Microfluidic Devices in Nanotechnology: Applications

MICROFLUIDIC DEVICES IN NANOTECHNOLOGY Applications Edited by CHALLA S. KUMAR

MICROFLUIDIC DEVICES IN NANOTECHNOLOGY

MICROFLUIDIC DEVICES IN NANOTECHNOLOGY Applications Edited by CHALLA S. KUMAR

Copyright Ó 2010 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada 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, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www. wiley.com. Library of Congress Cataloging-in-Publication Data: Microfluidic devices in nanotechnology. Applications / edited by Challa S. Kumar. p. cm. Includes bibliographical references and index. ISBN 978-0-470-59069-0 (cloth) 1. Microfluidic devices. 2. Nanofluids. 3. Nanotechnology. 4. Fluidic devices. I. Kumar, C. S. S. R. (Challa S. S. R.) TJ853.4.M53M5325 2010 2009051009 620.10 06–dc22 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

CONTENTS

Preface

vii

Contributors

xi

1

1

Microfluidics For Nanoneuroscience Pamela G. Gross and Emil P. Kartalov

2

Nanoporous Membrane-Based Microfluidic Biosensors

47

Shalini Prasad, Yamini Yadav, Manish Bothara, Vindhya Kunduru, and Sriram Muthukumar

3

Nanoparticle-Based Microfluidific Biosensors

91

Giovanna Marrazza

4

Microfluidic Enzymatic Reactors Using Nanoparticles

125

Chunhui Deng and Yan Li

5

Microfluidic Devices for Nanodrug Delivery

187

Clement Kleinstreuer and Jie Li

6

Microchip and Capillary Electrophoresis Using Nanoparticles

213

Muhammad J. A. Shiddiky and Yoon-Bo Shim

7

Pillars and Pillar Arrays Integrated in Microfluidic Channels: Fabrication Methods and Applications in Molecular and Cell Biology

255

Jian Shi and Yong Chen

v

CONTENTS

vi

8

Nanocatalysis in Microreactor for Fuels

281

Shihuai Zhao and Debasish Kuila

9

Microfluidic Synthesis of Iron Oxide and Oxyhydroxide Nanoparticles

323

Ali Abou-Hassan, Olivier Sandre, and Val
erie Cabuil

10

Metal Nanoparticle Synthesis in Microreactors

361

Peter Mike G€ unther, Andrea Knauer, and Johann Michael K€ ohler

Index

395

PREFACE

I hope you had an opportunity to go through the first volume. It gives me immense satisfaction in placing the second volume of the two-volume book series— Microfluidic Devices for Nanotechnology: Applications—in your hands. The second volume is the first book ever to be published that covers nanotechnology applications using microfluidics in a broad range of fields, including drug discovery, biosensing, catalysis, electrophoresis, enzymatic reactions, and synthesis of nanomaterials. While the first volume, Microfluidic Devices for Nanotechnology: Fundamental Concepts, in its combined form provides readers an up-to-date knowledge of the fluid and particle kinetics, spatiotemporal control, fluid dynamics, residence time distribution, and nanoparticle focusing within microfluidics, the second volume primarily captures upto-date applications. The book fills in a long-term gap that existed for the real-time measurement of biomolecular binding in biosensors and justification for incorporating nanoporous membranes into “lab-on-a-chip” biosensing devices. Focusing on lab-ona-chip systems for drug delivery (also called bio-MEMS), separating bioanalytes using electrophoresis, genomics, proteomics, and cellomics, the book is a must for biologists and biochemists. Highlighting the importance of nanoneuroscience, the book educates the reader on the discipline of microfluidics to study the nervous system at the single-cell level and decipher physiological processes and responses of cells of neural origin. For a nanomaterials chemist interested in novel approaches for synthesis of nanomaterials, this book is an excellent source of information covering a wide variety of microfluidic-based approaches for synthesis of metallic and nonmetallic nanomaterials. Finally, opening a window for the next-generation alternative energy portable power devices, nanocatalyst development for industrially useful reactions in silicon-based microreactors is discussed especially in the context of syngas conversion to higher alkanes, which could solve current difficulties of storage and transportation vii

viii

PREFACE

by converting natural gas into liquid fuels. Overall, the book contains reviews by world-recognized microfluidic and nanotechnology experts providing strong scaffolding for futuristic applications utilizing synergy between microfluidics and nanotechnology. Chapter 1 by Drs. Pamela G. Gross and Emil P. Kartalov focuses on the application of microfluidic devices to study the nervous system at single-cell level using nanotechnologies. This chapter describes various aspects of microfluidic chips used to decipher physiological processes and responses of cells of neural origin with examples of novel research not previously possible. Continuing on a similar theme, Chapter 2 by Professor Shalini Prasad et al. provides a detailed account of realtime biomolecular sensing through incorporation of nanoporous membranes, manmade as well as natural, into “lab-on-a-chip” biosensing devices. In addition to nanoporous membranes, simple spherical nanoparticles are finding novel applications when incorporated within the microchannels. Chapter 3 by Professor Giovanna Marrazza reviews the most recent applications of nanoparticles within microfluidic channels for electrochemical and optical affinity biosensing, highlighting some of their technical challenges and the new trends. Chapter 4 by Professors Chunhui Deng and Yan Li presents the recent advances in the field of immobilized microfluidic enzymatic reactors (IMERs), which constitutes a new branch of nanotechnology. In view of the increasing use of lab-on-a-chip systems in the healthcare industry, there is a growing demand for discovery, development, and testing of active nanodrug carriers within the microfluidic environment for controlled drug delivery. Chapter 5 by Professor Clement Kleinstreuer and Jie Li provides a comprehensive treatise on fundamentals and applications of microfluidics and bio-MEMS with respect to nanodrug targeting and delivery. Capillary electrophoresis (CE) and microchip electrophoresis (MCE) are two promising separation techniques for analyses of complex samples, in particular, biological samples. Not surprisingly, these techniques have been profoundly influenced by the advances in nanotechnologies. Chapter 6 by Muhammad J. A. Shiddiky and Professor Yoon-Bo Shim covers the recent developments and innovative applications of nanomaterials as stationary and/or pseudostationary phases in CE and MCE. This chapter illustrates the importance of various types of nanomaterials, including metal and metal oxide nanoparticles, carbon nanotubes, silica nanoparticles, and polymeric nanoparticles, in enhancing the separation of biological samples using CE and MCE. The examples we have seen so far involve externally fabricated nanomaterials, which are later on utilized for a number of applications within the microfluidic channels. Chapter 7 by Drs. J. Shi and Yong Chen discusses pillars and pillar arrays integrated into microfluidic chips in the fabrication process itself. This chapter demonstrates how such an approach provides a large variety of functionalities for molecule and cell biology studies. The applications we have seen so far in the first seven chapters range from biology to drug delivery. Chapter 8 by Shihuai Zhao and Professor Debasish Kuila is uniquely placed in the book as it brings out the recent recognition for microreactor as a novel tool for chemistry and chemical process industry, such as fuel industry. This chapter presents silicon-based microreactors for the development of nanocatalysts for

PREFACE

ix

industrially useful reactions. For example, methanol steam reformer to produce H2 and CO purifier is described in detail for potential microreactor applications in the next generation of alternative energy for portable power devices. The last example that the book provides is the application of microfluidic reactors for the synthesis of nanomaterials. With the increase in the demand for high-quality metal nanoparticles with narrow size, shape distribution, and homogeneous composition, the continuous-flow microfluidic processes are gaining attention as they are particularly suited for realizing constant mixing, reaction, and quenching conditions necessary for production of high-quality metallic nanomaterials. Chapter 9 by Dr. Ali Abou-Hassan et al. reviews the recent scientific literature concerning the use of microfluidics for the synthesis of the iron oxides nanomaterials. Chapter 10 by Professor J. Michael K€ ohler and coworkers is a fitting conclusion to the book delineating a number of promising opportunities and challenges for the application of microreaction technology for the synthesis and manipulation of metallic nanoparticles. In combination with the Chapter 9 in Volume 1, this will provide a strong platform from both theoretical and experimental perspectives on synergism between microfluidics and nanotechnology for automated microreactor-based controlled synthesis and engineering of nanomaterials for a number of applications. In conclusion, the two volumes bring out a clear understanding of theoretical and experimental concepts of microfluidics in relation to nanotechnology in addition to providing a seamless transition of knowledge between and micro- and nanofluidics. The contributors for both the volumes are world-renowned experts exploiting the synergy between microfluidics and nanotechnology. I am very much grateful to all of them for sharing my enthusiasm and vision by contributing high-quality reviews, on time, keeping in tune with the original design and theme of both the volumes. You will not be having this book in your hand but for their dedication, perseverance, and sacrifice. I am thankful to my employer, the Center for Advanced Microstructures and Devices (CAMD), who has been supporting me in all my creative ventures. Without this support, it would be impossible to make this venture of such magnitude a reality. No words can express the understanding of my family in allowing me to make my home a second office and bearing with my spending innumerable number of hours in front of the computer. It is impossible to thank everyone individually in this preface; however, I must make a special mention of the support from Wiley in general and the publishing editor Anita Lekhwani in particular, who has been working closely with me to ensure that this project becomes a reality. I am grateful for this support. Note: Additional color versions of selected figures are available on ftp://ftp.wiley. com/public/sci_tech_med/microfluidic_devices_concepts Baton Rouge, LA, USA November 15, 2009

CHALLA S. S. R. KUMAR

CONTRIBUTORS

Ali Abou-Hassan, Laboratoire de Physicochimie des Electrolytes Colloy¨des et Sciences Analytiques (PECSA), UMR 7195, Equipe Colloy¨des Inorganiques, ˜ Paris 6, Paris Cedex 5, France UniversitO Manish Bothara, Department of Electrical and Computer Engineering, Portland State University, Portland, OR, USA Vale´rie Cabuil, Laboratoire de Physicochimie des Electrolytes Colloy¨des et Sciences Analytiques (PECSA), UMR 7195, Equipe Colloı¨des Inorganiques, Universite Paris 6, Paris Cedex 5, France Yong Chen, Institute for Integrated Cell-Material Sciences, Kyoto University, Kyoto, Japan Chunhui Deng, Department of Chemistry, School of Pharmacy, Fudan University, Shanghai, China Pamela G. Gross, Student Health and Wellness Center, University of Nevada at Las Vegas, Las Vegas, NV, USA Peter Mike Gu¨nther, Department of Physical Chemistry and Microreaction Technology, Institute of Micro- and Nanotechnologies, Ilmenau University of Technology, Ilmenau, Germany Emil P. Kartalov, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA

xi

xii

CONTRIBUTORS

Clement Kleinstreuer, Department of Mechanical and Aerospace Engineering and Department of Biomedical Engineering, North Carolina State University, Raleigh, NC, USA Andrea Knauer, Department of Physical Chemistry and Microreaction Technology, Institute of Micro- and Nanotechnologies, Ilmenau University of Technology, Ilmenau, Germany Johann Michael Ko¨hler, Department of Physical Chemistry and Microreaction Technology, Institute of Micro- and Nanotechnologies, Ilmenau University of Technology, Ilmenau, Germany Debasish Kuila, Institute for Micromanufacturing, Louisiana Tech University, Ruston, LA, USA; Department of Chemistry, North Carolina A&T State University, Greensboro, NC, USA Vindhya Kunduru, Department of Electrical Engineering, Arizona State University, Tempe, AZ, USA Jie Li, Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC, USA Yan Li, Department of Chemistry, School of Pharmacy, Fudan University, Shanghai, China ´ di Firenze, Via della Giovanna Marrazza, Dipartimento di Chimica, UnivesitA Lastruccia, Sesto Fiorentino, Italy Sriram Muthukumar, Intel Corporation, Chandler, AZ, USA Shalini Prasad, Department of Electrical Engineering, Arizona State University, Tempe, AZ, USA Olivier Sandre, Laboratoire de Physicochimie des Electrolytes Colloı¨des et Sciences Analytiques (PECSA), UMR 7195, Equipe Colloı¨des Inorganiques, Universite Paris 6, Paris Cedex 5, France Jian Shi, Ecole Normale Superieure, Paris, France Muhammad J. A. Shiddiky, Department of Chemistry and Institute of Biophysio Sensor Technology, Pusan National University, Busan, South Korea Yoon-Bo Shim, Department of Chemistry and Institute of Biophysio Sensor Technology, Pusan National University, Busan, South Korea Yamini Yadav, Department of Electrical and Computer Engineering, Portland State University, Portland, OR, USA Shihuai Zhao, Institute for Micromanufacturing, Louisiana Tech University, Ruston, LA, USA; Tianjin University, Tianjin, China

1 MICROFLUIDICS FOR NANONEUROSCIENCE PAMELA G. GROSS Student Health and Wellness Center, University of Nevada at Las Vegas, Las Vegas, NV, USA

EMIL P. KARTALOV Keck School of Medicine, University of Southern California, Los Angeles, CA, USA

1.1 INTRODUCTION The nervous system of an organism is like the information technology department of an organization. Each of the billions of building blocks of the nervous system, called neurons, is a multistate device similar to the transistors of a microprocessor. But while transistors are binary state devices, neurons are capable of being in many thousands of states, and this adds many orders of magnitude to the complexity of possible connections within a nervous system. In addition, each neuron has multiple connections with other neurons, and some of these connections are bundled into tracts and nerves that travel within brain and spinal cord, and out to peripheral locations. In computers, disconnection of one network cable, or disabling of the electronic circuits in the server, can seriously compromise the function of the organization. Similarly, traumatic injuries or neurodegenerative processes such as multiple sclerosis, Alzheimer’s disease, or Parkinson’s disease can significantly impair the functionality of an individual by damaging the neurons, tracts, and nerves. However, unlike computer systems, medical repair processes do not yet exist because we do not yet understand how the system operates in the healthy state. This may change in the near future as cell biologists pursue stem cell interventions to regenerate or remodel Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

1

2

MICROFLUIDICS FOR NANONEUROSCIENCE

damaged areas of the nervous system. Simultaneously, engineers are teaming up with biologists to design electronic implants and prostheses that can interface with functioning tissue on either side of a damaged connection and act as a bridge to allow restoration of injured neuronal circuits. Pharmaceutical researchers are using nanotechnologies to create novel systems capable of delivering targeted drugs and other agents across the previously impenetrable blood–brain barrier,1,2 a feature of nervous systems that chemically separates the system from the rest of the organism. All these advances may be accelerated by knowledge derived from studies of cellular physiology using tools designed to study biological processes at the single cell level. As our ability to fabricate tools on the micro- and nanoscale levels has progressed, we can now study cellular processes at a scale compatible with cell size, and this is revealing new information about their operational responses, including how they respond to physical and chemical cues from their immediate environment. It is important that neuroscience researchers be aware of these new technologies, so that their use can be optimized. Recent advances in biological applications of micro- or nanotechnology have included novel micro- or nanoscaled carriers for drug delivery,3–6 quantum dots that operate as nanoscaled sensors at the cellular level,7–11 and nanoelectrodes.12 In addition, self-assembled monolayers and scaffolding, as well as carbon nanotubes, have been used as artificial nanotechnology matrices for cell culture.13–19 In neuroscience specifically, nanoparticles have been used for free radical scavenging in ischemic and neurodegenerative diseases.20 Scaffolds made of self-assembling nanofibers are being developed to enhance neuroregeneration.21 The blood–brain barrier has been successfully breached by drugs attached to special nanoparticles.22 High-resolution studies of the topography and material properties of live nervous system cells are being carried out by atomic force microscopy (AFM) (Figure 1.1).23,24 Single-molecule tracking using quantum dots has revealed details about the structure and function of membrane receptors.10,25,26 Finally, nanotubes, nanowires, and nanoneedles are being developed for use as relatively nontraumatic intracellular electrodes.12,27,28 On a slightly larger scale, microfabrication technology has been used to create microfluidic platforms that have been employed for a variety of nanoneuroscience studies, and these platforms will now be discussed. Microfluidics refers to a technology that utilizes microscale channels to manipulate fluid and suspended objects in a controlled manner at the nanoliter scale. Most microfluidic chips are designed and constructed using the same techniques as used in the development of microelectronic circuitry. Microfluidics has been advancing rapidly over the past decade and has progressed from basic devices, for example, a channel,29 a valve,30 and a pump,30 to large-scale two-dimensional integration of components,31 three-dimensional architectures,32 and nonlinear autoregulatory systems.32 Simultaneously, the development of the fundamental technology has enabled the advent of a plethora of specialized devices that have miniaturized important macroscale applications such as protein crystallization,33,34 DNA sequencing,35 and PCR (polymerase chain reaction), a technique for DNA detection and amplification.36,37 The same development has also enabled the advent of novel techniques to conduct fundamental research in a scale that was never previously possible.

INTRODUCTION

3

FIGURE 1.1 Atomic force microscopy images of neural lineage cells. (a) Three-dimensional rendering of an oligodendrocyte differentiated from a murine neural stem cell. Fixed sequentially with 100% ethanol and 4% PFA, air dried, and then imaged on an Asylum Research MFP 3D AFM using an Olympus AC160 cantilever in AC mode in air. Note the detailed process formation. Scan size is 90 mm
90 mm. (b) Three-dimensional rendering of a portion of a living astrocyte derived from a human embryonic stem cell on a polyornithine/laminin-coated substrate, imaged in media, in AC mode with an Olympus Biolever, and on an Asylum Research MFP 3D AFM. The image shows cytoskeletal fibrous elements visible through the cell membrane in the proximal thicker area of the cell as they enter a broad, flat attachment area. Scan size is 30 mm
30 mm (unpublished data, Pamela G. Gross).

More recently, some microfluidic chips incorporate other microtechnology and nanotechnology hardware, such as electrodes,38–44 magnetic coils,45,46 and surface-emitting lasers,47 to enhance their capabilities beyond fluid handling. Many of the first applications of microfluidic chips involved studying the physics of fluid dynamics at the microscale (characterized by low Reynolds numbers, laminar flow, and fast diffusion), which is quite different from the flow characteristics of bulk fluid at the macroscale (characterized by higher Reynolds numbers, turbulence, and slow diffusion). The unusual behavior of fluid traversing microchannels has allowed creation of new methodologies to manipulate molecules, in order to synthesize novel nanomaterials and chemical/pharmaceutical moieties, and this has been described in other chapters. For biologists, microfluidic platforms have emerged as invaluable tools to study biology at small scales, even down to the single cell level. For neuroscientists, these “lab-on-a-chip” platforms have enabled a novel approach for experiments on the cellular physiology of the nervous system. Their usefulness in deciphering the complicated interactions involved in the differentiation, growth, and maintenance of neurons in health and in disease has become increasingly apparent within the past 5 years, as more research in this field continues to be reported. As more neuroscientists become familiar with this technology, we anticipate a rapid evolution of the field. This chapter will review pertinent contributions in the use of microfluidics to study the physiology and pathophysiology of neurons and their support cells and will hopefully serve as a primer for neuroscientists unfamiliar with this technology, inspiring some to develop new applications of microfluidics to the field of neuroscience.

4

MICROFLUIDICS FOR NANONEUROSCIENCE

Microfluidic platforms typically contain a series of chambers and channels that each measure in the range of 1 mm to a few hundred microns and are used to process fluid at a microscopic scale. For in vivo applications, microfluidic technology has been integrated with neural implants for precise delivery of solutions.48 Three-dimensional electrodes with bundled microfluidic channels that can be implanted into severed nerves to guide and monitor their regeneration while allowing infusion of drugs are also under development.49 However, the most common biological application of microfluidics has been for in vitro studies, such as the delivery and processing of biochemical reactants for DNA sequencing35 and protein analysis,50 the sorting, counting, and analysis of cells by flow cytometry,51 the delivery of cell adhesives and cells for substrate micropatterning of cell populations,52,53 the development of biomimetic three-dimensional tissues, complete with stromal support molecules,18,19 and the isolation and nurtured maintenance of individual cells to study basic cell physiology and cell–cell interactions on a single (or near-single) cell basis.54–59 In addition, microfluidic platforms have been used to study the effect of laminar flow and shear forces on the function of endothelial and other types of cells,60,61 to provide artificial circulation through various organ-simulating cell culture chambers in order to determine the pharmacokinetics of prospective pharmaceutical agents,62 and to deliver test samples containing potential toxins to cells acting as biosensors (also known as “lab-in-a-cell” technology).63–65 Finally, microfluidics can be used to study physiology within small organisms, such as the effects of anesthetics on the regrowth of severed axons, or the recovery of axonal synapses after laser ablation in Caenorhabditis elegans nematodes that have been captured and immobilized in microfluidic chips.66,67 Microfluidic-based cell studies are a useful adjunct to conventional in vitro techniques or mini culture systems68 because microfluidic chambers have the ability to control both the amount of material (media, growth factors, etc.) used for cell study and their exact distribution over well-defined periods. This can permit better control of the experiment by limiting unanticipated extraneous factors and diffusion constraints that can occur in larger systems. The effects of cell population variability will also be more limited in smaller systems and therefore individual differences among similar cells will be less likely to influence results. From an economic standpoint, small culture volumes allow cost savings since the required volume of expensive media, hormones, and growth factors is orders of magnitude less than that used in typical culture flasks. These platforms can also be designed for high throughput and compatibility with automated laboratory equipment such as plate readers. In addition, the hardware is portable and it can be mass produced so inexpensively that it can be very cost-effective to perform massively parallel microfluidic platform-based experiments, in order to confirm results or test the effects of numerous agents simultaneously. These parallel experiments are necessary to verify results obtained on individual cells since it is known that there can be significant variation in the behavior of particular cells, even if they are cloned from the same precursor cell.69 Similarly, it will be imperative that the effect of microenvironment parameters such as mechanical forces, shear stress, effective culture volume, and material interfaces be well

INTRODUCTION

5

understood and controlled before interpreting single cell study results so that these factors do not contribute to misleading conclusions.70 Nevertheless, observations derived from studies of individual cells in a controlled microenvironment may be much more likely to reveal true cellular physiology responses than those derived from studying the responses of populations of cells simultaneously, as is done with conventional in vitro studies. Although the development of this technology has progressed significantly over the past 5–7 years, its utility as a tool is just beginning to be appreciated by biologists. There are many published reviews on the general topic of microfluidics for biological applications,69–85 but there have only been a few that have focused on microfluidic applications in neuroscience.80–82 This chapter will update the reader on the discipline of microfluidics to study the nervous system at the single cell level. Specifically, it will report on microfluidic chips used to decipher physiological processes and responses of cells of neural origin, and it will also focus on examples of systems that combine microfluidic chambers with other technologies for novel research not previously possible. Section 1.2 will begin with a description of current microfluidic chamber construction techniques, starting with a discussion of the characteristics of the PDMS polymer used in many microfluidic chambers and then moving on to cover stepby-step fabrication processes. Various architectural designs of use for cellular studies will then be introduced, followed by a description of alternative applications of PDMS to create tools that are useful in customizing the substrate of microfluidic chips for specific experiments. Practical limitations of microfluidic techniques will then be discussed to present a balanced view of the topic. In Section 1.3, gradient-generating designs will be reviewed, along with examples of how they have been used to study cellular responses. Methods of incorporating electrophysiological measurements into chip design, including patch clamping, will be examined and then use of other integrated micro- and nanoscaled analytical devices will be considered. The theory and methodology used for in vivo tissue simulation will be evaluated, since the natural behavior of cells is ultimately what most biological research is attempting to discern. Following this, a literature review of neuroscience research involving microfluidic platforms will be detailed in Section 1.4, starting with cell identification and separation tools, which is essential for researchers requiring specific subpopulations of neural lineage cells. Studies on microfluidic analysis of neuropeptide release will follow, which is of interest to individuals studying synapse formation and function. The use of microchips to study the effects of physical and chemical guidance cues on single cells will then be considered since this is a key to understanding how neural cells interact with their environment and with each other. Section 1.5 will focus on electrophysiology studies that use multielectrode arrays (MEAs) as microfluidic chamber substrates. This is a popular field of endeavor since these two technologies seem to be complementary and can allow studies on action potential characteristics and propagation in single axons. The effect of growth factors on neuronal responses of microfluidically cultured and isolated cells will be covered after this, given its significance in understanding cell differentiation and maturation.

6

MICROFLUIDICS FOR NANONEUROSCIENCE

The use of microfluidic chambers for gene therapy studies on neural cells will be subsequently discussed. Although this is a relatively new area of study, preliminary results are very promising and future research will likely take advantage of the unique capabilities that microfluidic chips offer to this field. The final area of research to be covered involves studies based on the microfluidic isolation of axons and neural cell bodies. This approach to neural research is gaining great interest, given the potential applications for those studying neural degeneration and regeneration processes, in addition to those interested in axonal transport mechanisms, and synapse formation and physiology. A general discussion with consideration of future perspectives will complete the chapter. It is hoped that the reader will gain an appreciation for the future potential of these platforms to uncover previously hidden cell-based interactions in the nervous system, and this will stimulate new applications of microfluidics for their specific research programs.

1.2 PDMS MICROFLUIDIC DESIGN AND FABRICATION 1.2.1 Characteristics of PDMS Initially, most microfluidic chambers were constructed on silicon wafers using “hard” lithography. Since those early studies, “soft” lithography has been developed and various polymers and fabrication techniques have been investigated.86 Now, softsided chambers made of polydimethylsiloxane (PDMS) are gaining increased popularity, especially for biological applications. PDMS is a silicon-type elastomer and can be purchased commercially as Sylgard 184 by Dow Corning or RTV by General Electric. It can be molded into many different shapes to form valves, chambers, and channels. PDMS is advantageous for biological studies since it is biocompatible, optically transparent down to wavelengths as low as 280 nm, permeable to gases needed for cellular respiration, autoclavable, and naturally inhibitory to cellular adhesion.87,88 PDMS has therefore proven very handy for cellular studies by allowing long-term cultures, optical microscopy, and fluorescent/chemiluminescent studies, while the cells are still in situ in the chip.64 A final advantage of this material for use with cell culture systems is that PDMS has been shown to be an excellent protective coating for on-chip solid-state analytical devices (such as surface-emitting lasers), since PDMS is optically transparent yet prevents the detrimental effects of ions migrating from the culture medium into sensitive electrical junctions.47 Native PDMS is hydrophobic, and this influences many of its surface properties, including its interactions with fluid and molecules that are in contact with it. These properties can be altered by physical and chemical treatments that can change the hydrophobicity of the surface of the PDMS channels and change its adhesive properties if this is desired.74 For example, the pretreatment of the PDMS channels with bovine serum albumin (BSA) will assist in blocking cell adhesion to its surface.54 Alternatively, PDMS can be made hydrophilic and supportive of cell growth by treatment with oxygen plasma,88 or UV/ozone,89 that acts by changing the moieties on the PDMS surface to increase the number of silanol groups and decrease the number

PDMS MICROFLUIDIC DESIGN AND FABRICATION

7

of siloxane groups. Polyethylene glycol (PEG) can also be used to alter the surface chemistry of PDMS.35 The surface interactions of PDMS with adjacent molecules will also depend on local flow conditions. Experimentally, proteins such as collagen and fibrinogen adhere to both hydrophobic and hydrophilic (oxygen plasma-treated) PDMS. But under flow conditions, the oxygen plasma-treated hydrophilic surfaces experienced only temporary adhesion, followed by rapid detachment of any adherent cells, whereas the hydrophobic PDMS channels became permanently clogged with protein and cells.88 Therefore, systems that are designed to have continuous exposure to protein-laden media and cells will likely benefit from pretreatment of the PDMS with oxygen plasma to increase the functional lifetime of the channels. 1.2.2 PDMS Chip Fabrication Protocol Microfluidic chip fabrication uses many of the same techniques used in electronic circuit production. The process begins with the creation of an architectural design using a computer-aided design (CAD) software program. The design is printed on a transparency using a high-resolution printer, since the feature size on the final chip will be determined by the resolution of features on the transparency. This transparency acts as a photomask during the next step, in which it is placed over a substrate (silicon wafer or glass) that is precoated with a thin layer of photoresist, a photocurable epoxy. UV exposure polymerizes exposed areas for photoresists such as SU-8, so the developer solution can strip away the unexposed areas because they are not polymerized, while the polymerized exposed structures remain. This type of photoresist is called negative photoresist because the result is the reverse, or “negative,” of the mask. On the other hand, photoresists such as 5740, SPR-220, and the AZ family are called “positive photoresists” because the result corresponds to the mask; that is, the result is “positive” to the mask. UV exposure makes a chemical change in positive photoresists that results in the material becoming more soluble, for example, in a strong base. Thus, the developer solution removes the material from the exposed areas, while the structures in the unexposed areas remain. In both cases, the result is a mold where the features are built in photoresist. Since photoresist is softer than silicon, the resulting mold is softer than traditional molds, and so the technique has been named “soft lithography.” In the next step, PDMS is combined with its catalyst in a 10:1 proportion and the mixture is degassed in a vacuum chamber to remove bubbles. It is then poured onto the master, allowed to cure, and then peeled off the mold. Access ports are punched after casting (or silicon tube ports are placed during casting) to create connections to input and drainage tubes. The PDMS slab is then placed onto a substrate such as a silicon wafer or a glass slide to create the final microfluidic platform. The PDMS forms a reversible conformal seal to the substrate, but optional treatment of the PDMS with plasma oxidation of the PDMS surface after curing will render the surface more hydrophilic and allow the PDMS to irreversibly bond to the substrate. After sterilization by means of autoclaving, UV treatment, or immersion in 70% ethanol, the system is ready for use.

8

MICROFLUIDICS FOR NANONEUROSCIENCE

Although PDMS can reproduce features down to 10 nm in size,75 actual fabricated channels in PDMS have not yet achieved a cross-sectional area smaller than 1 mm2.74 This is because the feature size of the PDMS is determined by the printed resolution on the photomask, which is determined by the printer used to create it. For example, standard printers that have a resolution of 5080 dpi can reproduce features on the photomask down to 25 mm resolution, whereas photoplotters that print at 20,000 dpi can achieve a resolution down to 8 mm.75,90 To reproduce smaller features, chrome masks may be used, which are created with e-beam or laser writing and are much more costly. In addition, the relative softness of PDMS makes it difficult to maintain uniformly high quality of the reproduced features when the linear scale is decreased below a few microns. 1.2.3 Architectural Designs of Microfluidic Platforms The physical behavior of a fluid flowing through microscale channels is very different from the flow characteristics of the same fluid flowing through larger channels. For example, fluid flow through microfluidic channels is laminar, so mixing does not occur between solutes placed at different locations in the channel cross section, except by the slow process of diffusion. Without turbulence, solute gradients will remain relatively intact as fluid traverses down channels of uniform width. If cells are localized at certain areas of this channel, their exposure to specific concentrations of solute can be tightly controlled. In fact, different parts of the cell can even be exposed to different and controlled concentrations of the solute. This laminar flow behavior can also be used to pattern and deposit specific solute concentrations onto the substrate or to pattern cell adhesives and repellents next to each other onto the substrate prior to introduction of cells. Alternatively, if mixed patterns are desired, deliberate oblique grooving of the floor of the channel can be employed to create turbulence in order to mix solutes,91–93 and nanotopographic features can also be added to the platform substrate to influence cell adhesion.94 Most microfluidic chips use some type of dynamic flow conditions, with flow achieved by the use of syringe pumps, gravity-driven reservoirs,95 electrokinetic control,96 or other more complicated functional PDMS valve structures. These valves are designed by layering “control channels” that act as bladders across flow channels. Application of pneumatic pressure in the “control channels” can then cause controlled collapse of the underlying fluidic lumen, and this controlled deformation of the flow channel’s lumen creates a functional valve.30 Digital control and sequential coordination of these valves can create peristaltic pumps. Rotary pumps based on similar mechanisms have been designed and used for applications that require repeated cycling of fluid for mixing, such as on-chip PCR, used in amplification and identification of DNA strands in genetic engineering.97 Various structures have been devised to immobilize cells within microfluidic chips. These architectures must be able to catch and retain a cell from a passing stream of media, while minimizing damage to the cell. Sieves have been used within the culture chamber to retain cells while also producing a nutrient gradient.98 Channel walls can be constructed at partial height to create a dam that allows flow from one channel to

PDMS MICROFLUIDIC DESIGN AND FABRICATION

9

another while gently transporting and immobilizing cells for later analysis.99 Inverted T junctions that have small docks with tiny drain channels at the junction have been used to immobilize single cells and then perform rapid on-chip calcium flux assays.54 Curved docking areas that can balance the forces exerted on cells (fluid flow versus gravitation) have also been used to isolate individual cells for culture and study of calcium mobilization.56,57 Gravity-induced flow has been combined with dielectrophoresis to trap and sort cells without physically contacting them.100 For the capture of cells with variable dimensions such as pancreatic islets (used in diabetes research), designs have combined one semiellipsoidal wall and one movable wall to create an adjustable holding area that will allow studies on the regional effect of infused glucose and drugs.59 Studies on pairs of cells have used intersecting channels that have been designed to trap pairs of cells from different populations to study intercellular communication via gap junctions between their cell membranes.55 Finally, the surface of PDMS has been microstructured with arrays of wells and coupled to a microfluidic system to create a test platform for parallel experiments on single cells or small groups of cells.101 As described above, a significant advantage of microfluidic chambers is that the architectural design and the dimensions of the channels and chambers can be customized for the morphology of the cells to be studied and to the task to be accomplished. As new researchers enter this field, we expect to see a wealth of new designs for novel applications. 1.2.4 PDMS Tools In many cases, it is desirable to have a microfluidic chip substrate that is patterned with different molecules prior to assembly of the chip. This can be easily achieved by creating a separate PDMS tool that contains the substrate pattern and can be used as a stencil or a stamp. This tool is fabricated using the same techniques as outlined above. After completion of the tool, it can be used for microcontact printing by dipping the patterned area into a fluid with the desired concentration of solute molecules and then transferring this pattern to the substrate. PDMS can also be formed into a twodimensional stencil sheet that allows patterned deposition of selected proteins or agents onto the underlying substrate, and this PDMS stencil has the advantage of being useful on irregular or curved substrate surfaces. Once the protein pattern has been created on the substrate, the remaining platform can be constructed by applying the matched PDMS chip so that its channels are complementary to it. With this arrangement, future cell attachment and differentiation can be guided, and cocultures can be created in controlled geometric patterns.102–104 Since this technique can help control the exact position of neurons on a substrate, the resultant controlled neuronal patterns can be very helpful in studying neural networks and interactions occurring within synapses.69,102,103,105 These techniques have also been combined with selective oxygen plasma treatment to create long-term and short-term cell repellent areas to coculture cells in controlled geometric patterns.104 In this case, cell repellent polymers were homogeneously deposited on a substrate, and a PDMS stencil was used to selectively protect the repellent from plasma treatment in certain areas. Unprotected areas lost their repellent nature and could then be treated with adhesives like

10

MICROFLUIDICS FOR NANONEUROSCIENCE

fibronectin and short-term repellents like BSA to create patterns of relative adhesivity over time that could then be seeded as desired with different cell types. 1.2.5 Practical Considerations and Limitations As with all new technologies, there are certain practical design considerations and limitations that must be recognized before planning a microfluidic chamber for cell studies. Cell viability has definitely been correlated with channel size and proportions, closed versus open-channel configurations, and static versus dynamic media flow. For example, it has been shown that in contrast to cells grown in conventional tissue culture flasks, the proliferation rates of cells grown in microfluidic channels without media flow depend on the height of the channel.106 This is likely due to loss of convective movement of cell-expressed inhibitory factors away from cells, rather than lack of nutrients or change in osmolarity or pH of the culture medium. In a static system where there is no flow of media or connection of the media in the channels to a bulk container, secreted factors can only be dissipated by diffusion, and this can be insufficient to remove their often deleterious effects.106 If continuous or intermittent flow is designed into the system, the flow rate must be optimized to provide nutrients and remove wastes, without producing excessive shear stress that can change morphology or migration of the cells or even detach the cells.98,107 Similarly, certain secreted factors may be essential for cellular health, and if the flow is too high, then these factors may be washed out. Pretreatment of the PDMS prior to the introduction of cells can have significant effects on cell culture success. For example, Matsubara et al. showed that different treatments to make the PDMS hydrophilic affected both the morphology and the density of mast cells.64 Similarly, Prokop et al. found that extracellular matrix deposition and plasma treatment of the PDMS improved subsequent cell cultures.98 Other important considerations include recognition of the fact that the tiny volume of microfluidic culture systems confers a much less stable homeostatic system in terms of temperature, carbon dioxide concentration, and humidity control compared to standard Petri or tissue culture flasks. These chips equilibrate much more rapidly with their environment than larger systems given their larger surface area to volume ratio, so each time these chips are removed from the incubator, they are prone to more rapid alteration of their temperature, atmosphere, and humidity. Temperature alone is known to directly influence gene expression, biochemical reactions, and diffusion speed. To maintain a stable system, steps must be taken to minimize losses of environmental stability. Similarly, if media is fed into these chips via tubing that is outside the incubator (i.e., if connected to a syringe pump), or if the tubing is part of a “mini” culture system, it is important that media temperature and CO2 content do not change during transport through the tubing.68 Therefore, although these microscale systems are technically portable, control systems for their ambient environment may be necessary if they require transport outside the incubator for time lapse imaging or other interventions. We witnessed this effect directly when an isolated axon in a microfluidic chip was observed to shrink back significantly within a few minutes of removal from the incubator and placement onto the cold microscope stage

PDMS MICROFLUIDIC DESIGN AND FABRICATION

11

(unpublished data). Control of humidity is also critical because water diffuses into PDMS according to Fick’s law of diffusion.108 From there, it can evaporate and lead to increased osmolarity of the cell culture medium and premature cell death. Methods to minimize evaporation, such as coating the PDMS with a thin layer of parylene, have been successfully implemented and shown to prolong cell viability.108 The last practical consideration in using PDMS as a culture chamber in microfluidic systems involves the high ratio of chamber surface area to the volume of culture medium and number of cells compared to standard culture flasks. This increased surface area to volume ratio can lead to increased interactions between chamber contents (media, cells) and chamber walls. For example, it is known that small hydrophobic molecules may partition into the PDMS and therefore be less bioavailable when studying their effects on cells.109 This extraction of media solutes by the PDMS can significantly change the concentration of some agents in media within microfluidic channels by many orders of magnitude. The magnitude of this change depends on the partition coefficient of the substance, the pH of the culture medium, and the counterion pairing in the media. Major decreases in the media concentration of neurotransmitters, hormones, and growth factors could change experimental outcomes, and the cost savings for using small volumes of these agents in microfluidic platforms can be lost if much larger quantity of the substance has to used to achieve the same effect. Therefore, many individuals are now experimenting with different surface treatments to decrease the porosity of the PDMS and avoid some of these effects. In addition to taking up biomolecules, the PDMS may also release potentially toxic agents from its polymer matrix.110 These can then be concentrated in a proportionally smaller volume of culture medium and may affect more sensitive cells. Certainly, neurons from different sources vary in their hardiness, and culturing sensitive neurons at low density and in serum-free conditions can be difficult, even in the most tightly controlled environments. For example, we have personally had difficulty maintaining the viability of human neural stem cells in PDMS chips, whereas rat dorsal root ganglion cells thrived in the same conditions (unpublished data). One possible explanation of this phenomenon may be found in the work of Millet et al., who hypothesized that there might be seepage of toxins from the PDMS.110 In their research, they tried to improve neuron survival in open- and closed-channel microfluidic chips by treating the PDMS with serial solvent-based extraction processes (to remove potentially cytotoxic uncross-linked oligomers and residual platinum catalyst in the PDMS) or with autoclaving (to drive cross-linking and outgas solvents). They found that treatment with extraction improved neuron survival, increased the development of neurites, and lowered platinum levels in the PDMS more than did autoclaving. Specifically, the ratio of neuron survival was 3:28:51 for native PDMS, autoclaved PDMS, and extracted PDMS, respectively. Overall cell viability in low density, small volume, serum-free studies in closed-channel devices was improved from less than 2 days in native PDMS to over 7 days for extracted PDMS. If gravity-driven flow was added, survival could be further increased to over 11 days by improving nutrient delivery and waste removal compared to static systems. These extraction processes will be imperative for future studies on

12

MICROFLUIDICS FOR NANONEUROSCIENCE

individual cells or cell-to-cell interactions in low-density cultures in microscale culture volumes. Perhaps the biggest impediment to the general acceptance of microfluidic platforms as tools for biological investigation will be the initial need for interdisciplinary teams of researchers that include both engineers who can devise and fabricate the chips and biologists who know when and where to best apply the technology. As with all new technology, the developers may not be able to recognize its most useful niche, and the users may not be aware of the technology or have the auxiliary tools and expertise to operate it correctly. However, as more individuals take the steps to experiment with the technology, it would become more commonplace and better utilized. On the other hand, the novelty of the underlying technology and of the general approach of combining neuroscience with microfluidics offers unique and exciting opportunities to address fundamental problems with new tools in new ways. These technologies thereby carry the immense promise of important breakthroughs and new insights both in fundamental neuroscience and in its extensions to biomedical practice in improving the treatment of many neurological diseases.

1.3 DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS 1.3.1 Gradient-Generating Designs As discussed above, the laminar flow that occurs in microscale channels can be used advantageously to create high-resolution gradients of solutes and special factors within the cell culture chamber, in order to assess the effects of these gradients on the behavior of individual cells. These designs employ two or more inputs—one for the studied factor and one for the diluting medium, with each connected to its own syringe pump. The inputs connect to a network of serpentine, interconnected channels that repeatedly split and remix, with each generation of splitting channels increasing in number, until they finally coalesce back into a single larger channel (Figure 1.2). At each branch point, some mixing occurs so that there is a gradient of concentrations of the studied factor(s) that is oriented perpendicular to the flow direction at the final exit channel, and this gradient has a range of resolution spanning from several microns to hundreds of microns. After exiting the gradient-generating device, the established gradient is maintained by laminar flow. By varying the flow rate into one input, dynamic and asymmetric gradients of variable shape (smooth, step, or multiple peaks) can also be created.52,111,112 These devices have been used to study the effects of IL8 (interleukin 8) on neutrophil chemotaxis,113EGF (epidermal growth factor) on breast cancer cell chemotaxis,114 and various growth factors on neural stem cells.115 These gradient-generating devices are commonly used for research on the physiology of neural lineage cells since the devices provide precise control over exposure of growth and inhibitory factors to these cells. In addition, these same devices have been used to etch a controlled gradient into the surface topology of chip substrates by injecting etching reagents, or to lay down gradients of adhesives, self-assembled monolayers (SAMs), and dyes.111 Finally, these gradient generators have been

DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS

13

FIGURE 1.2 (a) Photograph showing a microfluidic device used for generating gradients of green and red dyes in solution. The three incoming channels (top part of the photograph) were connected to syringes via tubings (not visible). After combining the streams into a single, wide channel (bottom of the photograph shown in (a)), a gradient was formed across the channel, perpendicular to the direction of flow. (b) Schematic explaining the nomenclature used for the mathematical description of the network. (c) Schematic demonstrating the application of the formulas governing the splitting ratios at the branching points. The dotted lines indicate the boundary between the two combined streams. The concentrations at the end of the serpentine channels can be calculated by multiplying the concentration of the incoming streams (cp, cq, cr) with the corresponding numbers of the splitting ratio (Vp þ 1)/B, (B  Vq)/B, (Vq þ 1)/B, and (B  Vr)/B, as indicated). Reprinted with permission from Ref. 112. Copyright 2001 American Chemical Society.

combined with large chip-based arrays of cell culture chambers (10
10) to simultaneously perform 100 parallel tests on the effect of an agent’s various dilutions.116 Since these chambers each had four individual access ports, repeated growth/passage cycles of the cells could be performed on-board by microfluidic control, so the cell cultures could be maintained over long periods. As demonstrated with the studies described above, these gradient devices are very useful tools for investigations into cellular responses to varying concentrations of specific factors, whether these factors are substrate bound or dissolved in media.

14

MICROFLUIDICS FOR NANONEUROSCIENCE

1.3.2 Integrated Electrophysiology For electrically active cells such as neurons and muscle cells, integrated electrical recording is a very valuable addition to microfluidic platforms. An early example of this was a unique system that was designed as a self-contained, portable unit for field use as a cell-based biosensor. The unit incorporated a hybrid glass/PDMS/silicon chamber for cell culture with integrated microfluidics, a microelectrode array substrate modified with fibronectin and gelatin for cell growth, a temperature regulation system, on-chip electronics for acquisition, analysis and display of action potentials, and a transparent cover that makes the unit amenable to microscopic inspection.65 This approach of creating a stand-alone unit with its own environmental controls may eventually be required of many platforms in the future; however, most neuronal studies have employed much simpler hardware, typically using commercial MEAs as the substrate for a PDMS microfluidic chip. These will be discussed in more detail in Section 1.4.4. The “gold standard” for electrophysiological studies has always been patch clamping, and many microfluidic platforms incorporating arrays of patch clamp electrodes have been engineered and successfully demonstrated. Conventional patch clamps use suction to attach the tip of a glass micropipette to a cell membrane, and then break the membrane and record the intracellular potential using a conductive fluid in the micropipette. In early microfluidic designs, these systems used PDMS microfluidic channels to guide cells to pores micromachined into silicon wafers,117 or they used cell-trapping pores in a horizontal PDMS substrate.118–120 These pores simulated the tip of a conventional glass micropipette and were used to create a highresistance seal to the cell wall for subsequent electrical recording. Ionescu-Zanetti et al. improved on this design by incorporating pores on a vertical channel wall of the PDMS to facilitate the use of optical and fluorescent microscopy to monitor the procedure.121 This vertical approach allows both the cell and the capillary tube leading to the pore to be in the same plane of focus, and, therefore, it permits easier guidance of the selected cell to the pore (using a combination of flow in the cell chamber and suction from the pore). It also permits visual monitoring of the cell condition and position during the recording. Each pore is connected to a capillary tube that applies negative pressure (suction) to attach the cell and break the membrane and to a silver/silver chloride electrode that then connects to a multiplexer circuit to process the recorded signals. Using CHO (Chinese hamster ovary) cells, the seal resistance between pores and cells was an average of 300 megaohms, and the system was able to record currents down to 20 pA. Individual cell trapping could be achieved in less than 3 s, and the seal was stable for 20–40 min (Figure 1.3). This group further updated this system by raising the trapping pore above the chamber floor to avoid deformation of the trapped cell. They also opened the ceiling of main chamber to ensure rapid fluidic access for high-throughput drug profiling on the clamped cell. For these authors, this microfluidic approach to patch clamping represented a much more efficient system for pharmaceutical analysis than traditional patch clamp technology.122 Unfortunately, patch clamp resistance seals in the megaohm range as reported above are not ideal, and other groups have been modifying their techniques to

DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS

15

FIGURE 1.3 Patch clamp array on a microfluidic platform. (a) Cell trapping is achieved by applying negative pressure to recording capillaries that open into a main chamber containing cells in suspension. Attached cells deform, protruding into the capillaries. Patch clamp recordings are obtained by placing AgCl electrodes in each of the capillaries, as well as in the main chamber. Signals are fed through a multiplexing circuit and into the data acquisition system. (Multiplexer setup and microscope objective are not to scale.) The device is bonded to a glass coverslip for optical monitoring. (b) Scanning electron micrograph of three recording capillary orifices as seen from the main chamber. The capillary dimensions are 4 mm
3 mm, with a site-to-site distance of 20 mm. (c) Dark-field optical microscope image of cells trapped at three capillary orifices. Trapping was achieved by applying negative pressure to the recording capillaries. The device consists of 12 capillaries arrayed 6 along each side of the main chamber fluidic channel along a 120 mm distance. Reprinted with permission from Ref. 121. Copyright 2005 National Academy of Sciences USA.

improve this. Chen and Folch used e-beam lithography to create a 1 mm cell attachment aperture in their patch clamp chip. This method was combined with standard photolithography using high-resolution photomasks to create larger suction channels. They also used O2 plasma treatment of their master and PDMS chip to smooth the edges on the aperture, and they achieved reliable gigaohm seals and signal quality that was similar to that obtained with traditional glass pipette patch clamps.123 Commercial forms of these microfluidic patch clamp technologies will likely be available in the near future. 1.3.3 Other Integrated Sensors and Microfluidic Capabilities Microfluidic chips have employed many other complementary microtechnologies in recent years for application to biological studies. Although they have not all been used specifically for neuroscience studies, they do have this potential and they are presented here for the interested and motivated reader. For example, as an alternative to using electrodes, charged membrane-permeable, potential-sensitive dyes have been used in a microfluidic device to determine the membrane potential of cells in a rapid, highly

16

MICROFLUIDICS FOR NANONEUROSCIENCE

sensitive manner, with minimal consumption of reagents.124 Various electrical parameters including amperometry, impedance measurement, and potentiometry have also been used to analyze cells and their ionic secretions in microfluidic chambers (see the excellent review on this topic by Bao, Wang, and Lu).85 The relative acidity of the contents of microfluidic channels has been monitored by pH-sensitive fluorescently tagged monolayers (SAMs) that are bound to the substrate of the microfluidic platform.125 An alternative pH meter with higher sensitivity has used a different technology in microliter flow chambers to measure pH changes down to 0.5
103 pH units. This device has been useful in the study of cellular processes that alter ATP levels, such as receptor activation and signal transduction.126 The oxygen content in microscale cell cultures is another important parameter that can be monitored by an on-chip oxygen sensor based on fluorescent quenching of ruthenium dye particles encapsulated in the PDMS of the microfluidic culture device.62 Silicon chips containing multiple microsensors for bulk detection of extracellular pH, oxygen consumption rates, and cell morphological alterations have also been developed.127 and although not yet applied to microfluidic single cell studies, it is reasonable to expect that they might be adaptable to this purpose in the future. More advanced technology has also been miniaturized for on-chip use. For example, single nonperfused neurons have been studied with NMR (nuclear magnetic resonance) microcoils, and NMR spectroscopy has been used to determine their metabolite content, but the need for continuous perfusion to prevent cell death was noted.45 To address this issue, planar NMR probes have been incorporated into microfluidic platforms and preliminary studies on their functionality are underway.46 Other advanced technologies such as surface-enhanced Raman scattering and confocal microscopy have been combined with microfluidics to study real-time intracellular chemical dynamics of single live cells with high spatial and temporal resolution.128 Apoptosis (programmed cell death) is an important cellular process that is well studied by both biologists and pharmaceutical companies since it is critical to understanding how to control cancer and cell growth in general. Microfluidics have been used to study the multiple morphologic and biochemical changes associated with apoptosis at the single cell level.72 Tamaki et al. noninvasively monitored the change in cytochrome c distribution that occurred during apoptosis of single neuroblastoma–glioma hybrid cells confined in quartz glass microfluidic chambers by using scanning thermal lens microscopy without the need for any labeling materials.129 Finally, on-chip single cell genetic evaluation and manipulation will be useful for neuroscience cellular studies. One successful technique for this involved a combined microfluidic/microelectroporation chip that could isolate and temporarily immobilize individual prostate cancer cells in a channel, prior to application of a 10 V, 100 ms electric pulse to puncture the cell membrane and insert green fluorescent protein genes into them.130 In more recent work, a lower applied voltage of only 0.8 V for 6.5 ms was focused at one location on the membrane of individually trapped HeLa cells and resulted in successful electroporation.131 Because of the lower voltage requirement and because this design was able to monitor the permeation of the membrane by recording accompanying jumps in electrical current across the cell membrane, it represented a definite improvement over previous technology.

DESIGNS AND DEVICES FOR NEUROSCIENCE APPLICATIONS

17

Another example of a platform for genetic studies used microfluidics to isolate cells and then lyse them prior to purifying and recovering their mRNA, the genetic instruction codes that cells use to synthesize proteins.132 Further development of this technology from this same laboratory has yielded chips capable of performing 72 parallel, 450 pL reverse-transcriptase PCR reactions that could detect mRNA levels down to 34 RNA templates.133 They have also used microfluidics to synthesize cDNA from subpicogram mRNA templates isolated from single cells134 and performed gene ligation with plasmids and successfully transformed the plasmid DNA into competent cells.135 This technology has potential utility for neuronal studies. 1.3.4 Simulating In Vivo Tissues with Microfluidics To draw reasonable conclusions about in vivo processes using data derived from our in vitro experimental models, these models must simulate real tissues as closely as possible. This is why many researchers have tried to consolidate multiple cell types and extracellular matrix proteins into a three-dimensional architecture. Otherwise, the data may be misleading and oversimplified. For example, most muscle cells grown in vitro have different morphology and function compared to those grown in vivo. However, one group of researchers found that they could culture cardiac muscle myocytes with more typical morphology if they cocultured fibroblasts alongside of them on linear, intersecting patterns of collagen deposited within microfluidic channels.136 Similarly, a perfused microfluidic platform built upon a substrate with alternating cell adhesive (matrigel on poly-D-lysine) and cell repellent stripes (polyacrylamide and polyethylene glycol) was used to grow and fuse myoblasts into realistic multinucleated myotubes.91 Based on studies such as these, it is now believed that many environmental parameters can directly influence a cell’s cytoskeleton and subsequently alter cell behaviors such as proliferation, motility, and migration, to name a few. To better understand these interactions, the effect of variably sized and shaped microfabricated cell culture wells has been studied. Preliminary work shows that these single cell wells (treated with cell adhesive material via PDMS microcontact printing) have been successful in altering the three-dimensional shape of the cell contained within them.137 Further study in this field may reveal how cell shape in vivo alters cell function. In addition to controlling individual cell shape, environmental cues from cell attachment matrices containing self-assembling proteins and gel-like substances have influenced the three-dimensional shape of a group of cells in vitro.17 Microfluidics have been used to build up realistic vascular tissue by sequentially depositing layers containing different cell types and extracellular matrices (collagen, matrigel, etc.) within a platform.18,19 These types of microfluidic platforms have the capability of creating more biomimetic in vitro systems, but they may also encounter the same type of limitations on study that using live tissue sections do. Rather than trying to simulate real tissue, it may be more important in the in vitro study of individual cell physiology to create a substrate with the right characteristics. For example, to be truly “physiologic,” in vitro substrates must re-create several characteristics of normal in vivo extracellular matrices, including mechanical properties (such as elasticity,

18

MICROFLUIDICS FOR NANONEUROSCIENCE

rigidity, and strain), chemical properties (such as ligand density and orientation), and topographic properties (such as surface curvature and fibrous contact guidance). These characteristics control cell distribution in tissues and guide cell morphology, behavior, gene expression, proliferation, differentiation, and apoptosis, presumably through interactions with transmembrane integrin receptors.17,138,139 This theory is reinforced by studies that show that neurons grow better on a “soft” bed of astrocytes than on glass.139 As research in this field progresses, we expect to see more physiologic substrates being incorporated into microfluidic systems. Researchers investigating neural prosthesis development are very interested in trying to optimize substrate topography and chemistry. Their goal is to find the best substrate that will prevent astrocytes from overgrowing the implanted electrodes since these cells can interfere with signal transmission by insulating the electrodes from the neurons. Results from research on microfluidic substrates for neural growth and on prosthesis optimization will likely benefit both fields. For example, studies have shown that certain cells do prefer certain nanotopographies, as demonstrated by the finding that astroglial cells preferentially attach to pillars over wells and respond to the topography by changing their expression of cytoskeletal proteins such as actin and vinculin.140 Independent chemical cues also have different effects on different cell types. For example, substrate-bound peptides with the amino acid sequence IKVAV preferentially promote neural adhesion, whereas the sequence RGD promotes fibroblast and glial cell adhesion.141 In contrast to the goals of prosthesis development to limit astrocyte attachments, neural stem cell researchers desire a controlled bed of astrocytes to generate a permissive environment for the differentiation of neural stem cells. Research shows that combining topographic cues (in the form of substrate grooves) with chemical cues (in the form of adsorbed laminin molecules) can orient over 85% of astrocytes in the direction of the grooves.142 This can help to control the differentiation of neural stem cells cultured over the astrocytic bed and hopefully will permit directed axon regeneration in future studies. A final limitation of developing truly three-dimensional cultures in vitro has been the difficulty of maintaining cell viability when cell density approaches the order of magnitude seen in live tissues. Cullen et al. theorized that this might be due to lack of adequate perfusion to supply nutrients and remove waste products.143 They used microfluidics to create a cylindrical PDMS culture plate that had multiple inlet ports at the base and peripheral (circumferential) outlet ports along the edge. The plate was covered with FEP (fluorinated ethylene propylene) membrane to minimize evaporative losses while allowing gas exchange, and the ports were connected to a syringe pump to keep the culture volume constant, while using forced interstitial convection at various perfusion rates. Cells were loaded into a 500 mm thick 3D matrigel matrix preloaded in the plate. The researchers were able to demonstrate that a perfusion rate of 10–11 mL min1 allowed greater than 90% viability in neuronal or neuronal/astrocyte cocultures, with cell densities that more closely matched the density of the brain than prior successful models (although still lower than that found in the brain cortex). It will likely be imperative that three-dimensional systems incorporate excellent perfusion systems to truly simulate an in vivo experience.

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

19

FIGURE 1.4 Schematics showing the microscaffold system. (a) An 8
8 array of hollow microtowers with functional cross-connects along the x-direction and structural cross-connects along the y-direction. The microtowers in this schematic project 1.5 mm out the topside and backside of the Si orifice plate. (b) A 3D cross section of the microscaffold device encased in the PDMS fluid manifold. The fluid manifold allows continuous perfusion of the cells growing within the active microscaffold. Reproduced with permission from the Royal Society of Chemistry.144

Rowe et al. approached the perfusion problem from a structural engineering standpoint and actually engineered a three-dimensional scaffolding system made of SU-8 photoresist and gold electrodes.144 The design included a system of integral microchannels and ports within the major support struts, to simulate microvascular perfusion, and the entire structure was encased in a PDMS manifold for fluid delivery. Their preliminary results found that their 3D substrate successfully supported neuronal cultures and the resultant neural networks that developed showed more complexity than those grown on two-dimensional electrode arrays. Future studies planned to include electrical recordings from the gold electrodes (Figure 1.4). Regardless of whether future in vivo-type culture systems employ three-dimensional support structures built up chemically or structurally, or using a combination of these techniques, they will likely figure prominently in research that is trying to decipher cell behavior in complex geometries and communities. 1.4 NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS 1.4.1 Cell Separation Tools One prerequisite for in vitro studies on cells of neuronal lineage is starting with the right cells. When harvested from the nervous system of donor animals, samples of cells contain a mixture of large neurons and generally smaller glial support cells. A separation process must then be used to isolate the desired cell type. Wu et al. devised

20

MICROFLUIDICS FOR NANONEUROSCIENCE

a method of microfluidic cell separation using the change in spatial distribution that occurs when fluid streams of different viscosity are mixed across an expanding channel.145 They suspended the mixture of cell types in an aqueous solution of sodium alginate that supported neuronal survival rates of greater than 90% during the separation process. They then mixed a stream of this suspension with an eluent stream traveling at a different flow rate. As the flow rate increases, the viscosity decreases since the alginate polymers become more linearly aligned with the microfluidic channel walls. As the flow rate decreases, the viscosity increases since the polymer chain bunches and intertwines with itself. If the eluent stream has a higher viscosity than the stream in which the cells are suspended, then the interface between the two streams is moved in such a way that the cell suspension stream becomes narrower and more closely applied to the channel wall. When this cell suspension stream width approaches a certain minimal value, the larger cells are picked up by the higher viscosity eluent stream and the smaller cells are left behind in the original cell stream. When these streams then encounter a branching point, the cells can be separated. The researchers were able to successfully separate neurons and glia at a rate nearing 100% when flow rates and interface distributions were stringently controlled. Stem cells that are fated to differentiate into neurons or astrocytes cannot be separated based on their size since they are indistinguishable by their morphology and size. Separation by flow cytometry is based on the elaboration of different antigens that can be fluorescently tagged, but these may not be present early in the differentiation process. Flanagan et al. devised a microfluidic chip that used the dielectric properties of stem cells to characterize their fate bias.146 Electrodes applied an alternating electric field to cells in a microfluidic channel. At low frequencies, the cells are repelled from the electrodes, but as the frequency of the switching field is increased, they are attracted to the electrodes. The frequency at which they are “trapped” between repulsion and attraction is characteristic for the cell’s fate bias, and these differentiating dielectric properties can be demonstrated prior to the development of protein markers. In this study, live cells could be distinguished from dead cells that do not “trap.” Astrocytes were trapped at 0.3 MHz, neurons trapped at 5 MHz, and neural stem cells trapped at 1 MHz. In addition, neural stem cells harvested from embryos at earlier developmental ages trap at a higher frequency than those harvested from embryos at later developmental ages, consistent with the fact that more stem cells from young embryos differentiate into neurons and more stem cells from older embryos differentiate into astrocytes. This technique can be used to measure heterogeneity in cell cultures and has several advantages over flow cytometry and FACS since it can be done on small numbers of cells, it can exclude dead cells, it is sensitive to minor differences in cells, and it needs no antibody. 1.4.2 Neuropeptide Release Neuropeptides are endogenous chemicals that play a major role in the differentiation, maturation, and communication between cells of neuronal lineage. A significant portion of neurophysiology research involves deciphering the factors involved in their

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

21

release, uptake, and induced responses in local and distant cells. The effect of neuropeptides on developing neuromuscular junctions has been studied using microfluidic chips. This research is exemplified by the work of Tourovskaia et al. who used PDMS masks to apply micropatterns of cell adhesive and cell repellent molecules to isolate single myotubes in thin parallel lines.147 A microfluidic device was then placed over this, myoblasts were allowed to attach to the substrate, and streams of media containing agrin were microfluidically perfused perpendicular to the axis of the cells. Agrin is a molecule released by the tip of growing axons when they contact muscle cells. The study showed that focal application of agrin stabilized the acetylcholine receptor aggregations in the myotubes, consistent with its presumed actions involved in the early development of neuromuscular synapses (Figure 1.5). Indirect measurement of neurotransmitter release from isolated single cells was demonstrated by Huang et al.148 They used PDMS microchannels to guide single pheochromocytoma PC12 cells into a chamber etched into glass. When this cell was stimulated with nicotine, they could detect dopamine release by amperometric monitoring using a carbon fiber microelectrode. Gold-covered single cell wells in silicon chips were also used to record catecholamine release from adrenal chromaffin cells.149 PDMS microchannels were used to immobilize PC12 cells and use amperometry to record calcium-induced dopamine and norepinephrine release.150 In similar work, Sun and Gillis were able to record quantal exocytosis of catecholamines after stimulation of chromaffin cells in microfluidic channels with potassium solution.151 They recorded amperometric spikes using indium tin oxide (ITO) electrodes when the catecholamines were oxidized on the electrode surface. When trying to differentiate the complex interactions that occur between different cells in mixed-type or pure-type cultures, it is important to determine the cell’s chemical response to stimulation, in addition to its electrical and morphological responses. Although the latter responses have been more readily studied at the single cell or near-single cell level using microfluidics, the analysis of released chemicals at this scale has been more difficult to achieve, given the extremely minute concentrations of the chemicals to be studied, in addition to the fact that the exact chemical species may not be known in advance. Jo et al. designed a microfluidic chip that would allow off-line analysis of neuropeptides released in response to chemical stimulation of neurons with potassium chloride (KCl).152 They used multichambered microfluidic chips that contained a cell culture chamber functionalized with poly-L-lysine to allow attachment of Aplysia bag cell neurons. Valves were used to selectively connect this chamber to three other chambers, each functionalized with a SAM that could adsorb molecules (released neuropeptides) by hydrophobic interactions. The cell culture chamber was exposed to the KCl that activated release of the neuropeptides, and the cell culture fluid was sequentially flushed from the culture chamber into each of the SAM-containing chambers before, during, and after KCL stimulation. Once the connecting valves were closed and the solutions were allowed time to adsorb to the SAM, the PDMS was peeled from the chip, exposing the SAMs for MALDI mass spectrometry measurements and subsequent imaging. The results of the research confirmed that this methodology could successfully detect two different released neuropeptides and that the majority of the released peptides adsorbed onto the SAM

22

MICROFLUIDICS FOR NANONEUROSCIENCE

FIGURE 1.5 Synaptogenesis on a chip. (a) During development, neurons release agrin at the site of contact between nerve and muscle. (b) Fluorescence micrograph of a portion of the myotube microarray after staining the AChRs with Alexa Fluor 488-conjugated a-bungarotoxin (BTX ). Scale bar is 50 mm. (c) Three high-magnification fluorescence micrographs of myotubes stained with BTX , showing that aneural AChR clusters display intricate shapes similar to those found in vivo. (d) Phase contrast micrograph of the microfluidic device containing a ladder micropattern of myotubes during stimulation by a laminar stream of agrin (spiked with Allura red dye for visualization). The black dashed box corresponds to the area shown in (b). Reproduced from Ref. 147. Copyright 2008 Elsevier.

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

23

layer and not onto the PDMS walls. This innovative combination of microfluidic control of cell bathing solution, with off-line mass spectrometry, has significant potential to study known and previously unknown chemical responses of neurons spatially and temporally. Since both the concentrations and the volumes of released neuromediators are so low, microfluidic-based on-chip analysis of these agents can represent an attractive alternative to standard laboratory techniques. Mourzina et al. devised and optimized an on-chip capillary electrophoresis system to separate neuromediators.89 They experimented with various PDMS treatments and separation buffers to improve electroosmotic pumping and decrease adsorption of the neuromediators onto the PDMS surface. With the addition of field-amplified sample stacking, they were able to achieve separation of fluorescently labeled neuropeptides (including oxytocin, serotonin, glutamic acid, and others) within tens of seconds at 110 pL volume. Fieldamplified sample stacking utilizes a principle where the analyte is dissolved in a dilute ionic solution that is sandwiched between higher concentration ionic solutions. Application of fluid flow and electrical current causes formation of a stepped electric field, resulting in migration of the analyte into the boundary area between these solutions for easier separation. This work demonstrates the utility of using microfluidic chips to process biologically relevant samples at minute scales. Along similar lines, unpublished data from other researchers (Phillips and Wellner) has demonstrated the adaptation of commercial microfluidic micromixer chips to detect proteins such as proinflammatory cytokines in tiny samples of blood or perspiration from patients with depression using recycling immunoaffinity chromatography (online communication at http://www.nibib.nih.gov/HealthEdu/eAdvances/ 30Jan09). Further refinement of these techniques will likely be forthcoming. 1.4.3 Physical and Chemical Guidance Cues The ultimate goal of most neuroscience research involves learning ways of preventing degeneration and promoting repair and regeneration of neurons and their processes. One key prerequisite for this is an understanding of how physical and chemical guidance cues affect neurite growth. This is also of great interest to individuals who study neural network design. Many studies in this field use microchannels to isolate and observe axonal responses. An early study demonstrated the relative importance of chemical guidance when neurites from chick spinal neurons that were otherwise physically confined in “v”-shaped channels and pits on a silicon nitride substrate were able to grow out of the channels if the substrate was pretreated with polylysine.53 The effects of physical cues alone were demonstrated by studying neurite elaboration by cells confined to square bottomed channels (constructed from polyimide walls placed on a glass substrate).153 Narrow channels (20–30 mm in width) caused fewer neurites to be elaborated from each cell, and each neurite was longer and more likely to be oriented parallel to the channel wall. These changes might be due to inflexibility of the cytoskeleton. The effect of isolated chemical cues is the subject of many ongoing studies. In one such study, a PDMS device with parallel stripes of channels was used to deliver

24

MICROFLUIDICS FOR NANONEUROSCIENCE

poly-L-lysine (PLL) or collagen to a substrate pretreated with sequential 3-aminopropyltriethoxysilane (APTES) and glutaraldehyde.154 This pretreatment allowed PLL or collagen to bind covalently to the substrate, so it was more stable and structurally homogeneous than those attached by simple protein adsorption. The PDMS was then removed and Aplysia neurons were applied and monitored as their neurites developed. Standard electrophysiology and mass spectrometry were used to investigate any differences between neurons cultured on patterned versus uniform layers of protein. The results showed that patterned substrates caused shorter, thicker, less branched, and slower growing neurites and caused changes in the cell’s electrical activity compared to neurons grown on uniform proteins. The effect of substance gradients on neurite extension was also specifically studied by Whitesides and colleagues.52 This group utilized the serpentine network of microfluidic channels discussed in Section 1.3.1 to study and quantify the effect of different laminin concentrations on neurite sprouting, differentiation (into dendrite or axon), and directionality from hippocampal neurons. They deposited a layer of poly-L-lysine onto channels in a plasma-treated PDMS chip and then created a gradual gradient that ranged from pure BSA to pure laminin. This treated PDMS channel was then cut from its substrate, inverted in a Petri dish, and seeded with hippocampal neurons. Cells that grew alone in the center of the gradient (center of channel) were observed and the length of their neurites was measured to determine axonal specification (axons were typically four times the length of dendrites). The results showed that approximately 60% of axons were oriented within a 120 arc in the direction of increasing laminin concentration, and although the laminin concentration did not guide the axonal growth, it did specify which early neurite would become the axon. In addition, the concentration of laminin required to influence axonal specification was determined. It is of particular importance to understand the in vitro effects of laminin since it has been shown to play a critical role in axonal pathfinding in the embryonic CNS in vivo.155 The effect of the absolute concentration of a substance on axonal growth is just one parameter that has been studied in microfluidic chips. The slope of the substance’s concentration gradient also exerts effects on axonal growth on specific types of neurons. Lang et al. demonstrated this with ephrin A5, which is a repulsive axon guidance molecule.156 They used silicon wafers with etched microfluidic channels to create multiple stripes of varying concentrations of ephrin A5. The molecules in these solutions were transferred to a PDMS stamp and ultimately into polystyrene culture dishes into which chick retinal ganglion cells originating from nasal or temporal locations on the retina were cultured. They tested the effects of both steep gradient variations and shallow gradient variations on axonal growth. They found that axons growing from neurons originating on the nasal portion of the retina do not respond to ephrin A5 gradients at all. However, axons growing from neurons originating from the temporal portion of the retina are inhibited, grow farther into shallow gradients than steep gradients, and halt their growth at a lower total ephrin A5 concentration and total exposure in shallow gradients of this molecule than in steep gradients. This research indicates the complicated inputs involved in the growth of axons of differentiated cells and the usefulness of microfluidics in deciphering the signals controlling axonal

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

25

FIGURE 1.6 Temporal RGC axons stop in substrate-bound gradients produced by mFN. (a) Fluorescence images of a stepwise gradient of ephrin A5 spanning a distance of 320 mm, and the corresponding countergradient of Fc and temporal axons stained with phalloidin invading the gradient. In the original article with color images, in (b–e) phalloidin-stained axons are shown in black and antibody-stained Fc in green. (b) Temporal axons stopping in a steep ephrin A5 gradient. (c) The stop zone shifts further into the gradient in a shallow gradient. (d) Nasal axons in a steep gradient do not stop. (e) Temporal axons growing on laminin lanes without underlying gradient. For scale, see (a). With kind permission from Springer Science þ Business Media.156

responses (Figure 1.6). These researchers published a detailed protocol on their methodology, including the use of a second active protein to set up an overlapping or countergradient.157 Similarly, gradient mixers have been used to create substrate-bound gradients with multiple agents of defined concentrations and slopes.112 Li et al. used a syringe pump connected to a PDMS gradient mixer to combine laminin, chondroitin sulfate proteoglycans (CSPGs), and/or BSA to form various patterns on a glass substrate.158 After 12 h of adsorption, the PDMS superstructure was removed and DRG neurons were applied. The study showed that cells adhered more strongly to higher laminin and lower CSPG concentrations, neurites grew toward higher laminin and lower CSPG concentrations, and double opposing gradients provided the strongest guidance cues. This research confirmed that neurites can detect and respond to both the slope and the fractional concentration change of substrate-bound gradients.

26

MICROFLUIDICS FOR NANONEUROSCIENCE

In reality, it may be an oversimplification to study the effect of individual agents on neurons since it is more likely that they are programmed to respond to complex, interacting, and synergistic forces, including topography, electromagnetic fields, and chemical/biological cues. For example, when adult rat hippocampal progenitor cells were cocultured with astrocytes aligned on laminin-coated substrate grooves, there was enhanced neuronal differentiation and alignment of neurites parallel to the astrocyte processes and substrate grooves. It is presumed that the astrocytes align themselves on the grooves and then secrete soluble factors that are concentrated locally by the topography, resulting in facilitated neuronal differentiation of the progenitor cells.159 1.4.4 Electrophysiology and Microfluidics Studies on neurons frequently involve monitoring of multiple physiologic parameters, including their electrical activity, as well as their morphology and expressed proteins. Research in this field has been made easier by the availability of commercial MEAs, which typically employ a glass substrate with electrically conductive microcontacts and leads made of gold, platinum, or transparent ITO. Heuschkel et al. used an ITO MEA and engineered a microfluidic chamber on its surface by layering negative photoresist and using photolithography to pattern buried microchannels in the resist.38 Once it was processed and baked, laminin and polyornithine were applied and chick embryonic motoneurons were introduced via the microchannels. As the neurons grew, their electrical activity was monitored via the microelectrodes. Although the experiment was successful and the photoresist was found to be biocompatible, most investigators now use PDMS to fabricate microfluidic channels intended for cell culture. For example, Morin et al. aligned the wells and channels of a PDMS microfluidic chip to commercial and custom planar microelectrode array substrates.39 PolyL-lysine or laminin was applied to the chips prior to addition of chick or murine cortical neurons. These cells remained viable and electrically active for weeks, as demonstrated by optical microscopy and electrical responses to stimulation. Despite a few problems with isolating potentials from single cells, inhomogeneity of the cells in each well (neuronversus glial), and PDMS adhesion to the commercial microelectrode substrate, the authors felt that this system showed potential for the development of neuronal networks. In similar work, Claverol-Tinture et al. used PDMS chips with channels and wells over poly-L-lysine-coated planar ITO microelectrode arrays and manually placed individual neurons in the wells.40 Once the axons grew, they were able to record single cell spikes from the soma (contained in the well), or the axon (contained in the channels), depending on how the microfluidic channels were aligned on the electrodes (Figure 1.7). In follow-up studies, they were able to achieve a signalto-noise ratio of 20 dB when recording electrical spike activity of up to 300 mV amplitude from multiple sites on single neurites extending in microchannels.41 Thiebaud et al. also developed a PDMS microfluidic chamber that incorporated microelectrode arrays.42 The first step in the fabrication process involved using a PDMS microcontact stamp to deposit laminin onto an MEA substrate in parallel

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

27

FIGURE 1.7 (a) Phase contrast image of a confined bipolar neuron after 12 days in vitro sprouting two neurites along a microchannel. (b) Extracellular potentials associated with an action potential and recorded by the eight microelectrodes shown in (a). (c) Raster plot showing KCl and glutamate dose responses. Reproduced with permission from Ref. 41. Copyright 2007 IEEE.

stripes whose width was consistent with the size of microelectrode arrays. A PDMS microfluidic apparatus with aligned microchannels then delivered culture medium containing neuronal cells to the laminin stripes to establish the culture. Once the cells attached, agents were injected into the parallel channels and delivered to the established lines of cells via laminar flow to study the electrophysiological effects of various pharmaceutical agents on the neurons. Commercial microelectrode arrays have also been integrated into microfluidic devices and used to study the effect of temperature changes on the electrical activity of a subpopulation of cold-sensitive cells derived from the dorsal root ganglion.44 In this research, a microfluidic chip was used to deliver polylysine over the array prior to the application of the cell suspension. There were two inlets to the chip—one

28

MICROFLUIDICS FOR NANONEUROSCIENCE

providing culture medium at 45 C, and the other providing medium at 4 C. By varying the flow rate of each input, they could rapidly (<1 s) switch the culture medium temperature flowing over cells from 35 to 16 C. Using a multichannel multiprocessor recording system, they recorded action potentials from the array of electrodes and found that certain cells consistently changed their firing rate from a mean of 0.028 spikes per second to a mean of 0.94 spikes per second in response to a switch to a colder temperature. These cells had the same morphology as the others that were not sensitive to temperature changes. This system demonstrated that the combined use of two unique microtechnologies (microfluidics and microelectrodes) could characterize and identify a special subpopulation of cells based on their electrophysiological responses. In a follow-up study, these researchers modified their microfluidic design by adding small reservoirs with flexible membrane covers to the side channels.43 This essentially introduced a “switch” that could rapidly perturb the relative flow from each input and vary the temperature flowing over the cells over a 50 ms pulse. Future physiologically relevant research will benefit from this potential to rapidly and transiently deliver agents to cultured cells on a timescale that more closely matches a cell’s innate responsiveness capabilities. 1.4.5 Growth Factor Effects As discussed in Section 1.3.1, microfluidic channels can be arranged to create gradient generators to study the effects of special factors on cell physiology. One research study examined the effect of various concentrations and combinations of growth factors (epidermal growth factor, fibroblast growth factor 2, and platelet-derived growth factor) on neural stem cell proliferation and differentiation.115 Chung et al. first demonstrated that they could successfully culture human neural stem cells in chambers that were precoated with poly-L-lysine and laminin, and kept under constant low flow (0.1 mL min1) of culture medium for at least 7 days. This low flow would help minimize autocrine and paracrine effects of secreted factors. When growth factor gradients were added, they demonstrated that neural stem cells proliferated in direct proportion to the growth factor concentration, whereas astrocytes differentiated in inverse proportion to the growth factor concentration, and all cells demonstrated increased migration toward areas with higher concentrations. The use of microfluidic gradients in this study presented a significant advantage over routine in vitro culture techniques and helped to elucidate the action of these specific factors on these cells. Wittig et al. also investigated the use of microfluidic channels to deliver graded concentrations of growth factors.160 They applied a reusable PDMS microfluidic culture medium and a factor delivery apparatus onto a standard poly-L-lysine/laminincoated Petri dish. Two channels delivering different additives in media coalesced in a “Y” shape that was placed adjacent to a neonatal spiral ganglion explant culture area. The base of the “Y” allowed growing neurites to sample two different media choices and then to decide which arm of the “Y” they preferred to grow into. They demonstrated that neurites preferred to grow toward culture medium containing neurotrophin-3 and they anticipated that this approach could be very useful in studying

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

29

the various effects of growth and inhibitory factors on proliferation, neurite extension, and cell migration. In addition to allowing precise delivery of controlled concentrations of growth factors spatially, microfluidic platforms with control valves have also been used to control the timing of the delivery of these substances to cells to determine the differentiation of cells. Nakashima and Yasuda used a microfluidic control valve to release nerve growth factor (NGF) through nanopores to control the differentiation and axonal growth of adrenal pheochromocytoma cells.161 They were able to switch the microvalve on and off with controlled frequency and duty cycles to guide cellular differentiation. They anticipated that the addition of an electrode to monitor real-time cellular response to their pulsed release of growth factors would be very useful in studying the physiological responses of cells during axonal regeneration processes. 1.4.6 Gene Therapy An alternative method of delivering growth factors to cells involves the use of microfluidic methods to genetically manipulate neurons by directly delivering the DNA codes for a given growth factor to them. Houchin-Ray et al. cultured neurons on a substrate that had been microfluidically patterned with a mixture of lipoplexes and plasmid DNA that coded for NGF and green fluorescent protein.162 They were able to achieve a transfection efficiency of 25% using a vector concentration 10 times less than typically used in culture media. Neurons cultured on the patterned areas had improved survival and enhanced neurite outgrowth, indicating that the NGF DNA had been incorporated into the cells and expressed. They also determined that pretreatment of the PDMS microfluidic chip with pluronic (an amphiphilic copolymer with surfactant properties) improved transfection rates since it made the PDMS less likely to bind the cationic lipid/DNA complexes. The NGF concentration gradient could be adjusted by changing the plasmid density in the solution and the size of the microchannels in the PDMS chip used to pattern the substrate. The potential application of patterned gene delivery and expression to specific cells in culture has significant promise for studies of neural physiology, cell-to-cell interactions, and regenerative medicine. Microfluidic chips that allow fluidic isolation of parts of neurons (Section 1.4.7) have enabled selective delivery of nonviral DNA to either neurites or the cell body of neurons to study the differences in processing that occurs at these locations. This is important for eventual in vivo studies since access to both the soma and the axon may not be possible given the long lengths of many axons. Since many studies are aimed at axon regeneration after spinal cord injury, treatments may need to be specialized for the site that is accessible. Bergen and Pun illustrated this point with their research that microfluidically delivered DNA in culture medium to the isolated neurites or to the isolated soma of PC12 neuron-like cells.163 The DNA was attached to either a lipidbased carrier (lipofectamine) or PEI (poly(ethylenimine)), a polymeric nanoparticle. In general, both had 4–5 times the uptake if delivered at the soma, compared to the neurite, although the lipid-based carriers had an uptake that was 5–7 times that for the PEI carrier. Uptake seemed to be mediated by vesicle formation. When delivered to the isolated neurites, the lipoplexed DNA could be taken up, but not transported,

30

MICROFLUIDICS FOR NANONEUROSCIENCE

whereas the PEI-based DNA was taken up and retrogradely transported in a saltatory fashion, but never made it into the soma. Therefore, gene expression occurred only if the gene-complexed carrier was delivered directly to the soma, and the transfection efficiency was much higher for the lipofectamine than for the PEI. This research sheds light on the difficulties inherent in repairing a damaged spinal cord. 1.4.7 Axonal Isolation As demonstrated in the above literature review, microfluidic chambers have been used to isolate small numbers of cells of neuronal lineage and to study their individual responses to stimuli, often with the aid of on-board analytical devices. However, none of these studies were able to fluidically isolate cell segments (like axons) from their soma, or from neighboring cells of various lineages. The only method to achieve this in the past involved the use of Campenot chambers that used nerve growth factor or brain-derived neurotrophic factor to “artificially” stimulate axonal growth from macro-scaled cultures across grease layers. However, Jeon’s group at the University of California at Irvine has successfully designed and implemented a microfluidic platform that incorporates tiny grooves of adequate size and length to allow fluidic isolation of axons from the regular cell culture chamber, so that their physiology can be studied independently.164 Specifically, they designed a PDMS chip with two chambers that were separated by a series of grooves that each had dimensions of 10 mm width, 3 mm height, and 150 mm length. These grooves were used to guide neurite growth from a cell culture chamber into a second chamber that could be used to study the isolated axons. The narrowness of the grooves prevented cell bodies from penetrating them, and the length of the grooves prevented the typically shorter dendrites from emerging into the axonal isolation chamber. A key design element of the tiny grooves was their high resistance to fluid transport that allowed them to achieve temporary (15 h) fluidic isolation of the somal compartment from the axonal compartment by applying a slightly higher hydrostatic pressure in the somal compartment. On doing this, there was one-way flow only, going from the somal to the axonal compartment, in a very slow and restricted manner. This allows independent chemical manipulation of the axon, without any direct effects on the soma, unless the axon directly transports the agents to the soma in a retrograde fashion. These researchers used a patterned poly-L-lysine substrate to keep the axons aligned in parallel stripes, so that they could be more easily identified along with their respective cell body.164–166 They used several alternative tools for the patterning, including a PDMS mold with tiny channels that could wick the agent in by capillary action (micromolding in capillaries), a PDMS stamp to transfer the agent by microcontact printing, and a PDMS mask to selectively protect and preserve a uniform, preapplied, dried agent during plasma etching. In follow-up research, Jeon’s group lengthened the microgrooves to 450 mm to allow longer studies (14 days) that still isolated axons from dendrites (Figure 1.8).164,165 They used this design to investigate axonal injury, regeneration, and interactions with cocultured oligodendrocytes during myelination. In addition, they proved that their isolated compartments could permit detection of purely axonal

NEUROPHYSIOLOGY EXPERIMENTS USING MICROFLUIDIC CHIPS

31

FIGURE 1.8 The microfluidic-based culture platform directs axonal growth of CNS neurons and fluidically isolates axons. (a) The culture chamber consists of a PDMS mold containing a relief pattern of somal and axonal compartments (1.5 mm wide, 7 mm long, 100 mm high) connected by microgrooves (10 mm wide, 3 mm high). The optically transparent PDMS adheres to a polylysine-coated coverslip. Rat CNS neurons (medium gray spots) are added to the somalside reservoir and are drawn into the somal channel (black) by capillary action. Within 3–4 days, axonal growth is guided into the axonal side (light gray) through the microgrooves. (b) A volume difference between the somal side and the axonal side (50 mL) allows chemical microenvironments to be isolated to axons for over 20 h owing to the high fluidic resistance of the microgrooves. Similarly, the volume difference can be reversed to isolate a chemical microenvironment to the somal side. (c) Fluidic isolation of Texas red dextran (top panel) to the axonal compartment demonstrates that axonal or somatic microenvironments can be independently manipulated using this culture platform. Axonally restricted application of CellTracker Green (middle panel) backtracked neurons from their isolated axons. The bottom image is the merged figure. Scale bar, 100 mm. (d) Counts of radioactivity in samples from somal and axonal compartments after [35S]methionine was localized to the axonal compartment for over 20 h. Counts in the somal compartment (3.7 c.p.m.  1.5 s.e.m.) were similar to background levels. Error bars, s.e.m. (n ¼ 3). Reprinted with permission from Ref. 165. Copyright 2005 Macmillan Publishers Ltd.

32

MICROFLUIDICS FOR NANONEUROSCIENCE

mRNA, and they demonstrated changes in gene expression in the soma in response to axonal chemical and physical manipulations. This chip has now become an important tool used in research programs at different academic settings, given its potential to reveal details of neuronal pathophysiology in neurodegenerative diseases and traumatic nerve damage. When neuronal axons are damaged, the distal portion experiences a process called Wallerian degeneration, in which the proximal portion of the axon shrinks back toward the cell body. Regeneration of damaged axons is possible but seems to be hampered by many complicated and interacting factors, including glial scar tissue that prevents the regrowing axons from finding their target cells. One active inhibitory component of the scar tissue is a group of molecules composed of core proteins with carbohydrate side chains, known as CSPGs. Neuroscientists studying axonal regeneration have postulated that treatments that limited the production of these molecules at sites of nerve damage might permit axonal regrowth and reestablishment of normal function. Jeon’s group tested this hypothesis in their microfluidic chip by isolating axons in a chamber with pre-applied stripes of alternating inhibitory and permissive molecules.167 The inhibitory stripes contained these aggrecan proteins to mimic glial scars that are generated after spinal cord injury. These inhibitory areas were effectively neutralized when chondroitinase was added to the axonal chamber of the microfluidic chip. This molecule acts to dissolve the carbohydrate glycosaminoglycan side chains of the CSPGs, but left the core protein intact. Since the addition of the chondroitinase allowed axons to cross onto the CSPG stripe, they concluded that the core protein of the CSPG was not inhibitory, but the carbohydrate side chains were the active inhibitory component of the CSPG molecule (Figure 1.9). Other proteins inhibitory to axon growth have also been tested by this research laboratory. Park et al. used their axon isolation chip to test the effects of two myelinassociated proteins on axon regeneration.168 After the axons grew into the isolation chamber (about 7 days), they were severed with vacuum aspiration and their regeneration was monitored in the presence and absence of various concentrations of NOGO-66 and MAG protein. Each of these proteins decreased the length of regenerated axons by 75–80% compared to controls. Higher concentrations of NOGO lead to increased inhibition of regeneration, with the effect saturating at a concentration of 10 nM. The effect of toxins on the electrophysiology and morphology of neurons has been studied using similar axonal isolation microfluidic chips that are combined with MEA substrates. Ravula et al. mated a PDMS superstructure onto a glass substrate with a patterned MEA and recorded the spontaneous and stimulated electrical activities from neurons that they cultured in this platform.169 They were able to record action potentials from both the soma and the isolated axons when they added potassium chloride. When the sodium channel blocker tetrodotoxin was added to the axonal compartment, only axonal action potentials ceased. In follow-up work, Ravula et al. tested the effects of other chemicals on the electrophysiology of fluidically isolated neuronal soma and axons. They found that low-dose vincristine caused no effect if directly applied to the soma of the neuron, but axonal application sequentially caused a decreased excitability of the axon, an initial increase in excitability of the soma, and

DISCUSSION AND FUTURE PERSPECTIVES

33

FIGURE 1.9 Effect of chondroitinase ABC treatment on axons growing in microfluidic strip assay device. 10 DIV cortical neurons growing on aggrecan–PLL patterned strips on glass. Axons avoid inhibitor-coated areas until a drug is applied to help axons overcome the inhibition or break down the surface-bound inhibitors. Axons avoid aggrecan strips and are confined to PLL strips during entire experimental period. Striking changes in growth cone morphology and axonal projections were observed when the pattern was treated with chondroitinase ABC, an enzyme that removes the CSPGs’ side chains without degradation of the core protein. After 30 min ChABC treatment, axons are observed to randomly extend across the pattern. Higher resolution micrographs indicate that aggrecan inhibition is overcome not only in the growth cones but also in the middle of the stripes. Reprinted from Ref. 167. Copyright 2008 Elsevier.

an eventual degeneration of the axon.170 Higher dose vincristine caused degeneration if applied to either compartment. The change in electrophysiology occurred approximately 6 h after vincristine exposure and was an earlier indicator of degeneration than were morphological changes. The degradation of the electrical response to a depolarizing dose of potassium chloride is first noted in the distal axon and then progresses proximally. Eighteen hours after exposure, morphological degeneration begins and progresses at a rate of 1–2 mm per day. All the above studies indicate that microfluidic axonal isolation chips, with or without electrical monitoring ability, hold promise for high-throughput screening of many pharmaceutical agents that have effects on neuronal health and might be useful in enhancing axon regeneration. 1.5 DISCUSSION AND FUTURE PERSPECTIVES Neuroscientists have typically been very open to the use of cutting edge technology for the study of neuronal physiology and have incorporated this technology into their daily research. The challenges of applying novel micro- and nanofabricated hardware to classic neurophysiology experiments will hopefully be matched by their potential

34

MICROFLUIDICS FOR NANONEUROSCIENCE

yield. The strength of microfluidics is in the highly efficient utilization of the reductionist approach in well-defined, tightly focused environments and problems. Another advantage of microfluidics is the integrability with multiple methods of interrogation, including chemical, optical, and electrical methods. Microfluidic platforms have the ability to deliver individual cells to specific locations, and then allow study of the effects of temporally and spatially controlled environmental perturbations on isolated parts of the cells. Arrays of these platforms can be arranged to allow large parallel experiments that can multiply experimental yield. When analytical components (such as NMR coils, pH, and electrical and chemical sensors) are incorporated into the microfluidic chips, real-time information on individual live cells can be recorded and followed. It is possible that the conclusions drawn from studies of conventional cell culture populations may not be borne out at the single cell level. Determining which in vitro cellular behaviors are most consistent with in vivo realities will then become imperative since it is likely that the choice of culture techniques may influence experimental results. Initiating a microfluidic-based research program is much simpler than one might think. Most clean room fabrication facilities at universities already have the technology to create the master for the PDMS platform since it is the same as that used to create silicon wafer-based electronics. If there is no access to these services locally, commercial businesses and some universities (such as the University of California at Irvine Integrated Nanosystems Research Facility Foundry171) do offer these services for a fee. Once the master is available, the supplies and the equipment to cast the PDMS chips are common and inexpensive. For example, PDMS (sold as Sylgard 184) is available from Fisher Scientific for about $60 for a 1.1 pound kit, which will be sufficient for dozens of chips. Vacuum pumps (to degas the PDMS prior to pouring), and ovens (to bake the mold), will improve the quality of the casting, but these are not necessary for chips that lack very small features. The only other tools that may be necessary are punches to gain access to the channels in the chip, and syringe pumps, if called for in the design. Therefore, once a master is obtained, the ability to produce multiple chips is within easy reach to life scientists. For those who prefer to purchase their PDMS chips, complete microfluidic fabrication foundries (such as those located at Stanford University and California Institute of Technology) offer commercial services to build chips based on original designs sent by researchers. Plasma treatment equipment confers the advantage of temporarily altering the surface of the PDMS to make it more hydrophilic and able to irreversibly bond to the substrate. However, reversible bonds are sufficient for most low-pressure cell culture applications, and, if desired, the PDMS surface can be made hydrophilic by simple exposure to media containing 10–15% serum. Commercial multielectrode arrays and supporting electronic modules are available from many sources and can be used as the substrate for microfluidic chips. Substrates with specialized nanotopographic features can be produced by most clean room fabrication facilities. Patterned cell adhesives are easily obtained using dedicated PDMS stencils or stamps. Three-dimensional scaffolding such as collagen sponges from BD Biosciences or Inamed Biomaterials can be used within PDMS growth chambers for studies on tissue engineering within microfluidic chips.

DISCUSSION AND FUTURE PERSPECTIVES

35

Interdisciplinary collaborations between neuroscientists and engineers may offer the best chance for success for life scientists who plan to initiate a microfluidic research program. In this setting, the engineers are often happy to provide technical advice and hardware in exchange for biological applications for their technology. This may be especially fruitful for each member of the team since federal funding agencies are recognizing the importance of interdisciplinary work for innovative groundbreaking fundamental research and industrial applications. As we look toward the future, material science issues may become a critical factor for the progress of microfluidic neuroscience. The literature cited above validates the potential for microfluidics, but issues of possible chemical interactions between the PDMS and the cells or the culture medium will have to be resolved. It is known that many materials that are used in standard in vitro studies may interact with culture contents. For example, proteins are known to adsorb onto polystyrene culture flasks but the relative ratio of the culture medium volume to the polystyrene surface area usually makes the impact of this phenomenon negligible for most applications. However, molecular and fluid interactions with PDMS on tiny culture volumes will be more likely to play a role in influencing results. Potential toxic effects of the PDMS on sensitive neurons will also need to be controlled. It may be that PDMS will have to be layered with other polymers, like parylene, to limit its interaction with media. Or PDMS may have to be replaced with other polymers that are not as “porous” to water and solutes. Finally, environmental controls may have to be built into the chips so that their temperature, osmolarity, and atmosphere are preserved during manipulation and imaging of the chip. Given the platform’s small volume and lower homeostatic reserve, the heat, humidity, and carbon dioxide levels established in an incubator may otherwise change very quickly as the chip is removed for microscopic examination or media changes. The larger ratio of surface area to volume of these small culture chambers may also lead to a greater tendency toward evaporation and osmotic concentration of the media. All these issues can result in significant effects on the cell’s physiological responses and viability. So, to take advantage of the platform’s ability to precisely control the milieu of the cells it contains, we may first need to engineer a controllable environment for the platform itself. Despite these challenges, the application of microfluidic technology to appropriate in vitro neuroscience research has the potential to offer new insights and to augment ongoing conventional in vitro and in vivo investigations. As with all technology, there are certain limitations and challenges to microfluidic studies, as outlined in Section 1.2.5. The “ground-up” approach to studying cellular interactions may yield truly novel results, but caution must be exercised before generalizing information from single cell studies to either homogeneous or heterogeneous populations of cells. Moreover, as demonstrated in the field of tissue engineering, generalizing conclusions from any two-dimensional cell culture (whether microfluidic or conventional in scale) to in vivo settings may also be misleading. But, as long as we keep our perspective, we can combine the results obtained from both conventional and alternative methodologies to obtain a greater understanding of the processes we are trying to decipher.

36

MICROFLUIDICS FOR NANONEUROSCIENCE

In conclusion, the science of studying individual neuronal cells is still in its infancy, but we now have the tools needed for this endeavor in the form of microfabricated microfluidic channels and electronic sensors that provide the platform for cell proliferation, separation, differentiation, and monitoring. Newer designs also allow isolation of individual axons on these cells, which will enable study of localized axonal physiology and its effects on the cell body. Combining microfluidics with other nanotechnologies will enable truly novel experiments never previously possible. As this chapter outlines, there are many applications of microfluidics to neuroscience, and individuals who can make use of this technology may pioneer entire new areas of research, including studies of individual axons, individual synapses, or the interactions of single neurons with other isolated CNS cells (microglia, oligodendrocytes, and astrocytes). Advances in other complementary microtechnologies, such as noninvasive single cell electrical recording, monitors for other key markers of cellular physiology (such as pH and ionic currents), tools for NMR spectroscopy, mass spectrometry, and designs to allow on-chip genetic manipulation and evaluation, should allow rapid advances to be made toward understanding neuronal physiology. Once we understand how these cells operate under controlled conditions, we may be able to use this information to gain control of cell responses in disease, and design treatments and possible cures for patients suffering from neurologic diseases.

ACKNOWLEDGMENT We thank Raymond W. Glover, MD, for reviewing and editing the manuscript.

REFERENCES 1. Silva, G.A. Nanotechnology approaches for the regeneration and neuroprotection of the central nervous system. Surg. Neurol. 2005, 63, 301–306. 2. Silva, G.A. Neuroscience nanotechnology: progress, opportunities and challenges. Nat. Rev. Neurosci. 2006, 7, 65–74. 3. Moshfeghi, A.A.; Peyman, G.A. Micro- and nanoparticulates. Adv. Drug Deliv. Rev. 2005, 57, 2047–2052. 4. Rosi, N.L.; Giljohann, D.A.; Thaxton, C.S.; Lytton-Jean, A.K.R.; Han, M.S.; Mirkin, C.A. Oligonucleotide-modified gold nanoparticles for intracellular gene regulation. Science 2006, 312, 1027–1030. 5. Svenson, S.; Tomalia, D.A. Dendrimers in biomedical applications—reflections on the field. Adv. Drug Deliv. Rev. 2005, 57, 2106–2129 (commentary). 6. Yokoyama, M. Drug targeting with nano-sized carrier systems. J. Artif. Organs 2005, 8, 77–84. 7. Gomez, N.; Winter, J.O.; Shieh, F.; Saunders, A.E.; Korgel, B.A.; Schmidt, C.E. Challenges in quantum dot-neuron active interfacing. Talanta 2005, 67, 462–471.

REFERENCES

37

8. Michalet, X.; Pinaud, F.F.; Bentolila, L.A.; Tsay, J.M.; Doose, S.; Li, J.J.; Sundaresan, G.; Wu, A.M.; Gambhir, S.S.; Weiss, S. Quantum dots for live cells, in vivo imaging, and diagnostics. Science 2005, 307, 538–544. 9. Pinaud, F.; Michalet, X.; Bentolila, L.A.; Tsay, J.M.; Doose, S.; Li, J.J.; Iyer, G.; Weiss, S. Advances in fluorescence imaging with quantum dot bio-probes. Biomaterials 2006, 27, 1679–1687. 10. Vu, T.Q.; Maddipati, R.; Blute, T.A.; Nehilla, B.J.; Nusblat, L.; Desai, T.A. Peptideconjugated quantum dots activate neuronal receptors and initiate downstream signaling of neurite growth. Nano Lett. 2005, 5, 603–607. 11. Winter, J.O.; Liu, T.Y.; Korgel, B.A.; Schmidt, C.E. Recognition molecule directed interfacing between semiconductor quantum dots and nerve cells. Adv. Mater. 2001, 13, 1673–1677. 12. Esplandiu, M.J.; Bittner, V.G.; Giapis, K.P.; Collier, C.P. Nanoelectrode scanning probes from fluorocarbon-coated single-walled carbon nanotubes. Nano Lett. 2004, 4, 1873–1879. 13. Albrecht, D.R.; Underhill, G.H.; Wassermann, T.B.; Sah, R.I.; Bhatia, S.N. Probing the role of multicellular organization in three-dimensional microenvironments. Nat. Methods 2006, 3, 369–375. 14. Hu, H.; Ni, Y.C.; Montana, V.; Haddon, R.C.; Parpura, V. Chemically functionalized carbon nanotubes as substrates for neuronal growth. Nano Lett. 2004, 4, 507–511. 15. Mattson, M.P.; Haddon, R.C.; Rao, A.M. Molecular functionalization of carbon nanotubes and use as substrates for neuronal growth. J. Mol. Neurosci. 2000, 14, 175–182. 16. Mohammed, J.S.; DeCoster, M.A.; McShane, M.J. Micropatterning of nanoengineered surfaces to study neuronal cell attachment in vitro. Biomacromolecules 2004, 5, 1745–1755. 17. Stevens, M.M.; George, J.H. Exploring and engineering the cell surface interface. Science 2005, 310, 1135–1138. 18. Tan, W.; Desai, T.A. Layer-by-layer microfluidics for biomimetic three-dimensional structures. Biomaterials 2004, 25, 1355–1364. 19. Tan, W.; Desai, T.A. Microscale multilayer cocultures for biomimetic blood vessels. J. Biomed. Mater. Res. A 72A, 2005, 146–160. 20. Dugan, L.L.; Lovett, E.G.; Quick, K.L.; Lotharius, J.; Lin, T.T.; O’Malley, K.L. Fullerenebased antioxidants and neurodegenerative disorders. Parkinsonism Relat. Disord. 2001, 7, 243–246. 21. Yang, F.; Murugan, R.; Ramakrishna, S.; Wang, X.; Ma, Y.X.; Wang, S. Fabrication of nano-structured porous PLLA scaffold intended for nerve tissue engineering. Biomaterials 2004, 25, 1891–1900. 22. Tiwari, S.B.; Amiji, M.M. A review of nanocarrier-based CNS delivery systems. Curr. Drug Deliv. 2006, 3, 219–232. 23. McNally, H.A.; Ben Borgens, R. Three-dimensional imaging of living and dying neurons with atomic force microscopy. J. Neurocytol. 2004, 33, 251–258. 24. McNally, H.A.; Rajwa, B.; Sturgis, J.; Robinson, J.P. Comparative three-dimensional imaging of living neurons with confocal and atomic force microscopy. J. Neurosci. Methods 2005, 142, 177–184. 25. Dahan, M.; Levi, S.; Luccardini, C.; Rostaing, P.; Riveau, B.; Triller, A. Diffusion dynamics of glycine receptors revealed by single-quantum dot tracking. Science 2003, 302, 442–445.

38

MICROFLUIDICS FOR NANONEUROSCIENCE

26. Pathak, S.; Cao, E.; Davidson, M.C.; Jin, S.H.; Silva, G.A. Quantum dot applications to neuroscience: new tools for probing neurons and glia. J. Neurosci. 2006, 26, 1893–1895. 27. Guillorn, M.A.; McKnight, T.E.; Melechko, A.; Merkulov, V.I.; Britt, P.F.; Austin, D.W.; Lowndes, D.H.; Simpson, M.L. Individually addressable vertically aligned carbon nanofiber-based electrochemical probes. J. Appl. Phys. 2002, 91, 3824–3828. 28. Patolsky, F.; Timko, B.P.; Yu, G.H.; Fang, Y.; Greytak, A.B.; Zheng, G.F.; Lieber, C.M. Detection, stimulation, and inhibition of neuronal signals with high-density nanowire transistor arrays. Science 2006, 313, 1100–1104. 29. Duffy, D.C.; McDonald, J.C.; Schueller, J.A.; Whitesides, G.M. Rapid prototyping of microfluidic systems in poly(dimethylsiloxane). Anal Chem. 1998, 70, 4974–4984. 30. Unger, M.A.; Chou, H.P.; Thorsen, T.; Scherer, A.; Quake, S.R. Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 2000, 113–116. 31. Thorsen, T.; Maerkl, S.J.; Quake, S. Microfluidic large-scale integration. Science 2002, 298, 580–584. 32. Kartalov, E.P.; Walker, C.; Taylor, C.R.; Anderson, W.F.; Scherer, A. Microfluidic vias enable nested bioarrays and autoregulatory devices in Newtonian fluids. PNAS 2006, 103, 12280–12284. 33. Hansen, C.L.; Skordalakes, E.; Berger, J.M.; Quake, S. A robust and scalable microfluidic metering method that allows protein crystal growth by free interface diffusion. PNAS 2002, 99, 16531–16536. 34. Hansen, C.L.; Sommer, M.O.A.; Quake, S. Systematic investigation of protein phase behavior with a microfluidic formulator. PNAS 2004, 101, 14431–14436. 35. Kartalov, E.P.; Quake, S.R. Microfluidic device reads up to four consecutive base pairs in DNA sequencing-by-synthesis. Nucleic Acids Res. 2004, 2873–2879. 36. Liu, J.; Enzelberger, M.; Quake, S. A nanoliter rotary device for polymerase chain reaction. Electrophoresis 2002, 23, 1531–1536. 37. Liu, J.; Hansen, C.L.; Quake, S. Solving the “world-to-chip” interface problem with a microfluidic matrix. Anal. Chem. 2003, 75, 4718–4723. 38. Heuschkel, M.O.; Guerin, L.; Buisson, B.; Bertrand, D.; Renaud, P. Buried microchannels in photopolymer for delivering of solutions to neurons in a network. Sens. Actuators B Chem. 1998, 48, 356–361. 39. Morin, F.; Nishimura, N.; Griscom, L.; LePioufle, B.; Fujita, H.; Takamura, Y.; Tamiya, E. Constraining the connectivity of neuronal networks cultured on microelectrode arrays with microfluidic techniques: a step towards neuron-based functional chips. Biosens. Bioelectron. 2006, 21, 1093–1100. 40. Claverol-Tinture, E.; Ghirardi, M.; Fiumara, F.; Rosell, X.; Cabestany, J. Multielectrode arrays with elastomeric microstructured overlays for extracellular recordings from patterned neurons. J. Neural. Eng. 2005, 2, L1–L7. 41. Claverol-Tinture, E.; Cabestany, J.; Rosell, X. Multisite recording of extracellular potentials produced by microchannel-confined neurons in vitro. IEEE Trans. Biomed. Eng. 2007, 54, 331–335. 42. Thiebaud, P.; Lauer, L.; Knoll, W.; Offenhausser, A. PDMS device for patterned application of microfluids to neuronal cells arranged by microcontact printing. Biosens. Bioelectron. 2002, 17, 87–93.

REFERENCES

39

43. Pearce, T.M.; Williams, J.J.; Kruzel, S.P.; Gidden, M.J.; Williams, J.C. Dynamic control of extracellular environment in in vitro neural recording systems. IEEE Trans. Neural. Syst. Rehabil. Eng. 2005, 13, 207–212. 44. Pearce, T.M.; Wilson, J.A.; Oakes, S.G.; Chiu, S.Y.; Williams, J.C. Integrated microelectrode array and microfluidics for temperature clamp of sensory neurons in culture. Lab Chip 2005, 5, 97–101. 45. Grant, S.C.; Aiken, N.R.; Plant, H.D.; Gibbs, S.; Mareci, T.H.; Webb, A.G.; Blackband, S.J. NMR spectroscopy of single neurons. Magn. Reson. Med. Sci. 2000, 44, 19–22. 46. Massin, C.; Vincent, F.; Homsy, A.; Ehrmann, K.; Boero, G.; Besse, P.A.; Daridon, A.; Verpoorte, E.; de Rooij, N.F.; Popovic, R.S. Planar microcoil-based microfluidic NMR probes. J. Magn. Reson. 2003, 164, 242–255. 47. Gourley, P.L.; Copeland, R.G.; Cox, J.D.; Hendricks, J.K.; McDonald, A.E.; Peterson, S.L.; Sasaki, D.Y. Biocompatible semiconductor optoelectronics. J. Biomed. Opt. 2002, 7, 546–554. 48. Lee, K.; He, J.P.; Clement, R.; Massia, S.; Kim, B. Biocompatible benzocyclobutene (BCB)-based neural implants with micro-fluidic channel. Biosens. Bioelectron. 2004, 20, 404–407. 49. Suzuki, T.; Kotake, N.; Mabuchi, K.; Takeuchi, S. Bundled microfluidic channels for nerve regeneration electrodes. Proceedings of the 3rd International IEEE EMBS Conference on Neural Engineering, 2007, pp. 17–18. 50. Kartalov, E.P.; Zhong, J.F.; Scherer, A.; Quake, S.R.; Taylor, C.R.; Anderson, W.F. Highthroughput multi-antigen microfluidic fluorescence immunoassays. Biotechniques 2006, 40, 85–90. 51. Fu, A.Y.; Chou, H.P.; Spence, C.; Arnold, F.H.; Quake, S.R. An integrated microfabricated cell sorter. Anal. Chem. 2002, 74, 2451–2457. 52. Dertinger, S.K.W.; Jiang, X.Y.; Li, Z.Y.; Murthy, V.N.; Whitesides, G.M. Gradients of substrate-bound laminin orient axonal specification of neurons. Proc. Natl. Acad. Sci. USA 2002, 99, 12542–12547. 53. Martinoia, S.; Bove, M.; Tedesco, M.; Margesin, B.; Grattarola, M. A simple microfluidic system for patterning populations of neurons on silicon micromachined substrates. J. Neurosci. Methods 1999, 87, 35–44. 54. Wheeler, A.R.; Throndset, W.R.; Whelan, R.J.; Leach, A.M.; Zare, R.N.; Liao, Y.H.; Farrell, K.; Manger, I.D.; Daridon, A. Microfluidic device for single-cell analysis. Anal. Chem. 2003, 75, 3581–3586. 55. Lee, P.J.; Hung, P.J.; Shaw, R.; Jan, L.; Lee, L.P. Microfluidic application-specific integrated device for monitoring direct cell–cell communication via gap junctions between individual cell pairs. Appl. Phys. Lett. 2005, 86. 56. Peng, X.Y.; Li, P.C.H. A three-dimensional flow control concept for single-cell experiments on a microchip. 2. Fluorescein diacetate metabolism and calcium mobilization in a single yeast cell as stimulated by glucose and pH changes. Anal. Chem. 2004, 76, 7400–7400. 57. Peng, X.Y.; Li, P.C.H. A three-dimensional flow control concept for single-cell experiments on a microchip. 1. Cell selection, cell retention, cell culture, cell balancing, and cell scanning. Anal. Chem. 2004, 76, 5273–5281. 58. Peng, X.Y.; Li, P.C.H. A three-dimensional flow control concept for single-cell experiments on a microchip. 2. Fluorescein diacetate metabolism and calcium mobilization in

40

59.

60.

61. 62.

63.

64.

65. 66.

67.

68.

69. 70. 71. 72. 73. 74. 75. 76.

MICROFLUIDICS FOR NANONEUROSCIENCE

a single yeast cell as stimulated by glucose and pH changes. Anal. Chem. 2004, 76, 5282–5292. Rocheleau, J.V.; Walker, G.M.; Head, W.S.; McGuinness, O.P.; Piston, D.W. Microfluidic glucose stimulation reveals limited coordination of intracellular Ca2 þ activity oscillations in pancreatic islets. Proc. Natl. Acad. Sci. USA 2004, 101, 12899–12903. Ma, N.; Koelling, K.W.; Chalmers, J.J. Fabrication and use of a transient contractional flow device to quantify the sensitivity of mammalian and insect cells to hydrodynamic forces. Biotechnol. Bioeng. 2002, 80, 428–437. Gray, B.L.; Lieu, D.K.; Collins, S.D.; Smith, R.L.; Barakat, A.I. Microchannel platform for the study of endothelial cell shape and function. Biomed. Microdevices 2002, 4, 9–16. Sin, A.; Chin, K.C.; Jamil, M.F.; Kostov, Y.; Rao, G.; Shuler, M.L. The design and fabrication of three-chamber microscale cell culture analog devices with integrated dissolved oxygen sensors. Biotechnol. Prog. 2004, 20, 338–345. Rothert, A.; Deo, S.K.; Millner, L.; Puckett, L.G.; Madou, M.J.; Daunert, S. Whole-cellreporter-gene-based biosensing systems on a compact disk microfluidics platform. Anal. Biochem. 2005, 342, 11–19. Matsubara, Y.; Murakami, Y.; Kobayashi, M.; Morita, Y.; Tamiya, E. Application of onchip cell cultures for the detection of allergic response. Biosens. Bioelectron. 2004, 19, 741–747. DeBusschere, B.D.; Kovacs, G.T.A. Portable cell-based biosensor system using integrated CMOS cell-cartridges. Biosens. Bioelectron. 2001, 16, 543–556. Guo, S.X.; Bourgeois, F.; Chokshi, T.; Durr, N.J.; Hilliard, M.A.; Chronis, N.; Ben-Yakar, A. Femtosecond laser nanoaxotomy lab-on-a-chip for in vivo nerve regeneration studies. Nat. Methods 2008, 5, 531–533. Allen, P.B.; Sgro, A.E.; Chao, D.L.; Doepker, B.E.; Edgar, J.S.; Shen, K.; Chiu, D.T. Single-synapse ablation and long-term imaging in live C. elegans. J. Neurosci. Methods 2008, 173, 20–26. Ho, C.L.; Mou, T.Y.; Chiang, P.S.; Weng, C.L.; Chow, N.H. Mini chamber system for long-term maintenance and observation of cultured cells. Biotechniques 2005, 38, 267–273. Chen, C.S.; Jiang, X.Y.; Whitesides, G.M. Microengineering the environment of mammalian cells in culture. MRS Bull. 2005, 30, 194–201. Walker, G.M.; Zeringue, H.C.; Beebe, D.J. Microenvironment design considerations for cellular scale studies. Lab Chip 2004, 4, 91–97. Li, N.; Tourovskaia, A.; Folch, A. Biology on a chip: microfabrication for studying the behavior of cultured cells. Crit. Rev. Biomed. Eng. 2003, 31, 423–488. Qin, J.H.; Ye, N.N.; Liu, X.; Lin, B.C. Microfluidic devices for the analysis of apoptosis. Electrophoresis 2005, 26, 3780–3788. Vilkner, T.; Janasek, D.; Manz, A. Micro total analysis systems. Recent developments. Anal. Chem. 2004, 76, 3373–3385. Kartalov, E.P.; Anderson, W.F.; Scherer, A. The analytical approach to PDMS microfluidic technology and its biological applications. J. Nanosci. Nanotechnol. 2006, 6, 2265–2277. Sia, S.K.; Whitesides, G.M. Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies. Electrophoresis 2003, 24, 3563–3576. Park, T.H.; Shuler, M.L. Integration of cell culture and microfabrication technology. Biotechnol. Prog. 2003, 19, 243–253.

REFERENCES

41

77. Andersson, H.; van den Berg, A. Microtechnologies and nanotechnologies for single-cell analysis. Curr. Opin. Biotechnol. 2004, 15, 44–49. 78. Andersson, H.; van den Berg, A. Microfluidic devices for cellomics: a review. Sens. Actuators B Chem. 2003, 92, 315–325. 79. Andersson, H.; van den Berg, A. Microfabrication and microfluidics for tissue engineering: state of the art and future opportunities. Lab Chip 2004, 4, 98–103. 80. Pearce, T.M.; Williams, J.C. Microtechnology: meet neurobiology. Lab Chip 2007, 7, 30–40. 81. Gross, P.G.; Weiner, L.P.; Kartalov, E.P.; Scherer, A. Microfluidic techniques for studying the nervous system. Crit. Rev. Neurobiol. 2005, 17, 119–144. 82. Gross, P.G.; Kartalov, E.P.; Scherer, A.; Weiner, L.P. Applications of microfluidics for neuronal studies. J. Neurol. Sci. 2007, 252, 135–143. 83. Meyvantsson, I.; Beebe, D.J. Cell Culture Models in Microfluidic Systems. Annu. Rev. Anal. Chem. 2008, 1, 423–449. 84. Paguirigan, A.L.; Beebe, D.J. Microfluidics meet cell biology: bridging the gap by validation and application of microscale techniques for cell biological assays. Bioessays 2008, 30, 811–821. 85. Bao, N.; Wang, J.; Lu, C. Recent advances in electric analysis of cells in microfluidic systems. Anal. Bioanal. Chem. 2008, 391, 933–942. 86. Becker, H.; Gartner, C. Polymer microfabrication methods for microfluidic analytical applications. Electrophoresis 2000, 21, 12–26. 87. De Silva, M.N.; Desai, R.; Odde, D.J. Micro-patterning of animal cells on PDMS substrates in the presence of serum without use of adhesion inhibitors. Biomed. Microdevices 2004, 6, 219–222. 88. Peterson, S.L.; McDonald, A.; Gourley, P.L.; Sasaki, D.Y. Poly(dimethylsiloxane) thin films as biocompatible coatings for microfluidic devices: cell culture and flow studies with glial cells. J. Biomed. Mater. Res. A 72A, 2005, 10–18. 89. Mourzina, Y.; Kalyagin, D.; Steffen, A.; Offenhausser, A. Electrophoretic separations of neuromediators on microfluidic devices. Talanta 2006, 70, 489–498. 90. Linder, V.; Wu, H.K.; Jiang, X.Y.; Whitesides, G.M. Rapid prototyping of 2D structures with feature sizes larger than 8 mu m. Anal. Chem. 2003, 75, 2522–2527. 91. Tourovskaia, A.; Figueroa-Masot, X.; Folch, A. Differentiation-on-a-chip: a microfluidic platform for long-term cell culture studies. Lab Chip 2005, 5, 14–19. 92. Stroock, A.D.; Dertinger, S.K.; Whitesides, G.M.; Ajdari, A. Patterning flows using grooved surfaces. Anal. Chem. 2002, 74, 5306–5312. 93. Stroock, A.D.; Dertinger, S.K.W.; Ajdari, A.; Mezic, I.; Stone, H.A.; Whitesides, G.M. Chaotic mixer for microchannels. Science 2002, 295, 647–651. 94. Martines, E.; McGhee, K.; Wilkinson, C.; Curtis, A. A parallel-plate flow chamber to study initial cell adhesion on a nanofeatured surface. IEEE Trans. Nanobioscience 2004, 3, 90–95. 95. Zhu, X.; Yi Chu, L.; Chueh, B.H.; Shen, M.; Hazarika, B.; Phadke, N.; Takayama, S. Arrays of horizontally-oriented mini-reservoirs generate steady microfluidic flows for continuous perfusion cell culture and gradient generation. Analyst 2004, 129, 1026–1031. 96. Erickson, D.; Liu, X.Z.; Venditti, R.; Li, D.Q.; Krull, U.J. Electrokinetically based approach for single-nucleotide polymorphism discrimination using a microfluidic device. Anal. Chem. 2005, 77, 4000–4007.

42

MICROFLUIDICS FOR NANONEUROSCIENCE

97. Liu, J.; Enzelberger, M.; Quake, S. A nanoliter rotary device for polymerase chain reaction. Electrophoresis 2002, 23, 1531–1536. 98. Prokop, A.; Prokop, Z.; Schaffer, D.; Kozlov, E.; Wikswo, J.; Cliffel, D.; Baudenbacher, F. NanoLiterBioReactor: Long-term mammalian cell culture at nanofabricated scale. Biomed. Microdevices 2004, 6, 325–339. 99. Yang, M.; Li, C.W.; Yang, J. Cell docking and on-chip monitoring of cellular reactions with a controlled concentration gradient on a microfluidic device. Anal. Chem. 2002, 74, 3991–4001. 100. Boettcher, M.; Jaeger, M.; Kischbaum, M.; Mueller, T.; Schnelle, T.; Duschl, C. Gravitation-driven stress-reduced cell handling. Anal. Bioanal. Chem. 2008, 390, 857–863. 101. Lovchik, R.; von Arx, C.; Viviani, A.; Delamarche, E. Cellular microarrays for use with capillary-driven microfluidics. Anal. Bioanal. Chem. 2008, 390, 801–808. 102. Chang, J.C.; Brewer, G.J.; Wheeler, B.C. A modified microstamping technique enhances polylysine transfer and neuronal cell patterning. Biomaterials 2003, 24, 2863–2870. 103. Cornish, T.; Branch, D.W.; Wheeler, B.C.; Campanelli, J.T. Microcontact printing: a versatile technique for the study of synaptogenic molecules. Mol. Cell. Neurosci. 2002, 20, 140–153. 104. Tourovskaia, A.; Barber, T.; Wickes, B.T.; Hirdes, D.; Grin, B.; Castner, D.G.; Healy, K.E.; Folch, A. Micropatterns of chemisorbed cell adhesion-repellent films using oxygen plasma etching and elastomeric masks. Langmuir 2003, 19, 4754–4764. 105. Nam, Y.; Chang, J.C.; Wheeler, B.C.; Brewer, G.J. Gold-coated microelectrode array with thiol linked self-assembled monolayers for engineering neuronal cultures. IEEE Trans. Biomed. Eng. 2004, 51, 158–165. 106. Yu, H.M.; Meyvantsson, I.; Shkel, I.A.; Beebe, D.J. Diffusion dependent cell behavior in microenvironments. Lab Chip 2005, 5, 1089–1095. 107. Gu, W.; Zhu, X.Y.; Futai, N.; Cho, B.S.; Takayama, S. Computerized microfluidic cell culture using elastomeric channels and Braille displays. Proc. Natl. Acad. Sci. USA 2004, 101, 15861–15866. 108. Heo, Y.S.; Cabrera, L.M.; Song, J.W.; Futai, N.; Smith, G.D.; Takayam, S. Characterization and resolution of evaporation-mediated osmolality shifts that constrain microfluidic cell culture in poly(dimethylsiloxane) devices. Anal. Chem. 2007, 79, 1126–1134. 109. Toepke, M.W.; Beebe, D.J. PDMS absorption of small molecules and consequences in microfluidic applications. Lab Chip 2006, 6, 1484–1486. 110. Millet, L.J.; Stewart, M.E.; Sweedler, J.V.; Nuzzo, R.G.; Gillette, M.U. Microfluidic devices for culturing primary mammalian neurons at low densities. Lab Chip 2007, 7, 987–994. 111. Jeon, N.L.; Dertinger, S.K.W.; Chiu, D.T.; Choi, I.S.; Stroock, A.D.; Whitesides, G.M. Generation of solution and surface gradients using microfluidic systems. Langmuir 2000, 16, 8311–8316. 112. Dertinger, S.K.W.; Chiu, D.T.; Jeon, N.L.; Whitesides, G.M. Generation of gradients having complex shapes using microfluidic networks. Anal. Chem. 2001, 73, 1240–1246. 113. Lin, F.; Nguyen, C.M.C.; Wang, S.J.; Saadi, W.; Gross, S.P.; Jeon, N.L. Effective neutrophil chemotaxis is strongly influenced by mean IL-8 concentration. Biochem. Biophys. Res. Commun. 2004, 319, 576–581.

REFERENCES

43

114. Wang, S.J.; Saadi, W.; Lin, F.; Nguyen, C.M.C.; Jeon, N.L. Differential effects of EGF gradient profiles on MDA-MB-231 breast cancer cell chemotaxis. Exp. Cell Res. 2004, 300, 180–189. 115. Chung, B.G.; Flanagan, L.A.; Rhee, S.W.; Schwartz, P.H.; Lee, A.P.; Monuki, E.S.; Jeon, N.L. Human neural stem cell growth and differentiation in a gradient-generating microfluidic device. Lab Chip 2005, 5, 401–406. 116. Hung, P.J.; Lee, P.J.; Sabounchi, P.; Lin, R.; Lee, L.P. Continuous perfusion microfluidic cell culture array for high-throughput cell-based assays. Biotechnol. Bioeng. 2005, 89, 1–8. 117. Pantoja, R.; Nagarah, J.M.; Starace, D.M.; Melosh, N.A.; Blunck, R.; Bezanilla, F.; Heath, J.R. Silicon chip-based patch-clamp electrodes integrated with PDMS microfluidics. Biosens. Bioelectron. 2004, 20, 509–517. 118. Sigworth, F.J.; Klemic, K.G. Microchip technology in ion-channel research. IEEE Trans. Nanobioscience 2005, 4, 121–127. 119. Klemic, K.G.; Klemic, J.F.; Reed, M.A.; Sigworth, F.J. Micromolded PDMS planar electrode allows patch clamp electrical recordings from cells. Biosens. Bioelectron. 2002, 17, 597–604. 120. Klemic, K.G.; Klemic, J.F.; Sigworth, F.J. An air-molding technique for fabricating PDMS planar patch-clamp electrodes. Pflugers Arch. 2005, 449, 564–572. 121. Ionescu-Zanetti, C.; Shaw, R.M.; Seo, J.G.; Jan, Y.N.; Jan, L.Y.; Lee, L.P. Mammalian electrophysiology on a microfluidic platform. Proc. Natl. Acad. Sci. USA 2005, 102, 9112–9117. 122. Lau, A.Y.; Hung, P.J.; Wu, A.R.; Lee, A.P. Open-access microfluidic patch-clamp array with raised lateral cell trapping sites. Lab Chip 2006, 6, 1510–1515. 123. Chen, C.; Folch, A. A high-performance elastomeric patch clamp chip. Lab Chip 2006, 6, 1338–1345. 124. Farinas, J.; Chow, A.W.; Wada, H.G. A microfluidic device for measuring cellular membrane potential. Anal. Biochem. 2001, 295, 138–142. 125. Mela, P.; Onclin, S.; Goedbloed, M.H.; Levi, S.; Garcia-Parajo, M.F.; van Hulst, N.F.; Ravoo, B.J.; Reinhoudt, D.N.; van den Berg, A. Monolayer-functionalized microfluidics devices for optical sensing of acidity. Lab Chip 2005, 5, 163–170. 126. Hafner, F. Cytosensor  microphysiometer: technology and recent applications. Biosens. Bioelectron. 2000, 15, 149–158. 127. Brischwein, M.; Motrescu, E.R.; Cabala, E.; Otto, A.M.; Grothe, H.; Wolf, B. Functional cellular assays with multiparametric silicon sensor chips. Lab Chip 2003, 3, 234–240. 128. Zhang, X.; Yin, H.; Cooper, J.M.; Haswell, S.J. Characterization of cellular chemical dynamics using combined microfluidic and Raman techniques. Anal. Bioanal. Chem. 2008, 390, 833–840. 129. Tamaki, E.; Sato, K.; Tokeshi, M.; Aihara, M.; Kitamori, T. Single-cell analysis by a scanning thermal lens microscope with a microchip: direct monitoring of cytochrome c distribution during apoptosis process. Anal. Chem. 2002, 74, 1560–1564. 130. Huang, Y.; Rubinsky, B. Flow-through micro-electroporation chip for high efficiency single-cell genetic manipulation. Sens. Actuators A Phys. 2003, 104, 205–212. 131. Khine, M.; Lau, A.; Ionescu-Zanetti, C.; Seo, J.; Lee, L.P. A single cell electroporation chip. Lab Chip 2005, 5, 38–43.

44

MICROFLUIDICS FOR NANONEUROSCIENCE

132. Hong, J.W.; Studer, V.; Hang, G.; Anderson, W.F.; Quake, S.R. A nanoliter-scale nucleic acid processor with parallel architecture. Nat. Biotechnol. 2004, 22, 435–439. 133. Marcus, J.S.; Anderson, W.F.; Quake, S. Parallel picoliter RT-PCR assays using microfluidics. Anal. Chem. 2006, 78, 956–958. 134. Marcus, J.S.; Anderson, W.F.; Quake, S.R. Microfluidic single cell mRNA isolation and analysis. Anal. Chem. 2006, 78, 3084–3089. 135. Hong, J.W.; Chen, Y.; Anderson, W.F.; Quake, S.R. Molecular biology on a microfluidic chip. J. Phys. Condens. Matter 2006, 18, 1–11. 136. Camelliti, P.; McCulloch, A.D.; Kohl, P. Microstructured cocultures of cardiac myocytes and fibroblasts: a two-dimensional in vitro model of cardiac tissue. Microsc. Microanal. 2005, 11, 249–259. 137. Dusseiller, M.R.; Schlaepfer, D.; Koch, M.; Kroschewski, R.; Tester, M. An inverted microcontact printing method on topographically structured polystyrene chips for arrayed micro-3-D culturing of single cells. Biomaterials 2005, 26, 5917–5925. 138. Schwarz, U.S.; Bischofs, I.B. Physical determinants of cell organization in soft media. Med. Eng. Phys. 2005, 27, 763–772. 139. Discher, D.E.; Janmey, P.; Wang, Y.-l. Tissue cells feel and respond to the stiffness of their substrate. Science 2005, 310, 1139–1143. 140. Turner, A.M.P.; Dowell, N.; Turner, S.W.P.; Kam, L.; Isaacson, M.; Turner, J.N.; Craighead, H.G.; Shain, W. Attachment of astroglial cells to microfabricated pillar arrays of different geometries. J. Biomed. Mater. Res. 2000, 51, 430–441. 141. Massia, S.P.; Holecko, M.M.; Ehteshami, G.R. In vitro assessment of bioactive coatings for neural implant applications. J. Biomed. Mater. Res. 68A, 2004, 177–186. 142. Recknor, J.B.; Recknor, J.C.; Sakaguchi, D.S.; Mallapragada, S.K. Oriented astroglial cell growth on micropatterned polystyrene substrates. Biomaterials 2004, 25, 2753–2767. 143. Cullen, D.K.; Vukasinovic, J.; Glezer, A.; Laplaca, M.C. Microfluidic engineered high cell density three-dimensional neural cultures. J. Neural Eng. 2007, 4, 159–172. 144. Rowe, L.; Almasri, M.; Lee, K.; Fogleman, N.; Brewer, G.J.; Nam, Y.; Wheeler, B.C.; Vukasinovic, J.; Glezer, A.; Frazier, A.B. Active 3-D microscaffold system with fluid perfusion for culturing in vitro neuronal networks. Lab Chip 2007, 7, 475–482. 145. Wu, Z.; Hjort, K.; Wicher, G.; Svenningsen, A.F. Microfluidic high viability neural cell separation using viscoelastically tuned hydrodynamic spreading. Biomed. Microdevices 2008, 10, 631–638. 146. Flanagan, L.A.; Lu, J.; Wang, L.; Marchenko, S.A.; Jeon, N.L.; Lee, A.P.; Monuki, E.S. Unique dielectric properties distinguish stem cells and their differentiated progeny. Stem Cells 2008, 26, 656–665. 147. Tourovskaia, A.; Li, N.; Folch, A. Localized acetylcholine receptor clustering dynamics in response to microfluidic focal stimulation with agrin. Biophys. J. 2008, 95, 3009–3016. 148. Huang, W.-H.; Cheng, W.; Zhang, Z.; Pang, D.-W.; Wang, Z.-L.; Cheng, J.-K.; Cui, D.-F. Transport, location and quantal release monitoring of single cells on a microfluidic device. Anal. Chem. 2004, 76, 483–488. 149. Chen, P.; Xu, B.; Tokranova, N.; Feng, X.; Castracane, J.; Gillis, K.D. Amperometric detection of quantal catecholamine secretion from individual cells on micromachined silicon chips. Anal. Chem. 2003, 75, 518–524.

REFERENCES

45

150. Li, M.W.; Spence, D.M.; Martin, R.S. A microchip-based system for immobilizing PC12 cells and amperometrically detecting catecholamines released after stimulation with calcium. Electroanalysis 2005, 17, 1171–1180. 151. Sun, X.; Gillis, K.D. On-chip amperometric measurement of quantal catecholamine release using transparent indium tin oxide electrodes. Anal. Chem. 2006, 78, 2521–2525. 152. Jo, K.; Heien, M.L.; Thompson, L.B.; Zhong, M.; Nuzzo, R.G.; Sweedler, J.V. Mass spectrometric imaging of peptide release from neuronal cells within microfluidic devices. Lab Chip 2007, 7, 1454–1460. 153. Mahoney, M.J.; Chen, R.R.; Tan, J.; Saltzman, W.M. The influence of microchannels on neurite growth and architecture. Biomaterials 2005, 26, 771–778. 154. Romanova, E.V.; Fosser, K.A.; Rubakhin, S.S.; Nuzzo, R.G.; Sweedler, J.V. Engineering the morphology and electrophysiological parameters of cultured neurons by microfluidic surface patterning. FASEB J. 2004, 18, 1267–1269. 155. Cohen, J.; Burne, J.F.; McKinlay, C.; Winter, J. The role of laminin and the laminin/ fibronectin receptor complex in the outgrowth of retinal ganglion cell axons. Dev. Biol. 1987, 122, 407–418. 156. Lang, S.; von Philipsborn, A.C.; Bernard, A.; Bonhoeffer, F.; Bastmeyer, M. Growth cone response to ephrin gradients produced by microfluidic networks. Anal. Bioanal. Chem. 2008, 390, 809–816. 157. von Philipsborn, A.C.; Lang, S.; Jiang, Z.; Bonhoeffer, F.; Bastmeyer, M. Substrate-bound protein gradients for cell culture fabricated by microfluidic networks and microcontact printing. Sci. STKE 2007, 414, 1–11. 158. Li, G.N.; Liu, J.; Hoffman-Kim, D. Multi-molecular gradients of permissive and inhibitory cues direct neurite outgrowth. Ann. Biomed. Eng. 2008, 36, 889–904. 159. Recknor, J.B.; Sakaguchi, D.S.; Mallapragada, S.K. Directed growth and selective differentiation of neural progenitor cells on micropatterned polymer substrates. Biomaterials 2006, 27, 4098–4108. 160. Wittig, J.H.; Ryan, A.F.; Asbeck, P.M. A reusable microfluidic plate with alternate-choice architecture for assessing growth preference in tissue culture. J. Neurosci. Methods 2005, 144, 79–89. 161. Nakashima, Y.; Yasuda, T. Cell differentiation guidance using chemical stimulation controlled by a microfluidic device. Sens. Actuators A Phys. 2007, 139, 252–258. 162. Houchin-Ray, T.; Whittlesey, K.J.; Shea, L.D. Spatially patterned gene delivery for localized neuron survival and neurite extension. Mol. Ther. 2007, 15, 705–712. 163. Bergen, J.M.; Pun, S.H. Analysis of the intracellular barriers encountered by nonviral gene carriers in a model of spatially controlled delivery to neurons. J. Gene Med. 2007, 10, 187–197. 164. Taylor, A.M.; Rhee, S.W.; Tu, C.H.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L. Microfluidic multicompartment device for neuroscience research. Langmuir 2003, 19, 1551–1556. 165. Taylor, A.M.; Blurton-Jones, M.; Rhee, S.W.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L. A microfluidic culture platform for CNS axonal injury, regeneration and transport. Nat. Methods 2005, 2, 599–605. 166. Rhee, S.W.; Taylor, A.M.; Tu, C.H.; Cribbs, D.H.; Cotman, C.W.; Jeon, N.L. Patterned cell culture inside microfluidic devices. Lab Chip 2005, 5, 102–107.

46

MICROFLUIDICS FOR NANONEUROSCIENCE

167. Vahidi, B.; Park, J.W.; Kim, H.J.; Jeon, N.L. Microfluidic-based strip assay for testing the effects of various surface-bound inhibitors in spinal cord injury. J. Neurosci. Methods 2008, 170, 188–196. 168. Park, J.W.; Vahidi, B.; Joon, K.H.; Rhee, S.W.; Jeon, N.L. Quantitative analysis of CNS axon regeneration using a microfluidic neuron culture device. BioChip J. 2008, 2, 44–51. 169. Ravula, S.K.; McClain, M.A.; Wang, M.S.; Glass, J.D.; Frazier, A.B. A multielectrode microcompartment culture platform for studying signal transduction in the nervous system. Lab Chip 2006, 6, 1530–1536. 170. Ravula, S.K.; Wang, M.S.; McClain, M.A.; Asress, S.A.; Frazier, B.; Glass, J.D. Spatiotemporal localization of injury potentials in DRG neurons during vincristineinduced axonal degeneration. Neurosci. Lett. 2007, 415, 34–39. 171. Minteer, S.D., Ed. Microfluidic Techniques: Reviews and Protocol, Humana Press: Totowa, NJ, 2006.

2 NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS SHALINI PRASAD AND VINDHYA KUNDURU Department of Electrical Engineering, Arizona State University, Tempe, AZ , USA

YAMINI YADAV AND MANISH BOTHARA Department of Electrical and Computer Engineering, Portland State University, Portland, OR, USA

SRIRAM MUTHUKUMAR Intel Corporation, Chandler, AZ, USA

2.1 INTRODUCTION There has been a tremendous interest in the past decade in the use of nanoporous membranes to develop sensors. These membrane-based technologies offer multiple advantages in sensing: due to the nanoporosity of membranes that enables size-based trapping of biomolecules, they can be manufactured using standard processes with tremendous quality control in membrane specifications, they are biocompatible, and they can be easily surface functionalized to ensure specificity in immobilization. This chapter reviews the advances and current state of the art in nanomembrane technology in the context of biosensing; key membrane manufacturing technologies have been reviewed and the methodologies of incorporating them into microelectronic/microfabricated platforms for developing biosensor devices have been reviewed. This chapter explains in detail the need for real-time measurement of biomolecule binding in biosensors and justification for incorporating nanoporous membranes into “lab-on-a-chip” biosensing devices. Prior to understanding the complex procedure of

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

47

48

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

construction of a membrane-based biosensor, the basic concept of a biosensor and different techniques to embed the nanoporous membrane have been reviewed. The techniques of biomolecule entrapment including encapsulation, covalent bonding, and adsorption that play an important role in determining the functionality and performance of a nanoporous membrane-based biosensor have been looked at. We have evaluated a number of applications for the nanoporous membrane-based biosensors in healthcare and environmental monitoring. A wide range of nanoporous membranes made from organic, inorganic, and hybrid nanomaterials have been examined. Methods such as anodization, ion track etching, phase separation, sol–gel rapid thermal annealing, focused ion beam, and lithography used for synthesis of nanoporous membrane have been described in this chapter. We have reviewed a range of electrochemical detection methodologies that have been most commonly adopted to perform detection with these membranes. The incorporation or integration of these nanoporous membranes into microscale platforms has been reviewed. We have also compared the multiple detection mechanisms based on constraints such as sensitivity, selectivity, time, and cost. Nanoporous membrane-based biosensors for detection of microorganism such as bacteria, pathogens, and viruses and analysis of concentration for body fluids such as glucose and cholesterol have been discussed. Finally, prospects of the development of natural and biomimetic nanomembranes have been discussed. The chapter has been organized to demonstrate the need for integration of nanoporous membranes into microfluidics with specific applications in biosensing. One of the emerging needs in the domain of biosensing is the real-time measurement or the kinetics of detection. This has a far-reaching impact on diagnostics in the areas of healthcare and the environment. The next section articulates the motivation for developing nanoporous microfluidic biosensors. The next section addresses the types of biosensors and how label-free biosensors can be developed using nanoporous membranes in conjunction with microfluidics. The next section focuses on addressing the need for nanoporous membranes for detection of both small molecules and molecular entities. The next section identifies the types of nanoporous membranes that have been used for biosensing and analyzes the merits and disadvantages of the multiple material systems. The next section focuses on methods of fabrication and processing of the nanoporous membranes and the methods of integrating them into platforms relevant to biosensing applications. The last section focuses on a number of healthcare and environmental applications that employ nanoporous microfluidic biosensors. The most distinguishing feature of this chapter is that it evaluates a new class of miniaturized lab-on-a-chip platforms—the nanoporous membrane-based microfluidic biosensors. Contemporary leading reviews focus either on microfluidics or on application for biosensors. This chapter is the first attempt to look at a new class of biosensors that have been quietly emerging in the domain of lab-on-a-chip devices. 2.2 NEED FOR REAL-TIME MEASUREMENTS Monitoring human health for early detection of disease conditions or health disorders is vital for maintaining a healthy life. Many tissues, microorganisms, organelles, cell receptors, enzymes, antibodies, nucleic acids, biomolecules, and so on help determine

BASIC CONCEPTS OF BIOSENSORS

49

the physiological state of a disease condition. In addition, analysis of food and environment for perturbants such as pesticides and river water contaminants with harmful biomolecules has also become invaluable for health diagnosis. Thus, there is an ongoing need for rapid analysis, active continuous-time monitoring systems with substantial accuracy for detecting biomolecules. A “real-time” biosensor that detects the analytes of interest in a near-continuous-time manner plays an important role in effective data generation and data processing, supporting real-time decision making, and rapid manipulation. In order to meet these multiple requirements from multiple environments, one of the standard approaches that have been adopted is developing hybrid biochemical analysis systems. These multiscale biosensors are versatile because they can monitor specific analytes from a wide range of environments at ultralow concentrations. They comprise a combination of nanomaterials integrated with microfluidic capabilities. This approach is similar to that followed by the semiconductor industry in integrated circuits. Microfluidic research involves the study of several fluid manipulation, detection, and separation techniques. Often, these different components are integrated into essential electronics to develop a complete “on-chip” analysis system. Well-established fabrication techniques are adapted from the semiconductor industry such as micromachining, injection and replica molding, soft lithography, wet etching, and photolithography. These techniques enable miniaturization of fluid handling systems to palm-held “micrototal analysis systems” or “lab-on-a-chip” devices that can perform a myriad of diagnostic and analysis tasks associated with a standard clinical laboratory assay. Hence, a lot of the present research in this area focuses on the integration of these complex requirements of real-time measurements with on-chip detection capabilities to build multifunctional biosensors. One of the innovative approaches adopted to address these requirements has been the use of nanoporous membranes embedded into microfabricated structures to generate multiscale platforms. These platforms are integrated into the appropriate microfluidics to achieve size-based trapping of analytes of interest, which are then immobilized and interrogated using a number of label-free methods that have been discussed in the following sections of the chapter. Hence, nanoporous membrane embedded with a microfluidic device provides a powerful tool for real-time measurement of “lab-on-achip” biosensors. The following section explains the basic concepts of biosensors. 2.3 BASIC CONCEPTS OF BIOSENSORS A label-free biosensor is a means of detecting biological agents such as antibodies, nucleic acids, tissues, cells, microbes, and metabolites. The working principle consists of binding bioanalytes of interest to bioreceptors, which in turn modulate the physiochemical signal associated with binding. Later, the electrochemical or optical transducer captures and converts the physiochemical signal into an electrical signal. The variation in signal such as electric potential, current, conductance, impedance, intensity and phase of the electromagnetic radiation, mass, temperature, and viscosity is monitored. The analysis of variation in one or more of these parameters quantifies the presence or absence of bioagents. This quantification is achieved without using

50

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

Electric potential Electric current Electric conductance Electric impedance Intensity and phase of EM radiation Mass Temperature Viscosity

Electrical signal

Transducer

Signal processing

Transduction

Physiochemical signal Bioreceptors Enzymes Antibody Nucleic acid Tissue Microbial Polysaccharide

Recognition

Bioanalytes

FIGURE 2.1

Biomolecular complex

Schematic prototype of a standard biosensor.

FIGURE 2.2 Integration of bioelements and transducer to construct a nanoporous membrane-based biosensor.3

fluorescent tags or labels. Figure 2.1 demonstrates a schematic diagram of biosensors. In nanoporous membrane-based biosensors shown in Figure 2.2, the porous nanomembrane acts as an intermediate layer between biological agents and the physicochemical detector component, or biological agents and transducer are combined with a nanoporous membrane to construct a biosensor. In the area of nanomaterial-based biosensor, one of the genres of nanomaterials that are incorporated in the sensors is the nanomembrane. The huge interest in nanomembranes is driven by their many desirable properties, particularly the ability to tailor the size and structure and thereby optimize signal transduction properties for sensing systems. These nanomembrane-based electrical biosensors produce signal amplification, leading to lower limit of detection. Applications in both clinical and nonclinical environments are discussed below. 2.4 APPLICATIONS OF NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS Most common applications of nanoporous membranes are self-regulated drug delivery, biomolecular separation devices, and biosensors. In biosensors, often the

APPLICATIONS OF NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

51

nanoporous membrane is incorporated for biomolecular separation. These nanoporous membrane-based microfluidic biosensors are significantly smaller in size than conventional fluid manipulation systems, rendering them portable and extremely useful in the areas of nanobiotechnology, bioanalysis, pharmaceuticals, medicine, and diagnostics. Advances in nanotechnology have impacted research in biotechnology with the development of “smart devices” capable of molecular manipulation. Microfabricated bioanalytical devices offer highly efficient platforms for genomic, proteomic, and metabolic studies. The nanoporous membrane plays an important role in fabrication of biocompatible and cost-effective lab-on-a chip devices. Moreover, analogous to the miniaturization of computer chips, bioanalytical sensors are also undergoing reduction in their size. This offers the benefit of shorter reaction speeds with faster analysis times. Multiple active sites on the microfluidic chip offer parallel operation modes, thereby resulting in multiplexed analysis and higher throughput. These membrane-based biosensors can be costeffectively produced on a large scale with a promise of higher analysis rates and better efficiency, owing to the compactness and better process control; lower analyte consumption lowers the cost incurred on expensive reagents and is environment friendly during disposal. Figure 2.3 illustrates the flowchart for different applications of biosensors. The important part of membrane-based biosensor is their thin porous membrane. The need for this nanoporous membrane is explained in following sections. 2.4.1 Need for Nanoporous Membrane Overall sensitivity of a sensor depends on signal transduction and mass transport effect. Miniaturization of a sensor increases signal-to-noise (S/N) ratio, an inherent advantage for signal transduction. It has been reported that mass transport of analytical solution through the sensor surface plays an important role in determining sensitivity.4 Detection limit for bioassays depends upon the amount of biomolecule interaction with the sensor surface. Whitman’s group at Naval Research Laboratory, Washington, DC, reported

FIGURE 2.3

Flowchart representation of a biosensor.3

52

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

that femtomolar detection limits for bioassays are likely. They predicted that the limit of detection will be due to analyte transport limitation, not due to signal transduction limitation, and without directed transport of biomolecules, individual nanoscale sensors will be limited to picomolar-order sensitivity for practical timescales.4 Total flux to the sensor was studied as a function of sensor geometry and volumetric flow. Enhancing mass transport by conventional methods of decreasing height decreased volumetric flow rate, which in turn decreased total flux of the sensor. On the other hand, directly injecting the analyte into the sensor rather than merely streaming it past increased mass transport effects, which in turn increased total flux of the sensor.4 It was found that the sensor flux could be increased by using a nanoporous membrane. Thus, the sensitivity of the sensor can be enhanced by incorporating nanoporous membrane into the microfluidic biosensors. 2.4.2 Efficient Size Sorting Membrane-based technology has been identified as a useful method for the separation of biomaterials including viruses, owing to its efficiency, ease of implementation, and cost effectiveness. Due to nanometer-sized domain, and relatively thin membrane thickness and narrow size distribution, the nanoporous membrane shows high flux, mechanical strength, and high selectivity for biomaterial separation. Comparative studies were performed between track-etched polycarbonate (PC) and anodized aluminum oxide (AAO) membranes with a uniform pore size for virus separation; it was found that the nanoporous AAO membrane yields an excellent selectivity and high solution flux. 2.4.3 Design Considerations for Molecular Sorting One of the important aspects of the porous membrane typically exploited in microfluidic devices is the sorting of biomolecules on the basis of size from different bioanalytes for downstream analysis of the bioagent on biosensor platform. Depending on size and diffusion rate, pore size is determined. Adiga et al. explained that the nanoporous membrane is evaluated in terms of porosity of the membrane that is defined by solvent flux through unit area of the membrane under a unit pressure difference.1 !
 r2p Ak Lp ¼ 8h Dx where rp is the pore radius, 8h is the solvent viscosity, and (Ak/Dx) is the ratio of surface porosity to the pore length. Earlier, electrophoresis was used as a mechanism for the separation of biomolecules.5 Increasing demand of real-time measurement and miniaturization of biosensors to build a lab-on-a-chip device has led to the integration of the nanoporous membrane into the microfluidic device. Kuo et al. have built a simple, rapid prototyping of porous nanostructures inside the microchannels for the separation of

TYPES OF NANOPOROUS MATERIALS

FIGURE 2.4

53

Schematic diagram of a nanoporous structure in a microfluidic channel.6

DNA molecules.6 Periodic porous nanostructures with a cavity size of 300 nm and interconnecting pore of 30 nm were fabricated inside the microchannels. The fabrication process step using negative photoresist SU-8 is explained in Figure 2.4. First, silica colloidal is grown inside the SU-8 microchannel, then the space between colloidal crystals is filled with SU-8, which is then cured using ultraviolet light. The silica nanoparticles are finally removed using BOE, creating pores in the SU-8 structure. 2.5 TYPES OF NANOPOROUS MATERIALS Porous membranes can be broadly classified depending upon their intrinsic materials such as organic, inorganic, polymer, and composite membranes. Depending upon the barrier structure, membranes are nonporous or have micropores, mesopores, and macropores with pore diameters of <2 nm, 2–50 nm, and 50–500 nm, respectively.7 Lehrstuhl f€ur Technische Chemie II explained that porous barriers could be used for very precise continuous permselective separations based on subtle differences in size, shape, and/or functional groups.7 Many scientists are developing novel composites and polymers to tailor well-defined porous membranes in terms of pore size to increase their functionality and selectivity. Furthermore, depending on the cross section, membranes are classified as isotropic, anisotropic, bimultilayer, thin layer, or matrix of mixed composite. Yang et al. synthesized a nanoporous membrane for virus filtration with good dimensional stability under high pressures maintaining high selectivity.8 The membrane consists of a double layer. The upper layer is a nanoporous film with a pore size of 17 nm and a thickness of 160 nm, which was prepared by polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) copolymer. The lower layer consists of conventional microfiltration membrane to enhance mechanical strength. Phase separation, sol–gel method, interface reaction, stretching extrusion, track etching, and microfabrication technique can also be used for the classification of porous membranes. Inorganic membranes are made from oxides, ceramics, and metals, while the most commonly used polymeric materials are Nafion, polycarbonate, polyethylene terephthalate, or polysulfone. In comparison to polymer membranes, inorganic membranes are versatile, can be used at very high temperatures, and are much more resistant to chemical attack. Different strategies for fabricating and integrating the reviewed nanoporous membranes into sensing devices are explained briefly in the following section.

54

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

2.6 FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES INTO MICROFLUIDIC DEVICE 2.6.1 Anodization Anodization is an “electrochemical etching” process. Traditionally, nanoporous AAO membranes are fabricated by two ways: mild anodization (MA) or hard anodization (HA). The MA method produces self-ordered pore structures with a limitation in terms of slow process; the processing time requires several days, the oxide growth rate is 2–6 mm h
1, and the process requires a narrow range of processing conditions called self-ordering regimes. Being a slow process, MA is not suitable for mass production and industrial processes. Another technique that is widely found in the literature is HA. HA is performed at a high voltage by using sulfuric acid at relatively low temperatures and high current densities resulting in the rapid growth of a thick porous oxide layer of approximately 50–100 mm h
1 for various applications. Comparative study shows that porous oxide films formed by HA are disoriented compared to the pores formed by MA. Moreover, they are mechanically unstable due to a strong tendency to develop cracks under the influence of even weak mechanical forces. These aspects have caused serious problems in practical application of anodic films, especially in nanotechnology research. A limitation with HA process is in controlling important structural parameters, such as pore size, interpore distance, and the aspect ratio of the nanopores of the resulting alumina membranes. A comparison between MA and HA based on porosity, interpore distance, pore diameter, and pore density is provided in Table 2.1. A schematic diagram provided in Figure 2.5 illustrates the hexagonal structural layout of AAO by the anodization method. Researchers at Max Planck Institute of Microstructure Physics in Halle have developed a novel approach for structural engineering of nanoporous alumina using a pulse anodization method with oxalic acid.9 The process sequence of the pulse anodization method is shown in Figure 2.6. In the pulse anodization method, low and high potential pulses were applied alternatively to achieve mild and hard anodization conditions, respectively. The new process is an effective way of improving mechanical stability with a thickness of >100 mm of hard anodized alumina. In addition, it also provides a unique opportunity for producing individual alumina nanotubes with uniform diameter and length by taking advantage of cracking phenomena in the hard anodization process. The interpore distances of the AAO are approximately Dint ¼ 200–300 nm that can be achieved only by mild anodization

TABLE 2.1

Comparison Between MA and HA in 0.3 N H2C2O49

Parameter Porosity (P, %) Interpore distance (Dint, nm) Pore diameter (Dp, nm) Pore density (p; pores cm
2)

MA

HA

10 100 40 1.0  1010

3.3–3.4 220–300 49–59 1.3–1.9  109

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES

55

FIGURE 2.5 Schematic diagram demonstrating a porous oxide film produced by the anodization method.

FIGURE 2.6 Scheme for the fabrication of porous alumina with modulated pore diameters by a combination of MA and HA of a prepatterned aluminum substrate.9

56

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.7 Graph illustrating different conditions for synthesis of hard and mild anodization nanoporous membranes.9

processes. It offers substantial advantages over conventional anodization processes in terms of processing time, allowing 2500–3500% faster oxide growth with improved ordering of nanopores. Perfectly ordered alumina membranes with high aspect ratios (>1000) of uniform nanopores with periodically modulated diameters have been realized. In addition to sulfuric acid (H2SO4), oxalic acid (H2C2O4) and phosphoric acid are also used for making AAO. Figure 2.7 illustrates different conditions for the synthesis of hard and mild anodization nanoporous membranes. Shalini Prasad’s group from Portland State University demonstrated uniform formation of pores onto an alumina membrane fabricated using the anodization method.10 A combination of oxalic (0.1 M) and sulfuric acid (0.3 M) was used as an oxidizing acid. Alumina membrane fabricated had a thickness of 250 nm with a distance of separation between each pore of 10–15 nm, and the diameter of each pore was 200 mm. The volume of each pore was 8  10
21 mL. The SEM characterization of AAO is shown in Figure 2.8. 2.6.2 Ion Track Etching Track etching technology consists of chemical and physical modifications of thin films induced by energetic ion irradiation.11 The ion tracking process consists of irradiation of polymeric thin foils with heavy ions and subsequent chemical etching of the particle tracks. Cylindrical, conical, funnel-like, or cigar-like pores with diameters ranging from tens of nanometers to the micrometer range can be obtained.12 To obtain nanopores, the etching is performed in an ion accelerator and controlled by monitoring etchable ion tracks of swift heavy ions. The energy of these specific ion tracks should

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES

57

FIGURE 2.8 Scanning electron micrographs showing the alumina nanomembrane. (a) Density of the pores is clearly visible from these micrographs. Panel (b) shows the uniformity of the pores and the size distribution of each pore. (c and d) The pore goes all the way through the membrane and creates uniform wells.

be greater than 1 MeV per nucleon. Each ion passing though the thin polymeric film produced single ion track, which subsequently represents one pore. By controlling the number of ion tracks, one can produce 109 pores cm
2 to 1 pore per sample.13 A single ion is hit into a membrane to produce single pores. A metallic plate with a thickness of 0.2 mm and an aperture of 0.2 mm is inserted in front of the track-etched templates for support. A semiconductor detector is placed behind the sample to detect each ion passing through the aperture and then through the sample. At such low fluxes, the probability of an ion passing through the aperture is 1 event s
1.14 This gives enough time to the automated system to switch off the beam using a fast chopper after an ion hit is detected by the detector. The next step after irradiating the polymer is chemical etching of the ion track to produce pores of desired shape, with size up to 5–100 nm. Length to diameter ratio in the range 10–1000 can be easily achieved. Chemical composition of the chemical etchants, optimum irradiation, postirradiation treatments, and etching temperature conditions determine shape, size, and density of the pores. There are two parameters that determine the shape and size of the pore: 1. vb: the bulk etch rate (etching rate for nonirradiated material). 2. vt: the track etch rate (the etching rate along the ion track). Typically, the etching results in a conical or double conical pore (depending on whetheroneorbothfacesofthefoil,respectively,areexposed totheetchingbath)withan

58

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

opening angle a: tga ¼ vb/vt. Alternatively, in the case of high-selectivity etching conditions, that is, vt  vb, the shape of the pores can be approximated to be cylindrical. Polycarbonate and polyethylene terephthalate are some of the polymers used for synthesis of nanoporous membrane. Even though former polymers have lower chemical, mechanical, and thermal stabilities, both cylindrical and conical pores can be easily etched in comparison to other polymers. PC also has advantages in terms of dissolving capability into several organic solvents such as dichloromethane. Using this method, 100 nm pores were created in a PC membrane. The group of Voss, Germany, investigated asymmetric nanopores in PET and polyimide (Kapton) membranes.15 It was observed that the transient properties of the pores depend both on the chemical structure of the polymer and on the irradiation and etching procedures used. However, this method cannot be applied to metal foils, except for some special cases of amorphous “glassy” metals, because swift heavy ions passing through a metal do not generate tracks that could be etched to form holes. 2.6.3 Phase Separation The phase separation method is the most common method for preparation and production of polymeric porous membranes. There are three types of phase separation techniques that are typically used to generate a polymeric membrane, which include the wet phase separation method, the thermally induced phase separation method, and the nonsolvent-induced phase separation method. In the first method, a cast thin layer of a polymer solution is immersed in a liquid nonsolvent, which is miscible with solvent.16 The exchange of the solvent from a thin layer of polymer solution with a nonsolvent from the coagulation bath produces thermodynamic instability in ternary systems. The thermodynamic instability is resolved by separation into polymer-rich and polymer-lean phases.17 The polymer-rich phase forms a solid membrane matrix, while the polymer-lean phase leaves a porous structure by leaching out of the system. By varying the polymer concentration, the thickness of the solution and coagulation medium, and temperature, a wide variety of asymmetrical porous membranes with a very large variety of properties can be synthesized. The ternary systems most commonly comprise polymers such as cellulose acetate, polysulfone, poly(methyl methacrylate), polyamide, and polyurethane, solvents such as acetone, N,N-dimethylformamide, and acetamide, and nonsolvent such as water.18 Kawakami’s group from the Tokyo Metropolitan University fabricated threedimensional fluorinated polyimide microporous membranes of cylindrical structures by the wet phase inversion process, which is formed by a ternary system of polyimide–solvent–water.19 Moreover, microporous polystyrene and polycaprolactone (PCL) porous membranes were also synthesized by the phase separation process. PCL has several advantages, including low cost, biocompatibility, and biodegradability. Moreover, PCL is an FDA-approved material for implantable devices. Thus, it is a superior material to fabricate affordable devices. In polymeric systems, phase separation can also be induced by solvent evaporation, temperature, or addition of a nonsolvent. The process where optimization of porous

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES

59

property is performed by controlling the thermal conditions is called the thermally induced phase separation (TIPS) process.20 The TIPS technique has been utilized for making microporous materials such as porous membranes and foams from semicrystalline polymers. In the nonsolvent-induced phase separation method, the concentrations of polymer, solvent, and nonsolvent are highly critical for fabrication of the nanoporous membrane. Kuo’s group from the Center of Membrane Technology at Taiwan prepared a polyvinylidene fluoride (PVDF) microporous membrane by using alcohol as nonsolvent (coagulant).21 To prepare the PVDF membrane, the wet phase inversion process was carried out by using water and n-propanol as nonsolvents. An increase in the pore size of the PVDF membrane was observed with an increase in the immersion time in the n-propanol bath. Comparative studies showed that TIPS could help a homogeneous solution reach liquid–solid phase separation more quickly.22 Using this technique, phase-separated solutions spend less time in the liquid–liquid separation region. Fast solidification can avoid pore coalescence. Thus, porous membranes with high porosity can be prepared with TIPS. Moreover, nanoscale pore size can be obtained via NIPS. Therefore, nanoporous membranes with high porosity can be prepared via the combination of TIPS and NIPS. Recently, Kurt et al. developed a polymerization-induced phase separation (PIPS) technique.23 Porous membranes were developed from a monomer and solvent mixture using PIPS. This novel membrane has a potential application in flow-through biosensors based on protein or DNA microarrays. Depending on the requirements of the application, often high demands are put on specific properties such as the control over pore size, pore size distribution, morphology, and surface functionality, which is controlled by adjusting monomer concentration and UV intensity. 2.6.4 Lithography At present, novel nanoporous micromachined membranes using lithography techniques are being developed for biosensor applications. Ferrari’s group from the University of Illinois fabricated porous silicon membranes using simple standard lithography techniques. These membranes are highly reproducible, extremely stable, and have the ability to be integrated into the silicon-based platform technology. In addition, these membranes exhibit selective permeability and low biofouling. Filtration of biomolecules is desirable for biosensors. Membranes with pore sizes as small as 20 nm are available. Even so, the filtration at these dimensions is far from absolute. The use of ion track etching for the synthesis of membranes yields millipores with low porosities (<109 pores cm
2) and limited pore sizes, and the pores are randomly distributed across the surface. In comparison, the anodization technique yields alumina pores of higher densities, such as 1010 pores cm
2, but pore sizes greater than 20 nm, and the pore configurations and arrangements are difficult to control. The technology of micromachining allows fabrication of membranes with multiple pore configurations and arrangements. Several microfabrication methods have been used to create pore sizes in the tens of nanometers on silicon substrates using

60

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.9 Cross-sectional schematic depiction of the microfabrication process for nanoporous silicon membrane fabrication.24

photolithography and deposition/selective removal of sacrificial layers.24 Figure 2.9 demonstrates a schematic diagram for the microfabrication of nanoporous membranes with a pore thickness of 24.5 nm. In the above example of the lithography process, a nitride layer is grown onto a silicon wafer, which functions as an etch stop layer. On the top of the nitride layer, a base layer of polysilicon is deposited that acts as a structural support. The holes are then defined into the base layer by chlorine plasma. Later, thin sacrificial oxide is grown using thermal oxidation. The thickness of the sacrificial layer determines the final pore size. These pores are then completely filled with the polysilicon plug layer, which is planarization through plug layer to base. The final step is to deposit nitride layer and backside patterning. The protective, sacrificial, and etch stop layers are removed by etching in HF. Similarly, the process sequence shown above was also used by Desai’s group from the Boston University for fabrication of porous membranes with pore sizes ranging from 5 to 100 nm.25 In another method for fabrication of micro- and nanoporous membranes, pillar-like templates are used. These pillar templates are fabricated using photolithography and e-beam lithography. Grant’s group from the University of Texas at Austin built organic polystyrene (PS) and inorganic (ZnO) pillar arrays as a template and polysulfone (PSf) porous films were fabricated.26 Figure 2.10 illustrates the complete process sequence of template-based porous membrane fabrication.

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES

FIGURE 2.10 method.26

61

Schematic diagram for fabrication of a porous membrane using the template

2.6.5 Focus Ion Beam Etching A nanopore is most effective as a single-molecule detector when the diameter of the pore is close to the diameter of the molecule, typically 2–10 nm, being detected.27 Nanoporous membranes have pore sizes ranging from <1 to 100 nm in diameter. The focused ion beam (FIB) technique has been conventionally used for preparing nanopores in thin films.28 This technique offers very good resolution and is able to directly pattern arrays of well-defined pores. The anodization method has disadvantages in terms of fabrication of pore sizes greater than 20 nm and it is nearly impractical to etch a pore below 30 nm in diameter reproducibly in terms of size and shape using commercial FIB systems. It is possible to etch nanopores using a high-energy focused ion beam. However, Losic’s group from the University of South Australia at Adelaide, SA, overcame the system and pore size limitations and nanofabricated pore size less than 10 nm onto AAO porous membranes using the FIB technique.29 The FIB milling technique can be successfully used for removing the oxide barrier film and controlled pore opening of AAO to form a single nanopore or nanopore arrays. Figure 2.11 demonstrates the process sequence for fabricating the nanoporous membranes. Bezryadin’s group from the University of Illinois at Urbana-Champaign has demonstrated the fabrication of symmetric sub-5 nm nanopores using focused ion and electron beams.27 FIB scans reduce the size of the pore. During the sequence of ion scans, the gradual shrinking of the pore size and the change in pore geometry in a regular fashion are observed. The beam provides some mobility to the material of the membrane around the pore and allows the pores to shrink in diameter due to surface tension. This sculpting process also allows an array or a pattern consisting of multiple nanopores to be fine-tuned at the same time.

62

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.11 Schematic diagram demonstrating fabrication of a single pore to an array of nanopores onto the AAO membrane using the FIB technique.29

Similar to ion beam milling, electron beams are also used for micromachining finely precise nanopore arrays in the nitride membrane. Meller from the Boston University fabricated nanopores in thin Si3N4 films using the intense e-beam of field emission.30 Si3N4 films were formed using standard lithography techniques and low-pressure chemical vapor deposition (LPCVD). Different irradiation conditions enabled nanopore fabrication in the range of 2–20 nm with exceptional size control and greater than 0.5 nm variability, thus resulting in an effective nanopore thickness of 17 nm. Similarly, David C. Bell from the Center for Nanoscale Systems, Harvard University at Cambridge, fabricated small nanopores in SiN on the order of 1 nm diameter with great accuracy and reproducibility.31,32 2.6.6 Rapid Thermal Annealing Using the rapid thermal annealing (RTA) technique, Striemer et al. were able to fabricate ultrathin porous nanocrystalline silicon (pnc-Si) membranes.33 These pnc-Si porous membranes are approximately 10 nm in thickness. These nanomembranes are highly fragile and the technique is very complex; rapid thermal annealing is not commercially used for large-scale production. With RTA, the average pore size created varies approximately from 5 to 25 nm. The pore size distributions in pncSi membranes can be controlled through adjustment and by varying the temperature. In the RTA process, during crystallization, voids are often spontaneously formed as nanocrystals nucleate and it grows into a 15 nm thick amorphous silicon (a-Si) film. The process sequence for the fabrication of the ultrathin pnc-Si membrane is shown in Figure 2.12.

FABRICATION AND INTEGRATION OF NANOPOROUS MEMBRANES

63

FIGURE 2.12 Schematic block diagram representing formation of a nanoporous membrane using rapid thermal annealing.33

2.6.7 Sol–Gel Technique Sol–gel technology offers a cheap and quick alternative for producing bioactive porous surfaces for various biosensors. The sol–gel thin film technique offers a number of advantages including low-temperature processing, ease of fabrication, and precise microstructural and chemical control.34 Tailoring the thickness of the membrane, pore size, and density provides an additional advantage to biocompatibility. Highly porous, supercritically dried sol–gel low-density membranes formed using the sol–gel technique are called aerogels. These biocompatible membranes are commonly used in biosensors and electrochemical biosensors. Various types of inorganic aerogels such as silica, carbon, and alumina are fabricated. Similarly, inorganic aerogels such as SEAgel can also be synthesized. The starting materials used in the preparation of the sol are usually inorganic metal salts or metal organic compounds such as metal alkoxides [M(OR)n], where M represents a network forming element such as Si, Ti, Zr, Al, B, and so on and R is typically an alkyl group.35,36 The most commonly used precursors are tetramethyl orthosilicate (TMOS) and tetraethyl orthosilicate (TEOS) in the sol–gel process.37 The basic sol–gel reaction begins when metal alkoxide is mixed with water and a mutual alcoholic solvent in the presence of acid/base catalyst. Thin films can be produced on a piece of substrate by dip, spin, and spray coating. During the sol–gel transformation, the viscosity of the solution gradually increases as the sol becomes interconnected to form a rigid, porous network of gel.38 During the drying process at an ambient pressure, the solvent liquid is removed and substantial shrinkage occurs. Extracting the liquid component of a gel through supercritical drying, a highly porous

64

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

network known as an aerogel is produced. Silica aerogels are biocompatible materials mostly used for bioapplications. Power et al. from the University of Virginia at Charlottesville prepared sol–gel-based silica macroporous membranes. These porous membranes were on the order of 10–100 mm.39 These silica membranes were used to trap viral bacteria Escherichia coli (pET-gfp). Similarly, Tiwari et al. prepared a biopolymer–SiO2 nanocomposite aerogel. The resulting composite aerogel consists of a mesoporous material that was controlled by thermal curing.40 Recently, Li et al. synthesized a protein-encapsulated bioaerogel, in which a recombinant red fluorescent protein, DsRed, was chosen as a model protein.41 It was prepared using sol–gel polymerization of TEOS with an ionic liquid as the solvent and pore-forming agent. A bioaerogel formed showed high porosity. Recently, scientists from the University of Leeds, UK, for the first time explored the use of self-assembled peptide organogels and hydrogels as starting materials for the creation of new nanostructured aerogels.42 The novel aerogels impart biological-like functionality for sensing applications. Peptides such as P7-2, P9-2, P11-2, P24-2, and K2 were studied. 2.7 FUNCTIONALITY OF MEMBRANE IN BIOSENSORS Membranes with various pore sizes, lengths, morphologies, and densities have been synthesized from diverse materials for size exclusion-based separation. Specific bioagent immobilization and detection remains a great technical challenge in many biosensors. To achieve this, a material with controllable pore diameter, length, and surface chemistry is needed. Selective capture requires two steps: collection and immobilization. Membranes are well suited for this because of their enhanced probability of interaction of the surface with the liquid being analyzed. In order to create an excellent biosensor, the biological component has to be attached to transducers. This process is known as immobilization. Membrane entrapment, physical absorption, matrix entrapment, and covalent bonding are four ways to couple a biosensing element to the membrane (Table 2.2). The bioagents are bonded to the sensing element by one of these four ways. Figure 2.13 illustrates schematic block diagram demonstrating functionality of the membrane in the biosensor. The adsorption method is the simplest and involves minimal preparation. In this method, although the bonding between bioagent and sensing transducer is weak, it is extremely useful for exploratory research. The phenomenon of physical adsorption via TABLE 2.2

Different Immobilization Techniques

Immobilization technique Physical

Adsorption, entrapment, confinement, and encapsulation

Chemical

Covalent binding and cross-linking

FUNCTIONALITY OF MEMBRANE IN BIOSENSORS

65

FIGURE 2.13 Schematic block diagram demonstrating physical and chemical functionality of a nanoporous membrane in biosensors.3

van der Waals bonds is demonstrated by Prasad and coworkers from the Portland State University in nanomonitor protein biosensors.10 The AAO membrane was used to adsorb enzymes onto the sensing microelectrode array surfaces. Similarly, an antibody-based conductometric biosensor using porous filter membranes, developed by Muhammad-Tahir and Alocilja, has been shown to detect bacteria and bovine viral diarrhea virus antigens.43 In addition, Song et al. incorporated absorption techniques in biosensors to build membrane-based assay devices.44 Other forces such as hydrogen bonds, hydrophobic forces, and ionic forces are used to attach a biomaterial to the surface of the sensor. At a single instance, multiple forces can also be used in a single biosensor. However, the adsorption technique has a disadvantage: under mild conditions, through the pores the enzyme drains from the carrier and alters the change in pH or ionic strength. Entrapment of the bioagent in the nanomembrane creates a good biorecognition layer close to the transducer, which is mainly a gold microelectrode. The membrane separates the analyst and the bioelement. This facilitates a better biosensor. Kueng et al. immobilized a biorecognition element using an enzyme entrapment technique to integrate amperometric ATP microbiosensors.45 A dual microdisk electrode configuration was integrated to immobilize the enzymes at one of the microdisk electrodes. Shan et al. constructed a novel glucose biosensor by electrochemical entrapment of glucose oxidase (GOD) into porous poly(acrylonitrile-co-acrylic acid), which was

66

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

synthesized via radical polymerization of acrylonitrile and acrylic acid.46 Similarly, scientists from the Kyushu Institute of Technology in Japan used nanoporous template electrodes to enable efficient enzyme entrapment by simple physical adsorption.47 Template electrodes made of porous carbon were efficient nanomaterials for entrapment of bioagents. The integrated biosensor using entrapment provides an increased surface area and sensitivity to sense the biochemical reaction; however, the technique has also demonstrated problems such as an easy leakage, serious diffusion constraints, and lower stability48. In the microencapsulation technique, the porous entrapment scheme forms a porous encapsulation matrix around the biological analytes that helps binding to the sensor transducer. The porous membrane facilitates transfer of electrons and ions. Darder et al. demonstrated encapsulation of enzymes by alumina membranes of controlled pore sizes.49 These AAO porous membranes created by anodization were used for immobilization of the biologically active elements. Hexagonally structured AAO membranes allowed higher amounts of glucose oxidase (GOx) enzyme uptake in a thin film. This idea facilitated the construction of amperometric biosensors with the nanoporous Al2O3 membrane. Similarly, in choline biosensors, enzyme choline oxidase (ChOx) was immobilized and encapsulated in a hybrid mesoporous membrane with 12 nm pore diameter.50 Moreover, Kim et al. developed a glucose biosensor based on a sol–gel-derived zirconia/Nafion composite film as an encapsulation matrix.51 To conclude, in all these biosensors, the nanoporous structure film greatly enhances the active surface area available for protein immobilization. In covalent bonding techniques, biological elements are linked to biosensing membranes by strong covalent bonds. Sensing surfaces are treated with reactive chemical groups. De Stefano et al. created a covalent bond between the porous silicon surface and the biomolecules, which specifically recognize the unknown analytes.52 It has been demonstrated that porous silicon-based optical microsensors help detect L-glutamine from E. coli using the covalent bonding method. However, covalent bonding and cross-linking produce more stable immobilization, but may degrade the activity of enzyme by including drastic synthesis environment in the immobilization process.48 2.8 DETECTION MECHANISM Diagnostic biosensors are probably the largest area of research in the field of bionanotechnology. Over the past few years, many new ideas and technologies have been proposed in the literature. In this section, a comparative study of these detection mechanisms will be discussed. Each detection mechanism significantly differs in the design of the device, the fabrication methodology adopted for such a design, and the area of application. Each mechanism has the same goals in mind that include reducing the sensing elements to be equivalent to the size of the target species, improving the sensitivity, reducing the reagent volumes, offsetting the costs of the reagents, moving toward a real-time system to acquire the results and simultaneously getting a lower detection limit, miniaturizing the entire system, and improving

DETECTION MECHANISM

67

portability. Based on the basic principles of detection, biosensor devices can be broadly classified into two classes, namely, labeled detection and label-free detection. 2.8.1 Membrane-Based Labeled Detection A label is generally a chemical such as fluorophore, which is tagged along with the biochemical under observation for detection. In the labeled detection mechanism, as demonstrated in Figure 2.14, a biomolecule that does not react to incident light, but the chemical (label) reactive to light, is tagged to the biomolecule of interest. When the complex matrix of biomolecule and label is irradiated with light, the complex gets excited to give an output signal at the photodetector. Hence, detection of biomolecule of interest by indirect detection of the label is called labeled detection. The most popular labeled detection method is fluorescence detection and ELISA (enzymelinked immunosorbent assay). Immobilized biomolecules are emerging as popular analytical tools given their reusability and sensitivity. In a nanomembrane-based labeled detection biosensor, immobilization of the antibody is performed by simply inserting the porous nanomembrane into the reaction chamber. Membrane-based immunochemical methods are gaining wide acceptance as they offer the advantages of sensitivity, specificity, rapidity, simplicity, and cost effectiveness, which is important for routine testing. However, they have significant disadvantages that label-free detection has been striving to overcome and replace ELISA over the past decade. 2.8.2 Membrane-Based Label-Free Detection When the biomolecule can be directly detected without the help of external labels, it is called label-free detection. In the label-free mechanism, detection and quantification of the properties of interest of the biomolecule are achieved without the help of external labels. This has a huge advantage as it is possible to measure the direct interaction of the biomolecule with the substrate rather than its interaction with an

FIGURE 2.14 Schematic diagram demonstrating the labeled detection technique.

68

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

external label. It also eliminates multiple steps of attaching the label to the biomolecule and the control steps to eliminate nonspecific binding of labels. It also eliminates contamination due to external chemicals. The electrochemical or electrical method of detection is a good example of label-free detection wherein the electrical properties of the biomolecules are observed and correlated for analysis. Label-free detection has many advantages over labeled detection. It has lower cost per assay, lower contamination, higher sensitivity, and significantly shorter detection time. However, one of the important issues faced by label-free detection is the lack of throughput. Most biomolecules have different yet very similar physical characteristics when it comes to label-free detection. Because of the nature and size of the biomolecules, it was difficult to distinguish them using detection systems that were many orders of magnitude bigger. A large amount of sample volume is also required to achieve high sensitivity, which also proved impractical, as it is not possible to have tens of milliliters of blood every time the patient needs to be tested. These issues were solved with the advent of nanotechnology. With the advent of nanofabrication techniques and nanomaterials, device setups around the same size of the biomolecules could be designed. Much higher surface area of interaction with the biomolecules is possible, which enables high sensitivity in detection. Size matching of the measuring arrangement to the biomolecules significantly reduces the sample volume required and increases detection limit to very low analyte concentrations. Hence, the label-free detection technique has become the most active area of research in the field of proteomics in the past decade because of its promising trend to replace ELISA as the predominantly used technique for protein detection. Label-free nanoporous membrane-based biosensors can be classified on the basis of detection mechanism, namely, electrical detection, optical detection, and mechanical detection. Electrical biosensors can be further classified on the basis of the electrical measurement, which includes voltammetric, amperometric/coulometric, potentiometric, and impedance. 2.8.2.1 Voltammetry The current–potential relationship of an electrochemical cell provides the basis for voltammetric sensors. Amperometric sensors are also based on the current–potential relationship of the electrochemical cell, which can be considered a subclass of voltammetric sensors.53 In amperometric sensors, a fixed potential is applied to the electrochemical cell, and a corresponding current due to the reduction or oxidation reaction is obtained. This current can be used to quantify the species involved in the reaction. The key consideration of an amperometric sensor is that it operates at a fixed potential. However, a voltammetric sensor can operate in other modes such as linear or cyclic voltammetric modes. Consequently, the respective current–potential response for each mode will be different. This technique measures the current associated with electrons, which are on the surface of biomolecules during redox processes. Shimomura et al. using amperometric mechanism detected choline with enzyme immobilized in a hybrid mesoporous membrane.50 Such an ability of the hybrid mesoporous membrane F127M suggests great promise for effective immobilization of enzyme useful for electrochemical biosensors. Liu et al. developed a sensitive

DETECTION MECHANISM

69

amperometric biosensor based on gold nanoelectrode array (NEA). The gold nanoelectrode array was fabricated in the template of PC membranes.54 A conventional three-electrode system was used in all the above measurements. NEA and NEA/GOx were used as a working electrode, Ag/AgCl was used as a reference electrode, and a spiral platinum wire acted as a counter electrode. The enzyme electrode exhibits an excellent response performance to glucose with a linear range from 10
5 to 10
2 M and a fast response time of 8 s. Figure 2.15 demonstrates immobilization of the PC membrane on the Au electrode. 2.8.2.2 Potentiometry Potentiometric sensors are fundamentally based on the potential developed between two electrodes. When a redox reaction takes place at an electrode surface, a potential is developed at the electrode–electrolyte surface. This potential is found and used to characterize the activity of the species involved in the reaction. Electrodes in potentiometric sensors can be inert or active. An inert electrode merely provides the surface for an electron transfer process. However, an active electrode either donates or accepts ions in the reaction. One electrode is always used as the reference electrode to complete the circuitry for the potentiometric sensor. A noninterference half-cell reaction occurs on this electrode. Potentiometric or voltammetric techniques are generally not used for protein detection because most proteins cannot be detected based on oxidation states and redox reactions. It is quite difficult to exchange charges between proteins and a medium for a change in oxidation state. These methods are used to signify the importance of metals in detection, but are not actively used for detection in biosensors in general. Shishkanova et al. demonstrated functionalization of the PVC membrane with single-stranded oligonucleotides for a potentiometric biosensor.55 Reddy et al. estimated triglycerides by porous silicon-based potentiometric biosensors.56 Lipase, an enzyme that hydrolyzes triglycerides, was immobilized on PS and was thermally oxidized. Upon hydrolysis, the triglycerides result in the formation of fatty acids, which changes the pH of the solution. The enzyme

Au electrode

PC membrane Teflon cap

FIGURE 2.15 Integration of the PC membrane with the gold electrode.54

70

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

solution-oxidized PS–crystalline silicon structure was used to detect changes in pH during the hydrolysis of tributyrin as a shift in the capacitance–voltage (C–V) characteristics. 2.8.2.3 Impedance/Capacitance This method of detection employs true electrical parameters for detection and is based more on perturbations in the electrical components than on charge transfers or redox reactions. It calculates the amount of perturbation introduced by the biomolecule to the system more than the inherent change in the properties of the biomolecule. The most popular methodology used is electrochemical impedance spectroscopy (EIS). EIS is the plot of the overall impedance and phase between two electrodes with respect to the frequency applied to the circuit. This method of detection requires at least a twoelectrode arrangement for a closed circuit. Interdigitated electrodes or working electrode–counter electrode setups are the most popular and efficient electrode arrangements. Interdigitated electrodes offer much higher surface area of interaction than normal planar electrodes and therefore are more preferred. In a working electrode–counter electrode setup, the counter electrode is built with much higher surface area than the working electrode. This difference in surface area creates a large difference in the interaction between the electrodes and the biochemicals, thereby providing a highly sensitive detection. Using one of such two-electrode arrangements, the impedance or capacitance is measured between those electrodes. Capacitance is preferred as a parameter of detection due to its high accuracy and reliability. It can be greatly enhanced with an increase in surface area of interaction and plays a crucial role in determining selectivity. This method is well suited for detecting proteins, as it needs to measure the amount of perturbation introduced by the proteins to the system and not charge transfers or half-cell reactions. For such small protein biomolecules to perturb an electrical component, a large surface area for interaction is required. For just planar electrodes, such as planar interdigitated electrodes, there might not be enough surface area for the proteins to bind. The surface also needs to be highly sensitive to protein interaction for significant perturbation. Bothara and coworkers designed a membrane-based nanomonitor protein biosensor using an impedance and capacitive measurement technique.57Figure 2.16 shows an electric double-layer electrode on which the nanoporous structure exists. Electrons in the electrode cause the free ions in the solution to adsorb at the metal–liquid interface creating the double layer. Some of the solvated ions occupy these spaces and the proteins, which are also charged, reach close to these surfaces, which modify the double layer causing changes to the frequency at which the peaks in energy occur. Figure 2.17 illustrates a schematic diagram of an impedance/capacitance porous membrane-based biosensor. The measurement is achieved by the redox reactions at the surface for optimal charge transfer. The nonfaradaic conductance of the electrical double layer formed at the electrode surface is sensitive to reactions and is the basis of nanomonitors. This technique is advantageous since it does not require addition of any redox probes. Furthermore, conductance measurements at different bias voltages can reveal much information about dielectric and charge environment at the interface. The

DETECTION MECHANISM

71

FIGURE 2.16 Schematic diagram illustrating charge distribution across the liquid–electrode interface forming a double layer.57

nanomonitors comprise multiple sensing sites with each sensing site containing approximately a quarter million nanowells. Using the nanomonitor immunoassay technique, we were able to detect CRP and MPO with the present lower limit of detection at 10 and 20 ng mL
1, respectively. The upper limit of detection for both the antigens was 100 mg mL
1. The dynamic range for CRP was 100 mg mL
1 to 10 ng mL
1 and that for MPO was 100 mg mL
1 to 20 ng mL
1. 2.8.2.4 Optical Detection Optical detection techniques have been widely used in the field of biosensors for their high sensitivity. The most common examples include ELISA and fluorimetry, as discussed earlier. The detection mechanism involved is usually fluorescence or chemiluminescence. Fluorescence detection techniques are based on fluorescent markers that emit light at specific wavelengths when light is incident on it. The change in intensity of light emitted or its absolute value, as in fluorescence resonance energy transfer (FRET), determines occurrence of the binding reaction. Certain advances in this technique have been able to identify as low as single-molecule detection levels. Fluorescence-based detection has been used in chips designed for

72

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.17 The schematic diagram of an impedance/capacitance porous membranebased biosensor.57

the purpose of microarrays. These techniques are true lab-on-a-chip devices with integrated fluidic channels to direct the biomolecules. Chemiluminescence is the generation of energy in the form of light when a chemical reaction takes place. In synthetic compounds, such chemiluminescence takes place when a highly oxidized species such as peroxide emits energy during a chemical reaction. Sometimes, synthetic compounds are added to the biomolecule for detection, which in turn forms a conjugate that releases energy in the form of light. Bioluminescence is another type of luminescent technique, which has been reported by using the firefly luciferase/ luciferin as the synthetic compound. Surface plasmon resonance (SPR) is a technique that looks at the surface activity. In this technique, a longitudinal wave of a certain charge density is propagated along the surface of the metal and the dielectric. Due to the total internal reflection of light against materials such as gold or silver, the evanescent wave or field created at the surface penetrates the interface into the dense medium (metal). This evanescent light is able to couple with the free electrons on the surface, called plasmons, and this creates a resonance wave, which is recorded. Hence, in the presence of any biomolecule layer located on the surface of the metal, the SPR adsorption profile is obtained. Due to the properties of biomolecules, each SPR spectrum is different, hence making it a valid detection technique. The biggest asset

73

DETECTION MECHANISM

and the biggest drawback of optical detection system is light. Optical detection requires huge infrastructures, light sources, and other bulky instruments, making it very expensive and not portable. It is also very difficult to integrate and calibrate the light for detection. A tiny jitter in the system could make it completely inoperable and take hours for a highly skilled personnel to fix it. Specificity is also an issue with many optical detection techniques, but ELISA is considerably well established in this respect. De Stefano et al. demonstrated single-stranded DNA sensor using porous silicon surface.58 Using photochemical functionalization process, the porous silicon passivated the surface of optical biosensors. Fluorescence measurements have been used to investigate the stability of the DNA single strands bound to the nanostructured material. A dose–response curve in the 6–80 mM range for an optical label-free biosensor has been realized. Similarly, Bonanno and DeLouise demonstrated a membrane-based optical biosensor.59 The sandwich assay scheme incorporated in the sensor comprises a linking biotin/streptavidin to attach biotinylated anti-rabbit IgG receptor to detect rabbit IgG. The schematic diagram illustrating the working principle is shown in Figure 2.18. The final detection range of rabbit IgG was 0.07–3 mg mL
1 (0.23–9.8 mg mm
2). In the detection mechanism, a normal incident beam of white light (spot size of <13 mm2) was exposed to the sensor surface. The optical reflectance spectrum was measured. The optical shift due to specific binding interactions within the porous matrix was measured. 2.8.2.5 Mechanical Detection Mechanical detection for biochemical entities and reactions is generally through the use of nanomembrane and micro- or nanoscaled cantilever sensors. They can be further divided into mass sensors or stress sensors. Microgravimetric transducers

Target Rabbit IgG

Bioreceptor Biotin-α-rabbit IgG B

Streptavidin

B B B

B B

B

Biotin SiO2 APTMS

PSi

FIGURE 2.18 Schematic diagram of a silicon porous membrane-based optical label-free biosensor.59

74

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

monitor mass changes that occur during the binding of target analytes to the surfaceconfined recognition layer. Xue and Cui micropatterned carbon nanotube (CNT) and cantilever arrays fabricated with layer-by-layer nano-self-assembly that has application toward biosensors.60 In recent years, the areas of biomicroelectromechanical systems (BioMEMS) and nanotechnology have gained a high level of prominence and have become almost inseparable from biological applications including detection, diagnostics, therapeutics, and tissue engineering. In the nanoporous membrane-based mechanical biosensor, the porous nanomembrane acts as a filter for a bioagent in the microfluidic channel; later, the biomolecules are detected using a cantilever or a piezoelectric crystal. In a mass biosensor, change in the mass caused by chemical binding to small piezoelectric crystals is detected. Payen et al. demonstrated an MEMS rf-interrogated biosensor. Figure 2.19 demonstrated the step-by-step sequence for measuring the level of pH and eventually glucose concentration (Brix) in grapes.61 The biosensing structure consists of a microneedle, a functionalized gel, a tuned tank oscillator circuit, and a miniature antenna. A microneedle used was to puncture the organism. A two-membrane matrix was integrated into the membrane–filter system. The membrane system was used to separate the wet part of the hydrogel from the dry portion in which the tank oscillator circuit is located. The first membrane was used to allow diffusion of the grape juice into the hydrogel. The second compliant membrane was impermeable. Its function was to separate the wet region of the hydrogel from integrated circuit and antenna. A rigid porous membrane was incorporated to protect the hydrogel from biological fluids and to prevent contamination by molecules. The change in pH induces a swelling of the hydrogel. The hydrogel displacement depends on the hydrogel sensitivity and the flexible membranes. The membrane displacement induces a change in capacitance. The LC tank circuit frequency interrogation system determines the optimal frequency for data collection.

Bioelements (crop) Grapes, fruits, roots, etc.

Filter system

Fluidic delivery Microchannel Membrane matrix

Transducer

Biosensor Hydrogel

Transduction mechanism Capacitor

FIGURE 2.19

Process sequence for the detection of pH and glucose in grapes.61

POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM

75

2.9 POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM 2.9.1 E. coli Biosensor Meat products are increasingly contaminated by foodborne pathogens, thereby increasing product recalls in the United Stated. These products are contaminated from a number of sources, including the environment and the animal itself.62 E. coli serotype O157:H7 are harmful and deadliest bacteria found in meat and dairy products. These O157:H7 bacteria cause food poisoning, gastroenteritis, urinary tract infections, and neonatal meningitis. Thus, the microorganisms need to be detected at an early stage of infection in human beings. Many researchers are conducting experiments and spending billions of dollars to fabricate biosensors to detect the disease in a short time frame and with a high degree of accuracy. Wang et al. fabricated a novel nanoporous biosensor based on single-stranded DNA (ssDNA) probe functionalized AAO nanopore membranes for E. coli O157:H7 DNA detection.63 These membrane-based biosensors offered low detection limit for DNA in picomoles and rapid label-free and easy-to-use bacteria detection that holds the potential for future lab-on-a-chip devices. Similarly, disposable porous filter membranes were used by Abdel-Hamid et al. to develop a flow-through amperometric immunofiltration assay system for rapid detection of E. coli O157:H7.64 In the experiment, nylon membranes were used as a solid support for immobilization of antibodies. Two types of membranes, Biodyne B and C, were tested on which antibodies to E. coli O157:H7 were immobilized. These porous membranes were used for filtering the antigen-containing solution through the antibody-coated filter membrane. This results in excellent antigen–antibody binding, thereby significantly reducing the assay time. The effect of membrane pore sizes of 0.45, 1.2, and 3 mm determined the amount of immobilized anti-E. coli antibodies. The detection limit of E. coli O157:H7 cells is 100 cells mL
1 and the working range is 100–600 cells mL
1. A complete immunoassay is carried out in 30 min. Again, Liu et al. conjugated CdSe/ZnS core–shell dendron nanocrystals with the corresponding antibodies and then passed through the microporous membrane where they attached to the membrane antigen–antibody.65 The membrane antigen–antibody conjugated with the nanocrystals facilitated an efficient and stable photoluminescence. The biosensor built using this technique was used to detect not only E. coli but also hepatitis B with a limit of detection as low as 2.3 CFU mL
1 and 5 ng mL
1, respectively. Due to unique optical properties, nanocrystals were used as fluorescent labels. A microporous membrane was chosen for its mechanical strength and biocompatibility. The biosensor system comprises a flow chamber with an immunofilter interface with sensitive and robust dendron nanocrystals as the detection indicators. By using a matrix of the membrane, the scientists observed an increase in the efficiency of the immunoreaction between antibodies and pathogens, decrease in the detection time, and reduction in the detection limit. The pore size was optimized such that there was an excellent immunoreaction between antibodies and pathogens in a small hole and the liquid solution could be drained out. In the biosensor system, as

76

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.20 Schematic pictorial representation of a porous membrane-based biosensor for detection of E. coli bacteria.65

shown in Figure 2.20, the antibodies are immobilized onto the microporous membrane using covalent bonding. Later, the E. coli antigens are captured onto antibodies to form immunocomplexes. The immunocomplexes formed on the surface of the membrane continue to react with the dendron nanocrystal-conjugated antibodies and form “sandwich” immunocomplexes. By measuring the photoluminescence from the dendron nanocrystals, the targets, that is, E. coli antigens, are detected. Then, the nanocrystal-labeled antibody solution was injected to form a sandwich immunocomplex of immobilized antibody–E. coli O157:H7–nanocrystal-labeled antibody. The complex immunostructure using photoluminescence was measured by a spectrofluorometer. A common practice for the E. coli pathogen detection is to use an antibody-coated filter porous membrane to immobilize E. coli O157:H7, and different detection mechanism are used to detect the immunoreactions. 2.9.2 Salmonella enteritidis A commercially produced foodstuff containing raw eggs such as ice cream often gets contaminated with one of the deadliest pathogens S. enterica. These pathogens suppress the human immunity system. Therefore, scientists are developing and fabricating biosensors for detection of S. enteritidis. Zhang and Alocilja investigated a label-free DNA electrochemical biosensor for the detection of S. enteritidis.66 A nanoporous silicon-based DNA biosensor consists of a porous silicon surface, which was functionalized with DNA probes specific to the gene of S. enteritidis. Electrical property of DNA, redox indicators, and cyclic voltammetry were used for

POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM

77

the characterization of the biosensor. Porous silicon was fabricated using an anodization process in an electrochemical Teflon cell. The anodization was carried out by hydrofluoric acid. A uniform pore structure, with pore sizes ranging from 10 to 30 nm pores, was connected, and an interpore space of 10–30 nm was observed. These porous membranes were functionalized in two steps, which involved silanization of the chip surface followed by DNA probe immobilization. By immobilizing the specific DNA probe onto the PS layer, the biosensor had the ability to capture complementary DNA (cDNA). Finally, a cyclic voltammogram of target DNA concentrations on porous silicon chips was obtained. This DNA probe has high selectivity and affinity for the target DNA. The redox marker has a greater affinity for dsDNA, and therefore, a greater electrochemical response was observed when hybridization occurred. When the concentration of target DNA increased, the charge transfer between the redox marker and the PS electrode was enforced so that the peak current increased with DNA concentration. The detection limit of the PS-based label-free DNA biosensor was 1 ng mL
1. 2.9.3 Virus Detection Multiple diseases are caused by a lethal submicroscopic infectious agent known as virus. In the modern world, serious diseases such as Ebola, AIDS, avian influenza, and SARS are caused by viruses. Therefore, all around the world many researchers are developing biosensors to detect viruses. Riccardi et al. designed a novel label-free electrochemical detection system of DNA hybridization for detecting hepatitis C virus.67 A porous polycarbonate membrane was integrated to detect short sequence (18-mer) target DNA after diffusion. A voltammetric microbiosensor based on immobilization of the 18-mer HCV-1 DNA probe was applied in combination with SECM line scans to evaluate hybridization of DNA fragments diffusing through a porous polycarbonate membrane. Rossi et al. were able to detect bacteriophage virus MS2 using a porous silicon biosensor.68 For immobilization of the virus, covalent bioconjugation of antibodies inside porous silicon films was carried out. By fluorescence, 2  107 plaque-forming units per milliliter (pfu mL
1) were detectable. The nanoporous membrane had an average pore size of approximately 50 nm. The internal surfaces of porous silicon film are hydrogen terminated for bioconjugation chemistry. The hydrogen is then replaced by a functional organic group, which can be linked by a desired protein molecule using a cross-linker. Onto this linker, the unlabeled anti-MS2 antibodies are binded. The conjugation of either Alexa 488 to the antibodies or Alexa 532 fluorophores to MS2 viruses was carried out depending on the functionalization methods. Once again, the porous silicon surfaces were used to immobilize unlabeled anti-MS2 antibodies. The amounts of antibody and virus bound to the porous silicon surface were evaluated by fluorescence intensity at the emission maximum. Suggestions were made to improve the sensor performance by further optimization of porous layer structure and thickness. Gyurcs
anyi mentioned about “passive” nanopore counters for detecting icosahedral chlorella virus viruses.69 A real-time monitoring of the antibody–virus binding was able to detect concentrations as low as 5  107 particles mL
1 by a label-free technique. Reichmuth developed a lab-on-a-

78

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

chip device for rapid and portable diagnostics for detecting zoonotic diseases.70 A microchip-based electrophoretic immunoassay with an integrated nanoporous membrane incorporated into an open-channel electrophoresis and laser-induced fluorescence detection with a labeled antibody was carried out to detect influenza virus. The functionality of polymer membrane filtration eliminates the need for washing, commonly required in surface-based immunoassays, increasing the speed of the assay. Yang et al. were able to filter human rhinovirus type 14 major pathogen of the common cold in humans using nanoporous block copolymers.8 These nanoporous membranes showed excellent resistance to all organic solvents. 2.9.4 Glucose Detection For the treatment of diabetes mellitus, the amount of glucose present in a mammal’s blood is analyzed. Saha et al. fabricated a bioelectronic biosensor for the detection of glucose.71 A nanoporous cerium oxide (CeO2) thin film deposited on a platinum (Pt)coated glass plate using pulsed laser deposition (PLD) has been utilized for immobilization of glucose oxidase (GOx). Differential pulsed voltammetry (DPV) and optical measurements show that the GOx/CeO2/Pt bioelectrode exhibits linearity in glucose concentration ranging from 25 to 300 mg dL
1. Immobilization of GOx onto CeO2 matrix was achieved by electrostatic interaction of positively charged CeO2 and negatively charged GOx enzyme at pH 7.0. Fourier transform infrared (FTIR), atomic force microscopy (AFM), and DPV techniques helped investigation of the CeO2/Pt electrode and GOx/CeO2/Pt bioelectrodes. Wei et al. had a novel approach for detection of glucose oxidase.72 Glucose oxidase is entrapped into a complex nanocomposite film of chitosan/nanoporous ZrO2/multiwalled carbon nanotubes (MWNTs). Nanoporous ZrO2 helped to enhance the stability of the immobilized enzyme. A wide linear response range from 8 mmol L
1 to 3 mmol L
1 was obtained by an amperometric glucose biosensor. The ability of CNTs to promote the electron transfer of hydrogen peroxide (H2O2) suggested a promising idea for the construction of oxidase-based amperometric biosensors. The organic–inorganic CHIT/ZrO2/ MWNT nanocomposite had an added advantage in terms of toughness of CHIT and chemical and thermal stability of nanoporous ZrO2. Another approach by Wang et al. demonstrated a nonenzymatic electrochemical glucose sensor based on nanoporous PtPb networks.73 A reproducible one-step hydrothermal method helped grow PtPb networks on Ti substrates. Voltammetry and amperometric methods are used to evaluate the electrooxidation activities of the synthesized electrodes toward nonenzymatic glucose oxidation in neutral media in the absence and presence of chloride ions. The nanoporous PtPb electrodes have strong and sensitive current responses to glucose. The excellent performance of the PtPb electrode was achieved at an optimal PtPb composition of 50%. Not only thin films with pores are used as a platform in a sensing device but also beads with pores are used as a mechanism to trap GOx. Vamvakaki and Chaniotakis utilized porous silica beads with pore sizes of 10 nm for the immobilization and stabilization of the GOx with diameters on the order of 7 nm.74 The confinement of the GOx leads to enhanced enzyme stability. Silica beads and porous polymer beads were embedded for the

POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM

79

development of novel and highly stable glucose biosensor systems. In order to construct the GOx biosensor, the beads with the immobilized enzymes were placed on platinum electrodes through Nafion membranes. The glucose biosensor used the electrochemical measurement technique, which consists of a three-electrode system of silver/silver chloride double junction reference electrode and a platinum counter electrode. Silica beads have a well-defined pore size of 10 nm and a particle size of 60–80 mesh, while polymer beads (PLRP-S) with a spherical porous structure have pore sizes of 10 and 30 nm and an average particle diameter of 50–70 mm. The comparative study showed that the response time of the free enzyme biosensors was between 30 and 60 s and that of the bead-adsorbed enzyme biosensors was between 60 and 90 s. The free GOx biosensor had a remaining activity of 60% after 70 h of continuous operation, while bead-adsorbed enzyme biosensors did not lose any of their initial activity even after 70 h of continuous operation. In addition, it was evident in the experiments that the size matching between the pore size and the molecular diameter of the enzymes is very important to achieve high enzymatic activity and prevent enzyme leaching. Ekanayake et al. enhanced the adsorption of glucose oxidase by introducing artificial porosity into polypyrrole-based glucose biosensors.75 It also enhanced the performance of the sensor in terms of increasing high enzyme loading, stability, sensitivity, reproducibility, and repeatability. The immobilization was done by physical adsorption. Glutaraldehyde was used for cross-linking, while in enzyme adsorption, the response current is increased. The reaction that produces H2O2 for sensing is shown below. GOx

Glucose þ O2
! Gluconic acid þ H2 O2 H2 O2 ! O2 þ 2H þ þ 2e
PF6
dopant helps produce microporous PPy films. This improves the porous structure and the amperometric response of the electrodes. The thickness of the nanoporous film is about 30–40 mm. PF6
dopant introduced to the PPy led to a significant improvement in sensor characteristics. Most recently, Fink et al. were able to detect glucose using a reusable enzyme-modified ion track membrane sensor.76 The enzyme glucose oxidase covalently linked to nanopores by covalent linking. The detection range of glucose concentrations is between 10 mM and 1 M. The principle for glucose sensor is described in Figure 2.21. The glucose biosensor consists of PET foils of 4  106 cm
2 conical nanopores with a pore size of approximately 1.0  0.1 mm at the wide end and approximately 30  20 nm at the tip. The measurement compartment was separated by a sensing membrane and electrodes were inserted into both parts of the vessel. The voltage and transmission ion current were measured by applying sinusoidally shaped AC voltage of up to 5 Vpeak–peak operating at 1 Hz. On the basis of the current–voltage characteristics obtained, a diode-like rectification curve was obtained due to ion flow. The soluble reaction products produced due to enzyme reaction diffuse away from pore tip and avoid blocking of the pore, thereby rendering these sensors reusable. Using the

80

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

FIGURE 2.21

The pictorial representation of a glucose sensor.76

enzyme glucose oxidase (GOx) as the catalyst, the oxidation of glucose in the presence of oxygen is performed for glucose biosensing purposes. The glucose was oxidized by GOx and not by any of the other components of the device. The linearity between the measured ion current through the sensor and the applied glucose concentration was recorded. 2.9.5 Cholesterol Understanding and knowing cholesterol level in our bloodstream is extremely essential to reduce the risk of a heart attack or stroke. Li et al. have fabricated a cholesterol biosensor. Cholesterol oxidase was entrapped in a silicic sol–gel matrix.77 The half-life of the biosensor is about 35 days. The cholesterol biosensor has a high sensitivity and selectivity and can determine cholesterol oxidase ranging from 1  10
6 to 8  10
5 mol L
1 with a detection limit of 1.2  10
7 mol L
1. They also fabricated a selective cholesterol biosensor based on the composite film-modified electrode for amperometric detection.78 A concentration range of approximately 10
510
4 mol L
1 with a detection limit of 6  10
7 mol L
1 was determined. The excellent sensitivity and selectivity were attributed to the PB/PPy layer on the biosensor. Singh et al. immobilized cholesterol oxidase (ChOx) onto zinc oxide (ZnO) nanoporous thin films grown on gold surface.79 The porous thin film was fabricated using rf magnetron sputtering. A cyclic voltammetric measurement method was used for detection of ChOx and the sensitivity of detection was in range 25–400 mg dL
1. In this voltammetric method, the ChOx/ZnO/Au bioelectrode was found to detect cholesterol. Arya developed a ChOx biosensor using optical measurement. The ChOx molecule was covalently bonded to the sensing ODT electrode. The life of the ODT electrode-based biosensor was 2 months.80 Ansari et al. derived a sol–gel nanoporous cerium oxide film.81 A nanostructured cerium oxide (NS-CeO2) film deposited on the indium tin oxide (ITO)-coated glass substrate was used to detect cholesterol oxide.

POROUS MEMBRANE-BASED BIOSENSOR FOR DETECTION OF LIVING ORGANISM

81

2.9.6 New Diverse Sensors Recently, biomolecules were immobilized onto the sensing porous microelectrodes, which are used as a means of detecting bioagents. These microelectrodes are porous in nature, which help localize the bioagent for sensing. Song et al. developed a biosensor for diagnosis and monitoring of liver disease. An electrochemical array of nanoporous silicon electrodes is integrated for constructing these sensors.82 The liver plays a major role in metabolism, digestion, detoxification, and elimination of various substances from the body. Biomarkers such as cholesterol, bilirubin, and aminotransferases present in the serum help analyze the status of the liver. Using the silanization technique, sensitivities of the device were recorded to be 0.2656 mA mM
1 for cholesterol, 0.15354 mA mM
1 for bilirubin, 0.13698 mA (U l
1)
1 for alanine aminotransferase (ALT), and 0.45439 mA (U l
1)
1 for aspartate aminotransferase (AST). Compared to traditional analyte measurement procedures, the novel analytical device demonstrated high level of sensitivities for the analyses of multiple samples and analytes without a marked cross-interference effect. A single device containing the multiarray electrodes for sensing three different analytes is shown in Figure 2.22. A readout within minutes of application of microvolumes of a sample, reduced physical dimensions of the device, relative stability of the reagents used, and simple electronic component assembly useful for point-of-care biomarker liver analyses are some of the advantages of the liver biosensor discussed above.83 Measurement of urea in real urine samples is performed by a urea biosensor developed by Yang et al.84 The urease was immobilized onto nanoporous alumina membranes prepared by the two-step anodization method. Simple physical adsorption and cross-linking technique used for immobilization of these enzymes and piezoelectric mechanical detection mechanism were used for analysis. In the urea biosensor, urease was immobilized onto nanoporous alumina membranes, and the frequency responses of the ESPS/FIA detection system with the alumina electrode and the alumina/urease electrode were measured after the injection of 1.0 and 0.1 mM urea solutions. The catalytic reaction of the urease–urea system can be described by the equation below, thereby increasing the conductivity due to change of uncharged urea molecule to three ions. The frequency decrease is thereby attributed to the increase in conductivity. urease

NH2 CONH2 þ 2H2 O
! 2NH4 þ þ CO3 2


FIGURE 2.22

Schematic representation of a biosensor for liver diagnosis.83

82

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

The fabricated urea biosensor presented high-selectivity monitoring of urea, better reproducibility (SD ¼ 0.02, n ¼ 6), 30 s shorter response time, wider linear range from 0.5 mM to 3 mM, lower detection limit of 0.2 mM, and good longterm storage stability with about 76% of the enzymatic activity retained after 30 days. The activity of enzyme increases with increasing pore length for large pore size, while for correspondingly small pore size, enzymatic activity slightly depends on pore length.85 A nanoporous membrane can also be used to interface biological materials with a biohybrid system. Wolfrum et al. suspended nanoporous membranes as interfaces for neuronal biohybrid systems. Specification of the porosity parameters showed change in the transconductance of the nanopores and therefore helped control diffusion of molecules through the membranes. The alumina nanoporous membranes are biocompatible with both primary vertebrate and insect neurons. The thin aluminum film with 500 nm nanopores was integrated onto silicon nitride/silicon oxide to create stable suspended nanoporous Si3N4/SiO2 membranes by simple lithography anisotropic chemical silicon etching from the backside of the wafer in combination with anodization of thin aluminum films. Cells are genetically engineered and are placed on alumina membrane, which serves as a voltage-gated potassium ion channel. The cells showed adhesion and grew on the surface. The nanoporous membranes act as a cell interface and facilitate control of the cell environment with minute quantities of chemicals. 2.10 MICROFLUIDIC BIOSENSOR SYSTEMS A long-term goal in the field of microfluidics is to create integrated, portable clinical diagnostic devices for home and bedside use. Nanoporous membrane-based biosensors are integral components of a microfluidic system. This section describes the applications of microfluidic systems that have demonstrated the incorporation of nanoporous membranes to develop sensor systems. The chip used for this system is composed of an inexpensive and biocompatible polydimethylsiloxane (PDMS) layer. This overlays a Pyrex glass substrate that contains arrays of microelectrodes, which are used to detect the electrical signal in the biological environmental system. These chips consist of a microchannel to dispense or flow enzymes and analytes into the sensing area. These microfluidic biosensor systems provide low cost, portable, and high-throughput analytical systems. The approach applied for the construction of the microfluidic module provided a precise sample handling in terms of volume and flow rates, minimal dead volume at the inlet and outlet holes, that is, no sample losses during the analysis and approximately 100% waste disposal, and the ability for quick interchange of channels with different geometries and dimensions.86 Joo performed integration of a thin nanoporous platinum film into a microfluidic system for nonenzymatic electrochemical glucose sensing.87 The schematic diagram of a biosensor embedded into a microfluidic is shown in Figure 2.23.

MICROFLUIDIC BIOSENSOR SYSTEMS

83

FIGURE 2.23 Schematic representation of a microfluidic system.87

Glucose sensors cause intrinsic problems in manufacturing, storage, and distribution due to a serious, unacceptable dependence on temperature and humidity. Moreover, quality control difficulty and a short lifetime ultimately mean sizable costs. Therefore, the glucose biosensor was integrated into microfluidic systems. Programmed fluidic control of multiple reservoirs yields high-throughput analysis, automatic calibration, and multiple uses. Similarly, Metz et al. using micromachining and ion track technology fabricated polyimide microfluidic devices with integrated nanoporous filtration systems.88 Figure 2.24 illustrates a schematic block diagram of a microfluidic biosensor with embedded porous membrane. The device consists of inlet and outlet openings with the buried channels. The porous area is the filtration platform. One of the key advantages of a biosensor-based microfluidic is the limited contamination from outer atmosphere. Maeng et al. invented a novel microfluidic biosensor based on an electrical detection system for alpha-fetoprotein,89 whereas Goral et al. detected nucleic acid sequences using an electrochemical microfluidic biosensor.90 The microfluidic biosensor system was made up of PDMS and was integrated into an interdigitated ultramicroelectrode array, and microchannels were fabricated in a glass chip. Encapsulation of biosensor area into the closed system eliminated background signals to absolute nil and the IDUA responded in a highly reversible manner to the injection of various volumes and various concentrations of the electrochemical marker. Thus, the limit of detection was pushed to 1 fmol per assay and the dynamic range was 1–50 fmol. Similarly, Zaytseva et al. developed a microfluidic biosensor module with fluorescence detection for rapid and reliable identification of pathogenic organisms and

FIGURE 2.24 Schematic diagram demonstrating a polyamide microfluidic device for biosensor application.88

84

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

viruses.86 The microfluidic biosensor included a network of microchannels fabricated using soft lithography. As explained in the previous example, PDMS, which is the most commonly used polymeric substance, is used for fabricating the biomodule. Integration of the micro total analysis system into a sensing element enabled sample preparation and detection steps onto a single platform. A network of microchannels comprising the microbiosensor is modeled using computer-aided design software, nanofabricated in a silicon wafer by photolithography and etching techniques and molded into a PDMS elastomer by replica molding. In addition, Kwakye et al. developed an electrochemical microfluidic biosensor for the quantification of RNA. The concept was to demonstrate detection of dengue virus RNA.91 The detection mechanism was based on nucleic acid hybridization and liposome signal amplification. In addition, integration of a minipotentiostat, an interdigitated ultramicroelectrode array, and a microfluidic biosensor was successfully exhibited. Microchannel fluidic devices were fabricated using standard photolithography and dry etching. Using soft lithography, the microchannels were realized in PDMS covered with a glass slide consisting of interdigitated ultramicroelectrode arrays and packaged in a Plexiglas housing. Incorporating microfluidic systems yields a potentially inexpensive, portable, and disposable biosensor device that is easy to assemble and use. Yu et al. demonstrated a nanoporous impedance sensing transducer for fast bacteria patterning and detection at a low-frequency spectrum.92 A poly(ethylene glycol) (PEG) hydrogel microfluidic chip included patterned nanoporous aluminum oxide membrane (AOM) that allowed the detection limit to improve from 104 to around 102 CFU mL
1, in comparison to the conventional impedance measurement system. Vargas-Bernal presented multiple topologies based on the microfluidic device in order to optimize the biosensor design for evaluation and detection of pesticides.93 The aim was to reduce the size of sample, reduce both the analytical system size and the test time, automate the operation, and finally ease the transportation. These requirements were met by using smart matrices for enzyme immobilization, microfluidic systems, more sensitive enzymes, enzymes with more specific roles, and different detecting methods. 2.11 SUMMARY AND FUTURE PERSPECTIVE In conclusion, we believe that nanoporous membrane-based microfluidic systems have a wide range of applications from development of biosensors for detection of living microorganisms that cause harmful diseases to fabrication of membrane-based drug delivery systems. These nanoporous membranes have opened a new field in the area of membrane-based biosensors. These nanoporous microfluidic membranebased biosensors not only are tested in scientific laboratory but are also being tested in clinical environments. Due to these reasons, there is a continuous need for rapid improvement in sensitivity, stability, ease of formation of solvent-free membranes, robustness, arrays, and cost. These requirements have led to various scientists exploring new methods and strategies in membrane fabrication and development. Natural membranes are now being used for fabrication of biosensors. Most common,

REFERENCES

85

naturally found nanopores membranes are lipid membranes,94 and recently, microscopic diatoms are being investigated in biosensors. These biomimetic lipid membranes consist of natural greasy molecules that are found in real membranes and some molecules that are not usually found in the body. These membranes stick tightly to metal surfaces, and this makes them rugged enough to provide a good home for membrane proteins even in industrial applications. Diatoms are microscopic, single-celled algae that possess rigid cell walls composed of hard and porous amorphous silica. Diatom biosensors are devices incorporating a biological molecular recognition component connected to a transducer capable of outputting a signal proportional to the concentration of the molecule being sensed.95 These low-cost and largely available natural materials found in microorganisms are now being investigated as transducer elements for optical biosensors or as targeting microcapsules for drug delivery.96 In addition, instead of nanoporous films, porous electrodes are being designed and fabricated for building biosensors that possess high sensitivity, outstanding selectivity, repeatability, and cost effectiveness. Researchers are discovering porous structures of any form that are biocompatible and can be embedded into a transducer for forming biosensing systems. In conclusion, there is a paradigm shift from manufactured membranes toward using natural and biomimetic membranes. REFERENCES 1. Adiga, S.; Curtiss, L.; Elam, J.; Pellin, M.; Shih, C.-C.; Shih, C.-M.; Lin, S.-J.; Su, Y.-Y.; Gittard, S.; Zhang, J.; Narayan, R. Nanoporous materials for biomedical devices. JOM J. Miner. Met. and Mater. Soc. 2008, 60, 26–32. 2. Whitesides, G.M. The origins and the future of microfluidics. Nature 2006, 442, 368–373. 3. Mohanty, S.P.; Kougianos, E. Biosensors: a tutorial review. IEEE 2006, 6, 35–40. 4. Sheehan, P.E.; Whitman, L.J. Detection limits for nanoscale biosensors. Nano Lett. 2005, 5, 803–807. 5. Cheng, K.-L.; Sheng, Y.-J.; Jiang, S.; Tsao, H.-K. Electrophoretic size separation of particles in a periodically constricted microchannel. J. Chem. Phys. 2008, 128, 101101. 6. Kuo, C.-W.; Shiu, J.-Y.; Wei, K.H.; Chen, P. Monolithic integration of well-ordered nanoporous structures in the microfluidic channels for bioseparation. J. Chromatogr. A 2007, 1162, 175–179. 7. Ulbricht, M. Advanced functional polymer membranes. Polymer 2006, 47, 2217–2262. 8. Yang, S.Y.; Jinhwan, J.P.; Moonhor, Y.; Key, R.S.; Kim, J.J.K. Virus filtration membranes prepared from nanoporous block copolymers with good dimensional stability under high pressures and excellent solvent resistance. Adv. Funct. Mater. 2008, 18, 1371–1377. 9. Lee, W.; Ji, R.; Gosele, U.; Nielsch, K. Fast fabrication of long-range ordered porous alumina membranes by hard anodization. Nat. Mater. 2006, 5, 741–747. 10. Reddy, R.K.; Bothara, M.G.; Barrett, T.W.; Carruthers, J.; Prasad, S. Nanomonitors: protein biosensors for rapid analyte analysis. IEEE Sens. J. 2008, 8, 720–723. 11. Apel, P. Track etching technique in membrane technology. Radiat. Meas. 2001, 34, 559–566.

86

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

12. Awasthi, K.; Kulshrestha, V.; Acharya, N.K.; Singh, M.; Vijay, Y.K. Ion transport through track etched polypropylene membrane. Eur. Polym. J. 2006, 42, 883–887. 13. Dobrev, D.; Neumann, R.; Angert, N.; Vetter, J. Formation of metal membranes by direct duplication of etched ion-track templates. Appl. Phys. A: Mater. Sci. Process. 2003, 76, 787–790. 14. Enculescu, I. Nanowires and nanotubes prepared using ion track membranes as templates. Digest J. Nanomater. Biostruct. 2006, 1, 15–20. 15. Siwy, Z.; Apel, P.; Dobrev, D.; Neumann, R.; Spohr, R.; Trautmann, C.; Voss, K. Ion transport through asymmetric nanopores prepared by ion track etching. Nucl. Instrum. Methods Phys. Res. B: Beam Interact. Mater. Atoms 2003, 208, 143–148. 16. Vogrin, N.; Stropnik, C.; Musil, V.; Brumen, M. The wet phase separation: the effect of cast solution thickness on the appearance of macrovoids in the membrane forming ternary cellulose acetate/acetone/water system. J. Membr. Sci. 2002, 207, 139–141. 17. Lee, H.J.; Jung, B.; Kang, Y.S.; Lee, H. Phase separation of polymer casting solution by nonsolvent vapor. J. Membr. Sci. 2004, 245, 103–112. 18. Stropnik, C.; Kaiser, V. Polymeric membranes preparation by wet phase separation: mechanisms and elementary processes. Desalination 2002, 145, 1–10. 19. Taketani, Y.; Nagaoka, S.; Kawakami, H. Fabrication of three-dimensionally ordered microporous membrane by wet phase separation. J. Appl. Polym. Sci. 2004, 92, 3016–3021. 20. Li, D.; Krantz, W.B.; Greenberg, A.R.; Sani, R.L. Membrane formation via thermally induced phase separation (TIPS): model development and validation. J. Membr. Sci. 2006, 279, 50–60. 21. Kuo, C.-Y.; Lin, H.-N.; Tsai, H.-A.; Wang, D.-M.; Lai, J.-Y. Fabrication of a high hydrophobic PVDF membrane via nonsolvent induced phase separation. Desalination 2008, 233, 40–47. 22. Fu, X.Y.; Sotani, T.; Matsuyama, H. Effect of membrane preparation method on the outer surface roughness of cellulose acetate butyrate hollow fiber membrane. Desalination 2008, 233, 10–18. 23. Kurt, R.; Simon, L.; Penterman, R.; Peeters, E.; de Koning, H.; Broer, D.J. Control over the morphology of porous polymeric membranes for flow through biosensors. J. Membr. Sci. 2008, 321, 51–60. 24. Leoni, L.; Boiarski, A.; Desai, T.A. Characterization of nanoporous membranes for immunoisolation: diffusion properties and tissue effects. Biomed. Microdevices 2002, 4, 131–139. 25. Lopez, C.A.; Fleischman, A.J.; Roy, S.; Desai, T.A. Evaluation of silicon nanoporous membranes and ECM-based microenvironments on neurosecretory cells. Biomaterials 2006, 27, 3075–3083. 26. Yan, X.; Liu, G.; Dickey, M.; Willson, C.G. Preparation of porous polymer membranes using nano- or micro-pillar arrays as templates. Polymer 2004, 45, 8469–8474. 27. Lo, C.J.; Aref, T.; Bezryadin, A. Fabrication of symmetric sub-5 nm nanopores using focused ion and electron beams. Nanotechnology 2006, 17, 3264–3267. 28. Patterson, N.; Adams, D.P.; Hodges, V.C.; Vasile, M.J.; Michael, J.R.; Kotula, P.G. Controlled fabrication of nanopores using a direct focused ion beam approach with back face particle detection. Nanotechnology 2008, 19, 1–10.

REFERENCES

87

29. Lillo, M.; Losic, D. Ion-beam pore opening of porous anodic alumina: the formation of single nanopore and nanopore arrays. Mater. Lett. 2009, 63, 457–460. 30. Kim, M.J.; McNally, B.; Murata, K.; Meller, A. Characteristics of solid-state nanometre pores fabricated using a transmission electron microscope. Nanotechnology 2007, 18, 1–5. 31. Kim, M.J.; Wanunu, M.; Bell, D.C.; Meller, A. Rapid fabrication of uniformly sized nanopores and nanopore arrays for parallel DNA analysis. Adv. Mater. 2006, 18, 3149–3153. 32. Bell, D.C. Fabrication and application of nanopores using TEM, STEM and ion beams. Microsc. Microanal. 2008, 14, 244–245. 33. Striemer, C.C.; Gaborski, T.R.; McGrath, J.L.; Fauchet, P.M. Charge- and size-based separation of macromolecules using ultrathin silicon membranes. Nature 2007, 445, 749–753. 34. Balamurugan, A.; Kannan, S.; Rajeswari, S. Structural and electrochemical behaviour of sol–gel zirconia films on 316L stainless-steel in simulated body fluid environment. Mater. Lett. 2003, 57, 4202–4205. 35. D
ıaz-Garc
ıa, M.E.; La
ınn˜o, R.B. Molecular imprinting in sol–gel materials: recent developments and applications. Microchim. Acta 2005, 149, 19–36. 36. Kumar, A.; Gupta, R. Bioactive materials for biomedical applications using sol–gel technology. Biomed. Mater. 2008, 3, 1–16. 37. Bhatia, R.B.; Brinker, C.J.; Gupta, A.K.; Singh, A.K. Aqueous sol–gel process for protein encapsulation. Chem. Mater. 2000, 12, 2434–2441. 38. Collinson, M.M. Recent trends in analytical applications of organically modified silicate materials. Trends Anal. Chem. 2002, 21, 31–39. 39. Power, M.; Hosticka, B.; Black, E.; Daitch, C.; Norris, P. Aerogels as biosensors: viral particle detection by bacteria immobilized on large pore aerogel. J. Non-Crystalline Solids 2001, 285, 303–308. 40. Tiwari, A.; Mishra, A.P.; Dhakate, S.R.; Khan, R.; Shukla, S.K. Synthesis of electrically active biopolymer–SiO2 nanocomposite aerogel. Mater. Lett. 2007, 61, 4587–4590. 41. Li, Y.K.; Chou, M.J.; Wu, T.-Y.; Jinn, T.-R.; Chen-Yang, Y.W. A novel method for preparing a protein-encapsulated bioaerogel: using a red fluorescent protein as a model. Acta Biomater. 2008, 4, 725–732. 42. Scanlon, S.; Aggeli, A.; Boden, N.; Koopmans, R.; Brydson, R.; Rayner, C.M. Peptide aerogels comprising self-assembling nanofibrils. Micro Nano Lett. 2007, 2, 24–29. 43. Yueming, Z.; Chakrabartty, S.; Muhammad-Tahir, Z.; Pal, S.; Alocilja, E.C. Spatiotemporal processing for multichannel biosensors using support vector machines. IEEE Sens. J. 2006, 6, 1644–1651. 44. Song, X.; Kaylor, R. Membrane-Based Assay Device, Kimberly-Clark Worldwide, Inc: Neenah, WI, 2002; p 24. 45. Kueng, A.; Kranz, C.; Mizaikoff, B. Imaging of ATP membrane transport with dual microdisk electrodes and scanning electrochemical microscopy. Biosens. Bioelectron. 2005, 21, 346–353. 46. Shan, D.; He, Y.; Wang, S.; Xue, H.; Zheng, H. A porous poly(acrylonitrile-co-acrylic acid) film-based glucose biosensor constructed by electrochemical entrapment. Anal. Biochem. 2006, 356, 215–221.

88

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

47. Ekanayake, E.M.I.M.; Preethichandra, D.M.G.; Kaneto, K. Fabrication and characterization of nanostructured conducting polymer electrodes for glucose biosensor applications. In Second International Conference on Industrial and Information Systems, 2007, pp. 63–66. 48. Park, J.; Kim, H.K.; Son, Y. Glucose biosensor constructed from capped conducting microtubules of PEDOT. Sens. Actuators B: Chem. 2008, 133, 244–250. 49. Darder, M.; Aranda, P.; Hern
andez-V
elez, M.; Manova, E.; Ruiz-Hitzky, E. Encapsulation of enzymes in alumina membranes of controlled pore size. Thin Solid Films 2006, 495, 321–326. 50. Shimomura, T.; Itoh, T.; Sumiya, T.; Mizukami, F.; Ono, M. Amperometric determination of choline with enzyme immobilized in a hybrid mesoporous membrane. Talanta 2009, 78, 217–220. 51. Kim, H.-J.; Yoon, S.H.; Choi, H.N.; Lyu, Y.-K.; Lee, W.-Y. Amperometric glucose biosensor based on sol–gel-derived zirconia/Nafion composite film as encapsulation matrix. Bull. Korean Chem. Soc. 2006, 27, 63–70. 52. De Stefano, L.; Rotiroti, L.; Rendina, I.; Moretti, L.; Scognamiglio, V.; Rossi, M.; D’Auria, S. Porous silicon-based optical microsensor for the detection of L-glutamine. Biosens. Bioelectron. 2006, 21, 1664–1667. 53. Bronzino, J.D. The Biomedical Engineering Handbook, Boca Raton, FL: CRC Press, 2000. 54. Liu, Y.; Zhu, Y.; Zeng, Y.; Xu, F. An effective amperometric biosensor based on gold nanoelectrode arrays. Nanoscale Res. Lett. 2009, 4, 210–215. 55. Shishkanova, T.V.; Volf, R.; Krondak, M.; Kr
al, V. Functionalization of PVC membrane with ss oligonucleotides for a potentiometric biosensor. Biosens. Bioelectron. 2007, 22, 2712–2717. 56. Reddy, R.R.K.; Basu, I.; Bhattacharya, E.; Chadha, A. Estimation of triglycerides by a porous silicon based potentiometric biosensor. Curr. Appl. Phys. 2003, 3, 155–161. 57. Reddy, R.K.K.; Prasad, S.; Barrett, T.; Carruthers, J. Electrical immunoassays toward clinical diagnostics: identification of vulnerable cardiovascular plaque. JALA 2008, 33–39. 58. De Stefano, L.; Arcari, P.; Lamberti, A.; Sanges, C.; Rotiroti, L.; Rea, I.; Rendina, I. DNA optical detection based on porous silicon technology: from biosensors to biochips. Sensors 2007, 7, 214–221. 59. Bonanno, L.; DeLouise, L. Steric crowding effects on target detection in an affinity biosensor. Langmuir 2007, 23, 5817–5823. 60. Xue, W.; Cui, T. Carbon nanotube micropatterns and cantilever arrays fabricated with layer-by-layer nano self-assembly. Sens. Actuators A: Phys. 2007, 136, 510–517. 61. Payen, S.; Walther, D.C.; Pisano, A.P. A MEMS RF-interrogated biosensor. In Third IEEE/ EMBS Special Topic Conference on Microtechnology in Medicine and Biology, 2005, pp. 76–79. 62. Alocilja, E.C. Biosensors for detecting pathogenic bacteria in the meat industry. Meat Biotechnol. 2008, 335–359. 63. Wang, L.; Liu, Q.; Hu, Z.; Zhang, Y.; Wu, C.; Yang, M.; Wang, P. A novel electrochemical biosensor based on dynamic polymerase-extending hybridization for E. coli O157:H7 DNA detection. Talanta 2009, 78, 647–652. 64. Abdel-Hamid, I.; Ivnitski, D.; Atanasov, P.; Wilkins, E. Flow-through immunofiltration assay system for rapid detection of E. coli O157:H7. Biosens. Bioelectron. 1999, 14, 309–316.

REFERENCES

89

65. Liu, Y.; Brandon, R.; Cate, M. Sensitive and rapid detection of pathogens using luminescent CdSe/ZnS dendron-nanocrystals and a porous membrane immunofilter. Anal. Chem. 2007, 79, 8796–8802. 66. Zhang, D.; Alocilja, E.C. Characterization of nanoporous silicon-based DNA biosensor for the detection of Salmonella enteritidis. IEEE Sens. J. 2008, 8, 775–780. 67. Riccardi, C. d.S.; Kranz, C.; Kowalik, J.; Yamanaka, H.; Mizaikoff, B.; Josowicz, M. Label-free DNA detection of hepatitis C virus based on modified conducting polypyrrole films at microelectrodes and atomic force microscopy tip-integrated electrodes. Anal. Chem. 2008, 80, 237–245. 68. Rossi, A.M.; Wang, L.; Reipa, V.; Murphy, T.E. Porous silicon biosensor for detection of viruses. Biosens. Bioelectron. 2007, 23, 741–745. 69. Gyurcs
anyi, R.E. Chemically-modified nanopores for sensing. Trends Anal. Chem. 2008, 27, 627–639. 70. Reichmuth, D.S.; Wang, S.K.; Barrett, L.M.; Throckmorton, D.J.; Einfeld, W.; Singh, A.K. Rapid microchip-based electrophoretic immunoassays for the detection of swine influenza virus. Lab Chip 2008, 8, 1319–1324. 71. Saha, S.; Arya, S.K.; Singh, S.P.; Sreenivas, K.; Malhotra, B.D.; Gupta, V. Nanoporous cerium oxide thin film for glucose biosensor. Biosens. Bioelectron. 2009, 24, 2040–2045. 72. Wei, W.-Z.; Zhai, X.-R.; Zeng, J.-X.; Gao, Y.-P.; Gong, S.-G. New amperometric glucose biosensor by entrapping glucose oxidase into chitosan/nanoporous ZrO2/multiwalled carbon nanotubes nanocomposite film. J. Central South Univ. Technol. 2007, 14, 73–77. 73. Wang, J.; Thomas, D.F.; Chen, A. Nonenzymatic electrochemical glucose sensor based on nanoporous PtPb networks. Anal. Chem. 2008, 80, 997–1004. 74. Vamvakaki, V.; Chaniotakis, N.A. Immobilization of enzymes into nanocavities for the improvement of biosensor stability. Biosens. Bioelectron. 2007, 22, 2650–2655. 75. Ekanayake, E.M.I.M.; Preethichandra, D.M.G.; Kaneto, K. Polypyrrole nanotube array sensor for enhanced adsorption of glucose oxidase in glucose biosensors. Biosens. Bioelectron. 2007, 23, 107–113. 76. Fink, D.; Klinkovich, I.; Bukelman, O.; Marks, R.S.; Kiv, A.; Fuks, D.; Fahrner, W.R.; Alfonta, L. Glucose determination using a re-usable enzyme-modified ion track membrane sensor. Biosens. Bioelectron. 2009, 2418, 2702–2708. 77. Li, J.; Peng, T.; Peng, Y. A cholesterol biosensor based on entrapment of cholesterol oxidase in a silicic sol–gel matrix at a Prussian Blue modified electrode. Electroanalysis 2003, 15, 1031–1037. 78. Li, J.-P.; Gu, H.-N. A selective cholesterol biosensor based on composite film modified electrode for amperometric detection. J. Chin. Chem. Soc. 2006, 53, 575–582. 79. Singh, S.P.; Sunil, K.A.; Pratibha, P.; Malhotra, B.D.; Shibu, S.; Sreenivas, K.; Vinay, G. Cholesterol biosensor based on rf sputtered zinc oxide nanoporous thin film. Appl. Phys. Lett. 2007, 91, 063901. 80. Arya, S.K.; Solanki, P.R.; Singh, S.P.; Kaneto, K.; Pandey, M.K.; Datta, M.; Malhotra, B.D. Poly-(3-hexylthiophene) self-assembled monolayer based cholesterol biosensor using surface plasmon resonance technique. Biosens. Bioelectron. 2007, 22, 2516–2524. 81. Ansari, A.A.; Kaushik, A.; Solanki, P.R.; Malhotra, B.D. Sol–gel derived nanoporous cerium oxide film for application to cholesterol biosensor. Electrochem. Commun. 2008, 10, 1246–1249.

90

NANOPOROUS MEMBRANE-BASED MICROFLUIDIC BIOSENSORS

82. Song, M.-J.; Yun, D.-H.; Min, N.-K.; Hong, S.-I. Electrochemical biosensor array for liver diagnosis using silanization technique on nanoporous silicon electrode. J. Biosci. Bioeng. 2007, 103, 32–37. 83. Song, M.-J.; Yun, D.-H.; Min, N.-K.; Hong, S.-I. Electrochemical biosensor array for liver diagnosis using a silanization technique on nano-porous silicon electrode. In Nanotechnology Materials and Devices Conference, NMDC 2006, IEEE, 2006, pp. 362–363. 84. Yang, Z.; Si, S.; Dai, H.; Zhang, C. Piezoelectric urea biosensor based on immobilization of urease onto nanoporous alumina membranes. Biosens. Bioelectron. 2007, 22, 3283–3287. 85. Yang, Z.; Si, S.; Zhang, C. Study on the activity and stability of urease immobilized onto nanoporous alumina membranes. Microporous Mesoporous Mater. 2008, 111, 359–366. 86. Zaytseva, N.V.; Goral, V.N.; Montagna, R.A.; Baeumner, A.J. Development of a microfluidic biosensor module for pathogen detection. Lab Chip 2005, 5, 805–811. 87. Joo, S.; Park, S.; Chung, T.D.; Kim, H.C. Integration of a nanoporous platinum thin film into a microfluidic system for non-enzymatic electrochemical glucose sensing. Anal. Sci. 2007, 23, 277–281. 88. Metz, S.; Trautmann, C.; Bertsch, A.; Renaud, P. Polyimide microfluidic devices with integrated nanoporous filtration areas manufactured by micromachining and ion track technology. J. Micromech. Microeng. 2004, 14, 324–331. 89. Maeng, J.-H.; Lee, B.-C.; Ko, Y.-J.; Cho, W.; Ahn, Y.; Cho, N.-G.; Lee, S.-H.; Hwang, S.Y. A novel microfluidic biosensor based on an electrical detection system for alpha-fetoprotein. Biosens. Bioelectron. 2007, 23, 1319–1325. 90. Goral, V.N.; Zaytseva, N.V.; Baeumner, A.J. Electrochemical microfluidic biosensor for the detection of nucleic acid sequences. Lab Chip 2006, 6, 414–421. 91. Kwakye, S.; Goral, V.N.; Baeumner, A.J. Electrochemical microfluidic biosensor for nucleic acid detection with integrated minipotentiostat. Biosens. Bioelectron. 2006, 21, 2217–2223. 92. Yu, J.; Liu, Z.; Liu, Q.; Yuen, K.T.; Mak, A.F.T.; Yang, M.; Leung, P. A polyethylene glycol (PEG) microfluidic chip with nanostructures for bacteria rapid patterning and detection. Sens. Actuators A: Phys. 2009, 154(2), 288–294. 93. Vargas-Bernal, R. Topologies based on microfluidics of pesticides biosensors. In Proceedings of the Electronics, Robotics and Automotive Mechanics Conference, IEEE Computer Society, 2007. 94. Nikolelis, D.P.; Raftopoulou, G.; Nikoleli, G.-P.; Simantiraki, M. Stabilized lipid membrane based biosensors with incorporated enzyme for repetitive uses. Electroanalysis 2006, 18, 2467–2474. 95. Parkinson, J.; Gordon, R. Beyond micromachining: the potential of diatoms. Trends Biotechnol. 1999, 17, 190–196. 96. De Stefano, L.; Lamberti, A.; Rotiroti, L.; De Stefano, M. Interfacing the nanostructured biosilica microshells of the marine diatom Coscinodiscus wailesii with biological matter. Acta Biomater. 2008, 4, 126–130.

3 NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS GIOVANNA MARRAZZA Dipartimento di Chimica, Univesit a di Firenze, Via della Lastruccia, Sesto Fiorentino, Italy

3.1 INTRODUCTION Over the past decade, many important advances have been made in the use of nanotechnology for biomolecular detection. The use of nanoscale materials for biosensing has seen explosive growth in recent years following the discovery of carbon nanotubes (CNTs) by Sumio Ijima in 1991. Recently, many advances were achieved in the electrochemical and optical detection of DNA and immunoreactions, through the use of innovative detection schemes (e.g., microfluidic platform) and new materials, particularly the use of nanoparticles (NPs), nanotubes and nanowires. This advanced technology has been extended throughout the field of biosensors and biochips. Specifically, nanoparticles, made from metals, semiconductor, carbon, and polymeric materials, have been widely investigated to enhance the reaction signal of bioreceptors such as enzymes, antibodies, and oligonucleotides. Use of nanoparticle labels has proved to be particularly advantageous in sensing and biosensing applications due to the fact that single biorecognition events (i.e., DNA hybridization or immunoreaction) are typically translated into a significant effect on its optical (change of the light absorption or emission) or electrochemical properties (oxidation or reduction current) onto a transducing platform, offering novel options for bioanalysis. Moreover, the application of nanoparticles in biosensors strongly relates to their properties that derive to a certain extent from synthesis and later modifications (chemical and biological).

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

91

92

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

Although the use of NPs in bioanalysis is a recent area of research, there are many publications on their use as immobilization platform or labels for detecting numerous analytes. In the past few years, several excellent reviews have been published on the application of nanoparticles1,2 and particularly on the use of metal nanoparticles (gold nanoparticles)3,4 for the improvement of biosensing performance. Tamanaha et al.5 report an overview of the various approaches developed for magnetic labeling and detection as applied to biosensing. Moreover, nanoparticles have been applied widely to microanalytical systems (lab on a chip). For a biosensor to be used outside the laboratory, it either has to be as simple or automated with regard to sample processing and reagent addition. Automation and miniaturization of biological analytical techniques, as well as the development of on-line and remote measurement equipment, can be achieved through biosensor technology. Microfluidic systems are well recognized for their ability to move small volumes of fluids through different processes and over a sensing surface. The use of microfluidic technology presents many advantages, for example, cost reducing because of reduced fabrication expense and decreased requirements for costly reagents, reducing process and assay times that are proportional to liquid volumes. Moreover, the microfluidic devices can be realized in different steps of analytical procedure: (1) the microfluidic devices can be used for target preconcentration and target separation from other sample components, (2) the sample pretreatment (e.g., cell disruption or sample homogenization) can be performed in a microfluidic system, (3) both active and passive mixers are available for combining sample and assay reagents, (4) techniques for using solid-phase materials for separations in microflow have been identified, and finally (5) on-chip temperature control is available for temperature-dependent reactions such as PCR or simply for maintaining stability of the system in harsh environments. Therefore, microfluidic devices have found great application in the fields of biosensor technology. An overview has been published by Merkoci and coworkers on important aspects of microfluidic chip platforms as new materials for electrochemical sensing. It describes important trends in constructing detectors and their electronics on microfluidic chip platforms, the importance of selecting appropriate detector materials, and the different detection modes in on-chip amperometry, conductometry, and potentiometry.6 Rios et al. have presented a general overview on the potential of analytical microsystems. They discuss the issues involved in the analytical process and the different steps involved in chemical analysis. Moreover, they identify challenges in applying analytical microsystems to these uses.7 Synergism between nanoparticle-based and microfluidics technologies may bring to new miniaturized analytical devices for multiple targets in bioanalytical assays.8 Choi et al. have reviewed important nanotechnologies such as the application of nanoparticles for the detection of biomolecules, the immobilization of biomolecules at nanoscale, nanopatterning technologies, and the microfluidic system for molecular diagnosis.9 In a recent study, Ligler has discussed emerging science and technology that will enable the creation of more efficient application-specific optical biosensors.

INTRODUCTION

93

Biological recognition and signal amplification strategies, nanotechnology for geometric control of the biochemistry and signal enhancement, microfluidic for automated reagent delivery and reaction control, and emergence of optical elements amenable to improved systems integration will play a critical role in this evolution. The use of nanoparticles therefore allows miniaturization of biosensor, development of microfluidic systems, and increase in the sensitivity of bioassays.10 Microfluidic approaches provide one of the most promising strategies to interface nanoengineered biosensors in a wide spectrum of clinical and biomedical applications. Most biomedical samples are naturally in a liquid environment, so the sensors must be combined with compelling fluid handling systems. Appropriate microfluidic delivery systems can be used to eliminate contamination, minimize analysis times, and enable portable systems. Here, without pretending to being exhaustive, the most recent applications of nanoparticles in microfluidic format for electrochemical and optical affinity biosensors have been reported, highlighting some of their technical challenges and the new trends by means of a set of selected recent applications. 3.1.1 Microfluidic in Bioassays In recent years, the need for microfluidic devices has driven their continuous development. Microfluidic-based microchips can improve analytic efficiency by reducing analysis time and sample volume, increase both sensitivity and selectivity, and allow for miniaturization of analytic devices. Because of these advantages, microchips are widely used in clinical diagnosis, environmental toxicant detection, bimolecular separation, and cell handling systems. Microfluidic involves the manipulation, transport, and analysis of fluids in micrometer-sized channels. Integration into flow injection systems, capillary electrophoresis, and microfluidic platforms is just the latest, logical step in the direction of automation. All these trends are important and should occur in parallel in the future development of bioassay methods for medical and clinical applications. One of the most important issues to develop microfluidic devices is the selection of material because surface effects become enlarged as the device is miniaturized. The materials for a microfluidic biodevice ought to be considered by the following issues as well as process ability: (1) chemical stability; (2) price, for disposable device to minimize contamination; (3) surface properties for biofouling; and (4) thermal stability for nucleic acid amplification such as PCR. Other properties, like optical transparency, can be preferred for testing modules. For example, for optical transducers, glass-utilizing devices are preferred over silicon devices owing to their optical properties. Silicon, glass, and polymers are the three main types of materials used for microfluidic fabrication. Although metals are one of the most widely used materials in industries, many limitations in micromachining prevented the extensive use of metal. The micro- and nanodimensions required by these devices can only be easily fabricated with semiconductor technology; thus, silicon became one of the first

94

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

materials to be used in the early 1990s. However, silicon is opaque and this prevented the use of fluorescent labels for detection, which are very popular with immunoassays. Biological molecules also tend to adsorb to silicon surfaces and these limitations prompted the search of other fabrication materials. Naturally, glass became the next material as it is transparent to nearly all absorption and emission wavelengths of fluorescent labels. However, the difficult fabrication techniques and toxic chemicals involved did not make glass a popular choice among researchers and manufacturers. Recently, researchers have turned their attention to the use of polymers. Polymers offer the advantages of being optically clear, non-toxic, and low cost. In addition, easy fabrication techniques and a variety of surface modification methods are available to improve the efficiency of these devices. Polycarbonate, polymethylmethacrylate (PMMA), polyethylene, polypropylene, and polystyrene are some examples of polymers used widely in all fields of research and industries. One of the most extensively used polymers in the past few years is polydimethylsiloxane, also known as PDMS. The elastomer PDMS has attracted attention as a material suitable for the easy and rapid fabrication of microfluidic devices using soft lithography. PDMS has a number of advantages: (1) features on the micrometer scale can be reproduced with high accuracy, (2) it is optically transparent down to 280 nm, (3) it cures at low temperatures, (4) it is not toxic, and (5) it can seal reversibly to itself and a range of other materials by making molecular (van der Waals) contact with the surface, or it can seal irreversibly after exposure to an air plasma by formation of covalent bonds. These characteristics made PDMS very compatible with biological studies. Lim and Zhang have reviewed into some fabrication materials and techniques available for microfluidic and have elaborated on the advantages of these devices for immunoassays.11 3.1.2 Nanoparticles Used in Bioassays Disease biomarkers and biological agents are often present at ultralow levels and require ultrasensitive methods for detection. Different strategies have been employed for amplifying the transducing signals of bioassays. Most conventional amplification strategies have relied on the use of labels, such as enzymes, electroactive molecules, redox complexes, and metal ions. The emergence of nanotechnology is opening new horizons for the use of nanomaterial labels in signal amplification. The applications of nanomaterials in bioassays can be classified into two groups according to their functions: (1) nanomaterial modified electrochemical transducers to facilitate antibody/acid nucleic immobilization or improve properties of transducers and (2) nanomaterial–bimolecular conjugates as labels for bioassays. In particular, nanomaterial labels are showing the greatest promise for developing ultrasensitive bioassays. Antibodies or nucleic acids labeled with nanomaterials can retain their bioactivity and interact with their counterparts, and based on the electrochemical and optical detection of those nanomaterials, the amount or concentration of analytes can be determined. The enormous signal enhancement associated with the use of nanomaterial amplifying labels provides the basis for ultrasensitive bioassays.11,12

95

INTRODUCTION

Therefore, NPs have been used extensively in affinity biosensors for the detection of nucleic acids and proteins. These particles are unique because their nanometer size gives rise to a high reactivity and beneficial physical properties (electrical, electrochemical, optical, and magnetic) that are chemically tailorable. The composition of NPs determines the compatibility and the suitability of the probes with analytes and many assays are possible. Table 3.1 shows some of the NPs that have been more frequently described in analytical biosensors, together with the detection systems used. The noble metal NPs, mainly gold nanoparticles (AuNPs), have been the most extensively used for this purpose involving optical or electrochemical detection. They exhibit bright colors due to the presence of a plasmon absorption band that is not present in the spectrum of the bulk metal, which is a result of the resonance of the incident photon frequency with the collective excitation of the conductive electrons of the particle. This effect is termed localized surface plasmon resonance (LSPR) and depends on the size, shape and composition of the nanoparticles, the distance between nanoparticles, and the refractive index of the environmental medium. Another feature of these nanoparticles is their capability to produce surface-enhanced Raman scattering (SERS) effects. The different and interesting properties of AuNPs have been widely explored in bioassays using a variety of detection systems. The redox properties of AuNPs have led to their widespread use particularly as electrochemical labels in protein and nucleic acid detection, with numerous configurations being explored.3 The application of underpotential deposition of Ag monolayer as a means of enhancing the electrochemical sensitivity of biomolecular reaction has been successfully applied to the development of a new bioaffinity platform via metal-enhanced electrochemical detection (MED sensors). This is based on the discovery that immobilized metal layer, as continuous film, particle, colloids, or monolayer, significantly amplifies the electrochemical signals, while reducing the reorganization energy TABLE 3.1 Bioassays

Nanoparticles and Detection Systems Commonly Used in Analytical

Nanoparticles

Detection systems

Noble metals (Au, Ag, Pt)

Photometry, fluorimetry, Rayleigh and Raman scattering, surface plasmon resonance, potentiometry, amperometry, conductimetry, stripping voltammetry, quartz crystal microbalance Photometry, fluorimetry, FRET, stripping voltammetry

Quantum dots Silica or polystyrene Dye doped Lanthanide chelate doped Ruthenium chelate doped Carbon nanotubes Dendrimers Source: Adapted from Ref. 2.

Fluorimetry, phosphorimetry Fluorimetry, FRET Electrogenerated chemiluminescence Electrochemical Fluorimetry

96

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

following molecular recognition at sensor. Silver enhancement scheme has been utilized in the recent biobarcode approach and provided the lowest detection limit to date for both DNA (500 zM, PCR less) and protein targets ((3–30 aM). Quantum dots (QDs) are inorganic semiconductor nanocrystals with interesting luminescent and electrochemical properties extensively used in numerous bioassays. Briefly, these NPs show broad excitation profiles and narrow emission peaks and can emit in a range of wavelengths by changing their size and composition. Also, they lack photobleaching and have long fluorescence lifetimes. However, QDs can show blinking characteristics when they are excited with high-intensity light, which could be a limiting factor for fast scan systems, such as flow cytometry. Other limitations are toxicity, size variation, agglomeration, and nonspecific binding. Surface oxidation of QDs can occur under combined exposure to aqueous/UV light excitation, which can lead (e.g., in CdSe-based QDs) to the release of cadmium ions, so that these NPs are inadequate for in vivo applications, such as in vivo drug delivery assays. However, they offer better imaging results than those achieved by organic dyes in cell-based or tissuebased drug studies. A large number of bioassays have used dye-doped silica NPs that consist of luminescent organic or inorganic species dispersed inside a silica matrix. These NPs enable significant amplification of the analytical signal due to the numerous dye molecules inside each NP. The silica-based NPs functionalized for coupling and containing stable lanthanide and silica NPs containing ruthenium(II) chelates, mainly tris(2,20-bipyridyl)dichlororuthenium(II) (RuBpy), have also been used in fluorescence resonance energy transfer (FRET)-based assays and fluorescence or chemiluminescence (CL) detection. Carbon nanotubes represent an important group of nanomaterials with attractive geometrical, electronic, and chemical properties. The structure of CNTs comprises concentric cylinders, with a diameter of a few nanometers to hundreds of lumens in length, that have interlinked hexagonal carbon rings. In addition to favorable electronic properties, they show a large surface area and an electrocatalytic effect that have been used in constructing electrochemical biosensors. Conducting polymer nanowires (CPNWs) are attractive alternatives to silicon nanowires and carbon nanotubes because of their tunable conductivity, flexibility, chemical diversity, and ease of processing. CPNWs can be prepared using a variety of protocols, such as chemical synthesis, template electrochemical synthesis, and electrospinning, and some chemical and biological sensors based on CPNWs have been reported. Wang et al. introduced a new approach for the in situ electrochemical fabrication of an individually addressable array of CPNWs positioned within an integrated microfluidic device and also demonstrated that such an integrated device can be used as a chemical sensor immediately after its construction.14 There are other NPs that have so far found fewer applications in bioassays. This is the case for dendrimers that are hyperbranched, tree-like structures having three different regions (i.e., core, branches, and surface). They have been used in some bioassays as reagents by adsorbing, caging, or covalently binding active molecules, such as fluorescent dyes, inside or onto their surface. Magnetic particles, which respond to an external magnetic field, have been used extensively for separation and preconcentration purposes and in electrochemical

FUNDAMENTALS OF BIOSENSORS

97

biosensors. Their unique properties allow magnetic particle-conjugated molecules to be quickly agglomerated or resuspended in the medium according to the external magnetic force, thus making them suitable for purifying biologically active compounds, such as nucleic acids and proteins.5 They comprise a metal or metallic oxide core, encapsulated in an inorganic or polymeric coating that renders the NPs biocompatible and stable and that may serve as a support for biomolecules. Also, the development of hybrid NPs, such as magnetic AuNPs or magnetic QDs, which combine sample manipulation and sensitive detection, is a very promising recent field of research. Another nanoparticles used in biosensor technology are liposomes. Liposomes are composed of a lipid bilayer with the hydrophobic chains of the lipids forming the bilayer and the polar headgroups of the lipids oriented toward the extravesicular solution and the inner cavity. The sizes of the liposomes vary, ranging from nanometers to several micrometers, which depends on the synthesis conditions. Owing to its high surface area, large internal volume, and capability to conjugate bilayer lipids with a variety of biorecognition elements, liposomes have been widely used as bioassay labels by encapsulating enzymes, fluorescent dyes, electrochemical and chemiluminescent markers, DNA, RNA, ions, and radioactive isotopes. An excellent review of the uses of liposomes in immunoassays is available in the literature.15 There is a clear trend toward developing multiplexed bioassays using nanobarcodes, which are based on the synthesis of particles that contain a mixture of NPs functionalized with the corresponding recognition agents. These agents can be antibodies or oligonucleotide sequences that recognize the targets of interest in protein or nucleic acid detection, respectively. In fact, it is often necessary to monitor or quantize several components in a complex system. For example, due to the limited specificity and sensitivity of biomarkers for clinical diagnosis, the measurement of a single biomarker is usually insufficient for diagnostic purpose. Some studies have shown that the measurement of biomarkers panel can avoid false positive or false negative results to improve their diagnostic value. Traditionally, bioassay of analytes panel is performed as discrete tests, that is, one analyte per assay run, and several runs are needed to detect all components in a complex system. Great consumptions of time, reagent, and labor limit the application. To overcome these limitations, multiplexed bioassays can measure two or more analytes in a single run. Compared to parallel single analyte methods, multiplexed bioassay offers some remarkable advantages, such as high sample throughput, improved assay efficiency, low sample consumption, and reduced overall cost per assay.16 3.2 FUNDAMENTALS OF BIOSENSORS A biosensor can be defined as an integrated receptor transducer device that is capable of providing selective quantitative or semiquantitative analytical information using a biological recognition element. A biosensor converts a biological event into a detectable signal by the action of a transducer and a processor. The usual aim of a biosensor is to produce either discrete or continuous digital electronic signals that are

98

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

FIGURE 3.1

Scheme of biosensors.

proportional to a single analyte or a related group of analytes. The principal components and functions of biosensors are shown in Figure 3.1. In a biosensor, the physical–chemical transformation due to the interaction between the biological element and the analyte target is converted into a usable signal by the transducer. The main purpose of the recognition system is to provide specificity to the biosensor, thus creating a device capable to detect either a specific target or a related family of compounds. Immobilization of the biological component can be performed using a variety of methods such as chemical or physical adsorption, physical entrapment within a membrane or gel, cross-linking of molecules, or covalent binding. In general, biosensors are classified by either their biological element or the transducer used. Biosensors can be subdivided into two classes based on the type of biorecognition molecule. Catalytic biosensors employ enzymes and microorganisms as the biorecognition molecule that catalyses a reaction involving the analyte to give a product. Common analytes for catalytic biosensors are small organic molecules such as glucose. The other category of biosensors is affinity biosensors. Biorecognition molecules commonly used in affinity biosensors include antibodies, DNA, peptides, and lectins. Affinity biosensors are characterized by a binding event between the biorecognition molecule and the analyte (the affinity reaction), often with no further reaction occurring. Hence, the challenge then becomes transducing the biorecognition event. As this class of biosensor is compatible with the detection of virtually all biological agents, it is this challenge that faces researchers attempting to develop portable devices for detecting toxins, microbes, and viruses. Transduction of affinity biosensors has been achieved using labeled species and label-free approaches. If transduction is achieved using labeled species, the principles are very similar to immunoassay, with the amount of analyte detected being inferred from the amount of label that binds to the interface. Label-free methods most frequently involve evanescent wave-based optical methods or using mass-sensitive acoustic wave devices that monitor molecules binding to, or desorbing from, a transducer surface. The most common transducers for detecting labeled species are optical, where an optically active label is detected, or electrochemical, where the label is electroactive.

FUNDAMENTALS OF BIOSENSORS

99

Electrochemical transduction transforms the effect of electrochemical interaction between analyte and electrode into a primary signal. Such effects may be stimulated electrically or result in a spontaneous interaction at the zero current condition. Voltammetric, amperometric, potentiometric, solid electrolyte, gas sensor, and chemically sensitized field effect transistor (CHEMFET) are distinct subgroups. Optical transducers are based on various technologies involving optical phenomena that are the result of an interaction of analyte with receptor. This group may be further subdivided according to the optical properties that have been applied in sensing (i.e., absorbance, reflectance, luminescence, fluorescence, refractive index, surface plasmon resonance (SPR), optothermal effect, and light scattering). A large number of optical transduction techniques can be used for biosensor development. In the case of both electrochemical and optical methods, one of the key factors that limits biosensor performance is nonspecific binding. The problem of nonspecific binding highlights the importance of interfacial design in a biosensor. 3.2.1 Affinity Biosensors Affinity biosensors are a subclass of biosensors: they are analytical devices comprising a biological or biomimetic affinity element (receptor). The sensing element is a highly specific receptor, and it is generally biologic, for example, a receptor of natural origin (bioreceptors): enzymes, antibodies, and nucleic acids. In the past few years, a great interest in biomimetic biosensors has risen: they use artificial or semiartificial receptors. This class of synthetic receptors includes PNA (peptide nucleic acid), LNA (locked nucleic acid), MIPs (molecular imprinted polymers), oligopeptides, and aptamers. In this chapter, immunosensors and DNA biosensor using electrochemical and optical transducers have been considered. 3.2.2 Immunoassay Immunoassays are currently the predominant analytical technique for the quantitative determination of a broad variety of analytes in clinical, medical, biotechnological, and environmental significance. The recognition elements are immunochemical antibody–antigen (Ab–Ag) interactions. This type of device combines the principles of solid-phase immunoassay with physical–chemical transduction elements (electrochemical, optical, piezoelectric, EW, and SPR). 3.2.2.1 Antibodies The use of highly specific antibodies is very popular not only in biosensor research but also in bioanalytical chemistry. The most important applications of antibodies are immunoassay, immunosensor, and immunoaffinity columns. Antibodies are serum molecules produced by B lymphocytes; they represent the soluble form of specific receptors for the antigen expressed by B lymphocytes.

100

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

In plasma subjected to electrophoresis, antibodies are found as proteins related to the gamma fraction, the highest in molecular weight (globulins) portion. Hence, the protein fraction that contains antibodies is usually called immunoglobulin or Ig. Different types of Ig exist: IgG, IgA, IgM, IgD, and IgE. All the antibodies have the same base structure, but they differ in the region that binds the antigen. Immunoglobulin G (IgG) is the most abundant immunoglobulin species in serum and also the most commonly used antibody in sensor applications. The molecule consist of four polypeptide chains, two identical heavy (H) chains and two identical light (L) chains joined to form a “Y”-shaped unit (Figure 3.2a). The length of the two chains is 450 amino acids for the H chain and 212 amino acids

FIGURE 3.2 (a) Scheme of a conventional IgG antibody. (b) Different bioassay formats: (A) direct label-free detection of the target protein binding to immobilized antibodies, (B) detection of labeled target proteins, (C) competitive assay, and (D) sandwich assay. Adapted from Ref. 28.

FUNDAMENTALS OF BIOSENSORS

101

for the L chain. The two identical H chains are connected via disulphide bridges. The connection between the L chain and the H chain also consists of disulphide bonds. Since all these bonds connect two chains, they are named interchain disulphide bridges. Both chains, L and H, also have intrachain disulphide bridges. The globular structure of the protein that is responsible for the name immunoglobulin is a result of these intrachain bonds. The binding site (paratope) is located within the VH and VL domains and each arm contains one binding site. In the variable regions, amino acid sequences can vary from antibody to antibody and allow the specific adaptation to certain antigens. The exact regions within these variable regions that have very high amino acid variability are called hypervariable regions, also known as complementary determining regions (CDRs). Three CDRs are integrated into the L chain and three into the H chain, resulting in six CDRs for each arm. A molecule used to induce an immune response is called an immunogen. An antigen is a molecule that can bind to an antibody. The region of the antigen that binds to an antibody is called epitope or antigenic determinant, and the corresponding part in the antibody is named paratope. Antigens can be big or small molecules, but the small moleculesareantigenic only whenthey are coupledto aprotein carrier. Such compounds are named haptens. The most used protein carriers are the bovine serum albumin (BSA) and ovalbumin (OVA). A spacer bridge is used to conjugate the hapten to the protein, in order to have the correct distance between them and to allow the correct immunization of the organism against the target. The spacer bridge needs to be a molecular chain lacking in substituents groups to avoid immunization against the bridge. The use of carrier proteins is not restricted to small nonimmunogenic compounds; in fact, small proteins of low native immunogenicity have been coupled to carrier proteins and successfully used to generate antibodies. 3.2.2.2 Antigen/Antibody Interaction Antigenic molecules are surrounded and trapped inside a pocket formed by the light and heavy chains of an antibody. This is called the combinatory site and is where a protein can be captured. The union between an antigen and an antibody is the result of noncovalent interactions between the amino acidic residuals of the antigens and of the combinatory site of the antibody. These bonds are weak interactions of different types such as hydrogen bonds, electrostatic forces, van der Waals forces, and hydrophobic interactions, but they all contribute to a binding of relevant energy. The strength of these bonds depends critically on the distance d between reagent groups. This strength is proportional to 1/d2 in the case of electrostatic forces and 1/d7 in the case of van der Waals forces. The affinity of an antibody toward an antigen can be expressed by the strength of repulsive and attractive forces. An antibody with high affinity for an antigen fits perfectly to its specific antigen. The law of mass action can be used to calculate the affinity given by the equilibrium constant K (equation (3.1)). Ab þ Ag $ Ab
Ag ka K¼ kd

ð3:1Þ

102

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

where [Ab] represents the antibody concentration, [Ag] the antigen concentration and [Ab–Ag] the antigen–antibody complex concentration. When the antibody is added in a solution containing the antigen, binding sites of antibody are involved in the antigen interaction; this reaction is regulated by antigen concentration in the sample and by the equilibrium constants. This assumption can be used when antibody and antigen are homogeneous, when the antigen has only one epitope and the antibody has only one binding site, and the separation between the bound and free forms is complete. Affinity constants are a quantitative measurement of the affinity between the two reagents and range from 106–1012 M
1. 3.2.2.3 Conventional Immunoassays Immunoassay makes use of the specific and sensitive antibody–antigen interaction and can be used to determinate either antibody or antigen concentration. To monitor the interaction, one of the species is conjugated with one of the many available labels. In general, in an immunoassay, one of the reaction partners, antibody or antigen, is bound to a surface and the binding of the second immunoreagent is detected by measuring the concentration of the attached label. Various assay formats are used for different applications. Immunoassays can be catalogued based on the type of label in these categories: (1) enzyme immunoassay (EIA); (2) radioimmunoassay (RIA); (3) fluorescent immunoassay (FIA); (4) enzyme-linked immunosorbent assay (ELISA). ELISA test is the most commonly used format for analysis. The factors contributed to the success of this format are speed, sensitivity, selectivity, and no use of radioactive materials. A typical ELISA is carried out in 96-well microtiter plate automated plate; shakers, washers, and readers are available. Immunoassays are well established in many fields of analytical interest for screening purposes or accurate results; in fact, they are cheap, sensitive, and allow many simultaneous analyses. Immunoassay procedures have been optimized for numerous analytes, and commercial kits are available for a wide variety of analytical compounds. Immunoassays can be set up in a variety of formats (Figure 3.2b) and the most important are competition, sandwich, and displacement formats. The main difference between the formats is the immobilized species, such as antibody, antigen, or hapten conjugate, and the number of layers used. The decision which of the formats is used for a particular analytical problem is influenced by the nature of the analyte, the cost and availability of antigen and antibody, and the required sensitivity. Most of the developed immunosensors are based on either competitive or sandwich assay when applied to the detection of low (herbicides, toxins) and high (proteins, cells) molecular weight molecules, respectively. The sandwich assay is used when the molecular target has multiple epitopes that is able to bind more than one antibody at the same time; this is possible when the epitopes are spatially well separated. In a direct sandwich assay, the first antibody (called capture antibody) is immobilized on the solid phase, and then the antigen (analyte) is added. The solid phase is washed to remove the unreacted components and then

FUNDAMENTALS OF BIOSENSORS

103

incubated with a secondary labeled antibody, able to react with the other epitope. After the washing step, the amount of secondary antibody bound to solid phase is determined. A competitive assay is generally used when the target has only one epitope and, as the name indicates, a competition reaction between two reagents for a third one exist. In an indirect competitive assay, the solid phase is modified by the immobilization of a spacer-linked antigen and labeled antibody and free antigen (analyte) are added. The solid-phase-bound and free antigens, present in the tested solution, compete for the antibody binding site. The extent of the affinity reaction is detected by adding a secondary labeled antibody able to bind the first one. When the labeled antibody concentration is kept constant, an increasing analyte concentration occupies more and more binding sites. This means that less labeled antibody can bind to the surface-bound antigen. Increasing analyte concentrations cause decreasing signals. In a direct competitive assay, the solid phase is modified by immobilization of the specific antibody and free and labeled antigens are added. The concentration of the labeled antigen is kept constant. Free antigen is added in a dilution series or as an unknown sample. Free and labeled antigens compete for the binding site of the immobilized antibody. High analyte concentrations occupy most of the binding sites and results in a low signal. Low analyte concentration allows the labeled antigen to bind. Increasing analyte concentrations cause a decrease in signal. 3.2.3 Immunosensor Competitive binding assay formats are most used in immunosensors using optical or electrochemical transducers. The main limitation of electrochemical techniques is the detection of the immunoreactions because it is necessary to use enzymes that will generate electrochemically active compounds. In general, a vast number of optical transduction techniques can be used for biosensor development. These may employ linear optical phenomenon (e.g., adsorption, fluorescence, phosphorescence, and polarization) or nonlinear phenomena (e.g., second harmonic generation). The choice of a particular optical method depends on the analyte and the sensitivity needed. Optical detection, absorbance, fluorescence, chemiluminescence, or evanescent wave monitoring, is the most common form of detection used with immunoassays. Evanescent field-coupled refractometric optical sensors have increased in popularity, as they monitor immunological hybridization reactions in real time. In a recent review, the state of the art and the recent developments in immunosensor have been described.17 Homogeneous immunosensor, heterogeneous immunosensor, integrated immunosensor, and biochip format immunosensor are presented based on optical, electrochemical, magnetic, or mechanical detection/ transduction systems. Most of the developed immunosensors include a sensing layer supporting a particular immobilized antigen or antibody. The solid support used is generally in close contact with a transducer needed for the detection of the formed immune complex.

104

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

In protein-sensing devices, the immobilized compound determines the specificity of the device, and the immobilization method frequently influences parameters such as lower detection limit, sensitivity, dynamic range, reusability, or liability for unspecific binding. Thus, varieties of immobilization methods are described that are applicable to different supports onto which the compound has to be immobilized (immobilization substrate).18 The choice of the immobilization substrate depends on the chosen assay format and detection principle. Many immunosensors in recent years have used different NPs as labels and have given rise to the return of interest in metalloimmunoassays, taking into account the metallic character of most of these NPs. There is a trend to develop multiplexed assays, but so far most of the immunoassays described only allow individual determinations, partly because of the cross-reactivity limitations of the antibodies. However, some recent examples show that NPs are useful for simultaneous determinations, which are described below. Although other methods have been reported, the assays selected can give an overview of the usefulness, as well as the versatility and the applicability of NPs in immunoassays. 3.2.3.1 Electrochemical Immunosensor Electrochemical immunosensors have been widely used in amperometric, potentiometric, and conductimetric configurations. Over the past few years, DiagnoSwiss has developed polymer devices with the advantages of microfluidic for bioanalytical applications.19 A new microfluidic biosensor platform dubbed GRAVI is commercially available for running assays with paramagnetic nanoparticles. Capture antibody is immobilized on nanoparticles, which can be preincubated with sample, in tube. After incubation, the mixture is flowed through the microchannels, and the paramagnetic nanoparticles are trapped near the electrodes by virtue of a magnet array. The biological reactions (occasionally requiring longer incubation times) can thus be freely adjusted in function of the assay. Reactions leading to the formation of the immune complex can be performed in tube or in the microchip, while washing steps and detection of the enzymatic reaction take place in the microchannels. In this manner, the microchip merely serves as physical biosensor, in which tests with a variety of microbeads can be performed in successive runs. The flexibility to consecutively use nanobeads functionalized with different capturing moieties translates into a full random access solution. Characterized by dramatically reduced time to result (<10 min) and significantly decreased sample/reagent consumption, the cost-efficient biosensor instrumentation allows performing multimenu analysis with minimal laboratory infrastructure. Coupled to a robotic liquid handler, the system dispenses samples and reagents from conventional plates or tubes into microchannels of a microchip in which assays are processed and results readout. As in conventional 96-well microtiter plates, the microchannels have a standard spacing of 9 mm to facilitate automation. With solely gravity- and capillary force-driven fluidics within the microchannels, liquids are free to flow, while magnetic beads, functionalized with the antibody of choice, are trapped near incorporated electrodes by virtue of a magnet array. Following assay performance

FUNDAMENTALS OF BIOSENSORS

105

and electrochemical signal detection in the parallel microchannels, chips are regenerated by magnet release and rinsing of beads out from the microchannels.20 Biosensors have emerged as a new technique for monitoring cancerous cells or their specific interaction with different analytes. Microfluidic-based microchips have become the focus of research interest for immunoassays and biomarker diagnostics. This is because the conventional immunoassays require relatively long assay times and large, complicated detection devices. Some important cancer biomarkers, alphafetoprotein (AFP), carcinoembryonic antigen (CEA), and prostate-specific antigen (PSA), have been detected by microfluidic-based immunosensing microchips for the diagnosis of liver, colon, and ovarian cancers, respectively. An excellent review has been published on the biosensor technology available today, areas that are currently being developed and researched for cancer markers diagnosis, and a consideration of future prospects for the technology.21 Maeng et al. have developed and characterized an immunoassay methodology comprised of microbeads and microbiochips. In this method, microbeads are used to filter and immobilize antibodies, and an immunogold silver staining (IGSS) method is then used to amplify electrical signals that correspond to the bound antibodies. The chip used for this system is composed of an inexpensive and biocompatible PDMS layer over a Pyrex glass substrate that contains a platinum (Pt) microelectrode used to detect the electrical signal in this system. The microelectrode is fabricated on the substrate and a microchannel and pillar-type microfilter is formed in the PDMS layer. A sandwich immunoassay approach was applied to detect AFP, a cancer biomarker, using this system. The results of this study showed that the time required for a complete assay was reduced by 1 h and a detection limit as low as 1 ng/mL was attained when this system used, which indicates that similar bead-based electrical detection systems could be used for the diagnosis of many forms of cancer.22 In clinical diagnosis, detection of one biomarker cannot provide sufficient clinical information for various cancer-related diseases, and the clinical information obtained from biomarkers is often related to the stage of tumorigenesis, monitoring of treatment, and the state of the patient. Therefore, it is important to develop the microchip system with high multiplexing capabilities as well as an efficient detection method. Researches related to immunosensing microchips have achieved efficient multiplex detection of biomarkers. A biobarcode assay (BCA) capable of achieving low detection limits and high specificity for both protein and DNA targets was developed by Goluch et al.23 The BCA utilizes AuNPs functionalized with oligonucleotides (the so-called biobarcodes that serve as surrogate targets and amplifying agents) and a target recognition element that may be an antibody for protein detection or an unique oligonucleotide sequence for nucleic acid detection. The BCA also uses functionalized magnetic microparticles (MMPs) adorned with antibodies that bind to the target. In the presence of targets (protein or oligonucleotide molecules) in solution, the MMPs form a sandwich complex with targets and gold nanoparticles, which can be localized and collected under an applied magnetic field. The barcode oligonucleotide molecules are then chemically released, identified, and quantified. The realization of a BCA in a microfluidic format presents unique opportunities and challenges. A modified

106

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

FIGURE 3.3 Schematic diagram of the surface immobilized biobarcode assay protocol.24 (a) The walls of the capture region are coated with antibodies, (b) samples are flowed through the capture region, (c) the target molecules attach to the antibodies on the channel walls, and (d) the target proteins are tagged with cofunctionalized nanoparticles containing polyclonal antibodies and unique barcode DNA oligonucleotides. (e) The barcode DNA is then released from the nanoparticles and transferred to the detection region where the complementary sequence is patterned. Steps (f)–(h) illustrate a scanometric detection protocol. (f) The barcode molecules attach to the complementary sequences in the appropriate regions, (g) universal nanoparticle probes are attached to the barcode DNA, and (h) the universal probes are silver stained to facilitate visualization in the visible spectrum. The upper channels represent the target capture region while the lower channels mark the barcode detection region of the device. Pneumatic control channels are inserted for directing the flow of fluid.

form of the BCA called the surface immobilized biobarcode assay (SI-BCA) was developed by the same research group.24 The SI-BCA employs microchannel walls functionalized with antibodies that bind with the intended targets (Figure 3.3). Compared to the conventional BCA, it reduces the system complexity and results in shortened process time, which is attributed to significantly reduced diffusion times in the microscale channels. Raw serum samples, without any pretreatment, were evaluated with this technique. Prostate-specific antigen in the samples was detected at concentrations ranging from 40 pM to 40 fM. The entire assay, from sample injection to final data analysis, was completed in 80 min. Wilson and Nie have developed a microchip for seven cancer biomarkers using an electrochemical detection method. The above-mentioned microchip-based multiplex immunosensing devices require a small quantity of sample and are time saving and convenient. Nevertheless, some shortcomings, such as requiring high power source, poor reproducibility, and no real-time monitoring, still remain.25 Yoomin and coworkers have described the development of a microchip-based multiplex

FUNDAMENTALS OF BIOSENSORS

107

electroimmunosensing system for simultaneous detection of cancer biomarkers using gold nanoparticles and silver enhancer. The microchip is composed of biocompatible poly (PDMS) and glass substrates. To fix the antibody immobilized microbeads, they used pillar-type microfilters within a reaction chamber. An IGSS method was used to amplify the electrical signal that corresponded to the immune complex. To demonstrate this approach, the authors simultaneously assayed three cancer biomarkers, AFP, CEA, and PSA, on the microchip. The electrical signal generated from the result of the immunoreaction was measured and monitored by a PC-based system. The overall assay time was reduced from 3–8 h to about 55 min when compared to conventional immunoassays. The working range of the proposed microchip was from 10
3 to 10
1 mg/mL of the target antigen.26 Tang et al. have reported a novel method for the detection of tumor markers, such as a-fetoprotein, CEA, cancer antigen 125 (CA 125), and CA 15-3 can be found in the body (usually blood or urine) when cancer is present. They synthesized magnet core/shell NiFe2O4/SiO2 nanoparticles and fabricated an electrochemical magnetic controlled microfluidic device. The immunoassay system consisted of five working electrodes and an Ag/AgCl reference electrode integrated on a glass substrate. Each working electrode contained a different antibody immobilized on the NiFe2O4/SiO2 nanoparticle surface and was capable of measuring a specific tumor marker using noncompetitive electrochemical immunoassay. Under optimal conditions, the multiplex immunoassay enabled the simultaneous detection of four tumor markers. The sensor detection limit was <0.5 mg/L for most analytes.27 The affinity biosensors have been widely applied for foodborne pathogen detections with the goal to overcome problems associated with traditional microbiological detection techniques. These are very elaborate, time-consuming, and have to be completed in a microbiology laboratory and are therefore not suitable for onsite monitoring. Rapid and reliable detection methods of pathogenic, toxin producing bacteria such as Salmonella spp., Listeria monocytogenes, Escherichia coli 0157:H7, or Staphylococcus aureus that are responsible of some of the major worldwide foodborne outbreaks are required. Biosensor-based tools offer the most promising solutions and address some of the modern-day needs for fast and sensitive detection of pathogens in real time or near real time. In particular, electrochemical biosensors for the detection of food pathogens have the advantage of being highly sensitive, rapid, inexpensive, and amenable toward microfabrication and were reviewed by Tamiya and coworkers.28 As a principle of transduction, the impedance technique has been applied in the field of microbiology as a means to detect and/or quantify foodborne pathogenic bacteria. The integration of impedance with biological recognition technology for detection of bacteria has led to the development of impedance biosensors that are finding widespread use in the recent years. Yang and Bashir have reviewed the progress and applications of impedance microbiology, particularly the new aspects that have been added to this subject in the past few years, including the use of interdigitated microelectrodes, the development of chip-based impedance microbiology, and the use of equivalent circuits for analysis of the impedance systems.29

108

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

Some researchers have continuously improved the impedance biosensor methods by integrating newly developed nanoparticles and microfuidics with interdigitated microelectrodes. Varshney and Li30 have realized an IDA-based impedance biosensor coupled with magnetic nanoparticle–antibody conjugates for rapid and specific detection of E. coli O157:H7 in ground beef samples. Instead of immobilizing antibodies directly on the electrode surface, antibodies were immobilized on magnetic nanoparticles. Magnetic nanoparticles (Fe3O4, 145 nm diameter) were conjugated with anti-E. coli antibody through biotin–streptavidin chemistry. The conjugates were then used to separate and concentrate E. coli cells from ground beef samples. The nanoparticle–cell complexes in 0.1 M mannitol solution were measured by impedance using IDA microelectrodes with 50 pairs of finger electrodes, each measuring 15 mm in width and space. When 2 mL of the complexes solution was spreading on the IDA electrode surface, nanoparticle–cell complexes were concentrated into the active layer of the IDA with the assistance of a magnet field. The lowest detection limits of this biosensor system for detection of E. coli O157:H7 in pure culture and ground beef were 7.4  104 and 8.0  105 cfu/Ml, respectively. This biosensor method has been later refined into a microfluidic chip-based biosensor by the same group.31 The microfluidic chip had a small detection chamber (60 nL) formed by a PDMS with embedded gold interdigitated microelectrodes on the bottom of the chamber. Magnetic particle–cell complexes in mannitol solution were injected into the detection chamber for sensitive impedance measurement. This microfluidic impedance biosensor was able to detect as low as 1.6  102 and 1.2  103 cfu/mL of E. coli cells present in pure culture and ground beef samples, respectively. Boehm et al. have developed an on-chip microfluidic biosensor for E. coli detection and identification. In this microfluidic biosensor, anti-E. coli antibodies were immobilized on the glass surface that served as the bottom of the microfluidic chamber; the impedance detection electrodes were however on the top cover of the chamber. Bacteria in suspension passing through the microfluidic chamber were selectively recognized and captured by the immobilized antibodies, thereby increasing the measured impedance within the chamber. This biosensor was able to detect about 104 cfu/mL of E. coli when a shallow chamber (2 mm) was used.32 The use of nanoparticles can improve the capture efficiency of antibodies to target cells.31 The microfluidic-based sensors allow continuous injection/perfusion of bacteria samples and accumulation/concentration of bacterial cells inside the impedance detection chamber over time, which can enhance the detection sensitivity and is particularly useful for detecting low concentrations of bacteria.32 These studies have brought attention to the impedance techniques suitable for label-free detection and have important advantages such as speed, de-skilled analysis, fewer numbers of steps, and the potential for the multianalyte detection. Baeumner and coworkers33 have developed microfluidic biosensors for detecting cholera toxin subunit B (CTB) as a model analyte using electrochemical and optical transducers. They employed liposome-based signal amplification, encapsulating labels within the liposomes. The microfluidic devices were made from PDMS using soft lithography from silicon templates. The polymer channels were sealed with a glass plate and packaged in a polymethylmethacrylate housing that provided leakproof

FUNDAMENTALS OF BIOSENSORS

109

sealing and a connection to a syringe pump. In the electrochemical format, an interdigitated ultramicroelectrode array (IDUA) was patterned onto the glass slide using photolithography, gold evaporation, and lift-off processes. For CTB recognition, CTB-specific antibodies were immobilized onto superparamagnetic beads and ganglioside GM1 was incorporated into liposomes. The fluorescence dye sulforhodamine B (SRB) and the electroactive compounds potassium hexacyanoferrate (II)/hexacyanoferrate(III) were used as detection markers that were encapsulated inside the liposomes for the fluorescence and electrochemical detection formats, respectively. The limits of detection (LOD) of both assay formats for CTB were found to be 6.6 and 1.0 ng mL
1 for the fluorescence and electrochemical formats, respectively. Changing the detection system was very easy, requiring only the synthesis of different markerencapsulating liposomes, as well as the exchange of the detection unit. It was found that in addition to a lower LOD, the electrochemical format assay showed advantages over the fluorescence format in terms of flexibility and reliability of signal recording. Electrochemical biosensors have been studied for many years at a development and research level, successfully applied in industries in the past few years, and now accepted as a standard method for screening some bacterial cells in food samples. Nanomaterials are claimed to improve electrochemical biosensor sensitivity. In respect to other transducing principles, electrochemical techniques are much easier to use and allow the miniaturization for the integration in handheld devices. 3.2.3.2 Optical Immunosensor Optical biosensors have been used widely over the past decade to analyze biomolecular interactions providing detailed information on the binding affinity and kinetics of interaction. A novel interference localized surface plasmon resonance (iLSPR) biosensor for the label-free detection of biomolecules in an arbitrary solution is reported by Tamiya and coworkers. The experimental and simulation analysis of an original nanostructure design constructed with plasmonic gold nanoparticles and photonic thin-film multilayers of silicon dioxide (500 nm in thickness) and silicon on a substrate was presented. The nanostructure substrate showed a high sensitivity for various refractive index solutions and a prominent capacity for functionalizing alkanethiol molecules on the gold surface and demonstrates great potential in the development of a microfluidicbased biosensor for monitoring biotin–avidin interactions in real time.34 A reflection-based localized surface plasma resonance fiber-optic probe for chemical and biochemical sensing, called fiber-optic localized plasma resonance (FO-LPR), has been proposed.35 Biomolecular recognition has detected the unique optical properties of self-assembled gold nanoparticles on the unclad portions of an optical fiber whose surfaces are modified with a receptor. To enhance the performance of the sensing platform, the sensing element is integrated with microfluidic chips to reduce sample and reagent volume, to shorten response time and analysis time, and to increase sensitivity. The main purpose of the present study is to simulate the biochemical assays in the FO-LPR microfluidic chip and to investigate the effects of parameters, such as inlet concentrations of analyte or the flow rate on the biochemical binding kinetics. The geometry of the grooved channel is also proposed to enhance the biochemical binding on the unclad optical fiber. The results reveal that

110

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

the chaotic mixing generated by the grooves enhances the biochemical binding when the injected flow rate is high and, because of this, limits the performance of the molecular mixing. The enhancement of biochemical binding performance was significant, especially at the low injected concentration of analyte. Xu and coworkers have described a rapid and ultrasensitive detection method using a microfluidic chip for analyzing benzodiazepines. Benzodiazepines are mainly used in the treatment of insomnia in clinical cases. Clonazepam (CZP) is a benzodiazepine derivative, and 7-aminoclonazepam (7-ACZP) is the major urinary metabolite (target metabolite) of clonazepam. A microfluidic chip-based immunoassay with laserinduced fluorescence (LIF) detection based on the water-soluble denatured bovine serum albumin (dBSA)-coated CdTe quantum dots was prepared for the ultrasensitive detection of 7-ACZP. The detection of 7-ACZP could be completed within 5 min. This method was compared with ELISA and showed a good correlation. The results were confirmed by high-performance liquid chromatography and tandem mass spectrometry (LC-MS/MS).36 The immuno- and affinity assays using mobile support allow the probes to move freely through a liquid media. The mobile support usually consists of microspheres made of latex or magnetic materials, or nanospheres such as QDs or gold nanoparticles. When using microspheres for multiplex assays, antibodies with fluorescent labels are attached to the microsphere or fluorescent microspheres are used for identification. Original bioassay for multiple types of antibodies (multiplex assay) was presented by Yoon and coworkers.37 They used an immunoassay (a type of immuno- and affinity assays) with mobile support. The QDs were conjugated onto microspheres both to enable multiplex assays and to enhance the limit of detection. This configuration was called “nano-on-micro” or “NOM.” Upon radiation with UV light (380 nm), a stronger light scattering signal is observed with NOMs than QDs or microspheres alone. In addition, NOMs are easier to handle than QDs. Since QDs also provide fluorescent emission, they are able to utilize an increase in light scattering for detecting antigen–antibody reaction and a decrease in QD emission to identify which antibody (or antigen) is present. Two types of NOM combinations were used. One batch of microspheres was coated with QDs emitting at 655 nm and mouse IgG (mIgG) and the otherwith QDs emitting at 605 nm and BSA. A mixture of these two NOMs was used to identify either anti-mIgG or anti-BSA. NOM particles and target solutions were mixed in a microfluidic device (using highly carboxylated microspheres as previously demonstrated by the samegroup) and on-chip detection was performed using proximity optical fibers. Forward light scattering at 380 nm was collected. With the positivetarget, the scattering signal was increased. The LOD was as low as 50 ng ml
1 (330 pM). Fluorescent emission (655 or 605 nm) was simultaneously collected. With the positive target, the emission signal was attenuated. Therefore, they were able to detect two different antibodies simultaneously with two different detection protocols. Chan’s group has created a diagnostic system capable of multiplexed, highthroughput analysis of infectious agents in human serum samples. They have demonstrated, as a proof-of-concept, the ability to detect serum biomarkers of the most globally prevalent bloodborne infectious diseases (i.e., hepatitis B, hepatitis C,

FUNDAMENTALS OF BIOSENSORS

111

and HIV) with low sample volume, rapidity, and 50 times greater sensitivity than that of currently available FDA approved methods.16 3.2.4 DNA Bioassay The field of molecular diagnostics has expanded rapidly over the past decade. Applications include the detection of mutations responsible for human inherited disorders, disease-causing and food-contaminating viruses, and research into bacteria and forensics. Detection of infectious species and genetic mutations at the molecular level opens up the possibility of performing reliable diagnosis even before any symptom of a disease appears. In addition, the development of novel therapeutics based on the regulation of gene expression provides revolutionary new opportunities in the area of pharmaceutical science. To improve patient care, molecular diagnostics laboratories have been challenged to develop new tests that are reliable, cost-effective, and accurate and to optimize existing protocols by making them faster and more economical. Conventional methods for the analysis of specific gene sequences are based on either direct sequencing or DNA hybridization. Because of its simplicity, DNA hybridization is more commonly used in the diagnostic laboratory than the direct sequencing method. In DNA hybridization, the target gene sequence is identified by a DNA probe that can form a double-stranded hybrid with its complementary nucleic acid with high efficiency and extremely high specificity in the presence of a mixture of many different, non-complementary nucleic acids. DNA probes are single-stranded oligonucleotides, labeled with either radioactive or non-radioactive material, that provide detectable signals indicating DNA hybridization. Radioactive labels are extremely sensitive but have the obvious disadvantages of short shelf life, risks associated with exposure of personnel to radiation, cost, storage, and disposal problems. On the other hand, non-radioactive probes, such as enzymatic or luminescence labels, are less sensitive and flexible in terms of design and application but are clearly safer and more environmentally friendly. Over the past few years, advances in robotics, microfluidics, electronics, and high-resolution optics have driven the impressive development of both DNA microarrays and real-time PCR systems. The majority of commercial microarrays (pioneered by Affymetrix Inc. with their GeneChip )38 and all real-time PCR instruments (e.g., ABI Prism 7900HT Sequence Detection System)39 rely upon the detection and quantitation of a fluorescent reporter, whose signal increases proportionally to the amount of hybridized target or amplified PCR product. Detection thus requires imaging equipment or fluorescence readers that are generally very expensive. In addition to these technologies already impacting on the market, new commercial research tools are expected to have a major influence in the coming years. 3.2.4.1 Nucleic Acids Structures The double helix structure of double-stranded DNA (dsDNA) is well known and is easily recognized not just within the scientific community. Although common, the

112

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

double helix is not the only possible configuration for DNA to take, with some configurations being radically different from the Watson and Crick proposed structure. In physiological conditions, A-DNA, B-DNA (Watson and Crick structure), and ZDNA have all been observed. Figure 3.4a shows the crystal structures for two DNA sequences, one is in the A conformation and the other is in the B. The conformation assumed by a length of DNA is dependent on a variety of factors including the base pair sequence and the supporting environment. These three conformations consist of two antiparallel strands bound together through hydrogen bonding. These single strands (ssDNA) consist of two main parts, namely, the phosphate–deoxyribose backbone that forms the chain and the associated nucleobases (commonly referred to as bases). The bases are carbon–nitrogen ring structures, which are of two types, the purines (adenine (A) and guanine (G)) consisting of two fused rings and the pyrimidines (thymine (T) and cytosine (C)) consisting of only one ring. These bases are joined to the phosphate backbone and the order in which they occur provides the basic coding for genetic material. The asymmetric ends of the DNA strands are labeled as 50 that terminates in a phosphate group and 30 that terminates with a hydroxyl group, and the nomenclature refers to the position of the terminal carbon in the (deoxy)ribose ring. As a

FIGURE 3.4 (a) The crystal structures for two DNA sequences: one is in the A conformation and the other is in the B.48 (b) Chemical structures of the “Watson and Crick” paired bases; adenine and thymine as well as guanine and cytosine.

FUNDAMENTALS OF BIOSENSORS

113

consequence of both space constraints, and so as to maximize hydrogen bonding, guanine preferentially binds to cytosine and adenine to thymine—this is known as Watson and Crick base pairing, Figure 3.4b shows the chemical structures of these nucleobases in their Watson and Crick pairs. Unlike the other two major conformations, Z-DNA differs in exhibiting a left-handed double helical structure (i.e., the helix rotates around the axis in the opposite sense); this structure is found to often occur with G–C rich sequences in low salt conditions. With B-DNA the interwinding of the DNA strands leads to a structure in which there are two distinct grooves spiraling around the DNA duplex. The larger groove is called the “major groove” and the smaller one the “minor groove”; these structural features are labeled in Figure 3.4a. The conformation of DNA while attached to an electrode surface has been shown to depend upon the strength of the interactions between the interface and the DNA strand. Through heating dsDNA and breaking the weak hydrogen bonds, the two strands may be separated. This process is known as “denaturation.” The melting temperature (Tm) at which this occurs depends on both the length of the DNA chain and the constituent bases. On cooling ssDNA, with a complementary sequence, the bases pair together to form dsDNA. This is known as “hybridization.” It is possible for strands of DNA that are not fully complementary to pair but the number of mismatches affects the stability of the dsDNA and consequently lowers the melting temperature. 3.2.5 DNA Biosensors DNA biosensors are analytical devices that result from the integration of a sequencespecific probe (usually a short synthetic oligonucleotide) and a signal transducer. The probe, immobilized onto the transducer surface, acts as the biorecognition molecule and recognizes the target DNA, while the transducer is the component that converts the biorecognition event into a measurable signal. Assembly of numerous (up to a few thousand) DNA biosensors onto the same detection platform results in DNA microarrays (or DNA chips), devices that are increasingly used for large-scale transcriptional profiling and single-nucleotide polymorphisms (SNPs) discovery. As clinical diagnostics and other applications (e.g., environmental screening) do not generally need the massive data accumulation typical of gene chips, alternative technologies are in development whose promise is to provide flexible and economical alternatives for applications that require relatively fewer measurements. The DNA biosensor can be broadly divided into two main groups: label-free systems and labeled. Those methods that use solution-phase reagents (e.g., metal complexes or organic dyes) as markers of the hybridization process will be referred to as label free. Label-free approaches typically rely on the measurement of changes in the electrical and optical characteristics of sensing layer before and after the hybridization reaction. By contrast, when organic and organometallic electroactive compounds, nanoparticles, and catalytic and redox enzymes are permanently bound (e.g., covalently or via (strept)avidin–biotin interactions) to one of the constituents of the surface-tethered duplex, the method will be considered as label based. Sensitivity and reliability of label-based approaches are often still unrivalled, as also witnessed by the choice of these methodologies for the microarray platforms now on the market.

114

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

Tremendous research activities have been carried out to miniaturize the conventional DNA analytical procedures in microchip platforms. These microdevices enjoy the miniaturization advantages of small size, low sample, reagent and power consumption, enhanced analytical performance (e.g., shorter assay time), and high level of integration. An ideal microanalyzer should feature sample in result out kind of automated operation, without any human intervention between individual assay steps. There are three essential components in a complete DNA assay protocol that include sample preparation, target amplification, and product detection. The following sections aim to give an outline of the wide variety of methods that have to date been employed to electrochemically and optically detect specific DNA sequences. Typically, the design of an electrochemical DNA biosensor involves the following steps: 1. Immobilization of the DNA probe 2. Hybridization with the target sequence 3. Labeling and electrochemical investigation of the surface Optimization of each step is required to improve the overall performance of the DNA sensing. 3.2.5.1 DNA Probe As the specificity of the hybridization reaction essentially depends on the biorecognition properties of the capture oligonucleotide, design of the capture probe is undoubtedly the most important preanalytical step. Thus, a number of probes, variable for chemical composition and conformational arrangement, have been used to assemble DNA biosensors. Oligonucleotides (ODNs) are regularly used as the DNA probe in a biorecognition layer. ODNs are short sequences of DNA generally 20 bases or less in length. Design of linear probes takes now great advantage of decades of experience, which has led to many commercially available types of software. DNA probes are typically short (18–40-mer) oligonucleotides that are able to hybridize with specific target sequences. While earlier work employed simple DNA probe sequences as a model (e.g., oligo d(G)20), recent reports describe the use of disease- or microorganism-related oligonucleotide sequences. Some alternative methods use PNA and LNA such as DNA analogues. PNA possess an uncharged pseudopeptide backbone (instead of the charged phosphate–sugar backbone of natural DNA). Because of their neutral backbone, PNA probes offer greater affinity in binding to complementary DNA and improved distinction between closely related sequences (including single-base mismatches). Such mismatch discrimination has a particular importance in the detection of disease-related mutations. Locked nucleic acids (LNA)40 are a class of nucleic acid analogues in which the ribose ring is “locked” by a methylene bridge connecting the 20 -O atom with the 40 -C atom. DNA oligos incorporating LNA nucleosides show increased thermal and further improved discriminative power with respect to single-base mismatched targets.

FUNDAMENTALS OF BIOSENSORS

115

The probe immobilization step plays the major role in determining the overall performance of an electrochemical DNA biosensor. The achievement of high sensitivity and selectivity requires maximization of the hybridization efficiency and minimization of nonspecific adsorption, respectively. Control of the surface chemistry and coverage is essential for assuring high reactivity, orientation, accessibility, and stability of the surface-confined probe, as well as for minimizing nonspecific adsorption events. 3.2.5.2 Hybridization Reaction The kinetics and mechanism of the hybridization reaction in solution has been widely studied. Hybridization involves a two-step process: nucleation and zippering. Nucleation is the rate-limiting step. It is assumed that the nature of the hybridization reaction at solid surfaces closely approximates that of the solution-phase reaction, but its rate is about 10–100 times slower. Efficient hybridization of a target to surface-bound probes can be impeded by several phenomena. For example, the immobilized probe may be not accessible for hybridization because of some steric hindrance. The rate of hybridization and the stability of the duplex depend on several factors, such as salt concentration, temperature, use of accelerating agents, base composition (G þ C content), and length of the probe sequence. The salt concentration markedly affects the rate of hybridization reaction. Below 0.1 M NaCl, a twofold increase of salt concentration increases the hybridization rate by 5–10-fold or even more. The rate levels off when the concentration exceeds 1.2 M NaCl. However, since this high salt concentration stabilizes mismatched duplexes, the use of high ionic strength solutions is not recommended for single-base mutation analysis. The rate of hybridization strongly depends on the temperature. The maximum rate is observed 20–25 C below Tm of the duplex. However, depending on salt concentration, annealing may effectively occur at temperatures well below the optimum value. The overall sensitivity of a hybridization assay is strongly influenced by the hybridization time. Moreover, the hybridization process can be facilitated using appropriate reagents. The presence of guanidine HCl in the target solution was shown to highly increase the rate of hybridization. The stringency of hybridization can be additionally altered using formamide. Formamide decreases the Tm of nucleic acid hybrids. Use of 30–50% formamide in the hybridization solutions allows the incubation temperature to be reduced to 30–42 C. The effect of sequence length on hybridization rate is well known. The lower hybridization yield of assays in which long probes and/or targets are used is attributed to steric hindrance and also to the slower mass transport rate of the target toward the surface immobilized probe. 3.2.5.3 Hybridization Detection Once the target DNA has been captured onto the sensor surface, a range of different approaches can be used for transducing the biorecognition event. The transducing principles can be broadly divided into reagentless, label-free, and label-based

116

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

schemes. Labeling methods allow high sensitivity, and these approaches are developed to the point that they give reproducible results. Among a range of options, current DNA hybridization detection methods have mainly employed fluorescent labels, quantum dots, or heavy atom complex nanoparticle labels. 3.2.5.4 Electrochemical DNA Biosensor The recent appearance on the diagnostic market of electrochemical DNA microarrays (e.g., Motorola eSensor DNA Detection System41 and Xanthon Xpression Analysis System42) demonstrates the enormous potential of electrochemical DNA-based biosensors. Electrochemical devices are highly sensitive, inexpensive, easy to use, portable, and compatible with microfabrication technologies. Moreover, in contrast to optical detection schemes, the electrical responses are independent of sample turbidity. To make DNA testing more convenient, more economically feasible, and ultimately more widely used, the appealing promise of electrochemical detection technologies is thus driving an intense research effort by hundreds of laboratories worldwide. In the past few years, several excellent reviews have been published on both electrochemical and DNA sensing.43–47 Moreover, the underlying physicochemical properties of DNA and its hybridization as a basis for understanding how present electrochemical methods may enable the detection of specific DNA sequences are reported by Compton and coworkers.48 In this section, only analytical procedures based on microfluidic platform coupled to nanoparticles for hybridization electrochemical detection will be considered. Liposomes encapsulating electroactive molecules have been used as labels for DNA biosensor. A biosensor based on nucleic acid hybridization and liposome signal amplification with an integrated microfluidic system and a minipotentiostat for the quantification of dengue virus RNAwas reported by Baeumner and coworkers.49,50 An electrochemical microfluidic biosensor with an integrated minipotentiostat for the quantification of RNA was developed based on nucleic acid hybridization and liposome signal amplification. Specificity of the biosensor was ensured by short DNA probes that hybridize with the target RNA or DNA sequences. The reporter probe was coupled to liposomes entrapping the electrochemically active redox couple potassium ferri/ferrohexacyanide. The capture probes were coupled to superparamagnetic beads that were isolated on a magnet in the biosensor. Upon capture, the liposomes were lysed to release the electrochemical markers that were detected on an interdigitated ultramicroelectrode array in the biosensor just downstream of the magnet. The current was measured, stored, and displayed by miniaturized instrumentation (miniEC). The same authors have presented an optimization of their studies.51 In the previous study, IDUA were fabricated on Pyrex 7740 as a substrate and overlaid PDMS channels. In general, gold has poor adhesion properties to most surfaces including poly (methyl methacrylate). An intermediate adhesion layer of titanium is commonly used between the substrate and the gold layer. However, for an electrochemical detection system, a bimetallic system results in a galvanic cell with the less noble of the two metals being solubilized. Since it cannot be guaranteed that the adhesion layer is not

FUNDAMENTALS OF BIOSENSORS

117

coming in contact with the solution, it can result in limited lifetime of the electrode. Alternatively, mercaptopropyltrimethoxysilane (MPTMS) has been used as an effective alternative to a metallic adhesion layer by providing the substrate surface with a thiol monolayer. The gold electrodes are then adhered using gold–thiol interactions. Although limitations were found for the usable applied potential, the electrodes were more stable and had a longer lifetime compared to metallic adhesion systems. In the recent study, cystamine was conjugated to the UV modified PMMA surface using water-soluble carbodiimide chemistries, resulting in a thiolated surface. A liposomal detection system was employed to aid in signal amplification. The liposome was tagged with a DNA probe complementary to the target RNA. Superparamagnetic beads tagged with a target complementary capture probe were used to immobilize the target and the liposome complex over the IDUA (Figure 3.5a and b). One of the major trend lines toward the research of novel diagnostic systems is the concept of DNA chips (or DNA microarrays), usually associated to microfabrication of diagnostic kits by screen printing techniques, inspired by planar, silicon-based technologies. The miniaturization of DNA analytical platforms has many advantages over the conventional benchtop counterparts. These include low sample/reagent consumption (volume of microliter down to picoliter) as well as short assay time (minutes rather than days). Most important, they permit the integration of a number of functions including sample preparation, target amplification, and product detection, thus enabling a fully automated operation that can be used by untrained individuals. Amplification of nucleic acids in biomedical and biochemical researches could be used for diagnosing disease, sequencing, genotyping, and evolutionary studies. Such applications of PCR require highly sensitive, fast, selective, and accurate detection methods. Therefore, there has been recent interest in developing an accurate, sensitive, selective, and fast detection method for PCR amplification. Shiddiky et al.52 developed an electrochemical method for analyzing PCR amplification through the detection of inorganic phosphates (Pi). This method coupled a microchip to a nanoparticle comprising poly-5,20 -50 ,200 -terthiophene-30 -carboxylic acid (poly-TTCA)/pyruvate oxidase (PyO) modified microbiosensor. It detects Pi produced from the pyrophosphate (PPi), which is released as a by-product of PCR. After completion of PCR, PPi is hydrolyzed to Pi by inorganic pyrophosphatase. On the microbiosensor surface, pyruvate was converted to H2O2 by PyO in the presence of Pi and oxygen, and subsequently, the anodic current of enzymatically generated H2O2 was detected at 10.5 V versus Ag/AgCl. The CE-EC analysis was completed within 2 min. An excellent operation stability of poly-TTCA/PyO was observed for a long period of analysis. Tin-doped indium oxide (ITO) is the material of choice for fabrication of a number of optoelectronic devices. However, ITO electrodes have also found application for DNA sequence-specific detection. An array of individually addressable ITO electrodes was used as the transduction element in an integrated analytical device employed for the multiplexed detection of E. coli and Bacillus subtilis cells. The choice of an electrochemically driven immobilization strategy (electrochemical copolymerization of pyrrole and pyrrole–oligonucleotide conjugates) allowed individual positioning of

118

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

FIGURE 3.5 (a) Firstly, sandwich hybridizations bind the liposomes to magnetic beads.51 Then, the bead/liposome complex is captured by a magnet placed over the channel. Finally, a detergent is passed through the channel resulting in the lysis of the liposomes. The liposome contents are then pumped over the IDUA for concentration determination. (b) (A) SEM of a gold IDUA formed on a PMMA substrate.51 (B) A PMMA sheet containing a hot embossed channel was then bonded to the PMMA containing the IDUA. The finished device contained a 500mm channel positioned along the IDUA. (C) The finished chip containing two channels.

the predefined probes at selected ITO surfaces. Notably, such immobilization chemistry provided the probes sufficient stability to tolerate the repeated thermal cycling of in situ PCR amplification of the target DNA.53 3.2.5.5 Optical DNA Biosensor In a recent review, Krull and coworkers have discussed the application of QDs, gold nanoparticles, and molecular switches in optical nucleic acid diagnostics. The size-

FUNDAMENTALS OF BIOSENSORS

119

dependent optical properties of nanoscale materials, as well as the ability to tailor both material and surface composition of NPs, create exciting new possibilities in nucleic acid analyses. Similarly, molecular assemblies of nucleic acids that generate “on–off” responses at the single-molecule (or particle) level offer significant advantages in diagnostics 54. In nature, DNA and RNA are found at extremely low concentrations. As a result, determining the concentration of DNA, especially without amplification, is a great challenge. Many different single-molecule optical methods have been used to quantify and characterize DNA and RNA. Today, an ideal biosensor platform is required to be not only miniaturized and costefficient but also capable of simultaneous detection of multiple analytes. The current trend toward creating point-of-care molecular diagnostic biosensors and massively parallel biorecognition arrays (microarrays) has introduced new technical challenges for the probes, transducers, and their detection apparatus. In the case of nucleic acid analyses, the idea is that each probe oligonucleotide has an associated spectral code. The spectral code is created by assigning combinations of different fluorophores to a particular probe oligonucleotide. The most common form of barcodes are NP or microparticle carriers that are doped with a specific combination of fluorescent dyes or QDs. QDs are particularly well suited to barcoding since multiple colors can be excited with a single wavelength and since each QD offers a narrow, symmetric emission profile. Hybridization can be detected by labeling target with a fluorophore of either shorter (e.g., fluorescein) or longer wavelength (e.g., Cy5) than the encoding QDs. Hybridization assays have been carried out using QD-barcode technology. Combining a biobarcode with microfluidic chip-based format, Mirkin and coworkers have developed a new version of the biobarcode assay named as Genomic Bio Bar Code Assay 55 (Figure 3.6). The assay utilizes oligonucleotide-functionalized magnetic microparticles to capture the target of interest from the sample. A critical step in the new assay involves the use of blocking oligonucleotides during heat denaturation of the double-stranded DNA. These blockers bind to specific regions of the target DNA upon cooling and prevent the duplex DNA from rehybridizing, which allows the particle probes to bind. Following target isolation using the magnetic particles, oligonucleotide-functionalized gold nanoparticles act as target recognition agents. The oligonucleotides on the nanoparticle (barcodes) act as amplification surrogates. The barcodes are then detected using the scanometric method. The limit of detection for this assay was determined to be 2.5 fM, and this is the first demonstration of a barcode-type assay for the detection of doublestranded genomic DNA. B. subtilis was chosen as a model system since it is a close family member of the lethal bacterium B. anthracis, which in its spore form is the biological weapon anthrax. This work paves the way for the transition of the biobarcode assay from a laboratory technique to one that can be deployed in the field for the rapid and accurate detection of biological terrorism agents. The method proposed by Nie et al. used color-coded nanoparticles and dual-color coincidence detection to quantify an array of biomolecules including DNA, RNA,

120

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

FIGURE 3.6 Genomic Bio Bar Code Assay.56 The first step is to isolate the genomic DNA from the bacterial cells and cut it with a restriction enzyme. This cut prevents the DNA from supercoiling during heating and gives smaller target pieces. The next step is to introduce blocking oligonucleotides designed to flank the probe binding sites and prevent strand rehybridization after thermal denaturation. The blocking oligonucleotides are used in excess to outcompete the native strand during hybridization. The target region is now “propped” open and accessible for probe binding. Magnetic microparticles (oligo-MMPs) are used to capture the targets from the sample and then washed. An excess of oligonucleotide modified gold nanoparticle probes (oligo-AuNPs) is added to the assay solutions, which results in the sandwiching of the target with the oligo-MMP. Unbound oligo-AuNPs are removed by washing. The barcodes are chemically released for scanometric detection and quantification.

proteins, and viruses within a microfluidic channel. Red and green nanoparticles bind to specific sites on the target molecule. Real-time coincidence measurements allow discrimination of nanoparticle-bound target molecules and individual unbound nanoparticles as they flow through the microfluidic channel. This method allows the precise quantification of the biomolecules without the need for separation or amplification.56 3.3 CONCLUSIONS AND FUTURE TRENDS In this chapter, the new biosensor designs based on microfluidic and nanoparticles have been presented. Compared to current state-of-the art DNA detection using fluorophore labels and protein detection using ELISA, significant advances have been made in terms of sensitivity, selectivity, and multiplexing capacity. Nanotechnology is revolutionizing the development of biodevices, and it is increasingly being used to design novel bioassays with high performance. Nanotechnology-based biosensors have been integrated within tiny biochips with on-board electronics, sample handling, and analysis. This greatly improves functionality by providing devices that are small, portable, easy to use, low in cost, disposable, and highly versatile diagnostic instruments. The combination of microfluidic technologies with nanoparticles is a very promising biosensor platform, and several examples have been presented.

REFERENCES

121

A special emphasis has placed on the challenges of integrating detection platforms for encoded nanoparticles into microdevices for performing multiplexed assay. This is highly important in situations where the amount of sample is very limited, such as in the analysis of blood from newborns, tumor tissue from biopsies, and so on. In addition, multiplexing allows more efficient and therefore less expensive use of reagents, and because the different targets are screened simultaneously, they experience equal conditions at each step of the assay procedure. Overall, these microfluidic bioassays lead to inexpensive screening for multiple diseases or biomolecule states in a simultaneous fashion, which should quickly change the way in which medicine is practiced, possibly leading to a prevention mindset rather than a response after a diagnosis. Such applications will become more numerous. In any case, the impact of nanoscale sensors will have a profound effect on medical, food, and environmental testing. REFERENCES 1. de la Escosura-Muniz A, Ambrosi A, Merkoci A. Electrochemical analysis with nanoparticle-based biosystems. Trends Anal. Chem. 2008; 27(7); 568–584. 2. Gomez-Hens A, Fernandez-Romero JM, Aguilar-Caballos MP. Nanostructures as analytical tools in bioassays. Trends Anal. Chem. 2008; 27(5); 394–406. 3. Pingarron JM, Yanez-Sedeno P, Gonzalez-Cortes A. Gold nanoparticle-based electrochemical biosensors. Electrochim. Acta 2008; 53; 5848–5866. 4. Wang Z, Ma L. Gold nanoparticle probes. Coord. Chem. Rev. 2009; 253; 1607–1618. 5. Tamanaha CR, Mulvaney SP, Rife JC, Whitman LJ. Magnetic labeling, detection, and system integration. Biosens. Bioelectron. 2008; 24; 1–13. 6. Pumera M, Merkoci A, Alegret S. New materials for electrochemical sensing VII. Microfluidic chip platforms. Trends Anal. Chem. 2006; 25(3); 219–235. 7. Rıos A, Escarpa A, Gonzalez MC, Crevillen AG. Challenges of analytical microsystems. Trends Anal. Chem. 2006; 25(5); 467–479. 8. Derveaux S, Stubbe BG, Braeckmans K, Roelant C, Sato K, Demeester J, De Smedt SC. Synergism between particle-based multiplexing and microfluidics technologies may bring diagnostics closer to the patient. Anal. Bioanal. Chem. 2008; 391; 2453–2467. 9. Choi J-W, Oh B-K, Kim Y-K, Min J. Nanotechnology in Biodevices. J. Microbiol. Biotechnol. 2007; 17(1); 5–14. 10. Ligler FS. Perspective on optical biosensors and integrated sensor systems. Anal. Chem. 2009; 81(2); 519–526. 11. Lim CT, Zhang Y. Bead-based microfluidic immunoassays: the next generation. Biosens. Bioelectron. 2007; 22; 1197–1204. 12. Liu G, Lin Y. Nanomaterial labels in electrochemical immunosensors and immunoassays. Talanta 2007; 74; 308–317. 13. Tansil NC, Gao Z. Nanoparticles in biomolecular detection. Nanotoday 2006; 1(1); 28–37.

122

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

14. Wang J, Bunimovich YL, Sui G, Savvas S, Wang J, Guo Y, Heath JR, Tseng H-R. Electrochemical fabrication of conducting polymer nanowires in an integrated microfluidic system. Chem. Commun. 2006; 3075–3077. 15. Edwards KA, Baeumner AJ. Liposomes in analyses. Talanta 2006; 68; 1421–1432. 16. Klostranec JM, Xiang Q, Farcas GA, Lee JA, Rhee A, Lafferty EI, Perrault SD, Kain KC, Chan WCW. Convergence of quantum dot barcodes with microfluidics and signal processing for multiplexed high-throughput infectious disease diagnostics. Nano Lett. 2007; 7(9); 2812–2818. 17. Marquette CA, Blum LJ. State of the art and recent advances in immunoanalytical systems. Biosens. Bioelectron. 2006; 21; 1424–1433. 18. Bilitewski U. Protein-sensing assay formats and devices. Anal. Chim. Acta 2006; 568; 232–247. 19. www.diagnoswiss.com. 20. Rossier JS, Baranek S, Morier P, Vollet C, Vulliet F, De Chastonay Y, Reymond F. GRAVI: Robotized microfluidics for Fast and Automated Immunoassays in Low Volume. JALA 2008; 13(6); 322–329. 21. Tothill IE. Biosensors for cancer markers diagnosis. Semin. Cell Dev. Biol. 2009; 20; 55–62. 22. Maeng JH, Lee BC, Ko YJ, Cho W, Ahn Y, Cho NG, Lee SH, Hwang SY. A novel microfluidic biosensor based on an electrical detection system for alpha-fetoprotein. Biosens. Bioelectron. 2008; 23; 1319–1325. 23. Goluch ED, Nam J-M, Georganopoulou DG, Chiesl TN, Shaikh KA, Ryu KS, Barron AE, Mirkin CA, Liu C. A bio-barcode assay for on-chip attomolar-sensitivity protein detection. Lab Chip 2006; 6(10); 1293–1299. 24. Goluch ED, Stoeva SI, Lee J-S, Shaikh KA, Mirkin CA, Liu C. A microfluidic detection system based upon a surface immobilized biobarcode assay. Biosens. Bioelectron. 2009; 24; 2397–2403. 25. Wilson MS, Nie W. Multiplex measurement of seven tumor markers using an electrochemical protein chip. Anal. Chem. 2006; 78; 6476–6483. 26. Ko YJ, Maeng JH, Ahn Y, Hwang SY, Cho NG, Lee SH. Microchip-based multiplex electro-immunosensing system for the detection of cancer biomarkers. Electrophoresis 2008; 29; 3466–3476. 27. Tang D, Yuan R, Chai Y. Magnetic control of an electrochemical microfluidic device with an arrayed immunosensor for simultaneous multiple immunoassays. Clin. Chem. 2007; 53 (7); 1323–1329. 28. Ahmed MU, Hossain MM, Tamiya E. Electrochemical biosensors for medical and food applications. Electroanalysis 2008; 20(6); 616–626. 29. Yang L, Bashir R. Electrical/electrochemical impedance for rapid detection of foodborne pathogenic bacteria. Biotechnol. Adv. 2008; 26; 135–150. 30. Varshney M, Li Y. Interdigitated array microelectrode based impedance biosensor coupled with magnetic nanoparticle–antibody conjugates for detection of Escherichia coli O157: H7 in food samples. Biosens. Bioelectron. 2007; 22; 2408–2414. 31. Varshney M, Li Y. Double interdigited array microelectrode-based impedance biosensor for detection of viable Escherichia coli O15:H7 in growth medium. Talanta 2008; 74; 518–525.

REFERENCES

123

32. Boehm DA, Gottlieb PA, Hua SZ. On-chip microfluidic biosensor for bacterial detection and identification. Sens. Actuators B Chem. 2007; 126; 508–514. 33. Bunyakul N, Edwards KA, Promptmas C, Baeumner AJ. Cholera toxin subunit B detection in microfluidic devices. Anal. Bioanal. Chem. 2009; 393(1); 177–186. 34. Hiep HM, Yoshikawa H, Saito M, Tamiya E. An interference localized surface plasmon resonance biosensor based on the photonic structure of Au nanoparticles and SiO2/Si multilayers. ACS Nano 2009; 3(2); 446–452. 35. Jen C-P, Huang C-T, Lu Y-H. Simulation of biochemical binding kinetics on the microfluidic biochip of fiber-optic localized plasma resonance (FO-LPR). Microelectron. Eng. 2009; 86; 1505–1510. 36. Chen W, Peng C, Jin Z, Qiao R, Wang W, Zhu S, Wang L, Jin Q, Xu C. Ultrasensitive immunoassay of 7-aminoclonazepam in human urine based on CdTe nanoparticle bioconjugations by fabricated microfluidic chip. Biosens. Bioelectron. 2009; 24; 2051–2056. 37. Lucas LJ, Chesler JN, Yoon JY. Lab-on-a-chip immunoassay for multiple antibodies using microsphere light scattering and quantum dot emission. Biosens. Bioelectron. 2007; 23; 675–681. 38. www.affymetrix.com. 39. www.appliedbiosystems.com. 40. www.exiqon.com. 41. www.motorola.com/lifesciences/esensor/. 42. www.xanthoninc.com. 43. Lucarelli F, Tombelli S, Minunni M, Marrazza G, Mascini M. Electrochemical and piezoelectric DNA biosensors for hybridisation detection. Anal. Chim. Acta 2008; 609; 139–159. 44. Sadik OA, Aluoch AO, Zhou A. Status of biomolecular recognition using electrochemical techniques. Biosens. Bioelectron. 2009; 24; 2749–2765. 45. Teles FRR, Fonseca LP. Trends in DNA biosensors. Talanta 2008; 77; 606–623. 46. Navani NK, Li Y. Nucleic acid aptamers and enzymes as sensors. Curr. Opin. Chem. Biol. 2006; 10; 272–281. 47. Lee TM, Hsing IM. DNA-based bioanalytical microsystems for handheld device applications. Anal. Chim. Acta 2006; 556; 26–37. 48. Batchelor-McAuley C, Wildgoose Gregory G, Compton RG. Biosens. Bioelectron. 2009; doi: 10.1016/j.bios.2009.01.045. 49. Goral VN, Zaytseva NV, Baeumner AJ. Electrochemical microfluidic biosensor for the detection of nucleic acid sequences. Lab Chip 2006; 6(3); 414–421. 50. Kwakye S, Goral VN, Baeumner AJ. Electrochemical microfluidic biosensor for nucleic acid detection with integrated minipotentiostat. Biosens. Bioelectron. 2006; 21(12); 2217–2223. 51. Nugen SR, Asiello PJ, Connelly JT, Baeumner AJ. PMMA biosensor for nucleic acids with integrated mixer and electrochemical detection. Biosens. Bioelectron. 2009; 24; 2428–2433. 52. Shiddiky MJ, Rahman MA, Park JS, Shim Y-B. Analysis of polymerase chain reaction amplifications through phosphate detection using an enzyme–based microbiosensor in a microfluidic device. Electrophoresis 2006; 27; 2951–2959.

124

NANOPARTICLE-BASED MICROFLUIDIFIC BIOSENSORS

53. Yeung S-W, Lee TM-H, Cai H, Hsing I-M. A DNA biochip for on-the-spot multiplexed pathogen identification. Nucleic Acids Res. 2006; 34(18); e118. 54. Algar WR, Massey M, Krull UJ. The application of quantum dots, gold nanoparticles and molecular switches to optical nucleic acid diagnostics. Trends Anal. Chem. 2009; 28(3); 292–306. 55. Hill HD, Vega RA, Mirkin CA. Non-enzymatic detection of bacterial genomic DNA using the bio-barcode assay. Anal. Chem. 2007; 79; 9218–9223. 56. Agrawal A, Zhang C, Byasses T, Tripp RA, Nie S. Counting single native biomolecules and intact viruses with color-coded nanoparticles. Anal. Chem. 2006; 78; 1061–1070.

4 MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES CHUNHUI DENG

AND

YAN LI

Department of Chemistry, School of Pharmacy, Fudan University, Shanghai, China

4.1 INTRODUCTION Enzymes are one of the catalysts that are useful for substance production in an environment-friendly way and have high potential for analytical applications.1 The use of enzymes for cleavage, synthesis, or chemical modification represents one of the most common processes used in biochemical and molecular biology laboratories. The continuing progress in medical research, genomics, proteomics, and related emerging biotechnology fields has led to an exponential growth of applications of enzymes and the development of modified or new enzymes with specific activities.2 In proteomic study, the rapid development of mass spectrometry and the associated coupled technologies has provided an opportunity to perform protein and peptide separation and identification in a highly automated and rapid fashion; hence, more comprehensive protein mapping in cells or tissues can be obtained. Prior to protein characterization by mass spectrometry analysis, it is necessary to perform the controlled enzymatic degradation in a short time, resulting in well-defined and reproducible peptide patterns. The ability of some proteases to cleave polypeptide chains at specific cleavage sites makes them important tools of proteomics for elucidation of the protein primary structure or post-translational modifications.3 However, the conventional techniques of in-solution digestion of proteins suffer from limitations of long digestion time, chronic autodigestion of enzyme, and sample loss, severely affecting the determination of comprehensive proteomic profiles.

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

125

126

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

On the other hand, microfluidic devices have become powerful tools for performing chemical or biological assays in recent years. Miniaturization to the microliter scale reduces the requirements for reagents and solvents, providing flexibility in use and faster reaction times, thereby supporting high-throughput experimentation. Translation of processes from discovery to actual production could be accelerated by using automated microscale processing methods that could lead to a reduction in time and thus in cost for the product to reach the market. Microstructured reactors minimally integrate the functional microfluidic element in a suitable and appropriately interfaced housing.4 Fluid flow as well as mass and heat transport are more easily controlled on the microscale, often resulting in enhanced yield and improved selectivity compared to conventional reactors.5–7 One of the promising solutions is the incorporation by patterning processes of enzymes, such as proteases, within a microchannel to form a microfluidic enzymatic reactor to carry out highly efficient and low-level protein digestion. Implementation of enzymatic reactions in microchannels allows a decrease in the amount of consumables and sample by several orders of magnitude. Detection sensitivity can also be improved for a sample with small total volume since no dilution is necessary, and the smaller scale increases the speed of diffusion-limited reactions allowing faster assays. Finally, microfluidic enzymatic reactor may allow long-term higher automation and better reproducibility.3 The aim of this chapter is to summarize recent advances in the field of immobilized microfluidic enzymatic reactors (IMERs), which constitutes a new branch of nanotechnology. Without claiming to be exhaustive, instead, this chapter focuses on IMERs fabricated with enzyme-immobilized nanoparticles, while other formats of IMERs are briefly discussed. Enzyme immobilization techniques and the main applications of IMERs in the fields of peptide mapping, biosensing, and kinetic study are also described. 4.2 ENZYME IMMOBILIZATION TECHNIQUES Typically, enzymes have high efficiency under mild conditions and are highly selective, but they are not stable for extended time in solution and during storage their activity gradually decreases. Significant improvements in both the reaction rates (much higher enzyme/substrate ratio can be achieved) and the storage stability can be achieved with enzymes immobilized on the surface of a suitable carrier material. Immobilized enzyme reactors are considerably more stable and catalytically active for a much longer duration than free enzymes.8,9 Immobilized molecules are more resistant to the unfolding of their native structure that may be caused by heat and pH changes. Besides, although solution-phase enzymatic reactors are simple, these approaches can yield autodigestion, and may require a delicate separation between the enzyme and the reaction products. In contrast, the catalyst could be easily removed from the reaction mixture when using immobilized enzyme, thus facilitating separation of product and recycled while its activity is preserved as well as avoiding contamination of the digestion products by free enzyme molecules, which can be

ENZYME IMMOBILIZATION TECHNIQUES

127

very detrimental to analysis.10 Also, continuous processes with the biocatalyst immobilized on a solid support packed in a reactor could be designed. These implementations were believed to significantly lower the cost of the biocatalyst used per unit of product, which might represent significant savings in operations requiring expensive enzymes. Although immobilization may result in some changes in enzymatic activity, optimum pH or affinity for the substrate, elimination of enzyme from the reaction mixture, enhanced stability, reduced mixing and dilution-related problems, and the possibility to reuse the reactors outweigh these changes. Avariety of methods are now available for immobilization of enzyme on the surface of the fused silica capillary or the channel of a microfluidic chip or covalent binding to the activated supports, physical adsorption of the enzyme on a solid matrix and copolymerization of the enzyme with the polymers. The specific immobilization chemistry depends on a variety of factors, character of the support, activation methods, and coupling procedure.2 4.2.1 Covalent Immobilization The most intensively used immobilization technique is the formation of covalent bonds between the protein and the support matrix,11 for example, immobilization in the presence of carbodiimides, cross-linking by glutaraldehyde, or cyanogen bromide activation of the support material. The main advantage of covalent binding to the activated support is that it can prevent the desorption of enzyme from the support matrix in the presence of substrates and solutions of high ionic strength and of reducing spontaneous enzyme deactivation rates, as in proteases autodigestion. These benefits, involving a longer IMER lifetime, are counterbalanced by the more easily altered native tertiary structure of the enzyme with subsequent decrease in catalytic activity. In addition, the use of a covalent binding mode involves a higher enzymatic thermal stability since the strong interaction of enzyme to support causes rigidity of the protein structure and consequently limits the thermal movement of the protein at a high temperature. Therefore, the attached enzyme unfolds with difficulty, inactivation is not so easily observed, and a higher reaction rate and fewer diffusional restrictions can be achieved.12 Proteins usually have a number of potential immobilizing sites, corresponding to particular functionalities on the molecules. The functional groups of the proteins suitable for covalent binding include (i) the a-amino groups of the chain and the e-amino groups of lysine and arginine, (ii) the a-carboxyl groups of the chain end and the b- and g-carboxyl groups of aspartic and glutamic acids, respectively, (iii) the phenol ring of tyrosine, (iv) the thiol group of cysteine, (v) the hydroxyl groups of serine and threonine, (vi) the imidazole group of histidine, and (vii) the indole group of tryptophan. The most common covalent immobilization procedures are summarized in Figure 4.1.2 The most widely used, among all these methods, is based on the activation of amino supports by using glutaraldehyde. The covalent Schiff’s base formation between the aldehydic group and the e-amine group of lysine residue (Figure 4.1) is obtained under

128

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.1 Schemes of immobilization procedures for covalent attachment of proteins: (a) immobilization after support activation using cyanogen bromide; (b) immobilization after support activation using trichlorotriazine; (c) immobilization of glycoprotein via their carbohydrate moieties; (d) immobilization after support activation using glutaraldehyde; (e) immobilization via epoxy group; (f) immobilization via azlactone group; and (g) immobilization using carbodiimide as “zero linker.” Reprinted from Ref. 2, with permission.

mild reaction conditions (T, pH, and stirring) in accordance with those required for the optimal catalytic activity and enzymatic stability. However, the formation of Schiff’s base is known to be reversible and could lead to gradual release of enzyme during prolonged exposure to buffer solutions, particularly at elevated pH. Therefore, reduction of Schiff’s base double bonds using a suitable reducing agent such as NaCNBH3 has been proposed in order to produce a stable secondary amine that can tolerate pH variations.13,14 Another preferable method for the covalent immobilization of enzymes on supports is via epoxide groups. A model study with beads having identical chemistry revealed that e-amino group of lysine, indole group of tryptophan, and phenol group of tyrosine residues were mostly involved in the reaction with epoxides forming a covalent bond between the protein and the support.15 While at neutral pH the reaction is slow and the binding can take several days,16 it is much faster at pH above 9.11 Besides one-step immobilization, enzyme can also be immobilized through a multistep binding procedure. One of the most popular multistep immobilization techniques utilizing the epoxide group involves the modification of epoxide group with a diamine followed by activation using a glutaraldehyde.17,18 The disadvantage of this immobilization reaction is a potential for production of undesirable by-products, for example, homoconjugates and various polymers.11 Another multistep binding procedure involves hydrolysis of epoxide groups using hydrochloric acid or sulfuric acid.19 The hydrolysis was followed by oxidation of hydroxide groups and reaction with TPCKtrypsin molecule. To suppress the reversibility of the formed Schiff’s base and stabilize the bond with the enzyme, the immobilization was performed in the presence of a reducing agent, sodium cyanoborohydride.20 Other functional groups that usually take

ENZYME IMMOBILIZATION TECHNIQUES

129

part in covalent binding are carboxyl, diol, phenolic groups, and so on (as shown in Figure 4.1). To prevent modification of the enzymatic activity or complete inactivation of the immobilized protein, it is important that the catalytic functional groups of the enzyme are not involved in the covalent linkage to the support. Unfortunately, many of the reactive groups suitable for immobilization are often situated in the active center of the enzyme. This problem can be sometimes eliminated by immobilization in the presence of the substrate21 or competitive inhibitor22 of the enzyme. This also helps stabilize the tertiary structure of the enzyme during immobilization. Immobilization of small molecules on a substrate is typically easy; however, the active center of larger proteins may no longer be accessible after immobilization. In these cases, improvement can be achieved by introducing a spacer molecule.23 Good steric accessibility of active sites can be obtained by oriented immobilization of glycoprotein enzymes through their carbohydrate moieties.24 At the end of immobilization (alternatively during this process), unreacted active groups of solid support must be blocked by reaction with inert moieties providing that the same group was used for ligand immobilization. This blocking reaction is necessary to prevent further nonspecific reactions between support and ligand that could decrease its stability or specificity.13,25 4.2.2 Physical Adsorption Physical adsorption of the enzyme onto a solid support is probably the simplest way of preparing immobilized ligand molecule.26,27 The method is based on nonspecific physical adsorption between the enzyme molecule and the surface of the supports. Binding forces involve ionic interactions, hydrogen bonds, van der Waals forces, hydrophobic interactions, and so on. Because no reactive species are involved, the conformational changes that might result in change in the biological activity are less significant. An advantage of adsorption is that usually no reagents, and only a minimum of activation steps, are required. Unfortunately, the stability of the adsorbed layer is typically much weaker than in the case of covalent bond, and desorption of the ligand resulting from changes in temperature, pH, or ionic strength is often observed. This situation can be partially overcome by a simple regeneration achieved by the removal of the deactivated enzyme and by reloading with a fresh active catalyst.28–30 One of the commonly used physical adsorption method is based on the enzyme binding to the support media via the Lewis acid–base interaction through the divalent cation chelators such as iminodiacetic acid (IDA), which is chemically bound to the matrix. The conventional procedures for immobilization of chelator on silica matrix would be (i) the derivatization of silica with the silane agent and (ii) the chemical linkage of the metal chelator to the silane-modified silica material. Unfortunately, this immobilization approach results in low ligand densities because the secondary reaction has to be performed below pH 8 for protecting the siloxane bonds between silane and silica matrix, yet the optimum pH at 10–12.31 To solve this problem, the chelator silane reagent was synthesized at a high pH (pH ¼ 11) first, and then

130

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

immobilized them on the inner capillary wall to increase the density of immobilized IDA.30,32 The metal ion of copper and subsequently enzyme was specifically adsorbed onto the capillary30,33 or microchip32 surface to form the IMER. 4.2.3 Layer-by-Layer Assembly The technique of layer-by-layer (LBL) assembly has been developed as a versatile method to functionalize surfaces.34,35 The process is based on the sequential deposition of interactive polymers from their solutions by electrostatic, van der Waals, hydrogen bonding, and charge transfer interactions.36 Since the layer-assembled microstructures with tailored composition and architecture can be used to incorporate functional biomolecules, such as proteins, enzymes, and drug molecules,37,38 it is attractive for applications in biocatalysis, immunosensing, and other biochemical analysis. The process is superior to other techniques for preparing multilayer thin films. The assembly is based on spontaneous adsorptions, no stoichiometric control is necessary to maintain surface functionality, and the assembled films have a good thermal and mechanical stability. The physicochemical properties of films could be controlled by adjusting the deposition conditions or the outermost layer of the films.39 A pair of biomacromolecules, positively charged chitosan and negatively charged hyaluronic acid, was assembled on the surface of a PET microfluidic chip using layerby-layer deposition for the formation of a microstructured and biocompatible scaffold to immobilize trypsin40 (as shown in Figure 4.2). The constructed microreactor provides a large surface area to volume ratio and a confined microenvironment, resulting in an increased reaction rate for the sensitive proteolysis of standard proteins at lower detection limits and also of real biological samples.41 4.2.4 Biospecific (Affinity) Adsorption Compared to nonspecific adsorption, much better results can be obtained by using biospecific (affinity) adsorption, for example, the biotin–avidin or streptavidin technique.23,42 The bonds between the water-soluble vitamin, biotin, and the egg white protein, avidin, or its bacterial counterpart, streptavidin, are among the strongest known and can be used for oriented immobilization. The strength of the binding between biotin and avidin is so strong (Kd ¼ 10
15 M) that it can tolerate extreme conditions of temperature, pH, and different solvent systems.42 The biotinylated polylysine was physically immobilized on a glass surface to capture streptavidin-conjugated alkaline phosphatase.43 This microreactor was

FIGURE 4.2 Depiction of layer-by-layer process for enzyme immobilization. Reprinted from Ref. 41, with permission.

ENZYME IMMOBILIZATION TECHNIQUES

131

applied for rapid determination of enzyme kinetics. Biotinylated lipid bilayer44 and partial biotinylation by photopatterning on fibrinogen45 were also used for immobilization. The main advantage of the oriented immobilization is good steric accessibility of the active binding site; however, theses methods are not suitable for long-term use because of their instability. In addition, these applications are limited to streptavidinconjugated enzymes. Another method of immobilization using bioaffinity interaction is based on protein A. Protein A, a coat protein extracted from the bacterium Staphylococcus aureus, has the unique capacity to bind mammalian immunoglobulin G (IgG), especially Fc constant region of IgG. One can also use other bacterial Fc binding proteins, instead of protein A, such as streptococcal protein G, which has the advantage of binding to a wider range of IgG species and subclasses.24 Pearson and coworkers developed a new flow cytometry method allowing the rapid assessment of a large number of particle-bound antibodies. Protein G-derivatized POROS beads were used to bind affinity-purified antibodies specific for synthetic peptides designed from human plasma proteins.46 4.2.5 Sol–Gel Encapsulation Other approaches for immobilization may involve a physical step, for example, using the low-temperature sol–gel technologies for protein encapsulation.47,48 The reaction involves the hydrolysis and polycondensation of alkoxysilane monomers. During this process, biomolecules are entrapped in the growing gel network rather than being chemically or physically attached to the surfaces.49,50 There is no chemical bond formation between protein and polymer matrix, but enzymes can be chemically modified without loss of their activity by conjugation to one of the monomers of the polymerization mixture prior to polymerization leading to covalent bond to the matrix.51 Because encapsulation occurs under mild conditions, biomolecules retain their structure and biological activity. With regard to stability, biomolecules entrapped in sol–gel typically exhibit improved resistance to thermal and chemical denaturation and increased storage and operational stability.49,52 Owing to the properties of sol–gel such as large microstructured surface area, porous morphology, and hydrophilicity, immobilization of enzymes in the matrix could result in high amount of loading with a large extent of bioactivity remaining in the microfluidic device.53 So far, most sol–gel technologies for protein encapsulation describe the use of a silica-based sol–gel. For example, Toyo’oka’s team developed a revised version of tetramethoxysilane hydrogel for enzyme encapsulation.47 Trypsin encapsulation was carried out in a single step under mild conditions within a capillary. In a similar methodology by the same authors, the tetramethoxysilane hydrogel was applied to fabricate a trypsin-encapsulated reactor within a sample reservoir of PMMA microchip.48 This encapsulated trypsin has proven to be able to digest model peptides, which were simply electrokinetically driven through the gel, or full proteins, which were allowed to stay in the gel for 1 h. On-chip capillary electrophoresis (CE) with laserinduced fluorescence (LIF) detection was then performed to analyze the digestion products. Encapsulated trypsin was found to have a 19-fold higher activity than free

132

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

trypsin, and showed increased stability even after continuous use, compared to that in free solution. Despite the many advances that have been achieved in silica sol–gel processes for bioencapsulation, the problem of miniaturization of the sol–gel structures on microfluidic chips has not yet been resolved due to the fragility of the final gel structure, manifested by shrinkage of the gel, pore collapse, and/or poor adhesion to the substrate remains.54 Protein encapsulation techniques using the modified versions of titania and alumina sol–gel matrixes were developed by Liu et al.55,56 They could get rid of the fragility observed in silica gel structures and successfully perform the encapsulation of enzyme to construct microfluidic enzymatic reactors for peptide mapping.57 Although sol–gel encapsulation possesses many advantages, there are still some problems. The coating materials of sol–gel may hinder the conformational transition of enzyme and the transport of substrate and product,58,59 resulting in low biocatalytic activity. A promising solution is to fabricate a single-enzyme containing capsulate with a thin, permeable coating.60,61 Kim and Grate have fabricated enzyme nanoparticles via a multistep procedure including surface modification, lyophilization, polymerization in organic solvent, and shell condensation, and successfully obtained enhanced enzyme stability at an insignificant increase in mass transfer resistance.62 Ouyang et al.63 presented a two-step procedure including surface acryloylation and in situ aqueous polymerization to encapsulate a single enzyme in nanogel. Horseradish peroxidase (HRP), an enzyme widely used in bioassay and biosynthesis but fragile to phase transfer, was chosen as the model enzyme. Compared to the free HRP, the HRP nanogel exhibited similar biocatalytic behavior while significantly improving stability at high temperature and in the presence of polar organic solvent.

4.3 FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS 4.3.1 Immobilization of Enzyme on Microchannel Surface So far, open-channel reactor designs that directly immobilize enzymes on a microchannel surface have been used most frequently for fabrication of IMERs, owing to the relatively simple fabrication process. The continuing progress in microchip-based bioanalysis will depend on the development of novel surface modification technologies in a simple and reliable fashion. The chemical patterning of a biocompatible interface within microfluidic channels must be efficient to immobilize domains of antibodies, enzymes, and other important biologically active compounds for a highly sensitive detection. In particular, as protein analysis continues to push the limits, the availability of new strategies will become more critical.41 Many of the reported microreactors are based on immobilization of enzyme directly on the surface of a fused silica capillary. Amankwa and Kuhr8,42 immobilized the enzyme on the inner surface of a 50 mm ID aminoalkylsilane-treated fused silica capillary via biotin–avidin–biotin coupling. Because the enzyme was coated on the capillary wall, a very low flow rate (e.g., 40 nL min
1) was needed to permit time for

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

133

diffusion of the protein sample to the immobilized enzyme. Kuhr’s group64,65 later on found that the proteolysis reaction rate could be enhanced by applying low-power acoustic vibration to the capillary, with digestion carried out in a batch-wise procedure. Efficient tryptic digestion of large proteins has been carried out in as few as 30 min.66 The capillary microreactors were used for protein characterization by trypsin, pepsin, and carboxypeptidase Y digestion. The introduction of a functional group on the microchannel surface was also used for covalent binding of enzyme. Laurell et al. reported an enzyme reactor with high aspect ratio channels.67,68 The glucose oxidase was immobilized on the whole of the sidewalls of the microchannels by sequentially pumping the silane, glutaraldehyde, and glucose oxidase through the microchannels. Trypsin was also covalently attached to the ultraviolet (UV)-modified PMMA surface using coupling reagents N-(3-dimethylaminopropyl)-N0 -ethylcarbodiimide hydrochloride (EDC) and hydroxysulfosuccinimide (sulfo-NHS).69 The bioreactor provided efficient digestion of a test protein, cytochrome c, at a flow rate of 1 mL min
1, producing a reaction time of 24 s to give adequate sequence coverage for protein identification. Although the direct immobilization methods are easy, they are limited by both the low surface area to volume (S/V) ratio of the capillary or microchannel and long diffusion times. Because of relatively low S/V ratio and long diffusion distances, a very low flow rate was needed to guarantee sufficient time (half an hour or more) for diffusion of the protein molecules to the immobilized enzyme. The S/V ratio can be increased by using very narrow bore capillaries or microchannels, for example, with an inner diameter of 10 mm or less.70 Alternatively, problems related to long diffusion times and low S/V ratio of the open tubular bioreactor can be avoided by capillary or microchannel surface modification. The fused silica capillary was pretreated with NH4HF2, in order to increase the surface area and wettability, permitting a homogeneous spreading of the hydrophilic carboxylsilyl layer.30 The silanol groups on the roughed surface of capillary were then activated using HCl accompanied by removal of the metal ions from the capillary inner surface. After that, GLYMO-IDA-silane was introduced to react with the capillary wall, and then Cu2 þ solution and the enzyme buffer solution were introduced to form an immobilized enzyme capillary microreactor. The time necessary to complete the digestion of standard proteins in a 100 cm long capillary IMER was about 15–30 min. Ekstr€om et al. developed a modified sol–gel technique to form nanostructures on a silica microchannel surface71 that modifies the microchannel surface first with silanization using a polymerized copolymer of (3-aminopropyl)triethoxysilane and/or methylsilane, followed by glutaraldehyde activation, and final enzyme coupling. Using this method, an increased surface area was obtained, and at least 10 times more enzymes can be immobilized on these nanostructures by covalent cross-linking through amide bond formation, disulfide or His-tag, or by using a modifying succinate spacer, compared to single-layer immobilization.72 A microreactor with immobilized cucumisin on the nanostructured surface could process substrate 15 times faster than the batch-wise reaction.72 The surface of polymeric microchannel is commonly hydrophobic that results in poor wettability with aqueous solvents and promotes nonspecific protein adsorption.

134

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

It is also relatively inert to direct chemical modification. Therefore, surface modification must be performed before enzyme immobilization in polymeric microchip. Chemical modification could be performed by introducing carboxyl groups to poly (dimethylsiloxane) (PDMS) surface based on ultraviolet graft polymerization of acrylic acid.73 The covalent and physical immobilization of trypsin was carried out using activation reagent 1-ethyl-3-(3-dimethyl aminopropyl)carbodiimide (EDC)/N-hydroxysuccinimide (NHS) and a coupling reagent poly(diallyldimethylammonium chloride) (PDDA), respectively. The lab-made devices provided effective digestion of several model proteins even at the fast flow rate of 3.5 mL min
1 for the EDC/NHS-made device and 0.8 mL min
1 for the PDDA-made device, which afforded very short residence times of 5 s and 20 s. Avinyl group-containing PDMS plate was fabricated with liquid silicon rubber containing pyrogenic silicic acid as a filler.4 Apart from its effect on elastomer properties, the silicic acid is expected to provide additional silanol groups for surface chemistry. Bas-relief microstructures were incorporated every 2.5 mm along each microchannel of a PDMS plate, alternating between the left and the right wall4 (see Figure 4.3a). They were included as flow obstacles to improve mass transfer to and from the microchannel surface through a passive mixing effect.6,74 Covalent protein attachment was then utilized, via crosslinking with glutaraldehyde on the amino-silanized microstructured surface of the reaction plate. Comparison of the activated and the untreated microchannels using

FIGURE 4.3 (a) Photograph of the microstructured PDMS multichannel plate. An SEM picture of a passive mixing element is shown in the inset; electron micrographs of identical sections of (b) microstructured plates without treatment; (c) after aminosilanization and activation with glutaraldehyde; and (d) characteristically uneven distribution of immobilized enzymes across the channels of the microstructured plate. Reprinted from Ref. 4, with permission.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

135

SEM analysis revealed surface modification with a layered structure whose formation is readily explained by the reaction of the surface silanol groups and most likely requires the latter (see Figure 4.3b and c). SEM analysis was also used to characterize b-glycoside hydrolase CelB immobilization with respect to spatial distribution of enzyme along the microchannels. Protein binding was not uniform, and clusters of aggregated protein were observed in the microchannels, often coating the passive mixing elements (see Figure 4.3d)4. Poly(methyl methacrylate) (PMMA) is another most commonly used polymer substrate for IMER fabrication. Employing the strategy of silane-based chemistry to introduce a variety of active groups, a craft copolymer has been designed and synthesized, which afforded a firm, stable but easy-access silane-functionalized chemical scaffold on a PMMA-based microchannel. These silane functional groups readily reacted with the sol–gel to form a stable bonding through a silicon–oxygen–silicon bridge. Thus, anchorage of proteins could be realized on the hydrophobic PMMA microchannels while preserving bioactivity.53 Surface modification methods incorporating the sol–gel technique were also developed for IMERs fabricated with polymer substrates. A trypsin-encapsulated titania (titanium dioxide) and alumina gel matrix was immobilized through the SiOH group formed on a PDMS surface by plasma oxidation.57 These SiOH groups act as anchors on the microchannel wall linked covalently to the hydroxyl groups of trypsinencapsulated sol matrix. As a result, the trypsin-encapsulated gel matrix was anchored to the wall of the microchannel, and the leakage of gel matrix from the microchannel was effectively prevented. Using this device, digestion times were significantly shortened (2 s) and the application for high-throughput protein identification was realized. On the other hand, current chip fabrication protocols generally do not allow or will destroy organic coatings when the two halves of the system are bonded. This means that organic coatings must be applied to the fully fabricated channel system, commonly by flowing reactants into the chip. The problem with this approach is that reagents are dispersed throughout the channel system and all channels are coated. This is undesirable in the case of some types of immobilized enzyme reactors. The problem was addressed by using electroosmotic flow to direct reagents to specific channels in a channel network.75 The route of transport, and thus the specificity of channel coating, was controlled by the well to which negative potential was applied. Flow in a multichannel network took the shortest route between the electrodes delivering the motive potential. Different reagents in the reaction were delivered from different wells and took different paths through the channel network. Only the separation channel was in the flow path of all the reagents used in the coating process and thus had channel-specific immobilization of the enzyme (Figure 4.4). Rubloff et al. developed a methodology that enabled the programmable assembly of biomolecules on localized assembly sites in microchannels using electrodeposition of the amine-rich polysaccharide chitosan to direct the assembly.76 They further demonstrated that a metabolic pathway enzyme, S-adenosylhomocysteine nucleosidase (Pfs), could be assembled in this way and that its catalytic action was retained in the microfluidic environment, shown by conversion of substrate S-adenosylhomocysteine

136

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.4 The flow direction of (a) the first-step chemical modification and (b) the second-step chemical modification. The arrow indicates the direction of electroosmotic flow. Reprinted from Ref. 75, with permission.

(SAH) to products S-ribosylhomocysteine (SRH) and adenine (Figure 4.5a).77 While promising as a methodology to replicate metabolic pathways and search for inhibitors as drug candidates, these investigations also revealed unintended (or parasitic) effects, including products generated by the enzyme (1) in the homogeneous phase (in the liquid) or (2) nonspecifically bound to microchannel surfaces. To reduce homogeneous reactions, a new packaging and assembly strategy was developed that

FIGURE 4.5 Minimize parasitic reactions by eliminating interconnect reservoirs and by separating sequential flow directions in cross channels. To test the background signal by parasitic reactions, Pfs enzyme solution was introduced without electroassembly followed by buffer rinsing, then enzymatic substrate SAH was introduced, and products were collected downstream to be analyzed by HPLC. (a) Single channel with interconnect reservoirs. (b) Single channel without interconnect reservoirs. (c) Cross channel without interconnect reservoirs. Reprinted from Ref. 78, with permission.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

137

eliminated fluid reservoirs that were commonly used for fluidic interconnects with external tubing (Figure 4.5b). To suppress reactions by nonspecifically bound enzyme on microchannel walls, a cross-flow microfluidic network design was implemented so that enzyme flow for assembly and substrate/product for reaction shared only the region where the enzyme was immobilized at the intended reaction site, as illustrated in Figure 4.5c.78 4.3.2 Packing Microchannels with Enzyme-Immobilized Micro/Nanoparticles To overcome the limitations caused by low specific inner surface of microchannels, besides surface modification, another alternative is to immobilize enzyme molecules on the surface of different supports (micro/nanoparticles), followed by fixing them in the microchannel by weirs, frits, or membranes. The procedures for fabrication of IMERs with packing micro/nanoparticles include choosing suitable supports, immobilization of enzyme on the support surface, and assembling the supports into the microchannel. Immobilization methods have been described in Section 4.2. Immobilization of enzymes on supports can be obtained namely “in situ” or “in batch.” When employing the in batch process, the enzyme is first immobilized on the support and then packed into the microchannel using a slurry packing technique, whereas in the in situ approach the enzyme is directly immobilized on the prepacked column. Although the in-batch process takes the advantages of reduced cost due to mass producibility, better reproducibility, and quality control, it could result in a loss of catalytic activity. Massolini et al.28 have compared in batch and in situ techniques of penicillin G acylase (PGA) immobilization on various derivatized silica supports concluding that the in situ technique is the best way to obtain satisfactory results in terms of bound amount of PGA and enzymatic activity retention. Choosing the right supports is vital and gets more and more complicated because of the increase in the number of natural and synthetic supports available that greatly differ in mechanical and physical properties. Furthermore, it is also necessary, at the same time, to meet many requirements such as low cost, non toxicity, maximum activity, high retention of catalytic activity over a long period, enzymatic stability, ease of protein availability, and immobilization. The surface of the supports, on which the enzyme is immobilized, has an important role to play in retaining the tertiary structure of the enzyme that highly influences the thermal stability and catalytic activity of the immobilized enzyme. Indeed an immobilized enzyme is known to acquire novel kinetic properties that can modify the Michaelis–Menten constant (Km) and maximum velocity (Vmax) and cause a shift of the pH and temperature-activity profile. Likewise, groups involved in the attachment of proteins to the support must be different from the active sites of enzymes. Therefore, the choice of both the support and the technique depends on the nature of the enzyme, on the nature of the support, and on its ultimate application. For this reason, it is not possible to recommend any universal immobilization methods. The toxicity of immobilization reagents should also be considered when final applications concern the food processing and pharmaceutical industries.12 An important factor in the preparation of a bioactive reactor is the structure of the support since this determines accessibility of active sites to substrates. The ideal

138

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

support must be inert, stable, and resistant to mechanical strength. However, the other physical properties, such as form, shape, porosity, pore size distribution, swelling capability, and charges, are also very important because they influence the kinetic process. Indeed, the reaction rates of the immobilized enzyme depend on the enzymatic intrinsic activity, on the substrate accessibility to interact with the active sites, on the amount of the loaded enzyme, and on the substrate concentration and diffusivity. The substrates must be able to diffuse from the bulk phase toward the surface and product away (external diffusion) and within the pores of immobilized enzyme particles (internal diffusion). These effects of diffusion, which are enhanced when the enzyme is entrapped within a matrix, limit the reaction rate because they affect the concentration of the substrate/cofactor in the vicinity of the enzyme. In this way, a diffusional layer around the immobilized enzymes is formed and its thickness is correlated with mass transfer effects. Thin diffusion layer, as opposed to a thick layer, results in a low diffusional resistance. Methods to minimize the diffusional effects could be the decreasing of enzyme loading and the increasing of substrate concentration and diffusivity. The latter feature is strongly influenced by a hydrodynamic parameter such as the flow rate that, as it increases causes a decrease in the diffusional layer. Care must be taken also in selecting the support materials because their characteristics strongly influence, as previously mentioned, the accessibility of active sites to substrates. For example, the more the pore diameter and size distribution increase the more the surface area decreases. Therefore, it is generally preferable to choose pores with a small diameter if the substrate has similar molecular dimensions. If substrates with high molecular weight are implicated in the enzymatic reaction and their diffusion in the active site is sterically hindered, a significant intraparticular mass transfer resistance, which in turn significantly decreases the overall reaction rate, must be evaluated. Another feature influencing the enzymatic activity is the particle size, as is known, the bigger the particle size, the greater the effect of diffusion control and less the activity. To make the correct choice, it is also important to consider the relation of particle size with pressure drop that are correlated in an inverse mode. Therefore, the evolution of microchannel packing material is to minimize the diffusional limitations by decreasing the size and optimizing the geometry of immobilized biocatalyst particles, by decreasing the substrate concentration, by enhancing the flow rate, by increasing the porosity, and by optimizing the biocatalyst distribution in the beads. 4.3.2.1 Nonmagnetic Supports Inorganic Supports The main supports used in enzyme immobilization are porous inorganic solids such as the controlled pore glass (CPG) and silica. CPG presents a higher thermal stability and a resistance to acids, whereas silica is characterized by a larger specific surface area. Both supports must be derivatized with functional groups that can interact covalently with enzymes. This feature may be obtained in the

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

139

laboratory or by using a commercial derivatized support.12 When uncoated CPG is employed, the researchers use 3-aminopropyltriethoxysilane (APTES)79,80 to obtain an aminopropyl-CPG that could be further activated with glutaraldehyde and then reacted with enzyme.81 Microreactors with enzymes immobilized on glass beads have been prepared by filling the reaction chamber with beads; such a device was used for the determination of xanthine using chemiluminescent detection82 and for online protein digestion.81 In the case of uncoated silica, glicidoxypropyltrimethoxysilane83 is used to obtain epoxy-silica. Organic Supports When the enzyme requires an alkaline medium, polymeric supports are employed. Among these organic solids, poly(vinyl alcohol) (PVA) activated with tresyl group is most used in in situ enzyme immobilization. The performance of these phases depends on the kind of enzyme and on the number of injections. In the case of leucine and 3-hydroxybutyrate dehydrogenase coupled to NADH oxidase, a satisfactory reproducibility is obtained within 10 days after 300 and 400 injections, respectively, of samples stored at 4 C when not in use.84 Crooks and Seong developed advanced analytical microreactors using enzyme-immobilized microbead mixing,85 and efficiently performed multistep enzyme reactions using glucose oxidase and horseradish peroxidase immobilized on polystyrene. Furthermore, the immobilization of enzymes on nickel–nitrilotriacetic acid (Ni–NTA) agarose beads has also been reported and was applied to immobilize bacterial P450;86 this immobilized enzyme was less denaturated because binding of the enzyme was achieved using a histidine (His)-tag. Many methods have been reported for packing enzyme-immobilized supports into microchannels. Microfabricated weirs87,88 or elevated structures (“bead stopper”)89 have been developed for keeping the supports packing in place. Andersson et al. reported micromachined chambers surrounded by filter-like structures used for solidphase DNA sequencing.90 Sato et al. fabricated a barrier in a microchannel for blocking derivatized beads used in chip-based immunoassays.91 A PDMS polymer device for an on-chip, fritless, capillary electrochromatography has been realized, characterized by a tapered column in which stationary phase particles were retained.92 Fabrication using deep reactive ion etching was required to produce the silicon master. Although successful to various extents, reported approaches often encountered difficulties in the fabrication of on-chip packed column reactors, and procedures commonly involved the use of sophisticated equipment, such as deep reactive ion etching92 or microfabrication techniques involving multiple exposures to achieve multiple-layered etching.89 Zhang et al. developed a novel fabrication approach that enabled the enzyme-immobilized microbeads with controlled sizes to be placed at any desired position on the microchip.93 The location of entrapped beads acting as the support of enzymes can be easily controlled by spotting the slurry of beads at desired position on the separation channel. The length of the immobilized beads region was determined by the diameter of the spot, the low limitation of which is 0.5 mm using a 10 mL pipette tip. Different linear ranges of the biosensor can be obtained with varied lengths. And the width of the region was controlled by the width of the separation channel (200 mm). The concentration of microbeads slurry controlled the coverage

140

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

degree of microbeads immobilized on the surface of microchannels. The coverage degree increased with the concentration of the slurry (Figure 4.6). Even when successfully produced, the high backpressure generated even at low flow rates often limits the use to relatively short column beds. To reduce the backpressure, functional groups could be directly attached to the walls of open-channel reactors. However, such an approach significantly reduced the reactive capacity of the reactor even after the surface modification. To decrease the flow resistance of the nanoparticle-packed IMERs while maintaining reasonably high reactive area, enzyme-immobilized nanoparticles could be immobilized on the microchannel walls, and left considerable space above the particles for low-resistance flow. CPG reactive particles were immobilized on PDMS microreactor beds94 using a three-layer design that was soft enough to allow implanting of the particles by applying gentle pressure over the particles (Figure 4.7). Following full polymerization, the particles were permanently immobilized. The curing at a temperature of 37 C for 5 h produced no deleterious effects on the activity of the immobilized glucose oxidase (GOD) particles. Silica nanoparticles were immobilized on the surface using slow evaporation of the particle suspension in a filled-in microchannel.95 The resulting microchannel was subjected to treatment with 3-aminopropyltriethoxysilane, and immobilization of enzyme was achieved by covalent cross-linking through an amino group. Although physical stability needs to be improved, a lipase-immobilized microreactor prepared by this method showed 1.5 times faster kinetics than those of a microreactor obtained by sol–gel surface

FIGURE 4.6 (a) Optical (left, A and C) and corresponding fluorescence images (right, B and D) of the immobilization of microbeads with different concentrations. The concentrations in A and C were 1 and 30 mg mL
1, respectively. The white dots in optical images and the green dots in fluorescence photos were the immobilized microbeads bonded with FITC-BSA. (b) Optical image of microchannels with multireactors. The white part in the image represents the enzyme reactors. Reprinted from Ref. 93, with permission.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

141

FIGURE 4.7 CCD images of (a) the GOD-CPG particles (200–400 mesh) immobilized on the PDMS surface of a section of the reactor and (b) a cross section of the immobilized particle bed. Reprinted from Ref. 94, with permission.

modification.96 This result showed good correlation with the surface area: particle arrangement has approximately 1.5 times larger surface area and could immobilize more enzymes. Because of its good biocompatibility, charge-stabilized gold nanoparticles (AuNPs) provide a mild microenvironment similar to that of proteins in native states and give protein molecules more freedom in orientation. They are biocompatible and nontoxic and can offer large specific surface areas for ready binding of a large range of biomolecules such as amino acids, proteins, and antibodies.97 PDDA/AuNP multilayer films containing protease were assembled on the surface of poly(ethylene terephthalate) (PET) microchannels to obtain a flow-through protein digestion biochip. Sequential alternate adsorption of the cationic polyelectrolyte PDDA and the anion-coated AuNPs led to the formation of a biocompatible and large specific surface to volume ratio network to immobilize the enzymes.98 Zeolite nanoparticles have been widely studied in the last decade and have drawn much interest due to their large external surface area compared to conventional zeolite crystals, high dispersibility in both aqueous and organic solutions, high thermal and hydrothermal stabilities, and tunable surface properties such as adjustable surface charge and hydrophilicity/hydrophobicity.99 The unique properties make nanozeolites promising candidates for microfluidic surface modification and enzyme immobilization. Silicalite-1 (S-1, all-silica MFI-type zeolite nanoparticles) was selected and successfully used to modify the PMMA surface. The silanol groups were introduced and readily reacted with sol–gel to form stable microstructure matrices in microchannels (Figure 4.8). Trypsin was then stably immobilized within the PMMA microchannel to fabricate an enzymatic on-chip microreactor.100 4.3.2.2 Magnetic Supports As mentioned above, difficulties were often encountered in packing of the enzymeimmobilized nanoparticles in the microchannel. Besides the requirements of elaborate in-capillary/microchannel chemistry, the reproducible filling of the microchannel still remains a challenge. During the past decade, magnetic nanoparticles are gaining increasing attention due to their ease of manipulation and recovery. On this basis, magnetic nanoparticles have many unique magnetic properties such as superparamagnetic, high coercivity, low Curie temperature, high magnetic susceptibility, and so on. Therefore, they are of great interest for researchers from a broad range of disciplines,

142

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.8 SEM images of microchannel modified with silicalite-1. (a) Planform of the channel at low magnification. (b) Planform of the channel at high magnification. (c) Cross section of modified microchannel at high magnification. Reprinted from Ref. 100, with permission.

including magnetic fluids, data storage, catalysis, and bioapplications.101–105 Thanks to the unique magnetic properties, the magnetic nanoparticles offer the advantage of straightforward and fast handling by using magnets or magnetic coils. Accordingly, the use of magnetic nanoparticles, which does not require elaborate in-channel chemistry or the use of sophisticated equipment for reversible packing of the supports in the microchannel, is shown to have significant potentials in microfluidic device.106,107 Generally, magnetic nanoparticles were prepared by encapsulating inorganic magnetic particles (usually magnetite or maghemite) with organic materials or inorganic materials, such as polymers, silica, metals, metal oxide, and so on.108,109 Synthesis and surface fictionalization strategies of magnetic nanoparticles were discussed by Jiang et al.110 Polymer Encapsulation Among the different organic materials that can be used to encapsulate the magnetic nanoparticles, polymers are of particular interest because of their wide range of properties. Polymer coating will increase repulsive forces to balance the magnetic and the van der Waals attractive forces acting on the magnetic nanoparticles. In addition, polymer coating on the surface of magnetic nanoparticles offer a high potential in the application of several fields. To use these materials for

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

143

fundamental or applied research, access to well-defined magnetic nanoparticle samples whose properties can be “tuned” through chemical modification is necessary. In a number of cases, it has now been shown that, through careful choice of the passivating and activating polymers and/or reaction conditions, can produce magnetic nanoparticles with tailored and desired properties. The use of polystyrene-encapsulated superparamagnetic beads (2.8 mm diameter) has been explored with commercial CE instrumentation for performing enzymatic and inhibition assays, as well as for analysis of biological molecules such as antigens and substrates.107 Trypsin was immobilized on nonporous COOH-functionalized polystyrenic magnetic particles (626 nm) and was used for protein digestion.111 Magnetic particles coated with poly (N-isopropylacrylamide), polystyrene, poly(2-hydroxyethyl methacrylate-co-ethylene dimethacrylate), poly(glycidyl methacrylate), [(2-amino-ethyl)hydroxymethylen]biphosphonic acid or alginic acid were utilized for trypsin immobilization and organized by an inhomogeneous external magnetic field in the microchannel.3 Silica Encapsulation Silica is also used for preparing the functionalized magnetic nanoparticles. Easily replaceable and regenerable IMERs have been fabricated with packing bed of magnetic silica nanoparticles that immobilized trypsin by metal-ion chelated adsorption32,33 or covalent binding.13 Magnetic nanoparticles with small size (300 nm in diameter) and high magnetic responsivity to magnetic field (68.2 emu g
1) were synthesized and modified with tetraethyl orthosilicate (TEOS) (Figure 4.9a and b).

FIGURE 4.9 Schematic illustration of trypsin immobilization on magnetic silica nanoparticles with (a) metal-ion chelated adsorption,32 and (b) covalent binding,13 and (c) aminefunctionalized magnetic nanoparticles.14 Reprinted from Refs 32, 13, and 14, respectively, with permission.

144

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

For metal-ion chelated adsorption of enzyme, the metal chelating agent of iminodiacetic acid was reacted with glycidoxypropyltrimethoxysilane (GLYMO) before its immobilization on the surface of magnetic silica nanoparticles. The metal ion of copper and enzyme were subsequently adsorbed onto the surface (Figure 4.9a).32,33 For covalent binding of enzyme, aminopropyltriethoxysilane and glutaraldehyde (GA) were introduced to functionalize the magnetic silica nanoparticles. Trypsin was then stably immobilized on the magnetic silica nanoparticles through the reaction of primary amines of the proteins with aldehyde groups on the magnetic silica nanoparticles (Figure 4.9b).13 However, despite popularity of silica stationary phases in chromatography, application of silica-based supports for the immobilization of proteins traditionally lags behind the use of organic polymers. The reason may be the limited arsenal of reactive chemistries available for silica, the danger of nonspecific interactions with surface silanols, and the limited hydrolytic stability of the support. The above method proved to be effective and has been successfully used for protein digestion. However, when using these methods, multiple steps of surface modification on the magnetic microspheres were required prior to trypsin immobilization, which resulted in a complicated and time-consuming procedure. Koh et al.112 modified commercially available magnetic nanoparticles with APTES and then activated them with glutaraldehyde prior to the immobilization of proteins on the nanoparticle surface. Their work offered a simpler way of preparing protein-immobilized magnetic nanoparticles; however, five reaction steps are still needed. The work done by Nishimura et al.113 involves the in situ preparation of magnetic nanoparticles (mixture of Fe3O4/g-Fe2O3) in the presence of trypsin at 4 C by chemical coprecipitation of FeCl2 and FeCl3 using NH4OH as precipitator). This approach can directly lead to trypsin-modified magnetic nanoparticles; however, because the synthesis temperature is too low, the obtained magnetic nanoparticles have poor crystallization as indicated by the X-ray diffraction spectrum. As a result, the magnetic nanoparticles possess poor magnetic response that may influence the practical application. In addition, because the magnetic nanoparticles were synthesized in the trypsin aqueous solution, the location of trypsin molecules in the trypsin-modified magnetic nanoparticles is ill defined. Li et al. reported a novel and facile way of the preparation and application of trypsin-immobilized magnetic nanoparticles with superparamagetism.14 First, aminefunctionalized magnetic nanoparticles were prepared through facile one-pot solvothermal synthetic strategy. Then, magnetic nanoparticles were functionalized with numerous aldehyde (–CHO) groups followed by immobilization of trypsin through reaction of the aldehyde groups with amine groups of trypsin (Figure 4.9c). These trypsin-immobilized magnetic nanoparticles were also successfully used for the preparation of an easily replaceable on-chip enzymatic microreactor.114 One of the most interesting properties of a suspension of superparamagnetic particles is its ability to self-organize in a magnetic field. When exposed to a uniform external magnetic field, the magnetic particles acquire a magnetic moment. The resulting dipole interactions cause an instantaneous self-organization of the suspension into a structure consisting of a columnar clustering in the direction of the field. These columns are in turn organized in the direction perpendicular to the field, in

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

145

structures that depend on the container geometry, particle density, and magnetic field history. The magnetic field can be provided by permanent magnets or magnetic coils. The enzyme-immobilized magnetic nanoparticles were locally packed into the microchannel by the application of one permanent magnet115 or two magnets. Two permanent magnets could be placed in a repulsive conformation, with the polarization making a small angle (30 ) with a straight channel that created a magnetic field parallel to the flow with a strong gradient pointing through the center of the chip channel. In the beginning, particles self-organize in chain-like columns along the channel direction (Figure 4.10, inset). When the concentration increases, the plug becomes opaque, probably due to the formation of a “labyrinth like” structure116 made of tortuous and ramified “walls” with one direction collinear to the field. The distance between walls is maintained by dipole–dipole repulsion, keeping in the bulk of the plug channels collinear to the flow, with a thickness of a few micrometers.111 Things were different when the two magnets were placed with opposite poles facing each other perpendicular to the channel axis.3 When approaching the magnets, the nanoparticles got magnetized and arranged in free-floating growing chains oriented perpendicular to the flow field. In about 60 min, these chains close to the center finally staggered to a dense plug of particle clusters (Figure 4.11).3 The time needed for formation of packing bed could be reduced within 1 min by using magnetic nanoparticles with higher magnetic responsivity to magnetic field.13 4.3.3 Monoliths In recent years, monolithic phases have emerged as an attractive and increasingly more popular alternative to packed columns due to simplicity of preparation and virtually unlimited choice of chemistries they offer. In addition, there is no need for retaining frits, and very fast separations can be achieved due to the typically lower flow resistance even with smaller pore sizes. Perhaps the most appealing aspect of

FIGURE 4.10 Microreactor with the plug of magnetic beads maintained between the two magnets; the inset is a 1006 microphotograph of the columns at the beginning of the formation of the plug. Reprinted from Ref. 111, with permission.

146

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.11 Filling of microchannel by magnetic nanoparticles. (a–d) Obtained after 10, 30, 45, and 60 min; flow rate of mobile phase 1 mL h
1. Reprinted from Ref. 3, with permission.

monolithic materials is the ease of preparation. The simple polymerization process starts from liquid precursors (polymerization mixture) and is performed directly inside the capillary or a microfluidic chip. In contrast to packed beds, monolithic structures exhibit excellent dimensional stability. The through-pores of monolithic materials can be easily controlled allowing high-speed flows at low backpressures and the surface of the monoliths can be easily chemically modified. Such flexibility is ideal for designing and developing the enzymatic reactor tailored for specific applications. Petro et al.17 described comparative studies in which trypsin was immobilized on both macroporous poly(glycidyl methacrylate-co-ethylene dimethacrylate) (poly(GMA-co-EDMA)) beads and on chemically analogous monolith. Monolith and beads were modified by a multistep process involving the modification of epoxide groups with ethylenediamine followed by activation using glutaraldehyde and final modification with trypsin. Despite the relatively small size of the monodisperse beads used to minimize the diffusional path length, the processivity of the enzyme immobilized on the monolithic material was nearly two times higher compared to that of the bead-based conjugates.17 Immobilized proteolytic enzyme reactors with monoliths as supports have been summarized by Svec.117 The current monoliths can be subdivided in two categories (inorganic monoliths and organic synthetic polymer-based monoliths) according to the material from which they are prepared. The most frequently used inorganic monoliths are silica monoliths, while some other inorganic monolithic materials have also been exploited for enzyme immobilization. Yi et al. developed a novel immobilized trypsin reactor with titania monolith as the carrier.118 The material was prepared from biocompatible precursors

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

147

using aqueous processing conditions involving the formation of a glycerol–titania composite sol and titania condensation. By adding poly(ethylene oxide), macroporous titania monolith was obtained. g-Glutamyl transpeptidase, a clinically relevant protease, was then entrapped in the monolithic network. 4.3.3.1 Silica Monolith The conventional silica-based monolithic columns are usually prepared by sol–gel approach, in which a porous silica rod could be formed by the hydrolytic polycondensation of alkoxysilane. Due to the existence of micrometer-size flow-through pores constituting a macroporous network and nanometer-size mesopores on the skeleton, as shown in Figure 4.12,119 silica-based monoliths have various merits, such as low backpressure drop across the column, good permeability, and fast mass transfer kinetics. So far, two different approaches have been used for the immobilization of proteins on silica monoliths: activation of preformed silica monoliths followed by enzyme immobilization or entrapment via sol/gel. Activated Silica Column Calleri et al.120 used a thoroughly dried commercial 4.6 mm ID monolithic silica column and activated its pore surface by reaction with 3-glycidoxypropyltrimethoxysilane in toluene. A combination of these two methods enabled good characterization of the surface functionalities. b-Glucuronidase was also immobilized on a silica monolith modified with 3-APTES and activated with N, N0 -disuccinimidyl carbonate, and was used for determination of dextromethorphan and dextrorphan in urine.121 Less typical is the immobilization of ascorbate oxidase by physical immobilization on plain silica monolith prepared in situ in poly(ether ether ketone) (PEEK) capillary and the use of the reactor for monitoring dopamine in the presence of ascorbic acid.122 Presumably, the enzyme interacts with acidic silanol functionalities. However, these coulombic forces are not very strong and the lifetime of the reactor can be significantly impaired. Encapsulation in Sol–Gel The simplest, yet least used approach to immobilization via sol/gel transition is encapsulation of the enzyme within the newly formed silica matrix. Kawakami123 used immobilized protease P in a monolith formed in PEEK capillary to afford a reactor for transesterification of vinyl butyrate. Kato and

FIGURE 4.12 SEM picture of the typical porous structure of (a) monolithic silica columns, (b) the mesoporous structure of the silica skeleton, and (c) the macropores or through-pores. Reprinted from Ref. 119, with permission.

148

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

coworkers developed a simple in situ encapsulation procedure to prepare the immobilized trypsin reactor.124 After mixing with a fully or partially hydrolyzed silane, trypsin could be well encapsulated in the hydrogel after several days. The enzymatic activity of the resultant monolithic reactor was about 700 times higher than that in free solution. It was noteworthy that, by immobilizing trypsin in the monolith located at the upstream of a separation capillary, the authors enabled the enzymatic digestion and CE separation in a single capillary. Thereafter, they improved this technique by coating trypsin-containing gel on a porous silica monolith,125 which was subsequently fitted into a 96-well plate for high-throughput proteome analysis. It was found that the encapsulated trypsin within the gel matrix could offer high catalytic turnover rate due to the large surface area of monoliths. All these experiments were carried out in the batch mode relaying on slow diffusion to achieve the mass transfer of both substrate and products. As a result, the digestion is slow. Although interesting, this approach does not fully utilize the potential of monolithic supports well demonstrated in flowthrough applications. Photopolymerized Sol–Gels In contrast to the classical preparation of inorganic monoliths using hydrolytically initiated polycondensation of alkoxysilanes, Zare’s group126 introduced a novel approach to inorganic monolithic columns for CEC combining photopolymerization and sol–gel transition. The monolith is obtained in a single step during which both addition polymerization and polycondensation of [3-(trimethoxysilyl)propyl] methacrylate simultaneously proceed in the presence of a porogen. This photochemical route facilitates the exact placement of the monolith within the device and the resulting material exhibits a high mechanical strength. Perhaps the major advantage of this approach is that there is no need for drying at high temperature that may lead to cracking of the monolith. This technique was demonstrated with a capillary integrating protein digestion into an immobilized enzyme reactor with electrophoretic separation of the digest and mass spectrometric detection of peptides.127 Avery high percentage of protein in this solution (25%) was required to achieve sufficient activity of the immobilized pepsin. Later on, they enhanced the activity of enzymatic reactor by covalently bonding trypsin to such monolithic silica via Schiff chemistry at room temperature, in which an alkoxysilane reagent with an aldehyde functional group links to an inactive amine on trypsin to form an imine bond.128 The results suggested that the proteolytic activity of such an immobilized trypsin was increased by 2000-fold compared to that obtained in solution. Another elegant approach includes the preparation of a 1 cm long monolithic plug in a capillary via photopolymerization of a mixture of condensed [3-(trimethoxysilyl) propyl] methacrylate and PEG dimethacrylate followed by functionalization with a toluene solution of (trimethoxysilyl) butyraldehyde passing through the monolith for 2 h.128 After a thorough wash with ethanol, the pores were filled with trypsin solution and the immobilization reaction allowed to proceed for 19 h. 4.3.3.2 Organic Polymer Monoliths Due to a wide variety of chemistries and formats readily available, organic polymers are very popular supports for enzyme immobilization.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

149

Acrylamide Copolymers The use of acrylamide gels is almost synonymous with enzyme immobilization via entrapment.129 Polyacrylamide is hydrophilic, that is, protein-friendly; its cross-linking density is readily controlled through the percentage of bisacrylamide cross-linker, and the redox-initiated polymerization proceeds at the room temperature. Mersal and Bilitewski130 used polyacrylamide gel in both capillaries and microfluidic chips for entrapment of glucose oxidase and determination of glucose. They also admixed acrylic acid in the polymerization mixture to generate EOF and drive the analytes through the device. Palm and Novotny131 developed an interesting approach with single-step fabrication of immobilized trypsin reactor in a capillary using their acrylamide-based monolith. The polymerization mixture consisting of monomers acrylamide, methylenebisacrylamide, and N-acryloylsuccinimide dissolved in buffer solution together with PEG and a redox initiator (TEMED and ammonium peroxodisulfate) was mixed with trypsin. The polymerization reaction and enzyme immobilization could be completed within 2 h for the high activity of NAS to enzyme. This kind of monolithic reactor offered high flow permeability and biocompatibility. Using the same singlestep technique, Palm and Novotny132 also immobilized peptide-N-glycosidase and used the conjugate for deglycosylation of proteins. However, the content of monomers in the polymerization mixture was restricted by the poor solvability of monomers in aqueous solution, resulting in a loose structure and low amount enzyme immobilized.133 Obviously, two processes related to the enzyme can be envisioned to simultaneously proceed during the polymerization: (i) reaction of lateral nucleophilic functionalities of trypsin with succinimide moieties in both monomeric and already polymerized form and (ii) entrapment of the enzyme in the acrylamide matrix. The extent of each of these processes is difficult to judge since no comparative experiments without the succinimide monomer were performed. Glycidyl Methacrylate Copolymers DIRECT IMMOBILIZATION VIA EPOXY FUNCTIONALITIES Epoxide groups represent a common tool in the field of enzyme immobilization. Using this reaction, Ben
cina et al.134 immobilized protein A, deoxyribonuclease, and trypsin on 3 mm thick 12 mm diameter poly(GMA-co-EDMA) disks using both static (disk immersed in the protein solution) and dynamic (solution of protein pumped through the disk) techniques. Comparative experiments clearly demonstrated significant effect of benzamidine on the activity of the immobilized enzyme. HYDROLYSIS FOLLOWED BY OXIDATION Since the direct reaction of epoxide functionalities with proteins is slow, several alternative approaches have also been developed. One of the oldest, which found its inspiration in the area of polysaccharide-based supports,135 includes hydrolysis of the epoxide ring to a 1, 2-diol and its oxidation using periodate. Detailed studies of both these reactions were carried out with poly (GMA-co-EDMA) beads almost three decades ago.136,19 The protein predominantly reacts with this support via its lateral primary amine groups of lysine residues. Since the imine --C¼N-- double bond that forms is not very stable and easily undergoes

150

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

hydrolysis liberating again the immobilized protein, hydrogenation with sodium cyanoborohydride is often used to convert this group to a secondary amine functionality --C--NH--. Luo et al.137 used this method to activate poly(GMA-co-EDMA) monolith and immobilized papain on this support. The amount of papain immobilized on the 50  4 mm ID monolith was 7.1 mg g
1, but its efficiency was rather low. This was ascribed to the steric hindrance resulting from the close vicinity of the enzyme to the surface. Other explanation could be a multipoint immobilization that would change the tertiary structure of the enzyme and deform the active site with the concomitant decrease in activity. K
renkova et al. prepared a capillary enzymatic reactor with covalent immobilization of trypsin on poly(GMA-co-EDMA) monolith using this hydrolysis followed by oxidation method.20 For comparison, they also immobilized trypsin on monolith by a single-step binding procedure. Since the reaction of carboxylic functionalities with epoxides is not very efficient, they have attempted to mask the epoxide groups with amino groups of aspartic acid to increase the hydrophilic character of the pore surface and immobilize trypsin through reactions of epoxide functionalities with amino groups of the protein molecule. No significant difference in the enzymatic activity or stability was observed for the reactors prepared in these two methods.20 Although inspired in recent years, the above methods still have some disadvantages: the hydrophilicity of diol functionalities originating from the hydrolyzed poly (GMA-co-EDMA) monolith is not sufficient to avoid adsorption of hydrophobic albumin in a highly aqueous mobile phase that limits the application of the IMERs for digestion of high molecular protein. To solve this problem, Svec’s group recently demonstrated a novel approach to modification of the surface chemistry of the poly (GMA-co-EDMA) monolith.138 The monolith was first hydrophilized via photografting of poly(ethylene glycol) methacrylate followed by photografting of a 4-vinyl-2,2dimethylazlactone to provide the pore surface with reactive functionalities required for immobilization. This new approach reduced the undesired nonspecific adsorption of proteins and peptides and facilitated control of both the enzyme immobilization and the protein digestion processes. AMINOLYSIS FOLLOWED BY DIALDEHYDE ACTIVATION Another traditional and often used path facilitating immobilization of glycidyl methacrylate-based supports comprises aminolysis of the epoxide ring using ammonia or a diamine, followed by activation with dialdehyde most often glutaraldehyde.139 The aldehyde functionality is then used for the reaction with an enzyme, and similarly to the previous technique, the labile imine double bond must be hydrogenated. Poly(GMA-co-EDMA) monolith prepared using thermally initiated polymerization in a 50 mm  4 mm ID column was modified with 1,6-diaminohexane and glutaraldehyde.137 This reaction path was believed to afford a spacer arm on which papain could reside thus decreasing the steric constrains. Indeed, compared to the approach comprising hydrolysis followed by oxidation to aldehyde described above, the immobilized papain exhibited twofold higher effectiveness as measured by digestion of human IgG despite the lower amount of attached protein. Yet, the enzymatic activity of the immobilized papain was only 17% of that observed with free enzyme. As expected, the immobilized enzyme was

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

151

significantly more stable at a temperature of 75 C. Ye et al.18 developed a microreactor containing poly(GMA-co-EDMA) monolith prepared using thermally initiated polymerization in a capillary. Aminolysis of the epoxide functionalities with 29% aqueous ammonia solution afforded primary amine groups that were then activated with glutaraldehyde. Finally, trypsin was immobilized in this monolith and stabilized by reduction with sodium cyanoborohydride. The disadvantage of this immobilization reaction is that it has a potential for producing undesirable by-products, for example, homoconjugates and various polymers. HYDROLYSIS FOLLOWED BY CARBONYLDIIMIDAZOLE ACTIVATION Carbonyldiimidazole (CID) activation of hydroxyl groups containing supports for immobilization of proteins was first suggested by Hearn et al.140 at the end of the 1970s, in response to problems encountered at that time with widely used cyanobromide activation method such as the formation of undesired ionized functionalities leading to nonspecific interactions. Ben
cina et al.134 used this approach for immobilization of trypsin on monolithic disks and achieved excellent activity. Once the immobilization was complete, unreacted imidazole carbamate functionalities were quenched. The activated support contained 0.8 mmol mL
1 imidazole carbamate functionalities. This is said to be much less than the content of epoxide groups in the monolith. However, no direct comparison between these two values can be made since a number of epoxide functionalities are buried within the matrix. As the immobilization occurs only at the pore surface, inaccessible groups cannot contribute to binding. In contrast, activation with CID is likely to occur mostly on the pore surface where are located the most accessible hydroxyl groups. Vinylazlactone Copolymers Beads prepared by copolymerization of 1-vinyl-4, 4dimethylazlactone and methylenebisacrylamide using an inverse suspension process were developed by 3M Company at the beginning of the 1990s.141 Their properties and applications were summarized by Heilmann et al.142 in an excellent review. Due to their enhanced reactivity to amine and thiol groups, they also found applications in immobilization of various enzymes.143,144 An additional benefit of this chemistry is the linkage of the protein through a dipeptide spacer that can contribute to enhancement in activity. Svec and coworkers introduced this chemistry in the field of monoliths and used it for immobilization of trypsin.145 The initial monolith was prepared from a mixture of monomers that afforded reactivity (1-vinyl-4, 4-dimethylazlactone), hydrophilicity (acrylamide), and cross-linking (ethylene dimethacrylate) dissolved in porogenic solvent (tetradecanol). Azobisisobutyronitrile was used as the thermal initiator. The porogen enabled formation of monoliths with a pore size of about 2.5 mm. Trypsin was immobilized on this monolith located in a 20 mm  1 mm PEEK tube by pumping its solution through the device at a flow rate of 0.2 mL min
1 for 60 min. The unreacted azlactone functionalities were then quenched by reaction with 2-aminoethanol. This protocol led to a conjugate with 38.8 mg mL
1 or 90.7 mg g
1 of immobilized trypsin. They extended this azlactone chemistry to both capillary and microfluidic formats.146–148

152

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.13 Scheme of the preparation technique affording monolith with well-defined size and location. (a) Empty capillary (or microfluidic chip); (b) capillary filled with the polymerization mixture consisting of monomers, porogenic solvents, and a photoinitiator; (c) capillary with attached photomask; (d) irradiation of the capillary contents through the mask with UV light in the range of 220–330 nm for 10 min to fabricate the monolith; (e) removal of unreacted polymerization mixture from the dark parts and washing with a solvent. Reprinted from Ref. 149, with permission.

Since the ultimate goal is the fabrication of complex devices, the commonly used thermally or redox-initiated free radical polymerization modes are not best suited for the preparation of monoliths within a specific part of the capillary or a microchip. To solve this problem, Svec further developed the photolithographic-like technique involving photopolymerization through a mask.149 This simple approach shown schematically in Figure 4.13 facilitates the formation of reactive porous monoliths within a specific part of a microsystem. The as-prepared reactor afforded suitable degrees of digestion of proteins even after very short residence time of less than 1 min. 4.3.4 Membrane The incorporation of polymeric membranes into microfluidic networks has been employed for enhancement in device functionality for years.150 Due to the porous structures of membrane media, polymeric membranes exhibit a large surface to volume ratio that serves to facilitate rapid solution exchange. Extremely large surface area, at least 200 cm2 of internal surface per cm2 of frontal surface, is available for protein adsorption and immobilization. Thus, the membranes containing adsorbed proteins can be employed as miniaturized enzyme reactors.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

153

FIGURE 4.14 Schematic representation of the miniaturized trypsin membrane reactor (a) by fabricating the PDMS microfluidic channel and coupling it to a PVDF membrane; (b) by placing the PVDF membrane inside the capillary fitting. Reprinted from Refs 26 and 27, respectively, with permission.

Gao et al. reported a miniaturized membrane reactor by fabricating the PDMS microfluidic channel on the PDMS substrate and coupling it to a poly(vinylidenefluoride) (PVDF) membrane providing large internal surface area for enzyme adsorption (Figure 4.14a).26 Despite the large S/V ratio of porous membrane media, this reactor had a high total dead volume due to capillary connections with the microchannel. This problem was eliminated by placing a hydrophobic and porous PVDF membrane around the end of a polymer sleeve (Figure 4.14b).27 The assembly of capillary fitting, containing a length of fused silica capillary, provided the necessary flow paths and the membrane media for performing rapid and effective proteolytic digestion. This membrane-based proteolytic reactor can be directly coupled with nano-ESI-MS for achieving speedy protein identification in seconds instead of hours. In the case of membrane clogging, the membrane-based reactor can be regenerated by simply replacing the old membrane with a new PVDF strip, followed by adsorption of a fresh enzyme solution. Hisamoto et al. reported that a nylon membrane could be formed at the interface of two solutions formed in a microchannel (Figure 4.15a). Peroxidase was immobilized on this membrane, which was used as a chemicofunctional membrane;151 however, immobilization of the membrane is technically difficult, and application of this method is limited because the nylon membrane is unstable in organic solvents. Maeda and coworkers have developed a method for the preparation of an enzymeimmobilized microreactor by simple loading of the enzyme solution and a mixture of glutaraldehyde and paraformaldehyde into the microchannel forms a cross-linked enzyme aggregate membrane on the microchannel wall (Figure 4.15b).152 4.3.5 Other Formats Protein digestion can also be performed on a trypsin-immobilized MALDI probe with the advantage of obviating the need to handle samples before carrying out MALDITOF-MS (matrix-assisted laser desorption ionization time-of-flight mass spectrometry) measurements. As described by Nelson et al.,153,154 enzyme was covalently attached to the MALDI probe via a gold-coated stainless steel sample target. Proteins were digested on enzyme-linked probes by depositing the sample directly on the active

154

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.15 (a) Single and parallel dual nylon membranes; (b) cross-linked enzyme aggregate membrane prepared inside the microchannel. Reprinted from Refs 151 and 152, respectively, with permission.

surface. Digestion was terminated by adding the MALDI matrix prior to MS analysis. Picomoles of proteins could be efficiently digested in less than 30 min. Houston and Reilly155 used this technique for hemoglobin characterization. Recently, Lubman and coworkers156 developed a method combining capillary monolithic RP-HPLC with onplate enzymatic digestion for obtaining protein identifications for human breast cancer cells, which is a simple protocol with the advantage of effectively minimizing sample loss. However, because the matrix used to enhance ionization was applied directly to the probe, the enzyme-linked MALDI probe could not be reused for consecutive digestion. To solve this problem, Li et al. introduced trypsin-linked magnetic nanoparticles into the on-probe digestions (Figure 4.16).25 Due to their magnetic property, the trypsin-linked nanoparticles could be easily removed from the probe after digestion, which would benefit sample–matrix cocrystallization and avoid causing possible contamination on the ion source chamber in MS. What is more important is its feasibility for reuse of MALDI probe for consecutive digestion. More recently, Ota et al. developed one kind of elegant trypsin immobilized monolithic silica with pipette-tip formula for high-throughput protein digestion.157 The silica-based monolith was first chemically modified by 3-aminopropyltrimethoxysilane, and then fixed into a 200 mL pipette tip by supersonic adhesion. After the carriers were activated by disuccinimidyl suberate (DSS), trypsin was finally immobilized. The tip enabled the digestion of reduced and alkylated protein within 20 times operation, and the enzymatic activity of the immobilized trypsin tip was about 50 times higher than that of the conventional in-solution format.

FABRICATION METHODS OF IMMOBILIZED MICROFLUIDIC ENZYMATIC

155

FIGURE 4.16 (a) Trypsin-linked magnetic nanospheres were added to the protein solution. (b) Nanospheres could be easily removed from the plate with a magnetized needle. (c) After the removal of the nanospheres, the plate is ready for MALDI-MS analysis. Reprinted from Ref. 25, with permission.

Except for the above formats, Fan and Chen developed a fiber-packed channel bioreactor for protein digestion by immobilizing trypsin on the fiberglass bundles embedded in the substrate of a PMMA microchip (Figure 4.17a).158 A UV-sensitive prepolymerized methyl methacrylate (MMA) molding solution containing a UV initiator was sandwiched between a PMMA cover plate and a PMMA base plate bearing glycerol-permeated fiberglass bundles and was exposed to UV light. During polymerization, the fiberglass bundles were embedded in the PMMA substrate to form fiberglass-packed microchannels. When the glycerol in the fiberglass bundles sealed inside the PMMA substrate was flushed away with water, the obtained porous fiberglass-packed microchannels could be employed as a support to immobilize trypsin with the aid of chitosan (CTS) and glutaraldehyde. However, when the enzyme activity decreased to some extent, the trypsin-immobilized layers were permanently modified in the channel. Therefore, they further fabricated a core-changeable needle enzymatic reactor by inserting a piece of trypsin-immobilized glass fiber into the needle of a syringe (as shown in Figure 4.17b),159 which could be regenerated by changing the core composed of a piece of glass fiber and a layer of enzyme-entrapped polymer coating. The in-needle fiber bioreactor has been coupled with MALDI-TOF MS for the digestion and peptide mapping of model proteins.

FIGURE 4.17 (a) SEM image of the cross section of a fiberglass-packed microchannel in the PMMA substrate.158 (b) Schematic diagrams of the core-changeable needle bioreactor.159 Reprinted from Refs 158 and 159, respectively, with permission.

156

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

4.4 APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS The fields of IMER application are becoming wider every year. A considerable number of papers have been published reporting successful application of enzymatic microreactors in chemistry and biochemistry. The principal field of application of microreactors is tryptic digestion of proteins. Enzymatic microreactors also facilitate characterization of enzyme activity as a function of substrate concentration, and enable fast screening of new biocatalysts and their substrates. They may constitute key parts of lab-on-a-chip and mTAS, assisting the analysis of biomolecules.160 Intentionally, in order to narrow the scope of this chapter, only applications in peptide mapping, biosensing, and kinetic study are discussed in this section. Readers interested in other application areas are advised to see comprehensive reviews12,160–163 and recent publications. 4.4.1 Peptide Mapping The ability to rapidly and efficiently digest and identify an unknown protein is of great utility for proteome studies. Identification of proteins via peptide mapping is generally accomplished through proteolytic digestion with enzymes such as trypsin. Limitations of this approach consist in manual sample manipulation steps and extended reaction times for proteolytic digestion. The use of immobilized trypsin for cleavage of proteins is advantageous in comparison with application of its soluble form. Thus, the greatest number of recent applications of IMERs refer to protein analysis by peptide mapping, and for determination of post-translational modifications (PTMs) such as phosphorylation, glycosylation, and lipidylation, which are essential in modulating biological functions of cells and can be associated with a number of diseases. Peptide mapping is typically performed using enzymatic cleavage of the protein and the peptide fragments in the resulting mixture are identified using electrospray ionization mass spectrometry (ESI-MS) or matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). In either case, separation of the peptide mixture, for example, by micro-high-performance liquid chromatography (m-HPLC) or capillary electrophoresis, prior to mass spectrometric analysis, minimizes the ionization suppression and improves the sequence coverage.164 One of the limiting steps in peptide mapping is the manual sample manipulation and extended reaction times for proteolytic digestion. Traditionally, enzymatic cleavage is performed in a homogeneous solution consisting of a mixture of the proteolytic enzyme and the protein. K
renkova and Foret summarized the most common enzymes used for protein digestion and their sites of cleavage in an excellent review (Table 4.1).2 Among these enzymes, the most frequently used is trypsin, which catalyzes the process of protein digestion through hydrolysis of peptide bonds at the C-end of the Arg and Lys residues and typically provides peptides in a mass range suitable for high-resolution/high-sensitivity mass mapping through mass spectrometry. A huge body of literature has been reported with trypsin-immobilized IMER for peptide mapping. Pepsin, an enzyme that

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

TABLE 4.1

157

Some of the Most Common Enzymes Used for Protein Digestion

Enzymes

IUBMB Enzyme Nomenclaturea

Trypsin Chymotrypsin

EC 3.4.21.4 EC 3.4.21.1

Endoproteinase LysC Endoproteinase AspN Endoproteinase GluC (S. aureus V8, pH 4) Endoproteinase GluC (S. aureus V8, pH 8) Endoproteinase ArgC (clostripain) Thermolysin

EC 3.4.21.50 EC 3.4.24.33 EC 3.4.21.19

C-terminus of Arg and Lys C-terminus of Phe, Tyr, Trp, Leu, and Met C-terminus of Lys N-terminus of Asp C-terminus of Glu

EC 3.4.22.19

C-terminus of Glu and Asp

EC 3.4.21.8

C-terminus of Arg

EC 3.4.24.27

Pepsin

EC 3.4.23.1

N-terminus of Leu, Ile, Val, Phe, Met, and Ala C-terminus of Phe, met, Leu, and Trp

Site of Cleavage

Reprinted from Ref. 2, with permission. a

IUBMB, International Union of Biochemistry and Molecular Biology.

cleaves proteins at pH 2 at C-end of Phe and Leu residues, can be seen as a complementary tool that may be used to both determine and confirm the respective order of each peptide fragment obtained from the tryptic digestion. Since pepsin digests proteins at acidic pH, it facilitates the direct coupling of the pepsin-immobilized reactor with ESI-MS. The specificity of pepsin is lower than that of trypsin, and some amino acid sequences are not cleaved despite the presence of Phe and Leu, while other sequences can be cleaved even in their absence. Although this would be detrimental for protein identification, this lack of specificity can be utilized for quantitative protein analysis.165,166 Due to miscleavages, pepsin digestion generates longer peptides that provide a signature of a specific protein. Thus, pepsin digestion has been successfully applied to quantitative protein analysis affording the reproducible formation of specific peptide markers.166 Endoproteinases such as GluC, ArgC, and LysC have been gaining more popularity due to their highly specific points of cleavages in recent years. They offer additional benefits such as digestion resulting in a smaller number of larger, information-rich peptide fragments that simplifies their separation and facilitates protein identification. These enzymes are also useful in mapping of post-translational modifications such as methionine oxidation since the large peptides allow lower quantization limits to be achieved using typical UV detection. The larger peptides formed during LysC digestion are also more likely to contain multiple sites allowing charging such as internal arginines, which form ions better suited for MS/MS analysis. Endoproteinase LysC is also a more robust enzyme that, compared to trypsin, maintains its activity even at relatively high concentrations of denaturants.138 To achieve efficient and reproducible digestion results, maintaining optimum pH, temperature, protein to enzyme ratio, and reaction time is critical. The variations in

158

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

size, structure, type, and level of PTM make proteins significantly different in their susceptibility to enzymatic digestion. Depending on the accessibility of the cleavage sites, complete digestion may require times ranging from several minutes to overnight. In principle, the time of digestion can be reduced using high concentration of the free enzyme;167 however, such an approach has several disadvantages. Besides the added cost, the enzymes often lose their activity and specificity, and the enzyme autodigestion results in undesirable formation of additional peptides, which may lead to the ionization suppression in MS analysis and complicate the interpretation of data. Immobilizing the enzyme on a solid support eliminates unwanted autodigestion and an extremely high local concentration of proteolytic enzyme providing rapid catalytic turnover. Moreover, due to the relatively long incubation time the current workflow of protein analysis includes protein digestion as an off-line step. With miniaturized IMER, complete protein digestion can be performed in less than 1 min and direct online coupling with further peptide concentration and separation steps become possible. For high-throughput proteome analysis, it is important that separation and detection are coupled online to proteolysis. The online digestion of proteins with IMER can enable faster and more automated protein identification. Moreover, the use of IMER avoids manual sample handling and the consequent possible contamination of the sample, and the reactor can be coupled online to separation and detection with the advantage of automation. Several possibilities can be envisioned for coupling IMER with a separation and identification system and they are outlined in Figure 4.18.161 4.4.1.1 IMER Coupled with MS Enzymes can be bound to the inner walls of the capillary (see Section 4.3.1). In this case, there are no backpressure constraints during sample injection. Such microreactors can be easily attached to an electrospray ionization (ESI) interface with mass spectrometry (MS), providing an online system.147,168 A typical experimental setup of IMER hyphenation with ESI-MS is depicted in Figure 4.19.20 In this online system developed by Foret’s group, the protein solution was pumped through the IMER at selected flow rates (50–300 nL min
1), and the digestion products were mixed in a Tjoint with a constant flow rate (1.2 mL min
1) of the spray solution (50% aqueous ACN, 1% formic acid) supplied by the second syringe pump. The resulting stream was analyzed with the microspray interface supplied with the MS instrument. Recently, the same group prepared IMERs with L-1-tosylamido-2-phenylethyl chloromethyl ketone (TPCK)-trypsin and pepsin A covalently immobilized on the wall of a 10 mm ID fused

FIGURE 4.18 Modes of combining enzyme reactors with separation and identification systems. Reprinted from Ref. 161, with permission.

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

159

FIGURE 4.19 Schematic diagram of IMER hyphenated online with ESI-MS. Reprinted from Ref. 20, with permission.

silica capillary. The optimized reactors were served as the nanospray needle in a liquid junction interface for CE-ESI/TOF-MS analysis of protein mixtures. On-line digestion of proteins on the capillary wall enabled faster and fully automated protein identification using peptide mass fingerprinting.70 Enzymes can also be immobilized on a variety of nanoparticles and packed into the microcolumn (see Section 4.3.2) and the product fractions may be sampled prior to offline analysis by means of MALDI-TOF-MS.30,32,33,13 The primary advantage of the MALDI approach compared to ESI, when coupling to microfluidic chips, is the potential for multiplexing. The directional control of the ESI spray can be difficult with a high-density array of spray tips. Furthermore, microfluidic chip interfaces with ESI can suffer from stability problems when sprays are started or stopped. This limits the speed of moving from one sample to another and therefore limits the throughput. In most cases, digested peptides eluted from the IMERs were collected and deposited on the MALDI probe followed by deposition of MALDI matrix. Alternatively, Lee et al. developed an off-line MALDI interface to combine matrix addition and deposition on a MALDI target using a robotic plate spotter modified to accept the effluent from the microfluidic chip automatically.69 Although less developed, online coupling with MALDI-MS can also be achieved. For example, Ekstr€ om et al.71 described a device that integrated the m-chip IMER with a sample pretreatment robot and a microfabricated microdispenser to transfer digested protein directly to a MALDI target plate for automated MS analysis (Figure 4.20). Anodic etching in a hydrogen fluoride/ethanol solution was used to produce a porous surface on the digestion chip. The use of porous silicon provided a 170-fold increase in enzymatic activity compared to nonporous reactor.169,170 This increase in surface area resulted in increased digestion efficiency and extremely fast digestions. The m-chip IMER allowed online enzymatic digestion of protein samples (1 mL) within 1–3 min,

160

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

FIGURE 4.20 Microfluidic system for MALDI protein analysis. (a) Automated sample pretreatment and injection; (b) m-chip IMER (the photo inset shows a SEM picture of the lamella structure with the porous layer); (c) microdispenser used to deposit sample into mvials; (d) shallow nanovials (300 mm  300 mm  20 mm) on the MALDI target plate; and (e) automated MALDI-TOF-MS analysis. Reprinted from Ref. 71, with permission.

about 200–1000 faster than digestion in solution. This integrated system provided a throughput of 100 samples in 3.5 h. The most common strategy for protein analysis is based on two-dimensional polyacrylamide gel electrophoresis (2D PAGE). Typically, the identification of a protein mixture involves separation, extraction, proteolytic digestion, and MS analysis of each protein spot. New technology developed by Cooper and Lee171 involved an online combination of electrophoretic protein transfer from a polyacrylamide gel with proteolytic digestion in membrane-based IMER. After electrokinetic-based protein extraction and stacking, real-time proteolytic cleavage of extracted protein, and direct deposition of protein digest onto the MALDI target, the peptides were identified by MS analysis (Figure 4.21). The sensitivity of the technology was demonstrated both by the detection of standard proteins from a gel protein loading as low as 1 ng and by the identification of low-abundance proteins in complex yeast cell lysates. 4.4.1.2 IMER Coupled with Liquid Chromatography System Liquid chromatography in combination with tandem mass spectrometry (LC–MS/ MS) can overcome many of the limitations of 2D gel electrophoresis in proteomic studies. The advantage of LC separation systems is that the automated operation is much easier to perform. Therefore, it is not surprising that a lot of publications deal with the development of IMERs compatible with liquid chromatographic systems. In most cases, IMERs are coupled with LC systems with several valves. Calleri et al. reported the immobilization of trypsin on silica-based monoliths via epoxy groups. With such an immobilized enzyme reactor, they achieved the hyphenation of online digestion with HPLC via a switching valve. It was found that the cleavage efficiency (aminoacidic recovery, %AA) achieved in 20 min by the online protocol was at least

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

161

FIGURE 4.21 Scheme of combined protein electronic transfer and miniaturized trypsin membrane digestion for gel protein identification using MALDI-MS. (a) Electronic protein transfer followed by (b) introduction of extracted proteins into a membrane reactor and deposition of protein digest on a MALDI target. Reprinted from Ref. 171, with permission.

comparable, or even better than the conventional off-line 4 h consuming method of digestion. By using the online system, protein digestion and genetic variant identification in serum samples were performed by the same group, and mutation sites in betalactoglobulin A and B variants were successfully located.172 Markides’s group173 integrated IMER with a trapping column so that the peptides produced from the enzymatic capillary reactor could be efficiently trapped and desalted that helped to avoid detrimental signal suppression in the following ESI process. The peptides eluted from the precolumn were then separated on an analytical capillary column by a buffer suitable for the ESI-MS process (Figure 4.22a). Different trapping columns were tested in order to avoid losses of hydrophilic peptides during the elution from the trypsin reactor on the separation column. The performance of the online system was compared to that of the classical digestion in solution, with reference to peptide

FIGURE 4.22 Schematic diagrams of IMERs coupled with LC. (a) Setup of online proteolysis, peptide trapping/desalting, separation, and ESI-MS analysis. (b) Setup of online system for peptide mapping of post-translational modified proteins. Reprinted from Refs 173 and 175, respectively, with permission.

162

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

sequence coverage and sensitivity. The peptide sequence coverage was increased from about 25% to 55%. Hsieh et al.174 developed a fully automated five-column chromatography system coupled to an ESI-MS/MS for the isolation, digestion, and characterization of human hemoglobin. The employment of the trypsin reactor and the application of a multivalved apparatus allowed automated, direct transfer of analytes between the various operation units. For analysis of post-translational modified proteins, Riggs175 described an automated multidimensional chromatographic system consisted of different steps: (i) reduction and alkylation of the proteins were achieved in the autosampler; (ii) proteolytic digestion was carried out on trypsin column; (iii) specific classes of peptides were selected by affinity chromatographic column; (iv) the selected peptides were transferred to a reversed-phase chromatographic column and further fractionated. The analytes were transferred between columns through valves (Figure 4.22b). The chromatographic effluent was analyzed by ESI-MS for the signature peptide approach. The model system chosen for this study was phosphorylated milk proteins, and the total analysis time in the tandem column mode of operation was under 2 h. Geiser et al. have developed an online system containing capillaries with two different porous polymer monoliths for protein digestion with immobilized pepsin, peptide preconcentration, and nanoliquid chromatography separation coupled to electrospray ionization mass spectroscopy (nLC–ESI-MS).148 The first monolith with well-defined porous properties was prepared by in situ copolymerization of 2-vinyl-4, 4-dimethylazlactone and ethylene dimethacrylate (poly(GMA-coEDMA)) and used as support for covalent immobilization of pepsin. The second, poly(laurylmethacrylate-co-ethylene dimethacrylate) (poly(LMA-co-EDMA)) monolith with a different porous structure served for the preconcentration of peptides from the digest and their separation in reversed-phase liquid chromatography mode (Figure 4.23a). With no need for adding a trapping column to the system, the top of the separation capillary was also served as a preconcentrator, thus enabling the digestion of very dilute solutions of proteins in the bioreactor and increasing the sensitivity of the mass spectrometric detection of the peptides using a time-of-flight mass spectrometer with electrospray ionization. The schematic diagram of the online system is provided in Figure 4.23b. Myoglobin, albumin, and hemoglobin were digested to demonstrate feasibility of the concept of using the two monoliths online. Successive protein injections confirmed both the repeatability of the results and the ability to reuse the bioreactor for at least 20 digestions. More recently, on the basis of the above work, they further modified the surface of the first monolithic poly(GMA-co-EDMA) via multistep/multilayer photografting to obtain support for enzyme immobilization with both reactive azlactone functionalities as well as largely eliminated nonspecific adsorption of proteins and peptides, and extended the application of this online system to analysis of high molecular weight human IgG.138 4.4.1.3 IMER Coupled with Capillary Electrophoresis System Capillary electrophoresis (CE) has become an important separation technique for peptide mapping because of its simplicity, speed, high separation efficiency, and low

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

163

FIGURE 4.23 (a) SEM micrographs of the cross section of 100 mm ID monoliths in capillaries: (A) support for the enzymatic reactor; (B) HPLC column prepared. (b) Scheme of the online system: (A) protein solution is injected into the immobilized pepsin reactor in 2% aqueous formic acid, digested, and peptides are trapped on the top of HPLC column; (B) the reactor is bypassed and peptides are separated in a gradient of acetonitrile. Reprinted from Ref. 148, with permission.

sample consumption. Different research groups have explored the possibility of performing trypsin hydrolysis online with an electrophoretic separation. However, online digestion of proteins before CE separation is difficult to accomplish due to the challenges of manufacturing a bioreactor compatible with the small sample volumes used in CE and because of the problems encountered in interfacing the reactor with the separation capillary. The two most common configurations are biocatalytic capillary located upstream of the separation capillary and trypsin immobilized in the first part of the capillary devoted to separation. IMER Coupled with CE Capillary In an on-capillary protein hydrolysis for peptide mapping, a stop-flow incubation period within the capillary is required to allow numerous hydrolysis products to accumulate before accomplishing their electrophoretic separation. The resolving power of the separation suffers when diffusional broadening of the reaction mixture zone becomes intolerable after lengthy incubation

164

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

periods, and increasing dilution of the reaction zone with incubation time results in a concomitant reduction in the sensitivity of the on-capillary detection. Traditionally, relatively slow reaction kinetics for protein hydrolysis generally require long incubation times, which can limit the usefulness of on-capillary proteolytic reactions. Another critical consideration is the suitability of the separation buffer to the enzyme-mediated chemistry. One must also have the capability to mix the enzyme with the chosen substrate(s), maximize conversion of substrate to products, and accomplish a separation of the resulting reaction mixture. These processes are generally optimal under different solution conditions (pH, ionic strength, buffer additives) in the capillary. The success of on-column peptide mapping is particularly vulnerable when the optimal pH of the proteolysis reaction differs from the pH at which optimal selectivity is obtained for the separation of the peptide. Kuhr’s group demonstrated that trypsin can be readily immobilized on the surface of a fused silica capillary via biotin–avidin–biotin technology42 and can be coupled to the separation capillary to enable online digestion and separation through an open fluid junction (Figure 4.24a).8 The enzyme-modified fused silica microreactor was coupled through a 100 mm solution gap to the separation capillary. Very little diffusional sample loss in the gap

FIGURE 4.24 Schematic presentation of online coupling of enzymatic reactor (a) coupled with CE capillary via an open fluid junction8; (b) coupled with an SPE preconcentrator and CE177; (c) coupled with transient isotachophoresis (CITP)/CZE-ESI-MS26; (d) CE-enzymatic microreactor-CE-MS/MS.178 Reprinted from Refs 8,26,177, and 178, respectively, with permission.

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

165

was observed at low ionic strength since the voltage gradient across the gap is large and the sample had little time for radial diffusion out of the junction. The utility of this approach was that it not only offered an efficient mass transfer of the sample by electromigration but also allowed subsequent separation of the transferred sample in the second capillary without the need to move either capillary. This configuration allowed independent optimization of digestion and separation conditions in an online peptide mapping procedure. The enzyme-modified capillary was filled with the protein solution and allowed to incubate at room temperature for approximately 2 h, during which proteolysis occurs. Subsequently, an aliquot of the digest was injected into the separation capillary by applying a potential across the two free ends of both capillaries. The injected sample was then separated by applying the CZE separation potential across the gap solution and the other end of the separation capillary. Using this approach, they were able to perform online digestion and separation of picomole quantities of protein by CZE in less than 3 h. The same coupled capillary CZE instrument was used for further experiments and three microreactors were prepared with two proteases (trypsin and pepsin) and a peptidase, carboxypeptidase-Y.64 These proteolytic enzymes are distinguished by the large differences in their specificities toward peptide bond cleavages and their pH of greatest catalytic efficiency. Offline protein digestions in carboxypeptidase-Y and trypsin-modified microreactors yielded C-terminal sequences and mass fingerprints, by either MALDI-MS or plasma desorption mass spectrometry, which were compared with protein fragment databases for identification. On-line trypsin digestion followed by CE/ESI-MS was also demonstrated to give the greatest efficiency in peptide mapping analysis.65 Efficient digestion of the oxidized insulin b-chain occurred within 40 min, and the entire peptide mapping required about 1 h. The long digestion time required in this system can be due to the fact that enzymes were directly immobilized on the capillary inner wall with limited surface area that leads to a low surface to volume ratio of the microreactor. Bonneil et al.176 fabricated a microreactor with commercially available enzymeimmobilized controlled pore glass beads packed in a capillary for peptide mapping by capillary electrophoresis. The proteins were perfused through the microreactor for about 2 h applying low pressure at the inlet. The digest was collected at the microreactor outlet and the tryptic fragments were separated by CE and detected by UV absorbance using a diode array detector. Subsequently, the microreactor was further coupled online with an SPE preconcentrator and CE in an effort to improve mapping sensitivity by minimizing sample handling that leads to peptide losses.177 The proposed system depicted in (Figure 4.24b) allowed digestion of proteins and preconcentration, separation, and detection of the peptides in 4 h. However, the separation efficiency was poor due to the multiple-valve design of the system and to the dispersion of the 60 nL desorption plug. Nevertheless, the maps were fairly reproducible in terms of migration time. On-line coupling of a miniaturized membrane reactor for proteolytic digestion with transient isotachophoresis (CITP)/CZE-ESI-MS was proposed by Gao et al.26 Trypsin was adsorbed on a PVDF membrane. As reported in the schematic setup of Figure 4.24c, proteins were forced by a syringe pump through pores into this

166

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

membrane, and the digested proteins were focused by transient isotachophoresis, separated by CE, and detected by ESI. The microfluidic system enabled rapid identification of proteins in minutes and consumed very little sample (nanogram). The rather large dead volume of this device resulted in poor separation efficiency for the digested peptides; theoretical plate counts were about 1500 for the peptides resulting from digestion of cytochrome c. However, the abilities of the platform to concentrate and resolve peptide mixtures enhanced the sensitivity and the dynamic range of ESI-MS detection for the identification of minor proteins. In order to reduce the dead volume, analysis time, and sample consumption, the same research group employed a commonly used capillary fitting for directly housing a miniaturized trypsin membrane reactor.27 The nanoscale trypsin reactor was integrated into a platform for the concentration and separation of the proteolytic digest using chromatographic and electrophoretic methods prior to mass spectrometry analysis. The application of sample stacking and separation techniques contributed to further enhancement in the dynamic range and detection sensitivity for the analysis of complex protein mixtures. The proposed nanoscale reaction system enabled rapid proteolytic digestion in seconds instead of hours for a protein concentration of less than 10
8 M. Dovichi and coworkers have described the production of a macroporous monolith with immobilized trypsin that was coupled to capillary electrophoresis for peptide mapping, which produced over 300,000 theoretical plates for peptide separation.18 That system employed postcolumn labeling with fluorescence detection but did not separate proteins before digestion. Recently, they further developed a fully automated bottom-up approach to protein characterization.178 Proteins were first separated by capillary electrophoresis. A pepsin microreactor was incorporated into the distal end of this capillary. The reason for choosing pepsin rather than trypsin as the digestion enzyme was that pepsin allows to use a volatile acetic acid–ammonium acetate buffer that is beneficial for combination with ESI-MS. Peptides formed in the reactor were then transferred to a second capillary, where they were separated by capillary electrophoresis and characterized by mass spectrometry. While peptides generated from one digestion were being separated in the second capillary, the next protein fraction underwent digestion in the microreactor (Figure 4.24d). Enzyme Immobilized in the First Part of the CE Capillary Construction of an IMER coupled with CE capillary is a quite complex operation; its reproducibility in fabrication is low, and ensuring electrical and fluidic connections is not trivial either. As an alternative, the enzyme can be directly immobilized on a portion of a capillary column; the first part containing the immobilized enzyme then acts as a microreactor and the second part of the column is devoted to the separation of the peptides. This approach can in principle reduce systematic errors associated with moving the sample from the microreactor into the separation zone. One of the most serious issues that restricted the application of this approach is the difficulty in controlling the position and the size of the patch of immobilized enzymes. Another critical consideration is that since the enzymatic reaction and the separation were performed in the same capillary column, the buffers are exactly the same for both procedures. It is known that the

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

167

FIGURE 4.25 Schematic illustration of an online enzyme reactor integrated into CE. Reprinted from Ref. 47, with permission.

optimal pH of the enzymatic reaction probably differs from the pH at which optimal selectivity is obtained for the separation of the peptide. Therefore, the buffer conditions should be carefully selected considering from both standpoints. Trypsin was encapsulated in tetramethoxysilane-based hydrogel in a single step under mild conditions within a capillary, and a 1.5-cm length of gel was formed at the inlet of the capillary.47 The average amount of trypsin in the capillary was about 0.90 mg cm
1 gel. The resultant monolithic reactor showed enzymatic activity approximately 700 times higher than that in free solution, without stopping the flow. The substrates were introduced electrokinetically from the inlet of the capillary and then cleaved into products while they flowed through the trypsin-encapsulated gel by electrophoresis and EOF. Unreacted substrates and products were separated by electrophoresis (Figure 4.25). The buffer condition was considered for both enzymatic reaction and CE separation and 50 mM Tris–HCl (pH 7.5) was finally selected. The system was used successfully for hydrolyzing a-N-benzoyl-L-arginine ethyl ester (BAEE) and separating its product without stopping the flow. However, the method was tested only on peptides, as proteins larger than trypsin could not pass through the hydrogel and were not digested, severely limiting the application of this system. Enzyme can also be immobilized on a portion of silica capillary through a photocoupling reaction. Bossi et al. constructed a CE-microreactor for peptide mapping using this technology.179 The bioreactor was characterized by being a single piece, thus ensuring no fluidic or electrical leakage typical of the reactors constructed as multiassemblies. The immobilization procedure was optimized, and the activity and stability of the reactor were tested with proteins of different dimensions (cytochrome c, hemoglobin, and carbonic anhydrase). Mapping online in the CE-microreactor was quite competitive in terms of time (completed map within 15 min) and exhaustive for

168

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

the mapping of small proteins. However, the CE-microreactor was not coupled to MS for complete identification of the peptides mapped. 4.4.2 Immobilized Enzyme Biosensors Immobilized enzymes have been used as component of detection device in various analytical and biomedical applications. Some of these devices have been associated with miniaturized electrochemical detectors and fluidic systems. Since numerous oxidase and dehydrogenase enzymes generate oxidizable products (hydrogen peroxide and NADH, respectively), enzyme-based biochip assays were integrated with various amperometric detectors by Wang et al.180–182 Such electrochemical detectors offer additional advantages for CE microchips, including compatibility with micromachining technologies, miniaturization of both the detector and control instrumentation, and high sensitivity and selectivity. On-line precolumn and on-column reactions of glucose oxidase and alcohol dehydrogenase have been employed both for selective measurements of glucose (in the presence of ascorbic acid and uric acid)180 and for the simultaneous measurements of glucose and ethanol (in connection with electrophoretic separation of the peroxide and NADH products).181 On-chip assay of amino acids based on their electrophoretic separation, postcolumn reaction with amino-acid oxidase and amperometric detection of the hydrogen peroxide product was also developed by the same group.182 L’Hostis et al.183 tested the enzymatic microreactor for glucose detection via glucose oxidase immobilized to glass beads. Glucose detection was carried out by electrochemical measurements of hydrogen peroxide generated enzymatically with a platinum electrode. Glucose was also measured with microfluidic biosensors based on immobilizing glucose oxidase in poly(dimethylsiloxane) electrophoretic microchips.93 An immobilized enzymatic fluorescence capillary biosensor was developed for determination of sulfated bile acid in urine.184 A microelectromechanical systems (MEMS) device consisting of two identical freestanding polymer diaphragms, resistive heaters, and a thermopile between the diaphragms was fabricated.185 Enzymes specific to a metabolic analyte system were immobilized on microbeads packed in the chambers. When a sample solution containing the analyte was introduced to the device, the heat released from the enzymatic reactions of the analyte was detected by the thermopile. The device had been tested with glucose solutions at physiologically relevant concentrations and shown its potentiality for continuous monitoring of glucose and other metabolites. De Boer et al. designed and implemented a continuous-flow microfluidic assay for screening (complex) mixtures for bioactive compounds.186 The microfluidic chip featured two microreactors (1.6 and 2.4 mL) in which an enzyme inhibition and a substrate conversion reaction were performed, respectively. Enzyme inhibition was detected by continuously monitoring the products formed in the enzyme–substrate reaction by electrospray ionization mass spectrometry. In order to enable the screening of mixtures of compounds, the chip-based assay was coupled online to capillary reversed-phase high-performance liquid chromatography with the HPLC column being operated either in isocratic or in gradient elution mode. In order to improve the

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

169

FIGURE 4.26 (a) Schematic overview of the online continuous-flow system: 1, pump and autosampler; 2, trapping column; 3, switching valve; 4, analytical column; 5, syringe pumps; 6, microfluidic chip; 7, mass spectrometer. (b) The microfluidic chip as used for bioactivity screening: 1, substrate solution; 2, LC effluent; 3, enzyme solution; 4, open tubular microreactor with a volume of 1.6 mL; 5, open tubular microreactor with a volume of 2.4 mL; 6, flow toward mass spectrometer. Reprinted from Ref. 186, with permission.

detection limits of the current method, sample preconcentration based on a micro online solid-phase extraction column was employed (Figure 4.26). IMERs are also utilized for screening inhibitors as potential drugs.187,188 Brennan and coworkers recently prepared protein-doped monolithic silica columns for immobilized enzyme reactors, which allowed the screening of enzyme inhibitors with MS as the detector.189 Kang and coworkers created an immobilized capillary acetylcholinesterase (AChE) reactor based on a layer-by-layer assembly for inhibitor screening. A 0.5 cm long plug of solution of the cationic polyelectrolyte polydiallyldimethylammonium was injected into the capillary to produce a positively charged coating on the surface of the capillary; subsequently, the enzyme solution with the same plug length was injected into the capillary and incubated for 10 min to immobilize the enzyme on the capillary wall via electrostatic interaction; third, PDDA solution with the same plug length was injected again into the capillary to cover the immobilized enzyme by forming a PDDA–AChE–PDDA sandwich-like structure. The substrate solution was injected and incubated for a short time, followed by applying a voltage to separate the product from the unreacted substrate

170

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

(Figure 4.27). Screening a small compound library containing 4 known AChE inhibitors and 42 natural extracts was demonstrated, and species with inhibition activity can be straightforwardly identified with the system.188 Two cytochrome P450 (CYP)-based immobilized enzyme reactors were developed to perform automated online phase I drug metabolism studies.190 For this purpose, biotinylated recombinant CYP2D6 or CYP3A4 reconstituted systems were anchored to the surface of two monolithic minicolumns (2 mm  6 mm ID), which had been covalently grafted with NeutrAvidin. After optimization of immobilization conditions, the obtained IMERs were integrated online into a LC hyphenated to an

FIGURE 4.27 Schematic representation of the immobilized capillary enzyme reactor with CE separation for inhibitor screening. Reprinted from Ref. 188, with permission.

APPLICATION OF IMMOBILIZED MICROFLUIDIC ENZYMATIC REACTORS

171

electrospray ionization MS/MS system. Studies with probe substrates and a known competitive inhibitor were performed, showing the potential of CYP-based IMERs in drug metabolism. 4.4.3 Kinetic Studies Microreactors offer significant advantages for online monitoring of biocatalysis and characterization of kinetics of supported enzymes. First, application of IMER offers a great advantage by shortening the analysis time. In many cases, an enzymatic reaction is very fast and can reach equilibrium within a single passage of substrate stream through the microreaction channel. However, several biotransformations, for example, those catalyzed by lipases, are slower. Scaling down the dimensions of the microreactor, and immobilizing the enzyme (lipase) inside a fused silica capillary, leads to very short times for the hydrolysis,191 while in batch reactions, completion of enzyme-catalyzed transesterification may take days for some supported lipases.192 Second, application of IMER greatly decreases the amount of biocatalyst used. In comparison with standard assays, the amount of enzyme used was very small: Seong et al. estimated that only 200 pmol (3  109 molecules of enzyme) was required for the analysis.193 Moore et al. presented an assay for 500 lipase molecules that could be applied to single cells.194 The kinetics model described by Lilly et al.195 is appropriate for systems with continuous flow of the substrate and under steady-state conditions, and can be summarized by the following equation: PS0 ¼ K 0 m lnð1
PÞ þ C=Q where P is the fraction of substrate reacted in the column, S0 is the substrate concentration at the beginning, K 0 m denotes the apparent Michaelis constant, C is the reaction capacity of the reactor, and Q is the flow rate of the substrate. This formula allows determination of the apparent Michaelis constant of the catalytic process when all other parameters are known. The Michaelis constant is typically measured with a series of experiments at different substrate concentrations in a well-mixed container. Seong et al. showed that the Michaelis constant determined with a microfluidic device with immobilized horseradish peroxidase was similar to the value obtained during homogeneous catalysis in batch mode.193 If any mass transfer effects contribute to the dynamics, an extrapolation to zero flow rate is required to obtain the value of the Michaelis constant for comparison with that of free enzyme.193 An interesting method for determining Km and nmax was presented by Jiang et al., who applied online frontal analysis of peptides originating from the digestion by trypsin immobilized on glycidyl methacrylate-modified cellulose.196 The Lineweaver–Burke diagrams were easily constructed based on the effects of injection of different concentrations and variation of flow rate of the substrate solution. Ristenpart et al. demonstrated a microfluidic technique for measuring Michaelis–Menten rate constants with only a single experiment.197 Enzyme and substrate were brought together in a coflow microfluidic device, and they established analytically and numerically that the initial concentration of

172

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

product scales with the distance x along the channel as x5/2. Measurements of the initial rate of product formation, combined with the quasi-steady rate of product formation further downstream, yielded the rate constants. They corroborated the x5/2 scaling result experimentally using the bioluminescent reaction between ATP and luciferase/luciferin as a model system. Mass transfer is always an important issue when considering enzymes entrapped in supports, and the ideal situation is when diffusion of substrate and product into and out of the bulk solution is not the rate limiting process. Koh and Pishko determined Michaelis constants of enzymes entrapped in hydrogel micropatches in microfluidic channels using Lineweaver–Burke graphs.198 Values were found to be lower, by approximately an order of magnitude, than those obtained from experiments using the homogeneous enzymes. The influence of entrapment in the hydrogel nanostructure on the kinetic properties of the enzymes was discussed. 4.5 SUMMARY AND FUTURE PERSPECTIVE In this chapter, we have focused on the devolvement of immobilized microfluidic enzymatic reactors using nanoparticles. Avariety of methods are now available for the immobilization of enzymes on nanoparticles. Immobilization techniques, such as physical adsorption, covalent binding, and copolymerization of enzyme with the polymers are discussed in this chapter. The specific immobilization chemistry depends on a variety of factors, characters of the supports, activation methods, and coupling procedure. The ideal supports for immobilization process should have the following characteristics: (i) large surface area, (ii) permeability, (iii) hydrophilic character, (iv) insolubility, (v) chemical, mechanical, and thermal stability, (vi) high rigidity, (vii) chemical reactivity for coupling of the ligands, and (viii) resistance to microbial and enzymatic attack.2 With the rapid acceleration of research in the area of nanotechnology, new nanoparticles (nonmagnetic or magnetic) that can serve as enzyme immobilization supports emerge in increasing numbers with lower cost, nontoxicity, higher activity, and stability, as well as easier enzyme availability and immobilization. Intentionally, several types of supports are now commercially available for immobilization processes. Poros particles (poly(styrene–divinylbenzene)) possessing large through-pores that allow analyte molecules to “perfuse” rapidly through the interior of the particles, as well as very short “diffusive” pores, have been introduced by Applied Biosystems (Foster City, CA, USA) into the market as Poroszyme Enzyme trypsin bulk media and cartridges (30 mm  2.1 mm ID) to perform rapid online tryptic digestion of protein.173–175 CIM Disk Monolithic Columns developed by BIA Separations (Ljubjana, Slovenia) are also utilized for trypsin immobilization and online protein digestion.199,200 In the light of current developments in nanotechnology, it is to be hoped that the availability of new support materials will lead to faster, more effective, and more stable IMERs. Due to the flow-through format of IMERs, they can be easily combined with other flow-through techniques such as HPLC or CE separation with MS detection, allowing their broad applications in tremendous areas. However, some drawbacks such as band

REFERENCES

173

broadening due to nonspecific binding to the enzyme and the influence of the flow rate on the substrate conversion efficiency in the case of specific mass transfer regimen can limit the application of IMERs in an online system. Now with the introduction of monolithic supports, it is possible to overcome all diffusion limitations so that the enzymatic kinetic parameters are flow-unaffected. IMERs are commonly used in medical diagnostics and therapy, enzyme-based electrodes, organic synthesis, kinetic study, biosensors, and many other applications such as removal of waste metabolites, blood detoxification, and/or corrections of inborn metabolic deficiency. Among these applications, most of the IMER applications are aimed at protein analysis by peptide mapping. High immobilized enzyme concentrations, in combination with the superior solute transport properties of perfusive media, allow achievement of complete digestion in minutes or even seconds rather than in hours when compared to solution-phase digest. Unlike solution-phase methods, the degree of protein digestion can be controlled through the flow rate in the IMER. The highly stable enzyme immobilization chemistry enables the reuse of the same IMERs for times of assays, with reproducible results. The integration of different steps (digestion, separation, and identification) in a single system is particularly attractive for the analysis of complex protein mixtures. However, it is noticed that, although numerous publications related to application of IMER contained online system in protein analysis have been reported so far, most of these studies concerned digestion of standard proteins, such as myoglobin, cytochrome c, R-lactalbumin, bovine serum albumin, and holo-transferrin. In contrast, few works concerned analysis of proteins in real samples (tissue, cell, etc.). Further technical improvements are needed to enable the IMERs contained online system for the detection and characterization of low-abundance analytes, and to increase the throughput of the methodologies. The final goal is to find robust, automated, and sensitive high-throughput analytical tools in bioanalytical science and in applied biotechnology.

REFERENCES 1. Miyazaki, M.; Maeda, H. Microchannel enzyme reactors and their applications for processing. Trends Biotechnol. 2006, 24, 463–470. 2. K
rrenkova, J.; Foret, F. Immobilized microfluidic enzymatic reactors. Electrophoresis 2004, 25, 3550–3563. 3. Bılkova, Z.; Slovakova, M.; Minc, N.; F€ utterer, C.; Cecal, R.; Horak, D.; Benesˇ, M.; Potier, I.; K
renkova, J.; Przybylski, M.; Viovy, J.-L. Functionalized magnetic micro- and nanoparticles: optimization and application to m-chip tryptic digestion. Electrophoresis 2006, 27, 1811–1824. 4. Thomsen, M.S.; Nidetzky, B. Microfluidic reactor for continuous flow biotransformations with immobilized enzymes: the example of lactose hydrolysis by a hyperthermophilic bglycoside hydrolase. Eng. Life Sci. 2008, 8, 40–48. 5. J€ahnisch, K.; Hessel, V.; L€ owe, H.; Baerns, M. Chemistry in microstructured reactors. Angew. Chem., Int. Ed. 2004, 43, 406–446.

174

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

6. DeMello, A.J. Control and detection of chemical reactions in microfluidic systems. Nature 2006, 442, 394–402. 7. Watts, P.; Wiles, C. Recent advances in synthetic micro reaction technology. Chem. Commun. 2007, 5, 443–67. 8. Amankwa, L.N.; Kuhr, W.G. Online peptide mapping by capillary zone electrophoresis. Anal. Chem. 1993, 65, 2693–2697. 9. Cobb, K.A.; Novotny, M. High-sensitivity peptide mapping by capillary zone electrophoresis and microcolumn liquid chromatography using immobilized trypsin for protein digestion. Anal. Chem. 1989, 61, 2226–2231. 10. Davis, M.T.; Lee, T.D.; Ronk, M.; Hefta, S.A. Microscale immobilized protease reactor columns for peptide mapping by liquid chromatography/mass spectral analyses. Anal. Biochem. 1995, 224, 235–244. 11. Hermanson, G.T. Bioconjugate Techniques; Academic Press: San Diego, CA, 1996. 12. Girelli, A.M.; Mattei, E. Application of immobilized enzyme reactor in on-line high performance liquid chromatography: a review. J. Chromatogr. B 2005, 819, 3–16. 13. Li, Y.; Yan, B.; Deng, C.-H.; Yu, W.-J.; Xu, X.-Q.; Yang, P.-Y.; Zhang, X.-M. Efficient onchip proteolysis system based on functionalized magnetic silica microspheres. Proteomics 2007, 7, 2330–2339. 14. Li, Y.; Xu, X.-Q.; Deng, C.-H.; Yang, P.-Y.; Zhang, X.-M. Immobilization of trypsin on superparamagnetic nanoparticles for rapid and effective proteolysis. J. Proteome Res. 2007, 6, 3849–3855. 15. Drobnık, J.; Vlasak, J.; Pila
r, J.; Svec, F.; Kalal, J. Synthetic polymers in the study of protein immobilization on glycidylmethacrylate carriers. Enzyme Microb. Technol. 1979, 1, 107–112. 16. Vodopivec, M.; Berovic, M.; Jancar, J.; Podgornik, A.; Strancar, A. Application of convective interaction media disks with immobilised glucose oxidase for on-line glucose measurements. Anal. Chim. Acta 2000, 407, 105–110. 17. Petro, M.; Svec, F.; Frechet, J.M.J. Immobilization of trypsin onto “molded” macroporous poly(glycidyl methacrylate-co-ethylene dimethacrylate) rods and use of the conjugates as bioreactors and for affinity chromatography. Biotechnol. Bioeng. 1996, 49, 355–363. 18. Ye, M.; Hu, S.; Schoenherr, R.M.; Dovichi, N.J. On-line protein digestion and peptide mapping by capillary electrophoresis with post-column labeling for laser-induced fluorescence detection. Electrophoresis 2004, 25, 1319–1326. 19. Svec, F.; Hrudkova, H.; Kalal, J. Reactive polymers. 20. Preparation of a carrier with aldehyde groups by oxidation of a macroporous glycidyl methacrylate-ethylenedimethacrylate copolymer with periodic acid. Angew. Makromol. Chem. 1978, 70, 101–108. 20. K
renkova, J.; Bilkova, Z.; Foret, F. Characterization of a monolithic immobilized trypsin microreactor with on-line coupling to ESI-MS. J. Sep. Sci. 2005, 28, 1675–1684. 21. Jiang, H.-H.; Zou, H.-F.; Wang, H.-L.; Ni, J.-Y.; Zhang, Q.; Zhang, Y.-K. On-line characterization of the activity and reaction kinetics of immobilized enzyme by highperformance frontal analysis. J. Chromatogr. A 2000, 903, 77–84. 22. Xie, S.; Svec, F.; Frechet, J.M.J. Design of reactive porous polymer supports for high throughput bioreactors: poly(2-vinyl-4,4-dimethylazlactone-co-acrylamide-co-ethyl dimethacrylate) monoliths. Biotechnol. Bioeng. 1999, 62, 30–35.

REFERENCES

175

23. Nouaimi, M.; M€ oschel, K.; Bisswanger, H. Immobilization of trypsin on polyester fleece via different spacers. Enzyme Microb. Technol. 2001, 29, 567–574. 24. Turkova, J. Oriented immobilization of biologically active proteins as a tool for revealing protein interactions and function. J. Chromatogr. B 1999, 722, 11–31. 25. Li, Y.; Yan, B.; Deng, C.-H.; Tang, J.; Liu, J.-Y.; Zhang, X.-M. On-plate digestion of proteins using novel trypsin-immobilized magnetic nanospheres for MALDI-TOF-MS analysis. Proteomics 2007, 7, 3661–3671. 26. Gao, J.; Xu, J.; Locascio, L.E.; Lee, C.S. Integrated microfludic system enabling protein digestion, peptide separation, and protein identification. Anal. Chem. 2001, 73, 2648–2655. 27. Cooper, J.W.; Chen, J.; Li, Y.; Lee, C.S. Membrane-based nanoscale proteolytic reactor enabling protein digestion, peptide separation, and protein identification using mass spectrometry. Anal. Chem. 2003, 75, 1067–1074. 28. Massolini, G.; Calleri, E.; De Lorenzi, E.; Pregnolato, M.; Terreni, M.; Felix, G.; Gandini, C. Immobilized penicillin G acylase as reactor and chiral selector in liquid chromatography. J. Chromatogr. A 2001, 921, 147–160. 29. Calleri, E.; Temporini, C.; Furlanetto, S.; Loiodice, F.; Fracchiolla, G.; Massolini, G. Lipases for biocatalysis: development of a chromatographic bioreactor. J. Pharm. Biomed. Anal. 2003, 32, 715–724. 30. Guo, Z.; Xu, S.-Y.; Lei, Z.-D.; Zou, H.-F.; Guo, B.-C. Immobilized metal-ion chelating capillary microreactor for peptide mapping analysis of proteins by matrix assisted laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 2003, 24, 3633–3639. 31. Anspach, F.B. Silica-based metal chelate affinity sorbents. I. Preparation and characterization of iminodiacetic acid affinity sorbents prepared via different immobilization techniques. J. Chromatogr. A 1994, 672, 35–49. 32. Li, Y.; Xu, X.-Q.; Yang, B.; Deng, C.-H.; Yu, W.-J.; Yang, P.-Y.; Zhang, X.-M. Microchip reactor packed with metal-ion chelated magnetic silica microspheres for high efficient proteolysis. J. Proteome Res. 2007, 6, 2367–2375. 33. Li, Y.; Yan, B.; Xu, X.-Q.; Deng, C.-H.; Yang, P.-Y.; Shen, X.-Z.; Zhang, X.-M. Oncolumn tryptic mapping of proteins using metal-ion-chelated magnetic silica microspheres by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun. Mass Spectrom. 2007, 21, 2263–2268. 34. Lvov, Y.; Decher, G.; Mohwald, H. Assembly, structural characterization, and thermal behavior of layer-by-layer deposited ultrathin films of poly(vinyl sulfate) and poly (allylamine). Langmuir 1993, 9, 481–486. 35. Decher, G. Fuzzy nanoassemblies: toward layered polymeric multicomposites. Science 1997, 277, 1232–1237. 36. Serizawa, T.; Yamaguchi, M.; Akashi, M. Enzymatic hydrolysis of a layer-by-layer assembly prepared from chitosan and dextran sulfate. Macromolecules 2002, 35, 8656–8658. 37. Ichinose, I.; Takaki, R.; Kuroiwa, K.; Kunitake, T. Electrostatic desorption of cytochrome c on ultrathin ZrO2-gel layers and preparation of alternate multilayers. Langmuir 2003, 19, 3883–3888. 38. Salloum, D.S.; Schlenoff, J.B. Protein adsorption modalities on polyelectrolyte multilayers. Biomacromolecules 2004, 5, 1089–1096.

176

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

39. Zhou, X.-C.; Wu, L.-Y.; Zhou, J.-Z. Fabrication of DNA microarrays on nanoengineered polymeric ultrathin film prepared by self-assembly of polyelectrolyte multilayers. Langmuir 2004, 20, 8877–8885. 40. Liu, Y.; Lu, H.-J.; Zhong, W.; Song, P.-Y.; Kong, J.-L.; Yang, P.-Y.; Girault, H.H.; Liu, B.H. Multilayer-assembled microchip for enzyme immobilization as reactor toward lowlevel protein identification. Anal. Chem. 2006, 78, 801–808. 41. Liu, Y.; Zhong, W.; Meng, S.; Kong, J.-L.; Lu, H.-J.; Yang, P.-Y.; Girault, H.H.; Liu, B.-H. Assembly-controlled biocompatible interface on a microchip: strategy to highly efficient proteolysis. Chem. Eur. J. 2006, 12, 6585–6591. 42. Amankwa, L.N.; Kuhr, W.G. Trypsin-modified fused-silica capillary microreactor for peptide mapping by capillary zone electrophoresis. Anal. Chem. 1992, 64, 1610–1613. 43. Gleason, N.J.; Carbeck, J.D. Measurement of enzyme kinetics using microscale steadystate kinetic analysis. Langmuir 2004, 20, 6374–6381. 44. Mao, H.-B.; Yang, T.-L.; Cremer, P.S. Design and characterization of immobilized enzymes in microfluidic systems. Anal. Chem. 2002, 74, 379–385. 45. Holden, M.A.; Jung, S.-Y.; Cremer, P.S. Patterning enzymes inside microfluidic channels via photoattachment chemistry. Anal. Chem. 2004, 76, 1838–1843. 46. Anderson, N.L.; Haines, L.R.; Pearson, T.W. An effective and rapid method for functional characterization of immunoadsorbents using POROS beads and flow cytometry. J. Proteome Res. 2004, 3, 228–234. 47. Sakai-Kato, K.; Kato, M.; Toyo’oka, T. On-line trypsin-encapsulated enzyme reactors by the sol–gel method integrated into capillary electrophoresis. Anal. Chem. 2002, 74, 2943–2949. 48. Sakai-Kato, K.; Kato, M.; Toyo’oka, T. Creation of on-chip enzyme reactors by encapsulating trypsin in sol–gel on a plastic microchip. Anal. Chem. 2003, 75, 388–393. 49. Dave, B.C.; Dunn, B.; Valentine, J.S.; Zink, J.I. Sol–gel encapsulation methods for biosensors. Anal. Chem. 1994, 66, 1120–1127. 50. Gill, I.; Ballesteros, A. Bioencapsulation within synthetic polymers. Part 1. Sol–gel encapsulated biologicals. Trends Biotechnol. 2000, 18, 282–296. 51. Hahn, R.; Podgornik, A.; Merhar, M.; Schallaun, E.; Jungbauer, A. Affinity monoliths generated by in situ polymerization of the ligand. Anal. Chem. 2001, 73, 5126–5132. 52. Kim, Y.-D.; Park, C.B.; Clark, D.S. Stable sol–gel microstructured and microfluidic networks for protein patterning. Biotechnol. Bioeng. 2001, 73, 331–337. 53. Qu, H.-Y.; Wang, H.-T.; Huang, Y.; Zhong, W.; Lu, H.-J.; Kong, J.-L.; Yang, P.-Y.; Liu, B.H. Stable microstructured network for protein patterning on a plastic microfluidic channel: strategy and characterization of on-chip enzyme microreactors. Anal. Chem. 2004, 76, 6426–6433. 54. Park, C.B.; Clark, D.S. Sol–gel encapsulated enzyme arrays for high-throughput screening of biocatalytic activity. Biotechnol. Bioeng. 2002, 78, 229–235. 55. Liu, Z.-J.; Liu, B.-H.; Kong, J.-L.; Deng, J.-Q. Probe trace phenols based on mediator-free alumina sol–gel derived tyrosinase biosensor. Anal. Chem. 2000, 72, 4707–4712. 56. Jiang, D.-C.; Tang, J.; Liu, B.-H.; Yang, P.-Y.; Kong, J.-L. Ultrathin alumina sol–gelderived films: allowing direct detection of the liver fibrosis markers by capacitance measurement. Anal. Chem. 2003, 75, 4578–4584.

REFERENCES

177

57. Wu, H.-L.; Tian, Y.-P.; Liu, B.-H.; Lu, H.-J.; Wang, X.-Y.; Zhai, J.-J.; Jin, H.; Yang, P.-Y.; Xu, Y.-M.; Wang, H.-H. Titania and alumina sol–gel-derived microfluidic enzymaticreactors for peptide mapping: design, characterization, and performance. J. Proteome Res. 2004, 3, 1201–1209. 58. Frenkel-Mullerad, H.; Avnir, D. Sol–gel materials as efficient enzyme protectors: preserving the activity of phosphatases under extreme pH conditions. J. Am. Chem. Soc. 2005, 127, 8077–8081. 59. Itoh, T.; Ishii, R.; Ebina, T.; Hanaoka, T.; Fukushima, Y.; Mizukami, F. Encapsulation of myoglobin with a mesoporous silicate results in new capabilities. Bioconj. Chem. 2006, 17, 236–240. 60. Kumar, R.; Maitra, A.N.; Patanjali, P.K.; Sharma, P. Hollow gold nanoparticles encapsulating horseradish peroxidase. Biomaterials 2005, 26, 6743–6753. 61. Ma, D.; Li, M.; Patil, A.J.; Mann, S. Fabrication of protein/silica core–shell nanoparticles by microemulsion-based molecular wrapping. Adv. Mater. 2004, 16, 1838–1841. 62. Kim, J.; Grate, J.W. Single-enzyme nanoparticles armored by a nanometer-scale organic/ inorganic network. Nano Lett. 2003, 3, 1219–1222. 63. Yan, M.; Ge, J.; Liu, Z.; Ouyang, P.-K. Encapsulation of single enzyme in nanogel with enhanced biocatalytic activity and stability. J. Am. Chem. Soc. 2006, 128, 11008–11009. 64. Licklider, L.; Kuhr, W.G. Optimization of online peptide mapping by capillary zone electrophoresis. Anal. Chem. 1994, 66, 4400–4407. 65. Licklider, L.; Kuhr, W.G.; Lacey, M.P.; Keough, T.; Purdon, M.P.; Takigiku, R. Online microreactors/capillary electrophoresis/mass spectrometry for the analysis of proteins and peptides. Anal. Chem. 1995, 67, 4170–4177. 66. Licklider, L.; Kuhr, W.G. Characterization of reaction dynamics in a trypsin-modified capillary microreactor. Anal. Chem. 1998, 70, 1902–1908. 67. Laurell, T.; Rosengren, L. A micromachined enzyme reactor in (110)-oriented silicon. Sens. Actuators B 1994, 19, 614–617. 68. Laurell, T.; Drott, J.; Rosengren, L. Silicon-wafer integrated enzyme reactors. Biosens. Bioelectron. 1995, 10, 289–299. 69. Lee, J.; Musyimi, H.K.; Soper, S.A.; Murray, K.K. Development of an automated digestion and droplet deposition microfluidic chip for MALDI-TOF MS. J. Am. Soc. Mass Spectrom. 2008, 19, 964–972. 70. K
renkova, J.; Kleparnık, K.; Foret, F. Capillary electrophoresis mass spectrometry coupling with immobilized enzyme electrospray capillaries. J. Chromatogr. A 2007, 1159, 110–118. ¨ nnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; Marko-Varga, G. 71. Ekstr€ om, S.; O Integrated microanalytical technology enabling rapid and automated protein identification. Anal. Chem. 2000, 72, 286–293. 72. Miyazaki, M.; Kaneno, J.; Uehara, M.; Fujii, M.; Shimizu, H.; Maeda, H. Simple method for preparation of nanostructure on microchannel surface and its usage for enzyme immobilization. Chem. Commun. 2003, 5, 648–649. 73. Wu, H.-L.; Zhai, J.-J.; Tian, Y.-P.; Lu, H.-J.; Wang, X.-Y.; Jia, W.-T.; Liu, B.-H.; Yang, P.Y.; Xu, Y.-M.; Wang, H.-H. Microfluidic enzymatic-reactors for peptide mapping: strategy, characterization, and performance. Lab Chip 2004, 4, 588–597.

178

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

74. Hessel, V.; L€ owe, H.; Sch€ onfeld, F. Micromixers: a review on passive and active mixing principles. Chem. Eng. Sci. 2005, 60, 2479–2501. 75. Xiong, L.; Regnier, F.E. Channel-specific coatings on microfabricated chips. J. Chromatogr. A 2001, 924, 165–176. 76. Park, J.J.; Luo, X.-L.; Yi, H.-M.; Valentine, T.M.; Payne, G.F.; Bentley, W.E.; Ghodssi, R.; Rubloff, G.W. Chitosan-mediated in situ biomolecule assembly in completely packaged microfluidic devices. Lab Chip 2006, 6, 1315–1321. 77. Luo, X.; Lewandowski, A.T.; Yi, H.; Payne, G.F.; Ghodssi, R.; Bentley, W.E. Programmable assembly of a metabolic pathway enzyme in a pre-packaged reusable bioMEMS device. Lab Chip 2008, 8, 420–430. 78. Luo, X.-L.; Berlin, D.L.; Buckhout-White, S.; Bentley, W.E.; Payne, G.F.; Ghodssi, R.; Rubloff, G.W. Design optimization for bioMEMS studies of enzyme-controlled metabolic pathways. Biomed. Microdevices 2008, 10, 899–908. 79. Tomer, S.; Dorsey, J.G.; Berthod, A. Nonionic micellar liquid chromatography coupled to immobilized enzyme reactors. J. Chromatogr. A 2001, 923, 7–16. 80. Shu, H.-C.; Wu, N.-P. A chemically modified carbon paste electrode with D-lactate dehydrogenase and alanine aminotranferase enzyme sequences for D-lactic acid analysis. Talanta 2001, 54, 361–368. 81. Lim, L.W.; Tomatsu, M.; Takeuchi, T. Development of an on-line immobilized-enzyme reversed-phase HPLC method for protein digestion and peptide separation. Anal. Bioanal. Chem. 2006, 386, 614–620. 82. Richter, T.; Shultz-Lockyear, L.L.; Oleschuk, R.D.; Bilitewski, U.; Harrison, D.J. Bienzymatic and capillary electrophoretic analysis of non-fluorescent compounds in microfluidic devices determination of xanthine. Sens. Actuators B 2002, 81, 369–376. 83. Bertucci, C.; Petri, A.; Felix, G.; Perini, B.; Salvatori, P. Lipase-based HPLC stationary phase: enantioselective synthesis of 2-substituted 1,3-propanediol monoacetates. Tetrahedron: Asymmetry 1999, 10, 4455–4462. 84. Kiba, N.; Saegusa, K.; Furosawa, M. Post-column enzyme reactors for chemiluminometric detection of glucose 1,5-anhydroglucitol and 3-hydroxybutyrate in an anionexchange chromatographic system. J. Chromatogr. B 1997, 689, 393–398. 85. Seong, G.H.; Crooks, R.M. Efficient mixing and reactions within microfluidic channels using microbead-supported catalysts. J. Am. Chem. Soc. 2002, 124, 13360–13361. 86. Srinivasan, A.; Bach, H.; Sherman, D.H.; Dordick, J.S. Bacterial P450-catalyzed polyketide hydroxylation on a microfluidic platform. Biotechnol. Bioeng. 2004, 88, 528–535. 87. Oleschuk, R.D.; Shultz-Lockyear, L.L.; Ning, Y.-B.; Harrison, D.J. Trapping of beadbased reagents within microfluidic systems: on-chip solid-phase extraction and electrochromatography. Anal. Chem. 2000, 72, 585–590. 88. Kerby, M.B.; Legge, R.S.; Tripathi, A. Measurements of kinetic parameters in a microfluidic reactor. Anal. Chem. 2006, 78, 8273–8280. 89. Nakamura, H.; Murakami, Y.; Yokoyama, K.; Tamiya, E.; Karube, I. A compactly integrated flow cell with a chemiluminescent FIA system for determining lactate concentration in serum. Anal. Chem. 2001, 73, 373–378. 90. Andersson, H.; van der Wijngaart, W.; Enoksson, P.; Stemme, G. Micromachined flowthrough filter-chamber for chemical reactions on beads. Sens. Actuators B 2000, 67, 203–208.

REFERENCES

179

91. Sato, K.; Tokeshi, M.; Odake, T.; Kimura, H.; Ooi, T.; Nakao, M.; Kitamori, T. Integration of an immunosorbent assay system: analysis of secretory human immunoglobulin A on polystyrene beads in a microchip. Anal. Chem. 2000, 72, 1144–1147. 92. Ceriotti, L.; de Rooij, N.F.; Verpoorte, E. An integrated fritless column for on-chip capillary chromatography with conventional stationary phases. Anal. Chem. 2002, 74, 639–647. 93. Zhang, Q.; Xu, J.-J.; Chen, H.-Y. Glucose microfluidic biosensors based on immobilizing glucose oxidase in poly(dimethylsiloxane) electrophoretic microchips. J. Chromatogr. A 2006, 1135, 122–126. 94. Xu, Z.-R.; Fang, Z.-L. Composite poly(dimethylsiloxane)/glass microfluidic system with an immobilized enzymatic particle-bed reactor and sequential sample injection for chemiluminescence determinations. Anal. Chim. Acta 2004, 507, 129–135. 95. Miyazaki, M.; Kaneno, J.; Kohama, R.; Uehara, M.; Kanno, K.; Fujii, M.; Shimizu, H.; Maeda, H. Preparation of functionalized nanostructures on microchannel surface and their use for enzyme microreactors. Chem. Eng. J. 2004, 101, 277–284. 96. Miyazaki, M.; Kaneno, J.; Yamaori, S.; Honda, T.; Briones, M.P.P.; Uehara, M.; Arima, K.; Kanno, K.; Yamashita, K.; Yamaguchi, Y.; Nakamura, H.; Yonezawa, H.; Fujii, M.; Maeda, H. Efficient immobilization of enzymes on microchannel surface through His-tag and application for microreactor. Protein Pept. Lett. 2005, 12, 207–210. 97. Merkoci, A.; Marta, A.; Tarraso’n, G.; Eritja, R.; Alegret, S. Toward an ICPMS-linked DNA assay based on gold nanoparticles immunoconnected through peptide sequences. Anal. Chem. 2005, 77, 6500–6503. 98. Liu, Y.; Xue, Y.; Ji, J.; Chen, X.; Kong, J.-L.; Yang, P.-Y.; Girault, H.H.; Liu, B.-H. Gold nanoparticles assembly microfluidic reactor for efficient on-line proteolysis. Mol. Cell. Proteomics 2007, 6, 1428–1436. 99. Tosheva, L.; Valtchev, V.P. Nanozeolites: synthesis, crystallization mechanism, and applications. Chem. Mater. 2005, 17, 2494–2513. 100. Huang, Y.; Shan, W.; Liu, B.-H.; Liu, Y.; Zhang, Y.-H.; Zhao, Y.; Lu, H.-J.; Tang, Y.; Yang, P.-Y. Zeolite nanoparticle modified microchip reactor for efficient protein digestion. Lab Chip 2006, 6, 534–539. 101. Patel, D.; Moon, J.Y.; Chang, Y.; Kim, T.J.; Lee, G.H. Poly(D,L-lactide-co-glycolide) coated superparamagnetic iron oxide nanoparticles: synthesis, characterization and in vivo study as MRI contrast agent. Colloid Surf. A 2008, 313, 91–94. 102. Zhao, M.; Josephson, L.; Tang, Y.; Weissleder, R. Magnetic sensors for protease assays. Angew. Chem., Int. Ed. 2003, 42, 1375–1378. 103. Mornet, S.; Vasseur, S.; Grasset, F.; Veverka, P.; Goglio, G.; Demourgues, A.; Portier, J.; Pollert, E.; Duguet, E. Magnetic nanoparticle design for medical applications. Prog. Solid State Chem. 2006, 34, 237–247. 104. Stevens, P.D.; Fan, J.; Gardimalla, H.M.R.; Yen, M.; Gao, Y. Superparamagnetic nanoparticle-supported catalysis of Suzuki cross-coupling reactions. Org. Lett. 2005, 7, 2085–2088. 105. Jun, Y.-W.; Choi J.-S.; Cheon, J. Heterostructured magnetic nanoparticles: their versatility and high performance capabilities. Chem. Commun. 2007, 1203–1214. 106. Choi, W.; Oh, K.W.; Thomas, J.H.; Heineman, W.R.; Halsall, H.B.; Nevin, J.H.; Helmicki, A.J.; Henderson, H.T.; Ahn, C.H. An integrated microfluidic biochemical

180

107.

108. 109.

110. 111.

112.

113.

114.

115. 116. 117.

118.

119. 120.

121.

122.

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

detection system for protein analysis with magnetic bead-based sampling capabilities. Lab Chip 2002, 2, 27–30. Rashkovetsky, L.G.; Lyubarskaya, Y.V.; Foret, F.; Hughes, D.E.; Karger, B.L. Automated microanalysis using magnetic beads with commercial capillary electrophoretic instrumentation. J. Chromatogr. A 1997, 781, 197–204. Xu, X.; Friedman, G.; Humfeld, K.D.; Majetich, S.A.; Asher, S.A. Superparamagnetic photonic crystals. Adv. Mater. 2001, 13, 1681–1684. Lu, Y.; Yin, Y.; Mayers, B.T.; Xia, Y. Modifying the surface properties of superparamagnetic iron oxide nanoparticles through a sol–gel approach. Nano Lett. 2002, 2, 183–186. Wu, W.; He, Q.-G.; Jiang, C.-Z. Magnetic iron oxide nanoparticles: synthesis and surface fictionalization strategies. Nanaoscale Res. Lett. 2008, 3, 397–415. Slovakova, M.; Minc, N.; Bılkova, Z.; Smadja, C.; Faigle, W.; F€ utterer, C.; Taverna, M.; Viovy, J.-L. Use of self assembled magnetic beads for on-chip protein digestion. Lab Chip 2005, 9, 935–942. Koh, I.; Wang, X.; Varughese, B.; Isaacs, L.; Ehrman, S.E.; English, D.S. Magnetic iron oxide nanoparticles for biorecognition: evaluation of surface coverage and activity. J. Phys. Chem. B 2006, 110, 1553–1558. Nishimura, K.; Hasegawa, M.; Ogura, Y.;et al. 4 C preparation of ferrite nanoparticles having protein molecules immobilized on their surfaces. J. Appl. Phys. 2002, 91, 8555–8556. Liu, J.-Y.; Lin, S.; Qi, D.-W.; Deng, C.-H.; Yang, P.-Y.; Zhang, X.-M. On-chip enzymatic microreactor using trypsin-immobilized superparamagnetic nanoparticles for highly efficient proteolysis. J. Chromatogr. A 2007, 1176, 169–177. Hayes, M.A.; Polson, N.A.; Phayre, A.N.; Garcia, A.A. Flow-based microimmunoassay. Anal. Chem. 2001, 73, 5896–5902. Lawrence, E.M.; Ivey, M.L.; Flores, G.A.; Lui, J.; Bibette, J.; Richard, J. Field-induced structure of confined ferrofluid emulsion. Int. J. Mod. Phys. B 1994, 8, 2765–2777. Svec, F. Less common applications of monoliths. I. Microscale protein mapping with proteolytic enzymes immobilized on monolithic supports. Electrophoresis 2006, 27, 947–961. Yi, Y.Y.; Chen, Y.; Brook, M.A.; Brennan, J.D. Development of macroporous titania monoliths by a biocompatible method. Part 2. Enzyme entrapment studies. Chem. Mater. 2006, 18, 5336–5342. Cabrera, K. Applications of silica-based monolithic HPLC columns. J. Sep. Sci. 2004, 27, 843–852. Calleri, E.; Massolini, G.; Lubda, D.; Temporini, C.; Loiodice, F.; Caccialanza, G. Evaluation of a monolithic epoxy silica support for penicillin G acylase immobilization. J. Chromatogr. 1031, 2004, 93–100. Calleri, E.; Marrubini, G.; Massolini, G.; Lubda, D.; de Fazio, S.S.; Furlanetto, S.; Wainer, I. W.; Manzo, L.; Caccialanza, G. Development of a chromatographic bioreactor based on immobilized beta-glucuronidase on monolithic support for the determination of dextromethorphan and dextrorphan in human urine. J. Pharm. Biomed. Anal. 2004, 35, 1179–1189. Chen, Z.; Hayashi, K.; Iwasaki, Y.; Kurita, R.; Niwa, O.; Sunaawa, K. On-line monolithic enzyme reactor fabricated by sol–gel process for elimination of ascorbic acid while monitoring dopamine. Electroanalysis 2005, 17, 231–238.

REFERENCES

181

123. Kawakami, K.; Sera, Y.; Sakai, S.; Ono, T.; Ijima, H. Development and characterization of a silica monolith immobilized enzyme micro-bioreactor. Ind. Eng. Chem. Res. 2005, 44, 236–240. 124. Sakai-Kato, K.; Kato, M.; Toyo’oka, T. On-line trypsin-encapsulated enzyme reactor by the sol–gel method integrated into capillary electrophoresis. Anal. Chem. 2002, 74, 2943–2949. 125. Kato, M.; Inuzuka, K.; Sakai-Kato, K.; Toyo’oka, T. Monolithic bioreactor immobilizing trypsin for high-throughput analysis. Anal. Chem. 2005, 77, 1813–1818. 126. Dulay, M.T.; Quirino, J.P.; Bennett, D.C.; Kato, M.; Zare, R.N. Photopolymerized sol–gel monoliths for capillary electrochromatography. Anal. Chem. 2001, 73, 3921–3926. 127. Kato, M.; Sakai-Kato, K.; Jin, H.M.; Kubota, K.; Miyano, H.; Toyo’oko, T.; Dulay, M.T.; Zare, R.N. Integration of on-line protein digestion, peptide separation, and protein identification using pepsin-coated photopolymerized sol–gel columns and capillary electrophoresis/mass spectrometry. Anal. Chem. 2004, 76, 1896–1902. 128. Dulay, M.T.; Baca, Q.J.; Zare, R.N. Enhanced proteolytic activity of covalently bound enzymes in photopolymerized sol gel. Anal. Chem. 2005, 77, 4604–4610. 129. Kostner, A.; Mandel, M. In Immobilized Enzymes; Mosbach, K., Ed.; Academic Press: New York, 1976; pp 191–195. 130. Mersal, G.A.M.; Bilitewski, U. Development of monolithic enzymatic reactors in glass microchips for the quantitative determination of enzyme substrates using the example of glucose determination via immobilized glucose oxidase. Electrophoresis 2005, 26, 2303–2312. 131. Palm, A.; Novotny, M.V. Analytical characterization of a facile porous polymer monolithic trypsin microreactor enabling peptide mass mapping using mass spectrometry. Rapid Commun. Mass Spectrom. 2004, 18, 1374–1382. 132. Palm, A.; Novotny, M.V. A monolithic PNGase F enzyme microreactor enabling glycan mass mapping of glycoproteins by mass spectrometry. Rapid Commun. Mass Spectrom. 2005, 19, 1730–1738. 133. Hilder, E.F.; Svec, F.; Frechet, J.M.J. Polymeric monolithic stationary phases for capillary electrochromatography. Electrophoresis 2002, 23, 3934–3953. 134. Ben
cina, K.; Podgornik, A.; Sˇtrancar, A.; Ben
cina, M. Enzyme immobilization on epoxyand 1,10 -carbonyldiimidazole-activated methacrylate-based monoliths. J. Sep. Sci. 2004, 27, 811–818. 135. Hermanson, G.T.; Mallia, A.K.; Smith, P.K. Immobilized Affinity Ligand Techniques; Academic Press: New York, 1992. 136. Hrudkova, H.; Svec, F.; Kalal, J. Reactive polymers. 14. Hydrolysis of epoxide groups of copolymer glycidylmethacrylate-ethylenedimethacrylate. Br. Polym. J. 1977, 9, 238–240. 137. Luo, Q.-Z.; Mao, X.-Q.; Huang, X.; Zou, H.-F. High-performance affinity chromatography for characterization of human immunoglobulin G digestion with papain. J. Chromatogr. B 2002, 776, 139–147. 138. K
renkova, J.; Lacher, N.A.; Svec, F. Highly efficient enzyme reactors containing trypsin and endoproteinase LysC immobilized on porous polymer monolith coupled to MS suitable for analysis of antibodies. Anal. Chem. 2009, 81, 2004–2012. 139. Valentova, O.; Marek, M.; Svec, F.; Sˇtamberg, J.; Vodra
zka, Z. Comparison of different methods of glucose oxidase immobilization. Biotechnol. Bioeng. 1981, 23, 2093–2104.

182

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

140. Hearn, W.; Bethell, G.S.; Ayers, J.S.; Hancock, W.S. Application of 1,10 -carbonyldiimidazone-activated agarose for the purification of proteins. II. The use of an activated matrix devoid of additional charged groups for the purification of thyroid proteins. J. Chromatogr. 1979, 185, 463–470. 141. Coleman, P.L.; Walker, M.M.; Milbrath, D.S.; Stauffer, D.M.; Rasmussen, J.K.; Krepski, L.R.; Heilmann, S.M. Immobilization of protein A at high density on azlactone-functional polymeric beads and their use in affinity chromatography. J. Chromatogr. A 1990, 512, 345–363. 142. Heilmann, S.M.; Rasmussen, J.K.; Krepski, L.R. Chemistry and technology of 2-alkenyl azlactones. J. Polym. Sci., Polym. Chem. 2001, 39, 3655–3677. 143. Drtina, G.J.; Haddad, L.C.; Rasmussen, J.K.; Gaddam, B.N.;et al. Azlactone-reactive polymer supports for immobilizing synthetically useful enzymes. II. Important preliminary hydrogen bonding effects in the covalent coupling of penicillin G acylase. React. Funct. Polym. 2005, 64, 13–24. 144. Heilmann, S.M.; Drtina, G.J.; Haddad, C.; Rasmussen, J.K.;et al. Azlactone-reactive polymer supports for immobilizing synthetically useful enzymes. Part I. Pig liver esterase on dispersion polymer supports. J. Mol. Catal. B 2004, 30, 33–42. 145. Xie, S.; Svec, F.; Frechet, J.M.J. Design of reactive porous polymer supports for high throughput bioreactors: poly(2-vinyl-4,4-dimethylazlactone-co-acrylamide-co-ethylenedimethacrylate) monoliths. Biotechnol. Bioeng. 1999, 62, 30–35. 146. Peterson, D.S.; Rohr, T.; Svec, F.; Frechet, J.M.J. Enzymatic microreactor-on-a-chip: protein mapping using trypsin immobilized on porous polymer monoliths molded in channels of microfluidic devices. Anal. Chem. 2002, 74, 4081–4088. 147. Peterson, D.S.; Rohr, T.; Svec, F.; Frechet, J.M.J. High-throughput peptide mass mapping using a microdevice containing trypsin immobilized on a porous polymer monolith coupled to MALDI TOF and ESI TOF mass spectrometers. J. Proteome Res. 2002, 1, 563–568. 148. Geiser, L.; Eeltink, S.; Svec, F.; Frechet, J.M.J. In-line system containing porous polymer monoliths for protein digestion with immobilized pepsin, peptide preconcentration and nano-liquid chromatography separation coupled to electrospray ionization mass spectroscopy. J. Chromatogr. A 2008, 1188, 88–96. 149. Svec, F. High-throughput peptide mass mapping using an integrated capillary device coupled to a mass spectrometer. LC-GC Eur. 2005, 18, 17–20. 150. Wang, P.-C.; DeVoe, D.L.; Lee, C.S. Integration of polymeric membranes with microfluidic networks for bioanalytical application. Electrophoresis 2001, 22, 3857–3867. 151. Hisamoto, H.; Shimizu, Y.; Uchiyama, K.; Tokeshi, M.; Kikutani, Y.; Hibara, A.; Kitamori, T. Chemicofunctional membrane for integrated chemical processes on a microchip. Anal. Chem. 2003, 75, 350–354. 152. Honda, T.; Miyazaki, M.; Nakamura, H.; Maeda, H. Immobilization of enzyme on a microchannel surface through cross-linking polymerization. Chem. Commun. 2005, 5062–5064. 153. Dogruel, D.; Williams, P.; Nelson, R.W. Rapid tryptic mapping using enzymatically active mass-spectrometer probe tips. Anal. Chem. 1995, 67, 4343–4348. 154. Nelson, R.W.; Dogruel, D.; Krone, J.R.; Williams, P. Peptide characterization using bioreactive mass-spectrometer probe tips. Rapid Commun. Mass Spectrom. 1995, 9, 1380–1385.

REFERENCES

183

155. Houston, C.T.; Reilly, J.P. Toward a simple, expedient, and complete analysis of human hemoglobin by MALDI-TOFMS. Anal. Chem. 1999, 71, 3397–3404. 156. Zheng, S.P.; Yoo, C.; Delmotte, N.; Miller, F.R.; Huber, C.G.; Lubman, D.M. Monolithic column HPLC separation of intact proteins analyzed by LC-MALDI using on-plate digestion: an approach to integrate protein separation and identification. Anal. Chem. 2006, 78, 5198–5204. 157. Ota, S.; Miyazaki, S.; Matsuoka, H.; Morisato, K.; Shintani, Y.; Nakanishi, K. Highthroughput protein digestion by trypsin-immobilized monolithic silica with pipette-tip formula. J. Biochem. Biophys. Methods 2007, 70, 57–62. 158. Fan, H.-Z.; Chen, G. Fiber-packed channel bioreactor for microfluidic protein digestion. Proteomics 2007, 7, 3445–3449. 159. Wang, S.; Chen, Z.; Yang, P.-Y.; Chen, G. Trypsin-immobilized fiber core in syringe needle for highly efficient proteolysis. Proteomics 2008, 8, 1785–1788. 160. Urban, P.L.; Goodall, D.M.; Bruce, N.C. Enzymatic microreactors in chemical analysis and kinetic studies. Biotechnol. Adv. 2006, 24, 42–57. 161. Massolini, G.; Calleri, E. Immobilized trypsin systems coupled on-line to separation methods: recent developments and analytical applications. J. Sep. Sci. 2005, 28, 7–21. 162. Laurent, N.; Haddoub, R.; Flitsch, S.L. Enzyme catalysis on solid surfaces. Trends Biotechnol. 2008, 26, 328–337. 163. Ma, J.-F.; Zhang, L.-H.; Liang, Z.; Zhang, W.-B.; Zhang, Y.-K. Recent advances in immobilized enzymatic reactors and their applications in proteome analysis. Anal. Chim. Acta 2009, 632, 1–8. 164. Foret, F.; Preisler, J. Liquid phase interfacing and miniaturization in matrix-assisted laser desorption/ionization mass spectrometry. Proteomics 2002, 2, 360–372. 165. Julka, S.; Regnier, F. Quantification in proteomics through stable isotope coding: a review. J. Proteome Res. 2004, 3, 350–363. 166. Bruyneel, B.; Hoos, J.S.; Smoluch, M.T.; Lingeman, H.; Niessen, W.M.A.; Irth, H. Trace analysis of proteins using postseparation solution-phase digestion and electrospray mass spectrometric detection of marker peptides. Anal. Chem. 2007, 79, 1591–1598. 167. Lazar, I.M.; Ramsey, R.S.; Ramsey, J.M. On-chip proteolytic digestion and analysis using “wrong-way-round” electrospray time-of-flight mass spectrometry. Anal. Chem. 2001, 73, 1733–1739. 168. Peterson, D.S.; Rohr, T.; Svec, F.; Frechet, J.M.J. Dual-function microanalytical device by in situ photolithographic grafting of porous polymer monolith: integrating solid-phase extraction and enzymatic digestion for peptide mass mapping. Anal. Chem. 2003, 75, 5328–5335. 169. Drott, J.; Rosengren, L.; Lindstr€ om, K.; Laurell, T. Pore morphology influence on catalytic turn-over for enzyme activated porous silicon matrices. Thin Solid Films 1998, 330, 161–166. 170. Drott, J.; Rosengren, L.; Lindstr€ om, K.; Laurell, T. Porous silicon carrier matrices in micro enzyme reactors-influence of matrix depth. Mikrochim. Acta 1999, 131, 115–120. 171. Cooper, J.W.; Lee, C.S. Integrated and ultrasensitive gel protein identification. Anal. Chem. 2004, 76, 2196–2202. 172. Calleri, E.; Temporini, C.; Perani, E.; Palma, A.D.; Lubda, D.; Mellerio, G.; Sala, A.; Galliano, M.; Caccialanza, G.; Massolini, G. Trypsin-based monolithic bioreactor

184

173.

174. 175. 176.

177. 178. 179.

180.

181.

182.

183.

184.

185. 186.

187.

188.

MICROFLUIDIC ENZYMATIC REACTORS USING NANOPARTICLES

coupled on-line with LC/MS/MS system for protein digestion and variant identification in standard solutions and serum samples. J. Proteome Res. 2005, 4, 481. Samskog, J.; Bylund, D.; Jacobsson, S.P.; Markides, K.E. Miniaturized on-line proteolysis–capillary liquid chromatography–mass spectrometry for peptide mapping of lactate dehydrogenase. J. Chromatogr. A 2003, 998, 83–91. Hsieh, Y.L.F.; Wang, H.; Elicone, C.; Mark, J.; Martin, S.A.; Regnier, F. Automated analytical system for the examination of protein primary structure. Anal. Chem. 1996, 68, 455–462. Riggs, L.; Sioma, C.; Regnier, F.E. Automated signature peptide approach for proteomics. J. Chromatogr. A 2001, 924, 359–368. Bonneil, E.; Mercier, M.; Waldron, K.C. Reproducibility of a solid-phase trypsin microreactor for peptide mapping by capillary electrophoresis. Anal. Chim. Acta 2000, 404, 29–45. Bonneil, E.; Waldron, K.C. On-line system for peptide mapping by capillary electrophoresis at sub-micromolar concentrations. Talanta 2000, 53, 687–699. Schoenherr, R.M.; Ye, M.-L.; Vannatta, M.; Dovichi, N.J. CE-microreactor-CE-MS/MS for protein analysis. Anal. Chem. 2007, 79, 2230–2238. Bossi, A.; Guizzardi, L.; D’Acuto, M.R.; Righetti, P.G. Controlled enzyme-immobilisation on capillaries for microreactors for peptide mapping. Anal. Bioanal. Chem. 2004, 378, 1722–1728. Wang, J.; Chatrathi, M.P.; Tian, B.; Polsky, R. Microfabricated electrophoresis chips for simultaneous bioassays of glucose, uric acid, ascorbic acid, and acetaminophen. Anal. Chem. 2000, 72, 2514–2518. Wang, J.; Chatrathi, M.P.; Tian, B. Microseparation chips for performing multienzymatic dehydrogenase/oxidase assays: simultaneous electrochemical measurement of ethanol and glucose. Anal. Chem. 2001, 73, 1296–1300. Wang, J.; Chatrathi, M.P.; Iban˜ez, A.; Escarpa, A. Micromachined separation chips with post-column enzymatic reactions of “class” enzymes and end-column electrochemical detection: assays of amino acids. Electroanalysis 2002, 14, 400–404. L’Hostis, E.; Michel, P.E.; Fiaccabrino, G.C.; Strike, D.J.; de Rooij, N.F.; Koudelka-Hep, M. Microreactor and electrochemical detectors fabricated using Si and EPON SU-8. Sens. Actuators B 2000, 64, 156–162. Li, Y.-S.; Liu, W.-P.; Gao, X.-F.; Chen, D.-D.; Li, W.G. Immobilized enzymatic fluorescence capillary biosensor for determination of sulfated bile acid in urine. Biosens. Bioelectron. 2008, 24, 538–544. Wang, L.; Sipe, D.M.; Xu, Y.; Lin, Q. A MEMS thermal biosensor for metabolic monitoring applications. J. Microelectromech. Syst. 2008, 17, 318–327. De Boer, A.R.; Bruyneel, B.; Krabbe, J.G.; Lingeman, H.; Niessen, W.M.A.; Irth, H. A microfluidic-based enzymatic assay for bioactivity screening combined with capillary liquid chromatography and mass spectrometry. Lap Chip 2005, 5, 1286–1292. Tang, Z.-M.; Wang, Z.-Y.; Kang, J.-W. Screening of acetylcholinesterase inhibitors in natural extracts by CE with electrophoretically mediated microanalysis technique. Electrophoresis 2007, 28, 360–365. Tang, Z.-M.; Wang, T.-D.; Kang, J.-W. Immobilized capillary enzyme reactor based on layer-by-layer assembling acetylcholinesterase for inhibitor screening by CE. Electrophoresis 2007, 28, 2981–2987.

REFERENCES

185

189. Besanger, T.R.; Hodgson, R.J.; Green, J.R.A.; Brennan, J.D. Immobilized enzyme reactor chromatography: optimization of protein retention and enzyme activity in monolithic silica stationary phases. Anal. Chim. Acta 2006, 564, 106–115. 190. Nicoli, R.; Bartolini, M.; Rudaz, S.; Andrisano, V.; Veuthey, J.-L. Development of immobilized enzyme reactors based on human recombinant cytochrome P450 enzymes for phase I drug metabolism studies. J. Chromatogr. A 2008, 1206, 2–10. 191. Kaneno, J.; Kohama, R.; Miyazaki, M.; Uehara, M.; Kanno, K.; Fujii, M.; et al. Development of surface modification method and its application for preparation of enzyme-immobilized microreactor. Kagaku Kogaku Ronbunshu 2004, 30, 154–158. 192. Kamal, A.; Sandbhor, M.; Ramana, K.V. One-pot lipase-catalyzed synthesis of enantiopure secondary alcohols from carbonyl compounds: a new protocol for lipase-mediated resolution. Tetrahedron: Asymmetry 2002, 13, 815–820. 193. Seong, G.H.; Heo, J.; Crooks, R.M. Measurement of enzyme kinetics using a continuousflow microfluidic system. Anal. Chem. 2003, 75, 3161–3167. 194. Moore, B.D.; Stevenson, L.; Watt, A.; Flitsch, S.; Turner, N.J.; Cassidy, C.; et al. Rapid and ultra-sensitive determination of enzyme activities using surface-enhanced resonance Raman scattering. Nat. Biotechnol. 2004, 22, 1133–1138. 195. Lilly, M.D.; Hornby, W.E.; Crook, E.M. Kinetics of carboxymethylcellulose ficin in packed beds. Biochem. J. 1966, 100, 718–723. 196. Jiang, H.; Zou, H.; Wang, H.; Ni, J.; Zhang, Q.; Zhang, Y. On-line characterization of the activity and reaction kinetics of immobilized enzyme by high-performance frontal analysis. J. Chromatogr. A 2000a, 903, 77–84. 197. Ristenpart, W.D.; Wan, J.; Stone, H.A. Enzymatic reactions in microfluidic devices: Michaelis–Menten kinetics. Anal. Chem. 2008, 80, 3270–3276. 198. Koh, W.-G.; Pishko, M. Immobilization of multi-enzyme microreactors inside microfluidic devices. Sens. Actuators. B 2005, 106, 335–342. 199. Nicoli, R.; Gaud, N.; Stella, C.; Rudaz, S.; Veuthey, J.-L. Trypsin immobilization on three monolithic disks for on-line protein digestion. J. Pharm. Biomed. Anal. 2008, 48, 398–407. 200. Nicoli, R.; Rudaz, S.; Stella, C.; Veuthey, J.-L. Trypsin immobilization on an ethylenediamine-based monolithic minidisk for rapid on-line peptide mass fingerprinting studies. J. Chromatogr. A 2009, 1216, 2695–2699.

5 MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY CLEMENT KLEINSTREUER Department of Mechanical and Aerospace Engineering and Department of Biomedical Engineering, North Carolina State University, Raleigh, NC, USA

JIE LI Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC, USA

5.1 INTRODUCTION Nanodrug delivery, employing microscale devices, is a broad and complex research topic with several major application areas. For example, cost-effective drug discovery, development, and testing are of great concern to the pharmaceutical industry, while clinical diagnostics and drug delivery, ideally in combined form, are of interest to healthcare providers. The associated microfluidic devices include lab-on-a-chip (LOC) systems for drug discovery/development and bio-MEMS (biological/biomedical microelectromechanical system) for controlled biological processing and optimal (nano-) drug delivery. Powered by microfluidics, the use of LOC devices can be a robust and fast method to discover, refine, and test a drug. This is important in light of the fact that presently only one-tenth of the drug compounds that enter the clinical trial phase succeed in becoming commercially available (see 03/31/07 Report at BioMarket Research.com). Bio-MEMSs are being used for controlled biological processes, such as cell sorting and multinodal bioimaging/identification, as well as for targeted drug delivery. The

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

187

188

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

latter entails biochemical or mechanical methodologies, that is, either a passive mode or different active delivery modes.1 1. Passive multifunctional nanoparticle systems (MFNPSs) include injected porous (micrometer) particles carrying nanodrugs and releasing them near/at the desired site. 2. Active nanodrug carriers (NDCs) of engineered size, shape, and surface characteristics circulate in the bloodstream and may actively attach to diseased cells/tissues. 3. Active (mechanical) drug delivery systems (DDSs) concentrate on 100% targeting methodologies, nanofluid flow in microchannels, nanodrug mixing, and microdevice optimization. This chapter focuses on both biochemical and mechanical drug delivery systems with an emphasis on experimental/computational simulation aspects of bio-MEMSs, where some function as implanted microfluidic devices for controlled nanodrug release. In any case, the overriding optimization objectives include biocompatibility, controlled nanodrug release, performance accuracy and reliability, minimization of side effects, size reduction, and cost-effectiveness. Previous reviews concentrated on particular topics. For example, Suh et al.2 discussed biological MFNPSs in nanotechnology, stressing nanotoxicity concerns and nanodrug applications to neuroscience, while Emerich and Thanos3 outlined the potential of nanomedicine enabling targeted delivery of diagnostic and therapeutic agents, and Kim et al.4 provided a past-to-future overview on nanotechnology in drug delivery. Riehemann et al.5 outlined recent developments and applications in nanomedicine. Parallel reviews on nanodrug (and gene) delivery systems are treatment specific toward particular diseases or organs. For example, Kasuya and Kuroda6 summarized nanomedicine for the human liver and Subramani7 considered nanodrug treatment applications for cancer and diabetes, while Kleinstreuer et al.8 reviewed targeted delivery of inhaled drug aerosols to predetermined sites to combat lung tumors or even systemic diseases, outlining the underlying methodology of a smart inhaler system.

5.2 MICROFLUIDIC DEVICES To appreciate the mechanics of microfluidic devices as well as ongoing modeling and simulation aspects, this section starts out reviewing a few basic elements of microfluidics and microsystems as well as their modeling assumptions. This brief discourse is especially useful for readers interested in a state-of-the-art sample application given in Section 5.4. The main focus of microscale research and development is on device fabrication and expansion of microsystem application areas, which implies innovative advances in the material sciences, manufacturing technology, as well as supportive design software creation. Electromechanical components of consumer goods, vehicles, and machinery, as well as entire devices, especially medical implants and laboratory test

MICROFLUIDIC DEVICES

189

equipment, are being built on a microscale. Examples include MEMS, microheat sinks, iPods, and appliance control parts, as well as sensors and drug-release patches in medicine, or lab-on-a-chip units and reactors in biomedical and chemical engineering. Clearly, it is the low production cost, compactness with a very high surface-to-volume ratio, rapid throughput with very small sample volumes, and integrated multifunctionality, for example, nanodrug mixing or particle separation or stream positioning, that make microscale fluid devices attractive alternatives to conventional flow systems.9 5.2.1 Microfluidics and Microsystems Microfluidics is the study of transport processes in microchannels. Of interest are methods and devices for controlling and manipulating fluid flow, finite liquid volume delivery, and particle transport on a nano- and microscale. Although microfluidics deals with fluid behavior in systems with “small” length scales, conventional (i.e., macroscale) flow theory is typically applied, at least for liquid flows in microchannels with Dhydraulic
10 mm and standard gas flows when Dh
100 mm. However, for microchannel gas flows in the slip regime, that is, 0:001  Kn ¼ l=L  0:1 (where the Knudsen number is the ratio of the molecular mean free path over a system length scale), modification to the velocity and temperature boundary conditions has to be made. Clearly, when the Knudsen number is above 0.1, alternative system equations and numerical solution techniques have to be considered. Microfluidic devices (or microsystems, or bio-MEMS) typically consist of reservoirs, channels, actuators, pumps, valves, mixers, sensors, controllers, filters, and/or heat exchangers. Associated with microfluidic devices are the following R&D areas: . . . . . . .

Microfabrication of components or entire devices, using silicon, glass, polymer, or steel Microfluidic transport phenomena, including mechanical micropumps as well as nonmechanical surface effects Task-specific devices, such as micrototal analysis systems (mTAS), LOC, or DDSs Reliable detection and measuring systems Power systems and microdevice packaging Data communication, including telemetry for monitoring system performance Biocompatibility and adherence to regulations

Bio-MEMSs for drug delivery are of interest in this chapter, where we focus on nanodrug transport phenomena in microchannels. Such devices offer a number of advantages, such as controlled drug release, reliable accurate dosing, targeted treatment, and automated feedback control, all resulting in small size and operational convenience, efficacy, and cost-effectiveness. Basic background information, including microscale device manufacturing methods, may be found in the books by Tabeling,10 Nguyen and Wereley,11 Saliterman,12 and Tesar.13 Reviews of engineering

190

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

flows in small devices have been provided by Stone et al.,14 Hilt and Peppas,15 Whitesides,16 Hu and Zengerle,17 and Geipel et al.,18 to name a few. 5.2.2 Microsystem Modeling Assumptions One of the key elements of all microsystems is the microchannel (soon becoming a nanochannel) with hydraulic diameters, that is, circular-tube-equivalent diameters, typically ranging from 10 to 500 mm. This is rather small in light of the fact that the diameter of the human hair is about 80 mm. When considering fluid flow in such tiny conduits, we should recall that the underlying macroscale modeling assumptions are valid only when 1. the fluid is “infinitely divisible,” that is, the fluid forms a continuum, and hence we can use the conservation laws in terms of the continuity, momentum, and heat/mass transfer equations, often summarized for all practical purposes as the Navier–Stokes equations; 2. all flow quantities are in local thermodynamic equilibrium, that is, no velocity or temperature jumps at fluid–wall interfaces. Concerning the continuum assumption (1), the two main classes of fluids, that is, gases (in case of nanospray delivery) and liquids (primarily nanodrugs in aqueous solutions) differ primarily by their densities and by the degrees of interaction between the constituent molecules. Focusing on aqueous solutions, water density is typically rliquid  103 kg m3 with an intermolecular distance lIM ¼ 0:3 nm. Now, if the key macroscopic length scale, for example, microchannel effective diameter (or height or width), is of the order of 10 mm or more, fluids with those characteristics appear continuous and hence the Navier–Stokes equations hold. The local thermodynamic equilibrium condition (2) implies that all macroscopic quantities within the fluid have sufficient time to adjust to their surroundings. That process depends on the time between molecular collisions and hence the magnitude of the mean free path traveled. Clearly, a rarefied gas in a small microchannel does not form a continuum and hence would exhibit velocity and temperature jumps at the channel walls, requiring more exotic solution methods, for example, the lattice Boltzmann method (LBM), direct simulation Monte Carlo (DSMC), or molecular dynamics simulation (MDS). The conventional driving force for flow in microchannels is still the net pressure force, using micropumps, when substantial flow rates, that is, Re ¼ vh=u > 1:0, are desired, as for rapid nanodrug mixing and delivery. However, certain microfluidic devices for biomedical, chemical, and pharmaceutical applications employ more esoteric driving forces, such as surface tension (i.e., capillary or Marangoni effects) and electrokinetic phenomena (i.e., electrophoresis or electroosmosis). In general, the surface-to-volume ratio varies as the inverse of the system’s length scale, that is, 1/L, and hence microsystems with relatively large surface areas may cause significant viscous resistance. In turn, it would require relatively powerful actuators, including pumps, valves, and so on, to operate a microfluidic device. In order to have such pumps/actuators/valves as integral parts of the microfluidic device, new principles

MICROFLUIDIC DEVICES

191

had to be employed. Thus, complementary to mechanical actuators with moving parts, microscale phenomena were used when the inlet Reynolds number was low (Re  O(1)), such as electrokinetic pumping (e.g., electroosmosis) and capillary surface tension effects, electromagnetic force fields, and acoustic streaming. Another contrasting macroscale versus submicrometer scale consideration is that conventional fluid flow is described by velocity and pressure fields and by its properties. Hence, they are characterized as interacting groups, such as kinematic (i.e., velocity and strain rate), thermodynamic (i.e., pressure and temperature), transport (i.e., viscosity, conductivity, and diffusivity), and miscellaneous parameters (i.e., surface tension, vapor pressure, etc.). However, on the submicrometer scale, matter, that is, solid, liquid, or gaseous, is more realistically described in terms of interacting molecules. For example, molecules in a solid are densely packed and arranged in a lattice, where each molecule is held in place by large repulsive forces according to the Lennard-Jones (L-J) potential.19 Nevertheless, when solving problems of fluid flow in microchannels, the continuum mechanics assumption is preferred over any molecular approach. For the latter approach, the state of each molecule in terms of position and velocity has to be known, and then one has to evolve/simulate that state forward in time for each molecule. That implies the solution of Newton’s second law of motion with the L-J force (i.e., the spatial derivative of the L-J potential) for billions of molecules. In contrast, when continuous fluid flow behavior can be assumed, that is, system length scales Lgas > 100 mm (or Kn  0.1) and Lliquid > 10 mm, we just numerically solve the conservation laws subject to key assumptions and appropriate boundary conditions, as exemplified in Section 5.4.1. In summary, it is not surprising that fluid flow in microchannels may differ from macrochannel flow behavior in terms of entrance, wall, and thermal flow effects.20 Specifically, because of the typically short microchannel length, entrance effects (i.e., developing 2D or 3D flows) may be dominant. At the microchannel wall, the “noslip” conditions may not hold for hydrophobic liquids, electrokinetic forces may come into play, and surface roughness effects may be substantial. Early onset of laminarto-turbulent flow transition may occur and viscous dissipation of heavy liquids in high shear rate fields may increase the fluid temperature measurably. 5.2.3 Categories of Microfluidic Devices The development and use of microfluidic devices, including bio-MEMS (see Figure 5.1), are naturally being driven by application areas, that is, drug delivery routes and targets for specific clinical treatment needs, and ultimately by business interests. For successful treatment, rapid administration of the right dosage of medication is important: too low a dosage may be ineffective and too high may be harmful. Furthermore, dose frequency and duration, drug toxicity and interaction, as well as allergies must all be considered on a patient- and disease-specific basis. A smart drug delivery system (SDDS) connects a patient, that is, the specific disease site, with an appropriate drug. An SDDS is a formulation (or device) with which nanomedicine is introduced into the body, released at a controlled rate, and

192

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

FIGURE 5.1 Bio-MEMS components and flow chart.

subsensitively transported across cell membranes for therapeutic action with minimal side effects. The two most common delivery routes are oral, that is, drugs taken by mouth and swallowed and hence a device is not needed, and via injection (i.e., parenteral administration) directly into the bloodstream or affected area. More modern routes include targeted drug aerosol inhalation for lung, sinus, and systemic diseases employing smart inhaler systems, transdermal delivery via microneedles, and body implanted microfluidic devices with controlled nanomedicine release. The oral route is the typical way to deliver drugs into the body because it is cheap and most convenient. However, it may not be very efficient because of drug absorption and/ or degradation before it reaches the bloodstream and ultimately the affected area or organ. While injection is effective for relatively high quantities of large-molecule drugs, it is also inconvenient, somewhat expensive, and not easy to control. Hence, with the advent of nanodrugs and gene therapy, new delivery devices, including bioMEMS, had to be developed. Key components of such systems include microchannels, micropumps (i.e., mechanical and electrokinetic), microvalves, microreservoirs, and micromixers.21–25 Application-driven drug delivery devices can be categorized into several groups—for example, microneedles providing active medicine infusion

MICROFLUIDIC DEVICES

193

through transdermal patches; drug-eluting stents maintaining patency of, say, coronary arteries; smart inhalers, with which drug aerosol streams (nasal sprays) are controlled, to improve targeted particle deposition; and self-contained external or implantable bio-MEMS. Possibly, the largest group directly benefitting from bio-MEMS is diabetic patients ensuring exact glucose control via embedded monitors and insulin dispensers. Patients with pacemakers or defibrillators may receive needed medication from an implanted bio-MEMS during an arrhythmia event. Severe asthma patients typically requiring several drugs, such as bronchodilators and anti-inflammatory medicine, may rely on real-time disease analysis and subsequently controlled, targeted drug release. Pain management can be handled by a programmable pump for, say, small-dose intraspinal morphine administration to block neurotransmitters from reaching the brain. Clearly, a lot of R&D work and clinical testing have to be accomplished before most conceived bio-MEMS gains public acceptance. The book edited by Desai and Bhatia26 discusses bio-MEMS for drug delivery in several chapters, while the review by Elman et al.27 briefly summarizes next-generation bio-MEMS, which the authors classified as passive or active delivery devices. Microneedles, made out of silicon, polymer, steel, or metal oxides, are only a few hundred micrometers long; that is, they generate microconduits past the outer skin permeation barrier without encountering a nerve. In array formation, connected to a liquid drug reservoir, they allow for dispersion and systemic uptake of macromolecular drugs, possibly replacing hypodermic injections, say, for vaccinations and insulin delivery.28–30 Drug-eluting stents are slow-release nanodrug implants that mainly function as scaffolds to keep arteries open after coronary angioplasty, reduce the likelihood of restenosis, and reject foreign object. After initially very positive response worldwide, they have recently encountered mixed reviews because of postoperative complications, such as late stent thrombosis in some patients.31–35 Micropumps are vital for direct drug delivery when connected to a microreservoir, or for nanodrug mixing in microchannels (see Section 5.4.1). Other micropump applications include movement of nano- to microliter solutions in LOC and mTAS devices, molecular particle sorting with microfilters or via hydrodynamic focusing, and flow measurements with microsensors. However, one should note that a majority of hydrodynamic microscale and certainly most nanoscale processes are driven by electrokinetic flow or surface-mediated transport.36 Thus, due to the very high frictional resistance, pressure-driven flow in a nanofluidic device is inappropriate. The reason is that Dp  L=D4h , where Dp is the pressure drop across the conduit of length L and hydraulic diameter Dh ¼ 4A/P, with A being the cross-sectional area and P the wetted perimeter. For example, Zahn37 reviewed the physics and fabrication of mechanical and nonmechanical micropumps. Hundreds of microreservoirs can be embedded into a single silicon microchip that is covered by a thin metal or polymer membrane. The microreservoirs may contain any combination of drugs, chemicals, and/or biosensors, where the membrane seal can be activated for controlled drug release, using preprogrammed microprocessors, wireless telemetry, or biosensor feedback. Clearly, these microchips can store and release

194

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

nanodrugs in a controlled fashion, and they are advantageous because of their small size, low power consumption, and absence of moving parts.

5.3 NANODRUG DELIVERY It is apparent from the previous sections that drug delivery systems based on bioMEMS are just beginning to reach the market. Self-regulated insulin delivery systems, drug-eluting stents, and microneedle arrays with reservoirs on a chip are some of the more mature examples. In this section (Figure 5.2), nanodrug carriers for biochemical drug/gene delivery systems as well as associated clinical application areas and mechanical nanodrug delivery methodologies are discussed. 5.3.1 Nanodrugs Nanodrugs (or genetic material) embedded in nano/microspheres are promising candidates for treatment of various diseases, such as cancer, infections, metabolic and autoimmune diseases, and diseases related to the brain (http://nano.cancer.gov).

FIGURE 5.2

Strategies of targeted nanodrug delivery systems.

NANODRUG DELIVERY

195

Such nanomedicine carriers can be microparticles made of soluble, insoluble, or naturally biodegradable polymers, microcapsules, porous particles, cells, liposomes, and so on. Emerich and Thanos3 provided an overview of typical nanoparticles used in drug and gene delivery, while Kasuya and Kuroda6 summarized the desirable properties and characteristics of nanomedicine carriers. Their modified and updated lists are given below. 5.3.1.1 Solid Nanoparticles Ceramic nanoparticles are made from inorganic nonmetallic compounds with porous characteristics such as oxides, that is, silica (SiO2), alumina (Al2O3), hydroxyapatite (HA), and zirconia (ZrO2). They are stable in the typical range of temperatures and pH encountered in the body and can be used to deliver proteins and genes. However, their lack of biodegradation and slow dissolution raises safety questions. For example, it was found that those made of silica can efficiently transport therapeutic genes to the spleen and trigger a potent immune response capable of attacking tumors.38 The results released in 2008 in Chemical & Engineering News (http://pubs.acs.org/ isubscribe/journals/cen/86/i35/html/8635scic.html#6) showed that iron oxide nanoparticles caused little DNA damage and were nontoxic, zinc oxide nanoparticles were slightly worse, and titanium dioxide caused only DNA damage. Carbon nanotubes (CNTs) are extremely small tubes that can be categorized as single-walled nanotubes (SWNTs) and multiwalled nanotubes (MWNTs). These compounds have become increasingly popular in various fields simply because of their small size and amazing optical, electric, and magnetic properties when used alone or with other materials, such as drugs. Carbon nanotubes have potential therapeutic applications in the field of drug delivery, diagnostics, and biosensing. For example, SWNTs have been shown to shuttle various cargos across cellular membrane without cytotoxicity. SWNTs can be used as a platform for investigating surface–protein and protein–protein binding. These nanotubes can act as highly specific electronic sensors for detecting clinically important biomolecules such as antibodies associated with human autoimmune diseases. Functionalized carbon nanotubes can also act as vaccine delivery systems. The basic concept is to link the antigen to carbon nanotubes while retaining its conformation, thereby inducing antibody response with the right specificity. Overall, the future use of carbon nanotubes in drug delivery systems may enhance detection sensitivity in medical imaging, improve therapeutic effectiveness, and decrease side effects. Nanocrystals are aggregates of molecules that can be combined in a crystalline form of the drug surrounded by a thin surfactant coating. High dosages can be achieved and poorly soluble drugs can be formulated for improved bioavailability. Both oral and parenteral delivery systems are possible and the limited carrier in the formulation reduces potential toxicity. Limitations include poor drug stability. Polymers such as albumin, chitosan, and heparin occur naturally and have been a material of choice for the delivery of oligonucleotides, DNA, protein, and drugs. The drug is physically entrapped in the polymer capsule. The characteristics can be summarized as follows: (i) water soluble, nontoxic, and biodegradable; (ii) surface modification (pegylation); (iii) selective accumulation and retention in tumor tissue;

196

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

and (iv) specific targeting of cancer cells while sparing receptor-mediated targeting of normal cells with a ligand. Polymer nanostructured fibers, core–shell fibers, hollow fibers, and nanorods and nanotubes provide a platform for a broad range of applications. For example, biological objects, including drugs, of different complexities carrying specific functions can be incorporated into such nanostructured polymer systems. Biosensors, tissue engineering, drug delivery, and enzymatic catalysis are just a few applications. Another example is superparamagnetic particles, known to display strong interactions with external magnetic fields leading to large saturation magnetization. By using periodically varying magnetic fields, the nanoparticles can be heated to provide a trigger for drug release. Solid lipid nanoparticles are lipid-based submicron colloidal carriers. They have a solid hydrophobic core surrounded by a monolayer of phospholipid. The system is stabilized by the inclusion of fairly high levels of surfactants. They are less toxic than polymer nanoparticles and can be used to deliver drugs orally, topically, or via the pulmonary route. While stability is a concern, it is better than that observed with liposomes. 5.3.1.2 Colloidal Soft Matter Dendrimers are artificial, polymer-based molecules formed from monomers such that each layer of branching units doubles or triples the number of peripheral groups (i.e., they look like a foam ball). The void area within a dendrimer, its ease of modification/ preparation, and size control offer great potential for targeted gene and drug delivery. Improvements in cytotoxicity profiles, biocompatibility, and biodistribution are needed. Dendrimers are repeatedly branched molecules. They are emerging as a rather new class of polymeric nanosystems with applications in drug delivery. The properties of dendrimers are dominated by the functional groups on the molecular surface. Dendritic encapsulation of functional molecules allows for the isolation of the active site, a structure that mimics the structure of active sites in biomaterials because dendritic scaffolds separate internal and external functions. For example, a dendrimer can be water soluble when its end group is a hydrophilic group, like a carboxyl group. It is theoretically possible to design a water-soluble dendrimer with internal hydrophobicity, which would allow it to carry a hydrophobic drug in its interior. Another property is that the volume of a dendrimer increases when it has a positive charge. If this property can be applied, dendrimers can be used for drug delivery systems that can give medication to the affected part inside a patient’s body directly. Hydrogels (also called aquagels) are a network of polymer chains that are water insoluble, and sometimes found as a colloidal gel in which water is the dispersion medium. Hydrogels are superabsorbent (they can contain over 99% water) natural or synthetic polymers. Hydrogels also possess a degree of flexibility very similar to natural tissue due to their significant water content. Hydrogels are responsive to specific molecules, such as glucose or antigens, that can be used as biosensors as well as in drug delivery system. Liposomes were tiny bubbles (vesicles) made out of the same material as a cell membrane. Liposomes are small spherical systems that are synthesized from

NANODRUG DELIVERY

197

cholesterol and nontoxic phospholipids. Liposomes can be filled with drugs and used to deliver drugs for cancer and other diseases. Membranes are usually made of phospholipids, which are molecules having a head group and a tail group. The head is attracted to water and the tail, which is made of a long hydrocarbon chain, is repelled by water. Because they are natural materials, liposomes are considered attractive, harmless drug delivery carriers that can circulate in the bloodstream for a long time. Another interesting property of liposomes is their natural ability to target cancer. The endothelial wall of all healthy human blood vessels is encapsulated by endothelial cells that are bound together by tight junctions. These tight junctions stop any large particle in the blood from leaking out of the vessel. Tumor vessels do not contain the same level of seal between cells and are diagnostically leaky. This ability is known as the enhanced permeability and retention effect. Liposomes of certain sizes, typically less than 400 nm, can rapidly enter tumor sites from the blood but are kept in the bloodstream by the endothelial wall in healthy tissue vasculature. Anticancer drugs such as doxorubicin (Doxil), camptothecin, and daunorubicin (DaunoXome) are currently being marketed in liposome delivery systems. Despite a relatively long history of investigation, liposomes exhibit limited stability and have not made significant medical impact. Micelles provide considerable advantages among drug carrier systems for their solubilization to contribute the increasing bioavailability of poorly soluble drugs and the characteristics to stay in the blood long enough to afford a gradual accumulation in a particular area. A micelle is an aggregate of surfactant molecules dispersed in a liquid colloid. A typical micelle in aqueous solution forms an aggregate with the hydrophilic “head” regions in contact with surrounding solvent, sequestering the hydrophobic tail regions in the micelle center. This type of micelle is known as a normal phase micelle (oil in water micelle). Inverse micelles have the head groups at the center with the tails extending out (water in oil micelle). Micelles are approximately spherical in shape. Other phases including shapes such as ellipsoids, cylinders, and bilayers are also possible. Micelle formation is essential for the absorption of fat-soluble vitamins and complicated lipids within the human body. When membrane phospholipids are disrupted, they can reassemble themselves into tiny spheres, smaller than a normal cell, either as bilayers or as monolayers. The bilayer structures are liposomes and the monolayer structures are called micelles. Microemulsions have great thermodynamic stability, which allows for self-emulsification at a wide range of temperatures and affords easy preparation. The structural variability of microemulsions together with its composition and the pH of the environment can obviously influence the drug release rate. Other positive characteristics include the low viscosity of the majority of the system. Organogels are systems resembling the structure of microemulsions but are semisolid. Most organogels utilized in pharmaceuticals are lecithin, gelatin, or sorbitan ester-based systems in biocompatible solvents. Complex aqueous phases such as vesicle suspensions entrapping drugs can be eventually incorporated into the organogel systems. They can entrap hydrophilic or hydrophobic drugs and antigens; it is possible to achieve controlled release systems.

198

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

5.3.1.3 Desirable Characteristics of Nanodrug Carriers With this large family of nanodrugs available (see Sections 5.3.1.1 and 5.3.1.2), it is important to consider suitable (or even optimal) attributes of nanodrugs and their carriers. Specifically, the following properties should be optimized for the in vivo use of conventional nanoparticle carriers for drug or gene delivery. 1. Acceptability to Versatile Payloads: The nanoparticle carrier (NPC) should allow payloads of various therapeutic materials. This facilitates the development of a general purpose carrier and concurrent administration of different drugs. 2. Low or No Toxicity: The NPC should not be made of or contain potentially dangerous materials. 3. Active Targeting: The NPC should recognize and attach target cells and tissues by a sensor molecule displayed on its surface. NPC of about 100 nm are likely to accumulate spontaneously in tumors after systemic injection, but this passive targeting mechanism based on the enhanced permeability and retention effect39 is excluded from these criteria. 4. Appropriate Size: The size of NPC should be about 40–150 nm in diameter. Too small nanoparticles (<40 nm) and too large nanoparticles (>150 nm) are nonspecifically removed from blood circulation by the function of kidney and reticuloendothelial system in the liver, respectively.40 NPC of about 40–150 nm could be used to target both tumors utilizing the enhanced permeability and retention effect and hepatocytes by passing through the fenestrae in liver endothelial cell.41 5. Appropriate Surface Charge: The surface charge of NPC is known to affect severely the stability and biodistribution of systemically administrated nanoparticle carriers. For ideal delivery, the surface of nanoparticle carriers should be optimized so as not to be entrapped by unexpected tissues. For example, one positively charged nanoparticle (i.e., polyplex of polyethyleneimine and DNA) was efficiently accumulated in the lung after systemic administration.42 6. Efficient Cell-Penetrating Activity: The nanoparticle carriers should possess cell-penetrating activity for active and rapid intrusion across the plasma membrane or the endosomal membrane of target cells and tissues because many therapeutic materials (particularly, genes and siRNAs) function intracellularly. 7. Mechanism of Intracellular Targeting: The nanoparticle carriers in target cells should bring and release payloads to the intracellular destination precisely (e.g., genes and siRNAs should be released in the nucleus and cytoplasm, respectively, not in endosomes). 5.3.2 Nanodrug Targeting As mentioned, targeting is the ability to direct in controlled fashion the drug particles, or a drug-loaded system, to the predetermined site of interest. There are mechanical and biochemical methodologies of drug targeting (Figures 5.2 and 5.3a and b). As Kaparissides et al.43 mentioned, in biochemical targeting two approaches can be

NANODRUG DELIVERY

199

FIGURE 5.3 (a) Passive drug targeting by nanodrug carriers in blood vessels and (b) active nanodrug targeting with conjugated antibodies.

distinguished, that is, passive and active targeting. They give as an example for passive targeting the observed preferential accumulation of chemotherapeutic agents in “solid” tumors as a result of the enhanced vascular permeability of tumor tissues compared to healthy tissue (Figure 5.3a). As a variation to nanodrugs cruising in the bloodstream and hopefully reaching the right site, drug carriers with surface functionalities, for example, ligands interacting with tumor cell receptors, can seek out, bind to, and penetrate target cells (Figure 5.3b). In contrast, mechanical drug targeting is the delivery of a controlled fluid–particle stream from an optimal release (or arterial injection) point to a predetermined deposition site.44,45 Specifically, Kleinstreuer45

200

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

discussed several mechanical drug delivery applications within the framework of fluid–structure interaction simulations. For example, targeted drug aerosol delivery can be potentially accomplished with a smart inhaler system.46 Another application is the targeting of liver tumors with radioactive microspheres,47 where the particles are released, via a microcatheter, from an optimal inlet position of the hepatic artery. Most of the drug targeting literature deals with biochemical targeting, especially “colloidal soft matter” (see Section 5.3.1.2) as reviewed by Bonacucina et al.48 Such “soft matter,” for example, microemulsions and organogels (as passive drug carriers) as well as liposomes, micelles, and dendrimers (as active drug carriers), can increase drug solubility and bioavailability and can attach/penetrate tumor cells, especially where micelles are used. Clearly, all present biochemical targeting efforts are disease/ treatment specific.1,49–54 Drug targeting via mechanical delivery devices include reservoirs, pumps, inhalers, catheters, needles, drug-eluting stents, and implants releasing drugs (Section 5.2.3). The next two sections provide some physical insight into the microfluidics of bioMEMSs, focusing on nanodrug mixing, microchannel flow, and device optimization. 5.4 Bio-MEMS APPLICATIONS To illustrate some aspects of bio-MEMS research and development, nanodrug mixing and microchannel designs are discussed in the next two sections. 5.4.1 Nanofluid Flow Simulations Microfluidics deals with methods and devices for manipulating and controlling fluid–particle flow in microchannels.55 A recent application area is nanomedicine with the goal of controlled nanodrug delivery to specified target areas.12,56 A key aspect of this goal is the development of integrated drug delivery systems to monitor and control target cell responses to pharmaceutical stimuli, to understand biological cell activities, or to facilitate drug development processes. An important part of such drug delivery systems, belonging to the family of bio-MEMS, is active or passive micromixers25,57 to assure near-uniform nanodrug concentrations. Static micromixers, not requiring any external energy source, rely on chaotic advection and/or enhanced diffusion, typically to mix two fluids.58,66 For example, Hardt et al.61 reviewed recent developments in micromixing technology, focusing on liquid mixing with passive micromixers. Four kinds of mixers that employed different hydrodynamic principles are discussed: hydrodynamic focusing, flow separation, chaotic advection, and split and recombine flows. Diffusive mixing can be improved by increasing the interfacial contact area between the different fluids and reducing the diffusion length scale. Thus, selecting the right type of micromixer for a specific application is very important. Li and Kleinstreuer67 analyzed rapid nanoparticle mixing in a carrier fluid, employing low-cost micromixers and heat transfer to achieve two system design goals, that is, uniform exit particle concentration and minimum required channel length. Specifically, a microfluidics device for controlled nanofluid flow in

Bio-MEMS APPLICATIONS

201

microchannels (Figure 5.4) is investigated for basic nanomedicine applications. Presently planned for laboratory-scale testing, uniform, predetermined concentrations of a stimulus (e.g., cocaine particles) should be delivered via multiple microchannels to an array of wells containing brain cells to measure cell responses (e.g., dopamine production levels). Their study focuses on device miniaturization in light of the ultimate goal of bio-MEMS implantation into the diseased brain region of, say, Parkinson’s patients. Most important, the impact of two types of static micromixers (Figure 5.4c) is analyzed to achieve uniform nanoparticle concentrations at the exit of a representative microchannel of minimum length. Figure 5.5 shows Lmin(Pe) for the different scenarios. Clearly, any micromixer module reduces Lmin significantly for all P
eclet numbers. While an increase in slotted baffle plates reduces Lmin, the simple three-sided injection unit performs best.

FIGURE 5.4 Microfluidics system: (a) laboratory-scale nanodrug supply device, (b) representative microchannels, and (c) static mixer inserts.

202

FIGURE 5.5

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

Micromixer influence on minimal uniformity length/system dimension.67

Alternative to the P
eclet number, which is based on the average velocity of nanofluid plus carrier fluid, the Reynolds number ratio of nanofluid to carrier fluid is a suitable operational parameter. The associated Reynolds numbers are Rei ¼

ðuDh Þi ni

ð5:1Þ

with i ¼ 1 indicating the carrier fluid in the solution channel and i ¼ 2 denoting the nanofluid channel (Figure 5.4b). For the given system, Figure 5.6 indicates that the main microchannel length can be below 4 mm when employing an injection micromixer, that is, a 70% reduction in channel length. The addition of nanoparticles and certainly the installment of micromixers increases the pressure drop in both channels and hence the pumping power requirements. Pumping power is defined as the product of the pressure drop across the channel (Dp) and the volumetric flow rate (Q), that is, P ¼ Dp  Q

ð5:2Þ

The pressure differences occur between the nanofluid inlet or carrier fluid inlet and the system outlet, that is, required minimal length. The volumetric flow rate is the sum of the volumetric flow rates of both nanofluid and carrier fluid. Figure 5.7a and b depicts the relationship of pressure drop and pumping power for the two cases. Clearly, the added micromixer increases the local pressure drop, but the decreased system length may reduce any negative effect caused by the micromixer. As shown in Figure 5.8a and b, the power requirement even decreases in some cases,

203

Bio-MEMS APPLICATIONS

FIGURE 5.6

Minimal uniformity length versus Reynolds number ratio.67

that is, when employing the injection micromixer. The employment of baffle-slit micromixers slightly increases the pressure drop; however, when the pumping power/ volumetric flow rate gets larger and larger, the negative effect appears to be less and less. For example, for the two- or four-baffle micromixer, the pressure drop is even smaller than that without any micromixer when the nanofluid supply rate is increased to 8 mm s1. In summary, employing an appropriate micromixer decreases the system dimension and the associated power requirement. A heat flux was used to ensure that mixture delivery to the living cells occurs at a required temperature of 37  C. The change of fluid properties and nanoparticle diffusivity, caused by the added heat flux, also benefits system miniaturization. As shown in Figure 5.6, the added heat flux greatly decreases the system dimension; that is, an average 35% reduction is observed. 5.4.2 Device Optimization It is obvious that the better an engineering device performs, the lower the (irreversible) losses, that is, the closer it operates at an isentropic efficiency. This directly implies that S_ gen should be minimized as part of any device/process design or improvement. For example, considering simple heat transfer from an ambient reservoir at T0 ¼ ¢, the entropy balance equation can be written as @S  S_ gen ¼ @t

X Q_ T0



X

_ in þ mSj

X

_ out > 0 mSj

ð5:3Þ

204

FIGURE 5.7 inlet.67

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

Pressure drop versus pumping power: (a) carrier fluid inlet and (b) nanofluid

205

Bio-MEMS APPLICATIONS

FIGURE 5.8

Heat flux influence on minimal uniformity length.67

_ loss ; that is, in general, power loss is due to system, heat Effectively, S_ gen  W transfer, and fluid flow irreversibilities: Ploss


 X @S Q_ X _ _ _ in  _ out    W loss ¼ T0 Sgen ¼ T0 mSj mSj @t T0

ð5:4Þ

Focusing on friction (or viscous effects) as the main cause of irreversibilities and hence entropy generation, the rate of irreversible conversion from flow energy to heat can be expressed as tij

S_ gen @vi Ploss ¼ T0  mF ¼ @xj 8 8

ð5:5Þ

Clearly, to minimize entropy production in pressure-driven microchannels, we have to reduce the viscous dissipation function F, that is, achieve minimization of "
2
2
2 # S_ gen m m @u @v @u @v þ2 þ2 ¼ F¼ þ 8 T T @y @x @x @y

ð5:6Þ

In general, and exclusively for fully developed flow, the term ðm=TÞð@u=@yÞ2 is most important. If wall slip is significant, then u(y) velocity profile is greatly affected and hence the channel pressure drop.

206

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

5.4.2.1 Liquid Flow in a Microchannel Minimization of entropy generation as a design tool to determine best device geometry and operation, especially for heat exchangers, has been established for macroscale configurations.68–75 However, fluid flow in microchannels exhibits dominant features often nonexisting or less influential in macrochannels, for example, wall slip velocities for some gases, entrance effects because of the short conduit length, significant surface roughness in relation to microchannel height (or hydraulic diameter), and so on.11 Thus, application of entropy generation minimization principles may assist in the optimal design of microchannel heat sinks and bio-MEMS in light of geometric and operational conditions.67,76–79 Classical methods for enhanced heat transfer, for example, an increase of heat transfer area and/or inlet Reynolds number, are limited options for microchannel flow. Thus, the use of nanofluids as coolants, for example, CuO or Al2O3 nanospheres with diameters in the range of 5 nm < dp < 150 nm in water, oil, or ethylene glycol, is a third option. In case of nanomedicine delivery with bio-MEMS, nanofluid flow analysis is important and measurable reduction of entropy generation is desirable. In this section, entropy generation is minimized for steady laminar pure water and nanofluid flows in a representative trapezoidal microchannel in terms of most suitable channel aspect ratio and Reynolds number range. One effective operational parameter is the inlet Reynolds number; Figure 5.9 indicates a desirable range of 425  Re  1100 for all fluids and aspect ratios considered, ignoring “slit flow” for AR ¼ 0.9337. Due to slightly enhanced frictional

FIGURE 5.9

System entropy generation versus Reynolds number.80

207

Bio-MEMS APPLICATIONS

nanofluid water effects (due to the increase of viscosity), ^ SG;total > ^SG;total . The overall flow field entropy is generated for different scenarios in terms of the following integral form:

1 ^ SG;total ¼ _ p mc

ððð ^ Sgen dV V 8

ð5:7Þ

An important geometric design parameter is the aspect ratio. Figure 5.10 shows ^SG;total ðARÞ for three fluids and different inlet Reynolds numbers. Specifically, for Uin ¼ 4 m s1 (implying Re ¼ 425, 437, and 466 for water, 1% CuO nanofluids, and 4% CuO nanofluids, respectively), the larger aspect ratio generates smaller ^SG;total values, while for Uin ¼ 10 m s1 (implying Re ¼ 1063, 1092, and 1165), the lower aspect ratio generates smaller ^ SG;total values. Clearly, the 1% CuO–water pairing yields more favorable results than the nanofluid with 4% CuO particles. Figure 5.10 also reveals that there are significant trend changes of ^SG;total (AR, Re) for all three fluids occurring at critical values, that is, AR  0.55 and Re  700. The main reasons are that with elevated Reynolds numbers, temperature gradients are reduced and the frictional effects become dominant, even more pronounced as AR ! 1.0. The additional test run for pure water at Re ¼ 638 confirms that Re  700 is critical, while Figure 5.9 has these trend changes in ^SG;total embedded as well.

FIGURE 5.10 System entropy generation as a function of aspect ratio and Reynolds number.80

208

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

5.5 CONCLUSIONS AND FUTURE PERSPECTIVE The development of nanodrugs and their carriers and the design of microscale devices for optimal drug targeting are presently very important research topics. Particularly, lab-on-a-chip systems for drug discovery, development and testing, active nanodrug carriers that attach to tumor cells, and mechanical drug delivery systems (also called bio-MEMSs) have attracted much attention and enjoy increasing use in the healthcare industry. This chapter summarizes some fundamentals and applications of microfluidics and bio-MEMS. Then, it discusses types of nanodrugs and characteristics of their carriers, as well as biochemical and mechanical nanodrug targeting. To illustrate some aspects of bio-MEMS components and applications, computational results for nanodrug mixing and nanofluid flow in microchannels and optimal microchannel design/ operation are presented. Future work will concentrate on the improvements in smart drug delivery systems on the market, ranging from nanodrug carriers to microfluidic devices for optimal drug targeting. One of the more challenging tasks will be to combine biochemically active NDC with bio-MEMSs. They could deliver up to 100% of the NDCs to the predetermined, disease-related target where the NDCs attach to the tumor cells, penetrate them, and release the nanodrugs. The underlying principle of particle targeting is based on the backtracking methodology45 that can be applied to smart delivery of inhaled drug aerosols and injected radioactive microspheres, NDCs, and so on.

REFERENCES 1. Jain, K.K. Drug Delivery Systems. Methods in Molecular Biology, Humana Press: 2008; Vol. 437. 2. Suh, W.H.; Suslick, K.S.; Stucky, G.D.; Suh, Y.-H. Nanotechnology, nanotoxicology, neuroscience. Prog. Neurobiol. 2009, 87, 133–170. 3. Emerich, D.F.; Thanos, C.G. Targeted nanoparticle-based drug delivery and diagnosis. J. Drug Target. 2007, 15(3), 163–183. 4. Kim, S.; Kwon, I.I.K.; Kwon, I.C.; Park, K. Nanotechnology in drug delivery: past, present, and future. In Nanotechnology in Drug Delivery; de Villiers, M.M.; Aramwit, P.; Kwon, G.S., Eds.; Springer/AAPS Press: 2009. 5. Riehemann, K.; Schneider, S.W.; Luger, T.A.; Godin, B.; Ferrari, M.; Fuchs, H. Nanomedicine: challenge and perspectives. Angew. Chem. Int. Ed. 2009, 48, 872–897. 6. Kasuya, T.; Kuroda, S. Nanoparticles for human liver-specific drug and gene delivery systems: in vitro and in vivo advances. Expert Opin. Drug Deliv. 2009, 6(1), 39–52. 7. Subramani, K. Applications of nanotechnology in drug delivery systems for the treatment of cancer and diabetes. Int. J. Nanotechnol. 2006, 3(4), 557–580. 8. Kleinstreuer, C.; Zhang, Z.; Donohue, J.F. Targeted drug-aerosol delivery in the human respiratory system. Annu. Rev. Biomed. Eng. 2008, 10, 195–220. 9. Bayraktar, T.; Pidugu, S.B. Characterization of liquid flows in microfluidic systems. Int. J. Heat Mass Transfer 2006, 49, 815–824.

REFERENCES

209

10. Tabeling, P. Introduction to Microfluidics, Oxford University Press: Oxford, UK, 2005. 11. Nguyen, N.-T.; Wereley, S.T. Fundamentals, Applications of Microfluidics, Artech House: Boston, MA, 2006. 12. Saliterman, S.S. Fundamentals of bioMEMS and Medical Microdevices, Wiley– Interscience/SPIE Press: Bellingham, WA, 2006. 13. Tesar, V. Pressure-Driven Microfluidics, Artech House: Norwood, MA, 2007. 14. Stone, H.A.; Stroock, A.D.; Ajdari, A. Engineering flows in small devices: microfluidics toward a lab-on-a-chip. Annu. Rev. Fluid Mech. 2004, 36, 381–411. 15. Hilt, J.Z.; Peppas, N.A. Microfabricated drug delivery devices. Int. J. Pharm. 2005, 306, 15–23. 16. Whitesides, G.M. The origins and the future of microfluidics: insight overview. Nature 2006, 442, 368–373. 17. Hu, M.; Zengerle, R. Microfabricated devices for controlled biochemical release. In 10th International Conference on New Actuators, Bremen, Germany, 2006, pp 263–271. 18. Geipel, A.; Goldschmidtboeing, F.; Jantscheff, P.; Esser, N.; Massing, U.; Woias, P. Design of an implantable active microport system for patient specific drug release. Biomed. Microdevices 2008, 10, 469–478. 19. Bird, G.A. Molecular Gas Dynamics and Direct Simulation of Gas Flow, Oxford University Press: Oxford, UK, 1994. 20. Koo, J.; Kleinstreuer, C. Liquid flow in microchannels: experimental observations and computational analysis of microfluidics effect. J. Micromech. Microeng. 2003, 13, 568–579. 21. Dutta, D.; Ramachandran, A.; Leighton, D.T. Jr Effect of channel geometry on solute dispersion in pressure-driven microfluidic systems. Microfluid. Nanofluid 2006, 2, 275–290. 22. Zhang, C.; Xing, D.; Li, Y. Micropumps, microvalves, and micromixers within PCR microfluidic chips: advances and trends. Biotechnol. Adv. 2007, 25, 483–514. 23. Oh, K.W.; Ahn, C.H. A review of microvalves. J. Micromech. Microeng. 2006, 16, R13–R19. 24. Tsai, N.-C.; Sue, C.-Y. Review of MEMS-based drug delivery and dosing systems. Sens. Actuators 2007, 134, 555–564. 25. Nguyen, N.-T. Micromixers: Fundamentals, Design and Fabrication, William Andrew: Norwich, NY, 2008. 26. Desai, T.; Bhatia, S. Therapeutic Micro/Nano Technology, BioMEMS and Biomedical Nanotechnology; Springer: 2006. 27. Elman, N.M.; Patta, Y.; Scott, A.W.; Masi, B.; Duc, H.L.H.; Cima, M.J. The next generation of drug delivery microdevices. Clin. Pharmacol. Ther. 2009, 85, 544–547. 28. Kim, K.; Lee, J.-B. MEMS for drug delivery. In Bio-MEMS: Technologies and Applications; Wang, W.; Soper, S.A. Eds.; CRC Press/Taylor & Francis: Boca Raton, FL, 2007. 29. Roxhed, N.; Samel, B.; Nordquist, L.; Griss, P.; Stemme, G. Painless drug delivery through microneedle-based transdermal patches featuring active infusion. IEEE Trans. Biomed. Eng. 2008, 55(3), 1063–1071. 30. Wu, Y.; Qiu, Y.; Zhang, S.; Qin, G.; Gao, Y. Microneedle-based drug delivery: studies on delivery parameters and biocompatibility. Biomed. Microdevices 2008, 10, 601–610.

210

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

31. Nakazawa, G.; Finn, A.V.; Kolodgie, F.D.; Virmani, R. A review of current devices and a look at new technology: drug-eluting stents. Expert Rev. Med. Devices 2009, 6(1), 33–42. 32. Melikian, N.; Wijns, W. Drug-eluting stents: a critique. Heart 2008, 94, 145–152. 33. Wykrzykowska, J.J.; Onuma, Y.; Serruys, P.W. Advances in stent drug delivery: the future is in bioabsorbable stents. Expert Opin. Drug Deliv. 2009, 6(2), 113–126. 34. Kukreja, N.; Onuma, Y.; Daemen, J.; Serruys, P.W. The future of drug-eluting stents. Pharmacol. Res. 2008, 57, 171–180. 35. Jamshidi, P.; Mahmoody, K.; Erne, P. Covered stents: a review. Int. J. Cardiol. 2008, 130, 310–318. 36. Karniadakis, G.; Beskok, A.; Aluru, N. Microflows and Nanoflows: Fundamentals and Simulation, Springer: 2005. 37. Zahn, J.D. Micropump applications in bio-MEMS. In Bio-MEMS: Technologies and Applications; Wang, W.; Soper, S.A.Eds.; CRC Press/Taylor & Francis: Boca Raton, FL, 2007. 38. Tan, K.; Cheang, O.; Ho, I.A.; Lam, P.Y.; Hui, K.M. Nanosized bioceramic particles could function as efficient gene delivery vehicles with target specificity for the spleen. Gene Ther. 2007, 14, 828–835. 39. Maeda, H.; Wu, J.; Sawa, T.; Matsumura, Y.; Hori, K. Tumor vascular permeability and the EPR effect in macromolecular therapeutics: a review. J. Control. Release 2000, 65, 271–284. 40. Drummond, C.D.; Meyer, O.; Hong, K.; Kirpotin, D.B.; Papahadjopoulos, D. Optimizing liposomes for delivery of chemotherapeutic agents to solid tumors. Pharmacol. Rev. 1999, 51(4), 691–744. 41. Snoeys, J.; Lievens, J.; Jacobs, F.; Duimel, H.; Collen, D.; Frederik, P.; Geest, B.D. Species differences in transgene DNA uptake in hepatocytes after adenoviral transfer correlate with the size of endothelial fenestrae. Gene Ther. 2007, 14, 604–612. 42. Bragonzi, A.; Diana, G.; Villa, A.; Galori, G.; Biffi, A.; Bordignon, C.; Assael, B.M.; Conese, M. Biodistribution and transgene expression with nonviral cationic vector/DNA complexes in the lungs. Gene Ther. 2000, 7(20), 1753–1760. 43. Kaparissides, C.; Alexandridou, S.; Kotti, K.; Chaitidou, S. Recent advances in novel drug delivery systems. J. Nanotechnol. 2006, 2, 1–11. 44. Kleinstreuer, C. Biofluid Dynamics: Principles and Selected Applications, Taylor & Francis: Boca Raton, FL, 2006. 45. Kleinstreuer, C. Optimal drug delivery based on fluid-particle dynamic simulations. ANSYS Advantage Vol. 3 (4), 2009. 46. Kleinstreuer, C.; Zhang, Z.; Li, Z.; Roberts, W.L.; Rojas, C. A new methodology for targeting drug-aerosols in the human respiratory system. Int. J. Heat Mass Transfer 2008, 51, 5578–5589. 47. Kennedy, A.; Kleinstreuer, C.; Basciano, C.A.; Dezarn, A. Computer Modeling of 90Y Microsphere Transport in the Hepatic Arterial Tree to Improve Clinical Outcomes. Int. J. Radiol. Oncol. Biol. Phys. (Red Journal), 2010, 76(2), 631–637. 48. Bonacucina, G.; Cespi, M.; Misici-Falzi, M.; Palmieri, G. Review: colloidal soft matter as drug delivery system. J. Pharm. Sci. 2009, 98(1), 1–42. 49. Tomioka, H.; Tatano, Y.; Yasumota, K.; Shimizu, T. Recent advances in anti-tuberculosis drug development and novel drug targets. Expert Rev. Respir. Med. 2008, 2(4), 455–471.

REFERENCES

211

50. Muthu, M.S.; Singh, S. Targeted nanomedicines: effective treatment modalities for cancer, AIDS and brain disorders. Nanomedicine 2009, 4(1), 105–118. 51. Lammers, T.; Hennink, W.E.; Storm, G. Tumor-targeted nanomedicines: principles and practice. Br. J. Cancer 2008, 99, 392–397. 52. Tosi, G.; Bostantino, L.; Ruozi, B.; Forni, F.; Vandelli, M.A. Polymeric nanoparticles for the drug delivery to the central nervous system. Expert Opin. Drug Deliv. 2008, 5(2), 155–174. 53. Polyak, B.; Friedman, G. Magnetic targeting for site-specific drug delivery: applications and clinical potential. Expert Opin. Drug Deliv. 2009, 6(1), 53–70. 54. Bohmer, M.R.; Klibanov, A.L.; Tiemann, K.; Hall, C.S.; Gruell, H.; Steinbach, O.C. Ultrasound triggered image-guided drug delivery. Eur. J. Radiol. 2009, 70(2), 242–253. 55. Kleinstreuer, C.; Li, J.; Koo, J. Microfluidics of nanodrug delivery. Int. J. Heat Mass Transfer 2008, 51, 5590–5597. 56. Labhasetwar, V.; Leslie-Pelecky, D.L. Biomedical Applications of Nanotechnology, Wiley–Interscience: Hoboken, NJ, 2007. 57. Nguyen, N.-T.; Wu, Z. Micromixers: a review. J. Micromech. Microeng. 2005, 15, R1–R16. 58. Stroock, A.D.; et al. Chaotic mixer for microchannels. Science 2002, 295, 647–651. 59. Kim, D.S. A barrier embedded chaotic micromixer. J. Micromech. Microeng. 2004, 14, 798–805. 60. Munson, M.S.; Yager, P. Simple quantitative optical method for monitoring the extent of mixing applied to a novel microfluidic mixer. Anal. Chim. Acta 2004, 507, 63–71. 61. Hardt, S.; Drese, K.S.; Hessel, V.; Schonfeld, F. Passive micromixers for applications in the microreactor and uTAS fields. Microfluid. Nanofluid. 2005, 1, 108–118. 62. Floyd-Smith, T.M.; Golden, J.P.; Howell, P.B.; Ligler, F.S. Characterization of passive microfluidic mixers using soft lithography. Microfluid. Nanofluid. 2006, 2, 180–183. 63. Chang, C.-C.; Yang, R.-J. Electrokinetic mixing in microfluidic systems. Microfluid. Nanofluid. 2007, 3, 501–525. 64. Chung, C.K.; Shih, T.R. Effect of geometry on fluid mixing of rhombic micromixers. Microfluid. Nanofluid. 2008, 4, 419–425. 65. Kang, T.G.; Singh, M.K.; Kwon, T.H.; Anderson, P.D. Chaotic mixing using periodic and aperiodic sequences of mixing protocols in a micromixer. Microfluid. Nanofluid. 2008, 4, 589–599. 66. Brotherton, C.M.; Sun, A.C.; Davis, R.H. Computational modeling and comparison of three co-laminar microfluidic mixing techniques. Microfluid. Nanofluid. 2008, 5, 43–53. 67. Li, J.; Kleinstreuer, C. Microfluidics analysis of nanoparticle mixing in a microchannel system. Microfluid. Nanofluid. 2009, 6, 661–668. 68. Bejan, A. Entropy Generation Minimization: The Method of Thermodynamic Optimization of Finite-Size System and Finite-Time Processes, CRC Press: Boca Raton, FL, 1996. 69. Bejan, A. Fundamentals of exergy analysis, entropy generation minimization, and generation of flow architecture. Int. J. Energy Res. 2002, 26, 545–565. 70. Sahin, A.Z. A second law comparison for optimum shape of duct subjected to constant wall temperature and laminar flow. Heat Mass Transfer 1998, 33, 425–430. 71. Sahin, A.Z. Entropy generation in a turbulent liquid flow through a smooth duct subjected to constant wall temperature. Int. J. Heat Mass Transfer 2000, 43, 1469–1478.

212

MICROFLUIDIC DEVICES FOR NANODRUG DELIVERY

72. Mahmud, S.; Fraser, R.A. The second law analysis in fundamental convective heat transfer problems. Int. J. Thermal Sci. 2003, 42, 177–186. 73. Mansour, R.B.; Galanis, N.; Nguyen, C.T. Dissipation and entropy generation in fully developed forced and mixed laminar convection. Int. J. Thermal Sci. 2006, 45, 998–1007. 74. Khan, W.A.; Culham, J.R.; Yovanovich, M.M. Optimal design of tube banks in crossflow using entropy generation minimization method. J. Thermophys. Heat Transfer 2007, 21(2), 372–378. 75. Ko, T.H.; Wu, C.P. A numerical study on entropy generation induced by turbulent forced convection in curved rectangular ducts with various aspect ratios. Int. Commun. Heat Mass Transfer 2009, 36, 25–31. 76. Chein, R.; Chuang, J. Analysis of microchannel heat sink performance using nanofluids. Appl. Thermal Eng. 2005, 25, 3104–3114. 77. Heris, S.Z.; Etemad, S.G.; Esfahany, M.N. Experimental investigation of oxide nanofluids laminar flow convective heat transfer. Int. Commun. Heat Mass Transfer 2006, 33, 529–535. 78. Jang, S.P.; Choi, S.U.S. Cooling performance of a microchannel heat sink with nanofluids. Appl. Thermal Eng. 2006, 26, 2457–2463. 79. Kleinstreuer, C.; Li, J. Microscale Cooling Devices, Encyclopedia of Micro and Nanofluidics, Li, D.; Ed.; Springer-Verlag: Heidelberg, 2008. 80. Li, J.; Kleinstreuer, C. Entropy generation analysis for nanofluid flow in microchannels. ASME J. Heat Transfer, 2010, in press.

6 MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES MUHAMMAD J. A. SHIDDIKY AND YOON-BO SHIM Department of Chemistry and Institute of Biophysio Sensor Technology, Pusan National University, Busan, South Korea

6.1 INTRODUCTION Nanoparticles (NPs), generally defined as materials with a particle size of less than 100 nm in at least one dimension, are of interest because their chemical and physical behavior is unprecedented and remarkably different from those in bulk form. They have great potential for applications in electronic, chemical, or mechanical industries, as well as in technologies, including superconductors, catalysts, drug carriers, sensors, magnetic materials, pigments, separation science, and in structural and electronic materials. In separation science alone, significant advances have been made in electrophoresis and microchip separations employing nanoparticles.1,2 The small size of nanoparticles is responsible for their novel, unrevealed properties (electrical, magnetic, chemical, optical, and mechanical).3 In addition, some nanoparticles possess advantages such as a large surface area, good chemical and thermal stability, significant mechanical strength, ease of modification, and biomolecular compatibility. Owing to these unique properties, nanoparticles have attracted immense attention in separation science. Chemical separations are inevitable for analysis of complex samples, in particular biological samples. In this vein, capillary electrophoresis (CE) and microchip electrophoresis (MCE) are, among all, two promising separation techniques.4–6

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

213

214

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

Since 1989, when nanoparticles were first applied as a pseudostationary phase in CE by Wallingford and Ewing,1 new separation media containing a large variety of nanoparticles with a wide range of chemically useful characteristics were developed for CE and MCE analyses. A pseudostationary phase with a large surface area, in combination with an electroosmotic flow (EOF)-driven system, has great potential in a highly efficient separation technique. For example, commercially available silica nanoparticles have been used as separation buffer additives in CE by Fujimoto and Muranaka.7 In the past decade, extensive research on nanomaterial applications in CE and MCE has evolved and different types of nanoparticles have been tested, such as gold nanoparticles (AuNPs), silica nanoparticles, magnetic nanoparticles, polymer nanoparticles, metal oxide nanoparticles, and carbon nanotubes. Until recently, nanoparticle-mediated CE and MCE have been studied in the analysis of proteins, DNA separation and sequencing, separation of drugs and drug delivery analysis in biomedical sciences, separation of environmental pollutants (organic and inorganic), and many other small molecule analyses. Much of the earlier works regarding use of nanoparticles in CE and MCE separation has been summarized in several reviews.8–13 This chapter gives an overview of the new developments and innovative applications of nanoparticles, including metal and metal oxide nanoparticles, carbon nanotubes, silica nanoparticles, and polymeric nanoparticles as stationary and/or pseudostationary phases in CE and MCE. The use of nanoparticles in the pseudostationary phases of CE and MCE is discussed in detail.

6.2 MICROCHIP ELECTROPHORESIS 6.2.1 Microfluidic Devices Since the conception of the microfluidic bioanalytical device over a decade ago by Widmer’s group,14 it has become a very efficient platform for the analysis of biologically, clinically, and environmentally important analytes. The microfluidic device frequently referred to either as lab-on-a-chip or micrototal analysis systems (m-TAS), typically consists of microchannel networks, miniaturized chambers/reactors, microseparation/detection units, and combinations thereof to accommodate the sensitive and large-scale analysis of analytes. Among many other components in a microfluidic device, valves and pumps, mixers, injectors, microreactors, preconcentrators and microseparators, filters, detectors, and temperature measurement units are the major components.15 6.2.2 Advantages and Applications of Microfluidic Devices The microfluidic devices have numerous advantages over traditional systems based on robotics and conventional analysis instrumentation. First, mature microfabrication techniques adapted from the semiconductor industry allow mass production of lab-ona-chip devices, thereby reducing cost per device and allowing massive parallel systems to be constructed easily. Other advantages include high speed, high efficiency, high

215

MICROCHIP ELECTROPHORESIS

throughput, low sample/reagent consumption, low waste generation, portability, and disposability.16 Possible applications include clinical instrumentation for point-ofcare testing, real-time monitoring for biodefense, methods for glucose monitoring and the subsequent release of insulin, forensic applications including DNA analysis at crime scenes, and environmental applications such as on-site testing of explosives or phenolic pollutants in groundwater.16–20 6.2.3 Separation Techniques in Microfluidic Devices Separation in a microfluidic device is based on a variety of principles, from simple microchip zone electrophoresis (MZE) to more complex methods, such as microchip gel electrophoresis (MGE), micellar electrokinetic chromatography (MEKC), isotachophoresis, microchip electrochromatography, isoelectric focusing, and microemulsion electrokinetic chromatography. 6.2.3.1 Microchip Zone Electrophoresis Zone electrophoresis is the simplest form of microchip electrophoresis. In MZE, the species to be separated are dissolved in buffer only, so that the ions can freely move along the solution by diffusion and/or under the influence of an electric field. Every charged molecule can be characterized by its electrophoretic mobility (mep), which is largely determined by its size radius (r) and charge (q): mep ¼ q=6pr h

ð6:1Þ

where h is the viscosity of the buffer system. When there is no electroosmotic flow, the final velocity (u) of a charged molecule is derived as u ¼ uep ¼ mep
E

ð6:2Þ

where E is the electric field strength. However, most often, electroosmotic flow occurs as well, and the final velocity is found as the vector sum of the electrophoretic velocity (uep) and the electroosmotic velocity (ueo). Because the electroosmotic flow, and hence the velocity associated with it, is a bulk property, it uniformly affects all molecules. In addition, the absolute magnitude of the electroosmotic velocity is larger than any of the electrophoretic velocities, and therefore all species, positively charged, uncharged, or negatively charged, move in the same direction, namely, the direction defined by the electroosmotic flow (Figure 6.1). A sample containing both cations and anions can be injected at one end of a capillary or microchannel and then, after separation has occurred, the individual bands of ions can be detected at the other end. This makes the necessary experimental set up for MZE much simpler. Separation in such systems is governed by the charge to size ratio of the molecules. Therefore, the small, highly charged cations are first to migrate through the channel and arrive at the detector, followed by larger, less charged cations, then all uncharged molecules, followed by larger, less charged anions, with the highly charged small anions arriving last at the detector (Figure 6.1). Due to the high

216

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.1 (a) An open-ended microchannel extends between two reservoirs, across which a high voltage is applied. This voltage causes analytes to migrate from the site of sample application at the cathode buffer through a detector to the anode. The EOF that results from wall pumping drives the separation of analytes. Reproduced from Ref. 21, with permission. (b) Vector addition of electrophoretic mobilities of individual ions and the electroosmotic mobility (meo) to yield total mobility (mtot).

separation power of MZE, even small differences in the charge to mass ratio can be sufficient to achieve separation. In MZE, due to the absence of the other interaction equilibrium, and ignoring contributions from the injection plug width and detector geometry, the only source of dispersion is diffusion, the effects of which can be minimized by decreasing the time from injection to detection. A large number of MZE separations have been performed mainly due to the relative ease of implementation and good separation efficiency.22 6.2.3.2 Separation of DNA by Microchip Gel Electrophoresis The mobility of DNA depends on its charge to mass ratio. In a free solution, in trisacetate or tris-borate buffer (pH 8.4), the difference of its charge to mass ratio is almost

MICROCHIP ELECTROPHORESIS

217

identical for fragments from approximately 400 to 48,500 bp. DNA is a biopolymer consisting of many repeating units called nucleotides, and the addition and subtraction of such units change only the charge and absolute mass or size but not the ratio of these two parameters. Because of this property, the separation of DNA fragments from approximately 400 to 48,500 bp is not possible in a free solution.23 Thus, to separate DNA molecules, channels are typically filled with a polymer gel that acts as a sieving matrix. This sieving matrix includes noncross-linked polymers, such as linear polyacrylamide, polyethylene glycol, and cellulose derivatives, as well as crosslinked polymers or gels, such as polyacrylamide and agarose. The entangled polymer network inside the microchannel serves as a molecular sieve in which smaller DNA and protein molecules migrate faster than larger ones. Electrophoresis of DNA in a gel matrix can be thought of as a type of “gel filtration” where a mixture of DNA molecules of different sizes is forced to move through the pores of the matrix under the influence of an electric field. Any DNA fragment to be fractionated encounters the gel network of polymer threads or pores that increases the effective friction and consequently lowers the velocity of the molecules. Small molecules move more rapidly through the gel, while large molecules move relatively slow. The existence of the gel medium significantly contributes to the observed electrophoretic mobility of the DNA molecules. The pore size in the gel matrix plays a critical role in determining the relative electrophoretic mobility and separation efficiency of the DNA fragments. Unlike MZE, electroosmotic flow is mostly undesired in MGE, particularly for separation of negatively charged DNA fragments. In bare silica channels, a large magnitude of the EOF drives DNA molecules toward the cathode and therefore distorts the separation of the DNA. In some cases, electroosmotic forces within the channel may cause the gel to migrate out of the channel. In addition, analyte–wall interactions due to ionic interactions and hydrogen bonding significantly interfere with the separation as well. Because of these issues, it is desirable to suppress EOF within the channel for separation of DNA by MGE. EOF can be controlled by applying a proper coating, either covalent or dynamic, at the inner surface of the channel.24 The coatings can increase, decrease, reverse, or eliminate the EOF depending on the presence or absence of certain functional groups contained within the coating materials. Coated microchannels also prevent sample interaction with the channel and feature good separation efficiencies and excellent run-to-run reproducibility in migration times. Most DNA separations by MGE are conducted in coated channels where the EOF is completely eliminated or significantly reduced to achieve highresolution separations. 6.2.3.3 Micellar Electrokinetic Chromatography MEKC is particularly useful for separating small and neutral molecules, which has been impossible by gel or zone electrophoresis.25 The separation mechanism is based on partitioning of the analyte between the micelle and the surrounding aqueous phase. MEKC can be performed by dissolving an ionic surfactant in the CE running solution at a concentration higher than the critical micelle concentration (cmc), with no

218

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.2 Separation principle of MEKC. Reproduced from Ref. 25, with permission.

instrumental modification. In general, neutral or alkaline buffer solutions are used to create conditions for a strong EOF that moves the entire liquid stream in the capillary toward the cathode (Figure 6.2). Therefore, even anionic micelles such as sodium dodecyl sulfate (SDS) migrate toward the cathode. The neutral analyte is not solubilized by or is free from the micelle and thus migrates at the same velocity as that of the EOF; the analyte that is totally incorporated into the micelle migrates at the same velocity as that of the micelle. Other neutral analytes are detected between t0 and tmc, migration time of the EOF marker and the micelle, respectively. The interval between t0 and tmc is called the migration time window. The wider the window, the larger the peak capacity, which is the number of peaks that can be separated during a run. Migration time can be measured by using markers such as methanol for EOF and Sudan III for the micelle. The retention factor k can be defined as k ¼ nmc =naq

ð6:3Þ

in which nmc and naq are the number of moles of the analyte in the micelle and surrounding aqueous phase, respectively; k can be measured by k ¼ ðtR t0 Þ=ft0 ð1tR =tmc Þg

ð6:4Þ

in which tR is the migration time of the analyte.26 The difference between this equation and the conventional one used in chromatography is the limited migration time window in MEKC. Although the micelle is not fixed inside the capillary, it plays the same role as the stationary phase in chromatography and is therefore called the pseudostationary phase. The MEKC resolution Rs, equation is Rs ¼

p

N=4½ða1Þ=a
½k2 =ð1 þ k2 Þ
ð½ð1t0 =tmc =½1 þ ðt0 =tmc Þk1 Þ

ð6:5Þ

APPLICATION OF NANOPARTICLES IN CE AND MCE

219

in which N is the plate number and a the selectivity factor equal to k2/k1.27 Equation 6.5 is similar to the one used in conventional chromatography, except for the addition of the last term on the right-hand side. This variable comes from the migration of the micelle or pseudostationary phase inside the capillary, thus, the migration of the pseudostationary phase causes reduction of the column length.28 If the micelle migration is completely suppressed or tmc is infinity, the resolution equation is the same as the conventional one. The third term on the right-hand side of equation (6.5) is the retention factor term. It is not independent of the other variable because it includes k. The optimum kopt value to maximize resolution is easily determined by25 kopt ¼

p

ðtmc =t0 Þ

ð6:6Þ

which differentiates the product of the last two terms in equation (6.5). Under neutral or acidic conditions, tmc/t0 is 3–4 and kopt is 1.7–2.0.25 To adjust k in MEKC, the concentration of the surfactant can be increased or decreased since k can be expressed as k ¼ KVmc =Vaq ¼ KvðCsf  cmcÞ

ð6:7Þ

in which K is the distribution coefficient of the analyte between the micelle and the aqueous phase, Vmc and Vaq are the volume of the micelle and the aqueous phase, respectively, v is the partial specific volume of the micelle, and Csf is the surfactant concentration.27 As shown by equation (6.7), k is linearly proportional to the surfactant concentration, an advantage of MEKC because the Csf needed to obtain a given k can be calculated provided the cmc and k at a certain Csf are known. MEKC rarely separates extremely hydrophobic analytes with high k-values. However, several strategies are possible. Adding an organic solvent significantly reduces k and gives better resolution of extremely hydrophobic analytes. The organic aqueous solution has a higher viscosity, and the migration times will be long. Adding too much organic solvent may destroy the micellar structure and/or decrease the migration time window because the electrophoretic mobility of the micelle is reduced, a likely result of the reduced charge on the micelle or the increased size of the micelle due to swelling caused by the organic solvent.25

6.3 APPLICATION OF NANOPARTICLES IN CE AND MCE 6.3.1 Why Nanoparticles The emergence of nanotechnology is opening up new horizons for the application of nanoparticles’ physics, chemistry, biology, medicine, materials science, and interdisciplinary fields. In particular, nanoparticles are of considerable interest in separation sciences due to their attractive physical and chemical properties. The unique

220

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

properties of nanoparticles offer excellent prospects for the development of separation media because of the following reasons: 1. Nanoparticles exhibit higher ratios of surface area to volume than their bulk counterparts for organofunctional groups that can interact with the channel/ capillary surface, target analytes, or both leading to enhanced separation selectivity and efficiency. 2. Nanoparticles offer additional interaction sites where the solute can interact with the running buffer additives and therefore lead to a higher selectivity for target molecules. 3. Some nanoparticles, particularly metal nanoparticles, can be conjugated with polymeric materials (e.g., poly(diallyldimethylammonium chloride)) and also act as an excellent pseudostationary phase that offers excellent separation performances with advanced functional properties for constructing separation media for DNA and proteins. 4. The use of the nanoparticles in detection is compatible with the miniaturization and integration of the analytical instruments, along with the inherent simplicity, speed, high selectivity, and excellent catalytic activity. The application of nanoparticles in CE and MCE has been studied mainly due to the enhanced separation performances (selectivity and efficiency) owing to these important features. 6.3.2 Nanoparticle-Mediated Capillary Electrophoresis Using nanoparticles as run buffer additives is similar to using micelle additives in micellar electrokinetic chromatography.25–27,29 In both cases, the purpose of using the additive is to provide additional interaction sites with which the solutes can interact. Both MEKC and nanoparticle-mediated electrophoresis (NME) offer some definite advantages. (i) The presence of nanoparticles and micelles in the run buffer avoids the need to pack the capillary with a stationary phase. (ii) Since there is no conventional stationary phase, the need for frits and other retaining techniques is eliminated. (iii) Micelles and nanoparticles in the run buffer move through the capillary with an apparent mobility that takes into account the effects of both the electroosmotic flow and their own native electrophoretic mobility. As a result, there is a constant turnover in the interacting media. The solutes are moving under the influence of the electric field and are separated on the basis of their different effective charges and by differential partitioning between the aqueous buffer and the nanoparticle or micellar pseudostationary phase. This additional partitioning effect increases the separation degrees of freedom and allows separation not only of the charged solutes but also of the neutral ones. Thus, the user can control the nature of interactions with additives and tailor the CE system to specific analytes. Although the use of nanoparticles and micelles in CE and MCE has many similarities, there are also some important differences between these two additives. One main difference between MEKC and

APPLICATION OF NANOPARTICLES IN CE AND MCE

221

NME is that in the former approach the solutes can penetrate the core of the micelles, whereas in NME, the interactions with the solutes occur always at the outer surface of the nanoparticles. These interactions can be either with the surface itself or with the organic moieties attached to the surface of the nanoparticle. In addition, NME extends the operable range to much higher fields. The operable field in CE is limited to a large extent by thermal effects induced by the applied voltage and the generated current. Therefore, it is advantageous, whenever possible, to operate CE at a low ionic strength. In the case of MEKC, it is necessary to use high surfactant concentrations that exceed the critical micelle concentration. Stabilized nanoparticles do not suffer from this drawback and stable colloidal solutions are attained even without dissolved species. 6.3.2.1 Gold Nanoparticles The presence of gold nanoparticles in the run buffer in CE has the ability to significantly affect the apparent mobility (map) of a certain solute and change the electroosmotic mobility (meo) of the run buffer. Ovadia et al.29 used citrate-stabilized AuNPs in conjunction with channels treated with poly(diallyldimethylammonium chloride) (PDDC) to manipulate the selectivity between solutes in CE. Modification of the channel wall with only PDDC covers the silanol groups with positive quaternary ammonium groups that adsorb the negatively charged gold nanoparticles. The adsorption of PDDC on the capillary and the subsequent introduction of the citrate-stabilized gold nanoparticles all change the apparent mobilities. Both the apparent selectivities and the observed selectivities, which take into account the effect of the electroosmotic flow, are affected. The extent of the change depends on the charge of the solutes and their functional groups. With the negatively charged solutes, the introduction of PDDC decreases the absolute value of the apparent mobilities. The introduction of the citrate-stabilized gold nanoparticles to the run buffer resulted in a further decrease in the apparent mobility of the solutes. Increasing the concentration of the nanoparticles in the buffer caused little change to the map values. The electrophoretic selectivity (aef) is defined here as the ratio of the electrophoretic mobilities (mef) of two neighboring solutes in an electropherogram: aef ¼ mef;2 = mef;1

ð6:8Þ

Similarly, the apparent selectivity (aap) is the ratio of the two neighboring apparent mobilities: ð6:9Þ aap ¼ map;2 = map;1 In the presence of the electroosmotic flow (meo), the observed selectivity (aobs) can be substantially different from the electrophoretic or apparent selectivity. The EOF vector can either enhance or diminish the electrophoretic or apparent selectivity. The observed selectivity is described by aobs ¼ tm;2 =tm;1 ¼ mobs;2 =mobs;1 ¼ map;2 =map;1 þ meo =meo

ð6:10Þ

where tm represents migration times and indexes 1 and 2 denote the solutes. For the positively charged solutes, aobs decreases slightly with increasing meo. For the

222

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

negatively charged solutes, since |meo| < |map|,1 the selectivities increase with increasing meo. The presence of the PDDC-gold coating modifies the electroosmotic mobility and the observed mobility of the solutes. These changes in the mobilities are manifested in selective alterations and allow obtaining separations that cannot be achieved without nanoparticles. AuNPs have been employed as pseudostationary phases in AuNP-coated capillaries for CE separation of acidic and basic proteins at low pH (3), achieving high plate numbers and high run-to-run reproducibility.31 The AuNPs were first protected by a bilayer of didodecyl dimethylammonium bromide (DDAB). The inner wall of the capillary was then coated with these nanoparticles to prevent protein adsorption. Another effect of the adsorbed, positively charged nanoparticle coating was a reversal in EOF. In addition, the DDAB-capped gold nanoparticles were noncovalently coated with polyethylene oxide (PEO) and dispersed in the electrolyte and used as a pseudostationary phase to enhance interactions with proteins. PEO was attached through the hydrophobic regions with the hydrophilic groups able to interact with proteins. At neutral pH, plate numbers of 23,000 were achieved for bovine serum albumin, which rose to 62,000 at pH 3.5. For lysozymes, respective plate numbers of 81,000 and 450,000 were achieved. Introducing PEO also raised the theoretical plates to over 1000,000 for lysozymes. Li and coworkers reported the synthesis of bifunctional Au-Fe3O4 nanoparticles that possess high catalytic activity for separation and detection of proteins.32 AuFe3O4 nanoparticles combined the merits of both gold and Fe3O4 nanoparticles and were formed by chemical bond linkage. Owing to the introduction of AuNPs, the bifunctional Au-Fe3O4 nanoparticles can be easily modified with other functional molecules to realize various nanobiotechnological separations and detections. For example, Au-Fe3O4 nanoparticles can be modified with nitrilotriacetic acid molecules through Au–S interaction and used to separate proteins simply with the assistance of a magnet. Bradford protein assay and sodium dodecyl sulfate–polyacrylamide gel electrophoresis were performed to examine the validity of the separation procedure, and the phosphate determination method suggested that the as-separated protein maintained catalytic activity. This result showed the efficiency of such a material in protein separation and suggested that its use could be extended to magnetic separation of other biosubstances. The author concluded that the developed synthetic strategy could be useful for facile preparation of diverse bifunctional and even multifunctional nanomaterials. A method for enrichment and separation of acidic and basic proteins using centrifugal ultrafiltration, followed by nanoparticle-filled capillary electrophoresis, has been described by Tseng and coworkers.33 To improve stacking and separation efficiencies of proteins, a separation buffer containing 1.6% PDDA was added to the AuNPs, PEO, cetyltrimethylammonium bromide (CTAB), and poly(vinyl alcohol). The use of AuNP as additives exhibited better efficiency in separation, stacking, and analysis time. The separation efficiencies of acidic and basic proteins remained greater than 10,000 and 100,000 plates m1, respectively. To further enhance detection sensitivity, protein samples were enriched using centrifugal ultrafiltration, followed by our proposed stacking method. The detection sensitivity was improved

APPLICATION OF NANOPARTICLES IN CE AND MCE

223

up to 314-fold compared to normal hydrodynamic injection. In addition, the limits of detection for most proteins were down to a nanomolar range. The proposed method was also successfully applied to the analysis of egg-white proteins. More recently, Tween 20-capped AuNPs were used as selective probes for extraction and preconcentration of aminothiols from an aqueous solution (Figure 6.3).34 Tween 20 molecules were noncovalently attached to the surface of AuNPs to form Tween 20–AuNPs. These modified-AuNPs were then used for selective extraction of aminothiols through

FIGURE 6.3 Online concentration and separation of urine samples by PDDA-filled CE (a) before and (b and c) after extraction of 5.0 mL urine using 100 Tween 20–AuNPs (200 mL). Urine samples (b) without and (c) with spiking of 5 mM GSH, Cys, and HCys were hydrodynamically injected by raising the capillary inlet 20 cm high for 180 s. Analytes attached to the surface of AuNPs are released upon the addition of 500 mM DTT (20 mL). A 50 cm capillary (20 cm to detector) is filled with 1.6% v/v PDDAC solution, which is prepared in 20 mM phosphate solution at pH 2.0. The applied voltage is 8.5 kV whereas the electric current is 80 mA. The detection wavelength is set at 200 nm. Reproduced from Ref. 34, with permission.

224

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

the formation of Au–S bonds. After extraction and centrifugation, aminothiols were detached from the surface of AuNPs by adding DTT in high concentration. This probe was then used in combination with CE and UV absorption detection. On-line concentration and separation of the released aminothiols were performed using 1.6% v/v PDDA as an additive in CE. Under optimal extraction and stacking conditions, the detection limit for glutathione (GSH), cysteine (Cys), and homocysteine (HCys) were 28, 554, and 456 nM, respectively. In comparison to the normal injection without the extraction procedure, approximately 2280-, 998-, and 904-fold improvement in sensitivity was observed for GSH, Cys, and HCys, respectively. It is believed that this approach had a significant potential to be extended to clinical diagnosis. Spherical AuNPs with sizes ranging from 2.0 to 3.0 nm were used for open-tubular CE for separation of thiourea, naphthalene, biphenyl, and four polycyclic aromatic compounds with high efficiency.35 The capillary was first etched with ammonium hydrogen difluoride, followed by prederivatization with (3-mercaptopropyl)-trimethoxysilane and immobilization of the dodecanethiol-capped AuNPs. An etching process was used to increase the surface area of the capillary inner wall, by a factor of up to 1000. The methodology in the article presented high separation efficiencies (up to 200,000 plates m1), although the etching process is time consuming and complicated. For the development of a low-viscosity polymer solution with high sieving and selfcoating abilities in DNA sequencing by CE, Wang and coworkers prepared AuNPs with particle sizes of approximately 20, 40, and 60 nm and added them to a quasiinterpenetrating network (quasi-IPN) composed of linear polyacrylamide (LPA) with different viscosity-average molecular masses of 1.5, 3.3, and 6.5 MDa, and poly-N,Ndimethylacrylamide (PDMA) to form polymer/metal composite matrices, respectively.36 These matrices improve ssDNA sequencing performances due to interactions between AuNPs and polymer chains and the formation of physical cross-linking points as demonstrated by intrinsic viscosities and glass transition temperatures. Addition of AuNPs to quasi-IPN containing 1.5 MDa LPA could significantly improve sieving ability for both small and large DNA fragments. The sieving ability of quasi-IPN/ AuNPs with lower molecular mass LPA and AuNPs approximated those of quasi-IPN with higher molecular mass LPA without AuNPs. Thus, the use of quasi-IPN/AuNPs with lower molecular mass LPA could avoid problems in relation to LPA with higher molecular mass, such as difficult preparation and very high viscosity, and thus promote full automation. As a result of the interactions of AuNPs with polymer chains, more dilute solutions with lower viscosities possessed much improved sieving performances in terms of resolution and migration time than relatively concentrated quasi-IPN, without AuNPs, containing the same LPA. Understanding nanoparticle behavior in the presence of an electric field will have a significant impact on separation science where nanoparticles can serve to improve either the mobility or detection sensitivity of target molecules. In this effort, Haes and coworkers exploited nanoparticle behavior in the presence of an electric field and investigated their impacts on separation science where nanoparticles are employed as pseudostationary phase comprising only 2% of the total capillary volume.37 The

APPLICATION OF NANOPARTICLES IN CE AND MCE

225

optical properties of covalently functionalized gold nanoparticles have been used to investigate both the stability of nanoparticles and the mobility of dopamine, epinephrine, and pyrocatechol in capillary electrophoresis, with either primarily covalently functionalized carboxylic acid (Au-COOH) or amine (Au-NH2) surface groups being characterized using extinction spectroscopy, transmission electron microscopy, and zeta potential measurements and utilized in separating neurotransmitters. The authors anticipated at least three nanoparticle-specific mechanisms, those effecting the separations. First, the degree of nanoparticle–nanoparticle interactions is quantified using a new parameter termed the critical nanoparticle concentration (CNC). CNC is defined as the lowest concentration of nanoparticles that induces predominant nanoparticle aggregation under specific buffer conditions and is determined using dual-wavelength photodiode array detection. Once the CNC has been exceeded, reproducible separations are no longer observed. Second, nanoparticle–analyte interactions are dictated by electrostatic interactions that depend on the pKa of the analyte and surface charge of the nanoparticle. Finally, nanoparticle–capillary interactions occur depending on surface chemistry. As a result of these three nanoparticlespecific interactions, the mobility of neurotransmitters increases in the presence of amine-terminated nanoparticles but decreases slightly with carboxyl-terminated nanoparticles. Furthermore, these interactions also directly influence the mobility of nanoparticles. They also found that run buffer viscosity was influenced by the formation of a nanoparticle steady-state pseudostationary phase along the capillary wall. Despite differences in buffer viscosity leading to changes in neurotransmitter mobilities, no significant changes in electroosmotic flow were observed. 6.3.2.2 Silica Nanoparticles Chang and coworkers described two methods for the analysis of biologically active amines using silica nanoparticles as a pseudostationary phase modifier in conjunction with laser-induced native fluorescence detection.38,39 The methods were applied for the analysis of urine samples. In the first method, the CE capillary wall was dynamically coated with poly(vinylpyrrolidone) and poly(ethylene oxide) to suppress EOF and minimize interactions between the capillary wall and analytes. Addition of silica nanoparticles to the electrolyte caused an increase in EOF, thereby enhancing separation speed, due to adsorption of silica nanoparticles onto the capillary wall. In addition, the adsorption of the silica nanoparticles onto the capillary wall suppressed nonspecific interactions of analytes with the inner wall of the capillary, enabling sharper peaks and thus achieving higher plate numbers. In the second method, capillaries were dynamically coated with silica nanoparticles and poly(L-lysine). The EOF direction was controlled by varying the outermost layer of the capillaries with poly(L-lysine) and SiO2 NPs. Over the pH range 3.0–5.0, the (poly(Llysine)–SiO2NP)n–poly(L-lysine) capillaries had an EOF toward the anodic end and were more suitable for the separation of acids with respect to speed, while the (poly(L-lysine)–SiO2NP)n capillaries had an EOF toward the cathodic end and were more suitable for the separation of biogenic amines regarding speed and sensitivity. The separation of standard solutions containing five amines and two acids by CE

226

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

with LIF detection using (poly(L-lysine)–SiO2NP)2–poly(L-lysine) and (poly(Llysine)–SiO2NP)3 capillaries was accomplished within 10 and 7 min, providing plate numbers of 38,000 and 50,000 plates m1 for 5-hydroxytryptamine, respectively. A CE separation method employing aminopropyl-modified nanoparticles as a pseudostationary phase was evaluated for improving separation efficiency using a mixture of six aromatic acids.40 An average theoretical plate value of 10,000 plates m1 was achieved. Excellent reproducibility of the migration time, peak area, and peak height were obtained. The ability to separate six aromatic acids with a plate number greater than 50,000 plates m1 and 100-fold enhanced sample capacity and sensitivity suggested their method may hold great promise to be incorporated into multidimensional separation approaches. 6.3.2.3 TiO2 For separation of oligopeptides and proteins in open-tubular CE,41 TiO2-based nanoparticles were attached to the inner wall of the capillary by a condensation reaction. Nanoparticles with a size of 10 nm were stabilized by polyethylene glycol. It was suggested that the main separation mechanism was based on the ligand exchange of the analytes with the phosphate ion groups adsorbed onto the TiO2-based nanoparticles. The CE system was used for separation of the angiotensin-type oligopeptides in phosphate buffer (pH 8.0), with an average separation efficiency of 31,000 plates m1. In another work, the system was used for separation of conalbumin, apo-transferrin, ovalbumin, and bovine serum albumin.42 Egg-white proteins (lysozyme, conalbumin, and ovalbumin a and b) were separated with an efficiency of 10,000 plates m1 for ovalbumin, with five different glycol isoforms of ovalbumin resolved. This was an interesting approach for protein separation at neutral pH and moderate ionic strength with reasonable selectivity. However, plate numbers were low and separation time was long. In addition, the capillary coating procedure is time consuming and complicated. 6.3.2.4 Carbon Nanotubes Carbon nanotubes (CNTs) have been widely used in a variety of areas, including biosensors, solar cells, field emission devices, and molecular electronics.43 In electrochemical assays, CNTs play a dual role in both the recognition and the transduction events, namely, as a building block material for biomolecule attachment via covalent bond formation between the carboxylic acid group of the CNTs and the amine groups of the biomolecules, and as molecular wires to allow electrical communication between the underlying electrode and the enzyme labels attached to the ends of the CNTs.44 CNTs also possess a large surface area, good chemical stability, and significant mechanical strength. These unique properties of CNTs make them extremely attractive for CE and MCE separation. The first use of CNTs as a pseudostationary phase in CE separation was described by Wang et al.45 who used carboxylic single-wall nanotubes (SWNTs) for separation of caffeine and theobromine with improved resolution. Xu and Li described the use of multiple-wall carbon nanotubes (MWCNs) for the separation of DNA fragments by a CE-contactless conductivity detection method with improved resolution.46 Na et al.

APPLICATION OF NANOPARTICLES IN CE AND MCE

227

compared the use of four different kinds of b-cyclodextrin (CD)-modified nanoparticles, MWNTs, polystyrene nanoparticles, TiO2-based nanoparticles, and Al2O3based nanoparticle, as pseudostationary phases for enantioseparation of the b-blocker clenbuterol.47 The use of surfactants was necessary to form stable nanoparticle suspensions. The use of b-CD-modified nanoparticles improved resolution compared to the use of the same amount of free b-CD due to the orientation of the b-CD adsorbed onto the nanoparticles, which allowed a better contact with analytes. MWNTs were also used for separation of seven purine and pyrimidine bases in CE.48 The method was also applied to determine purine and pyrimidine bases in yeast RNA. Separation of the adenine and thymine bases was accomplished in an electrolyte containing carboxylated MWNTs, previously impossible in the absence of MWNTs. MWCNTs interact with molecules via electrostatic interaction and hydrogen bonding. A network of MWNTs with molecular sieving properties was proposed to be responsible for separation. It was suggested that the presence of MWNTs decreased analyte–capillary wall interactions and stacked the analytes into narrower zones due to their large surface area and abundance of functional groups on the surface. The use of such systems could be favorable for protein separation where analyte-capillary wall interaction is an issue. 6.3.2.5 Lipid-Based Liquid Crystalline Nanoparticles Nanoparticle was utilized by Nilsson et al.49 for separation of proteins with similar mass to charge ratio at neutral pH without organic modifier using a hydrophobic interaction chromatography-based mechanism. Lipid-based liquid crystalline nanoparticles were prepared and used as pseudostationary phase. These nanoparticles have potential benefits including high biocompatibility, ease of preparation, and suspension stability at high concentrations. Using laser-induced fluorescence enabled detection at high nanoparticle concentrations. Green fluorescent protein (GFP) and mutants of GFP harboring single or double amino acid substitutions with the same charge were separated in the described system but not in conventional capillary electrophoresis (Figure 6.4). Separation was achieved by increasing the salt concentration to promote hydrophobic interactions by shielding the repulsive electrostatic interactions. In addition, the method was adapted to a capillary with an effective length of 6.7 cm, enabling fast separations and future applications on chip. Charged characteristics of biodegradable nanoparticles would influence not only the particulate flocculation during preservation but also the colloidal drug-releasing behavior and their interaction with biological cells. Hence, the electrophoretic mobility and the zeta potential of bovine knee chondrocytes (BKCs) and the three synthetic biocompatible NPs were evaluated under various ionic strengths and ionic species. Kuo and Lin50 have studied the electrophoretic mobility, zeta potential, and fixed charge density of BKCs, methyl methacrylate-sulfopropyl methacrylate (MMASPM) nanoparticles, polybutylcyanoacrylate (PBCA) NPs, and solid lipid nanoparticles (SLNs) under the influences of Na þ , K þ , and Ca2 þ with various ionic strengths. Results revealed that, for a specific cationic species, the absolute values of the electrophoretic mobility, the zeta potential, and the fixed charge density decreased with an increase in ionic strength. For a constant ionic strength, the effect

228

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.4 (a) 3D structure of native green fluorescent protein (PDB ID: 1QYO) visualized by PyMOL molecular graphics system (DeLano Scientific, Palo Alto, CA). (b) Separation of GFP, ( þ )GFP, ()GFP, and (H)GFP on a nondenaturing 5% polyacrylamide gel in 0.75 M Tris–HCl, pH 8.8. Electrophoresis was conducted in 192 mM glycine and 8.3 mM Tris–HCl, pH 8.3, for 30 min at 150 V. The fluorescence was visualized on a UV table set to an excitation wavelength of 365 nm. (c) Electropherogram showing separation of GFP, ( þ )GFP, and () GFP by capillary electrophoresis. (d) Electropherograms showing separation of GFP and (H) GFP by capillary electrophoresis. (c and d) Capillary: 17 cm effective length, 24 cm total length, 50 mm ID and 375 mm OD; electrolyte: 100–250 mM tricine, pH 7.5; separation voltage: 10 kV (normal polarity); detection: laser-induced fluorescence, excitation at 488 nm and emission at 520 nm; sample: GFP, ( þ )GFP and ()GFP (c) and GFP and (H)GFP (d), 0.01 mg mL1 of each protein in 50 mM tricine, pH 7.5; injection: 3 kV, 3 s. Reproduced from Ref. 49, with permission.

of ionic species on the reduction in the absolute values of the electrophoretic mobility, the zeta potential, and the fixed charge density followed the order Na þ > K þ > Ca2 þ for the negatively charged BKCs, MMA-SPM NPs, and SLNs. The reverse order is true for the positively charged PBCA NPs. 6.3.3 Nanoparticle-Mediated Microchip Electrophoresis Microchip capillary electrophoresis emerged in the early 1990s as a novel approach to the high-speed separation of biological compounds, including DNA and proteins. Since the early development in this area, growth in the research field has exploded and now includes chemists and engineers focused on developing new and better microchips, as well as biologists and biochemists who have begun to apply this exciting and

APPLICATION OF NANOPARTICLES IN CE AND MCE

229

still relatively new methodology to real-world problems. A significant advance has been achieved in microchip electrophoresis using nanoparticles. 6.3.3.1 Gold Nanoparticles Wang and coworkers showed the use of AuNPs in conjunction with microchip electrophoresis to improve the selectivities between solutes and to increase the efficiency of separation.51 They coated the channel wall with a layer of PDDC, followed by coating with citrate-stabilized AuNPs. The resolution and plate numbers of the solutes were almost double in the presence of AuNPs. Such selectivity improvements reflected changes in the observed mobility accrued from interaction of solutes with the particle surfaces. They discussed that the addition of the AuNPs in the separation buffer in MCE adds another separation vector to the orthogonal electrophoretic vector. The coexistence of these two vectors resulted in resolution and plate number enhancement of the solutes. In this vein, Chang and coworkers tested DNA separation in AuNP-filled microfluidic channels by MCE. To avoid the aggregation of the AuNPs and allow strong interactions with the DNA molecules, AuNPs were modified with a polymer matrix, poly(ethylene oxide) (PEO), via noncovalent bonding.52 Coating the separation channels on a poly(methyl methacrylate) plate with PEO/AuNP composites was effective in improving reproducibility and selectivity (Figure 6.5). To obtain more hydrophilic and stable channels with high EOFs, Chen and coworkers described a layer-by-layer assembly technique to coat the PDMS microchip channel with PDDA and silica nanoparticles.53 The cationic polymer PDDA was strongly attached to the surface of the PDMS channel; negatively charged silica nanoparticles were then immobilized by electrostatic interactions. In a similar manner, they also modified the PDMS surfaces with AuNPs and polyethyleneimine.54 Both channels displayed a long-time stability and good reproducibility and selectivity. Dopamine and epinephrine were separated with good selectivity, employing the coated channels. An EOF-switchable PDMS microfluidic channel modified with cysteine has been developed by Zhang and coworkers, PDMS channels were coated with PDDA and gold nanoparticles with a layer-by-layer technique to immobilize cysteine.55 The electroosmotic mobility, and hence the observed mobility inside the PDDA-goldcoated channels, can be reversibly switched between the cathodic (high pH) and anodic direction (low pH) by varying the pH of the running buffer. This pH-dependent selectivity can be divided into three surface charge states depending upon the protonated degree of the amino and carboxyl groups of the cysteine: positively charged, neutral, and negatively charged. At pH 5.0, near the isoelectric point of the chemisorbed cysteine, the surfaces of the channels are neutral. When the pH is above 6.0 or below 4.0, the magnitude of the observed mobility varies within a narrow range. Separation of dopamine and epinephrine, as well as arginine and histidine, was performed employing these AuNP-modified channels. Liu et al. described a microchip reactor coated with an AuNP network entrapping trypsin.56 The reactor is designed for the proteolysis of low-level proteins and complex extracts originating from mouse macrophages. The nanostructured surface coating was assembled via a layer-by-layer electrostatic binding of PDDC and AuNPs. The

230

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.5 Separations of a mixture of equal volume of 10 mg mL1 DNA markers V and VI under different conditions using a three-layer coated PMMA microdevice. PEO(GNPs) (1.5%) solutions containing 0.5 mg ml1 EtBr were prepared in 100 mM glycine buffer, pH 9.1 in (a); 100 mM glycine buffer, pH 9.1, containing 1 M urea in (b); and 100 mM glycine–citrate buffer, pH 9.2 in (c). Current: 20 mA in (a) and (b), 30 mA in (c). The separations were conducted at 2400 V. Hydrodynamic injections were conducted by dipping the DNA sample with a 30 cm  350 mm ID capillary. The separations were conducted at 2800 V. Reproduced from Ref. 52, with permission.

PDDC/AuNP multilayer assembly constructed on the surface of a PET microfluidic chip offered a biocompatible interface with a large surface area, desirable for the controlled adsorption of trypsin. Due to a high concentration of trypsin confined to the microchannel, low levels of the standard protein samples are rapidly digested on the microchip reactor within a few seconds. Furthermore, the digestion of real protein mixtures isolated from mouse macrophages illustrated the performance of the online microchip bioreactor. The protein mixtures extracted from the mouse

APPLICATION OF NANOPARTICLES IN CE AND MCE

231

macrophages were efficiently identified by online digestion and LC–ESI-MS/MS analysis. This method is feasible for the characterization of various real protein extracts. Microfluidic-based bioreactors have become the focus of research interest for immunoassays and biomarker diagnostics. Lin et al.57 and Ahn et al.58 developed electroimmunosensing microchip bioreactors based on AuNPs for real-time measurement of antigen–antibody reactions. Electrochemical impedance spectroscopy was used by Lin et al. to detect the antigen–antibody interaction, where a 10-fold detection sensitivity was observed employing AuNPs. In Ahn’s approach, using a microfilter and microbeds, target proteins were immobilized in the detection zone of the microreactor where the microelectrodes were located. The immunoreaction was detected by measuring the electrical resistance between the microelectrodes using AuNPs with silver enhancement. The electrical resistance varied according to the concentration of the target antibody; the detection limit of the method was 10 ng mL1. In another electroimmunosensing microchip bioreactor, Ahn et al. utilized pillar-type microfilters within a reaction chamber and immunogold silver staining to amplify the electrical signal that corresponded to the immune complex.59 To demonstrate this approach, they simultaneously assayed three cancer biomarkers, namely, alpha-fetoprotein, carcinoembryonic antigen, and prostatespecific antigen (PSA), on the microchip. The electrical signal generated as a result of immunoreaction was measured and monitored within 55 min and the working range of the proposed microchip was 103–101 mg mL1 of the target antigen. A nanomosaic network of AuNPs for the detection of ultralow concentrations of proteins was reported by Girault and coworkers using two planar microelectrodes embedded in a microchip (Figure 6.6) that permitted generation of capacitive coupling to the nanomosaic system without the need for direct electrical contact with the channel.60 By tailoring the microchannel surface using a sandwich configuration of polyethylene terephthalate/AuNPs/poly(L-lysine), the surface charge was modified following the biomolecule interactions and monitored using a noncontact admittance technique. The main process governing the electrical properties of the sensor was the adsorption of charged protein onto a highly sensitive polyethylene terephthalate/ AuNPs/poly(L-lysine) sandwich network. Under experimental conditions, the b-lactoglobulin is negatively charged and its adsorption on the surface poly(L-lysine) layer caused a change in the charged region of the Gouy–Chapman layer. The surface charge is modified during adsorption of macromolecules and monitored using capacitive admittance tomoscopy. This nanodevice system behaves like a tunable capacitor and can be employed for the detection of any kind of molecule. The femtomolar detection of an anionic protein, such as b-lactoglobulin in phosphate buffered saline medium, was taken as an example. AuNPs with fully matched DNA duplexes on their surfaces aggregate together without molecular cross-linking at high salt concentrations. The mechanism of this noncross-linking interaction between the duplex with different DNA sequence-modified AuNPs and a duplex-modified flat gold surface has been explained by Hosokawa and coworkers in a recent study.61 They immobilized

232

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.6 SEM images of (a) the PET photoablated microchannel with a cross section of 45 mm  100 mm and a length of 1.4 cm; (b) trapezoidal section of the microchannel; and (c) the planar microelectrodes. Vertically, the two band electrodes beneath the horizontal flow channel. The detection zone represents the area where capacitive coupling takes place. (d) Side view of the contactless zone, which is about 5 mm and represents the thickness separation between the flow microchannel and the planar microelectrodes. Reproduced from Ref. 60, with permission.

15-base single-stranded DNA (ssDNA) on the surfaces of AuNPs with a diameter of 40 nm and on a flat gold substrate. AuNPs were hybridized with 15-base ssDNA at a low salt concentration. A microfluidic device was used for simultaneous delivery of the following three components onto the gold substrate: the duplex-modified AuNPs, 15-base ssDNA to be hybridized onto the substrate, and NaCl at a high concentration. Adsorption of the AuNPs onto the substrate was monitored using surface plasmon resonance imaging. When the AuNPs and the substrate had an identical sequence, the adsorption behavior was analogous to the aggregation behavior of AuNPs in test tubes. Furthermore, they also investigated cases in which the AuNPs and the substrate had completely different sequences, and obtained results suggesting the noncross-linking attraction force primarily depends on the terminal base pairs of the duplexes. They claimed that the main mechanism of the noncross-linking interaction is likely to be interduplex base stacking rather than formation of Holliday junctions.62 6.3.3.2 Magnetic Nanoparticles Zhang and coworkers described a microchip enzymatic microreactor based on the glass microchip with trypsin-immobilized superparamagnetic nanoparticles.63 Magnetic nanoparticles with a small size (d ¼ 50 nm) and strong magnetism were synthesized. At first, amine-functionalized magnetic nanoparticles are prepared

APPLICATION OF NANOPARTICLES IN CE AND MCE

233

applying a facile one-pot strategy. Magnetic nanoparticles are then functionalized with numerous aldehyde groups by treating the amine-functionalized magnetic nanoparticles with glutaraldehyde. Finally, immobilization of trypsin on aldehydefunctionalized magnetic nanoparticles is achieved through reaction of the aldehyde groups with the amine groups of the trypsin. The magnetic nanoparticles are then locally packed in the glass microchip by application of a strong magnetic field using a magnet to form an on-chip magnetic nanoparticle packing bed. Capability of the proteolytic microreactor is demonstrated by cytochrome c, bovine serum albumin, and myoglobin as model proteins. Complete protein digestion is achieved in a short time (10 s) under a flow rate of 5.0 mL min1. These results are expected to open up new possibilities for proteolysis analysis as well as a new application of magnetic nanoparticles. Replacement of nanoparticles and new microreactor construction is facile, and the packing bed can be used at least five times without any treatment. Since the preparation and surface functionality of magnetic nanoparticles is low-cost and reproducible, the preparation method and application approach of magnetic nanoparticles may find much potential in proteome research. Chen and coworkers described an integrated microfluidic sorting device that utilized sugar-encapsulated magnetic nanoparticles to separate a specific strain of bacteria from a mixture solution.64 In this system, a microfluidic device consisting of two inlets and an electromagnet or permanent magnet is constructed by a soft lithography process. The magnetic field generated by either the electromagnet or the permanent magnet is strong enough to attract the bacteria bound to the magnetic nanoparticles to cross the stream boundary of the laminar flow. The sorting efficiency is found to depend on both flow rate and strength of the magnetic field. The maximum sorting efficiency was measured to be higher than 90% with selectivity near 100%. The reactor was able to separate 1000 bacterial cells within 1 min with more than 70% sorting efficiency. However, the reactor has a drawback. The formation of large aggregates at high bacterial and nanoparticle concentrations could block the microchannels. A stimuli-responsive magnetic nanoparticle-based microsystem for diagnostic target capture and concentration has been developed for microfluidic lab card settings by Stayton and coworkers.65 Telechelic poly(N-isopropylacrylamide) (PNIPAAm) polymer chains were synthesized with dodecyl tails at one end and a reactive carboxylate at the opposite end by the reversible addition fragmentation transfer technique. These PNIPAAm chains self-associate into nanoscale micelles and are used as dimensional confinements to synthesize the magnetic nanoparticles. The resulting superparamagnetic nanoparticles exhibited a Fe2O3 core (5 nm) with a layer of carboxylate-terminated PNIPAAm chains as a corona on the surface. The carboxylate group was used to functionalize the magnetic nanoparticles with biotin and subsequently with streptavidin. The functionalized magnetic nanoparticles can be reversibly aggregated in solution as the temperature is cycled through the PNIPAAm lower critical solution temperature (LCST). While the magnetophoretic mobility of the individual nanoparticles below the LCST is negligible, the aggregates formed above the LCST are large enough to respond to an applied magnetic field. Magnetic nanoparticles can associate with biotinylated targets as individual particles,

234

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

and then subsequent application of a combined temperature increase and magnetic field can be used to magnetically separate the aggregated particles onto the poly (ethylene glycol)-modified polydimethylsiloxane channel walls of a microfluidic device. When the magnetic field is turned off and the temperature is reversed, the captured aggregates redisperse into the channel flow stream for further downstream processing (Figure 6.7). The dual magnetic- and temperature-responsive nanoparticles can thus be used as soluble reagents to capture diagnostic targets at a controlled time point and channel position. They can then be isolated and released after the nanoparticles have captured the target molecules, overcoming the problem of low magnetophoretic mobility of the individual particles while retaining the advantages of

FIGURE 6.7 Particle capture and release scheme (a) and the corresponding micrographs (b). The PNIPAAm mNP capture/release was demonstrated in PEGylated PDMS microfluidic channels whose channel width was 500 mm. The magnetic field was introduced by embedding a magnet at the lower side of the channel. The mNP solution (4 mg mL1) was injected into the channels with a constant flow (1 mL min1) during the entire experiment. mNPs are soluble and free flowing in the PEGylated channels when temperature is below the LCSTof PNIPAAm. As they flow into the heated region, the temperature is above the LCST of the PNIPAAm, and the mNPs aggregate but do not stick to the nonfouling, PEGylated channel walls in the absence of an applied magnetic field. mNPs are captured onto the PEGylated channel walls only when the temperature is raised above the LCST and the magnetic field is applied. The reversal of the temperature and applied magnetic field results in the redissolution of the aggregated magnetic nanoparticles and their diffusive re-entry into the flow stream. Reproduced from Ref. 65, with permission.

APPLICATION OF NANOPARTICLES IN CE AND MCE

235

a high surface to volume ratio and faster diffusive properties during target capture. These dual magnetic- and temperature-responsive magnetic nanoparticles are designed to facilitate diagnostic target isolations and assays in point-of-care microfluidic diagnostic devices. A faster method of preparing an easily replaceable protease microreactor for a microchip application is described.66 Magnetic particles coated with poly(Nisopropylacrylamide), polystyrene, poly(2-hydroxyethyl methacrylate-co-ethylene dimethacrylate), poly(glycidyl methacrylate), [(2-amino-ethyl)hydroxymethylen]biphosphonic acid, or alginic acid, with immobilized trypsin, were utilized for heterogeneous digestion. To obtain the highest digestion efficiency, submicrometer spheres were organized by an inhomogeneous external magnetic field perpendicular to the direction of the channel. Kinetic parameters of the microchip-immobilized magnetic enzyme reactor were determined. The capability of the proteolytic reactor was tested with five model glycoproteins, ranging in molecular mass from 4.3 to 150 kDa. Digestion efficiency of proteins in various conformations was investigated using SDSPAGE, HPCE, RP-HPLC, and MS. The compatibility of the microchip-immobilized magnetic enzyme reactor system with total and limited proteolysis of high molecular weight glycoproteins was discussed, and subsequently paves the way for automated, high-throughput proteomic microchip applications. 6.3.3.3 Zeolite Nanoparticles Zeolite nanoparticles have drawn much interest in microchip-based applications due to their excellent properties including large external surface areas compared to conventional zeolite crystals, high dispersibility in both aqueous and organic solutions, high thermal and hydrothermal stabilities, and tunable surface properties such as adjustable surface charge and hydrophilicity/hydrophobicity. Yang and coworkers fabricated an enzymatic microreactor based on the poly(methyl methacrylate) (PMMA) microchip surface modified with zeolite nanoparticles.67 The hydrophobic–hydrophobic bonding interaction is used to immobilize zeolite nanoparticles on the PMMA surface due to the surface hydrophobic property of silicalite-1, with an aluminum-free framework. Immobilization of proteins on the silicalite-1/PMMA surface occurred via a silicon–oxygen–silicon bridge. Using MALDI-TOF mass spectrometry, the silicalite-1/PMMA microreactor provided an efficient digestion of cytochrome c and bovine serum albumin at a flow rate of 4.0 mL min1, affording a reaction time of less than 5 s. An on-chip microreactor has been developed for the acceleration of protein digestion through the construction of a nanozeolite-assembled network.68 The nanozeolite microstructure was assembled using a layer-by-layer technique based on PDDA and zeolite nanocrystals. Adsorption of trypsin in the nanozeolite network was theoretically studied based on the Langmuir adsorption isotherm model. It was found that the controlled trypsin-containing nanozeolite networks assembled within a microchannel could act as a stationary phase with a large surface to volume ratio for the highly efficient proteolysis of both proteins at low levels and with complex extracts. The maximum proteolytic rate of the adsorbed trypsin was measured to be 350 mM min1 mg1, much faster than that in solution. Moreover, due the large

236

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

surface to volume ratio and biocompatible microenvironment provided by the nanozeolite-assembled films, as well as the microfluidic confinement effect, lowlevel proteins down to 16 fmol per analysis were confidently identified using the as-prepared microreactor within a very short residence time, coupled to matrixassisted laser desorption time-of-flight mass spectrometry. The approach was further demonstrated in the identification of complex extracts from mouse macrophages integrated with two-dimensional liquid chromatography–electrospray ionizationtandem mass spectrometry. This microchip reactor is promising for the development of a facile means of protein identification. The electrophoretic mobility of AuNPs with different sizes was studied using platinum-coated alumina membranes in a microfluidic device.69 The electrophoretic mobility of gold nanoparticles depends on the nature of the mobile phase and interfacial properties of the alumina channels. The transport performance of nanoparticles are improved with the addition of SDS to the mobile phase because SDS not only decreases the physical adsorption of gold nanoparticles onto the nanochannel wall of the alumina membrane but also reduces the thickness of the electric double layer (decreasing apparent particle size). When the alumina membranes were modified with 6-aminohexanoic acid, it was further confirmed that the physical adsorption played a key role in the electrophoretic mobility of AuNPs. 6.3.3.4 TiO2 Nanoparticles Cremer and coworkers developed a facile and simple method for patterning metal nanoparticle films of arbitrary geometry inside sealed PDMS/TiO2/glass microfluidic devices and illustrated the ability to biofunctionalize these films with ligands for protein capture (Figure 6.8).70 A 6.0 nm TiO2 film is first deposited onto a planar Pyrex or a silica substrate subsequently bonded to a PDMS mold. UV light is then exposed through the device to reduce the metal ions in an aqueous solution to create a monolayer-thick film of metal nanoparticles of varying sizes by independently controlling the solution conditions in each microchannel where the film is formed. In terms of simplicity and design flexibility, this method has advantages over multiphase laminar flow-based assays. In addition, functionalizing nanoparticle films inside microfluidic channels may afford new opportunities for biosensors or screening assays. The ability to address individual ligands atop nanoparticle films inside microfluidic devices could be combined with such technologies as transmission surface plasmon resonance spectroscopy or surface-enhanced fluorescence, allowing the development of powerful lab-on-a-chip devices with label-free detection or fluorescence detection with enhanced sensitivity. 6.3.3.5 Polymer Nanoparticles Langer and coworkers demonstrated the interaction of PEGylated poly(lactic acid) nanoparticles/microparticles and similar particles conjugated to aptamers that recognize the transmembrane prostate-specific membrane antigen (PSMA), with cells seeded in microchannels.71 Binding of particles to the cells that expressed or did not express the PSMA (LNCaP or PC3) was evaluated with respect to changes in fluid shear stress, PSMA expression on target cells, and particle size. At static and low shear,

APPLICATION OF NANOPARTICLES IN CE AND MCE

237

FIGURE 6.8 Schematic diagram for the deposition of a silver nanoparticle film. First, a AgNO3 solution is introduced into the microchannel. Next, UV radiation is passed through a photomask onto the backside of the TiO2 thin film. Ag þ ions adsorbed at the interface are selectively reduced by photoelectrons, which grow into nanoparticle films. This process can be used in combination with thiol chemistry inside sealed microfluidic channels to address surface chemistries in almost any desired location or pattern. Reproduced from Ref. 70, with permission.

238

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

nanoparticle–aptamer bioconjugates selectively adhered to LNCaP, and not PC3 cells, but not under higher shear conditions. Control nanoparticles and microparticles lacking aptamers and microparticle–aptamer bioconjugates did not adhere to LNCaP cells, even under very low shear conditions. This model can be used to determine the ideal particle size and ligand density on the particle surface for binding to target cells under fluid flow conditions. The author claims that similar microfluidic models can be designed to simultaneously test multiple parameters and therefore maximally optimize the physical and chemical properties of therapeutic and diagnostic particles prior to their in vivo evaluation. Ishihara and coworkers have reported a method of preparing polymer nanoparticles for the selective capture of a specific protein from a mixture with high effectiveness.72 The nanoparticle surface was covered with hydrophilic phosphorylcholine groups and active ester groups for easy immobilization of antibodies. Phospholipid polymers (PMBN) composed of 2-methacryloyloxyethyl phosphorylcholine, n-butyl methacrylate, and p-nitrophenyloxycarbonyl polyethyleneglycol methacrylate were synthesized for the surface modification of poly(L-lactic acid) nanoparticles. Surface analysis of the nanoparticles using laser-Doppler electrophoresis and X-ray photoelectron spectroscopy revealed that the surface of nanoparticles was covered with PMBN. Protein adsorption was evaluated with regard to nonspecific adsorption onto the nanoparticles that was effectively suppressed by phosphorylcholine groups. The immobilization of antibodies on nanoparticles was carried out under such physiological conditions as to ensure specific binding of antigens. The antibody immobilized on the nanoparticles exhibited high activity and strong affinity for the antigen similar to that exhibited by an antibody in a solution. The selective binding of a specific protein as an antigen from a protein mixture was relatively high compared to that observed with conventional antibody-immobilized polymer nanoparticles. The authors found that the nanoparticles with both phosphorylcholine and active ester groups for antibody immobilization have strong potential for use in highly selective separation based on biological affinities between biomolecules. 6.3.3.6 Carbon Nanotubes In recent years, CNTs have attracted much attention as a novel monolithic stationary phase for high-performance liquid chromatography and capillary electrochromatography.73 Both the retention and the separation efficiencies were enhanced by incorporation of CNTs into the stationary phase. Carboxylic CNTs were used as a pseudostationary phase in CE, as described above, where electrodiffusion and adsorption were greatly suppressed between the capillary wall and solutes and thus led to better peak shapes of isomers. Nevertheless, CNTs have attracted much interest for their biocompatibility, especially for conjugation with proteins. Kong and coworkers tested a PMMA microfluidic chip for enantioseparation of tryptophan enantiomers using bovine serum albumin-conjugated CNTs as stationary phase.74 Successful separation of tryptophan enantiomers was achieved within 70 s with a resolution factor of 1.35, utilizing a separation length of 32 mm. The theoretical plate number was 24,000 plates m1 for D-tryptophan and 7700 plates m1 for L-tryptophan, respectively.

APPLICATION OF NANOPARTICLES IN CE AND MCE

239

6.3.3.7 AuNP-Mediated On-Chip Preconcentration The use of the MCE method in trace analysis of analytes are somewhat restricted owing to the small sample injection volume and short path length available for optical measurements.75,76 To overcome this limitation, one promising alternative is to preconcentrate the sample prior to detection. In order to further enhance the separation selectivity and detection sensitivity of the preconcentration method, one can modify the separation and preconcentration buffers with charge-stabilized metal nanoparticles. A detailed description of the experimental procedures and principles associated with this method are included in a paper by Shiddiky and Shim.77 They demonstrated an on-chip preconcentration method for DNA preconcentration, separation, and EC detection employing AuNP-modified buffers and electrodes (Figure 6.9). The device consisted of three parallel channels: the first two were the field-amplified sample stacking (FASS) and subsequent field-amplified sample injection (FASI) steps, the third was for microchip gel electrophoresis with electrochemical detection step. The stacking and separation buffers containing the hydroxypropyl cellulose (HPC) matrix were modified with AuNPs. The conducting polymer/AuNP-modified electrode was used to detect amplified DNA based on their direct oxidation in a solution phase. When both the FASS and the FASI methods were sequentially applied, total sensitivity was improved by approximately 25,000-fold compared to a conventional MGE-ED analysis. In the FASS step, DNA stacking occurs at the boundary between the low- and high-conductivity buffers due to the sharp decreases in the velocity of DNA migrating from the higher field strength (in low-conductivity sample buffer) to a relatively lower (in high-conductivity stacking buffer) zone. In the FASI step, stacking occurs in low-pH/high-conductivity stacking buffers containing HPC/AuNP matrix. DNAs were stacked first at the boundary between the water and the low-pH/highconductivity buffer, possibly for the same reason as mentioned above for the FASS step. The initially stacked DNA samples were subjected to a second stacking upon interaction with the HPC/AuNP matrix present at the concentration boundary. The second stacking occurred mainly because of retardation of the DNA movement by the HPC/AuNP matrix. In addition, the presence of trisodium citrate ions and citratestabilized AuNPs in the stacking buffer solution might have affected the enhancement of stacking efficiency due to three possibilities, the increase in relative conductivity of the stacking buffer; adsorption of DNA onto HPC/AuNP surfaces through the protruding part of the HPC matrix that adsorbed onto the AuNP surfaces; or minimization of the DNA adsorption on the wall of the channel in the presence of strongly adsorbed AuNPs. 6.3.3.8 Colloidal Au Self-Assembly in MCE Separation Self-assembled colloidal crystals have attracted interest in the fields of separation and polymer dynamics owing to the relatively simple fabrication and uniquely ordered porous structure. The microfluidic-based colloidal self-assembly technique dramatically reduces preparation time and avoids formation of cracks caused by drying. Using this technique, robust colloidal lattices of various pore sizes and materials can be readily incorporated into microfluidic devices for rapid separation of biomolecules

240

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.9 Schematic illustration of the preconcentration, separation, and electrochemical detection of DNA. (1) Sample loading: a voltage of þ 100 V cm1 was applied to R1 while R2 was grounded and R3 was left floating. (2) FASS step: (i) a potential of þ 200 V cm1 was applied to R2 for 130 s while R3 was grounded and R1 and R4 were left floating. (ii) Water/ stacking buffer injection: during FASS step, a water plug was injected hydrodynamically from R5 to R4 at a flow rate of approximately 0.1 mL min1 for 110 s. Thereafter, the stacking buffer was injected for 60 s at the same flow rate. (3) FASI step: initially preconcentrated sample was then injected into channel 2 by applying a voltage of approximately 150 V cm1 to R3 with R5 grounded, leaving all other reservoirs floating. (4) Sample loading and injection: a voltage of þ 100 V cm1 was applied to R7 for 40 s, while R4 was grounded and R3, R5, R6, and R8 remained floating. Injection was effected by applying an injection voltage of þ 200 V cm1 for 5 s to R4. (5) Separation and detection: MGE-ED was performed by applying separation field strength of 340 V to R6 and þ 1500 V to R8 with R3, R4, R5, and R7 floating. Amperometric detection potential: þ 0.8 V versus Ag/AgCl. Reproduced from Ref. 77, with permission.

APPLICATION OF NANOPARTICLES IN CE AND MCE

241

FIGURE 6.10 Fabrication and characterization of self-assembled colloidal arrays within microfluidic systems. (a) Schematic illustration of microfluidic colloidal self-assembly in a one-dimensional separation microchip (PDMS chip layout: (1) buffer, (2) sample, (3) sample waste, (4) buffer waste). (b) Optical micrograph of a translucent 0.9 mm silica sphere array growing inside a microchannel (as indicated by the left arrow), showing a convex evaporation interface at the channel opening. (c) Digital images of a PDMS chip packed with 0.9 mm silica spheres before drying. When illuminated by white light vertically from the bottom, the array exhibited monochromic transmitted light at various angles due to Bragg diffraction. (d and e) SEM images of a matrix of 330 nm silica spheres at different magnifications. (f) SEM image of a hexagonally closed packed 2 mm PS colloidal array fabricated within a microchannel. The arrows indicate lattice defects. The scale bars are 200, 2, and 10 mm in (d–f), respectively. Reproduced from Ref. 78, with permission.

with a wide size distribution. Recently, Zeng and Harrison developed a facile, microfluidic, colloidal self-assembly strategy to create ordered, robust, three-dimensional nanofluidic sieves within microfluidic devices, with which a fast separation of DNA and proteins of a wide size range was achieved.78 Fabrication and characterization of self-assembled colloidal arrays within microfluidic systems is schematically illustrated in Figure 6.10. This approach offers a significantly greater assembly speed over conventional colloidal deposition approaches, which usually take tens of hours to days to construct colloidal arrays within microdevices. The high assembly speed can be attributed to the large surface area of the protruding evaporation interface. Compared to the design of microfabricated barrier structures with a gap size smaller than the particles to be trapped, this approach uses the air–liquid interface as a virtual frit to retain the microspheres, offering advantages in terms of a much larger

242

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

evaporation area and simplified device fabrication. The high-speed assembly allows the packing of large particles of various materials. Separation of Proteins and DNA Nanometer-sized interstices in the close-packed sphere lattice create a nanofluidic sieve that consists of voids interconnected by narrower pores, with an equivalent diameter approximately 15% of the sphere size. Molecules experience a loss in entropy by steric constriction while traveling through a constraining pore.79 Thus, the porous structure imposes periodically modulated, free-energy barriers to molecular transport that are presumably responsible for sizedependent separation of molecules with dimensions comparable to the narrower pore sizes. Harrison’s approach thus provided separation of biomolecules with a wide size distribution, ranging from proteins (20–200 kDa) to dsDNA (0.05–50 kbp). For example, sieving-based protein separation in these porous beds was an obvious objective, made challenging by the very small pore size required. Figure 6.11a exhibits separation of SDS-denatured protein markers using differently sized silica particles. Four proteins of 20–205 kDa were separated in a matrix of 330 nm silica spheres, as shown in Figure 6.11a. The resolution was 2.64, between the 20.1 and the 116 kDa proteins, and 3.92, between the 116 and the 205 kDa proteins, respectively, which indicate a better size selectivity for larger proteins than smaller ones. Four consecutive runs of a low DNA mass ladder containing six equimolar dsDNA fragments from 100 to 2000 bp were performed in a 0.9 mm silica particle array; the electropherograms are superimposed in Figure 6.11b. In this case, the detector was located 5.0 mm from the injection point. The device was operated for approximately 5 h before this reproducibility test. The standard deviation of the migration time between these runs was less than 2%, indicating the stability of the self-assembled colloidal sieve under the applied electric field. The bed stability is likely a result of van der Waals forces between silica particles and the contribution from hydrogen bonds between the surfaces of closepacked silica particles.

FIGURE 6.11 (a) Separation of four proteins using 330 nm silica particles (E ¼ 19.1 V cm1, L ¼ 8 mm): (1) trypsin inhibitor, 20.1 kDa; (2) BSA, 66 kDa; (3) b-galactosidase, 116 kDa; and (4) myosin, 205 kDa. (b) Four consecutive runs of a low DNA mass ladder obtained using 0.9 mm silica beads in a device that had been operated for approximately 5 h (E ¼ 19.2 V cm1; L ¼ 5 mm, and the microchip has approximately100 mm wide and approximately 20 mm deep microchannels with a cross-injection design). Reproduced from Ref. 78, with permission.

APPLICATION OF NANOPARTICLES IN CE AND MCE

243

6.3.3.9 Surface Displacement Reactions on Colloidal Gold in Microfluidic Channels Spatial imaging of the fluorescence from derivatized organomercaptans in a flowing microfluidic channel has been developed by Bohn and coworkers to monitor the surface displacement kinetics of organomercaptans to the surface of AuNPs.80 In these experiments, surface displacement of the tagged reagent to the surface of the colloid is accompanied by a fluorescence decrease associated with quenching of fluorescence from the adsorbed fluorophore by coupling to the surface plasmon of Au. Several characteristic features of the hybrid nanofluidic–microfluidic devices have been exploited while making these measurements. The ability to electrokinetically inject fluid across a nanocapillary array membrane separating two microfluidic channels provides a convenient means of reaction initiation, and once begun, the constant velocity of the microfluidic flow allows a distance–time conversion so that spatial images can be translated directly into temporal plots of reactant concentrations. As test cases, the surface displacement kinetics of two fluorescently tagged species, a small organomercaptan (SAMSA), and an octapeptide (CDWAK WAD) have been measured. Excellent fits of the kinetic data to Langmuir kinetic models were obtained in all cases, obviating the need to invoke more complicated kinetic models. However, the surface displacement rate constants determined for thiol surface displacement on Au nanoparticles were roughly one order of magnitude larger than those measured for similar thiols on planar Au surfaces, indicating faster kinetics in the colloid–adsorbate system. These results highlight the utility of colloidal Au nanoparticles as molecular carriers for the sequestration of analytes, allowing the manipulation of mass-limited samples and ultimately the capture and delivery of selected analytes from a microfabricated device to an off-line detector. 6.3.3.10 Microchip-Based Bio-Barcode Assay A single disposable chip has been developed for carrying out a multistep process that employed nanoparticles for single protein marker detection.81 To illustrate the capability of this bio-bar code assay (BCA), Liu et al tested the presence of PSA in buffer solution and goat serum. The proposed BCA protocol can be divided into two stages—target protein separation (stage 1) and barcode DNA detection (stage 2). Major steps are illustrated in Figure 6.12. In stage 1, magnetic microparticles are introduced into a microfluidic channel reactor. The sample fluid is then flowed into the channel along with functionalized AuNP probes. Hybridized magnetic microparticle– protein–AuNPs sandwiches form when target proteins are present in the sample fluid. The magnetic microparticles and magnetic microparticle–protein–AuNPs conjugates are then immobilized to the channel wall with a magnet while the supernatant is washed away with several column volumes of buffer. Subsequently, barcode DNA strands are released from the AuNP probes by applying deionized water that dissociates the barcode DNA from the AuNPs. For stage 2, the released barcode DNA strands are transferred to a detection channel. The bottom surface of the detection channel is functionalized with capture strands that are half complementary to the barcode DNA. A second set of AuNP probes, functionalized with the remaining complimentary sequence, is then introduced. It should be noted that the AuNP probes used here are

244

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.12 Implementation of the bio-barcode assay within a microfluidic device. First, magnetic particles functionalized with monoclonal PSA antibodies are introduced into the separation area of the chip. The particles are then immobilized by placing a permanent magnet under the chip, followed by introduction of the sample and AuNPs that are decorated with both polyclonal antibodies and barcode DNA. The barcode DNA is then released from the AuNPs and is transported to the detection area of the chip. The detection area of the chip is patterned with capture DNA. Salt and a second set of AuNPs functionalized with complementary barcode DNA sequences are introduced into the detection area to allow hybridization. Finally, the signal from the AuNPs is amplified using silver stain. Reproduced from Ref. 81, with permission.

not the same as the AuNP probes employed in the previous stage. The barcode DNA molecules allow the functionalized NP probes to be hybridized to the surface. Chipimmobilized AuNP probes thus signify the presence of the barcode DNA. The amount of chip-immobilized AuNP probes can be detected with a number of methods. One method involves detecting light scattering off the surface-bound AuNP probes. Exposing the captured gold AuNP probes to a silver staining solution further enhances the detectable optical signal.82 Gold serves as a catalyst for silver staining, hence enlarging the size of the gold nanoparticles. Enhanced particles are visible to the naked eye and can be detected using commercially available scanners. Detection was accomplished at PSA concentrations as low as 500 aM. This corresponds to only 300 copies of protein analytes using 1 mL total sample volume. 6.3.3.11 Microfluidic Fabrication of Biopolymer Micro- and Nanoparticles Due to the advantages of the formation of polymer-based microparticles using microdevices over conventional emulsification techniques, with respect to size control and polydispersity of the final sample, fabrication of simple, reliable, and robust methods in microfluidic devices has gained interest. Kumacheva and colleagues

APPLICATION OF NANOPARTICLES IN CE AND MCE

245

described two methods for the formation of monodispersed, polymer-based microparticles in microfluidic devices from biologically derived and synthetic polymers.83,84 In the former method, they reported an approach generating capsules of biopolymer hydrogels. Droplets of an aqueous solution of a biopolymer were emulsified in an organic phase comprising a cross-linking agent. Polymer gelation was achieved in a microfluidic chip by diffusion-controlled, ionic cross-linking of the biopolymer, following the transfer of the cross-linking agent from the continuous phase to the droplets. Collecting particles in a large pool of cross-linking, agent-free liquid quenched gelation. Microgel structure, from capsules to gradient microgels to particles with a uniform structure, was controlled by varying residence time of the droplets on the microfluidic chip and concentration of the cross-linking agent in the continuous phase. The described approach was applied to prepare capsules of several polysaccharides such as alginate, carrageenan, and carboxymethylcellulose. For continuous and scalable production of core shell droplets and polymer capsules in microfluidic devices, they employed a capillary instability-driven breakup of a liquid jet formed by two immiscible fluids to achieveprecise control over emulsification of each liquid, allowing the production of highly monodispersed core shell droplets with a predetermined core diameter and shell thickness. Polymer particles with various shapes and morphologies, including spheres, truncated spheres and hemispheres, and single and multicore capsules were obtained via fast photopolymerization of monomeric shells. Lin and coworkers described the manipulation of Ca-alginate microspheres, using a microfluidic chip, for the encapsulation of gold nanoparticles, based on hydrodynamic focusing on the formation of a series of self-assembling sphere structures, the so-called water-in-oil (w/o) emulsions, in the cross-junction microchannel.85 These fine emulsions, consisting of aqueous Na-alginates, are then dripped into a solution of 20% calcium salt to accomplish Ca-alginate microspheres in an efficient manner. Experimental data show that microspheres with diameters ranging from 50 to 2000 mm, with a variation less than 5%, were precisely generated. The size and gap of the droplets are tunable by adjusting the relative sheath/sample flow rate ratio. Furthermore, the application of these particles toward encapsulated AuNPs was successfully performed. Developed microfluidic chip fabrication and setup are proficient and easily programmed to generate a large set of ordered Ca-alginate microspheres. Uniform, spherical, and molecularly imprinted polymer beads were prepared via controlled suspension polymerization in a spiral-shaped microchannel using mineral oil and a perfluorocarbon liquid as continuous phases.86 Monodisperse droplets containing the monomers, template, initiator, and porogenic solvent were introduced into the microchannel; particles of uniform size were produced by subsequent UV polymerization, quickly and without wasting polymer materials. Change in flow conditions within the microfluidic device allowed for variation in droplet/particle size. The diameter of the resulting products typically had a coefficient of variation below 2%. The specific binding sites created during the imprinting process were analyzed by radioligand binding analysis. The molecularly imprinted microspheres produced in the liquid perfluorocarbon continuous phase had a higher binding capacity compared to the particles produced in the mineral oil continuous phase, though it should be noted

246

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

FIGURE 6.13 Diagram showing the two processes used for producing alginate microparticles and nanoparticles. (a) In process 1, the alginate solution and the CaCl2 solution are injected into the channel. The two solutions mix together before the droplet formation with DMC as the continuous phase. As the droplets flow downstream, two processes are competing, the water diffuses out of the drops that harden due to the alginate cross-linking reaction. The particles are collected in a CaCl2 solution to achieve full extension of cross-liking and to store the particles. (b) In process 2, the same alginate solution is injected via the two Y-shaped inlet channels that join together to form the main flow channel. DMC is injected further downstream via the secondary Y-shaped channels to generate alginate droplets at their junction with the main flow channel. As the droplets flow downstream, the water is diffusing from the drops into the DMC. The condensed alginate particles are collected in a highly concentrated CaCl2 solution. Reproduced from Ref. 87, with permission.

that the aim of this study was not to optimize or maximize imprinting performance but rather to demonstrate broad applicability and compatibility with known MIP production methods. The successful imprinting against a model compound using two very different continuous phases (one requiring a surfactant to stabilize the droplets, the other not) demonstrates the wide-ranging of this approach. Rondeau and Cooper-White described a microfluidic-based methodology for the synthesis of monodispersed biopolymer (alginate) particles utilizing a multiphase microfluidic device that relied upon a novel pseudoequilibrium solvent diffusion process during laminar flow in PDMS-based microdevices.87 The two alternative routes have been explored in detail in order to understand the parameters influencing the resultant product, solvent diffusion, and competing condensation and/or crosslinking reactions (Figure 6.13). A detailed description of the experimental procedures and principles associated with this method, including the dependence of droplet generation, solvent diffusion, and cross-linking kinetics on changes in device and flow parameters and on polymer solution properties, is available in Ref. 87. These microdevice formats were used to make microparticles of controlled size and polydispersity and to reproducibly produce nanoparticles from alginate. The results indicate the process will allow production of particles from any synthetic or biologically derived polymer that can be solvated within a liquid partially miscible with

247

CONCLUSIONS

another liquid. The size of particles can be explicitly controlled through varying the relative solvency of the two liquids, physical properties of the polymer solution, including polymer concentration, configuration, ionic state, and solvent conditions, and processing parameters, including flow rate, temperature, and device design. A sensor for characterization of nanoparticle colloidal suspensions employing a diffraction grating under total internal reflection for investigation of nanodisperse fluids, passing through an integrated microfluidic channel, has been developed.88 Dispersions containing polymeric, metallic, and ferromagnetic nanoparticles were studied and accurately determined in real time with the specific refractive index for nanoparticle suspension and nanoparticle concentration. Nanoparticle concentrations were calculated at a resolution of 0.3–0.5 wt % for polymeric nanoparticles, 0.03–0.05 wt % for metallic nanoparticles, and 0.05–0.1 wt % for ferromagnetic nanoparticles. This translated to effective refractive indices determined with an accuracy of 7  104 for the polymeric and 2  104 for the metallic and ferromagnetic dispersions. 6.3.3.12 In-Channel Modification of Au Microelectrodes Enhancing the coulometric efficiency (Ceff) of amperometric detection interfaced with capillary electrophoresis is of considerable importance. The Ceff is defined as the percentage of the injected analyte detected at the amperometric detector, governed by the following equation: Ceff ¼ ðNd =Ni Þ  100%

ð6:11Þ

where Nd and Ni correspond to the number of moles detected and injected, respectively. Nd can be calculated based on Faraday’s equation: QC ¼ nFNd

ð6:12Þ

where n is the number of electrons, F is Faraday’s constant, and Qc the total charge (Qc ¼ i  t). Qc can be determined by integrating the area of the corresponding peak in the electropherogram. However, a smooth surface restrains the interaction between the analytes and the electrode, which results in lower sensitivity. Jankowiak and coworkers described a method for in-channel electrochemical deposition of gold particles on a PDMS/glass microfluidic device in order to vertically increase the surface area of the Au-sensing microelectrode.89 The modified electrodes provided well-resolved separation of dopamine and catechol in less than 60 s with enhanced Ceff; no peak broadening was observed, which can be attributed to vertical enlargement of the electrode’s size. The electroplated Au microelectrode offered a stable detection background current (<2 nA) with minimal noise level (<15 pA), which lowered the LOD by a factor of 2 for dopamine and a factor of 3 for catechol. 6.4 CONCLUSIONS Nanoparticles offer unique opportunities for designing excellent separation media for analyte separation in CE and MCE. The studies described above demonstrate the broad

248

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

potential of nanoparticles for new separation media with excellent selectivity and efficiency. Although significant advances have been made in the area of nanoparticles, many theoretical and technical problems still need to be solved. The successful realization of these new systems requires proper attention to aggregation of nanoparticles that commonly controls the performances of the method. The remarkable selectivity of the new nanoparticle-based CE and MCE protocols opens up the possibility for detecting proteins, DNA, peptides, amino acids, or numerous target agents that cannot be separated by conventional methods. All these approaches with various nanoparticles have been designed with nanometer dimensions and may be fully implemented in ultrasensitive detection and monitoring of biological events, where the role of mechanics in these events will be fully appreciated. REFERENCES 1. Wallingford, R. A.; Ewing, A. G. Capillary electrophoresis. Adv. Chromatogr. 1989, 29, 1–76. 2. Hayat, M. A.,Ed. Colloidal Gold: Principles. In Methods and Applications; Academic Press: 1991. 3. Schmid, G. G., Ed. Clusters and Colloids; VCH: Weinheim, Germany, 1994. 4. Shiddiky, M. J. A.; Kim, R.-E.; Shim, Y.-B. Microchip capillary electrophoresis with a cellulose-DNA-modified screen-printed electrode for the analysis of neurotransmitters. Electrophoresis 2005, 26, 3043–3052. 5. Shiddiky, M. J. A.; Won, M.-S.; Shim, Y.-B. Simultaneous analysis of nitrate and nitrite in a microfluidic device with a Cu-complex-modified electrode. Electrophoresis 2006, 27, 4545–4554. 6. Lee, K.-S.; Shiddiky, M. J. A.; Park, S.-H.; Park, D.-S.; Shim, Y.-B. Electrophoretic analysis of food dyes using a miniaturized microfluidic system. Electrophoresis 2008, 29, 1910–1917. 7. Fujimoto, C.; Muranaka, Y. Electrokinetic chromatography using nanometer-sized silica particles as the dynamic stationary phase. J. High Resolut. Chromatogr. 1997, 20, 400–402. 8. Wang, Y.; Ouyang, J.; Baeyens, W. R. G.; Delanghe, J. R. Use of nanomaterials in capillary and microchip electrophoresis. Expert Rev. Proteomics 2007, 4, 287–298. 9. Nilsson, C.; Birnbaum, S.; Nilsson, S. Use of nanoparticles in capillary and microchip electrochromatography. J. Chromatogr. A 2007, 1168, 212–224. 10. Nilsson, C.; Nilsson, S. Nanoparticle-based pseudostationary phases in capillary electrochromatography. Electrophoresis 2006, 27, 76–83. 11. Guihen, E.; Glennon, J. D. Nanoparticles in separation science—recent developments. Anal. Lett. 2003, 36, 3309–3336. 12. Lin, Y.-W.; Huang, M.-F.; Chang, H.-T. Nanomaterials and chip-based nanostructures for capillary electrophoretic separations of DNA. Electrophoresis 2005, 26, 320–330. 13. Zhang, Z.; Wang, Z.; Liao, Y.; Liu, H. Applications of nanomaterials in liquid chromatography: opportunities for separation with high efficiency and selectivity. J. Sep. Sci. 2006, 29, 1872–1878.

REFERENCES

249

14. Manz, A.; Graber, N.; Widmer, H. M. Miniaturized total chemical analysis systems: a novel concept for chemical sensing. Sens. Actuators B 1990, 1, 244–248. 15. Kutter, J. P.; Geschke, O. In Microsystem Engineering of Lab-on-a-Chip Devices; Geschke, O.; Klank, H.; Tellemann, P., Eds.; Wiley-VCH: Weinheim, Germany, 2004. 16. West, J.; Becker, M.; Tombrink, S.; Manz, A. Micro total analysis systems: latest achievements. Anal. Chem. 2008, 80, 4403–4419. 17. Dittrich, P. S.; Tachikawa, K.; Manz, A. Micro total analysis systems. Latest advancements and trends. Anal. Chem. 2006, 78, 3887–3908. 18. Wang, J. Real-time electrochemical monitoring: toward green analytical chemistry. Acc. Chem. Res. 2002, 35, 811–816. 19. Shiddiky, M. J. A.; Park, D.-S.; Shim, Y.-B. Detection of polymerase chain reaction fragments using a conducting polymer-modified screen-printed electrode in a microfluidic device. Electrophoresis 2005, 26, 4656–4663. 20. Shiddiky, M. J. A.; Rahman, M. A.; Park, J.-S.; Shim, Y.-B. Analysis of polymerase chain reaction amplifications through phosphate detection using an enzyme-based microbiosensor in a microfluidic device. Electrophoresis 2006, 27, 2951–2959. 21. Linhardt, R. J.; Toida, T. Ultra-high resolution separation comes of age. Science 2002, 298, 1441–1442. 22. Dolnik, V.; Liu, S.; Jovanovich, S. Capillary electrophoresis on microchip. Electrophoresis 2000, 21, 41–54. 23. Stellwagen, N. C.; Gelfi, C.; Pier Giorgio, R. The free solution mobility of DNA. Biopolymers 1997, 42, 687–703. 24. Kato, M.; Gyoten, Y.; Sakai-Kato, K.; Nakajima, T.; Toyo’oka, T. Cationic starch derivatives as dynamic coating additives for analysis of amino acids and peptides using poly(methyl methacrylate) microfluidic devices. Anal. Chem. 2004, 76, 6792–6796. 25. Terabe, S. Micellar electrokinetic chromatography. Anal. Chem. 2004, 76, 240A–246A. 26. Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Electrokinetic separations with micellar solutions and open-tubular capillaries. Anal. Chem. 1984, 56, 111–113. 27. Terabe, S.; Otsuka, K.; Ando, T. Electrokinetic chromatography with micellar solution and open-tubular capillary. Anal. Chem. 1985, 57, 834–841. 28. Zhang, C.-X.; Sun, Z.-P.; Lin, D.-K. Micellar electrokinetic capillary chromatography theory based on conventional chromatography. J. Chromatogr. A 1993, 655, 309–316. 29. Neiman, B.; Grushka, E.; Lev, O. Use of gold nanoparticles to enhance capillary electrophoresis. Anal. Chem. 2001, 73, 5220–5227. 30. Kenndler, E. In High Performance Capillary Electrophoresis; Khaledi, M. G., Ed.; Wiley: New York, 1998; pp 25–76. 31. Yu, C. J.; Su, C. L.; Tseng, W. L. Separation of acidic and basic proteins by nanoparticlefilled capillary electrophoresis. Anal. Chem. 2006, 78, 8004–8010. 32. Bao, J.; Chen, W.; Liu, T.; Zhu, Y.; Jin, P.; Wang, L.; Liu, J.; Wei, Y.; Li, Y. Bifunctional Au-Fe3O4 nanoparticles for protein separation. ACS Nano 2007, 1, 293–298. 33. Lin, C.-Y.; Liu, C.-H.; Chang, H.-C.; Tseng, W.-L. Enrichment and separation of acidic and basic proteins using the centrifugal ultrafiltration followed by nanoparticle-filled capillary electrophoresis. Electrophoresis 2008, 29, 3024–3031.

250

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

34. Li, M.-D.; Cheng, T.-L.; Tseng, W.-L. Nonionic surfactant-capped gold nanoparticles for selective enrichment of aminothiols prior to CE with UV absorption detection. Electrophoresis 2009, 30, 388–395. 35. Yang, L.; Guihen, E.; Holmes, J. D.; Loughran, M.; O’Sullivan, G. P.; Glennon, J. D. Gold nanoparticle-modified etched capillaries for open-tubular capillary electrochromatography. Anal. Chem. 2005, 77, 1840–1846. 36. Zhou, D.; Wang, Y.; Yang, R.; Zhang, W.; Shi, R. Effects of novel quasi-interpenetrating network/gold nanoparticles composite matrices on DNA sequencing performances by CE. Electrophoresis 2007, 28, 2998–3007. 37. Ivanov, M. R.; Bednar, H. R.; Haes, A. J. Investigations of the mechanism of gold nanoparticle stability and surface functionalization in capillary electrophoresis. ACS Nano 2009, 3, 386–394. 38. Kuo, I. T.; Huang, Y. F.; Chang, H. T. Silica nanoparticles for separation of biologically active amines by capillary electrophoresis with laser-induced native fluorescence detection. Electrophoresis 2005, 26, 2643–2651. 39. Huang, Y.-F.; Chiang, C.-K.; Lin, Y.-W.; Liu, K.; Hu, C.-C.; Bair, M.-J.; Chang, H.-T. Capillary electrophoretic separation of biologically active amines and acids using nanoparticle-coated capillaries. Electrophoresis 2008, 29, 1942–1951. 40. Liu, D. N.; Wang, J.; Guo, Y. G.; Yuan, R. J.; Wang, H.; Bao, J. J. Separation of aromatic acids by wide-bore electrophoresis with nanoparticles prepared by electrospray as pseudostationary phase. Electrophoresis 2008, 29, 863–870. 41. Hsieh, Y. L.; Chen, T. H.; Liu, C. P.; Liu, C. Y. Titanium dioxide nanoparticles-coated column for capillary electrochromatographic separation of oligopeptides. Electrophoresis 2005, 26, 4089–4097. 42. Hsieh, Y. L.; Chen, T. H.; Liu, C. Y. Capillary electrochromatographic separation of proteins on a coated titanium dioxide nanoparticles. Electrophoresis 2007, 27, 4288–4294. 43. Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chemistry of carbon nanotubes. Chem. Rev. 2006, 106, 1105–1136. 44. Gooding, J. J.; Wibowo, R.; Liu, J.; Yang, W.; Losic, D.; Orbons, S.; Mearns, F. J.; Shapter, J. G.; Hibbert, D. B. Protein electrochemistry using aligned carbon nanotube arrays. J. Am. Chem. Soc. 2003, 125, 9006–9007. 45. Wang, Z.; Luo, G.; Chen, J.; Xiao, S.; Wang, Y. Carbon nanotubes as separation carrier in capillary electrophoresis. Electrophoresis 2003, 24, 4181–4188. 46. Xu, Y.; Li, S. F. Y. Carbon nanotube-enhanced separation of DNA fragments by a portable capillary electrophoresis system with contactless conductivity detection. Electrophoresis 2006, 27, 4025–4028. 47. Na, N.; Hu, Y. P.; Jin, O. Y.; Baeyens, W. R. G.; Delanghe, J. R.; Taes, Y. E. C.; Xie, M. X.; Chen, H. Y.; Yang, Y. P. On the use of dispersed nanoparticles modified with single layer b-cyclodextrin as chiral selector to enhance enantioseparation of clenbuterol with capillary electrophoresis. Talanta 2006, 69, 866–872. 48. Xiong, X.; Ouyang, J.; Baeyens, W. R. G.; Delanghe, J. R.; Shen, X. M.; Yang, Y. P. Enhanced separation of purine and pyrimidine bases using carboxylic multiwalled carbon nanotubes as additive in capillary zone electrophoresis. Electrophoresis 2006, 27, 3243–3253. 49. Nilsson, C.; Becker, K.; Harwigsson, I.; B€ ulow, L.; Birnbaum, S.; Nilsson, S. Hydrophobic interaction capillary electrochromatography of protein mutants. Use of

REFERENCES

50.

51. 52.

53.

54.

55.

56.

57.

58.

59.

60. 61.

62. 63.

64.

251

lipid-based liquid crystalline nanoparticles as pseudostationary phase. Anal. Chem. 2009, 81, 315–321. Kuo, Y.-C.; Lin, T.-W. Electrophoretic mobility, zeta potential, and fixed charge density of bovine knee chondrocytes, methyl methacrylate-sulfopropyl methacrylate, polybutylcyanoacrylate, and solid lipid nanoparticles. J. Phys. Chem. B 2006, 110, 2202–2208. Pumera, M.; Wang, J.; Grushka, E.; Polsky, R. Gold nanoparticle-enhanced microchip capillary electrophoresis. Anal. Chem. 2001, 73, 5625–5628. Lin, Y.-W.; Huang, M.-J; Chang, H.-T. Analysis of double-stranded DNA by microchip capillary electrophoresis using polymer solutions containing gold nanoparticles. J. Chromatogr. A 1014, 2003, 47–55. Wang, W.; Zhao, L.; Zhang, J.-R.; Wang, X.-M.; Zhu, J.-J.; Chen H.-Y. Modification of poly(dimethylsiloxane) microfluidic channels with silica nanoparticles based on layer-bylayer assembly technique. J. Chromatogr. A 2006, 1136, 111–117. Wang, A.-J.; Xu, J.-J.; Zhang, Q.; Chen, H.-Y. The use of poly(dimethylsiloxane) surface modification with gold nanoparticles for the microchip electrophoresis. Talanta 2006, 69, 210–215. Wang, W.; Zhao, L.; Zhou, F.; Zhu, J.-J.; Zhang, J.-R. Electroosmotic flow-switchable poly (dimethylsiloxane) microfluidic channel modified with cysteine based on gold nanoparticles. Talanta 2007, 73, 534–539. Liu, Y.; Xue, Y.; Ji, J.; Chen, X.; Kong, J.; Yang, P.; Girault, H. H.; Liu, B. Gold nanoparticle assembly microfluidic reactor for efficient on-line proteolysis. Mol. Cell. Proteomics 2007, 6, 1428–1436. Li, M.; Lin, Y.-C.; Su, K.-C.; Wang, Y.-T.; Chang, T. C.; Lin, H. P. A novel real-time immunoassay utilizing an electro-immunosensing microchip and gold nanoparticles for signal enhancement. Sens. Actuators B 2006, 117, 451–456. Ko, Y. J.; Maeng, J. H.; Ahn, Y.; Hwang, S. Y.; Cho, N. G.; Lee, S.-H. Real-time immunoassay with a PDMS-glass hybrid microfilter electro-immunosensing chip using nanogold particles and silver enhancement. Sens. Actuators B 2008, 132, 327–333. Ko, Y.-J.; Maeng, J.-H.; Ahn, Y.; Hwang, S. Y.; Cho, N.; Lee, S.-H. Microchip-based multiplex electro-immunosensing system for the detection of cancer biomarkers. Electrophoresis 2008, 29, 3466–3476. Gamby, J.; Abid, J.-P.; Tribollet, B.; Girault, H. H. Nanomosaic network for the detection of proteins without direct electrical contact. Small 2008, 4, 802–809. Sato, Y.; Hosokawa, K.; Maeda, M. Detection of non-cross-linking interaction between DNA-modified gold nanoparticles and a DNA-modified flat gold surface using surface plasmon resonance imaging on a microchip. Colloids Surf. B: Biointerfaces 2008, 62, 71–76. Ortiz-Lombardia, M.; Gonzalez, A.; Eritja, R.; Aymami, J.; Azorin, F.; Coll, M. Crystal structure of a DNA Holliday junction. Nat. Struct. Biol. 1999, 6, 913–917. Liu, J.; Lin, S.; Qi, D.; Deng, C.; Yang, P.; Zhang, X. On-chip enzymatic microreactor using trypsin-immobilized superparamagnetic nanoparticles for highly efficient proteolysis. J. Chromatogr. A 2007, 1176, 169–177. Shih, P.-H.; Shiu, J.-Y.; Lin, P.-C.; Lin, C.-C.; Veres, T.; Chen, P. On chip sorting of bacterial cells using sugar-encapsulated magnetic nanoparticles. J. Appl. Phys. 2008, 103, 07A316/ 311–307A316/313.

252

MICROCHIP AND CAPILLARY ELECTROPHORESIS USING NANOPARTICLES

65. Lai, J. J.; Hoffman, J. M.; Ebara, M.; Hoffman, A. S.; Estournes, C.; Wattiaux, A.; Stayton, P. S. Dual magnetic-/temperature-responsive nanoparticles for microfluidic separations and assays. Langmuir 2007, 23, 7385–7391. 66. Zuzana B
ılkov
a, Z.; Slov
akov
a, M.; Minc, N.; F€ utterer, C.; Cecal, R.; Hor
ak, D.; Benesˇ, M.; Potier, I. I.; Krenkov
a, J.; Przybylski, M.; Viovy, J.-L. Functionalized magnetic micro- and nanoparticles: optimization and application to m-chip tryptic digestion. Electrophoresis 2006, 27, 1811–1824. 67. Huang, Y.; Shan, W.; Liu, B.; Liu, Y.; Zhang, Y.; Zhao, Y.; Lu, H.; Tang, Y.; Yang, P. Zeolite nanoparticle modified microchip reactor for efficient protein digestion. Lab Chip 2006, 6, 534–539. 68. Ji, J.; Zhang, Y.; Zhou, X.; Kong, J.; Tang, Y.; Liu, B. Enhanced protein digestion through the confinement of nanozeolite-assembled microchip reactors. Anal. Chem. 2008, 80, 2457–2463. 69. Yuan, H.; Cheow, P.-S.; Ong, J.; Toh, C.-S. Characterization of the electrophoretic mobility of gold nanoparticles with different sizes using the nanoporous alumina membrane system. Sens. Actuators B 2008, 134, 127–132. 70. Castellana, E. T.; Kataoka, S.; Albertorio, F.; Cremer, P. S. Direct writing of metal nanoparticle films inside sealed microfluidic channels. Anal. Chem. 2006, 78, 107–112. 71. Farokhzad, O. C.; Khademhosseini, A.; Jon, S.; Hermmann, A.; Cheng, J.; Chin, C.; Kiselyuk, A.; Teply, B.; Eng, G.; Langer, R. Microfluidic system for studying the interaction of nanoparticles and microparticles with cells. Anal. Chem. 2005, 77, 5453–5459. 72. Goto, Y.; Matsuno, R.; Konno, T.; Takai, M.; Ishihara, K. Polymer nanoparticles covered with phosphorylcholine groups and immobilized with antibody for high-affinity separation of proteins. Biomacromolecules 2008, 9, 828–833. 73. Li, Y.; Chen, Y.; Xiang, R.; Ciuparu, D.; Pfefferle, L. D.; Horv
ath, C.; Wilkins, J. A. Incorporation of single-wall carbon nanotubes into an organic polymer monolithic stationary phase for m-HPLC and capillary electrochromatography. Anal. Chem. 2005, 77, 1398–1406. 74. Weng, X.; Bi, H.; Liu, B.; Kong, J. On-chip chiral separation based on bovine serum albumin-conjugated carbon nanotubes as stationary phase in a microchannel. Electrophoresis 2006, 27, 3129–3135. 75. Gong, M.; Wehmeyer, K. R.; Limbach, P. A.; Arias, F.; Heineman, W. R. On-line sample preconcentration using field-amplified stacking injection in microchip capillary electrophoresis. Anal. Chem. 2006, 78, 3730. 76. Shiddiky, M. J. A.; Park, H.; Shim, Y.-B. Direct analysis of trace phenolics with a microchip: in-channel sample preconcentration, separation, and electrochemical detection. Anal. Chem. 2006, 78, 6809–6817. 77. Shiddiky, M. J. A.; Shim, Y.-B. Trace analysis of DNA: preconcentration, separation, and electrochemical detection in microchip electrophoresis using Au nanoparticles. Anal. Chem. 2007, 79, 3724–3733. 78. Zeng, Y.; Harrison, D. J. Self-assembled colloidal arrays as three-dimensional nanofluidic sieves for separation of biomolecules on microchips. Anal. Chem. 2007, 79, 2289–2295. 79. Giddings, J. C.; Unified Separation Science; Wiley: New York, 1991. 80. Kirk, J. S.; Sweedler, J. V.; Bohn, P. W. Nanofluidic injection and heterogeneous kinetics of organomercaptan surface displacement reactions on colloidal gold in a microfluidic stream. Anal. Chem. 2006, 78, 2335–2341.

REFERENCES

253

81. Goluch, E. D.; Nam, J.-M.; Georganopoulou, D. G.; Chiesl, T. N.; Shaikh, K. A.; Ryu, K. S.; Barron, A. E.; Mirkin, C. A.; Liu, C. A bio-barcode assay for on-chip attomolar-sensitivity protein detection. Lab Chip 2006, 6, 1293–1299. 82. Taton, T. A.; Mirkin, C. A.; Letsinger, L. R.; Scanometric, DNA array detection with nanoparticle probes. Science 2000, 289, 1757–1760. 83. Nie, Z.; Xu, S.; Seo, M.; Lewis, P. C.; Kumacheva, E. Polymer particles with various shapes and morphologies produced in continuous microfluidic reactors. J. Am. Chem. Soc. 2005, 127, 8058–8063. 84. Zhang, H.; Tumarkin, E.; Peerani, R.; Nie, Z.; Sullan, R. M. A.; Walker, G. C.; Kumacheva, E. Microfluidic production of biopolymer microcapsules with controlled morphology. J. Am. Chem. Soc. 2006, 128, 12205–12210. 85. Huang, K.-S.; Lai, T.-H.; Lin, T.-C. Manipulating the generation of Ca-alginate microspheres using microfluidic channels as a carrier of gold nanoparticles. Lab Chip 2006, 6, 954–957. 86. Zourob, M.; Mohr, S.; Mayes, A. G.; Macaskill, A.; P
erez-Moral, N.; Fielden, P. R.; Goddard, N. J. A micro-reactor for preparing uniform molecularly imprinted polymer beads. Lab Chip 2006, 6, 296–301. 87. Rondeau, E.; Cooper-White, J. J. Biopolymer microparticle and nanoparticle formation within a microfluidic device. Langmuir 2008, 24, 6937–6945. 88. Sarov, Y. E.; Capek, I.; Ivanov, T. B.; Ivanova, K. Z.; Sarova, V. A.; Rangelow, I. W. On total internal reflection investigation of nanoparticles by integrated microfluidic system. Nano Lett. 2008, 8, 375–38. 89. Dawoud, A. A.; Kawaguchi, T.; Jankowiak, R. In-channel modification of electrochemical detector for the detection of bio-targets on microchip. Electrochem. Commun. 2007, 9, 1536–1541.

7 PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS: FABRICATION METHODS AND APPLICATIONS IN MOLECULAR AND CELL BIOLOGY JIAN SHI ECOLE NORMALE SUPeRIEURE, PARIS, FRANCE

YONG CHEN Institute for Integrated Cell-Material Sciences, Kyoto University, Kyoto, Japan

7.1 INTRODUCTION Nanofabrication and microdevice technologies are now becoming more and more important for the continuous improvement of our basic understanding in cell biology and our ability in diagnosis, drug discovery, environment control, and clinical treatment.1–3 Among many others, nanofabrication can be used to pattern functional surfaces to control cell adhesion, migration, differentiation, and so on, whereas microdevices particularly designed for the manipulation of small volume of liquid samples, that is, microfluidic chips, can be used for highly efficient cell handling, biomolecule separation, high-throughput screening, and so on. The reasons are obvious: the sizes of the fabricated elements (patterned features and channels) not only are comparable to the size of cells and large biomolecules but can also be well designed for highly integrated microsystems.4–9 In addition, it will be possible to

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

255

256

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

produce such systems with high throughput and low cost, leading to a similar impact generated by CMOS-based semiconductor industry during the last few decades. Pillars are certainly the simplest and universal pattern elements. Nature uses pillars for different purposes. At micrometer scale, for example, one can observe dense pillar arrays on lotus leaves for self-cleaning, whereas more complicated pillar structures can be found on the butterfly wings showing splendid colors for seduction and/protection10,11 (Figure 7.1). With the rapid progress in nanofabrication tools, one can now easily produce high-density pillars of different sizes and different materials, also showing biomimetic or designed functionalities. In particular, patterned pillars can be used as artificial gels for biomolecule separation or preconcentration. They can also be used as artificial extracellular matrix (ECM) for controlling and/or monitoring cell growth. On the other hand, microfluidic chips are ideal systems for the integration of pillar arrays, although only a few investigations have been reported previously. Microfluidics itself is an emerging field with great promises.12–14 Indeed, the already demonstrated application potential has attracted a continuous growing interest

FIGURE 7.1 Photograph of lotus (a) and butterfly (c), and scanning electronic micrograph (SEM) of pillars on the lotus leave (b) and wings of the butterfly (d).10,11

PATTERNING TECHNIQUES

257

during the past 15 years. A large variety of functionalities can now be integrated for molecular and cell biology studies. One can, for example, use integrated microelectrodes to trap, separate, or fuse cells.15–17 One can also use integrated fiber or gratings for highly sensitive detection.18,19 The integrated pillars are often used as barriers to trap cells or as filters to separate biomolecules. It should also be possible to explore more functionalities based on their unique electric, magnetic, or mechanic properties. In this chapter, first we discuss pillars and pillar arrays integrated into microfluidic chips. Then, fabrication methods are reviewed and application examples are described to illustrate different aspects of their utilities. 7.2 PATTERNING TECHNIQUES Both top-down and bottom-up techniques can be used for the fabrication of highdensity pillars. The so-called “top-down” techniques have been issued from semiconductor industry, where lithography and pattern transfer techniques play the most essential role in the manufacturing of integrated circuits. By using standard photolithography, for example, the designed features are transferred from a mask to a resist layer coated on a substrate. Then, the features are produced in substrate by etching, regrowth, doping, and liftoff to obtain desired functionalities. In contrast, the “bottomup” techniques are based on self-assembling at nanoscale or molecular levels. Obviously, the two approaches can be combined and used for the fabrication of highly integrated microdevices with enhanced functionalities. 7.2.1 Lithography and Related Techniques Conventional lithography methods refer to optical lithography, electron beam lithography, focused ion beam lithography, X-ray lithography, and extreme UV lithography that are commonly used or studied by semiconductor industry and research laboratories for different purposes. During the past decades, efforts have been extensively devoted to the improvement of these methods in terms of resolution and throughput. However, there is still no high-throughput method for the manufacturing of integrated features of sizes down to 50 nm. Moreover, even for the fabrication of large features, the ownership cost is excessively high and the required environment is not suitable for different applications other than semiconductor industry. Electron beam lithography, for instance, is a high-resolution technique based on sequential writing. Typically, a high-energy (1–200 keV) electron beam is formed and projected on the sample with an electron beam optical column. The resolution of electron beam lithography is limited by scattering of second electrons in the resist layer and substrate, providing a feature size down to 10 nm. Although electron beam lithography is flexible and of high resolution, it is not suitable for mass production because of the limitation of its writing speed. Similarly, focused ion beam lithography provides excellent resolution but low processing speed. As high-energy ions penetrate the material, they lose their energy at a rate several orders of magnitude higher than that of electrons because of their large

258

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

mass values. For this reason, it is generally used for micromachining or localized implantation. Optical lithography is the common method for pattern reproduction. The resolution of projection optical lithography is defined by the Rayleigh criterion R ¼ k1l/NA, where l is the wavelength of the light, NA the numerical aperture of the optical system, and k1 an empirical factor depending on details of experimental conditions. To improve the resolution, sophisticated mask design, off-axial illumination, and top surface image techniques are used, but most important is to utilize short-wavelength exposure sources. The early projection systems worked with mercury lamp for a wavelength range between 350 and 400 nm. Now, advanced production systems use 248 or 193 nm wavelength radiation provided by krypton fluoride or argon fluoride excimer laser. To further improve the resolution, liquid is injected into the gap between the focusing lens and the substrate to obtain a larger NA (immersion lithography). Finally, extreme UV lithography has been studied for many years as candidate of the so-called next-generation lithography because of its much reduced exposure wavelength (11–13 nm). This technique is extremely complex and expensive, and it is not clear yet when this technique will be applied for the mass production. Unlike conventional methods, more recent techniques including soft lithography and nanoimprint lithography are not based on the use of photon, electron, and ion beams. Accordingly, these nonconventional methods are flexible, low cost, and promising for a large variety of applications. Soft lithography has been proposed in 1998 by Whitesides,20 which is based on the use of PDMS (polydimethylsiloxane) as lithography template. Typically, it is used for microcontact printing where molecules are released from template and then self-assembled in the contact area. The printed features can be used for different purposes, but they might be resistant for wet etching. The resolution of microcontact printing is limited by the mechanical stability of the template and the diffusion of inking molecules around the contact area. In practice, features of sizes down to 200 nm can be obtained and it is possible to achieve a higher resolution by using PDMS templates of improved stiffness and molecules of smaller diffusion length. Actually, soft lithography refers to a large class of patterning and device technologies that use PDMS features obtained by casting. Nanoimprint lithography was proposed by Chou in 199521 and soon it became a widespread method for both academic and industrial applications. Nanoimprint lithography is based on pressure-induced deformation with a high-resolution mold. In most cases, a resist layer deposited on substrate is patterned using a mold fabricated by electron beam lithography. After separation of the mold, the resist pattern is treated by reactive ion etching to remove the residual part of the imprinted area before further processing. Nanoimprint lithography is intrinsically of high resolution, high throughput, and low cost. No apparent limiting factor does exist, such as diffraction in optical lithography or scattering in electron beam lithography. Nanoimprint tools can be relatively simple since neither optical lens system for projection nor high-energy column for beam manipulation is required. Finally, once a mold is fabricated, it can be used many times, which makes this technique attractive for low cost and mass production. Currently, two types of nanoimprint methods coexist, that is, thermal nanoimprint lithography initially proposed by Chou and UV nanoimprint lithography

PATTERNING TECHNIQUES

259

(UV-NIL) that uses photocurable resin and UV light for solidification.22,23 UV imprinting works at room temperature with a low pressure, which should be more suitable for fine alignment compared to the thermal imprinting technique. For both thermal and UV-based processes, the mold fabrication is a critical issue because of the high cost. Pattern replication over large area is also difficult because of the nonflatness of wafers. Although large pressure can be applied, it is not reasonable to work with fragile sample and fine aligning. One solution is to use a step-and-repeat strategy but this may require a much more sophisticated imprinting tool. For these reasons, it is interesting to use soft molds for a more conformable imprinting over large wafer area.24,25 Soft molds such as PDMS can be easily produced by casting so that many copies can be obtained at low cost. Furthermore, PDMS has low surface energy that facilitates the mold separation after imprint. PDMS has, however, low Young’s module that prevents the replication of both high resolution (<50 nm) and very large features (>100 mm) because of the pressure-induced mechanic deformation. Fortunately, pillar arrays of a typical feature size in the range between 50 nm and 100 mm are useful for the most recent applications. Anyway, large features can be easily replicated by photolithography and small features can be produced by electron beam lithography. Then, mix-and-match techniques can be applied to cover all feature sizes in a more cost-effective way.26 To illustrate the replication performance, we show in Figure 7.2 a few examples of the fabrication ability of both thermal and soft UV-NIL techniques.

FIGURE 7.2 SEM images of fabricated high-density dot and pillar arrays. (a and c) Nickel dots (period 60 nm) and SiO2 pillars (period 300 nm) obtained by thermal nanoimprint lithography; (b and d) PMMA pillars on silicon (diameter 120 nm) and in plexiglas (height 3 mm) obtained by soft UV nanoimprint lithography.24–27

260

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

7.2.2 Auto Assemblage Nanopillars can also be produced by nonlithographic methods. One solution is to use the so-called bottom-up approach where a functional material can be constructed by self-assembly of elementary building blocks into ordered arrays. These building blocks can be typically metallic or polymer nanoparticles. After assembling, the pattern of the ordered arrays can be transferred to substrate by etch or other pattern transfer techniques. Similarly, copolymer demixing can also be used for the fabrication of nanopillar arrays. Monodispersed nanospheres made of silica, metal, and polymers are now commercially available and a large number of investigations have already been reported on their organization onto flat or structured surfaces. Most of the methods used so far were based on capillary force assisted auto organization. In the simplest case where a droplet of the colloid suspension is dried slowly on an unpatterned polar surface, the particles aggregate at the rim of the droplet because of the attractive capillary forces between particles. Several methods including controlled deposition, dipping, and microfluidic flowing have been proposed. By controlling more accurately the capillary force and the evaporation, nanoparticles could be assembled on patterned substrates with various geometries.27 Once highly ordered and large area particles are assembled, they can be used as mask for etching or liftoff. When reactive ion etching technique is applied, the space between the pillars can be controlled by adjusting the etch parameters. Otherwise, pillar arrays of different periods can be obtained by using nanospheres of different sizes. When liftoff technique is applied, a thin metal layer is deposited and the resulted hexagonal triangle pattern serves as etch mask for the substrate. By changing the angle of metal deposition (shadow masking), other forms of patterned arrays could also be obtained.28 Although both fabrication processes are simple, which can be used for a number of applications, it is in general difficult to produce perfect and large surface pillar arrays without defect. Block copolymer demixing has been proposed to achieve ultrahigh-density nanostructure patterning. A block copolymer consists of two chemically distinct polymer chains covalently bound at one end. Because of the tendency of phase separation for unlike chains and the constraint imposed by chain connectivity, the block copolymer can be self-organized to form periodic domains on a molecular length. In general, the competition between the interfacial and chain stretching energies governs the bulk equilibrium phase behavior, and the relative volume fractions of the blocks control the curvature, size, and the periodicity of the nanodomains. In most cases, functionalized block copolymers can be used for spontaneous organization on a flat substrate. However, the nanostructure formation can be improved by using a patterned substrate. Other kinds of physical or chemical means such as mechanic sliding and electric field induction can also be used for the improvement of the pattern formation, resulting in regular line and dot arrays with a feature size down to a few nanometers. Finally, the self-organized copolymers can be used for selective ion etching or liftoff for further device processing.

OTHER FABRICATION ASPECTS

261

7.2.3 Growth Chemical vapor deposition (CVD), physical vapor deposition (PVD), and many other derived deposition methods can be used for the fabrication of small features on a substrate. Indeed, with or without mask, a large variety of materials could be deposited as monocrystalline, polycrystalline, or amorphous, including silicon, silicon–germanium, silicon dioxide, silicon carbide, silicon nitride and silicon oxynitride, carbon fiber and carbon nanotubes, tungsten, titanium nitride, and various high-k dielectrics. Electroplating using either electrolytic or electroless deposition techniques can be applied to obtain high-density metallic pillars. With electrolytic deposition, the surface on which the metal is to be deposited is used as cathode and deposition occurs when the metal ions travel under the influence of the electric field in the aqueous solution of metal salts. A DC voltage is applied between the substrate and the metal to be deposited. With a patterned substrate, it is important to accurately control the current density that strongly influences the deposition rate, plating adherence, and plating quality. It is also important to note that in the case of highdensity nanopillar growth, the current density cannot be homogeneous over the plating surface. In general, the current density can be significantly larger at the edge than in the center of a patterned area.

7.3 OTHER FABRICATION ASPECTS Depending on the application domain and the manufacturing cost, materials of different types can be used. Among them, silicon-based materials and polymers are widely used because of their well-known physical and chemical properties and the availability of their processing parameters. In many cases, surface fuctionalization is needed to enhance the functionality of the fabricated pillar arrays. 7.3.1 Material Choice Silicon and silicon-based materials are mainly used for the production of integrated circuits by semiconductor industry. They are also commonly used for the manufacturing of microelectromechanical systems (MEMS). More recently, silicon-based materials are studied for the fabrication of integrated optoelectronic devices. Silicon nanopillars can be easily obtained by lithography and reactive ion-etching techniques, which explains why most nanopillars fabricated so far were based on silicon processing. A thin layer of silicon dioxide can be easily obtained on the surface of silicon by native or controlled oxidation, which is thermally stable and chemically resistant to organic solvents and acids. Then, surface fuctionalization can be easily done for different molecules. Compared to the polymers, silicon-based materials are, however, more expensive and their batch processing sequences are in general more complicated than the one-step replication methods. Finally, silicon is not desirable for in vivo applications and it is not suited for applications such as electrophoresis with a high electric field, observation with an inverted light microscope, and so on.

262

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

The fabrication of polymer pillars can be simple and cost effective using nonconventional techniques. Another advantage of using polymers is the availability of a broad range of material choices, including PMMA, PS, PC, cyclic olefin, elastomer, and others. In practice, the material choice depends on many factors such as the processing ability, cost, optical transparency, mechanic stiffness, and dielectric properties. When applied to microfluidics, appropriate sealing methods, such as lamination, ultrasonic welding, thermal bonding, have to be carefully studied to achieve the desired performance. Due to the limitation of the polymers’ thermal stability and the stability against organic solvents, the fabricated pillars can only be operated at temperature below the glass transition temperature and they have to avoid many types of organic solvents also. Finally, a surface treatment is generally required for microfluidic application since most of the polymers are not hydrophilic. Materials of other types are also studied to demonstrate more specific functionalities. For example, metallic pillar structures have shown unique electrical, magnetic, and optical properties. Gold pillar arrays can be used, in particular, for label-free detection because of much enhanced surface plasmon effect. Although magnetic nanopillars are generally considered as media for ultrahigh-density data recording, they can also be used for bioprocessing or bioanalyses. Semiconductor nanopillars of ZnO have shown not only interesting piezoelectric properties but also the usefulness in the study of UV lasers, dye-sensitized solar cells, antireflection coating, and so on.29–31 Therefore, functional materials should be more attractive than silicon and polymers, and fortunately they can also be patterned and integrated into microdevices made of silicon or polymers. 7.3.2 Surface Fuctionalization To endow the nanopillars with more specific applications, one can apply different physical or chemical methods to treat the surface of pillars made of silicon, silicon dioxide, or polymers. In general, a treated surface has to be stable in time and the applied method has to be compatible with the whole process of device fabrication. It is also desirable to have a high resolution when a local surface modification is applied. Typically, a hydrophilic surface treatment improves the device aqueous solubility and the device biocompatibility. When appropriate biomolecules are bound to the pillars’ surfaces, the fabricated devices can be used for biosensing, sorting, or purification. The surface treatment of pillars can be performed using a variety of techniques for different purposes. Physical modification can be done by CVD, PVD, spin coating, solution cast, and plasma processes, whereas chemical modification can be obtained by grafting and self-assembly. Both techniques can be applied for the improvement of device biocompatibility and the reduction or elimination of solute interactions with device surfaces. They can also be used for the modification of electroosmotic flow and surface immobilization of reactive biomolecules such as enzymes, antibodies, proteins, DNA, and so on. Finally, they can be used as sieving matrices in separation devices.

OTHER FABRICATION ASPECTS

263

7.3.3 Integration into Microdevices Integration of pillar arrays is a critical issue for the functional device fabrication for which different methods are studied. Conventional lithography methods such as electron beam lithography and focused ion beam lithography can be used for device prototyping, but they are not cost-effective and cannot be used for large-scale manufacturing. Nanoimprint lithography and self-assembling are able to pattern nanostructures cost-effectively, but it is more difficult to use them for the fabrication of complex and multilevel devices. One solution is to combine both methods, that is, to produce nanopillar arrays by nanoimprint lithography or self-organized assembling and other large features (cavities, channels, posts, etc.) by photolithography and soft lithography. High-density nanopillars could be integrated into microfluidic devices by in situ polymerization cast molding.32 The more general approach can be based on “mix-and-match” lithography methods.26 Indeed, we obtained high-density nanopillars by soft UV nanoimprint lithography and microfluidic cross-channels and reservoirs by standard optical lithography. As shown in Figure 7.3, soft UV-NIL is first applied. After the first liftoff with a thin film of nickel, larger features such as microfluidic channels are defined by photolithography and the second liftoff of a thin layer of nickel. Then, the substrate is etched by reactive ion etching and high-density and high aspect ratio nanopillars are obtained simultaneously with microfluidic channels. Finally, the device is sealed by a PDMS cover plate with connection holes. Clearly, this mix-and-match approach has the following advantages: (1) It is easy to control the nanoimprint process so that high-quality pattern definition can be achieved without considering particular mold and process design. (2) It is easy to align microfluidic channel pattern with the nanopillar areas, since the excess areas of the nanopillar arrays will be simply masked by Ni layer of the second liftoff. (3) It is

FIGURE 7.3 Mix-and-match fabrication process for the integration of pillar arrays into a microchannel.

264

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

easy to cover the patterned structures with a flat PDMS plate, and there will be no leakage problem since the nanopillar areas are completely integrated into the channel with the same etched height. In a particular case, we demonstrated large DNA separation by using this kind of microfluidic device. Obviously, the mix-and-match approach is applicable for the fabrication of other types of microdevices with integrated nanopatterns. 7.4 APPLICATION EXAMPLES Pillar arrays can be used for different purposes in cell biology. Although their whole application potential cannot be predicted at the present stage, we select the following examples to illustrate the current efforts on several aspects without paying particular attention on the pillars’ geometric design or their device configuration. Indeed, all fabricated patterns should be scalable and all pillar arrays could be integrated into microfluidic devices. 7.4.1 Pillars for Genomics and Proteomics Genomic and proteomic information is of great importance in cell biology, cancer diagnostics, and drug discovery. Although the human genomics project was finished earlier than expected, there are still remaining challenges for both fundamental research and clinic uses. One of the challenges is the fabrication of miniaturized biomedical devices for high-quality and high-throughput analyses. In such a context, nanopillar arrays integrated into microfluidic devices hold a great promise for improved genomic and proteomic investigation. 7.4.1.1 Biomolecule Preconcentration One critical issue in genomics and proteomics is the sample preparation. It is known that more than 10,000 different biomolecule species with concentrations varying over nine orders of magnitude exist in a typical blood sample. Such diversities of molecules, as well as their huge differences in concentration, cause a big problem for the sample preparation. In general, extracted and purified genomic materials are limited in both concentration and volume, which makes it difficult for both analysis processing and detection. For proteomics, advanced detection methods such as laserinduced fluorescent and mass spectroscopy are often required. This issue is not exceptional for the microfluidic analysis systems, although it is able to handle and manipulate the liquid sample in pL–nL scale with high efficiency. Besides, in proteomics, this problem is exacerbated by the fact that information-rich signaling molecules are present only in trace concentrations (nM–pM range). Furthermore, there is no signal amplification technique, such as polymerase chain reaction (PCR), for proteins and peptides. Efficient tools for sample amplification or preconcentration are required for the manipulation and detection of highly diluted analytes in extreme small volumes. For this particular application, the integrated nanopillar arrays may have some advantages.

APPLICATION EXAMPLES

265

Silicon microstructures with high surface to volume ratios have been used for capturing, washing, and eluting short- (500 bp) and medium-size (48,000 bp) DNA.33 In this study, chaotropic (GuHCl) salt solutions were used as binding agents, whereas ethanol-based solutions and water were used as wash and elution agents, respectively. DNA quantities approaching 40 ng cm
2 of binding area were captured from input solutions in the 100–1000 ng mL
1 concentration range. For dilute samples of interest for pathogen detection, PCR and gel electrophoresis were used to demonstrate extraction efficiencies of about 50%, and concentration factors of about 10 using bacteriophage l DNA as the target. Alternatively, the ability to purify single-stranded DNA (ssDNA) sequencing ladders has been demonstrated by using microfluidics-based solid-phase and reversible immobilization technology.34 Here, hot embossing has been used for the fabrication of polycarbonate (PC) containing microposts with large surface areas. An immobilization bed was prepared by exposing PC surfaces to UV radiation, which resulted in the formation of surface carboxylate groups through a photooxidation reaction. Such functionalities can serve as a capture medium for a wide range of molecular weight-sized DNAs for the isolation, preconcentration, or purification of DNA embedded in complex sample matrices. As observed by UV spectroscopy, a load of 7.6  1.6 mg mL
1 of gDNA was immobilized onto the PPC bed. The recovery of DNA following purification was estimated to be 85  5%. Considerable efforts have also been made to the problem of microfluidics-based protein preconcentration. In particular, porous silica membranes were used to preconcentrate protein samples prior to electrophoretic separations.35 The preconcentration rate in such devices, however, could not be easily predicted because of the problems of membrane clogging and aging. Most of the recent works have focused on nanochannel-based preconcentration. It is known that co-ions of the surface charges can be excluded from the nanochannel while counterions are enriched. If an electric field is applied, the enhanced counterionic transport can be used for electropreconcentration of charged biomolecules. Four preconcentration regimes can then be distinguished, and results suggested that both the mobility and the valence of the species are important parameters in the determination of the preconcentration rates.36 Alternatively, protein samples can be prepared by isoelectrical focusing (IEF) with patterned substrates.37 By using PDMS pillar arrays fabricated by soft lithography, a protein mixture with pI ranging from 4.7 to 10.6 has been successfully separated with good resolution. Compared to the classical gel-based system, this method considerably reduces the separation time from several hours to 10 min. The pillar chips can be reused several times while classical gels are disposable. 7.4.1.2 DNA Stretching and Separation Separation of DNA fragments by size is at the heart of genome mapping and sequencing. It can substantially enhance the capabilities of diagnosis, pharmacogenetics, and forensic tests. Compared to the commonly used gel-based electrophoresis, microfluidic channel electrophoresis provides clear advantages of improved separation speed, reduced consumption of reagents, and ease of automation.38–40 However, it is difficult to introduce high-viscous polymer solution into microscale channel and use

266

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

such simple device for long DNA separation. Microfluidic devices with integrated pillar arrays can then be used. On the one hand, using nanopillars obviates the introduction of gel matrix. On the other hand, during electrophoresis, DNA molecules will not be trapped by nanopillar arrays as that happened in gel matrix. With such a device, new separation mechanisms can be studied in correlation with theoretical modeling. The conformation change of long DNA molecules plays an important role during electrophoretic separation. Normally, a long DNA molecule in its relax state has a spherical shape. When migrating under electric field in nanopillars’ sieving matrix, the long DNA molecules are stretched. This deformation is not entropically favored and the stretched DNA molecules try to escape from the trap. Indeed, real-time images of a single T4 DNA molecule in the nanopillar region under an electric field show the step-by-step changes in conformation of the DNA molecule from spherical to linear.41 In the same nanopillar arrays, the migration mobility of different long DNA molecules is different. By varying the size and geometry of the nanopillar arrays, one can easily adjust the migration speed of long DNA molecules and hence separate them in a short time limit. By using electron beam lithography and reactive ion etching, Kaji et al. have integrated high aspect ratio (100–500 nm diameter and 500–5000 nm tall) nanopillars inside a microchannel42 (Figure 7.4). Then, DNA fragments of 1–38 kbp were separated into clear bands in a detection window of 1450 mm from the entrance of the nanopillar channel (25 mm in width and 2.7 mm in height) in only 170 s. The technique has also been applied to the separation of long DNA molecules (l-phage DNA: 48 kbp, T4 DNA: 165.6 kbp) in less than 30 s under a DC electric field. By using nanoimprint lithography, it is also possible to produce some devices. Thermal nanoimprint lithography has been first used for the fabrication of a pillar array

FIGURE 7.4 (a and b) Photography of a microfluidic chip for electrophoresis. (c) SEM image of integrated quartz pillars. (d and e) Fluorescence images of migrating single l DNA and T4 DNA molecules in the pillar region. (f) Electropherograms of l DNA and T4 DNA recorded at two detection points.42

APPLICATION EXAMPLES

267

of 150 nm diameter and 320 nm period into microfluidic channels.43 Alternatively, a nanoembossing technique has been developed to demonstrate the feasibility of onestep replication of nanopillar arrays and microfluidic channel for DNA molecule separation.44 More recently, a highly parallel and mix-and-match fabrication method based on soft UV nanoimprint lithography and contact photolithography has been proposed.26 Two pillar arrays with same period but different pore size are integrated into microchannels. Single DNA molecule migration behavior has been demonstrated and, consequently, l DNA and T4 DNA were successfully separated in a couple of minutes. The advantage of this method relies on the much improved process latitude of the soft UV-NIL and easy implementation of whole fabrication steps. It is also versatile and relevant to manufacture this type of nanodevice at low cost and high throughput. 7.4.1.3 DNA and Protein Arrays Currently, DNA and protein arrays are widely used for different purposes. For example, the microarray chips are used to obtain new insights into how genes and proteins work and how they are linked to the disease. They are also used for cancer diagnostic and disease analyses. To improve the performance of the microarray chips, one can, for instance, increase the density of the array, increase the signal to noise ratio, and develop technologies based on “lab-on-chip.” Among other possibilities, surface patterning can be used to increase both array density and signal to noise ratio. Previously, we performed a proof of concept experiment by integrating high aspect ratio and high-density nanopillars in the hybridization zones of a microarray chip, all surrounded by superhydrophobic surfaces. The nanopillars were produced by soft nanoimprint lithography and reactive ion-etching techniques. Then, by using the protocol described in Ref. 45, IgG molecules could be immobilized on the surfaces of nanopillars and fluorescentlabeled antigens could be easily attached through the bioaffinity interaction between antibody and antigen. As a result, the nanopillar areas show a higher fluorescent intensity than that in the surrounding areas (Figure 7.5a). By changing the size and

FIGURE 7.5 (a) SEM images of a SiO2 pillar array of 800 nm period and 1.6 mm height obtained by soft UV nanoimprint lithography. (b) Fluorescence image (lower) and intensity profile (upper) of top view of a microchannel with integrated pillars, observed after a surfacedependent chemical reaction.47

268

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

height of the nanopillars, a linear dependence of the fluorescence intensity versus the effective surface could be obtained (Figure 7.5b). To increase the detection sensitivity, the surrounding areas were also patterned, which become superhydrophobic to prevent the undesirable deposition or the cross-contamination.46 For demonstration, the fabricated device has been used to show that the presence of the high-density pillars and the superhydrophobic areas improved both the bioreaction efficiency and the detection sensibility. The surface-dependent enzymatic reaction has also been studied in microfluidic channels. Figure 7.5c shows that increasing the effective surface per unit area (by etching) leads to an increase of fluorescence intensity, which corresponds to the increase of the total amount of the enzyme attached to the microfluidic channel wall.47 7.4.2 Pillars for Cellomics Cells are the elementary building blocks of mammalians and many other living systems. For in vitro studies, cells can be cultured, proliferated, and differentiated in a culture dish or on other types of supports. In particular, patterned substrates can be used for the control of cell organization and cell kinetics. The size of a cell is typically several tens of micrometers so that their growth behavior is largely influenced by the topographic or chemical patterns of the comparable sizes. It is also known that many key functionalities of cells are also regulated by nanoscale complexes of subcellular systems. For example, the building blocks of both cytoskeleton (microtubules and actin filaments) and extracellular matrix have dimensions in the range of 10 nm. Therefore, it is interesting to investigate the influence of synthetic surface made of high-resolution patterns. Knowing that patterned micro- and nanostructures can be easily integrated into microfluidic devices, a large variety of cell culture and manipulation techniques such as on-chip cell sampling, trapping, sorting, characterization, and so on can be developed. 7.4.2.1 Cell Trapping Pillar arrays can be used for size-dependent and/or affinity-dependent cell trapping. Nagrath et al. presented isolation of circulating tumor cells (CTCs) in cancer patients with a simple microfluidic platform embedded with pillar arrays (Figure 7.6).48 The pillar arrays are first coated with antiepithelial cell adhesion molecule antibody. Under precisely controlled laminar flow conditions, viable CTCs can be isolated from peripheral whole blood samples, mediated by the interaction of target CTCs with antibody without requisite of prelabeling or processing of samples. Indeed, CTCs in the peripheral blood of patients with metastatic lung, prostate, pancreatic, breast, and colon cancer can be trapped in 115 of 116 (99%) samples with a concentration in the range of 5–1281 CTCs per milliliter and 50% purity. In addition, CTCs were isolated in 7/7 patients with early stage prostate cancer. Magnetic cell trapping can be done by using magnetic beads, also mediated by antibody–antigen interaction. By using microfluidic devices with patterned magnetic microcolumns, cancer cells could be most efficiently trapped.49 In this study, a hexagonal array of nickel pillars was fabricated by standard photolithography

APPLICATION EXAMPLES

269

FIGURE 7.6 Pillar-based trapping of circulating tumor cells: (a) the workstation setup, showing the manifold housing of CTC chip, a rocker for continuous mixing, and a pneumatic pressure-regulated pump for controlling the sample flow; (b) SEM image of a captured NCI-H1650 lung cancer cell spiked into blood. The inset shows a high magnification view of the cell.48

and electroplating. Then, they were integrated into microfluidic channels. Using a solenoid coil with an iron core, a large magnetic field gradient could be generated in the pillar areas so that supermagnetic beads were trapped reversibly. After an in situ surface modification with wheat germ agglutinin, A549 cancer cells were effectively captured from cell mixtures with trapping efficiency in the range from 62% to 74%. Cell trapping can also be done based on dielectrophoretic interaction. More generally, the dielectrophoretic forces allow manipulating single cell by changing the voltage, frequency, or phase of the electrical signal applied to the electrodes of the devices. For example, a negative dielectrophoretic (nDEP) force can be applied to generate traps at the stagnation points of cylindrical pillars arranged in a regular array.50 By carefully controlling the dielectrophoretic and hydrodynamic forces, both single-particle traps (capable of discriminating particles based on size) and multiparticle traps (capable of controlling the number of particles trapped) could be achieved with high precision. Although most of the previous studies were performed with polystyrene beads, such a trapping mechanism is certainly applicable for cells. 7.4.2.2 Cell Cultivation Not surprisingly, the most recent works of cell cultivation on patterned surface were motivated by tissue engineering, regenerative medicine, and implantable medical devices. It now becomes also important to integrate them into microfluidic chips. The advantage of using microfluidic devices for cell culture is multifold. Apart from the most evident facts of single-cell manipulation and high-throughput screening with much reduced samples and biochemical reagents, microfluidics allows the creation of a controllable and biomimetic microenvironment with high spatial and temporal resolution. In addition to the control of soluble cell factor, microfluidic devices also allow to integrate different types of materials and patterns as extracellular matrix. In

270

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

FIGURE 7.7 LRM55 cells after 6 h grown on strips of pillars. (a) Reflectance image, (b) fluorescence image, and (c) merged image of (a) and (b) (scale bar: 50 mm).51

particular, nanopillars can now be produced at low cost and high throughput, providing a new dimension for enhanced cell adhesion, proliferation, and differentiation. Finally, hybrid micro-, nano-, and chemical patterns are useful for the controlled formation of cellular networks. For example, Turner et al. have studied attachment of astroglial cells on smooth silicon and silicon pillars and strips with various width and separations.51 Fluorescence, reflectance, and confocal light microscopes, as well as scanning electron microscopy, were used to observe the cell morphology and the distribution of cytoskeletal proteins such as actin and vinculin on different surfaces (Figure 7.7). It turned out that both actin and vinculin distributions were highly polarized when cells were placed on the pillar arrays. Scanning electron microscopy clearly demonstrated that cells made contact with the top of the surface pattern and they did not reach the bottom even when the patterned features were separated from each other with a distance as large as 5.0 mm. These experiments support the use of surface topography to direct the attachment, growth, and morphology of cells. Investigations on other types of materials also showed the very significant differences of the cell behaviors on flat and patterned surfaces, indicating that the conventional dish-based (two-dimensional) culture will not be sufficient for understanding or generating mimetic (three-dimensional) tissues. Nomura et al. have reported cell culture on nanopillar arrays.52 By using nanoimprint lithography, nanopillars could be produced as a new type of cell culture medium. Then, HeLa cells were cultured and analyzed, showing that the cell division and proliferation on nanopillars were significantly different from that found with culture Petri dishes. In addition, the use of nanopillar substrates allows an easy subculture without conventional trypsinization. More recently, Kim et al. used nanopillar arrays made of a poly(ethylene glycol) (PEG) hydrogel to guide a 3D construct of primary rat cardiomyocytes.53 They observed that the PEG nanopillars not only guided the extension of cell membranes but also led to the 3D growth of cardiomyocytes with a new topographical guiding mechanism. In colonizing cardiomyocytes, the PEG nanopillars stimulated selfassembled aggregates among the contacting cells, in comparison to those on the bare PEG and the glass control. The myocytes cultured on the PEG nanostructure also

APPLICATION EXAMPLES

271

FIGURE 7.8 (a) SEM of aggregated cardiomyocytes cultured on PEG nanopillars, (b and c) electric properties of rat cardiomyocytes cultured on the pillar array and glass using the whole cell patch clamp technique (scale bars: 20 mV/100 ms). (d and e) Representative trace of action potentials recorded with PEG pillar substrate and glass (scale bars: 20 mV/2 s).53

exhibited substantial beating capability with higher amplitude when compared to those cultured on controls (Figure 7.8). Finally, protein patterns can be integrated into microfluidic chambers for cell attachment at single-cell level and long-term cell culture.54 After patterning and device integration, cells were seeded in and cultured for more than one week. In comparison to more conventional culture, significant changes of the cell morphology and protein expression could be observed: cells can form highly elongated features that cannot be observed with a protein pattern or in a microchamber alone. 7.4.2.3 Cell Characterization In living systems, cells adhere and interact with the extracellular matrix via cell–cell, cell–materials contacts. They can sense and respond to the signal arising in their local environment through ion channels and receptors present in their membranes, which are often associated with groups of proteins linked to the cytoskeleton. To obtain more clear insights into the signaling pathways activated by the membrane proteins, new functionalities are integrated into microfluidic devices. These integrated functionalities can be used not only to monitor the microenvironment or the inner parameters of cells but also to generate local stimulations for more precise determination of cellular reactivity. For example, integrated microelectrodes are used for analysis of intercellular or intracellular parameters such as pH, conductivity and the concentration of signaling molecules, and the electrochemical signature of peroxynitrite oxidation.55 For chemical sensing, cell adhesion behaviors could be regulated by mono- or mixed chemical patterns defined by different patterning techniques.56–58 In addition, the pattern size could be turned from micrometers to nanometers. For cellular mechanical properties, different techniques such as magnetic tweezers, atomic microscopy, microplates, and so on are used.59–61 Pillar arrays are also used for cell mechanical sensing. Elastomeric pillars made of PDMS are first used to measure the cell traction forces by Tan et al. (Figure 7.9). Based on the known physical parameters of the PDMS

272

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

FIGURE 7.9 (a) SEM of a smooth muscle cell attached to an array of posts uniformly coated with fibronectin. (b) Schematic of microcontact printing of protein. (c) Differential interference contrast (upper) and immunofluorescence (lower) graphs of the same region, only a 2  2 array of posts printed with fibronectin. (d) SEM of a smooth muscle cell attached to posts where only the tops of the posts have been printed with fibronectin. (e–g) Confocal images of immunofluorescence staining of a smooth muscle cell on posts. The force exerted by cells (white arrows) was calculated through the position change of the posts. The force map was spatially correlated to immunofluorescence localization of the focal adhesion protein vinculin. (h) Plot of the force generated on each post as a function of total area of focal adhesion staining per post (scale bars indicate 10 mm).62

pillars, deflections can be measured and forces can be derived more accurately. Then, they found that the intracellular force generated in a cell varied with cell spreading such that well spread cells exerted more average force per post than the less spread counterparts. They also confirmed earlier studies that the magnitude of the force exerted by cells correlated with the size of adhesion formed by attaching cells.62 By using the same technique, Roure et al. measured dynamic traction forces exerted by epithelial cells migration.63 In these two examples, the pillars are used as probe to measure the forces that were exerted by the cells. Compared to the microbeads or cantilever-based sensing, the pillar-based sensing is simpler and more accurate, allowing the determination over a large range the forces exerted on cells. Pillars can also be used to generate forces on cells to mimic the forces encountered in vivo. Sniadecki et al. employed magnetic nanowires embedded in the PDMS pillars to activate and measure the traction forces in cells.64 External forces up to 45 nN were achieved with a uniform magnetic field. Also, it was verified that the magnetic pillars can be driven at frequencies up to 5 Hz without significant damping effects.

APPLICATION EXAMPLES

273

7.4.3 Pillars for Biosensing In biomedical research, healthcare, and environmental monitoring, analysis tools are generally expensive but not enough flexible. Microfluidic devices are developed for fast and cost-effective analyses. Then, the integration of different sensing elements becomes important and many efforts have been devoted to this particular area. Optofluidics, for example, has been developed by integrating both active and passive optical elements into microfluidic chips.65 It is known that fluorescence detection is the most sensitive one but it requires specified molecular labeling. The interaction between the target and recognition molecules could be characterized by the intensity of the fluorescence. Although the detection limit can reach a single-molecule level, the contaminating materials in the sample can frequently interfere with the detection system. For this reason, highly sensitive, label-free detection is desired for both in vivo and in vitro analyses. As nanopillar arrays can be designed to have very particular electromagnetic or optical properties, they can be used to improve the sensibility of label-free detection in a microfluidic device. Among many others, three types of substrates are now widely studied for improved label-free detection, including surface plasmon detection, surface enhanced Raman spectroscopy (SERS), and photonic crystal-based analyses, all relying on the fabrication of high-density nanopatterns. For surface plasmon-based sensors, Chen and Jiang introduced hybrid arrays of metallic nanostructures obtained by depositing a silver film on the fused silica nanopillars.66 Such a structure was used to monitor the evaporation process of the absolute ethanol on the sample surface, showing two distinct peaks in the extinction spectrum of the p-polarized incident light, one at 585.3 nm and the other at 493.6 nm. With the addition of absolute ethanol on the sample surface, a redshift of 32.9 nm was observed for the higher peak while a blueshift of 42.3 nm was observed for the lower peak. It was also found that narrowest extinction peak appeared at normal incidence, while the polarization of the incident light did not affect the experimental result due to the symmetrical distribution of the nanostructures. SERS-based biosensing also depends on the shape of individual pillars. In general, sharp corners provide more pronounced SERS signals. Raman spectroscopy is also an ideal label-free optical detection technique for chemical and biomolecules. Lee and Liu described a low-cost and ultrasensitive substrate for SERS measurement in a microfluidic device (Figure 7.10).67 PDMS nanopillars were fabricated by soft lithography. After selective deposition of Ag thin film, the pillars were integrated into a glass-based microfluidic chip with a suitable optical window for SERS spectroscopic imaging. Rhodamine 6G and adenosine SERS spectra were then obtained by using a 785 nm wavelength laser excitation. As a result, the observed Raman scattering signal enhancement on the nanopillar-based Ag SERS substrate is more than 107 times higher than the control sample. Photonic crystals have recently been demonstrated as optical biosensors. Due to the highly localized confinement of the coupled light, photonic crystal sensors can be incorporated into microfluidic devices to facilitate localized measurements of the change in refractive index. Wu et al. presented the concept of using three-dimensional photonic crystals for refractive index sensing in a microfluidic channel.68 It was

274

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

FIGURE 7.10 (a) Schematic diagram of the chip with integrated patterns and the Raman imaging system. (b and c) Optical image of a microchannel with a laser focal spot illuminated on the surface of Ag/PDMS substrate with and without patterns. (d) SERS spectra of 1 mM R6G molecules taken on the area with (upper) or without (lower) patterns (1 s integration time). (e) SERS spectra of, from top to bottom, 1 mM, 10 nM, 100 pM, 1 pM, and 10 fM adenosine molecules taken from the Ag/PDMS nanowell SERS substrate, DI water, and the Raman spectrum of 10 mM adenosine molecules taken from the smooth Ag/PDMS substrate, respectively.67

demonstrated that a change in the refractive index of the fluid in a microchannel results in a shift in the band gap or band gap defect position of the photonic crystal. According to Fourier transform infrared spectroscopy of the photonic crystal sensor, a change of 6  10
3 in the refractive index of the fluid can be detected. 7.4.4 Other Applications Pillar arrays integrated microfluidic chips can have many other applications. Even for the most explored domain, only a few aspects have been investigated. Sometimes, one can use the same material and the same design to study different properties of a fabricated pillar array. For example, a nanopillar array made of zinc oxide (ZnO) can be used for optical detection, piezoelectric sensing, field emission, and so on. Owing to its high chemical stability, Bie et al. used ZnO nanopillars as gas sensor for hydrogen and ethanol.70 Their samples exhibited large responses of 18.29 and 10.41–100 ppm ethanol and hydrogen, respectively. Alternatively, Pradhan et al. obtained ZnO nanopillars by electrodeposition on indium tin oxide (ITO) glass substrates without any template catalyst or seed layer.69 They showed that these pillars can be used to achieve an excellent field emission performance, with a low turn-on electric field of 3.2 V mm
1 for 1.0 mA cm
2 and a threshold field of 6.6 V mm
1 for 1.0 mA cm
2.

REFERENCES

275

7.5 CONCLUSION AND FUTURE OUTLOOK We have attempted to review the current research topics on pillars and pillar arrays integrated into microfluidic devices. Examples were given to illustrate their high potential in genomics, proteomics, and cellomics studies and we expect a fast growing interest in this emerging field. From the fabrication point of view, new materials and new surface protocols can be used to enhance the functionality of pillars and integrated pillar arrays. From the application point of view, new topics are to be explored for both fundamental research and commercial purposes. With rapid progress in cell and molecular biology, nanofabrication and microdevice technologies will be spread as a general undertaking in coming years. ACKNOWLEDGMENTS This work was partially supported the European Commission through project contract NMP4-CT-2003-505311 (Nabis) and project contract CP-FP 214566-2 (Nanoscales). The authors would also like to thank colleagues and students of CNRS-LPN (Marcoussis, France) and ENS (Paris, France) for collaboration and assistance. REFERENCES 1. Roco, M.C.; Williams, R.S.; Alivisators, P. Biological, medical and health applications. In Nanotechnology Research Directions; Kluwer Academic Publishers: 2000; Chapter 8. 2. Niemeyer, C.M.; Mirkin, C.A. Nanobiotechnology: Concepts, Applications and Perspectives, Wiley: 2004. 3. Roco, M.C. Nanotechnology: convergence with modern biology and medicine. Curr. Opin. Biotechnol. 2003, 14, 337–346. 4. Whitesides, G.M. The “right size” in nanobiotechnology. Nat. Biotechnol. 2003, 18, 760–763. 5. Fortina, P.; Krick, L.J.; Surrey, S.; Grodzinski, P. Nanobiotechnology: the promise and reality of new approaches to molecular recognition. Trends Biotechnol. 2005, 23, 168–173. 6. Ball, P. Natural strategies for the molecular engineer. Nanotechnology 2002, 13, R15–R28. 7. Huang, S. Gene expression profiling, genetic networks, and cellular states: an integrating concept for tumorigenesis and drug discovery. J. Mol. Med. 1999, 77, 469–480. 8. Ito, Y. Surface micropatterning to regulate cell functions. Biomaterials 1999, 20, 2333–2342. 9. Girard, P.P.; Cavalcanti-Adam, E.A.; Kemkemer, R.; Spatz, J.P. Cellular chemomechanics at interfaces: sensing, integration and response. Soft Matter 2007, 3, 307–326. 10. Sun, M.H.; Luo, C.X.; Xu, L.P.; Ji, H.; Qi, O.Y.; Yu, D.P.; Chen, Y. Artificial lotus leaf by nanocasting. Langmuir 2005, 21, 8978–8981. 11. Kertesz, K.; Balint, Z.; Vertesy, Z.; Mark, G.I.; Lousse, V.; Vigneron, J.P.; Rassart, M.; Biro´, L.P.; Phys. Rev. E 2006, 74, 1539–3755.

276

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

12. Duffy, D.C.; McDonald, J.C.; Schueller, O.J.A.; Whitesides, G.M. Rapid prototyping of microfluidic systems in poly(dimethylsiloxane). Anal. Chem. 1998, 70, 4974–4984. 13. Squires, T.M.; Quakes, S.R. Microfluidics: fluid physics at the nanoliter scale. Rev. Mod. Phys. 2005, 77, 977–1026. 14. Unger, M.A.; Chou, H.P.; Thorsen, T.; Scherer, A.; Quake, S.R. Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 2000, 288, 113–116. 15. Andersson, H.; van den Berg, A. Microfluidic device for cellomics: a review. Sens. Actuators B 2003, 92, 315–325. 16. McClain, M.; Culbertson, C.T.; Jacobson, S.C.; Allbritton, N.L.; Sims, C.E.; Ramsey, J.M. Microfluidic devices for the high-throughput chemical analysis of cells. Anal. Chem. 2003, 75, 5646–5655. 17. Gao, J.; Yin, X.F.; Fang, Z.L. Integration of single cell injection, cell lysis, separation and detection of intracellular constituents on a microfluidic chip. Lab Chip 2004, 4, 47–52. 18. Galas, J.C.; Torres, M.; Belotti, M.; Kou, Q.; Chen, Y. Microfluidic tunable dye laser with integrated mixer and ring resonator. Appl. Phys. Lett. 2005, 86, 264101–264104. 19. Galas, J.C.; Peroz, C.; Kou, Q.; Chen, Y. Microfluidic dye laser intracavity absorption. Appl. Phys. Lett. 2006, 89, 224101–224103. 20. Xia, Y.N.; Whitesides, G.M. Soft lithography. Angew. Chem. Int. Ed. 1998, 37, 550–575. 21. Chou, S.Y.; Krauss, P.R.; Renstrom, P.J. Imprint of sub-25 nm vias and trenches in polymers. Appl. Phys. Lett. 1995, 67, 3114–3116. 22. Haisma, J.; Verheijen, M.; Heuvel, K.V.D.; Berg, J.V.D. Mould-assisted nanolithography: a process for reliable pattern replication. J. Vac. Sci. Technol. B 1996, 14, 4124–4128. 23. Chen, Y.; Carcenac, F.; Ecoffet, C.; Lougnot, D.J.; Launois, H. Mold-assisted near-field optical lithography. Microelectron. Eng. 1999, 46, 69–72. 24. Chen, Y.; Roy, E.; Kanamori, Y.; Belotti, M.; Decanini, D. Soft nanoimprint lithography. Proc. SPIE 2005, 5645, 283–288. 25. Roy, E.; Kanamori, Y.; Belotti, M.; Chen, Y. Enhanced UV imprint ability with a tri-layer stamp configuration. Microelectron. Eng. 78/79, 2005, 689–694. 26. Shi, J.; Fang, A.P.; Malaquin, L.; Pepin, A.; Decanini, D.; Viovy, J.L.; Chen, Y. Highly parallel mix-and-match fabrication of nanopillar arrays integrated in microfluidic channels for long DNA molecule separation. Appl. Phys. Lett. 2007, 91, 153114–153117. 27. Kraus, T.; Malaquin, L.; Delamarche, E.; Schmid, H.; Spencer, N.D.; Wolf, H. Closing the gap between self-assembly and microsystems using self-assembly, transfer, and integration of particles. Adv. Mater. 2005, 17, 2438–2442. 28. Kosiorek, A.; Kandulski, W.; Glaczynska, H.; Giersig, M. Fabrication of nanoscale rings, dots, and rods by combining shadow nanosphere lithography and annealed polystyrene nanosphere masks. Adv. Mater. 2005, 1, 439–444. 29. Tang, Z.K.; Wong, G.K.L.; Yu, P.; Kawasaki, M.; Ohtomo, A.; Koinuma, H.; Segawa, Y. Room-temperature ultraviolet laser emission from self-assembled ZnO microcrystallite thin films. Appl. Phys. Lett. 1998, 72, 3270–3272. 30. Rensmo, H.; Keis, K.; Lindstrom, H.; Solbrand, A.; Hagfeldt, A.; Lindquist, S.E. High light-to-energy conversion efficiencies for solar cells based on nanostructured ZnO electrodes. J. Phys. Chem. B 1997, 101, 2598–2601.

REFERENCES

277

31. Fujibayashi, T.; Matsui, T.; Kondo, M. Improvement in quantum efficiency of thin film Si solar cells due to the suppression of optical reflectance at transparent conducting oxide/Si interface by TiO2/ZnO antireflection coating. Appl. Phys. Lett. 2006, 88, 183508–183510. 32. Chen, G.F.; McCandless, G.T.; McCarley, R.L.; Soper, S.A. Integration of large-area polymer nanopillar arrays into microfluidic devices using in situ polymerization cast molding. Lab Chip 2007, 7, 1424–1427. 33. Christel, L.A.; Petersen, K.; McMillan, W.; Northrup, M.A. Rapid automated nucleic acid probe assays using silicon microstructures for nucleic acid concentration. J. Biomech. Eng. 1999, 121, 22–27. 34. Witek, M.A.; Llopis, S.D.; Wheatley, A.; McCarley, R.L.; Soper, S.A. Purification and preconcentration of genomic DNA from whole cell lysates using photoactivated polycarbonate (PPC) microfluidic chips. Nucleic. Acids Res. 2006, 34, e74. 35. Foote, R.S.; Khandurina, J.; Jacobson, S.C.; Ramsey, J.M. Microfabricated porous membrane structure for sample concentration and electrophoretic analysis. Anal. Chem. 2005, 77, 57–63. 36. Plecis, A.; Nanteuil, C.; Haghiri-Gosnet, A.M.; Chen, Y. Electropreconcentration with charge-selective nanochannels. Anal. Chem. 2008, 80, 9542–9550. 37. Dauriac, V.; Descroix, S.; Chen, Y.; Peltre, G.; Senechal, H. Isoelectric focusing in an ordered micropillar array. Electrophoresis 2008, 29, 2945–2952. 38. Harrison, D.J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C.S.; Manz, A. Micromachining a miniaturized capillary electrophoresis-based chemical analysis system on a chip. Science 1993, 261, 895–897. 39. Tegenfeldt, J.O.; Prinz, C.; Cao, H.; Huang, R.L.; Austin, R.H.; Chou, S.Y.; Cox, E.C.; Sturm, J.C. Micro- and nanofluidics for DNA analysis. Anal. Bioanal. Chem. 2004, 378, 1678–1692. 40. Fu, J.; Schoch, R.B.; Stevens, A.L.; Tannenbaum, S.R.; Han, J. A patterned anisotropic nanofilter array for continuous-flow separation of DNA and proteins. Nat. Nanotechnol. 2006, 2, 121–128. 41. Han, J.; Craighead, H.G. Separation of long DNA molecules in a microfabricated entropic trap array. Science 2000, 288, 1026–1029. 42. Kaji, N.; Tezuka, Y.; Takamura, Y.; Ueda, M.; Nishimoto, T.; Nakanishi, H.; Horiike, Y.; Baba, Y. Separation of long DNA molecules by quartz nanopillar chips under a direct current electric field. Anal. Chem. 2004, 76, 15–22. 43. Studer, V.; Pepin, A.; Chen, Y. Nanoembossing of thermoplastic polymers for microfluidic applications. Appl. Phys. Lett. 2002, 80, 3614–3616. 44. Pepin, A.; Youinou, P.; Studer, V.; Lebib, A.; Chen, Y. Nanoimprint lithography for the fabrication of DNA electrophoresis chips. Microelectron. Eng. 61/62, 2002, 927–932. 45. Zhang, Z.L.; Crozatier, C.; Le Berre, M.; Chen, Y. In situ bio-functionalization and cell adhesion in microfluidic devices. Microelectron. Eng. 78/79, 2005, 556–562. 46. Guo, S.S.; Sun, M.H.; Shi, J.; Liu, Y.J.; Huang, W.H.; Combellas, C.; Chen, Y. Patterning of hydrophilic micro arrays with superhydrophobic surrounding zones. Microelectron. Eng. 2007, 84, 1673–1676. 47. Shi, J.; Peroz, C.; Peyrade, D.; Salari, J.; Belotti, M.; Huang, W.H.; Chen, Y. Tri-layer soft UV imprint lithography and fabrication of high density pillars. Microelectron. Eng. 2006, 83, 1664–1668.

278

PILLARS AND PILLAR ARRAYS INTEGRATED IN MICROFLUIDIC CHANNELS

48. Nagrath, S.; Sequist, L.V.; Maheswaran, S.; Bell, D.W.; Irimia, D.; Ulkus, L.; Smith, M.R.; Kwak, E.L.; Digumarthy, S.; Muzikansky, A.; Ryan, P.; Balis, U.J.; Tompkins, R.G.; Haber, D.A.; Toner, M. Isolation of rare circulating tumor cells in cancer patients by microchip technology. Nature 2007, 450, 1235–1239. 49. Liu, Y.J.; Guo, S.S.; Zhang, Z.L.; Huang, W.H.; Baigl, D.; Xie, M.; Chen, Y.; Pang, D.W. A micropillar-integrated smart microfluidic device for specific capture and sorting of cells. Electrophoresis 2007, 28, 4713–4722. 50. Cui, H.H.; Lim, K.M. Pillar array microtraps with negative dielectrophoresis. Langmuir 2009, 25, 3336–3339. 51. Turner, A.M.P.; Dowell, N.; Turner, S.W.P.; Kam, L.; Isaacson, M.; Turner, J.N.; Craighead, H.G.; Shain, W. Attachment of astroglial cells to microfabricated pillar arrays of different geometries. J. Biomed. Mater. Res., Part A 2000, 51, 430–441. 52. Nomura, S.; Kojima, H.; Ohyabu, Y.; Kuwabara, K.; Miyauchi, A.; Uemura, T. Cell culture on nanopillar sheet: study of HeLa cells on nanopillar sheet. Jpn. J. Appl. Phys. 2005, 44, 1184–1186. 53. Kim, D.H.; Kim, P.; Song, I.; Cha, J.M.; Lee, S.H.; Kim, B.; Suh, K.Y. Guided threedimensional growth of functional cardiomyocytes on polyethylene glycol nanostructures. Langmuir 2006, 22, 5419–5426. 54. Wang, L.; Lei, L.; Ni, X.F.; Shi, J.; Chen, Y. Patterning biomolecules for cell attachment at single cell levels in PDMS microfluidic chips. Microelectron. Eng., doi: 10.1016/j.mee. 2009.01.030. 55. Amatore, C.; Arbault, S.; Bruce, D.; De Oliveira, P.; Erard, M.; Vuillaume, M. Characterization of the electrochemical oxidation of peroxynitrite: relevance to oxidative stress bursts measured at the single cell level. Chem. Eur. J. 2001, 19, 4171–4179. 56. Singhvi, R.; Kumar, A.; Lopez, G.P.; Stephanopoulos, G.N.; Wang, D.I.C.; Whitesides, G. M.; Ingber, D.E. Engineering cell shape and function. Science 1994, 264, 696–698. 57. Thery, M.; Racine, V.; Pepin, A.; Piel, M.; Chen, Y.; Sibarita, J.B.; Bornens, M. Adhesion controls the division axis of animal cells. Nat. Cell Biol. 2005, 7, 947–953. 58. Guo, L.J.; Cheng, X.; Chou, C.F. Fabrication of size-controllable nanofluidic channels by nanoimprinting and its application for DNA stretching. Nano Lett. 2004, 4, 69–73. 59. Amblard, F.; Maggs, A.C.; Yurke, B.; Pargellis, A.; Leibler, S. Subdiffusion and anomalous local viscoelasticity in actin networks. Phys. Rev. Lett. 1996, 77, 4470–4473. 60. Ando, T.; Kodera, N.; Naito, Y.; Kinoshita, T.; Furuta, K.; Toyoshima, Y.Y. A high-speed atomic force microscope for studying biological macromolecules in action. ChemPhysChem 2003, 4, 1196–1202. 61. Thoumine, O.; Ott, A. Time scale dependent viscoelastic and contractile regimes in fibroblasts probed by microplate manipulation. J. Cell Sci. 1997, 110, 2109–2116. 62. Tan, J.L.; Tien, J.; Pirone, D.M.; Gray, D.S.; Bhadriraju, K.; Chen, C.S. Cells lying on a bed of microneedles: an approach to isolate mechanical force. PNAS 2003, 100, 1484–1489. 63. du Roure, O.; Saez, A.; Buguin, A.; Austin, R.H.; Chavrier, P.; Silberzan, P.; Ladoux, B. Force mapping in epithelial cell migration. PNAS 2005, 102, 2390–2395. 64. Sniadecki, N.J.; Anguelouch, A.; Yang, M.T.; Lamb, C.M.; Liu, Z.; Kirschner, S.B.; Liu, Y.; Reich, D.H.; Chen, C.S. Magnetic microposts as an approach to apply forces to living cells. PNAS 2007, 104, 14553–14558.

REFERENCES

279

65. Psaltis, D.; Quake, S.R.; Yang, C.H. Developing optofluidic technology through the fusion of microfluidics and optics. Nature 2006, 442, 381–386. 66. Chen, X.; Jiang, K. A large-area hybrid metallic nanostructure array and its optical properties. Nanotechnology 2008, 19, 215303–215308. 67. Liu, G.L.; Lee, L.P. Nanowell surface enhanced Raman scattering arrays fabricated by soft-lithography for label-free biomolecular detections in integrated microfluidics. Appl. Phys. Lett. 2005, 87, 074101–074103. 68. Wu, J.; Day, D.; Gu, M. A microfluidic refractive index sensor based on an integrated threedimensional photonic crystal. Appl. Phys. Lett. 2007, 92, 071108–071110. 69. Pradhan, D.; Kumar, M.; Ando, Y.; Leung, K.T. Fabrication of ZnO nanospikes and nanopillars on ITO glass by templateless seed-layer-free electrodeposition and their fieldemission properties. ACS Appl. Mater. Interfaces 2009, 126, 604–608. 70. Bie, L.J.; Yan, X.N.; Yin, J.; Duan, Y.Q.; Yuan, Z.H. Nanopillar ZnO gas sensor for hydrogen and ethanol. Sens. Actuator B 2007, 126, 604–608.

8 NANOCATALYSIS IN MICROREACTOR FOR FUELS SHIHUAI ZHAO1,2

AND

DEBASISH KUILA1,3

1

Institute for Micromanufacturing, Louisiana Tech University, Ruston, LA, USA Tianjin University, Tianjin, China 3 Department of Chemistry, North Carolina A&T State University, Greensboro, NC, USA 2

8.1 INTRODUCTION The first computer occupied several rooms when it was born. Thanks to the development of microelectronic technology, small personal computers with dramatically improved processing powers and memory capacity have now become common at home and in the workplace. Even the palm-sized computer is not a fairy tale anymore. Chemists and chemical engineers are trying to bring the same kind of revolution to the chemical plants. Traditionally, the chemical industry operates reactors in a few large facilities to achieve economies of scale. Today, a small group of research institutes and companies in the chemical process industries are working in the opposite direction, that is, developing micrometer-scaled reactors, the socalled microreactor, as well as integrating microstructured heat exchangers, pumps, valves, and other devices to go with them. The idea of chemical processing on a chip is brought step by step into reality by employing the microelectronic and micromachining technology. Microreactors exhibit many practical advantages when compared with conventional reactors, not least is the demand for a high standard of safety, such as the transportation and storage of toxic, explosive, or harmful materials.1 In such cases, microreactors offer the capability to carry out production on site at the point of demand. The removal of potentially significant large-scale plant accidents associated with thermal runaway reactions could also be envisaged Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

281

282

NANOCATALYSIS IN MICROREACTOR FOR FUELS

due to inherent thermal dissipation possible in microreactor devices. Indeed, it has been demonstrated that reactions can be performed beyond their current explosive limits by adopting microreactor technology. In addition to the environmental and safety benefits of dealing with smaller quantities of material, microreactors offer many performance benefits compared to traditional batch-scale reactors. Most of these advantages stem from the high surface to volume ratio, which is a consequence of the decrease in fluid layer thickness in microscale reactors. The decrease of linear dimensions in a microreactor changes the properties of chemical processing, such as temperature, pressure, and so on, which refers to better heat transfer and mass transport. The heat management in microscale, enabling mass and heat transfer to be extremely rapid, leads inevitably to a higher level of reaction control and reactant manipulation at any one point within a microreactor. In particular, microreactors’ narrow channels (typical width of 5–500 mm) and thin walls (typical thickness of 5–100 mm) make them desirable candidates for studying potentially explosive reactions. The small channels and the large surface to volume ratios serve to both inhibit gas-phase free radical reactions and improve heat transfer for exothermic reactions. Microreactors exhibit different fluid dynamics from conventional reactors due to operations at the microscale. Microchannels have large length to depth ratios proportional to the number of molecular collisions with channel walls. Thus, the reaction efficiency is improved with the enhanced collision frequency. Microreactors are also able to promote the performance of some reactions by shortening the residence time and thereby avoiding side reactions.2 Microreactors promise a relatively simple and quick means for commercializing a chemical process. Conventional scale-up needs going from laboratory scale to a single large reactor unit through a series of costly laboratory experiments, pilot plant stages, and simulations. In contrast, with microreactor system, the desired product output can be achieved simply by combining a number of microreactors in parallel, allowing quick fabrication with low cost. The functional unit of a microreactor, such as mixing zone, is repeated, and fluid connection between these units can be achieved by using distribution lines and flow equipartition zones. Adopting a scale-out philosophy coupled with large-scale microreactor fabrication technology, it is possible to extend the optimization of reaction conditions on a single microdevice.1 Hence, the reaction efficiency and throughput capacity allowing the production of material on a supply and demand basis could be achieved without the need for redesigning the reaction methodology. A parallel array of microreactors guarantees that desired features of a single unit will remain while the whole system size is increased. Meanwhile, a larger number of units result in higher flexibility in adapting production range to varying demand since a certain number of systems can be turned off or additional systems may be added to the production plant. Thus, microreactors offer greater reaction control and selectivity in the field of chemical and biochemical synthesis, which in turn can be optimized through a scale-out methodology creating a safe and efficient approach to discovery and production. The development, optimization, and improvement of catalysts for use in commercially viable processes provides major challenges for scientific research, both with

INTRODUCTION

283

respect to the complexities of the systems being investigated and the practical challenges of the research effort.3 Commercial catalysts, particularly in heterogeneous systems, are often complex multielemental, multicomponent systems that are usually prepared via multistep procedures with the conditions under which each intermediate step is performed often known as impact performance. The advent of combinatorial chemistry in drug discovery, driven by the need to synthesize and screen a large number of candidates for a variety of applications, has provided a new opportunity for catalyst development and screening. A parallel array of microreactors is also a novel tool for screening inorganic materials, catalysts, and so on. These new parallel approaches aim to address the fundamental need to more effectively and efficiently evaluate complex systems to a degree that still provides useful results. On the basis of the advantages of microreactors discussed above, we could obtain potential benefits, such as fast transfer from research results to production, easier scale-up of production capacity, easier start of production at lower costs, and more flexible response to market demand.4 The continuous process in microreactors shortens the contact time of reactants and reaction times due to fast transfer in thin fluid layers. Furthermore, conversion rates may increase due to short diffusion distances, and selectivity may be improved due to more accurate control of contact times in microreactors for gas and liquid-phase reactions covering heterogeneous and homogeneous catalysis, catalytic oxidation, heterocyclic synthesis, and photochemical reactions. The chemical process miniaturization (CPM) technologies will find applications in a variety of ways including environmental sensing and control, improved operation of chemical processes, stronger economic performance through reduced costs, and increased safety for processing hazardous materials, as well as for research and teaching applications across a wide range of scientific disciplines. One major application of microreactors is in battery replacement for portable electric power devices. Various portable electronic devices have recently come into use, and the performance of these devices has improved remarkably.5 However, greater performance leads to greater consumption of electrical power, and it has become difficult to secure a sufficiently long-lasting power source for these portable electronic devices. This is often true even when a conventional lithium-ion battery is used, despite the higher energy density available compared to other secondary batteries. This energy demand is also expected to become more severe with the advent of broadband network devices, and the higher power consumption contributes to environmental pollution arising from the mass disposal of expended batteries. One of the most promising fuel cell technologies appears to be the proton exchange membrane fuel cell (PEMFC) fuelled by hydrogen. However, hydrogen fuel cells present obvious problems with transportation, storage, and distribution. Because of safety, onboard H2 generation by reforming a liquid fuel is being extensively studied for commercialization. Among the liquid fuel to be reformed, methanol remains a prominent candidate because of its low cost, ease of handling, and high energy density. However, as a by-product of stream reforming of methanol (SRM), carbon monoxide can poison the electrode catalyst (Pt) of fuel cells. Thus, a purification process, preferential oxidation of carbon monoxide in H2-rich fuel cell feed, has to follow the SRM process.

284

NANOCATALYSIS IN MICROREACTOR FOR FUELS

The other important application of microreactors is on-site production or analysis of chemicals. In petroleum engineering, the natural feedstocks are transported, sometimes over long distances, to a central plant and converted therein to more valuable products. The small size and remote location of a vast number of feedstocks render exploitation not profitable since neither plant construction nor transport in pipelines is economical.4 Processing in microreactors may be less expensive than using conventional equipment. Gas-to-liquid (GTL) technology, for example, can utilize microreactors on-site to produce higher alkanes from syngas (H2 and CO) for easy transportation and storage. Installation and removal of microreactor systems may be sufficiently fast and flexible toward high productivity due to numbering up of the single unit. This chapter focuses on nanocatalyst development for industrially useful reactions in silicon-based microreactors. Design and fabrication of microreactors were performed using simulation and micromachining techniques used in microelectromechanical systems (MEMS) and semiconductor industry. The use of cyclohexene for hydrogenation and dehydrogenation was considered as a prototype reaction to compare different catalyst coating methods in Si microreactors. The synthesized catalysts for different reactions were characterized using modern tools of surface science and chemical methods. Methanol steam reformer to produce H2 and CO purifier is included for potential microreactor applications in next generation of alternative energy for portable power devices. We have investigated Fischer–Tropsch (FT) synthesis in microreactors, also known as GTL technology, for syngas conversion to higher alkanes. It could solve current difficulties of storage and transportation by converting natural gas into liquid fuels. A parallel array system of microreactors was designed and constructed for screening different nanocatalysts in GTL technology.

8.2 DESIGN OF MICROCHANNEL REACTORS: MICROMIXING 8.2.1 Principles and Types of Micromixing Mixing has long been a critical issue in chemistry and chemical engineering. If the mixing is poor, the reaction process may be slowed down by local shortage of one of the reactants or thermal nonuniformities. Microreactors have the potential to be useful tools for chemical syntheses and kinetics studies. The reduction of reactor dimensions leads to a large surface to volume ratio of the reaction channel, which increases heat and mass transfer efficiencies. This feature allows microreactors to suppress “hot spots” and to be free from problems of mass transfer limitations. The reduction of reactor dimensions also leads to a laminar flow with small Reynolds number at each reactor channel. In other words, the mixing in microreactors is mainly driven by molecular diffusion without any assistance of turbulence. The diffusion process is almost accomplished between thin fluid layers that are formed by division of a mainstream into many small substreams or reduction of the channel width along the flow axis for one channel. Thus, large contact surfaces and small diffusion paths are generated in microchannels.

DESIGN OF MICROCHANNEL REACTORS: MICROMIXING

285

The degree of mixing of reactants greatly influences the product composition for multiple reactions and the kinetic measurements for a very fast reaction whose reaction time is shorter than the mixing time. To achieve better mixing, there are two basic principles to induce mixing at the microscale as defined by Hessel.6,7 In active mixing, the external energy sources, such as ultrasound, acoustic, bubbleinduced vibrations, piezoelectric vibrating membranes, magnetohydrodynamic action, and so on, are used to enhance the mixing efficiency. However, this type of mixing requires additional design and construction of force-driving devices or structures, which raise the complication and the cost of the whole mixing system. For passive mixing, flow energy, such as pumping or hydrostatic potential, is applied to reform the flow for faster and better mixing. Several types of passive mixing have been developed, such as chaotic mixing, microplume injection mixing, secondary flow mixing, distributive mixing, and split-and-recombination mixing, and so on. However, these techniques have their advantages and disadvantages. In general, the influences of mixing on microchannel reactors’ performance are investigated using computational fluid dynamics (CFD) simulations. Fluent is an example of such CFD codes and solves the conservation equations for mass, momentum, and energy using the control volume method. The reduction of the diffusion length is essential for fast mixing in microreactors, since mixing time is proportional to the square of diffusion length. In many micromixers, which are important parts of microreactors, the reactant flow is split into many laminar segments to shorten the mixing time. However, few studies have quantitatively addressed the relationship between the size of laminar segments in microreactors and the product composition of multiple reactions, and how the size of laminar segments affects the precision of the measurements of rate constants. These are very important factors for establishing a design method of microreactors for industrial applications. Thus, it is necessary to examine the beneficial effects of feeding reactants with a form of lamination segments for multiple reactions and show the lamination width in microreactors necessary to exhibit desired performance. 8.2.2 Omega Structure—A Novel Micromixing Technique In this section, we discuss briefly about several microreactor designs, more specifically, a novel micromixing design termed omega channel mixing. Its mixing performance is compared to that in straight and zigzag channels, the structures created for secondary flow mixing developed by our group previously. The dominant effect in micromixing is molecular diffusion, where the mixing length is proportional to the mixing time and diffusion coefficient. Especially, in the microscale reactor, the flow in microchannel is laminar; the mixing species is poorly diffusive. Therefore, it is important to create chaotic and turbulent flows in microchannels. A novel omega-shaped microchannel reactor has been designed to create profound chaotic flows in microchannels.8,9 A 3D fluidic dynamic simulation was used to evaluate the mixing performance of the omega-shaped microchannel, which is compared to common straight and zigzag microchannels. The 3D models of microreactors with different microchannel structures were simulated using the

286

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.1

2D top view of velocity contours in omega channel microreactors.9

MemCFD module of CoventorWareTM software (Coventor Inc.). The flow of a fluid through a fixed volume is governed by the Navier–Stokes equation. For fixed volume, the flow is subjected to conservation of momentum. The so-called omega channel microreactor consists of a network of omega-shaped channels and they are integrated as shown in Figure 8.1. The channels adopt the cellular structure of honeycombs. Figure 8.1 shows that the obstacles formed by the shape of the omega channel induce a high velocity value. They force the flow streamlines to comingle from center to the boundary channel wall and from boundary channel wall to the center periodically. The flow velocity in the omega channel varies significantly from point to point and the variance between the velocity values is large, which is bigger than that in the straight and zigzag channels. A gradual increase and decrease of path width in omega channel, the curve of channel, and the changes of flow direction help uneven the local flow velocity and generate vortices of the flow. Each omega channel has an impeding curve that impedes the oncoming flow and split it into two streams, and the two streams will unite with other flows from the adjacent omega channels. As a consequence, the chaotic motions within the omega channel are created without active moving parts or external source. The chaotic advection promotes the rapid changes (exponentially) of stretching and folding of the mixing species interfaces by either unsteady flow or complex structure of omega channel and rapidly increases the mixing efficiency. The transport of vortices can be calculated from the conservation equations of momentum. The incensement in vorticity is caused by promotion of stretching and turning of the vortex lines and incensement of diffusion is caused by viscosity and density inhomogeneity. Comparing the vorticity profile of the omega channel with that of zigzag channels in Figure 8.2 supportively shows that the omega channel generates higher vorticity than the zigzag channels. The vortices generated in omega channel tend to break up the streams into layers and each layer curls in a different manner. These breaking and curling actions and the comingling of center streamline and the edge streamline increase the chances of diffusion between the molecules of two liquids in a mixing process. However, straight channels do not have such advantages. In omega channel, the fluids flowing through channel are rotated as intended. There are critical differences between the low and high Reynolds number flows. As fluids pass through the omega channels with a periodically rotational shape, the fluids are

DESIGN OF MICROCHANNEL REACTORS: MICROMIXING

287

FIGURE 8.2 2D cross-sectional view of vorticities in (a) zigzag channel microreactor and (b) omega channel microreactor.9

stretched and folded due to the inertial force. This convection widened the interfacial area, inducing rapid mixing of fluids at a higher Reynolds number. As the Reynolds number becomes bigger (but still in laminar domain), fluids are mixed more and the interface of fluids becomes more distorted, recirculation occurs in the flow of omega channel, and the streamline turns unsymmetrical. The backmixing as shown in Figure 8.3 provides uniform distribution, and by increasing the residence time, a nondiffusion character can be introduced by the obstacles in the microreactor. The convection mixing action of the elements rapidly eliminates temperature gradients, reducing thermal degradation. To study the transitional properties of the mixing species, stochastic approach is employed for presenting the cumulative probabilities of one particle exiting the microreactor. The model presented is based on modeling the axial position of the particle by means of a Markov chain: the probability distribution of the axial position of the particle at a given time step n depends only on its position at time step n
1 and a set of transition probabilities. The transition probabilities are quantified in accordance with the particle transport processes.9 As expected, a molecule in the omega channel will take the longest time to exit the reactor among the three sets of reactors. The omega channel microreactor has a higher mean residence time than the other two microreactors and the variance of the residence time is the biggest due to turbulence.

FIGURE 8.3

Unsymmetrical, recirculation flow in omega channels at Re > 200.8

288

NANOCATALYSIS IN MICROREACTOR FOR FUELS

To evaluate simulation results described above, the microreactors with omegashaped, straight, and zigzag channels were fabricated and tested using Fischer–Tropsch synthesis (H2 þ CO) to higher alkanes, which is one of the major gas-to-liquid technologies to convert natural gas or other gaseous hydrocarbons into longer chain hydrocarbons for easy transportation and storage (discussed in Section 8.6). The experimental results of conversion rates as shown in Figure 8.4 are in agreement with the previous simulation predictions. The conversion rates in omega reactors are always higher than that of the straight and zigzag reactors under various conditions. The conversion efficiency for the omega-shaped reactor is 17% greater than that for the conventional straight channel microreactor and 12% greater than that for the zigzag-shaped channel microreactor. The longer residence time provides the molecules of different reaction species more mixing time and breaking of the C
O bond, thus promotes hydrocarbon chain (
CH2
)n growth, leads to more collisions between reactant molecules and the catalyst, and encourages more CO molecule consumption in the reaction. The advantage of omega reactor lies in the mixing performances, and consequently this mixing privilege contributes to higher conversion rate in omega reactor than that in the straight and zigzag channel reactors. In short, as described above, a network of omega-shaped microchannels has been designed and simulated by a Markov model for optimizing the channel geometry to improve mixing efficiency in microreactors. The simulation results are in agreement with the experimental data and show longer residence time and better mixing in omega channels than in common straight channels and zigzag channels of microreactors. In general, several mixing principles in microscale has been proposed. Most of them have been verified by making devices and testing them with simple flow visualization techniques and compared with predictions of CFD simulation. In future, a detailed benchmarking may be needed for proper mixing comparison between the different micromixers themselves, while dimensionless parameters are preferred. Moreover, the microdevices need to be benchmarked to conventional equipment.

FIGURE 8.4 CO conversion in straight, zigzag, and omega microreactors with different H2/CO gas ratios at H2/CO flow rate of 0.3/0.2, 0.2/0.1, and 0.3/0.1 sccm.8

FABRICATION OF MICROCHANNEL REACTOR

289

Finally, mixing efficiency may not be the only factor when designing or fabricating microreactors. The important factors, such as reactor functionality, reliability, and cost, should also be considered to establish commercial chemical production processes.

8.3 FABRICATION OF MICROCHANNEL REACTOR 8.3.1 Materials for Microreactor Construction The advent of microreactor technology was initially supported by silicon microfabrication techniques originally developed in the microelectronics and MEMS industry. As discussed in the previous section, computer simulation studies are performed prior to microfabrication. To date, a variety of materials and fabrication methods are available for construction of microreactors and microprocess components. In general, four types of materials are used as substrates for fabrication of microreactors: metal, glass, polymer, and silicon. Research at the Institut f€ ur Mikrotechnik Mainz (IMM) was initially focused on metal-based multichannel microreactors that could be produced using punching and conventional machining techniques.4 Small channel dimensions and the relatively high thermal conductivity of certain stainless steels result in extraordinary heat transfer characteristics. Generally, the choice of a particular metal or alloy depends on the application. Typical factors needed to be considered include resistance to corrosion, thermal properties and thermal response, and ability to handle mechanical stresses induced by the local process environment. Microfabrication techniques for metal microreactors include mechanical micromachining, laser micromachining, wet chemical etching, and selective laser melting. The combined effects of both material cost and ease of fabrication create a significant advantage for various polymers (plastics) over metals. Microfluidic plastic devices can be mass-produced using fabrication techniques such as hot embossing, injection molding, and casting. The biggest disadvantage of plastic devices is the maximum temperature that can be used without inducing material softening with subsequent creep or flow, which is typically less than 200 C.10 However, for applications where the maximum operating temperature is below the thermal stability limits of the particular plastic materials used in the device, such as those used in pharmaceuticals and bio-based material applications, plastics are an attractive alternative to metals. A special-purpose photostructured glass called FOTURAN, which is based on lithium aluminum silicate, is especially useful for creating microchannels and related structures with high aspect ratios, width of walls from 50 to 500 mm, length of walls up to 100 mm, and diameter of holes down to 50 mm. Sawing, grinding, polishing, and edge-forming machining are the techniques used to produce the basic glass plates.11 Photolithographic methods (resist coating, exposure, and development) and the subsequent chemical etching processes generate microstructures within the glass. The etched components can be upgraded by different coating processes (sputtering,

290

NANOCATALYSIS IN MICROREACTOR FOR FUELS

evaporation, screen printing, electroplating). Different bonding techniques (thermal bonding, gluing, soldering) add to its technical potential. The disadvantages of glass microreactors are that they are not suitable for high-temperature reactions (>400 C) and it is difficult to integrate them with heaters, sensors, and so on. Compared to glass, silicon has played a prominent role in microreactor fabrication due to the significant micromachining experience developed in both the integrated circuit (IC) and MEMS industries. Silicon is an excellent material because of its large operating temperature range and chemical inertness. It has been shown that the latter feature can be improved further by modifying the surface. The high thermal conductivity of single-crystal silicon (236 W m
1 K
1) versus aluminum (157 W m
1 K
1) also generates an advantage from a conduction heat transfer perspective.10 In addition, it is possible to integrate silicon-based microreactors with onboard sensors, such as those used for measurement of flow, temperature, and pressure. This creates an advantage of more precise control and performance information. This information comes at a significant cost since the machining and fabrication require specialized silicon processing equipment. Mass production of silicon microreactors is possible, but high volumes are needed to reduce the costs. Some research groups have focused on the use of silicon-based microreactors because it allows sensors, actuators, and microstructures to be directly integrated with microreactors.12 These features produce reactors with system characteristics that cannot be duplicated either on the macroscale or on the microscale using other materials, such as thermal time constants on the order of milliseconds, and the ability to control temperature and size within a length scale of a few micrometers. 8.3.2 Microfabrication of Silicon Microchannel Reactors The fabrication of silicon microreactors starts with cleaning of silicon surface by trichloroethane, acetone, and isopropyl alcohol, followed by photolithography, which is a well-developed microfabrication technique. The principle of photolithography is similar to exposure and development of photos taken by traditional cameras. The material to be exposed and developed to create intended patterns is called photoresist. There are two types of photoresists: positive and negative. Photoresist is spun on to a masking layer on a substrate, typically silicon oxide or nitride on a clean, polished silicon wafer. A pattern is formed by placing a chromium-patterned glass mask between a UV source and the silicon wafer. The resist is then developed to form a patterned protected layer. The unprotected areas of the material are then etched using either wet or dry etching when positive photoresist is used. There are two types of wet etchants for silicon. The isotropic etchants etch at equal rates in all directions, resulting in slightly rounded features. However, it is difficult to etch to high precision. Anisotropic etchants etch at different rates in different crystallographic planes.13 Although precise etching is more easily obtained in anisotropic etching where etch rates are low, it can result in rough surfaces. As an alternative etching method, dry etching comprises plasma- or dischargedbased techniques. They etch accurately at small dimensions resulting in less undercutting and broadening of features to give good pattern transfer. The various methods

FABRICATION OF MICROCHANNEL REACTOR

291

include chemical plasma etching, reactive ion etching, and ion beam etching.13,14 Recently, a deep reactive ion etching (DRIE) process was developed that utilizes a chlorine-based system in conjunction with passivation to achieve structures with large aspect ratios. These processes are very sensitive to various parameters. For example, temperature has a very strong effect and is one of the major causes of inconsistency. Small amounts of contaminants can drastically affect the product by reacting with the target or by changing the plasma chemistry. Combinations of dry and wet etching techniques can allow fabrication of complicated structures. One of the DRIE etching processes, inductively coupled plasma (ICP) etching, was utilized in our study for micromachining the reactor channels on the silicon wafer.14 The process underlying this technology is the so-called Bosch process, which is well developed and currently widely used on ICP equipment for DRIE technology in silicon micromachining. On the basis of Bosch process, Alcatel’s deep plasma etch technology was designed to deliver superior process performance and to meet the needs of a broad range of deep silicon etch applications. Its applications in MEMS and microfluidic device fabrication include high aspect ratios, etch depths of as great as 500 mm, etching through the wafer, etching into a buried cavity in the wafer, or etching onto buried oxide. The etch technology uses a patented high-density ICP source and a fluorine-based noncorrosive etch chemistry. The photolithography process can be replaced by LIGA, in which X-rays are used to perform lithography on thick resist.4,10,13 After development, the gaps between the remaining resist are filled by electroplating. The resist is then removed to leave a metal mold insert for injection or reaction molding. The use of highly coherent X-rays results in submicrometer resolution and large aspect ratios. Poly(methylmethacrylate) (PMMA) was the traditional resist of choice due to its excellent contrast and stability. However, due to its low sensitivity, a new negative resist based on a novolak resin was developed. A large range of metals, alloys, and dispersion composites have been used for electroplating, such as nickel, gold, copper, nickel, and cobalt. Suitable polymers for molding include PMMA, polyoxymethylene (POM), polyvinyldenefluoride (PVDF), polyaryl ether ether ketone (PEEK), and polycarbonate. The microfabricated reactors along with different microchannel structures such as straight, zigzag, and omega channels are shown in Figure 8.5. As mentioned above, ICP etching allows high aspect ratios of the microchannels. It has been widely embraced for MEMS processing due to its high etching selectivity, its ability to precisely transfer photoresist patterns into silicon substrates, and its cleanliness and compatibility with vacuum processing technologies. However, ICP etching has its own specific problems that include “grass” formation, etching uniformity, mask selectivity, and so on. Etching gas cycle time, bias power, and chamber pressure are the parameters adjusted to solve these unfavorable phenomena.14,15 The periodic change of different gases for etching (SF6) and passivation (C4F8) can lead to very high aspect ratios and very high etch rates. The surface morphologies of the wafer with respect to the increase of SF6 gas flow time change dramatically. The “grass” formation on the surface is the result of particulate material sticking inadvertently on the silicon surface. This material can locally mask the silicon during etching and can be formed due to redeposition and growth of polymer material from the sidewall passivation step.

292

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.5 Fabricated microreactors (a) with SEM pictures of (b) straight channels, (c) zigzag channels, and (d) omega channels.9

By increasing the cycle time of SF6, the “grass” formation was markedly decreased (Figure 8.6). As the SF6 gas flow time increased, the sidewall at the base of the reactor channel assumed a slight retrograde profile. This is caused by thin sidewall polymer layers at the bottom part of channel. When the cycle time of SF6 gas was increased, lateral ion bombardment to the sidewall also increased. The passivated polymer layer was thinner at the base of channels than at the top of the channels. Thus, the etching profile of trenches could be broadened at the bottom. It is useful to achieve high etch rates, but often at the expense of sidewall broadening problems. The etching characteristics at low chamber pressure (Table 8.1) show that by increasing the SF6 cycle time to 10 s the overall etching rate increased, but the etching uniformity dropped to 3.5%. The highest etching selectivity between the silicon and the photoresist was determined to be 167:1 and the aspect ratio was as high as 40:1. In the long SF6 gas cycle time region, the overall etching rate and etching uniformity are

FIGURE 8.6

SEM photos after etching with different SF6 gas flow times.15

293

FABRICATION OF MICROCHANNEL REACTOR

TABLE 8.1

ICP Etching Characteristics with Variable Parameters15

SF6:C4F6 (s)

Pressure (Pa)

Bias Power (W)

5 5 5 5 5 5 3

30 30 30 30 40 50 30

4:2 6:2 8:2 10:2 4:2 4:2 4:2

Etching rate (mm min
1) 4.13 5.39 5.90 6.40 3.91 4.86 3.16

Etching Uniformity (%) 11.6 3.1 2.4 3.5 8.7 8.3 6.8

Grass Formation Lots Little None None Lots Lots Lots

improved, and also the etching rate is nearly independent of the feature size. Increasing bias power causes higher average etching rate but does not contribute to improving etching uniformity to acceptable levels. Increasing bias power enhanced overall etching rate, but the etching uniformity was not improved. The surface morphology of the silicon substrate was strongly affected by etching gas species and the gas flow sequence time. However, the “grass” formation was not reduced at the low chamber pressure. The formation of micro “grass” on the silicon surface could result from several factors. Etching gas sequence time, bias power, and chamber pressure can all greatly change the surface morphology of silicon surface. Grass formation was dramatically reduced by increasing SF6 gas flow time. For the optimum conditions of the ICP etching to fabricate silicon microchannel reactors, longer cycle time of SF6 gas flow, high bias power, and low chamber pressure are all recommended. Most of the fabrication methods can produce structures but cannot form sealed structures that are imperative if the chemicals are to be contained in the channels. The use of gaskets and a suitable housing that is kept tightly clamped is a common sealing method, offering the advantage of easy assembly and disassembly. However, there are no suitable procedures for microscale assembling and handling. The most popular method for silicon microstructures is anodic bonding.14 Two wafers, one of which must be at least semiconducting, are heated to 180–500 C and a direct current (DC) voltage of 0.2–1 kV is applied across them. Silicon and Corning 7740 Pyrex glass are most frequently used. Wafers must be cleaned and polished before the anodic bonding procedure is started. The materials used must have similar thermal expansion coefficients. Alternatively, silicon fusion bonding and lamination method also allow the bonding of different materials. In general, the composition of a specific material of construction will partially dictate the preferred fabrication method. In many cases, a variety of materials may be equally applicable due to their similar chemical and thermal compatibility characteristics for the given process fluids as well as their ability to safely function over the desired operating ranges for both temperature and pressure within the given process safety settings.10 In these instances, the least expensive material with the most costeffective fabrication method is typically chosen. Ultimately, the cost and process application requirements will both dictate the final choice for materials of construction and fabrication methods. However, the enabling characteristics of miniaturization

294

NANOCATALYSIS IN MICROREACTOR FOR FUELS

must also be considered. If the selected material and fabrication technology cannot produce a microreactor that has step change advantages over a larger conventional reactor, the additional expense for miniaturization is not justified. Hence, when choosing a fabrication technology, the following factors should be considered by users: process cost, accuracy, process reliability, material choice, and process time. It is worthwhile after finishing a conceptual design of a specific microreaction system to compare the different technologies with respect to the mentioned parameters. Commonly, decisions are made on an overall evaluation of technological potential.

8.4 NANOCATALYST DEPOSITION ON THE MICROCHANNELS 8.4.1 Pretreatment of the Substrate and Coating Methods Structured catalysts and reactors are gaining more and more importance due to industrial applications. The microchannels in a microreactor have high surface to volume ratio and its geometric surface may be directly used for performing catalytic reactions, which is not the case in traditional reactors. However, to enhance activity and efficiency of a microreactor, it is necessary to increase the specific surface area by treatment of the microchannel walls, usually by applying porous coatings. The porous layer can be catalytically active or serve as a support for a catalytic phase. Micropacked beds of powder catalysts can sometimes be used, but, in general, a very thin layer of catalyst that sticks to the reactor wall is preferred because of mass and/or heat transfer improvement. Different methods can be used to deposit a catalyst layer on a surface, depending on the properties of the surface and the catalyst that has to be deposited. The pretreatment of the substrate to coat is important because it increases adhesion of the catalytic layer and thus the lifetime of the structured catalyst. Two common pretreatment methods are anodic oxidation and thermal treatment.16 The anodic oxidation method is generally applied to structures containing aluminum with the objective to obtain a porous alumina layer at the surface. The method is used either as a pretreatment before applying another coating or as a way to obtain a thin porous layer that can be directly impregnated. Like anodic oxidation, thermal oxidation is not really a deposition method but a surface modification technique. However, it can be used either as a pretreatment step to increase the catalyst adhesion or as a catalyst support. The chemical oxidation of the substrate is sometimes carried out by acid treatment to form metal oxide layers, such as Al2O3 layer, or to form a pseudolayer accessible to chemisorption of small charged particles. For silicon and titanium-based substrates, etching and/or oxidation of the surface can be obtained using an alkali treatment procedure. In general, two types of coating methods, chemical and physical, are used for nanocatalysts in the microchannels. The chemical methods include suspension, sol–gel, hybrid method between suspension and sol–gel, electrophoretic deposition (EPD), electrochemical deposition, electroless plating, and chemical vapor deposition (CVD).14,16 The physical methods consist of a mechanical method such as plasma vapor sputtering (PVD) and thermal methods such as evaporation and electron beam

NANOCATALYST DEPOSITION ON THE MICROCHANNELS

295

evaporation.14,16 These fabrication techniques are used in MEMS and semiconductor industry and often referred to as silicon coating procedures. In plasma vapor sputtering, capacitive plasma is generated between the surface to coat and a target made of the material to be deposited. Sputtering is performed under vacuum, the structured surface is operated as the anode, and the coating material is operated as the cathode that emits atoms to the surface.16,17 The catalytic metal (Pt, Fe, Co, etc.) is often sputtered without a prior oxide layer. In electron beam evaporation, the electron beam with high kinetic energy is directed to the target material for evaporation. Upon impact, the high kinetic energy is converted into thermal energy allowing the evaporation of the material. In pulsed laser deposition process, also known as pulsed laser ablation deposition, a laser is used to ablate particles from a target in a deposition chamber under reduced pressure and at elevated temperature. The number of laser pulses is related to the thickness of the film deposited on the substrate. For flame-assisted vapor deposition, the deposition process can take place in an open atmosphere without requiring the use of complex deposition chamber and/or vacuum system like that in CVD or PVD methods. The atomized chemical precursor of the catalyst is burned in a flame. The method can thus be considered as a “dry” way of deposition for the substrate that is placed in the combustion zone at controlled distances and temperatures. CVD technique requires the use of chemical precursors of the desired deposited material.17 The precursor can be the same as that used in sol–gel method (discussed below), but no solvent is required. Only the volatile precursor and the structured object are present in the deposition chamber. To enhance the deposition rate, the use of low pressures and high temperatures may be required. Plasma-assisted CVD (PACVD) also allows to perform the deposition at lower temperature and higher deposition rate. EPD is a colloidal process in which a DC electric field is applied across a stable suspension of charged particles attracting them to an oppositely charged electrode.16 One electrode (cathode) consists of the substrate to coat, the anode being either an aluminum foil or stainless steel. The thickness of the coating depends on the distance between the two electrodes (ca. 10 mm), the DC voltage (can vary from 10 to 300 V), the properties of the suspension (e.g., pH), and the duration. This technique is often used to deposit a layer of aluminum oxide (by oxidation of an aluminum layer) as a precoating, to favor the adhesion of a catalyst, deposited in the second time by dip coating in a suspension. Electrochemical deposition and electroless plating use ionic solutions.16 The first method, also called “electroplating,” produces a coating of metal on a surface by the action of electric current. The deposition of a metallic coating onto an object is achieved by putting a negative charge on the object to be coated (cathode) and immersing it into a solution that contains a salt of the metal to be deposited. When the positively charged metallic ions of the solution reach the negatively charged object, it provides electrons to reduce the positively charged ions to pure metal. All methods based on the dispersion of a finished material (catalyst support or catalyst itself) have been gathered under the term “suspension method.” In some preparations, the difference with sol–gel method is small because the suspension method often implies some gelification steps. It is the most widely used method,

296

NANOCATALYSIS IN MICROREACTOR FOR FUELS

namely, for ceramic monoliths. Powder (catalyst support or catalyst itself), binder, acid, and water (or another solvent) are the standard ingredients. The concentration of all ingredients varies largely from one experiment to another and also depends on the nature of the surface to coat and on the desired layer thickness. The size of the suspended particles has a great influence on the adhesion to the substrate. In sol–gel method, the starting point is a solution of a chemical precursor of the material to be deposited.14–16 It involves the transition of a system from the liquid sol to a solid gel and creates a three-dimensional network of inorganic matter known as “gel” from a colloidal or molecular solution of the precursor (sol) by low-temperature polymerization. It mainly involves the production of metal oxides by the hydrolysis of appropriate precursor compounds. The hydrolytic reaction is utilized to provide oxides in the form of thin films on the substrates or powders. Hydrolysis of the metal alkoxides to metal hydroxides followed by dehydration to metal oxides can be shown in the following equations: MðORÞn þ nH2 O ! MðOHÞn þ nROH; where M ¼ Si; Al; Ti; Zr M
OH þ HO
M0 !
M
O
H0 þ H2 O One important factor in sol–gel technology is the aging time allowing the gelation of the sol. It can vary from a few minutes to several weeks, depending on the concentrations in the sol and the characteristic size of the object to coat. The conditions during sol formation have to be chosen to obtain oligomers with desired degree of branching. Sols with high viscosities, obtained after long aging time, allow depositing thicker layer but easy to crack. A compromise has to be found for each preparation and substrate to coat. Sol–gel can also, in certain cases, be used to deposit a primer on the support to coat. In contrast, impregnation is often used (as a posttreatment) to deposit a catalytic active phase on the washcoat and does not differ from powder impregnation. Hybrid method does not differ very much from suspension method and sol–gel method. In the present case, a sol not only acts as the binder but also participates in the chemical and textural properties of the final deposited layer. In general, the suspension and the sol–gel methods are applied to the structured object using dip coating. An alternative to dip coating is spray coating. In the case of coating microreactor channels, the drops of the sol–gel can be deposited (drop coating) with possible simultaneous heating of the microreactor channels. Spin coating can also be used for a silicon microreactor. In this deposition method, the film thickness is related to the sol viscosity and the spin speed.14,18 When the different methods for thin films are compared, it appears that sol–gel technique produces layers less than 10 mm thick, whereas PVD methods yield layer thinner than 1 mm. The suspension methods can generate layers from 1 to 100 mm, but it is in general used to obtain thicker layers than that obtained by sol–gel. The sol–gel method allows porosity of the foam material, whereas the use of the suspension technology can result in blocking pores.16 To avoid the penetration of the oxide precursor in the pores, a hybrid method between suspension and sol–gel is preferred over sol–gel alone. The hybrid procedure indeed combines the advantages of the sol

NANOCATALYST DEPOSITION ON THE MICROCHANNELS

297

(precise control and tuning of the catalyst microstructure) and that of the suspension (ease of deposition). 8.4.2 Comparison of PVD with Sol–Gel Method Using Pt Catalyst We can compare the chemical method (e.g., sol–gel technology) with the physical method (e.g., PVD) to obtain metal (or metal oxide) catalysts on the surface of microchannels. Platinum, used as a catalyst in the hydrogenation and dehydrogenation of cyclohexene to cyclohexane and benzene (see below), has been deposited and studied in a silicon microreactor.14,18 Silica or alumina, prepared by a sol–gel process, is chosen as support because of its high specific surface area. We compared dip coating, drop coating, and spin coating by depositing SiO2 on the channels of the microreactor. The selective deposition of catalysts was applied to both PVD and sol–gel methods.14 The silica and alumina films were characterized by scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), and Brunauer–Emmett–Teller (BET) surface area measurement techniques. The microreactor was prepared by general lithography and ICP etching as discussed in previous section. The conversion of cyclohexene was used for testing activity and efficiency of the catalyst. Reaction results of supported Pt were compared to results obtained in unsupported Pt microreactors. The basic principle involved in the synthesis of the alumina coating solution is the hydrolysis of aluminum alkoxide to form its hydroxide, peptization, and polymerization to form the alumina. To allow the formation of a good three-dimensional network, the reflux step was performed after acid addition to the resultant solution and before coating on the channels of the microreactor. The silica sol–gel was prepared using tetraethoxysilane (Si(OC2H5)4) as the precursor. We chose SiO2 acid instead of SiO2 basic because the surface area is expected to be approximately 10 times higher.19 Dip coating, spin coating, and drop coating were used to apply SiO2 or Al2O3 on the channel region of the microreactor.14,18 Negative photoresist was chosen as the mask for the selective deposition (Figure 8.7) because the unexposed channel part of negative photoresist can be completely removed by developer, which minimizes residual photoresist. The photoresist was first spin coated on the silicon wafer, and then it was exposed in the UValigner. The development step opened the channel region of microreactor. After coating SiO2 or Al2O3 solution, a 200 C, 20 min baking step was required to dry the sol–gel coating for good adhesion. Pt was deposited on the catalyst support in the form of Pt þ 2 by chemical deposition, known as ion impregnation.16,20 Platinum(II)-2,4-pentanedionate (Pt(C5H702)2) was dissolved completely in toluene and the solution was dropped into the channel area on a hot plate at 70 C, with continued baking at the same temperature for 12 h. The impregnated Pt þ 2 on the catalyst support was reduced to Pt0 in the presence of gas mixture (40% H2 in N2, 0.2 L min
1) at 400 C for 4.5 h. The chemical state and the atomic concentration of the elements present in the coating of the catalyst support were studied using XPS. The ratio of Si:O in the silica film prepared by the sol–gel method is 1:2, while the ratio of Al:O in Al2O3 is 1:1.242, which is different from the stoichiometric ratio of 1:1.5. XPS analysis shows that the

298

NANOCATALYSIS IN MICROREACTOR FOR FUELS

Step 1: Before PR coating

Step 2: After PR coating

Step 3: After exposing photoresist

Step 4: After SiO 2 coating

Step 5: Strip PR with acetone and coat Platinum

Step 6: Bond reactor with Pyrex

FIGURE 8.7 The process for selective deposition of catalyst in the microchannels.18

peak for Pt þ 2 after reduction (Pt0) shifts from 73.125 to 70.9 eV (Figure 8.8a). The distance between the separation of the two peaks (70.9 and 74.2 eV) after reduction is 3.3 eV, which is in agreement with that of the standard peaks. The XPS depth profile of Pt shows the uniform concentration and equal distribution of Pt inside silica film. SEM analysis was performed on the channels that were coated with silica or alumina to image the catalyst support and also to verify thickness of the film. The SEM images in Figure 8.9 confirm that the adhesion of alumina to silicon sidewall is worse than that of silica. Some parts of the alumina film peeled off sidewall to yield bad coating. Figure 8.10a shows that the particles of silicon film in the microchannels are around 100 nm in diameter, which is confirmed by AFM image shown in Figure 8.10b. These nanoparticles create the pore structure of silica with the specific surface area

FIGURE 8.8 (a) Binding energy of platinum before and after reduction and (b) XPS depth profile of Pt deposited by sputtering.14

NANOCATALYST DEPOSITION ON THE MICROCHANNELS

FIGURE 8.9

299

(a) Silica on 100 mm channels and (b) alumina on 100 mm channels.14

(SSA) of 500 m2 g
1. The SSA study at different aging times and calcination temperatures show that longer aging time and higher calcination temperature decrease the specific surface area. SiO2, which has different aging time, was coated on 5 and 100 mm channel micreactors by dip-coating, drop-coating, and spin-coating techniques. The reactors with dip coating have larger surface area than those obtained with drop-coating and spin-coating methods. Under similar conditions, the reactor with 5 mm channels has higher surface area compared to that with 100 mm channels because there are more channels in the former. After the characterization was completed, the prototype reaction on conversion of cyclohexene was studied to test the activity and efficiency of supported catalyst. C6 H12 $ C6 H10 þ H2 $ C6 H6 þ 3H2 From the reaction results, we conclude: 1. For 100 mm channel reactor, the conversion increases with increasing residence time. Effect of residence time on conversion is much smaller on 5 mm channel than that on 100 mm, as shown in Figure 8.11a. 2. For 100 mm channels, the conversion on silica-supported Pt is much higher than that on sputtered Pt (Figure 8.11b) due to the high surface area of silica support with porous structure shown in Figure 8.10a.

FIGURE 8.10

(a) SEM picture of silica particles and (b) AFM image of silica particles.14

300

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.11 Reaction results on conversion of cyclohexene: (a) conversion of cyclohexene on sputtered Pt in 5 and 100 mm channels; (b) conversion of cyclohexene on sputtered Pt and silica-supported Pt in 100 mm at the same design point of the reaction.18

3. The conversions with silica-supported Pt and 100 mm channel are very close to the conversion with sputtered Pt in 5 mm channel when Figure 8.11a and b are compared. Thus, we may choose high reactant flow rate in 100 mm channel with sol–gel-supported Pt to receive more products at shorter time, which creates high efficiency and high productivity. 4. Compared to silica-supported Pt, alumina-supported Pt gives higher conversion of cyclohexene, but it deactivates faster. Moreover, the conversion on silicasupported Pt increases when temperature rises, but keeps constant on aluminasupported Pt and drops fast when the temperature is kept below 120 C. We discussed previously two kinds of catalysts: supported (sol–gel method) and unsupported (sputtering method). Sputter deposition is popular for coating thin, uniform layers on the channels of microreactors. We can get different compositions of catalyst layers by this method. But the disadvantage is that there is no surface area enhancement. Supported catalysts can have high surface area, which provides more reaction surface for high throughput. The sol–gel deposition process is cost-effective since no expensive equipment is required. Due to its easy formulation and coating processes, the high throughput of catalysts is the other benefit compared to sputter deposition. In general, some methods concern only the oxide deposition (which can further be impregnated by a catalyst precursor) and others concern the direct deposition of a metal on substrate, without any oxide layer. The ease and the speed of the PVD process are very advantageous in the case of parallel screening because it allows one to obtain an important catalyst library in a few hours and thus provides rapid information on the active metals to catalyze a reaction. However, due to their low porosity, the activity of the obtained catalysts by PVD methods leads to low activity catalysts compared to catalysts prepared by wet chemical procedures, such as sol–gel method. The sol–gel method presents some advantages: it can be automated and it can also be applied to closed microreactors. However, liquid-phase handling to coat a microreactor is not preferable due to nonuniform removal of solvent. In other words, to overcome the problems of low

HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR

301

activity of catalysts from CVD and PVD methods, the use of flame spray synthesis is recommended since it yields porous catalysts without handling a liquid precursor. The main data based on current research concern metal-on-oxide catalysts for which many methods exist. Some consider physical treatment of the surface to coat (anodization, plating, PVD, etc.), others involve a more or less complex chemical preparation (suspension and sol–gel). The properties of the deposited layer vary to a large extent, for example, the thickness, from nanometer (PVD) to near millimeter scale (suspension). The textural properties of the oxide supports can in certain cases reach that of traditional catalysts (suspension, sol–gel, powder plasma spraying methods). “Physical” methods in general lead to more adhesive layers, but to less active catalysts. The most widely used method involves direct synthesis on the surface. The nanostructure and microstructure and the size of the supports affect conversion efficiency, space availability, catalyst loading, and so on. We can simulate commercial catalyst practice by sol–gel and impregnation methods to realize optimized coating situations for future industrial implementation. 8.5 HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR 8.5.1 Hydrogen Production from SRM: Current Status The demand for power sources with superior performance has increased as a result of the rapid growth of the portable electronics market. Greater performance leads to greater consumption of electrical power, and it has become difficult to secure a sufficiently long-lasting power source for these portable electronic devices, even using lithium-ion batteries. The proton exchange membrane fuel cell (PEMFC), using hydrogen as a energy source, has attracted much attention for portable electronic devices since it has higher high volumetric energy density (for example, 2500 Wh L
1 for liquid hydrogen) than lithium-ion battery (400 Wh L
1).21 Thus, hydrogen production is getting a lot of attention from today’s researchers, and also due to consumption of gasoline and environmental concern.22–26 Steam reforming is an alternative process to produce hydrogen from organic sources with the aid of a catalyst. Of many candidates being considered for hydrogen fuel sources, methanol, ethanol/bioethanol, gasoline, and diesel are readily available and currently being investigated. The use of methanol for steam reforming is attractive due to its high energy density, low cost, easy transportation, and low reforming temperature. The main reactions involved in steam reforming of methanol may be presented by the following equations. 1. Steam reforming of methanol: CH3 OH þ H2 O Ð 3H2 þ CO2 ;

DHr ¼ 49:5 kJ mol
1

2. Methanol decomposition: CH3 OH Ð 2H2 þ CO;

DHr ¼ 90:6 kJ mol
1

302

NANOCATALYSIS IN MICROREACTOR FOR FUELS

3. Water–gas shift reaction: CO þ H2 O Ð H2 þ CO2 ;

DHr ¼
41:2 kJ mol
1

Considerable work already exists in the literature on SRM for hydrogen production using conventional macroscale reactors. It has disadvantages in terms of size and weight that affect the compactness of fuel processor. The catalyst in packed bed form also exhibits high-pressure drop and possible channeling of gases. The low gas velocity is maintained in the reformer to achieve high conversion, but it lowers the effective thermal conductivity of the catalyst bed. For an endothermic reaction such as methanol steam reforming, temperature gradient in the packed bed leads to lowered catalyst activity and falsified kinetics. In addition, hot and cold spots are commonly encountered in the catalyst bed that results in overall poor performance. The use of microreactors for steam reforming of methanol can compete with the conventional reactors due to the advantages of microreactor systems, including lightweight, compactness, rapid heat and mass transport due to large surface to volume ratio, and precise control of process conditions with higher product yields.24–26 The high heat exchange efficiency in microreactors allows one to carry out reactions under isothermal conditions. Also, microchannel reactors working under laminar flow conditions show low pressure drop compared to randomly packed bed reactors. The short radial diffusion time in microreactors leads to narrow residence time distribution of reaction gases, which allows an optimum contact time between reactants and catalysts avoiding the formation of unwanted by-products. The catalysts for SRM can be divided into two types: non-noble metal (typically nickel) and noble metals from Group VIII elements (typically platinum). Due to severe mass and heat transfer limitations, conventional steam reformers are limited to an efficiency factor for the catalyst, which is typically less than 5%. Therefore, kinetics and thus the activity of the catalyst are rarely the limiting factors with conventional steam reformers. The mass and heat transfer limitations have been shown to be overcome by employing microchannel-based reactors, enabling intrinsic kinetics of steam reforming to be exploited. In these systems, the noble Group VIII metals are preferred since they exhibit much higher specific activities than nickel catalysts. However, the high cost of Group VIII metals is driving some researchers to develop alternative catalysts. The catalyst performance is greatly influenced by the type of supports. A number of commercial and research derived noble metals (such as Pt and Rh) loaded onto metal oxide supports (such as CeO2, ZnO, MgO, Al2O3, SiO2, ZrO2, and TiO2) including more than one type of catalyst support have been tested for hydrogen production from methanol and ethanol. Although a number of catalysts have shown promising results, problems with deactivation of the catalysts with an ensuing decrease in hydrogen and carbon dioxide and an increase in carbon monoxide production have been reported. Currently, there is little understanding of the behavior of these catalysts in steam reforming reactions carried out in microreactors. Thus, more basic research is necessary to find the optimized combination of catalyst and support for hydrogen production.

HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR

303

8.5.2 Non-Noble Nanocatalysts for SRM Reactions to Produce Hydrogen This section focuses on the development of Ni and Co non-noble nanocatalysts on silica support in microchannel reactors for SRM reactions to produce hydrogen. The SRM microreactor is a silicon-based microdevice with the dimension of 1.6 cm  3.1 cm. It consists of vias, feed inlet, product outlet, and reaction zone with a volume of 9.6 mm3 and 120 straight channels of 50 mm width and 100 mm depth.27,28 The microreactors coated with silica-supported Co or Ni nanocatalysts were prepared by a sol–gel procedure as described in Section 8.4. The precursors for Co and Ni catalysts are cobalt nitrate and nickel nitrate. The catalyst-coated microreactor was dried by gentle heating and treated with 10% NH4OH solution for 30 min to form hydroxides of metal catalysts, followed by washing with DI water to remove residuals from ammonia treatment. Calcination completed the formation of oxides from hydroxides. The oxides of cobalt and nickel were finally reduced with hydrogen to active metals before packaging of the microreactor. The calcination temperature for SiO2 supported catalyst was obtained by differential thermal analysis (DTA). For Ni/SiO2 (Figure 8.12), endothermic peaks are observed at 100 C, which can be attributed to evaporation of water. Due to ammonia treatment during synthesis of the sol–gel encapsulated catalysts, most of metal nitrate salts are converted to metal hydroxides that are further converted to metal oxides by heating in air. The broad endothermic peaks observed at 100–150 C can be attributed to water loss and some of the metal nitrate salts getting decomposed. One exothermic peak may be either due to the metal hydroxides getting converted to the oxides or due to structural changes of the surface species. Thus, we may conclude that all of the metal hydroxides are converted to the metal oxides in temperature range of 350–400 C, which can be considered the minimum temperature for calcination. To optimize the calcination temperature for Ni/SiO2, X-ray Diffraction (XRD) analysis was performed by annealing the samples from 450 to 1000  C.27 The results show that the sharpness and intensity of the XRD peaks increased due to formation of larger crystallites during the increase of the annealing temperature. This represents evolution of the particle size and corresponds well to crystal growth. The higher

FIGURE 8.12

Differential thermal analysis of silica sol–gel-supported Ni catalyst.27

304

NANOCATALYSIS IN MICROREACTOR FOR FUELS

Magnetization (emu g–1)

0.2 0.15 0.1

As-made Reduced After reaction

0.05 0

–0.05 –0.1 –0.15 –0.2 –10

–5

0

5

10

Applied Magnetic filed (kG)

FIGURE 8.13 Room-temperature magnetization curves of Co/SiO2 before reduction, after reduction, and after SRM reaction using a vibrating sample magnetometer.27

crystalline characteristics of the samples calcined at 1000  C indicates that heating the catalysts at very high temperatures may result in large crystallite sizes due to aggregation. It has been reported that NiO crystallite in small size favors large Ni surface area after reduction, which implies that higher calcination temperature is not preferred for nanocatalysts. Similar observations of DTA and XRD were made for Co/SiO2. Thus, we selected 450  C as the optimized calcination temperature for both Ni and Co catalysts. Since pure metallic Co and Ni are ferromagnetic, one can study the magnetic properties of Co and Ni catalysts to understand the reduction efficiency during hydrogenation and formation of chemical compounds of metal catalysts during catalytic reactions. The saturation magnetization of the ferromagnetic component in magnetic curves obtained from VSM was used along with the energy dispersive Xray (EDX) results to estimate the pure metallic Co and Ni in the catalysts. Magnetization studies of the silica-supported Co catalyst before reduction with hydrogen show paramagnetic behavior as cobalt is in its oxide forms (Figure 8.13), which is also confirmed from the XRD and DTA results. Hydrogenation of the catalyst reduces most, if not all, of the Co oxide to pure metal (the active phase for SRM reaction), thus giving the catalyst the ferromagnetic behavior. The ferromagnetic nature almost disappears in the postreaction catalyst sample as most of the metallic Co yields nonferromagnetic species. Hence, the magnetization results were used for ferromagnetic catalysts to estimate the pure metal percentage from the saturation magnetization value of the ferromagnetic component obtained at different stages. The elemental analysis using EDX indicates both Ni and Co catalysts in the microchannels show loadings of 5–6%. The EDX analysis at different locations of the sample shows uniform distribution of the catalyst in sol–gel matrix. From transmission electron microscopy (TEM) analysis shown in Figure 8.14, we can estimate the size of the Co particles in silica sol–gel to be <10 nm. The BET analyses show that

HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR

305

FIGURE 8.14 TEM image of Co nanoparticles in silica synthesized by sol–gel method.27

specific surface area is 450 m2 g
1 for Ni/SiO2 and 340 m2 g
1 for Co/SiO2. The Barrett–Joyner–Halenda (BJH) analyses show that the pore size of silica support is  32 A. With SiO2 porous structure, Ni and Co can be effectively distributed to obtain larger surface area available for catalytic SRM reactions. SRM reactions were conducted over Co or Ni nanocatalyst supported by silica sol–gel in the temperature range of 180–240 C under atmospheric pressure. The flow rates of 1:1 methanol/water between 5 and 20 mL min
1 were controlled using a syringe pump (Figure 8.15). A cold trap was used to separate gaseous products from aqueous methanol and water. Methanol conversion was calculated from volume difference between the fed methanol–water mixture and the unreacted methanol–water mixture cooled with liquid N2 in the trap. The gaseous products containing H2, CO2, and CO were diluted with helium and analyzed using a mass spectrometer (MS) coupled with a residual gas analyzer (RGA) (QMS 200 gas analyzer from Stanford Research Systems). Hydrogen selectivity was calculated on the basis of partial pressures (proportional to moles) of different products.

FIGURE 8.15 Experimental setup for SRM reactions with Ni/SiO2 or Co/SiO2 catalyst.28

306

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.16 Methanol conversion (CH3OH:H2O ratio of 1:1) as a function of flow rate at 200 C using silica sol–gel-supported Ni/SiO2 and Co/SiO2 nanocatalysts in 50 mm channel microreactor.27

The main products of SRM reaction are hydrogen and carbon dioxide with a small amount of carbon monoxide. The methanol conversion decreases as the flow rate is increased (Figure 8.16) for both Co/SiO2 and Ni/SiO2 nanocatalysts with slightly higher conversion for Ni/SiO2. While our studies with Ni/SiO2 show 53% methanol conversion, Co/SiO2 shows only 37% conversion at 5 mL min
1 flow rate and 200 C. This may be attributed to larger specific surface area of Ni/SiO2 catalyst. However, the catalysts’ behavior in microreactors for SRM is not fully understood. The decrease in methanol conversion with increasing flow rate may be explained by lower residence time of the reactants in the microreactor at higher flow rates. The hydrogen selectivity (Figure 8.17) also decreases with increasing flow rate. The maximum hydrogen

FIGURE 8.17 Hydrogen selectivity (CH3OH:H2O ratio of 1:1) as a function of flow rate at 200 C using silica sol–gel-supported Ni/SiO2 and Co/SiO2 nanocatalysts in 50 mm channel microreactor.27

HYDROGEN PRODUCTION AND PURIFICATION IN A MICROREACTOR

307

selectivity is 74% for Ni/SiO2 catalyst and 67% for Co/SiO2 at 5 mL min
1 flow rate and 200 C. The temperature in the range of 180–240 C does not have a significant effect on methanol conversion for both catalysts, but the hydrogen selectivity is maximum at 200 C for Ni/SiO2 catalyst and at 220 C for Co/SiO2 catalyst. All the reactions described above were carried out within 8 h and no significant deactivation of the catalysts was observed during this period. However, when these reactions were carried out over 10 h, deactivation of the catalysts was noticed, which is consistent with the VSM analysis. The VSM results from the postreaction catalyst sample (Figure 8.4) provide an estimate of 90% Co and 85% Ni being converted to nonferromagnetic species after SRM reactions over 10 h. These species may be Ni or Co compounds such as their oxides, carbonyls, and carbides. 8.5.3 Platinum Catalyst for Preferential Oxidation of CO in Hydrogen As mentioned previously, there is a little amount of carbon monoxide (CO) produced during steam reforming of methanol to produce H2. In general, CO poisons the electrode catalyst of fuel cell. Thus, it is necessary to develop robust catalysts for preferential oxidation of CO in an atmosphere of H2.29–31 The reactions for preferential oxidation of CO are shown below: Pt

Desired reaction: CO þ O2
! CO2 Pt

Undesired reaction: H2 þ O2
! H2 O Platinum was chosen as the catalyst, and a sol–gel support was utilized to maximize the surface area. High conversion of CO is required to reduce its concentration to a level that is not detrimental to a proton exchange membrane (PEM)-based fuel cell. High selectivity to CO2 is desired because hydrogen is used to generate electricity in the fuel cell. The side reaction on oxidation of hydrogen to water reduces the energy available for the fuel cell. The microreactor used for CO oxidation was 3.1 cm long and 1.6 cm wide with 119 microchannels of 25 mm in width and 100 mm in depth. The reactant gases were premixed before being allowed to flow in the channels of the microreactor. An inlet manifold was designed to promote a uniform distribution of flow among the reaction microchannels. The inlet manifold consisted of two channels symmetrically off-axis from the single outlet.32 The effect of the ratio of O2/CO on CO oxidation was studied by keeping the mixed gas at 1 sccm and changing the airflow rate from 0.02 to 0.4 sccm at a constant temperature of 200 C. As expected, the selectivity of oxidation of CO decreases as the O2 to CO ratio is increased above 2. Surprisingly, the conversion of CO did not appear to significantly increase as the ratio of O2 to CO was increased above 2. The effect of temperature showed that conversion and selectivity were both maximized at the minimum temperature tested, 120 C. The lower stoichiometric excess of O2 displayed a higher sensitivity of conversion with respect to increase in temperature. The sensitivity of conversion and selectivity on reactor residence time was tested to

308 Conversion or selectivity

NANOCATALYSIS IN MICROREACTOR FOR FUELS 95.00% 90.00% 85.00% 80.00% 75.00% 70.00% 65.00% 60.00%

Conversion of CO

55.00%

Selectivity to CO2

50.00% 100

120

140

160

180

200

220

Temperature

FIGURE 8.18 Temperature effect on the conversion of CO and selectivity to CO2 at a low total flow rate of 0.2 sccm and O2/CO ratio of 0.5.32

show that this particular catalyst and reactor combination favored the lowest space velocity that could be achieved, 13 h
1. Additional experiments were then conducted at WHSVof 13 h
1 to finalize the optimal temperature of operation. The ratio of O2/CO was held constant at a stoichiometric ratio of 0.5. Figure 8.18 shows that 160 C is optimal temperature of operation at these conditions. The catalyst deactivation study showed that after 50 h of reaction both conversion of CO and selectivity to CO2 dropped significantly. We speculate that this could be due to two likely mechanisms: (1) oxidation of the platinum and (2) sintering of the platinum nanoclusters into larger particles. While limited recovery of the catalyst activity through repeated reduction treatments suggests the former is more likely, the irreversible reduction in exposed surface area through sintering of the platinum nanoclusters at the temperature of operation is the likely cause for the remaining drop in activity. Carbon deposition is not thought to be a major contributor due to the lack of a possible mechanism as predicted by Chemkin simulations.33 Improvement of the catalyst activity and efficiency will be a significant issue for future research and commercialization of this technology in fuel cell applications. The research described above for the development of catalysts for H2 production and elimination of CO is very important for the development of microfuel cells. Microfuel cells are high-energy-density sources for next-generation power portable products such as personal digital assistant (PDA), laptop, cellular phone, and so on. Fuel cells can provide more reliable, longer portable power than batteries. Microfuel cell products compete with traditional power systems that utilize both direct and indirect energy conversion methods. Direct methanol systems have the advantage of room-temperature operation but offer only relatively low power density due to methanol crossover through the proton exchange membrane and the low reaction rate of methanol oxidation over the anode electrocatalyst. In contrast, reforming systems generate electrical energy in the fuel cell from concentrated hydrogen produced by steam reforming, for example, from methanol. Reforming systems achieve high power density but are difficult to miniaturize due to the complexity

MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY

309

of the required structure that includes not only a fuel reformer but also a vaporizer, a CO removal unit, and various peripheral parts. However, the microchannel fabrication on a silicon substrate allows the various components of microfuel cell, such as steam reformer, hydrogen purifier, driving units, sensors, and microvalves, to be integrated by semiconductor technology and MEMS technology. It may also be possible, through the use of these technologies, to achieve sufficient thermal insulation between the reformer (with microreactor) and the periphery of the system, where the surface temperature should be reasonable to touch. It is therefore expected that the small SRM systems with high power density, sufficient to power portable electronic devices, will be realized in the near future. 8.6 MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY Catalyst development for GTL technology using FT synthesis is a very active area of research since it converts unconsumed natural gases into hydrocarbons that are subsequently converted to fuels by hydrocracking for low-cost production and easy transportation. Although FT synthesis has been developed by several worldwide petroleum corporations such as Exxon Mobil, Shell, Sasol, Syntroleum, and Rentech to help deliver the potential of the world’s untapped natural gas resources, a lot more work is necessary to develop robust and stable catalysts in this regard. The catalysts for FT synthesis generally contain metals such as cobalt, iron, nickel, and copper with support materials of oxides and zeolites. More specifically, the influence of various support materials such as titania, alumina, silica, and ceria on the activity of cobalt catalysts for CO conversion have been studied.34–36 Alumina is the primary choice as a support for cobalt catalysts due to its thermal stability, although reduction of cobalt oxide to cobalt is limited due to strong interaction between the support and cobalt oxides. Addition of a second catalyst like iron may improve the selectivity to heavier alkanes with less selectivity to methane.37 Recent studies have shown that the catalyst activity depends on the number of active sites located on the surface of the support formed by the reduction process, determined by the particle size, loading, and the degree to which the metal has been reduced. 8.6.1 Iron–Cobalt Mixed Catalysts for Fischer–Tropsch Synthesis As discussed in earlier sections, microreactors are ideal systems for catalyst development as they provide unique advantages such as low consumption of reactants, greater speeds in catalyst characterization, and easy integration with other devices. More significantly, high surface area of microchannels to volume of the reactor inhibits gasphase free radical reactions and improves heat transfer for exothermic reactions, such as FT synthesis to higher alkanes as shown below. Silicon microreactors with either 5 or 25 mm wide channels were coated with mixed metal Fe–Co in alumina sol–gel for conversion of syngas (CO þ H2) to higher alkanes. The volume of the microreactors is 9 mm3 for both 5 mm (1200 in number) and 25 mm (240 in number) wide channel reactors.38–40

310

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.19 Comparative studies on conversion of CO to alkanes in 5 and 25 mm channel reactors at different temperatures. Fixed H2/CO ratio of 3:1; 1 atm.38

200
300 C

CO þ H2
! CH4 þ C2 H6 þ C3 H8 þ higher alkanes Fe=Co on Al2 O3 support

Initial studies were performed in 5 mm channel reactors to study the conversion rates at different ratios of H2:CO. The conversion of CO reaches maximum (32%) at the H2:CO ratio of 3 with a total flow rate of 0.6 sccm and 230 C in 5 mm channel reactors. The low CO conversion is likely due to insufficient coating of sol–gel in 5 mm channels of the reactor. This is indeed supported by higher conversion in 25 mm channel reactors (Figure 8.19) where sol–gel can penetrate easily (shown in Figure 8.20b) allowing the reactants to have more interactions with the catalyst. This may also explain the observed lower loading (7%) than the intended loading (12%) of supported catalysts from our EDX studies. We believe that the sol–gel granules (less than 100 nm indicated by atomic force microscopy) and high viscous nature of alumina sol–gel cause blocking of 5 mm channels as shown in Figure 8.20a.

FIGURE 8.20 SEM image of alumina sol–gel encapsulated Fe/Co catalyst deposited on (a) 5 mm wide channel microreactor and (b) 25 mm channel microreactor with 100 mm depth.38

MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY

311

FIGURE 8.21 Selectivity of alkanes: methane, ethane, and propane in 5 and 25 mm wide channel reactors at 230 C; H2/CO ratio 3:1; 1 atm.38

However, the blockages of catalyst active sites in the microchannels may also occur due to carbon formation from the water–gas shift reaction and can also decrease CO conversion at higher temperature. The study of temperature effect on CO conversion shows initial increase with rising temperature and decrease after 230 C due to exothermicity of the reaction as the competition between activation and exothermicity progresses (Figure 8.19). The effect of flow rate of syngas was also investigated. The reaction results show that the conversion of CO decreases with increase in total flow rate as a result of decrease in residence time of the reaction. Although the conversion of CO is influenced by the width of the channels, the selectivity toward different alkanes is not. Figure 8.21 shows the selectivity to methane, ethane, and propane as a function of temperature in 5 and 25 mm channel reactors. Propane has been the major product of the reaction in both cases with selectivity up to 78%. The residence time does not have any significant effect on selectivity to alkanes in the microreactor. The use of Fe–Co catalyst on alumina support for FT synthesis to higher alkanes is based on two experimental observations reported in the literature: while cobalt catalyst inactivates water–gas shift reaction to some degree, iron has an advantage of improving selectivity to higher alkanes with less selectivity to methane. Due to its thermal stability, alumina with porous structure as a support for dispersing the metal catalyst to a large total surface area with high catalytic conversion has been investigated by Iglesia.34 However, reduction of cobalt oxide to cobalt is limited due to strong interaction between alumina and cobalt oxide35 and formation of Co aluminates.34 Furthermore, higher viscous nature of alumina sol makes the deposition

312

NANOCATALYSIS IN MICROREACTOR FOR FUELS

FIGURE 8.22 FE-STEM image of silica-supported Co synthesized by sol–gel method in 25 mm microchannel showing the particle size of 6 nm.41

process in Si microreactor difficult and results in nonuniform coating. Thus, alumina has been replaced with silica as the catalyst support to improve catalyst coating performance and to increase syngas conversion to alkanes due to good adhesion between silica and Si microchannels (Figure 8.9). The particle size of silica in 25 mm channels is <100 nm. The field emission scanning transmission electron microscopy (FE-STEM) analyses indicate that the particle sizes of the nanocatalysts in the silica support is 6 nm (Figure 8.22). The CO conversions increase from 52% with alumina (Figure 8.19) to 63% (Figure 8.23) with silica in 25 mm channel microreactors.

80

% CO conversion

75 70 65 60 55 50 H2/CO 3:1, without Ru H2/CO 2:1, with Ru H2/CO 2:1, without Ru H2/CO 3:1, with Ru

45 40 35 30 190

200

210

220

230

240

250

Temperature (ºC) FIGURE 8.23 Comparative studies on CO conversion to alkanes at different temperatures and two H2:CO ratios, using silica-supported Fe–Co catalysts with and without Ru at 1 atm and a total flow rate of 0.4 sccm.41

313

MICROREACTOR FOR GAS-TO-LIQUID TECHNOLOGY

8.6.2 Ruthenium Added as a Promoter to Iron–Cobalt Catalyst for CO Conversion Ruthenium as a promoter was added to the mixed Fe–Co catalyst to improve CO conversion since it is one of the most active catalysts and can work at lower temperatures.36,37,41 The highest conversion of CO on Ru–Fe–Co/SiO2 catalyst reaches 78% at 220 C by adding only 0.4 wt.% Ru at H2:CO ratio of 3:1. This is much higher than that observed for Fe–Co/SiO2 catalyst (63%). The effect of H2:CO ratio on CO conversion is less with Ru than that observed without Ru (Figure 8.24). With Ru, CO conversion is almost same for both 3:1 and 2:1 ratios of H2:CO at 200 C, whereas the H2:CO ratio effect is significant without Ru (40% at 2:1 to 49% at 3:1). This indicates that the role of Ru is more important in CO conversion than that of H2: CO ratio. Ru can improve CO conversion at lower temperature more efficiently than at higher temperature for both H2:CO ratios of 2:1 and 3:1 (Figure 8.23). Addition of Ru not only increases CO conversion but also affects the selectivity to alkanes. While Ru–Fe–Co/SiO2 catalyst at H2:CO ratio of 3:1 yields 70% selectivity to propane, 20% to ethane, and 10% to methane (Figure 8.24), these values with Fe–Co/SiO2 catalyst are 80%, 17%, and 3%, respectively. The lower selectivity to higher alkanes for Ru–Fe–Co/SiO2 may be due to possible side reactions to carbon formation with all the FT catalyst metals (Fe, Co, Ru). For FT synthesis, at 200 C or below, the consecutive reaction of carbon hydrogenation to the CH2 monomer and its consumption for chain growth inhibits CH4 formation and a carbon phase on the catalyst is not allowed to grow.42 The selectivity changes mainly to methane at elevated temperature, with possible deposition of carbon and catalyst deactivation (especially with iron) reducing average chain length of the product molecules.41 The results from our studies are consistent with the steady-state isotopic transient kinetic analysis (SSITKA), which indicates higher amount of CH4 in the presence of 90 80

Selectivity (%)

70 60 50 40

Methane with Ru Propane with Ru Ethane without Ru

Ethane with Ru Methane without Ru Propane without Ru

30 20 10 0 1.5

2

2.5

3

3.5

H2:CO ratio FIGURE 8.24 Selectivity to various alkanes at different H2:CO ratios using silica-supported Fe–Co catalysts with and without Ru in 25 mm wide channel reactors at 230 C, 1 atm, and a total flow rate of 0.4 sccm.41

314

NANOCATALYSIS IN MICROREACTOR FOR FUELS

Ru.41 Based on the increase in H2 chemisorption, Ru promotion increases the reaction by increasing the Co metal dispersion, resulting in an increase in the number of reaction sites. Similarly, Ru promotion of silica and titania-supported Co catalysts exhibit a synergistic effect on CO hydrogenation, and higher hydrogenation ability of Ru is due to a cleansing effect during CO hydrogenation that prevents the formation of carbon deposits on the catalyst surface. Meanwhile, it has been reported that Ru promoter causes higher reducibility of Co ion and better dispersion of Co metal in the catalyst support leading to higher CO conversion.34,35 In contrast, our magnetization studies show a minimal effect of Ru on Co ion reduction and corroborate that Ru rather has a synergistic effect on catalytic activities due to better dispersion of Co in the catalyst support.41 8.7 PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST SCREENING The development, optimization, and improvement of catalysts for use in commercially viable processes face major challenges for scientific research with respect to the complexities of the systems being investigated. Commercial catalysts, particularly heterogeneous systems, are often complex multielemental, multicomponent systems that are usually prepared via multistep procedures with the conditions under which each intermediate step is performed often impact performance.3 Starting from pharmaceutical research with respect to the identification and optimization of drugs, combinatorial methods for synthesis and screening have become increasingly important for other chemical and biological systems as well, for example, regarding homogeneous or heterogeneous catalysts or other types of materials. The goal of the development and application of parallel approaches to catalyst development is to provide equivalent tools and techniques that significantly expand the ability to explore the large complex parameter space addressed, while still generating results of an adequate quality in terms of resolution and accuracy. The achievement of this goal depends on a good understanding of how to integrate the parallel approach into the target activities. There are generally two approaches to combinatorial catalyst screening developed among different research groups: (1) high sample throughput with less information gained on each sample and (2) lower sample count with more complete information obtained on each sample, depending on the preparation and screening techniques.43 The first approach normally pursues a large number of candidates (>100) to extract 10 or so “leads” based on relatively limited screening information. In contrast, the second approach, which collects more detailed reaction information for a group of leads, then facilitates further elimination of less effective candidates from this group. In addition, the approach also yields information useful for process scaling and deployment in actual industrial processes. Specific advantages of using microreactors compared to conventional catalyst test systems have been described previously. Due to unique mass and heat transfer properties, as well as the uniform flow distribution, the microreactors are ideally

PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST SCREENING

315

free of any dead zones. Micoreactors provide a well-defined setting of operating conditions, and due to their small volume, fast changes in operating conditions can be performed with minimal time required to reach equilibrium. More specifically, microscale dimensions result in ultralow transport resistances such that the heat and mass transfer are extremely fast. The capability of microreactors to test a number of catalysts in separate reaction chambers, without any interference by flow mixing, guarantees high accuracy and reliability of the results, while maintaining a high analysis speed using compact devices. The microreactors consume reactant species only very slowly, thus allowing extremely expensive or rare chemical systems to be studied more economically. The small footprint of the microreactor and its peripherals need less infrastructure for operation, including floor space, energy supply, and support personnel.43 This compactness also allows an experimental approach whereby an array of microreactors are connected to share flow and analysis equipment, allowing a number of catalyst analysis experiments to be performed in parallel, increasing experiment throughput with reducing time required for development. These technical advantages are achieved with added benefit that the small reactant volumes are highly safe, being nearly exempt from explosion even when operating in what would normally be considered explosive regimes. In addition, the environmental hazards due to leakage are minimal. Based on the combinatorial screening approach and our experience with microfabricated reactor (microreactor) research described in previous sections, a silicon microreactor-based parallel catalyst analysis system was developed for industrially important reactions to dramatically decrease the catalyst development period and reduce overall operating costs (Figure 8.25). The flow of the reactant was controlled by Cole Parmer mass flow controller. These flow meters also indicate the absolute pressure in the line and are therefore useful for the alignment of the microreactors.44 Due to the low reactant flow rates required for the microscale reactions, the pressure in the system was built by carrier gas, helium, controlled by an Aalborg gas flow

FIGURE 8.25 A parallel array setup includes the following devices: (1) mass flow controllers, (2) heating blocks with microreactors inside, (3) pressure gauges, (4) gas flow controller of carrier gas, (5) multiline switching valve, (6) electrical connection, (7) mass spectrometer, (8) gas chromatography, and (9) gases for calibration.44

316

NANOCATALYSIS IN MICROREACTOR FOR FUELS

controller. This flow controller can provide a maximum of 10 sccm of flow rate. The switching between the four lines was done by VALCO multiposition valve that can switch between a maximum of 16 reaction lines with the help of an electronic actuator controlled both manually and automatically through the LabVIEW program. The pressure inside the setup is controlled by a Cole Parmer relief valve with the help of LabVIEW and the inputs from the mass flow meter and pressure gauges are used to open/close this relief valve such that the pressure inside the system is maintained at the set point provided by the user. The analysis of the product stream can be performed with the help of Stanford Research System Residual Gas Analyzer/Mass Spectrometer or a Varian Gas Chromatograph (GC). The injection to the GC was done with help of an 8-port valve. To achieve high-pressure synthesis, pressure has to be applied with the help of the carrier gas controlled though an Aalborg valve (maximum 10 sccm) at downstream of the multiposition valve. Since there is a pressure drop in nonselected streams by the VALCO multiposition valve, the downstream lines from the multiposition valve were vented out directly, and the carrier gas connection was made to the analyte line, thus maintaining pressure for the active reaction line. The actual parallel array setup and the complete process and instrument diagram (P&ID) are shown in Figures 8.25 and 8.26, respectively. The parallel array system of micoreactors with automatic control has been utilized for FT synthesis (CO and H2) to alkanes. As discussed in the previous section, today’s research groups in academia and industry have extensive interest to find a better catalyst to increase productivity, to control hydrocarbon product distribution, and to lengthen the catalyst life. The idea of using a parallel analysis system as a fast, economic, and easy-to-scale-up solution for FT catalyst development fits these requirements appropriately.

FIGURE 8.26

The P&ID of the parallel array setup of microreactors.44

PARALLEL MICROREACTOR SYSTEM FOR NANOCATALYST SCREENING

317

TABLE 8.2 CO Conversion on Three Different Catalysts in the Temperature Range of 100–250 C44 Reaction Temperature ( C)

CO conversion on Co catalyst CO conversion on Fe catalyst CO conversion on Ru catalyst

100

150

200

250

31% 74% 31%

85% 80% 45%

92% 82% 52%

90% 80% 62%

To fully understand the activities of Fe/Co/Ru in FT synthesis described in the previous section, we have tried to examine each catalyst in the microreactor separately using the parallel array system. Cobalt catalyst shows the highest CO conversion (Table 8.2).44,45 From the various characterization studies,44 cobalt catalyst is found to be well dispersed in the catalyst support matrix. This dispersion of catalyst allows a larger surface area of the active metal to be in contact with the feed gases, and hence higher conversion. The low conversion on the Ru catalyst (Table 8.2) can be attributed to the low dispersion of the catalyst; also the agglomeration of the catalyst particles reduced the amount of surface metal particles in contact with the feed gases. The absence of water–gas shift reaction with cobalt catalyst compared to iron and ruthenium is an indication of a different reaction mechanism. Iron and ruthenium catalyst can be seen to promote the same reaction mechanism that can be the CO insertion mechanism. The reaction temperature only increased the conversion of the reaction and had no prominent effect on the selectivity to products formed. The increase in residence time increased the production of long-chain compounds. The smaller and better dispersed the metal particles are, the more surface area it provides, and hence higher CO conversion is observed. From the above results, we can conclude that in reactions involving nanocatalysis, dispersion, particle size, and activity of supported metal catalysts play very significant roles. In general, parallel catalyst screening approaches for heterogeneously catalyzed gas-phase reactions have gained increasing popularity within the past years, as the development of novel and better catalysts for chemical processes is still mainly an empirical process utilizing existing technical know-how and experiences. Simultaneous handling of many samples and large amounts of data impose a need for the development of workflows and associated tool sets that minimize bottlenecks. The key components of the workflows are preparation system for synthesis, formation, and treatment of arrays of samples, reactor system for evaluating the performance of sample arrays for particularly target applications, characterization system for parallel characterization of key properties of the sample arrays, and informatics system for handling information and data flow between the various operations. Technical and costs constraints have in general limited how effectively the major reactor sections can be parallelized and for most groups have resulted in a systematic approach to the implementation of parallel methodologies to catalyst screening.3

318

NANOCATALYSIS IN MICROREACTOR FOR FUELS

Typically, the highest level of parallelization functions as a primary screening tool, providing evaluation of larger arrays of catalyst samples with a range of solutions compromising on the degree to which they approximate conventional systems. The secondary screening tools, which include the parallel array system described above, aim to provide almost the same level of individual control of conditions and detailed assessment of catalyst performance as typically performed on a bench-scale system. The described parallel catalyst screening setup has the flexibility to optimize the process conditions with accuracy equivalent to conventional units. This strategy of integrating detailed catalyst screening and flexible process control within a large parallel array system may help in the discovery of new commercial catalysts, the results of which may be successfully validated at all levels from the miniaturized parallel scale to final commercial operations. 8.8 SUMMARY Microreactors are recognized in recent years as a novel tool for chemistry and chemical process industry, such as fuel industry. The examples presented in this chapter only represent a small fraction of the many studies for microreactors being pursued by research groups worldwide. Invention of the next-generation process technologies in fuel processing will strongly depend upon discovery of new catalyst systems that result in attractive process economics with notably reduced environmental impact. The role of microreactor technology in the catalyst development and commercialization is expected to gain more importance due to its various advantages over conventional approaches. These advantages include improved safety characteristics, enhanced rates of heat and mass transfer, reduced hardware footprint, lower reagent costs, and ease of creating parallel systems for higher data throughput and improved workflow efficiency. In contrast, micropacked-bed reactors are easy to fabricate, but usually have a high-pressure drop during the passage of gases. Therefore, the walls of microreactors are more suitable for catalysis in this regard. We have used a few reactions to illustrate the advantages of performing chemical reactions in microreactors, which are particularly suited for highly exothermic and fast reactions. However, some issues such as high cost of fabrication facilities, limitation to highpressure reactions, and catalyst coating still exist. Extensive efforts are being made to address these problems. Most of the new microreactor developments described here are the results from studies of new chemistries with alternate synthetic routes, or as a result of scale-down from a previous, more conventional multiphase reactor.10 As a result, a greater opportunity exists to demonstrate the utility of microreactor as a robust enabling technology for the discovery and development of new multiphase catalyzed reactions. Significant progress in scaling up microchannel process technology to commercial scale has been made very recently. Focusing on solving challenges around device fabrication, flow distribution and catalyst integration are the keys to success. Various enabling technologies are allowing new microreactor designs to be fabricated. Advances in MEMS and microelectronic industry from the perspectives of design

REFERENCES

319

methodology, multifunctionality, new materials of construction, and fabrication techniques have allowed the development of more sophisticated 3D geometries at smaller length scales with increased spatial resolution. Adaptation of this knowledge into microreactor technology will be an ongoing challenge, but it should create an opportunity to combine various functionalities, such as onboard sensing and control systems, with microreactors into an integrated package. Results from today’s flow distribution models closely match experiments and are an integral tool for scaling microchannel reactors to commercial capacities. Since the mechanism in microfluidic flow is unique and distinct from that in conventional reactors, novel simulation methods and tools are expected to be discovered and invented. Uniform coatings of supported catalysts in the shape of thin films have been investigated extensively, but the ideal solution is still unknown, which encourages chemists and chemical engineers to pursue discovery of new technologies and effective modification of current methods. The larger task of integrating microreactors into functional microprocess systems for small-scale fuel processing is in the early stages of development. Different corporations and research centers, such as Sony, Casio, Exxon Mobil, and so on, have made significant progress in microreactor technology. The microfuel processor market can be divided into three device categories, portable electronics and portable units for military and healthcare segments.23 There is a great potential for microfuel processor to be integrated with microfuel cells to deliver more energy per volume or weight than conventional batteries. The initial focus is to use it as a recharger. The use of very small fuel cells in cell phones is a goal of many wireless carriers. When low cost, high efficiency, and reasonably small size reach the required level in today’s industry, and small enough to fit inside a cell phone, fuel cell powered electronic devices would be the favorite of many electronics customers. ACKNOWLEDGMENTS We gratefully acknowledge the financial support of NSF-EPSCoR and Louisiana BoR- RCS competitive Grant (to D.K.). We thank Drs. R. Besser, J. Palmer, and S. Naidu for help and suggestions and the graduate students, V. S. Nagineni, A. Potluri, W. Cao, Y. Liang, K. Shetty, and S. Mehta for their work described in this chapter. We are indebted to these students and Mr. J. Fang who have helped us in simulation, microfabrication, and experimental studies. We also thank Dr. K. Varaharamyan for his support of this project. REFERENCES 1. Haswell, S. J.; Skelton, V. Chemical and biochemical microreactors. Trends Anal. Chem. 2000, 19,6, 389–395. 2. Gavriilidis, A.; Angeli, P.; Cao, E.; Yeong, K. K. Y.; Wan, S. S. Technology and applications of microengineered reactors. Trans. IChemE 2002, 80, 3–30, Part A.

320

NANOCATALYSIS IN MICROREACTOR FOR FUELS

3. Karlsson, A.; Akporiaye, D. E.; Plassen, M.; Gillespie, R.; Holmgren, J. S. Parallel heterogeneous reactor systems for catalyst screening. Microreactor Technology and Process Intensification; ACS Symposium Series; Oxford University Press, 2005, Chapter 4. 4. Ehrfeld, W.; Hessel, V.; L€ owe, H.; State of the art of microreaction technology. Microreactors: New Technology for Modern Chemistry; Wiley-VCH: Weinheim, 2000; p 1–12. 5. Kawamura, Y.; Ogura, N.; Yamamoto, T.; Igarashi, A. A miniaturized methanol reformer with Si-based microreactor for a small PEMFC. Chem. Eng. Sci. 2006, 61, 1092–1101. 6. Hessel, V.; L€ owe, H.; Sch€ onfeld, F. Micromixers—a review on passive and active mixing principles. Chem. Eng. Sci. 2005, 60, 2479–2501. 7. Hessel, V.; L€ owe, H. Mixing principles for microstructured mixers: active and passive mixing. Microreactor Technology and Process Intensification; ACS Symposium Series; Oxford University Press, 2005, Chapter 21. 8. Liang, Y.; Nassar, R.; Fang, J.; Kuila, D.; Varahramyan, K. Investigation of a novel microreactor for enhancing mixing and conversion. Chem. Eng. Commun. 2008, 195(7), 745–757. 9. Liang, Y. Development of a novel microreactor for improved chemical reaction conversion, PhD dissertation, 2005. 10. Mills, P. L.; Quiram, D. J.; Ryley, J. F. Microreactor technology and process miniaturization for catalytic reactions—a perspective on recent developments and emerging technologies. Chem. Eng. Sci. 2007, 62, 6992–7010. 11. http://www.mikroglas.com/index.php?PAGE_ID¼538&LANG_ID¼9. 12. Jensen, K. F. Microreaction engineering—is smaller better? Chem. Eng. Sci. 2001, 56, 293–303. 13. Fang, J.; Wang, W.; Zhao, S.; Fabrication of 3D microfluidic structures. Encyclopedia of Microfluidics and Nanofluidics; Springer, 2008. 14. Zhao, S., Nano-scale platinum and iron-cobalt catalysts deposited in microchannel microreactors for use in hydrogenation and dehydrogenation of cyclohexene, selective oxidation of carbon monoxide and Fischer–Tropsch process to higher alkanes, PhD dissertation, 2003. 15. Shin, W. C.; McDonald, J. A.; Zhao, S.; Besser, R. Etching characteristics of a micromachined chemical reactor using inductively coupled plasma. Proceedings of the 6th International Conference on Microreaction Technology (IMRET VI), p. 357, AIChE, New Orleans, LA, 2002. 16. Meille, V. Review on methods to deposit catalysts on structured surfaces. Appl. Catal. A 2006, 315, 1–17. 17. Madou, M. Pattern transfer with additive techniques. Fundamentals of Microfabrication; CRC Press, 1997, Chapter 3, p 89–144. 18. Zhao, S.; Besser, R. Selective deposition of supported platinum catalyst for hydrogenation in a micromachined reactor. Proceedings of the 6th International Conference on Microreaction Technology (IMRET VI), p. 289, AIChE, New Orleans, LA, 2002. 19. Okuzaki, S.; Okude, K.; Ohishi, T. Photoluminescence behavior of SiO2 prepared by sol–gel processing. J. Non-Cryst. Solids 2008, 265, 61–67. 20. Campanati, M.; Fornasari, G.; Vaccari, A. Fundamentals in the Preparation of Heterogeneous Catalysts. Catal. Today 2003, 77, 299–314.

REFERENCES

321

21. Balat, M. Potential importance of hydrogen as a future solution to environmental and transportation problems. Int. J. Hydrogen Energy 2008, 33, 4013–4029. 22. Holladay, J. D.; Hu, J.; King, D. L.; Wang, Y. An overview of hydrogen production technologies. Catal. Today 2009, 139, 244–260. 23. Kundu, A.; Jang, J. H.; Gil, J. H.; Jung, C. R.; Lee, H. R.; Kim, S.-H.; Ku, B.; Oh, Y. S. Micro-fuel cells—current development and applications. J. Power Sources 2007, 170, 67–78. 24. Taegyu Kim, T.; Kwon, S. MEMS fuel cell system integrated with a methanol reformer for a portable power source. Sens. Actuators A 2008, 204–211, available online. 25. Kundu, A.; Park, J. M.; Ahn, J. E.; Park, S. S.; Shul, Y. G.; Han, H. S. Micro-channel reactor for steam reforming of methanol. Fuel 2007, 86, 1331–1336. 26. Men, Y.; Gnaser, H.; Zapf, R.; Hessel, V.; Ziegler, C.; Kolb, G. Steam reforming of methanol over Cu/CeO2/g-Al2O3 catalysts in a microchannel reactor. Appl. Catal. A 2004, 277, 83–90. 27. Shetty, K.; Zhao, S.; Seetala, N. V.; Kuila, D. Synthesis and characteriztion of non-noble nano-catalysts for hydrogen production in microreactors. J. Power Sources 2007, 163(2), 630. 28. Shetty, K.; Zhao, Cao W. S.; Seetala, N. V.; Kuila, D. Silica sol-gel supported nickel nanocatalyst for hydrogen production using microreactors. Mater. Res. Soc. Symp. Proc. 2006, 885, 259. 29. Nikolaidis, G.; Baier, T.; Zapf, R.; Kolb, G.; Hessel, V.; Maier, W. F. Kinetic study of CO preferential oxidation over Pt–Rh/g-Al2O3 catalyst in a micro-structured recycle reactor. Catal. Today, 2008, available online. 30. Srinivasa, S.; Dhingrab, A.; Imb, H.; Gularia, E. A scalable silicon microreactor for preferential co oxidation: performance comparison with a tubular packed-bed microreactor. Appl. Catal. A 2004, 274, 285–293. 31. Snytnikov, P. V.; Popova, M. M.; Menc, Y.; Rebrov, E. V.; Kolb, G.; Hessel, V.; Schouten, J. C.; Sobyanin, V. A. Preferential CO oxidation over a copper–cerium oxide catalyst in a microchannel reactor. Appl. Catal. A 2008, 350, 53–62. 32. Zhao, S.; Hu, J.; Kuila, D.; Besser, R.; Nassar, R.; Palmer, J. Nano-scale platinum catalyst in microreactors for preferential oxidation of CO amelioration in hydrogen fuel cell feeds. Microreactor Technology and Process Intensification; ACS Symposium Series; Oxford University Press, 2005, Chapter 8. 33. Ouyang, X.; Bednarova, L.; Ho, P.; Besser, R. S. Preferential oxidation of carbon monoxide in a thin-film catalytic microreactor: advantages and limitations. AIChE J. 2005, 51,6 1758–1772. 34. Iglesia, E. Design, synthesis, and use of cobalt-based Fischer–Tropsch synthesis catalysts. Appl. Catal. A 1997, 161, 59–78. 35. Iglesia, E.; Soled, S. L.; Fiato, R. A.; Via, G. H. Bimetallic synergy in cobalt-ruthenium Fischer–Tropsch synthesis catalysts. J. Catal. 1993, 143, 345–368. 36. Diehl, F.; Khodakov, A. Y. Promotion of cobalt Fischer–Tropsch catalysts with noble metals: a review. Oil Gas Sci. Technol. Rev. IFP 2009, 64,1, 11–24. 37. Li, S.; Krishnamoorthy, S.; Li, A.; Meitzer, G. D.; Iglesia, E. Promoted iron-based catalysts for the Fischer–Tropsch synthesis: design, synthesis, site densities, and catalytic properties. J. Catal. 2002, 206, 202–217.

322

NANOCATALYSIS IN MICROREACTOR FOR FUELS

38. Nagineni, V. S.; Zhao, S.; Siriwardane, U.; Seetala, N. V.; Fang, J.; Palmer, J.; Kuila, D. Microreactors for syngas conversion to higher alkanes: characterization of sol–gel encapsulated nano-scale Fe–Co catalysts fabricated in the microchannels. Indus. Eng. Chem. Res. 2005, 44(15), 5602. 39. Nagineni, V. S.; Zhao, S.; Indukuri, H.; Liang, Y.; Potluri, A.; Siriwardane, U.; Seetala, N.; Fang, J.; Kuila, D. Characterization of alumina and silica sol–gel encapsulated Fe/Co/Ru nanocatalysts in microchannel reactors for F–T synthesis of higher alkanes. Mater. Res. Soc. Symp. Proc. 2004, 820, P51. 40. Zhao, S.; Nagineni, V. S.; Hu, J.; Liang, Y.; Fang, J.; Nassar, R.; Siriwardane, U.; Naidu, S. V.; Varahramyan, K.; Palmer, J.; Kuila, D. Microreactor research and development at LA Tech University: fabrication of silicon microchannel reactors for catalyst studies on conversion of cyclohexene as a prototype and syngas to alkanes. Microreactor Technology and Process Intensification; ACS Symposium Series; Oxford University Press, 2005, Chapter 5. 41. Zhao, S.; Nagineni, V. S.; Seetala, N. V.; Kuila, D. Microreactors for syn-gas conversion to higher alkanes: the promoter effect of ruthenium on silica supported iron–cobalt catalysts. Ind. Eng. Chem. Res. 2008, 47(5), 1684. 42. Zhang, Y.; Hou, L.; Tierney, J. W.; Wender, I. Addition of acetylene to the Fischer
Tropsch reaction. Energy Fuels 2007, 21(2), 640–645. 43. Ouyang, X.; Besser, R. S. Development of a microreactor-based parallel catalyst analysis system for synthesis gas conversion. Cataly. Today 2003, 84, 33–41. 44. Mehta, S.; Thesis, M. S. Comparative studies of silica sol–gel supported iron, cobalt and ruthenium catalysts for Fischer–Tropsch reaction in microreactors, Louisiana Tech University, 2007. 45. Mehta, S.; Zhao, S.; Fang, J.; Kuila, D. Comparative studies of silica sol–gel supported Fe, Co, and Ru catalysts for F–T reaction in Si-microreactors. Preprints Am. Chem. Soc., Div. Fuel Chem. 2008, 53(2), 584–585.

9 MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES
ALI ABOU-HASSAN, OLIVIER SANDRE, AND VALERIE CABUIL

Laboratoire de Physicochimie des Electrolytes Collo€ıdes et Sciences Analytiques (PECSA), UMR 7195, Equipe Collo€ıdes Inorganiques, Universit
e Paris 6, Paris Cedex 5, France

9.1 INTRODUCTION Iron oxides are widespread in nature1 and are present almost everywhere in the global system, even in Mars’ soil.2 They are used for various applications in industry as colored pigments, magnetic materials, ferrofluids, catalysts, and so on. There are 16 species of iron oxides, hydroxides, or oxyhydroxides, which will be collectively referred to as iron oxyhydroxides in this chapter (Table 9.1). For more details on different iron oxyhydroxides, the reader can refer to the book by Cornell and Schwertmann.3 All the iron oxyhydroxides are of great interest and have numerous applications, but this chapter will focus only on materials widely studied in the past few years, that is, magnetic iron oxides such as magnetite Fe3O4 and maghemite g-Fe2O3 nanoparticles. These materials have wide-ranging technological applications when they are divided into nanoparticles, ranging from navigation with magnetite (or Lodestone) to modern high-density magnetic recording media and read head devices. Magnetite Fe3O4 is a black ferrimagnetic mineral containing both FeII and FeIII and has an inverse spinel structure. Maghemite g-Fe2O3 is a red-brown ferrimagnetic material, isostructural with magnetite but with cation-deficient sites. When the size of

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

323

324

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

TABLE 9.1

The Main Iron Oxyhydroxides

Oxide hydroxides and hydroxides

Oxides

Goethite a-FeOOH Lepidocrocite g-FeOOH Akagan
eite b-FeOOH Schwertmannite Fe16O16(OH)y(SO4)z
nH2O d-FeOOH Feroxyhyte d0 -FeOOH High-pressure FeOOH Ferrihydrite Fe5HO8
4H2O Bernalite Fe(OH)3 Fe(OH)2

Hematite a-Fe2O3 Magnetite Fe3O4 Maghemite g-Fe2O3 b-Fe2O3 e-Fe2O3 W€ ustite FeO

Readapted from Ref 3.

magnetite or maghemite particles is reduced below about approximately 15 nm, the particles are magnetic monodomains. It means that they have a permanent magnetic moment whose intensity is proportional to their volume, but the direction depends on the spontaneous fluctuations inside the grain. This magnetic moment is due to the crystalline order, a characteristic of the spinel-like structure. For ultrasmall particles (diameter smaller than a few nanometers), the surface disorder leads to a very significant decrease in the moment. Thus, nanometric ferro- or ferrimagnetic particles behave very differently from the corresponding bulk materials and their magnetic behavior is called superparamagnetism.4 The main characteristic of superparamagnetism is the spontaneous fluctuation of the direction of the magnetic moment in the small magnetic grain, which is due to the fact that, for very small ferromagnetic particles, the magnetic anisotropy energy (KV) responsible for keeping the magnetization oriented in the easy axis of magnetization is comparable to the thermal energy (kT). This results in zero magnetization in zero field if the fluctuations are averaged over a timescale larger than their typical time t.5 The fluctuation time t generally varies over a very broad timescale, depending on the size of the particles (Figure 9.1) Superparamagnetism also refers to the extremely large magnetic moments that these nanoparticles bear (typically a few tens of thousand Bohr magnetons mB) compared to the moment of isolated ions (5.4 mB for Fe2 þ and 5.9 mB for Fe3 þ ). When placed in external magnetic fields, the magnetization of a superparamagnetic suspension of nanoparticles is about 104 times larger than the magnetization of a paramagnetic solution with an equivalent iron salt concentration. Ferromagnetic bulk materials, once magnetized, show remanence (i.e., remain partially magnetized even in the absence of an applied field) and therefore are used as recording materials. In contrast, superparamagnetic materials differ from ferromagnetic bulk substances because they do not retain any magnetization once the external field is removed.6 Among others, superparamagnetic nanoparticles are largely used in magnetic storage media,7 for biosensing applications,8 for medical applications, such as

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

325

FIGURE 9.1 (a) For very small ferromagnetic (FM) particles, the magnetic anisotropy energy (responsible for keeping the magnetization oriented in certain direction) is comparable to the thermal energy (kT). When this happens, the particles become superparamagnetic, as thermal fluctuations randomly flip the magnetization direction between parallel and antiparallel orientations. (b) Typical magnetization curve for superparamagnetic nanoparticles (Langevin’s curve). Under a zero magnetic field, the magnetic moments are randomly oriented, but they progressively align parallel to the field direction when a magnetic field is applied. When all the magnetic moments are aligned with the magnetic field, the curve attains a saturation value Ms, which is the product of the volume fraction F by the specific magnetization ms of the material (e.g., 3  105 A m1 for colloidal maghemite, which corresponds to 33 Bohr magnetons per nm3).

targeted drug delivery,9 as contrast agents in magnetic resonance imaging,10 and as ferrofluids.11–14 For most of these applications, it is necessary to control the production of magnetic nanoparticles, their monodispersity, and their states of aggregation, as these physical parameters control their physical and physicochemical properties. In the past few years, several research groups have proposed to use microfluidic systems as a promising strategy for obtaining high-quality nanoparticles with high monodispersity in a single-shot process without any subsequent size selection. This chapter reviews the recent scientific literature concerning the use of microfluidics for the synthesis of the iron oxide nanomaterials over the past 5 years. After a review of the main synthesis methods used to prepare these materials in bulk chemistry, a few works related to the synthesis of ferric oxide nanoparticles in microfluidics will be introduced. 9.2 MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES The chemistry of iron oxyhydroxides is very diversified and rich. Almost all the species can be formed from solutions by a polycondensation mechanism, which is the

326

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

main topic of this section. For more details on the mechanisms and kinetics of the precipitation from ionic solution, the reader can refer to Refs 15 and 16. 9.2.1 Metallic Cations in Solution and Polycondensation Metal cations Mz þ in water are solvated by dipolar water molecules giving rise to aquo cations [M(OH2)6]z þ .17 In the particular case of iron salts (chloride, nitrates, etc.), dissolution in water produces hexacoordinated aquo complexes [Fe(OH2)6]z þ , where z ¼ 2 or 3. The polarization of coordinated water molecules in the coordination sphere strongly depends on the oxidation state and size of the cation. Charge transfer occurs via the Fe–OH2 s-bond and electron density is transferred from the bonding 3a1 molecular orbital of coordinated water molecules toward empty orbitals of the metal cations.18 This charge transfer results in weakening of the O–H bond within the water molecule, and the aquo complexes manifest Brønsted acid–base properties leading to deprotonation of coordinated water molecules: ½FeðOH2 Þ6 z þ þ hH2 O ! ½FeðOHÞh ðOH2 Þ6h ðzhÞ þ þ hH3 O þ The higher the oxidation state of the cation, the lower its size and the higher the acidity of the complex. This makes the ferric aquo complexes more acidic than ferrous complexes and hydroxylation of the cations occurs at very distinct ranges of pH, as indicated in Figure 9.2. The hydroxylation ratio h of a complex increases when the pH increases and aquohydroxo or oxohydroxo complexes are formed. In general, hydroxylated cation monomers are instable in solution. They spontaneously condense because of the nucleophilic character of the OH ligands and the electrophilic character of cations. Depending on the nature of the coordination sphere, two basic mechanisms of the cations are proposed for the condensation of hydroxylated complexes.19 Aquohydroxo complexes condense through a nucleophilic substitution that proceeds

FIGURE 9.2 Speciation of [Fe(OH)h(OH2)6h](zh) þ complexes of (a) Fe(II) and (b) Fe(III). Reprinted with permission from Ref. 15. Copyright 2004 the Royal Society of Chemistry.

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

327

when the coordination number of the hydroxo ligand is increased and water molecules are eliminated. This mechanism is called olation.

For oxyhydroxo complexes, there is no water molecule in the coordination sphere of the complexes and therefore no leaving group. The condensation mechanism proceeds in that case by two steps: .

First, association of the oxohydroxo complexes.

.

Then, elimination of water molecule and formation of oxo bridges.

As long as charged complexes exist in solution, the condensation is limited and polycationic species of 10–20 cationic atoms are formed. For iron complexes, which are very reactive, ferric species condense very rapidly as soon as pH  1 and it is difficult to isolate polycationic species. On the contrary, ferrous complexes condense only above pH 6, so some ferrous polycationic species have been isolated. When only zero-charge complexes exist, the condensation is unlimited and a solid phase precipitates: 0
n MðOHÞz ðOH2 ÞNz ! ½MðOHÞz n þ nH2 O The precipitation is accompanied by the elimination of all the coordinated water molecules and results in hydroxide formation. When the precipitated hydroxide is unstable, it dehydrates spontaneously to form oxides and oxyhydroxides. For example, alkalinization at room temperature of an aqueous solution of ferric ions quasi-instantaneously leads to a poorly defined highly hydrated phase called ferrihydrite. Depending on the pH of the solution, the ferrihydrite suspension evolves either to the oxyhydroxide phase a-FeOOH (goethite) or to the oxide phase a-Fe2O3 (hematite).

328

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

9.2.2 Kinetic Steps for the Precipitation Process To understand why precipitation leads to the formation of nanometric particles and how microfluidics can be used as a tool to elucidate nanoprecipitation mechanisms, we examine in this section the kinetics of the polycondensation process. We refer to the solid precursor as the zero-charged complex [Fe(OH)z(OH2)Nz]0 obtained by hydroxylation of the iron salt solution by an alkaline solution, as we introduced previously. In the following, we abbreviate the zero-charged complexes by P. Three steps are usually considered to describe the formation of solid particles: nucleation, growth (primary growth), and aging (secondary growth). Nucleation can be homogeneous nucleation, heterogeneous nucleation, or secondary nucleation.20 We consider here the simplest case of a homogeneous nucleation, also called classical nucleation theory (CNT), which occurs in the absence of a solid interface and consists in combining solute molecules to produce nuclei. This step leads to the formation of small clusters with a number of iron atoms sufficiently large to overcome the nucleation barrier. The global rate of such a process can be written as v ¼ k[P]a, where the values of a can range between 4 and 10, as proposed by Nielsen.16 These high a-values mean that the nucleation process is not an elementary reaction but the result of many chemical elementary steps. As the concentration of the precursor P generated by hydroxylation increases, and possibly reaches a critical concentration, the condensation rate increases, leading to the formation of many nuclei. This induces a decrease in the precursor concentration and in the condensation rate, which can be annulled if the concentration of P is very low. For any chemical process, the driving force behind the homogeneous nucleation is the total free energy of the supersaturated solution DG. The overall free energy of the nucleation phenomena can be written as DG ¼ DG1 þ DG2, where DG1 is the volume contribution resulting from the difference between the chemical potential of ions in the nuclei (mPn) and in solution (mP) and DG2 is the contribution of the interfacial energy (g) when a solid–liquid interface of surface area (A) is created.20 If we suppose that the global reaction leading to a nuclei formation from p precursors is pP , Pp then DG1 and DG2 can be written as DG1 ¼ pðmPn mP Þ ¼ pRT ln S

and

DG2 ¼ gA

S is called “supersaturation,” which represents the ratio of the precursor concentration in the solution to the solubility Cb of the macroscopic (bulk) solid, that is, S ¼ [P]/Cb. Spontaneous nucleation can occur if S > 1 (DG1 < 0), while no nuclei can form when S < 1 (DG1 > 0). However, even if S > 1, the nuclei can disappear if their size is

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

329

not sufficient to overcome the energy barrier due to the competition with interfacial energy DG2. DG2 ¼ gp2=3 ð36pv 2 Þ1=3 The total free enthalpy is thus DG ¼ pRTLnS þ gp2=3 ð36pv 2 Þ1=3 The variation in the total free enthalpy with respect to p reaches a maximum when q(DG)/qp ¼ 0. This allows to define a critical number of precursor molecules p , a critical spherical radius r beyond which the growth of nuclei is spontaneous, and the energetic barrier that the system should overcome to reduce its surface energy and to minimize the total free energy DG, DG : p* ¼

32pv 2 g3 3ðRTLnSÞ

; 3

r* ¼

2v g ; RTLnS

and

DG* ¼

16 pv 2 g3 p* RTLnS ¼ 2 3 ðRTLnSÞ2

From the above equations, it follows that the higher the saturation ratio S, the smaller the critical nuclei size r and the higher the DG . Indeed, for a given value of S, all particles with r > r will grow and all particles with r < r will dissolve. Figure 9.3 illustrates this thermodynamic approach for the nucleation process for several cases of supersaturation. When the concentration of the precursor reduces below the minimum concentration for nucleation, the latter stops, whereas the growth continues until the saturation equilibrium concentration of the precipitated species is reached (i.e., the solubility Cb of the bulk solid). In the classical ion-mediated crystal growth, growth occurs by addition of soluble species to the solid phase. The uniformity of the size distribution can be achieved through a short nucleation period that generates all the particles obtained at the end of the nucleation followed by a self-sharpening growth process. At this stage, the system is under kinetic control; the smaller particles grow more rapidly than the larger ones because the free energy driving force is larger for smaller particles than for larger ones if the particles are slightly larger than the critical size r. Figure 9.4 shows the variation in the precursor concentration with time during the precipitation in the ideal case when growth successively follows the nucleation step. This is the

ΔG ΔG *

(a) (b)

(c)

p*

p

FIGURE 9.3 Illustration of the overall free energy DG as a function of the number of the precursors p in the nuclei. (a) S < 1; (b and c) S > 1 and Sc > Sb.

330

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.4 Cartoon illustration of nucleation and growth during the preparation of monodisperse nanoparticles. Reprinted with permission from Ref. 23. Copyright 2004 the Royal Society of Chemistry.

famous model proposed first by LaMer and Dinegar to explain the mechanism of formation of sulfur sols.21,22 But in most systems, depending on the concentration of the precursor and the relative rates of the precursor formation and nucleation, nucleation and growth can occur successively or at the same time. Ideally, a requirement to achieve the monodispersity of the nanoparticles is that nucleation and growth are separated, in time or in space (in separate vessels). In practice, nearly monodisperse size distribution can be obtained by quickly stopping the nucleation growth (thermal quenching) or by supplying a reactant source to keep saturated conditions during the whole reaction.24 Growth processes are traditionally referred to as ripening or coarsening. Two primary growth mechanisms, illustrated in Figure 9.5, are commonly active to varying degrees during the ripening process.25 In the first growth mechanism, known as Ostwald ripening, larger particles grow at the expense of smaller ones, which are less stable because the solubility of a particle depends on its dimension, according to the Gibbs–Thomson equation26 Cr ¼ Cb expð2sVm =rRTÞ where Cr and Cb are the solubility values of the nanocrystals and the corresponding bulk solid, respectively, s is the interfacial tension, Vm is the molar volume of the materials, r is the radius of particles, R is the gas constant, and T is the temperature. The coefficient 2sVm =RT called “capillary length” is usually on the order 1 nm.27 The

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

331

Nucleation clusters Crystal growth Primary nanoparticles > 3 nm

Amplifications

Mesoscale assembly

Single crystal

Iso-oriented crystal

FIGURE 9.5 Alternative mechanisms of growth for single crystals. The classical crystal growth model is described by the path on the left. The path on the right involves the arrangement of primary nanoparticles into an iso-oriented crystal via oriented attachment, which can form a single crystal upon fusion of the nanoparticles and elimination of the grain boundaries. Partially adapted from Ref. 30.

solubility based on the Gibbs–Thomson equation describes the solubility of colloidal particles whose radius is larger than about 20 nm. For nanoparticles with r ¼ 1–5 nm, the value of the capillary length approaches the particle radius and the particle solubility ln(Cr) becomes strongly nonlinear against r1,28 presumably because the interfacial tension s for a particle with a small number of atoms can no longer be approximated by the value of the macroscopic solid phase. In addition, it is not necessary that the nuclei are made of the thermodynamically stable crystalline phase; they could be made either of an amorphous phase or of a metastable allotropic phase. The kinetics of Ostwald ripening crystal growth can usually be described by the following power law:29 DðtÞ ¼ D0 þ k
t1=n where D0 is the initial particle size (diameter), D(t) is the size at time t, and k is a rate constant for the limiting step. The exponent n is determined by the nature of the ratelimiting step. It is equal to 1 when the rate of the growth is controlled by diffusion in solution, equal to 2 when it is controlled by diffusion at the particle surface, and equal to 3 when it corresponds to the interface dissolution/precipitation step. In the second growth mechanism, known as Smoluchowski ripening, particles grow by coalescence through convection or active mixing. The nanoparticles themselves act as the building blocks for crystal growth. An oriented pair of nanoparticles aggregates and fuses (like in thermal sintering), eliminating the crystalline defects at their interface and releasing interfacial energy.31

332

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

This less known mechanism for nanoparticles growth through oriented attachment was first discovered in the samples of hydrothermally treated nanocrystalline TiO2.32–34 Natural nanoparticles of iron oxy/hydroxyl oxides were also found to grow by this mechanism under certain geochemical conditions.35 Despite its great popularity (565 citations according to ISI Web of Science in March 2009), the classical model proposed by LaMer seems to be rigorously appropriate only to the system it was developed for (sulfur sols) and other closely analogous systems of inorganic molecular crystals.36 Shortly after LaMer’s work, Turkevich called it a “theory of great tradition” but ultimately rejected it to interpret the kinetics of the synthesis of gold sols by reduction of chloroauric acid with citrate ions.37 Another issue in generalizing the LaMer mechanism to other systems is that at the end of the nucleation period, nuclei aggregation or Ostwald ripening may have already started to generate larger nuclei that are more stable than smaller ones, resulting in a range of nuclei sizes.36,38 As another limitation of LaMer’s theory and its variants, diffusion is not always the rate-determining step in particle growth.39 Microfluidics appears as a very relevant tool to study the kinetics of the synthesis of particles. Indeed, in a reactor as the one illustrated in Figure 9.6, due to the steady-state laminar flow, an almost linear relation exists between the position in the reactor and the time.40 A direct observation of the reaction mixture at several points of the reactor, if any suitable detection method is available, provides information about the kinetics of the nucleation and growth processes.41 Mixing and observing at the same time will reduce the dead time that is, even for highly efficient mixers, very long compared to the rate of precipitation reactions. Mixing in continuous-flow microreactors operating under laminar flow occurs by diffusion of the species at the point of confluence.42 Compared to bulk chemistry, where mean concentrations are used to describe the chemical system, chemistry in laminar flows needs to account for the local

FIGURE 9.6 Cartoon showing a typical Y-shaped continuous-flow microfluidic reactor operating under laminar flow, where the reagents mix by diffusion. At the interface, nanoparticles nucleate and grow.

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

333

concentration of species and their resulting gradients. In the case of the nucleation process, a local precursor concentration can be defined at every point of the reactor, giving rise to a local supersaturation and then to different nucleation phenomena. It means that different nuclei with different sizes will be formed depending on the local conditions. This normally should favor the increase in polydispersity of nanoparticles, unless a fast mixing of the reagents can occur, creating a burst nucleation followed by a fast growth. Anyway, microfluidics appears as a versatile tool to screen the effect of several parameters on the size and shape of particles. Indeed, mixing and residence times can be easily manipulated. 9.2.3 Case of Magnetite and Maghemite Nanoparticles Numerous bulk chemical methods can be used to synthesize magnetic nanoparticles: coprecipitation of iron salts,43–46 sol–gel synthesis,47 hydrothermal reactions,48 hydrolysis and thermolysis of precursors,49 synthesis in microemulsions,50 flow injection synthesis,51 and electrospray synthesis.52 Contrary to the case of silica or titania, for which a large variety of organometallic precursors exist, providing a good control of the precipitation kinetics, organometallic iron precursors are less abundant and highly reactive; thus, synthesis involving these precursors cannot be used for the synthesis of iron oxide particles. Until now, only one process that has been extended to microfluidics for the synthesis of magnetic nanoparticles is the most used technique in bulk chemistry, that is, coprecipitation of ferrous and ferric salts in alkaline medium. As polyol processes and thermal decomposition processes can be easily (with some conditions) extended to microfluidics, they will also be discussed here. 9.2.3.1 Coprecipitation Coprecipitation is a facile and a convenient way to synthesize iron oxides (either Fe3O4 or g-Fe2O3) in water from a stoichiometric aqueous Fe2 þ /Fe3 þ salt solution by adding a base under inert atmosphere at room temperature or elevated temperature. The chemical reaction of Fe3O4 formation may be written as Fe2 þ þ 2Fe3 þ þ 8OH ! Fe3 O4 þ 4H2 O This procedure is in fact a polycondensation process, with nucleation and growth steps. Quantitative data on nucleation and growth of hydrous metal oxides or hydroxides are rather limited. The reason is that during the precipitation of solids, competing reactions such as hydrolysis, condensation, and anion coordination take place concurrently. The elucidation of the processes is even more difficult when several solute complexes become involved in solid phase formation.53 In the case of iron oxides such as magnetite and maghemite, due to the high reactivity of Fe(II) and Fe(III), fast hydrolysis and condensation occur, leading to concurrent nucleation and growth (primary and secondary) and thus wide size distribution.

334

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

According to thermodynamics of the precipitation reaction, complete precipitation of Fe3O4 should be expected at a pH between 8 and 14, with a stoichiometric ratio of 2:1 (Fe3 þ /Fe2 þ ) in a nonoxidizing media.54 Experimental results show that the size, shape, and composition of the magnetic nanoparticles strongly depend on the type of anions associated with the ferric and ferrous cations (e.g., chlorides, sulfates, and nitrates), the molar ratio (Fe3 þ /Fe2 þ ), the reaction temperature, the pH value, and the ionic strength of the synthesis medium. Magnetite Fe3O4 nanoparticles are sensitive to oxidation. Magnetite evolves into maghemite g-Fe2O3 in the presence of oxygen. The latter is chemically stable in alkaline and acidic media. During oxidation of magnetite to maghemite, various electron or ion transfers are involved depending on the pH of the suspension. Oxidation under alkaline conditions involves the oxidation of the surface of particles, while under acidic and anaerobic conditions, surface Fe2 þ ions are desorbed as hexa-aqua complexes in solution. Rapid and complete oxidation can be achieved in acidic medium, as described by Massart and Cabuil.13 The main advantage of the coprecipitation process is that a large amount of nanoparticles can be synthesized, without any surfactant. However, achieving by this process a narrow particle size distribution without performing any size sorting is still a challenge. The first controlled preparation of superparamagnetic iron oxide particles by alkalinization of an aqueous mixture of FeCl3 and FeCl2 salts was performed by Massart in the 1980s.12 The synthesized nanoparticles were roughly spherical and XRD measurements showed a diameter of 8 nm. Different parameters of this process were largely studied to demonstrate the influence of the pH value, the base (ammonia, CH3NH2, and NaOH), the added cations (N(CH3)4 þ , CH3NH3 þ , Na þ , Li þ , K þ , and NH4 þ ), and the Fe2 þ /Fe3 þ ratio, denoted x, on the coprecipitation yield, the diameter, and the polydispersity of the nanoparticles. By modulating the different parameters, magnetic nanoparticles with a mean diameter ranging between 16 and 4 nm were prepared with a good reproducibility.13 The same results were obtained by Vayssieres et al.55 and Jolivet et al.54,56–58 The latter explained the shape tailoring by the variation of the electrostatic surface density of the nanoparticles determined by the chemical composition of the crystal surface, the pH, and the ionic strength. Babes et al.59 investigated the effect of iron concentration and the molar ratio x. When x increased, the mean size of particles increased but the synthesis yield decreased. The particles synthesized by Massart’s process have been coated with a wide range of molecular species such as amino acids, a-hydroxyacids (citric, tartric, and gluconic acids),60 hydroxamate (arginine hydroxamate),61dimercaptosuccinic acid (DMSA),44,62 or phosphoryl choline.63 Bee et al.45 investigated the effect of the concentration of citrate ions on the size of maghemite particles synthesized by Massart’s process. Increasing the amount of citrate ions allows a decrease in the diameter of citrate-coated nanoparticles from 8 to 3 nm. The authors explained these results by the chelation effect of the citrate on the ferric and ferrous cations, preventing nucleation, and by the adsorption of citrate on the nuclei, inhibiting the growth of the latter. Also, the authors took advantage from the adsorbed citrate species to stabilize

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

335

the nanoparticles in aqueous dispersion at neutral pH.64 The effect of citrate during the synthesis of iron oxide was also studied by Liu and Huang.65 The crystallinity of the synthesized nanoparticles decreased when the concentration of citrate was increased during the synthesis, and the presence of citrate induced changes in the surface geometry of the nanoparticles. Similarly, Barker et al.25 showed that by capping the magnetite nanoparticles during the synthesis with heptanoic acid in trioctylamine solvent, they were able to slow the ripening process, thus reducing the defects in the nanoparticles. Thus, the size and shape of the nanoparticles can be tailored by adjusting the pH, the ionic strength, the temperature, the nature of the iron salts, and the Fe2 þ /Fe3 þ molar ratio or by adding chelating organic anions (caboxylate, citric, gluconic, or oleic acid). Other factors such as the mixing rate or the mixing manner can also affect the size and polydispersity of particles. For example, a decrease both in the size and in the polydispersity is observed when the base is added to the aqueous solution of metallic salts compared to the opposite process where the solution of iron salts is added to the alkaline solution.13 Surprisingly, injection flux rates do not seem to have a preponderant influence on the nanoparticle synthesis.59 For magnetite nanoparticles, there were much more studies on the growth mechanism and its consequences on the magnetic properties than on the nucleation step. In the case of iron oxyhydroxide particles, dissolution–crystallization plays an important role in growth mechanisms. It depends on several parameters such as particle size, pH, ionic strength, presence of additives, and so on. Different values are provided for the solubility products (Ks) of the several iron oxyhydroxides, according to the authors. This may be due to differences in the characteristics of particles (size, shape, surface state, etc.). In general, the Ks values of different iron oxyhydroxides range from 1044 to 1034.3 Concerning the Fe3O4 solubility, there are in fact large discrepancies in its solubility especially in alkaline medium.66 This is probably due to the dissolution mechanism that involves the reduction of FeIII to FeII.67 As a result, the solubility is a function of the reduction potential of the system, which is a real problem under alkaline conditions as dissolved O2 is an extremely oxidizing agent and kinetic effects may be important. Ferrihydrite precipitation and aging in solution illustrates nicely how pH controls the solubility and thus the mechanisms of evolution of a population of nanoparticles in suspension (Figure 9.7). The evolution of the small amorphous nuclei of ferrihydrite obtained by alkaline precipitation of iron(III) salts (nitrate, chloride, etc.) strongly depends on pH (thus on solubility): in the range 5  pH  8, the insoluble ferrihydrite germs transform by in situ dehydration and local rearrangement into very small acicular particles of hematite a-Fe2O3, whereas for a higher solubility in acidic (pH < 4) or alkaline (pH > 8) media, the transformation proceeds more easily via a dissolution–crystallization process, leading to large goethite needles. It seems that both Ostwald ripening and coalescence are involved in the growth of the magnetite nanoparticles. Vayssieres et al.55 showed that the size of magnetite precipitated in aqueous solution can be adjusted and stabilized against ripening by controlling the pH and the ionic strength, the latter being imposed by a noncomplexing salt in the precipitation medium.

336

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.7 Influence of pH on the solubility of iron and ferric (hydro) oxide crystal structure. Reprinted with permission from Ref. 15. Copyright 2004 the Royal Society of Chemistry.

9.2.3.2 Thermal Decomposition Decomposition of organometallic compounds in high-boiling organic solvents containing stabilizing surfactants is a procedure that has been widely used to produce magnetic nanoparticles because these nanoparticles are obtained with a high level of monodispersity and size control. The iron organic precursors are [Fe(acac)3] (acac ¼ acetylacetonate), Fe(Cup)3 (Cup ¼ N-nitrosophenylhydroxylamine, C6H5N(NO)O), or Fe(CO)5. Hexadecylamine, oleic acid, and fatty acids are often used as surfactants. The size and morphology of the nanoparticles can be controlled by adjusting the reaction times, as well as the aging period, the temperature, the concentration and ratios of the reactants, the nature of the solvent and the precursors, and the addition of seeds. The decomposition of iron pentacarbonyl (Fe(CO)5) in a mixture of octyl ether and oleic acid and at 100 C, followed by oxidation by trimethylamine oxide ((CH3)3NO) at elevated temperature, resulted in the formation of monodisperse maghemite nanocrystals with a size of approximately 13 nm.68 The decomposition of [Fe(acac)3] in the presence of 1,2-hexadecanediol, oleylamine, or oleic acid in phenol ether leads directly to oxides.69 The use of iron(III) chloride salts as a iron

MAIN BULK PROCEDURES FOR THE SYNTHESIS OF IRON OXIDE NANOPARTICLES

337

source has been proposed for the preparation of magnetic nanoparticles.70,71 For details, the reader can refer to the reviews by Tartaj and Sato et al.72,73 The nanoparticles obtained by this procedure are dispersible in different organic solvents (hexane and toluene) but not in water, and sophisticated postpreparative methods are needed to make these nanocrystals water soluble. Until now, microfluidic reactors have not been used for this kind of synthesis, but the possibility to manipulate small volumes could be of great interest as soon as the structure of microreactors and the volumetric rate flow of different reagents permit the control of the reaction times and the aging periods of the chemical reactions. Also, due to the small dimensions of the channels, a precise control of the temperatures and the temperature gradients could be more precise than in bulk. However, even if thermal decomposition enables the synthesis of monodisperse nanoparticles, this process must be largely improved to be suitable for microfluidic preparation, especially because the different organic solvents usually used in this chemistry and the high temperatures needed for the decomposition of the precursors are both incompatible with typical PDMS channels and will require the use of quartz microreactors. And, even if quartz microreactors can be used, the extension of this chemistry in microchannels has to face the problem of the bubbling of the boiling solvents and the elimination of effluents (sometimes toxic) resulting from the decomposition of organometallic precursors. Finally, even if nontoxic precursors can be used, the nanoparticles generated through this process are dispersible in organic solvents, although the main applications of magnetic nanoparticles nowadays require water-soluble particles, for example, in biotechnology. 9.2.3.3 The Polyol Process The polyol process refers to the use of polyols (e.g., ethylene glycol, diethylene glycol, etc.) as solvents for the synthesis of metal or metal oxide nanoparticles. Owing to their high dielectric constants, polyols act as solvents able to dissolve inorganic compounds. They offer a wide range of operating temperature for producing inorganic compounds due to their relatively high boiling points.74 They also play the role of reducing agents, producing the metal particles from the precursor, and of stabilizers, allowing control of the growth of particles and preventing interparticle aggregation.75 In this method, the metal precursor is suspended in a liquid polyol and the solution is heated to a temperature close to its boiling point. This chemical approach has been described for the preparation of well-defined shapes and controlled sizes of oxides nano- and microparticles.76–83 Cai and Wan84 successfully synthesized magnetite nanoparticles in several polyols (ethylene glycol, diethylene glycol, triethylene glycol, and tetraethylene glycol) from Fe(acac)3 at high temperatures (Figure 9.8). The mixture was slowly heated to 180 C and kept at that temperature for 30 min, and then quickly heated to reflux (280 C) and kept at reflux for another 30 min. Joseyphus et al.85 reported the synthesis of Fe nanoparticles in polyols, their magnetic properties, and the influence of the nature of polyols on the formation of Fe nanoparticles, but no precisions either on the temperature gradients or on the final temperatures were given by the authors.

338

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.8 Formation of metal-chelated complexes and their decomposition yielding colloidal transition metal ferrites. Reprinted with permission from Ref. 86. Copyright 2004 the American Chemical Society.

Caruntu et al.86 described a new method for the synthesis of nanocrystalline iron(II,III) oxide, based on the elevated temperature hydrolysis of chelate ion alkoxide complexes in solutions of corresponding alcohol, diethylene glycol (DEG), and N-methyl diethanolamine (NMDEA). Polyol processes seem easy and efficient for the synthesis of iron and iron oxide nanoparticles, but all the publications describe the manipulation and control of temperature and temperature gradients. This control can be facilitated by using microfluidic reactors due to the high surface to volume ratio. But at the same time, if a polyol process has to be transposed in microreactors, one has to take into account several important points. The precursor and the polyol have to be chemically compatible with the reactor materials, which must also accept the high boiling temperatures needed for this procedure as in the thermal decomposition of metal complexes. 9.2.3.4 Synthesis in Constrained Environments Due to the importance of producing magnetic monodisperse nanoparticles, numerous methods have been developed to obtain nanoparticles of more uniform dimensions and well-defined size in constrained environments. These constrained environments include reversed micellar structures of surfactants in nonpolar solvents,87–89 vesicles,90 dendrimers,91 and cyclodextrins,92 and so on. Here, we present a few examples of the synthesis in reverse micelles as they are based on the same idea as the digital microfluidics, opposite to the synthesis in direct micelles, using surfactants for which the counterion is the metallic cation.93 The idea is that the imprisonment of the reactions in small micro/nanoreactors can impose kinetic

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

339

and thermodynamic constraints on particle formation and provide a confinement that limits particle nucleation and growth.90 The first synthesis of magnetic nanoparticles in micelles was reported by Inouye et al.,94 who prepared g-Fe2O3 and Fe3O4 by oxidation of Fe2 þ salts. Recently, Lee et al.95 described the use of the reverse micelles technology for the large-scale synthesis of uniform and highly magnetic nanocrystals. The particle size is tuned by varying the relative proportion of the iron salts, the surfactant, and the solvent. Vidal-Vidal et al. presented a one-pot microemulsion method to produce monodisperse and coated small nanoparticles. The nanoparticles were formed by the coprecipitation reaction of ferrous and ferric salts with two organic bases cyclohexylamine and oleylamine in a water-in-oil microemulsion. As the last example, the synthesis of magnetic nanoparticles inside phospholipidic vesicles was reported in the literature (magnetovesicles). Magnetoliposomes of 25 nm were prepared directly using the phospholipid vesicle encapsulating FeII ions. The slow diffusion of the hydroxide ions inside the vesicles causes the formation of magnetic nanoparticles.90

9.3 MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES g-Fe2O3 superparamagnetic iron oxide nanoparticles have been prepared for the first time by our group in a continuous coaxial-flow microreactor that achieves small diffusion distances and fast mixing times.1 In the same year, Frenz et al.96 reported the use of droplet-based microreactor for the synthesis of g-Fe2O3 nanoparticles. Later on, the initial coaxial-flow setup was improved by separating a nucleation reactor from an aging reactor to synthesize another iron oxide, the antiferromagnetic goethite nanolaths a-FeOOH.97 This section reviews the different methods used for the preparation of these nanoparticles in microfluidic reactors. 9.3.1 Synthesis of g-Fe2O3 Nanoparticles in Microfluidic Reactors 9.3.1.1 Synthesis in Continuous-Flow Microreactor As previously mentioned, there are several processes for the synthesis of magnetic nanoparticles. Among them, only coprecipitation has been transposed in microreactors, certainly because reactions occur in aqueous solution at room temperature. It allows using PDMS microreactors without any sophisticated chemical engineering. The chemical reaction summarizing the synthesis of magnetite nanoparticles is þ þ þ 2Fe3ðaqÞ þ 8OH Fe2ðaqÞ ðaqÞ ! Fe3 O4ðsÞ þ 4H2 OðlÞ

The first trials of the synthesis were run in a typical two-dimensional Y- or T-shaped microreactors, made by lithography in polydimethyl siloxane (PDMS, Sylgard 184) (Figure 9.9). In such microreactors, even when the concentration and the contact times of the different reagents were varied, a magnetic precipitate appears at the interface

340

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.9 Synthesis of magnetic nanoparticles inside a 2D Y-shaped microreactor showing the clogging at the interface. In a typical test, a solution of total iron salts with different concentrations and 0.5 as molar ratio Fe(II)/Fe(III) was injected in one microreactor arm and a solution of the alkaline solution tetramethylammonium hydroxide ((CH3)4NOH, TMAOH) in the other arm.

and clogs the channels, leading to noncontinuous synthesis process. Clogging is probably due to the adsorption of the magnetic nanoparticles, which are in contact with the PDMS walls on the top and the bottom of the 2D channel. To avoid the technical problems of adsorption and clogging, a 3D coaxial-flow microreactor to mix the two coaxial flows of miscible fluids, one containing the iron “precursor salts” and the other one a strong base, has been designed (Figure 9.10). It offers the opportunity to enable a precision positioning of the precursors flow at the center of the channel in both longitudinal and lateral dimensions, and on the other hand, it avoids adsorption of any precipitate species onto the PDMS walls as the latter are totally wetted by the alkaline outer flow.1 The length of the capillary from the confluence region to the outlet was 3 cm. A polytetrafluoroethylene (PTFE) tube (500 mm ID and 10 cm long) leading to a sample

FIGURE 9.10 Coaxial flow device operating under laminar regime. The inset shows the outlet of the inner capillary with the solution of iron(II) and iron(III) flowing into the stream of TMAOH alkaline solution. Reprinted with permission from Ref. 1. Copyright 2008 the Royal Society of Chemistry.

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

341

vial was connected to the reactor outlet. Depending on the two flow rates Qin and Qout, the residence times ranged between 10 and 48 s. The outer capillary with 1.7 mm diameter (d) was shaped by the molding of a cylindrical tubing (Upchurch Scientific) in a Petri dish with PDMS (Sylgard 184) and subsequent removal when the resin is cured. The central capillary with 150 mm ID and 360 mm OD was obtained by fixing a glass capillary (Plymicro, usually used for capillary electrophoresis) inside the tip of a micropipette (Gilson), the conical shape of which enables a precise centering. The iron (II/III) solution of total concentration 102 mol L1 (Fe(II)/Fe(III) ¼ 0.5) was injected into the inner flow with a volumetric rate flow Qin (1 < Qin < 100 mL min1). The alkaline solution of TMAOH of concentration 0.172 mol L1 was injected in the outer flow with a volumetric rate flow Qout (100 < Qout < 400 mL min1). TMAOH was chosen prior to any other base as the TMA þ cations afford enhanced stability of colloidal oxide dispersion.43 9.3.1.2 Flow and Transport Modeling in the Continuous Flow Microreactor Microreactors’ modeling is frequent in chemical engineering to deduce the right hydrodynamic and chemical parameters needed for the chemical synthesis. The objective of the present section is to describe some aspects of the flow behavior in the microreactor that are useful for the synthesis of iron oxide nanoparticles. Indeed, there were a few studies on the flow and mass transport in coflow mixers and they reported fundamentally the analytical aspects of mixing under laminar flow, rather than the influence of mixing on the physicochemical parameters of a reaction. Andreev et al.98 developed a mathematical model to describe the hydrodynamic and mass transfer during an acid–base reaction and showed that the maximum of the mean value product is obtained when the inner flow is considerably higher than the outer flow rate. The same group99 studied again the mixing in a coflow mixer for injection flow analysis and deduced that the mixing time is independent of the total volumetric rate flow Qtot after the point of confluence but depends on the volumetric rate flow ratio of the outer flow to the inner flow a, a ¼ Qout/Qin. Mixing times decreased when a was increased. Confocal laser scanning microscopy compared to the numerical model for pH mapping in the microreactor has been described in detail in our submitted work.100 Here, we review in brief some of the results important for the synthesis of the nanoparticles. Transport Modeling: pH Gradients Studies in bulk (see Section 9.2) have shown that for the synthesis of high-quality magnetic nanoparticles and the good reproducibility of the results, parameters such as the ratio Fe(II)/Fe(III), the pH, the mixing manner, the temperature, and so on need to be controlled. pH is the most important parameter, as it controls the hydrolysis, the polycondensation, and then the precipitation of the iron oxides. As Fe(II) and Fe(III) hydrolysis occurs in two different pH ranges, and as Fe(II) absorption is responsible for the magnetic properties of the magnetite,18 a very fast elevation of the pH (and thus a fast mixing) in the alkaline zone would minimize the formation of nonmagnetic iron(III) oxyhydroxides and increase the yield in magnetic nanoparticles.101

342

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

To understand how different volumetric rate flows affect pH gradients in a microreactor, a theoretical approach based on solving the underlying mass transport for different chemical species in the reactor, coupled with the comparison to experimental images from confocal laser scanning microscopy (CLSM) experiments, was performed. First, as precipitation reactions are complex, pH variations occurring during the mixing in the microreactor described above due to a model reaction that is the neutralization of a strong acid, HCl, by the strong base, TMAOH, were studied. HCl and TMAOH concentrations were the same as those used in the typical chemical synthesis of magnetic nanoparticles, that is, 0.794 and 0.172 mol L1, respectively. The experimental methodology consists in using a micromolar concentration of pH-dependent dye (fluorescein) in the inner stream containing the HCl solution to map the pH changes in the central jet stream where the acid–base reaction proceeds. The experimental results were compared with those of the modeling, the pH distribution in the reactor and the local fluorescein concentration being predicted by modeling the underlying mass transport of the various species in the system. The evolution of the concentration Ci ¼ Ci ðr; tÞ of solute particles i follows the continuity equation qCi þ divðCi u þ ji Þ ¼ si qt where u is the hydrodynamic velocity in the channel, calculated from the Navier–Stokes equation together with the compressibility condition. For a dilute system, it can be identified as the local average velocity of the solvent. si is the creation term and represents the local creation rate of particles due to the chemical reactions. ji is the diffusive flux of species i. The general expression of ji can be obtained from nonequilibrium thermodynamics. For a dilute solution, cross-correlations are negligible and it reduces to the Nernst–Planck expression: ji ¼ Di grad Ci þ

Di Ci Zi eE kB T

The electric field E is chosen to satisfy the Henderson field condition. Consequently, at any time the Xcharge distribution has the time to relax and the local electroneutrality condition Z C ¼ 0 is valid. i i i The various solute species i are H þ (aq), OH(aq), Cl, TMA þ , and fluorescein. Diffusion coefficients were estimated from the values at infinite dilution: DH þ ¼ 9:2  109 m2 s1, DCl ¼ 2:03  109 m2 s1, DOH ¼ 5:28  109 m2 s1, and DTMA þ ¼ 2  109 m2 s1. To simulate the different fluorescein species during the acid–base reaction between HCl and TMAOH, the following assumptions were made: (i) Fluorescein is diluted in the hydrochloric acid, so no need to account for its charge. Convection diffusion equation with no chemical reaction can be solved to calculate the local concentration map of the total fluorescein denoted Flu. (ii) Fluorescein is diluted compared to H þ

343

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

(a) O

HO

OH

O

HO

O

COOH

O

HO

pKa 2 = 6.27

Neutral

Monoanion

Dianion

F0

F-1

F-2

F+1 100

O

COO

H

COO

pKa1 = 4.31

pKa0 = 2.08 Cation

O

OH

H

COOH

H

O

O

OH

OH

(b)

% Light Intensity

80 60 40 20 0 2

4

6

pH

8

10

12

FIGURE 9.11 (a) Chemical structures of fluorescein based on different pH values. Fluorescein is a cation at pH < 2.08, neutral at 2.08 < pH < 4.31, an anion at 4.31 < pH < 6.27, and a dianion at pH > 6.27. (b) Percentage light intensity relative to the value at pH 11 as a function of pH for a solution of 6 mM disodium fluorescein. Copyright 2009 the American Chemical Society.

concentration in the acidified fluorescein solution and cannot interfere during the acid–base reaction. (iii) Different ionic species of the fluorescein have the same diffusion coefficient D 0.2  109 m2 s1. After calculating the local pH in the microreactor, the convection diffusion equation was solved for the fluorescein dye. Since the acid–base reactions are virtually instantaneous, it has been assumed that chemical equilibria are locally achieved for different fluorescein species. Confocal laser scanning microscopy experiments were used to confirm these modeling results. Fluorescein is a well-known pH-sensitive dye that progressively deprotonates when pH increases, as shown in Figure 9.11, producing the dianion F2, which is the only fluorescent species (monoanion F is quite less fluorescent). As the pH rises in aqueous solution, the fluorescence signal increases starting at pH approximately 5 and saturates at a maximum value above pH approximately 7. To verify the pH dependence of the signal, a stock solution of disodium fluorescein was prepared and diluted at 6 mM in different pH-buffered solutions. These solutions were perfused in both the inner and the outer capillaries, and the corresponding fluorescence images were captured with the CLSM at the median plane of the channel (i.e., far from the walls that produce an artifact due to possible adsorption of fluorescein onto PDMS). The images appeared uniform across a 300 mm  300 mm view field. By

344

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.12 (a) CLSM image of the acid–base reaction in the presence of 6 mM fluorescein solution in the inner flow with an applied volumetric rate flow a ¼ 400. The image was recorded in X–Y plane. (b) The three images are X–Z images, taken at the locations indicated by (1), (2), and (3), constructed from a “Z-stack.” The Z-stack consisted of 20 slices, which were spaced at 20 mm intervals. Copyright 2009 the American Chemical Society.

averaging the intensity over this constant area, a calibration plot has been drawn up in Figure 9.11b. The results are in good agreement with cited work.102 Figure 9.12 shows the steady-state 2D fluorescence profile obtained by confocal slicing in the middle of the outlet of the inner capillary where acidified fluorescein solution (6 mM) flows in the center surrounded by the outer alkaline TMAOH solution upon application of a volumetric ratio a ¼ 400. The fluorescence burst in the central jet was interpreted as to have stemmed from the deprotonation of fluorescein by the hydroxide ions diffusing toward the center that remain in excess after reaction with H þ . The hollow cylindrical shape on the fluorescence image (cross sections (1) and (2) of the central stream) illustrates the transition from a still acidic core (below pH 4) of the stream to a neutralized “skin” near the pKa of 6.2 of fluorescein. It is exactly in this boundary region near the pH equivalence that the coprecipitation of the iron salts in a synthesis experiment is expected. To illustrate the good agreement between the CLSM experiment and the model, Figure 9.13 compares the 2D fluorescence intensity map of fluorescein represented by a scale from cold (low intensity, acid) to warm colors (high levels, basic). The symmetry between the experimental part of the image on the left and the calculated one on the right is a clear evidence that the assumptions made to calculate the diffusion of the fluorescein species and the acid–base reaction are acceptable. Finally, the flow rates ratio a was tuned between 40 and 400 and the corresponding fluorescence profiles along the r ¼ 0 axis were compared in Figure 9.14a. The good matching between the predicted and the experimental intensity curves for several values of a clearly validates the proposed simulation method. Fluorescein is just a pH reporter dye, and the shape of the pH curves is in good agreement with the titration of a strong acid (HCl) by a strong base (TMAOH). When a increases, the squeezing effect of the inner stream by the outer stream increases; the inner stream containing the fluorescein dye and HCl is focalized, thus OH diffusion pathways to the H þ ions are reduced. This decrease in the diffusion distance between the reagent species implies a faster mixing and a steeper jump of pH near the equivalence point (pH 7). In view of these results and in order to synthesize magnetic nanoparticles, a ¼ 400 was found to be the best suitable case because it offers the

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

345

FIGURE 9.13 Comparison of the experimental and predicted fluorescence intensities in X–Y plane with (a) the left half-plane representing the experimental fluorescence intensity and (b) the right half-plane with the predicted fluorescence intensity for a ¼ 400. Copyright 2009 the American Chemical Society.

advantage of a fast mixing and a sharp pH jump. The choices a ¼ 80 and 40 would lead to a decrease in the yield in magnetic nanoparticles due to the precipitation of antiferromagnetic iron hydroxides. These results were used for the preparation of the stable colloidal and magnetic nanoparticles reported in our work1 using the coaxialflow microreactor and that will be resumed in the next section. Application for the Synthesis of g-Fe2O3 Nanoparticles For nanoparticle synthesis experiments, the inner solution was a mixture of iron salts with a total ferric and ferrous salt concentration of c ¼ 102 mol L1 and a molar ratio Fe(II)/Fe(III) of 0.5, prepared by mixing FeCl3 and “fresh” FeCl2
4H2O salts in diluted and degassed hydrochloric acid (pH 0.10). The outer flow was an alkaline solution of tetramethylammonium (TMAOH, 0.172 mol L1), which was injected with an outer volumetric rate flow Qout. The reaction was “quenched” by fast solvent extraction (using didodecyl dimethyl ammonium bromide in cyclohexane) to prevent any aging of the nanoparticles in the aqueous solution. The suspensions obtained in cyclohexane were always stable and the nanoparticles produced in the channel were fairly spherical with an average size around 7 nm. The evidence of their crystallinity was provided by the electron microdiffraction pattern in the inset of Figure 9.15, which shows the presence of the maghemite phase g-Fe2O3. Although suspensions obtained in cyclohexane were stable in a zero magnetic field, they sediment in the presence of a magnetic field gradient (e.g., on a strong permanent

346

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.14 (a) Evolution of the experimental fluorescence profiles (&) when a ¼ 40, 80, and 400 along the symmetry axis (r ¼ 0). The lines represent the predicted fluorescence intensities calculated for each case from the simulated F2 concentrations. (b) Simulated pH profiles along the symmetry axis (r ¼ 0) for a ¼ 40, 80, and 400. Copyright 2009 the American Chemical Society.

magnet), which suggests a magnetic character. This observation was confirmed by magnetization measurements (using a vibrating sample magnetometer) on a stable unquenched (aqueous) suspension: the magnetization curve (Figure 9.16) followed the Langevin law typical of superparamagnetism, calculated for an assembly of nanoparticles with a rather narrow distribution of diameters fitted by a log-normal law of parameters d0 ¼ 6 nm and s ¼ 0.2. By measuring both the volume fraction of nanoparticles f ¼ 5.7  105 (from iron titration by atomic spectroscopy) and the saturation magnetization Msat ¼ 7.9 A m1 for the suspension, the specific magnetization of the materials was deduced ms ¼ Msat/f ¼ 1.4  105 A m1, which is much below the bulk value of maghemite g-Fe2O3 (3.5  105 A m1), but not so far from the ms value of about 2.6  105 A m1 usually obtained for nanoparticles of approximately the same sizes prepared with the standard large-scale synthesis. Therefore, it can be deduced that nanoparticles prepared within few seconds in a millifluidic channel exhibit only a small decrease in ordering of their magnetic moments compared to particles obtained within about 30 min in bulk.

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

347

FIGURE 9.15 TEM image of nanoparticles prepared in the channel (for flow rates Qin ¼ 100 mL min1 and Qout ¼ 400 mL min1). The inset shows the electron microdiffraction pattern with the Miller indices of g-Fe2O3. Reprinted with permission from Ref. 1. Copyright 2008 the Royal Society of Chemistry.

FIGURE 9.16 Magnetization curve of a stable suspension in water of nanoparticles produced in the millifluidic device. The inset curves represent the fitting log-normal laws for the number distribution (solid line) and the volume distribution (dotted line) of diameters. Reprinted with permission from Ref. 1. Copyright 2008 the Royal Society of Chemistry.

348

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.17 (a) Pairing module. Two aqueous phases are injected by the outer channels and are synchronously emulsified by the central oil channel. The flow rates are Qo ¼ 800 mL h1 for the oil, and Qx ¼ 400 mL h1, Qy ¼ 100 mL h1 for the aqueous phases. (b) Fusion module. Paired droplets can be coalesced by applying an electrical voltage U between the two electrodes. Qo ¼ 650 mL h1, Qx ¼ 100 mL h1, Qy ¼ 60 mL h1. Reprinted with permission from Ref. 96. Copyright 2008 Wiley.

9.3.2 Synthesis in Microdroplet Reactor The use of a droplet-based microreactor for the synthesis of magnetic nanoparticles by coprecipitation of iron(II) and iron(III) by an alkaline solution of ammonium hydroxide was reported by Frenz et al.96 The microfluidic device consisted of two hydrodynamically coupled nozzles (Figure 9.17). During droplet formation in one of the nozzles, the aqueous stream blocks the oil coming from the central channel, leading to an increased oil flow through the second nozzle. Once the droplet is released, the oil flow switches back to the first channel, allowing droplet pairing at various flow rates. Iron chloride solution was flushed into one arm of the nozzle and ammonium hydroxide into the second arm, which led to droplet pairs containing the two reagents (Figure 9.18). To start a reaction, the droplet pairs can be coalesced by applying an electric field between the two on-chip electrodes. Transmission electron microscopy (TEM) and electron microdiffraction pattern showed that synthesized nanoparticles are monocrystalline and that the phase is g-Fe2O3. The average particle size deduced from TEM images is smaller for the fast compound mixing (4  1 nm) than for bulk mixing (9  3 nm). The superparamagnetic character of the nanoparticles is confirmed by the absence of hysteresis in the magnetization curve. The authors present their methods as a reliable way to produce magnetic nanoparticles. However, this method uses oils and surfactants to achieve the formation of the droplets and their fusion. These “additives” can affect the nucleation and growth mechanisms of the particles. Compared to microdroplet reactors, continuous-flow reactors are easier to handle and are more representative of the conditions of the bulk synthesis, with improved homogeneity, thus offering a better reproducibility of the synthesized particles. The authors defend their use of microdroplets by the

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

349

FIGURE 9.18 Characterization of the iron oxide particles produced. (a) TEM image of the nanoparticles. Inset: HRTEM image of a particle showing (220) spinel planes. (b) Electron diffraction pattern indicating different planes of the spinel structure. (c) Magnetization M/Ms (Ms is the saturation magnetization) as a function of the magnetic field H. Reprinted with permission from Ref. 96. Copyright 2008 Wiley.

enhancement in mixing by convection and the decrease in the reagent dispersion due to the droplets that act as spatially isolated microreactors. This concept is in fact the same as the one evoked for the synthesis of nanoparticles inside vesicles or microemulsions.93–95 The faster mixing time reported in this system is 2 ms, which is far larger than the nucleation time. Moreover, this time is difficult to define as it is totally arbitrary and depends on the concentration of reagents. 9.3.3 Synthesis of a-FeOOH Nanoparticles in Microfluidic Reactors Another interesting iron oxyhydroxide phase is goethite (a-FeOOH), which is widely found in iron-rich soils.35 This clay mineral constitutes the natural ochre pigment, and because of its elongated shape, synthetic goethite is often used as a precursor of a-Fe “hard magnet” particles for magnetic recording.103 Because of this elongated shape, suspensions of antiferromagnetic goethite/plate-like nanostructures (nanolaths) exhibit an original magneto-optical effect and spontaneously self-assemble into a nematic liquid–crystal phase above a threshold concentration.104 The importance of particle shape for the improvement of magnetic properties, or the control of the particle assembly, requires control of the synthetic conditions of these particles.105 The bulk methods reported for the synthesis of acicular (needle-like) goethite particles are based on the aging of ferrihydrite nanoparticles obtained by alkalinization of iron(III) salt solutions.106 They are indeed easily transferable to microfluidic devices as illustrated by Abou-Hassan et al.97

350

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

As discussed earlier, the alkalinization at room temperature of a solution of ferric salts by an alkaline solution leads to the precipitation of an amorphous oxide hydroxide precipitate of ferrihydrite. At its minimum of the solubility, ferrihydrite can evolve to hematite through an internal dehydration process, while in high-solubility domains (very acidic or very alkaline solutions), the transformation via a dissolution–precipitation mechanism is possible and leads to the formation of goethite. Goethite is the most thermodynamically stable phase, but due to the low solubility of iron oxides, the transformation into goethite is very slow, offering the possibility of a good separation between nucleation and growth. Another possible mechanism for the formation of goethite nanoparticles is the oriented attachment of iso-oriented ferrihydrite nanoparticles and then crystallization into goethite nanoparticles.107 A complex microfluidic device (Figure 9.19) was proposed by Abou-Hassan et al.97 in order to physically separate the process of nucleation of the ferrihydrite nanoparticles from their growth, leading to goethite particles. The nucleation of the primary ferrihydrite nanoparticles is induced by diffusive mixing at room temperature in a microreactor that is based on coaxial-flow geometry (R1). This mixing reactor is the same as described for the synthesis of magnetic nanoparticles and is based on a threedimensional coaxial-flow device of two streaming reagents. At the outlet of this micromixer, the suspended ferrihydrite nanoparticles are directly injected into the microtubular aging coil R2, which consists of a transparent PTFE tube of 1.7 mm inner diameter and 150 cm total length continuously heated in a water bath at 60 C. Temperature profiles were calculated to determine the tubing length (and thus the time) required for the fluid to reach a steady state. At the outlet of R1 (before aging) and

FIGURE 9.19 The experimental setup used for the preparation of the ferrihydrite and goethite nanoparticles. TMAOH ¼ tetramethylammonium hydroxide. Reprinted with permission from Ref. 97. Copyright 2009 Wiley.

MICROFLUIDIC SYNTHESIS OF IRON OXYHYDROXIDE NANOPARTICLES

351

FIGURE 9.20 (a) TEM picture of the sample taken after precipitation in the microreactor R1 (before aging), showing ferrihydrite nuclei of 4  1 nm diameter. The selected area diffraction pattern (inset) is typical of two-line ferrihydrite. (b) HRTEM image of an individual ferrihydrite nanoparticle with a 2.5 A lattice fringe. Reprinted with permission from Ref. 97. Copyright 2009 Wiley.

R2 (after aging), the resulting suspension is collected and analyzed by TEM and by high-resolution TEM (HRTEM). TEM pictures of the particles obtained after R1 show well-defined spherical ferrihydrite nanoparticles (nanodots) about 4  1 nm in size (Figure 9.20a). HRTEM measurements (Figure 9.20b) show that the nanoparticles are monocrystal line, exhibiting atomic planes with an interplanar distance of about 2.5 A, which is consistent with ferrihydrite nanoparticles. Under the given flow rate, the ferrihydrite solution reaches 60 C in about 1 s, that is, within the first centimeter after it has entered the heated zone of the tubing. The effective residence time is about 15 min, as estimated from the length of the tubing along which the fluid has reached the stationary temperature of 60 C. After aging for 15 min under continuous flow in the aging coil R2, goethite platelike nanostructures were observed with an average length L ¼ 30  17 nm and width w ¼ 7  4 nm (Figure 9.21a). This short aging time appeared to be sufficient for the growth of crystalline and anisotropic goethite nanoparticles that differ only in smaller sizes compared to those obtained after complete aging (1 day at 60 C).108 Moreover, the presence of few remaining ferrihydrite nuclei undergoing aggregation in the batch after 15 min and even after 24 h at 60 C (data not shown) supports the idea that goethite nanoparticles were formed by the aggregation mechanism rather than by dissolution/reprecipitation. Thus, the use of microfluidic device allows to significantly accelerate the synthesis of goethite nanoparticles from ferrihydrite nuclei. The novelty of this approach lies in the separation of the nucleation of the primary particles (ferrihydrite) and the growth of the goethite nanoparticles in two independent microreactors operating in different conditions. In the nucleation microreactor, the streaming reagents are mixed by molecular diffusion at room temperature in a flow-focusing geometry. The homogeneity of the mixture is ensured by the fast mixing time and the technical difficulty of microchannel clogging owing to the fact that precipitation onto the walls is avoided

352

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

FIGURE 9.21 (a) TEM image of the nanolaths after aging for 15 min in the microtubular loop R2, produced at pH 13 and under laminar flow. (b) HRTEM image of a nanorod particle. Lattice fringe spacing is consistent with goethite. The dashed lines serve to highlight the morphology and texture of the particle. Reprinted with permission from Ref. 97. Copyright 2009 Wiley.

by 3D geometry. The use of a microfluidic device for aging, by minimizing local temperature gradients, ensures a regular laminar flow, and finally leads to crystalline plate-like nanostructures. These particles have approximately the same values of aspect ratio and polydispersity index as those obtained in bulk synthesis (bulk synthesis usually yields goethite nanoparticles with a typical length of about 250 nm and a width of 40 nm, with a polydispersity index of about 50% for both dimensions109) but are smaller in size. The time required for aging decreases to 15 min (for a velocity of 0.1 cm s1) compared to bulk synthesis (several hours or days). This may originate from the small diameter of the aging reactor, causing a shear stress that prealigns the primary ferrihydrite nanoparticles and speeds up their oriented aggregation process. 9.4 PERSPECTIVES Among all the ferric oxide nanoparticles, superparamagnetic ones (superparamagnetic iron oxide nanoparticles, SPIONs) are of special interest because of their applications in the field of imagery and therapy. At present, the screening of the relation properties/structure is neither very easy nor economic in bulk, and continuous microfluidic systems, enabling to add reagents along the entire length of the channel, can be a very useful tool for screening the different parameters (size, surface functionalization, and aggregation) allowing to optimize given properties. This idea is illustrated in Figure 9.22 and can be summarized by three operations: adding, mixing, and reacting. Indeed, if several microreactors Ri are associated to form a series of microunit operations, the synthesis of SPION nanoparticles and their surface modification would

PERSPECTIVES

FIGURE 9.22

353

Cartoon illustrating the idea of online synthesis of functional nanomaterials.

be possible in an unique online process. The surface coating of the SPIONs by silica is a good example. Silica has been extensively exploited as a coating material for magnetic nanoparticles110–112 in order to get a protective, biocompatible, inert, and hydrophilic surface with excellent anchoring points for derivatizing molecules.113 Several methods have been reported in the literature for the formation of superparamagnetic iron oxide silica nanocomposites including reactions performed under St€ober conditions,114,115 microemulsions,116 emulsions,117 and aerosol pyrolysis.118 It seems that core–shell SPION silica nanoparticles and SPION luminescent silica nanoparticles can be obtained in microfluidic reactors without the use of any surfactant (Abou-Hassan, unpublished results). In the field of on-chip magnetic separation, there were many works devoted to the separation of micrometric magnetic particles, but to date no separation of nanometric particles on the microfluidic scale has been reported.119 A continuous-flow method capable of both separating magnetic from nonmagnetic particles and separating different magnetic particles from each other can be very helpful in synthetic chemistry. But the most important challenge in the field of nanomaterials’ synthesis in microfluidic lies in developing online characterization methods. For quantum dots or metallic nanoparticles, optical characterizations allowing establishment of a simple relation with particle size are available. That is not the case for ferric oxide nanoparticles. As the latter have magnetic properties, online magnetic measurements can perhaps be designed. On the same kind of idea, but of course generally for any kind of materials, online characterizations using small-angle X-ray or neutron scattering have to be developed. These are thus important in designing online nanoparticle characterization techniques.

354

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

REFERENCES 1. Abou-Hassan, A.; Sandre, O.; Cabuil, V.; Tabeling, P. Synthesis of iron oxide nanoparticles in a microfluidic device: preliminary results in a coaxial flow millichannel. Chem Commun. 2008, 1783. 2. http://marsrover.nasa.gov/. 3. Cornell, R.M.; Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses, Wiley-VCH: 2003. 4. Dormann, J.L.; Firoani, D. Magnetic Properties of Fine Particles, Amsterdam: NorthHolland, 1992. 5. Eisenmenger, J.; Schuller, I.K. Magnetic nanostructures: overcoming thermal fluctuations. Nat Mater. 2003, 2, 437. 6. Ferrucci, J.T.; Stark, D.D. Iron oxide-enhanced MR imaging of the liver and spleen: review of the first 5 years. Am J. Roentgenol. 1990, 155, 943. 7. Sun, S.; Murray, C.B.; Weller, D.; Folks, L.; Moser, A. Monodisperse FePt nanoparticles and ferromagnetic FePt nanocrystal superlattices. Science 2000, 287, 1989. 8. Miller, M.M.; Prinz, G.A.; Cheng, S.F.; Bounnak, S. Detection of a micron-sized magnetic sphere using a ring-shaped anisotropic magnetoresistance-based sensor: a model for a magnetoresistance-based biosensor. Appl Phys. Lett. 2002, 81, 2211. 9. Chourpa, I.; Douziech-Eyrolles, L.; Ngaboni-Okassa, L.; Fouquenet, J.F.; CohenJonathan, S.; Souce, M.; Marchais, H.; Dubois, P. Molecular composition of iron oxide nanoparticles, precursors for magnetic drug targeting, as characterized by confocal Raman microspectroscopy. Analyst 2005, 130, 1395. 10. Martina, M.S.; Fortin, J.P.; Menager, C.;Clement, O.; Barratt,G.; Grabielle-Madelmont, C.; Gazeau, F.; Cabuil, V.; Lesieur, S. Generation of superparamagnetic liposomes revealed as highly efficient MRI contrast agents forin vivo imaging. J Am. Chem. Soc. 2005, 127, 10676. 11. Massart, R.;French Patent 79-188-42, 1979. 12. Massart, R. Preparation of aqueous magnetic liquids in alkaline and acidic media. IEEE Trans. Magn. 1981, 17, 1247. 13. Massart, R.; Cabuil, V. Synthese en milieu alcalin de magn
etite collo€ıdale: controˆle du rendement et de la taille des particules. J. Chim. Phys. 1987, 7, 84. 14. Massart, R.; Dubois, E.; Cabuil, V.; Hasmonay, E. Preparation and properties of monodisperse magnetic fluids. J Magn. Magn. Mater. 1995, 149, 1. 15. Jolivet, J.-P.; Chaneac, C.; Tronc, E. Iron oxide chemistry. From molecular clusters to extended solid networks. Chem. Commun. 2004, 481. 16. Nielsen, A.E.; Kinetics of Precipitation, Oxford: 1964. 17. Baes, C.F.; Mesmer, R.E. The Hydrolysis of Cations, Wiley: New York, 1976. 18. Jolivet, J.P. Metal Oxide Chemistry and Synthesis. From Solution to Solid State, Wiley: Chichester, 2000. 19. Jolivet, J.P.; De la solution  a l’oxyde, EDP Sciences: 1998. 20. Kashciev, D. Nucleation Basic Theory with Applications; Butterworth-Heinemann; 2000. 21. LaMer, V.K.; Dinegar, R.H. Theory, production and mechanism of formation of monodispersed hydrosols. J Am. Chem. Soc 1950, 72, 4847. 22. KerKer, M.; LaMer, V.K. Particle size distribution in sulfur hydrosols by polarimetric analysis of scattered light. J. Am. Chem. Soc 1950, 72, 3516.

REFERENCES

355

23. deMello, A.J.; deMello, J.C. Microscale reactors: nanoscale products. Lab Chip 2004, 4, 11N. 24. Zhang, Q.; Liu, S.-J.; Yu, S.-H. Recent advances in oriented attachment growth and synthesis of functional materials: concept, evidence, mechanism, and future. J Mater. Chem. 2009, 19, 191. 25. Barker, A.J.; Cage, B.; Russek, S.; Stoldt, C. Ripening during magnetite nanoparticle synthesis: resulting interfacial defects and magnetic properties. J Appl. Phys. 2005, 98, 063528. 26. Sugimoto, T. Preparation of monodispersed colloidal particles. Adv Colloid Interface Sci. 1987, 28, 65. 27. Kabalnov, A.S.; Shchukin, E.D. Ostwald ripening theory: applications to fluorocarbon emulsion stability. Adv Colloid Interface Sci. 1992, 38, 69. 28. Talapin, D.V.; Rogach, A.L.; Haase, M.; Weller, H. Evolution of an ensemble of nanoparticles in a colloidal solution: theoretical study. J Phys. Chem. B. 2001, 105, 12278. 29. Borg, R.; Dienes, G.; The Physical Chemistry of Solids, Academic Press: Boston, MA, 1992. 30. C€ olfen, H.; Mann, S. Higher-order organization by mesoscale self-assembly and transformation of hybrid nanostructures. Angew. Chem., Int. Ed. 2003, 42, 2350. 31. Gilbert, B.; Zhang, H.; Huang, F.; Finnegan, M.; Waychunas, G.; Banfield, J. Special phase transformation and crystal growth pathways observed in nanoparticles. Geochem Trans. 2003, 4, 20. 32. Penn, R.L.; Banfield, J.F. Morphology development and crystal growth in nanocrystalline aggregates under hydrothermal conditions: insights from titania. Geochem Cosmochim. Acta 1999, 63, 1549. 33. Penn, R.L.; Banfield, J.F. Oriented attachment and growth, twinning, polytypism, and formationofmetastable phase:insightsfromnanocrystallineTiO2.AmMiner.1998,83,1077. 34. Oskam, G.; Hu, Z.; Penn, R.L.; Pesika, N.; Searson, P.C. Coarsening of metal oxide nanoparticles. Phys Rev. E 2002, 66, 11403. 35. Banfield, J.F.; Welch, S.A.; Zhang, H.; Ebert, T.T.; Penn, R.L. Aggregation-based crystal growth and microstructure development in natural iron oxyhydroxide biomineralization products. Science 2000, 289, 751. 36. Watzky, M.A.; Finke, R.G. Transition metal nanocluster formation kinetic and mechanistic studies. A new mechanism when hydrogen is the reductant: slow, continuous nucleation and fast autocatalytic surface growth. J. Am. Chem. Soc. 1997, 119, 10382. 37. Turkevich, J.; Stevenson, P.C.; Hillier, J. A study of the nucleation and growth processes in the synthesis of colloidal gold. Discuss Faraday Soc. 1951, 11, 55. 38. Matijevic, E.Preparation andpropertiesofuniform sizecolloids.Chem Mater. 1993,5,412. 39. Overbeek, J.Th.G. Monodisperse colloidal systems, fascinating and useful. Adv Colloid Interface Sci. 1982, 15, 251. 40. Salmon, J.-B.; Dubrocq, C.; Tabeling, P.; Charier, S.; Alcor, D.; Jullien, L.; Ferrage, F. An approach to extract rate constants from reaction–diffusion dynamics in a microchannel. Anal Chem. 2005, 77, 3417. 41. Sounart, T.L.; Safier, P.A.; Voigt, J.A.; Hoyt, J.; Tallant, D.R.; Matzke, C.M.; Michalske, T.A. Spatially-resolved analysis of nanoparticle nucleation and growth in a microfluidic reactor. Lab Chip 2007, 7, 908.

356

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

42. Knight, J.B.; Vishwanath, A.; Brody, J.P.; Austin, R.H. Hydrodynamic focusing on a silicon chip: mixing nanoliters in microseconds. Phys Rev. Lett. 1998, 80, 3863. 43. Cabuil, V.; Massart, R.; Bacri, J.C.; Perzynski, R.; Salin, D. Ionic ferrofluids: towards fractional distillation. J Chem. Res. 1987, S, 130. 44. Fauconnier, N.; Pons, J.N.; Roger, J.; Bee, A. Thiolation of maghemite nanoparticles by dimercaptosuccinic acid. J Colloid Interface Sci. 1997, 194, 427. 45. Bee, A.; Massart, R.; Neveu, S. Synthesis of very fine maghemite particles. J Magn. Magn. Mater. 1995, 149, 6. 46. Neveu-Prin, S.; Cabuil, V.; Massart, R.; Escaffre, P.; Dussaud, J. Encapsulation of magnetic fluids. J Magn. Magn. Mater. 1993, 122, 42. 47. Biddlecombe, G.B.; Gun’ko, Y.K.; Kelly, J.M.; Pillai, S.C.; Coey, J.M.D.; Venkatesan, M.; Douvalis, A.P. Preparation of magnetic nanoparticles and their assemblies using a new Fe(II) alkoxide precursor. J Mater. Chem. 2001, 11, 2937. 48. Wan, J.; Cai, W.; Feng, J.; Meng, X.; Liu, E. In situ decoration of carbon nanotubes with nearly monodisperse magnetite nanoparticles in liquid polyols. J Mater. Chem. 2007, 17, 1188. 49. Kimata, M.; Nakagawa, D.; Hasegawa, M. Preparation of monodisperse magnetic particles by hydrolysis of iron alkoxide. Powder Technol. 2003, 132, 112. 50. Chin, A.B.; Yaacob, I.I. Synthesis and characterization of magnetic iron oxide nanoparticles via w/o microemulsion and Massart’s procedure. J Mater. Process. Technol. 2007, 191, 235. 51. Salazar-Alvarez, G.; Muhammed, M.; Zagorodni, A.A. Novel flow injection synthesis of iron oxide nanoparticles with narrow size distribution. Chem Eng. Sci. 2006, 61, 4625. 52. Basak, S.; Chen, D.R.; Biswas, P. Electrospray of ionic precursor solutions to synthesize iron oxide nanoparticles: modified scaling law. Chem Eng. Sci. 2007, 62, 1263. 53. Bell, A.; Matijevic, E. Growth mechanism of hydrous chromium(III) oxide spherical particles of narrow size distribution. J Phys. Chem. 1974, 78, 2621. 54. Jolivet, J.P.; Froidefond, C.; Pottier, A.; Chan
eac, C.; Cassaignon, S.; Tronc, E.; Euzen, P. Size tailoring of oxide nanoparticles by precipitation in aqueous medium. A semiquantitative modelling. J. Mater. Chem. 2004, 14, 3281. 55. Vayssieres, L.; Chan
eac, C.; Tronc, E.; Jolivet, J.P. Size tailoring of magnetite particles formed by aqueous precipitation: an example of thermodynamic stability of nanometric oxide particles. J Colloid Interface Sci. 1998, 205, 205. 56. Jolivet, J.P.; Belleville, P.; Tronc, E.; Livage, J. Influence of Fe(II) on the formation of the spinel iron oxide in alkaline medium. Clays Clay Miner. 1992, 40, 531. 57. Jolivet, J.P.; Vayssieres, L.; Chan
eac, C.; Tronc, E. Synthesis of iron oxide-based magnetic nanomaterials and composites. Mater Res. Symp. Proc. 1997, 432, 145. 58. Jolivet, J.P.; Chan
eac, C.; Tronc, E. Synthesis of iron oxide-based magnetic nanomaterials and composites. Comp. Rend. Series 2002, 5, 659. 59. Babes, L.; Denizot, B.; Tanguy, G.; Le Jeune, J.J.; Jallet, P. Synthesis of iron oxide nanoparticles used as MRI contrast agents: a parametric study. J. Colloid Interface. Sci. 1999, 212, 474. 60. Fauconnier, N.; Bee, A.; Roger, J.; Pons, J.N. Adsorption of gluconic and citric acids on maghemite particles in aqueous medium. Prog Colloid Polym. Sci. 1996, 100, 212.

REFERENCES

357

61. Fauconnier, N.; Bee, A.; Roger, J.; Pons, J.N. Synthesis of aqueous magnetic liquids by surface complexation of maghemite nanoparticles. J Mol. Liq. 1999, 83, 233. 62. Roger, J.; Pons, J.N.; Massart, R.; Halbreich, A.; Bacri, J.C. Some biomedical applications of ferrofluids. Eur Phys. J. Appl. Phys. 1999, 5, 321. 63. Denizot, B.; Tanguy, G.; Hindre, F.; Rump, E.; LeJeune, J.J.; Jallet, P. Phosphorylcholine coating of iron oxide nanoparticles. J Colloid Interface Sci. 1999, 209, 66. 64. Fauconnier, N.; B
ee, A.; Roger, J.; Pons, J.N. Synthesis of aqueous magnetic liquids by surface complexation of maghemite nanoparticles. J Mol. Liq. 1999, 83, 233. 65. Liu, C.; Huang, P.M. Atomic force microscopy and surface characteristics of iron oxides formed in citrate solutions. Soil Sci Soc. Am. J. 1999, 63, 65. 66. Tremaine, P.R.; LeBlanc, J.C. The solubility of magnetite and the hydrolysis and oxidation of Fe2 þ in water to 300 C. J Sol. Chem. 1980, 9, 415. 67. Sweeton, F.H.; Baes, C.F. The solubility of magnetite and hydrolysis of ferrous ion in aqueous solutions at elevated temperatures. J Chem. Thermodyn. 1970, 2, 479. 68. Hyeon, T.; Lee, S.S.; Park, J.; Chung, Y.; Na, H.B. Synthesis of highly crystalline and monodisperse maghemite nanocrystallites without a size-selection process. J Am. Chem. Soc. 2001, 123, 12798. 69. Sun, S.; Zeng, H. Size-controlled synthesis of magnetite nanoparticles. J Am. Chem. Soc. 2002, 124, 8204. 70. Jana, N.R.; Chen, Y.; Peng, X. Size- and shape-controlled magnetic (Cr, Mn, Fe, Co, Ni) oxide nanocrystals via a simple and general approach. Chem. Mater. 2004, 16, 3931. 71. Park, J.; An, K.; Hwang, Y.; Park, J.-G.; Noh, H.-J.; Kim, J.-Y.; Park, J.-H.; Hwang, N.-M.; Hyeon, T. Ultra-large-scale syntheses of monodisperse nanocrystals. Nat Mater. 2004, 3, 891. 72. Tartaj, P.; Serna, C.J. Synthesis of monodisperse superparamagnetic Fe/silica nanospherical composites. J Am. Chem. Soc. 2003, 125, 15754. 73. Sato, S.; Murakata, T.; Yanagi, H.; Miyasaka, F.; Iwaya, S. Hydrothermal synthesis of fine perovskite PbTiO3 powders with a simple mode of size distribution. J Mater. Sci. 1994, 29, 5657. 74. J
ez
equel, D.; Guenot, J.; Jouini, N.; Fie
evet, F. Submicrometer zinc oxide particles: elaboration in polyol medium and morphological characteristics. J. Mater. Res. 1995, 10, 77. 75. Feldmann, C.; Jungk, H.-O. Polyol-mediated preparation of nanoscale oxide particles. Angew. Chem., Int. Ed. 2001, 40, 359. 76. Tzitzios, V.K.; Petridis, D.; Zafiropoulou, I.; Hadjipanayis, G.; Niarchos, D. Synthesis and characterization of L10 FePt nanoparticles from Pt–Fe3O4 core–shell nanoparticles. J Magn. Magn. Mater. 2005, 294, e95. 77. Chow, G.M.; Kurihara, L.K.; Kemner, K.M.; Schoen, P.E.; Elam, W.T.; Ervin, A.; Keller, S.; Zhang, Y.D.; Budnick, J.; Ambrose, T. Structural, morphological, and magnetic study of nanocrystalline cobalt–copper powders synthesized by the polyol process. J Mater. Res. 1995, 10, 1546. 78. Viau, G.; Ravel, F.; Acher, O.; Fivet-Vincent, F.; Fivet, F. Preparation and microwave characterization of spherical and monodisperse Co20Ni80 particles. J. Appl. Phys. 1994, 76, 6570.

358

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

79. Viau, G.; Ravel, F.; Acher, O.; Fivet-Vincent, F.; Fivet, F. Preparation and microwave characterization of spherical and monodisperse Co–Ni particles. J Magn. Magn. Mater. 1995, 140, 377. 80. Viau, G.; Fivet-Vincent, F.; Fivet, F. Nucleation and growth of bimetallic CoNi and FeNi monodisperse particles prepared in polyols. Solid State Ionics 1996, 84, 259. 81. Mercier, D.; Lvy, J.C.S.; Viau, G.; Fivet-Vincent, F.; Fivet, F.; Toneguzzo, P.; Acher, O. Magnetic resonance in spherical Co–Ni and Fe–Co–Ni particles. Phys. Rev. B: Condens. Matter 2000, 62, 532. 82. Feldmann, C. Preparation of nanoscale pigment particles. Adv Mater. 2001, 13, 1301. 83. Giri, A.K.; Chowdary, K.M.; Majetich, S.A. AC magnetic properties of compacted FeCo nanocomposites. Mater. Phys. Mech. 2000, 1, 1. 84. Cai, W.; Wan, J. Facile synthesis of superparamagnetic magnetite nanoparticles in liquid polyols. J Colloid Interface Sci. 2007, 305, 366. 85. Joseyphus, R.J.; Kodama, D.; Matsumoto, T.; Sato, Y.; Jeyadevan, B.; Tohji, K. Role of polyol in the synthesis of Fe particles. J. Magn. Magn. Mater. 2007, 310, 2393. 86. Caruntu, D.; Caruntu, G.; Chen, Y.; O’Connor, C.J.; Goloverda, G.; Kolesnichenko, V.L. Synthesis of variable-sized nanocrystals of Fe3O4 with high surface reactivity. Chem Mater. 2004, 16, 5527. 87. Dresco, P.A.; Zaitsev, V.S.; Gambino, R.J.; Chu, B. Preparation and properties of magnetite and polymer magnetite nanoparticles. Langmuir 1999, 15, 1945. 88. Santra, S.; Tapec, R.; Theodoropoulou, N.; Dobson, J.; Hebard, A.; Tan, W. Synthesis and characterization of silica-coated iron oxide nanoparticles in microemulsion: the effect of nonionic surfactants. Langmuir 2001, 17, 2900. 89. Meldrum, F.C.; Heywood, B.R.; Mann, S. Magnetoferritin: in vitro synthesis of a novel magnetic protein. Science 1992, 257, 522. 90. Sangregorio, C.; Wieman, J.K.; Connor, C.O’.; Rosenzweig, Z. A new method for the synthesis of magnetoliposomes. J Appl. Phys. 1999, 85, 5699. 91. Strable, E.; Bulte, J.W.M.; Moskowitz, B.; Vivekanandan, K.; Allen, M.; Douglas, T. Synthesis and characterization of soluble iron oxide–dendrimer composites. Chem Mater. 2001, 13, 2201. 92. Bonacchi, D.; Caneschi, A.; Dorignac, D.; Falqui, A.; Gatteschi, D.; Rovai, D.; Sangregorio, C.; Sessoli, R. Nanosized iron oxide particles entrapped in pseudo-single crystals of g-cyclodextrin. Chem Mater. 2004, 16, 2016. 93. Pileni, M.P.; Duxin, N. Micelle technology for magnetic nanosized alloys and composites. Chem Innov. 2000, 30, 25. 94. Inouye, K.; Endo, R.; Otsuka, Y.; Miyashiro, K.; Kaneko, K.; Ishikawa, T. Oxygenation of ferrous ions inreversed micelle and reversedmicroemulsion. JPhys. Chem. 1982, 86,1465. 95. Lee, Y.; Lee, J.; Bae, C.J.; Park, J.G.; Noh, H.J.; Park, J.H.; Hyeron, J.H. Large-scale synthesis of uniform and crystalline magnetite nanoparticles using reverse micelles as nanoreactors under reflux conditions. Adv Funct. Mater. 2005, 3, 503. 96. Frenz, L.; El-Harrak, A.; Pauly, M.; B
egin-Colin, S.; Griffiths, A.D.; Baret, J.-C. Dropletbased microreactors for the synthesis of magnetic iron oxide nanoparticles. Angew. Chem., Int. Ed. 2008, 47, 6817. 97. Abou-Hassan, A.; Sandre, O.; Neveu, S.; Cabuil, V. Synthesis of goethite by separation of the nucleation and growth processes of ferrihydrite nanoparticles using microfluidics. Angew. Chem., Int. Ed. 2009, 48, 2342.

REFERENCES

359

98. Andreev, V.P.; Koleshko, S.B.; Holman, D.A.; Scampavia, L.D.; Christian, G.D. Hydrodynamics and mass transfer of the coaxial jet mixer in flow injection analysis. Anal Chem. 1999, 71, 2199. 99. Scampavia, L.D.; Blankenstein, G.; Ruzicka, J.; Christian, G.D. A coaxial jet mixer for rapid kinetic analysis in flow injection and flow injection cytometry. Anal Chem. 1995, 67, 2743. 100. Abou-Hassan, A.; Dufr^eche, J.F.; Sandre, O.; M
eriguet, G.; Bernard, O.; Cabuil, V. Confocal laser scanning microscopy for pH mapping in a coaxial flow microreactor: application in the synthesis of superparamagnetic nanoparticles. J. Phys. Chem. C 2009, 113, 18097. 101. Kim, D.K.; Mikhaylova, M.; Zhang, Y.; Muhammed, M. Protective coating of superparamagnetic iron oxide nanoparticles. Chem Mater. 2003, 15, 1617. 102. Margulies, D.; Melman, G.; Shanzer, A. Fluorescein as a model molecular calculator with reset capability. Nat Mater. 2005, 4, 768. 103. Nunez, N.O.; Morales, M.P.; Tartaj, P.; Serna, C.J. Preparation of high acicular and uniform goethite particles by a modified-carbonate route. J Mater. Chem. 2000, 10, 2561. 104. Vroege, G.J.; Thies-Weesie, D.M.E.; Petukhov, A.V.; Lemaire, B.J.; Davidson, P. Smectic liquid-crystalline order in suspensions of highly polydisperse goethite nanorods. Adv. Mater. 2006, 18, 2565. 105. Lemaire, B.J.; Davidson, P.; Ferr
e, J.; Jamet, J.P.; Petermann, D.; Panine, P.; Dozov, I.; Jolivet, J.P. Physical properties of aqueous suspensions of goethite (a-FeOOH) nanorods. Part I. In the isotropic phase. Eur. Phys. J. E 2004, 13, 291. 106. Schwertmann, U.; Murad, E. Effect of pH on the formation of goethite and hematite from ferrihydrite. Clays Clay Miner. 1983, 31, 277. 107. Penn, R.L.; Banfield, J.F. Imperfect oriented attachment: dislocation generation in defectfree nanocrystals. Science 1998, 279, 1519. 108. Krehula, S.; Popovic, S.; Music, S. Synthesis of acicular a-FeOOH particles at a very high pH. Mater Lett. 2002, 54, 108. 109. Thies-Weesie, D.M.E.; de Hoog, J.P.; Hernandez Mendiola, M.H.; Petukhov, A.V.; Vroege, G.J. Synthesis of goethite as a model colloid for mineral liquid crystals. Chem Mater. 2007, 19, 5538. 110. Stjerndahl, M.; Andersson, M.; Hall, H.E.; Pajerowski, D.M.; Meisel, M.W.; Duran, R.S. Superparamagnetic Fe3O4/SiO2 nanocomposites: enabling the tuning of both the iron oxide load and the size of the nanoparticles. Langmuir 2008, 24, 3532. 111. Alcala, M.D.; Real, C. Synthesis based on the wet impregnation method and characterization of iron and iron oxide–silica nanocomposites. Solid State Ionics 2006, 177, 955. 112. Gushikem, Y.; Rosatto, S.S. Metal oxide thin films grafted on silica gel surfaces: recent advances on the analytical application of these materials. J Braz. Chem. Soc. 2001, 12, 695. 113. Sharma, P.; Brown, S.; Walter, G.; Santra, S.; Moudgil, B. Nanoparticles for bioimaging. Adv Colloid Interface Sci. 2006, 123, 471. 114. Barnakov, Y.A.; Yu, M.H.; Rosenzweig, Z. Manipulation of the magnetic properties of magnetite–silica nanocomposite materials by controlled St€ ober synthesis. Langmuir 2005, 21, 7524. 115. Deng, Y.H.; Wang, C.C.; Shen, X.Z.; Yang, W.L.; An, L.; Gao, H.; Fu, S.K. Preparation, characterization, and application of multistimuli-responsive microspheres with fluorescence-labeled magnetic cores and thermoresponsive shells. Chem Eur. J. 2005, 11, 6006.

360

MICROFLUIDIC SYNTHESIS OF IRON OXIDE AND OXYHYDROXIDE NANOPARTICLES

116. Lu, A.H.; Salabas, E.L.; Schueth, F. Magnetic nanoparticles: synthesis, protection, functionalization, and application. Angew. Chem., Int. Ed. 2007, 46, 1222. 117. Xu, Z.Z.; Wang, C.C.; Yang, W.L.; Fu, S.K. Synthesis of superparamagnetic Fe3O4/SiO2 composite particles via sol–gel process based on inverse miniemulsion. J. Mater. Sci. 2005, 40, 4667. 118. Tartaj, P.; Gonzalez-Carreno, T.; Bomati-Miguel, O.; Serna, C.J.; Bonville, P. Magnetic behavior of superparamagnetic Fe nanocrystals confined inside submicron-sized spherical silica particles. Phys Rev. B: Condens. Matter 2004, 69, 094401. 119. Pamme, N. Magnetism and microfluidics. Lab Chip 2006, 6, 24.

10 METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS PETER MIKE G€ uNTHER, ANDREA KNAUER, AND JOHANN MICHAEL K€ oHLER Department of Physical Chemistry and Microreaction Technology, Institute of Micro- and Nanotechnologies, Ilmenau University of Technology, Ilmenau, Germany

10.1 INTRODUCTION Metal nanoparticles are the oldest nanomaterials used by mankind for achieving specific functional material properties. Colloidal gold was generated in ancient glass matrix for coloring vine glasses or windows in middle age churches. Therefore, the tendency of gold to form homogeneous distributed nanoparticles under moderate reducing conditions and the efficient specific plasmon absorption were used. The optical properties of a nanocomposite material were the first reason for the synthesis and application of metal nanoparticles. Liquid colloidal solutions of noble metals such as gold were also known since the middle age. In the nineteenth century, Michael Faraday discovered basic laws of electrochemistry, investigated colloidal solutions of gold, and explained their nature based on the fundamental of understanding redox processes and charged particles. The particles inside the colloidal solutions were stabilized by charges that led to a high thermodynamic stability. For this reason, colloidal solutions were stable over decades and centuries not only in a rigid inorganic glass matrix but also inside a highly movable liquid. Metal nanoparticles are characterized by high surface to volume ratios, special chemical activities of surface sites, and as possessing different shapes and sizes like nanoparticles of other material classes. The particular interest in metal nanoparticles is due to the delocalization of electrons in metals. In contrast to dielectric solids, metallic Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

361

362

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

solids can be described as a system consisting of cations fixed in a crystal lattice and a cloud of highly movable electrons compensating the positive charges of the multicationic lattice and mediating the bonds inside the whole system. The typical metallic properties such as metallic shine and reflectivity, high thermal and electrical conductivity, and high refractive index are caused by the long-range delocalization of outer electrons. The particular properties of electrons, their high movability, and energy states make metallic nanoparticle completely different from nanoparticles of other materials. The small dimensions of nanoparticles restrict the delocalization range and the number of interacting electrons. This leads to a significant change of electronic and optical properties if the size of a metallic solid is shrinked down to the submicron and further down to the nanometer scale. Separated metal nanoparticles are charge confinements. Electron density exchange between metal nanoparticles depends on barriers at the particle surface. Thin adsorbate films, monomolecular capping layers, or a narrow vacuum gap can act as barriers that can be passed by electrons using the tunneling effects. Metal nanoparticles inside a chain, a matrix, or a network of such tunnel contacts act as chargeable islands in a charge transfer system controlled by Coulomb blockade effects. Thicker insulating surface films or the embedding of metal nanoparticles can suppress the charge transport between particles completely. The cloud of movable electrons inside the metal nanoparticles can be polarized by electric fields. This polarization occurs by static fields as well as by alternating fields of lower and higher frequencies. A temporal or oscillating polarization with very high frequency is also caused by interaction of metal nanoparticles with the photons of electromagnetic fields. The restricted number of atoms and movable electrons inside a metal nanoparticle lead to the formation of distinct energy levels in contrast to the energetic continuum represented by the band structure of a bulky conductive solid. The number of distinct energy levels decreases with decreasing number of atoms and electrons inside the particle. The specific interaction of nanoparticle with photons and the distinct energy levels for electron states are responsible for the specific optical properties of metal nanoparticles and their dependence on elementary composition, size, and shape. The strong dependence of electronic and optical properties on the particle parameters is an important reason for all recent demands for methods of producing particle populations of very high homogeneity. This demand is valid for nanoparticles of pure metals, but still more important for composite nanoparticles, where the spatial distribution of components represents an additional parameter with high impact on the electronic behavior of the particles and their interaction with the electromagnetic field. The demand for high-quality metal nanoparticles with narrow size and shape distribution as well as homogeneous composition can be satisfied by highly controlled preparation techniques. Continuous-flow processes are particularly suited for realizing constant mixing, reaction, and quenching conditions. The introduction of microreaction technology gives the possibility for realizing mixing, thermal activation, or cooling down within shortest time intervals. Consequently, microcontinuous-flow processes are of particular interest for the synthesis of metal nanoparticles in research

MECHANISM OF METAL NANOPARTICLE FORMATION

363

and development and for using this technique for production of high-quality metallic nanomaterials, too.

10.2 MECHANISM OF METAL NANOPARTICLE FORMATION 10.2.1 Nucleation For the synthesis of metal nanoparticles in the liquid phase, solutions of metal salts or complex compounds (as sources for the nanoparticle formation) are used. In these reactants, the metal is always present in an oxidized state (positive oxidation number). Therefore, the formation of metal nanoparticles from the liquid phase must always include a reduction step. One or more electrons have to be transferred from the reducing agent of the suited redox potential to the metal ion or the complex compound. So, in particular the first step of particle formation, namely, the nucleation, is a rather complex process. This step realizes the transition from the molecular disperse state (chemical solution) to the initial phase of the solid state (colloidal solution). It includes the following partial processes: . . .

Electron transfer from the reducing agent to the metal ion Exchange or removal of ligands from the primary coordination sphere of the metal ion Aggregation of a small group of metal atoms forming a nucleus

All three processes must occur simultaneously. So, the reaction conditions must suit to all these different reactions. On the one hand, the ligand exchange removal is normally strongly dependent on pH value. Low pH values lead to the destruction of many complexes since the protons can successfully compete with the metal ions. For example, ammonium or nitrogen-containing ligands that are applied frequently are protonated at low pH values resulting in a strong decrease in the coordination ability. At higher pH values, the complexes are more stable. On the other hand, the reduction power of reducing agents increases frequently with increasing pH. Thus, the nucleation probability decreases in many cases with lowering pH due to the increase in redox potential of the reducing agent. Furthermore, the pH value, together with existing ligands, surfactants, and the solution ionic strength, is responsible for the charge of forming metal nuclei. Uncharged nuclei will come together fast and can bind, aggregate, and sediment. As a consequence, large particles and irregular precipitates are formed. Charging is responsible for negative electrostatic interaction between the forming particles and contributes to the suppression of nanoparticle aggregation. The formation of small, well-defined, stable nuclei frequently depends on specific ligands supporting the complex mechanism of nucleation. It has been shown that appropriate derivatives of phosphoric acid can be applied to selectively form cluster species with exactly 11 or 55 gold atoms.1 Also, in case of particles of other sizes, ligands adsorbed at the particle surface support the electrostatic stabilization of the particles in the colloidal-dissolved state.

364

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

In contrast to nucleation, the growth of metal nanoparticles can be described by an electrochemical process, a mechanism well known from macroscopic heterogeneous chemistry: The electron transfer for conversion of metal ions or complex compounds into the neutral metal atoms of the solid state is divided into two partial processes that are electrically coupled by the electrochemical potential of the forming nanoparticle. The whole system has no electron exchange by a power supply. The formed electrochemical potential results from the superposition of an anodic and a cathodic process, which electrically compensate each other. The nanoparticle acts as a mixed electrode. The cathodic partial process is represented by the reduction of the metal ions to metal. Its potential depends on the position of the metal in the electrochemical series. It is influenced by the ligands of an original complex compound and by other species acting as ligands in the cathodic partial process. The anodic partial process is represented by the oxidation of the reducing agent. The electrode potential of this process is determined by the electrochemical standard potential of this oxidation process, the concentration of the reducing agent in the solution, and the pH. Despite the general importance of redox properties of involved species in both processes, the reaction rates for nucleation and nanoparticle growth may differ considerably. Nucleation is normally strongly nonlinear dependent on the reactant concentration. It goes on only above a critical concentration. The consumption of metal ions and reducing molecules leads to a decrease in the concentrations of these essential reaction partners, so that they can drop below the nucleation threshold. In contrast, the rate of the electrochemical processes in crystal growth of the forming nanoparticles follows the electrochemical laws and monotonously falls with decreasing reactant concentrations. The formation of nanoparticles of equal size can be expected if the nucleation occurs rapidly. After a short time interval with nucleation, the nucleation threshold should be achieved and then the further process is marked by the nanoparticle growth exclusively. A homogeneous growth of all nanoparticles can be expected if the time for crystal growth is much longer than the time for nucleation and if the local reactant concentrations are equal for all parts of the reaction solution during nanoparticle growth. Therefore, techniques are required for ensuring a very fast mixing of the reactants. This requirement is met by microreaction technology. Under homogeneous conditions, static micromixers can be used to realize fast mixing. So, multilamination mixers possess very short diffusion lengths, which results in short diffusion times. A certain disadvantage of processes under homogeneous-phase conditions is the fluid dispersion, leading to a broad distribution of residence times. A second disadvantage is the wall contact, leading to wall adhesion of particles, in particular under the high surface to volume ratios of microchannels. Both problems are overcome by introduction of the segmented flow technique for nanoparticle synthesis (see below). 10.2.2 Particle Growth The microreaction technology is suited for a combination of fast nucleation and growth initiation and moderate or slow nanoparticle growth. Different shapes of

MECHANISM OF METAL NANOPARTICLE FORMATION

365

nanoparticles can be addressed by choosing between transport or surface-controlled metal deposition during the growth of nanoparticles: A low electrochemical potential at the nanoparticle surface caused by high concentrations of a strong reducing agent possessing a low electrochemical overpotential at the particle surface leads to a fast reduction of metal ions on the surface. This results in concentration gradients in the environment of the nanoparticles and a transport control in the anodic partial process. Under these conditions, local differences in the surface energy are negligible for metal deposition. As a consequence, the particles grow as spheres. A slower growth of nanoparticles can be realized by weaker reducing agents and lower concentration as well as by the application of ligands forming more stable complexes with the metal cations or are preferentially bonded to certain binding places at the particle surface. The electrochemical open circuit potential of the growing nanoparticles is higher and nearly no concentration gradients appear under these conditions. So, small differences in local chemical properties and surface energy of the single binding places at the particle surface play an important role for the cathodic partial process, with implications for the metal deposition. As a result, crystallographic growth can be observed and the possibility of formation of triangles, hexagons, and cubes increases. The shape of nanoparticles and the ratio of different nanocrystal types can be influenced by the choice of reducing agent, ligands, solvent, and temperature. 10.2.3 Surface Capping The surface state of nanoparticles is important for the energetic state of the whole nanoparticle as well as for its chemical behavior. So, the ligand shell of a nanoparticle or adsorbed polymer molecules influence the plasmon absorption spectrum of the particle and modify the charge transfer behavior, the electrochemical properties, and the specific binding behavior between the particle and other particles, molecules, or surfaces. In addition, surface capping is always related to a change in the electrical particle charge. So, the aggregation behavior, the stability of colloidal solutions, the solubility properties, and precipitation are altered by surface modification. Surface capping is very important for the chemical and physical properties of nanoparticles and can be used for the extension of tuning ranges for nanoparticle parameters. The deposition of monomolecular layers, molecular bilayers, and multilayers allow controlling precisely the electronic and energetic coupling between a metallic nanoparticle and its environment. So, well-defined barriers for electron tunneling can be realized, which are of interest, for example, for single-electron devices. Different electronic properties of capping layers can be generated by application of monomolecular films containing organic conductors and semiconductor molecules. The orientated incorporation of asymmetrically functionalized organic molecules into capping layers allows the construction of diode-like or transistor-like systems with metal nanoparticles. The integration of dye molecules can be used for addressing of electronic energy between the metal core and the organic chromophores. This is used for resonance Raman spectroscopy. Plasmon energy of the metal core can be transferred into electronically excited states of organic chromophores and vice versa. In the future, such systems could also be of interest for nanoparticle-based

366

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

optoelectronic devices for nanometer-sized light sources, for solar cells, or for artificial photosynthesis. But all these applications require a high reproducibility in the formation of capping films and, therefore, well-controlled formation procedures. Micromixing is a convenient method for fast addition of surface capping reagents to colloidal solutions. It is of particular interest for two-step and multistep reactions with short residence times for fast switching of reaction conditions. So, nanoparticles, ligands, or parameters such as salt concentrations, pH, and ion strength can be varied quickly by fast mixing. Microreaction technology offers best possibilities for fast mass transfer to get highest homogeneity in the formation of capping layers. In addition, the thermal activation and the temperature dependence of adsorption equilibriums can be easily and quickly addressed by fast heat transfer and temperature changes under microfluidic conditions. Fast local convection, small channel diameters, and thin walls support the fast heat transfer in microreactors or in microsegmented flow. The synthesis of capping layer is often connected with a complete change in the solubility of nanoparticles. Frequently, metal nanoparticle capping is coupled with a transfer of nanoparticles between different solvents. Surfactants are used as phase transfer catalysts and form shells of surfactants or support the formation of shells of additional molecules.2 There are several approaches of microreaction technology for such phase transfer procedures. They can be realized by microcontactors, by segmented flow, and in particular by segmented flow three-phase systems consisting of an aqueous and a water-immiscible organic phase, both embedded in an inert carrier phase of a perfluorinated liquid. A third strategy for phase transfer processes in nanoparticle synthesis is based on the application of static micromixers for generation of emulsions and following reaction and phase separation.

10.3 NP PRODUCT AND PROCESS CHARACTERIZATION 10.3.1 Spectrometry The plasmon absorption of metal nanoparticles, first theoretically have been described by Mie,3 provides an opportunity for optical spectrometry of nanoparticle dispersions. Plasmons are density oscillations of charge carriers in metals or semimetals. In quantum mechanics, they act as quasiparticles (quantums) with energy EPl (given by equation (10.1)), where vPl is the frequency of the plasmon and ¯h the reduced Planck constant. The resulting energy EPl is valid for the so-called volume plasmons of solids. h EPl ¼ vPl


ð10:1Þ

For small particles, the plasmon is described by the Mie theory and called Mie plasmon.4 Mie plasmons are surface plasmons of a sphere and possess discrete spectra. Plasmon absorptions can be described as resonance of oscillating electrons of the conduction band with the exciting electromagnetic field (e.g., radiation in the range of

367

NP PRODUCT AND PROCESS CHARACTERIZATION

visible light). Therefore, it is possible to calculate the absorbance of colloidal solutions of metal nanoparticles if the size of nanoparticle is much lower than the wavelength of exciting light radiation (e.g., fulfilled by Au-NP with diameter of 25 nm and less) and known dielectric function of the nanoparticle material. For these small particles, the dipole approximation is valid5 and leads to equation (10.2). 3=2

el ¼

18pNP VeD e2 lðe1 þ 2eD Þ2 þ e22

ð10:2Þ

where Np is the number of nanoparticles, V is the volume of nanoparticles, eD is the dielectric constant of dispersion medium, l is the wavelength of absorbed light, and e1 and e2 are real part and imaginary part of dielectric function of the nanoparticle material, respectively. The extinction coefficient el of nanoparticle dispersions grows with the volume of particles and can reach values much higher than organic dyes, but low nanoparticle numbers in dispersions reduce el to the known values of metal organic compounds of about 3000–4000 L mol
1 cm
1. A very important part of plasmon absorptions is the fact of dependency from the particle diameter (Figure 10.1). This means that the particle size influences the absorption behavior, leading to shift of radiation absorption.6 Gold nanoparticles with diameter in the order of magnitude of 20 nm show a characteristic plasmon absorption with a maximum at about 526 nm. Larger gold particles and particle aggregates are characterized by an increase in the absorption at wavelengths above 600 nm. A shift in the optical properties can also be used for the detection or monitoring of the surface state of particles. Changing particle charges and the formation of adsorbate films influencing the electronic state of the particle result in changed plasmon absorption. Thereby, frequently a bathochromic shift can be observed. Silver nanoparticles as well as the formation of silver shells on gold cores can be easily detected by means of the sharp plasmon absorption at about 410 nm. Additional

FIGURE 10.1 Au and Ag nanoparticle dispersions with different nanoparticle sizes showing different colors.

368

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

longer wavelength shoulders or maxima indicate larger or nonspherical silver nanoparticles. Alloy-like binary particles have absorption maxima between the gold and the silver peak. Core/shell particles or other composite particles show, in general, broader peaks, double peaks, or shoulders. The size dependency of plasmon absorption opens an easy way to determine quickly the particle dimension or to observe the kinetics of particle growth by using common UV–Vis spectrometer. But the plasmon spectra are strongly dependent on the nanoparticle composition,7 the morphology of the particles,8 the chemical composition, the particle surface7,9 (e.g., type of surface ligands), the dispersion agent itself, and the grade of aggregation.6 A further interesting spectrometric method of nanoparticle characterization is the MALDI-TOF (time-of-flight) mass spectrometry,10 especially for ligand characterization on the particle surface. However, this method is limited to very small Au nanoparticles (<2 nm) as well as Au clusters due to mass restriction. Nanoparticles with higher diameter are too heavy to fly and reach the detectors of TOF mass spectrometer. 10.3.2 Optical Microscopy Abbe’s theory of imaging (equation (10.3)) describes the “problem” of optical microscopes with nanoparticles. In the range of visible light (400–780 nm), optical microscopes are not able to resolve dimensions of nanoparticles with diameters smaller than or equal to 100 nm. s¼

l l ¼ n sin a NA

ð10:3Þ

where s is the resolving capacity of a microscope, l the wavelength of used light, n the refractive index of medium, and NA the numerical aperture. In the best case (l ¼ 400 nm, high NA of about 2), it is possible to detect nanoparticle with minimum size greater than or equal to 200 nm as single object. A special method to detect nanoparticle with sizes smaller or equal to 100 nm is the dark-field microscopy. This technique uses the ability of very small objects to scatter light perpendicular to the beaming direction. To get dark-field images, a special darkfield condenser lens system is used that leads the incoming light in very small angle to the microscope objective. Due to this small angle, only scattered light from the nanoparticles passes the objective lens and is detectable. Nanoparticle scatters light very well due to the Mie scattering. The obtained images (Figure 10.2) show a dark background and very bright scattering centers. But these bright spots do not give correct information about the nanoparticle size and morphology. It is, however, important that the microscopic sample possess a high purity since impurities act as strong scattering centers. The detection and analysis of the scattered light from a single metallic nanoparticle (single plasmon resonance methods) with a spectrometer provides information about the particle size, morphology, and the chemical environment of the nanoparticle.11–14

NP PRODUCT AND PROCESS CHARACTERIZATION

369

FIGURE 10.2 Dark-field image of Au/Ag core/shell nanoparticles embedded in a thin polymer film on glass substrate (glass slide).

10.3.3 Ultramicroscopy Typical methods to determine the nanoparticle size and morphology are electron microscopic methods such as SEM (scanning electron microscope) and TEM (transmission electron microscope), as well as scanning probe microscopic methods like AFM (atomic force microscopy). Electron microscopic methods use a beam of electrons with certain energy to scan the object (e.g., Au nanoparticles in Figure 10.3) through interactions of the electron

FIGURE 10.3

SEM image from Au nanoparticles on silicon substrate.

370

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

beam with the material of the sample. These electron beam interactions depend on the material density, the chemical composition of the sample, electric properties of the sample material, and the energy of the electron beam. Electron microscopic methods can be used only for electroconductive samples. In the case of metal nanoparticle, this requirement is fulfilled. The difference between SEM and TEM, besides resolution varieties, is the manner of sample illumination. TEMs transmit the electron beam through the sample, whereas SEMs illuminate the sample surface (compare the difference in Figures 10.3 and 10.4). The sample preparation is therefore different and in case of TEMs quite laborious. For this reason, the SEM is the preferred method for routine analyses. TEMs and SEMs deliver two-dimensional pictures of samples (Figures 10.3 and 10.4). AFMs are further important microscopes for the determination of nanoparticle properties such as size (Figure 10.5) and morphology. They work with a mechanical probe (cantilever) in several working modes. The most important modes are the tapping mode with an oscillating cantilever tip in nanometer distance to sample surface and the contact mode where the cantilever pin gets in direct contact with the sample surface. In contrast to electron microscopes, the scanning speed is much lower, but the AFM gives real three-dimensional information of samples with a resolution down to few nanometers. Furthermore, AFMs can determine interactions between nanoparticles and their chemical environment.15 With chemically modified AFM tips, it should be possible to get information about the ligand sphere on the nanoparticle surface.

FIGURE 10.4 TEM image of Au/Ag core/shell nanoparticles. The Au cores are distinguishable in bigger particle (darker than Ag shell) due to stronger interactions with the electron beam.

NP PRODUCT AND PROCESS CHARACTERIZATION

371

FIGURE 10.5 Three-dimensional AFM image of Au nanoparticles.

10.3.4 Differential Centrifugal Sedimentation The differential centrifugal sedimentation (DCS) is used to determine the size distribution of particle-supported systems such as dye pigment suspensions. The size of typical detectable particle ranges from 20 nm up to 30 mm. For materials with a high density like Au nanoparticles, the lower detection limit is about 3–5 nm (Figure 10.6). Detection times depend on sample’s size distribution and density. Typical detection times are between some minutes (e.g., for metal nanoparticle) and hours (e.g., for polymer nanoparticles).

FIGURE 10.6 DCS spectrum of Au nanoparticles with sizes below 5 nm (device: DC 20000, CPS Instruments, USA).

372

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

Sedimentation methods use the fact of particle separation by g-forces (through the earth gravity or centrifuges) in media with known density and viscosity. For particle size determination, parameters such as the migration time or velocity are important to know. The active principle of sedimentation methods is described by the Stokes equation (10.4). The sedimentation velocity ns is dependent on particle diameter Dp, particle density rp, sedimentation medium density rs, and viscosity hs, as well as the acting g-forces a.

ns ¼


 aD2p rp
rs 18hs

ð10:4Þ

Low differences between sedimentation medium density and particle density (e.g., polymer nanoparticle) lead to a decrease of the sedimentation speed, whereas very high-density differences in case of metal nanoparticle dispersions increase the sedimentation rate. The Stokes equation is valid in case of smooth, spherical, and inelastic particles if the volume of sedimentation medium is much greater than the particle volume and the molecules of sedimentation medium are much smaller than the particle, as well as under nonturbulent or convective conditions. To separate particles smaller than 100 nm, the g-forces have to be higher than 9.81 m s
2 of the earth gravity force. For this reason, the nanoparticle sedimentation can be done in special centrifuges with several ten thousands rpm. In this case, the Stokes equation needs a modification because the g-forces change the value depending on the distance from the rotation center. If the separation medium temperature and angular speed vc of the centrifuge are constant over measurement time and the start point R0t, as well as end point RE of sedimentation is known, the particle diameter Dp depends on the sedimentation time ts (equation (10.5)). vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u18h lnðR =R Þ 1 u E  0 pffiffiffiffi Dp ¼ t
s ts rp
rs v2c

ð10:5Þ

A density gradient is used inside the centrifuge to support the sedimentation process and suppress sedimentation instabilities like the so-called streaming.

10.4 METAL NP SYNTHESIS IN HOMOGENEOUS FLUIDS In the past 5–6 years, an increasing interest in nanoparticle syntheses in microreactors or microfluidic devices is observed.16–18 Several types of nanoparticles were synthesized in microstructure containing arrangements ranging from semiconductor particles such as CdS,19,20 inorganic pigments21,22, and metallic nanoparticles such as Ag and Au nanoparticles with several morphologies23,24,26–28 to Au/Ag core/shell nanoparticles25 as well as Pt nanoparticles.29

METAL NP SYNTHESIS IN HOMOGENEOUS FLUIDS

373

But what is the most important reason to use the microreaction technology? There are some known advantages of nanoparticle syntheses in microreactor in contrast to batch syntheses. Microreactors or microfluidic channels allow more effective heat as well as mass transfer due to larger surface to volume ratios. Smaller dimensions of microchannels result in shorter mixing times compared to common batch reactors or streaming tubes, and in shorter mean diffusion length of molecules. In microchannels, the conditions for nanoparticle formation (nucleation and growth) are fulfilled as discussed in Sections 10.2.1 and 10.2.2 required. Due to better mixing, the nucleation step is preferred instead of the growth step. This results in a higher amount of nanoparticles in contrast to batch syntheses. Furthermore, the half-width of size distributions of microreactor produced nanoparticle was lowered to 50% of batch produced nanoparticles (Figure 10.7). This effect depends on the flow rates in the microchannels. In general, higher flow rates result in smaller size distributions and normally smaller nanoparticles (Figure 10.8). The size distribution and the nanoparticle size are affected not only by the flow rate but also by chemical conditions (e.g., pH value of reaction solution) or additives such as polymers (e.g., polyvinyl pyrrolidone (PVP)), or surfactants (e.g., sodium dodecyl sulfate (SDS)), which influence the diffusion behavior of the reactants and particularly the surface tension. But an important aspect using microreactors to form nanoparticles is the so-called reactor fouling. This means the precipitation or deposition of the nanoparticle-forming materials on the reactor channel walls. One reason is the formation of nucleation

FIGURE 10.7 Size distributions of Au nanoparticle synthesized by reduction of tetrachloroauric acid with ascorbic acid in a batch experiment and in a microreactor; cHAuCl4 ¼ 0:001 mol L
1 , cascorbic acid ¼ 0:020 mol L
1 , pH 9.5, and 0.25% PVP.

374

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

FIGURE 10.8 Particle diameter and half-width of size distribution of Au nanoparticles prepared in continuous-flow microreactor syntheses: cHAuCl4 ¼ 0:001 mol L
1 , cascorbic acid ¼ 0:020 mol L
1 , pH 9.5, and 0.025% PVP.

centers not only inside the reaction mixture but also on the channel surface. To avoid the reactor fouling, a few approaches are known. The simplest way is to minimize the residence time of reactants through high flow rates. In this case, the microreactor or micromixer just provides a good mixing but the nuclei formation takes place outside. The procedure works if the residence time is much lower than the nucleation kinetics. A second approach is to use hydrophobic channels for aqueous solutions. This can be realized by using hydrophobic materials (e.g., silicon chips made of PDMS) or by means of hydrophobization of glass channels. In this case, the formation of nucleation centers on the channel walls is reduced by the high interface energy (surface tension). The increased value of surface tension avoids the contact of the reactants with the channel surface. However, one disadvantage is low chemical resistance of hydrophobized glass channels. A third approach is to adjust the chemical behavior of the reactant solution or formed nanoparticles and the channel walls in such a manner that no nucleation on the walls can occur. One example for this is the preparation of Au nanoparticles by reduction of HAuCl4 with citric or ascorbic acid in glass microreactors. Using high pH values during the formation of the negative stabilized Au nanoparticles (due to the negatively charged surface ligands ascorbate or citrate), the reactor fouling is reduced. The reason is the deprotonation of the superficial OH groups of the glass surface, which causes a negative surface charge. In this case, the negatively charged nanoparticles are repulsed by the negative charge of the glass channel surface. Another method to avoid reactor fouling is using the segmented flow principle (see Section 10.5). As per description in Section 10.3.1, the characteristic plasmon absorption of nanoparticles offers a convenient way for noninvasive monitoring of the metal nanoparticle formation in microcontinuous-flow synthesis. Experiments with a time-resolved multichannel monitoring show that the formation and growth of Au

METAL NP SYNTHESIS UNDER SEGMENTED FLOW CONDITIONS

375

FIGURE 10.9 Scheme of setup for online monitoring by flow-through microspectrophotometry in Au/Ag core/shell nanoparticle syntheses under continuous segmented flow conditions.

and Au/Ag nanoparticles cannot be simply described by a continuous increase of a certain plasmon absorption. The nucleation and the growth of nanoparticles is always connected with a change in the electronic and optical properties of the particles. This effect is due to the size dependence of the electronic states of particles. The situation becomes complicated by the formation of binary or ternary metal nanoparticles because the changing composition and the growing size of the particles affect the change of the plasmon absorption during the particle formation. The response of optical properties of the reaction solution on the changing particle properties is well suited for the monitoring and detection of intermediate states. Changing optical properties including shifting as well as increasing plasmon absorption, for example, were measured by a six-channel arrangement of microflow-through spectrometers (see Figure 10.9) in a continuousflow synthesis of Au/Ag core/shell particles under segmented flow regime. 10.5 METAL NP SYNTHESIS UNDER SEGMENTED FLOW CONDITIONS 10.5.1 Specificity of Segmented Flow Conditions Liquid/liquid two-phase systems as emulsions are well known for long time and applied in a lot of special materials and technical processes. Normally, they are marked by a more or less statistical distribution of the single phases. In contrast, a well-organized series of segments connected with a homogeneous size and regular distance between the single phases was introduced for FIA analysis about 30 years ago.30,31 Droplets of analyte were embedded in a phase of a chemically inert and immiscible carrier liquid. So, well-defined portions of analyte solution can be transported by an ideal plug flow and the sequence of droplets can be used for precise analytical measurements. This principle became particularly important for

376

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

serial sample handling in high-throughput biomolecular analytical procedures,32–34 for miniaturized cellular screenings,35–38 and is of particular interest for a miniaturized chemistry.39–42 The microsegmented flow is marked by several specific features making this approach particularly suited for nanoparticle synthesis. An ideal decoupling in the reaction solution can be achieved if the inner surface of the wall has well-adapted properties to optimal interaction of the carrier liquid and the wall and the interaction of the embedded phase with the wall is suppressed. In case of mineral oil or aliphatic separation phase, an alkyl silanization is a possibility to reduce the wall wetting by an aqueous phase considerably. The silanization supports a good wall wetting by the lipophilic phase. The wetting behavior determines the geometry of the liquid/liquid interface. High wettability for the organic phase and low wettability for the aqueous phase result in a high contact angle for the aqueous phase. Segments can be completely released from the channel surface under fluid actuation and segment motion. A second important advantage of segmented flow is the residence time behavior. Homogeneous fluids moving through a channel form a laminar flow in case of lower Reynolds numbers, which is typical in the case of microfluidics. The consequence of laminar flow is that fluid volumes near the wall are moved much slower than fluid volume elements at the center of the channel. This effect leads to a strong fluidic dispersion. A sharp rectangular concentration signal applied to a microfluidic channel—for example, a reactant for nanoparticle formation—is reduced and spread over a larger region during the fluidic transport. As a consequence, large distributions of residence time are observed. The residence time distribution increases with increasing channel length. Segmented flow shows the opposite behavior. The transport of droplets or slugs is an ideal plug-like motion. The residence time behavior is very sharp and independent of the length of the transport path. The order of segments remains constant. All processed segments have the same residence time. 10.5.2 Mixing in Segmented Flow A third important advantage is the convective behavior related to the plug-like transport of fluid segments. The constant motion of the segment and suppression of normal laminar flow structure must be compensated by a radial motion component. The motion of segments induces a strong segment-internal conversion. Inside the fluid segments, pairs of vortices are formed. This increased convection must be compensated by an increase in fluidic pressure drop.43–45 However, the energy for segmentinternal flow-induced convection represents a unique opportunity for fast mixing inside fluid segments. Segment-internal convection leads to very efficient mixing, in particular at higher flow rates. Besides mixing, the heat transfer from the walls into the liquid is also strongly supported by the segment-internal convection. The strong radial component of liquid motion inside segments contributes very efficiently to the transport of heat between the inner part of the liquid and the walls. The heat transfer by segment-internal convection also increases with increasing flow rate.

METAL NP SYNTHESIS UNDER SEGMENTED FLOW CONDITIONS

377

The application of microsegmented flow for nanoparticle synthesis is a compromise between a maximum mixing rate or heat transfer efficiency and a high reproducibility in segment formation. In general, higher homogeneity in segment formation can be achieved at lower flow rates because the hydrodynamic conditions are more stable than in case of higher flow rates. In contrast, the transfer rates during mixing as well as the convection-induced heat transfer are intensified by increasing flow velocities. A homogeneous segment formation, dosing, or fusion behavior is an essential precondition of homogeneous product quality and narrow distribution of particle properties. Fluctuations in segment size and segment distances cause a shift in the concentration ratios of educts after dosing or droplet fusion, which results in changing conditions in particle formation. These problems are well illustrated by the consequences of formation of alternating droplet sizes or alternating droplet distances in some experiments at high flow rates. Such a behavior is responsible for formation of segments with periodically changing reactant ratios. The colloidal product solutions of such synthesis experiments are frequently marked by a bimodal or a multimodal particle size distribution (Figure 10.10) depending on droplet size distribution. Changing droplet sizes are a result of inhomogeneous droplet formation (Figure 10.12), which leads to longer segments with higher concentrations and shorter segments with lower concentrations of the added compound. 10.5.3 Process Homogeneity and Product Quality The control of segment parameters is important for the optimization of flow rate and reaction conditions inside the microfluidic segments. The quality of segment sequences can be characterized by fast microphotometric measurements applied for a direct monitoring of segment formation. The key parameters such as “segment size” and “segment distance” are determined by an automatic procedure. The quality of the whole process can be characterized by segment size/segment distance plots. They show the modality and the distribution width of the generated segments for the

FIGURE 10.10 Bimodal size distribution of Au/Ag core/shell nanoparticles due to alternating droplet sizes.

378

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

complete segment sequence generated and used in a synthesis experiment. Examples are given in Figure 10.11a and b. Alternating segment distances result in alternating particle sizes in the single segments (Figure 10.12), which is well reflected in the centrifugal sedimentation spectra (Figure 10.10). High homogeneity of particles can be achieved if the conditions of formation of microfluid segments, dosing, and mixing are well controlled. Monomodal particle distributions were obtained if the segments are formed regularly and with constant frequency. The product quality can be enhanced by increasing flow rate. High total flow rates of carrier liquid are always connected with high segment-internal convection. An increase of the local convection reduces the time needed for diffusive mixing considerably. So, optimal conditions for fast and homogeneous nucleation, fast termination of nucleation, and homogeneous particle growth can be realized.

Segment distance (ms)

(a) 2900 2400 1900 1400 900 400

99.0 % 100

300

500

700

900

Segment length (ms) (b) Segment distance (ms)

1700 1400 1100 800

7% 500

80.3%

7%

200 50

150

250

350

450

550

Segment length (ms)

FIGURE 10.11 Scatterplot of segment size and segment distance with (a) monomodal segment distribution and (b) trimodal segment distribution. Percentage shows the amount of segments in the respective scatterplot.

METAL NP SYNTHESIS UNDER SEGMENTED FLOW CONDITIONS

379

FIGURE 10.12 Schematic of segment formation with alternating composition caused by alternating segment distances at constant injector flow rate.

The specific advantage of microfluidics and, in particular, of segmented flow for the formation of nanoparticles is the control of fast mixing for process initiation and for process quenching. So, fast reaction steps under microcontinuous-flow conditions can be combined with very precise and narrow residence time for growth processes. One strategy for composite metal nanoparticle synthesis can be the combination between fast continuous-flow nucleation or growth initiation and a following slow growth process in batch. Another strategy is the combination of fast steps with longer residence loops for the realization of a combination of fast initiation and a slow surface-specific deposition process under complete microcontinuous-flow conditions. 10.5.4 Core/Shell Metal Nanoparticles Generated by Use of Microsegmented Flow The microsegmented flow is of particular interest for the formation of core/shell metal particles with high homogeneity. Fast nucleation of the core material and the fast initiation of deposition of the shell material can be achieved by fast mixing in microfluidic segments. It is very important to silanize the internal device surfaces if glass/glass or glass/silicon double injector chips are used. Then, very homogeneous segment formation and homogeneous dosing into performed segments can be realized. Core/shell metal nanoparticles can also be synthesized in tube arrangements. T-junctions of PTFE are well suited for segment generation and for injection and dosing. The reproducibility of segment formation is not as high as in the case of silanized chip devices, but it is, however, sufficient for forming high-quality core/shell nanoparticles. The flow conditions and the rate of mixing not only are important for the initial nucleation and growth of nanoparticles but also influence the quality of core/shell particles. Flow rate-dependent particle diameters were found if gold seeds are introduced into microfluid segments and are mixed with reducing agent and silver

380

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

FIGURE 10.13 Au/Ag core/shell nanoparticles prepared by reduction of tetrachloroauric acid and silver nitrate in aqueous solution at ambient temperature by application of microcontinuous-flow synthesis for starting silver shell formation: (a) initial Au/Ag core/shell particles with size of about 60 nm, (b) enforcement of silver shell at lower silver salt concentration resulting in particle diameters of about 120 nm, (c) enforcement of silver shell at higher silver salt concentration; Au seed particles with a diameter of 50 nm were enforced by 1.4 mmol L
1 AgNO3 solution by reduction with ascorbic acid (4.0 mmol L
1) in the presence of thiourea (1.0 mmol L
1) resulting in larger particles of about 250 nm diameter.

CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY

381

salts. This effect is obviously caused by a selection between gold seeds for silver deposition. The percentage of gold particles covered by silver is enhanced with increasing flow rate. At lower flow rates, larger particles were obtained while smaller particles were found at high flow rates. It is assumed that a fast mixing of the colloidal solution of gold seeds with the silver salt is responsible for a homogeneous start of silver deposition on the gold seeds. Different thicknesses of silver shells can be realized by enforcement of primary formed gold particles through further silver deposition (Figure 10.13). Slower mixing in case of lower flow rates results in concentration gradients inside the microfluid segments in timescales comparable to the silver shell growth, which could explain the observed thicker shells at lower flow rates. The development of microfluidic synthesis for binary and ternary nanoparticles in microfluid segments always demands for an adaptation of flow conditions for high segment quality and fast mixing at the same time. Unfortunately, the regularity of segment formation often decreases with increasing flow rate. This fact is obviously due to a bifurcation in the flow and droplet formation behavior, convective fluctuations, and also transition to turbulent conditions at very high flow rates. It leads to changes in segment size and in particular in segment distances causing differences in concentrations and also in concentration ratios if dosing by injection of reactants into preformed segments is applied. Larger particle size distribution and bimodal or multimodal nanoparticle peaks are obtained. Thus, a compromise for flow rates has to be found allowing sufficient high segment-internal convection on the one side and stable and regular segment formation on the other side.

10.6 CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY 10.6.1 Addressing of Morphological Classes and Yield Improvement for Nonspherical Particles Future demands on the technical development of microreactors, microfluidics, and new miniaturized procedures for synthesis of metal nanoparticle result from the general requirements and vision of nanoparticle fabrication and from the specific properties and possible applications of metal nanoparticles. Microreaction technology should contribute by their specific features as fast heat and mass transfer and fast switching of reaction conditions to the further development of nanomaterials. So, microreaction technology will probably become one of the most important classes of technological procedures for production of high-quality nanoparticles. Microreaction technology has to meet the requirements for different metals, alloys, and compounds for a large spectrum of new composite functional materials. Besides the elementary composition, the addressing of specific nanoparticle shapes and the reproducibility of morphology and size of the generated particles are a big challenge. Despite the recent progress in synthesis of metal and semiconductor nanoparticles by microcontinuous-flow synthesis, the development of synthesis protocols and

382

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

microreaction tools for a larger spectrum of metals and even less noble elements must be intensified. Protocols and reactor arrangements must be qualified in order to become suited for the continuous-flow generation of a large spectrum of metals and alloys and a large spectrum of size classes and morphological types. A particular need is the generation of nonspherical nanoparticles with high yield. The advantages of microreaction technology should be utilized consequently for realizing a selection of particle types by suitable mixing, reaction, and flow conditions in order to become able to synthesize a certain shape type exclusively. A high selectivity in nanoparticle shape is necessary to avoid any later selection or separation steps. In principle, microreaction technology should allow to find reaction conditions and optimal process chains resulting in a free addressability of shape and size classes of nanoparticles.

10.6.2 Improvement of Homogeneity, Shape-Identical, and Size-Identical Particle Populations The technical application of nanomaterials is mainly dependent on the properties of single components. So, nanoparticle production should be able to ensure constant quality in the particle parameters. The demands result from the homogeneity requirements in production, processing, and application. The properties and the parameter distribution of nanoparticles determine the chemical behavior, the density, and the mechanical, optical, and other physical parameters of composite materials. For automated processing and larger technical applications, constant properties of nanoparticles and constant distribution are absolutely needed. Longtime stability, aging effects, and response of materials against any kind of stress depend on the properties of the nanoparticles, their spatial distribution, and the distribution of their properties. It is well known that the application of suitable ligands can be used for the synthesis of metal clusters consisting of an exactly defined number of metal atoms. These metal clusters are a special type of small nanoparticles with absolutely monodispersity in composition, size, and shape. The definition of atom number and position of atoms inside these particles make them an analogue to molecular objects. In contrast, the overwhelming number of metal nanoparticles is formed by a more or less statistical deposition of metal atoms on the surface of a growing core resulting always in a certain size and shape distribution. A strict control of local transport and reaction conditions could help to achieve a nucleation and particle growth that goes on always following the same atomic interactions and the same scheme of deposition of atoms at the particle surface. Such a process can be thought in analogy to the epitactic growth of single-crystalline films known from solid-state semiconductor techniques using gas phase deposition. It is imaginable that the strict control of fluid motion, diffusion, and rates of surface reactions at different crystallographic planes of a growing particle can result in a strong reduction of statistical effects and an important improvement of reproducibility in growth of the individual nanoparticles.

CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY

383

The ultimate demand would be the synthesis of nanoparticles of a certain type of shape with an absolute identical number and arrangement of the involved atoms. The idea is to synthesize nanoparticles having the quality similar to that of molecules of a pure substance. Nanoparticles of this quality are small solids with the character of a three-dimensional macromolecule from the point of view of the definition of spatial organization. It is assumed that the goal of obtaining nanoparticle populations of complete identical shape and size can only be realized by the application of microreaction technology. 10.6.3 Continuous-Flow Synthesis of Composite Nanoparticles by Two and Multistep Reactions The architecture of composite nanoparticles is influenced by the character of the components, the applied precursors and reagents, and the temporal order of particle forming reactions. So, the reaction protocol can decide if an alloy or doped nanoparticle or a nanoparticle consisting of two regions with different elementary compositions is formed. Complex composite nanoparticles with different regions are obtained, for example, by deposition of one material after the completion of nucleation and deposition of the first material. Such synthesis succeeds if the nucleation of the second material occurs only at a selected surface areas of the primary particle formed by the first material. Normally, these surface-sensitive processes are slow reactions. It is not the transport but the surface reaction that must be the rate-determining step. A high reproducibility can be expected if the second material is deposited only on planes with a specific crystallographic orientation. Slow processes are normally carried out in batch synthesis. A combination of fast steps of reaction initiation and formation of stable nuclei must be combined with a surface control of primary particle growth. Therefore, it could be of importance to have well-defined conditions of local convection and particle motion inside the reaction mixture in order to have optimal condition of particle growth and to avoid aggregation, wall adsorption, and sedimentation. That is why microreaction technology could also be interesting for the slow steps of the whole process chain. These reaction steps must be realized by longer residence loops. This demand corresponds well with the principle of slowly moved fluid segments for the synthesis of different types of nanoparticles in the so-called SFTR (segmented flow tube reactor).46 This technique is the key for slow ongoing microcontinuous-flow reactions that can be combined with fast mixing steps or fast temperature changes to either get nucleations with high rates in a small time frame or change quickly the conditions of surface binding of effector molecules modulating the deposition of metals on different crystallographic planes. The definition of nucleation and growth phases of a second, third, or further material on the surface of preformed initial nanoparticles in a microcontinuous-flow process demand for good insights into the mechanisms. Changes in pH, temperature, or ligand concentration as well as the application of reducing agents with different redox potentials are tools for controlling the anodic and cathodic partial processes of open circuit metal deposition. Under certain conditions, the dissolution of metal

384

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

particles or a competition between metal deposition and dissolution can occur. The control of outer currentless electrode potential of nanoparticles is essential for giving the formation process of composite metal nanoparticles the right direction and performance. Obviously, the application of ligands binding selectively to certain crystallographic planes can control the deposition or redissolution of metal in the critical potential range. Besides the study of the elementary reaction steps, process monitoring can help to optimize the reaction conditions, the temporal order, and the speed of the single reaction steps. The types of obtained composite particles can vary considerably depending on the deposition conditions. Same crystallographic planes on the surface of the primary formed nanoparticle should undergo identical nucleation and growth for the second material if statistical effects are not dominating for the nucleation. Asymmetric shapes can be obtained if a nucleation of one crystallographic plane will alter the nucleation conditions drastically enough for inhibiting a nucleation of another surface area of the same crystallographic type. Such conditions can result from stronger changes in the electrochemical potential of the nanoparticle due to the first nucleation step. 10.6.4 Regional Functionalization and Three-Dimensional Self-Assembling of Nanoparticle A second possibility for generating composite nanoparticles is the connection of two or more preformed nanoparticles. In principle, nanoparticles can be moved in liquids by the thermally actuated Brownian motion. They stick together if the binding energy is higher than the thermal activation. The formation of stable chemical bonds leads to irreversible aggregation. In principle, all nanoparticles tend to aggregate. Aggregation takes place if solvation and electrostatic repulsion are too weak to inhibit a direct contact of nanoparticles. Metal nanoparticles are particularly sensitive against aggregation due the possibility of forming metal bonds. Metal bonds are much stronger than van der Waals interactions, for example, and can take place in all cases of direct contact between “naked” metal surfaces. Thus, the problem for building of nanoparticle assemblies is not to aggregate them, but to avoid undesired aggregation and enable the particles to a highly selective binding behavior. The construction of well-defined particle aggregates demands stable colloidal solutions of reactants on the one side and specific chemical interaction between particles on the other side. The simplest strategy of binary assembling was found by the principle of complementary surface functionalization. This principle was realized, for example, by covering two populations of gold nanoparticles with complementary chains of thiolated oligonucleotides.47–51 The cooperative effect of hybridization with a sufficient high number of hydrogen bonds leads to an overcoming of repulsion due to negative excess charge between the colloidal particles. The surface functionalization of nanoparticles can be performed in batch as well as in microreactors. Besides the formation of particle dimers, more types of aggregates can be found due to the coverage of the complete particle surface by the binding molecules (Figure 10.14).

CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY

385

FIGURE 10.14 Formation of nanoparticle aggregates depending on binding molecules.

A well-defined geometric condition for particle binding requires a regional functionalization. This demand means that the most important challenge for nanoparticle self-assembling is the realization of regional complementary functionalized nanoparticles. Nanoparticle pairs are formed by two particle types with one complementary region. Two complementary regions on the same particle or double regionally functionalized particles of two complementary types are forming chains or rings. Three or more binding domains will allow the formation of three-dimensional nanoparticle assemblies (Figure 10.14). A regional functionalization can be based in principle on selective reaction of different crystallographic planes of surfaces of a nanocrystal, but spherical nanoparticles or particles with low differences should be functionalized differently in two or three surface regions. How microreaction technology can support the formation and the assembling of regionally functionalized nanoparticles? Regional functionalization could be supported by enrichment of nanoparticles at liquid/liquid interfaces. A two-region functionalization could be realized if the affinity to both phases is regionally enhanced by two different chemical surface reactions in the two different liquid phases. The surface modification will enhance the stability of particles at the interface (Figure 10.15). The process could be realized under continuous-flow conditions if emulsions or microfluid segments are applied as liquid/liquid two-phase system. A three-region functionalization is more difficult to achieve. It could be realized by application of two different succeeding liquid pairs. Two differently functionalized surface regions are generated in the first step. One of the two regions must then additionally be functionalized by a third reaction modifying only a part of this region by application of a second liquid pair. Therefore, the wetting conditions in all three liquids must be well adapted. The three-region functionalization allows in principle the construction of any kind of two- or three-dimensional objects by self-assembling if a suitable complementarity of surface functionalization is chosen (Figure 10.16). So,

386

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

FIGURE 10.15 Stabilization of particles at the liquid/liquid interface by phase-specific surface modification.

more or less compact objects consisting of nanoparticles can be realized and more or less dense three-dimensional nanoparticle networks can be formed. 10.6.5 Nanoparticle-Based and Catalytic Nanomachines The vision of nanoparticle-based nanomachines could be realized if the self-assembling of nanoparticles leads to combinations of rigid and flexible structure elements in one nanosystem. In analogy to biomacromolecules, such nanoparticle machines could be actuated by molecular energizers and operated in dependence on substrates, initiators, promoters, and inhibitors. From a chemical point of view, such systems would work like a catalyst. From a mechanical point of view, they would work as a mechanical machine. Metals and metal compounds are well known for a lot of different catalytic effects. They could

FIGURE 10.16 Examples of different assembly types in case of particles with three different functions.

CHALLENGES OF METAL NP SYNTHESIS IN MICROREACTION TECHNOLOGY

387

support, for example, redox processes or the transfer of atoms or atom groups in molecules or from one molecule to another. In principle, different metals or alloys can be combined in composed nanosystems by nanoparticle self-assembling. The construction of nanoparticle-based catalytic assemblies and the connection of chemical activity on specific reaction sites with movability of the nanoenvironment of this reaction site would open a lot of possibilities for well-designed artificial catalysts. The so-formed catalytic nanosystems are thought to operate similarly to biological enzymes. Many highly efficient biocatalysts contain metal complexes in their active centers. It could be imagined to combine nanoparticle self-assembling also with more or less movable molecular structures. So, a large spectrum of new nanomachines for a huge number of different functions is imaginable. Additional electrochemical effects can be achieved by using the metal nanoparticles inside the particle assemblies for redox processes with local charge transfer via the nanoparticle surface. The nanoparticle acts as an open circuit electrode with mixed potential in this case. Photochemical processes and localized photothermal processes can be supported, in addition, by using the plasmon absorption band of suitable metal nanoparticles that are incorporated into the nanoparticle assembly. All these processes require well-controlled chemical and electrochemical conditions in the environment of the nanoparticle-based machines. Microfluidics and microreaction technology are challenged to supply the needed local fluidic conditions for operating the nanomachines. Microreaction technology is also required for the realization of compartments and for switching operation conditions to realize different catalytic processes in parallel and the desired temporal order. 10.6.6 Electronic and Optoelectronic Devices Formed by Self-Assembled Metal Nanoparticles Metal nanoparticles are of particular interest, in general, due to their specific electronic and optical properties. It is assumed that combinations of metal nanoparticles and molecules will play an important role after the era of integrated solid-state electronics using semiconductors. Future electronic devices will probably have to meet requirements of highly locally movable electrons, fast switching, well-defined tunneling barriers, and strong localization. The basics for one kind of future electronics were already created in the frame of single-electron tunneling devices. Networks of metal nanoparticles, dielectric shells, and molecular components will probably best suit for the realization of this type of future information processing system. Metal nanoparticles are also of interest for new classes of optoelectronic devices. The higher number of possible electron states distinguishes the metal nanoparticles from molecules and the discretization of electron states instead of a continuous band structure that distinguishes them from bulk metals or semiconductors. So, resonant interactions of metal nanoparticles and nanoparticle/molecule architectures with light lead to phenomena of molecular excitation and relaxation as well as solid-state effects, as known from photoelectrochemistry. So, nanoparticle systems including metal components seem to be suitable for conversion of light into electrical energy and

388

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

vice versa. This is of interest for light emitters and optical sensors as well as for nanolocal energy conversion and for development of new information processing systems integrating the handling and conversion of single photons, excitons, plasmons, bipolarons, and electric charges. 10.6.7 Microfluidic Strategies for Particle-Based Information Processing and Multichannel Interfaces The connection of three-dimensional nanoparticle assembling with fluidics could also open the way for highly integrated three-dimensional systems for information processing and storage. On the one hand, nanoparticle assembling could help construct nanochannels and compartments for nanofluidics. On the other hand, nanofluidics can support the transport and addressing of nanoparticles. Probably, nanoparticle functionalization, nanoparticle self-assembling, and nanofluid-based particle handling will become key elements for future information processing systems. A simple three-dimensional operation of information in microfluidic networks could be realized by encoding information by nanoparticle combinations in microfluid segments. The whole system could consist of a hierarchy of information storage and handling including the following levels: . . . .

Fluid strings: segments of liquid embedded in an immiscible carrier liquid inside micro- or nanochannels Larger nanoparticles inside the fluid segments Small nanoparticles assembled in the larger particles Molecules bond on the surface of smaller nanoparticles

Presently, the typical segment size in microfluidics is in the range between about 1 mL and some dozen of nanoliters. But there are many reports on smaller liquid portions handled by the principle of microsegmented flow. So, it is possible to scale down this principle to the femtoliter range. A femtoliter segment is large enough for containing 1000 nanoparticles with a diameter of 50 nm. Each of these nanoparticles can be composed of hundreds of smaller nanoparticles with diameters of 5 nm, which can carry hundreds of molecules on their surface. So, each femtoliter segment could, in principle, contain 107–108 elementary information units corresponding to about 100 MB. The space need for a femtoliter fluid segment in the plane is 10 mm2. That means that about 1000 TB could be stored in a planar microchannel system of 1 cm2. A three-dimensional microchannel system with 1000 fluid microchannel levels would possess an estimated storage capacity of about 1 billion TB. New strategies in liquid handling and nanoparticle management are necessary for writing and reading in this nanoparticle-based information handling system and for logic operations. Probably, the liquid handling operations will be similar to recent strategies of generation, switching, splitting, and fusion in segmented flow microfluidics. It can be expected that the development of microreaction technology for synthesis and modification of nanoparticles will prepare the first steps into a hypothetic future “fluidoparticle informatics” (Figure 10.17).

389

CONCLUSIONS

FIGURE 10.17

Schematic of a possible “fluidoparticle informatics.”

10.7 CONCLUSIONS There are a lot of very promising opportunities and challenges for the application of microreaction technology for the synthesis and manipulation of nanoparticles. This technique is under development for the generation of nanoparticles with very high homogeneity in particle properties. The possibility of sharp residence time distribution, fast mixing, and fast heating offers the best conditions for a homogeneous nucleation and short nucleation phases during the nanoparticle formation. Thus, narrow distributions of particle sizes can be achieved. The high quality of micro- and nanoparticles generated by the segmented flow tubular reactor (SFTR) is mainly attributed to these effects. Further effort in understanding mechanisms of nucleation and control of nucleation processes should help improve this critical process step for different types of nanoparticles. In all cases, the nucleation should start under same conditions in each volume element of a reaction mixture, go on under the same conditions, and be finished after the same time interval. This time interval should be as short as possible to avoid different starting conditions and different times for the growth of nanoparticles following the nucleation. So, for the different nanoparticles, special protocols are realization for realization of simultaneous initiation of nucleation and for very fast nucleation processes. The possibility of generating homogeneous particle qualities is not restricted on chemically uniform nanoparticles only. The advantages of microfluidics can also be used for producing binary and other more complex nanoparticles with constant structural and composition features. So, the mixing and reaction activation conditions can be optimized under microcontinuous flow, so that only one of several possible nanoparticle types is formed. For example, the formation of binary metal nanoparticles might be restricted by reaction condition in such a way that alloy nanoparticles, core/ shell particles, twin particles, or higher aggregates are formed.

390

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

The application of different strategies for microreactor-based nanoparticle synthesis opens a wide spectrum of reaction conditions for addressing different particle properties with narrow parameter distribution. It is expected that these possibilities will be used in future for further improvement in synthesis methods for metal nanoparticles with specific catalytic, optical, and electronic properties; for specifically functionalized metal nanoparticles with selective aggregation and binding behavior; and for very different applications integrating the specific advantages of moleculelike, solid-like, and mesoscale behavior of nanoparticles. These specific features are very important for new generations of electronic and optoelectronic devices as well as for new approaches in catalyst design and nanochemistry. It is assumed that microfluidics will help in future to qualify the automated microsynthesis and modification of metal nanoparticles into new principles of nanoengineering, combinatorial development of particle-based nanoarchitectures, developing of nanodevices, and nanoparticle-based information handling.

ACKNOWLEDGMENT We would like to thank J. Wagner, M. Brust, and T. R. Tshikhudo for their cooperation, S. Schneider and F. M€ oller for their support of the experimental work, and F. Jahn and H. Romanus for ultramicroscopic images. REFERENCES 1. Foos, E. E.; Twigg, M. E.; Snow, A. W.; Ancona, M. G. Competition between thiol and phosphine ligands during the synthesis of Au nanoclusters. J. Cluster Sci. 2008, 19, 573–589. 2. Levy, R.; Thanh, N. T. K.; Doty, R. C.; Hussain, I.; Nichols, R. J.; Schiffrin, D. J.; Brust, M.; Fernig, D. G. Rational and combinatorial design of peptide capping ligands for gold nanoparticles. J. Am. Chem. Soc. 2004, 126, 10076. 3. Mie, G. Ann. Phys. 1908, 25(3), 377–455. 4. Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles, Wiley–Interscience, 1983. 5. Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters, Springer: Berlin, 1995. 6. Link, S.; El-Sayed, M. A. Size and temperature dependence of the plasmon absorption of colloidal gold nanoparticles. J. Phys. Chem. B 1999, 103, 4212–4217. 7. Sun, Y. G.; Xia, Y. N. Increased sensitivity of surface plasmon resonance of gold nanoshells compared to that of gold solid colloids in response to environmental changes. Anal. Chem. 2002, 74(20), 5297–5305. 8. Link, S.; El-Sayed, M. A. Spectral properties and relaxation dynamics of surface plasmon electronic oscillations in gold and silver nanodots and nanorods. J Phys. Chem. B 1999, 103, 8410–8426. 9. Linnert, T.; Mulvaney, P.; Henglein, A. Surface chemistry of colloidal silver: surface plasmon damping by chemisorbed I-, SH-, and C6H5S-. J. Phys. Chem. 1993, 97, 679–682.

REFERENCES

391

10. Dass, A.; Stevenson, A.; Dubay, G. R.; Tracy, J. B.; Murray, R. W. Nanoparticle MALDI-TOF mass spectrometry without fragmentation: Au25(SCH2CH2Ph)18 and mixed monolayer Au25(SCH2CH2Ph)18
x(L)x. J. Am. Chem. Soc. 2008, 130, 5940–5946. 11. S€ onnichsen, C.; Franzl, T.; Wilk, T.; von Plessen, G.; Feldmann, J.; Wilson, O.; Mulvaney, P. Drastic reduction of plasmon damping in gold nanorods. Phys. Rev. Lett. 2002, 88(7), 0774021–0774024. 12. Mock, J. J.; Oldenburg, S. J.; Smith, D. R.; Schultz, D. A.; Schultz, S. Composite plasmon resonant nanowires. Nano Lett. 2002, 2(5), 465–469. 13. Kalkbrenner, T.; Hakanson, U.; Sandoghdar, V. Tomographic plasmon spectroscopy of a single gold nanoparticle. Nano Lett. 2004, 4(12), 2309–2314. 14. Song, Y.; Zhang, T.; Yang, W.; Albin, S. Correlating plasmon resonance spectra with threedimensional morphology of single silver nanoparticles. J Phys. Chem. C 2008, 112, 10767–10772. 15. McNamee, C. E.; Pyo, N.; Higashitani, K. Atomic force microscopy study of the specific adhesion between a colloid particle and a living melanoma cell: effect of the charge and the hydrophobicity of the particle surface. Biophys. J. 2006, 91(5), 1960–1969. 16. deMello, A. J. Control and detection of chemical reactions in microfluidic systems. Nature 2006, 442(7101), 394–402. 17. Watts, P.; Wiles, C. Recent advances in synthetic micro reaction technology. Chem. Commun. 2007, 5, 443–467. 18. Song, Y.; Hormes, J.; Kumar, C. S. S. R. Microfluidic synthesis of nanomaterials. Small 2008, 4(6), 698–711. 19. Edel, J. B.; Fortt, R.; deMello, J. C.; deMello, A. J. Microfluidic routes to the controlled production of nanoparticles. Chem. Commun. 2002, 10, 1136–1137. 20. Uehara, M.; Nakamura, H.; Maeda, H. Preparation of ZnS/CdSe/ZnS quantum dot quantum well by using a microfluidic reactor. J. Nanosci. Nanotechnol. 2009, 9(1), 577–583. 21. Kockmann, N.; Kastner, J.; Woias, P. Reactive particle precipitation in liquid microchannel flow. Chem. Eng. J. 2008, 135(S1), S110–S116. 22. Su, Y. -F.; Kim, H.; Kovenklioglu, S.; Lee, W. Y. Continuous nanoparticle production by microfluidic-based emulsion, mixing and crystallization. J. Solid State Chem. 2007, 180 (9), 2625–2629. 23. Wagner, J.; K€ ohler, J. M. Continuous synthesis of gold nanoparticles in a microreactor. Nano Lett. 2005, 5(4), 685–691. 24. Boleininger, J.; Kurz, A.; Reuss, V.; Soennichsen, C. Microfluidic continuous flow synthesis of rod-shaped gold and silver nanocrystals. Phys. Chem. Chem. Phys. 2006, 8(33), 3824–3827. 25. K€ ohler, J. M.; Held, M.; H€ ubner, U.; Wagner, J. Formation of Au/Ag nanoparticles in a two step micro flow-through process. Chem. Eng. Technol. 2007, 30(3), 347–354. 26. Wagner, J.; Tshikhudo, T. R.; K€ ohler, J. M. Microfluidic generation of metal nanoparticles by borohydride reduction. Chem. Eng. J. 2008, 135(1), 104–109. 27. Shalom, D.; Wootton, R. C. R.; Winkle, R. F.; Cottam, B. F.; Vilar, R.; deMello, A. J.; Wilde, C. P. Synthesis of thiol functionalized gold nanoparticles using a continuous flow microfluidic reactor. Mater. Lett. 2007, 61, 1146–1150.

392

METAL NANOPARTICLE SYNTHESIS IN MICROREACTORS

28. Tsunoyama, H.; Ichikuni, N.; Tsukuda, T. Microfluidic synthesis and catalytic application of PVP-stabilized, similar to 1 nm gold clusters. Langmuir 2008, 24(20), 11327–11330. 29. Niesz, K.; Hornyak, I.; Borcsek, B.; Darvas, F. Nanoparticle synthesis completed with in situ catalyst preparation performed on a high-pressure high-temperature continuous flow reactor. Microfluid. Nanofluid. 2008, 5(3), 411–416. 30. Ruzicka, J.; Hansen, E. H. Flow injection analyses. 1. New concept of fast continuous-flow analysis. Anal. Chim. Acta 1975, 78(1), 145–157. 31. Ruzicka, J.; Hansen, E. H. Flow injection analyses. 10. Theory, techniques and trends. Anal. Chim. Acta 1978, 99(1), 37–76. 32. Petrlova, J.; Mikelova, R.; Stejskal, K.; Kleckerova, A.; Zitka, O.; Petrek, J.; Havel, L.; Zehnalek, J.; Adam, V.; Trnkova, L.; Kizek, R. Simultaneous determination of eight biologically active thiol compounds using gradient elution-liquid chromatography with Coul-Array detection. J. Sep. Sci. 2006, 29(8), 1166–1173. 33. Alonso, A.; Almendral, M. J.; Curto, Y.; Criado, J. J.; Rodriguez, E.; Manzano, J. L. Determination of the DNA-binding characteristics of ethidium bromide, proflavine, and cisplatin by flow injection analysis: usefulness in studies on antitumor drugs. Anal. Biochem. 2006, 355(2), 157–164. 34. Chen, X. W.; Chen, M. L.; Chen, S. A.; Wang, J. H. Flow-based analysis: a versatile, powerful platform for DNA assays. Trends Anal. Chem. 2008, 27(9), 762–770. 35. Ivnitski, D.; Abdel-Hamid, I.; Atanasov, P.; Wilkins, E. Biosensors for detection of pathogenic bacteria. Biosens. Bioelectron. 1999, 14(7), 599–624. 36. Borland, L. M.; Kottegoda, S.; Phillips, K. S.; Allbritton, N. L. Chemical analysis of single cells. Annu. Rev. Anal. Chem. 2008, 1, 191–227. 37. Funfak, A.; Br€ osing, A.; Brand, M.; K€ ohler, J. M. Micro fluid segment technique for screening and development studies on Danio rerio embryos. Lab Chip 2007, 7, 1132–1138. 38. Martin, K.; Lemke, K.; Henkel, T.; Grodrian, A.; K€ ohler, J. M.; Metze, J.; Roth, M. Neue Kultivierungstechnologie zur Suche nach seltenen Mikroorganismen. Biospektrum 2006, 7, 743–745. 39. K€ ohler, J. M.; Henkel, Th.; Grodrian, A.; Kirner, Th.; Roth, M.; Martin, K.; Metze, J. Digital reaction technology by micro segmented flow—components, concepts and applications. Chem. Eng. J. 2004, 101(1–3), 201–216. 40. Fukuyama, T.; Rahman, T.; Sato, M.; Ryu, I. Adventures in inner space: microflow systems for practical organic synthesis. Synlett 2008, 2, 151–163. 41. Salazar-Alvarez, G.; Muhammed, M.; Zagorodni, A. A. Novel flow injection synthesis of iron oxide nanoparticles with narrow size distribution. Chem. Eng. Sci. 2006, 61(14), 4625–4633. 42. Teh, S. Y.; Lin, R.; Hung, L. H.; Lee, A. P. Droplet microfluidics. Lab Chip 2008, 8(2), 198–220. 43. Vanapalli, S. A.; Banpurkar, A. G.; van den Ende, D.; Duits, M. H. G.; Mugele, F. Hydrodynamic resistance of single confined moving drops in rectangular microchannels. Lab Chip 2009, 9(7), 982–990. 44. Adzima, B. J.; Velankar, S. S. Pressure drops for droplet flows in microfluidic channels. J. Micromech. Microeng. 2006, 16(8), 1504–1510.

REFERENCES

393

45. Groß, G. A.; Thyagarajan, V.; Kielpinski, M.; Henkel, T.; K€ ohler, J. M. Viscositydependent enhancement of fluid resistance in water/glycerol micro fluid segments. Microfluid. Nanofluid. 2008, 5, 281–287. 46. Donnet, M.; Bowen, P.; Jongen, N.; Lemaıˆtre, J.; Hofmann, H. Use of seeds to control precipitation of calcium carbonate and determination of seed nature. Langmuir 2005, 21, 100–108. 47. Ding, Y.; Gu, G.; Xia, X. H.; Huo, Q. Cysteine-grafted chitosan-mediated gold nanoparticle assembly: from nanochains to microcubes. J. Mater. Chem. 2009, 19(6), 795–799. 48. Dillenback, L. M.; Goodrich, G. P.; Keating, C. D. Temperature-programmed assembly of DNA: Au nanoparticle bioconjugates. Nano Lett. 2006, 6(1), 16–23. 49. Hussain, I.; Brust, M.; Barauskas, J.; Cooper, A. I. Controlled step growth of molecularly linked gold nanoparticles: from metallic monomers to dimers to polymeric nanoparticle chains. Langmuir 2009, 25(4), 1934–1939. 50. Park, S. Y.; Lytton-Jean, A. K. R.; Lee, B.; Weigand, S.; Schatz, G. C.; Mirkin, C. A. DNAprogrammable nanoparticle crystallization. Nature 2008, 451(7178), 553–556. 51. Lee, J. S.; Lytton-Jean, A. K. R.; Hurst, S. J.; Mirkin, C. A. Silver nanoparticleoligonucleotide conjugates based on DNA with triple cyclic disulfide moieties. Nano Lett. 2007, 7(7), 2112–2115.

INDEX

A549 cancer cells, 269 Abbe’s theory of imaging, 368 acid-base reaction, 344 CLSM image, 344 Acrylamide copolymers, 149 N-Acryloylsuccinimide, 149 Activated silica column, 147 Active nanodrug carriers (NDCs), 188 Affinity biosensors, 99, 107 Alginate microparticles/nanoparticles, 246 diagram, 246 Alzheimer’s disease, 1 Amine-terminated nanoparticles, 225 7-Aminoclonazepam (7-ACZP), 110 Aminolysis, 150 Aminopropyl-modified nanoparticles, 226 3-Aminopropyltriethoxysilane (APTES), 24, 139 Ammonium/nitrogen-containing ligands, 363 Anodic partial process, 364 Anodization, 54–56

Anodized aluminum oxide (AAO) membranes, 52 Antibodies, 99–100 Antibody-antigen (Ab-Ag) interactions, 99 Anticancer drugs, 197 camptothecin, 197 daunorubicin (DaunoXome), 197 doxorubicin (Doxil), 197 Aplysia neurons, 24 Apoptosis, 16, 18 Aquagels, See Hydrogels Astrocytes, 3, 18, 20, 26, 28, 36 Atomic force microscopy (AFM), 2, 3, 369, 370 AuNPs, see Gold nanoparticles (AuNPs) a-N-Benzoyl-l-arginine ethyl ester (BAEE), 167 Bio-barcode assay (BCA), 105, 243 implementation, 243 Biodegradable nanoparticles, characteristics, 227

Microfluidic Devices in Nanotechnology: Applications, Edited by Challa S. Kumar Copyright
2010 John Wiley & Sons, Inc.

395

396 Biological/biomedical microelectromechanical system (Bio-MEMS), 74, 187, 191, 193, 209 applications, 200–208 components/flow chart, 192 experimental/computational simulation aspects, 188 family of, 200 goal, 201 use, 187 Bioluminescent reaction, 172 Biopolymer-SiO2 nanocomposite aerogel, 64 Biosensing, 126 pillars for, 273 Biosensors, 49–51 for liver diagnosis, 81 for membrane-based labeled detection, 67 for membrane-based label-free detection, 67–68 impedance/capacitance, 70–71 mechanical detection, 73–74 optical detection, 71–73 potentiometric, 69–70 voltammetry, 68–69 functionality of membrane in, 64–66 Biotin-avidin-biotin coupling, 132 Biotinylated polylysine, 130 Block copolymer, 260 Blood-brain barrier, 2 Bottom-up approach, 257, 260 Bovine knee chondrocytes (BKCs), 227 fixed charge density, 227 zeta potential, 227 Bovine serum albumin (BSA), 6, 101 Bradford protein assay, 222 Brownian motion, 384 Caenorhabditis elegans, 4 Capacitance, 70, 72, 74 Capillary electrophoresis (CE), 131, 213 ESI/TOF-MS analysis of protein mixtures, 159 microreactor for peptide mapping, 167 nanomaterial applications, 214 pseudostationary phase, 214 separation buffer additives, 214 using nanoparticles, 213 Carbon nanotubes (CNTs), 91, 195, 226, 238 bovine serum albumin-conjugated, 238

INDEX

multiwalled nanotubes (MWNTs), 195, 226, 227 single-wall nanotubes (SWNTs), 195, 226 use, 226 Carbonyldiimidazole (CID) activation, 151 Catalytic nanosystems, 387 Cathodic partial process, 364, 365 role, 365 Cell-based biosensor, 14 Cellomics, pillars for, 268 cell characterization, 271–272 cell cultivation, 269–271 cell trapping, 268 Cell viability, 10 Ceramic nanoparticles, 195 Cetyltrimethylammonium bromide (CTAB), 222 Chemically sensitized field effect transistor (CHEMFET), 99 Chemical vapor deposition (CVD), 261 Chemiluminescence (CL), 72 detection, 96 Chip-based immunoassays, 139 Chitosan (CTS), 130, 155 Cholera toxin subunit B (CTB), 108 Circulating tumor cells (CTCs), 268 isolation, 268 pillar-based trapping, 269 Clonazepam (CZP), 110 CMOS-based semiconductor industry, 256 Colloidal self-assembly strategy, 241 Commercial multielectrode arrays, 34 Competitive binding assay formats, 103 Composite nanoparticles, 383, 384 architecture, 383 continuous-flow synthesis, 383 formation process, 384 Computational fluid dynamics (CFD) simulations, 285 Computer-aided design (CAD), 7 Conducting polymer nanowires (CPNWs), 96 Confocal laser scanning microscopy (CLSM), 341–343 Continuous-flow processes, 362 Conventional lithography methods, 263 Conventional silica-based monolithic columns, 147–148 Coulometric efficiency, 247 definition, 247

INDEX

397

CoventorWareTM software, 286 Critical nanoparticle concentration (CNC), definition, 225 Cross-linking agent, concentration, 245 Crystal lattice, metallic properties, 362 Curie temperature, 141 Current density, 261 b-Cyclodextrin (CD)-modified nanoparticles, 227 Cyclohexene, 299–300 Cytochrome P450 (CYP)-based immobilized enzyme reactors, 170 based IMERs in drug metabolism, 171

protein arrays, 267 sequencing, 4 DNA biosensors, 113–114 electrochemical, 116–118 hybridization, 115 detection, 115–116 reaction, 115 optical, 118–120 probe, 114–115 Drug, 198 conventional nanoparticle carriers, 198 Drug delivery systems (DDSs), 188, 208 Drug-loaded system, 198

Dendrimers, 196 properties, 196 water-soluble dendrimer, 196 Designs and devices. See also PDMS microfluidic design for neuroscience applications for generating gradients of green and red dyes, 13 gradient-generating designs, 12–13 integrated electrophysiology, 14–15 Diagnostic biosensors, See Biosensor Dialdehyde activation, 150 Didodecyl dimethylammonium bromide (DDAB), 222 capped gold nanoparticles, 222 Diethylene glycol (DEG), 338 Differential centrifugal sedimentation (DCS), 371 Diffusion coefficient, 285, 342, 343 Dipole-dipole repulsion, 145 Direct simulation Monte Carlo (DSMC), 190 Disuccinimidyl suberate (DSS), 154 DNA, 216, 217, 229, 240 bioassay, nucleic acids structures, 111–113 dsDNA fragments, 242 electrophoresis, 217 hybridization, 111 mobility, 216 molecules, 217 role, 266 preconcentration, 239 on-chip preconcentration method, 239 separation/electrochemical detection, 240 probes, 114–115

Electrical biosensors, See Biosensors Electrochemical DNA biosensor, 116–118 Electrochemical immunosensors, 104–109 Electrochemical impedance spectroscopy, 231 Electrochemical transduction, 99 Electron beam interactions, 370 Electron beam lithography, 257, 266 Electron microscopic methods, 369, 370 scanning electron microscope (SEM), 369 transmission electron microscope (TEM), 369 Electroosmotic flow (EOF)-driven system, 214, 217, 218, 225 Electrophoresis, microfluidic chip, 266 Electrophysiology and mass spectrometry, 24 and microfluidics, 26–28 change in, 33 effect of toxins on, 32 integrated, 14–15 Electroporation, 16 Electrospray ionization mass spectroscopy, 162 Electrostatic forces, 101 Entropy balance equation, 206 Enzyme immobilization techniques, 126–127 biospecific (affinity) adsorption, 130–131 covalent immobilization, 127–129 layer-by-layer assembly, 130 physical adsorption, 129–130 Enzyme immunoassay (EIA), 102 Enzyme-linked immunosorbent assay (ELISA), 102

398 Enzymes, 125 immobilized magnetic nanoparticles, 145 modified fused silica microreactor, 164 Escherichia coli, 64, 107 biosensor, 75–76 Esoteric driving forces, surface tension, 190 Extinction coefficient, 367 Extracellular matrix (ECM), 256 Fabricated high-density dot/pillar arrays, 259 SEM images, 259 Fabricated urea biosensor, 82 Faraday’s constant, 247 Faraday’s equation, 247 Fast diffusion, 3 Femtoliter segment, 388 Fiber-optic localized plasma resonance (FO-LPR) 109 microfluidic chip, 109 Field-amplified sample injection (FASI), 239 Field-amplified sample stacking (FASS), 239 Flow cytometry, 4 Fluidoparticle informatics, 389 schematic presentation, 389 Fluorescein, chemical structures, 343 Fluorescence detection, 273 Fluorescence resonance energy transfer (FRET), 71 Fluorescent immunoassay (FIA) analysis, 102, 375 Fluorinated ethylene propylene (FEP) membrane, 18 Focused ion beam (FIB), 61 Gel-based electrophoresis, 265 Gene therapy, 29–30 Genomics/proteomics, pillars for, 264–268 biomolecule preconcentration, 264–265 DNA and protein arrays, 267–268 DNA stretching and separation, 265–267 g-forces, 372 Glicidoxypropyltrimethoxysilane, 139 Globulins, 100 Glucose biosensor, 79, 80 Glucose detection, See Porous membranebased biosensor Glucose oxidase (GOx), 80 biosensor, 79 g-Glutamyl transpeptidase, 147

INDEX

Glutaraldehyde (GA), 24, 144 Glycidoxypropyltrimethoxysilane (GLYMO), 144 Glycidyl methacrylate copolymers, 149 GOD-CPG particles, 141 Goethite (a-FeOOH), nanoparticles, 349 Gold nanoparticles (AuNPs), 95, 141, 214, 221–225, 229–232, 244, 367 core/shell, 370, 375 bimodal size distribution, 377 flow-through microspectrophotometry, 375 preparation, 380 TEM image, 370 dark-field image of, 369 DCS spectrum, 371 dispersions, 367 electrophoretic mobility, 236 nanomosaic network, 231 percentage, 381 probes, 243, 244 SEM image, 369 size distributions, 373, 374 thiol surface displacement, 243 three-dimensional AFM image, 371 use, 214, 229 Gold pillar arrays, 262 Green fluorescent protein (GFP), 227 3D structure, 228 mutant, 227 Growth factor, 28–29 Harrison’s approach, 242 Heat flux, 205 influence on minimal uniformity length, 205 HeLa cells, 16, 270 High refractive index, 362 High-throughput method, 257 High-throughput screening, 269 Homogeneity, improvement, 382 Horseradish peroxidase (HRP), 132 Human genomics project, 264 Hybridization, 115–116 Hybrid nanofluidic-microfluidic devices, 243 characteristic features, 243 Hydrogels, 196 Hydrogen production. See also Platinum catalyst

399

INDEX

from SRM, 301–302 non-noble nanocatalysts for SRM reactions to, 303–307 Hydrophobic-hydrophobic bonding interaction, 235 Hydrophobic PMMA microchannels, 135 3-Hydroxybutyrate dehydrogenase, 139 Hydroxypropyl cellulose (HPC) matrix, 239 Hydroxysulfosuccinimide (sulfo-NHS), 133 Iminodiacetic acid (IDA), 129 based impedance biosensor, 108 Immobilization techniques, 64 Immobilized capillary acetylcholinesterase (AChE) reactor, 169 Immobilized enzymes, 168 biosensors, 168–171 Immobilized glucose oxidase (GOD) particles, 140 Immobilized microfluidic enzymatic reactors (IMERs), 126 application, peptide mapping, 156–158 capillary electrophoresis system, 162–168 liquid chromatography system, 160–162 MS, 158–160 online coupling, 164 Immobilized protease P, 147 Immobilizing sites, of proteins, 127 Immunoassays, formats, 100, 102–103 Immunoglobulin G (IgG), 100 Immunosensors, 103–104 Impedance/capacitance, 70–71 Indium tin oxide (ITO) electrodes, 21, 274 glass substrates, 274 Integrated sensors, 15 Interdigitated ultramicroelectrode array (IDUA), 109 Interference localized surface plasmon resonance (iLSPR) biosensor, 109 Iron oxide (Fe2O3) nanoparticles, 323 g-Fe2O3 nanoparticles, 339, 345 for magnetite and maghemite nanoparticles, 333 coprecipitation, 333–335 influence of pH, 336 polyol process, 337–338

synthesis in constrained environments, 338–339 thermal decomposition, 336–337 synthesis, 325 metallic cations and polycondensation, 326–327 precipitation process, kinetic steps, 328–333 Iron oxyhydroxide nanoparticles, 323–324 microdroplet reactor, synthesis in, 348–349 microfluidic synthesis, 339 synthesis of a-FeOOH nanoparticles, 349–352 synthesis of g-Fe2O3 nanoparticles, 339–348 Isoelectrical focusing (IEF), 265 Knudsen number, 189 Label-free biosensor, 49–50 Lab-in-a-cell technology, 4 Lab-on a-chip devices, 214 mass production, 214 Labyrinth like structure, 145 Laminar flow, 3, 12 Langmuir adsorption isotherm model, 235 Langmuir kinetic models, 243 Laser-Doppler electrophoresis, 238 Laser-induced fluorescence (LIF), 110 detection, 131 Lattice Boltzmann method (LBM), 190 Lennard–Jones (L-J) potential, 191 Liftoff technique, 260 Linear polyacrylamide (LPA), 224 Lipid-based liquid crystalline nanoparticles, 227–228 Liposomes, 97, 196, 197 property, 197 Lithography, 59–60 Localized surface plasmon resonance (LSPR), 95 Locked nucleic acids (LNA), 99, 114 Lodestone, 323 Lower critical solution temperature (LCST), 233 Low-pressure chemical vapor deposition (LPCVD), 62 LRM55 cells, 270

400 Macroscopic heterogeneous chemistry, 364 Magnetic anisotropy energy, 325 Magnetic cell trapping, 268 Magnetic microparticles (MMPs), 105 Magnetic nanoparticles, 232–235 Magnetite, 323 Fe3O4 nanoparticles, 334 Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), 156 probe, 153–154 TOF (time-of-flight) mass spectrometry, 235, 368 for the digestion and peptide mapping, 155 Mass spectrometry (MS), 158 Membrane-based biosensor, 48 Membrane-based technology, 52. See also Nanoporous membrane Mercaptopropyltrimethoxysilane (MPTMS), 117 Metal-chelated complexes, 338 Metal nanoparticles, 362, 364, 366, 382, 384 colloidal solutions, 366 core/shell, generated by microsegmented flow use, 379–381 formation mechanism, 363 nucleation, 363–364 particle growth, 364–365 surface capping, 365–366 growth, 364 interaction, 362 plasmon absorption, 366 synthesis, 372, 375 in homogeneous fluids, 372–375 in microreaction technology, challenges, 381 under segmented flow conditions, 375–381 overwhelming number, 382 Methyl methacrylate (MMA), 155 Methyl methacrylate-sulfopropyl methacrylate (MMA-SPM) nanoparticles, 227 Micellar electrokinetic chromatography (MEKC), 215, 217–219 advantage, 219 separation principle, 218 Micelle, 197 vs. drug carrier systems, 197

INDEX

Michaelis–Menten constant (Km), 137, 171 Microchannels and microfluidics, 189 blockages of catalyst active sites in, 311 coated, 217 degree of microbeads immobilized on, 140 enzymatic reactions in, 126 flow rates in, 373 fluid-particle flow in, 200, 206 high surface area of, 309 nanocatalyst deposition on, 294 nanodrug transport phenomena in, 189 network comprising of, 84 network of omega-shaped, 288 optical image of, 140 packing with, 137 selective deposition of catalyst in, 298 silica and alumina on, 299 turbulent flows in, 285 unusual behavior of fluid traversing, 3 Microchip-based bio-barcode assay, 243 Microchip electrophoresis (MCE), 213 microfluidic devices, 214–219 advantages and applications, 214 separation techniques, 215 size, 213 using nanoparticles, 213 Microchip enzymatic microreactor, 232 Microchip for cancer biomarkers, 106 Microchip gel electrophoresis (MGE), 215 separation of DNA, 216 Microchip-immobilized magnetic enzyme reactor system, 235 compatibility, 235 Microchip zone electrophoresis (MZE), 215–216 Microdroplet reactor, 348 Microelectromechanical systems (MEMS), 261 rf-interrogated biosensor, 74 Microemulsions, 197 Microfluidic channels, 255 application examples, 264–268, 274 pillars for biosensing, 273–274 pillars for cellomics, 268–272 pillars for genomics and proteomics, 264–268 introduction, 255–257 other fabrication aspects, 261–264

INDEX

integration into microdevices, 263 material choice, 261–262 surface fuctionalization, 262 patterning techniques, 257–261 auto assemblage, 260 growth, 261 lithography and related techniques, 257–259 pillars and pillar arrays, 255 Microfluidic chips, 15, 29, 255 chip based biosensor, 108 combining a biobarcode with, 119 fabrication, 7 FO-LPR, 109 for analyzing benzodiazepines, 110 for electrophoresis, 266 for microelectronic circuitry, 2 glass-based, 273 ideal systems for, 256 interfaces with ESI, 159 manipulation of Ca-alginate microspheres using, 245 miniaturization of the sol-gel structures on, 132 multiple active sites on, 51 neurophysiology experiments using, See Neurophysiology pillar arrays integrated, 274 PMMA for enantioseparation, 238 to improve neuron survival, 11 Microfluidic devices, 187, 191, 214–219, 266, 268 advantages and applications, 214 Bio-MEMS system, 187 categories, 191–194 development, 191 lab-on-a-chip (LOC) systems, 187 mechanics, 191 microfluidics and microsystems, 189–190 microsystem modeling assumptions, 190–191 R&D areas, 189 separation techniques, 215–219 DNA separation by microchip gel electrophoresis, 216–217 micellar electrokinetic chromatography, 217–219 microchip zone electrophoresis, 215–216 use, 191

401 Microfluidic platforms, 4 architectural designs, 8–9 Microfluidics system, 83, 201, 241, 388, 389 advantages, 389 analysis systems, 264 based cell studies, 4 based colloidal self-assembly technique, 239 based culture platform, 31 based microchips, 231 biosensor systems, 82–84 chambers for gene therapy, 6 for MALDI protein analysis, 160 in bioassays, 93–94 operation of information, 388 self-assembled colloidal arrays, fabrication/characterization, 241 sorting device, 233 synthesis, development, 381 technology, defined, 2 to synthesize cDNA, 17 Micro-high-performance liquid chromatography (m-HPLC), 156 Micromixer baffle-slit, 203 influence, 202 pumping power, 202 Micromixing, 366 omega structure, technique, 285–289 principles and types, 284–285 Microreaction technology, 364, 366, 381, 385, 387 advantages, 382 approaches, 366 development, 388 Microreactors, 361, 381 based nanoparticle synthesis, 390 construction, materials for, 289–290 for gas-to-liquid technology, 309–314 iron-cobalt mixed catalysts, 309–312 ruthenium as promoter, 312–314 for kinetic studies, 171–172 for steam reforming of methanol, 302 hydrogen production and purification, 301 metal nanoparticle synthesis, 361 technical development, 381 Microscaffold system, 19 Micrototal analysis systems (m-TAS), 214 Mie plasmon, 366

402 Mie theory, 366 Migration time window, 218 Mild anodization (MA), 54 Mix-and-match lithography methods, 259, 263, 264 advantages, 263 Mobility apparent mobility, 221 electroosmotic mobility, 221 Molecular dynamics simulation (MDS), 190 Monolithic phases, 145–147 Multichannel interfaces, 388 microfluidic strategies for, 388 Multielectrode arrays (MEAs), 5 Multiple sclerosis, 1 Nanocatalyst calcination temperature, 304 deposition on microchannels, 294 coating methods, 294–297 CVD technique, 295 electrochemical deposition, 295 pretreatment of substrate, 294–297 PVD vs. sol-gel method, 297–301 sol-gel method, 296 in GTL technology, 284 non-noble, for SRM reactions to produce hydrogen, 303 particle sizes of, 312 screening, parallel microreactor system for, 314–318 SRM reactions conducted over, 305 Nanocrystalline iron (II, III) oxide, 338 Nanocrystals, 62, 75, 76, 96, 195, 330, 336, 337 Nanodevice system, 231 Nanodrug delivery systems, 187, 194–200 bio-MEMS applications, 200–208 device optimization, 203 nanofluid flow simulations, 201–203 conclusions/future perspective, 208 introduction, 187–190 microfluidic devices, 188–194 categories, 191–194 microfluidics and microsystems, 189–190 microsystem modeling assumptions, 190–191 strategies, 194

INDEX

Nanodrugs, 194–200 carriers, passive drug targeting, 199 colloidal soft matter, 196–198 development, 211 nanodrug carriers, desirable characteristics, 198 solid nanoparticles, 194–196 targeting, 198–200 Nanoelectrodes, 2 Nanoimprint lithography, 258, 263, 266 Nanolaths, TEM image, 352 Nanomaterials, application, 382 Nanometric ferro/ferrimagnetic particles, 324 Nanoneedles, 2 Nanoparticle-mediated capillary electrophoresis, 220 carbon nanotubes, 226 gold nanoparticles, 221–225 lipid-based liquid crystalline nanoparticles 227–228 silica nanoparticles, 225 TiO2 226 Nanoparticle-mediated electrophoresis (NME), 220, 221 Nanoparticle-mediated microchip electrophoresis, 228 Au microelectrodes, in-channel modification, 247 AuNP-mediated on-chip preconcentration, 239 biopolymer micro/nanoparticles, microfluidic fabrication, 244 carbon nanotubes, 238–239 colloidal Au self-assembly, 239–243 gold nanoparticles, 229–232 magnetic nanoparticles, 232–235 microchip-based bio-barcode assay, 243 polymer nanoparticles, 236–238 surface displacement reactions, 243 TiO2 nanoparticles, 236 zeolite nanoparticles, 235–236 Nanoparticles (NPs), 213, 219, 362, 364, 383 aggregation, 248 aggregates formation, 385 application in CE and MCE, 219 nanoparticle-mediated capillary electrophoresis, 220–228

403

INDEX

nanoparticle-mediated microchip electrophoresis, 228 based catalytic assemblies, construction, 387 based/catalytic nanomachines, vision, 386 based information handling system, 388 carrier, 198 size, 198 surface charge, 198 colloidal suspensions, characterization, 247 definition, 213 filled capillary electrophoresis, 222 formation, 364 growth, 365 plasmon absorption, 374 polybutylcyanoacrylate (PBCA), 227 product/process characterization, 366–372 differential centrifugal sedimentation, 371–372 optical microscopy, 368–369 spectrometry, 366–368 ultramicroscopy, 369–371 properties, 220 regional functionalization, 384 shape, 365 silica-based, 96 small dimensions, 362 surface capping, 365 surface state, 365 synthesis, 373, 383 advantages, 373 TEM image, 347 three-dimensional self-assembling, 384 used in bioassays, 94–97 Nanopillars, 260 arrays, 267 Nanoporous membrane, 47 aluminum oxide membrane (AOM), 84 based microfluidic biosensors, application, 50–51 efficient size sorting, 52 molecular sorting, design considerations for, 52–53 fabrication and integration into microfluidic device, 54 anodization, 54–56 focus ion beam etching, 61–62 ion track etching, 56–58 lithography, 59–60

phase separation, 58–59 rapid thermal annealing (RTA) technique, 62 sol-gel technology, 63–64 need for, 51–52 Si3N4/SiO2 membranes, 82 types of, 53 Nanotechnology hardware, 3 Nanowires, 2, 91, 272 Navier–Stokes equation, 286, 342 Negative dielectrophoretic (nDEP) force, 269 Nerve growth factor (NGF), 29 Nervous system, 1 Neural lineage cells, 3 Neural prosthesis development, 18 Neuroblastoma-glioma hybrid cells, 16 Neurodegenerative processes, 1 Neurons, 1 Neuropeptides, 20 Neurophysiology experiments using microfluidic chips, 19 axonal isolation, 30, 32–33 cell separation tools, 19–20 electrophysiology, 26–28 gene therapy, 29–30 growth factor effects, 28–29 neuropeptide release, 20–23 physical and chemical guidance cues, 23–26 Newton’s second law of motion, 191 N-hydroxysuccinimide (NHS), 134 Nickel-nitrilotriacetic acid (Ni-NTA), 139 N-methyl diethanolamine (NMDEA), 338 NMR spectroscopy, 16 Noncross-linking interaction mechanism, 231 Nonspherical particles, 381 generation, 382 morphological classes, 381 yield improvement, 381 Nucleation mechanism, 363–364, 383 definition, 383 centers, formation, 374 Nucleic acids structures, 111–113 Oligonucleotides (ODNs), 114 Omega channel microreactors, 286–288 On-chip microreactor, 235 Open-channel reactor designs, 132

404 Open-ended microchannel, 216 Optical detection techniques, 71–73 Optical DNA biosensor, 118–120 Optical immunosensor, 109–111 Optical lithography, 258 Organic polymer monoliths, 148–152 Organogels, 197 Organomercaptans, 243 surface displacement kinetics, 243 Ovalbumin (OVA), 101 Oxidase-based amperometric biosensors, 78 Packing microchannels, with micro/ nanoparticles, 137–138 magnetic supports, 141–145 nonmagnetic supports, 138–139 organic supports, 139–141 Parkinson’s disease, 1 Particles, 386 assembly types, 387 based information processing, 388 microfluidic strategies for, 388 capture and release scheme, 234 stabilization, 386 Particle-supported systems, 371 dye pigment suspensions, 371 Passive multifunctional nanoparticle systems (MFNPSs), 188 Patch clamp array, 15 Patterning techniques, 257–261 auto assemblage, 260 growth, 261 lithography and related techniques, 257–259 Peclet numbers, 201, 202 Penicillin G acylase (PGA), 137 Pepsin, 156 Peptide mapping, 126 microfluidic enzymatic reactors for, 132 Phase transfer processes, 366 Photonic crystal sensor, 274 Fourier transform infrared spectroscopy, 274 Photopolymerization, 148 Physical vapor deposition (PVD), 261 Platinum catalyst for preferential oxidation of CO in hydrogen, 307–309 in PDMS, 11

INDEX

Polyacrylamide, 149 Polycarbonate (PC), 52, 58, 94 Poly(diallyldimethylammoniumchloride) (PDDA), 134 filled CE, 223 made device, 134 Poly(diallyldimethylammonium chloride) (PDDC), 221 adsorption, 221 Polydimethylsiloxane (PDMS), 6, 82, 258, 271 architectural designs, 8–9 based microdevices, 246 characterstics, 6–7 chip fabrication protocol, 7–8 culture plate, 18 elastomer, 94 growth chambers, 34 microchannels, 21 microchip channel, 229 microfluidic design, 6–7 practical considerations and limitations, 10–12 tools, 9–10 use, 258 Poly(ether ether ketone) (PEEK), 147 Polyethylene, 94 terephthalate, 58 Poly(ethylene glycol) (PEG) hydrogel, 7, 270 microfluidic chip, 84 nanopillars, aggregated cardiomyocytes, 271 Poly(ethylene oxide) (PEO), 222, 229 Poly(ethylene terephthalate) (PET) microchannels, 141 microfluidic chip, 230 photoablated microchannel, 232 SEM images, 232 Poly(ethylenimine) (PEI), 29 based DNA, 30 Poly(glycidyl methacrylate-co-ethylene dimethacrylate), 146, 150 Polyimide (Kapton) membranes, 58 Poly-L-lysine (PLL), 24 Polymerase chain reaction (PCR), 264 Polymer-based microparticles, advantages, 244 Polymer encapsulation, 142–143

INDEX

Polymeric membranes, incorporation into microfluid, 152–153 Polymerization-induced phase separation (PIPS), 59 Polymer (s), 195, 262 advantage, 262 characteristics, 195 fabrication, 262 nanoparticles, 236 pillars, 262 Polymethylmethacrylate (PMMA), 94, 135, 235 Poly-N,N dimethylacrylamide (PDMA), 224 Polyols, 337 Polypropylene, 94 Polystyrene, 94 Polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) copolymer, 53 Polysulfone (PSf) porous films, 60 Poly(vinyl alcohol) (PVA), 139 Poly(vinylidenefluoride) (PVDF) membrane, 59, 153 Polyvinyl pyrrolidone (PVP), 373 Porous membrane-based biosensor as cholesterol biosensor, 80 E. coli biosensor, 75–76 for detection of S. enteritidis, 76–77 for treatment of diabetes mellitus, 78–80 for virus detection, 77–78 new diverse sensors, 81–82 Post-translational modified proteins, 162 Potassiumchloride (KCl), 21 Potentiometric sensors, 69–70 Pressure drop vs. pumping power, 207 Programmed cell death, See Apoptosis Prostate-specific membrane antigen (PSMA), 236 Protein digestion, 139, 153 enzymes for, 157 Protein encapsulation techniques, 132 Proteolysis analysis, 233 Proton exchange membrane fuel cell (PEMFC), 301 Pyrex 7740, 116 Quantum dots (QDs), 2, 96 CdSe-based, 96 Radioimmunoassay (RIA), 102

405 Raman imaging system, 274 chip with integrated patterns, schematic diagram, 274 Raman spectroscopy, 273 Rapid thermal annealing (RTA), 62 Rayleigh criterion, 258 Reactive ion etching technique, 260 Real-time biosensor, 49 Real-time measurements, 48–49 Retention factor, definition, 218 Reynolds number, 3, 191, 202, 209, 210, 284 ratio vs. minimal uniformity length, 203 vs. system entropy generation, 206, 207 Ruthenium dye particles, 16 S/V ratio, 133 S-adenosylhomocysteine (SAH), 135–136 Salmonella enteritidis, 76–77 Scanning electron microscope (SEM), 270, 369, 370 images of microchannel modified with silicalite-1, 142 micrographs of ID monoliths, 163 picture of typical porous structure, 147 Scanning probe microscopic methods, 369 atomic force microscopy (AFM), 369 Schiff’s base, 127 Sedimentation methods, 372 principle, 372 Segmented flow conditions, 375 advantage, 376 mixing in, 376 process homogeneity/product quality, 377 scatterplot, 378 segment formation, 379 specificity, 375 Segmented flow tube reactor (SFTR), 383, 389 Self-assembled metal nanoparticles, 387 electronic/optoelectronic devices, 387 Self-assembled monolayers (SAMs), 12 analysis, 135 Silica, See Silicon dioxide (SiO2) Signal-to-noise (S/N) ratio, 51 Silicalite-1, 141 Silicon based materials, 261 microchannel reactors, 290–294 microstructures, 265

406 Silicon (Continued ) nanopillars, 261 porous membrane-based optical label-free biosensor, 73 Silicon dioxide (SiO2), 109, 242, 261, 262, 299 AFM image, 299 encapsulation, 143–145 nanoparticles, 225 photonic thin-film multilayers, 109 pillar array, SEM images, 267 SEM picture, 299 Silver nanoparticles, 367 core/shell, 370 bimodal size distribution, 377 flow-through microspectrophotometry, 375 preparation, 380 TEM image, 370 dark-field image of, 369 dispersions, 367 film deposition, schematic diagram, 237 Simulation, in vivo tissues with microfluidics, 17–19 Single plasmon resonance methods, 368 Single-stranded DNA (ssDNA), 232 sequencing ladders, 265 Smart drug delivery system (SDDS), 192 Smooth muscle cell, SEM, 271 Sodium dodecyl sulfate (SDS), 218, 373 denatured protein markers, 242 Soft lithography process, 233 Soft nanoimprint lithography, 267 UV lithography, 267 Sol-gel encapsulation, 132 Solid lipid nanoparticles (SLNs), 196, 227 Solid-phase immunoassay, 99 antibody in sensor applications, 100–101 antigen/antibody interaction, 101–102 Solid-state semiconductor techniques, 382 Solid tumors, chemotherapeutic agents, 199 SPION nanoparticles, 352 S-ribosylhomocysteine (SRH), 136 Staphylococcus aureus, 131 Stem cells, 20 Stokes equation, 372 Streaming process, 372 Streptococcal protein G, 131

INDEX

Superparamagnetic nanoparticles, 196, 324 Superparamagnetism, 324 Surface-dependent enzymatic reaction, 268 Surface enhanced Raman spectroscopy (SERS), 273 based biosensing, 273 Surface immobilized biobarcode assay protocol, 106 Surface plasmon-based sensors, 273 Surface plasmon resonance (SPR), 99 Surface-sensitive processes, 383 Surfactants, 366 Synaptogenesis on chip, 22 Synergism, 92 Tetraethyl orthosilicate (TEOS), 63 Tetramethyl orthosilicate (TMOS), 63 Thermal energy (kT), 324 Thermally induced phase separation (TIPS), 59 Tin-doped indium oxide (ITO), 117 TiO2-based nanoparticles, 226, 236 Top-down techniques, 257 L-1-Tosylamido-2-phenylethyl chloromethyl ketone (TPCK)-trypsin, 128, 158 Transmission electron microscope (TEM), 348, 369, 370 Traumatic injuries, 1 Tris(2,20 -bipyridyl)dichlororuthenium(II) (RuBpy), 96 Trypsin encapsulation, 131 in tetramethoxysilane-based hydrogel, 167 immobilized MALDI probe, 153 linked magnetic nanospheres, 155 membrane reactor, 153 modified magnetic nanoparticles, 144 Tumor cell receptors, 199 Tumor markers, detection, 107 Two-dimensional polyacrylamide gel electrophoresis (2D PAGE), 160 Ultraviolet (UV)-modified PMMA surface, 133 UV lithography, 258

407

INDEX

UV nanoimprint lithography (UV-NIL), 258–259 UV spectroscopy, 265 UV-Vis spectrometer, 368

Voltammetric sensors, 68–69 Water-in-oil (w/o) emulsions, 245 X-ray photoelectron spectroscopy, 238

van der Waals bonds, 65 van der Waals forces, 101, 242 Vinylazlactone copolymers, 151–152

Zeolite nanoparticles, 235 Zinc oxide (ZnO), 274