Novel Microﬂuidic Devices Based on a Thermally Responsive ...12419/FULLTEXT01.pdf · x Novel Microﬂuidic Devices Based on a Thermally Responsive PDMS Composite The contribution

Novel Microfluidic DevicesBased on a Thermally

Responsive PDMS Composite

Bjorn Samel

MICROSYSTEM TECHNOLOGY LABORATORY

ISBN 978-91-7178-732-3ISSN 1653-5146

TRITA-EE 2007:31

Submitted to the School of Electrical Engineering,KTH - Royal Institute of Technology, Stockholm, Sweden,

in partial fulfillment of the requirements for the degree of Doctor of PhilosophyStockholm 2007

ii Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

Copyright c©2007 by Bjorn Samel

All rights reserved for the summary part of this thesis, including all pictures andfigures. No part of this publication may be reproduced or transmitted in any form orby any means, without prior permission in writing from the copyright holder.

The copyrights for the appended journal papers belong to the publishing houses ofthe journals concerned. The copyrights for the appended manuscripts belong to theirauthors.

Printed by Universitetsservice US AB, Stockholm 2007.

Akademisk avhandling som med tillstand av Kungliga Tekniska Hogskolan framlaggestill offentlig granskning for avlaggande av teknologie doktorsexamen i elektrisk matteknikfredagen den 7 september 2007 klockan 10:00 i horsal F3, Lindstedtsvagen 26, Stock-holm.

ABSTRACT iii

Abstract

The field of micro total analysis systems (µTAS) aims at developments toward minia-turized and fully integrated lab-on-a-chip systems for applications, such as drugscreening, drug delivery, cellular assays, protein analysis, genomic analysis and hand-held point-of-care diagnostics. Such systems offer to dramatically reduce liquid sampleand reagent quantities, increase sensitivity as well as speed of analysis and facilitateportable systems via the integration of components such as pumps, valves, mixers,separation units, reactors and detectors. Precise microfluidic control for such systemshas long been considered one of the most difficult technical barriers due to integrationof on-chip fluidic handling components and complicated off-chip liquid control as wellas fluidic interconnections. Actuation principles and materials with the advantagesof low cost, easy fabrication, easy integration, high reliability, and compact size arerequired to promote the development of such systems.

Within this thesis, liquid displacement in microfluidic applications, by means ofexpandable microspheres, is presented as an innovative approach addressing someof the previously mentioned issues. Furthermore, these expandable microspheres areembedded into a PDMS matrix, which composes a novel thermally responsive siliconeelastomer composite actuator for liquid handling. Due to the merits of PDMS andexpandable microspheres, the composite actuator’s main characteristic to expand irre-versibly upon generated heat makes it possible to locally alter its surface topography.The composite actuator concept, along with a novel adhesive PDMS bonding tech-nique, is used to design and fabricate liquid handling components such as pumps andvalves, which operate at work-ranges from nanoliters to microliters. The integrationof several such microfluidic components promotes the development of disposable lab-on-a-chip platforms for precise sample volume control addressing, e.g. active dosing,transportation, merging and mixing of nanoliter liquid volumes. Moreover, microflu-idic pumps based on the composite actuator have been incorporated with sharp andhollow microneedles to realize a microneedle-based transdermal patch which exhibitson-board liquid storage and active dispensing functionality. Such a system representsa first step toward painless, minimally invasive and transdermal administration ofmacromolecular drugs such as insulin or vaccines.

The presented on-chip liquid handling concept does not require external actuatorsfor pumping and valving, uses low-cost materials and wafer-level processes only, ishighly integrable and potentially enables controlled and cost-effective transdermalmicrofluidic applications, as well as large-scale integrated fluidic networks for point-ofcare diagnostics, disposable biochips or lab-on-a-chip applications.

This thesis discusses several design concepts for a large variety of microfluidiccomponents, which are promoted by the use of the novel composite actuator. Resultson the successful fabrication and evaluation of prototype devices are reported hereinalong with comprehensive process parameters on a novel full-wafer adhesive bondingtechnique for the fabrication of PDMS based microfluidic devices.

iv Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

Bjorn SamelMicrosystem Technology Laboratory, School of Electrical Engineering,KTH - Royal Institute of Technology, SE-100 44 Stockholm, Sweden.

ABSTRACT v

True knowledge exists in knowing that you know nothing. And in knowing that youknow nothing, that makes you the smartest of all.

Socrates469–399BC, Greek Philosopher of Athens

vi Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

CONTENTS vii

Concentrate all your thoughts on thetask at hand. The sun’s rays do notburn until brought to a focus.

Alexander Graham Bell1847–1922, Scottish-born Canadian

Inventor, Educator, Telephone Pioneer

Contents

Abstract iii

List of papers ix

1 Introduction 11.1 A paradigm shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Introduction to PDMS based microfluidics 52.1 Polymers for microfabrication . . . . . . . . . . . . . . . . . . . . . . . 52.2 Poly(dimethysiloxane) . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3 Soft lithography, rapid prototyping and replica molding . . . . . . . . 62.4 PDMS microfluidics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Sealing of microfluidic structures 113.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.2 Adhesive wafer bonding . . . . . . . . . . . . . . . . . . . . . . . . . . 113.3 Sealing of PDMS structures . . . . . . . . . . . . . . . . . . . . . . . . 123.4 Selective PDMS bonding . . . . . . . . . . . . . . . . . . . . . . . . . . 133.5 Full-wafer adhesive PDMS bonding . . . . . . . . . . . . . . . . . . . . 133.6 Selective full-wafer adhesive PDMS bonding . . . . . . . . . . . . . . . 17

4 Expandable microspheres 214.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214.2 Material characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . 214.3 Expandable microspheres in microfluidics . . . . . . . . . . . . . . . . 224.4 Incorporation of expandable microspheres into a paste . . . . . . . . . 254.5 A thermally responsive PDMS composite . . . . . . . . . . . . . . . . 264.6 Composite characteristics . . . . . . . . . . . . . . . . . . . . . . . . . 27

5 Composite based microfluidic devices 335.1 Design and fabrication of fluidic components . . . . . . . . . . . . . . 335.2 Single-use pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345.3 Single-use valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365.4 Microfluidic networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365.5 Microliter range pumps . . . . . . . . . . . . . . . . . . . . . . . . . . 395.6 Liquid aspiration and dispensing . . . . . . . . . . . . . . . . . . . . . 425.7 Transdermal drug delivery . . . . . . . . . . . . . . . . . . . . . . . . . 44

viii Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

5.8 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

6 Summary of appended papers 49

7 Conclusions 53

Acknowledgements 55

References 57

Paper reprints 69

LIST OF PAPERS ix

Education is not the filling of a pail, butthe lighting of a fire.

William Butler Yeats1865–1939, Irish Poet

List of papers

The presented thesis is based on the following journal papers:

1. A Compact, Low-cost Microliter-range Liquid Dispenser based on ExpandableMicrospheresNiclas Roxhed, Susanna Rydholm, Bjorn Samel, Wouter van der Wijngaart,Patrick Griss and Goran StemmeJournal of Micromechanics and Microengineering, vol. 16, no. 12, pp. 2740–2746, December 2006.

2. A Thermally Responsive PDMS Composite and its Microfluidic ApplicationsBjorn Samel, Patrick Griss and Goran StemmeIEEE/ASME Journal of Microelectromechanical Systems, vol. 16, no. 1, pp. 50–57, February 2007.

3. A Disposable Lab-on-a-chip Platform with Embedded Fluid Actuators for ActiveNanoliter Liquid HandlingBjorn Samel, Volker Nock, Aman Russom, Patrick Griss and Goran StemmeBiomedical Microdevices, vol. 9, no. 1, pp. 61–67, February 2007.

4. Wafer-level Process for Single-use Buckling-film Microliter-range PumpsBjorn Samel, Julien Chretien, Ruifeng Yue, Patrick Griss and Goran StemmeIEEE/ASME Journal of Microelectromechanical Systems, vol. 16, no. 4, pp. 795–801, August 2007.

5. Painless Drug Delivery through Microneedle-based Transdermal Patches featur-ing Active InfusionNiclas Roxhed, Bjorn Samel, Lina Nordquist, Patrick Griss and Goran StemmeIEEE Transactions on Biomedical Engineering, accepted for journal publication,2007.

6. Liquid Aspiration and Dispensing Based on an Expanding PDMS CompositeBjorn Samel, Niklas Sandstrom, Patrick Griss and Goran StemmeSubmitted for journal publication, 2007.

7. The Fabrication of Microfluidic Structures by means of Full-wafer AdhesiveBonding using a Poly(dimethylsiloxane) CatalystBjorn Samel, Mohammad Kamruzzaman Chowdhury and Goran StemmeJournal of Micromechanics and Microengineering, vol. 17, no. 8, pp. 1710–1714,August 2007.

x Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

The contribution of Bjorn Samel to the publications listed above is:

1 part of design, part of writing2 all design, all fabrication, all experiments, major part of writing3 all design, all fabrication, major part of experiments, major part of writing4 all design, all fabrication, major part of experiments, all writing5 part of design, part of writing6 all design, major part of fabrication, major part of experiments, all writing7 all design, major part of fabrication, major part of experiments, all writing

In addition, the author has contributed to the following journal paper :

8. New Materials for Micro-scale Sensors and Actuators An Engineering Review;Section: Expandable microsphere compositesBjorn Samel, Wouter van der Wijngaart, Steve Wilson and Chris Bowen (Eds.)Materials Science and Engineering: R: Reports, vol. 56, no. 1–6, pp. 1–129,June 2007.

The work has also been presented at the following international conferences :

9. Expandable Microspheres incorporated in a PDMS Matrix: a Novel ThermalComposite Actuator for Liquid Handling in Microfluidic ApplicationsBjorn Samel, Patrick Griss and Goran StemmeProceedings of the 12th IEEE International Solid-State Sensors and ActuatorsConference (Transducers), Boston, MA, USA, June 8–12, 2003, pp. 1558–1561.

10. Low Cost Device for Precise Microliter Range Liquid DispensingNiclas Roxhed, Susanna Rydholm, Bjorn Samel, Wouter van der Wijngaart,Patrick Griss and Goran StemmeProceedings of the 17th IEEE International Conference on Micro Electro Me-chanical Systems (MEMS), Maastricht, The Netherlands, January 25–29, 2004,pp. 326–329.

11. Single-use Microfluidic Pumps and Valves based on a Thermally ResponsivePDMS CompositeBjorn Samel, Jessica Melin, Patrick Griss and Goran StemmeProceedings of the 18th IEEE International Conference on Micro Electro Me-chanical Systems (MEMS), Miami, FL, USA, January 30–February 3, 2005,pp. 690–693.

LIST OF PAPERS xi

12. Nanoliter Liquid Handling On A Low Cost Disposable With Embedded FluidActuatorsBjorn Samel, Volker Nock, Aman Russom, Patrick Griss and Goran StemmeProceedings of the 13th IEEE International Solid-State Sensors and ActuatorsConference (Transducers), Seoul, South Korea, June 5–9, 2005, pp. 356–359.

13. Compact, Seamless Integration of Active Dosing and Actuation with Micronee-dles for Transdermal Drug DeliveryNiclas Roxhed, Bjorn Samel, Lina Nordquist, Patrick Griss and Goran StemmeProceedings of the 19th IEEE International Conference on Micro Electro Me-chanical Systems (MEMS), Istanbul, Turkey, January 22–26, 2006, pp. 414–417.

14. Disposable Single-use Pumps based on a Wafer-level ProcessBjorn Samel, Julien Chretien, Ruifeng Yue, Patrick Griss and Goran StemmeProceedings of the 10th International Conference on Miniaturized Systems forChemistry and Life Sciences (µTAS), Tokyo, Japan, November 5–9, 2006, pp.1226–1228.

15. Active Liquid Aspiration and Dispensing based on an Expanding PDMS Com-positeBjorn Samel, Patrick Griss and Goran StemmeProceedings of the 20th IEEE International Conference on Micro Electro Me-chanical Systems (MEMS), Kobe, Japan, January 21–25, 2007, pp. 565–568.

16. A Laterally Operating Liquid Aspiration and Dispensing Unit based on an Ex-panding PDMS CompositeBjorn Samel, Niklas Sandstrom, Patrick Griss and Goran StemmeProceedings of the 11th International Conference on Miniaturized Systems forChemistry and Life Sciences (µTAS), Paris, France, October 7–11, 2007, ac-cepted for publication.

The author’s work has also been presented at the following workshops :

17. Novel Technology Platforms with Expandable MicrospheresNiclas Roxhed, Bjorn Samel, Patrick Griss, Gran StemmeProceedings of the 5th Micro Structure Workshop (MSW), Ystad, Sweden, March30–31, 2004, pp. 57–60.

18. A Low-cost Disposable Nanoliter Liquid Handling System with Embedded FluidActuatorsBjorn Samel, Volker Nock, Aman Russom, Patrick Griss and Goran StemmeProceedings of the 6th Micro Structure Workshop (MSW), Vasteras, Sweden,May 9–10 , 2006, pp. 55.

19. Miniaturized Drug Delivery System for Painless Trans- and Intradermal Injec-tionsNiclas Roxhed, Bjorn Samel, Lina Nordquist, Patrick Griss and Goran StemmeProceedings of the 6th Micro Structure Workshop (MSW), Vasteras, Sweden,May 9–10 , 2006, pp. 31.

xii Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

1 INTRODUCTION 1

Imagination is more important thanknowledge, for knowledge is limitedwhile imagination embraces the entireworld.

Albert Einstein1879–1955, German-born American

Physicist

1 Introduction

1.1 A paradigm shift

Historically, engineers have been fascinated by a sole maxim above all others thatfired their imagination by far and beyond. Equivalent to the Olympic motto “Citius,Altius, Fortius”, the aim has always been trying to surpass prevailing ideas in order tocreate faster cars, airplanes and trains, higher bridges and taller skyscrapers as well asstronger trucks and boats. Scientists and engineers continuously cross the line of whatis feasible at present by inventing new materials and fabrication techniques. Biggerwas, until not long ago, thought to be better when a new trend made an appearance.In today’s society we feel the urge to possess handy gadgets of ever smaller size,such as cell phones, computers and cameras – at its best all-in-one. Evidently, sizereduction has become an important aspect of our lives.

Similar to the progress in the semiconductor industry, other research areas haveexperienced a great effort of miniaturization as well, among them, the fields of an-alytical chemistry and biology. The research field that has most impact on thesedisciplines, from a size reductive point of view, is referred to as microfluidics anddeals with the science and technology of systems that have to do with the behavior,control and manipulation of small amounts of liquid volumes, typically in the range ofpicoliters to nanoliters and microliters. The miniaturization of analytical techniquesby means of microfluidics is associated with several beneficial aspects; the ability toreduce liquid sample and reagent quantities as well as the overall size of the usedanalytical instrumentation; the capability to conduct separations and detections withincreased sensitivity at high resolution within very short time. In addition, microflu-idics potentially allows parallel analyses of multiple samples on a single chip and offersthe possibility of portable as well as disposable devices.

To achieve this, microfluidic devices are produced by means of microfabricationtechniques [1–3]. In early developments, conventional materials such as silicon andglass were used along with their associated fabrication techniques which were adaptedfrom the semiconductor industries [4–7]. These primarily utilized standard pho-tolithography, wet and dry etching and various other micromachining techniqueswhich all require the use of a cleanroom facility and expensive laboratory equip-ment. The use of silicon as a material for the fabrication of microfluidic systems isnot only costly but also inappropriate since it does not allow conventional optical

2 Novel Microfluidic Devices Based on a Thermally Responsive PDMS Composite

detection methods to be used due to its opaqueness to visible and ultraviolet light.Glass on the other hand will supposedly be used in emerging developments due to itsbeneficial material characteristics, such as optical transparency, compatibility withsolvents, surface stability as well as its established and comprehensive fabrication andapplication techniques [8–10]. Nonetheless, microfluidic systems must contain a seriesof integrated components such as pumps, valves, mixers, separation units, reactorsand detectors to conduct the standard analytical operations. The fabrication of these,using materials such as silicon and glass makes this a particularly delicate and costlytask.

Today, techniques and materials adapted from the polymer forming industries,such as replica molding, injection molding and hot embossing are widely used withinthe filed of microfluidics. Yet, the conventional micromachining techniques are of-ten used in the context of fabricating a master from which polymers can be repli-cated [11, 12]. Using polymers is very attractive because they offer the possibility tobe engineered within a wide range of physical and chemical properties. Furthermore,they allow for mass manufacturing of polymeric microfluidic devices and are cheapand easy to duplicate when using the above mentioned fabrication techniques.

Researchers and scientists active in the microfluidic field have committed them-selves to developing a device which is able to perform numerous tasks that are gen-erally conducted in a standard analytical laboratory [13]. Their ultimate vision is toshrink down such a laboratory onto a single chip, eventually about the size of a palm.Such a “lab on a chip” (LOC) is envisioned to be easy to use and could at the sametime be operated remotely by a patient instead of visiting a hospital in order to geta diagnostic test carried out. The devices will be capable of diagnosing multiple con-ditions or pathogens at point-of-care using only a small bodily fluid sample and theresults could be available shortly after. The possibility to use such LOCs in a vastlyremote environment has been perceived as a powerful application, not at least to ad-dress the challenges for global health which involve the diagnosis and measurementof patients’ health statuses on an individual as well as a quantitative basis [14–17].Besides, such portable chemical detection systems are also of great interest to safety,forensic and defense authorities.

The field of microfluidics, maiden as it is, still lacks the indispensable to revealits powers concealed within; complete integrated systems that may be used by the“Average Joe” to monitor each and everybody’s condition in a home environment.Holding great potential – conceivably even revolutionary – yet a lot of effort is requiredin the future to make it live up to its expectations and literally shrink down thecentralized “lab on a chip”.

1.2 Organization

The thesis is organized as follows. Chapter 2 provides a brief overview of polymerbased microfluidics, with particular emphasis on microfluidic device fabrication usingPoly(dimethysiloxane).

An overview of applicable wafer bonding methods for sealing of microfluidic devicesis given in Chapter 3, along with the most common techniques that are used topermanently seal PDMS based microfluidic devices. The bulk of this chapter describes

1 INTRODUCTION 3

a novel technique for sealing of such devices adhesively by means of PDMS curing-agent. While this approach originally was pursued with the purpose of eliminatingbonding issues, it may be useful in complex and novel applications due to its manyprocessing advantages compared to alternative methods. Results of adhesively andselectively sealed microstructures, as well as process parameters thereof are discussedfor various applications.

Chapter 4 provides an introduction to the use of expandable microspheres withinthe microfluidic field, which has culminated in the development of a thermally re-sponsive PDMS composite. Fabrication and application techniques, as well as basiccharacteristics of this novel PDMS composite, are discussed and presented herein.

The successful demonstration of functional microfluidic components and devicesbased on the new thermally responsive PDMS composite are presented in Chapter 5.The diversity of components presented in this chapter ranges from valves, to variouskinds of pumps, to mixers and to a complete lab-on-a-chip platform covering themanipulation of nanoliter to microliter volumes. Detailed results of prototype devicesare illustrated along with the fabrication issues. This chapter moreover shows afeasibility study of a complete patch-like transdermal drug delivery system basedon the PDMS composite, which is electrically controllable and straightforward tofabricate. The proposed device represents a first step toward painless and convenientadministration of macromolecular drugs, such as insulin or vaccines.


2 INTRODUCTION TO PDMS BASED MICROFLUIDICS 5

It is easy to build a philosophy – itdoesn’t have to run.

Charles F. Kettering1876–1958, American Engineer, Inventor

2 Introduction to PDMS based microfluidics

2.1 Polymers for microfabrication

Microfluidic systems are growing increasingly attractive for the field of micro totalanalysis systems (µTAS) [4, 7, 13, 18] and aim at miniaturized and integrated lab-on-a-chip applications toward drug screening, drug delivery, cellular assays, proteinanalysis, genomic analysis and handheld point-of-care diagnostics [19–22]. As earlierdescribed, such systems potentially allow to reduce sample and reagent volumes, in-crease sensitivity and speed of analysis, offer portability and reduce production coststhrough high volume production and batch fabrication.

In such microfluidic systems various components such as pumps, valves, mixersand reactors are integrated on a single microfluidic chip in order to manipulate liq-uids in a controlled manner. In early developments materials well known from theconventional semiconductor industry, such as silicon and glass, were used [4–6,23,24].However, the complexity of devices based on these materials and their associatedfabrication techniques makes integration into such systems difficult and costly. Therequirements on microfluidic components are various; they should be easy to inte-grate, straightforward to fabricate, easy to use, robust, free of dead volume and cheapto duplicate.

Therefore, polymer microfabricated devices based on new materials and actuationmethods have gained increased attention from the MEMS society. Using polymers isvery attractive because of a number of reasons. They are cheap compared to single-crystalline silicon and easy to process allowing various molding as well as embossingtechniques. The fabrication of polymer based devices at low cost is attractive for theproduction of single-use applications, e.g. to reduce the risk for contamination inmedical applications. Moreover, polymers offer the possibility to tailor their physicaland chemical properties.

A considerable number of polymers have recently entered the MEMS field. Amongthese are thermoplastic polymers, polyimide, polycarbonate, acrylic plastics parylene,polymethyl methacrylate (PMMA), SU-8, etc. [25–30]. Furthermore, several promis-ing polymer based actuation principles have been utilized to address the above men-tioned issues and to potentially promote the development of integrated microfluidicsystems. They exploit a variety of both mechanical and chemical actuation princi-ples such as external pressure sources [31, 32], stimuli responsive polymers [33–36],centrifugal forces on a rotating disc [37], electrochemical gas generation through elec-trolysis [38, 39], electroosmosis with ion exchange membranes [40, 41], decompositionof hydrogene-peroxide [42] or AIBN (azobis-isobutyronitrile) [43], melting of paraf-


fin [44–46], thermal expansion of gas [47, 48], and expansion of a condensed gas en-capsulated in a microcavity [49] or in chemically inert microspheres [50].

2.2 Poly(dimethysiloxane)

One polymer in particular has become one of the most popular polymer materials forfabricating microfluidic systems: Poly(dimethylsiloxane) (PDMS). This elastomericpolymer has enabled the fabrication of microfluidic systems with two-dimensional aswell as more complex three-dimensional structures, such as integrated microfluidicchannels, pumps, valves, switches, µCE systems and chaotic mixers [31, 32, 51–56].PDMS is also used in the fabrication of micromachined mechanical and chemicalsensors as e.g. the spring material in accelerometers or the ion selective membrane inISFETs [57,58]. Other applications can be as an adhesive in wafer bonding, as a covermaterial in tactile sensors or as a mechanical decoupling zone in sensor packagings [59].

The use of PDMS offers several beneficial characteristics; it is elastomeric with ashear elastic modulus G ≈ 250 kPa due to one of the lowest glass transition tempera-tures of any polymer (Tg ≈ −125◦C), optically transparent down to 240 nm, durable,chemically inert and nontoxic. It allows for the use of several detection methods andthe integration of functional components, enables straight-forward interfacing to themacroscopic world as well as the fabrication of microscopic features via soft litho-graphic and rapid prototyping techniques [51, 60]. Other advantages of PDMS arethat it provides unique permeability characteristics [61–63] and the possibility to al-ter its surface chemistry, which is useful when creating reversible or irreversible bondswith other smooth surfaces. Without surface chemistry the intrinsic surface propertyof PDMS exhibits hydrophobic surface characteristics [64]. The structural formulafor PDMS is given in Figure 1, where n is the number of repeating monomer units.

Si

CH3

CH3

O

n

Figure 1. Structural formula of PDMS (n can be in the order of thousands).

One commercially available and commonly used silicone rubber is supplied in twoparts (Sylgard R© 184 Silicone Elastomer Kit, Dow Corning); a liquid base oligomer(dimethylsiloxane with the vinyl functional group, SiH=CH2), and a catalyst orcuring-agent (pre-mixture of a platinum complex and copolymers of methylhydrosilox-ane and dimethylsiloxane, SiH). When mixing base and curing-agent at a typical ratioof 10 parts base to one part curing-agent, by weight, the catalyst promotes a hydrosi-lylation reaction between base and curing-agent [65, 66].

2.3 Soft lithography, rapid prototyping and replica molding

One particularly important contribution to the microfluidic field has been a methodfor fabricating prototype devices in PDMS, namely soft-lithography [64]. Soft-lithogra-


phy represents a non-photolithographic approach to micro- and nanofabrication whichis based on self-assembly as well as replica molding and enables rapid prototyping ofdevices. It typically involves a soft polymer (PDMS) for patterning relief structures onits surface to facilitate, e.g. microcontact printing (µCP), replica molding (REM), mi-crotransfer molding (µTM), micromolding in capillaries (MIMIC) and solvent-assistedmicromolding (SAMIM) [65]. These processes are particularly attractive since theyallow microfabrication of devices in the range of 20-100 µm at very low cost andwithout the need for expensive laboratory equipment. Using these methods, pro-totype devices can be fabricated in a time period that is considerably shorter thanwhat is generally needed in silicon processing (approximately two days compared toone month or more). Figure 2 shows the principal fabrication steps used in rapidprototyping of PDMS.

��

��

��

��

��

��

��

��

��

��

� ��

��

��

��

��

��

Figure 2. Rapid prototyping process scheme.

In this work rapid prototyping and replica molding are the primarily used tech-niques for the fabrication of microfluidic prototype devices. The process begins bydesigning a desired pattern via computer-aided-design (CAD) software which is thenprinted on a high resolution transparency. This transparency is used as a photomaskto carry out standard photolithographic processes [2, 3]. In photolithography a thinphotosensitive polymer layer (photoresist) is illuminated with UV-light through thephotomask to initiate a polymerization process in a pattern at defined regions. Theunpolymerized photoresist is then dissolved in a developer solution. This leaves arelief structure on the wafer which is then used as a master to cast PDMS and createan imprint of the relief structure.

After fabrication of the master, the surface of the wafer needs to be treated to


eventually enable removal of the PDMS after casting and curing. In this work, thefabricated masters are covered with a teflon-like layer prior to replica molding. Toaccomplish this, the master wafers are treated in an inductively coupled plasma (ICP)using C4F8 gas for about one minute, which leaves its surface hydrophobic. The pro-cess is based on the degradation of C4F8 in several steps to (CF )x which polymerizesat the wafer surface. This is enforced by an induced magnetic field which triggersthe release of electrons from Ar. The free electrons initiate the degradation of C4F8.The deposited layer prevents a permanent bond to be formed between the silicon andPDMS, which enables the cured PDMS layer to be easily peeled off the master.

Using such a master, the negative relief structures are imprinted in PDMS byreplica molding. This involves casting of the mixed and degassed PDMS pre-polymeron the master, heat curing and eventually peeling it off the master. The replicatedPDMS layer can then be bonded to other flat surfaces reversibly or irreversibly, whichseals the chip.

2.4 PDMS microfluidics

Another particularly important contribution to the microfluidic field has been thedevelopment of pneumatically actuated valves by means of soft-lithographic fabrica-tion techniques [31]. These valves make use of a two-level PDMS channel networkwhere one level of channels contains the fluid to be controlled and the other level ofchannels pneumatically controls the adjacent fluid channels. The operation of suchvalves resembles the idea of someone stepping on a water hose and squeezing the hoseuntil water cannot flow through any longer. When applying pressure to the controlchannel and increasing it, a membrane, at the position where channels of both levelscross (seen from above), deforms into the fluid channel, which reduces the size ofthe channel until the fluid channel is completely blocked. When the pressure in thecontrol channel is reduced, the membrane moves back to its original state due to theelasticity of PDMS, which renders the fluid channel normally open.

��

��

(a)

��

��

��

(b)

Figure 3. Photograph of a two-level PDMS based microfluidic valve commercialized byFluidigm Corporation in (a) open state and (b) closed state. Control and fluid channel are100 µm wide channels crossing perpendicularly. Adapted and modified from [67]

Figure 3 shows photographs of such a microvalve in its normally open state as wellas the closed state from above. By configuring such valves in a multiple arrangement,effective microfluidic components such as pumps, mixers and biochips are feasible


to perform chemical and biological experiments [31,68,69]. Nevertheless, using thesevalves requires an off-chip controller for each independent valve structure. This meansthat the number of independent valves that may be integrated on a lab-on-a-chipdevice is limited by the use of these off-chip controllers due to their size, cost, andpower consumption. A microfluidic latching valve instead remains in its open or closedstate without the need for continuous external actuation, which can reduce the amountof external controls. PDMS latching valves have shown to sustain their state for morethan ten minutes until depressurization occurs due to the permeability of PDMS. Suchlatching valves have been presented in networks for microfluidic large scale integration(mLSI), similar to electronic integrated circuits, to realize microfluidic multiplexersand memory where 20 pneumatic control channels are sufficient to individually controlas many as 1024 fluidic channels [69, 70].

Figure 4. A demultiplexer using vacuum-latched valves in four rows of valves, where eachvalve is alternately connected to off-chip vacuum or pressure by 4/2 solenoid valves. The 16vacuum latch valves at the bottom of the device are actuated via the demultiplexer. Adaptedand modified from [71].

Normally closed membrane seat valves have been presented as latching valvesbased on a three- or four-valve network to further reduce the amount of external con-trols [72]. These valves comprise a PDMS membrane which is sandwiched in betweenetched glass wafers containing the fluidic channel network. The latching microflu-idic valve structures can be controlled independently using an on-chip demultiplexer,which makes them suitable for integration into large-scale, high-throughput lab-on-a-chip devices [71–73]. Figure 4 shows a four-bit binary demultiplexer which addressessixteen independent latching valves, distributing pressure and vacuum pulses to eachof them in turn, by a single input connection (pressure/vacuum) at the top of thedevice.

The technology of microfluidic large-scale integration enables hundreds of assaysto be performed in a highly automated parallel manner with multiple reagents. It is apotential candidate to replace the established methods of biological fluidic automationsuch as pipetting robots.


3 SEALING OF MICROFLUIDIC STRUCTURES 11

The true delight is in the finding outrather than in the knowing.

Isaac Asimov1920–1992, Russian-born AmericanBiochemist, Science Fiction Writer

3 Sealing of microfluidic structures

3.1 Introduction

In MEMS fabrication, bonding of substrates to one another has become an impor-tant process which utilizes a diverse range of wafer-bonding techniques. Among themost commonly used semiconductor bonding techniques are anodic and direct bond-ing [74–76], eutectic bonding [77], low-temperature melting glass bonding [78], solderbonding [79], thermocompression bonding [80], ultrasonic bonding [81] as well as ad-hesive bonding [82].

The rapid development of microfluidic systems toward µTAS has triggered theneed for new fabrication techniques to realize microfluidic channels, devices and sys-tems, which may even be integrated with complimentary metal oxide silicon (CMOS)circuits, e.g. for on-chip signal analysis of large amounts of data in such systems.Surface micromachining techniques, as well as various kinds of bonding methods forchip- and wafer-bonding, have been introduced to fabricate such microfluidic struc-tures, which most commonly are based on materials such as silicon, glass and poly-mers [83–86].

All bonding techniques are based on a common principle: if two materials arebrought into sufficiently close contact, they fuse and adhere to each other. A numberof basic bond types, i.e. covalent bonds, van der Waals bonds, metallic bonds andionic bonds, ensure that atoms and molecules in a solid or between two solids cohereand adhere, respectively. These bond types are based on electromagnetic (coulombic)forces, which result from the attraction of opposing electrical charges. Most waferbonding techniques are dominated by the bonding mechanism of covalent bonds andvan der Waals bonds. These mechanisms occur when atoms of two opposing surfacesare less than 0.5 nm apart from one another. Direct coulombic forces arise whenevercharging occurs. These are usually strong and dominating, unless water vapor orwater is present in the environment, which partly neutralizes the charged surfaces.

3.2 Adhesive wafer bonding

Adhesive wafer bonding is widely used in the manufacturing industries, e.g. for theproduction of airplanes and cars, to join materials of various kinds. The techniqueutilizes organic as well as inorganic intermediate layers to create a bond betweentwo wafer surfaces [83–85, 87–90]. Like other bonding techniques, adhesive bondingoccurs when atoms and molecules are brought into sufficiently close contact, which is


promoted by the adhesive material deforming and wetting the surfaces to be bonded.

The intermediate adhesive layer can be deposited on one or both surfaces by spincoating, laminating, spraying or other suitable deposition techniques. To finally createa bond, the wafers are brought into close contact and the adhesive layer is generallycured by applying heat and pressure. The exact procedure is dependent on whichadhesive material is used. Various organic adhesives (mainly polymers) with differentmaterial properties and chemistries are available [91]. When using polymer adhesives,the bonding temperatures may vary between room temperature and 450◦C, whichmakes adhesive bonding compatible with CMOS. A detailed discussion of differentadhesive bonding materials and the relevant bonding parameters can be found in [92].

3.3 Sealing of PDMS structures

Bonding of PDMS layers with imprinted microchannels is a critical step in the micro-fabrication process. PDMS based microfluidic device fabrication commonly utilizesthe property that PDMS can conformally contact and reversibly seal to other smoothsurfaces, which is provided by simple van der Waals contact. Such contact is water-tight up to fluid pressures of 35 kPa [64].

The most commonly used technique to form an irreversible seal is to expose PDMSshortly to oxygen plasma [53, 64]. Plasma oxidation is commonly used to alter thesurface chemistry of PDMS. Surface-oxidized PDMS can seal to itself, glass, siliconand others [93]. Most of the research on the chemistry happening during such plasmatreatment indicates the generation of silanol groups (Si-OH) on the surface of thePDMS at the expense of methyl groups (-CH3), which renders the PDMS surfacehydrophilic. The concentration of hydroxyl groups is increased upon the oxidation ofthe surface layer, which leads to the formation of strong intermolecular bonds uponcontact [64, 93, 94]. These bonds form the basis of a tight irreversible seal that canwithstand pressures of 350-500 kPa [95]. The formation of such irreversible seals inPDMS have recently been presented by means of a handheld corona system [96].

When using this bonding technique it may be necessary to try various exposureparameters for a given plasma setup. Furthermore, it requires technical skill to aligntwo surfaces and bring them into contact quickly after oxidation (<1 min), since thesurface of the oxidized PDMS reconfigures in air [60,97]. In case of fabricating multi-layered microfluidic systems the use of micromanipulators is required, since oxidativesealing occurs upon contact. Experience shows that this sealing approach works bestin a clean environment, when the samples themselves are clean and dry, and thesurfaces to be bonded are not bent during handling.

As alternative methods for sealing PDMS irreversibly one approach utilizes anexcess of the monomer to one layer of PDMS and an excess of curing-agent to theother [31]. The layers are brought into conformal contact and again cured at highertemperatures to form the seal. This method has the advantage that it allows the twolayers to be aligned properly, since sealing is initiated upon heat. On the other handit is limited to sealing of layers consisting of PDMS.


3.4 Selective PDMS bonding

In standard wafer bonding applications it is often attractive to only bond specific areasof a wafer, whereas others remain unbonded, in particular in the fabrication of three-dimensional or multilayered structures. In adhesive wafer bonding a localized bondformation is achieved by simply applying the adhesive only at the designated positions.Methods to deposit the polymer adhesive at certain positions involve spraying with ashadow mask, lamination of a patterned polymer sheet, local dispensing and screen-printing of a liquid polymer precursor.

As PDMS conventionally is bonded using an oxygen plasma, and not an adhesiveintermediate layer, the use of such methods is not an option. To selectively bondPDMS to a substrate, designated areas would need to be masked during the plasmaoxidation process. On chip-level such selective bonding would be rather straightfor-ward, as long as the smallest feature sizes are in a feasible range to pick and placethe mask accurately. However, to create areas in between bonded PDMS structureson wafer-level, that are separable from one another, we introduce an anti-sticktionpattern into the system, which remains on the structures after the plasma oxidationprocess. To achieve this, the idea used in adhesive wafer bonding, of using a shadowmask and spraying the adhesive, is picked up. Instead of spraying an adhesive wesputter thin Au-patterns, at a thickness in the range of nanometers, through theshadow mask on both PDMS surfaces. The formation of a permanent bond at thesedesignated positions is prevented, since the surfaces are not oxidized properly dur-ing the plasma oxidation step. At the positions where Au is present no permanentbond is formed, whereas the surrounding areas form a proper bond. This allows forthe layers at these areas to be separable from one another after the bonding processis completed. Paper 6 describes this bonding process by means of fabricating out-of-plane liquid handling devices. There a sheet of Scotchpak release liner (3M) ispatterned at designated areas, by using a standard drilling machine, and laminatedon the PDMS prior to sputtering of Au. This approach is limited to relatively largefeature sizes (∼1mm) and circular patterns. In case smaller feature sizes with im-proved precision and/or other geometries are desirable, a polymer can be patternedvia photolithography and used as the shadow mask. The use of SU-8 photoresistwould provide high precision as well as the possibility to achieve thick layers. A thicklayer of SU-8 photoresist is straightforward to fabricate and involves photolithograph-ical fabrication steps only. The mask could be fabricated on a release layer which isremoved after the lithographic steps to release the SU-8 sheet. Other ways that wouldenable the fabrication of an accurate shadow mask would be to use Parylene-C [98],epoxy-functionalized-PDMS [99] or photosensitive-PDMS [100].

3.5 Full-wafer adhesive PDMS bonding

Adhesive bonding, utilizing PDMS-prepolymer, has been described previously for thepurpose of fabricating microfluidic chips [101], where the PDMS-prepolymer is di-luted with the organic solvent toluene. However, PDMS based microfluidic deviceshave shown limited compatibility with various organic solvents such as toluene, dueto swelling [102]. Some organic solvents, e.g. methanol and glycerol, swell PDMSonly to a small extent so that they could still be used within channels of PDMS based


microfluidic devices. Others swell PDMS to a greater extent, thus they are not com-patible with the use in PDMS based systems. The ratio of swelling is approximatelyinversely related to the solubility parameter of the solvent that is used.

Therefore, a straightforward PDMS bonding technique, utilizing solely the curing-agent of the two-component silicone rubber (Sylgard R© 184) as the intermediate layerfor adhesive bonding, is evaluated. Its applicability is verified by means of fabricatingmicrofluidic structures using two different approaches.

One utilizes a technique where the structural layer is placed directly on the in-termediate layer, as depicted in Figure 5(a). For that purpose a structural PDMSlayer with integrated microchannels is fabricated through replica molding using stan-dard photolithographical means. Another unstructured PDMS layer is fabricated ona plane silicon wafer, coated with curing-agent at a desired speed and transferred ontothe structured PDMS layer. Finally, heat curing forms an irreversible seal at the inter-face. A cross-sectional view of the resulting microchannel, where the inside geometryhas slightly curved edges, is shown in Figure 5(b). We believe that after bringingboth layers into contact the curing-agent is wetting the inside of the microchannel,due to surface tension, and thereafter reacts with uncured monomers dissolved in thePDMS bulk. This renders the edges curved after polymerization. This phenomenoncould be useful in microfluidic devices that require round and multi-depth channels.A similar approach has been presented to fabricate round channels and a dual-depthserpentine micromixer, using the surface tension of PDMS [103].

��

��

��

��

��

��

��

��

��

��

��

��

(a) (b)

Figure 5. (a) Principal fabrication scheme for fabricating directly bonded microfluidicstructures. (b) Cross-sectional picture of an imprinted microchannel in PDMS bonded toanother PDMS layer which was spin coated with curing-agent prior to bonding. The edges ofthe microchannels become slightly curved. The dashed line indicates the interface betweenthe two layers.

Another approach uses a patterned adhesive bonding technique [104] to fabricatemicrofluidic structures. Its fabrication scheme is depicted in Figure 6(a). A structuralPDMS layer with imprinted microchannels is again fabricated through replica mold-ing. Spin-coating curing-agent on a carrier substrate allows the structural PDMSlayer to be stamped onto the coated carrier substrate to pick up the curing-agent.Then this PDMS layer can be transferred onto another unstructured PDMS layerand finally heat cured to generate a permanent bond at the interface. Figure 6(b)shows a cross-sectional view of the resulting microchannel, where the inside geometry


of the microchannel remains unchanged.

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

(a) (b)

Figure 6. (a) Principal fabrication scheme for patterned adhesive bonding of microfluidicstructures. (b) Cross-sectional picture of an imprinted microchannel in PDMS which wasstamped onto a curing-agent coated substrate prior to bonding to another PDMS layer. Thegeometry of the microchannel remains untreated when using this patterned adhesive bondingmethod. The dashed line indicates the interface between the two layers.

��

��

��

��

��

��

��

��

��

��

��

��

(a)

��

��

(b)

Figure 7. (a) Schematic of the fabrication of test devices. Flanged capillary tubing iscast in PDMS and cured. Curing-agent is spin coated on a substrate at various spinningspeeds. The fabricated parts are brought into conformal contact and heat cured at differenttemperatures.(b) Picture of the test device after pulling apart the flanged capillary tubing.The picture shows that the PDMS tears off through the bulk instead of separating at thebond interface. Clearly visible is the previously unbonded area, at the position of the flangedcapillary tubing, as well as the previously bonded area, which was torn off.

To further characterize this bonding technique, bond strength measurements areconducted on test devices, using the destructive blister test method [105]. A schematicof the fabrication of test devices used for the blister test is shown in Figure 7(a). Acompressed air supply and pressure transducer are connected to the fabricated devices.Increasing the pressure at the inlet eventually leads to failure of the bond interface ata critical pressure. The crack propagates through the bond interface and a distinctpressure drop can be noted. Table 1 shows the results for bond strength measurementsat different fabrication parameters and shows that the resulting bond strength iscomparable to, or even stronger than, plasma assisted PDMS bonding [95]. Foreach fabrication parameter several experiments have been conducted. Additionally,


Table 1. Bond strength measurements using the blister test method. The bond is failingand a crack propagates through the bond interface and a distinct pressure drop can be noted.

Bondedmaterial

Spinspeed

Curingtemper-ature a

Bondstrength

Relativestandarddeviation

/ rpm / ◦C / kPa / %

PDMStoPDMS

2000

65 455 13.1

100 538 9.7

120 462 4.5

4000

65 545 1.9

100 393 1.7

120 365 7.9

6000

65 531 13.4

100 372 9.0

120 531 6.0

PDMStoSilicon

2000

65 572 15.3

100 600 7.4

120 669 5.2

4000

65 800 5.6

100 593 5.3

120 696 3.8

6000

65 400 14.5

100 414 9.2

120 400 3.8

PDMStoGlass

2000

65 407 7.6

100 345 5.0

120 414 9.7

4000

65 393 5.8

100 338 10.4

120 400 18.4

6000

65 290 8.6

100 269 6.9

120 386 6.0

aCuring times according to typical values for PDMS curing: 4 h@65◦C, 1 h@100◦C, 1/2 h@120◦C.

test devices have been pulled apart manually, as shown in Figure 7(b). Instead ofseparating at the bond interface the device tears off through the PDMS bulk, whichindicates that the bond strength is comparable to the cohesion of bulk PDMS.

The presented experiments show that the adhesive PDMS bonding technique is ap-plicable for wafer-level fabrication. Using this technique is straightforward and utilizessolely the curing-agent of the standard two-component silicone rubber (Sylgard R© 184)


as the intermediate layer for adhesive bonding. Spin coating the curing-agent on var-ious substrates, such as PDMS, silicon or glass, allows microfluidic components tobe fabricated and sealed irreversibly at very low cost without the need for expen-sive laboratory equipment. Bond strength measurements show that the bond createdusing this bonding technique is comparable to, or stronger than, plasma assistedbonding [95]. Moreover, the fabrication of round channels is feasible due to surfacetension of the curing-agent. Utilizing a straightforward patterned adhesive bondingtechnique instead [104], the inside of the bonded microchannels remain untreated.During the fabrication of multilayered microfluidic chips, the structural layers can bealigned accurately since curing is initiated upon heat, which is an advantage over thewidely used plasma assisted PDMS bonding technique. Additionally, the presentedadhesive bonding technique is compatible with room-temperature curing and allowsthe functionalized immobilization of biomolecules in microstructures prior to sealing.Table 2 shows an evaluating comparison of various PDMS bonding techniques andtheir practical applicability.

Figure 8. Photograph of a rolled micro tube based on PDMS which contains enclosedmicrochannels.

Using this adhesive PDMS bonding technique furthermore allows the fabricationof rolled micro tubes made of PDMS, as shown in Figure 8. Such rolled PDMS microtubes could contain microchannels, immobilized biomolecules, sensors and circuitry,etc. taking advantage of the unique PDMS characteristics, such as permeability. Itsfabrication would start on an initially flat PDMS layer prior to rolling and sealing,which provides the possibility for fabricating e.g. sensors on both the inside andoutside of the micro tube. Then a flexible layer could be rolled to form micro tubeswith various diameters. A similar approach of fabricating biosensors on a spirallyrolled micro tube by means of a flexible Kapton substrate has been presented recentlyfor microcatheter-based cardiovascular in vivo monitoring [106].

3.6 Selective full-wafer adhesive PDMS bonding

Within wafer bonding it is often attractive to bond specific areas of a wafer, whereasothers remain unbonded. The adhesive PDMS bonding technique described in Sec-tion 3.5 has shown promising results, where the bonds created using this method areas strong as, or even stronger than, the bonds created by the widely used plasmaassisted PDMS bonding technique. Furthermore, using such an adhesive approachallows PDMS surfaces to be bonded selectively to another by means of a suitable


deposition technique.

��

��

� � �� !�� ! ��"

� ��

��

��

�� "

�� #��!��

��

�"

�"

(a)

��

��

��

$��%�

�� !

��

��

�� !

��

��

(b)

Figure 9. (a) Schematic of the principal fabrication steps when selectively bonding PDMSmicrochannels utilizing microcontact printing of PDMS curing-agent. (b) Picture sequenceof a liquid filled microchannel containing a selectively bonded area with the channel inlet at(I) zero pressure , (II) pressurized, and (III) depressurized, subsequently.

To selectively bond PDMS, by means of this adhesive PDMS bonding method,we utilize a softlithographic approach of microcontact printing PDMS curing-agentonto one of the surfaces to be bonded, as depicted in Figure 9(a). First, a PDMSstamp for microcontact printing as well as a structural PDMS layer with imprintedmicrochannels are fabricated using the replica molding technique, as described inSection 2.3. Then PDMS curing-agent, that will be used as the intermediate layerfor adhesive bonding, is spun on another substrate. The previously fabricated PDMSstamp is pressed on the coated substrate to pick up the curing-agent and again placedon another substrate in order to transfer a pattern and prepare for bonding. Thisrenders specific areas on the substrate coated with curing-agent according to the”inked” stamp pattern. The transferred curing-agent represents a patterned inter-mediate layer for adhesive bonding, which fulfills the bonding function at designatedareas after curing. Finally, another PDMS layer is placed on the ”inked” substrateand a permanent bond, at the interface where curing-agent is present, is accomplishedby heat- or room-temperature curing. Figure 9(b) shows a picture sequence of an un-bonded area before and after external pressure is applied to the channel inlet. Theapplicability of this selective full-wafer adhesive PDMS bonding method, via micro-contact printing PDMS curing-agent, is shown by means of fabricating in-plane liquidhandling devices, as shown in Paper 6.


Table 2. Comparison of various alternative PDMS bonding methods. The different PDMSbonding techniques are evaluated according to various attributes that are of importancefor PDMS microfluidic device fabrication. If a technique features beneficial characteristicstoward a certain attribute, it is marked as positive, and in the contrary as negative.

AttributesPlasmaoxidation[53,64]

Handheldcorona dis-charge [96]

DifferentPDMScomposi-tion [31]

Curing-agentadhesivebonding[Paper 7]

Sensitivitytoenvironment

– – – + ++

Urgency ofpost-processing

– – – ++ ++

Equipmentrequirements

– ◦ ++ ◦

Selectivebondformation

◦ ◦ ◦ ++

Wafer-levelapplicability

+ – – ++ ++

Layeralignment

– – ◦ + ++

Versatilitytowardnovelapplications

◦ ◦ + ++

++ = very good, + = good, ◦ = neutral, – = poor, – – = very poor.


4 EXPANDABLE MICROSPHERES 21

When inspiration does not come to me,I go halfway to meet it.

Sigmund Freud1856–1939, Austrian Physician, Founder

of Psychoanalysis

4 Expandable microspheres

4.1 Introduction

Thermally expandable microspheres (XB) are small spherical polymeric particleswhich have been industrially produced for several years. Thermoplastic microsphereswere first described by Dow Chemical Co. [107] and developed further by others [108–110]. They are used in a wide variety of application areas, e.g. as additives in printinginks to form topographical patterns on paper, wallpapers and textiles, in polyurethaneapplications to compensate for normal shrinkage and as blowing agent to decrease thedensity of plastics or to achieve a controlled and predictable foaming process while atthe same time improving shock, vibration and sound absorption properties. Moreover,expandable microspheres have been used to produce highly porous ceramics with highuniformity of the cell size and shape or in car protection applications for improvedanti-corrosion resistance and increased acoustical insulation levels [111–113]. Thedominating technique to manufacture expandable microspheres is based on suspen-sion polymerization of vinyl type monomers with in situ encapsulation of a volatileorganic propellant in an aqueous medium [114–116].

4.2 Material characteristics

Expandable microspheres are micrometer-size beads that increase their volume uponheat. Their gastight thermoplastic shell is typically made of polymers or copoly-mers of vinyl chloride, vinylidene chloride, acrylonitrile, methacrylonitrile, methylmethacrylate and styrene. This thermoplastic shell typically encloses small amountsof condensed liquid hydrocarbon, such as n-pentane, isobutane, isopentane, hexane,butane etc. The particle size of unexpanded microspheres is typically within 5 µmand 50 µm in diameter. When heated above a certain temperature the thermoplasticshell softens and the enclosed hydrocarbon changes phase from liquid to gas. Thisphase transition results in a considerable increase of internal hydrocarbon pressure,thus the microspheres are literally inflated, much like a balloon, which is illustratedin Figure 10(a). This volume increase proceeds until a new equilibrium between in-ternal hydrocarbon pressure and stress in the shell is reached. After cooling downthe microspheres retain in their expanded state since the thermoplastic shell hardensagain, thus this volume increase is irreversible.

An example of commercially available expandable microspheres is Expancel (AkzoNobel, Sundsvall, Sweden). In this work solely dry and unexpanded Expancel R© mi-


crospheres (820DU) have been used. Various types of expandable microspheres arecommercially available, featuring expansion temperatures in the range from 80-190◦C(Figure 10(b)), whereby the initial size of the microspheres is increased with higherexpansion temperatures. Upon heating, the volume of the microspheres can increaseup to 40 times, which leads to a decrease of its initial true density, typically from1000 kg/m3 down to 30 kg/m3 [111]. During the expansion process, the initial par-ticle size of a few microns in diameter is changed to several tenths of microns afterexpansion. At temperatures above the expansion temperature Tmax of the micro-spheres the thermoplastic shell gradually collapses and the encapsulated propellantvolatilizes, which leads to collapsing of the microspheres. Furthermore, Expancel R©

microspheres show excellent chemical resistance and can be used in contact with vari-ous chemicals, including many solvents, without negatively effecting their properties.All available grades are highly compressible, and retain their original volume afternumerous pressure loading and unloading cycles.

��

� �

��

(a)

(b)

Figure 10. a) Heating of the microspheres causes the propellant inside the spheres to gasifywhich results in a dramatic volume increase of up to 40 times. b) Graph showing differentexpansion intervals for various types of commercially available expandable microspheres.

4.3 Expandable microspheres in microfluidics

Microspheres, also known as beads, are commonly used as the solid phase for separa-tion, synthesis and detection of molecules in various fields, such as medical diagnostics,microbiology, cancer research, immunology and molecular biology. These beads of-fer to modify their surface chemistry via functional groups, which can change surface


properties to hydrophilic, hydrophobic, fluorescent or active for special ligand bindingproteins.

Utilizing functionalized beads as the solid phase, allows to chemically modify andat the same time increase the reactive surface area in e.g. a reaction chamber ormicrochannel. The effective use of functionalized beads in microfluidic devices requirestheir localization at specific positions in the system. Several designs addressing thisissue utilize integrated mechanical barriers made from different materials, such asglass, quartz and silicon, and aim at bead-based DNA analysis, chromatography andimmunoassay applications [117–120].

Expandable microspheres exhibit the possibility to expand upon heat and at thesame time possess the features of standard beads, which could be very attractive formicrofluidic applications. Two first approaches of how to utilize expandable micro-spheres as a novel component in the microfluidic field have been described in 2002.

One approach suggests methods for selective immobilization of expandable micro-spheres without the use of a mechanical barrier, including patterning by means ofphotolithography as well as microcontact printing combined with self-assembly basedon surface chemistry [121].

Patterning by photolithography is achieved by mixing expandable microsphereswith standard positive photoresist from Shipley (Microdeposit S1813). Such a resincan be spin coated on a substrate and subsequently be patterned by standard pho-tolithographical means. When spin coating a single layer of this resin, some of theexpandable microspheres remain at the exposed area even though the photoresist isdissolved completely after developing, as depicted in Figure 11. The remaining ex-pandable microspheres could be removed by overdeveloping the wafer, however thisin turn would decrease the selectivity to the unexposed area drastically. Therefore, athin layer of pure S1813 photoresist is first spin coated on a substrate, serving as arelease layer, following spin coating of the mixed resin on top. This enables completeremoval of the expandable microspheres at the exposed areas as depicted in Figure 11.

(a) (b)

Figure 11. Images showing expandable microspheres immobilized and patterned on asurface by means of photolithography. (a) Some microspheres remain at the exposed areaeven after dissolving the photoresist. (b) When using a release layer of pure photoresistbeneath, the surface is patterned successfully. Adapted and modified from [121].


Alternatively, the microspheres can be patterned and immobilized by self-assemblybased on surface chemistry [122]. This is accomplished by chemically patterning thesurface via microcontact printing and utilizing a self-assembly process of the micro-spheres. Immobilization of hydroxy-functionalized beads on silicon substrates hasbeen presented earlier by means of surface chemistry [123, 124]. The formation ofcovalent ester bonds between anhydride-functionalized surfaces and hydroxyl groupson beads is assumed, since beads do not attach on non-functionalized surfaces, i.e.the printed 4-(chlorosulfonyl) benzoic acid is not transformed into anhydride. Al-though, the surface chemistry of expandable microspheres was not known, a similarmethodology as for the hydroxy-functionalized beads has been applied to immobilizeexpandable microspheres. Figure 12 shows an SEM image of an immobilized self-assembled monolayer of expandable microspheres on a functionalized surface beforeand after expansion. No microspheres could be identified outside the printed anhy-dride pattern, which shows that expandable microspheres can be immobilized usingthis methodology.

Preliminary experiments show that it is also possible to immobilize monolayers ofexpandable microspheres in deep reactive ion etched channels. In such an applicationthe expandable microspheres might be used to coat the internal walls of microfluidicchannels, in order to enlarge the reactive surface area.

(a) (b)

Figure 12. SEM images of expandable microspheres using immobilization via self-assemblybased on surface chemistry. An immobilized monolayer of expandable microspheres before(a) and after (b) expansion. Adapted and modified from [121].

In another approach expandable microspheres are directly mixed in deionized wa-ter, which is utilized as a liquid actuator for the purpose of active liquid handling in amicrofluidic system [50]. The liquid actuator is localized at a specific position on thechip by means of a mechanical filter structure inside a microchannel which entrapsthe expandable microspheres. Heating is achieved by either integrated heaters or byheating the entire system which triggers microsphere expansion and accomplishes flu-idic actuation. Pumping and valving of liquids, using this approach, can be achievedwith nearly identical microstructures. Figure 13 depicts the design of the microfluidicdevice which determines whether the system works as a pump or a valve.


��

��

��

��

(a)

��

��

��

��

�� !�

(b)

Figure 13. Schematic exemplifying the concept of expandable microspheres used as a pump(a) or valve (b). Adapted and modified from [50].

4.4 Incorporation of expandable microspheres into a paste

The incorporation of expandable microspheres into other liquid materials simplifiestheir handling characteristics to a great extent. Expandable microspheres are mixedtogether with glycerin, which resembles a paste, allowing for volume expansion ofseveral times its original volume. This paste is used as an actuator material to re-alize a compact, low-cost liquid dispenser with high precision, which is suitable fordisposable drug delivery applications. Glycerin is chosen as the carrier liquid forthe microspheres, since its thermal properties promote good heat distribution and itshows a lower heat capacity compared to water. Additionally, its low vapor pres-sure prevents evaporation and keeps the microsphere mixture intact. Figure 14 showsan exploded view of a prototype device which is used to evaluate the approach ofconfining microspheres in a paste.

�"��#�"#��$��"�

%��$"��

��%"��&

'(��$��)��

��#��"

*+��

Figure 14. Exploded view of the prototype device. Ring, membrane and the expandablepaste are clamped into the notch of the liquid container, which stretches the vinyl membrane.

The paste is separated from a liquid container by a thin elastic vinyl membranewhich is stretched over a plastic ring and clamped into a notch in the container. A


(a) (b) (c)

Figure 15. Photographs of the expandable paste on a PCB. (a) Unexpanded state with aninitial volume of 35µl of expandable paste. (b) Expanded state after removal of the liquidcontainer. (c) Expanded state without a liquid container attached during operation.

lithographically defined heater on a standard printed circuit board (PCB) is used toenclose the expandable paste from the other side. As electrical power is applied tothe heater on the PCB, heat is generated which causes the microspheres inside thepaste to expand. The expanding paste dispenses liquid from the container through asmall opening, which capillary tubing can be attached to. Photos of different statesof the actuator paste are shown in Figure 15 and depict the expansion potential ofthis approach. The device is entirely fabricated using low-cost materials, allows foreasy-integration and is able to deliver liquid in the range of microliters to milliliters.

4.5 A thermally responsive PDMS composite

Expancel R© microspheres take the initial shape of a fine white powder, which makeshandling of small volumes of expandable microspheres as well as the issue of localizingthem in a microfluidic system rather time consuming and complicated [50, 121, 125].Instead of placing the expandable microspheres at specific positions in the system,e.g. by means of mechanical filter structures, surface chemistry or manually, we in-corporate them into a PDMS matrix. This simplifies their handling characteristics toa great extent and at the same time addresses the issue of localization. The result-ing novel thermally responsive silicone elastomer composite actuator combines themerits of PDMS and expandable microspheres (XB) and can be applied as a homo-geneous layer over the entire wafer, forming a matrix which confines the expandablemicrospheres within.

The composite is obtained by mixing PDMS base (Sylgard R© 184 Silicone Elas-tomer Kit, Dow Corning) with expandable microspheres (Expancel R© 820DU) andsubsequent degassing in vacuum. To achieve a resin ready for spinning on or castingagainst a wafer the desired amount of curing-agent is added and mixed thoroughly,followed by a degassing step in vacuum. To prevent the incorporated expandablemicrospheres from premature expansion during cross-linking of the PDMS matrix,curing of the composite is performed at a maximum of 65◦C for 4 hours or alter-natively at room temperature for at least 24 hours [126], i.e. below the actuationtemperature of the expandable microspheres. Figure 16 shows the distribution andsize of incorporated microspheres inside the cured composite before and after expan-sion.


(a) (b)

Figure 16. SEM photographs showing the size and distribution of expandable microspheres(DU 820) inside the composite before (a) and after (b) actuation.

4.6 Composite characteristics

In this section the expansion characteristics of the novel elastic polymer compos-ite are presented along with fabrication and application techniques as an integratedactuator material. The uncured composite has novel application possibilities andbeneficial processing properties which allow wafer-level processes and the well-knownrapid prototyping techniques to be used [60]. Figure 17(a) depicts the replica moldingprocedure: pouring the uncured composite on a master, degassing, heat curing andsubsequently peeling off the master. Figure 17(b) shows an SEM image of a high-resolution imprint of a replica-molded structure in the fully cured composite layer.The composite retains the characteristic of pure PDMS to replicate inverse featureswith high fidelity when cast against a master and subsequently peeled off.

#�"��,%��-.�"��#�

�,%��-.��(

/�)�)�

��$)�"��

(a)

0+ �

(b)

Figure 17. (a) Schematic of the replica molding process resulting in imprinted structures inthe PDMS-XB composite. (b) SEM picture of a high-resolution imprint of a replica moldedstructure in PDMS-XB composite using the well-established fabrication techniques knownfor pure PDMS.

Yet, the composite actuator’s main characteristic is to expand irreversibly upongenerated heat, which makes it possible to alter its surface topography locally withoutpatterning it. A thin composite film, spun on a wafer and covered with a rigid reser-voir, serving as a precisely defined volume, is used to show the expansion capability


of the composite actuator. Upon heat the composite fills out deep as well as smallvoids, which is revealed after removal of the rigid reservoir (Figure 18). The heightof the expanded composite structures conforms to the height of the structures usedas reservoirs.

��

(a)

0� �

(b)

Figure 18. Initially flat layers of the PDMS-XB composite fill out voids upon heat. (a) SEMpicture of the expanded composite after inducing heat to the entire system and removingthe rigid reservoir. A 80 µm deep replicated plastic reservoir was used. (b) SEM picture ofthe expanded composite after using a 30 µm deep DRIE etched microchannel as the rigidreservoir.

To properly control the expansion of the composite actuator at specific positionson a wafer, integrated heaters are defined lithographically. Adequate power appliedto the heaters introduces heat to the composite and hence expansion is effectuatedlocally. A thin, initially flat, composite film on a microheater covering an area of1 mm x 100 µm enables local expansion according to the shape of the heater andreaches 110 µm in height as can be seen in Figure 19.

��"��#��

�#��&��1 �2

003 345 �� 0�� 3��

�*� �

��*� �

��*� �

��*� �

5�*� �

%��)6 ��(��#��)��

�(��#��)��

Figure 19. Concept of an integrated resistor generating a heat pulse which initiates theexpansion of the composite. The graph represents the resulting local expansion of the com-posite (inset) measured with a profilometer along the length axis.

To evaluate spin-coating characteristics of the composite, as well as its operationalparameters, both the resulting film thickness of the spin-coated composite as well asits expansion behavior are measured. At various mixing ratios of PDMS prepolymer


and expandable microspheres, the resulting composite film thickness is evaluated asa function of spinning speed. After spin coating and curing various composite resins,film thickness measurements are conducted by cutting out pieces from the compositelayer and using a surface profiler across the cut area. Figure 20 shows the dependenceof the resulting composite film thickness for various mixing ratios as a function ofspin speed.

� +�� +�� 5�� 5+�� 0��

+�

��

+�

5��

5+�

0��

0+�

��

�+�

+��

��7�"��

��8��#9��))�7� �

+��&7��0��&7��+��&7��

Figure 20. Resulting composite thickness for various mixing ratios of expandable micro-spheres incorporated into the uncured PDMS matrix as a function of spin speed.

0��

#�"��,%��-.

��

(a)

/��!��

��'(��)��7�:

.��#��"��,%��"�(�7��&7�� +� 5�� 5+� 0�� 0+� �� +� +�� ++� 3�� 3+�

5�

��

3�

��

5��

55�

5��

53�

5��

0��

(b)

Figure 21. Relative volume change as a function of different mixing ratios of expandablemicrospheres. The volume change is determined by measuring the dimension variations afterheating to 80◦C and subsequent cooling.

The expansion behavior of the composite is evaluated according to the relativevolume change when heat is applied to the composite. The diameter and thicknessof circular discs, punched out from plane composite films, is measured with a caliper


gauge before and after expansion to calculate the relative volume change. Whenheated, the microspheres expand and deform the elastic composite drastically, asdepicted in Figure 21(a). Different concentrations of incorporated expandable mi-crospheres in the PDMS matrix are evaluated and show a maximum volume changeof approximately 270% for a concentration of 600 mg microspheres per 1 ml PDMSprepolymer at a typical mixing ratio of ten parts base to one part curing-agent, whichis depicted in Figure 21(b). The higher the bead concentration in the PDMS pre-polymer, the more viscous the composite resin becomes. Too high concentrations ofmicrospheres in the PDMS degrade its simple handling characteristics, e.g. for spin-on processing on wafer-level. The microsphere concentration in the PDMS matrixhas a significant influence on the expansion of the composite. However, varying themixing ratio for the PDMS prepolymer (base:curing-agent) shows only little influenceon the overall expansion. The relatively high standard deviation could originate fromthe measurement method (caliper gauge) and the elastomeric nature of PDMS. How-ever, the use of the actuator in a digital mode compensates for possible variationsin volume change of the material. On-chip liquid reservoirs, holding a specific liquidvolume, can be accurately defined during fabrication and, in one shot, the completeliquid volume inside the reservoir can be released. For such digital use no compli-cated feedback system to deliver a certain dose is required, since the liquid dose ispredefined.

! " # $ % &! &" &# &$ &% "!"$

"%

'!

'"

'#

'$

'%

#!

#"

##

(��)��

��*��

(��

��)�

� �

��+,��

��

-(�

-(�

��

�� +,��

��

��

��

��

��

��

Figure 22. Measured average temperature response of the NTC on top of the compositefor various heating times; (solid line) heating for 2 sec, (dotted line) heating for 1 sec onlyaccording to configuration (a) and (dashed line) heating for 2 sec according to configuration(b).

To investigate applications where the thermal behavior of the system is of im-portance we evaluated the impact of the thermal actuation energy. The dynamicheat response of the composite actuator is characterized in order to understand heatconduction through the composite with a small mounted device (SMD) negative tem-perature coefficient (NTC) thermistor placed on top of the PDMS-XB composite.Two different setups are used to measure the temperature response. First, a sim-plified measurement is conducted on the bare composite on top of a glass substrate


with an integrated heater, which constitutes only natural heat convection and ther-mal conduction to the glass substrate, as depicted in Figure 22(a). This results in ameasured average maximum temperature of 43◦C (solid line) and 35◦C (dotted line)when heating for two seconds or one, respectively. Figure 22(b) shows a more repre-sentative configuration of a liquid handling platform based on the composite actuator.The NTC is cast into PDMS to measure the temperature response inside a reservoirmade of pure PDMS which additionally introduces thermal conduction to the PDMSreservoir layer. The dashed line shows the temperature inside the PDMS mold reach-ing an average maximum temperature of 32◦C when heating for two seconds. Forall experiments the expansion of the composite actuator was validated after the mea-surement. These measurements show that the thermal actuation concept can be usedfor liquid handling components. The thermal load on liquid transported within sucha system is expected to lie well below the temperature of 80◦C, which is needed toactuate the expandable microspheres.


5 COMPOSITE BASED MICROFLUIDIC DEVICES 33

Impossible is a word only to be found inthe dictionary of fools.

Napoleon Bonaparte1769–1821, French Soldier, Emperor

5 Composite based microfluidic devices

5.1 Design and fabrication of fluidic components

Microfluidic components are designed according to the characteristic behavior of thecomposite actuator. As discussed in Section 4.6, the standard deviation of the com-posite expansion is relatively high. To compensate for this, microfluidic componentsare designed to operate in a digital mode, i.e. in the case of a micropump the liquidvolume which is enclosed in a reservoir is being released completely upon compositeexpansion. On-chip liquid reservoirs holding a certain liquid volume can be preciselydefined during the fabrication process. The use of the composite actuator in digitalmode has further advantages, such that a complicated feedback system to deliver acertain dose is not required, since the liquid dose is predefined by the size of the liq-uid reservoir. To achieve a certain dose, a multitude of such units can be arrayed inparallel. Using the composite actuator in digital mode decreases the overall system-complexity and increases delivery precision.

Devices using the composite actuator are fabricated in a multilayer fashion, whichinvolves the following steps. First, integrated heater structures defining the actuationposition are fabricated on standard FR4 printed circuit boards (PCB) using stan-dard photolithographical means; exposure to UV-light, developing of photoresist andetching of copper. All photolithographical fabrication steps use high-resolution trans-parencies as photomasks. Microheaters are fabricated in a meander shaped fashion tocover a desired area and their geometry can be adapted to generate a tailored heat pro-file. Finite element modeling (FEM) simulation, based on the Comsol multiphysics R©

simulation package, is used to optimize such tailored heat profiles. In contrast tomicroheaters with constant geometry, where the highest temperature occurs at thecenter of the heater and hence induces composite expansion at this location first, theheater geometry can be adapted to match a desired heat profile. Such heaters couldbe designed to generate e.g. a flat temperature profile which induces, spatially as wellas temporally, even composite expansion across the heater area or, a sloped tempera-ture profile which corresponds to a gradually progressive expansion of the composite,vertically as well as laterally.

After preparation of the heater structures, the composite actuator is prepared asdescribed previously in Section 4.5 and spin coated onto the PCBs to achieve a desiredcomposite thickness.

Thereafter, reservoirs and microchannel networks are formed on a master by us-ing the the replica molding techniques described in Section 2.3. Since expansion ofthe composite actuator creates rounded profiles, which eventually fill out reservoirs,


the structures on the master should preferably exhibit round edges to enable properdepletion of a microfluidic reservoir during operation. Structures that are used forreplica molding, ranging from micro- to millimeters in size, are fabricated to enablemicrofluidic components for various work ranges. A rather straightforward reflowprocess, based on standard photoresist, is utilized to create rounded profiles in themicrometer range. The creation of rounded structures in the size range of millimetersinstead, by means of standard micromachining techniques, is yet far more challeng-ing. Therefore, a simple process for fabricating such a reservoir master is developed,where UV-curing epoxy droplets are dispensed with a commercially available pipette.After the master is readily fabricated, liquid PDMS prepolymer is spin coated ontoit, cured accordingly and eventually peeled off. An exploded view of the multilayerstructure is shown in Figure 23(b).

As a last fabrication step, the replica molded PDMS structures and the previouslyfabricated stack of PCB/PDMS-XB have to be sealed permanently. Such a perma-nent seal is achieved utilizing either the plasma oxidation treatment (Section 3.3)or alternatively the full-wafer adhesive PDMS bonding technique described in Sec-tion 3.5. Figure 23(b) depicts a cross-sectional view of a fabricated device showingthe semi-circular microchannel and the expandable microspheres inside the PDMSmatrix. Finally, the devices are fully operational and ready to be filled with liquid.

&'�� (��

��)'��

�� (� � ��

(a)

��.

�,%��-.

�,%��#"�#��

(b)

Figure 23. (a) Schematic exemplifying the exploded view of an application, such as a pumpor a valve. (b) Crossectional SEM image of the PCB/PDMS-XB/PDMS stack, depictingthe semi-circular microchannel molded in PDMS and the PDMX-XB matrix.

5.2 Single-use pumps

The operating principle of single-use microfluidic pumps is described in Figure 24.Precise spatial and temporal control of the composite expansion is achieved by meansof lithographically defined integrated heaters. This allows expanding the compositelocally without patterning it. Such controlled composite expansion results in thefilling of a reservoir causing initially stored liquid to be released into a microchannel.

To investigate the performance of the fabricated pumps the on-chip liquid reser-voirs are filled under vacuum or manually with either paraffin oil or glycol and ethanolbased colored dyes diluted in deionized water. Vacuum filling is conducted by eithersubmerging the device into liquid [127] or adding a liquid droplet at the microchannelinlet after vacuum treatment of the device for 1 hour [128]. As sufficient power is


�,

��+,

��.��

�� .��/�0�

�1�� +,

��%!��

Figure 24. Schematic of the working principle, showing the device before and after actu-ation. Upon heat the PDMS-XB composite expands and consequently fills out a reservoir,resulting in dispensing of initially stored liquid.

applied to the system, the composite expands at the location of the integrated heaterunderneath and the stored liquid inside the reservoir is consequently released into themicrochannel.

In a prototype device with a 100 µm thick composite layer and reservoir dimensionsof 1.5 mm in diameter, holding a liquid volume of 24 nL, the liquid is released intothe microchannel upon applying 850 mW of mean power during 3 seconds. Thecomplete release of stored liquid from the pump reservoir within 1.6 seconds is shownin Figure 25. Even against back pressures of up to 100 kPa, liquid could be pumpedefficiently.

�2

$2

#2

�2

�2

��)

�+��)

��)

5��)

3��)

��

Figure 25. Picture sequence (a-e) showing the release of liquid from a reservoir. Thereservoir is completely depleted after 1.6 seconds.


5.3 Single-use valves

The use of the device as a normally open single-use valve is based on the same workingprinciple used for the pump. Due to the volume expansion of the composite actuatora microchannel section can be closed off effectively. Figure 26 shows the effect ofexternal pressure on an earlier closed valve. Increasing the external pressure at oneof the channel inlets makes the liquid front proceed into the blocked microchanneland cleaves the valve area until breakthrough occurs, which reopens the microchan-nel. The pressure where the microchannel is being reopened is highly dependent onthe power input during the valve closure. This phenomenon could be used to de-sign microfluidic networks with integrated valves for directional liquid transportationaccording to the power input at a certain location on the chip.

;��9��

;+�9��

;3�9��

;4�9��

�9��

��"�"�&��

��

�9��2

$2

#2

�2

�2

<2

#��#��

Figure 26. Picture sequence of the valve in its open (a) and closed (b) state. The closedmicrochannel can be reopened by applying external pressure to one of the inlets as shownin (c-f).

5.4 Microfluidic networks

The interaction of several pumps and valves allows fabrication of highly integratedmicrofluidic networks, e.g. for lab-on-a-chip applications. When designing such net-works it is important that liquid is released completely from a reservoir into thedesignated outlet. When using microheaters with constant geometry, where the high-est obtained temperature occurs at the center of the heater, liquid pinch-off couldoccur due to inaccurate reservoir to heater alignment during fabrication, as shown in


Figure 27(a). To prevent such residual liquid trapping during operation, the use ofa special microheater pattern, i.e. a decreasingly dense resistor meander to gener-ate a well defined sloped temperature profile, is suggested. Such a heater geometrygenerates a well defined sloped temperature profile, which corresponds to a gradu-ally progressive expansion of the composite actuator, vertically as well as laterally.This progressive expansion guides the liquid into a desired direction, according to theheater orientation as depicted in Figure 27(b). Additionally embedding active valvesprevents backflow during liquid transportation. Using this increasingly dense micro-heater pattern makes alignment of the microfluidic channel and reservoir network tothe underlying heater structure uncomplicated during the fabrication process of thedevices and ensures complete liquid release during operation.

�

�

��

��.��

(a)

�

�

��

(b)

Figure 27. (a) Exemplifying the problem of liquid pinch-off when a heater with constantgeometry is used, which leads to a maximum temperature across the heater occurring at thecenter. Therefore composite expansion is induced at this location first and might lead toliquid pinch-off. (b) The solution to the problem is a decreasingly dense heater meander,which leads to a directional gradually progressive composite expansion and thus preventingliquid pinch-off.

The design of such microfluidic networks includes these special microheater ge-ometries, as well as integrated valves for sealing off on-chip liquid reservoirs after thefilling procedure. Thereafter, the chip is ready for further use, its handling is easyand convenient, e.g. when conducting a biochemical reaction under a microscope.When applying sufficient power to the system, the composite expands locally at theintegrated heaters so that liquid stored in the reservoirs is consequently released intothe microchannels. Figure 28 shows transportation, merging and mixing of liquidsamples on such a microfluidic network.


��

)��"�)�"!��"

�=�)�# ��

�(�� ,%�� #��)��

�=55)�#

"��#��#��$�"

�� =50)�#

��"��>��(��,%��#��)��

�� =5�)�#

��)�"��"�)�"!��"

<��>��(��,%��#��)��

��

Figure 28. Video snapshots showing the operation scheme: (I) sample loading; (II) trans-portation into reaction chamber; (III) further transportation, (IV) of mixed liquid into down-stream reservoir for increased mixing. Note the evolving expansion profile of the compositein (III) and (IV).

The applicability of such microfluidic networks as a disposable lab-on-a-chip is con-firmed by demonstrating an enzyme reaction experiment based on a bioluminometricreaction. Detectable light is produced during a reaction mediated by the enzymeluciferase. Luciferase has been used extensively in molecular and cell biology, in par-ticular for the efficient detection and quantification of ATP, and as a reporter enzymefor studies of gene regulation and expression [129]. The enzyme also plays a centralrole in DNA pyrosequencing technology. In the presence of adenosine triphosphate(ATP), luciferine is oxidized into an exited state and light is emitted upon the returnto the ground state.

After preparation of the solutions, ATP is loaded into one and the pyrosequencingsolution into another reservoir. When adequate voltage is applied to the system, thesamples are pumped into a reaction chamber and come into contact, which instantlyinitiates the light emitting reaction. The enzymatic reaction is carried out in a darkenvironment and recorded with a CCD camera in real time. Post data processingallows to create false-color snapshots of the real-time reaction, which is shown inFigure 29(a). The instantly initiated reaction is indicated by the rapid increase oflight intensity upon contact of the samples, as depicted in Figure 29(b). The reactionkinetics of the system is comparable to the luminometric assay conducted on anearlier reported silicon based microfluidic platform [130]. Although only the luciferasereaction is shown, all necessary components to perform allele-specific extension forsingle-nucleotide polymorphism analysis based on pyrosequencing chemistry are there.Bioluminometric assays are limited to the optimal temperature of firefly luciferase(22◦C-28◦C) and thermal denaturation of firefly luciferase has been reported to occurat about 39◦C [131]. This shows that biological assays with limited thermostabilitycan be manipulated effectively on such a microfluidic platform.


� ��

� !� �� *+��

*+��

(a)

! '! $! 2! &"! &3!!

3!!

&!!!

&3!!

"!!!

"3!!

(��)��

4��

��)��5��

0��

6( 6�

0�� 71��

��89

(b)

Figure 29. (a) False-color snapshot of the light emission during the bioluminescence re-action on the LOC platform. (b) The graph corresponds to the light emission during thereaction. The reaction is initiated instantly as the solutions are transported and come intocontact, indicated by the rapid increase of light intensity.

5.5 Microliter range pumps

The size of the liquid volume that is manipulated by a microfluidic system can vary toa great extent and is determined by its practical application. Single-cell monitoringfor example, may require handling of liquid volumes of about one picoliter whereasthe administration of therapeutic agents, such as insulin or vaccines, requires themanipulation of liquid volumes ranging from hundreds of picoliters up to hundreds ofmicroliters.

Previously presented applications are capable of manipulating liquids in the rangeof nanoliters, where the composite expansion leads to filling of a reservoir merelywith solid matter. This holds restrictions in terms of practically achievable maximumstroke lengths, which limits its applicability for microliter range applications. Torealize the manipulation of liquid volumes in the range of microliters instead, a methodof buckling of the composite actuator film is adopted. The buckling phenomenon ofthe composite actuator film is based on the characteristic bond between the PDMS-XB and PCB, which is of moderate and reversible nature. The character of purePDMS to bond reversibly to plastics [60] is preserved by the PDMS-XB composite.The expandable microspheres incorporated inside the composite actuator layer expandisotropically upon heat, which introduces compressive stress into the composite andleads to unseaming of the bond between the PDMS-XB layer and the PCB substrate.The expansion induces buckling of the composite layer, which allows for larger strokelengths and eventually displacement of larger volumes. The operating principle ofmicroliter range pumps is described in Figure 30, where precise control is also achievedby means of lithographically defined integrated heaters. The maximum stroke length,i.e. the buckled state where the layer eventually finds its stress equilibrium is highlydependent on the power input, heater area, composite layer thickness and the amountof expandable microspheres in the composite layer and can be tailored towards adesired stroke.

Tests of only expanding the bare composite actuator film on the PCB substrate,i.e. no reservoir is attached, show that the expansion profile results in a spherically


�,

��+,

��.��

�� .��/:0�

�1�� +,

��%!��

Figure 30. Schematic of the working principle for microliter range pumping by means ofthe buckling composite actuator film, showing the device before and after actuation. Uponheat the PDMS-XB composite expands and consequently fills out a reservoir, which resultsin dispensing of initially stored liquid.

shaped structure, which is induced by the buckling of the composite actuator film,indicated in Figure 31(a). To ensure device performance and proper depletion ofmicrofluidic pumps, the shape of the used liquid reservoirs should conform to thespherical profile of the buckling composite actuator as well as possible. The fabricationof such structures, by means of standard micromachining techniques in the range ofmillimeters, is a challenging task. Therefore, a simple fabrication process for thereservoir master for PDMS replica molding is developed, where UV-curing epoxydroplets are dispensed with a commercially available pipette and subsequently curedunder UV-light. Figure 31(b) depicts the fabricated master showing the sphericallyshaped reservoirs on the wafer substrate. The resulting spherical shape of the liquidreservoirs, fabricated by the pipetting method, is beneficial for conformal filling bythe buckling unimorph composite.

"��

(a) (b)

Figure 31. (a) Photograph showing the resulting spherical profile of the buckling unimorphcomposite film after actuation. Note that no reservoir was used during actuation. (b) Pho-tograph showing the fabricated master mold for replica molding. UV-curing epoxy (EpotekOG 142-13) droplets are placed over the entire wafer, using an Eppendorf pipette.

To investigate the thermal load on a liquid sample, the dynamic temperatureresponse during liquid sample transportation is measured using the method of flu-orescence based thermometry. This has effectively been presented earlier, using astandard fluorescence microscope and a charge-coupled device (CCD) camera [132].A fluorophore, with a strongly temperature dependent quantum yield, is used to mea-sure the temperature response of an aqueous solution. Rhodamine B is chosen as the


fluorophore because its quantum yield is strongly temperature dependent in the rangeof 0-100 ◦C, which makes it ideal to determine the temperature by comparing its flu-orescence intensity at an unknown and potentially nonuniform temperature with itsintensity at a known, uniform temperature. Figure 32 depicts a fluorescence micro-scope image of a micropump, which is filled with rhodamine dye solution and showsthe area where temperature values are extracted during image analysis.

Figure 32. Fluorescence microscope image of the filled pump reservoir and access holeprior to dynamic temperature measurements. The dashed window depicts the area wheremeasurement data was extracted during image analysis.

First a calibration curve for the relationship between relative change of fluores-cent intensity and temperature has to be recorded. To generate such a calibrationcurve the average fluorescent intensity of the recorded images is determined via a cal-ibration setup for each corresponding temperature after background subtraction andnormalized by intensity at room temperature. Figure 33(a) shows the normalizedfluorescent intensity as a function of temperature, which is due to the quantum yieldof Rhodamine B. Open circles in the graph correspond to averaged measured datapoints, while the solid line is a third-order polynomial fit of the following form

T = p1 · I3 + p2 · I2 + p3 · I + p4 (1)

where T is the temperature and I the fluorescent intensity normalized by its valueat room temperature, p1 = (−71 ± 7) ◦C, p2 = (195 ± 13) ◦C, p3 = (−233 ± 11) ◦C,and p4 = (132 ± 3) ◦C. These calibration results correspond well to earlier pre-sented data [132]. The calibration data is then used to calculate the temperatureof the diluted Rhodamine B solution inside the system by measuring the intensityof fluorescence during operation of a micropump. The micropumps are filled withRhodamine B solution under vacuum and dynamic temperature measurements areconducted based on the recorded fluorescent images during the experiment and theearlier measured calibration data.

Temperature values are extracted during operation of the pump in an area acrossthe access hole. The temperature of liquid passing through the access hole reachesa maximum temperature of 50 ◦C, which is shown in Figure 33(b). The graph ad-ditionally depicts the dynamic liquid depletion of the pump reservoir, which showsthat the temperature of two thirds of the delivered liquid volume is below 43 ◦C.


5� 0� �� +� 3� 4� ��*5

�*0

�*�

�*+

�*3

�*4

�*�

�*

*

8��"��"��7�?�

��"�*��)��>�7��"$*��)

(a)

� � 5� 0� �� +� 3� 4� ��5�

5+

0�

0+

��

�+

+�

8��7�)

8��

��"��"��7��

�*�

�*5

�*3

�*�

�

��!��

��7��"$*��)

(��

0�.��

��&�#>#��

(b)

Figure 33. (a) Normalized fluorescent intensity of Rhodamine B dye solution as a functionof measured temperature during calibration. The solid line represents a third-order polyno-mial fit to the averaged data points. (b) Dynamic temperature response inside the accesshole during operation of the pump as well as the liquid volume remaining inside the pumpreservoir as a function of time.

This is below the critical temperature for many bioliquids, above which thermal de-naturation occurs. For micropumps handling lower volumes of liquid the maximumtemperature is expected to drop to lower values due to considerably shorter heatingcycles. Temperature sensitive liquids could also be manipulated indirectly, by usinga buffer liquid in direct contact to the composite actuator instead.

5.6 Liquid aspiration and dispensing

The irreversible behavior of expandable microspheres gives the composite actuator itscharacteristic to expand once. The use of the expandable composite actuator is veryconvenient and the realization of microfluidic pumps and valves rather straightfor-ward. The one-time expanding nature of the composite is restraining its applicabilityto uncomplicated devices.

Extra functionality may be added due to the fact that the composite actuatoroffers the possibility to locally alter its surface topography by means of generatedheat. Such active surface patterning is adopted to fabricate devices for controlled in-plane and out-of-plane liquid volume aspiration and release by creating an enclosedcavity in the system. The expansion of the composite actuator leads to generationof negative pressure when such a cavity is created in the system. Such an enclosedcavity may be achieved by means of lithographically defined integrated heaters. Whenvoltage is applied to the integrated heaters, the composite actuator expands along thegeometry of the heaters, which in turn leads to the lifting of the device-lid and cavitygeneration. To enable this lifting procedure, and accomplish cavity formation, severalfabrication and design aspects have to be addressed.

Firstly, an appropriate cover-lid needs to be implemented. A lid made of PDMSwould be too flexible, due to its elastomeric nature, and would not allow adequatelifting during operation to form a cavity. Using glass as the cover material wouldrestrict the wafer-level fabrication process of the devices. Instead, a reinforced cover-lid is fabricated by spinning a PDMS layer on a 3M-polyester-film, which counteracts


the low young’s modulus of PDMS, giving additional stiffness to the lid. This enablesthe lifting process during operation and at the same time retains its straight-forwardfabrication characteristics.

Secondly, to ensure device functionality, the reinforced cover-lid and the bottompart of the device have to be bonded together in such a manner that there are re-versibly as well as irreversibly bonded areas on the same wafer. To achieve an areawhere no bonding occurs, i.e. the area where a cavity is being created during de-vice operation, the formation of a permanent bond at these specific positions has tobe prevented in order not to impair the lifting of the cover-lid. To accomplish thisfunctionality, two different PDMS bonding techniques are applied, as described pre-viously in Section 3.3, to fabricate liquid aspiration units to operate either verticallyor laterally.

One technique utilizes the widely used PDMS plasma bonding process, where apatterned anti-sticktion layer is deposited onto the PDMS surface prior to plasmaexposure, which prevents areas from being oxidized during plasma oxidation. Theuse of this technique is shown by fabricating vertically operating liquid aspirationunits by means of a patterned Au anti-sticktion layer. These units are based on thecomposite actuator and a heater design consisting of an outer heater ring and aninner heater structure. As adequate voltage is applied to the outer heater ring, thecomposite expands along this heater geometry which results in lifting of the reinforcedcover-lid and hence formation of a cavity. As a consequence, negative pressure isgenerated and the cavity fills with liquid, which is present at the access port as shownin Figure 34(I-II). Addressing the inner heater of the device subsequently results incomposite expansion into the cavity and release of the before aspirated liquid volumeas can be seen in Figure 34(III-IV).

The fabrication of laterally operating liquid aspiration units is shown by means ofanother selective PDMS bonding technique. This approach utilizes a novel full-waferadhesive bonding technique, where PDMS curing-agent is used as the intermediatelayer for adhesive bonding. To selectively bond PDMS by means of this technique asoftlithographic approach of microcontact printing PDMS curing-agent onto one ofthe PMDS surfaces is utilized. The transferred curing-agent represents a patternedintermediate layer for adhesive bonding, which fulfills the bonding function at specificregions after curing.

Similarly to the vertically operating units a cover-lid is being lifted upwards uponthe event of aspiration and consequently a cavity is being created. The design isbased on the composite actuator and utilizes a reinforced cover-lid with an imprintedmicrochannel and integrated heaters, comprising of inlet and outlet valve heaters,outer heaters as well as an inner heater structure. The initial state of the deviceinlet is normally open, whereas the channel outlet is normally closed prior to liquidaspiration. This is accomplished via a supplementary barrier in the microchanneloutlet which is reversibly bonded at the PDMS-XB/PDMS interface, much like thearea where the cavity is being created during operation.

Figure 35(I) shows the initial structure after filling prior to device operation. Liq-uid aspiration is induced as voltage is applied to the outer heaters, which makes thecomposite expand along this heater geometry resulting in lifting of the reinforcedcover-lid and cavity formation, as shown in Figure 35(II). Consequently, this cavity


��

��

��,%�

@�

"��<�"#��

��

<��#�!��>

�(��#��)��

�(��#��)��

�(��#��)��

��<"��)��"�

��<"��!��&

��"�)

��<"��!��&��"�)

��<"��

Figure 34. Picture sequence showing the operation scheme of a vertically operating device:I) filled capillary tubing after sample loading; II) triggering the outer heater ring leads toliquid aspiration into the created cavity with liquid front moving downwards; III) Subsequentliquid release by triggering the inner heater and liquid front moving upwards; IV) liquid frontmoves back to its original position.

fills with liquid present at the device inlet. Addressing the inlet valve heater subse-quently results in closing off the device inlet, whereas the two outlet heaters along themicrochannel outlet are triggered to open the channel outlet by lifting the supplemen-tary barrier, as depicted in Figure 35(III). Triggering the inner heater of the devicenext, results in composite expansion into the cavity and release of the previouslyaspirated liquid volume through the channel outlet, as shown in Figure 35(IV).

5.7 Transdermal drug delivery

In the 1990s silicon based microfabrication techniques enabled the development ofmicrometer-sized needles [133]. Their use for drug delivery as minimally invasivetransdermal administration tools has been anticipated for delivery of macromoleculardrugs that do not passively diffuse through the dermal layer [134–137]. This admin-istration form is painless [138,139] since microneedles, a few hundred micrometers inlength, would only be inserted through the outmost skinlayers, yet not deep enoughto reach the nerve receptors in the lower reticular dermis. Such a microneedle-basedtransdermal patch may offer attractive and patient-friendly administration of vaccinesor insulin.

Until now, research has focused on the development of microneedles which havebeen tested for delivery of various drug formulations [139–146], but also as a gen-eral tool to circumvent the barrier function of the skin, as for example in EEG-measurements [147] or impedance-based skin cancer detection [148]. Both solid (i.e.


��

��

��

��

��

��)��)��&

�(��#��)��

�(��#��)��

�(��#��)��

<��#"�#��

<��#�!��>

Figure 35. Picture sequence showing the operation of a laterally operating device: I) filledmicrochannel inlet after sample loading in vacuum; II) triggering the outer heaters leads toliquid aspiration into the created cavity; III) Subsequently triggering the integrated valveheaters closes off the device inlet and opens the channel outlet; IV) Triggering the innerheater results in release of before aspirated liquid into the microchannel outlet.

needles without a bore) and hollow microneedle designs have been used for the pur-pose of active and passive drug delivery, respectively [140–142, 145, 146]. Active ad-ministration is the preferred solution to deliver a significant volume of a drug overa limited period, however only pharmacodynamic response and blood flux, withoutquantification of the delivered drug, have been reported [139,143,144].

The use of a minimally invasive transdermal patch would potentially have advan-tages in terms of injection safety, acceptability and therefore patient compliance [149].Additionally, intradermal injections intended for immunization purposes offer to di-rectly target the highly immuno-competent Langerhans’ cells of the dermis and epi-dermis [150].

In contrast, their use for rapid bolus administration is difficult since the amountof liquid that can be absorbed by the skin tissue is highly limited. Therefore, such in-jections would require an extended period to accomplish delivery of a particular dose.A patch-like system with integrated dispensing functionality, that slowly delivers thedrug into the tissue could overcome this limitation. Moreover, such an integratedand electrically controllable system would potentially allow variable infusion ratesfor advanced delivery schemes by means of feedback control for, e.g. insulin deliv-ery. Besides comfort and patient-compliance issues, reliability and cost-efficiency areimportant requirements on active patch controlled solutions.

To realize such a complete patch-like system with active dispensing functionalityfor the controlled release of liquid, the composite actuator, designed as a microliter


range pump, has been implemented in conjunction with minimally invasive, side-opened, microneedles, which is depicted in Figure 36(a). This approach addressesseveral important fabrication and handling issues and represents a complete patch-liketransdermal drug delivery system. The patch is electrically controllable, straightfor-ward to fabricate, exhibits low system-complexity, constitutes a disposable unit andyet offers to deliver liquid volumes both at high and low flow-rates. Furthermore,the insertion procedure of the patch is simple, due to the integrated ultra-sharp mi-croneedles, which have been developed for reliable skin penetration [151].

(a) (b)

Figure 36. (a) Exploded view of the microfabricated transdermal patch. (b) Photographof an assembled transdermal patch.

To evaluate the fabricated system, which is shown in Figure 36(b), an experimentalstudy comparing active and passive in-vivo insulin delivery into diabetic rats wasperformed. Two different delivery schemes, active as well as passive, are evaluated onrats receiving infusions at 2 µl/h as well as rats receiving diffusion-driven infusionsthrough actuated and non-actuated delivery devices, respectively. After 3 h infusiontime the devices are removed from the rats. Blood glucose levels as well as bloodplasma insulin concentrations are measured hourly throughout the infusion time atshifted time periods and additionally one hour after device removal.

Insulin delivery using active dispensing through the patch-like system shows sta-tistically significant decrease in blood glucose levels throughout the measurement,whereas passive (diffusion-driven) insulin delivery shows statistically significant de-crease in blood glucose levels after 120 min of delivery, as depicted in Figure 37(a).Consistent with the decrease of blood glucose for active delivery, the blood plasmainsulin concentration for active delivery clearly increases, as shown in Figure 37(b),whereas passively delivered plasma insulin concentrations show no significant increase.After a 3 h delivery period the insulin concentration is five times larger compared topassive delivery. This effect is related to the transport mechanism, where active in-fusion delivers insulin into the tissue for subsequent systemic distribution, whereaspassive delivery is limited to diffusive mass transport at the needle openings. Themeasured insulin concentration solely reflects administration of synthetic insulin ana-logue.

This study shows the feasibility of a patch-like system with on-board liquid storageand active dispensing capability. The proposed device represents a first step towardpainless and convenient administration of macromolecular drugs, such as insulin or


� +� �� +� 5�� 5+��

+

�

+

5�

@#��!��<�)��5� �7��))�!��<�)��8��#��"��

8��7��

.��&��#�)��7��%

(a)

� +� �� +� 5�� 5+��

5�

��

3�

��

��

5�

8��7��

A�)��#��#��"��7� �B7��

@#��!��<�)��5� �7��))�!��<�)��

(b)

Figure 37. Results of insulin delivery to diabetic rats using the patch system. (a) Graphshowing the measured blood glucose levels during the experiments. (b) Graph showing themeasured insulin concentration from blood plasma samples during the same experiments.Solid lines indicate active delivery through the transdermal patch, using a delivery rate of2 µl/h (n = 4). Dash-dotted lines indicate passive delivery through the transdermal patch,without an active delivery rate (n = 4). Dotted line in (a) represents vehicle-infused timecontrols (n = 4).

vaccines, and allows its application onto the body without particular skin prepara-tion, using moderate finger-force only. Although, rather low flow-rates are requiredfor leak-free delivery, the effect of insulin delivery is clearly detectable and representsa significant improvement compared to passive delivery. This suggests a future ad-ministration form where drugs, such as insulin or vaccines, can be delivered in anefficient, convenient as well as patient-friendly manner.

5.8 Discussion

In Table 3, the developed actuation methods presented in this thesis are evaluatedaccording to several features. These include fabrication issues, performance of theactuator as well as its versatility toward a wide range of different applications. Ta-ble 3 presents a comparison between the expandable microsphere actuator conceptsand several other actuation principles that clearly shows the advantages of the PDMScomposite actuator concept. The PDMS composite actuator exhibits beneficial char-acteristics due to its simple fabrication and integration as well as its flexibility to beused for various applications. This has been shown by means of fabricating variousmicrofluidic components for different purposes, e.g. positive and negative liquid dis-placement pumps, valves, as well as fully integrated networks and microneedle-basedtransdermal patches. The thermal impact on the manipulated liquid during opera-tion is moderate and could be further improved by introducing a buffer liquid into thesystem. Its low price and the possibility to fabricate highly integrated microfluidicnetworks makes this actuator interesting for applications in the medical field, suchas controlled and cost-effective transdermal microfluidic applications, point-of carediagnostics, disposable biochips or lab-on-a-chip applications.


Table 3. Comparison of various actuation methods. The different actuation principles areevaluated according to various attributes, such as fabrication, application and performance.If a technique features beneficial characteristics toward a certain attribute, it is marked aspositive, and in the contrary as negative.

Actuatorprinciples

Largescale inte-gration

High-volumeFabrication

Versatilitytoward ap-plications

Flexiblework-range

ForceImpacton liquid

Paraffin[44,46]

◦ ◦ ◦ + ++ ◦

Hydrogel[33,35,36]

+ + + + ◦ +

AIBN[43]

– – – + + ◦

Air-burst[49]

◦ ◦ ◦ + ++ +

Electrolysis[38,39]

– – ◦ ◦ + ◦ ◦

PneumaticPDMSvalves [31]

++ ++ ++ + + ++

Thermalexpansion[47,48]

– – ◦ ◦ + + ◦

XB in liquid[50]

– – – – + ++ – –

XB in paste[Paper 1]

– – ◦ ◦ + ++ ◦

PDMS-XB[Paper 2−6]

+ ++ ++ ++ + ◦

++ = very good, + = good, ◦ = neutral, – = poor, – – = very poor.

6 SUMMARY OF APPENDED PAPERS 49

They always say that time changesthings, but you actually have to changethem yourself.

Andy Warhol1928–1987, American Artist, Author

6 Summary of appended papers

Quick summary

Paper 1 shows the development of a liquid dispenser for dispensing of liquid volumesin the range from microliters to milliliters.

Paper 2 describes a novel thermally responsive composite actuator, which consistsof expandable microspheres incorporated in a PDMS matrix for controlled liquidactuation in microsystems.

Paper 3 reports on a disposable liquid handling lab-on-a-chip (LOC) platform withsmart microfluidic manipulation and on-chip active microfluidic components forapplications in analytical chemistry.

Paper 4 investigates disposable single-use microfluidic pumps for integrated activedosing in the microliter range, which are entirely based on a wafer-level fabrica-tion scheme.

Paper 5 shows the development of an electrically controlled microneedle-based trans-dermal patch, capable of controlled liquid release at low flow-rates.

Paper 6 reports on vertical, as well as lateral, liquid aspiration and dispensing units,which are based on an expanding PDMS composite actuator.

Paper 7 investigates a novel adhesive PDMS full-wafer bonding technique, whichutilizes PDMS curing-agent as the intermediate layer for adhesive bonding.

Detailed summary

Paper 1: A Compact, Low-cost Microliter-range Liquid Dispenser Based on Expand-able Microspheres

This paper presents a low-cost liquid dispenser for dispensing of microliters tomilliliter liquid volumes. The dispensing mechanism is based on a thermal actu-ator, where highly expandable microspheres expand into a liquid reservoir conse-quently displacing stored liquid. All device components are made out of low-costmaterials and the fabrication process has the potential for high volume batch


manufacturing. The device utilizes the property of the expandable microspheresto form a heat insulating layer between the heat source and the delivered liq-uid. Moreover, it does not require any feed back or complicated flow metering.The device was successfully tested for accurate and precise dispensing of liquidvolumes. The dispenser is capable of liquid delivery against counter pressures.The device can also function as a low flow-rate dispenser, as demonstrated in amicrofluidic dye laser application, where the flow rate is controlled by adjustingthe actuation power.

Paper 2: A Thermally Responsive PDMS Composite and Its Microfluidic Applica-tions

This work describes a novel composite actuator for controlled liquid actuation inmicrosystems, which is based on a thermally responsive elastomer. The compos-ite actuator consists of expandable microspheres incorporated in a polydimethyl-siloxane (PDMS) matrix and entails the merits of both PDMS and expandablemicrospheres. The main characteristic of the composite actuator is to expandirreversibly upon heat. The composite was used to fabricate single-use microflu-idic pumps for the displacement of liquids in the range of nanoliters, even againstcounter pressures. Moreover, liquid flow in microchannels was entirely blockedby means of the integrated valves. The devices are fabricated using low-cost ma-terials only, and the composite actuator allows using wafer-level processing. Thefluidic components based on the novel composite are highly integrable and do notrequire external actuators.

Paper 3: A Disposable Lab-on-a-chip Platform with Embedded Fluid Actuators forActive Nanoliter Liquid Handling

This paper reports on the development of a disposable liquid handling lab-on-a-chip (LOC) platform with embedded actuators for applications in analytical chem-istry. The proposed platform for nanoliter liquid handling is based on a thermallyresponsive silicone elastomer composite, consisting of PDMS and expandable mi-crospheres. In this LOC platform, we integrate active dosing, transportation andmerging of nanoliter liquid volumes. The disposable platform successfully demon-strates precise sample volume control with smart microfluidic manipulation andon-chip active microfluidic components. It is entirely fabricated from low-cost ma-terials using wafer-level processing. Moreover, enzymatic reaction and real-timedetection was successfully conducted to exemplify its applicability as an LOC.

Paper 4: Wafer-level Process for Single-use Buckling-film Microliter-range Pumps

This paper presents the development of disposable single-use microfluidic pumpsentirely based on a straight-forward wafer-level fabrication scheme, which allowsfor precise and accurate integrated active dosing in the microliter range. To ac-complish stroke-lengths needed for microliter range applications we utilize a newmethod of bending of a unimorph composite actuator film. The unimorph com-posite actuator consists of a temperature sensitive silicone elastomer composite,i.e. PDMS with incorporated expandable microspheres. The fabricated microp-

6 SUMMARY OF APPENDED PAPERS 51

umps successfully demonstrated precise liquid volume control, both at low andhigh flow-rates. Moreover, the method of fluorescent thermometry was used tomeasure the thermal load on liquid volumes dispensed with the micropumps. Thepresented fully integrated single-use micropumps are electrically controllable, donot require external means for liquid actuation, are made of low-cost materialsonly and might potentially be used in drug delivery applications.

Paper 5: Painless Drug Delivery through Microneedle-based Transdermal Patchesfeaturing Active Infusion

This paper reports on the first microneedle-based transdermal patch with inte-grated active dispensing functionality. The electrically controlled system consistsof a low-cost dosing and actuation unit capable of controlled release of liquid, inthe microliter range, at low flow-rates, through minimally invasive, side-opened,microneedles. The system has been tested successfully in-vivo by insulin ad-ministration to diabetic rats. Active infusion of insulin is compared to passive,diffusion-driven, delivery. Continuous active infusion causes significantly higherinsulin concentrations in blood plasma. Consistent with insulin concentrations,actively administered insulin results in a significant decrease of blood glucose lev-els. Additionally, insertion and liquid injection is verified on human skin. Thestudy shows the feasibility of a patch-like system with on-board liquid storageand dispensing capability. The proposed device represents a first step towardspainless and convenient administration of macromolecular drugs, such as insulinor vaccines.

Paper 6: Liquid Aspiration and Dispensing Based on an Expanding PDMS Compos-ite

This paper presents the development of active liquid aspiration and dispensingunits designed for vertical as well as lateral liquid aspiration. The devices arebased on a single-use thermally expanding PDMS composite actuator consistingof expandable microspheres incorporated into a PDMS matrix, which allows forlocally altering its surface topography by means of generated heat. Liquid aspira-tion is induced by creating an enclosed cavity upon the expansion of the compositeactuator, which in turn leads to generation of negative pressure in the system.To accomplish this functionality, two different techniques of bonding PMDS se-lectively on wafer-scale have been developed. Such selective PDMS bonding isachieved utilizing plasma assisted bonding with a patterned anti-sticktion layeror adhesive bonding via microcontact printed PDMS catalyst as the intermediatelayer for adhesive bonding. Fabricated devices successfully demonstrated activelycontrolled vertical as well as lateral liquid aspiration and release of liquid volumes.The devices are entirely fabricated from low-cost materials using wafer-level pro-cesses only and do not require external means for liquid actuation.

Paper 7: The Fabrication of Microfluidic Structures by means of Full-wafer AdhesiveBonding using a Poly(dimethylsiloxane) Catalyst

An adhesive bonding method, using PDMS curing-agent as the intermediate layer,


is presented for adhesive full-wafer bonding suitable for fabrication of microfluidicstructures. The curing agent of the two-component silicone rubber (Sylgard 184)is spin coated on a substrate, brought into contact with another PDMS layer andheat cured to create an irreversible seal which is as strong as, or even strongerthan, plasma assisted PDMS bonding. The applicability of the new PDMS adhe-sive bonding method is verified by means of fabricating microfluidic structures.Using this method allows for wafer-level bonding of PDMS to various materialssuch as PDMS itself, glass or silicon and more importantly to selectively bonddifferent layers by using a patterned adhesive bonding technique. Moreover, pre-cise alignment of the structural layers is facilitated, since curing is initiated uponheat, which is an advantage when fabricating multilayer microfluidic devices.

7 CONCLUSIONS 53

As for the future, your task is not toforsee it, but to enable it.

Antoine de Saint-Exupery1900-1944, French Aviator, Writer

7 Conclusions

This thesis presents novel microfluidic components based on the use of expandablemicrospheres, primarily incorporated in a PDMS matrix, which results in a thermallyresponsive composite actuator. Microfluidic components are fabricated in a multi-layer fashion consisting of a patterned PCB with integrated heaters, the thermallyresponsive actuator and a structural PDMS layer comprising microfluidic networks orpumps. This approach is characterized by the following advantages:

• simplified and convenient handling of expandable microspheres

• capability to design microfluidic pumps operating over a wide range of liquidvolumes (nanoliters to microliters)

• highly integrable actuator allows for on-chip integration of pumps, valves, mixersand reactors

• handling of temperature sensitive liquids

• precise microfluidic control

• straightforward fabrication process

• wafer-level fabrication for high volume production

• entirely fabricated using low-cost materials allowing for disposable applications

• no external liquid control with complicated fluidic interconnections

• low system complexity

• high reliability

The thesis

• introduces and discusses the concept of incorporating expandable microspheresinto other materials

• presents results on the characterization of a novel thermally responsive compos-ite actuator

• provides information on design and fabrication issues concerning the compositeactuator


• reports on the performance of fabricated microfluidic prototype components

• shows the versatility of the composite actuator toward various microfluidic com-ponents

Prototype devices utilizing expandable microspheres show great potential for the mi-crofludic field, despite the single-use characteristic of the actuator. The thermally re-sponsive PDMS composite was found to be suitable for single-use microfluidic devices,which could be used in the medical field as therapeutic or diagnostic applications.

Furthermore, this thesis presents a novel idea of bonding PDMS by an adhesivefull-wafer bonding approach utilizing the curing-agent of the two-component siliconerubber (Sylgard R© 184) as the intermediate layer for adhesive bonding:

• The process parameters of this adhesive bonding technique were investigated.The technique was successfully applied to fabricate microchannels inside PDMSlayers.

• Various fabrication techniques were applied to create directly bonded, patternedadhesive bonded and selectively bonded microchannels.

• Selective PDMS bonding has been accomplished via microcontact printing ofPDMS curing-agent, as the intermediate layer, for full-wafer adhesive bonding.

• PDMS bonding of various materials such as PDMS, silicon and glass has suc-cessfully been demonstrated on a wafer-scale.

• Bond strength measurements have been conducted using the blister test methodand show that using this method enables the formation of a bond interface thatis comparable to, or stronger than, plasma assisted bonding.

• When applying this bonding technique precise alignment of structural layersis facilitated, since curing is initiated upon heat, which is an advantage whenfabricating multilayer microfluidic devices.

• This adhesive bonding technique is compatible with room-temperature curingand would therefore allow for the functionalized immobilization of biomoleculesin microstructures prior to sealing.

• Spin coating the curing-agent allows microfluidic components to be fabricatedand sealed irreversibly at very low cost without the need for expensive laboratoryequipment.

Adhesive full-wafer PDMS bonding utilizing solely the curing-agent of a two-component silicone rubber was found to be a simple process compared to the com-monly used plasma assisted bonding process. No expensive laboratory equipment isneeded and the time-consuming adjustment of process parameters is no longer re-quired.

ACKNOWLEDGEMENTS 55

Great discoveries and improvementsinvariably involve the cooperation ofmany minds. I may be given credit forhaving blazed the trail, but when I lookat the subsequent developments I feelthe credit is due to others rather thanto myself.

Alexander Graham Bell1847–1922, Scottish-born Canadian

Inventor, Educator, Telephone Pioneer

Acknowledgements

At this day I look back and realize that all this would not have been possible withoutthe help and guidance of so many people. I am deeply grateful to everybody who hashelped to pave my way, whether it may have been a laugh we spent together duringa quick break or simply a discussion about work or life.

The work presented in this thesis has been carried out at the Microsystem Tech-nology Laboratory at the School of Electrical Engineering at the Royal Institute ofTechnology in Stockholm, Sweden. Financial support was provided in part by theSwedish Foundation of Strategic Research (SFF) and in part by the FP6-IST-SABIOProject (026554), which is funded by the European Commission.

I would like to take the opportunity to show my heartfelt thanks to a number ofpeople in particular:

First of all, I would like to express my sincere gratitude to my main supervisorGoran Stemme for always having his door open, giving me support and guidanceas well as the freedom to develop my own ideas, and not least for giving me theopportunity to work in his group. It has been a great honor and a true privilege towork together with you.

Patrick ’Schliifer’ Griss for being my supervisor, although being at such a remotelocation and for giving me guidance, support as well as the freedom to make my ownmistakes.

Sjoerd ’Sjoerdje’ Haasl, whom I trusted blindly as a thesis student and who sobrazenly dragged me into the MEMS field with his unique charm and enthusiasm.You in the first place inspired me to pursue a Ph.D. degree.

Wouter ’Rambo’ van der Wijngaart for passionately trying to give me an under-standing of the fundamental effects in microfluidics and for his respectful attitude.

Stefan ’the German’ Braun for establishing an outpost in the midst of the yet socontrolling Swedish sector; the ’German office’, which invites so much to linger on.Thank you for being my office mate, creating a great atmosphere and being such agood friend.

Niclas ’Gunther’ Roxhed, for fruitfully and truthfully discussing work and life andhow to make a difference.


Joachim ’Incremental’ Oberhammer for letting me use the thesis template, sopatiently answering all my LATEX related questions and for discussing the delicateissue, amongst others, of what is regarded novel and what is not.

Frank ’Jacuzzi’ Niklaus for relaxed discussions at work and for reading stories totwo grateful children.

Thomas ’Old Salt’ Frisk for keenly teaching me naughty as well as nautical vo-cabulary.

Petra ’Petrice’ Palmquist for so accurately proofreading this thesis and for sopassionately trying to teach me sundry grammatical rules.

Furthermore, I would like to express my honest gratitude to the people that puttheir trust in me as thesis or project students: Volker Nock, Julien Chretien, StephanSchroder, Mohammad Kamruzzaman Chowdhury, Niklas Sandstrom. It has been agreat pleasure working together with you and I hope it was as rewarding for you asit was for me; I learned more from you than you can imagine.

Ruifeng Yue, whom I had the pleasure of fruitfully working together with duringhis stay in Sweden.

All my former and present colleagues and friends at our group for creating sucha great working environment and for sharing so many memorable moments: Adit,Aman, Anna, Calle (We miss you!), Erika, Farizah, Frank, Fredrik C., Fredrik F.,Goran, Hans, Jessica, Joachim, Kjell, Kristinn, Martin, Mikael, Niclas, Niklas, Nu-tapong, Patrick, Patrik, Peter, Sjoerd, Stefan, Thomas, Ulla, Wouter. It has alwaysbeen a great pleasure working together with you, talking shop and sharing differentviews on life in Sweden.

Thanks to my colleagues and the laboratory staff in Kista. In particular I wouldlike to thank Cecilia Aronsson for never letting me down whenever asking yet sostupid questions and for so persistently teaching me your language.

Everybody that joined the weekly “Tisdagspub” and all the lads I had the pleasureof playing “Innebandy” with. It has always been a great pleasure meeting you guys.

Werner Zapka, Peter Lachmann, Georg Schulz and Joachim Werner for makingmy first visits to Sweden possible.

The Germans for keeping up the good friendship and their willingness to conveneevery time I am back home, even on short notice: Benjamin, Birgit, Daniel, Denise,Dierk, Marcus, Miriam, Nici, Peter, Silke, Simone, Stefan, Stefan, Sven, Thomas,Tina, Tine.

My parents Barbara and Wolfgang, my brother Dierk and my sister Christine fortheir constant support and understanding, teaching me the important things in lifeand for the home where I am always welcome.

Finally, I wish to express my deepest thank you to my beloved, for giving meendless love and support. I want to thank Nadja, Noah and Petra, not only forteaching me the true meaning of emancipation, but also for putting a smile on myface every single day. You have been the greatest that has ever happened to me.“Thank you!”

REFERENCES 57

He who has imagination without learn-ing has wings but no feet.

Joseph Joubert1754–1824, French Essayist, Moralist

References

[1] K. Petersen, “Silicon as a mechanical material,” Proceedings of the IEEE,vol. 70, no. 5, pp. 420–457, 1982.

[2] G. T. Kovacs, Micromachined Transducers Sourcebook. WCB/McGraw-Hill,New York, 1998.

[3] M. J. Madou, Fundamentals of Microfabrication: the Science of Miniaturiza-tion. CRC Press, USA, 2001.

[4] A. Manz, D. J. Harrison, E. M. J. Verpoorte, J. C. Fettinger, A. Paulus,H. Ludi, and H. M. Widmer, “Planar chips technology for miniaturizationand integration of separation techniques into monitoring systems : Capillaryelectrophoresis on a chip,” Journal of Chromatography A, vol. 593, no. 1-2, pp.253–258, Feb. 1992.

[5] D. J. Harrison, K. Fluri, K. Seiler, Z. Fan, C. S. Effenhauser, andA. Manz, “Micromachining a miniaturized capillary electrophoresis-basedchemical analysis system on a chip,” Science, vol. 261, no. 5123, pp. 895–897,1993.

[6] S. Shoji and M. Esashi, “Microflow devices and systems,” Journal of Microme-chanics and Microengineering, vol. 4, pp. 157–171, 1994.

[7] A. Manz and H. Becker, Eds., Microsystem Technology in Chemistry and LifeSciences, Topics in current chemistry 194. Springer-Verlag, Berlin, Germany,1998.

[8] K. B. Mogensen, N. J. Petersen, J. Hubner, and J. P. Kutter, “Monolithicintegration of optical waveguides for absorbance detection in microfabricatedelectrophoresis devices,” Electrophoresis, vol. 22, pp. 3930–3938, 2001.

[9] J. Hubner, K. Mogensen, A. Jorgensen, P. Friis, P. Telleman, and J. Kutter, “In-tegrated optical measurement system for fluorescence spectroscopy in microflu-idic channels,” Review of Scientific Instruments, vol. 72, no. 1, pp. 229–233,2001.

[10] K. B. Mogensen, H. Klank, and J. P. Kutter, “Recent developments indetection for microfluidic systems,” Electrophoresis, vol. 25, no. 21-22, pp.3498–3512, 2004.

[11] H. Becker and C. Gartner, “Polymer microfabrication methods for microfluidicanalytical applications,” Electrophoresis, vol. 21, no. 1, pp. 12–26, 2000.


[12] H. Becker and L. E. Locascio, “Polymer microfluidic devices,” Talanta, vol. 56,no. 2, pp. 267–287, Feb. 2002.

[13] A. Manz, N. Graber, and H. Widmer, “Miniaturized total chemical analysissystems: A novel concept for chemical sensing,” Sensors and Actuators B:Chemical, vol. 1, no. 1-6, pp. 244–248, Jan. 1990.

[14] H. Varmus, R. Klausner, E. Zerhouni, T. Acharya, A. S. Daar, and P. A.Singer, “Public health: Enhanced: Grand challenges in global health,” Science,vol. 302, no. 5644, pp. 398–399, 2003.

[15] A. S. Daar, H. Thorsteinsdottir, D. K. Martin, A. C. Smith, S. Nast, and P. A.Singer, “Top ten biotechnologies for improving health in developing countries,”Nature Genetics, vol. 32, no. 2, pp. 229–232, Oct. 2002.

[16] C. D. Chin, V. Linder, and S. K. Sia, “Lab-on-a-chip devices for global health:Past studies and future opportunities,” Lab Chip, vol. 7, pp. 41–57, 2007.

[17] B. H. Weigl, J. Gerdes, P. Tarr, P. Yager, L. Dillman, R. Peck,S. Ramachandran, M. Lemba, M. Kokoris, M. Nabavi, F. Battrell, D. Hoekstra,E. J. Klein, and D. M. Denno, “Fully integrated multiplexed lab-on-a-cardassay for enteric pathogens,” I. Papautsky and W. Wang, Eds., vol. 6112,no. 1. SPIE, 2006, p. 611202.

[18] D. J. Harrison and A. V. D. Berg, Eds., Micro Total Analysis Systems 98.Kluwer Academic Publishers: Dordrecht, 1998.

[19] D. R. Reyes, D. Iossifidis, P.-A. Auroux, and A. Manz, “Micro total analysissystems. 1. introduction, theory, and technology,” Analytical Chemistry, vol. 74,pp. 2623–2636, 2002.

[20] P.-A. Auroux, D. Iossifidis, D. R. Reyes, and A. Manz, “Micro total analysissystems. 2. analytical standard operations and applications,” Analytical Chem-istry, vol. 74, pp. 2637–2652, 2002.

[21] T. Vilkner, D. Janasek, and A. Manz, “Micro total analysis systems. recentdevelopments,” Analytical Chemistry, vol. 76, no. 12, pp. 3373–3386, 2004.

[22] P. Dittrich, K. Tachikawa, and A. Manz, “Micro total analysis systems.latest advancements and trends,” Analytical Chemistry, vol. 78, no. 12, pp.3887–3908, 2006.

[23] A. van den Berg and T. Lammerink, Microsystem Technology in Chemistry andLife Sciences, Topics in current chemistry 194. Springer Verlag, Berlin, 1998,ch. Micro Total Analysis Systems: Microfluidic Aspects, Integration Conceptand Applications, p. 21 49.

[24] H. Andersson, W. van der Wijngaart, P. Nilsson, P. Enoksson, and G. Stemme,“Valve-less diffuser micropump for microfluidic analytical systems,” Sensors andActuators, B: Chemical, pp. 259–265, Feb. 2001.

REFERENCES 59

[25] D. Armani and C. Liu, “Microfabrication technology for polycaprolactone,a biodegradable polymer,” Journal of Micromechanics and Microengineering,vol. 10, pp. 80–84, 2000.

[26] X. Wang, J. Engel, and C. Liu, “Liquid crystal polymer (lcp) for mems: pro-cesses and applications,” Journal of Micromechanical Microengineering, vol. 13,pp. 628–633, 2003.

[27] M. Dokmeci and K. Najafi, “A high-sensitivity polyimide capacitive relativehumidity sensor for monitoring anodically bonded hermetic micropackages,”Journal of Microelectromechanical Systems, vol. 10, no. 2, pp. 197–204, 2001.

[28] H. Klank, J. P. Kutter, and O. Geschke, “Co2-laser micromachining and back-end processing for rapid production of pmma-based microfluidic systems,” LabChip, vol. 2, pp. 242–246, 2002,.

[29] S. Balslev, A. M. Jorgensen, B. Bilenberg, K. B. Mogensen, D. Snakenborg,O. Geschke, J. P. Kutter, and A. Kristensen, “Lab-on-a-chip with integratedoptical transducers,” Lab Chip, vol. 6, pp. 213–217, 2006.

[30] K. Mogensen, J. El-Ali, A. Wolff, and J. Kutter, “Integration of polymer waveg-uides for optical detection in microfabricated chemical analysis systems,” Ap-plied Optics, vol. 42, no. 19, pp. 4072–4079, 2003.

[31] M. A. Unger, H.-P. Chou, T. Thorsen, A. Scherer, and S. R. Quake,“Monolithic microfabricated valves and pumps by multilayer soft lithography,”Science, vol. 288, pp. 113–116, 2000.

[32] R. F. Ismagilov, D. Rosmarin, P. J. A. Kenis, D. T. Chiu, W. Zhang, H. A.Stone, and G. M. Whitesides, “Pressure-driven laminar flow in tangentialmicrochannels: an elastomeric microfluidic switch,” Analytical Chemistry, vol.73 (19), pp. 4682–4687, 2001.

[33] D. J. Beebe, J. S. Moore, J. M. Bauer, Q. Yu, R. H. Liu, C. Devadoss, andB.-H. Jo, “Functional hydrogel structures for autonomous flow control insidemicrofluidic channels,” Nature, vol. 404, no. 6778, pp. 588–590, Apr. 2000.

[34] R. Liu, Q. Yu, and D. Beebe, “Fabrication and characterization of hydrogel-based microvalves,” Journal of Microelectromechanical Systems, vol. 11 (1), pp.45–53, 2002.

[35] C. Yu, S. Mutlu, P. Selvaganapathy, C. Mastrangelo, F. Svec, and J. Frechet,“Flow control valves for analytical microfluidic chips without mechanical partsbased on thermally responsive monolithic polymers,” Analytical Chemistry,vol. 75, no. 8, pp. 1958–1961, 2003.

[36] A. Richter, D. Kuckling, S. Howitz, T. Gehring, and K.-F. Arndt,“Electronically controllable microvalves based on smart hydrogels: magnitudesand potential applications,” Journal of Microelectromechanical Systems, vol. 12,no. 5, pp. 748 – 53, 2003.


[37] D. Duffy, H. Gillis, J. Lin, N. Sheppard, and G. Kellogg, “Microfabricatedcentrifugal microfluidic systems: Characterization and multiple enzymaticassays,” Analytical Chemistry, vol. 71, no. 20, pp. 4669–4678, 1999.

[38] S. Bohm, W. Olthuis, and P. Bergveld, “An integrated micromachinedelectrochemical pump and dosing system,” Biomedical Microdevices, vol. 1,no. 2, pp. 121–130, 1999.

[39] J. Munyan, H. Fuentes, M. Draper, R. Kelly, and A. Woolley, “Electricallyactuated, pressure-driven microfluidic pumps,” Lab Chip, vol. 3, no. 4, pp.217–220, 2003.

[40] A. Brask, J. Kutter, and H. Bruus, “Long-term stable electroosmotic pumpwith ion exchange membranes,” Lab Chip, vol. 5, pp. 730–738, 2005.

[41] A. Brask, D. Snakenborg, J. Kutter, and H. Bruus, “Ac electroosmotic pumpwith bubble-free palladium electrodes and rectifying polymer membrane valves,”Lab Chip, vol. 6, pp. 280–288, 2006.

[42] Y. H. Choi, S. Son, and S. S. Lee, “Novel micropump using oxygen as pumpingsource,” in Proc. of MEMS 2003, 2003, pp. 116–119.

[43] C.-C. Hong, S. Murugesan, S. Kim, G. Beaucage, J.-W. Choi, and C. Ahn,“A functional on-chip pressure generator using solid chemical propellant fordisposable lab-on-a-chip,” Lab Chip, vol. 3, no. 4, pp. 281–286, 2003.

[44] E. Carlen and C. Mastrangelo, “Electrothermally activated paraffin microac-tuators,” Journal of Microelectromechanical Systems, vol. 11 (3), pp. 165–174,2002.

[45] ——, “Surface micromachined paraffin-actuated microvalve,” Journal ofMicroelectromechanical Systems, vol. 11 (5), pp. 408–420, 2002.

[46] R. Boden, M. Lehto, U. Simu, G. Thornell, K. Hjort, and J.-A. Schweitz, “Apolymeric paraffin actuated high-pressure micropump,” Sensors and ActuatorsA: Physical, vol. 127, no. 1, pp. 88–93, Feb. 2006.

[47] X. Yang, C. Grosjean, Y.-C. Tai, and C.-M. Ho, “A mems thermopneumaticsilicone rubber membrane valve,” Sens. Actuators A, Phys., vol. A64, no. 1,pp. 101–108, 1998.

[48] C. Cooney and B. Towe, “A thermopneumatic dispensing micropump,” Sens.Actuators A, Phys., vol. A116, no. 3, pp. 519 –524, 2004.

[49] C.-C. Hong, J.-W. Choi, and C. H. Ahn, “An on-chip air-bursting detonator fordriving fluids on disposable lab-on-a-chip systems,” Journal of Micromechanicsand Microengineering, vol. 17, no. 2, pp. 410–417, 2007.

[50] P. Griss, H. Andersson, and G. Stemme, “Expandable microspheres for thehandling of liquids,” Lab Chip, vol. 2, no. 2, pp. 117–120, May 2002.

REFERENCES 61

[51] M. K. N. Jessamine, I. Gitlin, A. D. Stroock, and G. M. Whitesides, “Compo-nents for integrated poly(dimethylsiloxane) microfluidic systems,” Electrophore-sis, vol. 23, pp. 3461–3473, 2002.

[52] H. Wu, T. Odom, D.T.Chiu, and G. Whitesides, “Fabrication of complex three-dimensional microchannel systems in pdms,” Journal of American ChemicalSociety, vol. 125, pp. 554–559, 2003.

[53] B.-H. Jo, L. Van Lerberghe, K. Motsegood, and D. Beebe, “Three-dimensionalmicro-channel fabrication in polydimethylsiloxane(pdms) elastomer,” Journalof Microelectromechanical Systems, vol. 9 (1), pp. 76–81, 2000.

[54] B. Samel, P. Griss, and G. Stemme, “A thermally responsive pdms compositeand its microfluidic applications,” Journal of Microelectromechanical Systems,vol. 16, no. 1, pp. 50–57, 2007.

[55] B. Samel, V. Nock, A. Russom, P. Griss, and G. Stemme, “A disposable lab-on-a-chip platform with embedded fluid actuators for active nanoliter liquidhandling,” Biomedical Microdevices, vol. 9, no. 1, pp. 61–67, 2007.

[56] M. L. Adams, M. L. Johnston, A. Scherer, and S. R. Quake, “Poly-dimethylsiloxane based microfluidic diode,” Journal of Micromechanics andMicroengineering, vol. 15, no. 8, pp. 1517–1521, 2005.

[57] J. C. Lotters, W. Olthuis, P. H. Veltink, and P. Bergveld, “Polydimethylsiloxaneas an elastic material applied in a capacitive accelerometer,” Journal ofMicromechanics and Microengineering, vol. 6, no. 1, pp. 52–54, 1996.

[58] M. Antonisse, J. Engbersen, and D. Reinhoudt, “Nitrate and bicarbonate se-lective chemfets,” in Eurosensors IX - Transducers ’95. The 8th InternationalConference on Solid-State Sensors, Actuators and Microsystems, vol. 1, 1995,pp. 867–869.

[59] J. C. Lotters, W. Olthuis, P. H. Veltink, and P. Bergveld, “The mechanicalproperties of the rubber elastic polymer polydimethylsiloxane for sensorapplications,” Journal of Micromechanics and Microengineering, vol. 7, no. 3,pp. 145–147, 1997.

[60] J. McDonald and G. Whitesides, “Poly(dimethylsiloxane) as a material forfabricating microfluidic devices,” Accounts of Chemical Research, vol. 35, no. 7,pp. 491–499, 2002.

[61] A. Singh, B. D. Freeman, and I. Pinnau, “Pure and mixed gas acetone/nitrogenpermeation properties of polydimethylsiloxane [pdms],” Journal of PolymerScience Part B: Polymer Physics, vol. 36, no. 2, pp. 289–301, 1998.

[62] R. Prabhakar, T. Merkel, B. Freeman, T. Imizu, and A. Higuchi, “Sorption andtransport properties of propane and perfluoropropane in poly(dimethylsiloxane)and poly(1-trimethylsilyl-1-propyne),” Macromolecules, vol. 38, no. 5, pp.1899–1910, 2005.


[63] M. A. Eddings and B. K. Gale, “A pdms-based gas permeation pump foron-chip fluid handling in microfluidic devices,” Journal of Micromechanics andMicroengineering, vol. 16, no. 11, pp. 2396–2402, 2006.

[64] J. McDonald, D. Duffy, J. Anderson, D. Chiu, H. Wu, O. Schueller, andG. Whitesides, “Fabrication of microfluidic systems in poly(dimethylsiloxane),”Electrophoresis, vol. 21, pp. 27–40, 2000.

[65] Y. Xia and G. M. Whitesides, “Soft lithography,” Angewandte ChemieInternational Edition, vol. 37, no. 5, pp. 550–575, 1998.

[66] S. Thorslund, “Microfluidics in surface modified pdms, towards minaturized di-agnostics tools,” Ph.D. dissertation, Department of Engineering Sciences, Up-psala Universitet, 2006.

[67] FluidigmCorporation, “Nanoflex valves,” http://www.fluidigm.com.

[68] H.-P. Chou, M. A. Unger, and S. R. Quake, “A microfabricated rotary pump,”Biomedical Microdevices, vol. 3, no. 4, pp. 323–330, Dec. 2001.

[69] J. Liu, C. Hansen, and S. Quake, “Solving the ”world-to-chip” interfaceproblem with a microfluidic matrix,” Analytical Chemistry, vol. 75, no. 18, pp.4718–4723, 2003.

[70] T. Thorsen, S. J. Maerkl, and S. R. Quake, “Microfluidic large-scaleintegration,” Science, vol. 298, no. 5593, pp. 580–584, 2002.

[71] W. Grover, R. Ivester, E. Jensen, and R. Mathies, “Development and mul-tiplexed control of latching pneumatic valves using microfluidic logical struc-tures,” Lab Chip, vol. 6, pp. 623 – 631, 2006.

[72] W. H. Grover, A. M. Skelley, C. N. Liu, E. T. Lagally, and R. A. Mathies,“Monolithic membrane valves and diaphragm pumps for practical large-scaleintegration into glass microfluidic devices,” Sensors and Actuators B: Chemical,vol. 89, no. 3, pp. 315–323, Apr. 2003.

[73] J. Melin and S. R. Quake, “Microfluidic large-scale integration: The evolutionof design rules for biological automation,” Annual Review of Biophysics andBiomolecular Structure, vol. 36, no. 1, pp. 213–231, 2007.

[74] S. Iyer and A. Auberton-Herve, Eds., Silicon Wafer Bonding Technology forVLSI and MEMS Applications. INSPEC, London, 2002.

[75] Q.-Y. Tong and U. Gosele, SemiConductor Wafer Bonding: Science and Tech-nology. John Wiley & Sons, Inc., New York, 1999.

[76] M. Alexe and U. Gosele, Eds., Wafer Bonding: Applications and Technology.Springer-Verlag, Berlin, Heidelberg, New York, 2004.

[77] Y. Cheng, L. Lin, and K. Najafi, “Localized silicon fusion and eutectic bond-ing for mems fabrication and packaging,” Journal of MicroelectromechanicalSystems, vol. 9, no. 1, pp. 3–8, 2000.

REFERENCES 63

[78] S. Audet and K. Edenfeld, “Integrated sensor wafer-level packaging,” inTRANSDUCERS ’97 Chicago, International Conference on Solid-State Sen-sors, Actuators and Microsystems, vol. 1, 1997, pp. 287–289 vol.1.

[79] D. Sparks, G. Queen, R. Weston, G. Woodward, M. Putty, L. Jordan,S. Zarabadi, and K. Jayakar, “Wafer-to-wafer bonding of nonplanarizedmems surfaces using solder,” Journal of Micromechanics and Microengineering,vol. 11, no. 6, pp. 630–634, 2001.

[80] M. Maharbiz, M. Cohn, R. Howe, R. Horowitz, and A. Pisano, “Batch mi-cropackaging by compression-bonded wafer-wafer transfer,” in Micro ElectroMechanical Systems, 1999. MEMS ’99. Twelfth IEEE International Conferenceon, 1999, pp. 482–489.

[81] J. Kim, M. Chiao, and L. Lin, “Ultrasonic bonding of in/au and al/al for her-metic sealing of mems packaging,” in Proceedings IEEE International Confer-ence on Micro Electro Mechanical Systems (MEMS), 2002, pp. 415–418.

[82] F. Niklaus, G. Stemme, J. Q. Lu, and R. J. Gutmann, “Adhesive waferbonding,” Journal of Applied Physics, vol. 99, no. 3, pp. 1–28, 2006.

[83] R. Smith and S. Collins, “Micromachined packaging for chemical microsensors,”Electron Devices, IEEE Transactions on, vol. 35, no. 6, pp. 787–792, 1988.

[84] R. Jackman, T. Floyd, R. Ghodssi, M. Schmidt, and K. Jensen, “Microfluidicsystems with on-line uv detection fabricated in photodefinable epoxy,” Journalof Micromechanics and Microengineering, vol. 11, pp. 263–269(7), 2001.

[85] K. Oh, A. Han, i. S. Bhansal, and C. Ahn, “A low-temperature bondingtechnique using spin-on fluorocarbon polymers to assemble microsystems,”Journal of Micromechanics and Microengineering, vol. 12, pp. 187–191(5),2002.

[86] T. T. Veenstra, J. W. Berenschot, J. G. E. Gardeniers, R. G. P. Sanders,M. C. Elwenspoek, and A. van den Berg, “Use of selective anodic bonding tocreate micropump chambers with virtually no dead volume,” Journal of TheElectrochemical Society, vol. 148, no. 2, pp. G68–G72, 2001.

[87] W. P. Eaton, S. H. Risbud, and R. L. Smith, “Silicon wafer-to-wafer bondingat t ¡ 200 ◦c with polymethylmethacrylate,” Applied Physics Letters, vol. 65,no. 4, pp. 439–441, 1994.

[88] V. Spiering, J. Berenschot, M. Elwenspoek, and J. Fluitman, “Sacrificial waferbonding for planarization after very deep etching,” Journal of Microelectrome-chanical Systems, vol. 4, no. 3, pp. 151–157, 1995.

[89] S. Matsuo, K. Tateno, T. Nakahara, H. Tsuda, and T. Kurokawa, “Use ofpolyimide bonding for hybrid integration of a vertical cavity surface emittinglaser on a silicon substrate,” Electronics Letters, vol. 33, no. 13, pp. 1148–1149,1997.


[90] P. Absil, J. Hryniewicz, B. Little, F. Johnson, K. Ritter, and P.-T. Ho, “Ver-tically coupled microring resonators using polymer wafer bonding,” IEEE Pho-tonics Technology Letters, vol. 13, no. 1, pp. 49–51, 2001.

[91] W. M. Alvino, Plastics for Electronics: Materials, Properties, and Design Ap-plications. McGraw-Hill, Inc., New York, 1995.

[92] F. Niklaus, “Adhesive wafer bonding for microelectronic and microelectrome-chanical systems,” Ph.D. dissertation, Royal Institute of Technology (KTH),Sept. 2002, tRITA-ILA-0204.

[93] D. Duffy, J. McDonald, O. Schueller, and G. Whitesides, “Rapid prototyping ofmicrofluidic systems in poly(dimethylsiloxane),” Analytical Chemistry, vol. 70,no. 23, pp. 4974–4984, 1998.

[94] M. J. Owen and P. J. Smith, “Plasma treatment of polydimethylsiloxane,” Jour-nal of adhesion science and technology, vol. 8, pp. 1063–1075, 1994.

[95] S. Bhattacharya, A. Datta, J. Berg, and S. Gangopadhyay, “Studies on surfacewettability of poly(dimethyl) siloxane (pdms) and glass under oxygen-plasmatreatment and correlation with bond strength,” Journal of Microelectromechan-ical Systems, vol. 14, no. 3, pp. 590–597, 2005.

[96] K. Haubert, T. Drier, and D. Beebe, “Pdms bonding by means of a portable,low-cost corona system,” Lab Chip, vol. 6, pp. 1548– 1549, 2006.

[97] M. Morra, E. Occhiello, R. Marola, F. Garbassi, P. Humphrey, and D. Johnson,“On the aging of oxygen plasma-treated polydimethylsiloxane surfaces,”Journal of Colloid and Interface Science, vol. 137, pp. 11–24, 1990.

[98] S. Selvarasah, S. H. Chao, C. L. Chen, D. Mao, J. Hopwood, S. Ryley,A. Khademhosseini, A. Busnaina, and M. R. Dokmeci, “A high aspect ratio,flexible, transparent and low-cost parylene-c shadow mask technology for mi-cro patterning applications,” in Proceedings IEEE International Conference onSolid-State Sensors, Actuators, and Microsystems (Transducers), vol. 1, 2007,pp. 533–536.

[99] C. F. Carlborg, T. Haraldsson, G. Stemme, and W. van der Wijngaart, “Reliablebatch manufacturing of miniaturized vertical vias in soft polymer replica mold-ing,” in Proceedings Micro Total Analysis Systems (muTAS), 2007, accepted forpublication.

[100] A. A. S. Bhagat, P. Jothimuthu, and I. Papautsky, “Photosensitivepoly(dimethylsiloxane) (photopdms) for rapid and simple polymer fabrication,”in Proceedings IEEE International Conference on Solid-State Sensors, Actua-tors, and Microsystems (Transducers), vol. 1, no. 537–540, 2007.

[101] H. Wu, B. Huang, and R. H. Zare, “Construction of microfluidic chips usingpolydimethylsiloxane for adhesive bonding,” Lab Chip, vol. 5, pp. 1393–1398,2005.

REFERENCES 65

[102] J. Lee, C. Park, and G. Whitesides, “Solvent compatibility ofpoly(dimethylsiloxane)-based microfluidic devices,” Analytical Chemistry,vol. 75, no. 23, pp. 6544–6554, 2003.

[103] L. Kangsun, K. Choong, S. S. Kyeong, L. Jin Woo, J. Byeong-Kwon,K. Tae Song, L. Seung-Ki, and K. Ji Yoon, “Fabrication of round channelsusing the surface tension of pdms and its application to a 3d serpentinemixer,” Journal of Micromechanics and Microengineering, vol. 17, no. 8, pp.1533–1541, 2007.

[104] J. Oberhammer and G. Stemme, “Bcb contact printing for patterned adhesivefull-wafer bonded 0-level packages,” Journal of Microelectromechanical Systems,vol. 14, no. 2, pp. 419–425, Apr. 2005.

[105] H. Dannenberg, “Measurement of adhesion by a blister method,” Journal ofApplied Polymer Science, vol. 5, pp. 125–134, 1961.

[106] C. Li, J. Han, and C. H. Ahn, “Flexible biosensors on spirally rolled micro tubefor cardiovascular in vivo monitoring,” Biosensors and Bioelectronics, vol. 22,no. 9-10, pp. 1988–1993, Apr. 2007.

[107] D. Morehouse, M. Midland, and R. Tetreault, “Expansible thermoplastic poly-mer particles containing volatile fluid foaming agent and method of foaming thesame,” United States Patent 3,615,972, October 1971.

[108] J. Petersen, “Production of expandable beads from the thermoplastic materi-als,” European Patent 4,105,596, August 1978.

[109] J. Lundqvist, “Expandable thermoplastic microspheres and a method for theproduction and use thereof.” European Patent 0,486,080, May 1992.

[110] M. R. Soane, David S.; Houston, “Method for synthesizing thermo-expandablepolymeric microspheres,” United States Patent 6,617,364, September 2003.

[111] A. http://www.expancel.com, Expancel, “Expancel microspheres: An introduc-tion,” online, last accessed 2007.01.27.

[112] Y.-W. Kim, S.-H. Kim, H.-D. Kim, and C. B. Park, “Processing of closed-cell silicon oxycarbide foams from a preceramic polymer,” Journal of MaterialsScience, vol. 39, no. 18, pp. 5647–5652, 2004.

[113] M. Tomalino and G. Bianchini, “Heat-expandable microspheres for carprotection production,” Progress in Organic Coatings, vol. 32, no. 8, pp. 17–24,December 1997.

[114] P. Dowding and B. Vincent, “Suspension polymerisation to form polymerbeads,” Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol.161, pp. 259–269(11), 30 January 2000.

[115] M. Munzer and T. E., Polymerization processes. John Wiley & Sons, NewYork, US, 1977, ch. Chapter 5: Polymerizations in Suspension, pp. 106–142.


[116] H. Yuan, G. Kalfas, and W. Ray, “Suspension polymerization,” Journal ofMacromolecular Science - Reviews in Macromolecular Chemistry and Physics,vol. C31, no. 2-3, pp. 215–299, 1991.

[117] R. Oleschuk, L. Shultz-Lockyear, Y. Ning, and D. Harrison, “Trappingof bead-based reagents within microfluidic systems: On-chip solid-phaseextraction and electrochromatography,” Analytical Chemistry, vol. 72, no. 3,pp. 585–590, 2000.

[118] K. Sato, M. Tokeshi, T. Odake, H. Kimura, T. Ooi, M. Nakao, and T. Kitamori,“Integration of an immunosorbent assay system: Analysis of secretory humanimmunoglobulin a on polystyrene beads in a microchip,” Analytical Chemistry,vol. 72, no. 6, pp. 1144–1147, 2000.

[119] H. Andersson, W. van der Wijngaart, P. Enoksson, and G. Stemme,“Micromachined flow-through filter-chamber for chemical reactions on beads,”Sensors and Actuators B, vol. 67, p. 203208, 2000.

[120] H. Andersson, W. van der Wijngaart, and G. Stemme, “Micromachinedfilter-chamber array with passive valves for biochemical assays on beads,”Electrophoresis, vol. 22, no. 2, pp. 249–257, 2001.

[121] H. Andersson, P. Griss, and G. Stemme, “Expandable microspheres - surfaceimmobilization techniques,” Sensors and Actuators B, vol. 84, pp. 290–295,2002.

[122] A. Kumar, H. A. Biebuyck, and G. M. Whitesides, “Patterning self-assembledmonolayers: Applications in materials science,” Langmuir, vol. 10, no. 5, pp.1498–1511, 1994.

[123] H. Andersson, C. Jonsson, C. Moberg, and G. Stemme, “Self-assembled andself-sorted array of chemically active beads for analytical and biochemicalscreening,” Talanta, vol. 56, no. 2, pp. 301–308, Feb. 2002.

[124] H. Andersson, C. Jonsson, C. Moberg, and G. Stemme, “Patternedself-assembled beads in silicon channels,” Electrophoresis, vol. 22, no. 18, pp.3876–3882, 2001.

[125] N. Roxhed, S. Rydholm, B. Samel, W. van der Wijngaart, P. Griss, andG. Stemme, “A compact, low-cost microliter-range liquid dispenser based onexpandable microspheres,” Journal of Micromechanics and Microengineering,vol. 16, no. 12, pp. 2740–2746, 2006.

[126] DowCorning, “Sylgard 184 silicone elastomer kit product information.”

[127] J. Monahan, A. A. Gewirth, and R. G. Nuzzo, “A method for filling complexpolymeric microfluidic devices and arrays,” Anaytical Chemistry, vol. 73, no. 13,pp. 3193–3197, 2001.

REFERENCES 67

[128] K. Hosokawa, K. Sato, N. Ichikawa, and M. Maeda, “Power-freepoly(dimethylsiloxane) microfluidic devices for gold nanoparticle-based dnaanalysis,” Lab Chip, vol. 4, pp. 181–185, 2004.

[129] S. Gould and S. Subramani, “Firefly luciferase as a tool in molecular and cellbiology,” Anal. Biochem., vol. 175, p. 513, 1988.

[130] A. Russom, W. Wijngaart, H. Ehring, N.Tooke, H. Andersson, and G. Stemme,“Detection of single nucleotide incorporation using pyrosequencing in a microflu-idic device,” in Micro Total Analysis Systems, Nara, Japan, 2002, pp. 308–310.

[131] H. Matsuki, A. Suzuki, H. Kamaya, and I. Ueda, “Specific and non-specific bind-ing of long-chain fatty acids to firefly luciferase: cutoff at octanoate,” Biochim.Biophys. Acta, vol. 1426, pp. 143–150, 1999.

[132] D. Ross, M. Gaitan, and L. E. Locascio, “Temperature measurement in mi-crofluidic systems using a temperature-dependent fluorescent dye,” AnalyticalChemistry, vol. 73, pp. 4117–4123, 2001.

[133] M. Prausnitz, “Microneedles for transdermal drug delivery,” Adv. Drug DeliveryReviews, vol. 56, pp. 581–587, 2004.

[134] M. Gerstel and V. Place, “Drug delivery device,” United States Patent3,964,482, June 1976.

[135] B. Barry, “Novel mechanisms and devices to enable successful transdermal drugdelivery,” Eur. J. of Pharmaceutical Sci, vol. 14, pp. 101–114, 2001.

[136] P. Griss and G. Stemme, “Side-opened out-of-plane microneedles for microflu-idic transdermal liquid transfer,” IEEE Journal of Microelectromechanical Sys-tems, vol. 12, no. 3, pp. 296–301, Jun 2003.

[137] M. Prausnitz, S. Mitragotri, and R. Langer, “Current status and future potentialof transdermal drug delivery,” Nature Reviews Drug Discovery, vol. 3, pp. 115–124, 2004.

[138] S. Kaushik, A. Hord, D. Denson, D. McAllister, S. Smitra, M. Allen, andM. Prausnitz, “Lack of pain associated with microfabricated microneedles,”Anesth Analg., vol. 92, no. 2, pp. 502–504, 2001.

[139] R. Sivamani, B. Stoeber, G. Wu, H. Zhai, D. Liepmann, and H. Maibach, “Clin-ical microneedle injection of methyl nicotinate: stratum corneum penetration,”Skin Res. Technol., no. 11, pp. 152–156, May 2005.

[140] W. Lin, M. Cormier, A. Samiee, A. Griffin, B. Johnson, C. Teng, G. Hardee, andP. Daddona, “Transdermal delivery of antisense oligonucleotides with micropro-jection patch (macroflux) technology,” Pharmaceutical Res., vol. 18, no. 12, pp.1789–1793, 2001.


[141] J. A. Matriano, M. Cormier, J. Johnson, W. Young, M. Buttery, K. Nyam,and P. Daddona, “Macroflux microprojection array patch technology: A newand efficient approach for intracutaneous immunization,” Pharmaceutical Res.,vol. 19, no. 1, pp. 63–70, 2002.

[142] J. A. Mikszta, J. B. Alarcon, J. M. Brittingham, D. E. Sutter, R. J. Pettis, andN. G. Harvey, “Improved genetic immunization via micromechanical disruptionof skin-barrier function and targeted epidermal delivery,” Nat Med, vol. 8,no. 4, pp. 415–419, Apr 2002.

[143] D. V. McAllister, P. M. Wang, S. P. Davis, J.-H. Park, P. J. Canatella,M. G. Allen, and M. R. Prausnitz, “From the Cover: Microfabricated needlesfor transdermal delivery of macromolecules and nanoparticles: Fabricationmethods and transport studies,” PNAS, vol. 100, no. 24, pp. 13 755–13760,2003.

[144] H. Gardeniers, R. Luttge, E. Berenschot, M. de Boer, S. Yeshurun, M. Hefetz,R. van’t Oever, and A. van den Berg, “Silicon micromachined hollow micronee-dles for transdermal liquid transport,” IEEE Journal of MicroelectromechanicalSystems, vol. 12, no. 6, pp. 855–862, Dec 2003.

[145] W. Martanto, S. Davis, H. Holiday, J. Wang, H. Gill, and M. Prausnitz, “Trans-dermal delivery of insulin using microneedles in vivo.” Pharm Res., vol. 21, no. 6,pp. 947–952, 2004.

[146] S. Davis, W. Martanto, M. Allen, and M. Prausnitz, “Hollow metal microneedlesfor insulin delivery to diabetic rats,” IEEE Trans. Biomed. Eng., vol. 52, no. 5,pp. 909–15, 2005.

[147] P. Griss, H. Tolvanen-Laakso, P. Merilainen, and G. Stemme, “Characteriza-tion of micromachined spiked biopotential electrodes,” IEEE Transactions onBiomedical Engineering, vol. 49, no. 6, pp. 597–604, Jun 2002.

[148] P. Aberg, P. Geladi, I. Nicander, J. Hansson, U. Holmgren, and S. Ollmar,“Non-invasive and microinvasive electrical impedance spectra of skin cancer -a comparison between two techniques,” Skin Res. and Technol., vol. 11, no. 4,pp. 281–286, 2005.

[149] M. M. Levine, “Can needle-free administration of vaccines become the norm inglobal immunization?” Nat Med, vol. 9, no. 1, pp. 99–103, Jan 2003.

[150] S. A. Plotkin, “Vaccines: past, present and future,” Nat Med, Apr 2005.

[151] N. Roxhed, P. Griss, and G. Stemme, “Reliable in-vivo penetration and trans-dermal injection using ultra-sharp hollow microneedles,” in Proceedings IEEEInternational Conference on Solid-State Sensors, Actuators, and Microsystems(Transducers), vol. 1, 2005, pp. 213–216.

Documents

Novel Microﬂuidic Devices Based on a Thermally Responsive ...12419/FULLTEXT01.pdf · x Novel Microﬂuidic Devices Based on a Thermally Responsive PDMS Composite The contribution