POLYMER MICROELECTROMECHANICAL SYSTEMS: FABRICATION AND APPLICATIONS IN BIOLOGY
AND BIOLOGICAL FORCE MEASUREMENTS
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Nicholas J. Ferrell, B.S.
*****
The Ohio State University 2008
Dissertation Committee: Approved by Professor Derek J. Hansford, Advisor
Professor L. James Lee ___________________________________
Professor Allen Yi Advisor Biomedical Engineering Graduate Program
ABSTRACT
Polymer materials are increasingly being utilized in biomedical micro- and
nanotechnolgy applications. This trend has been driven by a several factors ranging from
materials compatibility to cost. The manufacturing techniques used to produce these
devices are considerably less mature than their silicon-based counterparts. New
manufacturing techniques are needed to address unique processing challenges posed by
polymer materials. To this end, we have developed a set of soft lithography based
micromolding techniques for fabrication of polymer microstructures and devices from a
wide range of materials. Materials include common thermoplastic polymers such as
poly(methyl methacrylate) (PMMA) and polystyrene as well as functional materials such
as conducting polymers. The processing techniques developed through this work are
capable of producing a wide range of structures including continuous microstructured
films, isolated polymer microstructures, and suspended structures. The nature of the
materials and the non-cleanroom based micromolding processes makes these techniques
considerably more cost effective with respect to both materials and processing costs.
In addition to developing processing techniques, characterization of the processes
as well as the materials is a critical step for implementation of polymers in practical
device applications. Process characterization was performed by systematically varying
process parameters and evaluating the resulting microstructures using common micro-
ii
and nanoscale characterization techniques. Scanning electron microscopy, atomic force
microscopy, and optical microscopy were all used to evaluate the resulting polymer
structures. Nanoindentation techniques were used to characterize the mechanical
properties of the materials. Elastic modulus, hardness, creep, scratch resistance, and yield
strength of several polymer MEMS materials were evaluated.
Application of these techniques for development of functional devices is
ultimately the goal. We have used the processing techniques that we have developed to
fabricate and test three polymer MEMS devices for biological applications. The first is a
microfabricated membrane system for isolation of individual cells or cell clusters. This
device could be utilized in a variety of cell biology applications including single cell
experimentation, cell cluster biology, and tissue engineering. The other two devices were
developed for measuring low magnitude biological forces. A polymer cantilever force
sensor was developed for measuring contractile forces produced by fibroblast cells. This
device could be used in cell mechanobiology studies, drug evaluation, and cell-based
biosensing. The final device is an adapted polymer cantilever sensor for measuring forces
produced by protein aggregates known a forisomes. This unique biomaterial could be
utilized as a valve or actuator in microdevices.
iii
Dedicated to my parents
iv
ACKNOWLEDGMENTS
I would like to thank my advisor, Derek Hansford, for his support and
encouragement dating all the way back to my undergraduate days. I thank Dr. Hansford
for his advice and mentorship. Working under his guidance over the years has been a
pleasure.
I thank all the members of our research group, past and present. Their support,
training, mentoring, and patients were indispensable. Specifically, I would like to
acknowledge Jay Woodard for his painstaking work with the design and simulation of the
cell force sensor. Special thanks to Jingjiao Guan, Joe Kitzmiller, Rob Short, Kelly
Larkin, and Jason Sakamoto for showing me the ropes and getting me started on my way
down this path. I am grateful to have had the opportunity to work closely with Yang Sun
and Randy Butler. I especially thank Daniel Gallego and Natalia Higuita for their
enthusiasm and constant willingness to help me out in the lab.
I am grateful to Profs. L. James Lee and Allen Yi for serving on my committee. I
would like to thank the technical staff and the Nanotech West Laboratory, particularly
Derek Ditmer and Paul Steffan, for technical support and training.
This research was supported in part by the Air Force Office of Scientific Research
MURI award (F49620-03-1-0421). Portions of this work were funded by the National
Science Foundation (NSF) Center for Affordable Nanoengineering of Polymer
v
Biomedical Devices (EEC-0425626). I also acknowledge the NSF Integrative Graduate
Research and Education Traineeship (IGERT) program (0221678) for my graduate
fellowship.
vi
VITA
September 27, 1979.……………………..Born - Van Wert, OH
2003….…………………………………..B.S. Mechanical Engineering, The Ohio State University
2003-present...…….…………….……….Graduate Research Fellow, The Ohio State University
2007….…………………………………..Visiting Research Scientist, Instituto de Enginharia Biomédica (INEB), Universidade do Porto
PUBLICATIONS
Research Publications
1. Nicholas Ferrell, James Woodard, Derek Hansford, “Fabrication of Polymer Microstructures for MEMS: Sacrificial Layer Micromolding and Patterned Substrate Micromolding.” Biomedical Microdevices, 9, 815-821 (2007).
2. Jingjiao Guan, Nicholas Ferrell, Bo Yu, Derek J. Hansford, L. James Lee, “Simultaneous Fabrication of Hybrid Arrays of Nanowires and Micro/Nanoparticles by Dewetting on Micropillars.” Soft Matter, 3, 1369-1371 (2007).
3. Manuel Palacio, Bharat Bhushan, Nicholas Ferrell, Derek Hansford, “Adhesion Properties of Polymer/Silicon Interfaced for Biological Micro-/nanoelectromechanical Systems Applications.” Journal of Vacuum Science and Technology A, 25, 1275-1284 (2007).
4. Nicholas Ferrell, Derek Hansford, “Fabrication of Polymer Micro and Nanostructures by Soft Lithography and Spin Dewetting.” Macromolecular Rapid Communications, 28, 966-971 (2007).
5. Manuel Palacio, Bharat Bhushan, Nicholas Ferrell, Derek Hansford, “Nanomechanical Characterization of Polymer Beam Structures for BioMEMS Applications.” Sensors and Actuators A: Physical, 135, 637-650 (2007).
vii
6. Randall T. Butler, Nicholas J. Ferrell, Derek J. Hansford, “Spatial and Geometrical Control of Biosilification Using a Patterned Poly-L-Lysine Template.” Applied Surface Science, 252, 7337-7342 (2006).
7. Jingjiao Guan, Nicholas Ferrell, L. James Lee, Derek J. Hansford, “Fabrication of Polymeric Microparticles for Drug Delivery by Soft Lithography.” Biomaterials, 27, 4034-4041 (2006).
8. Guohua Wei, Bharat Bhushan, Nicholas Ferrell, Derek Hansford, “Fabrication and Nanomechanical Characterization of Polymer MEMS for Biological Applications.” Journal of Vacuum Science and Technology A, 23, 811-819 (2005).
9. Jennifer Lewis, Mark Kotur, Omar Butt, Sumant Kulkarni, Alyssa Riley, Nick Ferrell, Kathryn Sullivan, Mauro Ferrari, “Biotechnology Apprenticeship for Secondary-level Students: Teaching Advanced Cell Culture Techniques for Research.” Cell Biology Education, 1, 26-42 (2002).
FIELD OF STUDY
Major Field: Biomedical Engineering
viii
TABLE OF CONTENTS
Page
Abstract…………………………………………………………………………………… ii Dedication………………………………………………………………………………... iv Acknowledgments…………………………………………………………………………v Vita……………………………………………………………………………………….vii List of Tables ………………………………………………………………………...….xiii List of Figures ...…………………………………………………………………………xiv Chapters Page
1. Microfabrication and Microelectromechanical Systems (MEMS) ................................. 1
1.1 Introduction............................................................................................................... 1 1.2 Microfabrication and MEMS.................................................................................... 2
1.2.1 Photolithography................................................................................................ 2 1.2.2 Silicon MEMS fabrication ................................................................................. 6 1.2.3 Applications of silicon MEMS ........................................................................ 10 1.2.4 Silicon-based BioMEMS ................................................................................. 10 1.2.5 Microfluidics.................................................................................................... 14 1.2.6 Polymer microfabrication ................................................................................ 17 1.2.7 Applications of polymer MEMS...................................................................... 21
1.3 Conclusion .............................................................................................................. 24
2. Cell Mechanics.............................................................................................................. 31
2.1 Introduction............................................................................................................. 31 2.2 Cell mechanics and cell forces................................................................................ 32
2.2.1 The cytoskeleton .............................................................................................. 32 2.2.2 Forces generated by adherent cells .................................................................. 34 2.2.3 The cytoskeleton and disease........................................................................... 37 2.2.4 Drugs and cell mechanics ................................................................................ 38
2.3 Measuring cell mechanics....................................................................................... 39 2.3.1 Appling forces to cells ..................................................................................... 39 2.3.2 Measuring the mechanical properties of cells.................................................. 40 2.3.3 Measuring forces generated by cells................................................................ 41 2.3.4 BioMEMS for measuring cell forces ............................................................... 43
ix
2.4 Conclusion .............................................................................................................. 46
3. Fabrication of Polymer Microstructures by Double Stamp Micromolding .................. 51
3.1 Introduction............................................................................................................. 51 3.2 Materials and Methods............................................................................................ 53
3.2.1 Double stamp micromolding process............................................................... 53 3.2.2 Process characterization................................................................................... 54
3.3 Results and discussion ............................................................................................ 55 3.4 Conclusion .............................................................................................................. 60
4. Spin Dewetting of Polymers on Polydimethylsiloxane (PDMS) Molds ..................... 62
4.1 Introduction............................................................................................................. 62 4.2 Materials and methods ............................................................................................ 65
4.2.1 Materials .......................................................................................................... 65 4.2.2 PDMS molding ................................................................................................ 65 4.2.3 Spin dewetting and pattern transfer ................................................................. 65 4.2.4 Process characterization................................................................................... 68
4.3 Results and discussion ............................................................................................ 68 4.4 Conclusion .............................................................................................................. 75
5. Lift-Off Processing for Fabricating Micropatterned Sulfonated Polyaniline ............... 78
5.1 Introduction............................................................................................................. 78 5.2 Materials and methods ............................................................................................ 80
5.2.1 SPAN synthesis................................................................................................ 80 5.2.2 Micromolding and Lift-off processing............................................................. 80 5.2.3 Process characterization................................................................................... 82
5.3 Results and discussion ............................................................................................ 83 5.4 Conclusion .............................................................................................................. 87
6. Microfabricated Membranes for Cell Isolation............................................................. 90
6.1 Introduction............................................................................................................. 90 6.2 Materials and methods ............................................................................................ 93
6.2.1 Materials .......................................................................................................... 93 6.2.2 Membrane fabrication...................................................................................... 93 6.2.3 Cell culture and filtration ................................................................................. 95 6.2.4 Characterization ............................................................................................... 97
6.3 Results and discussion ............................................................................................ 97 6.4 Conclusion ............................................................................................................ 104
7. Fabrication of Suspended Polymer Microstructures by Patterned Susbstrate Micromolding and Sacrificial Layer Micromolding....................................................... 107
7.1 Introduction........................................................................................................... 107 x
7.2 Materials and methods .......................................................................................... 109 7.2.1 PDMS mold fabrication ................................................................................. 109 7.2.2 Sacrificial layer micromolding ...................................................................... 110 7.2.3 Patterned substrate micromolding.................................................................. 113
7.3 Results and discussion .......................................................................................... 115 7.4 Conclusion ............................................................................................................ 120
8. Measuring the Mechanical Properties of Polymer Microstructures by Nanoindentation......................................................................................................................................... 123
8.1 Introduction........................................................................................................... 123 8.2 Materials and methods .......................................................................................... 128
8.2.1 Fabrication of polymer microstructures and thin films.................................. 128 8.2.2 Mechanical characterization .......................................................................... 129
8.3 Results and discussion .......................................................................................... 134 8.3.1 Hardness and elastic modulus by CSM nanoindentaion................................ 134 8.3.2 Creep behavior by CSM nanoindenation....................................................... 135 8.3.3 Scratch resistance from nanindentation scratch test ...................................... 136 8.3.3 Elastic modulus from normal beam bending ................................................. 138 8.3.4 Effects of aqueous environment and temperature.......................................... 141 8.3.5 Yield and breaking strength from normal beam bending .............................. 143 8.3.6 Lateral beam bending..................................................................................... 145
8.4 Conclusion ............................................................................................................ 146
9. Desing, Simulation, and Fabrication of a Polymer Sensor for Measuring Single Cell Forces.............................................................................................................................. 149
9.1 Introduction........................................................................................................... 149 9.2 Materials and methods .......................................................................................... 151
9.2.1 Solid modeling for device visualization ........................................................ 151 9.2.2 Finite element analysis................................................................................... 151 9.2.3 Device development and prototyping process ............................................... 151 9.2.4 Device fabrication.......................................................................................... 152 9.2.5 Device characterization.................................................................................. 154
9.3 Results and discussion .......................................................................................... 154 9.3.1 Design considerations .................................................................................... 154 9.3.2 Device overview ............................................................................................ 155 9.3.3 First generation cell force sensor ................................................................... 157 9.3.4 Second generation cell force sensor............................................................... 162 9.3.5 Third generation cell force sensor.................................................................. 170 9.3.6 Determining the force vectors from displacement data ................................. 177
9.4 Conclusion ............................................................................................................ 181
Biological Testing of the Polymer MEMS Cell Force Sensor........................................ 183
10.1 Introduction......................................................................................................... 183
xi
10.2 Materials and method.......................................................................................... 185 10.2.1 Sample preparation and surface modification.............................................. 185 10.2.2 Experimental setup....................................................................................... 186 10.2.3 Cell culture techniques................................................................................. 188 10.2.4 Cell placement ............................................................................................. 188 10.2.5 Image acquisition and analysis .................................................................... 189
10.3 Results and discussion ........................................................................................ 190 10.3.1 Sample preparation ...................................................................................... 190 10.3.2 Experimental results with first and second generation devices ................... 192 10.3.3 Results with third generation sensors .......................................................... 195 10.3.4 Effects of chemical exposure on cell forces ................................................ 205 10.3.5 Sources of error and measures to minimize error ........................................ 207
10.4 Conclusion .......................................................................................................... 209
11. Measurement of Mechanical Forces Generated by Plant P-Protein Aggregates (Forisomes) ..................................................................................................................... 212
11.1 Introduction......................................................................................................... 212 11.2 Materials and methods ........................................................................................ 213
11.2.1 Design .......................................................................................................... 213 11.2.2 Finite element simulations ........................................................................... 214 11.2.3 Fabrication and Characterization ................................................................. 214 11.2.4 Forisome preparation ................................................................................... 215 11.2.5 Experimental setup....................................................................................... 216
10.3 Results and discussion ........................................................................................ 216 11.4 Conclusion .......................................................................................................... 222
Conclusion and Future Outlook ...................................................................................... 224
Appendices
A. Design and Simulation Data for the Cell Force Sensor ............................................ 228
B. Polymer Microfabrication as a Tool for Processing Inorganic Materials .................. 239
List of References ........................................................................................................... 258
xii
LIST OF TABLES
Table Page
1.1 Typical etch processes for silicon based materials………...……………………...9
8.1 Mechanical properties of PPMA, PMMA, PS, and PS/Clay ……………...……140
10.1 Cell force calculations for second generation cell force sensor. ………………..194
10.2 Cell force calculations for second generation cell force sensor. ………………..195
A.1 Equations for calculating force angle……………………………………...….. 234
A.2 Equations for calculating deflection per unit force………………...…….……..237
A.3 Spreadsheet for calculating force vectors from displacement data……………..238
xiii
LIST OF FIGURES
Figures Page
Figure 1.1 Schematic diagram of the photolithography process..........................................2 Figure 1.2 SEM micrographs of photoresist patterns ………………………….….……....6 Figure 1.3 Etch profiles for silicon etch processes….………………………….….……....9 Figure 1.4 Silicon MEMS drug delivery device ……………………………….…...…....12 Figure 1.5 Immunoisolation biocapsule …………….………………………….…...…....12 Figure 1.6 Microcantilever sensor …………….……………………………….…...…....14 Figure 1.7 Microscale hot embossing system …………..….………………….…...…....16 Figure 1.8 Schematic diagram of the PDMS molding process ……………….…...…....18 Figure 1.9 Soft lithography processes ………………………………………….…...…....20 Figure 1.10 Polymer/metal tactile sensor ………...…………………………….…...…....22 Figure 1.11 Microfabricated polymer devices for controlled drug delivery …...…...…....23 Figure 1.12 Microfabricated tissue engineering scaffolds .…………………….…...…....24 Figure 2.1 Cytoskeletal proteins ……………………………………………….…...…....33 Figure 2.2 Molecular components of focal adhesions ...……………………….…...…....35 Figure 2.3 Cell motility process …………….....……………………………….…...…....36 Figure 2.4 Traction force microscopy ………………………………………….…...…....42 Figure 2.5 Silicon MEMS heart cell force sensor ..…………………………….…...…....44 Figure 2.6 Silicon cantilever fibroblast cell force sensor ..…………………….…...…....45
xiv
Figure 2.7 PDMS pillar cell force sensor ………...…………………………….…...…....45 Figure 3.1 Schematic diagram of the microtransfer molding process... ……….…...…....52 Figure 3.2 Schematic diagram of the double stamp micromolding process …...…...…....54 Figure 3.3 Optical micrographs of polystyrene spin coated on PDMS. ...………….…....56 Figure 3.4 SEM micrographs of four steps in the double stamp micromolding process ...57 Figure 3.5 Characterization of feature distortion versus polymer concentration …...…....59 Figure 3.6 Characterization of feature distortion versus transfer pressure ………....…....59 Figure 4.1 Schematic diagram of spin dewetting and pattern transfer. .……….…...…....67 Figure 4.2 AFM images of polystyrene structures fabricated by spin dewetting on a 2 µm diameter pillar mold …………………………...……………………………….…...…....69 Figure 4.3 Optical micrographs of PPMA dewetting on 20 µm diameter pillars…...…....70 Figure 4.4 Particles size versus (a) solution concentration and (b) mold feature size .......72 Figure 4.5 SEM micrographs at each step in the spin dewetting and pattern transfer process……………………………………………………………………………………73 Figure 4.6 Dewetting of polystyrene is PDMS wells ………………………….…...…....74 Figure 4.7 Nanoscale polystyrene features fabricated by spin dewetting ……...…...…....74 Figure 5.1 Schematic diagram of the lift-off process for SPAN patterning.. ….…...…....81 Figure 5.2 SEM micrographs of sacrificial layers, SPAN deposition of sacrificial layers, and SPAN patterns after lift-off ………………………………………………..…...…....83 Figure 5.3 Film thickness versus time for static and dynamic SPAN deposition ......…....85 Figure 5.4 Roughness versus time for static and dynamic SPAN deposition…........…....85 Figure 5.5 SEM micrographs of 20 µm circular SPAN patterns ……………….......…....86 Figure 6.1 Schematic diagram of the cell isolation membrane fabrication process. .…....95 Figure 6.2 Experimental setup for cell isolation …………………………………………96
xv
Figure 6.3 SEM micrographs of cell isolation membranes ……………………........…....98 Figure 6.4 SEM micrographs of NIH 3T3 cells isolated in circular, hexagonal, and square membrane features …………………………………………………………….......…....99 Figure 6.5 SEM micrographs of NIH 3T3 cells after removing the membrane ......…....100 Figure 6.6 Fluorescent micrographs of NIH 3T3 cells at low seeding density.. ......…....101 Figure 6.7 Fluorescent micrographs of NIH 3T3 cells at high seeding density ..………101 Figure 6.8 (a) SEM micrograph and (b) fluorescent micrograph of non-adherent THP-1 cells in circular membrane features …..……………………………………….......…....102 Figure 6.9 (a,b) Fluorescent micrographs and (c) SEM micrograph of C3A liver cell clusters in circular membrane features …………………………………………………103 Figure 7.1 Schematic diagram of the sacrificial layer micromolding process.. ………...111 Figure 7.2 Schematic diagram of the patterned substrate micromolding process ……...114 Figure 7.3 SEM microstructures fabricated by sacrificial layer micromolding ………...115 Figure 7.4 Polymer cantilevers (a) before and (b) after release of the sacrificial layer. ..116
Figure 7.5 Polymer cantilever deflected with a micropipette……...…………………...117 Figure 7.6 Polymer beams fabricated by patterned substrate micromolding …………...118 Figure 7.7 AFM images of features with and without proper selection of processing parameters ………………………………………………….……...…………………....120 Figure 8.1 Schematic diagram of the continuous stiffness measurement loading cycle..125 Figure 8.2 Schematic diagram of the process for fabricating cantilevers for lateral bending ………………………………………………………………………………….129 Figure 8.3 SEM micrographs of polymer beams for bending and nanoindentation testing…………………………………………………………………………………...130 Figure 8.4 Nanoindentation experimental setup …………………...…………………...133 Figure 8.5 Hardness and elastic modulus measured using CSM nanoindentation ...…...135 Figure 8.6 Creep measurements for PPMA, PMMA, PS, and PS/Clay………………...136
xvi
Figure 8.7 Scratch resistance and corresponding SEM micrographs showing film damage………………………………………………………………………………….137 Figure 8.8 Load-displacement plot for beam bending with an uncoated tip….………...138 Figure 8.9 Load-displacement plots for beam bending with PPMA, PMMA, PS, and PS/Clay beams ………………………………………………...…...…………………...139 Figure 8.10 Effects of aqueous environment and increased temperature on polymer properties……………………………………………...…….……...…………………...142 Figure 8.11 Load-displacement plots for beam bending at high loads .………………...144 Figure 8.12 Lateral bending tests with PS and PS/Clay …………...…………………...145 Figure 9.1 Schematic diagram of the fabrication process for the cell force sensor.. …...153 Figure 9.2 Circular configuration of the first generation cell force sensor …...………...156 Figure 9.3 Linear configuration of the first generation cell force sensor ….…………...157 Figure 9.4 L-beam cantilever and direction-angle convention ..…...…………………...158 Figure 9.5 Deflection plot for the first generation sensor ……….....…………………...160 Figure 9.6 SEM micrographs of the first generation cell force sensor .………………...161 Figure 9.7 Optical micrographs of cantilever bending ……..……...…………………...162 Figure 9.8 L-beam line of highest sensitivity ……………………...…………………...163 Figure 9.9 Hinge in the second generation beam design …...……...…………………...164 Figure 9.10 CAD images of second generation cell force sensor .....…………………...165 Figure 9.11 FEA simulation of second generation cantilever beam .…………………...166 Figure 9.12 Deflection plot for the second generation sensor …......…………………...167 Figure 9.13 Comparison of the deflection plots for the first and second generation sensor…………….………………………… …………………………………………..168 Figure 9.14 SEM micrographs of the second generation sensors made from PPMA and polystyrene ………………………………………………….……...…………………...169
xvii
Figure 9.15 Phase contrast micrographs of the sensor after sacrificial layer release ......170 Figure 9.16 Area of HT 1080 and 3T3 cells before and after spreading ..……………...171 Figure 9.17 Angle convention for third generation device ………...…………………...172 Figure 9.18 FEA loading conditions ………………………..……...…………………...173 Figure 9.19 Deflection plots for different FEA loading conditions ..…………………...174 Figure 9.20 Deflection plots for different measurement locations ...…………………...175 Figure 9.21 Comparison of the deflection plots for the first, second, and third generation cell force sensors …………………………………………………...…………………...176 Figure 9.22 SEM micrographs of the third generation sensor …......…………………...177 Figure 9.23 Plot of force angle versus beam deflection …….……...…………………...178 Figure 9.24 Plot of deflection per unit force versus force angle …...…………………...179 Figure 10.1 Experimental setup for force measurements ……..…...…………………...187 Figure 10.2 Contact angle for native, tissue culture treated, and O2 plasma modified polystyrene ………………………………………………….……...…………………...191 Figure 10.3 Optical micrographs of a fibroblast on the second generation sensor …......194 Figure 10.4 Optical micrographs of a fibroblast on the third generation sensor ……….197 Figure 10.5 Force magnitude versus time plot for a WS-1 fibroblast cell.. …………….199 Figure 10.6 Force direction versus time plot for a WS-1 fibroblast cell ……………….199 Figure 10.7 Force magnitude versus time plot for a WS-1 fibroblast cell.. …………….201 Figure 10.8 Force direction versus time plot for a WS-1 fibroblast cell ……………….201 Figure 10.9 Force and direction versus time plots for a WS-1 fibroblast cell ………….203 Figure 10.10 Force and direction versus time plots for a WS-1 fibroblast cell ..……….204 Figure 10.11 Optical micrographs of a fibroblast cell on the sensor before and after exposure to cytochalasin-D ………………………………………………….………….206
xviii
Figure 10.12 Force versus time plot for a WS-1 fibroblast cell exposed to cytochalasin-D ………………………………………………………………….………207 Figure 11.1 Forisome force sensor …………………..………………………………….214 Figure 11.2 SEM micrographs of the forisome force sensor ..………………………….217 Figure 11.3 Comparison of FEA and analytical solutions for forisome sensor force-deflection response …………………………………………………………..………….219 Figure 11.4 Optical micrographs of forisome before and after actuation …...………….220 Figure 11.5 SEM micrograph of a forisome on the sensor after testing …….………….221 Figure 11.6 Radial versus longitudinal forisome forces …………………….………….222 Figure A.1 Deflection plot for third generation cell force sensor ………………………229 Figure A.2 Curve fits for calculating force direction …………………………………...229 Figure A.3 Curve fits for calculating force per unit deflection …………………………234 Figure B.1 Schematic diagram of the silicon wet etch process ………………………...241 Figure B.2 SEM micrographs of silicon after TMAH etching …………………………242 Figure B.3 SEM micrographs of NaCl crystal on PDMS pillars ……………………….244 Figure B.4 SEM micrographs of sucrose particles on PDMS pillars …………………..245 Figure B.5 Plot of crystal/particle size versus NaCl and sucrose concentration ……….246 Figure B.6 SEM micrographs of ammonium heptamolybdate hydrate particles PDMS pillars……………………………………………………………………………………246 Figure B.7 SEM micrograph of poly-l-lysine particles on PDMS pillars………………247 Figure B.8 Schematic diagram of the peptide mediated deposition process.. ………….249 Figure B.9 SEM micrographs and EDS spectra of peptide mediated deposition of silica…………………………………………………………… ……………………….250 Figure B.10 SEM micrographs of peptide mediated gold deposition …………………..250 Figure B.11 Schematic diagram of the sol-gel micromolding process …..……………..252
xix
Figure B.12 SEM micrographs of silica micropatterns fabricated by sol-gel micromolding…………………………………………………………………………...252 Figure B.13 SEM micrographs and EDS spectra for sol-gel micromolded silica/HA nanoparticle composites ………………………………………………………………...253 Figure B.14 SEM micrographs and EDS spectra for selective sol-gel micromolding of silica/HA nanoparticle composites ……………………………………………………..255 Figure B.15 Optical micrographs of bone marrow cells grown on patterned and flat silica…………………………………………………………………………………….256 Figure B.16 SEM micrographs of bone marrow cells grown on micropatterned silica and silica/HA nanoparticles composites …………………………………………………….256
xx
CHAPTER 1
MICROFABRICATION AND MICROELECTROMECHANICAL SYSTEMS (MEMS)
1.1 Introduction
Micro and nanoscale devices hold great potential in a wide range of technological
fields. Micro and nanotechnology concepts have infiltrated almost every scientific
discipline including chemistry, physics, engineering, materials science, and medicine.
Research efforts in these areas are continuing to add to a growing body of knowledge and
push toward applying these concepts in ways that could have a significant impact on
society. Micro and nanoscale materials and devices have already found commercially
applications ranging from the automotive to the clothing industries. As these research and
commercialization efforts continue, no doubt additional progress will be made toward
reaching the potential that has been predicted for this field.
Traditionally micro and nanoscale devices, often referred to as micro or
nanoelectromechanical systems (MEMS and NEMS), have been fabricated from silicon-
based materials using processing techniques adapted from the semiconductor industry.
Polymer materials are increasingly being utilized in micro and nanoscale biomedical
devices. This trend is driven by a number of factors. Many polymer materials are known
to be relatively biocompatible, which allows them to be used either as implant materials
or for in vitro applications with a minimal detrimental effect on the host tissue or cells.
1
Polymers also have a wide range of properties with respect to bulk chemistry, surface
chemistry, and mechanical and physical properties. This broad range of properties
provides a large materials selection pool and allows materials to be chosen based on the
appropriate behavior in a specific application. In the case of micro and nanodevices,
processing costs can become a prohibitive factor due to the specialized materials,
equipment, and facilities needed to manufacture such devices. Polymer processing can
offer some distinct advantages in minimizing both materials and processing costs. Many
polymer materials have significantly lower cost relative to traditional materials used in
micro and nanodevices. Much of the processing can also be performed outside of a
cleanroom environment without expensive equipment, significantly reducing processing
costs.
1.2 Microfabrication and MEMS
1.2.1 Photolithography
Microfabrication is a term used to collectively describe a set of techniques used to
fabricate miniaturized devices for use in a variety of applications. One of the most
widely used techniques in microfabrication is the photolithography process.
Photolithography is a process of patterning photosensitive polymer (photoresist)
structures. The pattern is created by selectively exposing the photoresist to ultra violet
(UV) light. The UV exposure induces a chemical reaction in the photoresist that renders
the materials either soluble (positive resist) or insoluble (negative resist) in a chemical
developer. The photolithography process schematic for a negative photoresist is shown in
Figure 1.1.
2
Figure 1.1 Schematic diagram of the photolithography process for a negative photoresist.
The photolithography process can be broken down into six basic steps: substrate
preparation, pre-bake, exposure, post-exposure processing, developing, and
cleaning/drying. The first step is preparation of the substrate. Generally single crystal
silicon wafers are used, but glass, GaAs, and other materials are also used depending on
the application. The wafers are usually cleaned using piranha or the RCA cleaning
process.1 Piranha is a mixture of sulfuric acid (H2SO4) and hydrogen peroxide (H2O2) at
approximately a 3:1 ratio, respectively. The RCA cleaning procedure is an aqueous
3
mixture of hydrogen peroxide and ammonia (NH4OH) or hydrochloric acid (HCl). For
certain photoresists it is necessary to apply an adhesion promoter to improve the adhesion
between the wafer surface and the resist. Often this is done by vapor phase deposition of
a silane such as hexamethyldisilazane (HMDS).
The next step in the process is deposition of the photoresist. This is done by spin
coating the resist at speeds ranging from 1000-6000 rpm. The spin coating process can be
more complicated than one might expect. In its simplest form, the spin coating process is
a balance of the rotational inertial forces and the viscous shear of the fluid. The
equilibrium between these two factors leads to a uniform film thickness after the coating
process. Of course, several other factors contribute to the process. The use of volatile
solvents in the photoresists leads to a time varying viscosity of the fluid. Exhaust flow,
edge effects, and equipment variations also need to be considered. A number of
mathematical models of varying complexity have been developed to describe the spin
coating process.2,3 Process datasheets from the photoresist manufacturer usually contain
spin speed versus film thickness curves. These give a good estimate of spin coating
properties. However, given the differences in equipment and processing condition, it is
usually necessary to create spin speed versus film thickness curves empirically for each
photoresist. After spin coating, the photoresist is based to remove any residual solvent.
After coating and baking, the wafer is selectively exposed to UV through a
photomask. The mask is generally a high quality glass plate that is selectively coated with
chrome or another absorbing metal layer. For larger features, a high quality printed mask
(transparency mask) can be substituted for the chrome/glass mask.4 The mask is first
loaded into the exposure system (mask aligner). The wafer is placed below the mask and
4
the two pieces are either brought close to each other (proximity mode) or brought
together (contact mode). Other contact settings can be adjusted to improve the interface
between the mask and the wafer. Hard contact mode applies a stronger force to hold the
mask and wafer together and vacuum contact mode uses a vacuum to pull the wafer and
mask into tighter contact. Improving the contact between the mask and the wafer
improves the resolution of the pattern, but harder contact exposure increases wear of the
mask. Photoresist can also be transferred from the wafer to the mask, resulting in possible
defects in the pattern as well as requiring the mask to be cleaned more often. Post
exposure steps are often needed to complete the process. For example, negative tone SU8
photoresists (Microchem Corp.) require a post exposure baking process to complete the
cross linking reaction that is initiated by the UV exposure.
After exposure and post-exposure processing, the wafer is ready for the
development step. This process involves either immersion or spraying with a chemical
developer to remove the unwanted portions of the photoresist. For positive resists, the
exposed regions are removed, and for negative resist the unexposed regions are removed.
The developer chemistry depends on the photoresist. Negative tone resists are generally
developed in organic solvents, while positive resists are generally soluble in basic
solutions (e.g. NaOH). After development, the wafers are cleaned in deionized (DI)
water and dried.
Scanning electron micrographs of patterns created by photolithography are shown
in Figure 1.2. The images show features fabricated from both negative and positive
photoresist. Figure 1.2 (a) is SU8 negative photoresist and (b) is S1813 positive resist
positive resist (Shipley).
5
Figure 1.2 SEM micrographs of (a) 5 µm diameter SU8 pillars and (b) 5 µm diameter S1813 pillars fabricated by photolithography.
1.2.2 Silicon MEMS fabrication
The original microfabrication processes were developed for the microelectronics
integrated circuit (IC) industry. By combining IC manufacturing techniques with silicon-
based deposition and machining techniques, it is possible to integrate electrical and
mechanical elements (e.g. sensors, actuators, motors, pumps, etc.) on a single silicon
platform.
6
Processing techniques for silicon-based MEMS fabrication vary widely depending
on the materials and applications being considered. MEMS processing can broadly be
separated into two categories, deposition and machining. Deposition techniques allow
different silicon-based materials to be used as either sacrificial or structural materials in
MEMS. Some commonly used structural materials are polycrystalline silicon
(polysilicon) and silicon nitride (Si3N4). Silicon oxide (SiO2) is generally used as a
sacrificial material but can be used as a structural material for certain applications. Other
materials, both silicon and non-silicon based, have begun to gain attention for specific
applications. For example, significant research has focused on silicon carbide (SiC)
MEMS for high temperature, extreme environment applications.5 Other examples include
shape memory alloys6 and magnetic materials.7
Deposition of silicon-based materials is usually done via chemical means.
Polysilicon and Si3N4 are usually deposited using chemical vapor deposition (CVD) or
low pressure chemical vapor deposition (LPCVD). CVD and LPCVD depositions are
performed in tube furnaces at high temperature. The reaction of gases introduced to the
furnace leads to deposition of a solid material on the silicon wafer surface. Silicon oxide
can also be deposited via CVD or LPCVD, but it can also be grown by thermal oxidation
of a silicon substrate.
Physical methods are also used to deposit materials for MEMS applications.
Physical vapor deposition (PVD) processes do not rely on any chemical reaction for the
deposition. PVD processes include sputter coating and electron beam (e-beam)
evaporation and are often used for metal deposition in MEMS processing.
After thin film deposition and photolithography, materials are selectively
removed, or etched, to produce the structural or mechanical elements of MEMS.
Processes for selectively removing materials are referred to as micromachining
techniques. Theses processes are categorized as either surface or bulk micromachining. In
the case of bulk micromachining, the wafer material itself is used as the structural
material for the device. In surface micromachining, a material deposited on the surface of
the wafer is used as the structural material.
Methods of micromachining silicon-based materials can be separated into several
categories depending on the nature of the etching process and the resulting etch profile.
Wet versus dry etching refers to the physical conditions in which the etching is
7
performed. Wet processes use liquid chemical etchants that dissolve the solid material.
Dry etches use reactive gases to etch the materials by either chemical (plasma etching) or
a combination of physical and chemical processes (reactive ion etching). The resulting
etch profile determines if the etch is anisotropic or isotropic. Isotropic etches remove
material at an equal rate independent of direction. Anisotropic etches preferentially
remove material in a specific direction. The selectivity can either be based on the
crystallographic orientation of the substrate or on the directionality of the etch process.
For example, potassium hydroxide (KOH) etches silicon around 400 times faster in the
<100> crystallographic direction compared to the <111> direction.8 The resulting etch
profile for a <100> wafer gives a V-shaped groove with a 54.74º angle between the wafer
surface the sidewall of the groove. Directionality in dry etching (plasma and RIE) is not
based on the crystallography of the substrate, but is controlled by the etch conditions.
Isotropic or anisotropic etching can be achieved depending on the etching parameters.
The difference in the etch profiles for isotropic and anisotropic etches are shown in
Figure 1.3. Various gases are used for dry etching processes. Some common examples
are SF6, CF4, and Cl2. Gas mixtures that incorporate O2, He, and other gases are also used
to increase etch rates, control etch directionality, or for cooling.9 Table 1.1 gives some
examples of different silicon based etching processes and their characteristics. Note that
this table is not a comprehensive list of etch processes, and only provides a few
commonly used examples. For a more complete list, see Madou’s book Fundamentals of
Microfabrication.8
8
Figure 1.3 Typical profiles for silicon based etches.
Etch Chemistry Etch Material Mask Material Wet or dry
HNA Si Si3N4 wet
KOH Si Si3N4/SiO2 wet
TMAH Si SiO2Si3N4 wet
HF, Buffered HF SiO2 Photoresist wet
CF4 SiO2 Photoresist dry
CHF3 Si3N4 Photoresist dry
SF6 Si/PolySi Photoresist dry
Cl2 Si/PolySi Photoresist dry
Table 1.1 Typical etch processes for silicon based materials.
Another important step in the micromachining process is selecting a masking
material. The mask protects the underlying material during the etching process. A high
degree of selectivity is desirable, especially for long or deep etches. The selectivity refers
to the relative etch rate of the substrate material compared to the masking material. Some
9
examples of typical masking materials for a given etch are shown in Table 1.1. In the
case of dry etches with photoresist etch masks, the photoresists typically have relatively
poor selectivity. This requires that the etch depth be rather small or that a thick
photoresist layer be used in order to provide sufficient masking for longer etches.
1.2.3 Applications of silicon MEMS
Silicon MEMS have found commercial applications in a number of different
industries. According to a report by BCC Research, the market for MEMS processing
equipment and devices was estimated $5 billion in 2005 and is expected to grow by more
than 20% per year to more than $12.5 billion by 2010.10 Another report by Yole
Développement estimates a MEMS devices market of around $6 billion in 2006 with an
increase to approximately $11 billion dollars in 2011.11 One example of a commercially
successful device is the MEMS accelerometer, which is widely used in automotive airbag
deployment system. MEMS devices have also found commercial applications in
manufacturing of ink jet printer heads, optical devices, pressure sensors, and various
radio frequency device (RF MEMS).12 Significant research efforts in industry and
academia continue to focus on silicon MEMS device development, processing, materials,
and reliability.
1.2.4 Silicon-based BioMEMS
Medicine and biology have been significantly impacted, both from a research and
commercial perspective, by the development of MEMS technology. Silicon devices have
been considered for drug delivery devices, immunoisolation devices, and sensors, just to
10
name a few. Several review articles are available that summarize research efforts in these
areas.13-16
Figure 1.4 shows an example of a silicon MEMS device for drug delivery.17 The
device consists of an array of drug containing microreservoirs that can be triggered to
release a specific quantity or type of drug. The release mechanism is based on
electrochemically induced dissolution of metal covers on top of the reservoirs. One of the
advantages of this device is that the release can be externally controlled to allow a
continuous or time varying release profile. Another advantage of the individually
addressable reservoirs is the ability to release multiple different drugs as needed. The
device was tested in a dog model and showed steady release of drug over the course of 6
months.18 A more recent version of this device is being commercialized by MicroCHIPS,
Inc.
11
Figure 1.4 Silicon MEMS microchip for controlled drug delivery (17).
Figure 1.5 Immunoisolation biocapsule for insulin delivery (19).
The device in Figure 1.5 is designed for immunoisolation of xenografted
pancreatic islet cells for the treatment of diabetes.19 The device uses a nanoporous
membrane with very small (down to 7 nm) pores to isolate the encapsulated cells from 12
the host immune response while allowing size specific exchange of certain materials. The
membrane with 18 nm pores was shown to significantly reduce Immunoglobulin G (IgG)
diffusion while maintaining insulin diffusion. The device was tested in a mouse model
and effectively provided immunoisolation of the encapsulated cells over a short period of
time (1 week).20
Silicon MEMS are also used in a number of biomedical sensors. Several different
detection mechanisms are used in bioMEMS sensors. Some of the more common are
mechanical, optical and electrochemical detection. As an example, we will consider
mechanical microcantilever sensors. Figure 1.6 shows a schematic of the detection
mechanism for this type of sensor.21 A target molecule is bound to the surface of the
microcantilever. Binding of a receptor molecule creates a stress that causes the cantilever
to deflect. The amount of bending is proportional to the amount of bound receptor. This
concept has been used to characterize DNA hybridization21 and was sensitive enough to
detect a single base pair mismatch between two ssDNA molecules. The same concept has
been used to measure receptor-ligand binding. In experiments by Wu et al.22
microcantilevers were used to measure low concentrations (down to 6 ng/ml) of prostate
specific antigen (PSA), a key marker for prostate cancer. Microcantilevers have also been
used to detect larger particles such as viruses.23 In this case, a change in the resonant
frequency of the cantilever was used as the detection mechanism. When virus particles
attach to the surface of the cantilever, the mass increases which leads to a decrease in the
resonant frequency. This sensor is capable of measuring the attachment of a single virus.
13
Figure 1.6 Detection mechanism for microcantilever sensors. Binding between receptor and target molecules creates a surface stress that bends the cantilever. (21)
1.2.5 Microfluidics
One of the most commercially successful and well-researched areas of the MEMS
field is microfluidic devices. Microfluidic devices are current being explored for
applications that include medical diagnostics,24 DNA and proteomic analysis,25-27 and
drug screening.27 The basic concept of microfluidics involves the miniaturization and
automation of analytical equipment to allow complete analysis on a single chip with little
or no user intervention. Miniaturization of analytical devices allows portability, parallel
analysis, rapid response, and improved efficiency, while automation allows devices to be
used by non-skilled operators. These devices could potentially allow point of care 14
medical diagnostics in a fraction of the time currently needed for traditional laboratory
testing. Microfluidic DNA and proteomics chips have the potential to vastly improve the
time and efficiency of DNA and protein sample preparation, separation, and analysis. It
is not difficult to imagine the potential improvements in patient care and laboratory
technology that could be achieved through these devices. However, the integration of all
the necessary control components and sensing mechanisms needed for a functional
microsystem continues to be an elusive problem.
In order to achieve the automation and miniaturization that are desired in the
microfluidic device, a host of different functional components must be simultaneously
incorporated into a single, small scale device. These components can include valves,
pumps, reaction chambers, separation mechanisms, and detectors.
Given the small size of many of the components of microfluidic systems,
microfabrication has logically been the method of choice for fabrication of microfluidic
systems. A number of different materials and microfabrication techniques have been
employed. Silicon and glass were the first materials to be considered due to their
common use in microfabrication and MEMS, but for cost considerations and ease of
fabrication, polymer based systems have begun to dominate the field.
Fabrication methods for polymer microfluidic devices vary significantly from
silicon MEMS. Polymer devices are usually fabricated from thermoplastic polymers like
poly(methyl methacrylate) (PMMA) or polycarbonate or from the elastomer
polydimethylsiloxane (PDMS). Devices made from thermoplastic polymers are
commonly fabricated by hot embossing or injection molding.28,29 The first step in the hot
embossing process is fabrication of a mold. The mold is usually made from a hard
15
material such as a metal (e.g. nickel) or silicon or from a polymer with a high glass
transition temperature relative to the substrate material. The mold is generally fabricated
using standard microfabrication processes (photolithography, etching, etc.) sometimes in
combination with other processes (e.g. electroplating). The mold is then used as a tool to
fabricate multiple polymer devices. The substrate and/or mold are heated above the glass
transition temperature of the substrate polymer. Pressure is then applied to the mold to
transfer the channel pattern into the polymer. An image of a commercially available
microscale hot embossing system (EVG 520, EV Group) is shown in Figure 1.7. The
system consists of two heated/cooled plates with computer controlled hydraulics that
apply pressure to the plates. The pressure and temperature versus time profiles can both
be programmed.
Figure 1.7 (a) Commercially available microscale hot embossing system (EVG 520), (b) close-up of the heated plates.
Another commonly used method for fabrication of polymer microfluidic is soft
lithography. Soft lithography covers a wide range of fabrication techniques that involve 16
replication of microstructures in a soft material (PDMS).30 The PDMS microstructures
can then be used directly in device applications (such as microfluidics) or as tools for
further fabrication processes. The use of PDMS microstructures as microfabrication tools
will be discussed in detail in a later section.
The Effenhauser group was also one of the first to use PDMS for microfluidics.31
They focused specifically on polydimethylsiloxane (PDMS) as the structural material in a
protein separation device. Duffy et al.4 further expanded the use of PDMS as a materials
in microfluidic systems. In addition to the reduction in cost resulting form the use of
PDMS, the photomask used in the initial fabrication process was printed on a high
resolution printer instead of the expensive methods used in traditional photomask
fabrication. They also developed a method for covalently sealing PDMS to a number of
substrates with a simple exposure to an oxygen plasma. Oxygen plasma also changes the
surface chemistry of the PDMS, rendering it hydrophilic, versus its native, highly
hydrophobic state. This allows for easier filling of the PDMS microchannels with polar
fluids such as water and biological fluids. The charged surface of oxidized PDMS (SiO-)
also provides a negatively charged surface that is conducive to electroosmotic flow.32 A
significant amount of research has since focused on PDMS microfluidic devices for a
wide range of applications. Several review papers have been published in this area.33-35
1.2.6 Polymer microfabrication
17
In addition to microfluidic devices, polymers are being used more frequently in
other microfabricated devices. Many polymer microfabrication techniques are based on
soft lithography. Figure 1.8 shows the process for fabrication of PDMS molds for soft
lithography. A photoresist master is first fabricated using photolithography or another
micro/nanofabrication procedure. The PDMS prepolymer is mixed with a curing agent
and poured over the master and allowed to cure. The master pattern is transfer to the
PDMS creating a negative replica of the original pattern.
Figure 1.8 Schematic diagram for fabricating a PDMS mold from a photoresist master.
Whitesides et al. developed the original soft lithography techniques and
subsequently used them to create micropatterns and microstructures from a variety of
materials. Microtransfer molding (µTM)36 and micromolding in capillaries (MIMIC)37
use microfabricated PDMS molds as templates for further replication of polymer
structures. In µTM, a prepolymer is applied to the PDMS mold. The excess polymer on
top of the mold is removed and the mold is applied to a substrate. The prepolymer is then
18
cured using UV or heat. This process will be discussed in more detail in Chapter 3. For
the MIMIC process, the mold is applied to the substrate and a prepolymer solution is
applied to one side of the mold. The prepolymer fills the features in the mold either by
capillary action or using vacuum. The prepolymer is then cured and the mold is removed.
Related soft lithography techniques include replica molding, solvent assisted
micromolding,38 and microcontact printing (µCP).39 In solvent assisted micromolding,
the PDMS is used to imprint a polymer that is dissolved in a solvent. The solvent diffuses
through the mold over time, leaving only the polymer in the microfeatures. In µCP, the
PDMS mold is “inked” with a specific chemistry, then the mold is brought into contact
with a substrate, selectively transferring the chemical pattern to the substrate. This
process is generally used for micropatterning self assembled monolayers (SAMs) or
proteins.40 Each of these processes are shown in Figure 1.9.
19
Figure 1.9 Schematics for soft lithography processes (38).
Since the development of these initial processing techniques, several variations or
related techniques have been developed.41,42 These processes have greatly broadened the
versatility of the soft lithographic microfabrication techniques. In addition to processing
traditional polymers, significant process has also been made in micropatterning
20
functional materials such as conducting polymers,43-45 piezoelectric polymers,46,47
biodegradable polymers,48,49 and hydrogels.50
While soft lithography processes have found broad applications, they are
generally limited to fabrication of two-dimensional (2D) or extruded 2D structures. For
fabrication of more complicated three-dimensional shapes or suspended mechanical
structures, more complicated techniques are required. In some cases, traditional
microfabrication techniques such as photolithography with SU8 photoresist and
specialized etching and sacrificial layer techniques can be used to fabricate structural
components in polymer MEMS.51,52 In other cases, more complicated techniques such as
two-photon lithography53,54 and microstereolithography55 can be used to fabricate 3D
polymer structures. However, these processes have additional limitations such as high
cost, diminished resolution, and low throughput.
1.2.7 Applications of polymer MEMS
Microfabricated polymer structures have found applications in a number of areas,
either in the form of all polymer devices or as components in MEMS. Polymers are used
in electronic and optical devices. For example, conducting polymers have been used to
make all organic field effect transistors (FETs).56,57 Soft lithography has been used to
fabricate polymer waveguides.58 Polymers have also been used as components in various
sensors (pressure sensors, accelerometers, etc.).59 Figure 1.10 shows a hybrid
polymer/metal tactile sensor capable of measuring hardness, thermal conductivity,
temperature and surface contour.60
21
Figure 1.10 Hybrid polymer/metal tactile sensor (60).
Polymer microfabrication has also been used extensively in the biomedical field.
One of the advantages of microfabrication and MEMS is the ability to interface with cells
on their size scale. This ability to interact with, manipulate, and perform measurements
on a single cell and biomolecular level has lead to new devices and therapeutic
approaches. Some examples include drug delivery devices,61,62 tissue engineering
scaffolds,46,47 and cell micromanipulation devices.63 Figure 1.11 shows several examples
of microfabricated drug delivery devices. Figure 1.11 (a) is a polymer microreservior
system for drug delivery. Biodegradable polymers are used to provide controlled drug
release. Figure 1.11 (b-d) shows microfabricated biodegradable polymer microparticles
for drug delivery. Microsphere delivery systems are used in several commercial drug
delivery systems to give long term, sustained release.64 Microfabrication provides a
method to produce monodisperse particles that could potentially provide a more stable
release profile than particles produced using current manufacturing methods.
Microfabrication also allows fabrication of multiple particle geometries as opposed to
22
only spheres. Flat square or circular particles, high and low aspect ratio cylinders, and
many other geometries can be fabricated. Particle geometry could be used to give specific
release profiles.
Figure 1.11 (a) Polymer microchip drug delivery device (65), (b) square poly(lactic-co-glycolic acid) (PLGA) microparticles fabricated by soft lithography (61), (c) high aspect ratio and (d) low aspect ratio PLGA microparticles fabricated by soft lithography (66).
Scale bars equal 50 µm, 5 µm, and 20 µm in (b), (c), and (d), respectively.
23
Figure 1.12 shows two polymer tissue engineering scaffolds fabricated using soft
lithography. Microfabrication is often used in tissue engineering to provide highly
controlled cellular environments at the microscale.67 This is an important factor in
determining cell behavior. Microfabrication techniques allow precise control of
microarchitecture to engineer structure with well-defined pore size and porosity. Figure
12 (a) shows a PLGA tissue engineering scaffold fabricated by embossing and CO2
assisted bonding68 and (b) shows a polycaprolactone (PCL) scaffold fabricated using
multilayer micromolding.49
Figure 1.12 Microfabricated (a) PLGA (68) and (b) PCL (49) tissue engineering scaffolds.
1.3 Conclusion
MEMS have found applications in wide range of industries and continue to be a
focus area in the research community. Silicon micromachining techniques including
24
deposition and silicon micromachining are well established and well characterized
processes. One of the major focuses of MEMS research and commercialization efforts
has been in the biomedical field. Using microdevices to interface with biology on a
cellular and molecular level has led to new diagnostic, therapeutic, and sensor
applications. As the field of MEMS has progressed, polymers have been increasingly
considered as structural and functional components for specific applications. Polymers
offer several advantages over silicon-based materials, including biocompatibility, diverse
chemical, physical, and mechanical properties, and low cost materials and processing
techniques. The push toward incorporating polymers into MEMS devices has led to the
development of new microfabrication techniques to meet the unique challenges of
polymer processing. Some of the focus areas for polymer MEMS research are
microfluidic analytical devices, sensors, drug delivery devices, and tissue engineering. As
research in these areas continues and new applications are explored, advances will no
doubt be made toward the development of low cost, functional polymer MEMS devices
for a wide range of applications in biomedicine and other fields.
References
1 W. Kern (ed.), Handbook of Semiconductor Wafer Cleaning Technology: Science, Technology, and Applications, Noyes, Park Ridge, NJ (1993). 2 A.G. Ernslie, F.T. Bonner, L.G. Peck, “Flow of a viscous liquid on a rotating disk.” Journal of Applied Physics 29 (1958) 858-862. 3 D. Meyerhofer, “Characteristics of resist films produced by spinning.” Journal of Applied Physics 49 (1978) 3993-3997. 4 D.C. Duffy, J.C. McDonald, O.J.A. Schueller, G.M. Whitesides, “Rapid prototyping of microfluidic systems in poly(dimethylsiloxane).” Analytical Chemistry 70 (1998) 4974-4984.
25
5 M. Mehregany, C.A. Zorman, “SiC MEMS: opportunities and challenges for applications in harsh environments.” Thin Solid Films 355-356 (1999) 518-524. 6 Y. Fu, H. Du, W. Huang, S. Zhang, M. Hu, “TiNi-based thin films in MEMS applications: a review.” Sensors and Actuators A 112 (2004) 395-408. 7 D. Niarchos, “Magnetic MEMS: key issues and some applications.” Sensors and Actuators A 109 (2003) 166-173. 8 M. Madou, Fundumentals of Microfabrication, CRC Press, Boca Raton, FL (1997) pp.171. 9 H. Janseny, H. Gardeniers, M. de Boer, M. Elwenspoek, J. Fluitman, “A survey on the reactive ion etching of silicon in microtechnology.” Journal of Micromechanics and Microengineering 6 (1996) 14-28. 10 Microelectromechanical Systems (MEMS) Technology: Current And Future Markets, BCC Research (2006). 11 Global MEMS/Microsystems - Markets and Opportunities, Yole Développement (2007). 12 W.H. Ko, “Trends and frontiers of MEMS.” Sensors and Actuators A 136 (2007) 62-67. 13 R.S. Shawgo, A.C. Richards Gradyson, Y.W. Li, M.J. Cima, “BioMEMS for drug delivery.” Current Opinions in Solid State Materials Science 6 (2002) 329-334. 14 A.C. Richards Grayson, R.S. Shawgo, Y. Li, M.J. Cima, “Electronic MEMS for triggered delivery.” Advanced Drug Delivery Reviews 56 (2004) 173-184. 15 S.L. Tao, T.A. Desai, “Microfabricated drug delivery systems: from particles to pores.” Advanced Drug Delivery Reviews 55 (2003) 315-328. 16 R. Bashir, “BioMEMS: state-of-the-art in detection, opportunities and prespects.” Advanced Drug Delivery Reviews 56 (2004) 1565-1585. 17 J. Santini, M.J. Cima, R.J. Langer, “A controlled-release microchip.” Nature 397 (1999) 335-338. 18 M. Staples, K. Daniel, M.J. Cima, R. Langer, “Application of micro- and nano-electromechanical devices to drug delivery.” Pharmaceutical Research 23 (2006) 847-863.
26
19 T.A. Desai, D. Hansford, M. Ferrari, “Characterization of micromachined silicon membranes for immunoisolation and bioseparation applications.” Journal of Membrane Science 159 (1999) 221-231. 20 T.A. Desai, W.H. Chu, G. Rasi, P. Sinibaldi-Vallebona, E. Guarino, “Microfabricated biocapsules provide short-term immunoisolation of insulinoma xenografts.”Biomedical Microdevices 1 (1999) 131-138. 21 J. Fritz, M.K. Baller, H.P. Land. H. Rothuizen, P. Vettiger, E. Meyer, H.-H. Güntherodt, Ch. Gerber, J.K. Gimzewski, “Translating biomolecular recognition into nanomechanics.” Science 288 (2000) 316-318. 22 G. Wu, R.H. Datar, K.M. Hansen, T. Thundat, R.J. Cote, A. Majumdar, “Bioassy of prostate-specific antigen (PSA) using microcantilevers.” Nature Biotechnology 19 (2001) 856-860. 23 A. Gupta, D. Akin, R. Bashir, “Single virus particle mass detection using microresonators with nanoscale thickness.” Applied Physics Letters 84 (2004) 1976-1978. 24 T.H. Schulte, R.L. Bardell, B.H. Weigl, “Microfluidic technologies in clinical diagnostics.” Clinica Chimica Acta 321 (2002) 1-10. 25 G.H.W. Sanders, A. Manz, ”Chip-based Microsystems for genomic and proteomic analysis.” Trends in Analytical Chemistry 19 (2000) 364-378. 26 Yu, X.; Zhang, D.; Li, T.; Hao, L.; Li, X. “3-D microarrys biochip for DNA amplification in polydimethylsiloxane (PDMS) elastomer.” Sensors and Actuators A 108 (2003) 103-107. 27 M.T. Cronin, T. Boone, A.P. Sassi, H. Tan, Q. Xue, S.J. Williams, A.J. Ricco, H.H. Hooper. “Plastic microfluidic systems for high-throughput genomic analysis and drug screening.” Journal of the Association for Laboratory Automation 6 (2001) 71-75. 28 H. Becker, C. Gärtner, “Polymer microfabrication methods for microfluidic analytical applications.” Electrophoresis 21 (2000) 12-26. 29 H. Becker, C. Gärtner, “Polymer microfabrication technologies for microfluidic systems.” Analytical and Bioanalytical Chemistry 390 (2008) 89-111. 30 Y. Xia, G.M. Whitesides, “Soft lithography.” Angewandte Chemie International Edition 37 (1998) 550-575. 31 C.S. Effenhauser, G.J.M. Bruin, A. Paulus, M. Ehart, Analytical Chemistry 69 (1997) 3451-3457.
27
32 J.S. Buch, P.-C. Wang, D.L. DeVoe, C.S. Lee, “Field-effect flow control in a polydimethylsoloxane based microfluidic system.” Electrophoresis 22 (2001), 3902-3907. 33 J.C. McDonald, D.C. Duffy, J.R. Anderson, D.T. Chiu, H. Wu, O.J.A. Scheuller, G.M. Whitesides, “Fabrication of microfluidic systems in poly(dimethylsiloxane).” Electrophoresis 21 (2000), 27-40. 34 S.K. Sia, G.M. Whitesides, “Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies.” Electrophoresis 24 (2003), 3563-3576. 35 J.M.K. Ng, I. Gitlin, A.D. Stroock, G.M. Whitesides, “Components for integrated poly(dimethylsiloxane) microfluidic systems.” Electrophoresis 23 (2002), 3461-3473. 36 X.-M. Zhoa, Y. Xia, and G.M. Whitesides, “Fabrication of three-dimensional micro-structures: microtransfer molding.” Advanced Materials 8 (1996) 837-840. 37 E. Kim, Y. Xia, and G.M. Whitesides, “Polymer microstructures formed by moulding in capillaries.” Nature 376 (1995) 581-584. 38 Y. Xia, G.M. Whitesides, “Soft lithography.” Annual Review of Materials Science 28 (1998) 153-184. 39 A. Kumar, G.M. Whitesides, “Features of gold having micrometer to centimeter dimensions can be formed through a combination of stamping with an elastomeric stamp and an alkanethiol “ink” followed by chemical etching.” Applied Physics Letters 61 (1993) 2002-2004. 40 R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, G.M. Whitesides, “Patterning proteins and cells using soft lithography.” Biomaterials 20 (1999) 2363-2376. 41 J.A. Rogers, R.G. Nuzzo, “Recent progress in soft lithography.” Materials Today 8(2) (2005) 50-56. 42 M. Wang, H.-G. Braun, T. Kratzmüller, E. Meyer, “Patterning polymers by micro-fluid-contact printing.” Advanced Materials 13 (2001) 1312-1317. 43 W.S. Beh, I.T. Kim, D. Qin, Y. Xia, G.M. Whitesides, “Formation of patterned microstructures of conducting polymers by soft lithography, and applications in microelectronic device fabrication.” Advanced Materials 11 (1999) 1038-1041. 44 J.A. Rogers, Z. Boa, V.R. Raju, “Nonphotolithographic fabrication of organic transistors with micron feature sizes.” Applied Physics Letters 72 (1998) 2716-2718.
28
45 T. Granlund, T. Nyberg, L.S. Roman, M. Svensson, O. Inganäs, “Patterning of polymer light-emitting diodes by soft lithography.” Advanced Materials 12 (2000) 269-273. 46 B. Xu, F. Arias, S.T. Brittain, X.-M. Zhao, B. Grzybowski, S. Torquato, G.M. Whitesides, “Making negative Poisson’s ratio microstructures by soft lithography.” Advanced Materials 11 (1999) 1186-1189. 47 D. Gallego, N. Ferrell, D. Hanford, “Fabrication of Piezoelectric Polyvinylidene Fluoride (PVDF) Microstructures by Soft Lithography for Tissue Engineering and Cell Biology Applications.” in Printing Methods for Electronics, Photonics and Biomaterials, edited by G. Gigli (Mater. Res. Soc. Symp. Proc. Volume 1002E, Warrendale, PA, 2007) 1002-N04-05. 48 G. Vozzi, C. Flaim, A. Ahluwalia, S. Bhatia, “Fabrication of PLGA scaffolds using soft lithography and microsyringe deposition.” Biomaterials 24 (2003), 2533-2540. 49 D. Gallego, N. Ferrell, Y. Sun, D.J. Hansford, “Multilayer micromolding of degradable polymer tissue engineering scaffolds.” Materials Science and Engineering C (in press). 50 J. Guan, H. He, D.J. Hansford, L.J. Lee, “Self-folding of three-dimensional hydrogel microstructures.” The Journal of Physical Chemistry B 109 (2005) 23134-23137. 51 G. Genolet, J. Brugger, M. Despont, U. Drechsler, P. Vettiger, N.F. de Rooij, D. Anselmetti, “Soft, entirely photoplastic probes for scanning force microscopy.” Review of Scientific Instruments 70 (1999) 2398-2401. 52 X. Wang, J. Engel, C. Liu, “Liquid crystal polymer (LCP) for MEMS: processes and applications.” Journal of Micromechanics and Microengineering 13 (2003) 628-633. 53 G. Witzgall, R. Vrijen, E. Yablonovitch, V. Doan, B.J. Schwartz, “Single-shot two-photon exposure of commercial photoresist for the production of three-dimensional structures.” Optics Letters 23 (1998) 1745-1747. 54 T. Baldacchini, C.N. LaFratta, R.A. Farrer, “Acrylic-based resin with favorable properties for three-dimensional two-photon polymerization.” Journal of Applied Physics 95 (2004) 6072-6076. 55 A. Bertsch, H. Lorenz, P. Renaud, “3D microfabrication by combining microstereolithography and thick resist UV lithography.” Sensors and Actuators A 73 (1999) 14-23. 56 W.S. Beh, I.T. Kim, D. Qin, Y. Xia, G.M. Whitesides, “Formation of patterned microstructure of conducing polymer by soft lithography, and applications in microelectronic device fabrication.” Advanced Materials 11 (1999) 1038-1041.
29
57 H. Sirringhaus, T. Kawase, R.H. Friend, T. Shimoda, M. Inbasekaran, W. Wu, E.P. Woo, “High-resolution inkjet printing of all-polymer transistor circuits.” Science 290 (2000) 2123-2126. 58 X.-M. Zhao, S.P. Smith, S.J. Waldman, G.M. Whitesides, M. Prentiss, “Demonstration of waveguide couplers fabricated using microtrasnfer molding.” Applied Physics Letters 71 (1997) 1017-1019. 59 C. Liu, “Recent developments in polymer MEMS.” Advanced Materials 19 (2007) 3783-3790.
60 J. Engel, J. Chen, Z. Fan, C. Liu, “Polymer micromachined multimodal tactile sensors.” Sensors and Actuators A 117 (2005) 50-61.
61 G. Guan, N. Ferrell, L.J. Lee, D.J. Hansford, “Fabrication of polymeric microparticles for drug delivery by soft lithography.” Biomaterials 27 (2006) 4034-4041.
62 S.L. Tao, T.A. Desai, “Microfabrication of multilayer, asymmetric, polymeric devices for drug delivery.” Advanced Materials 17 (2005) 1625-1630.
63 N. Chronis, L.P. Lee, “Electrothermally activated SU-8 microgripper for single cell manipulation in solution.” Journal of Microelectromechanical Systems 14 (2005) 857-863.
64 M. Ferrari, A. Lee, L.J. Lee (eds.), Biological and Biomedical Nanotechnology, Springer, Boston, MA (2006).
65 A.C. Richards Grayson, I.S. Choi, B.M. Tyler, P.P. Wang, H. Brem, M.J. Cima, R. Langer, “Multi-pulse drug delivery from a resorbable polymeric microchip device.” Nature Materials 2 (2003) 767-772.
66 A. Chakrapani, “Processing and characterization of polymer microparticles for controlled drug delivery systems.” Ph.D. Thesis, Ohio State University (2007).
67 S.N. Bhatia, C.S. Chen, “Tissue engineering at the microscale.” Biomedical Microdevices 2 (1999) 131-144.
68 Y. Yang, S. Basu, D.L. Tomasko, L. James Lee, S.-T. Yang, “Fabrication of well-defined PLGA scaffolds using microembossing and carbon dioxide bonding.” Biomaterials 26 (2005) 2585-2594.
30
CHAPTER 2
CELL MECHANICS
2.1 Introduction
Cell mechanics is a broad field that covers a range of mechanical processes,
structural elements, and mechanical properties as they relate to cells and cell behavior.
Cell processes such as adhesion and motility have a significant mechanical component.
Less intuitive and less understood is the role of mechanical forces in regulating other
biochemical processes that ultimately control cell behavior. This interrelationship
between mechanical and biochemical processes is currently an active area of research
often referred to as mechanotrasduction or chemomechanical transduction.1 Mechanical
cues act as regulatory signals to control intracellular biochemical processes. External
factors like extracellular matrix or substrate stiffness and external shear stress are
involved in mediating a host of cell processes including gene expression, differentiation,
and proliferation. Cells are also responsible for generating forces that are transmitted to
the external environment. In fibroblast cells, for example, intracellular forces are
transmitted to the extracellular space and play a critical role in cell movement, adhesion,
and signal transduction.
The breadth and depth of the field of cell biomechanics makes a comprehensive
review of the field impractical. A few specific areas of cell mechanics will be covered as
31
they relate to the development and testing of the polymer MEMS cell force sensor
described in Chapters 9 and 10. Topics such as muscle cell mechanics or mechanics of
non-adherent cells will not be discussed in detail. Several measurement devices and
techniques for evaluating and measuring cell mechanics will be covered.
2.2 Cell mechanics and cell forces
2.2.1 The cytoskeleton
The cytoskeleton is set of proteins in the cell that act as the primary load bearing
and force generating component of the cell. The cytoskeleton consists of three primary
proteins: actin, microtubules, and intermediate filaments. A host of associated proteins
and protein complexes are responsible for polymerization and depolymerization of
cytoskeletal components, linking cytoskeletal component to each other and to the
extracellular space, and regulating cytoskeletal function. Figure 2.1 shows the three
molecular components of the cytoskeleton. F-actin, which is the filament form of the
monomer G-actin, is the primary structural component of most cell types. Actin filaments
consist of twisted strands of actin and have a diameter of 7-9 nm. Actin filaments are
polar molecules with a positive barded end where polymerization takes place and a
negative pointed end where depolymerization occurs. Actin filaments are responsible in
large part for maintaining the structure and morphology of the cells. Polymerization and
depolymerization of actin filaments is also the primary mechanism of cell motility in
adherent cells such as fibroblast.
Microtubules are tubular dimeric proteins consisting of alternating strands of α
and β tubulin. Microtubules are larger than actin filaments, with an outside diameter of
32
about 25 nm. They are also polar, and polymerize faster from the plus end. Microtubules
also play a structural role in the cell as well as being involved cell division and transport
processes. Microtubules work in connection with the motor proteins dynein and kinesin
to provide directional transport of organelles.
Figure 2.1 Cytoskeletal proteins, (A) actin filament, (B) microtubule, and (C) intermediate filament (2).
33
Intermediate filaments are large subset of cytoskeletal proteins. In terms of size,
their diameter is between that of actin filaments and microtubules (~10 nm), thus the
name intermediate filaments. The assembly process for intermediate filaments has several
general steps as shown in Figure 2.1. Polypeptides first assemble into coiled dimers then
into tetramer. The individual tetramers assemble into a protofilament. Protofilments then
assemble in filaments. Unlike actin filaments and microtubules, intermediate filaments
are highly stable.
2.2.2 Forces generated by adherent cells
When adherent cells are attached to the extracellular matrix or a tissue culture
surface, stresses are generated within the cytoskeleton. Cytoskeletal stresses are
responsible for maintaining the structural integrity of the cell and regulating cellular
shape and spreading. Molecular linkages between the cytoskeleton and focal adhesion
complexes (FACs) transmit forces to the extracellular space. Figure 2.2 shows the
molecular mechanism of force transmission.3 Within FACs, actin filaments are connected
to transmembrane integrin receptors via protein linkages. Various proteins and protein
complexes are involved in linking actin filaments to integrins. Some examples include α-
actinin, talin, paxillin, and viculin.2,3 Integrins are then attached either to the extracellular
matrix (in vivo) or to proteins on the cell culture substrate (in vitro).
34
Figure 2.2 Molecular components of a focal adhesion (3).
Recent research has shown that mechanical interactions that occur at the cell-
extracellular matrix interface are critical in regulating a host of cellular processes. Cell
motility requires that cells generate the necessary forces in order to move.4 Figure 2.3
shows the basic processes involved in cell motility. Protrusions, in the form of filopodia
or lamellapodia are formed at the lead edge of the cell. These areas are rich in highly
dynamic actin filaments. Contractile forces are created in the cytoskeleton. Focal
adhesions at the trailing edge of the cell are released, thus propelling the cell in the
direction toward the leading edge.
35
Figure 2.3 Cell motility process (2).
Intracellular tension, force transduction across the cell membrane, and associated
processes such as cell shape and spreading have been show to affect a number of other
cellular processes in addition to cell motility.5,6 A number of interesting experiments have
been performed to illustrate this point. By applying mechanical stress directly to integrin
receptors via magnetic twisting, Wang et. al7 demonstrated assembly of focal adhesions
and stiffening of the cytoskeleton proportional to the applied stress. These experiments
are evidence that the integrins are directly involved in transmitting forces between the
cytoskeleton and the extracellular matrix. By chemically inhibiting actin filament,
microtubule, and intermediate filament function independently, they also showed that all
three elements of the cytoskeleton are involved in the mechanical signal transduction
process. Chen et al. used micropatterned cell adhesion molecules to control the degree of
cell spreading in bovine and human endothelial cells. These experiments showed that
restricting the cells ability to spread induced apoptosis.8 This work demonstrated that
cells have the ability to use geometrical factors as regulatory cues to determine cell fate.
Chrzanowska-Wodnicka and Burridge9 showed that rhoA, a GTP binding protein,
36
increases cell contractility corresponding to the formation of focal adhesions and stress
fibers in fibroblast cell. These experiments are just a sampling of the evidence for the bi-
directional link between mechanical and biochemical processes in cells. Several review
and commentary articles are available with more information on the theory and
experimental evidence of the role of cell and cytoskeletal mechanics in regulating cell
function.5,6,10-12
2.2.3 The cytoskeleton and disease
In addition to normal cell function, a number of pathologies have a direct effect
on cytoskeletal function. A prime example is the case of metastatic cancers. Metastasis is
largely dependant on the cells ability to migrate from one location to another. As
demonstrated previously, the process of cell motility is largely dependant on cytoskeletal
dynamics. In a number of cancers, the regulatory mechanisms that control the
cytoskeleton do not function properly. In some cases, direct mutations of cytoskeletal
actin have been observed. In other cases, mutation or overexpression of regulatory
proteins such as members of the WASP family of proteins or Rho GTPases could be
responsible for changes in cytoskeletal function.13,14
The cytoskeleton is directly related to a number of other diseases.15 Mutations in
keratin intermediate filaments in epithelial cells are responsible for a range of skin
disorders.16,17 These disorders are often characterized by fragile skin and blistering. This
illustrates the important role of intermediate filament in maintaining the structural
integrity of cells under stress. Keratin intermediate filaments have also been implicated in
various liver pathologies and could be a factor in intestinal disease.18,19 A lack of the actin
37
regulatory protein WASP has been implicated in the immune disorder Wiskott-Aldrich
syndrome.20
2.2.4 Drugs and cell mechanics
Given the number of pathologies associated with the cytoskeleton, it is no surprise
that drugs have been developed to target specific cytoskeletal components. Various
chemotherapeutic agents with cytoskeletal targets have been explored. Primarily these
drugs have targeted the microtubules due to their important role in cell division.
Paclitaxel (Taxol®) is now commonly used to treat various cancers including lung,
ovarian, and breast cancer. Paclitaxel stabilizes microtubules to hinder cell mitosis.21-22
Vinca alkaloids23 and colchicine24 also target microtubules and inhibit depolymerization
to reduce cell proliferation. Vinca alkaloids are used widely used as chemotherapeutics
and colchicine has been explored as an anti-cancer drug as well as being used as a
treatment for gout. Cucurbinacin E has shown anti-proliferative effects in prostate cancer
cells and been shown to induce significant disruption of the actin cytoskeleton.25
38
Several other chemicals, while not necessarily being actively explored for
therapeutic applications, still have a significant influence on the cytoskeleton and have
been indispensable in studying cytoskeletal behavior. Cytochalsins are a group of fungal
derived toxins that interfere with actin polymerization primarily by binding the barbed
end of actin filaments and decreasing the amount of monomeric actin in the cell.26
Cytochalasins, especially cytochalasin D, have been widely used to study the role of actin
in various cellular processes. Phalliodin is another fungal toxin with a significantly
different effect on actin. Phalloidin inhibits depolymerization of actin filaments.
Phalloidin binds strongly to actin filaments and much less to monomeric actin. This shifts
the equilibrium between F-actin filaments and monomeric G-actin toward the formation
of filaments.26 Jasplakinolide, a toxin derived from marine sponges, has been shown to
induce actin polymerization in vitro27, though some in vivo observations are less straight
forward.28
2.3 Measuring cell mechanics
A wide array of different techniques and experimental methods have been
developed to study cell mechanics. For the purposes of this discussion, the techniques
will be broadly grouped to three types of experimental methods: (1) applying forces to
cells, (2) measuring the mechanical properties of cells, and (3) measuring forces
generated by cells. Each of these is important in its own right and considerable crossover
exists between these methods with regards to their importance in understanding cell
behavior.
2.3.1 Appling forces to cells
Several methods have been developed to deliver different types of mechanical
stimuli to various cell types. Externally applied shear stress is an important regulatory cue
for many cell types, especially endothelial cells. Shear stress has been shown in induce
cytoskeletal remodeling, changes cell morphology and orientation, and affect biological
function in endothelial cells.29-31 One of the most common methods to study shear
induced endothelial cell response is using a parallel plate flow through systems. This is a
simple method for controlling shear stress on cells by controlling the flow rate through a
chamber with fixed dimensions. The other commonly used method is the cone plate
system. This method uses a rotating cone and a fixed plate to create a controlled shear
39
stress. These studies have been critical in understanding mechanotransduction pathways
as well as understanding the specific shear stress induced changes to biochemical
processes within endothelial cells. Mechanical forces have also been applied to cells
using substrate deformation techniques.32,33 For these experiments, a flexible substrate is
mechanically deformed by applying uniaxial tension, biaxial tension, bending, or a range
of other mechanical loading conditions. These experiments have been used to study
muscle, bone, cartilage, and a variety of other relevant cells.
2.3.2 Measuring the mechanical properties of cells
Another active area of cell mechanics involves measuring the mechanical
properties of cells and cellular components. Measurements have been carried out on
whole cells or on specific regions of the cells. Given that the cytoskeleton is largely
responsible for maintaining the structural integrity to the cell, there is a direct link
between the elastic properties of the cell and cytoskeletal mechanics. One of the most
commonly used tools to study cellular and subcellular mechanical properties is atomic
force microscopy (AFM). AFM has been used to create elasticity maps to evaluate the
local mechanical properties of several different cell types.34 These studies have been used
to show variations in the elastic properties of a cell depending on location. These
techniques has also been used to evaluate the effects of different chemicals on the
mechanical properties cells.35 A significant decrease was observed in the elastic modulus
of cells after exposure to actin disrupting agents (e.g. cytochalasins, jasplakinolide) at
sufficient concentrations, but no significant changes to the elastic properties were
observed in response to microtubule targeting agents (e.g. cholcicine, taxol).
40
Other methods for measuring the mechanical properties of whole cells include
micropipette aspiration and optical tweezers.36 In the case of micropipette aspiration, a tip
with an inside diameter smaller than the cell is used and a known pressure is applied to
pull the cell into the pipette tip. The deformation of the cell is observed in a microscopy
and the elastic and viscous properties of the cell can be calculated.37 Optical tweezers use
laser light to apply a known force to a bead. By attaching the bead to a cell, the cell can
be deformed and the force and deformation information can be used to calculate the
mechanical properties. Optical tweezers have very good force resolution (~1 pN), but a
limited range of force (~ 600 pN). Higher forces require higher laser power and local
heating can damage the cell. This technique has primarily been applied to study of red
blood cells38 and biomolecules.39
2.3.3 Measuring forces generated by cells
41
Some of the seminal work studying forces at the cell-surface interface was done
by Harris et al.40 who grew cells on cross-linked silicone rubber substrates and observed
the wrinkling of the polymer film. One limitation with this method is the difficulty in
quantifying the forces. Since this initial work, improved methods have been developed
that incorporate fluorescent microbeads into polyacrylamide gels.41 The displacement of
the beads and the stiffness of the gel are used to calculate forces exerted by the cell. This
method is commonly referred to as traction force microscopy.42 While calculation of the
forces is still not straightforward, this process does allow quantification of spatially
resolved forces applied to the substrate. Figure 2.3 shows (a) the bead displacement
vectors overlaid on a phase contrast image of the cell and (b) a corresponding color-
coded traction stress map for a NIH 3T3 fibroblast.
Figure 2.4 Traction force microscopy for measuring fibroblast forces; (a) Displacement vector field overlaid on a phase contrast image of the cell and (b) corresponding stress
map.
Some interesting work has been done looking at the effects of substrate
mechanical properties on cell behavior. Pelham and Wang43 used polyacrylamide gels
with a range of stiffness to show that the formation of focal adhesions and motility of
epithelial (normal rat kidney) and fibroblasts (3T3) cells is dependant on the stiffness of
the substrate. Lo et al.44 took this work a step further and showed that 3T3 cells
preferentially migrate from the softer regions to the stiffer regions of the substrate. Cells
also responded to locally induced stresses created with a microneedle and migrated
42
toward the regions of higher tension and away from region of compression. They used
particle tracking methods to measure the forces generated by the cells, and showed that
the forces generated by cells on the stiffer substrates were higher than that on the softer
materials. These experiments demonstrate several important issues. First, the cells sense
the mechanical properties of their surroundings and response accordingly. This work also
demonstrates a link between substrate and cytoskeletal mechanics given that the forces
exerted on the soft materials were less that those on the stiffer material.
2.3.4 BioMEMS for measuring cell forces
Several bioMEMS devices, both silicon and polymer, have been developed for
measuring cell mechanics. Figure 2.4 shows an SEM micrograph of a silicon MEMS
heart cell force transducer.45 The device consists of a clamp to hold the cell in place. The
clamp is connected to a set of thin polysilion beams that are then attached via a slider
restraint to the substrate. The heart cell is attached to the clamp using silicone sealant
When the cell is stimulated to contract (by varying [Ca2+]), the polysilicon beams are
bent toward the center of the device. By measuring the magnitude of the displacement,
the force can be calculated. While this device clearly demonstrates measurement of
mechanical forces generated by cells, it is designed to measure the relatively large forces
generated by cardiomyocytes. The general measurement methodology, namely using
cantilever beam deflection to measure cell forces, can be applied more broadly.
43
Figure 2.5 Polysilicon heart cell force sensor (45).
Galbraith and Sheetz developed a silicon-based cantilever device for measuring
fibroblast cell forces.46 The device consists of an array of single cantilever beams with a
pad at the end of each beam as shown in Figure 2.5. When the cell attaches to the pad and
exerts a force, the cantilever is bent proportional to the applied load. Using this device,
they were able to measure traction forces exerted by fibroblasts at various locations on
the cell (i.e. lead edge, nucleus, and trailing edge) to determine the spatial distribution of
cell forces. They were also able to measure the time varying force. One disadvantage of
this device is that deflection of the beam can only be measured in the direction
orthogonal to the cantilever direction. The force had to be calculated by making the
assumption that all of the force was oriented in the direction of the long axis of the cell
(i.e. the direction of movement).
44
Figure 2.6 Silicon-based cantilever force sensor for measuring fibroblast traction forces (46).
Figure 2.7 PDMS device for measuring fibroblast contractile forces. The image shows the PDMS pillars and the cell with the force vector for each pillar (47).
45
More recently, microfabricated polymer structures have been developed for
measuring cell forces.47 The device in Figure 2.6 shows an array of PDMS micropillars
that act as vertical cantilevers. When the cell attaches to the pillars and exerts a force, the
pillars are deflected. Using the stiffness of the pillar and the deflection, the force exerted
on each PDMS pillar was calculated. They were able to measure the time dependence of
the force and by micropatterning adhesion molecules on the surface of the pillars, they
showed that the force per pillar is proportional to the amount of cell spreading.
2.4 Conclusion
Cell mechanics plays a critical role in regulating cell function. Mechanical
processes are involved in cell adhesion, movement, signal transduction, and even play a
role regulating cell death. Studying the cytoskeleton both from a biochemical and
mechanical perspective has led to a greater understanding of cell mechanics and
mechanotransduction. A variety of measurement devices and methods have been
developed to investigate these processes. Continued work in this area will lead to a more
complete understanding of cell mechanics and how these factors play a larger role in
maintaining normal tissue function. Another critical area is understanding the role of cell
mechanics in disease and using this information to develop and test drugs and new
therapeutic approaches.
References 1 A.W. Orr, B.P. Helmke, B.R. Blackman, M.A. Schwartz, “Mechanisms of mechanotransduction.” Developmental Cell 10 (2006) 11-20. 2 M.R.K. Mofrad, R.D. Kamm (eds.), Cytoskeletal Mechanics, Cambridge University Press, Cambridge (2006). 46
3 A. Nicolas, B. Geiger, S.A. Safran, “Cell mechanosensitivity control the anisotropy of focal adhesions.” Proceedings of the National Academy of Sciences USA 101 (2004) 12520-12525. 4 M.P. Sheetz, D.P. Felsenfeld, C.G. Galbraith, “Cell migration: regulation of force on extracellular matrix-integrin complexes.” Trends in Cell Biology 8 (1998) 51-54. 5 M.E. Chicurel, C.S. Chen, D.E. Ingber, “Cellular control lies in the balance of forces.” Current Opinions in Cell Biology 10 (1998) 232-239. 6 C.G. Galbraith, M.P. Sheetz, “Forces on adhesive contacts affect cell function.” Current Opinions in Cell Biology 10 (1998) 566-571. 7 N. Wang, J.T. Butler, D.E. Ingber, “Mechanotransduction across the cell surface and through the cytoskeleton.” Science 260 (1993) 1124-1127. 8 C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Geometric control of cell life and death.” Science 276 (1997) 1425-1428. 9 M. Chrzanowska-Wodnicke, K. Burridge, “Rho-stimulated contractility drive the formation of stress fibers and focal adhesions.” Journal of Cell Biology 133 (1996) 1403-1415. 10 D.E. Ingber, “Cellular tensegrity revisited I: Cell structure and hierarchical systems biology.” Journal of Cell Science 116 (2003) 1157-1173. 11 D.E. Ingber, “Tensiegrity II. How structural networks influence cellular information processing networks.” Journal of Cell Science 116 (2003) 1397-1408. 12 G. Bao, S. Suresh, “Cell and molecular mechanics of biological materials.” Nature Materials 2 (2003) 715-725. 13 A.E. Cress, R.B. Nagle (eds.), Cell Adhesion and Cytoskeletal Molecules in Metastasis, Springer, Dordrecht, The Netherlands (2006). 14 H. Yamaguchi, J. Condeelis, “Regulation of the actin cytoskeleton in cancer cell migration and invasion.” Biochimica et Beophysica Acta 1773 (2007) 642-652. 15 F.C.S. Ramaekers, F.T. Bosman, “The cytoskeleton and disease.” Journal of Pathology 204 (2004) 351-354. 16 E.B. Lane, W.H.I. McLean, “Keratins and skin disorders.” Journal of Pathology 204 (2004) 355-366.
47
17 E. Fuchs, “The cytoskeleton and disease: genetic disorder of intermediate filaments.” Annual Review of Genetics 30 (1996) 197-231. 18 K. Zatloukal, C. Stumptner, A. Fuchsbichler, P. Fickert, C. Lackner, M. Trauner, H. Denk, “The keratin cytoskeleton in liver disease.” Journal of Pathology 204 (2004) 367-376.
19 D.W. Owens. E.B. Lane, “Keratin mutations and intestinal pathology.”Journal of Pathology204 (2004) 377-385.
20 Y. Calle, H.-C. Chou, A.J. Thrasher, G.E. Jones. “Wiskott-Aldrich syndrome protein and the cytoskeletal dynamics of dendritic cells.” Journal of Pathology 204 (2004) 460-469.
21 P.B. Schiff, S.B. Horwitz, “Taxol stabilizes microtubules in mouse fibroblast cells.” Proceedings of the National Academy of Sciences USA 77 (1980) 1561-1565. 22 M.V. Blagosklonny, T. Fojo, “Molecular effect of Paclitaxel: myths and reality.” International Journal of Cancer 83 (1999) 151-156. 23 V.K. Ngan, K. Bellman, B.T. Hill, L. Wilson, M.A. Jordan, “Mechanism of mitotic block and inhibition of cell proliferation by the semisynthetic vinca alkaloids vinorelbine and its newer derivative vinflunine.” Molecular Pharmacology 60 (2001) 225-232. 24 G.G. Borisy, E.W. Taylor, “The mechanism of action of colchicines.” The Journal of Cell Biology 34 (1967) 525-533. 25 K.L.K. Duncan, M.D. Duncan, M.C. Ally, E.A. Sausville, “Cucurbitacin E-induced disruption of the actin and vimentin cytoskeleton in prostate carcinoma cells.” Biochemical Pharmacology 52 (1996) 1553-1560. 26 J.A. Cooper, “Effects of cytochalasin and phalloidin on actin.” The Journal of Cell Biology 105 (1987) 1473-1478. 27 M.R. Bubb, A.M.J. Senderowiczf, E.A. Sausville, K.L.K. Duncan, E.D. Kern, “Jasplakinolide, a cytotoxic natural product, induces actin polymerization and competitively inhibits the binding of phalloidin to F-actin.” The Journal of Biological Chemistry 269 (1994) 14869-14871. 28 M.R. Bubb, I. Spector, B.B. Beyer, K.M. Fosen, “Effects of Jasplakinolide on the kinetics of actin polymerization.” The Journal of Biological Chemistry 275 (2000) 5163-5170.
48
29 T. Ohashi, M. Sato, “Remodeling of vascular endothelial cells exposed to fluid shear stress: experimental and numerical approach.” Fluid Dynamics Research 37 (2005) 40-59. 30 O. Traub, B.C. Berk, “Laminar shear stress: mechanicsms by which endothelial cells transduce an atheroprotective force.” Arteriosclerosis, Thrombosis, and Vascular Biology 18 (1998) 677-685. 31 C.R. White, J.A. Frangos, “The shear stress of it all: the cell membrane and mechanochemical tranduction.” Philosophical Transactions of the Royal Society B 362 (2007) 1459-1467. 32 K.J. Van Vliet, G. Bao, S. Suresh, “The biomechanics toolbox: experimental approaches for living cells and biomolecules.” Acta Materialia 51 (2003) 5881-5905. 33 T.D. Brown, “Techniques for mechanical stimulation of cell in vitro: a review.” Journal of Biomechanics 33 (2000) 3-14. 34 T.G. Kuznetsova, M.N. Starodubtseva, N.I. Yegorenkov, S.A. Chizhik, R.I. Zhdanov. “Atomic force microscopy probing of cell elasticity.” Micron 38 (2007) 824-833. 35 C. Rotsch, M. Radmacher, “Drug-induced changes of cytoskeletal structure and mechanics in fibroblasts: an atomic force microscopy study.” Biophysical Journal 78 (2000) 520-535. 36 M.R.K. Mofrad, R.D. Kamm (eds.), Cytoskeletal Mechanics: Models and Measurements, Cambridge University Press, Cambridge (2006). 37 R.M. Hochmuth, “Micropipette aspiration of living cells.” Journal of Biomechanics 33 (2000) 15-22. 38 M. Dao, C.T. Lim, S. Suresh, “Mechanics of the human red blood cell deformed by optical tweezers.” Journal of the Mechanics and Physics of Solids 51 (2003) 2259-2280. 39 A.D. Mehta, M. Rief, J.A. Spudich, D.A. Smith, R.M. Simmons, “Single-molecule biomechanics with optical methods.” Science 283 (1999) 1689-1695. 40 A.K. Harris, P. Wild, D. Stopak, “Silicone rubber substrata: a new wrinkle in the study of cell locomotion.” Science 208 (1980) 177-179. 41 M. Dembo, Y.-L. Wang, “Stresses at the cell-to-substrate interface during locomotion of fibroblasts.” Biophysical Journal 76 (1999) 2307-2316. 42 S. Munevar, Y.-L Wang, M. Dembo, “Traction force microscopy of migrating normal and H-ras transformed 3T3 fibroblasts.” Biophysical Journal 80 (2001), 1744-1757.
49
43 R.J. Pelham, Y.-L. Wang, “Cell locomotion and focal adhesions are regulated by substrate flexibility.” Proceedings of the National Academy of Sciences USA 94 (1997) 13661-13665. 44 C.-M. Lo, H.-B. Wang, M. Dembo, Y.-L. Wang, “Cell movement is guided by the rigidity of the substrate.” Biophysical Journal 79 (2000) 144-152. 45 G. Lin, K.SJ. Pister, K.P. Roos, “Surface micromachined polysilicon heart cell force transducer.” Journal of Microelectromechanical Systems 9 (2000) 9-17. 46 C.B. Galbraith, M.P. Sheetz, “A micromachined device provides a new bend on fibroblast traction forces.” Proceeding of the National Academy of Sciences USA 94 (1997) 9114-9118. 47 J.L. Tan, J. Tien, D.M. Pirone, D.S. Gray, K. Bhadriraju, C.S. Chen, “Cells lying on a bed of microneedles: An approach to isolate mechanical force.” Proceeding of the National Academy of Sciences USA 100 (1993), 1484-1489.
50
CHAPTER 3
FABRICATION OF POLYMER MICROSTRUCTURES BY DOUBLE STAMP MICROMOLDING
3.1 Introduction
Micromolding techniques are now widely used for fabrication of microscale
polymer components. Hot embossing1 and injection molding2 have been scaled down
from their use in macroscale polymer processing to accommodate microscale fabrication.
Several of the soft lithography based techniques also allow fabrication of polymer
microstructures. Of particular interest to this work is the process of microtransfer
molding. Microtransfer molding was introduced by Whitesides et al.3 in 1996. Since then,
the process has been used for fabricating micro and nanoscale devices including
waveguides,4 Schottky diodes,5 and magnetic separation devices.6
The process of microtransfer molding is shown in Figure 3.1. A polymer solution
is applied to the surface of a micropatterned PDMS mold. The mold is then placed in
contact with a substrate and the polymer solution is cured and the mold is removed. This
process provides a simple and cost effective method to replicate microfeatures over large
areas. One limitation of this technique is the residual film that remains between the
polymer features after the mold is removed. This film can be removed by additional
51
processing such as oxygen plasma treatment, but the process requires specialized
equipment and introduces additional processing costs.
Figure 3.1 Schematic diagram of the microtransfer molding (µTM) process (7).
52
The process of double stamp micromolding is a modified microtransfer molding
process that allows fabrication of individual polymer microstructures without the need for
additional processing steps. The process adapted from work by Guan et al.8,9 A solution
of a thermoplastic polymer is spin coated onto the surface of a PDMS mold. An initial
stamping process with low pressure is used to remove the polymer material from the
raised portions of the PDMS mold. A higher pressure stamping step is then used to
remove the material from the recessed portions of the mold. The double stamp
micromolding process is described in detail here, and the process is characterized with
respect to the morphology of the features as a function of the polymer solution
concentration and the transfer pressure.
3.2 Materials and Methods
3.2.1 Double stamp micromolding process
Polystyrene was used as the model polymer is this study, but several other
thermoplastic polymers have been patterned using minor variations of the same basic
process. Other polymers that have been tested include poly(methyl methacrylate)
(PMMA), poly(propyl methacrylate) (PPMA), polycaprolactone (PCL), and poly(lactic-
co-glycolic) acid (PLGA). The process is also flexible with respect to the substrate that
can be used for pattern transfer. Rigid substrates such as silicon and glass as well as
flexible substrates (polyimide and polycarbonate) have been used.
The double stamp micromolding process was performed as described in Figure
3.2. The PDMS mold was spin coated with a polymer solution as shown in Figure 3.2 (b).
The polymer on the surface of the mold is then removed using higher temperature and
lower pressure. In the case of polystyrene, the temperature for the first stamp was 200 ºC
and the pressure was ~2.5 psi. The polymer that remained in the recess features of the
mold was then transferred to the substrate using lower temperature and higher pressure.
The lower temperature of the second stamping was needed to maintain the resolution of
the features and higher pressure was used to ensure that the polymer in the recesses of the
mold made contact with the substrate during the transfer. A transfer temperature of 125
ºC and transfer pressures between 30 and 200 psi were used in the second stamping. The
stamping was performed at varying polymer solution concentrations and pressures to
determine if the coating condition or the transfer conditions have a more significant effect
on the outcome of the feature morphology.
53
Figure 3.2 Schematic diagram of the double stamp micromolding process. (a) uncoated PDMS mold, (b) mold spin coated with polymer, (c) inverted mold in contact with heated
glass to remove surface polymer, (d) polymer removed from the raised portions of the mold, (e) selectively coated mold, (f) inverted mold in brought into contact with the
substrate with heat and pressure, (g) polymer structures after transfer to the substrate.
3.2.2 Process characterization
Two different PDMS mold geometries were used to characterize the process. For
SEM and optical characterization, a 5.3 µm deep mold with 5 µm wide channels and 45
µm spacing between channels was used. The mold was spin coated with 3, 5, 7.5, and
10% (wt/wt) solutions of polystyrene in anisole. The solutions were spin coated onto the
mold at 3000 rpm. Optical micrographs were taken to examine the film prior to
stamping. For relatively low polymer concentrations, dewetting of the polymer is
observed on the mold. The dewetting process and its utility in microfabrication will be
54
discussed in detail in Chapter 4. SEM micrographs were taken of the mold at several
steps in the process. Images were taken of the empty mold prior to spin coating, the mold
after spin coating with a 7.5% solution of PS, the mold after the first stamping to remove
the surface materials, and the substrate after stamping the PS features.
For AFM characterization, a 5.3 µm deep mold could not be used due to the
limited z-range of the AFM. Instead, a 3.3 µm deep mold with 5 µm wide channels and 5
µm spacing was used. AFM scans were run to determine the morphology of the final
feature cross section at a given polymer concentration. The features were characterized
using tapping mode atomic force microscopy (AFM) with a 300 kHz tip, scan length of
25 µm, and scan rate of 0.5 Hz. We define the distortion of the feature according to the
sag, or the depression created in the center of the feature during pattern transfer. The
sagging was characterized using the equation 1:
100max
minmax ×−
=H
HHS (1)
where S is defined as the percent sag, Hmax is the maximum height of the feature cross
section, and Hmin is the minimum height of the cross section.
3.3 Results and discussion
Figure 3.3 shows the dewetting of the PS on the PDMS mold at concentrations of
1%, 3%, 5%, 7.5%, and 10%. The images show that at 1% the polymer dewets and
partially fills the channels in the mold. In addition, a line of polymer forms on the raised
features of the mold. As the concentration increases, the wetting of the channels
increases while the dewetted line on the raised features become wider. At 5% a small
discontinuity exist between the PS in the features and that on the raised portions of the
55
mold. At 7.5% no dewetting occurs (i.e. the polymer film is continuous). Despite the
fact that no dewetting occurs, the double stamping process can still be performed. In fact,
the yield and feature morphology are more favorable at 7.5% than at any other
concentration. The process can also be done with a 10% solution, however, the yield
decreases significantly due to removal of material within the channels during the first
stamping. At even higher polymer concentrations, the thickness of the film is such that
the entire film is removed during the first stamp. This is an effective method for making
continuous films of microstructures.
Figure 3.3 Optical micrographs of polystyrene at (a) 1%, (b) 3%, (c) 5%, (d) 7.5%, and (e) 10% PS concentration (wt/wt) in anisole.
Scanning electron micrographs of several steps in the process are shown in Figure
3.4. The empty PDMS mold is shown in (a). Figure 3.4 (b) shows the PDMS mold after
spin coating with a 7.5% solution of PS. The images show that the film is continuous but
56
there is a topographical difference in the film on the raised portions of the mold versus
the channels. Figure 3.4 (c) shows the mold after the first stamping with material only in
the recessed channels. Finally, Figure 3.4 (d) shows the features after removal from the
mold. The images show that the features are transferred with little distortion of the
original feature morphology.
Figure 3.4 Scanning electron micrographs of four steps in the double stamp micromolding process; (a) uncoated PDMS mold, (b) mold after spin coating, (c) mold after the first stamping with polymer in the recesses of the mold, (d) polymer features
after transfer to the substrate.
Figure 3.5 shows the sagging of the PS features after the second stamping as a
function of the polymer solution concentration. The plot shows an inverse relationship
between the feature sag and the polymer solution concentration. This is likely caused by
a combination of two factors. First, at low concentration, the features are not filled
57
completely. The unoccupied space allows the mold to defect during the transfer process,
leading to distortion of the features. Second, at low polymer concentration the channels
are filled with a more curved profile due to increased retention of material at the wall of
the channel versus in the bulk of the channel. As the concentration increases, these
effects become less prevalent and the channels fill more completely and more uniformly.
Figure 3.6 shows the effect of the transfer pressure on sag of the features for a
7.5% solution spin coated at 3000 rpm. The figure shows that there is a much less
significant effect on the polymer features as the pressure is increased and no trend in the
sag is observed with increased pressure. This indicates that the choice of polymer
solution concentration has a much more significant impact on the feature morphology
than transfer pressure used to remove the features from the mold. All scans were
performed in triplicate and the error bars are ± the standard deviation.
58
0
5
10
15
20
25
30
35
40
45
50
0 2 4 6 8 10
Polymer Concentration (%)
% S
ag
12
3%
5%
7.5%
10%
Figure 3.5 Plot of the feature distortion (sag) with increasing polymer concentration; the insets show cross sectional AFM scans at each of the four polymer concentrations tested.
0
5
10
15
20
25
30
35
40
45
50
0 50 100 150 200 250
Pressure (psi)
% S
ag
Figure 3.6 Plot of the feature distortion (sag) with increasing pressure; a 7.5% PS solution spin coated at 3000 rpm was used to fabricate the structures.
59
3.4 Conclusion
SEM, AFM and optical microscopy were used to characterize the double
stamping of PS lines. The double stamping process can be used due to topographical
differences between the polymer on the raised portions of the mold and polymer in mold
recesses. As polymer concentration is further increased, a lower yield of polymer
features is obtained due to removal of polymer from the channels during the first
stamping process. The results also show that the polymer fills the mold in a predictable
manner with increased filling of the features as the polymer solution concentration is
increased. The transfer pressure used to remove the materials from the mold has much
less effect on the overall feature morphology.
This work shows that the double stamping process can be used to fabricate high
quality polymer microfeatures. Characterization of the process is critical for obtaining
high yield features with quality retention of the original feature morphology.
References 1 H. Becker, U. Heim, Sensors and Actuators A: Physical 83, 130 (2000). 2 L.J. Lee, M.J. Madou, K.W. Koelling, S. Daunert, S. Lai, C.G. Koh, Biomedical Microdevices 3, 339 (2001). 3 X.-M. Zhoa, Y. Xia, and G.M. Whitesides, “Fabrication of three-dimensional micro-structures: microtransfer molding.” Advanced Materials 8 (1996) 837-840. 4 X.-M. Zhao, S.P. Smith, S.J. Waldman, G.M. Whitesides, M. Prentiss, “Demonstration of waveguide couplers fabricated using microtrasnfer molding.” Applied Physics Letters 71 (1997) 1017-1019. 5 J. Hu, T. Deng. R.G. Beck, R.M. Westervelt, G.M. Whitesides, Sensors and Actuators A: Physical 75, 65-69 (1999). 6 T. Deng, M. Prentiss, G.M. Whitesides, Applied Physics Letters 80, 461-463 (2002).
60
7 Y. Xia, G.M. Whitesides, “Soft lithography.” Annual Review of Materials Science 28 (1998) 153-184. 8 J. Guan, A. Chakrapani, D.J. Hansford, “Polymer microparticles fabricated by soft lithography.” Chemistry of Materials 17 (2005) 6227-6229. 9 J. Guan, N. Ferrell, L.J. Lee, D.J. Hansford, “Fabrication of polymeric microparticles for drug delivery by soft lithography.” Biomaterials 27 (2006) 4034-4041.
61
CHAPTER 4
SPIN DEWETTING OF POLYMERS ON POLYDIMETHYLSILOXANE (PDMS) MOLDS
4.1 Introduction
Dewetting is a process of spatial separation of a material caused by physically,
chemically, or thermally induced instability in a thin solid or liquid film. Dewetting is a
phenomenon that has been well documented in thin polymer films.1,2 In most cases, this
is an undesirable effect that results in formation of holes in polymer films. However,
researchers have begun to exploit this process to engineer surfaces that elicit a desired
dewetting behavior for creating regularly structured polymer features. Primarily, this has
been achieved by locally altering the chemical composition of the surface of interest.3-6
Dewetting of thin polymer films on chemically heterogeneous surfaces has been
studied both theoretically and experimentally. The theory of polymer dewetting on
chemically non-uniform surfaces has been described by Lenz et al.7 and Konnur et al.8
Microscale differences in the wettability of the substrate lead to instabilities in the film,
which in turn leads to formation of a pattern. Experimentally, microcontact printing has
been used for inducing specific dewetting behavior. By printing hydrophilic chemistries
on hydrophobic substrates, a hydrophobic polymer such as polystyrene will preferentially
62
coat the hydrophobic regions of the substrate, thus creating a pattern.3,4 A similar concept
applies to printing of hydrophobic chemistries on hydrophilic substrates.5,6
In addition to chemically induced dewetting, surface topography is also known to
be a factor in the dewetting process. The effects of physical, chemical, and
physiochemical heterogeneities have been studied theoretically by Konnur et al.9 Their
model for physical dewetting was based on a sinusoidally rough surface and predicted
polymer dewetting in the depressions of the rough surface. In contrast to chemically
induced dewetting, physically induced dewetting is dominated by dynamic interactions
between the polymer solution and the topographically modified surface.
Bao et al.10 used a combination of chemical patterning and topography for
fabricating micro and nanoscale poly(methyl methacrylate) (PMMA) and polycarbonate
(PC) features. They coated a topographically patterned SiO2 surface with a higher
surface energy silane on the raised surfaces and a lower surface energy silane on the
recessed surfaces. During spin coating, PMMA and PC coated the higher surface energy
areas. The polymer features were then removed from the raised features using heat and
pressure.
Here we study the process of spin dewetting on topographically patterned PDMS
molds. This combination of soft lithography and physically controlled dewetting allows
fabrication of polymer microfeatures in both thin (<100 nm) and thick (>5 micron)
polymer films without the need for chemical surface modifications. By spin coating a
topographically patterned mold with a polymer solution with the appropriate solvent and
polymer content, micro and nanoscale polymer features can be fabricated in a single spin
coating process. The spin coating process provides well-controlled coating conditions for
63
patterning of microscale features over a large pattern area (> 5 cm2). While patterning of
nanoscale features is less reliable over a large area, the same process shows promise for
nanoscale fabrication as well. The lower reliability of replicating nanoscale features is
likely due to inconsistencies in the original e-beam fabricated master and due to a
relatively small original pattern area.
The spin dewetting process allows fabrication of features with either the same or
different geometrical properties as the original mold depending on the polymer
concentration used. Several levels of dewetting can be observed by altering polymer
solution concentration. Complete dewetting takes place at relatively low polymer
concentrations, and refers to the complete lack of polymer solution on the raised features
of the PDMS molds. For complete dewetting, the polymer remains only in the recessed
feature after the spin coating process. Partial dewetting occurs at higher polymer
concentrations. In partial dewetting, the polymer separates during the spin process,
resulting in filling of the recessed portions of the mold in addition to a second, physically
separated polymer pattern that remains on the raised features of the mold. At even higher
polymer concentrations, a continuous polymer film is formed on the PDMS mold.
Spin dewetting allows precise fabrication of polymer structures at the micro and
nanoscale with a single spin process. The dewetting process is initiated purely by surface
topography, thus no chemical surface modifications are needed. This process could be
applied to a number of applications as a simple method for fabricating physically
independent polymer structures at the micro and nanoscale.
64
4.2 Materials and methods
4.2.1 Materials
Poly (n-propyl methacrylate) (PPMA, MW ~150,000) (Scientific Polymer
Products), poly(methyl methacrylate) (PMMA, MW ~75,000) (Scientific Polymer
Products), and polystyrene (PS, melt flow index 4.0) (Aldrich) were used as model
polymers. Polymers were dissolved in anisole (Sigma) at concentrations ranging from
0.125% to 20% (wt/wt). From here on, all solution concentrations are given as weight
percent
4.2.2 PDMS molding
PDMS molds were fabricated using standard soft lithography procedures
described previously. Silicon wafer masters were fabricated from three different
photoresists depending on the desired mold geometry and thickness. The photoresists
used in this study were S1813 positive tone resist (Shipley), SPR 220-7 positive tone
resist (Shipley) and SU8 2005 negative tone resist (Microchem Corp.). Nanoscale
structures were fabricated using electron beam (e-beam) lithography. Silastic T-2
poly(dimethylsiloxane) (PDMS) (Dow Corning) was mixed at a 10:1 ratio with T-2
curing agent and poured over the patterned wafers. The molds were cured at room
temperature for 48 hours before removal from the wafer.
4.2.3 Spin dewetting and pattern transfer
The spin dewetting process is illustrated schematically in Figure 4.1. The PDMS
mold was spin coated with a polymer solution at 3000 or 4000 rpm for 60 seconds. Spin
speed and time were held constant for a given mold geometry to determine the effect of
the pattern geometry and solution concentration on the resultant structures. Depending
65
on the concentration of the polymer solution, periodicity of the pattern, and depth of the
pattern, four potential coating scenarios are possible: (1) complete dewetting with
intermediate separation, (2) complete dewetting with no intermediate separation, (3)
partial dewetting, and (4) no dewetting. Complete dewetting with intermediate
separation between features is shown on the far left of Figure 4.1(a). In this case, the
polymer solution dewets into portions of the recessed PDMS structure, while some
recessed areas of the PDMS mold remain uncoated. Complete dewetting with no
intermediate separation occurs when the polymer solution coats the entire recessed
portion of the mold but does not coat any of the raised portions of the mold. For partial
dewetting, the entire recessed portion of the mold is coated and parts of the raise portion
are coated. In this case, there is a physical separation between the polymer within the
recesses of the mold and the polymer coating the raised portions of the mold. Finally, the
case of no dewetting occurs for higher polymer concentrations when the film completely
coats the recessed and raised portions of the mold. Each of these scenarios is illustrated
in a separate column in Figure 4.1(a).
66
Figure 4.1 Schematic diagram of the spin dewetting and pattern transfer process; (a) four coating scenarios, (b) pattern transfer, and (c) final pattern.
To remove the polymer structures from the mold, the mold is inverted and heat
and pressure are applied to remove the polymer from the mold onto the desired substrate.
This is illustrated in Figure 4.1(b). For this study, polymer micro- and nanostructures
were transferred onto glass or silicon substrates. The substrates were heated to 95 °C for
PPMA and 170 °C for PMMA and polystyrene. Low pressure (~ 2.5 psi) was used to
remove features from the raised portion of the mold, and higher pressure (~ 30 psi for
S1813 molds and 60 psi for SPR and SU8 molds) was used to remove the features from
the recessed portions of the molds.
As shown in Figure 4.1(c), the resulting structures can have significantly different
geometrical properties from the original mold. This allows features of multiple
geometries to be fabricated from a single PDMS mold by varying polymer
concentrations.
67
4.2.4 Process characterization
Resulting micro and nanostructures were characterized using scanning electron
microscopy (SEM) (Hitachi S-3000H), atomic force microscopy (AFM) (Veeco
Dimension 3100), and optical microscopy.
4.3 Results and discussion
Three dimensional AFM images of polystyrene microstructures fabricated using
spin dewetting on 2 µm wide, 1.4 µm tall pillars are shown in Figure 4.2. The spacing
between features is also 2 µm. Polymers solutions were spin coated at 3000 rpm for 60
seconds. The images show the evolution of the feature geometry as the polymer solution
concentration is increased. Figure 4.2(a) shows complete dewetting of a 0.125 wt.-%
polymer solution around the pillars with intermediate separation between the features.
The resulting pattern consists of an array of physically separated polymer rings with 2 µm
holes. At a solution concentration of 1%, the same basic structure is again observed, but
the rings are interconnected with each of the adjacent rings. As the solution
concentration is increased further, the wetting between the features increases until a
continuous film is produced at a solution concentration of 3% [Figure 4.2(c)]. The
pattern is a mesh with 2 µm through holes. The holes were verified by SEM (data not
shown). Between solution concentrations of 3% and 4%, there was a transition from
complete dewetting to no dewetting, and the polymer coated the entire surface of the
PDMS mold. A solid film with 2 µm wells made from a 10 % solution is shown in
Figure 4.2(d).
68
Figure 4.2 Three dimensional AFM images and line scans showing the morphological evolution of polystyrene film after removal from the PDMS mold; all scans are 20x20 µm; (a) .125 wt.-%, z-scale is ±100nm, feature height is ~86 nm, (b) 1 wt.-%, z-scale is ±100nm, feature height is ~91 nm, (c) 3 wt.-%, z-scale is ±500nm, feature height is ~290
nm, (d) 10 wt.-%, z-scale is ±1 µm, feature height is ~800 nm.
Dewetting behavior was also studied on a PDMS mold with deeper features. In
this case, 7.5 µm tall circular features fabricated from SPR220-7 were employed.
Solutions of PPMA ranging from 1-20 wt.-% were spin coated at 3000 rpm for 60
69
seconds. The coated molds were then characterized using optical microscopy. Evolution
of the film morphology varied slightly from the 2 µm pillars due to the differences in
both the thickness and lateral dimensions of the features. Figure 3(a) shows an uncoated
PDMS mold. Figure 4.3(b-f) shows optical micrographs of spin coated PDMS molds at
increasing polymer concentrations. The images show that for a low polymer
concentration (1%), the polymer only coated the perimeter of the 20 µm pillar and at 3%
the polymer coated the perimeter of the pillar and a portion of the area in between the
pillars. This behavior is similar to that observed on the 2 µm pillars at low polymer
concentrations. Partial dewetting was observed at concentrations of 5-15%. Figure 4.3
shows images of the coated mold at (d) 5% and (e) 15% polymer concentrations. The
recessed features of the mold were coated with a continuous film while particles formed
on the raised features of the mold.
70
Figure 4.3 Optical micrographs of PPMA dewetting on a PDMS mold with 20 µm diameter, 7.5 µm tall pillars; (a) Uncoated PDMS mold (b-f) Polymer solution
concentrations are (b) 1 wt.-% (c) 3 wt.-% (d) 5 wt.-% (e) 15% wt.-% (f) 20 wt.-%.
The diameter of the particles formed on the mold can be well controlled by
varying the polymer concentration or the lateral dimension of the features on the mold.
Figure 4.4(a) show that the particle diameter increased linearly from 2.4-3.6 µm with
increased solution concentration. Figure 4.4(b) shows the influence of the lateral
dimension of the mold features on the particle diameter. In this case, hexagonal features
with lateral dimensions between 5 µm and 20 µm were used. The particle size increased
linearly from 2.6-5.4 µm depending on the size of the hexagonal feature that was coated.
The solution concentration is held constant at 15%. The ability to form monodisperse
microparticles using spin dewetting could provide a simple method for production of
microparticles for application such as drug delivery.11,12
71
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 2 4 6 8 10 12 14 16Polymer Solution Concentration (%)
Parti
cle
Size
(µm
)
.
(a)
0123456
5 10 15 20 2Hexagon Size (µm)
Parti
cle
Size
(µm
) .
(b)
5
Figure 4.4 Particle size characterization for (a) varying PPMA solution concentrations on 20 µm diameter PDMS pillars and (b) varying hexagonal pillar sizes with a PMMA
solution concentration of 15%.
Figure 4.5 shows SEM micrographs of each step of the coating and removal
process for a partially dewetted PPMA film on 20 µm PDMS features. Figure 4.5(a) is
the original PDMS mold. Figure 4.5(b) shows the mold after spin coating. The image
shows that the recessed area of the mold is almost completely filled with polymer and
physically separate polymer particles form on top of the pillars. Figure 4.5(c) and (d)
72
show the polymer particles that were removed from the top of the pillars using heat and
low pressure and the mesh in between the features removed using heat and higher
pressure.
Figure 4.5 SEM micrographs of each step in the spin coating and stamping process for partially dewetted features; (a) uncoated PDMS mold, (b) mold after spin coating, (c)
substrate with stamped particles, (d) substrate with stamped mesh.
To this point, all of the experiments have described dewetting behavior on pillar
structures. In addition to pillars, dewetting can be initiated in well structures. Figure
4.6(a) shows an SEM micrograph of 3% PMMA spin coated on 10µm diameter circular
wells. The polymer completely dewets and coats only the perimeter of the wells. Figure
4.6(b) shows the polymer rings removed from the wells using the stamping process
described previously.
73
Figure 4.6 (a) PDMS mold after spin coating with 3% PMMA, (b) PMMA rings after removal from the mold.
Figure 4.7 SEM micrographs of nanoscale features; (a) no intermediate dewetting between features, (b) intermediate dewetting between features. P represents the primary
features from the mold recesses and I indicates the intermediate dewetted features.
The dewetting process was also demonstrated with nanoscale features. A pattern
was first created using e-beam lithography and the pattern was transferred to PDMS. The
74
PDMS mold was then spin coated at 4000 rpm with a 1% polystyrene solution.
Experiments indicate that in addition to the depth of the pattern, the lateral dimensions
and the spacing of the pattern are also important parameters governing dewetting
behavior. Figure 4.7 shows SEM micrographs of polystyrene features after removal from
the mold. For features with smaller lateral dimensions and a smaller center-to-center
distance between features, complete dewetting was observed. For larger features with
larger spacing, only partial dewetting was observed. The features in Figure 7(a and b) are
approximately 170 nm separated by 300 nm and 250 nm separated by 500 nm,
respectively. Note that no residual polymer is observed in (a) and small polystyrene
strips approximately 50 nm wide are located between the features in (b) for the same
polymer concentration.
These results indicate that the original PDMS mold geometry (height, lateral
feature dimensions, and spacing) and polymer solution concentration are all important
factors in the formation of polymer features by spin dewetting. These features can have
significantly different geometries from the original mold. This allows several different
types of polymer features to be fabricated from a single PDMS mold. By properly
engineering the mold surface, complicated polymer structures can be formed and tailored
to a given application.
4.4 Conclusion
75
The process of spin dewetting has potential as a versatile and simple method of
fabricating polymer micro and nanostructures. Three different mold geometries were
studied here. They can be categorized as thin film, thick film, and nanoscale features.
Polystyrene, PMMA, and PPMA were used as model polymers, but the process could
potentially be applied to various soluble thermoplastic materials with minor process
modifications. The process was demonstrated on both pillar and well structures.
In this study, the effect of polymer solution concentration and mold feature
geometry on final polymer microstructure was examined. It was found that dewetting
occurs on several different types of patterns at differing polymer concentrations. For a
given mold geometry, the dewetting transitions can be controlled by varying the
concentration of the polymer solution. This results in polymer structures that can have
geometries that are significantly different than the original mold. The process was
applied to fabrication of microscale polymeric patterns and shows potential for
application in polymer nanofabrication.
References 1 R. Günter, “Dewetting of thin polymer films.” Physical Review Letters 68 (1992) 75-78. 2 R. Xie, A. Karim, J.F. Douglas, C.C. Han, R.A. Weiss, “Spinodal dewetting of thin polymer films.” Physical Review Letters 81 (1998) 1251-1254. 3 A. Sehgal, V. Ferreiro, J.F. Douglas, E.J. Amis, A. Karim, “Pattern-directed dewetting of ultrathin polymer films.” Langmuir 18 (2002) 7041-7048. 4 E. Meyer, H. Braun, “Controlled dewetting processes on microstructured surfaces- a new procedure for thin film microstructuring.” Macromolecular Materials and Engineering 276/277 (2000) 44-50. 5 Z. Zhang, Z. Wang, R. Xing, Y. Han, “How to form regular polymer microstructures by surface-patterned-directed dewetting.” Surface Science 539 (2003) 129-136. 6 Z. Zhang, Z. Wang, R. Xing, Y. Han, “Patterning thin polymer films by surface-directed dewetting and pattern transfer.” Polymer 44 (2003) 3737-3743. 7 P. Lenz, R. Lipowsky, “Morphological transitions of wetting layers on structured surfaces.” Physical Review Letters 80 (1998) 1920-1923.
76
8 R. Konnur, K. Kargupta, A. Sharma, “Instability and morphology of thin liquid films on chemically heterogeneous substrates.” Physical Review Letters 84 (2000) 931-934.
9 K. Kargupta, A. Sharma, “Dewetting of thin films on periodic physically and chemically patterned surfaces.” Langmuir 18 (2002) 1893-1903. 10 L.-R. Bao, L. Tan, X.D. Huang, Y.P. Kong, L.J. Guo, S.W. Pang, A.F. Yee, “Polymer inking as a micro- and nanopatterning technique.” Journal of Vacuum Science and Technology B 21 (2003) 2749-2754. 11 J. Guan, A. Chakrapani, D.J. Hansford, “Polymer microparticles fabricated by soft lithography.” Chemistry of Materials 17 (2005) 6227-6229. 12 J. Guan, N. Ferrell, L.J. Lee, D.J. Hansford, “Fabrication of polymeric microparticles for drug delivery by soft lithography.” Biomaterials 27 (2006) 4034-4041.
77
CHAPTER 5
LIFT-OFF PROCESSING FOR FABRICATING MICROPATTERNED SULFONATED POLYANILINE
5.1 Introduction
Traditional polymeric materials generally have the property of being poor
electrical conductors. However, several unique polymer chemistries have been developed
to act as organic electrical conductors or conducting polymers. Due to their unique
properties, conducting polymers have been considered as potential functional materials in
several applications including biosensors,1,2 light emitting devices,3,4 organic
transistors,5,6 and actuators.7,8
Sulfonated polyaniline (SPAN) is of particular interest due to the advantages of
being self-doped, environmentally stable, and conductive over a wide pH range.9 While
sulfonated polyaniline exhibits considerably higher solubility in aqueous solution than
native polyaniline,10 solution based processing can still be problematic in certain
applications. In situ polymerization of polyaniline on surfaces has been a common route
for formation of polyaniline films on flat surfaces.11,12 This method eliminates the need to
dissolve the polymer after synthesis. In addition to flat films, in situ polymerization may
provide a convenient method for creating micropatterned conducting polymer films.
78
The ability to pattern conducting polymers on the micro and nanoscale is
particularly important for many applications. A number of methods have been introduced
for patterning conducting polymers.13 By carefully tailoring the conducting polymer
chemistry to introduce photosensitivity, photolithography has been used for
micropatterning conducting polymers.14,15 The use of photolithography can be used to
either locally change the conductivity of the polymer or change the solubility of the
polymer in a specific solvent. While photolithographic processing of conducting
polymers has been widely studied, these processes require expensive processing
equipment and fabrication facilities.
In order in reduce the cost and complexity of patterning, soft lithographic
techniques have been explored as alternative fabrication routes. Micromolding in
capillaries (MIMIC)16-18 and microcontact printing18 are two soft lithography based
methods that have been used for the fabrication of conducting polymer microstructures.
The method described here combines double stamp micromolding with in situ
polymerization in a lift-off technique for microfabrication of SPAN microstructures on
both rigid and flexible surfaces. Molding is used to pattern a sacrificial layer from a
thermoplastic polymer. SPAN is then polymerized on the surface and the sacrificial layer
is dissolved in organic solvent, selectively removing SPAN from the surface. Atomic
force microscopy (AFM) was used to measure the film thickness and RMS roughness as
a function of the reaction time. Scanning electron microscopy (SEM) was used to image
the substrates at various steps in the microfabrication process. This method provides an
efficient, low-cost, non-cleanroom method for micropatterning of sulfonated polyaniline
79
on a variety of surfaces. This process may have applications ranging from conducting
polymer electronics and sensors to biomedical actuators.
5.2 Materials and methods
5.2.1 SPAN synthesis
Sulfonated polyaniline was synthesized in a manner similar to that described
previously.19 Analine [AN (Aldrich)] and 3-aminobenzenesulfonic acid [metalinic acid
(MA) (Fluka)] were copolymerized in 1 M HCl in an ice bath at 5-15°C. Ammonium
persulfate [APS (Aldrich)] was used as the oxidant. The molar ratio of AN:MA was 1:1
and the molar ratio of oxidant to each monomer was 1:1. Substrates were first attached to
the sidewalls of the reaction vessel. MA was dissolved in HCl and AN was added
dropwise to the reaction vessel. APS then dissolved separately in DI water. The reaction
beaker was placed in an ice bath and the APS solution was added. The reaction was
performed under both static and dynamic conditions. Under static conditions, the reaction
components were mixed thoroughly and the reaction was allowed to proceed without
agitation. Under dynamic conditions, a magnetic stir plate and stir bar were used to
constantly agitate the solution during the reaction.
5.2.2 Micromolding and Lift-off processing
To begin the process, a photoresist master was first fabricated using standard
photolithographic techniques, and the negative of the photoresist pattern was transferred
into PDMS using the molding process described previously. Micropatterning of
sulfonated polyaniline was then performed by patterning a sacrificial layer using double
80
stamp micromolding. In situ polymerization on the micropatterned substrate was
performed, followed by solvent lift-off.
Figure 5.1 shows the process for patterning SPAN. The PDMS mold was first
spin coated with a layer of poly(n-propyl methacrylate) (PPMA) (Scientific Polymer
Products). PPMA was dissolved in anisole (Aldrich) at various concentrations depending
on the geometry of the PDMS mold features. For this study, three different mold
geometries were used: a 1.4 µm deep mold with 2 µm wide channels, a 4.8 µm deep mold
with a minimum feature size of 5 µm, and a 7.5 µm deep mold with 20 µm circular
pillars. The PPMA solution and spin condition for each mold were as follows: 5% PPMA
spin coated at 3000 rpm, 10% at 2000 rpm, and 15% at 3000 rpm, respectively.
Figure 5.1 Schematic diagram of sacrificial layer patterning, in situ polymerization, and lift-off; (a) uncoated PDMS mold, (b) PDMS mold spin coated with PPMA, (c) mold is
inverted and applied to heated glass to remove surface polymer, (d) PPMA removed from the raised surface of the PDMS mold, (e) PPMA in recessed features of the PDMS mold,
(f) mold inverted and heat and pressure applied to remove polymer from recessed features, (g) PPMA features after removal of the mold (h) SPAN polymerized directly onto the surface, (h) PPMA removed in acetone leaving the patterned SPAN features.
81
After spin coating [Figure 5.1(b)], the polymer on the raised surfaces of the mold
was removed by bringing the mold into contact with a hotplate at 180 ºC and applying
light pressure (~ 2.5 psi) [Figure 5.1(c)]. Figure 5.1(d) shows the material that was
removed from the raised portion of the PDMS mold. The mold was then applied to the
substrate at 95 ºC and pressure was applied to the top of the mold and held for 3-5
seconds. The pressure used for the 1.4 µm, 4.8 µm, and 7.5 µm deep molds were 35, 40,
and 50 psi, respectively. This transferred the PPMA from the recessed portions of the
mold onto the substrate. Three different substrate materials were used in this study: glass,
silicon, and polyimide.
The patterned PPMA layer acted as a sacrificial layer for patterning the SPAN.
After patterning the PPMA, the SPAN was deposited in situ on the substrates for times
ranging from 15 minutes to 2 hours. The substrates were then removed from the reaction
and the PPMA layer was removed by sonication in acetone for 10 seconds followed by
drying in filtered air.
5.2.3 Process characterization
82
Polymer films were characterized using atomic force microscopy (AFM) (Veeco
Dimension 3100) and scanning electron microscopy (SEM) (Hitachi S-3000H). Film
thickness and roughness were measured using tapping mode AFM. The film thickness
was measured by using the AFM stepheight function. The RMS roughness was measured
over a 5 µm x 5 µm scan area at 1.5 Hz and 512 X 512 pixel resolution. SPAN films
deposited on clean silicon substrates were used for thickness and roughness
characterization. SEM was used to visualize the substrates at various steps in the
fabrication process.
5.3 Results and discussion
Figure 5.2 shows SEM micrographs of patterned PPMA sacrificial layers (a,b),
sacrificial layers after in situ polymerization of SPAN (c,d), and SPAN patterns after
sacrificial layer lift-off (e,f). Figure 5.2(c) shows a low magnification image to
demonstrate that the process can be used for fabrication of features over a relatively large
area. The minimum line width for the pattern shown is nominally 5 µm.
Figure 5.2 SEM micrographs of; (a,b) patterned PPMA sacrificial layers, (c,d) PPMA layer with SPAN deposited over the entire surface, and (e,f) SPAN micropatterns after
lift-off of the sacrificial layer.
83
SPAN film thickness and roughness were characterized throughout the process in
order to optimize the processing conditions to obtain smooth films with well-controlled
thicknesses. The SPAN film thickness and RMS roughness versus reaction time are
shown in Figure 5.3 and Figure 5.4, respectively. The parameters were measured for both
static and dynamic deposition processes to determine the effect of reaction conditions on
the film properties. For a static deposition process, the film thickness increased to a
steady state thickness of approximately 100 nm after 60 minutes of deposition time.
Between 60 and 120 minutes, the film thickness remained relatively constant. Roughness
also remained fairly constant between 45 and 90 minutes, but a large increase in the
average roughness and standard deviation was observed between 90 and 120 minutes.
This is likely due to the formation of larger aggregates of SPAN on the surface. This
phenomenon was confirmed qualitatively by SEM observation.
84
0
20
40
60
80
100
120
140
160
180
0 20 40 60 80 100 120 140
Reaction Time (min.)
Film
Thi
ckne
ss (n
m)
Static Dynamic
Figure 5.3 SPAN film thickness versus reaction time for static and dynamic reaction processes.
0
20
40
60
80
100
120
140
0 20 40 60 80 100 120
Reaction Time (min.)
RM
S R
ough
ness
(nm
)
140
Static Deposition Dynamic Deposition
Figure 5.4 RMS roughness of SPAN films versus reaction time for static and dynamic deposition.
85
Dynamic deposition significantly altered the deposition process, particularly with
respect to the film thickness. For dynamic deposition, the film thickness was significantly
higher than static deposition for reaction times greater than 60 minutes. A possible
explanation for the difference is local depletion of the aniline monomers during static
deposition. The higher flux of SPAN to the substrate surface could lead to a larger steady
state film thickness. Dynamic deposition did not have a significant effect on the
roughness of the films. The same basic trend in the roughness was observed for both
deposition processes. An increase in the standard deviation of the roughness at 120
minutes again indicates that larger aggregates were beginning to form on the surface.
SEM images of 20 µm circular patterns deposited using dynamic deposition are
shown in Figure 5.5. Figure 5.5(a) shows a low magnification micrograph of independent
features to demonstrate patterning over a large area and a higher magnification image of
the pattern is shown in (b).
Figure 5.5 SEM micrographs of discontinuous SPAN patterns at (a) high and (b) low magnifications.
86
5.4 Conclusion
The process described here combines double stamp micromolding with in situ
polymerization for fabrication of microscale patterns of sulfonated polyaniline. The
fabrication method provides a cost effective alternative to conventional cleanroom
fabrication processes and allows fabrication on a variety of both rigid and flexible
substrates. In situ polymerization eliminates the need for post process dissolution of the
polymer and allows polymerization and deposition in a single step. The lift-off process is
analogous to conventional lift-off used for fabrication of metal electrodes. The
insolubility of SPAN in organic solvent allows lift-off process to be employed as a simple
micropatterning technique. The process is capable of producing both continuous (lines)
and discontinuous (circles) patterns. This process has potential for simple fabrication of
SPAN microstructures over a large area for applications in a variety of conducting
polymer applications.
References 1 M. Gerard, A. Chaubey, B.D. Malhotra, “Application of conducting polymers to biosensors.” Biosensors and Bioelectronics 17 (2002) 345-359. 2 S. Geetha, C.R.K. Rao, M. Vijayan, D.C. Trivedi, “Biosensing and drug delivery by polypyrrole.” Analytica Chimica Acta 568 (2006) 119-125. 3 L. Dai, B. Winkler, L. Dong, L. Tong, A.W.H. Mau, “Conjugated polymers for light-emitting applications.” Advanced Materials 13 (2001) 915-925. 4 D. Braun, “Semiconducting polymer LEDs.” Materials Today 5(6) (2002) 32-39. 5 C.J. Drury, C.M.J. Mutsaers, C.M. Hart, M. Matters, D.M. de Leeuw, “Low-cost all-polymer integrated circuits.” Applied Physics Letters 73 (1998) 108-110. 6 N. Stutzmann, R.H. Friend, H. Sirringhaus, “Self-aligned, vertical-channel, polymer field-effect transistors.” Science 299 (2003) 1881-1884.
87
7 E.W.H. Jager, E. Smela, O. Inganäs, “Microfabricated conjugated polymer actuators.” Science 290 (2000) 1540-1545. 8E. Smela, “Microfabrication of PPy microactuators and other conjugated polymer devices.” Journal of Micromechanics and Microengineering 9 (1999) 1-18. 9X.L. Wei, Y.Z. Wang, S.M. Long, C. Bobeczko, A.J. Epstein, “Synthesis and physical properties of highly sulfonated polyaniline.” Journal of the American Chemical Society 118 (1996) 2545-2555. 10J. Yue, Z.H. Wang, K.R. Cromack, A.J. Epstein A.G. MacDiarmid, “Effect of sulfonic acid group on polyaniline backbone.” Journal of the American Chemical Society. 113 (1991) 2665-2671. 11J. Stejskal, I. Sapurina, J. Prokeš, J. Zemek, “In-situ polymerized polyaniline films.” Synthetic Metals 105 (1999) 195-202. 12 I. Sapurina, A. Riede, J. Stejskal, “In-situ polymerized polyaniline films 3. Film formation.” Synthetic Metals 123 (2001) 503-507. 13 S. Holdcroft, “Patterning π-conjugated polymers.” Advanced Materials 13 (2001) 1753-1765. 14 D.G. Lidzey, M.A. Pate, M.S. Weaver, T.A. Fisher, D.D.C. Bradley, “Photoprocessed and micropatterned conjugated polymer LEDs.” Synthetic Metals 82 (1996) 141-148. 15 F.J. Touwslager, N.P. Willard, D.M. de Leeuw, “I-line lithography of poly-(3,4-ethylenedioxythiophene) electrodes and application in all-polymer integrated circuits.” Applied Physics Letters 81 (2002) 4556-4558. 16 W.S. Beh, I.T. Kim, D. Qin, Y. Xia, G.M. Whitesides, “Formation of patterned microstructures of conducting polymers by soft lithography, and applications in microelectronic device fabrication.” Advanced Materials 11 (1999) 1038-1041. 17 J.A. Rogers, Z. Boa, V.R. Raju, “Nonphotolithographic fabrication of organic transistors with micron feature sizes.” Applied Physics Letters 72 (1998) 2716-2718. 18 T. Granlund, T. Nyberg, L.S. Roman, M. Svensson, O. Inganäs, “Patterning of polymer light-emitting diodes by soft lithography.” Advanced Materials 12 (2000) 269-273. 19 I. Mav, M. Žigon, A. Šebenik, “Sulfonated polyaniline.” Synthetic Metals 101 (1999) 717-718.
88
20 I. Mav, M. Žigon, A. Šebenik, J. Vohlidal, “Sulfonated polyanilines prepared by copolymerization of 3-aminobenzenesulfonic acid and aniline: the effect of reaction conditions on polymer properties.” Journal of Polymer Science A 38 (2000) 3390-3398.
89
CHAPTER 6
MICROFABRICATED MEMBRANES FOR CELL ISOLATION
6.1 Introduction
The ability to control the location and morphology of individual cells is extremely
important in several cell biology applications. Isolation of individual cells is critical for
biological experiments such as single cell electroporation1,2 and planar patch clamping.3,4
Positioning of cells is also important in cell-based biosensors,5-7 where cells often need to
be located on electrodes in order for measurements to be carried out. In addition to cell
placement, it is also well established that cell morphology plays a critical role in a variety
of cellular processes including proliferation, differentiation, apoptosis, and motility.8-11
Engineering specific cell behaviors by controlling the cellular environment can be
particularly important for tissue engineering applications.11
Cells often need to be localized in colonies or clusters in order to function
properly. Two examples include liver cells and embryonic stem cells. Isolated
hepatocytes are known to quickly lose liver specific function during in vitro culture. Cell
aggregates, often in the form of spheroids, are needed in order to maintain proper
function.12 Not only the spheroid shape, but also the size of the cluster is important for
maintaining cell viability. This is primarily due to nutrient transport issue at the core of
the cell spheroid.13 Embryonic stem cells are also known to aggregate into spheroids
90
called embryoid bodies.14 The ability to control spheroid size and the number of cells per
spheroid could have significant implications in studying hepatocyte and stem cell biology
and in tissue engineering.
Microfabrication is an attractive approach for interacting with individual cells
because of the ability to fabricate structures and devices on the same size scale as cells.
Microcontact printing (µCP)15 has been widely used to locally alter surface chemistry for
cell patterning. µCP is a soft lithography process that uses a poly (dimethylsiloxane)
(PDMS) stamp “inked” with various chemistries including self-assembled monolayers
(SAM)16,17. The stamp is then brought into contact with the surface of interest to transfer
the chemical pattern. The spatial difference in the surface chemistry alters subsequent
protein adsorption, which leads to preferential cell attachment. Cell adhesion proteins
such as fibronectin have also been directly patterned on surfaces to alter cell
attachment.6,18 In addition to alteration of surface chemistry, a combination of surface
topography and chemical modifications have been used to pattern adherent cells.19-20
Several other devices have also been fabricated for isolating individual or groups
of cells.21-24 Elastomeric stencils have been used to pattern groups of adherent cells.22
Cells were patterned by culturing them on a PDMS membrane with through holes, and
removing the membrane after cell attachment. Other passive systems have been used for
cell isolation. Rosenbluth et al.23 used microfabricated SU8 wells for AFM experiments
on non-adherent cells. Another passive device was developed by Fukuda and
Nakazawa.24 They used microfabricated wells of various diameters to create hepatocyte
spheroids of different sizes. Both of these systems rely on passive loading of the cells via
sedimentation.
91
Here we present a simple method for active isolation of adherent and non-
adherent cells in microwells with well defined geometries. The device consists of a
micropatterned membrane fabricated on the surface of a commercially available porous
filter. For clarity, it should be noted that from here on the microfabricated structure will
be referred to as the “membrane” and the commercially available portion will be called
the “filter”. The membranes were fabricated from thermoplastic polymers using a spin
dewetting process and transfer molding as described previously.25 Several membrane
geometries were used including circular, square, and hexagonal patterns to control cell
morphology. Individual cells or groups of cells could be isolated within each well by
controlling the seeding density and surface area of the pattern. Cells were isolated by
placing the device in a filtration setup and using vacuum to pull the cells into the wells.
The devices were tested using three different types of cells. NIH 3T3 fibroblasts
cells were isolated as either individual cells or multiple cells by altering the seeding
density. These cells were also used to demonstrate the ability to control cell morphology
by changing the geometrical properties of the microwells. THP-1 lymphocytes were used
to show the ability to isolate non-adherent cells. Finally, C3A hepatocytes were isolated
as well-defined clusters, with the size of the cluster being dependant on the area (i.e.
diameter) of the microwell.
The device has several advantages compared with current cell isolation
technology: no change in surface chemistry is needed to spatially orient the cells; the
device can be used for isolation of adherent or non-adherent cells; and by changing the
area of the wells, cell clusters with controlled sizes can by formed. This device could find
92
applications in cell biology studies of individual cells and cell aggregates, cell-based
biosensor arrays, and tissue engineering.
6.2 Materials and methods
6.2.1 Materials
Membranes were fabricated from either poly(propyl methacrylate) (PPMA,
Scientific Polymer Products) or polystyrene (Aldrich). Polymer solutions were made by
dissolving the solid polymers in anisole (Aldrich) at various concentrations depending on
the desired membrane thickness. Several commercially available filters were used. Track
etched polycarbonate filters (Isopore®, Millipore) with 200 nm and 400 nm pores were
used due to their flexibility and relatively good temperature stability. Anodized alumina
filters (Anodisc®, Whatman) with 20 nm pores were used. However, cracking of the
Anodisc filters during the molding process resulted in low yields. Transparent track
etched polystyrene filters (Transwell®, Corning) were also used. These filters provided
the advantage of being optically transparent, which allows observation using an inverted
microscope. However, they can only be used with PPMA membranes due to temperature
stability issues during fabrication.
6.2.2 Membrane fabrication
93
Membranes were fabricated using a two-step soft lithographic micromolding
process combined with spin dewetting. The geometry of the membrane pores was first
defined in a photoresist using standard photolithography. Negative tone SU8 25
(Microchem Corp.) and positive tone SPR 220-7 (Shipley) photoresists were used in the
fabrication. Photoresists were spin coated on clean silicon at 3000 rpm and 2000 rpm for
the negative and positive photoresists, respectively. The resulting film thickness was 9.2
µm for SU8 and 7.5 µm for SPR 220-7. The photoresist films were exposed to ultraviolet
light through chrome/glass photomasks with various geometries, and films were
processed according to the manufacturer’s suggested parameters.
The photolithographically patterned silicon wafers were used as a master for
replication of a PDMS mold. The mold was made by first mixing a 10:1 (wt/wt) ratio of
PDMS base with curing agent (Silastic T-2, Dow Corning). The base and curing agent
were mixed thoroughly and poured over the patterned wafer. The PDMS was degassed in
a vacuum dessicator until all bubbles were removed. The PDMS was allowed to cure at
room temperature for 48 hours before removing the mold.
Figure 6.1 shows the remainder of the membrane fabrication process after making
the mold. The mold was spin coated with a solution of PPMA or polystyrene in anisole.
Solutions of 15% (wt/wt) were used on the 7.5 µm deep molds, and 20% (wt/wt)
solutions were used on 9.2 µm molds. Solutions were spin coated at 3000 rpm for 1
minute. After spin coating, dewetting of the polymer was observed as the polymer
separated to form particles on top of the pillar structures of the PDMS molds. The
particles were removed by contacting the polymer with a glass slide at 180ºC (PPMA) or
200ºC (PS) using low pressure (~2.5 psi). This is shown in Figure 6.1(c and d). After
removal of the surface material, the mold was brought into conformal contact with the
filter. The filter was then place on a hotplate at 95ºC (PPMA) or 125ºC (PS) and pressure
of 60 psi was used to transfer the remaining polymer onto the filter. The resulting
structure consisted of through holes in the membrane with access to the underlying filter.
94
Figure 6.1 Schematic diagram of the membrane fabrication process; (a) uncoated PDMS mold, (b) mold spin coated with polymer (particles on the top of the pillars form by spin dewetting), (c,d) the coated mold is brought into contact with heated glass to remove the particles, (e) selectively coated mold, (f) the mold is inverted and brought into contact with the heated filter by applying pressure, (g) the final structure after removing the
mold.
6.2.3 Cell culture and filtration
NIH 3T3 mouse fibroblasts were cultured in Dulbecco’s modified Eagle’s
medium (DMEM) supplemented with 10% calf bovine serum and 1% penicillin-
streptomycin. THP-1 monocytes were cultured in RPMI 1640 medium supplemented
with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. C3A cells were
cultured in Eagles minimum essential medium (EMEM) supplemented with 10% FBS
and 1% penicillin-streptomycin. All cell lines and culture chemicals were obtained from
95
the American Type Culture Collection (ATCC). Cells were incubated at 37ºC in a
humidified atmosphere with 5% CO2.
Figure 6.2 Experimental setup for cell isolation.
Cells were isolated in the wells using a glass microanalysis vacuum filter setup
(Fisher Scientific). The system is shown in Figure 6.2. The membrane/filter was clamped
between a glass funnel and the underlying vacuum filter flask (Fisher Scientific). A cell
suspension in culture medium, with the desired concentration, was then placed in the
funnel on top of the membrane/filter, and vacuum was applied to pull the cell culture
medium through the setup. This trapped the cells in the microfabricated wells of the
membrane. Additional medium was added if required and the device was placed in an
96
incubator for 30 minutes to allow cell attachment. For non-adherent cells, the membrane
was immediately removed from the setup and cells were prepared for examination.
6.2.4 Characterization
Microfabricated molds and devices were characterized using scanning electron
microscopy (SEM, Hitachi S3000H). Cell isolation was characterized using SEM and
fluorescence microscopy (Nikon TS 100). Cells were prepared for SEM by dehydration
in graded ethanol solutions (70%, 80%, 90%, and 100%) and hexamethyldisilazane
according to the procedure described by Braet et al.26 For fluorescence microscopy, cells
were fixed in 70% ethanol for 30 minutes at –20 ºC. Cells were then stained with
propidium iodide RNase (PI RNase, BD Biosciences) for 10 minutes at –4 ºC. Samples
were washed three times in phosphate buffered saline (PBS, ATCC) and placed on glass
slides for observation.
6.3 Results and discussion
The four membrane geometries used for NIH 3T3 and THP-1 cell isolation are
shown in Figure 6.3. The features included 20 µm diameter circular features, 20 µm
hexagons (edge to edge), 30 µm squares, and 10 µm diameter circles. All of the images
were taken using the Isopore® polycarbonate filters as the substrate. Figure 6.3(d) clearly
shows the track etched pores in the underlying filter. It is critical that access to the pores
is maintained throughout the fabrication process in order to provide an open pathway
through the entire device to facilitate vacuum filtration of the cells.
97
Figure 6.3 SEM micrographs of cell isolation membranes. Feature geometry and size are (a) 20 µm diameter circles. (b) 20 µm (side to side) hexagons. (c) 30 µm squares. And (d)
10 µm diameter circles.
Figure 6.4 presents SEM micrographs of NIH 3T3 cells isolated in the devices.
Figure 4(a) shows one NIH 3T3 cell in each of four 20 µm circular wells. Figure 4(c)
shows cells isolated in two hexagonal wells with two wells empty. Figure 4(e) shows
multiple cells isolated in 30 µm squares. Multiple cells are observed in the upper left and
lower right features, while single cells are isolated in the other two features. Figure
4(b,d,f) shows high magnification images of individual cells isolated in three different
membrane configurations. The images show that the filtration process is not detrimental
to cell viability, as the cells are still able to attach and spread on the filter material.
98
Figure 6.4 NIH 3T3 cells isolated in three different mold geometries; (a,b) 20 µm diameter circles, (c,d) 20 µm (side to side) hexagons, (e,f) 30 µm squares.
Figure 6.5 shows SEM images of NIH 3T3 cells isolated in (a) 30 µm square and
(b) 20 µm circular features, after removal of the microfabricated membrane from the
filter. The images show that the cells adhere well to the filter material and that the cells
conform to the geometry of the microfabricated features.
99
Figure 6.5 SEM images of NIH 3T3 cells after removing the polymer membrane from the surface of the filter; (a) 30 µm squares, (b) 20 µm diameter circles.
Fluorescent microscopy was also used to image the isolated cells. Figure 6.6
shows the three membrane geometries loaded with a relatively low seeding density (2.5 x
104 cells/ml). Primarily individual cells are isolated in the wells. By increasing the
seeding density to 5 x 104 cells/ml, multiple cells can be observed in many of the wells as
shown in Figure 6.7. The patterns are the same 20 µm circles and 30 µm squares as
Figure 6.7(a and b) with a higher seeding density. This demonstrates the ability to
qualitatively control both the efficiency of the isolation and the number of cells isolated
in each well for a given well geometry.
100
Figure 6.6 Fluorescent micrographs of NIH 3T3 cells isolated at low density in (a) 30 µm square, (b) 20 µm hexagonal, and (c) 20 µm circular wells.
Figure 6.7 NIH 3T3 cells isolated in (a) 20µm circular and (b) 30 µm square wells at higher seeding density.
101
Non-adherent THP-1 cells were isolated in 10 µm circular wells as show in Figure
6.8. The ability to isolate non-adherent cells is significant, giving the difficulties
presented in manipulating these cells. For these cells, the surface chemistry approaches
cannot be employed for isolation. Some passive techniques have been used for isolating
non-adherent cells, but to our knowledge this is the first active mechanism that has been
employed for isolation of multiple adherent and non-adherent cell types.
Figure 6.8 THP-1 cells isolated in 10 µm circular wells; (a) SEM micrograph of a single cell and (b) fluorescent micrograph of an array of wells.
Finally, the ability to isolate aggregates of cells was demonstrated with C3A liver
cells. Circular wells with 50 µm diameter were used. PPMA membranes were patterned
on Transwell filters. Cells were seeded at 5x105 cells/ml. Figure 6.9 shows low
magnification (a) and higher magnification (b) fluorescent images of groups of
hepatocytes isolated within the wells. An SEM micrograph of the cell aggregates after
removing the membrane is shown in Figure 6.9(c).
102
Figure 6.9 Images of C3A cell clusters; (a) low and (b) high magnification fluorescent images of cells in 50 µm diameter circular wells, (c) SEM micrograph of a cell clusters
after removing the patterned membrane.
103
6.4 Conclusion
Micromolding of polymer membranes was used for isolation of adherent and non-
adherent cell lines. The process provides a simple method for spatial control of cell
location as well as control of cell morphology in the case of adherent cells. The efficiency
of the isolation as well as the number of cells isolated in each well was controlled by the
density of the cell suspension. Adherent cells were able to attach and spread on the filter
material. In addition, non-adherent cells were also successfully isolated. These devices
could be used in a number of applications including biological studies of cell morphology
as well as in cell-based biosensors. These devices are also well suited for single cell
experiments where spatial isolation of cells in needed.
References 1 Y. Huang, B. Rubinsky, “Microfabricated electroporation chip for singe cell membrane permeabilization.” Sensors and Actuators A 89 (2001) 242-249. 2 M. Khine, A. Lau, C. Ionescu-Zanetti, J. Seo, L.P. Lee, “A single cell electroporation chip.” Lab on a Chip 5 (2005) 38-43. 3 K.G. Klemic, J.F. Klemic, M.A. Reed, F.J. Sigworth, “Micromolded PDMS planar electrode allows patch clamp electrical recordings from cells.” Biosensors and Bioelectronics 17 (2002) 597-604. 4 X. Li, K.G. Klemic, M.A. Reed, F.J. Sigworth, “Microfluidic system for planar patch clamp electrode arrays.” Nano Letters 6 (2006) 815-819. 5 P. Wang, G. Xu, L. Qui, Y. Xu, Y. Li, R. Li, “Cell-based biosensors and its application in biomedicine.” Sensors and Actuators B 108 (2005) 567-584. 6 M. Nishizawa, K. Takoh, T. Matsue, “Micropatterning of HeLa cells on glass substates and evaluation of respiratory activity using microelectrodes.” Langmuir 18 (2002) 3645-3649.
104
7 C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Micropatterned surface for control of cell shape, position, and function.” Biotechnology Progress 14 (1998) 356-363.
8 C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Geometric control of cell life and death.” Science 276 (1997) 1425-1428. 9 Y. Ito, “Surface micropatterning to regulate cell function.” Biomaterials 20 (1999) 2333-2342. 10 A. Brock, D. Chang, C. Ho, P. LeDuc, X. Jiang, G.M. Whitesides, D.E. Ingber, “Geometric determinant of directional cell motility revealed using microcontact printing.” Langmuir 19 (2003) 1611-1617. 11 S.N. Bhatia, C.S. Chen, “Tissue Engineering at the Micro-scale.” Biomedical Microdevices 2 (1999) 131-144. 12 M. Khalil, A. Shariat-Panahi, R. Tootle, T. Ryder, P. McCloskey, E. Roberts, H. Hodgson, C. Selden, “Human hepatocyte cell lines proliferating as cohesive spheroid colonies in alginate markedly upregulate both synthetic and detoxificatory liver function.” Journal of Hepatology 297 (2001) 68-77. 13 R. Clicklis, J.C. Merchuk, S. Cohen, Modeling mass trasfer in hepatocyte spheroids via cell viability, spheroid size and hepatocellular funtion.” Biotechnology and Bioengineering 86 (2004) 672-680. 14 S.M. Dang, M. Kyba, R. Perlingeiro, G.Q. Daley, P.W. Zandstra, “Efficiency of embryoid body formation and hematopoietic development from embryonic stem cells in different culture systems.” Biotechnology and Bioengineering 78 (2002) 442-453. 15 Y. Xia, G.M. Whitesides, “Soft Lithography.” Annual Review of Materials Science 28 (1998) 153-184. 16 M. Mrksich, L.E. Dike, J. Tien, D.E. Ingber, G.M. Whitesides, “Using microcontact printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver.” Experimental Cell Research 235 (1997) 305-313. 17 R.S. Kane, S.Takayama, E. Ostuni, D.E. Ingber, G.M. Whitesides, “Patterning proteins and cells using soft lithography.” Biomaterials 20 (1999) 2363-2376. 18 H. Kaji, K. Takoh, M. Nishizawa, T. Matsue, “Intracellular Ca2+ imaging for micropatterned cardiac myocytes.” Biotechnology and Bioengineering 81 (2003) 748-751. 19 K.Y. Suh, J. Seong, A. Khademhosseini, P.E. Laibinis, R. Langer, “A simple soft lithographic route to fabrication of poly(ethylene glycol) microstructures for protein and cell patterning.” Biomaterials 25 (2004) 557-563.
105
20 M.R. Dusseiller, D. Schlaepfer, M. Koch, R. Kroschewski, M. Textor, “An inverted microcontact printing method on topographically structured polystyrene chips for arrayed micro-3-D culturing of single cells.” Biomaterials 26 (2005) 5917-5925. 21 E.W.H. Jager, C. Immerstrand, K.H. Peterson, K. Magnusson, I. Lundström, O. Inganäs, “The cell clinic: closable microvials for single cell studies.” Biomedical Microdevices 4 (2002) 177-187. 22 A. Folch, B. Jo, O. Hurtado, D.J. Beebe, M. Toner, “Microfabricated elastomeric stencils for micropatterning cell cultures.” Journal of Biomedical Materials Research 52 (2000) 346-353. 23 M.J. Rosenbluth, W.A. Lam, D.A. Fletcher, “Force microscopy of nonadherent cells: a comparison of leukemia cell deformability.” Biophysical Journal 90 (2006) 2994-3003. 24 J. Fukuda, K. Nakazawa, “Orderly arrangement of hepatocytes spheroids on a microfabricated chip.” Tissue Engineering 11 (2005) 1254-1262. 25 N. Ferrell, D. Hansford, “Fabrication of micro- and nanoscale polymer structures by soft lithography and spin dewetting.” Macromolecular Rapid Communications 28 (2007) 964-967. 26 F. Braet, R. De Zanger, E. Wisse. Drying cells for SEM, AFM and TEM by hexamethyldisilazane: a study on hepatic endothelial cells. Journal of Microscopy 186 (1997) 84-87.
106
CHAPTER 7
FABRICATION OF SUSPENDED POLYMER MICROSTRUCTURE BY PATTERNED SUBSTRATE MICROMOLDING AND SACRIFICIAL LAYER
MICROMOLDING
7.1 Introduction
MEMS technology has traditionally been focused on silicon based fabrication
techniques. In order to broaden the potential applications of MEMS devices, novel
materials must be considered for some applications. Polymers are of particular interest
due to their unique and diverse material properties and low cost, relative to silicon.
Polymer MEMS (P-MEMS) also hold great potential for biological applications due to
the inherent biocompatibility of many polymer materials.1 While silicon based fabrication
techniques are well established and understood, new processing techniques need to be
developed for polymer materials. This chapter introduces two methods for fabrication of
freely suspended three dimensional polymer microstructures for bioMEMS applications:
sacrificial layer micromolding (SLaM) and patterned substrate micromolding (PSM).
Photolithography has been the standard in fabricating polymer microstructures
due to its prevalence in microelectronics processing. Recently, several processes have
been developed for microfabrication of polymeric microstructures and devices from
materials other than standard photosensitive polymers. Microscale hot embossing2
microinjection molding,3 as well as the soft lithography based fabrication methods,
107
including replica molding,4 solvent assisted micromolding4 (SAMIM), microtransfer
molding (µTM),5 and micromolding in capillaries (MIMIC)6 have all been developed for
fabrication of polymer microstructures. These processing methods have found utility in a
number of biomedical applications such as microfluidic diagnostic and analytical
devices,7,8 tissue engineering,9,10 and drug delivery .11 However, they are generally used
for fabrication of planar structures. For suspended structures, other techniques such as
etching3 or photolithography using negative photoresist12 must be used. These processes
are limited in the materials that can be used in fabrication and require extensive use of
expensive cleanroom facilities and equipment.
SLaM and PSM provide alternative methods for fabrication of suspended polymer
microstructures from a wide range of thermoplastic, biocompatible polymer materials
that are not traditionally used in microfabrication but are common in biomaterials
applications. In addition, polymer materials are generally inexpensive and a majority of
the process can be performed outside of a cleanroom with minimal use of expensive
microfabrication equipment, thus reducing the overall cost of the process. The
applicability of these processes to a wide range of structural materials also increases the
materials selection pool, allowing materials to be selected based on appropriate biological
interactions.
Both SLaM and PSM involve double stamp molding processes followed by
alignment and bonding to a micropatterned substrate. The processes are capable of
yielding polymer microstructures from thermoplastic polymers with resolution
comparable to the resolution of the photolithography process. In the present work,
PPMA, PMMA, and polystyrene were used as model polymers. PPMA was used because
108
of its ease of processing due to its low glass transition temperature (~37°C), and PMMA
and polystyrene were chosen due to their wide use in biomedical applications. The same
basic process could also be extended to a variety of other thermoplastic polymer
materials. The processes provide a simple and cost effective method for fabrication of
polymer devices for integration into MEMS based sensor and actuator systems.
7.2 Materials and methods
7.2.1 PDMS mold fabrication
The device geometries were initially fabricated from photoresist using standard
photolithography. A layer of either SU8-5 negative tone photoresist (MicroChem Corp.)
or S1813 positive tone photoresist (Shipley) was spin coated on a <100> p-type silicon
wafer (WaferNet). The photoresist and spin speed used in the initial process were
selected based on the desired final structure geometry and thickness. After coating, the
wafers were processed according to the manufacturers’ suggestion processing parameters.
The photoresist features were transferred into a poly(dimethlysiloxane) (PDMS)
elastomer for use in fabricating the final polymer structural features. The process of
PDMS molding from a photoresist master is described in detail elsewhere.13 Briefly, a
10:1 ratio of T-2 transparent base and curing agent (Dow Corning) was mixed and stirred
thoroughly. The mixture was then poured over the patterned silicon wafer and placed in
a vacuum dessicator to remove bubbles incorporated during mixing. The sample was
removed from the vacuum periodically, and a razor blade was used to remove surface
bubbles. After the bubbles were completely removed, the PDMS mold was allowed to
109
cure at room temperature for 48 hours before removing the mold from the wafer. The
wafer could be used multiple times for making PDMS molds.
7.2.2 Sacrificial layer micromolding
Figure 7.1 shows the fabrication process used in the sacrificial layer
micromolding process. Poly(vinyl alcohol) (PVA) (Sigma-Aldrich) was chosen as a
sacrificial layer due to its solubility in water, insolubility in organic solvents, and thermal
stability. A 10:1 ratio of deionized water and PVA were mixed and heated to 70 ºC to
promote dissolution. After the polymer had completely dissolved, the solution was
filtered through high flow rate filter paper to remove impurities. The PVA/water solution
was spin coated on silicon wafers for 60 seconds at a spin speed of 1000 rpm. After
coating, the wafers were baked at 95 ºC for five minutes to remove any residual water.
The resulting sacrificial layer thickness was ~750 nm. Sacrificial layer thickness was
characterized using tapping mode atomic force microscopy (AFM) (Veeco, Dimension
3100). The sacrificial layer thickness can be varied by changing the PVA concentration
and spin speed. A thin layer (~400 nm) of poly(methyl methacrylate) (PMMA)
(Scientific Polymer Products) was then spin coated on the PVA layer and baked at 115 ºC
for two minutes. The PMMA layer acted to protect the PVA from being dissolved during
development in the upcoming photolithography process. A layer of SPR220-7 positive
tone photoresist (Shipley) was spin coated on the PMMA layer at 4500 rpm resulting in a
photoresist thickness of ~4.5 µm. The photoresist was processed according to the
manufacturer’s processing parameters to expose the anchor regions of the underlying
PMMA and PVA layers.
110
Figure 7.1 Schematic diagram of the sacrificial layer micromolding method. (a) sacrificial layer patterning process, (a,i) substrate is coated with PVA, PMMA, and photoresist, respectively, and the photoresist is patterned via photolithography, (a,ii)
surface is exposed to O2 plasma to etch the PMMA and PVA layers, (a,iii) photoresist and PMMA are removed with acetone, (b) micromolding, (b,i) patterned PDMS mold is uniformly coated with the structural polymer, (b,ii) coated mold is brought into contact
with a heated glass plate to remove surface material, (b,iii) glass plate is removed leaving polymer only in the recessed portions of the mold, (c) alignment and bonding, (c,i)
sacrificial layer and patterned mold are aligned, (c,ii) sacrificial layer and structural layer are brought into contact and bonded under heat and pressure, (c,iii) mold is removed,
(c,iv) sacrificial layer is removed by immersion in water.
The PMMA and PVA layers were then etched with an oxygen plasma in a
benchtop reactive ion etcher (Technics, MicroRIE 800). The patterned SPR220-7 layer
acted as an etch mask. An oxygen flow rate of 15 sccm, power of 200W, and 163 mTorr
pressure were used during the etch process. After the exposed PMMA and PVA were
selectively removed and the substrate anchor points were exposed, the wafer was
111
immersed in acetone for two minutes to remove the PMMA and photoresist layers,
leaving only the patterned PVA. After removal from the acetone bath, samples were dried
with a gentle stream of nitrogen.
The structural portion of the devices was fabricated via a modified micro-transfer
molding process.14,15 Poly(n-propyl methacrylate) (PPMA) (Scientific Polymer Products),
polystyrene (Sigma-Aldrich), and PMMA (Scientific Polymer Products) were chosen as
structural materials. Polymers were dissolved in anisole (Sigma-Aldrich) at
concentrations ranging from 1-10% (wt/wt). The polymer solution was spin coated onto
the patterned PDMS mold at spin speeds from 1000-6000 rpm for 60 seconds. Devices
ranged from ~200 nm to >5 µm in thickness based on the above parameters. To remove
the polymer from the surface (not in the recessed features) of the PDMS, the mold was
brought into contact with a glass slide heated to 175 °C. The same temperature was used
for all three structural materials. The mold was held on the slide for 5-10 seconds under
the pressure of its own weight. The mold was immediately removed, with the glass slide
still in contact with the heat source. The process was repeated if necessary to completely
remove the surface material, leaving the polymer only in the recessed portions of the
PDMS mold.
The selectively patterned PDMS mold was then aligned with the sacrificial layer
under an optical microscope. The substrate was heated to 95 ºC for PPMA, 120 ºC for
polystyrene, and 175 ºC for PMMA and a pressure of 0.21 MPa (30 psi) was applied to
the backside of the mold and held for 10 seconds. The pressure brought the structural
material into contact with the substrate material and the heating process promoted
adhesion between the sacrificial and structural layers. The molds were removed, resulting
112
in the final unreleased structures. Resultant device thickness was a function of the PDMS
mold depth, polymer concentration, and spin speed. Devices were characterized using a
mechanical stylus profilometer (Veeco Dektak, Series 3), atomic force microscopy, and
scanning electron microscopy (Hitachi, S-3000H)
The devices were released by immersion in DI water, which dissolved the
sacrificial PVA layer. Stiction during the drying process prevents the devices from being
operated under dry conditions. Our studies have shown that the released devices remain
separated from the substrate so long as they are maintained in an aqueous environment.
7.2.3 Patterned substrate micromolding
The patterned substrate method is similar to the sacrificial layer method in that the
same microtransfer molding process is used for fabrication of the structural portion of the
devices. The method differs in choice of the substrate material. For the sacrificial layer
method, silicon or glass acted as the substrate material and anchor region for the devices,
and a sacrificial material is removed to release the device. For the patterned substrate
method, a layer of patterned photoresist acts as the substrate material, and voids in the
photoresist layer act as the suspended portions of the devices.
Figure 7.2 shows a schematic of the patterned substrate fabrication method. A
layer of SU8-25 negative tone photoresist was patterned via standard photolithography
processing. The photoresist was coated at 2000 rpm, yielding a film thickness of
approximately 25 µm. The model substrate geometry for this work consisted of 20-30 µm
wide channels with 500 µm spacing between the channels.
113
Following the photolithography process, the double stamp micromolding process
described earlier was performed with a PMDS mold of the desired device geometry. In
this case, a mold geometry with 5 µm wide channels with 45 µm spacing was used,
yielding final structures consisting of an array of 5 µm wide beams suspended over the
20-30 µm channels in the photoresist. This is shown in Figure 7.2(a). The selectively
coated mold and photolithographically patterned substrate were then brought into contact
with heat and pressure. This transferred the polymer in the PDMS mold onto the
photoresist substrate with polymer suspended across the gaps in the patterned substrate.
Figure 7.2 Schematic diagram of the micromolding process used in polymer beam fabrication; (a) selectively coated PDMS mold, (b) mold inverted and aligned with
patterned substrate, (c) heat a pressure used to stamp the features, and (d) mold removed.
114
7.3 Results and discussion
Scanning electron micrographs of devices resulting from the SLaM process are
shown in Figure 7.3. The images show that multiple and complex geometries are
achievable and that the integrity of the feature geometry can be maintained throughout
the PDMS molding and micromolding processes. The features in Figure 7.3(d) are
nominally 750 nm, indicating that the feature resolution of the process reproduces that of
the photolithography process. The devices are shown prior to release of the sacrificial
layer. Figure 7.4 shows optical micrographs of polymer cantilevers beams before and
after release of the sacrificial layer. Figure 7.4(a) shows 10 µm wide beams before the
sacrificial layer was removed, and 7.4(b) shows the suspended devices after removal of
the sacrificial layer.
Figure 7.3 Scanning electron micrographs of devices made by the sacrificial layer micromolding (SLaM) method. Devices are shown prior to release of the sacrificial layer.
Arrows indicated the structural PPMA regions, sacrificial PVA layer, fixed anchor regions, and underlying silicon substrate. Structures shown include (a) cantilevers, (b)
cross-bridges, and (c) cantilever with through holes. The features in (d) have nominally 750 nm line width and are shown to demonstrate process resolution.
115
Figure 7.4 Optical micrographs of polymer cantilevers fabricated using the sacrificial layer micromolding process (a) before removal of the sacrificial layer and (b) after
release.
Figure 7.5(a) shows a phase contrast micrograph of a released PPMA cantilever for
low magnitude lateral force measurement applications. The cantilever is 250 µm long, 5
µm wide, and 3.5 µm thick. A micropipette was used to apply a force to the cantilever as
shown in Figure 7.5(b). Deflection of the beam was then measure optically and related to
the force magnitude using the cantilever beam bending equation:
3
3LEIxF = (1)
where F is the applied load, x is the measured deflection, L is the distance from the base
of the beam to the applied load, E is the Young’s modulus, and I is the area moment of
inertia. The load was assumed to be a point load and was applied at 245 µm from the base
of the cantilever. The modulus was taken to be 0.7 GPa16 and the moment of inertia
calculated from the above geometry is 3.65x10-23 m4. The deflection of the beam was
measured at 20 µm using image analysis software (ImagePro Discovery). The magnitude
116
of the resulting force is 104 nN. The same force applied to a silicon beam (E=190 GPa)17
would result in a lateral deflection of only 74 nm. This deflection would not be detectable
using a standard optical microscopy, and a much more sophisticated measurement system
would be needed. This illustrates the potential of using polymer MEMS for fabrication of
highly sensitive devices that would not be functional if they were fabricated from
standard silicon based MEMS materials.
Figure 7.5 Phase contrast micrographs of 5 µm wide cantilevers made using the sacrificial layer micromolding process (a) after removal of the sacrificial layer and (b)
after application of a 104 nN force using a micropipette tip.
Resulting devices from the patterned substrate micromolding (PSM) process are
shown in Figure 7.6. Polystyrene beams are shown in Figure 7.6(a) and PPMA beams in
Figure 7.6(b). The images show that the devices remain suspended over the channels with
a flat beam profile between the anchor points. The ability to fabricate flat structures over
the channels is constrained by the width of the channel (i.e. the portion over which the
device is suspended). For the work described here, 30 µm was the maximum channel
117
width used for fabrication of 5 µm wide structural features. Further work is needed to
determine the maximum attainable functional region for this or other feature geometries.
The non-uniformity of the channels in the substrate is a result of the use of a printed
transparency mask14 as opposed to a chrome/glass mask. Resolution of the substrate
channels could be considerably improved by use of a chrome/glass mask, but a
significant increase in cost would also be incurred.
Figure 7.6 Scanning electron micrographs of (a) polystyrene and (b) PPMA beams made using the patterned substrate method. The substrate is photolithographically patterned
SU8 photoresist.
118
The proper choice of a polymer MEMS fabrication method is dependant on the
given application. Both fabrication methods described here are limited by practical
fabrication constrains. The most significant constraint for the sacrificial layer
micromolding process at present is the inability to operate devices in dry conditions due
to stiction issues. Work is current being done to develop a drying process to minimize
this effect, allowing the devices to remain suspended in air. The geometrical constrains of
the SLaM technique are minimal. Suspended regions of >200 µm can be achieve and
remain suspended in aqueous environment. In contrast, the PSM allows devices that can
be operated in dry condition, but the length of the suspended regions of the device are
limited to those that maintain their structural integrity.
Given that both fabrication methods use a soft mold for fabrication, distortion of
the features during the final stamping process results in an undesirable loss of feature
resolution if the processing parameters for a given device geometry are not appropriately
selected and characterized. Figure 7.7 shows AFM section scans of devices of similar
geometries with (a) properly selected and (b) poorly selected processing parameters. The
parameters that are important for minimizing mold distortion are mold depth, polymer
solution concentration, spin speed, and stamping transfer pressure. The processing
parameters must be adjusted for any given feature geometry. As an example, the feature
in Figure 7.7(a) was processed using a 1.4 µm deep mold, 5% polymer solution, 4000
rpm spin speed, and 0.21 MPa (30 psi) transfer pressure. The mold distortion is less
critical for the PSM techniques, since the functional regions of the devices are never not
brought into contact with the substrate, thus the functional portions maintain their
geometrical integrity even if processing parameters are not completely optimized.
119
Figure 7.7 AFM section analysis of features with (a) appropriately selected process parameters and (b) poorly chosen process parameters. Important parameters include
mold depth, solution concentration, spin speed, and transfer pressure. The top feature has a height of 875 nm and the bottom feature has a maximum height of 1.12 µm and a
minimum height of 760 nm.
7.4 Conclusion
Successful fabrication of suspended polymer microstructures was achieved using
two different fabrication methods. Both processes are soft lithography based and rely on a
two-step micromolding process. The choice of substrate material (i.e. sacrificial layer or
patterned substrate) is application dependant, as each method is capable of producing
devices of different geometries and for specific operating conditions. The processes are
relatively simple and low cost as compared to silicon fabrication techniques, and the
120
transition from silicon to polymer materials could potentially broaden both the
functionality and applicability of the devices in a biological setting.
References 1 B.D. Ratner, A.S. Hoffman, F.J. Schoen, and J.E. Lemons, Biomaterials Science An Intoduction to Materials in Medicine, Elsevier Academic Press, San Diego, CA (2004) p. 67. 2 H. Becker and U. Heim, “Hot embossing as a method for the fabrication of polymer high aspect ratio structures.” Sensors and Actuators A 83 (2000) 130-135. 3 L.J. Lee, M.J. Madou, K.W. Koelling, S. Daunert, S. Lai, and C.G. Koh, “Design and fabrication of CD-like microfluidic platforms for diagnostics: polymer-based microfabrication.” Biomedical Microdevices 3 (2001) 339-351. 4 Y. Xia, G.M. Whitesides, “Soft lithography.” Annual Review of Materials Science 28 (1998) 153-184. 5 X.-M. Zhoa, Y. Xia, and G.M. Whitesides, “Fabrication of three-dimensional micro-structures: microtransfer molding.” Advanced Materials 8 (1996) 837-840. 6 E. Kim, Y. Xia, and G.M. Whitesides, “Polymer microstructures formed by moulding in capillaries.” Nature 376 (1995) 581-584. 7 H. Becker, C. Gärtner, “Polymer microfabrication methods for microfluidic analytical applications.” Electrophoresis 21 (2000) 12-26. 8 S.K. Sia, G.M. Whitesides, “Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies.” Electrophoresis 24 (2003), 3563-3576. 9 G. Vozzi, C. Flaim, A. Ahluwalia, S. Bhatia, “Fabrication of PLGA scaffolds using soft lithography and microsyringe deposition.” Biomaterials 24 (2003), 2533-2540. 10 S.N. Bhatia, C.S. Chen, “Tissue engineering at the microscale.” Biomedical Microdevices 2 (1999) 131-144. 11 B. Ziaie, A. Baldi, M. Lei, Y. Gu, and R.A. Siegel, “Hard and soft micromachining for BioMEMS: a review of techniques.” Advanced Drug Delivery Reviews 56 (2004) 145-172
121
12 G. Genolet, J. Brugger, M. Despont, U. Drechsler, P. Vettiger, N.F. de Rooij, D. Anselmetti, “Soft, entirely photoplastic probes for scanning force microscopy.” Review of Scientific Instruments 70 (1999) 2398-2401.
13 D.C. Duffy, J.C. McDonald, O.J.A. Schueller, G.M. Whitesides, “Rapid prototyping of microfluidic systems in poly(dimethylsiloxane).” Analytical Chemistry 70 (1998) 4974-4984. 14 J. Guan, A. Chakrapani, D.J. Hansford, “Polymer microparticles fabricated by soft lithography.” Chemistry of Materials 17 (2005) 6227-6229. 15 J. Guan, H. He, D.J. Hansford, L.J. Lee, “Self-folding of three-dimensional hydrogel microstructures.” The Journal of Physical Chemistry B 109 (2005) 23134-23137. 16 G. Wei, B. Bhushan, N. Ferrell, D. Hansford, Journal of Vacuum Science and Technology A 23, 811-819 (2005). 17 K.E. Peterson, “Silicon as a mechanical material.” Proceedings of the IEEE 70 (1982) 420-457.
122
CHAPTER 8
MEASURING THE MECHANICAL PROPERTIES OF POLYMER MICROSTRUCTURES BY NANOINDENTATION
8.1 Introduction
Mechanical characterization of materials is critical for design, modeling, function,
and reliability of MEMS devices. This is particularly important in the case of polymer
MEMS, where micro and nanoscale polymer properties may differ significantly from
bulk material properties.1-3 In the case of device design and modeling, a good grasp of the
materials properties will lead to a better understanding of device behavior prior to
fabrication. This in turn leads to fewer iterations in the design, decreased cost, and faster
device development. In terms of function, characterization of the material properties is
important for ensuring that the device will perform as designed. Finally, characterization
of the failure properties of materials will aid in understanding and improving the long
term reliability of MEMS.
Some of the important properties that are relevant for polymer MEMS
development include microscale hardness, elastic modulus, creep, adhesion, scratch
resistance, yield strength, and breaking strength. For biomedical applications of polymer
MEMS, devices often need to be operated in aqueous condition and/or at elevated
123
temperatures. Therefore, the polymer properties need to be tested under the appropriate
environmental conditions to simulate operating conditions.
Many well-established techniques are available for bulk macroscale materials
characterization. However, these methods are not generally applicable to micro and
nanoscale polymer characterization. Specialized characterization techniques are needed
to accommodate these high resolution measurements and small sample volumes.
Nanoindentation is a powerful tool for micro/nanoscale materials characterization. In
nanoindentation, a nanoscale tip with a know geometry is used to penetrate a material.
Similar to macroscale indentation tests, this data can be used determine the hardness of
the material. In addition, by measuring both load and displacement simultaneously, other
materials properties, including elastic modulus, can be determined. Some of the
advantages of nanoindentation include the ability to make measurements on extremely
thin films. The high resolution of nanoindentation with respect to load and displacement
are is also advantageous for measuring properties at very low loads.4 Some disadvantages
of nanoindentation are the need to correct for various instrumental factors, including
determining the point of the tip-material contact, accounting for the compliance of the
load frame, and determining the area function. Additionally, materials issue such as
creep, material pile-up around the tip, and substrate effects must be accounted for. While
many of these issues can be minimized by proper calibration and data analysis, failure to
account for these effects can lead to complication or misinterpretation of data.5
Many nanoindenters are also equipped with the continuous stiffness measurement
(CSM) technique.6 The CSM technique allows the nanoindenter tip to be oscillated at
high frequency while the load applied by the tip is incrementally increased. A plot of the
124
CSM load cycle is shown in Figure 8.1. Unlike conventional nanoindentation, where data
can only be collected at the unloading point, CSM allows continuous measurement of the
mechanical properties throughout the loading cycle. In addition, the high frequency tip
oscillation minimizes the influence of creep on the hardness and elastic modulus
measurements. CSM also allows measurement of the creep properties of a material by
applying a constant load and measuring the change in the indentation depth over time.
Figure 8.1 Load-displacement plot for CSM nanoindentation. The inset shows the high frequency oscillation of the tip during the load cycle.
125
We performed detailed mechanical characterization of four different materials
that have been or could potentially be used in polymer MEMS devices. The materials
chosen for this study were poly(methyl methacrylate) (PMMA), poly(propyl
methacrylate) (PPMA), poly(styrene) (PS) and a polystyrene-nanoclay composite
(PS/Clay). PMMA and PS are widely used engineering polymers. Historically, PMMA
has been the polymer of choice in ophthalmologic devices due to its high refractive
index, hardness and biocompatibility. Its surface can be functionalized with proteins,
which promotes bonding of tissues for in vivo implants.7 PMMA is also employed in
chips and valve components for immunosensors and other lab-on-a-chip applications.8,9
PPMA has a lower glass transition temperature (Tg) (35-43 oC)10 than PMMA (104-
106oC)11,12 which allows for easier processing at a lower temperature.
Polystyrene is particularly desirable for BioMEMS due to its ubiquitous use in
tissue culture applications. Tissue culture treated polystyrene is the most commonly used
material for in vitro cell biology studies of adherent cells. Thus, there is a wealth of
information about cellular behavior on polystyrene. In addition to tissue culture
polystyrene, various surface modification techniques can be employed for functionalizing
the PS surface in order to promote cell attachment and proliferation.7 Oxygen plasma
modified polystyrene has been shown to improve cell growth, proliferation, and
expression of cellular adhesion proteins proportional to the surface oxygen
concentration.13 The previous knowledge of cellular interactions with polystyrene makes
it a logical choice to use in BioMEMS devices for cellular interactions.
Clay nanoparticles in a polymer matrix act to improve the mechanical and thermal
properties as compared to the native polymer.14,15 Nanoclay composites can also be used
to improve barrier resistance and improve ionic conductivity. The extent to which
polymer properties are affected is determined by the clay content, polymer/clay
interfacial strength, and dispersion of the clay particles. Improvements in these
properties have made polymer nanocomposites candidate materials for use in the
automotive and aerospace industries.16,17 In addition, polymer nanocomposites have been
126
considered for diverse applications such as food packaging, flame retardants and
batteries.18-20
In polymer MEMS applications, the ability to control material properties over a
range of values could significantly improve device design flexibility. The
polystyrene/clay nanocomposite is beneficial, as it may allow fine-tuning of the
mechanical properties by varying the filler content while maintaining the
biocompatibility of the PS matrix. Complete redesign of the device is avoided as the
stiffness of the material can be altered as needed. Since the device design and fabrication
procedure are not changed, the process becomes more cost-effective.
Here, we present microscale mechanical property characterization of PPMA,
PMMA, polystyrene, and polystyrene/clay nanocomposites. Hardness, elastic modulus,
and creep properties were measured by CSM nanoindentation on supported polymer
microstructures. Elastic modulus was also measured by performing force-deflection
measurements on suspended microbridges fabricated using patterned substrate
micromolding (PSM). Yield and breaking strengths were evaluated by normal beam
bending at elevated loads. Scratch tests were performed on polymer films by translating
the nanoindenter tip across the surface on the film while ramping the load. The relative
scratch resistance of the materials was evaluated using the tip displacement profile and
SEM images of the damage. Finally, lateral bending of PS and PS/Clay cantilever beams
is demonstrated for the first time. Hardness and elastic modulus experiments were
conducted on samples that were soaked in deionized water to assess the effect of the
aqueous medium. In addition, the indentation and bending response of the polymer beams
127
was examined at human body temperature (37.5 oC) by fitting a heating assembly into the
indenter’s sample stage.
8.2 Materials and methods
8.2.1 Fabrication of polymer microstructures and thin films
Polymer solutions (PPMA, PMMA, and PS) were made by dissolving the
materials in anisole at various concentrations depending on the application. To prepare
the nanocomposite, the clay was first dispersed in the PS matrix by melt compounding.
The composite was then dissolved in anisole and sonicated for at least 8 hours to dissolve
the polymer and re-disperse the particles. The resulting concentration of the clay was
10% wt/wt (clay/PS). The clay additive in PS/Clay is Cloisite® 20A surface modified
natural montmorillonite (Southern Clay Products, Inc.) with thickness of approximately 1
nm and lateral dimensions of 70-150 nm.
128
Polymer microbridges were fabricated using the patterned substrate micromolding
technique21 as described in Chapter 7. Polymer cantilevers for lateral force measurements
were fabricated by sacrificial layer micromolding with one process variation. The mold
for the lateral force cantilevers consisted of 50 µm wide channels that were
approximately 27 µm deep. The PDMS mold was coated with a 15% solution of PS or
10% solution of PS/Clay. The polymers were removed from the mold and transferred to
the substrate using the same process that was used for PS in the previous chapter, except
that the PS/Clay was removed at a higher temperature (175 °C) than PS (125 °C). The
primary difference in the processes was the addition of a reinforcing structure at the base
of the cantilevers as indicated in Figure 8.2(vii). The reinforcement was applied using the
same process, with the reinforcing structures oriented perpendicular to the original
structures. This provided the necessary reinforcement to ensure that the polymer
cantilever was not delaminated from the surface during the test.
The polymer films that were used to assess the scratch resistance were made by
dissolving the polymers in anisole at various concentrations to achieve a film thickness of
approximately 500 nm. The polymer solutions were then spin coated on silicon at 3000
rpm for 1 minute. The films were then baked at 95 °C for 2 minutes to remove any
residual solvent. The thickness of the film samples was evaluated by scratching the film
with a razor blade and measuring the thickness using a stylus profilometer.
Figure 8.2 Schematic diagram of the process for fabrication of polymer cantilevers for lateral force measurements.
8.2.2 Mechanical characterization
Figure 8.3 shows an SEM micrograph of a polymer beam with the testing
locations for hardness, elastic modulus, and creep. The figure also shows the location of
129
the bending tests that were used to evaluate the elastic modulus, yield strength, and
breaking strength. Hardness, elastic modulus, and creep of the supported section of the
microstructures were measured using the CSM technique with a NanoIndentor II® (MTS
Systems Corp) with a diamond Berovich tip. The hardness and elastic modulus were
determined continuously between contact and 500 nm using a peak-to-peak load
amplitude of 1.2 µN at a frequency of 45 Hz. Creep was measured using a 30 µN
constant load that was held for 600 seconds.
Figure 8.3 SEM micrograph of the microbridges showing the location of the indentation experiment on the substrate supported portion of the structure and bending tests on the
suspended region of the structure.
130
Details of the numerical calculations for determining the hardness and elastic
modulus from the nanoindentation data can be found in Oliver and Pharr22 or Bhushan
and Li.23 Hardness (H) is calculated using equation 1:
AP
H max= (1)
where Pmax is the maximum load and A is the projected contact area. The elastic modulus
(E) is calculated by first determining what is referred to as the reduced modulus (Er)
according to equation 2:
ASEr 2
π= (2)
where S is the stiffness. The elastic modulus for the material can then be calculated from
equation 3:
)/)1(()/1(1
2
2
ttr EEE
υυ−−
−= (3)
where υ is the Poisson’s ratio of the material being tested, and υt and Et are the Poisson’s
ratio and elastic modulus of the tip material, respectively. Poisson’s ratio values for
PMMA, PS, PS/Clay, and PPMA were taken as 0.35, 0.325, 0.33 and 0.33, respectively.
Data for the Poisson’s ratio of PS/Clay and PPMA were not available, so a value of 0.33
was estimated based on the value for typical thermoplastic polymers.11
For scratch testing, a conical tip with included angle of 90° and radius of
curvature of 1 µm was translated over the surface of the polymer films as the load was
ramped to a maximum of 0.5 mN. The scratch length was 500 µm and the velocity was 5
µm/s. The coefficient of friction and scratch depth were measured with respect to the
131
normal load. The damage to the polymer film was evaluated using scanning electron
microscopy.
The beam bending experiments were performed using a conical tip with 1 µm
diameter and 90° radius of curvature. A schematic diagram of the nanoindenter setup for
bending tests is shown in Figure 8.4. The tip was dip coated with 2% PMMA (wt/wt) to
avoid damaging the beams with the tip. The bending location was determined using a
1500x microscope objective to apply the load as close as possible to the center of the
beam. In order to maintain the proper location, the recalibration of the distance between
the indenter and the microscope was repeated frequently according to the manufacturer’s
specifications. This calibration was performed on a soft material, polycaprolactone, to
minimize the damage to the PMMA coating on the tip. Samples were also tested after
soaking in deionized water for 36 hours to determine the effect of prolonged exposure to
aqueous environment. The sample stage was also heated to 37 °C to test the material
properties at body temperature. Details of the heating stage can be found elsewhere.24
The elastic modulus was calculated from the load-displacement plot according to
equations 4 and 5:25
12
3bhI = (4)
mI
lE192
3
= (5)
where I is the area moment of inertia for the rectangular beam cross section, b is the
width of the beam and h is the height of the beam. The elastic modulus can then be
calculated using equation 5, where l is the length of the beam and m is the slope of the
load displacement curve. 132
Yield and breaking strength of the polymer beams were evaluated by normal
beam bending at higher loads than those used to evaluate the elastic modulus. The yield
and breaking strength (σ) were calculated from equation 6:25
243
bhFl
=σ (6)
where F is the load corresponding to either the onset of yield or total failure, and l, b, and
h are length, width, and height of the beam, respectively.
Figure 8.4 Nanoindentation setup for beam bending.
Lateral bending tests were performed on the polymer cantilevers. The cantilevers
were first released by immersion in deionized water for 1 hour. The samples were tested
133
in water to avoid stiction of the cantilevers during the drying process. The nanoindenter
tip was then position 20 µm from the tip of the free end of the cantilever at a distance of
approximately 200-300 µm away from the edge of the cantilever. A scratch test was
performed at a constant load of 400 µN. Any additional load was attributed to the lateral
force of the beam against the tip.
8.3 Results and discussion
8.3.1 Hardness and elastic modulus by CSM nanoindentaion
Typical plots of the hardness and elastic modulus as a function of the penetration
depth are shown in Figure 8.5. Each material was tested five times (n=5) and the values
were taken at a penetration depth of the 100 nm. The respective hardness and elastic
modulus were found to be 390 ± 50 MPa and 5.1 ± 0.4 GPa for PS/Clay, 340 ± 30 MPa
and 4.8 ± 0.5 GPa for PMMA, 290 ± 20 MPa and 3.6 ± 0.4 GPa for PS, and 110 ± 30
MPa and 1.7 ± 0.5 GPa for PPMA. The values are given as the average ± the standard
deviation.
134
Figure 8.5 Plots of hardness and elastic modulus as a function of contact depth for CSM nanoindentation measurements.
8.3.2 Creep behavior by CSM nanoindenation
The viscoelastic nature of polymer materials makes creep behavior an important
consideration for polymer devices. The creep behavior of the materials was tested by
holding the load constant at 30 µN and monitoring the displacement of the tip over time.
Figure 8.6 shows plots of the tip displacement, mean stress, and contact stiffness versus
time for each material. The plots show that the displacement increases and the mean
stress decreases over time for each of the materials. This behavior is indicative of creep,
as the polymer material deforms around the tip. The data shows that all of the materials
exhibit some creep behavior. However, the rate of creep is directly related to the hardness
and modulus of the materials, with the harder materials with higher modulus showing the
135
slowest creep rate. The addition of a ceramic component in the PS/Clay nanocomposite
also helps to decrease the creep.
Figure 8.6 Displacement, mean stress, and contact stiffness as a function of time for a constant applied load of 30 µN.
8.3.3 Scratch resistance from nanindentation scratch test
Results from scratch testing are shown in Figure 8.7. The plots give the
coefficient of friction (left axis) and scratch depth (right axis) as a function of the normal
load. SEM images were taken at the beginning (region A), middle (region B), and end
(region C) of the scratch to evaluate the damage to the films. PMMA showed the best
scratch resistance, as indicated by a low penetration depth and smooth scratch profile.
136
The addition of clay nanoparticles seemed to improve the scratch properties of the
PS/Clay material relative to unmodified polystyrene. The spikes in the depth profile for
PS/Clay are attributed to roughness in the film. Polystyrene showed signs of film
delamination and material pile-up around the tip at approximately 0.45 mN. This is
indicated by sudden instability in the depth profile and is confirmed by SEM observations
near the end of scratch. PPMA showed the lowest scratch resistance with delamination
and pile-up occurring between 0.2-0.25 mN. Again, this was confirmed by SEM imaging
of the scratch.
Figure 8.7 Plots of the coefficient of friction and depth profile for nanoindentation scratch tests (left) and SEM images at the beginning (region A), middle (region B), and end
(region C) of the scratch.
137
8.3.3 Elastic modulus from normal beam bending
Figure 8.8 shows an example of data from a bending experiment performed with
an uncoated diamond tip. The plot shows an obvious change in the load/displacement
profile between 75-100 nm of displacement and a load of approximately 2 µN. This
effect results from significant penetration of the tip into the surface of the beam
structures. To avoid this effect, the tip was coated with a thin layer of PMMA for all
bending test. While the compliance of the PMMA coating is not considered in the
calculation of the modulus, it is assumed to be negligible given the small thickness of the
coating.
Figure 8.8 Load-displacement plot for PPMA beam bending with an uncoated diamond nanoindenter tip.
All other bending tests were performed with a PMMA coated tip to minimize the
damage to the beam. Figure 8.9 shows typical plots of the loading and unloading cycle
for each of the four materials tested. The plots show the expected linear behavior for
138
elastic bending. The slope of the load-displacement plot was determined by a best-fit line
of the loading portion of the plot. The elastic modulus and standard deviation (n=5) for
each material are shown in the figure and are summarized in Table 8.1.
Figure 8.9 Load-displacement plots for beam bending with a maximum load of 10 µN.
139
The elastic modulus from the beam bending can be compared with those obtained
from CSM nanoindentation as well as the bulk values available in the literature. The bulk
elastic modulus for PMMA and polystyrene were found to be 3.1-3.3 GPa11,12 and 3.2-3.4
GPa11, respectively. These values are shown in Table 8.1. The values for PPMA and
PS/Clay were not available in the literature. The results show fairly good agreement
between the modulus obtained from all three sources. This suggests several things. First,
the variation in the data is attributed to differences between the experimental methods
and not to variations in the behavior of the materials at the microscale. In other words,
with respect to the elastic modulus, the materials seem to behave according to the bulk
properties at this size scale. Moreover, the consistency of the beam bending tests seems to
validate this method as a legitimate means to characterize the elastic modulus of
microscale polymer structures. Some error may be associated with assumptions on the
boundary conditions as well as some plastic deformation of the beam surface by the tip.
However, these errors do not have a prohibitive effect on the overall measurements.
Polymer Elastic Modulus (GPa)
Bulk (literature) Nanoindentation Normal beam bending PPMA 1.7 0.7 PMMA 3.1-3.311,12 4.8 1.9 PS 3.2-3.411 3.6 1.9 PS/Clay 5.1 4.6
Table 8.1 Summary of elastic modulus measured by nanoindentation and microbeam bending and comparison with bulk values.
140
8.3.4 Effects of aqueous environment and temperature
The materials were tested after soaking in an aqueous environment and at human
body temperature to determine if the materials behave differently under simulated
physiological conditions. Any major differences in the behavior have a significant impact
on the application of these materials for biological purposes. The materials were tested by
both CSM nanoindentation and beam bending. The results are summarized in Figure
8.10. No significant differences were seen in the properties of PMMA, PS, or PS/Clay in
response to aqueous environment or elevated temperature. PPMA showed a decrease in
elastic modulus in response to both soaking and elevated temperature. This is not
unexpected, given the lower glass transition temperature and more open molecular
structure of PPMA relative to the other materials. While PPMA may still be a potentially
useful material in polymer MEMS devices, the operating conditions must be considered
in the design and implementation of PPMA based devices.
141
Figure 8.10 Hardness and elastic modulus of PPMA, PMMA, PS, and PS/Clay in response to soaking in aqueous environment and temperature measured by (a) CSM
nanoindentation and (b) beam bending.
142
8.3.5 Yield and breaking strength from normal beam bending
Beam bending tests were performed at higher loads (3 mN and 10 mN) to
evaluate both the yield strength and breaking strength of the materials. These experiments
were also valuable for evaluation of the failure mechanisms of the polymer
microstructures. Figure 8.11 shows load-displacement plots for beams tested at 3 mN and
10 mN. Arrows in the plots indicate the yield point and breaking point. The plots at 3 mN
show the obvious onset of yield as the load-deflection plot becomes nonlinear. At this
load, however, only PS/Clay failed completely as indicated in the figure. At 10 mN, all of
the beams broke as indicated by a gap in the load-displacement plot. As expected, when
the beam breaks, there is no longer a force being applied to the beam, which in turn leads
to a gap in the data collection. The spike in the load-displacement plot after the beams
break results from the nanoindenter tip becoming wedged in the channel. The data and
SEM images show one significant difference in the failure between the materials: PPMA,
PMMA, and PS all exhibited significant plastic deformation prior to total failure,
indicating a ductile failure, while PS/Clay, although also exhibiting some plastic
deformation, showed a more brittle failure.
143
Figure 8.11 Load-displacement plots and SEM images for bending tests at 3 mN (top) and 10 mN (bottom).
144
8.3.6 Lateral beam bending
Results of lateral force bending tests and SEM images of the cantilevers after
testing are shown in Figure 8.12. The load-displacement plot is in general agreement with
the expected behavior of the materials. A linear region of the plot is followed by the onset
of yield. For PS, a decrease in the load following yield indicates cracking of the beam.
For PS/Clay, the gap in the data indicates breaking of the beam. While this new method
for evaluating materials shows some promise for materials evaluation, the results from
the tests do not match well with data obtained via other methods. Therefore, the results
from these tests are only considered to be qualitative. However, proof of principle for this
method of testing is established, and further refinement of this technique could make it a
valuable tool for evaluating materials and devices that are subjected to lateral forces.
Figure 8.12 Load-displacement plots and SEM images of PS and PS/Clay cantilever measured by lateral bending.
145
8.4 Conclusion
Mechanical properties of polymer microstructures and polymer films were
evaluated using several different techniques. CSM nanoindentation of supported polymer
microstructures was used to measure hardness, elastic modulus, and creep. The scratch
resistance of the materials was evaluated with the nanoindenter scratch testing with the
aid of SEM imaging. Beam bending experiments were performed on suspended polymer
microbridges and the results were used to calculate the elastic modulus, yield strength,
and breaking strength of the microstructures. Aside from breaking strength, the property
measurements were similar to bulk materials properties obtained from the literature.
Hardness and elastic modulus were evaluated under simulated physiological conditions.
Only PPMA showed a change in mechanical properties under these conditions. Lateral
force bending of the beams was demonstrated as a potential method for evaluating
materials, but further refinements to the technique are need in order to obtain meaningful
quantitative property measurements.
References 1 J.A. Forrest, D. Dalnoki-Veress, “The glass transition in thin polymer films.” Advances in Colloid and Interface Science 94 (2001) 167-196. 2 J.L. Keddie, R.A.L. Jones, R.A. Cory, “Size-dependent depression of the glass transition temperature in polymer films.” Europhysics Letters 27 (1994) 59-64. 3 B. Bhushan, A.V. Kulkarni, W. Bonin, J. Wyrobek, “Nano/picoindentation measurement using a capacitance transducer system in atomic force microscopy.” Philosophical Magazine A 74 (1996) 1117-1128. 4 S.J. Bull, “Nanoindentation of coating.” Journal of Physics D 38 (2005) R393-R413. 5 A.C. Fischer-Cripps, “Critical review of analysis and interpretation of nanoindentation test data.” Surface and Coatings Technology 200 (2006) 4153-4165.
146
6 X. Li, B. Bhushan, “A review of nanoindentation continuous stiffness measurement technique and it applications.” Materials Characterization 48 (2002) 11-36. 7 I.H. Loh, M.S. Sheu, A.B. Fishcer, Biocompatible polymer surfaces, in Desk Reference of Functional Polymers: Synthesis and Applications (edited by R. Arshady) American Chemical Society, Washington, D.C. (1997). 8 D.B. Holt, P.R. Gauger, A.W. Kusterbeck, F.S. Ligler, “Fabrication of a capillary immunosensor in polymethyl methacrylate.” Biosensors and Bioelectronics 17 (2002) 95-103. 9 A. van der Berg (ed.), Lab-on-a-chip Chemistry in Miniaturized Synthesis and Analysis Systems, Elsevier, Amsterdam (2003). 10 B. Ellis, Polymers: A Properties Database, Available on compact disk, CRC, Boca Raton, FL (2000). 11 J. Brandrup, E.H. Immergut, E.A. Grulke (eds.), Polymer Handbook, Wiley, New York, NY (1999). 12 J.E. Mark, Polymer Data Handbook, Oxford University Press, Oxford (1999). 13 T.G. van Kooten, H.T. Spijker, H.H. Busscher, “Plasma-treated polystyrene surfaces: model surface for studying cell-biomaterial interactions.” Biomaterials 25 (2004) 1735-1747. 14 M. Alexandre, P. Dubois, “Polymer-layered silicate nanocomposites: preparation, properties and uses of a new class of materials.” Materials Science and Engineering 28 (2000) 1-63. 15 S.S. Ray, M Okamoto, “Polymer/layered silicate nanocomposites: a review from preparation to processing.” Progress in Polymer Science 28 (2003) 1539-1641. 16 J. Njuguna, K. Pielichowski, “Polymer nanocomposites for aerospace applications: properties.” Advanced Engineering Materials 5 (2003) 769-778. 17 F. Gao, “Clay/polymer composites: the story.” Materials Today 7 (2004) 50-55. 18 G. Beyer, “Nanocomposites: a new class of flame retardants for polymers.” Plastic Additives and Composites 4 (2002) 22-28. 19 M. Kurian, M.E. Galvin, P.E.. Trapa, D.R. Sadoway, A.M. Mayes, “Single ion conducting polymer-silicate nanocomposite electrolytes for lithium battery applications.” Electrochimica Acta 50 (2005) 2125-2134.
147
20 R.Y. Lockhead, C.R. Haynes. S.R. Jones, V. Smith, “The high throughput investigation of polyphenolic coupler in biodegradable packaging materials.” Applied Surface Science 252 (2006) 2535-2548. 21 H. Liu, B. Bhushan, “Investigation of nanotribological properties of self-assembled monolayerss with alkyl and biphenyl spacer chains.” Ultramicroscopy 91 (2002) 185-202. 22 W.C. Oliver, G.M. Pharr, “An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments.” Journal of Materials Research 7 (1992) 1564-1583. 23 B. Bhushan, X. Li, “Nanomechanical characterization of solid surfaces and thin films.” International Materials Review 48 (2003) 125-164. 24 H. Liu, B. Bhushan, “Investigation of nanotribological properties of self-assembled monolayers with alkyl and biphenyl spacer chains.” Ultramicroscopy 91 (2002) 185-202. 25 W.C. Young, R.G. Budynas, Roark’s Formulas for Stress and Strain, Wiley, New York, NY (2002).
148
CHAPTER 9
DESIGN, SIMULATION, AND FABRICATION OF A POLYMER SENSOR FOR MEASURING SINGLE CELL FORCES
9.1 Introduction
One particular area where polymer MEMS could be effectively employed is in the
field of cellular biomechanics. When cells adhere to the extracellular matrix or to a
foreign material, mechanical forces are generated at the adhesive contacts. Following
adhesion, cytoskeletal organization occurs, and molecular linkages between the adhesion
contacts and cytoskeletal components transmit forces through intracellular stress fibers
and adhesion sites to the underlying substrate.1
One of the key barriers to studying cellular mechanics is a lack of sensitive
measurement devices with the ability to accurately evaluate static and dynamic cellular
forces. A few MEMS and non-MEMS devices have been explored for measuring cellular
adhesion2,3, traction forces4,5, cell contractile forces6, and intracellular mechanics7.
Polymer MEMS offer several advantages over current devices due to their inherent
biocompatibility, high sensitivity based on material properties, and low cost fabrication
methods.
We set out to design a polymer sensor to measure forces generated by single
adherent cells. The device is based on an array of polymer microcantilevers. When cells
149
attach to the cantilevers, forces are transmitted to the beams resulting in deflection. By
measuring the beam deflection and calculating the force-deflection response of the beam,
the force vector can be calculated.
Design and simulation are critical steps in the device development process. Proper
design and accurate simulation of the device behavior can be performed prior to
fabrication of a prototype device. An adequate understanding of the key design
parameters will increase the likelihood that the device will provide the necessary
functionality in practice. Well thought out simulations of the device behavior will ensure
that the device functions in the expected manner. Both of these steps ultimately lead to
fewer design iterations, decreased cost, faster device development, and improved
functionality.
The design and simulations of the polymer MEMS cell force sensor will be
discussed here. A general overview of the device will be given along with a detailed
discussion of the parameters that were considered in the design. The force-deflection
behavior of the device was modeled using finite element analysis. Improvements in the
device design for second and third generation devices will be discussed with respect to
improving the force measurement capabilities as well as improving the biological
interactions between the cell and the device. The data analysis process for converting
experimental beam deflection data into force vector information will also be discussed.
150
9.2 Materials and methods
9.2.1 Solid modeling for device visualization
SolidWorks three-dimensional computer aided drafting (CAD) software was used
to visualize the device during the design process. This allowed for visualization of
different design iterations to consider possible problems with respect to device
fabrication and function. Additionally, CAD drawings could be integrated directly into
the modeling and fabrication processes. SolidWorks data could be imported into the finite
element simulation package and used directly in the photomask fabrication process.
9.2.2 Finite element analysis
The response of each prototype design was simulated using ANSYS finite
element analysis software. A single probe was isolated and analyzed by applying a range
of forces to the end of the cantilever beam and observing the deflection simulations.
Several simulation parameters were considered: changes in the simulations with respect
to the mesh (i.e. solid versus surface mesh), number of mesh elements, boundary
conditions, nature of the applied load, and force application location. Further design
modifications were made based on the simulation results to arrive at a set of initial
designs.
9.2.3 Device development and prototyping process
The process of transitioning from design to prototype began with the photomask
preparation process. This involved first developing the layout of the mask in SolidWorks.
One important consideration in the layout is the number of different designs that are
considered for a single mask. Too few designs can result in a narrow range of operating 151
parameters, and may require new mask production after a short period of time. Too many
designs can be impractical given that a limited number of different devices can be tested
over a given period of time. Another important consideration is the spacing between the
devices. Proper spacing can make device fabrication and the alignment of different
device layers significantly easier. In addition, the number of repeats of each design is
important. This determines the frequency of fabrication runs and impacts the device
fabrication time and cost. After the layout was completed, the SolidWorks file was
exported as a .dxf (AutoCAD file extension) file that could be sent directly to the mask
manufacturer for processing.
9.2.4 Device fabrication
152
The devices were fabricated using sacrificial layer micromolding (SLaM). The
process will be described briefly here, but details of the process can be found in Chapter
7. The fabrication schematic for the sensor is also shown in Figure 9.1. All devices were
initially fabricated from SU8 5 or SU 2005 (Microchem Corp.) with minor changes in the
photoresist thickness depending on the desired beam thickness. Devices were transferred
to PDMS molds using the standard molding process. Polyvinyl alcohol (PVA) sacrificial
layers were produced by photolithography followed by reactive ion etching to expose the
anchor regions of the substrate. Initial devices were fabricated from PPMA. The polymer
was dissolved in anisole (Sigma-Aldrich) at concentrations ranging from 1.2-10%
(wt/wt). The polymer solution was spin coated at onto the patterned PDMS mold at spin
speeds from 2000-6000 rpm for 60 seconds. Resultant device thickness was a function
of the PDMS mold depth, polymer concentration, and spin speed. Devices ranged from
180 nm to 3.7 µm based on the above parameters. To remove the surface polymer, the
mold was brought into contact with a glass slide heated to 175-185 °C. The mold was
held on the slide for ten seconds under the pressure of its own weight and then removed.
The process was repeated if necessary to completely remove the surface PPMA, leaving
the polymer only in the recessed features. The selectively patterned PDMS mold was then
manually aligned with the sacrificial layer under an optical microscope at 40x
magnification. The substrate was heated to 95º C and a pressure of 30 lb/in2 was applied
to the backside of the mold and held for 10 seconds.
Later generations of the device were fabricated from polystyrene. In this case, a
7.5% (wt/wt) solution was spin coated on the mold at 3000 rpm. The first stamping was
done at 200 ºC and the temperature of the second stamp was 120 ºC and the pressure was
increased to 70 psi. The polystyrene devices were also annealed at 115 ºC for 15 minutes
to reduce any residual stress in the beams.
Figure 9.1 Fabrication of force sensors by sacrificial layer micromolding; (a) Sacrificial layer patterning process, (b) structural layer molding, and (c) alignment and bonding.
153
9.2.5 Device characterization
The thickness of each device was measured individually after testing. The
thickness was measured using either atomic force microscopy (AFM) or a stylus
profilometer (Dektak, Series 3). Scanning electron microscopy and optical microscopy
were used to image the devices.
9.3 Results and discussion
9.3.1 Design considerations
Several factors were taken into consideration in the design of the sensor. Factors
related to function, fabrication, materials selection, and cost were all considered
individually or in combination to determine a set of designs for the initial fabrication run.
In term of function, it was necessary to design a device capable of measuring forces with
high sensitivity and high resolution. The expected force range for most cells is in the
range of hundreds of nanonewtons (nN) and dynamic changes in these forces are often
small (on the order of tens of nN) over short periods of time. The device must be capable
of measuring forces and changes in force on this scale. Spatial resolution of the force
measurements was also considered. A sufficient number of measurement locations on
each cell is advantageous in providing information about the spatial distribution of forces.
Material selection was a critical parameter in the function, biological interactions,
and cost. Polymers offer some particular advantages over traditional MEMS materials.
The low elastic moduli of polymers are advantageous for low magnitude force
measurements. Cantilevers made from low modulus materials provide larger deflections
per unit force, thus increasing the sensitivity of the measurements. Many polymers are
154
also biocompatible and are currently used in a number of implant and in vitro cell biology
applications. This is particularly true with polystyrene, which is the most common
material for in vitro cell culture. Polymers are generally less expensive than other MEMS
materials for both the material and processing costs. Most of the fabrication techniques
described here can be performed outside of a cleanroom setting without the expensive
processing equipment needed for purely silicon based processing.
The beam geometry was considered in the overall cost of device development.
One of the most costly steps in the process is mask production. The overall cost of the
mask is determined primarily by the smallest feature size. Masks with a minimum feature
size of 5 µm or larger are significantly less expensive than masks with features smaller
than 5 µm. This was taken into consideration relative to the width of the beam and the
spacing between the cantilevers. A beam width of 5 µm was used and a minimum spacing
between cantilevers was also 5 µm.
The design also took into consideration the evaluation of future devices with
increased levels of sophistication. It was necessary to use a generalized fabrication
protocol that could be easily adapted to different materials, device geometries, and
applications. The ability to add higher-level functions to the basic device was also
considered. The first several generations of the devices are passive structures that respond
to the cell. The ability to incorporate fluidics and actuation mechanisms into an active
device could significantly increase functionality.
9.3.2 Device overview
155
The device is based on an array of compound polymer cantilever beams. The
device consists of three primary functional regions. These are shown in the first
generation device model in Figure 9.2. The first region is the cell active area. This is the
portion in the center of the device where the cell is located for the measurements. When
the cell attaches and spreads on the cell active area, forces are transmitted to the
underlying cantilever beam array. This leads to deflection of the beams proportional to
the magnitude of the applied force. The beam deflections are captured over time using a
CCD camera attached to an inverted phase contrast microscope. The force vector is then
calculated from the deflection data. The second region of the device is the cantilever
array. This consists of a set of suspended cantilever beams that span the distance between
the cell active area and the outside edge of the device. The cantilevers are configured
such that forces can be measured in any direction in the plane of the structure. The shape
of the beam is designed to give a specific force-deflection behavior as guided by finite
element simulations. The third portion of the device is the anchor region, where the
device is fixed to the underlying substrate.
Figure 9.2 Generation 1 cell force sensor with labeled functional regions.
156
9.3.3 First generation cell force sensor
The first generation of the device consisted of an array of compound cantilevers
with a single bend in the beam (L-beam) that provides the ability to measure force in
multiple directions. Two configurations of the first generation device were considered.
The circular configuration is shown in Figure 9.2 and the linear configuration is shown in
Figure 9.3. In the linear configuration, the beams are fixed at anchor region of the device
and converge in the cell active area. For the linear configuration, the beams are fixed in
two parallel sets of beams on either side of the device. The circular configuration is well
suited for essentially static cells, while the linear configuration was designed to allow
measurements on motile cells.
Figure 9.3 Linear configuration of the first generation cell force sensor with labeled functional regions.
The force-deflection behavior of the first generation device was simulated using
finite element analysis. Inputs to the finite element simulations included the geometry of
157
the structure and the mechanical properties of the structural material. The geometry of the
device was imported directly from the SolidWorks models, with the exception of the
beam thickness. A discussion of the method to correct for the thickness of the beam is
given in Section 9.3.6. The elastic modulus is also important for simulating the force-
deflection behavior. For PPMA devices, a modulus of 2.0 GPa was used. This value was
estimated based on being in the middle range between the values for the elastic modulus
obtained from normal beam bending and nanoindentation experiments (see Chapter 8).
Force inputs were applied to the beams in 30˚ increments over 360˚ as shown in Figure
9.4. Forces at three magnitudes (5 nN, 10 nN, and 20 nN) were applied to a circular area
in the center of the pad at the end of the cantilever. The x,y deflection profile for all three
forces are shown in Figure 9.5.
Figure 9.4 (a) Image of the L-beam cantilever used in the FEA simulations, (b) close up view of the end of the beam with the direction-angle convention.
158
The deflection profile shows several key features. Ideally the x,y deflection would
produce a circular profile with the direction of the deflection being in the same direction
as the force. The plot in Figure 9.5 shows a skewed elliptical deflection profile, indicating
that the beam responds non-linearly to forces applied in different directions. The
consequence is that the device has different resolution and sensitivity depending on the
direction of the deflection. For deflections in direction 6 and 12, the device has higher
sensitivity and gives a larger deflection for a given force. However, there is lower
resolution in determining the direction of the force. The opposite is true for deflection
corresponding to directions 3 and 4 (60˚-90˚) and directions 9 and 10 (240˚-270˚). In
these general directions, the device has lower sensitivity and gives smaller deflection for
a given force, but the ability to resolve the direction of the force improves. For the
direction convention, refer to Figure 9.4. In addition, the plot shows a significant
difference between the direction of the deflection and the direction of the force. This
complicates determination of the force vector direction from the deflection data. A
detailed discussion of determining the force vector magnitude and direction is given in
Section 9.3.6. The plot also shows the expected linearity between the magnitude of the
force and the magnitude of the deflection, as indicated by the concentricity and spacing
of the ellipses.
159
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
-3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
y-deflection (µm)
x-de
flect
ion
(µm
)
5 nN
15 nN
30 nN
Position 1
Position 4Position 7
Position 10
Figure 9.5 Deflection plot for the first generation cell force sensor.
The finite element simulations were also used to determine if the devices behave
in the linear elastic regime for the expected range of forces. Estimates of the yield
strength for PPMA and polystyrene can be found in chapter 8. The values for maximum
stress and strain found in the FEA simulations were well below even more conservative
estimates of stress and strain needed to induce yield of the materials.
Despite some complications in the L-beam design, this configuration still
provides adequate force-deflection characteristics to allow for meaningful measurement
to be taken. The first fabrication run was performed with the first generation device in
both the circular and linear configurations. Devices with 4, 6, 8, 10, and 12 probe circular
configurations and a linear device with 14 cantilevers were chosen for the initial
fabrication run.
160
Fabrication results for the first generation sensor are shown in Figure 9.6. The
SEM images in Figure 9.6 are shown prior to release of the sacrificial layer. An optical
micrograph of the device immersed in phosphate buffered saline after release is shown in
Figure 9.7(a). A microprobe attached to a micromanipulator was used to apply a force to
the beam to induce a deflection, showing that the beam is free to move and that the beam
returns to its initial location after the probe is removed as shown in Figure 9.7(b,c).
Figure 9.6 SEM micrographs of the (a) circular 12 probe device and (b) linear 14 probe device.
161
Figure 9.7 Optical micrographs of the circular 4 probe device (a) before bending with a microprobe, (b) during bending, and (c) after removing micromanipulator tip.
9.3.4 Second generation cell force sensor
Given the limitations in the L-beam design and lessons learned from initial
fabrication and biological testing with the first generation device, a second generation
device was developed to improve the force-deflection response of the beam, make
fabrication easier, and improve the cell-sensor interactions. The geometry of the
compound cantilever was redesigned to provide a more uniform deflection profile.
Several lessons learned from the L-beam were taken into consideration for the new
design. Most notably, the highest deflection per unit force was in the direction
perpendicular to the line between the fixed base of the beam and the end of the beam as
shown in Figure 9.8. While this line provides the highest sensitivity, it also gives the
162
largest offset between the direction of the deflection and the direction of the force.
Therefore, by combining two L-beams into a single beam, with specific angles between
the fixed base and the ends of the beams, it was possible to decrease the offset between
the direction of the deflection and the direction of the force. In addition, this indicates a
more homogeneous sensitivity. This was the guiding principle in the development of the
second generation beam design.
Figure 9.8 L-beam showing the line of highest sensitivity.
Several other factors were taken into consideration for designing the second
generation device. First, geometrical constraints imposed by the circular configuration
affects the ability to increase the number of cantilevers on the device without physical
interference (contact) between the beams. During the iterative design process, it was also
noted that the geometry of the bends in the cantilever affects of the response. This is
shown in Figure 9.9. By reducing the width of the beam and the bends, it is possible to
produce a hinge-like structure and reduce the non-linearity in the deflection response.
163
However, this is constrained by the width of the beam. It is advantageous to maintain the
5 µm minimum feature size to keep the cost of the device low.
Figure 9.9 Image of the beam joint showing the material that was removed to create a hinge-like bend.
With respect to the fabrication and biological aspects of the device, it was noted
in the first generation that the width of the anchor region small enough that it was
difficult to align the structural and sacrificial regions of the device. The width of the
anchor region was increased from 50 µm to 130 µm to minimize this issue. It was also
noted that it was difficult for the cells to attach and spread on the relatively small surface
area of the first generation devices. The cell active area of the second generation devices
was increased significantly to allow the cell to spread more naturally on the device.
The final geometry for the second generation beam is shown in Figure 9.10(a).
Figure 9.10(b-e) shows how the new beam design fits into a 4, 6, 8, and 10 probe systems
without interference between the beams. The images also show the increase in the area of
the anchor region to decrease alignment problems.
164
Figure 9.10 Images of (a) final design of the second generation compound beam (b) 4 probe device (c) 6 probe device, (d) 8 probe device, and (e) 10 probe device.
165
Figure 9.11 shows an example of the progression of the finite element simulations
for the second generation device. The geometry of the beam was first imported from
SolidWorks. The boundary conditions were established by placing a fixed support at the
end of the beam and the load was applied to top surface of the free end of the beam
[Figure 11(a)]. The mesh was then defined as shown in Figure 9.11(b). After imputing
the material properties, the simulation was run to solve for the x and y deflections of the
beam [Figure 11(c)]. The angle of the force input was then changed in 30º increments and
the simulation and analysis was repeated over 360º. The deflection of the end of the
cantilever was analyzed and used to create the deflection plots.
166
Figure 9.11 FEA simulations; (a) boundary condition and applied force, (b) mesh, and (c) deflection solution.
The x,y deflection plot for the second generation beam is shown in Figure 9.12.
The plot shows the profile for 5 nN, 10 nN, and 20 nN forces. The plot shows that there is
significantly less offset between the deflection direction and force direction. While the
plot is still elliptical, it is much closer to circular than the first generation device. This
gives a more uniform sensitivity to forces in any direction. A comparison between the
first and second generation beams is shown in Figure 9.13. The plot shows the
differences in the deflection profile for a 5 nN force. Improvements to the offset and the
uniformity are obvious. Sensitivity of the device is reduced in the direction corresponding
to 0º and 180º relative to the sensitivity of the first generation device. The overall
uniformity in the sensitivity is improved, thus providing better measurement of forces
independent of direction. The deflection profile for the second generation device also
improves the ability to resolve the force direction.
-1
-0.5
0
0.5
1
-1 -0.5 0 0.5 1
y-deflection (µm)
x-de
flect
ion
(µm
)
5 nN10nN20 nN
Pos.1
Pos.4
Pos.7
Pos.10
Figure 9.12 Deflection plot for second generation compound beam. 167
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
-0.75 -0.50 -0.25 0.00 0.25 0.50 0.75
x-de
flect
ion
(µm
)
y-deflection (µm)
Initial L-beamresponse
secondgeneration beamresponse
Figure 9.13 Comparison of the beam deflection response for the L-beam and second generation beam designs.
Fabrication results for the second generation devices are shown in the Figure 9.14.
The SEM micrographs show PPMA sensors with (a) 4, (c) 6, and (e) 10 probe
configurations. Figure 9.14(b,d,f) shows SEM micrographs of devices fabricated from
polystyrene. Polystyrene replaced PPMA as the structural material in order obtain a well-
understood and biocompatible surface for cell attachment. Figure 9.15 shows phase
contrast micrographs of immersed polystyrene sensors after release of the sacrificial
layer.
168
Figure 9.14 SEM micrographs of cell force sensors; (a,b) four probe sensors, (c) six probe sensor, (d) eight probe sensor, (e,f) ten probe sensors. The devices in the left
column are made from PPMA and the right column are made of polystyrene.
169
Figure 9.15 Phase contrast optical micrographs of (a) 4 probe sensor, (b) 10 probe sensor, and (c) 6 probe sensor after releasing the sacrificial layer in aqueous solution.
9.3.5 Third generation cell force sensor
Improvements to the sensor design for the third generation were primarily driven
by improving the biological interactions between the cell and the device. Initial
experiments with the second generation device revealed that significant improvements
could be made to improve the cells’ ability to attach, spread, and function normally on
the devices. Changes were made primarily to the cell active area. First, an adhesion pad
was placed in the center of the device and fixed to the substrate in the same manner as the
anchor region. This adhesion pad was designed to provide the cell with a surface for
initial attachment. The size of the pad was determined based on cell area measurements
taken from two cell lines of interest: HT 1080 fibrosarcoma cells and 3T3 mouse
170
fibroblasts. The pad area was designed to provide a surface for cell attachment by also
allow the cell to bridge between the pad and the cantilevers during cell spreading. The
area of the end of each cantilever was also increased to provide a larger area for the cell
to spread. It should be noted that all of the final experiments were performed with WS-1
skin fibroblasts instead of either of the cell lines that were analyzed. However, these
measurements gave a good estimate of the area needed to promote healthy cell behavior
on the sensor. Cell area measurement for HT 1080 and 3T3 cells are shown in Figure
9.16. The projected cell area for unattached cells and for cells after 12 hours of
attachment are given. Based on these measurements, the areas of the adhesion pads that
were chosen for the fabrication run were 400 µm2 (4 probe sensors-design 1), 420 µm2 (6
probe sensor-design 1), and 625 µm2 (4 and 6 probe sensors-design 2).
0
100
200
300
400
500
600
700
800
900
t=0 min. t=720 min.
Proj
ecte
d C
ell A
rea
(µm
2)
HT10803T3
Figure 9.16 Projected cell area for HT1080 and 3T3 cells before attachment (t=0 min.) and after spreading (t=720 min.).
171
Even though the basic geometry of the beam was not changed for the third
generation sensors, the increase in the size of the pad led to a decrease in the overall
length of the 5 µm wide section of the beam. This had a significant impact of the force-
deflection behavior, making it necessary to perform new FEA simulations and produce
new force-deflection plots for the third generation device. It should be noted that the
angle convention was changed for the simulations of the third generation device. The
angle convention is shown in Figure 9.17. The change in the angle convention only
changes the appearance of the profile (i.e. the orientation of the profile). It is critical,
however, that the same convention is maintained when experimental data is analyzed.
Figure 9.17 Angle convention for third generation device simulations.
Additional FEA simulations were run to determining the effects of the area and
location of the applied load, the location on the beam that the deflection is analyzed, and
the pad area on the deflection profile. Four different loading scenarios were tested as
shown in Figure 9.18. The figure shows the load applied to a (a) 5 µm diameter circular
area at the center of the pad, (b) 5 µm diameter circular area at the end of the pad, (c) line
at the end of the pad, and (d) the area of the front half of the pad. The difference in the
172
deflection profile for each of the different loading conditions is shown in Figure 19. The
plot does show a small difference in the behavior depending on the loading conditions. It
was determined that the load of the front half of the pad area would be used in the
analysis. This condition was chosen for two reasons. First, this loading condition seems
to fit most appropriately with the actual behavior of a cell on the device. Second, this
produces a response in the middle of the two extremes, thus providing a reasonable
estimation of the loading.
Figure 9.18 FEA loading conditions; (a) load applied to a 5 µm diameter circle at the center of the pad, (b) load applied to a 5 µm diameter circle at the center of the pad, (c)
load applied to the front edge line of the pad and (d) load applied to the front half area of the pad.
173
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
-0.2 -0.1 0 0.1 0.2
x-deflection
y-de
flect
ion
0.2
(a)(b)(c)(d)
Figure 9.19 Deflection plot for four different loading conditions; (a),(b),(c), and (d) correspond to the loading conditions in Figure 9.18.
The location on the cantilever where the deflection was measured was also tested
to determine if different locations affect the profile. The left and right side of the tip of
the cantilever were tested. These locations were chosen because they give the largest
displacement for a given force. The locations that were tested are shown in Figure 9.20(a)
and the difference in the profile is shown in (b). The plot shows that there is a difference
in the profile, with the right side measurement showing more deflection offset. Therefore,
the left side displacement was used for simulations and for all experimental
measurements.
174
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
x-deflection
y-de
flect
ion
Right SideLeft Side
(b)
Figure 9.20 (a) Convention for location of deflection analysis, (b) deflection plot for deflection analysis on the right and left side of the probe.
The area of the pad was also tested to determine if this had an effect on the
profile. This is important because the individual pad area is different for the 4 and 6
probe configurations. The profiles showed that the area of the pad did not have a
significant effect on the profile, provided that the length of the 5 µm wide section of the
beam was maintained constant.
A summary in the difference between the deflection plots for the first, second, and
third generation devices are shown in Figure 9.21. The angle conventions were
transformed into the convention of Figure 9.17 to allow for direct comparison of the
profiles. The first generation device showed significant angular offset between the force
175
direction and the displacement direction. The flat elliptical nature of the profile leads to a
higher device sensitivity in certain directions relative to others as well as lower resolution
in determining the force direction. The second generation device shows a significant
improvement in the profile, with better uniformity in the sensitivity for force in any
direction, decreased angular offset between the force and displacement direction, and
improved resolution in determining the force direction. The third generation device
shows that the overall beam is stiffer based on shortening the length of the 5 µm wide
portion of the beam. The profile shows even less angular offset between the force and
displacement directions based on choosing the most appropriate location to analyze the
displacement.
-0.400
-0.300
-0.200
-0.100
0.000
0.100
0.200
0.300
0.400
-0.600 -0.400 -0.200 0.000 0.200 0.400 0.600
x-deflection
y-de
flect
ion
Gen 1
Gen 2
Gen 3
Figure 9.21 Comparison of the deflection plot for first, second, and third generation cell force sensors.
176
Fabrication results for the third generation device are shown in Figure 9.22. A
lower magnification image of the overall device is shown in Figure 9.22(a). A higher
magnification SEM image (b) shows the cell active area with the increase in the size of
the beam ends and the inclusion of the central adhesion pad.
Figure 9.22 SEM micrographs of the third generation sensor at (a) low magnification and (b) high magnification.
9.3.6 Determining the force vectors from displacement data
To this point, the deflection profiles have been given only for a few specific
forces. For practical experimental purposes, it is necessary to calculate the force
magnitude and force direction from deflection measurements over a wide range of
magnitudes and directions. This was done in three basic steps. First, the x,y deflection
plot was made from FEA simulations at 10º increments in the force direction in the same
manner described previous. The deflection plot is given in Appendix A. The x,y
deflection data points were used to calculate the deflection angle for each simulated force
direction. The force angle was then plotted versus the beam angle. Parabolic curves were
fit to the data over a range of angles that provided a R2 value for the curve fit > 0.998. An
177
example for beam angles between 1.6º and 27.6º is shown in Figure 23. The plots for the
remaining angles are given in the Appendix A. This provided a set of equations to
transform the experimentally determined beam deflection angle into the force direction
angle. The set of equations is also available in Appendix A. The calculated force
direction is then used in the next step to calculate the magnitude of the force.
y = -0.055914x2 + 3.937841x - 6.182234R2 = 0.999931
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30
Beam Deflection Angle (°)
Forc
e A
ngle
(°)
Figure 9.23 Plot of force angle versus beam deflection angle from FEA simulations. The experimentally determined beam angle can then be used to calculate the force angle.
The next step in the process was to calculate the magnitude of the simulated beam
deflection by simply taking the sum of the squares of the x,y deflection for a specific
force magnitude (i.e. 5 nN). The magnitude of the beam deflection was then plotted
versus the force angle for a specific range of force angles. Parabolic curves were fit to
this data such that R2 > 0.998 over the force direction range. An example for force angles
from 0-60º is shown in Figure 9.24. The remaining plots and equations are given in
Appendix A. This gives a set of equations to calculate the expected deflection magnitude
178
per unit force in a given direction. Since the deflection magnitude scales linearly with the
force magnitude, the magnitude of the force can be calculated by taking the
experimentally determined deflection divided by the simulated deflection per unit force.
y = -0.00001578x2 - 0.00026049x + 0.15466464R2 = 0.99959381
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0 10 20 30 40 50 60 70
Force Angle (°)
Bea
m D
efle
ctio
n (µ
m)/5
nN
Figure 9.24 Plot of beam deflection per 5 nN of force versus the angle of the force.
The final step in the process was to correct the force magnitude for other variables
including the elastic modulus and the beam thickness. Since the force magnitude scales
linearly with elastic modulus, it was possible to correct for the elastic modulus by simply
multiplying the magnitude of the force by the ratio of the actual modulus to the modulus
used in the FEA simulation. The force magnitude does not scale linearly with the beam
thickness. To account for this factor, FEA simulations were performed for a range of
179
thicknesses (1-10 µm) at a specific force angle (i.e. 0º). The magnitude of the deflection
was then calculated from the x,y deflection data. The magnitude of the deflection was
then normalized to the deflection at a beam thickness of 3 µm. The normalized deflection
was then plotted again the beam thickness. An exponential curve was fitted to the data.
The equation for the curve could be used to calculate a correction factor to account to the
beam thickness. The magnitude of the force could then be calculated by dividing the
uncorrected force magnitude by the correction factor.
In summary, a methodology for calculating the force magnitude and direction
from experimental deflection data was given. This is done by first calculating the
direction of the force from the experimentally determined beam deflection angle. The
force magnitude is then calculated from the previously determine force angle, the
experimentally determined beam deflection magnitude, and the expected deflection per
unit force at a given force angle. The force magnitude was then corrected for elastic
modulus and beam thickness. The beam thickness correction factor was determined by
running FEA simulations over a range of beam thicknesses.
All of the equations for calculating the force vector and the correction factors
were integrated into an Excel spreadsheet. The spreadsheet was configured such that the
angle of the probe deflection and the magnitude of the deflection could be input into the
spreadsheet. Then for a given material and beam thickness, the angle and magnitude of
the force were automatically calculated. The layout of the spreadsheet is given in the
Appendix A.
180
9.4 Conclusion
A cantilever sensor for measuring forces generated by single adherent cells was
designed with the aid of SolidWorks 3-D CAD software and ANSYS finite element
simulations. The design consists of an array of cantilevers with a specific geometry to
allow for flexibility in measuring force in multiple directions with high sensitivity. A
number of other design consideration were also taken into account during the
development process. Some of the other key factors included ease of fabrication, material
properties and biocompatibility, and cost.
Three generations of the devices were designed with functional and biological
improvements being made for each generation. A much improved beam response was
achieved for the second generation device by considering the response behavior of the
first generations beam (L-beam) and performing an iterative design with the aid of FEA.
The third generation device was re-designed with biological improvement being the
primary goal. The simulated behavior of each device was discussed in detail along with
the methodology for determining force vectors from experiment data. Fabrication of each
generation of the force sensor was demonstrated.
References 1 C.G. Galbraith, M.P Sheetz, “Forces on Adhesive Contacts Affect Cell Function.” Current Opinions in Cell Biology 10 (1998) 566-571. 2 A. Yamamoto, S. Mishim, N. Maryuama, M. Sumita, “A new technique for direct measurement of the shear force necessary to detach a cell from a material.” Biomaterial 19 (1998) 871-879. 3 K.A. Athanasiou, B.S. Thoma, D.R. Lanctot, D. Shin, C.M. Agrawal, R.G. LeBaron, “Development of the cytodetachment technique to quantify mechanical adhesiveness of a single cell.” Biomaterial 20 (1999) 2405-2415.
181
4 C.G. Galbraith, M.P. Sheetz. “A micromachined device provides a new bend on fibroblast traction forces.” Proceeding of the National Academy of Sciences USA 94 (1997) 9114-9118. 5 J.L. Tan, J. Tien, D.M. Pirone, D.S. Gray, K. Bhadriraju, C.S. Chen, “Cells lying on a bed of microneedles: An approach to isolate mechanical force.” Proceeding of the National Academy of Sciences USA 100 (2003) 1484-1489. 6 R.M. Hochmuth, “Micropipette aspiration of living cells.” Journal of Biomechanics 33 (2000) 15-22. 7 D.A. Stenger, G.W. Gross, E.W. Keefer, K.M. Shaffer, J.D. Andreadis, W. Ma, J.J. Pancrazio, “Detection of Physiologically Active Compound Using Cell-based Biosensors.” Trends in Biotechnology 19 (2001) 304-309.
182
CHAPTER 10
BIOLOGICAL TESTING OF THE POLYMER MEMS CELL FORCE SENSOR
10.1 Introduction
Recent research has shown that cell forces are critical in regulating a host of
cellular processes1,2,3. Cytoskeletal components are responsible for regulating cellular
shape and spreading. Integrin rich focal adhesions provide a molecular linkage between
the intracellular and extracellular spaces that is imperative for establishing bi-directional
signal transduction pathways2,6. Adhesion forces at the cell membrane-substrate interface
are also responsible for remodeling the extracellular matrix. Transient changes in the
adhesion forces and the intracellular stress field are the primary mechanism of cellular
motility4,5,7. Cellular adhesion forces also have implication in tissue engineering and
biomaterials. The degree and strength with which cells adhere to tissue engineering
scaffolds and implant materials can be the limiting factor in their functional effectiveness.
By employing a polymer MEMS approach to measure cell forces, a better understanding
of the fundamental role of cellular mechanics in regulating cellular processes can be
obtained.
In addition to normal cell function, abnormalities is the cytoskeleton have been
implicated in a wide range of pathologies.8 Cytoskeletal protein abnormalities have been
associated with muscular and neurodegenerative disease, 9,10 cancer,11-13 infectious
183
disease14 and pathologies of the immune systems,15 among others. A better understand of
the cytoskeleton and cytoskeletal mechanics could lead to improved diagnostic and
therapeutic approaches to these diseases. Cytoskeletal proteins, especially actin, also play
a critical role in would healing.16,17
There are already a number of drugs and chemicals that directly target
cytoskeletal components. Prime examples are chemotherapeutics that target the
microtubule network to impede cell proliferation. Paclitaxel (Taxol) is a commonly used
drug for treatment of ovarian, breast, and lung cancer. Paclitaxel stabilizes the
microtubule network, thus inhibiting cell division.18 Another set of microtubule targeting
anticancer drugs are the vinca alkaloids. While they function via a different mechanism
of action than Taxol, this class of drugs also inhibits cell mitosis.19 Other chemicals,
including the cytochalasins20, phalloidon20, and jasplakinolide21, target the actin network
and affect polymerization and depolymerization of actin filaments. The ability to directly
test how these and other drugs affect cell mechanics could improve the therapeutic
effectiveness of the drugs, lead to new applications of cytoskeletal targeting agents in
disease therapy, and potentially provide a new avenue to drug testing and environmental
monitoring.
Here we describe measurement of fibroblast cell forces using the polymer MEMS
sensor described in the previous chapter. Sample preparation, experimental setup, and
testing procedure will be described. WS-1 skin fibroblast cells were tested. Forces were
measured with the second and third generation devices. Third generation devices were
used to measure dynamic force by analyzing time lapse images. The ability to measure
cellular response to chemical perturbations was demonstrated by exposing the cell to
184
cytochalasin-D and measuring the force response over time. This work demonstrates the
ability to use the polymer MEMS force sensor to measure dynamic cellular forces and
measure chemically induced changes in cellular mechanics. This device could find
applications in various cell biology studies, chemomechanical transduction studies, and
drug and chemical testing.
10.2 Materials and method
10.2.1 Sample preparation and surface modification
Initial experiments with PPMA were performed without any surface modification.
After changing to polystyrene, the sample surface was modified with an oxygen plasma
to promote cell adhesion. The devices were expose to O2 plasma in a benchtop reactive
ion etcher (Technics, MicroRIE 800). Samples were exposed for ten seconds with a
power of 100 W, O2 flow rate of 30 sccm, and pressure of 187 mTorr. Sample beams
were characterized by AFM before and after exposure to O2 plasma to determine if the
treatment had a significant effect on the beam thickness or roughness.
After O2 plasma treatment, the samples were placed in PDMS coated petri dishes.
The dishes were coated with a thin layer of PDMS to fix the samples in place during
testing. PDMS was used to avoid using any additional adhesives that could be toxic to the
cells. PDMS was prepared by mixing a 10:1 ratio of PDMS T-2 translucent base and T-2
curing agent. The PDMS was mixed thoroughly and a small amount was placed in the
dish and spread evenly. The PDMS was allowed to cure for a minimum of 48 hours. The
PDMS was sterilized with 70% ethanol for at least 30 minutes before use.
185
The surface modified devices were place in the PDMS coated dishes and cell
culture medium was added to the dish within 30 minutes of modification. The O2 plasma
surface modification is not stable over time, so devices were modified immediately
before testing. The petri dish was placed in the cell culture incubator at 37 ºC to promote
dissolution of the PVA sacrificial layer. After completely dissolving the sacrificial layer,
the cell culture was serially diluted three times. A majority of the medium was aspirated
from the dish and fresh medium was added. This was done to avoid drying the device
completely which would lead to collapse of the beams due to stiction.
10.2.2 Experimental setup
186
The experimental setup for testing the sensor is shown in Figure 10.1. The overall
setup is shown in Figure 10.1(a) and a close-up of the sample stage and micromanipulator
is given in Figure 10.1(b). The key features of the experimental setup are labeled. An
inverted phase contract microscope (Nikon, TS 100) was used for observation. The
microscope was placed on a vibration isolation platform [World Precision Instruments
(WPI)]. A custom built Plexiglas enclosure was fitted to the microscope. The images in
Figure 10.1 are shown with the doors of the enclosure removed to allow the setup to be
seen more clearly. Teflon tubing was used to connect a gas cylinder containing a blood
gas mixture (5% CO2, 20% O2, 75% N2) to a bulkhead fitting on the enclosure. A heat
source was placed in the enclosure to maintain the temperature at approximately 37 ºC. A
thermometer/hygrometer probe was placed in the enclosure to monitor the temperature
and humidity throughout the experiment. A three axis manual micromanipulator (WPI,
KITE-R) was used to place the cell in the cell active area of the device. The
micromanipulator was fixed to the vibration stage using a magnetic holding stand (WPI,
M10L). A 2 µm inside diameter glass micropipette with a Luer fitting (WPI, µTipTM) was
prefilled with cell culture medium using a micropipette filling needle (WPI, MicroFilTM).
The micropipette was attached to a 10 mL syringe via Tygon tubing. The micropipette
was placed in the micropipette holder arm and attached to the micromanipulator. A high
resolution (6.6 megapixel) digital camera (Pixelink, PL-A782) was attached to the
microscope. The camera was attached to a computer with digital image acquisition
software (PixeLINK Capture.OEM).
Figure 10.1 Experimental setup for cell force measurements. 187
10.2.3 Cell culture techniques
WS-1 human skin fibroblasts (ATCC) were used for testing. The cells were
cultured in T-75 or T-25 tissue culture flasks (Fisher Scientific) in Eagle’s Minimum
Essential Medium supplemented with 10% fetal bovine serum (FBS, ATCC) and 1%
penicillin-streptomycin (ATCC). For subculturing, the medium was aspirated from the
flask. Cells were detached with 0.25% trypsin-EDTA (ATCC) at 37 ºC. Fresh medium
was then added to the trypsin-EDTA solution and cells were centrifuged at 1200 rpm for
5 minutes. Trypsin containing medium was aspirated and cells were suspended in fresh
medium. The cell suspension was then split into the appropriate number of culture flasks
(cells were usually split at a 1:4 every 3-5 days). Cells were frozen and thawed as needed.
For sensor testing, the cells were detached from the culture flask in the same manner as
described previously. A small volume of cell suspension (~0.5 ml) was removed and
added to the petri dish with the sensors. The remaining cell suspension was either split or
placed back into a single flask depending on the cell density.
For studies with cytochalasin-D (Cyto-D, Calbiochem), the material was
dissolved in dimethylsulfoxide (DMSO). The cells were allowed to attach and spread on
the sensor as normal. Cyto-D was then injected into the medium via a 20 µl pipette to a
final concentration of 0.1 µM.
10.2.4 Cell placement
188
For testing, it was necessary to place a single cell on the cell active area. This was
done using the micropipette and three-axis micromanipulator. After adding the cell
suspension, the petri dish with sensors was immediately moved to the microscope stage
with the enclosure preheated to 37 ºC and flushed with cell culture gas. The micropipette
tip was place in the cell suspension and moved to a position just above the sensors. This
allowed the device and the micropipette tip to be imaged simultaneously through the
microscope optics. The difference in focal planes between the device and the
micropipette tip was used to estimate the distance between them to avoid damaging the
device with the tip or breaking the tip. After locating the appropriate cell, positive and
negative pressure applied via the syringe was used to move the cell to the center of the
device. The micropipette was then removed from the cell culture solution. After cell
placement, the enclosure was regularly flushed with cell culture gas to maintain the
proper gas concentration.
10.2.5 Image acquisition and analysis
After placing the cell on the device, digital images were captured at specific
intervals to monitor the deflection of the beams over time as the cell attached and spread
on the sensor. Images were captured using the digital image acquisition software. Images
were analyzed using ImageJ software. ImageJ is a free software download from the
National Institutes of Health and allows pixel mapping among a variety of the other
functions. The software can be downloaded from http://rsb.info.nih.gov/ij/.
189
For initial experiments, images were acquired in phase contrast mode. This
provided the best contrast between the device and the cell. However, this complicated
edge detection due to halos at the edges in the phase contrast images. For most of the
experiments with the third generation device, the cell placement was performed in phase
contrast mode then the microscope was switched to brightfield mode to capture the
images for analysis. Phase contract images were only taken for demonstration purposes
(i.e. Figures 10.3 and 10.9).
An initial image was acquired prior to cell attachment to determine the initial
locations of each of the beams. The images were analyzed by determining the x and y
pixel values for the tip of the beams. For the second generation devices, the end of the
arrow shaped pad was analyzed, and for the third generation devices the left side of the
end of the pad was analyzed (see chapter 9). A fixed point on each device was analyzed
to allow correction of any image shift, and a line between fixed points on the images was
also analyzed to ensure that there was no angular shift in the images. The subsequent
images were analyzed in a similar manner to determine the displacement of the beam
relative to the initial position. After correcting for image shift, these values were used to
calculate the pixel displacement magnitude and angle of displacement. The pixel
displacement magnitude was then converted to distance using the calibration factor for
the microscope at a specific magnification. The majority of the images were captured at
100x and the calibration factor was 3.6 pixels/µm. After determining the displacement
vector for each image, the force direction and magnitude were calculated according to the
procedure in Chapter 9.
10.3 Results and discussion
10.3.1 Sample preparation
Figure 10.2 shows the static contact angle of polystyrene films before and after
plasma surface modification. These values are compared with the static contact angle of
tissue culture treated polystyrene. The graph shows that the contact angle of native
polystyrene (94±1°) is much higher than that of tissue culture treated polystyrene
(46±2°). After modification, contact angle of the plasma treated polystyrene (31±2°)
190
surface had a contact angle even lower than tissue culture treated polystyrene. This
surface modification produces a surface similar to that of tissue culture treated
polystyrene and should provide a much more suitable surface for cell attachment and
spreading.
0102030405060708090
100
Polystyrene Tissue CulturePolystyrene
Plasma TreatedPolystyrene
Stat
ic C
onta
ct A
ngle
(°)
Figure 10.2 Static contact angle for native, tissue culture treated, and O2 plasma treated polystyrene.
AFM was used to determine if the plasma etch process had any significant effect
on the surface morphology or the height of the beam. AFM scans were performed on the
same area of the same beam before and after plasma exposure. The stepheight function of
the AFM image analysis software was used to measure the height of the beam before and
after exposure. In both cases, the beam was measured at 2.3 µm. Likely, there is a slight
etch due to the oxygen plasma exposure, but given the short exposure time (10 seconds) it
191
was not enough to produce a significant change in the beam dimensions. The images do
not show any noticeable change in the surface morphology. Overall, the plasma exposure
did not have a significant detrimental effect on the beam geometry.
10.3.2 Experimental results with first and second generation devices
Limitation in the experimental setup and the poor cell interactions with the first
generation device prevented any meaningful measurements from being performed.
However, lessons learned from the first generation device provided the necessary
foundation to improve the beam design, the microscope incubation system, and the image
acquisition setup.
Initial experiments with the second generation device were performed with
unmodified PPMA as the structural material. These experiment were not performed using
the stage incubator described previously. Instead, the samples were placed in a cell
culture incubator between images. This significantly complicated the image analysis
given that large horizontal, vertical and angular image shifts had to be corrected. This
procedure also limited the number of images that could be taken and analyzed. However,
these experiments did provide a proof of concept for the device functionality, as well as
providing further insight into improving the device design, experimental setup, and data
acquisition system. In addition, the first meaningful force measurements were
demonstrated.
192
Figure 10.3 shows phase contrast optical micrographs of a WS-1 skin fibroblast
on the PPMA sensor at times ranging from 0 minutes to 88 minutes. The probes are
labeled 1-4 to aid the discussion. The cell was able to attach and spread on three of the
four probes. At 49 minutes after cell seeding, the cell is still rounded, and there is little
observable deflection in the probes. There is evidence that the cell is beginning to attach
to probe 1 and extend some filopodia in various directions as indicated in Figure 10.3(b).
By 80 minutes, there is significant deflection of the beams. Probes 1 and 4 have been
pulled into contact with one another. There is also significant deflection of probes 1 and
2. By 88 minutes, probe 1 and 2 are also pulled into contact. Probe 3 was never engaged
during this experiment. The end of the cantilever was initially out of focus, which
indicates that it is not in the same focal plane as the other cantilevers or the cell. This is
likely due to damage during the fabrication, release, or cell seeding. Figure 10.4(a) shows
an overlay of the images at time=0 min. and time=88 minutes to show the relative
displacement of each beam more clearly. The results of the displacement analysis and
force calculations are shown in tables 10.1 and 10.2. The resulting forces are on the order
of hundreds of nanonewtons (nN).
193
Figure 10.3 Optical micrographs of WS-1 fibroblast on the second generation sensor at times (a) 0 min. (b) 47 min. (c) 80 min. and (d) 88 min.
Cell Force
Probe Direction (°) Magnitude (nN)
1 261 83
2 212 70
3 - -
4 269 177
Table 10.1 Cell force calculations for time=80 min. for the experiment corresponding to Figure 10.3.
194
Cell Force
Probe Direction (°) Magnitude (nN)
1 317 159
2 270 119
3 - -
4 270 286
Table 10.2 Cell force calculations for time=88 min. for the experiment corresponding to Figure 10.3.
10.3.3 Results with third generation sensors
Changes were made in both the design of the sensor and the material used in third
generation of the device. All devices were made from plasma modified polystyrene. In
addition to improving cell interactions with the device, this material is also much better
characterized with respect to its mechanical properties. This decreases the influence of
error due to estimation of the elastic modulus of the materials. The area of the cell
adhesion region was increased to allow the cell to spread more normally. In addition,
considerable process was made in improving the microscope incubation system and
image acquisition. This combination of improvements made it possible to make
significant process in acquiring meaningful, time resolved cell force measurements.
Figure 10.5 shows time lapse optical micrographs of a WS1 skin fibroblast on the
third generation device. The image in Figure 10.5(a) shows the cell after initial seeding
on the device. The rounded cell is easily visible on the device. At a time of 7.5 minutes,
the cell become less visible as it begins to attach and spread. By 15 minutes after seeding,
195
bending of all four of the beams can be observed. The magnitude of the bending increases
steadily until time 50 minutes, at which point the beams are in contact with one another
or with the adhesion pad.
196
Figure 10.4 Optical micrographs of WS-1 fibroblast on the third generation sensor at times (a) 0 min., (b) 7.5 min., (c) 15 min., (d) 22.5 min., (e) 30 min., (f) 37.5 min., (g) 45
min., and (h) 50 min.
197
The deflection of the beams was analyzed using image analysis. The direction of
the force was then calculated according to the FEA simulations. A plot of the magnitude
of the force versus time is given in Figure 10.5. The plot shows a steady increase in the
force on each beam over time. The plot indicates that the cell attached to the beam
relatively quick, with forces applied to three of the four beams 10 minutes after seeding
and forces on all four beams at 15 minutes after seeding. Forces were highest in beam 1
and reached a magnitude of 879 nN by 50 minutes. The maximum forces for the beams 2,
3, and 4 were 523 nN, 554 nN, and 761 µN, respectively. The maximum force on beams
2, 3, 4 were reach at times of 55, 45, and 50 minutes, respectively. A plot of the force
direction versus time is given in Figure 10.6. The plot shows that the forces are
concentrated in the direction between 80-120º. This indicated that the forces are directed
toward the center of the device for all four probes. This further validates that the
deflection of the beams is a result of contractile force exerted on the beams by the cells.
198
-200
0
200
400
600
800
1000
0 10 20 30 40 50 6
Time (min.)
Forc
e (n
N)
0
Probe 1 Probe 2 Probe 3 Probe 4
Figure 10.5. Force versus time plot for WS-1 fibroblasts on third generation sensor, plot corresponds to the pictures in Figure 10.4.
0
50
100
150
200
250
300
0 10 20 30 40 50 60 7
Time (min.)
Forc
e D
irect
ion
(°)
.
0
Probe 1 Probe 2 Probe 3 Probe 4
Figure 10.6 Force direction plot for WS-1 fibroblast cell corresponding to the force magnitude plot in Figure 10.5.
199
Figures 10.7 and 10.8 show the force direction and magnitude versus time plots
for another experiment. The force magnitude plot has several interesting features. First,
the force plot shows that the cell did not attached to the probe immediately and the cell
attached to the probes at different times. The cell began to exert forces on beam 1 at
approximately 30 minutes. The cell attached to probe 4 at around 40 minutes, probe 3 at
around 50 minutes. The maximum forces for beams 1, 3, and 4 were 1.00 µN, 0.90 µN,
0.16 µN, respectively. The forces on probes 1 and 3 were considerably higher than those
on probe 4. This is likely due primarily to the area of the cell that covered the beam. For
example, if the cell were off-center or only attached to the edge of the beam, one would
expect the forces on that beam to be much lower than if the cell spread over a large area
of the beam surface.
The cell did not attach to probe 2. Even though there are no forces exerted on
probe 2, this does give a good illustration of the noise in the measurements. The positive
and negative values were assigned to give a reference to the direction, with measurements
between 0º and 180º (i.e. toward the center of the device) assigned positive values and
forces between 180º and 360º (i.e. away from the center of the device) assigned negative
values. The plot shows that the force measurements on probe 2 fluctuated between
positive and negative values, further indicating that the measurements are due to noise
and not to actual forces exerted on the beam. The range of measured values was –42 to
+17 nN, so the range of the noise is approximately 60 nN in this case. It should be noted
that the images were captured at 100x magnification. If we assume that the noise is
primarily due to errors in image analysis, the noise range should decrease considerably
with increased image magnification.
200
-200
0
200
400
600
800
1000
1200
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00Time (min.)
Forc
e (n
N)
Probe 1 Probe 2 Probe 3 Probe 4
Figure 10.7 Force magnitude versus time plot for a WS-1 skin fibroblast.
0
50
100
150
200
250
300
350
400
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00
Time (min.)
Forc
e D
irect
ion
(°) .
Probe 1 Probe 2 Probe 3 Probe 4
Figure 10.8. Force direction versus time plot corresponding to the force magnitude plot of Figure 10.7.
201
The force direction plot in Figure 10.8 shows that after cell attachment, the forces
are concentrated toward the center of the device in the range between 70-110º. The plot
also shows that the force direction is essentially random for probe 2 and for the rest of the
probes prior to cell attachment. This is expected given that these measurements are due to
noise in the system.
202
0
200
400
600
800
1000
1200
1400
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00 200.00
Time (min.)
Forc
e (n
N)
Probe 1 Probe 2 Probe 3 Probe 4(a)
0
50
100
150
200
250
300
350
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00 200.00Time (min.)
Forc
e D
irect
ion
(°) .
Probe 1 Probe 2 Probe 3 Probe 4(b)
Figure 10.9 (a) Force and (b) direction versus time plots for a WS-1 fibroblast.
203
-200
0
200
400
600
800
1000
1200
0 50 100 150 200 250Time (min.)
Forc
e (n
N)
Probe 1 Probe 2 Probe 3 Series4(a)
0
50
100
150
200
250
300
350
400
0 50 100 150 200 250Time (min.)
Forc
e D
irect
ion
(°) .
Probe 1 Probe 2 Probe 3 Probe 4(b)
Figure 10.10 (a) Force and (b) direction versus time plots for a WS-1 fibroblast.
Figures 10.9 and 10.10 show force magnitude and direction plots for several other
experiments. In each case, similar behavior is observed with respect to the force
magnitude and direction plots. The point of cell attachment on each probe can be seen by 204
the increase in the magnitude of the force. After a period of random noise in the angle of
the force, the direction stabilizes in the general direction toward the center of the device
as expected given the contractile nature of the forces.
10.3.4 Effects of chemical exposure on cell forces
One of the potential applications of the cell force sensor is in evaluating the
effects of drugs, toxins, and other chemicals on cell mechanics. In order to show that the
sensor is capable of measuring chemically induced changes in the cell forces, the cell was
exposed to cytochalasin-D (cyto-D). Figure 10.11 shows optical micrographs of a cell
that was seeded on the device and allowed to function normally on the device for
approximately one hour. The images show that the cell exerted normal contractile forces
with the beams being deflected in the general direction toward the center of the device.
At 57 minutes after seeding, the cyto-D solution was introduced into the system. After
injection of cyto-D, the images show that the beams began to move apart and away from
the center of the device, indicating a significant decrease in the force applied to each
beam. This is confirmed by image analysis and force calculations as shown in Figure
10.13. The force magnitude plot shows an increase in the forces on each beam over time.
The maximum force on each beam was calculated to be 462 nN, 378 nN, 224 nN, and
123 nN for beams 4, 2, 3, and 1, respectively. After injecting cyto-D, the forces on each
beam decrease significantly. By 150 minutes, the forces on each beam decreased to 33
nN, 35 nN, 66 nN, and 40 nN for beam 4, 2, 3, and 1, respectively.
205
Figure 10.11 Time lapse optical micrographs of cell forces before and after exposure to cytochalasin-D. The chemical was administered at 57 min. The time point for the images are (a) 0 min., (b) 30 min., (c) 50 min., (d) 70 min., (e) 90 min., (f) 110 min., and (g) 130
min.
206
0
5 0
1 0 0
1 5 0
2 0 0
2 5 0
3 0 0
3 5 0
4 0 0
4 5 0
5 0 0
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0
T im e (m in .)
Forc
e (n
N)
P ro b e 1 P ro b e 2 P ro b e 3 P ro b e 4
C y to -D In jec tio n
Figure 10.12 Cell force magnitude versus time for a cell exposed to Cytochalasin-D. The chemical was injected at 57 minutes after seeding as indicated on the plot.
10.3.5 Sources of error and measures to minimize error
Errors in the measurement of the cell forces can be introduced in two main areas.
First, errors can be caused by materials and processing related factors. These include the
estimation of the elastic modulus and the measurement of the beam thickness. Despite the
fact that polystyrene is a fairly well characterized material, variations in the actual elastic
modulus of the material versus the value used in the simulations could still lead to some
error in the force calculations. The measurement of the beam thickness is also an
important factor. Each beam was measured individually and the beams should be quite
uniform in thickness given the fabrication processes used, so this factor should only lead
to a minimal amount of error. Finally, creep of the polystyrene during the experiment
could introduce a small amount of error into the system. These factors are difficult to
207
avoid. A direct calibration of the force-deflection behavior of the beams would eliminate
the need to estimate the elastic modulus; however, it is extremely difficult to find a good
calibration method for beams at this size scale and force measurement range. Therefore,
errors are likely to be introduced by the calibration process itself.
The second area that error can be introduced is in the image analysis. Errors in the
image analysis are the primary source of noise in the force magnitude plots and result in
errors in calculating both the force direction and magnitude. In order for the deflection
vector for the beam to be determined, a point at the end of each beam and a fixed point on
the device needed to be tracked over time. Ideally, the exact same pixel point on both the
end of the probe and the fixed point would be tracked consistently throughout the
analysis. However, given that the images were analyzed manually, an element of human
error is inevitably introduced into the system. One way to minimize this issue is to use
higher magnification images for analysis. For the first and second generation devices, the
images could be captured at a maximum magnification of 100x. This was a consequence
of not having any stationary points in the center of the device. This required that the
images have both the cell active area and the anchor region in the same image to provide
a reference point to correct for image shift. The fixed pad in the third generation device
not only serves as a point for cell attachment, but also provides a fixed reference point in
the image analysis. This allows the images to be collected at higher magnification. As a
result, each pixel represents a smaller surface area. Thus, errors in selecting the
appropriate pixel lead to less overall error in the measurement. A more sophisticated way
to deal with this issue would be to use an automated image analysis system. Not only
would this reduce the error in the pixel mapping process, it would also significantly cut
208
down on the time required for analysis. One downside of automated image analysis is the
high cost of the software.
10.4 Conclusion
The cell force sensor was tested using WS-1 skin fibroblasts as a model cells.
Various improvements were made from one generation of the device to the next in order
to improve the response of the beam and the behavior of the cell on the device. In
addition, various improvements were made to the experimental setup and the image
acquisition and analysis system. By the third generation of the device, time resolved
measurements of the cell force magnitude and direction could be made. In addition, the
effects of cytochalasin-D on cell force were determined. This is a significant step in the
implementation of the devices as a tool for evaluating the effects of drugs, toxins, and
other chemicals on cell mechanics.
References
1 C.G. Galbraith, M.P. Sheetz, “Forces on adhesive contacts affect cell function.” Current Opinions in Cell Biology 10 (1998) 566-571. 2 M.C. Beckerle, Cell Adhesion, Oxford University Press, Oxford (2001). 3 C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Geometic control of cell life and death.” Science 276 (1997) 1425-1428. 4 A.R. Horwitz, J.T. Parsons, “Cell migration—movin’ on.” Science 286 (1999) 1102-1103. 5 E. Crowley, A. Horwitz, “Tyrosine phosphorylation and cytoskeletal tension regulate the release of fibroblast adhesions.” Journal of Cell Biology 131 (1997) 525-537. 6 M. Chicurel, R. Singer, C. Meyer, D. Ingber, “Integrin binding and mechanical tension induce movement of mRNA and ribosomes to focal adhesions.” Nature 392 (1998) 730-733. 209
7 M.P. Sheetz, D.P. Felsenfeld, C.G. Galbraith, “Cell migration: regulation of force on extracellular-matrix-integrin complexes.” Trends in Cell Biology 8 (1998) 51-54. 8 F.C.S. Ramaekers, F.T. Bosman, “The cytoskeleton and disease.” Journal of Pathology 204 (2004) 351-354. 9 D. Paulin, A. Huet, L. Khanamyrian, Z. Xue, “Desminopothies in muscle disease.” Journal of Pathology 204 (2004) 418-427. 10 M. Pekny, M. Pekna, “Astrocyte intermediate filaments in CNS pathologies and regeneration.” Journal of Pathology 204 (2004) 428-437. 11 H. Yamaguchi, J. Condeelis, “Regulation of the actin cytoskeleton in cancer cell migration and invasion.” Biochimica et Biophysica Acta 1773 (2007) 642-652. 12 Z.Z. Wu, G. Zhang, M. Long, H.B. Wang, G.B. Song, S.X. Cai, “Comparison of the viscoelecastic properties of normal hepatocytes and hepatocellular carcinaoma cells under cytoskeletal perturbation.” Bioreology 37 (2000) 279-290. 13 M.J. Rosenbluth, W.A. Lam, D.A. Fletcher, “Force microscopy of nonadherent cells: a comparison of leukemia cell deformability.” Biophysical Journal 90 (2006) 2994-3003. 14 L.H. Miller, D.I. Baruch, K. Marsh, O.K. Doumbo, “The pathogenic basis of malaria.” Nature 415 (2002) 673-679. 15 Y. Calle, H.C. Chou, A.J. Thrasher, “Wiskott-Aldrich syndrome protein and the cytoskeletal dynamics of dendritic cells.” Journal of Pathology 204 (2004) 460-469. 16 G.D. Sempowski, M.A. Borrello, T.M. Blieden, K. Barth, R.P. Phipps, “Fibroblast heterogeneity in the healing wound.” Wound Repair and Regeneration 3 (1995) 120-131. 17 V.C. Sandulache, A. Parekh, H.-S. Li-Korotky, J.E. Dohar, P.A. Hebda, “Prostaglandin E2 differentially modulates human fetal and adult dermal fibroblast migration and contraction: implication for would healing.” Wound Repair and Regeneration 14 (2006) 633-643. 18 P.B. Schiff, S.B. Horwitz, “Taxol stabilizes microtubules in mouse fibroblast cells.” Proceedings of the National Academy of Sciences USA 77 (1980) 1561-1565. 19 V.K. Ngan, K. Bellman, B.T. Hill, L. Wilson, M.A. Jordan, “Mechanism of mitotic block and inhibition of cell proliferation by the semisyntheic vinca alkaloids Vinorelbine and its newer derivative Vinflunine.” Molecular Pharmacology 60 (2001) 225-232. 20 J.A. Cooper, “Effects of cytochalasin and phalliodin on actin.” The Journal of Cell Biology 105 (1987) 1473-1478.
210
21 M.R. Bubb, I. Spectro, B.B. Beyer, K.M. Fosen, “Effects of jasplakinolide on the kinetics of actin polymerization.” The Journal of Biological Chemistry 275 (2000) 5163-5170.
211
CHAPTER 11
MEASUREMENT OF MECHANICAL FORCES GENERATED BY PLANT P-PROTEIN AGGREGATES (FORISOMES)
11.1 Introduction
Forisomes are small chemomechanically active P-protein aggregates uniquely
found in the sieve cells of legumes (Fabaceae). Unlike most other motor proteins,
forisomes do not rely on adenosine triphosphate (ATP) to supply the energy for
actuation.1 Instead they contract or expand in respond to changes in specific ion
concentrations or pH. Given their unique actuation mechanism and ability to produce
large conformational changes, forisomes have been suggested for use as micro-actuators,
drug delivery devices, or micro-valves.1,2
During forisome contraction, free chemical reaction energy is converted to
longitudinal and radial mechanical work and volume change. Upon exposure to Ca(II),
Sr(II), and Ba(II) ions or during pH shifts from 7.3 to 4.5 or 7.3 to 11.0, the protein
aggregates contract by 10 to 40% of their original length and increase their diameter by
as much as 150%.3 The stability of the energy conversion depends on the dissolved
oxygen concentration and can be considerably improved by working under anaerobic
conditions.
212
The focus of this chapter is the development of a sensor with the ability to
measure the forces generated by forisomes in both the lateral and radial directions.
Previous work from Knoblauch et al.1 and Schwan et al.4 show that the longitudinal
contraction forces are between 50 nN and 120 nN. However, there is a need for a more
refined force measurement device for measuring these forces in both the longitudinal and
radial directions. Our approach to measuring forisome forces was the development of a
microfabricated polymer cantilever sensor.
Polymers are being employed more frequently in micro/nanoscale devices. This is
often due to their unique physical, chemical, and mechanical properties. In this case,
using a polymer with a relatively low elastic modulus provides a cantilever structure that
gives relatively large and measurable deflections, even in response to relatively low
magnitude forces.
11.2 Materials and methods
11.2.1 Design
213
The layout of the device is shown in Figure 11.1. The structure consists of four
cantilever probes, each 250 µm in length and 5 µm wide. Two of the probes are designed
to measure the force in the longitudinal direction of the forisome, and the spacing of the
probes is 30 µm to match the typical geometry of the forisome in the extended
confirmation. Similarly, the other two probes measure force in the radial direction and
have a spacing of approximately 3 µm. The device is fixed to the substrate at the edges of
the structure and the cantilevers are suspended approximately 1 µm above the substrate
using a water soluble poly(vinyl alcohol) (PVA) sacrificial layer that is dissolved prior to
testing.
Figure 11.1 (a) Cantilever sensor for measuring forisome forces, (b) close-up of the measurement region, and (c) cross-section of the center region of the sensor.
11.2.2 Finite element simulations
The behavior of the device was simulated using finite element analysis (ANSYS).
A fixed support was placed at the base of the beam, and forces ranging of 0-600 nN were
applied to a 3x3 µm surface at the free end of the beam. An elastic modulus of 3.2 GPa
and Poisson’s ratio of 0.325 were used for polystyrene.5 It was assumed that the
mechanical properties of the material were similar to the bulk properties as reported in
the literature. Previous research on the mechanical properties of polystyrene using
nanoindentation indicates that this was a reasonable assumption.6
11.2.3 Fabrication and characterization
Devices were fabricated using sacrificial layer micromolding. This process is a
soft lithography7 based micromolding technique that is capable of producing suspended
214
polymer structures such as cantilevers. The details of the process are described
elsewhere.8 Briefly, the device geometry was defined in SU8 2005 (MicroChem.)
photoresist using standard photolithography. The structure was then transferred into
poly(dimethylsiloxane) (PDMS). After curing, the PDMS mold was spin coated at 3000
rpm with a 7.5% (wt/wt) solution of polystyrene dissolved in anisole. The resulting film
coated both the recesses and raised features of the mold. To remove the material from the
raised features of the mold, the surface was brought into contact with a glass slide heated
to 200 ºC. This removed the polymer from the top surface of the mold, resulting in a
mold that was selectively filled with polystyrene in the recessed features. The selectively
coated mold was then aligned and brought into contact with the patterned sacrificial
layer. Heat (125 ºC) and pressure (~0.35 MPa) were used to transfer the structure onto the
sacrificial layer. During testing, a water-based conditioning solution was introduced into
the measurement chamber. This solution dissolved the sacrificial layer and released the
cantilever beams.
11.2.4 Forisome preparation
The forisomes tested in this work were isolated from Vicia faba. They had
original lengths between 30 and 50 µm and diameters between 2 and 3 µm. Forisomes
were prepared by mechanically isolating phloem tissue from the stems of plants that were
4 to 6 weeks old. The rind between the first and seventh internodes of the plants was
excised and put into a Ca2+-free solution containing 10 mM EDTA, 0.1 M KCl, and 10
mM Tris adjusted to pH 7.3. The separated and pre-dried phloem tissue was powdered
215
under liquid nitrogen and subsequently suspended in a Ca2+-free solution. The suspension
was filtered through a nylon sieve with a mesh size of 55 µm.
11.2.5 Experimental setup
After isolating the forisomes, they were transfer to a storage chamber. The storage
chamber was connected by a transfer channel to the measurement chamber. The flow and
exchange of conditioning and rinsing solutions were adjusted by computer-controlled
piston pumps in the measurement chamber. Two plunger pumps were combined to stock
solutions, adjusting the concentrations of Ca2+ ions and EDTA to switch forisomes
between the calcium loaded and calcium free states.
The force measurement chip was fixed in the measurement chamber. The
forisomes were transferred from the storage chamber and attached to the devices using a
micropipette attached to a three-axis micromanipulator. With the exception of the pumps,
the entire measurement set-up was mounted on an actively damped optical bench.
11.2.6 Data acquisition and analysis
During Ca2+ induced actuation of the forisomes, images of the devices were
captured on a CCD camera attached to the microscope. The displacement of each probe
was analysed using a digital greyscale image correlation program (ARAMIS).
10.3 Results and discussion
A SEM micrograph of the device prior to removal of the sacrificial layer is shown
in Figure 11.2. The overall device is shown in Figure 11.2(a). The anchor region can be
seen at the outer edge of the device and the beams are suspended over the PVA sacrificial
216
layer. Figure 11.2(b) shows the central region of the device where the forisome is
attached. The spacings between the beams are labeled.
Figure 11.2 Scanning electron micrograph of the forisome sensors before release of the sacrificial layer, (a) full view and (b) close-up of the center measurement region.
217
Results from FEA simulations are compared with the analytical solution for beam
bending in Figure 11.3. The analytical solution assumes a straight beam 250 µm long
with a point load applied to the end of the beam. The solution for the area moment of
inertia (I) is given in equation 1. Note that since the beam is deflecting laterally, the width
of the beam is the squared term in the equation. The force-deflection plot was then
determined using the relationship in equation 2:9
12
3hbI = (1)
EIFld3
3
= (2)
where b is the width of the beam and h is the height of the beam in equation 1 and F is
the applied load, l is the beam length, E is the elastic modulus, d is the deflection and I is
the moment area of inertia for a rectangular beam cross section in equation 2. The
boundary condition for the FEA simulations assumes that the load is applied to a 3x3 µm
surface at the end of the beam. The differences between the analytical and numerical
solution are attributed to differences in the loading conditions and the assumption of a
straight beam in the analytical solution.
218
Figure 11.3 Comparison of the finite element and analytical solutions for the force-deflection behavior of the beam.
Figure 11.4 shows optical micrographs of the forisome attached to the sensor (a)
before and (b) after actuation with 10 mM Ca2+. The images show that the forisome
contraction resulted in a significant bending of the cantilevers in both the longitudinal
and radial directions. The net displacement of both of the beams in a given axis was then
used to calculate the corresponding force according to the finite element simulation.
219
Figure 11.4 Optical micrographs of the sensor (a) before and (b) after actuation with Ca2+.
Figure 11.5 shows an SEM micrograph of the sensor with a forisome attached.
The forisome was placed with the long axis between the cantilevers with the larger gap
spacing. No additional adhesives were used to attach the forisome to the sensor. The
natural adhesion of the forisome to polystyrene was sufficient to maintain contact
between the forisome and the sensor during the measurements. The SEM image was
taken after the device was used and dried. The label A in the image shows damage to the
220
device caused by the micropipette tip. The label B points out that the cantilevers in the
images were attached to the substrate. This was a result of stiction of the device during
the drying process. The cantilevers remained suspended while in an aqueous
environment.
Figure 11.5 SEM micrograph of a forisome on the sensor after testing. The notation A on the image indicates damage to the beam caused by forisome placement and B shows that
the beam collapsed during the drying process due to stiction.
The resulting forces for each of seven different forisomes are shown in Figure
11.6. The longitudinal force is plotted versus the radial force. The forces ranged from 84-
136 nN with an average force of 117±19 nN (± standard deviation) in the longitudinal
direction of the forisomes. Forces in the radial direction were 22-61 nN with an average
force of 43±14 nN. The variation in the force is primarily attributed to differences in the
geometry and behavior of the individual forisomes. The results for the longitudinal force
magnitudes are in qualitative agreement with previous estimations. This is the first time
that the lateral forces generated by forisomes have been measured. While the radial forces
221
are smaller relative to the longitudinal force, these forces are sufficiently large to be
functionally incorporated into a microscale actuator or valve system.
Figure 11.6 Plot of radial versus longitudinal force for seven experiments.
11.4 Conclusion
In summary, we were able perform parallel measurements of both the radial and
longitudinal forces generated by forisomes in response to stimulation by Ca2+. The forces
were measured using a unique polymer cantilever sensors. Forces ranged from 84-136 nN
in the longitudinal direction and 22-61 nN in the radial direction. The ability to quantify
the bi-directional forces generated by forisomes is an important step in applying these
materials as functional components in microdevices.
222
References 1 M. Knoblauch, G. A. Noll, T. Müller, D. Prüfer, I. Scheider-Hüther, A. J. E. van Bel, “ATP-independent contractile proteins from plants.” Nature Materials 2 (2003) 600-603. 2 M. Knoblauch, W. S. Peters, “Biomimetic actuators: where technology and cell biology meet.” Cell and Molecular Life Sciences 61 (2004) 2497-2509. 3 S. Schwan, M. Fritzsche, A. Cismak, A. Heilmann, U. Spohn, “In vitro investigation of the geometric contraction behavior of chemo-mechanical P-protein aggregates (forisomes). Biophysical Chemisty 125 (2007) 444-452. 4 S. Schwan (unpublished data).
5 J. Brandrup, E. H. Immergut, E. A. Grulke, Polymer Handbook, Wiley, New York, NY (1999). 6 M. Palacio, B. Bhushan, N. Ferrell. D. Hansford, “Nanomechanical characterization of polymer beam structures for BioMEMS applications.” Sensors and Actuators A 135 (2007) 637-650. 7 Y. Xia, G.M. Whitesides, “Soft lithography.” Angewandte Chemie International Edition 37 (1998) 550-575. 8 N. Ferrell, J. Woodard, D. Hansford, “Polymer microfabrication for MEMS: sacrificial layer micromolding and patterned substrate micromolding.” Biomedical Microdevices 9 (2007) 815-821. 9 R. C. Hibbeler, Mechanics of Materials, Prentice Hall, New York, NY, 4th Edition (1997).
223
CHAPTER 12
CONCLUSIONS AND FUTURE OUTLOOK
Polymer MEMS have potential for a wide range of applications. One of the more
promising areas is the biomedical field. The biocompatibility, diverse material properties,
and low processing costs can all be taken advantage of for development of polymer
biomedical devices. Several areas of polymer microdevices have already found
commercial success. Specifically, polymer microfluidic devices are widely used for
diagnostic and analytical applications. Other areas are less mature, and require new and
improved methods of efficient and cost effective fabrication of functional polymer
structures and components. In addition to development of new microfabrication
techniques, characterization of polymer mechanical properties at the micro and nanoscale
is critical for proper implantation of polymers as functional components in MEMS.
Finally, there are a wide range of specific biomedical areas where polymer MEMS can be
used for practical applications. By directly interfacing with cells and biomolecules on
their size scale, new areas of cell biology can be explored. Polymer microfabrication
techniques have and will continue to provide a means to develop biomedical tools for cell
biology. In addition, polymer microdevices provide an avenue to develop new sensor
technologies with unprecedented resolution and sensitivity. MEMS devices have also led
224
to new diagnostic and therapeutic techniques. Continued research in these areas will lead
to new and even more exciting developments in biomedical research and clinical practice.
The work presented in this thesis focused on three primary areas of polymer
MEMS research: (1) development of new polymer microfabrication techniques, (2)
characterization of these techniques and of the properties of potential polymer MEMS
materials, and (3) development of devices for specific biomedical applications. This led
to the characterization of the double stamp micromolding technique and the development
and characterization of four other novel polymer microfabrication techniques. Spin
dewetting was used to fabricate microstructure from PPMA, PMMA, and polystyrene.
The process is capable of fabricating structures with either the same or very different
geometry than the original PDMS mold features. Proper design of the mold and selection
of polymer solution and spin coating conditions allows tight control over the geometry of
the resulting microstructures. Lift-off processing using micromolded PPMA sacrificial
layers was used to pattern conducting sulfonated polyaniline. Sacrificial layer
micromolding and pattern substrate micromolding were developed to allow fabrication of
suspended, mechanically independent polymer microstructures such as cantilevers.
A comprehensive characterization of the mechanical properties of potential
polymer MEMS materials was conducted using nanoindentation techniques. A new
method for conducting microscale beam bending tests was developed and used to
measure the elastic modulus, yield strength, and breaking strength of PPMA, PPMA, PS,
and PS/clay nanocomposites beams. Conventional nanoindentation techniques were used
to measure hardness, elastic modulus, and scratch resistance of the materials. Polymers
were also tested at body temperature and after exposure to aqueous environment to
225
determine if these factors have a significant effect on material properties. Results showed
that only PPMA had a significant change in properties under these conditions.
Three new polymer MEMS devices were developed and tested. The first device is
a microfabricated cell isolation devices. The device consists of a microfabricated
membrane bonded to a commercially available nanoporous filter. The device allows
active manipulation and isolation of adherent cells, non-adherent cells, and cell clusters.
In the case of adherent cells, morphology of individual cells can be controlled based on
the shape and size of the individual microwells. For cell clusters, the cluster size can be
controlled by changing the size of the wells. Two polymer devices were developed for
measuring low magnitude biological forces. The cell force sensor was designed and
simulated using finite element analysis. The device was fabricated using sacrificial layer
micromolding. The cell force sensor was tested using fibroblasts cells and dynamic forces
were measured. The effects of chemical disruption of the cytoskeleton on cell forces were
also determined. The third device was a modified force sensor for measuring forces
generated by plant protein aggregates called forisomes. The forces generated by
forisomes in response to changes in calcium concentration were measured for the first
time in both the longitudinal and radial directions.
The field of polymer MEMS is still in its relative infancy. The fabrication
processes developed through this work could find applications in the development of a
wide range of polymer devices for biomedical and other applications. The cell isolation
device could find applications in liver cell biology, stem cell biology, and tissue
engineering just to name a few. The cell force sensor could be used as a tool for cell
biologist to study cytoskeletal mechanics in normal and diseased cells. The device could
226
also be used as a drug evaluation and testing method for drugs that target cytoskeletal
components. This basic method of force measurement could find broader applications in
sensors for other applications.
227
APPENDIX A
DESIGN AND SIMULATION DATA FOR THE CELL FORCE SENSOR
228
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
x-deflection (µm)
y-de
flect
ion
(µm
)
0°
90°
180°
270°
Figure A.1 Deflection plot at 10º increments for third generation cell force sensor.
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30
Beam Deflection Angle
Forc
e A
ngle
Figure A.2 Curve fits for calculating force angle from beam deflection angle. (Continued)
229
Figure A.2 Continued
0
10
20
30
40
50
60
70
80
90
0 10 20 30 40 50 60 70
Beam Deflection Angle
Forc
e A
ngle
0
20
40
60
80
100
120
0 20 40 60 80 100 120 140
Beam Deflection Angle
Forc
e A
ngle
(Continued)
230
Figure A.2 Continued
0
20
40
60
80
100
120
140
0 50 100 150 200
Beam Deflection Angle
Forc
e A
ngle
0
20
40
60
80
100
120
140
160
180
155 160 165 170 175 180 185
Beam Deflection Angle
Forc
e A
ngle
(Continued)
231
Figure A.2 Continued
0
50
100
150
200
250
300
175 180 185 190 195 200 205 210
Beam Deflection Angle
Forc
e A
ngle
235
240
245
250
255
260
265
200 210 220 230 240 250
Beam Deflection Angle
Forc
e A
ngle
(Continued)
232
Figure A.2 Continued
255
260
265
270
275
280
285
0 50 100 150 200 250 300 350
Beam Deflection Angle
Forc
e A
ngle
275
280
285
290
295
300
305
310
315
300 310 320 330 340 350
Beam Deflection Angle
Forc
e D
irect
ion
300
310
320
330
340
350
360
370
340 345 350 355 360 365
Beam Deflection Angle
Forc
e A
ngle
233
Beam Angle (°) Equation R2 1.62-27.62 f= -0.055914b2 + 3.937841b - 6.182234 0.99993127.62-62.04 f = -0.009885b2 + 1.467356b + 27.012935 1.00000062.04-127.45 f = 0.00038811b2 + 0.23221066b + 64.09965349 1.000000127.45-157.03 f = 0.013060b2 - 3.039140b + 275.204737 1.000000157.03-179.02 f = 0.0564224b2 - 16.6828923b + 1348.4376082 0.999936170.02-207.80 f = -0.0550594b2 + 23.7377278b - 2315.1491175 0.999961207.80-242.04 f = -0.00999159b2 + 5.07882517b - 383.93813498 1.000000242.04-307.45 f = 0.0003649b2 + 0.1052494b + 213.1488368 1.000000307.45-343.92 f = 0.017022b2 - 10.275335b + 1830.226010 0.998718343.92-1.62 f = 0.062337b2 - 41.131991b + 7,082.698040 0.999792 b=beam angle (°) f=force angle (°)
Table A.1 Summary of equations for calculating force angle from the beam deflection angle.
Beam Deflection/5nN for Force Angle 0-60
y = -0.00001578x2 - 0.00026049x + 0.15466464R2 = 0.99959381
0.000000000.020000000.040000000.060000000.080000000.100000000.120000000.140000000.160000000.18000000
0 10 20 30 40 50 60 70
Force Angle
Bea
m D
efle
ctio
n/5n
N
Figure A.3 Curve fits for calculating force and from beam deflection angle. (Continued)
234
Figure A.3 Continued
Beam Deflection/5nN for Angle 60-90
y = 0.00003404x2 - 0.00647351x + 0.34882960R2 = 0.99773130
0.000000000.010000000.020000000.030000000.040000000.050000000.060000000.070000000.080000000.09000000
0 20 40 60 80 100
Beam Deflection/5nN for Force Angle 90-120
y = 0.00002622x2 - 0.00397301x + 0.18712398R2 = 0.99853986
00.010.020.030.040.050.060.070.080.090.1
0 20 40 60 80 100 120 140
Beam Deflection/5nN for Angle 120-180
y = -0.00001731x2 + 0.00630147x - 0.41971182R2 = 0.99954678
0.000000000.020000000.040000000.060000000.080000000.100000000.120000000.140000000.160000000.18000000
0 50 100 150 200
(Continued)
235
Figure A.3 Continued
Beam Deflection/5nN for Force Angle 180-240
y = -0.00001638x2 + 0.00568310x - 0.33847476R2 = 0.99922685
0.000000000.020000000.040000000.060000000.080000000.100000000.120000000.140000000.160000000.18000000
0 50 100 150 200 250 300
Beam Deflection/5nN for Force Angle 240-270
y = 0.00003409x2 - 0.01876177x + 2.62208024R2 = 0.99762382
0.000000000.010000000.020000000.030000000.040000000.050000000.060000000.070000000.080000000.09000000
235 240 245 250 255 260 265 270 275
Beam Deflection/5nN for Force Angle 270-300
y = 0.00002621x2 - 0.01340333x + 1.75048365R2 = 0.99854329
0.000000000.010000000.020000000.030000000.040000000.050000000.060000000.070000000.080000000.090000000.10000000
265 270 275 280 285 290 295 300 305
(Continued)
236
Figure A.3 Continued
Beam Deflection/5nN for Force Angle 300-360y = -0.00001723x2 + 0.01249192x - 2.11000316
R2 = 0.99960531
0.000000000.020000000.040000000.060000000.080000000.100000000.120000000.140000000.160000000.18000000
290 300 310 320 330 340 350 360 370
Force Angle Equation R2 0-60 DF = -0.00001578f2 - 0.00026049f + 0.15466464 0.99959460-90 DF = 0.00003404f2 - 0.00647351f + 0.34882960 0.99773190-120 DF= 0.00002622f2 - 0.00397301f + 0.18712398 0.998540120-180 DF = -0.00001731f2 + 0.00630147f - 0.41971182 0.999547180-240 DF = -0.00001638f2 + 0.00568310f - 0.33847476 0.999227240-270 DF = 0.00003409f2 - 0.01876177f + 2.62208024 0.997624270-300 DF = 0.00002621f2 - 0.01340333f + 1.75048365 0.998543300-360 DF = -0.00001723f2 + 0.01249192f - 2.11000316 0.999605 f=force angle (°) DF=deflection/5nN force
Table A.2 Equations for calculating deflection per unit force from beam deflection angle.
237
Input Beam Direction (°) 90 Formula 0-27.62 0 -104.5702 27.62-62.04 0 79.0104 62.04-127.45 88.14131 88.14131 127.45-157.03 0 107.4717 157.03-179.02 0 303.981 179.02-207.8 0 -624.7421 207.8-242.02 0 -7.7813 242.04-307.45 0 225.57249 307.45-343.02 0 1043.311 343.02-360 0 3885.772 Cell Force Direction 88.14131 Input Probe Deflection (µm) 2.5 0-60 0.00 0.009146096 1366.703315 60.01-90 293.78 0.042548504 293.7823636 90.01-120 0.00 0.040614885 307.7689372 120.01-180 0.00 0.001198899 10426.23046 180.01-240 0.00 0.035152638 355.5920908 240.01-270 0.00 1.238698981 10.09123297 270.01-300 0.00 0.773029067 16.17015522 300.01-360 0.00 -1.142973022 -10.93639111 Cell Force Magnitude 293.78 Adjustments Modulus Correction Young's Modulus (MPa) 3300 293.7823636 Thickness Correction Beam Thickness (micron) 2.5 1.21 243.5810949 Width Correction Beam Width (micron) 6 1.00 243.5323884
Final Results Cell Force Direction (°) 88.14 Cell Force Magnitude (nN) 243.53
Table A.3 Excel program for calculating the cell force direction and magnitude of experimental beam deflection data.
238
APPENDIX B
POLYMER MICROFABRICATION AS A TOOL FOR PROCESSING INORGANIC MATERIALS
239
B.1 Introduction
The techniques described in this thesis have been geared toward fabrication of
polymer structures for biological applications. However, several of the basic processing
techniques have also been utilized effectively in processes ceramics, ceramic composites,
and other inorganic materials. This appendix will describe several processes that were
derived from the techniques given in the previous chapters.
B.2 Spin dewetting of PMMA etch masks for silicon wet etching
240
Chapter 4 describes the process of spin dewetting of polymer materials on PDMS
molds. In addition to being used directly in a device application, as was the case with the
cell isolation device in Chapter 6, the molded polymer structures can be used etch masks
for further silicon processing. PMMA has previously been suggested as a masking
material for silicon etching with potassium hydroxide (KOH) and tetramethyl ammonium
hydroxide (TMAH).1-3 Spin dewetting provides a simple method for depositing
micropatterned PMMA films to act as etch masks. We demonstrated the use of
micromolded PMMA as a etch mask for anisotropic etching of <100> silicon in TMAH.
Molding of polymer etch masks could significantly reduce both the time and cost of
silicon etching as compared to typical masking techniques. The low etch rate of PMMA
in TMAH and KOH makes it a strong candidate as a etch mask for silicon processing.3
However, delamination of the PMMA film during etching can be a significant problem.
We used baking techniques as well as silicon surface modification with
hexamethyldisilazane (HMDS) to increase the adhesion between silicon and PMMA.
While PMMA masking time was significantly increased for lower temperature etching
(60 ºC), the masking time was considerable less at higher temperatures (80 ºC). This
study suggests that PMMA patterned using this technique is adequate as an etch mask for
lower temperature silicon etching.
The etch process is shown in Figure B.1. The PMMA films were fabricated via
spin dewetting and pattern transfer with 8% (wt/wt) PMMA solutions in anisole spin
coated on the PDMS molds with 30 µm square pillars with height of 9 µm. The
substrates were placed in TMAH with constant stirring. After etching, the mask was
removed by sonication in acetone.
Figure B.1 Process for wet etching of silicon in TMAH using a micromolded PMMA etch mask.
241
Figure B.2 SEM micrographs of silicon after etching in TMAH.
Figure B.2 shows silicon after etching in 60 ºC TMAH for 270 minutes. PMMA
showed adequate adhesion for masking of silicon in 60 ºC TMAH for etch times up to
480 minutes. HMDS deposition and heat treatment were essential to increase the PMMA
adhesion. Samples that were not coated with HMDS or annealed showed PMMA
delamination after approximately 5 minutes. Samples that were only coated with HMDS
or only annealed had masking times of 120 minutes and 40 minutes, respectively.
Masking times were taken as the time needed to delaminate the PMMA from the silicon
substrate. While PMMA was effective for low temperature etching, as the TMAH bath
temperature was increased to 80 ºC, delamination occurred much more quickly, usually
within 30 minutes. These results indicate that PMMA molding using spin dewetting
could find applications in specific silicon etch processes, but is not suitable for all
applications.
B.2 Spin dewetting of inorganic materials on PDMS Molds
The spin dewetting process is not limited to organically soluble thermoplastic
polymers.4 Spin dewetting was also demonstrated using aqueous solutions of NaCl,
242
sucrose, ammonium heptamolybdate tetrahydrate, and poly-l-lysine. The crystal size or
particles size for NaCl and sucrose were characterized versus the solution concentration.
SEM micrographs of dewetted crystals of NaCl are shown in Figure B.3. A solution of
NaCl in DI water was spin coated on the mold at 3000 rpm for one minute. The mold
features were 10 µm diameter pillars with 10 µm spacing between pillars and height of
7.5 µm. Sucrose particles were formed using the same mold and spin coating process
(Figure B.4). Note the difference in the morphology of the resulting structures. NaCl
formed regular single cubic crystals while sucrose formed hemispherical particles. The
crystal or particle size versus concentration of the NaCl and sucrose are given in Figure
B.5. NaCl crystals as small as 600 nm were formed using low concentration solutions.
Figures B.6 and B.7 show particles of ammonium heptamolybdate tetrahydrate and poly-
l-lysine made using the same process.
243
Figure B.3 SEM micrographs of NaCl crystals formed by spin dewetting at (a) 0.5% NaCl, (b) 10% NaCl, and (c) 25% NaCl.
244
Figure B.4 SEM micrographs of sucrose particles formed by spin dewetting at (a) 0.5% sucrose, (b) 10% sucrose, and (c) 25% sucrose.
245
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 5 10 15 20 25 30
Concentration (%)
Cry
stal
/Par
ticle
Siz
e (µ
m) .
NaCl Sucrose
Figure B.5 Crystal/particle size versus concentration (wt/wt) for NaCl and sucrose dewetting on 10 µm diameter pillars at 3000 rpm.
Figure B.6 Ammonium heptamolybdate hydrate particles formed from 1% (wt/wt) solution.
246
Figure B.7 Poly-l-lysine particles formed from 0.1% (wt/wt) solution.
B.3 Peptide mediated deposition of silica and gold
In Chapter 5, double stamp micromolding was used to fabricate sacrificial layers
that were removed after in situ polymerization of sulfonated polyaniline. The basic
process of using micromolded structures as sacrificial layers can be applied to other
materials. PPMA sacrificial layers were used for patterned peptide deposition. The
patterned peptides were then used to template selective deposition of silica and gold from
solution. This process allows deposition of materials at ambient temperature and near
neutral pH on a wide range of substrates. Silica deposition has been demonstrated on
silicon, polyimide, iron, and nickel.
The process schematic is shown in Figure B.8. Sacrificial layers were deposited
using double stamp micromolding. Poly-l-lysine, 3X Flag peptide, and bovine serum
albumin (BSA) were adsorbed on the surface. The sacrificial layers were removed in
organic solvent. Silica and gold were then selectively deposited out of solution on the
247
surface of the peptide. Details of the silica and gold deposition process can be found
elsewhere.5,6
Figure B.9 shows an SEM micrograph of patterned silica deposition on polyimide
with corresponding energy dispersive spectroscopy (EDS) of the pattern and the
substrate. The ESD spectra show that the material on surface is silica as indicated by the
Si peak and the increase in the magnitude of the O peak. The spectra also indicate that the
deposition is selective to the regions with the patterned peptide, given that no Si peak
appears in Scan Area 1. Figure B.10 shows patterned gold nanoparticle deposition using a
3X Flag peptide template.
248
Figure B.8 Schematic diagram of peptide mediated materials deposition process.
249
Figure B.9 SEM of silica deposition on a poly-l-lysine template. EDS spectra show Si and an increased amount of O in the peptide coated region, and no Si is present in the
uncoated region.
Figure B.10 Gold nanoparticles deposited on a 3X Flag peptide template.
250
B.4 Sol-gel processing of silica and silica/hydroxyapatite nanoparticle composites
A simple sol-gel micromolding process was developed for fabricating
microstructures from silica and silica/hydroxyapatite nanoparticle composites. This could
have applications for surface modification of dental implants for improving cell
interactions. The process is shown in Figure B.11. A sol (with or without HA
nanoparticles) was applied to a substrate and a PDMS mold was applied to the surface
and held under pressure. The sol was allowed to gel, and the mold was removed. The
substrate was then sintered at 350 ºC for 30 minutes. Figure B.12 shows 5x5 µm line and
5 µm diameter pillar patterns made of silica. EDS was used to characterize the chemical
makeup of the patterns with HA particles. Figures B.13 shows SEM micrographs and
corresponding EDS spectra for silica/HA nanoparticles composites. The EDS spectra
indicate that HA particles are present in the line features and in between the features as
indicated by the presence of Ca and P peaks. The HA in the pillar features is difficult to
detect using EDS. This is likely due to the thickness of the pillar features. If the HA is
buried within the feature, it is difficult to detect via EDS.
251
Figure B.11 Schematic diagram of the sol-gel microfabrication process.
Figure B.12 SEM micrographs of (a) 5x5 µm lines, (b) 5 µm diameter pillars, and (c) cross section view of 5 µm diameter pillars.
252
Figure B.13 SEM micrographs and EDS spectra for silica/HA patterns.
253
A process was also developed to give semi-selective deposition of high
concentrations of HA particles within the features. In this case, the HA particles were
suspended in ethanol. The solution was applied to the patterned PDMS mold and dried. A
glass slide was used to spread the particle coating and move the particles into the
recessed features of the mold. Tape was used to remove the particles from the surface of
the mold. The sol was applied to the substrate and the particle coated mold was applied to
the surface. The gelation and sintering processes were repeated as described above. SEM
micrographs of line and pillar patterns are shown in Figure B.12 along with EDS spectra
of the features and in between the features. The spectra show significant Ca and P peaks
on the features but little or no Ca and P in between the features. While this indicates that
the process is reasonably selective with regard to the location of the HA particles, it is
still possible that some HA is present between the features, but not at high enough
concentration to be detected by EDS.
Cell culture studies with primary human bone marrow osteoblasts were performed
to access the effects of the micropatterns on cell behavior and morphology. Figure B.13
shows optical micrographs of cells grown on a line pattern and flat SiO2. The images
show that the lines induce significant changes in cell morphology, with cells aligning and
elongating in the direction of the lines. Similar behavior is seen in the SEM images in
Figure B.14. All patterns (i.e. line, pillars, and pillars with HA) induced significant
changes in cell shape and orientation, with cells exhibiting a long, narrow morphology
guided by the direction of the features.
254
Figure B.14 SEM micrographs and EDS spectra for silca/HA composites fabricated using semi-selective deposition.
255
Figure B.15 Optical micrographs of cells grown on (a) micropatterned lines and (b) flat SiO2.
Figure B.16 SEM micrographs of cells grown on SiO2 (a) lines, (b) pillars, and (c) SiO2/HA pillars.
256
References 1 M. Ilie, B. M„rculescu, N. Moldovan, N. Nastase, M. Olteanu, “Adhesion between PMMA mask layer and silicon wafer in KOH aqueous solution.” in Materials and Device Characterization in Micromachining, edited by C. R. Friedrich and Y. Vladimirsky, (Proc. of SPIE 3512, Santa Clara, CA, 1998) pp. 422-430. 2 D. S. Bodas, S. K. Mahapatra, and S. A. Gangal, “Comparative study of spin coated and sputtered PMMA as an etch mask material for silicon micromachining.” Sensors and Actuators A 120 (2005) 582-588. 3 D. S. Bodas and S. A. Gangal, “Poly(methyl methacrylate) as masking material for microelectromechanical system (MEMS) fabrication.” Journal of Applied Polymer Science 102 (2006) 2094-2098. 4 J. Guan, N. Ferrell, B. Yu, D.J. Hansford, L.J. Lee, “Simultaneous fabrication of hybrid arrays of nanowires andmicro/nanoparticles by dewetting on micropillars.” Soft Matter 3 (2007) 1369-1371. 5 R.T. Butler, N.J. Ferrell, D.J. Hansford, “Spatial and geometrical control of silicification using a patterned poly-L-lysine template.” Applied Surface Science 252 (2006) 7337–7342. 6 H. Borteh, N. Ferrell, R. Butler, S. Olesik, D. Hansford, “Electroless Deposition of Gold Nanoparticles Over Silicon-based Substrates.” in Materials, Processes, Integration and Reliability in Advanced Interconnects for Micro- and Nanoelectronics, edited by Q. Lin, E. T. Ryan, W. Wu, D.Y. Yoon, (Mater. Res. Soc. Symp. Proc. 990, Warrendale, PA, 2007) 0990-B08-17.
257
LIST OF REFEFNCES
1. W. Kern (ed.), Handbook of Semiconductor Wafer Cleaning Technology: Science, Technology, and Applications, Noyes, Park Ridge, NJ (1993).
2. A.G. Ernslie, F.T. Bonner, L.G. Peck, “Flow of a viscous liquid on a rotating disk.” Journal of Applied Physics 29 (1958) 858-862.
3. D. Meyerhofer, “Characteristics of resist films produced by spinning.” Journal of Applied Physics 49 (1978) 3993-3997.
4. D.C. Duffy, J.C. McDonald, O.J.A. Schueller, G.M. Whitesides, “Rapid prototyping of microfluidic systems in poly(dimethylsiloxane).” Analytical Chemistry 70 (1998) 4974-4984.
5. M. Mehregany, C.A. Zorman, “SiC MEMS: opportunities and challenges for applications in harsh environments.” Thin Solid Films 355-356 (1999) 518-524.
6. Y. Fu, H. Du, W. Huang, S. Zhang, M. Hu, “TiNi-based thin films in MEMS applications: a review.” Sensors and Actuators A 112 (2004) 395-408.
7. D. Niarchos, “Magnetic MEMS: key issues and some applications.” Sensors and Actuators A 109 (2003) 166-173.
8. M. Madou, Fundumentals of Microfabrication, CRC Press, Boca Raton, FL (1997) pp.171.
9. H. Janseny, H. Gardeniers, M. de Boer, M. Elwenspoek, J. Fluitman, “A survey on the reactive ion etching of silicon in microtechnology.” Journal of Micromechanics and Microengineering 6 (1996) 14-28.
10. Microelectromechanical Systems (MEMS) Technology: Current And Future Markets, BCC Research (2006).
11. Global MEMS/Microsystems - Markets and Opportunities, Yole Développement (2007).
258
12. W.H. Ko, “Trends and frontiers of MEMS.” Sensors and Actuators A 136 (2007) 62-67.
13. R.S. Shawgo, A.C. Richards Gradyson, Y.W. Li, M.J. Cima, “BioMEMS for drug delivery.” Current Opinions in Solid State Materials Science 6 (2002) 329-334.
14. A.C. Richards Grayson, R.S. Shawgo, Y. Li, M.J. Cima, “Electronic MEMS for triggered delivery.” Advanced Drug Delivery Reviews 56 (2004) 173-184.
15. S.L. Tao, T.A. Desai, “Microfabricated drug delivery systems: from particles to pores.” Advanced Drug Delivery Reviews 55 (2003) 315-328.
16. R. Bashir, “BioMEMS: state-of-the-art in detection, opportunities and prespects.” Advanced Drug Delivery Reviews 56 (2004) 1565-1585.
17. J. Santini, M.J. Cima, R.J. Langer, “A controlled-release microchip.” Nature 397 (1999) 335-338.
18. M. Staples, K. Daniel, M.J. Cima, R. Langer, “Application of micro- and nano-electromechanical devices to drug delivery.” Pharmaceutical Research 23 (2006) 847-863.
19. T.A. Desai, D. Hansford, M. Ferrari, “Characterization of micromachined silicon membranes for immunoisolation and bioseparation applications.” Journal of Membrane Science 159 (1999) 221-231.
20. T.A. Desai, W.H. Chu, G. Rasi, P. Sinibaldi-Vallebona, E. Guarino, “Microfabricated biocapsules provide short-term immunoisolation of insulinoma xenografts.”Biomedical Microdevices 1 (1999) 131-138.
21. J. Fritz, M.K. Baller, H.P. Land. H. Rothuizen, P. Vettiger, E. Meyer, H.-H. Güntherodt, Ch. Gerber, J.K. Gimzewski, “Translating biomolecular recognition into nanomechanics.” Science 288 (2000) 316-318.
22. G. Wu, R.H. Datar, K.M. Hansen, T. Thundat, R.J. Cote, A. Majumdar, “Bioassy of prostate-specific antigen (PSA) using microcantilevers.” Nature Biotechnology 19 (2001) 856-860.
23. A. Gupta, D. Akin, R. Bashir, “Single virus particle mass detection using microresonators with nanoscale thickness.” Applied Physics Letters 84 (2004) 1976-1978.
24. T.H. Schulte, R.L. Bardell, B.H. Weigl, “Microfluidic technologies in clinical diagnostics.” Clinica Chimica Acta 321 (2002) 1-10.
259
25. G.H.W. Sanders, A. Manz, ”Chip-based Microsystems for genomic and proteomic analysis.” Trends in Analytical Chemistry 19 (2000) 364-378.
26. Yu, X.; Zhang, D.; Li, T.; Hao, L.; Li, X. “3-D microarrys biochip for DNA amplification in polydimethylsiloxane (PDMS) elastomer.” Sensors and Actuators A 108 (2003) 103-107.
27. M.T. Cronin, T. Boone, A.P. Sassi, H. Tan, Q. Xue, S.J. Williams, A.J. Ricco, H.H. Hooper. “Plastic microfluidic systems for high-throughput genomic analysis and drug screening.” Journal of the Association for Laboratory Automation 6 (2001) 71-75.
28. H. Becker, C. Gärtner, “Polymer microfabrication methods for microfluidic analytical applications.” Electrophoresis 21 (2000) 12-26.
29. H. Becker, C. Gärtner, “Polymer microfabrication technologies for microfluidic systems.” Analytical and Bioanalytical Chemistry 390 (2008) 89-111.
30. Y. Xia, G.M. Whitesides, “Soft lithography.” Angewandte Chemie International Edition 37 (1998) 550-575.
31. C.S. Effenhauser, G.J.M. Bruin, A. Paulus, M. Ehart, Analytical Chemistry 69 (1997) 3451-3457.
32. J.S. Buch, P.-C. Wang, D.L. DeVoe, C.S. Lee, “Field-effect flow control in a polydimethylsoloxane based microfluidic system.” Electrophoresis 22 (2001), 3902-3907.
33. J.C. McDonald, D.C. Duffy, J.R. Anderson, D.T. Chiu, H. Wu, O.J.A. Scheuller, G.M. Whitesides, “Fabrication of microfluidic systems in poly(dimethylsiloxane).” Electrophoresis 21 (2000), 27-40.
34. S.K. Sia, G.M. Whitesides, “Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies.” Electrophoresis 24 (2003), 3563-3576.
35. J.M.K. Ng, I. Gitlin, A.D. Stroock, G.M. Whitesides, “Components for integrated poly(dimethylsiloxane) microfluidic systems.” Electrophoresis 23 (2002), 3461-3473.
36. X.-M. Zhoa, Y. Xia, and G.M. Whitesides, “Fabrication of three-dimensional micro-structures: microtransfer molding.” Advanced Materials 8 (1996) 837-840.
37. E. Kim, Y. Xia, and G.M. Whitesides, “Polymer microstructures formed by moulding in capillaries.” Nature 376 (1995) 581-584.
260
38. Y. Xia, G.M. Whitesides, “Soft lithography.” Annual Review of Materials Science 28 (1998) 153-184.
39. A. Kumar, G.M. Whitesides, “Features of gold having micrometer to centimeter dimensions can be formed through a combination of stamping with an elastomeric stamp and an alkanethiol “ink” followed by chemical etching.” Applied Physics Letters 61 (1993) 2002-2004.
40. R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, G.M. Whitesides, “Patterning proteins and cells using soft lithography.” Biomaterials 20 (1999) 2363-2376.
41. J.A. Rogers, R.G. Nuzzo, “Recent progress in soft lithography.” Materials Today 8(2) (2005) 50-56.
42. M. Wang, H.-G. Braun, T. Kratzmüller, E. Meyer, “Patterning polymers by micro-fluid-contact printing.” Advanced Materials 13 (2001) 1312-1317.
43. W.S. Beh, I.T. Kim, D. Qin, Y. Xia, G.M. Whitesides, “Formation of patterned microstructures of conducting polymers by soft lithography, and applications in microelectronic device fabrication.” Advanced Materials 11 (1999) 1038-1041.
44. J.A. Rogers, Z. Boa, V.R. Raju, “Nonphotolithographic fabrication of organic transistors with micron feature sizes.” Applied Physics Letters 72 (1998) 2716-2718.
45. T. Granlund, T. Nyberg, L.S. Roman, M. Svensson, O. Inganäs, “Patterning of polymer light-emitting diodes by soft lithography.” Advanced Materials 12 (2000) 269-273.
46. B. Xu, F. Arias, S.T. Brittain, X.-M. Zhao, B. Grzybowski, S. Torquato, G.M. Whitesides, “Making negative Poisson’s ratio microstructures by soft lithography.” Advanced Materials 11 (1999) 1186-1189.
47. D. Gallego, N. Ferrell, D. Hanford, “Fabrication of Piezoelectric Polyvinylidene Fluoride (PVDF) Microstructures by Soft Lithography for Tissue Engineering and Cell Biology Applications.” in Printing Methods for Electronics, Photonics and Biomaterials, edited by G. Gigli (Mater. Res. Soc. Symp. Proc. Volume 1002E, Warrendale, PA, 2007) 1002-N04-05.
48. G. Vozzi, C. Flaim, A. Ahluwalia, S. Bhatia, “Fabrication of PLGA scaffolds using soft lithography and microsyringe deposition.” Biomaterials 24 (2003), 2533-2540.
261
49. D. Gallego, N. Ferrell, Y. Sun, D.J. Hansford, “Multilayer micromolding of degradable polymer tissue engineering scaffolds.” Materials Science and Engineering C (in press).
50. J. Guan, H. He, D.J. Hansford, L.J. Lee, “Self-folding of three-dimensional hydrogel microstructures.” The Journal of Physical Chemistry B 109 (2005) 23134-23137.
51. G. Genolet, J. Brugger, M. Despont, U. Drechsler, P. Vettiger, N.F. de Rooij, D. Anselmetti, “Soft, entirely photoplastic probes for scanning force microscopy.” Review of Scientific Instruments 70 (1999) 2398-2401.
52. X. Wang, J. Engel, C. Liu, “Liquid crystal polymer (LCP) for MEMS: processes and applications.” Journal of Micromechanics and Microengineering 13 (2003) 628-633.
53. G. Witzgall, R. Vrijen, E. Yablonovitch, V. Doan, B.J. Schwartz, “Single-shot two-photon exposure of commercial photoresist for the production of three-dimensional structures.” Optics Letters 23 (1998) 1745-1747.
54. T. Baldacchini, C.N. LaFratta, R.A. Farrer, “Acrylic-based resin with favorable properties for three-dimensional two-photon polymerization.” Journal of Applied Physics 95 (2004) 6072-6076.
55. A. Bertsch, H. Lorenz, P. Renaud, “3D microfabrication by combining microstereolithography and thick resist UV lithography.” Sensors and Actuators A 73 (1999) 14-23.
56. H. Sirringhaus, T. Kawase, R.H. Friend, T. Shimoda, M. Inbasekaran, W. Wu, E.P. Woo, “High-resolution inkjet printing of all-polymer transistor circuits.” Science 290 (2000) 2123-2126.
57. X.-M. Zhao, S.P. Smith, S.J. Waldman, G.M. Whitesides, M. Prentiss, “Demonstration of waveguide couplers fabricated using microtrasnfer molding.” Applied Physics Letters 71 (1997) 1017-1019.
58. C. Liu, “Recent developments in polymer MEMS.” Advanced Materials 19 (2007) 3783-3790.
59. J. Engel, J. Chen, Z. Fan, C. Liu, “Polymer micromachined multimodal tactile sensors.” Sensors and Actuators A 117 (2005) 50-61.
60. G. Guan, N. Ferrell, L.J. Lee, D.J. Hansford, “Fabrication of polymeric microparticles for drug delivery by soft lithography.” Biomaterials 27 (2006) 4034-4041.
262
61. S.L. Tao, T.A. Desai, “Microfabrication of multilayer, asymmetric, polymeric devices for drug delivery.” Advanced Materials 17 (2005) 1625-1630.
62. N. Chronis, L.P. Lee, “Electrothermally activated SU-8 microgripper for single cell manipulation in solution.” Journal of Microelectromechanical Systems 14 (2005) 857-863.
63. M. Ferrari, A. Lee, L.J. Lee (eds.), Biological and Biomedical Nanotechnology, Springer, Boston, MA (2006).
64. A.C. Richards Grayson, I.S. Choi, B.M. Tyler, P.P. Wang, H. Brem, M.J. Cima, R. Langer, “Multi-pulse drug delivery from a resorbable polymeric microchip device.” Nature Materials 2 (2003) 767-772.
65. A. Chakrapani, “Processing and characterization of polymer microparticles for controlled drug delivery systems.” Ph.D. Thesis, Ohio State University (2007).
66. S.N. Bhatia, C.S. Chen, “Tissue engineering at the microscale.” Biomedical Microdevices 2 (1999) 131-144.
67. Y. Yang, S. Basu, D.L. Tomasko, L. James Lee, S.-T. Yang, “Fabrication of well-defined PLGA scaffolds using microembossing and carbon dioxide bonding.” Biomaterials 26 (2005) 2585-2594.
68. A.W. Orr, B.P. Helmke, B.R. Blackman, M.A. Schwartz, “Mechanisms of mechanotransduction.” Developmental Cell 10 (2006) 11-20.
69. M.R.K. Mofrad, R.D. Kamm (eds.), Cytoskeletal Mechanics, Cambridge University Press, Cambridge (2006).
70. A. Nicolas, B. Geiger, S.A. Safran, “Cell mechanosensitivity control the anisotropy of focal adhesions.” Proceedings of the National Academy of Sciences USA 101 (2004) 12520-12525.
71. M.P. Sheetz, D.P. Felsenfeld, C.G. Galbraith, “Cell migration: regulation of force on extracellular matrix-integrin complexes.” Trends in Cell Biology 8 (1998) 51-54.
72. M.E. Chicurel, C.S. Chen, D.E. Ingber, “Cellular control lies in the balance of forces.” Current Opinions in Cell Biology 10 (1998) 232-239.
73. C.G. Galbraith, M.P. Sheetz, “Forces on adhesive contacts affect cell function.” Current Opinions in Cell Biology 10 (1998) 566-571.
74. N. Wang, J.T. Butler, D.E. Ingber, “Mechanotransduction across the cell surface and through the cytoskeleton.” Science 260 (1993) 1124-1127.
263
75. C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Geometric control of cell life and death.” Science 276 (1997) 1425-1428.
76. M. Chrzanowska-Wodnicke, K. Burridge, “Rho-stimulated contractility drive the formation of stress fibers and focal adhesions.” Journal of Cell Biology 133 (1996) 1403-1415.
77. D.E. Ingber, “Cellular tensegrity revisited I: Cell structure and hierarchical systems biology.” Journal of Cell Science 116 (2003) 1157-1173.
78. D.E. Ingber, “Tensiegrity II. How structural networks influence cellular information processing networks.” Journal of Cell Science 116 (2003) 1397-1408.
79. G. Bao, S. Suresh, “Cell and molecular mechanics of biological materials.” Nature Materials 2 (2003) 715-725.
80. A.E. Cress, R.B. Nagle (eds.), Cell Adhesion and Cytoskeletal Molecules in Metastasis, Springer, Dordrecht, The Netherlands (2006).
81. H. Yamaguchi, J. Condeelis, “Regulation of the actin cytoskeleton in cancer cell migration and invasion.” Biochimica et Beophysica Acta 1773 (2007) 642-652.
82. F.C.S. Ramaekers, F.T. Bosman, “The cytoskeleton and disease.” Journal of Pathology 204 (2004) 351-354.
83. E.B. Lane, W.H.I. McLean, “Keratins and skin disorders.” Journal of Pathology 204 (2004) 355-366.
84. E. Fuchs, “The cytoskeleton and disease: genetic disorder of intermediate filaments.” Annual Review of Genetics 30 (1996) 197-231.
85. K. Zatloukal, C. Stumptner, A. Fuchsbichler, P. Fickert, C. Lackner, M. Trauner, H. Denk, “The keratin cytoskeleton in liver disease.” Journal of Pathology 204 (2004) 367-376.
86. D.W. Owens. E.B. Lane, “Keratin mutations and intestinal pathology.” Journal of Pathology204 (2004) 377-385.
87. Y. Calle, H.-C. Chou, A.J. Thrasher, G.E. Jones. “Wiskott-Aldrich syndrome protein and the cytoskeletal dynamics of dendritic cells.” Journal of Pathology 204 (2004) 460-469.
88. P.B. Schiff, S.B. Horwitz, “Taxol stabilizes microtubules in mouse fibroblast cells.” Proceedings of the National Academy of Sciences USA 77 (1980) 1561-1565.
264
89. M.V. Blagosklonny, T. Fojo, “Molecular effect of Paclitaxel: myths and reality.” International Journal of Cancer 83 (1999) 151-156.
90. V.K. Ngan, K. Bellman, B.T. Hill, L. Wilson, M.A. Jordan, “Mechanism of mitotic block and inhibition of cell proliferation by the semisynthetic vinca alkaloids vinorelbine and its newer derivative vinflunine.” Molecular Pharmacology 60 (2001) 225-232.
91. G.G. Borisy, E.W. Taylor, “The mechanism of action of colchicines.” The Journal of Cell Biology 34 (1967) 525-533.
92. K.L.K. Duncan, M.D. Duncan, M.C. Ally, E.A. Sausville, “Cucurbitacin E-induced disruption of the actin and vimentin cytoskeleton in prostate carcinoma cells.” Biochemical Pharmacology 52 (1996) 1553-1560.
93. J.A. Cooper, “Effects of cytochalasin and phalloidin on actin.” The Journal of Cell Biology 105 (1987) 1473-1478.
94. M.R. Bubb, A.M.J. Senderowiczf, E.A. Sausville, K.L.K. Duncan, E.D. Kern, “Jasplakinolide, a cytotoxic natural product, induces actin polymerization and competitively inhibits the binding of phalloidin to F-actin.” The Journal of Biological Chemistry 269 (1994) 14869-14871.
95. M.R. Bubb, I. Spector, B.B. Beyer, K.M. Fosen, “Effects of Jasplakinolide on the kinetics of actin polymerization.” The Journal of Biological Chemistry 275 (2000) 5163-5170.
96. T. Ohashi, M. Sato, “Remodeling of vascular endothelial cells exposed to fluid shear stress: experimental and numerical approach.” Fluid Dynamics Research 37 (2005) 40-59.
97. O. Traub, B.C. Berk, “Laminar shear stress: mechanicsms by which endothelial cells transduce an atheroprotective force.” Arteriosclerosis, Thrombosis, and Vascular Biology 18 (1998) 677-685.
98. C.R. White, J.A. Frangos, “The shear stress of it all: the cell membrane and mechanochemical tranduction.” Philosophical Transactions of the Royal Society B 362 (2007) 1459-1467.
99. K.J. Van Vliet, G. Bao, S. Suresh, “The biomechanics toolbox: experimental approaches for living cells and biomolecules.” Acta Materialia 51 (2003) 5881-5905.
100. T.D. Brown, “Techniques for mechanical stimulation of cell in vitro: a review.” Journal of Biomechanics 33 (2000) 3-14.
265
101. T.G. Kuznetsova, M.N. Starodubtseva, N.I. Yegorenkov, S.A. Chizhik, R.I. Zhdanov. “Atomic force microscopy probing of cell elasticity.” Micron 38 (2007) 824-833.
102. C. Rotsch, M. Radmacher, “Drug-induced changes of cytoskeletal structure and mechanics in fibroblasts: an atomic force microscopy study.” Biophysical Journal 78 (2000) 520-535.
103. R.M. Hochmuth, “Micropipette aspiration of living cells.” Journal of Biomechanics 33 (2000) 15-22.
104. M. Dao, C.T. Lim, S. Suresh, “Mechanics of the human red blood cell deformed by optical tweezers.” Journal of the Mechanics and Physics of Solids 51 (2003) 2259-2280.
105. A.D. Mehta, M. Rief, J.A. Spudich, D.A. Smith, R.M. Simmons, “Single-molecule biomechanics with optical methods.” Science 283 (1999) 1689-1695.
106. A.K. Harris, P. Wild, D. Stopak, “Silicone rubber substrata: a new wrinkle in the study of cell locomotion.” Science 208 (1980) 177-179.
107. M. Dembo, Y.-L. Wang, “Stresses at the cell-to-substrate interface during locomotion of fibroblasts.” Biophysical Journal 76 (1999) 2307-2316.
108. S. Munevar, Y.-L Wang, M. Dembo, “Traction force microscopy of migrating normal and H-ras transformed 3T3 fibroblasts.” Biophysical Journal 80 (2001), 1744-1757.
109. R.J. Pelham, Y.-L. Wang, “Cell locomotion and focal adhesions are regulated by substrate flexibility.” Proceedings of the National Academy of Sciences USA 94 (1997) 13661-13665.
110. C.-M. Lo, H.-B. Wang, M. Dembo, Y.-L. Wang, “Cell movement is guided by the rigidity of the substrate.” Biophysical Journal 79 (2000) 144-152.
111. G. Lin, K.SJ. Pister, K.P. Roos, “Surface micromachined polysilicon heart cell force transducer.” Journal of Microelectromechanical Systems 9 (2000) 9-17.
112. C.B. Galbraith, M.P. Sheetz, “A micromachined device provides a new bend on fibroblast traction forces.” Proceeding of the National Academy of Sciences USA 94 (1997) 9114-9118.
113. J.L. Tan, J. Tien, D.M. Pirone, D.S. Gray, K. Bhadriraju, C.S. Chen, “Cells lying on a bed of microneedles: An approach to isolate mechanical force.” Proceeding of the National Academy of Sciences USA 100 (1993), 1484-1489.
266
114. R. Günter, “Dewetting of thin polymer films.” Physical Review Letters 68 (1992) 75-78.
115. R. Xie, A. Karim, J.F. Douglas, C.C. Han, R.A. Weiss, “Spinodal dewetting of thin polymer films.” Physical Review Letters 81 (1998) 1251-1254.
116. A. Sehgal, V. Ferreiro, J.F. Douglas, E.J. Amis, A. Karim, “Pattern-directed dewetting of ultrathin polymer films.” Langmuir 18 (2002) 7041-7048.
117. E. Meyer, H. Braun, “Controlled dewetting processes on microstructured surfaces- a new procedure for thin film microstructuring.” Macromolecular Materials and Engineering 276/277 (2000) 44-50.
118. Z. Zhang, Z. Wang, R. Xing, Y. Han, “How to form regular polymer microstructures by surface-patterned-directed dewetting.” Surface Science 539 (2003) 129-136.
119. Z. Zhang, Z. Wang, R. Xing, Y. Han, “Patterning thin polymer films by surface-directed dewetting and pattern transfer.” Polymer 44 (2003) 3737-3743.
120. P. Lenz, R. Lipowsky, “Morphological transitions of wetting layers on structured surfaces.” Physical Review Letters 80 (1998) 1920-1923.
121. R. Konnur, K. Kargupta, A. Sharma, “Instability and morphology of thin liquid films on chemically heterogeneous substrates.” Physical Review Letters 84 (2000) 931-934.
122. K. Kargupta, A. Sharma, “Dewetting of thin films on periodic physically and chemically patterned surfaces.” Langmuir 18 (2002) 1893-1903.
123. L.-R. Bao, L. Tan, X.D. Huang, Y.P. Kong, L.J. Guo, S.W. Pang, A.F. Yee, “Polymer inking as a micro- and nanopatterning technique.” Journal of Vacuum Science and Technology B 21 (2003) 2749-2754.
124. J. Guan, A. Chakrapani, D.J. Hansford, “Polymer microparticles fabricated by soft lithography.” Chemistry of Materials 17 (2005) 6227-6229.
125. M. Gerard, A. Chaubey, B.D. Malhotra, “Application of conducting polymers to biosensors.” Biosensors and Bioelectronics 17 (2002) 345-359.
126. S. Geetha, C.R.K. Rao, M. Vijayan, D.C. Trivedi, “Biosensing and drug delivery by polypyrrole.” Analytica Chimica Acta 568 (2006) 119-125.
127. L. Dai, B. Winkler, L. Dong, L. Tong, A.W.H. Mau, “Conjugated polymers for light-emitting applications.” Advanced Materials 13 (2001) 915-925.
267
128. D. Braun, “Semiconducting polymer LEDs.” Materials Today 5(6) (2002) 32-39.
129. C.J. Drury, C.M.J. Mutsaers, C.M. Hart, M. Matters, D.M. de Leeuw, “Low-cost all-polymer integrated circuits.” Applied Physics Letters 73 (1998) 108-110.
130. N. Stutzmann, R.H. Friend, H. Sirringhaus, “Self-aligned, vertical-channel, polymer field-effect transistors.” Science 299 (2003) 1881-1884.
131. E.W.H. Jager, E. Smela, O. Inganäs, “Microfabricated conjugated polymer actuators.” Science 290 (2000) 1540-1545.
132. E. Smela, “Microfabrication of PPy microactuators and other conjugated polymer devices.” Journal of Micromechanics and Microengineering 9 (1999) 1-18.
133. X.L. Wei, Y.Z. Wang, S.M. Long, C. Bobeczko, A.J. Epstein, “Synthesis and physical properties of highly sulfonated polyaniline.” Journal of the American Chemical Society 118 (1996) 2545-2555.
134. J. Yue, Z.H. Wang, K.R. Cromack, A.J. Epstein A.G. MacDiarmid, “Effect of sulfonic acid group on polyaniline backbone.” Journal of the American Chemical Society. 113 (1991) 2665-2671.
135. J. Stejskal, I. Sapurina, J. Prokeš, J. Zemek, “In-situ polymerized polyaniline films.” Synthetic Metals 105 (1999) 195-202.
136. I. Sapurina, A. Riede, J. Stejskal, “In-situ polymerized polyaniline films 3. Film formation.” Synthetic Metals 123 (2001) 503-507.
137. S. Holdcroft, “Patterning π-conjugated polymers.” Advanced Materials 13 (2001) 1753-1765.
138. D.G. Lidzey, M.A. Pate, M.S. Weaver, T.A. Fisher, D.D.C. Bradley, “Photoprocessed and micropatterned conjugated polymer LEDs.” Synthetic Metals 82 (1996) 141-148.
139. F.J. Touwslager, N.P. Willard, D.M. de Leeuw, “I-line lithography of poly-(3,4-ethylenedioxythiophene) electrodes and application in all-polymer integrated circuits.” Applied Physics Letters 81 (2002) 4556-4558.
140. I. Mav, M. Žigon, A. Šebenik, “Sulfonated polyaniline.” Synthetic Metals 101 (1999) 717-718.
141. I. Mav, M. Žigon, A. Šebenik, J. Vohlidal, “Sulfonated polyanilines prepared by copolymerization of 3-aminobenzenesulfonic acid and aniline: the effect of
268
reaction conditions on polymer properties.” Journal of Polymer Science A 38 (2000) 3390-3398.
142. Y. Huang, B. Rubinsky, “Microfabricated electroporation chip for singe cell membrane permeabilization.” Sensors and Actuators A 89 (2001) 242-249.
143. M. Khine, A. Lau, C. Ionescu-Zanetti, J. Seo, L.P. Lee, “A single cell electroporation chip.” Lab on a Chip 5 (2005) 38-43.
144. K.G. Klemic, J.F. Klemic, M.A. Reed, F.J. Sigworth, “Micromolded PDMS planar electrode allows patch clamp electrical recordings from cells.” Biosensors and Bioelectronics 17 (2002) 597-604.
145. X. Li, K.G. Klemic, M.A. Reed, F.J. Sigworth, “Microfluidic system for planar patch clamp electrode arrays.” Nano Letters 6 (2006) 815-819.
146. P. Wang, G. Xu, L. Qui, Y. Xu, Y. Li, R. Li, “Cell-based biosensors and its application in biomedicine.” Sensors and Actuators B 108 (2005) 567-584.
147. M. Nishizawa, K. Takoh, T. Matsue, “Micropatterning of HeLa cells on glass substates and evaluation of respiratory activity using microelectrodes.” Langmuir 18 (2002) 3645-3649.
148. C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Micropatterned surface for control of cell shape, position, and function.” Biotechnology Progress 14 (1998) 356-363.
149. Y. Ito, “Surface micropatterning to regulate cell function.” Biomaterials 20 (1999) 2333-2342.
150. A. Brock, D. Chang, C. Ho, P. LeDuc, X. Jiang, G.M. Whitesides, D.E. Ingber, “Geometric determinant of directional cell motility revealed using microcontact printing.” Langmuir 19 (2003) 1611-1617.
151. M. Khalil, A. Shariat-Panahi, R. Tootle, T. Ryder, P. McCloskey, E. Roberts, H. Hodgson, C. Selden, “Human hepatocyte cell lines proliferating as cohesive spheroid colonies in alginate markedly upregulate both synthetic and detoxificatory liver function.” Journal of Hepatology 297 (2001) 68-77.
152. R. Clicklis, J.C. Merchuk, S. Cohen, Modeling mass trasfer in hepatocyte spheroids via cell viability, spheroid size and hepatocellular funtion.” Biotechnology and Bioengineering 86 (2004) 672-680.
153. S.M. Dang, M. Kyba, R. Perlingeiro, G.Q. Daley, P.W. Zandstra, “Efficiency of embryoid body formation and hematopoietic development from embryonic stem cells in different culture systems.” Biotechnology and Bioengineering 78 (2002) 442-453.
269
154. M. Mrksich, L.E. Dike, J. Tien, D.E. Ingber, G.M. Whitesides, “Using microcontact printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver.” Experimental Cell Research 235 (1997) 305-313.
155. H. Kaji, K. Takoh, M. Nishizawa, T. Matsue, “Intracellular Ca2+ imaging for micropatterned cardiac myocytes.” Biotechnology and Bioengineering 81 (2003) 748-751.
156. K.Y. Suh, J. Seong, A. Khademhosseini, P.E. Laibinis, R. Langer, “A simple soft lithographic route to fabrication of poly(ethylene glycol) microstructures for protein and cell patterning.” Biomaterials 25 (2004) 557-563.
157. M.R. Dusseiller, D. Schlaepfer, M. Koch, R. Kroschewski, M. Textor, “An inverted microcontact printing method on topographically structured polystyrene chips for arrayed micro-3-D culturing of single cells.” Biomaterials 26 (2005) 5917-5925.
158. E.W.H. Jager, C. Immerstrand, K.H. Peterson, K. Magnusson, I. Lundström, O. Inganäs, “The cell clinic: closable microvials for single cell studies.” Biomedical Microdevices 4 (2002) 177-187.
159. A. Folch, B. Jo, O. Hurtado, D.J. Beebe, M. Toner, “Microfabricated elastomeric stencils for micropatterning cell cultures.” Journal of Biomedical Materials Research 52 (2000) 346-353.
160. M.J. Rosenbluth, W.A. Lam, D.A. Fletcher, “Force microscopy of nonadherent cells: a comparison of leukemia cell deformability.” Biophysical Journal 90 (2006) 2994-3003.
161. J. Fukuda, K. Nakazawa, “Orderly arrangement of hepatocytes spheroids on a microfabricated chip.” Tissue Engineering 11 (2005) 1254-1262.
162. N. Ferrell, D. Hansford, “Fabrication of micro- and nanoscale polymer structures by soft lithography and spin dewetting.” Macromolecular Rapid Communications 28 (2007) 964-967.
163. F. Braet, R. De Zanger, E. Wisse. Drying cells for SEM, AFM and TEM by hexamethyldisilazane: a study on hepatic endothelial cells. Journal of Microscopy 186 (1997) 84-87.
164. B.D. Ratner, A.S. Hoffman, F.J. Schoen, and J.E. Lemons, Biomaterials Science An Intoduction to Materials in Medicine, Elsevier Academic Press, San Diego, CA (2004) p. 67.
270
165. H. Becker and U. Heim, “Hot embossing as a method for the fabrication of polymer high aspect ratio structures.” Sensors and Actuators A 83 (2000) 130-135.
166. L.J. Lee, M.J. Madou, K.W. Koelling, S. Daunert, S. Lai, and C.G. Koh, “Design and fabrication of CD-like microfluidic platforms for diagnostics: polymer-based microfabrication.” Biomedical Microdevices 3 (2001) 339-351.
167. B. Ziaie, A. Baldi, M. Lei, Y. Gu, and R.A. Siegel, “Hard and soft micromachining for BioMEMS: a review of techniques.” Advanced Drug Delivery Reviews 56 (2004) 145-172.
168. G. Wei, B. Bhushan, N. Ferrell, D. Hansford, “Microfabrication and nanomechanical characterization of polymer microelectromechanical systems for biological applications.” Journal of Vacuum Science and Technology A 23 (2005) 811-819.
169. K.E. Peterson, “Silicon as a mechanical material.” Proceedings of the IEEE 70 (1982) 420-457.
170. J.A. Forrest, D. Dalnoki-Veress, “The glass transition in thin polymer films.” Advances in Colloid and Interface Science 94 (2001) 167-196.
171. J.L. Keddie, R.A.L. Jones, R.A. Cory, “Size-dependent depression of the glass transition temperature in polymer films.” Europhysics Letters 27 (1994) 59-64.
172. B. Bhushan, A.V. Kulkarni, W. Bonin, J. Wyrobek, “Nano/picoindentation measurement using a capacitance transducer system in atomic force microscopy.” Philosophical Magazine A 74 (1996) 1117-1128.
173. S.J. Bull, “Nanoindentation of coating.” Journal of Physics D 38 (2005) R393-R413.
174. A.C. Fischer-Cripps, “Critical review of analysis and interpretation of nanoindentation test data.” Surface and Coatings Technology 200 (2006) 4153-4165.
175. X. Li, B. Bhushan, “A review of nanoindentation continuous stiffness measurement technique and it applications.” Materials Characterization 48 (2002) 11-36.
176. I.H. Loh, M.S. Sheu, A.B. Fishcer, Biocompatible polymer surfaces, in Desk Reference of Functional Polymers: Synthesis and Applications (edited by R. Arshady) American Chemical Society, Washington, D.C. (1997).
271
177. D.B. Holt, P.R. Gauger, A.W. Kusterbeck, F.S. Ligler, “Fabrication of a capillary immunosensor in polymethyl methacrylate.” Biosensors and Bioelectronics 17 (2002) 95-103.
178. A. van der Berg (ed.), Lab-on-a-chip Chemistry in Miniaturized Synthesis and Analysis Systems, Elsevier, Amsterdam (2003).
179. B. Ellis, Polymers: A Properties Database, Available on compact disk, CRC, Boca Raton, FL (2000).
180. J. Brandrup, E.H. Immergut, E.A. Grulke (eds.), Polymer Handbook, Wiley, New York, NY (1999).
181. J.E. Mark, Polymer Data Handbook, Oxford University Press, Oxford (1999).
182. T.G. van Kooten, H.T. Spijker, H.H. Busscher, “Plasma-treated polystyrene surfaces: model surface for studying cell-biomaterial interactions.” Biomaterials 25 (2004) 1735-1747.
183. M. Alexandre, P. Dubois, “Polymer-layered silicate nanocomposites: preparation, properties and uses of a new class of materials.” Materials Science and Engineering 28 (2000) 1-63.
184. S.S. Ray, M Okamoto, “Polymer/layered silicate nanocomposites: a review from preparation to processing.” Progress in Polymer Science 28 (2003) 1539-1641.
185. J. Njuguna, K. Pielichowski, “Polymer nanocomposites for aerospace applications: properties.” Advanced Engineering Materials 5 (2003) 769-778.
186. F. Gao, “Clay/polymer composites: the story.” Materials Today 7 (2004) 50-55.
187. G. Beyer, “Nanocomposites: a new class of flame retardants for polymers.” Plastic Additives and Composites 4 (2002) 22-28.
188. M. Kurian, M.E. Galvin, P.E.. Trapa, D.R. Sadoway, A.M. Mayes, “Single ion conducting polymer-silicate nanocomposite electrolytes for lithium battery applications.” Electrochimica Acta 50 (2005) 2125-2134.
189. R.Y. Lockhead, C.R. Haynes. S.R. Jones, V. Smith, “the high throughput investigation of polyphenolic coupler in biodegradable packaging materials.” Applied Surface Science 252 (2006) 2535-2548.
190. H. Liu, B. Bhushan, “Investigation of nanotribological properties of self-assembled monolayerss with alkyl and biphenyl spacer chains.” Ultramicroscopy 91 (2002) 185-202.
272
191. W.C. Oliver, G.M. Pharr, “An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments.” Journal of Materials Research 7 (1992) 1564-1583.
192. B. Bhushan, X. Li, “Nanomechanical characterization of solid surfaces and thin films.” International Materials Review 48 (2003) 125-164.
193. H. Liu, B. Bhushan, “Investigation of nanotribological properties of self-assembled monolayers with alkyl and biphenyl spacer chains.” Ultramicroscopy 91 (2002) 185-202.
194. W.C. Young, R.G. Budynas, Roark’s Formulas for Stress and Strain, Wiley, New York, NY (2002).
195. A. Yamamoto, S. Mishim, N. Maryuama, M. Sumita, “A new technique for direct measurement of the shear force necessary to detach a cell from a material.” Biomaterial 19 (1998) 871-879.
196. K.A. Athanasiou, B.S. Thoma, D.R. Lanctot, D. Shin, C.M. Agrawal, R.G. LeBaron, “Development of the cytodetachment technique to quantify mechanical adhesiveness of a single cell.” Biomaterial 20 (1999) 2405-2415.
197. D.A. Stenger, G.W. Gross, E.W. Keefer, K.M. Shaffer, J.D. Andreadis, W. Ma, J.J. Pancrazio, “Detection of Physiologically Active Compound Using Cell-based Biosensors.” Trends in Biotechnology 19 (2001) 304-309.
198. M.C. Beckerle, Cell Adhesion, Oxford University Press, Oxford (2001).
199. C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, “Geometic control of cell life and death.” Science 276 (1997) 1425-1428.
200. A.R. Horwitz, J.T. Parsons, “Cell migration—movin’ on.” Science 286 (1999) 1102-1103.
201. E. Crowley, A. Horwitz, “Tyrosine phosphorylation and cytoskeletal tension regulate the release of fibroblast adhesions.” Journal of Cell Biology 131 (1997) 525-537.
202. M. Chicurel, R. Singer, C. Meyer, D. Ingber, “Integrin binding and mechanical tension induce movement of mRNA and ribosomes to focal adhesions.” Nature 392 (1998) 730-733.
203. D. Paulin, A. Huet, L. Khanamyrian, Z. Xue, “Desminopothies in muscle disease.” Journal of Pathology 204 (2004) 418-427.
204. M. Pekny, M. Pekna, “Astrocyte intermediate filaments in CNS pathologies and regeneration.” Journal of Pathology 204 (2004) 428-437.
273
205. H. Yamaguchi, J. Condeelis, “Regulation of the actin cytoskeleton in cancer cell migration and invasion.” Biochimica et Biophysica Acta 1773 (2007) 642-652.
206. Z.Z. Wu, G. Zhang, M. Long, H.B. Wang, G.B. Song, S.X. Cai, “Comparison of the viscoelecastic properties of normal hepatocytes and hepatocellular carcinaoma cells under cytoskeletal perturbation.” Bioreology 37 (2000) 279-290.
207. M.J. Rosenbluth, W.A. Lam, D.A. Fletcher, “Force microscopy of nonadherent cells: a comparison of leukemia cell deformability.” Biophysical Journal 90 (2006) 2994-3003.
208. L.H. Miller, D.I. Baruch, K. Marsh, O.K. Doumbo, “The pathogenic basis of malaria.” Nature 415 (2002) 673-679.
209. G.D. Sempowski, M.A. Borrello, T.M. Blieden, K. Barth, R.P. Phipps, “Fibroblast heterogeneity in the healing wound.” Wound Repair and Regeneration 3 (1995) 120-131.
210. V.C. Sandulache, A. Parekh, H.-S. Li-Korotky, J.E. Dohar, P.A. Hebda, “Prostaglandin E2 differentially modulates human fetal and adult dermal fibroblast migration and contraction: implication for would healing.” Wound Repair and Regeneration 14 (2006) 633-643.
211. J.A. Cooper, “Effects of cytochalasin and phalliodin on actin.” The Journal of Cell Biology 105 (1987) 1473-1478.
212. M.R. Bubb, I. Spectro, B.B. Beyer, K.M. Fosen, “Effects of jasplakinolide on the kinetics of actin polymerization.” The Journal of Biological Chemistry 275 (2000) 5163-5170.
213. M. Knoblauch, G. A. Noll, T. Müller, D. Prüfer, I. Scheider-Hüther, A. J. E. van Bel, “ATP-independent contractile proteins from plants.” Nature Materials 2 (2003) 600-603.
214. M. Knoblauch, W. S. Peters, “Biomimetic actuators: where technology and cell biology meet.” Cell and Molecular Life Sciences 61 (2004) 2497-2509.
215. S. Schwan, M. Fritzsche, A. Cismak, A. Heilmann, U. Spohn, “In vitro investigation of the geometric contraction behavior of chemo-mechanical P-protein aggregates (forisomes). Biophysical Chemisty 125 (2007) 444-452.
216. S. Schwann (unpublished data).
274
275
217. M. Palacio, B. Bhushan, N. Ferrell. D. Hansford, “Nanomechanical characterization of polymer beam structures for BioMEMS applications.” Sensors and Actuators A 135 (2007) 637-650.
218. R. C. Hibbeler, Mechanics of Materials, Prentice Hall, New York, NY, 4th Edition (1997).
219. N. Ferrell, J. Woodard, D. Hansford, “Fabrication of polymer microstructures for MEMS: sacrificial layer micromolding and patterned substrate micromolding.” Biomedical Microdevices 9 (2007) 815-821.
220. M. Ilie, B. M„rculescu, N. Moldovan, N. Nastase, M. Olteanu, “Adhesion between PMMA mask layer and silicon wafer in KOH aqueous solution.” in Materials andDevice Characterization in Micromachining, edited by C. R. Friedrich and Y.Vladimirsky, (Proc. of SPIE 3512, Santa Clara, CA, 1998) pp. 422-430.
221. D. S. Bodas, S. K. Mahapatra, and S. A. Gangal, “Comparative study of spin coated and sputtered PMMA as an etch mask material for silicon micromachining.” Sensors and Actuators A 120 (2005) 582-588.
222. D. S. Bodas and S. A. Gangal, “Poly(methyl methacrylate) as masking material for microelectromechanical system (MEMS) fabrication.” Journal of Applied Polymer Science 102 (2006) 2094-2098.
223. J. Guan, N. Ferrell, B. Yu, D.J. Hansford, L.J. Lee, “Simultaneous fabrication of hybrid arrays of nanowires andmicro/nanoparticles by dewetting on micropillars.” Soft Matter 3 (2007) 1369-1371.
224. R.T. Butler, N.J. Ferrell, D.J. Hansford, “Spatial and geometrical control of silicification using a patterned poly-L-lysine template.” Applied Surface Science 252 (2006) 7337–7342.
225. H. Borteh, N. Ferrell, R. Butler, S. Olesik, D. Hansford, “Electroless Deposition of Gold Nanoparticles Over Silicon-based Substrates.” in Materials, Processes, Integration and Reliability in Advanced Interconnects for Micro- and Nanoelectronics, edited by Q. Lin, E. T. Ryan, W. Wu, D.Y. Yoon, (Mater. Res. Soc. Symp. Proc. 990, Warrendale, PA, 2007) 0990-B08-17.