Miniaturization of Analytical Systems: Principles, Designs and Applications

Miniaturization of Analytical Systems: Principles, Designs and Applications ANGEL RIOS University of Castilla-La Mancha, Ciudad Real, Spain ALBERTO ESCARPA University of Alcal a, Madrid, Spain BARTOLOME SIMONET University of C ordoba, C ordoba, Spain

Miniaturization of Analytical Systems: Principles, Designs and Applications

Miniaturization of Analytical Systems: Principles, Designs and Applications ANGEL RIOS University of Castilla-La Mancha, Ciudad Real, Spain ALBERTO ESCARPA University of Alcal a, Madrid, Spain BARTOLOME SIMONET University of C ordoba, C ordoba, Spain

This edition first published 2009 Ó 2009 John Wiley & Sons, Ltd Registered office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com. The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for every situation. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom. The right hand circular image on the cover design is reproduced from The Essential Guide to Analytical Chemistry, Georg Schwedt, John Wiley & Sons, Ltd, 1997. Original German, Georg Thieme Verlag, 1996. Library of Congress Cataloging-in-Publication Data Rios Castro, Angel. Miniaturization of analytical systems : principles, designs and applications / Angel Rios, Alberto Escarpa, Bartolome Simonet. p. cm. Includes bibliographical references and index. ISBN 978-0-470-06110-7 1. Chemistry, Analytic. 2. Miniature electronic equipment. I. Escarpa, Alberto. II. Simonet, Bartolome. III. Title. QD75.3.R56 2009 543–dc22 2009016260 A catalogue record for this book is available from the British Library. ISBN 978-0-470-06110-7 Set in 11/13pt Times by Thomson Digital, Noida, India. Printed and bound in Great Britain by CPI Antony Rowe Ltd, Chippenham, Wiltshire.

Contents Preface

xi

1 Miniaturization in Analytical Chemistry 1.1 Introduction 1.2 Miniaturization as One of the Critical Trends in Modern Analytical Chemistry 1.3 Evolution in the Field of Analytical Miniaturization 1.4 Classification of Miniaturized Analytical Systems and Definition of Terms 1.5 Theory of Miniaturization 1.6 Features of Miniaturized Analytical Systems 1.7 Incidences of Miniaturization in the Analytical Process 1.7.1 Miniaturization of the Steps of the Analytical Process 1.7.2 Integrated (Micro)systems for the Performance of the Entire Analytical Process 1.8 Outlook References

1 1

2 Tools for the Design of Miniaturized Analytical Systems 2.1 Introduction 2.2 Miniaturized Analytical Processes: The Downsizing and Integrating Phenomena 2.2.1 Transport within Microfluidic Systems 2.2.2 Microsystem Design from Transport Parameter Information 2.3 Microfluidic Devices 2.3.1 Microvalves 2.3.2 Moving Liquids in Miniaturized Systems 2.3.3 Mixers 2.3.4 Volume-dispensing and Sample-introduction Devices 2.3.5 Detection Systems for Analytical Microsystems 2.4 Microtechnology 2.4.1 Computer Simulations in Microfluidics 2.4.2 Micromachining 2.4.3 Packaging of Microsystems

2 8 10 15 17 19 20 30 33 36 39 39 40 42 45 45 45 51 59 63 64 66 66 68 74

vi

Contents

2.5 MEMS and NEMS 2.5.1 Fabrication and Characterization 2.5.2 Functionalization 2.5.3 Detection Methods 2.6 Outlook References 3 Automation and Miniaturization of Sample Treatment 3.1 Introduction 3.2 Simplification of Sample Treatment: Microextraction Techniques 3.2.1 Calibration in Microextraction Processes 3.2.2 Solid Phase Microextraction (SPME) Techniques 3.2.3 Liquid Phase Microextraction (LPME) Techniques 3.2.4 Comparison of Solid and Liquid Phase Microextraction Techniques 3.3 Simplification of Sample Treatment: Continuous Flow Systems 3.3.1 Coupling to Gas Chromatography 3.3.2 Coupling to Liquid Chromatography 3.3.3 Coupling to Capillary Electrophoresis (CE) References

74 75 76 81 84 88 93 93 94 94 95 114 117 119 120 123 127 135

4 Miniaturized Systems for Analytical Separations I: Systems Based on a Hydrodynamic Flow 4.1 Introduction 4.2 The Earliest Example of Miniaturization of a Gas Chromatograph and Some Other Developments 4.3 Capillary Liquid Chromatography (CLC) 4.4 Liquid Chromatography on Microchips 4.4.1 The Agilent HPLC Chip 4.4.2 Other Approaches to Microchip HPLC 4.4.3 Some Selected Applications References

141 144 150 150 155 157 163

5 Miniaturized Systems for Analytical Separations II: Systems Based on Electroosmotic Flow (EOF) 5.1 Introduction 5.2 CE on the Microchip Format 5.3 Modes and Theories of CE Microchips 5.4 Microfabrication Techniques 5.4.1 Microfabrication of Glass CE Microchips 5.4.2 Microfabrication of Polymer CE Microchips

165 165 168 170 175 177 178

139 139

Contents

5.5 Basic Fluidic Manipulation/Motivation: Electrokinetic Injection and Separation Protocols 5.6 Electrochromatography in Microchip Format: Designs and Applications 5.7 Comparison of Hydrodynamic and Electroosmotic Flow-driven Miniaturized Systems 5.8 Analytical Applications 5.8.1 DNA Analysis 5.8.2 Protein Analysis 5.8.3 Small-molecule Analysis 5.9 Outlook References

vii

181 184 187 189 190 196 200 206 208

6 Detection in Miniaturized Analytical Systems 6.1 Introduction 6.2 Laser-induced Fluorescence (LIF) Detection 6.2.1 Lamp-based Fluorescence Detection 6.2.2 Fluorescence Excited by Light-emitting Diodes 6.2.3 (Electro)chemiluminescence Detection 6.3 Electrochemical Detection (ED) 6.3.1 End-channel Detection 6.3.2 In-channel Detection 6.3.3 Off-channel Detection 6.3.4 Electrode Materials 6.3.5 Modes of Detection 6.4 Microfluidics–MS Interfacing 6.4.1 Microfluidics-based Electrospray Ionization (ESI) Sources 6.4.2 Microfluidics–MALDI-MS Interfacing 6.5 Unconventional Detection Methods 6.5.1 Absorbance Detection 6.5.2 Surface Plasmon Resonance (SPR) Detection 6.5.3 Thermal Lens Detection 6.5.4 Other Detection Methods 6.6 Outlook References

213 213 214 218

234 243 248 248 249 251 253 254 254

7 Miniaturization of the Entire Analytical Process I: Micro(nano)sensors 7.1 Introduction 7.2 Evolution of Sensors with Nanotechnology

263 263 264

218 219 220 221 224 225 228 230 234

viii

Contents

7.3 Micro(nano)sensors 7.3.1 Optical Sensors 7.3.2 Electrochemical Sensors 7.3.3 Magnetic Sensors 7.3.4 Mass Sensors 7.4 Nanoprobes for In Vivo Bioanalysis References 8 Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 8.1 mTAS, Microfluidics and Lab-on-a-Chip: Concepts and Terminology 8.2 Basic Concepts of Microfluidics: The Design of Analytical Microsystems 8.2.1 Types of Transport 8.2.2 Laminar and Turbulent Flows 8.2.3 The Design of Microfluidic Systems 8.3 The Basics of Downscaling in Microsystems 8.4 Microfluidic Platforms: Types, Principles and Classification 8.4.1 Capillary-driven Test Strips 8.4.2 Microfluidic Large-scale Integration (LSI) 8.4.3 Centrifugal Microfluidics 8.4.4 Electrokinetic Platforms 8.4.5 Droplet-based Microfluidic Platforms 8.5 Microfluidic Devices for Analytical Lab-on-a-Chip Applications 8.5.1 DNA Analysis Integrated on Microfluidic Devices 8.5.2 Real Clinical Sample Analysis on Microfluidic Devices 8.5.3 Real Environmental and Related Sample Analysis on Microfluidic Devices 8.5.4 Real Food Sample Analysis on Microfluidic Devices 8.6 Outlook References 9 Portability of Miniaturized Analytical Systems 9.1 Introduction 9.2 Portable Gas Analysers 9.3 Portable Electrochemical Analysers 9.4 Portable Optical Analysers 9.5 Portable Lab-on-a-Chip Analysers References

266 266 270 272 273 275 278

281 281 283 284 285 286 290 293 294 296 297 299 300 302 305 310 323 328 338 339 345 345 346 351 351 352 354

Contents

ix

10 Analytical Performance of Miniaturized Analytical Systems 10.1 Introduction 10.2 Quality Control in Miniaturized Systems 10.3 Validation of Microsystems 10.4 Qualification of Microsystems 10.5 Robustness of Microsystems Further Reading

357 357 358 360 361 362 363

Index

365

Preface Without question, the main drivers of modern analytical techniques are the simplification of procedures and the improvement of measurement quality. To reach these goals, modern analytical techniques try to reach lower detection limits, improve selectivity and sensitivity, and achieve faster analysis time, higher throughput and less expensive analysis systems with ever-decreasing sample volumes. These very ambitious goals are exacerbated by the need to reduce the overall size of the device and the instrumentation. These items are termed ‘analytical miniaturization concepts’. The miniaturization of analytical systems is a rapidly growing area. It is associated with performing the analytical process on a small scale (sometimes, a very small scale). Different terms have been used to represent this idea: ‘miniaturized analytical systems’, ‘analytical micro(nano)systems’ and, more ambitious, ‘micro total analysis systems’ (mTAS), also called ‘lab-on-a-chip’. Originally, microsystems came from the need to perform online control and monitoring of industrial processes. This mainly resulted in the development of (bio)chemical sensors. However, after this initial phase, society’s needs increased and sensors became insufficient to respond to new specific problems, as they suffer from poor selectivity for many applications. As a consequence, the concept of the total analysis system appeared in analytical chemistry. The idea of such a system is to integrate all the steps of analytical process (sampling, sample treatment, separation of analytes and detection) in the same device. More recently, the interest in portable instruments, allowing field tests to be carried out with ease, has increased the practical usefulness of miniaturized analytical systems. Between sensors and lab-on-a-chip devices, a wide range of microsystems, which can affect either the entire analytical process or only a part of it, have been described. This exciting challenge has guided our effort to offer a book with a general approach to miniaturization in analytical chemistry, including the principles, designs and applications of miniaturized systems. Through ten chapters, the different issues characterizing such systems are critically discussed. The first two chapters include the basic concepts behinds miniaturization in analytical chemistry, as well as the mechanical and electronic tools needed for designing and fabricating these systems. It is very important to give an integrated classification of the systems and to define the different terms associated with miniaturization, in order to provide a systematic view of both the different levels of miniaturization and the main objectives of the downsizing developments.

xii

Preface

Chapters 3 to 6 represent the solid core of the book. Taking as their basis the analytical process, these chapters deal with: the miniaturization of sample treatment (including the consequent automation), with sections devoted to the problems associated with sample introduction in micro(nano)systems; miniaturized systems for analyte separation, divided into two chapters according to the forces involved in moving the flow; and detection in micro-size environments. Through these chapters, practical aspects such as the representativeness of the portion of sample analysed, the analytical potential of micro- and nanochromatography, the advantages of miniaturized capillary electrophoresis with special attention to microchip format, and both well-established and new approaches to detection in miniaturized systems, are comprehensively studied. Chapters 7 and 8 deal with the miniaturization of the entire process: from the sample introduction to the generation of the corresponding analytical results. Thus, when possible, Chapter 7 considers the use of sensors and biosensors in an online approach, or as micro(nano)probes, very useful for in vivo analyses. The objective of this chapter is not to give a wide report on the sensor field (different books address this topic), but to cover the integration of micro(nano)sensors in miniaturized technology. Chapter 8 covers the miniaturization of the entire analytical process under the philosophy of the mTAS approach. From a practical point of view, it is clear that mTAS entails many different challenges, but this trend in analytical chemistry is very attractive and will be the reality of future analytical work (both inside and outside the lab). Microfluidic concepts and lab-on-a-chip systems will make up the content of this chapter, in which a rich discussion of real samples is offered. The last part of the book (Chapters 9 and 10) deals with two aspects directly connected to the usefulness of miniaturized analytical systems: the design of portable miniaturized systems (very interesting for the performance of field tests) and how to assure the practical reliability of micro(nano)systems (quality control tests, performance and validation activities, as well as the robustness of the miniaturized depicted systems). The ruggedness of micro(nano)systems is briefly discussed and related to the tools used for the design and fabrication described in the first chapters of the book. Finally, the authors wish to thank the broad group of researchers who have contributed to analytical miniaturization developments over the past two decades. They have been cited throughout this book, where their works have been selected, studied and strategically related between them in order to give the reader a novel textbook on analytical miniaturization as a whole. The authors hope that the reader enjoys this book and finds it useful in their own teaching and research developments.

1 Miniaturization in Analytical Chemistry 1.1

Introduction

Miniaturization is rapidly growing, with novel ideas emerging in recent years. Like other fields, analytical systems have been affected by this new technology. Concretely, the capacity to carry out laboratory operations on a small scale using miniaturized devices is very appealing. Micro total analysis systems (mTAS), also called lab-on-a-chip, have renewed interest in scaling laws in the last 10–15 years. A small scale reduces the time required to synthesize and analyze a product, as greater control of molecular interactions is achieved at the microscale level. In addition, reagent cost and the amount of chemical waste can be very much reduced. Now, at the beginning of the twenty-first century, it is clear that the lab-on-a-chip approach is starting to be considered as a potential analytical tool in many application fields. Nevertheless, additional efforts must be addressed to two main points: (i) the laws at nanometre scale must be established, as basic physical and chemical fundaments cannot be applied; and (ii) more applications demonstrating the real use of these systems must be developed, particularly in the area of complex samples analysis. There is no doubt that miniaturized chemical analysis systems have a tremendous potential. For instance, it is foreseeable that such devices will allow the study and analysis of complex cellular processes, facilitate the development of new diagnostic abilities that could revolutionize medicine, and have applications in environmental monitoring, food analysis and industry. Some miniaturized analytical systems, such as capillary gas chromatography, microliquid chromatography and microcapillary electrophoresis – which can be Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

2

Miniaturization of Analytical Systems

considered as intermediate levels of miniaturization – have been consolidated in routine laboratories for the analysis of complex samples. There is no doubt that the majority of real analysis requires an appropriate sample treatment step. In this way, important efforts have been made to reduce the sample volume and its feasible manipulation in a miniaturized environment. The development of flow-processing devices for analyte purification/preconcentration, both on-column and on-fibre solid phase microextraction, hollow-fibre liquid phase microextraction, sorptive stir bars and so on are clear examples of miniaturization of the sample treatment step. The direct coupling of these methodologies (using at-, in- or on-line modes) to the instruments even allows an additional reduction of sample volume. At the same time, these coupled approaches reduce human manipulation and increase the degree of integration of the different analytical steps. The extension of this concept results in the modern concept of miniaturization: a miniaturized instrument integrating sample input, pre- and post-column reaction chambers, separation columns and detection units on to a single and small device. In this chapter, an overview of miniaturization in analytical science is presented. In addition to preliminary aspects dealing with miniaturization, other issues such as the classification and definition of terms, the features and main challenges of miniaturized analytical systems, the incidence of miniaturization in the different steps of the analytical process and present trends in miniaturization are covered in this chapter.

1.2 Miniaturization as One of the Critical Trends in Modern Analytical Chemistry Modern analytical chemistry is more and more a scientific discipline connected to the real world. International standards require a certain quality of client service from analytical laboratories, such as the norms ISO 9001 and, more specifically, ISO 17025. Hence, laboratories are encouraged to obtain client feedback in order to identify their real information needs. As Figure 1.1 represents, the developments and trends of modern analytical chemistry must be coherent with this feedback. Trends toward simplification (‘ease of implementation or use’), automation (‘electromechanical self-operation’, involving a feedback loop to control the system without human participation) and miniaturization (‘small scale’; ‘construction to a very small scale’) are well recognized. A common interface between these three characteristics can be identified as an attractive area for interesting developments, bringing about the so-called (re)-evolution of analytical chemistry from the end of the twentieth century onwards. This trend has a strong influence on present analytical chemistry. On the other hand, a high degree of simplification and automation is intrinsically involved in miniaturized systems. The simplification, mechanization/automation and miniaturization of the analytical process must assure the fitness for purpose of the analytical information generated with respect to customer information needs.

Miniaturization in Analytical Chemistry

MODERN ANALYTICAL CHEMISTRY (trends)

3

CLIENT INFORMATION NEEDS

SIMPLIFICATION AUTOMATION

ANALYTICAL PROCESS

FITNESS FOR PURPOSE

MINIATURIZATION

TECHNOLOGICAL DEVELOPMENTS

ANALYTICAL INFORMATION

Figure 1.1 General trends in modern analytical chemistry and their connection with the analytical process and client information needs

Transportantion to the detector

SAMPLE

Calibration

Sub-sampling SAMPLING

Sample treatment and conditioning

DATA ACQUISITION AND PROCESSING

Introduction

DETECTION

REAL WORLD

INFORMATION NEEDS

Analytical microsystems, when used to provide this analytical information in analytical laboratories, are no exception (Chapter 10 deals with issues related to the performance and reliability of the information provided by analytical microsystems). Following the scheme shown in Figure 1.1 and the idea of implementing different degrees of miniaturization of the analytical process, Figure 1.2 shows the number of steps connecting the real world (the client information needs) with the final results generated by the application of the measurement process. In this figure,

RESULTS

ANALYTICAL PROCESS ANALYTICAL VALIDATION REPRESENTATIVENESS

Figure 1.2 Scheme of the analytical process noting the different steps and activities, with special attention to miniaturization aspects

4

Miniaturization of Analytical Systems

the main steps and activities requiring especial attention for miniaturization purposes have been identified. Thus, among these activities must be considered the sampling (and the corresponding subsampling), calibration and validation of the entire analytical process; steps such as introduction of samples, treatment and conditioning of samples, transportation to the detection point, measurement, and data acquisition and processing; as well as final goals such as quality and representativeness of the analytical results. The possibilities and difficulties inherent in the miniaturization of each of these steps and activities are quite different. Even the level of miniaturization is an important issue. By considering the entire analytical process (fortunately, miniaturization implies reduction or elimination of some of the mentioned steps), the main problems for miniaturization affect the preliminary operations (sampling– sample introduction and sample treatment steps). Conversely, detection and signal transduction, as well as data acquisition and processing, are steps which can achieve a high degree of miniaturization. It seems clear that the final objective of miniaturized systems in the analytical domain is represented by mTAS [1]. In fact, ideally a TAS performs all the analytical steps (sample preparation, analyte separation and analyte detection) in an integrated instrument. The philosophy of TAS has been the enhancement of online and automated analysis, as well as of the analytical performance; however, significant drawbacks still exist, which should be the subject of future work in this field (for example, sample introduction and the successive analysis of a set of samples, slow sample transport, and the necessity of fabricating interfaces between the different components). The mTAS concept was developed from the modification of the TAS by downsizing and integrating its analytical multiple steps (sample preparation, separation and detection) on to single monolithic devices [1]. In essence, a mTAS is a device that improves the performance of an analysis by virtue of its reduced size. But not just analytical tasks can be performed in a miniaturized system; other chemical functions such as synthesis can also be performed. For this reason, today the mTAS concept is also known as ‘lab-on-a-chip’. After almost two decades since the concept was introduced, now is a very exciting time for mTAS, due to the large bulk of advances and challenges that appeared in the last revision published in Analytical Chemistry [2]. Some guidelines may be stated concerning the approach of miniaturization as a whole: (i) Miniaturization implies a micro-size environment. It is very important to keep in mind that miniaturization is not only a decreasing of scale, but that other forces and phenomena are present in micro-size environments [3–5]. (ii) Miniaturization requires technology facilities. Thus, to understand what the revolution of miniaturization in analytical chemistry means, it is necessary to consider the synergic marriage (relationship) between some part of the technology and the objectives of modern analytical chemistry. In other words,

Miniaturization in Analytical Chemistry

5

in their fundamental sense, the objectives or demands of analytical chemistry should answer to some questions related to miniaturization, such as why, when and how miniaturization is performed. Obviously, the answer to the last question comes from the existing technology facilities [6]. Figure 1.3 shows the outlines of this complementation. From an analytical point of view, traditionally chemical analyses have been performed in central laboratories since they require skilled personnel and specialized analytical instrumentation. However, the trend today is to move chemical analyses close to the ‘customer’ or the bulk sample (in-situ analysis). Miniaturization plays a prominent role in the decentralization of chemical analysis. As a consequence of this, miniaturized analytical devices should be portable, easy to operate and reliable. The automation of chemical analysis also requires the real implementation of mTAS in analytical laboratories. In fact, analytical microsystems could ultimately be represented as a black box, where the role of the users basically consists in providing the samples and pushing a start button. From a technological point of view, several potential benefits of analytical microsystems can be observed (see Figure 1.3): very well-understood knowledge about microfabrication techniques and the chemistry of materials; the possibility for fabricating sophisticated microcircuits according to analytical demands/objectives; and high compatibility with mass production. Another advantage lies in the possibility of opening new directions in microfluidics, such as the presence of laminar flow. On the other hand, one weak point of analytical microsystems comes

Technologycal tools

Analytical demands

G

D OO

L MP CO

IM O PR M VE TS EN

- Automation - Simplification: No skilled personnel No sophisticated instrumentation - Decentralization of analyses - Samples: Very low volumes High throughput - Reagents: Low consumption

TY RI TA N E EM

- Microelectronic developments - Microfabrication techniques - Microfluidics - Chemistry of materials - Cleanroom facilities

E ED NE D

-Real-world interfaces: Sampling and sample introduction Incomplete miniaturization Conditioning of the microsystem Change of sample between analyses

Figure 1.3 Relationships between present analytical demands and technological developments for miniaturization

6

Miniaturization of Analytical Systems

from the present technology; sometimes the complete miniaturization of all the electronics and mechanical parts of the system is not allowed. Moreover, these elements are expensive and the required technology is not always available (for instance, cleanroom facilities). The microtechnology used in the miniaturized environments could also be understood as the integration between the possibilities offered by microfabrication and microfluidics. Both microfabrication (micromachining) and microfluidics are inversely connected with the grade of complexity in the analytical chemistry process. Indeed, while microfabrication plays the prominent role in the separation and detection systems, the physical control of the process has a prominent role in all steps of sample preparation. In addition, the required microfabrication is very sophisticated in the sample preparation step. Chapter 2 deals with the tools for designing miniaturized analytical systems. Two other intrinsic characteristics of analytical microsystems have clear connections with technology developments: the extremely low sample volumes used, and the presence of laminar flows. Small volumes of both sample and reagent (pl–nl levels) are representative of most miniaturized systems. This characteristic has clear advantages associated with cost and analytical throughput, but it also presents disadvantages, such as the suitability of detection techniques. Consequently, much research effort has been focused on the development of miniaturized and sensitive detection units [7], and today the detection improvements are still one of the most important research focuses [8]. As a consequence of the low volumes required, a very precise handling of sample is crucial in microanalytical systems, and a high dependence of the surface properties of microchannel manifolds and interconnections and dead volumes is observed [3]. Complex fluid manipulation at femtolitre and nanolitre scales is readily achieved without any mechanical valves or external pumps by using the electrokinetic phenomena [9]. In this way, the focus has been centred on the integration of functional components within monolithic systems using both lithography and micromoulding technologies. Micropumps are other typical devices used to propel fluids in microchannels. A recent review has been published by P. Woias [10]. A more sophisticated approach has been proposed by M.S. Anderson [11], who presented a combined atomic force microscope (AFM) and Raman spectrometer as a microfluidic device for sampling and trace chemical analysis. On the other hand, the miniaturization of valves for microfluidics [12], or to set up a miniaturized lab-on-valve system [13], constitutes an additional tool for fluid manipulation in microchips. The main examples of integrated processing within microfabricated devices have been directed at linking analytical principles (i.e. detection/CE separation, reactors, designs of microcircuits). As previously mentioned, miniaturization is more than simply the scaling down of well-known systems. The relative importance of forces and processes changes with the scale. Thus, as a consequence of their miniaturized scale, another feature of the analytical

Miniaturization in Analytical Chemistry

7

microsystems is the presence of laminar flow (Reynolds number is typically very low), where viscous forces dominate over inertia. This means that turbulence is often unattainable and the transport of molecules is only produced by diffusion, which has direct consequences on the design of this type of microsystem. This constitutes one of the most attractive features, since in most analytical microsystems the diffusion process is very fast, as diffusion effects are inversely proportional to the square of the length. This consequence will be especially relevant during the sample preparation step. However, the ability to efficiently process raw samples (as in classical laboratory tests, directly from the body of a person or an animal, or in field tests), and subsequently to perform the required analytical operations on-chip, will be a key aspect in both the definition of the eventual success and the application of microfluidic systems [14,15]. Whereas the previous points are specific for mTAS approaches, other general requirements can be stated, from an analytical point of view, for a successful miniaturized approach: (i) The developed analytical microsystems must be close to real-world demands (the objective is to solve analytical problems). Just one of the main challenges of analytical microsystems is dealing with the real-world interface. That means understanding the information needs of the client (identification of the problem and its translation to an analytical level), which is a general challenge of any analytical work. But, more important for microsystems, it also means ensuring the representativeness of the results through an appropriate sampling plan and a suitable method of analysis at the particular miniaturized level. This includes achieving the selective and sensitivity requirements through sample treatment and separation/detection. In addition, analytical microsystems offer a significant decrease in costs, by dramatically reducing the volume of samples and reagents needed to perform a chemical analysis. This feature also opens up the possibility of processing samples in parallel, which is very useful when the same chemical analysis must be performed many times, as is the case in routine laboratories. This approach is very useful when high-throughput screening is needed. In conventional analyses, handling and processing of the sample is frequently done manually, at least in part, and often in specialized laboratories. However, mTAS allows chemical analyses to be brought close to the place where they need to be performed, independent of both the laboratory and the laboratory personnel. Thus, these integrated analytical systems are very suitable for online measurements. (ii) The analytical microsystems must perform reliably, in order to be coherent with the present quality assurance requirements in analytical laboratories. In this respect, issues related to the calibration and validation of the methodologies carried out by the analytical microsystems must be taken into account. As the

8

Miniaturization of Analytical Systems

final objective will be the analysis of real samples, the validation of such a method has to be performed using the samples for which the method is intended.

1.3 Evolution in the Field of Analytical Miniaturization A. Manz et al. established different periods in the evolution of analytical microsystems in a publication in 2002 [16]. Based on this source, and completed with recent developments, Figure 1.4 shows the schematic history of miniaturization in the analytical field. The most representative milestones are briefly described below. The first period (1975–1989) was characterized by works addressed to the miniaturization of components, such as micropumps, microvalves and chemical sensors, based on silicon technology. The integration of such silicon microcomponents is remarkable for the fabrication of two miniaturized instruments. The first was a miniaturized gas chromatographic analyser [17]. The basic chromatographic device included, in a single silicon wafer, an injection valve and a separation column

S m CONVENTIONAL ANALYTICAL SYSTEMS

Analytical systems partially miniaturized

mm

I

Z

E µm

nm

MICRODEVICES MINIATURIZED DIVICES

1970

Microchips

NANODEVICES

Nanochips

‘The Early Days’ (1975-1989)

Y E A R S

1980

Miniaturized gas chromatographic analyser Fabrication of micropumps and microvalves Chemical sensors Miniaturized coulometric acid-base titration system

1990 ‘The Renaissance’ (1990-1993)

Miniaturized open-tubular liquid chromatograph Introduction of µTAS concept Miniaturization of FIA systems Electroosmotic pumping Electrophoresis in planar chips First applications related to handling biomolecules and cells Microchips for flow cytometry

‘Growing to Critical Mass’ (1994-1997)

2000

Consolidation as a potential analytical alternative for solving real problems Nanotechnologycal approaches

Chip-based analyses Different miniaturized commercial products On-chip liquid chromatography with EC detector Surface chemical modification of the microchannels Miniaturized mass spectrometer Advances in microfabrication, design, separation, biochemical reactors and detection

Molecular or molecular-sized devices Analytical nanotechnology

Figure 1.4 Evolution of miniaturized analytical systems and main milestones (adapted from [16] and completed for years since its publication)

Miniaturization in Analytical Chemistry

9

1.5 m long. A thermal conductivity detector was fabricated on a separate silicon wafer and mechanically clamped on the wafer containing the column. The second type of miniaturized instrument was a coulometric acid–base titration system, employing a solid-state pH-sensitive sensor to determine the acid or base concentration in the sample [18]. The second period (1990–1993) saw the fabrication of silicon-based analysers in 1990, thanks to new developments producing a miniaturized open-tubular liquid chromatograph fabricated on a silicon wafer [19]. This work presented a 5 · 5 mm silicon chip containing an open-tubular column and a conductometric detector, connected to an off-chip conventional LC pump and valves in order to perform highpressure liquid chromatography. It was also in 1990 that Manz et al. proposed the concept of mTAS [20], in which silicon chip analysers incorporating sample pretreatment, separation and detection played a key role. According to this author, the main reason for this miniaturized approach was to enhance the analytical performance of the existing sensors, due to the poor results in terms of selectivity and lifetime showing at this time. Therefore, the main objective of mTAS, initially, was not the reduction of size, although the advantages of miniaturization were recognized. The miniaturization of a flow-injection analysis (FIA) system, based on stacked modular devices in silicon and plexiglass, was also reported in this period [21,22]. This key device (less than 1 cm3) conformed to a 3D structure in which more than 10 chips were integrated. Despite the new developments in micropump systems and microvalves for microflow arrangements, the high pressures necessary for transporting in microchannels complicated their practical use. In this way, electro-osmotic pumping was an attractive and feasible tool for the movement of aqueous media in mTAS, particularly when separation was needed. In fact, electro-osmotic pumps are characterized by the absence of mechanicallymoving parts and the lack of a specific location of the pump in the manifold. Moreover, the flow in the interconnected channels can be controlled by switching voltages, without the need for valves. These new strategies were critical to the development of electrophoresis in planar chips, used for the first time in 1992 [23,24]. It was also in this period that applications related to the reaction and handling of biomolecules and cells started; for instance, the use of microfabricated chambers to carry out DNA amplification (PCR) or flow citometry. An important increase in the number of publications related to mTAS took place in the third period (1994–1997). A great variety of chip-based analyses were reported, and different miniaturized commercial products were offered by important instrumentation companies. The modular concept of mTAS was revisited, involving both electrochemical and optical detection, and at the same time, surface chemical modification of the microchannels opened interesting possibilities for miniaturized chromatography. Thus, Cowen and Craston developed an on-chip liquid chromatographer with an electrochemical detector [25]. Another interesting development was carried out by Feustel et al. [26], consisting of a miniaturized mass spectrometer

10

Miniaturization of Analytical Systems

incorporating an integrated plasma chamber for electron generation, an ionization chamber and an array of electrodes acting as the mass separator. Additionally, in this period, significant contributions to microfabrication (covering a broader range of applications and using new materials), separation modes, detection devices and new applications (focussed on biological species, mainly) were made [16]. The last 10 years, since 1998, can be considered a period of consolidation of miniaturized developments, which can be seen as potential analytical alternatives for solving real problems. The wide number and variety of publications is decisive proof of this, although in many cases the developments remain in the research world, with limited implementation in control or routine laboratories. As microfabrication technologies, and particularly the design and production of microfluidic systems, constitute one of the milestones of analytical miniaturization, the important developments in flow-control devices (micropumps and microvalves), interconnections and interfaces, together with bonding techniques and surface modifications, have played a key role in the present analytical miniaturization scene. This is the subject matter of part of Chapter 2 of this book. The application fields of analytical miniaturized devices have been clearly expanded, with a particular impact in the bioanalytical area. Thus, many uses deal with protein and peptide analyses, DNA separation, PCR and performance of inmunoassays or clinical diagnosis. A rise in publications related to cellular applications can also be observed. Selected examples have been reported throughout the various chapters of this book.

1.4 Classification of Miniaturized Analytical Systems and Definition of Terms Miniaturized systems can be classified according to different criteria. Figure 1.5 presents a classification distinguishing between general criteria and specific criteria applied to analytical work. Thus, from a general point of view, miniaturization is clearly associated with the reduction of size, although the term can be somewhat confusing, and it does not always involve a ‘small scale’. It is a relative term, depending on what is understood by the ‘normal scale’. This conflict can be resolved by giving to the prefix ‘mini-’ a broad meaning, representing devices, instruments and systems at dimensions greater than 1 mm (even at the cm range). This is the starting point in Figure 1.6, where the basic classification of the miniaturized systems according to the intrinsic size is represented at two fields: the general and the specific analytical field. From a general point of view, the three basic levels of miniaturization are established by using the prefix mini- (higher than 1 mm, or at the cm scale), micro- (lower than 1 mm) and nano- (lower than 1 mm). Due to the frequent use of these systems with liquid (aqueous) media, the corresponding volumes associated with these levels are included in Figure 1.6, as well as the type of technology involved in the fabrication of the corresponding

Miniaturization in Analytical Chemistry ‘NORMAL’(MACRO)SYSTEMS

11

GENERAL STUDIES

MINISYSTEMS SYNTHESIS MICROSYSTEMS

SIZE

OBJECTIVE ANALYSIS

NANOSYSTEMS

GENERAL CRITERIA FOR CLASSIFYING

MINIATURIZED SYSTEMS ANALYTICAL CRITERIA FOR CLASSIFYING

ANALYTICAL PURPOSE

COMPLEXITY OF SAMPLE AND INFORMATION NEDEED

SAMPLE HANDLING AND TREATMENT DIRECT MEASUREMENT OF ONE COMPONENT

(BIO)CHEMICAL REACTIONS SEPARATION OF ANALYTES

MEASUREMENT OF FEW COMPONENTS WITH SOME SAMPLE TREATMENT

DETECTION COMPLEX SAMPLES REQUIRING THE SEPARATION OF THE SAMPLE COMPONENTS

Figure 1.5 Different criteria used for classifying miniaturized systems

systems. Thus, conventional technology has commonly produced the wide variety of minisystems existing so far, whereas micro- and nanotechnology were used to produce micro- and nanosystems, respectively. Microsystems are microstructured devices, integrated as structures in the micrometric range, produced by using microfabrication techniques; the term ‘nanosystem’ refers to a system that has its main structures in the 1–100 nm range. MEMS (microelectromechanical systems) and NEMS (nanoelectromechanical systems) are representative devices often used in micro- and nanoscale works, as well as micro- and nano-fluidic structures. Analytical miniaturization could be viewed as ‘the fact of making to a small scale a part or the whole of the analytical process (see mTAS), or of reducing the size of the different devices involved in the analytical process (see micropumps, microsensors) or the analytical technique itself (see microgravimetry)’. Consequently, when the general-sized concepts are applied to analytical science, the corresponding three levels of miniaturization can be called: (i) Analytical minisystems or minitechniques, involving some specific devices such as minireactors or minicolumns; or the application of minitechniques, for which commonly the prefix mini- has been replaced by micro- (although not appropriately, according to the general classification). This is the case for

12

Miniaturization of Analytical Systems

MINIATURIZATION

smaller scale

Mini-

Micro-

Nano-

>> 1 mm (µL)

< 1 mm (nL)

< 1 µm (pL, fL, aL)

CUASI-CONVENTIONAL FABRICATION TECHNIQUES MINISYSTEMS

MICROTECHNOLOGY

NANOTECHNOLOGY

MICROSYSTEMS

NANOSYSTEMS

MEMS Fluidic microstructures

NEMS Nanofluidics

MINIATURIZATION IN ANALYTICAL SCIENCE ANALYTICAL MINISYSTEMS or MINITECHNIQUES Minireactors Minicolumns (Micro)gravimetry (Micro)titration

Micro-GC Micro/nano HPLC

ANALYTICAL MICROSYSTEMS Microreactors Micropumps Microvalves Capillary columns Microsensors Microactuators Array microsystems µFIA µCE µTAS

ANALYTICAL NANOSYSTEMS Nanoparticles Nanotubes Nanofibers Quantum dots Etc.

(longer scale)

Figure 1.6 Classification of miniaturized systems based on size criteria

microgravimetry and microtitration. Other, longer-scale equipment (chromatographic instruments, mainly) has been named micro-GC or micro-/nanoHPLC, not due to the size of the ‘miniaturized’ equipment, but for the use of minidevices (such as pumps, valves or capillary columns) for handling microor nanovolumes, and for introducing separation advantages against normalsized equipment. From this point of view, the commercially-available capillary electrophoresis (CE) equipment (fabricated at high scale size) could be associated with micro-/nano-CE. Just in this case, the term ‘micro-CE’ is correctly used when the electrophoretic separation is performed in microchips. This fact is not so frequently used for GC and LC in microchips. (ii) Analytical microsystems are microstructured devices, integrated as analytical structures in the micrometric range, produced by using microfabrication techniques. They include microdevices (microreactors, micropumps, microvalves, capillary columns, microsensors and microactuators, array microsystems), microtechniques (mFIA, lab-on-a-chip valve, mCE), and the complete analytical process (mTAS). (iii) Analytical nanosystems are, in fact, systems at nanometer size, built with atomic precision by using nanotechnology facilities. Nanotechnology is a

Miniaturization in Analytical Chemistry

13

multidisciplinary field of applied science and technology working at approximately 1–100 nm and dealing with the fabrication of devices at this size. Useful materials, devices and systems at this size range may be obtained by two different approaches. In the ‘bottom-up’ approach, materials and devices are built from molecular components which assemble themselves chemically by principles of molecular recognition. In the ‘top-down’ approach, nanoobjects are constructed from larger entities without atomic-level control. Examples of analytical nanosystems are nanoelectromechanical systems, nanoelectrodes, etc. In addition to the proper size classification of the miniaturized systems, from a chemical point of view, miniaturization can be addressed to three main objectives (Figure 1.5): (i) synthesis of compounds; (ii) analysis of samples; and (iii) biochemical studies of cells, microorganisms, etc. In this context, the article by D. Belder presenting the integration of chemical synthesis and analysis on a chip as a potential trend is interesting [27]. This author recognizes that while complex integrated lab-on-a-chip systems have been described for many biochemical applications, similar achievements remain to be made in synthetic chemistry. The main reason for this, besides the often more challenging reaction conditions for synthetic chemistry, is the lack of appropriate detection techniques. Thus, fluorescence detection, which is commonly applied for on-chip detection in bioanalytics, is often not applicable to classical chemical reactions and corresponding products of interest. A promising approach to carrying out chemical reactions and analysis is the use of micro-CE. An example of an integrated reaction/separation device applied to synthetic chemistry and catalyst development has been presented by D. Belder et al. [28]. Figure 1.7 shows a scheme of the microfluidic chip, combining a microfluidic reactor with microchip electrophoresis. The chip layout is basically a merged design of the common cross-channel design for on-chip electrophoresis and a typical meandering channel for mixing and reaction. On the other hand, miniaturization technology provides facilities for creating tools with feature sizes matching the dimensions of cells, and enables integration of cell-handling and fluidmanipulation elements [29]. Therefore, microsystems create new opportunities for the spatial and temporal control of cell growth [30]. The additional integration with bioanalytical platforms creates multifunctional microdevices showing great promise for basic biomedical and pharmaceutical research, as well as robust and portable point-of-care devices. The analytical purpose of the miniaturization can be focussed on sample handling and treatment, the development of (bio)chemical reactions in miniaturized environments, the separation process or the detection. The different incidences of miniaturization in the analytical process steps are discussed in Section 1.7. Microarrays are probably the most representative example of microanalytical systems based on selective and specific (bio)chemical reactions. Their use normally involves

14

Miniaturization of Analytical Systems

Figure 1.7 Integrated synthesis/analysis chip (channel width 50 mm) showing the different parts used for introduction of educts through microvials (1), merging zone (2), mixing and reaction microcoil (3), injection cross for the introduction of the reaction products to the microelectrophoretic system (4), and microchannel for the electrophoretic separation of the products (fluorimetric detection). (Reprinted from [28] with permission from Wiley VCH)

biochemical reagents such as antibodies or genetic material, or specific chemical reagents such as molecular imprinted polymers (MIPs). The second group is constituted by those systems involving (electro)chromatographic separations. Portable GC equipment is a typical example, which can be used to detect volatile compounds in foods. In the case of liquid chromatography, miniaturization results in a high pressure in the system, which makes its use difficult. For this reason, the major tendency is to apply electrochromatographic separations, and hence electrophoretic chips are probably the most developed and studied miniaturized system. The complexity of samples and the intended information required are two factors taken into account by J.P. Kutter and O. Geschkes [31] for distinguishing between: (i) the direct measurement of one or a few components with no or little sample preparation; (ii) measurement of one or a few components requiring some treatment of sample; and (iii) more complex samples (separation of the components). Undoubtedly, these categories are based on sample complexity and selectivity criteria. Accordingly, different analytical microsystems can be associated with them, for example (bio)sensors (probe-type sensors, flow-through sensors and microsensors in a progressive scale of miniaturization) for the first group. FIA and mFIA are the most appropriate (micro)systems for the second group, whereas m-LC and m-CE are ideal for performing microseparations. There is no doubt that information from objects and systems is a relevant part of the requirement for making well-founded and timely decisions [32]. As a response to this high need of information, analytical sciences have developed rapid, low-cost screening methods

Miniaturization in Analytical Chemistry

15

that allow the classification of samples as positive or negative [33,34]. Most of the screening methods provide a total index of global response, which permits the classification of samples from the legal, toxicological or quality point of view.

1.5

Theory of Miniaturization

Theoretical considerations can first be established in terms of similarity and proportionality. The similarity approach uses dimensionless parameters to consider similarity, whereas the proportionality approach uses the characteristic length of known systems versus scaled-down systems to consider proportionality [16]. The former can be used to correlate, in an easy way, experimental results when a great deal of variables are involved, and they are defined in terms of parameters that are assumed to be constant through the whole system. The proportionality approach provides very useful information related to the behaviour of a simple flow system when miniaturized. It assumes the miniaturization as a downscaling process in three dimensions. In fact, as the key factor of miniaturization is the size of the devices, the typical device length, d, can be used as the basis of fundamental changes in the characteristics of miniaturized devices [35]. Thus, Table 1.1 summarizes the basic characteristics of devices at three different sizes: 1 mm, 100 mm and 10 mm. The corresponding volumes associated with these typical lengths clearly decrease, with values of 1 mL, 1 nL, and 1 fL, respectively; as do the number of molecules for a particular concentration of the flowing solution. More important are the three final parameters in Table 1.1. Thus, for instance, as the number of units that can be arranged on a given surface increases with 1/d2, for a typical length of 10 mm, 2.5 · 105 devices/cm2 (250 000 units) can theoretically be integrated per cm2. The smaller dimensions have a further impact on analytical standard operations such as mixing, separation and detection, because turbulent flow mixing is avoided. Thus, on the microscale, viscous forces dominate over inertial forces, leading to a laminar flow regime. Under this condition, two liquids can co-flow without turbulent

Table 1.1 Device characteristics for different typical-length d values (extracted from [35]) Typical length:

1 mm

100 mm

10 mm

Volume Number of moleculesa Diffusion time Unit density (devices/cm2) Information density (values/min/cm2)

10
6 l 6 · 1011 15 min 25 1.5

10
9 l 6 · 108 10 s 2500 250

10
12 l 6 · 105 100 ms 2.5 · 105 2.5 · 106

a

For a concentration 1 mM.

16

Miniaturization of Analytical Systems

mixing. Reynolds number characterizes the type of flow, and can be calculated by the following equation: rrh v Re ¼ ; ð1:1Þ h where r is the fluid density, rh is the hydraulic diameter of the capillary, u is the average fluid velocity and Z is the dynamic viscosity. The lower Re, the closer the behaviour of the flow to a laminar flow. In general, systems are considered to be in the laminar flow regime for Re < 2000. This value is even lower than 1 for microfluidic devices. Therefore, diffusion is an important parameter for these microsystems. Thus, as the time that a molecule needs to travel a length d decreases with d2, for a small molecule (diffusion coefficient 10
9 m2/s), this results in diffusion times decreasing with the typical length of the device (Table 1.1). For instance, diffusion time is only 100 ms for devices of a typical length of 10 mm. As the diffusion times can be related to the capacity for exchanging molecular information according to the expression: Information ¼

1 ; tdiffusion

ð1:2Þ

when the number of volumes is taken into account, an information density parameter can be calculated as a function of the time and surface area, given the values shown in Table 1.1. These values theoretically increase with the fourth power of d. This is a critical aspect, because it is the basis of the high-throughput capability which characterizes miniaturized analytical systems. The influence of relevant parameters on the behaviour of a miniaturized system can be evaluated from proportionality considerations. For this purpose, the parameter of interest must be defined as a function of space (d) and time (t), which are the key variables of the miniaturization. This allows knowledge of the characteristics of a parameter after downscaling, without having knowledge of other constants associated with the material. Under this model, miniaturization is viewed as a simple 3D downscaling process based on d factor. Depending on whether or not the other key variable, time, is constant, two different situations can be distinguished: (i) Time-constant system, for which diffusion is of lesser importance. In this case, the timescale is the same for the miniaturized system as for the large system. Variables characterizing the time parameters, such as transport time, response time and analysis time, remain the same. This is the situation characterizing simple transportation systems and FIA, where an analyte is injected into a flowing carrier stream for subsequent analysis. According to A. Manz et al. [36], when a time-constant system is scaled down, the linear flow rate decreases by a factor d, volume flow rate by d3 and voltage by d2 (Table 1.2). Hence, the main advantage is the reduced consumption of sample, carrier and reagent solutions.

Miniaturization in Analytical Chemistry

17

Table 1.2 Scaling factors for miniaturized analytical systems, where d represents a typical length in the system (extracted from [36]) Parameter

Time-constant system

Diffusion-controlled system

Space Time Linear velocity Volume flow rate Pressure dropa Voltageb Plate number

d constant d d3 constant d2 —

d d2 1/d d 1/d 2 constant contant

a

Laminar flow conditions; b For electro-osmotic pumping.

(ii) Diffusion-controlled system, in which the time is regarded as a surface and proportional to d2. Parameters such as molecular diffusion, heat diffusion and flow characteristics are important because they control the separation efficiency. Therefore, in these cases, timescale needs to be considered. As the diffusion time on the travel length is proportional to d2, a 10-fold downscaling produces a 100-fold reduction of related time parameters (for instance, the analysis time). On the other hand, as Table 1.2 shows, the pressure drop scale is proportional to 1/d2, meaning that for a 10-fold downscaled system, a 100 times higher pressure drop is required to generate the same flow. If an electrical potential is applied to produce an electro-osmotic flow (EOF), the generated flow remains the same when the electric field is kept constant. The separation efficiency (expressed through the plate numbers) is also unaffected by miniaturization, but is directly proportional to the applied voltage. Obviously, the applied voltage is limited by the microsystem heating (Joule effect), although for miniaturized systems, lower currents are used, and the higher surface-tovolume ratio allows a good heat dissipation. Based on this fact, ultrafast, highefficiency separations can be achieved in miniaturized diffusion-controlled systems.

1.6

Features of Miniaturized Analytical Systems

The previous section pointed out some of the features and advantages of miniaturized systems. Now, Figure 1.8 summarizes some of the main characteristics of analytical microsystems, establishing their relationship with the quality of the generated information. Automation (human participation reduction) and simplification are two features commonly associated with miniaturization. The other characteristics listed in Figure 1.8 show clear influences on the speed of result generation (high throughput), the amount of information (simultaneous or multiparametric)

18

Miniaturization of Analytical Systems CHARACTERISTICS OF ANALYTICAL MICROSYSTEMS

LOW SAMPLE VOLUME

HIGH THROUGHPUT

RAPID ANALYSIS HIGH EFFICIENCY ROBUST

SIMULTANEOUS DETERMINATIONS OF ANALYTES

LOW COST PORTABLE

MORE REPRESENTATIVE RESULTS

HIGH LEVEL OF INFORMATION

FIELD TESTS

LOW HUMAN PARTICIPATION

QUALITY OF RESULTS Figure 1.8 Main characteristics of microsystems and their effects on the quality of the analytical results provided

and the autonomy given for microsystems allowing field tests (portability). These characteristics represent very advantageous aspects for the corresponding analytical methods using miniaturized systems. The development of such microsystems requires the miniaturization of the electronic and mechanical parts of the system. The microtechnology used in the miniaturized environments could also be understood as integrating the possibilities offered by microfabrication and microfluidics. The term ‘microfluidic’ is used to denote any process that involves the use of small amounts of fluids, such as 10
9–10
18 litres, in channels with a small diameter (tens to hundreds of micrometres). Hence, mTAS are examples of microfluidic systems. From an analytical point of view, these microsystems offer a number of useful capabilities: (i) the ability to use small quantities of samples and reagents; (ii) the possibility of carrying out separations with a high resolution; (iii) the use of low-cost setups; and (iv) the short time taken to perform complete analyses. In addition, microfluidics also exploits characteristics of fluids in microchannels, such as laminar flow. This allows new capabilities in the control of the molecules in space and time. One of the most important characteristics of microfluids is the dramatic difference between the physical properties of fluids moving in large channels and those travelling through micrometer channels, mainly the turbulence or its absence (laminar flow). On large scales, fluids mix convectively, as inertia is often more important than viscosity. However, in microsystems, the fluids do not mix convectively and, consequently, two fluidic streams that merge in a microchannel will

Miniaturization in Analytical Chemistry

19

flow in parallel, without any turbulence. Therefore, the diffusion of molecules across the interface is the unique phenomenon responsible for mixing. When effective mixing is required, it is necessary to introduce specific devices to accomplish this. Another important characteristic is the presence of electro-osmotic flow, which, as in CE, is a consequence of the ionization or surface charge of the microchannels. Therefore, when an electrical potential is applied across the microchannel, the fluid moves as a plug, with a characteristic planar flow profile.

1.7

Incidences of Miniaturization in the Analytical Process

(Bio)chemical analysis is performed through the so-called analytical process, which integrates a group of steps and substeps connecting the sample with the corresponding results. Miniaturization can affect a single step/substep, various integrated steps/substeps or the entire process. Figure 1.9 illustrates these possibilities, taking the analytical process as the central part, and considering it the ‘analytical black box’, which can be the basic subject of integrated analytical (micro)systems; or the specific sequence of the three main steps: preliminary operations, signal measurement and transduction, and data acquisition and processing. Hence, the upper part of the figure represents a (micro)system performing the whole process, whereas the lower part includes the different analytical standard operations involved in the process. This view allows a description of the incidence of miniaturization in the analytical process as: (i) the partial miniaturization of step(s), devices or equipment;

INTEGRATED MICROSYSTEMS

MINIATURIZATION (simplification)

µTAS

AUTOMATION SIMPLIFICATION INTEGRATED SYSTEMS

TAS

IN VIVO MEASUREMENT DEVICES

PORTABLE EQUIPMENT

S A M P L I N G

SAMPLE

SUBS A M P L I N G

ANALYTICAL ‘BLACK BOX ’

ANALYTICAL PROCESS PRELIMINARY OPERATIONS

SIGNAL MEASUREMENT AND TRANSDUCTION

RESULTS DATA ACQUISITION AND PROCESSING

Sample introduction Sample preparation and conditioning Chemical reaction

Chromatographic & Electrophoretic Separations

DETECTION

Microprocessor (software facilities)

ANALYTICAL STANDARD OPERATIONS

Figure 1.9 Incidence of miniaturization in the whole analytical process or individually in the different analytical standard operations involved in the process

20

Miniaturization of Analytical Systems

or (ii) the miniaturization of the integrated systems performing the entire analytical process, in which miniaturization should be viewed in combination with automation and simplification of the process. 1.7.1

Miniaturization of the Steps of the Analytical Process

The main approaches used for the miniaturization of the steps making up the analytical process are given below, pointing out the present challenges and evaluating the main strengths/advantages and weaknesses/disadvantages existing in each. Taking into account the natural development of the miniaturization devices in analytical works, the logical sequential steps of the analytical process have been inverted, in order to present them in accordance with the increasing difficulty of their miniaturization. Detection Techniques Detection has been one of the main challenges for analytical microsystems, since very sensitive techniques are needed as a consequence of the ultra-small sample volumes used in micron-sized environments. In principle, a wide variety of detection alternatives can be used in microfluidic systems [37]. Laser-induced fluorescence (LIF) was the original detection technique and is the most used detection scheme in CE microchips, due to its inherent sensitivity [38,39], given by the ease of its focusing. This characteristic, together with the fact that molecules of biochemical relevance are fluorescent in many cases, is an important reason for the wide use of LIF in microfluidics. Today, LIF has attained a position as a standard detection technique for microchip separation [39]. However, the high cost and the large size of the instrumental setup of LIF are sometimes incompatible with the concept of mTAS. Also, tedious derivatization schemes are needed to use LIF with nonfluorescent compounds. The most important alternative to LIF detection is, without any doubt, electrochemical detection (ED). ED is very important because of its inherent miniaturization without loss of performance and its high compatibility with microfabrication techniques. Likewise, it has a high sensitivity, its responses are not dependent on the optical path length or sample turbidity, and it has few power-supply requirements, which are all additional advantages. As proof of the prominent role of ED, see the recent publication of excellent reviews [40–43] and others papers [44–48]. The main challenge in CE–ED coupling has been the conflict between the high voltages used in the electrophoretic separation and the potential used for the detection. However, at microscale this drawback is over. Three strategies have been employed: end-channel, in-channel and off-channel electrochemical detections. In end-channel detection the electrode is placed just outside of the separation channel, which involves the alignment of the electrode. Separation voltage has a

Miniaturization in Analytical Chemistry

21

minimal influence on the applied potential in the electrochemical detector because most of the voltage is dropped across the channel. For in-channel detection, the electrode is placed directly in the separation channel using an isolated potentiostat. Off-channel detection involves grounding the separation voltage before it reaches to the detector by means of a decoupler. Electrode placement in off-channel detection is similar to that in in-channel detection, but the separation voltage is isolated from the amperometric current through the use of a decoupler. Conceptually speaking, the decoupler effectively shunts the separation voltage to ground and a field-free region is created where analytes are pushed past the electrode by the EOF generated before the decoupler. Since no decoupler is necessary, end-channel configurations offer the following advantages: they are simple and rugged, electrode replacement is feasible, and microfabrication facilities are not strictly required. However, the main drawbacks are the alignment of the electrode with the outlet of the channel, and the loss of separation efficiency due to the distance between the end of the channel and the working electrode. This separation distance is also crucial for the signal-to-noise ratio obtained and can lead to a complete loss of the analytical current. In both in- and off-channel configurations, the analytes migrate over the electrode while they are still confined to the channel, thus eliminating the band broadening often observed with end-channel alignments. However, in these configurations, microfabrication facilities are usually needed and in addition the nature of on-chip miniaturized electrodes limits the ability to modify and periodically clean the electrode surface. Conductimetric detection on microfabricated devices has recently been developed. It constitutes an important possibility for detection in analytical microsystems. In comparison to amperometry, conductimetric detection is less sensible, but it is a universal detection technique and has been applied as a detection mode in CE microchips, both in the galvanic (a pair of electrodes is placed in the separation channel for liquid impedance measurement) [49,50] and in the contactless (no contact between the pair of electrodes and separation channel solution) [51–54] mode. Contactless conductimetric detection is preferred for three main reasons: (i) the electronic circuit is decoupled from the high voltage applied for separation (no direct DC coupling between the electronics and the liquid in the channel); (ii) the formation of glass bubbles at the metal electrodes is avoided; and (iii) electrochemical modification or degradation of the electrode surface is prevented. Both contacts and contacless detectors integrated into a microchannel require physical connection to read out electronics placed inside or even outside the microdevice. Conductimetric designs are often very sophisticated and microfabrication facilities are required; however, simple and easy alternatives have also been proposed. These alternatives involve the deposition of conducting electrodes on the cover plate of microchips only, avoiding the clean-room laboratory infrastructure and showing significant advantages over the other approaches due to their simplicity [51].

22

Miniaturization of Analytical Systems

As already mentioned, the inherent strength of ED versus other detection modes is its compatibility with miniaturization (without loss of performance), plus advantages in microtecnology requirements (microfabrication facilities). It also constitutes, itself, an important improvement on LIF implementation, in which only one setup is possible. The possibilities in the implementation of ED can be understood as follows (see Figure 1.10): (i) analytical configurations showing complete integration of the electrochemical cell in the separation system (hard lithography); (ii) a partial integration, in which the separation system is microfabricated in one layer using soft lithography and the electrochemical detection is first deposited in another one; and (iii) the partial integration of the electrochemical cell (reference and auxiliary electrodes) with the separation system, allowing a replacement of the working electrode. In addition to LIF and ED, there are other detection approaches for analytical microsystems, but, from a realistic point of view, they are less developed than those discussed above. All of them are the focus of a recent and very interesting investigation. For instance, integrated UV and visible light absorption detectors have been tried on microchips [55]. Recently, the use of thermal lens microscopy (TLM) in combination with microchips has demonstrated an interesting analytical potential, as T. Kitamori et al. have clearly reported [56]. TLM is one of the most powerful absorption microspectrometries, and its sensitivity is comparable to LIF. TLM can be applied to both fluorescent and nonfluorescent analytes, thus expanding its versatility. Typical applications of microchip TLM include extraction, immunoassay, flow injection,

(a)

CE microchip (Glass)+ Electrochemical cell Complete integration (one piece): hard lithography

(b)

CE microchip (PDMS)

Electrochemical cell

Partial integration (2 layers): soft lithography

(c)

CE microchip (Glass or Polymer)

Electrochemical cell

Electrochemical cell + set up integration

Figure 1.10 Micromachining of electrochemical detectors integrated in microchips. (Reprinted from [101] with permission from Elsevier)

Miniaturization in Analytical Chemistry

23

enzymatic assay, in vitro bioanalysis and CE methods [56]. Complicated (bio) chemical processes can also be carried out by microchip TLM, as has recently been demonstrated [57]. Although the evolution of TLM instruments has been impressive, the nonminiaturized environment, the sophistication and the cost of such instruments are the main disadvantages of TLM microsystems. Relevant challenges in coupled detection techniques are focused on the interfaces, particularly on making the micro flows from microsystems compatible with the flow requirements of detection units. Mass spectrometry (MS) and ICP are examples of approaches in this area. MS, as one of the most powerful detection and identification techniques, has been successfully interfaced with microdevices, showing significant possibilities in the corresponding microchips from a practical point of view [55]. Some attempts have also been made to devise an interface between a chip and ESI (electrospray ionization) [58]. Some reports described an efficient sample introduction into the mass spectrometer. However, its size and price remained a challenge to its widespread use. A very interesting new interface, microchip capillary electrophoresis with inductively coupled plasma spectrometry for metal speciation, has been also reported [59,60], opening the possibility of new applications involving inorganic analytes. Detection in miniaturized analytical systems is systematically covered in Chapter 6. Separation Techniques The potential benefits of miniaturization were quickly applied to separation techniques. As M. Szumski and B. Buszewski reported, miniaturized separation systems are currently divided into ‘column’ and ‘chip’ systems [61]. The former are related to the miniaturization of column chromatographic systems, while in the latter the separation is performed in the channel of a chip device. The first group is characterized by the presence of micro- and nano-HPLC [62], whereas the second is more related to mTAS. As we mentioned before, diffusion effects are very fast in micron-sized environments. These effects have a direct advantage in separation techniques, since the reduction of the size of microchannels accelerates samplestationary phase equilibria. In general, miniaturized separation techniques such as chromatography, electrophoresis and electrochromatography have great advantages over techniques at conventional scale. The decrease of the scale provides very rapid separations, versatile channel designs, very small sample volumes and low reagent consumption. The increase of the molecular interaction also makes the chemical process highly efficient [55,63–65]. CE was one of the earliest examples of mTAS, and constitutes one of the most representative examples of an analytical microsystem. Using CE microchips, analysis times can be reduced to seconds and extremely high separation efficiencies can be achieved. Over the past decade, the most active field (as judged by publication outputs) of analytical microsystems development has been the transference from

24

Miniaturization of Analytical Systems

conventional separation techniques (macro scale) to planar chip formats. Therefore, without any doubt, CE microchips have a prominent role in the miniaturization field. In fact, they can be considered a synergic combination of miniaturized CE and microchip technological developments. On one hand, this combination involves those features derived from the analytical performance itself, such as the ability to simultaneously assay hundreds of samples in a matter of minutes (or less); rapid analysis combined with massively parallel analysis arrays, which should yield ultrahigh throughput; and the low volume of sample needed (at picolitre level), potentially prepared onboard for complete integration of sample preparation and analysis function (i.e., derivatization). On the other hand, the synergic combination conceptually involves those features derived from important aspects involved in their miniaturization. That means the easy microfabrication of a network of channels using materials of well-known chemistry, exhibiting by themselves a good electro-osomotic flow, and the possibility of using the electrokinetic phenomena to move fluid – namely valveless microdevices – and subsequently increasing their analytical possibilities. Since electrokinetics is easily applied (just a pair of electrodes is needed), EOF-driven flow has been successfully implemented using different types of material in the manufacturing channel, with glass the most commonly used. Microfabrication on polymers is faster and cheaper than that on glass, so these materials have great possibilities for mass production. In contrast, glass chip presents the best EOF, and the chemical modification on the surfaces of its microchannels is better understood than that in polymers. These features involve larger versatility in chemical analysis. The two major polymers used on chips, PDMS and PMMA, both present good optical transparency and low EOF. However, if PDMS is oxidized in a plasma discharge it presents similar EOF to glass material. Furthermore, as PDMS is an elastomer, the process of bonding substrate to the cover plate is easier, whether reversible or irreversible bonding (by oxidation in a plasma discharge) is chosen. Nowadays, PDMS is overtaking PMMA in microchip use. Another important advantage it has is the possibility of obtaining disposable CE microchips [66]. In addition, micro-CE has shown distinct advantages when compared to conventional CE, such as reduced analysis times and extremely high separation efficiencies obtained. Indeed, it has been theoretically indicated that analysis times can be reduced through a reduction in channel length or an increase in separation voltage [65]. Many of the benefits mentioned for CE microchips could equally be applied to downsized chromatographic techniques; however, the literature covers relatively few examples of chip-based chromatographic instruments compared with chipbased CE devices, even though the first analytical instrumentation on-chip was a gas chromatograph [67] and later a liquid chromatograph [68]. This is unsurprising as CE is almost perfectly suited to miniaturization, while the miniaturization of chromatographic systems involves some technical challenges, such as the microfabrication of valves and pumps, which are generally not faced in CE. On the other

Miniaturization in Analytical Chemistry

25

hand, the early on-chip liquid chromatography (LC) examples showed the potential advantages that miniaturization could provide [68]. These advantages included superior efficiency compared to conventional LC, facile positing of detection cells and low unit cost. In addition, the use of an opened tubular system instead of a normal packed column was deemed advantageous due to the shorter analysis times and low pressure drops for a given performance. The common conventional-scale alternatives of packed and tubular columns in LC have also been transferred to the microscale. Although packed columns may be desirable, the introduction of stationary phase material into microfabricated channels is not a trivial process. Frits must be fabricated within the channel structure to retain the packing and a high-pressure interface between the chip and an external pump must be applied. In addition, due to the reduced channel, the packing process is difficult to carry out and can often lead to nonuniform density particles at channel walls, reducing separation efficiencies. Consequently, the majority of initial on-chip chromatographic methods employed a tubular approach [65]. The difficulties associated with packing microfabricated channels can be eliminated if the packed bed is replaced by a porous bed formed by in situ polymerization from organic monomers. The process of bed formation is easy, since a low-viscosity monomer solution can be introduced by vacuum or pressure into the microfluidic channel prior to initiation [69]. In this way, it is important to note that electricallydriven chromatographic separations are especially attractive within chip-based systems, due to a lack of pressure gradients and reduced fabrication complexity. Due to the fact that the mobile phase runs by electro-osmosis, the flat flow profile significantly reduces band-broadening when compared to conventional LC methods. Chapter 4 presents the miniaturized systems for analyte separation based on hydrodynamic flow, while the microsystems based on electro-osmotic flow are described in Chapter 5. Since, as stated before, one of the primary advantages of micromachining analytical instrumentation is the ability to facilitate processes or fabricate structures that are extremely difficult or even impossible to recreate on the microscale, future research is focussed in this area. The concept, which can be thought as in situ micromachining, involves the creation of micron-sized particle support structures on the surface of a planar wafer [70,71]. This line of research is now underway, and is opening exciting new analytical possibilities. Sample Preparation One of the first approaches of miniaturization in sample treatment can be identified with the use of microdevices/units for performing nonchromatographic/electrophoretic separation techniques as sample treatment tools [72]. Solvent microextraction (liquid phase micro-extraction, LPME) and solid phase microextraction (SPME) procedures are two examples.

26

Miniaturization of Analytical Systems

As stated earlier, most advances have been made in separation and detection schemes. However, many of the necessary analytical steps (such as sampling and sample pretreatment) are still performed outside the chip. An important effort has been made to move them on-chip in the last few years. Thus, some excellent reviews on this topic have been published [73,74]. Although this analytical process step is less developed than the separation and detection steps, potentially it has a very exciting future. The combination of two important aspects underlines this statement: the possibility of designing complex layouts in connection with the presence of fast diffusion (statistic transport), and the simplicity of applying electrokinetics (directed transport). Table 1.3 lists the main strengths and weaknesses in the integration of sample preparation in analytical microsystems, and has been taken as the basis of the discussion of this section. Although microfabrication and microfluidics are very well adapted to many sample preparation steps, several challenges – such as miniaturization of components, direct analysis of raw samples, sampling and sample introduction – are still present. However, the good current knowledge of microfluidics and the possibility of creating sophisticated layouts can be drawn as the main technological strengths. In addition, derivatization schemes in different formats, using the well-implanted detection modes (LIF and ED), have already been successfully introduced [7,41,64]. Sample treatment Filtration, extraction/preconcentration (including clean-up in some cases) and derivatization are the most common analytical operations that can be carried out in sample preparation. Prior sample filtration is, probably, the most essential step for analyses using microfluidic systems. Raw samples and other already-treated samples (in aqueous media) require basic filtration before analysis. Due to the small dimensions of typical microstructures, particulates can cause serious operational problems, providing sites for blockage. The simplest solution is

Table 1.3 Strengths and weaknesses in the integration of the sample preparation in analytical microsystems (A) MAIN TECHNOLOGICAL SUPPORTS Microfabrication (layouts) Microfluidics (fast diffusion, electrokinetics) (B) MAIN STRENGTHS Derivatization schemes successfully introduced (pre- and post- strategies/LIF and electrochemical) (C) MAIN WEAKNESSES (challenges) Miniaturization of some components Direct analysis of real samples Representativeness of the portion of sample analysed Sampling/sample introduction in the microsystem Changes of samples within a set of analysis

Miniaturization in Analytical Chemistry

27

to filter reagents and samples prior to their introduction. Therefore, it is desirable to integrate a sample filtration system on-chip prior to the analysis. Two approaches have been employed: structurally-based filtration (controlled by the manufacturing process) and diffusion-based filtration (controlled with diffusion). Structurallybased or microfabricated filters have been proposed in popular architectures such as frits, pillar structures and flow restrictions within fluidic channels to mimic conventional filters [75]. H-filter structures have been most used in diffusion-based filtration processes [76]. Filtration can be induced by allowing the analytes of interest to migrate across a laminar boundary (between a sample and a solvent stream) while unwanted heavier particulates are retained in the original fluid stream. It is very important to underline that both approaches are principally linked to miniaturization itself: the first because of the well-known technology integrating multiple micropieces into a single microdevice; and the second because it works on the basis of an inherent property in microfluidics: presence of laminar flow. However, the limit of the structurally-based filters is the resolution limit of the manufacturing process, while the great advantage of the diffusion-based filters is that the whole process is controlled by physics, although a very sophisticated technology is also required. Liquid–liquid extraction could play a prominent role in miniaturized systems. The high surface-to-volume ratios and the short diffusion distances typical within microfluidic environments, combined with laminar flow conditions, offer the possibility of performing liquid–liquid extraction within microchannels without the need for agitation. The main challenge is to induce appropriate electro-osmotic flows when common organic solvents are used. The few examples found in the literature have also employed H-filter strategies [76]. Two approaches to performing solid phase extraction in microfluidics have been proposed: first, coating channel walls with a high-affinity stationary phase, and second, packing the microchannels with the stationary phase material [66,72,73]. The advantages and disadvantages are clearly defined in miniaturized environments. Coating-channel approaches depend on the available surface area for interaction, and in micron-sized channels the contact surface is very small. A simple way to increase the surface area is to pack the microchannels with stationary phase; however, the packing process is not easy, and this route is often avoided. A very attractive possibility, compatible with miniaturized dimensions, is to replace conventional stationary phase materials with a continuous porous bed in situ formed from polymerization of organic monomers. The process of bed formation is easy, since a low-viscosity monomer solution can be introduced by vacuum or pressure into the microfluidic channel prior to initiation [66,75]. An excellent paper showing the power of integration filtration, concentration and separation was recently published [77]. In this paper, filtration, concentration and separation are integrated on to microchip. Filtration consists of an array of seven thin channels (1 mm deep) which come together into one channel (5 mm deep). The input of the thin channels is

28

Miniaturization of Analytical Systems

communicated with sample reservoir and the sample loading is electrokinetically carried out. Sample concentration is performed by solid phase extraction. The separation channel is coated with C18 (1.5–4 mm) particles. Then, simply by increasing injection times, the analytes are retained and concentrated in the separation channel, and afterwards eluted with the appropriate solvent. In some cases, these basic operations for sample treatment have been seen as micro unit operations (MUO) [78], and used for developing microintegrated chemical systems on the basis of a concept similar to electronics: a chemical process can be constructed like an electronic circuit, but in place of the resistor, capacitor and diode, mixing, extraction, phase separation and other operations are integrated. Field-amplified sample stacking is a common method for sample preconcentration in electrophoretic systems. However, its implementation into a chip platform is not easy. The initial difficulty associated with sample stacking in a microfluidic format is the control of the analyte zone during the stacking and separation procedures. This often requires the use of relatively complex channel networks and voltage programmes to stack the analyte zone [73,74]. Derivatization on microchip is well established in analytical microsystems. The reason for this development can be found in the traditional use of LIF as a sensitive detection system, since the earliest times. Again, the role of microfabrication in the design of complex microcircuits offers a unique route in sample pretreatment and especially in derivatization schemes, where the process is carried out before or after analyte separation and before analyte detection. In other words, the high degree of functional integration (reagent mixing, product separation and post-column labelling) provides an elegant indication of the potential benefits of microfluidic systems. Derivatization schemes carried out prior to the separation (precolumn) or immediately prior to the detection (post-column), have been proposed using LIF and ED detections. Thus, derivatization in fluorescence has been well implemented into microchip in connecting with LIF detection [79]. A group of works uses a very attractive strategy based on the combination of suitable pre- and post-channel layouts with bioreagents such as enzymes and antibodies, along with electrochemical detection [80,81]. It is important to remark that, in some cases, these are truly lab-on-a-chip devices [82,83]. In fact, the generation of electroactive products using selective bioreagents such as specific enzymes, class-enzymes and antibodies in adequate pre- and post-column derivatization schemes has revealed the potential of these strategies in the simplification of treatments of complex samples. This line can be used for developing other sophisticated layouts for sample treatment when real samples are analysed in these microsystems. Sampling and sample introduction Sample introduction and the representativeness of the small size of the sample introduced are two important aspects of microfluidic systems. Most chip-based systems still adopt a discrete and often

Miniaturization in Analytical Chemistry

29

manual approach to on-chip sample loading and sample changing. Such practice seriously lowers the overall throughput and counteracts the advantage of achieving fast separations. On-chip loading and changing of a series of samples has become a limiting factor in real-world applications. Efficient assays of real-world samples will require the incorporation of a continuous sampling capability (from the external environment) or rapid sampling of multiple discrete samples. Such an ability to continuously introduce real samples into micro-sized environments should make analytical microsystems compatible with real-life applications. Different devices, based on pressure-driven [84–86], hydrodynamic [87] and autonomous polymerloading [88] techniques have been reported for sample injection. Finally, one of the most important aspects to consider is the continuous introduction of samples, which requires an interfacing system between the macroenvironment (mL–mL) and the microchip (nL–pL) in order to achieve an efficient sample change with low carryover. Hence, a suitable design of interfaces, similar to the hyphenated detection techniques, is necessary. Although bibliography is poor in papers dealing with this subject, various approaches have been proposed in excellent works. In one, using a microfluidic matrix device, sample introduction was performed through a wide-bored microchannel, drilling an access hole through the sample reservoir [89]. Very recently, the use of a sharp sampleinlet tip was reported [90]. An excellent report dealing with trends relating to different configurations of sample introduction interfacings was recently written by Fang, showing a critical vision of different strategies used by his research group [91]. This report shows that the trend is to separate the sample-introduction channel (SIC) with hydrodynamic flow from the separation channel with electrokinetic flow, thus avoiding totally pressure-driven flows. This was achieved by introducing a sample-loading channel (SLC) between the SIC and the separation channel. The SIC acts as a flow-through sampling reservoir and the sample is electrokinetically loaded into the SLC from the SIC. The SLC connects with the separation channel. This split-flow strategy is improved by a flow-through sampling reservoir featuring a guided overflow design. This array avoids Poiseuille flows, maintaining equal liquid levels for the sample, the buffer and the waste reservoirs. A fast and simple sample introduction was recently reported [88]. It consists of the use of a sharp sample-inlet tip alternately placed in the sample and the buffer vials. This tip was fabricated by sharpening the inlet side of the chip with a diamond saw. A good reproducibility was obtained (100 repetitive flow injection measurements resulted in a response with an RSD of 3.7%). C.-W. Huang and G.-B. Lee have proposed an interesting contribution consisting of microautosamplers for discrete sample injection and dispensation in microchips [92]. It is a promising approach toward realizing the mTAS concept. All these aspects related to the miniaturization of sample treatment are reported in more depth in Chapter 3.

30

Miniaturization of Analytical Systems

1.7.2

Integrated (Micro)systems for the Performance of the Entire Analytical Process

As Figure 1.11 shows, an ideal analytical measurement system should be characterized by: (i) Portability and self-operation, in order to avoid or reduce the sampling step and allow the possibility of performing field tests. (ii) Either the noninvasive measurement approach or the performance of in-line measurements. These features allow the minimum perturbation of the sample. (iii) Facilities for the maintenance of the system and large lifetimes. Alternatively, reusability at low price. (iv) Self-calibration incorporated in the system. (v) The possibility for the incorporation of quality-control activities, in order to assure a reliable response. Sensors were one of the first approaches to performing the entire analytical process. They represent by themselves a miniaturization and simplification of the analytical process, although the initial expectation of a wide field of application was demonstrated to be unfounded by limitations in selectivity and lifetime. But in any case, sensors constitute a growing area of interest with a strong social and industrial impact. In fact, sensors provide fast response with reduced costs. Ideally, a sensor is a device that provides an analytical signal for a specific compound present in a raw sample in a direct, reversible, continuous and reliable manner. From a practical point of view, most of the sensors reported in the literature fall short of the requirements of the above definition. Hence, a more realistic definition of a sensor is that it is a sensitive microzone where a (bio)chemical reaction occurs that is connected to or integrated in a physical (optical, electrochemical, thermal, masssensitive) transducer [93]. Through this connection to an instrument, analytical information can be produced in situ in real time. The problem of interferences is partially solved by the so-called ‘biosensors’, which are defined by IUPAC as selfcontained integrated devices that are capable of providing specific quantitative or semiquantitative analytical information using a biological recognition element (biochemical receptor) in direct contact with a transducer element. Biosensors including transducers based on integrated circuit microchips are known as ‘biochips’. The biological recognition element used in a biosensor can be an enzyme, an antibody/antigen, a nucleic acid, a cellular structure or a biomimetic receptor. The term ‘microsensor’ describes a sensor whose active sensing area is in the micro range, while the sensing device itself is much larger for easier handing. ‘Microelectrode’ and ‘ultramicroelectrode’ are terms used in electrochemistry used instead of microsensor, although electrodes with smaller radii (nanometric size) are currently used [94]. The small size of these electrodes makes diffusional mass

Figure 1.11 Different approaches to the implementation of the entire analytical process, combining various degrees of automation, simplification and miniaturization (see text for details)

Miniaturization in Analytical Chemistry 31

32

Miniaturization of Analytical Systems

transport extremely efficient and double-layer capacitance very small, attaining greater signal-to-noise ratios and lower detection limits than those of macroelectrodes [95,96]. Chapter 7 reports the miniaturization of the entire analytical process through micro(nano)-sensors. The TAS approach is addressed to meet these ideal features, integrating the different analytical standard operations of the analytical process. Portable instruments can be considered as a particular case of TAS, designed to operate outside the laboratory (miniaturized portable analytical systems are reported in Chapter 9). Both are integrated analytical systems, with a high level of automation and simplification for the intended analyses. Miniaturization commonly exists in these systems, but in a relative sense, with respect to other laboratory alternatives using conventional equipment to carry out the same application. On the other hand, the following step giving integrated microsystems (mTAS or in vivo measurement devices) is strongly characterized by miniaturization, although simplification and automation are implicit in these microsystems. Chapter 8 is devoted to the miniaturization of the entire analytical process through mTAS approaches. As illustrative examples, Figure 1.12 shows four common categories of integrated systems miniaturizing the entire analytical process. The first one (Figure 1.12a)

Figure 1.12 Examples of different levels of miniaturized analytical systems: (a) portable X-ray fluorescence equipment (Reprinted from [97] with permission of American Chemical Society, Copyright 2007); (b) small analyser for SPR with optical biosensor detection (Reprinted from [98] with permission from Elsevier); (c) portable miniaturized equipment for health care (Reproduced from [99] with permission of IEEE, Copyright 2007); (d) mTAS for micro-ELISA analysis (Reproduced from [78] with permission of IEEE, Copyright 2007)

Miniaturization in Analytical Chemistry

33

is a portable total reflection X-ray fluorescence spectrometer. Despite the macrosize (cm level), it represents a clear miniaturization with respect to a conventional X-ray fluorescence spectrometer operating at laboratory-routine work [97]. The second category (Figure 1.12b) is represented by small-sized laboratory equipment used for the determination of environmental organic pollutants based on a surface plasmon resonance (SPR) optical biosensor device [98]. The third (Figure 1.12c) is a type of portable miniaturized equipment for health care used for home medical diagnosis. It corresponds to a calorimetric three-item measurement chip (triglyceride, total-cholesterol and HDL) combined with an electronic blood-collection system (6 mL of whole blood) [99]. The final category (Figure 1.12d) is represented by mTAS, as the multichannel micro-ELISA chip in the image shows [78]. From a practical point of view, the reliability of the information generated by these integrated microsystems is a key factor. Performance tests for the proper microsystem and the appropriate validation of the whole analytical method must be carefully carried out. Within these activities, calibration must be planned in tandem with the miniaturized equipment work, especially in the case of portable systems. In this context, the integration of the calibration with the particular manifold or scheme of the microsystem should be taken into account. Major difficulties can be found in miniature analysis instruments for measuring trace levels of gases, commonly used for field tests. In these cases, standard gas in situ generation for onsite calibration is a very useful alternative. Thus, recently S.-I. Ohira et al. described one of these systems for on-site checking or onsite calibration of a micro gas analysis system, mGAS [100]. The key part of this arrangement is shown in Figure 1.13. The source reagent (R1 in Figure 1.13) was mixed with the desorbing solution (R2) and reacted in the mixing coil (MC), then introduced into a microchannel gas generator. Generated vapour permeated through the membrane and was extracted into the air stream flowing on the opposite side of the membrane. This microsystem was used to generate H2S, SO2, CH3SH and NH3.

1.8

Outlook

Analytical microsystems constitute an important trend within analytical chemistry science. The high number of scientific contributions published in this field so far is a good reason to imagine that one of the next objectives will be dealing with application issues and expanding their use to laboratories at bench level. Miniaturization presents different difficulties and problems regarding the different steps involved in the analytical process [101]. Detection schemes, such as LIF and ED (amperometry), are well established in analytical microsystems; nevertheless, novel designs, especially in ED, are being developed. Although MS was studied from earliest times because it is a very powerful technique [102,103], its use in mTAS is nonetheless not widely proposed due to its size and price. Recently, F. Foret and

34

Miniaturization of Analytical Systems

Figure 1.13 Liquid flow system for gas generation with microchannel gas desorber. R1: source reagent; R2: desorbing solution; MP: liquid minipums; MC: mixing coil; GD: gas desorber; BP: back pressure tube. (Reprinted from [100] with permission from Elsevier)

P. Kus
y published a review on microfluidics for multiplex MS analysis [104]. As the authors recognize, the development of microfluidics–MS coupling is still in its infancy, despite the analytical potential of these microsystems. They pointed out that ‘the future success will depend on the willingness of both the instrument manufacturers and the users to adopt this technology in practice’. This will require a clear demonstration of the technology’s superiority and robustness. Regarding the detection capabilities, LIF and amperometry are not very suitable for small molecules, including inorganic species. Alternative detection schemes, such as conductivity, ICP-AEE or ICP-MS, are necessary and consequently research on new detection systems for analytical microsystems is still open. Additionally, the use of TLM for detection and imaging studies performed in microchips will open interesting new possibilities. Separation on chips is primary developed using CE, which constitutes the most representative example of mTAS. In microseparation, new tools such as the design of new analysis schemes and strategies including the use of nanoparticles, employment of specially designed polymers and new cover materials will be developed. Sample preparation at microscale level has been less developed than both detection and separation techniques. In general, some of the most relevant promises of lab-on-a-chip devices, such as integration of all laboratory functions in order to get practical sample treatments, commercialization, truly handheld operation and ease of use, are coming, but today are not a reality. However, the possibilities of using electrokinetics as directed fluid transport, of creating suitable layouts for

Miniaturization in Analytical Chemistry

35

problematic needs, and of laminar flow, allow us to think of future developments as a near reality. Total integration [105] and, especially, world-to-chip interfacing are considered the major challenges, particularly in high-throughput applications requiring frequent sample change, such as online continuous process monitoring. More specifically, various world-to-chip interfacing schemes are being developed to meet different requirements of sample and microfluidic chips, in terms of sampling rate, sample introduction, sample consumption, precision, stability and degree of automation. Two clear trends are the development of more integrated, rugged, portable and fully automated sample introduction systems, and systems based on various hydrodynamic principles outside of the electrokinetic environment. On the other hand, a basic trend with respect to analytical applications is the exploration of the possibilities of these microsystems in other important analytical areas, ranging from medical diagnosis and pharmaceutical screening to food and environmental control [106]. This expanded use can only be fulfilled by versatile and highly sophisticated analytical devices. One exciting trend in miniaturization is the use of micro/nanodevices for biological and bioanalytical studies of cells, genes, proteins and individual molecules. Different alternatives are available for cell trapping in microfluidic chips [107], demonstrating that miniaturization technology provides facilities for creating tools with feature sizes matching the dimensions of cells, and enables integration of cell-handling and fluid-manipulation elements and the chemical analysis of single cells [108]. Also, microsystems create new opportunities for the spatial and temporal control of cell growth. Further, as recently reported by J. El-Ali et al. [109], typical unit operations such as cell growth, treatment, selection, lysis, separation and analysis have been demonstrated to be implemented using mTAS. But, as they pointed out, robust approaches to fabrication, integration and packaging (such as communication with the macroenvironment) remain major areas of research. Gene analysis on a single chip is now a possibility, although with some challenges [110]. Microfabricated electrophoresis devices offer several advantages over conventional methods for rapid genotyping [111] and DNA sequencing [112]. Proteomic analysis based on microfluidics is also the subject of interesting developments. Electrophoresis devoted a special issue to this topic in 2006 [113], and other interesting contributions have appeared since. In an exciting article, H. Craighead discussed future lab-on-a-chip technologies for interrogating individual molecules [114]. As he reported, these advances allow the manipulation and measurement of individual molecules, and ‘the adaptation of these approaches to lab-on-a-chip formats is providing a new class of research tools for the investigation of biochemistry and life processes’. It can be said that the first generation of lab-on-a-chip devices is starting to work, and in the future, successive (more sophisticated) generations will introduce themselves as a new reality in analytical laboratories. In this new generation, a further integration of sample collection and sample preparation, improving

36

Miniaturization of Analytical Systems

lab-on-a-chip for world connections in order to reach the goal of ‘total integration’, will be necessary. New approaches to avoid the major shortcomings of chip-based analytical systems, such as the risk of the analyte adsorption on walls and at interfaces (which is especially high in low-volume analytical systems), and the optical interference at the walls of chips in the detection point, will help to solve practical problems. The concept of digital microfluidics in the late 1990s can be addressed in this way, as it involved the manipulation of discrete volumes of liquids on a surface by mechanisms such as electowetting, thermocapillary transport and surface acoustic wave transport [115]. Sample levitation can open the exciting new approach of lab-on-a-drop, representing an alternative for preparing nano-pico sample volumes without contamination from solid walls or other external objects, or between samples [116].

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

A. Manz, N. Graber, H.M. Widmer, Sens. Actuators, B1 (1990) 244. P.S. Dittrich, K. Tachikawa, A. Manz, Anal. Chem., 78 (2006) 3887. P.A. Greenwood, G.M. Greenway, Anal. Chem., 21 (2002) 726. B.H. Weigl, R.L. Bardell, C.R. Cabrera, Adv. Drug Del. Rev., 55 (2003) 349. O. Geschke, H. Klank, P. Tellesmann (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. B. Bhushan (Ed.), Handbook of Nanotechnology, Part A, Springer, Heidelberg, 2007. M.A. Schwarz, P.C. Hauser, Lab Chip, 1 (2001) 1. K. Uchiyama, H. Nakajima, T. Hobo, Anal. Bioanal. Chem., 379 (2004) 375. L. Szekely, R. Freitag, Electrophoresis, 26 (2005) 1928. P. Woias, Sens. Actuators B, 105 (2005) 28. M.S. Anderson, Anal. Chem., 77 (2005) 2907. D.B. Weibel, M. Kruithof, S. Potenta, S.K. Sia, A. Lee, G.M. Whitesides, Anal. Chem., 77 (2005) 4726. H. Erxleben, J. Ruzicka, Anal. Chem., 77 (2005) 5124. M.A. Burns, B.N. Johnson, S.N. Brahmasandra, K. Hanique, J.R. Webster, M. Krishnan, T.S. Sammarco, P.M. Man, D. Jones, D. Heldsinger, C.H. Mastrangelo, D.T. Burke, Science, 282 (1998) 484. E.T. Lagally, I. Medintz, R.A. Mathies, Anal. Chem., 73 (2001) 565. D.R. Reyes, D. Iossifidis, P.-A. Auroux, A. Manz, Anal. Chem., 74 (2002) 2623. S.C. Terry, J.H. Jerman, J.B. Angell, IEEE Trans. Electron Devices, ED-26 (1979) 1880. B. van der Schoot, P. Bergveld, Sens. Actuators, 8 (1985) 11. A. Manz, Y. Miyahara, J. Miura, Y. Watanabe, H. Miyagi, K. Sato, Sens. Actuators, B1 (1990) 249. A. Manz, N. Graber, H.M. Widmer, Sens. Actuators, B1 (1990) 244. A. Manz, J.C. Fettinger, E. Verpoorte, H. L€ udi, H.M. Widmer, D.J. Harrison, Trends Anal. Chem., 10 (1991) 144. J.C. Fettinger, A. Manz, H. Ludi, H.M. Widmer, Sens. Actuators B, 17 (1993) 19. A. Manz, D.J. Harrison, E.M.J. Verpoorte, J.C. Fettinger, A. Paulus, H. Ludi, H.M. Widmer, J. Chromatogr., 593 (1992) 253. D.J. Harrison, A. Manz, Z.H. Fan, H. Ludi, H.M. Widmer, Anal. Chem., 64 (1992) 1926. S. Cowen, D.H. Craston, Proc. Micro Total Analysis Systems, (1994) 295. A. Feustel, J. Muller, V. Relling, Proc. Micro Total Analysis Systems, (1994) 299.

Miniaturization in Analytical Chemistry [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70]

37

D. Belder, Anal. Bioanal. Chem., 385 (2006) 416. D. Belder, M. Ludwing, L.-W. Wang, M.T. Reetz, Angew Chem. Int. Ed., 45 (2006) 2463. R.M. Johann, Anal. Bioanal. Chem., 385 (2006) 408. J. El-Ali, P.K. Sorger, K.F. Jensen, Nature, 442 (2006) 403. J.P. Kutter, O. Geschkes, In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004, p. 213. M. Valc
arcel, A. R
ıos, Anal. Chim. Acta, 400 (1999) 425. A. R
ıos, D. Barcel
o, L. Buydens, S. C
ardenas, K. Heydorn, B. Karlberg, K. Klemm, B. Lendl, B. Milman, B. Neidhart, R. Stephany, A. Townshend, A. Zschunke, M. Valc
arcel, Accred. Qual. Assur., 8 (2003) 68. M. Valc
arcel, S. C
ardenas, Trends Anal. Chem., 24 (2005) 67. A. Manz, J.C.T. Eijkel, Pure Appl. Chem., 73 (2001) 1555. A. Manz, D.J. Harrison, E. Verpoorte, H.M. Widmer, Adv. Chromatography, 33 (1993) 1. K.B. Mogensen, H. Klank, J.P. Kutter, Electrophoresis, 25 (2004) 3498. M.A. Schwarz, P.C. Hauser, Lab chip, 1 (2001) 1. K. Uchiyama, H. Nakajima, T. Hobo, Anal. Bioanal. Chem., 379 (2004) 375. W.R. Vandaveer IV, S.A. Pasas, R.S. Martin, S.M. Lunte, Electrophoresis, 23 (2002) 3667. J. Wang, Talanta, 56 (2002) 223. J. Tanyanyiwa, S. Leuthardt, P.C. Hauser, Electrophoresis, 23 (2002) 3659. A.J. Blasco, A. Escarpa, In: M.L. Marina, A. R
ıos and M. Valc
arcel (Eds), Analysis and Detection by Capillary Electrophoresis, Elsevier, Amsterdam, 2005, p. 703. R.S. Keynion, T.J. Roussel, Jr., M.M. Crain, D.J. Jackson, D.B. Franco, J.F. Naber, K.M. Walsh, R.P. Baldwin, Anal. Chim. Acta, 507 (2004) 95. J. Wang, G. Chen, M.P. Chatrathi, M. Musameh, Anal. Chem., 76 (2004) 298. Y. Liu, J.A. Vickers, C.S. Henry, Anal. Chem., 76 (2004) 1513. N.A. Lacher, S.M. Lunte, R. Scott Martin, Anal. Chem., 76 (2004) 2482. P. Ertl, C.A. Emrich, P. Singhal, R.A. Mathies, Anal. Chem., 76 (2004) 3749. R.M. Guijt, E. Baltussen, G. van der Steen, R.B.M. Schasfoort, S. Schlautmann, H.A.H. Billiet, J. Frank, G.W.K. van Dedem, A. van den Berg, Electrophoresis, 22 (2001) 235. M. Galloway, W. Stryjewski, A. Henry, S.M. Ford, S. Llopis, R.L. McCarley, S.A. Soper, Anal. Chem., 74 (2002) 2407. M. Pumera, J. Wang, F. Opekar, I. Jel
ınek, J. Feldman, H. L€ owe, S. Hardt, Anal. Chem., 74 (2002) 1968. J. Wang, M. Pumera, Anal. Chem., 74 (2002) 5919. J. Wang, G. Chen, A. Muck, Jr., Anal. Chem., 75 (2003) 4475. X. Bai, Z. Wu, J. Josserand, H. Jensen, H. Schafer, H.H. Girault, Anal. Chem., 76 (2004) 3126. C.P. Palmer, V.T. Remcho, Anal. Bioanal. Chem., 372 (2002) 35. T. Kitamori, M. Tokeshi, A. Hibara, K. Sato, Anal. Chem., 76 (2004) 52A. M. Goto, K. Sato, A. Murakami, M. Tokeshi, T. Kitamori, Anal. Chem., 77 (2005) 2125. K. Uchiyama, H. Nakajima, T. Hobo, Anal. Bioanal. Chem., 379 (2004) 375. Q.J. Song, G.M. Greenway, T. McCreedy, J. Anal. At. Spectrom., 18 (2003) 1. Q.J. Song, G.M. Greenway, T. McCreedy, J. Anal. At. Spectrom., 19 (2004) 883. M. Szumski, B. Buszewski, Crit. Rev. Anal. Chem., 32 (2002) 1. J.P. Chervet, M. Ursem, J.P. Salzmann, Anal. Chem., 68 (1996) 1507. V. Dolnik, S. Liu, S. Jovanovich, Electrophoresis, 21 (2000) 41. G.J.M. Bruin, Electrophoresis, 21 (2000) 3931. A. de Mello, Lab Chip, 2 (2002) 48N. S.A. Soper, S.M. Ford, S. Qi, R.L. McCarley, K. Kelly, M.C. Murphy, Anal. Chem., 72 (2000) 643A. S.C. Terry, J.H. Jerman, J.B. Angell, IEE Trans Electron Devices, 26 (1979) 1880. A. Manz, Y. Miyahara, J. Miura, Y. Watanabe, H. Miyage, K. Sato, Sens. Actuators, B1 (1990) 249. F. Svec, J.M.J. Frechet, Anal. Chem., 64 (1992) 820. B. He, N. Tait, F. Reignier, J. Pharm. Biomed. Anal., 17 (1998) 925.

38 [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116]

Miniaturization of Analytical Systems B. He, N. Tait, F. Reignier, Anal. Chem., 70 (1998) 3790. L. Ramos, J.J. Ramos, U.A.Th. Brinkman, Anal. Bioanal. Chem., 381 (2005) 119. J. Lichtenberg, N.F. de Rooij, E. Verpoorte, Talanta, 56 (2002) 233. A.J. de Mello, N. Beard, Lab Chip, 3 (2003) 11N. B. He, L. Tan, F. Reignier, Anal. Chem., 71 (1999) 1464. B.H. Weigl, P. Wager, Science, 283 (1999) 346. B.S. Broyles, S.C. Jacobson, J.M. Ramsey, Anal. Chem., 75 (2003) 2761. T. Kitamori, Transducers & Eurosensors ’07, Lyon, France, 2007, pp. 11–16. A.G. Hadd, S.C. Jacobson, J.M. Ramsey, Anal. Chem., 71 (1999) 5206. J. Wang, M.P. Chatrathi, B. Tian, Anal. Chem., 72 (2000) 5774. J. Wang, M.P. Chatrathi, A. Ib
a~nez, A. Escarpa, Electroanalysis, 14 (2002) 400. J. Wang, A. Ib
a~nez, M.P. Chatrathi, A. Escarpa, Anal. Chem., 73 (2001) 5323. J. Wang, A. Ib
a~nez, M.P. Chatrathi, J. Am. Chem. Soc., 125 (2003) 8444. M. Tabuchi, M. Ueda, N. Kaki, Y. Yamasaki, Y. Nagasaki, K. Yoshikawa, K. Kataoka, Y. Baba, Nat. Biotechnol., 22 (2004) 337. D. Solignac, M.A.M. Gijs, Anal. Chem., 75 (2003) 1652. N.Y. Lee, M. Yamada, M. Seki, Anal. Sci., 20 (2004) 483. U. Backofen, F.M. Matysik, C.E. Lunte, Anal. Chem., 74 (2002) 4054. T. Ito, A. Inoue, K. Sato, K. Hosokawa, M. Maeda, Anal. Chem., 77 (2005) 4759. C. Lin, G. Lee, S. Chen, Electrophoresis, 23 (2002) 3550. G. Chen, J. Wang, Analyst, 129 (2004) 507. Q. Fang, Anal. Bioanal. Chem., 378 (2004) 49. C.-W. Huang, G.-B. Lee, Electrophoresis, 26 (2005) 1807. R. Kellner, J.-M. Mermet, M. Otto, M. Valc
arcel, H.M. Widmer, Analytical Chemistry, WileyVCH, Weinheim, Germany, 2004, Chapters 29 and 33. S. Alegret, Integrated Analytical Systems, Comprenhensive Analytical Chemistry Vol. XXXIX, Elsevier, Amsterdam, 2003. A.J. Bard, L. Faulkner, Elctrochemical Methods: Fundamentals and Applications, 2 Ed., John Wiley and Sons, Ltd, New York, 2001. R.J. Forster, Chemical Society Reviews, 23 (1994) 289. S. Kunimura, J. Kawai, Anal. Chem., 79 (2007) 2593. E. Mauriz, A. Calle, A. Montolla, L.M. Lechuga, Talanta, 69 (2006) 359. Y. Horiike, H. Ogawa, M. Nagai, H. Koda, C.-H. Chang, S. Hashioka, M. Takai, Y. Morimoto, Transducers & Eurosensors ’07, Lyon, France, 2007, pp. 347–350. S.-I. Ohira, K. Someya, K. Toda, Anal. Chim. Acta, 588 (2007) 147. A. R
ıos, A. Escarpa, M.C. Gonz
alez, A.G. Crevill
en, Trends Anal. Chem., 25 (2006) 467. S. Benetton, J. Kameoka, A.M. Tan, T. Wachs, H. Craighead, J.D. Henion, Anal. Chem., 75 (2003) 6430. N. Sillon, R. Baptist, Sens. Actuators B, 83 (2002) 129. F. Foret, P. Kus
y, Electrophoresis, 27 (2006) 4877. S. Alegret, Integrated Analytical Systems, Comprehensive Analytical Chemistry Vol. XXXIX, Elsevier, Amsterdam, 2003, Chapter 1. A. R
ıos, A. Escarpa, B.M. Simonet, M. Zougagh, Analytical Microsystems for Complex Matrices, DEStech Publications, Lancaster, USA, in press. R.M. Johann, Anal. Bioanal. Chem., 385 (2006) 408. A.K. Price, C.T. Culbertson, Anal. Chem., 79 (2007) 2614. J. El-Ali, P.K. Sorger, K.F. Jensen, Nature, 442 (2006) 403. J. Alper, Anal. Chem., 79 (2007) 809. E. Sz
antai, A. Guttman, Electrophoresis, 27 (2006) 4896. Y. Shi, Electrophoresis, 27 (2006) 3703. C.S. Lee, Electrophoresis, 27 (2006) 3501. H. Craighead, Nature, 442 (2006) 387. R.B. Fair, Microfluid Nanofluid, 3 (2007) 245. F. Priego, M.D. Luque de Castro, Trends Anal. Chem., 25 (2006) 856.

2 Tools for the Design of Miniaturized Analytical Systems 2.1

Introduction

This chapter deals with the tools needed to design analytical microsystems at micro and nano scale, according to the classification shown in Figure 1.6. Micro(nano) fluidics and micro(nano)electromechanical systems (MEMS and NEMS, respectively) are the characteristic scenarios in which the analytical process occurs at a miniaturized scale. Miniaturization at this scale is completely different from miniaturization at the mini scale, of which some examples were given in Chapter 1. Many of the phenomena that we are used to living with in the macro world, and which we take for granted, have next to no significance for fluids in the micro world. For instance, inertia and turbulent flow play an important role in the macroscale world (transport in conventional pipes and tubes, rivers, oceans, etc.) but mean nothing at micro- and nanofluidic scales. By contrast, diffusion (having practically no transport significance at the macro flow scale) becomes the dominant process at micro(nano)fluidic scales. In addition, surfaces become an ever more important factor as dimensions are reduced. In fact, the ratio of surface to volume increases drastically as this happens. The objective of this chapter is to give the basic principles of the behaviour of microfluidic systems for an analytical intended use, as well as the main microfluidic devices needed to develop the analytical process at this downsizing scale. For further information about the engineering side of microfluids, readers should consult more specialized books and publications (see e.g. [1–4]). Figure 2.1 shows a simplified scheme of the main steps from planning to the final analytical use of Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

40

Miniaturization of Analytical Systems

DEFINITION OF ANALYTICAL PROBLEM

TECHNICAL SPECIFICATIONS

MICROSYSTEM DESIGN

SIMULATION AND OPTIMIZATION

FABRICATION: (1) PROTOTYPE (2) MASSIVE FABRICATION

ANALYTICAL USE

PERFORMANCE TEST

Computerized software

Cleanroom facilities Microfluidic devices Microelectronic facilities

Verification of technical specifications NO

OK OK

ANALYTICAL VALIDATION

NO

Figure 2.1 The main steps to be followed in the development of an analytical microsystem, and the different tools and devices involved

a specific microsystem. This figure also summarizes the main points to be described in this chapter: (i) design of the microsystem according to some previous technical specifications (connected to the analytical problem to be solved); (ii) simulation and optimization by use of an appropriate computerized software; (iii) the proper fabrication as a previous prototype using cleanroom, microfluidic and microelectronic facilities for micromachining and final integration by packaging; and (iv) technical verification of the prototype and, if it passes, analytical validation for its intended use. If the expected technical and analytical purposes are achieved, massive fabrication can be planned if needed. As stated in the previous chapter, the two most important phenomena characterizing microfluidic devices are the almost exclusive presence of laminar flow and the dependence on diffusion as the major available transport mechanism.

2.2 Miniaturized Analytical Processes: The Downsizing and Integrating Phenomena Fluid behaviour has been studied extensively for several centuries, and an important number of monographs and articles have been published on this subject. At macro

Tools for the Design of Miniaturized Analytical Systems

41

and mini scales, density, the pressure of the liquid and its viscosity are the characterizing parameters. Knowing density and viscosity, other predictions can be made for liquids in motion, particularly about what kind of flow will be found in microsystems (meaning the flow in channels at typical dimensions on the micro scale). The dimensionless Reynolds number equation, Re, establishes the relation between the magnitudes of the inertial and the viscous forces: Re ¼

rrh v ; h

ð2:1Þ

as was introduced in Chapter 1. Numbers larger than about 2300 correspond to turbulent flows. Under this regime, inertial forces are dominant (this is the behaviour we typically know in everyday life). According to Equation 2.1, large Reynolds numbers are attained at higher liquid densities, higher flow velocities, larger typical length scales or lower viscosities. A Reynolds number between 2000 and 3000 is in the so-called regime of transitional flow, and the region below about 2000 is referred to as the laminar flow regime. This type of regime is obtained at lower velocities, smaller dimensions, smaller densities or higher viscosities. It can be demonstrated that flows in capillaries with dimensions smaller than 100–150 mm need a rather impossible pumping velocity, n, value in order to reach the transitional flow regime, and hence the laminar flow regime is the characteristic situation in this microfluidics. For a capillary with a cylindrical cross-section, following the Hagen–Poiseuille theory, the volume flow, Q, is given by the expression: Q¼

DV pR4 DP; ¼ 8hL t

ð2:2Þ

where R is the radius of the capillary, L is its length and DP is the pressure drop across this length (also called hydraulic pressure). The velocity profile (i.e. the velocities n(r) at different radial positions between the centre (r ¼ 0) and the wall (r ¼ R)) is given by the equation: vðrÞ ¼ ðR2 r 2 Þ

DP ; 4hL

ð2:3Þ

clearly describing a parabolic flow profile. The term 8hL=pR4 (the reciprocal appears in Equation 2.2) is also called the fluidic resistance. Therefore, the fluidic resistance dramatically increases as the channel dimensions are reduced, and hence higher pressure drop values are necessary to move liquid through smaller conduits. For channels with noncylindrical cross-sections, similar expressions to Equation 2.2 can be found [5]. Another useful relation for the design of microfluidic channels is based on the continuity equation, describing the behaviour of flow in channels with changing cross-sections. This equation states that the product of the cross-sectional area and the flow velocity is a constant: A1 v1 ¼ A2 v2 ¼ const:

ð2:4Þ

42

Miniaturization of Analytical Systems

On the other hand, Bernoulli’s equation also considers pressure and height differences in flows, and plays a key role as a direct application of the law of energy conservation, relating pressure, kinetic energy and potential energy according to the equation: 1 Pþ rv2 þrgh ¼ const: 2

ð2:5Þ

It is very useful to know the value of n because it can give an indication of the transit time through a microfluidic channel network, and therefore helps to assess whether there is enough time for a chemical reaction or mixing to take place. 2.2.1

Transport within Microfluidic Systems

The motion of a fluid in a channel can be driven by internal forces that are determined by fluid and channel properties, such as interfacial forces, or by an external field (electric, thermal, magnetic, etc.). At the micro and nano scales the relative importance of forces is given by the following order: buoyancy < inertial force or gravitation force < viscous force  interfacial force. Therefore, interfacial forces become prominent upon system downscaling. There are two different types of transport within microfluidic systems, depending on the nature of the driving agent behind the transport: directed transport and statistical transport. Directed transport is transport that is controlled by exerting work on the fluid. This results in a volume flow of the fluid (characterized by a direction and a flow profile). When the work is generated by a pump, a pressuredriven flow is produced, whereas if the work is generated by a voltage, the so-called electro-osmotic flow is established. By contrast, statistical transport is not a directly controlled transport, as it is based on an entropy-driven transport. Thus, the transport occurs only if a fluid is more disordered after transport than before. This is the case with transport by diffusion: movement of molecules from more concentrated zones to less concentrated ones in a liquid medium. In practice, both types of transport take place, particularly because commonly directed transport meets with some gradients in the flow, such as a temperature gradient and a concentration gradient. By definition, heat transfer by mass transport is called convention. Free or natural convention is caused by a difference of temperature, but fluids can also be moved with external forces to create a directed flow (forced convention). There are several ways in which forced convention can generate directed flow, such as capillary flow or other forces (gravity, pressurized air, etc., which create a single instant of pressure difference in the microsystem). Migration is the directed transport of molecules in response to an electric field. This is the original basic force involved in (micro)capillary electrophoresis and its different modes. The electrically charged molecules, in a polar solvent medium (aqueous

Tools for the Design of Miniaturized Analytical Systems

43

solutions, mainly), experience a Coulomb force, F, due to the electric field: F ¼ qE;

ð2:6Þ

where q is the charge on the molecule and E the strength of the electric field. When the terminal velocity of the ions is reached, the Coulomb force is balanced by the Stokes force: F ¼ 6phrv;

ð2:7Þ

where Z is the viscosity of the liquid, r the hydrodynamic radius of the molecule and n is the speed of the charged molecules. At the terminal speed, both forces are equalized: qE ¼ 6phrn;

ð2:8Þ

q E; 6phr

ð2:9Þ

and hence: n¼

where the mobility of the molecules, m, is given by: q m¼ 6phr

ð2:10Þ

Diffusion is a pure statistical transport process. Thus, in a liquid or gas medium, all molecules move in all directions as long as no external forces are applied. This phenomenon takes place when there is a concentration gradient of one kind of molecule within a fluid. The statistical movement of a single molecule in a fluid is a random movement characterized by the Einstein–Smoluchowski equation: pffiffiffiffiffiffiffiffi x ¼ 2Dt; ð2:11Þ where x is the average distance moved after an elapsed time t between molecule collisions, and D is a diffusion constant characteristic for each molecule. If x is half of the channel width, W, it is possible to estimate the time a molecule takes to cross this distance: tcross ¼

W2 8D

ð2:12Þ

The consequence of this equation is the key influence of the channel width on the mixing of liquids inside it. On the other hand, the overall effect of the random walk of all molecules can be described as temporal and spatial changes in concentration. These changes are governed by Fick’s laws. The fact that molecules travel from areas of high molecular concentration to areas of low concentration is expressed by using the concept of particle flux, j (the number of molecules crossing a certain area during

44

Miniaturization of Analytical Systems

a time t):

@c ð2:13Þ @x This equation means that the molecular flux is stronger when the concentration gradient is steeper. Additionally, it can be stated that when molecules leave a test volume, the number of molecules in the test volume is correspondingly lower. Thus, the continuity equation can be expressed as: j ¼ D

@c @j ¼ ; @t @x

ð2:14Þ

and therefore, by combining Equations 2.13 and 2.14: @c @2c ¼ D 2; @t @x

ð2:15Þ

which is Fick’s second law of diffusion, showing that the variation of concentration with time is directly proportional to the diffusion coefficient of the molecules. D is affected by different variables in a (micro)fluidic environment, according to the Stokes–Einstein equation: kT D¼ ; ð2:16Þ 6phr where k is the Boltzmann constant and T the temperature. The denominator is commonly known as the frictional constant (depending on the viscosity of the medium and the hydrodynamic radius): f ¼ 6phr

ð2:17Þ

It is important to remark that under different circumstances and for different geometries of microfluidic systems, either directed or statistical flow can dominate, or both types may be of equal importance. To evaluate the various flow situations, the ratio between the mass transport due to directed flow and that due to diffusion can be estimated. This ratio gives a dimensionless number called the P
eclet number (Pe), which is expressed by: vd Pe ¼ ; ð2:18Þ D where d is a characteristic length of the microfluidic system. Thus, the P
eclet number, for a particular microchannel, can be calculated, and the diffusion compared to the directed flow can be evaluated. This is important information for the design of microfluidic systems that need to keep control of diffusion, such as chemical separation systems. If the P
eclet number is much smaller than 1 then diffusion dominates the microfluidic flow (directed flow is of secondary importance), whereas if the P
eclet number is much larger than 1, the molecules mainly flow according to the externally applied driving force (diffusion has only a minor influence).

Tools for the Design of Miniaturized Analytical Systems

45

Another term used for transportation in minichannels is dispersion, commonly applied to flow-injection analysis (FIA) systems. In this case, a sample plug containing a high concentration of certain molecules moves along channels of 300–700 mm (lower for micro-FIAs). The concentration boundaries become vaguer due to diffusion of molecules, and the widening of the sample plug is generally called dispersion. In addition to the band broadening caused by diffusion, there is another bandbroadening causedbythe parabolicvelocityflow profile, whichoccurswhen the flow is pressure-driven. This superposition phenomenon is called Taylor dispersion. 2.2.2

Microsystem Design from Transport Parameter Information

The previously defined dimensionless P
eclet number is useful in designing microsystems. In microsystems, the flow velocities are usually small, and the crucial variable that determines the P
eclet number, according to Equation 2.18, is therefore the channel length, d. For long enough channels, the P
eclet number is always larger than 1, and the flow is consequently directed. On the other hand, if the P
eclet number is much smaller than the length-to-width ratio of the microchannel, d/w, then Taylor dispersion is observed. By contrast, if the P
eclet number is much larger than this ratio, then diffusion is not the main agent of dispersion. Thus, narrow microchannels allow transverse diffusion to play a significant role, which does not happen in relatively broad channels. Therefore, microsystem designers have to decide the length, width and flow speed. They then have to adjust the microchannel volume by choosing the depth of the channels. As the two most important phenomena found in microfluidic devices are the almost exclusive presence of laminar flow and the dependence on diffusion as the major available transport mechanism, the possibilities for developing microfluidic devices come from the application of these physical principles to the practice. These aspects will be considered in the next section of this chapter.

2.3

Microfluidic Devices

Different devices are used to generate and control the hydrodynamic flow in microfluidics. Among the most important devices are pumps and valves, but other elements such as injection/metering devices, mixers, heaters, etc., are commonly integrated in the microfluidic layout. Devices dealing with the detection are also key elements in microfluidics. Although detection in microsystems is considered in this chapter, it will be systematically described in Chapters 6 and 7. 2.3.1

Microvalves

Valves are designed to manipulate the direction of the motion of the fluid, and they are seldom combined with micropumps in microfluidics. Basically, they introduce

46

Miniaturization of Analytical Systems

directionality into the flow, and according to their type of work can be classified as passive or active valves. The ideal characteristics of microvalves can be summarized by the following parameters: zero leakage, zero power requirements, zero dead volume, high differential pressure capability, zero response time, chemical resistance and not being sensitive to particulate contamination. Of course, any valve meets all these requirements, and the selection of the appropriate valve for a particular use becomes a key factor. A large number of different valves are commonly used in microfluidic systems, as reviewed by K.W. Oh and C.H. Ahn [6]. Following these authors, valves can be classified in two main categories, as Table 2.1 shows. Passive microvalves can use mechanical or nonmechanical moving parts, whereas active microvalves can use both mechanical or nonmechanical moving parts and external systems. Different physical principles are involved in the functioning of both categories of valves, given a variety of types of microvalve, as reported in Table 2.1.

Table 2.1 Classification of microvalves Category

Mode

Base of actuation

Type

Passive

Mechanical

Check valve

Nonmechanical

Capillary

Mechanical

Magnetic

Flag Membrane Ball In-line mobile structure Diffuser Abrupt Liquid-triggered Brust Hydrophobic valve External magnetic fields Integrated magnetic inductors Electrostatic Electrokinetic Piezoelectric Bimetallic Thermopneumatic Shape memory alloy Bistable Electrochemical Hydrogel Sol-gel Parafin Electrorheological Ferrofluids Built-in Rotary Membrane In-line

Active

Electric Piezoelectric Thermal Nonmechanical

Bistable Electrochemical Phase change Rheological

External

Modular Pneumatic

Tools for the Design of Miniaturized Analytical Systems

47

Passive valves do not need any external energy for their function. Those with moving parts cannot be opened or closed without changing the geometry of flow paths, such as with check valves, pressure control valves or flap valves. Other passive, nonactuated, valves have no moving parts, and they are the more interesting for combination with autonomous capillary systems [7]. A simple passive valve can be built by using a cantilever in which a thin strip of silicon bends when enough pressure is applied from one side (valve flap type) [8]. The simple work of this check valve is schematically shown in Figure 2.2. It is based on mechanical action and does not require any external energy for its operation (only flow pressure in the microchannel). There are also passive valves not based on mechanical action but using other forces, such as surface-tension or capillary valves [9,10], hydrophobic valves [11] and pH-sensitive valves [12], among others. These types of valve are designed for a single use, because their function depends on an air–liquid interface, which is typically present only when first filling a microchip. Indeed, they influence how a liquid fills the region of the valve, but not how a liquid afterward flows through the valve. The simpler version of this approach is a restriction in the microchannel. Liquid cannot penetrate into the next channel segment because of the increased surface tension around the restriction. Only when a higher pressure is applied will the liquid break through the restriction into the next section of the channel network. Other basically passive valves are based on the use of hydrogels, which are sensitive to some external stimulus. Commonly, they suffer a change in their volume. Depending on the stimulus generated by the solution flowing through the channel, the hydrogel swells or stays in the shrunken state. As many of these

Figure 2.2

Scheme of a simple passive microvalve based on a cantilever

48

Miniaturization of Analytical Systems

hydrogels are pH-sensitive, this chemical stimulus can easily be used. Recently, H.J. Cho and coworkers have developed passive valves for microfluids based on the use of a nanostructured functional polymer surface [13]. They have fabricated and tested two fully integrated microfluidic valves, one with a superhydrophobic polymer surface, and the other with a switchable, thermosensitive polymer surface. The passive valve with the superhydrophobic polymer surface selectively inhibits the flow of water-based reagents and passes aqueous solutions containing surfactants. In the case of the thermosensitive valve, the switchable polymer surface becomes hydrophobic when heated to temperatures exceeding 65  C, thus inhibiting the flow of water, and becomes hydrophilic at room temperature, thus allowing the flow of water. Figure 2.3(a) shows the general scheme of the microvalve, while Figure 2.3(b) is a photograph of the fabricated structure, micromachined from the designed scheme. In these figures, the reservoirs for the inlet and outlet microchannels can be seen, as can the location of the polymer film. The same design was used for both polymers. The microchannels were fabricated by standard

Figure 2.3 Scheme of a passive valve for microfluids based on the use of a nanostructured functional polymer surface (In, inlet; O1, outlet 1; O2, outlet 2) (a) and the corresponding photograph (b). The fabrication of a passive microfluidic valve with a superhydrophobic surface (see text for details) (c) and the fabrication process of a thermosensitive valve (see text for details) (d). (Reprinted from [13] with permission from Elsevier)

Tools for the Design of Miniaturized Analytical Systems

49

photolithography and wet-etching techniques, while the polymer surface for both the valves was fabricated using the layer-by-layer deposition technique, in which multiple layers of polyelectrolytes (poly(allylamine hydrochloride), PAH) were coated in a channel wall, followed by silica nanoparticle treatment. For the thermosensitive valve, the polymer surface was further coated with the thermosensitive polymer poly(N-isopropylacrylamide) (PNIPAAm). Figure 2.3(c) shows the steps followed for the fabrication of the passive microfluidic valve with the superhydrophobic surface: (i) A glass slide is lithographically patterned to form the T-junction microchannel shown in Figures 2.3(a) and (b). (ii) The patterned glass slide is wet etched by using the photoresist as an etch mask. (iii) The glass slide is photolithographically patterned to form an opening for the initial polyelectrolytes. (iv) Positively charged PAH is dip-coated on the opening. (v) Negatively charged poly(acrylic acid) (PAA) is then dip-coated. This layerby-layer deposition is continued successively until 101 bilayers of PAH and PAA are formed. (vi) A poly(dimethylsiloxane) (PDMS) slab is bonded to the glass slide after oxygen plasma treatment. For the fabrication of the thermosensitive valve, the stages involved are schematically shown in Figure 2.3(d): (i) An etched glass slide with a lithographically patterned opening for the initial polyelectrolites is dip-coated with positively charged PAH. (ii) Negatively charged silica nanoparticles are then dip-coated. This layerby-layer deposition is continued successively until 40 bilayers of PAH and silica nanoparticles are formed. This is followed by a second run of layerby-layer deposition to decorate the polymer surface with silica nanoparticles. (iii) The glass chip with the rough polymer surface is annealed at 400  C for 2 hours. (iv) After a third run of layer-by-layer deposition to create two bilayers of PAH and PAA, the glass chip is dipped into the initiator solution for 2 hours. Finally, PNIPAAm is grafted on to the initiator-coated surface to create the switchable, thermosensitive polymer surface. Both passive valves were successfully tested for passing/stopping liquid samples depending on their wettability. Active valves, by contrast, require an actuator to provide the mechanical action, allowing or stopping the flow passage. They are normally designed to be in one or other of these states (‘open’ or ‘closed’) without actuation, while energy must be applied to pass or maintain the opposite state. Therefore, actuators are decisive

50

Miniaturization of Analytical Systems

Figure 2.4 Illustrations of the actuation principles of active microvalves with mechanical moving parts: (a) electromagnetic; (b) electrostatic; (c) piezoelectric; (d) bimetallic; (e) thermopneumatic; (f) shape memory alloy actuation. (Reprinted from [10] with permission of Elsevier, Copyright 2007)

elements. They can be based on different principles: pneumatic, electrostatic, piezoelectric, electromagnetic, etc., but are typically characterized by four features: (i) the pressure they can build up; (ii) the stroke displacement they generate; (iii) their response time; and (iv) their reliability. Each type of actuator presents better characteristics in one feature or the other. Figure 2.4 illustrates the actuation principles widely employed in active microvalves based on mechanical principles. Most of these active microvalves couple a flexible membrane to some actuation methods (magnetic, electric, piezoelectric, etc.). Traditionally, these active microvalves are accomplished using MEMS-based bulk or surface micromachining technologies. Active microvalves based on nonmechanical operation are of particular interest in terms of their simple device structure and disposability, making them well suited for applications in life sciences. External active microvalves are actuated by the aid of external systems such as built-in modular [14,15] or pneumatic means [16,17]. Many other microvalves can be developed using different designs, materials and actuation principles. Thus, in addition to the examples described, others can be found in the literature using different principles. G.M. Whitesides and coworkers have described torque-actuated valves for controlling the flow of fluids in microfluidic channels [18]. These valves consist of small machine screws (about

Tools for the Design of Miniaturized Analytical Systems

51

500 mm) embedded in a layer of polyurethane cast above microfluidic channels fabricated in PDMS. Turning the screws actuates the valve by collapsing the PDMS layer between the valve and the channel, controlling the flow of fluids in the underlying channels. These valves do not require power to retain their setting (on/off), and additionally they allow setting between ‘on’ and ‘off’. They can be integrated into portable, disposable microfluidic devices. J. Wang et al. recently proposed the use of nanowires for the fabrication of switchable microchip devices [19]. They used functional nickel nanowires for switching on-demand operation of microfluidic devices. Controlled reversible magnetic positioning and orientation of these nanowires at the microchannel outlet offers modulation of the detection and the separation processes, respectively. These devices have been used in electrophoretic separations in microchips. 2.3.2

Moving Liquids in Miniaturized Systems

Pumps are devices to set fluids into motion. Many of these devices have been developed based on a large variety of principles. Interesting reviews can be found for wider information about pumping devices in microfluidics [20–22]. In this section we have selected some representative examples, and Chapter 4 covers the miniaturized systems based on a hydrodynamic flow. A special case is pumping fluids by electro-osmosis, which is briefly described in this chapter and developed in more detail in Chapter 5, where the miniaturized systems based on an electro-osmotic flow are reported. Following recent reviews on micropumping techniques, pumping mechanisms can be classified as shown in Table 2.2. Two main categories can be identified. The first is mechanical displacement micropumps, defined as those that exert oscillatory or rotational pressure forces on the working fluid through a moving solid–fluid (vibrating diaphragm, peristaltic, rotary pumps) or fluid–fluid (ferrofluid, phase change, gas permeation pumps) boundary. The second category, nonmechanical micropumps, covers electro- and magnetokinetic micropumps, defined as those that provide a direct energy transfer to pumping power and generate constant/steady flows by the continuous addition of energy (electro-osmotic, electrohydrodynamic, magnetohydrodynamic, electrowetting, etc.) [21]. Micropumps in the above categories can be further divided into subcategories based on their actuation principle. As A. van den Berg et al. have stated [23], the flow of fluids within enclosed channels of glass, silicon or plastic can be actuated spontaneously, nonmechanically or mechanically. As the hydrodynamic resistance increases due to confinement, the most frequently used pumping method (syringe pumping) is increasingly difficult to use in micro(nano)channels due to the reduction in dimension. Mechanical Micropumps These are truly microfabricated pumps, used for moving liquids in microsystems. They need to meet different basic requirements to assure accurate analytical results:

52

Miniaturization of Analytical Systems

Table 2.2 Classification of micropumps with different actuation methods Category

Mode of operation

Type

Mechanical micropumps

Diaphragm

Electrostatic Piezoelectric Electromagnetic Thermal Pneumatic Polymer materials Peristaltic Ferrofluid Phase change Gas boundary Rotating gear Viscous force Induction Injection Polarization Ion drag DC AC

Fluid Rotary Nonmechanical micropumps

Electrohydrodynamic

Electroosmotic Magnetohydrodynamic Electrowetting Others

. . . .

Flexural plate wave Optoelectrostatic Bubble type

stability, with little pulsation; controllable flow rate at a preset range; resistance to aggressive reagents and high temperatures; reasonable cost.

Sometimes these requirements are extremely hard to meet, or imply the substitution of the traditional silicon by alternative materials for the microfabricated pumps. Normally, these pumps use the same type of actuators as active valves. The combination of the actuators chosen for the pumping and the valves determines the overall performance, in terms of generated pressure, stroke displacement, response time and reliability. The most common version of a mechanical micropump consists of a pump chamber and two valves: an inlet valve and an outlet valve. The pump chamber is mostly a combination of a diaphragm or membrane and an actuator to displace the membrane or diaphragm. This displacement, together with the action of the valves, results in alternating flux into and out of the chamber (pumping action). Both passive and active microvalves can be used in these arrangements. These types of micropump employ many different actuation mechanisms, as Table 2.2 shows. Figure 2.5 gives some selected examples of mechanical displacement micropumps.

Tools for the Design of Miniaturized Analytical Systems

53

Figure 2.5 Examples of different types of mechanical micropumps: (a) vibrating diaphragm micropump; (b) peristaltic micropump; (c) dynamic, active diaphragm valve; (d) principles of operation of different rotating microvalves. (Reprinted from [21] with kind permission from Springer)

The first (Figure 2.5(a)) is the scheme of a simple diaphragm displacement micropump, as described above. During the expansion stroke, the pumping chamber expands, resulting in a corresponding decrease in chamber pressure. When the inlet pressure is higher than the chamber pressure, the inlet valve opens and liquid fills the expanding chamber. During the compression stroke, the volume of the chamber decreases with the moving diaphragm, causing the internal pressure to increase, whereby liquid is discharged through the outlet valve. The different types of actuation mechanism reported in Table 2.2 can be used to vibrate the diaphragm. One of these possibilities is a peristaltic mechanism. Pumps based on this principle incorporate the peristaltic motion of actuators in series to generate pumping action (Figure 2.5(b)). Thus, they can be considered a subset of the vibrating diaphragm pumps, utilizing the different types of transducer (piezoelectric, pneumatic, etc.). When the first diaphragm is actuated, it restricts the flow to the inlet of the pump. As the second diaphragm is actuated, fluid is pushed toward the third pumping chamber. Similarly, actuating the third diaphragm in succession pushes the fluid through the outlet of the pump. This sequence is continuously repeated for pumping

54

Miniaturization of Analytical Systems

action from left to right. Recently, L.-S. Jang and Y.-Ch. Yu have presented a driving system for a peristaltic micropump that is based on piezoelectric actuation [24]. This micropump with an integrated driving system is portable and has the potential to be integrated with other components, such as biosensors. Y.-C. Hsu et al. have developed an interesting peristaltic antithrombogenic micropump for in vitro and ex vivo blood transportation tests [25], very useful for studies in clinical environments. Dynamic-geometry valves can also be used as pumping microdevices, as they provide flow direction by deformation, motion or deflection. Figure 2.5(c) shows the schematic functioning of such active valves working as pumping devices. Traditional rotary micropumps constitute another category of mechanical micropump, in which a toothed gear rotating in a fluid chamber with an inlet and an outlet port produces the fluid flow. However, there are several additional micropumps in which the fluid is driven with a rotating component (internal or external). Figure 2.5(d) shows different possibilities based on rotating-gear and viscous-force pumping mechanisms. Typically, rotating-gear micropumps drive the finned or toothed gear with an electric motor for rotation. Fluid becomes entrapped between the gear teeth while turning and thereby is transported from the inlet to the outlet position. Fluid displacement using viscous forces generated by a rotating component has been also developed, using different configurations, as Figure 2.5(d) indicates. Eccentric placement of a rotating shaft in a straight channel is one of the possibilities. When the cylinder rotates, a net force is transferred to the fluid due to unequal shear stress on opposite sides of the rotor. Spiral-channel viscous pumps operate in a similar way to Couette flow, in that the movement of the boundary induces viscous stress on the fluid near the wall. As a final example (Figure 2.5(d)), disc viscous micropumps have been developed, in which the flow is generated by the rotation of a disc, which also acts as a boundary to the channel flow. Many other micropumps based on mechanical operation modes have been described. Table 2.3 briefly summarizes the main actuation forces involved in the most common mechanical micropumps. Nonmechanical Micropumps Nonmechanical micropumps require the conversion of nonmechanical energy to kinetic energy to supply the fluid with momentum. This phenomenon is practical only at the microscale level. In contrast to mechanical micropumps, nonmechanical pumps generally have neither moving parts nor valves, so geometry design and fabrication techniques in this type of pumping in microsystems are relatively simpler. These micropumps directly convert electrical and magnetic forms of energy into fluid motion. Since these pumping processes occur in a continuous manner, the resulting flow is generally quite constant. Whereas electrokinetic pumps often utilize an electric field to pull ions within the pumping channel,

Based on the Coulomb attraction force between oppositely charged plates. The membrane is forced to deflect in either direction as appropriate voltage is applied on the two opposite electrostatic plates located on each side. A piezoelectric disc attached on a diaphragm, a pumping chamber and valves. Actuated by the deformation of the piezoelectric materials. The actuation involves the strain induced by an applied electric field on the piezoelectric crystal. A chamber which is full of air inside is expanded and compressed periodically by a pair of heater and cooler devices. The periodic change in volume of chamber actuates the membrane with a regular movement for fluid flow. Consists of soft magnetic cores activated by currents in energized coils, or of permanent magnets. Electromagnetic actuation requires an external magnetic field, usually in the form of a permanent magnet. Based on the differences in thermal expansion coefficients of materials. When dissimilar materials are bonded together and subjected to temperature changes, thermal stresses are induced, providing a means of actuation. The diaphragm is made of two different metals, which exhibit different degrees of deformation during heating. Micropumps make use of the shape memory effect in SMA materials such as titaniumnickel. The shape memory effect involves a phase transformation between two solid phases. Actuated by stress gradient by ionic movement due to electric field. ICPF is composed of polyelectrolyte film, with both sides chemically plated with platinum. Due to the application of an electric field, the cations included in the two sides of the polymer molecule chain will move to the cathode. At the same time, each cation will take some water molecules to move towards the cathode. This ionic movement causes the cathode of the ICPF to expand and the anode to shrink. When there is an alternating voltage signal, the film bends alternately.

Electrostatic

Ion-conductive polymer film (ICPF)

Shape memory alloy (SMA)

Bimetallic

Electromagnetic

Thermopneumatic

Piezoelectric

Foundation

Type of actuation

Table 2.3 Common actuation forces involved in mechanical micropumps

[47,48]

[45,46]

[44]

[40–43]

[37–39]

[30–36]

[26–29]

Selected references

Tools for the Design of Miniaturized Analytical Systems 55

56

Miniaturization of Analytical Systems

magnetokinetic pumps typically utilize the Lorentz force on the bulk fluid to drive the microchannel flow. Due to these characteristics, in strict terms these ‘pumps’ are mechanisms to produce the movement of flow in microchannels without the existence of pumping devices as in mechanical micropumps. Electrohydrodynamic pumping This modality utilizes electrostatic forces acting on dielectric liquids to generate flow. The different types of electrohydrodynamic pumping (Table 2.2) are based on the method by which the charged particles are introduced into the fluid. Thus, induction pumping requires a gradient in either the electrical conductivity or the permittivity of the working fluid. This is typically achieved by anisotropic fluid heating or by discontinuities in properties, which occur for layers of nonmixing fluids or suspended particles in the fluid. Alternating voltages are imposed on the electrodes present on the boundary of the fluid channel. These voltages vary in time, creating a travelling wave that moves through the working fluid, perpendicular to the gradient in conductivity. By contrast, in micropumping by injection, electrochemical reactions at the electrodes cause the injection of free ions into the bulk liquid. These ions experience Coulomb forces due to the presence of the electric field. This causes the movement of ions, which in turn carry the bulk fluid with them. In the polarization type of micropumping, a nonhomogeneous electric field is used through the working fluid to create a variation in the fluid electric field density, causing the pumping action. Electro-osmotic flow Electro-osmotic flow (EOF) [49] is the liquid flow originating in the presence of an electric field when an ionic solution is in contact with a charged solid surface. Oppositely charged ions in the fluid shield the surface charge and can be manipulated with a DC or AC electric field. DC electro-osmotic pumping is normally used, particularly for electrophoretic separtions [50], and is described in more detail in Chapter 5. Therefore, only a basic treatment is given in this section. To give a well-known example, in a silica material in contact with an aqueous electrolytic solution the solid surface has an excess of negative charge due to the ionization of the surface’s silanol groups. A high number of counterions of these anions are found on the interphase between the capillary wall and the solution originating the so-called electric double layer, which is formed by a stagnant layer adjacent to the capillary wall (Stern layer) and a mobile diffuse layer (Gouy– Chapman layer) (see Figure 2.6). The cationic counterions in the diffuse layer migrate toward the cathode. As they are solvated, they drag solvent with them, originating the EOF. The linear velocity of the EOF, veo, depends on the potential originated across the double layer and the so-called zeta potential, z, through the following equation: veo ¼

eEz ; 4ph

ð2:19Þ

where e is the dielectric constant of the solution. Since e, z and Z depend on many properties of solids and liquids, a great number of experimental variables control the

Tools for the Design of Miniaturized Analytical Systems

57

Figure 2.6 Formation of the double layer and generation of the EOF. C, electrical potential; z, zeta potential. (Copyright 2000 Agilent Technologies. Reproduced with permission)

EOF. The extremely small size of the double layer leads to flow at the walls of the capillary, resulting in a flat profile which enables high separation efficiencies with respect to those separation techniques with hydrodynamic flow profiles (pressuredriven techniques). EOF modifies the migration of a species, since if it moves in the same direction as EOF its velocity will increase, and if it moves against the EOF its velocity will decrease. Therefore, it is possible to create a flow by filling a microchannel with a buffer solution and applying a suitable voltage at the channel ends. This is an easier way to move fluids in microchannels than the use of micropumps, but the particular flow profile resulting from EOF is another interesting advantage over hydrodynamic flows. In fact, contrary to hydrodynamic flows characterized by a parabolic distribution of the flow velocities, EOF is generated close to the wall. Therefore, it produces a plug-like profile with a very uniform velocity distribution across the entire cross-section of the channel. Probably the main disadvantage of EOF is its strong dependence on the chemical factors involved in the microsystem. This dependence makes EOF hard to control, as any change affecting zeta potential has a direct effect on the EOF. The use of buffers allows some control of EOF. EOF can also be manipulated by changing the pH and by coating the walls of the capillary. Thus, at low pH, electro-osmotic flow reduces in magnitude, and it can be suppressed totally for lower pH values; the permanent or periodical coating of internal walls of the capillary with nonchargeable compounds cancels the

58

Miniaturization of Analytical Systems

electro-osmotic flow. Other major limitations of this pumping technique are the high voltage and the electrically conductive solution required. A wide variety of examples using DC electro-osmotic flow can be found. C.H. Chen and coworkers have described planar electro-osmotic micropumps [51,52], while L. Chen et al. have reported the fabrication and characterization of a multistage electro-osmotic micropump [53]. Other authors have developed microsystems with packed particles or monoliths using high-pressure electro-osmotic pumping [54,55]. Recently, F.-Q. Nie et al. developed a robust, compact, on-chip, electro-osmotic micropump for microflow analysis based on parallel, encased, monolithic silica capillary columns [56]. AC electro-osmotic flow has emerged as a viable microscale pumping mechanism for conductive or electrolytic solutions. Unlike the deprotonation on the channel surface for DC-EOF, electrodes positioned on the channel boundary provide the charge to establish the electric double layer in the AC case. The advantage of AC-EOF pumping is that relatively high velocities can be achieved with less than 10 V [57,58]. In addition, flow reversal can be achieved, making this approach bidirectional [59,60]. B.P. Cahill et al. have used this modality of EOF with interdigitated electrodes for portable lab-on-a-chip systems [61]. Other nonmechanical micropumping techniques Electrowetting involves wettability change due to applied electric potential. In electrowetting, the fluid is transported using surface tension (an interfacial force which dominates at microscale). Voltage is applied on the dielectric layer, decreasing the interfacial energy of the solid and liquid surface, which results in fluid flow. Thus, when applying the electric potential along the interface of a liquid metal (mercury, generally) droplet in an electrolyte, charge redistribution occurs, resulting in a gradient in surface tension at the interface, which causes movement of the droplet to regions of lower surface tension. Switching the direction of the applied potential also changes the direction motion. F. Mugele and J.C. Baret have reviewed electrowetting, including a discussion of the challenges of this technique and a review the state of the art [62]. Furthers works on models of droplet manipulation with electrowetting on flat and rough surfaces have also been carried out by V. Bahadur and S.V. Garimella [63,64]. The pumping effect in bubble-type micropumps is based on the periodic expansion and collapse of a bubble generated in microchannel. This modality always needs to be heated, so the application scope is limited in that heating may not always be desired or allowed. Geng et al. reported a bubble-based micropump for electrically conducting liquids [65], while Zahn et al. reported microneedles integrated with an on-chip MEMS bubble micropump for continuous drug-delivery applications [66]. In another article, Z. Yin and A. Prosperetti reported data obtained on a simple micropump based on the periodic growth and collapse of a single vapour bubble in a microchannel [67].

Tools for the Design of Miniaturized Analytical Systems

59

Other, less common nonmechanical micropumping techniques are based on optoelectrostatic microvortex [68] and the use of flexural plate wave pumps [69]. A recent, more in-depth evaluation and discussion of all these micropumping technologies can be found in the review published by B.D. Iverson and S.V. Garimella [21], which includes exhaustive tables showing the main developments reported in the literature. The review published by A. Nisar et al. is mainly addressed to biomedical applications [22]. A systematic study of multiphase flow in microand nanochannels, taking special interest in multiphase confined fluid manipulation, has been reported [23] (systems in nature are rarely either single-phase or nonconfined). 2.3.3

Mixers

Mixing is one of the challenges in microfluidic devices. The goal is to obtain a homogeneous blend of two solutions in as little time as possible. Passive and active mixers can be distinguished by the strategy used to achieve the homogenous mixture. Passive mixing on microdevices depends solely on diffusion as the transport mechanism. Hence, to improve mixing using this alternative, a reduction of the diffusion distances or the introduction of some kind of extra movement needs to be considered. For active mixers on the other hand, external energy is used to induce turbulence. As C.-C. Chang and R.-J. Yang stated [70], electrokinetic mixing represents one of the most practical alternatives for mixing analytes and reagents in microfluidic systems, due to the difficulties derived from the low Reynolds number flows in microscale devices. Electrokinetic mixing approaches can also be categorized as either active or passive in nature (Figure 2.7). Active mixers use either time-dependent (AC or DC field switching) or time-independent (DC field) external electric fields to achieve mixing, while passive mixers achieve mixing in DC fields simply by virtue of their geometric topology and surface properties, or electrokinetic instability flows. Passive Mixers The simplest examples of passive mixers are T- and Y-junctions, where two different streams meet. At the macroscopic level, the mixture of the two streams is almost instantaneous, because of the presence of a turbulent flow. Nevertheless, in microchannels, the only lateral transport mechanism that allows the mixture is diffusion (Figure 2.8). Although different parameters affect diffusion, the main design parameters for effective mixtures are the width of the main channel (resulting channel) and the linear flow velocity. Thus, the worst-case diffusion distance corresponds to half the width channel, whereas the flow velocity determines the contact time between the two streams and therefore translates into a minimum length of the main channel at any given flow velocity. By an opposite approach, it is possible to select the conditions (channel width, flow velocity and channel length)

60

Miniaturization of Analytical Systems

Figure 2.7 Classification scheme for microfluidic mixing based on electrokinetic mixing. (Reprinted from [70] with kind permission from Springer)

in which hardly any mixing occurs and in which the two streams flow almost unperturbed side by side down the channel. Hence, for practical purposes, channel width, flow velocity and main channel length are decisive parameters for designing microfluidic systems. However, the fabrication of narrow and deep channels requires special machining technology, and they introduce a higher fluidic resistance. One possible strategy is to work with standard-sized channels, split each channel into an array of smaller channels and then merge them again in such a way that the split flows of one solution get interlaced with the split flows of the other. Electrokinetically passive mixing is enhanced by the particular geometry topologies, surface properties or instability phenomena which occur naturally under a static (DC) electric field. According to the classification shown in Figure 2.7, electrokinetic passive mixing can be categorized as either lamination or chaotic mixing [70]. In the former case, mixing mainly relies on the molecular diffusion effect between two or more parallel streams, while the latter usually involves the interfacial contact area between two mixing streams. This area is greatly increased and the diffusion length is reduced by means of the repeated stretching and folding

Tools for the Design of Miniaturized Analytical Systems

61

Figure 2.8 Scheme of two mixer/filtering microdevices. (a) a T-junction where two liquids meet and flow downstream side by side, slowly mixing by diffusion only (laminar flow); (b) mixingfiltering microdevice in which laminar flow and diffusion are involved

of the fluids in a microchannel. The previously mentioned T- or Y-shaped mixing configurations are, in fact, a type of laminar mixing based on multiple parallel electro-osmotic streams (mixing by molecular diffusion) [71,72]. More recent developments based on this principle can be found in the literature [73–75]. In contrast with multistream lamination mixing, split-and-recombine lamination mixing relies on a multistep procedure, in which three basic steps are required [76]: fluid element splitting, recombination and rearrangement. The three steps can be completed when the two mixing streams flow through an out-of-plane microchannel structure. Grooved surface-enhanced electro-osmotic mixing was proposed by Stroock et al. [77]. It consists of a chaotic mixing channel with patterned grooves for a pressure-driven flow system. Transverse flows can be produced in the grooved channels, which results in a helical flow motion at a low Reynolds number regime. On the other hand, a heterogeneous surface charge patterning can produce an enhancement in electro-osmotic mixing. This principle has been used by different authors for electrokinetic passive mixing in a variety of applications [78–81]. Further, the electrokinetic instability phenomenon produced in microfluidics [82] can be used as another passive mixing alternative. Despite limitations found with this modality, several uses have been reported [83–85].

62

Miniaturization of Analytical Systems

Active Mixers Whereas passive mixers rely solely on energy from within the system to facilitate mixing, active mixers can be designed in which external energy of some form induces chaotic behaviour to accelerate the mixture. This modality refers to the enhancement of mixing in electrokinetically-driven microfluidic systems using a time-dependent electric field, or in pressure-driven flow systems by means of an externally time-dependent or -independent electrical force. Chaotic mixing can be achieved by several different methods (Figure 2.7). In general, the choice of driving amplitudes and frequencies, and optimization of operation conditions, is a challenge in designing an active micromixer. Under appropriate operation conditions, the majority of electrokinetic active mixing schemes in the literature are able to induce chaotic mixing, with the exception of periodically-switching electro-osmosis mixing enhancement. Dielectrophoretic-based mixing refers to the polarization of a particle or biological element relative to a suspension medium in a nonuniform electric field which results in motion [86]. Positive dielectrophoretic forces move particles to regions with a higher electric field, while the particles are repelled from the electrode edges to a lower electric field by negative dielectrophoretic forces. Positive or negative dielectrophoretic force is only induced by applying the specific range of the frequency of the AC field. Electrowetting-on-dielectric droplet-based mixing is based on the wetting properties of a droplet in contact with an insulated electrode, which can be altered by means an electric field. If an electric field is applied nonuniformly on an electrode array, a surface energy gradient is induced, which can be used to manipulate the motion of the droplet [87,88]. Micromixing by electrohydrodynamic instability generally refers to the force generated by the interaction between electric field and net charge density near the solid–liquid interfaces. A microfluidic mixer has been developed by El Moctar et al. utilizing the electrohydrodynamic instability phenomenon [89]. Oddy et al. first presented an active micromixer in which an AC electric field induces a more chaotic flow field to enhance the mixing of two pressure-driven flow streams [82]. A highintensity AC electric field from side channels was still required to induce electrokinetic instability to enhance mixing in this work. Many researchers have investigated a periodically switching electro-osmotic mixing approach for an efficient active mixing [90–93]. It is performed by in some way pulsing the velocities or pressures of two pressure-driven flow streams working at low Reynolds number. Field-induced electro-osmosis enhanced mixing, presenting various modalities, has also been widely applied to the active control of electroosmotic flows and liquid pumping in microchannels, but only scarcely applied in micromixing [94,95]. A novel strategy for mixing solutions and initiating chemical reactions in microfluidic systems has been proposed by A.N. Hellman et al. [96]. This method utilizes highly focused nanosecond laser pulses form a Q-switched Nd:YAG laser

Tools for the Design of Miniaturized Analytical Systems

63

to generate cavitation bubbles within 100 and 200 mm-wide microfluidic channels containing the parallel laminar flow to two fluids. The bubble expansion and subsequent collapse within the channel disrupts the laminar flow of the parallel fluid streams and produces a localized region of mixed fluid. This approach to generating the mixing of adjacent fluids may present advantages in many microfluidic applications, as it requires neither tailored channel geometries nor the fabrication of specialized on-chip instrumentation. 2.3.4

Volume-dispensing and Sample-introduction Devices

A very important function in microfluidics is the ability to dispense very welldefined small amounts or volumes of solutions. In practice, this means introducing a volume whose size is not exactly known but which is always the same, or a volume whose size is exactly known and is given by a particular geometric arrangement of channel segments. The latter case is called ‘metering’. For many analytical techniques it is necessary to inject small, well-defined amounts of sample solutions. Popular implementations of an aliquoting or injection function on microchips make use of a channel cross or a double T arrangement, along with electrokinetic fluid manipulation, to achieve this objective. Section 5.5, in Chapter 5, deals with this method of introducing very small volumes into microchips. An overview of developments in high-throughput microfluidic sample-introduction techniques based on a capillary sampling probe and a slotted-vial array has recently been published [97]. These sample introduction systems have received attention in miniaturized flow injection analysis, sequential injection analysis, capillary electrophoresis and liquid–liquid extraction. Introduction of samples to a microfluidic chip requires an interfacing system between samples with volumes in the ml–ml range from the macro world, and microfluidic systems handling volumes in the nl–pl range. Without any doubt, world-to-chip interfacing is considered a major challenge. Q. Fang et al. [97] summarize the different possible modes of sample introduction for microfluidic chips (Figure 2.9). As these authors explain,

Figure 2.9 Different designs for microfluidic world-to-chip interfacing. (Reprinted from [97] with permission from Elsevier)

64

Miniaturization of Analytical Systems

in mode A, sample and reagent solutions are loaded in the on-chip reservoirs connected with microchannels. This approach has been employed frequently in most reported chip-based analytical systems. It presents a high degree of integration, but the sample-change operation, which usually requires the sample reservoir to be manually emptied, rinsed and refilled with a new sample solution, is timeconsuming and difficult to automate. Mode B (Figure 2.9) is characterized by employing a split-flow interface fabricated on the chip to achieve continuous sample introduction. Different sample solutions are sequentially delivered to the chip via connecting tubing through the split-flow interface. Various designs have been reported based on this interfacing model [98–100], allowing a high throughput, and continuous and automated sample introduction. Mode C (Figure 2.9) involves the use of an on-chip sampling probe for continuous sample introduction, which replaces the on-chip sample reservoir in mode A and the sample-introduction channel in mode B. In this case, a sampling probe is directly connected to the chip microchannel, and sample introduction is performed simply by inserting the inlet of the sampling probe into an off-chip sample vial [101]. The main problem with this mode is that sample-change operations are manually performed by moving the inlet tip of the sample inlet tubing from one sample vessel to the next. Some improvements have been achieved by Q.-H. He et al. with the fabrication of a monolithic sampling probe system for automated and continuous sample introduction in CE microchips [102]. Mode D (Figure 2.9) is characterized by the use of autosamplers based on x-y-z stages to automate sample introduction [103]. Although these approaches are very effective, the cost and size of the systems are significantly increased. Other dispensing devices have recently been developed. C.K. Byun et al. have reported an electro-osmosis-based nanopipettor [104]. This consists of a microfabricated electro-osmotic flow pump, a polyacrylamide grounding interface and a nanoliter-to-picoliter pipet tip. One advantage of this nanopipettor is that it has no moving parts, and it works with good levels of accuracy and precision. M.L. Kovarik and S.C. Jacobson reported an attolitre-scale dispensing unit for nanofluidic channels [105]. This dispenses to the nanochannels volumes at the fl range. The transport was electrokinetically produced by applying up to 10 V directly from an analogue output board. 2.3.5

Detection Systems for Analytical Microsystems

Although detection in miniaturized systems is the subject of Chapter 6, this section exists to complete the list of the basic microdevices involved in analytical microfluidic systems. Section 2.6.3 also includes detection in MEMS. The most common detection systems used in microfluidics are based on optical and electrochemical sensing. B. Kuswandi et al. recently reviewed the optical sensing

Tools for the Design of Miniaturized Analytical Systems

65

systems for microfluidic devices [106]. They distinguished between the coupling approach (off-chip approach) and the integration approach (on-chip approach). The off-chip approach combines macroscale optical detection with micro-sized detection areas by using pinholes at focus points along the optical path or optical fibre technology. Although very low levels of background signal can be achieved, the reduction of the path length within the device can decrease the sensitivity of the method. Absorbance, fluorescence and chemiluminescence are the most common detection modes. On the other hand, the integration of optical components or functions in a microfluidic platform that should be able to perform all chemical functions and detections in a single device requires increased integration of not just fluidic elements, but also electrical or other types of elements (on-chip approach). MEMS have demonstrated the integration of mechanical and electrical functionalities into small structures for diverse applications (Section 2.6.3). The development of optofluidic technology, as the fusion of microfluidics and optics (the use of optical microscopy to study microorganisms in aqueous solutions, the common use of optical methods to analyse liquids and liquid solutions, etc.), has opened interesting possibilities [107]. The development of a large body of research to accommodate this knowledge at a miniaturized scale has produced a variety of elements allowing real optofluidic integration for microanalysis [108]. Electrochemical detection is another powerful tool for detection in microchips. This offers great advantages in realizing the lab-on-a-chip concept, such as inherent miniaturization, low power requirements, low limits of detection and compatibility with advanced micromachining systems. A wide variety of materials have been described for electrochemical sensing in microfluidic chip platforms [109]. Amperometric, conductimetric and potentiometric modes are the normal detection alternatives in microchips, although the first is clearly the most common. Amperometric detection on a microfluidic chip platform is typically accomplished using DC amperometry, which simply involves holding the working electrode at a particular potential sufficient to oxidize or to reduce the compounds of interest, and measuring the current that results. This method has proved to be sensitive, although a decoupler system must be used [110]. Conductivity detection is a simple and universal detection technique in miniaturized electrophoretic systems. This detection can be accomplished in electrophoretic systems either by direct galvanic contact of the run buffer and the sensing electrodes or by the contactless mode, in which the electrodes do not contact the solution. Potentiometry on a microfluidic chip platform is applied less as an electrochemical technique, most probably because microfluidic chips usually contain an inherent separation step (generally by electrophoresis or electrochromatography), so there is no need for discrimination of one analyte with respect to other components of the sample by ion-selective electrodes (ISEs).

66

Miniaturization of Analytical Systems

2.4 Microtechnology Following the description of the different microfluidic devices in the previous section, microtechnology must now be reviewed in order to explain the integration of the different elements, at microsystem scale, to produce different layout designs. The journey from electronics to microelectronics can be seen as following a similar evolution to that from fluids to microfluidics. Thus, the main milestones in the electronics–microelectronics evolution (invention of the transistor, improvements in semiconductor technology, introduction of the integrated circuit, etc.) have been used to inform a general strategy toward fabricating miniaturized systems in other disciplines (mechanics and optics, for instance). Microelectromechanical systems (MEMS), described in Section 2.5, are gaining more and more analytical interest through the development of microcantilevers. Microfluidics is another area of interest, following the introduction of the mTAS concept by Manz and coworkers in 1990, as was described in Chapter 1. It is necessary to briefly describe the manipulation of small amounts of samples and reagents on microchip, which allows the fabrication of microsystems for chemical analysis in silicon, glass and plastics, the packaging of such microsystems and the management of detection. 2.4.1

Computer Simulations in Microfluidics

As the fabrication of microsystems is very complex and requires infrastructure facilities (particularly cleanroom facility) and the participation of experts in different scientific and technological fields, computer simulations are extremely convenient. Simulations do not only provide a more complete understanding of the physical and chemical processes involved in the microsystem, but are very useful for the optimal development of the designs and for minimizing the risk of fabricating unsuccessful materials. G. Goranovic and H. Bruus have systematically described the different aspects (and stages) necessary to appropriately implement simulations in microfluidics [111]. A brief summary is given below. Physical Aspects and Design Different aspects must be taken into account: Dimensions: A microsystem is a network of fluidic channels and other additional components occupying an area of about one square centimetre, with channel widths typically on the order of 100 mm. Geometry: The basic component of a microfluidic network is the channel. Its main features include length, cross-section and surface properties. After a channel is made in a substrate, it is covered with a bonded lid, which can be made of different materials.

Tools for the Design of Miniaturized Analytical Systems

67

Surface: Chemical groups at the surface of channel walls, such as silanol (SiOH) groups in glass, can react with the ions in an electrolyte solution and create a very thin polarized layer, creating an EOF in an applied electric field (as explained in Section 2.3.2). Fluid Properties: Density, viscosity, electrical and thermal conductivity, diffusion coefficients and the surface tension of the liquids filling the channels are important to take into account. Heating: Joule heating is generated when an electrical current flows through a channel. In some applications, such heating may cause a negative effect in the samples analysed or the systems studied (e.g. biological samples). If this is the case, heating should be taken into account in the simulation. Software and Hardware Requirements Software packages are essential for simulations. Capability and price are the basic guidelines for selection. CFDRC (www.cfdrc.com) offers one package, a patent named Integrated Microfluidic System Design Using Mixed Methodology Simulations. This is a simulation-based method for rapidly designing, evaluating and/or optimizing complex microfluidic systems and biochips. It simulates processes including pressure-driven and electro-osmotic flows, electrophoretic separation, analyte dispersion, mixing of analytes, heat transfer, electric and magnetic fields, and biochemical reactions in order to complete the design of a particular microsystem. Coventor (www.coventor.com) offers another software package specifically oriented toward microtechnology, including MEMS and microfluidics. Coventor 3D MEMS software presents three different products. CoventorWare is used for MEMS R&D, design and product development. MEMulator enables MEMS designers and process engineers to visualize the effects of a design and process modifications in 3D before fabricating an actual device. EM3DS is used for electromagnetic modelling. All these packages describe the hardware requirements for their operation. Any commercial computational fluid dynamics package typically consists of a pre-processor, a solver and a post-processor. The pre- processor is the interface between the user and the numerical solver. Different factors must be considered before the solver can be started. Thus, geometry, grid generation, model, fluid parameters, boundary conditions and initial conditions have to be fixed and/or defined [111]. Numerical parameters must be adjusted so that solvers can solve the problem at hand most efficiently. Solvers are numerical algorithms intended to solve the governing equations. Solver settings include type of solver, numerical schemes, number of iterations, convergence criteria, etc. Both surface boundary conditions and solver setting must be specified for the simulation. Solvers commonly work by following the finite-volume method, based on the concept of transport equations, or the finite-element method, which uses simple piecewise functions (for instance, linear or quadratic functions) to approximate the exact solution. The first method is

68

Miniaturization of Analytical Systems

used mostly for fluid dynamics, and the second is more suitable for mechanical stress analysis and simulations of MEMS. Finally, post-processors are used to visualize results, following different techniques. Reliability of the Microsystem It is very convenient, after describing the various stages of the simulation process, to summarize the numerical errors and uncertainties involved. In computational fluid dynamics, the following errors and uncertainties can be identified: Model Uncertainty: Real flow in comparison to exact solution of modelled equations. Discretization or Numerical Error: Exact solution with respect to the numerical solution of the equations. Iteration or Convergence Error: Fully converged compared to not fully converged solutions. Round-off Errors: Parameter values that are below machine accuracy. Application Uncertainties: Lack of available data. User Errors: Mistakes and carelessness. Code Errors: Bugs in the software. Interpretation and Evaluation of Simulations Correct interpretation of numerical results is one of the most important issues, because of the complexity of simulations. In particular, the visualizing capabilities of computational fluid dynamics programs are of special importance, as well as their complete understanding by the user. The definitive test for the simulation results is the experiments: thus, one can see to what extent the simulation results represent reality. The problem is that in microfluidic systems it is rather difficult to make detailed measurements, because of the small size of these systems. One experimental possibility is the use of fluorescent tracers and a CCD camera to monitor the position of the tracer in two consecutive instants by illuminating it with a light source. A nonexperimental alternative is to compare the simulation with well-established theoretical models, such as the equivalent circuit theory. This theory describes a fluidic network as an equivalent electrical network by expressing a linear relationship between a pressure drop (equivalent to a voltage drop) and the corresponding flow rate (equivalent to a current) [112]. 2.4.2

Micromachining

After the design, to particular specifications, and the revision of the computer simulation, fabrication of the microsystem is carried out. Two main aspects have to be considered: the material for the fabrication and the specific facilities required (mainly the cleanroom processing with the appropriate tools). Without any doubt,

Tools for the Design of Miniaturized Analytical Systems

69

silicon is the most common substrate because it is a well-known material, it can be obtained in the highest purity (99.99999%) and it is of great interest to the semiconductor industry. Nevertheless, other materials such as glass and polymers are also used to fabricate analytical microsystems. Silicon and Silicon-compatible Materials Silicon substrates are used in many of the devices fabricated so far for analytical microsystems, and probably will continue to be used to a large extent in the future. It is the most-studied material and its micromachining is at a very advanced state. The substrates used are almost exclusively single-crystalline silicon, in the format of disc-like wafers (normally with a thickness of between 200 and 650 mm and a diameter between 75 and 150 mm). Wafers are usually cut to give top surfaces that roughly correspond to the main crystal planes in silicon, given by the Miller indices (100), (110) and (111). After cutting, the wafers are subjected to a lapping step, to give a surface roughness of about 0.5 mm and to remove about 50 mm of material from both sides of the wafer. Then one of the sides is further polished, using finer mechanical procedures, to produce a surface roughness commonly between 10 and 100 nm. Wafers are characterized by several parameters, conforming to different types of specifications of interest for the particular application to be developed. After the substrate preparation, a set of steps has to be followed to produce the silicon wafer at the packaging level [113]: . . .

optical lithography; deposition; etching/removal.

Lithography is the technique used to transfer a computer-generated pattern on to a substrate, such as silicon, but also glass or other materials. Although photolithography is the most common lithography technique in microelectronic fabrication, electrobeam (e-beam) and X-ray lithography are in progressive use in the MEMS and nanofabrication areas. Figure 2.10 summarizes the main steps in a photolithography process. Most of the fabricated microstructures contain materials other than that of the substrate, which are obtained by various deposition techniques or by modifying the substrate. Most of the thin films deposited have different properties to those of their corresponding bulk substrate. Different thin-film deposition and doping strategies have been reported, such as surface oxidation (film of SiO2 from Si), doping with certain impurities (for instance, boron to form p-type regions; phosphorus or arsenic to form n-type regions), chemical vapour deposition, physical vapour deposition (evaporation and sputtering), electroplating and pulsed laser deposition [114]. Thin-film and bulk-substrate etching is another fabrication step that is of fundamental importance to both the very large-scale integration (VLSI) processes and micro/nanofabrication. Thus, various conducting and dielectric films deposited

70

Miniaturization of Analytical Systems

(a)

(b)

1 – Oxidize the substrate SiO2

1

Silicon substrate

2

Deposit thin film

3

Spin photoresist

Substrate 2 – Spin the photoresist and soft bake Photoresist Substrate 3 – Expose the photoresist Light

4

Soft bake

5

Align the mask

Photomask

Substrate

6

Expose the waver

7

Develop the resist

4 – Develop the photoresist and hard bake

Substrate 5 – Etch the oxide

8 9

Hard bake

Substrate 6 – Strip the photoresist

End of lithography Substrate

Figure 2.10 (a) Main steps in a photolithography process, and (b) scheme of a photolithography with a positive PR

for passivation or masking purposes need to be removed at some point. The different etching techniques can be grouped into wet and dry categories. J.L. Perry and S.G. Kandlikar have reviewed the fabrication methods of nanochannels for single-phase liquid flow [115]. They distinguish four alternatives: (i) blulk nanomachining and wafer bonding; (ii) surface nanomachining; (iii) buried channel technology; and (iv) nanoimprint lithography. In bulk nanomachining and wafer bonding, features are created of the bulk of a silicon wafer. This can be done by reactive ion etching (RIE) or by a wet anisotropic etchant with aqueous KOH- or ethylenediamine-based solutions. Figure 2.11(a) describes this process, where wet anisotropic etching of one-dimension nanochannel is performed [116]. A (110) silicon wafer is used, which has a thin native oxide. The wafer is then lithography patterned and the oxide mask is etched with an HF solution. The silicon is anisotropically etched at an elevated temperature by a developer solution, which is essentially a water-dilute solution of TMAH (tetramethyl ammonium hydroxide). Next, the oxide mask is stripped and bonded to a borofloat glass wafer. Nanochannels can also be fabricated by surface nanomachining. This involves embedding the structures in a layer of appropriate sacrificial material on the surface of the substrate. The sacrificial material is dissolved, leaving a complete nanochannel.

Tools for the Design of Miniaturized Analytical Systems

71

Figure 2.11 (a) Fabrication process for bulk nanomachining with wafer bonding (left side, reproduced from [115] with permission from Springer, Copyright Springer) and cross-section of a silicon wafer with a 50 nm deep channel bonded to a borofloat wafer (right side, reproduced from [116] by permission of IOP) following this fabrication technique. (b) Schematic cross-section of an a-Si nanochannel array fabricated by Stern et. al [117] (left side) and the corresponding picture showing the surface nanomachined channels of 0.5 mm wide  100 nm high and 1 mm wide  100 nm high (right side), Copyright Springer.

Figure 2.11(b) is a schematic cross-section of an amorphous Si (a-Si) nanochannel array made by Stern et al. [117]. First, a thick layer of thermal oxide is grown for the electric isolation of electronic devices. Then a set of 0.05 mm microlayers (tetraethyl ortho-silicate, TEOS; Si3N4) are put down to form the lower-channel dielectric layer. These alternating layers help maintain the channel dimensions because each imposes an opposite thin-film stress (the TEOS layer is compressive, while the nitride layer is tensile). Afterwards, a thin a-Si film of nanometer thickness is grown in a furnace. The a-Si is patterned lithographically to define the nanochannels. The Si3N4 film below the a-Si is used as an etch stop during chemical wet etching. The top-channel dielectric layers are then deposited over the patterned a-Si layer and capped with a thick phosphosilicate glass (PSG) to protect the structure during channel etching. Additionally, reservoir regions are created at the ends of the nanochannels. Buried-channel technology consists of a set of steps, as Figure 2.12 shows [118]. In step 1, a bare substrate is covered with a suitable masking material and

72

Miniaturization of Analytical Systems

Figure 2.12 Fabrication sequence for conduit using buried channel technology (left side) and a picture of microchannels formed with this technology [118] (right side). Channels are closed with silicon nitride, which is buried underneath the silicon surface. Figures selected from reference [115]. SPRINGER-VERLAG

lithographically patterned. An isotropic etchant is then used to make a rounded-out feature (step 2). Then a trench is etched in the substrate by deep RIE (step 3). In step 4, the trench is conformally coated with a material to prevent lateral etching of the sidewalls in step 6. In step 5 (Figure 2.12), the coating is removed only at the bottom of the trench, and the structure is etched in the bulk of the substrate again with an isotropic etchant (step 6). After stripping the coating in step 7, the structure is closed by filling the trench with a suitable material, as step 8 shows. For nanoimprint lithography, the fabrication of a previous mould (by lithography, mainly) is necessary. Then two basic steps are followed (Figure 2.13(a)). In the first (the imprint step), a mould with nanostructures on its surface is pressed into a thin

Figure 2.13 (a) Schematic of nanoimprint lithography process and (b) comparison between the conventional and the template approaches. (c) Pictures of nanoimprinted fluidic channels with template used to enclose the nanochannels [120]. Figures selected from reference [115]. SPRINGER-VERLAG

Tools for the Design of Miniaturized Analytical Systems

73

polymer on a substrate. This step duplicates the nanostructures on the mould in the polymer film. The second step is the pattern transfer, where an anisotropic etching process is used to remove the residual polymer in the compressed area. This step transfers the thickness contrast pattern into the entire polymer. As was pointed out by J.L. Perry and S.G. Kandlikar [115], who presented an interesting comparison of these nanochannel fabrication methods, one last important aspect necessary for the fabrication of working nanofluidic systems is the enclosing of the channels. This is not an easy task, and several different sealing techniques to close up nanochannels have been developed. Thus, for instance, Cao et al. sputter silicon dioxide over the nanochannels [119], whereas Guo et al. have developed a more practical solution [120]. Their technique simply imprints a channel template into a thin polymer film while on a glass substrate. This makes it easy to control the nanochannel dimensions by a simple relationship involving the initial polymer layer thickness and the mould pattern configuration. Figure 2.13(b) compares the typical nanoimprint lithography process with the template approach. This fabrication process can be controlled to give predictable channel heights, as shown in Figure 2.13(c). Other Substrates Used in Analytical Microsystems Glass and polymers can form alternative substrates to silicon for microsystems fabrication. Glass is a monolithic noncrystalline solid consisting mainly of silicon dioxide (SiO2). Glass often contains additives or impurities that modify properties such as mechanical stability and the glass transition temperature. Glass can be used instead of silicon in analytical microsystems in chip formats because of its unique properties: it is resistant to many chemicals, optically transparent (optical detection and visual inspection are allowed) and a dielectric material (useful for creating electrokinetic-driven flows and separations). In addition, glass presents other advantages such as its hardness, high thermal stability and relative biocompatibility (applications in bioanalysis). Borofloat glass and quartz wafers are most commonly used as they are compatible with many cleanroom processes. Some traditional silicon microfabrication techniques, such as photolithography and wet chemical etching, have been adapted to glass processing. On the downside, micromachining of glass is less versatile than that of silicon, due to its noncrystalline structure and our limited experience with glass as a material for microsystems. In some cases, polymers are better suited to fabricating microsystems than is silicon. Many types of polymer show better resistance to chemical treatment and better biocompatibility. Moreover, they can be produced at significantly lower cost than silicon microsystems. There are a wide variety of plastics, with very different characteristics, including transparency in the ultraviolet (PDMS), visible (polymethylmethacrylate, PMMA; polycarbonate, PC) and infrared (polyetherimide) regions of the electromagnetic spectrum. The two major methods for machining polymers are replication from a master and direct machining. Replication methods

74

Miniaturization of Analytical Systems

often produce a microstructure by allowing a polymer workpiece to form an inverse copy of a mould. Direct machining methods remove small amounts of polymer in places where microstructures, such as microchannels or microwells, should be located. 2.4.3

Packaging of Microsystems

The functions of packaging are to protect the devices form the environment and the environment from the devices. The appropriate package is essential for an efficient operation of the microsystem. Aspects such as microfluidic interconnects (connections between microdevices, but also the interface between a macrocomponent and a microdevice), electrical connections, optical interconnections, and mounting and encasing devices that provide an appropriate interface to other devices or to the external environment, must be taken into account. Packaging also includes the protection of the microsystem from humidity and the possible introduction of foreign species, as well as the sealing techniques to assure hermeticity and structural integrity. A systematic treatment of microsystem packaging has been reported by G. Perozziello [121]. D. Erickson and D. Li reviewed the integrated microfluidic devices for a broad range of application areas [122]. More recently, G.T. Roman and R.T. Kennedy have published a review of fully integrated microfluidic separations systems for biochemical analysis [123]. In this review, the authors systematically discuss the integration of detection transducers into microfluidic devices, of pumps and power supplies for fluid manipulation, and of complete systems for biological analysis. In same cases, control of the assembly of microsystems is carried out by image processing [124]. A new fluidic interconnection approach for packaging of microsystems has recently been proposed by O. Geschke and coworkers [125]. They proposed a reversible, integrated fluidic interconnection composed of custommade cylindrical rings integrated in a polymer house next to the fluidic network. R. Lo and E. Meng have also described integrated and reusable in-plane microfluidic interconnects [126].

2.5 MEMS and NEMS The fabrication of micro- and nanoelectromechanical systems (MEMS and NEMS, respectively), has clearly been growing during the last decades. A wide variety of MEMS devices have been produced, and some are commercially used for different industrial, consumer and biomedical applications. From an analytical point of view, several types of sensor have been developed. Perhaps biosensors (BioMEMS and BioNEMS) and gas sensors have been given the most attention. MEMS and NEMS are an integration of mechanical elements, sensors and actuators, and electronics on a common silicon substrate through microfabrication technology [127]. The

Tools for the Design of Miniaturized Analytical Systems

75

invention of scanning tunnelling microscopy (STM) in 1981, and later atomic force microscopy (AFM) in 1986 by Binning et al., were the first milestones in the development of nanoscience and nanotechnology, and examples of the use of MEMS and NEMS. In fact, STM and AFM are not simply microscopy tools, but are also extremely useful techniques for chemists and biochemists. In AFM, a microcantilever, which is the most simple MEMS-based device, measures the small force acting between a sample surface and a sharp tip at the apex of a thin beam clamped at its side. The force on the tip bends the cantilever, which acts as a force transducer. Cantilever-based devices have been shown to be highly versatile sensors using mechanical, optical, electrostatic and electromagnetic methods to actuate or sense cantilever motion in order to detect gases, chemicals or biological entities. They also have wide applications in the field of medicine, specifically for the screening of diseases, detection of point mutations, blood glucose monitoring and detection of chemical and biological warfare agents. Different issues must be taken into account to develop microcantilevers as miniaturized analytical sensors. These are briefly discussed below. 2.5.1

Fabrication and Characterization

Silicon, silicon nitride, silicon oxide and polymers are the main materials used in the fabrication of microcantilevers. Cantilevers are batch-fabricated using wellestablished thin-film processing technologies that provide low cost, high yield and good reproducibility. The fabrication techniques generally include thin-layer deposition, photolithographic patterning, etching, and surface and bulk micromachining. Usually, a sacrificial layer is deposited on a prepatterned substrate before the structural material of the cantilever is deposited. This structural layer must be free of a stress gradient, otherwise the cantilevers will have problems in initial bending. It is also possible to fabricate arrays of thousands of identical cantilevers on one wafer [128]. In most cases only Si is used as a starting/base material and for the large-size passive structures/components such as the pedestal of the cantilever, but the material of the components can be replaced by other materials with even lower cost that are easily machined and differently functioned. Thus Wakayama et al. [129] proposed a new type of cantilever consisting of Si beams (with a small slab) and a ceramic pedestal for AFM and gas-sensor applications. This new cantilever has a lower cost and better quality, by which they mean improved homogeneity and reproducibility, especially for the tip/stylus part [129]. Young’s modulus is a key factor that affects the spring constant of a cantilever, which is directly related to the properties of the cantilever material. Silicon or related materials were traditionally used to make cantilevers, but they have a high Young’s modulus. If cantilevers were made of a softer material, they would be more sensitive for static deflection measurements; hence, polymers with a much lower Young’s modulus than silicon have been used as a substitute material for the fabrication of

76

Miniaturization of Analytical Systems

cantilevers [130,131]. Among polymers, SU-8 has been shown to be very sensitive, exhibiting a Young’s modulus about 40 times lowers than that of silicon. Theoretically, SU-8 cantilevers can reach the same sensitivity as previously fabricated Si cantilevers, and metallic resistors are known to have a very low electrical noise. 2.5.2

Functionalization

The development and application of chemical and biochemical cantilevers has been one of the fastest growing fields over the last decade or so [132,133]. A key step in developing a sensor is the immobilization of sensing agents. A good immobilization method should meet the following requirements: (i) be simple and fast; (ii) be nonspecific, i.e. the method should be usable for the immobilization of various sensing agents; (iii) produce immobilized reagents that are stable and do not leach from the substrate; and (iv) produce immobilized reagents that retain their chemical and biochemical activities. There are three widely used methods for immobilization [134]: (i) adsorption of sensing agents on to a solid substrate; (ii) covalent binding, which involves the formation of permanent chemical bonds between sensing agents and a support; and (iii) encapsulation or entrapment of sensing agents within a polymeric matrix. The receptor layer deposited on the cantilever surface directly affects the selectivity, reproducibility and resolution. We want to deposit a thin (to avoid changes in mechanical properties of the cantilever), uniform (to generate a uniform stress) and compact (to avoid interactions with the solid substrate beneath) layer of receptor molecules. This layer should be stable and robust: hence the receptors should be covalently anchored to the surfaces; on the other hand, receptor molecules should have enough degrees of freedom to freely interact with their specific ligand in the environment. The receptor activity should last over time and possibly stand regeneration of the sensing layer if the sensor has to be reused several times. Most of these requirements are common to other sensors. In fact, the proposed coating techniques and procedures are shared with other transducing principles. Self-assembled Monolayers (SAMs) The use of the self-assembling properties of alkane chain molecules with thiol (-SH) groups on gold substrates [135] or silane (-SiOX) groups on silicon substrates [136] is an easy and popular method for creating monolayers on cantilever surfaces. They spontaneously form uniform, densely packed, robust (covalent binding) monolayers. They can be synthesized with different chain lengths and end groups with specific chemical properties. They are therefore ideally suited to act as cross-linkers to anchor the receptor molecule to the substrate. SAMs on gold substrates The most frequently used technique to prepare organosulfur SAMs is self-assembly from organic or aqueous solvents. The process

Tools for the Design of Miniaturized Analytical Systems

77

simply involves immersion of a clean, gold-coated substrate into a solution of thiol, sulfide or disulfide. It is focussed on the use of the x-substituted alkanethiols, since they have been used most frequently. The effects of single-component x-substituted alkanethiolate SAMs on gold depend primarily on the temperature, the solvent used, the purity of the thiol, the immersion time and the quality of the gold substrate. The purity of the compound may at first be considered important when dealing with thiols since it is well-known that thiols readily dimerize to form a disulfide when exposed to oxygen. However, a small fraction of the corresponding disulfide in the solution is not critical for the formation of single-component monolayers, because disulfides are known to undergo S–S bond reductive cleavage and adsorb as thiolates on gold. As an example, a biotinylated silicon nitride cantilever has been reported [135]. One side was coated with 3 nm of chromium and 40 nm gold layer. The thin layer was used to increase the adhesion of gold to silicon nitride. The cantilever was incubated in a mixture of simple thiol alcohol and complex biotinylated analogue. A stable, dense and highly structured thiol monolayer on the freshly coated gold was then formed, while on the opposite side, physisorbed thiols were removed by rinsing the cantilever in ethanol. In this case, incubation of the whole cantilever structure in a thiol solution was suitable. If the deposition has to be on one side only, spraying or other multistep procedures can be employed. Ji et al. developed microcantilevers modified with a SAM to respond sensitively to specific ion concentrations [137]. The authors report the detection of trace amounts of CrO42 using microcantilevers modified with a SAM of triethyl12-mercaptododecylammonium bromide. The SAM was prepared on a silicon microcantilever coated with a thin layer of gold on one side. The predicted surface structure of the modified microcantilever used in this experiment is shown in Figure 2.14(a). When CrO42 was passed through the flow cell housing the modified cantilever, it was adsorbed on to the surface, displaced the halide ion and combined with the cationic ammonium group, forming a stable ionic association that altered the surface stress, causing a cantilever deflection. In other work [138], the same group demonstrated that the 1,6-hexanedithiol SAM on a gold coated cantilever acts as an unusually specific recognition agent for CH3Hgþ. Zuo et al. reported a novel SiO2 microcantilever-based sensor operated in static mode and functionalized with a self-assembled bilayer of Cu2þ/11-mercaptoundecanoic acid (11-MUA), which can specifically adsorb organophosphorus compounds [139]. This composite layer can specifically recognize P¼O-containing compounds with the formation of P¼O-Cu2þ coordination structure on the surface. SAMS on silane substrates Silanes are a group of molecules that consist of a silicon atom covalently attached to four variable groups. One or more of these groups can be readily substituted by the oxygen of a hydroxyl group at a silicon oxide surface. Thus, these molecules are capable of forming covalent bonds with a variety of glasses and silicon substrates. As stated, some of the most thoroughly

78

Miniaturization of Analytical Systems

Figure 2.14 Schematic representation of the functionalization of a microcantilever with (a) SAM on the gold surface; (b) SAM on the glass surface of the microcantilever; and (c) nanotubes on the microcantilever surface

studied and well-characterized SAMs in this category are prepared from alkyltrichlorosilanes (or chlorosilanes), e.g. octadecyltrichlorosilane (OTS). Other silanes commonly used include the aminosilanes, e.g. aminoethylaminopropyltrimethoxysilane (which is also known as trimethoxylsilylpropylethylene diamine). Maraldo et al. describe a silanization procedure with 3-aminopropyl-triethoxysilane (APTES) prior to the immobilization of activated antibody [140]. The procedural steps used to functionalize the cantilever sensor surface and activate the target antibody are schematically represented in Figure 2.14(b). The glass surface is sequentially cleaned with methanol–hydrochloric acid solution (1:1, v/v), concentrated sulfuric acid, hot sodium hydroxide and finally boiling water (step 1). After cleaning, the glass surface is silanylated with APTES in deionized water at pH 3.0 (adjusted by hydrochloric acid, 0.1 M) and 75  C for 2 hours (step 2). Covalent coupling of the stable intermediate with the amine functionalized glass surface is carried out by flowing the activated antibody past the glass surface (step 3). Sol-gel Functionalization Sol-gel glass is an optically transparent glass-like material produced by the hydrolysis and polycondensation of organometallic compounds at a low temperature.

Tools for the Design of Miniaturized Analytical Systems

79

The organometallic compounds commonly used are silicon alkoxides, e.g. Si(OCH3)4, tetramethylorthosilicate. Silica glass prepared by the sol-gel method is a porous matrix that contains interconnected ‘bottleneck-like’ pores (or so-called cavities and cages) formed by a 3D SiO2 network. The pores are usually in the size range of 1.5 to 10 nm, depending on the composition of the precursor and the conditions of preparation. It is found that the large sensing agents can be securely trapped, but small analytes can diffuse readily into and out of the pores of the sol-gel matrix. The process that is used to produce sol-gel glass is called the sol-gel process. It can be divided into the following steps: mixing (to form a solution), gelation, aging and drying. In a typical procedure to prepare a silica glass by the sol-gel technique, one starts with an appropriate alkoxide precursor, e.g. Si(OCH3)4, which is mixed with water, an acidic catalyst such as HCl and a mutual solvent such as methanol to form a solution (the sol) by stirring or sonication. Hydrolysis results in the formation of silanol groups (Si-OH). Linkage of silanol with siloxane occurs as a polycondensation reaction and eventually leads to the formation of an SiO2 network (with silanol groups on the surface). The resulting amorphous gel contains water and methanol. Aging of a gel involves maintaining the gel immersed in liquid for a period of time, from hours to days. During aging, polycondensation continues and the strength of the gel thereby increases. Finally, in the drying process, the solvents (water and methanol) are removed from the interconnected pore network. A sensing agent is added to the mixture at some time during the formation of the sol or gel. A sol-gel is a viscous solution of colloidal silica particles in a liquid. The sol-gel solution containing a sensing agent can be coated on to cantilever sensors as a sensing element. In the sol-gel glass, the reagents are trapped inside the pores of the matrix, where they may move freely or interact with silanol groups on the inner surfaces of the pores. The analytes to be determined, usually smaller, can diffuse into the pores and react with the sensing agents. On the other hand, the sol-gel glass has a large porosity (about 30%) and a very large specific surface area (>300 cm2/g) [134]. Therefore, a substantial fraction of the entrapped reagents may be exposed to a neighbouring phase and intrapore volume and can react with the analytes therein. The use of sol-gel materials for the development of chemical sensors and biosensors is of intense current interest [141]. In most applications, the sol-gel material functions as a porous support matrix in which analyte-sensitive species are entrapped. The sol-gel process is attractive because gelation can occur at low temperatures, which allows for the encapsulation of a variety of temperaturesensitive molecules, including biological reagents. A glucose sensor based on the entrapment of the protein glucose oxidase within a sol-gel matrix has been developed [142], as has an optical sensor based on the entrapment of high pKa dyes within a sol-gel matrix, which is used to determine pH in highly acidic and highly alkaline environments [143,144]. Raiteri et al. have used thin films of sol-gel

80

Miniaturization of Analytical Systems

coated on silicon nitride microcantilevers via a dip-coating process to show the change in response of the microcantilever to different mixtures of ethanol and water [145]. Fagan et al. modified and evaluated silicon microcantilevers with thin films of sol-gel [141]. Thin films of sol-gel were deposited on to microcantilevers using a spin-coating procedure. The method of film formation requires less equipment and is less expensive than conventional techniques such as CVD, evaporation or sputtering [146]. The authors have also studied the ability to modify the selectivity of the microcantilevers by modifying the chemical nature of the surface of the sol-gel film through reaction with organosilanes. Biofunctionalization Many bioanalytical methods, in particular biosensor microcantilevers, require a high density of functional molecules, low nonspecific protein adsorption, long-term stability and durability. Proteins – including antibodies and enzymes – and cells are immobilized on the sensor surface. This immobilization process should be uniform, avoiding any change in the mechanical properties of the cantilevers, and allow accessibility by the target molecule. Adsorption, entrapment and covalent attachment are the leading techniques employed for immobilization of biomolecules on to sensor surfaces. Carrascosa et al. give a good discussion of DNA and protein immobilization on cantilevers and array systems [128]. To ensure long-term stability, a chemical surface functionalization has to be applied to deposit a receptor layer on the cantilever surface. The ideal case for surface functionalization is a closed monolayer, in which the receptor is covalently anchored to the surface with high density but with enough space between the receptors to interact with the specific ligand. The receptor activity should be maintained over time and resist several regeneration cycles. If one wants to use the deflection detection method for cantilevers, only one side has to be functionalized. By using thiol monolayers, this can be done by evaporating gold on one side of the cantilever only or by incubating only one side of a double-sided, gold-coated cantilever in the thiol solution. Incubating only one side causes stress, which has to be taken into account [147]. Raiteri et al. reported a procedure for coating each side of the cantilever with a different thiol [148]. Nevertheless, surface functionalization is as easy as it sounds; however, if the deflection detection is used, the experimenter has to ensure that the specific interaction takes place at one side of the cantilever only and, even more importantly, that unspecific interaction is blocked. Not all biomolecules which may be used as recognition sites are commercially available as thiols. Yan et al. reported a method to apply a nanoassembly layer-by-layer technique on one side of a cantilever only [149]. This method produces ultrathin organized polymer films to which the recognition sites can be attached either electrostatically or by chemical reaction.

Tools for the Design of Miniaturized Analytical Systems

81

Functionalization Using Carbon Nanotubes By integrating nanotube growth into batch-fabricated microsystems, NEMS and MEMS are able to create sensors and develop methods and devices for direct and reliable measurement of nanotube transducer properties. Hierold et al. recently published a good review concerning MEMS and NEMS sensors based on carbon nanotubes (CNTs) [150]. This article presented the fabrication techniques and methods used for the integration of CNTs in these sensors. Jungen et al. describe a series of processing steps required to achieve successful integration of single-wall nanotubes (SWNTs) directly into a MEMS chip [151]. The process is schematically shown in Figure 2.14(c). An e-beam lithography resist (PMMA) of 400 nm thickness is spin coated on top of the 2 · 2 mm chip (step 1). The written structures are openings, typically 2 mm in feature size, defining placeholders for the catalyst material and resulting in a locally defined spot for catalysts for spatial control of nanotube growth. After development, a droplet of a catalytic solution based on iron nitride dissolved in methanol is placed on the chip and evaporated using a hotplate (40  C) (step 2). In order to grow tubes that span a 2 mm trench, the catalyst coating should come with a high surface area presenting nonaggregated nanoparticle-sized sites for the assembly of carbon molecules into nanotubes. The liftoff process is completed by stripping the PMMA in acetone under ultrasonic agitation (step 2). The chip is placed into a 4 inch quartz-tube LPCVD reactor. The furnace is heated under argon to 900  C. Methane fed at 1000 SCCM and kept at 100 mbar is provided as a carbon feedstock for 10 minutes. The reactor is evacuated and cooled to at least 300  C before venting to atmosphere using nitrogen (step 3). A second e-beam lithography and liftoff step are used immediately after this to deposit a Cr/Au layer for electrical connectivity (step 4). Since the growth occurs on the surface of the Poly 2 layer, most of the CNTs are freestanding prior to the HF release of the MEMS. The second spin coating has been proven not to destroy the CNTs. After a standard HF release the MEMS chip is ready for actuation. 2.5.3

Detection Methods

In general there are three different methods for transducing the recognition event into micromechanical motion: first, the frequency change due to additional mass loading or changing in the force constant can be measured (i.e. the cantilever is used as a microbalance); second, the binding of the bimetallic cantilever can be used as a temperature sensor (e.g. to sense calorimetric effects upon adsorption); and third, cantilevers can work as stress sensors by measuring the bending due to changes in the surface stress at one side of the cantilever. Adsorption of molecules, when they are restricted to one of the cantilever surfaces, produces differential surface stress that bends the cantilever. At the same time, the resonant frequency of the cantilever also varies due to mass loading. The bending and the changes in resonant frequency can be monitored by several techniques, using optical reflection, capacitive and

82

Miniaturization of Analytical Systems

piezoresistive/piezoelectric techniques. Changes in resonant frequency can be detected by measuring the thermal noise of the cantilever. However, to achieve great sensitivity, especially when working in liquids, it is necessary to pre-energize the cantilevers by using alternating electric, magnetic or acoustic fields. Optical Detection The method most often used to measure cantilever bending is the optical reflection technique. In this technique, deflection of the cantilever is measured by reflecting a beam from a solid-state laser diode on to the free apex of the cantilever and then off the back of the cantilever. The reflected laser beam from the cantilever surface is collected by a position-sensitive detector (PSD) consisting of two closely spaced photodiodes whose output signal is collected by a differential amplifier. When the cantilever bends, the reflected laser beam moves on the photodetector surface, and the distance traveled is proportional to the cantilever deflection. The advantage of this detection system is that it is capable of detecting deflection in the subnanometer range and can be implemented easily. But it also has its disadvantages. First, the presence of a focused laser beam in a liquid cell environment can result in additional thermal management issues, giving rise to extraneous readings. Second, the alignment system is expensive and involves great precision, which can ultimately raise the cost of the whole diagnostic kit. In addition, it also reduces the kit’s portability. Further, the implementation of an optical method for readout of arrays is technologically challenging, as it requires an array of laser sources with the same number of elements as the cantilever array. This technique is employed in the optically-based commercial array platforms, but sequential switching, on and off, of each laser source is necessary to avoid overlap of the reflected beams on the photodetector. This problem can be elegantly solved using a scanning laser source, where the laser beam is scanned along the array in order to illuminate the free ends of each microcantilever sequentially [152]. Lechuga et al. introduced a new type of optical waveguide cantilever that does not need an array of laser sources as the cantilever itself acts as the waveguide in conducting the light [153]. At the exit of the optical cantilever, light can be collected by another waveguide or by a photodetector. This new device has been shown to have a good performance, and it offers an interesting approach to further integration in lab-on-a-chip microsystems. Another implemented optical deflection method is interferometry detection. This is based on interference of a reference laser beam by the one reflected by the cantilever. The cleaved end of an optical fibre is brought close to the cantilever surface. One part of the light is reflected at the interface between the fibre and the surrounding media, and the other part is reflected at the cantilever back into the fibre. These two beams interfere inside the fibre, and the interference signal can be measured with a photodiode. Interferometry detection is highly sensitive and provides a direct and absolute measurement of the displacement. The light has

Tools for the Design of Miniaturized Analytical Systems

83

to be brought close to the cantilever in order for enough to be reflected. For this purpose, Rugar et al. positioned the cleaved end of an optical fibre a few micrometers away from the free end of the cantilever [154]. By doing this, they could measure deflections in the 0.01 A range. Bonaccurso et al. have proposed an interferometric system based on two cantilevers for parallel surface-stress monitoring of two cantilevers only 200 mm apart [155]. However, a few technical problems arise. Positioning the fibres is a demanding task, and interferometry works for only small displacements: absolute deflection is defined only within a single wavelength, and measurements in liquids are, in general, less sensitive. An alternative approach is to use interdigitated cantilevers as an optical diffraction grating [156,157]. In this case, the reflected laser light forms a diffraction pattern whose intensity is proportional to cantilever deflection. An interferometric sensor for the AFM with a cantilever micromachined into the shape of interdigitated fingers to form a diffraction grating has been presented [156,157]. Cooper et al. demonstrated a promising type of microfabricated accelerometer based on the optical interferometer, and this interdigitated cantilever can also be used for infrared imaging and chemical sensing [158]. Piezoresistive/Piezoelectric Detection The piezoresistive method involves the embedding of some piezoresistive material near the top surface of the cantilever to record the stress change occurring there. As the microcantilever deflects, it undergoes a stress change that applies strain to the piezoresistor, thereby causing a change in resistance that can be measured by electronic means. Piezoresistive detection shows several advantages compared to the optical sort: no expensive macroscopic optical components and no timeconsuming laser alignments are needed and read-out electronics can, in principle, be integrated on to the same silicon chip supporting the cantilevers. Optical techniques can also be affected by artefacts due to changes in the optical properties of the medium surrounding the cantilever (e.g. a change in the refracting index when exchanging two different solutions), which can move the laser spot on the photodetector surface. Piezoresestive/piezoelectric detection does not suffer from this problem and can work in nontransparent solution. A disadvantage is that the deflection resolution for the piezoresistive readout system is only one nanometre, compared with one angstrom with the optical detection method. Another disadvantage with this method is that a piezoresistor has to be embedded in the cantilever. Fabricating such a composite cantilever is complicated. Capacitive Detection Although the most common detection mechanism reported in the literature is the optical beam deflection method, capacitive sensing can also be used. This does not require a complex adjustment and alignment system, as the optical beam does, and

84

Miniaturization of Analytical Systems

provides an integrated readout, which is very convenient for portable low-cost systems. The capacitive method is based on the principle that when the cantilever deflection takes place due to the adsorption of the analyte, this deflection is detected by the capacitance variation between the cantilever itself, acting as a moveable electrode, and the fixed electrode placed underneath. This deflection technique is highly sensitive and provides absolute displacement, but is not suitable for measuring large displacements. Moreover, it does not work in electrolyte solutions due to the faradic currents between the capacitive plates. Therefore, it is limited in its sensing applications. Interferometric Detection Interferometric detection of cantilever deflection is based on constructive and destructive interferences that occur when a collimated beam of light reflects off two surfaces displaced from one another [159]. The intensity of light is measured by a photodiode array. Although this is not a very common detection technique, it is capable of measuring very small deflections, but has a very limited dynamic range. Interferometric detection is used for high-temperature vibration sensors [160], as well as for studying the different resonance modes of cantilevers in arrays [161].

2.6 Outlook Fully integrated microfluidic devices have been widely developed during the last few years. A variety of platforms have been proposed for different applications. Without doubt, biochemical analysis has been the area that has received most attention. In fact, as P. Abgrall and N.T. Nguyen recently reported [162], the proximity of the dimension of nanochannels and the Debye length, the size of biomolecules such as DNA or proteins, and even the slip length, added to the excellent control of the geometry, give unique features to nanofluidic devices. These devices not only find applications wherever there is less well-defined porous media (electrophoresis gels, mainly) but give new insight into the sieving mechanisms of biomolecules and the flow of fluid at the nanoscale. Moreover, the control of the geometry allows smarter design, resulting in new separation principles (among other things) through taking advantage of the anisotropy. This fact connects with the trend of bioanalytical chemistry to analyse the profiles and dynamics of molecular components and subcellular structures in living cells. For these purposes, microfluidic devices have come to be of marked interest, because of their advantages in performing analytical functions such as controlled transportation, immobilization and manipulation of biological molecules and cells. Further, their ability to carry out separation and mixing processes, as well as the dilution of chemical reagents in a controlled way, allows the analysis of intracellular parameters and detection of cell

Tools for the Design of Miniaturized Analytical Systems

85

metabolites (even on a single-cell level). Therefore, microfluidics technology is an excellent tool for the manipulation and analysis of biological cells [163]. Cell manipulation on chips commonly uses different forces. In magnetic manipulation, magnetic particles are selectively attached to cells. This is a common method for cell separation or purification in microfluidic devices. The optical manipulation of biological species has received increasing interest due to its noncontact and contamination-free manipulation process. On the other hand, the separation of target cells on microfluidic structures for culture and assay purposes is an important application of on-chip mechanical manipulation. Different microstructures have been developed for this goal [163]. Electrical manipulation through dielectrophoresis is another alternative which has been used as a microfluidic cell trap. Cell treatment is crucial for the integration of cell treatment steps on-chip to develop microfluidic devices for cell analysis. In this context, cell lysis is absolutely necessary for the analysis of cell constituents. The ability to integrate both steps (lysis and analysis) greatly increases the power and portability of many microfluidic devices [164]. Moreover, microfluidic devices are especially suitable for cell culture, as the scale of these devices allows the creation of locally formed stable microenvironments for this purpose [165]. Compared to traditional culture tools, microfluidic platforms provide much greater control over cell microenvironment and rapid optimization of media composition while using relatively small numbers of cells. For other cell-based assays, in which the introduction of hydrophilic or membrane-impermeant molecules into cells is necessary, electroporation, electrofusion and optoporation processes have been successfully applied in microfluidic systems for cell treatment (e.g. [166–168]). Cell analysis on microchips introduces numerous benefits, such as reduced cell consumption, automated and reproducible reagent delivery, and improved performance. Consequently, the progressive development of microfluidic devices for specific bioanalytical applications of cell processing has become a reality. M. Yang and coworkers have reviewed numerous publications dealing with these applications [163]. Within these applications, cytometry has received much attention. Cytometry refers to the measurement of the physical and chemical characteristics of cells. This technique is commonly restricted to specialized laboratories, and is generally expensive. The development of readily adaptable chip technology allows the flow cytometer unit to be adaptable to a wide variety of potential uses, with portability and low cost. Differing from conventional flow cytometers, which usually employ fluorescence detection, microfluidic flow cytometers can be mounted on a variety of detection instruments. In particular, chemical cytometry, referring to the use of highly sensitive analytical tools to characterize chemical composition of single cells, is essential for understanding the molecular mechanisms of many fundamental biological processes. But it is also crucial for the study of serious health disorders, because these studies require simultaneous analysis of a large number of chemical species within single cells. These chemical species

86

Miniaturization of Analytical Systems

cover small molecules generally related to cell metabolism, large molecules such as proteins, and compounds of genetic interest such as nucleic acid. Using this technology, S.R. Quake and coworkers developed a microfluidic chip for automated nucleic acid purification from small numbers of bacterial or mammalian cells [169]. Microfluidic systems also show great potential for the analysis of intracellular parameters and the detection of the presence of cell metabolites, even on the singlecell level. Thus, a representative number of applications, including monitoring of ATP, NAD, FAD, IgE, pH and Ca2þ, among other compounds, have been reported by M. Yang et al. [163]. One interesting use of microfluidic systems is as reactors for DNA amplification [170], particularly via polymerase chain reaction (PCR). This enzyme-catalysed reaction allows any nucleic acid sequence to be generated in abundance in vitro, and has become a fundamental tool in molecular biology. Using this process with conventional thermal cyclers is slow and inefficient due to the large thermal masses associated with instrumentation. To avoid this problem, many microfabricated devices for PCR have been reported, showing a clear reduction in the thermal mass of the system. Thus, different heating mechanisms have been used for effective and rapid PCR in volumes ranging from 1 pL to 50 mL. As an example of the progress made in this crucial area, Figure 2.15 shows the integrated microfluidic bioprocessor for PCR assays developed by R.G. Blazej et al. [171]. This demonstrates that

Figure 2.15 Nanolitre-scale microfabricated bioprocessor integrating thermal cycling, sample purification and capillary electrophoresis for complete Sanger sequencing from 1 fmol of DNA template. [171]. Reprinted with permission of Nature Publishing Group

Tools for the Design of Miniaturized Analytical Systems

87

complex biological processing can be performed at higher speeds and better efficiencies than previously reported. Microfluidic systems also allow miniaturization and integration of complex functions, which could move sophisticated diagnostic tools out of the developedworld laboratory. Thus, microfluidic diagnostic technologies for public health can be seen as a reality, as T. Edwards et al. report [172]. These authors present an example of an integrated disposable diagnostic card, which is reproduced in Figure 2.16. Without doubt, microfluidics opens a wide field of applications and possibilities in a variety of different contexts, as previous examples and references demonstrate. But, as G.M. Whitesides states [173], ‘the field is still at an early stage of development,’ and despite these developments, other issues must be addressed. Real problem-solving and commercialization are two important challenges to date.

Figure 2.16 Integrated disposable diagnostic card: (a) image of the card, where the O-rings are for interfacing with off-card components (valves, pumps, etc.); (b) scheme of the card. This card accepts filtered saliva from the syringe and contains an H-filter for further sample conditioning, a herringbone mixer for mixing antibodies with the sample, and channels with gold-coated surfaces for detection of analyte in the sample using an SPR imaging-based immunoassay. (Reprinted from [172] with permission of Nature Publishing Group)

88

Miniaturization of Analytical Systems

It is important to identify exactly which applications will need these types of miniaturized formats, because of their advantages or, even more importantly, because alternatives cannot be used (much reduced size of samples, in vivo monitoring, etc.). As A. Escarpa et al. report [174], ‘real (or non-ideal) sample analysis performed on microfluidics conceptually involves the complete integration of sample preparation, analyte separation, and detection on these platforms’. Transition from the behaviour of analytical microsystems with synthetic standard solutions (academic research) to natural (‘real’) samples is not a simple change of the inlet vial of a microfluidic device; in many cases it is a complete change of behaviour. From this applied-practical side of analytical microsystems, it is clear that the main achievements have been reached in the bioanalytical field. Despite the excellent reviews in the fields of environmental [175,176], explosives [177] and food [178] analysis, many challenges remain for the routine implementation of microfluidics in these areas. At the end of the day, as usual in the analytical method, analytical validation, involving not only target analytes but also objective samples, will be the definitive demonstration of the true usefulness of miniaturized analytical systems.

References [1] O. Geschke, H. Klank, P. Tellesmann (Eds.), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. [2] B. Bhushan (Ed.), Handbook of Nanotechnology, Springer, Heidelberg, Germany, 2007. [3] D. Li (Ed.), Encyclopedia of Microfluidics and Nanofluidics, Springer, Heidelberg, Germany, 2008. [4] K. Dill, R. Liu, P. Grodzinsky, Microarrays: Preparation, Microfluidics, Detection Methods, and Biological Applications, Springer, Heidelberg, Germany, 2009. [5] J.P. Kutter, H. Klank, In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. [6] K.W. Oh, C.H. Ahn, J. Micromech. Microeng., 16 (2006) R13. [7] M. Zimmermann, P. Hunziker, E. Delamarche, Microfluid Nanofluid, 5 (2008) 395. [8] C.T.A. Kovacs, Micromachined Transducers Sourcebook, McGraw-Hill, New York, USA, 1998. [9] J. Chen, P. Huang, M. Lin, Microfluid. Nanofluid., 4 (2008) 427. [10] H. Cho, H.Y. Kim, J.Y. Kang, T.S. Kim, J. Colloid Interface Sci., 306 (2007) 379. [11] C.M. Lu, Y.B. Xie, Y. Yang, M.M.C. Cheng, C.G. Koh, Y.L. Bai, L.J. Lee, Anal. Chem., 79 (2007) 994. [12] D.J. Beebe, J.S. Moore, J.M. Bauer, Q. Yu, R.H. Liu, C. Devadoss, B. Jo, Nature, 404(6778) (2000) 588. [13] G. Londe, A. Chunder, A. Wesser, L. Zhai, H.J. Cho, Sens. Actuators B, 132 (2008) 431. [14] K.W. Oh, C. Park, K. Namkoong, J. Kim, K.-C. Ock, S. Kim, Y.-A. Kim, Y.-K. Cho, C. Ko, Lab. Chip., 5 (2005) 845. [15] Z. Yang, R. Maeda, J. Chromatogr. A, 1013 (2003) 29. [16] H. Takao, M. Ishida, Microelectromech. Syst., 12 (2003) 497. [17] A. Luque, J.M. Quero, C. Hibert, P. Fluckiger, A.M. Ganan-Calvo Sens. Actuators B, 118 (2005) 144. [18] D.B. Weibel, M. Kruithof, S. Potenta, S.K. Sia, G.M. Whitesides, Anal. Chem., 77 (2005) 4726.

Tools for the Design of Miniaturized Analytical Systems [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61]

89

E. Piccin, R. Laocharoensuk, J. Burdick, E. Carrilho, J. Wang, Anal. Chem., 79 (2007) 4720. P. Woias, Sens. Actuators B, 105 (2005) 28. B.D. Iverson, S.V. Garimella, Microfluid. Nanofluid., 5 (2008) 145. A. Nisar, N. Afzulpurkar, B. Mahaisavariya, A. Tuantranont, Sens. Actuators B, 130 (2008) 917. L. Shui, J.C.T. Eijkel, A. van den Berg Sens. Actuators B, 121 (2007) 263. L.-S. Jang, Y.-Ch. Yu, Microsyst. Technol., 14 (2008) 241. Y.-C. Hsu, S.-J. Lin, Ch.-Ch. Hou, Microsyst. Technol., 14 (2007) 31. J.W. Judy, T. Tamagawa, D.L. Polla, Proc. MEMS, 91 (1991) 182. R. Zengerle, M. Richter, H. Sandmaier, Proceedings of IEEE, Microelectromechanical System, 1992, 19. R. Zengerle, J. Ulrich, S. Kluge, M. Richter, A. Richter, Sens. Actuators A, 50 (1995) 81. C. Cabuz, W.R. Herb, E.I. Cabuz, S.T. Lu, Proc. IEEE MEMS, (2001) 519. H.T.G. Van Lintel, F.C.M. van de Pol, S. Bouwstra, Sens. Actuators, 15 (1988) 153. C.G.J. Schabmueller, M. Koch, M.E. Mokhtari, A.G.R. Evans, A. Brunnschweiler, H. Sehr, J. Micromech. Microeng., 12 (2002) 420. K. Junwu, Y. Zhigang, P. Taijiang, C. Guangming, W. Boda, Sens. Actuators A, 121 (2005) 156. G.H. Feng, E.S. Kim, J. Microelectromech. Syst., 14 (2005) 192. A. Doll, M. Heinrichs, F. Goldschmidtboeing, H.J. Schrag, U.T. Hopt, P. Woias, Sens. Actuators A, 130/131 (2006) 445. Y.C. Hsu, S.J. Lin, C.C. Hou, Microsystem Technologies, 14 (2008) 31. T. Suzuki, Y. Teramura, H. Hata, K. Inokuma, I. Kanno, H. Iwata, H. Gotera, Microsystem Technologies, 13 (2007) 1391. O.C. Jeong, S.S. Yang, Sens. Actuators A, 83 (2000) 249. J.H. Kim, K.H. Na, C.J. Kang, Y.S. Kima, Sens. Actuators A, 120 (2005) 365. O.C. Jeong, S.W. Park, S.S. Yang, J.J. Pak, Sens. Actuators A, 123/124 (2005) 453. Q. Gong, Z. Zhou, Y. Yang, X. Wang, Sens. Actuators A, 83 (2000) 200. C. Yamahata, C. Lotto, E. Al Assaf M.A.M. Gijs, Microfluid. Nanofluid., 1 (2005) 197. C. Yamahata, M. Chastellain, V.K. Parashar, A. Petri, H. Hofmann, M.A.M. Gijs, J. Microelectromech. Syst., 14 (2005) 96. T. Pan, S.J. McDonal, E.M. Kail, B. Ziaiel, J. Microelectromech. Syst., 15 (2005) 1021. Y. Yang, Z. Zhou, X. Ye, X. Jiang, Bimetallic Thermally Actuated Micropump Vol. 59, American Society of Mechanical Engineers, Dynamic Systems and Control Division (Publication) DSC, 1996, p. 351. D. Xu, L. Wang, G. Ding, Y. Zhou, A. Yu, B. Cai, Sens. Actuators A, 93 (2001) 87. G. Shuxiang, T. Fukuda, Proc. IEEE Int. Conf., 2 (2004) 1616. S. Tadokoro, S. Yamagami, M. Ozawa, Proc. IEEE Int. Conf. on Microelectromechanical Systems, (1999) 37. S. Guo, K. Asaka, Proc. IEEE Conf. on Robotics & Automation, (2003) 1830. M.L. Marina, M. Torre, Talanta, 41 (1994) 1411. M.L. Marina, A. R
ıos, M. Valc
arcel, In: D. Barcel
o (Ed.), Comprehensive Analytical Chemistry, Elsevier, Amsterdam, 2005. S. Zeng, C.H. Chen, J.C. Mikkelsen, J.G. Santiago, Sens. Actuators B, 79 (2001) 107. C.H. Chen, J.G. Santiago, J. MEMS, 11 (2002) 672. L. Chen, H. Wang, J. Ma, C. Wang, Y. Guan, Sens. Actuators B, 104 (2005) 117. Y. Takemori, S. Horiike, T. Nishimoto, H. Nakanishi, T. Yoshida, Proc. 13th Int. Conf. on Solid State Sensors, Actuators and Microsystems, (2005). P. Wang, Z. Chen, H.C. Chang, Sens. Actuators B, 113 (2006) 500. F.-Q. Nie, M. Macka, L. Barron, D. Connolly, N. Kent, B. Paull, Analyst, 132 (2007) 417. M. Mpholo, G.C. Smith, A.B.D. Brown, Sens. Actuators B, 92 (2003) 262. S. Debesset, C.J. Hayden, C. Dalton, J.C.T. Eijkel, A. Manz, Lab On A Chip, 4 (2004) 396. S. Vijendran, C.G. Smith, M.I. Mpholo, Proc. SPIE, 6112 (2006) 61120. P. Garc
ıa-S
anchez A. Ramos, N.G. Green, H. Morgan, IEEE Trans Dielectrics Electrical Insulation, 13 (2006) 670. B.P. Cahill, L.J. Heyderman, J. Gobrecht, A. Stemmer, Sens. Actuators B, 110 (2005) 157.

90 [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107]

Miniaturization of Analytical Systems F. Mugele, J.-C. Baret, J. Phys. Condensed Matter, 17 (2005) 705. V. Bahadur, S.V. Garimella, J. Micromech. Microeng., 16 (2006) 1494. V. Bahadur, S.V. Garimella, Langmuir, 23 (2007) 4918. X. Geng, H. Yuan, H.N. Oguz, A. Prosperetti, J. Micromech. Microeng., 11 (2001) 270. J.D. Zahn, A. Deshmukh, A.P. Pisano, D. Liepmann, Biomed. Microdev., 6 (2004) 183. Z. Yin, A. Prosperetti, Micromech. Microeng., 15 (2005) 643. M. Nakano, S. Katsura, G.G. Touchard, K. Takashima, A. Mizuno, IEEE Trans. Ind. Appl., 43 (2007) 232. N.-T. Nguyen, A.H. Meng, J. Black, R.M. White, Sens. Actuators A, 79 (2000) 115. C.-C. Chang, R.-J. Yang, Microfluid Nanofluid, 3 (2007) 501. A.E. Kamholz, B.H. Weigl, B.A. Finlayson, P. Yager, Anal. Chem., 71 (1999) 5340. R.F. Ismagilov, A.D. Stroock, P.J.A. Kenis, G.M. Whetesides, H.A. Stone, Appl. Phys. Lett., 76 (2000) 2376. R.-J. Yang, C.C. Chang, IMECE’04, (2004) 61441. C.H. Wu, R.-J. Yang, Biomed. Microdevices, 8 (2006) 119. D. Kohlheyer, G.A.J. Besselink, R.G.H. Lammertink, S. Schlautmann, S. Unnikrishnan, R.B.M. Schasfoort, Microfluid Nanofluid, 1 (2005) 242. F. Sch€onfeld, V. Hessel, C. Hofmann, Lab Chip, 4 (2004) 65. A.D. Stroock, S.K.W. Dertinger, A. Ajdari, I. Mezic, H.A. Stone, G.M. Whitesides, Science, 295 (2002) 647. D. Erickson, D. Li, Langmuir, 18 (2002) 1883. C.-C. Chan, R.-J. Yang, J. Micromech. Microeng., 14 (2004) 550. K. Fushinobu, M. Nakata, Trans ASME J. Electronic Packaging, 127 (2005) 141. G.H. Tang, Z. Li, J.K. Wang, Y.L. He, W.Q. Tao, J. Appl. Phys., 100 (2006) 094908. M.H. Oddy, J.G. Santiago, J.C. Mikkelsen, Anal. Chem., 73 (2001) 5822. J. Park, S.M. Shin, K.Y. Huh, I.S. Kang, Phys. Fluids, 17 (2005) 118101. C.-H. Tai, R.-J. Yang, M.-Z. Huang, C.-W. Liu, C.-H. Tsai, L.-M. Fu, Electrophoresis, 27 (2006) 4982. M.-Z. Huang, R.-J. Yang, C.-H. Tai, C.-H. Tsai, L.-M. Fu, Biomed. Microdevices, 8 (2006) 309. T.B. Jones, Electromechanics of Particles, Cambridge University Press, Cambridge, UK, 2005. J. Lee, H. Moon, J. Fowler, T. Schoellhammer, C.-J. Kim, Sens. Actuators A, 95 (2002) 259. H. Moon, S.K. Cho, R.L. Garrell, C.J. Kim, J. Appl. Phys., 92 (2002) 4080. A.O. El Moctar, N. Aubry, J. Batton, Lab Chip, 3 (2003) 273. X. Niu, J.-K. Lee, J. Micromech. Microeng., 13 (2003) 454. I. Glasgow, S. Lieber, N. Aubry, Anal. Chem., 76 (2004) 4825. I. Glasgow, J. Batton, N. Aubry, Lab Chip, 4 (2004) 558. A. Dodge, A. Hountondji, M.C. Jullien, P. Tabeling, Phys. Rev. E, 72 (2005), 056312. D. Lastochkin, R. Zhou, P. Wang, Y. Ben, H.-C. Chang, J. Appl. Phys., 96 (2004) 1730. M.Z. Bazant, Y. Ben, Lab Chip, 6 (2006) 1455. A.N. Hellman, K.R. Rau, H.H. Yoon, S. Bae, J.F. Palmer, K.S. Phillips, N.L. Allbritton, V. Venugopalan, Anal. Chem., 79 (2007) 4484. Q. Fang, X-T. Shi, W.-B. Du, Q.-H. He, H. Shen, Z.-L. Fang, Trends Anal. Chem., 27 (2008) 521. S. Attiya, A.B. Jemere, T. Tang, G. Fitzpatrick, K. Seiler, N. Chiem, D.J. Harrison, Electrophoresis, 22 (2001) 318. Y.-H. Lin, G.-B. Lee, C.-W. Li, G.-R. Huang, S.-H. Chen, J. Chromatogr. A, 937 (2001) 115. Q. Fang, G.-M. Xu, Z.-L. Fang, Anal. Chem., 74 (2002) 1223. Q.H. He, Q. Fang, W.-B. Du, Z.-L. Fang, Analyst, 130 (2005) 1052. Q.-H. He, Q. Fang, W.B. Du, Z.-L. Fang, Electrophoresis, 28 (2007) 2912. S. Ekstrom, P. Onnerfjord, J. Nilsson, M. Bengtsson, T. Laurell, G. Marko-Varga, Anal. Chem., 72 (2000) 286. C.K. Byun, X. Wang, Q. Pu, S. Liu, Anal. Chem., 79 (2007) 3862. M.L. Kovarik, S.C. Jacobson, Anal. Chem., 79 (2007) 1655. B. Kuswandi, N. Nuriman, J. Huskens, W. Verboom, Anal. Chim. Acta, 601 (2007) 141. D. Psaltis, S.R. Quake, C. Yang, Nature, 442 (2006) 381.

Tools for the Design of Miniaturized Analytical Systems [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151]

91

H.C. Hunt, J.S. Wilkinson, Microfluid Nanofluid, 4 (2008) 53. M. Pumera, A. Merko i, S. Alegret, Trends Anal. Chem., 25 (2006) 219. D.M. Osbourn, C.E. Lunte, Anal. Chem., 75 (2003) 2710. G. Goranovic, H. Bruus, In: O. Geschke, H. Klank, P. Tellesmann (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. W.E. Morf, O.T. Guenat, N.F. de Rooij, Sens. Actuators B, 72 (2001) 266. A.M. Jorgensen, K.B. Mogensen, In: O. Geschke, H. Klank, P. Tellesmann (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. B. Bhushan (Ed.), Handbook of Nanotechnology, 2 Ed, Springer. Heidelberg, Germany, 2007, Chapter 7. J.L. Perry, S.G. Kandlikar, Microfluid Nanofluid, 2 (2006) 185. J. Haneveld, H. Jansen, E. Berenschot, N. Tas, M. Elwenspoek, J. Micromech. Microeng., 13 (2003) S62. M.B. Stern, M.W. Geis, J.E. Curtin, J. Vac. Sci. Technol. B, 15 (1997) 2887. M.J. de Boer, J. Microelectromech. Syst., 9 (2000) 94. H. Cao, Appl. Phys. Lett., 81 (2002) 174. L.J. Guo, X. Cheng, C. Chou, Nano Lett., 4 (2004) 69. G. Perozziello, In: O. Geschke, H. Klank, P. Tellesmann (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, Germany, 2004. D. Erickson, D. Li, Anal. Chim. Acta, 507 (2004) 11. G.T. Roman, R.T. Kennedy, J. Chromatog. A, 1168 (2007) 170. R. Schmitt, S. Driessen, B. Engelmann, Microsyst. Technol., 12 (2006) 640. G. Perozziello, F. Bundgaard, O. Geschke, Sens. Actuators B, 130 (2008) 947. R. Lo, E. Meng, Sens. Actuators B, 132 (2008) 531. G. Binnig, C.F. Quate, C. Gerber. Phys. Rev. Lett., 56 (1986) 930. L.G. Carrascosa, M. Moreno, M. Alvarez, L.M. Lechuga, Trends Anal. Chem., 25 (2006) 196. T. Wakayama, T. Kobayashi, N. Iwata, N. Tanifuji, Y. Matsuda, S. Yamada, Sens. Actuators A, 126 (2006) 159. A. Johansson, M. Calleja, P.A. Rasmussen, Sens. Actuators A, 111 (2005) 123. J. Thaysen, A.D. Yalcinkaya, P. Vettiger, A. Menon, J. Phys. D: Appl. Phys., 35 (2002) 2698. W.R. Seitz, CRC Crit. Rev. Anal. Chem., 19 (1988) 135. M.E. Collison, M.E. Meyerhoff, Anal. Chem., 62 (1990) 425A. J. Lin, C.W. Brown, Trends Anal. Chem., 16 (1997) 200. R. Raiteri, M. Grattarola, H.J. Butt, P. Skl
adal, Sens. Actuators B, 79 (2001) 115. T. Cass, F.S. Ligler, Immobilized Biomolecules in Analysis, Oxford University Press, Oxford, UK, 1998. H.F. Ji, T. Thundat, R. Dabestani, G.M. Brown, P.F. Britt, P.V. Bonnesen, Anal. Chem., 73 (2001) 1572. H.F. Ji, Y. Zhang, V.V. Purushotham, S. Kondu, B. Ramachandran, T. Thundat, D.T. Haynie, Analyst, 130 (2005) 1577. G. Zuo, X. Li, P. Li, T. Yang, Y. Wang, Z. Cheng, S. Feng, Anal. Chim. Acta, 580 (2006) 123. D. Maraldo, R. Mutharasan, Sens. Actuators B, 123 (2007) 474. B.C. Fagan, C.A. Tipple, Z. Xue, M.J. Sepaniak, P.G. Datskos, Talanta, 53 (2000) 599. P.C. Padney, S. Upadhyay, H.C. Pathak, Sens. Actuators B, 60 (1999) 83. L.R. Allain, K. Sorasaenee, Z.B. Xue, Anal. Chem., 69 (1997) 3076. L.R. Allain, Anal. Chem., 72 (2000) 1078. R. Raiteri, S. Martinoia, F. Molinari, G. Carlini, D. Ricci, M. Grattarola, Sensors Microsystems, 1998 (1999) 51. C.J. Brinker, G.W. Scherer (Eds), Sol-gel Science, Academic Press, San Diego, USA, 1990. R. Berger, E. Delamarche, H.P. Lang, C. Gerber, J.K. Gimzewski, Science, 276 (1997) 2021. R. Raiteri, M. Grattarola, H.J. Butt, P. Skladal, Sens. Actuators B, 79 (2001) 115. X. Yan, Y. Lvov, H.F. Ji, A. Singh, T. Thundat, Org. Biomol. Chem., 1 (2003) 460. C. Hierold, A. Jungen, C. Stampfer, T. Helbling, Sens. Actuators A, 136 (2007) 51. A. Jungen, C. Stampfer, J. Hoetzel, V.M. Bright, C. Hierold, Sens. Actuators A, 130 (2006) 588.

92

Miniaturization of Analytical Systems

[152] J. Tamayo, M. Alvarez, L.M. Lechuga, European Patent PCT/EP2005/002356 (2004). [153] K. Zinoviev, C. Domınguez, L.M. Lechuga, J.A. Plaza, V. Cadarso, European Patent PCT/EP05/380137 (2005). [154] D. Rugar, H.J. Mamin, P. Guethner, Appl. Phys. Lett., 55 (1989) 2588. [155] E. Bonaccurso, H.J. Butt, V. Franz, M. Stepputat, R. Raiteri (Eds), Cantilever Sensors and Nanostructures, Heidelberg, Germany, 2000. [156] G.G. Yaralioglu, A. Atalar, S.R. Manalis, C.F. Quate, Appl. Phys., 83 (1998) 7405. [157] E.B. Cooper, E.R. Post, S. Griffith, J. Levitan, S.R. Manalis, M.A. Schmidt, C.F. Quate, Appl. Phys. Lett., 76 (2000) 7405. [158] E.B. Cooper, E.R. Post, S. Griffith, J. Levitan, S.R. Manalis, M.A. Schmidt, C.F. Quate, Appl. Phys. Lett., 76 (2000) 3316. [159] M. Helm, J.J. Servant, F. Saurenbach, R. Berger, Appl. Phys. Lett., 87 (2005) 064101/1. [160] P.M. Nieva, N.E. McGruer, G.G. Adams, Micromech. Microeng., 16 (2006) 2618. [161] J. Reed, P. Wilkinson, J. Schmit, W. Klug, J.K. Gimzewski, Nanotechnology, 17 (2006) 3873. [162] P. Abgrall, N.T. Nguyen, Anal. Chem., 80 (2008) 2326. [163] C. Yi, C.-W. Li, S. Ji, M. Yang, Anal. Chim. Acta., 560 (2006) 1. [164] H. Andersson, A. van den Berg, Sens. Actuators B, 92 (2003) 315. [165] M.J. Mahoney, R.R. Chen, J. Tan, W.M. Saltzman, Biomaterials, 26 (2005) 771. [166] Y.C. Lin, M. Li, C.C. Wu, Lab Chip, 4 (2004) 104. [167] G. Tresset, S. Takeuchi, Biomed. Microdevices, 6 (2004) 213. [168] J.S. Soughayer, T. Krasieva, S.C. Jacobson, J.M. Ramsey, B.J. Tromberg, N.L. Allbritton, Anal. Chem., 72 (2000) 1342. [169] J.W. Hong, V. Studer, G. Hang, W.F. Anderson, S.R. Quake, Nat. Biotechnol., 22 (2004) 435. [170] A.J. de Melo, Nature, 442 (2006) 394. [171] R.G. Blazej, P. Kumaresan, R.A. Mathies, Proc. Natl. Acad. Sci. USA, 103 (2006) 7240. [172] P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M.R. Tam, B.H. Weigl, Nature, 442 (2006) 412. [173] G.M. Whitesides, Nature, 442 (2006) 368. [174] A. Gonz
alez, M. Herv
as, M.A. L
opez, M.C. Gonz
alez, A. Escarpa, Talanta, 74 (2007) 342. [175] G. Chen, Y.H. Lin, J. Wang, Talanta, 68 (2006) 497. [176] L. Marle, G.M. Greenway, Trends Anal. Chem., 24 (2005) 795. [177] M. Pumera, Trends Anal. Chem., 29 (2008) 269. [178] A. Escarpa, M.C. Gonz
alez, A.G. Crevillen, A.J. Blasco, Electrophoresis, 28 (2007) 1002.

3 Automation and Miniaturization of Sample Treatment 3.1

Introduction

In recent years, analytical instrumentation has evolved exponentially. A wide range of instruments capable of performing highly accurate determinations based on a variety of physical and/or chemical principles currently exists as a result. Such instruments include molecular and atomic absorption and emission spectrophotometers, vibrational and mass spectrometers, and electrochemical detectors such as conductimeters, to name a few. These advances in instrumentation have run in parallel with substantial improvements in separation science, particularly as regards liquid chromatography and capillary electrophoresis (CE). Also, the growing trend to miniaturization has facilitated the development of analytical equipment, enabling accurate, expeditious separation of a variety of components by using reduced amounts of samples and reagents. Despite the above advances, work at routine analytical laboratories continues to be slowed down by the preliminary operations of the analytical process. In fact, pretreating complex samples is often labour-intensive and time-consuming. Therefore, automating or integrating such operations could help simplify routine analyses. In this context, miniaturization of sample treatment devices provides an elegant, effective way of automating sample treatment. This chapter is concerned with sample treatment simplification through miniaturization and automation. Miniaturized sample preparation methods have been regarded as the most attractive techniques for the pretreatment of complex sample mixtures prior to the determination process. Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

94

Miniaturization of Analytical Systems

The effective online coupling of the miniaturized sample preparation and the instrumental separation technique (chromatographic or electrophoretic) enables several advantageous features. These advantages can be itemized as follows: . . . . .

low cost of operation due to the extremely low solvent consumption; high-speed analysis due to the simplification of the process; high efficiency; high selectivity when tailored systems are designed; development of environmentally friendly analytical procedures.

3.2 Simplification of Sample Treatment: Microextraction Techniques There is no doubt that miniaturization in the field of sample treatment has led to a simplification of this tedious step of the analytical processes. The miniaturization of sample treatment devices has resulted in the use of a lower sample volume, low sample processing time and low consumption of reagents [1]. In this context, it is important to present two trends in microextraction techniques: solid phase microextraction (SPME) and liquid phase microextraction (LPME). Microextraction techniques were developed to address the need to facilitate rapid sample preparation both in the laboratory and onsite at the system under investigation. In microextraction techniques, a small amount of the extracting phase (solid or liquid) is exposed to the sample for a well-defined period of time. A partitioning equilibrium is then reached between the sample and the extraction phase. This approach requires long times to reach equilibrium but is not affected by the convection conditions. Another interesting approach is based on utilizing a short partitioning time (pre-equilibrium conditions). This approach reduces the sample treatment time but requires exhaustive control of the convention/agitation conditions; in this way the amount of analyte extracted is related to time and can be quantified [1–3]. This extraction process generally follows the profile shown in Figure 3.1. 3.2.1

Calibration in Microextraction Processes

Calibration based on equilibrium extraction is very simple and has been the method most used to date. However, the fact that equilibrium times can range from seconds to months must be taken into account. When the equilibrium time is long, the calibration can be performed in the linear range (see Figure 3.1) by using kinetic calibration methods. Calibration based on equilibrium extraction is widely used in onsite sampling. The initial concentration of the target analyte can be calculated according to the

Extracted amount of analyte

Automation and Miniaturization of Sample Treatment

95

Near equilibrium

Kinetic regime

linear regime

t50

t95

Time

Figure 3.1 Typical extraction profile for microextraction techniques

following equation: CS ¼

Ce n ¼ ; Kes Kes Ve

ð3:1Þ

where Cs and Ce are the concentration of the target analytes in the sample and of the extraction phase respectively, n represents the total amount of analyte extracted, Ve is the volume of the extraction phase and Kes is the distribution coefficient of the analyte between the phases. This calibration only requires that the distribution coefficient is known. It is recommended for rapid equilibrium times. When the analysis is performed in the linear range (see Figure 3.1), quantification can be performed by assuming that the rate of mass transfer remains constant during the extraction process. Under this condition, the following equation is verified: nt Cs ¼ ð3:2Þ RS As can be seen, the concentration of the target analytes in the sample, Cs, is related to the total amount extracted, n, through the sampling rate of the transfer between phases Rs and the extraction time t. Typically, the calibration methods in microextraction techniques are limited to either the equilibrium regime or the linear regime. Recently, kinetic calibration methods that can be used for the entire extraction profile have been proposed [4–7]. However, these calibration procedures are not commonly used in methods involving a microextraction step. They are most used when the microextraction step is used for both preconcentration of the target analytes and to perform the sampling step, for example in environmental matrices [7,8]. 3.2.2

Solid Phase Microextraction (SPME) Techniques

In SPME processes, a small amount of extracting phase dispersed or bounded on a solid support is exposed to the sample. After a well-defined period of time,

96

Miniaturization of Analytical Systems

an amount of the target analyte is preconcentrated on the solid phase. Quantification, as was indicated in Section 3.2.1, can be performed based on equilibrium or on timed accumulation of analytes in the solid phase. Today, there are several procedures to implement an SPME, namely: Beds: Extraction of analytes on dispersed microparticles. Fibres: Extraction of analytes on coated fibres. Vessels: Extraction of analytes on the wall of chemically modified vessels. Stir Bar: Extraction of analytes on the wall of stirrers. In Tubes: Extraction of analytes on the wall of capillaries. Microextraction in Packed Syringe. Discs/Membranes: Extraction of analytes on the surface of disc or membrane filters. Figure 3.2 illustrates several implementations of SPME. In SPME techniques it is possible to distinguish between liquid and solid coatings. As shown in Figure 3.3, these coatings have resulted in two types of extraction mechanism, called absorptive and adsorptive extraction. In liquid coatings the target analytes penetrate on to the whole volume of the coating thanks to the salvation of the analytes by the coating molecules [9]. The penetration of the analytes is limited by the diffusion coefficient in the liquid coating. This type of extraction is called absorption. In the case of solid sorbents, the extraction occurs only on the surface. This is called adsorption [9].

Figure 3.2 Configurations of SPME systems. (Adapted from [2] with permission from Elsevier, Copyright 2000)

Automation and Miniaturization of Sample Treatment

97

Support (fibre core) Coating

Analyte molecule

Absorption

Adsorption–large pores

Adsorption–small pores Figure 3.3 Schematic representation of the extraction processes of analytes on solid phases. Two mechanisms can be distinguished: adsorptive and absorptive extraction. (From [9] reproduced by permission of the Royal Society of Chemistry)

Extraction Modes with Coated Fibre Three basic types of extractions can be performed using SPME, namely direct extraction, headspace configuration and a membrane protection approach. Figure 3.4 illustrates the differences between these modes. In the direct extraction mode, the coated fibre is directly inserted into the sample and the analytes are transported from the sample matrix to the extracting phase. To facilitate extraction, it is important to agitate the solution in order to transport the analytes from the bulk solution to the vicinity of the fibre [10,11]. In the headspace mode, the analytes need to be transported through the barrier of air before they can reach the coating. The main advantage of this mode is the protection of the fibre from high molecular mass and other nonvolatile interferences present in the sample matrix [12–14]. Another important advantage is that it allows modification of the matrix, for example to an extreme pH, without damaging the fibre. The amounts of analyte extracted into the coating from the sample vial at equilibrium using direct and

98

Miniaturization of Analytical Systems

Figure 3.4 Modes of SPME operations: (a) direct extraction; (b) headspace configuration; (c) membrane-protected configuration

headspace sampling are identical so long as the sample and gaseous headspace volumes are the same. This is because the equilibrium concentration is independent of fibre location in the sample/headspace system. If the above condition is not satisfied, a significant sensitivity difference between the direct and headspace approaches exists for very volatile analytes only [15]. The choice of sampling mode has a very significant impact on extraction kinetics, however. When the fibre coating is in the headspace, the analytes are first removed from the headspace, then indirectly extracted from the matrix. Overall, mass transfer to the fibre is typically limited by mass transfer rates from the sample to the headspace. Therefore, volatile analytes are extracted faster than semivolatiles since they are at a higher concentration in the headspace, which contributes to faster mass transport rates through it [16]. Temperature exerts a significant effect on the kinetics of the process by determining the vapour pressure of the analytes. In fact, the equilibrium times for volatiles are shorter in the headspace SPME mode than for direct extraction under similar agitation conditions. This outcome is produced by two factors: (i) A substantial portion of the analyte is in the headspace prior to extraction. (ii) Diffusion coefficients in the gaseous phase are typically four orders of magnitude larger than in liquid media. Since concentrations of semivolatiles in the gaseous phase at room temperature are typically small, however, overall mass transfer rates are substantially lower and result in longer extraction times. They can be improved by using very efficient agitation or by increasing the extraction temperature [8,17]. Extraction Modes with In-tube Solid Phase Extraction Figure 3.5 shows that there are two fundamental approaches to in-tube SPME: active or dynamic, when the analytes are passed through the tube; and passive or

Automation and Miniaturization of Sample Treatment (a)

99

(b)

attachment hub

sealing septum

septum-piercing needle

microtubing

coated fused silica fibre

sample

sample flow

Figure 3.5 Comparison of (a) passive versus (b) dynamic modes of in-tube extractions

static, when the analytes are transferred into the sorbent using diffusion. In each of these approaches, the coating may be supported on a fused-silica rod, or on the inside of a tube or capillary [18]. In the dynamic configuration, we can assume for example the use of a piece of fused-silica capillary coated with a thin film of extracting phase. In these geometric arrangements, the front of analyte migrates through the capillary with a speed proportional to the linear velocity of the sample, and inversely related to the partition ratio [19]. For short capillaries with a small dispersion, the extraction time can be assumed to be similar to the time required for the centre of the band to reach the end of the capillary. The extraction time is proportional to the length of the capillary and inversely proportional to the linear flow rate of the fluid [20]. The extraction time also increases with an increase in the coating/sample distribution constant and with the thickness of the extracting phase but decreases with an increase in the void volume of the capillary. An increase in the coating/sample distribution constant produces an increase in the absolute amount extracted. It has been shown that increases in the amount extracted can be achieved, in many cases, by preconditioning the capillary with methanol or some other appropriate solvent prior to extraction. Enhancement has even been observed when a plug of methanol is aspirated into the capillary before the sample is drawn in, and following the use of a sample plug in the capillary during the extraction aspirate/dispense steps. This is analogous to the solvent preconditioning used to enhance solid phase extraction (SPE).

100

Miniaturization of Analytical Systems

Figure 3.6 Schematics of different implementations of in-tube SPME. (From [24] with permission from American Chemical Society, Copyright 1997)

In practice, in-tube SPME is implemented by replacing a section of the tubing in a commercially available autosampler and then programming the autosampler to pass sample in and out of the extraction capillary until equilibrium or a suitable extraction level is reached [21–24]. Several options for implementing in-tube SPE are summarized in Figure 3.6. The removal of analytes from a tube is an elution problem analogous to frontal chromatography and has been discussed in detail [24,25]. In general, if the desorption temperature of a GC is high and thin coating is used, all the analytes will be in the gaseous phase as soon as the coating is placed in the injector, and the desorption time corresponds to the elution of two void volumes of capillary. For liquid desorption, for example into liquid chromatography equipment, the desorption volume can be even smaller since the analytes can be focused at the front of the desorption solvent [26–28]. In addition to the measurement of analyte concentration, it is possible to obtain an integrated sampling with a simple SPME system [29,30]. This is particularly important in field measurements, where changes of analyte concentration over time and space must often be taken into account.

Automation and Miniaturization of Sample Treatment

101

When the extracting phase is not exposed directly to the sample, but is contained in a protective tubing (needle) without any flow of the sample through it, the extraction occurs through the static gas phase present in the needle. The integrated system can consist of the extraction phase coating the interior of the tubing, or it can be externally coated fibre withdrawn into the needle. These geometric arrangements represent a very powerful method able to generate a response proportional to the integral of the analyte concentration over time and space (when the needle is moved through the space). In these cases, the only mechanism of analyte transport to the extracting phase is diffusion through the gaseous phase contained in the tubing. During this process, a linear concentration profile is established in the tubing between the small needle opening, characterized by surface area and the distance between the needle opening and the position on the extracting phase. Fibre Coatings As was indicated above, the efficiency of the extraction process is dependent on the distribution constant. This is a characteristic parameter that describes the properties of a coating and its selectivity toward the analyte versus other matrix components. Specific coatings can be developed for a range of applications. Coating volume determines method sensitivity. Therefore, it is important to use the appropriate coating for a given application. This is clearly demonstrated in Figure 3.7, which

Figure 3.7 Total ion current GC/MS chromatogram of bencene, toluene, ethylbenzene and o, m, p-xylene (BTEX) and phenol in water extracted with (a) a poly(dimethylsiloxane) coating and (b) a poly (acrylate) coating. (From [31] with permission from American Chemical Society, Copyright 1994)

102

Miniaturization of Analytical Systems

compares the performance of two different coatings for analysis of polar and nonpolar compounds from aqueous matrix. The distribution constant and the sensitivity of the method drop over two orders of magnitude for o-Xylene, and increase by an order of magnitude for 2,4-dichlorophenol when the film is changed from nonpolar poly(dimethylsiloxane) to polar poly(acrylate) polymer [31]. Coating selection and design can be based on chromatographic experience. To date, several experimental coatings have been prepared and investigated for a range of applications. In addition to liquid polymer coatings such as PDMS for general applications, other more specialized materials have been developed. For example, ion-exchange coatings were used to remove metal ions and proteins from aqueous solutions [32–34], liquid crystalline films to extract planar molecules and carbowax for polar analytes [35,36], metal rods to electrodeposit analytes, pencils to extract pesticides [37] and Nafion coating to extract polar compounds from nonpolar matrices [38,39]. Polypyrole coatings have recently been developed to extract polar or even ionic analytes and possibly to explore the conductive polymer properties of the polymer [40]. This might involve applying a charge to the polymer during extraction in order to selectively extract analytes of interest and then reversing the charge to facilitate desorption. Coating polymers are versatile materials in which molecular/analyte recognition can be achieved in different ways, including: (i) the incorporation of counterions that introduce selective interactions; (ii) the utilization of the inherent and unusual multifunctionality (hydrophobic, base–acid and p–p interactions, polar functional groups, ion-exchange, hydrogen bonding, electroactivity, etc.) of the polymer; (iii) the introduction of functional groups to the monomers; (iv) the co-deposition of metals or other monomers within the polymer; and (v) the application of appropriate electrochemical potentials. A very pronounced difference in selectivity toward target analytes and interferences can be achieved by using surfaces common to affinity chromatography. Using the method of polymer imprinting [41], antibody mimics can be generated with specifities to an analyte of choice. Briefly, the desired affinity can be introduced by adding an amount of the compound of interest to the polymerization reaction. This ‘pattern’ chemical may be removed after polymerization, leaving vacant sites of a specific size and shape, suitable for binding the same chemical again from an unknown sample. In this technique, we have observed that nonspecific binding should be controlled for, but that enhancements in sensitivity are seen, particulary at low analyte concentrations. Commercial Devices The first commercial version of the laboratory SPME device was introduced by Supelco in 1993 (see Figure 3.8). This was later improved by the inclusion of

Automation and Miniaturization of Sample Treatment

103

Figure 3.8 Design of the first commercial SPME device made by Supelco

an adjustment for the depth of the fibre with respect to the end of the needle, which allows control of the exposure depth in the injector and extraction vessel. The device incorporates such useful features as colour marking of the fibre assemblies, to distinguish between various coating types. In addition to standard PDMS coatings of various thicknesses and polyacrylates (PAs), Supelco developed new mixed phases based on solid/liquid sorption, such as Carbowax–DVB and PDMS–DVB. Supelco also introduced a high-performance liquid chromatography (HPLC) interface (Figure 3.9) that integrates the original concept with the injection valve (Figure 3.10). Varian has developed an SPME autosampler based on its 8000 GC autosampler system, taking advantage of the facts that the SPME device is analogous to a syringe in its operation and that after desorbtion the coating is cleaned and ready for reuse [42]. The major challenge is to incorporate agitation and temperature control, as well as other enhancements such as fibre internal cooling or dedicated injectors. One improvement is an SPME system that incorporates an agitation mechanism consisting of a small motor and a cam to vibrate the needle. The fibre in this design works as a stirrer. The vibration causes the vial to shake and the fibre to move with respect to the solution; the result is a substantial decrease in equilibration times compared to a static system. This mode of agitation simplifies fibre handling because it does not require the introduction of foreign objects into the sample prior to extraction. Recently, Varian has reached an agreement with CTC Analytics of Switzerland to incorporate SPME sampling on its CombiPal autosampler. This is a robotic system with a great deal of flexibility for programming SPME analyses. Samples are loaded

104

Miniaturization of Analytical Systems SPME fibre holder

Septum-piercing needle Needle guide Double tapered ferrule

Compression union

SPME fibre Solvent desorption chamber To interface

Static desorption (no flow) HPLC column

Solvent from syringe Waste

Valve

Mobile phase from pump

From interface

To interface From interface

Sample injection HPLC column

Solvent from syringe

Waste

Valve

Mobile phase from pump

Figure 3.9 Schematic diagram of the Supelco SPME–HPLC interface

on to trays accomodating five vial sizes, and samples are heated and agitated during extraction, using a separate sample-preparation chamber. To facilitate agitation, the extraction chamber is rotated at a programmable rotation speed during extraction. Additional vials/stations are present to accommodate wash solutions, derivatizing agents, temperature control, derivatization and fibre conditioning, in order to facilitate operation of SPME at optimal conditions. While the built-in software can be used to perform basic SPME analyses, extra programming flexibility is provided by the Cycle Composer software, available as an accessory. We expect that this new instrument, with its greatly enhanced flexibility, will significantly expand the range of SPME methods amenable to automation. New coatings and devices are expected to follow, as interest in SPME grows along with the unprecedented number of new applications appearing in the literature. SPME coupled to high-speed gas chromatography is a good combination for the performance of rapid and cost-effective investigations in the field, even of complex organic samples. As discussed above, SPME is particularly suited to fast GC as it is solvent-free and the thin coatings can provide very fast desorption of analytes at high temperatures. Some instrumental modifications were recently performed in order to achieve successful fast separations. A portable system was optimized for SPME–fast GC field investigations and was commercialized by SRI Instruments [43]. The instrument was tested in combination with a flame ionization detector (FID), a photoionization detector (PID) and a dry electrolytic conductivity detector (DELCD). A dedicated injector, presented in Figure 3.11, was mounted on the portable system in order to use SPME for

Automation and Miniaturization of Sample Treatment

105

Figure 3.10 SPME–HPLC interface: (a) 1/16 in. stainless steel (SS) tee; (b) 1/16 in. SS tubing; (c) 1/16 in. PEEK tubing (0.02 in. ID); (d) two-piece, finger-tight PEEK union; (e) PEEK tubing (0.005 in. ID) with a one-piece PEEK union

Figure 3.11 Design of dedicated injector. 1, modified Swagelok fitting; 2, stainless steel tubing; 3, moulded septum; 4, nut; 5, needle guide; 6, nut; 7, blind ferrule; 8, stainless steel tubing; 9, 0.53 mm ID fused-silica capillary; 10, contact

106

Miniaturization of Analytical Systems

high-speed separation. The injector guarantees very fast thermal desorption of the analytes from the SPME fibre [43]. The injector for high-speed GC should produce as narrow an injection band as possible. Internal volumes of regular injector ports are too large (e.g. split/splitless injector), since they have been designed to accommodate large volumes of gaseous samples or vapours produced by solvent injection. Thermal focusing for separation improvement is not convenient for fast separations, since temperature programming is impractical for high-speed GC, hence an injector port with a small internal volume was required for this application. Also, very fast thermal desorption from the SPME fibre was required to produce a narrow injection band and achieve effective separation. In the dedicated injector for SPME–fast GC, the injector port was maintained cold during needle introduction and was rapidly heated only when the fibre was exposed to the carrier gas stream. Small custom modifications of the commercial devices can open new possibilities. For example, the addition of input and output connections to the autosampler vial allows the system to be used to continuously monitor flowing streams, as shown in Figure 3.12. The flow-through design facilitates agitation of the sample. Alternatively, when a connection is added directly to the needle of the autosampler syringe, the system can analyse samples present in the vial without needing to expose the fibre. The fibre containing the extracted analytes in its coating can then be introduced to the instrument for desorption. SPME has been introduced as a modern alternative to traditional sample preparation technology, and it is able to address many of the requirements put forward by analytical researchers. This technique eliminates the use of organic solvents, and substantially shortens the time of analysis and allows convenient

SPME-FIBRE INLET

OUTLET

FLOW-CELL

Figure 3.12 Representation of an automated SPME device. Inlet and outlet to be connected through the flow cell can be selected from the carousel.

Automation and Miniaturization of Sample Treatment

107

automation of the sample preparation step. It can integrate sampling with sample preparation, which makes it suitable for onsite analysis and process monitoring. The configuration and operation of SPME devices are very simple. For example, for the coated fibre implementation of the technology, anyone who knows how to use a syringe is able to operate the SPME device. In the case of automated in-tube extraction for HPLC, fitting a piece of the GC capillary into the system and then turning on the autosampler is all that is required to start its operation. The technology is designed to greatly simplify sample preparation. This feature, however, creates the false impression that the extraction is a simple, almost trivial process. This misunderstanding frequently results in disappointments. It should be emphasized that the fundamental processes involved in SPME are similar to those in more traditional techniques and therefore the difficulty of developing successful methods is analogous. The nature of target analytes and the complexity of the sample matrix determine the level of difficulty in accomplishing a successful extraction. The simplicity, speed and convenience of the extraction devices primarily impact the costs of practical implementation and automation of the developed methods. The potential savings in analysis time, reduced solvent use and apparent simplicity of SPME techniques will continue to attract interest among analytical chemists searching for improved analysis methods. So long as analysts have a sound understanding of the theory and principles behind this technique, good accuracy and precision will follow. Stir Bar Sorptive Extraction (SBSE) Stir bars are coated with a layer (typically 0.5–1 mm thick) of polydimethylsiloxane and then used to stir aqueous samples, thereby extracting and enriching solutes into the polydimethylsiloxane coating. After extraction, either thermal desorption or liquid desorption can be used. This technique is named stir bar sorptive extraction (SBSE) [3]. The basic principles of SBSE are thus identical to SPME using polydimethylsiloxane-coated fibres, but the volume of extraction phase is 50–250 times larger. Early SBSE applications have been reviewed by David et al. [44], and more recently Kawagushi et al. reviewed SBSE applications in environmental and biomedical analysis, mainly focusing on SBSE in combination with in situ derivatization [45]. In this review article, the principle of SBSE is discussed, some practical issues are considered and an overview is given of SBSE in different application areas, including environmental analysis, food analysis and life science and biomedical analysis. Figure 3.13 shows the influence of Ko/w and phase ratio on extraction efficiency. For a given phase ratio (sample volume/PDMS volume), an ‘S-shape’ curve is obtained, where the position of the curve depends on the b ratio. For SPME, the volume of polydimethylsiloxane is approximately 0.5 ml. For a sample of 10 ml, the

108

Miniaturization of Analytical Systems

Figure 3.13 Theoretical recovery (%) in function of solute logKo/w for SPME (100 m fibre, 0.5 ml PDMS) and SBSE (1 cm  0.5 mm df, 25 ml PDMS) and 10 ml sample volume. Equilibrium sampling is assumed. (Reprinted from [3] with permission of Elsevier, Copyright 2007)

phase ratio is thus 20 000. This results in poor recoveries for solutes with low Ko/w values. A solute with logKo/w ¼ 3 (e.g. naphthalene, Ko/w ¼ 3.17) is only recovered at 4.8%. The increased recovery obtained by SBSE in comparison with SPME has been demonstrated by different groups, using PAHs [46,47] and pesticides [48] as test solutes. Stir bars of 1 or 2 cm length coated with a 0.5 or 1 mm layer have been made commercially available. A magnetic rod is encapsulated in a glass jacket, on which a polydimethylsiloxane coating is placed. Direct contact between metal and PDMS has been found to catalyse polymer degradation during thermal desorption. Most applications have been performed using pure polydimethylsiloxane coatings. Recently, other phases have also come under development and been tested. Liu et al. [49,50] described the use of sol-gel technology to obtain thin (30 mm) layers of PDMS on stirring rods. Lambert et al. [51] coated restricted-access material (RAM) on stir bars for the extraction of caffeine (logKo/w ¼ 0.1) and metabolites in biological fluids. The RAM-coated stir bars were used in combination with liquid desorption, followed by liquid chromatography. Bicchi et al. [52] described the use of a dual-phase stir bar both in SBSE (immersion) mode and in headspace (HSSE) mode. These new stir bars consisted of an outer PDMS coating with a carbonadsorbent material inside. Magnetic stirring is possible, with two small magnets placed at the ends of the stir bar. This dual-phase device, whereby sorption is combined with adsorption (on the carbon material), showed increased recovery of very volatile compounds emitted from plant material and of more polar solutes in water samples. Alternative designs have also been used [53–55]. Large ‘PDMS rods’ or glass tubes (without magnet) with dimensions up to 8 cm long and coated

Automation and Miniaturization of Sample Treatment

109

with 250 ml PDMS have been used, in both SBSE (using shaking) and HSSE mode. Zhao et al. [55] used a PDMS rod for passive sampling, and time-weighted average (TWA) concentrations of polycyclic aromatic hydrocarbons were measured in a water stream. SBSE of a liquid sample is performed by placing a suitable amount of sample in a headspace vial or other container. The stir bar is added and the sample is stirred, typically for 30–240 minutes. The extraction time is controlled kinetically, determined by sample volume, stirring speed and stir bar dimensions, and must be optimized for a given application. Optimization is normally accomplished by measuring the analyte recovery as a function of the extraction time. The highest recovery is obtained under equilibrium conditions, where no additional recovery is observed when the extraction time is increased further. However, as with SPME, selected sampling times are often much shorter than the time needed to reach full equilibrium. These non-equilibrium conditions often result in sufficient sensitivity and good repeatability, while the extraction time is not excessively long. Normally, SBSE is applied to the extraction of aqueous samples containing low concentrations of organic compounds. Samples containing high concentrations of solvents, detergents, etc. should be diluted before extraction. For the extraction of highly apolar solutes, such as PAHs and PCBs from water samples, an organic modifier is often added to minimize wall adsorption. The organic modifier will affect the PDMS–water partitioning, but for these applications the overall effect is higher recovery. For the extraction of solutes covering a broad polarity range, optimization of the organic modifier concentration is necessary, and even dual extractions (with and without modifier) are used. SBSE is also used, in combination with in situ derivatization. Polar solutes with low Ko/w values will normally result in very low recoveries. In general, derivatives have higher Ko/w values, and consequently higher recovery and higher sensitivity are obtained. Typical derivatization reactions that can be performed in aqueous media include acylation of phenols using acetic anhydride, esterification of acids and acylation of amines using ethyl chloroformate, and oximation of aldehydes and ketones using pentafluorobenzyl hydroxylamine (PFBHA). Solutes can also be derivatized after extraction in order to improve chromatographic performance and/or detectability. This can, for instance, be performed by silylation. Silylation cannot be performed in aqueous media, but is possible in combination with thermal desorption. Several examples of in situ derivatization and post-extraction derivatization are discussed in a recent review by Kawaguchi et al. [56]. After extraction, the stir bar is removed, dipped on a clean paper tissue to remove water droplets, and introduced into a thermal desorption unit. In some cases, it is recommended that the stir bar be rinsed a little with distilled water to remove adsorbed sugars, proteins and other sample components. This step will avoid the formation of nonvolatile material during the thermal desorption step. Rinsing does not cause solute loss as the sorbed solutes are present inside the

110

Miniaturization of Analytical Systems

polydimethylsiloxane phase. Alternatively, liquid desorption can be used. Typically, the stir bar is placed in a small vial (2 ml, or a vial with insert) and desorption is performed with apolar solvents (hexane), followed by GC analysis, or with polar solvents (methanol, acetonitrile), followed by LC analysis. It has been demonstrated [57] that loaded stir bars can be stored at 4  C for 1 week without loss of solutes. This opens interesting prospects for onsite sampling and extraction. The loaded stir bars are sent to the laboratory for analysis, not the samples. After thermal or liquid desorption, the stir bars can be reused. Typically, the lifetime of a single stir bar is 20 to more than 50 extractions, depending on the matrix. For headspace sampling, the stir bar can be placed above a liquid or solid sample, and special devices are available to hold it. PDMS-coated stir bars have also been applied for passive air sampling or for continuous monitoring of persistent organic pollutants in water. Paschke et al. [58] used a stir bar inside a semipermeable membrane. The solutes were enriched based on the diffusion through the membrane, then sorbed into the PDMS material. SBSE was first applied in environmental analysis. The main advantage of the technique is that it can be applied to volatile organic compounds (VOCs) and semivolatile compounds [59]. Most applications, however, deal with semivolatile compounds, typically eluting after decane on an apolar column in GC. For the determination of PAHs in water samples, sensitivities at 1 ng/l can be obtained with RSDs below 15% (at 10 ng/l) [60]. An important aspect in the extraction of these very apolar solutes (logKo/w of benzo[a]pyrene ¼ 6.11) is the reduction of wall adsorption in the extraction vessel. For this purpose, methanol (5–10%), sodium chloride or hyamine (ionic tenside) are added prior to extraction. Leon et al. [61] also demonstrated that a high desorption flow (100 ml/minute) is needed for efficient thermal desorption of the high-molecular weight PAHs (dibenzo[ah] anthracene, benzo[ghi]perylene). SBSE has also been successfully applied to the extraction of pesticides and polychlorinated biphenyls (PCBs) [62]. Popp et al. [63] demonstrated the analysis of PCBs including coplanar congeners (PCB 77, PCB 126 and PCB 169) in water at sub-ng/l levels. For this analysis, 20% methanol was added to the sample to reduce wall adsorption and interaction with humic acids. A multiresidue method for PAHs, PCBs and pesticides was validated by Leon et al. [64] according to ISO/EN 17025. A 100 ml sample was saturated with sodium chloride and extracted (overnight) over 14 hours using a 2 cm stir bar coated with a 0.5 mm thick PDMS film (0.5 mm df). Using thermal desorption combined with GC–MS in scan mode, LODs were in the order of 0.1–10 ng/l (lowest for apolar solutes with high MS response, highest for more polar solutes such as simazine) and repeatability was in the order of 7% at 50 ng/l. The method was used in an interlaboratory trial and the results showed excellent agreement with those obtained by classical methods [65]. Recently, the analysis of odorous compounds in drinking water has also received lots of attention. Compounds such as 2-methylisoborneol (MIB), geosmin and

Automation and Miniaturization of Sample Treatment

111

chlorinated and brominated anisoles, originating from the corresponding phenols, have odour thresholds of less than 10 ng/l. Using SBSE, these compounds can be extracted with high recovery from drinking water samples. In comparison to labourintensive techniques such as closed loop stripping, the SBSE method provides a much higher sample throughput with better sensitivity, reproducibility and accuracy [65–67]. Nakamura et al. [65] demonstrated that for the target compounds, the sensitivity was 10–50 times better than that of SPME. The method was validated according to an official norm (AFNOR XP T 90–210) by Benanou et al. [67]. A 100 ml sample was extracted using a 2 cm · 0.5 mm df stir bar over 120 minutes. After extraction, thermal desorption and GC–MS in SIM mode were used. Limits of quantification were below odour threshold for all target solutes (0.5 ng/l for geosmin, 1 ng/l for methyl-isoborneol and 0.1 ng/l for trichloroanisole), while the repeatability was better than 15% at 2 ng/l. Parallel sniffing was also used to detect other compounds responsible for off-odours in water [68]. SBSE can be combined with in situ derivatization. Although some papers have described the extraction of phenols without derivatization [69], better recoveries are obtained after in situ derivatization. This can be achieved by adding the derivatization reagent (acetic anhydride) to the sample after pH adjustment (potassium carbonate). The extraction can be performed simultaneously with derivatization. The acylated phenols also perform better during GC analysis, and detection limits down to ng/l levels can be obtained [70,71]. The extraction was performed by SBSE using a 10 mm · 0.5 mm df PDMS stir bar on a 10ml sample after addition of 0.5 g K2CO3 and 0.5 mL acetic anhydride. After extraction, the phenyl acetates were thermally desorbed in splitless mode and analysed on a 30 m · 0.25 mm inner diameter (ID) · 0.25 mm df HP-5MS column using MS detection in selected ion monitoring (SIM) mode. The chromatogram shows the extracted ion chromatograms for the mono-chlorophenols (as acetates) (ion 128, 3 isomers), dichlorophenols (ion 164, 6 isomers, 2 isomers not separated), trichlorophenols (ion 196, 5 isomers), tetrachlorophenol (ion 232, 1 isomer) and pentachlorophenol (ion 266). The excellent sensitivity is clearly illustrated. The same derivatization method was used by Itoh et al. [72] for the analysis of hydroxy-PAHs. Applications of SBSE in food analysis can be divided into three categories: the analysis of minor constituents (volatiles, additives), the determination of trace compounds responsible for off-odours (aldehydes, haloanisoles) and the analysis of trace contaminants (pesticides). In food analysis, SBSE is used for the analysis of volatiles in plant material [73–81] and in fruit, including strawberries [82], grapes [83], snake fruit [84] and raspberries [85]. Both SBSE (immersion) mode and headspace (HSSE) mode were used. SBSE and HSSE were also used for the analysis of volatile constituents and aroma compounds in coffee [86], beer [87], wine [88–91] and whiskey [92]. Volatiles in vinegar [93], volatiles emitted by fungi (including mycotoxin-producing fungi) [94] and decomposition products of peptides and amino acids [101] were measured using SBSE. In general, sorptive

112

Miniaturization of Analytical Systems

extraction allows the detection of important flavour and fragrance compounds below olfactory threshold values, as demonstrated by the analysis of compounds released by oak into wine [95]. These compounds included furfural, guaicol, vanillin and whiskey lactone. Unlike in other techniques, no artefact formation, such as transformation of vanillin into the corresponding acetals, was observed. Alves et al. [91] used chemometric data treatment on the results obtained by SBSE–TD–GC–MS to classify Madeira wines according to variety and age. The information obtained on trace levels of monoterpenes in particular was found to increase the discriminating power. Bicchi et al. [96] studied the impact of phase ratio, PDMS volume and sampling time, and temperature on the recovery of volatile organic compounds in the essential oils. It was found that by using HSSE, a concentration factor of 1000 in comparison to static headspace could be obtained, and limits of detection were in the nmol/l range. Besides the analysis of relatively volatile compounds, preservatives such as benzoic acid, sorbic acid and parabens were also measured. For benzoic acid (pKa ¼ 4.21), the sample was adjusted to low pH (pH 2). At pH > pKa, the ionic form results in very low recoveries (logKo/w < 0). Analysis was performed by thermal desorption in combination with GC–MS. Under these conditions, excellent sensitivities (8 ppb) and repeatability (RSD < 5%) were obtained using a 1 cm 1 mm df stir bar in a 25 ml sample [97]. Nonvolatile compounds have also been analysed, such as hop bitter acids in beer, using SBSE followed by liquid desorption and micellar electrokinetic chromatography (MEKC) [98]. A special application was the use of a stir bar, placed in a special holder, for the detection of aroma compounds released in the mouth during wine tasting [106]. Besides the analysis of minor constituents, SBSE has also been used for trace analysis of off-odours. The most important application is the analysis of trichloroanisole (TCA) in wine. This compound is responsible for cork taint. Using SBSE, TCA can be detected at sub-ng/l concentrations, below the odour threshold values [99]. Trichloroanisole can also be detected in cork, using HSSE and thermostating the cork samples at 100  C [100]. As for food applications, SBSE can be applied to biological samples, including urine, plasma, saliva, etc., for the analysis of both volatiles and semivolatiles. In the biochemical/life-science application area, special attention should be paid to the possibilities of in situ derivatization combined with SBSE, since often the target compounds are quite polar (e.g. metabolites). A very interesting application area is described by Soini et al. [101]. SBSE was used to characterize chemical signal compounds in animal urine samples and gland excretion. Based on the detailed profiles obtained, gender and age could be differentiated using chemometric data processing. Wahl et al. described the analysis of barbiturates in urine [102]. Other applications of SBSE for the determination of environmental contaminants in biological samples include the determination of PCBs in sperm [103] and the determination of phenols and chlorophenols in urine [104]. For the determination

Automation and Miniaturization of Sample Treatment

113

of pharmaceuticals and drugs of abuse (including the metabolites), both direct SBSE and SBSE combined with in situ derivatization are used. More recently, SBSE with in situ derivatization was used by Stopforth et al. in various diagnostic analyses, including the determination of tuberculostearic acid in sputum (a marker of tuberculosis) [105], the determination of the oxidative stress marker 4-hydroxynonenal in urine [106] and the determination of estrone and estradiol (as markers in female urine) [107]. For this application, 1 ml urine was extracted by SBSE using in situ derivatization with PFBHA (oximation of aldehyde). The extracted oxime was further derivatizated by placing the loaded stir bar in the headspace of the acylating reagent (acetic anhydride), thereby derivatizing the hydroxyl function. Finally, thermal desorption–GC–MS in SIM mode was used. The PFBoxime-acetate derivative of hydroxyl-nonenal could be detected in urine at pg/ml levels [108]. Microextraction in a Packed Syringe Microextraction in a packed syringe (MEPS) [108–116], or in a packed pipette tip, is a relatively new technique [108–116]. A small amount (1 mg) of the SPE material is inserted into a syringe (100–250 ml) or pipette tip as a plug, which is secured by frits at either end. The sample is withdrawn through the solid phase plug, and the analytes adsorb on the SPE material. The SPE material can then be washed before being eluted with a suitable organic solvent. The procedure can be performed automatically by an autosampler and even connected online with a GC injector if large-volume injection (LVI) techniques are used. Most current MEPS applications involve online connection with LC rather than GC, since with the automated procedure it is not easy to dry the SPE material before elution and small amounts of water get injected into the GC instrument. In addition, the elution is typically performed with relatively polar solvents such as methanol, which in GC can cause problems. Online connection with GC is possible, but a very careful optimization of injection parameters is then required. Packed pipette tips are typically used in offline mode. The selection of SPE material is the most critical parameter in the optimization of the extraction. In the case of offline procedures, more than one washing step can be carried out, and drying can be done by placing the pipette tip into vacuum or by using a drying agent to remove water from the final extract. The elution solvent, type of injector and injection conditions are critical if MEPS is to be coupled online. In an automated system, where drying typically is not possible, the eluent should be miscible with water. It should also be sufficiently volatile. Methanol is used in most applications, though it cannot be considered a very good choice, especially for LVI to GC. The programmable temperature vaporizer (PTV) interface is used in most applications, usually with solvent split technique. The choice of injection rate, injection temperature and flow rate is particularly critical for volatile analytes.

114

3.2.3

Miniaturization of Analytical Systems

Liquid Phase Microextraction (LPME) Techniques

A miniaturized version of traditional liquid–liquid extraction (LLE) has been developed. It is called either liquid phase microextraction (LPME), single-drop microextraction (SDME), solvent microextraction or liquid–liquid microextraction. In this technique, a microdrop of solvent is suspended from the tip of a conventional microsyringe and then either exposed to the headspace of the sample or immersed in a sample solution in which it is immiscible. The operation of SDME has been described in several literature reviews [117]. Headspace SDME is similar to traditional headspace sampling in that volatiles are sampled from the vapours above the sample, so that interference from the sample matrix is avoided. A variety of methods and specialized pieces of equipment are available for this purpose. Another type of microscale LLE is dispersive liquid–liquid microextraction (DLLME). In this method, the appropriate mixture of extraction solvent and disperser solvent is injected rapidly into an aqueous sample, resulting in a cloudy solution. The cloudy solution consists of minute droplets of disperser solvent and extraction solvent in aqueous solution. In fact, this technique can be regarded as multiple-drop microextraction. In DLLME, the instantaneous mixing of the three components ensures equilibration within a few seconds due to the infinitely large interface between the fine extractor droplets and the aqueous solution. Thus, transfer of analytes from the aqueous phase to the organic phase is very short compared with typical equilibrium extraction times for SPME and LPME. To separate the phases, centrifugation is required. Miscibility in organic phase (extraction solvent) and aqueous phase is the main criterion for the selection of the disperser solvent. Typical disperser solvents are acetone, acetonitrile and methanol [117–119]. Liquid phase extraction with microdrops can be carried out as a two-phase or three-phase system. The three-phase system can be regarded as a micro LLE-back-extraction system which exploits the acid–base character of the analytes to achieve simultaneous enrichment and cleanup. The simplest mode of LPME is the SDME, in which analytes are extracted from a stirred aqueous sample into a drop of organic solvent (1–3 ml) suspended from the needle of a microsyringe. Once the extraction is complete, the drop is retracted into the syringe and injected into the chromatographic system for analysis. The main limitation of LPME is the drop instability. The droplet may be lost from the needle tip of the syringe during the extraction, particularly when samples are stirred vigorously, or emulsion may form, particularly in the extraction of biological samples. The advantages of the single-drop technique include the suitability of a wide range of solvents, low solvent consumption and low cost. In addition, no preconditioning is required, unlike in SPME, where the fibre must be pretreated for efficient extraction. Also, memory effects are avoided. However, because solvents must have a relatively high boiling point, the method is not suitable for the extraction of highly volatile analytes.

Automation and Miniaturization of Sample Treatment

115

Parameters that have been considered for SDME include solvent type, size of drop, shape of needle tip, temperature of sampling, equilibration and sampling (extraction) time, effect of stirring and ratio of headspace volume to sample volume [120,121]. Naturally, pH and salt addition and other parameters that affect the LLE process must be considered as well. The choice of solvent is critical. In headspace analysis, the solvent should not be too volatile, because otherwise the drop will evaporate during the extraction. At the same time, the solvent should not elute together with analytes, and it should not, therefore, be too nonvolatile. Neither should it be insoluble in the sample solvent. Naturally, the solvent in the drop should dissolve well to analytes. Typically, n-octyl acetate, isoamyl alcohol, undecane, octane, nonane and ethylene glycol are used as solvents. Drop volumes are typically 1–2 ml, as solvent drops of these volumes are relatively stable. In most cases, the extraction is done at room temperature, because even though increase of temperature usually increases the extraction efficiency, it also leads to drop instability and thereby to lower reproducibility. Stirring must also be done at low speed because otherwise it affects the stability of the drop. One way of minimizing the problems of drop instability in LPME is to add a polymeric membrane to serve as a support for the extracting solvent. This not only enables the use of larger volumes but also acts as a physical barrier between the phases. The membrane is usually made of polypropylene or other porous hydrophobic material. Polypropylene is highly compatible with a broad range of organic solvents and, owing to a pore size of 0.2 mm, the membrane strongly immobilizes the organic solvent in the pores. A further benefit of the use of a membrane is that, owing to the pore structure, the concentration of high-molecular mass compounds in the sample extract is reduced. Membrane-assisted liquid extraction can thus be considered a subtechnique of LPME. Membrane-assisted liquid extraction can be performed in a hollow fibre, a flat-sheet membrane module or a membrane bag [124,125]. The membranes can be microporous, where the pores are filled with an organic solvent, or nonporous. In the latter case, the extraction mechanism is different, as analytes must diffuse through the membrane before extraction to the acceptor phase. Figure 3.14 shows different systems utilized in membrane-assisted liquid extraction. The extraction can be static or dynamic, depending on the membrane module. As in LPME, two-phase and three-phase systems are possible. Membrane extraction systems based on hollow fibres closed at both ends are also available, and can be used for onsite sampling [126]. The two-phase mode is typically utilized in microporous membrane LLE (MMLLE) and the three-phase mode in supported liquid membrane LLE (SLM). The MMLLE version is mostly used in combination with GC, while SLM, in which the final extract is typically aqueous, is used with LC and electro-driven separations. In two-phase extraction, the extraction is based on diffusion of the analytes from the aqueous sample (donor) to the organic acceptor solution. The extraction process depends on the partition coefficients of the analytes. Static membrane-assisted extraction typically employs

116

Miniaturization of Analytical Systems

Figure 3.14 Different solutions for membrane-assisted liquid extraction: (a) a U-shaped hollow fibre; (b) a rod-like hollow fibre; (c) a membrane bag; (d) dynamic MMLLE. (Reprinted from [122] with permission from Elsevier, Copyright 2008)

hollow fibres or membrane bags, both normally disposable. The acceptor volume, i.e. the volume of the organic solvent, is from 5 to 25 ml, while sample volumes range from 0.5 to 4 ml. In dynamic MMLLE, the membrane unit is either a planar membrane or a hollow-fibre membrane. In the planar configuration, the membrane is clamped between two blocks and separates two flow channels: the donor and acceptor channels. In the hollow-fibre membrane module, the acceptor phase flows inside and the donor phase outside the membrane. Membrane-assisted solvent extraction offers a means to combine conventional LLE online with GC. The interface between the extraction unit and the GC is typically an on-column interface or a PTV. From an instrumental point of view, online MMLLE–GC is straightforward. The MMLLE is virtually a microsystem for continuous LLE and is used for the extraction of nonionizable compounds with nonpolar organic solvents. Also, the solvent volume is typically small and well suited for online transfer. Miniaturized membrane extraction systems have also been reported [127], where the extraction is performed in a system installed directly to a GC injector (Figure 3.15). The interfacing of membrane extraction and GC can be done directly or via a sample loop located in a multiport valve. The main disadvantage of all membrane extraction techniques is that the extraction tends to be non-exhaustive. The recoveries are usually at a similar level as those in SPME, and calibration must be carefully carried out. Quantitative recoveries are obtained

Automation and Miniaturization of Sample Treatment

117

Figure 3.15 Miniaturized online MMLLE–GC system, installed to a GC injector port. The parts numbered are: 1, sample pump; 2, sample pipette; 3, sample tray; 4, pipette needle port; 5, extraction card holder, called a card guard (CG); 6, waste valve; 7, waste bottle; 8, solvent pump; 9, 4 ml solvent vial; 10, GC needle; 11, solvent cup; 12, washing fluid

for highly hydrophobic analytes, and also when either the flow rate of the donor is kept sufficiently low or the sample is circulated across the extraction system. If quantitative recoveries are not required, the extraction time can be decreased by increasing the flow rate of the sample, or the sensitivity can be increased by using an excessive amount of sample. Selection of the solvent is critical, but the literature data on conventional LLE can be utilized in the selection. The solvent should be immiscible with water and strongly immobilized in the pores of the membrane. It should also be sufficiently volatile, though too volatile a solvent (e.g. pentane) can cause problems in the extraction. Selective extraction can be achieved by a careful choice of the material and the pore size of the membrane, as well as the extraction solvent. Pore size is important because in addition to the LLE process, size exclusion takes place in the extraction, increasing the selectivity. 3.2.4

Comparison of Solid and Liquid Phase Microextraction Techniques

Choosing a suitable extraction technique requires consideration of a range of factors, including the efficiency of the extraction, the sample throughput time, the

118

Miniaturization of Analytical Systems

complexity (cost) of equipment, the complexity of method development, the amount of organic solvent used and the range of applicability. SPME and SBSE instrumentation is commercially available, and SPME in particular has been widely applied to the analysis of several types of volatile compound. SPME instrumentation is simpler, as the injection in SBSE requires a special interface for thermal desorption. However, the extraction efficiency is clearly better in SBSE than in SPME, and SBSE is thus more suitable for trace analysis. Both techniques are solvent-free and easy to use, and the enrichment factors are typically high. In addition, SPME and SBSE devices are easily stored and transported, and they can be applied even in onsite sampling. SPME has also been used for in vivo sampling. Field sampling is effective because only the fibre or stir bar with the absorbed analytes needs to be brought back to the laboratory. Transportation of large sample volumes is avoided, and no sampling accessories – such as pumps or filters (as found in onsite SPE) – are required. SBSE is the more rugged system for onsite sampling because SPME fibres are quite fragile. In addition, if trace amounts are to be extracted, SBSE gives higher recoveries and thus better sensitivities. Both techniques are very useful for samples where the sample volume is limited, as for example in the determination of the composition of pore water in marine sediments. Although derivatization can improve the extraction of polar analytes, both SPME and SBSE are best suited to the extraction of relatively nonpolar and reasonably volatile analytes. Online SPE–GC and MEPS and pipette-tip SPE are adsorptive extraction techniques and they give quantitative recoveries, unlike the abovementioned sorptive extraction techniques, and are therefore well suited to trace analysis. Awide range of SPE materials are available, and not only nonpolar but also polar analytes can be extracted efficiently. These SPE techniques tend to be less selective, and matrix components are often extracted as well. On the other hand, the rather low selectivity is an advantage for a profiling type of analysis, i.e. when all or most of the sample components are of interest. SPE–GC can be performed in a fully automated and closed system, where the whole extract is transferred for the GC analysis. This improves the sensitivity and reliability of the analysis, as problems with sample loss and contamination are minimized. Although use of the instrumentation is fairly straightforward, the system is complex, particularly in the method development. Pipette-tip SPE is simple to use, but as an offline technique is performed manually and is user-dependent and less repeatable than the online system. Also, because only part of the extract is analysed, the sensitivity is not as good as in the online system. The optimization is easy and flexible, only small volumes of solvents are needed, and as a new pipette tip can be used for each extraction, there are no problems with memory effects. MEPS can be employed as an automated online technique, but the lack of a drying step makes the interfacing with GC demanding. Unlike the other techniques described here, LPME does not require any special instrumentation. It is not particularly rugged, however, and the repeatability is very much dependent on the user. Although it is often said that

Automation and Miniaturization of Sample Treatment

119

LPME has been developed as an alternative to SPME, the applications are not quite the same. In LPME, the solvent must be relatively nonvolatile, and thus highly volatile compounds cannot be analysed because of coelution with the solvent peak in GC. In contrast, (relatively) nonvolatile analytes can be extracted and desorbed without problem, unlike in SPME and SBSE. Thus, this technique is best suited for semivolatile to (relatively) nonvolatile analytes. The membrane-assisted extraction techniques, which apply the same extraction mechanism as LPME, are likewise best suited for semivolatile to (relatively) nonvolatile analytes. The choice of solvent is less critical than in LPME. The repeatability is typically better than in LPME because there are no problems with drop stability as there are in LPME, and larger volumes can be extracted. Static membrane-assisted extraction, with either hollow fibres or membrane bags, is inexpensive and simple to use, and since the membranes are typically for single use, there are no problems with cross-contamination. The flat-sheet modules are best suited to dynamic extraction of larger sample volumes and to online connection with GC. In the flat-sheet modules the membranes are normally used for several extractions, and careful cleaning between extractions is required to minimize the risk of cross-contamination. The main advantage of the miniaturized SPE methods over SPME, SBSE and LPME is that because the extraction is typically quantitative, the calibration is straightforward. This is in contrast with the other techniques, where various calibration methods are employed, including classical calibration relying on equilibrium extraction, or more novel kinetic calibration.

3.3

Simplification of Sample Treatment: Continuous Flow Systems

Flow systems provide an elegant, effective way of automating sample treatment. This section is concerned with how sample processing can be improved by using a flow system connected to a discrete sample introduction system. As stated in other chapters, flow systems are especially relevant to sample treatment and automated method calibration. Online combinations of automated sample pretreatment units and discrete sample introduction devices are highly attractive in this context as they allow the whole analytical sequence to be developed in a single instrumental assembly while avoiding the typical problems associated with conventional sample treatment approaches. Simplification also influences the quality of the information delivered by analytical laboratories. Flow analysis systems can help consolidate this as a trend. In fact, the intrinsic advantages of flow systems, which typically include low costs, high flexibility and throughput, and the ability to easily minimize errors, make them especially suitable for developing fast-response analytical systems. A flow analyser provides binary information that can be used for sample classification and qualification in accordance with preset threshold values imposed by either clients or

120

Miniaturization of Analytical Systems

legislation [128]. To this end, the flow system can be online coupled to a detector (vanguard configuration) or a high-performance instrument (vanguard/rearguard configuration), the flow-processing unit being used to treat samples [28]. Vanguard/ rearguard configurations are the more flexible inasmuch as the sample treatment unit and the high-performance instrument can either share the same detector or use different ones, which facilitates alteration of the system as required to perform additional or alternative confirmation measurements. Depending on the degree of human participation and the particular hardware used, samples can be treated for analysis by using batch procedures, integrated devices or coupled devices. In the latter case, devices can be coupled at-line, online or in-line. At-line coupling involves the use of a robotic interface to link the flow system to the measuring instrument. Usually, the robotic interface operates either by placing a fraction of treated or conditioned sample into an autosampler vial or by inserting the sample directly into the injector of the analytical instrument. Unlike at-line coupling, online coupling involves direct contact between a flowing stream coming from the flow system, a valve (or an appropriate interface) and a part of the instrument, usually the injection region. The flow system and the instrument are usually linked via a flow element such as a valve, T-connector or split-flow interface. Finally, in-line coupling involves complete, close integration of the flow system and the instrument. This is accomplished by incorporating a critical element of the flow system (e.g. an extraction unit) into the instrument. Manufacturers have favoured at-line coupling in commercial instruments. One typical example is the device used to couple SPME in fibres with liquid or gas chromatographs. Because the commercial instrument is of the closed type, the analyst has no access to specific parts, which restricts the potential for linking to other devices. This is probably why flow systems and instruments are most often coupled by inserting a fraction of sample into the injector of the measuring instrument. 3.3.1

Coupling to Gas Chromatography

The main difficulty in coupling flow systems to gas chromatographs arises from the different aggregation states of the fluids involved and the relatively low sample volume that needs to be inserted (a few microlitres) in order to maintain chromatographic resolution [129]. This problem can be avoided by using the LVI technique, which affords injections of hundreds of microlitres with a negligible effect on resolution. This section deals with the online coupling of flow configurations of the low- and high-pressure type to gas chromatographs. Most existing configurations have evolved from previous offline arrangements, which can be deemed at-line systems when an autosampler is used to inject samples.

Automation and Miniaturization of Sample Treatment

121

Properly coupling a flow-processing device to a gas chromatograph entails consideration of the following: (i) the volatility and thermal stability of the target analytes; (ii) the compatibility of the medium to be used with liquid samples; (iii) the extract volume, which should be as low as possible in order to minimize analyte dilution and avoid unduly decreasing the sensitivity as a result; and (iv) the transfer of the analytes from the automated pretreatment module to the gas chromatograph, which should be quantitative. A number of interfaces have been used for this purpose, the three most popular being based on on-column injection, solvent vaporization and multiport valves. In the on-column interface, a continuous high-pressure flow system based mainly on SPE is coupled to a gas chromatograph via a six-port injection valve controlled by the GC software; the valve transfers the analyte, in an appropriate medium, via a metal or fused-silica capillary typically 20–30 cm long · 0.1 mm ID that is inserted into the septum of the on-column injector. Because this type of system is usually employed with the LVI technique to inject large volumes of solvent (typically 100 ml), the on-column injector must be connected to a retention gap (a 3–5 m long · 0.53 mm ID deactivated fused silica capillary) followed by a 1–2 mm · 0.25 mm ID retaining precolumn and finally the analytical column. Usually, the precolumn and the analytical column contain the same stationary phase. Also, a solvent vapour exit (SVE) kit is installed between the two columns in order to remove most of the injected solvent and minimize the loss of the most volatile compounds. Alternatively, a programmable temperature vaporizer (PTV) can be used instead of the on-column injector, the SVE being kept even though it is unnecessary as excess organic solvent can be removed in the split valve. When a multiport injection valve is used, the organic phase that leaves the flow-processing device, which contains the highest analyte concentrations, is used to fill the loop of a high-pressure injection valve (typically 5 ml volume). A desiccating column is placed in front of the valve in order to prevent water traces from reaching the chromatographic column. The carrier gas of the chromatograph is also used to transfer the analytes from the loop to the injector via a PTFE or stainless steel tube furnished with an injection needle, which is inserted into the septum of the split/splitless injector. Depending on the particular instrument and injector configuration, the carrier gas should either be split (between the injection valve and the chromatograph) or not in order to maintain peak resolution. In some cases, additional heating of the interface is required. The following section describes the uses of nonchromatographic separation techniques for sample treatment in flow systems and discusses the most salient advantages and disadvantages of the previous interfaces. The main purpose of coupling a flow-processing device online to a gas chromatograph is usually to facilitate the development of a fully automated sample pretreatment process. Obviously, a continuous nonchromatographic separation technique will often be required for this purpose. SPE is the most common choice. This is hardly surprising since SPE is compatible with gas chromatography, because

122

Miniaturization of Analytical Systems

the analytes, which are isolated on a packed sorbent material, are eluted with a few microlitres of an organic solvent that can be directly transferred to the instrument for separation prior to detection. Water traces can be removed from the sorbent by drying prior to elution. This online trace enrichment operation can be performed using various flow configurations. The earliest approaches used a low-pressure continuous flow system in which the sorbent was packed in a laboratory-made miniaturized column accommodated in the loop of an injection valve, the sample and reagents (conditioning solvents, eluent and derivatizing reagents, if needed) being directly aspirated into the manifold. A wide variety of applications have been developed following this basic configuration, including the determination of pollutants, vitamins, fatty acids and drugs in samples of environmental, clinical, toxicological and agrifood interest [129]. The flow manifold can also be constructed by using a combination of highpressure injection valves and pumps in order to deliver the sample and organic solvents required to condition the columns, along with a syringe pump for the eluent. The sorbent column is hand-packed (usually with a polymer) and placed between the high-pressure injection valves for the sample and eluent. Typical determinations with these systems include pesticides [130] and endocrine disruptors [131] in water. The extraction selectivity can be increased by using an immunosorbent material, such as that employed in the determination of s-trizine in aqueous samples, with the Prospekt (Programmable Online Solid Phase Extraction Technique) system (commercially available from Spark Holland) for sample preparation, a Midas autosampler (also from Spark Holland) and an HPLC pump for analyte elution. A PTV interface can be used in combination with the LVI technique for the determination of pesticides in water, basically [132]. Dynamic microwave-assisted extraction [133] and dynamic sonication-assisted solvent extraction [134] can be coupled online prior to SPE– LVI–GC in order to determine organophosphate esters in air samples. Analytical pervaporation has proved a reliable alternative to headspace techniques for the direct analysis of volatile fractions when coupled online to gas chromatography [135]. Usually, the pervaporation module consists of a lower unit for sample introduction and an upper part insulated by a membrane permeable to the target analytes. A spacer is usually needed to facilitate formation of the headspace. For coupling, the upper chamber is placed in the loop of an injection valve in order to allow the acceptor gas (namely the chromatographic carrier gas) to pass through and drive the analytes to the injector. LLE can also be implemented in a continuous-flow manifold. However, it is not the first choice as it provides poorer enrichment factors due to the need to maintain the organic-to-aqueous phase ratio as close to unity as possible in order to ensure efficient phase separation, and also because of its lack of robustness, the weakest element in the system being the phase separator (a T-piece, membrane or sandwich connector). In any case, online coupled LLE has been used with or without

Automation and Miniaturization of Sample Treatment

123

derivatization to determine phenols, fatty acids, pesticides and cholesterol in waters, fats and dairy products. Emerging nonchromatographic separation techniques have also been online coupled to gas chromatography with a view to circumventing the shortcomings of some well-established extraction techniques. Such is the case with MMLLE, which has some advantages over LLE such as the formation of zero emulsions, the use of smaller sample volumes and the obtainment of clean extracts. Also, selectivity can be increased by choosing an appropriate membrane material and pore size. Reported applications use planar flat-sheet polypropylene membranes sandwiched between two blocks 10–100 ml in volume. The extraction unit and the GC instrument are interfaced via a multiport valve accommodating the sample loop. The membrane should be replaced every 50–100 uses, depending on the particular application, in order to minimize potential adsorption problems. Configurations typically use LVI and SVE. These systems have been employed to determine organochlorine pesticides in water [136] and red wine [137]. Hollow-fibre membranes, which provide increased extraction surfaces – and improved extraction efficiency as a result – have been used for the extraction of polycyclic aromatic hydrocarbons and pesticides from soils and grapes, respectively, following online pressurized hot-water extraction. The water phase, containing the analytes, was driven to the donor side of the membrane and the analytes were extracted to the acceptor solution, the concentrated extract then being online transferred to the GC instrument via an on-column interface. Flow-processing devices for coupling to gas chromatographs use a low- or highpressure mode in combination with split/splitless injection or LVI. In general, they afford the introduction of untreated samples, which simplifies the analytical process and results in improved productivity-related analytical properties. The most common nonchromatographic techniques can easily be coupled to GC via an appropriate interface; alternatively, a tandem flow-system interface can be used for the online coupling of a previous auxiliary energy-based extraction system. 3.3.2

Coupling to Liquid Chromatography

The combined use of flow-processing devices and liquid chromatographs has proven a powerful tool for solving a number of problems in analytical science. The nature of the two partners facilitates their coupling, which produces a synergistic effect that enhances the analytical properties of the ensuing methods. The similarities between the two partners are apparent from their associated equipment. Flow-processing devices and liquid chromatographs can be coupled into two main types of configuration, namely: (i) Precolumn arrangements, where the flow system is used mainly to improve sensitivity and selectivity by preconcentration or cleanup, but can also be

124

Miniaturization of Analytical Systems

employed for other purposes such as saving reagents or introducing problematic or hazardous samples. (ii) Post-column arrangements, which are intended to facilitate detection of the target analytes by derivatization in the flow system following separation on the chromatographic column. This section focuses on precolumn arrangements, where the flow device is used to introduce treated sample aliquots into the chromatographic system. The highpressure injection valve is the central element of the combined system as it acts as the interface between the low-pressure (flow) and high-pressure (chromatographic) lines. The autoinjector of a liquid chromatograph constitutes the simplest example of a precolumn arrangement (in the at-line mode); an automatic syringe similar to those employed in sequential-injection (SI) analysis is used to fill the sample loop of the injection valve. In practice, coupled flow-processing/liquid chromatography systems depart markedly from the above-described example. In fact, the flow-processing device is used not only to introduce samples into the chromatographic system but also for sample pretreatment. The pretreatment, which is mainly intended to adapt the sample to the chromatographic separation conditions, involves one or more operations including a nonchromatographic procedure. Treated sample aliquots are then driven to the high-pressure injection valve, which is actuated to introduce the plug into the column (see Figure 3.16). Reproducible injection into the chromatograph relies on accurate timing and synchronized operation of the whole system. A wide variety of sample-treatment protocols can be used to increase the sensitivity and selectivity of determinations. Using an SPE protocol for analyte preconcentration and cleanup allows a different type of coupling to be implemented (see Figure 3.17). A precolumn packed with sorbent material is accommodated in the loop of the high-pressure injection valve to retain the target analytes. The flow system allows the initial steps of the SPE protocol (conditioning, equilibration, sample loading and sorbent cleanup) to be developed in an automatic manner. Elution is carried out by switching the valve to its injection position. Fully automated SPE-based configurations such as the Prospekt [138] or Aspec [139] can also be online coupled to a liquid chromatograph by using a high-pressure injection valve as an interface between the two.

HPP IV FPD

Chromatographic column

D

waste

Figure 3.16 General flow device–liquid chromatographic coupling. FDP, flow-processing device; HPP, high-pressure pump; IV, high-pressure injection valve; D, detector

Automation and Miniaturization of Sample Treatment

125

HPP IV IV

Chromatographic column

FPD FPD

D D

waste SPE column

Figure 3.17 Continuous SPE coupled to liquid chromatography. FDP, flow-processing device; HPP, high-pressure pump; IV, high-pressure injection valve; D, detector

Other coupled systems use two or three injection valves. Thus, a configuration with two valves was used to implement sample screening and confirmation methods in the same manifold; this operational mode has been referred to as the dual use of flow configurations in liquid chromatography. To this end, an SPE column is placed in the first valve and the chromatographic column in the second, the two being online connected as shown in the general scheme of Figure 3.18. The second valve is essential for switching between the screening (load position) and confirmation (inject position) modes. Essentially, the flow-injection (FI) system is first employed to screen all samples for the target analytes or analyte families, and then those with total concentrations close to or above the preset cutoff level are individually analysed by liquid chromatography. In the confirmation step, the FI system can be used either to preconcentrate samples or as a post-column derivatization system. Before analytical injection techniques were developed, continuous segmented flow configurations were used in a precolumn combination with liquid chromatography. This allowed samples to be processed in an automatic manner. Such a combination can thus be considered a precursor to current flow approaches based on sample injection. FI techniques evolved from a straightforward operational basis, used inexpensive hardware, were easy and convenient to operate and highly flexible, and provided a high sample throughput and cost-effective performance. Sequential-injection (SI) techniques evolved from FI and shared some of their basic features. Both FI and SI manifolds have been widely used to design flow-processing devices coupled online to liquid chromatographs. Both techniques can be used in combination to assemble complex manifolds for specific applications.

Chromatographic column

HPP

IV

IV FPD

D

waste SPE column

Figure 3.18 Flow-processing device coupled online to liquid chromatograph involving both a screening and a confirmation method. FDP, flow-processing device; HPP, high-pressure pump; IV, high-pressure injection valve; D, detector

126

Miniaturization of Analytical Systems Sample Waste D Mobile phase

CC HC

SV

SP

Figure 3.19 Sequential-injection chromatography. SP, syringe pump; HC, holding coil; SV, selection valve; CC, chromatographic column; D, detector

The recently developed technique of sequential-injection chromatography (SIC) [140] testifies to the highly symbiotic nature of flow manifolds and liquid chromatographs. The SIC technique (see Figure 3.19) uses an SI manifold to introduce samples and deliver the mobile phase, in addition to a monolithic chromatographic column to separate analytes at a low pressure. This combination has been used to determine naphazoline nitrate and methylparaben [141], and lidocaine and prilocaine [142], in pharmaceuticals. The flexibility of FI and SI techniques has been exploited to efficiently pretreat samples for chromatographic separation, implementing several nonchromatographic techniques. Membrane-based separation techniques include dialysis and gas diffusion. These have been used for the determination of anions, and of ammonia and methylamines in natural water, respectively. A typical sandwich diffusion cell is placed in the low-pressure line and the acceptor stream is driven to the loop of the high-pressure injection valve. ASTED systems (automated sequential trace enrichment of dialysate), which perform dialysis and preconcentration, have been successfully used to determine tetracycline in eggs, and ioxidanol in human, rat and monkey plasma. Online membrane extraction, both alone and in combination with pervaporation, has been used to preconcentrate analytes prior to HPLC processing. In recent years, microdialysis coupled online to HPLC has proven effective for the analysis of some species in complex matrices, showing operational ease, expeditious isolation of components from the sample matrix, potential enrichment and the use of little or no organic solvent. The determination of sulphonamides in milk and of organic acids in milk fermentation products constitute two salient examples [143]. The SPE/HPLC couple is a consolidated choice that has been applied to the determination of pesticides in water, lanthanides in rocks, and caffeine and selected anilines and phenol compounds in aquatic systems. This operational approach is in continuous development, particularly as regards sorbent materials and protocols.

Automation and Miniaturization of Sample Treatment

127

Some authors have used molecular imprinted polymers (MIPs) [144] and carbon nanotubes [145] as sorbents for selective analyte retention. Similarly, new protocols such as renewable SPE are being automated and combined with HPLC. The synergetic hyphenation between flow-processing devices and liquid chromatography has been extensively exploited in analytical sciences, since it is useful and easy to implement, taking into account the nature of the partners. Flowprocessing devices provide an evident selectivity and sensitivity improvement to the developed methods. Several nonchromatographic techniques have been automated and coupled to liquid chromatography, proving the evident versatility of the described arrangement. 3.3.3

Coupling to Capillary Electrophoresis (CE)

CE is a highly efficient, flexible separation technique that has become a serious competitor to other separation techniques, including chromatographies. However, sample requirements are more stringent in CE than in other separation techniques. Appropriate sample treatment is a critical step toward obtaining accurate, reproducible results; this entails avoiding clogging of the capillary and adsorption of macromolecules on its walls, among other things. Electro-osmotic flow is one of the driving forces in CE. This phenomenon arises from the presence of surface charges on capillary walls. The result is a net flow of buffer solution in the direction of the negative electrode. Electro-osmotic flow is quite robust and occurs at a rate of 0.5–4 nl/s depending on the buffer pH. One other important factor is the small inner diameter of the capillary, which affords the use of very low volumes of samples and reagents, and also effective miniaturization. The ability to couple flow-processing devices to CE equipment is limited by the following factors, all of which warrant careful consideration: (i) Compatibility of hydrodynamic flow in the processing device with electroosmotic flow in the capillary. (ii) Compatibility of the high flow rates typically used in processing devices with the low rates of electro-osmotic flow in a CE system. (iii) Compatibility of the sample plug coming from the processing device with the small sample volume to be introduced in the CE capillary. (iv) Compatibility of the sample composition with the electrophoretic system. (v) Decoupling of the high voltages and currents applied to the electrophoretic separation system and the flow-processing device. The following sections describe selected coupled systems and their interfaces, and also some of their more salient uses. Flow-processing devices have been coupled in-, at- and online to CE equipment [146]. Overall, online coupled systems are the most commonplace. Interested readers can find a review of sample treatment devices used in combination with commercially available CE equipment in [147].

128

Miniaturization of Analytical Systems

Coupling a flow-processing device at-line with CE equipment entails the use of a robotic arm interface to place the sample coming from the former in an empty vial in the latter [147]. This combination is subject to several restrictions, the most significant of which are the accessibility of the CE autosampler and the minimum volume that can be delivered to the vial. Also, it requires the use of an electronic interface and appropriate control software in order to synchronize the movements of the robotic interface and autosampler. As can be seen in Figure 3.20, the needles of the programmable robotic arm can be positioned in two ways with respect to the autosampler vials. While the flow-processing device is working, the needles are down and the sample is prepared and transferred to the CE vial. When the flowprocessing device stops, the needles are lifted and CE analysis is started by moving the previously filled sample vial to the position where the capillary end and electrode are located. As shown in Figure 3.20, the programmable arm can be fitted with one or two needles. In the latter case, the needles can be identical or different in length, the longer one being used to fill the autosampler vials and the shorter one to drain them in order to maintain a preset liquid level. When a single needle is used, a constant, preset volume is delivered to each vial (e.g. by using air as carrier). As an alternative to a programmable robotic arm, processing devices and CE equipment can be coupled at-line via the replenishment system, which is used to empty vials and fill them with fresh buffer in some commercial instruments [148]. This entails disconnecting the replenishment needles from the Teflon tube coming from the replenishment bottles and reconnecting them to one coming from the flowprocessing device [147]. The most salient advantage of at-line coupled systems is that the CE equipment is run in its normal operation mode. This allows samples to be introduced hydrodynamically or electrokinetically into the capillary. It also facilitates other CE operations such as capillary conditioning. Coupling a flow-processing device online to CE involves inserting the capillary end of the latter in a continuous stream of the former. Because the flow-processing device and the electrophoretic system operate at different flow rates, the two require a split-flow interface for coupling. The split-flow interface was developed simultaneously in a vertical configuration by Fang’s group [149] and in a horizontal configuration by Karlberg’s [150]. Figure 3.21 compares the two types of interface, which possess a low dead volume and are electrically grounded. Samples and electrolyte solutions are introduced in the electrophoretic capillary by the effect of electroosmotic flow and electrophoretic mobility. Hydrodynamic flow should be avoided by ensuring that the liquid level at the interface (the capillary inlet) coincides with that in the capillary outlet. A split-flow interface can be readily constructed from dielectric materials such as Teflon, methacrylate or Plexiglas. Alternatively, the electrophoretic capillary can be inserted into a piece of Tygon tubing and attached to it with glue.

ROBOTIC INTERFACE

Waste

1

INTERFACE WITH TWO NEEDLES

Treated sample

Pump

ROBOTIC INTERFACE

CE EQUIPMENT

CE autosampler

Liquid level

2

Capillary

CE autosampler

(a)

FLOW SYSTEM

AIR

Treated sample

FlowFlow processing device device

SAMPLE

ROBOTIC INTERFACE

• CE system - ON

Needles up : • Flow system – OFF

2

• CE system - OFF

Capillary

• Flow system – ON

Needles down :

ROBOTIC INTERFACE

Waste

1

INTERFACE WITH ONE NEEDLE

Waste

Waste

CE autosampler

Capillary

CE EQUIPMENT

CE autosampler

2

Capillary

(b)

Figure 3.20 Schematic depiction of the at-line coupling of flow-processing devices to CE equipment. Configuration of robotic interface in (1) sample-collection mode and (2) sample-injection mode for electrophoretic analysis. Description of flow system manifold for (a) robotic interfaces with two needles and (b) robotic interfaces with one needle

FLOW SYSTEM

Waste

FlowFlow processing device device

SAMPLE

Waste

1

Automation and Miniaturization of Sample Treatment 129

Flow

Flow

Capillary

(a)

Treated sample

Electrode

Waste

Waste

Capillary

(b)

Waste

Interface grounded

Pt

D

Waste

Pt

CE equipment ’s pressure system

Loop (high resistence)

Closed valve

Pressure system::

X

–

Electrolyte

Detector

Liquid level

Capillary

HV

Figure 3.21 Schematic depiction of the online coupling of flow-processing devices to CE equipment. Detail of split-flow interface (a) in vertical design and (b) in horizontal design. Description of systems for performing hydrodynamic injection mode for electrophoretic analysis

Waste

Electrode

ELECTROLYTE

Pump

Flow Flowprocessing device

SAMPLE

+

130 Miniaturization of Analytical Systems

Automation and Miniaturization of Sample Treatment

131

The flow-processing device is made compatible with the electrophoretic system by grounding the interface. This avoids voltage differences between the interface and the flow system. Electrophoretic separation is accomplished by applying a voltage difference to the capillary end, as shown in Figure 3.20. This configuration is highly recommended when an optical detector is used with the electrophoretic system, but hardly useful with other types of detector (e.g. mass spectrometers). According to Valc
arcel’s group, the split-flow interface must be subjected to a voltage difference in order to obtain one in the electrospray needle of the electrospray ionization interface of the mass spectrometer [151]. Figure 3.22 shows two configurations with mass spectrometers, one with a grounded electrospray needle and one with an electrospray chamber. The current arising from the voltage difference between the grounded flow-processing device and the split-flow interface – which is connected to a voltage source – must be interrupted in order to avoid arc discharges in valves and pumps in the flow system. Alternatively, the current can be interrupted by using a plug of a substance with a high electrical resistance, such as pure water or air. It is also advisable to insert a safety line consisting of a grounded electrode immediately in front of the pumps. In principle, the use of online coupled systems linked via a split-flow interface is limited to the electrokinetic introduction of samples. However, this injection mode

MS Capillary

ESI

SAMPLE

HV

FlowFlow processing device device

HV H.V. HV H.V.

x MS

Split-flow interface

ESI Waste

Current interruption:: • Plug of air • Plug of pure water CE EQUIPMENT

MS DETECTOR

Figure 3.22 Schematic design of the online coupling of a flow-processing device to CE–mass spectrometry equipment through a split-flow interface. This configuration requires voltage isolation through a plug of air or a large plug of pure water

132

Miniaturization of Analytical Systems

provides biased results in some cases. This has prompted the use of hydrodynamic injection instead. Thus, Pu et al. [152] used a split-flow cell affording hydrodynamic injection of samples into the CE capillary by electro-osmotic flow. The interface included a Nafion joint to connect the CE capillary to the tube of the flow system (see Figure 3.23(a)). An electronic time-relay system was used to control switching of the high voltage. Kuban et al. [150] inserted a valve at the end of a horizontal splitflow interface. Switching the valve off caused the sample to be forced into the capillary. The greatest limitations of this approach arise from the need to control the time during which the valve is on and off, and also the pressure generated in the system. Santos et al. [151] proposed controlling the pressure by using a valve loop capable of withstanding higher pressures. With commercial CE instruments, one can also connect the pressure system [147] (see Figure 3.20). It should be noted that commercial CE equipment affords injection at a low pressure (0.5 psi) with a high precision (0.01 psi). The split-flow interface can be used for the online coupling of both FI and SI systems. Following a common approach, split-flow interfaces, microsequential injection systems and integrated lab-on-a-valve systems have all been coupled to CE. SI systems have also been online coupled via a microvalve allowing the insertion of a constant volume of sample into the capillary. However, the loop volume is high relative to those typically used in CE work and must be reduced by switching the valve or flushing the sample from the interface in order to facilitate electrokinetic insertion of a portion of sample (see Figure 3.23(b)). A modified version of the split-flow interface was recently used to accommodate a typical SPME fibre precisely at the capillary end [153]. Although samples were treated on the fibre and the flow system was only used to couple SPME and CE, they could also be processed at the interface connecting the fibre to the flow system. In-line coupling a flow-processing device to CE involves either inserting the capillary end (inlet region) into the treated sample or vice versa. This can be done by using hollow fibres to perform LPME. As shown in Figure 3.24, the electrophoretic capillary can be inserted into the lumen of a hollow fibre, or the extraction unit into the capillary. In the latter case, the fibre can be fitted to the capillary by heating. Flow systems have also been coupled to microchip electrophoretic systems [154], using a capillary as the interface between an SI system and the microchip; for this purpose, one end of the capillary is attached to the microchip via a Teflon fitting. The principal shortcoming of this combination is the presence of residual hydrodynamic flow in the microchip separation channel [154]. The use of flow-processing devices constitutes one of the most reliable ways of improving performance in analytical methods through automation, miniaturization and simplification of the preliminary operations of the analytical process – which are critical in CE work. Coupling flow-processing devices to CE equipment facilitates three types of task, namely: automatic calibration, automatic screening of samples and sample preconcentration and cleanup [146].

Figure 3.23 Alternatives to automatic hydrodynamic sample introduction for electrophoretic analysis in online flow system–CE combinations: (a) use of Nafion membranes; (b) use of injection valve

Automation and Miniaturization of Sample Treatment 133

134

Miniaturization of Analytical Systems

Sample Capillary Capillary

Sample

Waste

FlowFlow processing device device

Waste

Extraction extraction unit

Waste

Extraction extraction unit

FlowFlow processing device device

Electrolyte

Waste

Electrolyte

Figure 3.24 equipment

Schematic depiction of the in-line coupling of flow-processing devices to CE

Throughput in this context is limited by electrophoretic time. In fact, the high applied voltage must be interrupted and the capillary rinsed before each new run. Kuban et al. [150], however, have demonstrated that consecutive runs are feasible. The coupled equipment described in the literature includes extraction, filtering, dialysis, membrane, gas extraction and gas diffusion units; hollow fibres; and even auxiliary devices. The principal conclusion is that choosing an appropriate flowprocessing device can facilitate the treatment and analysis of virtually any type of sample (biomedical, pharmaceutical, environmental, food). Readers interested in a more detailed description of the types of device coupled to CE so far are referred to [146,147]. One of the most critical factors in selecting a flow-processing device is the compatibility of the chemical composition of the conditioned sample with the electrophoretic analysis to be performed. A number of sample-treatment techniques (e.g. SPE, LPE) require the use of organic solvents such as methanol, direct insertion of which into an electrophoretic buffer can result in current interruptions. This problem can be avoided by changing the buffer pH, adding a modifier (e.g. a surfactant to the organic solvent) or supplying the buffer with a small amount of solvent. In this way, at-line coupled systems afford slight modification of treated or conditioned samples by collection into vials containing a modifier. Alternative approaches involve modifying samples within the flow system. Other influential factors to be considered include the sample volume and dilution by effect of diffusion phenomena occurring in the flow system. Coupling flow-processing devices to microelectrophoretic chips or electrophoretic systems is of special interest when developing in vivo determinations; this is

Automation and Miniaturization of Sample Treatment

135

facilitated by the ability of electrophoretic systems to use small volumes of sample and provide rapid analyses. For example, a microdialysis needle online coupled to CE enabled the in vivo monitoring of the evolution of a drug [155]. In fact, this approach allows the characteristic pharmacokinetic response to the drug of a simple animal to be recorded. One other useful advantage is the ability to analyse special samples such as brain tissue [156]. In general, the use of flow-processing devices improves the overall efficiency, selectivity and sensitivity of CE methods. In addition, it facilitates special determinations such as in vivo pharmacokinetic tests. No doubt, miniaturization of flow-processing devices will help further expand the potential of this combination.

References [1] M.A.Z. Arruda, Trends in Sample Preparation, Nova Science Publishers, New York, USA, 2007. [2] H. Lord, J. Pawliszyn, J. Chromatogr. A, 885 (2000) 153. [3] F. David, P. Sandra, J. Chromatogr. A, 1152 (2007) 54. [4] S.N. Zhou, W. Zhao, J. Pawliszyn, Anal. Chem., 80 (2008) 481. [5] G. Ouyang, J. Pawliszyn, Anal. Chem., 78 (2006) 5783. [6] G. Ouyang, J. Cai, X. Zhang, H. Li, J. Pawliszyn, J. Sep. Sci., 31 (2008) 1167. [7] G. Ouyang, W. Zhao, J. Pawliszyn, Anal. Chem., 77 (2005) 8122. [8] Z. Mester, R. Sturgeon, J. Pawliszyn, Spectrochim. Acta B, 56 (2001) 233. [9] T. Gorecki, X. Yu, J. Pawliszyn, Analyst, 124 (1999) 643. [10] Z. Zhang, J. Pawliszyn, Anal. Chem., 65 (1993) 1843. [11] M. Portillo, N. Prohibas, V. Salvad
o, B.M. Simonet, J. Chromatogr. A, 1103 (2006) 29. [12] C. Basheer, H.K. Lee, J. Chromatogr. A, 1047 (2004) 189. [13] Z. Zhouyao, J. Poerschmann, J. Pawliszyn, Anal. Commun., 33 (1996) 219. [14] F.M. Musteata, J. Pawliszyn, J. Pharm. Pharmaceut. Sci., 9 (2006) 231. [15] V. Larroque, V. Desauziers, P. Mocho, J. Environ. Monit., 8 (2006) 106. [16] P.A. Martos, J. Pawliszyn, Anal. Chem., 71 (1999) 1513. [17] T.E. Siebert, H.E. Smyth, D.L. Capone, C. Neuwohner, K.H. Pardon, G.K. Skouroumounis, M.J. Herderich, M.A. Sefton, A.P. Pollnitz, Anal. Bioanal.Chem., 381 (2005) 937. [18] P. Campins-Falco, J. Verdu-Andres, A. Sevillano-Cabeza, C. Molins-Legua, R. HerraezHernandez, J. Chromatogr. A, 1211 (2008) 13. [19] D. Globig, C. Weickhardt, Anal. Bioanal. Chem., 381 (2005) 656. [20] Y. Fan, Y.Q. Feng, Z.G. Shi, J.B. Wang, Anal. Chim. Acta, 543 (2005) 1. [21] H. Kataoka, H.L. Lord, S. Yamamoto, S. Narimatsu, J. Pawliszyn, J. Microcolumn. Sep., 12 (2000) 493. [22] A.R. Raghani, K.N. Schultz, J. Chromatogr. A, 995 (2003) 1. [23] J. Wu, C. Tragas, H. Lord, J. Pawliszyn, J. Chromatogr. A, 976 (2002) 357. [24] R. Eisert, J. Pawliszyn, Anal. Chem., 69 (1997) 3140. [25] D. Globig, C. Weickhardt, Anal. Bioanal. Chem., 381 (2005) 656. [26] A. Martin-Calero, J.H. Ayala, V. Gonzalez, A.M. Afonso, Anal. Bioanal. Chem., 390 (2008) 227. [27] A. Aresta, F. Palmisano, C.G. Zambonin, J. Pharm. Biol. Anal., 39 (2005) 643. [28] X. Liu, Y. Ji, Y. Zhang, H. Zhang, M. Liu, J. Chromatogr. A, 1165 (2007) 10. [29] G. Ouyang, J. Pawliszyn, Trends Anal. Chem., 25 (2006) 692. [30] Dj. Djozan, T. Baheri, R. Farshbaf, Sh. Azhari, Anal. Chim. Acta, 554 (2005) 197. [31] Z. Zhang, M.J. Yang, J. Pawliszyn, Anal. Chem., 66 (1994) 844A. [32] J.L. Liao, C.M. Zeng, S. Hjerten, J. Pawliszyn, J. Microcol. Sep., 8 (1996) 1.

136 [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75]

Miniaturization of Analytical Systems V. Kaur, J.S. Aulakh, A.K. Malik, Anal. Chim. Acta, 603 (2007) 44. A. Malik, V. Kaur, N. Verma, Talanta, 68 (2006) 842. A.M. Freitas, P. Costa, C. Parreira, L. Vilas-Boas, Chromatographia, 54 (2001) 647. S. Villanueva, A. Vegas, J.A. Fernandez-Escudero, M. Iniguez, M.L. Rodriguez-Mendez, J.A. de Saja, Sensors Actuators B, 120 (2006) 278. H.B. Wan, H. Chi, M.K. Wong, C.Y. Mok, Anal. Chim. Acta, 298 (1994) 219. Y.N. Hsieh, P.C. Huang, I.W. Sun, T.J. Whang, C.Y. Hsu, H.H. Huang, C.H. Kuei, Anal. Chim Acta, 557 (2006) 321. T. Gorecki, P. Martos, J. Pawliszyn, Anal. Chem., 70 (1998) 19. J. Wu, X. Yu, H. Lord, J. Pawliszyn, Analyst, 125 (2000) 391. A. Kloskowski, M. Pilarczyk, A. Przyjazny, J. Namiesnik, Critical Rev. Anal. Chem., 39 (2009) 43. C. Arthur, L. Killam, K. Buchholz, J. Berg, J. Pawliszyn, Anal. Chem., 64 (1992) 1960. T. Gorecki, J. Pawliszyn, Field Analyt. Chem. Technol., 1 (1997) 277. F. David, B. Tienpont, P. Sandra, LC–GC N. Am., 21 (2003) 108. M. Kawaguchi, R. Ito, K. Saito, H. Nakazawa, J. Pharm. Biomed. Anal., 40 (2006) 500. P. Popp, C. Bauer, L. Weinrich, Anal. Chim. Acta, 436 (2001) 1. C. Blasco, M. Fernandez, Y. Pico, G. Font, J. Chromatogr. A, 1030 (2004) 77. B. Tienpont, F. David, C. Bicchi, P. Sandra, J. Microcol. Sep., 12 (2000) 577. W.M. Liu, H. Wang, Y. Guan, J. Chromatogr. A, 1045 (2004) 15. W.M. Liu, Y. Hu, J. Zhao, Y. Xu, Y. Guan, J. Chromatogr. A, 1095 (2005) 1. J.P. Lambert, W.M. Mullett, E. Kwong, D. Lubda, J. Chromatogr. A, 1075 (2005) 43. C. Bicchi, C. Cordero, E. Liberto, P. Rubiolo, B. Sgorbini, F. David, P. Sandra, J. Chromatogr. A, 1094 (2005) 9. J. Pettersson, A. Kloskowski, C. Zaniol, J. Roeraade, J. Chromatogr. A, 1033 (2004) 339. B.V. Burger, B. Marx, M. le Roux, W.J.G. Burger, J. Chromatogr. A, 1121 (2006) 259. W. Zhao, G. Ouyang, M. Alaee, J. Pawliszyn, J. Chromatogr. A, 1124 (2006) 112. M. Kawaguchi, R. Ito, K. Saito, H. Nakazawa, J. Pharm. Biomed. Anal., 40 (2006) 500. D. Benanou, F. Acobas, M.R. de Roubin, F. David, P. Sandra, Anal. Bioanal. Chem., 376 (2003) 69. A. Paschke, K. Schwab, J. Brummer, G. Sch€uurmann, H. Paschke, P. Popp, J. Chromatogr. A, 1124 (2006) 187. E. Baltussen, P. Sandra, F. David, C.A. Cramers, J. Microcol. Sep., 11 (1999) 737. B. Kolahgar, A. Hoffmann, A.C. Heiden, J. Chromatogr. A, 963 (2002) 225. V.M. Leon, B. Alvarez, M.A. Cobollo, S. Munoz, I. Valor, J. Chromatogr. A, 999 (2003) 91. P. Serodio, J.M.F. Nogueira, Anal. Chim. Acta, 517 (2004) 21. P. Popp, P. Keil, L. Montero, M. R€uckert, J. Chromatogr. A, 1071 (2005) 155. V.M. Leon, J. Llorca-Porcel, B. Alvarez, M.A. Cobollo, S. Munoz, I. Valor, Anal. Chim. Acta, 558 (2006) 261. S. Nakamura, N. Nakamura, S. Ito, J. Sep. Sci., 24 (2001) 674. N. Ochiai, K. Sasamoto, M. Takino, S. Yamashita, S. Daishima, A.C. Heiden, A. Hoffmann, Analyst, 126 (2001) 1652. D. Benanou, F. Acobas, M.R. de Roubin, Water Sci. Technol., 49 (2004) 161. D. Benanou, F. Acobas, M.R. de Roubin, F. David, P. Sandra, Anal. Bioanal. Chem., 376 (2003) 69. M. Kawaguchi, K. Inoue, M. Yoshimura, R. Ito, N. Sakui, H. Nakazawa, Anal. Chim. Acta, 505 (2004) 217. L. Montero, S. Conradi, H. Weiss, P. Popp, J. Chromatogr. A, 1071 (2005) 163. M. Kawaguchi, Y. Ishii, N. Sakui, N. Okanouchi, R. Ito, K. Saito, H. Nakazawa, Anal. Chim. Acta, 533 (2005) 57. N. Itoh, H. Tao, T. Ibusuki, Anal. Chim. Acta, 535 (2005) 243. R. Rodil, P. Popp, J. Chromatogr. A, 1124 (2006) 82. P.A. Martos, In: J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction, Royal Society of Chemistry, Cambridge, 1999, pp. 159–168. M. Kreck, S. P€uschel, M. W€uust, A. Mosandl, J. Agric. Food Chem., 51 (2003) 463.

Automation and Miniaturization of Sample Treatment

137

[76] D. Burkhardt, A. Mosandl, J. Agric. Food Chem., 51 (2003) 7391. [77] M. Salinas, A. Zalacain, F. Pardo, J. Agric. Food Chem., 52 (2004) 4821. [78] C. Bicchi, C. Cordero, E. Liberto, P. Rubiolo, B. Sgordini, P. Sandra, J. Chromatogr. A, 1071 (2005) 111. [79] D. Hampel, A. Mosandl, M. W€ust, Phytochemistry, 66 (2005) 305. [80] N. Scascighini, L. Mattiacci, M.D. D’Alessandro, A. Hern, A.S. Rott, S. Dorn, Chemoecology, 15 (2005) 97. [81] J. Weidenhamer, J. Chem. Ecol., 31 (2005) 221. [82] M. Kreck, A. Scharrer, S. Bilke, A. Mosandl, Eur. Food Res. Technol., 213 (2001) 389. [83] F. Luan, A. Mosandl, M. Gubesch, M. W€ust, J. Chromatogr. A, 1112 (2006) 369. [84] C.H. Wijaya, D. Ulrich, R. Lestari, K. Schippel, G. Ebert, J. Agric. Food Chem., 53 (2005) 1637. [85] S. Sewenig, D. Bullinger, U. Hener, A. Mosandl, J. Agric. Food Chem., 53 (2005) 838. [86] C. Bicchi, C. Iori, P. Rubiolo, P. Sandra, J. Agric. Food Chem., 50 (2002) 449. [87] T. Kishimoto, A. Wanikawa, N. Kagami, K. Kawatsura, J. Agric. Food Chem., 53 (2005) 4701. [88] J. Diez, C. Dominguez, D.A. Guillen, R. Veas, C.G. Barroso, J. Chromatogr. A, 1025 (2004) 263. [89] A. Buettner, J. Agric. Food Chem., 52 (2004) 2339. [90] J. Marin, A. Zalacain, C. De Miguel, G.L. Alonso, M.R. Salinas, J. Chromatogr. A, 1098 (2005) 1. [91] R.F. Alves, A.M.D. Nascimento, J.M.F. Nogueira, Anal. Chim. Acta, 546 (2005) 11. [92] J.C.R. Demyttenaere, J.I.S. Martinez, R. Verhe, P. Sandra, N. De Kimpe, J. Chromatogr. A, 985 (2003) 221. [93] E.D. Guerrero, R.N. Marin, R.C. Mejias, C.G. Barroso, J. Chromatogr. A, 1104 (2006) 47. [94] J.C.R. Demyttenaere, R.M. Morina, P. Sandra, J. Chromatogr. A, 985 (2003) 127. [95] C.Y. Lu, Z.G. Hao, R. Payne, C.-T. Ho, J. Agric. Food Chem., 53 (2005) 6443. [96] C. Bicchi, C. Cordero, E. Liberto, P. Rubiolo, B. Sgordini, P. Sandra, J. Chromatogr. A, 1071 (2005) 111. [97] A.G.J. Tredoux, H.H. Lauer, T. Heideman, P. Sandra, J. High Resol. Chromatogr., 23 (2000) 644. [98] A. De Villiers, G. Vanhoenacker, F. Lynen, P. Sandra, Electrophoresis, 25 (2004) 664. [99] A. Miki, A. Isogai, H. Utsunomiya, H. Iwata, J. Biosci. Bioeng., 100 (2005) 178. [100] C. Lorenzo, A. Zalacain, G.L. Alonso, M. Rosario Salinas, J. Chromatogr. A, 1114 (2006) 250. [101] H.A. Soini, K.E. Bruce, D. Wiesler, F. David, P. Sandra, M.V. Novotny, J. Chem. Ecol., 31 (2005) 377. [102] H.G. Wahl, C. Peterfi, R. Werner, H.M. Liebich, Clin. Chem., 48 (2002) B9. [103] T. Benijts, J. Vercammen, R. Dams, H. Pham Tuan, W. Lambert, P. Sandra, J. Chromatogr. B, 755 (2001) 137. [104] M. Kawaguchi, K. Inoue, N. Sakui, R. Ito, S. Izumi, T. Makino, N. Okanouchi, H. Nakazawa, J. Chromatogr. B, 799 (2004) 119. [105] A. Stopforth, A. Tredoux, A. Crouch, P. van Helden, P. Sandra, J. Chromatogr. A, 1071 (2005) 135. [106] A. Stopforth, B.V. Burger, A.M. Crouch, P. Sandra, J. Chromatogr. B, 834 (2006) 134. [107] S. Sisalli, A. Adao, M. Lebel, I. Le Fur, P. Sandra, LC–GC Eur., 19 (2006) 33. [108] A. Stopforth, B.V. Burger, A.M. Crouch, P. Sandra, J. Chromatogr. B, 834 (2006) 134. [109] A. El-Beqqali, A. Kussak, M. Abdel-Rehim, J. Chromatogr. A, 1114 (2006) 234. [110] M. Abdel-Rehim, J. Chromatogr. B, 801 (2004) 317. [111] M. Abdel-Rehim, LC–GC Eur, 22 (2009) 14. [112] L.G. Blomberg, Anal. Bioanal. Chem., 393 (2009) 797. [113] Z. Altun, M. Abdel-Rehim, Anal. Chim. Acta, 630 (2008) 116. [114] P. Wynne, Y. Huang, R. Hibbert, D. DiFeo, P. Dawes, LCGC North America, 26 (2008) 16. [115] M. Abdel-Rehim, A. Andersson, A. Breitholtz-Emanuelsson, M. Sandberg-Staell, K. Brunfelter, K.J. Pettersson, C. Norsten-Hoeoeg, J. Chromatogr. Sci., 46 (2008) 518. [116] R. Said, Z. Hassan, M. Hassan, M. Abdel-Rehim, J. Liq. Chromatogr. Rel. Technol., 31 (2008) 683.

138 [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156]

Miniaturization of Analytical Systems L. Xu, C. Basheer, H.K. Lee, J. Chromatogr. A, 1152 (2007) 184. J. Romero, P. Lopez, C. Rubio, R. Batlle, C. Nerin, J. Chromatogr. A, 1166 (2007) 24. M. Saraji, A. Bidgoli, H. Akbar, J. Chromatogr. A, 1216 (2009) 1059. T. Ligor, B. Buszewski, Anal. Bioanal. Chem., 391 (2008) 2283. B. Wielgomas, W. Czarnowski, Anal. Bioanal. Chem., 390 (2008) 1933. T. Hyoetylaeinen, M.L. Riekkola, Anal. Chim. Acta, 614 (2008) 27. J.B. Quintana, T. Reemtsma, J. Chromatogr. A, 1124 (2006) 22. T. Einsle, H. Paschke, K. Bruns, S. Schrader, P. Popp, M. Moeder, J. Chromatogr. A, 1124 (2006) 196. C. Basheer, J.F. Obbard, H.K. Lee, J. Chromatogr. A, 1068 (2005) 221. C. Basheer, M. Vetrichelvan, S. Valiyaveettil, L.H. Kee, J. Chromatogr. A, 1139 (2007) 157. T. Hyoetylaeinen, K. Luethje, M. Rautiainen-Raemae, M.L. Riekkola, J. Chromatogr. A, 1056 (2004) 267. M. Valc
arcel, M. C
ardenas, M. Gallego, Trends Anal. Chem, 18 (1999) 685. M. Valc
arcel, M. Gallego, A. R
ıos, Fresenius J. Anal. Chem., 362 (1999) 58. E. Pocurrull, C. Aguilar, F. Borrull, R.M. Marce, J. Chromatogr. A, 818 (1998) 85. L. Brossa, R.M. Marce, F. Borrull, E. Pocurrull, J. Chromatogr. A, 963 (2002) 287. S.H.G. Brondi, F.C. Spoljaric, F.M. Lanas, J. Sep. Sci., 28 (2005) 2243. M. Ericsson, A. Colmsj€o, Anal. Chem., 75 (2003) 1713. C. S
anchez, M. Ericsson, H. Carlsson, A. Colmsj€ o, J. Chromatogr. A, 993 (2003) 103. D.W. Bryce, A. Izquierdo, M.D. Luque de Castro, Anal. Chem., 69 (1997) 844. K. L€uthje, T. Hy€otyl€ainen, M.L. Riekkola, Anal. Bioanal. Chem., 378 (2004) 1991. T. Hy€otyl€ainen, K. L€uthje, M. Rautiainen-R€am€a, M.L. Riekkola, J. Chromatogr. A, 1056 (2004) 267. Y.C. Barret, B. Akinanya, S.Y. Chang, O. Vesterqvist, J. Chromatogr. B, 821 (2005) 159. K. Halme, E. Lindfors, K. Peltronen, J. Chromatogr. B, 845 (2007) 74. D. Sat
ınsk
y, J. Huclov
a, P. Solich, R. Karl
ıek, J. Chromatogr. A, 1015 (2003) 239. P. Chocholous, D. Satinsky, P. Solich, Talanta, 70 (2006) 408. J. Klimundov
a, D. S atinsk
y, H. Sklen
arov
a, P. Solich, Talanta, 69 (2006) 730. M.C. Wei, C.T. Chang, J.F. Jen, Chromatographia, 54 (2001) 601. R. Koeber, C. Fleischer, F. Lanza, K.S. Boos, B. Sellergren, D. Barcelo, Anal. Chem., 73 (2001) 2437. G.Z. Fang, J.X. He, S. Wang, J. Chromatogr. A, 1127 (2006) 12. B.M. Simonet, A. R
ıos, M. Valc
arcel,In: M.L. Marina, A. R
ıos, M. Valc
arcel (Eds), Comprehensive Analytical Chemistry: Analysis and Detection by Capillary Electrophoresis, 45, Elsevier, 2005, pp. 173–223. B. Santos, B.M. Simonet, A. R
ıos, M. Valc
arcel, Trends Anal. Chem., 25 (2006) 968. B. Santos, B.M. Simonet, A. R
ıos, M. Valc
arcel, Electrophoresis, 25 (2004) 3427. Z.L. Fang, H.W. Chen, Q. Fang, Q.S. Pu, Anal. Sci., 16 (2000) 197. P. Kuban, B. Karlberg, Trends Anal. Chem., 17 (1998) 34. B. Santos, B.M. Simonet, B. Lendl, A. Rios, M. Valc
arcel, J. Chromatogr. A, 1127 (2006) 278. Q.S. Pu, Z.K. Fang, Anal. Chim. Acta, 398 (1999) 65. B. Santos, B.M. Simonet, A. Rios, M. Valc
arcel, Electrophoresis, 28 (2007) 1312. Y. Chen, W. Lu, X. Chen, Z. Hu, Electrophoresis, 28 (2007) 33. L. Wang, Z. Zhang, W. Yang, J. Pharma. Biol. Anal., 39 (2005) 399. C.M. Ciriacks, M.T. Bowser, Anal. Chem., 76 (2004) 6582.

4 Miniaturized Systems for Analytical Separations I: Systems Based on a Hydrodynamic Flow 4.1

Introduction

Analytical separation techniques play a prominent role in analytical science. Without the participation of these techniques, very few reliable analyses could be carried out. Separation techniques are involved in both the treatment of samples (use of the so-called nonchromatographic separation techniques, meaning techniques without instrumental detection) and the detection of analytes after chromatographic or electrophoretic separation. Aspects related to the automation and miniaturization of separation techniques used for sample treatment have been reviewed in Chapter 3. Chapters 4 and 5 deal with miniaturized systems for analyte separation, using either a hydrodynamic flow (Chapter 4) or an electroosmotic flow (Chapter 5). Therefore, these two chapters cover the miniaturization tendencies of both chromatographic (liquid chromatography (LC), basically) and electrophoretic separation techniques; they also connect with Chapter 8, which is devoted to the miniaturization of the entire analytical process through the mTAS approach. The miniaturization of analytical separation techniques can be reviewed in parallel with the evolution of these techniques. Figure 4.1 summarizes this evolution, from the classic modes to their current alternatives. LC and electrophoresis basically involve the movement of fluids and/or analytes for the performance of separation. Planar and column chromatography, performed at atmospheric pressure (open- and macrocolumns), have given way to more efficient approaches such as Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

140

Miniaturization of Analytical Systems ANALYTICAL SEPARATION TECHNIQUES ‘miniaturization’

Planar & Column Chromatography -Paper -Think layer -(Macro)column

Contemporary modes Present modes / tendencies Column Chromatography HPLC COLUMN HPTLC

High pressure Flow rate: ml/min Sample volume: µl

Capillary Liquid Chromatography µHPLC

nl/min nl

CAPILLARY Capillary Electrophoresis CE

Conventional Electrophoresis

HPLC chip

nano-HPLC

Flow rate: µl/min Sample volume: µl

CLC CEC MEKC CZE

Mic r Nan ofluidi c oflu idic

Classic modes

CE-Chip CE chip

High potential Sample volume: nl to fl

G GC

Micro-GC

Figure 4.1 A general view of analytical separation techniques and the tendency to miniaturization

high-performance think-layer chromatography (HPTLC) in the planar mode and high-performance liquid chromatography (HPLC) in the column mode. The improvements were based on the stationary phases, the detection system and the use of closed columns (reduced size with respect to classical LC) working at high pressure. Microcolumns were also needed in ulterior developments, and particularly for the development of new alternatives addressing the use of capillary columns. The replacement of microcolumns by capillaries has led to the development of very interesting modern alternatives in analytical separation science. Thus, capillary liquid chromatography (CLC), through so-called micro- and nano-HPLC, and capillary electrophoresis (CE), have introduced a true revolution in instrumental separation techniques. Capillaries constitute a significant size reduction, but more importantly have allowed the merging of chromatographic and electrophoretic techniques in interesting analytical modes. Modes such as micellar electrokinetic chromatography (MEKC) and capillary electrokinetic chromatography (CEC) are hybrid methodologies with electrophoretic and chromatographic foundations, providing a wide analytical potential. Thus, it is possible to cover everything from pure CLC to pure electrophoresis using the capillary zone electrophoresis (CZE) mode – a wide variety of separation alternatives with an impressive scope of application. Despite the clear improvements and possibilities these modes represent, miniaturization of the equipment as a whole is not truly evident. Miniaturization affects many of the components and devices integrated in the commercialized equipment, but the equipment itself is not considered miniaturized at all. The true pass to miniaturization is given when chromatographic or electrophoretic separations are carried out in chips. In these cases, all the principles of micro- and

Miniaturized Systems for Analytical Separations I

141

Table 4.1 Common characteristics associated with LC techniques Name of technique

Column i.d.

Flow rate

Conventional LC Microbore LC Micro-LC Capillary-LC (micro) Nano-LC

3–6 mm 1–3 mm 0.2–1.5 mm 150–500 mm 10–150 mm

0.5–2.0 ml/min 100–500 ml/min 10–100 ml/min 1–10 ml/min 10–1000 nl/min

Injection volume level ml ml ml ml Nl

nanofluidics described in Chapter 2 are involved. Thus, today HPLC chips and CE chips are truly miniaturized systems connecting to the lab-on-a-chip approach. Based on the evolution shown in Figure 4.1 toward the miniaturized hydrodynamic separation techniques, in this chapter liquid chromatographic tendencies focussed on capillary HPLC and HPLC chip are reported. The common characteristics associated with LC techniques are summarized in Table 4.1.

4.2

The Earliest Example of Miniaturization of a Gas Chromatograph and Some Other Developments

Because of its foundation and equipment requirements, gas chromatography (GC) is an exceptional case, as was noted in Chapter 1. Although the objective of this and the following chapter is to describe the miniaturization of hydrodynamic analytical systems, in order to give a complete view of miniaturized chromatographic separations, gas phase separations must first be briefly commented upon. There is not much in the literature about gas phase separation on chips, despite the fact the first working microchip-based chromatographic system was a miniaturized gas chromatograph in 1979 [1]. As J.P. Kutter said [2], this development was hardly pursued afterwards, probably because the analytical community was not yet ready to embrace this new technology. Unsatisfactory results, due to difficulties in producing homogeneous stationary phases, may have contributed to this feeling. However, miniaturized gas chromatographs are of great interest in several application fields. These microsystems could potentially be used for breath analysis, indoor air-quality monitoring and warfare agent detection, allowing the onsite and portable monitoring of gases in a fast and reliable way. Miniaturization of a GC system requires the miniaturization of the gas transportation system that is necessary for the efficient extraction and manipulation of the analytes in the gaseous samples. Due to the ineffective transportation produced by gas micropumps, micro-GC (mGC) systems have had to rely on off-chip, highvolume and high-power gas transportation systems, such as syringe pumps and large mechanical pumps. Such dependence on off-chip systems has introduced difficulties for the creation of a truly portable miniaturized GC analyser. H. Kim et al. have

142

Miniaturization of Analytical Systems

developed a micropump-driven high-speed MEMS GC system for volatile organic compound (VOC) separation and detection [3]. Micropump operation is based on the use of a four-stage micropump with integrated microvalves, working as a peristaltic micropump. Commonly, micromachined GC channels are coupled with devices for preconcentration, injection and/or detection of gas phase analytes [4–6]. But, obviously, microcolumns are the heart of the GC technique, and will be for mGC systems too. The first works were concentrated on columns arranged with a spiral geometry, which produce lower dispersion than serpentines in CE microchips [7,8]. More recently, in several published articles, MEMS have been considered as columns for mGC [9–12]. R.I. Masel et al. have compared microcolumns fabricated with different geometries and have demonstrated that, for mGC, serpentine columns give the lowest dispersion, since the small spiral columns generate on overall higher Dean vortice when integrated over the entire column [13]. On the other hand, an efficient stationary phase contained within a microfabricated channel is crucial to the development of a complete mGC. The deposition of the stationary phase into the microchannel can be a complicated task. The most common types of deposition reported for open-tubular microfabricated columns use traditional static and dynamic coatings [14,15] and vapour deposition of a polysiloxane phase [16]. Fullerenes and carbon nanotubes have been demonstrated to have very interesting properties as stationary phases in GC [17]. In particular, multiwall carbon nanotubes have been shown to perform well as a GC stationary phase, in both packed and open-tubular approaches. In fact, compared to traditional stationary phases, they have the higher surface-to-volume ratio characteristic of nanostructurated materials, as well as better thermal and mechanical stability [18,19]. Because of the high thermal stability, this material is ideal for temperatureprogrammed separation. In addition, carbon nanotube stationary phases can be deposited easily by lithography. In this way, O. Bakajin and coworkers have developed an ultrafast GC methodology based on the use of single-wall carbon nanotube stationary phases in microfabricated channels [20]. They used an integrated heater for temperature programming and a synchronized dual-valve for rapid injections. The authors see this development as a field-test sensor microsystem that can be used to continuously monitor the gas composition of atmospheric environments. The integration of all these elements allows the fabrication of mGC microsystems. Researchers from the Engineering Research Center for Wireless Integrated MicroSystems (WIMS ERC) have developed an integrated microanalytical system for complex vapour mixtures, which is a good example of a microfabricated gas chromatograph [21]. This particular mGC is illustrated in Figure 4.2. These authors developed a WIMS mGC prototype, which includes the following parts: (i) An inlet particle filter, consisting of a macroporous silicon membrane with tortuous pores, which exhibits a high filtering efficiency.

Miniaturized Systems for Analytical Separations I

143

(ii) A calibration vapour source, using a dual-layer structure whose base contains a reservoir for retaining the volatile-liquid calibrant. This microdevice produces a constant vapour generation for several days. (iii) Smart latching valves, manipulating transportation into the system. (iv) Dual-separation microcolumns, based on a convolved square–spiral geometry (3.0 m length each). Wafer-level low-pressure anodic bonding is used to seal the channel with a Pyrex cover plate, and fused-silica capillaries are epoxied into recessed side ports for fluidic connections. The compounds of the sample are separated according to their reversible partition equilibriums between the mobile phases and the thin polymeric stationary phases lining the walls of the columns. (v) An integrated chemiresistor sensor array coated with thin films of different gold-thiolate monolayer-protected nanoparticles (MPNs), whose responses vary with the nature of the analyte vapour. (vi) Commercial microspectrometers, which can be interfaced to the microcolumn. (vii) A distributed vacuum pump for micropumping. (viii) A preconcentrator.

Figure 4.2 Scheme and specific images of the analytical components of the WIMS mGC prototype. (From [21] with permission of IEEE, Copyright 2007)

144

Miniaturization of Analytical Systems

Figure 4.3 Photography of the WIMS mGC prototype (top) and separation of a set of common air contaminants (FID detection) (bottom). (From [21] with permission of IEEE, Copyright 2007)

Figure 4.3 shows a view of the WIMS mGC prototype, as well as an example of a chromatogram in which 19 air contaminants were separated in 4 minutes. Microfabricated GC has the potential to achieve superior performance to traditional GC. Other advantages lie in its capability to perform parallel analysis with low cost, low power consumption, small thermal mass (which allows fast temperature programming rates), portability and short analysis time.

4.3 Capillary Liquid Chromatography (CLC) ‘Capillary liquid chromatography’ (CLC) is a general term that includes, in many cases, mLC and nano-LC. This definition is accepted in this chapter, because the separation is carried out into capillaries in both cases. Table 4.1 reports the features of CLC modes. The high resolution intended in CLC with the downsizing of the separation element containing the stationary phase (which follows the sequence shown in Figure 4.4) has the limitation of the pressure drop across the initial packed columns. Although open-tubular capillary columns and packed microcapillary columns present a higher permeability, and have allowed the development of basic CLC separations with a higher resolution, the problem of the pressure still remains.

Miniaturized Systems for Analytical Separations I

COLUMN

CAPILLARY

145

CHIP (microchannels)

High-pressure pumps and micropumps Electrically-driven pumping and electro-osmotic flow

Conventional LC

---

CLC (micro- and nano-) - - - Electrochromatography modes -- - - Conventional CE (CZE) ----

LC chip CE chip

Figure 4.4 Key factors for downsizing LC systems and main analytical techniques

As T. Takeuchi recognized [22], one possible reason for the limited development of CLC (which began with Ishii’s group, Japan, in 1974 [23]) lies in the use of electrically-driven pumping (Figure 4.4) for separation methods such as CZE and the other electrochromatographic modes (described in Chapter 5). In fact, these electrochromatographic techniques have the potential to produce much higher theoretical plates than pressure-driven separation methods. The features of CLC are attributed to the use of smaller-diameter columns and lower eluent flow rate. The latter results in the saving of solvents, reagents and packing materials, in comparison with conventional LC. Mass sensitivity is generally improved due to the small volume detection in the capillary, which is of particular interest when only very low sample size is available (biological samples, for instance). The low flow rate in CLC is another advantage for the wider manipulation of the mobile phases, as well as for connection to detectors requiring low flow of sample. This is the case with CLC–MS coupling, presenting a higher compatibility than conventional HPLC–MS systems. Finally, the low heat capacity of the capillary columns facilitates the control of column temperature and hence more effective and easier temperature programming. All these features and advantages are summarized in Figure 4.5, adapted from reference [24]. As stated before, bioanalytical applications are one of the biggest potential uses of CLC, particularly when it is coupled to a mass spectrometry (MS) detector. Recently, J.M. Saz and M.L. Marina have reviewed the application of CLC to the determination and characterization of bioactive and biomarker peptides [25]. The combination of CLC with MS is very selective and sensitive, enabling the analysis of new (even not-yet-discovered) peptides and permitting the simultaneous analysis of a great number of peptides. Both CLC and MS are frequently interfaced with micro- or nano-electrospray ionization (ESI). Miniaturization of the ESI interface improves sample ionization efficiency in the MS system. Additionally, miniaturized ESI devices work with low flow rates of the same order of magnitude as those provided by the CLC system. Here, the use of capillary columns improves separation efficiency and, hence, increases the overall selectivity

146

Miniaturization of Analytical Systems Low consumption of mobile phase

LOW FLOW RATE

Easier manipulation of mobile phase Better compatibility with MS detectors

SMALLER COLUMN DIAMETER (CAPILLARY)

SMALLER PARTICLE SIZE (stationary phase)

Lower stationary phase, solvent and sample volume consumption

Improved mass sensitivity

LOW HEAT CAPACITY

Temperature programming

Figure 4.5 The main effects, and the corresponding advantages, produced by the reduction of the column (capillary) diameter

of the CLC–ESI–MS arrangement. In this review [25], an interesting example was selected to show the potential of CLC–ESI–MS for the analysis of succinylated or trimethylammonium (TMAB)-labelled endogenous peptides extracted from the pituitaries of mice [26]. To detect and quantitate the labelled peptides, a microHPLC (300 mm inner diameter (ID), C18 3 mm particles)–ESI–MS system was used.
To identify the peptides, a trapping column (5 mm · 300 mm ID, C18 5 mm, 100 A) to preconcentrate and desalt the sample and a nano-HPLC (150 mm · 75 mm ID, C18
5 mm, 300 A)–ESI–MS system were used. As expected, due to the larger mass difference between the heavy and light forms, peptides labelled with the TMAB reagent generally showed much better resolution in the MS than the succinylated peptides, as Figure 4.6 illustrates. In each of these CLC–ESI–MS systems it is very common to use online preconcentration steps following the scheme shown in Figure 4.7, as already described by A.J. Oosterkamp et al. [27]. Particularly at nano-level LC, the use of precolumns is highly recommended (in addition to the cleanup and preconcentration objectives), since capillaries can easily be blocked at the inlet when real samples have to be analysed. The special compatibility of CLC with particular detectors has been studied by P. Chaminade and coworkers [28]. They compare three commercial universal detectors that allow a direct detection of lipids. These detectors are: the charged aerosol detector (CAD), the evaporative light-scattering detector (ELSD) and the ion trap (IT) mass spectrometer with atmospheric pressure chemical ionization (APCI) and ESI (see above) sources. The detectors are compared in terms of response intensity, linearity and limit of detection, while working at high temperatures. CLC offers interesting possibilities for the determination of lipids, as these compounds are not very sensitive in other instrumental separation techniques.

Miniaturized Systems for Analytical Separations I

147

Figure 4.6 Mass spectra of labelled peptides obtained from CLC–ESI–MS analysis of succinylated or TMAB-labelled endogenous peptides extracted from the pituitaries of mice. Left-column spectra correspond to a phosphorylated fragment of chromagranin B. Middlecolumn spectra correspond to joining peptide-Gly-Lys-Arg. Right-column spectra correspond to VS-Gly-Lys-Arg. The upper panels correspond to succinylated peptides and the lower panels to TMAB-labelled peptides. From [26], reproduced with permission of John Wiley & Sons, Ltd

Figure 4.7 Typical setup for online SPE preconcentration CLC–ESI–MS (adapted from [27])

148

Miniaturization of Analytical Systems

ELSD is the primary detector for the analysis of lipids because of its compatibility with a large range of solvents and elution gradients. ELSD can detect all solutes that are less volatile than the mobile phase. Recently, CAD has been developed as a new detector. It is a universal hybrid detector useful for HPLC applications, initially presented by Dixon and Peterson in 2002 [29]. CAD works in two steps: the first involves nebulizing the LC column effluent and evaporating the solvents (similarly to ELSD); the second is the ionization of the aerosol particles by impact with the positively charged nitrogen cation obtained by corona discharge. The amount of ion charged is afterwards detected by an electrometer (similar to the principle of APCI in MS). It is interesting to compare the sensitivity (through the limits of detection, LOD, estimated on the basis of three times the signal-to-noise ratio; S/N ¼ 3) for representative lipid compounds using the different detectors [28]. Table 4.2 reproduces these results. Without any doubt, these less popular detectors such as CAD and ELSD open very interesting possibilities when used for CLC separations. Thus, ELSD has been coupled to CLC since 1999 [30], is already commercialized and, particularly for lipids, presents analytical advantages with respect MS detection (as Table 4.2 demonstrates). Again, biological samples not available in large amounts can be analysed by CLC–ELSD. An interesting application has been developed by L. Quinton et al. [31]. These authors propose a microanalytical system, based on CLC–ELSD, for the separation of stratum corneum ceramides. Stratum corneum lipids were obtained from the volar side of forearm skins of healthy subjects. A sample injection volume, as low as 1 ml, was used for the chromatographic analysis. Figure 4.8 shows the separation of the different lipid classes of a stratum corneum sample. The main groups of lipids and the chromatographic region for the analysis of ceramide classes can be seen. Table 4.2 LOD a for three representative lipids (in ng) obtained with different detectors coupled to CLC. (Reprinted from [28] with permission from Elsevier) Detector ELSD

CAD

ESI-MS (full scan)

API-MS (full scan)

a

Temperature (
C) 100 125 150 100 125 150 100 125 150 100 125 150

Not detected with this detector.

Cholesterol

Ceramide III B

Squalene

16.00  1.10 10.00  1.00 5.00  0.20 15.00  1.20 7.50  0.80 2.50  0.10 N/Aa N/Aa N/Aa N/Aa N/Aa N/Aa

30.00  2.12 20.00  0.87 10.00  1.00 40.00  1.90 36.00  0.30 30.00  1.69 4.00  0.12 1.50  0.06 1.00  0.01 0.03  0.01 0.04  0.02 0.05  0.01

1.20  0.08 1.00  0.01 0.60  0.02 1.00  0.10 0.70  0.02 0.18  0.01 N/Aa N/Aa N/Aa 0.50  0.10 0.80  0.40 0.84  0.30

Miniaturized Systems for Analytical Separations I 375

mV

149

Sq

300 Cho

Free fatty acids

Triglycerides

225 150

Ceramides classes

75 min

–15 0.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

Figure 4.8 Separation of lipid classes of a stratum corneum sample. Sq, squalene; Cho, cholesterol. (Reprinted from [31] with permission from Wiley VCH)

Separation capabilities are not only associated with the reduction size of the chromatographic column. The manipulation of the stationary phase can significantly improve the selectivity of many applications. Basically, three main dispositions of stationary phase in chromatographic columns can be identified for CLC: open-tubular capillary columns [32], packed microcapillary columns [33] and monolithic silica capillary columns [34]. Additionally, fused silica capillaries [35] can be used for electrochromatographic methods. Among these, columns of packed particles are still the most popular for HPLC because of their great utility, excellent performance and wide variety. Recently, Q.-S. Qu et al. have proposed the use of gold microspheres (AuMSs) modified with octadecanethiol for CLC [36]. Gold nanoparticles (AuNPs) are becoming increasingly attractive in many scientific fields because of their long-term stability, high surface area-to-volume ratio and ease of chemical modification. With this background, AuMSs have great potential to become substitutes for silica-based materials as stationary phase for CLC. This is due to their stability at high pH, reasonable rigidity and ease of manipulation from a physicochemical point of view. This pure AuMS-based C18 stationary phase may be a promising alternative to conventional C18 material for CLC. F. Fanali and coworkers have reviewed recent applications of CLC (nano-LC for these authors; they refer to the use of capillaries of 10–100 mm ID) [37]. As they report, the main areas of application of CLC are proteomic, pharmaceutical and environmental. One of the most exciting applications of CLC is in proteomics, especially in peptide mapping, protein sequencing and the study of proteins. Nanoscale proteomics studies commonly require CLC–MS technology, employing the interfaces reported above (ESI or APDC) as well as CLC–tandem mass spectrometry (CLC–MS/MS) [38]. Even the use of nano-HPLC–ICPMS for the quantification of sulfur-containing peptides has been described [39]. Figure 4.9 shows the high capacity for separating 24 different peptides using a combination of nano-HPLC–ICP IDMS with nano-HPLC–ESI MS/MS.

150

Miniaturization of Analytical Systems

Figure 4.9 (a) Sulfur mass chromatogram obtained from human serum albumin (HAS) tryptic digest by precolumn isotope dilution analysis with nano-HPLC–ICP–MS. (b) Table with the assignments of the peaks. (From [39] with permission, Copyright ACS)

4.4

Liquid Chromatography on Microchips

The miniaturization of the column inner diameter (Table 4.1 and Figure 4.5) and volumetric flow rates in LC (HPLC) is an ongoing trend that is mainly driven by the need to handle small volumes of complex sample, particularly in the context of highthroughput screening technologies. Recent progress toward HPLC in lab-on-a-chip includes the integration of individual operations (basically, reaction, preconcentration, separation and the corresponding detection) into mass-produced and low-cost device development. This recent progress toward separations in microchip HPLC has the potential to become a more powerful tool than nanoLC for the analysis of complex samples [40]. This is of particular interest for LC–MS or LC–tandem mass spectrometry (LC–MS/MS) analyses, because of its compatibility with the flow-rate requirements of a nano-electrospray interface for the online coupling of the microchip HPLC to MS. This coupling facility is one important advantage over microchip CE, which has clear difficulties with interfacing to MS. Agilent Technologies now markets a microfluidic chip that integrates a trapping column, separation column and electrospray source within a single device. 4.4.1

The Agilent HPLC Chip

H. Yin and K. Killeen, from Agilent Technologies, have recently reviewed the fundamental aspects and applications of the Agilent HPLC Chip [41]. This is a polyimide device which is simultaneously a nano-electrospray interface to a mass spectrometer, an analytical chromatographic column of appropriate size to the nano-electrospray flow rate, and an enrichment column for online sample concentration in advance of the analytical column. In this device, there are no

Miniaturized Systems for Analytical Separations I

151

Figure 4.10 View of the commercial Agilent HPLC Chip (top right) and traditional nano-LC components (around the image of the chip). Equivalent locations are indicated by arrows. Details of the electrospray tip are shown at the top of the figure as the HPLC Chip–MS interface. The photography of the Agilent nano-CLC–MS equipment shows the Agilent HPLC ChipCube where microchip is loaded. (Reproduced by permission of Agilent Technologies)

fittings, adapters, connectors or any other dispersive flow elements negatively affecting the performance in nano-LC. The commercial Agilent HPLC Chip is shown in Figure 4.10; traditional nano-LC components are displayed around this image, indicating their location in the chip. The figure is completed with the overall Agilent equipment in which the chip must be assembled, and details of the electrospray tip. Polyamide was chosen as the chip substrate because it is an extremely heat- and cold-stable material and has good dimensional stability (coefficient of thermal expansion of 20 ppm/
C). It is quite insoluble in most organic media and is not appreciably basic or acidic, which is a fundamental requirement for reversed phase LC. In contrast to glass or fused silica, channels can be formed by a laser ablation process, as this material absorbs the light. Thus, according to the information given by these authors [41], the microfabrication process consists of laser ablation of the channels, holes, chambers and columns. The sample-enrichment column and the analytical column are slurry packed with a variety of chromatography media. The basic HPLC Chip has an LC channel of 50 mm (d) · 75 mm (w) · 50 mm (l), with a 40 nl enrichment channel. Particles conforming the stationary phase are retained in the column space by the ‘keystone effect’, avoiding the need for frits, under operating pressure of over 120 bar [42]. The chip-packing process is adapted and modified from methods originally developed for the packing of fused silica capillaries. Both the enrichment and the chromatographic channels are packed using an input capillary that is connected to the chip by a chip-valve interface [43]. Transfer

152

Miniaturization of Analytical Systems

volume between the enrichment column and the analytical column (or other on-chip functions) is minimized by installing the HPLC Chip within a two-position rotary valve. Thus, the chip is sandwiched between the stator and the rotor of the valve, establishing a micro-to-macro interface. The valve assembly, the support plate and the attached valve are mounted on a two-axis stage via a rotating bracket that allows the chip to be loaded into the valve assembly (Figure 4.11a and Figure 4.11b), and then to be rotated into place such that the chip extends into a spray chamber mounted on the mass spectrometer (Figure 4.11c). The two-axis stage allows vertical and axial adjustment of the chip spray tip with respect to the cones on the mass spectrometer inlet (Figure 4.11d). The chip is interposed between the rotor and the stator with enough precision so that three basic running steps can be performed. In the initial position, mobile phase is pumped to the chip, passing through the enrichment unit and separation column

Figure 4.11 (a) Scheme of the chip with clamp in open position, including the stator, clamp and valve assembly. (b) Photography showing the chip in the clamped position ready for rotation into the spray chamber. (c) Details of the chip and valve assembly in operational position on the MS interface. (d) Photography of the spray formed in the nano-LC–MS interface after the flow exits by the electrospray tip. (Reproduced from [43], Copyright 2007 American Chemical Society, and by permission of Agilent Technologies)

Miniaturized Systems for Analytical Separations I

153

Figure 4.12 Scheme of the chip–rotor interface in the LC run mode (a) and in the loading process: the sample-loading configuration (b) and the LC-running configuration (c) of the rotor channels. The upper-right part of the figure shows the scheme of the LC run mode with the sixway positions of the rotary valve. (Reproduced from [43], Copyright 2007 American Chemical Society, and by permission of Agilent Technologies)

and then reaching the spray (base-line signal); whereas sample flow is driven to waste (Figure 4.12a). In the enrichment position of the valve, sample is introduced to the chip flow from the autosampler, driven through the enrichment column, and then to waste (Figure 4.12b). When the rotor is again switched 60
(injection of the preconcentrated analytes), the flow from the nanopump enters the enrichment column and sweeps the analytes into the analytical column (Figure 4.12c). At the end of this column the flow passes electrical contacts which allow the biasing of the effluent for electrospray. Agilent has developed the HPLC ChipCube to optimize the performance of the entire system. The valve is designed to form a face seal with the HPLC Chip with minimum mechanical wear on the valve rotor. Figure 4.13 shows details of the entrance of the chip into the cube through the sandwiched position between the stator and the rotor of the valve. The chromatographic performance of the analytical column on an HPLC Chip was studied by G. Rozing et al. [44]. The results obtained for a group of peptides showed that chromatographic peaks resulting from HPLC Chip were narrower and more symmetrical than those obtained from a packed fused-silica capillary column. This fact is in agreement with the recordings shown in the article published by S. Ehlert and U. Tallarek [40], reproduced from a presentation by Vollmer and Miller [45]. These are given in Figure 4.14. As can be seen, the HPLC Chip offers increased resolution and superior peak shape compared to conventional nano-LC

154

Miniaturization of Analytical Systems

Figure 4.13 Introduction and sandwiched position of the chip in the stator–rotor unit. (Reproduced by permission of Agilent Technologies)

Figure 4.14 Comparison of microchip LC–MS with nano-LC–MS by analysis of a yeast gel band. (Reproduced from [40] with permission from Springer)

Miniaturized Systems for Analytical Separations I

155

using the same adsorbent particles and similar packed-bed dimensions. The worst results were obtained using a longer nano-LC column packed with smaller adsorbent particles. 4.4.2

Other Approaches to Microchip HPLC

Other microchip LCs have been described in the literature. T.D. Lee and coworkers developed a microfluidic chip that integrates all the fluidic components of a gradient LC system [46], designed as a platform for an LC–MS/MS microsystem. This chip was batch fabricated on a silicon wafer using photolithographic processes, with Parylene as the main structural material. The chip includes three electrolysis-based electrochemical pumps, one for loading the sample and the other two for delivering the solvent gradient; platinum electrodes for delivering current to the pumps and establishing the electrospray potential; a low-volume static mixer; a column packed with silica-based reversed phase support; integrated frits for bead capture; and an electrospray nozzle. Figure 4.15 shows a photograph of the chip and the corresponding schemes. As Figure 4.15b illustrates, the chip is basically conformed on a 500 mm silicon wafer layer with various layers on top deposited for the fabrication of the different chip elements (50 mm thicknesses). In this top layer is packed the chromatographic column (RP column), the mixer, the exit for the electrospray and the chamber for the pump or the sample. The chip is mounted into a polyetherimide jig that couples a port at the front of the column to Teflon tubing via a poly(dimethylsiloxane) (PDMS) gasket. A slurry of the stationary phase material is forced into the column through an access port between it and the mixer until its entire length is filled. Once packed, the chip is mounted in a different holder that utilizes a 5 mm-thick polyetherimide cover. Chambers are machined into this cover piece to form reservoirs for the sample and solvents. The PDMS gasket allows the

Figure 4.15 (a) Photograph of the LC chip. (b) Diagram showing the placement of the different elements. (c) Chip-holder assembly for the inlet in an Agilent MSD ITMS. (Reproduced from [46] with permission of American Chemical Society, Copyright 2005)

156

Miniaturization of Analytical Systems

sealing between the cover and the chip. For the analyses, the completed chip assembly is mounted on an Agilent electrospray ion source housing using a modified probe that incorporates a 3D positioner (Figure 4.15c). A. Ishida and coworkers have developed a microchip for reversed phase LC using porous monolithic silica [47]. The chip consists of a double T-shaped injector and a 40 cm separation channel (in a serpentine configuration). Figure 4.16 shows a photograph and details of this LC chip. This microchip is fabricated using glass wafer as the substrate and the cover plates. Standard photolithography, wet chemical etching and bonding techniques are used in the different steps of the fabrication process. The layout and the dimensions of the different sections of the chip are shown in Figure 4.16b. Two photomasks are prepared for the fabrication. The first defines the double T-type injector and separation channel pattern, and the second defines the grooves which will serve as connectors. The overall dimensions of this chip are 35 mm · 35 mm. The typical channel width is about 400 mm, and the channel depth is about 30 mm. The connection grooves are etched to a depth of 150 mm. After rinsing with water, the etched glass chip is aligned to the cover plate

Figure 4.16 (a) Photograph of the LC chip. (b) Schematic layout of the chip with details of the channels and inlet/outlet positions. (c) Overall configuration including the LC chip. S, sample; MP, mobile phase; ED, electrochemical detector; W1, mobile phase waste; W2, sample drain; V1, mobile phase inlet valve; V2, mobile phase outlet valve; V3, sample inlet valve; V4, sample outlet valve. (Reprinted from [47] with permission from Elsevier)

Miniaturized Systems for Analytical Separations I

157

and then thermally bonded in a programmable oven. Then fused silica capillaries are inserted into the grooves. Monolithic stationary phase in the microchip is created by introducing reactant solutions by syringes. Figure 4.16c shows the schematic configuration of the chip LC system. The ends of the capillaries inserted in the microchip are connected to two-way valves (V1, V3 and V4 in Figure 4.16c) via heatshrinkable tubes. Valves V1 and V3 are connected to syringes allowing the introduction of the mobile phase (V1) and the sample (V3), whereas valves V2 and V4 drive to waste. The end of the separation channel in the microchip is connected to an electrochemical detector (ED) using conventional amperometric detection. Different catechin compounds can be separated in this chip with good resolution, although long analysis times are required. The sample-loading capacity of microchip LC systems can be enhanced by introducing parallel multichannel flow pass, analogous to multicapillaries or fibrepacked capillaries, as T. Greibrokk and coworkers pointed out [48]. This possibility will increase the sample throughput and the speed of the analysis because of the better resolution of separation in microchips. 4.4.3

Some Selected Applications

LC analysis using the microchip approach is practically addressed to the analysis of very small sample sizes, in which appropriate resolutions and sensitivity must be achieved. As the rest of the auxiliary elements, including the detector, are commonly not miniaturized, portability is not the objective of such analytical microsystems. Therefore, biological or related bioanalytical applications have received the main attentionofmicrochip LC, mainlythe use ofMSdetectionfor reliablescreening ofthe numeroustargetcompoundsin thesesamples.Inthissection, some selectedexamples illustrate the practical potential of microchip LC systems. In the pharmaceutical field, for example drug metabolism and pharmacokinetic studies, it is common to use small animals for testing prior to human assays. The animal size sets the limit on the blood volume that can be drawn per unit time, and serial bleeding of a few or even only one animal, which reduces the variability in the pharmacokinetic profile, results in even lower volumes per bleed. Consequently, studies with small animals set extreme demands on the analytical system with respect to both the capacity to handle small sample volumes and the detection sensitivity. For these studies, LC–MS is often used for the analysis of samples. Here, the HPLC chip/MS system fulfils these requirements. In this context, S. Buckenmaier et al. have used the HPLC Chip interfaced to a triple quadrupole (QQQ) mass spectrometer for the monitoring of atenolol, atropine, metroprolol and imipranine in blood plasma samples [49]. M. Vollmer and S. Buckenmaier have coupled the HPLC Chip with a time-offlight (TOF) mass spectrometer detector for the determination of dextromethorphan (DEM) and its metabolites in human plasma and urine samples [50]. The

158

Miniaturization of Analytical Systems

metabolism of DEM has been extensively investigated and has been used as a model drug to distinguish between ‘poor’ and ‘efficient’ cytochrome P450D6 metabolizers. During the metabolism (Figure 4.17a), DEM (m/z ¼ 272) is either demethylated to 3MM (m/z ¼ 258) or converted to the isobar dextrorphan (DOR). The latter compound may be further metabolized to 3OM (m/z ¼ 244) or directly glucoronidated to dextrorphan glucoronide (DORGlu, m/z ¼ 434). 3OM is further metabolized to 3-hydroxymorphinan glucuronide (3OMGlu, m/z ¼ 420). The samples are analysed with two different gradients. Figure 4.17b shows the typical elution profile for the fast gradient. Proteomics is a field in which the LC microchip is paid more attention. Initially, proteomics had the objective of identifying the proteins of biological systems, but more and more this discipline is moving to targeted strategies aiming at identifying key proteins (biomarkers, for instance) that can provide reliable diagnostic and prognostic indicators of disease progression or treatment effects. The problem is that biological samples are complex mixtures of proteins and the detection of biomarkers (existing at very low abundance) in these mixtures needs sensitive and robust analytical methods. One of the conventional and most powerful proteomics approaches involves combining high-resolution separations with high-accuracy MS. Since LC is complementary to MS/MS with respect to the population of peptides that may be detected, an interesting approach is to combine different pieces of information obtained by these two techniques. Thus, for instance, the retention time (RT) obtained by CLC clearly defines a peptide, as does its mass; the characterization of a peptide by a couple of pieces of data (mass measurement

Figure 4.17 (a) Metabolism of DEM by citochrome P4502D6. (b) Analysis of DEM metabolites by HPLC Chip–TOFMS (100 nM in vitro assay; concentration on column 100 fmol; gradient length 2.5 minutes; total run time 9.5 minutes). (Reproduced from [50] with permission from Agilent Technologies)

Miniaturized Systems for Analytical Separations I

159

and the corresponding RT, noted as mass/retention time, MRT, for m/z-RT) is very attractive. In this way, a protein will be characterized by different MRT values, which constitute a specific peptide map. Therefore, both the RT and the m/z values must be as accurate and reproducible as possible. The replacement of the CLC equipment by a HPLC Chip system can be of great value in these cases, as microfluidics chromatographic separations provide high separation efficiency and deliver precise and repeatable flow rates. Using these advantages of HPLC Chip–MS systems, J. Hardouin et al. have investigated different biomarkers signatures [51]. In this case, an ITMS was used. These authors carried out the identification of human autoantigens after obtaining autoantigen mixtures by affinity separation (in order to simplify the complexity of the sample) of Caco-2 cell proteins on the immunoglobulins from a group of healthy volunteers. These mixtures were digested by trypsin, and just the tryptic digest-resulting solutions were then analysed by HPLC Chip–MS coupled with an IT. In each sample, around 20 proteins were identified using the software package MASCOT. The measurement of MRTs revealed the identity of a peptide in different situations, such as the lack of fragment information, measured m/z above the authorized mass tolerance, or the presence of different peptides sequences with close m/z ratios. Anti-doping laboratories use a test to detect the misuse of recombinant erythropoietin (rhEPO) based on its different migration pattern on isoelectric focusing (IEF) gel compared to endogenous human erythropoietin (hEPO), which can be explained by structural differences. While there is definitely a need to identify those differences by LC–MS/MS, the extensive characterization that was achieved for the rhEPO was never performed on hEPO because its standard is not available in sufficient amount. Taking this problem into account, P.E. Groleau and coworkers developed an analytical method to detect pmol amounts of N-linked and O-linked glycopeptides of the recombinant hormone (used as a model), using a nanoflow HPLC Chip–ESI–ITMS [52]. The diagnostic ion at m/z 366 of oligosaccharides was monitored in the product ion spectra to identify the four theoretical glycosylation sites, Asn24, Asn38, Asn83 and Ser126 on glycopeptides 22–37, 38–55, 73–96 and 118–136, respectively. The method described by these authors provides a means to detect glycopeptides from commercially available pharmaceutical preparations of rhEPO with the sensitivity required to detect pmol amounts of hEPO, which allows the identification of structural differences between the recombinant and the human forms of the hormone. Proteins in general do not act individually in biological processes. Most cellular events are controlled by multiple proteins, involving – in many cases – noncovalent interactions between proteins and other molecules. Traditional methods used to study noncovalent protein interactions include ultracentrifugation, calorimetry and various types of spectroscopy. Commonly, these methods require large amounts of sample, are relatively slow and can be fairly nonselective. For these reasons, ESI–MS has recently been used to study noncovalently-associated

160

Miniaturization of Analytical Systems

complexes. Despite the advantages of technique for this purpose, care must be taken to ensure the protein complex does not denature in the solvents used for ESI. Moreover, other instrumental parameters (temperature, voltage and pressure) must be controlled to avoid fragmentation of the complex. For identification purposes, high-resolution, accurate-mass and high-mass scanning are critical features in the analysis of macromolecular assemblies. TOFMS posseses all of these capabilities, and hence is a good choice for the study of noncovalent associations. In addition to these advantages of TOF detectors, HPLC Chip provides an extremely stable nanospray and significantly improves upon the ease of use and reliability of nanoflow ESI. G.W. Kilby demonstrated this suitability for the study of apomyoglobin and myoglobin noncovalently-associated complexes using an HPLC Chip–ESI–TOFMS arrangement [53]. The analysis was carried out in less than five minutes. N. Tang and P. Goodley carried out the characterization of protein phosphorylation using HPLC Chip electron transfer dissociation (ETD) ITMS [54]. Protein phosphorylation is a type of post-translational modification which plays an important role in the regulation of many cellular functions. The precise determination of phosphorylation sites within a protein is crucial to the understanding of cell regulation mechanisms. Collision-induced dissociation (CID) and ETD can be used in the same LC–MS run to analyse samples for protein phosphorylation studies. Figure 4.18a depicts the scheme for the ETC process. The ETC reactant is generated by a small negative chemical ionization (NCI) source that is mounted directly upon the inlet section of the second octapole ion guide on the ITMS. The NCI source is filled with fluoranthene, which is sublimed and combined with methane gas in the presence of electrons emitted from a filament. The electrons are slowed by collision with the methane gas and captured by the fluoranthene molecules to make fluoranthene radical anions. From the ESI chamber, the peptide ions are generated from the nanospray chip tip. All the positive ions are allowed to enter the MS ion inlet, through the ion optics and ultimately into the IT. The positive multiplycharged peptide ions are isolated and all other ions are ejected. During this process, the flow of the reactant radical anions is closed by a voltage-gating process. Alternatively, the NCI source can be gated to allow negative ions to enter the IT as a packet. The two packets of ions (negative and positive) exist in the ITat the same point in time. After a few milliseconds, the positive peptide ions and the negative ions become interactive and the electron from the fluoranthene radical anions is transferred to the positive peptide ion. The electron transfer is rapid and sufficiently energetic, yielding a series of positive amino acid residue fragments. The IT is subjected to the scan process in the same manner as CID and the positive ETD product ions are scanned out of the trap, yielding the ETD MS/MS spectrum [54]. As an example of the results obtained by this methodology, Figure 4.18b illustrates the differences between CID and ETD fragmentation for the synthetic phosphopeptide TTHyGSLPQK using HPLC Chip nanospray infusion. As can be seen, CID has very

Miniaturized Systems for Analytical Separations I

161

Figure 4.18 (a) Diagram of the Agilent 6340 ITMS with ETD module. (b) Comparative spectra obtained with CID and ETD for the phosphopeptide TTHyGSLPQK (m/z ¼ 404.7) using the HPLC Chip. (Reproduced from [54] with permission from Agilent Technologies)

few fragment ions, while ETD produces nearly complete sequence coverage, with indication of the phosphotyrosine location. The majority of the protein drugs in existence are glycoproteins. It has been demonstrated that the efficacy of a glycoprotein drug (commonly used to treat and prevent diseases) critically depends on the type and extent of glycosylation of the protein. Therefore, it is of great interest to characterize glycoproteins as an essential part of drug quality control. P.D. Perkins has explored the ability of HPLC Chip coupled to a trap XCT ultra-MS system to detect and identify oligosaccharides and N-linked glycans cleaved from two commercially available glycoproteins [55]. The HPLC Chip used in this case contains a porous graphitized carbon column

162

Miniaturization of Analytical Systems BPG, 700-2200 m/z Glycans from 1 pmol starting glycoprotein

Gal Man GlcNAc NeuAc Fuc

or G1

G0

G2

A B

C A: Isoform of Go B: Isoforms of G1 C: Isoform of G2

10

12

14

16

18

20

22 Time (min)

Figure 4.19 Base peak chromatogram of released reduced N-linked glycans from human serum IgG. (Reproduced from [55] with permission from Agilent Technologies)

Table 4.3 Relative percentage amounts of the IgG N-linked glycans identified by HPLC Chip/Trap XCT Ultra MS system from Agilent Technologies. (Reproduced from [55] with permission from Agilent Technologies) Glycan (including isoforms)

Structure

m/z value Peak height of Relative of alditol [Mþ2H]2þ averaged spectrum percent

G0

733.4

46886

14.3

G1

814.4

103172

31.4

G0 þ bisecting GlcNAc

835.2

30969

9.4

G2

895.4

35483

10.8

G1 þ bisecting GlcNAc

916.0

25222

7.7

G2 þ bisecting GlcNAc - fuc

924.0

7623

2.3

G1 þ NeuAc

959.9

16101

4.9

G2 þ bisecting GlcNAc

997.0

24573

7.5

1041.0

38886

11.8

G2 þ NeuAc

Miniaturized Systems for Analytical Separations I

163

specifically designed for oligosaccharide separation applications. The combination of the chip with the sensitivity and scan speed of the trap MS detector provide the appropriate platform to characterize N-linked glycans. After a reasonable reduction of the number of plausible candidate compounds, base peak chromatograms, such as that shown in Figure 4.19, were obtained. As can be seen in this figure, the known major glycans in IgG, designated by G0, G1 and G2, were found. The two linkage isoforms of G1 were separated, due to the high resolving power of the porous graphitized carbon column for oligosaccharide separation. Some minor isoforms were detected whose MS/MS spectra also matched those for G0, G1 and G2. Additionally, the data produced by this system enabled the calculation of the relative distribution of glycans in the sample. Thus, Table 4.3 shows the identities and relative percentage distributions of the IgG N-linked glycans.

References [1] S.C. Ferry, J.H. Jerman, J.B. Angell, IEEE Trans. Electron. Devices, ED-26 (1979) 1880. [2] J.P. Kutter, Trends Anal. Chem., 19 (2000) 352. [3] H. Kim, W.H. Steinecker, S. Reidy, G.R. Lambertus, A.A. Astle, K. Najafi, E.T. Zellers, L.P. Bernal, P.D. Washabaugh, K.D. Wise, Transducers & Eurosensors ’07, 3A3.4 (2007) 1505. [4] A. de Mello, Lab on a Chip, 2 (2002) 48N. [5] P.R. Lewis, IEEE Sensors J., 6 (2006) 784. [6] M. Stadermann, Anal. Chem., 78 (2006) 5639. [7] Y. Wang, Q.A. Lin, T. Mukherjee, Lab on a Chip, 4 (2004) 453. [8] S.C. Jacobson, R. Hergenroder, L.B. Koutny, R.J. Warmack, J.M. Ramsey, Anal. Chem., 66 (1994) 1107. [9] S. Zampolli, I. Elmi, J. Sturman, S. Nicoletti, L. Dori, G.C. Cardiani, Sen. Actuators B, 105 (2005) 400. [10] M. Agah, G.R. Lambertus, R. Sacks, K. Wise, J. Microelectromechanical Syst., 15 (2006) 1371. [11] S. Reidy, G. Lambertus, J. Reece, R. Sacks, Anal. Chem., 78 (2006) 2623. [12] P. Ostman, L. Luosujarvi, M. Haapala, K. Grigoras, R.A. Ketola, T. Kotiaho, S. Franssila, R. Kostiainen, Anal. Chem., 78 (2006) 3027. [13] A.D. Radadia, R.I. Masel, M.A. Shannon, Transducers & Eurosensors ’07, 3EI1.P (2007) 2011. [14] G. Lambertus, A. Elstro, K. Sensenig, J. Potkay, M. Agah, Anal. Chem., 76 (2004) 2629. [15] G. Lambertus, R. Sacks, Anal. Chem., 77 (2005) 2078. [16] J. Muller, O. Krusemark, U. Lehman, Micro Total Analysis Systems 2000, Kluwer Academic Publisher, Boston, USA, 2000 pp 167–170. [17] L.A. Kartsova, A.A. Makarov, Russ. J. Appl. Chem., 75 (2002) 1725. [18] Q. Li, D. Yuan, J. Chromatogr. A, 1003 (2003) 203. [19] C. Saridara, S. Miltra, Anal. Chem., 77 (2005) 7094. [20] M. Stadermann, A.D. McBrady, B. Dick, V.R. Reid, A. Noy, R.E. Synovec, O. Bakajin, Anal. Chem., 78 (2006) 5639. [21] E.T. Zellers, S. Reidy, R.A. Veeneman, R. Gordenker, W.H. Steinecker, G.R. Lambertus, H. Kim, J.A. Potkay, M.P. Rowe, Q. Zhong, C. Avery, H.K.L. Chan, R.D. Sacks, K. Najafi, K.D. Wise, Transducers & Eurosensors ’07, 3A3.1 (2007) 1491. [22] T. Takeuchi, Chromatography, 26 (2005) 7. [23] D. Ishii, Jasco Report, 11 (1974) 1. [24] T. Takeuchi, Anal. Bioanal. Chem., 375 (2003) 26. [25] J.M. Saz, M.L. Marina, J. Sep. Sci., 31 (2008) 446.

164

Miniaturization of Analytical Systems

[26] F.Y. Che, L.D. Fricker, J. Mass Spectrom., 40 (2005) 238. [27] A.J. Oosterkamp, M. Carrascal, D. Closa, G. Escobar, E. Gelpı, J. Abian, J. Microcolumn Sep., 13 (2001) 265. [28] A. Hazotte, D. Libong, M. Matota, P. Chaminade, J. Chromatogr. A, 1170 (2007) 52. [29] R.W. Dixon, D.S. Peterson, Anal. Chem., 74 (2002) 2930. [30] S. Heron, A. Tchapla, J. Chromatogr. A, 1140 (1999) 95. [31] L. Quinton, K. Gaudin, A. Baillet, P. Chaminade, J. Sep. Sci., 29 (2006) 390. [32] K. Hibi, D. Ishii, I. Fujishima, T. Takeuchi, T. Nakanishi, J. High Resolut. Chromatogr. Commun., 1 (1978) 21. [33] T. Tsuda, M.V. Novotny, Anal. Chem., 50 (1978) 632. [34] N. Ishizuka, H. Minakuchi, K. Nakanishi, N. Soga, N. Tanaka, J. High Resolt. Chromatogr., 21 (1998) 477. [35] R.D. Dandeneau, E.H. Zerenner, J. High Resolut. Chromatogr. Chromatogr. Commun., 2 (1979) 351. [36] Q.-S. Qu, X.-X. Zhang, Z.-Z. Zhao, X.-Y. Hu, C. Yan, J. Chromatogr. A, 1198–, 1199 (2008) 95. [37] J. Hernandez-Borges, Z. Aturki, A. Rocco, S. Fanali, J. Sep. Sci., 30 (2007) 1589. [38] Y. Shen, N. Tolic, C. Masselon, L. Pasa-Tolic, D.G. CampII, M.S. Lipton, G.A. Anderson, R.D. Smith, Anal. Bioanal. Chem., 378 (2004) 1037. [39] D. Schaumioffel, P. Giusti, H. Preud’Homme, J. Szpunar, R. Lobinski, Anal. Chem., 79 (2007) 2859. [40] S. Ehlert, U. Tallarek, Anal. Bioanal. Chem., 388 (2007) 517. [41] H. Yin, K. Killeen, J. Sep. Sci., 30 (2007) 1427. [42] M. Mayer, E. Rapp, C. Marek, G.J.M. Bruin, Electrophoresis, 20 (1999) 43. [43] H. Yin, K. Killeen, R. Brennen, D. Sobek, M. Werlich, T. van de Goor, Anal. Chem., 77 (2007) 527. [44] G. Rozing, T. van de Goor, H. Yin, K. Killeen, J. Sep. Sci., 27 (2004) 1391. [45] M. Vollmer, C. Miller, Poster Presentation in the ABRF Meeting, Savannah, GA, USA (2005). [46] J. Xie, Y. Miao, J. Shih, Y.-C. Tai, T.D. Lee, Anal. Chem., 77 (2005) 6947. [47] A. Ishida, T. Yoshikawa, M. Natsume, T. Kamidate, J. Chromatogr. A, 1132 (2006) 90. [48] Y. Saito, K. Jinno, T. Greibrokk, J. Sep. Sci., 27 (2004) 1379. [49] S. Buckenmaier, M. Vollmer, L. Trojer, C. Emotte, The Column, May, (2008) 20. [50] M. Vollmer, S. Buckenmaier, Agilent Application Note, 5989-5938EN (2006). [51] J. Hardouin, R. Joubert-Caron, M. Caron, J. Sep. Sci., 30 (2007) 1482. [52] P.E. Groleau, Ph. Desharnais, L. Cote, C. Ayotte, J. Mass Spectr., 43 (2008) 924. [53] G.W. Kilby, Agilent Application Note, 5989-5596EN (2006). [54] N. Tang, P. Goodley, Agilent Application Note, 5989-5158EN (2006). [55] P.D. Perkins, Agilent Application Note, 5989-5157EN (2006).

5 Miniaturized Systems for Analytical Separations II: Systems Based on Electroosmotic Flow (EOF) 5.1

Introduction

The most commonly used separation techniques in analytical microdevices are chromatography and electrophoresis. Chromatography is suitable for the separation of both charged and uncharged molecules. The separation makes use of physicochemical differences between the components, such as different affinities for a stationary phase or different distribution coefficients between a stationary phase and a mobile phase. Chromatographic techniques require the presence of a stationary and mobile phase. Electrophoresis, on the other hand, can be used for charged molecules (ions), which, in the simplest mode of electrophoresis, are separated in an electric field according to their charge-to-mass ratios. Besides this electrophoretic motion, an additional movement is often imparted by electro-osmotic flow (EOF) phase, through which the mobile phase and the mixture to be separated are percolated. Movement of mobile phase is induced by pressure (chromatography) or electro-osmosis (capillary electrophoresis). With the separation technique selected, three steps must be executed in order to perform a successful, useful separation in a miniaturized format: (i) injection, in which a defined portion of sample is introduced into the separation system by pressure or by electrokinetic means; (ii) separation, during which the different components in the sample are moved through the separation system and separated into individual bands or zones; and (iii) detection, by which the individual bands are Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

166

Miniaturization of Analytical Systems

registered at the outlet of the separation system based on various physical principles. Though the first analytical microdevices were chromatographic systems [1,2], it was soon realized that it would be easier from an engineering point of view to create electrophoretic systems first [3–5]. In Chapter 4 we discussed the miniaturization of systems for analyte separation based on hydrodynamic flow; in this chapter we are going to study the miniaturization of capillary electrophoresis. The miniaturization of capillary electrophoresis (CE microchips) was one of the earliest examples of a mTAS system [6]. This technology emerged as an important new analytical technique in the early 1990s with the introduction of the mTAS concept [6] and the seminal work of Manz, Harrison, Verpoorte and Widmer [7]. This new technique was a result of the marriage of the ability of conventional CE to analyse ultrasmall volumes (nl) with microfabrication techniques perfected in the semiconductor industry to produce very small structures in silicon. CE microchips employ channels etched into a planar substrate, which are based upon microfabrication techniques developed in the semiconductor industry. Early applications of CE microchips employed glass or quartz for chip substrates because they are optically transparent and exhibit EOF properties similar to those of fused silica. A number of other materials, including polymers/plastics, have now been investigated [3,4,8]. Moreover, CE microchips have the potential to simultaneously assay hundreds of samples in a matter of minutes or less. This rapid analysis combined with massively parallel analysis arrays should yield an ultrahigh throughput. These microchips typically consume only picolitres of sample and they may potentially be prepared onboard for a complete integration of sample preparation, separation and detection. Also, a key advantage of the use of these planar devices/substrates is the ease of manufacture and the fact that microfabrication techniques can be adapted for mass production. Therefore, these features make CE microchips an attractive technology for the next generation of CE instrumentation [8]. The key features of CE microchips have been widely stated in the literature and are summarized in Table 5.1 [9]. Very fast analysis without loss of performance (high efficiency) and the possibility of performing several assays simultaneously (parallelization of analysis) are among the most attractive characteristics. Very low sample consumption and waste generation are additional important features, sometimes connected with the low cost, which is mostly due to the use of inexpensive materials (plastics) in single-use microsystems (disposability). A comprehensive overview of these features will be given in the following sections. However, one of the most unique characteristics is the possibility of integrating different analytical steps into the single device (highly-multiplexed systems or lab-on-a-chip). Although this will be touched upon in this chapter, it will be discussed in more depth in Chapter 8. In addition, in this chapter we will study the designs, fundamentals and applications of CE microchips, giving a concise but in-depth overview of the

High efficiency

Ultrafast analysis

Very low sample/reagent consumption

Extremely low waste generation Portability

Disposability

Highly multiplexed systems Lab-on-a-chip

Ultrahigh throughput

High efficiency

Rapid analysis

Low sample

Low waste generation

No disposability

Integration

High sample throughput

No portability

Analytical features

Analytical features

Conventional CE

CE microchip

Miniaturization of CE conceptually offers portability, as a consequence of the size of the analytical instruments Single-use polymeric/plastic microchips can be readily produced at low cost, so that they can be disposed of after use System integration of components of various functions on a single chip enables the construction of a multifunctional analytical laboratory (see Chapter 8) Example for CE microchips: parallel separation channels can be fabricated on a single chip device, enabling high-throughput assays on microchips (see Section 5.8.1)

The miniaturized feature of microchip devices reduces the consumption of samples and reagents, and in consequence extremely low wastes are generated

Applicability of high field strength results in high efficiency per unit length of separation channel (see Section 5.3) The efficient heat dissipation of microchip channels allows application of high electric field strength, which provides high-speed electrophoretic separation (see Section 5.3)

Remarks

Table 5.1 Relevant keys for conventional CE and CE microchips

Facility to use the electrokinetic phenomena (valveless microsystems)

Existence of EOF (glass and polymers)

Easy fabrication of network microchannels with zero dead volume interconnections (microchip format concept) The chemistry of glass and polymers (PDMS, PMMA) is well known

Technical features

Miniaturized Systems for Analytical Separations II 167

168

Miniaturization of Analytical Systems

heterogeneous aspects involved in them, both technological and analytical. For more details, see the excellent reviews [3–5,8] and some informative books [10–12].

5.2 CE on the Microchip Format CE miniaturized into microchip format is an analytical microsystem constituted at least by an injector (where a sample plug is critically loaded) and one separation microchannel (where electrophoretic separation of analytes is performed), interfaced suitably to reservoirs (where different solutions/samples are deposited). Microchannels and reservoirs are fabricated in microchips using photolithography or micromolding to form channels for sample injection, CE separation and analyte detection. Once all solutions, including those of the samples, are loaded, the samples are typically transferred electrokinetically into an injector region. Then their components are separated by application of a high voltage, and afterwards detected using a suitable detection system. The system-integration concept on the microchip platform eliminates the necessity of most fluidic connections that otherwise link microfluidic components. Avoiding such connections greatly reduces sample dispersion, delay times and dead volumes between the different microchip compartments, significantly increasing the separation power of such integrated miniaturized systems. The typical layout of a CE microchip (both a simple cross-T injector (a) and a twin-T injector (b)) is depicted in Figure 5.1. It has a network of channels with (a)

(b) Reservoirs (100–250 µl)

RB

S

RB

SW

SW S

L=3–10 cm Separation channel

10–100 µm

20 µm 50 µm

W

Simple Cross-T Injector

Detection cell

W

Twin-T injector

Figure 5.1 Common microchip layouts

Highvoltage supply

Miniaturized Systems for Analytical Separations II

169

widths varying from 10 to 100 mm, with typical straight separation channels between 3 and 10 cm. In a typical setup, buffer solutions are introduced on sample (S), running buffer (RB) and waste (W) reservoirs. The reservoir volume is often defined by their capacities (100–250 ml). Electric fields are applied to the reservoirs by high-voltage power supplies (1–5 kV) and platinum electrodes are placed in the reservoirs for injection and separation as samples are normally electrokinetically driven into the network channels. In microchip electrophoresis and related separation techniques based on EOF, the primary interface between the regular laboratory environment and the microworld is the buffer reservoirs [13]. Their accessibility and volume are key issues in microchip design and development. One of the considerations is the reduction of buffer separation/reagent volume and sample consumption. Early microchip designs had very low-volume self-contained reservoirs, and increases in glass cover-plate thickness increased the reservoir volume [14]. The two main reasons for increasing reservoir volume were to counter solvent evaporation and to suppress the effect of electrolysis, as the latter can cause unpredictable pH changes in the buffer reservoirs. To keep the pH stable during the electrophoresis process, higher buffer capacities were also considered; however, this led to higher current and concomitantly higher Joule heat generation. To reduce these effects, reservoirs are now designed and fabricated with highly reproducible sizes and shapes and with chemical incompatibility. Also, drilled reservoirs vary somewhat in size and shape in terms of both reservoir-to-reservoir reproducibility and consistency of the inner diameter throughout the reservoir depth. Thus, a robust and reproducible manufacturing process for these connections is crucial to the wide applicability of microseparation devices. With respect to microchip layouts, in general the design of microchips for CE has undergone significant development from simple single-channel structures to increasingly complex ones [3,4,13]. Current designs of CE chips allow reactions on-chip and separations in multiple channels. Microreactors can also be added to perform online reactions such as (bio)chemical reactions and post-column derivatizations (see Section 5.8). Separation channel length and geometry must also be properly designed. The change of the CE column format from long capillaries to shorter microfabricated channels on microchips brings new opportunities, limitations and challenges. Miniaturization, smaller injection plugs and shorter separation paths enable rapid separations without significant peak-broadening due to diffusion. With respect to channel geometry, as mentioned above, CE microchips have channel depths of 15–40 mm and widths of 60–200 mm. The small cross-section of separation channels and the large thermal mass of microchips allow Joule heat to be dissipated efficiently. Thus, high electric fields (over 2 kV/cm) can be applied on microchips. These issues will be discussed in the next section.

170

Miniaturization of Analytical Systems

5.3 Modes and Theories of CE Microchips The transfer of the knowledge gained in conventional CE to microchip format theory is not new. Basically all electrophoretic modes have been successfully transferred to microchip format. Capillary zone electrophoresis (CZE) is the most commonly used separation mode in CE microchip due to its versatility, ease of operation and separation power. However, in principle, it has the limitation of being useful only for separation of charged species (neutral species cannot be separated). In CZE the channel contains only a buffered medium and the eletrophoretic mobilities of the analytes depend on their charge/mass ratio. Indeed, the separation mechanism of the components of the sample is based on these differences (charge/mass ratios), which allow them to have different electrophoretic mobilities and therefore different velocities, and it is possible to separate cations and anions if the EOF is strong, in order to lead anionic substances to the detector located at the cathode (direct polarity). Figure 5.2 shows the separation principle in connection with the microchip design. In general, this mode has been preferred in small molecules analysis (see Section 5.8). Capillary gel electrophoresis (CGE) combines the good resolution of classical gel electrophoresis with the simple instrumentation used in CZE. The separation principle is: a capillary is filled with a gel (with pores of controlled size) that acts as a molecular sieve. A scheme of the separation mechanism is shown in Figure 5.3(a). It is based on the different mobilities of the components of the sample through the pores of the gel due to their different molecular sizes (molecular sieving). The compounds of small size will traverse the gel matrix

µep + anode

< µep

< buffer µeo

µep

Detection – cathode

Figure 5.2 Principle of CZE on microchip. (Reprinted from [15] with permission and adapted)

Miniaturized Systems for Analytical Separations II

171

DETECTOR

(a) GEL

anode

cathode –

+

pl1

(b)

pl2

pl3

H3PO4

NaOH

anode

cathode

pH gradlent

+

– Small pH bajo

(c)

High pH alto

pl

µec

µPG

anode +

O P

A A

cathode –

Figure 5.3 Principles of CE modes on microchip. (Reprinted from [15] with permission and adapted)

easier and will have shorter analysis times than those substances of large size, which traverse the gel more slowly. DNA analysis is the golden example of this electrophoretic mode in microchip (see Section 5.8). Capillary isoelectric focusing (CIEF) is employed to achieve separation of proteins, peptides, amino acids and other substances of amphoteric character. In CIEF the separation is not based on the differences in electrophoretic mobilities but on the differences in pI of solutes. Separations by CIEF are based on the electrophoretic migration of amphoteric substances in a pH gradient, as shown in Figure 5.3(b). The pH gradient needed though the capillary is obtained by using a solution of ampholytes (substances of zwitterionic character with different pKa values (carrier ampholyte)). The sample can be added to the carrier electrolyte. When the electric field is applied, a pH gradient is established within the channel from the anode (low pH) to the cathode (high pH). Sample components will move until they find a pH value equal to their pI, since at this pH they lose their charge and then stop (this phenomenon is called focusing). In isotachophoresis (ITP), a volume of analyte is placed between a leading electrolyte (LE) and a terminating electrolyte (TE). The LE is composed of highmobility ions and so has a higher mobility than the sample ions and the TE. Conversely, the TE is composed of low-mobility ions and so has lower mobility than the LE and sample ions. During the electrophoretic separation, analytes in the

172

Miniaturization of Analytical Systems

sample are arranged into discrete bands in order of mobility. The velocity and concentration of these bands are adapted to a value governed by the leading ion. Consequently, both separation and concentration adaptation occur simultaneously. Electrokinetic chromatography has also been transferred to microchip format. Micellar electrokinetic chromatography (MEKC) is characterized by the use of micelles as pseudophase. Micelles are aggregates that are formed by adding surfactants to the separation buffer at a concentration above their CMC. Ionic surfactants form micelles that move at a different velocity to the EOF. In this case, the separation of neutral molecules is produced by their distribution between the aqueous and the micellar phases. When the micellar phase moves to the detector, the elution range is between the time corresponding to EOF and the migration time of the micelles, and all analytes will migrate between these two limits depending on their distribution between the two phases (see Figure 5.3(c)). Part of CZE, this has often been used on microchip format. Capillary electrochromatography (CEC) combines the advantages of HPLC and CE, such as the high efficiency of CE (movement of solutes by electrical forces) and the high selectivity of HPLC (chromatographic interactions). Analytes are separated by the combined action of partitioning an LC-type stationary phase and mobile phase and, if charged, by differences in electrophoretic mobility. In CEC the solutes are transported through the column by the EOF of the solvent and/or by their electrophoretic mobility. This approach will be examined in Section 5.6, since it involves specific technical challenges and deserves its own discussion. On the other hand, the theory and mechanisms of action for microchip CE are also based on fundamental knowledge gained in the development of conventional CE [11,16]. Two electrokinetic phenomena – electro-osmosis and electrophoresis – can contribute to the movement of species in a microchannel as a consequence of the application of an electric field between the ends of the channel. Electro-osmosis can be described as ‘the movement of liquid relative to a stationary charged surface,’ while electrophoresis is ‘the movement of a charged species (i.e. dissolved or suspended material) relative to a stationary liquid’ [11]. While electro-osmosis results in nonselective transport of charged or neutral molecules, electrophoresis is effective in separating ions of differing charge-to-mass ratios. The directed transport of molecules in response to an electric field is called migration. In most cases, the moving molecules are ionized in a polar solvent such as water. These electrically charged molecules experience a Coulomb force due to the electric field. This phenomenon is termed electrophoresis. The charged molecules first accelerate toward one of the electrodes, but then slow down and reach a terminal velocity, because they experience a drag caused by friction with the liquid. The Coulomb force (F) is given by: F ¼ qE;

ð5:1Þ

Miniaturized Systems for Analytical Separations II

173

where q is the charge on the molecule and E the strength of the electric field. Once the molecules reach their terminal velocity, the Coulomb force is balanced by a drag force, the Stokes force: F ¼ 6phrn;

ð5:2Þ

where Z is the viscosity of the liquid, r is the so-called hydrodynamic radius of the molecule, which indicates its size, and v is the velocity of the molecule. The terminal velocity of the molecule is reached when both forces are equal and opposite, so that: qE ¼ 6phrn

ð5:3Þ

From Equation 5.3, the terminal velocity is calculated as: n ¼ mE;

ð5:4Þ

where m is the mobility of the molecules, given by: m¼

qE ; 6phr

ð5:5Þ

where q is the charge of the molecule, E is the applied field strength (V/m), Z is the viscosity of the mobile phase and r is the radius of the molecule, which is related to its mass. On the other hand, electro-osmosis generates the EOF. EOF is a bulk solution flow phenomenon that occurs in capillaries filled with mild ionic solutions (typically <100 mM) when a voltage is established across the capillary, and in consequence it is also present in miniaturized systems. Although, as we have mentioned before, the theory and mechanisms of action for microchip CE are not new, some relevant key points should be made in a simple and clear way [16,17]. One important idea that we have to keep in mind is that miniaturization is highly desired when it adds new benefits without loss of those previously gained. In other words, the primary gain of miniaturization is enhanced analytical performance. In consequence, with the definitions given above, in the simplest mode (CZE) in CE separation of microchips, the velocity of a given analyte is determined by the mobility of the analyte plus the EOF, as shown in Equation 5.6: v ¼ ðme þ meof ÞE;

ð5:6Þ

where n is the velocity of the analyte, me is the electrophoretic mobility of the analyte, meof is the EOF and E is the field strength. Therefore, the electrophoretic flow is a constant for a specific set of run conditions for all molecules being separated. The electrophoretic mobility of the analyte molecules, therefore, is the most important factor for separation. The two most important factors influencing electrophoretic separations in zone electrophoresis are the analyte mass and the net charge, as was shown in Equation 5.5.

174

Miniaturization of Analytical Systems

Because the mass of a molecule is constant, the easiest way to influence electrophoretic separation is through changes in the mobile phase pH, which affect the charge on the molecules being separated. In chromatography and electrophoresis, analytical performance is normally judged through the analysis time t and the separation efficiency (expressed as the number of theoretical plates, N). Using CE as an example, the analysis time is given by Equation 5.7: t¼

L2 ; mV

ð5:7Þ

where L defines the distance between the point of injection and the point of detection, m is the vector sum of the electro-osmotic and electrophoretic mobilities of a given species (me þ meof) and V is the potential drop across the separation channel. Inspection of Equation 5.7 clearly demonstrates that a reduction in L will significantly reduce the total analysis time. On the other hand, the number of theoretical plates describes the ability of a separation system to discriminate between different species. Longitudinal diffusion (caused by intrachannel concentration gradients) and extrachannel band broadening due to sample injection and detection act in real systems to reduce this ability [16,17]. Consequently, N can be defined according to Equation 5.8: L2 N ¼ 2; ð5:8Þ s where s2 is the sum of the variance contributions of longitudinal diffusion, injection and detection. All three contributions to band broadening are unavoidable in a real system. However, injection and detection variances can usually be made much smaller than a quarter of the diffusional variance. Consequently, in a simple treatment it is valid to consider s2 as being solely due to diffusional processes. In this case, with an analysis time t, the band variance at the detector is given by: 2DL ; ð5:9Þ v where D (m2/s) is the diffusion coefficient of the molecular species and u is its total migration velocity (m/s). Substitution of Equation 5.9 into Equation 5.8 yields a more complete expression for N: s2 ¼

N¼

Lv LmE mV ¼ ¼ 2D 2D 2D

ð5:10Þ

This expression is significant since it states that although the separation efficiency increases as a function of applied voltage, there is no dependency on the channel dimensions. In other words, miniaturization should not directly influence separation efficiencies.

Miniaturized Systems for Analytical Separations II

175

A closer analysis of this situation demonstrates that this is not the whole truth. Although Equation 5.10 shows that N can be improved by increasing the potential drop across the separation channel, this cannot occur indefinitely. In practical applications, Joule heating effects impose a maximum limit on V. At high applied potentials, radial temperature gradients are produced within the channels. These act to increase radial diffusion and thus increase s2. The heat Q (W/m) generated per unit length in a rectangular channel of height h and aspect ratio a is given by: Q¼

V 2 h2 a lc; L2

ð5:11Þ

where l (m2/molO) is the molar conductivity of the buffer and c (mol/m3) its concentration. This expression shows that Q increases as a function of V, as expected [17]. However, a reduction in both h and a (especially h) will reduce Q and thus allow higher field strengths to be used (which will afford an increase in N, if desired). In fundamental terms, since heat is generated uniformly within the fluid volume element but can only be dissipated across the immediate boundaries of the electrophoresis channel, increasing the surface-to-volume ratio provides the dominant route to improved heat dissipation. Eijkel et al. [17] have extended the treatment above and determined the maximum rate of plate generation as:   N 1 m2 ¼ 2 lc ð5:12Þ t max h plcD This reinforces the idea that the maximum rate of plate generation can be dramatically increased by decreasing channel heights. Interestingly, it has been noticed that the rate of plate generation is independent of channel width, implying that wide channels can be used without any loss of performance. In conclusion, this simple treatment shows that it is possible to summarize the enhancement of analytical performance expected in miniaturized CE systems thus: (i) analysis times can be reduced through a reduction in channel length or an increase in separation voltage; (ii) plate numbers can be increased through an increase in separation voltage; and (iii) heat generation can be reduced by reducing channel heights.

5.4

Microfabrication Techniques

An accurate and in-depth description of microfabrication techniques is obviously outside the scope of this chapter, but readers can find more information in [18–24]. CE microchips are mainly fabricated using various glass substrates, from inexpensive soda lime glass to high-quality quartz. Glass substrates are the most

176

Miniaturization of Analytical Systems

common because of their good optical properties (optical transparency, which allows optical detection and visual inspection), resistance to many chemicals, well-understood surface characteristics (exhibiting an EOF close to fused silica 9.5 · 104 cm2/Vs at pH >9), dielectric properties (the possibility of working with the high voltages used in CE) and well-developed microfabrication methods (adapted from the silicon microfabrication industry). Other advantages of glass are its hardness, high thermal stability and biocompatibility (wide range of applications: DNA separations, enzyme/immunoassays, cell biology and small molecules). Polymeric substrates have mainly consisted of poly(methylmethacrylate) (PMMA) (by injection moulding or hot embossing) and poly(dimethylsiloxane) (PDMS) (by casting). These materials have found favour due to their ease of fabrication. PMMA is an example of a thermoplastic material (because a wide range of microfabrication methods are available for this compound), and PDMS is an example of an elastomer. PDMS has been widely discussed due to several characteristics: (i) it is suitable for optical detection down to 320 nm; (ii) it cures at low temperatures and moulding can easily be replicated through the process of prototyping, master formation and soft lithography; (iii) it can seal reversible to itself and other materials by van der Waals contact with a clear smooth surface at room temperature; (iv) its surface chemistry can be controlled to form EOF using a plasma technique. Polymers also show good resistance to chemical treatment and good biocompatibility. It is possible to find a polymer that has the optical qualities desired for any given application (e.g. PDMS is transparent to the UV region of the electromagnetic spectrum, while most thermoplastic polymers, such as PMMA and PC (polycarbonate), are transparent in the visible region). The large EOF in glass is partly due to the high surface charges because of the deprotonation of silanol groups. However, many polymers do not contain ionizable functional groups and thus would be expected to produce much smaller EOFs, a clear disadvantage. Indeed, under controlled microfabrication techniques, it is possible to generate an EOF very acceptable for PMMA, of 1.2 · 104 cm2/Vs. Although native PDMS exhibits a very small EOF value, subjecting it to oxidation in a plasma discharge changes the surface characteristics. OSi(CH3)2O groups are transformed to OSi(OH)4n producing SiOH groups, which leads to surfaces that can be readily deprotonated as glass or fused silica. Also, polymer microchips are of increasing interest because their potentially low manufacturing costs may allow them to be disposable. As PDMS is transparent in the visible light range, LIF detection is readily applicable. Among the few disadvantages of PDMS are poorly defined EOF and adsorption of nonpolar hydrophobic samples. A comparison between microchip materials is given in Table 5.2 (adapted from [21]).

Miniaturized Systems for Analytical Separations II

177

Table 5.2 Comparison of materials in CE microchip. (Adapted from [21]) Feature

Glass/Fused silica

Polymera

Electronic conductivity Thermal conductivity (cal/cm-s- C) Bioassay compatibility

none: OK for CE 2 · 103

none: OK for CE 4.5 · 104b

fair

Optical detection

glass: very good fused silica: excellent isotropic wet etching only

fair to very good (varies according to polymer choice) poor to very good (varies according to polymer choice) Si or glass mastering plus replication techniques; direct methods (ablation, dry etching) dependent on mastering and replication methods: <0.2–20 inexpensive for large numbers of devices

Microfabrication

Feature aspect ratio (depth/width of microchannels) Cost a b

<0.5 moderately expensive

acrylics, polycarbonates, polyolefins and polydimethylsiloxane have been most utilized and studied value for polymethylmethacrylate

5.4.1

Microfabrication of Glass CE Microchips

Structures on glass substrates are usually generated using standard photolithographic technologies [20,22,23]. Figure 5.4 presents an example diagram of such a procedure. Briefly, first a layer of photoresist is spin-coated on top of the sacrificial mask layer (Cr/Al) (deposition). Photoresist is a polymer that becomes soluble (positive) or insoluble (negative) in developer solutions after exposure to light (exposition light). In the next step the photoresist spun on the microchip is exposed in the region defined by a photomask, typically using an aligner. The photomask is a plate with a user-designed pattern that is transparent while the background is opaque to the exposition light and vice versa. After the microchip is baked to harden unexposed resist, the exposed photoresist is dissolved with a developer solution (development). The sacrificial mask layer (Cr/Al) of the exposed region is removed using the appropriate etchants (sacrificial layer removing). During this time, the sacrificial layer underneath the unexposed photoresist remains intact. After pattern transfer and development, the portions of the microchip that are to be etched are unmasked and are now ready for chemical etching (glass etching). HF is used as the primary etchant and can be prepared in various solutions including HF/ NH4F, HF/HNO3 and concentrated HF. The etching rate of HF for glass is readily controllable if the temperature is controlled. Following etching of the microchannels, the photoresist and sacrificial mask layer are stripped (stripping) and the

178

Miniaturization of Analytical Systems

Figure 5.4 Microfabrication of glass microchips. (Reprinted from [11] with permission of Wiley-VCH)

access holes drilled (not shown in Figure 5.4). The access holes can be drilled on the etched substrate or on another blank glass wafer. When holes are drilled on the etched substrate, aligning two substrates for bonding is much easier. Finally, the substrate is bonded to another piece of substrate to form a finished microchip (bonding). 5.4.2

Microfabrication of Polymer CE Microchips

When a microchip substrate is a polymer, the two major methods of machining are replication from a master (moulding methods) and direct machining [19,21,24]. Replication methods often produce a microstructure by allowing a polymer workpiece to form an inverse copy ofa mould. These methods are: hotembossing, injection moulding and casting. The formation of microchannels and other structures using moulding methods generally involves two primary steps: (i) moulder (also known as

Miniaturized Systems for Analytical Separations II

179

master) fabrication; and (ii) channel pattern transfer from the moulders to polymer substrates. These methods have in common that a very soft or even liquid form of polymer is poured or pressed into a mould, after which the material is hardened and removed from the mould. Direct machining methods on the other hand remove small amounts of polymer in places where the microstructures (microchannels, reservoirs) should be located. This type of micromachining involves laser ablation. Hot embossing is the process of pressing a mould into heated, softened thermoplastic, followed by cooling, producing an inverted replicate of the mould. The first step in hot embossing consists of heating the mould and the polymer to glass transition temperature. Once the polymer begins to soften it takes on the form and shape of the mould. The mould and the polymer are then cooled below the glass transition temperature to harden the polymer, and afterwards the polymer is removed from the mould. Embossing can take several minutes per device and can be a useful tool in rapid prototyping devices. In the injection moulding technique the molten polymer material is injected into the mould. For microinjection moulding, the mould is heated to the softening temperature of the polymer to prevent the injected polymer material from hardening too early. After injection of the molten polymer into the mould, it is slowly cooled, so that the polymer hardens and the polymer microstructure can be removed from the mould. Injection moulding allows very high-throughput production with low production costs (moulded devices are released every 5–10 seconds). On the other hand, laser micromachining (such as photoablation) is a direct machining method based on the removal of polymer material by using intense UVor infrared radiation provided by a laser. The photoablation process involves absorption of a short-wavelength laser pulse to break covalent bonds in long-chain polymer molecules with production of a shock wave that ejects decomposed polymer fragments [18]. Many commercially available polymers can be photoablated, including polycarbonate, poly(methylmethacrylate) (PMMA), polystyrene, nitrocellulose and poly(tetrafluoroethylene). The laser energy can be specially patterned by using a mask with subsequent generation of microcavities and channels in various geometries or by controlling the position of the laser with x–y stages. The resulting structures are generally characterized as having little thermal damage, straight vertical walls and well-defined depth. However, laser ablation does not lend itself well to mass production. A particular example of polymer machining of great significance is the PDMS microchip [25]. PDMS is a widely used polymer substrate for soft lithographybased microfabrication and nowadays has a relevant role in the miniaturization scene not only as microchip material but as ideal material for building lab-on-a-chip systems (soft lithography), as we will discuss in Chapter 8. Apart from the features described above, the major advantage of this material is the ability to rapidly prototype complex structures.

180

Miniaturization of Analytical Systems

Casting uses a chemical process to harden the polymer [18]. Two components, a base and a hardener or cure, are mixed just prior to use. Immediately upon mixing, the chemical curing process starts. After some time, this process results in hardening of the polymer (at atmospheric pressure and temperature). The liquid mixture is poured in the mould, and as the polymer sets it takes the shape of the mould. The polymer structure can then be removed from the mould. This technique is very popular with elastomers such as PDMS because they form hermetic, reversible seals and smooth surfaces like glass or silicon by adhesion. Casting is the simplest of the three moulding processes, but requires contact with the mould for minutes or hours. A schematic representation of this type of replica moulding is shown in Figure 5.5 [25]. Channels in PDMS are easily formed by replica moulding, which is simply the generation of a negative replica of the master in PDMS by pouring the prepolymer over the master and curing it at atmospheric pressure at somewhat elevated temperatures. In the last step, the PDMS replica containing the channel network is sealed irreversibly with a second flat slab of PDMS. In the example shown, this is due to oxidation of both surfaces in an oxygen plasma discharge.

Figure 5.5 Microfabrication of PDMS microchips. (Reprinted from [25] with permission of ACS, Copyright 1998)

Miniaturized Systems for Analytical Separations II

181

Oxidation results in a stronger seal than untreated PDMS, and it seems to be easier to fill oxidized PDMS channels with liquids. As explained above, higher cathodic EOFs were generated, which is consistent with a more negatively charged surface of PDMS after oxidation. However, instead of irreversible oxidation, glass plates are usually used to seal the channels, which also physically support certain integrated elements like electrodes. When the PDMS and glass parts are ready and cleaned with methanol, the two surfaces are brought into contact by a reversible seal. For irreversible bonding, as described in the example above, the two surfaces are placed under plasma and brought into contact after appropriate surface activation.

5.5

Basic Fluidic Manipulation/Motivation: Electrokinetic Injection and Separation Protocols

Basically, two fluidic manipulations have to be used in CE microchips: injection of a defined plug of sample and analyte separation. Without question, a very important fluidic handling function is the ability to dispense very well-defined and small volumes of solutions (in order to keep selectivity constant) with reproducibility. In particular, the injection of a well-defined, reproducible injection plug into the separation channel is of paramount importance to finding the right tradeoff between separation efficiency and detection sensitivity. Also, the reduced injection volumes used in microchip format (pl) help maintain resolution. These small volumes are very important in many analytical applications. As mentioned in Section 5.2 (see Figure 5.1), the integrated injectors are usually either cross-channel pieces, formed by orthogonally intersecting the separation channel with a channel connecting the sample to waste (T injectors) (Figure 5.1(a)), or twin-T injectors (Figure 5.1(b)), where the two arms of the sample-to-waste channel are offset to form a large injector region. Integrated sample injectors permit volume-defined electrokinetic sample injection of short injection plugs with reproducibility [27,28]. The type of injection employed depends on the sample matrix, the precision required and the number of voltage sources available. Three different established injection methodologies have been employed with CE microchips: no-pinched, pinched and gated injections. In each of these approaches the detection reservoir is held at constant ground. The easiest injection procedure is the no-pinched approach, which is shown in Figure 5.6(a) (in general, it is the strategy normally used). This principle can be carried out using just a single power supply. A high voltage is applied to the sample reservoir for a short time, with the electrochemical reservoir held at ground (injection). Sample is introduced directly into the separation channel by electrokinetic injection. After the injection is completed, the high voltage is switched back to the buffer reservoir and the separation is initiated (run/separation). This approach

182

Miniaturization of Analytical Systems

Figure 5.6 Electrokinetic injection protocols. (Reprinted from [11] with permission of Wiley-VCH)

does not use pushback voltages to keep sample in the main separation channel and can result in irreproducible injections and large plugs. Pinched injections require voltage control of each reservoir (Figure 5.6(b)). The sample reservoir is placed at the end of one of the side arms, and prior to injection the sample is continuously pumped across the intersection to the waste reservoir (load). Additionally, the voltages are arranged so that there is also flow

Miniaturized Systems for Analytical Separations II

183

from the buffer reservoir and the reservoir at the end of the main channel (electrochemical reservoir) toward the intersection and on to the sample waste reservoir. This is to avoid premature bleeding of the sample into the main channel. By adjustment of the voltages (flow rates), one can ‘pinch’ the sample stream more or less on its passage through the intersection (inject and run/separation). This is why this technique is called ‘pinched injection’. Figure 5.6(c) shows the principle of gated injection, which also requires voltage control over each reservoir, but with another configuration. Before injection (preinjection), the main waste reservoir at the end of the separation channel (electrochemical reservoir) is set to ground. Thus, it functions as the anode to which the EOF is directed. All other reservoirs are at higher positive potentials. The topmost reservoir contains the sample and, to ensure pumping out of the reservoir, the potential is set to a somewhat large value. For the same reason, the sample waste reservoir is set to a relatively low voltage to receive the sample flow. The remaining reservoir is filled with a buffer and set at a slightly higher voltage than the sample reservoir. This allows flow out of this reservoir, to keep the sample from entering the main channel by flowing right across the junction and into the sample waste reservoir. At the same time, this reservoir also provides a flow of fresh buffer into the main channel. The simplest way to facilitate injection is to remove the power supplies at the sample waste and the buffer reservoir from the electric circuit (injection step). Voltage is still applied at the other reservoirs, allowing the sample to flow into the main channel. The injection step is finished as soon as the original voltage distribution is established again, thereby cutting off the sample flow into the main channel and releasing a plug of sample flowing to the anode (run and separation steps). This procedure is known as ‘gated injection’ because the injected volume is mainly determined by the timing of the injection sequence and the sample can enter the main channel for as long as this electrokinetic valve is open. In this scheme, the injection is biased, meaning that during the time of injection, positively charged ions are injected to a larger extent than neutral species, which are injected to a larger extent than negatively charged ions. This fact needs to be taken into consideration when designing a method for quantitative analysis involving one or more gated injection steps. The main differences between these three procedures are, first, in pinched injection volume is predefined and fixed, and second, in pinched injection if sufficient time is given to the loading step, the composition of the sample solution present at the intersection is the same as the composition of the original sample solution, and consequently there is no electrokinetic bias upon injection. Nopinched injection becomes the easiest but least reproducible scheme, although several reports include good relative standard deviations (less than 5%). In this way, the robustness of the commercially available simple cross-integrated injector of glass microchips has been carefully studied [29]. Intra-injector precision was defined as the relative standard deviation obtained in the repetitive analysis carried

184

Miniaturization of Analytical Systems

out using the same injector, and inter-injector precision as the relative standard deviation obtained in the repetitive analysis carried out using both injectors. Very good results with respect to migration time, peak height and peak area were obtained with RSD values at less than 4% when just one injector was used (intra-injector precision). When the results obtained from both injectors (inter-injector runs) were compared, differences of less than 7% were obtained. Therefore, the results obtained indicate good reproducibility between injectors and open the possibility of working indiscriminately and/or sequentially with both injectors, at least when the same sample is deposited in both reservoirs.

5.6 Electrochromatography in Microchip Format: Designs and Applications Although microchip zone electrophoresis (micro-CZE) offers high separation efficiency due to short separation channels, there is a major need for an additional separation mechanism that would manipulate selectivity in the chip-based separation systems. Such selectivity manipulation can be achieved by microchip capillary electrochromatography (micro-CEC). As stated above, this approach requires a special discussion since it involves technical challenges absent in the other electrophoretic modes. CEC is an electric field-mediated separation technique in which the liquid flow is generated by the electric field itself toward a stationary phase. As we will discuss in the next section, the main advantage of using electric field over pressure for flow generation is the flat flow profile of the EOF; thus, CEC is one of the best candidates for the construction of novel and high-efficiency microanalytical devices. Microchip-based electrochromatography is a hybrid method of micro-CZE and chip-based liquid chromatography (micro-LC) and combines the best characteristics of both methods. As has been widely discussed in Chapter 4 and in sections of the current chapter, the separation mechanism in micro-CZE is based on the differences between mobilities of solutes, while in micro-LC it is based on the differences in partition coefficients between two phases. Combining micro-CZE with micro-LC creates a powerful analytical tool capable of separating both ionic and neutral compounds. Since 1994, when the first micro-CEC work was published by Ramsey and coworkers [30], microchip electrochromatography has received considerable attention [31,32]. As expected, the stationary phases developed for conventional HPLC applications were the bases for the generation of novel micro- and nanocapillary columns, and the same preparation protocols and methods were applied for microchip-based systems. One of the most difficult tasks is to fill and retain the packing material in the small-diameter channels. As Pumera clearly states [31], the main modes of micro-CEC are (i) open-channel electrochromatography, in which

Miniaturized Systems for Analytical Separations II

185

the walls of microchannel are coated with stationary phase; (ii) packed-channel electrochromatography, in which microchannels are filled with typical RP–HPLC packing materials; and (iii) continuous rod electrochromatography, in which channels contain either monolithic stationary phase prepared by in situ polymerization or monoliths directly microfabricated using lithographic methods. The contribution of resistance to mass transfer decreases as the dimensions of the channel decrease, as predicted by theory. This principle is greatly exploited in openchannel electrochromatography. In this strategy the walls of a microchip channel are coated with the chosen selector. Different distribution equilibria between the running buffer and coated stationary phase are responsible for tailoring the selectivity and improved resolution of the solutes. As the diameter of the open channel decreases, the resistance to mass transfer decreases as well, but other problems arise. For example, the smaller the surface of the modified channel, the more problems there are with the separation channel overloading. Also, injection and detection volumes must be scaled down proportionally to prevent excessive contributions to band broadening and therefore may fall below the limit of detection. To minimize these problems, channels with a high aspect ratio of channel width to channel height are fabricated. This allows improved mass transfer in one dimension, while the sample load and detection volume are maintained in the other [30]. C18 and C8 (most popular), gold nanoparticles (AuNPs) and bulk microchip material modifications are the main strategies used in this micro-CEC approach. The chemistry of the octadecyl (C18) group bonding on a glass surface is well documented. The ability to design and fabricate chip manifolds with low-volume connections offers great possibilities for precise on-chip mixing of mobile phase and therefore opens many possibilities for not only isocratic but also high efficient gradient elution. On-chip mixing is a great advantage of microchip-based systems over conventional CE/CEC, where the gradient elution is a very complicated issue [33]. The use of AuNPs in conjunction with microchip-based electrophoresis to improve the selectivities between solutes and to increase the efficiency of the separation has also been reported [34]. The authors of [34] coated the glass wall of a channel in a microfluidic device with a layer of poly(diallyldimethylammonium chloride), and 10 nm citrate-stabilized AuNPs were collected on this channel. In contrast to the surface modifications of glass- or polymer-based devices, the incorporation of the appropriate selector molecules into the whole bulk material of microchip by copolymerization has also been reported [35,36]. Planar microchip format facilitates fully automated repetitive loading and removal of stationary phase in packed-channel electrochromatography. Since EOF velocity is independent of particle size, it is possible to use stationary phase with small diameters down to 1.5 mm without experiencing the large pressure drop typical for LC. Further lowering mass transfer resistance in open-channel electrochromatography requires a reduction of the dimensions of its channels.

186

Miniaturization of Analytical Systems

However, there are considerable problems with using open channels with widths of 3 mm or less, such as difficult manufacture, clogging and low detection volume [37]. In contrast, packed beds have the benefit of providing low mobile phase mass transfer resistance and a wide variety of stationary phases. Furthermore, packed-channel electrochromatography can utilize commercially prepared, reproducible stationary phase materials developed for HPLC. Stationary phase is usually held in position by frits or weirs. Both glass and polymer have been used as microchip material in this approach. Different physical structures such as stationary phases have been considered for packed-channel microchip electrochromatography with electrokinetically loaded and removed beads [38], as shown in Figure 5.7(a). PDMS microchips with built-in integrated hexagonal structures – which serve as inlet and outlet frits [39] allowing the entrapment of 5 mm C18 beads, and which are introduced by a special side-loading channel using pressure applied by syringe as well as by a fritless device [40] using particles with diameter 3 mm – were spatially retained and packed without the need for frit structures [41] in the tapered geometry at the beginning of the PDMS channel. Microspheres covered with C18 groups were loaded using a vacuum and stopped at the taper, but the simple use of a vacuum was not sufficient to make them stable for micro-CEC separation.

Figure 5.7 Relevant examples of CEC on microchip. (a) Packed-channel microchip electrochromatography with electrokinetically loaded and removed beads: (i) at an initial stage of electrokinetic packing; (ii) after it is completely filled with beads. (Reprinted from [38] with permission of ACS, Copyright 2000); (b) Schematic of the polymer-based monolithic on microchip. The inset shows a scanning electron micrograph of a channel cross-section filled with photoinitiated acrylate polymer monolith. The mean pore diameter is 1 mm. (Reprinted from [42] with permission of ACS, Copyright 2002); (c) Scheme of a typical monolithic approach separation column with detailed inlet section. The whole column was microfabricated using PDMS substrate. (Reprinted from [43] with permission from Elsevier)

Miniaturized Systems for Analytical Separations II

187

Achieving the uniform packing of microchannels with beads can be a very challenging task in packed electrochromatography. Problems in channel packing can lead to irregularities in the packed bed and in poor system efficiency. Peak broadening can be caused by the inlet frit. Moreover, frits contribute to the formation of bubbles. These problems can be overcome by preparing a microchannel consisting of a block of porous solid, called a monolith or rod, manufactured by in situ polymerization or by a microfabrication technique. Both monoliths prepared by in situ polymerization [42] and monoliths prepared by microfabrication [43] appear to be common (see Figure 5.7(b),(c)). Indeed, in monolithic microCEC, photoinitiated polymerization permits the creation of a stationary phase at the specific places where a microfluidic network allows tailored applications. The inherent method of microfabrication of a microfluidic network by photolithography opens new avenues for simultaneous microfabrication of built-in monolithic columns. In the first approach, the preparation of the continuous bed is relatively easy, since a monomer solution with low viscosity is simply pressed into the channels. This way, highly efficient columns can be fabricated without the limitations of packed-channel electrochromatography. Monomer mixture composition allows an easy manipulation of separation parameters such as hydrophobicity, pore size and charge. Monolithic separation phases for micro-CEC can be manufactured in two different ways: by a chemical initiation, which results in a microchip with channel continuously filled with polymer; or by photopolymerization, which allows the creation of a polymer bed at specific places in the microfluidic network. In the second approach, the stationary phase support structure dimensions are microfabricated and then introduced into the microsystems. As we have already indicated, Figure 5.7 shows relevant and representative examples of miniaturized electrochromatography approaches.

5.7

Comparison of Hydrodynamic and Electroosmotic Flow-driven Miniaturized Systems

In the miniaturization scene, EOF-driven systems are clearly dominant over hydrodynamic ones. The reason for this dominance lies partially in the inherent simplicity of their fabrication and operation, combined with unique features with respect to separation speed, sample injection and reagent consumption. In consequence, pumping by electro-osmosis is a unique and widely applied principle that has deserved and deserves special attention. It is obviously possible to create a flow (induce pumping) by filling a microchannel with a buffer solution and applying suitable voltage at the channel ends. This relatively easy method is clearly an advantage of electro-osmotic pumping,

188

Miniaturization of Analytical Systems

Figure 5.8 Comparison between hydrodynamic and electro-osmotic flows. (Reprinted from [44] reproduced with permission from Wiley VCH)

especially compared to the effort it takes to fabricate and run microfabricated pumps and valves. However, the main advantage of using electric field over pressure for flow generation is the flat flow profile of the EOF, which generates high-efficiency separations. In contrast to hydrodynamic flows, where one finds a parabolic distribution of the flow velocities, with the largest velocity at the centre of the channel and zero velocity at the walls, EOF is generated close to the wall and therefore produces a plug-like profile with a very uniform velocity distribution across the entire cross-section of the channel (Figure 5.8, from reference [44]). Another advantage of EOF systems is in CZE: due to the absence of other interaction equilibria, and ignoring contributions from the injection plug width and detector geometry, the only source of dispersion is diffusion, and the effects of diffusion can be minimized by decreasing time from injection. The main disadvantage of EOF is its strong dependence on the chemistry of the system, which is of course a consequence of the strong dependency of the zeta potential on the chemical state of the system. Indeed, zeta potential formation is a characteristic feature of solid (channel wall)–liquid (mobile phase) interfaces. The composition and chemical characteristics of the two phases play an important role. The buffer pH is the most sensitive parameter in aqueous mobile phases, due to the fact that electrolysis processes take place in the microchip reservoirs. This dependency makes EOF hard to control: every change in pH, dielectric constant, concentration and so on due to reactions or the mixing process has an immediate effect on the magnitude of the EOF. On the other hand, this strong dependency also means that there are many parameters that can be exploited in order to control and manipulate EOF. Table 5.3 summarizes the key points of this comparison.

Miniaturized Systems for Analytical Separations II

189

Table 5.3 Comparison between hydrodynamic and electro-osmotic flows Feature

Hydrodynamic approach

Electro-osmotic approach

Driving force Stationary phase Sample consumption Sample injection

gradient of pressure always low

gradient of voltage sometimes extremely low

hydrodynamic sample volume fixed by loop

Separation profile Technology aspects

parabolic microfabrication of valves and pumps required. highly sophisticated technologies required

Facilities

cleanroom and other facilities required dead volumes not equal to zero high reproducibility

electrokinetic sample volume easily changed toward voltage (versatile/ variable loop) quasi-flat free pumping without any moving parts (valveless microsystems) higher voltages can be used for the separation, and the dissipation of heat is increased compared to macroscopic systems due to the higher surface-to-volume ratio cleanroom facilities required for glass microchips interfaces reservoirs. dead volumes equal to zero less reproducibility strong depen dence of chemistry

Connections/ interfaces Analytical aspects

5.8

Analytical Applications

Two important points must be made regarding this section. First, the volume of research found on the analytical applications of CE microchips is outside the scope of this section; if you wish to know more, we invite you to read several specific books [10–12,18], chapters [15] and reviews [3,4,45]. Second, the borders between applications using CE microchips and applications belonging to the lab-on-a-chip domain are sometimes blurred. As the fundamentals of miniaturized separation and detection systems are studied in depth in Chapter 8, the applications discussed in this section tend to fall into the lab-on-a-chip domain. We have selected the most representative applications in terms of their ‘historical significance’, focusing mainly on the bioanalytical field (using laser-induced fluorescence (LIF) as the detection route; see Chapter 6). A brief overview will be given in three broad sections: DNA, protein and small-molecule analysis. Although there is a great deal of literature dealing with these applications, overviews are scarce [45]. The separation of DNA for both sequencing and fragment analysis has taken on an important role in biological research, clinical measurement and forensics.

190

Miniaturization of Analytical Systems

Arguably, one of the most successful application areas for microchip CE has been DNA separation. The second major area of application is protein analysis. There is a great need to develop new analytical methodologies for the measurement of proteins in clinical research, pharmaceutical development and to better understand fundamental biological processes. With the sequencing of the human genome, it is anticipated that thousands of new proteins will be discovered and/or synthesized. Traditional methods are often too slow to be able to handle the high-throughput analysis that will be required with this number of analytes. The last major area of development in microchip CE has been the analysis of small molecules. Reviews on amino acids [46], explosives [47], environment [48,49] and food [50,51] have recently appeared in the literature. 5.8.1

DNA Analysis

Since DNA is the blueprint for life in all its forms, it is no wonder that nucleic acids are of such interest to scientists. The Human Genome Project, completed in 2003, produced a wealth of information and some unexpected results, such as the very small percentage of actual coding sequences (1–2%) and the small number of genes (25 000), together with the fact that almost 50% of the genome consists of repetitive sequences which do not possess any known physiological functions [52]. On the other hand, sequence information on DNA sections immediately surrounding the genes has an apparent role in the determination of when and how many copies of transcripts/proteins are generated, referred to as gene expression regulation. Even a slight change in sequence can influence the structure and function of the corresponding expressed proteins. In this sense, it is becoming increasingly evident that a wide range of diseases and disorders are associated with local abnormalities or mutations in the sequence of nucleotides encoded in an organism’s DNA [53]. It is clear that one of the most important issues in medicine today is the analysis of the differences between individual DNA sequences. These differences are called polymorphisms. Genetic variance between two unrelated persons is on average 0.1%. The most frequent polymorphisms are single-nucleotide polymorphisms (SNPs), which involve the substitution, deletion or insertion of a single nucleotide in the genomic sequence [54]. The number of SNPs is estimated to be in the range of 3–10 million in the human genome [55]. There is no doubt that the ability to detect these sequence variations in a rapid and inexpensive manner using small sample and reagent volumes offers the potential to catalyse a revolution in medicine by identifying the pathways through which genomics can lead to disease [56,57]. Electrophoresis (both conventional and capillary) is the most widely used method for the analysis of genetic polymorphisms. The recently developed microfluidic electrophoresis devices offer several

Miniaturized Systems for Analytical Separations II

191

advantages, including higher separation speed and, most importantly, the capability for system integration and multiplexing. Completion of the Human Genome Project was the first challenge for microchip CE separations. The reduction in the time and cost (lowering consumption of sample and reagent while maintaining high-resolution separations) of DNA analysis was the greatest contribution of this planar microfluidic electrophoretic platform with respect to capillaries and slab gels. In the post-genomic era, DNA separationbased technologies must attain certain performance characteristics to solve new targeting applications, essentially focused on clinical diagnostics, forensics applications and small-scale sequencing of individual genomes for personalized medicine initiatives. This means development of low-cost instrumentation that can be easily operated by novice users, automation of the entire DNA processing pipeline, preferably into a single instrument, and disposable fluid-handling devices that can be mass-produced at low cost; all of this potentially addressed by microfluidic chips [58]. DNA analysis is a very complex topic to try to describe in a general book on miniaturization. The literature is enormous, and many disagreements can be found. In this section, we will limit ourselves to providing a description of the main focuses of DNA analysis: genotyping applications for clinical diagnostics, DNA forensics and DNA sequencing. An overview of the grade of integration of DNA analysis on to the microfluidic platform will be given in Chapter 8. In this matter especially, the terms ‘CE microchips’ and ‘microfluidic devices’ mean almost the same thing. Excellent literature can be found for DNA analysis. The pioneer work of Mathies’ group and other book chapters [60,61] and review papers [52,53,58,62–64] have been published on various applications of microchips. In addition, many companies (such as Caliper, Aclara, Shimadzu, Gyros, Cepheid, Fluidigm, Micronics and Agilent) have commercialized microchip technology, and a few products are already commercially available. Finally, DNA separation is performed in a manner very similar to traditional CGE experiments. The same gel matrices are often used and detection is achieved using LIF. Rapid, high-resolution on-chip separations have been demonstrated for oligonucleotides, RNA and DNA fragments, and in genotyping and sequencing applications. A variety of sieving media developed for CE, such as linear polyacrylamide and its derivatives (e.g. polydimethylacrylamide), polyethylene oxide, polyvinylpyrrolidone (PVP), polyethylene glycols, polysaccharides and cellulose derivatives, have been successfully adapted to microchip electrophoresis of nucleic acids. Genotyping As has already been mentioned, identification of genetic sequence variations in the population is key to unravelling the complex interactions that are the underlying

192

Miniaturization of Analytical Systems

causes of many medical disorders and diseases. Therefore, the need is evident for advanced technology to perform mutation DNA analysis (genotyping) in a less expensive and higher throughput way than is currently attainable. There are many different techniques currently in use for mutation analysis in the capillary format; those which have been adapted to the microfluidic platform will be presented in this chapter. Genotyping consists in identifying differences (polymorphisms) in the sequence of nucleotides within various regions of the genome, which are produced by either environmental or inherited conditions. Broadly, it is possible to distinguish between scanning strategies, which are employed to identify unknown variations within larger regions of known sequences (the locus of the polymorphism is not known), and diagnostic (detection) strategies, which are used to analyse specific variations existing at known locations (the locus is known within a certain gene). Scanning strategies Common scanning techniques that are used in conjunction with electrophoresis for genotyping include single-strand conformation polymorphism (SSCP), denaturing or temperature gradient gel electrophoresis (DGGE, TGGE) and heteroduplex analysis (HDA). These different techniques are going to be briefly explained and some relevant examples from the literaturewill be presented. It is important to remark that all of them follow similar steps in the analysis profile. After polymerase chain reaction (PCR) amplification of a region of interest, depending on the specific methodology, mutations or differences in sequence imply principally conformational perturbations or changes in the melting properties, which mean alteration of the mobility in the electrophoretical channels and allow the mutant allele to be distinguished from the ‘wild’ fragments. It is important to emphasize that those examples where sample preparation and electrophoresis sorting are integrated in the same microdevice will be discussed in Chapter 8. Single-strand conformational polymorphism is a simple method that combines DNA amplification of the specific fragments by PCR, denaturation of the doublestranded amplicons to generate single-stranded DNA and analysis of the denatured fragments by electrophoresis. Single-stranded DNA fragments, under the appropriate experimental conditions, possess sequence-specific 3D conformations, where just a single-base mutation will alter the mobility of the DNA fragments and allow them to be distinguished. Variables including fragment length, type, extent of mutation and guanine–cytosine content all influence the sensitivity of this technique. Samples subjected to mutation scanning surveys using SSCP typically include the mutagenic DNA as well as their known wild-type counterparts from a PCR-defined region. As such, the electrophoretic technique must be able to afford high-resolution capabilities to observe the presence of the mutation in a high level of background in ‘normal’ or wild-type DNA sequences [65–70]. Despite the shorter separation times typically afforded by microchips, a single-channel microdevice cannot match the throughput of multiple-lane CAE

Miniaturized Systems for Analytical Separations II

193

instruments, which are now capable of simultaneously separating hundreds of samples. Fortunately, the micromachined format is well suited to forming arrays of separation channels, because reservoir spacing can be designed for automated parallel sample loading and multiple separation lanes can be bundled together to facilitate detection. The work of Tian et al. [68] demonstrated the ability to perform a high-throughput SSCP analysis of genes associated with hemochromatosis and hypertrophic cardiomyiopathy (heart muscle abnormality) using a 48 lane section of a 384 lane glass-based MCE device. A total of 21 SNPs from HFE, MYL2, MYL3 and MHY7 genes were fully discriminated. After off-chip PCR amplification using fluorescently labelled primers, 1 ml of each product was loaded in a high-density microchip contained 8 cm long straight channels that were radially configured about a centred common anode. Separations of mutant alleles from their wild-type counterparts could be performed in less than 400 seconds using 5% polydimethylacrylamide, 10% glycerol and 15 % urea as sieving material. Denaturing gradient gel electrophoresis and temperature gradient gel electrophoresis are two analogous techniques used to discriminate heteroduplexes containing mismatched base pairs from fully matched homoduplexes. They rely on sequence-dependence associates with the melting behaviour of double-stranded DNA fragments. Variations in temperature (TGGE) or chemical denaturant conditions within the sieving gel (DGGE) in a gradient format during the migration of solutes through the separation channel make even a single-base sequence change to modify the conformational stability of double-stranded DNA and as such perturb electrophoretic mobility. From a practical point of view, gels possessing identical gradient properties are difficult to achieve in microseparation formats, while TGGE has proven to be more amenable in microchip CE, where an internal tapered microheater allows spatial and temporal temperature gradients along the lengths of the separation channels [71,72]. Heteroduplex analysis is based on detecting conformational differences in the migration behaviour of DNA fragments during gel electrophoresis. Heteroduplexes form as a combination of wild and mutant single-strand DNA fragments, incorporating conformational perturbations which induce detectable changes in electophoretic migrations with respect to homoduplexes (wild:wild or mutant:mutant combinations). Heteroduplexes contain bulges within the rehybridized DNA fragments where bases are mismatched. With appropriate separation parameters, heteroduplexes migrate slower than corresponding homoduplexes. This method is often used for DNA species containing insertions or deletions, which typically yield more pronounced mobility shifts due to significant bulges resulting from mismatched base pairs in duplexed DNA [73–76]. Diagnostic strategies As previously mentioned, these methodologies are geared toward detection of the specific location of genomic variations. The assays

194

Miniaturization of Analytical Systems

associated with the identification of specific point mutations include allele-specific PCR (AS-PCR), ligase detection reaction (LDR), single-base extension (SBE) and restriction fragment length polymorphism (RFLP). Allele-specific PCR is a DNA hybridization-based technique that makes use of primers specific to the locus or loci being interrogated. In this case, an allele-specific primer hybridizes to the SNP site, then undergoes amplification using a DNApolymerase enzyme. Amplification only occurs if the primer perfectly matches the template. Different electrophoretical microchips, including some made of glass or polymeric materials, are reported for AS-PCR mutation analysis [77–82]. Ligase detection reaction employs a mixture of two complementary detecting oligonucleotide primers designed to hybridize a polymorphic target DNA strand, such that they are separated by a single-base mutation site. Only if both oligos are perfectly matched to the target will they be joined by a highly specific and thermally stable DNA ligase enzyme. After a series of repeated thermally activated ligation cycles, the product concentration is increased to a level that can be detected by gel electrophoresis. Often, the LDR assay is coupled to a primary PCR to amplify the gene of interest. In this sense, PCR can be used to generate products for subsequent interrogation using microchip CE [83–86]. Single-base extension, also known as minisequencing or primers-guided nucleotide incorporation, is based on the activity of DNA polymerase. In this technique, an oligonucleotide primer is hybridized with its 3’ end immediately upstream of the locus to be genotyped. The SBE reaction is analogous to Sanger cycle sequencing, except that only chain terminators (dideoxynucleoside triphosphates, ddNTPs) are included in the reaction. The DNA polymerase incorporates the ddNTP complementary to the target base on the template, and further extension of the DNA chain does not occur. Avariety of different detection schemes can be used to determine the identity of the ddNTP that was incorporated and the genotype of the target allele, particularly electrophoresis with four-colour LIF [87–89]. Restriction fragment analysis is based on the use of various endonucleases (restriction enzymes) that cleave recognition sequences usually 4–8 bp in length within double-stranded DNA. Upon treatment, DNA fragments having unique sizes corresponding to the distance between the restriction sites cut by the restriction enzymes are generated. Secondary treatments of the cleaved fragments with other restriction enzymes can provide further specific fragmentation patterns that can be subjected to electrophoretic sorting. If a mutation or SNP exists at particular restriction site, it can change the restriction pattern generated [90,91]. DNA Forensics The numerous advantages of analytical microchip technology take on special importance in the case of forensic DNA analysis. In the forensics community, decreased sample handling is a key feature, as there is less opportunity for sample

Miniaturized Systems for Analytical Separations II

195

contamination during the processing steps. The significant reductions in reagent and sample consumption, and associated reduction in analysis costs, are also of particular benefit in forensic interrogations. Similarly, the potential of the analytical microchip to incorporate all processing steps (DNA extraction, quantization, amplification and separation) on to microdevices for fast (one order of magnitude lower) and automated processing makes this technology an ideal and revolutionary platform for forensic DNA analysis [62]. An important aspect of forensic DNA testing is the identification of humans based on DNA profiling. Usually, this profiling is achieved through the analysis of highly polymorphic variations presented in the form of short tandem repeat fragments (STRs). These are specific sequences, from two to seven base repeating patterns, found in the genome. STRs are preferably distributed throughout junk DNA regions, with the number of repeating units typically very specific for a particular individual, especially when occurring at several loci. Genetic fingerprints are compared to profiles within the Combined DNA Index System (CODIS) databank to identify a match. DNA testing can be used in other forensics areas, for example in establishing the paternity of children, inferring the geographic origin of unknown human DNA samples, identifying potential microbial or viral pathogens for biodefence or recognizing species-specific sources of unknown trace evidence [58]. Some relevant examples of these methodologies are summarized in the excellent review by Harsman et al. [62,92–95] DNA Sequencing With the completion of the human genome sequence, it looks like the era of genomic sequencing (at least large-scale genomic sequencing) is coming to an end. Nonetheless, great effort is being made by researches to develop new technologies capable of decreasing the costs and increasing the data acquisition rate, with the goal of attaining personalized medical care. Full or partial genome sequence determination will ultimately help to realize the promise of medical care tailored to an individual’s unique genetic identity. There is no doubt that the millions of SNPs in the human genome contribute greatly to phenotypical variations among individuals. Nevertheless, other types of larger-scale genetic variations such as insertions, deletions and large-scale genomic rearrangements are also likely to play a prominent role in determining our species’ genetic diversity. In this sense, new human genome resequencing, along with an analysis of copy number polymorphism, loss of heterozygosity and chromosomal rearrangement, will be necessary to fully characterize any one person’s genetic blueprint [63]. Microchip CE devices require some performance features in order to compete favourably with CAE in large-scale sequencing. This means: dispersion-limiting

196

Miniaturization of Analytical Systems

channel turn geometries to extend channels within a small footprint for increased DNA read lengths; multiline formats for high-throughput analyses; and especially, the capacity to integrate sample preparation and DNA sequencing fragment separation on to one device. This last aspect will be treated in detail in Chapter 8, where lab-on-a-chip devices are presented. From the pioneering work of Woolly et al., reported in 1995 [96], in which the sequencing was conducted on a 3.5 cm long glass channel and DNA sequencing traces of 150 bp were performed with 97% base calling accuracy in 540 seconds, new improvements have been made, getting better base read length and higher throughput and resolution in lower analysis time [97,98]. As already mentioned, a great effort is being put into the production of highly integrated sequencing units. Sample preparation and cleanup, together with the electrophoretic separation, are being integrated to minimize the labour and overhead time associated with DNA sequencing. These strategies constitute an exciting challenge for the development of microfluidic platforms. Finally, in order to show the potency of CE miniaturization in DNA analysis, special attention must be given to the development of capillary array electrophoresis technology by Mathies’ group. The timeline of this evolution is shown in Figure 5.9 as a golden example of the role of CE microchips in DNA analysis (see [60] for details). 5.8.2

Protein Analysis

Separation of protein has always been an important part of the analysis of biological systems with CE microchips. Excellent reviews and book chapters can be found in

Figure 5.9 Impress timeline of the capillary array electrophoresis for DNA analysis. (Reprinted from [60] with permission from Elsevier)

Miniaturized Systems for Analytical Separations II

197

the literature [45,99,100]. A variety of protein analysis methods can be developed on-chip. Direct separation of fluorescently labelled proteins has been demonstrated numerous times for both the characterization of new CE microchip devices and the measurement of the purity of protein samples. Capillary electrophoretic immunoassays have also been performed at microscale. Immunoassays allow for the direct measurement of protein analytes in complex samples such as blood and urine, making this a very clinically useful tool. Enzyme assays have several important applications, including the determination of functional pathways in cellular systems and the amplification of signals to improve the sensitivity of detection for routine analytical measurements. Also, 2D impress separations have also been reported. Capillary SDS gel electrophoresis as a method for protein separation has a long and successful history. Various polymers have been demonstrated to be suitable sieving matrices for SDS electrophoresis, including dextran and linear polyacrylamide (LPA), poly(ethylene oxide) and guaran. Three sieving matrices have been developed into commercial products and are currently sold by Beckman Coulter, Bio-Rad and Sigma. Capillary SDS electrophoresis has been performed in conventional CE instrumentation with UV detection. SDS microchip electrophoresis demands the elimination of EOF and an increase in detection sensitivity. Complete elimination of EOF has been a challenging task, while improvement of the detection sensitivity is achieved primarily by using LIF detection. SDS microchip electrophoresis was first demonstrated by Peter Schultz’s group [101]. A commercial system for SDS electrophoresis of proteins on a microchip has been developed by Caliper Technologies [102]. The microchip contains 16 wells for both samples and electrolytes. The chips are made from soda-lime glass. Channels (12.5 mm long, 36 mm wide and 13 mm deep) are produced by etching. The wells have a diameter of 2 mm and, after dicing, the size of each chip is about 17.5 mm2. Wall coating turns out to be unnecessary due to the presence of poly(dimethylacrylamide) (PDMA) sieving matrix that reduces electro-osmotic mobility. Capillary isoelectric focusing (CIEF) is another important method for the separation of proteins [44]. Separating proteins by their pI has achieved very high separation capacity. During separation by CIEF, proteins migrate in a pH gradient generated by carrier ampholytes and focus at various positions along the pH gradient corresponding to their pIs. Multidimensional electrophoresis of proteins (2D of proteins) is a laborious method for high-resolution separation of proteins. This method, which typically consists of IEF in IPG followed by SDS electrophoresis on polyacrylamide slab gel, is very difficult to automate. However, it is considered a gold standard for separation of proteins and its transfer to a microchip format has been seen as solution to many problems. One of the most representative 2D CE systems has been developed by the Ramsey group [103]. A chip for serial coupling of OCEC and CE is shown in Figure 5.10, together with a 2D plot of the separation of a complex peptide mixture.

198

Miniaturization of Analytical Systems

Figure 5.10 2D protein analysis. (Reprinted from [103] with permission of ACS, Copyright 2001)

Affinity CE(ACE) is a method which separates ligands and their substrates (typically proteins). A protein changes its mobility and peak shape when it binds a ligand. These changes can be used to measure stoichiometry between the ligand and the receptor protein, as well as calculate binding kinetics and equilibrium constants. The equilibrium constants provide a quantitative value of the affinity of the ligand to a substrate. There are several derived CE methods that involve immunoaffinity interactions. These interactions can also be used to concentrate analytes from the sample [104]. Inflammatory cytokines were analysed in CSF of patients with head injury by immunoaffinity CE on chip [105]. After incubation with fluorescent dye, cytokines were separated by CE in less than 2 minutes. Since proteins in this type of application have rather low mobilities, the short separation path and thus faster analysis is a significant advantage [105,106]. Enzyme immunoassays are a group of analytical methods with tremendous power due to the combination of their inherent selectivity (antigen–antibody reaction) with the high sensitivity of the relevant detection routes coupled with enzyme amplification. In consequence, these applications are close to the border of lab-on-a-chip. The first electrochemical enzyme immunoassay on microchip platform was reported using antimouse IgG and mouse IgG, as shown in Figure 5.11. A CE microchip with a layout previously designed (fitness-to-purpose) to integrate precolumn reactions of alkaline phosphatase-labelled antibody with antigen, followed by the electrophoretic separation of the free antibody and antibody–antigen complex was strategically used. The separation was followed by a post-column

Miniaturized Systems for Analytical Separations II

Ab-E

199

RB

IRC

B

Ag S

CE

Y Y Y Y Y Y YY Y

Ag-Ab-E

Y Y Y YY Y Y Y Y

Ab-E

RE WE

Figure 5.11

Immunoassay on CE microchip. (Reprinted from [107] with permission from ACS)

reaction of the enzyme tracer, with 4-aminophenyl phosphate substrate, and a downstream amperometric detection of the liberated 4-aminophenol product. A remarkable detection limit (1.7 · 1018 M) was obtained [107]. With respect to peptide analysis, peptides are a large class of molecule that differ in size, charge, conformation, hydrophobicity and the ability to form biospecific complexes. Peptides play an important role in many physiological processes, including the regulation of pain, blood pressure and immune response, and can act as antibiotics, coenzymes/enzyme inhibitors, drugs, growth stimulators, hormones, neurotransmitters and toxins. The role of neuropeptides in the regulation of the central nervous system and signal transduction has been extensively investigated. Identification of specific peptides and their functions may permit the design of pharmacologically active synthetic analogues that can be used for the treatment of neurological diseases such as Parkinson’s and Alzheimer’s. Peptide analysis is also important for proteomics research. Digestion of a protein yields distinct peptide mixtures, which are separated and mapped in order to characterize the parent molecule. In some cases, multidimensional separations are needed to achieve resolution and identification of all peptide components. The determination of peptide concentrations in complex biological samples, protein digests and drug formulations demands the use of highly efficient, sensitive and selective analytical separations. Electrophoretic separation methods are routinely employed because of their fast analysis times, high separation efficiencies and compatibility with sensitive detection techniques. Electrophoretic-based separation modes have been used for the determination of peptide mobility, isoelectric point, relative mass and charge, dissociation/association constants of complexes, diffusion coefficients and hydrophobic qualities. For enzyme immunoassays where biological activity must be maintained, buffers of physiological pH and temperature can be used. For hydrophobic peptides, the

200

Miniaturization of Analytical Systems

addition of the surfactants to the run buffer can aid peptide solubility. Coating the capillary wall minimizes adsorption of analyte to the capillary surface. Finally, because most peptides are amphoteric, analyte charge can be altered by adjusting the pH. This can change mobility, as well as enhancing solubility and reducing adsorption to the negative wall. Many of the methods developed for peptide analysis on conventional CE systems are now being transferred to microchip platforms, owing to the many additional advantages of miniaturized analytical system. 5.8.3

Small-molecule Analysis

Bionalytical and Clinical Analysis Applications in bioanalytical and clinical analyses have mostly been covered in Sections 5.8.1 and 5.8.2. However, bioanalytical and clinical analyses are of crucial significance around the world and many of the unique possibilities of miniaturized systems are very important in the analysis of other small molecules with high health impact, such as glucose, carbohydrates, lactic acid and renal markers, among others. Indeed, one of the first successful reports on separations of microchips described the electrophoretic separation of a mixture of fluorescent-labelled amino acids [108]. A small amount of sample (pl–nl) was injected and transported down a main channel, where separation occurred because of the different migration velocities of the components of the sample. To detect individual bands of molecules as they passed the detector, a laser with a suitable wavelength was focused on a single point in the microchannel, where it excited the fluorescent-labelled molecules. This type of setup (relatively simple channel layout, (programmable) power supplies to facilitate injection and separation, and LIF) has been and still is the most commonly used microchip setup for analytical separations (because of its historical significance, this work will be studied in detail in Chapter 6). Homocystein, which is an amino acid involved in the synthesis and metabolism of the essential amino acid methionine [109]; uric acid [110]; and renal markers (creatine, creatinine, p-aminohippuric acid and uric acid) [111] are among the studied compounds of clinical significance. Different strategies involving enzymes and different analytical strategies have been creatively proposed for the detection of small molecules of significance to health. Figure 5.12 shows representative applications involving each chip layout with electrochemical detection: on-column, precolumn and post-column concepts. First, using the on-column concept, the simultaneous bioassay of glucose, uric acid, ascorbic acid and acetaminophen has been demonstrated [112]. The beauty of this strategy consists in the mixing of the enzyme glucose oxidase with the sample on-column liberating hydrogen peroxide after the sample reaction with glucose. Neutral hydrogen peroxide was separated from the other negative charged analytes, as can be observed in the electrophoregrams shown in Figure 5.12(a). By incorporating a

Miniaturized Systems for Analytical Separations II

10 nA a

b

201

c 3 nA c a b de

D

b 2 nA

fg h

C

a

i

c d

B A 100

200 Time’s

On-column

300

0

100

200 Time’s

300

Pre-column

400

0

100 200 300 400 500 Time’s

Post-column

Figure 5.12 CE microchip layouts for small molecules of bioanalytical relevance. (Reprinted from [112,113] with permission of ACS and [114] with permission of Wiley-VCH)

precolumn reactor into the chip design – an integrated microfabricated device that performs the online derivatization/biocatalytic reactions of multiple enzymes and substrates – the electrophoretic separation of the involved analytes and the electrochemical detection process on a chip platform have been illustrated for the electrophoretic separation of eight amino acids (analysis time, 6 minutes; LODs down to the 2.5 · 106 M level) (Figure 5.12(b)) [113]. A new biochip strategy for performing post-column reaction of a ‘class’ enzyme (amino acid oxidase) on CE microchip platforms has also been described for the rapid and sensitive measurement of amino acids (arginine, isoleucine, alanine, phenylalanine) in connection to a postcolumn biogeneration of electroactive hydrogen peroxide [114] (Figure 5.12(c)). The beauty of this work is that the concept can be extended to different target analytes based on the post-column reactions of other ‘class’ enzymes, thus improving selectivity and sensitivity in their analytical performance. Figure 5.13 also shows an impress application [115] dealing with the on-chip integration of enzymes and immunoassays. This novel concept was applied to the simultaneous measurement of insulin and glucose. As illustrated in the figure, the implementation of such a dual-bioassay single-channel microchip protocol involves the judicious coupling of multiple enzymatic and immunoreactions through integration of electrophoretic separations with relevant pre- and post-column reaction steps. The operation taking place in the biochip includes precolumn reactions of the enzyme-labelled antibody (antihuman insulin) with its antigen (insulin) and the reaction of the glucose-dehydrogenase (GDH) enzyme with its glucose substrate in the presence of NADþ cofactor; these reactions are followed by the electrophoretic separation of the free antibody, antibody–antigen complex and

202

Miniaturization of Analytical Systems

Ab-E2 + E1

RB B

RC

Y Y Y Y Y Y YY Y

Ag + S1 S2

Insulin Immuno complex

NADH

Y Y Y YY Y Y Y Y CE

Free label IgG

RE WE

Figure 5.13 Schematic of the CE biochip used for simultaneous immunological and enzymatic assays. (Reprinted from [115] with permission from ACS)

NADH product of the glucose enzymatic reaction, and a post-column reaction of the alkaline phosphatase (ALP) enzyme tag with its p-nitrophenylphosphate (p-NPP) substrate. Both the NADH and the p-nitrophenol products are monitored at the downstream amperometric detector. Such a coupling of different bioassays on the same microfluidic device reflects the versatility and integration features of such devices. Environmental Analysis Environmental applications are one of the most important analytical areas of CE microchips related to the demands of ‘green analytical chemistry’, due to the negligible waste generated by these microsystems, which are termed ‘environmentally friendly’ [48,49]. The utility of CE microchips for the rapid separation (4 minutes) of seven priority chlorophenolic pollutants down to a concentration level of 1 · 106 M, for the detection of toxic organophosphate nerve agents such as paraoxon, methyl parathion, fenitrothion and ethyl parathion in 140 seconds, and for the separation and detection of organic peroxides in less than 150 seconds has been noted in the last few years [49]. In addition, aromatic amines, hydrazines, nitroaromatic compounds and inorganic and organic ions have been studied and reviewed [49]. Another environmental-miscellaneous group of relevant analytical applications can be found in the literature [116]. A very attractive strategy using a singlechannel chip-based analytical microsystem has allowed the rapid measurement of the total content of relevant compounds (organic explosives or nerve agents), as well as detailed micellar chromatographic identification of each [116]. Figure 5.14

Miniaturized Systems for Analytical Separations II

203

Individual (a) 10 nA

(b)

1 2 3 4 56

5 nA 1

explosives

3

(ii)

nerve 20 nA agents

100 nA

RB with SDS

Total

RB without SDS Sample

(i)

(i)

0

+2 kv

2

(ii)

50 100 150 0

50

100

150

Time (s) +5 kv

Figure 5.14 Schematic of individual and total assays on CE microchips and electrophoregrams depicting the total (i) and the individual (ii) assays for mixtures of nitroaromatic (a) nerve agents (b). (Reprinted from [116] with permission of ACS)

shows the strategy schemes employed and the electrophoregrams obtained for both individual and total (less than 1 minute!) measurements of each compound class studied. This strategy reveals the very unique integration of important modern concepts in analytical chemistry, such as screening methods for the measurement of total index in order to detect ‘alarm’ concentrations (vanguard strategies), the separation/confirmation of individual ones (rearguard strategies) and the use of two separation protocols (MECK and CZE), on to a single microchip. Food Analysis An important development has been noticed for this field of application. Emerging applications in food analysis using CE microchips were focused on separation of analytes of significance in food such as natural antioxidants, vitamins, flavours, preservatives and amino acids using typical microchip CE layouts [50]. Lately, the well-known complexity of food matrices has been approached using CE microchips with different strategies to improve the selectivity and sensitivity of the analysis by avoiding sample preparation or making it as simple as possible: (i) enhancing the peak capacity in order to perform direct injection; (ii) using the microchip platform to measure one target analyte/group of analytes with or without separating other related interferences; (iii) integrating sample preparation steps on the microchip platform; and (iv) integrating new analytical tools from nanotechnology in the detection stage [51]. New analyte separations of significance in food involving DNA probes, biogenic amines, vanilla flavours and dyes have been reported as successfully breaking new barriers in areas of high impact in the market, such as transgenic food analysis

204

Miniaturization of Analytical Systems

and the detection of frauds and toxins. A simple microchip layout has been the most common design used, though sophisticated new ones are emerging. Fast detection of food-related phenolic acids (chlorogenic, genistic, ferulic and vanilic acids) [117] as well as seleno-amino acids of significance in food (Se-methionine, Se-ethionine and Se-methyl cysteine) have been reported [118]. C.S. Henry et al. have demonstrated the suitability of PDMS microchip in the analysis of flavan-3-ols [119]. Separation of (þ)-catechin, epigallocatechin gallate, ()-epicatechin and epicatechin gallate was achieved in 300 seconds using the MECK mode, avoiding the use of organic solvents; compare this to HPLC method, which needs 25 minutes and large organic solvent volumes. Simultaneous detection of prominent natural antioxidants including ascorbic acid and flavonoids has been reported in less than 200 seconds [120], and watersoluble vitamins (B and C) have been separated in less than 130 seconds using single-channel microchip electrochemistry platforms [121]. Separation of up to five analytes involving natural vanillin markers and ethyl vanillin – commonly used as an artificial flavour – was obtained in less than 200 seconds. These separations are opening exciting new analytical possibilities in studies of adulteration and authenticity in food samples [122,123]. Prominent amino acids from green tea (Arg, The and Gln) were separated in less than 2 minutes and were really detected in green tea samples using plastic microchips with LIF detection [124]. The separation of commonly used preservatives (such as benzoate, sorbate and vitamin C) on PMMA microchip has also been reported [125,126]. The transference of the adopted buffer from conventional to microchip format gave a poor resolution, mainly due to the short effective length of the microchip compared to conventional capillary, and another buffer, prepared with cyclodentrins HP-b-CD and CTAB, was needed to achieve suitable separation. These studies pointed out that the transference from conventional to microchip format is not just a decrease of scale, and that additional investigations should be made before a practical adaptation. Other relevant examples are, among others: the ultrafast detection and simultaneous analysis of genetically modified maize [127]; the analysis of neuroactive amines in fermented beverages using a portable and versatile microfabricated CE microchip system, which has been used in the identification of biomarkers such as amino acids, amines, amino sugars and nucleobases in other fields [128]; artificial dyes [129,130]; and toxic alkaloids (colchicines, aconitine, strychnine and nicotine) [131,132]. Figures 5.15–5.17 show selected examples of microchip layouts and analyte separations of significance in food on microchip format. These figures clearly show the main strength of the miniaturization approach: the dramatic decrease of the analysis timescale without a loss of analytical performance. These examples exhibit the main trends and the work done in the area using CE microchips. Finally, an important review dealing with quantitative analysis using CE microchips indicates the analytical potency of these miniaturized devices [133]. Table 5.4 gives a basic overview of the relevant applications and selected literature.

Miniaturized Systems for Analytical Separations II

205

Figure 5.15 CE microchip layouts used in separation of small molecules of food significance. ((a) Reproduced from [29] with permission from Elsevier; (b) From [119] reproduced by permission of the RSC; (c) From [125] with permission from Wiley-VCH; (d) Reprinted from [124] with permission from Elsevier)

Figure 5.16 Fast separation of vanilla fingerprint markers for sample screening and analyte confirmation of frauds in one single run. (Adapted from [123] with permission of the RSC)

206

Miniaturization of Analytical Systems

Figure 5.17 Sample preparation integrated on CE microchip. (a) Microchip layout; (b) detection of typical dyes. (Reprinted from [132] with permission of Wiley-VCH)

5.9 Outlook CE on microchips is undoubtedly the most rapidly advancing and fruitful branch to date in miniaturized separation systems. The reason for this dominance lies partially in the inherent simplicity of its fabrication and operation, combined with its unique features with respect to separation speed, sample injection and reagent consumption. CE microchips employ channels etched into a planar substrate, which are based upon microfabrication techniques developed in the semiconductor industry. Early applications of CE microchips employed glass or quartz for chip substrates because they are optically transparent and exhibit EOF properties similar to those of fused silica. A number of other materials including polymer/plastics have since been investigated. Photolithography and micromolding are used to fabricate microchannels for sample injection, CE separation and analyte detection. Once all solutions, including those of the samples, are loaded, the samples are typically transferred electrokinetically into an injector region. Their components are then separated by application of a high voltage, and afterwards detected with a suitable detection system (the latest will be studied in Chapter 6). Moreover, CE microchips have the potential to simultaneously assay hundreds of samples in a matter of minutes or less. The rapid analysis combined with massively parallel analysis arrays should yield an ultrahigh throughput. These microchips typically consume only picolitres of sample and may potentially be prepared onboard for a complete integration of sample preparation, separation and detection.

Analytes studied (selected)

Glass (mainly) and polymer microchips Simple cross-channel (mainly CZE and MEKC) Glass/PDMS microchips Simple/double-T cross channel

Environmental (pollutants, anti-terrorist controls) Food (control of frauds, food safety)

Chlorophenols Nerve agents Explosives Organic peroxides Natural antioxidants Vitamins Flavours Preservatives

Glass and polymer microchips Pre-column channel Post-column channel (mainly CZE and MEKC)

(mainly gel electrophoresis) multiline formats

Highly sophisticated (reaction chamber, digestion, PCR amplification)

Features of microchip and modes

Proteomic & immunoassays Small-molecule analysis Clinical & bioanalytical Aminoacids (point-of-care Glucose, carbohydrates testing) Renal markers Neurotransmitters

DNA & protein analysis Genotyping (scanning DNA, proteins, peptides applications; diagnostic applications) Forensics

Field of application

Table 5.4 Selected overview of relevant analytical applications of CE microchips

Medium degree of maturity with integrated protocols involving enzymatic reactions using pre- and post-column approaches Fast separations are mainly focused on standards of environmental/food significance Low degree in validation methods and quantitative domains

Fast separations. High degree of maturity (instrumentation commercially available)

Remarks

[50,51]

[48,49]

[2,4,16,45]

[4,45,99,100] (protein analysis)

[4,52,53,58,60–64] (DNA analysis)

References (selected)

Miniaturized Systems for Analytical Separations II 207

208

Miniaturization of Analytical Systems

Also, a couple of key advantages of the use of these planar devices/substrates are their ease of manufacture and the fact that microfabrication techniques can be adapted for mass production. These features make CE microchips an attractive technology for the next generation of CE instrumentation. A wide array of applications involving mainly DNA and protein analysis have been investigated, indicating the power of these miniaturized devices. In addition, the suitability of these devices in new applications involving small molecules in fields such as clinical, environmental and food has also been well documented. In summary, these microsystems have a very well-defined, prominent role in the analytical miniaturization scene.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

S.C. Terry, J.H. Jerman, J.B. Angell, IEE Trans Electron Devices, 26 (1979) 1880. U. Bilitewski, M. Genrich, S. Kadow, G. Mersal, Anal. Bioanal. Chem., 377 (2003) 556. V. Dolnik, S. Liu, S. Jovanovich, Electrophoresis, 21 (2000) 41. G.J.M. Bruin, Electrophoresis, 21 (2000) 3931. A. de Mello, Lab Chip, 2 (2002) 48N. A. Manz, N. Graber, H.M. Widmer, Sens. Actuators B, 1 (1990) 244. A. Manz, D.J. Harrison, E.M.J. Verpoorte, J. Chromatogr. A, 593 (1992) 253. P.S. Dittrich, K. Tachikawa, A. Manz, Anal. Chem., 78 (2006) 3887. S. Liu, A. Guttman, Trends Anal. Chem., 23 (2004) 422. R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003. O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004. C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Human Press, Totopwa, New Jersey, 2006. L. Sz
ekely, A. Guttman, Electrophoresis, 26 (2005) 4590. Z. Huang, N. Munro, A.F.R. Huhme, J.P. Landers, Anal.Chem., 71 (1999) 5309. M.A. Garc
ıa, M.L. Marina, A. R
ıos, M. Valc
arcel, In: M.L. Marina, A. R
ıos, M. Valc
arcel (Eds), Analysis and detection by capillary electrophoresis, Elsevier (CAC Series), Amsterdam, 2005, pp. 31–134. S.C. Jakeway, A.J. de Mello, E.L. Russell, Fresenius J Anal. Chem., 366 (2000) 525. J.C.T. Eijkel, A.J. de Mello, A. Manz, In: M.H. Schryver (Ed.), Organic Mesoscopic Chemistry, Blackwall Science, 1999, pp. 185–219. M.J. Madou (Ed.) Fundamentals of Microfabrication (The Science of Miniaturization), CRC Press LLC, Boca Raton, 2000. H. Becker, C. G€artener, In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-chip. Miniaturized systems for (bio) Chemical analysis and synthesis, Elsevier, Amsterdam, 2003, pp. 21–35. J.G. Gardieneers, R.E. Oosterbroek, A. van der Berg, In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003, pp. 37–64. G. Binyamin, T.D. Boone, H.S. Lackritz, A.J. Ricco, A.P. Sassi, S.J. Williams, In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003, pp. 83–112. A.M. Jorgensen, K.B. Mogensen, In: O. Geschke, H. Klank, P. Telleman (Eds),) Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004, pp. 117–160.

Miniaturized Systems for Analytical Separations II

209

[23] D. Petersen, K.B. Mogensen, H. Klank, In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004, pp. 161–168. [24] H. Klank, In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-aChip Devices, Wiley-VCH, Weinheim, 2004, pp. 169–182. [25] D.C. Duffy, O.J.A. Schueller, G.M. Whitesides, Anal. Chem., 70 (1998) 4974. [26] H. Cui, K. Horiuchi, P. Dutta, C.F. Ivory, Anal. Chem., 77 (2005) 1303. [27] N.A. Polson, M.A. Hayes, Anal. Chem. (2001) 313A. [28] J.P. Kutter, K.B. Mogensen, H. Klank, O. Geschke, In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004, pp. 39–77. [29] A.G. Crevill
en, I. Barrigas, A.J. Blasco, M.C. Gonz
alez, A. Escarpa, Anal. Chim. Acta, 562 (2006) 137. [30] S.C. Jacobson, R. Hergenr€oder, L.B. Koutny, J.M. Ramsey, Anal. Chem., 66 (1994) 2369. [31] M. Pumera, Talanta, 66 (2005) 1048. [32] L. Sz
ekely, A. Guttman, Electrophoresis, 26 (2005) 4590. [33] M.J. Sepaniak, D.F. Swaile, A.C. Powell, J. Chromatogr. A, 480 (1989) 185. [34] M. Pumera, J. Wang, E. Grushka, R. Polsky, Anal. Chem., 73 (2001) 5625. [35] W. Xu, K. Uchiyama, T. Shimosaka, T. Hobo, J. Chromatogr. A, 907 (2001) 279. [36] W. Xu, K. Uchiyama, T. Hobo, Chromatography, 23 (2002) 131. [37] J.P. Kutter, S.C. Jacobson, N. Matsubara, J.M. Ramsey, Anal. Chem., 70 (1998) 3291. [38] R.D. Oleschuk, L.L. Schultz-Lockyear, Y. Ning, D.J. Harrison, Anal. Chem., 72 (2000) 585. [39] K.W. Ro, W.J. Chang, H. Kim, J.M. Koo, J.H. Hahn, Electrophoresis, 24 (2003) 3253. [40] L. Ceriotti, N.F. Rooij, E. Verpoorte, Anal. Chem., 74 (2002) 639. [41] G.A. Lord, D.B. Gordon, P. Myers, B.W. King, J. Chromatogr. A, 768 (1997) 9. [42] D.J. Throckmorton, T.J. Shepodd, A.K. Singh, Anal. Chem., 74 (2002) 784. [43] B.E. Slentz, N.A. Penner, F. Regnier, J. Chromatogr. A, 948 (2002) 225. [44] J. Vindevogel, P. Sandra, Electrophoresis, 15 (1994) 842. [45] V. Dolnik, S. Liu, J. Sep. Sci., 28 (2005) 1994. [46] M. Pumera, Electrophoresis, 28 (2007) 2113. [47] M. Pumera, Electrophoresis, 27 (2006) 244. [48] L. Marle, G.M. Greenway, Trends Anal.Chem., 24 (2005) 795. [49] G. Chen, Y.H. Lin, J. Wang, Talanta, 68 (2006) 497. [50] A. Escarpa, M.C. Gonz
alez, A.G. Crevill
en, A.J. Blasco, Electrophoresis, 28 (2007) 1002. [51] A. Escarpa, M.C. Gonz
alez, M.A. L
opez Gil, A.G. Crevill
en, M. Herv
as, M. Garc
ıa, Electrophoresis, 29 (2008) 4852. [52] E. Sz
antal, A. Guttman, Electrophoresis, 27 (2006) 4896. [53] M. Handal, V.M. Ugaz, Expert. Rev. Mol. Diagn., 6 (2006) 29. [54] E.S. Lander, L.M. Linton, B. Virren, C. Nusbaum, Nature, 409 (2001) 860. [55] Y. Sowa, Sci. Technol. Trends, 18 (2006) 13. [56] F.S. Collins, E.D. Green, A.E. Guttmacher, M.S. Guyer, Nature, 422 (2003) 835. [57] A. Syv€anen, Nature Genet., 37 (2005) S5. [58] R. Sinville, S.A. Soper, J. Sep. Sci., 30 (2007) 1714. [59] R.T. Kelly, A.T. Woolley, Anal. Chem., 1 (2005) 97A. [60] R.G. Blazej, B.M. Paegel, C.A. Emrich, R.A. Mathies, In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003, pp. 231–247. [61] A.W. Chow, In: C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Human Press, Totopwa, New Jersey, 2006, pp. 129–143. [62] K.M. Horsman, J.M. Bienvenue, K.R. Blasier, J.P. Landers, J. Forensic Sci., 52 (2007) 784. [63] C.P. Fredlake, D.G. Hert, E.R. Mardis, A.E. Barron, Electrophoresis, 27 (2006) 3689. [64] S. Liu, A. Guttman, Trends Anal. Chem., 23 (2004) 422. [65] H.J. Tian, A. Jaquins-Gerstl, N. Munro, M. Trucco, L.C. Brody, J.P. Landers, Genomics, 63 (2000) 25. [66] G. Vahedi, K. Kaler, C.J. Backhouse, Electrophoresis, 25 (2004) 2346.

210

Miniaturization of Analytical Systems

[67] D.P. Manage, Y. Zheng, M.J. Somerville, C.J. Backhouse, Microfluidic Nanofluidic, 1 (2005) 364. [68] H.J. Tian, C. Emrich, J. Scherer, Electrophoresis, 26 (2005) 1834. [69] Y. Endo, L. Zhang, R. Katashima, M. Itakura, E.A.S. Doherty, A.E. Barron, Y. Baba, Electrophoresis, 26 (2005) 3380. [70] S.H. Kang, S. Jang, S.K. Park, Bull Korean Chem. Soc., 27 (2006) 1346. [71] J.S. Buch, C. Kimball, F. Rosenberger, W.E. Highsmith, D.L. DeVoe, C.S. Lee, Anal. Chem., 76 (2004) 874. [72] J.S. Buch, C. Kimball, F. Rosenberger, W.E. Highsmith, D.L. DeVoe, C.S. Lee, Lab Chip., 5 (2005) 392. [73] H.J. Tian, L. Brody, J.P. Landers, Genome Res., 10 (2000) 1403. [74] H.J. Tian, J.P. Landers, Anal. Biochem., 309 (2002) 212. [75] T. Footz, M. Samerville, R. Tomaszewski, Genetic Testing, 7 (2003) 283. [76] T. Footz, M. Samerville, R. Tomaszewski, Analyst, 129 (2004) 25. [77] A.P. Sassi, A. Paulus, I.D. Cruzado, J. Chromatog. A, 894 (2000) 203. [78] W.C. Sung, G.B. Lee, C.C. Tzeng, S.H. Chen, Electrophoresis, 22 (2001) 1188. [79] X.M. Zhou, S.J. Shao, G.D. Xu, R.T. Zhong, D.Y. Liu, J.W. Tang, Y.N. Gao, S.J. Chen, B.C. Lin, J. Chromatog. B, 816 (2005) 145. [80] I. Mendintz, W. Wong, G. Sensabaugh, R.A. Mathies, Electrophoresis, 21 (2000) 2352. [81] I. Mendintz, W.W. Wong, L. Berti, L. Shiow, J. Tom, J. Scherer, G. Sensabaugh, R.A. Mathies, Genome Res., 11 (2001) 413. [82] K. Karasawa, H. Arakawa, T. Igarashi, N. Goto, M. Maeda, J. Chromatog. B, 810 (2004) 41. [83] X.J. Lou, N. Panaro, P. Wilding, Biotechniques, 37 (2004) 392. [84] Y. Wang, B.V. Aidya, H. Farquar, W. Stryjewski, R.P. Hammer, R.L. McCarley, S.A. Soper, Y.W. Cheng, F. Barany, Anal. Chem., 75 (2003) 1130. [85] G. Thomas, R. Sinville, S. Sutton, H. Farquar, R.P. Hammer, S.A. Soper, Y.W. Cheng, F. Barany, Electrophoresis, 25 (2004) 1668. [86] H. Hashimoto, M.L. Hupert, M.C. Murphy, S.A. Soper, Y.W. Cheng, F. Barany, Anal. Chem., 77 (2005) 3243. [87] N. Herbert, S.A. Brazil, Lab Chip, 3 (2003) 241. [88] Z.P. Li, H. Tsunoda, K. Okano, K. Nagai, H. Kambara, Anal. Chem., 75 (2003) 3345. [89] W.N. Vreeland, R.J. Meagher, A.E. Barron, Anal. Chem., 74 (2002) 4328. [90] E.S. Roddy, M. Price, A.G. Ewing, Anal. Chem., 75 (2003) 3704. [91] A. Guttman, H.G. Gao, R. Haas, Clin. Chem., 47 (2001) 1469. [92] D. Schmalzing, L. Koutny, A. Adourian, P. Belgrader, P. Matsudaira, D. Ehrlich, Proc. Natl. Acad. Sci. USA, 94 (1997) 10273. [93] D. Schmalzing, L. Koutny, L. Chisholm, A. Adourian, P. Matsudaira, D. Ehrlich, Anal. Biochem., 270 (1999) 148. [94] Y.N. Shi, Electrophoresis, 27 (2006) 3703. [95] S.H. Yeung, S.A. Greenspoon, A. McGuckian, C.A. Crouse, C.A. Emrich, J. Ban, R.A. Mathies, J. Forensic Sci., 51 (2006) 740. [96] A.T. Woolley, R.A. Mathies, Anal. Chem., 67 (1995) 3676. [97] J.H. Aborn, S.A. El-Difrawy, M. Novotny, E.A. Gismondi, R. Lam, P. Matsudaira, B.K. McKenna, T. O’Neil, P. Streechon, D.J. Ehrlich, Lab Chip, 5 (2005) 669. [98] T. Boone, Z.H. Fan, H. Hooper, A. Ricco, H.D. Tan, S. Willians, Anal. Chem., 74 (2002) 78A. [99] B.A. Fogarty, N.A. Lacher, S.M. Lunte, In: C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Human Press, Totopwa, New Jersey, 2006, pp. 159–186. [100] A.W. Chow, In: C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Human Press, Totopwa, New Jersey, 2006, pp. 145–158. [101] S. Yao, D.S. Anex, W.B. Caldwell, D.W. Arnold, K.B. Smith, P.G. Schultz, Proc. Natl. Acad. Sci. USA, 96 (1999) 5372. [102] L. Bousse, S. Mouradian, A. Minalla, H. Yee, K. Williams, R. Dubrow, Anal. Chem., 73 (2001) 1207. [103] N. Gottschlich, S.C. Jacobson, C.T. Culbertson, J.M. Ramsey, Anal. Chem., 73 (2001) 2669.

Miniaturized Systems for Analytical Separations II [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133]

211

N. Guzman, T.M. Phillips, Anal. Chem., 77 (2005) 61A. T.M. Phillips, Electrophoresis, 25 (2004) 1652. A.R. Stettler, M.A. Schwarz, J. Chromatogr. A, 1063 (2005) 217. J. Wang, A. Ib
a~nez, M.P. Chatrathi, A. Escarpa, Anal. Chem., 73 (2001) 5323. C.S. Effenhauser, A. Manz, H.M. Widmer, Anal. Chem., 65 (1993) 2637. S.A. Pasas, N.A. Lacher, M.I. Davies, S.M. Lunte, Electrophoresis, 23 (2002) 759. J.C. Fanguy, C.S. Henry, Electrophoresis, 23 (2002) 767. J. Wang, G. Chen, A. Muck, D. Shin, A. Fujishima, J. Chromatogr A, 1022 (2004) 207. J. Wang, M.P. Chatrathi, B. Tian, R. Polsky, Anal. Chem., 72 (2000) 2514. J. Wang, M.P. Chatrathi, B. Tian, Anal. Chem., 72 (2000) 5774. J. Wang, M.P. Chatrathi, A. Ib
a~nez, A. Escarpa, Electroanalysis, 14 (2002) 400. J. Wang, A. Ib
a~nez, M.P. Chatrathi, J. Am. Chem. Soc., 125 (2003) 8444. J. Wang, M. Pumera, M.P. Chatrathi, A. Escarpa, M. Musameh, Anal. Chem., 74 (2002) 1187. M. Scampicchio, J. Wang, S. Mannino, M.P. Chatrathi, J. Chromatogr. A, 1049 (2004) 189. J. Wang, S. Manino, C. C
amara, M.P. Chatrathi, M. Scampicchio, J. Zima, J. Chromatogr. A, 1091 (2005) 177. R.W. Hompesch, C.D. Garc
ıa, D.J. Weiss, J.M. Vivanco, C.S. Henry, Analyst, 5 (2005) 694. A.J. Blasco, I. Barrigas, M.C. Gonz
alez, A. Escarpa, Electrophoresis, 26 (2005) 4664. A.G. Crevill
en, A.J. Blasco, M.C. Gonz
alez, A. Escarpa, Electrophoresis, 27 (2006) 5110. ´ vila, M.C. Gonz
alez, M. Zougagh, A. Escarpa, A. R
ıos, Electrophoresis, 28 (2007) 4233. M. A ´ vila, M. Zougagh, A. Escarpa, A. R
ıos (in press). M. A M. Kato, Y. Gyoten, K. Sakai-Kato, T. Toyo’oka, J. Chromatogr. A, 1013 (2003) 183. P. Kuban, P.C. Hauser, Electrophoresis, 26 (2005) 3169. W.S. Law, P. Kub
an, J.H. Zhao, F.Y. Li, P.C. Hauser, Electrophoresis, 26 (2005) 4648. K.S. Kumar, S.H. Kang, Electrophoresis, 28 (2007) 4247. C.N. Jayarajah, A.M. Skelley, A.D. Fortner, R.A. Mathies, Anal. Chem., 79 (2007) 8162. N. Dossi, E. Piccin, G. Bontempelli, E. Carrilho, J. Wang, Electrophoresis, 28 (2007) 4240. N. Dossi, R. Toniolo, A. Pizzariello, S. Susmel, F. Perennes, G.J. Bontempelli, Electroanal. Chem., 601 (2007) 1. C.I.D. Newman, B.C. Giordano, C.L. Copper, G.E. Collins, Electrophoresis, 29 (2008) 803. K.S. Lee, M.J.A. Shiddiky, S.H. Park, D.S. Park, Y.B. Shim, Electrophoresis, 29 (2008) 1910. T. Revermann, S. G€otz, J. K€unnemeyer, U. Karst, Analyst, 133 (2008) 167.

6 Detection in Miniaturized Analytical Systems 6.1

Introduction

Detection in analytical microsystems is a subject of paramount importance. Its development is intrinsically linked to the birth of the concept of mTAS in the early 1990s [1]. Indeed, detection has been one of the main challenges for analytical microsystems, since very sensitive techniques are needed as a consequence of the ultrasmall sample volumes used in micron-sized environments. In consequence, without question, in order to speak of challenges in the miniaturization domain we must speak of detection schemes on these devices [2]. In principle, a wide variety of detection alternatives can be used in analytical microsystems, which can be categorized into optical and electrochemical detections, mass spectrometry and other unconventional detection schemes. Optical detectors are commonly used due to their accessibility in laboratories and the simplicity of microfluidics–detector interfaces. Among them, laser-induced fluorescence (LIF) was the original detection technique, and it is the most often used detection scheme in microsystems because of its inherent sensitivity [3–6]. Electrochemical detection (ED) is very important because of its inherent miniaturization without loss of performance and its high compatibility with microfabrication techniques. Likewise, it possesses high sensitivity, its responses are not dependent on the optical path length or sample turbidity, and it has low power supply requirements, all of which are additional advantages [7–10]. Mass spectrometry (MS), the third approach most commonly used, has been covered in some excellent book chapters [11,12]. Apart from LIF, ED and MS, there are other detection Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

214

Miniaturization of Analytical Systems

approaches for analytical microsystems, called ‘unconventional detection routes’, but these have been developed to a lesser extent. They have recently been revised and include, among others, ultraviolet (UV), surface-enhanced Raman spectroscopy (SERS), infrared (IR), resonance magnetic nuclear (RMN), surface plasmon resonance (SPR) and thermal lens microscopy (TLM) [13]. In this chapter the focus will be on the more relevant detection routes, such as LIF and ED. An overview of MS coupled with microsystems and outlines of the more exotic detection routes will also be given.

6.2 Laser-induced Fluorescence (LIF) Detection It is convenient to incorporate conventional optical devices into microfluidic detection systems because no interconnections are required between the detector unit and the microfluidicdevice. Microfluidicsinvolves onlyphotons, sominimaldisturbance is caused to the reaction system as long as it is not susceptible to photochemical decomposition. Therefore, optical detection is highly compatiblewith microfluidics. However, the key problem with optical detection in microfluidics is its relatively poor detection limit, due to the small amount of sample detected at a particular point in time in the miniaturized channels. The application of universal absorbance detection is hindered by the short path lengths of small microchannels. Therefore, fluorescence detection is the dominant optical detection in analytical microsystems. LIF is one of the most important detection methods used in microchip separations, because it can be adapted best to low-concentration/low-volume systems. Recent advances in laser technology have produced stable light sources that cover a rapidly increasing range of wavelengths from the UV to the IR region of the electromagnetic spectrum, that are relatively inexpensive and that can easily be focused on to micron-sized detection areas. The use of a pinhole at the focus point along the optical path (confocal LIF) even allows for 3D focusing and further reduction of the background signal (scattered light, autofluorescence of microchip material). These very low levels of background signal combined with very sensitive photon detection techniques – photomultiplier tubes (PMTs), photon counting systems, charge-coupled devices (CCDs) – result in the lowest limits of detection (LOD) of all microchip detection systems. The fact that only a few compounds exhibit native fluorescence is another reason for the low background signals in LIF, especially in comparison with absorption spectroscopy. On the other hand, labelling with fluorescent markers is required for all nonfluorescent analytes. Although commercial instruments that can be used for microchip LIF (such as the Agilent Bioanalyzer 2100 system) are available, most researchers depend on homemade LIF detection systems. These usually consist of the following basic parts: a laser excitation source, an optical system to focus and collect the light, and a photosensitive detector, generally a PMT.

Detection in Miniaturized Analytical Systems

215

Figure 6.1 Seminal LIF works. (a) Schematic of an optical detection system. An ion laser light is focused with a lens (1) on to the separation channel, which is held in place with a Plexiglas holder (2). Fluorescence emission (3) is collected. With microscope objective (4) focused on to an air slit (5), filtered (6) and then detected with a PMT. (Reproduced from [14] with permission of American Chemical Society, Copyright 1993); (b) Seminal LIF works. Impress separation of a mixture of six FITC-labelled amino acids at separation length of 24 mm. (Reproduced from [15] with permission of ACS, Copyright 1993)

As the conceptual review of Schwarz and Hauser states [2], since the introduction of the concept, the established way of carrying out fluorescence detection has been based on an argon-ion laser with an emission line at 488 nm, a microscope lens for collecting the fluoresced light and a PMT for detection [14,15]. Figure 6.1 shows selected information from the seminal works developed by the parents of mTAS: a schematic of the optical detection system used and the impressive separation of six FITC-labelled amino acids in a few seconds, indicating the power of this detection system in terms of efficiency and sensitivity. Post-column labelling carried out on-chip with the aid of an auxiliary reagent channel eliminates the manual analyte derivatization step prior to injection. Liu et al. have recently optimized this method for protein determination by the use of a dye which binds noncovalently and possesses the fast reaction kinetics necessary for rapid analysis and high sensitivity [16]. Lu et al. [17] have extended fluorescence detection on-chip to inorganic species by determining uranium using a selective calixarene ligand which has been specifically derivatized with a rhodamine moiety in order to render it fluorescent. In order to make fluorescence detection more universal, a number of researchers have investigated the use of indirect fluorescence detection. For indirect detection, a

216

Miniaturization of Analytical Systems

Figure 6.2 (a) Common setup used for a LIF detection system, including laser, excitation filter and focusing lens: (i) emission is collimated by an objective followed by an emission filter and PMT detector; (ii) epifluorescence setup, including wavelength-dependent (dichroic) beam splitter. (b) Integrated optical fibres for LIF with two-channel PMT detection. (c) Orthogonal LIF setup with collimation of emission light through the sidewall of the chip. (Adapted from [6] with permission from Springer)

fluorescent substance with a charge of the same sign as that of the analyte is added to the background buffer and its displacement (in order to maintain electroneutrality) is measured in the form of negative going peaks. Detections of amino acids [18], phenols of environmental significance [19], explosives [20] and photographic developers [21,22] have been performed using this approach. Taking into account all the relevant literature, and following the method in the excellent reviews of G€ otz and Kurst [6], as well as Viskari and Landers [13], two main optical system designs can be found, as depicted in Figure 6.2(a). The first (i) uses a lens or an objective to focus the laser on to the microchip channel, typically at an angle of 45
or 37
(Brewster’s angle), while the collimation of emission light is performed by an objective perpendicular to the chip plane [23–26]. The second (ii) uses the same objective to focus and collect the light, and a dichroic mirror for wavelength separation [27–34]. For fast operation and fast optical alignment, all of the optical components of these (epifluorescence) systems can be incorporated into a fluorescence microscope [35–39]. A pinhole (ranged from 40 mm to 2 mm) is commonly placed at the focal point along the light path to obtain spatial filtering of the emission and to reduce scattered and background light. The PMT has become the most widely used detector in LIF; it gives excellent sensitivity, a wide dynamic range and a high detection frequency. With the increasing separation speed obtained on microchips, peak widths are growing smaller and smaller. To monitor separations with peak widths below 500 ms, the very high detection frequency provided by PMTs is needed. Flavin metabolites are a rare example of analytes that possess native fluorescent properties and can be excited in the visible area of the electromagnetic spectrum. They have therefore frequently been used as model analytes for LIF detection [40,41]

Detection in Miniaturized Analytical Systems

217

However, if excited with UV light, a large number of biological compounds are fluorescent, especially proteins containing tryptophan or tyrosine. The ability to exploit these useful native properties is hindered by the fact that optical components are commonly made of glass, which makes them nontransparent to light with a wavelength below 330 nm [42]. Incorporating optical fibres into the microchip is one way of simplifying the detection system, by minimizing the number of required optical components. An optical fibre coupled to a blue diode-pumped laser [43,44] has been incorporated into a microchannel network. This channel ended 190 mm away from the separation channel and enabled excitation without the need for any focusing optical components. Beneath the chip, a 400 mm pinhole, a holographic notch filter (476 nm) and an interference filter (535 nm) were attached. The emission light was then detected directly with a PMT from below the chip without any collimating optics. Fluorescein isothiocyanate (FITC)-labelled epinephrine and dopamine were used as model analytes. A natural extension of a dual-wavelength detection approach using integrated optical fibres has been presented by Lee et al. [45]. A glass/PDMS hybrid chip with two sets of integrated optical fibres facing each other (Figure 6.2(b)) has been proposed. Each of the two channels was excited by a different laser source, enabling two different wavelengths to be produced (488 and 632 nm). The two optical fibres were connected to two PMT modules equipped with appropriate bandpass filters to exclude excitation light. Although no separation was performed, a protein sample plug (BSA) labelled with two different fluorescent dyes (FITC and Cy5) could be observed with the two-wavelength detection setup. LOD of 200 ppm for labelled BSA is claimed. An interesting way of reducing the background signal caused by scattered excitation light is presented by Fang et al. [46]. They use a LIF detection system based on an orthogonal optical arrangement, as shown in Figure 6.2(c). Using an excitation beam perpendicular to the chip, highly sensitive detection was achieved by collimating the emission light from the microchannel through the sidewall of the chip. A special microchip layout was used to bring the separation channel in close proximity with the sidewall (1.5 mm), which was intensively polished to form a highly transparent surface. The emitted light was collimated with a microscope objective and focused on to a 900 mm pinhole. Two filters excluded the remaining excitation light, while the actual detection was performed by a PMT. The amount of scattered light, as well as the fluorescence emission, was found to be dependent on the viewing angle in the plane with the chip. Studies of both intensities with regards to the viewing angle revealed the most favourable ratio between the fluorescence signal and the scattered light at a collection angle of 45
, which was then used for further measurements. The limit of detection for fluorescein was 1.1 pM. The performance of the system was further demonstrated by the separation of FITClabelled amino acids. In addition, an approach to the simultaneous LIF detection of multiple capillary electrochromatography (CEC) separations was proposed by Lin

218

Miniaturization of Analytical Systems

and coworkers [47]. They developed a microchip with four parallel separation channels, enabling the parallel detection of four different analyte solutions. 6.2.1

Lamp-based Fluorescence Detection

Because lamp-based excitation is the most common method used with fluorescence microscopy to image biological samples, these light sources can be applied very easily to the detection of microchip separations using commercially available microscope setups. The common epifluorescence microscope uses condensing optics to collimate and parallelize the light generated by the lamp. After reflection by a dichroic mirror, the light is focused with an objective on to the microscope stage. Fluorescence emission is collected by the same objective and is usually detected by a PMT after encountering a dichroic mirror and an emission filter. Following LIF detection, lamp-based approaches form the second largest group of optical detection techniques used for microchip separations. Two different light sources can be found. High-pressure xenon arc lamps exhibit a rather homogeneous emission spectrum ranging from the UV up to near-IR light. Mercury lamps, in contrast, show the typical line spectrum of excited mercury. The line intensity ratios and the underlying background are dependent on the pressure in the applied mercury lamp and can result in a fairly high radiation output at several fixed wavelengths. If, however, the lines do not match the excitation wavelength of the respective analyte, the flexibility of the broad spectrum of the Xe lamp can deliver more favourable results. Several examples of lamp-based excitation using epifluorescence microscopes combined with PMT detection have been reported [48–51], including the detection of subsecond chiral separation [52] and the fast separation of amino acids in green tea [53] and FITC-labelled amino acids [54]. 6.2.2

Fluorescence Excited by Light-emitting Diodes

Light-emitting diodes (LEDs) have recently been introduced as an inexpensive and powerful alternative light source for fluorescence detection. The new generation of LEDs exhibit very high output power covering the whole visible wavelength range and even UV light. LEDs are the most effective light sources available and require only low-power driving currents. These advantages, combined with their very compact dimensions, mean that they are perfectly suited for integration into miniaturized devices. However, the half-bandwidth of the light emitted from LEDs often exceeds 20 or 30 nm, which implies the need for additional filters and frequently generates a higher level of background signal. Since the photons in a high-output LED are usually generated over an area of a few square millimetres, LEDs are not considered to be point light sources. This fact, combined with the divergence of the emitted

Detection in Miniaturized Analytical Systems

219

light, means that sophisticated collimation and focusing optics are needed in order to make use of the whole radiation power of the LED. An approach to LED-induced fluorescence detection with an integrated LED and optical fibre was presented by Uchiyama and coworkers [55]. The distance between the diode and the detection area plays a major role in the divergence of light emitted from LEDs. By separating an LED from its epoxy lens, the authors yielded an LED with a flat surface, which was incorporated into the microchip during fabrication. Although no focusing optics were necessary due to the extreme proximity of the LED, the broad emission spectra of the LEDs used made the introduction of a thin excitation filter necessary. An optical fibre was moulded into the PDMS chip at a right angle to the microchannel. While a shorter distance to the separation channel would increase the amount of emission light detected, it was found that an unwanted deformation of the channel walls occurred if the optical fibre was placed too close. The optimum distance was found to be 100 mm. The emission light was guided through the optical fibre through a bandpass emission filter and was detected by a PMT. LOD was significantly improved compared to previous results [56]. For fluorescein and rhodamine dyes, they were found to be in the mid-nanomolar range, but a microscope-based LIF system gave a 20-fold improvement in sensitivity. This is explained by the fact that all of the laser light is easily focused on to the microchannel, while despite its close proximity, only 25% of the emitted light of the LED could be used for detection. Further improvements in sensitivity are anticipated when an LED with a smaller light-emitting face and optimized driving conditions is used. 6.2.3

(Electro)chemiluminescence Detection

Chemiluminescence (CL) has proven to be a very sensitive and selective detection method for common CE separations. Because the light is generated by a chemical reaction, no excitation light source is required and a filter system to reduce the background is not necessary. This simplified setup is obviously attractive when attempting to integrate CL detection on to microchips. However, because the CL reagent needs to be mixed with the separated analytes before detection, a more complex microchip layout is required. Liu et al. [57] developed several chip layouts and evaluated their performance with respect to different CL model systems, including the metal ion-catalysed luminol–peroxide reaction and the dansyl species-conjugated peroxalate–peroxide reaction. The separation of Cr(III), Co(II) and Cu(II) ions as well as the chiral recognition of dansyl–phenylalanine enantiomers could be accomplished within one minute with low micromolar and even submicromolar LOD. A comparison between the different chip layouts showed that the pattern with the Y-shaped junction was preferred by the luminol–peroxide system, while a V-shaped junction yielded better results with the peroxalate– peroxide system. The V-shaped design was later applied to the submicromolar

220

Miniaturization of Analytical Systems

detection of ATP and ATP-conjugated metabolites using a firefly luciferin– luciferase bioluminescence system [58]. Wang and coworkers applied an electrochemiluminescence (ECL) detection method to a microchip separation. In ECL, an electroactive compound, in this case tris(2,20 -bipyridyl)ruthenium(II) ([Ru(bpy)3]2þ), is oxidized by applying avoltage to additional electrodes in the separation channel, and it subsequently reacts with the analytes with the emission of photons. In the work presented, thin-film indium tin oxide (ITO) electrodes were added during microchip fabrication [59]. The transparency of the ITO material makes these electrodes superior to the more widely used platinum-based electrodes, because none of the photons generated are absorbed or deflected. The method developed was applied to the detection of proline and the determination of lincomycin in urine down to a concentration of 9 mmol/l [60].

6.3 Electrochemical Detection (ED) ED is the most important alternative to the LIF detection route in analytical microsystems, due to its inherent features of miniaturization without loss of performance and its compatibility with advanced micromachining. High sensitivity and responses not dependent on optical path length or sample turbidity are additionally very attractive features. Also, many more compounds exhibit significant electroactivity than native fluorescence. The detector design should ensure well-defined mass transport, minimal band broadening and electrical isolation (decoupling) from the high separation voltage (typically 1–5 kV). The latter property is attributed to the fact that the current associated with the high separation voltage is usually several orders of magnitude larger than that measured at the electrochemical detector (ED). High sensitivity, selectivity (via the applied potential and electrode material), simple handling, longterm stability and rigidity are additional requirements. To meet these requirements, some different and very creative approaches have been published in the relevant literature [61–64]. In a conceptual way, the main approaches proposed take into account the relative position between both working electrode-separation channels [62], where the configurations can be classified as: end-channel, in-channel and off-channel detection (Figure 6.3(a)). In end-channel detection (both externally mounted and internally mounted), the electrode is placed just outside the separation channel. For in-channel detection, the electrode is placed directly in the separation channel. Offchannel detection involves grounding the separation voltage before it reaches the detector by means of a decoupler. In addition, taking into account the relative position between both working electrode-flow directions [64], we can find three configurations, as shown in Figure 6.3(b): flow by, where the direction of flow is parallel to the electrode

Detection in Miniaturized Analytical Systems

221

Figure 6.3 (a) Common configurations of EDs for CE microchips, based on different capillary/ working-electrode arrangements: end-channel (external or internal), in-channel and off-channel. (Reproduced with permission from [7]) (b) Common configurations of EDs for CE microchips, based on different positions of the electrode relative to the flow direction: flow by (parallel using two plates); flow on to (with the surface normal to the flow direction); flow through (with the detector placed directly on the channel exit). (Reprinted from [64] with permission from Elsevier)

surface (which means using two plates); flow on to, where the electrode surface is normal to the flow direction; and flow through, where the electrode is placed directly on the channel exit. 6.3.1

End-channel Detection

End-channel detection involves the alignment of the electrode at the end of the channel (tens of micrometres) and means that the electrode is outside the channel. Separation voltage has a minimal influence on the potential applied in an ED because most of the voltage is dropped across the channel [62,64,65]. The main advantage of this design is that no decoupler is necessary, and the whole system is simpler and more rugged because the channel is all in one piece. However, the main disadvantage is a loss of separation efficiency due to the relatively short length of the channel and the large distance between the end of the channel and the working electrode. This separation distance is also crucial for the noise signal obtained and can lead to a complete loss of the analytical current. Next, the most relevant and representative CE-ED microchip end-channel configurations will be described. Figure 6.4 shows selected CE-ED microchip end-channel designs.

222

Miniaturization of Analytical Systems

Figure 6.4 Selected electrochemical end-channel designs for miniaturized systems. Mathies’ design: integrated electrochemical detection cell: (a) full chip view; (b) expanded view of the integrated ED; (c) scanning electron micrograph of the detection region showing the location of the working electrode in the exit channel. (Reproduced from [66] with permission of ACS, Copyright 1998); Wang’s design: chip, chip holder and electrochemical cell. (Reproduced from [67] with permission of ACS, Copyright 1999); Lunte’s design: full chip view. (Reproduced from [71] with permission of ACS, Copyright 2000); Chen’s design: chip, chip holder and electrode alignment. (Reproduced from [72] with permission of ACS, Copyright 2002)

Mathies’group described the first CE-ED microchip[66] (Figure 6.4), showing the integration of microfabricated CE on glass substrates with ED in only one microdevice. Photolithographic placement of the working electrode just outside the exit of the electrophoresis channel provided high-sensitivity ED with minimal interference from the separation high voltage. In this approach, flow by was the configuration used in relation to flow direction-detector position. The suitability of the microsystem was illustrated, giving examples of separations of neurotransmitters in less than 100 seconds with a good resolution, and detection limits of 3.7 mM for dopamine.

Detection in Miniaturized Analytical Systems

223

Wang and coworkers (Figure 6.4) have reported a glass CE-ED microchip system that utilizes a single working electrode using end-channel configuration. This configuration has been described by the coupling of a single carbon ink screenprinted working electrode [67] (flow on to configuration), and a sputtered gold electrode (not shown in the figure) [68] (flow through configuration). A Plexiglas holder was designed to hold the microchip and to integrate the detector and reservoirs. The ease of replacement/alignment of the screen-printed working electrode is one of the most attractive features, as is the adaptation to other complex microchip formats and other electrodes using the same electrochemical cell detection/Plexiglas setup. The reliability of the new setup was demonstrated using model analytes (dopamine and catechol) and nowadays its versatility has been widely demonstrated. The same external end-channel configuration has been extended by the same group using other polymer microchips (PMMA [69], PDMS [70]) with excellent results. The LOD obtained for model analytes, such as epinephrine and catechol, by using the PMMA end-channel configuration was lower than that obtained in glass chips and PDMS-based carbon electrodes. These improvements were attributed to effective isolation of the working electrode from the high separation voltage (noise decreasing) and to the enhanced mass transport of analyte due to the combined flow on to/flow by flux. However, these materials have been used to a much lesser extent than glass because of the excellent analytical performance of glass microfluidic chips. The third end-channel configuration (Lunte’s design, Figure 6.4) to discuss consists of a top layer of PDMS containing the separation and injection channels and a bottom glass layer on to which four gold detection electrodes have been deposited. It was also the first report of dual-electrode ED in a PDMS microchip format [71]. The performance of the chip was evaluated using catechol as a test compound (detection limits were of 4 mM). Also, the benefit of the dual detection in terms of detection selectivity was illustrated in the detection of an unresolved mixture (ascorbic acid and catechol) by application of suitable detection potentials. End-channel detection with a replaceable microelectrode has also been proposed [72], involving fixing the working electrode (carbon fibre, Pt, Au) through a guide tube on the end of the separation channel (Chen’s design, Figure 6.4). The alignment of the electrode was carried out accurately and reproducibly, allowing a low noise and good reproducibility with detection limits for dopamine of 2.4 · 107 M. The interesting thing about this configuration is that this kind of electrode can be fabricated in general laboratories and can be replaced quickly with electrodes of different materials. To overcome the limitations of end-channel detection systems, a microfabricated CE chip containing an integrated sheath-flow ED has been proposed [73]. In this approach, the sheath-flow channels join the end of the separation channel from each side, and gravity-driven flow carries the analytes to the EDs placed at different working distances from the separation channel exit.

224

6.3.2

Miniaturization of Analytical Systems

In-channel Detection

In-channel strategies involve the placement of the working electrode directly within the separation channel. The analytes migrate over the electrode while still confined to the channel, thus eliminating the band broadening often observed with endchannel alignments [64]. Martin and coworkers [74] have reported work, which involves the development of an electrically isolated potentiostat, that has made it possible to place the working electrode in the separation channel. The in-channel configuration concept versus conventional end-channel configuration is shown in Figure 6.5(a). The miniaturized compact potentiostatic system (4 · 9 · 2 cm) holds promise for integration of the entire detection system on to a chip. The in-channel configuration helps to eliminate some of the negative separation performance characteristics encountered with the end-channel configuration, especially with respect to the alignment of the working electrode at the end of the separation channel and to separation efficiencies. In fact, this configuration improved the separation efficiency (number of plates, by a factor of 1.3) and exhibited similar separation performance to those found for LIF in terms of plate height and peak symmetry. A simple and sensitive electrode design for CE-ED microchip has been presented by Liu et al. [75]. The new design employs metal microwires as the working

Figure 6.5 (a) In-channel configuration of electrophoresis microchip with integrated electrochemical detection: chip layout (top) and electrode alignments used in these studies (bottom). (Reproduced from [74] with permission of ACS, Copyright 2002); (b) Off-channel configuration of electrophoresis microchip with integrated electrochemical detection: layout of the electrophoretic chip, a photograph of the microfluidic chip and an SEM micrograph of the PDMS substrate around the cross-region. (Reprinted from [79] with permission from Elsevier)

Detection in Miniaturized Analytical Systems

225

electrodes for ED. This approach aligns a solid metal microwire through the separation channel, allowing integration of multiple electrodes and the use of different electrode materials (copper, gold and platinum). The size of the working electrode and the relative position of the flow with respect to electrode surface are also discussed. As influent variable, electrode size had an important influence on the separation efficiency and the detection limits obtained. In fact, the lowest detection limit reported to date for dopamine (100 nM) was achieved without the use of a decoupler, as a result of the higher collection efficiency with the new design. 6.3.3

Off-channel Detection

Off-channel detection has also been employed to overcome the drawbacks of endchannel detection. Electrode placement in off-channel detection is similar to that in in-channel detection (Figure 6.3), but the separation voltage is isolated from the amperometric current through the use of a decoupler. Conceptually speaking, the decoupler effectively shunts the separation voltage to ground, and a field-free region is created where analytes are pushed past the electrode by EOF generated before the decoupler. This strategy has been used less than end- and in-channel configurations [76,77]. One of the strategies proposed uses a microchip composed of two different polymers, while the decoupler is made by creating holes in the top layer. This allows the placement of the working electrode directly in the separation channel after the decoupler [76]. Another strategy uses a palladium metal film decoupler in a Plexiglas-based microchip [77]. Other new mixed designs using microfabrication/decoupler strategies have also been reported. Wu et al. [78] have described for the first time the use of a microfabrication technique to simultaneously integrate a three-electrode ED and an electric platinized decoupler with an oxygen plasma-treated PDMS layer containing a CE channel. This work combines the microfabrication advantages with the decoupler concept. At first, the material of all of the electrodes was Au. Platinum nanoparticles were electrically deposited on the ground electrode and the reference electrode, in order to make them the decoupler and the pseudoreference electrode, separately. The decoupler was located in front of the three-electrode detector to isolate the interference from high-voltage separation/detection potential. Also, Lai et al. [79] have described an amperometric detector including an in-channel dual electrode with a palladium film decoupler. The microchip cleverly integrates a Pd-film decoupler just before the two Au working electrodes. This design, integrating the off-channel approach, is shown in Figure 6.5(b). In addition, the new development of a cellulose acetate decoupler for on-column ED in a CE microchip has also been described [80]. The decoupler was constructed by aligning a series of 20–30 mm holes through the cover plate of the microchip with the separation channel. Then a thin film of cellulose acetate was cast within the holes. Detection limits of 25 nM were achieved for dopamine, revealing the

226

Miniaturization of Analytical Systems

excellent isolation of the detection circuit for the separation currents. Lunte’s group [81] described the fabrication and evaluation of a fully integrated palladium decoupler/ED using a hybrid glass/PDMS device. Both a Pd decoupler and a working electrode were deposited on to a glass plate prior to the reversible sealing of the glass layer to a PDMS substrate containing the separation channel network. The effect of the Pd decoupler size on its ability to remove hydrogen was also investigated. 500 mm was determined to be the optimum decoupler size and the microsystem was evaluated using dopamine as the model analyte. To date, the work reported has primarily focused on the miniaturization of individual components such as the fluidic circuit or the electrodes rather than on a holistic approach aimed at miniaturizing the entire system. In order to create true lab-on-a-chip systems (see Chapter 8 for an explanation of the concept), the electronics as well as the microfluidic circuit and detection scheme need to be miniaturized into a compact system. In this way, another contribution has presented an overview of this approach (total miniaturization) and, in particular, has demonstrated some of the unique advantages of totally microfabricated systems [82]. This contribution deals with a multidisciplinary approach in order to develop a self-contained and transportable CE-ED system that (i) incorporates all necessary electrodes directly on the chip and (ii) utilizes miniaturized supporting electronics designed specially for the purpose of supporting CE-ED microchips designs. Finally, an illustrative example revealing the inherent benefits of ED (inherent miniaturization and compatibility with microfabrication techniques) has been given in [83]. A detailed outline/illustration delineating the processing steps for the microfabrication of capillary electrophoresis (CE) with integrated ED platforms for analyte separation and detection has been proposed. The processing steps outlined are appropriate for the production of lab-on-a-chip prototypes using obtained glass substrates and common microfabrication techniques. Microfabrication provides a major advantage over existing macroscale systems by enabling precise control over electrode placement and integration of all required CE and ED electrodes directly on to a single substrate with a small footprint. In the processing sequences presented, top and bottom glass substrates are photolithographically patterned and etched using wet chemical processing techniques (see Chapter 5). The bottom substrate contains seven electrodes required for CE/ED operation, whereas the top substrate contains the microchannel network. The flush planar electrodes are created using sputter deposition and liftoff processing techniques. Finally, the two glass substrates are thermally bonded to create the final lab-on-a-chip device (see Figure 6.6). In summary, the two general strategies for integration of electrodes into CE microchips are end-channel configuration and microfabrication. End-channel configurations allow electrode cleaning and use of chemically modified electrodes; however, these designs generally lack portability and the ability to incorporate

Figure 6.6 Integration of fabrication of electrochemical cell and microchip channels: (a) mask and microchip layout; (b) microfabrication steps for electrodes and microchannels. (Adapted from [83] with permission from Humana Press)

Detection in Miniaturized Analytical Systems 227

228

Miniaturization of Analytical Systems

multiple electrodes. Microfabrication allows the incorporation of multiple electrodes on-chip and is easily made portable, but requires the use of expensive metallization and cleanroom facilities, and integration of more than one electrode material is challenging. To overcome these drawbacks, in-channel and off-channel approaches are useful alternatives. 6.3.4

Electrode Materials

The detector performance is strongly influenced by the material of the working electrode, since this is where the electrochemical reaction of the analyte occurs. The selection of the working electrode depends primarily on the redox behaviour of the target analytes and the background current over the applied potential region. One important aspect is that selectivity can be obtained through a judicious choice of the working electrode material and the applied potential. Native materials and surface modifications have been explored in the EDs used in microfluidic devices. Carbon, platinum and gold have been used as electrode materials [61–64]. Carbon (including carbon paste or ink, and glassy carbon) is the most widely used material, due to its minimal fouling and lower overpotential, larger potential window and low background. Carbon-based (carbon fibre) [84] and carbon paste electrodes in PDMS-based chips for the detection of amino acids [85] have also been employed. Wang’s group [67], since the introduction of carbon ink screen-printed electrodes, has reported a large number of possibilities using this type of electrode. The electrodes are printed using patterned stencils. The working electrodes are deposited on a ceramic plate and cured. Contact silver ink is applied to overlap the carbon. An insulating layer is printed to cover the carbon–silver junction and is used to define the electrode area. A large number of analytes have been studied, from model analytes (dopamine and catechol) through explosives and pollutants to a wide range of analytes of biological/clinical significance. A technique termed micromoulding of carbon inks for use in microfluidics has also been reported [86]. This approach uses PDMS microchannels to define the size of the microelectrode. First, PDMS microchannels of the approximate dimensions for the microelectrode are made using soft lithography. The PDMS is then reversibly sealed to a substrate and the microchannels are filled with carbon ink. After a heating step the PDMS mould is removed, leaving a carbon microelectrode with a size slightly smaller than the original PDMS microchannel. The resulting microelectrode can be reversibly sealed to a PDMS-based flow channel. Platinum electrodes are useful for detecting peroxides, alcohols, aldehydes and hydrazines [87]; gold (sputtered and disk electrode) has also been used as an electrode material [68,88,89]. The power of amperometric detectors can be enhanced through a deliberate modification of the electrode surface. Electrocatalytic surfaces can be used to

Detection in Miniaturized Analytical Systems

229

accelerate the electron-transfer reactions of species with slow electron-transfer kinetics. A palladium-coated screen-printed carbon electrode for the electrocatalytic detection of various hydrazines [90], a palladium layer that allows a significant (>0.3 V) lowering of the applied potential and leads to sharper and larger peaks, and a cobalt-phthalocyanine (CoPC)-modified carbon-paste detector for the electrocatalytical detection of thiols [85] have all been proposed. Other attractive routes include selective transport or preferential accumulation, in connection with permselective coatings and preconcentrating surfaces, respectively. Wang’s group has used a chemically vapour-deposited boron-doped diamond film band electrode for end-channel amperometric detection [91]. This material offers enhanced sensitivity, lower noise levels and sharper peaks for several groups of important analytes such as phenols, organophosphate nerve agents and nitroaromatic explosives. Detection limits of 70 and 110 ppb were obtained for explosives using this electrode material. The performance of pyrolysed photoresist carbon films used as electrode material for ED in microchips has been described [92]. The most attractive features of this material are an exceptionally smooth surface (similar to polished glassy carbon) and a low oxygen:carbon ratio, which enables low background current levels. The pyrolysed photoresist carbon films have been introduced as planar carbon electrodes in PDMS–quartz hybrid microchip devices and their utility has been demonstrated by the separation and detection of various neurotransmitters. This microchip scheme, coupled with sinusoidal voltammetry detection, has allowed a detection limit of 100 nM for dopamine. Without question, nanotechnology also offers valuable tools for microfluidics. Indeed, analytical nanotechnology is the natural solution to well-known microfluidic challenges such as selectivity (peak capacity) and in particular sensitivity, which is the main challenge in microfluidics, as discussed in the introduction to this chapter. The new generation of analytical microsystems should take into account the nanotechnology tools integrated in its architecture. Apart from their high sensitivity and inherent miniaturization, they offer the opportunity to modify surfaces with nanomaterials, enabling us to cross frontiers in new applications. An excellent example of this is carbon nanotubes (CNTs). CNTs are a new group of nanomaterial with unique geometrical, mechanical, electronic and chemical properties which offer notably exciting possibilities involving the large active surface at electrodes of small dimensions, the enhancement of electronic transfer and strong sorption capacity [93]. These properties have a clear influence on analytical sensitivity, which is enhanced by the use of these nanomaterials. There are two main types of carbon nanotube that can have high structural perfection: single-walled carbon nanotubes (SWCNTs), which consist of a single graphite sheet seamlessly wrapped into a cylindrical tube, and multiwalled carbon nanotubes (MWCNT), which comprise an array of such nanotubes concentrically nested like the rings of a tree [94–96]. Such advantages were illustrated for the first time [97] in connection

230

Miniaturization of Analytical Systems

with several types of hydrazine, dopamine, catechol and ascorbic acid, phenol and purine. However, microfluidic devices using CNT materials (SWCNTs and two MWCNTs) for real-world analysis involving detection of a wide group of analytes (dietary antioxidants, water-soluble vitamins, vanilla flavours and isoflavones) have rarely been proposed [98,99]. Ultrafast separations resulted in well-defined and resolved peaks with enhanced voltammetric current in comparison with those obtained from unmodified electrodes, making CNTs an ideal material for electrochemical sensing in real analysis (see Chapter 8). Peak efficiency and even resolution were dramatically increased, enhancing in consequence the peak capacity of these microsystems. 6.3.5

Modes of Detection

Of the three general modes of ED, amperometry, conductimetry and potentiometry, amperometry is the one most used in the microchip format [62,64]. Conductimetry has also been extensively investigated for microchip-based separations [63], while potentiometry is not implemented and is not discussed here. Amperometry/Voltammetry Amperometry is accomplished by measuring the resulting current after applying a constant potential to the working electrode. A reference and auxiliary electrode are also present. Amperometry mode has the advantage of generally good detection limits but is restricted to electroactive species. Selectivity is achieved through a judicious choice of the detection potential, with the optimal detection potential normally determined by hydrodynamic voltammetry. It is the most widely reported electrochemical method for microchannel-based separations (see also the designs previously described in which amperometry was used as the detection mode). However, other variations have been proposed. Sinusoidal voltammetry (SV) employs a large amplitude sine wave as the excitation waveform. The resulting faradic current response is nonlinear and contains both the signal components at the excitation frequency and the harmonics. The charging current associated with scanning the electrical double layer, which is the major background component, is primarily linear and therefore resides predominantly at the excitation frequency. There is an increase in the discrimination between signal and background at the higher harmonics, resulting in greater sensitivity and selectivity of the amperometric response [100,101]. Sinusoidal voltammetry was found to be more sensitive (nM) than constant potential amperometry (mM) for the neurotransmitters studied (dopamine, isoproterenol and L-dopa). Moreover, pulsed amperometric detection on a hybrid PDMS–glass microchip with a platinum working electrode for the ED of glucose, maltose and xylose has been reported [102]. The direct amperometric detection of carbohydrates at noble metal electrodes is characterized by electrode fouling. Carbohydrate detection at a

Detection in Miniaturized Analytical Systems

231

platinum electrode involves adsorption and subsequent anodic dehydrogenation of the molecule at the electrode surface. However, the adsorption of additional molecules quickly becomes limited, as the dehydrogenation product remains adsorbed on the surface of the electrode. As the surface becomes contaminated with the reaction product, the signal rapidly decreases. Pulse amperometry maintains electrode activity via a triple pulse sequence including detection, electrode cleaning and electrode reactivation potentials. Cleaning of the electrode surface via oxidative removal of the adsorbed hydrocarbons occurs at a sufficiently positive potential, while regeneration of the electrode surface takes place at a negative potential. Pulse amperometric detection has also allowed the separation and detection of underivatized carbohydrates, amino acids and sulfur-containing antibiotics in CE microchip format [103]. Although the utility of amperometry has been widely demonstrated, a unique concept involving coupling voltammetry on microfluidic chip platforms has also been reported [104]. CE microchips can be understood as ideal platforms for the performance of microscale voltammetric analysis. This concept deals with the introduction of voltammetric protocols as detection/electrochemical characterization after pumping the analyte using the CE microchips as an ‘injection/sample preparation/liquid manipulation system’, because it allows the use of ultrasmall detection volumes. This approach is advantageous over nanovial voltammetry, which lacks the sample preparation, liquid-handling and fluid control or manipulation capabilities. But the most attractive feature is that this coupling enhances the power of microchip devices, as it adds a new dimension of analytical information and opens up highly sensitive detection schemes. The utility of linear scan, hydrodynamic modulation, rapid square-wave and adsorptive stripping has been demonstrated using catechol, hydrazine compounds and nickel as model analytes and screen-printed detectors. Conductimetry Conductimetric detection on microfabricated devices has also been developed [63,105–115]. Conductimetric detection is a less sensible but universal detection technique that has been applied as a detection mode in microchip format in both galvanic (a pair of electrodes is placed in the separation channel for liquid impedance measurement) [105,106] and contactless (no contact between the pair of electrodes and separation channel solution) [107–115] modes. Contactless detection is preferred for three reasons: (i) the electronic circuit is decoupled from the high voltage applied for separation (no direct DC coupling between the electronics and the liquid in the channel); (ii) the formation of glass bubbles at the metal electrodes is prevented; and (iii) electrochemical modification or degradation of the electrode surface is prevented. Both contact and contactless detectors integrated into a microchannel require physical connection to read out electronics placed

232

Miniaturization of Analytical Systems

Figure 6.7 Conductivity designs on microchip platforms. (a) Configuration of electrophoresis microchip with integrated galvanic conductivity detection. (Reproduced from [105] with permission of Wiley-VCH); (b) Configuration of electrophoresis microchip with integrated contactless conductivity detection. Microchip (top) and enlarged view of the detector (bottom). (Reproduced from [110] with permission of ACS, Copyright 2002)

inside or even outside the microdevice. These connections, however, should not influence the bonding process or the fluid flow in the microchannel. Figure 6.7 shows two representative designs involving both detection formats, galvanic (a) and contactless (b). In each strategy, microfabrication also plays an important role in the integration of conductivity detectors. On the one hand, with respect to the galvanic mode, van der Berg and coworkers have developed glass chips for zone electrophoresis and demonstrated the possibility of detecting inorganic ions and fumaric, malic and citric acids with embedded electrodes (see Figure 6.7(a)) [105]. Also, Spoer and coworkers [106] have designed a device in PMMAwith Pt wire electrodes integrated and have shown the separation and sensitive detection of amino acids, proteins and DNA fragments by a built-in conventional contacting conductivity detector using CZE, MECK and electrochromatography, respectively. On the other hand, Dedem et al. carried out one of the first and most interesting reports involving contactless conductivity detectors [106]. In this work, a contactless capacitively coupled conductivity detector integrated within a CE microchannel was presented and evaluated. The four-electrode conductivity measurement made use of two outer and two inner electrodes. Capacitive coupling with the liquid inside the channel was obtained by covering the electrodes with a thin dielectric layer such as silicon carbide. Other works dealing with integrated contactless conductivity have also been reported [107–109]. Special attention should be paid to Pumera and Wang’s design [110] because their approach allows a significant simplification of the manufacturing process (see Figure 6.7(b)).

Detection in Miniaturized Analytical Systems

233

Indeed, they have demonstrated contactless detection using another design by placing electrodes on top of a cover plate rather than embedding them in the microdevice. The new contactless conductivity microchip detector is based on placing two planar-sensing aluminum-film electrodes on the outer side of a PMMA microchip and measuring the impedance of the solution in the separation channel. The analytical suitability of this approach has been extended by other interesting analytical applications. First, the proposal to use dual detection (the combination of a contactless conductivity detector with an end-column thick-film amperometric detector) for simultaneous determination of both ionic (small cations) and electroactive (nitroaromatic explosives) species as well as for the confirmation of peak identity. Second, the suitability of the use of contactless conductivity detection for monitoring separated aliphatic amines in UV-absorbing solvents (dimethylformamide, dimethylacetamide, dimethyl sulfoxide, propylene carbonate media), revealing the suitability of this detector in nonaqueous media [112]. Third, the proposal to use a movable contactless-conductivity detector [113]. This system relies on positioning the detector at different points along the separation channel by ‘sliding’ the electrode holder. This enables rapid switching between ‘total’ (unresolved solutes) and ‘individual’ (resolved/fingerprints) signals by placing the detector at the beginning and the end of the separation channel, respectively. Hauser’s group has contributed important developments in conductivity detection. An excellent example of this is a work dealing with a high-voltage approach to capacitively coupled contactless conductivity detection for use on microfabricated planar electrophoresis devices [114]. The authors demonstrate that better sensitivity is obtained by placing the electrodes in troughs that allow tighter coupling to the separation channel (detection limits for small ions Kþ, Naþ and Mg2þ: 0.49, 0.41 and 0.35 mM, respectively). In this work, the detection of heavy metals such as Mn2þ, Zn2þ and Cr3þ was also described. The universal nature of the new design was illustrated by the detection of citric and lactic acids, which are of interest in food and beverage analysis, as well as the detection of three antiinflammatory drugs. A sophisticated and interesting on-chip design for contactless four-electrode conductivity detection for CE microdevices has been reported [115]. In this work, four contactless electrodes were deposited and patterned on the second glass substrate for on-chip conductivity detection. This new configuration improved the analytical performance over the classical two-electrode detection setup. The novel configuration was illustrated with inorganic ions (Kþ, Naþ and Liþ) and with six organic acids. Bai et al. [116] have proposed passive conductivity detection applied to CE microchips. The separation electrical field is used to generate a potential difference between two electrodes located along the channel. For constant-current electrophoresis, the generated signal is proportional to the resistance of the solution

234

Miniaturization of Analytical Systems

passing between electrodes. This principle takes direct advantage of the separation field and the signal is simply measured by a high-impedance voltmeter. The concept has been demonstrated by the separation of three alkali ions on a polymer microchip.

6.4 Microfluidics–MS Interfacing MS allows the ionization of intact molecules for a highly accurate determination of their molecular weight, making identification of molecules easier, and it is becoming a universal detection method. This detection route has become increasingly important as an analysis tool in the proteomics area in particular, because of its sensitivity. The two microfluidic–MS coupling methods mainly reported in the selected literature are ESI [11,117–127] and to a lesser extent MALDIMS [12,128–131]. 6.4.1

Microfluidics-based Electrospray Ionization (ESI) Sources

The mass spectrometer is a powerful analytical tool because of its high resolving power, ion isolation capability and ability to measure the mass-to-charge ratio for both parent and fragment ions. When thinking about MS, an image of a large bulky piece of instrument comes to mind. The seemingly unlikely marriage between MS and microfluidic devices is actually a good one because the mass spectrometer and the microchip are well matched, due to the similarity between the flow rates generated by the microchip and those required for electrospray ionization (ESI). Electrospray ionization mass spectrometry (ESI-MS) is concentration (not mass) sensitive. Therefore, miniaturizing the sample introduction system achieves the highest sample concentrations and the lowest detection limits. Microfluidic devices interfaced with ESI-MS provide a convenient, miniaturized sample preparation/ introduction system. Solutions are commonly transported through microfluidic channels by application of an electric potential (electrokinetically) or by application of pressure (hydrodynamically). Flow control on microfluidic devices can also be governed electrokinetically, although the development of reliable mechanical valves is desirable. As expected, because of the inherent properties of electrokinetics (as studied in-depth in Chapter 5), most systems interfaced with mass spectrometers are electrically driven. One critically important process is that of ionization. ESI has become one of the most versatile ionization methods for MS. Liquid, under the influence of a high electric field, is sprayed into a mass spectrometer, where the mass-to-charge ratio of the analyte is determined. The production of ions in the gas phase from the solution phase is critical for direct coupling of fluidic devices with the mass spectrometer, and the ESI sources are well suited for this application. Micro- and nanoflow ESI

Detection in Miniaturized Analytical Systems

235

interfaces are ideal for these applications as they cover flow-rate ranges from hundreds to a few nanolitres per minute, which are the same as the flow rates used by most microfluidic devices. A critical parameter is the diameter of the electrospray emitter and the flow rate. In early ESI experiments, electrospray emitters with relatively large internal diameters were used (usually on the order of 100 mm) at several microlitres per minute [132,133]. Droplets generated from such emitters are relatively large, requiring a large number of Coulombic explosions before gas phase ions are formed. Later, ‘microelectrospray’ was introduced, operating at sub-ml/minute flow rates with tip diameters around 50 mm [134]. Higher flow rates (ml/minute range) could be sprayed with a coaxially nebulizing gas, called ‘ionspray’ or ‘pneumatically assisted electrospray’ [135,136]. Mann and coworkers developed a novel way of performing electrospray, called ‘nanospray’, where small droplets are generated from emitters with an inner diameter (ID) of several microns [137,138]. Decreasing the ID of emitters has several advantages, including lower electrospray voltages, which allow the emitter to be positioned closer to the inlet of the MS. The number of ESIgenerated ions that are introduced to the MS is increased in this way, i.e. the yield of analyte ions that enter the MS is higher than in ESI performed from large diameter emitters. Nano-ESI consumes fewer samples and can be coupled online with nano-LC techniques due to the low flow rate. Pure water solutions can be sprayed without discharges, and higher ionization efficiency can be obtained due to the smaller initial ratio of droplets generated. Another important reason to use nanospray is that ESI-MS is a concentration-sensitive technique. This implies that the diameter of an emitter can be miniaturized and the flow rate reduced without losing mass spectrometric sensitivity [139]. An additional advantage of reducing the diameter of the emitter is that ion suppression effects are reduced when going from electrospray to nanospray. The number of analyte molecules per droplet in regular ESI is large and may lead to competition for charges when mixtures of analytes are present. In the case of nanospray, each droplet formed will contain on average a single or a few analyte molecules, eliminating ion suppression effects and making nanospray a very attractive technique. Another reason why nanospray is preferred by many researchers is that the number of charges available per analyte molecule is much higher than in electrospray. This enhances the probability of ionizing an analyte. A drawback of reducing the dimensions of the various components necessary for a successful nano-LC experiment is that the influence of very small dead volumes, which normally have a minor impact on micro-LC separations, becomes significant in nano-LC. The quality of the nano connections depends a great deal on the experience of the experimenter. Also, variations in tip diameter will have a significant influence on the appearance of the mass spectrum. Microtechnology makes it possible to fabricate chips with submicrometre precision and accuracy, and with reproducible minimal or zero dead volume connections. It has therefore

236

Miniaturization of Analytical Systems

become an important technology for nano-ESI and nano-LC-ESI experiments. Besides the absence of dead volumes, the main reason for most researchers to work with chips is that multiple functionality, such as pumps and valves, can be monolithically integrated on to a small device using small sample volumes (see Chapter 8 for a description of the integration concept). Ultimately, chips should be cheap to make and disposable in order to avoid cross-contamination. Small sample volumes (often sub-ml) are required, and as a consequence of the advantageous small dimensions of the channels, analysis can be performed faster, leading to increased sample throughput. Flow rates typically obtained in microfluidic channels are on the order of several tens of nanolitres per minute, compatible with nanospray. For these reasons, microfluidics offers researchers a route to new tools for the performance of experiments that were difficult if not possible to perform previously. Several methods have appeared in the literature for coupling a microfluidic chip to a mass spectrometer, using electrospray as the ionization technique. These can be roughly divided into three general approaches [127]. The first is monolithic, with electrospray being performed directly from microfluidic channels opened at the edge of a chip. Though this approach is simple, the position of the electrospray along the chip edge is often difficult to control. To overcome this problem, a second approach, which involves inserting fused-silica capillaries directly at the ends of channels in a chip, to serve as electrospray tips, has been proposed. However, this approach is technically difficult to realize, inspiring researchers to try still a third approach, which is to monolithically integrate emitters in the fabrication process. These three generations of chips are described in this section, together with some illustrative examples. The advantages and disadvantages of each approach will be discussed. Figure 6.8 shows the main schemes of the configurations, as reported in [127]. Electrospray Directly from the Chip The first groups to explore the possibility of coupling online analysis in microfluidic chips with MS were those of Karger et al. [140] and Ramsey and Ramsey [141]. Both reported, at about the same time, microfluidic chips that allowed ESI to be performed from the edge of the chip. These contributions correspond to the schemes shown in Figure 6.8(a) and (b), respectively. Karger et al. [140] reported a glass chip with nine microfluidic channels for direct infusion experiments. This device has 60 mm wide and 25 mm deep channels. The spray was formed directly from the edge of the chip using an external syringe pump to generate a pressure-driven flow of 100–200 nl/minute. One of the main issues with this device, however, was that the liquid leaving the microfluidic channel tended to spread along the edge of the chip.

Detection in Miniaturized Analytical Systems

237

Figure 6.8 Coupling configurations for microchip–MS. (Reproduced from [127] with permission of the Royal Society of Chemistry)

Ramsey and Ramsey used a chip with 10 mm deep and 60 mm wide channels in which pumping was performed electro-osmotically [141]. The chip contains a main channel with an exit in the edge of the chip where electrospray is generated. A side channel or arm is connected to the main channel near the exit. An ionic current high enough to induce an electro-osmotic flow (EOF) in the main channel is generated through the channels by applying a 4 kV potential difference between the buffer inlet reservoir and side-arm reservoir. EOF in the side channel is suppressed by a polyacrylamide coating that increases the surface viscosity and reduces the zeta potential of the glass walls. The coated side arm therefore cannot accommodate the entire EOF generated in the uncoated main channel, resulting in an increase in pressure at the intersection of the main channel and the side arm. The side-arm channel is longer than the remainder of the main channel to the channel opening, and thus has a higher resistance to flow. This, combined with the difference in EOF capacity of the two channels, results in an indirect pressure-driven flow toward the channel opening in the edge of the chip. This flow is high enough to obtain a constant electrospray current. A schematic overview of the electrical configuration of the chip is shown in Figure 6.8(b).

238

Miniaturization of Analytical Systems

Figure 6.9 Picture of microchip–MS coupling (left). Effects of droplet formation (right) (Reproduced from [141] with permission of ACS, Copyright 1997)

Using these strategies, large droplets are generated at the exit and cause excessive band broadening and sample dilution, precluding the use of this interface for onchip separations. Attempts to minimize droplet size have included coating, derivatization of the orifice with hydrophobic reagent (hydrophilic liquids have a low contact angle with glass and will tend to easily wet glass surfaces) or pneumatic assistance of the droplet formation. Figure 6.9 illustrates this effect in connection with the scale of one chip–MS coupling. Although the fabrication of these chips with spray orifices directly on the edge is relatively simple and truly monolithic in nature, several disadvantages to this approach make it unsuitable for a number of applications. First of all, hydrophilic liquids have a low contact angle with the hydrophilic glass, and therefore tend to spread along the edge of glass devices. The same is true for hydrophobic liquids, which wet hydrophobic chips made of plastic. This makes it difficult to predict the exact position from which the spray is generated. Small sharp objects, such as small splinters of substrate material left on the side of the chip after fabrication, experience a high local electrical field strength [142]. Moreover, the total volume of liquid wetting the edge of the chip is relatively large and may act as a dead volume to negatively influence the performance of separation chips. To avoid the spreading of hydrophilic liquids, hydrophobic coatings have successfully been tested [140,143], but were found not to be stable over time. Coatings are also not a solution to the wetting problem for on-chip separations performed in gradients of solvents with different polarities. A possible approach to overcome some of the problems described here is to close the exit of the microfluidic channel with a porous PTFE membrane from which small Taylor cones are generated [144]. Similarly, the

Detection in Miniaturized Analytical Systems

239

formation of a porous monolithic phase in the emitter can lead to stable electrospray from the end of the emitter [145,146]. Electrospray from an Attached Capillary Sprayer To overcome the problems that arise when spraying from the edge of a chip, several researchers have described connecting standard fused-silica emitters to microfluidic channels. Harrison and coworkers, following the scheme of Figure 6.8(c), were among the first to describe a novel approach integrating a fused-silica electrospray emitter with their chip [147]. Glass devices for CE separations were made and 200 mm holes were drilled in the edge of the chip using a tungsten carbide drill. The glue was also used to seal the connection between chip and needle. A capillary with an outer diameter of 185 mm and inner diameter of 50 mm was then inserted into the hole, connecting with the 13 mm deep and 40 mm wide microfluidic channel. The influence of the connection on the CE performance, i.e. the plate number, was studied by separating a mixture of fluorescently labelled amino acids using LIF as the detection technique. Two chips were compared, each with a different dead volume at the channel–capillary junction due to differences in the drill-bit geometries used. One of the drills was pointed and the other flat-tipped. When the hole was drilled using a pointed drill, a dead volume of 0.7 nl resulted (see Figure 6.10), which in turn meant broader peaks. Plate numbers for the separation of fluorescently labelled amino acids were measured to be 40 000 on the chip at the end of the

Figure 6.10 Example of electrospray from an attached capillary sprayer. (Reproduced from [147] with permission of ACS, Copyright 1999)

240

Miniaturization of Analytical Systems

microfluidic channel, i.e. the spray capillary was not taken into account. When the same compounds were separated on the device and detected at the end of the inserted electrospray needle, plate numbers around 15 500 were obtained, even though the separation length was almost doubled. Better results were obtained with the flat-tipped drill, where the inner capillary butted right up against the microfluidic channel, leading to a negligible dead volume. In this case, the plate numbers for the separation of the same compounds were 117 000 at the end of the microfluidic channel and 71 000 at the end of the inserted spray needle. This study demonstrates very clearly that the integration of electrospray emitters in chips requires special attention. When a similar chip was coupled to a mass spectrometer by this group for the analysis of low-nanomolar peptide mixtures [148,149], the configuration shown in Figure 6.8(c) was used. The above method for coupling emitters with microfluidic devices relies on applying the high voltage to a metallic coating on the emitter; the configuration schematically shown in Figure 6.8(c) was used [150]. In another approach, the electrical field is applied between one of the inlet reservoirs on the chip and the inlet of the mass spectrometer, as shown in Figure 6.8(d) [151]. Note that this works well for applying the electrospray potential, but when separations based on CE need to be performed, it will not work due to the electrical current limitation imposed by the ESI process. An exception to this has been demonstrated by van der Greef and coworkers, who obtained a stable spray and good separation of b-agonists by applying the potential at one of the inlet reservoirs, using the inlet of the mass spectrometer as the counter electrode [152]. They made use of a low-conductivity buffer that exhibits a large potential drop over the microfluidic separation channel, enabling separation by CE. However, it is often not desirable to use a lowconductivity buffer when performing separations. This is because a large potential difference between the end of the microfluidic channel and one of the inlets of the microfluidic device would be needed to perform a separation, resulting in the generation of heat. In other approaches, the electrical field is applied to a stainlesssteel tube connected to the exit of a chip through which the eluent is guided [153–155] or to electrodes integrated in the microfluidic channel, which may be made of gold [156–159], conductive carbon or conductive epoxy [160] (see Figure 6.8(e)). All these approaches have the advantage that compounds eluting from the column are directly electrosprayed without dilution. A different configuration consists of a liquid junction frequently making use of a sheath flow of makeup buffer. This approach has the advantage that the composition of the separation buffer can be modified to make it suitable for ESI. Also, when the flow rate from an external LC column is incompatible with the ESI source used, a makeup solvent can be added [161], though dilution of the analyte will take place. Most liquid junctions are made by inserting a spray capillary into the device and coupling it to an existing ion source [162–171]. Couplings where the potential is applied to a side channel in contact with the separation channel through a porous

Detection in Miniaturized Analytical Systems

241

(glass) membrane [172–175] or a dialysis membrane made of polysulfone [176,177] (see Figure 6.8(f)) have also been reported. The coupling of (nano)electrospray emitters with microfluidic channels improves the ESI-MS performance significantly compared to spraying from the edge of a chip. This approach has also allowed microfluidic devices to be coupled to commercial electrospray ion sources. However, several difficulties with using this approach should be mentioned. Dead volumes, having a negative effect on the separation performance, are easily introduced upon coupling, though they can be avoided by using complex drilling approaches involving gluing [178]. These drilling procedures may not be the best choice if a microfluidic device needs to be disposable and therefore cheap to fabricate. Manually inserting spray needles in chips is, from a labour-intensity and cost point of view, not the cheapest way to make devices. Additionally, though not reported, glue may release compounds that interfere in the mass spectrum under certain buffer/organic solvent conditions. The next section will describe a third approach to making chips where the electrospray emitter is made during one of the chip-fabrication process steps. Integrated Emitters As discussed in the previous section, one of the main problems with manual insertion of spray needles into microfluidic channels is that this approach does not lend itself to mass fabrication [160]. A complete study of this is outside of the scope of this chapter book; for details, see reference [127] and related. In brief, two approaches have demonstrated that chips can be made in a highly integrated fashion, where both microfluidic channel and emitter can be formed simultaneously during the same fabrication process, eliminating dead volumes. The first approach makes use of technologies that allow fabrication of the microfluidic channel and emitter in a single machining step, one chip at a time, i.e. serially. Examples are based on laser machining and milling. The second approach makes use of photolithographic processes that allow large numbers of structures to be made in a single step and can therefore be used to make many chips simultaneously. This approach is more parallel in nature. Examples of chips made using these approaches have been illustrated in several excellent works [179–187]. Two representative examples of microfluidic devices with integrated emitters are illustrated in Figures 6.11 and 6.12. High-throughput Analysis As we have already mentioned, the role of microfluidics–MS coupling in sample preparation is very useful. Its role in high-throughput analysis will now be briefly discussed. Parallel analysis for high throughput is of interest to the analytical chemist for economical reasons. However, a mass spectrometer can in principle handle only one analysis at a time. The high-throughput approach for MS analysis is shown in Figure 6.13 [153]. A way to work around this has been presented by

242

Miniaturization of Analytical Systems

Figure 6.11 Designs of electrosprays with integrated emitters I. (Reproduced from [180] with permission of ACS, Copyright 2003)

Figure 6.12 Designs of electrosprays with integrated emitters II. (Reproduced from [183] with permission of ACS, Copyright 2002)

Detection in Miniaturized Analytical Systems

243

Figure 6.13 Design for high-throughput analysis. (Reproduced from [153] with permission of ACS, Copyright 1998)

de Biasi and coworkers, who developed a four-channel electrospray interface for the analysis of the compounds eluting from four LC columns [188]. Four separations are performed simultaneously, but the eluates are analysed in a multiplexed manner. This can of course have the disadvantage that compounds eluting from one column may not be registered while mass spectra are being acquired from another column. Similar chip-based experiments in which multiple channels are monitored simultaneously using a mass spectrometer have not been performed. However, several studies do describe the use of multiple channels integrated on a chip, each connected to an individual nozzle, for high-throughput analysis in a serial fashion (one channel at a time). These devices have the particular advantage that cross-contamination is avoided, since each channel/nozzle is used for only one sample. The first chip with multiple channels/emitters, presented by Karger et al. [143], contained nine parallel channels, each spraying from an individual orifice at the edge of the chip. One impressive example was presented by Liu and coworkers, who connected each well of a 96-well plate to an individual microfluidic channel with manually inserted ESI needles to avoid cross-contamination [160]. Each well could be individually pressurized with nitrogen gas to induce flow in the microfluidic channels. The chip was positioned on a computer-controlled translational stage in front of a mass spectrometer. Flow through one microfluidic channel at a time was induced. By moving the chip with the stage, other reservoirs could be sequentially pressurized to induce a flow through the corresponding microfluidic channel and electrospray emitter. A stationary high-voltage electrode was positioned such that during the movement of the translational stage, the high voltage was connected only to the pressurized channel. Although sample cross-contamination did not occur, several channels were blocked during the manual insertion and gluing of the 96 ESI needles. 6.4.2

Microfluidics–MALDI-MS Interfacing

At the beginning of the MS section, we mentioned that microfluidic systems are an important component of the ongoing push toward miniaturization and integration of

244

Miniaturization of Analytical Systems

analytical platforms for proteome characterization. Effective coupling between microfluidics and MS is clearly a critical part of this trend. Interfacing with MALDI-MS is of particular interest as an offline approach to protein characterization, enabling on-chip processes to be effectively decoupled from back-end MS analysis, while also offering excellent sensitivity, wide mass range analysis and overall high throughput. However, as we have already mentioned, this interface has been used to a lesser extent than ESI, appearing in just one review [131]. In one sense, traditional MALDI target preparation already involves microfluidic processing, with nanolitre-scale sample deposition performed using advanced spotting tools to limit droplet volume on the target surface. Here we have limited ourselves only to those technologies which have been realized through the use of microfabrication techniques, including direct microfabrication of MALDI targets, the use of microfluidic systems for the achievement of improved off-chip MALDI target preparation and the integration of microfluidic systems and MALDI targets within a single substrate. Microfabricated MALDI targets are already having a realworld impact, with a variety of microstructured targets commercially available. From the perspective of integrated systems, the latter two categories are of greater interest. Combining microfluidic networks (for lab-on-a-chip functionality) with on- or off-chip MALDI target preparation offers substantial promise for integrated and high-throughput bioanalysis. The direct integration of microfluidics and MALDI targets within a single chip appears to be a particularly natural combination, since both technologies generally employ planar substrates. The growing interest in this level of integration is reflected in the range of approaches that have been demonstrated for delivery of analyte to on-chip MALDI target surfaces. Unlike ESI-MS, offline MALDI-MS employs an alternative soft ionization method that is typically applied to the analysis of solid phase analyte cocrystallized with an energy-absorbing matrix material on the surface of a supporting target plate. Modern MALDI-MS uses a nitrogen laser to ionize samples cocrystallized with an organic matrix component [189], with detection of generated ions typically performed using a TOF-MS instrument (MALDI-TOF-MS). Compared to ESI-MS, MALDI-MS tends to be more tolerant of salts and other sample contaminants, offers excellent detection limits and produces mass spectra which are relatively simple to interpret because of the absence of the multiple charge states seen in ESI-MS. Although generally limited to the analysis of higher molecular weight species because of the interference from matrix components below 500 Da, MALDI-MS can be applied to a wide range of sample masses, with good sensitivity above 300 kDa, compared to less than 100 kDa for ESI-MS. Following target preparation, MALDI-MS analysis is a high-throughput process that can be performed with minimal user training. From the point of view of microfluidic systems, the offline nature of MALDI-MS analysis represents a key advantage over ESI-MS. For example, the timescales for biomolecular separations and online MS data acquisitions are often incompatible. However, offline MALDI-MS enables on-chip sample processing steps to be fully

Detection in Miniaturized Analytical Systems

245

decoupled from the back-end MS analysis. In addition, offline MALDI-MS analysis provides a powerful method for coupling multiplexed parallel on-chip analyses with MS detection. While simultaneous ESI-MS from multiple microchannels is generally not practical, parallel deposition of samples from multiplexed microchannels on to MALDI target spots may be readily achieved. It should be noted that although MALDI-MS itself is a serial process, its high duty cycle enables reasonably high throughput for a large number of deposited samples. As we have already mentioned, technologies for coupling microfluidics with MALDI-MS analysis may be organized into three broad categories. The first category discussed here addresses microfabricated MALDI target plates for passive containment of nanolitre-scale sample volumes on the target surface. The second encompasses microfluidic systems designed to enable off-chip MALDI target preparation. The final category includes a number of microfluidic systems that have been developed to directly integrate on-chip MALDI targets into the microfluidic substrate, enabling tasks such as sample cleanup and proteolytic reactions to be seamlessly combined with target preparation. Selected developments in each of these areas are described in the following sections. Microfabricated MALDI Targets MALDI target preparation is typically carried out by one of the several proven liquid phase deposition methods, such as the dried-droplet [190], fast solvent evaporation [191], sandwich [192] or two-layer [193] methods. These approaches involve pipetting or spotting sample and matrix solutions on to the MALDI target, then airdrying the deposited spot. As a result of the liquid phase deposition, lateral spreading of the sample over the target can lower the local analyte concentration, thereby reducing MALDI-MS sensitivity [191,194]. Final sample spots typically measure from 5 to 15 mm2, with only a small fraction of this area irradiated during MALDI-MS analysis [195]. For an approximate laser spot diameter of 200 mm, less than 1% of the deposited sample is represented in each measured spectrum. In addition to affecting the sensitivity of MALDI-MS analysis, variations in mass resolution, intensity and selectivity can occur because of surface wetting, preventing meaningful quantitative MALDI-MS analysis [196–198]. As a result, it is typically necessary to acquire numerous spectra from single laser shots on a given deposited sample in order to search for ‘sweet spots’ where analyte concentration is maximized, thereby constraining overall analysis throughput and requiring laborious manual inspection of each sample for optimal results. Thus, reducing the diameter of deposited sample spots has been recognized as an important goal. Microfabrication technologies have been widely used to produce MALDI target plateswithmodifiedsurfacesthat serve toconstrainthe lateralspreadofanalyte during deposition and matrix cocrystallization. The most common approach is the use of patterned hydrophobic regions surrounding hydrophilic target pads [199–201].

246

Miniaturization of Analytical Systems

Off-chip MALDI Target Preparation One strategy for interfacing microfluidics to MALDI-MS is to deposit the sample from the microfluidic device on to a standard off-chip target plate. While direct mechanical spotting of the sample from a microchannel is feasible, precise control of the spotting volume can be difficult, and in the case of high-resolution separations, the spotting action may produce bulk flow within the microchannel, which can impact separation efficiency [202]. Furthermore, direct spotting is difficult for multiplexed devices which contain arrays of closely-spaced microchannels, since crosstalk between adjacent channels can occur during droplet deposition. As an alternative to direct spotting, microfluidic piezoelectric dispensers fabricated in bulk-etched silicon [203–208] have been used to deliver subnanolitre volumes of sample or matrix solutions to MALDI targets. A number of studies have investigated the use of electrospray deposition of analytes from capillaries on to MALDI targets as an alternative to mechanical deposition. Capillary electrospray has been reported for MALDI target preparation following CE [209,210], RP LC [211] and size-exclusion chromatography [212]. Using polycarbonate microchannels with integrated hydrophobic membrane electrospray tips [213], model peptides and proteins were deposited on to a standard stainless-steel target, then MALDI-MS analysis was undertaken. A scheme of the system is shown in Figure 6.14. By reducing the gap between the electrospray tip and the target surface, sample spots as small as 170 mm were readily produced, with significantly improved spectral reproducibility compared to traditional target preparation methods [214]. A similar chip design containing three parallel electrospray tips spaced only 150 mm apart was used to simultaneously deposit bradykinin, fibrinopeptide and angiotensin from the adjacent channels, with no measurable

Figure 6.14 Off-chip MALDI target preparation design. (Reproduced from [214] with permission of Wiley-VCH)

Detection in Miniaturized Analytical Systems

247

crosstalk observed between deposited spots. A further benefit of this approach was demonstrated by employing the hydrophobic membrane used for the electrospray tip as a real-time proteolytic reactor. By saturating the hydrophobic binding sites on the membrane with trypsin, complete digestion of intact proteins was achieved during the electrospray process, potentially enabling on-chip protein separations without the need for on-target digestion prior to MALDI-MS. Microfluidic/Target Integration In the context of MALDI-MS analysis, an important feature of microfluidic systems is that they generally possess a planar geometry in which microchannels are arranged within one or more planes of a flat substrate. This planar configuration makes it convenient for microfluidic substrates to be used directly as MALDI target plates. In this approach to interfacing microfluidics with MALDI-MS, manipulation of sample and matrix in the liquid phase may be combined with on-chip crystallization, thereby eliminating the need for deposition of the sample on to a separate substrate prior to MALDI-MS analysis. For a microfluidic chip to act as a target substrate, analyte molecules must be accessible by the MALDI laser source for ionization. A simple approach demonstrated by Liu et al. [215] consisted of fabricating open microchannels within a glass substrate such that liquid within the channels was exposed to air throughout the chip. Electrophoresis microchannels, 1.6 cm long and 250 mm deep and wide, were machined into a glass chip using a wafer saw to form a standard cross-injection configuration. Peptide and oligosaccharide samples were premixed with matrix, and a plug of the combined solution was injected and separated by CE in the open channels. Following separation, solvent was allowed to evaporate, leaving behind cocrystallized solid phase sample and matrix in the channels. The MALDI laser could then be rastered along the length of the separation channel to generate spatially resolved mass spectra of the separated analyte. While the preceding examples describe approaches to direct online MALDI-MS analysis from microfluidic substrates, there is growing interest in performing onchip target preparation followed by offline MALDI-MS. Ekstr€om et al. [216] have described a simple but promising microfluidic technology for offline MALDI analysis which combines sample cleanup and enrichment with on-chip sample/ matrix cocrystallization. The system consists of a bulk-etched silicon wafer containing an array of nanovials packed with RP beads pre-incubated with protein digest. After washing the 30–40 nl bead volume, the sample is eluted through a 15 mm hole in the wafer backside using a mixture of acetonitrile and matrix solution. The enriched sample deposits on the backside of the wafer in a 500 mm diameter ring around the exit hole. The resulting peptide/matrix sample target is directly interrogated on the supporting silicon substrate.

248

Miniaturization of Analytical Systems

6.5 Unconventional Detection Methods As stated in the previous sections, there is no question about the predominance of LIF and ED, followed by MS, as detectors coupled to analytical microsystems. However, other routes have also been published in the literature, such as ultraviolet (UV), surface-enhanced Raman spectroscopy (SERS), infrared (IR), resonance magnetic nuclear (RMN), surface Plasmon resonance (SPR) and thermal lens microscopy (TLM) [13]. Here we will briefly outline these alternative detection routes. 6.5.1

Absorbance Detection

Absorbance detection is the most universal detection mechanism for samples in CE and HPLC systems. However, this popularity has not translated to the microchip format as the on-chip measurement of absorbing species has proven to be challenging, mainly because of the short optical path length (channel depth) in miniaturized systems and the difficulties in coupling the light into and out of these channels. Nevertheless, several recent reports have been published detailing attempts to overcome these limitations and showing the feasibility of absorbance detection in microsystems. The first approach to increasing the detection sensitivity involves using the incident light more effectively. Some works [217] have attempted this by using waveguides to manipulate the light path on a silica/PDMS microchip with a liquidcore waveguide and a 5 mm path length. They used Teflon fluoropolymers as the cladding material and tested the effectiveness of their system for measuring crystal violet under flow injection conditions. In turn, other works [218] have fabricated pure-silica planar waveguides with an effective 1.2 mm optical path length using more real-world samples; they electrophoretically separated a mixture of caffeine, paracetamol and ketoprofene with detection at 254 nm. The second approach to waveguiding for defining light paths is to transport light from the source through fibre optics. Some investigators [219] have used a miniature fibre-optic spectrometer, a bundle of seven 400 mm UV-visible fibres and a tungsten/halogen lamp as a source in their absorbance detector setup on glass/PDMS microchip. The system was used to determine calcium ions with arsenazo III as a complexing agent. With the goal of applying this to clinical analysis, they showed its effectiveness for detecting calcium in urine. Others have performed on-chip electrochromatography with fibre-optic absorbance detection [220]. They separated three peptides and detected them on a chip positioned between two fibres. The third approach deals with the integration of miniaturized detectors such as fibre optics directly into microchannels [221,222] (similar approaches to those shown in Figure 6.2(b) for LIF detection). In these works, an optical path length of 10 mm for detecting different concentrations of phosphate in water was proposed, showing good

Detection in Miniaturized Analytical Systems

249

sensitivity and reproducibility. The attractiveness of this approach is that it provides a modular detector that potentially could be integrated with any microanalytical system. In addition, planar hydrodynamic chromatography in microchip format with UV absorption detection was performed for the first time by Blom et al. [223]. They analysed polystyrene latex nanoparticles in a chip fabricated from fused silica. The light was transported through optical fibre and collected at a photo cell that was connected to a standard HPLC detector. In other, more peculiar work, Vegvari and Hjerten [224] coupled a hybrid microdevice with UV detection. They fabricated a chip that contained short silica capillaries as separation channels coupled to a modified commercial UV detector; detection of a variety of analytes, including peptides, proteins, DNA, bacteria and viruses, was shown. Comparisons to conventional CE were made and, in some cases, a similar resolution was achieved with significantly decreased analysis times. On the other hand, several groups have utilized commercial microchip instrumentation equipped with a linear, whole-channel UV absorbance detection system [225,226], testing its applicability in transforming affinity capillary electrochromatography (ACE) techniques to microchip format [227], in the rapid separation of antimicrobial metabolites with high efficiency [228] and in chiral separations of pharmaceuticals by electrokinetic chromatography (EKC) [229,230]. The same system was employed in developing a method to test for nitrogen monoxide metabolites in biological fluids [231], along with a single-step quantitation of DNA [232], a chip gel electrophoresis method for high-sensitivity DNA detection [233] and a monosaccharide derivative separation [234]. The value of absorbance detection as a universal detector is clear; however, while this detection technique is often utilized with the capillary format, it is probably not going to succeed with microchips, at least not to the same extent, since the associated short path lengths make it troublesome to perform sensitive detection. Additional difficulties involved with fabricating and aligning the optical light paths cause considerable amounts of light to be lost. For absorbance detection in the UV range, there is the additional requirement of UV-transparent microdevices (quartz or fused silica), which significantly increases the cost of the devices. 6.5.2

Surface Plasmon Resonance (SPR) Detection

SPR is a physical process that can occur when photons of plane-polarized light strike a reflecting surface under total internal reflection (TIR) conditions. The reflecting surface is typically a thin film of metal, such as gold, which allows the electrical field of the photons to extend about half a wavelength beyond the metal surface. When the energy of the photon electrical field is optimized, it can interact with the free electron constellations in the gold surface. The photons from the incident light are absorbed and converted into the surface plasmons used for

250

Miniaturization of Analytical Systems

Figure 6.15 Principle and micrographs of the fibre-optic SPR microsensor (a) and the magnified apex region (b), which is explained in schematic (c). Experimental setup for the SPR microsensors (d) that monitors changes in the RI of liquids flowing in the microfluidic device. PD, photodetector; PC, personal computer. (Reprinted from [237] with permission from Elsevier)

detection. SPR is a powerful technique for measuring biomolecular interactions in real time and in a label-free environment. When reactants in solution (e.g. highaffinity ligand or antigen) are exposed to a surface where a sensor molecule (e.g. receptor or antibody) is immobilized, binding results in a change in the refractive index (RI) at the sensor surface, which is measured as a change in resonance angle or resonance wavelength, and normally converted to arbitrary units and displayed in a plot termed a sensorgram. Figure 6.15 shows the principle of the SPR approach. Most of the recent work involving miniaturized SPR detection has focused on designing immunosensors in which interaction of species is required on the sensing surface. Typically, this kind of system uses flow of analytes as the sample introduction mechanism. Relevant works dealing with SPR detection have been appearing. Whelan and Zare [235] have used a miniaturized SPR sensor to detect a mixture of three high-RI materials with an untreated gold surface, demonstrating the functionality of the system by selectively analysing electrophoretically delivered IgG with proteins immobilized on the gold surface. They also developed a fabrication technique in PDMS for smaller-volume flow cells for the SPR detector [236]. A positive photoresist was used to fabricate these microdevices, which were utilized in detecting benzyl alcohol in bulk solution and deposition of IgG on the treated sensing surface. An interesting variation on SPR detection was created by Kurihara et al. [237], who fabricated a conical fibre-optic microsensor by coating a gold film on a chemically etched single-mode fibre with a conical tip on the end. They applied their design to real-time monitoring of a four-component mixture flowing in a microfluidic device. Other investigators have studied DNA hybridization on a solid support [238], where SPR detection was compared with two other detection

Detection in Miniaturized Analytical Systems

251

techniques on different detection-specific devices. An SPR-based immunosensor for the detection of the main metabolite of heroin and morphine [239] and another approach using a gold-binding polypeptide in an indirect competitive immunoassay in an SPR-based biosensor for the detection of a dioxin precursor [240] have also been published. Microarray-based immunoassays with SPR imaging have been used for labelfree and real-time analysis with the goal of increasing throughput. Kanda et al. [241] used PDMS microfluidic devices as a surface to pattern the gold and adsorb antigens from the sample, with the potential to produce arrays of 64 or more spots. Xinglong et al. [242] studied rabbit IgG and its antibody by 2D detection in different microarrays. The same group [243] also used imaging interferometry to obtain information on biochemical reactions by extracting the phase change from the SPR interference patterns. Palumbo et al. [244] developed a multichannel SPR imaging system with no moving parts and applied it to monitoring polyelectrolyte thin-film thicknesses. SPR detection is ideal for microchips in many respects. It is universal such that any sample which produces a change in the RI can be detected. It offers very sensitive real-time label-free detection, which requires only small amounts of sample. Also, the surface plasmons can probe the sample solution more efficiently due to the small dimensions of the microchannels. The number of applications using SPR detection is likely to increase in the future, especially for bionsensors involving specific biomolecular interactions. 6.5.3

Thermal Lens Detection

Thermal lens detection takes place when coaxial excitation and probe beams are focused into a liquid sample. The energy of the excitation beam is absorbed by the sample species, which undergo a transmission to excited states. The excitedstate molecules subsequently lose their energy in the form of heat through nonradiative relaxation processes, causing localized temperature increases to occur in the sample. The heating of the sample correlates with a change in the RI, producing an effect which appears equivalent to the formation of a transient optical lens within the medium – this phenomenon is called the ‘thermal lens effect.’ Since there is a linear dependence on the concentration of the absorbing molecules, quantitative analysis can be carried out by measuring changes in the divergence or convergence of the probe beam due to the thermal lensing. Articles have recently been published that collectively provide a comprehensive overview of the technique and instrumentation involved in microchip TLM [245–247]. The principle of TLM is shown in Figure 6.16. In the literature, works have appeared dealing with the fundamental aspects of the coupling of this type of detection. A 3D glass microchannel network for the performance of mixing, reaction, solvent extraction and TLM detection on this

252

Miniaturization of Analytical Systems

Figure 6.16 Principles of a thermal lens microscope. (a) TLE is generated by the excitation beam and detected by the probe beam. Because the TLE is concave, it shifts the focal point of the probe beam; the shift is detected in the probe beam intensity after passing through the pinhole. (b) Illustration of how changing the excitation and probe beams affects the focal point difference, which in turn affects signal detection. Proper adjustment of the focal point difference is required for optimal detection. (c) Light introduction part of a micro-TLM unit. (Reproduced from [247] with permission of ACS, Copyright 2004)

microchip with pressure-driven fluidics has been published [248]. This analysed samples of Fe(II) and Co(II) ions and concluded that the system could be used for analysis of micromolar sample concentrations. Proskurnin et al. [249] compared the possibilities for thermal lens detection in capillaries and microchips based on theoretical calculation and experimental confirmation. Matching the best LOD expected by UV detection in CE, they demonstrated a 108 M LOD for ferroin. Uchiyama et al. [250] optimized an interface chip for the coupling of CE capillaries with TLM detection. They investigated the dependence of a number of theoretical plates on the detector location and the channel dimensions by separating several derivatized amino acids. Theoretical plate numbers around 100 000 were reached for the optimized system. Ghaleb and Georges [251] studied the effects of the excitation beam size on thermal lens spectrometry detection in microfluidic channels. They concluded that small laser beam waists can reduce the edge effects and provide better spatial resolution, allowing flow velocities higher than 1 mm/s to be measured in the channels. Tamaki et al. [252] built a thermal lens spectrometer with a Xe lamp as an excitation source. They compared the spectra from turbid solutions with normal clear solutions and found that light scattering was not a disturbance in their microchip-based system, allowing the analysis of turbid samples such as cell lysis solutions, nanoparticles, colloids and vesicles. Tokeshi et al. [253] developed a miniaturized thermal lens detection system with two different configurations that were based on an optical fibre and SELFOC microlens. Detection limits of 108 and 109 M were determined for two test dyes, namely sunset yellow and NiP, respectively. Mawatari et al. [254] also worked on miniaturizing a thermal lens spectrometer with a precise focusing system. They achieved a detection limit of 30 nM for xylenecyanol solution, with a

Detection in Miniaturized Analytical Systems

253

sensitivity of 20–100 times that obtained by spectrophotometry. In addition, the microchip-TLM approach has been proposed for the bioassay of cultured cells [255]. With cell culture, chemical and enzymatic reactions, and detection integrated into a single glass microdevice, nitricoxide release from peritoneal macrophage cells could be monitored and the total assay time reduced from 24 to 4 hours. Impressively, the LOD was improved two orders of magnitude over conventional batch methods. TLM is one of the most common detectors in microsystems. It is widely applicable and especially suitable for microchannel detection because of the rectangular geometries involved, which are easier to align for detection purposes than round capillaries. Unlike absorbance detection, the sensitivity of TLM detection is usually independent of the optical path length, resulting in considerably higher sensitivity, up to one-molecule detection. The only limitations that could slow TLM detector usage with future microsystems are that careful alignment of the beams is needed for optimal configuration and two lasers are needed for excitation and probing of the sample, contributing to the overall cost. However, the availability of small diode lasers will make the detection system less expensive and more portable. 6.5.4

Other Detection Methods

Although it may have unique capabilities, IR detection is likely to have limited applicability with future microdevices. While it can provide valuable information about reaction rate constants, folding states and secondary structures of proteins in solution, these advantages are offset by the lack of chip materials that are IRtransparent. In addition, radiation sources would need to be improved in terms of intensity and ease of coupling with the microdevice, e.g. through fibre optics. Some works dealing with IR detection route coupled to microfludioc devices can be found [256–258]. Raman detection has few limitations that would constrain its application as a microdevice detection technique in the future. The substrate and its thickness must be carefully selected for the microdevice in order to minimize the background fluorescence signal from the glass. Also, in small-volume microchannels some analytes emit very weak Raman signals requiring high concentrations and/or long acquisition times. However, Raman methods should have a better outlook than IR techniques since water is virtually Raman transparent and adsorption by water molecules poses no interference. In addition, surface or resonance enhancement techniques can be used to improve the sensitivity. Finally, laser light sources over a wide range of wavelengths are available and excitation close to near-IR can be utilized to suppress the fluorescence background signal. Although Raman has more advantages than IR, as stated above, not many works have been published [259–264].

254

Miniaturization of Analytical Systems

NMR is a superior tool for the provision of direct solution-based determination of molecular structure. Unfortunately, at the same time it is one of the least sensitive analytical techniques. Currently, microscale NMR can only be applied to samples that are present in appreciable concentrations. The microcoils have not yet developed to the level where they could be applied to samples with small amounts of analytes. Better spectral resolution could be obtained by improving the magnetic field homogeneity of the microcoils, but not enough research has been conducted yet to determine whether microchip NMR can become a major analysis tool [265–268].

6.6 Outlook Both LIF and ED are the standard detection methods in analytical microsystems. LIF is the most sensitive option; however, ED becomes the more versatile, allowing different creative routes as well as offering an inherent high compatibility with the microfabrication techniques. Despite first impressions, MS is a very suitable detection route because of the high compatibility between the flows used by MS and those employed under microfluidic scale, and because it is a universal detection method. With respect to unconventional routes, for detection methods that have been translated from the capillary to the microchip format, most seem to have retained their effectiveness, with the exception of absorbance detection, which suffers from sensitivity problems. Without a doubt, absorbance detection will continue to be used within the microformat, but it will be limited in its applicability to certain sample concentrations. It can be envisioned that TLM and SPR will become much more common in microfluidic detection. Instrumentation for the utilization of both of these techniques with microdevices has already been developed and miniaturized for portability. Furthermore, they are both universal in their response, requiring no chemical treatment of the analytes to make them detectable. Both methods require lasers as radiation sources but many laboratories already have LIF detectors and the same sources can be used. Some of the other alternative microchip detection techniques are more exotic and require external instrumentation that can be prohibitively expensive for many laboratories. Also, MS techniques are established in chip-based detection to the extent that the same information, if not more, can often be obtained by using MS. Many of these detection techniques have great potential in specific applications, however.

References [1] A. Manz, N. Grabner, H.M. Widmer, Sens. Actuators B Chem., 1 (1990) 244. [2] M.A. Schwarz, P.C. Hauser, Lab Chip, 1 (2001) 1. [3] K. Uchiyama, H. Nakajima, T. Hobo, Anal. Bioanal. Chem., 379 (2004) 375.

Detection in Miniaturized Analytical Systems [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48]

255

M.E. Johnson, J.P. Landers, Electrophoresis, 25 (2004) 3513. C. Yi, Q. Zhang, C.W. Li, J. Yang, J. Zhao, M. Yang, Anal. Bioanal. Chem., 384 (2006) 1259. S. G€otz, U. Karst, Anal. Bioanal. Chem., 387 (2007) 183. W.R. Vandaveer, S.A. Pasas-Farmer, D.J. Fischer, C.N. Frankenfeld, S.M. Lunte, Electrophoresis, 25 (2004) 3528. J. Wang, Electroanalysis, 17 (2005) 1133. M. Pumera, A. Merkoc¸i, S. Alegret, Trends Anal. Chem., 25 (2006) 219. A.J. Blasco, A. Escarpa,In: M.L. Marina, A. Rıos, M. Valcarcel (Eds), Analysis and detection by capillary electrophoresis, Elsevier CAC Series, Amsterdam, 2005, pp. 703–758. D.J. Harrison, R.D. Oleschuk, P. Thibault,In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-ona-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003, pp. 249–270. T. Laurell, G. Marko-Varga, J. Nilsson,In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-aChip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003 pp. 271–284. P.J. Viskari, J.P. Landers, Electrophoresis, 27 (2006) 1797. K. Seiler, D.J. Harrison, A. Manz, Anal. Chem., 65 (1993) 1481. C.S. Effenhauser, A. Manz, H.M. Widmer, Anal. Chem., 65 (1993) 2637. Y. Liu, R.S. Foote, S.C. Jacobson, R.S. Ramsey, J.M. Ramsey, Anal. Chem., 72 (2000) 4608. Q. Lu, J.H. Callahan, G.E. Collins, Chem. Commun., (2000) 1913. N.J. Munro, Z. Huang, D.N. Finegold, J.P. Landers, Anal. Chem., 72 (2000) 2765. M. Arundell, P.D. Whalley, A. Manz, Fresenius’ J. Anal. Chem., 367 (2000) 686. S.R. Wallenborg, C.G. Bailey, Anal. Chem., 72 (2000) 1872. S. Sirichai, A.J. de Mello, Analyst, 125 (2000) 133. S. Sirichai, A.J. de Mello, Electrophoresis, 22 (2001) 348. K.W. Ro, J.H. Hanh, Electrophoresis, 26 (2005) 4767. S.H. Lee, S.I. Cho, C.S. Lee, B.G. Kim, Y.K. Kim, Sens. Actuators B, 110 (2005) 164. P.J. Obeid, T.K. Christopoulos, P.C. Ioannou, Electrophoresis, 25 (2004) 922. N.A. Lacher, N.F. de Rooij, E. Verpoorte, S.M. Lunte, J. Chromatogr. A, 1004 (2003) 225. Z.D. Sandlin, M. Shou, J.G. Shackman, R.T. Kennedy, Anal. Chem., 77 (2005) 7702. J. Qin, N. Ye, L. Yu, D. Liu, Y. Fung, W. Wang, X. Ma, B. Lin, Electrophoresis, 26 (2005) 1155. J. Qin, F.C. Leung, Y. Fung, D. Zhu, B. Lin, Anal. Bioanal. Chem., 381 (2005) 812. B.C. Giordano, L. Jin, A.J. Couch, J.P. Ferrance, J.P. Landers, Anal. Chem., 76 (2004) 4705. A.B. Jemere, R.D. Oleschuk, D.J. Harrison, Electrophoresis, 24 (2003) 3018. E.S. Roddy, J.A. Lapos, A.G. Ewing, J. Chromatogr. A., 1004 (2003) 217. S. Etoh, T. Fujimura, R. Hattori, Y. Kuroki, Jpn. J. Appl. Phys. Pt., 1 (42) (2003) 4093. X. Yang, X. Zhang, A. Li, S. Zhu, Y. Huang, Electrophoresis, 24 (2003) 1451. Y.Y. Ling, X.F. Yin, Z.L. Fang, Electrophoresis, 26 (2005) 4759. N.P. Beard, J.B. Edel, A.J. Demello, Electrophoresis, 25 (2004) 2363. J. Gao, X.F. Yin, Z.L. Fang, Lab Chip, 4 (2004) 47. A.M. Skelley, R.A. Mathies, J. Chromatogr. A, 1021 (2003) 191. K. Hata, Y. Kichise, T. Kaneta, T. Imasaka, Anal. Chem., 75 (2003) 1765. B.F. Liu, H. Hisamoto, S. Terabe, J. Chromatogr. A, 1021 (2003) 201. J. Qin, Y. Fung, D. Zhu, B.C. Lin, J. Chromatogr. A, 1027 (2004) 223. P. Schulze, M. Ludwig, F. Kohler, D. Belder, Anal. Chem., 77 (2005) 1325. H.F. Li, J.M. Lin, R.G. Su, K. Uchiyama, T. Hobo, Electrophoresis, 25 (2004) 1907. H.F. Li, Z.W. Cai, J.M. Lin, Anal. Chim. Acta, 565 (2006) 183. S.K. Hsiung, C.H. Lin, G.B. Lee, Electrophoresis, 5 (2005) 1122. J.L. Fu, Q. Fang, T. Zhang, X.H. Jin, Z.L. Fang, Anal. Chem., 78 (2006) 3827. Z. Shen, X. Liu, Z. Long, D. Liu, N. Ye, J. Qin, Z. Dai, B. Lin, Electrophoresis, 27 (2006) 1084. Y. Mourzina, A. Steffen, D. Kalyagin, R. Carius, A. Offenh€ausser, Electrophoresis, 26 (2005) 1849.

256

Miniaturization of Analytical Systems

[49] S.J.O. Varjo, M. Ludwig, D. Belder, M.L. Riekkola, Electrophoresis, 25 (2004) 1901. [50] D.A. Wicks, P.C.H. Li, Anal. Chim. Acta, 507 (2004) 107. [51] D.E. Starkey, Y. Abdelaziez, C.H. Ahn, J. Tu, L. Anderson, K.R. Wehmeyer, N.J. Izzo, A.N. Carr, K.G. Peters, J.J. Bao, H.B. Halsall, W.R. Heineman, Anal. Biochem., 316 (2003) 181. [52] N. Piehl, M. Ludwig, D. Belder, Electrophoresis, 25 (2004) 3848. [53] M. Kato, Y. Gyoten, K. Sakai-Kato, T. Toyo’oka, J. Chromatogr. A, 1013 (2003) 183. [54] Q. Yan, R.S. Chen, J.K. Cheng, Anal. Chim. Acta, 555 (2006) 246. [55] K. Miyaki, Y. Guo, T. Shimosaka, T. Nakagama, H. Nakajima, K. Uchiyama, Anal. Bioanal. Chem., 382 (2005) 810. [56] Y. Guo, K. Uchiyama, T. Nakagama, T. Shimosaka, T. Hobo, Electrophoresis, 26 (2005) 1843. [57] B.F. Liu, M. Ozaki, Y. Utsumi, T. Hattori, S. Terabe, Anal. Chem., 75 (2003) 36. [58] B.F. Liu, M. Ozaki, H. Hisamoto, Q. Luo, Y. Utsumi, T. Hattori, S. Terabe, Anal. Chem., 77 (2005) 573. [59] H. Qiu, J. Yan, X. Sun, J. Liu, W. Cao, X. Yang, E. Wang, Anal. Chem., 75 (2003) 5435. [60] X. Zhao, T. You, H. Qiu, J. Yan, X. Yang, E. Wang, J. Chromatogr. B, 810 (2004) 137. [61] J.J. Xu, A.J. Wang, H.Y. Chen, Trends Anal. Chem., 26 (2007) 125. [62] W.R. Vandaveer, S.A. Pasas, R.S. Martın, S.M. Lunte, Electrophoresis, 23 (2002) 3667. [63] J. Tanyanyiwa, S. Leuthardt, P.C. Hauser, Electrophoresis, 23 (2002) 3659. [64] J. Wang, Talanta, 56 (2002) 223. [65] S.R. Wallenborg, L. Nyholm, C.E. Lunte, Anal. Chem., 71 (1999) 544. [66] A.T. Woolley, K. Lao, A.N. Glazer, R.A. Mathies, Anal. Chem., 70 (1998) 684. [67] J. Wang, B. Tian, E. Sahlin, Anal. Chem., 71 (1999) 5436. [68] J. Wang, B. Tian, E. Sahlin, Anal. Chem., 71 (1999) 3901. [69] J. Wang, M. Pumera, M.P. Chatrathi, A. Escarpa, R. Konrad, A. Griebel, W. D€ orner, H. L€ owe, Electrophoresis, 23 (2002) 596. [70] J. Wang, M. Pumera, M.P. Chatrathi, A. Rodriguez, S. Spillman, R.S. Martın, S.M. Lunte, Electroanalysis, 14 (2002) 1251. [71] R.S. Martin, A.J. Gawron, S.M. Lunte, C.S. Henry, Anal. Chem., 72 (2000) 3196. [72] Y. Zeng, H. Chen, D.W. Pang, Z.L. Wang, J.K. Cheng, Anal. Chem., 74 (2002) 2441. [73] P. Ertl, C.A. Emrich, P. Singhal, R.A. Mathies, Anal. Chem., 76 (2004) 3749. [74] R.S. Martin, K.L. Ratzlaff, B.H. Huynh, S.M. Lunte, Anal. Chem., 74 (2002) 1136. [75] Y. Liu, J.A. Vickers, C.S. Henry, Anal. Chem., 76 (2004) 1513. [76] J.S. Rossier, R. Ferrigno, H.H. Girault, J. Electroanal. Chem., 492 (2000) 15. [77] D.C. Chen, F.L. Hsu, D.Z. Zhan, C.H. Chen, Anal. Chem., 73 (2001) 758. [78] C.C. Wu, R.G. Wu, J.G. Huang, Y.C. Lin, H.C. Chang, Anal. Chem., 75 (2003) 947. [79] C.C.J. Lai, C.H. Chen, F.H. Ko, J. Chromatogr. A, 1023 (2004) 143. [80] D.M. Osbourn, C.E. Lunte, Anal. Chem., 75 (2003) 2710. [81] N.A. Lacher, S.M. Lunte, R.S. Martin, Anal. Chem., 76 (2004) 2482. [82] R.S. Keynion, T.J. Roussel, Jr., M.M. Crain, D.J. Jackson, D.B. Franco, J.F. Naber, K.M. Walsh, R.P. Baldwin, Anal. Chim. Acta, 507 (2004) 95. [83] M.M. Crain, R.S. Keynton, K.M. Walsh, T.J. Roussel, R.P. Baldwin, J.F. Naber, D.J. Jackson, In: C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Humana Press, Totopwa, New Jersey, 2006 pp. 13–26. [84] A.J. Gawron, R.S. Martin, S.M. Lunte, Electrophoresis, 22 (2001) 242. [85] R.S. Martin, A.J. Gawron, B.A. Fogarty, F.B. Regan, E. Dempsey, S.M. Lunte, Analyst, 126 (2001) 277. [86] M.L. Kovarik, N.J. Torrence, D.M. Spence, R.S. Martin, Analyst, 129 (2004) 400. [87] P. Gavin, A.G. Ewing, J. Am. Chem. Soc., 118 (1996) 8932. [88] J. Wang, A. Escarpa, M. Pumera, J. Feldman, J. Chromatogr. A, 952 (2002) 249. [89] A. Hilmi, J.H.T. Luong, Anal. Chem., 72 (2000) 4677. [90] J. Wang, M.P. Chatrathi, B. Tian, R. Polsky, Electroanalysis, 12 (2000) 691. [91] J. Wang, G. Chen, M.P. Chatrathi, A. Fujishima, D.A. Tryk, D. Shin, Anal. Chem., 75 (2003) 935.

Detection in Miniaturized Analytical Systems

257

[92] N.E. Hebert, B. Snyder, R.L. McCreery, W.G. Kuhr, S.A. Brazill, Anal. Chem., 75 (2003) 4265. [93] S. Iijima, Nature, 354 (1991) 56. [94] J. Wang, Analyst, 130 (2005) 421. [95] A. Merkoc¸i, M. Pumera, X. Llopis, B. Perez, M. del Valle, S. Alegret, Trends Anal. Chem., 24 (2005) 826. [96] C.E. Banks, R.G. Compton, Analyst, 131 (2006) 15. [97] J. Wang, G. Chen, M.P. Chatrathi, M. Musameh, Anal. Chem., 76 (2004) 298. [98] A.G. Crevillen, M. Avila, M. Pumera, M.C. Gonzalez, A. Escarpa, Anal. Chem., 79 (2007) 7408. [99] A.G. Crevillen, M. Pumera, M.C. Gonzalez, A. Escarpa, Lab Chip, 9 (2009) 346. [100] N.E. Hebert, B. Snyder, R.L. McCreery, W.G. Kuhr, S.A. Brazill, Anal. Chem., 75 (2003) 4265. [101] N.E. Hebert, W.G. Kuhr, S.A. Brazill, Anal. Chem., 75 (2003) 3301. [102] J.C. Fanguy, C.S. Henry, Analyst, 127 (2002) 1021. [103] C.D. Garcıa, C.S. Henry, Anal. Chem., 75 (2003) 4778. [104] J. Wang, R. Polsky, B. Tian, M.P. Chatrathi, Anal. Chem., 72 (2000) 5285. [105] R.M. Guijt, E. Baltussen, G. van der Steen, R.B.M. Schasfoort, S. Schlautmann, H.A.H. Billiet, J. Frank, G.W.K. van Dedem, A. van den Berg, Electrophoresis, 22 (2001) 235. [106] M. Galloway, W. Stryjewski, A. Henry, S.M. Ford, S. Llopis, R.L. McCarley, S.A. Soper, Anal. Chem., 74 (2002) 2407. [107] R.M. Guijt, E. Baltussen, G. van der Steen, H. Frank, H. Billiet, T. Schalkhammer, F. Laugere, M. Vellekoop, A. Berthold, L. Sarro, G.W.K. van Dedem, Electrophoresis, 22 (2001) 2537. [108] A. Berthold, F. Laugere, H. Schellevis, C.R. de Boer, M. Laros, R.M. Guijt, P.M. Sarro, M.J. Vellekoop, Electrophoresis, 23 (2002) 3511. [109] J. Lichtenberg, N.F. de Rooij, E. Verpoorte, Electrophoresis, 23 (2002) 3769. [110] M. Pumera, J. Wang, F. Opekar, I. Jelınek, J. Feldman, H. L€ owe, S. Hardt, Anal. Chem., 74 (2002) 1968. [111] J. Wang, M. Pumera, Anal. Chem., 74 (2002) 5919. [112] J. Wang, M. Pumera, Anal. Chem., 75 (2003) 341. [113] J. Wang, G. Chen, A. Muck, Jr., Anal. Chem., 75 (2003) 4475. [114] J. Tanyanyiwa, P.C. Hauser, Anal. Chem., 74 (2002) 6378. [115] F. Laugere, R.M. Guijt, J. Bastemeijer, G. van der Steen, A. Berthold, E. Baltussen, P. Sarro, G.W.K. van Dedem, M. Vellekoop, A. Bossche, Anal. Chem., 75 (2003) 306. [116] X. Bai, Z. Wu, J. Josserand, H. Jensen, H. Schafer, H.H. Girault, Anal. Chem, 76 (2004) 3126. [117] R.D. Oleschuk, D.J. Harrison, Trends Anal. Chem., 19 (2000) 379. [118] A.J. de Mello, Lab Chip, 1 (2001) 7N. [119] P.A. Limbach, Z. Meng, Analyst, 127 (2002) 693. [120] D. Figeys, D. Pinto, Electrophoresis, 22 (2001) 208. [121] N. Lion, T.C. Rohner, L. Dayon, I.L. Arnaud, E. Damoc, N. Youhnovski, Z.Y. Wu, C. Roussel, J. Josserand, H. Jensen, J.S. Rossier, M. Przybylski, H.H. Girault, Electrophoresis, 24 (2003) 3533. [122] S.L.S. Freire, A.R. Wheeler, Lab Chip, 6 (2006) 1415. [123] A.D. Zamfir, L. Bindila, N. Lion, M. Allen, H.H. Girault, J. Peter-Katalinic, Electrophoresis, 26 (2005) 3650. [124] F. Foret, P. Kusy, Electrophoresis, 27 (2006) 4877. [125] I.M. Lazar, J. Grym, F. Foret, Mass Spectrom. Rev., 25 (2006) 573. [126] W.C. Sung, H. Makamba, S.H. Chen, Electrophoresis, 26 (2005) 1783. [127] S. Koster, E. Verpoorte, Lab Chip, 7 (2007) 1394. [128] M. Brivio, R.H. Fokkens, W. Verboom, D.N. Reinhoudt, N.R. Tas, M. Goedbloed, A. van den Berg, Anal. Chem., 74 (2002) 3972. [129] M. Brivio, N.R. Tas, M. Goedbloed, H. Gardeniers, W. Verboom, A. van den Berg, D.N. Reinhoudt, Lab Chip, 5 (2005) 378.

258

Miniaturization of Analytical Systems

[130] J. Liu, K. Tseng, B. Garcia, C.B. Lebrilla, E. Mukerjee, S. Collins, R. Smith, Anal. Chem., 73 (2001) 2147. [131] D.L. DeVoe, C.S. Lee, Electrophoresis, 27 (2006) 3559. [132] M. Yamashita, J.B. Fenn, J. Phys. Chem., 88 (1984) 4671. [133] M. Yamashita, J.B. Fenn, J. Phys. Chem., 88 (1984) 4451. [134] P.E. Andren, M.R. Emmett, R.M. Caprioli, J. Am. Soc. Mass Spectrom., 5 (1994) 867. [135] A.P. Bruins, T.R. Covey, J.D. Henion, Anal. Chem., 59 (1987) 2642. [136] J. Abian, A.J. Oosterkamp, E. Gelpi, J. Mass Spectrom., 34 (1999) 244. [137] M.S. Wilm, M. Mann, Int. J. Mass Spectrom. Ion Process., 136 (1994) 167. [138] M. Wilm, M. Mann, Anal. Chem., 68 (1996) 1. [139] A.P. Bruins, J. Chromatogr. A, 794 (1998) 345. [140] Q. Xue, F. Foret, Y.M. Dunayevskiy, P.M. Zavracky, N.E. McGruer, B.L. Karger, Anal. Chem., 69 (1997) 426. [141] R.S. Ramsey, J.M. Ramsey, Anal. Chem., 69 (1997) 1174. [142] T.C. Rohner, J.S. Rossier, H.H. Girault, Anal. Chem., 73 (2001) 5353. [143] Q. Xue, Y.M. Dunayevskiy, F. Foret, B.L. Karger, Rapid Commun. Mass Spectrom., 11 (1997) 1253. [144] Y.X. Wang, J.W. Cooper, C.S. Lee, D.L. DeVoe, Lab Chip, 4 (2004) 363. [145] T. Koerner, R.D. Oleschuk, Rapid Commun. Mass Spectrom., 19 (2005) 3279. [146] T. Koerner, K. Turck, L. Brown, R.D. Oleschuk, Anal. Chem., 76 (2004) 6456. [147] N.H. Bings, C. Wang, C.D. Skinner, C.L. Colyer, P. Thibault, D.J. Harrison, Anal. Chem., 71 (1999) 3292. [148] J. Li, C. Wang, J.F. Kelly, D.J. Harrison, P. Thibault, Electrophoresis, 21 (2000) 198. [149] J. Li, J.F. Kelly, I. Chernushevich, D.J. Harrison, P. Thibault, Anal. Chem., 72 (2000) 599. [150] G.A. Valaskovic, N.L. Kelleher, D.P. Little, D.J. Zeserud, F.W. McLafferty, Anal. Chem., 67 (1995) 3802. [151] I.M. Lazar, R.S. Ramsey, J.M. Ramsey, Anal. Chem., 73 (2001) 1733. [152] E.X. Vrouwe, J. Gysler, U.R. Tjaden, J. van der Greef, Rapid Commun. Mass Spectrom., 14 (2000) 1682. [153] D. Figeys, S.P. Gygi, G. McKinnon, R. Aebersold, Anal. Chem., 70 (1998) 3728. [154] D. Figeys, Y. Ning, R. Aebersold, Anal. Chem., 69 (1997) 3153. [155] W. Shui, Y. Yu, X. Xu, Z. Huang, G. Xu, P. Yang, Rapid Commun. Mass Spectrom., 17 (2003) 1541. [156] N. Lion, J.O. Gellon, H. Jensen, H.H. Girault, J. Chromatogr. A, 1003 (2003) 11. [157] N. Lion, J.O. Gellon, H.H. Girault, Rapid Commun. Mass Spectrom., 18 (2004) 1614. [158] J.S. Rossier, N. Youhnovski, N. Lion, E. Damoc, S. Becker, F. Reymond, H.H. Girault, M. Przybylski, Angew. Chem., Int. Ed., 42 (2003) 53. [159] L. Bindila, M. Froesch, N. Lion, Z. Vukelic, J.S. Rossier, H.H. Girault, J. Peter-Katalinic, A.D. Zamfir, Rapid Commun. Mass Spectrom., 18 (2004) 2913. [160] H. Liu, C. Felten, Q. Xue, B. Zhang, P. Jedrezejewski, B.L. Karger, F. Foret, Anal. Chem., 72 (2000) 3303. [161] E. Gelpi, J. Mass Spectrom., 37 (2002) 241. [162] B. Zhang, F. Foret, B.L. Karger, Anal. Chem., 72 (2000) 1015. [163] B. Zhang, F. Foret, B.L. Karger, Anal. Chem., 73 (2001) 2675. [164] Y. Tachibana, K. Otsuka, S. Terabe, A. Arai, K. Suzuki, S. Nakamura, J. Chromatogr. A, 1011 (2003) 181. [165] J. Li, P. Thibault, N.H. Bings, C.D. Skinner, C. Wang, C. Colyer, D.J. Harisson, Anal. Chem., 71 (1999) 3036. [166] Y. Deng, J. Henion, J. Li, P. Thibault, C. Wang, D.J. Harrison, Anal. Chem., 73 (2001) 639. [167] Y. Deng, H. Zhang, J. Henion, Anal. Chem., 73 (2001) 1432. [168] J.H. Chan, A.T. Timperman, D. Qin, R. Aebersold, Anal. Chem., 71 (1999) 4437. [169] J. Kameoka, H.G. Craighead, H. Zhang, J. Henion, Anal. Chem., 73 (2001) 1935. [170] Y. Tachibana, K. Otsuka, S. Terabe, A. Arai, K. Suzuki, S. Nakamura, J. Chromatogr. A, 1025 (2004) 287.

Detection in Miniaturized Analytical Systems

259

[171] F. Foret, H. Zhou, E. Gangl, B.L. Karger, Electrophoresis, 21 (2000) 1363. [172] I.M. Lazar, L. Li, Y. Yang, B.L. Karger, Electrophoresis, 24 (2003) 3655. [173] I.M. Lazar, R.S. Ramsey, S.C. Jacobson, R.S. Foote, J.M. Ramsey, J. Chromatogr. A, 892 (2000) 195. [174] G.E. Yue, M.G. Roper, E.D. Jeffery, C.J. Easley, C. Balchunas, J.P. Landers, J.P. Ferrance, Lab Chip, 5 (2005) 619. [175] J. Wen, Y. Lin, F. Xiang, D.W. Matson, H.R. Udseth, R.D. Smith, Electrophoresis, 21 (2000) 191. [176] J. Gao, J. Xu, L.E. Locascio, C.S. Lee, Anal. Chem., 73 (2001) 2648. [177] Y. Jiang, P.C. Wang, L.E. Locascio, C.S. Lee, Anal. Chem., 73 (2001) 2048. [178] N.H. Bings, C. Wang, C.D. Skinner, C.L. Colyer, P. Thibault, D.J. Harrison, Anal. Chem., 71 (1999) 3292. [179] M. Schilling, W. Nigge, A. Rudzinski, A. Nezer, R. Hergenr€ oder, Lab Chip, 4 (2004) 220. [180] M. Svedberg, A. Pettersson, S. Nilsson, J. Bergquist, L. Nyholm, F. Nikolajeff, K. Markides, Anal. Chem., 75 (2003) 3934. [181] L. Licklider, X.Q. Wang, A. Desai, Y.C. Tai, T.D. Lee, Anal. Chem., 72 (2000) 367. [182] M. Svedberg, M. Veszelei, J. Axelsson, M. Vangbo, F. Nikolajeff, Lab Chip, 4 (2004) 322. [183] J. Kameoka, R. Orth, B. Ilic, D. Czaplewski, T. Wachs, H.G. Craighead, Anal. Chem., 74 (2002) 5897. [184] S. Arscott, S. Le Gac, C. Rolando, Sens. Actuators B, 106 (2005) 741. [185] G.A. Schultz, T.N. Corso, S.J. Prosser, S. Zhang, Anal. Chem., 72 (2000) 4058. [186] K. Tang, Y. Lin, D.W. Matson, T. Kim, R.D. Smith, Anal. Chem., 73 (2001) 1658. [187] J. Sj€odahl, J. Melin, P. Griss, A. Emmer, G. Stemme, J. Roeraade, Rapid Commun. Mass Spectrom., 17 (2003) 337. [188] V. de Biasi, N. Haskins, A. Organ, R. Bateman, K. Giles, S. Jarvis, Rapid Commun. Mass Spectrom., 13 (1999) 1165. [189] M. Karas, U. Bahr, Trends Anal. Chem., 9 (1990) 321. [190] M. Karas, F. Hillenkamp, Anal. Chem., 60 (1988) 2299. [191] O. Vorm, P. Roepstorff, M. Mann, Anal. Chem., 66 (1994) 3281. [192] L. Li, R.E. Golding, R.M. Whittal, J. Am. Chem. Soc., 11 (1996) 662. [193] Y.Q. Dai, R.M. Whittal, L. Li, Anal. Chem., 71 (1999) 1087. [194] R.R. Hensel, R.C. King, K.G. Owens, Rapid Commun. Mass Spectrom., 11 (1997) 1785. [195] M. Schuerenberg, C. Luebbert, H. Eickhoff, M. Kalkum, H. Lehrach, E. Nordhoff, Anal. Chem., 72 (2000) 3436. [196] K. Strupat, M. Karas, F. Hillenkamp, Int. J. Mass Spectrom. Ion. Proc., 111 (1991) 89. [197] M. Karas, F. Hillenkamp, Anal. Chem., (1998) 2299. [198] F.M.L. Amado, P. Domingues, M.G. Santana-Marques, A.J. Ferrer-Correia, K.B. Tomer, Rapid Commun. Mass Spectrom., 11 (1997) 1347.
[199] T. Redeby, J. Roeraade, A. Emmer, Rapid Commun. Mass Spectrom., 18 (2004) 1161. [200] S. Xiong, Q. Ding, Z. Zhao, W. Chen, G. Wang, S. Liu, Proteomics, 3 (2003) 265. [201] R.L. Gundry, R. Edward, T.P. Kole, C. Sutton, R.J. Cotter, Anal. Chem., 77 (2005) 6609. [202] E. N€agele, M. Vollmer, Rapid Commun. Mass Spectrom., 18 (2004) 3008. [203] S. Ekstr€om, D. Ericsson, P. Onnerfjord, M. Bengtsson, J. Nilsson, G. Marko-Varga, T. Laurell, Anal. Chem., 73 (2001) 214. [204] S. Ekstrom, J. Nilsson, G. Helldin, T. Laurell, G. Marko-Varga, Electrophoresis, 22 (2001) 3984. [205] H.R. Aerni, D.S. Cornett, R.M. Caprioli, Anal. Chem., 78 (2006) 827. [206] S. Ekstr€om, P. Onnerfjord, J. Nilsson, M. Bengtsson, T. Laurell, G. Marko-Varga, Anal. Chem., 72 (2000) 286. [207] T. Miliotis, G. Marko-Varga, J. Nilsson, T. Laurell, J. Neurosci. Methods, 109 (2001) 41. [208] P. Onnerfjord, J. Nilsson, L. Wallman, T. Laurell, G. Marko-Varga, Anal. Chem., 70 (1998) 4755. [209] G.S. McLeod, J. Axelsson, R. Self, P.J. Derrick, Rapid Commun. Mass Spectrom., 11 (1997) 214.

260

Miniaturization of Analytical Systems

[210] G.L. DeVault, M.J. Sepaniak, Electrophoresis, 21 (2000) 1320. [211] D.B. Wall, S.J. Berger, J.W. Finch, S.A. Cohen, K. Richardson, R. Chapman, D. Drabble, J. Brown, D. Gostick, Electrophoresis, 23 (2002) 3193. [212] X. Lou, J.L.J. van Dongen, J. Mass Spectrom., 35 (2000) 1308. [213] Y. Wang, J. Cooper, C.S. Lee, D.L. DeVoe, Lab Chip, 4 (2004) 263. [214] Y.X. Wang, Y. Zhou, B.M. Balgley, J.W. Cooper, C.S. Lee, D.L. DeVoe, Electrophoresis, 26 (2005) 3631. [215] J. Liu, K. Tseng, B. Garcia, C.B. Lebrilla, E. Mukerjee, S. Collins, R. Smith, Anal. Chem., 73 (2001) 2147. [216] S. Ekstr€om, L. Wallman, J. Malm, C. Becker, H. Lilja, T. Laurell, G. Marko-Varga, Electrophoresis, 25 (2004) 3769. [217] M.P. Duggan, T. McCreedy, J.W. Aylott, Analyst, 128 (2003) 1336. [218] K.B. Mogensen, F. Eriksson, O. Gustafsson, R.P.H. Nikolajsen, J.P. Kutter, Electrophoresis, 25 (2004) 3788. [219] N. Malcik, J.P. Ferrance, J.P. Landers, P. Caglar, Sens. Actuators B, 107 (2005) 24. [220] R. Jindal, S.M. Cramer, J. Chromatogr. A, 1044 (2004) 277. [221] S. Bargiel, A. Gorecka-Drzazga, J.A. Dziuban, P. Prokaryn, M. Chudy, A. Dybko, Z. Brzozka, Sens. Actuators A, 115 (2004) 245. [222] Z. Liang, N. Chiem, G. Ocvirk, T. Tang, K. Fluri, D.J. Harrison, Anal. Chem., 68 (1996) 1040. [223] M.T. Blom, E. Chmela, R.E. Oosterbroek, R. Tijssen, A. van Den Berg, Anal. Chem., 75 (2003) 6761. [224] A. Vegvari, S. Hjerten, Electrophoresis, 24 (2003) 3815. [225] H. Nakanishi, T. Nishimoto, M. Kanai, T. Saitoh, R. Nakamura, T. Yoshida, S. Shoji, Sens. Actuators A, 83 (2000) 136. [226] H. Nakanishi, T. Nishimoto, A. Arai, H. Abe, M. Kanai, Y. Fujiyama, T. Yoshida, Electrophoresis, 22 (2001) 230. [227] A.R. Stettler, M.A. Schwarz, J. Chromatogr. A, 1063 (2005) 217. [228] E. Guihen, J.D. Glennon, J. Chromatogr. A, 1071 (2005) 223. [229] F. Kitagawa, S. Aizawa, K. Otsuka, Anal. Sci., 21 (2005) 61. [230] M. Ludwig, F. Kohler, D. Belder, Electrophoresis, 24 (2003) 3233. [231] T. Miyado, Y. Tanaka, H. Nagai, S. Takeda, K. Saito, K. Fukushi, Y. Yoshida, S. Wakida, E. Niki, J. Chromatogr. A, 1051 (2004) 185. [232] F. Xu, M. Jabasini, B. Zhu, L. Ying, X. Cui, A. Arai, Y. Baba, J. Chromatogr. A, 1051 (2004) 147. [233] Z. Xu, T. Nishine, A. Arai, T. Hirokawa, Electrophoresis, 25 (2004) 3875. [234] S. Suzuki, Y. Ishida, A. Arai, H. Nakanishi, S. Honda, Electrophoresis, 24 (2003) 3828. [235] R.J. Whelan, R.N. Zare, Anal. Chem., 75 (2003) 1542. [236] A.R. Wheeler, S. Chah, R.J. Whelan, R.N. Zare, Sens. Actuators B, 98 (2004) 208. [237] K. Kurihara, H. Ohkawa, Y. Iwasaki, O. Niwa, T. Tobita, K. Suzuki, Anal. Chim. Acta, 523 (2004) 165. [238] T. Livache, E. Maillart, N. Lassalle, P. Mailley, B. Corso, P. Guedon, A. Roget, Y. Levy, J. Pharm. Biomed. Anal., 32 (2003) 687. [239] P.P. Dillon, A.J. Killard, S.J. Daly, P. Leonard, R. O’Kennedy, J. Immunol. Methods, 296 (2005) 77. [240] N. Soh, T. Tokuda, T. Watanabe, K. Mishima, T. Imato, T. Masadome, Y. Asano, S. Okutani, O. Niwa, S. Brown, Talanta, 60 (2003) 733. [241] V. Kanda, J.K. Kariuki, J.D. Harrison, M.T. McDermott, Anal. Chem., 76 (2004) 7257. [242] Y. Xinglong, W. Dongsheng, W. Dingxin, O.J. Hua, Y. Zibo, D. Yonggui, L. Wei, Z. Xinsheng, Sens. Actuators B, 91 (2003) 133. [243] Y. Xinglong, W. Dingxin, W. Xing, D. Xiang, L. Wei, Z. Xinsheng, Sens. Actuators B, 108 (2005) 765. [244] M. Palumbo, C. Pearson, J. Nagel, M.C. Petty, Sens. Actuators B, 90 (2003) 264. [245] M.J. Navas, A.M. Jimenez, Crit. Rev. Anal. Chem., 33 (2003) 77.

Detection in Miniaturized Analytical Systems

261

[246] M. Tokeshi, Y. Kikutani, A. Hibara, K. Sato, H. Hisamoto, T. Kitamori, Electrophoresis, 24 (2003) 3583. [247] T. Kitamori, M. Tokeshi, A. Hibara, K. Sato, Anal. Chem., 76 (2004) 52A. [248] Y. Kikutani, H. Hisamoto, M. Tokeshi, T. Kitamori, Lab Chip, 4 (2004) 328. [249] M.A. Proskurnin, A.V. Pirogov, M.N. Slyadnev, A.G. Borzenko, Y.A. Zolotov, J. Anal. Chem., 59 (2004) 828. [250] K. Uchiyama, M. Tokeshi, Y. Kikutani, A. Hattori, T. Kitamori, Anal. Sci., 21 (2005) 49. [251] K.A. Ghaleb, J. Georges, Appl. Spectrosc., 58 (2004) 1116. [252] E. Tamaki, A. Hibara, M. Tokeshi, T. Kitamori, Lab Chip, 5 (2005) 129. [253] M. Tokeshi, J. Yamaguchi, A. Hattori, T. Kitamori, Anal. Chem., 77 (2005) 626. [254] K. Mawatari, Y. Naganuma, K. Shimoide, Anal. Chem., 77 (2005) 687. [255] M. Goto, K. Sato, A. Murakami, M. Tokeshi, T. Kitamori, Anal. Chem., 77 (2005) 2125. [256] S. Kulka, N. Kaun, J.R. Baena, J. Frank, P. Svasek, D. Moss, M.J. Vellekoop, B. Lendl, Anal. Bioanal. Chem., 378 (2004) 1735. [257] T. Pan, R.T. Kelly, M.C. Asplund, A.T. Woolley, J. Chromatogr. A, 1027 (2004) 231. [258] T.M. Floyd, M.A. Schmidt, K.F. Jensen, Ind. Eng. Chem. Res., 44 (2005) 2351. [259] R.M. Connatser, R.A. Riddle, M.J. Sepaniak, J. Sep. Sci., 27 (2004) 1545. [260] F.T. Docherty, P.B. Monaghan, R. Keir, D. Graham, W.E. Smith, J.M. Cooper, Chem. Commun., (2004) 118. [261] T. Park, S. Lee, G.H. Seong, J. Choo, E.K. Lee, Y.S. Kim, W.H. Ji, S.Y. Hwang, D.-G. Gweon, S. Lee, Lab Chip, 5 (2005) 437. [262] S.A. Leung, R.F. Winkle, R.C.R. Wootton, A.J. de Mello, Analyst, 130 (2005) 46. [263] P.D.I. Fletcher, S.J. Haswell, X. Zhang, Electrophoresis, 24 (2003) 3239. [264] K. Ramser, J. Enger, M. Goksor, D. Hanstorp, K. Logg, M. K€all, Lab Chip, 5 (2005) 431. [265] C. Massin, F. Vincent, A. Homsy, K. Ehrmann, G. Boero, P.-A. Besse, A. Daridon, E. Verpoorte, N.F. de Rooij, R.S. Popovic, J. Magn. Reson., 164 (2003) 242. [266] B. Sorli, J.F. Chateaux, M. Pitaval, H. Chahboune, B. Favre, A. Briguet, P. Morin, Meas. Sci. Technol., 15 (2004) 877. [267] J.H. Walton, J.S. de Ropp, M.V. Shutov, A.G. Goloshevsky, M.J. McCarthy, R.L. Smith, S.D. Collins, Anal. Chem., 75 (2003) 5030. [268] H. Wensink, F. Benito-Lopez, D.C. Hermes, W. Verboom, H.J.G.E. Gardeniers, D.N. Reinhoudt, A. van den Berg, Lab Chip, 5 (2005) 280.

7 Miniaturization of the Entire Analytical Process I: Micro(nano)sensors 7.1

Introduction

This chapter provides an overview of the principles and developments behind micro (nano)sensors. It will focus on the evolution of sensors through nanotechnology, in agreement with the IUPAC definition, ‘a chemical sensor is a device that transforms chemical information into an analytically useful signal’. Normally the origin of the chemical information is a chemical reaction which involves the analyte. In some cases, the chemical information is provided by a physical property of the system. Traditionally, a chemical sensor is characterized as presenting two basic functional units: the receptor and the transducer. The receptor is the part of the sensor in which selective recognition takes place and the transducer is the part that converts the chemical information into a measurable signal, which is processed and recorded. According to the recognition process, three types of sensor can be distinguished: (i) the so-called physical sensors, in which the response produces a change in a physical property such as mass, conductivity, pressure, temperature, etc.; (ii) chemical sensors; and (iii) biochemical sensors. Ideally, a sensor must provide good selectivity and sensitivity. In addition, it is recommended that it presents a high stability and robustness. A short response time is also an important characteristic when dealing with real samples. Finally, it must be integrated in a self-contained device. This aspect is important because it allows the simplification of the analytical process. Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet Ó 2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

264

Miniaturization of Analytical Systems

In recent years, big changes have occurred in the development of sensors, thanks to the incorporation of nanoparticles. New phenomena are being observed and exploited in research and applications incorporating nanoparticles into sensors. Some of these new phenomena are itemized below: .

.

.

.

. .

Atomic diffusion through nanoparticle interfaces becomes an efficient mechanism of transport of matter at relatively low temperatures compared to conventional matter. This effect can be used to enhance the sensitivity of a gas sensor. If the nanoparticle is crystalline and its size is smaller than that of the electron mean free path, the electronic conductivity is found to decrease because of grain boundary scattering. Phonon spectra are modified by the effects of surface. This is referred to as the ‘phonon confinement effects’. Size-induced control of luminescence relaxation and/or the optical properties takes place. Such effects are interesting in optoelectronic device applications. Surface effects in magnetic materials lead to more efficient magnetic sensors. Tribiological properties are changed tremendously. These changes facilitate reduced friction and wear in microelectromechanical systems (MEMS)-application microsystems.

7.2

Evolution of Sensors with Nanotechnology

Nowadays a wide range of sensors with different transduction mechanisms exist. There are surface acoustic wave sensors, infrared (IR) gas detectors, potentiometric and amperometric sensors, paramagnetic cells, flame ionization detectors, etc. Notwithstanding this large list, sensors have experienced an exponential evolution with the use of nanocomponents. An example is the development of a gas sensor through the use of metallic nanoparticles. The sensing mechanism is based on a change in the potential barrier of the metal oxide when it is exposed to a gas atmosphere [1]. A change in the resistance is measured using a pair of electrodes placed under the metal oxide layer. The SnO2 nanoparticles of the sensors were catalytically modified by different Pd loadings, which act as active filters (Figure 7.1). In this way, the sensor response was tuned to present a higher sensitivity to CO than to O2. Nanoparticles increased the sensitivity but also the selectivity of the sensor response compared to commercial sensors. This is an example of how nanoparticles can improve the recognition element of a sensor. Nanoparticles can also have a positive influence on the transducer element of the sensor. An example is carbon nanotubes (CNTs) in electrochemical sensors. As shown in Figure 7.2, within the electronic bulk capacitance there is an asymmetric capacitor established between the carbon nanoparticle and the solution. For electrochemical nanosensors, single-walled carbon nanotubes (SWCNTs) are

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

265

Figure 7.1 Miniaturized gas sensor modified with nanoparticles. In addition to enhancing sensitivity, the SnO2 nanoparticles allow CO and O2 to be distinguished. (Reprinted from [1] with permission from Elsevier)

Figure 7.2 Scheme of the transduction mechanism of CNTs in solid contact ion-selective electrodes. (Extracted from [2] with permission of ACS, Copyright 2009)

266

Miniaturization of Analytical Systems

especially suitable because all the atoms are located at the surface and small variations in the local chemical environment can be detected. In addition, CNTs are very hydrophobic, therefore no water layer is formed at the interface between the polymeric membrane and the network of CNTs and they do not involve redox reactions. These properties make the instrumental response of the SWCNT electrodes (the sensor) very stable and insensitive to electroactive reagents. This is a clear example of how nanoparticles – CNTs – can be used to improve a transducer.

7.3 Micro(nano)sensors Chemical sensors can be broadly classified according to the recognition element and the transduction technique. In this section, micro(nano)sensors will be classified according to transducer technique. 7.3.1

Optical Sensors

Basically, optical sensors are sensors based on the interaction of light with matter. An optical sensor consists of a sensitive element placed in contact with an optical transmitter and coupled to a light source and a photodetector that converts the light into an electrical signal. In recent years, nanoparticles have been widely used to develop optical sensors because of their optical properties. Quantum dots (QDs) are among the most used nanoparticles in optical sensing [3]. QDs are nanostructured materials, also known as zero-dimensional materials, semiconductor nanocrystals or nanocrystallites. These colloidal nanocrystalline semiconductors, comprising elements from the periodic groups II–VI, III–V or IV–VI, are roughly spherical, with sizes typically in the range 1–12 nm diameter. At such reduced sizes (close to or smaller than the dimensions of the exciton Bohr radius within the corresponding bulk material), these nanoparticles behave differently to bulk solids due to quantum-confinement effects [4,5]. Quantum confinements are responsible for the remarkable attractive optoelectronic properties exhibited by QDs, including their high-emission quantum yields, size-tuneable emission profiles and narrow spectral bands [5,6]. Moreover, their strong sizedependent properties result in a tunability emission. The size-dependent emission is probably the most striking and most studied optical property of QDs. As the emission properties of semiconductor nanocrystals depend strongly upon the energy and the density of the electron states, they can be altered by engineering the size and shape of these tiny structures (Figure 7.3). Figure 7.4 shows the fluorescence spectra of CdSe QDs with different nanoparticle diameters. As can be seen, differently-sized CdSe nanocrystals can be tuned in the 500–700 nm range. As the luminescence of QDs is very sensitive to their surface states [7], it is reasonable

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

267

Conduction band

Energy

P

e–

Conduction band Bandgap

SP

3

σ

S

Trap

Energy

σ*

E ≥ Eg

Luminescence

Valence band

Trap

h+

Valence band

Atomic orbitals Hybrid orbitals Molecular orbitals Density of states

Figure 7.3 Atomic orbitals versus molecular orbitals of semiconductor nanoparticles, and an energy-level diagram describing the promotion of one electron

Fluorescence λ=551nm

λ=590 nm

λ=647 nm

hν(λ=400 nm)

∼3 nm

∼4 nm

∼7 nm

Normalized fluorescence intensity

1 0.8 0.6 0.4 0.2

0 500

550

600 Wavelength, nm

650

700

Figure 7.4 Size-tuneable fluorescence spectra of CdSe QDs. The diameter sizes of the nanoparticles are shown over the fluorescence spectrum. (Extracted from [2] with permission of ACS, Copyright 2009)

268

Miniaturization of Analytical Systems

to expect that the chemical or physical interactions between a given chemical species and the surface of the nanoparticles will result in changes in the efficiency of the core electron–hole recombination. This has been the basis of the increase in research activity on the development of novel optical sensors based on QD probes. Also, the photoluminescence activation effect (attributed to passivation of surface trap sites that are either ‘filled’ or energetically moved closer to the band edges by this simple chemical process) is induced after adding Zn and Mn ions to colloidal solutions of CdS or ZnS QDs [7,8]. This behaviour provides the basis for optical sensing of such metallic cations with QDs. Besides the activation effect, QD-based optical sensing quenching strategies (based on quenching by the analyte that affects the luminescence emission of the nanoparticle) have been proposed. Quenching mechanisms that explain how metal ions quench fluorescence of QDs include inner filter effects, nonradiative recombination pathways, electron-transfer processes and ion-binding interactions. Examples are the determination of triethylamine and benzylamine in gas [9] and the determination of paraoxon in water [10]. In addition to increasing or decreasing fluorescence, QDs can be exploited for sensing in two other ways. One is the use of nanoparticles in sensor or chemical assay applications based on photochemically-induced fluorescence (or F€orster) resonance energy transfer (FRET) mechanisms and the other is the measurement of the surface-plasmon resonance. As an example, specific binding of different proteins was observed via measurements of FRET between a CdSe-ZnS QD donor attached to one of the proteins under study and some organic acceptor dyes attached to the other. In the presence of specific interactions between both proteins, strong enhancement of the acceptor dye fluorescence was observed [11]. Colloidal metal nanoparticles can have optical absorption spectra with an absorption peak that looks similar to the absorption peak of colloidal semiconductor particles. However, this absorption does not derive from transitions between quantized energy states. Instead, in metal particles, the collective mode of motion of the electron gas can be excited. The members of the electron gas are referred to as surface plasmons. The peak in the absorption spectrum is the resonance frequency for the generation of surface plasmons. The size dependence of the plasmon frequency is negligible. For instance, the absorption maximum for colloidal gold nanocrystals hardly shifts for nanoparticles in the size range 5–30 nm. This contrasts with the dramatic size effect in the absorption of colloidal semiconductor nanocrystals when the diameter is changed by only fractions of a nanometer. In addition to metallic or semiconductor nanoparticles, polymer nanoparticles containing fluorescent dyes have been also used [12]. In a metallic nanoparticle, incident light can couple to the plasmon excitation of the metal. The excitation involves the light-induced motion of all the valence electrons. These nanoparticles provide a powerful and evolving toolkit for biological detection. It has been demonstrated that plasmon resonance is strongly dependent on shape and size [13,14]; therefore it is possible to make a wide range of light

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

269

scatterers that can be detected at different wavelengths. Using specific organic molecules [15] or DNA [16–18], it appears to be possible to sensitively design the particle arrangement, providing a rich system for detection. The electromagnetic field in the near-field region around a metallic nanoparticle is greatly enhanced, providing important new mechanisms for detection. When many gold nanoparticles are located near to each other, their plasmon resonances couple to each other via the near field, shifting the plasmon resonance to higher energy. This change in the optical response from isolated to aggregated metallic nanoparticles can be used to sensitively develop chemical sensors. One example is depicted in Figure 7.5; it is used in the detection of nucleic acids, using gold nanoparticles coated with a high density of oligonucleotides on the surface [19]. It is extremely desirable to be able to optically detect a fingerprint spectrum, but ordinarily this is only possible with vibrational (IR and Raman) magnetic resonance spectroscopies, and none of these has sensitivity anywhere close to what is possible with luminescence. However, the large field enhancement in the proximity of a gold nanoparticle is known to lead to the surface-enhanced Raman spectroscopy (SERS) effect [20]. It is possible to detect a wide range of biological events involving gold nanoparticles coated with specific molecules that offer a distinct Raman spectrum. The SERS effect has been known for some time to provide an enhancement of as great as 105 in Raman cross-section for molecules on a rough gold surface.

Figure 7.5 Gold nanoparticle-based biosensor. The DNA-induced aggregation of gold nanoparticles leads to a shift in the plasmon resonance. (Extracted from [19] reprinted with permission from the American Association for the Advancement of Science)

270

7.3.2

Miniaturization of Analytical Systems

Electrochemical Sensors

The operational modes of electrochemical sensors are based on traditional laboratory electoanalytical principles. In general, two configurations can be considered, with the electrodes commonly designated as indicator and reference electrodes (potentiomatric sensors) and as working and reference electrodes (voltammetric electrodes). These configurations are subjected to the generation of current densities low enough to avoid disturbing the equilibrium potential of the reference. In addition, the ohmic drop through the solution does not significantly alter the applied voltage. Both situations can be easily avoided by the introduction of a third electrode through the current flows. There is no doubt that the advances in microelectronic fabrication techniques have provided impetus to the development of new types of electrochemical sensor. But the incorporation of nanoparticles has, thanks to their properties, also facilitated the improvement of sensitivity and the miniaturization of the sensor. Pseudo-1D nanostructures, such as semiconductor nanowires [21,22], cabon nanotubes [23] and metallic nanoparticles [24], offer the greatest chance yet of creating robust, sensitive and selective electrical detectors. Current flow in any 1D system is extremely sensitive to minor perturbations, and in nanowies and nanotubes the current flows extremely close to the surface. For example, semiconductor nanowires can be functionalized with biological macromolecules and incorporated into electrical circuits [25]. In this configuration, the current that flows in the nanowire is very sensitive to binding events of the macromolecule (Figure 7.6). The precise mechanism on many occasions is not well established. In the example in Figure 7.6 it is supposed that the wire is surrounded by a double layer of ions, and any perturbation to this ionic environment may alter the field experienced by the wire. Similar experiments may be possible with nanotubes, although in such cases care must be taken to separate the complex mixture of metallic and semiconducting nanotubes that is created in the generation process. CNTs are of especial interest, owing to their unique structure-dependent electronic and mechanical properties [27]. CNTs can be classified into SWCNTs and multiwalled carbon-nanotubes (MWCNTs). SWCNTs possess a cylindrical nanostructure (with a high aspect ratio), formed by rolling up a single graphite sheet into a tube. SWCNTs can thus be viewed as molecular wires with every atom on their surface. MWCNTs comprise an array of such nanotubes, concentrically nested like the rings of a tree. The remarkable properties of CNTs suggest the possibility of developing superior electrochemical sensing devices. The redox centre of most oxidoreductases is electrically insulated by a protein shell. Because of this shell, the enzyme cannot be oxidized or reduced at an electrode at any potential. The possibility of direct electron transfer between enzymes and electrode surfaces could pave the way for superior reagentless

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

271

Figure 7.6 Nanowire- and nanotube-based electrical biosensor. Scheme showing silicon nanowires (a) functionalized with biotin and (b) on exposure to streptavidin, and (c) a nanowire that is not functionalized. Scheme showing nanotubes functionalized with (d) biotins (show similar changes) and (e) a quartz-based microbalance signal, and (f) the electrical signal of nanotubes after addition of different concentrations of streptavidin. (Adapted from [26])

biosensing devices, as it obviates the need for cosubstrates or mediators and allows efficient transduction of the biorecognition event. The aligned CNTs in the nanoforest, prepared by self assembly, can act as molecular wires to allow electrical communication between the underlying electrode and redox proteins. In fact, redox proteins can easily be attached to the ends of CNTs [28,29]. Willner’s group demonstrated that aligned reconstituted glucose oxidase (GOx) on the edge of SWCNTs can be linked to an electrode surface [30]. The system is described in Figure 7.7. Such enzyme reconstitution on the end of CNTs represents an extremely efficient approach. Electrons have thus been transported along distances greater than 150 nm, with the length of the SWCNTs controlling the rate of electron transport. The catalytic properties of metal nanoparticles have also facilitated the electrical contact of redox centres of proteins with electrode surfaces. Nanomaterials that transport ions rather than electrons may also form extremely interesting artificial electrical biosensors. Perhaps the most well known proposal is to rapidly sequence a single DNA molecule by electrically sensing the base pairs as they pass by electrodes that are embedded around a nanopore in a semiconductor material [32]. Different polynucleotides can be distinguished from one another as they pass through an alpha hemolysin channel [33], and entirely artificial nanopore/ detector schemes are under active investigation [34]. Nanostructured materials such as 3D-ordered macroporous carbon [35] and SWCNTs [36] have appeared very recently as ion-to-electron transducers in ionselective electrodes with internal solid contact. These sensors based on macroporous

272

Miniaturization of Analytical Systems

Figure 7.7 Assembly of SWCNTelectrically-contacted GOx electrode: linking the reconstituted enzyme (on the edge of the FAD-functionalized SWCNT) to the electrode surface. (From [31], reproduced by permission of the Royal Society of Chemistry)

carbon as transducer were found to show excellent long-term potential stability and insensitivity to O2 and light. Furthermore, these carbon-based materials did not show any evidence of the formation of a water layer between the ion-selective membrane and the transducer. Other performance parameters were comparable to the rest of the ion-selective electrodes with internal solid contact [37]. The nanostructured characteristics of these materials, which are associated with their high surface-tovolume ratio, together with the inert character of the carbon structures, seem to be the intrinsic reason for their outstanding transducing properties. 7.3.3

Magnetic Sensors

More complex physical behaviour, beyond quantum-confined semiconductor systems (single electron-like behaviour), metals (with collective plasmon excitations) or even ordinary electrical devices, arises in systems with correlated electron behaviour, including nanoscale magnetic systems and superconductors. Magnetic

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

273

crystals behave as a single magnetic domain, with all the spins in a crystal coupled together to create a giant magnetic moment. In a very small crystal, at a high enough temperature, this moment wanders randomly (supermagnetic). However, above a critical size, this moment becomes locked in a fixed direction (ferromagnetic). Magentic nanocrystals are widely employed in artificial biological detection, serving important roles as magnetic resonance contrast enhancement agents [38,39]. Superconductors already play an important role in biomedicine, as they form the basis for the magnets used in magnetic resonance imaging (MRI). 7.3.4

Mass Sensors

Microcantilever sensors are most promising for mass micro(nano)sensors. This class of highly sensitive sensors can perform local, high-resolution and label-free molecular recognition measurements [40–42]. They are derived from the microfabricated cantilevers used in atomic force microscopy and are based on the bending induced in the cantilever when, for example, a chemical or biochemical interaction takes place on one of its surfaces. The microcantilevers translate the molecular recognition of biomolecules into nanomechanical motion [43] (from a few to hundreds of nanometres), which is commonly coupled to an optical or piezoresistive readout system [44,45] (Figure 7.8). Microcantilevers are typically made of silicon/silicon nitride or polymer materials, with dimensions of tens to hundreds of nanometres length, some tens of nanometres width and hundreds of nanometres thickness. Moreover, these devices can be fabricated in arrays comprising tens to thousands of microcantilevers, so they

Figure 7.8 (a) Scheme of a general cantilever with biomolecular recognition. (b) Scheme of the optical readout method for a cantilever-bending evaluation. (c) Scheme of the piezoresistive readout and the Wheatstone bridge configuration. (Reprinted from [42] with permission from Elsevier)

274

Miniaturization of Analytical Systems

are a promising alternative to current DNA and protein chips because they may permit parallel, fast, realtime monitoring of thousands of analytes (e.g. proteins, pathogens and DNA strands) without any need for labelling. When fabricated at the nanoscale (nanocantilevers), the sensitivity goes down and expected limits of detection (LODs) are in the femtomole to attomole range, with the astonishing possibility of detection at the single-molecule level in real time [46]. Optical readout is one of the most common schemes for detecting the movement of microcantilevers, as derived from standard AFM. The displacement of the free end of the cantilever is measured using the optical deflection of an incident laser beam on a position-sensitive photodetector, which allows the absolute value of the cantilever displacement to be calculated. This method provides subangstrom  (sub-A) resolution and can be implemented easily. The main disadvantage of this readout technique is that it requires external devices for deflection measurements, and their continuous alignment and calibration is very time consuming. Piezoresistive readout is based on the changes observed in the resistivity of the material of the cantilever as a consequence of a surface-stress change. The sensor measures the variation in film resistance with respect to surface stress caused by the specific binding of molecules. One of the most important advantages of this sensor is that it can be work with nontransparent solutions. In addition, no time is needed for laser alignment. Further, it is compatible with miniaturization and array fabrication. The main disadvantage is the intrinsic noise level, which directly affects the resolution and the sensitivity when compared to optically detected cantilevers [47], although reducing the thickness of piezoresistive cantilevers might increase sensitivity. The covalent adsorption of proteins on gold surfaces of cantilevers can be achieved by a wide variety of chemical procedures, ensuring the reproducibility and the stability of the protein coating. The main disadvantage of chemical adsorption is that some of the functional groups could randomize the orientation of the active sites of the protein, preventing binding with the analyte. Key is the choice of an appropriate immobilization procedure in order to avoid a low rate of activity in the protein receptor. Several strategies could be employed. A very interesting one is covalent immobilization of carboxylate-terminated alkanethiols followed by esterification of the carboxylic groups with a catalyst that facilitates the formation of amide bonds between carboxylic acids and amines, for example the amine function of proteins, to yield the amide bond and the final immobilization of the protein on the surface-coated cantilever. Microcantilever sensors have been widely developed based on DNA hybridization [48,49]. They have been used for more accurate detection of proteins involved in cancer [50] and other diseases [51,52], and in environmental sciences [53]. Cantilever sensors have also been used to detect chemicals such as volatile compounds [54], warfare pathogens [55], explosives [56] and glucose [57], and ionic species such as calcium ions [58].

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

7.4

275

Nanoprobes for In Vivo Bioanalysis

In biochemical science there is no doubt about the importance of developing nanoprobes for in vivo bioanalysis. This type of analysis consists of the use of marked nanopaticles. QDs are probably the most widely used nanoparticles for the development of fluorescent probes for biomedical applications. Compared with organic dyes and fluorescent proteins, semiconductor QDs offer several unique advantages, such as size- and composition-tuneable emission from visible to IR wavelengths, large absorption coefficients across a wide spectral range, and very high levels of brightness and photostability [59]. Owing to their broad excitation profiles and narrow/symmetric emission spectra, high-quality QDs are also well suited for combinatorial optical encoding, in which multiple colours and intensities are combined to encode thousands of genes, proteins or smallmolecule compounds [60–62]. Despite their relatively large size (2–8 nm diameter), bioconjugated QD probes behave like fluorescent proteins and do not suffer from serious binding kinetic or steric hindrance problems. QDs also have a greater surface area and more functionalities that can be used to link to multiple diagnostic and therapeutic agents. Furthermore, polymer-encapsulated QDs have been found to be essentially nontoxic to cells and animals, an essential requirement for future clinical applications.

Figure 7.9 The structure of a multifunctional QD probe. Schematic illustration showing the capping ligand TOPO, an encapsulating copolymer layer, tumour-targeting ligands (such as peptides, antibodies or small-molecule inhibitors) and polyethylene glycol (PEG). (Adapted from [66] with permission from Elsevier)

276

Miniaturization of Analytical Systems

Figure 7.10 Methods for conjugating QDs to biomolecules. (a) Traditional covalent crosslinking chemistry using ethyl-3-dimethyl amino propyl carbodiimide as a catalyst. (b) Conjugation of antibody fragments to QDs via reduced sulfhydryl-amine coupling. (c) Conjugation of antibodies to QDs via an adaptor protein. (d) Conjugation of histidine-tagged peptides and proteins to Ni-NTA-modified QDs, with potential control of the attachment site and QD:ligand molar ratios. (Adapted from [66] with permission from Elsevier)

High-quality QDs are typically prepared at elevated temperatures in organic solvents, such as tri-n-octylphosphine oxide and hexadecylamine. These hydrophobic organic molecules not only serve as the reaction medium, but coordinate with unsaturated metal atoms on the QD surface to prevent the formation of bulk aggregates. As a result, the nanoparticles are capped with a monolayer of the organic

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

277

Figure 7.11 (a) In vivo simultaneous imaging of multicolour QD-encoded microbeads injected into a live mouse. (b) Molecular targeting and in vivo imaging of a prostate tumour in a mouse using a QD–antibody conjugate. (Adapted with permission from [69])

ligands and are soluble only in nonpolar hydrophobic solvents such as chloroform. For biological imaging applications, these hydrophobic dots can be solubilized by amphiphilic polymers. Several polymers have been reported, including octylaminemodified low-molecular-weight polyacrylic acid, PEG-derivatized phospholipids, block copolymers and polyanhydrides [63–68] (see Figure 7.9). Figure 7.10 shows the principal methods for conjugating QDs to biomolecules [66]. In vivo imaging with QDs has been reported for lymph-node mapping, blood-pool imaging and cell-subtype isolation (Figure 7.11) [69]. Ballou and coworkers injected PEG-coated QDs into mouse bloodstream and investigated how the surface coating would affect their circulation lifetime [67]. In contrast to small organic dyes, which are eliminated from the circulation within minutes of injection, PEG-coated QDs were found to stay in the blood circulation for an extended period of time (half-life more than 3 hours). This long-circulating feature can be explained by the unique structural properties of QD nanoparticles. PEGcoated QDs fall within an intermediate size range: they are small enough and sufficiently hydrophilic to slow down opsonization and reticuloendothelial uptake, but are large enough to avoid renal filtration. Webb and coworkers took advantage of this property and reported the use of QDs and two-photon excitation to image small blood vessels [68]. They found that the two-photon absorption cross-sections of QDs are two to three orders of magnitude larger than those of traditional organic fluorophores. As a final remark, QDs have already fulfilled some of their promise as a new class of molecular imaging agents. Through their versatile polymer coatings, QDs have also provided a ‘building block’ for the assembly of multifunctional nanostructures and nanodevices.

278

Miniaturization of Analytical Systems

References [1] J.C. Belmonte, J. Manzano, J. Arbiol, A. Cirera, J. Puigcorbe, A. Vila, N. Sabate, I. Gracia, C. Cane, J.R. Morante, Sens. Actuactors B, 114 (2006) 881. [2] G.A. Crespo, S. Macho, J. Bobacka, F.X. Rius, Anal. Chem., 81 (2009) 676. [3] J.M. Costa-Fernandez, R. Pereiro, A. Sanz-Medel, Trends Anal. Chem., 25 (2006) 207. [4] C.J. Murphy, J.L. Coffer, Appl. Spectrosc., 56 (2002) 16A. [5] A.P. Alivisatos, Science, 271 (1996) 933. [6] H. Weller, Angew. Chem. Int. Ed. Engl., 32 (1993) 41. [7] D.E. Moore, K. Patel, Langmuir, 17 (2001) 2541. [8] K. Sooklal, B.S. Cullum, S.M. Angel, C.J. Murphy, J. Phys. Chem., 100 (1996) 4551. [9] A.Y. Nazzal, L. Qu, X. Peng, M. Xiao, Nano Lett., 3 (2003) 819. [10] X. Ji, J. Zheng, J. Xu, V.K. Rastogi, T.-C. Cheng, J.J. DeFrank, R.M. Leblanc, J. Phys. Chem. B, 109 (2005) 3793. [11] D.M. Willard, L.L. Carillo, J. Jung, A. Van Orden, Nano Lett., 1 (2001) 469. [12] J.R. Taylor, M.M. Fang, S. Nie, Anal. Chem., 72 (2000) 1979. [13] R.C. Jin, Science, 294 (2001) 1901. [14] J.J. Mock, M. Barbic, D.R. Smith, D.R. Schultz, S. Schultz, J. Chem. Physics, 116 (2002) 6755. [15] J.P. Novak, D.L. Feldheim, J. Am. Chem. Soc, 122 (2000) 3979. [16] C.M.M. Daniela Zanchet, W.J. Parak, D. Gerion, A.P. Alivisatos, Nano Lett., 1 (2001) 32. [17] C.J. Loweth, W.B. Caldwell, X.G. Peng, A.P. Alivisatos, Ange. Chem. Int. Ed., 38 (1999) 1808. [18] A.P. Alivisatos, Nature, 382 (1996) 609. [19] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, C.A. Mirkin, Science, 277 (1997) 1078. [20] K. Kneipp, H. Kneipp, I. Itzkan, R.R. Dasari, M.S. Feld, J. Phys. Condens. Matter, 14 (2002) R597. [21] J.T. Hu, T.W. Odom, C.M. Lieber, Acc. Chem. Res, 32 (1999) 435. [22] Y.Y. Wu, Chem. Eur. J., 8 (2002) 1261. [23] T.W. Ebbesen, Annu. Rev. Mater. Sci., 24 (1994) 235. [24] J. Wang, L. Wang, J. Di, Y. Tu, Talanta, 77 (2009) 1454. [25] Y. Cui, Q.Q. Wei, H.K. Park, C.M. Lieber, Science, 293 (2001) 1289. [26] P. Alivisatos, Natur. Biotechnol., 22 (2004) 47. [27] R.H. Baughma, C. Satishkumar, A. Govindaraj, M. Nath, Chem. Phys. Chem., 2 (2001) 79. [28] J.J. Gooding, R. Wibowo, J.Q. Liu, W. Yang, D. Losic, S. Orbons, F.J. Mearns, J.G. Shapter, D.B. Hibbert, J. Am. Chem. Soc., 125 (2003) 9006. [29] F. Papadimitrakopoulos, J.F. Rusling, Electrochem. Commun., 5 (2003) 408. [30] F. Patolsky, Y. Weizmann, I. Willner, Angew Chem. Int. Ed., 43 (2004) 2113. [31] J. Wang, Analyst, 130 (2005) 421. [32] D.W. Deamer, D. Branton, Acc. Chem. Res., 35 (2002) 817. [33] A. Meller, L. Nivon, E. Brandin, J. Golovchenko, D. Branton, Proc. Natl. Acad. Sci., 97 (2000) 1079. [34] O.A. Saleh, L.L. Sohn, Nano Lett., 3 (2003) 37. [35] C.Z. Lai, M.A. Fierke, A. Stein, P. Buhlmann, Anal. Chem., 79 (2007) 4621. [36] G.A. Crespo, S. Macho, F.X. Rius, Anal. Chem., 80 (2008) 1316. [37] C.Z. Zai, M.M. Joyer, M.A. Fierke, N.A. Petkovich, A. Stein, P.J. B€ uhlmann, J. Sol. State Electrochem., 13 (2009) 123. [38] D. Hogemann, L. Josephson, R. Weissleder, J.P. Basilion, Bioconjugate Chem., 11 (2000) 941. [39] C. Wilhelm, F. Gazeau, J.C. Bacri, Eur. Biophys. J., 31 (2002) 118. [40] R. McKendry, J. Zhang, Y. Arntz, T. Strunz, M. Hegner, H.P. Lang, M.K. Baller, U. Certa, E. Meyer, H.J. G€untherodt, C. Gerber, Proc. Natl. Acad. Sci. U S A, 99 (2002) 9783. [41] Y. Arntz, J.D. Seelig, H.P. Lang, J. Zhang, P. Hunziker, J. Ramseyer, E. Meyer, M. Hegner, Ch. Gerber, Nanotechnology, 14 (2003) 89. [42] L.G. Carrascosa, M. Moreno, M. Álvarez, L.M. Lechuga, Trends Anal. Chem., 25 (2006) 196.

Miniaturization of the Entire Analytical Process I: Micro(nano)sensors

279

[43] J. Fritz, M.K. Baller, H.P. Lang, H. Rothuizen, P. Vettiger, E. Meyer, H.-J. G€ untherodt, C. Gerber, J.K. Gimzewski, Science, 288 (2000) 316. [44] R. Raiteri, M. Grattarola, H.-J. Butt, P. Skladal, Sens. Actuators B, 79 (2001) 115. [45] P.G. Datskos, T. Thundat, N.V. Lavrik, Encyclopedia Nanosci. Nanotechnol., 10 (2004) 1. [46] B. Ilic, H.G. Craighead, S. Krylov, W. Senaratne, C.K. Ober, P. Neuzil, J. Appl. Phys., 95 (2004) 3694. [47] X. Yu, J. Thaysen, O. Hansen, A. Boisen, J. Appl. Phys., 92 (2002) 6296. [48] R. McKendry, J. Zhang, Y. Arntz, T. Strunz, M. Hegner, H.P. Lang, M.K. Baller, U. Certa, E. Meyer, H.-J. G€untherodt, C. Gerber, Proc. Natl. Acad. Sci., 99 (2002) 9783. [49] K.M. Hansen, H.-F. Ji, G. Wu, R. Datar, R. Cote, A. Majumdar, T. Thundat, Anal. Chem., 73 (2001) 1567. [50] G. Wu, R.H. Datar, K.M. Hansen, T. Thundat, R.J. Cote, A. Majumdar, Nat. Biotechnol., 19 (2001) 856. [51] Y. Arntz, J.D. Seelig, H.P. Lang, J. Zhang, P. Hunziker, J. Ramseyer, E. Meyer, M. Hegner, Ch. Gerber, Nanotechnology, 14 (2003) 86. [52] A. Savran, S.M. Knudsen, A.D. Ellington, S.R. Manalis, Anal. Chem., 76 (2004) 3194. [53] M. Alvarez, A. Calle, J. Tamayo, L.M. Lechuga, A. Abad, A. Montoya, Biosens. Bioelectron., 18 (2003) 649. [54] M.K. Baller, H.P. Lang, J. Fritz, C. Gerber, J.K. Gimzewski, U. Drechsler, H. Rothuizen, M. Despont, P. Vettiger, F.M. Battiston, J.P. Ramseyer, P. Fornaro, E. Meyer, H.J. Guntherodt, Ultramicroscopy, 82 (2000) 1. [55] B.L. Weeks, J. Camarero, A. Noy, A.E. Miller, L. Stanker, J.J. De Yoreo, Scanning, 25 (2003) 297. [56] L.A. Pinnaduwage, A. Gehl, D.L. Hedden, G. Muralidharan, T. Thundat, R.T. Lareau, T. Sulchek, L. Manning, B. Rogers, M. Jones, J.D. Adams, Nature, 425 (2003) 474. [57] A. Subramanian, P.I. Oden, S.J. Kennel, K.B. Jacobson, R.J. Warmack, T. Thundat, M.J. Doktycz, Appl. Phys. Lett., 81 (2002) 385. [58] H.F. Ji, T. Thundat, Biosens. Bioelectron., 17 (2002) 337. [59] W.C.W. Chan, D.J. Maxwell, X.H. Gao, R.E. Bailey, M.Y. Han, S.M. Nie, Curr. Opin. Biotechnol., 13 (2002) 40. [60] M.Y. Han, X.H. Gao, J.Z. Su, S.M. Nie, Nat. Biotechnol., 19 (2001) 631. [61] X.H. Gao, S.M. Nie, J. Phys. Chem. B, 107 (2003) 11 575. [62] X.H. Gao, S.M. Nie, Anal. Chem., 76 (2004) 2406. [63] B. Dubertret, P. Skourides, D.J. Norris, V. Noireaux, A.H. Brivanlou, A. Libchaber, Science, 298 (2002) 1759. [64] X.H. Gao, Y.Y. Cui, R.M. Levenson, L.W.K. Chung, S.M. Nie, Nat. Biotechnol., 22 (2004) 969. [65] T. Pellegrino, L. Manna, S. Kudera, Nano Lett., 4 (2004) 703. [66] X. Gao, L. Yang, J.A. Petros, F.F. Marshall, J.W. Simons, S. Nie, Curr. Opin.Biotechnol., 15 (2006) 63. [67] B. Ballou, B.C. Lagerholm, L.A. Ernst, M.P. Bruchez, A.S. Waggoner, Bioconjug Chem, 15 (2004) 79. [68] D.R. Larson, W.R. Zipfel, R.M. Williams, S.W. Clark, M.P. Bruchez, F.W. Wise, Science, 300 (2003) 1434. [69] X.H. Gao, Y.Y. Cui, R.M. Levenson, L.W.K. Chung, S.M. Nie SM, Nat. Biotechnol., 22 (2004) 969.

8 Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 8.1

mTAS, Microfluidics and Lab-on-a-Chip: Concepts and Terminology

One of the oldest and most important challenges in analytical chemistry is the accurate measurement of a specific compound in a complex matrix. Generally, this can be carried out by following two main approaches: improving selectivity toward the detection system by using selective (bio)sensors or improving selectivity toward separation systems with nonspecific detectors coupled after the separation of sample mixture in distinct zones of single analytes. Miniaturization of the first approach was studied in Chapter 7, and the miniaturization of the second approach will be studied here. The micro total analysis system (mTAS) concept was developed from a modification of the total analysis system (TAS) approach by downsizing and integrating its multiple steps (injection, reaction, separation and detection) on to a single device, yielding a sensor-like system with a fast response time, low sample consumption, onsite operation and high stability [1,2]. The fact that miniaturized analysis systems contain all the elements needed to perform the required analysis is clearly reflected in the term ‘mTAS’. The recent strong interest in this approach is stimulated by the fact that the attempt to find solutions for chemical measurement problems through development of individual sensors for each of the desired parameters has not been very successful up to now [3]. The main reason for this probably lies in the wide Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

282

Miniaturization of Analytical Systems

variety of parameters and applications in existence, making one general approach for a sensor very difficult. A solution to this would be to develop a relatively expensive ‘overperforming’ analysis system instead of a simple sensor. Apart from its small size, the relevant/key feature of the mTAS concept is the possibility of handling fluidics on the nanolitre and even picolitre scale, which has widened the scope of mTAS so it is now called microfluidics. Microfluidics is the science and technology of systems that process or manipulate small amounts of fluidics (10
9–10
18 l), using channels measuring from tens to hundreds of micrometres [4]. For this reason, the term microfluidics better covers the research and emphasizes the strong impact that miniaturization and integration have on the fluidics and chemical engineering of analytical microsystems. However, miniaturization is more than simply the scaling down of well-known systems, since the relative importance of forces and processes changes with scale. As we will see in the next section, one of the most relevant characteristics of analytical microsystems is the omnipresence of laminar flow (Reynolds numbers are typically very low), in which viscous forces dominate over inertia. This means that turbulence is often unattainable and that molecule transportation only occurs through diffusion, which has direct consequences on the designs of these types of microsystem [5]. The ability to perform laboratory operations on a small scale using miniaturized devices is very appealing [6]. These aspects have been joined under the ‘lab-on-achip’ term, indicating that not only analytical tasks but also synthesis can be performed [7]. Small volumes reduce the time taken to synthesize and analyse a product; the unique behaviour of liquids at the microscale allows greater control of molecular concentrations and interactions; and reagent costs and the amount of chemical waste can be greatly reduced. In consequence, microfluidics exploits both its own most obvious characteristic – small size – and the less obvious characteristics of fluids in microchannels. It offers fundamentally new capabilities in the control of concentrations of molecules in space and time. Independent of the terminology used, mTAS, a microfluidic system and lab-ona-chip all share a series of generic components: a method of introducing reagents and samples; methods for moving these fluids around on the platform and for combining and mixing them; and various other devices (such as detectors for most microanalytical work, and components for purification of products for systems used in synthesis). This drives the integration concept [7,8]. Thus, it should be possible to combine components or subsystems from different developers/suppliers into one microsystem, where the natural tendency exists to look for monolithic integration. Another relevant key feature of the miniaturization of analytical systems, as we mentioned in Chapter 5, is that microfabrication techniques are much needed. Microfluidic devices have not developed as clones of silicon microelectronic devices. Glass and plastics are the main materials used in the fabrication process.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

283

Much of the exploratory research in microfluidic systems has been carried out in a polymer – poly(dimethylsiloxane) (PDMS) – the properties of which are entirely distinct from those of silicon [4]. PDMS is an optically transparent, soft elastomer. The ease with which new concepts can be tested in PDMS, and its ability to support certain very useful components (such as pneumatic valves), have made it the key material for exploratory research and research engineering at the early stages of development. Microelectronic technologies have, however, been indispensable for the development of microfluidics, and as the field has developed, glass, steel and silicon have again emerged as materials with which to build specialized systems that require chemical and thermal stability. Within this context, analytical microsystems are among the parents of microfluidics [4]. Indeed, the field of microfluidics began in microanalytical methods – gas phase chromatography (GPC), high-pressure liquid chromatography (HPLC) and capillary electrophoresis (CE) – which, in capillary format, revolutionized chemical analysis, as we have seen in previous chapters. These methods (combined with the power of the laser in optical detection) made it possible to simultaneously achieve high sensitivity and high resolution using very small sample amounts. With the successes of these microanalytical methods, it seemed obvious to develop new, more compact and more versatile formats for them, and to look for other applications of microscale methods in chemistry and biochemistry. Perhaps the most relevant example is the explosion of genomics in the 1980s, followed by the advent of other areas of microanalysis related to molecular biology, such as high-throughput DNA sequencing, requiring analytical methods with much greater throughput and higher sensitivity and resolution than had previously been contemplated in biology. One important piece of work was offered in Chapter 5 (see Section 5.8.1). Microfluidics offered ways of overcoming these problems. In this chapter we will study the miniaturization of the entire analytical process. Basic concepts of microfluidics will first be given. Then a classification of analytical systems will be proposed. Afterwards, a critical overview of the grade of miniaturization maturity in each analytical step and an overview of real applications will be explored in detail. In addition, the reader is invited to look at [9–11] for further information.

8.2

Basic Concepts of Microfluidics: The Design of Analytical Microsystems

In this section, we are going to introduce the most relevant concepts regarding microfluidic aspects such as types of transport and dimensionless numbers of significance in the design of microsystems. Some examples will be given. For further reading, please check the excellent book chapters [12,13].

284

8.2.1

Miniaturization of Analytical Systems

Types of Transport

There are two very different types of transport within microfluidic systems: directed transport and statistical transport. The difference lies mainly in the nature of the driving agent behind the transport and manifests itself in the form of the transport. Directed transport is controlled by exerting work on the fluid. The work results in a volume flow of the fluid, where the flow can usually be characterized by a direction and a flow profile. The work is often generated mechanically by a pump or electrically by a voltage. Flow that is driven mechanically is called pressure-driven flow, and flow driven by a voltage is called electro-osmotic flow (EOF) (as we have seen in Chapter 5, Figure 5.1). Examples of directed flows are migration and forced convection. Statistical transport differs from directed transport in that it is not directly controlled. Instead, statistical transport is entropy-driven, which means that transport occurs only if a fluid is more disordered after transport than before. A typical situation is at an interface between a liquid with a high concentration of one type of molecule and a liquid with zero concentration of the same molecule. This orderly situation leads spontaneously to a statistical transport of molecules from the side with high concentration to the side with zero concentration. Eventually, the concentration is equal in both liquids, the liquids are evenly mixed and the situation is less ordered than before. This statistical transport is called diffusion. The directed transport of molecules in response to an electric field is called migration. In most cases, the moving molecules are ionized in a polar solvent such as water. These electrically charged molecules experience a Coulomb force due to the electric field. The equation describing this phenomenon, which is based on electrophoresis, was given in Chapter 5 (Equation 5.1). The definition of convection in physics is heat transfer by mass transport. In everyday life, convection is experienced when warm layers of fluids move into colder areas due to the difference in density that is caused by a temperature difference. This phenomenon is also called free convection or natural convection. Fluids can also be moved around with external forces to create a directed flow, which would then be called forced convection. There are several ways in which forced convection can generate directed flow: if a microsystem is to be filled with its working liquid, but is still empty, then the first type of directed flow is a capillary flow, filling up the microsystem. In a similar fashion, other forces, such as gravity, a pressurized air bladder or the centripetal forces in a spinning disc, create a single instant of pressure difference in the microsystems. One very important aspect to consider is the role of diffusion in microdevices, most specifically diffusion perpendicular to the direction of flow of the fluid. This ‘transverse diffusion’ is important in all microchannels, but is often overlooked in devices that perform chromatographic separations because under the best of circumstances the plug of sample is uniform from wall to wall of the channel. In the devices to be discussed, the concentration is initially very much not uniform across

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

285

the channel, and typically the flow is induced by pressure gradients, so the flow itself is also not uniform. This makes for a very rich set of problems, and an equally rich set of behaviours that can be exploited for practical purposes. As we have already mentioned, the process of diffusion is statistical by nature. In a liquid or gas, all molecules move in all directions as long as no external forces are applied. Each molecule moves in a certain direction for a certain time until it is hit by another molecule, whereupon it changes direction. Because the molecules are indistinguishable from one another, no net flow is observable. Diffusion occurs when there is a concentration gradient of one kind of molecule within a fluid. For example, when a concentration gradient of large molecules is present in water, one may notice net movement of large molecules away from areas of high concentration toward areas of low concentration. The main reason for this is that there are more large molecules in areas of high concentration than in areas of low concentration, so many molecules move randomly in one direction, but only a few molecules move randomly back. The statistical movement of a single molecule in a fluid can be described as a ‘random walk’. This movement is characterized by the Einstein–Smoluchowski equation: pffiffiffiffiffiffiffiffi x ¼ 2Dt; ð8:1Þ where x is the average distance moved after an elapsed time t between molecule collisions and D is a diffusion constant that is characteristic for the given molecule. The most important conclusion that can be deduced from Equation 8.1 is that the moved distance is proportional only to the square root of time, in contrast to directed movement, in which the distance is directly proportional to time. This finding can be attributed to the fact that the molecules move randomly and not in a directed way. Equation 8.1 also shows that a larger diffusion constant means faster movement. In general, it is true that the larger a molecule, the smaller its diffusion constant. Equation 8.2 allows us to estimate the time tcross it takes a molecule to cross half the channel width W of the main channel of a T-junction: W2 ð8:2Þ 8D This means that the diffusion dimension, here the channel width, of a microchannel has a large influence on the mixing of liquids in that channel. tcross ¼

8.2.2

Laminar and Turbulent Flows

Two dimensionless numbers have special significance in analytical microsystems. The first is the Reynolds number, which informs the behaviour of the fluidic in terms of either chaotic or steadier, more well-behaved flows. The second is the P
eclet number, which informs the dominance of directed versus statistic transports.

286

Miniaturization of Analytical Systems

The dimensionless Reynolds number is given in Equation 8.3: Re ¼

rdn ; h

ð8:3Þ

where d is the typical length scale (e.g. the diameter or the channel depth) and v is the average velocity of the moving liquid. With this number we can obtain an impression of the flow behaviour in microsystems. From empirical observations, physicists and engineers have found that Reynolds numbers larger than about 2300 correspond to what is called turbulent flow. Under this regime, inertial forces are dominant. As we can see from Equation 8.3, large Reynolds numbers are attained at higher liquid densities, higher flow velocities, larger typical length scales or lower viscosities. The region in which the Reynolds number is between about 2000 and 3000 is called the regime of transitional flow, and the region in which it is below about 2000 is referred to as the laminar (or creeping) flow regime. Again, by examining Equation 8.3 we find that low Reynolds numbers are attained at lower velocities, smaller dimensions, smaller densities or higher viscosities. In consequence, a decrease of scale drives inherently to obtain a laminar flow. On the other hand, to evaluate the various flow situations, we can examine the ratio between the mass transport due to directed flow and that due to diffusion. This ratio is the dimensionless P
eclet number, given as: nd ; ð8:4Þ D where d is a characteristic length of the microfluidic system. Knowing the P
eclet number of a microchannel and its dimensions, the designer can distinguish between two major situations. First, if the P
eclet number is much smaller than 1, diffusion dominates the microfluidic flow and directed flow is of secondary importance. Second, if the P
eclet number is much larger than 1, the molecules of interest flow mainly according to the externally applied driving force and diffusion has only a minor influence. In microsystems, the flow velocities are usually comparatively small. The crucial variable that determines the P
eclet number is therefore the channel length, d. For long-enough channels, the P
eclet number is always larger than 1, and the flow is consequently directed. Pe ¼

8.2.3

The Design of Microfluidic Systems

Microfluidic devices generally operate in the presence of flow, so diffusional transport is complemented by convectional/direct transport. However, at low-Reynoldsnumber conditions (<1, which usually are present), the absence of turbulence both simplifies the transport and slows it substantially; transport perpendicular to flow relies entirely on diffusion. This is both a problem and an opportunity. The opportunity is to design experimental conditions in which the very slowness of

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

287

diffusion of larger solutes over defined distances can be exploited. It is this feature that distinguishes truly microfluidic devices and systems. From the previous two sections it is clear that the two most important phenomena we face in microfluidic devices are the almost exclusive presence of laminar flow and the significance/dependence on diffusion as the major transport mechanism. We should stress, however, that turbulent flow may occur under certain rare conditions, for example when geometries change rapidly and do not allow the flow to reach a steady state between changes. Let us now look at the possibilities for using laminar flow and diffusion in a constructive and creative way to implement functional elements and processes inside a microfluidic device along all analytical steps. We are going to introduce some examples selected for their ‘historical’ significance. Mixing is one of the challenges in microfluidic devices. All passive mixing (i.e. without external energy input) relies primarily on diffusion. The T-shaped channel is perhaps the simplest fluidic device for mixing two solutions [14] (Figure 8.1(a)). In a sense, the worst possible mixer is one in which two fluids are brought into contact and equilibration occurs simply by diffusion perpendicular to the direction of flow. A perfect device for performing such mixing

Figure 8.1 Selected microfluidic devices. (a) A simple interdiffusional mixer with two inlets and one outlet [14]. (b) A ‘filtering’ device utilizing laminar flow and diffusion (termed H-filter) (Reproduced from [12] with permission of Wiley-VCH). (c) Flow focusing by use of two laminar side flows (i) and various flow ratios (ii–iv) (Reprinted from [16] and adapted with permission of American Chemical Society, Copyright 1997)

288

Miniaturization of Analytical Systems

of two solutions by diffusion only would consist of two blocks of gelatin, each containing a different solution. The polymeric network of the gelatin eliminates convection in the fluids but does not inhibit the diffusion of either the solvents or moderately large solutes. It is possible to put the two blocks of gelatin into contact, beginning interdiffusion that will continue until the contents of the two blocks have become identical and homogeneous. The convection in real microfluidic channels leads to many secondary phenomena not present in the blocks of gelatin, but it is often helpful to idealize the microfluidic devices in this way to understand the most basic phenomena. With the knowledge of how diffusion time and contact time can be influenced by proper design of a microfluidic layout, we can easily devise ways to make use of this. A simple yet effective example is the so-called H-filter [15] (Figure 8.1(b)). It has two inlet channels, a section where the two liquids are laminated and flow side by side, and then two exit channels. If liquid A contains a mixture of two different molecular species, one with a small diffusion coefficient and one with a rather large diffusion coefficient, and liquid B contains pure solvent, the following happens at the H-crossbar where both liquids are laminated: depending on the width of the crossbar and the flow velocity, only a very tiny fraction of the molecules with a small diffusion coefficient diffuse into liquid B in the given contact time. On the other hand, a much larger fraction of the molecules with a large diffusion coefficient can diffuse into liquid B. Another very popular design is a layout that features three inlet channels merging into a common channel, and then, depending on the intended use, one to several exit channels [16]. It is not important how the channels are arranged geometrically, whether merging at an acute angle or in a right-angle geometry. On the inlet side, the liquid of interest typically enters the middle channel, and two flanking liquids can be used to guide the centre liquid and even form or compress it slightly – all by using the fact that only laminar flow occurs in microsystems. Control is achieved fairly easily by merely adjusting the relative or absolute flow rates (volume flow) of all the flows involved. For example, to form or narrow the centre flow to a greater or lesser extent, the laminating flow rates can be increased or decreased. In this way it is even possible to increase the concentration of a chemical species present in the centre flow, right at the point where the two focusing side streams join the centre stream (Figure 8.1(c)). The high fluorescent background problem in protein analysis by CE has been cleverly solved in a microfluidic chip using an integrated on-chip ‘destaining’ step [17]. After the protein fractions have been sufficiently resolved in the separation column, an intersecting channel structure is introduced. The side channels are filled with the same polymer-sizing matrix as in the main separation channel, but the buffer solution contains no sodium docecyl sulphate (SDS). By applying an electric current from the side channels into the main channel, two focusing streams containing no SDS are induced, creating a concentration gradient

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

289

in SDS transverse to the flow direction downstream of the intersection. By transverse diffusion, free SDS micelles in the centre of the channel diffuse laterally outward into the zones of low SDS concentration, causing a breakup of the micellar structure when the local concentration of SDS drops below CMC (critical micellar concentration). As free micelles break up, fluorescent dyes are released. Using dyes that do not fluoresce when they are dissociated from the micelles, the high background drops to a low level. This destaining process inside the microchannel network is photographically captured on the right-hand side. Bound SDS micelles, on the other hand, cannot migrate outward as readily, owing to the large protein size, and they tend to remain stable because of the association with the proteins. Consequently, dyes captured in the protein–SDS complexes remain highly fluorescent after destaining. Finally, let us look at an exciting and impressive example of how to use lamination and diffusion to create truly tiny structures within already-microfabricated channels. As two liquids are laminated, a reaction can take place at the interface between them. The ‘reaction zone’ is limited to a small region around the interface, defined by diffusion. In this way it is possible to demonstrate, for example, the etching of a small trench into the bottom of a microchannel (Figure 8.2(a)) [18]. To achieve this, the researchers laminated a solution of potassium fluoride and a solution of hydrochloric acid, creating hydrofluoric acid at the interface. This acid

Figure 8.2 Microfluidics and microfabrication (Reprinted from [18] with permission of American Association for the Advancement of Science). (a) In situ generation of hydrofluoric acid to etch a trench in the middle of a microchannel. (b) In situ creation of a silver wire inside a microchannel. (c) In situ creation of an electrochemical cell with working, counter and reference electrodes inside a microchannel

290

Miniaturization of Analytical Systems

etched the substrate material, producing a trench at the lateral position of the interface between the two solutions. In a similar way, the researchers could deposit a tiny silver wire at the interface of a silver salt solution and a reducing solution (Figure 8.2(b)) [18]. Beyond these mere demonstrations, they produced a miniaturized electrochemical cell within a channel using the approaches described. They first selectively etched a gold patch on the bottom of the channel by laminating a gold etching solution with two buffer streams (thereby protecting part of the gold), which were to become the working and counter electrodes. Then, again by laminating two suitable solutions, they deposited a silver wire in between the separated gold patches, thereby creating a reference electrode and finishing the ED cell (Figure 8.2(c)) [18].

8.3 The Basics of Downscaling in Microsystems The two main approaches used to understand the downscaling of analytical systems are dimensionless and similarity approaches [2]. Dimensionless parameters define units such as volume, column length, linear flow rate, retention time and pressure drop in terms of quantities that can be assumed to be constant over the entire system. For decades, engineers have used dimensionless parameters to correlate experimental results when large numbers of significant variables are involved. In the case of chromatography using an open-tubular column (in which the columns are not packed, but instead the walls of the column support the stationary phase and a hollow core allows the mobile phase to move freely), the constant quantities include inner diameter, mobile phase viscosity, average diffusion coefficient of a sample in the mobile phase and Poiseuille number for a circular crosssection. The dimensionless-parameters approach can be applied as long as the contribution of other physical phenomena not considered in the parameter can be neglected. Nonlinear behaviour can be usefully described with the aid of dimensionless parameters such as the P
eclet number (previously introduced); n, which is the ratio of axial bulk flow to diffusion mass transport (effectively describing the relative importance of convective or diffusive transport, or change in flow rate); the Fourier number, t, which describes the average number of times a molecule contacts the wall of the capillary (elution time or mixing efficiency); and the Bodenstein number, P, which characterizes the backmixing within a system (or pressure drop). These parameters make it possible to extrapolate results obtained for one system to other similar systems through multiplication by constant factors, as shown in Table 8.1. So even before starting an experiment, we can establish whether or not the device will function. For instance, miniaturization to channel sizes of 1.5 mm results in a pressure drop of 3900 bars, which is difficult if not impossible to provide.

2225 188 47

13 400

6750

10 200

1.3

17

670

3900

0.67 0.28

1.5

Reduced variance in terms of volume, sv ¼ sv =dc3

Reduced variance in terms of time, st ¼ st Dm =dc2

Reduced column length, l ¼ L/dc Fourier number (reduced retention time), t ¼ tD=dc2 Bodenstein number (reduced pressure drop), P ¼ pdc2 =FhmDm Reduced variance of elution profile in terms of length, sx ¼ sx/dc

Relationship

Dimensionless parameters

350

7.5

450

280

450 000 7500

Calculated values

Calculated parameter sets for an open-tubular column liquid chromatography (LC) system with one million theoretical plates at zero retention as capillary diameter changes (P
eclet number, v ¼ 38). The diffusion coefficient is assumed to be Dm ¼ 10
9 m2 s
1; the viscosity is hm ¼ 10
3 Ns m
2; and the Poiseuille number, F, has a value of 32 for a circular cross-section. An open-tubular LC system with zero retention was chosen for numerical simplicity (adapted from [2])

350

10

Pressure difference, Dp (bar) Peak standard deviation in terms of length, sx (mm) Peak standard deviation in terms of time, st (ms) Peak standard deviation in terms of volume, sv (pl)

2.3 3.13

13.5 122

Length, L (m) Time, t (min)

5

30

Parameters

Diameter, dc (mm)

Table 8.1 Examples of the dimensionless-parameters approach

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 291

292

Miniaturization of Analytical Systems

As we have already mentioned, because the Reynolds number is proportional to the characteristic length d, for miniaturized systems a small value will be obtained, indicating laminar flow. Such parameters are useful for controlling separations and for mixing to optimize the output of chemical reactions. In the similarity approach, useful information about the behaviour of simple miniaturized flow systems can be attained by considering the proportionalities within a system. The parameters of interest, such as flow rate, pressure drop and electric field, are viewed as a function of the variables to be miniaturized in space and time. Using this approach, the major trends a parameter undergoes during its downscaling become apparent, with no knowledge of material constants (such as viscosity or heat capacity) required. For any system, it is possible to proceed from a given point, such as an experimental result, and extrapolate an estimated magnitude for the variables in a downscaled system. Changes in geometry only influence this estimation by a constant factor. Although this approach cannot be used to predict the feasibility of a system, it can in principle allow physically impossible cases to be rejected and provide an idea of the order of magnitude of relevant variables. If it is assumed that miniaturization is a simple three-dimensional downscaling process characterized by a typical length parameter d [1], we can easily predict the behaviour of the relevant physical variables. The typical length d represents the scaling factor of the miniaturization. A single degree of freedom for the mechanical parameters remains: time. For simplicity, only two important cases are considered and discussed below. If the timescale is the same for the miniaturized system as for the full-scale system, it is referred to as a time-constant system (Table 8.2). Relevant time variables (such as analysis time, transport time and response time) remain the same. However, linear flow rate in a tube would decrease by a factor d, volumetric

Table 8.2 Examples of the similarity approach Parameter-related system

Time-constant system

Diffusion-related system

Space, x Time, t Linear flow rate, u Volume flow rate, F Pressure drop (laminar flow), Dp Voltage (electro-osmotic flow), U Electric field, U/L Reynolds number, Re Peclet number (reduced flow rate), v Fourier number (reduced elution time), t Bodenstein number (reduced pressure), P Reduced voltage, V

d Constant d d3 Constant d2 d d2 d2 1/d 2 d2 Constant

d d2 1/d d 1/d 2 Constant 1/d Constant Constant Constant Constant Constant

Proportionality factors are given for mechanical parameters in relation to a characteristic length, d (adapted from [2])

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

293

flow rate by d3 and the Reynolds number by d2. By contrast, the pressure drop needed to maintain the desired flow rate would remain the same. This time-constant behaviour is important for simple transportation and for analytical techniques such as flow-injection analysis (FIA) systems. Diffusion would certainly have a more dominant role in mass flow in miniaturized systems. The main advantage in downscaling simple transport or FIA systems lies in the conservation of carrier and reagent solutions. A 10-fold decrease in size, for example, would cause a 1000fold decrease in carrier or reagent consumption. The diffusion-related system becomes important when molecular diffusion, heat diffusion or flow characteristics dictate the separation efficiency in a given system. In such instances, the timescale is treated as a surface that is proportional to d2. This behaviour is in perfect agreement with standard chromatographic and electrophoretic band-broadening theory. All dimensionless parameters, including the P
eclet number, Fourier number and Bodenstein number, remain constant regardless of the size of the system. In other words, hydrodynamic diffusion, heat diffusion and molecular diffusion effects behave in the miniaturized system exactly as in the original system. This means that downscaling to one-tenth of the original tube diameter reduces related time variables such as analysis time and required response time for a detector to one-hundredth of their original magnitudes. The pressure drop requirements increase by a factor of 100, but the voltage requirements (for electrophoresis or electro-osmosis) remain unchanged. The main advantage of the diffusioncontrolled system is that miniaturization achieves faster separations while maintaining comparable separation efficiency. An initial experiment in this field was conducted in 1992 [19]. Finally, the predictions of similarity laws (for example, the same performance of a separation in a shorter time) were experimentally validated when electrophoresis-based separations of amino acids with up to 75 000 theoretical plates were obtained in about 15 seconds using a microchip system [20]. But where is the limit? The absolute limit of miniaturization with respect to separation techniques depends on the maximum applicable pressure (which is dependent on the stability of the used material) in the case of chromatography; on the maximum voltage that can be applied in the case of electrophoresis; and on the molecular characteristics that affect, for instance, sterical behaviour, viscosity and surface tension. We emphasize that it is impossible to miniaturize further than to the level of a single molecule.

8.4

Microfluidic Platforms: Types, Principles and Classification

A microfluidic platform provides a set of fluidic unit operations (fluidic handling), which are designed for easy combination within a well-defined fabrication

294

Miniaturization of Analytical Systems

technology, where the monolithic approach is the most commonly used. The platform allows the implementation of different application-specific systems (assays) in an easy and flexible way, based on the same fabrication technology. In consequence, a microfluidic platform (very often planar and monolithic) cleverly integrates the required operations during the analysis. These units are metering, mixing, incubating and so on; in other words, all physical and chemical operations for the performance of sample introduction, preparation, separation and detection. In this section, we intend to give an overview of microfluidic platforms, focusing only on platforms for lab-on-a-chip applications in the analytical chemistry domain, while being aware that there are other possible fields of application, such as chemical synthesis microprocess engineering and microdosage systems. Even in the field of lab-on-a-chip systems we cannot cover all the microfluidic platforms known from the literature. In addition, we would like to underline that the term ‘labon-a-chip’ involves both electrokinetic and pressure-driven systems, although some literature gives ‘lab-on-a-valve’ for pressure systems. From a technological point of view, four to five types of analytical microsystem in microfluidic platforms could be properly given [21], as summarized in Table 8.3. For each, a brief description of their principles, analytical and technological features, and main analytical fields of application is given. An in-depth description of unit operations such as metering, mixing and so on is outside the scope of this chapter and other readings are recommended [9,10,21]. In addition, the fundamentals of the main microfluidic platforms have been given in Figure 8.3. An introduction to the basic principles and the main advantages and disadvantages of each type of platform will now be given, from the early test strips to the impressive ‘labs on droplets’. For further information, the reader is invited to check the excellent review in [21]. 8.4.1

Capillary-driven Test Strips

Test strips, or ‘lateral flow assays’ as they are also called, have been known in the diagnostic field since the 1960s, representing the state of the art with billions of units being produced at the lowest costs, and embracing the concept of lab-on-a-chip. Although they can be regarded as the most successful microfluidic platform for labon-a-chip applications in terms of the number of commercialized products (e.g. diabetes testing, pregnancy testing, etc.), hardly any publications exist from a microfluidic point of view, despite the fact that the complexity of test strips varies from a single fleece (i.e. non-woven material featuring high capillarity), for example for pH measurement, to very complex and partially microstructured configurations of multiple fleeces that enable the implementation of more complex tests like immunoassays. However, these platforms can be looked at as mirrors in which the concept of lab-on-a-chip is clearly shown, including the idea of use by nonspecialized people at the ‘point of care’.

Planar surface

Electrowetting-driven Surface acousticdriven

Presssure-driven

Electric and eletro-osmotic forces

Electrokinetic

Droplet-based Channel-based

Directed centrifugal force

Passive liquid transport via capillary forces Direct capillary filling Presure-driven system

Principle forces & weaknesses

Centrifugal

Microfluidic largescale integration (LSI)

Capilllary-driven strips

Microfluidic platform

High flexibility since liquid processing paths can be freely controlled.

No flexibility: channel geometry is fixed

Flexible and configurable technology (versatility) Inherent limitations of PDMS material External pumping and valves required Modular setup of the system with cheap, disposable and easily exchangeable plastic cartridges No flexibility Flexible and configurable technology (versatility) High voltages Valveless Microsystems Hard to accomplish in handheld devices High-sensitivity detection systems required (a few molecules reach detector area)

Energy independence Handheld device

Technical features

Table 8.3 Microfluidics platforms in analytical chemistry

Miniaturization, integration and automation

High-resolution analysis and extremely low sample consumption Parallel analysis

Parallel analysis Integration of multiple liquid-handling steps Parallel analysis

Less sensistivity and reproducibility

Analytical features

High-throughput screening applications

Clinical, bionalytical, environment, security, food DNA analysis

Drug screening Clinical (alcohol in blood)

Clinical assays (point-of-care testing) gold standard Bioanalytical, DNA extraction

Fields of analytical application

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 295

296

Miniaturization of Analytical Systems

Figure 8.3 Simple schemes of the main principles of microfluidic devices: (a) hydrodynamic (i) and electrokinetic (ii); (b) centrifugal; (c) droplet-based microfluidic platforms – channel-based (i) and surface-based (ii) (Reproduced from [21] with permission of the Royal Society of Chemistry)

The basic principle of the platform is passive liquid transport via capillary forces within the capillaries of a fleece or a microstructured layer. The liquid samples are loaded into a start reservoir, from which they penetrate the underlying fleeces. Another method, particularly applied in patient self-testing applications, is the direct capillary filling of the strip from the sampling point. The main advantage of this platform is the possibility of performing an automated onsite measurement using a cheap, small, disposable device, combined with the simple actuation principle, which means it does not need any energy supply, giving it a huge potential for point-of-care and patient self-testing applications. Drawbacks arise from its simplicity: assay protocols within capillary-driven systems follow a fixed process scheme, imprinted in the microfluidic channel design. Passive liquid propulsion by capillary forces only cannot be influenced actively once the process is started and as a consequence the exact timing of the assay steps depends on variations in the viscosity and surface tension of the sample. Other crucial unit operations are metering and incubation, the accuracies of which are limited, and mixing, which cannot be accelerated on the test strip platform. A further critical point is the long-term stability of the wetting properties inside the fleeces or microstructures, which depends on their surface properties and temperature. 8.4.2

Microfluidic Large-scale Integration (LSI)

The best example of a pressure-driven microfluidic system is the microfluidic largescale integration (LSI) platform, which arose together with a novel fabrication technology for microfluidic channels, called soft lithography. Using that technology, the monolithic fabrication of all necessary fluidic components within one single elastomer material (PDMS) became possible, in a similar way to the siliconbased technology in microelectronics. As we have already mentioned in Chapter 5, PDMS is an inexpensive but still powerful material, offering several advantages

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

297

compared with silicon or glass. It is a cheap, rubber-like elastomer with good optical transparency and biocompatibility. It can be structured using the soft lithography technique based on replication moulding on micromachined moulds. It was first used by George Whitesides’ group for the fabrication of optical devices [22] and stamps for chemical patterning [23,24]. Thereafter, microfluidic devices were manufactured using the PDMS technology [25–29]. Since then, however, PDMS has been used as a merely passive material for the construction of microfluidic channels only, as mentioned in Chapter 5. The strength of the technology really became obvious when it was expanded into the multilayer soft lithography (MSL) process [30,31]. With this technology, several layers of PDMS can be hermetically bonded on top of one another, resulting in a monolithic, multilayer PDMS structure. Today, this technology is pushed forward by the Fluidigm Corporation, CA [32]. Based on the high elasticity of PDMS, the basic microfluidic unit operation is a valve which is made of a planar glass substrate and two layers of PDMS on top of each other. The lower elastomer layer contains the fluidic ducts and the upper elastomer layer features pneumatic control channels. The microfluidic LSI platform certainly has the potential to become one of the foremost microfluidic platforms for highly integrated applications. It is a flexible and configurable technology which stands out due to its suitability for LSI. The PDMS fabrication technology is comparably cheap and robust and can be used to fabricate disposables. Reconfigured layouts can be assembled from a small set of validated unit operations, and design iteration periods for new chips are on the order of days. Some of the system functions are hardware-defined by the fluidic circuitry, but others, such as process sequences, can easily be programmed from outside. Limitations of the platform are related to the material properties of PDMS: for example, chemicals that are not inert to the elastomer cannot be processed, and elevated temperatures such as those in microreaction technology are not feasible. Also, the implementation of applications in the field of point-of-care diagnostics, where often a handheld device is required, does not seem to be beneficial using the LSI platform. The external pressure sources and valves have to be shrunk to a smaller footprint, which is technically feasible of course, but the costs would be higher than in other platform concepts. Figure 8.4(a) shows a representative example of a microsystem based on this approach. 8.4.3

Centrifugal Microfluidics

A new generation of centrifugal devices has emerged from the technical capabilities offered by microfabrication and microfluidic technologies based on directed centrifugal force [33] (see Figure 8.3b). Length scales of the fluidic structures in the range of a few hundred micrometres allow parallel processing of up to one hundred units assembled on a disc. This enables a high throughput of many tests by highly parallel and automated liquid handling. In addition, the new opportunities

298

Miniaturization of Analytical Systems

Figure 8.4 Selected examples of microfluidic platforms (adapted). (a) Construction of the main unit operations on the multilayer PDMS-based LSI platform. The NanoFlex valve, as depicted on the left, can be closed by applying a pressure p to the control channel: (i) microfluidic valves; (ii) peristaltic pumps; (iii) mixing structures (Reproduced from [21] with permission of the Royal Society of Chemistry); (b) Microfluidic realization of CE analysis on the electrokinetic platform. After the sample has been transported to the junction area (1) it is metered by the activated horizontal flow and injected into the separation channel (2). There, the sample components are electrophoretically separated (3) and read out by their fluorescence signal (4). The complete microfluidic CE chip is depicted on the right. (Copyright Agilent Technologies Inc. Reproduced with permission)

arising from the miniaturization of the centrifugal fluidics cut down assay volumes to less than 1 ml. Fields such as drug screening, in which precious samples are analysed, will particularly benefit from the low assay volumes. The modular setup, with cheap, disposable and easily exchangeable plastic cartridges, is certainly one major advantage of the centrifugal microfluidic platform. The cost-efficient fabrication predominantly originates in the simple and passive microfluidic elements, which can easily be combined monolithically within the same fabrication process. These elements allow the implementation of all necessary unit operations in order to perform complex assay protocols in an automated way. Owing to the rotational symmetry of the discs, a high degree of parallelization can optionally be achieved. All processes are controlled by the frequency of rotation of one single macroscopic rotary engine. As soon as any additional actuation or sensing function is required on the module while it is rotating, things become tricky from a technical point of view if a contact-free

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

299

interface is not applicable. The platform also lacks flexibility compared with others that allow online programming of fluidic networks within a single piece of hardware. Most of the logic functions, as well as their critical frequencies, are permanently imprinted into the channel network. 8.4.4

Electrokinetic Platforms

Electrokinetic pumping (and particle manipulation) principles are based on surface forces and thus gain impact within the microdimensions due to the increased surface-to-volume ratio. This advantage, combined with the simple setup of electrokinetic systems, which basically consist of microfluidic channels and electrodes, explains the early advent of microfluidic lab-on-a-chip applications based on the electrokinetic platform. These focused on the analysis of chemical compounds via electrophoretic separation within microchannels (CE) [19,34–36]. In consequence, CE microchip is conceptually a miniaturized separation technique (see Chapter 5) as a particular case of an electrokinetic platform which conceptually involves more than just electrophoretic separations. Similarly, as we have already explained in Chapter 5, the principle of fluid propulsion (fluid transport) on the electrokinetic platform is based on the movement of the liquid layer right at the interface to the solid phase (electric double layer), initiated by an external voltage. Standard channel materials for electrokinetic actuation are glass and polymers that possess negatively charged surfaces, thus causing a surplus of positively charged liquid molecules in the double layer close to the channel walls. As soon as an electric potential is applied along the channel, the positively charged liquid molecules are attracted by electrostatic forces and thus move toward the negative electrode. As a result of the viscous coupling, the bulk liquid is dragged by the moving layer and a planar velocity profile (v) evolves. This is a major difference compared to the parabolic velocity profile of pressure-driven flows. Consequently, a rectangular-shaped liquid plug with limited dimensions along a microchannel gets increasingly widened within a pressure-driven flow, while it keeps its original flat shape within an EOF. So sample dispersion is drastically reduced in EOFs, making EOF the method of choice for separation analysis. Besides electro-osmosis, other effects also occur within the electric field as soon as the liquid contains electrically charged particles or molecules. These are attracted (F) by one of the electrodes, depending on their charge and valence, and consequently move toward the electrode in a stationary surrounding liquid. The velocity of the molecules depends on their charge and size, enabling a distinction between different species. This effect is called electrophoresis and is used for separation. The electro-osmotic actuation of liquid flows enables pulse-free pumping without any moving parts. Additionally, no dispersion occurs in the EOF and thus sample plugs are not broadened during separation. Miniaturization of electrophoretic

300

Miniaturization of Analytical Systems

analysis enables the automation and parallelization of tests with small dead volumes, thus reducing the required amount of sample. Furthermore, higher voltages can be used for the separation, and the dissipation of heat is increased compared with macroscopic systems due to the higher surface-to-volume ratio. Overall, miniaturized electrophoresis enables the fast and efficient analysis of (bio)molecules. Drawbacks can be seen in the need for high-performance detection technologies due to the reduced volumes and thus signals. Also, technical problems arise in CE systems due to pH gradients arising over time (chemistry of the system) or during operation, streaming currents which counteract the external electric field and gas bubbles that can occur due to electrolysis at the electrodes. In addition, high voltages are needed, which can hardly be generated in mobile handheld devices, for example. Figure 8.4(b) shows a representative example of a microsystem based on an electrokinetic platform. 8.4.5

Droplet-based Microfluidic Platforms

The principal idea behind droplet-based microfluidic systems is the use of single droplets as reaction confinements for (bio)chemical reactions/analyses. Dominant interfacial and surface-tensional forces in the microdimension enable the precise generation and spatial stabilization of these droplets. Since the droplets are kept isolated within an immiscible surrounding fluid, lateral dispersion (Taylor dispersion) can be avoided while moving them to different locations. A multitude of parallel screening reactions, each consuming only a minute amount of reagent, are enabled inside several small-sized droplets on the platform. Droplet-based microfluidic systems can be divided into two basic setups, the channel-based approach and the planar surface approach, as described in Figure 8.3(c). The channel-based systems are mostly pressure-driven, with the droplet generation and manipulation relying on actuation via liquid flows within closed microchannels. On the planar surface-based platforms, droplets can be arbitrarily moved in two dimensions, representing planar programmable laboratories onchip. They are actuated by electrowetting or surface acoustic waves (SAWs), respectively. Channel-based Approach The pressure-driven, droplet-based platform relies on a two-phase fluid flow through the microchannels. The two immiscible phases are dispersed into each other so that a sample fluid (e.g. aqueous solution) forms plugs of a certain length, separated by the carrier fluid (e.g. oil) along the channel. This flow scheme is called segmented flow, since the size of the inner phase droplet exceeds the cross-sectional dimensions of the channel leading to the squeezed fluid plugs. The two-phase flow is pumped throughout the channels by an externally-applied pressure.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

301

The most elementary unit operation on the pressure-driven, droplet-based platform is the initial generation of the droplets. This step can also be considered as metering, since the liquid volumes involved in the latter reaction within the droplet are defined during the droplet formation process. Generally, two different microfluidic structures have been reported for a controlled droplet generation: the flow-focusing structure [37] and the T-shaped junction [38]. The size of the droplet is influenced by the channel dimensions and the strength of the shear forces at the channel junction (higher shear forces lead to smaller droplets) in both droplet formation mechanisms. In order to use droplets inside channels as reaction confinements, the different liquid streams have to be loaded into the droplets first. A method for combining three different sample liquid streams by a sheath flow arrangement with subsequent injection as a common droplet into the carrier fluid has been shown [39]. Different concentrations and ratios of two reagent substreams plus a dilution buffer merge into one droplet and perform a so-called on-chip dilution. The mixing ratios can be adjusted by the volume flow ratio of the three streams. Planar Surface Approach The electrowetting effect was first described by Gabriel Lippmann in 1875, while recent developments were initiated in the early 1990s by the introduction of the idea of using thin insulating layers to separate the conductive liquids from the metallic electrodes in order to eliminate electrolysis. That paved the way for the application of the electrowetting effect as a liquid propulsion principle for lab-on-a-chip systems. The electrowetting-on-dielectric (EWOD) [40] technology is based on a liquid droplet placed between two electrodes covered with insulating dielectric layers. An applied voltage between the two electrodes changes the contact angle. The droplet, which is enclosed between the two electrode plates, features a sufficient volume to cover parts of two addressable electrodes at all times. If a voltage is applied to one of the control electrodes covered by the droplet, the contact angle at this part of the droplet is reduced, initiating a movement of the droplet along the paths given by the activated pads. Because the path of the droplets is determined by the pattern of electric fields, the EWOD-driven, droplet-based platform is easily programmable. This allows an operator to perform different assays, run by different programs on the same piece of hardware. Besides aqueous solutions, several other liquids, such as organic solvents, ionic liquids and aqueous surfactants solutions [41], as well as biological fluids, such as whole blood, serum, plasma, urine, salvia, sweat and tears [42], have been successfully transported on the EWOD droplet-based platform. An alternative to the electrowetting-based transport of droplets on a plane surface has been also proposed [43].This approach is based on SAWs, which are mechanical

302

Miniaturization of Analytical Systems

waves with amplitudes of typically only a few nanometres. The SAWs are generated by a piezoelectric transducer chip (e.g. quartz) fabricated by placing interdigital electrodes (interdigital transducer, IDT) on top of a piezoelectric layer. Liquid droplets situated on the hydrophobic surface of the chip can be moved by the SAWs if the acoustic pressure exerted on them is high enough [44]. This approach is also sometimes referred to as ‘flat fluidics’ because no cover or slit is required, unlike in the EOWD approach. Today, this technology is advanced by Advalytix, Germany [45]. General advantages of droplet-based microfluidics include the small liquid volumes of the droplets, reducing reagent and sample consumption and paving the way for high-throughput screening applications. Additionally, the batch mode operation scheme within the nanolitre-to-microlitre-sized droplets represents a consistent further development of the classic assay protocols in, for example, well plates. The pressure-driven approach combines these advantages with highthroughput capabilities in a quasi-continuous operational scheme. The completely enclosed liquid droplets furthermore allow the incubation and storage of liquid assay results over a long period of time without evaporation. However, the microfluidic functionality is engraved by the channel design and cannot be adopted during an assay, for example. In contrast, the surface-based actuation schemes (EWOD and SAW) come up with a high flexibility since liquid processing paths can be freely programmed. In addition, the simple setup without any moving parts can be fabricated very cost-efficiently using standard lithographic processes. Comparing the EWOD and SAW principles for droplet actuation on the planar surface, the electrical change of the contact angle depends on the liquid properties and could cause electrolysis, while the SAW principle allows easier adaption to the liquid properties. Evaporation of liquid and the long-term stability of the hydrophobic and hydrophilic surface coatings are the major drawbacks of the surfacebased techniques.

8.5 Microfluidic Devices for Analytical Lab-on-a-Chip Applications Analytical microsystems can be understood as the ‘main product’ – still in its infancy – of successful systems and applications in the field of mTAS and lab-on-achip. They have evolved both innovative analytical chemistry methods and microfluidics, and the two have been cleverly combined. The development of practical analytical microsystems [4] – especially those for bioanalysis – continues rapidly, although, given its early focus, this area has taken longer than expected to reach widespread routine use. Part of the problem is that there is limited technology in two parts of the analytical process: sample preparation (including analytical separation of the mixtures into single analyte bands) and analyte detection. Biological samples – particularly clinical samples (such as blood

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

303

or faeces) and those obtained by environmental and food sampling – are often dilute or complicated. Before these samples can be analysed by microfluidic devices, they must be converted to a form that is compatible with the intended analysis, and then introduced into the analytical device. The procedures required to complete these tasks are surprisingly sample-dependent and not necessarily ‘micro’ in scale. After a sample has been prepared, introduced into the analytical device and processed, it must then be detected. Having the microfluidic chip as just a small part of a system in which sample introduction and detection are much more complicated than the chip’s operation may be appropriate in some circumstances, but does detract from the potential advantages of microfluidic devices. Other standard problems, such as pumping, valving and on-chip reagent storage, also require better solutions than those available so far. In this way, not all analytical steps and/or microsystems are equally incorporated into the analytical scene of the labs, and the incidence level of miniaturization is not homogeneous. In fact, the detection and separation steps are the most investigated, while sample preparation is notably the least developed [6]. Although separation techniques and detection principles have been covered in detail in Chapters 4–6, we are next going to study the state of the art, but from a lab-ona-chip point of view. This involves integrating all the analytical steps and dealing with real samples. The philosophy of the above is depicted in Figure 8.5. Indeed, this figure shows the monotlithic approach as well as the intensity of integration of the main analytical steps in microfluidic platforms, represented by the analytical microsystem as final product.

Figure 8.5

Approach to the integrated concept of analytical microsystems

304

Miniaturization of Analytical Systems

From the beginning, as stated in Chapter 6, detection has been one of the main challenges for analytical microsystems, since very sensitive techniques are needed as a consequence of the ultrasmall sample volumes used in micron-sized environments. In principle, a wide variety of detection alternatives can be used in microfluidic systems (see Chapter 6), but laser-induced fluorescence (LIF) and electrochemical detection (ED), along with mass spectrometry (MS), are the routes most commonly used. LIF was the original detection technique and is the most common in CE microchips because of its inherent sensitivity [47]. However, its high cost and the large size of its instrumental setup have sometimes been incompatible with the concept of mTAS. In addition, tedious derivatization schemes are required to use LIF with nonfluorescent compounds. The principal alternative to LIF detection is, without a doubt, ED. ED is very important because of its inherent miniaturization without loss of performance and its high compatibility with microfabrication techniques. Likewise, it possesses high sensitivity, its responses are not dependent on the optical path length or sample turbidity, and it has low power supply requirements. Proof of the predominant role of ED can easily be found in the literature [48–50]. Apart from LIF and ED, there are other detection approaches for analytical microsystems, but these have been developed to a lesser extent [51]. With respect to separation techniques (see Chapters 4 and 5), micro-CE was one of the earliest examples of mTAS, and it constitutes one of the most representative examples of an analytical microsystem [52]. Using CE microchips, analysis times can be reduced to seconds and extremely high separation efficiencies can be achieved. The easy microfabrication of a network of channels using materials of well-known chemistry, which themselves have good EOFs, and the possibility of using the electrokinetic phenomena to move fluids are among the most important factors in understanding the relevance of CE microchips to miniaturization. Since electrokinetics is easy to apply (only a pair of electrodes are needed), EOF has been successfully implemented using different types of materials to manufacture the channels, with glass the most commonly used. Microfabrication on polymers is faster and cheaper than on glass, so these materials have great potential for mass production. However, glass chips present the best EOF, and chemical modification on the surface of glass microchannels is better understood and easier than that on polymers. Many of the benefits already discussed for CE microchips could equally be applied to downsized chromatographic techniques; however, relatively few examples of chip-based chromatographic instruments are covered in the literature compared with chip-based CE devices. This fact is not surprising since CE is almost perfectly suited to miniaturization. Also, the miniaturization of chromatographic systems involves a number of technical challenges, such as the microfabrication of valves and pumps, which are generally not encountered in CE [6]. In other words, and in a general sense, electrokinetically-driven systems are dominant in the pressure system.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

305

Sample preparation steps involving emerging technologies for sample preparation/pretreatment [53] and the macroworld-to-microworld interfaces [54,55] have been addressed. Because of its inherent complexity, the sample preparation step is less developed than those of separation and detection. However, a very exciting future in real analysis is conceivable because of (i) the inherent possibilities of microfabrication technology to create sophisticated designs (complex layouts according to sample requirements) and microstructures (filters) in connection with sample preparation requirements; (ii) the inherent possibilities offered by microfluidics – the presence of laminar flow and diffusion in action (see Section 8.2); and (iii) the ease of using electrokinetic phenomena to move fluidics into the network channels with accuracy (focusing on flow techniques). In addition, other important general challenges still await analytical solutions, such as automation, sampling, sample introduction and sample preparation (except derivatization schemes). Moreover, a breakthrough in the so-called real world interface is highly anticipated [6]. Along these lines, total integration and, especially, world-to-chip interfacing are considered the major challenges, particularly in high-throughput applications requiring frequent sample changes, such as continuous online process monitoring. The microfluidic literature is vast and rapidly expanding, with emerging applications to real samples. The focus of this section is on real (non-ideal) samples, in order to check how much progress is actually being made toward useful lab-on-achip devices. Extracting out examples of real sample analysis from the body of papers in this area has been a demanding task. The development of applications for a new technology is always a confirmation that the technology is maturing. The ability to efficiently process raw samples (from the laboratory, the body or the field) and subsequently perform the required analytical operations ‘on-chip’ will be key in defining the eventual success and application of microfluidic systems. Although the volume of research in this area is impressive, real sample analysis is still in its infancy and constitutes one of the major challenges of microfluidics. The following subsections provide in-depth coverage of the state of the art of microfluidic devices, focusing on the analysis of real samples as an indicator of the grade of maturity of these microsystems. In this context, although every sample is a ‘real sample’, we mean the analysis performed directly on the matrix containing the analyte (or those spiked). Section 8.5.1 obeys the classification of analytical applications. 8.5.1

DNA Analysis Integrated on Microfluidic Devices

The concept of mTAS applied to nucleic acid analysis involves different aspects such as nucleic acid extraction (e.g. tissue, blood, cells, etc.) amplification (polymerase chain reaction, PCR) and product analysis (real-time quantitative PCR, gel electrophoresis, capillary gel electrophoresis (CGE), DNA array) [56]. Many research

306

Miniaturization of Analytical Systems

groups have made a great effort to construct miniaturized analytical components devoted to some of these aspects. However, the assembly of the whole process on microfluidic platforms (glass, silicon or polymer microchip) is still in its infancy. Presently, there is no doubt that PCR is the most developed aspect of nucleic acid analysis miniaturization. Nevertheless, some pre-PCR modules such as sample purification and preconcentration, and post-PCR modules such as CGE and DNA microarray have been coupled to PCR on single microdevices. This section, offering a brief review of some of the most relevant studies reported in the literature, will be structured around three different topics: pre-PCR, PCR and post-PCR integration approaches. Coupling of some of these steps, even the whole gene process of analysis, on to microfluidic devices is presented. With respect to integrated pre-PCR, some of the treatments prior to analysis of nucleic acids include cell identification, isolation and sorting, nucleic acid extraction and analyte purification and preconcentration, since biological samples contain different types of cells and PCR requires relatively pure nucleic acid samples that are free of reaction-inhibiting contaminants. Isolation of different-sized cells into microfluidic devices can be accomplished by filters of appropriate dimensions [57,58]. For cells of similar size, hydrodynamic [59] or electrodynamic methods are available [60,62]. On the other hand, and in order to know the quantitative information on molecules per cell, prior to PCR, cell counting is crucial. Unfortunately, at this time coupling of cell counting and PCR on microdevices is still quite far off. This may be circumvented when genetic analysis is focused at the single-cell level, where mTAS can show its great suitability due to its inherent dimensions [63–65]. After sample collection, nucleic acid extraction to obtain target DNA or RNA can be accomplished through solid phase extraction (SPE) or isotachophoresis (ITP). With SPE, membranes or beads with high affinity to nucleic acids and very low affinity to proteins are embedded inside the microdevices [66,67]. ITP, a wellestablished technology for on-chip sample pretreatment, could be a potential tool for on-chip nucleic acid purification and concentration, given adequate electrolytes on chips where electric fields are employed [68,69]. Different strategies of purification and concentration on microchips and the whole integration of pre-PCR components with PCR on microfluidic devices are summarized in Table 8.4 [70–75]. As we already know, PCR is an enzyme-mediated process in which, in response to temperature cycling, a single DNA molecule can be rapidly amplified into many billions of molecules. This reaction was described in 1985 by Mullis and coworkers [76] and has revolutionized the field of molecular biology, being an indispensable part of genetic analysis. The conventional format of PCR involves the temperature cycling of the PCR components in thin-walled, plastic tubes with volumes typically in the 2–50 ml range. In microchip, the volume of PCR mixture can be reduced by one to two orders of magnitude and the speed with which

Substrate material

Glass

Silicon/glass

PC

PDMS/silicon

Silicon/glass

PCR type

Stationary chamber

Stationary chamber

Stationary chamber

Stationary chamber

Stationary chamber

E. coli

Listeria cells

Whole blood

Whole blood

Whole blood

DNA sample

Table 8.4 Pre-PCR integrated with PCR on a microchip

Microchip filter for white blood cell isolation White blood cell isolation on the filter section of the microchip Immunomagnetic bead-based cell capture DNA bound to silica-coated pillars in the presence of the chaotropic salt guanidinium isothiocyanate Cell lysis by laser-irradiated magnetic bead system, and magnetic beads for removing denatured proteins

Pretreatment technique

4

50

20

12

50

Volume (ml)

[75]

[73,74]

[72]

[71]

[70]

Ref.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 307

308

Miniaturization of Analytical Systems

thermocycling can occur is highly increased. Relevant studies of microchips based on integrated PCR have been extensively reviewed in the literature [77–81], where different performance and designs can be compared. In most cases, the significant increase in surface-to-volume ratio, together with surface properties of current materials in these microdevices, make a surface passivation method necessary in order to get reproducible good performance and stability in miniaturized PCR. Diverse treatments to overcome this drawback are mentioned in the literature, such as dynamic coating during PCR by adding passivation agents such as bovine serum albumin (BSA), poly(ethylene glycol) (PEG) or polyvinylpyrrolidone (PVP) into PCR solutions. These compounds are used to reduce the undesired adsorption of enzyme and DNA. On the other hand, chemical modification such as silanization before PCR is a better choice if the device is designed to be reusable [56]. At a higher grade of integration, real-time PCR (or quantitative PCR) on microfluidic devices involves the amplification of specific gene sequences coupled to quantification of the original target DNA. Detection is usually achieved by monitoring the increase in the fluorescence signal associated with the increase of double-stranded DNA (dsDNA) production after each round of thermocycling through the intercalation of a fluorescence probe. This technique presents the advantage of acquiring quantitative information such as the determination of unknown amounts of sample (a number of copies of the starting template have clinical relevance, e.g. viral titre) or PCR efficiency. Adaptation of real-time PCR to miniaturized platforms is rapidly evolving and the reported literature is presented in Table 8.5 [74,75,82–88]. Finally, the analysis of PCR products is the most developed area of nucleic acid analysis integration on microfluidic devices. This fact can be attributed to the complexity associated to sample handling during pre-PCR treatments. Various integrated post-PCR methods such as CGE, DNA microarray and immunoassay using mainly fluorescence or electrochemical detection have been adapted to the microchip format. Due to the low volumes used in miniaturized nucleic acid analysis, high-sensitivity detection systems are required. Sensitivity related to fluorescence detection makes the LIF detection system the most commonly adapted instrument to post-PCR analysis. Unfortunately, most of the reported works are quite far from miniaturized devices, since instrumental complexity involves the use of benchtop instruments. ED is a powerful alternative thanks to its easy miniaturization, low cost and portability. Nevertheless, the problems of electrode contamination and relatively low sensitivity need further work. CGE is the most commonly used method for post-PCR analysis and has been extensively studied in the microchip format, as we saw in Chapter 5. Integration of PCR amplification and CGE into a single microchip allows rapid analysis, reduces reagent consumption and minimizes sample handling, avoiding contamination. On the other hand, devices used for CGE can be adapted straightforwardly to the whole process. Sample reservoirs can be used as amplification chambers and, after PCR,

Substrate material

PDMS/Silicon

Silicon/glass Silicon

Silicon Silicon/glass PDMS PDMS PDMS

SU-8

Stationary chamber

Stationary chamber Stationary chamber

Stationary chamber Stationary chamber Stationary well Stationary channel Continuous flow

Stationary chamber

Thin-film heater Peltier element Thin-film heater and fan Joule heating Aluminium block with cartridge heater and temperature sensor Integrated Pt heater array

Thermoelectric heater/ cooler GeneSpector Micro PCR Thin-film heater

Thermal source

Real-time PCR integrated on a microchip

PCR type

Table 8.5

SYTOX Orange and TO-PRO-3

TaqMan probe Ethidium bromide and TaqMan probe TaqMan probe Molecular beacon probe TaqMan probe TaqMan probe TaqMan probe

SYBR green

Lab-on-a-chip strategy

25

50 50 7 0,42

4 50

50

Volume (ml)

[88]

[83] [84] [85] [86] [87]

[75] [82]

[74]

Ref.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 309

310

Miniaturization of Analytical Systems

amplified products can be directly injected into the separation microchannel for detection. Some examples of PCR–CGE coupling are listed in Table 8.6. DNA microarrays have been widely used in genome applications. Hybridization of gene-amplified fragments to oligonucleotide probes immobilized to a solid support is carried out. Despite differences in fragment lengths provided by CGE, DNA arrays offer information related to the sequences of amplicons, with a higher throughput. In the same way, immunoassay after PCR amplification has the potential to recognize specific sequences of nucleic acids in a fast and integrated mode. Different approaches reported in the literature are presented in Table 8.6 [89–108]. The use of mTAS as an integration system for nucleic acid analysis is an exciting challenge. However, in order to achieve really full quantitative analysis, some drawbacks need to be resolved. Although PCR amplification and product analysis are perfectly developed and coupled on microchip format, sample handling and pretreatment must be improved. 8.5.2

Real Clinical Sample Analysis on Microfluidic Devices

Nevertheless, as was already mentioned, not all the processes of an analytical procedure have been equally incorporated into a miniaturized device to date. In small-molecule analysis, many reports have been published on the detection step in microfluidic systems, whereas the sample preparation step tends to be accomplished off-chip [109,110]. Often, the analyte of interest is accommodated within an extremely complex matrix which requires different sample pretreatment techniques, making it difficult to incorporate into a microchip. This is the main obstacle to processing real samples and achieving a true mTAS or lab-on-a-chip device. In other words, working with real samples usually requires some degree of purification (isolation/cleanup) and preconcentration prior to the measurement step. Sensitivity (only a few nanolitres reach the detector) and selectivity can pose serious problems in the analysis of these raw samples. Thus, dealing with ‘real’ samples has been seen as an important issue for lab-on-a-chip, since sample pretreatments integrated on microfluidic devices are highly needed [109,110]. Clinical and biochemical analysis have been the most extensive fields of application for lab-on-a-chip systems and they continue to grow today [111–114]. The measurement of relevant analytes in physiological fluids for the diagnosis and prevention of diseases is a very exciting application, especially in point-of-care testing. The advantages (cited above) of downsizing from TAS to miniaturized devices are ideal for near-patient clinical analysis. Smaller patient sample volume, which is particularly important in geriatric and paediatric patients; reduced reagent consumption; shortened analysis time; integrated multiple processes in a single device (portability); and enhanced reliability and sensitivity through automation are just some of the achievable benefits which make this technology interesting. The

PC

Silicon/Glass PC Silicon/Glass

Silicon/Glass

Stationary chamber (isothermal) Stationary chamber

Stationary chamber Stationary chamber Stationary chamber

Stationary chamber

Multiple stationary chamber

Stationary chamber Stationary chamber Stationary chamber

Stationary chamber Stationary chamber Stationary chamber

PC Glass Glass Glass Glass Glass Glass Glass

Stationary chamber Stationary chamber Stationary chamber Multiple stationary chamber Stationary chamber Stationary chamber Stationary chamber Stationary chamber

Thermal source

Pt temperature sensors and heaters

Resistance sensor and heater Thermoelectric module Pt temperature sensors and heaters

Peltier element

Resistive heater Polysilicon heater Commercial thermocycler Commercial thermocycler Commercial thermocycler Thin film heater Resistive heater Integrated Ti/Pt heating elements with RTD PDMS/Glass Polieter element Silicon/Glass Aluminium thin film resistors Poly(cyclic olefin) Graphite ink-based heating resistor, m fan and RTD Glass Polieter thermocelectric module PDMS/Glass Thermoelectric module Glass chip for PCR, Ti/Pt thin-layer heater PMMA chip for CGE Glass Ti/Pt resistance temperature heater and sensors PMMA Block incubator

Substrate material

PCR type

Table 8.6 Post-PCR analysis integrated on a microchip

DNA hybridization assay DNA microarray Immunoassay Electrochemical detection Electrochemical detection

Multiple microchip CGE Microchip CGE

Microchip CGE Microchip CGE Microchip CGE

Microchip CGE Microchip CGE Microchip CGE

DNA microarray Microchip CGE Microchip CGE Microchip CGE Microchip CGE Microchip CGE Microchip CGE Microchip CGE

lab-on-a-chip strategy

8

3 8 8

38

10

0.38

3.5–4 2 10

30–50 3 or 10 0.029–0.084

20 20 10–25 12 12.5 0.28 0.28 0.2

Volume (ml)

[108]

[105] [106] [107]

[104]

[103]

[102]

[99] [100] [101]

[96] [97] [98]

[72] [89] [90] [91] [92] [93] [94] [95]

Ref.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 311

312

Miniaturization of Analytical Systems

reason the clinical field has been the most extensively developed is because the use of enzymes, antibodies and other biological reagents saves extensive sample pretreatment. The sensitivity associated with the use of enzymes allows samples to be diluted in order to eliminate interference from other constituents of the matrix and problems for proteins adsorbing to the capillary walls, without lowering detection limits. Conversely, the use of antibodies, enzymes or other affinity receptors is responsible for the selectivity required for these assays. A wide variety of bioanalytical and clinically relevant analytes, such as proteins (see Chapter 5), peptides, amino acids, neurotransmitters, carbohydrates and vitamins, have been determined in the microchip format. However, references describing analysis in complex matrices are few, most of them being analysed at the laboratory where the samples are manufactured. This section focuses on those analytes determined in true real or spiked samples. Selected examples will be described for different approaches, such as the use of CE, enzymes and antibodies to provide selectivity to the assay. The most remarkable features and the type of pretreatment accomplished will be outlined. Table 8.7 lists the most relevant information about the microfluidic approaches used in the analysis of clinical samples. The majority of the existing miniaturized systems require highly predictable and homogenous samples. In the case of heterogeneous samples, such as human whole blood, cell-free samples such as serum or plasma are preferable over those with problems associated with flow-path blockage and cell lysis. In such cases, wholeblood filtration is accomplished on-chip for the subsequent analysis. Thorslund et al. [115] have developed a microsystem for on-chip sample preparation by which testosterone is measured in diluted whole blood. This system uses a hydrophilic polypropylene membrane filter incorporated on to a PDMS device to remove blood cells with minimal adsorption of proteins. In clinical analysis, microchip CE is the most common technique used in miniaturized analytical devices, due to its special suitability for miniaturization [116–118]. Relevant analytes used to diagnose different diseases have been analysed by microchip CE. Lipoprotein analysis in serum samples was demonstrated by Verpoorte’s group [119,120]. In their first approach [119], low- and highdensity lipoproteins (LDLs and HDLs) were separated in a glass microchannel using a dynamic coating (methylglucamine) to prevent lipoprotein adsorption. NBD-ceramide was used as a label for LIF detection. In a later study [120], they demonstrated the analysis of different forms of clinically relevant LDLs without a permanent or dynamic coating of the glass microchannels. Adding SDS below the critical micelle concentration to the sample provided sufficient sample recovery to detect the lipoprotein at low concentrations. In both cases, ultracentrifugation of the serum and its subsequent dialysis were accomplished prior to microchip analysis. Neurotransmitters and their metabolites have been determined in biological fluids. Simultaneous amperometric and fluorescence detection is reported by

Testosterona

Lipoproteins (LDL and HDL)

Lipoproteins (LDL types)

Whole blood (spiked)

Serum

Serum

Neurotransmitters (dopamine and ascorbic acid)

Catecholamines and their cationic metabolites

Amino acids

Serum

Mouse brain and human urine

Urine

Cerebral spinal Dopamine, catechol, fluid (spiked) NBD-arginine and NBDphenylanaline

Analytes

Sample PDMS chip with membrane filter incorporated RIA detection off-chip Glass CE microchip LIF detection Glass CE microchip LIF detection Glass CE microchip LIF and EC detection

Microfluidic system

Glass CE microchip. Electrochemical detection Dilution and filtration Glass CE off-chip microchip Electrochemical detection Filtration and dilution Glass CE off-chip microchip LIF detection

No treatment

Ultracentrifugation and dialysis off-chip Ultracentrifugation and dialysis off-chip Centrifugation and dilution off-chip

Dilution off-chip

Sample preparation

Table 8.7 Lab-on-a-chip applications in clinical analysis

Separation time: 120 s LOD: 32.9 mM

LOD dopamine: 448 nM LOD cathecol: 1.52 mM LOD NBD-argin.: 16 mM LOD NBDphenyl.: 28 mM Separation time: 220 s 1.3 (%RSD) LOD: 0.032 mM Separation time: 150 s LOD: sub-mM

Separation time: 21 s

Separation time: 28 s

Sample preparation microsystem

Analytical characteristics

[122]

[121]

[120]

[119]

[115]

Ref.

(continued )

Simultaneous [123] determination. Carbon nanotube (CNT)modified electrodes Indirect fluorescence [124] detection approach. 19 amino acids detected

Cellulose-ssDNA modified SP electrode

Dual fluorescence and electrochemical detection

Uncoated glass microchip

Sample preparation on-chip. Removal of blood cells with minimal adsorption of proteins Methylglucamine as dynamic coating

Remarks

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 313

Analytes

Uric acid

Creatine, creatinine, p-aminohippuric acid and uric acid

Creatine, creatinine and uric acid

Lactate and glucose

Glucose

Sample

Urine

Urine

Urine (spiked)

Serum and blood

Plasma

Table 8.7 (Continued) Microfluidic system

PDMS CE microchip Electrochemical detection

Dilution and filtration Glass CE Off-chip microchip Post-column enzymatic reaction Electrochemical detection Dilution and filtration PDMS/glass CE off-chip microchip Eletrochemical detection

Dilution off-chip

Filtration and dilution PDMS/Glass CE off-chip microchip Electrochemical detection Dilution and filtration Glass CE off-chip microchip Electrochemical detection

Sample preparation

Separation time: 120 s LOD: 6 mM

Separation time: 340 s LOD creatine: 20 mM LOD uric acid: 40 mM LOD paminohippuric acid: 20 mM Separation time: 200 s LOD creatine: 250 mM LOD creatinine: 80 mM LOD uric acid: 270 mM LOD glucose: 30 mM LOD lactate: 45 mM

Separation time: 30 s LOD: 1 mM

Analytical characteristics

---

100-fold dilution avoid protein adsorption

[129]

[128]

[127]

[126]

Coupling of enzymatic bioassay and electrophoretic separation

Direct PAD detection

[125]

Ref.

50–75-fold dilution avoids potential interferences

Remarks

314 Miniaturization of Analytical Systems

Live cells

Tissue fluids

Saliva (spiked)

Serum (spiked)

Blood (spiked)

Serum (spiked)

Glass microchip Colorimetric enzymatic method High degree of integration Electrowetting for microdroplet actuation Analysis time: 30 s High degree of LOD: 0.5 mM integration

---

[140]

[138, 139]

[137]

[136]

[134]

[133]

[131]

[130]

(continued )

Integrated flow cell on silicon chip Chemiluminiscent enzymatic FIA system Ca2þ No treatment Glass/PDMS Analysis time: 30 s No significant microchip LOD: 26.8 mM interferences from Reflectance other ions presented measurements Haemoglobin Dilution Silicon microchip Concentration Optical path extension Photometric range: 12–17 g/dl with 45 mirrors at each end of the detection channel Tetanus antibody and No treatment Glass microchip. Separation time: Noncompetitive and tetanus toxin PAGE gel CE 3 min competitive immunoassay. LOD: 680 pM immunoassay format LIF detection Off-chip incubation MMP-8 Centrifuged and Glass microchip. Analysis time: Pretreatment on-chip diluted PAGE gel CE < 10 min (filtering, enrichment immunoassay. LOD: 130 ng/ml and mixing) LIF detection Neuropeptides Ultramicrodialyzer Glass microchip Separation time: 12 neuropetides are Immunoextraction Immunoaffinity CE 160 s determined on-chip LIF detection LOD: 0.6 pg/ml simultaneously Insulin Conditioning and Glass CE Competitive stimulation on-chip Microchip Assay time: 30 s immunoassay format Immunoassay LOD: 3 nM LIF detection

Dissolved

Serum (spiked)

Lactate

Dilution

Serum, plasma, Glucose urine and saliva (spiked)

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 315

Serum (spiked)

B-type natriuretic peptide

Plasma (spiked) Cardiac markers (CRP, Mb, cTnI and S100)

Serum

Whole blood

Serum (spiked)

Serum (spiked)

Histamine

Whole blood

Dilution

No treatment

Sample preparation

Microfluidic system

PDMS microchip Immunosorbent assay LIF detection Glass/PDMS microchip Immunosorbent assay SPR detection

PMMA microchip. Electrochemical detection Carcinoembryonic No treatment Glass microchip antigen Immunosorbent assay Termal lens detection Immunosupressive Dilution Quartz microchip acidic protein Immunosorbent assay Chemiluminiscence detection Botulinum neurotoxin Filtration on-chip PDMS platform Immunosorbent assay Electrochemical detection Human IgG Filtration and dilution Immunosorben off-chip assay LIF detection

Analytes

Sample

Table 8.7 (Continued)

Analysis time: 30 min LOD: 5 pg/ml

Differentiation of normal from elevated IgG concentrations Analysis time: 10 min LOD: 30 ng/ml

---

Analysis time: 2 min LOD: 0.1 mM

Analysis time: 2 min LOD: 200 ng/ml Analysis time: 35 min LOD: 0.03 ng/ml

Analytical characteristics

Ref.

Sandwich format Micromosaic Simultaneous determination Competitive format Off-chip incubation

Competitive format Biopassivated microchannells

Sandwich format on agarose beads High degree of integration

Competitive format on glass beads High degree of integration

[147]

[146]

[145]

[144]

[143]

Multichanneled matrix [141] column coated with cation-exchange resin Sandwich format on glass [142] beads Antibody-colloidal gold conjugate

Remarks

316 Miniaturization of Analytical Systems

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

317

Lapos et al. for the analysis of cerebrospinal fluid (CSF) samples [121]. The use of an orthogonal detection scheme allows LIF detection of NBD-arginine and NBD-phenylalanine, and ED of dopamine and catechol, increasing the possibility for the detection of analytes with different physical characteristics. The simultaneous use of two detectors reduces the run-to-run migration time irreproducibility for certain samples. CSF samples from patients with multiple sclerosis were centrifuged, diluted and spiked with electroactive species off-chip. Johirul et al. [122] have developed a microfluidic chip based on CE coupled with a cellulose-single-stranded DNA (ssDNA) modified screen-printed electrode for the simultaneous analysis of dopamine (DA), norepinephrine (NE), 3,4-dihydroxyL-phenylalanine (L-DOPA), ascorbic acid (AA) and 3,4-dihydroxyphenylacetic acid (DOPAC). The modification of the electrode with cellulose-ssDNA prevents surface passivation and amplifies electrocatalytic activity through the interaction of the analyte and ssDNA. Cellulose regulates the solubility of ssDNA in aqueous media and ssDNA acts as an efficient promoter of a direct electron transfer reaction and/or the concentration of analytes on the electrode surface by means of the electrostatic interaction. The applicability of the system to DA analysis in the presence of AA in serum samples without treatment has been demonstrated. Schwarz’s group [123] has developed a sensitive and selective method for the simultaneous determination of catecholamines and their cationic metabolites, without mutual interference, in mouse brain homogenates and human urine. They used a novel combination of additives (sodium dodecyl sulphate micelles, dendrimers forming a second pseudo-stationary phase and borate complexation) to achieve complete separation on the chip and a carbon nanotube (CNT)-modified electrode as a detector to enhance sensitivity. Sample pretreatment, carried out offchip, was necessary in both cases. For urine, sample dilution in 0.2 M HClO4 containing 0.1% of Na2EDTAwas required to prevent oxidation of the analytes and filtration. The brain tissue was homogenated, centrifuged and diluted in running buffer prior to injection. Landers et al. [124] have demonstrated the possibility of high-throughput screening for several amino acids in urine samples using CE on a microchip format. Indirect fluorescence detection has been adapted to the electrophoretic microchip in order to provide a fast analysis of amino acids. Although the detection limits for indirect detection are lower than those achievable using fluorescencelabelled analytes, the lack of sample preparation and simplified electrophoretic profiles make this technique attractive for many types of analysis. The analysis of urine samples requires filtration and dilution of the sample to avoid the high ionic values derived from the presence of NaCl and other inorganic ions. Despite the lower signal-to-noise ratio obtained when compared to conventional CE (significantly longer capillary), the method allows the quantitative determination of various amino acids in abnormal urine specimens.

318

Miniaturization of Analytical Systems

Microchip CE-ED to measure uric acid in physiological urine samples was reported by Fanguy et al. [125]. The sample preparation and the analysis of uric acid, with a clinically accepted enzymatic reaction, can be tedious. However, the proposed microchip system using ED is able to evaluate single-filtered diluted urine samples in the expected range of the concentration of the analyte. Due to the low detection limit of the method, uric acid determination is accomplished in 50–75fold dilution of the urine, without potential interference from other organic compounds with similar potential oxidation in the sample. The groups of Wang and Henry reported the simultaneous analysis of different renal markers in urine samples using microchip CE [126,127]. In the first paper [126], creatine, creatinine, p-aminohippuric acid and uric acid were determined by the coupling of an on-chip enzymatic assay, electrophoretic separation and ED. Hydrogen peroxide from the enzymatic reactions of creatinine and creatine on-chip was separated from p-aminohippuric and uric acids, and the three species were amperometrically detected at the end-column detector. Urine samples were filtered and diluted prior to analysis. Henry et al. [127] reported the separation of these renal markers (creatine, creatinine and uric acid) using 30 mM borate buffer (pH ¼ 9.4) plus 1 mM SDS and their direct detection using pulse amperometric detection. The three analytes were determined in diluted and spiked urine samples, and the creatinine results were validated using a commercially available assay kit based on the Jaffe reaction. Heightened levels of carbohydrates in biological fluids, mainly glucose, are especially relevant in diagnosing metabolic disorders. With this in mind, different microchip systems have been developed to determine carbohydrates in urine or blood samples. Wang and his group [128] reported an electrophoretic microchip that determines lactate and glucose in blood samples. This procedure combines the selectivity and amplification features of enzymatic assays with CE separation efficiency. Specifically, this microchip integrates electrophoretic separation of lactate and glucose in the separation channel, their post-column reaction with the corresponding oxidases and amperometric detection of the hydrogen peroxide liberated in the two separated product zones, as illustrated in Figure 8.6(a). Electrophoretic separation avoids potential electroactive interference, enhancing the advantages of enzyme flow-injection assays. Analysed samples (serum and blood samples) must be filtered, centrifuged and diluted in the running buffer offchip. Du et al. reported the electrophoretic separation and direct ED of glucose on a PDMS/glass microchip [129]. The device includes copper microelectrodes manufactured on to the electrode plate by selective electrodeless deposition. Plasma samples from healthy and diabetic patients were already diluted and filtrated, and glucose was determined in 120 seconds with a detection limit of 6 mM. Lactate and glucose were also determined using colourimetric enzyme assays from a different mTAS approach. Srinivasan et al. [130] reported an impressive ‘lab-

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

319

Figure 8.6 Labs-on-a-chip for clinical analysis. (a) Schematic of the CE-EC microchip, with the post-column enzymatic reaction, for the simultaneous measurement of glucose (G) and lactate (L). (Reprinted from [128] with permission from Elsevier); (b) Integrated digital microfluidic lab-on-achip for clinical diagnostics. This lab-on-a-chip, based on electrowetting for microdroplet actuation, integrates sample injection, on-chip reservoirs, droplet-formation structures, fluidic pathways, mixing areas and optical detection sites on the same substrate. Glucose is determined in spiked serum, plasma, urine and saliva as a proof of concept by the colourimetric enzymekinetic method (Trinder reaction). (Reprinted from [130] reproduced by permission of the Royal Society of Chemistry)

on-a-chip architecture’ for the automated analysis of glucose in physiological fluids. The microfluidic system they used is shown in Figure 8.6(b). This lab-on-achip, based on electrowetting for microdroplet actuation, integrates sample injection elements, reservoirs, droplet formation structures, fluidic pathways, mixing areas and optical detection sites all on the same substrate. Detection based on the colourimetric enzyme-kinetic method (Trinder reaction) allows glucose to be determined in spiked serum, plasma, urine and saliva samples with good concurrence to the reference method. Additionally, I. Karube’s group [131] fabricated a compactly integrated flow cell with a chemiluminescent FIA system to determine lactate concentration in serum. The cell was made by micromachining techniques and mainly consists of a reactor, a mixing cell and a spiral groove. Lactate was determined using lactate oxidase and was catalysed by pyruvate and hydrogen

320

Miniaturization of Analytical Systems

peroxide. Subsequently, the hydrogen peroxide reacted with the luminol-ferricyanide reagent and the resulting chemiluminescent product was detected. Lactate concentration was determined in undiluted spiked serum samples using this system. Moon et al. [132] developed an integrated digital microfluidic chip for multiplexed proteomic sample preparation for MALDI-MS analysis. The system integrates the handling of solutions and reagents with EWOD actuation, allowing the parallel processing of multiple sample droplets for high-throughput MALDI-MS. Processing four angiotensin peptide/urea samples demonstrated the suitability of the device for proteomic applications. Spectrophotometric detection is not usually employed in microfluidic devices, mainly due to its low sensitivity; however, Caglar and coworkers [133] have developed a glass-PDMS microchip to determine Ca2þ ions in serum samples by taking reflectance measurements of arsenazo III immobilized on the surface of polymer beads. The beads were used at the detection point of the microfluidic sensor device with a fibre-optic assembly for reflectance measurements. Despite potential interference from other ions, the microchip sensor allowed Ca2þ to be determined in untreated serum samples within the physiological range. In another approach, to overcome the usually low sensitivity associated with photometric microchips, Noda et al. [134] developed an optical path extension absorption-photometry microchip for highly sensitive haemoglobin measurements. In this device, the reflection in two 45 mirrors at each end of the fluidic channel allows the propagation of light from the source to the detector along a longer optical path (5 mm). Blood samples that were previously diluted and to which the haemolysis chemical (cyanogen solution) were added were then photometrically measured, allowing the determination of haemoglobin within the normal human concentration range (12–17 g/dl). As we know, immunoanalytical systems have become one of the most widely used methods for detection and determination of a wide variety of analyte molecules in clinical diagnosis. Miniaturization of immunoassays on to microchip platforms combines the power of microfluidic devices with the high sensitivity and specificity of antigen–antibody interactions. The inherent reduction in diffusion distances for the species and the larger surface area-to-volume ratio can increase the chances of the antigen and antibody binding within the microfluidic channels. However, in a miniaturized immunoassay, particular points that are more important than in conventional systems can be considered. These include the effect of the capillary surface’s chemical nature (nonspecific binding), the methods of liquid transportation because of high flow resistance, special consideration for mixing procedures due to laminar flow and specific detection principles because of the reduction in volume reaching the detector [112]. Taking into account the exceptional properties of miniaturized CE together with the selectivity power of immunoassays, the coupling of both techniques (mCEI) constitutes one of the most widely used approaches [135]. Herr et al. have developed microfluidic immunoassays based on gel electrophoretic separation and

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

321

the quantization of bound and unbound species. In an earlier paper [136], direct immunoassay (noncompetitive) and competitive formats were used in the detection of the tetanus antibody and the tetanus toxin, respectively, in spiked serum samples. On-chip photopatterned native PAGE gel was used for the electrophoretic separation of the fluorescently-labelled C-fragment of the tetanus toxin with the specific antibody from its complex after an off-chip incubation step. Recently [137], the same researchers improved their system by integrating sample pretreatment on-chip (filtering, enrichment, mixing) with an electrophoretic immunoassay to determine the collagen-cleaving enzyme matrix metalloproteinase-8 (MMP-8) in saliva as a diagnostic fluid. Relevant clinical concentrations (LOD 130 ng/ml) are measured in less than 10 minutes. Samples must be centrifuged and diluted to 20 ml before being loaded into the microfluidic system. Exploiting the capability of CE to analyse multiple analytes in a single sample, Phillips and Wellner [138,139] designed an immunoaffinity CE to measure concentrations of up to 12 inflammatory neuropeptides and cytokines in muscle and CSF. Using one reservoir of the microchip, Fab fragments from each of the specific antibodies were immobilized on their surfaces. After the incubation steps with sample and laser dye and the subsequent washing steps, the captured analytes were eluted and directed to the separation channel. In less than 160 seconds, 12 neuropeptides were resolved and determined in a 100 pl sample. Tissue fluid samples were first passed through an ultramicro dialyser. In the same paper [139], an electrokinetic immunoassay was developed for the same analytes. However, 280 seconds were needed to resolve the immune complex of the free-labelled antibody of one specific analyte, and higher LODs were obtained. Kennedy et al. demonstrated the continuous monitoring of insulin secretion from live cells (islets of Langerhans) using a CE competitive immunoassay [140]. Continuous sampling, online reaction and electrophoretic separation of the fluorescein isothiocyanate-labelled insulin (FITC-insulin) and FITC-insulin–Ab complex were accomplished in 30 seconds. A detection limit for insulin of 3 nM and continuous monitoring up to 30 minutes with no intervention were achieved. Matsunaga’s group [141] reported the use of a multichanneled matrix column coated with cation-exchange resin to separate immunocomplexes after a competitive binding event. Histamine is measured in 10 ml of whole blood directly extracted from patients by an electrochemical flow immunoassay. Based on different isoelectric points of immunocomplexes, histamine-antibody and histamine-derivative-antibody, only the first is retained in the column. The conjugation of the specific antibody with ferrocene allows ED of the immunocomplex in the eluent. Histamine concentration in the 200–2000 ng/ml range is measured in whole blood within 2 minutes. The determination of cancer markers is crucial for an early diagnosis of the disease and immunoassay is considered an indispensable technique for determining small amounts of a tumour marker in serum. Different papers have reported using

322

Miniaturization of Analytical Systems

microchip-based immunoassays for cancer diagnosis [142,143,145]. Sometimes the electrophoretic separation of bound and unbound species is not suitable in the microchip format. An interesting alternative would be the integration of the widelyused immunosorbent assay (ELISA), in which antigens or antibodies are fixed on a solid surface, on to a microchip known for its high sensitivity. Kitamori et al. [142] and Tsukagoshi et al. [143] have developed glass bead immunosorbent assays for the human carcinoembryonic antigen (CEA) and immunosuppressive acidic protein (IAP), respectively. In Kitamori’s work, a sandwich immunoassay format together with thermal lens detection of colloidal gold-labelled antibodies is used. Serum samples from patients are assayed in 35 minutes, with a determination limit of 0.03 ng/ml for CEA. Tsukagoshi uses chemiluminescence to detect unbound luminol-labelled analytes at the end of the microchannel, where a microperoxidase catalyses the hydrogen peroxide reaction. IAP spiked serum samples are determined in the relevant clinical range of concentrations within 2 minutes for each assay. Also based on a sandwich ELISA format, where the specific antibody is immobilized on functionalized agarose gel beads, Moorthy et al. [144] developed a microfluidic platform for detecting botulinum neurotoxin directly from whole blood. Sample preparations by filtration, mixing and incubation with reagents were carried out on the device. Clinically relevant amounts of the toxin can be detected by measuring the colour of an insoluble enzymatic product of the alkaline phosphatase reaction. Elevated levels of monoclonal human IgG in serum may be linked to cancer. Thus Thormann’s group [145] reported a heterogeneous competitive human serum IgG immunoassay. Immunoglobulin G and Cy5-IgG, used as a tracer, compete in order to bind to a specific antibody immobilized on a biopassivated microchannel. Using Cy3-mouse IgG as an internal standard allows samples from patients with normal IgG serum levels (8–16 mg/ml) to be differentiated from patients with elevated IgG concentrations (>16 mg/ml). A sophisticated approach is reported by Delamarche’s group [146], in which different cardiac markers are simultaneously detected. They combine microfluidic networks, in which the flow of liquids is driven by capillary forces (self-regulating), and micromosaic immunoassays based on patterning lines to capture antibodies exposed to solutions of analytes. Using secondary fluorescently-labelled antibodies, a micromosaic of fluorescent zones reveals the binding event in a single imaging step. In this case, four cardiac markers – C-reactive protein (CRP), myoglobin (Mb), cardiac troponin (cTnI), and S100 – are simultaneously determined in spiked plasma samples (1 ml volume) within 10 minutes. Recently, Niwa et al. [147] reported a microfluidic device combined with a portable surface plasmon resonance (SPR) sensor system to determine the cardiac marker B-type natriuretic peptide (BNP) in spiked serum samples. After competitive immunoreaction takes place, the thiol compound generated by the enzymatic

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

323

reaction with the trapped labelled antibody is accumulated on a thin gold film located in the microchannel. The extension of the chemisorption of thiol molecules can be monitored by the SPR angle shift and related to BNP concentration. A concentration range of 5 pg/ml to 100 ng/ml covers the usual levels of BNP in blood. 8.5.3

Real Environmental and Related Sample Analysis on Microfluidic Devices

Modern society demands the development of portable, robust and accurate analytical systems to monitor environmental compounds such as contaminants, explosives, chemical warfare agents and inorganic and organic ions. Portability allows analyses to be carried out outside the laboratory, preventing or minimizing the risk of contaminating the sample and leading to a faster response time at a lower cost. The ability to use portable equipment is an intrinsic characteristic of microfluidic systems, which is why they are an excellent response to this analytical demand. Sample introduction, the complexity of the sample matrix and the low concentration of the chemical species in the environment are possible challenges to the use of microfluidic devices for environmental analysis. There are excellent bibliographic reviews in the literature about microfluidic devices for environmental analysis [148,149], but no reviews are to be found on the analysis of real samples and their particular challenges. In this section, the most relevant papers in the literature that use microfluidic devices for environmental analysis have been classified and discussed according to the kinds of analytes studied: warfare agents and explosives, and pollutants and ions of environmental interest. Table 8.8 summarizes the most relevant information regarding the analysis of the main compounds of environmental significance in real samples using miniaturized devices. J. Wang and coworkers have developed a method for determining thiol-containing degradation products of V-type nerve agents in water samples [150]. The detection of three of these compounds was achieved in only 4 minutes, though only in a water sample spiked with its corresponding standards. The microsystem consists of a glass CE microchip with amperometric detection (screen-printed carbon ink electrode). A derivatization of the compounds with o-phthaldialdehyde was necessary, which was developed both on- and off-chip. When the process was performed on-chip, the complete reaction between derivatization agent and compounds was not achieved, resulting in decreased sensitivity. With the same microsystem [151], three organophosphate nerve agents were detected in 140 seconds from a river-water sample spiked with those analytes. In both cases [150,151], the LODs were in the micromolar range, but this seems to be insufficient for real applications. Better LODs were obtained by the same research group using contactless conductivity as its detection mode [152]. A detection limit within the

2-(dimethylamino) ethanethiol 2-(diethylamino) ethanethiol 2-mercaptoethanol Paraoxon methyl parathion Fenitrothion Methylphosphonic acid ethylphosphonic acid isopropylphosphonic acid Trinitrotoluene 2,4-dinitrotoluene 2,6.dinitrotoluene

2-nitrophenol 3-nitrophenol 4-nitrophenol 2,4-dinitrophenol 2-methyl-4,6 dinitrophenol

4-aminophenol 2-aminonaphtalene o-aminobenzoic acid

Phenol 2-chlorophenol 2,4-dichlorophenol 2,3-dichlorophenol

Tap and river water (spiked)

Groundwater (spiked)

River water (spiked)

River water (spiked)

Soil and groundwater

River water (spiked)

River water (spiked)

Analytes

Sample

Filtration off-chip

---

Filtration off-chip

Amperometry Carbon ink screenprinted electrode

Detection mode

Glass microchip CZE

Glass microchip CZE

Glass microchip MEKC

140

240

1.3–2

7–30

0.48–0.88

120

150

120

240

---

---

---

Comparison with LC procedure (method 8330) US EPA

---

---

---

Analysis Accuracy time (s) evaluation

0.005–0.006 130

0.9–3.8

6–11

LOD (mM)

Amperometry 1–2 Carbon ink screenprinted electrode modified with gold

Amperometry Boron-doped thin-film detector

Amperometry Glassy carbon electrode

Amperometry Gold electrode

Amperometry Carbon ink screenprinted electrode PMMA microchip Contactless CZE conductivity

Glass microchip MEKC

Glass microchip CZE

Microfluidic system

Extraction, centrifugation Glass microchip and filtration off-chip MEKC

Filtration off-chip

Filtration off-chip

Derivatization agent (o-phthaldialdehyde) in presence of Valina Off- and on-chip

Sample treatment

Table 8.8 Lab-on-a-chip applications in environmental analysis

[156]

[155]

[154]

[152]

[151]

[150]

Ref.

324 Miniaturization of Analytical Systems

Phenol 4,6-dinitro-o-cresol pentachlorophenol

DOC (dissolved organic carbon)

City water (spiked) Medicines

River water

Sulfite Nitrite

Cd(II), Pb(II), Co(II), Ni(II) Cd(II), Pb(II), Co(II), Ni(II)

Nitrite Fluoride Phosphate

River, pond and rain water

Plexiglas surface contaminated

River, tap, mineral water

Electroplating Pb(II) sludge reference material Water mimetic Cr(III) solution

Levoglucosan

Smoke particles

Derivatization agent N-(9-acridinyl) maleimide 2,3-diamoino naphthalene on-chip Chelating agent 2(5-bromo-2-pyridylazo)5-(N-prpyl-N-sulfopropylaminophenol C18 silica column off-chip off-line Preconcentration and elimination of interferences on-chip on-line

EDTA to remove interferences off-chip

Derivatization with isothiocyanate and filtration vvoff-chip Extraction, centrifugation and filtration off-chip

Aerosol particle collection and extraction off-chip ---

LIF

LIF

PMMA microchip Conductivity ITP-CZE column coupled

Glass microchip Colorimetric Nonaqueous CZE

PDMS-glass microchip FIA

PDMS-glass Chemo microchip luminescence FIA (mixer/reactor (luminol-H2O2) system)

PDMS microchip LIF Sensor

Glass microchip CZE

0.5–0.7

360

0.003–0.053 80

900

60

0.1

0.4–1

---

120

120

60

0.011

---

PDMS microchip Pulse amperometric 17 CZE detection Gold electrode PDMS microchip Pulse amperometric 0.9–2.2 CZE detection Gold electrode [159]

---

Recovery

Recovery

Recovery

[164]

[163]

[162]

[161]

Respect reference [160] material

---

Respect declared [158] content (label medicine)

Comparison with [157] GC/MS

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 325

326

Miniaturization of Analytical Systems

nanomolar range was achieved for the determination of three organophosphonate nerve agent degradation products from a river-water sample (spiked with standards). In this paper, a disposable PMMA microchip with integrated electrodes was used with excellent results (analysis time 130 seconds). Luong et al. were able to determine trinitrotoluene, 2,4-dinitrotoluene and 2,6dinitrotoluene in soil and groundwater samples by MEKC-EC on a glass microchip with a gold wire as an electrode [153]. The analysis time was 4 minutes and LODs were below the micromolar range. The accuracy of the method was validated by comparison with the LC procedure of the US Environmental Protection Agency (method 8330). For the soil sample, an extraction step off-chip was necessary. In all papers, the water samples were filtered off-chip. J. Wang and coworkers have also separated and detected three different groups of pollutants (nitrophenols, aromatic amines and chlorophenols) by CE on a glass microchip with amperometric detection using river- and groundwater samples. In this case, the detected compounds were added prior to the experiment. Five nitrophenol derivates were detected in 120 seconds with a glassy carbon electrode [154], three aromatic amines (4-aminophenol, 2-aminonaphthalene and o-aminobenzoic acid) in 150 seconds with a boron-doped diamond thin-film detector [155], and phenol and three chlorophenols in 120 seconds with a carbon SPE modified with gold [156]. In these papers, the LODs were within the micromolar range and accuracy was not evaluated. The samples were filtered before the analysis. C.D. Garcia et al. obtained good results in their analysis of environmental samples using a PDMS microchip with pulse amperometric detection (PAD) [157]. Levoglucosan (the largest single component of the water-extractable organics in smoke particles) was determined in smoke particles in only 1 minute [157]. The results obtained with this method are in accordance with those obtained with GC/ MS. The aerosol particles were collected and extracted off-chip and offline. The LOD achieved (17 mM) was adequate for the analysis of smoke samples but insufficient for atmospheric samples. In another paper [158], three of the most important pollutants (phenol, 4,6-dinitro-o-cresol and pentachlorophenol) were detected in a city water sample spiked with the corresponding standards. The LODs were within the micromolar range and accuracy was evaluated by determining phenol in two medicines with labelled content. Sample treatment was not necessary. An original method was developed for the evaluation of the environmental index DOC (dissolved organic carbon) [159]. The microsystem consists of a glass CE microchip with LIF as its detection mode. Derivatization with a labelling agent, fluorescein isothiocyanate, was required (overnight). A sample of river water was analysed, but the results were not compared with other methods. Anions and cations of environmental interest have also been explored. I.H. Chang et al. fabricated a sensor to monitor Pb2þ in environmental samples [160]. This

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

327

Figure 8.7 Labs-on-a-chip for environmental analysis. (a,b) Electrical bias configurations for fluidic control of lead and DNAzyme within an NAI/microfluidic device: (a) on state; (b) off state. (c–h) Temporal sequence of fluorescence images at the intersection of the crossed microchannels (Reprinted from [160] with permission of ACS, Copyright 2005)

microdevice was made on PDMS, contained a network of microfluidic channels that were fluidically coupled via a nanocapillary array interconnect (NAI) and used LIF as its detection mode. This sensor’s selectivity was achieved using a lead-specific DNAzyme constructed by the authors, who possessed a fluorophore fragment. In the presence of Pb2þ, the substrate DNA was cleaved, resulting in the release of fragments and a concomitant increase in fluorescence, as illustrated in Figure 8.7. Accuracy was evaluated with an electroplating sludge reference material, but no further samples were analysed. An excellent LOD was achieved (11 nM). Two different research groups have used CE microchips as FIA systems. In the first paper [161], Cr (III) was detected by chemiluminescence in a water mimetic solution in 1 minute. The catalytic property of Cr (III), the result of the reaction between luminol and H2O2, was taken advantage of to determine this cation. All parameters of the mixer/reactor microsystem were optimized and off-chip sample treatment with EDTAwas needed to remove interferences (other metal cations). In the second paper [162], sulfite and nitrite were detected by FIA-LIF in different water samples, but the use of two derivatization agents was needed (on-chip). LODs within the micromolar range and a huge analysis time (15 minutes) were obtained. The accuracy of both methods was evaluated by recovering the standards added in the studied samples. G. Deng and G.E. Collins separated and detected four metallic cations (Cd2þ, Pb2þ, Co2þ, Ni2þ) using a nonaqueous CE microchip

328

Miniaturization of Analytical Systems

with colourimetric detection on a contaminated Plexiglas surface prepared by the authors [162]. Sample treatment was carried out off-chip and offline and consisted of extraction with a chelating agent. The extract was then preconcentrated in a C18 silica microcolumn, and finally the sample was introduced on to the microchip, achieving LODs within the nanomolar range. Bodor et al. determined three anions (NO2
, F
, PO43
) in river, tap, and mineral water on a PMMA column-coupling microchip with conductivity detection [164]. The microchip consisted of two coupled columns: the first was used as an ITP column and the second as a capillary zone electrophoresis (CZE) column. In the ITP stage, the sample was preconcentrated and cleaned up, and in the CZE stage the anions were separated. Minimal sample treatment off-chip was required (filtration and degassing). LODs were within the micromolar range and the accuracy of the method was not evaluated. 8.5.4

Real Food Sample Analysis on Microfluidic Devices

The state of the art of food analysis on CE microchip platforms has recently been revised [165]. In food analysis, sensitivity and selectivity are the major drawbacks. In contrast to biochemical analysis, where very specific reagents (immunoassay/ enzymes) are frequently used and dramatically improve the selectivity in the analysis, in food analysis the use of bioreagents is obviously less developed and sample preparation is always needed, involving complex sample preparation protocols and analyte separation with advanced detection schemes. Microfluidics is now emerging in food analysis, since, as was mentioned before, sample preparation integration is less developed due to its inherent complexity. The microfluidics used are mainly CE microchips applied to both liquid and solid samples. The first approaches focused on exploring fast separation and suitable detection routes for prominent analytes with food significance. In all cases, the sample preparation step is carried out off-chip and microchips are merely used as fast analytical separation systems. In consequence, the first ‘real samples’ analysed are still ‘easy’, for example nutraceuticals and dietary supplements. Table 8.9 lists the relevant information about food analysis on, primarily, CE microchips and other microfluidic systems. Next, we are going to discuss the relevant keys to understanding the present role of microfluidics in food analysis. Liquid samples, as opposed to solid samples, require minor sample preparation, and they have been studied by direct injection using both amperometry and conductimetry. Thus, two single phenolic acids, vanillic and chlorogenic acid, were detected in red wine samples in less than 300 seconds using a CE microchip with amperometry detection [166]. The overall protocol involves simple sample preparation off-chip (filtration and dilution in an electrophoretic buffer), and the wine sample was directly injected on to the microchip. Selectivity and sensitivity were checked using different oxidation potentials in order to obtain relevant

Inorganic cations (water, juices, beer and milk): Ammonium Potassium Calcium Sodium Magnesium Inorganic anions (water): Chloride Nitrate Sulfate Vitamin C Preservatives Benzoate Sorbate

Beverages

Vitamin C tablets Soft drinks

Chlorogenic Vanillic

Analytes detected in food sample

Red wine (one)

Sample

Power & filtration off-chip

Dilution & filtration off-chip

Dilution & filtration off-chip

Sample preparation

Table 8.9 Lab-on-a-chip applications in food analysis

PMMA Simple cross Capacitively coupled contactless conductivity

PMMA microchip Contactlees conductivity

Glass CE microchip Electrochemical detection (SPE electrode)

Microfluidic approach

Separation time (<50 s) Good precision (less than 4%) External Calibration (r > 0.99) LODs 3–10 mg/l

Separation time (<40 s) LODs (90–150 mg/l)

Separation time (<300 s) Good precision Less than 7% in peak areas External Calibration (r > 0.99) LOD ¼ 10 mM

Analytical characteristics Ref.

(continued )

[168] Quantitative analysis is given for all analytes Comparison with CE conventional

[166] Use of red wine (high polyphenolic content) Use of different oxidation potentials during wine analysis High degree of [167] integration. Quantitative analysis is given for all analytes

Remarks

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 329

Arbutin Ascorbic acid

Water-soluble vitamins Pyridoxine B6 Vitamin C Folic acid

Vanillin Ethyl Vanillin

Pear pulps and commercial juices

Formulations

Vanilla pod and sugars

Glass CE microchip Electrochemical detection

Glass CE microchip Electrochemical detection

Miniaturized CE-ED Electrochemical detection (Cu electrode) PDMS CE Microchip Electrochemical detection

Microfluidic approach

Extraction & Glass CE maceration off-chip microchip Electrochemical detection

Solution & filtration off-chip

Juice filtration or solid-liquid extraction & filtration off-chip

Solid-liquid (þ)Catechin extraction & Epigallocatechin gallate filtration off-chip (
)Epicatechin Epicatechin gallate

Green tea extract (nutraceutical)

Filtration off-chip

Sample preparation

Tryptamine, tryptophan, tyramine

Analytes detected in food sample

Wine, beer

Sample

Table 8.9 (Continued) Remarks

Use of suited real sample (high concentration) Quantitative analysis is given for all analytes Separation of target antioxidant couples

[171]

[170]

[169]

Ref.

Calibration and [172] determination in less than 350s Quantitative analysis is given for all analytes Separation time (<200 s) Detection of frauds [173] Good precision less than 7% in peak heights

Separation time (<300 s) Good precision Less than 7% in peak areas External calibration (R2 > 0.99) LODs 8-10 mM Separation time (<200 s) Good precision Less than 7% in peak areas External Calibration (R2 > 0.99) Separation time (<130 s) High accuracy (errors less than 7%)

Separation time (<360 s) Homemade Good precision miniaturized (RSD 3.6%) CE system LODs 0.6–1.5 mM

Analytical characteristics

330 Miniaturization of Analytical Systems

Arg, The, Gly

Total isoflavones

Nitrite

Staphylococcus enterotoxin B

Green tea

Natural extract dietary supplements

Vegetables, fruits and meats

Milk (spiked)

Separation time (120 s) Precision chip-to-chip 3%

[177]

[176]

[175]

[174]

(continued )

Derivatization is not selective to amino acids Selectivity is improved by removing the high content of catechins in tea samples Validation of the method using HPLC Extraction & filtration Valveless Analysis time (60 s) Calibration off-chip microsystem integrated Accuracy is evaluated using a secondary standard (error less than 7%) Extraction & PMMA microchip Separation time (<120 s) Validation respect 732-cation resin (mFIA) Good precision AOAC official column & filtration Chemiluminisless than 4.1% method (Griess off-chip cence detection External calibration method) (r > 0.99) Interferents (Fe2þ, Ni2þ, V (V), LOD ¼ 4 mg/l Cu2þ, uric acid Hydrodynamic flow and vitamin C) Filtration off-chip PDMS microchip Sandwich-type --Fluorescence immunoassay detection Fluorescently labelled antibody LOD ¼ 0.5 ng/ml Hydrodynamic flow

Extraction & filtration PMMA microchip & derivatization of with LIF detection amino acids off-chip

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS) 331

Botulinum neurotoxin A

Folic acid

Infant formula

Analytes detected in food sample

Baby powder, milk, sucrose (spiked)

Sample

Table 8.9 (Continued)

Filtration off-chip

Filtration off-chip

Sample preparation

Commercial polyimide immuchip for ELISA Electrochemical detection (gold electrode)

Polycarbonate ELISA-chip Colourimetric detection

Microfluidic approach Sandwich-type immunoassay Cross-flow immunochromatography Horseradish peroxidase LOD ¼ 2 ng/ml Competitive immunoassay Hydrodynamic flow Analysis time ¼ 5 min RSD 7% alkaline phosphatase

Analytical characteristics

Ref.

Fivefold [178] enhancing of LOD respect the kit commercially established Validation respect [179] reference method (microbiological assay)

Remarks

332 Miniaturization of Analytical Systems

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

333

information using the wine sample. As expected, lower potentials provided greater selectivity, while higher ones led to increased sensitivity. Red wines are very complex and a lot of polyphenols are present since red wine is one of their most important sources. It is likely that only acid structures were detected because these compounds are prominent, while flavonoids did not show any interference because of the dilution performed. The elegant approach, based on the use of a universal detector (contactless conductivity) coupled to a microchip in the analysis of liquid samples, where selectivity is achieved by microchip separation, is illustrated in the papers of Hauser’s group [167]. These papers strategically combine the minor sample preparation requirements of these kinds of samples (just dilution and filtration) with the inherent characteristics and advantages of (contactless) conductivity. Indeed, water, wine, beer and fruit juice samples were simply diluted in electrolyte solution. Separations involving inorganic cations (ammonium, potassium, calcium, sodium, magnesium), inorganic anions (chloride, nitrate, sulphate) and the simultaneous analysis of up to 12 inorganic and organic anions (oxalate, tartrate, malate, citrate, succinate, acetate, lactate) in less than 100 seconds were obtained. In addition, a detailed evaluation of the methods including reliable quantitative data was given, indicating the power of these microdevices. The inherent advantages of miniaturization for food analysis using both liquid (beverages) and solid samples (vitamin C tablets) were also demonstrated, illustrating the separation of commonly used preservatives such as benzoate and sorbate, and vitamin C using a PMMA microchip coupled with contactless conductivity [168]. Ye et al. [169] fabricated a miniaturized CE-ED system by cutting a conventionally fused silica capillary and gluing it on to a Plexiglas plate. This system was used in the determination of three bioactive amines (tryptamine, tryptophan and tyramine) in wine and beer. Good reproducibility (3.6%), analysis time (6 minutes) and LOD (mM range) were obtained. Solid samples are very complex, and tedious off-chip extraction processes are always required before analysis. An analysis of five flavan-3-ols (catechins and derivatives) in commercially available green tea extract was performed by Henry’s group, using a PDMS microchip with MEKC and pulsed amperometric detection [170]. The detection of (þ)-catechin, epigallocatechin gallate, (
)-epicatechin and epicatechin gallate was achieved in less than 3 minutes in tea extracts (functional food). In addition, two more unidentified peaks were detected during analysis (probably attributable to other polyphenols). A quantitative analysis was performed and good accuracy was demonstrated. Sample preparation was still complex and off-chip (methanol-water, stirred extraction and filtration), and the sample matrix was still ‘easy’, since a nutraceutical containing high analyte concentrations was assayed; however, this paper was relevant because it clearly demonstrated the excellent analytical performance of these microsystems in sample quantitation.

334

Miniaturization of Analytical Systems

Escarpa’s group has proposed different strategies for the analysis of food samples, involving natural antioxidants (polyphenols and vitamins) using singlechannel microchip electrochemistry platforms with ED [171,172]. Polyphenolic analysis is very complex because many compounds can contribute to the complexity of the electropherogram obtained (as was also illustrated in Henry’s papers). Although fast detection of prominent natural antioxidants including ascorbic acid and polyphenols was obtained in less than 200 seconds, the analysis of pear samples was very complex and only arbutin in some pulps and ascorbic acid in juice (where up to four peaks were observed) were really detected [171]. Recovery studies using blank samples were carefully performed for both analytes in order to confirm their provisional identification. The detection of phlorizdin (the fingerprint analyte of apples, unexpected in pears) in the presence of arbutin (the fingerprint of pears) was also noticed, demonstrating the possibility of adulterations. Real detection and the quantitative analysis of water-soluble vitamins has been reported in several formulations containing one or more of the analysed vitamins at different concentration levels in less than 130 seconds [172]. The amount of each vitamin declared by the manufacturer (per capsule) on the label closely matched the values obtained using the microchip protocol (relative error ranging from 2 to 9%). The well-known problems associated with the inherent features of electrokinetic injection (modification of ionic strength, compatibility between buffers and solvents/matrices) made quantitation more difficult than expected in those samples with high mineral salt content. An approach for detecting fraud in vanilla samples based on the analysis of ethyl vanillin, vanillin and other related vanilla fingerprint markers has been proposed [173]. Vanillin and other fingerprint markers were really detected in vanilla beans (Vanilla planifolia and Vanilla tahitensis), confirming the authenticity of the samples and revealing the suitability of the approach. Ethyl vanillin was really detected in certain vanilla sugars, indicating the non-natural origin of the vanilla flavour used, since the presence of ethyl vanillin is unambiguous proof of fraud. Vanillin was really detected in the other sugar samples, in which other markers were not identified, indicating the non-natural origin. A detailed quantitative analysis was given and important differences in the quantitative levels of vanillin were found between natural and adulterated samples, revealing the possibilities of the method in control analysis. Three prominent green-tea amino acids (theanine (The), arginine (Arg) and glycine (Gly)), previously extracted and derivatized off-chip, were separated on to a PMMA microchip format using LIF detection in less than 2 minutes [174]. While sensitivity was not a challenge because of the detection route used, the selectivity of the analysis was improved by removing catechins, since the derivatization scheme involved both kinds of compounds, amino acids and catechins. Methodological calibration and analysis on board a planar microfluidic device to determine total isoflavones in soy samples was proposed [175]. Accuracy (systematic

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

335

error below 6%) was demonstrated for the first time using these microsystems with a secondary standard from the Drug Master File (SW/1211/03) as reference material. Ultrafast calibration and analysis of total isoflavones in soy samples were successfully integrated, taking only 60 seconds each, notably enhancing the analytical performance of these microdevices with an important decrease in overall analysis times (less than 120 seconds) and with an increase in accuracy by a factor of three. In addition, the method was applied to the analysis of soy tablets, with acceptable accuracy. D. He et al. [176] have developed a method to determine nitrite in different food samples (vegetables, fruits and meats) using a PMMA chip as a mFIA system with chemiluminescence detection (luminol-ferricyanide system). Exhaustive extraction and sample treatment (732-cation resin column) was needed per analysed solid sample. Good analytical performance was achieved with a reproducibility of 4.1%, a detection limit of 4 mg/l and an analysis time less than 2 minutes. An interference study was performed, discovering that Fe2þ, Ni2þ, V (V), Cu2þ, uric acid and vitamin C interfered with the analysis. Nevertheless, the method was compared with the AOAC official method (Griess method) with excellent results. To improve selectivity and sensitivity, some immunoassays have been developed on microfluidic devices as an alternative strategy for food analysis. Cheng et al. [177] determined staphylococcal enterotoxin B in spiked dry milk using a PDMS chip with fluorescence detection. Sandwich-type immunoassays were developed utilizing the chip as a chamber reaction. In the first step, the chip surface was functionalized with biotinylated antibodies. In the second step, the sample was hydrodynamically introduced on to the chip, and then the fluorescently-labelled secondary antibodies were added. Finally, the chip was imaged and quantified on a fluorescence scanner. An excellent LOD was achieved with this method (0.5 ng/ml). Botulinum neurotoxin A [178] was detected in spiked food samples (infant formula, milk, sucrose) on a polycarbonate ELISA chip with colourimetric detection. A cross-flow immunochromatographic assay was developed, in which a sandwich-type immunoassay with an enzyme-linked antibody was performed vertically, and a flow of the enzyme substrate revealed the presence of botulinum neurotoxin A horizontally. An excellent LOD (2.0 ng/ml) was achieved, enhancing five-fold the LOD established by the commercial kit. Another interesting ELISA on microfluidic devices was performed to determine folic acid in infant formula [179]. A commercial polyimide immunochip ELISA with eight channels and hydrodynamic flow was used. In this case, a competitive assay between free antigens (sample analyte) and fixed antigens on the surface of the chip was performed and the product of the enzyme reaction was monitored by ED (E ¼ 250 mV). Excellent analytical performance (reproducibility 7%, LOQ 2 ng/ml, analysis time 5 minutes) was obtained. The method was validated with the reference method (microbiological assay).

336

Miniaturization of Analytical Systems

Finally, one of the most important trends in miniaturization is the integration of nanomaterials in microfluidic systems. Lab-on-a-chip approaches based on the novel marriage between the electrokinetic microfluidic platform and nanotechnology have recently been proposed for analytical domains [180–182]. The analytical suitability of the electrokinetic microfluidic platform with CNTs as detectors has been proposed based on its dual format/use as both flow (with integrated calibration) and separation systems, independently, as shown in Figure 8.8. Two relevant applications of high significance, determination of total isoflavones and fast detection of antioxidant profiles, were chosen to demonstrate their analytical potential. Conceptually, the analytical challenges (poor selectivity and sensitivity, complexity of samples) from lab-on-a-chip devices can potentially be solved with the analytical promises offered by nanotechnologies (increase in selectivity and sensitivity). Thus, CNTs, as a representative example of modern analytical nanotechnology, offer an excellent opportunity in the building of lab-on-a-chip systems since they play a prominent role in selectivity (toward decrease of oxidation potential) and sensitivity (toward increase of the signal) of the analysis, avoiding complex sample preparation. To illustrate this interesting effect, Figure 8.9(a) shows the detection of dietary antioxidants using different CNT materials, where an enhancement in sensitivity independent of the CNT material used is clearly

Figure 8.8 Labs-on-a-chip coupled with nanodetectors for real analysis. Dual format of microfluidics with carbon nanotube detectors for analytical domains. (a) Flow-injection format for isoflavones determination. RB, running buffer; CR, calibration reservoir; SR, sample reservoir; (b) Separation format for fast detection of antioxidants. RB, running buffer reservoir; SR, sample reservoirs. (Reprinted from [181] reproduced by permission of the Royal Society of Chemistry)

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

337

Figure 8.9 Ultrafast detection of antioxidants in real samples with enhanced sensitivity. (a) Microchip electropherograms corresponding to a mixture of antioxidant standards with electrode materials studied: (i) SPE bare, (ii) SPE-SWCNT, (iii) SPE-MWCNT-B, (iv) SPEMWCNT-A. (b) Electropherograms corresponding to extract from apple peels: (i) SPE bare, (ii) SPE-MWCNT-A. (c) Analysis of apple extract with SPE-MWCNT-A: (i) pulp, (ii) peel. Peaks: (1) arbutin, (2) phloridzin, (3) (þ)-catechin, (4) rutin, (5) ascorbic acid. (Reproduced from [180] with permission of ACS, Copyright 2007)

observed in comparison to the detectors without CNT. Sharp and well-resolved peaks for all compounds were obtained, in contrast with tailing unresolved peaks for the same analytes using unmodified detectors. Since MWCNT-A gave the best sensitivity and the lowest noise, it was chosen as a material optimal for the analysis of antioxidants in food samples. Figure 8.9(b,c) shows selected examples of apple samples (peel and pulp). Figure 8.9(b) shows the surface area effect observed during the detection of (þ)-catechin and rutin from apple peels when MWCNT-Awas used. When unmodified detector was used, a smoothly tailing peak for (þ)-catechin was observed (electropherogram a). These results reveal the suitability of the modified electrodes for detection of both antioxidants. In addition, Figure 8.9(c) shows the electropherograms obtained for peel and pulp when MWCNT-A was used. As expected, rutin was only detected in peels (this analyte is a fingerprint of these

338

Miniaturization of Analytical Systems

matrices (Figure 8.9(b,c)), indicating again the suitability of the CNT as a novel material in antioxidant detection.

8.6 Outlook As has been clearly stated [2–5], microfluidics has seen the rapid development of new methods of fabrication, and of the components – the microchannels that serve as pipes, and other structures that form valves, mixers and pumps – that are essential elements of microchemical ‘factories’ on a chip. However, the electrokinetic approach is clearly dominant in analytical science. EOF benefits mTAS in three ways. First, it generates an essentially flat velocity profile, as compared to the parabolic profile of pressure-driven flow (in other words, separated zone analytes are flat). This results in less band broadening and hence an increase in separation efficiency. Second, the flow in multiple channels on a microchip can easily be controlled with a few electrodes. This further simplifies the apparatus by eliminating valves and pumps (dead volumes are zero and band broadening is reduced, yielding better separation efficiency). Third, scaling down the size of electrophoresis channels yields both a reduction in analysis time and an increase in separation efficiency. However, the impact of microfluidics on science has not yet been revolutionary. Revolutions in technology require both a broad range of different types of components and subsystems, and their integration into complete, functional systems. The field of microfluidics is in early adolescence, and still lacks both these essential requirements, in addition to the integration of components into systems that can be used by non-experts. As a field, it is a combination of unlimited promise, pimples and incomplete commitment. This is a very exciting time for the field, but we still do not know exactly what it will be when it grows up. Compact devices allow samples to be analysed at the point of need, rather than in a centralized laboratory. The advantages are compelling, but designing and making devices of reduced size that operate effectively is challenging. The pioneers recognized the huge financial input and research effort needed to realize the full potential of the concept. A clear proof of this is the state of the art in the analysis of real samples, given in this chapter. Currently, and strictly speaking, there are few papers that deal with the analysis of real samples using microfluidic devices, regardless of the application area. This clearly shows the enormous difficulties that the analysis of real samples presents on a microscale. Sensitivity and selectivity are the main challenges to developing a true lab-on-a-chip, with their improvement in terms of the use of ultrasensitive detectors and the integration of sample treatments on microfluidic devices being the main requirements. Moreover, additional efforts have to be made toward the validation of the methods in order to demonstrate the reliability of microfluidic systems. In spite

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

339

of these drawbacks, incredible analysis times, low sample and reagent consumptions, and interesting applications involving clinical, environmental and food analysis have been published. There is no doubt that the performance of sample preparation on-chip, the discovery of novel ultrasensitive detection routes and the systematic use of smart bioreactives and nanotechnology are the keys to the success and incorporation of microfluidic devices in our laboratories. An important analytical role for nanotecholgy in improving selectivity and sensitivity of lab-on-a-chip devices is highly expected [180,181].

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]

A. Manz, N. Grabner, H.M. Widmer, Sens. Actuators B., 1 (1990) 244. D. Janasek, J. Frazke, A. Manz, Nature, 442 (2006) 374. P.S. Dittrich, K. Tachikawa, A. Manz, Anal. Chem., 78 (2006) 3887. G.M. Whitesides, Nature, 442 (2006) 368. R. Daw, J. Finkelstein, Nature, 442 (2006) 367. A. Rios, A. Escarpa, M.C. Gonzalez, A.G. Crevillen, Trends Anal. Chem., 25 (2006) 467. A. Mello, Nature, 442 (2006) 394. A. Mello, Anal. Bioanal. Chem., 372 (2002) 12. O. Geschke, H. Klank, P. Telleman (Eds) Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004. R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003. M.J. Madou (Ed.), Fundamentals of Microfabrication (The Science of Miniaturization), CRC Press LLC, Boca Raton, 2000. J.P. Kutter, H. Klank,In: O. Geschke, H. Klank, P. Telleman (Eds), Microsystem Engineering of Lab-on-a-Chip Devices, Wiley-VCH, Weinheim, 2004, pp. 13–37. P. Yager, In: R.E. Oosterbroek, A. van der Berg (Eds), Lab-on-a-Chip. Miniaturized Systems for (Bio) Chemical Analysis and Synthesis, Elsevier, Amsterdam, 2003, pp. 115–150. A.E. Karnholz, Quantitative Analysis of Diffusion and Chemical Reaction in Pressure-driven Microfluidic Channels (Bioengineering), University of Washington, Seattle, 2002. J.P. Brody, P. Yager, Sensors and Actuators A, 58(1) (1997) 13. S.C. Jacobson, J.M. Ramsey, Anal. Chem., 69 (1997) 3212. A.W. Chow, In: C.S. Henry (Ed.), Microchip Capillary Electrophoresis: Methods and Protocols, Human Press, Totopwa, New Jersey, 2006, pp. 145–158. P.J.A. Kenis, R.F. Ismagilov, G.M. Whitesides, Science, 285 (1999) 83. A. Manz, D.J. Harrison, E.M.J. Verpoorte, J.C. Fettinger, A. Paulus, H. L€ uidi, M. Widmer, J. Chromatogr., 593 (1992) 253. D.J. Harison, K. Fluri, K. Seiler, Z. Fan, C.S. Effenhauser, A. Manz, Science, 261 (1993) 895. S. Haeberle, R. Zengerle, Lab Chip, 7 (2007) 1094. Y.N. Xia, G.M. Whitesides, Annu. Rev. Mater. Sci., 28 (1998) 153. Y.N. Xia, G.M. Whitesides, Angew. Chem., Int. Ed. Engl., 37 (1998) 551. C.S. Effenhauser, G.J.M. Bruin, A. Paulus, M. Ehrat, Anal. Chem., 69 (1997) 3451. E. Delamarche, A. Bernard, H. Schmid, B. Michel, H. Biebuyck, Science, 276 (1997) 779. A.Y. Fu, C. Spence, A. Scherer, F.H. Arnold, S.R. Quake, Nat. Biotechnol., 17 (1999) 1109. K. Hosokawa, T. Fujii, I. Endo, Anal. Chem., 71 (1999) 4781. D.C. Duffy, J.C. McDonald, O.J.A. Schueller, G.M. Whitesides, Anal. Chem., 70 (1998) 4974. S.K. Sia, G.M. Whitesides, Electrophoresis, 24 (2003) 3563.

340 [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74]

Miniaturization of Analytical Systems M.A. Unger, H.P. Chou, T. Thorsen, A. Scherer, S.R. Quake, Science, 288 (2000) 113. S.R. Quake, A. Scherer, Science, 290 (2000) 1536. Fluidigm Corporation, USA (www.fluidigm.com). D.C. Duffy, H.L. Gillis, J. Lin, N.F. Sheppard, G.J. Kellogg, Anal. Chem., 71 (1999) 4669. D.J. Harrison, A. Manz, Z.H. Fan, H. Ludi, H.M. Widmer, Anal. Chem., 64 (1992) 1926. D.J. Harrison, K. Fluri, K. Seiler, Z.H. Fan, C.S. Effenhauser, A. Manz, Science, 261 (1993) 895. Agilent Technologies, USA (www.agilent.com). M. Joanicot, A. Ajdari, Science, 309 (2005) 887. T. Nisisako, T. Torii, T. Higuchi, Lab Chip, 2 (2002) 24. H. Song, J.D. Tice, R.F. Ismagilov, Angew. Chem. Int. Ed. Engl., 42 (2003) 768. J. Lee, H. Moon, J. Fowler, T. Schoellhammer, C.J. Kim, Sens. Actuators A, 95 (2002) 259. D. Chatterjee, B. Hetayothin, A.R. Wheeler, D.J. King, R.L. Garrell, Lab Chip, 6 (2006) 199. V. Srinivasan, V.K. Pamula, R.B. Fair, Lab Chip, 4 (2004) 310. A. Wixforth, Superlattices Microstruct., 33 (2003) 389. A. Wixforth, C. Strobl, C. Gauer, A. Toegl, J. Scriba, Z. von Guttenberg, Anal. Bioanal. Chem., 379 (2004) 982. Advalytix AG, Germany (www.advalytix.de). Z. Guttenberg, H. Muller, H. Habermuller, A. Geisbauer, J. Pipper, J. Felbel, M. Kielpinski, J. Scriba, A. Wixforth, Lab Chip, 5 (2005) 308. M.A. Schwarz, P.C. Hauser, Lab Chip, 1 (2001) 1. W.R. Vandaveer, S.A. Pasas, R.S. Martin, S.M. Lunte, Electrophoresis, 23 (2002) 3667. J. Wang, Electroanalysis, 17 (2005) 1133. M. Pumera, A. Merkoc¸i, S. Alegret, Trends Anal. Chem., 25 (2006) 219. P.J. Viskari, J.P. Landers, Electrophoresis, 27 (2006) 1797. A. van der Berg, P. Bergveld, Lab Chip, 6 (2006) 1266. J. Lichtenber, N.F. de Rooij, E. Verpoorte, Talanta, 56 (2002) 233. S.K. Mohanty, D. Kim, D.J. Beebe, Electrophoresis, 27 (2006) 3772. C.K. Fredrickson, Z.H. Fan, Lab Chip, 4 (2004) 526. L. Chen, A. Manz, P.J.R. Day, Lab Chip, 7 (2007) 1413. L. Zhu, Q. Zhang, H. Feng, S. Ang, F.S. Chau, W.T. Liu, Lab Chip, 4 (2004) 337. P. Sethu, A. Sin, M. Toner, Lab Chip, 6 (2006) 83. J.S. Marcus, W.F. Anderson, S.R. Quake, Anal. Chem., 78 (2006) 3084. P. Gascoyine, C. Mahidol, M. Ruchirawat, J. Satayvivad, P. Watcharasit, F.F. Becker, Lab Chip, 2 (2002) 70. P. Gascoyne, J. Satayavivad, M. Ruchirawat, Acta Trop., 89 (2004) 357. E.T. Lagally, S.H. Lee, H.T. Soh, Lab Chip, 5 (2005) 1053. J. El-Ali, P.K. Sorger, K.F. Jensen, Nature, 442 (2006) 403. C.E. Sims, N.L. Allbritton, Lab Chip, 7 (2007) 423. B. Huang, H. Wu, D. Bhaya, A. Grossman, S. Granier, B.K. Kobilka, R.N. Zare, Science, 315 (2007) 81. J. Ueberfeld, S.A. El-Difrawy, K. Ramdhanie, D.J. Ehrlich, Anal. Chem., 78 (2006) 3632. L.A. Legendre, J.M. Bienvenue, M.G. Roper, J.P. Ferrance, J.P. Landers, Anal. Chem., 78 (2006) 1444. J.E. Prest, S.J. Baldock, P.J. Day, P.R. Felden, N.J. Goddard, B.J. Treves, J. Chromoatogr. A., 1156 (2007) 154. L. Chen, J.E. Prest, P.R. Fielden, N.J. Goddard, A. Manz, P.J. Day, Lab Chip, 6 (2006) 474. P. Wilding, L.J. Kricka, J. Cheng, G. Hwichia, M.A. Shofner, P. Fortina, Anal. Biochem, 257 (1998) 95. P.K. Yuen, L.J. Kricka, P. Fortina, N.J. Panaro, T. Sakazume, P. Wilding, Genome Res., 11 (2001) 405. R.H. Liu, J.N. Yang, R. Lenigk, J. Bonanno, P. Grodzinski, Anal. Chem., 76 (2004) 1824. N.C. Cady, S. Stelick, C.A. Batt, Biosens. Bioelectron, 19 (2003) 59. N.C. Cady, S. Stelick, M.V. Kunnavakkam, C.A. Batt, Sens. Actuators B, 107 (2005) 332.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

341

[75] J.G. Lee, K.H. Cheong, N. Huh, S. Kim, J.W. Choi, C. Ko, Lab Chip, 6 (2006) 886. [76] R.K. Saiki, S. Scharf, F. Faloona, K.B. Mullis, G.T. Horn, H.A. Erlich, N. Arnheim, N. Science, 230 (1985) 1350. [77] L.J. Kricka, P. Wilding, Anal. Bioanal. Chem., 377 (2003) 820. [78] P.A. Auroux, Y. Koc, A. de Mello, A. Manz, P.J.R. Day, Lab Chip, 4 (2004) 534. [79] M.G. Roper, C.J. Easley, J.P. Landers, Anal. Chem., 77 (2005) 3887. [80] C.S. Zhang, J.L. Xu, W.L. Ma, W.L. Zheng, Biotechnol. Adv., 24 (2006) 243. [81] P.J. Day, Expert Rev. Mol. Diagn., 6 (2006) 23. [82] M.A. Northrup, B. Benett, D. Hadley, P. Landre, S. Lehew, J. Richards, P. Stratton, Anal. Chem., 70 (1998) 918. [83] M.S. Ibrahim, R.S. Lofts, P.B. Jarhling, E.A. Henchal, V.W. Weedn, M.A. Northrup, P. Belgrader, Anal. Chem., 70 (1998) 2013. [84] A. Guliksen, L. Solli, F. Karlsen, H. Rogne, E. Hovig, T. Nordstrom, R. Sirevag, Anal. Chem., 76 (2004) 9. [85] Q. Xiang, B. Xu, R. Fu, D. Li, Biomed. Microdev., 7 (2005) 273. [86] G.Q. Hu, Q. Xiang, R. Fu, B. Xu, R. Venditti, D.Q. Li, Anal. Chim. Acta, 557 (2006) 146. [87] T. Nakayama, Y. Kurosawa, S. Furui, K. Kerman, M. Kobayashi, S.R. Rao, Y. Yonezawa, K. Nakano, A. Hino, S. Yamamura, Y. Takamura, E. Tamiya, Anal. Bioanal. Chem., 386 (2006) 1327. [88] Z. Wang, A. Sekulovic, J.P. Kutter, D.D. Bang, A. Wolff, Electrophoresis, 27 (2006) 5051. [89] A.T. Woolley, D. Hadley, P. Landre, A.J. deMello, R.A. Mathies, M.A. Northrup, Anal. Chem., 68 (1996) 4081. [90] L.C. Waters, S.C. Jacobson, N. Kroutchinina, J. Khandurina, R.S. Foote, J.M. Ramsey, Anal. Chem., 70 (1998) 158. [91] L.C. Waters, S.C. Jacobson, N. Kroutchinina, J. Khandurina, R.S. Foote, J.M. Ramsey, Anal. Chem., 70 (1998) 5172. [92] W.C. Dunn, S.C. Jacobson, L.C. Waters, N. Kroutchinina, J. Khandurina, R.S. Foote, M.J. Justice, L.J. Stubbs, J.M. Ramsey, Anal. Biochem., 277 (2000) 157. [93] E.T. Lagally, P.C. Simpson, R.A. Mathies, Sens. Actuators B, 63 (2000) 138. [94] E.T. Lagally, I. Medintz, R.A. Mathies, Anal. Chem., 73 (2001) 565. [95] E.T. Lagally, C.A. Emrich, R.A. Mathies, Lab Chip, 1 (2001) 102. [96] J.W. Hong, T. Fujii, M. Seki, T. Yamamoto, I. Endo, Electrophoresis, 22 (2001) 328. [97] I. Rodriguez, M. Lesaicherre, Y. Tie, Q.B. Zou, C. Yu, J. Singh, L.T. Meng, S. Uppili, S.F.Y. Li, P. Gopalakrishnakone, Z.E. Selvanayagam, Electrophoresis, 24 (2003) 172. [98] C.G. Koh, W. Tan, M.Q. Zhao, A.J. Ricco, Z.H. Fan, Anal. Chem., 75 (2003) 4591. [99] Z.M. Zhou, D.Y. Liu, R.T. Zhong, Z.P. Dai, D.P. Wu, H. Wang, Y.G. Du, Z.N. Xia, L.P. Zhang, X.D. Mei, B.C. Lin, Electrophoresis, 25 (2004) 3032. [100] G.V. Kaigala, R.J. Huskins, J. Preiksaitis, X.L. Pang, L.M. Pilarski, C.J. Backhouse, Electrophoresis, 27 (2006) 3753. [101] F.C. Huang, C.S. Liao, G.B. Lee, Electrophoresis, 27 (2006) 3297. [102] C.N. Liu, N.M. Toriello, R.A. Mathies, Anal. Chem., 78 (2006) 5474. [103] Y. Hataoka, L.H. Zhang, Y. Mori, N. Tomita, T. Notomi, Y. Baba, Anal. Chem., 76 (2004) 3689. [104] Y.J. Liu, C.B. Rauch, R.L. Stevens, R. Lenigk, J.N. Yang, D.B. Rhine, P. Grodzinski, Anal. Chem., 74 (2002) 3063. [105] D. Trau, T.M. Lee, A.I. Lao, R. Lenigk, I.M. Hsing, N.Y. Ip, M.C. Carles, N.J. Sucher, Anal. Chem., 74 (2002) 3168. [106] J. Wang, Z.Y. Chen, P. Corstjens, M.G. Mauk, H.H. Bau, Lab Chip, 6 (2006) 46. [107] T.M.H. Lee, M.C. Carles, I.M. Hsing, Lab Chip, 3 (2003) 100. [108] S.W. Yeung, T.M. Lee, H. Cai, I.M. Hsing, Nucleic Acids Res., 34 (2006) e118. [109] A.J. de Mello, N. Beard, Lab Chip, 3 (2003) 11N. [110] A.G. Crevillen, M. Herv
as, M.A. L
opez Gil, M.C. Gonzalez, A. Escarpa, Talanta, 27 (2007) 342. [111] L.J. Kricka, P. Fortina, Clin. Chem., 48 (2002) 1620. [112] A.J. Tudos, G.A.J. Besselink, R.B.M. Schasfoort, Lab Chip, 1 (2001) 83.

342 [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157]

Miniaturization of Analytical Systems E. Verpoorte, Electrophoresis, 23 (2002) 677. U. Bilitewski, M. Genrisch, S. Kadow, G. Mersal, Anal. Bioanal. Chem., 377 (2003) 556. S. Thorslund, O. Klett, F. Nikolajeff, K. Markides, J. Bergquist, Biomed. Microdev., 8 (2006) 73. S.F.Y. Li, L.J. Kricka, Clin. Chem., 5 (2006) 37. B. Baena, A. Cifuentes, C. Barbas, Electrophoresis, 26 (2005) 2622. W.R. Vandaveer, S. Pasas-Farmer, D.J. Fischer, C.N. Frankenfeld, S.M. Lunte, Electrophoresis, 25 (2004) 3528. B.H. Weiller, L. Ceriotti, T. Shibata, D. Rein, M.A. Roberts, J. Lichtenberg, J.B. German, N.F. de Rooij, E. Verpoorte, Anal. Chem., 74 (2002) 1702. L. Ceriotti, T. Shibata, B. Folmer, B. Weiller, M.A. Roberts, N.F. de Rooij, E. Verpoorte, Electrophoresis, 23 (2002) 3615. J.A. Lapos, D.P. Manica, A.G. Ewing, Anal. Chem., 74 (2002) 3348. M. Johirul, A. Shiddiky, R. Kim, Y. Shim, Electrophoresis, 26 (2005) 3043. M. Vlckova, M.A. Schwarz, J. Chromatogr. A, 1142 (2007) 214. N.J. Munro, Z. Huang, D.N. Finegold, J.P. Landers, Anal. Chem., 72 (2000) 2765. J.C. Fanguy, C.S. Henry, Electrophoresis, 23 (2002) 767. J. Wang, M.P. Chatrathi, Anal. Chem., 75 (2003) 525. C.D. Garcia, C.S. Henry, Analyst, 129 (2004) 579. J. Wang, M.P. Chatrathi, G.E. Collins, Anal. Chim. Acta, 585 (2007) 11. Y. Du, J.L. Yan, W.Z. Zhou, X.Y. Yang, E.K. Wang, Electrophoresis, 25 (2004) 3853. V. Srinivasan, V.K. Pamula, R.B. Fair, Lab Chip, 4 (2004) 310. H. Nakamura, Y. Murakami, E. Tamiya, I. Karube, Anal. Chem., 73 (2001) 373. H. Moon, A.R. Wheeler, R.L. Garrell, J.A. Loo, C.J. Kim, Lab Chip, 6 (2006) 1213. P. Caglar, S.A. Tuncel, N. Malcik, J.P. Landers, J.P. Ferrance, Anal. Bioanal. Chem., 386 (2006) 1303. T. Noda, H. Takao, K. Yoshioka, N. Oku, M. Ashiki, K. Sawada, K. Matsumoto, M. Ishida, Sens. Actuators B, 119 (2006) 245. W.S.B. Yeung, G.A. Luo, Q.G. Wang, J.P. Ou, J. Chromatogr. B, 797 (2003) 217. A.E. Herr, D.J. Throckmorton, A.A. Davenport, A.K. Singh, Anal. Chem., 77 (2005) 585. A.E. Herr, A.V. Hatch, D.J. Throckmorton, H.M. Tran, J.S. Brennan, W.V. Giannobile, A.K. Singh, PNAS, 104 (2007) 5268. T.M. Phillips, Electrophoresis, 25 (2004) 1652. T.M. Phillips, E. Wellner, J. Chromatogr. A, 1111 (2006) 106. M.G. Roper, J.G. Shackman, G.M. Kahlgren, R.T. Kennedy, Anal. Chem., 75 (2003) 4711. T. Lim, H. Ohta, T. Matsunaga, Anal. Chem., 75 (2003) 3316. K. Sato, M. Tokeshi, H. Kimura, T. Kitamori, Anal. Chem., 73 (2001) 1213. K. Tsukagoshi, N. Jinno, R. Nakajima, Anal. Chem., 77 (2005) 1684. J. Moorthy, G.A. Mensing, D. Kim, S. Mohanty, D.T. Eddington, W.H. Tepp, E.A. Johnson, D.J. Beebe, Electrophoresis, 25 (2004) 1705. V. Linder, E. Verpoorte, N.F. De Rooij, H. Sigris, W. Thormann, Electrophoresis, 23 (2002) 740. M. Wolf, D. Juncker, B. Michel, P. Hunziker, E. Delamarche, Biosens. Bioelectron., 19 (2004) 1193. R. Kurita, Y. Yokota, Y. Sato, F. Mizutani, O. Niwa, Anal. Chem., 78 (2006) 5525. G. Chen, Y.H. Lin, J. Wang, Talanta, 68 (2006) 497. L. Marle, G.M. Greenway, Trends Anal. Chem., 24 (2005) 795. J. Wang, J. Zima, N.S. Lawrence, M.P. Chatrathi, Anal. Chem., 76 (2004) 4721. J. Wang, M.P. Chatrathi, Anal. Chem., 73 (2001) 1804. J. Wang, M. Pumera, G.E. Collins, A. Mulchandani, Anal. Chem., 74 (2002) 6121. A. Hilmi, J.T. Luong, Environ. Sci. Technol., 34 (2000) 3046. J. Fischer, J. Barek, J. Wang, Electroanalysis, 18 (2006) 195. D. Shin, D.A. Tryk, A. Fujishima, A. Muck, G. Chen, J. Wang, Electrophoresis, 25 (2004) 3017. J. Wang, M.P. Chatrathi, B. Tian, Anal. Chim. Acta, 416 (2000) 9. C.D. Garcia, G. Engling, P. Herckes, J.L. Collett, C. Henry, Environ. Sci. Technol., 39 (2005) 618.

Miniaturization of the Entire Analytical Process II: Micro Total Analysis Systems (mTAS)

343

[158] Y. Ding, C.D. Garcia, Analyst, 131 (2006) 208. [159] S. Wakida, A. Chiba, T. Matsuda, K. Fukushi, H. Nakanishi, X. Wu, H. Nagai, S. Kurosawa, S. Takeda, Electrophoresis, 22 (2001) 3505. [160] I. Chang, J.J. Tulock, J. Liu, W. Kim, D.M. Cannon, Jr, Y. Lu, P.W. Bohn, J.V. Sweedler, D.M. Cropek, Environ. Sci. Technol., 39 (2005) 3756. [161] Y. Xu, F.G. Bessoth, J.C.T. Eijkel, A. Manz, Analyst, 125 (2000) 677. [162] S. Fujii, T. Tokuyama, M. Abo, A. Okubo, Anal. Sci., 20 (2004) 209. [163] G. Deng, G.E. Collins, J. Chromatogr. A, 989 (2003) 311. [164] R. Bodor, V. Madajova, D. Kaniansky, M. Masar, M. Johnck, B. Stanislawski, J. Chromatogr. A, 916 (2001) 155. [165] A. Escarpa, M.C. Gonz
alez, M.A. L
opez Gil, A.G. Crevill
en, M. Herv
as, M. Garc
ıa, Electrophoresis, 29 (2008) 4852. [166] M. Scampicchio, J. Wang, S. Mannino, M.P. Chatrathi, J. Chromatogr. A, 1049 (2004) 189. [167] P. Kuban, P.C. Hauser, Electrophoresis, 26 (2005) 3169. [168] W.S. Law, P. Kuban, J.H. Zhao, F.Y. Li, P.C. Hauser, Electrophoresis, 26 (2005) 4648. [169] Q. Chu, Y. Guan, C. Geng, J. Ye, Anal. Lett., 39 (2006) 729. [170] R.W. Hompesch, C.D. Garcia, D.J. Weiss, J.M. Vivanco, C.S. Henry, Analyst, 130 (2005) 694. [171] A.J. Blasco, I. Barrigas, M.C. Gonzalez, A. Escarpa, Electrophoresis, 26 (2005) 4664. [172] A.G. Crevillen, A.J. Blasco, M.C. Gonzalez, A. Escarpa, Electrophoresis, 27 (2006) 5110.
[173] M. Avila, M. Zougagh, A. Escarpa, A. R
ıos (in press). [174] M. Kato, Y. Gyoten, K. Sakai-Kato, T. Toyo’oka, J. Chromatogr. A, 1013 (2003) 183. [175] A.J. Blasco, A.G. Crevillen, P. de la Fuente, M.C. Gonzalez, A. Escarpa, Analyst, 132 (2007) 323. [176] D. He, Z. Zhang, Y. Huang, Y. Hu, Food Chem., 101 (2007) 667. [177] Y. Dong, K.S. Phillips, Q. Cheng, Lab Chip, 6 (2006) 675. [178] S. Han, J.H. Chob, I.H. Choa, E.H. Paek, H.B. Ohc, B.S. Kim, C. Ryu, K. Lee, Y.K. Kim, S.H. Paek, Anal. Chim. Acta, 587 (2007) 1. [179] D. Hoegger, P. Morier, C. Vollet, D. Heini, F. Reymond, J.S. Rossier, Anal. Bioanal. Chem., 387 (2007) 267.
[180] A.G. Crevill
en, M. Avila, M. Pumera, M.C. Gonz
alez, A. Escarpa, Anal. Chem., 79 (2007) 7408. [181] A.G. Crevill
en, M. Pumera, M.C. Gonz
alez, A. Escarpa, Lab Chip, 9 (2009) 346. [182] M. Pumera, A. Escarpa, Electrophoresis (in press).

9 Portability of Miniaturized Analytical Systems 9.1

Introduction

Miniaturization is a growing trend in the field of analytical chemistry. It is important that analytical methods can be taken out of the laboratory, into the field. To this end, analytical chemists are developing portable instruments for fieldwork. One example is electrochemical analysers, which satisfy many of the requirements for onsite and in situ measurements. Handheld glucose meters, widely used for home diabetes testing, represent an example of a commercially successful miniaturized electrochemical device. There is no doubt about the need for portable devices to be used in field applications and point-of-care diagnosis or environmental control. A miniaturized gas analysis system is an especially attractive approach. It can be placed in a hospital room but also next to a patient – even in a patient’s home. Miniaturization is the key to establishing instrumentation for onsite analysis. By employing miniaturization, sophisticated instruments can be readily used in the field as sensors. The merits of miniaturization and the use of microdevices include: . .

.

High-speed analyses. On some occasions, a high enrichment factor can be achieved in microdevices because of their high surface-to-volume ratio. This is especially relevant in gas analysis. Very fast heating/cooling. This is important in the development of gas sensors but also in the miniaturization of PCR reactions.

Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

346 .

.

Miniaturization of Analytical Systems

Power consumption is small. In fact, the entire system can be driven by a battery for use in the field. Environmental friendliness, because of the small consumption of chemicals.

9.2 Portable Gas Analysers Miniaturization of gas chromatographs (GCs) brings many advantages, such as portability for field applications and fast operation using the GC–GC system. Miniaturizations has been attempted with conventional fluidic tubes and connectors. Microfabrication is an attractive option for the development of greatly improved instruments, and many investigations have been reported. Portable gas chromatographs (PGCs) show potential as instruments for field monitoring. They allow qualitative and quantitative analysis of unknown gas mixtures. The key components of the simplest GC are: injector, chromatographic column(s) and detector(s). Existing GCs ensure high accuracy and reliability of analysis but are heavy and large, and suffer failures of their mechanical parts. These features mean that, normally, in classical fabrication procedures, GCs do not meet the conditions for outdoor industrial and environmental applications. They are not portable devices able to work anywhere and to make accurate, automatic, continuous (online) analyses of gas samples. It seems that the way to solve this problem is a miniaturization of the key components of GC and its fabrication by the use of micromechanical techniques, followed by future integration of the instrument on a single chip. The first integrated GC was developed in the late 1970s by Terry et al. [1]. This system presented many technical problems and was not implemented in mass production. In the 1980s, Agilent developed a PGC characterized by two important components: a microelectromechanical systems (MEMS)-like micromechanical silicon/glass injection system and a thermal conductivity detector, extremely miniaturized. Since then, several PGCs have been described. As an example of the potential of current technology, Figure 9.1 presents the PGC designed by Dziuban and coworkers [2]. The pneumatic and electrical components of the portable instrument are presented in Figure 9.2. This PGC has dimensions of 150 · 450 · 300 mm and weights about 3.5 kg. At the front panel of the PGC, connections for the inlet of the carrier gas (helium) supply, the inlet of the analyte, pressure and temperature regulators and the polycarbonate (PC) port are located. External small He gas bottles and portable PCs have to be used. Portable microchip GC was applied to the identification and reduction of fugitive emissions in a pharmaceutical manufacturing plant [3], and emissions of VOCs were monitored at landfills using a PGC [4]. Lu et al. developed a PGC with tuneable retention and sensor array detection for the determination of complex vapour mixtures [5]. Two capillary columns were connected in series and the column temperatures were individually controlled. Since the carrier gas was air and

Portability of Miniaturized Analytical Systems

347

Figure 9.1 Portable gas chromatograph: (a) the device with coworking equipment; note small He bottle, analyte reservoir and portable PC; (b) internal view of the chromatograph. (Reprinted from [2] with permission from Elsevier)

the detector was a surface acoustic wave (SAW) sensor array, this GC did not need a gas cylinder. The entire system was incorporated in a box just 35 · 35 · 10 cm. Ji et al. applied a similar PGC, coupled with SPME sampling, to water BTEX analysis [6]. Recently, some groups have been investigating new concepts in microchip GC technology. Figure 9.3 shows the typical components of a microdevice for GC determination [7]. Lu et al. integrated onboard calibration, sample preconcentratorfocuser thermal desorber, temperature-programmed separations and ‘spectral’ detection with an array of microsensors [8]. The components were connected with

348

Miniaturization of Analytical Systems

Figure 9.2 Structure of the PGC: (a) pneumatic block scheme; (b) electrical block scheme. (Reprinted from [2] with permission from Elsevier)

a capillary tube. Ambient air was used as a carrier gas, making the use of a cylinder gas unnecessary. Eleven types of VOC were separated in only 90 seconds and the LODs were in the low ppbvs for a 0.25 l sample. This device was also applied to analyses of markers of environmental tobacco smoke (ETS) [9]. The target ETS species were 2,5-dimethylfuran and 4-ethenylpyridine (4-EP) (as a surrogate for 3-EP), and their LODs were 0.58 and 0.08 ppb, respectively. The sample volume was 1 l and the analysis could be repeated every 15 minutes. Chemical sensors were used as the GC detector; the resistances of gold thiolate monolayers were measured and five types of thiolate were used in the sensor array. Very rapid analysis is available with a micro-GC system. Lambertus et al. integrated a preconcentrator and a GC separation/detection system [10]. The separation column in this device was 150 mm · 250 mm · 25 cm coiled in a 1.2 cm

Portability of Miniaturized Analytical Systems

349

Figure 9.3 Microsystem for GC determination. The top scheme shows typical micro-GC for airsample analysis. The pictures are examples of a micropreconcentrator (left), a microchip column (centre) and a micro-DMS detector (right). (Reprinted from [7] with permission from Elsevier)

square die. At 25 cm/s carrier velocity, 625 theoretical plates and a holdup time of 1.9 seconds were obtained. A mixture of 11 gases was separated into individual components in only 10 seconds, with a column temperature increasing at a rate of 10
C/s. To improve the separation, a longer column (3 m) with retained channel geometry was tested [11]. The wall of the channel was statically coated with PDMS by filling it with 0.4% dimethyl polysiloxane to form a thinner and more uniform coating than the dynamic coating. As a result, 12 500 theoretical plates were obtained. If a separation system is miniaturized, the detector must be much smaller than the separation column in order to obtain adequately separated signals. Also, the use of auxiliary gases is not suitable for field application. Hence, a typical detector such as FID is not suitable. A thermal conductivity detector (TCD) is simple, and a microfabricated TCD sensor is also available [12]. A micro-TCD consumes very little power and is suitable for use in field monitoring. However, its sensitivity is low and so its application is limited. New technologies for compact and highly sensitive detectors have been developed. For example, SAW devices were attached downstream of a separation column [13]. To improve performance, several kinds of SAW device were arranged in an array. Mass spectrometry (MS) is one of the most attractive detection methods for GC. The mass spectrometer itself can be used without a GC system. Increasing efforts are being directed at developing miniature MSs, including hand-portable ones, and at achieving the performance characteristics of traditional laboratory instruments [14]. MEMS-based MSs have been investigated as ion traps [15–17],

350

Miniaturization of Analytical Systems

asynchronous ion-shield mass filters [18] and for time-of-flight [19]. Also, a handheld MS including a vacuum system and a data process board has been developed [20]. The development of these micro-MS systems is in progress, and such devices are expected to be available for use in environmental and onsite applications in the near future. Commercialization of portable MSs has grown rapidly, with several companies offering products. A field-ready portable GC/MS, which uses a cylindrical ion trap (CIT) mass analyser and an electron ionization/chemical ionization (EI/CI) source has been released by Griffin Analytical Technologies, Inc. (West Lafayette, IN) and a portable CIT-based leak detection instrument has been developed by MKS Instruments, Inc. (Wilmington, MA). Portable GC/MS instruments utilizing quadrupole mass analysers are also available from Inficon (E. Syracuse, NY), Constellation Technology, Corp. (Largo, FL) and Bruker Daltonics, Inc. (Billerica, MA). A portable time-of-flight instrument, utilizing permeable membranes for sample introduction, is available from Kore Technology, Ltd (Ely, Cambridgeshire, UK). Most commercial equipment currently available for portable air monitoring employs GC in combination with MS, and while this offers an added degree of separation, it also increases the time required for analysis. In a typical GC/MS analysis, preconcentration followed by thermal separation of components on a capillary column may increase the analysis time to several minutes, with the actual MS analysis taking only a small percentage of the total analysis time, making MS/MS more appealing for analyses requiring high speed as well as specificity. This choice of requirements normally dictates that the mass analyser be an ion trap, since alternatives do not yet adequately meet portability needs. Several ionization techniques have been implemented on portable MS instruments, including EI and CI [21,22], matrix-assisted laser desorption ionization (MALDI) in the case of time-of-flight analysers [23,24], and reduced-pressure photoionization [25]. Multiple sample introduction methods have been investigated for both air and water samples, and membrane introduction MS inlets [26,27] and fibre introduction systems [28–30] have been particularly common. While many ionization and sample introduction technologies have been investigated on portable MSs, one shortcoming is direct air analysis through an atmospheric inlet. The ability to sample air directly via an atmospheric pressure inlet and to use the simple ionization methodology of atmospheric pressure chemical ionization (APCI) is ideal for portable instrumentation aimed at real-time trace analysis. APCI, first reported by Horning et al. [25], is known for its simplicity, sensitivity and wide range of applications [31–33]. Corona discharge (CD) is now commonly used to produce primary reagent ions, replacing the traditional radioactive 63Ni sources. Both of these sources create, through a chain of ion/molecule reactions starting with N2þ and O2þ, protonated water clusters that are ultimately responsible for protonating the analytes, leading to the ionic species observed in the positive ion mode, namely MHþ [34].

Portability of Miniaturized Analytical Systems

351

Figure 9.4 Details of the electrical array chip. The chip is equipped with 28 positions of 200 mm diameter. Of these, 14 positions have test structures and 14 have interdigitated gold ultramicroelectrode arrays (IDAs) with 0.8 mm gap and 0.8 mm width. The positions used for DNA assay are marked. Also shown is a scheme of the DNA-ELISA. (From [36] with kind permission from Springer)

9.3

Portable Electrochemical Analysers

The miniaturization and portability of electrochemical systems, along with their low power demands, make these microdevices especially interesting. Modern microfabrication technology allows us to replace the traditional bulky electrodes and cumbersome cells with easy-to-use miniaturized electrochemical systems. Portable electrochemical analysers have been widely used and exploited in medical applications. The determination of glucose is a clear example. Noninvasive approaches for continuous glucose monitoring represent a promising route for getting round the challenges of implantable devices and the inconvenience of fingerstick sampling. In particular, Cygnus, Inc. has developed an attractive, wearable glucose monitor, based on the coupling of reverse iontophoretic collections of glucose and biosensor functions [35]. The new GlucoWatch Biographer provides three glucose readings per hour for up to 12 hours. Electrochemical biochips have been widely developed as a tool for microarrays and continuous monitoring. Albers et al. [36] have developed a portable and miniaturized amperometric biosensor device capable of evaluating biomolecular interactions by measuring the redox recycling of ELISA products, as well as electrically monitoring metabolites. Figure 9.4 shows a detail of the electrical array chip and a scheme of the DNA-ELISA. As can be seen, target molecules bound specifically to the thiol-linked capture molecules and were than enzyme-labelled.

9.4

Portable Optical Analysers

Miniaturized and portable device based on optical measurements have also been described in the literature. One example is the development of a portable surface

352

Miniaturization of Analytical Systems

Figure 9.5 Portable SPR sensor prototype system including sensor, optics, electronics and flowdelivery system: (a) scheme of the SPR device; (b) commercial b-SPR platform. (Reprinted from [37] with permission from Elsevier)

plasmon resonance [37]. The scheme of the portable sensors is described in Figure 9.5. The sensing mechanism is based on variations of the refractive index of the medium adjacent to the metal sensor surface during the interaction of the analyte with its corresponding recognition element, previously immobilized at the sensor surface. Once the biomolecular interaction has taken place, the regeneration of the surface allows the reusability of the same sensor surface in a large number of assay cycles. Development in other fields such as electronics is critical to making portable equipment. For example, a compact fluorimetric sensor equipped with an LED source and a high-sensitivity PMT detector have been described [38]. The potential of this equipment has been demonstrated for the determination of aflotoxin AFM1 in liquid solutions. The instrument allows the rapid monitoring of analyte in milk, enabling the detection of concentrations up to the legal limit. Another interesting example is a portable total-reflexion X-ray fluorescence spectrometer [39]. The spectrometer mainly consists of a 1.5 W X-ray tube, a waveguide-type slit, a detector and a sample carrier (a quartz optical flat), and these components are contained in an attach
e case-type box. Continuum X-rays emitted from the low-power X-ray tube are used for the excitation of the X-ray fluorescence, and the minimum detection limit for Cr is a few nanograms.

9.5 Portable Lab-on-a-Chip Analysers The field of chip-based separation microsystems continues its rapid growth. Today, there is an urgent need to develop compatible detection modes. Much of the work on capillary electrophoresis (CE) microchips uses laser-induced fluorescence (LIF) detection. However, LIF detection requires a large, expensive off-chip supporting

Portability of Miniaturized Analytical Systems

353

optical system that greatly compromises the benefits of miniaturization and portability. The technology of portable optical analysers has still not been applied to microchips. Electrochemical detection has recently attracted considerable interest for miniaturized analytical systems [40,41]. Electrochemistry (EC) holds great promise for such microsystems, as it offers high sensitivity (approaching that of fluorescence); inherent miniaturization of both the detector and control instrumentation; independence of optical path length or sample turbidity; low cost; low power requirements;

Figure 9.6 Schematic diagram of the microchip CE chip and instrument. (a) The microfabricated chip consists of four folded separation channels 21.4 cm in length, each with a 0.6 cm long, 70 mm wide injection channel placed 0.6 cm from the anode reservoir. The crosshatch marks are included in the chip design for improved bonding. (b) This microchip is placed against the planar face of a composite objective through which the confocal excitation and fluorescence detection is performed. Electrophoresis power supplies, a thermoelectric cooler for temperature control and pneumatic valve actuators are also contained within the instrument. The 11 kg instrument measures 25.4  30.48  10.16 cm. (From [44] with permission of ACS, Copyright 2007)

354

Miniaturization of Analytical Systems

and high compatibility with advanced micromachining and microfabrication technologies. Related to microchips, portable CE equipment has also been described [42,43]. In such systems, the coupling of optical sensing is easier. For example, Haddad and coworkers [44] have used a commercial portable CE instrument to separate inorganic anions and cations found in postblast residues from improvised explosive devices. The CE instrument was modified for use with a miniaturized LED detector to enable sensitive indirect photometric detection. The system was useful for the determination of 15 anions and 12 cations. A portable microchip CE system (Figure 9.6) has also been used for the determination of neurologically active biogenic amines in fermented beverages. The target molecules are labelled on their primary amino groups with fluorescamine in a 10 minute reaction, and the samples analysed directly, producing a detailed electropherogram in only 120 seconds on a microfabricated glass CE device containing 21.4 cm long separation channels [44].

References [1] S. Terry, J. Jerman, J. Angell, IEEE Trans. El. Dev. ED, 26 (1979) 1878. [2] J.A. Dziuban, J. Mr
oz, M. Szczygielska, M. Małachowski, A. G
orecka-Drzazga, R. Walczak, W. Buła, D. Zalewski, Ł. Nieradko, J. Łysko, J. Koszur, P. Kowalski, Sens. Actuators A, 115 (2004) 318. [3] R.H. Lambert, J.A. Owens, Field Anal. Chem. Technol., 1 (1997) 367. [4] R.E. Chiriac, R. Lornage, L. Fine, J. Carre, J.L. Gass, T. Lagier, Int. J. Environ. Anal. Chem., 87 (2007) 43. [5] C. Lu, J. Whiting, R.D. Sacks, E.T. Zellers, Anal. Chem., 75 (2003) 1400. [6] J. Ji, C. Deng, W. Shen, X. Zhang, Talanta, 69 (2006) 894. [7] S.I. Ohira, K. Toda, Anal. Chim. Acta, 619 (2008) 143. [8] C. Lu, W.H. Steinecker, W. Tian, M.C. Oborny, J.M. Nichols, M. Agah, J.A. Potkay, H.K.L. Chan, J. Driscoll, R.D. Sacks, K.D. Wise, S.W. Pang, E.T. Zellers, Lab Chip, 5 (2005) 1123. [9] Q. Zhong, R.A. Veeneman, W.H. Steinecker, C. Jia, S.A. Batterman, E.T. Zellers, J. Environ. Monit., 9 (2007) 440. [10] M. Agah, G.R. Lambertus, R. Sacks, K. Wise, J. Microelectromech. Syst., 15 (2006) 1371. [11] G. Reidy, J. Lambertus, R. Reece, Sacks Anal. Chem., 78 (2006) 2623. [12] S.C. Terry, J.H. Jerman, J.B. Angell, IEEE T. Electron. Dev., 26 (1979) 1880. [13] P.R. Lewis, R.P. Manginell, D.R. Adkins, R.J. Kottenstette, D.R. Wheeler, S.S. Sokolowski, D.E. Trudell, J.E. Bynes, M. Okandan, J.M. Bauer, R.G. Manley, G.C. Frye-Mason, IEEE Sens. J., 6 (2006) 784. [14] E.R. Badman, R.G. Cooks, J. Mass Spectrom., 35 (2000) 659. [15] M.J. Madsen, W.K. Hensinger, D. Stick, J.A. Rabchuk, C. Monroe, Appl. Phys. B, 78 (2004) 639. [16] M.G. Blain, L.S. Riter, D. Cruz, D.E. Austin, G. Wu, W.R. Plass, R.G. Cooks, Int. J. Mass Spectrom., 236 (2004) 91. [17] S. Pau, C.S. Pai, Y.L. Low, J. Moxom, P.T.A. Reilly, W.B. Whitten, J.M. Ramsey, Phys. Rev. Lett., 96 (2006) 120801. [18] J. Hauschild, E. Wapelhorst, J. M€uller, Int. J. Mass Spectrom., 264 (2007) 53. [19] H.J. Yoon, J.H. Kim, E.S. Choi, S.S. Yang, K.W. Jung, Sens. Actuators A, 97–98 (2002) 441. [20] L. Gao, Q. Song, G.E. Patterson, R.G. Cooks, Z. Ouyang, Anal. Chem., 78 (2006) 5994.

Portability of Miniaturized Analytical Systems

355

[21] G.E. Patterson, A.J. Guymon, L.S. Riter, M. Everly, J. Griep-Raming, B.C. Laughlin, Z. Ouyang, R.G. Cooks, Anal. Chem., 74 (2002) 6145. [22] J. Moxom, P.T.A. Reilly, W.B. Whitten, J.M. Ramsey, Rapid Commun. Mass Spectrom., 16 (2002) 755. [23] R.J. Cotter, T.J. Cornish, J. Mass Spectrom., 34 (1999) 1368. [24] D. English, R.J. Cotter, J. Mass Spectrom., 38 (2003) 296. [25] E.C. Horning, M.G. Horning, D.I. Carroll, I. Dzidic, R.N. Stillwell, Anal. Chem., 45 (1973) 936. [26] L.S. Riter, B.C. Laughlin, E. Nikolaev, R.G. Cooks, Rapid Commun. Mass Spectrom., 16 (2002) 2370. [27] L.S. Riter, Y. Peng, R.J. Noll, G.E. Patterson, T. Aggerholm, R.G. Cooks, Anal. Chem., 74 (2002) 6154. [28] E.C. Meurer, D.M. Tomazela, R.C. Silva, F. Augusto, M.N. Eberlin, Anal. Chem., 74 (2002) 5688. [29] L.S. Riter, E.C. Meurer, I. Cotte-Rodriguez, M.N. Eberlin, R.G. Cooks, Analyst, 128 (2003) 1119. [30] I. Cotte-Rodriquez, E. Handberg, R.J. Noll, D.A. Kilgour, R.G. Cooks, Analyst, 130 (2005) 679. [31] J. Sunner, G. Nicol, P. Kebarle, Anal. Chem., 60 (1988) 1300. [32] W.M.A. Niessen, J. Chromatogr. A, 794 (1998) 407. [33] J. Abian, J. Mass Spectrom., 34 (1999) 157. [34] C.J. Proctor, J.F.J. Todd, Org. Mass Spectrom., 18 (1983) 509. [35] M. Tierney, H. Kim, M. Burns, J. Tamada, R. Potts, Electroanalysis, 12 (2000) 666. [36] J. Albers, T. Grunwald, E. Nebling, G. Piechotta, R. Hintsche, Anal Bioanal Chem, 377 (2003) 521. [37] E. Mauriz, A. Calle, A. Montoya, L.M. Lechuga, Talanta, 69 (2006) 359. [38] C. Cucci, A.G. Mignani, C. Dall’Asta, R. Pela, A. Dossena, Sensors and Actuators B, 126 (2007) 467. [39] S. Kunimura, J. Kawai, Anal. Chem., 79 (2007) 2593. [40] A. Woolley, K. Lao, A. Glazer, R. Mathies, Anal. Chem., 70 (1998) 684. [41] J. Wang, B. Tian, E. Sahlin, Anal. Chem., 71 (1999) 5436. [42] L.L. Yuan, H.P. Wei, H.T. Feng, S.F.Y. Li, Anal. Bioanal. Chem., 385 (2006) 1575. [43] T. Kappes, B. Galliker, M.A. Schwarz, P.C. Hauser, Trends Anal. Chem, 20 (2001) 133. [44] J.P. Hutchinson, C.J. Evenhuis, C. Johns, A.A. Kazarian, M.C. Breadmore, M. Macka, E.F. Hilder, R.M. Guijt, G.W. Dicinoski, P.R. Haddad, Anal. Chem., 79 (2007) 7005.

10 Analytical Performance of Miniaturized Analytical Systems 10.1

Introduction

The need for a consistent approach to analytical quality management has been recognized for both professional and regulatory bodies. As such, organizations such as the ISO (International Organization for Standardization), AOAC International (The Association of Analytical Communities), VAM (Valid Analytical Measurement), the ILAC (International Laboratory Accreditation Cooperation) and so on have published guidance on best practices. ISO has provided a large number of documentation and standardization practices. However, none of these has been adapted and specifically applied to miniaturized analytical systems. There are a number of parameters used to characterize the performance and quality of analytical methods. These parameters must also be applied to miniaturized analytical systems. In this chapter, the concepts behind and a framework for quality assurance are introduced and adapted to miniaturized analytical systems. From an analytical point of view, the core standard is ISO/IEC 17025 on ‘General Requirements for the Competence of Testing and Calibration Laboratories’. According to this guide, the quality of an analytical result depends on many factors. However, it must be remembered that the objective of an analytical process is to provide a result that is fit for its intended purpose. Because all measurements are subject to error, to meet the fitness-for-purpose requirement we need to establish measurement confidence. As will be pointed out below, in the case of miniaturized systems both sample preparation and sample representation are key aspects in assuring the quality of results. Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet Ó 2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

358

10.2

Miniaturization of Analytical Systems

Quality Control in Miniaturized Systems

Analytical performance can be defined as the overall control framework for a process. Therefore, two parts can be distinguished: quality assurance, which corresponds to that part of the analytical performance related to control and documentation of the process; and quality control, which is concerned with the analytical measurement process. The main objective of quality control is to assure the quality of calibration, standardization and equipment qualification (Figure 10.1). In this context, it is important to distinguish between validation and qualification. Qualification is a term used to refer to the proper functioning of an instrument or piece of equipment. Obviously, this process includes calibration of the instrument. In the case of miniaturized analytical systems, quality control can imply the calibration and verification of important and critical steps such as the determination of reproducibility of sample injection. The importance of establishing a welldefined quality control can be itemized as follows: .

. .

Miniaturization is already a large sector of industry, and it is expected to continue to grow in coming years. Precise control of the dimensions of objects in a key issue. Metrology is well established in research institutes and industry. However, implementation of metrology and quality assurance in miniaturized devices still faces a number of problems. The main differences can be identified in precision, cost and efficiency, and of course the conditions under which measurement takes KNOWLEDGE INFORMATION

GOOD DATA

QUALITY CONTROL

MINIATURIZED SYSTEMS TRADITIONAL SYSTEMS

STANDARIZATION

EQUIPMENT QUALIFICATION

RELATIVE IMPORTANCE

CALIBRATION

Figure 10.1 Components of quality control and its relation with the data-to-information chain. The relative importance of this parameter in traditional versus miniaturized systems

Analytical Performance of Miniaturized Analytical Systems

359

place. This aspect is even more critical when the measurement is performed on the nano scale or in a micro region. From an industrial point of view, a critical parameter is one that is important to the device function of the miniaturized system. Methods of measurement developed for conventional materials in many cases cannot be directly applied to micro scale. For such cases, special protocols need to be developed. Over the last few years, we have seen how interest in ensuring the quality of results has increased and how metrology has gradually developed into what it is today. Examples are the use of uncertainty calculation rather than precision studies and the ever-growing importance of determining the robustness of an analytical method. In order to approach an assurance of the quality of measurement-based analytical methods in miniaturized equipment or instruments, we must first ask ourselves, ‘What is the difference?’ and ‘Do miniaturized systems really need special treatment?’ Let us consider a simple chromatographic procedure. There is no doubt about the precision of today’s equipment. It has both a high level of precision in the injection volume and a high level of precision in migration times. However, if we carry out the same process in miniaturized equipment, questions arise such as, ‘Can injection volumes really be reproduced?’ and ‘What degree of reproducibility would we have in the creation of different miniaturized columns?’ We are accustomed to accepting the reproducibility of the chromatographic columns supplied by commercial companies, but can we do so in the case of miniaturized systems? We believe, in this respect, that the technology is still in development and not yet fully implemented and that we must therefore be stricter in our considerations and look into the repercussions these changes might have on our results. Without a doubt, the quality of results is something we must ensure at all times. As a general rule, we can state that small uncontrolled changes in miniaturized instruments can have a very great influence on results and that the controls we carry out must therefore be stricter. We can take an example from the case of capillary electrophoresis. Imagine we use a 25 mm diameter capillary and we have a tolerance of 2 mm. This would mean that the variation in the sample volume introduced in the capillary would be from 0.39 to 0.75 nl; a difference of 200%. It is therefore evident that correct calibration is of key importance when it comes to changing to a different capillary or, indeed, during the use of the same capillary. Small obstructions may easily arise in the capillary, due for instance to the uncontrolled evaporation of buffer within it, which will have a direct effect on the sample volume injected. Once again, the importance of our affirmation that stricter controls must be established becomes clear. This consideration of the dimensions of the capillary can also be applied to other aspects such as sample injection pressure, injection time and do on. Instrument graduation is therefore also very to take into account.

360

10.3

Miniaturization of Analytical Systems

Validation of Microsystems

In this section, we will look at the parameters that are going to characterize the quality of methods and the critical points that may have greater effects in the case of miniaturized systems. Detection limit is a very important quality parameter. We define this as the minimum quantity of analyte we are able to detect for a determined confidence interval or probability. In relation to this parameter, we also define the quantification limit, or the minimum amount we can quantify. In other words, the minimum quantity we are able to determine, with its corresponding confidence interval. On many occasions, these parameters are determined from idealized situations, giving values which are very different from reality. The detection limit is clearly a situation in which we cannot find ourselves in a real case, whereas the quantification limit should be a real quantity that we are able to determine. Many miniaturized systems, such as sensors, do not offer a linear answer regarding analyte concentration, and it is advisable to determine this parameter experimentally. For this purpose, samples with ever-decreasing concentrations of the analyte of interest must be analysed. It is important to carry out this analysis with the sample, given that the matrix may have a very considerable effect. In the case of miniaturized systems, we are faced with a dilemma: whether to express these limits as quantities of analyte or as concentrations. Generally speaking, detection limits expressed as concentrations are high, such that many authors prefer to express them as quantities of analyte. Analysed volumes being very small, generally of the order of nanolitres, very small quantities are expressed. Both forms are correct, but we must bear in mind that, on making the transformation with the analysed volume – for example, the volume injected into a microchip – the error (precision) of this volume must be added. Another parameter is precision. With precision, our intention is to assess the variability of results due to random errors. It is generally expressed as the standard deviation found in the results of a sample analysed on several occasions. It is important for this parameter to include all different sources of variability. In traditional methods, it is accepted that precision can be determined from five or six analyses of one sample. However, in the case of miniaturized systems, are six measurements enough to take into account all different sources of error? We must bear in mind that determinations are going to be made on very small quantities of sample. In some cases therefore, the number of determinations carried out must be increased in order to assure the representativeness of samples. In our opinion, robustness is the parameter that should be determined and assessed with the greatest care in the case of miniaturized systems. The control and precision of variables is critical. When we talk about variables, we are referring to both chemical variables and to physical variables of the instrument, such as pressure control, temperature control and so on.

Analytical Performance of Miniaturized Analytical Systems

USER REQUIREMENTS SPECIFICATIONS WHAT ARE THE PERFORMANCE MEASUREMENT AND ACCEPTANCE CRITERIA?

METHOD PERFORMANCE CRITERIA

WHICH VARIABILITY IS ACCEPTABLE?

IS RELATED TO

ARE RELATED TO

FITNESS FOR PURPOSE

PRELIMINARY VALIDATION

361

IS HARMONIZATON ASSURED?

ARE THERE APPROPRIATE STANDARDS?

ARE RELATED TO

ANALYTICAL RELIABILITY CRITERIA

INITIAL EVALUATION

IS THE ROBUSTNESS TRIAL WELL-DEFINED? ARE ALL THE VARIABLES?

METHOD DEVELOPMENT PROCESS

Figure 10.2 Analytical method development and validation process

Other characteristic parameters, typically involved in analytical calibration, require a similar treatment to the traditional approach. Nowadays, not only is linear regression used, but, thanks to chemometric tools, it is possible to approach the quantification and detection of analytes from a wide variety of angles. It must be kept in mind that validation is not a single event. Figure 10.2 represents the relationship between method development and the validation process.

10.4

Qualification of Microsystems

A key aspect in assuring the quality of analytical measurements is traceability. Traceability is the demonstration of the uninterrupted chain of the system, which relates the measurement to a primary standard whose value is known and accepted. In the case of the qualification of miniaturized instruments, this generally implies demonstrating their traceability to fundamental constants (for example spectral lines, time, length, etc.) or to a certified reference material. Imagine we wish to demonstrate the traceability of a balance, for instance. We can do so relatively easily be using certified weights. But now let us consider the case of a miniaturized instrument such as a cantilever. Although we are measuring mass, we cannot demonstrate and assure traceability so easily. Remember that ISO 9001 requires that ‘equipment be calibrated or verified at specified intervals or prior to use, against measurement standards traceable to international or national standards; where no such standards exist, the basis for calibration or verification shall be recorded’. More specifically, ISO/IEC 17025 requires that ‘equipment and its software used for testing, calibrating, and sampling shall be capable of achieving the

362

Miniaturization of Analytical Systems

accuracy required and shall comply with specifications relevant to the tests and/or calibrations concerned. Before being placed into service, equipment shall be calibrated or checked to establish that it meets the laboratory’s specification requirements and complies with relevant standard specifications. Note that these are usually not the same as the manufacturer’s specifications.’ The qualification of miniaturized instruments is a little-developed area. This is doubtless because of the fact that, despite its rapid advance, it has not yet been fully introduced into routine laboratories. Qualification should define at least the following aspects: (i) Technical specifications such as means of calibration, means of operation, maintenance operations, normal values of precision, accuracy, etc. (ii) Environmental specifications, such as work temperature intervals. (iii) Sample specifications; specifically, specifications, limitations and characteristics of the sample to be determined. Sampling indications may sometimes also be included. (iv) Data transformation and treatment specifications.

10.5

Robustness of Microsystems

There can be no doubt about the technological advances that have taken place over the last few years. Without these advances, it would not be possible to talk about miniaturization today. These technological advances do not only open new doors and create new possibilities for obtaining analytical information; they also give rise to new unknown factors. Without doubt, once of the most important of these factors is robustness. Robustness is the capacity of a method to provide the same analytical result despite the small variations arising in the analytical parameters which affect the measuring process. The critical values of said parameters are generally determined experimentally. For practical purposes, those parameters which the analyst can control or change with relative ease are the ones chosen as analytical parameters; the pH of the sample, for instance. In many cases, other parameters which are sometimes even more important, such as all the variables of the instrument, the time or the operator themself, are not taken into account. If we do take these variables into account, we are determining ‘ruggedness’. This term is much more important and decisive from a practical viewpoint if we are thinking of applying a particular procedure in different routine laboratories. We must not forget that the aim at all times is to achieve harmonization or comparability between laboratories. In order to assure ruggedness, it is essential to approach interlaboratory studies systematically. In some miniaturized systems, limitations arise when there is no

Analytical Performance of Miniaturized Analytical Systems

363

possibility of determining the analytical parameter by any other method. Imagine a miniaturized system such as a microelectrode capable of effecting a direct determination in the cytoplasm of a cell. In this example, it would not only be difficult to find another analytical technique, it would be difficult even to assure the homogeneity and reproducibility of the samples. Clearly, there are many degrees of miniaturization and we have chosen an extreme case in this example, our intention being simply to emphasise the added difficulties which have to be taken into consideration.

Further Reading [1] C. Burgess, Valid Analytical Methods and Procedures, The Royal Society of Chemistry, Cambridge, 2000. [2] B. Neidhart, W. Wegscheider (Eds), Quality in Chemical Measurements, Springer-Verlag, Berlin, 2001. [3] M. Parkany (Ed.), Quality Assurance and TQM for Analytical laboratories, The Royal Society of Chemistry, Cambridge, 1995. [4] F.E. Pritchard (Ed.), Quality in the Analytical Chemistry Laboratory, ACOL, John Wiley & Sons, Ltd, 1995. [5] AOAC International, Official Methods of Analysis of AOAC International, 17 Ed., 2002. [6] ISO 9001: Quality Management Systems: Requirements, Section 7.6(a), 2001. [7] ISO/IEC 17025: General Requirements for Competence of Testing and Calibration Laboratories, Section 5.5.2, 1999.

Index Page numbers in italics refer to figures; page numbers in bold refer to tables. The prefix “micro-” has frequently been omitted for clarity in the index; e.g. for “microextraction”, search under “extraction”. absorbance (UV) detection 248–9 Agilent Technologies, commercial products 150–5, 151, 161, 214 amperometry 65, 230–1 analysis see also TAS (total analysis system) approach client information needs 2–3, 3, 7–8, 119–20, 361 integration of operations 26, 30–3, 31, 302–3 process stages 3, 3–4, 19, 19–20 quality management 357–63 trends in technological demands 4–5, 5, 30 applications equipment, examples of 32, 32–3, 34 potential analytical fields 1, 87–8, 141 technological challenges 33–6, 302–5 arrays capillary electrophoresis 166, 192–3, 196 microcantilever 273–4 at-line coupling 120, 124, 128, 129 atomic force microscopy (AFM) 75 automation 2, 5 sampling 103–4, 106, 106, 119–20 Bernoulli’s equation 42 biomedical analysis see clinical sample analysis Bodenstein number 290 Boltzmann constant 44 bubble-based micropumps 58

bulk nanomachining and wafer bonding 70, 71 buried channel technology 71–2, 72 calibration 33, 358, 359 in sample extraction 94–5, 95, 119 cantilevers as passive valves 47, 47 in sensor devices 78, 273–4 assembly 75–81 detection methods 81–4, 273, 274 capacitive detection 83–4 capillaries see channels, micro(/nano)fluidic capillary-driven test strips 294, 296 capillary electrophoresis (CE) analytical applications 189–206, 207, 312, 317–18 CGE (capillary gel electrophoresis) 170–1, 171, 191–4, 197, 308 CIEF (capillary isoelectric focusing) 171, 171, 197 CZE (capillary zone electrophoresis) 170, 170, 188 interfacing compatibility 127–35, 131, 131 isotachophoresis (ITP) 171–2 microchip design and fabrication 168, 168–9, 175–81, 196, 298 miniaturization to microchip scale 23–4, 140–1, 166, 167, 299–300 capillary liquid chromatography (CLC) 140, 144–50

Miniaturization of Analytical Systems: Principles, Designs and Applications and Bartolome Simonet
2009 John Wiley & Sons, Ltd

Angel Rios, Alberto Escarpa

366

Index

carbon nanotubes (CNTs) 229–30 as detectors 336, 336–8, 337 as electrochemical transducers 264–6, 265, 270–2 as GC stationary phase 142 integration onto electromechanical systems 78, 81 casting, of PDMS microchips 180, 180 CEC (capillary electrokinetic chromatography) 140, 172, 184–7 cells (biological), micromanipulation 84–6, 306 centrifugal microfluidic devices 297–9 CGE (capillary gel electrophoresis) 170–1, 171, 191–4, 197, 308 channels, micro(/nano)fluidic closure of (bonding) 73, 178, 180, 181 dimensions influence on fluid flow 41–2, 59–60, 66, 285–6 in separation techniques 175, 185–6 droplet manipulation 300–1 laminar flow microfabrication 289, 289–90 surface properties 67, 176, 308 transport parameters 42–5 chemiluminescence detection 219–20 chromatography electrokinetic (LC/CE hybrid) modes 140, 172, 184–7 miniaturization 24–5, 139–41 use of preconcentration columns 146, 147 CIEF (capillary isoelectric focusing) 171, 171, 197 cleanroom facilities 40, 68, 189 clinical sample analysis drug metabolism 112–13, 134–5, 157–8, 159 monitoring and point-of-care equipment 32, 296, 310, 351 real samples and diagnosis 87, 87, 310–23, 313–16 in vivo nanoprobes 275–7 coated fibres, in microextraction 97–8, 98, 101, 101–2 computer simulations for design 66–8 conductivity detection 21, 65, 231–4 continuity equation (flow in channels) 41 continuous flow systems 119–35

Coulomb forces 56 equation 43, 172–3 coupling 120 at-line 124, 128, 129 in-line 132, 134 online 120–3, 130, 133 CZE (capillary zone electrophoresis) 170, 188

170,

decouplers 225–8 derivatization 28, 109, 215 design development steps 40, 40, 360–1, 361 electrochemical detectors, configuration options 220–1, 221 exploiting microfluidic properties 286–90 microchip layouts 168, 168–9 modification of commercial devices 104–6 optical detection systems 216, 216–17 prototype development 142–4, 143, 144 tools 66–8 detection techniques 20–3, 64–5, 81–4 alternatives, for microsystems 213–14, 254, 281, 304 chemiluminescence 219–20 electrochemical (ED) 20–2, 22, 65, 220–34 fluorescence lamp-based 218 laser-induced (LIF) 20, 214–18, 215, 216 LED-induced 218–19 infrared (IR) 253 mass spectrometry (MS) 234–47 nuclear magnetic resonance (NMR) 254 surface-enhanced Raman spectroscopy (SERS) 253, 269 surface plasmon resonance (SPR) 249–51, 250, 322–3, 351–2 thermal lens microscopy (TLM) 22–3, 251–3, 252 ultraviolet (UV absorbance) 248–9 detectors, compatible with CLC 146–8, 148 devices, microfluidic see also lab-on-a-chip devices characteristics related to size 15, 15–16, 290–3 detectors 64–5 dispensers 63–4 mixers 59–63, 287–8

Index

367

pumps 51–9 valves 45–51 diagnostic tools (medical) 87, 87, 310–23 diaphragm micropumps 52–4, 53 diffusion 16, 17 Fick’s laws (equations) 43–4 and fluid mixing 18–19, 59–60, 287–8 molecular motion parameters 43, 285 Stokes-Einstein equation 44 diffusion-controlled systems 17, 293 dispensers 63–4 dispersion 45 DNA analysis 190–1 forensics 194–5 genome sequencing 86, 86, 195–6, 271 genotyping 191–4 integrated, on microchip 305–10, 311 downscaling theory 15–17, 290–3, 291, 292 droplets in microextraction 114–15 used in microchip devices 300–2, 319

electrospray ionization (ESI) sources 131, 145–6, 234–6, 237 capillary sprayers 239, 239–41 direct, from chip 236–9, 238 integrated emitters 241, 242 electrowetting 58, 62, 300–1 ELISA (enzyme-linked immunosorbent assay) 32, 198–200, 199, 322, 335 end-channel detection 221–3, 222 environmental analysis 323–8, 324–5 atmospheric monitoring 142–4, 144, 346–8 toxic pollutants and residues 110–11, 123, 202–3, 203 etching, glass 49, 177–8 extraction 27–8 calibration 94–5, 95 choice of technique 117–19 liquid-phase (LPME) 114–17 solid-phase (SPE/SPME) 95–113, 118, 306

ease of use see simplification Einstein-Smoluchowski equation 43, 285 electro-osmotic flow (EOF) 56–8, 57, 127, 173, 299 compared with hydrodynamic 187–8, 188, 189 electrochemical detection (ED) 20–2, 22 amperometric 65, 230–1 channel/electrode configurations 220–8, 221 conductimetric 21, 65, 231–4 electrode materials 228–30 electrochemical sensors 270–2, 271, 272, 351, 351 electrodes for AC-EOF pumping 58 materials 220, 228–30 nanostructures, in electrochemical sensors 270–2, 272 electrohydrodynamic pumping 56 electrokinetic devices injection 128–31, 132, 181–4, 182 microfluidic chip 298, 299–300 electrophoresis see also capillary electrophoresis (CE) miniaturization stages 140–1 theoretical basis 42–3, 172–5

fabrication 6, 22, 24–5 CE chip assembly process glass 177–8, 178, 227 polymer 178–81, 180 facilities 66 integration (packaging) 74 layer-by-layer deposition 49 LC chip assembly 155–7 materials 69, 73–4, 75–6, 175–6, 177 within microchannels 289, 289–90 micromachining 68–74, 151 optimization (simulation) 66–8 filtration 26–7, 312 flow characterization 15–16, 41, 187–8, 188, 189 flow-injection analysis (FIA) systems 45, 293 fluid properties 67 fluidic resistance 41 fluidic manipulation 6, 36 see also devices, microfluidic droplets 300–2 flow focusing 287, 288–9 mixing, in microsystems 18–19, 59–63, 287–8

368

Index

fluorescence 20, 28, 216–17 see also detection techniques, fluorescence of quantum dots 266–8, 267 food analysis 111–12, 203–4, 205, 328–38, 329–32 forensic sample analysis 194–5 Fourier number 290 frictional constant 44 functionalization of sensors of carbon nanotubes (CNTs) 78, 81, 270–2, 271 of quantum dot nanoprobes 276, 276–7 self-assembled monolayers (SAMs) 76–8, 78, 80, 274 sol-gel materials 78–80 gas analysers, portable 33, 34, 346–50, 347 gas chromatography (GC) coupling with flow systems 120–3 microextraction coupling 104, 106, 116, 117 miniaturization 141–4, 346–9, 348 gas sensor, nanoparticles 264, 265 gated injection 182, 183 gel electrophoresis, capillary (CGE) 170–1, 171, 191–4, 197, 308 genome sequencing 86, 86, 195–6, 271 genotype scanning strategies 192–3 glass as chip material 175–6, 177 microchip fabrication method 177–8, 178 as substrate 73 gold microspheres/nanoparticles in chromatography 149, 185 nanoparticles, in biochemical sensors 268–9, 269 substrate for cantilever monolayers 76–7, 274 headspace sample extraction 97–8, 98, 110 health care equipment 32, 296, 310, 351 heating effects 67, 175 history of miniaturization 8, 8–10, 140, 166, 283 hot embossing 179

HPLC (high-performance liquid chromatography) capillary (CLC) 140, 144–50 chip-based 140–1, 150–63 hydrodynamic flow compared with EOF 187–8, 188, 189 principles 40–5 imaging, in vivo 277, 277 immunoassays 320–2 in-channel detection 224, 224–5 in-line coupling 120, 132, 134 in-tube SPME 98–101, 99, 100 in vivo bioanalysis 134–5, 275–7 infrared (IR) detection 253 injection moulding 179 injectors design 105, 106 electrokinetic 128–31, 132, 181–4, 182 for GC instruments 121–2 hydrodynamic 132 for LC instruments 124–5, 125 integration 26, 30–3, 31 see also coupling; TAS (total analysis system) approach interfaces 29, 64, 128–32, 130 on microchip detectors 224–5, 227 DNA analysis 305–10, 311 sample preparation 206 packaging 74 in portable equipment 346–54 sample extraction 100, 100–1, 103–6, 116–17, 117 interfacial forces 42 surface tension 47, 58 interferometric detection 84 ISO standards 2, 357, 361–2 isotachophoresis (ITP) 171–2, 306 lab-on-a-chip devices 226, 227, 282, 293–4, 295 capillary-driven test strips 294, 296 centrifugal 297–9 droplet-based 300–2, 319 electrokinetic 298, 299–300 large-scale integration (LSI) 296–7, 298 in portable instruments 352–4, 353 laminar flow 15–16, 41, 286–7

Index lamp-based fluorescence detection 218 large-scale integration (LSI) platforms 296–7, 298 large-volume injection (LVI) techniques 113, 120–1 laser ablation (photoablation) 179 laser-induced fluorescence (LIF) detection 20, 214–18, 215, 216 lateral flow assays 294, 296 light-emitting diodes (LEDs) 218–19 limits of detection (LODs) 274, 360 lipids analysis 146, 148, 148–9, 149 liquid chromatography (LC) see also HPLC (high-performance liquid chromatography) coupled with MS 145–6, 149, 150 coupling with flow systems 123–7, 124 miniaturization stages 139–41, 140, 141 liquid phase microextraction (LPME) 118–19 droplet-based (SDME and DLLME) 114–15 membrane-assisted (MMLLE and SLM) 115–17, 116, 123 lithography 69, 70, 177, 297 m TAS see lab-on-a-chip devices; TAS (total analysis system) approach machining 68–74 deposition 69 etching 49, 69–70, 177–8 laser ablation 151, 179 lithography 69, 70, 177 macromolecular characterization applications 159–63, 162 magnetic sensors 272–3 MALDI (matrix-assisted laser desorption ionization) targets 243–5 integration with microfluidic chips 247 microfabrication 245 off-chip preparation 246, 246–7 mass sensors 273, 273–4 mass spectrometry (MS) high-throughput analysis 241, 243, 243 interfacing with microfluidic devices 23, 145–6, 149, 150

369

electrospray ionization (ESI) 131, 131, 234–43 matrix-assisted laser desorption ionization (MALDI) 243–7 portable equipment 349–50 materials, for microfluidic devices 282–3 MEKC (micellar electrokinetic chromatography) 140, 171, 172 MEMS/NEMS (micro/nanoelectromechanical systems) 65, 264 applications 74–5, 142 design tools 67–8 detection methods 81–4, 273, 273–4 functional assembly 76–81 structural fabrication 75–6 MEPS see packed syringe microextraction (MEPS) microchip devices see also integration, on microchip; lab-on-a-chip devices CE 168, 175–81, 196, 199 GC 346–9, 349 HPLC 140–1, 150–63, 155, 156 microfluidics behaviour principles 39, 40–5, 285–6 definition 18, 282 devices 14, 45–65, 293–302, 296 transport 284–5 microporous membrane microextraction (MMLLE) 115–17, 116, 123 microtechnology 12, 12 design tools 66–8 devices 45–65 fabrication 68–74, 75–81 migration (electric field directed molecule transport) 42–3, 172, 284 miniaturized systems advantages of small scale 1, 17–18, 18, 146, 167 classification criteria 11 chemical objectives 13 complexity 14–15 process steps 13–14, 19–20 size classes 10–13, 12, 141 design steps 40, 40, 360–1, 361 history of miniaturization 8, 8–10, 140, 166, 283 scaling factors 16–17, 17, 292, 292–3 technological challenges 6–8, 33–6, 213

370

Index

minisystems, definition 10–12, 12 mixing 59, 60 active 62–3 passive 59–61, 61, 287–8 monolithic column (continuous rod) electrochromatography 186, 187 moulding methods, polymer microchips 178–9, 180, 180 multidimensional (2D) protein electrophoresis 197, 198 nanoimprint lithography 72, 72–3 nanotechnology see also carbon nanotubes (CNTs) biomedical QD nanoprobes 275, 275–7 definition 12, 12–13 fabrication methods 70–3, 289–90 nanoparticles, colloidal 266–9 nanostructures 229–30, 270–2, 271 sensor components 264–6 no-pinched injection 181–2, 182, 183 nuclear magnetic resonance (NMR) 254 off-channel detection 225–8 online coupling 120–3, 130, 133 open-channel electrochromatography 185–6 optical detection 64–5, 82–3, 214, 274 optical fibres, use in detectors 217, 219, 248 optical sensors 266–9, 351–2, 352 optimization by computer simulation 66–8 by recovery/measurement 109, 252 packaging, of microsystems 74 packed-channel electrochromatography 185–7, 186 packed syringe microextraction (MEPS) 113, 118 PCR (polymerase chain reaction) 86, 192–4, 305–8, 307, 309 PDMS (polydimethylsiloxane) 24, 180, 283, 296–7 Peclet number 44–5, 286, 290 peptides analysis 145–6, 147, 149, 199–200 performance error estimation 68, 360 micro-CE injector precision 183–4 testing framework 357–8 theoretical, for separation techniques 174–5

peristaltic micropumps 53, 53–4 pesticides (PAH/PCB) analysis 110, 123 pharmaceutical applications 157, 161–3 photolithography 69, 70, 177 piezoresistive detection 83, 274 pinched injection 182, 182–3 point-of-care testing 32, 296, 310, 351 polymerase chain reaction see PCR (polymerase chain reaction) polymers see also PDMS (polydimethylsiloxane) aqueous flow control 48, 48–9 as chip material 176, 177 microchip fabrication methods 178–81, 180 in microextraction 102 as microsystem substrate 73–4, 151 portable systems 5, 323 benefits of portability 345–6 electrochemical biosensors 351, 351 gas analysers 33, 34, 141–4, 346–50, 347 microchip devices 352–4, 353 optical sensors 351–2, 352 protein analysis 196–200 proteomics 149, 150, 158–9, 320 pumps 51, 52, 144–5, 145 mechanical 51–4, 53, 55, 141–2 nonmechanical 54, 56–9 qualification, of miniaturized instruments 358, 361–2 quality control, analytical 358, 358–9 quantum dots (QDs) 266–8, 267, 275, 275–7 Raman detection 253, 269 reliability (precision) 68, 360 see also performance replica moulding see casting, of PDMS microchips Reynolds number (equation) 16, 41, 286 robustness, analytical 360, 362–3 rotary micropumps 53, 54 ruggedness, interlaboratory 362–3 sample treatment introduction 28–9, 63, 63–4, 119–35 preparation 25–8, 94–119, 310, 333 scaling factors 16–17, 17, 292, 292–3 scanning tunnel microscopy (STM) 75 segmented flow 300–1

Index self-assembled monolayers (SAMs) 76–8, 78 semiconductors nanocrystalline 266–8, 267 nanowire 270, 271 sensors 30, 32, 263–6 cantilever-based 75–84 electrochemical 270–2, 271, 272, 351, 351 magnetic 272–3 mass (microcantilever) 273, 273–4 optical 266–9, 351–2, 352 separation efficiency, electrophoresis 174–5 separation techniques 23–5, 126, 139–41, 140, 165–6 sequential-injection chromatography (SIC) 126, 126 silanes, substrate for SAMs 77–8 silicon, as substrate 69–73 simplification 2, 28, 30, 107 simulations 66–8 single-drop microextraction (SDME) 114–15 small-molecule analysis 200–6 sol-gel glass, in sensors 78–80 solid phase microextraction (SPME) coated fibres 97–8, 98, 101, 101–2 coatings, properties of 101, 101–2 commercial devices 102–7 in-tube 98–101, 99, 100 packed syringes/pipette tips 113, 118 range of techniques 95–6, 96, 97 stir bars 107–13, 108, 118 split-flow interfaces 29, 64, 128–32, 130 stir bar sorptive extraction (SBSE) 107–13, 108, 118 Stokes force (equation) 43, 173 surface acoustic waves (SAWs) 301–2 surface-enhanced Raman spectroscopy (SERS) 253, 269 surface nanomachining 70–1, 71 surface plasmon resonance (SPR) detectors 32, 249–51, 250, 322–3, 351–2 nanosensors 268–9 surface properties, channels 67, 176, 308

371

surface tension in electrowetting pumps 58 as passive valve 47 synthetic chemistry 13 TAS (total analysis system) approach 4, 32–3, 281–2 see also lab-on-a-chip devices micro scale integration (mTAS) 302–5 Taylor dispersion 45 thermal lens microscopy (TLM) 22–3, 251–3, 252 thiol monolayers 76–7, 80 time-constant systems 16, 292–3 transitional flow 41, 286 transport, directed and statistical 42–5, 284–5 turbulent flow 41, 286 uncertainty (errors)

68, 360

validation (of methods) 360–1, 361 valves see also injectors active 49–51, 50 characteristics 45–6, 46 interface with HPLC chip 151–3, 153 passive 47, 47–9, 48 velocity (flow) profile EOF 57, 188, 188 in microfluidic channels 41, 299 volatile organic compounds analysis 111–12, 142, 346–8 voltammetry 230–1 volume dispensers 63–4 volume flow, Q (Hagen-Poiseuille theory equation) 41 wafers, silicon 69–73 water odours analysis 110–11 X-ray fluorescence equipment

32, 352