmorillonites heated at 300°C, lithium ions appear to move irreversibly into the clay framework. A multinuclear study indicates that lithium ions do not migrate to vacant octahedral positions but are located at the bottom of the pseudo-hexagonal cavities [69]. In a parallel study, Theng et al [70] have found that the quadrupolar coupling constant of Li increases upon heating, in agreement with the lower symmetry of the distorded ditrigonal cavities of the silicate layer. Upon heating at 300°C under high water pressure, the resulting sample shows a NMR spectrum similar to the original unheated clay, indicating that lithium ions have moved back in the interlamellar space [70]. 2D MAS NMR techniques [12] have been used to correlate interlayer nuclei and structural nuclei in a Na-montmorillonite and an Al-saponite . The ! H - 27A1 heteronuclear correlation spectra of an Al-saponite shows correlation between water protons and octahedral interlamellar aluminum ions (*= 0 ppm) whereas the tetrahedral Al(III) (« 67 ppm) are correlated to the hydroxyl protons. No differentiation are possible between structural octahedral aluminum in montmorillonites and interlamellar aluminum ions [71]. The direct dipolar interaction between two nuclei depends on their internuclear distance. Several double-resonance techniques able to determine this parameter have been published [12]. Thus, a spin-echo double resonance (SEDOR) experiment has been applied to dried Cd-montmorillonite. The 113Cd - 27A1 distance between the counterion and the octahedral atom has been estimated to be < 5A. Such a value implies a migration of Cd2+ ions in the ditrigonal cavities of the clay tetrahedral sheets [72]. With quadrupolar nuclei such as 23Na, the strong quadrupolar interaction induces a line splitting (triplet) of a few MHz. High resolution NMR spectra of quadrupolar nuclei often require twodimensional multiple quantum magic angle spinning (2D MQMAS) technique or appropriate probes to rotate the sample around two directions [12, 73]. Indeed, the MAS technique is unable to reduce significantly this interaction, and the radiofrequency pulse cannot excite homogeneously such a broad frequency range. Then, the central band (m -1/2 —> 1/2 transition) is usually observed under MAS to narrow the line. Second order quadrupolar interaction shifts and broadens the
Nuclear Magnetic Resonance Spectroscopy
231
detected signal, and special techniques are used to eliminate these further perturbations. A recent review deals with all these technical aspects [73]. 23Na 2D 3QMAS NMR spectra of Laponite and synthetic saponites with a layer charge < 0.60 per half-unit cell exhibit one single contour plot, suggesting one mean sodium cation environment (Figure 4).
Figure 4 - 23Na 2D 3QMAS spectrum of the saponite with 0.35 charge per half unit cell. Therefore, isomorphous cation distribution is uniformly distributed within the tetrahedral layer. When the clay charge is greater than 0.65, at least one tetrahedral aluminum is present per ditrigonal cavity. Such a high charge density can provide electrostatic repulsion and at least two distributed signals are shown on the spectrum of saponite (0.75). Although small shifts occur with the clay charge, the dominant sodium site is similar in both dimensions to that of the lower charge clay [74]. The isotropic dimension Fl shows the high resolution spectrum with narrow(s) signal(s). The projections on the two axes are correlated with the isotropic chemical shift and the second order quadrupolar-induced shift which is related to the intensity of the quadrupolar interaction. More details can be found in the literature [73, 74]. The 27A1 MAS NMR spectra of smectites pillared with aluminum polyhydroxy polymers (Ali3) exhibit two signals near 0 and 60 ppm from the 12 Al octahedral surrounding 1 Al tetrahedron, respectively. In pillared hectorite and Laponite, where the location of the charge is mainly octahedral, the clay layers retain their structure even after heating at 350°C, while in pillared beidellite, where the layer charge is mainly located in the tetrahedral sheet, the gallery polyoxyaluminum cation reacts with the tetrahedral layer of the beidellite [75]. Beidellite forms more regularly pillared structures than montmorillonite. A similar mechanism has been suggested for the formation of beidellite pillared with Ga13 and GaAl12- [76] and AlCe- polyoxy cations [77].
232
J. Grandjean
4.2 - Intercalated organic cations The weak natural abundance of nuclei such as I3C or 29Si (Table 1) implies weak NMR sensitivity. This situation can be improved in some way by a cross polarisation (CP) experiment. The rare spins are brought in contact with the 'Hreservoir by adjusting the two radiofrequency B 1 C or si and B 1 H fields. Under the Hartman - Hahn condition (yH B] H = Yc.si B1>c,si), the proton polarisation is transferred to the rare spin nucleus. After the removal of the BiCorSi field, the 13C (or 29Si) signal is detected while pursuing the 'H irradiation. The theoretical gain of sensitivity is proportional to the ratio of the gyromagnetic ratios TH/YCSI- The Hartman - Hahn condition is strictly valid for static spectra but remains a good approximation for spinning rates at the magic angle typically lower than ca. 5kHz. Fast MAS could be very useful in many cases and the above condition is no more fulfilled. Instead, polarisation transfer occurs at different frequencies, symmetric to the position of the Hartman - Hahn condition. As the relevant side bands are narrower than that in the static case, a careful calibration is required to get the maximum effect. Proton polarisation can also be transferred to quadnipolar nuclei. A recent paper deals with all these aspects [78]. During the cross polarisation sequence, the 13C magnetization M increases exponentially with the time constant T C H whereas the decrease of the proton magnetization with the relaxation time in the rotating frame (T]p(H)) leads to smaller cross polarisation reducing the 13C signal intensity. The variation of the carbon magnetization is ruled by the equation: M(t) = Mo [exp(-t/Tlp(H) - exp(-t/TCH)]/ [1 - TCH(Tlp(H))-1] (Eq. 10) where t is the contact time and Mo is the equilibrium magnetization. The cross polarisation is most efficient for static 13C (29Si) - 'H dipolar interaction and mobile carbon groups exhibit greater TCH values. This equation is valid assuming TCH/TI P (H) and TCH/Tlp(C) ratios are small. A full analysis of the plot of the magnetisation as a function of the contact time is required for quantitative measurements. Indeed, the intensity ratio changes drastically with the value of the contact time (figure 5). By contrast the equilibrium magnetization ratio from the two curve fittings is equal to 2.1, very close to the theoretical ratio. All these aspects are summarized in a recent review [79]. 13 C CP MAS NMR experiments are used to probe the molecular structure and dynamics of organic cations intercalated in smectites. One of the first study includes complexes of tetramethyl ammonium (TMA) and hexadecyltrimethylammonium (HDTA) cations with a montmorillonite and a vermiculite that have iron contents of 3.4 and 7.4%, respectively. Rapid motion is detected in these complexes, as seen from the averaging effect of the ! H - 13C heteronuclear dipolar interaction. The cation motion appears to be isotropic and anisotropic for intercalated TMA and HDTA, respectively [80]. The time constant TCH has been used to probe mobility of the dimethyldistearylammonium (DDSA) ion intercalated in montmorillonite. Here, the steric effects seem to be predominant since T C H values < 200 us indicate low amplitude and/or frequency of the CH2 motion. The adsorption of methanol vapor on the organoclay leads to larger values. A decrease of the dipolar interaction between the more distant CH2 groups of two neighbouring chains, brought by methanol insertion, enhances alkyl chain mobility [81].
233
Nuclear Magnetic Resonance Spectroscopy
The conformational heterogeneity and mobility of 1-octadecylamine (ODA) intercalated in montmorillonite has been described by NMR relaxation techniques and 2D WISE NMR (dispersion of proton lines along the 13C chemical shift scale).
Figure 5 - Plot of the signal intensity as a function of the contact time (DHEMHA: solid line: NCH3; dashed line: CH2OH) The proton line widths can directly reflect the mobility of the surfactant chain, supporting information from the TCH values. The main signal in the 35 - 30 ppm range is assigned to the inner CH2 groups of the long alkyl chain. The chemical shift of these groups depends on the conformation of the two y positions: trans - trans, trans - gauche, gauche - trans or gauche - gauche. The signals near 34 and 32 ppm are assigned to the trans and gauche conformations, respectively. The total content of gauche conformation has been calculated to be approximately 18%. The trans conformer is more rigid as indicated by shorter TCH values and broader line widths [82]. Table 3 - Most investigated surfactants Acronym BTA DDSA HEDMHA DHEMHA HDTA OA ODA PA TMA
Cation (CH3CH2)3N+CH2H5C6 (CH3)2N+((CH2)17CH3)2 (HOCH2CH2)N+(CH3)2(CH2),5CH3 (HOCH2CH2)2N+(CH3)(CH2)15CH3 (CH3)3N+(CH2)15CH3 H3N+(CH2)7CH3 H3N+(CH2)17CH3 H3N+(CH2)2H5C6 N+(CH3)4
Reference [88] [81] [87] [87] [80],[86] [90] [82],[85],[86] [88] [80],[91]
234
J. Grandjean
The proximity of organic cations to structural iron leads to a very effective 13C ('H) relaxation and an important line broadening. Indeed, the magnetic moment of the unpaired electron is 658 times larger than that of the proton, inducing a strong nuclear relaxation mechanism that results from direct or indirect electron — nucleus dipolar interaction. This paramagnetic effect in clay systems has been later studied in detail. Among different possible mechanisms, spin exchange between electrons on different ferric ions, leading to fluctuation of the dipolar interaction, is the dominant effect in the shortening of 'H and 13C relaxation times Tj and T2. Extensive line broadening (tiny T2, see figure 2) can give rise to a substantial signal loss. The paramagnetic effect results from statistical variation in local Fe concentration within a pair of clay layers rather than variations in the distance from the clay midplane [83,84]. Dynamics of cationic surfactants intercalated in montmorillonite and silylated montmorillonite have been studied. With the same iron content in the two clays, variation of the relaxation parameters means mobility difference. The larger relaxation times observed with the silylated materials imply a higher degree of surfactant mobility [85]. However, in any case, the paramagnetic effect generates line broadening with a partial loss of the signal intensity and resolution, preventing any analysis along the alkyl chain of the surfactant. Furthermore, the estimation of the paramagnetic effect on the nuclear relaxation process requires several assumptions. Therefore, NMR study of non-paramagnetic clay systems should provide more insight on the behaviour of the organic cations incorporated in clay galleries. Thus, the 13C CP MAS spectrum of the HDTA-Laponite exhibits wellresolved signals. The 'H relaxation times (in the laboratory frame and in the rotatory frame) do not show any significant variation along the long hydrocarbon chain, as spindiffusion evens out the differences in local magnetization during the measuring time. The spin-diffusion rate depends on the proton density which is not uniform in the case of partial exchange of the sodium cations. Thus, 29Si detected ! H relaxation times is described by a biexponential law, corresponding to silicon nuclei near exchanged (high proton density from the close organic cation) and unexchanged sites. In contrast to the *H relaxation time, the 13C relaxation time in the laboratory frame T^C) and the time constant TCH increase regularly from the head group to the terminal methyl group, according to a mobility increase along the hydrocarbon chain. Decrease of the Ri(C) values is characteristic of high mobility (figure 2) [86]. By contrast, the corresponding values obtained with HDTA-saponites are not significantly different whatever the carbon nucleus may be. More rigidity is indeed observed and Ri(C) values are near the maximum of the curve (figure 2). The 13C NMR spectra of HDTA- and ODA-Laponite exhibit much narrower lines than those obtained with the ODA-montmorillonite [82]. Paramagnetic effect (clay characteristics were not reported [82]), mobility reduction from the higher platelet size, higher charge density and/or charge originating from substitution in the tetrahedral layer may account for that observation. As indicated previously, the main signal in the 35 - 30 ppm range is assigned to the inner CH2 groups (C4 - C14) of HDTA. In the spectrum of the pure solid surfactant, the chemical shift at ca. 34 ppm is associated with the all-trans conformation of the alkyl chain (figure 6). Two partly resolved signals are observed in the spectra of HDTA- and ODALaponite, resulting from a mixture of trans (34.4 ppm) and gauche (32.7 ppm) conformations. In contrast to ODA-montmorillonite [82], the gauche conformer is dominant [86]. The same behavior has been obtained with HEDMHA- and DHEMHALaponite (figure 7).
Nuclear Magnetic Resonance Spectroscopy
235
Figure 6 - The all-trans conformation of the long alkyl chain of the DHEMHA iodide
Figure 7 - nC CP MASNMR spectrum of DHEMHA-Laponite The location of cation isomorphous substitution and the charge density influence the properties of the intercalated cation. Thus, the HDTA, HEDMHA and DHEMHA surfactants intercalated in the lowly charged saponites (0.3-0.4 charge per half-unit cell) show a bilayer structure with a high trans conformer content (ca. 70%) of the alkyl chains lying down on the clay surface . For higher charge clays, the mean surface area per cation decreases and the extended more rigid trans conformation of the surfactant cation is less easily formed. Accordingly, the population of the less densely packed gauche conformer increases (60-70%) as well as the basal spacing. The further shortening of the surface area per cation occurring with the highest charge saponites (0.75 and 0.80), changes the balance between repulsive and attractive interactions, leading to a different arrangement. The interchain Van der Waals interactions are enhanced and the trans conformation raises until ca. 90%. This is consistent with X ray diffraction data accounting for all-trans alkyl chains being tilted at 54.5° to the clay surface, the angle of the tilt that optimises the binding of the head group to the clay
236
J. Grandjean
surface [87]. 13C CP NMR has been used to identify a proton transfer and ion migration in a solid amine-clay mixture. The authors compare the spectra of phenethylammonium (PA) and benzyltriethylammonium (BTA) cations incorporated into hectorite or Laponite. The line width of the aromatic part is similar when PA or BTA is exchanged on hectorite and for a ground mixture of the clay and the chloride salt of PA. The ground mixture of the bromide salt of BTA is characterized by a much narrower line, similar to that of the pure salt. It has been suggested that proton transfer and subsequent ion migration occurring between clay protons and chloride ions are responsible for the broader line observed with PA [88]. During the synthesis of hectorite by hydrothermal crystallisation of a magnesium silicate, tetraethyl ammonium ions are used to aid crystallization and become incorporated as the exchange cations within the interlayer. This process has been followed by 13C MAS NMR to support a possible clay crystallization mechanism [89]. In clay suspensions, 2H NMR has revealed preferential orientation of molecules and ions near the mineral surface. Fast mobility strongly reduces the static quadrupolar splitting. Although slower motions are present in solids, the static 2H NMR spectrum remains sensitive to mobility variation. The ammonium protons of OA have been deuterated before intercalation in saponite. Variation of the deuteron splitting has been used to study the temperature influence on the head group dynamics. At 116K, a rather weak splitting value of 50 kHz has been determined, indicating internal rotation of the ND3+ group about its C3 axis is already operative. Cationic uniaxial rotation along the long axis of the surfactant cation gradually occurs in the 116-400K temperature range. The progressive reduction of the quadrupolar splitting is also indicative of a wide distribution of correlation times associated to this rotation. Large-amplitude motion of the whole ion in the interlayer space accounts for the quadrupolar splitting values obtained at higher temperatures [90]. ! H longitudinal relaxation times T! of TMAsaponite have been measured as a function of temperature. Under molecular isotropic rotational diffusion, the T! plot versus the inverse of temperature (proportional to the correlation time) shows one minimum (figure 2). A quite different plot is obtained for more complex motions. The relevant equation also predicts a co2 dependence of Tj when COTC> 1, whereas a much weaker dependence of « co1 ° below 140K is observed. This is indicative of a continuous distribution of the correlation times. The observed T! curve as a function of the inverse of temperature has been reproduced assuming two motional modes, isotropic cationic rotation as a whole and translational self-diffusion [91]. 4.3 - Intercalated (adsorbed) molecules As pointed previously, the presence of paramagnetic metal ions (mainly iron) in most natural clays induces line broadening that limits the suitability of NMR methods. For this reason, such studies have been restricted to synthetic samples or to natural smectites with very low iron content, especially hectorite. The NMR studies are concerned with characterization, structural and dynamic aspects of the clay-organic complexes. Sorption of triethylphosphate (TEP) on ion-exchanged smectites has been studied by 31P and 13C (CP) MAS NMR spectroscopy. TEP molecules progressively release coordinated water molecules from the solid interface. The 31P CP MAS spectrum of the adsorbed molecules shows two lines characterizing two motionrestricted phases attributed to monolayer and bilayer complexes. NMR (and IR) results suggest that the TEP molecules are directly coordinated to the interlayer cations via the
Nuclear Magnetic Resonance Spectroscopy
237
phosphoryl group P = O. The mobility of the sorbed species depends on the charge density of the clay [92,93]. Tributylphosphate (TBP) is one of the organic compounds which is used as ligating agent for nuclear fuel processing and is a widespread contaminant in ground water and soils surrounding processing facilities. The europium isotopes are among the release contaminants. Pure TBP exhibits a 31P NMR signal at 0.3 ppm, but in the Eu(NO3)3 - TBP complex, the signal is shifted to ca. -180 ppm. The pseudo-contact interaction between unpaired electrons of the (paramagnetic) cation and 31 P nuclei is responsible for it. Hectorite adsorbs this complex from solution, as shown by the signal near -180 ppm of the 31P MAS NMR spectrum. Another signal occurs at 5 ppm when TBP is in excess, corresponding to uncomplexed TBP that is not exchanged with the absorbed complex. When Eu-hectorite is put in contact with excess of TBP, no complex is formed. Based on these results, the authors conclude that actinides and lanthanides can enter a clay as a cation, presumably from a highly aqueous environment or as an organic complex formed prior to the sorption into the clay [94]. Small organic molecules such as urea, formamide or dimethyl sulfoxyde (DMSO) can also penetrate the kaolinite (phyllosilicate 1:1) galleries. NMR studies of these intercalates deals with the orientation, conformation and dynamics of the guest molecules. The 13C signals of the intercalated molecules are shifted downfield in response to increased hydrogen bonding after intercalation. The 13C spectrum of the kaolinite-DMSO intercalate shows two equally intense methyl signals which have been first assigned to two inequivalent methyl groups in the same DMSO molecule, one methyl group being keyed into the ditrigonal holes of the silicate layer and the other standing approximately parallel to the sheets [95,96]. 2H NMR has been used to study the molecular motion of the intercalated DMSO molecules. Another model has been proposed assuming two kinds of interlayer DMSO sites: one DMSO molecule has one methyl group keyed like in the previous model, and the other DMSO molecule, not keyed, adopts another orientation [97,98]. In both models the sulfonyl oxygen is hydrogen-bonded to the inner surface hydroxyls. More recently, a multinuclear NMR study has been performed in the temperature range 170-380 K [99-101]. The main conclusions of these papers can be summarized as: - Below 320 K, all interlayer molecules are equivalent with one methyl group keyed in the pseudohexagonal cavities of the silicate sheet. Above 320 K, some of the keyed methyl groups are released from the trapped holes giving rise to two coexisting DMSO sites in the interlayer space. Increasing temperature enhances this process and all the interlayer molecules are essentially free at 415 K. - The methyl groups of the intercalated DMSO molecules undergo free rotation around their C3 axis over the investigated temperature range. This symmetry axis is fixed at low temperature (ca. 160 K). A wobbling motion of the methyl groups is induced at higher temperature. After being released from the ditrigonal holes, the DMSO molecules undergo an anisotropic rotation of the whole molecules. Intercalation of formamide (FA), N-methylformamide (NMF), and Ndimethylformamide (DMF) has been studied similarly. In the kaolinite - FA system, outer surface and interacalated molecules are characterized by two carbonyl signals at 166.4 and 168.4 ppm, respectively. Molecules loosely adsorbed on the outer surface of the host have great mobility and produce negligible cross polarization efficiency. Accordingly, the outer surface signal at 166.4 ppm is not present in the 13C CP MAS NMR spectrum of the kaolinite - FA system. Among the three amide molecules,
238
J. Grandjean
intercalation of FA molecules induces the strongest downfield shift of the carbonyl signal. Weak Van der Waals interaction between the guest and the host results in a deshielding of the resonance, and the strength of the interaction is in the order FA > NMF » DMF. Therefore, mobility of the intercalated molecules, as deduced from the 13 C line widths, is in the reverse order FA < NMF < DMF. The *H chemical shifts confirm the existence of hydrogen-bonding between the amide protons of FA and the oxygens of the silicate sheets. No such hydrogen-bonding with NMF exists at room temperature. In contrast to intercalated DMSO molecules, the chemical shift value of the hydroxyl groups in the host kaolinite structure is not affected by intercalation. The structure of the three kaolinite intercalates is schematically shown in the figure 8 [102]. Table 4 - Studied molecules Acronym (name) Acetone Benzene DMF DMSO FA HEX MMA NMF PEG PEO PS Pyridine TBP Trichloroethylene TEP TNT
Molelecule (CH3)2C=O C6H6 (CH3)2NCOH (CH3)2S=O H2NCOH H2C=CH(CH2)3CH3 H 2 OC(CH 3 )COOCH 3 H3CNHCOH H(OCH2CH2)nOH (-CH2CH2O-)n (-CH2CH(C6H5)-)n C6H5N (CH3(CH2)3 O)3P=O CC12=CHC1 (CH3CH2O)3P=O (NO2)3H2C6CH3
Reference [107] [105,106] [102] [95-101] [102] [103] [117] [102] [110] [83],[111-116] [116] [107] [94] [107] [92],[93] [108]
Interaction of interlayer cations of Laponite and hex-1-ene (HEX) has been studied by 'H, 13C NMR, and 27A1 MAS NMR. All the expected I3C signal of HEX are detected with the Na-Laponite. By contrast, the peaks of the olefinic carbon atoms are not visible when the clay counterion is Al3+. This suggests some direct or indirect interaction of the double bond with aluminum ions leading to extensive line broadening whereas the saturated part of the hydrocarbon chain moves rather freely. Comparison between the 27A1 MAS NMR spectra of Al-Laponite and the clay-organo complex support this sorption mechanism [103]. Crown ethers and cryptands are known to complex cations. Their intercalation in montmorillonite and hectorite results in the formation of 1:1 and 2:1 ligand/cation interlayer complexes. In Na-hectorite, the apparent 23Na shift (second order effects have not been suppressed) varies with the nature of the macrocycle [104].
Nuclear Magnetic Resonance Spectroscopy
239
Figure 8 - Schematic representation of the structure of kaolinite intercalated compounds with (A) FA, (B) NMF, and (C) DMF. (Reprinted with permission from [102]. Copyright (1999) American Chemical Society). Macroscopic ordering of clay aggregates allows the determination of the arrangement of molecules intercalated between the clay platelets from the determination of the chemical shift anisotropies, because the principal axes of the chemical shift tensor are related to the molecular coordinate system. The 13C NMR spectra of benzene sorbed on Ag-exchanged hectorite show a single line whose position depends on the orientation of the clay platelets with respect to the external magnetic field. At room temperature, benzene molecules are mobile as inferred from the rather narrow signal. Molecular motion is restrained at lower temperature. Below 251 K, the anisotropy pattern is only detected. The equation (1) adapted to the chemical shift anisotropy is used to simulate the experimental spectrum. To reproduce the experimental patterns, it was necessary to tip the benzene C6 axis up out the ab plane of the clay platelet by about 15°. The proposed dynamic model includes rotation of the benzene molecules about the C6 axis and about the normal to the clay layer. The latter motion is quenched at 77 K [105]. More recently, local motion of benzene adsorbed on Ca-montmorillonite has been studied by 2H NMR. These results suggest that adsorbed benzene molecules first form n complexes with Ca2+ in the interlayer space of the clay. At or below 198 K, the quadrupolar splitting of the adsorbed molecules is close to half of the static-pattern splitting. From the equation (1), adapted to this system, local molecular motion can be deduced. Adsorbed benzene molecules undergo a small-angle wobbling of the C6 axis accompanied by discrete jumps about the hexad axis. At higher temperatures, benzene molecules start to perform large-angle wobbling of the C6 axis with extremely fast jumps around the axis, and eventually desorb from Ca2+ to tumble freely in the interlayer space of the clay. Fast motion cancels the quadrupolar interaction as shown by the presence of a central signal in the spectrum [106]. Using similar computer
240
J. Grandjean
simulations of 2H line shapes, this research group has also studied local motions of deuterium-labelled organic pollutants (trichlorethylene, acetone, pyridine, benzene or ethylene glycol) adsorbed on montmorillonite. Local mobility of the adsorbed trichloroethylene molecules remains partly restricted even above the 185 K melting point of the pollutant. This indicates a strong Cl2C=CDCl/Ca-montmorillonite interaction. Rapid and largely isotropic reorientation only occurs at room temperature when the spectrum exhibits a narrow single line. At similar low temperatures, the spectrum of adsorbed acetone molecules is narrowed by roughly a factor of 3 compared to the width of the adsorbed trichloroethylene spectrum. This is consistent with fast rotation of the CD3 group about the C-C axis but a composite motion consisting of three-site hop of the CD3 group and two-site hop about the C=O axis also accounts for the observed spectra. In the case of pyridine adsorbed on a substrate with acidic sites, one may expect at least some contributions from hydrogen-bonded complexes (N — HA). Motion about the Cpara — N —H (hydrogen bond) axis can reproduce the experimental spectrum at low temperature [107]. Dynamics of adsorbed trinitrotoluene has been studied by 2H MAS NMR. Magic angle spinning rate slower than the quadrupolar interaction gives rise to a narrow central band and spinning site bands spaced at intervals equal to the MAS rate [12]. Analysis of the spinning side band pattern leads to the mode of motion of the CD3 group. Intercalation of TNT in the acidtreated K10 montmorillonite provides a 2H NMR spectrum simulated by assuming C3 ring jumps (55%) and free rotation around the axis perpendicular to the aromatic ( and mineral siloxane) plane. The adsorbed TNT binding is weaker with Ca-montmorillonite, and free ring rotation dominates [108]. Intercalation of polymer molecules has attracted considerable attention during the last few years. NMR spectroscopy has been used for characterization and dynamics of the intercalated organic material. Thus NMF in NMF-kaolinite has been replaced by acrylamide which was then polymerised in situ. The process was followed by the change of the signal intensity of the ethylenic carbon atoms of the 13C CP MAS NMR. One-hour treatment is required to complete the polymerisation [109]. Polymerization of ethylene glycol intercalated in kaolinite was not observed. However, poly(ethylene glycol)-kaolinite intercalates have been prepared by releasing intercalated DMSO. 13C CP MAS NMR spectra, in combination with IR and X-ray diffraction, indicate that the intercalate polymer (PEG) is more constrained in the interlamellar space of the clay than it is in the bulk. The intercalated polymers are arranged in flattened monolayers where the ethyleneoxy groups show their oxygen atoms facing towards the hydroxyl surface of kaolinite [110]. Poly(ethylene oxide) (PEO) complexes are interesting materials due to their anisotropic ionic conductivity. The 13C CPMAS NMR spectra of Na-, K-, and Bahectorite/PEO complexes consist of a unique signal at « 70 ppm that has been assigned to a gauche conformation of the methylene groups, suggesting that the helicoidal conformation of the polymer PEO is maintained after intercalation. The 23Na NMR spectrum of the relevant complex shows also a unique peak, indicating an homogeneous environment inside the polymer helix and sodium ions are directly coordinated to the ethylenoxy units of the polymer [111]. Later, 2H solid-state NMR has been used to probe dynamics of the polymer in deuterated PEO/Li-fluorohectorite intercalates. The temperature dependence of the quadrupolar NMR powder patterns indicates that PEO, even at 220 K, possesses some small amplitude dynamics. At higher temperatures, even above the bulk polymer melting point, the spectra still retain residual powder patterns in
Nuclear Magnetic Resonance Spectroscopy
241
addition to an intense central line, indicating restricted polymer motion between silicate layers. In addition, the authors indicate that there is a competition between the PEO oxygens and the surface oxygens of the silicate layers for interaction with Li+ cations [112]. PEO intercalated in montmorillonite and hectorite has also been studied by solidstate NMR to probe cation dynamics in these nanocomposites. In montmorillonite, the paramagnetic iron content produces an extra broadening of 15 kHz, compared to the 7Li NMR spectrum of the hectorite nanocomposite at 220 K. As the temperature increases, the 7Li linewidths observed in the montmorillonite systems undergo narrowing beyond that associated to the paramagnetic effect. Ion dynamics, responsible for this additional narrowing, is approximately two orders of magnitude less than polymer orientation. Cation surroundings in the vicinity of Li+ differ from that found in the nanocomposites with larger alkali ions, as an effect of the greater cation - Fe3+ distance [113]. The theoretical background of the paramagnetic effect on the NMR parameters of these systems are described in the literature [83]. The OC - CO torsion angle of PEO with 13% 13C - 13C labelled units can be estimated from the simulation of the 13C twodimensional double quantum (2D DQ) NMR spectra. The gauche content is estimated to 90 + 5%, which provides valuable constraints on the possible conformation in the intercalation gap [114]. Numerous solid-state NMR techniques are currently used to characterize the structure and dynamics of solids. The PEO-hectorite intercalate has been considered in order to optimise the conditions of several NMR experiments dedicated for studying organic materials near the silicate surfaces [115]. The intercalation of poly(styrene - ethylene oxide) block polymers (PS-fe-PEO) into hectorite has been studied by multinuclear solid-state NMR. Polymer intercalation is assessed by 2D 'H-29Si heteronuclear correlation spectra. Two copolymers with similar PEO block lengths (7 and 8.4 kDa) but different PS block lengths (3.6 versus 30 kDa) are compared. While the PS block is found not to be intercalated in either copolymer, definite proof of PEO intercalation in the sample with the shorter PS block is provided by a 'H- 1 3 C heteronuclear correlation experiment. In the PS-rich sample, the amount of intercalated PEO is much smaller, and a significant fraction of PEO is not intercalated. A schematic model is shown in Figure 9 [116]. Intercalation of copolymer of methyl methacrylate (MMA) (Table 4) and MDEA (OCH3 of MMA replaced by O(CH2)2N+(CH2CH3)2CH3r) into hectorite has been performed in two ways: either the formed copolymer has been directly intercalated in the clay or copolymerisation has been realized in situ on MDEA-exchanged hectorite. Different 13C relaxation times in the laboratory and the rotating frames are measured for the solid copolymer (MMA/MDEA ratio of 8) and the copolymer directly intercalated in hectorite, indicating the intercalation process affects copolymer dynamics. By contrast, the 13C relaxation times in the laboratory and the rotating frames are similar for the solid copolymer (MMA/MDEA ratio of 6) and the copolymer-hectorite complex prepared in situ. This could be due to a small percentage of MMA copolymerisation between the clay layers, most of MMA unit sequences are formed mainly out of the layers [117]. Polyaniline can exist in several oxidation states ranging from the completely reduced base state to the completely oxidized state. The most studied form of the polymer is emeraldine, in which the number of reduced and oxidized units is equal. This insulator can be converted to a conducting form by proton doping. 2H MAS NMR studies of the polymer indicate that polarons play an important role in the conductivity mechanism. After intercalation into montmorillonite, the polymer spectrum is similar to
242
J. Grandjean
that of the non conductive emeraldine base. Failure to observe a (metallic) Knight shift in the nanocomposite, is consistent with the importance of 3D mechanisms of charge transfer for bulk conductivity in polyaniline [118].
Figure 9 - Schematic structural models of PS-b-PEO intercalated in hectorite (PEO blocks: dashed lines); PS blocks: solid lines); a and b are the side and top views for the copolymer with short chains of PS; c and d the corresponding views for the copolymer with long chains of PS. (Reprinted with permission from [116]. Copyright (2003) American Chemical Society). Interaction of polypeptides with clay minerals is at the basis of both natural biochemical processes and technological applications. Direct observation of polylysine side-chain interaction with montmorillonite interlayer surfaces is shown through 2D 'H27 A1 heteronuclear correlation NMR spectroscopy. Indeed, a peak correlates the NH3+ protons of the polylysine side chains and octahedral Al(III) of the mineral. Fast rotational motion around the C3 axis of the NH3+ group allows resolution of this small signal [119]. Adsorption of polylysine and polyglutamic acid on montmorillonite has been studied by 13C CP MAS NMR. The chemical shift of the backbone a- and carbonyl carbon nuclei have shown that both polypeptides, which exhibit a mixture of ct-helical and random conformation in the bulk, tend to unfold and adopt a more extended random coil structure on adsorption. Furthermore, analysis of the resonance line widths indicates an increase ordering of the positively-charged side chain of polylysine but not in the case of the negatively-charged side chain of polyglutamic acid [120]. 5 - Conclusions The diversity of the NMR experiments allows to study clay systems, ranging from water-rich suspensions to dried solids. The determination of the chemical shifts, quadrupolar splittings, relaxation rates or self-diffusion coefficients provides a detailed picture on the structure and dynamics of molecules and ions near a clay surface. These studies have also benefited from the more recent developments of NMR technology
Nuclear Magnetic Resonance Spectroscopy
243
such as two-dimensional or multiple quantum coherence experiments. The survey of literature indicates that NMR study of clay systems is a growing and promising area. Acknowledgments I am grateful to F.N.R.S. (Brussels) for grants to purchase a solid-state NMR spectrometer and for support in our studies on clay materials. 6 - References [I] T.R. Jones, Clay Miner. 18 (1983) 399. [2] D.P. Siantas, B.A. Feinberg and J.J. Fripiat, Clays Clay Miner., 42 (1994) 187. [3] J.T. Smith and R.N.J. Comans, Geochim. Cosmochim. Acta, 60 (1996) 995. [4] Y.O. Aochi and W.F. Farmer, J. Colloids Interface Sci., 161 (1993) 106. [5] L. Bailey, M. Keall, A. Audibert and J. Lecourtier, Langmuir, 10 (1994) 1544. [6] E.S. Boek, P.V. Coveney and N.T. Skipper, J. Amer. Chem. Soc, 117 (1995) 12608. [7] M. Ogawa and K. Huroda, Bull. Chem. Soc. Jpn., 70 (1997) 2593. [8] M. Zanetti, S. Lomakin and G. Camino, Macromol. Mater. Engn., 279 (2000) 1. [9] M. Alexandre and P. Dubois, Mater. Sci. Engn., 98 (2000) 1. [10] D. Canet, Nuclear Magnetic Resonance: Concepts and Methods, J. Wiley & Sons, Chichester, 1996. [II] M. H. Levitt, Spin Dynamics: Basis of Nuclear Magnetic Resonance, J. Wiley & Sons, Chichester, 2001. [12] D.D. Laws, H.-M.L. Bitter and A. Jerschow, Angew. Chem. Int. Ed., 41 (2002) 3096. [13] J. McConnell, The Theory of Nuclear Magnetic Relaxation in Liquids, Cambridge, University Press, Cambridge, 1987. [14] J.-J. Delpuech Ed., Dynamics of Solutions and Fluid Mixtures by NMR, J. Wiley & Sons, Chichester, 1995. [15] O. Sodermann and P. Stilbs, Progr. NMR Spectr., 26 (1994) 445. [16] W.S. Price, Annu. Rep. NMR Spectrosc, 32 (1996) 53. [17] J. Grandjean, Annu. Rep. NMR Spectrosc, 35 (1998) 217. [18] J. Grandjean, in Encyclopedia of Surface and Colloid Science, A. Hubbard Ed., M. Dekker, (2002) 3700. [19] D.E. Woessner and B.S. Snowden Jr., J. Chem. Phys., 50 (1968) 1516. [20] B. Halle and H. Wennerstram, J. Chem. Phys., 75 (1981) 1928. [21] J. Grandjean and P. Laszlo, J. Magn. Reson., 83 (1989) 128. [22] D.T. Edmons and A.L. Mackay, J. Magn. Reson., 20 (1975) 515. [23] D.T Edmons and A. Zussman, Phys. Lett., 41A (1972) 167. [24] J. Grandjean and P. Laszlo, ACS Symp. Ser., 415 (1990) 396. [25] J. Grandjean and P. Laszlo, Clays Clay Miner., 37 (1989) 403. [26] D. Petit, J.-P. Korb and A. Delville, J. Grandjean, P.Laszlo, J. Magn. Reson., 96 (1992)252. [27] J. Grandjean and P. Laszlo, Clays Clay Miner., 42 (1994) 652. [28] C.A. Weiss and W.V. Gerasimowicz, Geochim. Cosmochim. Acta, 60 (1996) 265. [29] J. Grandjean, J. Colloid Interface Sci., 185 (1997) 554. [30] A. Delville, J. Grandjean and P. Laszlo, J. Phys. Chem., 95 (1991) 1383. [31] J.A. Ripmeester, L.S. Kotlyar and B.D. Sparks, Colloids Surf. A, 78 (1993) 57. [32] J. Grandjean and P. Laszlo, J. Amer. Chem. Soc, 116 (1994) 3980.
244
J. Grandjean
[33] J. Grandjean and P.Laszlo, J. Colloid Interface Sci., 175 (1995) 207. [34] J. Grandjean and J.-L. Robert, J. Colloid Interface Sci., 187 (1997) 267. [35] J. Grandjean and J.-L. Robert, J. Magn. Reson., 138 (1999) 43. [36] J. Grandjean and P. Laszlo, J. Magn. Reson. Ser A, 118 (1996) 103. [37] C. Breen, J.O. Rawson and B.E. Mann, J. Mater. Chem., 6 (1996) 253. [38] P. Porion, M.P. Faugere, E. Lecolier, B. Gherardi and A. Delville, J. Phys. Chem., 102 (1998) 3477. [39] P.Porion, M. Al Mukhtar, A.M. Faugere, R.J.M. Pellenq, S. Meyer and A. Delville, J. Phys. Chem. B, 107 (2003) 4012. [40] P. Porion, M. Al Mukhtar, S. Meyer, A.-M. Faugere, J.R.C. van der Maarel and A. Delville, J. Phys. Chem. B, 105 (2001) 10505. [41] Y. Nakashima, Clays Clay Miner., 48 (2000) 603. [42] Y. Nakashima, Amer. Mineral., 86 (2001) 132. [43] Y. Nakashima, Clay Miner., 37 (2002) 83. [44] Y. Nakashima, Clays Clay Miner., 51 (2003) 9. [45] F.P. Duval, P. Porion, A.-M. Faugere and H. Van Damme, J. Colloid Interface Sci., 242(2001)319. [46] P. Porion, S. Rodts, M. Al Mukhtar, A.-M. Faugere and A. Delville, Phys. Rev. Lett., 87(2001)208302-1. [47] R.M.E. Valkenbrog, L. Pel and K. Kopinga, J. Magn. Reson., 151 (2001) 291. [48] S. Abbrent, S.H. Chung, S.G. Greenbaum, J. Muthu and E.P. Giannelis, Electrochim. Acta, 48 (2003) 2113. [49] S.K. Sur, J.F. Heinsberger and R.G. Bryant, J. Magn. Reson., 103 (1993) 8. [50] A. Delville and L. Letellier, Langmuir, 11 (1995) 1361. [51] T. Zavada, J. Grandjean, R. Kimmich and A. Kobelkov, J. Chem. Phys., 110 (1999)6977. [52] S. Rossi, P.F. Luckham, N. Green and T. Cosgrove, Colloids Surfaces A, 215 (2003)11. [53] J. Grandjean, Langmuir, 14 (1998) 1037. [54] C. Gevers and J. Grandjean, J. Colloid Interface Sci., 236 (2001) 290. [55] P.S. Hubbard, J. Chem. Phys., 53 (1970) 985. [56] T.E. Bull, J. Magn. Reson., 8 (1972) 344. [57] J.P.H. Zwetsloot and J.C. Leyte, J. Colloid Interface Sci., 181 (1996) 351. [58] U. Eliav, H. Shinar and G. Navon, J. Magn. Reson., 98 (1992) 223. [59] J. Grandjean, J. Colloid Interface Sci., 239 (2001) 27. [60] D. Tinet, A.M. Faugere and R. Prost, J. Phys. Chem., 95 (1991) 8804. [61] P. Di Leo and P.O'Brien, Clays Clay Miner., 47 (1999) 761. [62] P. Di Leo and J. Cuadros, Clays Clay Miner., 51 (2003) 403. [63] M.D. Alba, A.I. Becerro, M.A. Castro and A.C. Perdigon, Chem. Commun., (2000, 37. [64] V. Laperche, J.F. Lambert, R. Prost and J.J. Fripiat, J. Phys. Chem., 94 (1990) 8821. [65] J.F. Lambert, R. Prost and M.E. Smith, Clays Clay Miner., 40 (1992) 253. [66] C.A. Weiss Jr., R.J. Kirkpatrick and S.P. Altaner, Geochim. Cosmochim. Acta, 54 (1990) 1655. [67] C.A. Weiss Jr., R.J. Kirkpatrick and S.P. Altaner, Amer. Miner., 75 (1990) 970. [68] Y. Kim and R.J. Kirkpatrick, Geochim. Cosmochim. Acta, 61 (1997) 5199. [69] V. Luca, CM. Cardile and R.H. Meinhold, Clay Miner., 24 (1989) 115.
Nuclear Magnetic Resonance Spectroscopy
245
[70] B.K.G. Theng, S. Hayashi, M. Soma and H. Seyama, Clays Clay Miner., 45 (1997) 718. [71] M.D. Alba, A.I. Becerro, M.A. Castro and A.C. Perdig6n, Chem. Commun., (2001)249. [72] D.J. Sullivan, J.S. Shore and J.A. Rice, Clays Clay Miner., 46 (1998) 349. [73] J. Rocha, CM. Morais and C. Fernandez, Clay Miner., 38 (2003) 259. [74] L. Delevoye, J.-L. Robert and J. Grandjean, Clay Miner., 38 (2003) 63. [75] D. Plee, F. Borg, L. Gatineau and J.J. Fripiat, J. Amer. Chem. Soc, 107 (1985) 2362. [76] K.B. Brandt and R.A. Kydd, Chem. Mater., 9 (1997) 567. [77] M.J. Hernando, C. Pesquera, C. Blanco and F. Gonzalez, Chem. Mater., 13 (2001) 2154. [78] J.-P. Amoureux and M. Pruski, Molec. Phys., 100 (2002) 1595. [79] W. Kolodziejski and J. Klinowski, Chem. Rev., 102 (2002) 613. [80] T.K. Pratum, J. Phys. Chem., 96 (1992) 4567. [81] K. Khatib, M. Francois, P. Tekely, L.J. Michot, J.Y. Bottero and I. Baudin, J. Colloid Interface Sci., 183 (1996) 148. [82] L.-Q. Wang, J. Liu, J. Exharos, K.Y. Flanigan and R. Bordia, J. Phys. Chem. B, 104(2000)2810. [83] D.K. Yang and D.B. Zax, J. Chem. Phys., 110 (1999) 5325. [84] D.L. VanderHart, A. Asano and J.W. Gilman, Chem. Mater., 13 (2001) 3781. [85] J. Grandjean, J. Bujdak and P. Komadel, Clays Miner., 38 (2003) 369. [86] D. Kubies, R. Jerome and J. Grandjean, Langmuir, 18 (2002) 6159. [87] R. Miiller, J. FIrobarikova, C. Calberg, R. Jerome and J. Grandjean, Lagmuir, in press. [88] S. Bank and G. Ofori-Oki, Langmuir, 8 (1992) 1688. [89] K.A. Carrado, L. Xu, D.M. Gregory, K. Song, S. Seifert and R.E. Botto, Chem. Mater., 12(2000)3052. [90] M. Yamauchi, S. Ishimaru and R. Ikeda, Mol. Cryst. Liq. Cryst., 341 (2000) 315. [91] S. Ishimaru, M Yamauchi and R. Ikeda, Z. Naturforsch., 53 A (1998) 903. [92] P.O'Brien, C.J. Williamson and C.J. Groombridge, Chem. Mater., 3 (1991) 276. [93] C.J. Williamson and P.O'Brien, J. Mater. Chem., 4 (1994) 565. [94] C.J. Hartzell, S.-W. Yang, R.A. Parnell and D.E. Morris, J. Phys. Chem., 99 (1995) 4205. [95] J.G. Thompson and C. Cuff, Clay Clay Miner., 33 (1985) 173. [96] M. Raupach, P.F. Barron and J.G. Thompson, Clays Clay Miner., 36 (1987) 208. [97] M,J. Duer and J. Rocha, J. Magn. Reson., 98 (1992) 524. [98] M.J. Duer, J. Rocha and J. Klinowski, J. Amer. Chem. Soc, 114 (1992) 6867. [99] S. Hayashi and E. Akiba, Chem. Phys. Lett., 226 (1994) 495. [100] S. Hayashi, J. Phys. Chem., 99 (1995) 7120. [101] S. Hayashi, Clays Clay Miner., 45 (1997) 724. [102] X. Xie and S. Hayashi, J. Phys. Chem. B, 103 (1999) 5949. [103] D.T.B. Tenakoon, J.M. Thomas, W. Jones, T.A. Carpenter and S. Ramdas, J. Chem. Soc. Farad. Trans. I., 82 (1986) 545. [104] P. Aranda, B. Casal, J.J. Fripiat and E. Ruiz-Hitzky, Langmuir, 10 (1994) 1207. [105] H.A. Resing, D. Slotfeldt-Ellingsen, A.N. Garroway, T.J. Pinnavaia and K. Unger, in Magnetic Resonance in Colloid and Interface Science, J. P. Fraissard, H. A. Resing, Eds., D. Reidel, (1980) p. 239.
246
J. Grandjean
[106] J. Xiong and G.E. Maciel, J. Phys. Chem. B,103 (1999) 5543. [107] J. Xiong, H. Lock, I. Chuang, C. Keeler and G.E. Maciel, Environ. Sci. Technol, 33 (1999). [108] E.F. Emery, T. Junk, R.E. Ferrell Jr., R.De Hon and L.G. Butler, Environ. Sci. Technol., 35(2001)2973. [109] Y. Sugahara, S. Satokawa, K. Kuroda and C. Kato, Clays Clay Miner., 38 (1990) 137. [110] J.J. Tunney and C. Detellier, Chem. Mater., 8 (1996) 927. [111] P. Aranda and E. Ruiz-Hitzky, Chem. Mater., 4 (1992) 1395. [112] S. Wong, S. Vasudevan, R.A. Vaia, E.P. Giannelis and D.B. Zax, J. Amer. Chem. Soc, 117(1995)7568. [113] S. Wong and D.B. Zax, Electrochim. Acta, 42 (1997) 3513. [114] D.J. Harris, T.J. Bonagamba and K. Schmidt-Rohr, Macromolecules, 32 (1999) 6718. [115] S.-S. Hou, F.L. Berger and K. Schmidt-Rohr, Solid State NMR, 22 (2002) 110. [116] S.-S. Hou, T.J. Bonagamba, F.L. Berger, P.H. Madison and K. Schmidt-Rohr, Macromolecules, 36 (2003) 2769. [117] C. Forte, M. Geppi, S. Giamberini, G. Ruggeri, C.A. Veracini and B. Mendez, Polymer, 39(1998)2651. [118] Y.A. Goddard, R.L. Void and G.L. Hoatson, Macromolecules, 36 (2003) 1162. [119] R.D. Gougeon, M. Reinholdt, L. Delmotte, J. Miehe-Brendle, J.-M. Chezeau, R. Le Dred, R. Marchal and P. Jeandet, Langmuir, 18 (2002) 3396. [120] R.D. Gougeon, M. Soulard, M. Reinholdt, J. Miehe-Brendle, J.-M. Chezeau, R. Le Dred, R. Marchal and P. Jeandet, Eur. J. Inorg. Chem., (2003) 1366.
PESTICIDE-CLAY INTERACTIONS AND FORMULATIONS JUAN CORNEJO*, RAFAEL CELIS, LUCIA COX and M. CARMEN HERMOSIN Instituto de Recursos Naturales y Agrobiologia de Sevilla, CSIC. P.O. Box 1052. 41080 Sevilla - SPAIN. * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
248
J. Cornejo et al.
1 - Introduction During the last few years it has been reported that pesticides have been found in several compartments like soil, ground water, surface water, sediments, air, foods and even animal tissues. This situation is viewed with great concern due to the environmental problems arising from the use of enormous amount of pesticides and specially for those persistent and mobile molecules affecting soil-water ecosystems entering later in the trophic chain. Production and uses of pesticides are still increasing. However, there are also several significant developments that will have long term impact on pesticide usage and residues in water. There has been a steady decrease in the amount of herbicide needed to control weeds since 1940's (Figure 1), because the specificity of the new molecules seems to be more effective using lower dose than earlier ones [1]. The soil is the main recipient of organic agrochemicals used in plant protection to control or destroy weeds, insect, fungi, and other pests in a deliberated way. However, sometimes the soil receives those chemicals in accidental ways like spillages from broken containers, industrial wastes disposal, etc. Thus the soil plays a central role in the environmental fate of these chemicals and in the protection of ground and surface waters. The soil behaves as a filter, where the organics are chemically or/and biologically degraded but it can also be able to retain most of these chemicals avoiding or limiting their leaching to deeper soil horizons and ground waters. For these reasons the soil can not be regarded as a sink for chemicals or as a compartment with unlimited contaminant loading capacity or unlimited natural attenuation power.
Figure 1 - Historical trends in recommended herbicide application rates, by chemical class. After Nash and Leslie [I]. Once the pesticide is applied to targeted objectives (weeds, plants pest, soil, etc), most of the chemical finally reaches the soil either directly or from leaves or air deposition. After entering the soil, pesticides undergo different processes in the soil environment mainly of adsorption-desorption, transport and physico-chemical and biological degradation [2].
Pesticide-Clay Interactions and Formulations
249
The above-mentioned processes are of course dependent on three big groups of factors that are associated with the soil properties (structure, texture, pH, % Organic Matter (OM), microbiological activity, etc.) and soil management (traditional or conservation tillage, irrigation method, etc.) to the pesticide properties (water solubility, stability-pH-UV, formulation, etc.) and to the climate conditions (rainfall, temperature, UV, etc.). The organic fraction of the solid phase of the soil is often assumed to dominate the interactions of pesticides in the soil environment. This assumption is taken by modelers and specially in studies dealing with rich organic matter soils obscuring the important role played by the mineral surfaces on the fate of pesticides in soil. However, mineral surfaces may dominate the fate of organic chemicals arriving in the soil. It is very well known that in Mediterranean, semiarid and arid zones the soils are very poor in OM content being the mineral fraction the responsible colloidal particles for surface interactions. Even in normal soils and subsoils it is expected an important contribution of the mineral fraction to the interaction of polar or weakly polar organic molecules like some pesticides, due to the higher percentages of the mineral fraction over the organic one. It is also necessary to consider that both soil fractions are not generally isolated but forming organo-mineral associations with surface properties very different from those they own separately contributing to sorption process [3-5]. In summary, clay-pesticide interactions are here considered under two complementary points of view: pesticides interaction with the clay minerals of the soils and the design of clay-based formulations for an improvement of the agronomical and environmental use of pesticides. 2 - Pesticides Agricultural pesticides are often detected in natural waters, and therefore, they are an important group of organic pollutants which production and uses must be controlled to minimize the health and environmental problems. There is a strong relationship between the amount of pesticide applied and the amount detected in soil and water. It has been pointed out about the decrease of the amount of new pesticides needed for control some pests. However, it is also true that the total amount of pesticide marketed in the world increased because of the change in farming practices, beginning in the 90's. No-tillage or conservation tillage is being widely used in developed countries. This type of soil management needs higher amounts of herbicides with the corresponding environmental impact but reducing the soil erosion and increasing the water infiltration because seeds are drilled directly into the soil containing plants residues from the previous crop instead of plough the fields before planting. Before drilling the seed, all weeds are destroyed with the appropriated herbicide. Pesticide is a generic name for compounds used for pest control most of them used in agricultural practices. On a weight basis, agriculture is the largest user of pesticides (77%) the rest being used for industrial and commercial activities (16%) and for home and garden sectors (6%). The three main groups of pesticides are insecticides controlling insects, herbicides for weeds, and fungicides for plant diseases. There are some other small groups of chemicals with specific objectives like rodenticides, nematicides, fumigants, molluscicides and plant growth regulators. There is a large number of pesticides currently in use, with a wide range of physicochemical properties which determine its behavior in the environment. Molecular size, ionizability, water
250
J. Cornejo et al.
solubility, lipophilicity, polarizability and volatility are all important properties. Pesticides can be classified in many different ways. Figure 2 is a classification scheme for selected pesticides on the basis of their significant chemical properties and reported behavior in soils and water [6].
Figure 2 - Classification of pesticides by Gevao et al [6]. Clays and clay minerals are very special natural minerals with so many specific properties that they have originated a whole research world. In previous sections most of these properties have been shown but in any case most of these properties were known by the soil science studies. The interaction of clays and pesticides has been studied for many years. Cruz et al [7], Weber [8], Bailey and White [9], Mortland [10], White [11] and many others were the pioneer scientists who created the scientific basis of the actual knowledge of the new wonderful world of the applied clay science. 3 - The pesticide sorption process The natural process of sorption by soil solids, mainly those constituting the soil colloidal fraction, determines the amount of pesticide that can reach the target organism and the amounts available for other processes such as volatilization, degradation and leaching [12-14]. The sorption of pesticides by soil colloidal particles is also of interest in the transport of these compounds in runoff and surface waters, and even in ground waters, because this paniculate matter can act as a carrier of organic contaminants from point source [2,12,15-17]. As it was mentioned above, organic matter is recognized as the primary factor related to sorption of non-polar organic pesticides in soil or sediment/water systems [18,19], but for polar organic pesticides the behavior is not expected to be the same [20,21], specially for soils and sediments with low organic carbon content [22-29].
Pesticide-Clay Interactions and Formulations
251
Sorption of pesticides by diverse materials is also a method used for their elimination from water or their immobilization in contaminated soils [30,31] and a basis for controlled release formulations, which can decrease their contamination potential [32] and can be more economical than decomposing non-bioactive amounts of pesticide in the environment [33]. Recently, there has been an increasing interest in the use of natural and modified clays as supports or carriers to reduce the leaching of soil-applied herbicides or insecticides [34-40]. In this chapter we will first discuss pesticide-clay interactions, and second how these interactions can be used in pesticide formulations. 4 - Pesticide-clay interactions Soil clay fraction as a whole (mineral and organic components) has been shown to be the responsible for the sorption of many polar pesticides [26,28,41-45]. According to Mingelgrin and Gerstl [20], the mineral clay fraction of the soil increases its importance in pesticide soil sorption when organic matter content of the soils is low and when pesticides are ionic, ionizable or polar. For these pesticides, Hermosin et al [46] observed that variation coefficients decreased when sorption was considered in the clay fraction basis, indicating a more homogeneous behavior of the clay fraction in pesticide sorption as compared with the whole soil. For all pesticides considered in this study (Figure 3), variation coefficients were higher in the case of sorption coefficient Kj (Kj= pesticide sorbed/pesticide in solution) than in the case of sorption coefficient Kclay (Kclay= IQ/Vo clay * 100). For every pesticide considered, clay was the soil characteristic better or equal than organic carbon or organic matter correlated to pesticide soil sorption.
Figure 3 - Kj and K^ variation coefficients for selected pesticides (Ac= acephate, At= atrazine, Br~ bromacil, Ch= chlordimeform, Dq= diquat, Du~ diuron, Epeihylparaoxon, Ef= Ethofumesate, Fl=fluoridone, lm= imazethapyr).
252
J.Cornejoetal.
On the contrary, the organic matter associated to the colloidal fraction of soils seems to be less reactive in sorption of polar compounds [44,47]. This is clearly shown in Figure 4 for the triazinone herbicide metamitron: its affinity for Fe-saturated SWy montmorillonite is much higher than for commercial (Fluka) humic acid or soil humic acid [48]. Clay minerals play an important role in adsorption of polar organic pesticides mainly due to the high surface areas associated to their small particles size, and smectites are among all clay minerals the most important regarding sorption of organic compounds such as pesticides since they contribute most of the inorganic surface area in soils [25]. Smectites have swelling or expandable structure that makes the interlamellar surface accessible to interchange the cations and to water and polar organic molecules such as pesticides [9,17,43,49-54]. Clay minerals, due to isomorphous substitution in their structural framework, carry a permanent negative charge which, in their natural state, is compensated by inorganic cations. Hence, cationic pesticides can be adsorbed to clay minerals by ion exchange processes and sorption is directed by charge pattern interactions between cations and the surface charges of the clay mineral [33,55-57]. Anionic pesticides can also interact with clay minerals through positive edge charges (variable charge) of the silicate layers, through hydrogen bonds or attached to multivalent metal ions at exchange sites through cation bridges [33]. However, in general, anionic pesticides are weakly retained by most of soil or sediment components due to repulsion between clay minerals surface negative charges and organic anions [48,58,59]. Sorption capacity of clay minerals for hydrophobic pesticides is often considered considerably reduced, due to the highly hydrophilic environment provided by hydration water of exchangeable cations [10,60]. However, clay minerals also have "hydrophobic microsites" where pesticide molecules also sorb. In fact, even very highly hydrophobic organic pollutants, such as phenanthrene [61], have been shown to sorb on smectites. This was also proposed by Laird [62], who suggested that the herbicide atrazine initially sorbs on smectites as molecular species on these hydrophobic microsites on the clay siloxane surface. The same explanation was given by Celis et al [4] studying sorption of simazine and atrazine on model soil colloidal components. Moreover, van Oss and Giesse [63] used the surface thermodynamic theory to demonstrate the hydrophobicity of clay minerals. Figure 5 summarizes the different main sorption sites on a model smectite for model cationic (C), anionic (A), polar (P) or hydrophobic (H) pesticide molecules. The nature of the exchangeable cations of the clay mineral greatly influences sorption. The type of exchangeable cation adsorbed to the siloxane surface of kaolinite determined the adsorption of nitroaromatic compounds on kaolinite: significant adsorption was observed in the presence of weakly hydrated cations (i.e., Cs+, Rb+, K+, or NH4+) while strongly hydrated cations (H+, Na+, Ca2+, Mg2+ o Al3+) prevented any specific interaction [64]. Oxygen ligands at the external siloxane surface of kaolinite are accessible only in the presence of weakly hydrated cations. The same observation was made by these authors in sorption studies with other clay minerals such as illite and montmorillonite [65]. On another hand, sorption of polar molecules in the interlamellar spaces of montmorillonite by substitution of water molecules hydrating the exchangeable cation have been shown to be facilitated by small ionic potential of these cations [43,49,53,66,67], since the opening of the silicate layers is much easier and
Pesticide-Clay Interactions and Formulations
253
thereby the substitution of the hydration water molecules of the exchangeable cations by pesticide molecules is facilitated.
Figure 4 - Metamitron sorption isotherms on FeSWy montmorillonite and soil and commercial (Fluka) humic acids.
Figure 5 - Main sorption sites on a model smectite for model cationic (C), anionic (A), polar (P), and hydrophobic (H) pesticide molecules. An exception to this is the case of transition metals such as Fe3+ and Al3+, which has been shown to increase adsorption capacity of montmorillonite through the formation of coordination bonds between the electron donor groups of organic molecules and the metal ions and through protonation of groups such as NH. The last
254
J. Cornejo et al.
was observed for the triazine herbicide simazine [4] or the polar insecticide imidacloprid [54]. The high sorption of these pesticides on Fe saturated montmorillonite has been attributed to the acidic media provided by the saturating cation, which allows protonation of metamitron, simazine and imidacloprid molecules close to the montmorillonite surfaces and, thus, protonated molecules can easily be adsorbed on negatively charged surfaces on the interlayer space of the montmorillonite by a cation exchange mechanism. Figure 6 show this process in the case of simazine. The high polarizing power of the saturating cation allows protonation of simazine to H+-simazine which sorbs to a higher extent. Surface acidity also leads to further hydrolysis of sorbed simazine, since nucleophilic attack by water molecules is facilitated [4]. Iron oxides and humic acid coatings of the montmorillonite seem to favor this process, since they can act as proton donors [4].
Figure 6 - Protonation, sorption, and hydrolysis to hydroxysimazine of simazine molecule at the clay surface. Surface properties of the clay mineral determine their sorption capacity for pesticide molecules. Low CEC and layer charge of the clay mineral facilitates the opening of the silicate layer and the adsorption of polar organic molecules in the interlamellar space of these clay minerals [43,49,50,53]. Laird et al [68] found that atrazine sorption on 14 reference soil smectites ranged from 0 to 100 % depending on surface properties of the smectites and found a high negative correlation between sorption and surface charge density. This is clearly shown in Table 1 for three different organic compounds: two polar pesticides, the carbamate insecticide methomyl and urea herbicide thiazafluron, an a organotin biocide monobutyltin. Sorption of methomyl and thiazafluron is negatively correlated with layer charge and CEC (SAz montmorillonite > SWy montmorillonite> SH hectorite), whereas higher sorption of monobutyltin is observed in low layer charge montmorilloite SWy than in high layer charge smectite SAz. This table also shows the lack of sorption on kaolinite KGa measured in the case of the polar compounds and the lower sorption in the case of the MBT cation.
255
Pesticide-Clay Interactions and Formulations
Conversely to smectite, kaolinite has a rigid nonexpandable structure (Figure 7) which gives rise to a lower affinity for organic compounds including organic cations such as monobutyltin. Table 1 - Date obtained on different smectites calculated from Freundlich adjustment of isotherms. Smectite SSAf SHCa-1 SWy-Ca SAz-Ca Kga
(mV) 63 31 97 10
LC{ (molc unit cell"1) 0.31 0.68 1.13 0
CEC (mmolc kg"1) 439 764 1200 40
MET 24.66 10.21 6.96 0
Kf THIA 134.2 79.8 2.5 0
MBT nd§ 2332 1681 38
SSA = Specific surface area; LC = layer charge; CEC = cation exchange capacit. Sorption coefficients for: MET = carbamate insecticide methomyl [53], THIA = urea herbicide thiazafluron [43] and MBT = organotin cation monobutyltin [16]. f Van Olphen and Fripiat [69];} Jaynes and Boyd [70]; § not determined
Figure 7 - Structures of smectite and kaolinite. Many studies dealing with interactions of pesticides with pure minerals have been carried out in order to elucidate the possible binding mechanisms in soils [42,49,50,53,71], but very little information is available on adsorption of herbicides by the clays separated from soils, whose surface properties may be different than those for the pure minerals [25,42-44,47]. Although model adsorbents facilitate the study of pesticide adsorption, the actual mineral surfaces in soils can be interassociated with other soil components such as iron oxides or humic substances which might block adsorption sites on the clay minerals surfaces, giving rise to different adsorption capacities than that of pure minerals [43,48,72,73]. Association of iron oxide (ferrihydrite) to Ca-SWy montmorillonite increased sorption capacity of the clay mineral for the triazine herbicides simazine and atrazine [5]. This increase in sorption has been attributed to partially dissociated H2O molecules surrounding hydroxy Fe polymers in the interlayer of the montmorillonite promoting protonation of the herbicide molecules. The association of this montmorillonite with humic acid also increased sorption of these herbicides on the clay mineral and even the sorption capacity of the humic acid, which has been attributed to changes in
256
J.Cornejoetal.
conformation of the humic polymers increasing sorption on humic coatings [5]. Different results were obtained in the case of the urea herbicide thiazafluron of very high water solubility [74]. The association of iron oxide to clay mineral reduced sorption of thiazafluron by blocking the access to the interlamellar spaces of the smectite and by coating the external surface of montmorillonite. The same result was observed in the case of clay mineral association with humic acid. The high affinity of the highly polar thiazafluron for sorption sites in montmorillonite is reduced when montmorillonite surface is blocked by iron oxides and/or humic acid. In the case of the anionic herbicide 2,4-D [75], no sorption was measured on montmorillonite clay mineral due to repulsion between negatively charged clay surface and 2,-D anions, whereas ferrihydrite and humic acid coatings on the smectite surface provided sorption sites. The different behavior observed with these three different pesticides (Figure 8) reveals that the surface properties of soil clays are even more important than the relative amount present in soil, since these surfaces may not be accessible to organic compounds, due to interactions with other soil components such as humic substances, which may compete with pesticides for sorption sites [25,48,71,74,75].
Figure 8 - Molecular structures of the herbicides atrazine, thiazafluron, and2,4-D. 5 - Pesticide interactions with modified clays Because of the hydrophilic, negative character of their surfaces, clay minerals, in particular 2:1 phyllosilicates, have been shown to be very good adsorbents for cationic and highly polar pesticides, but their adsorption capacity for poorly soluble, nonionic pesticides is usually low [10,51,60,76]. The strong hydration of natural inorganic exchange cations produces a hydrophilic environment at the clay mineral surface that considerably reduces the sorptive capacity of clays for hydrophobic compounds. Replacement of natural metal-exchange cations with organic cations through ion exchange reactions has been shown to change the nature of the surface from hydrophilic to hydrophobic, and hence, this simple modification has been proposed for the improvement of the sorptive properties of clay minerals for hydrophobic organic compounds, including hydrophobic pesticides [77-79] (Figure 9). The organic cations most commonly used for clay mineral modification are quaternary ammonium ions of the general form [(CH3)3NR]+ or [(CH3)2NR2]+, where R is an aromatic or aliphatic hydrocarbon. Incorporation of large alkylammonium cations, such as octadecylammonium (ODA), dioctadecylammonium (DODA) and hexadecyltrimethylammonium (HDTMA), in the interlayers of smectitic clays has resulted in organoclays with enhanced affinity for neutral [38,80,81] and even acidic pesticides [51,77,82-85]. Celis et al [38] found that adsorption of the uncharged
Pesticide-Clay Interactions and Formulations
257
fungicide triadimefon by Arizona montmorillonite increased from 0 to > 90% after modification with HDTMA cations. Increases in adsorption after modification of montmorillonite with large alkylammonium cations have also been reported for acidic pesticides, such as 2,4-D [51,76], imazamox [85], bentazone [82], dicamba [83,84], and picloram [86] (Table 2), especially at low pH levels where the protonated form of the pesticide predominates [85]. It appears that the interlayer phase formed from large alkylammonium alkyl groups functions as a partition medium for nonionic organic compounds and effectively removes such compounds from water [60,77,87].
Figure 9 - Preparation of organoclays through ion exchange reactions. The adsorptive characteristics of organoclays formed using small quaternary ammonium cations, such as tetramethylammonium, are much different, since small organic cations exist as discrete species on the clay mineral surface and do not form an organic partition phase [78,87-89]. In these organoclays, the organic cations act as nonhydrated pillars that prop open the clay layers exposing the abundant siloxane surface area [89]. Low-charge montmorillonite modified with small alkylammonium cations were found to be particularly effective in removing alachlor, norflurazon, and hexazinone from aqueous suspensions [90,91]. Besides the size of the hydrocarbon chains, the surface charge of the clay mineral and the amount of organic cation in the interlayer have been shown to be major factors influencing the adsorptive properties of alkylammonium-exchanged clays (Table 2), because these parameters determine the arrangement of the organic cation in the clay mineral interlayer, and in turn the presence of space available to host pesticide molecules [38,60,85,87]. A recent research line on organoclays as adsorbents of pesticides is the selective modification of clay minerals with organic cations containing appropriate functional groups to maximize the affinity of the adsorbent for a given pesticide. Although this concept has been applied to develop organoclays with increased affinity and selectivity for heavy metal ions [92-94], little information exists on how organic cations with different functionalities influence pesticide adsorption by organoclays. Very recently, Nir et al [90] pointed out the importance of the structural compatibility between the pesticide molecule and the alkylammonium cation preadsorbed on the clay mineral in determining the performance of organoclays as adsorbents of pesticides. Alachlor and metolachlor, both containing a phenyl ring in their structure, displayed greater affinity for montmorillonite exchanged with organic cations with phenyl rings, such as benzyltrimethylammonium or phenyltrimethylammonium, than for alkylammonium-exchanged montmorillonites.
258
J. Cornejo et al.
Table 2 - Freundlich coefficients (Kf) for the adsorption of some acidic pesticides by unaltered and modified montmorillonite samples (data from references [83], [85] and [86])f. OctS (%)
Piclora m
Dicamb Kf Imazamox a
SWy-2 SAz-1 SWy-2 SWy-2 SWy-2
Main Interlayer cation Na+ Ca+ ODA ODA HDTMA
60 90 56
0 0
0 0
SWy-2
HDTMA
83
37
1
36
SAz-1
ODA
67
175
1
ODA HDTMA HDTMA
98 54 85
504 40 240
2 1 1
163 20 115 37 92
Sample
Montmorillonite
SW AS ODA-SW! ODA-SW2 HDTMA-
13 3 62 1 1
0 0 100 117
4
1 6 2
55
5
44
0 3
SWi
HDTMASW2 ODA-SAi
ODA-SA2 SAz-1 HDTMA-SA, SAz-1 HDTMA-SA2 SAz-1
2 1 2
5
167
2
352 77 272
1 0 24
SWy-2: Wyoming montmorillonite (CEC= 76 meq/WOg), SAz-1: Arizona montmorillonite (CEC= 120 meq/lOOg), ODA: octadecyl-ammonium, HDTMA: hexadecyltrimethyl-ammonium, OCtS: percentage of the CEC of the clay mineral compensated with organic cations.
Figure 10 - Simazine adsorption isotherms by untreated montmorillonite (SWy-2) and montmorillonite modified with different natural organic cations (data from reference 95).
Pesticide-Clay Interactions and Formulations
259
Based on this concept, Cruz-Guzman et al [95] hypothesized that natural organic cations with appropriate polar functional groups could allow one to selectively modify clay mineral surfaces to maximize their affinity for selected pesticides. Using Lcarnitine, L-cystine dimethyl ester and thiamine cations as modifying agents, the authors found that the chemical nature of the interlayer organic cation greatly influenced simazine adsorption by exchanged clays (Figure 10), probably through a combination of functionality and steric effects, and suggested the possibility to selectively modify clay mineral surfaces with organic cations containing appropriate functional groups to create an interlayer microenvironment designed to improve the affinity of the clay mineral for a given pesticide. The suitability of natural organic cations for this purpose was considered to be particularly interesting to minimize the environmental impact of the adsorbent when incorporated into natural ecosystems for practical applications. 6 - Pesticide-clay formulations The environmental problems associated with pesticide use, particularly the use of highly mobile pesticides, have become a current concern because of the increasing presence of these agrochemicals in ground and surface waters. To compensate for transport and degradation losses and to ensure adequate pest control for a suitable period, pesticides are applied at concentrations greatly exceeding those required for control of the target organisms, thus increasing the likelihood of runoff and leaching and hence the risk of surface and ground water contamination [96,97]. This problem is exacerbated in the case of highly soluble pesticides, because the risk of offsite movement from the intended target area increases as the pesticide is quickly dissolved in the soil solution [98]. Most pesticide formulations in current use deliver the bulk, if not all, of the active ingredient in an immediately available form that is readily released to the environment [99]. For highly soluble pesticides, these formulations may result in great pesticide losses shortly after application, before molecules have time to diffuse into soil aggregates and reach adsorption sites at the soil surfaces [39,100]. Recently, increased attention has been directed to reduce pesticide transport losses by the development of less hazardous formulations, such as slow-release formulations, in which only a part of the active ingredient is in an immediately available form; the bulk of the herbicide is trapped or sorbed in an inert support and is gradually released over time [101]. Beneficial effects related to the use of slow-release formulations include reduction in the amount of chemical required for pest control, decrease in the risk of environmental pollution as a result of pesticide transport losses (i.e., leaching or volatilization) (Figure 11), savings in manpower and energy by reducing the number of applications required in comparison to conventional formulations, increased safety for the pesticide applicator, and a general decrease in nontarget effects [97]. Among the various materials proposed as carriers in pesticide formulations [102], recently there has been a renewed interest in the use of natural soil constituents, like clays, iron oxides or humic acids [34,35, 96,98,99,103-105]. Clays have unique properties, such as their high specific surface areas associated with their small particle size, low cost, and ubiquitous occurrence in most soil and sediment environments. Margulies et al [34] used montmorillonite and sepiolite as carriers for the volatile herbicide S-ethyldipropylcarbamothioate (EPTC) and showed that volatilization losses of EPTC were greatly reduced by the use of the clay minerals as pesticide carriers. At 30°C, the half-life time (T1/2) of EPTC in its free form was 10 h, whereas when adsorbed
260
J. Cornejo et al.
to montmorillonite the r M was more than 5 days. Similarly, when EPTC was incorporated into soil, the TI/2 values were 4 and 9 days for the free and adsorbed forms, respectively. Control release of atrazine and alachlor was achieved by Johnson and Pepperman [99] by addition of selected minerals, including calcium bentonite, finegrind bentonite, montmorillonite K10, kaolinite, and iron oxide. Cox et al [39] used Fe(III)-treated Wyoming montmorillonite to enhance the affinity of the clay mineral for simazine and 2,4-D, and prepared clay-pesticide complexes that displayed reduced leaching of the pesticides in hand-packed soil columns compared to the free form of the pesticides. Similar results were reported by Celis et al [86,91] for the herbicides picloram and hexazinone preadsorbed on Fe(HI)-montmorillonite.
Figure 11 - Pesticide transport processes as affected by the formulation. Application as a clay complex enhances the soil sorption process, reducing pesticide losses by volatilization and leaching. In addition to the use of clay minerals as such, the possibility exists to selectively modify clay mineral surfaces, for instance through the incorporation of organic cations in the interlayers, to improve their adsorption capacity for selected pesticides and to control the desorption rate once added to the environment. Different types of pesticide formulations based on organoclays, particularly alkylammoniumexchanged clays, have been proposed [33,38,40,82,83,105,106]. However, while much attention has been given to describe the diversity of organoclay-pesticide interactions [51,75,77,81-84,86,90,107], information on the factors controlling the release rate and extent of pesticides from organoclay formulations is less abundant. Particularly limited are studies that validate the behavior of organoclay formulations of pesticides under real, field conditions, although the comparable or even improved weed control efficacy reported by several authors for organoclay-based formulations of herbicides compared to standard formulations seems promising for practical application of these formulations [37,108-110]. One major obstacle of many clay-pesticide formulations assayed has been observed to be the high amount of pesticide not released from the formulation, which would result in soil contamination and would require higher application rates for the same amount of active ingredient to be released [91,110]. Formulations of hexazinone
Pesticide-Clay Interactions and Formulations
261
supported on Fe(III)-montmorillonite released from 75% to only 20% of the herbicide associated with the clay mineral [91]. Gerstl et al [96] also observed that a large portion (35-70%) of the active ingredient remained adsorbed in controlled release formulations of alachlor at the end of release experiments. Similarly, Johnson and Pepperman [99] found that from 5 to 27% of alachlor and atrazine formulated with bentonite, montmorillonite, kaolinite and iron (III) oxide was not released, with the greatest retention by bentonite formulations. To circumvent the above-mentioned limitation, factors influencing the release rate and extent of pesticides from clays and organoclays need to be considered. The characteristics of the clay mineral, the amount and nature of exchangeable cations, the clay-pesticide ratio, and the procedure followed to prepare the formulation all affect the interaction of the pesticide with the sorbent and in turn the release rate and extent from the resulting formulation [37,40,96,91,105-107,110]. Thus, El-Nahhal et al [37] found that organoclay complexes with 0.5 mmol benzyltrimethylammomum/g of montmorillonite gave larger adsorbed amounts and better formulations of alachlor as compared to benzyltrimethylammonium preadsorbed up to the cation exchange capacity of the clay (0.8 mmol/g). Hermosin et al [105] showed that the release of fenuron from organoclays was inversely proportional to the adsorbent power of the organoclay and the fenuron-organoclay mixing time. Similarly, loosely-bound preparations of hexazinone with hexadecyltrimethyl-ammonium-montmorillonite (i.e., simple mechanical mixtures) have been shown to lead to greater amounts of herbicide released than formulations where an intimate association was promoted by the use of organic solvent [91].
Figure 12 - Hexazinone breakthrough curves in hand-packed soil columns after application as commercial formulation and HDTMA-montmorillonite formulations containing 20%, 10%, and4% active ingredient (a.i.).
262
J. Cornejo et al.
In a recent study, Celis et al [110] found hexazinone formulations based on hexadecyltrimethylammonium-montmorillonite displayed slow release properties in water, retarded herbicide leaching through soil columns and maintained a herbicidal efficacy similar to that of the free form of the herbicide, whereas formulations based on phenyltrimethylammonium-montmorillonite released the herbicide instantaneously and did not display slow release properties. High organoclay-herbicide ratios made the interaction of the herbicide with the organoclay more intimate and reduced the release rate of hexazinone as well as its leaching through soil columns (Figure 12). Therefore, the possibility exists to select all these variables to optimize the performance of the formulation for practical applications. 7 - Summary and conclusions Clay-pesticide interactions are considered under two complementary points of view: pesticide interactions with the soil clay minerals and the design of clay-based formulations for an improvement of the agronomical and environmental use of pesticides. Clay minerals increase in importance on pesticide soil sorption when organic matter content of the soils is low and when pesticides are ionic, ionizable or polar. Smectites are among all clay minerals the most important regarding sorption because they have swelling or expandable structure that makes the interlamellar surface accessible to interchange the cations, water and pesticide molecules. Surface properties of the clay minerals determine their sorption capacity for pesticide molecules. Although model sorbents facilitate the study of pesticide adsorption, the actual mineral surfaces in soils can be interassociated with other soil components such as iron oxides or humic substances which might block adsorption sites on the clay minerals surfaces, giving rise to different adsorption capacities than that of pure minerals. The surface properties of soil clays are even more important than the relative amount present in soil, since these surfaces may not be accessible to organic compounds, due to interactions with other soil components which may compete with pesticides for sorption sites. Replacement of natural metal-exchange cations with organic cations through ion exchange reactions has been shown to change the nature of the surface from hydrophilic to hydrophobic increasing clay mineral affinity for hydrophobic organic compounds, including hydrophobic pesticides. A recent research line on organoclays as adsorbents of pesticides is the selective modification of clay minerals with organic cations containing appropriate functional groups to maximize the affinity of the adsorbent for a given pesticide. These organoclays can be used in clay-pesticide formulations. Factors influencing the release rate and extent of pesticides from clays and organoclays are still issues which need to be addressed. The characteristics of the clay mineral, the amount and nature of exchangeable cations, the clay-pesticide ratio, and the procedure followed to prepare the formulation all affect the interaction of the pesticide with the sorbent and in turn its release from the formulation. 8 - References [1] R.G. Nash and A.R. Leslie, Eds., Groundwater Residues Sampling Design, American Chemical Society, Washington D.C., 1991. [2] J. Cornejo, P. Jamet and F. Lobnik . Pesticide/soil interactions: Some current research methods: Introduction, Eds. J. Cornejo and P. Jamet, INRA, Paris, 2000.
Pesticide-Clay Interactions and Formulations
263
[3] J. Comejo and M.C. Hermosin, Humic Substances in Terrestrial Ecosystem: Interaction of humic substances and soil clays, Ed. A. Piccolo, Elsevier, Amsterdam, 1996. [4] R. Celis, J. Cornejo, M.C. Hermosin and W.C. Koskinen, Soil Sci. Soc. Am. J., 61 (1997) 436. [5] R. Celis, J. Cornejo, M.C. Hermosin and W.C. Koskinen, Soil Sci. Soc. Am. J., 62 (1998) 165. [6] B. Gevao, K.T. Semple and K.C. Jones, Environ. Pollut, 108 (2000) 3. [7] M. Cruz, J.L. White and J.D. Russell, Israel J. Chem., 6 (1968) 315. [8] J.B. Weber, Res. Rev., 32 (1970) 93. [9] G.W. Bailey and J.L. White, Res. Rev., 32 (1970) 29. [10] M.M. Mortland, Adv. Agron., 23 (1970) 75. [11] J.L. White, Bound and Conjugated Pesticide Residues: Clay-particle interactions, Eds. Kaufman et al, ACS Symposium Series. Washington, 1970. [12] J.F. Mcarthy and J.M. Zachara, Environ. Sci. Technol., 23 (1989) 497. [13] A.J. Beck, A.E.J. Johnston and K.C. Jones, Crit. Rev. Env. Sci. Technol., 23 (1993) 219. [14] J. Cornejo and P. Jamet, Eds., Pesticides/soil interactions: Some current research methods, INRA, Paris, 2000 [15] M.C. Hermosin, J. Cornejo, J.L. White and F.D. Hess, J. Agric. Food Chem., 30 (1982) 728. [16] M.C. Hermosin, P. Martin, and J. Cornejo, Environ. Sci. Technol., 27 (1993) 2606. [17] L. Cox , M.C. Hermosin, W.C. Koskinen and J. Cornejo, Clay Miner., 36 (2001) 267. [18] S.W. Karickhoff, D.S. Brown and T.A. Scott, Water Res., 13 (1979) 241. [19] C.T. Chiou, SSSA Special Publication 22: Reactions and Movement of Organic Chemicals in Soils, Eds. B.L. Sawney et al, ASA and SSSA, Madison, 1989. [20] U. Mingelgrin and Z. Gerst, J. Environ. Qual., 12 (1983) 1. [21] U.A.T. Brinkman, Environ. Sci. Tecnol, 29 (1995) 79. [22] K.S. Reddy and R.P. Gambrell, Agric. Ecosyst. Environ., 18 (1987) 231. [23] R. Calvet, Environ. Health Persp., 88 (1989) 147. [24] W.R. Roy and I.G. Krapac, J. Environ. Qual., 23 (1994) 549. [25] D.A. Laird, P.Y. Yen, W.C. Koskinen, T.R. Steinheimer and R.H. Dowdy, Environ. Sci. Technol., 26 (1994) 1054. [26] I. Roldan, M.C. Hermosin and J. Cornejo, Sci. Total Environ., 132 (1993) 217. [27] L. Cox, M.C. Hermosin and J. Cornejo, Int. J. Environ. Anal. Chem., 58 (1995) 305. [28] L. Cox, M.C. Hermosin and J. Cornejo, Chemosphere, 32 (1996) 1381. [29] M.C. Fernandes, L. Cox, M.C. Hermosin and J. Cornejo, Pest Manag. Sci., 59 (2003) 545. [30] F.H. Yelverton, J.B. Weber, G. Peedin and W.D. Smith, Pub. AG-442 Agricultural Extension Service North Carolina State University, Raleigh, NC, 1990. [31] J. Wagner, H. Chen, B.J. Brownawell and J.C. Westall, Environ. Sci. Technol.,. 28 (1994)231. [32] D.J. Park, W.R. Jackson, I.R. McKinnon and M. Marzhall, Controlled-Release Delivery Systems for Pesticides, Ed., H.B. Scher, Marcel Dekker Inc, New York, 1999. [33] G. Lagaly, Appl. Clay Sci., 18 (2001) 205.
264
J. Cornejo et al.
[34] L. Margulies, T. Stern and B. Rubin, J. Agric. Food Chem., 46 (1994)1223. [35] E. Gonzalez-Pradas, M. Fernandez-Perez, M. Villafranca-Sanchez, F. MartinezLopez and F. Flores-Cespedes, Pestic. Sci., 55 (1999) 546. [36] M.C. Hermosin, J.P. Aguer, J. Cornejo and M.J. Calderon, The 1 lth International Clay Conference "Clays for our future". Book of Abstracts, Ottawa, 1997. [37] Y. El-Nahhal, S. Nir, T. Polubesova, L. Margulies and B. Rubin, J. Agric. Food Chem., 46 (1998) 3305. [38] R. Celis, W.C. Koskinen, M.C. Hermosin, M.A. Ulibarri and J. Cornejo, Soil Sci. Soc. Am. J., 64 (2000) 36. [39] L. Cox, R. Celis, M.C. Hermosin and J. Cornejo, J. Agric. Food Chem., 48 (2000) 93. [40] A. Nennemann, Y. Mishael, S. Nir, B. Rubin, T. Polubesova, F. Bergaya, H. van Damme and G. Lagaly, Appl. Clay Sci., 18 (2001) 265. [41] M.C. Hermosin and J. Cornejo, Toxicol. Environ. Chem., 25 (1989) 45. [42] L. Cox, M.C. Hermosin and J. Cornejo, Chemosphere, 27 (1993) 837. [43] L. Cox, M.C. Hermosin and J. Cornejo, Europ. J. Soil Sci., 46 (1995) 431. [44] E. Barriuso, P. Benoit and V. Bergheaud, Environmental Behavior of Pesticides and Regulatory Aspects: Role of soil fractionation in retention and stabilisation of pesticides in soils, Eds. A. Copin et al, COST 66 ACTION, Europen Study Service, Belgium, 1994 [45] M.J. Sanchez-Martin, L.R. Lorenzo, T. Crisanto, M. Arienzo, and M. SanchezCamazano, Commun. Soil Sci. Plant Anal., 26 (1995) 3243. [46] M.C. Hermosin, J. Cornejo and L. Cox, Pesticide/soil interactions: Some current research methods: Calculation and validation of Kclay as predictor for polar or ionic pesticide adsorption by soils, Eds. J. Cornejo and P. Jamet, INRA, Paris, 2000. [47] L. Cox, M.C. Hermosin, R. Celis and J. Cornejo, Water Res., 31 (1997) 1306. [48] R. Celis, L. Cox, M.C. Hermosin and J. Cornejo, Intern. J. Environ. Anal. Chem., 65(1996)245. [49] M.J. Sanchez-Martin and M. Sanchez-Camazano, Chemosphere, 16 (1987) 937. [50] M.C. Hermosin, I. Roldan and J. Cornejo, Chemosphere, 23 (1991) 473. [51] M.C. Hermosin and J. Cornejo, J. Environ. Qual, 22 (1993) 325. [52] G. Lagaly, Progr. Colloid Polym. Sci., 95 (1994) 61. [53] L. Cox, L., M.C. Hermosin and J. Cornejo, Clay Miner., 29 (1994) 767. [54] L. Cox, W.C. Koskinen, R. Celis, P.Y. Yen, M.C. Hermosin and J. Cornejo, Soil Sci. Soc. Am. J., 62(1998)911. [55] M.C. Hermosin and J.L. Perez-Rodriguez, Clays Clay Miner., 29 (1981) 143. [56] G. Lagaly, Development of ionic polymers, vol. 2: Smectitic clays as ionic macromolecules, Eds. A.D. Wilsom and H.J. Prosser, H.J., Elsevier, London, 1986. [57] G.A. Sezer, A. G. Tilrkmenolu and E. H. Goktiirk, Appl. Geochem., 18 (2003) 711. [58] D.D. Fontaine, R.G. Lehmann and R.J. Miller, J. Environ. Qual., 20 (1991) 759. [59] R.D. Wauchope, T.M. Buttler, P.W.M. Hornsby, A. Beckers, A. and J.P. Burt, Rev. Environ. Contam. Toxicol., 123 (1992) 1. [60] W.F. Jaynes and G.F. Vance, Soil Sci. Soc. Am. J., 60 (1996) 1742. [61] L.S. Hundal, M.L. Thompson, D.A. Laird and A.M. Carmo, Environ. Sci. Technol, 35(2001)3456. [62] D.A. Laird, ACS Symp. Ser., 630 (1996) 86.
Pesticide-Clay Interactions and Formulations
265
[63] C.J. van Oss and R.F. Giese, Clays Clay Miner., 43 (1995) 474. [64] S.B. Haderlein, S.B. and R.P. Schwarzenbach, Environ. Sci. Technol., 27 (1993) 316. [65] S.B. Haderlein, K.W. Weissmahr and R.P. Schwarzenbach, Environ. Sci. Technol., 30(1996)612. [66] P. Fusi, M. Franci and G.G. Ristori, Appl. Clay Sci., 4 (1989) 403. [67] M. Bosetto, P. Alfaioli and P. Fusi, Soil Sci., 155 (1993) 105. [68] D.A. Laird, E. Barriuso, R.H. Dowdy and W.C. Koskinen, Soil Sci. Soc. Am. J., 56 (1992)62. [69] H. Van Olphen and J.J. Fripiat, J.J., Eds., Data Handbook for Clay Materials and other Non-Metallic Minerals; Pergamon Press, Oxford, 1979. [70] W.F. Jaynes and S.A. Boyd, Soil Sci. Soc. Am. J., 55 (1991) 43. [71] A. Pusino, W. Liu and C. Gessa, Pest. Sci., 36 (1992) 283. [72] J.M. Oades, Soil Colloids and Their Association in Aggregates: Association of colloids in Soil Aggregates, Eds. De Bodt et al, NATO ASI Series; New York and London, 1990. [73] J. Cornejo and M.C. Hermosin, Environmental Behavior of Pesticides and Regulatory Aspects: Role of soil clay fraction in pesticide adsorption: defining a Kclay, Eds. A. Copin et al COST 66 ACTION, Europen Study Service, Belgium, 1994. [74] R. Celis, L.Cox, M.C. Hermosin and J. Cornejo, J. Environ. Anal. Chem., 26 (1997)472. [75] R. Celis, M.C. Hermosin, L. Cox and J. Cornejo, Environ. Sci. Technol., 33 (1999) 1200. [76] M.C. Hermosin and J. Cornejo, Chemosphere, 24 (1992) 1493 [77] S.A. Boyd, S. Shaobai, J.F. Lee and M.M. Mortland, Clays Clay Miner., 36 (1988) 125. [78] J.F. Lee, J.R. Crum and S.A. Boyd, Environ. Sci. Technol., 23 (1989) 1365. [79] S.L. Lemke, P.G. Grant and T.D. Phillips, J. Agric. Food Chem., 46 (1998) 3789. [80] J.P. Aguer, M.C. Hermosin, M.J. Calder6n and J. Cornejo, J.Environ. Sci. Health B., 35 (2000) 279. [81] L. Cox, W.C. Koskinen, M.C. Hermosin, J. Cornejo and E.L. Arthur, Intern. J. Environ. Anal. Chem., 82 (2002) 553. [82] M.J. Carrizosa, M.J. Calderon, M.C. Hermosin and J. Cornejo, Sci. Total Environ., 247 (2000) 285. [83] M.J. Carrizosa, W.C. Koskinen, M.C. Hermosin and J. Cornejo, Appl. Clay Sci., 18(2001)223. [84] H. Zhao, W.F. Jaynes and G.F. Vance, Chemosphere, 33 (1996) 2089. [85] R. Celis, W.C. Koskinen, A.M. Cecchi, G.A. Bresnahan, M.J. Carrizosa, M.A. Ulibarri, I. Pavlovic and M.C. Hermosin, J. Environ. Sci. Health B., 34 (1999) 929. [86] R. Celis, M.C. Hermosin, L. Cornejo, M.J. Carrizosa and J. Cornejo, Intern. J. Environ. Anal. Chem., 82 (2002) 503. [87] W.F. Jaynes and S.A. Boyd, Clays Clay Miner., 39 (1991) 428. [88] J.F. Lee, M.M. Mortland, C.T. Chiou, D.E. Kile and S.A. Boyd, Clays Clay Miner., 38(1990)113. [89] J.M. Brixie and S.A. Boyd, J. Environ. Qual., 23 (1994) 1283.
266
J. Cornejo etal.
[90] S. Nir, T. Undabeytia, D. Yaron-Marcovich, Y. El-Nahhal, T. Polubesova, C. Serban, G. Rytwo, G. Lagaly and B. Rubin, Environ. Sci. Technol, 34 (2000) 1269. [91] R. Celis, M.C. Hermosin, M.J. Carrizosa and J. Comejo, J. Agric. Food Chem., 50 (2002) 2324. [92] L. Merrier and C. Detellier, Environ. Sci. Technol., 29 (1995) 1318. [93] G. Sheng, S. Xu and S.A. Boyd, Soil Sci. Soc. Am. J., 63 (1999) 73. [94] R. Celis, M.C. Hermosin and J. Cornejo, Environ. Sci. Technol., 34 (2000) 4593. [95] M. Cruz-Guzman, R. Celis, M.C. Hermosin and J. Cornejo, Environ. Sci. Technol., (in press). [96] Z. Gerstl, A. Nasser and U. Mingelgrin, J. Agric. Food Chem., 46 (1998) 3797. [97] Z. Gerstl, A. Nasser and U. Mingelgrin, J. Agric. Food Chem., 46 (1998) 3803. [98] M. Fernandez-Perez, E. Gonzalez-Pradas, M. Villafranca-Sanchez, F. FloresCespedes and M.D. Urena-Amate, J. Environ. Qual., 29 (2000) 304. [99] R.M. Johnson and A.B. Pepperman, Pestic. Sci., 53 (1998) 233. [100] T.J. Gish, M.J. Schoppet, C.S. Helling, A. Schirmohammadi, M.M. Schnecher and R.E. Wing, Trans. ASAE, 34 (1991) 1738. [101] M. Fernandez-Perez, E. Gonzalez-Pradas, M.D. Urefta-Amate, R.M. Wilins and I. Lindrup, J. Agric. Food Chem., 46 (1998) 3828. [102] H.B. Scher, Ed., Controlled-Release Delivery Systems for Pesticides, Marcel Dekker Inc., New York, 1999. [103] M.E. Carr, R.E. Wing and W.M. Doane, Starch/Staerke, 46 (1994) 9. [104] Y. El-Nahhal, S. Nir, C. Serban, O. Rabinovitch and B. Rubin, J. Agric. Food Chem., 48 (2000) 4791. [105] M.C. Hermosin, M.J. Calderon, J.P. Aguer and J. Cornejo, Pest. Manag. Sci., 57 (2001) 803. [106] T. Undabeytia, S. Nir and B. Rubin, J. Agric. Food Chem., 48 (2000) 4767. [107] M.J. Carrizosa, M.C. Hermosin, W.C. Koskinen and J. Cornejo, Soil Sci. Soc. Am. J., 67(2003)511. [108] Y. El-Nahhal, S. Nir, T. Polubesova, L. Margulies and B. Rubin, Pestic. Sci., 55 (1999)857. [109]Y. El-Nahhal, Environ. Pollut, 124 (2003) 33. [110] R. Celis, G. Facenda, M.C. Hermosin and J. Cornejo, Pest Manag. Sci., (2004) (submitted).
PHARMACEUTICAL AND COSMETIC APPLICATIONS OF CLAYS ALBERTO LOPEZ-GALINDO "' and CESAR VISERAS 2 1
Instituto Andaluz de Ciencias de la Tierra (CSIC-UGR). Facultad de Ciencias, Campus Fuentenueva. 18071 - Granada - SPAIN. 2 Departamento de Farmacia y Tecnologia Farmaceutica, Facultad de Farmacia. Universidad de Granada. 18071, Granada - SPAIN. E-mail: [email protected] * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
268
A. López-Galindo and C. Viseras
1 - Introduction Clays are substances found throughout the earth's surface, as they are the main component of soils and pelitic sedimentary rocks. Because of their frequency of occurrence and their particular properties, they have been used by man since prehistoric times for therapeutic purposes, such as to cure wounds, relieve irritations or treat gastrointestinal disorders. In Europe, Asia, Africa and America, most ancient civilizations used some form of clays in this manner, the best known examples being those of Mesopotamia, Egypt, Greece and Rome as they were mentioned by numerous classical authors. The medicinal "earths" were normally named according to their place of origin, and were thus known as Egyptian, Nubian, Lemnian, Samian, Cimolian earths, Armenian bole, etc. Lemnian earth, from the Greek island of Lemnos, can be considered the first medicine recorded in history [1] and was in use until the beginning of the last century. Its importance is reflected in its being mentioned, among others, by Homer, Theofrastes, Pliny the Elder, and Galenus, who twice travelled to Lemnos in the Aegean Sea to study its preparation. In the Middle Ages the Arabs added new varieties to those familiar to the Greco-Roman world, with significant contributions by Avicena and Averroes. Later, both the Spanish king Alfonso X the Wise, in the collection of his texts and previous translations known as the Lapidario, and Agricola, in his De Re Metallica, dedicated extensive chapters to the properties and applications of medicinal earths. During the Renaissance, when the first Pharmacopoeia appeared, the use of these clays was regulated to a certain extent. In modern times, with the change of mentality brought about by scientific and technological progress, their use has become considerably more restricted, although they continue to be used as natural remedies for the prevention, relief or cure of certain pathologies of the skin, inflammations, dislocations, contusions and the treatment of wounds. Those interested can find in literature numerous examples of such applications in historic times for both health and beauty [2-6]. The development of some branches of sciences such as Mineralogy, Chemistry and Pharmacy in the 18th and 19th centuries was decisive for understanding the nature of these materials. But it was not until the beginning of the 20th century, with the improvement in instrumental techniques, in particular the discovery of X-ray diffraction, when the causes of the singularly useful properties of clay began to be understood. This are related to directly to their small particle size and their crystalline structure, which makes them suitable for application as absorbent, sterilising, antiinflammatory and detergent substances. At present, environmental awareness and interest in the use of natural products has led to an increase in the use of clayey geomaterials in medical and thermal treatment. In Europe, where the greatest number of spas and therapeutic centres using clays is found, Italy is probably the country with the longest tradition and most frequent use of these materials. For this reason courses and scientific meetings have recently been organised there on the subject [3,4] and where the protocols and norms qualifying the different materials used in fangotherapy are drawn up [7,8]. We should here point out that there is some confusion in the literature regarding the terms "clay mineral" and "clay". The former is a mineralogical term referring to part of a family (the phyllosilicates) consisting of hydrated aluminosilicates containing considerable amounts of Mg, K, Ca, Na and Fe and, occasionally, less
Pharmaceutical and Cosmetic Applications of Clays
269
common ions such as Ti, Mn, or Li. Despite their varied chemical composition, they can be classified in just a few major groups - smectites, micas, kaolin, talcum, chlorites, vermiculites, fibrous and interstratified. The word "clay", on the other hand, refers to natural materials composed of very fine-grained minerals, with some plasticity when mixed with water and which harden on drying. It is, therefore, applicable to all smallsized particles, normally < 2 (im, found in soils or sediments, including, apart from the phyllosilicates mentioned above, other minerals and/or organic products such as quartz, feldspars, carbonates, sulphates, Fe and/or Al oxides, humus, etc. The expression "healing clays" applies mainly to the second term and refers, therefore, to natural clays that, after appropriate treatment to bring out a particular property, are used for pelotheraphy in spa centres. 2 - Structure and texture Clay minerals are among the most widely used materials in pharmaceutical formulation, because of their properties as excipients and/or their biological activities [2,6,9-14). These features depend on both their colloidal dimensions and high surface areas (basic properties), resulting in optimal rheological characteristics and/or excellent sorption capacities. For these reasons, clays have been used for many years in the formulation of solid (tablets, capsules, and powders), liquid (suspensions, emulsions) and semisolid (ointments, creams) dosage forms, either for oral or topical administration. Only some clay minerals are used in pharmacy, including kaolin, talc, smectites (montmorillonite and saponite), and fibrous clays (palygorskite and sepiolite). The kaolin group is a family including kaolinite, halloysite, dickite and nacrite, of which kaolinite is the most common mineral, so that kaolin and kaolinite frequently become synonymous [13]. The smectite group includes, among others, montmorillonite, beidellite, nontronite and saponite, although rocks containing montmorillonite as main mineral are also referred to as bentonites [15]. Finally the palygorskite-sepiolite group includes two minerals - palygorskite (often known as attapulgite) and sepiolite. The particular use of a clay mineral for any specific pharmaceutical application depends firstly on its structure. The structural unit of clay minerals consists of a combination of Al or Mg octahedra and Si tetrahedra, resulting in layered structures that may be organised as consecutive strata of octahedral and tetrahedral sheets (T:O or 1:1 clays), or structures with one octahedral sheet "sandwiched" between two tetrahedral ones (T:O:T or 2:1 clays), allowing for an initial classification. The main difference in the behaviour of these two classes is their performance when dispersed in polar solvents. 1:1 clays do not swell, whereas 2:1 ones do, creating highly structured systems with interesting rheological properties. Further distinction can be made on the basis of chemical differences. In some of these minerals, isomorphic substitution in the octahedral or tetrahedral layers creates negative charges compensated by exchangeable ions in the interlayer space. The swelling properties of clay minerals are strongly affected by the type and hydration grade of the predominant exchangeable ion [16]. Finally, textural differences between structurally and chemically identical minerals affect their adsorptive and rheological properties [17]. As a result of their structural and chemical characteristics, both kaolinite (1:1 layered silicate of Al) and talc (1:1 layered silicate of Mg) show minimal layer charges, presenting low cation-exchange capacities (< 15-20 mEq/lOOg). On the other hand,
270
A. López-Galindo and C. Viseras
smectites are 2:1 layered silicates, characterised by octahedral and tetrahedral substitutions and high ion-exchange capacities (100-200 mEq/lOOg). Differences in the number of cations in octahedral sites lead to the division of smectites into di- and trioctahedral groups, montmorillonite falling into the first group and saponite into the second. Finally, sepiolite and palygorskite are 2:1 phyllosilicates, but, unlike other clay minerals, they have a fibrous morphology resulting from the 180° inversion occurring every six (sepiolite) or four (palygorskite) silicon tetrahedra, causing a structure of chains aligned parallel to the "a" axis, each of which has a 2:1 structure. This threedimensional ordering also causes open channels measuring 3.7 x 6.4 A (palygorskite) and 3.7 x 10.6 A (sepiolite) and containing zeolitic and crystallization water. Sepiolite has a BET surface area of approximately 300 m2/g and palygorskite 120-180 m2/g. These values can increase as the adsorbed and zeolitic water evaporates when the mineral is heated. Figure 1 shows the most common morphology of these minerals when observed under scanning electron microscope (SEM).
Figure 1 - Usual clay morphology observed by SEM. A) Smectite; B) Talc; C) Kaolinite; D) Palygorskite Although all the particles are small-size (always < 5 microns) the differences between the different types of phyllosilicates are clear. Smectites (A) are usually present
Pharmaceutical and Cosmetic Applications of Clays
271
as scarcely differentiated planar particles with quite irregular edges and a size of less than 2 microns, while the particles of talcum (B) are clearly individualised, with clear edges and a quite larger size. The particles of kaolinite (C) are better crystallised, with abundant pseudo hexagonal shapes, while palygorskite samples (D) are normally made up of large aggregates containing randomly arranged microfibres, although this mineral is sometimes found as small fibres covering other crystals. As a result of these basic characteristics, clay minerals are used in numerous industrial applications involving ceramics, plastics, paper, paint, catalysis, cosmetics, etc., as reviewed in the literature [18-26]. 3 - Use in pharmaceutical formulations 3.1 - Pharmaceutical denominations Both the European Pharmacopoeia (EP) and the United States Pharmacopoeia (USP) contain monographs regarding clay mineral materials. In the EP 4th [27], official monographs of "Aluminium Magnesium Silicate", mixture of montmorillonite and saponite (a), "Bentonite" (montmorillonite, b), "Kaolin" (c), "Magnesium Trisilicate" (sepiolite, d)) and "Talc" (e) are included. Besides these monographs the USP 25 [28] adds the following: "Activated Attapulgite" (f), "Alumina and Magnesium Trisilicate Oral Suspensions" (g), "Alumina and Magnesium Trisilicate Tablets" (h), "Bentonite Magma" (i), "Colloidal Activated Attapulgite" (j), "Magnesium Trisilicate Tablets" (k) and "Purified Bentonite" (1). Table I summarises the clay minerals included in EP 4th and USP 25, with their chemical and commercial correspondences. Some ambiguities between mineralogical, chemical and pharmaceutical names can be observed. The term "bentonite" is too generic, as it can be used both for a rock consisting mainly of smectites (mineralogy) or a material mainly containing montmorillonite (pharmacy). On the other hand, "Aluminium Magnesium silicate" (EP 4th) and "Magnesium Aluminum silicate" (USP 25) are not univocal names, creating some confusion. Both mainly refer to blends of montmorillonite and saponite, according to their specific Al/Mg ratio (between 95 and 105% w/w of that stated on the label). Moreover, palygorskite samples are frequently commercialised under these denominations, as well as attapulgite. Finally, sepiolite seems to correspond to the so-called "Magnesium Trisilicate", described as a blend of Si and Mg oxides prepared to meet the pharmacopoeia requirements. USP 25 requires not less than 20% w/w of MgO and not less than 45% w/w of SiO2, where EP 4* indicates not less than 29% w/w of MgO and not less than 65% w/w of SiO23.2 - Pharmaceutical specifications As intended for use in the preparation of medicines, clay mineral materials must fulfil certain requirements concerning their chemistry (stability and high chemical inertia), physical characteristics (texture, water content, dimensions) and toxicological nature (chemical safety, microbiological purity). Some of these properties, such as those regarding safety and stability, are vital. It must be remarked that minerals intended for use as pharmaceutical materials may contain crystalline silica (both quartz and cristobalite), that should be controlled and avoided as far as possible, as it is classified by the International Agency for Research on Cancer (IARC) as a product with sufficient evidence of carcinogenicity in laboratory animals and limited evidence in humans (group 1, IARC Monographs) [29]. On the other hand, amorphous silica, found in nature as biogenic silica (e.g.
272
A. López-Galindo and C. Viseras
diatomaceous earth) and as silica glass (volcanic genesis), is not classifiable as carcinogenic to humans (group 3) [29]. Regarding kaolin, USP 25 specifies that it must be "powdered and freed from gritty particles by elutriation". Impurities such as quartz, mica, hematite or pyrite are mainly contained in the coarse fraction of the rock, and should be eliminated. Concomitant administration of drugs, such as some antibiotics (amoxicillin, ampicillin, clindamycin), cimetidine, atropine, phenytoin, digoxin and quinidine could reduce drug absorption as a result of drug-kaolin interaction, and should be avoided [30-36]. Talc, presented as a white or almost white, impalpable and unctuous powder, may contain variable amounts of other minerals, such as hydrated Al silicate, magnesite, calcite and dolomite that can remain when used as excipient. Talc containing asbestos is, however, not suitable for pharmaceutical use because of its carcinogenic activity in humans. According to the IARC monograph [37,38], talc not containing asbestiform fibres is not classifiable as to its carcinogenicity to humans (group 3), while there is sufficient evidence for the carcinogenicity to humans of talc containing asbestiform fibres (group 1). In fact, the EP 2002 monograph on talc includes specific tests (infrared, X-ray diffractometry and optical microscopy analysis) to detect asbestos and to determine asbestos character in talc. Because of their cation-exchange capacity smectites can interact with certain drugs affecting their bioavailability. Nevertheless, this interaction could be advantageous in the formulation of controlled release systems, which is one of the most attention-grabbing fields of clay applications at present, as discussed below. Table 1 - Pharmaceutical, mineral, chemical and commercial correspondences among clays used in Pharmacy Clay Mineral
Rock
Pharmacopoeia] name
Kaolinite
Kaolin
Kaolin, Heavy (EP 4th)
Talc
Talc
Kaolin (USP 25) Talc (EP 4th and USP 25)
Montmorillomte
Beniorute
Bentonite (EP 4th and USP 25)
Group: Smectites Subgroup: diocthaedral Saponite
Purified Bentonite (USP 25)
Chemical name and CAS registry number Hydrated aluminium silicate (1332-58-7)
Empirical formula
Usual names
Al,Si,O s (OH)«
Talc (14807-96-6)
Mg 3 Si 4 O l() (OH)i
Aluminium magnesium silicate (1302-78-9) Aluminium magnesium silicate (12511-31-8)
(Na.Ca.KWAl.Mgfe Si 3 Oi 0 (OH)-.nH,O
China Clay, bolus alba. porcelain clay, weisserton, white hole Magsil osmanthus, Magsil star. powdered talc, purified french chalk, purtalc, soapstone. steatite Mineral soap, clay soap, taylorite, wtlkinite, Veegum HS, Albagel, mineral colloid
Bentonite
Aluminium magnesium silicate (EP 4th) Magnesium aluminium silicate (USP 25)
Aluminium magnesium silicate (12511-31-8) Magnesium aluminium silicate (1327-43-1)
(Ca,Na,K) W3 (Mg,Fe) 3 (Si, Al) a Oi 0 (OH) : .nH ; O
Veegum R-K-HV-T-F, Carrisorb, Gelsorb, Magnabites Colloidal, Colloidal complex
Suberoup: triocthaedral Palygorskite
Palygorskite
Attapugite (USP 25)
(Mg,Al.Fe) 5 (Si,Al) 8 0 : o (OH), (OH2), (H : O) a
Attapulgite, Attasorb, Pharmasorb
Sepiolite
Sepiolite
Magnesium trisilicate (EP 4th) and (USP 25)
Aluminium magnesium silicate (12511-31-8) Magnesium aluminium silicate (1327-43-1) Hydrated magnesium trisilicate (39365-87-2) Magnesium aluminium silicaie (1327-43-1) Anhydrous magnesium trisilicate, magnesium metasilicate, magnesium ortnsilicare
Mg 3 Si i : 0 3 0 (OHMOH : )4 (H2O)B
Silicic acid, hydrated magnesium salt, meerschaum, parasepiolite, sea foam, talcum plasticum.
Group: Smectites
Finally, regarding fibrous clays, although this is not a specific requirement of any Pharmacopoeia, the particle size of fibrous minerals must be carefully controlled because of its possible biological effect. In a preformulation study, Viseras et al [39] showed that samples corresponding to different clay minerals used in pharmacy
Pharmaceutical and Cosmetic Applications of Clays
273
presented particle sizes lower than the value generally accepted for defining a particle as a fibre (> 5 urn in length and a length/diameter ratio > 3:1) [40]. Moreover studies carried out on humans exposed to some sepiolite samples confirmed that exposure to these minerals involves no risks [41-43]. The I ARC clearly distinguishes between palygorskite and sepiolite [44,45]. Palygorskite samples are classified as long (>5um) clay fibres and short(<5um) clay fibres. Long palygorskite samples are possibly carcinogenic to humans (group 2B), while the short ones cannot be classified as to their carcinogenicity to humans (group 3). On the other hand, there is inadequate evidence in humans for the carcinogenicity of sepiolite (group 3), whatever the length of the fibres, although there is limited evidence in laboratory animals suggesting carcinogenicity of long(>5um) fibres. 3.3 - Use as excipients In the preparation of pharmaceutical products particular importance is attached to the selection of suitable excipients, i.e., auxiliary substances contained in the formulation with the purpose of providing the product with an adequate presentation. Excipients must facilitate the administration of the active ingredients, improve their efficiency and ensure stability until the expiry date for usage by the patients. The fundamental property for a product to be used as excipient is it being innocuous, while attention should also be paid to other attributes affecting the organoleptic characteristics of the end product, such as taste, smell and colour. Clays are regarded as essentially non-toxic and non-irritant materials at the levels used in pharmaceutical excipients and are included in the Inactive Ingredients Guide [46] published by the Food and Drug Administration (FDA). This guide contains all inactive ingredients present in approved (or conditionally approved) drug products marketed for human use. Table 2 shows a summary of the applications of clay minerals as pharmaceutical excipients in drug products as provided by this guide. 3.3; 1 - Solid dosage excipients In Table 3, clay minerals are classified according to their functionality as excipient in solid dosage forms. Kaolinite is mainly used as a diluent because of its white to greyish-white colour. Its suitability as pharmaceutical excipient greatly depends on the geological nature (sedimentary, residual, and hydrothermal) and mineral composition of the deposits, which have an important effect on texture and particle size distribution, and consequently, on the rheological properties (flow) of the powder mass [26,47,48]. Talc is mainly used as diluent, glidant and lubricant in tablet and capsule formulations. In addition, talc is used as an additive to promote film coating of tablets and particles [13,49,50]. Several studies have shown that the chemical composition and physical properties of talc depend on the source and the method of preparation [47,48,51-55]. Smectites, such as bentonite and Mg Al silicate, are used in solid dosage forms as tablet and capsule disintegrants, tablet binders and adsorbents. The use of bentonite in the formulation of tablets has been studied in the past by several authors [56-58]. Feinstein and Bartilucci [59] investigated the efficiency of disintegration of bentonite, concluding that its effectiveness is comparable to other typical disintegrants, such as cellulose derivatives. Wai et al [60] indicated that laminar clays are not good disintegrants when used as intragranular agents. In contrast, Fielden [61] proposed that a suitable technological procedure results in good disintegrant characteristics, even when
274
A. López-Galindo and C. Viseras
used as an intra-granular agent. Table 2 - Pharmaceutical applications of clays as excipients in drug products for human use. Pharmacopeial name
Administrati ion
Dosage form
Potency range (*)
Immediate release (IR) Kaolin or Heavy kaolin
Oral
and Modified release (MR)
— N o n specified
(delayed or sustained) tablets Film coated tablets
Talc
0.189-204mg
Oral
MR (Sustained or Repeat actiion) 0.2 - 3 mg (sustained) and 73.93 mg (repeat) tablets
Sublingual
Tablet
5 mg
Lotion Topical
Ointment Powder
Non specified
Capsules Oral Bentonite or Purified Bentonite
Tablets Suspensions
0.45 % w/w
Topical
Suspensions
2.1 %w/w
Transdermal
Film
Vaginal
Suppository
Non specified
Drops Granule Oral Mg Al or Al Mg silicate
Reconstitution granules
— N o n specified
Syrup Suspensions
0.15-2 %Wv
Tablets
8mg
Rectal
Suspension
Vaginal
Ointment
Non specified
Emulsion (creams) Topical
Lotion
1.5% w/w
IR Tablets Magnesium Trisilicate
Oral
Coated Tablets
Non specified
Sustained Release Tablets (*) POTENCY RANGE: Minimum and maximum amounts of inactive ingredient for each route/dosage form
The use of fibrous clays in the formulation of tablets is based on their properties as glidants and binders. The suitability of some laminar and fibrous phyllosilicates as additives in solid dosage forms was recently investigated [12,62]. Fibrous clays can also be used as disintegrants. Viseras et al [63] showed that, in comparison to other silicates, sepiolite could be used as direct compression disintegrant even at low concentration. Regarding binding properties, Angulo et al [64] showed that sepiolite considerably improved the durability and quality of pellets. Moreover, their
Pharmaceutical and Cosmetic Applications of Clays
275
high surface area allows fibrous clays to be used in solid formulations as adsorbents of liquid drugs. Finally, unlike palygorskite, sepiolite may be used as a pharmaceutical excipient for drugs subject to oxidative degradation, such as hydrocortisone. Sepiolite avoids degradation because of its lower ferric iron content in comparison with palygorskite [65,66]. Table 3 - Uses of clay minerals as excipients in solid dosage forms Excipient
Dosage forms
Functional category
Kaolin and Heavy Kaolin
Tablets and capsules
Diluent and adsorbent
Talc
Tablets, capsules and powders
Coating aid, lubricant, diluent and glidant
Tablets, capsules and granules
Adsorbent, 1Dinder and disintegrant
Tablets and capsules
Adsorbent, |jlidant, binder and disintegrant
Bentonite Magnesiurr i Aluminium Silicate Magnesiurr 1 Trisilicate
3.3.2 - Liquid and semisolid dosage excipients Pharmaceutical dispersions are shaken several times during their "life", leading to changes in the system structure, and when administered orally they encounter a special pH environment that may severely affect their properties. Both suspending and anticaking agents are used to prevent drastic changes in dispersion properties. Some types of laminar and fibrous clays are particularly useful as stabilisers because of their positive thixotropic nature [39,67-70]. Table 4 summarises the main uses of clays in liquid and semisolid formulations. Kaolinite and talc are employed in liquid formulations as suspending and anticaking agents [13]. Lagaly [71] pointed out the importance of particle morphology and surface charge in the rheological behaviour of kaolin suspensions. Yuan and Murray [72] compared the rheological characteristics of kaolin dispersions prepared with different crystal morphologies (planar kaolinite and tubular or spherical halloysite), concluding that particle morphologies strongly affected the dispersion viscosities. Bentonite and Magnesium Aluminum Silicate are commonly used as suspending and stabilising agents in the formulation of suspensions, gels, ointments and creams for oral or topical administration. USP 25 describes four types of Magnesium Aluminum Silicate (IA, IB, IC, IIA) with different viscosity and Al/Mg ratio contents. When laminar clays are dispersed in a polar medium, face-edge and face-face interactions are the two major mechanisms implied in the formation of a rigid network [73-78]. Recently, the colloidal and rheological properties of bentonite suspensions were reviewed by Luckam and Rossi [79], who emphasise that laminar silicate gels are sensitive to the addition of electrolytes. In addition, Ma and Pierre [80] considered the influence of Fe3+ ions on the colloidal behaviour of montmorillonite suspensions, concluding that both Fe3+ and its hydrolytic products acted as counter ions to neutralise the electric double layer around clay particles. By means of absorption on clay particles,
276
A. López-Galindo and C. Viseras
the hydrolysis products could also modify the surface charge of the clay thus improving suspension coagulation. The effect of ion type and ionic strength on the sol-gel transition of sodium montmorillonite dispersions was studied by Abend and Lagaly [81], who obtained phase diagrams of different states (sol, repulsive gel, attractive gel, sediment) of the dispersions, showing that the borderline between gel and sediment depends on the type of counter-ion and co-ion. Table 4 - Uses of clays as excipients in liquid and semisolid dosage forms Excipient
Dosage forms
Functional category
Creams and pastes
Emulsifying agent
Suspensions
Suspending and anticaking agent
Kaolin and Talc
Bentonite and Magnesium Aluminium Silicate
Magnesium Trisilicate
Ointments, Creams and Gels Emulsifying agent Suspensions
Suspending and anticaking agent
Suspensions
Suspending and anticaking agent
Fibrous clays dispersed in water form a three-dimensional structure composed of interconnecting fibres [82]. Fibrous clay gels retain their stability in the presence of high concentrations of electrolytes, thus making them ideal for such an application [8385]. Some investigations have focused on the effects of hydrodynamic factors, such as size and shape of the particles, on the final product properties. Viseras et al [39] assessed the effects of shear history on the rheology of laminar and fibrous clay dispersions, concluding that the degree of dispersion and the structural changes resulting from differences in particle shape significantly affect the rheological properties of the systems. A linear relation was found between mixing energy and apparent viscosity in the laminar systems, while apparent viscosity was related to mixing power for the fibrous ones. A subsequent study examined the filtration behaviour of some Spanish clay-water dispersions, the results of which were compatible with the rheological properties of the systems [70]. Some authors have evaluated the use of clay minerals in combination with other agents. Ciullo [86] showed a synergic effect of Veegum® and natural gums as stabilisers in the formulation of emulsions. Recently, Lagaly et al [87,88] studied the use of smectites in combination with non-ionic surfactants as stabilisers in the formulation of oil in water emulsions. The main mechanism of stabilisation was the formation of a mechanical barrier around the oil phase droplets, preventing their coalescence. The rheological behaviour of the emulsions was also investigated and a strong influence of the clay mineral and surfactant was found.
Pharmaceutical and Cosmetic Applications of Clays
277
3.4 - Use of clay minerals as active substances Clay minerals are also used in pharmacy because of their biological activity, both in the treatment of gastro-intestinal and topical diseases. Moreover, they are used in the treatment of some much more specific illnesses. Marketed preparations containing clays as active substances are summarised in Table 5. Table 5 - Uses of clays as active principles in marketed products Active
Therapeutic use
Antidiarrhoeal & gastrointestinal protectors
Kaolinite Antacid
Brand names EU: Dystomin-E, Entrocalm, Collis Browne's, Kaoprompt-H, Kaopectate, Kaopectate-N, Enterosan, Kaodene, Kalogeais, Pectipar, Carbonaphtine Pectinee, Kao-Pront USA: Kao-Spen, Kapectolin, K-P Generic, Kaopectate Other: Bipectinol, Donnagel-MB, Kaomagma, Kaomagma with Pectin, Chloropect EU: Neutroses Vichy, Neutroses, Kaobrol Simple, Kaomuth, Anti-H, Gastropax Other: De Witt's Antacid
Anti-inflammatory
EU: Cicafissan, Antiphlogistine USA: Mexsana
Homeopathic Product
Other: Alumina Silicata
Anti-rubbing
EU: Ictiomen, Aloplastine, Lanofene 5, Poudre T.K.C.
Anti-haemorrhoids
EU: Titanoreine
Talc Pleurodesis
Palygorskite (attapulgite)
Antidiarrhoeal
Formulated and prepared in hospitals just before their use USA: Diar-Aid, Diarrest, Diasorb, Diatrol, Donnagel, Kaopectate, Kaopectate Advanced Formula, Kaopectate Maximum Strength, Kaopek, K-Pek, Parepectolin, Rheaban and Rheaban Maximum strength, Quintess EU: Streptomagma, Actapulgite, Gastropulgite, Mucipulgite, Norgagil, Diasorb Others: Fowler's and Kaopectate
Magnesium Trisilicate
Smectite
Antacid
USA: Streptomagma, Kaopectate
Antacid
EU: Neutroses Vichy, Neutrose S. Pellegrino, Instatina, Masbosil, Silimag, De Witt's antacid, Anti-acide-GNR, Gastric Expanpharm, Gastropax, Magnesie compose Lehning, Triglysal, Contracide, Gelusil Other: Trisil,. De Witt's Antacid, Gasulsol Tab
Antacid Antipruritic and local anaesthetic Antidiarrhoeal & gastrointestinal protectors
EU: Smecta Others: Calamine Lotion
EU: Diosmectite
278
A. López-Galindo and C. Viseras
3.4.1 - Antidiarrhoeal uses Antidiarrhoeals are usually categorized in four groups; antiperistaltics, adsorbents, antisecretory and digestive enzymes. Palygorskite and kaolinite are included in the second group of adsorbent agents [89,90]. Kaolinite, used as oral adjunct in the symptomatic treatment of diarrhoea because of its adsorbent properties, is administered orally in doses of about 2-6 g every four hours [13]. It may be formulated alone or in combination with other actives, such as pectin, loperamide, aluminium and magnesium salts, belladonna extract and morphine. As regards palygorskite, it has been described as even more effective than kaolinite in the symptomatic treatment of diarrhoea because of its capacity to adsorb and retain water, bacterium and some toxins [91]. Cerezo et a! [92] evaluated the possibilities of fibrous clays as non specific anti-diarrhoeic agents, concluding that both palygorskite and sepiolite comply with the pharmacopoeial specifications and may be taken into account. The daily dose of palygorskite can be up to 9 g in the form of oral suspension, conventional and chewable tablets. For oral suspension and tablets, the usual dose is 1200 to 1500 milligrams (mg) taken after each loose bowel movement, with no more than 9000 mg being taken in twenty-four hours. For chewable tablets, the dose is slightly less, and no more than 8400 mg should be taken in twenty-four hours. However, no conclusive evidence is available to show that palygorskite use may reduce the duration of diarrhoea, stool frequency, or stool fluid losses [93]. Smectite is equally effective in the treatment of infectious diarrhoea as it reduces the duration and frequency of liquid stool by mechanisms including absorption of water and electrolytes in the intestine, decrease of mucolysis caused by bacteria and protection of the luminal surface against pathogenic bacteria [94,95]. Moreover, some authors have described the use of smectite in the treatment of acute diarrhoea, although this clay is not currently recommended for this purpose [96-98]. On the contrary, Carretero [6] recently illustrated the use of Na+ smectite as osmotic laxative, although no experimental evidence supports this statement. 3.4.2 - Gastrointestinal protector Mucus forms a 200 micra thick layer of gel on the gastro-duodenal mucous membrane [99], which acts as a physical barrier preventing direct contact between the gastric enzymes and the cells of the mucous membrane, thus avoiding digestion of the latter [100] and mechanical erosion. In patients suffering from peptic ulcer, the thickness of the mucus layer decreases, while the mucolytic activity of the gastric juices and the enzyme levels increase [101,102). Clays provide multiple gastro-intestinal protection mechanisms associated with the different etiologies of deterioration mechanical erosion, enzyme attack, bacterial toxins, drugs, alimentary allergies, genetic factors, environmental factors such as tobacco, alcohol, etc. Several adsorbent agents (bentonite, kaolinite, active carbon) present anti-endotoxemic activity both in vivo and in vitro that reduces the alteration of the mucous membrane to the levels of healthy individuals [103,104]. The protective effect of clays on intestinal barriers is related to their influence on the rheological properties of the mucus. As discussed previously, clay particles dispersed in a water solution medium greatly increase its viscosity and, consequently, its stability. On the other hand, the mucoadhesivity of clays, i.e., their positive interaction with and binding to glycoproteins present in the mucus, is probably an important protective mechanism. Deterioration of glycoproteins by reactive agents, such as free
Pharmaceutical and Cosmetic Applications of Clays
279
radicals, ethanol or some drugs, is reduced when the polymer is complexed to the clay [105-108]. 3.4.3 - Antacid uses Clay minerals such as palygorskite and magnesium trisilicate (sepiolite) may be used as symptomatic antacid agents, because of their capacity to neutralise acidity in the gastric secretions. They are used in combination with aluminium hydroxide and kaolinite in suspensions and chewable tablets. Magnesium trisilicate is given in doses from 1 to 4 g, reacting with hydrochloric acid to form magnesium chloride and silicon dioxide, with an H+ neutralising capacity of around 15 meq/g. It is indicated in the treatment of gastric and duodenal ulcers. Magnesium chloride resulting from the neutralizing action may induce diarrhoea in some cases. Kaolinite in combination with sodium bicarbonate and magnesium trisilicate is commercialized, having an H+ neutralising capacity of around 56 meq /g. 3.4.4 - Anti-inflammatory and antiseptic purposes Purified talc is used in dusting powders to calm irritation and prevent roughness, while kaolinite is used for sore throat symptoms, including tonsillitis, pharyngitis and stomatitis, and is responsible for the adsorption of waste products. Mixtures of kaolin, bentonite and palygorskite have been proposed for use as dressing for the treatment of skin injuries, especially burns [109]. Kaolinite is applied topically as kaolin poultice to reduce inflammation [110]. Finally, pastes of kaolin and salicylic acid are applied as percutaneous anti-inflammatory in the treatment of muscular pain and tendonitis. Fibrous clays are also used in the treatment of aqueous inflammations, adsorbing the aqueous fraction and probably also retaining the proteic fraction of the inflammation [111-113]. They are probably able to effectively retain toxins and bacteria as happens in the gastrointestinal tract. Bentonite is included in antipruritic and local anaesthetic preparations for topical use. 3.4.5 - Topical applications The use of clay minerals as actives in topical dosage forms (creams, milks and powders) has been proposed on the basis of their capacity to efficiently adsorb a variety of undesired substances, including greases, skin exudates and external agents such as bacterial toxins [114]. However, clay minerals in such formulations are normally employed as excipients, i.e., as auxiliary substances intended to maintain the dose of the active principle in the area to be treated by increasing the viscosity of the system, promoting skin-adhesivity and keeping a high concentration of drug in the proximity of the treated skin-area. 3.4.6 - Other uses Kaolin is included in human homeopathic preparations, when it is known as alumina silicate, in the form of drops, globules and oral granules. Bentonite may be used as adsorbent in paraquat poisoning [110]. Talc is concomitantly used with carrageenates in suppositories administered as mucoprotectors and lubricants of rectal mucosa in the treatment of haemorrhoids. Finally, this mineral is also indicated as the preferred treatment for pleural effusion, a complication in patients with malignant neoplasms caused by disturbance of the normal
280
A. López-Galindo and C. Viseras
reabsorption of fluid in the pleural space [115]. Talc pleurodesis for the treatment of malignant pleural effusion is an effective method, preventing recurrent effusion in 8090% of cases and being less painful than tetracycline [116]. Talc can be insufflated in a dry state or instilled as slurry. The dose should be restricted to 5 grams [117]. 4 - New uses in modified drug delivery systems Most of the clays used in pharmacy can interact with other components of the formulation and, in the specific case of drug-clay interaction, this can affect the bioavailability of the drug itself [118-120], among others. The best known cases are those of montmorillonite and saponite, which are fairly common, well studied smectites [121,122]. Later studies have evaluated the effect of several factors such as ionic strength, the dielectric constant of the medium and the addition of polymers, confirming that ionic exchange is the main mechanism involved in absorption [123-127]. More recently, Tolls [128] examined the influence of the molecule's lipophilia in the absorption by clays of various veterinary drugs. On the other hand, the oral administration of fibrous clays could also affect the bioavailability of some drugs, such as mebeverine, folic acid, contraceptive steroids, promazine, atropine, glycosides (digoxin, digitoxin), erythromycin, paracetamol, chloroquine, quinidine, propranolol and tetracycline. Moreover antimicrobial preservatives, such as parabens, could be inactivated [129-133]. In recent years, there has been discussion on how to take advantage of these interactions for aims that are biopharmaceutical (modification of drug release or solubility), pharmacological (prevention or reduction of side-effects) and chemical (increased stability) [134]. Ideally, a pharmaceutical form should be designed to fulfil the therapeutic requirements, while avoiding or minimising the side effects. Conventional pharmaceutical forms are designed to release the dose immediately and achieve rapid, complete absorption of the drug. However, immediate release forms require repeated administration to maintain efficient concentrations of the drug. To avoid this, modified release pharmaceutical forms attempt to fulfil the therapeutic requirements by optimising the time, rate and location of drug release [28] and are known as "sustained release" (reducing frequency of administration to at least half that of a conventional form), "delayed release" (releasing the drug over a predetermined period) and "sitespecific release" (releasing the drug at or near the place of physiological activity). These release objectives can be achieved by using products of drug-clay interaction. Delgado et al [135] examined the use of kaolinite samples with different degrees of crystallinity as vehicles for the controlled release of drugs and found a linear relation between the crystallinity of the mineral and the release of amilobarbitone. Halloysite, a tubular polymorph of kaolinite, has recently been proposed for pharmaceutical use and, specifically, the tubules of this mineral could act as natural vehicles for microencapsulation and controlled release of both hydrophilic and lipophilic agents [136-138]. Moreover, the alternative absorption of macromolecules of opposite charge, including proteins, clays and poly-ions, has been proposed for the preparation of immobilisation vehicles, characterised by their capacity to guarantee the biological activity of the enzyme [139]. The halloysite microtubes filled with NAD coenzyme were assembled with ADH (alcohol-dehydrogenase) coenzyme and used as sustained release vehicle of the cofactor of the immobilised enzyme [140]. Regarding the use of swelling clays, Cameroni et al [141] studied the effect of
Pharmaceutical and Cosmetic Applications of Clays
281
different factors on drug release from compounds of papaverine and Veegum® (a commercial smectite) and found that the amount released depended on the pH, the ionic strength of the dissolution fluid and the elimination rate of the drug from the medium. Moreover, optimisation of the formulation, obtained by surface deposit of the papaverine on the compound, gave in vitro absorption profiles with zero kinetics, together with rapid achievement of constant drug concentrations in the gastro-intestinal tract [142]. Finally, Forni et al [143] showed that montmorillonite affects release in matrices of polyvinyl alcohol by interaction with the drug. More recently, Oya et al [144] proposed the use of Ag / montmorillonite compounds instead of a conventional organic agent, as a thermostable inorganic agent with high antimicrobial and antifungic activity for the treatment of muco-cutaneous conditions. Similar results were found using chelates of Ag and Tiabendazol in montmorillonite [145]. Fouche [146] examined the use of antibiotic-clay compounds in the treatment of gastric ulcer determined by Helicobacter pylori, with the conclusion that the clay aided penetration of the drug through the gastro-intestinal barrier. Absorption of 5-fluorouracil by montmorillonite has been considered for the development of new therapeutic systems for oral administration in the treatment of colo-rectal cancer [147]. hi recent years, five Spanish clays, including smectites and fibrous minerals, have been evaluated with regard to enzyme immobilisation, with the conclusion that at least the fibrous minerals could be used as vehicles for biotransformations [148]. These same clays can be used to obtain compounds with different types of drugs (timolol, tetracyclines, imidazolic antifungics) in which the release profiles have suitable kinetics for use as modified release systems [134,149-151). Preparation of the compounds is carried out by interaction of the solid (clay) with solutions of the drug in different media. However, Rives-Arnau et al [152] proposed a new dry process for the preparation of drug-clay compounds, as an alternative to the more common wet process, consisting in complexing by grinding the clay and the drug together. A third formation mechanism of these compounds would involve contact between the drug and the clay at the melting temperature of the active agent (Viseras et al, unpublished). 5 - Use in cosmetics The pharmaceutical (treatment) and cosmetic (care and beauty) uses of clay minerals are normally mentioned together, even though their aims are very different. It is therefore advisable to specify the intended use of a clay, as this will determine not only the technical aspects of its treatment, but also legal questions or matters of code of practice. A "cosmetic product" is any substance or preparation intended to be placed in contact with the various external parts of the human body (epidermis, hair system, nails, lips and external genital organs) or with the teeth and the mucous membranes of the oral cavity with a view exclusively or mainly to cleaning, perfuming, changing their appearance and/or correcting body odours and/or protecting or keeping in good condition (Council Directive 76/768/EEC on the approximation of the laws of the Member States relating to cosmetic products). On the other hand, Council Directive 2001/83/EC on the Community Code relating to medicinal products for human use defines "a medicinal product" as any substance or combination of substances administered to human beings for treating or preventing disease, making a medical diagnosis or restoring, correcting or modifying physiological functions.
282
A. López-Galindo and C. Viseras
A detailed study of all the possibilities of clays in the field of cosmetics falls outside the scope of this review, and so we shall concentrate on some examples that show the close relationship between their cosmetic applications and the properties resulting from their high specific surface and small size, which have been extensively discussed above. In most cases, cosmetic preparations make use of clays' rheological properties with the aim of the physical stabilising of the end product, just as they are used as excipients in medicinal preparations. Similarly, the use of clay minerals as active in cosmetics is closely related to their adsorbent capacity. They are used in deodorant powders and creams as they eliminate the gases responsible for the bad smell [153], in bath powders and baby powders, where they absorb sweat and humidity, keeping the folds of skin lubricated and thus avoiding friction; in facial powders to reduce the shine of talcum and increase the adherence of the preparation; and, finally, in face packs to clean the skin of grease. Other cosmetic uses are related to their emulgent capacity, whose mechanism has already been discussed, examples being the use of palygorskite and smectite in dry shampoos, which are widely used in North African countries. They can also be used as protection against external agents, in particular solar radiation, as proposed by Del Hoyo et al [154], who determined the capacity of phenylsalicylate complexes in sepiolite to prevent sunburn. In cosmetic preparations the clays act as a physical barrier against UV radiation, considerably increasing the protection factor of the compound. This is a question of much interest at present, given the appreciable increase in skin pathologies caused by radiation. 6 - Topical use: clays in spas Applications of clays to the human body for therapeutic purposes (geotherapy, fangotherapy and pelotheraphy) are very ancient techniques which have become increasingly popular in recent times. The beneficial effects for particular rheumaticarthritic pathologies and sporting injuries, as well as in dermatological and cosmetic applications, are based on the rheological properties, the high capacity for cation exchange and absorption, and the slow cooling rate of clays when properly prepared using different types of water. The term "peloid" refers to the product resulting from the mixture of a liquid phase (salt, saline or mineral-medicinal water), a solid inorganic phase (clay minerals and other minerals such as quartz, calcite, feldspars, etc.) and a third organic phase (bacteria, algae, diatomeas, protozoa, arthropods, etc.), which is applied topically as a therapeutic agent in the form of poultices or baths [6]. The preparation of thermal muds and peloids (medical and not mineralogical or geological terms) from clayey materials rich in smectites and other clays requires a process known as "maturation" affecting the clays when they are brought into contact with thermal and/or mineral water [155-160]. Traditionally, sulphurous water is used when the aim is to produce dermatological masks and bromo-iodic water for thermal treatment of bone and muscular injuries [161]. The maturing process lasts from 3 to 20 months and causes important changes technical properties of the clays, whose plasticity, absorption capacity, cooling index and grain-size alter as a result of the profound interaction between the different phases involved and the biological activity of the organisms themselves and their metabolic products. The nature of both the mineral and organic components involved is decisive for the final properties of the therapeutic mud [7,162-166], which varies from spa to spa according to the type of clayey material used
Pharmaceutical and Cosmetic Applications of Clays
283
and the composition of the thermal-medicinal water. The peloid obtained after maturation is applied to the whole body or on selected parts of the patients for 10 to 15 days at a temperature of 40-45°C in 1 to 2 cm layers for 20-30 minute sessions. The application produces relaxing, anti-inflammatory and analgesic effects in the treated area due to vasodilation, perspiration and stimulation of the cardio-circulatory and respiratory systems. It is particularly beneficial in the treatment of degenerative arthropathies and the associated painful syndromes, bone and joint injuries, rheumatism and arthritis in different parts of the body, spondylosis, myalgia, neuralgia, chronic phlebopathy, certain skin ailments, etc. [167-169]. Although there is no specific protocol for qualification of any one "peloid" thermal mud, in recent years considerable progress has been made in this direction, particularly due to the various proposals of the Italian Group of the AIPEA [3,4]. In many spas, after the local reserves of clays are exhausted, artificial mixtures of clayey materials are used whose nature is not always clearly determined. The choice of a suitable material should be made with clear ideas as to factors such as mineral composition, chemistry, pH, grain-size, specific surface, cation exchange capacity (total and for the main cations Na+, K+, Ca2+, Mg2+), consistency parameters (liquid and plastic limits, plastic index), rheology (activity, adhesivity, viscosity, water retention), thermal behaviour (heat capacity and conductivity, cooling kinetics), and organic matter and micro-organisms content. The most suitable materials are those with a high content in swelling clay minerals, fine granulometry and a low amount of "abrasive" materials (quartz, feldspar) for a pleasant application of the "peloid" mud, good thermal, rheological and adhesive properties and a low content in hazardous trace elements and minerals (such as free silica and asbestiform minerals). In this sense, we should point out the importance of control of the contents in certain, potentially toxic trace elements and their mobility during the maturing process (such as As, Sc, Tl, Pb, Cd, Cu, Zn, Hg, Se and Sb) in order to avoid possible intoxication during treatment [164,170,171]. 7 - Concluding remarks The development of certain instrumental techniques during the second half of the 20th century led to the discovery of the enormous compositional and textural variability of clays, thus improving understanding of the different mechanisms involved in their physico-chemical properties, in particular, those related to their surface characteristics (adsorbent capacity and Theological properties). These theoretical advances, which helped understanding of the processes behind the traditional uses of clays since antiquity as natural products with therapeutic and cosmetic aims, also resulted in the development of new applications. Of all the applied sciences using these "new" materials, those concerning health seem to be where most future investigation will take place on clays, to determine their possibilities in the treatment of illnesses and in the care and protection of the human body. What is at present known, as briefly described in this survey, informs of the variety and number of applications in use and allows us to foresee important advances in the coming years, particularly in the development of new drug delivery systems. The global increase in standard of living also suggests that body care in specialized centres will become increasingly popular, involving a reconsideration of the geomaterials used in such centres, which will inevitably require a correct qualification of the materials used and the exchange mechanisms involved.
284
A. López-Galindo and C. Viseras
Acknowledgements This work received support from DGES project MAT2003-06606 and Group RNM-0179 of the Junta de Andalucia (Spain). 8 - References [I] J. Bech, in: La Cristalografia y la Industria Farmaceutica, ed. by Reales Sociedades Espafiolas de Fisica y Quimica, Univ. Granada, 15 (1996). [2] E. Galan, MJ. Liso and M. Forteza, Bol. Soc. Esp. Miner., 8 (1985) 369. [3] F. Veniale, ed, Atti Convegno "Argille Curative", 1996. [4] F. Veniale, Miner. Petrogr. Acta, XLII, (1999) 267. [5] W.R. Reinbacher, Clay Miner. Soc. USA News, (1999) 22. [6] M.I. Carretero, Appl. Clay Sci., 21 (2002) 155. [7] A. Bettero, M. Marcazzan and A. Semenzato, Miner. Petr. Acta, XLII, (1999) 277. [8] S. Cara, G. Carcangiu and M. Tamanini, Miner. Petr. Acta, XLII, (1999) 299. [9] J. Cornejo, in: Conferencias de la IX y X Reuniones de la Sociedad Espanola de Arcillas, ed. by E. Galan and M. Ortega, (1990) 51. [10] A. Oya, T. Banse, F. Ohashi and S. Otani, Appl. Clay Sci., 6 (1991) 135. [II] D.B. Braun, Over the Counter Pharmaceutical Formulations, Noyes Publications, New Jersey, 1994. [12] C. Viseras and A. L6pez-Galindo, Pharm. Dev. Techn., 5, 1 (2000) 47. [13] A.H. Kibbe, Handbook of Pharmaceutical Excipients. 3 rd Ed., American Pharmaceutical Association, Washington DC, 2000. [14] C. Viseras and A. Lopez-Galindo, App. Clay Sci., 14 (1999) 69. [15] R.E. Grim and N. Guven, Bentonites: geology, mineralogy, properties and uses, Elsevier, Amsterdam, 1978. [16] J.H. Denis, M.J. Keall, P.L. Hall and G.H. Meeten, Clay Min., 26 (1991) 255. [17] A. Yebra, Influencia de la mineralogia, quimismo y textura en las aplicaciones basicas industriales de la sepiolite. Ph.D Thesis, Universidad de Granada, 2000. [18] R.H.S. Robertson, Mineral use guide, Cleaver-Hume Press, London, 1957. [19] R.E. Grim, Applied Clay Mineralogy, McGraw-Hill Book Company, Inc., New York, 1962. [20] W.L. Haden and LA. Schwint, Ind. Engin. Chem., 59 (1967) 58. [21] A. Alvarez, in: Palygorskite-Sepiolite. Occurrences, Genesis and Uses, ed. by A. Singer and E. Galan, Elsevier, Amsterdam, 1984. [22] E. Galan, Proc. Int. Clay Conf. Denver 1985, ed. By L.G. Schulz, H. Van Olphen and F.A. Mumpton, The Clay Mineral Soc, Bloomington, 1987. [23] E. Galan, Clay Minerals, 31 (1996) 443. [24] F. Veniale, Sci. Geol. Mem., 89 (1990) 81. [25] J. Konta, App. Clay Sci., 10 (1995) 275. [26] H.H. Murray, App. Clay Sci., 17 (2000) 207. [27] European Pharmacopoeia 4th Edition, Strasbourg, France, 621 (a), 707 (b), 1429 (c), 1518 (d), 1999 (e), (2002). [28] United States Pharmacopoeia 25, Rockville, MD, 2573 (a), 2510 (b), 974 (c), 1043 (d), 1639 (e), 182 (f), 80 (g), 80 (h), 183 (i), 2512 (j), 1044 (k), 2511 (1), (2002). [29] Silica, in: IARC Monographs on the evaluation of carcinogenic risks to humans, IARC Scientific publications, Lyon, France, 68 (1997) 41.
Pharmaceutical and Cosmetic Applications of Clays
285
[30] C.W. Ridout, Pharm. Acta Helv., 43 (1968) 177. [31] K.S. Albert, K.A. DeSante, R.D. Welch and A.R. DiSanto, J. Pharm. Sci., 67 (1978) 1579. [32] K.S. Albert, J.W. Ayres, A.R. DiSanto, DJ. Weidler, E. Sakmar, M.R. Hallmark, R.G. Stoll, K.A. DeSante and J.G. Wagner, J. Pharm. Sci., 67 (1978) 1582. [33] F. Ganjian, A.J. Cutie and T. Jochsberger, J. Pharm. Sci., 69 (1980) 352. [34] J.C. McElnay, P.F. D'Arcy and O. Throne, Int. J. Pharm., 7 (1980) 83. [35] A.J. Bucci, S.A. Myre, H.S.L. Tan and L.S. Shenouda, J. Pharm. Sci., 70 (1981) 999. [36] S.A.H. Khalil, L.M. Mortada and M. El-Khawas, Int. J. Pharm., 19 (1984) 233. [37] Talc, in: I ARC Monographs on the evaluation of carcinogenic risks to humans, IARC Scientific publications, Lyon, France, 42 (1987) 185. [38] Talc not containign asbestiform fibres, in: IARC Monographs on the evaluation of carcinogenic risks to humans, IARC Scientific publications, Lyon, France, 7 (1987) 349. [39] C. Viseras, G.H. Meeten and A. Lopez-Galindo, Int. J. Pharm., 182 (1999) 7. [40] Ausschuss fur Gefahrstoffe, Bun. Des Arbeitsblatt 9 (1988) 84. [41] Y.L. Baris, A.A. Sahin and M.L. Erkan, Arch. Environ. Health, 35 (1980) 343. [42] K. McConnochie, C. Bevan, R.G. Newcombe, J.P. Lyons, W.J. Skidmore and J.C. Wagner, Thorax, 48 (1993) 370. [43] M. Governa, M. Valentino, L. Visona, F. Monaco, M. Amati, G. Scancarello, and G. Scansetti, Cell Biol. Toxic, 11 (1995) 237. [44] Palygorskite (attapulgite), in IARC Monographs on the evaluation of carcinogenic risks to humans, IARC Scientific publications, Lyon, France, 68 (1997) 245. [45] Sepiolite, in IARC Monographs on the evaluation of carcinogenic risks to humans, IARC Scientific publications, Lyon, France, 68 (1997) 267. [46] Inactive Ingredient Guide, Division of Drug Information Resources, Center for Drug Evaluation and Research, US Food and Drug Administration, Rockville, MD, 1996. [47] E. Gamiz, E. Caballero, M. Delgado-Rodrlguez and R. Delgado Calvo-Flores, Ann. pharmaceutiques francaise, 47 (1989) 53. [48] E. Gamiz, G. Delgado Calvo-Flores, J. Parraga and R. Delgado Calvo-Flores, Ann. pharmaceutiques francaise, 47 (1989) 33. [49] S. Dawoodbhai and C.T. Rhodes, Drug Dev. Ind. Pharm., 16, 16 (1990) 2409. [50] S. Dawoodbhai, E.R. Suryanarayan, C.W. Woodruff and C.T. Rhodes, Drug Dev. Ind. Pharm., 17, 10 (1991) 1343. [51] K. Lin and G.E. Peck, Drug Dev. Ind. Pharm., 20, 19 (1994) 2993. [52] K. Lin and G.E. Peck, Drug Dev. Ind. Pharm., 21, 4 (1995) 447. [53] K. Lin and G.E. Peck, Drug Dev. Ind. Pharm., 21,2 (1995) 159. [54] K. Lin and G.E. Peck, Drug Dev. Ind. Pharm., 22, 5 (1996) 383. [55] D.S. Phadke, M.P. Keeney and D.A. Norris, Drug Dev. Ind. Pharm., 20, 5 (1994) 859. [56] H.M. Gross and C.H. Becker, J. Am. Pharm. Assoc. Sci. Ed., 41 (1952) 157. [57] A. Firouzabadian and C.L. Huyck, J. Am. Pharm. Assoc. Sci., 43 (1954) 248. [58] A.D. Nair and V.N. Bathia, J. Am. Pharm. Assoc. Sci. Ed., 46 (1957) 131. [59] W. Feinstein and A.T. Bartilucci, J. Pharm. Sci., 55 (1966) 332. [60] K.N. Wai, H.G. Dekay and G.S. Banker, J. Pharm. Sci., 55 (1966) 1244. [61] Fielden, K.E. (1996). Water-dispersible tablets. USA Patent US 5556639,17 Sep.
286
A. López-Galindo and C. Viseras
[62] C. Viseras, A. Yebra, and A. Lopez-Galindo, Pharm. Dev. Techn., 5, 1 (2000) 53. [63] C. Viseras, F. Ferrari, A. Yebra, S, Rossi, C, Caramella and A. Lopez-Galindo, STP Pharma Sci., 11,2(2001)137. [64] E. Angulo, J. Brufau and E. Esteve, Animal Feed Sci. Technology, 53 (1995) 233. [65] J. Cornejo, M.C. Hermosin, J.L. White, G.E. Peck and S.L. Hem, J. Pharm. Sci., 69 (1980) 945. [66] M.C. Hermosin, J. Cornejo, J.L. White and S.L. Hem, J. Pharm. Sci., 70 (1981) 189. [67] M.K. El-Halabi, Estudio comparativo de bentonitas en relacion a sus aplicaciones en tecnologia farmaceutica. Ph.D Thesis, Universidad de Granada, 1977. [68] S.H. Chang, M.E. Ryan and R.K. Gupta, Rheol. Acta, 32 (1993) 263. [69] F. Miano and M.R. Rabaioli, Colloids and Surfaces, 84 (1994) 229. [70] C. Viseras, P. Cerezo, G.H. Meeten and A. Lopez-Galindo, Int. J. Pharm., 217 (2001)201. [71] G. Lagaly, Appl. Clay Sci., 4 (1989) 105. [72] J. Yuan and H.H. Murray, App. Clay Sci., 12 (1997) 209. [73] H. Van Olphen, An introduction to Clay Colloid Chemistry, John Wiley & Sons, New York, 1967. [74] Y. Permien and G. Lagaly, Clay Minerals, 29 (1994) 751. [75] Y. Permien and G. Lagaly, Clay Minerals, 29 (1994) 761. [76] Y. Permien and G. Lagaly, Colloid. Polymer Sci., 272 (1994) 1306. [77] Y. Permien and G. Lagaly, App. Clay Sci., 9 (1994) 251. [78] Y. Permien and G. Lagaly, Clays Clay Minerals, 43 (1995) 229. [79] P.F. Lukham and S. Rossi, Adv. Colloid Interface Sci., 82 (1999) 43. [80] K. Ma and A.C. Pierre, Colloids Surf. A, 155 (1999) 359. [81] S. Abend and G. Lagaly, App. Clay Sci., 16 (2000) 201. [82] T.C. Simonton, S. Komarmeni and R. Roy, App. Clay Sci., 3 (1988) 165. [83] V.V. Parkhomenko, V.Y. Tretinik and L.A.Kudra, J. App. Chem., 60 (1987) 2048. [84] G. Fadat, G. Engstrom and M. Rigdahl, Rheol. Acta, 27 (1988) 289. [85] U. Eriksson, G. Engstrom and M. Rigdahl, Rheol. Acta, 29 (1990) 352. [86] P.A. Ciullo, Cosmetics Chem., 32 (1981) 275. [87] G. Lagaly, M. Reese and S. Abend, App. Clay Sci., 14 (1999) 83. [88] G. Lagaly, M. Reese and S. Abend, App. Clay Sci., 14 (1999) 279. [89] M. Guerney, in: Handbook of gastrointestinal drug therapy, ed. by M.M. Van Ness and M. Guerney, Little Brown, Boston, 1989. [90] D.B. Braun, in: Over the Counter Pharmaceutical Formulations, ed. by D.B. Braun, Noyes Publications, New Jersey, 61, 1994 [91] H.L. DuPont, C D . Ericsson, M.W. DuPont, A. Cruz and J.J. Mathewson, Amer. J. Medicine, 88-6A (1990) 20S. [92] P. Cerezo, C. Viseras, A. Lopez-Galindo, F. Ferrari and C. Caramella, App. Clay Sci., 20 (2001) 81. [93] World Health Organization, The Rational Use of Drugs in the Management of Acute Diarrhoea in Children, Geneva. 1990. [94] J.G. Rateau, G. Morgant, M.T. Droy-Priot and J.L. Parier, Current Med. Res. Opinion, 8(1982)233. [95] C. Dupont, J.L. Moreno, E. Barau, K. Bargaoui, E. Thian and O. Plique, J. Pediatr. Gatroenterol. Nutr., 14, 4 (1992) 413. [96] A.A. Madkour, E.M. Madina, O.E. El-Azzouni, M.A. Amer, T.M. El-Walili and T.
Pharmaceutical and Cosmetic Applications of Clays
287
Abbass, J. Pediatr. Gastroenterol. Nutr., 17 (1993) 176. [97] L. Mahraoui, M. Heyman, O. Plique, M.T. Droy-Lefaix and J.F. Desjeux, Gut, 40 (1997)339. [98] A. Guarino, M. Bisceglia, G. Castellucci, G. Iacono, L.G. Casali, E. Bruzzese, A. Musetta and L. Greco, J. Pediatr. Gastroenterol. Nutr., 32, 1 (2001) 71. [99] M. Bickel and J.R. Kauffmann, Gastroenterology, 80 (1981) 770. [100] A. Allen, Eur. J. Gatroenterol Hepatol., 2 (1990) 169. [101] J.P. Pearson, R. Ward, A. Allen, N.B. Roberts and W. Taylor, Gut, 27 (1986) 243. [102] M.T. Droy-Lefaix, Revue Med., 138, 5 (1987) 411. [103] B. Ditter, R. Urbaschek and B. Urbaschek, Gastroenterology, 84 (1983) 1547. [104] K.R. Gardiner, N.H. Anderson, M.D. McCaigue, P.J. Erwin, M.L. Halliday and B.J. Rowlands, Gut, 34 (1993) 51. [105] J. Fioramonti, H. Navetat, M.T. Droy-Lefaix and L. Bueno, in: Veterinary pharmacology, toxicology and therapy in food producing animals, ed. by F. Simon, P. Lees and G. Semjem, Unipharma, Budapest, 1990. [106] M.T. Droy-Lefaix, O. Plique, G. Geraud and Y. Drouet, Hellen. J. Gastroenterol., 5 (1992) 70. [107] V. Theodorou, J. Fioramonti, M.T. Droy-Lefaix, O. Plique and L. Bueno, Aliment. Pharmacol. Ther. 8, 3 (1994) 295. [108] J.F. Peignet, P. Giral and O. Plique, Med. Chir. Dig, 26, 5 (1997) 233. [109] H.F. Kamp, USA Patent US 4748978, 7 June, 1988. [110] K. Parfitt, Martindale: The Extra Pharmacopoeia, 32th Ed., The Pharmaceutical Press, London, 1999. [ I l l ] M.A. Lizarbe, N. Olmo and J.G. Gavilanes, Biomaterials, 8 (1987) 35. [112] J.L. Herrera, N. Olmo, J. Turnay, A. Sicilia, A. Bascones, J.G. Gavilanes and M.A. Lizarbe, Biomaterials, 16 (1995) 625. [113] N. Olmo, J. Turnay, J.L. Herrera, J.G. Gavilanes and M.A. Lizarbe, J. Biomed. Mat. Res., 30 (1996) 77. [114] D.B. Braun, in: Over the Counter Pharmaceutical Formulations, ed. by D.B. Braun, Noyes Publications, New Jersey, 318, 1994. [115] L.F. Black, Mayo Clin. Proc, 47 (1972) 493. [116] L.S. Fentiman, R.D. Rubens and J.L. Hayward, Eur. J. Cancer. Clin. Oncol., 22 (1986) 1079. [117] Y. Aelony, R. King and C. Boutin, Ann. Intern. Med., 115 (1991) 778. [118] S.A.H. Khalil and M. Iwuagwu, J. Pharm. Sci., 67 (1978) 287. [119] J.E. Browne, J.R. Feldkamp, J.L. White and S.L. Hem, J. Pharm. Sci., 69 (1980) 816. [120] O. Al-Gohary, J. Lyall and J.B. Murray, Pharm. Acta Helv., 62, 3 (1987) 66. [121] J.W. McGinity and J.L. Lach, J. Pharm. Sci., 65 (1976) 896. [122] J.W. McGinity and M.R. Harris, Drug Dev. Ind. Pharm., 6, 1 (1980) 35. [123] M. Sanchez-Camazano, M.J. Sanchez, M.T. Vicente and A. Dominguez-Gil, J. Pharm. Sci., 69 (1980) 1142. [124] M.J. Sanchez, M. Sanchez-Camazano, M.T. Vicente and A. Dominguez-Gil, J. Pharm. Pharmacol., 33 (1981) 408. [125] M.J. Sanchez, M. Sanchez-Camazano, M.T. Vicente and A. Dominguez-Gil, Drug Dev. Ind. Pharm. 9, 6 (1983) 1019. [126] F. Forni, V. Iannuccelli, G. Coppi and M.T. Bernabei, Arch. Pharm., 322 (1989) 789.
288
A. López-Galindo and C. Viseras
[127] M.L. Salayero, M.J. Sanchez, M. Sanchez-Camazano and A. Dominguez-Gil, Drug Dev. Ind. Pharm. 11,11 (1985) 1909. [128] J. Tolls, Environmental Science & Technology, 35, 17 (2001) 3397. [129] M.C. Allwood, Int. J. Pharm., 11 (1982) 101. [130] M.A. Iwuagwu and A. Jideonwo, Int. J. Pharm., 65 (1990) 63. [131] O. Al-Gohary, Int. J. Pharm., 67 (1991) 89. [132] M.A. Iwuagwu and K.S. Aloko, J. Pharm. Pharmacol., 44 (1992) 655. [133] M.S. Arayne andN. Sultana, Pharmazie, 48 (1993) 599. [134] P. Cerezo, Mecanismos y cineticas de adsorcion-liberacion de farmacos en soporte arcilloso. Contribution al estudio de complejos timolol maleato con esmectitas, paligorskitas y sepiolitas. Ph. D. Thesis, Univ. Granada, 2003 [135] R. Delgado, G. Delgado, A. Ruiz, V. Gallardo and E. Gamiz, Clay Minerals, 29 (1994)785. [136] R.R. Price and B.P. Gaber, USA Patent US5651976, 29 July, 1997 [137] R.R. Price, B.P. Gaber and Y. Lvov, J. Microencapsulation, 18, 6 (2001) 713. [138] S.R. Levis and P.B. Deasy, Int. J. Pharm. 243 (2002) 125. [139] Y. Lvov, K. Ariga, I. Ichinose and T. Kunitake, Thin Solids Films, 284/285 (1996) 797. [140] Y. Lvov, R. Price, B. Gaber and I. Ichinose, Colloids and Surfaces: Engineering, 198-200 (2002) 375. [141] R. Cameroni, F. Forni, V. Iannuccelli, G. Coppi and M.T. Bernabei, II Farmaco, 40,9(1985)283. [142] F. Forni, V. Iannuccelli, G. Coppi, M.A. Vandelli and R. Cameroni, II Farmaco, 40,4(1985)101. [143] F. Forni, V. Iannuccelli, G. Coppi, M.A. Vandelli and R. Cameroni, Boll. Chim. Farm., 126,8(1987)342. [144] A. Oya, T. Banse and F. Ohashi, Appl. Clay Sci., 6 (1992) 311. [145] F. Ohashi, A. Oya, L. Duclaux and F. Beguin, App. Clay Sci. 12 (1998) 435. [146] C. Fouche, Patent number FR2677253. 12 November, 1992. [147] F.H. Lin, Y.H. Lee and C.H. Jian, Biomaterials, 23, 9 (2002) 1981. [148] I. De Fuentes, C. Viseras, D. Ubiali, M. Terreni and A. Alcantara, J. Molecular Catalysis B: enzymatic, 11 (2001) 657. [149] C. Aguzzi, C. Viseras, A. Garces, J. Cruz, P. Cerezo, F. Ferrari and C. Caramella, Proc. 1st EUFEPS Conference on Optimising Drug Delivery and Formulation: New Challenges in Drug Delivery, Versailles (F), 2003. [150] C. Viseras, C. Aguzzi, M. Zafra, A. Garces, P. Cerezo, C. Caramella and A. Lopez-Galindo, Proc. 1st EUFEPS Conference on Optimising Drug Delivery and Formulation: New Challenges in Drug Delivery, Versailles (F), 2003. [151] C. Viseras, P. Cerezo, C. Aguzzi, M.T. Viseras, C. Caramella, and A. Cerezo, Proc. 1st EUFEPS Conference on Optimising Drug Delivery and Formulation: New Challenges in Drug Delivery, Versailles (F), 2003. [152] V. Rives-Arnau, C. Del Hoyo and M.A. Vicente, ES Patent ES2126506, 16 March, 1999. [153] H. Ueda and M. Hamoyoshi, J. Mater. Sci., 27 (1992) 4997. [154] C. Del Hoyo, M.A. Vicente and V. Rives, Clay Minerals, 33 (1998) 467. [155] M. De Bernardi and G.M. Pedrinazzi, in: Argille Curative, ed. by F. Veniale, Proc. Meeting Salice Terme/PV, Gruppo Ital. AIPEA, 1996. [156] J. Yvon and T. Ferrran, in: Argille Curative, ed. by F. Veniale, Proc. Meeting
Pharmaceutical and Cosmetic Applications of Clays
289
Salice Terme/PV, Gruppo Ital. AIPEA, 1996. [157] L. Galzigna, C. Moretto and A. Lalli, Biomedicina & Pharmacotherapy, 50 (1996) 306. [158] N. Minguzzi, N. Morandi, S. Tagnin and F. Tateo, Miner. Petr. Acta, XLII, (1999) 287. [159] F. Veniale, M. Setti, F. Soggetti, M. Lofrano and F. Troilo, Miner. Petrogr. Acta, XLII, (1999)263. [160] C.J. Sanchez, J. Parras and M.I. Carratero, Clay Minerals, 37 (2002) 457. [161] C. Torresani, Cosmesi Derm., 30 (1990) 59. [162] T. Ferrand and J. Yvon, Appl. Clay Sci., 6 (1991) 21. [163] F. Veniale, in: Argille e minerali delle argille. Guida alia defmizione di caratteristiche e propieta per gli usi industriali, ed. by N. Morandi and M. Dondi, Corso di Formazione, Gruppo Italiano AIPEA, Rimini, 1997. [164] V. Summa and F. Tateo, Appl. Clay. Sci, 12 (1998) 403. [165] S. Cara, G. Carcangiu, G. Padalino, M. Palomba and M. Tamanini, Appl. Clay. Sci, 16(2000)117. [166] S. Cara, G. Carcangiu, G. Padalino, M. Palomba and M. Tamanini, Appl. Clay. Sci, 16 (2000) 125. [167] P. Barbieri, in: Argille Curative, ed. by F. Veniale, Proc. Meeting Salice Terme/PV, Gruppo Ital. AIPEA, 1996. [168] F. Benazzo and A. Todesca, in: Argille Curative, ed. by F. Veniale, Proc. Meeting Salice Terme/PV, Gruppo Ital. AIPEA, 1996. [169] G. Nappi, Medicina e clinica termale, Edizioni Selecta Medica, Pavia, 2001 [170] V. Summa and F. Tateo, Appl. Clay. Sci, 15 (1999) 477. [171] N. Mascolo, V. Summa and F. Tateo, Appl. Clay. Sci, 15 (1999) 491.
REMOVAL OF METALS BY NATURAL AND MODIFIED CLAYS CRISTINA VOLZONE Centro de Tecnologia de Recursos Mmerales y Ceramica - CETMIC (CIC-CONICET-UNLP) - CC 49, Cno. Centenario y 506, (1897) M.B. Gonnet Provincia de Buenos Aires - ARGENTINA E-mail: [email protected] / [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych andK.G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Removal of Metals by Natural and Modified Clays
291
1 — Introduction
It is well known that metals occur in nature and that from the early days man has been using some of them for different purposes. Copper, bronze and iron have been the most important ones in human history because they have marked an advance in civilisation as man started using tools made of them (approx. in 8,000 BC; 3,000 BC, and 1,200 BC, respectively). So important were these metals that historians took the name of some of them to describe periods, such as' Bronze Age' and 'Iron Age'. Metallic elements are classified on the basis of their physicochemical properties and these occupy three-quarters of the periodic table of elements. Growth in population and advances in science and technology have brought comfort, but a number of natural systems were altered and health and the environment have been affected as well. In our daily life, we are surrounded by metal ores and metallic elements that we have put to different uses. However, during the process of obtaining certain metallic elements from ores (metallurgical industry), a lot of waste is produced that is harmful to us. Acid mine water may show huge amounts of such metals, for example cadmium, copper, zinc, lead, mercury, etc. This water may occasionally contaminate the underground and natural water thus leading into a serious threat to living organisms. Therefore, the amount of metal in wastewater has to be reduced to prevent its accumulation in the biosphere. Heavy metals refer to high density metallic elements such as mercury, cobalt, copper, chromium (III), iron, etc. However, some of these metals including manganese, molybdenum, vanadium, strontium, zinc, etc, are essential to human, animal and plant life. On the contrary, excessive levels of some of them may be toxic. There are others, which are harmful to health even in low concentration such as mercury, lead, cadmium, chromium (VI), arsenic, and antimony. Treatment and location options for such heavy metals are to be taken into account to purify soils and waters. Toxic heavy metals such as cadmium and mercury may be included in the dietary habits of animals through environmental exposure, thus contaminating food products derived from those animals. There is an increasing interest around the world in cleaning up polluted rivers and lakes, and implementing systems that regulate the disposal of waste that contains metals. Among others, the International Environmental Protection Agency (EPA) has determined various levels in the concentration of metals beyond which organisms may be altered, even die, and produce corrosion on different solids. Most often, cadmium occurs in small quantities associated with zinc ores but also with copper and lead ores. Whenever cadmium compounds bind to sediments in rivers, they may be bio-accumulated or redissolved easily, and some cadmium compounds may leak through the soil and reach underground water. Environmental protection agencies have determined that the maximum level of cadmium in drinking water should not exceed 5 parts per billion (ppb) to overcome any health hazard. Copper is a metal found in natural deposits as ores containing other elements. It is widely used in household plumbing materials, and the maximum contamination level has been set at 1.3 parts per million (ppm). Copper is rarely found in underground water, but copper mining and smelting operations may be a source of contamination. The greatest percentage of total zinc in polluted soil and sediment is associated with iron and manganese oxides. Chromium is a metal found in natural deposits as ores containing other elements. It is mainly used in metal alloys such as stainless steel, protective coatings on metal, tannery industry, magnetic tapes, and pigments for paint products, cement, paper, rubber, floor covering, and other materials. The maximum level for chromium has been set at 0.1
292
C. Volzone
ppm. Mercury is a liquid metal found in natural deposits as ores containing other elements. Electrical products such as dry-cell batteries, fluorescent light bulbs, switches, and other control equipment account for half of the mercury in use. No more than 2 ppb of mercury is allowed in drinking water. Mercury is usually removed from wastewater by co-precipitation, coagulation, adsorption, ionic change, solvent extraction, electrooxidation, and flotation [1]. Lead is a metal found in natural deposits as ores containing other elements. It is generally the most widespread and concentrated contaminant at a lead battery-recycling site. Lead must be removed from drinking water, if it is above 15 ppb. Thus, it is imperative that processes for heavy metal decontamination have to be developed. These may be carried out through different techniques: chemical precipitation, solvent extraction, ultrafiltration, ionic exchange, adsorption, biosorption, photocatalysis, reverse osmosis, evaporation, non-conventional flotation, etc. [2,3]. Metal immobilization through precipitation and adsorption is considered a common mechanism to reduce metal in contaminated soils [4]. Different natural substances, like zeolites, carbon, and clays may be used as adsorbents for metal retention [5], although of late the preparation of modified clays and design of new adsorbent materials have been studied as a way to improving environmental conditions. This Chapter gives an overview of removal of such metallic impurities using both natural and modified clays. 2 — Retention of metals by clays The components of waste disposal are usually natural clay deposits or compacted bentonites liner owing to their high sorption capacity for cations and their low hydraulic conductivity. Aluminosilicates and oxides minerals are capable of removing many metals over a wide pH range and to much lower dissolved levels than the precipitation method [6]. Clays are the most important elements of the mineral kingdom and their use, mainly in ceramic products, dates from around 8,000 BC. Clay minerals are part of the soils and are essentially hydrous aluminosilicates, commonly known as phyllosilicates. The phyllosilicates contain bi-dimensional tetrahedral and octahedral sheets. The tetrahedral cations are normally Si4+, which may be replaced by Al3+ and/or Fe3+, Figure 1. The octahedral cations normally are Al3+, Mg2+, Fe2+/3+. The assemblage of one octahedral sheet between two tetrahedral sheets is known as the 2:1 layer (or T-O-T), whereas; the assemblage of one octahedral and one tetrahedral sheets form the 1:1 layer (or T-O).
Figure I - Schematic clay structure (2:1 layer).
Removal of Metals by Natural and Modified Clays
293
Kaolinite and smectite are examples of 2:1 and 1:1 layers, respectively. Charge imbalance due to isomorphic substitutions in the structure layers is compensated by cations such as Na+, K+, Mg2+, Ca2+, etc., placed in interlayer position, Figure 1.These cations are easily exchanged [7,8] and this property can be used for metal retention on clays. Nevertheless, the retention of metals by clays can also be controlled by properties such as surface area, surface charge, pH, ionic strength, etc. [6,9]. The different cation exchange capacity of natural clays plays an important role in the retention mechanism [10] and there is over-exchange when initial metal concentration exceeds the concentration corresponding to the cation exchange capacity of clay [11]. The amount of cation uptake by clays could be increased after different physical and/or chemical treatments. These treatments can modify structural (chemical composition, changes in interlayer distance, new species in interlayer position, etc.), textural (surface area, pore distribution, porosity, etc.) and/or acidic properties (Lewis and/or Bronsted sites) (Figure 2). The modifications on clays can be carried out by different methods: intercalation of inorganic and/or organic substances; intercalation of hydrolysed inorganic OH-cation species; pillaring clays; ligands intercalation, acid and alkaline treatments, etc. [12-16]. Pillared clays, also known us pillared interlayer clays (PILCs), are obtained by cation exchange of polynuclear hydroxy-cation species between the aluminosilicate layers followed by calcinations [17]. The cation of the polynuclear hydroxy-cation may be: Al, Cr, Fe, Zr, Ti, Ni ions, etc. There are many publications dealing with preparation, characterisation and uses of intercalated OH-cation onto clays, and pillaring clays [1933] in which textural, structural and acidic properties have also been analysed as well as different applications, such as in catalysis and as adsorbents. Acid modifications on clays are prepared by treatment of clays with different concentrations of acid (generally HC1, H2SO4) solutions at boiling temperature [34-39]. The acid treatment increases the surface area, porosity, pore, volume, and acid sites of the clay, and it can be used in catalysis, as bleaching, as adsorbents, etc.
Figure 2 - Schematic changes in clays after structural and textural modifications (M). To evaluate environmental impact, many researchers are studying in detail the interaction between clays and different metals from solutions. Studies about retention of metal ions by clays are important since they give further information on procedures in soil chemistry, hydrometallurgy, treatment of
294
C. Volzone
wastewater, etc. The main factors that affect the adsorption of metals are pH, temperature, ionic concentration, ionic strength, metal amount, etc. [40]. Soil clays and oxides have shown the ability to select different metal cations as Christensen [41], Zhu and Alva [42], Carey et al [43] mentioned. The retention of cadmium by clay has been studied under different conditions and this is depicted by the isotherms, which show the initial slope of the adsorption isotherm and the amount adsorbed by illite increased with pH [40]. The adsorption of cadmium by some soils is found to be pH-dependent and increases when increasing Cd concentrations [40,44]. Maximum adsorption values at unadjusted pH for soils containing different clays content ranged from 7.88mmol kg"1 to 64.8mmol kg"1. Clay content in soils is important for higher retention; however, Cd adsorption is related significantly to contents of organic carbon and low crystalline Fe. Adsorption of Cd onto phyllosilicate clays may occur both by specific and non-specific adsorption [45]. Cd complexation to the edge sites was studied by Zachara and Smith [46]. Cadmium sorbed on soil is strongly influenced by soil pH, cation exchange capacity (CEC), and organic content [47]. A nuclear magnetic resonance (NMR) study of Cd adsorption on montmorillonte indicates that Cd ion may be localised in interlayer and on the external surface. CdCl+ can also be adsorbed in the interlayer [48]. Similar results are found for kaolinite [49] where treatment with diluted Cd solution produces adsorption of Cd in interlayer, whereas in a treatment with a concentrate Cd solution, the ions can be situated in interlayer and on surface sites. And there is also adsorption of CdCl+ in the interlayer. The adsorption of cadmium on montmorillonite is low in highly concentrated chloride solutions ( > 1M), [50]. Lead tends to accumulate in the soil surfaces. In calcareous soils, Pb precipitates as Pb-carbonates, and all components of soils are responsible for adsorption of Pb [51,52]. Pure kaolinite retains Pb and Ca ions at low initial metal concentrations; and the adsorption of PbCl+ and CdCl+ becomes important with a higher metal concentration, where the most important retention mechanism is cation exchange [53]. The adsorption of Pb appears to progress the most slowly initially, but after equilibrium is reached the adsorption of Pb is the highest observed [54]. Modified clays, as phosphatic clays may be effective for immobilising heavy metals such as Pb, Cd, and Zn ions from aqueous solutions [4], The amount of metal desorbed onto phosphatic clays decreased in the following order: Pb>Cd>Zn. The significant differences between the amounts of metals sorbed from phosphatic clay suggest differences in their adsorption mechanisms. Results demonstrate that waste phosphatic clay could be a potential agent to treat Pb-contaminated soils [4]. Thiamont is one smectite covalent grafted with a chelating sulfhydryl functionality. It is found to be an effective adsorbent for Pb and Hg (70 and 65 mg metal /g adsorbent, respectively), but as a less effective adsorbent for Cd and Zn ions [55]. The adsorption of Cd (pH=6.9), Cu (pH=4.9), Pb (pH=4.9), and Zn (pH=4.9) by Al- and Zr-hydroxy intercalated bentonite is dominated by cation exchange. On the other hand, Al and, to a lesser degree, Zr-hydroxy intercalated and pillared bentonites exhibit high affinity for Zn ions, which is independent from the ionic strength of the solvent at neutral pH [9]. It was also demonstrated that from pH=4 to pH=6, selectivity for kaolinite appears to be Pb > Cd ions [53,54]. The metal adsorption in kaolinite is usually accompanied by the release of the hydrogen (H+) ion from the edge sites of the mineral [54]. Mercier and Pinnavaia [14] developed a porous clay heterostructure with uniform intragallery mesoporosity with important capacity for Hg retention. The Hg
Removal of Metals by Natural and Modified Clays
295
adsorption-desorption study by Yin et al [56] indicated that the contamination of any soils with a high concentration of Hg (II) could result in groundwater problems because a large fraction of Hg could eventually be leached out, however, clay content and organic matter play an important role in soil remediation. The adsorption of Hg (II) by kaolinite is initially influenced by pH, and the presence of Cl, Ni, and Pb reduce the Hg retention [57]. Ionic strength and the presence of SO4= and PO4= ions have relatively low impact on the adsorption of Hg (II). The silanol (SiOH) group is responsible for retaining the bulk of the adsorbed Hg (II), and both the silanol (SiOH) and aluminol (A10H) groups must be considered for adsorption of Hg (II) by kaolinite [57]. Zachara et al [58] and He et al [59] have also studied ion adsorption capacity on the clay surface by applying a triple layer model (TLM) as a composite of Si-OH and Al-OH sites. Copper, nickel cobalt and manganese retentions by natural kaolinite have been studied by Yavuz et al [60], which demonstrated the following adsorption affinity order for metal ions: Cu > Ni > Co > Mn. They concluded that kaolinite might be used to remove traces of heavy metals from an aqueous solution. Triantafyllou et al [11] demonstrated that bentonite retains an important amount of Ni and Co, but it presents higher affinity for Ni. Adsorption of heavy metals on Na-montmorillonite decreases when pH decreases and, at low pH values (2.5-3.5), the hydrogen ion competes with heavy metal [61]. Abollino et al [61] have found the following order for retention of metals by Na-montmorillonite: Cr > Ni > Mn > Zn > Cu > Co > Pb; and also have analyzed adsorption in the presence of different ligands (EDTA, tartaric acid, oxalic acid, citric acid, etc.) present in solution. Cu, Zn, Cd, Hg, Pb, Ca, and Na ion retentions by soils have been studied by Airoldi and Critter [44] who have suggested that the study of adsorption onto soil surfaces gives information about the exchange capacity of the matrix and that interactions occur by complex formation between the organic matter of the soil matrix and the cations dispersed in an aqueous solution. Adsorption of Cu and Zn by tropical peat soils indicate that Cu is retained more than Zn at the same pH value, and there is a relationship between proton release and Cu and Zn adsorption in the range of 1 to 2, suggesting that Cu and Zn replaced one or two protons from the sites [62]. Matthes et al [9], Volzone and Garrido [15,16] have studied the retention of heavy metals by pillared bentonites. The adsorptions of Cd, Cu, Pb, and Zn by bentonite were dominated by cation exchange. This tendency is similar for certain cations as a function of pH by using pillared bentonites or intercalated bentonites with hydrolysed cations [9]. The retention Zn at pH=6.9 suggests a complexation of Zn ions on surface hydroxyl groups of the intercalated polyhydroxy cations in bentonites and the pillared clays [9]. The Zn ions may also be adsorbed by altered tuffaceous material provided they contain smectite as a clay component [63]. Selective Zn adsorption by halloysite decrease when pH decreases, and all the Zn adsorbed was extracted with 0.1 M HC1 [64]. The adsorption capacity of the different types of clays for Zn ions follows the following order: sepiolite > bentonite > paligorskite > illite > kaolinite, where clays with the highest specific surface and cation exchange capacity show the strongest adsorption capacity for Zn [6]. Garcia Sanchez et al [6] have analysed the retentions of Cd, Cu, Zn, and Ni by different silicate minerals such as sepiolites, kaolinites, illites, bentonites, and palygorskite, containing different impurities, and have demonstrated that a factor like the
296
C. Volzone
reaction medium, such as pH and ionic strength, influence the adsorption process. Adsorptions of Pb, Zn, and Cd by kaolinite in different conditions of pH, metal concentration and exposure periods have been also analysed by Miranda-Trevino and Coles [54]. When different cations are present in equal amount in the solution, the cation with the smaller atomic radius and/or higher charge is preferentially adsorbed by the montmorillonite [9,65]. This behaviour is also observed even if the concentration of sodium is present twenty-two times or more in a higher molar concentration than chromium in a wastewater solution [65-67]. 3 - Special analysis: retention of chromium by clays There is not much scientific information about the interaction between chromium in solution and clays or soils solids. In 1976, Bartlett and Kimble [68,69] analysed the retention of chromium by soils. In 1977, Griffin et al [70] studied the adsorption of chromium by kaolinite and montmorillonite clays and in 1980 Kopperman et al[71], and Kopperman and Dillard [72] by using chlorite, illite and kaolinite clays. At the same time, Rengasamy and Oades, in 1977, [19] and Brindley and Yamanaka, in 1979, [20] started specific studies about the intercalation of polymeric chromium species onto montmorillonites. Later, from 1984 until today other researchers such as Pinnavaia et al [21], Tzou and Pinnavaia, [22], Carr, [23], Volzone et al [73], Volzone and Cesio [74-76], Volzone [77-79] and Yoong et al [49] continued with similar studies and also analysed the Cr-pillared clays. The most stable oxidation state of chromium is chromium (III). Chromium (II) compounds are reducing agents, and chromium (VI) compounds are strong oxidising agents. The trivalent chromium, Cr (III), and the tetravalent dichromate Cr (VI) in solution and in soils are the most important forms in the environment chemistry. The Cr (III) is low in toxicity and an essential trace ion for several biological activities whereas hexavalent chromium is highly toxic. Chromium is increasing as a pollutant, mainly due to industrial activities such as leather tanning, rubber, mineral mining, paint formulating, porcelain enamelling, electrical and electronic components, and non-ferrous metal manufacturing. According to EPA, chromium is one of the major threats to human health. Although each country has its own regulations on the maximum concentration of chromium in different media, in general, it is provided that the content of chromium (III) in solution should not be above 0.10 ppm [80].
Figure 3 - Flow diagram showing different steps in the analysis of retention of chromium by clays: (a) chromium solutions, (b) clay conditions, (c) combination of (a) and (b). Cr(s): Species chromium (III) solutions prepared in laboratory, Cr (w) chromium from -wastewater solutions.
Removal of Metals by Natural and Modified Clays
297
The evaluation of the removal of chromium by clays can be considered as a combined analysis of chromium in solution plus clays, which in schematic form is shown in Figure 3. Examples of chromium retention on clays or on modified clays by using chromium from an aqueous solution or from wastewater are shown in Figure 3 c. 3.1 - Chromium (III) solutions The chemistry of chromium in solution is complex and, whenever a researcher is evaluating chromium retention on solids, it is advisable to know which chromium species are present in solution. Polymeric chromium (III) species in solution have been meticulously studied by Ardon and Plane [81], Laswick and Plane [82]; Kolaczkowcki and Plane [83], Thompson and Connick [84], Finhol et al [85], Stunzi and Marty [86], Spiccia et al [87], Stunzi et al [88], Yoon et al [49]. Cr (III) in solution may be present in different hydrolysed forms as a function of pH, OH/Cr, hydrolysis time and temperature, etc.; it is possible then that the amount of chromium retention by clays may vary, even without interfering cation. Figure 4 shows the UV-visible spectra of three chromium (III) solutions prepared from chromium nitrate salt, Cr(NO3)3.9H2O, at different hydrolysed conditions [75,87]. The "M" solution is a fresh 0.05 M chromium nitrate salt. The "T" solution is prepared by mixing rapidly a 0.5 M chromium nitrate solution by the addition of 2 NaOH (OH/Cr=8) [73,75]. The operation has been carried out under certain conditions in order to obtain a chromite solution that has been then rapidly acidified with 2 N HC1O4 to produce the protoned oligomers in solution with total chromium [Cr3+] = 0.05M. The solution has been equilibrated at 25 °C during 30 hours before being added to the clay suspension. The P(l/60) solution is prepared from 0.1 M chromium nitrate solution by the addition of 0.2 M NaOH (OH/Cr=2) at 60°C, and hydrolysed for one day [73]. The total chromium in this solution is [Cr3+] = 0.05 M. The three solutions show two maximum absorption peaks at 408 nm and 575 nm for the "M" solution; at 421 nm and 581 nm for the "T" solution; and at 424 nm and 586 nm for the "P" solution. The bands of the "M" solution correspond to the presence of monomeric species, Cr(H2O)63+ [84]. The shifting of the peak to high wavelength is related to higher polymeric chromium species [84]. A high content of trimeric with low monomeric and dimeric species should be present in the "T" solution depending on the way of preparation [86,87]. Yoon et al [49] have been obtained trimeric chromium in organic media, such as the trimeric chromium oxyformate. The polymeric chromium solution, P(l/60), contains mainly trimeric-species, Cr3(OH)45+, followed by terra-, Cr4(OH)66+; mono-, Cr(H2O)3+, and dimer-Cr2(OH)24+ species [73]. The polyhydroxy chromium in solution may be obtained by adding NaOH to a nitrate chromium solution (OH/Cr=2) at different hydrolysed temperatures and time [73]. Figure 5 shows the pH of polymeric solutions vs. hydrolysis time as prepared at two different temperatures, 20°C and 60°C. The pH of 0.1 M chromium nitrate solution is 2.10, and it changes to pH=4.00 after a final addition of 0.2 M NaOH (OH/Cr=2). When hydrolysis time progresses, pH decreases close to constant values as shown in Figure 5, reaching pH values of 3.23 and 2.45 for hydrolysis temperatures of 20°C and 60cC, respectively. The absorption spectrophotometric characteristics for each solution are shown in Table 1, where Xmaxj and ^max2 are the maximum adsorption values, and Imax1/Imax2 are the ratio between the maximum intensities of the bands.
298
C. Volzone
Figure 4 - UVvisible spectra of chromium solutions: -'-'-' M" solution,— ' T " solution, "/"' solution [75].
Figure 5 - pH of polymeric solutions vs. hydrolysis time [73]. The absorption spectrophotometric characteristics for each solution are shown in Table 1, where Xmaxt and A.max2 are the maximum adsorption values, and Imax1/Imax2 are the ratio between the maximum intensities of the bands. In the same table, the M solution (a monomeric solution as mentioned in the previous paragraph) is included for comparison, where the ratio between the maximum absorption at 408 to 475 equals 1.17, which confirms the presence of monomeric Cr(H2O)63+ [33]. In the OH-Cr
Removal of Metals by Natural and Modified Clays
299
solution, hydrolysed at 20°C for one or 30 days, P(l/20) or P(30/20), trimeric and tetrameric species are abundant [75]. The P(l/60) solution contains mainly trimeric followed by tetra species, but if the hydrolysis time is longer, for example, 36 days, P(36/60) or more, e.g. 100 days, the P(100/60) spectrum characteristic (Tablel) changes and the monomeric species are the highest, although small amounts of tri- and dimer species were also found. Longer hydrolysis times at 60°C would then originate the depolymerisation of the species in solution. The hydrolysed chromium solution aged for 3 weeks at 25°C, prepared with chromium nitrate solution and Na2CO3, shows bands located at 420 and 580 nm, as reported by Tzou and Pinnavaia [22]. The same authors have analysed a similar solution but aged for 36 hours at 95°C. The spectrum of this sample has shown two bands situated at 420 and 586 nm. More polymerised chromium species may be present in the last solution. Table 1 - Maximum absorption values (Ajnax,, X,max2) and ratio between maximum intensity of the band (Imax,/Imax2) [73]. ^ma X l nm 408
A,max2 nm 575
Imaxi/Imax2
P(l/20) P(2/20) P(8/20) P(17/20) P(30/20)
424 424 424 423 423
583 583 583 584 584
1.34 1.36 1.38 1.39 1.45
P(l/60) P(6.5/60) P(36/60) P(67/60) P( 100/60)
424 423 422 420 419
586 586 586 586 586
1.39 1.33 1.36 1.32 1.32
Hydrolysed Cr solution P(day/°C) M:(NO 3 ) 3 Cr0.1 M
1.17
The basic chromium sulphate, [Cr(H2O)5(OH)SO4], is a widely used primary tanning agent. Figure 6 (a) shows the spectrum, in the UV-visible range, of the basic chromium sulphate solution prepared with 2,000 ppm Cr (mg L"1). It shows two bands at 422 and 586 nm with similar characteristics of solutions containing chromium species such as Cr2(OH)24+ [89]. This type of species is important for tanning processes. The spectrum of the tanning waste solution is shown in Figure 6 (b). This solution was kindly provided for this study by INESCOP (Technological Institute for Footwear and Leather, Spain). The tanning waste corresponds to the end of the tanning process in a tannery, and it contains 2,000 ppm of chromium, organic matter, solids in suspension, inorganic salts, and oil and fats [89]. The suspended solids have been separated by filtration before the spectrum. The maximum absorptions of the waste solution were at 419 and 576 nm, (Figure 6 b). The type of chromium species in waste solution is unknown. However, according to the characteristics of the spectrum, monomeric chromium species, Cr(H2O)63+, could be present [86]. The different types of chromium species among tanning salt and tanning waste is due to changes originated after the tanning process.
300
C. Volzone
Figure 6 - UV-visible spectra chromium solutions containing 2,000 ppm Cr. (a) Basic chromium sulphate salt solution, (b) tanning wastewater [89]. 3.2 — Retention of chromium from solutions by clays Different aspects are included in this section: influence of chromium (III) species, types of clays, and the effects of thermal treatment, all referred to retention of chromium on clays and modified clays (as shown in Figure 3 c, quadrants I and II). 3.2.1 - Influence of chromium (III) species on clays A montmorillonite clay after being treated with monomeric chromium, Cr(H2O)63+, from M solution; trimeric chromium species from T solution, and polymeric chromium (III) species, obtained at different hydrolysed conditions, shows similar hkl reflections except to 001 reflection, d(001) spacing, which is modified according to Cr cation or intercalated OH-Cr cation species intercalated, Table 2. The structure of the montmorillonite is preserved in spite of the different media used as indicated by the hkl reflection of the sample after treatment with blank solutions (a reactive without chromium) [75]. The original sample (M) and after treatment, monomeric chromium shows similar interlayer spacing values, M: 14.8 A and M-M: 15.0 A. Higher values (18.4 - 20.7 A) have been obtained in montmorillonite after treatment with trimeric, MT, and polymerised solutions at different time and temperature conditions (M-P(l/20), M-P(30/20) and M-P(l/60) solutions). The increase in the spacing is attributed to different polymerised OH-Cr-species retained in the montmorillonite. Montmorillonite treated with polymeric hydroxy-chromium solution prepared during a long hydrolysed time at 60°C, P(36/60) and P(100/60), shows two different interlayer spacings, 13.5 and 19.0 A for M-P(36/60), and 13.9 and 17.0 A for MP( 100/60) samples as shown in Table 2. Such solutions contain mainly monomeric Cr species with small amounts of dimer and trimer, as mentioned in 3.1. The higher interlamellar spacing of montmorillonite, d (001) spacing, is due to higher proportions of
Removal of Metals by Natural and Modified Clays
301
polymerised chromium species intercalated onto clay due to the different density per atom of chromium of the chromium species. Rengasamy and Oades [19], Brindley and Yamanaka [20], Pinnavaia et al[21], and Tzou and Pinnavaia [22] have obtained a maximum of d(001) spacing of 14.0, 16.8, 27.0, and 27.6 A, respectively, for OH-Cr-smectites by using OH-Cr solution with OH/Cr=2. Whereas, Tzou and Pinnavaia [22], using hydroxy-chromium solution from NaCC>3 with an hydrolysis temperature at 25 °C during 3 weeks, have obtained d(001) spacing of 17.7 A. On the other hand, Carr [23] has obtained a basal spacing of approximately 14.7 A using an OH/Cr =1 solution. These differences are attributed to the different procedures followed by the authors: i.e., preparation, hydroxy-chromium solution, hydroxy-chromium-smectite, and amount of Cr/sample. Table 2 - d(001) spacing, A, of montmorillonite after treatment with differently prepared OH-Cr solutions (M: monomeric, T: trimeric, and P: polimerics). Sample
M
d(001),A
14.8
M-M M-T 15.0
18.4
M-P (1/60) 20.7
M-P (36/60) 13.5, 19.0
M-P (100/60) 13.9, 17.0
M-P (1/20) 19.9
M-P (30/20) 19.9
Modified kaolins and bentonites have been used to retain chromium from a basic chromium sulphate salt solution. The spectrum of this salt in UV-visible range has been shown in Figure 6.a, and it contains a high proportion of dimeric species, Cr2(OH)24+ [89]. Two kaolins, A and E, with higher kaolinite content in sample A (further on this in 3.4) and two bentonites, Bl, and B2, with high Wyoming- and Chetotype montmorillonite contents, respectively, have been used. The clays A, E, and Bl have been modified with caustic potash solutions, and stabilised at pH 7 and temperature of 500 K (Ak, Ek, Blk). The B2 bentonite has been pillared with OH-A1 species and modified with hexamethaphosphate ligand (B2alh) [15]. Table 3 - Cr retention from basic chromium sulphate salt by clays.
Clay A Ak
Cr retention from salt, mg/g 1 day 8 days 4 5 18 26
E Ek
6 27
8 44
Bl Blk
8 30
8 45
B2 Balh
7 33
7.5 47
The retention of chromium by A, E, and B1 increases after treatment from 4, 6, and 8 mg Cr/g to 18, 27, and 28 mg Cr/g after a 1-day contact, respectively (Table 3).
302
C. Volzone
The different behaviour between Ak and Ek is due to the different mineralogical composition. The pH of the suspensions of modified clays and chromium salt solutions were close to 5.5. The B2alh adsorbent reached a pH suspension of 4 and it presents similar chromium retention than Blk. The original samples reach the maximum equilibrium of retention after one day of contact; however, in the modified clays retention increases after eight days, suggesting a different mechanism. i) Influence of the chromium added The interlayer spacing, d(001), of a montmorillonite treated with polymeric Cr species (from P(l/60) solution), increases from 14.8 to 20.7 A when the amount of Cr is increased in the range of 0.5-20 mmol Cr(III) per gram of clay (Table 4) [47]. The high spacings when adding 10 and 20 mmol Cr/g correspond to gallery heights of 10.9 11.lA. This is the difference between d(001) spacing and 2:1 layer (9.6 A). Two layers of trimer-(Cr3(OH)45+), tetramer-(Cr4(OH)66+), or dimer-(Cr2(OH)24+) species have perhaps been intercalated between the 2:1 layer of smectites because these species are 5, 6.5, and 4 A in height, respectively [77,87]. Figure 7 shows the characteristic spectra of the supernatants after montmorillonite treatments with different mmol Cr/g added (S-M-P0.5, S-M-P1.5, S-MP3.5, S-M-P5, S-M-P10, and S-M-P20) [77]. Table 4 - d(001) spacing after different Cr added. Sample d(001), A
M 14.8
M-P0.5 15.7
M-Pl .5 > 16i >
M-P3.5 18.6
M-P5 19.6
M-P10 20.5
M-P20 20.7
Figure 7 - Absorption in visible spectra and pH of the OH-Cr-solution P(l/60), and supernatant after treatment with different mmol Crper gram of sample [77].
Removal of Metals by Natural and Modified Clays
303
The starting solution, P(l/60), shows two bands at 424 and 586 nm. The supernatant after treatment with 0.5 mmol Cr/g is colourless and then does not show bands in the spectrum. As a consequence, the chromium added is retained by montmorillonite. The bands of S-S-P1.5 supernatant shifted the peaks to smaller wavelengths of 415 and 583 nm, this behaviour indicates that species in this solution are less highly polymerised forms. As a result, selectivity of the montmorillonite to retain specific polymerised species do occur but, if necessary, other species will be retained, as it has been demonstrated when S-M-0.5 did not show Cr species, because the montmorillonite has retained all species present in solution. After treatment of montmorillonite with the P(l/60) solution, the original interlayer cations are removed and then they are found in the supernatant solutions [77]. The different interlayer cations from montmorillonite [77] are now in the supernatants and the pH of the supernatants (Figure 7) does not allow evaluating the OH-Cr-species present in such solutions for certain. The textural characteristics of the original montmorillonite and after adding lOmmol Cr/g as monomeric (M-M10) and polymeric (M-P10) Cr-species are shown in Table 5, where the polymeric solution is P(l/60). The BET surface of the M montmorillonite increases, after intercalation with OH-Cr species, from 36 to 175 m2/g (similarly, the surface area calculated by t-method also increases from 38 to 170 m2/g). The micropore and external surfaces, calculated from t-plot [90], show that the main difference between the original montmorillonite, M, and after intercalation with OH-Cr species (M-P10), is attributed to the micropore contribution, because the external surfaces are similar (25 and 32 m2/g), whereas the micropore surfaces are 7 and 138 for M and M-P10, respectively. The BET surface of intercalated montmorillonite with monomeric species, M-M10, shows a similar value to the original smectite, 40 m2/g. The intercalation of polymeric chromium species onto montmorillonite increases textural characteristics such as micropore surface and consequently the total surface. Table 5 - Some textural characteristics of montmorillonite (M) and after treatment with polymeric (M-P10) and monomeric (M-M10) species, by addition of 10 mmolCr/g. Sine: surface from t plot, Se: external surface, Sm: micropore surface. Surface, m2/g BET surface Sme Se Sm
M 36 38 25 7
M-P10 175 170 32 138
M-M10 40 — — -
ii) Influence of the types of clays The adsorption isotherms for Cr(III) from chromium nitrate follow Langmuir equation according to Griffin et al [70], in which the chromium species is Cr(H2O)63+. They found the retention values of 79.5 meq/100 g and 15.1meq/100g for montmorillonite and kaolinite, respectively [70]. The cation exchange has been the main mechanism for cation adsorption. Similar behaviour has been observed by Volzone & Tavani [65] and Tavani & Volzone [66] after retention of chromium (III) from a tanning waste solution by kaolmite and montmorillonite clays. Cr (III) is also absorbed between
304
C. Volzone
30 and 300 times more than Cr(VI) by kaolinite and montmorillonite in the pH range of 1.5 to 4.0 as mentioned by Griffin et al [70]. The adsorption of chromium (III) from chromium nitrate solution by chlorite, illite, and kaolinite is pH-dependent [71]. The chromium is adsorbed as Cr(III) aqua at a pH value below 4, however, at a higher pH it is probable that polymerised chromium ion, Cr3(OH)45+, is present. The retention of chromium by illite has been similar to that of chlorite and both are higher those of kaolinite, in accordance with cation exchange capacity [71]. Koppelman and Dillard [72] have mentioned that hydrolysis of chromium occurs after adsorption in clays similar to what has been mentioned for smectites when they have been intercalated with OH-Cr species [20,22,23,73]. The basal spacing of a kaolinite and a vermiculite treated with polymerised chromium (OH-Cr species from the P(l/60) solution) is similar before treatment, 7.3 and 14.9 A, respectively [78], nevertheless the basal spacing for Cheto type montmorillonite (MCh), Wyoming type montmorillonite (MW), a beidellite (B), a nontronite (N) and a saponite (S) have been 20.6, 20.7, 20.4, 20.1, and 20.2 A, respectively. As a consequence, the chromium intercalated in smectites showed a separation of 10.9-11.1 A, calculated by the difference between measured basal spacing and the height of the corresponding layer (2:1 layer at 9.5 A). The chromium species are placed in these separations according the dimension of the species [73,88] and as mentioned in 3.1.2.2. The different basal spacing values between 16.8 and 27.8 A of the montmorillonite-OH-Cr complexes reported by Rengasamy and Oades [19], Brindley and Yamanaka [20], Pinnavaia et al [21], Tzou and Pinnavaia [22], and Volzone [77,79,91], may be attributed to different intercalation procedures (hydroxy-chromiumsolution, hydroxy chromium-smectite, Cr added, etc.). The N2 adsorption-desorption isotherms of the clays before and after OH-Crsolution treatment (not shown) corresponded to H3 type according to the classification by Gregg and Sing [92]. This type is a hysteresis loop with a vertical adsorption branch at a relative pressure very close to 1, and a desorption branch close to medium pressure. Such hysteresis loop may be formed due to slit-shaped pores [92], The pore shapes are preserved after treatment with OH-Cr-species, but with different volumes (although with the initial volume in the low pressure region of the isotherm), except for kaolinite which volume is unaffected [78]. Figure 8 shows the BET surface area of the clays before and after OH-Cr treatment, and also micropore surface/surface BET ratio. Micropore and the external surface area, and the micropore volume of the samples have been obtained by the t-plot method [90], and for kaolin clay the internal and external surface areas have been derived according to the method of Delon et al[93] that assumes slit shaped pores. The treatment of the clays with OH-Cr-species increased BET surface mainly for smectite clays. The intercalation of OH-Cr species in the clays increased the micropore surface in a higher proportion when Smic/SBET increased [78,91]. Micropore volumes in smectites increase, slightly affecting the micropore volume in vermiculite without affecting the micropore volume in kaolinite after OH-Cr treatment [78]. The micropore volume to total volume ratio increases after chromium species treatment in the vermiculite and the smectite clays, and remains unchanged in kaolinite (Figure 9). The amount of chromium (III) uptake (expressed as % Cr2O3) is a function of the octahedral charge of the original clays (Table 6), as shown in Figure 10. Similar correlation has been obtained after intercalating aluminium polymeric species in smectites [94]. The small amount of Cr retained by clays with no negative octahedral charge might be attributed to defects in clays.
Removal of Metals by Natural and Modified Clays
305
Figure 8 - BET (Brunauer-Emmett-Teller) surface of natural clays and after treatment with OH-Cr species solution. Smic: micropore surface, SBET-' BET surface.
Figure 9 - Total and micropore volume of the natural clays and after OH-Cr-species solutions treatment.
306
C. Volzone
Table 6 shows the composition of the half unit cell of each original clay that has been calculated from < 2 um fraction [26]. The major negative charges in smectites are originated in the octahedral sheet, whereas in kaolinite and vermiculite they are located in the tetrahedral sheet. This vermiculite shows positive charge in the octahedral sheet. Table 6 - Composition of half-unit cell of the natural clays [31,77,78].
Clay K
Si,v 4+ 3.955
V
2.890
MCh MW B N S
3.940 3.910 3.610 3.700 3.820
Te trahedral sh eet Charge A1 IV 3+ -0.046 0.046
AW +
1.100
-1.100
0.080
0.060
2.810
+0.040
0.060 0.090 0.390 0.300 0.180
-0.060 -0.090 -0.390 -0.300 -0.180
1.360 1.610 1.630 0.130 0.200
0.060 0.130 0.110 1.800 0.020
0.600 0.260 0.160 0.130 2.500
-0.600 -0.260 -0.160 -0.100 -0.190
3.852
Octahedr al sheet Mg 2 + Fe 3 + 0.018 0.003
charge 0
Figure 10 - Retention of chromium from the P(l/60) solution, expressed as a percentage ofCr2Os vs. the octahedral charge of the clays (Table 6). 3.2.2 - Effects of thermal treatment on Cr-clays The structural, textural and acidity changes in chromium clays due to thermal treatment allow, for example, the evaluation of the thermal stability of Cr-pillared clays, the possible immobilisation of chromium, etc. The larger endothermic peak of original smectites is produced by dehydration of the clay (water intercalated between layers) [74,95]. The size, shape, and temperature depend on the saturating cation. When the smectites are treated with the OH-Cr-solution,
Removal of Metals by Natural and Modified Clays
307
P(l/60) solution, the exchangeable cations are replaced by OH-Cr-species. The differential thermal analysis, DTA, diagrams of all OH-Cr- smectites show a double peak (140°C and 190°C), and correspond to dehydration of OH-Cr-species. This behaviour has been confirmed by means of a DTA diagram of the OH-Crmontmorillonite with the same montmorillonite [74]. The thermal stability of the interlayer spacing up to 600°C in air atmosphere of one intercalated montmorillonite with different chromium species such as monomeric, trimeric, and polymeric species are shown in Table 7. A typical reduction of interlamellar spacing has been observed in montmorillonite, M, which collapses from 14.8 to 9.3 A after thermal treatment. Similar behaviour is shown by the montmorillonite intercalated with monomeric species, M-M sample (15.0 to 9.4 A). The clay treated with trimeric species, M-T, is more stable than the M-M sample. A better thermal stability is shown for montmorillonite treated with chromium species with higher polymerised species, M-P(l/60), but all samples collapsed at 600cC (9.6 A). The intercalated montmorillonite with OH-Cr polymerised species by Tzou and Pinnavaia [22] and Brindley and Yamanaka [20] shiftted the d(001) spacing from 27.6 to 10.2 A, and from 16.8 to 9.8 A, after thermal treatment from 25 to 450°C, respectively (Table 7). As a consequence, the high polymeric species causes a larger spacing and better thermal stability in the smectite. Table 7 - Thermal stability of the interlayer spacing of the intercalated Cr species in montmorillonite. Temperature M-M M (°C) (A) (A) 14.8 15.0 25 14.2 14.0 200 12.5 12.0 300 9.3 9.4 450 9.3 9.4 600 * Brindley and Yamanaka [20] ** Tzou and Pinnavaia [22]
TP**
M-T
M-P(l/60)
M- P(30/20)
BY*
(A)
(A)
(A)
(A)
(A)
18.4 14.8 14.6 11.9 9.2
20.7 19.9 18.0 14.5 9.4
19.7 17.0 16.0 9.0, 14.0 9.0
16.8 16.8 10.4 9.8
27.6 24.0 22.0 10.2
i) Heating up to 450°C to obtain Cr-PILCs The PILCs are obtained by cation exchange of polynuclear hydroxy-cation species between the aluminosilicate layers followed by calcination [17]. Cr-PILCs have been prepared from smectites clay by Rengasamy and Oades [19], Brindley and Yamanaka [20], Vaughan and Lussier [17], Carr [23], Pinnavaia et al [21], Vaughan [18], Tzou and Pinnavaia [22], Drljara et al [96], Volzone [31]. Figure 11 shows that the d(001) spacing of the starting smectites (sm) is in the range of 13.5 - 15.5 A. These values are a function of natural saturation interlayer cations of the smectites [26,74]. After heating at 450°C (sm-450), the spacing of smectites decreased at around 9.5 A (the depth of the individual triple-sheet-layer) due to dehydration of smectites (water intercaled between layers) [31]. The intercalation of OHCr species increases the d(001) spacings of the smectites (OH-Cr-sm) to 20.1-20.7 A, Figure 11. The OH-Cr-sm are converted to Cr pillared smectites by heating at 450°C, at a heating rate of 2 °C/min in N2 atmosphere (Cr-PILCs/N, Figure 11), and the d(001) spacing of these products are in the range of 13.8 - 18.8 A. The difference between the
308
C. Volzone
values of the d(001) spacing of Cr-PILCs/N and the height of the 2:1 layer (9.5 A) corresponds to an interlayer separation or gallery in the range of 4.3 - 9.7 A, according to smectite members used to prepare Cr-PILCs/N. The treatment of the OH-Cr-sm at 450°C in air atmosphere (Cr-PILCs/A) originate a reduction of the d(001) spacing of around 15 A, except for pillared nontronite (N) that shows a low value (10 A) [Figure 11]. This behaviour is associated with the dehydroxylation of the 2:1 layer of the nontronite that occurs at 490°C [95]. The a-Cr2O3 phase is observed in the X-ray diffraction spectra of the OH-Cr smectites heated at 450°C in air atmosphere, and it coincides with an exothermic peak around 420°C in thermal gravimetry analysis that corresponded to the dehydration of OH-Cr-species and the crystallisation of Cr2O3 [74,77]. The Cr2O3 phase is not observed in the XRD of the OH-Cr smectites heated at 450°C in a nitrogen atmosphere [31,77] and therefore, the poly-oxo-hydroxycations may be present in the interlayer position of the Cr-PILCs/N. This agrees with Tzou and Pinnavaia [22] who have worked with Cr-montmorillonite.
Figure 11 - Interlayer d(001) spacing of the starting smectites and after different conditions. MCh: Cheto-type montmorillonite, MW: Wyoming type montmorillonite, B: beidellite, N: nontronite, S: saponite [31]. The textural characteristics of the Cr-PILCs obtained in both atmospheres are shown in Table 8. The BET surface area of the OH-Cr-sm after pillaring by heating at 450°C in N2 atmosphere (Cr-PILCs/N) reached 50 - 165 m2/g in a different way depending on the smectite members The contribution in micropore surface area, by tmethod [90], regarding total surface (micropore surface area + external surface area) of the Cr pillaring smectites in N2 atmosphere are in the range of 53 - 87 %. The total volume (Vtot) obtained at the end of the N2 adsorption isotherm branch (P/Po = 0.986), increases from 0.023 - 0.174 cnrVg for starting smectites to 0.072 - 0.304 cm3/g for Cr pillared smectites obtained in a N2 atmosphere. The micropore volume (Vmic) has been obtained by a high pressure branch extrapolated to the
309
Removal of Metals by Natural and Modified Clays
adsorption axis in the t-plot (90). The contribution in Vmic of the Cr pillared smectites (in N2 atmosphere) increases in the range of 1.37 to 10.3 times with respect to the original smectites. The micropore volume and the total volume ratio (Vmic/Vtot) increase after pillaring with respect to the ratio for the original smectites. The values of micropore surface area and micropore volume of the Cr-PILCs/N increases with the interlayer spacing [Figure 11 and Table 8]. Table 8 - Textural characteristics Cr-PILCs obtained in nitrogen (N) and air (A) atmospheres. SBET: BET surface area; Smic: micropore surface area; Se: external surface area; Vtot: total volume; Vmic: micropore volume; Vmic/Vtot: micropore volume/total volume ratio [31]. SBET
Sample MCh Cr-PILCMCh/N Cr-PILCMCh/A
m2/g 81 165 50
Smic m2/g 71 143 25
Se m2/g 7 34 12
Vtot cm3/g 0.174 0.140 0.058
Vmic Cm3/g 0.039 0.075 0.020
Vmic/Vtot % 22 54 34
MW Cr-PILCMW/N Cr-PILCMW/A
19 128 11
12 68 5
4 62 1.5
0.074 0.304 0.065
0.003 0.031 0.005
4 10 8
B Cr-PILCB/N Cr-PILCB/A
7 78 10
6 54 5.5
2 36 2
0.044 0.107 0.042
0.005 0.028 0.005
11 26 12
N Cr-PILCN/N Cr-PILCN/A
27 50 4
14 36 1.5
2 21 1.5
0.051 0.072 0.033
0.016 0.022 0.002
32 31 6
S Cr-PILCS/N Cr-PILCS/A
8 62 8
6 48 2
2 25 1.8
0.023 0.091 0.067
0.005 0.028 0.039
22 31 58
The Cr-PILCs obtained in air condition (Cr-PILCs/A) shows low textural characteristics, and this could be attributed to the reduction of interlayer spacing and/or the aggregation of Cr2O3 species in the products. The micropore surface area, the micropore volume, and d(001) spacing of the Cr pillaring smectites increases with the amounts of the chromium retained (1.33 - 2.48 mmol Cr/g) [31] Cr-PILCs obtained at 300°C and 450°C show Lewis (L) and Bronsted (B) acidity characteristics. L/(L+B) ratio of the Cr-PILCs increases with the tetrahedral negative charge of the smectite structures as demonstrated by Volzone [31]. The structural charges and compositions of the original smectites (Table 6) influence the structural, textural, and acidity characteristics of the Cr-PILCs. ii — Heating at high temperatures The dioctahedral smectites show an endothermic peak between 450-750°C in TGA diagram, a fact that proves the existence of dehydroxylation (loss of water of
310
C. Volzone
constitution) [95]. These differences must be related in some way to the energy to which the hydroxyl groups are bound in the lattice [97], and to its chemical composition, as it is also seen in Table 6. The nontronite, N shows one breadth endothermic dehydroxylation peak (490°C) below the one obtained for beidellite, B, (560°C), and for montmorillonites, MW, (640°C-710°C). Saponite, S, lost the OH at high temperature (720°C). This behaviour is characteristic of the trioctahedral smectites [95]. In montmorillonite, MW, a small S-shaped endothermic-exothermic peak at about 850°C-950oC corresponds to the destruction of the lattice and the recrystallisation into new phases respectively [95]. The endothermic peaks of nontronite, N, and beidellite, B, are absent, although the exothermic one occurs at 900°C for N and at 980°C for B. According to Kerr [98] such end-exothermic occurs simultaneously. The endothermic peak at around 490oC-710°C of dioctahedral smectites (N, B, and MW) are shifted to lower temperature after the OH-Cr-treatment, (460°C-660°C). This peak depends, among other factors, on the perfect stacking of the layers and on gross substitution [97]. Nevertheless, the peaks at high temperatures of all smectites are shifted to higher values. The absence of the original exchangeable cations, which are replaced by Cr-species, tends to shift peaks to higher temperatures [74]. The smectites phases developed at 1000°C, with and without OH-Cr treatment, are shown in Table 9. Table 9 - Phases of smectites with and without OH-Cr treatment after thermal treatment at 1000 °C in air atmosphere [74]. Sample MCh MCh-Cr
Phases at 1000 °C (3-quartz, crystobalite, anorthite P-quartz, a-Cr2O3
MW MW-Cr
a-quartz, albite a-quartz, a-Cr2O3
N N-Cr
a-quartz, hematite, anorthite a-quartz, hematite (low), a-Cr2O3
B B-Cr
mullite, a-quartz a-quartz, a-Cr2O3
S S-Cr
enstatite enstatite, a-Cr2O3
The phases of the smectites are determined by bulk composition (Table 6), and exchangeable cations. The calcium present in the interlayer position of the nontronite, N, has originated anorthite, whereas the sodium content of the MW montmorillonite has originated albite. The high structural aluminium content of the beidellite, B, favours mullite formation, whereas the high structural magnesium content in saponite, S, (Table 6) allows the presence of the enstatite phase. The iron content of nontronite has been the hematite phase. Quartz is present in these dioctahedral smectites.
Removal of Metals by Natural and Modified Clays
311
The a-Cr2O3 phase is present in all OH-Cr-smectites treated from 420cC up to 1000°C. No alteration of the Cr(III) oxidation number has been observed throughout all thermal treatment up to 1000°C. Anorthite and albite phases are not present in the XRD diagrams of the MCh-Cr and MW-Cr smectites, respectively, treated at 1000°C. This could be attributed to the fact that the exchangeable cations (Ca,Na) present in the original smectites have been replaced during the treatment with the OH-Cr-solution, and then those phases could not develop at higher temperatures. With respect to a- and (3-quartz, hematite and enstatite phases are present in smaller proportion. The mullite is not present in the beidellite-Cr (B-Cr) smectite up to 1000°C. As to the OH-Cr(III)-species retained by the smectites, no alteration in its oxidation number has been observed when subject to an air thermal treatment. Nevertheless, if the same chromium species in solution is supported by an inert substance and subjected to the same procedure, Cr(VI) has been found after thermal treatment [99]. A comparison of the products obtained from OH-Cr montmorillonite heated in air and nitrogen atmosphere up to 1000 °C is shown in Table 10. Table 10 - Phases of OH-Cr-M at different temperature up to 1000 °C in air and nitrogen atmospheres [76]. OH-Cr-M Temperature °C 25
Air atmosphere
Nitrogen atmosphere
montmorillonite (d(001) spacing: 20.5 A) ct-quartz (impurity) a-cristobalite (impurity)
montmorillonite (d(001) spacing: 20.5 A) a-quartz (impurity) a-cristobalite (impurity)
450
montmorillonite (d(001) spacing: 14.0 A ) a-quartz (impurity) a-cristobalite (impurity) aCr2O3 (eskolaite)
montmorillonite (d(001) spacing: 18.6 A) a-quartz (impurity) a-cristobalite (impurity)
1000
u-cordierite P-quartz (impurity) P-cristobalite a-Cr2O3 (eskolaite)
P-cristobalite Mg-Cr-spinel
The a-Cr2O3 phase has been found in OH-Cr-M by heating in air atmosphere, which appears at 450°C and remains up to 1000°C (the last temperature analysed). The a-Cr2O3 phase is absent in OH-Cr-M heated in a nitrogen atmosphere. Nevertheless, at 1000 °C thermal treatment shows that the Mg-Cr-spinel is present. The thermal treatment of the intercalated OH-Cr montmorillonite at 1000°C, where the montmorillonite has already collapsed, also presented differences in the products: eskolaite (a-Cr2O3) in air, and Mg-Cr-spinel in nitrogen, where the oxidised status of the chromium has been mantained, Cr(III).
312
C. Volzone
3.3 - Retention chromium from wastewater by clays This section analyses quadrants III and IV in Figure 3 (c). Although precipitation, in a physical-chemical treatment, is the most utilized method for retaining high chromium concentration in effluents, clays could be potential adsorbents of chromium in low concentration from effluents. Common wastewater is tanning waste where concentration of sodium is higher than that of chromium because high NaCl concentrations come from the tanning process. Volzone and Tavani [65] have studied the adsorption of Na and Cr(III) ions by smectite using tanning wastewater containing 13.11 and 0.81 mmol L"1 of each cation, respectively. Nevertheless, the highest adsorptions of Na and Cr (III) have been 0.30 and 0.80 mmol per gram of smectite, respectively. Such behaviour is attributed to the different density of the electric charge of both cations. The same authors [66] have also found a similar behaviour by using kaolinite as an adsorbent, where 0.008 and 0.036 mg per gram kaolinite of sodium and chromium have been retained. An illite clay also showed a similar behaviour [67]. Exchange capacity has been the mechanism proposed and the predominant chromium (III) species in wastewater have been monomeric chromium species. Modified kaolin with varied mineralogical composition influence chromium retention from tanning waste. The main three components, in different proportions, of raw clays (A, B, C, D, and E) are kaolinite, illite, and quartz. The composition of each of the clay is shown in a triangular plot in Figure 12 [89], where the C sample is a component rich in illite.
Figure 12 - Mineralogical composition of some clays used as raw materials [89]. The clays have been modified by treatments with caustic potash solutions; stabilised at pH 7.0; and there were a thermal treatment at 500 K (Ak, Bk, Ck, Dk, and Ek).
Removal of Metals by Natural and Modified Clays
313
The waste liquid used is a tanning waste solution provided by INESCOP as mentioned in section 3.2. Once suspended, solids have been removed by filtration. Then, this solution has been contacted with different clays at different times and at 25 °C and with a solid/solution ratio of 5.5 % in a batch system [89]. Figure 13 shows Cr retention by modified clays as a function of contact time; and Table 11 depicts retention values after 1 and 8 days. In general, extending contact time favours the uptake of chromium, although its amount increases in different proportion as a function of the mineralogical composition. A higher kaolinite content with a low quartz amount in raw clay favours retention in the modified sample (Ek). The mineralogical composition of the raw kaolin material (eg A and E), [Figure 12] influence the amount of retention chromium both from tanning and salt, as it is observed in Tables 3 and 11. The presence of illite and quartz in kaolin clays has a considerable influence on the value of retention chromium.
Figure 13 - Retention of chromium from Cr tanning wastewater by modified kaolinitic and illitic clays. Three bentonites with a high content of Wyoming (Bl) and Cheto (B2) types montmorillonites, and saponite (S), treated in similar conditions as kaolins, have also showed an important retention of chromium from wastewater [Table 11 and Figure 14]. Values are higher than those for modified kaolin, except for Ek, whose retention values after 1 and 8 days are similar to modified montmorillonite, Blk, Table 11. The pHs of the suspensions are close to 5.5 and some of the chromium (III) may precipitate on clay [71]. The B2 bentonite has been modified by OH-A1 intercalation, B2al, [15]. Cr retention from tannery waste by B2al is lower than the one by untreated bentonite (B2: 6.5 and B2al 6 mgCr/g after a one-day contact time). In general, OH-A1 intercalation reduces cation exchange capacity, thus hindering a high Cr retention [15]. To improve retention, hexametaphosphate ions (HPM) were added to B2al (B2alh). Later, the sample
314
C. Volzone
has been given thermal by heating at 500°C (B2alh5). Retention of Cr has increased as shown in Figure 14 and Table 11. Adsorption capacity increased because A1-0H groups were available for reaction with HMP; and the high amount of retained Cr may be related to the presence of surface phosphate groups. After heating at 500°C, structural properties remain unchanged but there is a possible dehydration of an external Al phosphate. The high amount of retained Cr at a low pH (close to 4) may be related to the presence of surface phosphate groups [15] in accordance with phosphatic soils [4]. The retention of Cr by modified Al-bentonite with HMP after heating at 500°C increases more than twice if compared to the original bentonite, Table 11. Table 11 - Cr retention from tannery waste by modified clays.
Sample Ak Bk Ck Dk Ek Blk B2k Sk B2al B2alh B2alh5
Cr retention from tanning wastewater mg/g 1 day 12.0 9.5 7.5 6.5 16.5 16.0 17.0 18.0 6.0 10.0 10.0
8 days 17.0 15.0 8.0 7.5 22.0 30.0 26.0 35.0 8.0 13.0 18.0
The B2k sample retains more chromium than the B2alh5 (Table 11). Nevertheless, mechanical resistance in pellet form is reversed. This property is very important if the solid adsorbent is used in percolation systems. Several experiments about retention chromium in solutions have been carried out using powdered clays and in pellets to evaluate mechanical resistance in a liquid media. More retained Cr (III) from chromium salt is obtained than from tanning wastewater by clays as shown in Tables 3 and 11. It can be seen from Figure 6, a basic chromium sulphate salt contains more polymerised Cr species than tanning wastewater [Figure 6(b)] [84]. As a consequence, more species from a salt solution are required to compensate the negative charge of the clay, due to the low charge per Cr atom. In general, chromium retention is a function of type of clay, activation treatment, and chromium solution used. Clays may also be used in dye and colloid retention from a diluted tanning wastewater [100]. Figure 15 (a) shows the spectra of a diluted chromium (50 ppm Cr) tanning waste containing dye and colloids after a 24-hour contact with two different clays. A natural montmorillonite, m, and the same montmorillonite after acid treatment with sulphuric acid (ml 8) are used as adsorbents. The spectrum of the initial solution shows bands at 450, 581, and 619 nm; the band at 581nm could correspond to the presence of chromium.
Removal of Metals by Natural and Modified Clays
315
Figure 14 - Chromium retention from Cr tanning wastewater by modified bentonites.
Figure 15 - Absorption spectra: (a). Tanning wastewater containing low chromium, dye, and colloids before and after treatments with natural montmorillonite (m) and acid montmorillonite (ml8). (b) An aniline solution prepared in laboratory after being treated with acid montmorillonite (+ml8). The high background in the spectrum is due to the colloids in the solution. Clays retain mainly the colloids. For a comparative analysis, one solution of dye, the aniline commonly used in tannery, has been reported [Figure 15 (b)]. The spectrum of the supernatant after the contact of acid montmorillonite with aniline solution is shown also in Figure 15 (b). The acid montmorillonite has been previously calcinated at 400 c C. Clays can interact with organic substances by intercalation, adsorption and cation exchange [101]. Organic substances, such as aniline, are strongly adsorbed by clay
316
C. Volzone
minerals [102], The acid nature (Bronsted acid) of the 2:1 layer silicates facilities the protonation of organic species [103]. Espantaleon et al [104] have studied removal dyes from solution by using acid-activated bentonite with a sulphuric acid solution at different concentration. The removal of dyes in this case has been better than with carbon and natural bentonite [104]. Acid treatment on clays should be controlled because acid attack may break the clay structure [36-38] thus loosing the adsorptive characteristics. In general, the best textural characteristics of acid bentonites are obtained when close to 75% of the octahedral cations is released from the structure by acid treatment. 4 - Can we use clays to improve the environment? There is no easy answer to this question. However, if we analyse the variables that play an important role in the retention of metals by clays carefully, it is possible to use clays to have a better environment, and then we would also be capable of controlling the disposal of metal in soils. Nowadays, there are many researchers all over the world studying the metal retention mechanism from solutions by clays and/or soils. A contribution about different aspects of the removal of chromium (III) from solutions by clays has been described in this chapter. However, the answer is not complete as there are more questions to be tackled by an efficient and thorough study, i.e. what do we make with a metal-clay product?; how do we use the resulting Cr-clay?; is it inert forever? An answer to the use of Cr-clays may be, for example, preparing pillared clay (PILCs), as long as we take into account the influence of the variables, i.e. impurity of clays, interference of chromium in solutions, etc. As to how we may obtain an inert material from chromium clay, the answer may be, for example, after an appropriate calcination of Cr-clays. It would then be possible to obtain refractory products. These studies should be carefully analysed since, depending on the different types of atmosphere (i.e. air, oxygen, nitrogen, etc.) used during heating, it is possible to obtain different products from the same Cr-clay as explained in this chapter. Mass balance should also be evaluated to check if chromium evaporates. It is important to take into account that chromium (III) species that cannot intercalate onto solids may be oxidised to Cr (VI) when the solid is treated at high temperature, as it occurs; for example, with a-Al2O3, where the chromium (III) species is only deposited on the surface [99]. It is necessary to keep in mind that it is important not to create pollutants in the environment from industrial activities, and to replace them by using new alternative technologies. Nevertheless, there is pollution in our planet that is harmful to human, animal, and plant life. Acknowlegement It was a pleasure for me to take part in Project V.6-CYTED because interesting experience on removal of chromium from wastewater was acquired. The invitation by Editor Prof. F. Wypych to write this chapter is gratefully acknowledged. I would like to thank all the publishers that allowed me to quote the tables, figures, and pieces of information in the references. I wish to thank Lie. A. M. Cesio for our discussions about chromium chemistry. Finally, I thank my family for their support, and I am particularly grateful to my friend and translator Alicia Bernatene for helping me editing this paper and to my son, Lucas Marchel, for helping me with some figures.
Removal of Metals by Natural and Modified Clays
317
5 — References [I] H.M. Lima and A.C. Araujo, in: Mineral Precessing and Environment, IV Meeting of the Southern Hemisphere on Mineral Technology, ed. by S.Castro and F. Concha. Univesidad de Concepcion, Concepcion, Chile, 1994. [2] G. Blanchard, M. Maunaye and G. Martin, Water Res., 18 (1984) 1501. [3] J. Rubio, I.A.H. Schneider and W. Aliaga, in: Clean Technology for the Mining Industry, ed. by M.A. Sanchez, F. Vergara, S.H. Castro, Universidad de Concepcion, Concepcion, Chile, 1996. [4] S.P. Singh, L.Q. Ma and W.G. Harris, J. Environ. Qual., 30 (2001) 1961. [5] G. Rodriguez Fuentes and P. Avila Garcia, Catalizadores y Adsorbentes Iberoamericanos para la Remocion de Metales Pesados de Efluentes Industriales, Ediciones CYTED, Espafla, 2000. [6] I. Garcia, A. Sanchez, E. Alvarez Ayuso and O. Jimenez De Bias, Clay Miner., 34 (1999)460. [07] S.W. Bailey, in: Crystal Structures of Clays Minerals and their X-ray identification, ed. by G.W. Brindley and G. Brown, Mineralogical Society, London, 1980. [08] D.M.C. MacEwan and M.J. Wilson, in: Crystal Structures of Clays Minerals and their X-ray identification, ed. By G.W. Brindley and G. Brown, Mineralogical Society, London, 1980. [09] W. Matthes, F.T. Madsen and G. Kahr, Clays Clay Miner., 47 (1999) 617. [10] J.M. Zachara and J.C. Westell, in: Soil Physical Chemistry, ed. by D.L. Sparks, 2nd ed., CRC Press, USA (1999). [II] H. Triantafyllou, S.E. Christodoulou and P. Neou-Syngouna, Clays Clay Miner., 47(1999)567. [12] M.L. Occelli and R.J. Rennard, Cat. Today, 2 (1988) 309. [13] H. Ming-Yuan, L. Zhonghui and M. Enze, Cat. Today, 2 (1988) 321. [14] L. Mercier and TJ.Pinnavaia, Microp. Mesop. Mat., 20(1998) 101. [15] C. Volzone and L.B. Garrido, Ceramica, 48, 307 (2002) 153. With kind permission from Ceramica for reproducing information, ceramicafajipen.br [16] C. Volzone and L.B. Garrido, VI Reunion Anual de SETAC Latinoamerica. Buenos Aires, Argentina, 20-23 de Octubre de 2003, Abs. (2003)113. [17] D.E.W. Vaughan and R.J. Lussier, in: Proc. 5th Int. Zeol. Conf, ed. by L.V.C. Rees, Heyden Press, London, 1980. [18] D.E.W. Vaughan, in: Perspectives in Molecular Sieve Science, ed. by W.H. Flank and T.E. White, American Chemical Society, Washington, D. C.,1988. [19] P. Rengasamy and J.M. Oades, Aust. J. Soil Res., 16 (1978) 53. [20] G.W. Brindley and S. Yamanaka, Am. Miner., 64 (1979) 830. [21] T.J. Pinnavaia, M-S. Tzou and S.D. Landau, J. Am. Chem. Soc, 107 (1985) 4783. [22] M.S. Tzou and T.J. Pinnavaia, Cat. Today, 2 (1988) 243. [23] M.R. Carr, Clays Clay Miner., 33 (1985) 357. [24] R. Burch, Catalysis today: Pillared Clays, Elsevier Science Publishers, Netherlands, 1988. [25] C. Volzone and L.B.Garrido, Ceramica, 42, 275 (1996) 217. [26] C. Volzone, Doctoral Thesis, Facultad de Ingenieria. Universidad Nacional La Plata, 1997. [27] L. B. Garrido, C. Volzone and R. M. Torres Sanchez, Colloids and Surfaces A: Physic. Engineer. Asp., 121 (1997) 163.
318
C. Volzone
[28] L.B. Garrido and C. Volzone, Anais do 42° Congresso Brasileiro de Ceramica, 3-6 de Junio, Pocos de Caldas, Minas Gerais, Brasil, vol I (1998) 158. [29] C. Volzone, L.B. Garrido, J. Ortiga and E. Pereira, in: Desarrollo de adsorbentes para la separation de gases, ed. by F.Rodriguez Reinoso-P.Andreu, Ediciones CYTED, Espafia, 1998. [30] C. Volzone andN.E. Hipedinger, Clays Clay Miner., 47 (1999) 109. [31] C. Volzone, Microp. Mesop. Mat., 49 (1-3) (2001) 197. With kind permission of Elsevier for reproducing information and parts of Figure 1, Table 1 and Table 3 of the paper (Figure 11, part of Table 6, Table 8 in this chapter), www.elsevier.com [32] GJ. Churchman and C. Volzone, Proceeding of the 12* International Clay Conference, Bahia Blanca, Argentina, July 22-28, 2001, eds E.A.Dominguez, G.R. Mas, F. Cravero. Elsevier, (2003) 31 [33] A.C. Vieira Coelho, P. Souza Santos, C. Volzone and L.D.V. Abreu, Proceeding of the 12th International Clay Conference, Bahia Blanca, Argentina, July 22-28, 2001, eds E.A.Dominguez, G. R. Mas, F. Cravero. Elsevier. (2003) 655. [34] A. Ackerman, Chemie at Industrie, 61 (1949) 29. [35] K. Bruckman, F. Fijal, J. Haber, Z. Klapyta, T. Wiltowski and W. Zabiniski, Miner. Polonia, 7 (1976) 5. [36] P. Souza Santos, Tecnologia de Argilas. Ed. Edgard Bltlcher Ltda., 1975. [37] C. Volzone, J.M. Porto Lopez and E. Pereira, Rev. Latinoam. Ing. Quim. Quim. Apl., 16(1986)205. [38] G.E. Christidis, P.W. Scott and A.C. Dunham, Appl. Clay Sci., 12 (1997) 329. [39] E.F. Folleto, C. Volzone and L.M. Porto, Braz. J. Chem. Eng., 20 (2003) 139. [40] J.C. Echeverria, E. Churio and J.J. Garrido, Clays Clay Miner., 50 (2002) 614. [41] T.M. Christensen, Water Air Soil Poll., 21 (1984)105. [42] B. Zhu and A.K. Alva, Soil. Sci., 155 (1993) 61. [43] P.L. Carey, R.G. McLaren and J.A. Adams, Water Air Soil Poll., 87 (1996) 189. [44] C. Airoldi and S.A.M. Crittes, Clays Clay Miner., 45 (1997) 125. [45] K.A. Bolton and L.J. Evans, Can. J. Soil Sci., 76 (1996) 183. [46] J.M. Zachara, S.C. Smith, C.T. Resch and C.E. Cowan, Soil Sci. Soc. Am. J., 56 (1992) 1074. [47] T. Undabeytia, S. Nir, G. Rytwo, E. Morillo and C. Maqueda., Clays Clay Miner., 46(1998)423. [48] P. Di Leo and P. O'Brien, Clays Clay Miner., 47 (1999) 761. [49] J-B. Yoon, S-H. Hwang and J-H. Choy, Bull. Korean Chem. Soc, 21 (2000) 989. [50] Y. Egozy, Clays Clay Miner., 28 (1980) 311. [51] G.F. Soldatini, R. Riffaldi and R.L. Minzi, Water, Air and Soil Poll, 22 (1976) 110. [52] H.A. Gharaie, M. Maftoun and N. Karimian, 17th WCSS, 14-21 August 2002, Thailand, (2002) 1961. [53] C.A. Coles and R.N. Yong, Appl. Clay Sci., 22 (2002) 39. [54] J.C. Miranda-Trevino and C.A. Coles, Appl. Clay Sci., 23 (2003) 133. [55] L. Mercier and C. Detellier, Environ. Sci. Technol, 29 (1995) 1318. [56 ] Y. Yin, H.E. Allen, C.P. Huang, D.L. Sparks and P.F. Sanders, Environ. Sci. Technol., 31(1997)496. [57] D. Sarkar, M.E. Essington and K.C. Misra, Soil Sci Soc. Am. J., 64 (2000) 1968. [58] J.M. Zachara, C.E. Cowan, R.L. Schmidt and C.C. Ainsworth, Clays Clay Miner., 36(1988)317.
Removal of Metals by Natural and Modified Clays
319
[59] L.M. He, L.W. Zelazny, V.C. Baligar, K.D. Ritchey and D.C. Martens, Soil Sci. Soc. Am. J., 61(1997)784. [60] O. Yavuz, Y. Altunkaynak and F.Guzel, Water Res., 37 (2003) 948. [61] O. Abollino, M. Aceto, M. Malandrino, C. Sarzanini and E. Mentasti, Water Res., 37 (2003) 1619. [62] K. Naganuma, M. Okazaki, K. Yonebayashi, K. Kyuma, P. Vijarnsorn and Z.A. Bakar, Soil Sci. Plant Nutr., 39 (1993) 455. [63] J.L. Venaruzzo, C. Volzone, J. Ortiga and A. Ortiz Ricardi, VI Reunion Anual de SETAC Latinoamerica, Buenos Aires, Argentina, 20-23 de Octubre de 2003, Abs. (2003) 96. [64] K. Wada and Y. Kakuto, Clays Clay Miner., 28 (1980) 321. [65] C. Volzone and E.L. Tavani, J. Soc. Leather Technol. Chem., 79 (1995) 148. [66] E.L.Tavani and C.Volzone, J. Soc. Leather Technol. Chem., 81 (1997) 143. [67] E. L. Tavani and C. Volzone, Jornadas SAM'98-IBEROMET V, Rosario, Argentina Tomol(1998)35. [68] R.J. Bartlett and J.M. Kimble, J. Environ. Qual., 5 (1976) 379. [69] R.J. Bartlett and J.M. Kimble, J. Environ. Qual., 5 (1976) 383. [70] R.A. Griffin, A.K. Au and R.R. Prost, J. Enviorn. Sci. Health, 12 (1977) 431. [71] M.H. Koppelman, A.B. Emerson and J.G. Dillard, Clays Clay Miner., 28 (1980) 119. [72] M.H. Koppelman and J.G. Dillard, Clays and Clay Miner., 28 (1980) 221. [73] C. Volzone, A.M. Cesio, R.M. Torres Sanchez and E. Pereira, Clays Clay Miner., 41 (1993) 702. With kind permission of The Clay Minerals Society for reproducing information, Figure 1 and Table 1 of the paper (Figure 5 and Table 1 in this chapter), http://cms.lanl.gov [74] C. Volzone and A.M. Cesio, Mat. Chem. Phys., 48 (1997) 216. With kind permission of Elsevier for reproducing information and part of Table 2 of the paper (Table 9 of this chapter) www.elsevier.com [75] C. Volzone and A.M. Cesio, Anales de la Asociacion Quimica Argentina, 47 (1-2) (1999) 59. With kind permission of the Asociacion Quimica Argentina, AQA, for reproducing information and Figure 1 of the paper (Figure 4 in this chapter), anales(a)inifta.unlp.ar [76] C. Volzone and A.M. Cesio, Mat. Chem. and Phys., 79, 1 (2003) 98. With kind permission of Elsevier for reproducing information and Table 1 of the paper (Table 10 in this chapter), www.elsevier.com [77] C. Volzone, Clays Clay Miner., 43 (1995) 377. With kind permission of The Clay Minerals Society for reproducing information, Figure 5 and part of Table 2 of the paper (Figure 7 and part of Table 6 in this chapter), http://cms.lanl.gov [78] C. Volzone, Aust. J. Soil Res., 36 (1998) 423. With kind permission of CSIRO Publishing for reproducing Table 2 (part of Table 6 in this chapter), http://www.publish.csiro.au/nid/84.htm [79] C. Volzone, in: Catalizadores y Adsorbentes Iberoamericanos para la Remocion de Metales Pesados en Efluentes Industrials, ed. by G. Rodriguez Fuentes y P. Avila Garcia. Ediciones CYTED, Madrid, Espafla. l ra Edition, 2000. [80] website: www.epa.gov/OGWDW/dwh/c-ioc/chromium.html [81] M. Ardon and R.A. Plane, J. Am. Chem. Soc, 81 (1959) 3197. [82] J.A. Laswick and R.A. Plane, J. Am. Chem. Soc, 81 (1959) 3564. [83] R.W. Kolaczkoncki and R.A. Plane, Inorg. Chem., 3 (1964) 322.
320
C. Volzone
[84] M.E. Thompson and R.E. Connick, Inorg. Chem., 20 (1981) 2279. [85] J.E. Finhol, M.E. Thompson, and R.E. Connick, Inorg. Chem., 20 (1981) 4151. [86] H. Stunzi and W. Marty, Inorg. Chem., 22 (1983) 2145. [87] H. Spiccia, W. Marty and R. Giovannelli, Inorg. Chem., 27 (1988) 2660. [88] H. Stunzi, L. Spiccia, F. P. Rotzinger and W. Marty, Inog. Chem., 28 (1989) 66. [89] C. Volzone, J. Ortiga, V. Segarra and E. Montiel, Asociacion Quimica Espanola de la Industria del Cuero (AQEIC), 54, 1(2003) 16. With kind permission of AQEIC for reproducing information and Figures 2 and 3 (Figures 12 and 6 in this chapter), http://www.valles.com/aqeic [90] B.C. Lippens and J.H. de Boer, J. Catal., 4 (1965) 319. [91] C. Volzone, Mat. Chem. Phys. 47 (1997) 13. With kind permission of Elsevier for reproducing part of information of the paper, www.elsevier.com [92] S.J. Gregg and K.S.W. Sing, Adsorption Surface Area and Porosity, Academic Press, 2nd Edition, London, 1991. [93] J.F. Delon, O. Lietard, J.M. Cases and J. Yvon, Clay Miner., 21(1986) 361. [94] C. Volzone and L.B. Garrido, Clay Miner., 36 (2001) 115. [95] R.C. Mackenzie and S. Caillere, in: Data Handbook for Clay Materials and other Nonmetallic Minerals, ed. by H. van Olphen and J.J. Fripiat, Pergamon, Oxford, 1979. [96] A. Drljara, J.R. Anderson, L. Spiccia and T.W. Turney, Inorg. Chem., 31 (1992) 4894. [97] R.H. Loeppert and M.M. Mortland, Clays Clay Miner., 27 (1979) 373. [98] P.F. Kerr, J. Laurence Kulp and P.K. Hamilton, Report No 3. American Petroleum Institute, Project 49, Clay Mineral Standards, Columbia University, New York, 1949. [99] C. Volzone and A.M. Cesio, J. Mat. Sci. Letters, 14 (1995) 658. [100] C. Volzone, VI Reuni6n Anual de SETAC Latinoamerica (Sociedad de Toxicologia y Quimica Ambiental). Buenos Aires, Argentina, 20-23 de Octubre. Abs. (2003) 22. [101] G. Lagaly, Phil. Trans. R. Soc. Lond. A, 311 (1984) 315. [102] C.T. Johnston, in: Organic Pollutants in the Environment. CMS Workshop lectures, ed. Brij L. Sawhney, The Clay Miner. Soc, Boulder, USA, 1996. [103] M.M. Mortland and K.U. Raman, Clays Clay Miner., 16 (1968) 393. [104] A.G. Espantaleon, J.A. Nieto, M. Fernandez and A. Marsal, Appl. Clay Sci., 24 (2003) 105.
CATALYTIC AND ADSORPTION PROPERTIES OF MODIFIED CLAY SURFACES ALEXANDER MORONTA Centra de Superficies y Catalisis, Facultad de Ingenieria, Universidad del Zulia, Maracaibo 4003-A - VENEZUELA. E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
322
A. Moronta
1 - Introduction Clay minerals are the most abundant sedimentary mineral group. They predominate in the colloidal fractions of soils, sediments, rocks and waters and are classified as phyllosilicates (usually hydrous aluminosilicates). In Geology the word clay is used in two ways: firstly as a rock classification which generally implies an earthy, fine-grained material that develops plasticity on mixing with limited amount of water. Secondly, it is used as a particle term, which describes clays as minerals which have a particle size <4um [1], although the modern tendency is to define clays as a naturally occurring material composed primarily of finegrained minerals, which is generally plastic at appropriate water content and will harden with dried or fire [2]. Clays have been one of the most important minerals for man for centuries. The first recorded application of clay as a catalyst, was reported by Bondt et al, [3] who investigated the dehydration of alcohol in 1797. Nowadays, clay minerals science has found a diverse variety of industrial applications such as emulsion stabilizers, detergents, ceramics, medical formulations, bleaching earth, food additives, etc [4]. Clay minerals are efficient adsorbents for the removal of organic pollutants from water when they are modified with organic cations [5-7] because of their hydrophobicity and high surface area and are good materials that catalyze many organic reactions, when acid activated [8-10], either in polar or non-polar media. Another type of heterogeneous catalyst based on clays are pillared clays (PILCs), which porosity, reactivity, selectivity and thermal stability are also widely applied to a variety of organic reactions [11-15]. This chapter is organized to cover 1) Some general aspects of clays' structure, 2) Isomorphous substitution in clay minerals, 3) Cation exchange capacity of clays, 4) Cation migration, 5) Swelling capacity, 6) acidity, 7) Acid activation process, 8) Ion exchange activation, 9) Pillars formation, 10) Clay-organic cation interaction and 11) Acid-activated organoclays. Details of the catalytic and adsorption properties of modified clays will be given each section, and before entering in this matter and in order to understand better the relationship between the clays and their catalytic uses, a brief discussion of the structure is given. 2 - Structure of smectite clay minerals
Figure 1 - Representation of the structure ofmontmorillonite.
Catalytic and Adsorption Properties of Modified Clay Surfaces
323
Smectite clays consist of one octahedral sheet sandwiched by two tetrahedral sheets, resulting in a 2:1 or T:O:T structure (Figure 1). The former includes di- or trivalent metal cations (Al3+, Fe3+, or Mg2+) surrounded by six oxygen or hydroxyl anions in a octahedral structure and in the latter Si4+ cations are surrounded by four oxygen anions in a tetrahedral structure [16-18]. Those layers attract each other by electronic forces. 3 - Isomorphous substitution In smectites clays there can be considerable substitution in either the octahedral or tetrahedral sheet, giving rise to a varying degree of substitution in both sheets (Figure 2) [19].
Figure 2 - Isomorphous substitution in the octahedral sheet (a) and tetrahedral sheets (b). This phenomenon is responsible for some important properties of clays minerals. Substitution in the octahedral sheet usually by Fe3+ and Mg2+ for Al3+, creates a de-localized charge deficiency in the layer [20]. Also, there can be substitution in the tetrahedral sheets of Al for Si, which again creates a more localized charge imbalance. In smectites this effect accounts for approximately 80% of the total cation exchange capacity (CEC). The generated charge deficiency is balanced by exchangeable cations between the unit layers and around the edges. In natural clays Ca, Mg, Na and K are the predominant exchangeable cations [21]. 3.1 - Broken edges Undercoordinated metal ions (Si4+, Al3+, Fe3+) on the broken edges on the surface of the clay can react with water molecules to form surface hydroxyl groups [19] in an attempt to complete their coordination sphere. The overall contribution of these edge sites to the CEC is approximately 20%, and depends strongly on the size and shape of the clay particle. As the particle size decreases, the contribution of broken sites to the reactivity of the clay particle becomes significant. 3.2 - Layer charge Depending upon the degree of isomorphous substitution, clay minerals are classified into three major categories:
324
A. Moronta
a - Neutral lattice structures (kaolinite and serpentine 1:1, talc and pyrophyllite 2:1, chlorite 2:1 + 1). In these structures the 2:1 or 1:1 units of interlinked tetrahedra and octahedra have a net charge of zero. The substitution within the sheets is cancelled electrostatically and the individual layers (2:1, 1:1 and 2:1 + 1) are bound into a crystal by low energy bonds (van der Waals type). The main bonding forces between the layers of kaolinite and halloysite are hydrogen bonds between -OH on one layer and a bridging -O- on the next [22]. b - High-charge mica structures (0.9-1.0 charge) In these minerals there is a charge imbalance due to isomorphous substitution in the basic 2:1 structures, which is compensated by interlayer ions between the layers bonding them together. The charge on the 2:1 layer is near 1. The interlayer ion is firmly held between the adjacent layers and it is an integral part of the structure that is extremely difficult to exchange. c - Low-charge 2:1 structures (0.2-0.9 charge) where there is a net charge imbalance on the tetrahedral-octahedral network of 0.2-0.9, compensated by loosely held ions in the interlayer position which can be easily exchanged in aqueous solution. These minerals swell attracting various types of molecules between the 2:1 layers by the electrostatic charge. The low charge means that the cations are not fixed between the layers. 4 - Cation exchange capacity (cec) To a greater or lesser degree, based on their layer charge, many clays have the ability to adsorb exchange cations from solution. It is this cation storage that makes clays such an important component of many soils. A typical montmorillonite can exchange over 100 millimoles of Mn+ cations per lOOg of clay. The ideal structures of the clay minerals depicted above are often deviated from a number of ways introducing charge imbalances into the structure, mainly isomorphous substitution. As mentioned before, the overall negative charge is balanced by adsorption of metal cations into the interlayer region of the clay mineral [21]. More information is available regarding cation exchange than anion exchange. In clay minerals the common exchangeable cations are calcium, magnesium, hydrogen, potassium, ammonium and sodium, approximately in that order of general relative abundance. The common anions in clay minerals are sulphate, chloride, phosphate, and nitrate. A typical montmorillonite have hydrated Na+, Ca2+ or Mg2+ cations in the interlamellar region. The interlayer cations can be easily exchanged with cations from an aqueous solution [23]. This sorption from solution is normally considered as a simple exchange process. The exchangeability of cations related with the clay by cations of a particular aqueous solution is determined by: The nature of the clay mineral. The nature of the cation, e.g., hydration energy, size, valency. The concentration of the electrolyte and pH of the exchange solution. The population of exchange sites on the clay. Solutions of small size/high valency metal cations are very effective at displacing the interlayer exchangeable cations of a clay [24,25].
Catalytic and Adsorption Properties of Modified Clay Surfaces
325
4.1 - Determination of the CEC The cation exchange capacity of clays is determined experimentally by different analytical methods, most of which are based on the principle of replacement of the original resident exchange ions in an aqueous clay suspension. Among the most well known methods are the ammonium ion exchange and the methylene blue tests [26,27]. Other methods include the use of mechanical extraction [28] and the adsorption of the colored Co(H2O)62+ ion [29]. More recently, a faster method for the determination of CEC has been reported, which involves the use of a Cu-ethylendiamine complex [30]. The CEC represents the isomorphous substitutions in the clay layers if all compensating cations are accessible for exchange. If the CEC and in particular the chemical composition of the clay are known, then it is possible to attribute the substituting ions to the tetrahedral and octahedral sheets. 5 - Cation migration The thermal treatment of a montmorillonite saturated by cations of small radii (<0.7 A) results in a charge reduction and a concomitant fall in the cation exchange capacity. This was first shown for lithium [31] and then for magnesium and aluminum [32,33].
Figure 3 - Total conversion vs aluminum content in charge reduced SAz-1. It is widely considered that, on heating, Li+ cations migrate from the interlayer space into the vacant octahedral [31-34] into the hexagonal holes of the tetrahedral sheet
326
A. Moronta
[36-38] or into both [39,40]. This Li+ fixation reduces the charge imbalance, due to isomorphous substitutions, and decreases the ability of the mineral to adsorb cations. Many studies have used various temperatures (in the range 200-300 °C), different reaction times and clays for the preparation of reduced-charge montmorillonites (RCMs), thus rendering direct comparisons rather difficult. Calvet and Prost [40] were able to achieve 31-92% of total Li+ in a Camp Berteaux montmorillonite heated at 108-220 °C. They stated, using IR, that Li+ cations moved from the interlayer space towards the octahedral vacancy, assuming that Li+ ions lie near the sites of isomorphous substitution. Madejova et al [41] and Komadel et al [42] prepared a series of RCMs starting from a Li-montmorillonite by heating to various temperatures (105210 °C) for 24 h, and studied the effect of acid treatment upon the dissolution of the clays. These authors demonstrated that the extent of acid dissolution decreased with increased amounts of Li fixed within the montmorillonite structure. Additionally, the Li+ content increased with increasing heating temperature. Infrared spectroscopy revealed that Li+ was trapped in the hexagonal cavities of the tetrahedral sheet at all temperatures. The reduced charge montmorillonites perform little catalytic activity as Brensted acid catalysts. However, the replacement of the residual interlamellar Li+ by catalytically active cations, such as Al3+, results in a more active material (Figure 3) [43-45]. 6 - The hydration and swelling capacity of clays The outstanding characteristic of the members of the smectite group is their capacity to absorb water and other polar molecules between the sheets, thus producing a marked expansion of the interlayer spacing. The presence of several layers of water molecules causes the basal spacing to increase from 10.5 A to values in the order of 12.5-20.0 A, depending on the type of clay and exchangeable cation (Figure 4) [46,47]. The swelling properties of smectites means that they are often called expanding clays. Only smectite clays have this particular property of increasing the interlamellar space; other 2:1 clay minerals such as mica and vermiculite do not expand as readily due to the excessively high layer charge.
Figure 4 - Swelling process of smectites in the presence of water vapor. The heating process, at a regular temperature rate, affects the weight loss of clays. These temperatures depend on the type of clay and the rate and time of heating. According to Bradley and Grim [48], below 300 °C, water and other molecules are adsorbed reversibly between the clay laminae. Above 500 °C the water of hydration is removed and the laminae collapse irreversibly [48,49]. After calcination of the clay at 600 °C no water is adsorbed in the interlamellar space.
Catalytic and Adsorption Properties of Modified Clay Surfaces
327
7 - Acidity of clays Clays minerals show both Bransted and Lewis acidity. Air-dried clays have been found to have high acidities when most of their water is removed to leave just one layer of intercalated water between the charged sheets [50]. Additionally, when natural occurring clays are exchanged with highly polarizing species such a M3+ cations, an increase in the acidity is observed due to the hydrolysis of the solvated water molecules (Equation 1) [50-52]. [M(OH2)n]m+ o
[MCOH^iOH]^ 1 ^ + H+
(Eq. 1)
The dissociation of water molecules is the most important source of Bronsted acidity, while Lewis acidity is normally associated with exposed Al3+ or Fe3+ at the broken crystalline edges which can be increased by heating the clay material to >300 °C [53]. However, the heat treatment can lead to the irreversible collapse of the clay layer [54] 7.1 - Determination of the Surface Acidity The first measurement of clay acidity was reported by Walling [55] who found that an acid treated clay imparted color to p-nitrobenzene-azo diphenylamine. The acidity was thought to be due to the residual acid from the activation treatment. Since this study, a variety of methods and techniques have been used to measure the acidity of catalysts and catalyst supports, including the use of indicators [52,56], measurement of the quantity of chemisorbed ammonia at various temperatures [57,58], the determination of the IR spectrum of chemisorbed ammonia [59,60], carbocation formation of arylmethanol [61] and adsorption of pyridine by IR [62-64]. The applications of these techniques have given conflicting and differing results on the same material. Nevertheless none of these methods, except the last one, distinguishes between protonic acidity (Bronsted) and aprotonic acidity (Lewis). This is because the IR spectrum of adsorbed pyridine on acid solids shows a band at 1548 cm"1 associated with the pyridinium ion formed when the molecule reacts with Bransted acid sites and bands in the region 1440-1460 cm"1 attributed to pyridine coordinately bound to Lewis acid sites [62]. The determination of the number of Bransted and Lewis acid sites (in arbitrary units) results by dividing the integrated absorbances in the two regions by the weights of the wafer, and by the molar extinction coefficients, 1.67 (cm ^mol"1) for Bronsted and 2.22 for Lewis acid sites [65]. Unfortunately, the use of all of these methods can be tedious, difficult and time consuming, particularly if the requirement is for a rapid evaluation of the acidity of the material produced. It has been shown that the thermogravimetric analysis of the desorption/decomposition of bases such as cyclohexylamine, butylamine and pyridine can be used for the rapid semi-quantitative determination of the number of acid centers in clays [66-68]. 8 - Acid activation of clays More than 50 years ago, Eugene Houndry [69] found that acid-modified smectites provide gasoline in high yield when used as petroleum-cracking catalysts. These modified clays were used extensively as commercial catalysts until the mid1960's, when they were replaced by more thermally stable and selective zeolite catalysts. However, smectite clays are still used today as commercial catalysts.
328
A. Moronta
Acid-treated clays are also commonly used commercially for decolorizing oils. The effect of acid attack on properties such as surface area and the decolorizing ability have been widely studied [70,71]. Another commercial usage of acid-treated clays is in color formation with leuco dyes in pressure sensitive recording paper [72,73]. The acid treatments increase the surface area (from ca. 40 m2g"' to ca. 500 m2g" ') [74] of clay minerals by disaggregation of clay particles, elimination of several mineral impurities, removal of metal-exchange cations, and proton exchange, reasons for which they are usually known as "acid activation treatments" [75,76]. The enhanced surface area depends significantly on the type of clay used; for example in non-swelling bentonites it is dramatically improved, but the opposite trend is observed in swelling bentonites [77]. In general, the surface area passes through a maximum beyond which further acid treatment reduces surface area [77-79]. The acid activation process is often quite severe and destroys much of the clay layer structure [80] as it removes iron, aluminum and magnesium from the octahedral sheet (Figure 5). The exchangeable cations are replaced mainly by Al3+ and H+-cations [81,82]. The CEC decreases with increasing acid treatment [77].
Figure 5 - Representation of the effect of acid activation. Treatment of montmorillonite clays with cold, dilute acids has little effect on the elemental composition of the host layer and results in an essentially protonexchanged clay, whereas activation with hot concentrated acids results in the removal of ions associated with the octahedral sheet and may not produce an exclusively protonexchanged clay [83-85]. It is now generally accepted that clays with a high octahedral Mg or Fe content leach more readily than those which have a high octahedral Al population [83,86-91]. Acid activation causes little damage to the silicate layer and consequently the structure in the center of the platelet, at the limit of acid attack, remains intact [71,81,83]. The rate of dissolution of the octahedral sheet is a first order process [86,92] which increases not only with increasing concentration of acid,
Catalytic and Adsorption Properties of Modified Clay Surfaces
329
temperature and contact time, but also with increasing Mg content in the octahedral sheet [70,71]. Results obtained after acid-leaching five clays of differing elemental composition showed that the type of clay used had little influence on the catalytic activity which was determined mainly by the extent of octahedral sheet depletion and proton content [83]. However, it is now imperative that the activation conditions are optimized for a given clay, although, it should be emphasized that the acid activation process really works well when selected bentonites are used. Other clays such as kaolin, attapulgite, etc., only improve slightly or not at all when they are treated with acids [77]. 9 - Ion-exchange activation The intercalated cation in naturally occurring clays is Na+ or Ca2+ and occasionally K+. Such clays are very weak catalysts, if at all. When, however, these ions are exchanged by the simple process of immersion in an aqueous solution of the relevant cation they become active. This process can be achieved using (i) active metal cations and (ii) hydroxy-metal polycations, "pillars". 9.1 - Metal cations The resident exchangeable cations, which are strongly hydrated in the presence of water, are replaced by highly polarizing species of small radius such as aluminum, chromium or iron [93,94]. Consequently, the high catalytic activity of Al3+- compared to Fe3+- and Cr3+-exchanged montmorillonites has been attributed to the enhanced polarization of water molecules in the primary co-ordination sphere of the Al3+ cation [95-97]. There are of course numerous cations other than Al3+, Fe3+ and Cr3+ that are also comparable in ionic radius, which also possess a polarizing power. Examples of these are Mg2+, Cu2+, Ni2+, Co2+ and Zn2+ [98-100]. The order of activity of ionexchange clays with these cations was Cu2+ > Zn2+ > Ni2+ > Co2+ ~ Mg2+, for both Brensted and Lewis acid catalyzed reactions [101]. These ion activated clays have been shown to be efficient catalysts for a variety of organic reactions, which include formation of di-alkylethers from alcohols [93], protonation of amines [100], transformation of alkenes [102,106], hydration of ethylene [97], and esterification of organic acids [107], among others. It is worth mentioning that the choice of the reaction environment plays an important role in the catalyzed process of ion-exchanged clays. For example, Al3+-, Cu2+-, Co2+-, Ni2+- and La3+-exchanged Tonsil showed, in general, lower activity for the Diels-Alder reaction between cyclopentadiene and methyl vinyl ketone % conversion) than that of Cr3+- and Fe3+-exchanged Tonsil (92+1% conversion) [108]. Certainly, the acidity of Al3+-exchanged clay is similar to that of Cr3+-exchanged clay and higher than that of M2+-exchanged clays [109]. This observation led Adams et al [108] to postulate that d orbitals of the exchanged transition metal cations are responsible, in this particular reaction, for the catalytic activity of the ion-exchanged clay minerals where the role of Bransted acid sites was discarded. 9.2 - Pillared clays When clays are compared as catalysts to the more rigid cage-like zeolites, they are found to be comparable at lower temperatures but, due to a tendency to dehydrate and undergo layer collapse, clays are usually inferior at higher temperatures (ca. >150 °C). One way to overcome this disadvantage is to incorporate large inorganic cations. Such a process is known as pillaring and, in addition to the great improvements in
330
A. Moronta
available surface area and structural integrity, which is often obtained, these materials have the added advantage that the pillars themselves may be catalytically active [110112]. Pillared clays (PILCs) are prepared by exchanging polycations into the interlayer region of expandable clay minerals (usually montmorillonite and saponite) which, following calcination, are transformed to metal oxide pillars fixed to the layers of the clay to yield a rigid cross-linked material (Figure 6) [113-115]. A variety of inorganic oxides, A12O3 [116], ZrO2 [117], TiO2 [118], Cr2O3 [119], Ga2O3 [120], and mixed-metal oxides Al2O3-Ga2O3 [121], and Al2O3-SiO2 [122,123] have been successfully pillared in smectites to generate high surface area catalysts.
Figure 6 - Schematic representation of the pillarization process. The catalytic properties of pillared clays are the result of the propping apart of the clay structure, which have an increased surface area (from ca. 50 to ca. 350 m2g"') and pore volume (from ca. 0.08 to 0.2-0.3 cm3 g"1). This exposes much of the interlayer region and any acid sites available to reactant molecules. The pillared structures are found to be stable up to 500-700 °C [116,123-125]. In contrast, the potential uses of PILCs demonstrate that the variation of the pillaring procedure can affect the properties of the resulting material [126,127]. One approach to varying the nature of the matrix is by acid activation, which further improves the surface area (up to 400 m2 g"') and gives higher acidity values. Such acidmodified materials can be subsequently pillared and the resulting pillared acid-activated clays (PAACs) possess enhanced chemical and physical properties compared to those of conventional PILCs [128-130]. This assertion has been confirmed by an enhanced catalytic activity of PAACs, which showed to be better catalysts that those derived from acid-treated clays or PILCs, towards the dehydration of butan-1-ol and pentan-1-ol, alkylation of benzene by dodec-1-ene and conversion of cumene [11,130,131], as well as improved adsorption of chlorophyll from edible oils [132]. Furthermore, the surface properties of pillared clays can be altered by incorporation of metal cations via impregnation methods [133]. In this regard, Ni/(La)Al-PILC are potential catalysts for methane reforming with CO2 to produce synthesis gas in a wide range of temperatures [134]. Additionally, alumina-chromia pillared saponites and Al-pillared clays impregnated with cobalt compounds are selective catalysts for the dehydrogenation of ethylbenzene to styrene with minor yield of other products (Figure 7) [135,136].
Catalytic and Adsorption Properties of Modified Clay Surfaces
331
Figure 7 - Product formed during the catalytic transformation of ethylbenzene. 10 - Clay-organic cation interactions The exchange cations can also be replaced with different organic cations by a simple ion exchange. A variety of organic cations may be used in this regard to form organoclays that, unlike natural clay, are effective adsorbents for removing organic contaminants from water [5-7,137-141]. One class of organic cations that have been widely used to synthesize organoclays are quaternary ammonium cations (QUATs) of the general form [142]: [(CH3)3NR]+ or [(CH3)2NRR']+ where R and R' are aromatic or alkyl hydrocarbon species. Substituting such organic cations for native metal cations drastically alters the surface properties of the clay, which change from hydrophilic to organophilic. This occurs because the heat of hydration of the organic cations is very low so that they do not attract water molecules, and because of the substantial amount of organic carbon associated with the clay surface and interlayers. Additionally, the intercalated organic cations act as pillars to prop open the aluminosilicate sheets resulting in greater interlayer spacing that do not change substantially in the presence of water. 10.1 - Nature of the organocation The interaction of quaternary alkylammonium cations (QACs) with clays is affected by the size and structure of the R group, the clay type, solution conditions and the nature of the exchange cation. In general, the interactions between clays and QACs are strong. The synthesis of organoclays using short-chain QACs is relatively straightforward. This is because short-chain length organocations have high solubilities in water and are sorbed on the clay exclusively by cation exchange. For QACs with large hydrophobic moieties [e.g., hexadecyltrimethylammonium, (HDTMA+)], there are two complicating factors (i) they have a very low solubility in water so that solvents
332
A. Moronta
such as methanol may be needed to dissolve the organic modifiers and (ii) both a cation exchange and non-exchange mechanism may contribute to the overall adsorption of such organocations by clays. The non-exchange adsorption of QACs is due to nonelectrostatic interactions between the alkyl moieties (tails) of QACs that have bound to the clay surface by cation exchange and the alkyl tails of QACs that have not undergone ion exchange. Depending on the size of the organic cations and the layer charge of the mineral, the alkyl chain of organic cations may form flat lying monolayer, bilayer, pseudotrimolecular layer, or paraffin complexes [138,143,144]. Cowan and White [145] studied the adsorption of straight-chain monoalkylammonium salts by Na-Bentonites. A linear relationship between the change in the free energy and the number of carbon atoms in the aliphatic chain was found. The increment in the free energy was ascribed to van der Waals interactions. Consequently, a mechanism of the exchange process was proposed, in which the length of the hydrocarbon chain was considered to play the most important role. Similarly, Theng et al [146] reported the replacement of the resident cations (Na+, Ca2+) on montmorillonite by different alkylammonium cations, finding that the affinity of the mineral for the organic molecules was linearly correlated to molecular weight with the exception of the smaller methylammonium and the larger quaternary ammonium ions such as tetra-n-propylammonium and tetra-n-butylammonium. Therefore, the greater the length of the alkylammonium chain, the greater is the contribution of physical, non-coulombic forces to adsorption. Within a group of primary, secondary and tertiary amines, the affinity of the alkylammonium ions for the clay decreases in the series R3NH+ > R2NH2+ > RNH3+. Similar conclusions were reached by Vansant and Peeters [147]. The differences were explained in terms of the size and shape of the ions. In the study carried out by Theng et al [146] Na+ was more easily exchanged by the organic cations than Ca2+. Quaternary alkylammonium ions are preferentially adsorbed on to the cationexchanges sites of montmorillonites. Thus solvents do not significantly displace the organocation from the clay, and the structure is stable in the presence of high concentrations of metal cations [148]. In this regard, Mortland and Barake [149] reported that the order or effectiveness in replacing ethylammonium ion was Al3+ > Ca2+ >Li + 10.2 - Organo-clay complexes Organic chemicals in surface and ground-water supplies have become a major environmental problem. Adsorption by activated carbon is widely used to remove these pollutants from drinking water. Although clays have been recognized as sorbents of such organic compounds; few studies have been conducted to determine whether certain clays could serve as practical sorbents in the treatment of water and waste-water. The literature dealing with interactions between alkylammonium montmorillonites and organic compounds suggests that modified montmorillonites may be potentially effective adsorbents to improve water quality. Interaction mechanisms of clay-organic complexes include ion exchange, co-ordination/ion dipole, hydrogen bonding and van der Waals forces. 10.3 - Ion exchange Organic complexes are intercalated in the clay mineral structure by ion exchange with cations neutralizing, the negative electrical charge responsible for the CEC of the mineral. The organic molecules are positively charged because of the protonation of an amine group as in the case of alkyl amines.
Catalytic and Adsorption Properties of Modified Clay Surfaces
333
In addition to the adsorption of organic compounds by ion exchange, many amines can be protonated at the clay surface [67-69]. The source of protons for this reaction are (i) exchangeable H+ occupying exchange sites, (ii) water associated with the metal cations at exchange sites or (iii) proton transfer from another cationic species already at the clay surface. 10.4 - Ion-dipole and coordination Many polar molecules can be adsorbed on clay minerals. Depending on the nature of the saturating cation, it serves as an adsorption site by ion-dipole or as a coordination type of interaction [150,151]. The greater the affinity that the exchangeable cations have for electrons, the greater will be the energy of interaction with polar groups of organic molecules capable of donating electrons. Hence, transition metal cations on the exchange complex having unfilled d orbitals will interact strongly with electron supplying groups [152]. In the case of molecules such as water and ammonia, the solvation of the exchangeable cation on the clay surface is the most energetic and therefore the primary mechanism of adsorption [153,154]. 10.5 - Hydrogen bonding This is an extremely important bonding process in many clay-organic complexes. While it is less energetic than coulombic interactions, it becomes very significant, particularly in large molecules where additive bonds of this type coupled with a large molecular weight may produce a relatively stable complex. Three classes of hydrogen bonds are distinguished: Water bridge: It involves the linking of a polar organic molecule (e.g., ketone, benzoic acid, etc.) to an exchangeable metal cation through a water molecule in the primary hydration shell [155]. Organic-organic hydrogen bonding: This phenomenon appears when the exchangeable cation on the clay is an organic cation, where the possibility of interaction with another species of organic compound through hydrogen bonding exists [156]. Clay mineral oxygens and hydroxyls: It requires the interaction of molecules capable of hydrogen bonding with oxygens or hydroxyls of the clay mineral surface. 10.6 - Van der Waals forces Van der Waals or physical forces are interactions operating between all atoms, ions or molecules, but are relatively weak. They result from attractions between oscillating dipoles in adjacent molecules. They decrease very rapidly with an increasing distance between the interacting species. They are quite significant in clay-organic complexes, particularly for organic compounds of high molecular weight [146,147]. 10. 7 - Adsorption behavior If a clay mineral has metal cations occupying exchange sites, its surface is hydrophilic because of the water molecules in the hydration shell solvating the cations. Such surfaces are not good adsorbents for removing hydrophobic, poorly water-soluble organic molecules from water. If certain organic cations are placed on the exchange complex by ion exchange, however, the surface becomes hydrophobic and, in turn, organophilic. Organic cations possessing long-chain alkyl groups are particularly able to impart the hydrophobic quality to the mineral surface [157]. Such organoclay
334
A. Moronta
complexes are able to sorb molecules which themselves are hydrophobic by what has been called hydrophobic bonding [5,158] (Figure 8). This process is essentially a nonpolar interaction between the organic phase of the clay organic complex and the hydrophobic organic molecule [158,159]. Numerous reports have been devoted to the study of organoclay complexes since the 1940's. However, only a small amount of data has been assembled regarding the quantitative aspects of the adsorption of organic molecules from aqueous solution by clay minerals. A number of scientists have shown that amines in the ionic form can penetrate between the layers of clay resulting in the production of interlamellar complexes. Cowan and White [160] showed dodecylammonium bentonite, in a series of ethyl- to octadecyl-ammonium bentonite, to be the most active adsorbent for mono- and dihydric phenols. Slabaugh and Carter [161] reported adsorption of 31 milligrams of methanol per gram of dodecylammonium montmorillonite. Stul et al [141] were able to adsorb 9 to 330 milligrams of hexanol and 40 to 530 milligrams of octanol per gram of dodecylammonium montmorillonite in studies conducted over a wide range of alcohol concentrations.
Figure 8 - Schematic representation of the adsorption of benzene over an organoclay. Mortland et al [162] showed that phenol and its chlorinated congeners were sorbed by hexadecyltrimethylammonium (HDTMA+)-smectite and + hexadecylpyridinium (HDPY )-smectite in proportion to the number of chlorine atoms on the phenol structure. Thus phenol itself was not adsorbed significantly by these complexes, but trichlorophenol was strongly sorbed. In other words, as the hydrophobicity of the molecule is increased, the sorption is also increased. Mortland et al [162] and Boyd et al [163] used quaternary alkylammonium cations exchanged on smectites to enhance the uptake of phenol and chlorophenols from water. They showed that in general the adsorptive capacity of modified clays exchanged with large quaternary alkylammonium ions was greatly increased, compared with that of the unmodified clays. Charge density effects were first noted by Lee et al [164] and Jaynes and Boyd [165], who found that low charge tetramethylammonium (TMA)- and Trimethylphenylammonium (TMPA)-montmorillonite generally adsorbed larger quantities of arenes, from aqueous solution, than did high-charge TMA- or TMPAmontmorillonites. A similar increase in aqueous-phase arene adsorption with decreasing
Catalytic and Adsorption Properties of Modified Clay Surfaces
335
surface-charge density, (and hence with an increasing area of uncharged siloxane surface) for a range of TMPA-montmorillonites led Jaynes and Boyd [137] to conclude that arenes adsorb preferentially onto the free uncharged siloxane surface. They further concluded that adsorbed TMPA+ ions are not involved directly in aqueous-phase arene adsorption, but prop open the clay layers so that arenes can penetrate the interlamellar region and adsorb on the uncharged siloxane surface. The adsorption capacity of arenes by high charge smectite was substantially improved when the clay was subjected to Li+ charge reduction prior to exchanging with TMPA+ ions [142]. Indeed, using this procedure, the surface area was enhanced and the organic carbon content in the reduced charge organoclay decreased steadily as the charge reduction process decreased. The increase in the adsorptive behavior was because the reduced charge clays contained less TMPA+, which appeared to have little direct interaction with the adsorbates, and only function to prop open the clay interlayer. 11 - Acid-activated organoclays (aaocs) As stated previously, the catalytic properties of natural clays can be greatly modified using acid treatment, which removes the desired fraction of exchangeable cations and produces materials with enhanced surface area, acidity and porosity making them powerful catalysts for a wide range of polar reactions. Useful modifications are also obtained by exchanging the resident cations by an active metal cation (e.g., Al3+). In this regard, Rhodes and Brown [166] studied the activity of acid-treated and aluminum-exchanged acid-treated clays using the formation of tetrahydropyranyl ether from 3,4-dihydropyran and methanol and the isomerization of a-pinene to camphene and limonene as test reactions (Figure 9). The comparison of the catalytic activities of Al3+-exchanged acid-treated clays with the unexchanged acid-treated clays, revealed that, there was little difference in the activity between the two clays at a treatment time > 2 h for both reactions. However, at short mid treatments times (<15 min for the polar and <1 h for the non-polar reaction), the Al3+-exchange forms were significantly more active. The catalytic activity in the non-polar medium of a-pinene was optimized when the external surface area reached a maximum, which occurs with extensive acid leaching (4 h acid treatment). The catalyst at this stage had an essentially hydrophobic surface that served to attract the non-polar molecules. The maximum activity in the polar media (shown by Al3+-exchanged montmorillonite) was optimized when the acidity and swelling ability of the catalyst was at a maximum because the hydrophilic clay attracts the polar reactants to the surface, where the acid centers reside. The main disadvantage of using extensively acidleached clays is that depletion of the octahedral sheet causes a significant reduction in the number of cation-exchange sites, which is where the protons reside [89,167], with the concomitant loss in catalytic activity. As discussed previously, the hydrophilic aluminosilicate surface of swelling clays can be rendered hydrophobic by displacing the natural occurring inorganic cations with organocations, resulting in expanded layers that creates favorable conditions for non-polar molecules to access the interlamellar spacing. Certainly, acid activation and organocation exchange can be combined to provide protons and expanded layers, which also produce a balance between the hydrophilic and organophilic character on the clay surface and produce effective catalysts that promote reactions in non-polar media.
336
A. Moronta
Figure 9 - Formation of tetrahydropyranyl ether from 3,4-dihydropyran andmethanol (top) and the isomerization of a-pinene (bottom). Breen et al [168] studied the effect of different acid treatments on the organoclays formed from clays of different initial composition and evaluated the resistance of organocation to acid leaching. The organocations used were tetramethylammonium (TMA+) (a), dodecyl-trimethylammonium (DDTMA+) (b) and octadecyltri-methylammonium (ODTMA+) (c). These authors tested the ability of the prepared catalysts using the isomerization of a-pinene to camphene and limonene. In general, the catalytic activity of the non-acid-treated organoclays was very low, and followed the order TMA+ > DDTMA+ > ODTMA+. Similarly, mild acid treatment alone (0.1 M HC1 for 1 h at 25 °C) did not to produce a large amount of camphene and limonene, but severe acid treatment (1 M HC1 for 1 h at 95 °C) did enhance it. Moreover, mild acid treatment on the organoclays resulted in a fourfold increase in the product yield. Additionally, the conversion was further increased as the severity of acid treatment increased. Similarly, a relative increased yield was only observed in organoclays containing the smallest amounts of DDTMA+ and 0DTMA + (0.25 CEC) and treated with mild acid conditions, the former being more active that the latter. The activity of organoclays prepared with small amounts DDMTA+ and ODTMA+ was particularly more appreciable when acid treatment became more severe. In general trend, DDMTA+ and ODTMA+ cations were more resistant than TMA+ to displacement byH + .
Catalytic and Adsorption Properties of Modified Clay Surfaces
337
The shape, charge and orientation of the organocation should be taken into consideration due to these parameters have a remarkable effect on the catalytic conversion of a particular organic molecule. This observation is shown in Figure 10, where the products distribution for the isomerization of oc-pinene is illustrated for aluminum activated organoclays prepared using tetramethylammoiun (TMA)+, the diprotonated form of l,4-diazabiciclo-(2.2.2) octane (DABCOH2)2+ and 1,5diaminopentane (DAPH2)2+. Evidently, the type of organocation utilized showed a significant effect on the total conversion obtained using the isomerisation of cc-pinene over SWy-2 clay. The lowest catalytic activity is registered in SW-A1/DAP, suggesting that the exchange with DAPH22+ cations results in a sterically restricted diffusion of pinene in the interlayer space of the catalyst and or no useful acidity. The activity in SW-A1/DABCO samples is considerably higher than for SWAl/DAP. This enhancement in activity can be ascribed to the greater openness in the interlamellar space for a-pinene to react, except for the internal surface fraction occupied by the immobilised DABCO ions. On the other hand, the vertical orientation of DABCOH22+ produces channels between the alkyldiammonium cations in the interlayer space, allowing access to reactant molecules, whereas in DAPH22+ there is a lack of height that may prevent the access of a-pinene to the interlamellar space [169]. In particular TMA+-exchanged clays, treated at room temperature with only 0.1 M HC1, proved to be effective catalysts for the conversion of a-pinene to camphene and limonene. Total conversions of 60 to 90% were obtained making them effective competitors for zeolites and pillared clays for this isomerization [170]. The enhanced activity was attributed to the spatial separation of the TMA+ ions in the interlamellar spacing, which probably controlled the effective size of the catalytic site [164]. This study indicated that (i) acid-activated organoclays (AAOCs) were more effective when the galleries were not congested with large organocations, hence AAOCs derived from TMA+-exchanged clays were found to be the most effective, (ii) the activity was influenced by the nature of the starting clay and (iii) that preadsorbedTMA+ cations appeared unexpectedly less resistant to subsequent displacement by protons. In a similar work, Breen and Watson [171] studied the influence of acid treatment on organoclays prepared using a polycation (Magnafol 206) in samples derived from a Na-montmorillonite (SWy-2) and a Ca-montmorillonite (SAz-1). The clays were first exchanged with the polycation to satisfy 0.25, 1 and 1.5 times the CEC of both clays and then acid activated using 6 M HC1 at 95 °C for 30, 90 and 180 min. Acid-activated samples, absent of the polycation, were also prepared either at 25 °C or at 95 °C. They found that hot acid treatment increased the total conversion of a-pinene for both SWy-2 (77%) and SAz-1 (65%) more than cold mild acid treatment (43% for SWy-2 and 15% for SAz-1). The highest yields were achieved using a mild acid activation treatment (90 min) in clays with the lowest polycation content (0.25 CEC), total conversions of 90% and 83% for SWy-2 and SAz-1, respectively. At more prolonged treatment time (180 min) the total conversion was catastrophically decreased (especially for SAz-1) because the samples presented a much-reduced polycation loading. Catalysts prepared with intermediate and high polycation loadings were, in general, less effective for the reaction process. The presence of polycation had a more marked influence on the activity of samples derived from SAz-1 increasing the yield from 25% for acid-activated SAz-1, in the absence of polycation, to 50% camphene for acid-activated polycation exchanged SAz-1, while the increase in percentage of camphene formed for SWy-2 was only from 42 to 52%. Breen and Watson [170] concluded that the enhancement in yield for SAz-1
338
A. Moronta
was due to the increased hydrophobicity of the polycation loaded clay, whereas the comparable yield for SWy-2 in the absence and presence of polycation may suggest that this clay disperses well in the non-polar a-pinene.
Figure 10 - Products distribution for catalysts derived from SWy-2 clay. Breen's studies [168,171] demonstrated that the amount of either organocation or polycation plays a key role as well as the severity of acid treatment and the type of clay utilized. Indeed, there is a competitive mechanism between protons and organocations for the clay surface (external and internal) that governs the hydrophobic/hydrophilic character and consequently the catalytic activity. Acid activated organoclays prepared combining TMA+ at 1 CEC of the studied clays and different concentrations of hydrochloric acid demonstrated that the catalytic activity towards the isomerization of ce-pinene is dramatically reduced when the incorporation of the organocation is >30% of the CEC the clay. The gallery surface accessible to reactant molecules in the presence of high TMA+ content is insufficient to allow access to the catalytic acid sites [172]. Similarly, the catalytic activity of aluminum activated clays (AlACs) and aluminum activated organoclays were studied to avoid depletion of the octahedral sheet of a Mg-rich saponite. The samples were prepared combining volumes of 1M TMA+ solutions and 0.1 M Al3+ in different ratios to satisfy the CEC of four clays [173]. In general the catalytic properties over AlACs and AlAOCs was very similar to those observed using H+ and H+/TMA+ [172]. Al+/TMA+-exchanged clays gave lower conversions than their Al+-exchanged counterparts when the Al3+ offered was low (10-40%), but values were similar at an Al3+ content of 50% CEC. Figure 11 shows the product distribution for one of the studied clays (SWy-2). Certainly, at high organocation content there was a reduction in acidity and therefore a decrease in activity was expected in the isomerization of a-pinene, but-1-ene and for the adsorption of hept-1-ene [172-175]. The high activity found was attributed to the ability of clays to keep the layers permanently apart thus facilitating the ingress of reactant molecules.
Catalytic and Adsorption Properties of Modified Clay Surfaces
339
Figure 11 - Products distribution for the catalysts derived from SWy-2 clay. 12 - Conclusions Clay minerals were originally very important catalysts for petroleum-cracking reaction, but they were replaced by more thermally stable zeolites catalysts. Nowadays, pillared clays have produced a cheap and competitive thermal stable solid. Additionally, pillared-organic clays have been found to be good adsorbents for the removal of organic pollutants from water. Useful catalysts are also obtained by modification of the original clay by combination acid activation with orgonocation intercalation for organic reaction. The catalytic activity of these modified materials can be carefully optimized for specific reactions. Changes in the surface are, cation exchange capacity, the nature of the exchange cation, the organic loading and the charge density influence activity in different ways. Acknowledgements The author acknowledges his postgraduate and engineering students for their continuous outstanding experimental work and for improving figures. Thanks are given to FONAC1T and CONDES-LUZ for financial support. Thanks are specially given to professor Jorge Sanchez for helpful discussions and carefully reading this manuscript. I acknowledge the contributions of the authors indicated in the citations. Finally, I would like to thanks the editor for the invitation to write the present chapter and for his forbearance. 13 - References [1] R.E. Grim, Ed., Applied Clay Mineralogy, McGraw Hill, New York, 1962. [2] S. Guggenheim and R.T. Martin, Clays Clay Miner., 43 (1995) 255. [3] N. Bondt, J.R. Deiman, P. van Troostwyk and A Lowrenberg, Ann. Chim. Phys., 21(1797)48. [4] H.H. Murray, Clay Miner., 34 (1999) 39. [5] T.A. Wolfe, T. Demirel and E.R. Baumann, J. Water Pollut. Control Fed., 58 (1986)68.
340
A. Moronta
[6] M.B. McBride, T.J. Pinnavaia and M.M Mortland, Adv. Envoron. Sci. Technol. 8 (1977) 145. [7] G. Sheng, S. Xu and S.A. Boyd, Clays Clay Miner. 45 (1997) 659. [8] H. Kaplan, U.S. Patent, 3,287,422, 4 (1966). [9] P. Komadel, M. Janek, J. Medejova, A. Weekes and C. Breen, J. Chem. Soc. Faraday Trans. 93 (1997) 4207. [10]C.T. O'Connor, L.L. Jacobs and M. Kojima, Appl. Catal. 40 (1988) 277. [11] R. Mokaya and W. Jones, J. Catal. 153 (1995) 76. [12] T.J. Pinnavaia, M.S. Tzou, S.D. Landau and R.H. Raythatha, J. Mol. Catal. 27 (1984) 195. [13] J. Pires, A.P. Carvalho, P.R. Pereira, and M. Brotas de Carvalho, React. Kinet. Catal. Lett. 65(1998)9. [14]P. Salermo, M.B. Asenjo and S. Mendioroz, Themochim. Acta 379 (2001) 101. [15] M. Lenarda, L. Storano, R. Ganzerla and R. Bertoncello, J. Mol. Catal. 144 (1999) 151. [16]U. Hoffmann, K. Endell and D. Wilm, Z. Krist. 86 (1933) 340. [17]C.E. Marshall, Z. Krist. 91 (1935) 433. [18]E. Maegdefrau and U. Hoffman, Z. Krist. 98 (1937) 299. [19]C.T. Johnston, Sorption of Organic Compounds on Clay Minerals: A surface Functional Group Approach, 1-44: Organic Pollutants in the Environment, Ed. B.L. Shawhney, Clay Mineral Society Workshop, Lectures 8, Boulder, USA, 1996. [20]R.W. Grimshaw, Chemistry and Physics of Clays, Benn, London, 1971. [21]G.W. Brindley and G. Brown, Crystal Structure of Clay Minerals and their X-ray Identification. Mineralogical Society, London, 1980. [22] J.P. Quirk, Nature 188 (1960) 253. [23]R.E. Grim and G. Kulbicki, Am. Miner. 46 (1961) 1329. [24]R.E. Grim and R.H. Bray, J. Am. Ceram. Soc. 19 (1936) 307. [25] A.L. Bar and H.J. Tenderloo, Kolloid-Beihefter 44 (1936) 97. [26] R.H.S. Robertson and R.M. Ward, J. Pharm. Pharmcol. 3 (1951) 27. [27] G.W. Brindley and T.D. Thompson, Israel J. Chem. 8 (1970) 409. [28] W.F. Jaynes and J.M. Bigham, Clays Clay Miner. 34 (1986) 93. [29] C.N. Rhodes and D.R. Brown, Clay Miner. 29 (1994) 799. [30]F. Bergaya and M. Vayer, Appl. Clay Sci. 12 (1997) 275. [31] U. Hoffman and R. Klemen, Z. Anorg. Chem. 262 (1950) 95. [32] R. Greene-Keller, Miner. Mag. 30 (1955) 604. [33] R. Greene-Keller, Clay Miner. Bull. 2 (1953) 52. [34] G. Sposito, R. Prost and J.P. Gaultier, Clays Clay Miner. 31 (1983) 9. [35] E. Srasra, F. Bergaya, and J.J. Fripiat, Clays Clay Miner. 42 (1994) 237. [36] R. Tettenhorts, Am. Miner. 47 (1962) 769. [37] V. Luca and C M Cardile, Phys. Chem. Miner. 16 (1988) 98. [38]R. Alvero, M.D. Alba, M.A. Castro and J.M. Trillo, J. Phys. Chem. 98 (1994) 7848. [39] J.D. Russell and V.C. Farmer, Clay Miner. Bull. 5 (1964) 443. [40] R. Calvet and R. Prost, Clays Clay Miner. 19 (1971) 175. [41] J. Madejova, J. Bujdak, W.P. Gates and P. Komadel, Clay Miner. 31 (1996) 233. [42] P. Komadel, J. Bujdak, J. Madejova, V. Sucha and F. Elsass, Clay Miner. 31 (1996) 333. [43] J.H. Purnell, J.M. Thomas, P. Diddams, J.A. Ballantine and W. Jones, Catal. Lett., 2(1989)125. [44] A. Moronta, Ph.D. Thesis, Sheffield Hallam University-UK, 1999.
Catalytic and Adsorption Properties of Modified Clay Surfaces
341
[45] C. Breen and A. Moronta, Clay Miner., 36 (2001) 467. [46]P.F. Low, I. Ravina and J.L. White, Nature, 226 (1970) 445. [47] B. Velde, Composition and Mineralogy of Clays: Origin and Mineralogy of Clays, Clays and the Environment, Ed. B. Velde, Springer, Berlin, 1995. [48] W.F. Bradley and R.E. Grim, Am. Miner., 36 (1951) 182. [49] J.M. Serratosa, Am. Miner., 45 (1960) 1101. [50JM.M. Mortland, and K.V. Raman, Clays Clay Miner., 16 (1968) 393. [51]M. Frenkel, Clays Clay Miner., 22 (1974) 435. [52]H.A. Benesi and H.C. Winquist, Adv. Catal., 27 (1978) 98. [53]B.K.G. Theng, The Chemistry of Clay-Organic Reactions, Ed. A. Hilger, London, 1974. [54] J.L. McAtee and C.B. Concilio, Am. Miner., 44 (1959) 1219. [55] C. Walling, J. Am. Chem. Soc, 72 (1950) 1164. [56]H.A. Benesi, J. Am. Chem. Soc, 78 (1956) 5990. [57]R.T. Barth and E.V. Ballou, Anal. Chem., 33 (1961) 1080. [58] Y. Kubokawa, J. Phys. Chem., 67 (1963) 769. [59] W.A. Pliskin and R.P. Eischens, J. Phys. Chem., 59 (1955) 1156. [60] M.M. Mortland, J.J. Fripiat, J. Chaussidon and J. Uytterhoeven, J. Phys. Chem., 67 (1963)248. [61] J.J. Helsen, Bull. Groupe Fr. Argiles, 22 (1970) 139. [62]E.P. Parry, J. Catal, 2 (1963) 371. [63] J. Connerton, T.W. Joyner and M.B. Padley, J. Chem. Soc. Faraday Trans., 91 (1995) 1841. [64] A.R. Swoboda and G.W. Kunze, Clays Clay Miner., 13 (1964) 277. [65]C.A. Emeis, J. Catal., 141 (1993) 347. [66]J.A. Ballantine, P. Graham, I. Patel, J.H. Purnell, K.J. Williams and J.M Thomas, Proc. Int. Clay Conf. Denver, 1985, Eds. L.G. Schultz, H. van Olphen and F.A. Mumpton, The Clay Mineral Society, Bloomington, Indiana (1987) 311. [67] C. Breen, A.T. Deane and J.J. Flynn, Clay Miner., 22 (1987) 169. [68] C. Breen, Clay Miner., 26 (1991) 473. [69]Ryland, L.B, Tamele, M.W., Wilson, J.N. in Catalysis. (Emmett, P.H, ed.) Reinhold, New York, 7, chap 1, 1960. [70]D.A. Morgan, D.B. Shaw, M.J. Sidebottom, T.C. Soon and R.S. Taylor, J. Am. Oil Soc, 62(1985)292. [71] I. Norvak and M. Gregor, Proc. Int. Clay Conf. Tokio, (1969) 851. [72] R. Fahn and K. Fenderl, Clay Miner., 18 (1983) 447. [73] A. Hirokawa, Nendo Kagaku, 20 (1980) 99. [74] S. Brunauer, P.H. Emmett and E. Teller, J. Am. Chem. Soc, 60 (1938) 309. [75]M.A. Vicente, M. Suarez, J.D. Lopez-Gonzalez and M.A. Banares-Munoz, Langmuir, 12 (1996) 566. [76] S. Mendioroz and J.A. Pajares, Langmuir, 3 (1987) 676. [77]D.R. Taylor and D.B. Jenkins, Soc. Min. Eng. AIME Trans, 282 (1986) 1901. [78] G. Lagaly, Clay Miner, 16 (1981) 1. [79] S.C. Kheok and E.E. Lim, J. Am. Oil. Chem. Soc, 59 (1982) 129. [80]H. Kaviratna and T.J. Pinnavaia, Clays Clay Miner, 42 (1994) 717. [81]G.A. Mills, J. Holmes and E.B. Cornelius, J. Phys. Colloid Chem, 54 (1950) 1170. [82]R.D. Heyding, R. Ironside, A.R. Norris and R.Y. Prysiazniuk, Can. J. Chem, 38, (1960) 1003. [83] C. Breen, F.D. Zahoor, J. Madejova and P. Komadel, J. Phys. Chem, 101 (1997) 5324.
342
A. Moronta
[84]C.L. Thomas, J. Hickey and G. Stecker, Ind. Eng. Chem., 42 (1950) 866. [85] I. Tkac, P. Komadel and D. Mttller, Clay Miner., 29 (1994) 11. [86]B.B. Osthaus, Clays Clay Miner., 4 (1956) 301. [87] I. Norvak and B. Cicel, Clays Clay Miner., 26 (1978) 341. [88] M. Janek and P. Komadel, Geol. Carpath. Ser. Clays, 44 (1993) 59. [89] C. Breen, J. Madejova and P. Komadel, Appl. Clay Sci., 10 (1995) 219. [90] C. Breen, J. Madejova and P. Komadel, J. Mater. Chem., 5 (1995) 469. [91]P. Komadel, J. Madejova, M. Janek, W.P. Gates, R.J. Kirkpatrick and J.W. Stucki, Clays Clay Miner., 44 (1996) 228. [92] W.T. Granquist and G. Gardner Sumner, Clays Clay Miner., 6 (1959) 292. [93] J.A. Ballantine, M. Davies, H. Purnell, M. Rayanakorn and J.M. Thomas, J. Chem. Soc. Chem. Commun., (1981) 427. [94] A. Vaccary, Catal. Today, 41 (1998) 53. [95]M.P. Atkins, D.J.H. Smith and D.J. Westlake, Clay Miner., 18 (1983) 423. [96]D.T.B. Tennakoon, R. Schlogl, T. Rayment, J. Klinowski, W. Jones and J.M. Thomas, Clay Miner., 18 (1983) 357. [97] R. Gregory, D.J.H. Smith and D.J. Westlake, Clay Miner., 18 (1983) 431. [98] J.H. Purnell, J. Williams and L. Yun, Catal. Lett., 10 (1991) 63. [99]R.P. Mitre and P.S. Sindhu, Clays Clay Miner., 19 (1971) 391. [100] C. Breen, Clay Miner., 26 (1991) 487. [101] D.R. Brown and C.N. Rhodes, Catal. Lett., 45 (1997) 35. [102] J.M. Adams, A. Bylina and S.H. Graham, J. Catal., 75 (1982) 190. [103] J.M. Adams, A. Bylina and S.H. Graham, Clay Miner., 16 (1981) 325. [104] E. Choren, A. Moronta, G. Varela, A. Arteaga and J. Sanchez, Clays Clay Miner., 45(1997)213. [105] E. Choren, A. Moronta, A. Arteaga and J. Sanchez, Clays Clay Miner., 45 (1997) 221. [106] A. Moronta, S. Taylor and C. Breen, Clays Clay Miner., 50 (2002) 265. [107] J.A. Ballantine, M. Davies, H. Purnell, M. Rayanakorn, J.M. Thomas and K.J. Williams, J. Chem. Soc. Chem. Commun., (1981) 8. [108] J.M. Adams, S. Dyer, K. Martin, W.A. Matear and R.W. McCabe, J. Chem. Soc. Perkin Trans., (1994)761. [109] J.M. Adams, T.V. Clapp and D.E. Clemet, Clay Miner., 18 (1983) 411. [110] S.L. Suib, J.F. Tanguay and M.F. Occelli, J. Am. Chem. Soc, 108 (1986) 6972. [ I l l ] K. Urabe, H. Sakurai and Y. Izumi, J. Chem. Soc. Chem. Commun., (1986) 1074. [112] S. Yamanaka and G.W. Brindley, Clays Clay Miner., 26 (1978) 21. [113] R.M. Barrer and D.M. Macleod, Trans. Faraday Soc, 51 (1955) 1290. [114] J. Shabtai, N. Frydman and R. Lazar, Proc. 6™ Int. Congr. Catal. Chemical Society, London, B5 (1976) 1. [115] T.J. Pinnavaia, Science, 220 (1983) 365. [116] G.W. Brindley and R.E. Semples, Clay Miner., 12 (1977) 229. [117] S. Yamanaka and G.W. Brindley, Clays Clay Miner., 27 (1979) 119. [118] S. Yamanaka, T. Nishihara and M. Hattori, Mater. Chem. Phys., 17 (1987) 87. [119] T.J. Pinnavaia, M. Tzou and S.D. Landau, J. Am. Chem. Soc, 107 (1985) 4783. [120] A. Bellaoui, D. Plee and P. Meriaudeau, Appl. Catal., 63 (1990) L7. [121] F. Gonzalez, C. Pesquera, C. Blanco, I. Benito and S. Mendioroz, Inor. Chem., 31(1992)727. [122] M.L. Occelli, Proc. Int. Clay Conf., Denver, (1985) 319. [123] J. Sterte and J. Shabtai, Clays Clay Miner., 35 (1987) 429.
Catalytic and Adsorption Properties of Modified Clay Surfaces
343
[124] E. Kikuchi and T. Matsuda, Catal. Today, 2 (1988) 297. [125] P.B. Malla and S. Komarneni, Clays Clay Miner., 41 (1993) 472. [126] K. Takahana, M. Yokoyama, S. Hirao, S. Yamanaka and M. Hattori, J. Mat. Sci., 27 (1992) 1297. [127] K. Ohtsuka, Y. Hayashi and M. Suda, Chem. Mater., 5 (1993) 1823. [128] R. Mokaya and W. Jones, J. Catal., 153 (1995) 76. [129] F. Cooli and W. Jones, J. Mater. Chem., 8 (1998) 2119. [130] J. Bovey and W. Jones, J. Mater. Chem., 5 (1995) 2027. [131] R. Mokaya and W. Jones, J. Chem. Soc. Chem. Commun., 8 (1994) 929. [132] R. Mokaya, W. Jones, M.E. Davies and M.E., Whittle, J. Mater. Chem., 3 (1993) 381. [133] R.T. Yang and W. Lee, J. Catal., 155 (1995) 414. [134] S. Wang, H.Y. Zhu and G.Q. Lu, J. Colloid Interface Sci., 204 (1998) 128. [135] M.A. Vicente, A. Meyer, E. Gonzalez, M.A. Banares-Munoz, L.M. Gandia and A. Gil, Catal. Lett., 78 (2002) 99. [136] E. Gonzalez and A. Moronta, Appl. Catal., (2003) In press. [137] G. Sheng, S. Xu and S.A. Boyd, Wat. Res., 30 (1996) 1483. [138] W.F. Jaynes and S.A. Boyd, Soil Sci. Soc. Am. J., 55 (1991) 43. [139] W.F. Jaynes and S.A. Boyd, S.A. Clays Clay Miner., 39 (1991) 428. [140] M.M. Socias-Viciana, M.C. Hermosin and J. Cornejo, Chemosphere, 37 (1998) 289. [141] M.S. Stul, J.B. Uytterhoevenand J.D. Bock, Clays Clay Miner., 27 (1979) 377. [142] S.A.Boyd and W.F. Jaynes, Role of the Layer Charge in Organic Contaminant Sorption by Organo-Clays, 48-77: Layer Charge Characteristics of 2:2 Silicate Clay Minerals, Ed. A.R. Mermut, Clay Mineral Society Workshop, Lecture 6, Boulder, USA, 1994. [143] G. Lagaly and A. Weiss, Proc. Int. Clay Conf., Tokyo, Ed. L. Heller, Israel Univ. Press, 1969. [144] G. Lagaly, Clays Clay Miner., 30 (1982) 215. [145] C.T. Cowan and D. White, Trans. Faraday Soc, 54 (1958) 691. [146] B.K.G. Theng, D.J. Greenland and J.P. Quirk, Clay Miner., 7 (1967) 1. [147] E.F. Vansant and G. Peeters, Clays Clay Miner., 26 (1978) 279. [148] T.R. Jones, Clay Miner., 18 (1983) 399. [149] M.M. Mortland and N. Barake, Trans. 8Th Int. Congr. Soil Sci., Bucharest, Publising House, Acad. Soc. Rep. Romania 3 (1964) 433. [150] B.L. McNeal, Soil Sci., 97 (1964) 96. [151] K.K. Bissada, W.D. Johns and F.S. Cheng, Clay Miner., 7 (1967) 155. [152] R.H. Dowdy and M.M. Mortland, Soil Sci., 105 (1968) 36. [153] J. Keay and A. Wild, Clay Miner. Bull., 4 (1961) 221. [154] J.W. Kijne, Soil Sci. Soc. Am. Proc, 33 (1969) 539. [155] R.L. Parfitt and M.M. Mortland, Soil Sci. Soc. Am. Proc, 32 (1968) 355. [156] M.M. Mortland, J. Agr. Food Chem., 16 (1968) 706. [157] R.M. Pashley, P.M. McGuiggan and B.W. Ninham, Science, 229 (1985) 1088. [158] M.B. McBride, T.J. Pinnavaia and M.M. Mortland, Adsorption of Aromatic Molecules by Clays in Aqueous Suspensions: Fate of Pollutants in the Air and Water Environments. Part I, Ed. I.H. Suffet, Wiley, New York, 1977. [159] M.M. Mortland, Adv. Agron., 22 (1970) 75. [160] C.T. Cowan and D. White, Clays Clay Miner., 9 (1962) 459. [161] W.H. Slabaugh and L.S. Carter, J. Colloid. Inter. Sci., 27 (1968) 235. [162] M.M. Mortland, S. Shaobai and S.A. Boyd, Clays Clay Miner., 34 (1986) 581.
344
A. Moronta
[163] S.A. Boyd, J.F. Lee and M.M. Mortland, Nature, 333 (1988) 345. [164] J.F. Lee, M.M. Mortland, C.T. Chiou, D.E. Kile and S.A. Boyd, Clays Clay Miner., 38 (1990) 113. [165] W.F. Jaynes and S.A. Boyd, J. Air Waste Mgnt. Assoc, 40 (1990) 1649. [166] C.N. Rhodes and D.R. Brown, Catal. Lett., 24 (1994) 285. [167] A. Corma, A. Mifsud and E. Sanz, Clay Miner., 22 (1987) 225. [168] C. Breen, R. Watson, J. Madejova, P. Komadel and Z. Klapyta, Langmuir, 13 (1997) 6473. [169] A. Moronta, Ciencias, 10 (2002) 48. [170] A. De Stefanis, G. Perez, O. Ursini and A.A.G. Tomlinson, Appl. Catal., 132 (1995)353. [171] C. Breen and R. Watson, Appl. Clay Sci., 12 (1998) 479. [172] C. Breen and A. Moronta, J. Phys. Chem. B, 103 (1999) 5675. [173] C. Breen and A. Moronta, J. Phys. Chem. B, 104 (2000) 2702 [174] A. Moronta, V. Ferrer, J. Quero, J. Sanchez and E. Choren, Appl. Catal. A, 230 (2002) 127. [175] A. Moronta, S. Taylor and C. Breen, Clays Clay Miner., 50 (2002) 265.
PREPARATION OF LAYERED DOUBLE HYDROXIDES EIJI KANEZAKI Department of Chemical Science and Technology, Faculty of Engineering, The University of Tokushima, 2-1 Minamijosanjima, Tokushima 770-8506 - JAPAN E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
346
E. Kanezaki
1 - Preparation of layered double hydroxide with interlayer carbonate. Layered double hydroxides (LDH) have a general formula of [MaM'b(OH)2(a+b)]A3bH2O, abbreviated as MM'/A-LDH where M and M' are a divalent and a trivalent metal cation in the layer of the hydroxide, respectively, and the excess positive charge of the layer is compensated by the negative one on the interlayer monoanion A, which can be replaced by a di-, tri- or other multivalent anion or in some cases a mixture of them, thus LDH is called an anionic clay mineral. LDH is a mimic of naturally occurring hydrotalcite (MgAl/CO3-LDH), has a similar layered structure, which is called hereafter the hydrotalcite-like layered structure, and is also called the hydrotalcite-like compound. Over the last two decades, interest has been growing in the availability for the intercalation of various organic anions having flexible or rigid molecular frameworks into LDH not only from the scientific but also from industrial viewpoints [1]. LDH is usually prepared in ordinary conditions of temperature, pressure and so on as a precipitate from solutions by means of environmentally benign methods. Therefore, it is easy to synthesize the LDH which has the desired anion(s) at the interlayer region when we carefully select the combination of the metals and the organic compound. Both the facility in the synthesis and the potential variety in the combination of the components promise the usefulness of LDHs for developing new-type of materials. A lot of studies have been done on the synthesis and characterization of LDH with the variety of divalent (Zn2+, Ni2+, Fe2+, Co2+, etc.) or trivalent metal cations (Fe3+, Cr3+, Sc3+, etc.) in the layers of LDH. It is outstanding that many kinds of anion, indifferent to mono- or multivalent, are intercalated at the interlayer of LDH because the volume of the interlayer gallery is variable and the interlayer distance between the adjacent layers is enhanced or reduced in proportion to the molecular size of the anion. The interlayer anion of LDH is organic, inorganic or coordination compounds which could be introduced into LDH by means of anion-exchange, coprecipitation or rehydration the last of which is very characteristic of the intercalation of LDH and is described below in detail. The rehydration method originates in a unique nature of LDH; the hydrotalcite-like layered structure is reproduced when the precursor LDH is calcined at high temperature to collapse the layered structure followed by swelling in aqueous solution containing anions to be intercalated. Collapsing the layered structure in the precursor LDH results in only amorphous phases of the mixed metal oxides of MO and M'2O3 in the calcined solid, all of which are absent in the X-ray powder diffraction patterns (XRD). When the intercalation of anions occurs at the interlayer gallery region of LDH, the XRD pattern usually changes drastically and the explicit layered structure appears again; the magnitude of the basal spacing, calculated from the lowest 29 angle of diffraction lines, indicates the molecular size of the intercalated anion perpendicular to the normal axis of the stacking layers. Therefore, in the study of the solid-state chemistry of LDH, the XRD measurement before and after the intercalation of a particular anion is essential with rare exceptions. Furthermore, the inspection of the thermal change of the XRD pattern together with that of differential thermal analysis/thermal gravimetry (DTA/TG) allows us the fruitful discussion since the collapse of the layered structure, which results from the dehydroxylation of the double hydroxide, takes place at a moderate temperature. However, thermally metastable solid phases are sometimes missed in conventional measurements in which samples are heated and cooled outside the sample chambers. Therefore, high temperature in situ measurements are necessary in order to investigate the thermal change of the solids
Preparation of Layered Double Hydroxides
347
precisely. In particular, the in situ high temperature XRD measurement (in situ HTXRD) is favorable for the study of LDH since the layered structure recovers very soon in an ordinary atmosphere owing to the adsorption of water and CO2 in air. Results of these measurements are illustrated below for two LDH compounds, which have the interlayer carbonate and are frequently used as the precursor for the intercalation of the particular
1.1 - Mg and Al layered double hydroxide with interlayer carbonate Figure 1 illustrates a general view of the in situ HTXRD patterns of MgAl/CO3-LDH in the temperature range from 30°C to 1000°C. There are three regions of temperature (T), which have the common HTXRD pattern; 30
348
E. Kanezaki
thickness of one layer of the double hydroxides (4.8 A) and the interlayer distance (3.0 A) [1]. The interlayer distance of this layered structure is determined by the dimension of the interlayer carbonate dianions which are located with their molecular plane parallel to the internal surface of the layer [1,2,4]. On the other hand, the interlayer distance of Phase II is calculated 1.8 A with the same assumption for the thickness of one double hydroxides layer as that of the hydrotalcite-like layered structure. The smaller value of the interlayer distance of Phase II than that of the hydrotalcite-like layered structure suggests strongly that the interlayer species are different from each other between these two solid phases.
Figure 1 - In situ HTXRD diffraction patterns of MgAl/CO3-LDH with the atomic ratio Mg/Al=2 between 30°C-1000°C; temperatures are 1000°C (top), 900°C, 800°C, 700°C, 600°C, 500°C, 400°C, 380°C, 360°C 340°C, 320°C, 300°C, 280°C, 260°C, 240°C, 220°C, 200°C, 180°C, 160°C, 140°C, 120°C, 100°C and 3 0°C (bottom). Reprinted from [40] with permission from Elsevier. Table 1 - Indexing of the in situ HTXRD pattern of Phase II produced by calcination of MgAl/CO3-LDH (Mg/Al=3) at 300°C; hexagonal lattice with ao=3.O7O A co=6.592 A is assumed. Reprinted from [40] with permission by Elsevier.
29/degree 13.40 26.96 33.86 36.40 43.74 60.24 61.90
observed I/Io 100 8 10 20 10 22 10
d/A 6.592 3.304 2.645 2.466 2.068 1.535 1.498
Calculated indexing d/A 6.592 (001) (002) 3.296 (100) 2.659 (101) 2.466 (102) 2.069 1.535 (110) (111) 1.495
Figure 2 gives the time course of the XRD pattern of Phase II, which is observed after the calcination of Mg/Al/CO3-LDH at 300°C in an electric furnace. Though the pattern does not change at one day after the calcination, the intensity of the Phase II diffraction decreases after 8 days and it becomes weak significantly at 56 days
Preparation of Layered Double Hydroxides
349
after the preparation when the calcined sample is placed in an air-tight silica gel desiccator after preparation; (see lines indicated by arrows in Figure 2). In contrast to the decrease of the Phase II diffraction, Phase I appears weakly after 8 days and becomes distinct at 56 days after the preparation. The same XRD pattern in Figure 2 (d) is also observed at three months after the calcination when the sample is still placed in the same desiccator. These results indicate clearly that Phase II is metastable and goes to an XRD silent amorphous phase slowly and that a small part of the amorphous phase goes Phase I in a dry condition. When the Phase II sample powder is placed in a humid atmosphere, the change is accelerated and Phase I is predominant in the XRD pattern after a few days. These results indicate that two processes are present in the change from (a) to (d) in Figure 2; first, the metastable Phase II degrades to an amorphous phase in a dry condition. Second, a small quantity of Phase I is formed from this amorphous phase owing to the contact to trace quantity of H2O (and probably of CO2) in a desiccator. Keeping the unchanged XRD pattern in Figure 2 (d) for three months after the sample preparation above is elucidated by the lack of enough H2O (and probably of CO2), which are necessary to return to Phase I, in the desiccator. A direct phase transition from Phase II to Phase I is doubtful. Remembering that the total structure of the layer is determined by the chemical status of the magnesium ions in the layer, it is concluded that the brucite-like layered structure having the octahedral unit [Mg(OH)6] is still maintained in Phase II since MgO is not observed in the temperature range where Phase II is alive (180
Figure 2 -X-ray diffraction pattern of Phase II; (a) immediately, (b) 1 day, (c) 8 days and (d) 56 days after sample preparation, positions of the most intense line of Phase I and that of Phase II in Figure 1 are indicated by arrows. Reprinted from [39] with permission by Kluwer Academic/Plenum Publishers. Differential thermal MgAl/CO3-LDH are illustrated in observed at 196°C (endotherm A) throughout the temperature range
analysis/thermal gravimetry (DTA/TG) of Figure 3. In the DTA curve, two endotherms are and at 379°C (endotherm B). The total mass loss agrees well with the summation of all volatile
350
E. Kanezaki
components (H2O and CO2) in the chemical formula. It is noted that temperatures of these two endotherms correspond to those of formation and degradation of Phase II; Phase II is formed at the temperature of endotherm A and is degraded at the temperature of endotherm B. Since the indexing of the pattern of Phase II in Table 1 suggests strongly that the layered structure is still maintained in this solid, it is reasonable to as-sume that the mass loss at endotherm A in the TG curve is due to the elimination of interlayer carbonate and water in the solid sample. Interlayer carbonate ((CO32-)inter.) in MgAl/CO3-LDH is thermally oxidized by a nearby water molecule and produces both volatile CO2 and interlayer hydroxyl anion ((OH)inter.) as in eq. (1) [5]. (CO32-),m=, + H2O -> 2(OH-)mter+ CO2 T
(Eq. 1)
The calculated mass loss 16.4% at endotherm A by using the chemical formula is close to the observed one 17% at the endotherm A in the TG curve of Figure 3 thus supporting the elimination of species in eq. (1). Namely the hydroxyl anions are produced at the interlayer region of LDH when Phase II is formed at the temperature of endotherm A. Less bulky nature of this anion than a carbonate explains the reduction of the interlayer distance in XRD patterns from 3.0 A (hydrotalcite-like layered structure) to 1.8 A (Phase II). This reduction may be related to the migration of Al3+ into the interlayer, which has been proposed recently [6]. In the rehydration method, the LDH with interlayer carbonate is used as the precursor and is calcined at the temperature which is very important because it is fatal to the successful reconstruction of the hydrotalcite-like layered structure in the rehydration method. In the calcination of the precursor LDH, the temperature should be higher than that of the layer collapse but at the same time lower than that of the the spinel phase formation because this solid phase is stable and is hard to be converted to LDH in water. Namely, for MgAl/CO3-LDH above, the temperature is usually set between 400°C and 700°C since the former temperature corresponds to the collapse of the layer of the double hydroxides and the latter to the beginning of the spinel formation. In X-ray photoelectron (XP) spectra of Mg/Al/CO3-LDH (not shown), several photoelectron peaks appear, which originate in core level electrons emitted from Mg, Al, C and O in the solid compound. Since Al3+ ions occupy the octahedral Mg2+ sites isomorphously, the coordination number of Al3+ in the layer is six (0 h Al). This is verified in the 27A1 MAS NMR spectra obtained by a Brucker Avance300 solid state NMR spectrometer of the University of Tokushima with the reference of A1(NO3)3 aqueous solution (0.3 moldm-3; 8= -0.1 ppm) after the data accumulation of 64-128 times. Sample powder was put in a capped ZrO-rotor (7 mm or 4 mm in diameter) spinning at the frequency in the range 4.0-13 kHz in order to avoid spinning side bands. Figure 4 illustrates 27A1 MAS NMR spectra of Mg/Al/CO3-LDH at room temperature for two spinning frequencies in order to eliminate the spinning side bands. Indifferent to the spinning frequency, only one absorption maximum corresponding to the change in the angular momentum of the nuclear spin in 27A1 Arn,=l is observed at 5=8.39 ppm with good S/N ratio. It has been reported that the chemical shift of this nucleus is 8.1 ppm for Mg/Al/(CO3+NO3)-LDH [7], 9.2 ppm for Mg/Al/B4O5(OH)4-LDH [8], 14.5 ppm for Zn/Al/CO3-LDH [9] and the same value for Cd/Al/NO3-LDH [10]. All of these values including the chemical shift in this study fall in the range of the aluminum nucleus having the octahedral hexa-coordination. It has been also reported that the absorption due to the
Preparation of Layered Double Hydroxides
351
nucleus having the tetrahedral four-coordination (TA Al) is weakly observed at about 5=80 ppm on elevating the sample temperature in situ at 100°C in 27A1 MAS NMR and that the intensity of the absorption increases (but does not predominate) with the increase of the temperature up to 400°C [7]. Caution should be paid on eliminating spinning side bands by means of changing the spinning frequency.
Figure 3 - DTA/TG curves of MgAl/CO3-LDH (Mg/Al=3). Reprinted from [39] with permission by Klnwer Academic/Plenum Publishers.
Figure 4 - Al MAS NMR spectra of MgAl/COrLDH (Mg/Al=2.6) at spinning frequency 4kHz (A) and 7kHz (B), maxima with an asterisk are spinning sidebands. [49]
352
E. Kanezaki
Figure 5 - High temperature in situ FT-1R spectra of MgAVCOi-LDH at room temperature (A), at 210°C (B), at 400°C (C) and at 500°C (D). [49]
Preparation of Layered Double Hydroxides
353
Figure 5 (A) shows the FT-IR spectrum of MgAl/CO3-LDH at room temperature, which has been reported by many authors [11-15]. A broad band which is located at 3500-3600 cm"1 is the O-H stretching vibration both of the hydroxide group in the layer and of the interlayer water molecules. The band located at 1652 cm"1 is the O-H bending of the interlayer water. Figures 5 (B)-(D) are FT-IR spectra of this compound at 210°C, at 400°C and at 500°C, respectively. On elevating the temperature, the transmittance of the sample disk becomes low at the high energy side in the spectra because the probe light is scattered increasingly owing to the proceeding of the microcrystallization of KBr. At 210°C (Figure 5(B)), two major differences are observed in comparison with the spectrum at room temperature; first, the band due to the O-H bending of the interlayer water disappears, which band is originally observed at 1652 cm"1 at room temperature in Figure 5(A). Since this band is not observed above this temperature, it is concluded that the interlayer water is lost up to 210°C. In other words, the interlayer water eliminates from the sample exhaustively at endotherm A in Figure 3. Comparing the water content in this solid sample with the mass loss in Figure 3, the mass loss at endotherm A has the other origin(s) than the interlayer water. Secondly, the intensity at around 1500 cm"1 increases because the v3 vibration of CO32" is doubly split; one component is located at 1414 cm"1 and the other one at 1505 cm"1 the latter of which components is overlapped by the weak band due to the water in air observed in the spectrum at room temperature. The splitting v3 vibration of the carbonate has been reported as the result of the coordination of oxygen atoms in the carbonate to a metal cation [16]. This is the case which is observed in Figure 5(B); the oxygen atom of carbonate makes chemical bond directly to a metal atom in the layer. The carbonate still maintains the trigonal symmetry at 210°C since the totally symmetric V! vibration, which is forbidden in the D3h symmetry [17], is not observed in Figure 5(B). Since the metastable Phase II is alive between 180°C and 380°C in this solid (Mg/Al=2.6), as is described in the result of XRD, it is concluded that the spectrum in Figure 5(B) is of Phase II. Constantino and Pinnavaia have reported the thermal shrinkage (1.85 A) of the basal spacing at 250°C concluding that the interlayer sulfonate changes thermally to the "grafting sulfonate" at the interlayer region on elevating the temperature [18]. Since the shrinkage (1.29 A) in this study from Phase I to Phase II is in the same order of magnitude it is inevitable to assume that the interlayer carbonate is grafting in the non-aqueous circumstance of the interlayer gallery region at 210°C in this study. This conclusion seems reasonable because it is difficult to suppose that the anion would be located solely at the interlayer gallery region at 210°C without cations. The grafting carbonate would be unstable since it has three equivalent C-O bonds directed to metal cations in the adjacent layers and at the same time it holds the trigonal symmetry. This unstable nature may explains the metastable profile of Phase II in the HTXRD pattern in Figure 1. The FT-IR spectrum at 400°C in Figure 5(C) shows four major differences; first, the intensity of the O-H stretching at 3500-3600 cm"1 decreases, secondly, the sharp band at 1645 cm"1 appears, thirdly, the intensity of the band at around 1500 cm"1 decreases, and in fourth, the band at 1083 cm"1 newly appears. The decrease in the intensity of the O-H stretching is elucidated by the thermal event in which the layered structure of Phase II collapses at 380°C owing to the thermal dehydroxylation of the double hydroxide. The grafting carbonate, which is observed at 210°C, decomposes on elevating the temperature up to 400°C and a part of the carbonate is oxidized to CO2. The band at 1645 cm"1 is
354
E. Kanezaki
assigned to the C-O stretching vibration (v3) of the coordinating CO2 molecule to the metal cation [16]. It is received that this vibration is observed at 2349 cm"1 for the free CO2 molecule and shifts drastically to the lower energy side as 1600-1800 cm"1 on the coordination to metal. Though the V] vibration of the coordinating CO2 locates in the region 1100-1300 cm"1 [16], it is hidden behind the steep increase of the absorption at 1414 cm"1 in Figure 5(C). It has been reported that the brisk CO2 evolution is observed at 300-400°C [10], which CO2 molecule presumably participates in the coordinating CO2. The formation of the coordinating CO2 is important because this species is expected as the useful Cl source in the catalytic reaction and the formation of the coordination bond to metals could open the possibility to go further reaction. Both of the latter two in the four differences described before in the spectrum at 400°C (Figure 5(C)) are due to another carbon species which is also produced after the grafting carbonate decomposes thermally. The band at 1083 cm"1 has been observed in the IR spectrum of CaCO3 (aragonite) as the v, vibration of the carbonate [16]. It has been reported that the Vi vibration of the carbonate is allowed in the trans-Cs symmetry of the O-O-C=O structure in CaCO3 (aragonite) although it is for-bidden in the planar trigonal structure of the dianion as mentioned before [17]. The band at 1410 cm"1 is the v3 vibration of the carbonate, which has the aragonite-like symmetry. Although it has been reported that the v3 band of the carbonate is doubly split in the trans-Cs symmetry [ 16,17], the counterpart may be hidden behind the band cluster around 1500 cm"1 in Figure 5(C). The sharp band located at 856 cm"1 and a shoulder at 878 cm"1 are of the doubly split v2 vibration of the carbonate and the weak band at 707 cm"1 and the band at 616 cm"1 are of the doubly split v4 vibration of the dianion. Thus, there are two carbon species at 400°C; the CO2 complex and the aragonite-like carbonate. In the IR spectrum at 500°C (Figure 5(D)), the intensity of all bands decreases. A broad band is still observed at 3500-3600 cm"1 which is probably due to the OH stretching of amorphous aluminum hydroxide in the form of A1(OH)3 or AIO(OH). Three bands which are located at 1470 cm"1 (broad), 852 cm"1 and at 671 cm"1 are the v3, v2 and the v4 vibration bands of the carbonate, respectively. The CO2 complex and the aragonite-like carbonate both of which are observed at 400°C decompose on elevating the temperature to 500°C; another carbonate having the CaCO3 (calcite)-like structure with the forbidden V] vibration of the carbonate appears in the IR spectrum in Figure 5(D) [16]. Since aragonite is metastable in an ordinary atmosphere and thermally transforms to calcite, the assignment above is reasonable. A sharp band located at 882 cm"1 is assigned [16] to the Mg-0 vibration band in MgO which is produced after the collapse of Phase II. Since this solid is also observed in HTXRD patterns above 400°C in Figure 1, it is likely that this Mg-O vibration overlaps the band at 878 cm"1 in the spectrum at 400°C (Figure 5(C)). The location of the v3 vibration of the carbonate at 500°C (1470 cm"1) shifts to the higher energy side than that at room temperature (1394 cm"1). In spite of the uncertainty of the band location due to the broadness in Figure 5(D), this shift is significant. In other words, the interlayer carbonate has smaller force constant in this vibration mode at room temperature than the calcite-like carbonate does at 500°C, probably owing to the presence of the hydrogen bonds in the interlayer gallery region at the room temperature as previously reported [13]. Both the carbonate and the hydroxylate which are left in the solid sample at 500°C would eliminate from the sample as CO2 and H2O, respectively, on further increase of the temperature although we cannot follow the spectral change above 500°C from the instrumental reason. The elimination of these species would elucidate the
Preparation of Layered Double Hydroxides
355
gentle mass loss observed in the TG curve above this temperature in Figure 3. In IR spectra of the other two solids with different atomic ratios Mg/Al in the layer, the parallel discussion is done. In conclusion, thermal change in the layered structure of MgAl/CO3-LDH associated with elevating temperature below 1000°C is understood by means of in situ HTXRD, DTA/TG analysis and the high temperature in situ FT-IR measurements in which the thermal change in the chemical status of interlayer species is studied in detail. These results indicate that the thermal shrinkage of the basal spacing in Phase II is due to the thermal oxidation of the anion mainly and the residual carbonate makes direct bonding to metal atoms in the layers. 1.2 - Zn and Al layered double hydroxide with interlayer carbonate When the element of divalent metal cation is changed, the different results are observed in these measurements. Figure 6 shows a general view of in situ HTXRD of ZnAl/CO3-LDH on increasing the temperature from 30°C and 1000°C. The hydrotalcite-like layered structure is observed up to the temperature of 160°C with the intense diffraction line at 26=11.3 A (d=7.8 A) associated with some discrete lines, which is very similar to that of MgAl/CO3-LDH in the temperature between 30°C and 180°C of Figure 1. A metastable phase appears at 180°C with diffraction lines slightly shifted to the higher 29 angle thus giving a smaller basal spacing d=7.4 A (26=11.9 A). The reduced basal spacing has been also reported at 120°C in the HTXRD measurement of ZnAl/Cl-LDH whereas the reduced value recovers on further increasing the sample temperature followed by the collapse of the layer structure [18]. The reason for appearing a metastable phase having smaller basal spacing is assumed to be the same as that of MgAl/CO3-LDH described before. It is common to the three LDHs above that the basal spacing is reduced at lower temperature than that of the layer collapse. This metastable phase disappears promptly at 200°C which is lower than the temperature that Phase II disappears in Figure 1, suggesting the thermal instability of the metastable phase of ZnAl/CO3-LDH. The layered structure collapses and the dehydroxylation occurs at this temperature producing metal oxides, finally a spinel phase appears above 800°C. The temperature at which the layered structure collapses differs depending on the divalent metal species in the layers; 200°C for Zn and 379°C for Mg. DTA/TG thermal analysis in Figure 7 also illustrates the thermal change of the hydrotalcite-like layered structure of ZnAl/CO3-LDH. A sharp endotherm at 180°C appears with a large mass loss. In the DTA curve of ZnAl/Cl-LDH, however, it has been reported that the intense endotherm splits in two; one is assigned to the elimination of interlayer water and the other to that of lattice one due to the layer collapse [19]. The endotherm in Figure 7, however, includes the elimination of the interlayer water/carbonate and dehydroxylation of the layers of the double hydroxide which induces the collapse of the layered structure. The endotherm at 180°C in Figure 7 corresponds to the metastable phase also observed at 180°C in Figure 6 and this phase promptly disappears on increasing the temperature. In Figure 8, three absorption bands are observed in 27A1 MAS NMR spectrum of the calcined ZnAl/CO3-LDH at 500°C. A sharp band located at 5=8.890 ppm, a shoulder at 5=50.015 ppm and a broad one at 5=72.289 ppm in this figure agrees with those bands which have been assigned to the 27A1 nuclei in OhAl, in the penta-coordinated Al [21,23] and in the tetra-coordinated one (r d Al) [20-24], respectively.
356
E. Kanezaki
Figure 6 - In situ HTXRRD patterns ofZnA/CC>3-LDH on increasing the temperature; at room temperature, 100°C, 120°C, 140°C, 160°C, 180°C, 200°C, 220°C, 240°C, 260°C, 280°C, 300°C, 320°C, 340°C, 360°C, 380°C, 400°C, 500°C, 600°C, 700°C, 800°C, 900°C and 1000°C (from bottom to top). [50]
Figure 7 - DTA/TG thermal analysis ofZnAl/CO3-LDH. [49] Both bands of the non-octahedrally coordinated Al3+ in this figure disappear in the NMR spectrum of the rehydrated LDH; all of the Al3+ ions uniformly locate in the octahedral coordination sphere after the intercalation of anions in the rehydration process. Although this uniformity has been also reported previously for some intercalated anions [20,21], no definite explanation has been made; abundant resources of donating atoms in
Preparation of Layered Double Hydroxides
357
aqueous solution containing anions likely cover the deficiency in the donating atom of non-octahedral Al3+ upon the rehydration.
Figure 8 - 27Al MAS NMR spectrum ofZnAl/C03-LDH calcined at 5OO°C measured with the spinning frequency of 7 kHz; chemical shifts of three maxima are indicated, peaks v/ith asterisks are spinning side bands. Reprinted from [47] with permission by Kluwer Academic/Plenum Publishers. 2 - Intercalation of organic anions to ldh 2.1 - Intercalation by rehydration method It has been reported that aliphatic [4] and aromatic [2, 25-30] molecular dianions are intercalated in hydrotalcite-like layered compounds. When these large organic ions are intercalated, an increase of the basal spacing is usually observed in powder XRD patterns as described before. Furthermore, the magnitude of the interlayer distance depends on the orientation of the interlayer organic molecules since organic molecules generally have spatial anisotropy in the dimension of molecules and since the dimension, which is measured along the direction perpendicular to the layer surface, reflects on the magnitude of the interlayer distance of the intercalated products. A careful examination of XRD patterns of these LDH gives possible orientation of the interlayer molecules and thereby insight into the interaction(s) between the molecules and other component(s) in the products as well as understanding the properties of the products for potential use. In this section, intercalation of naphthalenedisulfonates by the rehydration method is described, in which the calcined ZnAl/CO3-LDH is used as the precursor. Three isomers of naphthalenedisulfonates (naphthalene-2,6-, -1,5- and -2,7-disulfonates; abbreviated as N26DS, N15DS and N27DS, respectively, and collectively abbreviated as NijDS) are intercalated in the interlayer region of Zn and Al layered double-hydroxides. Figure 9 exhibits two XRD patterns of ZnAl/N26DS-LDH; the basal diffraction appears at 29=5.74 A (d=15,38 A) and is associated with some prominent (00/)-type diffractions in Figure 9A (the 15 A phase). It has been reported that the unit cell of ZnAl/CO3-LDH has hexagonal symmetry with a lattice constant ao=3.06 A and with the c axis perpendicular to the layers [2,13]. The XRD patterns of ZnAl/NijDS-LDH were thus analyzed on the assumption that the unit cell symmetry and the ao dimension was the same as those of ZnAl/CO3-LDH but that the dimension of the other lattice constant c0
358
E. Kanezaki
differs among the products. Indexing of the pattern in Figure 9 indicates that only one XRD-active phase exists and that the predominant crystal growth occurs in the direction parallel to the c axis in this solid sample, which points out well developed stacking of the layers of Zn and Al double hydroxide along this axis [1,2]. On the other hand, the basal diffraction shifts to a slightly smaller diffraction angle of 26=5.18 A (d=17.05 A) in Figure 9B (the 17A phase).
Figure 9 - XRD patterns ofZnAl/N26DSLDHs; (A) the 15Aphase with diffraction lines at 26=5.740 A (001), 11.58 A (002), 17.44 A (003) 23.36 A (004), 29.38 A (005), 34.54 A (006), 41.50 A (007), 46.74 A (008), 52.94 A (009), 60.20 A (00 10) and 61.50 A (110): (B) thel 7A phase with lines at 20=5.18 A (001), 10.50 A (002), 15.72 A (003), 21.12 A (004), 26.42 A (005), 37.32 A (007), 53.12 A (00 10) and 60.22 A (110); lines with asterisks are those of the host ZnAl/CO3-LDH without interlayer aromatic molecules. Reprinted from [48] with permission by Kluwer Academic/Plenum Publishers. The 17A phase of ZnAl/N26DS-LDH was obtained when the amount of the organic salt was increased relatively to that of aluminum in the calcined powder; the ratio of (organic salt)/Al=5 and 1 for the 17Aandthe 15A phases, respectively. In addition to the intense (001) diffraction of the 17A phase associated with some (00/)-type lines, another sequence of diffraction lines (marked with asterisks in Figure 9B) is present. Since the additional sequence agrees with the XRD pattern of the precursor, it is concluded that the synthetic condition for the 17A phase of the ZnAl/N26DS-LDH also favors the formation of the carbonate-intercalated product in which interlayer carbonate
Preparation of Layered Double Hydroxides
359
are originally solvated species in the reaction mixture. The XRD patterns in Figure 10A and B are of ZnAl/N15DS-LDH; the 15A phase with the basal diffraction at 20=5.82 A (d=15.17 A) and the 17A phase with the diffraction at 29=5.24 A (d=l 6.85 A), respectively. Both phases in this figure exhibit well developed stacking of the layers, though the 17A phase in Figure 10B appears together with the carbonate-intercalated product which is also observed in Figure 9B. Only the 15A phase in Figure 10A is formed after leaving the precipitate in the reaction mixture for 40 days, therefore, it is suggested that this phase is thermodynamically more stable than the 17A phase. The 15A phase has been observed within several kinds of LDH by some authors [25,30]. In contrast, only the 17A phase is observed in the XRD pattern (not shown) of ZnAl/N27DS-LDH, in which the basal diffraction appears at 26=5.34 A (d= 16.54 A). It should be noted that the magnitude of the interlayer distances of the 17A phases of all the NijDS-intercalated products in Table 2 is in good agreement with the estimated molecular size of the individual guest organic anions using MO calculations. This result is significant and will be discussed later.
Figure 10 - XRD patterns ofZnAl/N15DS-LDH: (A) the ISA phase with diffraction lines at 29=5.820° (001), 11.68° (002), 17.58° (003) 23.48° (004), 29.24° (005), 35.60° (006), 60.20° (110) and 61.72° (0010); (B) the 17A phase with lines at 29=5.24° (001), 10.54° (002), 15.80° (003), 21.18° (004), 26.54° (005), 37.36° (007), 56.48° (00 10) and 60.26° (110); lines with asterisks are the same as those in Figure 9. Reprinted from [48] with permission by Kluwer Academic/Plenum Publishers.
360
E. Kanezaki
The 17A phase of ZnAl/N26DS-LDH has two maxima located at Xmax= 277 nm and at Xmax=314 nm in the diffuse reflectance (DR) spectrum (not shown), which are assigned as a 1B2u(7t, TI*) and a 'B3u(7i, n*) transitions, respectively in the naphthalene moiety [14]. This profile in the DR spectra is commonly observed in all the spectra of ZnAl/NijDS-LDH in this study. Therefore, it strongly suggests that the interlayer organic anions retain the planar molecular skeleton of the naphthalene moiety with the stable conjugated p-electron framework. A wide scan XPS spectrum of the 15A phase of ZnAl/N26DS-LDH shows several peaks which are assigned to core level electrons emitted from Zn, Al, C, S and O atoms of the solid samples [32]. Table 2 - Summary of results and synthetic conditions of ZnAl/NijDS-LDH. Reprinted from [48] with permission by Kluwer Academic/Plenum Publishers. Ij
/M'VA
/L'VA
NijDS/OV'
Zn:Al:NijDS4)
MM"5)
notes
26
12.66 12.05
0.67 0.56 0.56 0.39 0.26
4:2:5 4:2:1 4:2:1 4:2:1 4:2:1
ZnAl
15
12.3 10.6 12.1 10.4 10.3 10.3 10.4 10.8 11.7
17Aphase[43] 15Aphase[26] 17Aphase[26] 15A phase [43] 15Aphase[25] 15Aphase[30] 15Aphase[30] 15Aphase[30] 17Aphase[26]
27
12.08
ZnAl MgAl ZnAl ZnCr CaAl ZnAl
1) Molecular size: twice the anionic radius of oxygen (2.8 A) plus the interatomic distance between two anionic groups of an aromatic dianion whose geometry is optimized in MO calculation. 2) Interlayer distance: the interplanar spacing d(001) obtained in XRD patterns minus the thickness of a layer (4.8A). 3) Molar ratio of intercalating NijDS and CO3 ' based on the chemical analysis of intercalated products with the composition of the layer [ZnAl050(OH)iM] and with compositions of the interlayer are (C03)o,is(N26DS)0 w, (C03)oj6(N26DS)oo9, (C03)0j6(N15DS)o.o9, (C03)OJS(Nl5DS)oo7for the 17A phase and for the 15A phase of ZnAl/N26DS-LDH and ofZnAl/N15DS-LDH, respectively, and (C03)02o(N27DS)o.osfor the 17A phase ofZnAl/N27DS-LDH; water and chloride are omitted. 4) Molar/atomic ratio in preparation. 5) Divalent (M) and trivalent (M1) metal ions within layers. In the precursor LDH with interlayer carbonate, the binding energy of the Al 2p electron, Eb(Al 2p), is 74.2 0.6 eV (1 eV=1.602xl0"19 J) which agrees with that of the octahedrally coordinated Al3+ ions previously reported [23]. No significant shift of Eb(Al 2p) is observed after intercalation of NijDS within experimental error; Eb(Al 2p) of the 15A and of the 17A phases are 74.9 0.6 and 75.0 0.6 eV for ZnAl/N26DS-LDH, 74.6 7 and 6 eV for ZnAl/N15DS-LDH, respectively, and 6 eV for the ZnAl/N27DS-LDH. Therefore, no dependence of Eb(Al 2p) on the magnitude of the basal spacing results in this study. Since the binding energy is a good measure of the chemical status of metals [32], it is concluded that the aluminum ions in the layers still exist as
Preparation of Layered Double Hydroxides
361
trivalent cations and are coordinated octahedrally in all phases of ZnAl/NijDS-LDH. In other words, the two values for the basal spacings of ZnAl/N26DS-LDH and ZnAl/N15DS-LDH do not originate from any change in the chemical status of the Al3+ ion in the layer. A typical DTA/TG thermal analysis curve of ZnAl/N27DS-LDH (Figure 11) shows two prominent endotherms (at 195°C and at 283°C) and two exotherms (at 520°C and at 565°C) in the DTA curve. This profile agrees well with that of the 15A phase of ZnAl/N26DS-LDH [26]; two endotherms at 185°C and at 308°C were both assigned to the elimination of water and CO2, and exotherms at 487°C and at 565°C to thermal decomposition of N26DS at the outer surface and in the interlayer region, respectively. The result of the previous study leads us to the parallel conclusion that the two endotherms are assigned to the elimination of water and CO2 molecules and that the two exotherms are due to the decomposition of N27DS anions at the outer surface (520°C) and in the interlayer region (565°C). It is important that two anion-sites at which NijDS ions locate, the outer surface and the interlayer region, are distinguished thermally [26]. The total mass loss in the TG curve in the temperature range up to 800°C is 35.8 % which corresponds to the sum of water, carbonate and N27DS contents (36.0 %) being estimated from the result of the chemical analysis in Table 2. All of the ZnAl/NijDS-LDHs give the similar DTA/TG profile, which indicates that the peak locations are insensitive both to the difference in the magnitude of the basal spacing and to the kind of guest isomer.
Figure 11 - DTA/TG analysis of ZnAl/N27DS-LDH in the range of (room temperature)
362
E. Kanezaki
in his early crystallographic work [13]. He concluded that the interlayer CO32" anion has the conformation in which the molecular plane of the ion lies parallel to the two-dimensional layers thus the magnitude of the interlayer distance is essentially equal to the diameter of the O2" anion. Although an orientation of interlayer organic anions, which is perpendicular to the layer surface, was concluded in earlier studies [25,27,28], a tilted orientation of the anions has been also suggested in recent XRD studies when basal spacings of the intercalated products are smaller than molecular sizes of the anions [2,26,29,30]. Since the interlayer distance of the 17A phase in Table 2 agrees with the estimated molecular size of the anion in all of the ZnAl/NijDS-LDHs, this result suggests strongly that the interlayer NijDS anion in this phase takes the bridging orientation which links a pair of Al3+ ions in the neighboring layers like a fastener zipping on these two layers. The line that connects two anionic oxygen atoms (the 0 - 0 line), each of which belongs to different sulfonate groups in one bridging NijDS, is parallel to the c axis in this orientation. In the 15A phase, however, the interlayer distance is smaller than the molecular size of the guest in Table 2. A plausible explanation is displayed in Figure 12 for the caseof interlayer N15DS; the guest molecule has the orientation in which the O-O line is not parallel to the c axis. It is the author's opinion that the variation in the magnitude of the basal spacing is caused by the variation in the magnitude of the enhanced interlayer distance owing to an alternate orientation by the interlayer molecular anions in the cases of N26DS and N15DS. The variation in the basal spacing does not result from the difference in the thickness of the layer because this variation is so small in both cases (1.7A) that it cannot be explained by the change of the stacking number of the octahedron unit. A pair of Al3+ ions in the 15A phase, which are linked by one N15DS molecule, do not locate directly opposite to each other but are separated in length (R) by 9 A measured along the inner surface of the layers (Figure 12) although R=0 in the 17A phase. The separation in this figure agrees with that value of 5 A previously reported for the case of the interlayer N26DS anion in the 15A phase [26] and both values are approximately equal to twice the interatomic distance between two adjacent metal ions (Ro=3.12 A [34] ) in the layers of Mg and Al double hydroxide. If it is assumed that interatomic distances in the layers of Mg and Al double hydroxide and in the layers of Zn and Al double hydroxide are the same, it is interesting that only two values of the ratio R/Ro are observed; (R/Ro)=2 for the 15 A phases and (R/R<))=0 for the 17A phases. Although the author cannot propose the reason why a value of 1 for the ratio is missing, it is assumed that the reason may be steric hindrance due to the bulky naphthalene moiety. The chemical compositions of the interlayers of the intercalated products in Table 1 shows both of carbonate and NijDS anions coexist in all solid samples of ZnAl/NijDS-LDH. It has been reported that some kinds of LDH with interlayer carbonate are synthesized [4]; the interlayer distance (ca. 3 A) of them, which is equal to the molecular size of carbonate, is much smaller than the estimated molecular size of NijDS in Table 2. Therefore, even when carbonate and NijDS ions coexist in the interlayer, we cannot recognize the coexistence of the two ions in XRD patterns because NijDS ions would act as pillars between two layers of double hydroxide, which is the case which we observe in the 15 A phases. In contrast, when two kinds of intercalated product are formed as a mixture, one is ZnAl/NijDS-LDH and the other is ZnAl/C0 3 - LDH, we can recognize this situation in the XRD measurement, which is the case which we observe in the 17A phases. No evidence to support the presence of superlattice is observed so far.
Preparation of Layered Double Hydroxides
363
The molecular ratio NijDS/CO3 is always larger in the 17A phase than in the 15A phase for ZnAl/N26DS-LDH and ZnAl/N15DS-LDH. Considering that the 17A phases of these two intercalated products are formed together with the ZnAl/CO3-LDH and that the compositions of the 17A phases in Table 1 are thus of a mixture of them, the real ratio in the pure phase is more larger than the value listed in Table 1 for both 17A phases.
Figure 12 - A plausible orientation of the interlayer N15DS between two layers ofZn and Al double hydroxide in the 15A phase; other anions and water molecules are omitted. Reprinted from [48] with permission by Kluwer Academic/Plenum Publishers. We therefore conclude that the guest molecules in the 17 A phase are packed in the interlayer region more closely than in the 15 A phase, which is realized by taking the orientation of the guest molecule in the 17 A phase as described above. The reason why more than one basal spacing is not observed in ZnAl/N27DS-LDH is not known. In the previous work, we reported that only the 15A phase was observed when N26DS is intercalated and that only the 17A phase was observed when N15DS or N27DS was intercalated [26] as listed in Table 2. Together with the results of Drezdzon [25] and of Meyn et al [30], this study reveals that the duality in the magnitude of the interlayer distance due to the orientation change of the interlayer molecules is commonly observed when N26DS or N15DS is intercalated between layers of Zn/ and Al double oxide. This duality has been also reported by our laboratory when AQ26 is intercalated between layers of Mg and Al double hydroxide [2] which is described in the next section. It is
364
E. Kanezaki
important that the magnitude of the basal spacing is changeable by means of controlled preparation conditions. With this changeable magnitude of the basal spacing, we would be able to design materials having desirable pore-sizes in the mesopore region, which is of value in reaction catalysis and in adsorption technology.
2.2 - Intercalation of organic anions by the coprecipitation method 2.2.1 - Intercalation of Naphthalenedisulfonate at the interlayer region of Mg and Al double hydroxide. In the high temperature in situ XRD patterns of MgAl/N26DS-LDH (Figure 13), there are three temperature regions which are classified by the common XRD pattern; room temperature
Preparation of Layered Double Hydroxides
365
interlayer water and v c o vibration of the interlayer carbonate are observed in 3500-3400 cm"1, at 1634 cm"1 and at 1364 cm"1, respectively, in this spectrum. The other sharp absorption bands located in the region 1500-600 cm"1 have been assigned to molecular vibrations of the interlayer N26DS [35,36]. Three S-O bonds in the sulfo group of the interlayer N26DS are classified in two categories; one bond has the oxygen atom coordinating to the metal cation in the layer and the other two bonds have not. The vSo vibration in the former bond is observed in the FT-IR spectrum of the free sodium salt of N26DS (not shown) at 973 cm"1 whereas this band shifts to the lower energy side in all of the FT-IR spectra of the interlayer N26DS up to 400°C (Figures 15(A)-(D)) [16].
Figure 13 - In situ HTXRD patterns of MgAl/N26DS-LDH at different temperature from bottom to top, room temperature, 100°C, 120°C, 140°C, 160°C, 180°C, 200°C, 220°C, 240°C, 260°C, 280°C, 300°C 340°C, 380°C, 400°Q 500°C, 600°C, 700°C, 800°C, 900°C and 1000°C: indexing of the diffraction at room temperature is 20 (hkl) = 5.26° (001), 10.56° (002), 15.86° (003), 21.24° (004), 26.44° (005), 34.68° (102) and 60.92° (110) under the assumption of a hexagonal unit cell with lattice constants ao=3.O4 A and co=16.77A[49] Figure 15(A) shows the FT-IR spectrum of MgAl/N26DS-LDH at room temperature. The broad VOH vibration of the hydroxyl group, the SHOH vibration of the interlayer water and vco vibration of the interlayer carbonate are observed in 3500-3400 cm"1, at 1634 cm"1 and at 1364 cm"1, respectively, in this spectrum. The other sharp absorption bands located in the region 1500-600 cm"1 have been assigned to molecular vibrations of the interlayer N26DS [35,36]. Three S-O bonds in the sulfo group of the interlayer N26DS are classified in two categories; one bond has the oxygen atom coordinating to the metal cation in the layer and the other two bonds have not. The vSo vibration in the former bond is observed in the FT-IR spectrum of the free sodium salt of N26DS (not shown) at 973 cm"1 whereas this band shifts to the lower energy side in all of the FT-IR spectra of the interlayer N26DS up to 400°C (Figures 15(A)-(D)) [16]. Two v s o vibrations in the latter two S-O bonds are observed at 1037 cm"1 and at 1234 cm"1 both in the spectra of the free salt and that of the interlayer N26DS (Figures 15(A)-(D)) although the latter band is not distinct accidentally in Figure 15(A). The coordination of the sulfo
366
E. Kanezaki
group(s) in the interlayerN26DS would have so large influence on the results of XRD and of FT-IR of MgAl/N26DS-LDH that its thermal profiles in both measurements are different from those of MgAl/CO3-LDH.
Figure 14 - 27Al MAS NMR spectra ofMgAl/N26DS-LDH at the spinning rate of 7 kHz; peaks with an asterisk are spinning side bands. [49] Figures 15(B)-(E) are the FT-IR spectra of MgAl/N26DS-LDH at 140°C, 210°C, 400°C and at 500°C, respectively. On elevating the temperature to 140°C, an absorption appears at 866-862 cm"1. Although this band locates near the v Mg0 vibration in the spectra of MgAl/CO3-LDH, it should not be the v Mg0 vibration because MgO is not observed in the XRD pattern of Mg/Al/N26DS-LDH below 500°C in Figure 13. Therefore, the band at 866-862 cm"1 is tentatively assigned to the VAIO vibration of the amorphous aluminum oxide. This assignment is reasonable because the observed energy (866-862 cm"1) agrees with one energy quantum of the vA10 vibration (a(Al-O)=865 cm"1) calculated by the equation (1) with the assumption that the force constant in the v^o vibration K(A1-0) is equal to the constant in the vMgo vibration K(Mg-O), CT(AI-O)/ a(Mg-O)=0.98 {K(Al-O)/K(Mg-O)}°5
(Eq. 2)
where cr(Mg-O) is one energy quantum of the v Mg0 vibration and is equal to 882 cm"1. Since A12O3 is not observed in the FT-IR spectra of MgAl/CO3-LDH it is likely that the presence of the interlayer N26DS is one of the reasons for observing this solid in Figure 15(B). It is supposed that the aluminum ion which is in the Oh coordination sphere of the layer and at the same time is coordinated by the interlayer N26DS at room temperature would be unstable because of the bulky nature of the coordinating sulfo group in the interlayer N26DS and thus the metal cation undergoes the change into the TA structure on elevating the temperature with keeping the coordination by the N26DS. Since the aluminum ion in the TA structure cannot remain in the layer, it moves to the interlayer region thus producing a vacancy in the layer. This thermal change lowers the regularity along the c-axis because the interlayer organic anion no longer supports the width of the
Preparation of Layered Double Hydroxides
367
interlayer distance whereas the regularity along the aft-plane has little influence by the change. The resulting T& Al3+ would be observed in the high temperature measurement of 27 A1 MAS NMR spectra. In the spectrum at 200°C (Figure 15(C)), the absorption by the interlayer water observed at 1634-1627 cm"1 in Figures 15(A) and (B) disappears because of the thermal elimination of water, which elimination is also described in the spectrum of MgAl/CO3-LDH at 210°C (Figure 5(B)). The disappearance of the interlayer water corresponds to the first endotherm at 191°C in DTA/TG thermal analysis curve of MgAl/N26DS-LDH. The spectral profile in the molecular vibrations of the interlayer N26DS at 200°C is almost the same as that at 140°C indicating no change in the chemical status of the organic anion. On elevating the temperature to 400°C (Figure 15(D)), the absorption due to the interlayer carbonate located at 1362-1364 cm"1 in the spectra below 200°C still remains. The interlayer carbonate in MgAl/N26DS-LDH does not change its chemical form on elevating the temperature to 400°C, which result differs from that of MgAl/CO3-LDH. This difference is elucidated from the fact that the interlayer carbonates in the latter compound are in the close contact with the metal cations in the layers because of the narrow spacing of the interlayer gallery region and hence the anions easily undergo the thermal reaction with the cations whereas the anions in the former compound are not. However, the intensity of the bands for the interlayer carbonate is more weak at 400°C than that at 200°C because a part of this anion eliminate from the sample at the former temperature. The elimination of the interlayer carbonate is indistinguishable in the DTA/TG thermal analysis curve of MgAl/N26DS-LDH as an independent event because it is included in the tail of the broad endotherm centered at 434°C. The impossibility of distinguishing the CO2 elimination in the thermal analysis curve has been also described in the result of MgAl/CO3-LDH [42]. Almost all of the molecular vibrations originated from the interlayer N26DS are still observed in the spectrum at 400°C (Figure 15(D)) except for the skeleton vibration of the naphthalene moiety at 1496-1500 cm"1 observed in the spectra below 200°C (Figures 15(A)-(C)) because of the thermal deformation of the skeleton of the moiety. This absence also occurs in the spectrum of the free sodium salt of N26DS at 400°C. In the IR spectrum at 500°C (Figure 15 (E)), however, the intensity of all bands decreases in common. The broad VOH vibration at 3500-3600 cm"1 becomes extremely weak because of the dehydroxylation of the layers, which agrees with the results in XRD and DTA/TG, where the layered structure disappears completely at 500°C in the former measurement and the endotherm at 434°C appears in the latter thermal analysis curve. Furthermore, the profile in the region of the molecular vibration of N26DS is changed considerably because of the thermal decomposition of this organic anion. The strong absorption located at 1180-1187 cm"1 in Figures 15(A)-(D) has been assigned to the 8CH vibration of the naphthalene moiety in the interlayer N26DS [35,36] and disappears in Figure 15 (E) as the result of the thermal decomposition of this moiety. The absorption located at 1032 cm"1 in Figure 15(E), which is assigned to one of the vSo vibrations of the non-coordinating S-0 bond in the sulfo group of the interlayer N26DS as described before, also becomes a weak shoulder at 500°C. The decomposition of the interlayer naphthalenedisulfonate is also observed both in the XRD pattern and in the DTA/TG thermal analysis curve; MgSO4 appears in the XRD pattern at 600°C as a reaction product between Mg2+ and the organic anion (Figure 13) and the broad endotherm centered at 749°C is observed in the latter thermal analysis curve (the result is not shown). In conclusion, the thermal change in the layered structure of
368
E. Kanezaki
MgAl/N26DS-LDH on elevating the sample temperature to 1000°C is understood by means of in situ HTXRD with the aid of DTA/TG, 27A1 MAS NMR and the high temperature in situ FT-IR study.
Figure 15 - High temperature in situ FT-IR spectra of MgAl/N26DS-LDH at room temperature (A), 140°C (B), 200°C (C), 400°C (D) and at 500°C (E). [49] 2.2.2 - Intercalation of 9,10-Anthraquinonedisulfonates at the interlayer region of Mg and Al double hydroxide In this section, the intercalation of 9,10-Anthraquinone-disulfonates (AQij; ij=15, 18, 26 and 27) between layers of Mg and Al double hydroxide is described. The synthetic method of the coprecipitation has been described previously [2,43]. Aqueous solution of MgCl2 and A1C13, metal sources of the double hydroxide, were simultaneously added to an aqueous solution of a sodium salt of AQij with the molar ratio of Mg: Al: AQij and with vigorous stirring at room temperature in air and with keeping pH followed by aging overnight at 73-74°C. The precipitate forms immediately after the addition of solutions altogether and is washed until a negative result in the test for Cl" with Ag+ for the filtrate. Owing to the intercalation of AQ27 between the layers of Mg and Al double hydroxide, the basal spacing in XRD patterns is enhanced to 19 A, 15 A or 12 A and the resulting solids are called hereafter as 19A phase, 15A phase and 12A phase, respectively. The patterns of these solids are indexed and also displayed altogether in Figure 16 with the assumption of the hexagonal unit cell symmetry. It is observed that the magnitude of
Preparation of Layered Double Hydroxides
369
one lattice constant ao (3.06 A) is the same as that of MgAl/CO3-LDH in Table 1 although the other lattice constant c0 is large. The symmetric and sharp basal diffraction indexed (001) is associated with a sequence of prominent (00/) lines in all the patterns and shows that each one of three phases are formed by the pure product. This observation indicates that the brucite-like layers (4.769 A in thickness [4]) in the intercalated products have the well developed stacking in the direction parallel to the c-axis. Basal spacings of all the phases are listed together with the other results in Table 3. Two values of the basal spacing are also observed when AQ26 is intercalated instead of AQ27 by means of the same method, although only the 12A phase has been obtained so far under the hydrothermal condition at C in aging precipitates. Only a single phase has been observed when AQ15 or AQ18 is intercalated. The position of two -SO3" groups in the interlayer AQij thus governs whether the multiplicity in the basal spacing appears or not as well as the degree of the multiplicity. Although the mechanism is not known, it is interesting that the multiplicity appears preferably with isomers of 2,6- and 2,7-substitutions both of which have two -SO3" groups at the most distant positions from the carbonyl groups in anthraquinone. An absorption at around ^.=315 nm has been reported due to the 1B2n(n,nt) transition of anthraquinone [44]. The guest AQij molecule is responsible for all the absorption maxima in the DR spectra of AQij-intercalated products except for a gradual signal rise to the short wavelength side. Considering that no significant change is noticed when the spectra are compared with each other among all the AQij-intercalated products, it is concluded that the interlayer AQij molecules commonly hold the planar framework of the conjugated 7i-electron system in the anthraquinone moiety. A weak absorption due to the 'Blg(n,7t*) transition to the lowest singlet state of anthraquinone [44] is located at around X=485 nm in the DR spectrum of the solid sodium salt of AQij. Instead of this band location, a broad absorption appears at Xmax=501-526 nm. The check does not confirm that this absorption is due to the intercalation of AQij. Judging from the comparable intensity of this band to the allowed (7C,7C*) transition described above, it is suggested that this absorption originates from electronic transition(s) between the conjugated 7t-electron orbital(s) and other orbital(s) which results from the interaction(s) between the intercalated AQij and other species in the intercalated product. XP wide scan spectra of the MgAl/AQij-LDHs commonly show several peaks which are assigned to core level electrons emitted from C, S, O, Al and Mg [32]. The Eb(Al 2p) is 9 eV and 73.7 0.9 eV for the 19A and the 12A phases, respectively; the variation of the energy due to the difference in the basal spacing is insignificant. These values agree with that of the octahedrally coordinated Al3+ within experimental errors [23]. Since MgAl/CO3-LDH has Eb(Al 2p)=73.6 eV [2,26], it is concluded that insignificant change takes place in the chemical status of this cation in the layers upon the intercalation of AQij. A very weak signal at around 200 eV is additionally observed in a few samples after they are etched by an Ar+ bombardment for 60 s. It corresponds to tiny quantity of Cl" detected by the potentiometric titration previously described. Typical results of MgAl/AQij-LDHs are listed in Table 3 including basal spacings and chemical compositions. Since each of these phases has never been produced as a mixture, it is concluded that the optimum synthetic condition for each phase is not overlapped. Although the 12A phase in MgAl/AQ26-LDH is preferable in the hydrothermal condition as previously described, both the phases in MgAl/AQ26-LDH and all of three phases in MgAl/AQ27-LDH are thermodynamically stable at room
370
E. Kanezaki
temperature because no mutual transformation is observed in XRD patterns after these solid samples are left in air for a month. The results of XRD, XPS and DR spectra force us to conclude that the multiplicity in the basal spacing is originated from the distinct dimension of interlayer distance which is governed by the size of the interlayer guest AQij molecules along the c-axis.
Figure 16 - Indexing in XRD patters of MgAl/AQ27-LDH by means of the Ni-filtered CuKa, line (X=1.5405 A); the 19 A phase (upper) with diffraction lines at 29= 4.780° (001), 9.500° (002), 14.20° (003), 18.82° (004), 24.10° (005) and 61.28° (110); the 15A phase (middle) with lines at 5.880° (001), 11.94° (002), 17.84° (003), 24.14° (004), 34.82° (006) and 60.70° (110); the 12A phase (lower) with lines at 26=7.22° (001), 14.34° (002), 22.18° (003), 35.60° (005) and 60.82° (110). Reprinted from [43] with permission by Taylor & Francis.
Preparation of Layered Double Hydroxides
371
The Al/(A1+Mg) ratio in Table 3 remains within the limit for Mg and Al double hydroxide reported previously [4]. The interlayer distance of the 19A phase of the AQ26-intercalated product is 14.26 A which is just less than the calculated guest molecular size /M=15.285 A in Table 3. This is rationalized if we assume that an orientation of the interlayer AQ26 molecule bridging as a bidentate between adjacent layers of Mg and Al double hydroxide, which has been proposed in the cases of N26DS between layers of Zn and Al double hydroxide (Figure 12), C1O4" and SO42" both between layers of Mg and Al double hydroxide [2,4,26,45,46]. This orientation is held by the attraction between the positive charge on Al3+ in the layer and the negative one on -SO3" in the AQ26 anion at the interlayer. A slightly tilted molecular plane of the interlayer AQ26 relative to the inner surface of the layer is suggested for the 19A phase and a more tilted one is supposed for the 12A phase. Similar orientations are proposed for interlayer AQ27 anions in three phases of intercalated products. Table 3 - Summary of synthetic and analytical results of MgAl/AQij-LDHs. Reprinted from [43] with permission by Taylor & Francis. d J) /A
AQij AQ15 AQ18 AQ26
E'VeV 74.0 9 8 9
12.882 10.444 15.285
AQ27
9
15.520
11.64 11.46 12.73 19.03 12.23 15.02 18.47
Formula MgAlo.3,(OH)2.93(C03)0.o6(AQ15)o.09 MgAlo.43(OH) 2 . 85 (AQ18) 021 MgAlo 47 (OH) 2 94 (CO 3 ) 0 i2(AQ26)o.,2 MgAlo.47(OH)295(C03)o.o4(A026)o.19 MgAlo.22(OH)2.39(AQ27)oi3
MgAl0.33(OH)2.63(AQ27)0.19 -
1) binding energy ofAl 2p electron inXPS. 2) molecular size: the sum of ionic diameter of anionic oxygen (2.56 A) and the interatomic distance between two anionic oxygens in different -SO3~ groups in AQij whose geometry is optimized by MO calculation, see text. 3) interplanar spacing obtained d(00l) in XRD patterns. This bridging model of the guest anions leads to the upper limit of the AQ26/A1 ratio is 0.5. When the ratio in the analysis is lower than this limit, the intercalation of CO3 " is not negligible as is indicated in the analytical formulae in Table 2. In the 19A phase, the ratio are larger than that in the 12A phase. It is plausible that the orientation of the interlayer AQ26 anions in the former phase is such that they take the close-packed structure thereby a high ratio is realized whereas the anions in the latter phase do not take the structure. Steric hindrance due to the bulky anthraquinone moiety is supposed for the reason of it although no reason is supposed for the AQ27-intercalated products. 3 - References [1] F. Cavini, F. Trifiro and A. Vaccari, Catalysis Today, 11 (1991) 173; A. Roy, C. Forano, K. E. Malki and J. Besse, Expanded Clays and Other Microporous Solids; Anionic Clays: Trends in Pillaring Chemistry, Eds. M. L. Occelli and H. Robson, Van Nostrand Reinhold, New York, 1992, chap. 7; D. O'Hare, ed., Inorganic Materials, 2nd ed., John Wiley & Sons, Chichester, 1997; V. Rives and M. A. Ulibarri, Coordination Chem. Rev.,181 (1999) 61; T. J. Pinnavaia and G. W.
372
E. Kanezaki
Beall, Eds., Polymer-Clay Nanocomposites, John Wiley & Sons, Chichester, 2000. [2] E. Kanezaki, S. Sugiyama and Y. Ishikawa, J. Mater. Chem., 5 (1995) 1969. [3] MgO (periclase) JCPDS card No.45-9846; MgAl2O4 (spinel) JCPDS card No.21-1152. [4] S. Miyata, Clays Clay Miner., 23 (1975) 369. [5] W. Yang, Y. Kim, P. K. T. Liu, M. Sahimi and T. T. Tsotsis, Chem. Engin. Sci., 57 (2002) 294 [6] M. Belloto, B. Rebours, O. Clause, J. Lynch, D. Bazin and E. Elkaim, J. Phys. Chem., 100 (1996) 8535. [7] M. J. Hudson, S. Carlino and D.C. Apperley, J. Mater. Chem., 5 (1995) 323. [8] L. Li, S. Ma, X. Liu, Y. Yue, J. Hui, R. Xu, Y. Bao and J. Rocha, Chem. Mater., 8 (1996)204. [9] S. Velu, V. Ramkumar, A. Narayanan and C. S. Swamy, J. Mater. Sci., 32 (1997) 957. [10] F.M. Vichi and O.L. Alves, J. Mater. Chem., 7 (1997) 1631. [11] S. Kannan and C.S. Swamy, J. Mater. Sci., 32 (1997) 1623. [12] M. Sato, H. Kuwabara and S. Sato, Clays Sci., 8 (1992) 309. [13] T. Hibino, Y. Yamashita, K. Kosuge, A. Tsunashima, Clays Clay Miner., 43 (1995) 427. [14] J. Santhanalakshmi and T. Raja, Appl. Catal., A147 (1996) 69. [15] F. Kooli, V. Rives and M. A. Ulibarri, Inorg. Chem., 34 (1995) 5122. [16] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, 5th ed., sers. A and B, John Wiley & Sons, New York, 1997. [17] F. A. Cotton, Chemical Application of Group Theory, 2nd ed., John Wiley & Sons, New York, 1971, chap. 10. [18] E. M. Moujahid, J. Besse and F. Leroux, J. Mater. Chem., 13 (2003) 258. [19] V. Prevor, C. Forano and J. Besse, Appl. Clay Sci., 18 (2001) 3. [20] S. Velu, V. Ramkumar, A. Narayanan and C.S. Swany, J. Mater. Sci., 32 (1997) 957. [21] C. Depege, F. El Metoui, C. Forano, A. de Roy, J. Duouis and J. Besse, Chem. Mater., 8(1996)952. [22] M.J. Hudson, S. Carlino and D.C. Apperley, J. Mater. Chem., 5 (1995) 323 [23] W.T. Reichele, S.Y. Kang and D.S. Everhardt, J. Catal., 101 (1986) 352. [24] S. Kannan and C.S. Swamy, J. Mater. Sci., 32 (1997) 1323. [25] M.A. Drezdzon, Inorg. Chem., 27 (1988) 4628. [26] E. Kanezaki, K. Kinugawa, and Y. Ishikawa, Chem. Phys. Lett., 226 (1994) 325. [27] K. Chibwe and W. Jones, J. Chem. Soc. Chem. Commum., (1989) 926. [28] I.Y. Park, K. Kuroda, and C. Kato, Chem.Lett., (1989) 2057. [29] K.R. Franklin, E. Lee and C.C. Nunn, J. Mater. Chem., 5 (1995) 565. [30] M. Meyn, K.Beneke, and G. Lagaly, Inorg. Chem., 29 (1990) 5201. [31] A. Streitwieser, Jr., Molecular Orbital Theory for Organic Chemists, Wiley, New York, (1961), Chap. 8. [32] D. Briggs and M.P. Seah, Practical Surface Analysis-Auger and X-ray Photoelectron Spectroscopy, J. Wiley & Son, New York (1982). [33] R. Allman, Chimica, 24 (1970) 99; ibid, Acta Crystallogr., B24 (1968) 972. [34] D.D. Elleman and D. Williams, J. Chem. Phys., 25 (1956) 742. [35] E. Kanezaki, J. Incl. Phen., 34 (2000) 447. [36] E. Kanezaki, Mater. Res. Bull, 34 (1999) 1435.
Preparation of Layered Double Hydroxides
373
[37] MgSO4; JCPDS card No. 21-0546. [38] E. Kanezaki, Inorg. Chem, 37 (1998) 2588. [39] E. Kanezaki, J. Mater. Sci. Lett., 17 (1998) 371. [40] E. Kanezaki, Solid State Ionics, 106 (1998) 279. [41] E. Kanezaki, Mater. Res. Bull., 33 (1998) 773. [42] V. Rives, Inorg. Chem, 38 (1999) 406. [43] E. Kanezaki, Mol. Cryst. Liq. Cryst, 286 (1996) 153. [44] E. Kanezaki, N. Nishi and M. Kinoshita, Bull. Chem. Soc. Jpn, 52 (1979) 2836. [45] E. Kanezaki, J. Mater. Sci, 30 (1995) 4926. [46] V.R.L. Constantino and T.J. Pinnavaia, Inorg. Chem, 34 (1995) 883. [47] E. Kanezaki, J. Incl. Phen, 46 (2003) 89. [48] E. Kanezaki, J. Incl. Phen, 24 (1996) 41. [49] E. Kanezaki, unpublished result. [50] E. Kanezaki et al, Mater. Res. Bull, submitted.
POLYOXOMETALATE HYDROXIDES
COMPLEXES
OF
LAYERED
DOUBLE
CHANGWEN HU7* and DANFENG LI2 ' Department of Chemistry, Beijing Institute of Technology, Beijing, P.R. CHINA, 100081 2 Institute of Polyoxometalate Chemistry, Northeast Normal University, Changchun, P.R. CHINA, 130024 * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych andK.G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Polyoxometalate Complexes of Layered Double Hydroxides
375
I - Introduction Studies on layered double hydroxides (LDH) materials as a new research field has attracted much attention in recent years [1,2]. The general formula for LDH is [M[_ 2+ 3+ x+ n 2+ and M3+ ions usually represent metallic cations x Mx (OH)2] Ax/n "-/MH2O, where M 2+ 2+ 2+ 2+ 2+ such as Mg , Zn , Cu , Mn , Co , Al3+, Cr3+, Fe3+ ions, etc. The central positions of octahedra of M(OH)6 are occupied by the cations, and An~ represents the interlayer anions such as NO3", CO32", C2O42", etc. The positively charged sheets are formed through M(OH)6 edge-sharing octahedra and balanced by An" anions in order to maintain the electrical neutrality of the complexes. So far, a large number of LDH has been synthesized, and the interlayer anions may be inorganic anions, organic anions, single-metal anionic complex, oxometalate complexes, etc [3]. If the interlayer anions An" are oxometalates, LDHs are generally classified into high-nuclearity system, medium-nuclearity system, and low-nuclearity system according to the nuclearities of A"". Polyoxometalates (POM)-LDHs are of the high-nuclearity system. As the name indicates, POMs are constituted primarily by d° early transition metal cations and oxide anions. The principal d° transition metal ions that form the molecular structural framework or scaffolding of POMs are W, Mo, V, Nb, Ta, Ti, etc. There are two generic families of POMs, the isopolyoxometalates, which contain only transition metal cations and oxide anions, and the heteropolyoxometalates which contain one or more p or d block elements as heteroatoms located at structurally well defined sites, in addition to the more numerous transition metal and oxide ions [4]. The basic structural principle for POMs is the same. Their structures are governed by the principle that each metal atom occupies an [MOX], coordination polyhedron and form a very large cluster system. Fused [MO6] octahedrons served as the constituent unit in the structure of isopolyoxometalates. The structures of isopolyoxometalates W7O24S~, Vi0O286~, and H2W1204o6" are shown in Figure 1 (a), (b), and (c), respectively. Keggin structure heteropolyoxometalates [XM12O40]"" can be denoted as X M n (X may be B, Si, Ge, P, As, and some other atoms; M is either molybdenum or tungsten). In which, twelve octahedrons are fused and four M 3 On groups are arranged around the heteroatom X which is tetrahedrally surrounded. A M3Oi3 group is an assembly of three octahedrons sharing edges so that they have a common corner, which is also a corner of the central tetrahedron, as shown in Figure 2. Isomer ct-XM12 has Td symmetry and isomer (3-XM[2 has C3v symmetry after rotation of one M3O13 group located at the top of the polyanions. One metal group MO of the above two isomers may be missing and yielding an I1 metal atom POM, [XM U O 39 ] 8 \ denoted as XM n . The vacant site can be occupied by other metal atom, then forms monosubstituted Keggin compound XM n Z. When three metal atoms have been taken off the Keggin structure XM12, compounds with three vacant metal sites XM9 can be formed. The three metal atoms may be taken from different skeletal plane of XM r2 , and there are two isomers A-XM9 and B-XM9, their structures are shown in Figure 3. When the three vacant sites have been occupied by other three metal atoms, then trisubstituted Keggin compound formed, such as XM9Z3. Figure 4 shows the structure of [U(a-SiWnO39)2]12~, in which the [SiWn] ligands are trans-conformation and connected by UIV center (white sphere).
376
Changwen Hu and Danfeng Li
Figure 1 - The structure of some typical isopolyoxometalates.
Figure 2 - The structure ofKeggin type heteropolyoxometalates. [Co4(H2O)2(AsW9O34)2]10" has a core of four fused cobalt-containing octahedrons, which is liquated to two AsW9O34 units by coordinating to one oxygen atom of AsO4 and six oxygen atoms of the nonatungstate unit, as shown in Figure 5. Dawson structure POM [X2M18O62]"~, denoted as X2M18, derives from the Keggin structure XM12 by withdrawing three metal atoms from three different M 3 Oo groups. Similar to their Keggin counterparts, lacunary or substituted Dawson derivatives X2M17 or X2M17Z also can be formed, as shown in Figure 6.
Polyoxometalate Complexes of Layered Double Hydroxides
311
Figure 3 - Lacunary or substituted derivatives of the Keggin POMs.
Figure 4 - [U(a-SiWuO19)2y2-
Figure - 5
[Co4(H2O)3(PW9O34)J.
[NaP5W30Ono]14 is known as the Preyssler compound. As shown in Figure 7, the POM has a five-fold axis of symmetry with a sodium ion encapsulated into as a large gray center. The five [PW6] groups is a cyclic assembly. An other bigger cluster, namely [H2Si4Nbi6O56]14" has been synthesized recently [5], as shown in Figure 8, the NbO 6 octahedra (NB1,3,5,6,1O and 16) constitute one 3/4-Keggin subunit, the dark NbO6 octahedra (NB2,4,8,11 and 15) constitute the second 3/4-Keggin subunit, and the two NbO6 octahedra are shared between the two "fused" subunits, SiO4 tetrahedra are gray. [SiNb1204o]16' and [Ti2O2]4+ chains is shown in Figure 9.
378
Changwen Hu and Danfeng Li
(a) X 2 M 1 8
(b) X 2 M 1 7 Z
Figure 6 - The structure ofDawson type POMs.
Figure 7 - Preyssler compound [NaPjW}oOllo]l4~. It is well known that POM compounds can be used as catalysts owing to their excellent acidic and oxidative functions [6]. Recent investigations show that POMLDHs are a kind of solid functional materials and may be used in many fields such as catalysis, pharmaceutical chemistry, adsorbents, ion exchangers, and environmental protection [7]. In view of the application on the field of catalysis which is particularly emphasized, it is of great value to intercalate POM into the interlayer of LDH to form POM-LDH compounds because (i), LDH as a class of solid layer materials is also widely used in catalysis [8]; (ii), the practical recycle problem when using soluble POM in the process of industrial catalysis may be solved [9]; (iii), The increased special surface area would be attributed to the increase of the catalytic activity of the POMLDH compounds [10]; (iv), the size-selectivity catalysis function may be produced by using the POM-piUared catalyst as interlayer is believed as a reaction field where anion exchange reaction between the interlayer guest species and other molecules or anions with suitable size may occur [11]. Their shape-selectivity is similar to that of the molecular sieves. In addition, some active catalysts may be formed from POM-LDH
Polyoxometalate Complexes of Layered Double Hydroxides
379
precursors. For example, as highly active catalysts for some organic reactions, the nanometer mixed oxides with definite structure can be obtained by calcination of POMLDH at various temperatures [12].
Figure 8 - The structure of[H2Si4Nb,6O56JN' cluster.
Figure 9 - The structure of the [SiNbl2OM]16' and [^Orf4'
chains
2 - Structural characteristics of POM-LDH The structure of POM-LDH involves (1) the pillared layered structure, i.e. the whole structure is composed of the layers of double hydroxides of metal ions (M2+ and M3+) and the anions in gallery; (2) the orientation and location of the polyanion that interacts with the layers among the layers [7]. A schematic view of the structure of the POM-LDH is shown in Figure 10. It can be seen from the figure that the gallery heights of the POM-LDH were calculated by subtracting the thickness of the host layer from basal spacing (c/Ooi) of the XRD data of the samples. Also, it can be seen that the size of the intercalated POM anions are responsible to the gallery height (interlayer spacing) of the POM-LDH. The layered structure of the POM-LDH is very similar to that of brucite (Mg(OH)2). In general, the divalent metallic cations (M2+) in the layers are partly replaced by highly charged cations (M3+) forming double hydroxides. These octahedrons sharing edges with OH form 0.47 run thick layers [2]. The intercalated
380
Changwen Hu and Danfeng Li
POM anions are linked with positively charged host layer by means of electrostatic force and hydrogen binding formed through the water molecules of the interlayer of the hydroxyl group in the layers so that the system turns to electrical neutrality. The pillared structure of POM-LDH and the structure of POM in the gallery may be examined by XRD and IR spectra [13-15]. The dOoi values of some reported POM-LDH are listed in Table 1 [9,10,16-28].
Figure 10 -A schematic view of the structure of the POM-LDH compounds. Table 1 also indicates that the d Ooi values of the POM-LDH are the same of Keggin type POM (1.47+0.02 nm). In a Keggin-type structure, twelve metaloxygen octahedrons form a shell surrounding tetrahedrally coordinated heteroatom (P, Si, B, etc.); the shape is close to spherical, with a diameter of ca. 1 nm. When the thickness of the host layer (0.47 nm) is subtracted from the c?Ooi value, the obtained gallery heights (1.00+0.02 nm) are in agreement with the sphere diameter of Keggin-type heteropoly anion [23]. When the elemental composition of POM anions having Keggin structure (central atom X and substituted coordinative atom Z, etc.) varies, the d0QX values have no changes as expected, because their spherical volumes are hardly changed. Other type POMs, including lacunary Dawson ion, a-[P2Wi7O6i]10~, Dawson ion, ct-[P2Wi8O62]6", Finke ion, [Co4(H2O)2(PW9O34)2]10", Preyssler ion, [NaP5W3oOno]14~, etc., have been successfully intercalated into Mg2Al, Zn3Cr or Zn2Al-LDHs. Compare the result with intercalation of Keggin ions, larger gallery heights can be achieved. On the contrary, the less size isopolyoxometalates, [Mo7024]6", [Vio028]6~, [H2Wi2O40]6" intercalated into the interlayer space, less gallery heights are obtained. The structural, thermal and textural properties of POM-LDH intercalates have been demonstrated on the basis of XRD, FTIR, 31P NMR and MAS-NMR, and N2 adsorption-desorption studies [10,14,25,26].
381
Polyoxometalate Complexes of Layered Double Hydroxides
Table 1- The basal spacing (0
dm /nm
[Mo 7 O 24 f [V 10 O 28 ] 6 -
[H 2 W 12 O 40 f a-l,2,3-[SiW 9 V 3 04o] 7 "
Ref.
1.44
Calculated gallery heights/nm 0.97
1.22
0.75
[17]
1.48
1.01
[18]
1.46
0.99
[10]
[SiW 9 O3 7 Z3(H 2 O) 3 ] - (Z=Co , Cu )
1.46-1.47
0.99-1.00
[19]
[GaW 9 O 3 7Z3(H 2 O) 3 ] n " (Z=Co 2+ , Fe 3+ )
1.46-1.47
0.99-1 1.01+0.02
[20]
2+
10
n
5+
2+
4+
2
[PW n ZO 3 9 ] - (Z=V , Ti )
[16]
0.98
[9] [10]
1
0.99 0.99+0.02
[21] [22]
1.4610.01
0.99+0.02
[AsW u O 3 9 Co] -
1.55
1.08
[9] [23]
[GeW u O 3 9 Z(H 2 O)] 6 XZ=Ni 2 + , Cu 2+ )
1.45+0.01
1
[21]
[BW 11 O 39 Z(H 2 O)] n "(Z=Al 3+ , Ni 2+ , Cu 2+ )
1.45+0.01
1
[9]
2
2
[24]
[SiW n O 3 9 f
1.45
[GeWnO 39 f
1.46
[ S i W n O 3 9 Z(H 2 O)] 6 -(Z=Ni 2+ ,Co 2+ , Mn 2+ ) [XW n 0 3 9 Co(H 2 0)] n (X=P 5 + ,Ge 4 + ,B 3 + ) 5
[Ln(XW n O 3 9 ) 2 ] n (Ln=La 3+ ,Ce 3+ ;X=Si 4+ ,B 3+ ) a-[P 2 W 18 O 62 ] 6 " [Co 4 (H 2 O) 2 (PW 9 O 3 4) 2 ] 10 [NaP 5 W 3 oO U o] 14 -
1.75 or 1.93 1.77+0.03
1.28 or 1.46
[25, 26]
1.30+0.03
[27]
2.17 1.71
1.70 or 1.24
[28]
or
Differences are also found due to different orientations that these anions can achieve in the interlayer space. If the ions between the layers possess different axis, different dool values will be obtained, as is shown in Figure 11. From Figure 11, it can be clearly seen that different orientations in the interlayer region in view of the exchange products are responsible to the difference of the JOoi values. Structurally, the special orientation and localization of POM anions in the pillared POM-LDH is a significant subject. Pinnavaia et al have predicted that between the double layers the anion of Keggin-type with C2 symmetrical axis is perpendicular to the layers consisting of M2+ and M3+ cations, but this prediction has not been verified experimentally [26]. According to the 27A1 and 31P MAS NMR for the Pillared ZnAl-PWnV, we have succeeded in confirming the orientation recently [14]. The peak position and the peak shape in 27A1 MAS NMR for the sample show no obvious change, indicating that the Al in the host layer still lies in the center of an octahedral before and after ZnAl-PW n V being intercalated into the gallery. It has been verified experimentally that PWnV anions have no isomers [23]. However, the 31P MAS NMR in the sample appears two
382
Changwen Hu and Danfeng Li
peaks with the area ratio of 2:1, which indicates that the PWUV anion with C2 axis as the symmetrical axis is perpendicular to the host layers, as shown in the two orientations in the gallery from Figure 12.
Figure 11 - Model of POM with different orientation in the interlay er ofPOM-LDH.
Figure 12 - Structure model of the orientation of mono-substituted Keggin [PWuVO39]4' anion in the interlayer ofZnAl-LDH. Also, two similar orientations of Dawson a- [P2W18O62]6" ion with its C3 axis parallel or perpendicular to the layers have been reported by Pinnavaia et. all via the meixnerite method [17], as shown in FigureD. The basal spacing was 1.93 nm and 1.75 nm, respectively. The irreversible decrease in the basal spacing upon heating at 100°C is believed due to the unstable condition for the former orientation. In the field of POM-LDHs, it is a very interesting scientific topic to study the orientations of POM clusters in the interlayer space. Figure 14 is our proposed structure model of two orientations of Preyssler-type [NaP5W3o0110l14" ion in the interlayer when different precursors are used. That is, the interlayer [NaP5W3oOnol14 ion is oriented with its C5 axis perpendicular and parallel to the brucite-like layer, respectively. The
Polyoxometalate Complexes of Layered Double Hydroxides
383
measured spacings for the 001 reflection are 2.17 and 1.71 nm, respectively. The above result is in agreement with the prediction of Evans et al, but they have only prepared Zn2Al-[NaP5W3o0110]14' with its C5 axis parallel to the brucite-like layer [28]. We hope that more and more POMs with complicated structures will be intercalated into the interlayer space of LDHs, and their orientations can be analyzed continuously with the development of science and technology.
Figure 13 - Reaction scheme for intercalation ofDawson POM in a LDH.
Figure 14 - Structure model of the two orientations of Preyssler [NaP5W30Ouo] ' ion in the interlayer space.
3 84
Changwen Hu and Danfeng Li
3 - Synthesis method of POM-LDH POM-intercalated LDH represent a new class of pillared materials for catalysis. However, it is not easy to synthesize POM-LDH in highly crystalline and pure form, because LDH hosts are basic, whereas most of POMs are unstable under basic condition. Thus, hydrolysis reactions of the POM may partly occur on the process of preparation of the POM-LDHs, resulting in amorphous and impure products. In addition, it is much important to intercalate effectively POM anion of large volume into the interlayer of LDH. As a matter of fact, studies on the POM-LDHs pillared by polyoxotungstates are relatively more active than those pillared by polyoxomolybdates or polyoxovanadates as polyoxotungstates are more stable than polyoxomolybdates or polyoxovanadates. Since 1988 Pinnavaia group reported the first isopolyoxometalatescontaining LDH complexes, namely ZnAl-Vi0O286", many new type POM-LDHs have been reported. Among these groups, Pinnavaia's [27], Evans' [28], Min's [29], and Hu's [30-32] are much active. The intercalated POMs include isopolytungstates and heteropolytungstates with different nuclearities, i.e., [W7O24]6" ion, [H2W1204o]6~ ion, the Keggin-type a-[SiW n O 3 9] 8 \ a-l,2,3-[SiV 3 W 9 O 37 ] 7 - and [XW,,O39Z(H 2 O)r ions, the Dawson-type a-[P 2 W 18 O 62 ] 6 ", a-[P 2 W 17 O 61 ] 10 " and [P2W17O61Mn(H2O)]8" ions, the Finke-type [Co 4 (H 2 O) 2 (PW 9 O 34 ) 2 ] 10 \ and [Zn4(H2O)2(AsW9O34)2]10" ions, the double Dawson-type [P 4 W 3 oZn 4 (H 2 0) 2 O n2 ] 16 " and the Preyssler-type [NaP 5 W 3 (Aio] 14 ' ions. These POM-LDHs with different gallery heights are generally synthesized by anion exchange reactions. The properties such as the basicity of the precursors and the charge density have effects on the crystallinity and the orientations of the products, which finally ensure the maximum interactions between OH groups on the host sheets and the interlayer guest POM anions. Up till now, six different methods have been widely used to synthesize POM-LDH. Pinnavaia's and Hu's group have summarized these synthesis methods recently, as described and compared in the following [17,33]. 3.1. - Organic-anion-pillared precursor-exchange (OPE) method In those previous methods, the anions encapsulated in gallery of LDH are small oxo anions, such as Cl~, CO32", NO3", etc., so that the interlayer gallery heights have been limited and the exchange reaction with POM is slow. Drezdzon has taken a novel approach to solve those problems [16]. By using the precursor MgAl-TA (TA '= terephthalate), Dredzon has prepared MgAl-[Mo 7 O 24 f" and MgAI-[V10O28]6". The principle method is as follows: At first, an organic-anion-pillared clay precursor MgAlTA was prepared. After that, MoO42" or VO3~ were polymerized into [Mo 7 02 4 ] 6 or [Vi0O28]6" when the slurry of MoO42" (or VO3") and Mg 2 Al-TA were acidified to pH 4.44.7. At the same time, TA2~ anions reacted with H+ coming from an aqueous solution, leading to the neutral molecular H 2 TA, which escaped from the electrostatic interaction with the host sheets. Finally, [Mo7O24]6" or [V10O28]6" entered the interlayer space to balance the positive charged sheets, i.e., the precursor was exchanged with POM to obtain POM-LDH having a large gallery height. The results of XRD indicated that the basal spacings for MgAl-[Mo 7 O 24 ] 6 " and MgAl-[V10O28]6" are 1.44 and 1.22 nm, respectively. The two intercalates also have high crystallinity. The process can be described as Formula 1. Mg2Al-TA + [Mo7O24]6"
pH= 4.4-4.7, 70cC »-
Mg2Al-[Mo7O24]6' + TA2'
(Formula 1)
Polyoxometalate Complexes of Layered Double Hydroxides
385
Recently, we also used Mg2Al-TA as the precursor to react with aqueous W7O246" anions to prepare Mg2Al-[W7O24]6" in order to obtain the solid [W7O24]6", which is difficult to be prepared by other methods [32], as shown in Formula 2 and Figure 15. Mg2Al-TA + [W7O24]6"
pH= 6 7 70
' '
°C>
Mg2Al-[W7O24]6" + TA2"
(Formula 2)
Figure 15 - The structure ofMg2Al-[W7O24]6'. 3.2. - Reconstitution method The reconstitution method for the preparation of POM-LDH utilizes the reconstitution of mixed metal oxide solid solutions [34]. In this method, a CO32"-LDH precursor was thermally decomposed to form the oxide solid solution, and then the oxide was reconstituted into a pillared LDH by hydrothermal reaction with the pillaring POM anion. The process can be described in Formula 3.
Mg2+-Al3+-CO32-
2+
Mg -AT -OH
5
° ° ° C ' 3 h > Mg-Al-0 (ox.de solid solution)
[C 6 H 8 0 4 f, glycerine
Mg 2 + -Al 3 + -[H 2 W 1 2 O 4 0 ] 8
2+
Mg - Al -C6H8O4
H
2°-25°C.
3Oh
,
[H 2 W I2 O 40 f, H 2 0 /glycerine +-
Formula (3)
This method has generally been used to prepare POM-LDH containing Keggin type POM anions with higher gallery heights [10,35]. 3.3. - Swelling agent (SA) method To obtain exceptionally well-ordered LDH containing carboxylate and other organic anions, which are excellent precursors for the preparation of regularly microporous materials pillared by POM anions of Keggin type, glycerol as a swelling
386
Changwen Hu and Danfeng Li
agent has been utilized to facilitate the reaction of an organic acid with the hydroxide exchange form of the LDH [18,36]. For example, meixnerite [Mg3Al(OH)8]OH-2H2O was selected as a hydroxide ion exchange form of an LDH. This compound was conveniently prepared from hydrotalcite, [Mg3Al(OH)8][C03]o5-2H20, by first calcining the hydrotalcite at 500°C for 3 h to form a mixed oxide solid solution, and then slurrying the mixed oxide in degassed water (25°C, 16 h) under a CO2-free nitrogen atmosphere to form the crystalline meixnerite. The addition of two volumes of glycerol to the slurry resulted in the extensive swelling of the LDH layers. The reactions of meixnerite in 1:2 (V/V) H2O glycerol with long-chain carboxylic acid or with adipic acid in stoichiometric ratios proceeded smoothly at ambient temperature to form the corresponding organic anion intercalates. Subsequently, the organic anions in the LDH interlayers were exchanged by Keggin species [SiWnC^] 8 " to yield the corresponding POM-Pillared MgAl-[SiWnO39]8". 3.4. - Direct coprecipitation (DC) method The DC method for the preparation of POM-LDH is described in the following example [37]. First, K 8 [SiW n 0 39 ]12H 2 O (1.2 mmol) was dissolved in 120 ml of CO2free water under Ar atmosphere. The pH of the POM solution was adjusted to 6.0 by the addition of 0.2 mol/L HNO3. Then, 20 ml of a mixed aqueous solution of Zn(NO3)2 (1.2 mmol) and A1(NO3)3 (0.60 mmol) was added dropwise to the rigorously stirred POM solution at 100°C. The pH of the mixture was maintained near 6.0 by simultaneous addition of 0.20 mol/L NaOH to the POM solution. After reaction for 1 h, the solid product was separated by centrifugation, washed thoroughly with deionized water, and dried in air at 90°C, the process can be described as Formula 4: „ [SiW11O39]8" + Zn(NO3)2/Al(NO3)3
pH=6.0
»- ZnAl-[SiW n O 39 r
(Formula4)
3.5. - NO3-LDH exchange (LE) method First, a NO3-LDH precursor is synthesized according to a conventional procedure. Then, in aqueous solution without assistance of other agents, the exchange reaction of the NO3-LDH with a POM of K+ salt can be performed effectively to obtain the product POM-LDH. We have prepared a group of ZnAl-POM clays by the LE method [14]. A typical procedure is as follows: under the N2 atmosphere, an aqueous solution of a POM (K+ salt, ca. 9.0 g in 30 ml distilled water) was added to a stirred suspension of NO3'-ZnAl (3.0 g) in 20 ml CO2-free, distilled water at 90°C, the resulting suspension was left to stir under these conditions for 10-15 h and then the product was separated by centrifugation, washed with hot water several times and finally the crystalline product was dried at 120°C for 10 h. Hu et al have developed a modified LE method in which the exchange reaction of the NO3"-LDH with a POM (K+ salt) was carried out in microwave conditions [38]. The experimental results show that the exchange reaction in microwave conditions can give a product rapidly and exchange completely. 3.6. - Phase-transfer method using surfactant This method was first reported by Valim et. all, as shown in Figure 16 [39].
Polyoxometalate Complexes of Layered Double Hydroxides
387
The mechanism involved a big basal spacing LDH, namely, ZnCr-DS (DS represents dodecylsulfate) precursor. Since the salt formed between an anionic surfactant such as DS and a cationic one such as A^-cetyl-AyvyV-trimethylammonium bromide (CTAB) is insoluble in the aqueous phase so easily moves from the aqueous medium and entering into an organic phase such as chloroform. Therefore, the surfactant anion is removed from the interlayer domain and the migration of the salt to the organic phase favors the concomitant intercalation of the anion of POM (A n ) with different size. Formation of the salt and its migration to the organic phase are fast, so is the anion exchange process. We have also attempted to use this new method to intercalate larger polyanions such as Keggin, Dawson or Presslyer-type POMs into the ZnCr-DS precursor, but only the Keggin-type POMs were intercalated into the interlayer with poor crystallinity. The possible reason is that CTAB has the opportunity to form a precipitate with the voluminous POMs so that the anion exchange reaction cannot be performed smoothly. Therefore, we conclude that preparation of the POMLDH intercalates by using Valim's method must be performed in a dilute solution and the intercalated POM should not be too large.
Figure 16 - Schematic representation of the phase-transfer method using surfactant In above-mentioned six methods, each of them displays respective advantages and shortcomings according to different objectives. The OPE, SA and the LE methods are normally quite convenient to carry out in the stage of ion exchange reactions, but they require the synthesis of a suitable LDH precursor. Moreover, exchange reactions may be slow, particularly when the exchanging anion is larger than the anion initially occupying the LDH galleries. Although the DC method has the advantages of rapidly preparing POM-LDH, it is difficult to ensure a product with good crystallinity and purity. In general, the reconstitution reactions require multi-step procedures and are even more inefficient than the ion exchange ones. Ulibarri et al carried out comparative study for these synthetic methods and have obtained some valuable results [36]. By these methods, a lot of POMs-LDH has been prepared and some of them are listed in Table 2. As is shown in Table 2, when Keggin saturated polyanions [SiW12O40]4" or [PW12O40]3" enter into the interlayer space, it is inevitable for them to hydrolyze partly and one water molecule occupies the position of the WO group which has been lost. So it is necessary to correct their compositions to be [SiWuO39(H2O)]8" and [PWnOsoCHjO)]7-.
388
Changwen Hu and Danfeng Li
Table 2 - List of POM-pillared LDH prepared with above 6 methods. Abbr. Formula
Intercalated POM
Ref.
MgAl-Mo 7 O 24 MgAl-V 10 O 28 MgAl-SiWn ZnAl-SiV 3 W 9 ZnAl-SiWu
[Mo7O24]6" [V 1 0 O 2 8 f [SiW H O 3 9 f
16 16 18 10 10,37 37 38 39 40 41 42 43 43 44 20 45,46 47 48 49 21 21 17,27 15 19 19 2 50
Z11AI-PV3W9
ZnAl-PZW n ZnCr-V 10 O 28 ZnAl-H 2 W 12 O 40 MgAl-PVnW12.n NiAl-PW 12 NiAl-PCoWn MgAl-SiW 12 MgCr-V 10 O 28 MgAl-GaZ 3 W 9 ZnAl-SiZW,, MnAl-PWnCu ZnAl-XW n Co ZnAl-PVnW12.n ZnAl-GeZW,i ZnAl-GeWn ZnAl-BVWuO^ ZnAl-BZWu ZnAl-SiZ 3 W 9 ZnAl-Ln(XWn) 2 ZnAl-NaP5W30 Ni(Co)Al-XZW n MgAl(Ga)-PW n O 39 , H 2 W I2 O 40 ZnAl-PZWn Mg (Zn)Al-POM Mg (Zn)Al-POM Mg (Zn)Al-POM
a-[SiV 3 W 9 O 3 7 ] n -
a- [SiW n O 3 9 f a- [PV 3 W 9 0 37 f [PZW n O 39 ] 5 "(Z= Co 2+ , Ni 2 + , Cu 2+ , Zn 2+ , Fe 2+ )
[v 10 o 28 ] 6 a-[H 2 W 1 2 O 4 0 f [PV n W 12 . n ] (n+3) -(n=0,2,3,4) [PW n O 3 9 (H 2 O)f [PCoW n O 39 ] 7 [SiW n O 3 9 (H 2 O)f [V 10 O 28 ] 6 [GaZ 3 W 9 O 37 ] n " (Z= Co 2+ , Fe 2+ ) [SiZWnO 39 ] 6 " (Z= Mn 2+ , Co 2+ , Ni 2+ , Cu 2+ , Zn 2+ , Fe 2+ )
[PW n CuO 3 9 f [XW n CoO 3 9 ] n " (X= Ge, B , Co) [PV n W 12 . n ] (n+3> (n= 1,2,3,4) [GeZW u O 3 9 ] 6 " (Z= Ni 2+ , Cu 2+ )
[GeW n O 3 9 f a-[BVW u O 39 ] 6 [BZW,,O39]n" (Z= Ni2+, Cu2+, Al3+) [SiZ3W9O37]n"(Z= Co2+, Cu2+) [Ln(XW n O 39 ) 2 ] n ", Ln= La3+, Ce3+, Ce4+; X= P, Si, B) [NaP5W30O110]14" [XZW,,O39(H2O)]n", (X= P, Si, Ge; Z=Cu2+, Co 2+ , Ni2+) [PW n O 39 ] 7 -, [H 2 W 12 O 40 f [PW u CrO 39 ] 4 -, [PW n ZO 40 ] n - (Z= Ti, V) a-[P 2 W, 8 O 62 f,a-[P 2 W 17 O 61 ] 10 -, Co4(H2O)2(PW9O34)2]10-, [NaSb9W21O86]18[Zn4(H2O)2(AsW9O34)2]10-, [WZn3(H2O)2(PW9O34)2]12-,[P4W30Zn4(H2O)2O112]16[W7O24]6", [P2W17MnO61(H2O)]8-
51 52 2,27 2,27 11,32
4 - Catalysis by POM-LDHs So far, POM-LDHs are mainly used as acid, redox catalyst and photocatalyst. In this chapter we will review in detail the application of POM-LDHs on these aspects. 4.1. - Acid and redox catalysis It is known that heteropolyacids and their salts, a class of POMs, have
Polyoxometalate Complexes of Layered Double Hydroxides
389
exhibited excellent acid and redox catalysis. Several industrial processes such as hydration of olefins and polymerisation of tetrahydrofuran have been operated utilizing heteropolyacids. However, POMs are expensive and have to be recovered after being used in reactions, thus rendering their repeated use difficult. POM-LDH derived from intercalating POM anions into hydrotalcites provides an important way to support POM on supporter. The reaction of acetic acid with n-butanol forming n-butyl acetate is of great interest in industry. We used recently some of POM-LDH materials in the state of solid acids as catalyst to investigate their catalytic activity for esterification in the liquid-solid biphase systems [53]. The reaction equation is as follows.
CH3COOH + CH3(CH2)3OH -> CH3COO(CH2)3CH3 + H2O
Under the same conditions, the catalytic activities of ZnAl-PWnZ (Z=V, Ti, Cr, Ni) were determined. It was found that the catalytic activity of ZnAl-PWnTi was the highest, even twice as high as that of H-type molecular sieves (HY) commonly used in industry. The sequence of esterification activities was found as follows: ZnAl-PWHTi (59.2%) > ZnAl-PWnV (48%) > ZnAl-PWnCr (42%) > ZnAl-PW,,Ni (36%), (data in the brackets are conversions for the reactions time 5 h). This indicates that these monosubstituted anions of Keggin-type create new active centres that ZnAl-NO3" precursor or PW n Z does not possess. It is considerably possible that the new active centres are derived from the interaction between the Keggin type anions and the host layers. In addition, we have performed the following experiment to test the stability and lifetime for the catalyst of ZnAl-PW n Ti. ZnAl-PW,,Ti (0.8 g) was used to catalyse the reaction for 2.5 h, then the catalyst was filtered and then put into another reaction of the same reactants, the reaction was also carried out for 2.5 h, while the mother liquor with the catalyst filtered was allowed to react under the same conditions for the same period. For the catalysts after being used for the second time, no reduction of their activities could be found, while the reaction of the mother liquor with the catalyst removed could hardly be found. The above operation was repeated for 5-8 times, and the same results were obtained. IR and XRD result for the PW n Ti being used as catalyst for 5-8 times showed that its structure was still remained in the ZnAl-PWuTi. This demonstrates that the pillared layered Keggin anions are stable in structure. The data of the catalytic activities and selectivities for the reaction of acetic acid with n-butanol, using the catalyst of ZnAl-GeWn, ZnAl-GeWuNi, and ZnAl-GeWnCu, etc. are listed in Table 3. By comparison with ZnAl-NO3 and GeW u , the catalytic activities of ZnAl-GeWn, ZnAlGeW n Ni, and ZnAl-GeWnCu are markedly higher in the reaction [21]. The selectivities of these catalysts are nearly the same, which are high for the esterification (>98.8%) and low for the etherification (<1.2%). ZnAl-XWnCo (X= Ge4+, B3+ and Co2+), synthesized recently by us, has also been used as catalysts in the reaction of acetic acid with n-butanol. The high activities of them compared with that of ZnAl-NO3, GeW n Co and HY are remarkable. The catalytic activity order is as follows: ZnAl-GeWuCo > ZnAl-BWnCo > ZnAl-CoWnCo. The order shows that the catalytic activity increases with the increase in valence of the central atom, which is in agreement with the catalytic activity order of their nonintercalated POM anions [54]. When the reaction was carried out for 5 h, the selectivity
390
Changwen Hu and Danfeng Li
of w-butyl acetate was almost 100% and without by-product being observed. The time courses of the esterification reaction for these POM-LDH catalysts are shown in Fig. 17. Table 3 - Esterification activity and selectivity of acetic acid with n-butanol". Catalyst*
Catalyst mass/ g
Acetic acid conversion (%)
Selectivity ("/ «-BuOAC
0.8002 0.2 99.8 66.5 ZnAl-GeWn 0.5 99.5 57.3 ZnAl-GeWnNi 0.8012 0.1 99.9 56.2 0.7995 ZnAl-GeWnU 1.2 98.8 17.2 0.8018 ZnAl-NO3 0.1 0.8021 99.9 17.8 K 8 GeW u O 39 " Reaction conditions: Acetic acid/n-Butanol/Acetophenone (solvent) = 0.05/0.05/0.25 (mol) Reaction temperature: 9(fC, reaction time: 5 h; * Catalysts were pretreated in N2for 2 h at 120"C before the reaction.
Figure 17 - Catalytic activity of ZnAl-XWuCo 1, ZnAl-GeW,,Co 2, ZnAl-BW,,Co 3, ZnAl-CoWuCo 4, HY5, GeWnCo 6, ZnAl-NO3 7, Without catalyst. The catalytic behaviour of POM-LDH for the alkylation of isobutane with butene has been investigated using ZnAl-[SiWuO39(H2O)]8", MgAl-[PW12O39(H2O)]7' and NiAl-[PCoWnO39]5~as catalysts, which were synthesized via both the reconstitution method and the LE method under microwave condition [35, 43]. The data of the alkylation activity is listed in Table 4. It is found from Table 4 that whether it is LDH or POM-LDH, both exhibit obviously high activity for the alkylation of isobutane with butene. For the LDH that are used as catalysts in the reaction, MA1-CO32" is more active than ZnAl-CO32" and MgAl-CO32\ Upon the catalyst of LDH-[SiWnO39(H2O)]8" prepared by the reconstitution method, the conversion of butene is much the same as
Polyoxometalate Complexes of Layered Double Hydroxides
391
upon LDH, but LDH-[SiWnO39(H2O)]8" can remain a stable conversion. It should be noticed that the catalyst of NiAl-[SiWnO39(H2O)]8~ creates a much higher conversion of butene, the C12 and C,6 content than that of the NiAl-[SiWnO39(H2O)]8". This is considered to result from the increase in acid sites of heteropolyacid on the catalyst after the interlayer water is lost. The C12 and C16 content increased are produced from the polymerised butene on the acid sites of interlayer H3PWi2O40. The conversion of butene can be catalysed by both LDH and POM-LDH, which shows that the alkylation reaction can occur on both the base sites of brucite and the acid sites of heteropolyacid. Therefore, it is realized that POM-LDH is a sort of acid base bi-functional catalyst. Table 4 - Alkylation activity of isobutane with butene".
Catalyst
Butene conversicD n (%)
Product s distribut ion (%) ( Mass fra<:tion)*
n
C9-n
C,2
Cifi
66.5
5.0
0.6
0
35.8
60.0
3.6
0.6
0
29.6
65.7
3.6
1.1
0
24.2
63.8
3.6
3.4
0
39.1
57.2
3.1
0.6
0
65.8
1.6
46.6
6.2
37.9
7.7
77.7
2.0
60.0
14.2
21.7
2.0
ZnAl-CO32" MgAl-CO32"
17.2 13.4
27.9
ZnAHSiWuC^CHzO)]8'
17.5
NiAl-CO32" MgAl- [SiWnO39(H2O)]8-
35.6 18.4
NiAl- [PWnO39(H2O)]7(300°C) NiAl-[PCoWnO39]5' (300°C)
oc
Reaction conditions: Reaction temperature: 100"C, Reaction pressure: 2.0MPa; Reaction time: 4 h; Isobutane/Butene 11/1 (mole); Catalyst volume 2.0 cm3 " Catalysts in reactor were pre-treated in N2 at 100"C, for NiAl-[PWI1OM(H2O)]7' (300°C) was at 300°C before reaction. * C7, C9.n, C12 and C,6: Products C number<7, 9-11,12 and 16. c C8° content in Cs were about 40-56% over all catalysts. Liu et al have prepared a POM-LDH system (MnAl-PWuCu) and investigated its performance for the alkylation of isobutane with butene [47]. After the catalyst was activated at 250°C, it was used to catalyse the alkylation reaction, from which a stable butene conversion (62%) was reached. The product compositions are listed in Table 4. It is seen from Table 4 that the C8° is the main proportion and the C9+ is less in the product composition. This indicates that the alkylation reaction on MnAl-PWuCu produces C8° component mainly, while it forms less alkenes and the higher hydrocarbon components. The isopropyl alcohol reaction in which isopropyl alcohol is dehydrated to form alkene on the acid centre and dehydrogenated to produce ketone on the base centre, is and effective probe reaction to examine the acid and base property of catalyst. The reaction equations are as follows.
392
Changwen Hu and Danfeng Li
Acidcat
CH3-CH-CH3 AIT UH
-
CH 3 <:H=CH 2 + H 2 O
-
CHj-f-CHj O
Base cat. '
+H2
For the isopropyl alcohol reaction, the catalytic activity of ZnAl-PWyZ (Z = Cu2+, Ni2+, Co2+) has been investigated [55]. These three catalysts had been pre-treated by calcining at 250 °C for 2 h (ZnAl-PWnNi and ZnAl-PWnCo) and for 4 h (ZnAlPWnCu). Yields of propylene and propanone over these catalysts are listed in Table 5. The same conclusion as ZnAl-SiW n Z is drawn from Table 5 that the ZnAl-P WnZ is mainly acidic catalyst with less base property in the isopropyl alcohol reaction. It is remarkable that ZnAl-PW n Co has a much higher activity than that of ZnAl-PWnNi and ZnAl-PWnCu. This difference in the activity is related to the different kinds of heteroatom Z in the PZW U . Table 5 - Yield of two products over different catalysts . Sample ZnAl-PWnNi ZnAl-PW,,Co ZnAl-PWuCu
Propylene (%) 36.1 91.1 50.1
Propanone (%) 1.1 0.0 3.9
* Reaction temperature: 200°C; Reaction time: I h Recently, Pinnavaia et al have used the synthesized film of ZnAl[NaP5W3oOi10]14" in the reaction of H2O2 oxidation of cyclohexene, and the turnover numbers obtained were 10-182 [2]. In addition, Dutta et al have reported the reaction of selective oxidation of o-xylene into o-tolualdehyde and found that the Li2Al-[V2O7]4" catalyst has a certain activity [56]. Min et al have studied the reaction of dioxygen oxidation of cyclohexene on the catalyst LDH pillared by Keggin [XWUZ]"" ion, i.e., ZnAl-[XWnZ]"", in a liquid phase. They thought that Z in the interlayer [XWHZ]"~ ions was the active sites [29]. Watanabe et al also reported the epoxidation of alkenes on the POM-LDHs [57]. Some typical oxidative reaction using POM-LDH as catalysts are listed on Table 6. The adsorption property and porosity of the POM-LDHs, which are effective on the reactivity of the catalysts, have been tested and summarized recently [11, 23]. BET specific surface area, pore volume, and median diameter of the pores for SiWnZ (Z = Co, Cu, Ni)-LDHs are shown in Table 7 as an example. From Table 8 it can be seen clearly that these kinds of POM-LDHs materials have similar BET specific surface area and the microporous structure. In view of the nitrogen adsorption/desorption isotherm (77 K) under the low relative pressure, the lagging circle begins to appear at p/po = 0.42, suggesting the existence of some mesopores in these compounds. The above results are very similar to those of Pinnavaia's for ZnAl-[V10O28]6" [27].
Polyoxometalate Complexes of Layered Double Hydroxides
393
Table 6 - Oxidation-catalytic activities of some POM-LDHs. Oxidative catalyst Li 2 Al-[V 2 O 7 ] 4 "
Zn 2 Al-[XW n Z] n "(X=Si, B, Ge; Z=Co 2+ , Cu 2+ , Ni 2+ )Zn 2 Al-[SiW 9 Z 3 ] 10 " (Z= Co 2+ , Cu 2+ ) Co(Ni)Al-[XW u Z] n ", (X= Si, B, Ge; Z= Co 2+ , Cu 2+ , Ni 2+ ) MgAl-[V 2 O 7 ] 4 ", MgAl[V 10 O 28 ] 6 Zn(Mg) 2 Al-[Mo 7 O 24 ] 6 ", MgAl-[W 7 O 24 ] 6 \ Zn 2 Al-
[H 2 W, 2 O 40 f ZnAl-[NaP5W3o0110]14" thin film ZnAl-[XWnZ]n'(X=P, Si; Z= Mn2+, Fe3+,Co2+, Cu2+, Ni2+) Zn(Ni)Al[SiW n O 39 (H 2 O)] 8 \ ZnAl-[SiWuO39(H2O)]8"
Reaction Selective oxidation of otolualdehyde from oxylene " Peroxide oxidation of benzaldehyde to benzoic acid" Peroxide oxidation of acetaldehyde to acetic acid" Oxidative dehydrogenation of propane * Peroxide oxidation of cyclohexene " Peroxide oxidation of cyclohexene " Oxygen epoxidization of cyclohexene " Epoxidization of alkenes"
Activity Having a certain activity
Ref. 2
Conversion of benzaldehyde, 16%80%, higher than that of POM Generation of acetic acid, 70-236 (mmoKg-cat.y'h1) Conversion of propane, 23-28%, (530°C) Conversion is 4-6 times more than that of pure POM Turnover number, 10-182 Activity is higher than precursor, Z is the active site Turnover number, 0.88-31
31, 53, 58-60 50
61
62
29 2
57
" Liquid-solid system, 'Gas-solid system Table 7 - BET specific surface area, pore volume and median pore diameter for nZ (Z=Co2+, Cu2+, Ni 2 >LDHs. POM-LDH BET specific Pore volumexlO" Median diameter of Vml-g 1 surface area/m2-g"' the pore/nm 8.24 20.00 9.71 ZnAl-SiW,, 22.04 11.21 7.87 ZnAl-SiWnCo 6.24 7.02 ZnAl-SiW n Cu 3.56 ZnAl-SiW,,Ni 12.95 5.59 9.26 8.05 31.13 ZnAl-NO3 15.46 4.2. - Photocatalysis Recently, dispersed solid particles of TiO2 have been extensively used as an efficient photocatalyst in the chemical decomposition of many organic substrates [6365]. It generally leads to total mineralization of organics into CO2. But TiO2 ultra fine particles are easy to lose and it is difficult to separate them from the reaction systems due to the formation of a milky dispersion of the TiO2 particles. POM-LDHs are generally used as oxidative and redox catalysts in the reported
394
Changwen Hu and Danfeng Li
references. However, their ability to undergo photo-induced multielectron transfer without changing structures is attractive for their use as photocatalysts in the oxidation of organic substrates [58]. The photoactivities of POM-LDHs originate from the unique structure of POMs [23]. They have a number of features in common with semiconductor metal oxide clusters and can be considered as the analogs of the latter [65]. It is often stated that the photochemistry of POM can be regarded as a model for the photochemical processes on semiconductor metal-oxide surfaces [66]. Both classes of materials are constituted by d° transition metal and oxide ions and exhibit similar electronic attributes including well-defined HOMO-LUMO gaps (band gaps). The "gaps" inhibit the recombination of electrons and holes that are generated by the irradiation of the surface of the photo-catalysts with the light energy higher than the band gaps. The electrons and holes thus formed are capable of initiating chemical reactions because of the formation of OH radicals resulting from the subsequent reaction of holes and OH" groups coming from water. Photocatalytic degradation of the organic pollutants by POMs such as [W7O24]6~, [W10O32]4~, [PW,iO39(H2O)]7", [SiWnO39(H2O)]8" and [P2Wi8O62]6" in homogeneous systems has been reported to be very effective [67,68]. However, it is difficult to separate them from the reaction systems for recycling. Therefore, synthesis of the insoluble POM-pillared compounds and using them as heterogeneous catalysts are of fundamental and practical significance. In the previous references, Pinnavaia et al have studied photo-oxidative dehydrogenation of isopropanol on the ZnAl-[V10O28]6", ZnAl-[BVWnO40]6, ZnAl[SiV3W904o]7" and ZnAl-[H2W1204o]6\ and the turnover numbers obtained were 4-10 [17,26]. Studies on the SM-LDHs pillared by polyoxotungstates are relatively more active than those on the SM-LDHs pillared by polyoxomolybdates or polyoxovanadates because polyoxotungstates are more stable than polyoxomolybdates or polyoxovanadates. More recently, we have used a series of synthesized POM-LDHs , namely, [W 7 O 24 f, [SiWnO39Z(H2O)]6" (Z=Cu2+, Co2+ and Ni2+), [P2W18O62f, [P2WnO61Mn(H2O)]8" and [NaP5W3o0110]14" on the organic pollutantsHexachlorocyclohexane (HCH) and pentachloroenitrobenzene (PCB) photodegradation reaction, and found those materials have high reactivity on the reactions [11]. As is well known, the two pollutants as organocholorine hydrocarbons threaten the health of human beings seriously and cannot be totally degraded by the microbiological method. Scientists have conducted a lot of practical work in order to eliminate the pollutants in water and create a green environment. By using the POMLDHs as the photocatalysts, the conversion of HCH into CO2 and HC1 was 30%-80% by irradiation of HCH slurry in the near-UV region for 4 h, and HCH can be totally mineralized by controlling the reaction conditions (see Figure 18). Figure 19 has shown the photodegradation of HCH in presence of Mg2Al[W 7 O 24 ] 6 \ Different photoactivities observed may be caused by the following reasons: (i) different amounts of the interior active sites (the bonds of W-O-W) existing in the different POM-LDH intercalates and (ii) different specific surface areas obtained for the different POM-LDH intercalates (see Table 8). Usually, a higher crystallinity or larger particle size results in a lower specific surface area. As for the LDH precursors, the photocatalytic activity was hardly observed, indicating that the photoactivity of the POM-LDHs originates from the interlayer POM ions rather than LDH precursor. In addition, it is very easy to separate the POM-LDH catalyst from the reaction systems indicating the catalysts are recyclable.
Polyoxometalate Complexes of Layered Double Hydroxides
395
Figure - 18 Activity patterns for the photocatalytic degradation ofHCH on the POMLDHs; [HCHJ0 =10 mg/L, [POM]= 0.5 mmol/L, irradiating time 4 h.
Figure 19 - Photodegradation ofHCH and formation ofCl ion with photocatalysis time in presence of Mg2Al-[W7O24]6'. [HCH]0=10mg/L, [HCH], and [Cl]~, represent the concentration ofHCH and CT ion at elapsed time of reaction, respectively. According to our studies, the model for the photocatalytic reaction on the POM-LDHs are proposed and shown in Figure 20. When the pillared compound is irradiated by light energy near the UV area, the photons penetrate the sheets. Subsequently, the interlayer POM molecules absorb the light energy and are photoexcited to their charge-transfer excited state (POM-LDH)*. At the same time, the
396
Changwen Hu and Danfeng Li
reaction molecules (HCH) diffuse into the interlayer space and are photoactivated by the (POM-LDH)*, and then are degraded into inorganic product molecules (CO2 and HC1) which subsequently escape from the interlayer space through diffusion. Table 8 - Chemical compositions, BET surface area and average particle size for the POM-LDH intercalates. Proposed formula BET surface Average particle area (m2/g) size (nm) Mg! 8Al(OH)6(W7O24)0.18-0.7 H2O 43.5 100 Zn2.9Cr(OH)8[SiW11039Ni(H20)]o.i9-0.2H20 98.1 20 ZriLgAKOHMSiWuOj^o 13-0.2 H2O 9.26 50 Zn2.8Cr(OH)8[SiWii039Mn(H20)]o.i9-0.2 H2O 93.5 45 Zn, 8Al(OH)6[P2W17MnO61(H2O)]014-0.7 H2O 34.0 100 Mg1.8Al(OH)6(NaP5W3o011o)o.o72-0.2H20 66.7 150 Zn2 oAl(OH)6(NaP5W3o0110)o 075-0.2 H2O 65.2 145
Figure 20 - Model ofphotocatalytic degradation of aqueous HCH on the POM-LDH. Our studies show that the disappearance of the HCH follows LangmuirHinselwood first order kinetics. From the results of GC-MS and IC, the intermediates produced during the process of photocatalysis were iden- phototified and detected. That is, phenylhexachloride, polyphenol, tetrahydroxybenzoquinone and benzoquinone were detected by GC-MS; acetic acid and formic acid were detected by IC. According to the distribution of the intermediates, the reaction mechanism for the photocatalytic degradation of aqueous HCH on the POM-LDH intercalate is proposed as follows: As a heterogeneous photocatalyst, the interlayer POM has the double aptitudes of adsorption of the reactants on the catalysts' surface and absorption of the photon with the light energy higher than its band gap. The photocatalytic reaction occurs in the photointerlayer space of the intercalates and the mode of activation of the catalyst is photonic activation by exciting the POM-LDH with suitable light energy, and then
Polyoxometalate Complexes of Layered Double Hydroxides
397
forms the excited species (POM-LDH)*. By using the semi-conductor notation, the excitation of the POM—LDH at the OMCT band can be presented by Equation 1: POM-LDH + to- <-> POM-LDH (e"+ h+)
(Eq. 1)
In principle, h causes oxidations and e" reductions. POM-LDH(e"+ h* ) has a more powerful redox ability by forming electron-hole pairs and their recombination is inhibited by the " g a p " of the POM. The reaction of the photoholes with the surface hydroxyl groups occurs and OH" radicals are generated (Eq. 2): POM-LDH (e + h' ) + H2O <- POM-LDH(e") + OH + H+
(Eq. 2)
The OH radicals are strong and unselective oxidant species in favor of totally oxidative degradation and mineralization of organic substrates such as HCH into inorganic products (CO2 and HC1). Then the products escape from the interlayer space through diffusion. For the degradation of HCH by the POM-LDH, the following general scheme (see Scheme 1) has been proposed according to the above studies. The overall reaction can be shown as follows:
C 6 H 6 Cl f i +6Oi
? 6CO, +6HC1
~ POMs-LDH
Accordingly, a pathway of photocatalytic degradation of aqueous HCH is shown on Scheme 1.
Scheme 1 - Pathway of photocatalytic degradation of aqueous HCH. As shown in Table 9, the particle sizes for the POM-LDH are 20-150 run as estimated by TEM, so they can be considered as the supermolecular layered nanometer materials. When the reactant molecules enter the space, they contact and react with the interlayer POM anions that have been photoactivated by the light energy in near UV area (expressed as POM*). Finally, the product molecules thus formed desorb from the POM* active sites, and then escape from the interlayer space through diffusion and enter solution (Eq. 2). This result has both theoretical and practical significance [40].
398
Changwen Hu and Danfeng Li
We have also synthesized Zn/Al/W(Mn) mixed oxides by calcinating these POM-LDHs precursors at 600-700°C, and used them on the HCH photodegradation reaction [12]. Compared with the POM-LDH precursors, the mixed oxides thus obtained exhibited higher photocatalytic activity, as is shown in Figure 21. The explanation for the results of Figure 21 is as follows. Monolacunary or -substituted Keggintype POM often has lower photocatalytic activity than that of its saturated unit [18]. However, the photocatalytic activity of SiWn and SiWuMn is improved when they are intercalated into the interlayer. The interlayer space provides the reaction place and easier accesses to the active sites and shorter difusion pathway. The ZnO- and AI2O3- like amorphous phases resulted in lower photocatalytic activity than their precursors. ZnO together with ZnAl2O4 mixed phases, pure ZnWO4 and Zn/Al/W(Mn) mixed oxides exhibited relatively high photocatalytic activities to the photodegradation of HCH. Transition metal oxides (e.g.TiO2,WO3, ZrO2 and ZnO), transition metaloxygen clusters (e.g. Keggin units), and the mixed oxides (e.g., ZnAl- SiWn- MO and ZnAl- SiW n Mn- MO) are constituted by d0 transition- metals and oxide ions, and they can be photoexcited and formed oxygen to metal charge transfer (OMCT) excited state with considerable oxidizing capacity. Therefore, HCH molecules were photoactivated by the excited metal oxides, metal- oxygen clusters or the mixed oxides, and then photooxidized into CO2 and H2O.
Figure 21 - Photocatalytic activity of various photocatalytic materials to the degradation of an aqueous HCH (0.14 mM). The content of each active component in the solids or aqueous solutions was ca. 0.015 mM. Turnover number was ca. 9.3. Irradiation time was 4 h. Conversion (%) = (l-[HCH]J[HCH]0)*100, where [HCH], and [HCH]0 represents t time and initial concentration of HCH, respectively. The UV absorption characteristics of the above photocatalytic materials are as follows: (i) pure SiW n Mn and SiW u Mn-LDHs have similar OMCT bands, i.e., 250260 run; (ii) pure ZnO and ZnO together with ZnAl2O4 mixed phases have UV absorption in the range of 200- 400 nm; (iii) pure ZnWO4 has UV absorption in the
Polyoxometalate Complexes of Layered Double Hydroxides
399
range of 200- 400 nm; and (iv) both of ZnAl- S i W n - MO and ZnAl- SiW n MnMO have the maximum UV absorption at 310- 320 nm (mainly originated from ZnWO4 phase). In general, lower band gaps (higher OMCT band) result in higher photocatalytic efficiency due to easier transition from the ground state to the excited state. In addition, the photocatalytic activity is also related to the surface physicochemical properties and the porous structures of the mixed oxides (Figure 22). The lowest photocatalytic activity of the X-ray amorphous phases is due to destruction of the layered structure, decomposition of SiW n or SiWnMn, and disappearance of the OMCT band. In addition, the content of ZnO in the amorphous phases is very low although it is an excellent photocatalyst. In the cases of as-synthesized Zn/Al/W(Mn) mixed oxides, their predominant phase had a ZnWO4 inverse spinel structure, while ZnWO4 itself is an excellent photoactive material. At the same time, the mesoporous structure of the composite allowed the reaction to take place in the pores, therefore, the host-guest interaction increased and the photocatalytic activity was improved compared with the precursor ZnAl- SiW n (Mn). The photocatalytic activity of the mixed phases of ZnO and ZnAl2O4 mainly originated from the photoactive ZnO (pure ZnO and ZnO together with ZnAl2O4 mixed phase had similar activities to photodegrade HCH). The reason for producing higher photocatalytic activity of ZnAl- SiW u Mn- MO than that of ZnAl- SiWn- MO is unclear now, and further research will explain it. In addition, separation and recovery of the photocatalytic is very easy.
Figure 22 - BET specific surface area, pore size and conversion evolutions with temperatures for ZnAl- SiW,, and ZnAl- SiW,,Mn (m). The physicochemical properties of the products were characterized by the methods of powder XRD, elemental analysis, scanning electron microscopy, UV diffusion reflectance spectroscopy, infrared spectroscopy, thermogravimetric analysis, electron spin resonance and N2 adsorption- desorption measurements. Phase changes during the process of calcination of the POM containing LDH precursors were studied, indicating that the precursors decomposed at 400°C and formed ZnO and Al2O3-like
400
Changwen Hu and Danfeng Li
phases together with an X-ray amorphous phase, and then yielded a mixture of wellcrystallized phase of ZnAl2O4 formal spinel and ZnWO4 inverse spinel at 600- 700°C. As for the thermally decomposed products, the changes of BET surface areas and pore sizes are related to their crystallinity and structures. 5 - Concluding remarks and expectation The newest research progress in the compositions, structures, synthesis methods and photocatalytic and oxidation-catalytic functions of the POM-LDHs complexes are reviewed from the viewpoint of host-guest chemistry, and the relationships among compositions, structures and property of POM-LDHs are expounded. We predict that new types of POM-LDHs materials with larger basal spacings will be obtained continuously with the appearance of new synthesis routes and characteristic methods. At the same time, the analyses of the structures will be made more deeply. By using the preparation technology of alternate adsorption film, those POM-LDHs can be made into ultra thin film materials. The research on their light, electronic, magnetic and catalytic properties in the state of nanometer ultra thin film will be one of the newest subjects. In addition, it is another active field for POM-LDHs to adsorb and separate organic and inorganic molecules. We hope the progresses in studying POM-LDHs will be made continuously in the cross-field of whole supermolecular-chemistry and solid chemistry. We predict that new types of POMLDHs materials with larger basal spacings will be obtained continuously with the appearance of new synthesis routes and characteristic methods. 6 - Reference [I] C.W. Hu, Q.L. He and E.B. Wang, Prog. Natur. Sci, 6, 5 (1996) 524. [2] A.G., Elizabeth, S.K., Yun and T.J., Pinnavaia, Appl. Clay. Sci., 13 (1998) 479. [3] V. Rives and M. A. Ulibarri, Cood. Chem. Rev., 181 (1999) 61. [4] C.L., Hill and CM., Prosser-McCartha, Cood. Chem. Rev., 143 (1995) 407. [5] M. Nyman and F. Bonhomme, T. Alam, Science, 297 (2002) 996. [6] I.V., Kozhevnikov, Chem. Rev., 98 (1998) 171. [7] F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 11 (1991) 173. [8] J. Twu and P. K. Dutta, J. Phys. Chem., 93 (1989) 7863. [9] C.W. Hu, Y.Y. Liu, Z.P. Wang et al, Sci. in China, Ser. B, 25, 9 (1995) 916. [10] E. Narita, P.D. Kaviratna and T.J., Pinnavaia, Chem. Lett., (1991)805 . [II] Y.H. Guo, D.F. Li, C.W. Hu, Y.H. Wang and E.B. Wang, Inter. J. Inorg. Mater., 3 (2001) 347. [12] Y.H. Guo, D.F. Li, C.W. Hu, E. Wang, Y.C. Zou, H.Ding and S.H.Feng, Microp. Mesop. Mater., 56 (2002) 153. [13] C.W. Hu, Y.Y. Liu, Z.P. Wang et al, Chin, J. Catal., 16, 4 (1995) 255. [14] C.W. Hu, Q.L. He, Y.F. Zhang, E.B. Wang et al, J. Chem. Soc, Chem. Commun., (1996) 121. [15] C.W. Hu, Q.L. He, Y.F. Zhang and E.B. Wang, Chin. Chem. Lett., 6, 7 (1995) 609. [16] M.A. Drezdzon, Inorg. Chem., 27 (1988) 4628. [17] T. Kwon, G.A. Tsigdinos and T.J. Pinnavaia, J. Am. Chem. Soc, 110 (1988) 3653. [18] E.D. Dimotakis and T.J. Pinnavaia, Inorg. Chem., 29 (1990) 2393. [19] Q.L. He, H. Zhen, C.W. Hu, Y.F. Zhang and E.B. Wang, Prog. Natur. Sci., 7 (1997) 229. [20] J.Y. Zhang, X.W. Yu, S. Zhang, C.W. Hu, Z.J. Liang and E.B. Wang, Chin. J. Mol.
Polyoxometalate Complexes of Layered Double Hydroxides
401
Catal. (in Chinese), 9 (1995) 241. [21] C.W. Hu, Q.L. He, T.D. Tang and E.B. Wang, Acta Physico-Chimica Sinica, 11 (1995) 193. [22] E.L. Salinas, PS. Castillo and Y. Ono, Mater. Res. Soc. Symp. Proc, 371 (1995) 163. [23] E.B. Wang, C.W. Hu and L. Xu, Introduction of the polyoxometalate chemistry (in Chinese), Beijing: Beijing Chemical Industry Press, 1998. [24] Q.L. He, C.W. Hu, Y.H. Zhang, et al, J. Northeastern. Norm. Univ., 3 (1996) 122. [25] R.G. Finke, M.V. Droege and P.J. Domaille, Inorg. Chem., 35 (1987) 3886. [26] T. Kwon and T.J. Pinnavaia, J. Mol. Catal., 74 (1992) 23. [27] S.K. Yun and T.J. Pinnavaia, Inorg. Chem., 35 (1996) 6853. [28] J. Evans, M. Pillinger and J. Zhang, J. Chem. Soc, Dalton Trans, 14 (1996) 2963. [29] J. Guo, Q.Z. Jiao and E.Z. Min, Catal. Lett., 40 (1996) 43. [30] C.W. Hu, D.F. Li and Y.H. Guo, Chin. Sci. Bull., 46, 13 (2001) 1061. [31] C.W. Hu, E.B. Wang, M. Misono et al, Catal. Today, 30 (1996) 141. [32] Y.H. Guo, D.F. Li and C.W. Hu, Appl. Catal. B, 30 (2001) 337. [33] C.W. Hu, Q.L. He and E.B. Wang, Prog. Natur. Sci., 6, 5 (1996) 524. [34] K. Chibwe and W. Jones, Chem. Mater., 1 (1989) 489. [35] Z.Xu, H.M. He, D.Z. Jiang and Y. Wu, Acta Physico-Chimica Sinica, 10 (1994) 6. [36] M.A. Ulibarri, F.M. Labajos, V. Rives, R. Trujillano, W. Kaguny and W. Jones, Inorg. Chem., 33 (1994) 2592. [37] E. Narita, P.D. Kaviratna and T.J., Pinnavaia, J. Chem. Soc, Chem. Commun., (1993) 60 . [38] H.H. He, T. Sun and D.Z. Jiang, Acta Physico-chimica Sinica, 10 (1994) 272. [39] E.L. Crepaldi, PC. Pavan and J.B. Valim, J. Chem. Soc, Chem. Commun., (1999) 155. [40] T. Kwon and T.J. Pinnavaia, Chem. Mater., 1 (1989) 381. [41] J. Wang, Y. Tian and R. Wang, Chem. Mater., 4 (1992) 1276. [42] S. Kasztelan and J.B. Moffat, J. Catal., 106 (1987) 512. [43] Z. Xu, Y. Wu, H.H. He and D.Z. Jiang, Chin. Sci. Bull, 39 (1994) 1392. [44] F. Kodi, V. Rives and M.A. Ulibarri, Inorg. Chem., 34 (1995) 5122. [45] J. Guo, T. Sun, J. Shen, D. Liu, D.Z. Jiang and E.Z. Min, Chem. J. Chin. Univ., 16 (1995) 346. [46] C. W. Hu, X. Zhang, Q.L. He and E.B. Wang, Trans. Met. Chem., 22 (1997) 197. [47] J. Liu, T. Sun and D.Z. Jiang, Chin. J. Catal. (in Chinese), 16 (1995) 241. [48] C.W. Hu, Y.Y. Liu, Z.P. Wang, J.Y. Zhang and E.B. Wang, Chin. J. Mol. Catal. (in Chinese), 16(1995)255. [49] J. Guo, T. Sun, J. Shen and D.Z. Jiang, Chem. J. Chin. Univ., 16 (1995) 512. [50] C.W. Hu, X. Zhang and L. Xu, Appl. Clay. Sci., 13 (1998) 495. [51] M.R. Weir, J. Moore and R.A. Kydd, Chem. Mater., 9 (1997) 1686. [52] Y.Y. Liu, C.W. Hu, Z.P. Wang, J. Zhang and E.B. Wang, Sci. in China, Ser. B, 39 (1996) 86. [53] C.W. Hu, Y.Y. Liu, Q.L. He, J.Y. Zhang and E.B. Wang, Chin. J. Mol. Catal. (in Chinese), 17(1996)237. [54] C.W. Hu, M. Hashimoto, T. Okuhara and M. Misono, J. Catal., 143 (1993) 437. [55] J.G. Liu, Q.Z. Jiao, P. Zhang, J. Guo and D.Z. Jiang, Chem. J. Chin. Univ., 15 (1994)275. [56] P. K. Dutta and J. Twu, J. Phys. Chem., 93 (1989) 7864.
402
Changwen Hu and Danfeng Li
[57] Y. Watanabe, K. Yamamoto and T. Tatsumi, J. Mol. Catal., Ser. A, General, 145 (1999)281. [58] C. W. Hu, Q.L. He and E.B. Wang, Chin. J. Mol. Catal. (in Chinese), 10, 1 (1996) 67. [59] Q.L. He, C.W. Hu and E.B. Wang, Chin. Sci. Bull., 42 (1997) 914. [60] L. Xu, C.W. Hu and E.B. Wang, J. Natur. Gas. Com., 2 (1997) 155. [61] K. Bahranowski, G. Bueno and K. Wcislo, Appl. Catal. A, 185 (1999) 65. [62] E.A. Gardner, S.K. Yun and T.J. Pinnavaia, Appl. Catal, Ser. A, 167 (1998) 65. [63] J.M. Herrmann, Catal. Today, 53 (1999) 115. [64] I. Ilisze, Z. Laszlo and A Dombi, Appl Catal A, 180 (1999) 25. [65] Y.H. Guo, Y.H. Wang, C.W. Hu, Y.H. Wang, E.B. Wang, Y.C. Zhou and S.H. Feng, Chem. Mater., 12, 11 (2000) 3501. [66] C.W. Hu, B. Yue and T. Yamase, Appl Catal A, 194-195 (1999) 99. [67] R.F. Renneke, M. Kadkhodayan, M. Pasquali and C.L. Hill. J. Am. Chem. Soc, 113(1991)8357. [68] G. Chester, M. Anderson and H. Read, J Photochem Photobiol A, 71 (1993) 291.
CATIONIC AND ANIONIC CLAYS FOR BIOLOGICAL APPLICATIONS
JIN-HO CHOY* and MAN PARK
National Nanohybrid Materials Laboratory (NNML) School of Chemistry and Molecular Engineering Seoul National University, Seoul, 151-747 - KOREA * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
404
J. Choy and M. Park
1 - Introduction Clay minerals have attracted a great deal of attention from a very wide range of scientific and industrial fields because both of natural abundance and of exhaustless and unlimited potentials [1-7]. Incessant researches have been conducted on the fundamental properties like basic structures, formation and weathering of framework structure, swelling (hydration), ion exchange capacity, etc. all of which leads to elucidation of their important roles in many academic fields like agriculture, biology, geology, and inorganic and environmental sciences [8-16]. Of course, they are extensively utilized in innumerable industrial fields like ceramics, paper, paint, plastics, drilling fluids, foundry bondants, chemical carriers, liquid barriers, decolorization, catalysis, adsorption, and so on. In particular, their biological applications for curative and protective purposes are as old as mankind itself, which clearly assures biocompatibility. To our surprise, there is still the increasing interest in further development of their applications for biological purposes, especially for pharmaceutical, cosmetic, and even medical purposes. A variety of clay minerals are found in nature. In general, clay minerals could be roughly divided into three classes by ion exchange property [17]. These are nonionic, cationic, and anionic clays. Non-ionic clays with no true ion exchange capacity include kaolinite, serpentine, chlorite, illite, pyrophyllite, and talc. Cationic clays with cation exchange capacity comprise many alumino-silicate clays such as vermiculites, smectites, and swelling micas. And the last class with anionic exchange capacity is represented by layered double hydroxides. The former two classes of clay minerals are widespread in nature, but fairly difficult to be synthesized. Whereas, the last class is rare in nature, but easily and inexpensively synthesized. Although crystal structure is a major criterion of the detailed classification into a number of groups and subgroups, all their structures are based on two-dimensional framework in common. And for the most part, they are usually hydrated or to be easily hydrated, which leads to one of the important features along with ion exchange properties. In fact, these two properties are most intensively studied, and also play a crucial role in biological applications, the key topic in this chapter. Therefore, this chapter mainly deals with biological applications of two classes of clays, cationic and anionic ones, describing their structure, physical and chemical properties, and preparation. 2 -Cationic clays Biological applications of cationic clays have been found in many fields such as pharmaceutics, medicine, cosmetics, spas, food, fodder, and pesticide into which cationic clays are directly involved as active principles or excipients [16, 18-20]. In these applications, their important roles are originated from the physical and chemical properties such as high adsorption capacity, plastic (soft) framework, chemical inertness and low or null toxicity. Although a few synthetic clays like laponite are employed in some cases, natural clays are generally preferred partially due to favorable concept of 'natural' and partially due to difficulty and high cost of their syntheses. 2.1 - Structure and classification Negative charge of cationic clay minerals could be generated from two different mechanisms, namely, isomorphic substitution and breakage of oxygen bridge [17. 21-22]. The former develops permanent negative charge in framework that is not
Cationic and Anionic Clays for Biological Applications
405
affected by bulk environments like pH and particle size, whereas the latter leads to temporary charge (or pH-dependent charge) derived from deprotonation of hydroxyl groups on cleaved edge or defect site of layer framework. The clay minerals possessing only temporary charge could belong to non-ionic clays rather than cationic clays because of their fundamental structure being non-ionic. On the other hand, the clay minerals like illite and muscovite could be also described as non-ionic clays due to lack of cation exchange property although they have isomorphic substitution in their framework. Thus, cationic clays considered here are arbitrarily confined to those, which allow their interlayer cations to be easily exchanged by other cationic species in solution, which make it possible for us to extract their common features such as micatype 2:1 structure, exothermic hydration, high internal surface area, high cation exchange capacity, some degree of cation selectivity, and typically interlayer displacement. Fundamental crystal structure of cationic clays is very similar to framework of mica. Unit layer consists of one octahedral sheet sandwiched between two tetrahedral sheets, as shown in Figure 1 [2-6].
Figure 1 - Crystal structure of 2:1 aluminosilicate clay The cations in the tetrahedral sheet are typically Si 4t and Al3+, while those in the octahedral sheet are Al3+, Fe3+, Mg2+ Fe2+ and etc. Because cations of both tetrahedral (Si4+) and/or octahedral (Al3+, Fe3+, Mg2+) sheets could be isomorphically substituted by lower valence cations, layer framework develops permanent negative charge that is usually delocalized over layer surface. Net negative charge is electrostatically neutralized by the cations in interlayer space between layers. And these interlayer cations prefer to be hydrated for thermodynamic reasons. In general, degree of net negative charge is evaluated by cation exchange capacity expressed in centiequivalent of exchangeable cation per kilogram of sample [17]. The cationic clays could be divided into several groups by charge density, as demonstrated in Table 1. Each group could be further separated by various criteria like octahedral occupancy, cation species of octahedral site and charge site, as described in Table 2 [17, 21-22].
406
J. Choy and M. Park
Table 1 - Classification of the 2:1 clays Layer charge per unit cell
Group name
Octahedral occupancy
Species
0.0
Pyrophillite
Dioctahedral
Pyrophillite
Talc
Trioctahedral
0.5-1.2
Smectite
Dioctahedral Trioctahedral
Talc Montmorillonite, beidellite Hectorite, saponite
1.2-1.8
Vermiculite
Dioctahedral
Dioctahedral vermiculite
Trioctahedral
Trioctahedral vermiculite
Dioctahedral
Muscovite, paragonite
Trioctahedral
Phlogopite, biotite
=2.0 s4.0
Mica Brittle mica
Variable
Chlorite
Dioctahedral
Margarite
Trioctahedral
Clintonite
Dioctahedral
Donbassite
Trioctahedral
Clinochlore
Table 2 - Subgroups of Smectite _ , Subgroup
_, Charge site
Ideal composition Jl TT Mint(Moct)(Mteta)O1(OH)m.nH 2O
Montmorillonite
Octahedral
Mx(Al2.xMgx)(Si4)O10(OH)2.nH2O
Beidellite
Tetrahedral
Mx(Al2)(Si4.xAlx)Oi0(OH)2.nH2O
Nontronite
Tetrahedral
Mx(Fe23+)(Si4.xAlx)O10(OH)2.nH2O
Saponite
Tetrahedral
Mx(Mg3)(Si4_xAlx)O10(OH)2.nH2O
Hectorite
Octahedral
Mx(Mg3.xLix)(Si4)O10(OH)2.nH2O
2.2 - Preparation Cationic clays for biological purposes are prepared mostly from natural sources with a few exceptions. Although there are the great variations in quality requirements for each application, biological applications typically require more strict and higher purity compared with that of other applications. Especially, the clays to be applied for pharmaceutics, medicine, and cosmetics should meet the strict requirements about phase purity, chemical composition, particle size, and kinds of impurity [16, 23]. However, overall quality of natural clays mainly depends on origin of the deposit. General purification methods for natural clays mainly depend on physical and magnetic
Cationic and Anionic Clays for Biological Applications
407
separations including sieving, filtration, and sedimentation. For some specific applications, acidification is also frequently employed. Synthetic cationic clays is available even though their practical applications are currently very limited. General methods for synthesis of clays are based on hydrothermal treatment of precursors at various temperatures under autogenous pressure. Several smectites are successfully synthesized in a high purity, including beidellite, hectorite (the trade name Laponite), saponite, and stevensite. However, most synthetic methods have some disadvantages like long reaction time (often several weeks to even months), high temperature (up to 300-600°C) and relatively low crystallinity [24-25]. 2.3 - Pharmaceutical Applications Cationic clays possess high sorption ability, high internal surface area, high cation exchange capacity and typically interlayer displacement from which most pharmaceutical applications have been benefited [16,26-27]. Traditionally, smectites have been most widely employed as both active principle and excipient for various pharmaceutical purposes [28-29]. As the active principle, major therapeutic effects of smectites are to eliminate the excess water in the feces as antidiarrhoeaics and to protect skin mechanically against the physical or chemical substances generated from both external and internal sources as dermatological protectors. Bentonite-water mixtures are known to be the best materials for spas to alleviate joint pain occurring between flares of joint disease. Local temperature elevation along with ion exchanges gives rise to pain-relieving effects. Their major function derives from high swelling properties and ability to retain large water percentages [28-29]. Slow delivery of heat to the underlying tissues ensures that temperatures up to 50°C can be applied without causing tissue injury [28]. On the other hand, as the excipient, they play an important role in dispersing active principles due to their ability to increase in volume in the presence of water, in buffering abrupt change of acidity, and in stabilizing emulsion, polar gel and suspension because of their colloidal characteristics to avoid the segregation of the pharmaceutical formulation's components. Although there are still innumerous efforts to enhance and expand the roles of clays in these traditional applications, recent attention rapidly shifts to exploration of their new potentials such as drug carrier, protecting matrix, release controlling agent and chemical modifier. These new attempts are dealt in detail here because they are indeed of hot issues in current times. New potentials of cationic clays for pharmaceutical applications arise mostly from recent hybridization of drugs with the clays either in ordered or disordered manner. Traditional applications mainly rely on the inherent physical, rheological, and chemical properties of the clay [16,30]. On the other hand, new potentials from hybridization take advantage of complexation of inorganic and organic compounds that leads to interesting characteristics of organic-inorganic hybrid distinguished from that of each component in simple physical mixtures. Hybridization is reported to offer the fascinating features such as protected delivery, controlled release, enhanced water solubility, and increased dispersion ability. Furthermore, it has a feasibility of target delivery [19,31-33]. In spite of these promising prospects suggested thus far, not so many works appear in literatures to date mainly because hybridization approaches have quite recently attracted intense attention. The recent studies on complexation of drug and clay are rapidly increased. A
408
J. Choy and M. Park
great number of studies have focused on preparation and characterization of surfactantclay hybrids to facilitate the advanced formulations of various drugs [17,19,34]. Polyethylene oxide and AMsopropylacrylamide are the representative organic polymers for the hybrids. On the other hand, only a few studies have been carried out on direct complexation of drug with clays. Ito et al recently found that complexation of indomethacin, an anti-inflammatory and as an analgesic agent, with smectite enhanced its penetration rate through skin [19]. Enhanced permeability was rationalized by the increase in both stability of amorphous indomethacin and water solubility by their complexation with smectite. Lin et al intercalated 5-fluorouracil, effective chemotherapeutic agent for colorectal cancer, into montmorillonite to diminish its severe side effect through its in situ release [17]. Lee and Fu reported that release properties of drugs could be controlled by loading them into nanocomposites of Nisopropylacrylamide and montmorillonite. Release property of loaded drug could be controlled by electrostatic interaction between the drug and nanocomposite. Electrostatic attraction decreases the release ratio, while electrostatic repulsion increases the release ratio. Our laboratory also recently found that hybridization of poorly water-soluble intraconozole, antifungal agent, with smectites led to remarkable improvement of its water solubility and bioavailability (Figure 2). It was suggested that molecular arrangement of intraconozole within nanosized interlayer space of smectites was greatly contributed to these enhancements.
Figure 2 - Effect of clays on pharmacokinetic parameters of itraconazole after oral administration. After magnabrite (Mag) or Montmorillonite (Mmt) was reacted with itraconazole (Itra) at a weight ratio of clay 0.7/Itra 0.3, the resulted solid was coated with either hydroxypropylmethyl cellulose (hpmc) or hydroxypropyl cellulose (hpc) at a weight ratio of solid 0.7/polymer 0.3. The parameters were compared with those of commercial Sporanox (maximum blood concentration (Cmax): 223 ng/ml; time to reach Cmax (Tmax): 1.8 hr).
Cationic and Anionic Clays for Biological Applications
409
Figure 3 - Stabilization of emulsions by solid particles, (a) stabilization by envelopes of particles around the oil droplets, (b) stabilization encapsulation oil in a threedimensional network of particles 2.4 - Cosmetic application Cationic clays have been widely employed as thickener and emulsion stabilizer in cosmetics [35-36]. They are also used as active principles for adsorption of substances such as greases, toxins, etc. and for antiperspiration to give the skin opacity, remove shine and cover blemishes. These applications are mainly based on high cation exchange capacity, excellent swelling property along with remarkable hydration ability, and structural plasticity. Particular attention has been given to the organic polymer-clay nanocomposites in which organic polymers are complexated with the layers of clay by chemical interactions like electrostatic attraction, hydrogen bonding, and Van der Waals interaction. The complexation enables one not only to enhance the above inherent properties but also to impart new functionalities, especially organophilicity of clay, high stability of organic component, and new rheological properties [37]. A recent review by Ray and Okamoto describes the detailed properties of various polymer/layered silicate nanocomposites [38-39]. Depending on the strength of interfacial interactions between the polymer matrix and layered silicate, they classified three different types of nanocomposites, intercalated, intercalated and flocculated, and exfoliated ones (Figure 4).
Figure 4 - Possible polymer/inorganic nanocomposites.
410
J. Choy and M. Park
These unique combinations of hydrophilic inorganic layers and hydrophobic organic polymers at sublattice and molecular level facilitate loading or incorporating chemically diverse active compounds into a broad range of cosmetic formulation. Furthermore, long lasting and high effectiveness could be also achieved by controlling their release property. 2.5 - Agricultural applications to animal feed and pesticides Clay minerals have been incorporated in animal feeds for multiple purposes mainly due to high binding (processing) ability, water sorption capacity, and cation retention capacity [40]. In animal feed, they act as an anti-caking and pelletizing aid in nonmedicated feeds, as an adsorbent of mycotoxins and gastrointestinal gases, and as a consolidating additive of feces [41]. Especially, incessant attention has been given to mycotoxin detoxication by the various clays. Phillips et al analyzed the in vitro binding capacities of different adsorbents, which were representative of the major chemical classes of aluminosilicates [42-43]. They claimed that smectites are the suitable candidate clay for in vivo trials concerning the prevention of aflatoxicosis in chicken because smectites were shown to have a high affinity for aflatoxin B1. The complex of smectite-aflatoxin Bl was stable at temperatures of 25 and 37°C, in a pH range of 2-10, and in an eluotropic series of solvents. A chemically modified montmorillonite also reported to exhibit a high binding capacity for zearalenone of 108 mg/g [44]. The clay modified with cetylpyridinium or hexadecyltrimethylammonium resulted in an increased hydrophobicity of the clay surface that led to a high affinity to the hydrophobic zearalenone. There is also an attempt of virus adsorption to clays, and the viruses most studied include poliovirus, encephalomyocarditis virus and reovirus. Although adsorbed viruses were not deactivated completely, these results suggested a potential of clay additive to give a distinct advantage in prevention of virus infection [20,45]. Application of clays to pesticide formulation is another important subject nowadays because both active ingredient and adjuvants of pesticide formulations cause serious environmental problems [46-48]. It is desperately pursued that a minimum amount of a pesticide exerts the maximum effect at the right moment and place. In order to increase the pesticide efficiency and to reduce their leaching into related environments like air and water, it has been suggested that reversible complexation of active pesticide ingredient with clay minerals would be one of the feasible solutions. However, most of researches have focused on adsorption of pesticides by clay minerals for their removal from water and immobilization in soils. To date, there are only a few attempts to explore the potential of clay minerals as carriers in pesticide formulations. El-Nahhal et al reported that adsorption of alachlor and metolachlor on a bentonite was very efficient when the clay mineral was modified by cation exchange with benzyltrimethyl ammonium (BTMA) ions [49,50]. The raw bentonite only adsorbed 3% of the amount of metolachlor added, whereas bentonite preadsorbed with 0.5 mmol BTMA and 0.8 mmol BTMA adsorbed, 25% and 20% of the amount added, respectively, to suggest a high feasibility for slow release formulation. Carrizosa also suggested that primary alkylammonium saturated clays were suitable as potential sorbents for slow release formulation [51]. These bentazone-modified clay complexes released 20 to 80% of their bentazone content. In addition to slow release formulations, clays also turned out to be very effective in protecting unstable pesticides against
Cationic and Anionic Clays for Biological Applications
411
volatilization and photodegradation that consequently forced farmers to increase frequency and dose of herbicide application. Recently, Lagaly evaluated feasibility of pesticide-clay formulations based on the binding mechanisms between pesticides and clays, and suggested a wide range of pesticides to be stably complexated with clays by direct electrostatic interaction as well as by induced interactions through surface modification of clays [52]. 3 - Anionic clays: layered double hydroxides Anionic clays are typically included into layered double hydroxides (LDHs) because no other layered clays have been found to possess exchangeable anions in their interlayer space. Similar to cationic clays, their interlayer could accommodate a variety of anionic guest components, which leads to a broad spectrum of potential biological applications such as pharmaceutics, medicine, cosmetics, and pesticides [53-57]. However, their practical applications to biological fields are currently very limited compared with those of cationic clays mainly due to shortage of accumulated understanding in their characteristics [58]. The representative commercial product based on LDH is an antacid made of hydrotalcite. As a matter of fact, their potentials for biological applications [59-63] are recently explored in spite of extensive studies on their catalytic activity [64-67] and anion exchange behavior [68]. It is worthy to note the trend that more and more attention has been given to their potentials for a wide range of biological fields. 3.1 -Structure and classification Fundamental structure of LDH is based on hydrotalcite. The structure of hydrotalcites was first elucidated by Allmann for the CO3-Mg,Fe system (pyroaurite and sjogrenite) in 1968 [69] (Figure 5). Depending on the arrangement of an octahedral layer and an interlayer, two polytypes of layer structure are well recognized. The one (hydrotalcite) has a rhombohedral unit cell containing three stacked repeat units, whereas the other (manasseite) has a hexagonal unit cell containing two stacked repeat units [70-71]. These two polytypes exhibit the same local topology of the layerinterlayer bonding, but different in the long-distance layer-layer interactions.
Figure 5 - Layer stacking of LDHs.
412
J. Choy and M. Park
On the other hand, two dimensional lattice structure could be constructed in two distinct types: hydrotalcite-type and hydrocalumite-type (Figure 6). The general formula of hydrotalcite-type LDHs is [M2+i.IM3+;r(OH)2]"(A'"");t/m,«H2O, where the M"+ are metal cations (M2+ = Mg2+, Zn2+, Ni2+, Cu2+, etc., M3+ = Al3+, Fe3+, etc.) and Ara" are interlayer anions (Am" = CO32", NO3", SO42", and other anionic species) [69-71]. The layer structure of LDHs is constructed with a stacking of brucite structure of Mg(OH)2 in which Mg(OH)6 octahedra are connected through edge sharing into 2dimensional sheets with layer thickness of 4.8 A. Some of divalent cations in the brucite layer are substituted by trivalent cations such as Al3+, which develop permanent cationic layer charge (Figure 7).
Figure 6 - Layer structures of hydrotalcite (a) and hydrocalumite (b)
Figure 7 - The structure of LDHs (a) brucite layer (b) LDH layer. They have a wide range of chemical compositions. In addition, their layered structures exhibit a variety of stacking faults to generate many different polytypes of crystals. On the other hand, Ca-containing LDHs exhibit hydrocalumite structure with corrugated brucite-like main layers [72]. Ca atoms are hepta-coordinated with six hydroxides and one interlayered water, which are edge-shared with octahedral trivalent metal cations. The framework layer is formed in an almost fixed molar ratio of 2 Ca2+/M3+, which results in the general formula of [Ca2M(OH)6]+ A" mH2O. Also, the kinds of M3+ of hydroxide layer is very limited, typically Fe3+ and Al3+. This unique
Cationic and Anionic Clays for Biological Applications
413
layer framework results in the distinct XRD pattern distinguished from the other usual LDHs, usually leading to better crystallinity. The interlayer space of LDHs is occupied by charge-balancing anions that are typically bound to the layer through hydrogen bonding with water molecules. Exchangeability of interlayered anions depends on thenelectrostatic interaction with positively charged layer. Except for carbonate ion, most organic and inorganic anions are known to be exchangeable. Thus LDHs are widely applicable to various supramolecular structure or heterogeneous hybrid systems. Another unique group of LDHs are layered hydroxide salts with anion exchange capacity [73-77]. The layered hydroxide salts can be classified into two structural types, based on the structure of either zinc or copper hydroxide nitrate with the typical compositions Zn5(OH)8(NO3)2)-2H2O and Cu2(OH)3NO3 (Figure 8) [77]. Interlayered nitrate is exchangeable. And a fraction of Zn or Cu cations could be isomorphically substituted for other divalent cations. These hydroxide salts are fundamentally built of brucite-like layers in which one forth of octahedral sites is vacant. Particularly, vacant zinc sites are occupied with tetrahedral zinc cations. Their theoretical anion-exchange capacity is similar to those of other LDHs (2-5 meq/g).
rm
* r»
W*»
B
Figure 8 - The structure ofZn5(OH)8(NO3)r2H2O and Cu2(OH)3NO3. 3.2 - Preparation LDHs are easily and economically synthesized in high purity and yield, which greatly increases the potential of LDH for various applications. In spite of great heterogeneity in framework composition, LDHs are exclusively synthesized by precipitation of metal hydroxides, typically coprecipitation of mixed metal cations by base titration either with or without hydrothermal treatment that usually enhances crystalline property. Because the contamination with air-born carbonate ions frequently occurs during synthesis procedures, a special caution is needed to prepare carbonatefree LDHs. On the other hand, two dimensional heterostructured LDH hybrids could be prepared by three different techniques, coprecipitation, anion exchange and reconstruction ones (Figure 9). Intensive attention has been given to coprecipitation technique because it allows lattice engineering through fine tuning of framework charge density, framework composition, anion properties, treatment condition, and etc. [70-71].
414
J. Choy and M. Park
Figure 9 - Strategic illustration for preparation of LDH-derived complexes and their application; (a) C60-LDH hybrid (b) metal complex-LDH hybrid (c) isomeric organic molecue-LDH hybrid and (d) bulky organic molecule-LDH hybrid and their delaminated product 3.3 - Pharmaceutical Applications Pharmaceutical applications of LDHs mainly depend on acid buffering effect and anion exchange property. Hydrotalcite-derived antacid and antipeptic are representative of their applications in pharmaceutics [55,78-81]. Hydrotalcite was also explored as potential adsorbents of intestinal phosphate [82-84]. In addition to simple acid buffering and anion adsorption ability, hydrotalcite was also suggested to have barrier properties similar to those of gastric mucous, and to afford mucosal protection by its ability to maintain or mimic the barrier properties of gastric mucous gel. Unfortunately, LDHs could be found in a narrow range of pharmaceutics currently, unlike natural cationic clays. However, recent advance in hybridization technique has brought about a dramatic increase in the attention to their pharmaceutical potentials. Pioneering works have been carried out by Choy et al. They demonstrated excellent potentials of LDHs as a reservoir and delivery carrier for genes and drugs by hybridizing with DNA and As-myc, and etc. [59,62]. X-ray diffraction analyses showed that the interlayer distance of LDH increased from 0.87 nm (for NO3") to 2.39 nm (DNA), 1.94 nm (ATP), 1.88 nm (FITC), and 1.71 nm (As-myc), respectively, upon intercalating corresponding biomolecules into hydroxide layers (Figure 10). They clearly showed the excellent efficiency of LDH as gene and drug delivery carrier as well as protective role. The intercalated DNA was safely protected against harsh conditions such as strong alkaline to weak acidic environments and against DNase attack. [59] HL-60 cells treated with As-myc/LDH hybrids exhibited time dependent
Cationic and Anionic Clays for Biological Applications
415
inhibition on cell proliferation, indicating nearly 65 % of inhibition on the growth compared to the untreated cells, after 4 days (Figure 11). On the other hand, LDH itself was noncytotoxic towards HL-60, indicating its biocompatibility, and thus the suppression effect of cancer cell growth is solely from As-wyc/LDH. [62].
Figure 10 - Powder X-ray diffraction patterns for (a) the pristine LDH, (b) DNA-LDH, (c) ATP-LDH, (d) FITC-LDH, and (e) As-myc-LDH
Figure 11 - Effect of As-myc/LDH hybrids and As-myc only on the growth of HL-60 cells.
416
J. Choy and M. Park
Controlled cells are incubated without any treatment. The final concentration of each material was 20 uM. These remarkable results have evolved into the advanced concepts of LDHs for pharmaceutical applications, that is, degradable inorganic drug delivery system and targeted delivery. The drugs intercalated into LDHs are easily released by ubiquitous carbonate ion due to extremely high affinity. Furthermore, distinguished from other inorganic matrixes like cationic clays, not only could LDH be prepared from biocompatible compositions with arbitrarily tailored physical and chemical properties but also it is completely decomposed by acidic body fluids. These characteristics could be exploited for novel target delivery system for a wide spectrum of drugs. The schematic illustration was also proposed, based on the above experimental results along with inherent characteristics of LDH (Figure 12).
Figure 12 - The schematic diagram of bio-LDH hybrid (a) pristine LDH host (b) biomolecules (c) bio-LDH hybrid for biomolecule reservoir or carrier (d) cellular uptake mechanism of bio-LDH hybrid for biomolecule carrier cell-line by apoptosis. Lanel; control, Lane 2; LDH, Lane 3; MTX, Lane 4; MTX-LDH.
Cationic and Anionic Clays for Biological Applications
All
control! LDH LJ MTX LJ Hybrid
Figure 13 - Electrophoresis analysis for detection ofDNA ladder formation ofSaOS-2
Figure 14 - The effect of MTX-LDH on normal cell growth at the concentration of 5.0 Hg/ml.
418
J. Choy and M. Park
Quite recently, Choy et al also reported a successful application of drug-LDH hybrid to in vitro cancer treatment in which LDH played an essential role hi protective delivery of methotrexate (MTX) [85] (Figure 13,14). The initial proliferation of SaOS-2 cell was more strongly suppressed by treatment with MTX-LDH hybrid than with MTX alone. A series of genetic and efficacy analyses indicated that LDH did not exert any appreciable harmful effects on both normal and cancer cells, and that the treatment mechanism of MTX was not affected by hybridization. These results strongly suggested that LDH not only plays a role as a biocompatible delivery matrix for drugs but also facilitates a protective delivery to result in a significant increase in the delivery efficiency. Furthermore, Kriven together with Choy et al [86] evaluated the in vivo safty of LDH for adult male Sprague Dawley rats, and found that LDHs could be veininjected without any considerable effects on tissues and organs below a dose rate of 100 mg/Kg. LDHs, when extravascular, are locally irritating and elicit an inflammatory response around the precipitated particles at a dose rate of more than 200 mg/Kg. The pioneering works of Choy et al have led to a rapid increase in the research on pharmaceutical applications of LDHs [59-63]. Khan et al showed that a series of pharmaceutically active compounds including diclofenac, gemfibrozil, ibuprofen, naproxen, 2-propylpentanoic acid, 4-biphenylacetic acid and tolfenamic acid could be reversibly intercalated into LDH for their storage and controlled release [55]. Ambrogi et al have reported the hybridization of LDH with ibuprofen exhibiting antiinflammatory activity through which its controlled release as well as watersolubility could be significantly enhanced [83]. Recently, Ambrogi et al also found that hybridization with LDHs resulted in a significant increase in solubility of the nonsteroidal anti-inflammatory drugs such as indomethacin, tiaprofenic acid, and ketoprofen [86]. In gastric juice of pH 1.2, their LDH hybrids exhibited much higher (up to twice in case of tiaprofenic acid) solubility than their free forms did. They suggested that the enhancement of solubility resulted from the lack in crystallinity of intercalated drugs, which is directly released in ionic form by the fast dissolution of LDHs in acidic medium. Increase in water-solubility of poorly soluble drugs also plays a crucial role in drug bioavailability and hence in drug formulation. Many different approaches have been developed to improve drug solubility, which include introduction of polar or ionizable group, preparation of soluble prodrugs, use of polymorphs and of amorphous form of the drug, complexation, formation of inclusion compound, solid dispersions, etc. Compared with these approaches, hybridization with LDHs offers outstanding advantages like established safety, simplicity, cost effectiveness, and high dispersion property. Current researches for the pharmaceutical applications are still concentrated mainly on simple intercalation and disintercalation of various drugs in LDHs. The drugs to be intercalated into LDHs are widely expanding from simple anionic forms to neutral and zwitterionic forms. However, there are the very limited reports on in vitro activities of the intercalated drugs. It is not so far away for LDHs to be practically utilized as a next generation drug carrier for a broad spectrum of drugs. 3.4 - Cosmetic applications Practical application of LDHs to cosmetics have not been found so far although LDHs are reported to offer many fascinating aspects like high adsorption capacity to remove skin exudates and to encapsulate skin-sensitive coloring and UV-screening
Cationic and Anionic Clays for Biological Applications
419
agents, excellent anion exchange ability to protectively deliver active substances for anti-wrinkling and skin-regenerating, and also stabilizing potentials to improve rheological properties of various formulations, especially emulsion. Even though very limited, there are dozens of references to explore the potentials of LDHs for cosmetic purposes. Choy et.al. reported the intercalation of vitamins A, E, and C into zinc-based double salt (ZBS) for their safe storage and delivery [53-54]. In particular, they carried out further study on ascobate-LDH hybrid for controlled release and safe delivery. Ascorbate was intercalated into ZBS by coprecipitation route, and then the hybrid was coated with silica to enhance ascorbate stability and dispersion property of the hybrid. Futhermore, they showed better skin permeation efficacy of intercalated ascorbate. Figure 15 shows the X-ray diffraction patterns of the vitamin C-zinc hydroxide hybrid obtained during the first encapsulation process (a) and the silica modified one (vitamin C-inorganic hybrid) and SEM of (b) [53]. The primary L-ascorbic acid-inorganic hybrid exhibited a layer character with the basal spacing of 14.5 A to indicate an intercalation with 1:1 layer sequence along the c-axis, where L-ascorbate molecules were encapsulated by inorganic layers as depicted in the inset.
Figure 15 - (A) Powder X-ray diffraction patterns of a) vitamm C-zinc hydroxide hybrid and (b) Vitabrid-C and (B) scanning electron micrograph of Vitabrid-Cpowder The silica deposition on the primary L-ascorbic acid-inorganic hybrid gave rise to a drastic suppression of long range ordering, as indicated by X-ray amorphous property of the modified hybrid. It was also noticed that the vitamin C molecules encapsulated in the interlayer space of inorganic layers were released in a timecontrolled manner gradually by foreign chloride anions via ion-exchange process in an aqueous solution of 0.08 % NaCl. The released vitamin C was confirmed to be the pure one by its UV-Vis spectrum. In skin permeation test, the overall features of permeation patterns are quite similar one another, suggesting the similar penetration mechanism irrespective of the sample forms. However, the absolute amounts of permeated vitamin C after 24 hrs are more or less different with the following order; the modified hybrid powder (12.0 mg/cm2) > the modified hybrid powder in w/o emulsion (10.4 mg/cm2) > pure vitamin C in o/w emulsion (7.9 mg/cm2). This clearly indicated that the inorganically encapsulated vitamin C shows higher penetration rate than the pure
420
J. Choy and M. Park
vitamin C. The proposed releasing and delivering mechanism of vitamin C molecules in vitamin C-inorganic hybrid is schematically represented in Figure 16. In vitamin Cinorganic hybrid, the vitamin C molecules are adsorbed and immobilized between inorganic layers with positive surface charge, and further coated with nano-sized silica particles, forming a nanoporous shell structure.
Figure 16 - The proposed releasing and delivering mechanism of vitamin C in VitabridC Due to its well developed nanoporous structure, the vitamin C-inorganic hybrid absorbs effectively the skin wastes, serums, and sweats discharged from the human skin. Actually, the hybrid shows a large oil absorption capacity more than 150 %. The absorption of chemical species such as NaCl and fatty acids in sweat and skin wastes into the nanopores of the hybrid gives rise to a release of vitamin C in the pore by the exchange reaction between them, in such a way that the vitamin C molecules could be slowly diffused out from the inorganic shell and delivered into the epidermis in skin. There are couples of other reports dealing with LDHs for cosmetic applications [56,88]. Hussen et al intercalated naphthol blue black into Mg-Al LDH for its encapsulation, which may make its formulation more easy and broad. They attempted to encapsulate the human skin several organic UV absorbents such as 4-hydroxy-3methoxybenzoic acid, 2-hydroxy-4-methoxybenzophenone-5-sulfonic acid, 4-hydroxy3-methoxycinnamic acid, 4,40-diaminostilbene-2,20-disulfonic acid, p-aminobenzoic acid and urocanic acid by the intercalation into Zn2Al layered double hydroxides (Zn2Al-LDHs). They found that the oxidation catalytic activity of the intercalated UV adsorbents for the air oxidation of castor oil greatly decreased along with enhancement of the UV absorption ability. Further elucidation on potential of LDH and consequent applications would be actively exploited in near future.
Cationic and Anionic Clays for Biological Applications
421
3.5 - Agricultural applications Layered double hydroxides are one of the idealized inorganic matrices for a wide range of agricultural fields because not only could their framework be decomposed into plant nutrients, but also their structures offer charming features such as accommodation and controlled release of various active anionic agro-substances, high buffering capacity, high water retention ability, and acid-neutralizing potential. However, their agricultural applications within biological scope are rarely found up to now although cationic clays have been widely utilized. Only a few references are available. One of the main reasons seems to be due to the fact that cationic clays are naturally abundant and cheap. For another reason, demand on anionic clays in agriculture was not so high and urgent enough to search for them. However, the present situation has changed since so many anionic compounds derived from agriculture are contaminating soil and water environments, intensively cultivated soils develop acidic property extensively, and advances in various techniques lead both to increased demand on anionic clays and to their cost-effective availability. In fact, the attempts to remove anionic pesticides by adsorption to LDHs have steadily increased recently, which is not a main topic here. LDHs possess the excellent potential as green carrier for plant nutrients, pesticides, and growth regulators and as active principle in animal feeds, although currently not so many researches are undergoing. Komarneni et al [89] suggested nitrate-LDH as a potential slow-release fertilizer by synthesizing nitrate-LDH in ambient condition without any considerable contamination of carbonate-LDH. Recently, a plant growth regulator a-naphthaleneacetate (NAA) was intercalated through coprecipitation route by Hussein et.al [90] to explore the protected storage and controlled release in natural environments. More attention has been given to pesticide formulation that consists mainly of various organic solvents. Lakraimi et al [57] prepared pesticide-LDH hybrid with 2,4-dichlorophenoxyacetate, a broad leaf herbicide, by ion exchange reaction with chloride form of ZnAl-LDH for slow-release formulation. It is expected, once their nontoxicity is confirmed, that LDHs could be found in animal feeds soon or later because their acid neutralizing potential and high anion adsorption capacity are highly required to animal feeds as a complement to cationic clays. 4 - Conclusive remarks Clays have served human life in various ways, and their contribution will be further expanded in future. To date, many natural clays have been incorporated into various commercial products as simple additives or adjuvants; whereas, their uses as active principles are very limited. However, recent attention has been rapidly shifted to their advanced applications. In particular, a new interdisciplinary field is to emerge from protective and controlled delivery of various functional components with both natural and synthetic clays. An increasing number of bio-clay hybrids are continuously design-made and ready to be applied as the new types of delivery systems. It is easily expected that more and more bio-clay hybrids will be applied to in vivo study sooner or later. This trend also renews the interest in synthetic clays because their physical and chemical identities could be precisely controlled. Therefore, human life will further benefit from intensive exploration on biological potentials of clays.
422
J. Choy and M. Park
5 - References [I] T. J. Pinnavaia, Science, 220 (1983) 220. [2] J.K. Thomas, Chem. Rev., 93 (1993) 301. [3] Vaccari, Catalysis Today, 41 (1998) 53 [4] T. Shichi and K. Takagi, J. Photochem.Photobiol. C , 1 (2000) 113 [5] M. Ogawa and K. Kuroda, Chem. Rev., 95 (1995) 399. [6] S.L. Swartzen-AUen and E. Matijevic, Chem. Rev., 74 (1974) 385 [7] K.G. Theng, The Chemistry of Clay-Organic Reactions, Adam. Hilger, London, 1974. [8] J.H. Choy, S. Y. Kwak, Y.S. Han and B.H. Kim, Mater. Lett., 33 (1997) 143. [9] M. Ogawa and K. Kuroda, Bull. Chem. Soc. Jpn., 70 (1997) 2593. [10] K.A. Carrado, Appl. Clay Sci., 17 (2000) 1. [II] K. Ohtsuka, Chem. Mater., 9 (1997) 2030. [12] G.W. Brindley and R.E. Sempels, Clay Miner., 25 (1977) 229. [13] N. Lahav, U. Shani and J. Shabtai, Clays Clay Miner., 26 (1978) 107. [14] D.E.W. Vaughan, R.J. Lussier and J.S. Magee, Jr. U.S. Patent 4, 176, 090, 1979. [15] S. Yamanaka and G.W. Brindley, Clays Clay Miner., 27 (1979) 119. [16] F.H. Lin, Y.H. Lee, C.H. Jian, J.M. Wong, M.J. Shieh and C.Y. Wang. Biomaterials, 23 (2002) 1981. [17] M.J. Wilson, A Handbook of Determinative Methods in Clay Mineralogy, Chapman and Hall, New York, 1987 [18] X. Xia, J. Yih, N.A. D'Souza and Z. Hu, Polymer, 44 (2003) 3389. [19] T. Ito, T. Sugafuji, M. Maruyama, Y. Ohwac and T. Takahashic, J. Supramol. Chem., 1(2001)217. [20] K. Mayura, M.A. Abdel-Wahhab, K.S. McKenzie, A.B. Sarr, J.F. Edwards, K. Naguib and T.D. Phillips, Toxi. Sci., 41 (1998) 175. [21] L. Fowden, R.M. Barrer, P.B. Tinker (Eds.), Clay Minerals: Their Structure, Behaviour and Use, The Royal Society, London, 1984. [22] A.C.D. Newman (Ed.), The Chemistry of Clays and Clay Minerals, Monograph 6 of the Mineralogical Society, Longman, London, 1987. [23] S. Poyard, N. Gaffrezic-Renault, C. Martelet, S. Cosnier, P. Labbe and L.L. Besombes, Sensors and Actuators B., 33 (1996) 44. [24] S. Komarneni, R. Roy, Science, 239 (1988) 1286. [25] N. Takahashi, M. Tanaka, T. Satoh, T. Endo, M. Shimada, Microporous Mater., 9 (1997)35. [26] M.I. Carretero, Appl. Clay Sci., 21 (2002) 155. [27] E. Gamiz, J. Linares and R. Delgado, Appl. Clay Sci., 6 (1992) 359 [28] D. Poensin, PH. Carpentier, C. Fechoz and S. Gasparini, Joint Bone Spine, 70 (2003) 367. [29] T. Ferrand and J. Yvon, Appl. Clay Sci., 6 (1991) 21.. [30] J.W. McGinity and J.I. Lach, J. Pharm. Sci., 66 (1977) 63. [31] J.L. White and S.L. Hem, Ind. Eng. Chem. Prod. Res. Dev., 22 (1983) 665. [32] W.J. Macrter and R. Broussean, Cem. Cone. Res., 20 (1990) 891. [33] A. Akelah and A. Moet, J. Appl. Polym. Sci. Appl. Polym. Symp., 55 (1994) 153. [34] W.F. Lee and Y.T. Fu, J. Appl. Polym. Sci., 89 (2003) 3652. [35] A.R. Elmore and F.A. Andersen, Int. J. Toxicol., 22 (2003) 37. [36] S.S. Ray and M. Okamoto, Prog. Polym. Sci., 28 (2003) 1539.
Cationic and Anionic Clays for Biological Applications
423
[37] L.M. Zhang, C. Jahns, B.S. Hsiao and B. Chu, J. Colloid Interface Sci., 266 (2003) 339. [38] S.S. Ray and M. Okamoto, Macromol. Rapid Comm., 24 (2003) 815. [39] K. Okamoto, S.S. Ray and M. Okamoto, J. Polym. Sci. Pol. Phys., 41 (2003) 3160. [40] P. Grant and T. Phillips, J. Agric. Food Chem.,.46 (1998) 599. [41] J. Ferrario, A. Byrne and D. Cleverly, Environ. Sci. Technol., 34 (2000) 4524. [42] T.D. Phillips, L.F. Kubena, R.B. Harvey, D.R. Taylor and N.D. Heidelbaugh, Poult. Sci., 67(1988)243. [43] J. Ramu, K. Clark, G.N. Woode, A.B. Sarr and T.D. Phillips, J. Food Protection, 60(1997)1. [44] S.L. Lemke, P.G. Grant and T.D. Phillips, J.Agric. Food Chem., 46 (1998) 3789 [45] T.C. Chen and E.J. Day , Ani. Feed Sci. Tech., 27 (1989) 83. [46] CD. Brown, C.J. Fryer and A. Walker, Pest. Manag. Sci., 57 (2001) 1127. [47] D.R. Taylor and R.G. Goss, N Eeng-Littleton, 52 (2000) 54. [48] A. Armstrong, K. Aden, N. Amraoui, B. Diekkruger, N. Jarvis, C. Mouvet, P. Nicholls and C. Wittwer, Agr. Water Managen., 1-3 (2000) 85. [49] Y. El-Nahhal, S. Nir, T. Polubesova, L. Margulies and B. Rubin, J. Agric. Food Chem., 46 (1988) 3305. [50] Y. El-Nahhal, S. Nir, L. Margulies and B. Rubin, Appl. Clay Sci., 14 (1999) 105. [51] M.J. Carrizosa U.M.J. Calderon, M.C. Hermosin and J. Cornejo, Sci. Total Environ., 247(2000)285. [52] G. Lagaly, Appl. Clay Sci., 18 (2001) 205. [53] J.H. Yang, S.Y. Lee, Y.S. Han, K.C. Park and J.H. Choy, Bull.Korean. Chem. Soc, 24 (2003) 499. [54] S.H. Hwang, Y.S. Han, and J. H. Choy, Bull. Korean Chem. Soc, 22 (2001) 1019. [55] A.I. Khan, L. Lei, A. J. Norquist and D. O'Hare, Chem. Commun., (2001) 2342 [56] M.Z.B. Hussein, Z. Zainal, A. Hj. Yahaya, and A.B. Abd. Aziz, Mate. Sci. Eng. B., 88 (2002) 98. [57] M. Lakraimi, A. Legrouri, A. Barroug, A. De Roy and J. P. Besse, J. Mater. Chem., 10 (2000) 1007. [58] F.J. Brocker and L. Kainer, German Patent 2, 024,2 82 (1970), to BASF AG. and UK Patent 1, 342, 020 (1971), to BASF AG. [59] J.H. Choy, S.Y. Kwak, J.S. Park, Y.J. Jeong and J. Portier, J. Am. Chem. Soc, 121 (1999) 1400. [60] J.H. Choy, S.Y. Kwak and J.S. Park, J. Am. Chem. Soc, 121 (1999) 1399. [61] J.H. Choy, S.Y. Kwak, J.S. Park and Y.J. Jeong, J. Mater. Chem., 11 (2001) 1671. [62] J.H. Choy, S.Y. Kwak, Y.J. Jeong, and J.S. Park, Angew. Chem. Int. Ed., 39 (2000)4041. [63] S.Y. Kwak, Y.J. Jeong, J.S. Park and J.H. Choy, Solid State Ionics, 151 (2002) 229. [64] E. Dumitriu, V. Hulea, C. Chelaru, C. Catrinescu, D. Tichit and R. Durand, Appl. Catal, A., 178(1999)145. [65] S. Velu and C.S. Swamy, Appl. Catal., 119(1994)241. [66] S. Ueno, K. Yamaguchi, K. Yoshida, K. Ebitani and K. Kaneda, J. Chem. Soc, Chem. Commun., (1998) 295. [67] F. Medina, D. Tichit, B. Coq, A. Vaccari and N. Thy Dung, J. Catal., 167 (1997) 142.
424
[68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90]
J. Choy and M. Park
J. Serrano, V. Bertin and S. Bulbulian, Langmuir, 16 (2000) 3355. R. Allmann, Acta Crystallogr. Sect. B., 24 (1968) 972. A.I.Khan and D.O'Hare, J. Mater. Chem., 12 (2002) 3191. F. Cavani, F. Trifiro and A. Vaccari, Catal. Today., 11 (1991) 173. I. Rousselot, C. Taviot-Gueho, F. Leroux, P. Leone, P. Palvadeau, and J-P Besse, J. Solid State Chem., 167 (2002) 137. W. Stahlin and H.R. Oswald, Acta. Cryst. B., 26 (1970) 860. S.P. Newman and W. Jones, J. Solid State Chem., 148 (1999) 26. H. Morioka, H. Tagaya, M. Karasu, J.I. Kadokawa and K. Chiba, J. Mater. Res., 13(1998)848. M. Meyn, K. Beneke and G. Lagaly, Inorg. Chem., 32 (1993) 1209. H. Nishizawa and K. Yuasa, J. Solid State Chem., 141 (1998) 229. P. Grubel, K.R. Bhaskar, D.R. Cave, P. Garik, H.E. Stanley and J.T. Lamont. Aliment Pharmacol Ther., 11 (1997) 139. S. Miyata, Kyowa Chemical Industry Co. Ltd (Tokyo, JP), United States Patent, 4,514,389, 1982. A.C. Playle, S. R. Gunning, and A.F. Llewellyn, Pharm. Acta Helv., 49 (1974) 298. H. Nakayama, N. Wada, and M. Tsuhako, Int. J. Pharm., (2004), in press. A. Ookubo, K. Ooi, and H. Hayashi, J. Pharmaceutic. Sci., 81 (1992) 1139. H. Nakayama, K. Takeshita, and M. Tsuhako, J. Pharmaceut. Sci., 92 (2003) 2419. V. Ambrogi, G. Fardella, G. Grandolini and L. Perioli, Int. J. Pharmaceut., 220 (2001)23. J.H Choy, J.S. Jung, J.M. Oh, M. Park, J.Y. Jeong, Y.K. Kang and O.J. Han, Biomaterials, (2004), in press. W.M. Kriven, S-Y. Kwak, M.A. Walling and J-H. Choy, MRS Bulletin, (2004), in press. V. Ambrogi, G. Fardella, G. Grandolini, M. Nocchetti and L. Perioli, J. Pharmaceut. Sci., 92 (2003) 1408. J. Olanrewaju, B.L. Newalkar, C. Mancino, and S. Komarneni, Mater. Lett., 45 (2000) 307. Q. He, S. Yin, T. Sato, J. Phys. Chem. Solid, (2003), in press. M. Z. bin Hussein, Z. Zainal, A.H. Yahaya and D.W.V. Foo, J. Contr. Release, 82 (2002)417.
ENVIRONMENTAL HYDROXIDES
REMEDIATION
INVOLVING
LAYERED
DOUBLE
CLAUDE FORANO Laboratoire des Materiaux Inorganiques, UMRCNRS 6002, Universite Blaise Pascal, 63177, Aubiere Cedex - FRANCE E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
426
C. Forano
1 - Introduction 1.1 - Environmental challenges In 2001, Bejoy [1] published a paper entitled "Hydrotalcite: the clay that cures". Obviously, this paper referred to the curative properties of hydrotalcite for medicinal applications. However, this curative character can be easily extended to the Environment, and this review is demonstrating that hydrotalcite like compounds and more generally Layered Double Hydroxides (LDHs) have curative or remediation properties for the Environment too. Indeed, investigations on the potential uses of LDHs for decontamination of environmental sites or prevention of pollutant dispersion in Nature are growing greatly since ten years [2-6]. A wide range of contaminants can be removed from industrial effluents or wastewater by anion exchange and adsorption processes or catalytic remediation, using LDHs, modified LDHs or calcined LDHs. Pesticides, toxic organic chemicals, greenhouse gases, heavy metals, undesirable inorganic substances are as much targeted molecules to be controlled for the preservation of earth and health security. 1.2 - Presentation of Layered Double Hydroxides The Layered Double Hydroxides (LDHs) display unique physical and chemical properties surprisingly close to the properties of clays. Their layered structure, their wide chemical compositions due to variable isomorphic substitutions of metallic cations, their variable layers charge density, their ion exchange properties, their very reactive interlamellar spaces often used as nanoreactors, their water swelling, rheological and colloidal properties makes them a mineral family that can be referred as clays and based on their anion exchange properties as anionic clays. Hydrotalcite, Mg6Al2(OH)16CO3.4H2O is one of the most representative mineral of the group and other minerals are often named as Hydrotalcite like compounds (HTlc). Its structure is related to the layered Mg(OH)2 brucite structure where part of Mg2+ cations of the layer have been substituted by Al3+. Carbonate anions are then intercalated within the layers to insure the structure electroneutrality. From this structural consideration, the chemical formula of hydrotalcite can be written as: Mgo.75Al0 25(OH)2(C03)o.5.0.5H20 or in an abbreviated way [Mg-Al-CO3] or [Mg-Al]. The other members of the family are built on a combination of divalent and trivalent metals and interlayer anions. Their general formula is [MII1_xMIIIx(0H)2][Xq" x/q.nHzO] where [Mn,_xMmx(0H)2]x+ represents the layer and [Xq"x/qnH2O] the interlayer chemical compositions (Figure 1).
Figure 1 - Structure of Layered Double Hydroxides
Environmental Remediation Involving Layered Double Hydroxides
427
In order to simplify the chemical description, the following abbreviated formula is often used: [MnR-Mm-X] (with R = (l-x)/x). A wide range of [M2+-M3+] associations have been incorporated in the structure: -
M2+ = Mg, Ca, Sr, Mn, Fe, Co, Ni, Cu, Zn, Cd M3+ = Al, Cr, Fe, Sc, Ga, Y, In, Ce M+ = Li M4+ = Zr, Pd, Sn
LiAl2(OH)6X.nH2O is the only example with a monovalent cation incorporated in the structure. Recent works have claimed the introduction of tetravalent metals in the brucite layer [7-9] has enable that tetravalent metal cations to be incorporated into the structure to some extent. But LDHs are not limited to a binary combination of metal cations; multimetal LDHs can also be synthesized. Concerning the chemical composition of the interlayer domains, a wide range of anionic species can be intercalated. This specific property of LDHs will be discussed when referring to their anion exchange capacities. Many reference papers have reviewed the methods of preparation, the physicochemical properties and the applications of anionic clays in recent years [10-21,21b]. 1.3 - Physical and chemical properties of LDHs in relation with remediation Anion Exchange Capacity The anion exchange properties of the LDHs are generated by the presence of charge compensating anions in the interlayer spaces of the structure (Figure 1). The amount of anions per unit formula (MII1.xMIIIx(OH)2Xx/q.2/3H2O) is fixed by x, the rate of substitution of the divalent metal by the trivalent ones. The anion exchange capacity, a.e.c. depends on x value and a straightforward definition of a.e.c. can be given for a monovalent anion containing LDH:
a.e.c. = — (meq. 1100 e) F.W. where F.W. is the formula weight and F.W. = (MMII + 46) + (M Mm + M x - MMn) x. The F.W. formula supposes a full occupancy (2/3) of the interlayer crystallographic sites for the water molecules, 0.66 H2O/Metal. Since the anion exchange capacity is directly related to the M n /M m molar ratio (R), it is constant for LDH materials with fixed R value such as [Li-Al2], [Zn2-Cr] and [Cu2-Cr], and highly tunable for [Mn-Mm] systems with variable M n /M m ratio, i.e. [MgR-Al] (1
(c.d.= —=—
)
and
the
free
cross
section
area
428
( S free
C. Forano
=
(nm
2
/ e ) ) P e r l a y e r f° r
a sel
"ies of anionic clays. The usual values
c.d. range from 200 to 400 meq/lOOg. Table 1 - a.e.c. for some synthetic anionic clays. [MU-MU1] [Mg-Al-Cl]
[Zn-Al-Cl]
[Mg-Al-CO3] [Mg-Al-NO3] [Li-Al-Cl]
X
0.20 0.25 0.33 0.20 0.25 0.33 0.33 0.33 0.33
F.W. (g/mol) 77.92 79.83 82.88 110.80 110.65 110.41 81.08 91.64 78.12
a.e.c. (meq/lOOg) 256.7 313.2 398.2 180.5 225.9 298.9 407.0 360.1 422.4
a (A) 3.060 3.054 3.042 3.09 3.08 3.07 3.042 3.042 3.070
c.d. (e/nm2) 2.47 3.09 4.24 2.49 3.13 4.16 4.24 4.24 4.16
Sfree
(nm 2 /e) 0.405 0.323 0.236 0.401 0.319 0.240 0.236 0.236 0.240
Comparison of these ion exchange capacities with those of the anti-type cationic clays clearly shows the greater ability of the anionic clays to incorporated interlayer ions in their structure and their flexibility in terms of a.e.c. variations. These exchange properties have been extensively used for the preparation of new LDH phases containing a wide range of anions. The anion exchange reaction can be described by the following equilibrium: [Mn1.xMIIIx(OH)2]X"-x/q.nH2O + xNaY% p « [Mn1.xMmx(OH)2]Y<>-x/p.nH2O + xNaX"-I/q A priori, no theoretical limit appears for the intercalation of all types of anions in the LDH structure and a large number of atoms or molecules can be intercalated in their anionic forms. Adding that, neutral polar molecules can also be co-intercalated with these anions, this gives rise to a very wide composition range of the interlayer domains. Typically, the following families of anions are concerned: > halides (F, Cl", Br", I"), > non-metal oxoanions (BO33\ CO32", NO3", Si2O52", HPO42\ SO42", C1O4", AsO 4 3 \ SeO42", BrO4", etc.), > oxometallate anions (VO 4 3 \ CrO42", MnO 4 \ V10O286\ Cr2O72", Mo7O246\ PWi2O403", etc.), > anionic complexes of transition metals (Fe(CN)62", etc.), > volatil organic anions (CH3COO", C6H5COO", C12H25COO", C2O42", C6H5SO3", etc.), >anionic polymers (PSS, PVS, etc.). Acid-Basic properties Layered double hydroxides and their calcined products display unique acidbasic properties. A great number of catalytic applications involved the basic character of these compounds [17,18]. This property has obviously a great interest when looking at new materials for the uptake of acid molecules such as acid gases (CO2 or SO2).
Environmental Remediation Involving Layered Double Hydroxides
429
Knoevenagel condensation, CO2 adsorption or acid titration experiments are commonly used to quantify the basicity of the materials [23-26]. It has been shown that calcined LDHs display a stronger basic character than the uncalcined precursors due to the presence strong O2' basic sites. Moreover, the basic character depend on the chemical composition of both the layer and the interlayer. Rousselot et al [23] have related this property to the electronegativity of the metal cations, showing that phases containing Mg, Ca or Ga display higher basic characters. Shape and morphology When regarding a material for its adsorption or catalytic properties, the study and the tuning of its textural properties appear necessary. Natural Hydrotalcite displays particles with a regular hexagonal shape and sizes ranging between 2 to 20 um. The morphology of synthetic LDH particles depends on the method of preparation (Figure 2). Particles with a high degree of crystal cohesion leading to well-resolved XRD patterns but with a wide range of particle sizes are obtained when using the standard constant pH coprecipitation method or the "urea synthesis" [27]. Fine grained crystals with rough surfaces, relatively high surface areas and mesopores with size in the range 50-300 A were obtained by the variable pH coprecipitation method. Edge-face crystal aggregation and cofacial layer stacking create a secondary particle aggregation with a sand-roses morphology and interparticle mesopores which provide voids for water. Morphology is also affected by the chemical composition of the structure [28].
Figure 2 - S.E.M. of Cu2Cr(OH)6Cl.2H2Oprepared by a) CuO/CrCl3 induced hydrolysis and b) coprecipitation (V. Prevot, unpublished results) and Mg2Al(OH)6(CO3)0.5-2H2O prepared by c) coprecipitation andd) urea method [27].
430
C. Forano
Surface and porosity properties The theoretical surface areas of one LDH monolayer can be easily calculated taking into account its composition and structural property: Stheo = a 2 V3 10"18N/(F.W.) N = Avogadro number; a = cell parameter (run); F.W. = formula weigth relative to the unit formula (g.mol-1). For Zn2Cr(OH)6C1.2H2O and Mg3Al(OH)8(C03)o.5.2H20, the calculated S ^ values are respectively 817 m2/g and 1285 m2/g. Practically, such high values cannot be measured for anionic clays because the internal surface is hardly accessible. Anionic clays generally display N2 isotherms features corresponding to mesoporous or nearly non-porous materials even when large pillaring anions (POM) are intercalated. The typical values of specific surface area (s.s.a.), measured by the B.E.T. technique, range from 20 - 85 m2/g. N 2 adsorption plots indicate type II N-adsorption isotherms with a narrow hysteresis cycle, due to pores open on both ends. However, specific surface area can be greatly improved when coprecipitation is performed in mixed water/alcohol media [29]. Synthetic hydrotalcite prepared in H2O/ethylene glycol (1/1) solution exhibits a s.s.a. value of 136 m2/g, larger than for the material prepared in standard conditions (45 m2/g). 2 - Remediation of inorganic contaminants by exchange/adsorption process Miyata [30] reported for the first time in 1983 the ion-exchange isotherms of a series of hydrotalcite like compounds leading to the ion selectivities of hydrotalcite for monovalent and divalent anions: OH" > F" > Cl" > Br" > NO3" > I" and CO32" > NYS2" > SO 4 2 \ Miyata concluded this work: "by utilising their characteristic ion selectivity, HTs are expected to find application in removal of acid dyes, HPO 4 2 \ CN", CrO42", AsO43", Fe(CN)63", etc. from waste waters". This work generated a great number of studies in which the anion exchange properties of LDHs and the anion adsorption ability under LDH reconstruction process have been investigated for the removal of environmentally undesirable anions. In 1995, Parker et al [31] published extended results on the use of hydrotalcite as anion adsorbents. They confirm the initial Miyata's results for the anion sorption capacity of a commercial [Mg-Al-CO3] sorbent : SO42" > F" > HPO42" > Cl" > B(OH)4" > NO3" > I" but stressed the point that the carbonate LDH is ineffective as anion exchanger. However, the calcined materials display higher adsorption properties even after various cycles of calcinations/rehydration. As a consequence of recent interests to develop the use of anionic clays for environmental remediation, anion exchange properties of LDH have been deeply reinvestigated. The main objective of these studies is to clearly characterize the adsorption properties of the materials under various solid/liquid interface conditions. The effect of sorbent composition, adsorbate concentration, proton concentration, solid/liquid ratio and competing anions on adsorption have been examined. Physisorption and chemisorption, surface and bulk adsorption, concentration of adsorption sites have been assessed. The adsorption capacity is deeply affected by the nature of the counter-anion of the LDH layers. Inacio et al [32] showed that the adsorption capacity for the [Mg-Al-X] sorbent series varies in the order X: CO3 < Cl < NO3. The carbonate anion seems to reduce strongly the adsorption capacity. Indeed, the
Environmental Remediation Involving Layered Double Hydroxides
431
adsorption of dodecylsulfate by [Mg-Al-CO3] and [Mg-Al-Cl] leads respectively to 0 and 100% adsorption. The fluid structure at the interface and in the interlayer spaces, the effective diffusion coefficients of surface-sorbed species, their surface lifetimes, rotational and translational dynamics are controlled by the structure and the composition of the surface of the host structure [33]. In a study of internal versus external uptake of anions, J. Boclair et al [34] showed that ferrocyanide does not displace carbonate from synthetic hydrotalcite but is taken up on the outside of the particles. Anion uptake is in this case controlled by specific hydrogen bonding requirements and not by charge density alone, a feature that can be used to control whether uptake will be both internal and external, or external only. Ulibarri et al [35] have shown that both the type of interlayer anion and the crystallinity of the minerals affects deeply the adsorption capacity. Experimental adsorption isotherms are usually processed with the Langmuir or Freundlich models developed for physisorption [32]. However, in most of the studies the adsorption mechanism is identified as an anion exchange process, showing that both theoretical models are inadequate. 2.1 - Removal of oxoanions Amongst oxoanions, adsorption of phosphate species by LDHs was much more examined because of their consequences on water pollution [36-40]. Removal of phosphate by a dissolution/coagulation process using a synthetic pyroaurite ([Mg-FeCO3]) and hydrocalumite-like material ([Ca-Fe-CO3]) was effective (above 80% from studied effluents) but strongly dependent on the buffering pH effect of both the sorbent and the solution. The rate of removal was well correlated with the amount of both dissolved sorbent and release cations [36,37,38]. Simultaneous uptake of PO43" and NH4+ can be obtained by the use of non-selective nanocomposite sorbent made from granular zeolite coated with hydrotalcite and an organic binder [40]. 2.2 - Removal of heavy metals Presence of heavy metals in the environment arises from both natural and anthropogenic sources. They are present in the effluents of many industries and contamination of natural waters and soils is a major environmental concern owing to their high toxicity. The adsorption technology is the most common process used for the uptake of heavy metals from polluted reservoirs. However, depending on the pH conditions metal can exist as cations or various anionic forms. LDHs appear then as good sorbents for the removal of the anionic species but have revealed also effective for the uptake of metal cations. 2.2.1 - Removal of heavy metals as oxometallate anions Since the years 90's, researches for the development of hydrotalcite as metal adsorbents have focussed mainly on the removal of selenate [41-46], arsenate [41,43,4651] and chromate [47,52-60] anions. Selenium is an essential element for life but has a toxic effect at a very low level of concentration. Selenium has been widely used in industries, e.g., metal refineries, glass works, electronics industries, and others. It is volatile along with Hg and has received environmental concerns (maximum admitted value in aqueous environment: 100 ug.L"1). In natural waters and soils, Se exists in selenite (SeO32") and
432
C. Forano
selenate (SeO42~) forms. You et al [44] recently reported a detailed adsorption study of both Se anionic forms by [Mg-Al] and [Zn-Al] LDHs. The L-type ion-exchange isotherms of both LDHs show a high affinity for SeO32': 123 cmol.kg"1 and 463 cmol.kg"1 for respectively [Mg-Al] and [Zn-Al] (Figure 3). No significant difference was observed between selenite and selenate adsorption. The amount of adsorbed species is directly related to the anion exchange capacity of the host structure and the adsorption mechanism involves an anion exchange process. For the [Zn-Al] material, nearly 97% of the a.e.c. was reached, less than 90% for the [Mg-Al] sorbents. The competing effect of the lattice anions increases in the adsorption order: HPO42" < SO42" < CO32" < NO3". Intercalation of SeO42" in LDHs leads to an increase in the basal spacing (from 0.773 nm to 0.905 nm), while no interlayer distance change was observed in the case of the SeO32" exchange. Das et al [45] have shown that the fraction of SeO32" uptake increases with a decrease in both pH and temperature. The negative value of the enthalpy AH°a
Figure 3 - A dsorption isotherms for SeO}~ adsorbed at pH 9 and 25 °C on Mg-Al LDH and Zn-Al LDH from You et al [44]. Contamination of potable water by arsenic arises often from the natural leaching of As(V) containing minerals. In the pH range from 7.0 to 11.5, AsO43" is the predominant species, while at lower pH the protonated form (HAsO42") dominates. Calcined hydrotalcite were able to remove more than 70% of As(V) anionic species in a large range of concentration [51]. Synthetic Hydrotalcite were also effective for the removal of As(III) [48,49]. Cr(VI) is present in the effluents of many industries (electroplating, mining, tanning, fertiliser). Because of its high toxicity (carcinogenicity, liver damage), the maximum limit prescribes for chromium (VI) in water by the US Environment Protection Agency is 50 ug.L'1. The problem of contamination by Cr(VI) is a growing environmental concern because of the increasing discharge of chromium containing wastes. Many studies have been focused on the treatment of chromate containing aqueous solutions using calcined and uncalcined hydrotalcite like materials ([Mg-Al], [Mg-Al], [Ni-Al], [Zn-Al], [Zn-Cr]). Both CrO42' and Cr2O72" species can be adsorbed. Sorptive flotation of metal-loaded particles was proposed by Lazaridis et al [50,59] as an alternative 2-stage process for the uptake of arsenate and chromate from
Environmental Remediation Involving Layered Double Hydroxides
433
wastewaters. The ability of calcined hydrotalcite to remove oxoanions [47] from solution decreased in the order: vanadate > arsenate > chromate. As often observed, the absorption capacity is reduced after few cycles of calcination-rehydration-anion exchange. 2.2.2 - Removal of heavy metal as cations Hydrotalcite like compounds act as alkaline buffers in aqueous solutions leading to the precipitation of some heavy metals. Consequently, they can be used as precipitating agents of heavy metallic cations for the decontamination of wastewaters. Zinc cations were removed by hydrotalcite [56] according this technology. Synthetic pyroaurite has been used in batch and columns experiments to remove lead from water [61] Grady et al [62] developed the used of hydrotalcite thin films for the remediation of wastes containing hazardous metal ions (Pb, Cr, Hg, Cd, Cu, Co, Ag). Tarasov et al [63] demonstrated that ethylenediaminetetraacetate intercalated in [LiAl2(OH)6]Cl.(H2O)0 5 LDH display chelation properties for Ni2+ cations (Figure 4). Time-resolved in-situ x-ray diffraction measurements revealed that the chelation/intercalation reactions proceed very quickly.
Figure 4 - Possible structures of EDTA4' intercalated [Li-Al] LDH a) before and (b) after solid-state metal chelation from Tarasov [63]. 2.2.3 - Removal of nuclear wastes Technological processes for the disposal of nuclear wastes commonly involve the immobilisation of the hazardous in cementitious materials. Hydrotalcite like compounds and hydrocalumite or ettringite, the calcium form of hydrotalcite [23], have been evidenced as earlier alteration products of cements or basaltic glasses. Consequently they are good material candidates as barrier materials for nuclear wastes [64,65]. They have been studied for their adsorption properties of actinides (Th, U, Np, Pu, and Am) [66,67]. In order to improve the reliability of underground disposal systems used in the retention of nuclear wastes for long periods of time, it is interesting to consider materials with higher adsorption capacity for anionic species than that of oxide and hydroxides materials already used. The high ability of layered double hydroxides and their calcined products to adsorb/exchange anions from solution was also used for the removal of 99Tc, ""Re, 99Mo in their anionic form (MO4~) from radioactive wastewaters
434
C. Forano
[68-70]. Up to two MO4" anions (M=Tc, Re) per [Mg6Al2(OH)18] meixnerite formula, corresponding to 100% exchange, were absorbed [67,67b]. For the molybdate analogues, X-ray diffraction shows that intercalation of the oxometallate anions in the LDH structure leads to a low basal spacing value (0.779 nm) as already previously reported [12,15,72] for anion exchange studies on LDHs. Synthetic Hydrotalcite provides, at neutral pH, a similar distribution coefficient (102 mL/g) for TcO4" adsorption than for other materials used in low level radioactive waste backfill [69]. Calcined Hydrotalcite can also be used for the separation of 99Tc04" / "MoQj 2 " species from liquid mixtures by means of an elution process. The high adsorption affinity for 99Mo species allows to envisage the replacement of alumina by hydrotalcite as adsorbant for M0O42" in 99mTc generator used for medical purposes [67,70]. Synthetic hydrotalcite has been investigated for the sorption of 129I containing anionic species (129F and 129IO3") [69,71-74], one of the long-term radioactive species in underground disposals. Calcined LDHs show higher adsorption capacities for iodide (3.8.10"3 mol/g) than pyrite or magnesium oxide [75] because the iodide anions are incorporated in the LDH structure under a reconstruction/exchange process. 129IO3" has been identified as the stable iodine species in aerobic and alkaline conditions. [Mg-AlCO3] adsorbs 129IO3" anions (9.2 meq/g) twice its anion exchange capacity, only at the surface of the particles but releases up to 80% under water desorption. 129IO3" intercalates irreversibly in the [Mg-Al-NO3] structure by anion exchange [74]. 2.3 - Adsorption of greenhouse gases Calcined hydrotalcites display unique and strong basic properties, which make them efficient scavengers for acid gases recovery from hot gas streams. The recovery of CO2 and SOx from power-plant flue gases is considered to be the first step in reducing total carbon and sulfur oxide emissions. Many papers and patents describe the use of calcined LDHs for the adsorption of carbon dioxide [76-82] and sulfur oxide [83-87] gases. The Mg-Al-0 system is the adsorbent the most investigated. Mg-Al-O adsorbent can be supported in macropores (l|am) of porous alumina tubes for a better stability and reversibility of the CO2 adsorption [80]. Synthetic meixnerite display also a high CO2 ability and can adsorb more than 10 mL of CO2/g [77]. Ding et al [78,79] used a bench-scale adsorption unit based on a cyclic and multi-bed technology to analyse the adsorption and desorption of CO2 by a potassium promoted hydrotalcite adsorbent. Application are envisaged in existing industrial processes in which recovered CO2 at elevated temperature can be used as a feedstock for further catalytic processing, e.g., dry methane reforming and carbon gasification. Typical industrial applications of SOx include sulfur removal from fluid catalyst cracking process, cold-side combustion gas sulfur abatement and cleaner coal gasification [85]. [Ca6Fe2(0H)16](CO3).xH2O, [Mg6Fe(OH)16)(CO3).xH2O and [Ca2Al(OH)2](NO3).xH2O are suitable sorbents for removing SO2 from the flue gas cold side of coal-burning power plants [83,84]. The sorbents are useful at 100-400°C. SO2 gas absorbs into a hydrotalcite structure as SO32' anions by replacing most of the gallery CO32" anions. Albers et al [85] showed that incorporation of organic acids in anionic forms in the LDH adsorbent increases the sorption of SOx. [Mg2-Al-NO3] or [Mg3-AlNO3] were also used for the removal of elemental sulfur present in fluids such as refined petroleum products, e.g., gasoline, jet, diesel, kerosene or fuel additives such as ethers
Environmental Remediation Involving Layered Double Hydroxides
435
[86-87]. Various technologies for the treatments of flue gases containing other environmental pollutant gaseous compounds such as silane [88,89], arsine [89], phosphine [89], and other hydrides [90] also use hydrotalcite like compounds as adsorbents. More recently, Okada et al [91] tested the simultaneous acidic and basic gas adsorption properties of composites materials based on layered double hydroxide / alumino-silicate xerogels. The adsorption properties were evaluated by NH3 and CO2 temperature programmed desorption. Composites materials prepared by the sol/precipitation method display very thin mineral particles and showed higher adsorption for NH3 and CO2 gases compared with pure LDH and aluminosilicagel. 3 - Remediation of organic pollutants by adsorption/exchange process In the field of environmental technology, much attention has been paid in the potential applications of synthetic materials as sorbents for organic and inorganic pollutants in water. If activated carbon is one of the most used sorbent, clay minerals and modified clay minerals appear as promising alternative materials [92-94]. However the sorptive capacity of clay minerals is low, particularly for hydrophobic molecules. The increasing sorbent demand is stimulating research to look for new sorbent materials [95]. Since the 90's, many groups have focussed their researches on the use of anionic clays (LDHs) for the adsorption of contaminant molecules and companies have started to published few patents on this subject [96-99]. 3.1 - Adsorption of organic molecules by LDHs Intercalation of organic anions in layered double hydroxides has been investigated since many years [100-105]. Organic anions intercalate in the LDH structure with their anionic groups (CO2", -SO3", -PO3', -OSO3", -OPO3") always interacting via strong hydrogen bonding with the hydroxylated planes. The hydrophobic hydrocarbon skeleton is being push far away from the hydrophilic layer surface, in a lowest energy conformation arangement. a,co-dianionic molecules have a pillaring effect, bridging two adjacent layers separated by a basal spacing proportional to the size of the hydrocarbon skeleton. An organoclay with a single layer of anion is then obtained. For monoanionic molecules bilayer or inter-twined layer arrangements lead to an increase in the interlamellar distance, usually greater than 15 A [103-107,107b]. In the case of organic anions more particularly, the packing of the anions in the interlamellar spaces dependent on relation between the charge and size of the anions and the free area per unit charge of the layer. When the cross section area of the anions meets this free layer area, optimal packing is insured. This is the case for dodecylsulfate anion containing [Zn2Al] LDH. The dodecylsulfate anions adopt the same packing than in the sodium salt phase, with electrostatic and hydrogen bonding between anionic heads and OH layers and van de Waals bonding between lateral hydrocarbon chains. Guest-Guest interactions are then monitored by the layer charge density. This becomes a determining factor when one envisages adsorption of neutral molecules, alkyl amines and alcohol in alkylsulfate-containing LDH [108], catalysis reactions or in-situ polymerization [109]. As a consequence of recent interests for the use of anionic clays in environmental remediation processes, anion exchange properties have been
436
C. Forano
reinvestigated at the solid/liquid interface [110-113]. In such studies, anion exchange properties are examined over a wide concentration range of the incoming anions, including very low concentration conditions. This experimental approach allows to differentiate surface and bulk exchange processes and to quantify the rate of exchange surface sites. Ulibarri et al [4,35] and J. Inacio et al [32] showed that both the type of interlayer anion and the crystallinity of the minerals affect deeply the adsorption capacity. Adsorption experiments on [Mg-Al-CO3] and [Mg-Al-Cl] for deodecylsulfate conducted in exactly the same conditions led to 0 and 100% adsorption respectively. Molecular dynamics and ion diffusion studies at surfaces and interfaces of layered double hydroxides [33], showed that the structure and the composition of the mineral surfaces control the fluid structure at the interface and in the interlayer space of these minerals, as well as the effective diffusion coefficients of surface-sorbed species, their surface lifetimes, rotational and translational dynamics. Experimental adsorption isotherms are usually processed with classical Langmuir or Freundlich models developed for physisorption [32] in order to determine sorption capacity coefficients. For amphoteric anions such as glyphosate ([N-(phosphonomethyl) glycine]) two different sorbent/sorbate interactions were identified, the electrostatic adsorption and the ligand exchange. Adsorption is limited to the surface and the distribution coefficients (Kd) depend on the pH of the solution [114]. In a study of internal versus external uptake of anions, J. Boclair et al [34] showed that ferrocyanide does not displace carbonate from synthetic hydrotalcite but is taken up on the outside of the particles. Anion uptake is in this case controlled by specific hydrogen bonding requirements and not by charge density alone, a feature that can be used to control whether uptake will be both internal and external, or external only. 3.2 - Removal of Pesticides and related organic compounds by LDHs Adsorption of phenol Phenols have been classified by the U.S. Environmental Protection Agency as priority pollutants because of their high toxicity and their widespread use. For examples, nitrophenols are involved in the production of dyes, pigments, preservatives, pesticides, Pharmaceuticals and rubber chemicals [115] and triphenol is found in wastewaters of munition factories. Phenols and particularly highly chloro- and nitro- substituted phenols are present in environmental pH conditions of natural waters as soluble phenolate anions. The main works on the adsorption of phenols by LDHs were published by Cornejo et al [4,116-121]. Adsorption of trichlorophenol (TCP) and trinitrophenol (TNP) on hydrotalcite like compounds [MgR-Mm-X] (with MUI = Al, Fe, X = CO3 and Cl, R = 2, 3, 4) and 500°C calcined derivatives were studied under various i) pH conditions ranging from 2.0 to 13.0, ii) phenol concentrations and iii) solid/liquid ratio. The adsorption isotherms for the untreated adsorbents are of L type according to the classification of Giles [122], typical of monofunctionnal solutes adsorbing on specific sites with an adsorption capacity of Kf = 6.7 urnol.g"1 for TCP on [Mg3-Al-CO3]. The difference of pKa values between TCP (6.9) and TNP (3.8) explains that amount of TNP removed by [Mg3-Al-CO3] is much higher than that of TCP in term of CO32' anion displacement and whatever pH. The effect of pH on the isotherms and adsorption capacities indicates that phenols are adsorbed through an anion exchange mechanism. At high equilibrium TNP concentration, the expansion of the basal spacing from 0.783 nm to 1.352 nm evidences the intercalation of the anion within the hydrotalcite layers.
Environmental Remediation Involving Layered Double Hydroxides
437
El Shafei et al [123] recently pointed out that the compensating anion have a marked influence on the adsorption properties of [M"-MIU] LDH (M11 = Ni, Mg, Zn and M ln = Al, Cr, Fe) toward 4-chlorophenol. In all cases, at low equilibrium concentration, the adsorption of 4-chlorophenol (pH =11) occurred at the edges of the layer structure accompanied by C17OH" exchange. The increasing order in the amount of 4chlorophenol adsorbed was matched by an increasing order in the interlayer space due to substitution. Adsorption of pesticides The extended contamination of soils and ground water from the widespread use of pesticides in modern agriculture is a current concern that is impelling research looking for remedies. On the other hand, adsorption is one of the key processes in determining the fate of pesticides in soils, and the study of the adsorption of pesticides on cationic clays has been the subject of numerous papers [124-126]. Amongst the various types of pesticides, the important group of molecules with ionizable functions such as -OH, -CO2H, -SO3H (2,4-D, 2,4,5-T, MCPA, Dicamba, Mecoprop, Imazamox) undergo acidic dissociation leading to the formation of highly soluble anionic species in water. With the hydrophilic and positive caracters of their surface, anionic clays are very good sorbents for cationic and highly polar organic pesticides. Besse et al studied the adsorption of pesticides from the phenoxyacetic acids family (MCPA, 2,4-D and 2,4,5-T) by [Mg-Al] [32] and [Zn-Al] [127,128]. The adsorption characteristics of the herbicide MCPA [32] were evaluated with particular attention to the effect of the layer charge, original interlayer anion (CO 3 2 \ NO3", Cl"), pH and sorbent morphology. The adsorption isotherms are described by the Freundlich model (S-type) as for imazamox adsorption on analogous sorbents [128]. The adsorption capacity increases with the layer charge density. MCPA adsorption on LDHs occurs by an anion exchange mechanism in two steps, first anion exchange occurs at the surface then followed by an interlayer exchange. The adsorption capacity depends on the nature of the starting anions, following the affinity order: NO 3 ' < Cl' < CO32" proposed by Miyata [30] and increases with the specific surface area. Similarly, Dicamba adsorption [130] is affected by competing anions, increasing in the order: SO42" < HPO42" < CO32" < NO3" ~ F" » Cl" « Br" « I". Carbonate LDHs display very low sorptive properties whatever the pesticides MCPA [32] or Imazamox [129]. Even organo-LDHs (alkylsulfate-LDHs) with enhanced hydrophobic properties hardly adsorb Imazamox compared to cationic clays modified by alkyl ammonium. Desorption in presence of common inorganic anions is reversible. In the case of MCPA adsorption study [32], internal exchange of the anionic pesticide with expansion of the LDH basal spacing (2.21 nm) was observed at high equilibrium concentration, indicating that MCPA anion is intercalated perpendicular to the layers as expected [104]. For amphoteric anions such as glyphosate ([N(phosphonomethyl) glycine]) two different sorbent/sorbate interactions were identified, the electrostatic adsorption and the ligand exchange. Adsorption is limited to the surface and the distribution coefficients (Kd) depend on the pH of the solution [114]. Adsorption of nonpolar or hydrophobic pesticides (triadimefon, linuron, atrazine, acephate, diazinon) by hydrophilic layered double hydroxides is very low [130,131,131b]. Alternatively modification of LDHs by anionic surfactants intercalation or by thermal treatment (200°C - 500°C) will change drastically their adsorption properties (see following chapters).
438
C. Forano
NMR is a powerful method to assess interactions of xenobiotics (pesticides, PAH, hazardous compounds) with soil components including soil organic matter and clay fraction of soils. Interactions of pesticides with a hydrated synthetic hydrotalcite used as a soil model have been analyzed by using liquid state HR (high-resolution)MASNMR[132]. 3.3 - Dye removal The effluent discharged by various textile industries contains a large number of dyes, increasing the total COD of wastewaters [133-136]. Although most of dye are non-toxic, many of them complexe highly toxic metals (e.g. Cr) with harmful consequences to the aquatic life in rivers. Moreover, the persistence of color appearance (at concentration above 1 mg/L) in treated wasterwaters prevents their re-use. Because of stringent government legislations, improvment of separation processes have gained much importance. Hydrotalcite like compounds have demonstrated, in this environmental application too, high ability to removed color or dye by adsorption reaction [137-145]. Indeed, LDHs display high adsorption capacities for dye molecules and can be very competitive with other sorbents. As an example, the uptake amount of Remazol Blue by LDHs (125 mg/g sorbent) is intermediate than that of organo-clays (240 mg/g sorbent) and clay carbon composite material (52 mg/g) [140]. Orthman et al [144] gives the removal capacity of a hydrotalcite sorbent for a series of synthetic dyes (table 2), using 1.5 g of clay in 20 h contacted with 1 liter of a solution containing O.lg/L of dye. In this study, the authors show that hydrotalcite has an adsorption capacity toward acid blue 29 comparable with that of commercial activated carbon but exhibits stronger adsorption at low concentration. Table 2 - Adsorption capability of hydrotalcite for various synthetic dyes from ref. [144]. Dye Solution Acid Blue 29 EosinB Reactive Blue 48 Dispersed Red 1 Basic Blue 66 Basic Blue 9
Removal (%) 96 98 8 86 54 44
Dye charge Anionic Anionic Anionic Nonnionic Cationic Cationic
Owing to the calcined LDH ability to recover the original LDH structure under rehydration, regeneration of the sorbents is a valuable property for cost reduction of the remediation process. Takashi [143] studied dye adsorption from simulated wastewater liquids and pointed out the competing effect of additives (dyestuff (dispersant salts, acids, alkali), Na2CO3, surfactant, polyvynilalcool - PVA) on the adsorptivity of the LDH. A lower adsorbability was observed in real dye wastewater than in model wastewater. As already reported by Porter et al [146], the adsorption equilibrium for acidic dyes is best described by the Langmuir equation at low solid/liquid ratio (e.g. a 5 g/L). However, the reduction of colour and the increase of adsorption rate are proportional to the increase of solid loading in the solution, the decrease of other dissolved ionic substance concentration, the pH reduction and Temperature increase.
Environmental Remediation Involving Layered Double Hydroxides
439
Figure 5 - (a) X-ray powder diffraction patterns of Zn-Al-HTlc samples having increaing methyl orange exchange percentages: a) 4%, b) 9%, c) 46%, d) 70%, e) 84% from ref. [147], (b) computer-generated models showing the most probable arrangement of methyl orange anions between the hydrated HTLc layers from ref. [147]. A detailed kinetic study of the adsorption of a bifluoro-functional-azo reactive dye (Yellow LS-R) by a commercial hydrotalcite was recently realized by Lazaridis et al [145]. The authors concluded that the adsorption rate of color onto hydrotalcite particles is very sensitive to solid load, pH and ionic strength of the solution. Evidence is provided that the the sorption is a complex process that involved various kinetic models, including at least three mechanisms, i.e. external surface enhancement or film diffusion, adsorption and diffusion processes. Zhu et al [141] mentioned that the colored substances can be adsorbed on the surface or enter the interlayer region of the clay by anion exchange. Costantino et al [147,148] reported the uptake of Methyl Orange and Fluorescein anions by Zn-Al-Hydrotalcite. They investigated the structural aspects of the Host and Guest interactions showing that Methyl Orange anions (MO) easily replace the chloride anions with an amount of MO exchange nearly 94% (2.76 mmol/g) of the anion exchange capacity in the best experimental conditions (Figure 5a). In case of fluorescein uptake, saturation of the [Zn-Al] surface is reached at 50 \\mo\lg. The basal spacing of the Host Structure expands up to 2.42 nm and 1.65 run for Methyl Orange (Figure 5b) and fluorescein anions respectively. The Guest orientation within the galleries of the host structure are driven by both electrostatic Guest-Host interactions between anionic groups and OH planes and TI-TI Guest-Guest interactions. In the case of MO containing LDH, the anions lie perpendicular to the layers, leading to an highly expanded structure.
440
C. Forano
3.4 - Adsorption of humic substances Humic substances (HS) are one of the major fraction of organic matter present in natural upland surface waters (rivers, lakes, pounds). They display a large range of distribution in molecular weigt (from a few hundred to several thousands) and size, acidic properties and are highly soluble in water. They form stable water soluble complexes with heavy metals or organic pollutants. In natural pH conditions humic substances have an overall negative charge. Moreover, their preferential adsorption on mineral oxides and clays leads to very stable natural nanocomposite humic-clay complexes. Their presence is a problem facing the water industry and the water purification processes. The contamination of water by humic substances affect the removal process by reducing the adsorption capacity of the adsorbent toward the target pollutants. Due to their opposite charge and hydrophilic properties, anionic clays appear to be the best adsorbents for the removal of humic substances. Humic acids removal by hydrotalcite-like sorbing agents ([Mg3-Al-CO3], [Z114Al-Cl], [Mg3-Al-NO3], [Cug-Al-Cl]) was firstly reported by Cockett et al [97,98] in 1993. Due to their anion exchange properties, hydrotalcite like compounds are more effective in humic substances removal than clays and even pillared clays (PILCs) [149,150]. Seida et al [151] suggested that the adsorption mechanism proceeds via intercalation of the natural polymer within the LDH layers but on the contrary a X-ray diffraction study realized by Mohd et al [152] did not show any increase in the basal spacing when LDH or calcined LDH are contacted aqueous solutions of humic substances. The [Mg-Al-CO3] and 500°C calcined [Mg-Al-CO3] materials adsorb up to 15 mg HS per gram of adsorbent from synthetic and natural peat water. The HS uptake follows the Langmuir model of adsorption. 3.5 - Adsorption by pillared anionic clays Cyclodextrins are cyclic oligosaccharides with a doughnut-shape ring suitable for the micro-encapsulation of volatile or toxic organic compounds. Intercalation of anionic cyclodextrins in LDHs [153] results in the formation of novel functional materials with new adsorption characteristics. This new hybrid sorbent can be seen as a organic pillared LDH. While sorptions of trichloroethylene, tetrachloroethylene, benzene, toluene, xylene, ethylbenzene, trichlorobenzene by [Mg3-Al-NO3] are negligible, the cyclodextrine containing LDH display highest adsorption capacities toward the same target chemicals. The study also showed that because the organic compound adsorption is controlled by diffusion in the structure and therefore by the molecular size/shape factor, then the sorption affinity of the sorbent tends to decrease with the molecular diffusion volume. 3.6 - Adsorption by organoclays 3.6.1 - Adsorption/Intercalation of surfactants by LDHs : through organoclays adsorbents Aqueous colloidal solutions of amphiphilic molecules display high adsorption capacities of non-ionic or apolar organic contaminant and great interest for environmental decontamination. Moreover, adsorption of surfactants by mineral oxide or clay surfaces also leads to stable colloids which are involved in many cleaning processes for detergency, dispersion/flocculation and enhanced oil recovery. In such
Environmental Remediation Involving Layered Double Hydroxides
441
processes the adsorbents are easily separated from the clean medium. Layered Double Hydroxides display strong surface and interlayer hydrophilic properties due to high content of hydroxyl groups and water molecules. These properties limit the adsorption and the intercalation of nonionic or hydrophobic contaminants. However, anionic surfactants such as alkyl carboxylate, alkylsulfate or alkyl sulfonate are easily adsorbed and exchanged in LDH structures [100,101,112,154-157]. The formation of these organoclays arise from the self assembling of organic and inorganic layers, interfaced through negative head of the surfactant layers and the positive hydroxylated layers of the mineral host structure. Such architecture leads to a hydrophobic environment between the layers and at the surface of the particles. The modification of the hydrophilicity of LDHs open new properties for adsorption of pollutants. Dekany et al [158,159] compared the adsorption properties of both cationic surfactant/smectites and anionic surfactant/LDH antitype systems. The higher charge density of the anionic clays explains their higher adsorption capacities in terms of ion exchange capacity percentage. The change in basal spacing during dodecylsulfate (DS) adsorption by hydrotalcite appears even at low DS equilibrium concentration (0.5mmol." '). The interlayer distance increases from 0.78 nm to 3.2 run in agreement with the orientation of DS anions nearly perpendicular to the sheets. Pavan et al [160-162] studied the adsorption of dodecylsulfate, octylsulfate, dodecylbenzenesulfonate and octylbenzenesulfonate by [Mg-Al] LDHs. Adsorption experiments and electrokinetic measurements showed that adsorption of anionic surfactants occurs following a two-step mechanism. In the first step, adsorption of DS is forced by electrostatic interactions between positively charged LDH particles and negatively charged sulfates groups of DS. Inversion of the zeta potential is observed as expected [162,163]. The second step involves the van der Waals attractions between hydrophobic tails of adsorbed and free DS and the formation of intermediate hemimicelles or admicelles. Effect of pH and Temperature on adsorption were studied. The amount of adsorbed surfactant increases with temperature. Depending on the LDH materials either adsorption or adsorption / intercalation of DS are observed. Comparison of adsorption isotherms (typical L-type curves) [164,165] (You, 2002a & b) allows to establish the order of affinity of organosulfate and organosulfonate surfactants towards [Mg-Al] LDHs: dodecylsulfate (SDS) > 4octylbenzenesulfonate (SOBS) > dodecylbenzenesulfonate (SDBS) > octylsulfate (SOS). The X-ray diffraction analysis revealed that the surfactants can adopt various configuration within the layers, forming either monolayers or bilayers arrangements. As expected all the organoclays exhibit strong decreases in the BET surface properties down to 1-2 m 2 .g'\ These results point out the potential applications of LDHs for surfactant removal from aqueous solutions. 3.6.2 - Adsorption properties of surfactants modified LDHs The hydrophobic properties induced by intercalation of surfactants in anionic clays permits the adsorption of nonanionic and hydrophobic organic contaminants'1581. Organo-LDHs modified with dodecyl sulfate are as good as organo-montmorillonites in adsorbing hydrophobic pesticides [130,166] or related organic pollutants [164,165,167]. Kopka et al [154] work were one of the first to evidence the take up of organics, precisely long-chain alkanols, by intercalated [Zn-Cr- CnH2n+iSO4"] (n=6, 8,..., 18). The sorption of pesticides by organo-LDHs appears a very important challenge
442
C. Forano
for the environmental management of detoxification of polluted natural waters or slowrelease formulation of pesticides containing compounds in order to reduce the pesticides spread over. A comparative adsorption study [166] demonstrated that polar pesticides such as imazamox and triadimefon are much more adsorbed on organoclays than on the corresponding natural or inorganic minerals. The results of this work indicate that the diverse sorbents assayed may find application as filters in water decontamination for imazamox anion and as filters and supports for slow release formulations in the case of imazamox and triadimefon. The adsorption of organic molecule occur provided the interlayer spacing (dis) of the modified LDHs is greater than the size of the molecule otherwise the adsorbant can not accommodate the pollutant. Indeed, Dutta et al [167] showed that pyrene can be adsorbed in myristate- and hexanoate-LDHs (dis = 2.16 and 1.78nm respectively) while it cannot penetrate the structure of succinate-LDH (d;s = 0.75nm). Partition of organochlorine compounds (1,1,1-trichloroethane, trichloroethylene, tetrachloroethylene, 1,2,4-trichlorobenzene) in Mg-Al LDH phases [164,165] was proved to be more effective than in octanol. Adsolubilisation of organic molecules in dodecylsulfate (DS) containing LDHs was realised [163]. It was shown that amount of adsolubilized 2-naphtol increases with decreasing DS concentration and increasing feed concentration of 2-naphtol. [Mg-Al] LDHs is also efficient for perfluorooctanoic acid recovering from wastewater generated by emulsion polymerisation of fluoropolymer (PTFE) production plants [168]. 3.7 - Adsorption by calcined LDHs 3.7.1 - The calcinations-Reconstruction of LDHs The thermal decomposition of hydrotalcite like compounds proceeds in the following steps: > dehydration of weak bonded molecules with conservation of the layered structure (T < 200°C), > dehydroxilation of the structure simultaneously to the decomposition of volatil interlayered anions leading to amorphous mixed oxides with enhanced surface and porosity properties (200°C < T < 500°C), > crystallization of well defined MO or M n M ln 2 O 4 spinel oxides (T > 500°C). The intermediated amorphous oxides (step 2) can recover the original LDH structure when it comes in contact with solutions containing anions. This well known calcination-reconstruction process [169,170] can be used advantageously to extract anions from aqueous solutions. Thus many studies on wastewater treatments involved the used of calcined LDHs as sorbents. Only LDHs subjected to a complete reconstruction are interesting materials, i.e. [Mg-Al], [Ni-Al] or [Zn-Al] compounds because they involved both anion chemi- and physi-sorption. Indeed, under regeneration of the original anionic clay structure, anions are incorporated in the interlamellar domains in order to counterbalance the positive charge of the layer. Excess of absorption at the surface can also occur. 3.7.2 - Adsorption by calcined LDHs Crepaldi et al [111,113] recently compared the adsorption efficiency of calcined and uncalcined [Mg-Al-CO3] LDHs for terephtalate (TA) and benzoate (BA)
Environmental Remediation Involving Layered Double Hydroxides
443
anions. These molecules are commonly found in industrial wastes of organics (terephtalic acid) production plants and they are of major environmental concern. The study clearly demonstrates that the uptake of TA and BA by calcined materials, in concentration conditions usually found in industrial wastewaters, is much more efficient than for the uncalcined LDH. The adsorption proceeds in two steps, 1) a fast rehydration of the mixed oxide with intercalation of OH" anions, 2) a slow anion exchange of the OH" by the terephtalate anions. Up to 83% of TA and 85% of BA can be removed from the solutions by the calcined LDHs, which retain a high rate of adsorption (90%) over further decontamination cycles. The same adsorption mechanism was already suggested by Narita et al [110] for the adsorption of 2-naphthol-3,6-disulfonate and m-benzenedisulfonate by calcined [Mg-Al-CO3] and [Zn-Al-CO3] LDHs. In both cases adsorption isotherms fitted the Freundlich model. Competition between aromatic anions and OH" (pH = 9-10) limited greatly the adsorption of organics. 4 Catalytic remediation by LDHs 4.1 Catalytic properties of LDH uses for environmental remediation Application of catalytic engineering to the resolution of environmental problems is an alternative to usual destroying technology of incineration of environmentally undesirable organic compounds such as organo-halogenated or oxygenated aromatic compounds, constituents of many fungicides, pesticides and week destroyers. As catalysts or catalysts supports, hydrotalcite compounds are active materials for a large number of reactions [17] including hydrogenation reactions, polymerization, CH4 reforming or partial oxidation, DeNOx and DeSOx processes. They can be considered as good candidates for the development of catalytic remediation processes. Moreover, taking into account that the replacement of liquid bases by more friendly solid catalysts is an environmental target, calcined LDHs appear to be very promising substitutes. 4.2 - Catalytic decomposition of organic molecules As we have just discussed, Layered Double Hydroxides and calcined Layered Double Hydroxides display a wide range of catalytic properties for various organic molecule transformations. Therefore, they should have large potential applications for the degradation of organic pollutants, such as volatile organic compounds (VOC) or wastewater treatments. However, few studies [171-173] have been devoted so far to the catalytic degradation of toxic organics by these materials. Total oxidation Barhanowski et al [172] studied the catalytic combustion of toluene and ethanol by [Cu-Cr], [Cu-Al], [Zn-Cr] mixed oxides prepared by calcination of LDH parent materials at 873 K. All samples display very high catalytic activities. For the best catalyst i.e. [Cu-Cr] system, the importance of the Cu/Cr ratio and the role of the interface boundaries in the CuO-CuCr2O4 system on the catalytic performance were established. Palladium compounds derived from [Mg-Al] hydrotalcite also show strong activity in total oxidation of toluene, even better than the conventional palladium catalyst [172]. The catalyst precursor ([Mgo.75Alo 25(OH)2]°25+ [PdCl2(OH)2]0.125-x2'
444
C. Forano
(CO3)x2\nH2O) is prepared by coprecipitation of [Mg-Al] LDH with [PdCl4]2" which transformes under intercalation into [PdCl2(OH)2]2". In a similar approach, total oxidation of nitrogen containing VOC was performed by multimetallic hydrotalcites containing Cu, Cr, V, Al and Zn [173]. Partial oxidation of organic molecules Phenol compounds are of major environmental concerns because of high toxicity. Partial oxidation can be seen as a first step in the degradation of such pollutants. Alejandre et al [174] studied the oxidation of phenol by [Cu-Al] mixed oxides with Cu/Al ratio ranging from 0.5 to 3.0. In order to reach a steady state under catalytic conversion (~ 50%), it is necessary to treat the mixed catalyst in order to transform the unstable CuO in the more stable and active CuAl2O4 phase. Catalytic hydroxylation of phenol into catechol or hydroquinone, with H2O2 as oxidant, were also conducted on Cu, Ni, Co LDH derived catalysts [20,175,176]. Metal porphyrins and metal phtalocyanines are homogeneous biomimetic catalysts for the oxidation and even reduction of organic substrates. Immobilisation of these active molecules on inorganic supports, and more particularly in LDHs, improves their stability, their selectivity and their lifetime. Ukrainczyk et al [177] studied the reductive dechlorination of aqueous solution of CC14 by exchanged Cobalt tetrakis (Nmethyl-4-pyridiniumyi)porphyrin LDH catalyst while Chibwe et al [178] reported the use of layered double hydroxide supported cobalt(II) phthalocyanines as possible environmental remediation oxidation catalysts. The latter evaluated the oxidation of 2,6di-tert-Bu phenol and 1-decanethiol and demonstrated the higher activity of lower charge LDH catalysts for the remediation of contaminated ground water and industrial effluents. Regarding that some of these materials are very good catalysts for oxidation reactions of organics [179], they offer promising perspectives of environmental remediation. 4.3 - Catalytic decomposition of NOx and SOx Nitrogen and sulfur oxides are two major atmospheric polluants owing to their potential involvement in i) stratospheric destruction of ozone layer, ii) formation of acid rains and iii) being greenhouse gases [180-182]. The major anthropogenic contributions to total NOx and SOx emissions arises from power generation stations (30% and 65% of NOx and SOx emissions respectively), petrochemical plants (10% and 7% of NOx and SOx emissions respectively) such as fluid catalytic cracking (FCC) units, adipic and nitric acid production plants and automotive exhaust emissions. As an example, N2O with a relatively long average life time of 150 years and a net greenhouse effect of about 300 times greater than CO2, has an atmospheric concentration which increases of about 0.37% per year mainly due to human activities. In addition to industrial emissions of N2O, agricultural sources of pollution due to anthropogenic nitrogen fixation are becoming noticeable [183,131b]. In order to reduce anthropogenic emissions of harmful gases, development of efficient technologies for emission control is a priority. The catalytic decomposition and the selective catalytic reduction (SCR) are so far the two promising processes to reach such environmental requirements. Reduction of NOx can be achieved either by a thermal catalytic decomposition following the simple scheme: 2NO -> N2 + O2
Environmental Remediation Involving Layered Double Hydroxides
445
or by catalytic reaction with selective reductors such as H2 or volatil hydrocarbons: (3n+l)NO + CnH2n+2 -> (3n+l)/2N2 + nCO2 + (n+l)H2O The removal of SOx can proceed either via an oxidative process through the conversion of SO2 into metal sulfate by reaction with basic oxides: SO2 + l/2O2 -> SO3 MOX + SO3 -» MXSO4 or via a reductive pathway under hydrogen leading the formation of H2S: SO2 + 3H2 - H2S + 2H2O or MXSO4 + 3H2 -> MXO + H2S + 3H2O The state of the art on the catalytic decomposition of N2O has been described in some reviews [184]. For the catalytic removal of NOx, various catalysts have been used including supported and unsupported metals, pure and mixed oxides and zeolites such as perovskite oxides [185,186], spinel oxides [187], ZnO [188], CeO2 and Ce3O4 [189], Rh containing zeolites [190,191,191b] or supported oxides [190,191,191b]. Zeolitic systems are very active for decomposition at low temperatures but their main limitation is due to an effective poisoning by different other gases (mainly SO2) which can be present in the gas stream. Some pure oxides display a high catalytic activity in nitrogen oxide decomposition but their low specific surface areas drastically reduce their industrial applications. Since few years, many groups focus their research on the study of the catalytic activity of mixed oxides prepared from the calcination of hydrotalcite like compounds [192-194]. The ability of the LDH precursors to accomodate in the layers a great variety of divalent and trivalent metal cations, homogeneously distributed at an atomic scale level and their low thermal stability leads to the formation of mixed oxides with tunable chemical composition and high BET surface areas. NO and N2O decomposition. The [Cun-Mg"-AlIn] mixed oxides system was one of the first investigated for the catalyic decomposition of Sox [195] and Nox [196] from the exhaust gas of the FCC units. The active catalyst is prepared by reduction of the calcined [CuII-MgII-Alm] LDH (M'VAI111 = 3 and 5% molar Cu11) in propane [196]. It is composed of a dispersion of Cu° metallic nanoparticles embeded in an amorphous matrix of MgO. The decomposition of NO in air by the catalyst is maximum at 750°C but it never exceeds 80% of total conversion and it decreases with operating time while under reducing conditions (addition of propane in the gas stream) total reduction of NO is obtained. Cu(0) for NO decomposition and Cu(0) and/or Cu(I) for NO reduction with a hydrocarbon were identified as the active sites of the catalyst by an in situ XPS/XAES study [197]. The ability of these materials to redisperse the copper active sites without destroying the structure of the material makes them different from other catalysts such as zeolites, and
446
C. Forano
is essential for suitable catalytic behavior at high temperatures. [Mg-Co] mixed oxides derived from LDH were also investigated [198]. In this case, decomposition of the double hydroxides into amorphous mixed oxides (MgO, CO3O4, Co3_xMgxO4) catalyst is performed at 350°C under N2O/He gas flow. Under catalytic conditions (N2O/He 1.5 dmV 1 gas flow, GHSV 3000 h"1), approximatively 6 moles of N2O per Kg of catalyst can be decomposed at 35O°C within lh, which is comparable to some of the most active catalysts. A higher activity is clearly attributed to the larger number of Mg-O-Co sites which is determined by the Mg/Co molar ratio in the LDH precursor. In order to control the Co3+/(M2++M3+) amount in the catalyst, [MgII-Co"-Com] LDH precursors can be prepared by in-situ generation of Co3+ from coprecipitation of Mg2+-Co2+ mixed salts [199,200]. From 25% up to 57% Co3+ molar content was introduced in the materials. The stabilizing effect of Mg2+ delays the dehydroxylation to higher temperature and increases the specific surface area of the calcined derivatives. Moreover, according to Perez-Ramirez et al [201,202] the presence of Mg in Ni and Co containing mixed oxides prevents the deactivation of the catalysts due to the preferential adsorption of nitrogen and sulfur oxides on MgO. The highest N2O conversion (6.2 mmol/g.h i.e. 70% at 400°C) is observed for the catalyst with the highest Co/(Mg+Co) ratio (0.75) and the lower Co3+/Co2+ ratio, prepared from the calcination (400°C) of the [Co2+2-Mg2+-Co3+] LDH precursor. This is explained in terms of variation in degree of inversion of the MgxCo3_xO4 spinel phase. In the decomposition reaction of NO and N2O, the desorption of O2 has been identified as the limiting step. In cases of cobalt and rhodium containing catalysts [203], the reaction mechanism of N2O decomposition involves several adsorbed states of oxygen as studied by TEOM microbalance. Energies of activation and rate coefficients were estimated for the main reaction steps of N2O decomposition at various N2O, O2, and water partial pressures by thermally treated CoLaAl-hydrotalcite catalysts [204]. Nano-size (5-7 run) supported ConAl-hydrotalcite-like catalysts on y-Al2O3 were synthesized [205] with a coprecipitation method in which the y-Al2O3 acts as a source of Al for coprecipitation and a support for the compound formation. Decomposition into the active supported spinel phase (CoIICo2.xmAlx04/Y-Al203) occurs then at much lower temperature (210°C). Kannan et al [194,206-209] studied the effect of the divalent metal and the M n /M ra ratio on the decomposition of N2O in industrial process streams simulated conditions. A [Mn-Mra] mixed oxides serie with M n = Mg, Co, Ni, Cu, Zn and M ln = Al, Fe, Cr was examined. Obalova et al [210] and Perez-Ramirez et al [201,202] recently confirmed the high catalytic performances of both [Mg-Co] and [Mg-Ni] materials. From kinetic data, the effect of the trivalent cation composition of the calcined LDH materials on activity was established in the following order: [M-Cr] < [M-Fe] < [M-Al] where M = Ni or Co. Over the [Mu-Alnl] serie the Ni and Co containing catalysts are the most active while the Mg and Zn analogues showed poor catalytic activities. Temperature required for 50% conversion of N2O increases in the series: Co-Al < Ni-Al < Cu-Al < Mg-Al. The results confirm the better activity of spinel like mixed oxides over pure divalent or trivalent metal oxides, which need higher temperature operating conditions for similar conversion percentages. The activity increases linearly with increase in the active surface metal ion concentration. However, the electronic states of the different metal cations are not assessed. In term of catalyst life-time, the best material ([Co-Al]) displays an unaffected activity after 175h time-onstream showing the viability of calcined LDHs catalysts for industrial exploitation.
Environmental Remediation Involving Layered Double Hydroxides
447
In order to improve the various properties (stability, life-time, activity) that makes the catalyst the ideal candidate for economical developement, it appears crucial to find the best chemical composition of the oxides in terms of active sites and adsorption properties. Because amongst the highest active materials for decomposition of nitrogen oxides, we found La, Ru, Rh, Pd, Pt, In supported on silicate or pure oxides, mixed oxides prepared from calcined LDHs containing non common metal cations [211-219] were tested for their catalytic activities on nitrogen oxides decomposition. Attention has been paid recently on the activity of Rh, Pd, La on supported Co containing mixed oxides [201,202,220]. According the authors, all precursors coprecipitated from different combinations of (Co, Mg, Al, Rh, La, Pd) salts crystallized in a single phase with a hydrotalcite structure. Addition of divalent and trivalent metal to the Co-Al catalyst leads to a strong improvement in the N2O decomposition. In particular, it is revealed that the less expensive Pd containing catalyst is more stable and as active as the Rh containing one provided its loading is 50 times higher than for the latter. However, interestingly, with such a high active sites content the catalyst becomes less sensitive to poisoning. Both Co-Rh,Al and Co,Pd-La,Al mixed oxide catalysts with Mg incorporated in the structure show any inhibition to SO2 and O2 in simulated FBClike flue-gas conditions at high temperature conditions because Mg-Al spinel acts as a SO2 scavenger. NO and N2O Selective catalytic Reduction. Catalytic decomposition of N2O has been studied more than the Selective Catalytic Reduction. However, the presence of other gases (O2, H2O, SO2) in realistic gas stream conditions often inhibited the catalysts activity and affect the nitrogen oxide decomposition rate [184,191]. N2O reduction by CO [221], hydrocarbons [222] and NH3 [223] have been examined. The NOx storage-reduction technique is developped for the treatment of exhaust gases [224-226]. Novel NOX storage-reduction catalysts for diesel-duty engine emissions with improved NOx storage and resistance to SO2 deactivation were prepared by impregnation of Pt or Cu,Pt on [Mg-Al] LDH precursor supports [225,226]. A dualbed catalytic system was developed by Perez-Ramirez et al [224] in which NOx and N2O are successively removed in two stages (i: NOX-N2O, ii: N2O-N2) from flue gas by selective catalytic reduction with propene over a activated carbon-supported Pt catalyst. Catalytic reduction of NO by ammonia as reducing agent on catalysts derived from LDHs were also investigated [227,228,228b]. Catalysts [227,228b] prepared from [Mg-Al] LDHs were Al3+ is partially substituted by V3+ and Mg2+ by Fe2+, showed up to 87% conversion at 65 8K in oxygen excess conditions. [Mg-Al] hydrotalcite-derived polyoxovanadate-intercalated (V10O286", V2O74") catalysts [228] display a satisfactory catalytic activity at high temperature complementary with classical Cu-Mg-Al hydrotalcite catalysts, whose performance prevailed in low temperature conditions. SOX decomposition. Sources of antropic sulfur oxide emissions are numerous, consequently appropriate technologies must be developped for each type. Indeed, if successful processes have been developped to reduce SOX effluent from thermal power stations [229] they are not relevant for emission removal by Fluid Catalytic Cracking (FCC) units because of very different operating conditions [222]. Corma et al have shown that Cu-Mg-Al Hydrotalcite based catalysts are not only efficient for SOX removal in FCC
448
C. Forano
units [195] but also for the simultaneous removal of SOx and NOx at low oxygen concentration [222]. In the presence of propane, SO2 is removed as H2S, under a reductive pathway, leading to the formation of sulfur based Cu(I) catalytic sites active for the reduction of NO. Then, both contaminants are removed from the stream. In the case of Co-Mg-Al systems addition of an oxidant (CeO2) is necessary in order to oxidize SO2 to SO3 [230]. LDHs or calcined LDHs materials can be used to purify S oxides-containing gases. They can serve as S oxides binding material for the conversion of sulfur containing feddstaocks [231]. Vanadium cations supported by LDHs have been proved to be active species for SO2 oxidation [227]. Presence of both V3+ and V5+ was evidenced by XPS and the catalytic conversion rate was directly related to the amount of vanadium in the material. DeSOx additives based on closed materials (Mg-Al-V-Ce hydrotalcite derivatives) [232] display a high adsorption capacity at the FCC operating conditions and also good regeneration ability. 4.4 - Photocatalytic decomposition by LDH's Photochemical remediation is an alternative process to chemical degradation particularly for water purification or control of toxic air contaminants. The main advantage of the photocatalytic detoxification is that it leads to total degradation of organics into harmless inorganic residus under mild conditions. Polyoxometalates display similar electronic features than metal oxides semiconductors, generating under irradiation OH radicals, suitable for photooxidation of organic molecules. Immobilisation of W7O246~ anions in LDH host structures leads to insoluble photocatalysts materials with an homogeneous dispersion of active species over the solid and higher specific surface areas compared to pure POMs. Intercalation of POMs in LDHs is obtained by direct anion exchange reaction on expanded terephtalate containing LDH provided the pH of the solution is adjusted to the existing conditions of the oxometalate in solution [233,234]. Guo et al demonstrated the photocatalytic degradation of aqueous organochlorine pesticides (hexachlorocyclohexane or HCH) by Polyoxometalates (POM) intercalated [Mg2-Al-W7O24] LDH (Guo, 2001) and calcined POM containing LDH's ([Zn2-Al-SiWn039] and [Zrij-Al-SiWuOjsMn^O)]) [235]. Complete mineralization of HCH was proved by the recovery of nearly 98% of chlorine. The mechanism of the organic molecule photooxidation involves an intramolecular O(2p)-W(5d) charge transfer, the formation of the interlayered excited state [W7O246"]* species and the generation of the strong and unselective photoactive OH radicals. The photochemical degradation is assisted by the ability of the solid catalyst to adsorb the reactants (organics and O2) and the photon. Calcination (600-700°C) of [Zn2-AlSiW u O 39 ] and [Zn2-Al-SiWi,O39Mn(H2O)]) [235] POM -LDHs systems into mixed oxides leads to an improvement of the photocatalytic activity because of the presence of photoactive ZnAl2O4 and ZnWO4 spinel compounds with lower band gaps compared to the pure products. Surface and porosity properties do not play a major role in the degradation efficiency. In a more reasonable approach, the transformation of organics in less toxic residues by heterogeneous photocatalysts can be preferred. Oxidation by soft and selective agents can be a first step through the transformation hi more valuable molecules. F. Van Laar et al [236] have evidenced the photocatalyic activity of [Mg-AlMoO4] LDHs for the peroxidation of methylcyclohexene, dimethylbutene and
Environmental Remediation Involving Layered Double Hydroxides
449
a-terpinene. Indeed, under irradiation of ligth the LDH materials react on H2O2 to generate the photoactive single state ('02) without the need of soluble base due to the basic characteristic of the support. Here again the performances of the catalyst can be improved by the tuning of the LDH properties (composition, hydrophilic properties etc.). Nanocomposites based on the association of co-exchange metal porphyrin [LiAl] organo-LDH with titanium oxide particles [237] have been prepared for potential applications in photooxidative remediation of aqueous organic pollutants. 5 - Conclusions To date, lots of investigations have been carried out to show that Layered Double Hydroxides and their derivatives compounds display suitable properties for remediation of environmental pollutants (heavy metals, toxic gases, organic pollutants, etc.). They appear to be competitive candidates for adsorption or catalytic processes. However, more efforts must be made in the future to transfer these results from laboratory scale to real environmental conditions and industrial developments. 6 - References [I] N. Bejoy, Hydrotalcite: the clay that cures, Resonance, 6 (2001) 57. [2] C.S. Swamy, S. Kannan and S. Velu, Hydrotalcite like Materials: Their possible role in Environmental Control, Main Goup Elements and Their Compounds, V.G. Kumar Das (Edt.), Narosa Publishing House, New Delhi, (1996) 112. [3] M.J. Hudson, Extraction of priority pollutants using inorganic ion exchangers Mineral Processing and the Environment, NATO ASI Series, Series 2: Environment 43 (1998) 223. [4] M.A. Ulibarri and M.C. Hermosin, Layered Double Hydroxides in Water Decontamination, in Layered Double Hydroxides: present and Future, Ed. V. Rives, Nova Sc. Publ. Inc., New York, (2001) 251. [5] Y. You, Use of layered double hydroxides and their derivatives as adsorbents for inorganic and organic pollutants, Avail. UMI, Order No. DA3053088, Diss. Abstr. Int. B, 63 (2002) 2279. [6] T. Hibino, Nendo Kagaku, 42 (2003) 139. [7] S. Velu, K. Suzuki, M. Okazaki, T. Osaki, S. Tomura and F. Ohashi, Chem. Mater., 11(1999)2163. [8] S. Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki and F. Ohashi, J. Catal., 194 (2000) 373. [9] D. Tichit, N. Das, B. Coq and R. Durand, Chem. Mater., 14 (2002) 1530. [10] F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 11 (1991) 173. II1] S. Miyata, Zeoraito, 8 (1991) 7. [12] A. de Roy, C. Forano, M. El Malki and J-P. Besse, Anionic clays: trends in pillaring chemistry. In: M.L. Occelli and H.E. Robson (Editors), Synthesis of microporous materials, Expanded Clays and Other Microporous Solids. Van Nostrand Reinhold, New York, (1992) 108. [13] G. Mascolo, Appl. Clay Sci., 10 (1995) 21. [14] F. Trifiro and A. Vaccari, In: Atwood J.L., Davies J.E.D., MacNicol D.D., Vogtle F. Eds., Comprehensive Supramolecular Chemistry, Vol. 7. Pergamon, Oxford, UK, (1996)251. [15] V. Rives, M.A. Ulibarri, Coord. Chem. Rev., 181 (1999) 61.
450
C. Forano
[16] T. Hibino, J. Uchisawa and A. Tsunashima, Shigen Kankyo Gijutsu Sogo Kenkyusho Hokoku, 28 (1999) 1. [17] A. Vaccari, Applied Clay Science, 14 (1999) 161. [18] A. Vaccari, Hydrotalcite-type anionic clays (layer double hydroxides) as precursors of both catalysts and catalyst supports, Advances in Science and Technology (Faenza, Italy), (1999) 16 (Ceramics: Getting into the 2000's, Pt. D), 571-584. [19] B.F. Sels, D.E. De Vos and P.A. Jacobs, Catal. Rev. - Sci. Engin., 43 (2001) 443. [20] V. Rives, Layered Double Hydroxides, Editor(s): Rives, Vicente, Publisher: Nova Science Publishers, Inc., Huntington, N.Y, (2001). [21] S.P. Newman and W. Jones, Layered double hydroxides as templates for the formation of supramolecular structures, Editor(s): Jones, W.; Rao, C.N.R., Supramolecular Organization and Materials Design (2002), 295. [21b] A. de Roy, C. Forano and J-P. Besse, Layered Double Hydroxides: synthesis and post-synthesis modification in Layered Double Hydroxides: Present and Future, Edt. Rives V., Nova science publishers, New York (2001) 1. [22] E. Lopez-Salinas, E. Torres-Garcia and M., Garcia-Sdnchez, J. Phys. Chem. Solids, 58(1997)919. [23] I. Rousselot, C. Taviot-Gueho, F. Leroux, P. Leone, P. Palvadeau and J-P. Besse J. Solid State Chem. 167 (2002) 137. [24] J. Shen, M Tu and C. Hu, J. Solid State Chem., 137 (1998) 295. [25] D. Tichit, A. Rolland, F. Prinetto, G. Fetter, M.J. Martinez-Ortiz, M.A. Valenzuela and P. Bosch, J. Mater. Chem., 12 (2002) 3832. [26] M.L. Occelli, J.P. Olivier, A. Auroux, M. Kalwei and H. Eckert, Chem. Mater. 15 (2003)4231. [27] M. Adachi-Pagano, C. Forano, J-P. Besse, J. Mater. Chem., 13 (2003) 1988. [28] M. del. Arco, V. Rives and R. Trujillano, Surface and textural properties of hydrotalcite-like materials and their decomposition products, Studies in Surface Science and Catalysis (Characterization of Porous Solids III) 87 (1994) 507. [29] F. Malherbe, C. Forano and J-P. Besse, Microp. Mat., 10 (1997) 67. [30] S. Miyata, Clays clay miner., 32 (1983) 305. [31] L.M. Parker, N.B. Milestone and R.H. Newman, Ind. Eng. Chem. Res., 34 (1995) 1196. [32] J. Inacio, C. Taviot-Gueho, C. Forano and J-P. Besse, Applied Clay Sci., 18 (2001) 255. [33] A.G. Kalinichev, R.J. Kirkpatrick and R.T. Cygan, Amer. Mineral. 85 (2000) 1046. [34] J.W. Boclair, P.S. Braterman, B.D. Brister, J. Jiang, S. Lou, Z. Wang, and F. Yarberry, Origin of Life and Evolution of the Biosphere, 31 (2001) 53. [35] M.A. Ulibarri, I. Pavlovic, C. Barriga, M.C. Hermosin and J. Cornejo, Appl. Clay Sci., 18(2001)17. [36] W-X. Zhang, H. Sakane, T. Hatsushika, N. Kinomura and N. Suzuki, Takashi Muki Materiaru, 4 (1997) 132. [37] Y. Seida and Y. Nakano, J. Chem. Engin. Jpn., 34 (2001) 906. [38] Y. Seida and Y. Nakano, Water Res., 36 (2002) 1306. [39] R. Nomura, T. Mori and E. Kanezaki, Intern. J. Modern Phys. B, 17 (2003) 1458. [40] T. Kitao, Y. Kiso, H. Imai and T. Hosoya, Jpn. Kokai Tokkyo Koho, 4 (2003), JP 2003299948 A2 20031021 Application: JP 2002-104726 20020408. [41] G. A. O'Neill, J.W. Novak and E.S. Martin, (Aluminum Co. of America, USA),
Environmental Remediation Involving Layered Double Hydroxides
451
Method for removing arsenic or selenium from an aqueous solution containing a substantial background of another contaminant, U.S. (1990), 10 pp. US Patent 4935146 A 19900619 [42] M. Tsuji, J. Matsunami and Y. Tamaura, Trans. Mat. Res. Soc. Jpn, 24 (1999) 357. [43] S. Miyata, M. Taketomi, T. Kojima, S. Tanizaki, A. Hashimoto, K. Kamishiro, M. Eguchi and S. Suzuki, Satoru. Jpn. Kokai Tokkyo Koho (2000), 5 pp. Patent No. JP 2000233188 [44] Y. You, G.F. Vance and H. Zhao, Appl. Clay Sci., 20 (2001) 13. [45] J. Das, D. Das, G.P. Dash and K.M. Parida, J. Coll. Interf. Sci., 251 (2002) 26. [46] C.R. Bryan, Y. Wang, H. Xu, P.S. Braterman and H. Gao, Mat. Res. Soc. Symposium Proceedings (Scientific Basis for Nuclear Waste Management XXVI) 757 (2003) 671. [47] F. Kovanda, E. Kovacsova and D. Kolousek, Scripta Geol., 28-29 (2000) 155. [48] G.N. Manju and T.S. Anirudhan, Indian J. Environ. Health, 42 (2000) 1. [49] Y.W. You, H.T. Zhao and G.F. Vance, Environ. Technol., 22 (2001) 1447. [50] N.K. Lazaridis, A. Hourzemanoglou and K. A. Matis, Chemosphere, 47 (2002) 319. [51] B. Dousova, V. Machovic, D. Kolousek, F. Kovanda and V. Dornicak, Water, Air, and Soil Pollution, 149 (2003) 251. [52] S.W. Rhee, M.J. Kang, H. Kim and C.H. Moon, Environ. Technol., 18 (1997) 231. [53] R.L. Goswamee, P. Sengupta, K.G. Bhattacharyya and D. K. Dutta, Appl. Clay Sci., 13(1998)21. [54] B. Houri, A. Legrouri, A. Barroug, C. Forano and J-P. Besse, Coll. Czech. Chem. Comm., 63 (1998) 732. [55] B. Houri, A. Legrouri, A. Barroug, C. Forano and J-P. Besse, J. Chim. Phys. Phys. ChimieBiol., 96(1999)455. [56] M. Lehmann, A.I. Zouboulis and K.A. Matis, Chemosphere, 39 (1999) 881. [57] G.N. Manju, M.C. Gigi and T.S. Anirudhan, T. S., Indian J. Chem. Tech., 6 (1999) 134. [58] F. Kovanda, E. Kovacsova and D. Kolousek, Coll. Czech. Chem. Comm., 64 (1999) 1517. [59] N.K. Lazaridis, K.A. Matis and M. Webb, Chemosphere, 42 (2000) 373. [60] N.K. Lazaridis and D.D. Asouhidou, Water Research, 37 (2003) 2875. [61] Y. Seida, Y. Nakano and Y. Nakamura, Water Research, 35 (2001) 2341. [62] T.L. Grady, T. Hess and H.G. McWhinney, Proceedings - NOBCChE, 22 (1995) 115. [63] K.A. Tarasov, D. O'Hare and V.P. Isupov, Inorg. Chem., 42 (2003) 1919. [64] A. Abdelouas, J.L. Crovisier, W. Lutze, B. Fritz, A. Mosser and R. Mueller, Clays Clay Miner., 42 (1994) 526. [65] C. Altenhein-Haese, H. Bischoff, L. Fu, J. Mao and G. Marx, J. Alloys and Comp.. 213-214(1994)554. [66] G. Marx, C. Altenhein-Haese, H. Bischoff, J. Engelhardt and F. Feldmaier, J. Radioanal. Nucl. Chem., 194 (1995) 95. [67] J. Serrano, V. Bertin and S. Bulbulian, J. Radioanal. Nucl. Chem., 254 (2002) 91. [67b] M.J. Kang, K.S. Chun, S.W. Rhee and Y. Do, Radiochimica, 85 (1999) 57. [68] M.J. Kang, S.W. Rhee, H. Moon, V. Neck and Th. Fanghaenel, Radiochim. Acta, 75 (1996) 169. [69] S.D. Balsley, P.V. Brady, J.L. Krumhansl and H.L. Anderson, J. Soil Contamin., 7
452
C. Forano
(1998) 125. [70] J. Serrano, V. Bertin and S. Bulbulian, Langmuir, 16 (2000) 3355. [71] M.T. Olguin, P. Bosch, D. Acosta and S. Bulbulian, Clays Clay Miner., 46 (1998) 567. [72] E. Narita, Y. Umezu and S. Takahashi, Jpn. Kokai Tokkyo Koho (1999), 10 pp, JP 11084084 A2 19990326 Heisei, Application: JP 97-251469 19970902. [73] G. Fetter, M.T. Olguin, P. Bosch, V.H. Lara and S. Bulbulian, J. Radioanal. Nucl. Chem., 241 (1999) 595. [74] T. Toraishi, S. Nagasaki and S. Tanaka, Appl. Clay Sci., 22 (2002) 17. [75] M.J. Kang, S.W. Rhee and P.S. Hahn, Hwahak Konghak, 39 (2001) 517. [76] T. Yamamoto, T. Kodama, N. Hasegawa, M. Tsuji and Y. Tamaura, Energy Conversion and Management, 36 (1995) 637. [77] M.A. Easley and W.E. Horn (Aluminum Company of America, USA), U.S. Carbon dioxide adsorption of synthetic meixnerite. (1999) 10 pp., Cont.-in-part of U.S. 5,645,810, US 5882622 A 19990316, Application: US 97-842934 19970409. Priority: US 95-473828. [78] Y. Ding and E. Alpay, Chem. Engin. Sci., 55 (2000) 3461. [79] Y. Ding and E. Alpay, Process Safety and Environmental Protection, 79 (2001) 45. [80] S. Kikkawa, Ceramic Transactions, 112 (2001) 589. [81] Z. Yong, V. Mata and A.E. Rodrigues, Separation and Purification Technology, 26 (2002) 195. [82] Z. Yong, and A.E. Rodrigues, Energy Conversion and Management, 43 (2002) 1865. [83] TJ. Pinnavaia, J. Amarasekera and C.A. Polansky, (Michigan State University, USA), Layered double hydroxide sorbents for the removal of sulfur oxides from flue gas and other gas streams. PCT Int. Appl. (1991) 53 pp., WO 9110505 Al 19910725. Application: WO 91-US343 19910117. Priority: US 90-466984 19900118. [84] TJ. Pinnavaia and J. Amarasekera, (Michigan State University, USA), Layered double hydroxides (LDH) as low temperature recyclable sorbents for removal of SOx from flue gases, U.S. (1994), 11 pp., US 5358701 A 19941025, Application: US 92-976543 19921116. [85] E.W. Albers and H.W. Burkhead Jr., (Contract Materials Processing, Inc., USA), Hydrotalcite sorbent for sulfur oxide removal PCT Int. Appl. (1997), 18 pp. WO 9748480 Al 19971224, Application: WO 97-US9809 19970619. Priority: US 96-20126 19960620. [86] M-A. Poirier and C. Detellier, (USA), Sulfur removal from hydrocarbon fluids by contacting said fluids with hydrotalcite-like adsorbent material. U.S. (1999), 4 pp. US 5951851 A 19990914 US Patent N° Application: US 97-961612 19971031. [87] M-A. Poirier, (Exxon Research and Engineering Co., USA), Sulfur removal from hydrocarbon fluids by mixing with organo mercaptan and contacting with hydrotalcite-like materials, alumina, bayerite or brucite. U.S. (2000), 6 pp., Cont.-in-part of U.S. 5,958,510. US 6027636 A 20000222, Application: US 98-131107 19980807. Priority: US 97-961612 19971031. [88] I. Harada, M. Aritsuka and M. Hoshikawa, (Mitsui Toatsu Chemicals, Japan), Agents for treating silane-containing gases and their use, Jpn. Kokai Tokkyo Koho (1996), 6 pp. JP 08318131 A2 19961203 Heisei. Patent Application: JP
Environmental Remediation Involving Layered Double Hydroxides
453
95-128290 19950526. [89] T. Watanabe, M. Yamada, A. Morita, H. Kikuchi and F. Endo, (Nippon Oxygen Co Ltd, Japan), Treatment of gases and the treating agents. Jpn. Kokai Tokkyo Koho (1996), 5 pp. Patent JP 08206432 A2 19960813 Heisei. Application: JP 95-17148 19950203. [90] F. Schmidt, J. Ladebeck and C. Heinisch, (Sued-Chemie A.-G., Germany, Absorption/oxidation agent for removing environmental pollutant gaseous hydrides and/or halogen compounds, Ger. Offen. (1998), 8 pp., DE 19626421 Al 19980108, Application: DE 96-19626421 19960701. [91] K. Okada A. Kaneda and Y.A. Kameshima, Mater. Res. Bull., 37 (2002) 209. [92] B.K.G. Theng., The chemistry of clay organic reactions, Wiley & Sons, New York (1994). [93] M. Sanchez-Camazano, M.J. Sanchez Martin, Water Air Soil Pollut, 74 (1994) 19. [94] M.J. Carrizosa, W.C. Koskinen, M.C. Hermosin and C. Cornejo, Appl. Clay Sci., 18(2001)223. [95] I. Pavlovic, M.A. Ulibarri, M.C. Hermosin and J. Cornejo, Fresenius Environ. Bull., 6(1997)266. [96] Kyowa Chemical Industry Co., Ltd., Japan, Adsorbent, Jpn. Kokai Tokkyo Koho (1983), 5 pp. JP 58214338 A2 19831213 Showa., Application: JP 82-94792 19820604. [97] K.R.F. Cockett, M. Concannon and M. Webb, (Crosfield, Joseph, and Sons, UK), Absorption of toxic organic compounds from water using hydrotalcite-like materials, Eur. Pat. Appl. (1993), 12 pp., EP 541358 Al 19930512, Application: EP 92-310113 19921104. Priority: EP 91-310191 19911105. [98] K.R. Cockett, M. Concannon, R.M. Hunter, A.L. Lovell, A. Nock, M. Webb and R.T. Whalley, Sorption water purification with hydrotalcite-like sorbing agents (Crosfield Joseph and Sons Ltd., UK) Eur. Pat. Appl. (1993), 34 pp, EP 566260 Al 19931020, [99] F.S. Williams and A. J. Perrotta, Light Metals, (1998) 81. [100] M. Meyn, K. Beneke and G. Lagaly, Inorg. Chem., 32 (1990) 1209. [101] M. Meyn, K. Beneke and G. Lagaly G., Inorg. Chem., 29 (1990) 5201. [102] H. Nijs, A. Clearfield and E.F. Vansant, Microp. Mesop. Mat., 23 (1998) 97. [103] V. Prevot, C. Forano and J-P. Besse J-P., Inorg Chem., 37 (1998) 4293. [104] V. Prevot, C. Forano and J-P. Besse, Applied Clay Science, 18 (2001) 3. [105] S. Aisawa and E. Narita, Zeoraito, 17 (2000) 101. [106] J. Sanz and J.M. Serratosa, J. Am. Chem. Soc, 106 (1987) 4790. [107] U. Costantino, M. Casciola, L. Massinelli, M Nocchetti and R. Vivani, Solid State Ionics, 97 (1997) 203. [107b] S. Bonnet, C. Forano, A. de Roy, J-P. Besse, P. Maillard and M Momenteau, Chem. Mater., 8(1996)1962 [108] H.P. Boehm, J. Steinle and C. Wieweger, Angew. Chem. Int. Ed. Engl., 16 (1997) 265. [109] F. Leroux and J-P. Besse, Chem. Mater., 13 (2001) 3507. [110] E. Narita, T. Yamagishi and K Suzuki, Nippon Kagaku Kaishi, 6 (1992) 676. [ I l l ] E.L. Crepaldi, J. Tronto, L.P. Cardoso, J.B. Valim, Colloids Surfaces A: Phys. Engin. Aspects, 211 (2002) 103. [112] E.L. Crepaldi, P.C. Pavan, J. Tronto and J.B. Valim, J. Coll. Interf. Sci., 248 (2002) 429.
454
C. Forano
[113] P. Cardoso, J. Tronto; E.L. Crepaldi and J.B. Valim, Mol. Cryst. Liquid Cryst., 390 (2003) 49. [114] M.J. Sanchez, M.V. Villa and M. Sanchez-Camazano, Clays Clay Miner., 47 (1999) 777. [115] J.W. Moore and S. Ramamoorthy, Phenols in organic chemicals in naturals waters, Springer, New York, (1984) 141-147. [116] M.C. Hermosin, I. Pavlovic, M.A. Ulibarri and J. Cornejo, J. Environ. Sci. Health A: Environ. Sci. Engin., 28 (1993) 1875. [117] M.C. Hermosin, I. Pavlovic, M.A. Ulibarri and J. Cornejo, Fresenius Environ. Bull., 4(1995)41. [118] M.C. Hermosin, I. Pavlovic, M.A. Ulibarri. and J. Cornejo, Wat. Research., 30 (1996) 171. [119] M.A. Ulibarri, I. Pavlovic, M.C. Hermosin and J. Cornejo, Applied Clay Sci., 10 (1995)131. [120] J. Cornejo, R. Celis,I. Pavlovic, M.A. Ulibarri and M.C. Hermosin, Clay Miner., 35(2000)771. [121] C. Barriga, M. Gaitan, I. Pavlovic, M.A. Ulibarri, M.C. Hermosin and J. Cornejo, J. Mat. Chem., 12 (2002) 1027. [122] G.H. Giles, T.H. MacEwan, S.N. Nakhwa and D. Smith, J. Chem. Soc, 43 (1960) 619. [123] G.M.S. El Shafei, Adsorption Sci. Techn., 20 (2002) 767. [124]L.Cox, M.C. Hermosin and J. Cornejo, Eur. J. Soil Sci., 46 (1995) 431. [125] R. Celis, J. Cornejo, M.C. Hermosin and W.C. Koskinen, Soil Sci. Soc. Am. J., 62 (1998)911. [126] J.P. Aguer, M.C. Hermosin, M.J. Calderon and J. Cornejo, J. Environ. Sci. Health B, 35 (2000) 279. [127] M. Lakraimi, A. Legrouri, A. Barroug, A. de Roy and J-P. Besse, J. Chim. Phys. Phys. Chimie Biol., 96 (1999) 470. [128] M. Lakraimi, A. Legrouri, A. Barroug, A. de Roy and J-P. Besse, J. Mater. Chem., 10 (2000) 1007. [129] R. Celis, W.C. Koskinen, A.M. Cecchi, G.A. Bresnahan, M.J. Carrisoza, M.A. Ulibarri, I. Pavlovic and M.C. Hermosin, J. Environ. Sci. and Health B, 34 (1999) 929. [130] M.V. Villa, M.J. Sanchez-Martin and M. Sanchez-Camazano J. Environ. Sci. and Health B., 34(1999)509. [131] R. Celis, W.C. Koskinen, M.C. Hermosin, M.A.Ulibarri and J. Cornejo, Soil Sci. Soc. Am. J., 64 (2000) 36. [131b] S. Kannan, Catalytic decomposition of nitrous oxide on new materials, Selected Studies in Heterogeneous Catalysis, Viswanathan B. (Ed.), Indian Institute of technology, Madras, India, (1996) pp. 62-85. [132] B. Combourieu, J. Inacio, A.M. Delort and C. Forano, Chem. Commun., 21 (2001)2214. [133] I.G. Laing, Rev. Prog. Coloration, 21 (1991) 56. [134] P. Mavros, A.C. Daniilidou, N.K. Lazaridis and L. Stergiou, Flotation Environ. Technol., 15(1994)601. [135] K. Skelly, Rev. Prog. Coloration, 30 (2000) 21. [136] T. Robinson, G. MucMullan, R. Marchant and P. Nigam, Bioresource Technol., 77(2001)247.
Environmental Remediation Involving Layered Double Hydroxides
455
[137] G.A. O'Neill and G.M. Goyak, (Aluminum Co. of America, USA), Removal of color contaminants from pulp and paper wastewater streams using adsorbents. US Patent 5378367 U.S. (1995), 6 pp. [138] M. Webb and R.T. Whalley, (Crosfleld Ltd., UK), Dye removal from wastewaters using hydrotalcite. PCT Int. Appl. (1995), 20 pp. Application: WO 94-EP4177 19941215. Priority: GB 93-26302 19931223. [139] M. Webb, Synthetic clay anion exchangers: their structure, modification and application in removing color and toxins from textile process waters, Special Publication - Royal Society of Chemistry (1996), 182 (Ion Exchange Developments and Applications), 135-142. [140] J. Bhatt, H.M. Mody and H.C. Bajaj, Clay Res., 15 (1996) 28. [141] H.Y. Zhu, J. Orthman and G.Q. Lu, Proceedings of the Asia-Pacific Conference, 3rd, Hong Kong, China, Dec. 3-6, 2000 (2001), Meeting Date 2000, 356-361. Publisher: World Scientific Publishing Co. Pte. Ltd., Singapore, Singapore. [142] M.Z. bin Hussein, Z. Zainal, I. Yaziz and T.C. Beng, J. Environ. Sci. Health A., 36 (2001) 565. [143] N. Takahashi, K. Kosuge, T. Nakai and Y. Sato, Mizu Kankyo Gakkaishi, 25 (2002) 553. [144] J. Orthman, H.Y. Zhu and G.Q. Lu, Sep. Purif. Technol., 31 (2003) 53. [145] N.K. Lazaridis, T.D. Karapantsios and D.Georgantas, Water Research, 37 (2003) 3023. [146] J.F. Porter, G. McKay and K.H. Choy K.H, Chem. Eng. Sci., 54 (1999) 5863. [147] U. Costantino, N. Coletti, M. Nocchetti, G.G. Aloisi and F. Elisei, Langmuir 15 (1999)4454. [148] U. Costantino, N. Coletti, M. Nocchetti, AG.G. Alois, F. Elisei and L Latterini, Langmuir, 16(2000)10351. [149] S. Amin and G.G. Jayson, Wat. Research, 30 (1996) 299. [150] G. Onkal-Engin, R. Wibulswas and D.A. White, Environ. Technol., 21 (2000) 167. [151] Y. Seida and Y. Nakano, Water Research, 34 (2000) 1487. [152] M.Z. bin Hussein, Z. Zainal, E.M. Choong, J. Por. Mater., 8 (2001) 219. [153] H. Zhao and G.F. Vance, Clays Clay Miner., 46 (1998) 712. [154] H. Kopka, K. Beneke and G. Lagaly, J. Coll. Interf. Sci., 123 (1988) 427. [155] M.Z. bin Hussein and T.K. Hwa J. Nanoparticle Res., 2 (2000) 293. [156] M. Adachi-Pagano, C. Forano and J-P. Besse, Chem. Commun., (2000) 91. [157] Z.P. Xu and P.S. Braterman, J. Mater. Chem., 13 (2003) 268. [158] I. Dekany, F. Berger, K Imrik and G. Lagaly, Colloid Polym. Sci., 275 (1997) 681. [159] I. Dekany and T. Haraszti, Colloids Surf. A, 123-124 (1997) 391. [160] P.C. Pavan, G.A. Gomes and J.B. Valim, Microp. Mesop. Mater., 21 (1998) 659. [161] P.C. Pavan, E.L. Crepaldi, G.A. Gomes and J.B. Valim, Colloids Surf. A, 154 (1999) 399. [162] P.C. Pavan, E.L. Crepaldi and J.B. Valim, J. Coll. Interface Sci., 229 (2000) 346. [163] K. Esumi, S. Yamamoto and K. Shinjuku-ku, Colloids Surf. A, 137 (1998) 385. [164] Y. You, H. Zhao and G.F. Vance, Colloids Surf. A, 205 (2002) 161. [165] Y. You, H. Zhao and G.F. Vance, J. Mater. Chem., 12 (2002) 907. [166] R. Celis, M.C. Hermosin, J. Cornejo and W.C. Koskinen, Organoclays and organohydrotalcites as sorbents for polar pesticides, Brighton Conference
456
C. Forano
Weeds, 2(1999)699-704. [167] P.K. Dutta and D.S. Robins, Langmuir,10 (1994) 1851. [168] H. Funaki, M. Eda, H. Kamiya, K. Omori and T. Kamiya, (Jemco Inc., Japan; Asahi Glass Co., Ltd.), Recovery of fluorine-containing emulsifier by using layered double hydroxide, Jpn. Kokai Tokkyo Koho (2003) 9 pp. JP 2003212922 A2 20030730 Application: JP 2002-17542 20020125. CAN 139:135238 AN 2003:582477 [169] T. Hibino and A. Tsunashima, Chem. Mater., 10 (1998) 4055. [170] T. Hibino and A. Tsunashima, J. Mater. Sci. Lett., 19 (2000) 1403. [171] K. Bahranowski, E. Bielanska, R. Janik, T. Machej and E.M. Serwicka, Clay Miner., 34(1999)67. [172] J. Carpentier, J.F. Lamonier, S. Siffert, E.A. Zhilinskayaand A. Aboukais, Appl. Catal. A, 234 (2002) 91. [173] J. Haber, K. Bahranowski, J. Janas, R. Janik, T. Machej, L. Matachowski, A. Michalik, H. Sadowska and E.M. Serwicka, Studies in Surface Sci. and Catal., 130A (International Congress on Catalysis, 2000, Pt. A) (2000) 599-604. [174] A. Alejandre, F. Medina, X. Rodriguez, P. Salagre and J.E. Sueiras, J. Catal., 188 (1999)311. [175] A. Dubey and V. Rives, J. Mol. Catal., 181 (2002) 151. [176] A. Dubey, S. Kannan, S. Velu and Suzuki, Applied Catal. A, 238 (2003) 319. [177] L. Ukainczyk, M. Chibwe, T.J. Pinnavaia and S.A. Boyd, Environ. Sci. Technol, 29(1995)439. [178] M. Chibwe, M.E. Perez-Bernal, R. Ruano-Casero and T.J. Pinnavaia, Editor(s): Churchman, G. Jock; Fitzpatrick, Rob W.; Eggleton, R. A., Clays Controlling the Environment, Proceedings of the International Clay Conference, 10th, Adelaide, July 18-23, 1993 (1995), 87-91. [179] M. Halma, F. Wypych, S.M. Drechsel and S. Nakagaki, J. Porph. Phthal., 6-8 (2002) 502. [180] P.L. Crutzen and CJ. Howard, Pure Appl. Geophys., 116 (1978) 497. [181] J.C. Kramlich and W.P. Linak, Prog. Energy Combust. Sci., 20 (1994) 149. [182] S.S. Cliff and M.H. Thiemens, Science, 278 (1997) 1774. [183] A.P. Kinzig and R.H. Socolow, Phys. Today, 47 (1994) 24. [184] F. Kapteijn, J. Rodriguez-Mirasol and J.A. Moulijn. Appl. Catal. B, 9 (1996) 25. [185] Y. Teraoka, H. Fukuda and S. Kagura, Chem. Lett., 1 (1990) 1. [186] H. Yasuda, N. Mizuno and N. Misoni, J. Chem. Soc. Chem. Commun., 19 (1990) 1094. [187] Z.P. Xu and H.C. Zeng, J. Mater. Chem., 8 (1998) 2496. [188] M. Iwamoto and H. Hamada, Catal. Today, 10 (1991) 57. [189] H. Hamada, Catal. Today, 22 (1994) 21. [191] G. Centi, A. Galli, B. Montanari, S. Perathoner and A. Vaccari, Catal. Today, 35 (1997) 113. [191b] K. Yuzaki, T. Yarimizu, K. Aoyagi, S. Ito and K Kunimori, Catal. Today, 45 (1998) 129. [192] S. Kannan and C.S. Swamy, Preparation and characterization of cobalt containing layered double hydroxides, in Studies in Surface Science and Catalysis, Poncelet, G., Martens, J., Delmon, B., Jacobs, P.A., Grange, P. (Eds.) (1994) Vol. 91, Elsevier, Amsterdam, pp. 903-914. [193] S. Kannan and C.S. Swamy, Appl. Catal. B, 3 (1994) 109.
Environmental Remediation Involving Layered Double Hydroxides
457
[194] S. Kannan, S. Velu, V. Ramkumar and C.S. Swamy, J. Mater. Sci., 30 (1995) 1462. [195] A. Corma, A.E. Palomares and F. Rey, Appl. CataL, B, 4 (1994) 29. [196] A. Shannon, F. Rey, G. Sankar J.M. Thomas, T. Maschmeyer, A.M. Waller, A.E. Palomares, A. Corma, A.J. Dent and G.N. Greaves J. Chem. Soc. Faraday Trans., 92(1996)4331. [197] F. Marquez, A.E. Palomares, F. Rey and A. Corma, J. Mater. Chem., 11 (2001) 1675. [198] M. Qian and H. C. Zeng, J. Mater. Chem., 7 (1997) 493. [199] Z.P. Xu and H.C. Zeng, Chem. of Mater., 12 (2000) 2597. [200] U. Chellam, Z.P. Xu and H.C. Zeng, Chem. Mater., 12 (2000) 650. [201] J. Perez Ramirez, J. Overeijnder, F. Kapteijn and J.A. Moulijn, Appl. Catal. B., 23 (1999) 59. [202] J. Perez-Ramirez, F. Kapteijn and J.A. Mouliin, Catal. Lett., 60 (1999) 133. [203] M.C. Romin-Martinez, F. Kapteijn, J. Overeijnder, I.A. Moulijn, Proceedings of the International Conference on Greenhouse Gas Control Technologies, 4th, Interlaken, Switz., Aug. 30-Sept. 2, 1998 (1999), 1998, 343-348. Publisher: Elsevier, Oxford, UK. [204] H. Dandl and G. Emig, Appl. Catal. A, 168 (1998) 261. [205] R. Xu and H.C. Zeng, Chem. Mater., 13 (2001) 297. [206] S. Kannan, A. Narayanan and C.S. Swamy, J. Mater. Sci., 31 (1996) 2353. [207] S. Kannan and C.S. Swamy, J. Mater. Sci., 32 (1997) 1623. [208] S. Kannan, Appl. Clay Sci., 13 (1998) 347. [209] S. Kannan and C.S. Swamy, Catal. Today, 53 (1999) 725. [210] L. Obalova, F. Kovanda, M. Chmielova, Z. Lacny and B. Smetana, Chem. Papers, 56 (2002) 387. [211] Y.F. Chang, J.G. McCarty and E.D. Wachsman, Appl. Catal. B, 4 (1994) 283. [212] J.N. Armor, T.A. Braymer, T.S. Farris, Y, Li, F.P. Petrocelli, E.L. Weist, S. Kannan and C.S. Swamy, Appl. Catal. B, 7 (1996) 397. [213] J. Oi, A. Obuchi, G.R. Bamwenda, A. Ogata, H. Yagita, S. Kushiyama and K. Mizuno, Appl. Catal. B, 12 (1997) 277. [214] H.C. Zeng and X.Y. Pang, Appl. Catal. B, 13 (1997) 113. [215] S. Kushiyama, Appl. Catal. B, 13 (1997) 197. [216] K. Yuzaki, T. Yarimizu, S. Ito and K. Kunimori, Catal. Lett., 47 (1997) 173. [217] S.C. Christoforou, E.A. Efthimiadis and LA. Vasalos, Catal. Lett., 79 (2002) 137. [218] Y. Li, SOx tolerant NOx trap catalysts and methods of making and using the same. (Engelhard Corporation, USA), PCT Int. Appl. (2002), 49 pp., WO 2002011866 A2 20020214, Application: WO 2001-US24033 20010731. Priority: US 2000-634228 20000809. [219] S. Alini, A. Bologna, F. Basile, T. Montanari and A. Vaccari, (Radici Chimica S.p.A., Italy), Process for the catalytic decomposition of nitrous oxide (N2O), Eur. Pat. Appl. (2002), 24 pp., EP 1262224 Al 20021204, Application: EP 2001-830354 20010530. [220] J. Perez-Ramirez, G. Mul, X. Xu, F. Kapteijn and J.A. Moulijn, Studies in Surface Science and Catalysis 13 0B (2000) 1445 (International Congress on Catalysis) [221] Y. Liu, X. Wang, X. Yang and Y. Wu, React. Kinet. Catal. Lett., 68 (1999) 399. [222] A. Corma, A.E.; Palomares, F. Rey and F. Marquez, J. Catal., 170 (1997) 140. [223] G. Carja and H. Niiyama, Environ. Engin. Manag. J., 1 (2002) 155.
458
C. Forano
[224] J. Perez-Ramirez, J.M. Garcia-Cortes, F. Kapteijn, M.J. Illan-Gomez, A. Ribera, C. Salinas-Martinez de Lecea and J.A. Moulijn, Appl. Catal. B, 25 (2000) 191. [225] G. Fornasari, F. Trifiro, A. Vaccari, F. Prinetto, G. Ghiotti and G. Centi, Catal. Today, 75 (2002) 421. [226] G. Centi, G. Fornasari, C. Gobbi, M. Livi, F. Trifiro and A. Vaccari, Catal. Today, 73 (2002) 287. [227] G. Carja, I. Popa, G. Ciobanu and N. Aelenei, Environ. Engin. Manag. J., 1 (2002) 307. [228] R. Dziembaj, L. Chmielarz, A. Wegrzyn and P. Kustrowski, Bull. Pol. Acad. Sci. Chem., 50 (2002) 237. [228b] M. Guyon, J.C. Beziat, F. Figueras, H. Kochkar, J-M. Clacens (Renault, Fr.), Catalyst composition for the storage and reduction of NOx from exhaust gases, Fr. Demande (2003), 14 pp., FR 2831468 Al 20030502, Application: FR 2001-14150 20011031. [229] J.N. Armor, Appl. Catal. B, 1 (1992) 221. [230] A.E. Palomares, J.M. Lopez-Nieto, F.J. Lazaro, A. Lopez and A. Corma, Appl. Catal. B, 20 (1999) 257. [231] E.H. Van Broekhoven, (AKZO N. V., Neth.) Catalyst composition and absorbent which contain an anionic clay, Eur. Pat. Appl. (1988) 16 pp., EP 278535 Al 19880817, Application: EP 88-200010 19880107. Priority: NL 87-56 19870113. [232] R. Cuevas, M.T. Bueno, J. Ramirez, C. Salcedo, B. Mar and F. Pedraza, Revista Mexicana de Ingen. Quim., 2 (2003) 1. [233] T. Kwon, G.A. Tsigdinos and T.J. Pinnavaia, J. Amer. Chem. Soc, 110 (1988) 3653. [234] Y.G. Guo, D.F. Li, C.W. Hu, Y.H. Wang, E.B. Wang, Y.C. Zhou and S.H. Feng, Appl. Catal. B, 30 (2001) 337. [235] Y.H. Guo, D.F. Li, C.W. Hu, E. Wang, Y.C. Zou, H.Ding and S.H.Feng, Microp. Mesop. Mater., 56 (2002) 153. [236] F. Van Laar, D. De Vos, D. Vanoppen, B. Sels, P.A. Jacobs, A. Del Guerzo, F. Pierard and A. Kirsh-De Mesmaeker, Chem. Commun., 2 (1998) 267. [237] D.S. Robins and P.K. Dutta, Langmuir, 12 (1996) 402.
LAYERED DOUBLE HYDROXIDE / POLYMER NANOCOMPOSITES FABRICE LEROUX* and JEAN-PIERRE BESSE Laboratoire des Materiaux Inorganiques, UMR 6002-CNRS, Universite Blaise Pascal, 24 av. des Landais, 63177 Aubiere cedex, FRANCE. * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych and K. G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
460
F. Leroux and J. -P. Besse
I - Introduction Commonly defined as nanocomposite, the system may be described as inorganic sheets lying on top of each other in which covalent forces maintain the chemical integrity and present an interlamellar gap filled up with the polymer guest, thus giving rise to a sandwich-like structure, but it is also extended to the situation where the inorganic filler (LDH) is dispersed into polymeric matrix. The incorporation of polymer between the galleries proceeds via different pathways such as coprecipitation, exchange, in-situ polymerization, surfactant mediated incorporation, hydrothermal treatment, reconstruction or restacking. The latter method, recently effective via the exfoliation of the LDHs layers, appears to be more favourable - in terms of crystallinity - to capture monomer entities than the whole polymer. The in-situ radical polymerization recently developed for these nanocomposites present potentially the advantage to tune the tacticity and the molecular weight of the generated polymer by playing on the layer charge density and the particle size of the host structure, respectively. This process can be defined as an endotactic reaction like those encountered for other relevant assemblies. Indeed, a large variety of LDH/polymer systems may be tailored considering the highly tunable intralayer LDH composition coupled to the choice of the organic moiety. Bio-related polymers and large bio-macromolecules have been incorporated within the galleries of LDH materials, such as poly(aspartate), alginate, deoxyribonucleic acid (DNA), an obvious interest for these bio-organoceramics is the drug release aspect, but also from a fondamental point of view a better understanding of the biomechanisms and other biomimetism phenomena. The purpose of this chapter is to present the state of the art and to identify new trends in term of applications for the materials composed of the assembly between Layered Double hydroxides (LDH) and Polymers. Many reviews and chapters of books devoted to various aspects of claynanocomposites including some of them include LDH materials, may be found in the literature [1-3]. The scope of the present chapter is first to update the LDH/polymer research domain by providing as much informations as needed concerning the synthesis and characterization in addition to this academic point of view to point out the potential applications of those systems. This chapter devoted to Layered Double Hydroxide / Polymer Nanocomposites is structured as follows: - a description of LDH materials including the natural occurence, chemical composition and the aspect of the stacking sequence is briefly discussed. In addition, directly related to the building of inorganic-organic assemblies, the layer charge density and the colloidal and exfoliation properties are presented. - a second section is devoted to the synthetic pathways for the LDH / polymer assembly. Taking into account the anion affinity and stability of LDH, special cares are needed to avoid the breakdown of the host structure and also to provide the best matching between monomeric repetition and layer charge density. It is illustrated by the example of the inorganic polymer poly(silicate) LDH nanocomposite. The methods for the assembly are listed and a non-exhaustive list of compounds is supplied. - the methods of characterization for these two-components materials, i.e. the host structure, and the state of polymer, as well as the nature of the interaction, are gathered in a next section. We will see that the simple model based on a hydrophobic/hydrophilic bilayer needs sometimes to be reformulated taking into account the tacticity of the formed polymer and the free space available between the layers of the host structure. The textural properties and behaviour in temperature will also be presented.
Layered Double Hydroxide / Polymer Nanocomposites
461
- a fourth section summarizes the most obvious domains of potential applications for the LDH polymer systems, namely the sensors and electrochemical devices, the preparation of carbonaceous replicla presenting high surface area with a controlled porosity, the hydrocalumite-type materials related to the MDF cements and finally the application of LDH as nanofillers dispersed in a polymeric matrix. - in the fifth section bio and bio-inspired LDH nanocomposites, the state of the art, the characterizations and processings are presented. - a final section is devoted to the perspectives and future developments. 2 - Brief description of LDH material 2.1 - Natural occurancy and chemical composition Natural hydrotalcite of composition Mg6Al2(OH)16(CO3) 4H2O was first identified in Sweden in the year 1842. Other minerals presenting different chemical natures and stacking sequence belong to this large family, such as manasseite, Mg3Al(OH)8 (CO 3 ) 05 2H2O (2H), meixnerite, Mg3Al(OH)8(CO3)0 5 2H2O (3R), sjogrenite Mg3Fe(OH)8(CO3)0,j 2,25H2O (2H), stichtite, Mg3Cr(OH)8(CO3)0,5 2H2O (3R), takovite Ni3Al(OH)8(CO3)0,5 2H2O (3R), pyroaurite, Mg3Fe(OH)g(CO3)0,5 2,25H2O (3R), hydrocalumite, Ca2Al(OH)6[(C032")on(OH)o78 2.38H2O (3R), wermlandite, Mg(Al, Fe)0,5SO4 2H2O (2H), [Fen4Fenl2(OH)12]2+.[SO4. nH2O]2", an iron hydroxysulphate commonly called green rust, etc. Thanks to synthetic procedures, the family has been largely extended, and various cations at different oxidation states were also stabilized giving rise to unusual compositions. It is the case for cations such as V(III) [4], Mn(II) [5], Co(III)Co(II) [6] or those incorporating La, Y [7]. 2.2 - Layer charge density The anionic exchange capacity (A.E.C.) is reported for some LDH compositions in Figure 1.
Figure 1 - Variation ofanion exchange capacity (meq/lOOg) as a function of the amount of trivalent cation reported per formula weight. The data for [Mgt.xGaJ samples are taken from ref. 11.
462
F. Leroux and J. -P. Besse
The values range between 450 to 200 meq/lOOg, lower values are not possible. The ratio M(III) to M(II) would be too low to maintain the LDH structure. For comparison, cationic clays present exchange capacity in the vicinity of 100 meq/lOOg associated with an area per charge of 70 A2/charge, whereas it is ranging between 25 and 40 A2/charge for LDH materials with compositions M(II)xMn(III) usually in the domain 2 < x < 4. This pictures layers tightly stacked via the attractive forces with the interlayer anions filling the gallery. Vacancies in the inter-sheets domain are not great taking into account the large packing of anions balancing the layers charge. This situation is unfavourable for either an ion-exchange reaction or an exfoliation process. The smaller the exchange capacity {i.e. the layer charge density), the easier the formation of nanocomposite is. The LDH materials do not present the property to be readily exfoliated such as smectite or MS2 type-chalcogenides (MoS2, NbSe2, etc.) [89], the delamination of LDH sheets requires elaborate synthesis. The difference in the charge density may be exemplified by amino acids using either Zn2Al hydrotalcite and montmorillonite as the guest molecules, L-tyrosine and Lphenylalanine, protonated or deprotonated may be incorporated as pillars in cation or anion exchanger layered compounds [10]. When using montmorillonite as host, the basal spacing is largely increased whereas the amplitude for the hydrotalcite is much smaller which shows the difference in the swelling properties between the two clays. 2.3 - Colloidal and exfoliation properties Small particle size and low layer charge density are important features for giving colloidal and/or exfoliated systems and therefore beneficial for subsequent polymer / host structure assembly [12]. This statement may explain the relatively small number of nanocomposites reported in the literature, added to the fact that each LDH composition leads to an unique material the properties of which (exchange, reconstruction or exfoliation) are not easily transferred from one to another. Beside to the layer charge density, the diffusion of cumbersome molecule such as polymer may be a major impediment, therefore the particle size needs to be minimized if one considers the incorporation via direct exchange. To overcome diffusion problems, two main options may be depicted : either the control of the particles size during the preparation or the delamination of the host framework once formed. A method involving separate nucleation and aging steps was reported [13]. It is based on the simultaneous addition of the reactants in a colloidal mill in which the forces prevent aggregation of the nucleated particles. In contrast to conventional preparation methods giving rise to a wide dispersion in the crystallite size, this technique affords smaller crystallites associated with a very narrow distribution of crystallite size, suitable for further anion exchange reaction. The effect of synthetic conditions on tailoring the size of hydrotalcite particles was studied by Oh et al [14]. Homogeneous precipitation of uniform hydrotalcite particles are reported utilizing urea hydrolysis [15]. Small particle size may be achieved by stabilization of emulsions by heterocoagulation of clay minerals and LDH using paraffin/water emulsions stabilized by colloidal particles without surfactants [16]. LHD material was found to be stabilized by forming envelopes round the oil droplets, and the addition of bentonites creates a three-dimensional network of particles with high elasticity which impedes coalescence of the oil droplets. Lagaly et al have studied the properties of colloidal magnesium aluminum hydroxides [17].
Layered Double Hydroxide / Polymer Nanocomposites
463
Other colloidal suspensions of LDH may be obtained after hydrolysis of methoxide-intercalate LDH, and are suitable to form transparent films [18]. One can also change the interlayer chemical nature, which may become more suitable for subsequent reactions with organic species. Hydrophobic LDH, prepared from hydrocalumite and anionic surfactants, adsorb n-heptane, benzene, toluene, and npropanol between the layers with considerable increase of the basal spacing [19]. Large molecules may be incorporated using a mixture of surfactants S+-S" during the exchange reaction [20]. Even if aqueous montmorillonite dispersions were the first platelet-like colloidal systems that showed nematic ordering [21], stable nematic phase of LDH was recently evidenced [22], making LDH a closer mirror image of smectites. The second option consists of delamination of LDH layers. Owing to the high layer charge, the LDH material does not present a natural tendency to exfoliate (see above), and hence many publications are found in this area. Nevertheless, few attempts have been successful with three distinguished routes to date: - delamination of Mg2Al first intercalated with dodecylsulfate anions (DDS) is achieved in polar acrylate monomer under high shear [23]. Suspensions are stable and the polymerization of 2-hydroxyethylmethacrylate (HEMA) gives poly(acrylate) with the inorganic component still in the delaminated form. - a combined use of alcohols and Zn2Al / DDS was used by Adachi-Pagano et al to achieve exfoliation [24]. As stipulated by Hibino and Jones [25], this process does not result from a strong driving force by solvent inclusion, but rather by the choice of the boiling point of the solvent. Instead of a solvent presenting high dielectric constant, known to facilitate delamination, Adachi-Pagano et al had preferred the use of reflux condition at temperature higher than water boiling point. An increase of the alkyl chain of alcohol increase the boiling point but decrease the dielectric constant. The drying process of the surfactant-modified precursor is of great importance to obtain a complete translucent colloidal solution [26]. - by the use of various combinations of amino acid anions and polar solvent, exfoliation of LDH sheets may be achieved [25]. Optimum results habe been obtained using glycine / formamide. A clear colloidal suspension with no original crystalline structure is obtained. As much as 3.5 g.L"1 of LDH can be delaminated by this way. As evidenced by Leroux et al [26], an optimum hydration state and/or water molecule / solvent displacement rate exists for successful delamination All these pathways open new promising possibilities for the incorporation of molecules as cumbersome as polymers and also for the dispersion of LDH materials. It is recently exemplified by Li et al [27]. Using the method developed by Hibino and Jones, they were able to entrap poly(vinyl alcohol) between the layers of Mg3Al. The ultrasound is found to improve slightly the crystallinity of the nanocomposite, although, a large contribution of PVA casting the crystallites cannot be avoided. Alternatively, Bubniak et al use a surfactant-modified LDH to disperse in PEG matrix [28]. 3 - Synthetic pathways for the assembly According to the type of interaction between the inorganic and organic moieties, hybrid materials may be classified in two distinct classes of materials [29]: the first concerns those for which the interactions are weak (van der Waals, electrostatic, or hydrogen bonding), the second concerns strong interactions (covalent bonding). In absence of grafting, the nanocomposite polymer / LDH may be regarded as belonging to first category.
464
F. Leroux and J. -P. Besse
3.1 - Special cares In the case of conjugated polymer, when the host structure presents a strong oxidizing capacity, such as xerogel V2O5, the in-situ polymerization is concomitant with the incorporation of the monomer. Therefore the process is called reductive intercalative polymerization (RIP) [30]. Concerning LDH materials, RIP process is irrelevant, other strategies have to be undertaken to induce the polymerization. Generally, conjugated polymers are polymerized in very acidic medium and with the use of dopant such as FeCl3 or NH4S2O8. Unfortunately, these conditions cannot be employed with LDH framework as low pH medium will dissolve the inorganic structure and anions presenting great affinity toward LDH such as chloride and disulfate anions will displace the monomer. 3.2 - The methods In this section, we consider the case of polymer LDH systems, composed of sheets lying on top of each other and in which covalent forces are maintaining the chemical integrity, whereas weak interlayer interactions are present between lamellae. To be completely sandwiched, the polymeric moiety has to diffuse between the inorganic layers. There are several possible strategies to incorporate the polymer at the core of the host material as underlined by Schollhorn in a review paper [31], who considers three principal options: (a) intercalation of the monomer molecules and subsequent in-situ polymerization, (b) direct incorporation of extended polymer chains in the host lattice via exchange reaction in the case of small molecular weight or via coprecipitation method, (c) transformation of the host material into a colloid system and its subsequent restacking in presence of the polymer. An additional route consists to take advantage of the memory effect of some LDH materials and to reconstruct the lamellar framework on the polymer. Another alternative is to use the guest displacement or solvent-assisted method to compatibilize the host structure galleries to the guest molecule. These different pathways are displayed in Figure 2. The in-situ polymerization (a) is generally a highly suitable method and is employed for the incorporation of various monomers between 2D host structure. Yet, the process is limited by two factors [2]: - the distance between monomers when they are strongly anchored (or grafted) to the host matrix, ;. e. the degree of freedom, which must somehow be in agreement with the layer charge, - the condition that the polymerization (temperature, pH or redox reaction) must leave the layered structure intact. Concerning LDH host structure, different monomers can be polymerized insitu. Acrylate / Mg2Al LDH hybrid material, obtained via exchange reaction with the interlayered anions (Cl~ or NO32"), is further polymerized after a thermal treatment at 80°C. The basal spacing was found to slightly decrease to 13.4 A from 13.8 A [32]. The IR spectroscopy provides information on the polymerization with the disappearance of the C=C vibration band. It is noteworthy that carbonate LDH phase does not react with acrylate. Acrylic acid was also intercalated in the lamellar structure of an ironsubstituted nickel LDH material [33-34], In this study, potassium persulphate is used as an initiator for the polymerization process. The resulting phase was although partially exchanged by SO42" anions. Insertion of polymer leads to a phase presenting a basal spacing of 12.6 A thinner than that of acrylate intercalated monomer phase (13.6 A). This was explained by the absence of electrostatic repulsion between the C=C double bonds. Recently, Vaysse et al report the in-situ polymerization of acrylate in iron-, cobalt-, or manganese-substituted nickel hydroxydes [33]. The preparation of the
Layered Double Hydroxide / Polymer Nanocomposites
465
inorganic framework either via solid state reaction or coprecipitation depending on the nature of the trivalent cations influencing strongly the arrangement of the macromolecule. For Co and Mn-containing LDH, the intercalation and polymerization process appear to proceed concomitantly. Although, the presence of isotactic poly(acrylate) and a basal spacing for the nanocomposite of 7.8 A seem to be questionable taking into account the layer charge density of the host structure and the free-space available for the polymer chain, respectively. Styrene sulfonate was also polymerized between Zn2Al LDH sheets [35], giving rise to well-defined nanocomposites. It was found that when the layers charge density is decreased, i.e. Zri2Al -> Zn4Al, the polymerization is not completed, this will be further discussed in section 4. 3.
Figure 2 - Scheme of the preparation of LDH / polymer nanocomposites: (a) in-situ polymerization, (b) polymer direct incorporation, (c) restacking or reconstruction, and (d) guest displacement or solvent-assisted method.
466
F. Leroux and J. -P. Besse
The direct polymer incorporation by coprecipitation or exchange reaction (b) may also be achieved. During the inorganic crystal growth, it is possible to form nanocomposites with polymers presenting an anionic function such as sulfonate groups. From the polymer point of view, it corresponds to an entrapment reaction, which from the inorganic seeds, it may be considered as a self-assembly assisted process. This gives rise generally to ill-defined materials such as those including poly(acrylic acid), poly(vinylsulfonate) and poly(styrenesulfonate) within layered double hydroxides [3637]. The crystallinity may be partially cured by hydrothermal synthesis as exemplified by PSS / Zn2Al nanocomposite [36]. It was found in this case that the coherence length along the stacking direction is largely increased but also that the inorganic sheets are less corrugated. Generally, the polymer is found not only to influence strongly the textural properties (see section 4.6) but also the intralayer composition of the inorganic structure. The oxide and salt method first developed by Boehm et al [38] and commonly used for the preparation of M2Cr (M = Cu and Zn) was reported to immobilize polymer by the formation of Zn2Cr layers, although, no further characterization were provided [36]. Thus for organo-modified smectite, when the chemical nature between the interlayer space and the guest is not compatible, often due to hydrophibicity, the guest displacement and/or solvent-assisted methods (d) are used. For instance, the presence of hydrophobic alkyl chain in some case makes it possible the incorporation of non functionnalized polymer like what was reported for poly(paraphenylene) (PPP) into MoO3 host structure [39], poly(ethylene oxide) into MnPS3 chalcogenide [40], poly(palanine) polycondensation and poly(vinylpyrrolidone) into ammonium acetate modified-kaolinite [41-42]. Concerning LDH material, previous studies had shown that a pre-intercalation may prepare the interlayer LDH galleries for subsequent incorporation, as evidenced by the early work of Drezdon with the pre-intercalation of terephthlate anions [43] or the swelling with glycerol for the incorporation of poly(oxometallates) (POM) such as H2W12O406" by Dimotakis and Pinnavaia [44]. Dodecylsulfate (DDS) LDH precursor was used for the incorporation of C60 without functionalizing the fullerene molecule [45]. C6o spherical molecules were introduced by dissolving the molecule into the interlayer hydrophobic phase. The solvent plays also an important role in the swelling process. Similarly, the same precursor was used for the direct incorporation of poly(ethylene oxide) [28], the organomodified LDH presents a basal-plane repeat distance of 26.2 A, which suggests a highly interdigitated situation of the alkyl chains. When incorporated PEO, the d-spacing is increased up to 38.2 A. The nanocomposite was characterized by XPS and FTIR spectroscopies. Immobilization of a guest molecule can be achieved by reconstruction of the layered framework by the so-called memory effect. The reconstruction method (c) was successfully employed by Yun et al for the preparation of silicate-intercalated LDH [46]. Meixnerite-type material of composition [MgxAl-OH] (x = 2,3,4) were calcined at 500°C under air and used as precursor for the incorporation of silicate. This was performed using tetraethylorthosilicate, Si(OC2H5)4 (TEOS). This procedure gives rise to samples better crystallized than those formed by ion-exchange [47] or direct coprecipitation methods from metasilicate source and Zn2M (M = Al, Cr) as LDH precursor [48]. An interlayer short chain silicate structure was evidenced by 29Si NMR experiments (section 4.3). Restacking of the layers on a polymer (PSS, Mw of 70000 g/mol) is reported by Leroux and Besse [2]. Unfortunately, the process gives rise to illdefined materials, and appears to be more suitable for the incorporation of smaller molecules.
Layered Double Hydroxide / Polymer Nanocomposites
467
Let us now discuss the case of conjugated polymer. When using the RIP process, from the point of view of the host structure, the final assembly may be considered like a polymer bronze, from the side of the polymer it is viewed as a dopant. Unlike the xerogel V2O5 [49], LDH materials cannot induce instantly the polymerization of poly(aniline), therefore the notion of dopant or polymer bronze is not well adapted in this case. Insertion of conjugated polymers into LDH framework was first reported by Challier and Slade [50]. Terephthalate and hexacyanoferrate exchanged Cu2Cr LDH phases are used as host matrices for the oxidative polymerization of aniline. The reaction performed under reflux condition gives rise to a rather poorly defined material with a basal spacing of « 13.5 A as well as by-products. An alternative method consists in incorporating a soluble anionic monomer such as aniline-2-sulfonate or metanilic acid (3-amino benzene sulfonic acid - H2NC6H4SO3H). Polymerization of the monomer requires less drastic conditions than that needed for aniline, giving rise to a relatively well-ordered system [51]. Indeed, the electrophilic function decreases the potential of polymerization [52-53] and the sulfonic acid ring-substituted polyaniline (PANIS) is capable of self-doping [54-56]. This is suitable since any doping using an external oxidizing agent induces preferentially an exchange with the counter-anions. The conductivity of PANIS is independent of the external protonation over a broad pH range, although, the presence of the sulfonate groups decreases the conductivity of the polymer in its conductive state [57]. It was found that the atmospheric oxygen is necessary for the ignition of the in-situ polymerization [51]. A recent work has shown that the location of the amino group, the length of the alkylsulfonate functional group or the presence of electron withdrawing group on the benzene ring, orientate strongly the rate of polymerization [58]. Polymerization of aniline carboxylic acid into LiAl2 LDH material was also reported, although, the reaction is not complete [59]. The photoinduced isomerization and polymerization of (Z, Z)-muconate anion in the gallery space of [LiAl2(OH)6]+ layers was recently reported [60]. Initially vertically orientated, the anions are polymerized in aqueous medium, while isomerization into more stable (E, E)-muconate only takes place in methanol suspension. The photoisomerization of indolinespirobenzopyran in anionic clay matrices of layered double hydroxides was also reported by Tagaya et a/ [61]. 3.3 - A non exhaustive list A growing number of LDH/polymer systems is reported in the literature. Some of them are displayed in Table 1. The preparation is indicated according to the classification (Fig. 2). 4 - Characterization of the two-components materials A combination of techniques is necessary to characterize these multicomponents assemblies, such as Infrared Fourier transform (FTIR), solid state nuclear magnetic resonance (NMR), Raman, electron spin resonance (ESR) and X-ray absorption (XAS) spectroscopies (X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS)) at a local scale, X-ray diffraction for the long range order, adsorption measurements, electronic microscopies (by electron scanning (SEM), electron transmission (TEM) or by atomic force (AFM)), and gas adsorption for the textural properties. 4.1 - The host structure The incorporation of polymer alters the order at different scale, from the atomic order to the microstructure. The breadth of the hydroxyl stretch region may be
468
F. Leroux and J. -P. Besse
explained by the presence of hydrogen bonded hydroxyl groups with organic moiety [65]. Generally, the incorporation of polymer caused the density to decrease, typically from 2 to 1.5 g/cm3. Table 1 - Polymer/LDH nanocomposites classification (see text).
Polymer /LDH
Synthetic pathiway
References
[50] Cu2Al / PANI (d) Cu2Cr /PANI (a) [51] [59] LiAl2 (a) [62] Ca2Al / PVA (b) [37] Mg3Al / PSS (b) (a) (a, b) Zn2Al / PSS [37], [51] MAI / PA, PVS (M = Mg, Co, Zn) (b) [63], [36] CaAl / PA, PVS, PSS [36] (b) NiFe / PA [32], [34] (a) M2Ni / PA (M = Mn, Fe, Co) [33] (a) (a, b, c) Zn2Al / PSS [2] [64] (b,d) M O 7 PEG-(DC and AS) (M =Cu, Zn) Polymer: (PANI) poly (aniline), (PVA) poly (vinyl) alcohol, (PSS) poly(styrene sulfonate), (PVS) poly (vinyl sulfonate), (PA) poly(acrylic acid), (PEG-DC) poly(ethylene glycol) dicarboxylic, (PEG - AS) poly(ethylene glycol) alkyl (3-sulfopropyldiether). Pathway: (a) in-situ polymerization, (b) polymer direct incorporation, (c) restocking or reconstruction, and (d) guest displacement method.
Figure 3 - X-ray diffraction patterns of two pristine materials (a), hydrotalcite-type Zn2Al and hydrocalumite, Ca2Al, and of their 4-styrene sulfonate (VBS) derivatives (b). The Miller indexing is given, Zn2Al and Ca2Al structures are described in R-3m and R-3 space group, respectively.
Layered Double Hydroxide / Polymer Nanocomposites
469
The presence of large number of harmonics [resp. small] on XRD diagram is indicative of a long-range [resp. short] ordering in the stacking sequence for the framework. From the diffraction line width and using Scherrer relationship, an average coherence length may be estimated, although one has to pay attention to the validity of the relation when the diffraction lines are ill-defined. A general trend may be drawn from the different works gathered in the literature. The host structure Ca2Al gives rise to more ordered structures then that for other LDH counterparts, independently of the polymer nature. It is generally attributed to the ordering of hydrocalumite phase (see next section). The incorporation of monomer molecule (or polymer) is proping apart the layered structure as evidenced by the shift of the harmonics in Figure 3. 4.2 - An endotactic reaction The host structure may accomodate very well a guest molecule, when the monomeric repetitions are close to specific distance found in the galleries. In this case, the incorporation may be defined as an endotactic reaction. In the first approach, one may take into account the layer charge density of the host and the projected surface area of the guest molecule. It is reminiscent of poly(pyrrole) present between the layers of host structures, the compatibility in distance is given by the comparison between 3 (NH) functions (8 A) and V 2 O 5 xerogel host in the crystallographic direction [1-30], or in the direction [201] for FeOCl [49]. Polyaniline may be accomodated into FeOCl galleries by the creation of H Cl stabilizing hydrogen bond taking into account the matching between the distances 10.2 A for (PANI) and 10.4 A for FeOCl on [101] [66].
Figure 4 - Ideal local order for a cation composition M(II) /M(III) of (a) 2:1, and (b) 3:1. The correlations between cations are indicated by P2 -} P6 (see text). Since only the average structure of LDH is known, except in few cases such as hydrocalumite and LiAl2, it is important to consider the local structure. According to an
470
F. Leroux and J. -P. Besse
ideal model based on edge-sharing octahedra [67], it is possible to define the local environment around each type of cations according to the layer charge density. Fig. 4 displays the local cation environment for M(II)XM(III), x= 2 and 3. The correlation between cations are as follows: a first correlation Me - Me at a distance of a noted as P2, then P3, P4, P5 and P6 at «n/3, 2a,
Figure 5 - Moduli of the Fourier transform for CoxAl at Co K-edge (x = 2 (black circle) andx = 3 (empty square)). The distances are not corrected from phase shifts.
Figure 6 - X-ray diffraction (left part) and absorption (right part) of the nanocomposites PSS / Zn2Al before (a) and after (b) hydrothermal treatment. The distances are not corrected from phase shifts) Reprinted from Ref. [35] with permission from Royal Society of Chemistry. The study of the local order, generally observed in LDH host structure such as Mg2Fe [68], Co2Fe,.yAly [69], M2Cr (M = Cu, Zn) [70], etc. is helpful to better
Layered Double Hydroxide / Polymer Nanocomposites
All
understand the potential matching between the LDH host structure and a guest molecule. An example of the local order is given for LDH samples of composition CoxAl (x = 2 and 3) (Fig. 5). An increase in the intensity of the peaks P2 and P4 has to be understood as a greater number of heavier backscattering atoms Co compared to Al atoms in the corresponding shells. To illustrate the concept of local order, nanocomposites obtained from poly(styrene sulfonate) via coprecipitation with or without subsequent hydrothermal treatment are compared. It was found that the lack of crystallinity for the resulting nanocomposite may be partially cured by such a process [35]. From XRD informations, it was shown that the initial ill-defined assembly is not due to a mismatching between the inorganic framework and the polymer (Fig. 6). After the hydrothermal treatment, the nanocomposite is not only more crystallized (/'. e. presence of greater number of stacked platelets) but also the sheets are more planar, as P4 intensity, due to the focusing effect between three atoms lying in a straight line, is amplified (Fig. 6 - right part). The polymer was found to enhance more the nucleation and less the crystal growth owing to its stiffness and textural morphology. 4.3 - Illustration of the endotactic process : the case of inorganic polymer The study of matching between two "sub-lattices" was carried out in the case of inorganic polymers, such as silicate layers. The interlayer poly(silicate) structure between LDH galleries is reported by several authors [46-48]. Supported by 29Si solid state NMR results which are able to address the degree of condensation of silicon-based tetrahedra, Si(OSi)x(OH)4_x and the substitution degree y of the second Si neighbors by another element Si(OSi)x.y(OM)y(OH)4_x, early insight concerning the interlayer silicate structure was provided by Schutz and Biloen [47].
Figure 7 - The proposed structure of silicate / LDH assembly according to ref. [48]. Yun et al confirmed the presence of short chain silicate structure with the presence of Q2, Q3 and Q4 SiO4 site [46] and explained the formation of the polymeric
472
F. Leroux and J. -P. Besse
entity as proceeding in three distinct steps, (i) hydrolysis of TEOS to form Si(OC2H5)4. precursors, (ii) condensation of silanol groups to form Si-O-Si linkages, and (iii) neutralization of some SiOH groups by condensation with the MOH groups present on the gallery surfaces of the LDH. In this study, NMR results preclude the formation of a 2:1 layered silicate. Starting either with Mg3Al or LiAl2 by anion exchange with silicic acid, Schutz and Biloen surmise that the silicate must be condensed into a two-dimensional six-ring structure. These authors proposed a model for the accomodation of the [HSi2O5]nn" twodimensional silicate system with Mg3Al(OH)8]+ host structure, based on the puckering of the SiO4 tetrahedra giving rise to SiOH and SiO- in same proportions. This model gives nice evidences of the matching between guest host structure and the hexagonal lattice of the brucitic LDH sheets (Figure 7). In a first approximation, a rapid calculation based on the concept of bond valence [71] may be predictive of the matching between the polymeric structure and LDH host. It is worth noting that the bond length Si-O (1.63 A) is close to fl/V3 (Fig. 7). LDH materials intercalated with POM Keggin anions such as Mo7O246" and Vio0286" have been studied for their potential application in catalysis, but the exact state of the anions is often surmised. Lopez and Ono have studied the change occuring for Vbased anions by ESR spectroscopy [72]. Initially accomodated with its C2 axis perpendicular to the brucite layers, the decavanate anions undergo a strong rearrangement upon thermal treatment, giving rise first to the reduction of V(V) in V(IV) and then to the formation of polymeric vanadate species [-VO3-]n"~.
y(OH)y
4.4 - State of the polymer Once immobilized between LDH layers, the slow diffusion of the polymer impedes any rapid anion exchange, thus providing a kinetic stability even in aqueous medium containing sodium carbonate. The dimension of the galleries is often consistent with those expected for the incorporation of bilayers of anionic polymers between LDH sheets. Nevertheless little attention is paid to better understand the reasons from the monomer / LDH precursors why in some cases the in-situ polymerization process is not complete and/or why strong contractions are generally observed after thermal treatment. The picture of two polymer chains running on either side of the galleries often questionable has to be modified when the free-space available is irrelevant taking into account the dimension of polymer, but also when the charge per surface of the polymer is mismatching the host charge density. A recent paper using hydrocalumite as ordered model had shown that the polymerization of 4 - styrene sulfonate gives rise to a syndiotactic polymer [73]. Indeed, the hydrocalumite structure is completely ordered: aluminium atoms are six-fold coordinated, whereas the calcium atoms present a [6+1] coordination, the seventh apex oxygen atom coming from the interlayer water molecule. The space group is changed from R-3 to C2/c after incorporation of the monomer supporting the bi-layer arrangement related by a two fold axis. The layer charge density and ordering of the hydrocalumite structure is literally ushering monomer molecule into a specific site.The distance between two consequent monomers up and down being close to the cell parameter a, they can swivel and connect each other only in a zig-zag manner. A model of the in-situ polymerization is provided in Fig. 8. The conductive state of conjugated polymer such as poly(aniline) was estimated from IR diagnostic by Challier and Slade [50]. The proportion between the two ring modes (quinoid and benzoic) enables them to conclude for the presence of poly(aniline) under its emeraldine form.
Layered Double Hydroxide / Polymer Nanocomposites
473
Electron spin resonance is a neat technique to address quantitatively the conductive state of a polymer. A positive charges localized along the chain of a polymer may correspond to the quinone diimine radical cations, or to the polarons obtained by pdoping.
Figure 8 - Schematic representation of syndiotactic poly(styrene) accomodated into LDH-type galleries.
Figure 9 - RPE signal in temperature of aniline - 2 - sulfonate acid intercalated into Cu2Cr LDH galleries, with a sweep width of (a) 6500 G and (b) 150G. In (c) line width (AHPP) variation in temperature of the nanocomposite formed after treatment at 473 K.
474
F. Leroux and J. -P. Besse
For sulfonate aniline derivative, the signal of the polymer is depicted by a fine line at 393 K (see Fig. 9) [51]. A shoulder is also visible at 373 K, indicating an ignition of the polymerization process at lower temperature. The fine line due to the polymer increases continuously in intensity without broadening, the associated linewidth is of 40 +/- 10 G and g of 2.0034. The linewidth (AHPP) of the broad signal (response of the LDH host strucuture) is independent on the heating process, whereas it increases in temperature for the fine signal. The positive dependence of AHPP observed in the temperature range 130-300 K, confirms the metallic behavior of at least a part of the nanocomposite [74]. Unfortunately, small conductive properties are reported for other PANI related nanocomposites, (PANI)o.8oCa2Nb3010 (less than 10"10 S.cm"1) [75], PANI / HMMoO6 (M=Nb, Ta) [76], PANI / HUO2PO4 [77] or PANI / VOPO4 [78]. The lack of conductivity was explained by the fact that the polyaniline is encapsulated inside the inorganic host. Even if a part of the polymer presents a metallic behavior, the whole material does although not possess such a property. The polymer is present as isolated conductive domains (islands) dispersed in a non-electronically conductive matrix or may also possess defects at its ends, as pointed out by Javadi et al [79]. Small conjugation lengths are generally measured, which associated to an electron withdrawing effect of the sulphonate groups may cause a poor electronic delocalisation. 4.5 - Measure of the interaction As interfacial properties may be of great interest for an application point of view, and since a thin polymer layers may even change drastically ion diffusion or electrical properties (see section 5), it is important to distinguish the part of polymer embedding the nanocomposite from the portion truly intercalated. A measure of the interaction between the two components may be indicative of grafting process or any structural change. Through measurements of the spin-lattice relaxation time (T]) using the inversion recovery sequence [n - % - nil -AQ], the proportion of C60 truly intercalated between Mg2Al LDH-type layers can be estimated from the part just adsorbed on the clay surface [45]. Once intercalated in LDH, a slowing of the rotational motion of C6o is observed. The variation in NMR lines width indicate also the presence of dipolar interaction between the 13C nuclei in C6o and proton containing species. The part of the polymer coating the nanocomposite can be carefully analyzed by adsorption isotherm measurements, as already performed on LDH materials with styrene sulfonate and poly(styrene sulfonate) [80] and surfactant molecules [81-82] adsorbed on their surface. The adsorption isotherms are classified according to the slope of the initial adsorption curve, and are further subdivided according to their curvature [83]. Two concepts, the Langmuir and the Freundlich models, are commonly applied to modelize the adsorption process. The linear form of the Langmuir model is represented by the relation:
=-L+
qe
Q° bQ°C e
=^+C^
qe
( Eq .l)
XmKe Xm
where Ce is the concentration of adsorbate at equilibrium (mg/1), qe the amount adsorbed at equilibrium (mg/g), and Xm = Qo the maximum amount of adsorbate that can be
Layered Double Hydroxide / Polymer Nanocomposites
475
adsorbed in a monolayer, Ke = b the Langmuir constant related to the interaction intensity of adsorption. The Freundlich model is given by the following relation : qe = K f .C e "°
linearised in
lnq e = l n K f + - l n C e n
(Eq. 2)
where kf and 1/n are constant related to the adsorption capacity and the intensity of the solvent, respectively. The adsorption of 4 - styrene sulfonate monomer onto Zn2Al (Cl) LDH surface is characteristic of the L(2) Langmuir type according to the classification of Giles et al It is usually observed for the adsorption of monofunctional polar solute on polar substrate in a polar solvent as, for instance, the adsorption of glucose on graphite, or of sulphanilic acid on wool. Since the monomer molecules adopt the Langmuir model, it was inferred that strong interactions are present between the adsorbed molecules and the outer-surface of the LDH material, while the adsorbat / adsorbat interaction was weak. Poly(styrene sulfonate) adsorption was found to follow the Freundlich model, with a curvature charateristic of the S(2) adsorption type. It is commonly observed in the case of monofunctional solute molecule presenting a moderate intermolecular attraction and meeting a strong competition to reach the substrate sites, generally from the molecules of the solvent.
Figure 10 - I3C CP-MAS NMR spectra of (a) sulfanilic acid and (b) 4-styrene sulfonate acid. The bottom spectra are relative to the respective intercalated phase. (MAS rotation at 10 kHz) Concerning the interaction of the guest molecule with the /w«er-surface of the galleries, this was illustrated by El Mostafa et al work, who studied the interaction
476
F. Leroux and J. -P. Besse
between 4-styrene sulfonate and sulfanilic acids with Zn2Al LDH materials by means of solid state NMR spectroscopy [35]. The incorporation of the organic molecule induces an up-field shift of the carbon atoms (C2), (C3) and (C5) (Fig. 10), corresponding to a shielding effect and consistent with an electrostatic interaction between the sulfonate function and the hydroxide layers. The interaction weakens the electrophilic character of the monomer through the carbon atom (C5). The shielding propagates through the benzene backbone down to (C2). It is interesting to note that the resonance peak of (Cl) (=CH2) is then deshielded, indicating that the n electrons are preferentially located on the terminal carbon. From electronic consideration, it may explain why the in-situ polymerization is achieved in the absence of chemical initiator and requires a soft thermal treatment only. Similar observations were made on the sulfanilic acid, the incorporation also inducing a shielding of (Cl) and deshielding of (C5). The interaction may be studied by FTIR. Shifts in the symmetric and asymmetric stretching modes of (SO3~) functional group constitute sheer evidences of a geometric disturbance. The observed down-shift in frequency corresponds to a weakening of the S=O bond strength. The loosening of the S=O bond suggests the presence of an electrostatic binding with the clay surface through hydrogen bond as follows S = O...H - O - Me (Me = Zn or Al) [35]. 4.6 - Textural properties Polymer macromolecules are known to modifiy the nucleation and the crystal growth of the inorganic colloids. The nanocomposite does not present the usual «sandrose» morphology of the inorganic LDH parent after in-situ polymerization of 4-styrene sulfonate. Concerning hydrocalumite host, the thermal treatment does not change significantly the morphology of the sample, yet the size of the platelets is substantially reduced (Fig. 11). The morphology for PSS / LDH nanocomposite differs with the presence of large chunks. It may be related to the synthesis pathway, LDH crystal being selfassembled directly on the polymer [2]. The microtexture of the nanocomposite poly(vinyl alcohol) (PVA) / Ca2Al displayed in Fig. ll(f) shows a crumpled flaky aspect. 4.7 - Behaviour in temperature It has been observed [62] that the presence of polymer not only affects the crystallinity but also the dimension and morphology of the pristine host material. In addition, it induces in several other effects in the assembly under the affect of temperature. First by holding together the layers, it enhances generally the thermal stability. The collapse of the lamellar structure is delayed under temperature. PVA/Ca2Al layered structure is found to be stable up to a temperature of 400°C [62]. The authors have speculated that the nature of the interface between the organic and inorganic components may be the reason of such improved thermal stability. The organoceramic transforms at high temperature into inorganic solids of different compositions compared to those resulting from the heat treatment of the pristine host material. SEM pictures show how the nanocomposite is degraded on heating (Fig. 12); the organic residue is encompassing the inorganic crystallites, thus preventing the crystallization of CaO. It is also well demonstrated in the system poly(styrene sulfonate) / Zn2Al, where the formation of ZnO crystal was found to be largely delayed due to increasing temperature (Fig. 12 (c) and (d)) [84].
Layered Double Hydroxide / Polymer Nanocomposites
All
Figure 11. Scanning electron micrographs ofZn2Al / Cl (a) and its (PSS) derivatives (b and c) and of Ca2Al hydrocalumite (d) and its (PSS) (e) and (PVA) (f) derivatives. (PSS)-based nanocomposites were synthesized via in-situ polymerization (b, e) or by direct polymer incorporation (c). The bar represents 5 \m, except for (f) 10 /jm. ((f) reprinted from Ref. 62 with permission from ACS). Secondly, the polymer may act as a protective shell, for not only it delays the crystallization of by-product (see above) but also it induces the formation of unusual solids after thermal treatment under inert atmosphere. The nature of these solids
478
F. Leroux and J. -P. Besse
depends evidently on the cations initially present in the LDH sheets but also on the organic function of the polymer. Chalcogenides such as ZnS or CaS [84-85] or nitride such as A1N [86] were thus obtained after thermal decomposition under N2 atmosphere.
Figure 12. SEM micrographs of nanocomposites PVA / Ca2Al treated at (a) 500°C and (b) 1000°C under air, and of (c) Zn2Al / Cl and (d) PSS / Zn2Al after thermal treatment at 600°C under N2 atmosphere. The bar represents 2 /m. ((a) and (b) reprinted from Ref [62] with permission from ACS, (c) and (d) reprinted from Ref. [84] with permission from RSC). When calcined residues are amorphous, XAS technique is useful to characterize the local order. For instance, concerning PSS / Zn2Al nanocomposite, XANES at sulfur K-edge was first examined to know whether or not the initial sulfonate group is maintained in temperature or if a grafting process occurs. The spectra are dominated by a single white-line feature reflecting the transitions 1 s to np and corresponding to localized, unfilled atomic or molecular states (Fig. 13). At 600°C, the white line for PSS / LDH phase initially located rigorously at the same energy than for PSS macromolecule, i.e. 2481.6 eV, is shifted either to higher energy after treatment in air, or to lower energy when treated under N2 atmosphere, indicating the formation of Na2SO4 or ZnS, respectively. It corresponds in the first case to the oxidation of the sulfonate to sulfate, in the second to the reduction reaction of sulfur atoms down to -2. EDX analysis shows that there is a phase segregation (Fig. 13 left part), with small particles arranged in stars composed mostly of Zn and S. The corresponding fine structure allows to shed more light, the features showing great similarities with ZnS, 2H - wurtzite phase, as evidenced by the
Layered Double Hydroxide / Polymer Nanocomposites
479
comparison of the two kx(k) signals (Fig. 13, right part (bottom)). ZnS (2H) crystallizes in a hexagonal symmetry (a =3.82 A, c = 6.26 A) with P63mc (186) space group. The authors conclude from EXAFS refinements that the first shell surrounding the sulfur atoms is associated to the presence of Zn atoms, with a distance Zn-S comparable to that observed in ZnS [84].
Figure 13 - Combined analysis of the products obtained after thermal treatment of poly(styrene sulfonate) / Z112AI system at 600°C under (a) N2 and (b) air atmosphere. On the left part, the EDX analysis is presented, and on the right part, XANES (above) and EXAFS (below) spectra are displayed in comparison to ZnS (c). By the guest displacement of Me(CH2)nOSO3" sulfonate anions (see section 3.2), the acrylonitrile was intercalated and subsequently polymerized. Acrylonitrile was polymerized in presence of an initiator benzoyl peroxide into a surfactant-modified Mg3Al LDH [86]. The thermal behavior of the nanocomposite was studied under N2 atmosphere. Below 1600°C under N2 atmosphere, the by-products were A1N, MgS and MgO, whereas above this temperature, single-phase submicron A1N grains were obtained. During the carbothermal reduction, the stratified polymer (PAN) acting as a shell impedes the layers to collapse on each other and to form mixed (Mg, Al) oxide, while the crystal growth is occuring in the interior of the particles, thus giving rise to submicron particles after decarbonization. The by-products were identified by X-ray diffraction. Another aspect is the formation of carbonaceous materials after demineralization (see section 5.3).
480
F. Leroux and J. -P. Besse
For Al-based LDH, the structural changes occuring during the thermal treatment are generally accompanied by a conversion of intra-sheets A1(OH)6 octahedra to inter-sheets A1O4 tetrahedra. This conversion can be evaluated by single pulse 27A1 solid-state or triple-quantum (3Q) 27A1 NMR both in magic angle condition MAS [8788]. As much as 10% of initially (^-coordinated Al nuclei can be thus converted to Td without the collapse of the structure. Generally the presence of polymer delays the temperature of conversion [85,35]. 5 - Domains of application Polymer (or LDH depending of which side) is expected to impart measurable benefits for the whole material. The properties of polymer / LDH nanocomposites are presented in the following. 5.1 - Organic inorganic hybrid assembly 5.1.1 - Adsorption, sensors and electrochemical sensors Surface and interfacial characterizations are of great importance for application in adsorption of molecules or bio-related molecules. Electrophoretic mobility may be measured at the stationary plane using the technique of electrophoretic light scattering [63,89]. A notable difference is observed between a carbonate inorganic phase and poly(vinyl sulfonate) (PVS) and poly(styrene sulfonate) (PSS) LDH nanocomposite derivatives. The latter presents a negative electrophoretic mobility over a large domain of pH without the presence of an isoelectric point (iep), whereas the inorganic phase displays an iep at pH of 11 exhibiting both positive and negative surface charge region below and above this value, respectively. This behaviour has to be understood by the amphoteric nature of the surface site consisting of both hydrous aluminum and magnesium hydroxide whereas no i.e.p. suggests that only one type of surface site is present in the case of the nanocomposites. Since the behavior is close to what was reported for sulfonate polystyrene latex [90], the authors concluded that sulfonate sites are in the near surface region. Anionic surfactants may be readily incorporated between LDH layers adopting various positions, perpendicular to the hydroxide sheets or a tilted position [91]. From dodecylbenzne sulfonate (DBS) organo-modified LDH, enhanced adsorption capacities of tri-, and tetra-chloroethylene is explained on the basis of the alkyl tails and benzene rings acting as an organic solvent phase, suitable for adsorption properties [92]. A patent reports the invention of an underneutralized superabsorbent polymer trapping Na+ cations and LDH Cl" anions, thereby removing the electrolytes from a solution and may find application as diapers, incontinence garments, sanitary napkins, etc [93]. Recent works on conjugated polymer / LDH have shown that the polymer once immobilized inside the galleries are still electroactive and may be of use for bio-sensor [94-95]. Electrochemical polymerization of aniline derivative such as 2-aniline sulfonate was found to be effective in non-aqueous solution without proton donor. Previous works performed by Miras et al [96] and Yamada et al [97] have shown the possibility to obtain electroactive polyaniline in such electrolyte. As observed in other confined cases such as zeolite [98] and montmorillonite [99], the potential of oxidation of the electroactive monomer is shifted from its initial position. 5.2 - Biological and environmental applications A Chinese patent reports the use of Mg-Al hydrotalcite-type materials as thermal insulating filler in polyethylene film. The nanocomposite exhibits greater
Layered Double Hydroxide / Polymer Nanocomposites
481
absorption of IR light than the common filler talc, and therefore may find applications in green-houses [100]. Among the numerous patent relative to the use of polyolefin resins as agricultural film, one report the mixing of the resin with 12 to 30 wt% of a LDH having a peak temperature for liberating water of crystallization, such as LiAl2 [101]. The laminated polyolefin resin film reported by the research group of Sumitomo Chemical0 comprises an inner layer which contains in majority the resin and 12 to 30 wt% of a layered double hydroxide. LDH materials may be present into the three different layers but, in order to avoid the deterioration of the film appearance (transparency), mostly inside the inner layer. During Ni sorption onto the pyrophyllite, a clay mineral, and using timeresolved characterization, some authors have observed a depletion of Ni from solution and release of Si from the pyrophyllite structure [102]. By diffuse reflectance and extended X-ray absorption spectroscopies, it was found that the sorbed intermediate phase consited of a Ni-Al LDH, and that upon ageing, interlayered nitrate anions were changed by silica polymers. The silica tetrahedra derived from the pyrophyllite structure migrate towards NiAl galleries. This "silication" reaction gradually transforms the LDH into a precusor Ni-Al phyllosilicate, and therefore enhances the structural stability of the surface precipitate. This study suggests a potential for long term Ni stabilization in soil. In the same way, Seida and Nakano report the removal of humic substances by iron containing LDH [103]. Only a few studies report the use of LDH / polymer systems in cosmetic and pharmaceutical fields. For instance, an occlusive gel comprising the LDH materials intercalated by monoalkyl (ether / benzene) sulfate anions and a vegetable oil and a secondary thickener may find application as a cosmetic composition [104]. The intercalation of polystyrene oligomers into hydrotalcite may also be considered as a soap-free emulsion polymerization [105]. Chibwe et al report the catalytic properties of biomimetic metallomacrocycles intercalated in layered double hydroxides [106]. Another paper reports the immobilization of penicillin G acylase on calcined layered double hydroxides [107]. 5.3 - Carbonaceous replica In 1988, Kyotani et al investigated the formation of highly orientated graphite from polyacrylonitrile by using a two-dimensional space between montmorillonite lamellae [108]. The organic moiety was incorporated under its monomer form between the layers, then the polymerization was carried out using y-radiation under nitrogen. The formed polymer decomposes between 400-500°C and yielding to a flat carbonaceous material. After further thermal treatments and strong-acid washing to get rid of the inorganic layers, a soft carbon was obtained, which can turn out to graphite under thermal treatment at high temperature (2800°C). From several studies, it was shown that constrained organic polymers give rise to carbonaceous materials presenting a high surface area associated to a rather good control of the porosity, the constrainement hampering the side-reaction between carbon species and increase the microporosity. The concept was generalized to 2D or 3D inorganic network, such as mesoporous silicate MCM48 with sucrose giving rise to a carbonaceous material presenting mostly a mesoporosity associated to a surface area of 1380 m2g"' [109-110]. Moreover, in this case, a mesoporous silicate adopting a new symmetry, may be regenerated from the carbon replica [111].
482
F. Leroux and J. -P. Besse
Figure 14 - Left part: N2 adsorption-desorption curves for the IPC obtained from (a) PSS and (b) Zn2Al / PSS. Right part : TEM micrograph of carbonaceous material obtained from PSS / Zn2Al nanocomposite after thermal treatment at 60°C under N2 atmosphere and subsequent acid treatment. The bar represents 100 nm. Layered double hydroxides were employed as the inorganic host structure and 1,5 naphthalene disulfonate and poly(styrene 4-sulfonate) (PSS) [91-93] as organic precursors. For the latter and after the charring process, the surface area of the carbonaceous material presents a maximum of 1020 m2/g. Conversely, PSS-Na displays a surface area of 370 m2/g only. This is exemplified with the typical N2 adsorptiondesorption curves of the two carbonaceous materials, also called intercalated polymerderived carbon (ICP) (Fig. 14). The benefit of confinement in terms of specific surface area and well-defined pore distribution size is clearly evidenced. The curvature during the desorption process for the ICP obtained from PSS / Zn2Al nanocomposite is characteristic of pores shaped like "bottleneck" form. TEM studies revealed (Fig. 14, right part), an open structure for ICP in agreement with the adsorption measurements. Indeed, Putyera et al have shown from CH4 and SF6 adsorption measurements that the extracted carbonized products present a heterogeneous micropore structure. When other organic molecules such as 1, 5 naphthalene disulfonate are used instead of polymer, there is no gain in the specific surface area from the constrainment [113], whereas the effect is clearly observed for the polymer. For instance, Na-PSS when confined into LDH galleries gives carbonaceous materials exhibiting a specific surface area of 1020 m2/g associated to a microporous
Layered Double Hydroxide / Polymer Nanocomposites
483
volume of 0.31 cm3/g [112], values more than twice those obtained in the absence of any constrain [115]. 5.4 - Cement-related materials The incorporation of polymer into hydrocalumite host is of current concern for cement application. Indeed, like calcium silicate (noted as C-S-H), these calcium aluminium hydroxide salts AFm also called Friedel salts occur in the hydration process of cement and their role in the mechanical properties may be of major importance. The research in this field is topical [116-117]. Few studies have reported computed models for the interaction of polymer with hydrocalumite type surface. From a theoretical approach [118-120], it was demonstrated that the cross-linking of polymer chains, polyacrylate and polyvinylalcohol, with either Ca or Al atoms could be achieved via the carboxylate or the terminaison -O-, respectively and that it may be a key factor for the filling of the large voids present initially within the cement. Those materials are defined as Macro Defect Free cements. On the other hand, the incorporation of poly(silicate) between LDH galleries may be viewed in terms of composition closely related to cement-based solids, C-S-H. Except application in catalysis for the dehydration/disproportionation of 2-methyl 3butyn-2-ol (MBOH) [46], no work relative to cement-base materials using LDH polymer nanocomposite has been reported so far, making this research quite incentive all the more so that the intercalation of polymer into CSH materials was found to be not effective, probably yielding to an entanglement rather than to a true intercalation [121]. 5.5 - Towards LDH nanofillers 5.5.1 - Enhanced mechanical properties General approach for the formation of nanocomposite from clays is to use a swelling agent, which can also compatibilize the inorganic layers with the polymer [122]. To do so, alkylammonium cations are incorporated between the layers of cationic clays. For different purposes, the complete exfoliated state is not wished [123], and some work has shown that small part of the polymer intercalated between the nanofiller sheets may greatly promote the properties of the whole [124]. This concept is not new. Indeed in 1974, Hawthorne et al studied the polymerization of vinylic monomer at the surface of mineral for mechanical purposes [125]. In 1979, Kato et al reported the thermal properties of a nanocomposite formed by nylon-6 obtained from aminocaproic acid and a montmorillonite [126]. Few years later, researchers at Toyota using the same assembly but utilizing the inorganic framework dispersed in the polymer, show a great enhancement of the mechanical properties [127129]. Since then, the research works in this field have not ceased, using different polymers but mainly the one clay mineral: montmorillonite. Nevertheless few recent papers have related the use of LDH materials as nanofillers. A nanocomposite LDH / poly(imide) was prepared from a poly(imide) precursor, poly(amic acid), and an organo-modified LDH- aminobenzoi acid (AB) [130]. Poly(imide) (PI) is known as a great engineering plastic. The preparation was adapted from a recent work reporting the reaction of imidization in presence of sodium kenyaite of composition Na2Si22O45.10H2O, a cationic clay [131].
484
F. Leroux and J. -P. Besse
Figure 15 - TEM micrographs of LDHs/epoxy nanocomposites with various LDHs contents : (a) 3wt%, (b) 5 wt%, and (c) 7 wt%. The bar length is 50 nm. (Reprintedfrom Ref. [132] with permission from Elsevier).
Layered Double Hydroxide / Polymer Nanocomposites
485
The (AB) organo-modified LDH was placed with pyromellitic anhydride (PMDA) and 4'-oxidianiline (ODA) in N, N-dimethylacetamide. This gives rise to poly(amic acid) precursors which transform into PI after imidization reaction. The idea of producing a multi-branched polymer grafted into LDH galleries is based on the reaction of the amine group with anhydride : one anhydride functional group of PMDA reacts with the amino group of AB and the other with ODA. ODA molecules which present two amino groups are acting as linkage agent. After having reached a high molecular weight, the further polymerization of PAA induces the exfoliation of the LDH-AB nanolayers. The tensile strength at break is increased with LDH-AB content and a maximum of 131 MPa is reached with 5 wt% for the nanocomposite LDH-AB/ PI, which is 43% higher than pure PI. The glass transition, Tg, is found to increase with the LDH-AB content, associated to the more severe restriction of the PI chain mobility in presence of LDH nanolayers. The rigid Mg/Al nanolayers enhance the stiffness of the nanocomposite and the thermal resistance, while the thermal expansion coefficient decreased. Finally, the nanocomposite may act as a gas barrier preventing volatile gas to permeate by creating a long path for the diffusion. Same authors report the preparation of LDH/epoxy nanocomposites [132]. The epoxy resin was prepared by mixing the epoxy, diglycidylether of bisphenol (DGEBA) with a diamine curing agent, a polyoxypropylene diamine Jeffamine® D400. The inorganic Mg2Al LDH layers are reacting with aminolauric acid (AL) during the coprecipitation and the solid presents a paraffin structure. The layers of AL/Mg2Al are exfoliated into the epoxy resin at different loadings as observed in Fig. 15. Like LDHAB/PI system, the obtained nanocomposite presents a same general trend in terms of increased properties. This is explained by the adhesion between the LDH nanolayers and the epoxy arising from the reaction between the amine groups of the intercalated amino laurate and the epoxy group. The formation process of the LDH/epoxy nanocomposites is represented in Figure 16.
Figure 16 - The process of formation ofLDHs/'epoxy nanocomposites. (Reprinted from Ref. [132] with permission from Elsevier).
486
F. Leroux and J. -P. Besse
In the same way, the preparation of poly(methyl methacrylate) LDH nanocomposite was recently reported [133]. An intercalated LDH phase glycine / Mg3Al was first delaminated in formamide, then a solution of acetone containing the polymer was added. From TEM observations, the brucite-like sheets of LDH were found to be individually dispersed in the polymer matrix. The effect of the carrier resin, dispersing agent and processing conditions on the dispersion mechanism of the masterbatch for LDH/polymer nanocomposites are discussed in a paper written in Chinese [134]. Modified-LDH with suitable anions such as alkyl- or alkylphenyl-carboxylic, sulfonic or -phosphonic acids, when mixed with a polymer matrix in presence of a second charge-carrying group such as cationic group yields to exceptionally stable and particularly homogeneous nanocomposites as claimed by the authors of the patent [135]. Those nanocomposite materials may contain layered double hydroxides in amount over 20 wt % and may find application as shaped moldings. 5.5.2 - Flame-retardant properties Clay-polymer nanocomposites have proven to be interesting candidates as fire retardants or gas barriers. Previous works report mainly the utilization of cationic clay, although LDHs are emerging in this type of application. MgAl LDH-type layers were also dispersed in poly(ethylene-grafter-maleic) (PE-g-MA) [136]. To do so, dodecylsulfate organo-modified Mg3Al prepared from reconstruction is placed with PE-g-MA in xylene under reflux condition. The molecular dispersion is reached since the authors observed the disappearance of harmonic diffraction lines. From TEM and selected area electron diffraction (SAED), it was shown that the inorganic layers are present as a disorder phase with sheets of about 70 run length or width and that the hexagoanl crystal structure is kept. The nanocomposite presents a slower thermo-oxidative behavior than pure PE-g-MA in temperature, and may be of use as flame-retardant material. A patent describes also the fireproof for halogen-free polymer compositions of thermoplastic, crosslinkable or crosslinked elastomeric and/or duroplastic polymers mixed with LDH and organo-modified layered silicate [137]. It is emphasized mostly with the use of Escorene™, an ethylene vinyl acetate copolymer. The study on combustion and thermal degradation behaviors of flame-retarding polyamide 6 (PA-6)/ polypropylene (PP) blends containing nano-layered double hydroxides and NH4-polyphosphate were studied [138]. The flame-retardancy is found to be improved by the synergistic effect of the mineral addition, LDH promotes crosslinking and char formation during the thermal degradation of the blends. 6 - Bio and bio-inspired LDH nanocomposites 6.1. The state of the art Few biopolymers were incorporated between LDH lamellae. For instance, the incorporation of poly(a, p aspartate) is reported [139]. It proceeds by the condensation process from the aminosuccinic acid via a polysuccinimide intermediate which rearranges to give polyaspartate at 220°C. It was found that the basal spacing decreases during the condensation process from 11.1 to 9.0 A, giving an available space of only 4.2 A for the accomodation of the polymer. As mentioned in section 3.2, the polymer may influence the intralayer composition: starting initially from a ratio Mg(II) to Al(III) of 2, the final product presents after incorporation of poly(a,p aspartate) (initially used as cosolute in the basic reaction solution) a ratio of 1.2 showing that the uptake of Al(III) cations is more favourable [139].
Layered Double Hydroxide / Polymer Nanocomposites
487
The presence of biopolymer influences strongly the textural properties of the hybrid systems. This is exemplified by alginate / Zn2Al LDH nanocomposite. Alginic acid, a biopolymer produced by brown seaweed a heteropolysaccharide having a non regular structure, has extensively been studied for its property in gel formation and its subsequent application in the food packaging and pharmaceutical industries. The glycuronan is considered as a linear copolymer with alternative sequences of guluronate (G) and mannuronate (M), presenting a glycosidic linkage, diaxial or diequatorial, respectively. The assembly between the biopolymer and the inorganic matrix gives rise to an intricate morphology (Fig. 17), with stacked layers still observable. The synthesis of LDH oriented by the biopolymer leads to unusual submicronic features [140], with the observation of a tubular shape in comparison to the sand-rose morphology of the chloride LDH phase (Fig. 17). The bio-polymer is acting as a glue to consolidate nanosized LDH particles into larger scale aggregates.
Figure 17 - SEMpicture of alginate / Zn2Al nanocomposite (left part) and the schematic representation (right part). (Reprintedfrom permission of Elsevier from Ref [140]) Biomolecules and large bio-macromolecules were also incorporated between the layers of LDH materials. They are nucleoside monophosphates: AMP, CMP, GMP, ATP [141]. The herring testis DNA, an adenosine triphosphate, were also incorporated in LDH materials [142]. Such bio-nanocomposites were prepared via exchange reaction with nitrate-LDH precursors. 6.2 - Characterization and processings The incorporation of the biomolecules is pushing apart the layers, the d-spacing is increased from 0.78 nm up to 1.94 nm for ATP-Mg2Al (LDH). The increase of the basal spacing is consistent with the thickness of DNA molecule in double helical conformation lying parallel to the basal plane, although the dimension of the the interlayer space suggests a confinement of the molecule and therefore strong electrostatic interactions with the host structure. A spreading of the vibration bands in the IR spectrum was also interpreted as an interaction between the biomolecule and the hydroxide sheets. 6.3 - Applications The role of LDH materials is first to provide the neutralization of the negatively charged biomolecules, making possible the transfer by cancelling the electrostatic repulsive interaction between the biomolecule and the negatively charged
488
F. Leroux and J. -P. Besse
cell. Then, once incorporated into the lyosome, the LDH moiety dissolves allowing a progressive release of the biomolecule. An incubation in CO2 atmosphere may give rise to the release of the molecule as well, induced by the strong affinity of LDH materials towards carbonate anions. For example, the potential delivery of DNA was tested from fluorophore using a laser scanning confocal microscope. By this way, it was shown that the cell can engulf the neutralized nanoparticles through phagocytosis or endocytosis [143] LDH-based bio-hybrids may provide new opportunities as reservoir and delivery carriers of functional biomolecules such as DNA, and therefore may find applications in gene therapy and drug delivery [144]. Choy et al have shown that a biomolecule such as ATP molecules can be exogenously introduced into eucaryotic cells. An overview of the chemistry and further delivery of these systems is provided on Fig. 18.
Figure 18 - Schematic illustration of the hybridization and expected transfer mechanism of the bio-LDH hybrid into a cell. (Reprintedfrom permission ofWiley-VCHfrom Ref. [143]). 7 - Future developments It is important to distinguish between the inorganic polymer layer assembly and the LDH nanolayers dispersed into a polymer. For the first type of materials, several further developments may be considered. Recent works on bio-polymer nanocomposites based on chitosan intercalated in montmorillonite have successfully been used in the development of bulk-modified electrode. The nanocomposite was found to provide a long-time stability [145]. This should give some ideas for in the detection of anions using LDH nanocomposite. Like
Layered Double Hydroxide / Polymer Nanocomposites
489
carbon aerogels [146], the ICPs may also constitute a new family of high-surface area carbon materials suitable for application as supercapacitors. The concept is based on charge accumulation present at the layers, the so-called Helmhotz layers. Indeed, carbon repliqua obtained from cationic clays, bentonite or sepiolite, and pyrene as organic precursor were tested as anode material in lithium battery [147]. High reversible capacity (up to 825 mAh/g corresponding to more than twice the capacity for graphite composition LiC6) is recovered during the first electrochemical cycle. The micropores may act as superficial sites for lithium storage. Unfortunately, it is associated to an irreversible loss of the capacity corresponding to 40% of the first discharge capacity and a large polarization. These two drawbacks are due to the presence of heteroatoms, mostly oxygen atoms which are forming carbonyls and lactones in the carbonaceous material, thus entrapping lithium ions by the formation of Li2O during the discharge process. For an application of the ICPs as anode material in Li ion battery, a special care is needed to avoid the presence of heteroatoms responsible for the polarization during the charge process Some recent research performed at the University of Cornell were dedicated to nanocomposite polymer electrolytes. Some of them include polymer / layered silicate. The clay or nanosized ceramic powder can perform as solid plasticizers for ionically conducting polymer membranes, and at the same time inhibiting the polymer crystallization [148]. Once again, LDH may play a role in this field. The polymer / hydrocalumite system may be considered as a promising nanocomposite in the field of MDF cement-related material. Formation of nanocomposite from calcium silicate layered structure requires a sol-gel approach using organotrialkoxysilane, similarly to phyllo(organo)silicate [149-150], the covalent bonding is taking place via the silanol groups, whereas incorporation of either monomer or polymer proceeds easily for the hydrocalumite framework. Considering the recent developments to exfoliate LDH layers, the use of LDH as nanofillers is an emerging domain of application, where LDH material may present advantages in comparison to smectites such as montmorillonite due to its versatility in the chemical composition and the tunable charge density, allowing multiple anchorages with the polymer. Moreover the presence of iron cations and alkyl ammonium molecules are drawbacks for the use of smectites-based nanocomposites, increasing the photodegradation and decreasing the stability in temperature, respectively. Using extended version of the discrete finite-layer rigidity model including intra- and interlayer rigidity effects, some authors conclude that graphite is much more floppy than Ni2Al LDH carbonate phase while vermiculite is more rigid [151]. Despite the fact that LDH materials appears to represent one of the least rigid clay systems, according to models that did not take into account organo-modified phases, polymer LDH nanofillers are thought to perform as well as smectites, as recently evidenced by the works of Hsueh and Chen [130,132]. Flame-retardant properties may be enhanced due to the presence of hydroxyl groups in great numbers. Some of these applications are patented, and we believe that with the ever-growing need of our society for more and more sophisticated multicomponent systems, new LDH nanocomposites based on copolymer mixtures such as Noryl™, poly(styrene - co - acrylonitrile) (SAN), poly(acrylonitrile -co- butadiene - co - styrene) (ABS) etc., will emerge in different field of applications, food packaging, car and tyre industries, etc. To rise all these new tasks, LDH materials have to meet important requirements, i.e. control of the particle size, porosity, and processability. An enhanced compatibility between polymer and LDH is crucial since the interfacial
490
F. Leroux and J. -P. Besse
properties are often the key factor to high stiffness, high modulus and heat resistant composites [152]. 8 - Summary From several studies, it is observed that the multi-components systems polymer / LDH are thermally more stable than the pristine inorganic compounds, leading for example to potential applications in flame-retardant composites. Similarly to other host structures, LDH materials provide a confinement for the organic moiety suitable after a charring process to the preparation of high surface area carbonaceous materials, which are of use in adsorption processes. The incorporation of polymer into hydrocalumite host, a LDH-type material is of current concern for cement application. Indeed, like calcium silicate (noted as C-SH), these calcium aluminium hydroxide salts called Friedel salts occur in the hydration process of cement and their role in the mechanical properties is of major importance. Moreover, the filling of the large voids present initially within the cement is a key factor to reinforce the whole structure, the assembly is defined as Macro Defect Free (MDF) cements. An other incentive aspect is the use of LDH materials as nanolayers for filler in polymeric matrix. Largely studied in the case of smectite-type materials, some recent results show similar trends for LDH nanofillers, i.e. an increase of the mechanical properties and of the polymer glass transition temperature. It shows that the polymer/LDH assembly, not yet extensively studied, constitutes an appealing new class of nanocomposites in numerous topical applications. Acknowledgments. The authors would like to thank PhD student El Mostafa Moujahid, Dr. Christine Taviot-Gueho and Dr. Marc Dubois for their fruitful discussions. 9 - References [I] Polymer-Clay Nanocomposites, Eds T. J. Pinnavaia and G. W. Beall, John Wiley and Sons (2000) [2] F. Leroux and J.-P. Besse, Chem. Mater., 13 (2001) 3507. [3] CO. Oriakhi, J. Chem. Edu., 77 (2000) 1138. [4] F.M. Labajos, M.D. Sastre, R. Trujillano and V. Rives, J. Mater. Chem., 9 (1999) 1033. [5] F. Malherbe, C. Forano and J.-P. Besse, J. Mater. Sci. Lett., 18 (1999) 1217. [6] B. Zapata, P. Bosh, G. Fetter, M.A. Valenzuela, J. Navarrete and V.H. Lara, Inter. J. Inorg. Mater., 3(2001)23. [7] J.M. Fernandez, C. Barriga, M.A. Ulibarri, F.M. Labajos and V. Rives, Chem. Mater., 9(1997)312. [8] M.G. Kanatzidis, R. Bissessur, D.C. DeGroot, J. Schindler and C.R. Kannewurf, Chem. Mater., 5 (1993) 595. [9] H.-L. Tsai, J.L. Schindler, C.R. Kannewurf and M.G. Kanatzidis, Chem. Mater., 9 (1997)875. [10] A. Fudala, I. Palinko and I. Kiricsi, Inorg. Chem., 38 (1999) 4653. [II] E. Lopez-Salinas, M. Garcia-Sanchez, J.A. Montoya, D.R. Acosta, J.A. Abasolo and I. Schifter, Langmuir, 13 (1997) 4748. [12] A.J. Jacobson, Mater. Sci. Forum, 152-153 (1994) 1. [13] Y. Zhao, F. Li, R. Zhang, D.G. Evans and X. Duan, Chem. Mater., 14 (2002) 4286. [14] J-M. Oh, S.-H. Hwang and J.-H. Choy, Solid State Ionics, 151 (2002) 285-291.
Layered Double Hydroxide / Polymer Nanocomposites
491
[15] M. Ogawa and H. Kaiho, Langmuir, 18 (2000) 4240. [16] S. Abend, N. Bonnke, U. Gutschner and G. Lagaly, Colloid Polym Sci., 276 (1998) 730 [17] G. Lagaly, O. Mecking and D. Penner, Colloid and Polymer Science, 279 (2001) 1090. [18] E. Gardner, K.M. Huntoon and T.J. Pinnavaia, Adv. Mater., 13 (2001) 1263. [19] I. Dekany, F. Berger, K. Imrik and G. Lagaly, Colloid Polym Sci., 275 (1997) 681. [20] E.L. Crepaldi, P.C. Pavan and J.B. Valim, Chem. Commun., (1999) 155. [21] W.J. McLaughlin, J.L. White and S.L. Hem, J. Colloid Interface Sci., 165 (1994) 41. [22] S. Liu, J. Zhang, N. Wang, W. Liu, C. Wang and D. Sun, Chem. Mater., 15 (2003) 3240. [23] S.O'Leary, D.O'Hare and G. Seeley, Chem. Comm., (2002) 1506. [24] M. Adachi-Pagano, C. Forano and J.-P. Besse, Chem. Com., (2000) 91. [25] T. Hibino and W. Jones, J. Mater. Chem., 11 (2001) 1321. [26] F. Leroux, M. Adachi-Pagano, M. Intissar, S. Chauviere, C. Forano and J.-P. Besse, J. Mater. Chem., 11 (2001) 105. [27] B. Li, Y. Hu, R. Zhang, Z. Chen and W. Fan, Mater. Res. Bull.Uncorrected proof on line (2003). [28] G.A. Bubniak, W.H. Schreiner, N. Mattoso and F. Wypych, Langmuir, 18 (2002) 5967. [29] C. Sanchez and F. Ribot, New J. Chem., 18 (1994) 1007. [30] M.C. Kanatzidis, C.G. Wu, H.H. Marcy, D.C. DeGroot and C.R. Kannewurf, Chem. Mater., 2 (1990) 222. [31] R. Schollhorn, Chem. Mater., 8 (1996) 1747. [32] M. Tanaka, I.Y. Park, K. Kuroda and C. Kato, Bull. Chem. Soc. Japan, 62 (1989) 3442. [33] C. Vaysse, L. Guerlou-Demourgues, E. Duguet and C. Delmas, Inorg. Chem., 42 (2003) 4559. [34] S. Rey, J. Merida-Robles, K.-S. Han, L. Guerlou-Demourgues, C. Delmas and E. Duguet, Polym. Intern., 48 (1999) 277. [35] El M. Moujahid, J.-P. Besse and F. Leroux, J. Mater. Chem., 12 (2002) 3324. [36] CO. Oriakhi, I.V. Faar and M.M. Lerner, J. Mater. Chem., 6 (1996) 103. [37] CO. Oriakhi, I.V. Faar and M.M. Lerner, Clays Clay Miner., 45 (1997) 194. [38] H.P. Boehm, J. Steinle and C. Vieweger, Angew. Chem. Int. Ed. Engl., 16 (1977) 265. [39] T.A. Kerr, F. Leroux and L.F. Nazar, Chem. Mater., 10 (1998) 2588. [40] CO. Oriakhi and M.M. Lerner, Chem. Mater., 8 (1996) 2016 [41] T. Itagaki, Y. Komori, Y. Sugahara and K. Kuroda, J. Mater. Chem., 11 (2001) 3291 [42] Y. Komori, Y. Sugahara and K. Kuroda, Chem. Mater., 11 (1999) 3. [43] A.M. Drezdon, Inorg. Chem., 27 (1988) 4628. [44] E.D. Dimotakis and T.J. Pinnavaia, Inorg. Chem., 29 (1990) 2393. [45] W.-Y Tseng, J.-T Lin, C.-Y. Mou, S. Cheng, S.-B. Liu, P.P. Chu and H.-W. Liu, J. Am. Chem. Soc, 118 (1996) 4411. [46] S.K. Yun, V.R.L. Constantino and T.J. Pinnavaia, Clays and Clay Miner., 43 (1995) 503. [47] A. Schutz and P. Biloen, J. Solid State Chem., 68 (1987) 360. [48] C. Depege, F.-Z. El Metoui, C Forano, A. de Roy, J. Dupuis and J.-P. Besse, Chem. Mater., 8(1996)952.
492
F. Leroux and J. -P. Besse
[49] M.C. Kanatzidis and C.G. Wu, J. Am. Chem. Soc, 111 (1989) 4139. [50] T. Challier and R.C.T. Slade, J. Mater. Chem., 4 (1994) 367. [51] El M. Moujahid, M. Dubois, J.-P. Besse and F. Leroux, Chem. Mater., 14 (2002) 3799. [52] X. Wei and A. J. Epstein, Synth. Met, 74 (1995) 123. [53] N. Kuramoto and E. M. Genies, Synth. Met., 68 (1995) 191. [54] S.-A. Chen and G.-W. Hwang, Polymer, 38 (1997) 3333 [55] H. Tang, A. Kitani, T. Yamashita and S. Ito, Synt. Metals, 96 (1998) 43 [56] S. Shimizu, T. Saitoh, M. Uzawa, M. Yuasa, K. Kano, T. Maruyama and K. Watanabe, Synt. Metals, 85 (1997) 1337. [57] J. Yue, Z.H. Wang, K.R. Cromack, A.J. Epstein and A.G. MacDiarmid, J. Am. Chem. Soc, 113(1991)2665. [58] Unpublished results [59] V.P. Isupov, L.E. Chupakhina, M.A. Ozerova, V.G. Kostrovsky and V.A. Poluboyarov, Solid State Ionics, 141-142 (2001) 231. [60] S.W. Rhee and D.-Y. Jung, Bull. Korean Chem. Soc, 23 (2002) 35-40 [61] H. Tagaya, S. Sato, T. Kuwahara, J. Kadokawa, K. Masa and K. Chiba, J. Mater. Chem., 4 (1994) 1907. [62] P.B. Messersmith and S.I. Stupp, Chem. Mater., 7 (1995) 454. [63] O.C., Jr. Wilson, T. Olorunyolemi, A. Jaworski, L. Borum, D. Young, A. Siriwat, E. Dickens, C. Oriakhi and M. Lerner, Applied Clay Sci., 15 (1999) 265. [64] F. Leroux, P. Aranda, J.-P. Besse and E. Ruiz-Hitzky, Eur. J. Inorg. Chem., 6 (2003) 1242. [65] M. del Arco, C. Martin, I. Martin, V. Rives and R. Trulillano, Spectrochim. Acta 49A, (1993) 1575. [66] C.-G. Wu, D.C. DeGroot, H.O. Marcy, J.L. Schindler, C.R. Kannewurf, T. Bakas, V. Papaefthymiou, W. Hirpo, J.P. Yesinowski, Y.-J. Liu and M.G. Kanatzidis, J. Am. Chem. Soc, 117 (1995) 9229. [67] W. Hofineister and H. von Platen, Cryst. Rev, 3 (1992) 3. [68] M. Vucelic, W. Jones and G.D. Moggridge, Clays Clay Miner, 45 (1997) 803. [69] M. Intissar, R. Segni, C. Payen, J.-P. Besse and F. Leroux J. Solid State Chem, 167 (2002) 508. [70] H. Roussel, V. Briois, E. Elkaim, A. de Roy and J.-P. Besse, J. Phys. Chem. B, 104 (2000) 5915. [71] N.E. Brese and M.O1 Keeffe, Acta Cryst. B, 47 (1991) 192. [72] S.E. Lopez and Y. Ono. Bull. Soc Japan, 65 (1992) 2465. [73] L. Vieille, C. Taviot-Gueho, J.-P. Besse and F.Leroux. Chem. Mater., 15 (2003) 4369.
[74] R. J. Elliot, Phys. Rev. B, 96 (1954) 266 [75] S. Uma and J. Gopalakrishnan, Mat. Sci. Eng. B, 34 (1995) 175. [76] N.S.P. Bhuvanesh and J. Gopalakrishnan, Mater. Sci. Eng. B, 53 (1998) 267. [77] H. Nakajima and G. Matsubayashi, Chem. Lett, 423 (1993). [78] N. Kinomura, T. Toyama and N Kumada, Solid State Ionics, 78 (1995) 281 [79] H.H.S Javadi,R. Laversanne, A.J. Epstein, R.K. Kholi, E.M. Scherr and A.G. Mac Diarmid, Synth. Met, 29 (1989) E439 [80] El M. Moujahid, J. Inacio, J.-P. Besse and F. Leroux, Micropor. Mesopor. Mater, 57 (2003) 37. [81] P.C. Pavan, G. de A. Gomes and J.B. Valim, Micropor. Mesopor. Mater, 21 (1998) 659.
Layered Double Hydroxide / Polymer Nanocomposites
493
[82] P.C. Pavan, E.L. Crepaldi, G. de A. Gomes and J.B. Valim, Colloids Surf. A, 154 (1999) 399. [83] C.H.Giles, T.H.MacEwan, S.N. Nakhwa and D. Smith, J. Chem. Soc, (1960) 3973. [84] El M. Moujahid, J.-P. Besse and F. Leroux, J. Mater. Chem., 13 (2003) 258. [85] L. Vieille, El M. Moujahid, C. Taviot-Gueho, J.-P. Besse and F. Leroux. J. Phys. Chem. Solids, in press [86] Y. Sugahara, N. Yokoyama, K. Kuroda and C. Kato. Ceramics Inter., 14 (1988) 163. [87] M.A. Aramendia, Y. Aviles, V. Borau, J.M. Luque, J.M. Marinas, J.R. Ruiz and F.J. Urbano, J. Mater. Chem., 9 (1999) 1603. [88] J. Rocha, M. del Arco, V. Rives and M.A. Ulibarri, J. Mater. Chem., 9 (1999) 2499. [89] O.C., Jr. Wilson, T. Olorunyolemi, A. Jaworski, L. Borum, D. Young, E. Dickens and C. Oriakhi, Ceramic Transactions,! 10 (2000) 93. [90] D. Bastos and F.J. de las Nieves, Colloid Polym. Sci., 271 (1993) 860. [91] H. Kopka, K. Beneke and G. Lagaly, K., J. Colloid Interface Sci., 123 (1988) 427. [92] Y. You, H. Zhao and G.F. Vance, J. Mater. Chem., 12 (2002) 907. [93] H. Chmielewski, PCT Int. Appl. (2001), patent number WO 2000-US30240. [94] J. Tronto, K.C. Sanchez, E.L. Crepaldi, Z. Naal, S.I. Klein and J.B. Valim, J. of Phys. Chem. Solids, in press. [95] El M. Moujahid, F. Leroux, M. Dubois and J.-P. Besse, Comptes Rendus Chimie, 6 (2003) 259. [96] M.C. Miras, C. Barbero, R. Kotz and O. Haas, J. Electrochem. Soc, 138 (1991) 335. [97] K. Yamada, K. Teshima, N. Kobayashi and R. Hirohashi, J. Electroanal. Chem., 394(1995)71. [98] K.L.N. Phani, S. Pitchumani and S. Ravichandran, Langmuir, 9 (1993) 2455. [99] H. Inoue and H. Yoneyama, Electroanal. Chem., 233 (1987) 291. [100] G. Xu, C. Guo, X. Duan and C. Jiang. Yingyong Huaxue, 16 (1999) 45 (In Chinese). [101] K. Kotani and T. Sakaya. European patent number EP0715954 (1995). [102] R.G. Ford, A.C. Andreas, K.G. Scheckel and D.L. Sparks, Environmental Sci. Techn., 33(1999)3140. [103] Y. Seida and Y. Nakano, Water Resources, 34 (2000) 1487. [104] K.R. Franklin and J.D. Hewitt-Jones. PCT Int. Appl. (1997) Patent number WO 1993-EP3097. [105] G.T.D. Shouldice, P.Y. Choi, B.E. Koene, L.F. Nazar and A. Rudin, J. Polymer Sci., 33 (1995) 1409. [106] M. Chibwe, L. Ukrainczyk, S.A. Boyd and T.J. Pinnavaia, J. Mol. Catalysis A: Chemical, 113(1996)249. [107] L. Ren, J. He, D.G. Evans, X. Duan and R. Ma, J. Mol. Catalysis B: Enzymatic, 16 (2001) 65. [108] T. Kyotani, N. Sonobe and A. Tomita, Nature, 28 (1988) 331. [109] R. Ryoo, S.H. Joo and S.J. Jun, Phys. Chem. B, 103 (1999) 7743. [110] D. Kawashima, T. Aihara, Y. Kobayashi, T. Kyotani and A. Tomita, Chem. Mater., 12(2000)3397. [ I l l ] J.Y. Kim, S.B. Yoon and J.-S. Yu, Chem. Mater., 15 (2003) 1932. [112] K. Putyera, T.J. Bandosz, J. Jagieo and J.A. Schwarz, Applied Clay Sci., 10 (1995) 177. [113] K. Putyera, T.J. Bandosz, J. Jagieo and J.A. Schwarz, Carbon, 34 (1996) 1559.
494
F. Leroux and J. -P. Besse
[114] T. Bandosz, J. Jagieo, K. Putyera and J.A. Schwarz, Chem. Mater., 8 (1996) 2023. [115] K. Putyera, J. Jagiello, T.J. Bandosz and J.A. Schwarz, Carbon, 33 (1995) 1047. [116] V. Fernon, A. Vichot, P. Colombet, H. van Damme and F. Beguin, Mater. Sci. Forum, 152-153 (1994) 335. [117] H. Kuzel and H Baier, H. Eur. J. Miner., 8 (1996) 129. [118] A.A. Bonapasta, F. Buda and P. Colombet, Chem. Mater., 12 (2000) 738. [119] A.A. Bonapasta, F. Buda and P. Colombet, Chem. Mater., 13 (2001) 64. [120] A.A. Bonapasta, F. Buda, P. Colombet and G. Guerrini, Chem. Mater., 14 (2002) 1016. [121] F. Merlin, H. Lombois, S. Joly, N. Lequeux, J.-L. Halary and H. van Damme, J. Mater. Chem., 12 (2002) 3308. [122] H. Ishida, S. Campbell and J. Blackwell, Chem. Mater., 12 (2000) 1260. [123] E.P. Giannelis, Adv. Mater., 8, (1996) 29. [124] K.E. Strawbecker and E. Manias, Chem. Mater., 12 (2000) 2943. [125] D. Hawthorne, J.H. Hodgkin, B.C. Loft and D.H. Solomon, J. Macromol. Sci. Chem., 8 (1974) 649. [126] C. Kato, K. Kuroda and M. Misawa, Clays Clay Miner., 27 (1979) 129. [127] A. Usuki, Y. Kojima, M. Kawasumi, A. Okada, Y. Fukushima, T. Kuruachi and O. Kamigaito, J. Mater. Res., 8 (1993) 1179. [128] O. Kamigaito, M. Kawasumi, M. Hohzaki, Y. Kojima and A. Okada, US Patent. 4810743 (1993). [129] K. Yano, A. Usuki, A. Okada, T. Kurauchi and O. Kamigaito, J. Polym. Sci. A, 31(1993)2493. [130] H.-B. Hsueh and C.-Y. Chen, Polymer, 44 (2003) 1151. [131] C.-M Leu., Z.-W. Wu and K.-H. Wei, Chem. Mater., 14 (2002) 3016. [132] H.-B. Hsueh and C.-Y. Chen, Polymer, 44 (2003) 5275. [133] B. Li, Y. Hu, J. Liu, Z. Chen and W. Fan, Colloid Polym. Sci., 281 (2003) 998. [134] L. Zhang, L. Darning, L. Wenbao and H. Wang, Suliao, 30 (2001) 22 (In Chinese) [135] H.R. Fisher and L.H. Gielgens, International Patent number WO99/35186 (1999). [136] W. Chen and B. Qu, Chem. Mater., 15 (2003) 3208. [137] N. Schall, T. Engelhardt, H. Simmler-Hubenthal and G. Beyer, PCT Int. App. (2000). Patent number 2000-EP2115. [138] J. Xue, J. Hao, Y. Zhao and X. Duan. Xiandai Huagong, 22 (2002) 34-37. (in Chinese). [139] N.T. Whilton, P.J. Vickers and S. Mann, J. Mater. Chem., 7 (1997) 1623. [140] F. Leroux, J. Gachon and J.-P. Besse. J. Solid State Chem., in press [141] J.-H. Choy, S.-Y. Kwak, J.-S. Park and Y.-J. Jeong, J. Mater. Chem., 11 (2001) 1671. [142] J.-H. Choy, S.-Y Kwak, J.-S. Park, Y.-J. Jeong and J. Portier, J. Am. Chem. Soc, 121 (1999) 1399. [143] J.-H. Choy, S.-Y Kwak, Y.-J. Jeong and J.-S. Park, Angew. Chem. Int. Ed., 39 (2000) 4042. [144] J.-H. Choy, J.-S. Park, S.-Y. Kwak, Y.-J. Jeong and Y.-S. Han, Mol. Crystals and Liquid Crystals, 341 (2000) 425. [145] M. Darder, M. Colilla and E. Ruiz-Hitzky, Chem. Mater., 15 (2003) 3774-3780. [146] J.M. Miller, B. Dunn, T.D. Tran and R.W. Pekala, J. Electrochem. Soc, 144 (1997) L309. [147] G. Sandi, R.E. Winans and K.A. Carrado, J. Electrochem. Soc, 143 (1996) L95. [148] F. Croce, G.B. Appetecchi, L. Persi and B. Scosati, Nature, 394 (1998) 456. [149] E. Muthusamy, D. Walsh and S. Mann, Adv. Mater., 14 (2002) 969.
Layered Double Hydroxide / Polymer Nanocomposites
495
[150] M.G. Fonseca and C. Airoldi, J. Mater. Chem., 10 (2000) 1457. [151] S.A. Solin, D. Hines, S.K. Yun, T.J. Pinnavaia and M. F. Thorpe, J. Non-Cryst. Solids, 182(1995)212. [152] M. Okamoto, S. Morita, H. Taguchi, Y.H. Kim, T. Kotaka and H. Tateyama, Polymer 41(2000)3887.
CATALYTIC PROPERTIES OF HYDROTALCITE-TYPE ANIONIC CLAYS SIMONE ALBERTAZZI, FRANCESCO BASILE and ANGELO VACCARI* Dipartimento di Chimica Industriale e dei Materiali, Alma Mater Studiorum Universita di Bologna, INSTM-UdR di Bologna, Viale del Risorgimento 4, 40136 Bologna - ITALY. * E-mail: [email protected]
Clay Surfaces: Fundamentals and Applications F. Wypych andK.G. Satyanarayana (editors) © 2004 Elsevier Ltd. All rights reserved.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
497
1 -Introduction Layered double hydroxides (LDHs) [or hydrotalcite-type (HT) anionic clays] are natural or synthetic mixed hydroxides with interlayer spaces containing exchangeable anions, relatively simple and cheap to synthesize on both laboratory and industrial scales [1-8]. They may be described by the general formula: [Mz+j.x M3+x (OH)2]b+ [An"b/n] mH2O (M= metal, A= interlayer anion, and b= x or 2x-l, for z = 2 or 1 respectively) and many names are used depending on the composition and polytype [9]. Feitknecht [10,11] first used the term LDHs, hypothesizing a structure with intercalated hydroxide layers, although many years later the single crystal XRD analysis [12,13] showed that all cations were localized in the same layer, with the anions and water molecules located in the interlayer region. On the other hand, the name HT compounds is probably due to the extensive characterization carried out on natural or synthetic hydrotalcite (Mg/Al hydroxycarbonate), while the term anionic clays was first used by Reichle [14] to underline the complementarity with the properties of cationic clays (or clay minerals). However, none of these terms is accepted by all. LDHs can be considered promising materials for a large number of possible applications due to their high versatility, low cost, easily manipulated properties, wide range of composition and/or preparation variables (Table 1), which make it possible to produce tailor-made materials to fulfil specific requirements. Table 1 - Composition and preparation variables in the synthesis of LDHs [1-7]. Composition variables Cation size Value of x Cation stereochemistry Cation nature and ratio Nature of balancing anions Amount of interlayer water Crystal morphology and size
Preparation variables pH Precipitation method Precipitation temperature Reagent concentration Aging Washing and drying Presence of impurities
Thus, it is not surprising that synthetic LDHs, as such or after thermal decomposition, find many industrial applications. Indeed, over the past few years there has been an exponential increase in such applications in both open and patent literature [1-3,5-8,15-21]. Furthermore, the possibility of technological upgrading, i.e. the transition from two- to three-dimensional structures by pillaring or intercalation processes, opens new prospects for the preparation of unusual materials [5,8]. Although the largest amounts of LDHs are used in the polymer industry, mainly to stabilize PVC [1,5-7,22], the most promising applications are as precursors of catalysts or catalyst supports, because of the specific features of the mixed oxides obtained by controlled thermal decomposition: 1) surface area values of 100-300m2/g 2) homogeneous and thermally stable interdispersion of the elements, which by reduction form small and stable metal crystallites 3) synergetic effects between the elements, which favour multifunctional properties 4) memory effect, with structure reconstruction under mild conditions. However, the research on LDHs and catalysis followed separate paths up to the year
498
S. Albertazzi, F. Basile and A. Vaccari
1970, when the first patent appeared that referred specifically to them as optimal precursors for hydrogenation catalysts [23] (Table 2). Table 2 - First examples of LDHs as precursors of industrial catalysts and their applications after calcination [23). LDH precursor Mg6Al2CO3(OH)16-4H2O Ni6Al2CO3(OH)16 4H2O Ni3Mg3Al2CO3(OH)16 4H2O Co3Mg3Al2CO3(OH)16 4H2O Co6Al2CO3(OH)16 4H2O Ni0.9oCoo75Cuo.35Mg4Al2C03(OH)164H20 Cu3Mg3Al2CO3(OH)16 4H2O Cu6Al2CO3(OH)16 4H2O Cu3Zn3Al2CO3(OH)16 4H2O Ni3Zn3Al2CO3(OH)16 4H2O Ni3Mg3Al18Cr0.2CO3(OH)16 4H2O
Catalytic application Dehydration, catalyst support Hydrogenation, dealkylation Hydrogenation, dealkylation, cracking Hydrogenation Hydrogenation Hydrogenation, dehydrogenation Dehydrogenation of sec-alcohols, hydrogenation of nitro groups Isomerization, hydrogenation of nitro groups, dehydrogenation of sec-alcohols Dehydrogenation of sec-alcohols, lowtemperature water gas shift conversion Hydrogenation Hydrogenation
Established catalytic applications of the mixed oxides obtained by controlled calcination of LDHs (polymerization of alkene oxides, aldol condensation of aldehydes and ketones, methane or hydrocarbon steam reforming, methanation, methanol synthesis, higher-alcohols or hydrocarbons synthesis, etc.) have been extensively treated in previous reviews [1-3,5-7,17-19,21], Thus, the attention here is focused on more recent applications, as promising areas of research. However, upgrading may also refer to new compositions or preparation methods. New compositions include LDHs without trivalent cations [24] or those containing unstable V3+ ions [25], noble-metal ions [2631] or tetravalent ions (Ti4+, Zr4+or Sn4+) [32-35]. Examples of new preparations include LDHs exhibiting sheet broadness to thickness ratios ranging from 100 to 2000 [36-38], sol-gel syntheses [39-41] and recent syntheses claiming significant material or process improvements [42-46]. 2 - Precursors of catalyst supports LDHs with different compositions, calcined at 523-723 K and partially or completely chlorinated, have been claimed as supports for Ziegler catalysts for the polymerization of olefins, showing higher activities than catalysts prepared from (MgCO 3 ) 4 Mg(OH) 2 H 2 O and better control of the molecular weight (Table 3) [47]. More recently, calcined Mg/Al LDHs have been reported to support CeO2 for SOx removal from the emissions of fluid catalytic cracking units (FCCU) (see also Section 5), with maximum activity for the Ce02/MgAl2O4-Mg0 system [48-50]. These catalysts are very stable even after severe steam treatment, showing hardness values very close to those of typical FCC catalysts, and good catalytic properties as well as easy catalyst regeneration. Mg/Al mixed oxides have also been successfully used as supports for transition metal oxides for the selective catalytic reduction (SCR) of NO by NH3, as alternatives to less thermally stable supports such as active carbons [51].
Catalytic Properties of Hydrotalcite-Type Anionic Clays
499
Table 3 - Activity of TiCl4 (VC14)/AIR3 catalysts supported on mixed oxides obtained by calcination of different LDHs (anions always carbonates) [47]. Calcination Calcination Active phase time (h) temperature (K) TiCl4/AlR3 563 15 Ni/Al 523 10 Mg/Zn/Al TiCVAIRs TiCl4/AlR3 603 30 Mg/Cr 473 5 Mg/Mn/Al TiCVAIRs VCI4/AIR3 503 10 Mg/Co/Mn/Cu/Al 593 20 VCI4/AIR3 Co/Cr 593 40 Mg/Al/Cr TiCVAlR3 723 3 Mg/Al ZrCl4/AlR3 LDH
Catalytic activity (g of polyethylene) 130 135 165 180 130 135 160 130
Among the three supported transition metal oxides, the best catalytic performances were obtained with CuO, while Cr2O3 formed high amounts of N2O and Fe2O3 exhibited low NO conversion, unlike that observed on active carbons at lower temperatures [52]. However, new applications continuously appear, such as for example that of V or Mo oxides supported on calcined Mg/Al LDHs for the oxidative dehydrogenation of n-butane [53,54! and propane [55,56] or the vapor phase synthesis of i-butyraldehyde from methanol and n-propanol [57]. In this latter reaction, since the synthesis took place through metal enolates as intermediates, the activity depended mainly on the acid-base properties and the selectivity on the method of V oxide loading. Calcined LDHs have also been reported as useful supports of noble metals. The conversion of «-hexane to aromatic hydrocarbons has been claimed using calcined LDHs (typically for 12 h at 873 K) prior to Pt or Pd impregnation [58,59], For this process, catalysts obtained by calcination of LDH precursors containing noble metal ions inside the structure have also been claimed [26]. Chemisorption techniques have shown that nanometric sized Pt particles supported on Mg/Al mixed oxides prepared from LDHs compared with those supported on A12O3 (with similar Pt dispersion) exhibited particular properties, such as a higher differential heat of H2 adsorption or the presence of most thermally-stable CO species, bridge-bonded on Pt through the C atom and with Mg2+ ions by the O atom [60]. Ni- and Pd-supported catalysts prepared by impregnation of Mg/Al LDHs calcined at 823 K for 18 h, followed by drying and further calcination at 673 K for 4 h, have been successfully employed in the one-step synthesis of methyl i-butyl ketone (MIBK) from acetone (DMK) and H2 at atmospheric pressure [61]. Basic properties of the support and density of metal sites were key factors for activity and product distribution, with better results than those reported for previous catalysts, although stability problems in the reaction conditions still remained. The same reaction has also been investigated using Ni/Mg/Al, Co/Mg/Al and Fe/Mg/Al mixed oxides obtained by calcination of LDHs at 673 K for 15 h, and successive reduction at 723 K for 24 h [62]. Ni/Mg/Al gave selectively MIBK, whereas Co/Mg/Al gave methyl i-butyl carbinol (MIBC) and Fe/Mg/Al produced mesityl oxide (MSO), which are all valuable chemicals (Fig. 1).
500
S. Albertazzi, F. Basile and A. Vaccari
Figure I - Reaction pathway for the conversion of acetone in the presence of H2 over some mixed oxides obtained by calcination and successive reduction ofLDHs (MIBK = methyl i-butyl ketone; MIBC = methyl i-butyl carbinol; MSO = mesityl oxide) [62]. In other cases, the noble-metal supported catalysts have been prepared by direct impregnation of LDHs followed by reduction, observing unusual behaviors. For example, Ru crystallites obtained by reduction at 548 K of carbonyl complexes supported on Mg/Al LDHs dried at 338 K, formed by CO hydrogenation high amounts of oxygenates (mainly methanol and lower amounts of C2-C4 alcohols), unlike analogous catalysts prepared using acid supports that were very selective in hydrocarbons (branched alkanes and internal alkenes) [63]. Moreover, catalysts obtained by impregnating RuCl3-xH2O on some commercial LDHs have also been claimed, after reduction at 423 K under 5.0 MPa of H2, for the partial reduction of monocyclic aromatic hydrocarbons to cycloolefins, with a key role of the support on the dispersion and stability of Ru crystallites [64]. More recently, catalysts prepared by impregnation of different Mg/Al LDHs with various Pd anionic precursors and following reduction at 573 K for 3 h, have been investigated in CO chemisorption at room temperature and phenol hydrogenation at 453 K [65,66]. These catalysts showed better performances than Pd/Al2O3 and Pd/MgO in phenol hydrogenation, with selectivity to cyclohexanone higher than 90%. The best values were obtained with LDHs prepared by the high-supersaturation method and with interlayer carbonates, impregnated with acidified PdCl2. It has been proposed that acidification led to PdCl42", which replaced carbonates on the external edge surface, leading to fine dispersion and, after reduction, small Pd particles. An Mg/Cr LDH heated in N2 at 723 K for 18 h and impregnated with a HCl-acidified aqueous PdCl2 solution has been recently proposed for the vapor phase hydrogenolysis of CC12F2 [67]. Finally, the most recent and interesting application is the use of calcined commercial LDHs as supports of Pt or Pt/Cu to prepare novel NOx storage-reduction (NOXSR) catalysts for diesel light-duty engine emissions [68-71]. These catalysts showed better performances in NOx storage than the already known lPt-15Ba/Al2O3 Toyota-type NOXSR catalyst at reaction temperatures lower than 523 K, due to the lower basicity of the Mg/Al mixed oxide in comparison to BaO, which induces both a
Catalytic Properties of Hydrotalcite-Type Anionic Clays
501
lower inhibition on Pt activity (NO to NO2 oxidation and/or hydrocarbon oxidation) and a lower thermal stability of the stored NOX. [69,70]. However, the nature of LDH precursor considerably affected the catalytic performances [69]. The co-presence of Pt and Cu slightly worsened the low temperature activity, but considerably promoted the resistance to deactivation in the presence of SO2 or after severe hydrothermal treatment (Fig. 2).
Figure 2 - Mean NOX conversion as a function of the reaction temperature before and after hydrothermal treatment for the catalysts: (i) lPt-4Cu/LDH [1 wt. % Pt on calcined LDH modified by impregnation of 4 wt.% CuO prior to loading ofPt], and (ii) lPt15Ba [1 wt.%Pt on 15 wt.% Ba-yAl2O3 (Toyota-type NOXSR catalyst)] [69-71]. These effects can be attributed to the possible formation of a Pt-Cu alloy, highlighted by FT-IR characterisation [68,70], and the modification of Mg/Al mixed oxide surface induced during Cu impregnation. In any case, the overall lPt-4Cu/LDH performances were superior to those of the lPt-15Ba/Al2O3 Toyota-type catalysts. 3 - Catalysts for organic syntheses LDHs as such exhibit generally poor basic properties, considerably lower than those of the mixed oxides obtained by their thermal decomposition [2,3,5,6]. However, good basic properties have been recently claimed for meixnerite, obtained by combining decarboxilation and controlled rehydration [72-75]. High activity and selectivity have been reported in the aldolisation of acetone (Fig. 3), already widely investigated using calcined LDHs [76-79], as well as in the aldolic condensation of benzaldehyde with acetone or acetophenone, although in this last reaction KF supported on alumina preliminarily treated at 723 K, was most active at 423 K, reaching 90% selectivity in chalcone at 80% conversion [80]. The presence of strong basic sites has also been claimed for LDHs activated below the structural decomposition temperature (< 523K) [81], suggesting that adsorbed water can inhibit the access to these basic sites on the surface. However, most applications refer to the mixed oxides obtained by thermal decomposition in the range between the decomposition temperature of the LDH structure and that of formation of stoichiometric phases, in which metastable poorly crystallized mixed oxides form, characterized by high surface area and non-stoichiometric compositions, with an excess of divalent cations [3,5,6].
502
S. Albertazzi, F. Basile and A. Vaccari
Figure 3 - Reaction pathway for the aldol condensation of acetone [3,76]. Mg/Al mixed oxides offer the possibility of replacing liquid bases, which pose severe environmental problems, with environmentally friendly solid catalysts that can be easily separated and recycled. A good recent example is the isomerization of eugenol and safrole, important intermediates in the synthesis of Pharmaceuticals and fragrances, using dried LDHs with different Mg/Al atomic ratios [82]. Mg/Al LDHs showed considerable activity, which varied with the composition: in the reaction of eugenol to ieugenol an Mg/Al atomic ratio of 4.0 showed maximum activity with the conversion of 73% ca. and a cis:trans ratio of 17:83, while the sample with Mg/Al ratio of 6.0 exhibited maximum activity in the isomerization of safrole to i-safrole with 75% ca of conversion and a cis:trans ratio of 15:85. Trans isomers were formed predominantly in both cases due to their better thermodynamic stability. The textural and surface properties of the Mg/Al mixed oxides as well as their activity may be modulated by an appropriate selection of the composition and preparation conditions [83-90]. For example, with increasing Mg/Al ratio, the number of basic sites with 9.0 < pK < 13.3 increases, whereas that of the stronger basic sites (13.3 < pK < 16.5), which catalyze for example, Michael addition or aldol condensation reactions decrease (Fig. 4) [91]. However, good yields in the Knoevenagel condensation of aromatic and aliphatic aldehydes with ethylacetonate and malonitrile have been recently reported at 333 K and in solvent-free conditions using a non-thermally activated Ni/Al LDH [92]. On the other hand, the significant increase of activity observed by calcination and rehydration of LDHs [72-75,87] has been attributed to the formation of a more irregular but still layered structure, with an enhanced activity of interlayer OH" close to disordered edges, obtained via rehydration, in comparison to interlayer OH" in a regular LDH structure [90]. One very recent study looks at the factors affecting the aldol condensation of model aromatic aldehydes with acetone using calcined and rehydrated Mg/Al LDHs, in which best conversion values for Fbenzaldehydes were observed, with a significant improvement of the productivity per kg of catalyst achieved by repeatedly adding new amounts of aldehyde to the reaction mixture, when the conversion was 80% ca., developing a semi-continuous process [93]. The acid-base properties depend significantly on composition, for example Zn/Al, Zn/Cr or Ni/Al LDHs are less basic than Mg/Al LDH [3,5], while LDHs pillared with polyoxometalate (POM) anions may be strongly acidic [94]. Acid-base catalysts
Catalytic Properties of Hydrotalcite-Type Anionic Clays
503
obtained by calcination of B-containing LDHs have been recently investigated for the vapor phase Beckmann rearrangement of cyclohexanone oxime to caprolactam [95-99], key step in the synthesis of Nylon 6, since basic sites on the cationic sheets promote desorption of lactam, consequently reducing the formation of tars, without impairing the acid function [100].
A) Knoevenagel condensation B) Michael addition C) Decarboxylation
D) Claisen condensation and cyclization E) Aldolic condensation F) Oxidation of benzaldehyde
Figure 4 - Reaction pathways between benzaldehyde and ethyl acetonate [91] The role of the processing conditions and the influence of different methods of acidity regulation were investigated in some depth [97,98]. However, at low B-contents, the performances were poorer than expected, because of the loss of acid properties due to grafting reaction during thermal treatments as well as to the low surface availability of boron anions [99]. Another interesting application is the vapor phase alkylation of phenol or mcresol to produce gasoline additives or chemical intermediates using calcined Mg/Al, Mg/Cr or Mg/Fe LDHs, with a direct correlation between composition and acid-base properties of the catalyst and the activity and selectivity between O- and C-alkylated products [101-103]. Recently, the vapor and liquid-phase methylation of m-cresol (Fig. 5) has been investigated in depth by Cavani and coworkers [41,104-106], which showed the role of the preparation method and catalyst composition on the catalytic performances. The basicity significantly affected the activity and the oxygen/carbon
504
S. Albertazzi, F. Basile and A. Vaccari
(O/C) alkylation ratio, while medium strength basic sites were identified as the active sites for the methylation of m-cresol, although the effect of the basicity was more relevant under gas-phase reaction conditions than in the liquid phase. In this field, we may also mention the selective methylation in the 523-623 K range of catechol to guaicol and veratrole, using dimethyl carbonate as reagent and Mg/Al LDHs calcined at 723 K for 6 h as catalyst [107].
Figure 5 - Reaction pathway for the methylation of m-cresol [41,106]. Recent years have seen an impressive increase in the number of catalytic applications of calcined LDHs in the synthesis of intermediates, such as the synthesis of glycol ethers by reaction of olefin oxides and alcohols [108-113], the benzylation and benzoylation of substituted benzenes over solid Ga and Mg-oxides and/or chlorides derived from Ga/Mg LDH by HC1 pre-treatment or calcination [114], the synthesis of methylamines on Cu/Mg/Al and Mg/Fe/Al calcined at 723 K for 5h [115], the hydrolysis of benzonitrile substituted derivatives on Mg/Al or Mg/Ga LDHs calcined at 773 or 873 K, in which the active sites were identified as Bransted basic sites of moderate strength [116] or the production of monoglycerides (food emulsifiers) by transesterification between glycerol and triglycerides, using MgO or calcined hydrotalcites with a low Al-content [117]. With regard to fine chemicals, the synthesis of 5-phenyl-3-methyl-2pentenenitrile (citronitril) is of interest in the perfume, detergent and soap industries (Fig. 6) [118]. In the first step, maximum activity was observed for a Mg/Al (3:1) hydrotalcite calcined at 723 K for 18 h, while the second step is a more demanding reaction from the catalyst point of view, with important roles played by the amount of water present and reaction temperature. When they were optimized, high conversion and the selectivity close to 50% was obtained in a one-pot reaction. On the other hand, the one-pot synthesis of 4-methyl-2-pentanone (methyl i-buthyl ketone or MIBK) has been proposed by gas phase reaction of acetone with H2 and using Pd/Cu catalysts prepared from LDHs, as a possible alternative to the homogeneous conventional threestep process, which generates huge amounts of waste [119]. It was found that the multifunctional transformation of acetone to MIBK was rate determined by the basic function.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
505
Another example, is the Claisen-Schmidt synthesis of chalcones and flavonoids (products of interest to the pharmaceutical industry) using calcined LDHs [120,121].
Figure 6 - Reaction pathway for the one-pot synthesis of 5-phenyl-3-methyl-2pentenenitrile (citronitril) [118] Table 4 - Meerwein-Ponndorf-Verley reduction of unsaturated aldehydes to the corresponding unsaturated alcohols using a Mg/Al (3:1) LDH heated in N2 at 723 [123]. Reacion time (h) Conversion (%) Selectivity (%) Aldehyde Citronellal 4 90 95 a 5 75 92 b Cinnamaldehyde 5 83 70 c Citral Catalyst = 0.140 g; reaction temperature = 355 K; aldehyde = 1 mmol; 2-propanol = 10 ml; stirring speed = 900 rpm. (a) Citronellol; (b) Cinnamyl alcohol; (c) Nerol + geraniol The reaction rate went through a maximum according to the composition, crystal size and calcination temperature or the water content in the solvent (although an excessively high amount of water in the reaction medium had an inhibiting effect, probably due to the competition of water with the reactants), suggesting a key role of strong basic sites: oxygens with a low coordination number and/or surface hydroxyl groups. An optimized catalyst (Mg/Al = 3:1, calcined at 723 K for 18 h) was investigated in the synthesis of 2',4',4-trimethoxychalcone (vesidryl), product of pharmacological interest owing to its diuretic and choleretic properties, obtaining yields of 55% and 85 %, after 20 h of reaction at 423 and 443 K, respectively [120]. Finally, noteworthy is the application of mixed oxides obtained from LDHs containing at least one element between Fe, Al, V, Cr and/or Ga, for the selective permethylation, under near- or super-critical conditions, of P-, y-, 5-tocopherol or their mixtures into octocopherol, the most valuable product having the highest vitamin E activity [122]. Although without industrial applications, catalysts obtained from LDHs were widely investigated in many organic reactions: for example Mg/Al LDHs heated in N2 up to 823 K are highly active, selective and regenerable catalysts for the liquid phase Meerwein-Ponndorf-Verley reduction of carbonyl compounds using 2- propanol as the hydrogen donor [123]. A Mg/Al (3.0 at. ratio) mixed oxide obtained by calcination at 723 K showed better activity in comparison to alumina, silica-alumina or Y-zeolite, attributed to a synergetic effect between strong Lewis basicity and mild acidity (Table
506
S. Albertazzi, F. Basile and A. Vaccari
4). On the other hand, after thermal decarbonation and rehydration the same LDH has been reported for the cyanoethylation of alcohols [124], with a high activity and air stability, unlike that occurring for other solid base catalysts. A similar calcined/ rehydrated Mg/Al (2.5 at. ratio) LDH has been reported in the selective Michael addition on methyl vinyl ketone, methyl acrylate, and simple and substituted chalcones by donors such as nitroalkane, malononitrile, diethyl malonate, cyanoacetamide and thiols, with quantitative yields under mild reaction conditions [125], Finally, as further examples, we may mention the Friedel-Crafts alkylation of benzene and other aromatic compounds using a Mg/Ga (2.0 at. ratio) calcined at 1073 K for 4h, with high benzylation activity even in the presence of moisture in the reaction mixture [126] or the catalytic hydroxylation of phenol over CuM(II)M(III) LDH, where M(II) = Ni or Co and M(III) = Al, Cr or Fe, prepared with the low supersaturation method, with M(II)/M(III) = 3.0 (at. ratio) and Cu/M(II) = 5.0 (at. [127]. Ueno et al [128] have reported the use of Mg/Al LDHs as efficient base catalysts for the epoxidation at low temperatures of olefins using H2O2 and benzonitrile and with MeOH as the solvent (Fig. 7a). Analogously, Mg/M1" (M m = Al, Ga or In) mixed oxides, obtained by heating in N2 at 773 K of the corresponding LDHs, have been reported in the epoxidation of limonene to limonene oxide in the presence of acetonitrile, butyronitrile or benzonitrile, observing maximum activity for the latter nitrile and the Mg/Al LDH containing interlayer carbonates [129,130]. Recently, the epoxidation of electron-deficient alkenes [131] and chiral electron-deficient alkenes [132] have been reported using only H2O2 and as catalysts Mg/Al mixed oxides obtained by calcination at 723 or 773 K. The highest activity was observed for an atomic ratio Mg/Al = 3.0 with the catalyst being recoverable and reusable at least twice without loss of activity. Kaneda et al [133-135] have applied LDHs in the heterogeneous Baeyer-Villiger oxidation of ketones using molecular O2 and benzaldehyde. Multicomponent LDHs containing a small amount of Fe, Ni or Cu led to higher yield values, due to cooperative action between the basic and the transition metal sites (Fig. 7b). Interesting reviews on the applications of HT anionic clays and their derivatives in the oxidation of organic compounds have recently been published by Mayoral and coworkers [136] and Choudary et al [137] with special emphasis on the ecofriendly options for the production of specialty and fine chemicals.
Figure 7 - Activity of some LDHs in (a) the epoxidation of olefins using H2O2 [128] and (b) the Bayer-Villinger oxidation of ketones [133-135].
Catalytic Properties of Hydrotalcite-Type Anionic Clays
507
In an innovative paper (although, to our knowledge, these results have not been confirmed by other authors), Indian researchers have reported that Mg/Al mixed oxides catalyze the diasteroselective synthesis from aldehydes and nitroalkanes of nitroalkanols, which can be easily hydrogenated with retention of configuration forming pharmacologically important derivatives (Table 5) [138]. Table 5 - Heterogeneous diasteroselective synthesis of nitroalcohol derivatives (Henry reaction) catalyzed by a Mg/Al = 3.0 (at. ratio) LDH calcined at 723 K[138]. Substrate
Nitroalkane
React. time (h) 6
Product
Yield a threo/ erytro 3.25:1 87
l-Phenyl-2nitropropan-1-ol l-(3-Nitrophenyl)-2- 95 3 -Nitrobenzaldehy de Nitroethane 12.5:1 6 nitropropan-1-ol l-(41.23:1 Nitroethane 8 4-Methoxy62 Methoxyphenyl)-2benzaldehyde nitropropan-1-ol l-(2-Chlorophenyl)- 89 1.53:1 6 2-Chloro-benzaldehyde 1-Nitropropane 2-nitrobutan-1 -ol 6 Nitroethane 1 -(4-Nitrophenyl)-2- 84 100:0 4-Nitro-benzaldehyde nitropropan-1-ol 6 100:0 l-(2-Chlorophenyl)- 82 2-Chloro-benzaldehyde Nitroethane 2-nitropropan-l-ol 6 1.5:1 74 Furan-2-carboxaldehyde Nitroethane l-(2-Furyl)-2nitropropan-1-ol 8 88 100:0 2-Chloroquinoline-3 - Nitroethane l-(2-Chloro-3carboxaldehyde quinoline) -2-nitropropan-1 -ol 6 Cynnamaldehyde Nitroethane l-Styryl-241 1.25:1 nitropropan-1-ol (a) Isolated after column chromatography. (b) Average ratios calculated from l3C-NMR signals (50.3 MHz). Maximum activity was observed for a Mg/Al (3:1) mixed oxide, which could be reused after new activation at 723 K. A wide range of aldehydes and nitroalkanes were investigated, without any effect of the catalyst/substrate ratio and, very relevant, with formation predominantly of the threo isomers, i.e. the most energetically favoured on the basis of a transition state model based on chair-like structures involving coordination between the two oxygen atoms and the Mg atoms. At the end of this section, we should like to mention the applications of pillared or intercalated LDHs (the two terms refer to the accessibility or not of the interlayer space) in the organic syntheses that up to now, however, have only been investigated on a laboratory scale. These materials have been claimed: i) to obtain shape-selective catalysts; ii) to stabilize homogeneous or biomimetic catalysts to increase their service life and allow easy recovery and recycling, and iii) to prepare supported catalysts with concentration of the active phase, stability and activity higher than those obtained with conventional support. For more detailed information on the specific properties of pillared or intercalated LDHs and their different preparation methods, references should Benzaldehyde
Nitroethane
508
S. Albertazzi, F. Basile and A. Vaccari
be made to previous reviews [3,5,7,8,17,18]. Further studies are required to determine whether these materials have practical applications and whether they can be used as such or only in the form of oxides obtained by thermal decomposition. Referring to the interlayer accessibility, only a small fraction of the claimed pillared LDHs can be considered correctly as such: the main examples are those containing polyoxometalate (POM) anions, which were first described in the patent literature by Wolterman [139]. Other compounds containing low-charge anions, metal complexes, etc. should be more properly defined as intercalated, to emphasize that the interlayer space is not accessible, and reaction occurs on the external edge-sites of the crystallites [140]. There is little information in the literature on the thermal evolution of pillared LDHs and the few data reported are not always in agreement, probably because of the different techniques employed or insufficient characterization [3,5]. However, pillared POMs show decomposition temperatures significantly lower than those of the corresponding salts, which may be attributed to the reaction with the mixed oxides obtained by dehydroxylation of the brucite-type sheets of LDHs [141,142]. This low thermal stability (< 573 K ca.) has been explained on the basis of the higher charge density of the LDH framework (4e"nm"2) [5]. On the other hand, it is evident that the thermal stability of anions such as metal complexes, phtalocyaninetetrasulphonate, etc. is not significantly modified by intercalation in the HT structure and such materials may be used only at T < 373 K ca. On the basis of the above considerations, pillared or intercalated HT anionic clays have only low temperature applications [3,5,143,144], while for other applications claimed at higher temperatures, serious doubt exists regarding their stability and they must be more correctly considered as an alternative way to prepare homogeneous mixed oxides [3,5,139,145,146-150]. LDHs pillared with iso or heteropolyanions showed significant photocatalytic activity in the oxidation of i-propanol to acetone in the presence of O2, despite scattering of the host particles [141,151]. Such results indicate that it may be possible to realize shape selective photochemical processes by controlling the distance between the pillars, i.e. the pore size, such as in zeolites. The ability of HT compounds to incorporate anionic molecular species while preserving their photochemical excitedstate lifetimes has been reported for Ru (4,7-diphenyl-l,10phenanthrolinedisulphonate)4" ions, and the use of layered host systems for the intercrystal immobilization of anionic photocatalysts was proposed [152]. There is little information in the literature on the thermal evolution of pillared LDHs and the few data reported are not always in agreement, probably because of the different techniques employed or insufficient characterization [3,5]. However, pillared POMs shows decomposition temperatures significantly lower than those of the corresponding salts, that may be attributed to the reaction with the mixed oxides obtained by dehydroxylation of the brucite-type sheets of LDHs [141-143]. This low thermal stability (< 573 K ca.) (Fig. 8) has been explained on the basis of the higher charge density of the LDH framework (4e"nm"2) [5]. On the other end, it is evident that the thermal stability of organic anions, metal complexes, phtalocyaninetetrasulphonate, porphyryntetrasulphonate, etc. is not significantly modified by intercalation in the LDH structure and such materials may be used only at T < 373 K ca. On the basis of the above considerations, pillared or intercalated LDHs have only low temperature applications [3,5,144,145], while for those claimed at higher temperatures serious doubt exists regarding their stability; they have to be considered as an alternative way to prepare homogeneous mixed oxides [3,5,56,139,141,146-150].
Catalytic Properties of Hydrotalcite-Type Anionic Clays
509
Figure 8 -Evolution with incresing temperature of an Li/Al-V2O7 pillared LDH [141] LDHs pillared with iso or heteropolyanions showed significant photocatalytic activity in the oxidation of i-propanol to acetone in the presence of O2, despite scattering of the host particles [142,151], Such results indicate that it may be possible to realize shape selective photochemical processes by controlling the distance between the pillars. The ability of LDHs to incorporate anionic molecular species while preserving their photochemical excited-state lifetimes has been reported for Ru(4,7-diphenyl-l,10phenanthrolinedisulphonate)4" ions, and the use of layered host systems for the intercrystal immobilization of anionic photocatalysts was proposed [152]. In the shape selective epoxidation of alkene with H2O2, different selectivities have been observed for 2-hexene or cyclohexene and attributed to different accessibility obtained intercalating POMs with different sizes [153]: it was concluded that (i) the basic nature of the brucite-like sheets favoured high selectivity for epoxides over diols, and (ii) a steric constraint imposed by the interstitial environment in the pillared catalysts improved the cis to trans ratio in the oxidation of 2-hexane in comparison to the homogeneous reaction. However, these conclusions have subsequently been refuted by Pinnavaia et al [144], who showed that the enhanced substrate selectivity cannot be attributed to molecular sieving based on the substrate size or shape, because in many cases the pillared structure was lost under the conditions used to dry the samples prior to catalytic tests (393 K), while structurally stable samples showed limited access to the solvated pillared POMs under reaction conditions. H2O2 has also been investigated in the liquid phase oxidation of acetaldehyde catalyzed by LDHs pillared by Keggin-type heteropolyanions [154], showing a selectivity in acetic acid of nearly 100% and a catalytic activity per m.mol of heteropolyanion from 3 to 9 times higher than that obtained using the potassium salts of the POM. The authors claimed stabilization of the pillared POMs against H2O2 and specific effects on the catalytic activity related to the brucite-like sheet and heteropolyanion composition. Surprisingly, however, dried (< 443 K) and calcined samples (> 723 K) showed the same activity, which decreased dramatically for intermediate temperatures. The authors attributed the two ranges of high activity to POM-LDHs or complex spinel-type oxides obtained by their calcination, respectively, although the hypotheses reported by Pinnavaia et al should also be considered [144]. 4 - Catalysts for hydrogenation or hydrogenolysis reactions Hydrogenation is one of the most useful, versatile, and environmentally suitable reaction routes for many organic syntheses (fine chemicals, intermediates, petrochemical industry, etc.) and for the treating of petrochemical feeds. There are two
510
S. Albertazzi, F. Basile and A. Vaccari
general families of materials used as catalysts for the hydrogenation of functional groups: i) noble metals (Pd, Pt, Rh and Ru) and ii) transition metals (Fe, Ni, Cu, Cr and Co). The former are more active, thus smaller reactors and less severe reaction conditions may be employed; however, they are more expensive and thus a tight metal control is strictly required to avoid process losses and an almost complete metal recovery. Since hydrogenation reactions present deactivation problems such as poisoning, coking and sintering, the choice of a suitable support is fundamental in the catalyst preparation. Nowadays, mixed oxides obtained from LDHs seem to be the most promising option for preparing active and selective hydrogenation catalysts, due to their texture properties, easy preparation and simple and safe handling. The hydrogenation of nitriles is a reaction of high industrial interest, since it is the main pathway for the production of primary amines, which are then converted into polymer and agrochemical intermediates (for example, the hydrogenation of adiponitrile to 1.6-hexanediamine to manufacture Nylon-6,6). The industrial processes are usually carried out in liquid phase at elevated H2 pressures and temperatures, using transition metal catalysts. Due to the high reactivity of imines, which are the partial hydrogenated intermediates, a conventional process gives a mixture of primary, secondary and tertiary amines (Fig. 9). The selectivity towards one type of amine is strongly determined by the catalyst: for example, Co-Raney, Ni-Raney and Ru are selective to primary amines, Cu and Rh are preferentially used in the preparation of secondary amines and, finally, Pt and Pd are highly selective towards tertiary amines.
Figure 9 - Reaction pathway for the hydrogenation ofnitriles to amines [157,158]. Multicomponent mixed oxides obtained from LDHs have been reported in the vapour or liquid phase hydrogenation of nitriles under mild reaction conditions, highlighting the requirements for a compromise between metal reducibility (that affects the activity) and the basicity of the catalyst (that affects the selectivity) [155-161]. Mg/Fe mixed oxides had high activity and selectivity in the reduction of aromatic nitrocompounds in mild conditions; furthermore they were cheap, regenerable, operated at atmospheric pressure and can be recycled without any loss of activity [155,156]. Ni/Mg/Al LDHs also constituted good precursors of catalysts for the selective hydrogenation of nitriles to primary amines, lowering the undesired consecutive transamination reaction, which occurred between imine- and amine-like species on both metal and acid sites, through a bifunctional mechanism [157]. Ni/Mg/Al mixed oxides
Catalytic Properties of Hydrotalcite-Type Anionic Clays
511
showed good catalytic performances in the acetonitrile hydrogenation: the activity as well as the formation of condensed by-products being controlled by finely tuning the Ni/Mg atomic ratio [158,159]. The introduction of Co in the catalysts [160] diluted the Ni-metal phase in small ensembles, less prone to accommodate neighbouring multibonded species, thus optimizing the hydrogenation activity and leading to higher performances than Ni/Mg/Al catalysts. Finally, Ni/Mg/Al mixed oxides having different atomic ratios were recently investigated in the liquid-phase hydrogenation of adiponitrile with high yield values in aminocapronitrile, a key intermediate in a new synthesis of caprolactam and, consequently, in the production of Nylon-6 [161]. The selective hydrogenation of acetylene to ethylene has been chosen as the model reaction for the hydrogenation of unsaturated compounds, also allowing the specific effects of the catalyst composition and properties [162-164] to be highlighted. Active Ni-containing catalysts were prepared by controlled calcination of LDHs; furthermore, the addition to the Ni/Al systems of Cr or Zn improved the catalytic performance. It was found that the addition of ZnO drastically reduced coke formation, and this decrease can be further enhanced by the addition of Cr. A Ni/Zn = 4.0 atomic ratio resulted optimum value, yielding high activity and selectivity values, with very low coke formation. While the addition of ZnO improved the catalytic performances, the increased Ni content decreased conversion and selectivity in ethylene, whereas it increased coke formation. For more detailed information on the hydrogenation of unsaturated hydrocarbons, a recent review by Monzon et al should be referred to [164]. Pd nanoparticles incorporated in a LDH host were studied in the mild and selective alkine partial hydrogenation [165]. The Pd nanoparticles were incorporated by anion exchange of a diluted suspension of LDH containing nitrates and a Pd-hydrosol stabilized by sodium dodecyl sulphate. Characterization by ICP-AES, XRD, and TEM analyses showed the deposition of almost monodispersed Pd particles, mainly on the external surface of the LDH sheets. These samples proved to be efficient catalysts for the liquid-phase partial hydrogenation under mild conditions of both terminal and internal alkynes. In the hydrogenation of phenylacetylene to styrene a selectivity of 100% was obtained, and the cis stereoselectivity for the hydrogenation of internal alkynes (4-octyne or 1-phenyl-l-pentyne) approached 100%. The activity increased with the Pd dispersion, while the selectivity in alkene was essentially unaffected. Calcined Mg/Al LDHs have been reported as efficient catalysts in the vapourphase reduction of a wide number of ketones to the corresponding alcohols: the presence of strong basic sites favoured the reduction of the ketone to the corresponding alcohol, whereas medium-strong acidic sites favoured the dehydration of the alcohol formed to alkene. Jyothi et al [166,167] obtained excellent yields in the conversion of aryl alkyl ketones to aryl alkenes. The catalytic activity of these calcined LDHs was attributed to the co-existence of acid-base pair sites with strong basic sites, which can promote hydrogenation transfer. The reduction of unsaturated compounds using an organic molecule instead of H2 or a metal hydride is known as "hydrogen transfer" and is highly selective and especially suitable for reducing ct,[3-unsaturated carbonyl compounds. In addition to the example already given [123], Aramendia et al [168-170] reported the catalytic transfer hydrogenenation of citral (Fig. 10) on basic mixed oxides. All the catalysts studied were found to provide excellent activity and selectivity in the hydrogenation process (conversion always exceeded 95% and selectivity 90% within the first 40 h of reaction). The catalytic activity was found to depend on the surface basicity and the catalysts were reused up to three times without any substantial loss of
512
S. Albertazzi, F. Basile and A. Vaccari
conversion or selectivity.
Figure 10 - Reaction pathway for the catalytic transfer hydrogenation of citral [168170] Vapour phase hydrogenation of maleic anhydride (MA) is the most direct way to produce y-butyrolactone (GBL), currently one of the most useful intermediates in the production of N-methyl pyrrolidone (by reaction with methyl amine), for use as an alternative to chlorine-based solvents, the use of which will be subjected to increasing restrictions due to their negative impact. Furthermore, the same industrial plant can produce according to reaction conditions other small volume commodities of high added value, such as (tetrahydrofuran (THF) and, mainly, 1,4 butanediol (BDO) (Fig. 11). Cu/Zn/Al obtained by calcination of LDHs have been reported as interesting alternatives to the currently employed reduced Cu chromite catalysts [171,172], which are becoming increasingly difficult to use because of the toxic nature of the spent catalysts, due to Cr(VI) and Ba(II) content. Cu/Zn/Al mixed oxides showed better performances than the Cr-containing catalysts, with highest yields in GBL, also favouring the formation of THF and reducing the amounts of low-cost by-products (Fig. 12). Furthermore, the relationships between surface area and porosity and the catalytic properties for the Cu/Zn/Al catalysts have been investigated, highlighting a decrease in the surface irreversible adsorption phenomena and improving the mechanical strength by pressing the catalyst powder without binders [172], with a behaviour similar to that claimed for other Cu/Zn/Al catalysts in the hydrogenation of coco methyl esters [173].
Figure 11 - Reaction pathway for the selective hydrogenation of maleic anhydride. At the present state-of-the-art, the selective vapour phase hydrogenation of phenol is the most suitable way to obtain cyclohexanone, a key intermediate for the production of caprolactam and, consequently, of Nylon-6 and polyamide resins. In
Catalytic Properties of Hydrotalcite-Type Anionic Clays
513
industrial plants, the selective hydrogenation is carried out over Pd supported on A12O3; however, this catalyst has a low resistance to deactivation by coke deposition. Mg/Al mixed oxides derived from LDHs have been reported as useful supports [65,66]. It was reported that the role of support is to adsorb phenol molecules near to the Pd particles; thus Pd adsorbed H2 molecules and supplied hydrogen atoms to the aromatic ring by a spillover mechanism. Depending on the acid-base properties of the support, the mode of adsorption of phenol varied, affecting selectivity to cyclohexanone or cyclohexanol. The catalytic behaviour of Pd supported on both uncalcined Mg/Al LDHs or the corresponding mixed oxides obtained by controlled calcination have been investigated. Pd supported on LDHs as such [174,175] showed very high conversion values (>95%) associated to high selectivities (>85%) in cyclohexanone. On the other hand, the basic sites of the calcined LDHs were very effective in affecting the adsorption of phenol, which was preferentially hydrogenated to cyclohexanone [176,177]. With a 0.3 wt.% of Pd content the selectivity higher than 95% was reached in cyclohexanone, with a conversion of 40%.
Figure 12 - Catalytic activity as a function of the catalyst composition (as atomic ratio) in the selective vapour phase hydrogenation of MA [172]. The removal of aromatic hydrocarbons and the increased cetane number in diesel fuels have received increasing attention in recent years, due to the increasing environmental restrictions governing the composition of diesel fuels both in Europe and the United States [178,179]. In particular, high aromatic content and low cetane numbers worsen the quality of diesel fuels and increase the particulate emissions. Thus, the development of catalysts for aromatic saturation and hydrogenolysis/ring-opening, characterised by good thio-tolerance, since small amounts of S-containing compounds are always present, is strongly wished for. Noble metals (Pt and/or Pd) supported on large-pore zeolites have been receiving more attention as aromatic hydrogenation catalysts [180,181], although the acidity of the support can increase the likelihood of undesirable cracking reactions, which increase the rate of coke deposition and yield in low-value gases. Pd/Pt supported on basic Mg/Al mixed oxides obtained by calcination at 773 K for 8 h of a commercial LDH (Sasol, D) have been recently proposed as an interesting alternative for the hydrotreating of oil fractions, giving high hydrogenation activities, associated to very good thio-tolerances, while the hydrogenolysis activity to
514
S. Albertazzi, F. Basile and A. Vaccari
high-molecular weight products (able to increase the cetane number) increased with the Pt content [182]. The increasing severity of the quality standards of drinking water and the toxicity of nitrates have generated the urgent need to develop new technologies for the removal of nitrates from water. Nowadays, from an environmental point of view, the most adequate way is to convert the nitrates into N2 by liquid-phase hydrogenation on noble metal catalysts. Palomares et al [183,184] showed that LDH-supported Pd/Cu catalysts exhibited very interesting activity for the liquid-phase hydrogenation of nitrates. These samples were more active in the removal of nitrates and produced lower amounts of ammonium ions than analogous catalysts Pd/Cu supported on alumina. These results can be related with the characteristics of the LDH, which after mildtemperature calcination, may regenerate its structure by contact with aqueous anion solutions. In this structure, the nitrates were located in the interlayer space, compensating the partial positive charge and reducing mass transfer problems. The activity and selectivity of the catalysts increased, mainly if the LDH was synthesised with Cu2+ ions inside the brucite-type sheets, resulting in a higher dispersion of the Cucontaining active sites. The selective catalytic hydrogenolysis of chlorine from chlorinated organic compounds is of increasing interest, since they are known to be harmful for the environment (depletion of the ozone layer and green house effect). Pd-containing LDH has been reported as precursors for the catalytic vapour-phase hydroconversion of chlorofluorocarbons (CFCs) and hydrochlorofluorocarbons (HCFCs), in particular CC12F2 (CFC-12) and CHC1F2 (HCFC-22), observing the highest conversion and selectivity for the heavily loaded Pd catalyst (4 as atomic ratio) [185]. On the other hand, different Pd supported on Mg/Cr mixed oxide catalysts have also been reported for the hydrogenolysis of CC12F2 [186], showing that 6 wt.% of Pd was optimum loading. The hydrogenolysis activity for CC12F2 followed the order: Pd on Mg/Cr mixed oxide > Pd on MgO > Pd on Cr2O3.; however, Pd on Mg/Cr mixed oxide yielded deep hydrogenation product (CH4) with higher selectivity, Pd on MgO yielded dechlorination product (CH2F2) and Pd/Cr2O3 showed a poor activity. To conclude, it was hypothesized that calcined Mg/Cr LDH had synergetic effects when used as a support for Pd and employed in the hydrogenolysis of CC12F2. Finally, Ni-based catalysts prepared from several Ni/Mg/Al LDHs with different atomic ratios have been tested in the vapour-phase hydrodechlorination of 1,2,4-trichlorobenzene [187,188]. All catalysts were active at the reaction temperatures tested, with an increase of the selectivity in benzene, by increasing the conversion value. The highest TOF and selectivity in benzene values were achieved at 523 K using the catalyst with the highest Mg content, which did not show any significant change in activity even at reaction times higher than 500 min. The catalytic behaviour was explained considering that MgO not only modified the electronic properties of the Ni particle causing the H2 desorption at lower temperatures, but also adsorbed the HC1 produced during the hydrodechlorination reaction, avoiding the deactivation process. 5 - Catalyst for environmental applications LDHs may be used as anion exchangers on account of the accessibility of the interlayer region, which depends on the nature of the anion present [2,3,5]. They display an exchange capacity (2-3 meq g"1) [2,189] similar to that of anion exchange resins, but are characterized by a higher resistance to temperature; LDHs were therefore utilized as
Catalytic Properties of Hydrotalcite-Type Anionic Clays
515
anion exchangers in some high temperature applications, such as in the treatment of the cooling water of nuclear reactors [190]. However, poor results have been reported on account of low selectivity coefficients and gradual desorption due to the slow substitution by CO2 from the atmosphere [191]; thus the use of calcined LDHs, operating by structure-reconstruction, has also been proposed [192]. LDHs, as such or after mild-temperature calcination, has been reported also as sorbent for many organic contaminants in water [20], for example nitro or chloro phenols [193,194], dodecyl benzylsulphonate [194], dyes [195], etc. The decontamination of soils or wastewaters from transition-metal ions may be achieved by their high and selective uptake by nitrate- or carbonate-containing Mg/Al LDHs [196]. It has been hypothesized that this selective cation uptake occurred by substitution in the LDH structure for Mg2+ ions through a process known as diadochy. Finally, it should be noted that the synthesis of LDHs can be directly applied to the treatment of waste waters [197,198]. About the catalytic applications to water treatment, in addition to the examples on the catalytic removal of nitrates reported in the previous Section 4 [182,183] interesting results have been obtained using Mg/Al LDHs intercalated with phatalocyanines in the catalytic oxidation of 1-decanethiol or 2,6-di-r-buthylphenol, probe molecules to check their potential use as immobilized catalysts for remediation of contaminated ground water and industrial effluents at ambient conditions [140,144,199201]. For both probe molecules, a significant increase in activity and stability was observed in comparison to the homogeneous catalysts, together with the possibility of recovering the catalyst by simple filtration (Fig. 13).
Figure 13 - O2 uptake plots for the autoxidation of 1-decanthiol in the presence of Co(II)-phthalocyaninetetrasulphonates: (O) homogeneous catalyst (deactivated after one reaction cycle), (v) catalyst intercalated in Mg/Al LDH (no loss of activity after five consecutive reaction cycles) [200]. The increase in stability was attributed to the inhibition of the phthalocyanine dimerization through intercalation in the LDH structure. Co(II)-phthalocyanine-LDH catalysts, prepared by direct synthesis, anion exchange or structure-reconstruction, have
516
S. Albertazzi, F. Basile and A. Vaccari
been recently reported for the oxidation of mercaptans in light oil sweetening, observing a higher activity than an industrial non-basic impregnated active charcoal with, furthermore, a good mechanical resistance [202]. The same catalyst has been reported also for the oxidation of 2-mercaptoethanol [203]. The elimination of thiols has also been investigated by oxidation with O2 at ambient conditions using intercalated Mo(VI) complexes [204,205]. The authors claimed that water solutions could be used, although the results referred only to tests carried out with ethanol as the solvent. Cu/Ni/Al LDHs with different atomic ratios and calcined at different temperatures between 373 and 1173 K have been investigated in the wet air oxidation of phenol aqueous solutions, using a trickle-bed reactor and air with an O2 partial pressure of 0.9 MPa at a reaction temperature of 413 K [206]. Non calcined LDHs were practically inactive, while after calcination at 623 and 673 K they showed high initial activities, which decreased continuously over the reaction time, because of the elution of the active phase. On the other hand, the spinel phases obtained by calination at 1023 K showed high conversion (among 40-75 %), without any loss of activity after a continuous working run of 15 days. Finally, we may mention the complete photocatalytic degradation of traces of an aqueous organochlorine pesticide, hexachlorinecyclohexane (HCH), by irradiation in the near UV area and using a suspension of Mg12Ai6(OH)36(W7O24).4H2O [207] as the catalyst. The model and mechanism (Fig. 14) for the photocatalytic degradation of HCH on the pillared LDH indicated that the interlayer space was the reaction field and that photodegradation of OH* radicals were responsible for the degradation pathway. For further information on LDH applications to water decontamination, reference should be made to a recent review by Ulibarri and Hermosin [208].
Figure 14 - Reaction pathway of the photocatalytic degradation of aqueous hexachlocyclohexane by Mg/Al LDH pillared with paratungstate (W7O2/~) ions [207]. Moving on now to air decontamination, it must be noted that calcined LDHs have been applied as catalysts in many reactions of high environmental relevance, such as SOx and NOx removal, N2O decomposition and, recently, the combustion of organic vapours. In the former case, they were investigated to reduce the emissions from FCCU, which represent the most relevant SOx source in petroleum refineries and for which, severe restrictions have been imposed by different countries [48-50,209-211]. These catalysts have to be capable of oxidizing the SO2 to SO3 (also using CeO2 as co-
Catalytic Properties of Hydrotalcite-Type Anionic Clays
517
catalyst) and to fix it as sulphate in the regenerator, while in the reducing atmosphere of the cracking zone the sulphate is decomposed to H2S, which can be recovered by available techniques (Fig. 15). As reported in Section 2 [48-50], compounds of intermediate basicity, such as Mg/Al mixed oxides obtained by calcination of LDHs have been proposed as supports to achieve better compromises between SOx uptake and catalyst regeneration. The insufficient regeneration of the earlier Ce/MgAl2O4MgO catalysts was overcome by the partial substitution of Al3+ ions by Fe3+, V3+ or Cr3+ ions, thus obtaining active and stable catalysts, which, furthermore, allowed the simultaneous control of SOx and NOx emissions from the FCC regenerating units [209].
Figure 15 - Pathway for the SOx removal reaction [48]. The above catalysts were Al-rich, thus with limited amounts of SOx captured, especially in high-temperature regenerators. Mg-Al mixed oxides with higher Mgcontents were subsequently investigated [210], observing after impregnation with 5 wt% of CeO2 very good SOx adsorption associated with a limited regeneration capacity. To overcome this constraint, the catalysts were further impregnated with 1 wt% (as metal oxide) of Co, Cu, Zn, Ni, Fe, Cr, V or Ti, obtaining better regenerability for V, Fe and Cu. Due to the constraints in the use of V and Fe in FCCU, attention was focused on Cu/Mg/Al mixed oxides obtained from HT precursors (again impregnated with 5wt% of CeO2) observing the best compromise between adsorption and regeneration capacity for 5 wt% CuO. It is noteworthy that Zn/Mg/Al mixed oxides obtained by mild-temperature calcination of LDHs have also been claimed as catalyst additives for the sulphur reduction in FCC naphtha [211]. Operating with an O2 excess < 0.6 %, Cu/Mg/Al LDHs calcined for 3 h at 1023 K and activated under H2 for 30 min at 823 K., simultaneously removed both SOx (by an oxidation and/or reducing reaction) and NOx (by a reducing and/or decomposition reaction) [212]. The active centres have been identified as Cu° or Cu+1 species and it was observed that the catalytic activity was not affected by the presence of coke and the formation of "copper-sulphured species". Similar results have been obtained with an analogous Co/Mg/Al catalyst, which was able to reduce NOx even in the presence of excess O2, although with a lower activity in SOx removal than the Cu-containing
518
S. Albertazzi, F. Basile and A. Vaccari
catalyst. Again, reduced cobalt species have been proposed as the active sites, as was highlighted by the absence of activity in the presence of a high O2 excess or if the catalyst was not previously reduced [213]. Fe/Mg/Al, Cr/Mg/Al and Cu/Mg/Al catalysts obtained from LDHs have been investigated in the selective catalytic reduction (SCR) of NO [214]. A general improvement in the catalytic performances in comparison with the analogous catalysts prepared by incipient wetness impregnation of a Mg/Al mixed oxide [51] was observed, related to better dispersion of the transition oxides. Considering the best behaviour of Cu-containing catalysts, Cu/Mg/Al HT LDHs were thoroughly investigated as a possible inexpensive alternative to Cu-zeolites for NO reduction by NH3 [215]. While no significant differences were observed in the absence of O2, it was found that the presence of stabilized copper ions improved the catalytic performances in the presence of excess O2 (such as that present in flue gases), with interesting catalytic performances for Cu-contents in the range 12 to 25 wt% ca. (as CuO), associated with high stability with time-on-stream (Fig. 16). The reaction mechanism and the nature of the active sites involved either in SCR of NO or ammonia oxidation (main side-reaction at high temperature), have been investigated using FT-IR spectroscopy, showing that the SCR reaction occurs between gas-phase NO and NH3 strongly adsorbed on Cu-containing phases and/or highly dispersed CuO clusters [216]. Moreover, the SCR activity of these catalysts did not involve BrOnsted acid sites, absent on the surface of these samples, thus indicating that Bronsted acidity is not always a key requirement for SCR activity.
Figure 16 - NO reduction as a function of time-on-stream, at 653 K with NHs and excess O2, for a Cu/Mg/Al mixed oxide (CuO = 12.5 wt.%) obtained from a LDH by calcination for 14 h at 923 K [(*) NO reduction, (v) selectivity in N2J [215]. A thermally decomposed Mg/Fe LDH has been reported to decompose NO at 673 K, probably via N2O, although with low conversion values (< 27%) and experimental conditions very far from those of practical interest [217]. On the other hand, a Cu/Mg/Al LDH, containing 9 wt% ca. of CuO, calcined at 1023 K and activated under H2 at 773 K for 30 min catalysed both NO decomposition and its reduction by propane [218]. In the NO decomposition, the initial conversion (95% ca.) decreased rapidly with time-on-stream, for example to 80% ca. at 973 K and to lower values when the reaction temperature was further decreased. The catalytic activity increased with
Catalytic Properties of Hydrotalcite-Type Anionic Clays
519
increasing temperature up to ca. 1023 K, while higher temperatures yielded lower conversions, due to thermodynamic limitations. On the contrary, in the reduction with propane, total conversion of NO was obtained at T > 873 K, thus indicating that the reaction was not thermodynamically limited. By using EXAFS spectroscopy and combined EXAFS/XRD under the operating conditions, the active species were identified as Cu1 for NO decomposition and Cu° for its reduction. The reduction of NO with propane in the presence of different contents of O2 has also been investigated [219,220], observing complete NO conversion at T > 723 K and for O2-contents lower than the stoichiometric amount required for the complete oxidation of propane. The catalytic behaviour depended significantly on the catalyst composition, with better performances for the Cu/Cr mixed oxides obtained by calcination for 3 h at 873 K [219] in comparison with those obtained with a decavanadate-exchanged Mg/Al (2:1) LDH calcined at 823 K for 3 h [220]. Between the Cu/Cr mixed oxides, the best performances were observed for an intermediate Cu/Cr ratio (2:1), which also withstood higher O2-contents in the feed. Operating at 80% of the stoichiometric O2-content, this catalyst showed 100% conversion of NO and 75% conversion of propane at 723 K, and an almost negligible formation of N2O above 623 K. Some of these authors [221 ] have also recently claimed the use of multicomponent mixed oxides in the total oxidation of N-containing organic vapours, using dimethylformamide (DMF) as a probe molecule, this being widely used as an industrial solvent. The catalytic performances depended on the catalyst composition and calcination temperature: for example, the best results were observed for Cu/Zn/Cr/Al/V or Cu/Cr/V LDHs calcined for 3 h at 673 K or 823 K, respectively. Calcined LDHs have been widely investigated in the decomposition of N2O, a greenhouse pollutant 270 times approximately more powerful than CO2 (on a weight basis), which, furthermore, contributes to catalytic destruction of stratospheric ozone. The continuous increase of N2O in the atmosphere is mainly caused by anthropogenic activities, such as cultivated soil, biomass burning, stationary and mobile combustion, and chemical processes (for example adipic acid synthesis). However, in the short term, only the emissions associated with combustion and chemical processes can be reduced, with a reasonable goal of achieving about 60-90% reduction by 2010 [222]. Kannan and Swamy [223,224] claimed calcined Co/Al, Cu/Al or Ni/Al LDHs as effective catalysts in the decomposition of N2O, although using a recirculating static reactor and a low partial pressure of N2O (6.7 kPa), i.e. with experimental conditions very far from those of practical interest. In fact, using a flow reactor and feeding a 1% v/v N2O/He gas mixture, both Cu/Al and Ni/Al mixed oxides showed very low activity with appreciable conversion only above 773 K [222]. Even worse performances were observed for calcined LDHs containing both elements, while a relatively higher activity was obtained by CuO supported on a Mg/Al mixed oxide prepared by calcining a LDH precursor. A significant increase in activity was observed when the Ni/Al sample was prereduced at 773 K for 30 min in a 20% H2/He flow and then reoxidized with N2O before the catalytic tests. The best performances were obtained using a Mg/Al mixed oxide containing a small amount of Rh, for which the change from 10 to 90% of N2O conversion occurred in a temperature range of 100 K ca. [222]. The role of the composition of the precursors has been investigated as a function of the reaction conditions and feed composition (mainly presence of O2 and/or H2O), with the aim to optimize these materials in a simulated process stream (Table 6) [27,225,226].
520
S. Albertazzi, F. Basile and A. Vaccari
Table 6 - Catalytic activity in N2O decomposition of different mixed oxides obtained by calcination of HT precursors [27,225,226]. N 2 O conversion (%) Atomic ratio
The addition of Pd to the Co/La/Al catalyst led to results comparable to those of the Rh-containing catalyst only for a Pd content 50 times higher than that of Rh. This effect is attributed to an increase in the rate of desorption of O2 from the catalyst surface, the most difficult step in N2O decomposition [231]. Significant increases in stability in SO2 and O2 mixtures were found mainly with the addition of Mg, which plays a key role although it is not itself active in the reaction. For example, a Co/Mg/Rh/Al catalyst did not show any deactivation even after 80 h of reaction, while for the corresponding Co/Rh/Al catalyst, conversion dropped from 80 to 60% during the first 10 h. The presence of Mg leads to an increase in the surface area of the calcined materials, provides stability against SO2 inhibition at low temperature, and eliminates the initial deactivation, but at the same time decreases the specific activity [232]. Some of the above authors [233] proposed a dual-bed catalytic system by which NOx and N2O may be subsequently removed from flue gases simulating leanburn engine conditions. NOx is removed in the first stage by SCR with propene over a Pt supported on activated carbon catalyst. The second bed decomposes the N2O formed in the first bed into N2 and O2 using either Co/Rh/Al and Co/Pd/La/Al mixed oxides obtained from LDHs or ion exchanged Fe-ZSM-5 and Pd-ZSM-5 zeolites. The Co/Rh/Al mixed oxide proved to be the most active and stable catalyst and at 475 K and 700 K in the first and second stage, respectively, molar conversions of 90% and 100% for NOx and N2O were achieved in a stable operation lasting 50 h. Recently, some of these authors [234] have carried out a detailed investigation on the interaction of N2O with a Co/Mg/Rh/Al (3:1:0.02:1) mixed oxide, obtained by calcining the LDH precursor for 18 h at 723 K, using a tapered oscillating microbalance coupled to a mass spectrometer (TEOM-MS) and testing the temperature range 303-598 K. Although the reaction conditions were, again, very far from those of practical interest, a possible model for the N2O interaction with the catalyst surface has been derived: O2 recombination/ desorption and oxidation of the surface sites appeared to be the slowest processes, while the removal of adsorbed oxygen by N2O from the gas phase was much faster.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
521
Recently, an extremely efficient catalytic process has been claimed for converting N2O, using noble-metal containing M/Mg (Co)/Al (M = mainly Rh, with a content between 0.5 and 5.0, as atomic ratio %) LDH-derived catalysts, which were previously reduced under H2 flow for 1-10 h at 773-1073 K [235]. The tests were performed using waste gas from an adipic acid production plant, which contained from 10 to 40 % of N2O and significant percentages of CO2, O2 and H2O, operating at a temperature ranging from 573 to 1073 K and reducing the exit N2O level to below 200ppm. The key parameters in the preparation of these active and stable catalysts have been investigated in depth, focusing the attention on the relationship between catalyst activity and composition and the possibility of developing synergetic effects with other elements (Pd or La) [236]. It was shown that activation by reduction improved both activity and stability significantly, with the complete N2O decomposition (< 50 ppm) reached at different oven temperatures as a function of the Rh-content (Fig. 17). On the contrary, the Rh/Co/Al mixed oxide exhibited poorer performances and quite poor stability; these differences were even clearer by feeding 2.6 wt.% H2O in the gas stream. H2O not only reduced the catalytic activity, but provided information on the role of the basicity of the support, since the catalyst with a higher Mg/Al ratio deactivated faster (Fig. 18) [236].
Figure 17 - Catalytic activity of reduced Rh/Co/Al (and Rh/Mg/Al mixed oxides with and without the presence of 2.6 wt.% H2O in the feed (oven temperature 723 K) [236]. Maximum surface area values were obtained by calcining the precursors below 593 K, with optimum activity and life at 723-773 K, while calcination at 1073 K was too severe. The tests were at first carried out with a low N2O concentration in the feed; the N2O conversion increased with increasing reaction temperature and for the Co/Al mixed oxides a maximum for an atomic ratio of 3.0 was found, although the conversion significantly worsened by co-feeding O2 or, mainly, O2 and steam. Improved activity at low temperatures was reported for a Co/Al/La (4:1:1) catalyst, although in this case the material did not reveal any HT structure by XRD analysis. Using this latter catalyst, Dandl and Emig [227] determined the reaction rate at various partial pressures of N2O, O2 and H2O in a temperature range from 573 K to 823 K, by estimating the energies of activation and rate coefficients for the main step of the reaction. However, the best performances [225,226] were obtained with a Co/Rh/Al catalyst, with the 0.7 wt% of Rh, which was also active at 573 K and did not show any deactivation effect due to O2
522
S. Albertazzi, F. Basile and A. Vaccari
or steam. Large beds of Co/Al catalyst (with an atomic ratio equal to 2.2) calcined at 773 K were shown to perform effectively for long times-on-stream under a simulated wet N2O process stream (10-15% N2O, 2% H2O and/or 2% O2), without any hydrothermally induced deactivation effect. In comparison with the already known CoZSM5 catalyst, the Co/Al catalyst showed a 75 K lower light-off temperature (corresponding to 50% conversion of N2O), i.e. reducing energy requirements and placing less wear and stress upon process equipment [225].
Figure 18 - Deactivation with time-on-stream of reduced Rh/Mg/Al catalysts in N2O decomposition in the presence of 2.6 wt.% H2O in the feed (oven temperature 723 K) [236] LDHs containing different combinations of bivalent (Co,Pd and Mg) and trivalent (Al, La, Rh) cations, previously calcined for 18h at 723 K, have been investigated to obtain less expensive systems with activity approaching those of Rhcontaining catalysts [228]. The formation of pure LDH was claimed also for an Al/La = 1.0 atomic ratio, in contrast with that previously reported [225] and the significantly higher ionic radius of the La3+ ions [229]. Furthermore, the thermal decomposition of the La-containing samples was different from that typical of LDHs and the calcined samples showed the presence of segregated La-containing phases. The tests were carried out with a diluted feed (0.1% N2O in He, with the addition in the deactivation tests of 0.01% ca SO2 and 3% O2 for 1 h), i.e. in conditions far from those of industrial interest. A significant increase in activity was observed in comparison with the Co/Al catalyst due to the presence of Rh (Co/Rh/Al = 3/0.02/1 as atomic ratio) or La (Co/La/Al = 3/1/1). In the case of La, the increase in activity was attributed to the high activity in N2O decomposition of La2O3 [230]. 6 - Catalysts for advanced natural gas exploitation Exploitation of natural gas is an important goal from both the industrial and academic points of view, considering its low price and high availability. Activation of CH4, its main component, is a key factor in promoting the utilization of natural gas, and currently the indirect way [i.e. the preliminary formation of CO and H2 (synthesis gas or syngas)] appears to be more promising [237-239] than either the direct partial oxidation to CH3OH or the coupling reaction to C2-hydrocarbons, which have so far shown low yields [240,241]. Furthermore, many important chemical processes require syngas, in
Catalytic Properties of Hydrotalcite-Type Anionic Clays
523
various H2/CO ratios (production of ammonia, hydroformylation reactions, synthesis of CH3OH, hydrocarbons, CH3COOH, etc.). The syngas production technologies (steam reforming, autothermal reforming, combined reforming and non-catalytic partial oxidation) [242-246], are well established and have a high thermal efficiency, but being capital and energy intensive, are still being continuously improved by careful investigations of the complex links between catalytic activity, material constraints and process schemes. The catalytic processes operate at significantly lower temperatures than the non-catalytic partial oxidation process (1623-1873 K), and Ni-containing catalysts very often prepared from LDHs have been widely employed, starting from the pioneering work on steam reforming of CH4, in which LDHs were claimed for the first time as catalyst precursors [246]. CH4 + H2O V CO + 3 H2 AH°298 = 206 kJ/mol (Eq. 1) AH°298 = - 36 kJ/mol (Eq. 2) CH4 + V2 O2 n CO + 2 H2 CH4 + CO2 V 2CO + 2 H2 AH°298 = 248 kJ/mol (Eq. 3) CO + H2O V CO2 + H2 AH°298 = -41 kJ/mol (Eq. 4) Steam reforming is strongly endothermic (Eq. 1) and produces higher H2/CO ratios (three or more since an excess of steam is generally used to reduce coke formation) than those required in current downstream processes. Another way of making syngas is direct catalytic partial oxidation (CPO) (Eq. 2), which is exothermic and gives rise to a more desirable H2/CO ratio. Steam or O2 may be replaced by CO2 (Eq 3), this reaction being of industrial interest because of the low H2/CO ratio in the product gas [248,249] and as a CO2 consuming reaction [250]. These three reactions may be combined to achieve the most suitable H2 to CO ratios. The above processes are complex and involve other reactions, the most relevant of them being the water gas shift reaction (Eq. 4), which has a considerable effect on the ratio between CO and H2. All three syngas syntheses can be catalyzed by Ni-supported [251-255] or noble metal containing catalysts [29-31,256-258]. With Ni-containing catalysts, the main problem is the particle size growth as a function of time-on-stream. The larger particle size promotes coke formation and catalyst deactivation. A possible solution is the addition of small amounts of H2S to the feed, blocking the sites for coke formation while maintaining sufficient sites for the reforming reaction (H. Topsoe SPARG process). Very interesting active and stable catalysts have been obtained starting from LDHs, which slow down the growth of the Ni particles, and hence excessive coke formation is avoided without using H2S. Since Ni + ions are randomly distributed in the layered structure of the precursor and somewhat insulated by Mg2+ and Al3+ ions, it was believed that Ni aggregation in LDH-derived catalysts is minimum. This effect, however, is not a general effect of the LDH structure, but depends on the composition as shown by comparing the behaviours of Ni/Cr and Ni/Al LDHs [259]. The application of Ni/Al or Ni/Mg/Al catalysts obtained from LDHs in the preparation of syngas by CPO of CH4, with or without steam (T = 973-1073 K; GHSV = 103-104 h"1), has been widely investigated showing excellent results [28,260-266]. Bhattacharyya and co-workers attributed the superior stability and higher activity of the catalysts obtained by calcination of LDHs in comparison with commercial Ni-supported catalysts to the fact that - in the former samples - the Ni particles are surrounded or "decorated" by crystallites of a spinel-type phase, in agreement with a structural model previously proposed by Clause et al [267]. Sintering and coke formation are reduced
524
S. Albertazzi, F. Basile and A. Vaccari
and there is a corresponding significant increase in the service life of the catalyst. Even though no specific example was reported in all the above patents, the possible application of noble metals (Rh, Ru, Pt or Pd) has also been claimed. Alternative preparation methods have been recently reported for Ni- and Co-containing catalysts, either to obtain a sheet-like morphology [37,38] or to use salt-free reactants, thus reducing the amount of co-products which have to be removed and sent to disposal [268]. The CPO reaction received increased interest after it was shown that high yields in CO and H2 can be achieved by operating at very high GHSV values (105-106 h" '), even in the absence of steam and/or CO2 in the feed [251,252,256]. In comparison to current technologies, this leads to the introduction of small CPO reactors and advantageous process conditions. Again, catalysts obtained by calcination of LDHs can be usefully employed, although for the Ni/Mg/Al catalysts some differences arise in comparison with the previous data obtained at lower GHSV values. In particular, catalysts with high Ni-contents required mild activation conditions, but deactivated rapidly with time-on-stream, while those with low Ni-contents required severe reduction conditions, but show high stability during the reaction [269,270]. Using a stable catalyst with low Ni-content, the roles of various feeds (CH4/O2/He = 2:1:4 or 2:1:20 v/v) and reaction temperatures (873-1273 K) were investigated. The best catalytic performance was observed at high temperature, approaching the thermodynamic equilibrium, while at lower reaction temperatures the specific effects of GHSV and feed composition were observed. At high GHSV values, the best catalytic performances have been observed with catalysts obtained from HT precursors containing small amounts of noble metals. Detailed investigations of the synthesis and thermal evolution of Rh-, Ir, Ru, Pd and Ptcontaining HT have been carried out [29-31,271]. Key factors in the synthesis are the ionic radius and preferential coordination, with the latter playing the more important role. Furthermore, regardless of the calcination temperature (923 or 1173 K), the following tendency to segregate as side metal and/or oxide phases was detected: Pt > Pd > Ru > Ir > Rh (with Rh not showing any segregation phenomena), which was related to the oxidation state and preferential structure of the noble metal ions [31]. Detailed crystal structure refinements of M/Mg/Al mixed oxides (5:71:24 as at. ratio and with M = Ru, Rh or Ir) obtained by calcination of the corresponding LDHs at 923 or 1173 K have been very recently carried out using both neutron and X-ray powder full pattern diffraction analysis [272]. MgO- and spinel-type oxides were detected as main phases, with different solubility of the three noble cations in the various oxides lattices. The Mg/Al ratio significantly affected the structure and distribution of the noble metals in the calcined materials with, furthermore, an increase in activity in the CPO of CH4 by decreasing the Al content (Fig. 19). In the CPO reaction, the syngas formation decreased according to the order Rh > Ru = Ir » Pt > Pd. The best catalytic performances were observed for a 1% Rh content (as atomic ratio) whereas higher Rh contents did not increase the activity, unlike what was observed for Ru-based catalysts (Fig.20) [271]. The Rh/Mg/Al (5: 24:71, as at. ratio) obtained from a LDH precursor has been detailed characterized, thus evidencing the formation of an oxide solid solution of Rh, Mg and Al, responsible for the highly dispersed and active Rh catalyst obtained after reduction [273]. The stability of the Rh on the surface after 100 h of time-onstream was shown by determining the dispersion of the Rh before and after reaction (using HRTEM) and demonstrating the unchanged particle size dispersion.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
525
Figure 19 - Comparison of the activity of some Rh/Mg/AI catalysts containing 5% Rh (as at. ratio) in the CPO of methane carried out at a residence time of 7 ms, with an oven temperature of 773 K and feeding a CH/O/He = 2:1:4 v/v gas mixture [272].
Figure 20 - Activity in CPO ofCH4 as a function of the Ru- or Rh-content (as at. ratio) for catalysts obtained from LDHs by calcination for 14 h at 1173 K and following reduction for 5 h at 1023 K in a an equimolar flow H2/N2 mixture. (Oven temperature = 1023 K; GHSV = 500,000 K''; CH/O2/He = 2:1:4 v/v) [271]. Some of these catalysts have been investigated also in the catalytic total oxidation (CTO) of CH4 [274]: in this case, the lower light-off temperatures were observed for Pd, Rh and Ir based catalysts, but good performances at high CH4 conversion were observed only for the Pd-based catalyst, while Rh- and Ir-based catalysts deactivated with timeon-stream due to either oxidation or sintering phenomena. It was reported that in the CTO of CH4 a mixture of PdO and Pd° was more active at low temperatures, while at higher temperatures (i.e. high conversion values) Pd° was the active species, showing a maximum for an intermediate crystal size. Key points in the CPO of CH4 at high GHSV values are the high exothermicity of the reaction and the presence of hot spot phenomena, that do not allow the accurate determination of the catalyst surface temperature and heat distribution along the catalytic bed. However, operating in autothermal conditions the gas and surface temperatures for Ni, Rh and Rh/Ni catalysts obtained from HT precursors have been determined simultaneously by using a thermocouple and IR thermography, thus
526
S. Albertazzi, F. Basile and A. Vaccari
evidencing very different heat distributions as a function of the catalyst composition [275,276]. Detailed investigations of the thermal profile of the catalytic bed and the gas/solid heat distribution as a function of the catalyst composition, GHSV values and reagent concentration, have been recently reported for these catalysts [276,277]. It was observed that Rh prevented the oxidation of the Ni in the first part of the catalytic bed and that the temperature at the end of the catalytic bed controlled both CH4 conversion and syngas selectivity (Fig. 21).
Figure 21 - IR thermography of the top of the catalytic bed during CPO ofCH4 carried out in autothermal conditions for the catalysts: a) Ni/Mg/Al (10:61:29 as at. ratio) and b) Rh/Ni/Mg/Al. (0.1/6/59/35 as atomic, ratio). (GHSV = 300,000 h-1; CH/O2/He = 2:1:4 v/v) [277]. Specific thermal profiles have been detected as a function of the catalyst composition and reaction conditions: for example, at high residence times a Ni/Mg/Al catalyst (10:61:29 as at. ratio) showed the best activity, due to the shift of the temperature maximum towards the end of the catalytic bed where the thermodynamic equilibrium is reached [278]. On the contrary, at low residence times (i.e. in O2 rich conditions) the Ni-containing catalyst was fully inactive, while a Rh/Ni/Mg/Al (0.1/6/59/35 as atomic ratio) catalyst was much more active than a Rh/Mg/Al (1:71:28 as at. ratio) catalyst in term of both CH4 and O2 conversion, notwithstanding the lower amount of Rh. This result indicates the positive effect of Rh which allows to maintain a higher amount of Ni in the reduced state also at low residence times. Thus, LDHs may be considered as optimum precursors to favour the development of synergetic effects between the elements in the final catalysts [275,278]. Finally, the significant improvement in mechanical and thermal stability in CPO of CH4 at high GHSV values is noteworthy as claimed very recently for tailormade materials obtained by calcination of Ni/Mg/Al, Rh/Mg/al and Rh/Ni/Mg/Al LDHs containing different amounts of silicates as interlayer anions [279,280]. The ex-LDHsilicates Rh, Ni and/or Rh/Ni catalysts were more active and selective at very high GHSV values than the analogous catalysts obtained from LDHs containing carbonates with, furthermore, a stable activity with time-on-stream, also operating in a pilot plant. Ni/Al, Rh/Mg/Al and Rh/Ni/Mg/Al catalysts obtained from LDH precursors have been applied in C02-reforming of CH4 [281,282]. Under severe operating conditions (T= 1133 K; P = 2,1 MPa and GHSV= 14,400 h"1) Ni/Al catalysts showed performances identical to those of commercial catalysts, with coke formation and
Catalytic Properties of Hydrotalcite-Type Anionic Clays
527
pressure build-up across the catalytic bed after a short period of time [281]. Under less severe operating conditions, the LDH-derived catalysts exhibited higher activity and stability than commercial catalysts, with very low coke formation (Fig. 22). No significant differences were observed in the LDH-derived catalyst as a function of a preliminary reducing treatment, while for lower Ni-contents, a longer induction time (24 h ca) was required to reach full activity.
Figure 22 - Conversion as a function oftime-on-stream in the CO2-reforming ofCH4for a (a) or (v) prereduced Ni/Al catalyst obtained from a LDH precursor and (u) a commercial catalyst, which showed indication of coke formation after one week on stream (T = 1133-1163 K; P = 0.7MPa,CO2/CH4 = 1.0 andH2O/CH4 = 0.5 v/v) [281]. Analogous catalysts were also investigated in CO2-reforming at high GHSV values (150-300,000 h"1, T= 945-960 K, P = 0.1 MPa), to highlight the contribution of CO2-reforming reaction in the CPO of CH4 and check the possibility of using smaller reactors [282]. Higher activity of Ni towards CO2 and of Rh towards CH4 was observed. The bimetallic catalysts Rh/Ni/Mg/Al exhibited enhanced conversion of both CO2 and CH4, thus confirming the existence of synergetic effects in the catalysts obtained from LDHs. Finally, these catalysts did not show any structural and/or physical modification after the catalytic tests with, furthermore, no deactivation phenomena. Finally, it is noteworthy that Mg/Al LDH - also intercalated with [Ni(edta)]2" species - has been reported as a precursor for an efficient catalyst of CO2-reforming of CH4 after calcination for 16 h at 773 K and reduction for 15 min at 873 K in a H2/N2 (10:35 v/v) flow [283]. The catalyst maintained high activity within 150 h time-on-stream at 1073 K and could be used repeatedly after regeneration. Furthermore, although coke deposition onto the catalyst surface attained 5-10 wt.%, it did not diminish conversion and selectivity values. Ni-containing catalysts obtained from LDHs have been reported as very active and selective also in the methanation of CO (inverse reaction of steam reforming) [3]. The Ni/Al catalysts obtained from LDHs containing CO32" as interlayer anions showed good stability under hydrothermal conditions, together with high activity and selectivity
528
S. Albertazzi, F. Basile and A. Vaccari
also in the absence of alkali ions [284], generally added to improve activity. Active and stable catalysts for methanation reaction have also been prepared from ternary Ni/Mg/Al HT precursors [285]. The role of LDHs as precursors of catalysts for the conversion of syngas to more valuable intermediates (CH3OH, higher alcohols and/or hydrocarbons) has already been reviewed [3], therefore only the main features affecting the catalytic behaviour and some recent investigations will be reported here. LDHs were used to prepare Zn/Cr catalysts for the synthesis of methanol at high temperature and pressure [286]. Moreover, Cu/M(II)/Cr [M(II) = Zn, Mg, Co or Mn] and, mainly, Cu/Zn/Al mixed oxides obtained by calcining LDHs were widely investigated in the synthesis of CH3OH at low temperature and pressure (523 K and 5.0 MPa) [3], although using Cu/Zn/Al catalysts the highest activity was obtained with catalysts with low Al-content, i.e. obtained from mixtures of hydroxycarbonates and LDHs [287-290]. It has been reported that aurichalcite and LDH gave rise to more active catalysts than rosasite [290,291]. While keeping the composition constant, the relative amounts of the LDH and rosasite phases were varied by using an ultrasonic generator during the precipitation and aging steps [291]. The amount of LDH increased with the ultrasound frequency, with a corresponding increase in the catalytic activity, attributed to the formation of highly dispersed Cu(Cu+), ZnO and A12O3 particles. On the other hand, in conditions close to those of the reaction, the presence in the most active Cu/Zn/Al catalyst of well dispersed Cu° and an amorphous dehydrated (Cu,Zn)6Al2(OH)16CO3 phase was reported [290]. Among the Cu/M(II)/Cr systems obtained from LDHs, it was observed that the addition of Co or Zn considerably modified the activity and/or selectivity of the Cu/Cr catalyst, due to the formation of cubic non-stoichiometric spinel-type mixed oxides, that - by reduction - gave rise to highly dispersed Cu° not detectable by XRD analysis and probably characterised by low-Miller-index surface planes [292,293]. Formation of segregated cobalt oxides and/or metallic cobalt was never detected, regardless of the Co-content. Zn significantly promoted the synthesis of methanol, while Co exhibited a dramatic poisoning effect when present in amounts up to 2% (as at. ratio), while further Co additions resulted in an increase in the activity, but with selectivity in hydrocarbons (Fig. 23). The low Co amounts poisoned the oxidizing capacity of the catalyst surface, while for higher Co-contents a synergetic effect occurred between the well-dispersed metallic copper and the Co-containing cubic spinel-type phase [293]. The nature and the catalytic activity in the Fischer-Tropsch synthesis of Co/Cu/Zn/Cr mixed oxides, mainly obtained from LDHs, have been investigated for a wide range of compositions [294]. For all catalysts the hydrocarbons were the main products and presented typical Schulz-Flory distributions. While the Co/Cr catalysts showed very low activity, a maximum was obtained for catalysts containing comparable amounts of Co and Cu, in agreement with the hypothesis on the role of synergetic effects between these elements. More recently [295], a high selectivity in C2-C4 paraffins [useful in obtaining liquefied petroleum gas (LPG) from syngas] has been achieved using a hybrid catalyst prepared by mixing in the weight ratio 1:1a Cu/Zn/Al (38/38/24 as at. ratio) catalyst for CH3OH, obtained by calcination for 24 h at 633 K h and successive reduction for 8 h at 633 K of a LDH precursor, with a Mn (4.65 wt.%)exchanged H-ZSM5 zeolite. Up to 523 K CH3OH was the main product, which however converted to hydrocarbons with increasing temperature; with the best C2-C4 selectivity observed at 548 K. The relative rate of formation on the zeolite of carbenium ions of
Catalytic Properties of Hydrotalcite-Type Anionic Clays
529
various chain lengths determined the selectivity in C2-C4 olefins, that underwent rapid hydrogenation to the corresponding paraffins on the CH3OH catalyst, thus preventing the formation of aromatics.
Figure 23 - Total productivity as a function of Co content for Cu/Co/Zn/Cr catalysts obtained by calination for 24 h at 653 K of LDHs (T = 555 K, P = 1.2 MPa, GHSV 15,000 A"'; H2/CO/CO2 = 65:32:3 v/v) [293,294]. Different classes of catalysts obtained by calcination of LDHs have been claimed for the synthesis of mixtures of CH3OH and higher molecular weight (HMW) alcohols, which may be used as high octane blending stock for gasoline. In some cases the claimed catalysts are those active in the synthesis of CH3OH at high (Zn/Cr) or low [Cu,Zn,Al(Cr)] temperature, doped with an alkali promoter to form mainly branched alcohols [3]. In the latter case, the most selective and stable catalysts were obtained from pure LDHs and required lower amounts of an alkali promoter, with a key role of the catalyst composition and reaction parameters [296]. In other cases the catalyst contained Fischer-Tropsch elements together with copper and produced mainly linear primary alcohols. Also these systems required the presence of small amounts of an alkali element to enhance the synthesis in HMW alcohols and avoid uncontrolled methanation. A chain growth scheme for the synthesis of HMW alcohols from syngas has been proposed by Smith and Anderson [297], assuming one or two carbon additions at the a- or p-carbon atom of the growing alcohol. Figure 24 shows the selectivity in the different classes of products as a function of the precursor composition for the alkalized ternary Co/Cu/Cr systems [298]. Intermediate compositions, especially in the range 1.0 < Cu/Co < 3.0, with Co/Cr > 0.5, gave rise to active and selective catalysts for the synthesis of C1-C6 alcohols, with high purity of the alcohol phase, since the presence of Co lowered the formation of esters, ketones and aldehydes, being C r C 6 hydrocarbons the main by-products obtained. It must be noted that for compositions typical of LDHs [i.e. M(I1)/M(II1) atomic ratios ranging from 2.0 and 3.0] CH3OH was the main product, while the HMW alcohol synthesis required an at. ratio ranging between 0.5 and 1.0, thus suggesting the formation also of an amorphous phase containing part of the Cr3+ ions. An interesting review of the most recent trends and developments in the science and technology of
530
S. Albertazzi, F. Basile and A. Vaccari
catalyzed hydrogenation of carbon oxides (CO and CO2) has been published recently [299].
Figure 24 - Main products obtained as a function of the composition with the Cu/Co/Cr catalysts [298]. Among the most recent applications of CH3OH, the production of H2-rich gaseous fuels have been investigated for spark ignition engines and coupled-fuel-cells electric engines [300,301]. This latter application requires a CO-free H2 stream, since CO irreversibly poisons the Ru/Pt electrocatalyst in polymer electrolyte CH3OH fuel cells and dramatically diminishes their performances. H2 can be obtained from CH3OH via three different processes: a) decomposition (Eq. 5); b) steam reforming (SR) (Eq. 6); and c) partial oxidation (PO) (Eq. 7): CH3OH ^ /2H 2 + C0 CH3OH H H2O A/3 H2 + CH3OH H- V2 0 2
A H 298 =
co2
n2H 2 + co 2
- 92 kJ/mol = 49 kJ/mol 298 = -192kJ/mol
A H 298 AH
(Eq. 5) (Eq. 6) (Eq. 7)
To date, the SR reaction of CH3OH has been the only process used for H2 production for fuel cell applications [302],This reaction, however, produces a considerable amount of CO as a by-product, which may be removed in a second-stage catalytic reactor via either the water gas shift reaction (Eq. 4) or even CO oxidation [303]. For fuel cell technology, the PO reaction of CH3OH is advantageous in comparison with SR reaction, because it uses air instead of steam and does not require any external heat. LDHs have been claimed for the preparation of catalysts active in all the above processes [304-311]. For example, Pd and Rh-based catalysts obtained by calcination of LDHs for 3 h at 723 or 923K and subsequent reduction for 1 h at 673 K have been investigated in the decomposition of CH3OH, in order to decrease the temperature of H2 production. Higher metal dispersion and activity was obtained with a Pd/Mg/Cr catalyst, significantly better than those of analogous impregnated catalysts (Fig. 25) [304].
Catalytic Properties of Hydrotalcite-Type Anionic Clays
531
Figure 25 - Arrhenius plot of the methanol decomposition rate over various Pd catalysts reduced at 673 K: (v) 3.0 wt.% Pd impregnated on MgO; (») 2.9:2.9 wt.% Pd/Cr impregnated on MgO; (n) 3.0 wt. % Pd impregnated on Mg/Cr LDH calcined for 3 h at 698 K; (A) 3.8 wt. % Pd/Mg/Cr LDH calcined for 3 h at 698 K [304]. On the other hand, a novel catalyst with high activity and selectivity to syngas, due to the large surface area and Pd dispersion, was prepared from a composite obtained by synthesizing the LDH precursor in a suspension of hexagonal mesopouros silica (HMS) [305]. Finally, in situ production of H2 by catalytic decomposition of different alcohols in a 10% (v/v) mixture with simulated gasoline has been recently reported using a Cu/Al (72:28 as at. ratio) catalyst, derived from the corresponding LDH during the reaction [306]. The onset of H2 formation generally occurred at 473-503 K, and was related to the formation of metallic Cu during in situ modification of the initial LDH. In alcohol/fuel mixtures, dehydrogenation of the alcohols appeared to be the major mechanism, with significant irreversible catalyst deactivation above 623-633 K. Cu/Zn/Cr and/or Cu/Zn/Al catalysts, obtained either from pure LDHs or mixtures of LDH and hydroxycarbonate phases, have been widely employed in the SR and PO reactions of CH3OH [301,307,308]. A Cu/Zn/Cr LDH calcined at 573 K showed a high activity in the SR reaction at 523 K, with negligible deactivation with time-onstream and, furthermore, optimum activity and stability also in the tests carried out at high pressure [307]. Cu/Zn/Al catalysts - derived from precursors containing mainly a LDH - exhibited good catalytic activity in the PO reaction, with at 573 K a CH3OH conversion of 40-60% and high selectivity in H2 (>90%) and CO2 (>95%) [308], although the possible presence of hot spot phenomena should be considered, giving rise to higher temperatures on the catalyst surface. The catalytic activity decreased with increasing (Cu+Zn)/Al atomic ratio in the precursor, while increasing amounts of hydroxycarbonates favoured the formation of considerable amounts of dimethyl ether as a by-product. The low temperature (423-673 K) SR reaction of CH3OH to produce H2 has been investigated also over Mg/Al, Cu/Al, Co/Al and Ni/Al catalysts formed in situ during the reaction from the corresponding LDHs [309]. The reducibility of the divalent cations present in the LDH was a crucial parameter in determining the SR activity of the catalysts. The Cu/Al catalyst was the most efficient, becoming active at 503 K ca., while Ni/Al and Co/Al catalysts required significantly higher (588-593 K) activation temperatures and Mg/Al was fully inactive. Furthermore, pre-activation of the Cu/Al
532
S. Albertazzi, F. Basile and A. Vaccari
LDH by calcination for 4 h at 673 K, followed by reduction for 1 h at 673 K in dilute H2 did not significantly change the catalytic activity (Fig. 26).
Figure 26 - Steam-reforming activity as a function of the temperature for the pretreatedCu/Al (3.1:1.0 as at. ratio) LDH-derived catalyst [309] A way of significantly reducing the hot spot phenomena is to combine the SR and PO processes by simultaneously cofeeding O2, steam and CH3OH to obtain oxidative CH3OH reforming (OSR). Furthermore, the ratio of three reactants can be chosen in order to supply the heat necessary to maintain the steam reforming reaction by the partial oxidation reaction, with an overall reaction heat near to zero [300,301]. Velu et al [310,311] have recently reported LDH-derived Cu/Zn/Al and/or Cu/Zn/Al(Zr) catalysts very active in OSR of CH 3 0H, producing CO-free H2 suitable for fuel cells, with up to 100% CH3OH conversion after 25 h of on-stream operation at 503 K ca. During the PO reaction, levels of CO above 3% are usually obtained, but the temperature and excess H2O present in the OSR process gives favourable conditions to transform CO into CO2 by the water gas shift reaction (Eq. 4) [301]. Among the Cu/Znbased catalysts, those containing Zr were the most active, with optimum O2/CH3OH and H2O/CH3OH molar ratios in the ranges 0.2-0.3 and 1.3-1.6, respectively [311]. Figure 27 compares the rates of H2 production and outlet CO concentrations in the SR and OSR reactions, showing that the latter reaction was more efficient since the rate of H2 production increased by six fold, while the outlet CO level only by a factor of two. Catalysts always obtained from LDHs (containing Ni2+, Zn2+ and/or Cu2+) have also been claimed for the steam reforming of dimethyl ether [312] or other oxygenated compounds (ethers, alcohols, C2-C4 aldehydes or ketones, although examples have been reported only for methanol) [313,314] to produce H2, together with other products (for example acetic acid), without the use of excess steam (T= 473-573 K; P = 0.5-2.5 MPa; GHSV= 2,000-3,000 h"1). Finally, Cu/Zn/Cr catalysts obtained from LDHs were investigated in the use of CH3OH for H2 storage and as a H2 carrier, showing high stability in CH3OH dehydrogenation to methyl formate. In these catalysts the formation of spinel-type mixed oxides avoided the severe deactivation observed with the Cu/ZnO based catalysts, due to the reduction of ZnO in the vicinity of Cu [315].
Catalytic Properties of Hydrotalcite-Type Anionic Clays
533
Figure 27 - Comparison of the efficiency of the CH3OH steam-reforming (SR) and oxidative-steam-reforming (OSR) reactions, operating at 503 K and a constant CH3OH conversion value of 50% [311]. 7 - Concluding remarks For the layered double hydroxides [LDHs, also called hydrotalcite-type (HT) anionic clays] a broad spectrum of catalytic applications already exists and even more will probably be found in the future in unexpected areas, due to the possibilities of designing catalysts tailored for specific reactions and/or substrates. Thus, the exponential increase in the number of publications referring to LDHs in recent years is not surprising. Even more significant is the percentage of these papers reporting catalytic applications (as such or, mainly, after thermal decomposition), thus showing the increasing academic and industrial interest for these materials and their relevance in catalysis [316]. In all cases, it must be noted the high flexibility and potential of synthetic LDHs as precursors of multicomponent catalysts, due to the large number of composition and preparation variables that may be adopted. One of these promising fields is the increase in the longevity by intercalation in LDHs of biomimetic catalysts, which have generally a very limited stability as homogeneous catalysts (see, for example, Sections 3 and 5). As previously mentioned, since the interlayer space is not accessible, the reaction occurs on the external edge-sites of the crystallites [140]. LDHs intercalated with Mn-porphyrins have been reported as suitable catalysts for many reactions [317], as the epoxidation of cyclohexene using PhIO, while very poor activity was observed in the hydroxylation of heptane, thus suggesting that the accessibility of the metal centres depended on substrate polarity [318]. Furthermore, mention has to be made of the application as biomimetic catalysts of Mg/Al or Ni/Al LDHs exchanged with tungstates (to an extent of 12% and with the WO42" anions mainly located in edge positions) [319]. These catalysts were used in selective oxidative bromination and bromide-assisted epoxidation reactions with an activity 100 times higher than the homogeneous analogue (Fig. 28). The low cost and heterogeneous nature of this system, together with its ability to operate under mild reaction conditions (298 K; 6 < pH < 8 ) using bromides rather than Br2, suggest interesting environmentally friendly routes to produce brominated chemicals and
534
S. Albertazzi, F. Basile and A. Vaccari
drugsand epoxide intermediates. As recently reported, the latter may be considered an interesting example of "engineered solid catalysts" based on ion exchange properties of cationic or anionic clays (LDHs) [320]. The polarity of these supports can be varied by an appropriate ion exchange, tailoring the catalytic activity, such as it has been reported for tungstate (hydrophilic) or tungstate/p-toluensolphonate (hydrophobic) exchanged Mg/Al LDHs in the olefin epoxidation with H2O2 as a function of the nature of the organic substrate [321].
Figure 28 - Catalytic pathway in bromination with WO/~ LDH [319]. H2O2 binds to tungstate to form peroxotungstate at the surface. Electrostatic attraction brings bromide to the surface, facilitating transfer of the activated oxygen atom from peroxotungstate to Bf. Reaction of 2-electron-oxidized Br species in solution: electrophilic bromination of phenol red into phenol blue (1) and bromide-assisted 'O2 generation from H2O2 [319]. Finally, it must be pointed out that the concept of catalytic applications of LDHs should have a wider meaning, including all the "catalytic devices", i.e. both usual heterogeneous catalysts and materials based on the exploitation of the catalytic properties, like gas sensors, porous catalytic membranes, semiconductor surface devices, etc. In this frame, mention has been made to the applications as electrode coatings, potentiometric sensors or in electrocatalysis [322-328]. Acknowledgments Financial support from the Ministero per 1' Istruzione, 1' Universita e la Ricerca [MIUR, Roma (I)] and the EU-Growth Project G5RD-CT2001-00537 is gratefully acknowledged. 8 - References [1] S. Miyata, Industrial use of hydrotalcite-like compounds, Kagaku Gijutsushi MOL 15(10) (1977) 32 (in Japanese) [2] W.T. Reichle, Chemtech 16 (1986) 58. [3] F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 11 (1991) 173. [4] A. de Roy, C. Forano, K. El Malki and J.-P. Besse, Synthesis of Macroporous Materials, vol. 2: Expanded Clays and Other Microporous Solids, Eds. M.L. Occelli, and H.E. Robson, Van Nostrand Reinhold, New York, USA, 1992, ch. 7. [5] F. Trifiro and A. Vaccari, Comprehensive Supramolecular Chemistry, vol. 7: SolidState Supramolecular Chemistry: Two and Three-Dimensional Inorganic Networks, Eds. J.L. Atwood, J.E.D Davies, D.D. MacNicol and F. Vogtle, Pergamon, Oxford, UK, 1996, ch. 8.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
535
[6] A.Vaccari, Catal. Today, 41 (1998) 53. [7] A. Vaccari, Appl. Clay Sci., 14 (1999) 161. [8] V. Rives and M.A. Ulibarri, Coord. Chem. Rev., 181 (1999) 61. [9] V.A. Drits, T.N. Sokolova, G.V. Sokolova and V.I. Cherkashin, Clays Clay Miner., 35(1987)401. [10] W. Feitknecht, Helv. Chim. Acta, 25 (1942) 131. [11] W. Feitknecht, Helv. Chim. Acta, 25 (1942) 555. [12] R. Allmann, Acta Crystallogr. B, 24 (1968) 972. [13] H.F.W. Taylor, Miner. Mag., 37 (1969) 338. [14] W. T. Reichle, Solid State Ionics, 22 (1986) 135. [15] W. Kagunya, Z. Hassan and W. Jones, Inorg. Chem., 35 (1996) 5970. [16] S. Carlino, Chemistry in Britain, 33 (9) (1997) 59. [17] F. Basile, G. Fornasari and A. Vaccari, Encyclopedia of Surface and Colloid Science, Ed. A. Hubbard, Marcel Dekker, New York, USA, 2002, p. 909. [18] F. Basile and A. Vaccari, Layered Double Hydroxides: Present and Future, Ed. V. Rives, Nova Science Publisher, New York, USA, 2001, ch. 10. [19] X. Han and Y. Wang, Progress in preparation and applications of hydrotalcite and hydrotalcite-like materials, Jiangsu Huagong, 31,2 (2003) 26 (in Chinese). [20] N. Takahashi and K. Kosuge, Applications of hydrotalcite-like compounds to water and wastewater treatment, Mizu Shori Gijutsu, 44, 4 (2003) 165 (in Japanese). [21] X. An, X. Xie and Z. Wang, Properties and catalytic applications of hydrotalcitelike compounds, Taiyuan Ligong Daxue Xuebao, 33, 5 (2002) 498 (in Chinese). [22] L. Van der Ven, M.L.M. van Gemert, L.F. Batenburg, J.J. Keera, L.H. Gielgens, T.P.M. Koster and H.R. Fischer, Appl. Clay Sci., 17 (2000) 25. [23] F. J. Brocker and L. Marosi, German Pat., 2, 024, 282 (1970) to BASF AG. [24] P.V. Kamath, G.H.A. Therese and J. Gopalakrishnan, J. Solid State Chem., 128 (1997)38. [25] V. Rives, F.M. Labajos, M.A. Ulibarri and P. Malet, Inorg. Chem., 32 (1993) 5000. [26] N.J. Blom and E. G. Derouane, Eur. Pat. Appl., 475, 357 (1992) to H. Topsoe A/S. [27] C.S. Swamy, S. Kannan, Y. Li, J.N Armor and T.A. Braymer, Eur. Pat. Appl., 640, 379 (1994) to Air Products and Chemicals Inc. [28] A. Bhattacharyya, M.S. Kleefisch and C.A. Udovich, US Pat., 5, 439, 861 (1995) to Amoco Co. [29] L. Basini, G. Fornasari, F. Trifiro and A. Vaccari, Eur. Pat., 725, 038B1 (1996) to Snamprogetti SpA. [30] F. Basile, L. Basini, G. Fornasari, M. Gazzano, F. Trifiro and A. Vaccari, J. Chem. Soc. Chem. Commun., (1996) 2453. [31] F. Basile, G. Fornasari, M. Gazzano and A. Vaccari, Appl. Clay Sci., 16 (2000) 185. [32] R.M. Taylor, Clay Miner., 19 (1984) 591. [33] S. Velu, V. Ramaswamy, A. Ramani, B.M. Chanda and. S. Sivasanker, J. Chem. Soc. Chem. Commun., (1997) 2107. [34] S. Velu, K. Suzuki, M. Okazaki, T. Osaki, S. Tomura and F. Ohasashi, Chem. Mater., 11(1999)2163. [35] S. Velu, K. Suzuki, M.P. Kapoor, S. Tomura, F. Ohashi and T. Osaki, Chem. Mater., 12(2000)719.
536
S. Albertazzi, F. Basile and A. Vaccari
[36] A.A. Schutz, L.A. Cullo and C.P. Kelkar, US Pat, 5, 399, 329 (1995) to Aristech Chemical Co. [37] C.P. Kelkar, A.A. Schutz and L. Cullo, US Pat, 5, 518, 704 (1996) to Aristech Chemical Co. [38] C. P. Kelkar and A.A. Schutz, Micropor. Mater., 10 (1997) 163. [39] T. Lopez, P. Bosch, E. Ramos, R. Gomez, O. Navarro, D. Acosta and F. Figueras, Langmuir, 12(1996)189. [40] F. Prinetto, G. Ghiotti, P. Graffin and D. Tichit, Micropor. Mesopor. Mater., 39 (2000) 229. [41] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego and M. Saba, Micropor. Mesopor. Mater., 66 (2003) 77. [42] K. Diblitz, K. Noweck, A. Brasch and J. Schiefler, Eur. Pat, 807, 086 (1996) to RWE-DEA Aktiengesellschaft fur Mineralogie und Chemie. [43] E.S. Martin, J.M. Stinson, V. Cedro III and W.E. Horn Jr., US Pat, 5, 730, 951 (1998) to Aluminum Co. of America. [44] A.A. Vierheilig, WO Pat, 9, 920, 389 (1999) to Bulldog Tecnologies. [45] D. Stamires, M. Brady, W. Jones and F. Kooli, WO Pat, 9, 941, 198 (1999) to Akzo Nobel N.V. [46] T. Kameda, T. Yoshioka, M. Uchida and A. Okuwaki, Phosphorus, Sulfur and Silicon and the Related Elements, 177 (2002) 1503. [47] H. Muller-Tamm, H. Frielingsdorf, G. Schweier and L. Reuter, German Pat, 2, 163,851 (1971) to BASF AG. [48] A.A. Bhattacharyya, G.M. Woltermann, J.S. Yoo, A. Karch and W.E. Cormier, Ind. Eng. Chem. Res, 27 (1988) 1356. [49] J.S.Yoo, A.A. Bhattacharyya and C.A. Radlowski, Ind. Eng. Chem. Res, 30 (1991) 1444. [50] R. Zhao, C. Yin, H. Zhao and C. Liu, Fuel Process. Technol, 81 (2003) 201. [51] M. Gazzano, B. Montanari and A. Vaccari, Actas XV Simp. Iberoam. Catalisis, vol. 3, Eds. E. Herrero, O. Annunziata and C. Perez, Univ. Nac. Cordoba, Cordoba (Arg), 1996, p. 1967. [52] J. Pasel, P. Ka)3ner, B. Montanari, M. Gazzano, A. Vaccari, W. Makowski, T. Lojewski, R. Dziembaj and H. Papp, Appl. Catal. B, 18 (1998) 199. [53] T. Blasco , J.M. Lopez Nieto, A. Dejoz and M.I. Vaquez, J. Catal, 157 (1995) 271. [54] J.M.Lopez Nieto, A. Dejoz and M.I. Vaquez, Appl. Catal. A, 132 (1995) 41. [55] K. Bahranowski, G. Bueno, V. Cortes Corberan, F. Kooli, E.M. Serwicka, R.X. Valenzuela and K. Wcislo, Appl. Catal. A, 185 (1999) 65. [56] P.C.H. Mitchell and S.A. Wass, Appl. Catal. A, 225 (2002) 153. [57] P. Dinka, K. Prandova and M. Hronec, M , Appl. Clay. Sci, 13 (1998) 467. [58] E.G. Derouane, R.J. Davis an N.J.; Blom, Eur. Pat. Appl, 476, 489 (1992) to H. Topsoe A/S. [59] E.G. Derouane, V. Jullien-Lardot, R.J. Davis, N. Blom and P.E. Hoilund-Nielsen, New Frontieres in Catalysis, vol. B, Eds. L. Guczi, F. Solymosi and P. Tetenyi, Akademiai Kiado, Budapest, H, 1993, p. 1031. [60] Z. Gandao, B. Coq, L.C. de Menorval and D. Tichit, Appl. Catal. A, 147 (1996) 395. [61] Y.Z. Chen, CM. Hwang and C.W. Liaw, Appl. Catal. A, 169 (1998) 207. [62] R. Unnikrishnan and S. Narayanan, J. Mol. Catal. A: Chemical, 144 (1999) 173.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
537
[63] T.J. Pinnavaia, M. Rameswaran, E.D. Dimotakis, E.P. Giannelis and E.G. Rightor, Faraday Discuss. Chem. Soc, 87 (1989) 227. [64] H. Fukuhara, F. Matsunaga and Y. Nakashima, Eur. Pat. Appl. 0, 323, 192 (1988) to Mitsui Petrochemical Industries Ltd. [65] S. Narayanan and K. Krishna, Appl. Catal. A, 174 (1998) 221. [66] S. Narayanan and K. Krishna, Catal. Today, 49 (1999) 57. [67] A.H. Padmasri, A. Venugopal, J. Krishnarmurthy, K.S. Rama Rao and P. Kanta Rao, J. Mol. Catal. A: Chemical, 181 (2002) 73. [68] L. Balduzzi, F. Prinetto, G. Ghiotti, A. Bianchini, M. Livi and A. Vaccari, Studies in Surface Science and Catalysis, vol. 140: Oxide-based Systems at the Crossroad of Chemistry, Eds. A. Gamba, C. Colella and S. Coluccia, Elsevier, Amsterdam, NL, 2001, p. 67. [69] G. Centi, G. Fornasari, C. Gobbi, M. Livi, F. Trifiro and A. Vaccari, Catal. Today, 73 (2002) 287. [70] G. Fornasari, F. Trifiro, A. Vaccari, F. Prinetto, G. Ghiotti and G. Centi, Catal. Today, 75(2002)421. [71] G. Centi, G.E. Arena and S. Perathoner, J. Catal., 216 (2003) 443. [72] R. Teissier, D. Tichit, F. Figueras and J. Kervennal, French Pat., 00, 094 (1995) to ELF Atochem. [73] D. Tichit, M.N. Benanni, F. Figueras, R. Tessier and J. Kervennal, Appl. Clay Sci., 13(1998)401. [74] F. Figueras, D. Tichit, M. Bennani Nacri and R. Ruiz, Catalysis of Organic Reactions, Ed. F.E. Herkes, Dekker, New York, USA, 1998, p. 37. [75] F. Prinetto, G. Ghiotti and D. Tichit, Studies in Surface Science and Catalysis, vol. 130: Proc. 12th Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, 3243. [76] W.T. Reichle, J. Catal., 63 (1980) 295. [77] W.T. Reichle, Eur. Pat. Appl., 095, 783 (1983) to Union Carbide Co. [78] C.P. Kelkar and A. A. Schutz, Appl. Clay Sci., 13 (1998) 417. [79] J.I. Di Cosimo, V.K. Diez and C.R. Apesteguia, Appl. Clay Sci., 13 (1998) 433. [80] J. Lopez, R. Jacquot and F. Figueras, Studies in Surface Science and Catalysis, vol. 130: Proc. 12* Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 491. [81] V.R.L. Constantino and T.J. Pinnavaia, Catal. Letters, 23 (1994) 361. [82] D. Kishore and S. Kannan, Green Chem., 4 (2002) 607. [83] A.L. McKenzie, C.T. Fishel and R.J. Davis, J. Catal., 138 (1992) 547. [84] A. Corma, V. Fornes and F. Rey, J. Catal., 148 (1994) 205. [85] J. Shen, J.M. Kobe, Y. Chen and J.A. Dumesic, Langmuir, 10 (1994) 3902. [86] D. Tichit, M.A. Lhouty, A. Guida, B.H. Chiche, F. Figueras, A. Auroux, D. Bartalini and E. Garrone, J. Catal., 151 (1995) 50. [87] K.K. Rao M. Gravelle, J. Sanchez Valente and F. Figueras., J. Catal., 173 (1998) 115. [88] J. Sanchez Valente, F. Figueras, M. Gravelle, P. Kumbhar, J.-P. Besse, J. Catal., 189(2000)370. [89] E. Dumitriu, V. Hulea, C. Chelaru, C. Catrinescu, D. Tichit and R. Durand, Appl. Catal. A, 178 (1999) 145. [90] J.C.A.A. Roelofs, D.J. Lensveld, A.J. van Dillen and K.P. de Jong, J. Catal., 203 (2001) 184.
538
S. Albertazzi, F. Basile and A. Vaccari
[91] A. Corma, V. Fornes, R.M. Martin-Aranda and F. Rey, J. Catal., 134 (1992) 58. [92] U. Constantino, M. Curini, F. Montanari, M. Nocchetti and O. Rosati, J. Mol. Catal. A: Chemical, 195 (2003) 245. [93] M. Campanati, S. Franceschini, O. Piccolo, A. Vaccari and A. Zicmanis, Catal. Commun., in press. [94] W. Kagunya, Z. Hassan and W. Jones, Inorg. Chem., 35 (1996) 5970. [95] S. Cheng and J.T. Lin, Synthesis of Macroporous Materials, vol. 2: Expanded Clays and Other Microporous Solids, Eds. M.L. Occelli, and H.E. Robson, Van Nostrand Reinhold, New York, 1992, NL, p. 170. [96] J.T. Lin and S. Cheng, Abstracts of 11th Int. Congress on Catalysis, Baltimore, USA, 1996, Poster 012. [97] L. Forni, G. Fornasari, R. Trabace, F. Trifiro, A. Vaccari and L. Dalloro, Studies in Surface Science and Catalysis, vol. 130: Proc. 12th Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 3513. [98] L. Forni, G. Fornasari, E. Miserazzi, R. Trabace, F. Trifiro, A. Vaccari and L. Dalloro, Proc. 9th Intern. Symp. Hetrogeneous Catalysis, Eds. L. Petrov, Ch. Bonev and G. Kadinov, Bulg. Acad. Sciences, Sofia, BG, 2000, p. 507. [99] R. Bechara, A. D'Huysser, M. Fournier, L. Forni, G. Fornasari, F. Trifiro and A. Vaccari, Catal. Letters, 82 (2002) 59. [100] Y. Izumi, S. Sato and K. Urabe, Chem. Letters, (1983) 1649. [101] S. Velu and C.S. Swamy, Appl. Catal. A, 145 (1996) 225. [102] S. Velu and C.S. Swamy, Appl. Catal. A, 162 (1997) 81. [103] S. Velu and S. Sivasanker, Res. Chem. Intermed., 24 (1998) 657. [104] F. Cavani, C. Felloni, D. Scagliarini, A. Tubertini, C. Flego and C. Perego, Studies in Surface Science and Catalysis, vol. 143: Scientific Bases for the Preparation of Hetrogeneous Catalysts, Eds. E. Gaigneaux, D.E. De Vos, P. Grange, P.A. Jacobs, J.A. Martens, P. Ruiz and G. Poncelet, Elsevier, Amsterdam, NL, 2002, p. 953. [105] M. Bolognini, F. Cavani, C. Felloni, D. Scagliarini, C. Flego and C. Perego, Catalysis of Organic Reaction, Ed. D.G. Morrell, Dekker, New York, USA, 2002, p. 115. [106] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego and M. Saba, Catal. Today, 75 (2002) 103. [107] T.M. Jyothi, T. Raja, M.B. Talawar and B.S. Rao, Appl. Catal. A, 211 (2001) 41. [108] H. Soo, B.C. Ream and H. Robson, US Pat., 5,064,804 (1991) to Union Carbide Chemicals & Plastics Technology Co. [109] M.P. Atkins and M. Chibwe, US Pat., 5, 110, 992 (1992) to The British Petroleum Co. [110] M.W. Forkner, US Pat., 5, 260, 495 (1993) to Union Carbide Chemicals & Plastics Technology Co. [111] H-C. Raths, W. Beuer, K. Friedrich and K. Hermann, US Pat., 5, 292, 910 (1994) to Henkel Kommanitgesellschaft auf Aktien. [112] M.P. Atkins, C.L. Depege and C.R.M. Forano, PCT WO 00, 842 (1997) to BP Chemicals Ltd. [113] J.A. Barhhorst, A.F. Elsasser. Jr. and C.W. Blewett, US Pat, 5, 907, 067 (1999) to Henkel Co. [114] V.R. Choudary, S.K. Jana and V.S. Narkhede, Appl. Catal. A, 235 (2002) 207.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
539
[115] G. Carja, R. Nakamura and H. Niiyama, Appl. Catal. A, 236 (2002) 91. [116] R. Prihod'ko, M. Sychev, I. Kolomitsyn, P.J. Stobbelaar, EJ.M. Hensen and R.A. van Santen, Micropor. Mesopor. Mater., 56 (2002) 241. [117] A. Corma, S. Iborra, S. Miquel and J. Primo, J. Catal., 173 (1998) 315. [118] A. Corma, S. Iborra, J. Primo and F. Rey, Appl. Catal., 114 (1994) 215. [119] M. de J. Martinez-Ortiz, D. Tichit, P. Gonzales and B. Coq, J. Mol. Catal. A: Chemical, 201 (2003) 199. [120] M.J. Climent, A. Corma, S. Iborra and J. Primo, J. Catal., 151 (1995) 60. [121] F. Figueras, A. Guida, M.H. Lhouty, D. Tichit and P. Geneste, Appl. Catal. A, 164(1997)251. [122] M. Breuninger, US Pat, 5, 932, 748 (1999) to Roche Vitamins Inc. [123] P.S. Kumbhar, J. Sanchez-Valente, J. Lopez and F. Figueras, J. Chem. Soc, Chem. Commun., (1998) 535. [124] P.S. Kumbhar, J. Sanchez-Valente and F. Figueras, J. Chem. Soc, Chem. Commun., (1998)1091. [125] B.M. Choudary, M.L. Kantam, C.R.V. Reddy, K.K. Rao and F. Figueras, J. Mol. Catal. A: Chemical, 146 (1999) 279. [126] V.R. Choudary, S.K. Jana and V.S. Narkhede, Catal. Commun., 2 (2001) 331. [127] A. Dubey, S. Kannan, S. Velu and K. Suzuki, Appl. Catal. A, 238 (2003) 319. [128] S. Ueno, K. Yamaguchi, K. Yoshida, K. Ebitani and Kaneda, J. Chem. Soc. Chem. Commun., (1998) 295. [129] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Luque, J.M. Marinas, FJ. Romero, J.R. Ruiz and FJ. Urbano, Studies in Surface Science and Catalysis, vol.130: Proc. 12* Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 1667. [130] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Luque, J.M. Marinas, J.R. Ruiz and FJ. Urbano, Appl. Catal. A, 216 (2001) 257. [131] J.M. Fraile, J.I. Garcia, D. Marco, J.A.Mayoral, E. Sanchez, A. Monzon and E. Romeo, Studies in Surface Science and Catalysis, vol.130: Proc. 12th Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 1673. [132] J.M. Fraile, J.I. Garcia, D. Marco and J.A. Mayoral, Appl. Catal. A, 207 (2001) 239. [133] K. Kaneda, S. Ueno and T. Imanaka, J. Mol. Catal. A: Chemical, 102 (1995) 135. [134] K. Kaneda and S. Ueno, S., ACS Symposium . Series, vol. 638: Heterogeneous Hydrocarbon Oxidation, Eds. B.K. Warren and S.T. Oyama, S.T., ACS, Washigton DC, USA, 1996, p. 300. [135] K. Kaneda and T. Yamashita, Tetrahedron Lett., 37 (1996) 4555. [136] J.M. Fraile, J.I. Garcia and J.A. Mayoral, Catal. Today, 57 (2000) 3. [137] B.M. Choudary, M. Lakshmi Kantam and P. Lakshmi Santhi, Catal. Today, 57 (2000) 17. [138] V.J. Bulbule, V.H. Deshpande, S. Velu, A. Sudalai, S. Sivasankur and V.T. Sathe, Tetrahedron, 55 (1999) 9325. [139] G.M. Woltermann, US Pat., 4, 454, 244 (1984) to Ashland Oil Inc. [140] M. Chibwe, L. Ukrainczyk, S.A. Boyd and T.J. Pinnavaia, J. Mol. Catal. A, 113 (1996)249. [141] J. Twu and P.K. Dutta, J. Phys. Chem., 93 (1989) 7863. [142] T. Kwon and T.J. Pinnavaia, J. Mol. Catal., 74 (1992) 23.
540
S. Albertazzi, F. Basile and A. Vaccari
[143] J. Wang, Y. Tian, R.C. Wang and A. Clearfield, Chem. Mater., 4 (1992) 1276. [144] T.J. Pinnavaia, M. Chibwe, V.R.L. Constantino and S.K. Yun, Appl Clay Sci., 10 (1995) 117. [145] E. Gardner and T.J. Pinnavaia, Appl. Catal. A, 167 (1998) 65. [146] M.A. Drezdon, US Pat., 4, 774, 212 (1988) to Amoco Co. [147] K. Hashimoto, Y. Matsumura, M. Fukuchi and Y. Kera, Catal. Letters, 19 (1993) 375. [148] J. Don Wang, G. Serrette, Y. Tian and A. Clearfield, Appl. Clay Sci., 10 (1995) 103. [149] F. Malherbe, C. Forano, B. Sharma, M.P. Atkins and J.P. Besse, Appl. Clay Sci., 13(1998)381. [150] R. Dula, K. Wcislo, J. Stoch, B. Grzybowska, E.M. Serwicka, F. Kooli, K. Bahranowski and A. Gawel, Appl. Catal. A, 230 (2002) 281. [151] T. Kwon, G.A. Tsigdinos and T.J. Pinnavaia, J. Am. Chem. Soc, 110 (1988) 3653. [152] E.P. Giannelis, D.G. Nocera and T.J. Pinnavaia, Inorg. Chem., 26 (1987) 203. [153] T. Tatsumi, K. Yamamoto, H. Tajima and H. Tominaga, Chem. Letters, (1992) 815. [154] C. Hu, X. Zhang, L. Xu, B. Mu, W. Zu and E. Wang, Appl. Clay Sci., 13 (1998) 495. [155] P.S. Kumbhar, J. Sanchez-Valente and F. Figueras, Tetrahedron Lett., 39 (1988) 2573. [156] P.S. Kumbhar, J. Sanchez-Valente, J.M.M. Millet and F. Figueras, J. Catal., 191 (2000) 467. [157] F. Medina, D. Tichit, B. Coq, A. Vaccari and N.T. Dung, J. Catal., 167 (1997) 142. [158] F. Medina, R. Dutartre, D. Tichit, B. Coq, N.T. Dung, P. Salagre and J.E. Sueiras J. Mol. Catal. A: Chemical, 119 (1997) 201. [159] N.T. Dung, D. Tichit, B.H. Chiche and B. Coq, Appl. Catal. A, 169 (1998) 179. [160] B. Coq, D. Tichit and S. Ribet, J. Catal., 189 (2000) 117. [161] D. Tichit, R. Durand, A. Rolland, B. Coq, J. Lopez and P. Marion, J. Catal., 211 (2002)511. [162] A. Monzon, E. Romeo, C. Royo, R. Trujillano, F.M. Labajos and V. Rives, Appl. Clay Sci., 13(1998)363.. [163] A. Monzon, E. Romeo, C. Royo, R. Trujillano, F.M. Labajos and V. Rives, Appl. Catal. A, 185(1999)53 [164] A. Monzon, E. Romeo and A.J. Marchi, Layered Double Hydroxides: Present and Future, Ed. V. Rives, Nova Science Publisher, New York, USA, 2001, ch. 11. [165] A. Mastalir and Z. Kiraly, J. Catal., 220 (2003) 372. [166] T.M. Jyothi, T. Raja, K. Sreekumar, M.B. Talawar and B.S. Rao, J. Mol. Catal. A: Chemical, 157(2000)193. [167] T.M. Jyothi, T. Raja and B.S. Rao, J. Mol. Catal. A: Chemical, 168 (2001) 187. [168] M.A. Aramendia, V. Borau, C. Jimenez, J. M. Marinas, J. R. Ruiz and F.J. Urbano, Appl. Catal. A, 206 (2001) 95. [169] M.A. Aramendia, V. Borau, C. Jimenez, J. M. Marinas, J. R. Ruiz and F.J. Urbano, J. Mol. Catal. A: Chemical, 171(2001) 153. [170] M.A. Aramendia, V. Borau, C. Jimenez, J. M. Marinas, J. R. Ruiz and F.J. Urbano, Appl. Catal. A, 249 (2003) 1.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
541
[171] G.L. Castiglioni, M. Ferrari, A. Guercio, A. Vaccari, R. Lancia and C. Fumagalli, Catal. Today, 27 (1996) 181. [172] G.L. Castiglioni, A. Guercio and A. Vaccari, J. Porous Mater., 2 (1995) 79. [173] D.S. Thakur, B.R. Roberts, T.J. Sullivan and A.L. Vichek, US Pat., 5, 155, 086 (1992) to Engelhard Co. [174] S. Narayanan and K. Krishna, Catal. Today, 49 (1999) 57. [175] S. Narayanan and K. Krishna, Appl. Catal. A, 147 (1996) 253. [176] Y.Z. Chen, C. W. Liaw and L.I. Lee, Appl. Catal. A, 177 (1999) 1. [177] S. Velu, M.P. Kapoor, S. Inagaki and K. Suzuki, Appl. Catal. A, 245 (2003) 317. [178] B.H. Cooper and A. Stanislaus, Catal. Rev. Sci. Eng., 36 (1994) 75. [179] B.H. Cooper and B.B.L. Donnis, Appl. Catal. A, 137 (1996) 203. [180] H. Yasuda and Y. Yoshimura, Catal. Letters, 46 (1997) 43. [181H. Yasuda, T. Sato and Y. Yashimura, Catal. Today, 50 (1999) 63. [182] S. Albertazzi, G. Busca, E. Finocchio, R. Glockler and A. Vaccari, J. Catal., submitted. [183] A.E. Palomares, J.G. Prato, F. Marquez and A. Corma, Applied Catalysis B, 41 (2003) 3. [184] A.E. Palomares, J.G. Prato, F. Rey and A. Corma, J. Catal., 221 (2004) 62. [185] A. Morato, C. Alonso, F. Medina, Y. Cesteros, P. Salagre, J.E. Sueiras, D. Tichit and B. Coq, Appl. Catal. B, 32 (2001) 167. [186] A.H. Padmasri, A. Venugopal, J. Krishnamurthy, K.S. Rama Rao and P. Kanta Rao, J. Mol. Catal. A: Chemical, 181 (2002) 73. [187] Y. Cesteros, P. Salagre, F. Medina, J.E. Sueiras, D. Tichit and B. Coq, Appl. Catal. B, 32(2001)25. [188] A. Morato, C. Alonso, F. Medina, Y. Cesteros, P. Salagre, J.E. Sueiras, D. Tichit and B. Coq, Appl. Catal. B, 32 (2001) 167. [189] F.J. Dent, L.A. Moignard, A.H. Eastwood, W.H. Blackburn and D. Hebden, Trans. Inst. Gas Eng., (1945-6) 602. [190] S. Miyata Eur. Pat., 152, 010 (1985) to Kyowa Chem. Ind. [191] D.W. Oscarson, H.G. Miller and R.L. Watson, An evaluation of potential additive to a clay-based buffer material for the immobilization of 1-129, Atomic Energy of Canada, Pinawa (C), 1986. [192] J. Serrano, V. Bertin and S. Bulbulian, J. Radioanal. Nuclear Chem., 254 (2002) 91. [193] M.A. Ulibarri, I. Pavlovic, M.C. Hermosin and J. Cornejo, Appl. Clay Sci., 10 (1995)131. [194] M.A. Ulibarri, I. Pavlovic, C. Barriga, M.C. Hermosin and J. Cornejo, Appl. Clay Sci., 18(2001)17. [195] N.K. Lazaridis, T.D. Karapantsios and D. Georgantas, Water Res., 37 (2003) 3023. [196] S. Komarneni, N. Kazai and R. Roy, Mater. Commun., 8 (1998) 1329. [197] P. Ulmgren, Nordic Pulp and Paper Research J., 1 (1987) 708. [198] T. Kameda, T. Yoshioka and A. Okuwaki, Development of new synthetic process for hydrotalcite using seawater and calcined dolomite and its application to wastewater treatment, Kagaku Kogyo, 52 (2001) 810 (in Japanese). [199] M.E. Perez-Bernal, R. Ruano-Casero and T.J. Pinnavaia; Catal. Letters, 11 (1991) 55. [200] M. Chibwe and T.J. Pinnavaia, J. Chem. Soc, Chem. Commun., (1993) 278.
542
S. Albertazzi, F. Basile and A. Vaccari
[201] T.J. Pinnavaia, M.E. Perez-Bernal, R. Ruano-Caserio and M. Chibwe, US Pat, 5, 302, 709 (1994) to Michigan State University. [202] I. Chatti, A. Ghorbel, P. Grange and J.M. Colin, Catal. Today, 75 (2002) 113. [203] V.I. Iliev, A.I. Ileva and L.D. Dimitrov, Appl. Catal. A, 126 (1995) 333. [204] A. Cervilla, A. Corma, V. Fornes, E. Llopis, P. Palanca, F. Rey and A. Ribera, J. Am. Chem. Soc, 116 (1994) 1595. [205] A. Cervilla, E. Llopis, A. Ribera, A. Corma, V. Fornes and F. Rey, J. Chem. Soc. Dalton Trans., 20 (1994) 2953. [206] A. Alejandre, F. Medina, X. Rodriguez, P. Salagre, Y. Cesteros and J.E. Sueiras, Appl. Catal. B, 30 (2001) 195. [207] Y. Guo, D. Li, C. Hu, Y. Wang, E. Wang, Y. Zhou and S. Feng, Appl. Catal. B, 30 (2001) 337. [208] M.A. Ulibarri and M.C. Hermosin, Layered Double Hydroxides: Present and Future, Ed. V. Rives, Nova Science Publisher, New York, USA, 2001, ch. 9. [209] J.S. Yoo, A.A. Bhattacharyya, C.A. Rodlowski and J.A. Karch, Appl. Catal. B, 1 (1992) 169. [210] A. Corma, A.E. Palomares and F. Rey, Appl. Catal. B, 4 (1994) 29. [211] T. Myrstad, H. Engan, B. Seljestokken and E. Rytter, Appl. Catal. A, 187 (1999) 207. [212] A. Corma, A.E. Palomares, F. Rey and F.; Marquez, J. Catal., 170 (1997) 140. [213] A.E. Palomares, J.M. Lopez-Nieto, FJ. Lazaro, A. Lopez and A. Corma, Appl. Catal. B, 20 (1999) 257. [214] T. Lojewski, R. Dziembaj, R. Chyzy, A. Vaccari and B. Montanari, Polish J. Environm. Studies, 6 (1997) 71. [215] B. Montanari, A. Vaccari, M. Gazzano, P. Ka(3ner, H. Papp, J. Pasel, R. Dziembaj, W. Makowski and T. Lojewski, Appl. Catal. B, 13 (1997) 205. [216] M. Trombetta, G. Ramis, G. Busca, B. Montanari and A. Vaccari, Langmuir, 13 (1997)4628. [217] R.S. Mulukula and C. Detellier, J. Mater. Science Lett., 15 (1996) 797. [218] I.J. Shannon, F. Rey, G. Sankar, J.M. Thomas, T. Maschmeyer, A.M. Waller, A.E. Palomares, A. Corma, A.J. Dent and G.N. Greaves, J. Chem. Soc, Faraday Trans., 92 (1996) 4331. [219] K. Bahranowski, E. Bielanska, J. Janas, T. Machej, L. Matachowski and E.M. Serwicka, Polish J. Environm. Studies, 6 (1997) 59. [220] K. Bahranowski, R. Dula, A. Gavel, J. Janas, T. Machej and E.M. Serwicka, Polish J. Environm. Studies, 6 (1997) 65. [221] J. Haber, K. Bahranowski, J. Janas, R. Janik, T. Machej, L. Matachowski, A. Michalik, H. Sadowska and E.M. Serwicka, Studies in Surface Science and Catalysis, vol. 130: Proc. 12th Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz, J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 599. [222] G. Centi, A. Galli, B. Montanari, S. Perathoner and A. Vaccari, Catal. Today, 35 (1997)113. [223] S. Kannan and C.S. Swamy, Appl. Catal. B, 3 (1994) 109. [224] S. Kannan and C.S.; Swamy, Catal. Today, 53 (1999) 725. [225] J.N. Armor, T.A. Braymer, T.S. Farris, Y. Li, F.P. Petrocelli, E.L. Weist, S. Kannan and C.S. Swamy, Appl. Catal. B, 7 (1996) 397. [226] S. Kannan, Appl. Clay Sci., 13 (1998) 347. [227] H. Dandl and G. Emig, Appl. Catal. A, 168 (1998) 261.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
543
[228] J. Perez-Ramirez, F. Kapteijn and J.A. Moulijn, Catal. Letters, 60 (1999) 133. [229] R.D. Shannon, Acta Crystallogr. A, 32 (1976) 751. [230] X. Zhang, A.B. Walters and M.A. Vannice, Appl. Catal. B, 4 (1994) 237. [231] F. Kapteijn, J. Rodriguez-Mirasol and J.A. Moulijn, Appl. Catal. B, 9 (1996) 25. [232] J. Perez-Ramirez, J. Overeijnder, F. Kapteijn and J.A. Moulijn, Appl. Catal. B, 23 (1999) 59. [233] J.Perez-Ramirez, J.M. Garcia-Cortes, F. Kapteijn, M.J. Man-Gomez, A. Ribera, C. Salinas-Martinez de Lecea and J.A. Moulijn, Appl. Catal. B, 25 (2000) 191. [234] M.C. Roman-Martinez, F. Kapteijn, D. Cazorla-Amoros, J. Perez-Ramirez and J.A.Moulijn, Appl. Catal. A, 225 (2002) 87. [235] S. Alini, A. Bologna, F. Basile, T. Montanari and A. Vaccari, Eur. Pat. Appl. 1, 262, 224 (2001) to Radici Chimica SpA. [236] S. Alini, F. Basile, A. Bologna, T. Montanari and A. Vaccari, Studies in Surface Science and Catalysis, vol. 143: Scientific Bases for the Preparation of Hetrogeneous Catalysts, Eds. E. Gaigneaux, D.E. De Vos, P. Grange, P.A. Jacobs, J.A. Martens, P. Ruiz and G. Poncelet, Elsevier, Amsterdam, NL, 2002, p. 131. [237] D.L. King, J.A. Cusumano and R..L. Garten, Catal. Rev. Sci. Eng., 23 (1981) 233. [238] M.J. Howard, M.D. Jones, M.S. Roberts and A.T. Taylor, Catal Today, 18 (1993) 325. [239] J.J. Spivey, S.K. Gangwal, J.R. Zoeller, J.C. Winslow and R.D. Srivastava (Eds.), Catal. Today, 58(2000)231. [240] M.J. Brown andN.D. Parkins, Catal. Today, 8 (1991) 305. [241] J.M. Fox, Catal. Rev. Sci. Eng., 35 (1993) 169. [242] J.R. Rostrup-Nielsen, Catalysis Science and Technology, vol. 5, Eds. J.R. Anderson and M. Boudart, M., Springer-Verlag, Berlin (D), 1984, p. 1. [243] A.Solbakken, Studies in Surface Science and Catalysis, vol. 61: Natural Gas Conversion, Eds. A. Holmen, K.-J. Jens and S. Kolboe, Elsevier, NL, 1991, p. 447. [244] D.E. Ridler and M.W. Twigg, Catalysis Handbook, 2nd ed., Ed. M.W. Twigg, Wolfe, London, UK, 1989, p. 225. [245] J.R. Rostrup-Nielsen, Catal. Today, 18 (1993) 305. [246] J.N. Armor, Appl. Catal. A, 176 (1999) 159. [247] J.R.H. Ross, M.C.F. Stel and A. Zeini-Isfahani, J. Catal., 52 (1978) 280. [248] J.H. Edwards and A.M. Matra, A.M., Studies in Surface Science and Catalysis, vol. 81: Natural Gas Conversion II, Eds. H.E. Curry-Hyde and R.F. Howe, Elsevier, NL, 1994, p. 291 [249] E. Ruckestein and Y.H. Hu, J. Catal., 162 (1996) 230. [250] A. Ashcroft, A.K. Cheetham, M.H.L. Green and P.D. Vernon, Nature, 352 (1991) 225. [251] V.R. Choudhary, A.S. Mamman and S.D. Sansare, Angew. Chem. Int. Ed. Eng., 31 (1992) 1189. [252] V.R. Choudhary, A.M. Rajput and B. Prabhakar, J. Catal., 139 (1993) 326. [253] J.R. Rostrup-Nielsen and J.-H. Bak Hansen, J. Catal., 144 (1993) 38. [254] R. Blom, I.M. Dahl, A. Slagtern, B. Sortland, A. Spjelkavik and E. Tangstad, Catal. Today, 21 (1994) 535. [255] Y.H. Hu and E. Ruckenstein, Catal. Letters, 36 (1996) 145. [256] D.A. Hickman and L.D. Schmidt, J. Catal., 138 (1992) 267.
544
S. Albertazzi, F. Basile and A. Vaccari
[257] D.A. Hickman, E.A. Haupfear and L.D. Schmidt, Catal. Letters, 17 (1993) 223. [258] G.A. Foulds and J.A. Lapszewicz, Catalysis, vol. 11, Eds. J.J. Spivey and S.K. Agarwal, Royal Society of Chemistry, London, UK, 1994, p. 413. [259] O. Clause, F. Trifiro, A. Vaccari and L. Zatorski, Appl. Catal., 73 (1991) 217. [260] A. Bhattacharyya and M.P. Kaminski, US Pat., 5, 354, 932 (1994) to AMOCO Co. [261] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A. Udovich, PCT WO 14, 700 (1994) to AMOCO Co. [262] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A: Udovich, US Pat., 5, 399, 537 (1995) to AMOCO Co. [263] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A. Udovich, US Pat., 5, 591, 238 (1997) to AMOCO Co. [264] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A. Udovich, US Pat., 5, 653, 774 (1997) to AMOCO Co. [265] S.B.J. Scott, PCT WO 37,390 (1997) to BG PLC. [266] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A. Udovich, US Pat., 5, 767, 040 (1998) to AMOCO Co. [267] O. Clause, B. Rebours, E. Merlin, F. Trifiro and A. Vaccari, J. Catal., 133 (1992) 231. [268] A. Bhattacharyya, W.D. Chang, M.S. Kleefisch and C.A. Udovich, PCT WO 15, 459 (1999) to AMOCO Co. [269] F. Basile, L. Basini, G. Fornasari, D. Matteuzzi, F. Trifiro and A. Vaccari, Actas XV Simp. Iberoam. Catalisis, vol. 3, Eds. E. Herrero, O. Annunziata and C. Perez, Univ. Nac. Cordoba, Cordoba, Argentina, 1996, p. 1843. [270] F. Basile, L. Basini, M. D' Amore, G. Fornasari, A. Guarinoni, D. Matteuzzi, G. Del Piero, F. Trifiro and A. Vaccari, J. Catal., 173 (1998) 247. [271] F. Basile, L. Basini, G. Fornasari, M. Gazzano, F. Trifiro and A. Vaccari, Studies Surface Science and Catalysis, vol. 118: Preparation of Catalysts VII, Eds. B. Delmon,P.A. Jacobs, R. Maggi, J.A. Martens, P. Grange and G. Poncelet, Elsevier, Amsterdam, NL, 1998, p. 31. [272] F. Basile, G. Fornasari, M. Gazzano and A. Vaccari, J. Mater. Chem., 12 (2002) 3296. [273] F. Basile, G. Fornasari, M. Gazzano, A. Kiennemann and A. Vaccari, J. Catal., 217(2003)245. [274] F. Basile, G. Fornasari, M. Gazzano and A. Vaccari, Appl. Clay Sci., 18 (2001) 51. [275] F. Basile, L. Basini, G. Fornasari, A. Guarinoni, F. Trifiro and A. Vaccari, Studies in Surface Science and Catalysis, vol. 119: Natural Gas Conversion V, Eds. A. Parmalina, D. Sanfilippo, F. Frusteri, A. Vaccari and F. Arena, Elsevier, Amsterdam, NL, 1998, p. 693. [276] F. Basile, G. Fornasari, F. Trifiro and A. Vaccari, Studies in Surface Science and Catalysis, vol. 130: Proc. 12* Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 449. [277] F. Basile, G. Fornasari, F. Trifiro and A. Vaccari, Catal. Today, 64 (2001) 21. [278] F. Basile, G. Fornasari, F. Trifiro and A. Vaccari, Catal. Today, 77 (2002) 215. [279] F. Basile, G. Fornasari, S. Albertazzi, A. Vaccari, D. Gary and P. Arpentinier, Eur. Pat. Appl. 2,291,304 (2002) to Air Liquide S.A.
Catalytic Properties of Hydrotalcite-Type Anionic Clays
545
[280] F. Basile, S. Albertazzi, P. Arpentinier, P. Del Gallo, G. Fornasari, D. Gary, V. Rosetti and A. Vaccari, Studies in Surface Science and Catalysis, Natural Gas Conversion 7, in press. [281] A. Bhattacharyya, W.D. Chang and D.J. Schumacher, Appl. Clay Sci., 13 (1998) 317. [282] F. Basile, G. Fornasari, E. Poluzzi and A. Vaccari, Appl. Clay Sci., 13 (1998) 329. [283] A.I. Tsyganok, K. Suzuki, S. Hamakawa, K. Takehira and T. Hayakawa, Catal. Letters, 77(2001)75. [284] BASF AG, UK Pat., 1, 462, 059-60 (1973). [285] J.H.R. Ross, Catalysis, Specialist Periodical Reports, vol. 7, Eds. G.C. Bond and G. Webb, Royal Society of Chemistry, London, UK, 1987, p. 1. [286] E. Errani, F. Trifiro, A. Vaccari, M. Richter and G. Del Piero, Catal Letters, 3 (1989) 65. [287] E.B.M. Doesburg, R.H. Hoppener, B. de Koning, X. Xiading and J.J.F. Scholten, Studies in Surface Science and Catalysis, vol. 31: Preparation of Catalyst IV, Eds. B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet, Elsevier, Amsterdam, NL, 1987, p. 767. [288] S. Gusi, F. Trifiro, A. Vaccari and G. Del Piero, J. Catal., 94 (1985) 120. [289] K. Richter, W. Kraus, G. Nolze and B. Peplinski, Studies in Surface Science and Catalysis, vol. 75: New Frontiers in Catalysis, Eds. L. Guczi, F. Solymosi and P. Tetenyi, P., Elsevier, Amsterdam, NL, 1993, p. 2805. [290] R.G. Herman, C.E. Bogdan, P.L. Kumler and D.M. Nuszkoski, Mater. Chem. Phys., 35(1993)233. [291] J.L. Li and T. Inui, Appl. Catal. A, 139 (1996) 87. [292] S. Gusi, F. Pizzoli, F. Trifiro, A. Vaccari and G. Del Piero, Studies in Surface Science and Catalysis, vol. 31: Preparation of Catalyst IV, Eds. B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet, Elsevier, Amsterdam, NL, 1987, p. 753. [293] G. Fornasari, A. D' Huysser, L. Mintchev, F. Trifiro and A. Vaccari, J. Catal., 135 (1992)386. [294] G. Fornasari, S. Gusi, F. Trifiro and A. Vaccari, I.&E.C. Research, 26 (1987) 1500. [295] G. Ravichandran and D.K. Chakrabarty, Catal. Letters, 22 (1993) 179. [296] G. Fornasari, S. Gusi, T.M.G. La Toretta, F. Trifiro and A. Vaccari, Studies in Surface Science and Catalysis, vol. 30: Catalysis and Automotive Pollution Control, Eds. A. Crucq and A. Frennet, Elsevier, Amsterdam, NL, 1987, p. 469. [297] K.J. Smith and R.B. Anderson, J. Catal., 85 (1984) 428. [298] Ph. Courty, D. Durand, E. Freund and A. Sugier, J. Mol. Catal., 17 (1982) 241. [299] Y. Borodko and G.A. Samorjai, Appl. Catal. A, 186 (1999) 355. [300] M.A. Pena, J.P. Gomez and J.L.G. Fierro, Appl. Catal. A, 144 (1996) 7. [301] J.L.G. Fierro, Studies in Surface Science and Catalysis, vol. 130: Proc. 12* Intern. Congress Catalysis, Eds. A:Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, p. 177. [302] B.A. Peppley, J.C. Amphlett, L.M. Kearns and R.F. Mann, Appl. Catal. A, 179 (1999)21. [303] K. Sekizawa, S. Yano, K. Eguchi and H Arai, Appl. Catal. A, 167 (1998) 291. [304] R. Shiozaki, T. Hayakawa, Y.-Y. Liu, M. Kumagai, S. Hamakawa, K. Suzuki, T.
546
S. Albertazzi, F. Basile and A. Vaccari
Itoh, T. Shishido and K. Takehira, Catal. Letters, 58 (1999) 131. [305] Y.-Y. Liu, K. Suzuki, S. Hamakawa, T. Hayakawa, K. Murata, T. Ishii and M. Kumagai, Chem. Letters, (2000) 486. [306] S.R. Segal, K.A. Carrado, C.L. Marshall and K.B. Anderson, Appl. Catal., in press. [307] L.M. Kearns, J.C. Amphlett, E. Haliop, H.M. Jenkins, R.F. Mann and B.A. Peppley, Prepr. Symp. Am. Chem. Soc, Div. Fuel Chem., 44 (1999) 905. [308] S. Velu, K. Suzuky and T. Osaky, Catal. Letters, 62 (1999) 159. [309] S.R. Segal, K.B. Anderson, K.A. Carrado and C.L. Marshall, Appl. Catal. A, 231 (2002)215. [310] S. Velu, K. Suzuky and T. Osaky, J. Chem. Soc, Chem. Commun., (1999) 2341. [311] S. Velu, K. Suzuki, M.P. Kapoor, F. Ohashi and T. Osaki, Appl. Catal. A, 213 (2001)47. [312] A. Bhattacharyya and A. Basu, US Pat, 5, 626, 794 (1997) to AMOCO Co. [313] H.J. Kohnke, German Pat, 19, 534, 434 (1995) to Daimler-Benz AG. [314] A. Bhattacharyya, PCT WO 03,779 (1999) to AMOCO Co. [315] K.D. Chung, M.S. Lee, O.S. Joo, S.H. Han and S.J. Uhm, Hydrogen Power: Theor. Eng. Solutions, Proc. Hyothesys Symp., Ed T.O. Saetre, Kluwer, Dordercht, NL, 1998, p. 43. [316] D. Tichit and A. Vaccari, Appl. Clay Sci., 13 (1998) 311. [317] S. Bonnet, C. Forano, A. de Roy, J.P. Besse, P. Maillard and M. Momenteau, Chem. Mater., 8 (1996) 1962. [318] L. Barloy, J.P. Laliter, P. Battioni, D. Mansuy, Y. Piffard, M. Tournoux, J.B. Valim and W. Jones, New J. Chem., 16 (1992) 71. [319] B.F. Sels, D.E. De Vos, M. Buntinx, F. Pierard, A. Kirsch-De Mesmaeker and P.A. Jacobs, Nature, 400 (1999) 855. [320] D.E. De Vos, B.F. Sels, W.M. Van Rhijn and P.A. Jacobs, Studies in Surface Science and Catalysis, vol. 130: Proc. 12th Intern Congress Catalysis, Eds. A. Corma, F.V. Mejo, S. Mendioroz and J.L.G. Fierro, Elsevier, Amsterdam, NL, 2000, 137. [321] B.F. Sels, D.E. De Vos D.E. and P.A. Jacobs, Tetrahedron Lett., 37 (1996) 8557. [322] P.D. Kaviratna and TJ. Pinnavaia, J. Electroanal. Chem., 385 (1995) 163. [323] S. Therias and C. Mousty, Appl. Clay Sci., 10 (1995) 147. [324] B. Ballarin, M. Gazzano, R. Seeber, D. Tonelli and A. Vaccari, J. Electroanal. Chem., 445 (1998) 27. [325] S. Therias, B. Lacroix, B. Schollhorn, C. Mousty and P. Palvadeau, J. Electroanal. Chem., 454 (1998) 91. [326] K. Yao, M. Taniguchi, M. Nakata, K. Shimazu, M. Takahashi and A. Yamagishi, J. Electroanal. Chem., 457 (1998) 119. [327] B. Ballarin, R. Seeber, D. Tonelli and A. Vaccari, J. Electroanal. Chem., 463 (1999) 123. [328] B. Ballarin, M. Morigi, E. Scavetta, R. Seeber and D. Tonelli, J. Electroanal. Chem., 492 (2000) 7.
Index A Absorption - 480 Acid activation - 327 Acid-activated organoclays - 335 Acid and base properties - 428 Acid and redox catalysis - 388 Acid-base interactions - 92 Acidity - 327 Active surface sites - 75 Active substances - 277 Activity coefficients - 170 Adsorption/desorption - 126, 153, 236, 321, 333, 430, 434,435, 436, 437, 440,482 Adsorption calorimetry - 132 Adsorption enthalpy - 101, 130, 141, 144 Adsorption isotherms - 128, 132, 258 Adsorption kinetics - 439 Agricultural applications - 410, 421 Aldolic condensation - 503 Alkylation-391 Alkylammonium cations - 257 Amine cations - 104,257, 331 Animal feed-410 Anion exchange capacity (AEC) - 427,428 Anion exchange process - 430 Anionic clays - 403, 411 Antacid - 279 Antidiarrhoeal - 278 Anti-inflammatory - 279 Antiseptic - 279 Arrhenius plot - 531 B Bayer-Villinger oxidation - 506 BET (Brunauer-Emmett-Teller) - 305,393, 396,399,482 Bio-LDH hybrid - 416, 488 Biological applications - 403,480 Biomolecule - 416, 487, 488 Biopolymer - 486 Broken edges - 323 Bromination - 534 Brucite structure - 2, 121, 412
548
Index
C Calcination - 442, 499, 503, 520 Carbonaceous replica - 481 Catalysts supports - 498, 499 Catalytic properties of clays - 321,443, 444 Catalytic properties of LDH-POM - 388, 390 Cation exchange capacity (CEC) - 76, 156, 324 Cation exchange reaction - 48, 175 Cationic clays - 403, 404 Cation migration - 325 Cement related materials - 483 Chemical modification - 1 Chemical reactivity - 1, 138 Chlorite structure- 123,140, 141 Clay structure - 4, 5, 6, 119, 120, 121, 122, 123, 186,253,255,269,322,323,326,328, 330,405 Claysen condensation - 503 Claysen cyclization — 503 Colloidal properties - 462 Copper(II) hydroxide acetate micrograph -18 Copper hydroxide nitrate structure - 413 Contact angle - 93, 94, 131, 143 Co-precipitation - 364, 386 Cosmetics-267, 281, 409, 418 CPO (catalytic partial oxidation) - 523, 524, 525 CTO (catalytic total oxidation) - 525 D Dawson polyoxometalates - 383 Debye-Huckel law - 170 Debye length- 170 Decarboxylation — 503 Deintercalation -206,209,210,211,213 Differential Scanning Calorymetry (DSC) - 10, 11, 16, 22, 31, 33, 36, 45, 361 Diffuse layer - 59, 163,170 Diffuse-Reflectance Infrared Fourier Transform (DRIFT) Spectroscopy - 187, 189, 191, 192, 196, 200, 203, 204,207,210, 213 DKR (Dubinin-Kaganer-Radushkevich) Equation - 112 DLVO (Derjaguin, Landau, Verwey and Overbeek) Theory - 82, 85, 98 Double layer - 172 DR (Dubinin-Radushkevich) Equation - 103 Drug delivery system — 274, 280 DTA (Differential Thermal Analysis) - 351, 356 Dye-179,438 E Electrokinetic of clay surfaces - 57 Electrokinetic properties - 58, 62 Electronegativity - 1 3 9
Index
549
Endotactic reaction - 469,471 Enthalpy- 145 Entropy-101, 145 Environmental - 426, 443, 514 EPR (Electron Paramagnetic Resonance) - 473 Esterification - 389, 390 EXAFS (Extended X-ray Absorption Fine Structure) - 479 Exchange reactions - 8, 20, 175, 386, 435 Excipients - 273,275 Exfoliation reaction - 19, 29, 49, 462 F Flame retardant - 486 Freundlich-258 FTIR (Fourier Transform Infrared) - 10, 11, 12, 17, 24,27, 31, 37, 38,46, 352, 368 G Gastrointestinal protector - 278 GC (Gouy-Chapman) theory - 155 GCS (Gouy-Chapman-Stern) model - 156 Gibbsite structure - 121 Grafting by displacement - 41 Grafting reactions - 2, 8, 14,20, 39,49 Greenhouse gases - 434,444,445,446,447, 517, 518, 520 Green rusts - 4, 461 Guest displacement - 465 H Hardness- 139 Henry reaction - 507 Herbicide - 248 Heteropolyoxometalate - 376, Homogeneous catalysis - 520 HT-FTIR (High temperature FTIR) - 368 HTLc (Hydrotalcite-like) - 426,496 HTXRD (High temperature X-ray Diffraction) - 348, 356, 365 Humic acid - 254, 440 Hybrid materials - 463, 480 Hydrated kaolinite - 42, 43, 45, 46 Hydrocalumite structure - 412 Hydrocalumite micrograph - 477 Hydrogenation - 509, 510, 512, 513 Hydrogenolysis - 509 Hydrogen bonding - 333 Hydrophobic/hydrophilic - 97 Hydrotalcite structure - 412 Hydroxysalts - 3, 13, 413 I Impregnation — 500
550
Index
Infrared spectroscopy - 134, 184 Infrared thermography - 526 Inorganic polymer - 471 In-situ polymerization - 464, 465 Intercalation by displacement - 34 Intercalation of organic anions - 357,364, 368 Interlayer or surface grafting - 8, 14, 20, 39, 49 Interlayer inorganic cations - 229 Intercalated molecules - 34, 42, 236 Intercalated organic cations - 232 Ion exchange reaction - 13, 329, 332 Isomerization of cc-pinene - 336 Isomorphous substitution - 61, 323 Isopolyoxometalate - 376 K Kaolinite-6, 32, 33, 75, 119, 120, 122, 186, 255, 270 Kaolinite intercalation - 32, 34, 42, 192, 195, 202, 206, 213, 237, 239 Kaolinite micrograph- 185, 270 Kaolinite structure - 6, 75, 119, 120, 122, 186, 255 Keggin compounds - 375 Keggin type heteropolyoxometalates - 376 Knoevenagel condensation - 503 L Layer charge density - 62, 76, 323, 461 LDH (Layered Double Hydroxide) - 4, 20, 345, 357, 374, 380, 382, 411, 414, 416, 426, 459, 460, 465 Layered double hydroxide micrograph - 429, 477, 478, 479, 482, 484, 487 Layered double hydroxide structure - 4, 380, 411, 412, 414, 426, 433, 439, 465, 473 Lithiophorite structure - 20 LW (Lifshitz-van der Waals) forces - 92 M MAS (Magic angle spinning) - 229, 351, 357, 366, 475 Mechanical properties - 483 Mechanochemical activation - 188, 191, 192,202 Mechanochemical reactions - 42 Meerwein-Ponndor-Verley reduction - 505 Memory effect — 464 Metal nanoparticles - 26 Methanol decomposition - 531 Methylation - 504 Michael addition - 503 Mixed oxides - 499, 500, 520, 521 Montmorillonite - 5, 97, 122,175, 176, 303, 327 Montmorillonite structure - 5, 122, 327 N Nanocomposite - 19, 29, 48, 49, 409, 459, 468, 480, 484, 487
Index Nanofillers - 48, 483, 484 Natural gas exploitation - 522 NMR (Nuclear Magnetic Resonance) - 216, 351, 357, 366,475 Noble metals supported catalysts - 500 Nuclear isotopes - 218 Nuclear spin interactions - 217 Nuclear wastes - 433 O Organic synthesis - 501 Organic-anion-pillared — 384 Organic pollutants - 435 Organoclay - 257, 331, 332, 440 OSR (Oxidative-steam-reforming) - 533 Oxidation - 393, 503 Oxidation-reduction reactions — 12 Oxometallates - 431 P Palygorskite micrograph - 270 Particle size - 124 PDI (Potential Determining Ions) - 60 Pesticide - 247, 249, 250, 251, 256, 259, 410, 436, 437 Pharmaceutics - 267, 271, 407, 414 Pharmaceutical denomination - 271 Pharmaceutical specification - 271 Pharmacokinetic parameters - 408 Phase-transfer method - 386 Photocatalysis - 395, 396, 398, 448, 516 PILCs (Pillared clays) - 49, 293, 306, 309, 329, 330, 440, 509 Pillared anionic clays - 388, 440, 509, 516 Polymer - 19, 29, 240, 242, 459, 468 Polymer direct incorporation - 465 POM (Polyoxometalate) - 374, 382, 385, 388, 393, 396 Porosity-430 Potentiometric titration - 74 Preyssler compounds - 377,378 PZC (Point of Zero Charge) - 69 Q Quadrupolar splitting — 220 Quaternary amine cations - 104 R Raman Spectroscopy - 184, 194, 195, 197, 198, 209, 211 Reconstruction method - 385, 442, 465 Reduction-505, 517 Reduction of oxides - 300 Reductive intercalative polymerization - 464 Reforming of methane - 527 Rehydration method - 357
551
552
Index
Relaxation - 219, 224 Remediation - 430, 435, 443 Removal of dyes - 438 Removal of heavy metals -290,292, 431,433 Removal of nuclear wastes - 433 Removal of oxoanions - 431 Removal of oxometallates - 431 Removal of pesticides - 436 Restacking- 19,465 S SAED (Selected area electron diffraction) - 28 SEDOR (Spin-echo double resonance) - 230 SEM (Scanning electron microscopy) - 18, 31, 33, 185, 186, 270, 419, 429, 477, 478, 487 Self diffusion-219, 223 Sensors - 480 Sepiolite structure - 79 Siloxane sheet-5, 121 Smectite structure - 75, 253, 255, 292, 323, 326, 328, 330, 334, 405 Smectite micrograph - 270 Solvation reaction - 1 2 Solvent assisted method - 465 Solid dosage - 273 Solvent exchange reaction - 13 Sorbent-157,440 Sorption process - 250 Sorption site - 253 SR (Steam-reforming) - 530, 532 SSA (Specific surface area) - 156, 393 Stern layer- 160 Surface acidity - 68, 327 Surface adsorption - 7, 20, 101, 108, 126, 128, 130, 132, 141, 144, 157 Surface area- 124 Surface charge - 60 Surface energy - 125, 136,143 Surface exchange reaction - 8 Surface potential - 60, 167 Surface properties - 430 Surface sites - 1 2 3 Surface tension - 91, 94 Surface thermodynamics - 90 Surfactants - 30, 108, 233,386, 441 Swelling - 146, 326 Swelling agents - 385 Swelling pressure - 59 T Talc micrograph - 270 Talc structure -123
Index
553
TEM (Transmission Electron Microscopy - 28, 482, 484 Textural properties - 476 Thermal reactions - 26, 48, 306, 309, 310, 311,499 Thermodynamics-91, 125 Thermogravimetry (TG) - 10, 11, 16, 22, 31, 33, 36, 45, 351, 356, 361 Topical use - 279, 282 TSDC (Thermal Simulated Depolarization Currents) - 134 Two-dimensional multiple quantum magic angle spinning (2D MAS) - 230 U UV-VIS (Ultraviolet-visible spectrophotometry) - 298, 300, 302 V Van der Waals forces - 333 Vibrational spectroscopy - 187 Vitabrid-C powder - 419, 420 W Wastewater-290, 292, 312,431, 433 Water adsorption - 118 Wettability-130 X XAS (X-ray Absorption) - 470 XANES (X-ray Absorption Near Edge Structure) - 479 XRPD (X-ray Powder Diffraction) - 9, 15, 21, 28, 30, 33, 35, 40, 41, 42, 43, 206, 349, 358, 359, 370, 381, 415, 419, 439, 470 Z Zeta potential - 58, 64, 67, 70, 71, 77, 78, 80 Zinc hydroxide nitrate structure - 3,413
This page is intentionally left blank