University of Massachusetts Amherst ScholarWorks@UMass Amherst

Doctoral Dissertations 1896 - February 2014

1-1-2006

I. Composite polymer coatings prepared in supercritical carbon dioxide. II. Chemical modification of atomic force microscope probes. III. Liquid mobility on surfaces with patterned chemistry and topography.

Kevin A. Wier University of Massachusetts Amherst

Follow this and additional works at: https://scholarworks.umass.edu/dissertations_1

Recommended Citation Wier, Kevin A., "I. Composite polymer coatings prepared in supercritical carbon dioxide. II. Chemical modification of atomic force microscope probes. III. Liquid mobility on surfaces with patterned chemistry and topography." (2006). Doctoral Dissertations 1896 - February 2014. 1088. https://scholarworks.umass.edu/dissertations_1/1088

This Open Access Dissertation is brought to you for free and open access by ScholarWorks@UMass Amherst. It has been accepted for inclusion in Doctoral Dissertations 1896 - February 2014 by an authorized administrator of ScholarWorks@UMass Amherst. For more information, please contact [email protected].

University of Massachusetts Amherst

I B R A R Y 1

:

5

!

)

I

This is an authorized facsimile, made from the microfilm master copy of the original dissertation or master thesis published by UMI.

The bibliographic information for this thesis is contained in UMFs Dissertation Abstracts database, the only central source for accessing almost every doctoral dissertation accepted in North America since 1861.

® Dissertation UMI Services

From:Pro(jsuest COMPANY

300 North Zeeb Road P.O. Box 1346 Ann Arbor, Michigan 48106-1346 USA

800.521.0600 734.761.4700 web www.il.proquest.com

Printed in 2006 by digital xerographic process on acid-free paper

I. COMPOSITE POLYMER COATINGS PREPARED IN SUPERCRITICAL CARBON DIOXIDE

II. CHEMICAL MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES

III. LIQUID MOBILITY ON SURFACES WITH PATTERNED CHEMISTRY AND TOPOGRAPHY

A Dissertation Presented

by

KEVIN A. WIER

Submitted to the Graduate School of the University of Massachusetts Amherst in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

May 2006

Polymer Science and Engineering UMI Number: 3215774

UMI

UMI Microform 3215774 Copyright 2006 by ProQuest Information and Learning Company.

All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 © Copyright by Kevin A. Wier 2006

All Rights Reserved I. COMPOSITE POLYMER COATINGS PREPARED IN SUPERCRITICAL CARBON DIOXIDE

II. CHEMICAL MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES

III. LIQUID MOBILITY ON SURFACES WITH PATTERNED CHEMISTRY AND TOPOGRAPHY

A Dissertation Presented

by

KEVIN A. WIER

Approved as to style and content by:

Thomas J. McCarthy, Chair

Alfred J. Crosby, Member

Anthony D. Dinsmore, Member

Shaw Ling Hsu, Department Head Polymer Science and Engineering I. COMPOSITE POLYMER COATINGS PREPARED IN SUPERCRITICAL CARBON DIOXIDE

II. CHEMICAL MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES

III. LIQUID MOBILITY ON SURFACES WITH PATTERNED CHEMISTRY AND TOPOGRAPHY

A Dissertation Presented

by

KEVIN A. WIER

Approved as to style and content by:

Thomas J. McCarthy, Chair

Anthony D. Dinsmore, Member

Shaw Ling Hsu, Department Head Polymer Science and Engineering

ACKNOWLEDGMENTS

My years spent in graduate school would not have been as successful or

enjoyable without the support and encouragement of many friends and colleagues. I would like to thank my advisor Professor Tom McCarthy for his guidance and direction during my graduate career. His enthusiasm and creative perspectives about science were a great example of how interesting and exciting research could be. Professor Al

Crosby not only made helpful remarks about my research as a member of my committee, but was also a valuable resource of information during my job search and

about graduate school in general. I also thank my other committee member Professor

Tony Dinsmore for his comments and critiques of my research.

The most rewarding part of graduate school for me was the wonderful people I

met. I want to thank the past and present members of the McCarthy group for making it

enjoyable to come to lab each day and for all their help and camaraderie throughout the years. Thanks to Jack, Jeff, Chris, Xinqiao, Margarita, Wei, Ebru, Taehyung, Zhixiang,

Sung In, like, Lichao, Yufeng, Scott, and especially Misha, Jung Ah, and Jay. I was also blessed with wonderful classmates and housemates who made my time spent in

Amherst particularly pleasant. There is more to life than just science and I want to thank Bryan, Greg, Lachelle, Drew, Matt, Sian, and Brad for providing much needed respites from the stresses of graduate school.

I am especially grateful for the love and support of my parents. Bob and Mary, and sisters, Kristin and Sherrie. Their faith and unwavering encouragement helped me throughout this life-changing journey.

iv ABSTRACT

I. COMPOSITE POLYMER COATINGS PREPARED IN SUPERCRITICAL CARBON DIOXIDE

II. CHEMICAL MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES

III. LIQUID MOBILITY ON SURFACES WITH PATTERNED CHEMISTRY AND TOPOGRAPHY

MAY 2006

KEVIN A. WIER, B.S., MICHIGAN TECHNOLOGICAL UNIVERSITY

Ph.D., UNIVERSITY OF MASSACHUSETTS AMHERST

Directed by: Professor Thomas J. McCarthy

The initial part of this dissertation describes the preparation of the first reported

poly(p-xylylene) polymer/polymer composites. Poly(/?-xylylene) (PPXN) and its derivatives, known collectively as parylenes, are solvent resistant and blends or composites cannot be easily made by conventional methods. Supercritical carbon dioxide was used as both a plasticizer and solvent to infuse and polymerize a variety of vinyl monomers inside the parylene films. Infrared spectroscopy, wide angle X-ray diffraction, and thermal gravimetric analysis were used to characterize the composites.

Multilayer coatings of PPXN and other polymer films were prepared and selectively modified with metal nanoparticles.

The second part details the modification of atomic force microscope (AFM) probes using a variety of monochlorosilanes to improve the chemical sensitivity of

AFM. X-ray photoelectron spectroscopy revealed that the silane reaction was successful. Adhesion force measurements between the modified probes and similarly modified silicon wafers were performed, but showed only a slight variance between the

V different tip chemistries. Surfaces with patterned chemistry were prepared and

examined with the modified probes using tapping mode AFM. The contrast in the

phase images was dependent on the tip chemistry.

The ability to confine and direct the motion of hquid droplets on a surface using

only gravity and differences in surface chemistry is discussed in the first chapter of Part in. A hydrophobic alkylsilane surface with low contact angle hysteresis was patterned with lines of a more hydrophobic fluoroalkyl silane. Liquid droplets moved easily on the low hysteresis matrix, but pinned at the more hydrophobic lines. A variety of patterns were used to demonstrate that a decrease in the hysteresis reduces the force needed to induce drop motion and also lowers the barriers that are needed to confine the droplets.

Condensation on a variety of ultrahydrophobic surfaces was examined in the

second chapter of Part III. Optical microscopy showed that water condensed between the hydrophobic surface features before being expelled to the top. The condensed liquid pinned the contact line of a macroscopic droplet and dynamic contact angle measurements revealed an increase in hysteresis which corresponded to a decrease in liquid mobility.

vi 81

TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS iv

ABSTRACT v

LIST OF TABLES xii

LIST OF FIGURES xiv

CHAPTER

1 . POLY(P-XYLYLENE) COMPOSITE COATINGS PREPARED IN

SUPERCRITICAL CARBON DIOXIDE 1

1.1 Introduction 1

1 .2 Background 3

1.2.1 Supercritical Carbon Dioxide 3 1.2.2 Poly(/?-xylylene) 6 1.2.3 Chemical Modification of Poly(p-xylylene) 10

1.3 Experimental 13

1.3.1 Materials 13

1.3.2 Experimental Procedures 14

1.3.2.1 Modification of Silicon Wafers 14 1.3.2.2 Chemical Vapor Deposition of Poly(/;-xylyene) 15 1.3.2.3 Absorption Kinetics 15 1.3.2.4 Polymer Composite Synthesis 15 1.3.2.5 Metal Composite Synthesis 16 1.3.2.6 PPXN/PECA Multilayer Coatings 17

1.3.3 Characterization 18

1.4 Results and Discussion 18

1 .4. 1 Vapor Deposition of Poly(p-xylylene) 1 1.4.2 Phase Behavior and Absorption Kinetics 20

1 .4.3 PPXN Polymer Composites 2

1 .4.4 Poly(chloro-/?-xylylene) Composites 32

1 .4.5 Polymer Composites on Silicon Wafers 33

Vll 1.4.6 Composite Characterization 34

1.4.6.1 Thermal Properties 34

1 .4.6.2 X-ray Diffraction 35

1.4.7 Parylene/Metal Composites 37 1.4.8 PPXN/PECA Multilayer Templates 40

1.5 Conclusions 41

1.6 References 43

2. MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES FOR IMPROVED SURFACE CHEMICAL IDENTIFICATION 51

2.1 Introduction 51 2.2 Background 52

2.2.1 Atomic Force Microscope 52

2.2. 1 . 1 Contact Mode AFM 52

2.2. 1 .2 Tapping Mode AFM 54

2.2.2 AFM Probe Surface Chemistry 55 2.2.3 AFM Probe Surface Modification 56 2.2.4 Surface Chemical Analysis Using AFM 58

2.2.4. 1 Force Curves 58 2.2.4.2 Lateral/Friction Force Images 60 2.2.4.3 Tapping Mode Phase Images 61

2.3 Experimental 63

2.3.1 Materials 63 2.3.2 Experimental Procedures 64

2.3.2.1 Silane Modification 64 2.3.2.2 Silicon Dioxide Growth and Modification 64 2.3.2.3 Microcontact Printing 65

2.3.3 Characterization 66

2.3.3.1 Chemical Force Microscopy 67 2.3.3.2 X-ray Photoelectron Spectroscopy 67

2.4 Results and Discussion 70

Vlll 2.4.1 Silane Modification of Probes and Surfaces 70 2.4.2 Adhesion Force Measurements 74 2.4.3 Tris(TMS)/Si02/Silane Surfaces 78

2.4.3.1 Silicon Dioxide Growth and Modification 78 2.4.3.2 Phase Images 83

2.4.4 Microcontact Printed Thiol Surfaces 85

2.4.4.1 Preparation 85 2.4.4.2 AFM Characterization 90

2.5 Conclusions 92 2.6 References 94

3. TWO DIMENSIONAL DROPLET FLUIDICS BASED ON DIFFERENTIAL LYOPHOBICITY 100

3.1 Introduction 100 3.2 Background 101

3.2.1 Contact Angle, Wettability, and Hysteresis 101 3.2.2 Liquid Mobility in Micro and Droplet Fluidic Devices 106

3.3 Experimental 108

3.3.1 Materials 108 3.3.2 Experimental Procedures 108

3.3.2.1 UV Pattern Formation 108 3.3.2.2 Characterization 109

3.4 Results and Discussion 1 10

3.4.1 Patterned Silane Surfaces 1 10

3.4.2 Characterization by XPS and Contact Angle 1 14

3.5 Conclusions 121 3.6 References 122

4. CONDENSATION ON ULTRAHYDROPHOBIC SURFACES AND ITS EFFECT ON WATER DROPLET MOBILITY 126

4.1 Introduction 126

ix 4.2 Background 126

4.2.1 Superhydrophobic Surfaces 126 4.2.2 Hysteresis on Rough Hydrophobic Surfaces 129 4.2.3 Wetting Transitions 130

4.3 Experimental 132

4.3.1 Materials 132 4.3.2 Experimental Procedures 133

4.3.2.1 Ultrahydrophobic Surfaces 133 4.3.2.2 Condensation Experiments 133

4.4 Results and Discussion 134

4.4.1 Preparation of Ultrahydrophobic Surfaces 134 4.4.2 Condensation on Ultrahydrophobic Surfaces 137

4.4.2.1 Optical Microscopy 137 4.4.2.2 Contact Angle Measurement 141 4.4.2.3 Capillary Pressure 149 4.4.2.4 Nanoscale Roughness 152 4.4.2.5 Low Hysteresis during Condensation 153

4.5 Conclusions 154 4.6 References 155

APPENDICES

A. SUPPLEMENTAL DATA FOR CHAPTER 2 AND 4 159

A.l Chapter 2 159

A. 1.1 Tris(TMS) Silylation in scCO. 159 A. 1.2 Phase Contrast on Block Copolymers 161

A.2 Chapter 4 163 A. 3 References 168

X B. EXPERIMENTS WITH PDMS 169

B.l Buckled PPXN on PDMS 169 B.2 PDMS in Anodized Aluminum Membranes 175 B.3 References 180

BIBLIOGRAPHY 182

xi LIST OF TABLES

Table Page

L L Physical properties of parylene polymers 10

2. 1 . XPS analysis of silane-modified silicon tapping mode AFM probes 73

2.2. XPS analysis of silane-modified silicon nitride contact mode AFM probes 74

2.3. Normalized adhesion force (N/m x 10') between alkylsilane- modified silicon nitride contact mode AFM probes and similarly modified silicon wafers, measured in air 75

2.4. Normalized adhesion force (N/m x 10"^) between alkylsilane- modified silicon nitride AFM contact mode probes and similarly modified silicon wafers, measured in water 77

2.5. Kinetics of the vapor phase reaction of Tris(TMS)Cl with a silicon wafer 79

2.6. Roughness parameters for Tris(TMS)Cl-modified surfaces determined by AFM 82

2.7. XPS analysis of chemically modified silica posts grown from an incomplete Tris(TMS) monolayer 82

2.8. Degrees of phase contrast between chemically modified AFM tips and similarly modified silica post surfaces 85

2.9. XPS analysis of gold surfaces before and after modification with alkanethiols 86

3.1. XPS analysis of an alkylsilane modified silicon wafer after being coated with a resist layer, exposed to UV radiation, developed, and plasma cleaned 112

4. 1 . Dynamic contact angle measurements on various hydrophobic silicon post surfaces before (dry) and after (wet) condensation 146

4.2. Experimental and calculated parameters describing the various hydrophobic post surfaces 151

Xll A. 1 . Comparison of vapor phase Tris(TMS)CI silylation of silicon wafers with the same reaction using hquid CO2 or scCOo as the solvent 161

A. 2. Degrees of phase contrast between the different phases of block copolymer films as detemiined using chemically modified AFM tips 162

xni LIST OF FIGURES

Figure Page

1 . 1 . Schematic diagram of the general procedure to prepare

polymer/polymer composites using SCCO2 1

1.2. General pressure-temperature phase diagram showing the supercritical region.'^ Reproduced by pemiission of The Royal Society of Chemistry 3

1.3. Carbon dioxide density as a function of pressure at different temperatures 4

1.4. (a) Poly(p-xylylene) and (b) poly(chloro-/?-xylylene) 6

1 .5. Vapor deposition polymerization of poly(/?-xylylene) 7

1.6. Vapor deposition polymerization apparatus 19

1.7. A representative desorption profile of CO2 from PPXC after a 2 hr

soak at 40 1 500 psi. Extrapolation to t=0 gives the mass uptake after the soak period 21

1.8. Infrared spectra of a PPXN film (a) immediately after reaction with

neat styrene and AIBN (40 T for 8 hrs / 80 "C for 13 hrs); (b),

(c) after subsequent 1 and 5 min rinses in toluene respectively 23

1.9. Infrared spectra of PPXN film (a) after reaction with styrene and

AIBN m scC02 (40 "C, 4 hrs / 80 T, 4 hrs); (b), (c) after 5 and 60 min rinses in toluene respectively 23

1.10. Infrared spectra showing a PPXN/PMMA composite after a 5 min rinse in THF and after one week of THF soxhlet extraction 25

1.11. IR spectra of PPXN homopolymer and composites with PMMA and PS. Arrows indicate characteristic peaks that were used to determine composite composition 25

1.12. Relative PMMA concentration in a PPXN/PMMA composite as a function of soak time (5 wt% MM A, 0.3 mol% AIBN) 29

1.13. Relative PMMA concentration in a PPXN/PMMA composite as a

function of reaction time (4 hr / 40 °C soak, 80 ''C reaction, 0.3 mol% AIBN) 29

xiv 1.14. Relative polystyrene concentration in a PPXN/PS composite as a

function of initial styrene concentration (4 hr / 40 T soak, 4 hr / 80 T reaction, 0.3 mol% AIBN initiator) 30

1.15. Relative PMMA concentration in a PPXN/PMMA composite as a

function of initial MMA concentration (4 hr / 40 soak, 4 hr / 80 °C reaction, 0.3 mol% AIBN initiator) 30

1.16. PMMA concentration in a PPXC/PMMA composite as a function

of reaction time (4 hr / 80 T soak, 5 wt% MMA, 0.3 mol% TBPB) 31

1.17. Comparison of PMMA concentration in a PPXC/PMMA composite as a function of initial MMA concentration using two different initiators, AIBN and TBPB (4 hr soak, 4 hr reaction, 0.3 mol% initiator). Inset shows AIBN data in more detail 31

1.18. Transmission IR spectra of PPXN/PMMA and PPXC/PMMA composites on silicon wafers. The carbonyl stretching vibration

at 1734 cm ' is highlighted 34

1.19. Thermal gravimetric analysis of PMMA, PPXC, and a PPXC composite film containing 30 wt% PMMA 35

1 .20. Wide angle X-ray diffraction data for PPXC and a composite containing 40 wt% PMMA 36

1.21. Hydrogen reduction of [Ag(COD)(hfac)] and [Pt(C0D)Me2] to deposit the con^esponding metal 38

1.22. Transmission FT-IR spectra of (a) PPXN, (b) [Ag(COD)(hfac)] and (c) [Ag(COD)(hfac)] in PPXN. Arrows indicate characteristic peaks for the silver precursor in the composite 39

1.23. TEM image showing the preferential deposition of platinum in the poly(ethyl-2-cyanoacrylate) layer surrounded by poly(p- xylylene).'"^ 40

2. 1 . Schematic diagram of an atomic force microscope 52

2.2. An example of a force-distance curve used to measure tip-sample adhesive forces 54

2.3. (a) Tapping mode cantilever and tip (b) Contact mode cantilever and tip 55

XV 2.4. Possible reactions of trichloro- and monochlorosilanes with surface silanols 57

2.5. Force-distance cui-ves for COOH-CH3, CH3-CH3, and COOH- COOH tip-samples interactions in ethanol. Reprinted with permission from reference 45. Copyright 1995 American Chemical Society 59

2.6. AFM images of a COOH-terminated thiol square in a CH3 matrix, (a) height image, (b) lateral force image with a COOH-terminated AFM tip (c) lateral force image with a CH3 terminated tip. Reprinted with permission from reference 45. Copyright 1995 American Chemical Society 60

2.7. Tapping mode AFM phase images of a COOH-terminated thiol open square in a CH3-teiTninated thiol matrix with a (c) COOH-

modified AFM tip (d) CH3-modified tip. (25 jam x 25 [xm image)

Reprinted with pemiission from reference 5 1 . Copyright 1 998 American Chemical Society 62

2.8. Tapping AFM phase images of a 20:80 blend of PMMA/polybutadiene changing only the tapping force: (a) rsp=0.84 (b) rsp=0.73 (c) rsp=0.58. Reprinted with permission from reference 54. Copyright 2000 American Chemical Society 62

2.9. Silanes used for chemical modification of AFM probes; (a) 3- aminopropyldimethylethoxysilane, (b) 3,3,3-

trifluoropropyldimethylchlorosilane, (c) n- propyldimethylchlorosilane, (d) tridecafluoro-1,1,2,2- tetrahydrooctyldimethylchlorosilane 70

2. 10. SEM images of a silicon nitride AFM tip (a) before and (b) after a 2 minute oxygen plasma cleaning 71

2.1 1. (a) SEM image of a commercial niobium film with sharp features that are smaller than an AFM tip,^" (b) Contact mode AFM height image of a similar film 72

2.12. Preparation of surfaces having chemically modified silica posts surrounded by Tris(TMS) 79

2.13. AFM tapping mode height images of (a) plasma cleaned silicon wafer, (b) silicon wafer modified with Tris(TMS)Cl for 3 days, (c) Si02 posts grown in the incomplete Tris(TMS) monolayers (0.5 X 0.5 \im scan size, 2 nm z scale) 81

XVI 1

2.14 AFM tapping mode phase images of SiO: posts grown from Tris(TMS) monolayers (a) before and (b) after a lowpass filter algorithm was applied to reduce background noise 84

2.15 Schematic diagram of the process to microcontact print alkanethiols onto gold 86

2.16. (a) SEM image of OCT microcontact printed onto gold (circles are not covered by the thiol monolayer), (b) Optical micrograph of a

similar sample after the exposed gold was etched 2 hours in a 1

mM NaCN, 1 M NaOH solution, (c) Optical micrograph of water condensation on the hydrophilic MDA features of a surface microcontact printed with 32x32 jim OCT and MDA squares in a checkerboard pattern 88

2.17. Tapping mode AFM height (left) and phase (right) images of microcontact printed thiols on gold consisting of MDA circles surrounded by OCT. The AFM tip chemistry was different for each image: (a) -CH3, (b) F13, (c) -NH2, (d) poly(vinyl alcohol) 89

2.18. AFM height (left) and lateral force (right) images of (a) OCT matrix microcontact printed onto gold leaving bare gold circles, (b) OCT matrix with MDA-modified circles 92

3.1. Stafic contact angle of a drop of liquid on a surface 1 0

3.2. Schematic diagram and photograph of (a) Cassie-Baxter type wetting (b) Wenzel type wetting. Reprinted with permission from reference 6. Copyright 2003 American Chemical Society 104

3.3. Schematic diagram of the UV lithography process used to prepare surfaces with patterned silane chemistry 112

3.4. XPS line scan of fluorine concentration on a silicon wafer patterned

on one side with DMDCS and on the other with FDCS 1 15

3.5. XPS multiplex spectra of the C Is region on the (a) DMDCS side and (b) FDCS side of a patterned sample similar to that shown in Figure 3.4 115

3.6. Silicon wafer chemically patterned with sections of DMDCS and FDCS: (a) Water contact angle measurements on the DMDCS (left) and FDCS (right) sections, (b) Same surface held vertical with a 2 |iL water droplet on the DMDCS side pinned at the FDCS interface, (c) Water droplet sliding along the DMDCS/FDCS interface. 117

xvn 3.7. (a) Silicon wafer after chemical pattern formation, (b, c) TMCS modified silicon wafers with FDCS lines after immersion in

ethanol 1 17

3.8. DMDCS modified silicon wafer with FDCS lines. Water droplets placed at each invisible vertex are sequentially combined as the

sample is slightly tilted and rotated clockwise 1 18

3.9. (a) Pattern prepared with FDCS lines on a DMDCS background that directs water droplets outward through the gap under the influence of centrifugal force, (b) Pattern of hydrophobic lines that could be used to combine liquid droplets 120

4. 1 . The critical contact angle hysteresis, AO, plotted versus the mean

contact angle, 0, for a water droplet with a 0.5 mm radius on a surface tilted 90°. Below the critical line the drop can move but

above it the drop remains pinned to the surface. Reprinted from reference 26 128

4.2. Diagram of a water droplet on a rough, hydrophobic surface that is in either the (a) Cassie-Baxter or (b) Wenzel wetting regime 129

4.3. Contact angle as a function of roughness for an initially flat surface having a static contact angle of 120°. Reprinted with permission from reference.'^^ Copyright 1964 American Chemical Society 130

4.4. (a) Diagram showing the dimensions used to designate the various silicon post surfaces: W is the width of the post, Y and Z are the dimensions of the unit cell, (b) Illustrations of the top of an indented square post, IP, and a four armed star post, StP, (c) Representative SEM of lithographically prepared silicon posts 135

4.5. Optical micrographs of two different static contact angles on the same rough, hydrophobic surface. Image (a) was taken after the droplet was carefully placed on the surface and (b) was taken after the two drops in image (a) merged 136

4.6. Sequential optical micrographs of condensed water growth on

hydrophobized silicon posts: 40 \xm tall, 2 |im wide, 4x8 fim unit ('Mmgp4,Hnm^ cell White arrows indicate drops that protrude above the tops of the posts 138

xviii 4.7. Sequential optical micrographs of water evaporating from

hydrophobized silicon posts: 40 fim tall, 2 fim wide, 4x8 |im unit

cell (2^*"^SP'^•^^'"'). White arrows indicate drops that transitioned from sitting on top of the posts to wetting between them 138

4.8. Sequential environmental SEM images of water condensation on

hydrophobized silicon posts: 40 \im tall, 8 \im wide, 16x32 ^m unit cell ('^"SP'^-'-n 140

4.9. Optical micrograph of large condensed water droplets on a surface

of hydrophobized silicon posts: 40 \im tall, 2 |im wide, 4x8 \im (^Mnisp4.8^n.^ unit cell arrows indicate distortions of the contact line. Darker areas, both underneath and between the drops, indicate where water penetrates between the posts 141

4.10. Advancing and receding contact angles on a surface with

hydrophobized silicon posts: 40 |im tall, 2 |im wide, 4x8 |im unit

cell (2^^'"SP'*'^^'"): (a) dry surface (b) surface with condensed water

(needle diameter is 0.5 mm) 142

4.1 1. Comparison of water droplet mobility on surface ^^""sp-'-^^'"^ (a) before and (b) after condensation 144

4.12. (a-d) Sequential optical micrographs of condensed water growth

on hydrophobized silicon posts: 40 \im tall, 8 ^m wide, 32x80 fim unit cell (^^^s?^-'^"^^""). (e,f) Water drops easily roll off the

tilted surface when it is dry (e), but stick after condensation (0 144

4.13. (a) Hydrophobic spot on a hydrophilic matrix, (a) Water contact angle on the hydrophobic spot (b) Contact angle on a mixed surface 148

4.14. (a) A hydrophobic surface with nanoscale roughness, (b,c) Receding contact angle measurements on the surface before and

after condensation respectively (needle diameter is 0.5mm) 152

A. 1 . (a-d) Sequential optical micrographs of condensed water growth

on hydrophobized silicon posts: 40 [im tall, 8 \im wide, 32x32 (8M-Sp32,32Mn.) ^im unit cell J ^3

A.2. (a-d) Sequential optical micrographs of condensed water growth

on hydrophobized silicon posts: 40 |j,m tall, 8 ^m wide, 16x32 (^^Mmgp 16,32^01^ [im unit cell ^^^^ ^^^^^ ^^.^p^ ^^^-j^ ^.^jj ^j^^

tilted surface when it is dry (e), but stick after condensation (f) 164

xix A. 3. (a-d) Sequential optical micrographs of condensed water growth

on hydrophobized silicon posts: 40 ^im tall, 8 |im wide, 16x32 unit cell('^"IP''-''n 165

A.4. (a-d) Sequential optical micrographs of condensed water growth

on hydrophobized silicon posts: 40 ^im tall, 8 \xm wide, 16x32 (^Mmstpi^-^^Mm^ ^im unit cell 165

A. 5. (a-d) Sequential optical micrographs of condensed water growth on hydrophobized silicon: 32 ^im hollow wells. (e,f) Water

drops easily roll off the tilted surface when it is dry (e), but stick

after condensation (f) 166

A. 6. Advancing and receding contact angles on a dimethyldichlorosilane modified silicon wafer: (a) dry surface (b) surface with condensed water. Hysteresis remains low even after

condensation (needle diameter is 0.5 mm) 167

B. l. (a) PDMS film held in tension, (b) PPXN deposited on constrained PDMS, (c) Buckled PPXN film on unconstrained PDMS 170

B.2. (a) Gradient width PDMS sample, with arrows indicating the direction the sample was stretched and constrained and letters showing the approximate location of the following optical micrographs, (b,c,d) Buckled PPXN film on PDMS after the tension was released 171

B.3. Buckled PPXN on a gradient thickness PDMS sample held under tension in the horizontal direction. The thickness of the underlying PDMS layer varies for each image: (a) 2.3 mm, (b)

1 .6 mm, (c) 1 .3 mm, (d) 1 mm 1 73

B.4. Buckled PPXN on PDMS that was not constrained during PPXN deposition 173

B.5. (a) Bottom surface and (b) cross section of a commercial AAM 1 76

B.6. SEM cross-section images of PDMS rods prepared in an AAM after the membrane was dissolved. Arrows indicate a film of

PDMS connecting all the rods 1 78

B.7. Top view of PDMS rods prepared in an AAM, shown after the membrane was dissolved 179

XX CHAPTER 1

P0LY(/7-XYLYLENE) COMPOSITE COATINGS PREPARED IN SUPERCRITICAL CARBON DIOXIDE

1 1 Introduction

Supercritical carbon dioxide (scCO:) can be used as a solvent/plasticizer to prepare polymer composites.' '^ A vinyl monomer and initiator are infused into a polymer substrate with the help of scCO: and subsequently heated to initiate free radical

polymerization as shown in Figure 1 . 1. The second polymer phase is deposited in the amorphous regions of a semicrystalline polymer substrate resulting in a kinetically

trapped morphology which is different from conventional blends. This method proved applicable to a wide range of polymer substrates. Further research showed that anionic"^ and ring-opening metathesis polymerizations^ could also be performed in SCCO2- swollen polymers. These earlier reports focused on the modification of free standing polymer substrates. Here we demonstrate that the technique can also be used to prepare polymer composite coatings.

SCCOt SCCO? M I M M M I M M M M

I M M 1 I soak react M 1 M M vent M M M M I M

Polymer sample Infuse monomer and Polymer/polymer initiator composite

Figure 1.1. Schematic diagram of the general procedure to prepare polymer/polymer composites using SCCO2.

1 Poly(p-xylylene) (PPXN) and poly(chloro-/7-xylylene) (PPXC) films were chosen as substrates for this work. Parylenes, as they are generally called, are semi- crystalline polymers prepared by vapor deposition polymerization (VDP), which produces pin-hole free, conformal coatings. They have found modest commercial success as biocompatible coatings for medical devices and as insulating layers for microelectronics.^ Parylenes show excellent resistance to common organic solvents and are chemically inert.

Parylenes are of interest for a variety of reasons. First, the VDP process allows the film thickness to be precisely controlled from a few nanometers to tens of microns.

Our group also has expertise in the controlled growth of other polymer thin films including poly( vinyl alcoholf layered polyelectrolytes^"'*^, and poly(ethyl-2- cyanoacrylate)". A multilayered structure combining two or more of these conformal coatings may be of interest as a template for further reactions. Second, parylenes are often used as barrier layers because of their low gas and moisture permeability. This family of polymers offers a test to the limits of the scCO^ composite process. Finally, parylenes are relatively inert to chemical reagents and solvents, making solution

blending difficult. PPXN decomposes near its melting point which makes traditional melt blending tricky. Supercritical carbon dioxide offers a new method to prepare conformal composite polymer coatings.

2 1.2 Background

1 .2. 1 Supercritical Carbon Dioxide

Above a critical temperature and pressure, materials are no longer classified as

solids, liquids, or gases, but are instead supercritical fluids (Figure 1.2). This interesting

phase of matter has properties that are intemiediate between the liquid and gas phase. A

supercritical fluid has no surface tension, expands to fill the volume of its container, and

has a diffusivity similar to a gas. However, under most conditions it has a liquid-like

density. These properties have made supercritical fluids of interest as solvents for

various applications.

un CO

Temperature

Figure 1 .2. General pressure-temperature phase diagram showing the supercritical region.'" Reproduced by permission of The Royal Society of Chemistry.

3 Carbon dioxide has a critical temperature, 31.1°C, and critical pressure, 1070 psi,

that are relatively easy to attain.'^ Carbon dioxide is also nonflammable, nontoxic, and

inexpensive which is why it has received so much attention as an environmentally

friendly alternative to conventional liquid solvents.'"^"' Removal of residual carbon

dioxide in the finished product is also less of a problem because it is a gas at ambient conditions and simply diffiases out.

1

0.9

0.8

^0.7 ^xx^^^^"^ •

-§ 0.6 w/N !' -;,*r ,

0.5

. m -I- I „ „ I 0.4 U20C I + X , 40 C ° 0.3 60 C X -i- # »:80 , X » © ^ C 0.2 • 100C -120C 0.1

1000 2000 3000 4000 5000

Pressure (psi)

Figure 1.3. Carbon dioxide density as a function of pressure at different temperatures.

Unlike conventional solvents, the solvent quality of supercritical carbon dioxide,

SCCO2, can be controlled. The solubility parameter of a fluid is related to its density.

Figure 1 .3 shows how the density of scCO:, and thus the solubility parameter, can be varied by changing the temperature and pressure. For example, at 5800 psi and 40 "C

scCO: has a solubility parameter of 7.3 which is similar to hexane.' ^ Carbon dioxide

has no dipole moment, but a large quadrupole moment contributes to its solubility

parameter. Many nonpolar molecules dissolve in scCO:, but few polar ones do.'^ It is

4 also a poor solvent for high molecular weight polymers with the noted exception of

amorphous fluoropolymers'^'^" '^' and silicones

The interaction of SCCO2 with polymers has been widely studied and

reviewed. Wliile it is a poor solvent for most polymers, it is soluble in many."

The less dense amorphous regions are more easily penetrated by scCO^ than the

crystalline regions of a semi-crystalline polymer. The scCO: plasticizes, i.e., lowers the

glass transition temperature of many amorphous polymers. The diffusion rates

of small molecules are increased in scC02-swollen polymer substrates and that principle

has been used to impregnate polymers with various organic molecules such as

fragrances and dyes.^"'"^^

Along the same lines, Watkins and McCarthy showed that polymer/polymer

composites could be prepared by polymerization of a monomer that had been infused

into a solid polymer substrate. ' A vinyl monomer and initiator were allowed to diffuse into the polymer with SCCO2 serving a dual purpose as both the solvent for the monomer and plasticizer for the polymer. The sample was then either heated in the presence of SCCO2 to initiate free radical polymerization or the SCCO2 was vented and the sample then heated. Polymerization in the presence of SCCO2 was found to produce samples with a higher and more uniform concentration of the second polymer phase.

The monomer could be replenished from the SCCO2 phase as it was polymerized inside the polymer. The second polymer phase was deposited into the amorphous regions of the semicrystalline substrate resulting in a kinetically trapped morphology that was different from conventional blends.

5 This method proved appHcable to a wide range of polymer substrates/ Further

37 38 * research showed that anionic polymerization and ROMP could also take place m

scC02-swollen polymers. Polymer/metal composites could be prepared by infusion of

an organometallic precursor followed by hydrogen reduction to form metal

particles.^^'"*^ These reports focused on the modification of bulk polymer samples. In this work we will demonstrate that this technique can also be used to prepare polymer composite coatings.

1.2.2 Polv(;?-xvlvlene)

Poly(p-xylylene) (PPXN) and poly(chloro-jO-xylylene) (PPXC), shown in Figure

1 .4, are semicrystalline polymers that are most commonly prepared by a chemical vapor deposition technique that produces pinhole free, conformal coatings. Parylenes, as they are generally called, have found modest commercial success as insulating layers for electronics and as biocompatible coatings for medial devices. They show excellent resistance to common organic solvents and are chemically inert. Several reviews cover

'^'"'*^ the synthesis, properties, and applications of parylenes.^

a b

Figure 1.4. (a) Poly(/?-xylylene) and (b) poly(chloro-/>-xylylene).

6 1

Although parylenes can be prepared by wet chemical methods , chemical vapor

deposition, or more specifically vapor deposition polymerization ( VDP), is the preferred method to prepare useful polymeric materials. In the late 1940s, Szwarc produced

poly(/?-xylylene) by flash pyrolysis ofp-xylene under reduced pressure.^^'"^*^ However, the polymers formed in low yield, were insoluble, and had low molecular weight. Other

drawbacks were that the material was crosslinked and pyrolysis took place at 900- 1 200

"C, which made it intolerant of functional groups on the starting monomer.

Figure 1.5. Vapor deposition polymerization of poly(/?-xylylene).

The major advancement in parylene polymerization occurred in the 1960s when

Gorham published a VDP process that improved on the Szwarc method.''^'^'^ Figure 1.5

shows a typical reaction scheme. [2.2]Paracyclophane is placed at one end of the

reaction chamber and the substrates to be coated are placed at the other (see Figure 1 .6).

The system is pumped down to a minimum of 100 millitorr. When the dimer is gently

heated, 60-150 °C, it sublimes and passes through a 600-650 °C tube fiimace. The bridging ethylene bonds are cleaved which relieves the considerable ring strain, 3 kcal/mol, present in the molecule. Two /j-xylylene monomers are produced and diffuse

7 into the deposition chamber where they condense on the substrate and polymerize. The

Gorham method produces high molecular weight, uncrosslinked polymer at a much lower pyrolysis temperature compared to the Szwarc method.

Several important points about this process need to be noted. The system is kept under vacuum to ensure a low monomer concentration in the vapor phase. The monomers can react in the vapor phase to form dimers, but the dimer is not thermodynamically stable and quickly reverts to monomer. The concentration of

monomer increases when it condenses on the surface leading to the rapid fomiation of

the stable diradical trimer that initiates the polymerization." ' " Since the monomer

(trimer) itself is the initiator, there are no byproducts or side reactions, which makes this distinct from most other polymerizations. In order for the monomer to condense, the

substrate must be cooled below a certain ceiling temperature, which is related to the molecular weight and volatility of the monomer. The "ceiling temperatures" for poly(/?- xylylene) and poly(chloro-jO-xylylene) are 30°C and 90°C respectively.'''^

The exact polymerization mechanism is still being investigated and modeled,

but it is similar to living free radical because high molecular weight polymer is formed

rapidly and there are no bimolecular termination reactions.^'^"'^^'^'^ The polymerization actually takes place below the surface so the monomer must diffuse down into the sample to the growth interface. Eventually the growing chains become buried and monomer can no longer reach them. Termination by atmospheric contaminants can occur after the system has been brought to ambient conditions.

There is no liquid intermediate in the reaction as the monomer goes directly

from the gas phase to a high molecular weight, solid polymer. This process is unique in

8 that the fihti conformally coats the surface even in small, high aspect ratio gaps.

The reason is that parylenes have a low sticking coefficient, defined as the probability for a molecular particle to fix onto a solid surface at a single collision. Since many

/7-xylylene molecules must adsorb and evaporate before one reacts, the monomer

concentration is more evenly distributed across the surface before polymerization takes place. Typical liquid coatings tend to pool in large gaps and span small ones to form an uneven coating.

Film thickness and properties depend on a number of parameters, some of which have already been mentioned.^ A higher sublimation temperature increases the monomer flux and speeds up film growth. Lower substrate deposition temperature causes more monomer condensation and increases film growth. Lower system pressure decreases monomer concentration and slows film growth, but produces better coverage on high aspect ratio features. Above a certain pressure the monomer concentration in

the vapor phase reaches a point where polymerization is possible and the polymer falls like snow. Obviously, increasing the reaction time and the initial dimer amount will increase the final film thickness.

Some of the physical properties of PPXN and PPXC are listed in Table 1.1.

Both are semicrystalline and insoluble in conventional solvents. PPXN is only soluble

in chlorinated biphenyls and benzyl benzoate above 200 °C, while PPXC is soluble in

alpha-chloronaphthalene at -150 °C. PPXN is the preferred coating for applications

where lubricity is needed and also has a dielectric constant that is independent of frequency. PPXC has better barrier properties and meets specifications for military use.

Some applications for parylenes include coatings for electronic circuitry, sensors, and

9 '

medical devices. PPXN and PPXC both meet the FDA's Class VI requirements and are

used for needles, catheters, and implants.

Table 1.1. Physical properties of parylene polymers.^

Poly(p-xylylene) Poly(chloro-p-xylylene)

Meltmg point (T) 410 290

Glass transition ("C) 13 36

Refractive index 1.661 1.639

Coefficient of friction 0.25 0.29

Dielectric constant (at i ,000 hz) 2.65 3.1

CO2 permeability 214 7.7

(cm^-mil/100 in^-24 hr- atm(23°C))

1.2.3 Chemical Modification of Poly(/;-xylylene )

Various attempts have been made to broaden the utility of parylene polymers through chemical modification. Post deposition processes typically modify only the surface because the polymer is insoluble in most solvents. VDP of a functionalized monomer, copolymerization, or codeposition varies the bulk properties of the polymer.

Copolymers can also be made by solution polymerization, but will not be discussed.

There are several reviews covering functional parylenes, so only a few examples will be

"-^*' given here.^'

Adding functional groups to the aromatic rings or the bridging ethylenes of paracyclophane can vary the bulk properties of parylene. A broad spectrum of monomers have been used to prepare the corresponding parylene polymer with

10 substituent groups including methyl , ethyl , bromo , chloro , fluoro" , cyano ,

**^'^"^'^^. acetyl^*^, hydroxy^'", trimethysiloxy^', phenyl^^"^'"\ amino^''*'^^, and others^

However, the functional groups must be able to withstand the 600+ °C temperature of

the pyrolysis oven.

Vapor deposition polymerization of parylenes starting from monomers besides

cyclophanes has received attention recently/"''"^^'"^'^ Cyclophanes are difficult to prepare

in high yield because of the same inherent ring strain that makes them easy to

polymerize. Monomers such as a-substituted-, a,a'-disubstituted-, or a,a-

disubstituted- /?-xylylenes are easier to synthesize with a wide range of functional

groups. They can also be used to prepare functional parylenes in the same VDP setup

as cyclophanes although the polymer yields tend to be lower.

44 Vapor deposition copolymerization has been tried with vinyl monomers.

Maleic anhydride^'^'^'', n-vinylpyrolidone^^, 4-vinylpyridine^'^, styrene"*^'^^, and

perfluorooctyl methacrylate^*^'^' comonomers have all been used with varying success.

The different volatility of the reagents makes copolymerization difficult to control. The

substrate temperature is an important parameter, with lower temperatures generally giving better results.

Polymer composites containing metal or inorganic particles are of interest for a variety of applications because of their unique electrical, optical and magnetic

^'^'^^ properties. Parylenes can be codeposited with other materials capable of CVD to prepare such materials. Silicon dioxide/PPXC composites were prepared by

simuhaneous CVD of the dichloro[2.2]paracyclophane and diacetoxy-di-/- butoxysilane.^'' The composition and properties could be controlled by adjusting the

11 vaporization temperature. Metal vapors can also be codeposited with parylenes.

' ' " Silver \ germanium , iron , nickel lead , and magnesium are some of the metals that have been used. As an example of a possible application of these materials,

84 85 PPXN/lead composites showed promise as a gas sensor for ammonia. '

Post deposition surface modification of parylenes has also been performed. A

technique often used in industry is to expose a polymer to plasmas, which introduce various functional groups, such as hydroxyls and carbonyls, that make the surface hydrophilic and often improves adhesion.**^"^' However, in most cases, the process

allows little control over the final surface chemistry. A few studies have been done to add specific chemical functionality to the surface. Hydroxyl groups were added to

PPXN using ammonium sulfide^^'^^, and sulfuric acid was used to add sulfonate groups''^"^"'^'^"^^. Recently, a wide variety of functional groups was added to PPXN via electrophilic aromatic substitution reactions.''^

Another common method used to vary polymer properties is blending. However, since PPXN and PPXC are insoluble under reasonable conditions in common solvents

and PPXN decomposes near its melting point, conventional blending techniques cannot be used. Polymerization in scC02-swollen polymer substrates will allow access to parylene composites.

12 1.3 Experimental

1.3.1 Materials

Methyl methacrylate and styrene were purchased from Aldrich and distilled from calcium hydride under vacuum. 2,2'-Azobis(isobutyronitrile) (AlBN) was purchased from Aldrich and recrystallized from methanol. /e/'Z-Butyl peroxybenzoate

(TBPB), [2.2]paracylophane, tetrahydrofuran, (l,5-cyclooctadiene)dimethylplatinum

(11) [Pt(C0D)Me2], ( l,5-cyclooctadiene)(hexafluoroacetylacetonato)silver(l)

[Ag(COD)(hfac)] and toluene were purchased from Aldrich and used as received.

Methacryloxypropyltrimethoxysilane and trivinylchlorosilane were obtained from

Gelest and used as received. Carbon dioxide (Coleman Grade, 99.99%) was purchased from Merriam Graves and passed through columns containing activated alumina and copper catalyst (Engelhard Q-5) to remove water and oxygen, respectively. A motorized syringe pump, ISCO model lOODM, was used to pressurize the carbon dioxide. Prepurified hydrogen was purchased from Merriam Graves and added to a custom assembled piston that could be pressurized using the CO2 from the syringe pump. Poly(chloro-/7-xylylene) film (-35 \im) was provided by Paratronix, Inc.

(Attleboro, MA) and used as received. Silicon wafers were purchased from

International Wafer Service (<100> orientation, P/B doped, resistivity from 20.0 to 40

Q cm). House-purified water (reverse osmosis) was further purified using a Millipore

Milli-Q system (18.2 MQ-cm).

13 1.3.2 Experimental Procedures

1 .3.2. 1 Modification of Silicon Wafers

Silicon wafers were modified with methacryloxypropyltrimethoxysilane

(MPTS) and trivinylchlorosilane (TVCS) to improve the adhesion of the parylene films.

The native wafers were cleaned overnight in a solution containing 7 parts sulfiiric acid with 4 wt% dissolved sodium dichromate and 4 parts hydrogen peroxide. This

oxidizing mixture is extremely reactive toward hydrocarbons. The sodium dichromate

was first dissolved in the sulfuric acid and then it was slowly added to a container holding the hydrogen peroxide and silicon wafers. The orange solution became green, warmed to -90 °C, and generated lots of oxygen bubbles. The wafers were rinsed with water and dried following cleaning.

Two different procedures were used to functionalize the surfaces. The TVCS was applied by suspending the silicon wafers above a pool of the silane in a closed container heated to 60 °C in an oil bath. The vapor phase reaction was allowed to

proceed for 1 day after which the samples were removed and rinsed with toluene, ethanol, and water. The MPTS was applied by a solution reaction according to a

97 literature procedure. Twenty milliliters of methanol, 0. 1 mL water, 20 jaL of MPTS

were mixed and allowed to sit at room temperature for 2 hours. Cleaned silicon wafers were then added for 10 minutes, rinsed with methanol, and air dried. Modified wafers were characterized by XPS, ellipsometry, and contact angle.

14 1 .3.2.2 Chemical Vapor Deposition of Polv(/;-xvlvene)

The PPXN films used in this study were prepared in a home-built reactor using

the Gorham method/*^ [2.2]Paracyclophane was sublimed at 1 10 "C and pyrolyzed at

650 °C. The polymer was deposited on 1 .5 cm x 1 .3 cm silicon wafers, both clean and silane modified, that were cooled to 10 "C under a 15 mTorr vacuum. Film thickness ranged from 2.5-5 ]im. Free standing PPXN films were prepared by carefully peeling

the polymer off the walls of the deposition chamber and cutting 1 .5 cm x 1.3 cm samples. PPXC films were also prepared in a similar manner, although most composites were prepared using the commercially obtained PPXC.

1.3.2.3 Absorption Kinetics

The equilibrium solubility of CO2 in PPXC was determined as described by

Berens. " Preweighed samples of PPXC were placed in a reaction vessel and allowed to

soak in SCCO2 at 40 °C and 1 500 psi for a given amount of time. The samples were transferred to a balance within 30-45 seconds after rapidly venting the vessel. Mass loss was recorded as a function of time via a computer connected to the balance. The data was plotted as weight percent CO: versus the square root of time. Extrapolation of the initial linear portion of the curve to zero time gave the CO2 weight percent at the end of the soak period.

1.3.2.4 Polvmer Composite Synthesis

All reactions were run in 10 cm long, 1 .5 cm inner diameter, ~15 mL, type 316 stainless steel vessels with a needle valve at one end and a plug on the other. Samples

15 of PPXN or PPXC (-1.5 cm x 1.3 cm) either free standing or coated on silicon wafers

were placed in the vessel. The free standing PPXN films were placed in custom made

aluminum holders to keep them from folding up. Monomer (methyl methacrylate or

styrene) and initiator (AIBN or TBPB) were mixed and added to the vessel using a

syringe. (Monomer concentration was determined as a weight percentage in CO2 and the initiator concentration was always 0.3 mol% relative to monomer.) The vessel was then purged with carbon dioxide gas, weighed, and allowed to equilibrate in a 40 "C water bath. Carbon dioxide was added via a syringe pump at 40 °C and a pressure of

1500 psi when AIBN was the initiator or 1350 psi when TBPB was used. The vessel was reweighed to determine the mass of carbon dioxide added. The samples were allowed to soak for a given period of time at a temperature at which the half-life of the

initiator was hundreds of hours: 40 °C (1500 psi) for AIBN and 80 T (3320 psi) for

TBPB. The temperature was then raised to initiate polymerization: 80 ''C (2650 psi) for

AIBN and 120 °C (3980 psi) for TBPB. After a certain reaction period the vessel was reweighed to make sure that there were no leaks during the experiment. The carbon dioxide was vented into ethanol and the films were recovered. Surface homopolymer was removed from the film by rinsing with THF or toluene until a constant mass or reproducible IR spectrum was obtained.

1.3.2.5 Metal Composite Synthesis

Samples of PPXN or PPXC were placed in a stainless steel reaction vessel with either (l,5-cyclooctadiene)dimethylplatinum (II) [Pt(C0D)Me2] or (1,5- cyclooctadiene)(hexafluoroacetylacetonato)silver(l) [Ag(COD)(hfac)] at a

16 concentration of 0.5 wt% relative to COj. Carbon dioxide was added via a syringe pump at 40 "C and a pressure of 1350 psi. The carbon dioxide was released after 2 hours and the films rinsed in diethyl ether to remove any surface adsorbed precursor.

The films were put into another vessel and hydrogen was added from a custom built

piston at 40^^C and 1350psi. The hydrogen was vented after 1 hour.

1.3.2.6 PPXN/PECA Multilayer Coatings

Multilayer films of poly(p-xylylene) and poly(ethyl-2-cyanoacrylate) were prepared in collaboration with fellow graduate student Zhixiang Lu. PPXN was deposited by VDP. Alkylation with propylene oxide was performed via a literature procedure.'^^ The film was placed in a flask with 5 equivalents of stannic chloride and

20 mL dichloromethane. Propylene oxide, in a 1:1 molar ratio versus stannic chloride,

was added dropwise with stirring over 1 minute. The reaction proceeded for 20 hours at

0 '^C and was terminated with methanol. Ethanol, hydrochloric acid, methanol, and water were used to rinse the samples.

The films were then modified with (N,N-diethyl 3-aminopropyl)- trimethoxysilane to place amine groups on the surface. The reaction took place in a 0.2

M solution of the silane in toluene at 70 °C for 3 days. The films were then rinsed with toluene, ethanol, 1:1 ethanol: water, water, and vacuum dried.

Poly(ethyl-2-cyanoacrylate) was grown from the tertiary amine surface by exposing the sample to ethyl-2-cyanoacrylate in a closed container at 50 °C for 30 minutes. The conditions were adjusted to get the desired thickness.

17 1.3.3 Characterization

Transmission and attenuated total internal reflectance infrared spectra were

recorded using a Bio-Rad FTS 1 75C FTIR spectrometer. A 45" gennanium internal reflection crystal was used to obtain ATR-IR spectra. The penetration depth over the wavenumber range of 4000-700 cm"' was determined to range from 0.17 to 0.99 \im, using a refractive index of 1.661 for PPXN and 4 for the Ge crystal. Wide-angle X-ray diffraction was performed on a Siemens D500 diffractometer. Film thickness was determined using a Rudolph Research AutoEL-II ellipsometer, Dektak"' surface

profilometer, or interference fringes in the IR spectra.*^*^ Blend composition was determined gravimetrically for PPXC composites. Thermal gravimetric analysis was performed on a TA Instruments Thermogravimetric Analyzer 2950 at a heating rate of

10 °C/min. UV-VIS spectra were recorded using a Perkin-Elmer Lambda 2 spectrometer. X-ray photoelectron spectra were obtained on a Physical Electronics

Quantum 2000 Scanning ESCA Microprobe. (A detailed description of XPS can be found in Chapter 2 section 2.3.3.2.)

1.4 Results and Discussion

1 .4. 1 Vapor Deposition of Polv(;;-xvlvlene)

The parylene films were prepared as previously described. Figure 1 .6 shows a schematic diagram of the deposition apparatus. The dimer was placed at one end of the reaction chamber with the cooled substrates on the other. After passing through the

18 tube furnace, the gas phase monomer condensed and polymerized on the samples. The temperature of each section was independently controlled.

CD vacuum

Sublimation Pyrolysis Deposition

Figure 1 .6. Vapor deposition polymerization apparatus.

The films were characterized by a variety of techniques. XPS showed the

expected shake-up peak at -291 eV which is characteristic of the 7i->7i transition of the

aromatic rings. Most films had a small amount of oxygen on the surface which is either the result of oxidation during storage or reaction of atmospheric oxygen with the

benzylic radical chain ends after deposition. PPXC had -11% chlorine which is the expected value for a monochlorinated repeat unit. Wide angle X-ray analysis of PPXN

showed strong reflections at d-spacings of 0.5 1 5 nm and 0.395 nm which correspond

well with the literature values for the (020) and (110) planes of a monoclinic unit cell.'^'^

PPXC had one strong diffraction peak at a d-spacing of 0.392 nm and two weaker peaks

at 0.522 nm and 0.28 1 nm also indicating a monoclinic unit cell. DSC was used to determine the crystallinity, with that of PPXN varying from -45-55% depending on deposition conditions and PPXC slightly lower, -35%.

19 1.4.2 Phase Behavior and Absorption Kinetics

Previous research has shown that both MMA and styrene form homogenous solutions in carbon dioxide under the conditions used in this study. The initiators,

AIBN and TBPB, are also soluble in scC02.'*^"^ A stainless steel vessel with quartz windows was used to confirm the homogeneous phase behavior for a few representative reactions.

The equilibrium mass uptake of carbon dioxide in PPXC was determined by soaking a piece of the polymer in SCCO2 for a certain amount of time, rapidly releasing the pressure, and using a balance to measure the mass loss of CO2 over time. Figure 1.7 shows a typical desorption experiment. Extrapolation of the initial linear portion of the curve back to zero time gives the mass uptake of CO2 at the end of the soak period. The soak period was varied from 30 minutes to 5 hours. The solubility of CO2 in PPXC was determined to be ~3 wt% at 40 °C and 1500 psi and equilibrium was achieved in less than 30 minutes.

A similar procedure was attempted to determine the solubility of CO2 in PPXN samples at 40°C and 1500 psi. However, the carbon dioxide diffused out of the samples so rapidly that by the time they were placed on the balance they had nearly returned to

their initial weight.

Solubility determinations of MMA and styrene in the PPXN and PPXC films were not attempted. As will be discussed, the monomer wets the samples after the

carbon dioxide is vented. The result is a surface adsorbed layer that makes mass uptake experiments more difficult for films than with millimeter thick samples.

20 3.5

3 -

0 10 20 30 40 50 60 70

Time^'^ (sec)^'^

Figure 1.7. A representative desorption profile of CO: from PPXC after a 2 hr soak at 40 °C, 1500 psi. Extrapolation to t=0 gives the mass uptake after the soak period.

1.4.3 PPXN Polymer Composites

Reaction conditions were chosen to maximize the incorporation of the second polymer component. The films were soaked in a scCO: solution of monomer and initiator at a temperature at which the half-life of the initiator was hundreds of hours.

The soak period allowed monomer and initiator to penetrate the substrate. The temperature was then raised to decompose the initiator and polymerize the monomer.

This process was shown to increase the incorporation of the second polymer compared to conditions where the carbon dioxide was vented after the soak period and the sample then heated under nitrogen to initiate polymerization." Polymerization in the presence of the SCCO2 solution allows monomer to partition from the SCCO2 fluid phase into the

21 substrate as it is consumed. This procedure was chosen not only because of the higher

incorporation of the second polymer phase, but also because venting could allow

reagents to desorb from the relatively thin films. A two step procedure was also

impractical because the excess monomer left after the soak period wetted the films,

making handling and transfer difficult.

A control experiment was done to determine if the PPXN films could be

modified without SCCO2. A film sample was placed in neat styrene with 0.3 mol%

AIBN at atmospheric pressure under nitrogen. The sample was allowed to soak for 8

hrs at 40 "C and then 13 hours at 80 ''C. Figure 1 .8 shows the FT-IR spectra of the

sample. As reacted, there is a distinct peak at 701 cm'' corresponding to the aromatic

out of plane bending vibration of polystyrene. However, after a few minutes rinse in

toluene the peak disappears. No composite was formed and only a surface layer of

polystyrene adsorbed to the film. PPXN placed in neat MMA also showed no

modification.

A similar experiment was perfonned, but this time scCOt was added to the

vessel. The peak at 701 cm ' in the FT-IR spectra in Figure 1.9 reveals that polystyrene

is present. The intensity of the polystyrene peak initially decreased when the sample

was first rinsed in toluene, but then remained constant. The surface adsorbed polystyrene was easily rinsed off leaving only polystyrene that was entrapped in the

PPXN substrate. Similar results were found for PMMA. It is also important to note that the FT-IR spectra of the PPXN/PMMA composite shows the disappearance of the

vinyl (-C=CH2) stretching vibration found at 1638 cm ' in the MMA monomer. This

indicates that polymer has formed rather than entrapment of MMA inside PPXN.

22 ) )

^ 1 1 1 1 1 r" 850 800 75iJ 700 650 600 550 VVavenumber (cm-1

Figure 1.8. Infrared spectra of a PPXN film (a) immediately after reaction with neat styrene and AIBN (40 T for 8 hrs / 80 for 13 hrs); (b), (c) after subsequent 1 and 5 min rinses in toluene respectively.

h 1 1 1 1 1 r- 850 800 750 700 650 600 550 VVavenumber (cm-1

Figure 1.9. Infrared spectra of PPXN film (a) after reaction with styrene and AIBN in

SCCO2 (40 T, 4 hrs / 80 °C, 4 hrs); (b), (c) after 5 and 60 min rinses in toluene respectively.

23 It has been shown in previous work that polymerization inside scCO^-Swollen substrates resuhs in a polymer with higher molecular weight compared to polymerization in the SCCO2 phase. '""^ However, some low molecular weight homopolymer can form in the scCO: solution and precipitate onto the outside of the

sample. It is important to remove the surface homopolymer to get an accurate composite composition, especially when dealing with films that have a high surface area to volume ratio.

Parylenes are very solvent-resistant so rinsing the samples in THF or toluene removed only the surface layer and left the bulk composite unchanged. An example

was already shown in Figure 1.9. This is an important point, because the composite composition detennination would be affected if the second polymer phase was easily extracted from the inside of the film while removing the excess homopolymer surface layer. Attenuated total internal reflection IR spectra were also recorded for PPXN/PS composite samples. The ATR-IR experiment probed the outer 200-900 nm of the

composite films. The characteristic polystyrene vibration at 701 cm ' was still apparent even after repeated solvent rinsing, proving that only the surface homopolymer, and

very little of the entrapped polymer, was removed.

Figure 1.10 shows the FT-IR spectra of a PPXN/PMMA composite. The film

was rinsed with THF for 5 minutes after the reaction and an IR spectrum was taken. It was then continuously extracted for a week in THF (soxhlet extraction) and another IR

' spectrum was recorded. The peak at 1 734 cm corresponds to the carbonyl stretching vibration of PMMA. The area under the peak was calculated for each spectrum and there was almost no change. Gravimetric analysis also revealed a negligible change.

24 )

Therefore, it was concluded that a quick solvent rinse was an efficient way to remove the surface adsorbed homopolymer without affecting the composite composition.

.6-

b mm .5&-

0) A c .5-C_J n(0 oi_ (/)

.4-

.3^

2000 1900 1800 1700 1600 1500 Wavenumber(cm-I)

Figure 1.10. Infrared spectra showing a PPXN/PMMA composite after a 5 min rinse in THF and after one week of THE soxhlet extraction.

2.5H

1020 PPXN O E (0 .o 1.5- ol_ (/)

< PPXN/PMMA 1 A"

PPXN/PS

3500 3000 2500 2000 1500 1000 500 Wavenumber (cm-1

Figure 1.11. IR spectra of PPXN homopolymer and composites with PMMA and PS. Arrows indicate characteristic peaks that were used to determine composite composition.

25 The free standing PPXN films were thin (2-5|j.m) and often tore during

composite preparation. As a result, composition by mass uptake was usually unreliable.

Instead, IR was used to qualitatively prove blend synthesis and semi-quantitatively

determine composition. Figure 1.11 shows the FT-IR spectra of PPXN and its

' composites with PS and PMMA. The carbonyl stretching vibration at 1734 cm

identifies the PMMA in the PPXN/PMMA blend while the peak at 701 cm"' is

characteristic of PS in the PPXN/PS blend. Relative composition was determined by

taking a ratio of the area of the characteristic peak of PMMA or PS and the

characteristic peak area for PPXN at 1020 cm'. This allowed comparison of different

reaction conditions while removing the effects of thickness variations from sample to

sample.

Quantitative determination of the composition requires a calibration curve to

compare the IR peak areas of reference samples with known composite composition.

Typically the references would be prepared by blending known amounts of the two

components. However, with parylenes blending is not possible because they do not

dissolve at reasonable conditions nor is melt processing possible. In fact, the purpose of

this work was to make the first polymer composites of poly(p-xylylene).

An attempt was made to prepare a calibration curve by spin coating PS or

PMMA onto PPXN. The thickness was measured by ellipsometry and profilometry which, along with the density, was used to calculate the weight percent. FT-IR spectra were taken of each standard and the ratio of the area of each characteristic peak was determined. The resulting calibration curve had a lot of scatter in the data and was not useful. The measured FT-IR peak intensities were likely affected by slight interferences

26 and scattering that can occur at the interface between the two components in the multilayer film. Another attempt at a calibration curve was made by cryogrinding the

PPXN film to make a physical blend with PS or PMMA. This proved difficult because the ground PPXN film was statically charged and stuck to everything making accurate weighing troublesome.

The sinusoidal baselines in the spectra in Figure 1.11 are an interference pattern

resulting from multiple internal reflections of the IR radiation in the thin films.*^*^

Although aesthetically displeasing, the wave pattern does not affect the peak intensities.

It is actually useful because the thickness of the films can be easily calculated using

Equafion 1.1:

2n^v (II)

where d is the film thickness. Am is the number of maxima minus 1, Av is the

wavenumber difference between the first and last maxima, and n is the refractive index of the film.

Initial experiments were done to determine the time needed for methyl methacrylate and styrene to reach an equilibrium concentration inside the PPXN films.

This required varying the soak period during which the monomer and initiator penetrate

the substrate at a temperature that results in very little initiator decomposition. Figure

1.12 shows an example for a PPXN/PMMA composite (5 wt% MMA, 0.3mol% AIBN).

The soak period was varied from zero to 4 hours at 40 °C while the polymerization reaction period was kept constant at 4 hr and 80 °C. There was no significant difference

27 in the amount of PMMA in the composite. The styrene system showed a similar result.

This suggests the reagents diffuse quickly into the PPXN substrate, which is reasonable given the high surface area to volume ratio of the films. All further reactions were done

at a soak time of four hours.

A longer reaction period increased the composite composition. Figure 1.13 shows the relative PMMA concentration as a function of reaction time for a

PPXN/PMMA composite prepared by soaking in SCCO2 for 4 hr at 40 ''C, with 5 wt%

MMA and 0.3 mol% AIBN. The PMMA concentration increased linearly with time.

As the reaction proceeds, monomer can partition from the SCCO2 phase into the

substrate ensuring that there is monomer at the site of the growing polymer chains. At

much longer reaction times the curve reaches a plateau as all the monomer has polymerized, either inside or outside of the substrate. PPXN/PS composites showed the same trend of a linear increase in composite composition with time.

The effect of the initial monomer concentration in scCOi on composite composition was determined by varying the amount from -5-20 wt%. (Higher monomer concentrations were attempted, but resulted in the films being irretrievably encased in homopolymer.) Figures 1.14 and 1.15 show results for PMMA and PS composites prepared by soaking PPXN films in SCCO2 with 0.3 mol% AIBN and varying amount of monomer for 4hr at 40"C followed by reaction for 4 hr at 80 "C. The concentration of both PMMA and PS increased in the respective composites as the concentration of monomer increased.

28 3.0-1

3 2.5 c

2.0 c o u c 1.5 o u

1.0-

> 0.5-

0^ 0.0

Soak time (hr)

Figure 1.12. Relative PMMA concentration in a PPXN/PMMA composite as a function of soak time (5 wt% MMA, 0.3 mol% AIBN).

3.0 -| ^ 2.8- 3 2.6- ^ 2.4-1 c o 2.2-

(0 2.0- c 1.8- u0) 1.6- c o 1.4- u 1.2- 1.0- 0.8- 0.6- >0) 0.4- ss. 0.2

0.0

Reaction time (hr)

Figure 1.13. Relative PMMA concentration in a PPXN/PMMA composite as a function of reaction time (4 hr / 40 °C soak, 80 °C reaction, 0.3 mol% AIBN).

29 1

1.6-,

^ 1.4- 3 1.2- c o 1.0- re c o0) 0.8 c o u 0.6-1 0) Q. o 0.4 >

0.2 DC

<- r- r- 0.0 1 — — — 2 10 12 14

Styrene concentration In CO (\Art%)

Figure 1.14. Relative polystyrene concentration in a PPXN/PS composite as a function of initial styrene concentration (4 hr / 40 °C soak, 4 hr / 80 °C reaction, 0.3 mol% AIBN initiator).

5.0-

4.5-

c 4.0- o

re 3.5- c o 3.0- u c o 2.5- u 1 < 2.0-

1.5- Ql 1.0- ive!

re 0.5-

MIVIA concentration in CO^ (wt%)

Figure 1.15. Relative PMMA concentration in a PPXN/PMMA composite as a function of initial MMA concentration (4 hr / 40 T soak, 4 hr / 80 "C reaction, 0.3 mol% AIBN initiator).

30 '

40

^ 35

C .2 25-

£ 20- u i 15 I 10-1

Q- 5

0 r- I— — — I 10 12 14 16 18

Reaction time (hr)

Figure 1.16. PMMA concentration in a PPXC/PMMA composite as a function of reaction time (4 hr / 80 °C soak, 5 wt% MMA, 0.3 mol% TBPB).

70 n 8. TBPB O AIBN 6 60- 4

50- (Wt' 2

ition 0 40- 0 4 8 12 16 20

*-' § 30 c o ^20

a. 10-1

0 —1—'—I——1— —1—'—I—'—r- —1— 10 12 14 16 18 20 22

MMA concentration in CO (wt%)

Figure 1.17. Comparison of PMMA concentration in a PPXC/PMMA composite as a fiinction of initial MMA concentration using two different initiators, AIBN and TBPB (4 hr soak, 4 hr reaction, 0.3 mol% initiator). Inset shows AIBN data in more detail.

31 1.4.4 Polv(chloro-p-xvlvlene) Composites

PPXC/PMMA composites were prepared using AIBN and TBPB initiators in a

manner similar to the PPXN composites. The PPXC films were thicker and more

robust than the PPXN, which allowed absolute composite composition to be determined

gravimetrically. All composite samples were rinsed with THF to remove surface-

adsorbed homopolymer until a constant mass was obtained.

Composite composition was controlled by varying the reaction time. Figure

1.16 shows the relative PMMA concentration as a flinction of reaction time for a

PPXC/PMMA composite prepared by soaking in scCOi for 4 hr at 80 "C with 5 wt%

MMA and 0.3 mol% AIBN followed by reaction at 120 °C. The PMMA concentration

increased linearly with time and could be varied from 5-40 wt% while keepmg the

initial monomer concentration constant. The slope of the curve decreased after ~16 hours as the monomer was exhausted.

The free radical initiator had a dramatic impact on the PMMA concentration in the PPXC composite film. The two curves in Figures 1.17 show the PMMA weight percent in the composite as a function of the initial monomer weight percent in CO2 with AIBN or TBPB as the initiator. As was observed for the PPXN composites, an increase in the initial monomer concentration produced an increase in the PMMA concentration. However, the TBPB initiator produced composites with a much higher

loading of PMMA. For example, at an initial monomer concentration of 20 wt% MMA the TBPB initiated composite had a PMMA concentration of -65 wt% while the AIBN initiated composite only had -6.5 wt%. Similar results were seen when PPXC/PS composites were prepared. The amount of PS in the composite increased from 3.5 to 27

32 wt% when the initiator was changed from AIBN to TBPB at the same initial monomer

concentration (5 wt%). A reason for the disparity may be that the initiators differ in

their solubility in SCCO2 and/or the substrate. AIBN may be more soluble in the carbon

dioxide phase and not partition very easily into the PPXN film. The TBPB

polymerizations were also run at a higher temperature, 120 °C versus 80 °C for AIBN.

This difference in temperature may have increased the solubility of TBPB in the PPXC

film.

1.4.5 Polymer Composites on Silicon Wafers

Composites were also prepared using parylene substrates coated onto silicon

wafers to show that this SCCO2 procedure is applicable to the films' main use as a

coating material. The parylene films did not always adhere well to the substrates so the

silicon wafers were modified with silanes containing unsaturated functional groups.

The parylene monomer can then possibly form a covalent bond with the surface via a

free radical reaction during deposition. The parylene films deposited onto the modified substrates rarely debonded.

The reaction conditions used to prepare the composites were similar to those used for the free standing films. Figure 1.18 shows the transmission FT-IR spectra

PPXN/PMMA and PPXC/PMMA composites prepared under similar reaction conditions: 4 hr soak at 40 ''C with 0.3 mol% AIBN and 10 wt% MMA followed by 4 hr reaction at 80 °C. The spectra were obtained through the silicon wafer. The carbonyl

vibration at 1734 cm ' identifies the PMMA in the composite. The composite films

33 )

adhered well to the silicon substrates. Debonding only occurred occasionally when the

CO2 was rapidly vented after the reaction.

T r 2OUIJ 1500

vVave n u rn b e r ( c m- 1

Figure 1.18. Transmission IR spectra of PPXN/PMMA and PPXC/PMMA composites

on silicon wafers. The carbonyl stretching vibration at 1734 cm ' is highlighted.

1.4.6 Composite Characterization

1.4.6.1 Thermal Properties

A PPXC/PMMA composite with ~30wt% PMMA was characterized by themial gravimetric analysis. Figure 1.19 gives the weight loss as a function of temperature for the composite and both homopolymers. The PMMA homopolymer decomposed at

~350''C while PPXC was thennally stable to ~-5()0"C. Two plateaus were observed for the composite corresponding to the decomposition of each of the homopolymers. The onset of PMMA decomposition in the composite appeared to be inhibited by about 50"C

34 compared to the homopolymer. It is likely that the PPXC film was simply slowing the diffusion of the PMMA decomposition products so the weight loss was not observed immediately as was the case with the homopolymer. A slower heating rate would probably alleviate this artifact. The TGA analysis provided further proof of composite formation rather than simply entrapment of MMA.

120

100+- —

80

\PMMA PPXC/PMMA . PPXC

60

^ 40

20

0

-20 200 400 600

Temperature (°C)

Figure 1.19. Thermal gravimetric analysis of PMMA, PPXC, and a PPXC composite film containing 30 wt% PMMA.

1.4.6.2 X-ray Diffraction

PPXN and PPXC are both semi-crystalline polymers. PPXN has two main

crystalline morphologies. The (3 phase is formed at low deposition temperatures (-78

°C) and the a phase is more stable at higher deposition temperatures (-1 7 "C to 26

°C)."^^''°^ Interestingly, a transition between the monoclinic a phase and the hexagonal

P phase can be induced by post deposition annealing above 220 °C.

35 —

The unit cell parameters of crystalline PPXN have been reported as follows:

a=0.59 nm, Z)=1.06 nm, c( fiber axis)=0.655 nm, p=135".'^'' The a/?-projection is face-

centered rectangular and as the aromatic ring is substituted only the Zj-value increases.

The value of b for PPXC is 1 .28 nm. It has been reported by Murthy and Kim that the

99 values of a and |3 remain invariant

10000- • unprocessed PPXC CO^ annealed (80°C/120°C) 9000- 110 A PPXC/PMMA composite 8000- A baseline 3 7000- 1

*-> 6000- '55 c 5000- (D 4000- 0) 020 200 3000-

TO 2000-

1000-

0- d=0.52nm :d=0.39nm :d=0.28nm

-< ' 1 1 r — 10 15 20 25 30 35 40

2e (degrees)

Figure 1 .20. Wide angle X-ray diffraction data for PPXC and a composite containing 40 wt% PMMA.

The wide angle X-ray diffraction pattern of the PPXC and the PPXC/PMMA

composite are shown in Figure 1 .20. The PPXC film as received, with no SCCO2 treatment, has one strong diffraction peak at a d-spacing of 0.392 nm and two weaker peaks at 0.522 nm and 0.281 nm. The weaker peaks' relative intensities increase upon

exposure to SCCO2, (4 hrs / 80 "C, 4 hrs / 120 °C), consistent with improved crystallinity.

36 Supercritical carbon dioxide has been shown to plasticize polymers, i.e., increase the

'^^"'"'^ chain mobility allowing enhancement of crystallinity as a fimction of annealing.

Introduction of -40 wt% PMMA into PPXC under similar scCO^ conditions does not

show a change in the crystallinity of the PPXC polymer as evidenced by the X-ray data.

However, it would appear that the addition of PMMA disrupts the annealing process

slightly as seen in the shift in d-spacing, 0.392 nm to 0.395 nm, and slight reduction in

relative intensity of the 110 diffraction peak. This evidence suggests that scCO^

transports methyl methacrylate mainly into the amorphous regions of the PPXC leaving the crystalline regions unmodified.

The three diffraction peaks observed for PPXC, d=0.522 nm, 0.392 nm, and

0.281 nm, are similar to the characteristic 020. 1 10, and 200 peaks of PPXN, d=0.530 nm, 0.391 nm, and 0.210 nm respectively.*^'^ In the diffraction pattern of the composite

(see Figure 1 .20) the relative ratio of the d-spacings is similar but measurably different suggesting that the shear angle of 135° may not be valid in our case.

1.4.7 Parylene/Metal Composites

Platinum and silver particles were deposited into both types of parylene film using a two step reaction procedure. A piece of film was soaked in a SCCO2 solution containing either [Pt(C0D)Me2] or [Ag(COD)(hfac)] (Figure 1.21). The CO2 was vented after the desired reaction time and the films placed in a second vessel with hydrogen gas. The hydrogen reduced the precursor, freeing the ligands and leaving behind silver or platinum. The SCCO2 was vented in order to limit the amount of metal deposited on the surface. A quick rinse with diethyl ether before reduction removed

37 most of the surface-adsorbed precursor. Unlike the surface-adsorbed polymer that was

easily removed from the polymer/polymer composites with a solvent rinse, platinum or

silver are much more difficult to dissolve without affecting the substrate.

The PPXN and PPXC films are clear and colorless before the reaction and

remain so after infusion of the precursors. Figure 1 .22 shows transmission FT-IR spectra of PPXN, [Ag(COD)(hfac)], and a composite of the two. The peaks at 738, 661,

and 577 cm ' in the composite spectrum are characteristic peaks for [Ag(COD)(hfac)] and reveal the presence of the unreduced precursor inside the film. The silver and platinum composite films become orange and gray respectively upon reduction. A broad absorbance peak at -450 nm in the UV-VIS spectrum was observed for both parylene/silver composites corresponding to the orange color of the films.

CF

Pt + 2 CH4 +

Figure 1.21. Hydrogen reduction of [Ag(COD)(hfac)] and [Pt(C0D)Me2] to deposit the corresponding metal.

X-ray diffraction gave no information about the platinum or silver clusters which was surprising given that both Watkins^*^ and Brown''^ were able to see diffraction peaks from the metals. The concentration of particles in the thin films may be too low to give a detectable diffraction pattern. Another explanation could be that the size of the particles was too small to be observed. The metal precursor may be

38 )

immobile inside the polymer substrate after the SCCO2 has been released. As the

precursor is reduced with hydrogen, the other molecules cannot diffuse to the site to increase the size of the particles.

X-ray photoelectron spectroscopy revealed the presence of platinum and silver in the parylene/metal composites. The atomic concentration of metal ranged from 0.5-

1%. Sputter depth profiling was done to determine if the metal was only on the surface or throughout the bulk of the film. In this process, argon ions were used to sputter material from the surface, exposing the interior of the sample prior to each successive spectral acquisition. The profiles revealed a slightly higher surface concentration of metal, however after five sputter cycles the concentration stopped decreasing and remained constant.

1600 1400 1200 1000 800 600 vVavenurnber (cm-1

Figure 1.22. Transmission FT-IR spectra of (a) PPXN, (b) [Ag(COD)(hfac)] and (c) [Ag(COD)(hfac)] in PPXN. Arrows indicate characteristic peaks for the silver precursor in the composite.

39 1.4.8 PPXN/PECA Multilayer Templates

Poly(ethyl-2-cyanoacrylate) (PECA) is widely used as an adhesive and is more

commonly known as super glue. The polymer can be formed by anionic polymerization

using even a weak nucleophile like water. Vapor phase deposition of the monomer onto

a surface modified with an anionic initiator permits the formation of a confomial

coating.

Our research group has expertise in the controlled growth and modification of

' conformal coatings of PECA. ' A collaboration was setup to prepare a multilayer structure of PECA and PPXN to see if one of the layers could be selectively modified.

PPXN was deposited and modified with a tertiary amine as described above. PECA was then grown from the surface and another layer of PPXN was added. The process could be repeated as many times as desired, but only a sandwich structure was prepared

initially.

50 nm ppxn

Figure 1.23. TEM image showing the preferential deposition of platinum in the poly(ethyl-2-cyanoacrylatc) layer surrounded by poly(/;-xylylenc).""

40 Platinum metal was deposited into the PPXN/PECA multilayer film using

supercritical carbon dioxide in a procedure similar to the one previously described for parylene/metal composites. However, a one step reaction was used, with the platinum precursor being both infused into the polymer sample and reduced with hydrogen in the

presence of scCO^. TEM of the sample. Figure 1 .23, revealed that the platinum particles were selectively deposited in the PECA regions and not PPXN. The exact reason has not been established. The platinum precursor may interact with the functional groups on PECA, with ligand exchange being possible. The platinum

precursor may also be more soluble in the PECA so the concentration is enhanced in

that phase. Research on such multilayer samples is continuing in our group and will be the subject of further investigations.

1.5 Conclusions

The first polymer composites of parylenes were prepared by radically polymerizing vinyl monomers inside scC02-swollen poly(p-xylylene) and poly(chloro- p-xylylene) substrates. The weight percentage of the second polymer phase could be increased with longer reaction times or higher initial monomer concentration. Initiator selection also proved important. Higher concentration composites could be prepared when tert-buty\ peroxybenzoate was used as the initiator compared to AIBN initiated samples under similar reaction conditions. Characterization by infrared spectroscopy, wide angle X-ray, and thermal gravimetric analysis confirmed composite preparation.

Platinum and silver particles were also deposited in the films using scCO: soluble precursors. Characterization of the composite samples provided mixed results.

41 Visual observation revealed a change in the color, but SEM, TEM, and WAXD failed to

show any particles. X-ray photoelectron spectroscopy established the presence of metal

in the composite films. The mixed results suggest that either the particles were too

small and at too low a concentration to be observed or the observed metal was only

confined to the surface of the film and did not penetrate the interior.

Finally, multilayer templates of PPXN and PECA were prepared. Both are

confomial polymer coatings that can be grown with precise thickness control. Platinum

metal was selectively deposited in the PECA regions and its presence was confirmed by

TEM.

42 1.6 References

( 1 ) Watkins, J. J.; McCarthy, T. J. Macromolecules 1994, 27, 4845-4847.

(2) Watkins, J. J.; McCarthy, T. J. Macromolecules 1995, 28, 4067-4074.

(3) Kung, E.; Lesser, A. J.; McCarthy, T. J. Abstracts ofPapers ofthe American

Chemical Society 1999, 218, U4 1 7-U4 1 7.

(4) Cao, C. T.; McCarthy, T. J. Abstracts ofPapers of the American Chemical 5oc/e/v 2000, 279, U440-U440.

(5) Beach, W.F.; Lee, C; Bassett, D. R.; Austin, T. M.; Olson, R. In Encyclopedia of Polymer Science and Engineering, 2nd ed.; Wiley: New York, 1989; Vol. 17, pp 990-1025.

(6) Kozlov, M.; Quannyne, M.; Chen, W.; McCarthy, T. J. Macromolecules 2003, 36, 6054-6059.

(7) Kozlov, M.; McCarthy, T. J. Langmuir 2004, 20, 9 1 70-9 1 76.

(8) Chen, W.; McCarthy, T. J. Macromolecules 1997, 30, 78-86.

(9) Hsieh, M. C; Fams, R. J.; McCarthy, T. J. Macromolecules 1997, 30, 8453- 8458.

(10) Phuvanartnuruks, V.; McCarthy, T. J. Macromolecules 1998, 31, 1906-1914.

(11) Lu, Z. X.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society^ 2004, 227, U379-U379.

( 1 2) Cooper, A. I. Journal ofMaterials Chemistry 2000, 1 0, 207-234.

(13) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction: Principles and Practice, 2nd ed.; Butterworth-Heinemann: Boston, 1994.

43 (14) Perrut, M. Industrial & Engineering Chemistry Research 2000, 59, 453 1-4535.

(15) Wells, S. L.; DeSimone, J. Angewandte Chemie-International Edition 2001, 40, 519-527.

(16) O'Neil, A.; Watkins, J. J. Green Chemistiy 2004, 6, 363-368.

1 95-2 ( 1 7) Andrade, C. K. Z.; Alves, L. M. Current Organic Chemistiy 2005, 9, 18.

(18) Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. Mo/ecw/ar Thermodynamics ofFhdd-Phase Equilibria, 2nd ed.; Prentice-Hall: New Jersey, 1986.

( 1 9) Hyatt, J. A. Journal of Organic Chemistiy 1984, 49, 5097-5 101.

(20) Krukonis, V. J. Polymer News 1985, 11, 7-16.

(21) DeSimone, J. M.; Guan, Z.; Elsbemd, C. S. Science 1992, 257, 945-947.

(22) Yilgor, I.; McGrath, J. E. Polymer Bulletin 1984, 12, 49 1 -497.

(23) Yilgor, I.; McGrath, J. E. Polymer Bulletin 1984, 12, 499-506.

(24) Kendall, J. L.; Canelas, D. A.; Young, J. L.; DeSimone, J. M. Chemical Reviews 1999, 99, 543-563.

(25) Kirby, C. F.; McHugh, M. A. Chemical Reviews 1999, 99, 565-602.

(26) Young, J. L.; DeSimone, J. M. Pure and Applied Chemistiy 2000, 72, 1357- 1363.

(27) Tomasko, D. L.; Li, H. B.; Liu, D. H.; Han, X. M.; Wingert, M. J.; Lee, L. J.; Koelling, K. W. Industrial & Engineering Chemistry Research 2003, 42, 643 1- 6456.

44 (28) Woods, H. M.; Silva, M.; Nouvel, C; Shakesheff, K. M.; Howdle, S. M. Journal ofMaterials Chem is tiy 2004, 14, 1663-1678.

(29) Wang, W. C. V.; Kramer, E. J.; Sachse, W. H. Journal ofPolymer Science Part

B-Polymer Physics 1982, 20, 1 37 1 - 1 384.

(30) Chiou, J. S.; Barlow, J. W.; Paul, D. R. Journal ofApplied Polymer Science 1985,30,2633-2642.

(31) Wissinger, R. G.; Paulaitis, M. E. Journal ofPolymer Science Part B-Polymer Physics 1987, 25, 2497-2510.

(32) Berens, A. R.; Huvard, G. S.; Korsmeyer, R. W.; Kunig, F. W. Journal of Applied Polymer Science 1992, 46, 231-242.

(33) Berens, A. R.; Huvard, G. S.; Korsmeyer, R. W.: United States Patent 4,820,752, 1989.

(34) Sand, M. L.: United States Patent 4,598,006, 1986.

(35) Kazarian, S. G.; Brantley, N. H.; Eckert, C. A. Chemtech 1999, 29, 36-41.

(36) Watkins, J. J.; Ph.D. Dissertation, University of Massachusetts, 1997.

(37) Kung, E.; Lesser, A. J.; McCarthy, T. J. Macromolecules 2000, 33, 8192-8199.

(38) Cao, C. T.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society 2001, 22 f U362-U362.

(39) Watkins, J. J.; McCarthy, T. J. Chemistry ofMaterials 1995, 7, 1 991 -&.

(40) Brown, G. D.; Watkins, J. J. Abstracts ofPapers of the American Chemical Society 2001, 227, U369-U369.

(41) Errede, L. A.; Szwarc, M. Quarterly Reviews 1958, 12, 301-320.

45 (42) Corley, R. S.; Haas, H. C; Kane, M. W.; Livingston, D. I. Journal ofPolymer Science\9S4J3A^'l-\56.

(43) Beach, W. F. In Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed.;

Kroschwitz, J. I.; Howe-Grant, M., Eds.; Wiley: New York, 1998; Vol. Supplement, pp 863-901.

(44) Greiner, A.; Mang, S.; Schaefer, O.; Simon, P. Acta Polymerica 1997, 48, 1-15.

(45) Itoh, T. Prog. Polym. Sci. 2001, 2(5, 1019-1 059.

(46) Szwarc, M. Nature 1947, 160, 403.

(47) Szwarc, M. Discussions of the Faraday Society 1947, 2, 46.

(48) Szwarc, M. Journal ofPolymer Science 1951, 6, 3 1 9-329.

(49) Gorham, W. F. Journal ofPolymer Science-Polymer Chemistiy Part A-1 1966, 4, 3027-3039.

(50) Gorham, W. F. Advances in Chemistry Series 1969, 643-&.

(5 1 ) Beach, W. F. Macromolecules 1978, 77, 72-76.

(52) Gaynor, J. F.; Desu, S. B.; Senkevich, J. J. Macromolecules 1995, 28, 7343- 7348.

(53) Fortin, J. B.; Lu, T. M. Chemistry ofMaterials 2002, 14, 1945-1949.

(54) Rogojevic, S.; Moore, J. A.; Gill, W. N. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1999, 77, 266-274.

Broer, J.; 5- (55) D. Luijks, W. Journal ofApplied Polymer Science 1981 , 26, 24 1 2422.

(56) Tolstopyatov, E. M. Journal ofPhysics D: Applied Physics 2002, 35, 1516-1 525.

46 (57) Kubo, M.; Iwatsuki, S. In Polymeric Materials Encyclopedia; Salamone, J. C, Ed.; CRC Press: Boca Raton, 1996.

(58) Greiner, A.; Mang, S.; Schafer, O.; Simon, P. Acta Polymerica 1997, 1-15.

(59) Chow, S. W.; Loeb, W. E.; White, C. E. Journal ofApplied Polymer Science 1969, 75, 2325-2332.

(60) Lahann, J.; Langer, R. Macromolecules 2002, 35, 4380-4386.

(61) Kubo, M.; Toyota, K.; Itoh, T.; Iwatsuki, S. Polym. Bull. 1995, 35, 615-619.

(62) Auspos, L. A.; Hall, L. A. R.; Hubbard, J. K.; Kirk, W. M.; Schaefgen, J. R.; Speck, S. ^. Journal ofPolymer Science 1955, 75, 9-17.

(63) Auspos, L. A.; Bumam, C. W.; Hall, L. A. R.; Hubbard, J. K.; Kirk, W. M.; Schaefgen, J. R.; Speck, S. B. Journal ofPolymer Science 1955, 75, 19-29.

(64) Lahann, J.; Klee, D.; Hocker, H. Macromol. Rapid Commun. 1998, 79, 441-444.

(65) Lahann, J.; Balcells, M.; Rodon, T.; Lee, J.; Choi, I. S.; Jensen, K. F.; Langer, R. Langmuir 2002, 18, 3632-3638.

(66) Greiner, A.; Simon, P.: United States Patent 5,3 10,858, 1994.

(67) Simon, P.; Gremer, A. Polymer Journal 1992, 24, 1317-1320.

(68) Simon, P.; Mang, S.; Hasenhindl, A.; Gronski, W.; Greiner, A. Macromolecules 1998, 37, 8775-8780.

(69) Sochilin, V. A.; Mailyan, K. A.; Alexandrova, L. N.; Nikolayev, A. A.;

Pebalk, A. V.; Kardash, I. Y. Doklady Akademii Nauk Sssr 1991, 379, 173-176.

(70) Gaynor, J. F.; Desu, S. B. J. Mater. Res. 1994, 9, 3 125-3 130.

(71) Gaynor, J. F.; Desu, S. B. J. Mater. Res. 1996, 77, 236-242.

47 (72) Beecroft, L. L.; Ober, C. K. Chemistry ofMaterials 1997, 9, 1302-1317.

(73) Shenhar, R.; Norsten, T. B.; Rotello, V. M. Advanced Materials 2005, 1 7, 657- 669.

(74) El-Sayed, M. A. Accounts of Chemical Research 2001, 34, 257-264.

(75) Godovski, D. Y. In Thermal and Electrical Conductivity ofPolymer Materials;

Springer-Verlag: Berlin, 1995; Vol. 1 19, pp 79-122.

(76) Godovsky, D. Y. In Biopolymers/Pva Hydrogels/Anionic Polymerisation Nanocomposites; Springer-Verlag: Berlin, 2000; Vol. 153, pp 163-205.

(77) Senkevich, J. J.; Desu, S. B. Chemistry ofMaterials 1999, 77, 1814-1821.

(78) Sergeev, G.; Zagorsky, V.; Petrukhina, M. Journal ofMaterials Chemistry 1995, 5,31-34.

(79) Gerasimov, G. N.; Sochilin, V. A.; Chvalun, S. N.; Volkova, L. V.; Kardash, I. Y. Macromolecular Chemistiy and Physics 1996, 797, 1387-1393.

(80) Gerasimov, G. N.; Nikolaeva, E. V.; Smimova, E. I.; Sochilin, V. A.;

Trakhtenberg, L. I. Doklady Physical Chemistiy 2001, 580, 223-226.

(81) Gerasimov, G. N.; Popova, E. L.; Nikolaeva, E. V.; Chvalun, S. N.; Grigoriev,

E. I.; Trakhtenberg, L. I.; Rozenberg, V. I.; Hopf, H. Macromolecular Chemistry and Physics 1998, 799, 2179-2184.

(82) Grigor'ev, E. I.; Zav'yalov, S. A.; Chvalun, S. N. Vysokomolekularnye

- Soedineniya. Ser.A Ser.B Ser. C Kratkie Soobshcheniya 2003, 45, 1 308- 1313.

(83) Zagorskii, V. V.; Ivashko, S. V.; Bochenkov, V. E.; Sergeev, G. B. Nanostructured Materials 1999, 12, 863-866.

(84) Sergeev, G.; Zagorsky, V.; Petrukhina, M.; Zavyalov, S.; Grigorcv, E.; Trakhtenberg, L. Analytical Communications 1997, 34, 113-114.

48 (85) Bochenkov, V. E.; Stephan, N.; Brehmer, L.; Zagorskii, V. V.; Sergeev, G. B. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2002, / 98- 200,911-915.

(86 Chan, C. M.; Ko, T. M.; Hiraoka, H. Surf. Sci. Rep. 1996, 24, 3-54.

(87 Grace, J. M.; Gerenser, L. J. J. Dispersion Sci. Teclmol. 2003, 24, 305-341.

(88 Lahann, J.; Klee, D.; Thelen, H.; Bienert, H.; Vorwerk, D.; Hocker, H. J. Mater. Sci.-Mater. Med 1999, 70, 443-448.

(89 Nowlin, T. E.; Smith, D. F., Jr. Journal of Applied Polymer Science 1980, 25, 1619-1632.

(90 Yu, Q. S.; Deffeyes, J.; Yasuda, H. Prog. Org. Coat. 2001, 41, 247-253.

(91 Martini, D.; Shepherd, K.; Sutcliffe, R.; Kelber, J.; Edwards, H.; San Martin, K. Applied Surface Science 1999, 14 f 89-100.

(92 Senkevich, J. J.; Yang, G. R.; Lu, T. M. Colloid Surf A-Physicochem. Eng. Asp. 2003, 216, 167-173.

(93 Senkevich, J. J.; Yang, G. R.; Lu, T. M. Colloid Surf. A-Physicochem. Eng. Asp. 2003,274,119-126.

(94 Kaufman, M. H.; Mark, H. F.; Mesrobian, R. B. Journal ofPolymer Science 1954, 13, 3-20.

(95 Sochilm, V. A.; Pebalk, A. V.; Semenov, V. 1.: Kardash, I. Y. Vysokomolekulyarnye Soedineniya Seriya A 1991, 33, 1536-1542.

(96 Herrera-Alonso, M.; McCarthy, T. J. Langmuir 2004, 20, 9184-9189.

(97 Stewart, D. D.; Union Carbide Corporation: United States Patent 3,600,216, 1971.

(98 Larena, A.; Pmto, G. Spectr. Lett. 1992, 25, 447-461.

49 (99) Murthy, N. S.; Kim, H. G. Polymer 1984, 25, 1093-1096.

(100) Suppes, G. J.; McHugh, M. A. Journal of Chemical and Engineering Data 1989, 3^,310-312.

(101) Lora, M.; McHugh, M. A. Fluid Phase Equilibria 1999, 757, 285-297.

(102) Kung, E.; Lesser, A. J.; McCarthy, T. J. Macromolecules 1998, 37, 4160-4169.

(103) Brown, C. J.; Farthmg, A. C. Journal of the Chemical Society 1953, 3270-&.

(104) Kubo, S.; Wunderlich, B., 1972; Vol. 10, pp 1949-1966.

(105) Isoda, S.; Tsuji, M.; Ohara, M.; Kawaguchi, A.; Katayama, K.-i., 1983; Vol. 24, pp 1155-1161.

(106) Iwamoto, R.; Wunderlich, B. J. Polym. Sci. Pt. B-Polym. Phys. 1973, 77, 2403- 2411.

(107) Chiou, J. S.; Barlow, J. W.; Paul, D. R. Journal ofApplied Polymer Science 1985, 30, 3911-3924.

1- (108) Beckman, E.; Porter, R. S. J. Polym. Sci. 1987, Part B: Polvm. Phys. 25, 15 1 1517.

(109) Takada, M.; Tanigaki, M.; Ohshima, M. Polymer Engineering and Science 2001,^7,1938-1946.

(110) TEM image courtesy of Zhixiang Lu.

50 CHAPTER 2

MODIFICATION OF ATOMIC FORCE MICROSCOPE PROBES FOR IMPROVED SURFACE CHEMICAL IDENTIFICATION

2. 1 Introduction

Since its discovery in the mid 1980s, atomic force microscopy has

revolutionized the characterization of surfaces.' AFM allows facile detennination of

sample topography in the nanometer range and under controlled conditions atomic

resolution is possible. It is a nondestructive technique requiring minimal sample preparation and can be carried out at ambient conditions. Complimentary techniques,

such as SEM, require the sample to have a conductive coating and must be done under vacuum.

As will be discussed, chemical infonnation about the samples can also be obtained by AFM. The nanometer spatial chemical resolution of AFM is unmatched by any other technique. Secondary ion mass spectroscopy, for example, has a lateral

resolution that is hundreds of nanometers at best, is destructive to the sample, and must

be done under vacuum. However, AFM suffers from a lack of chemical specificity. It is often easy to image chemical differences, but not unequivocally identify the chemical functionality. The goal of this project was to improve AFM chemical specificity by surface modification of AFM probes with monofiinctional silanes.

51 2.2 Background

2.2.1 Atomic Force Microscope

A schematic diagram of a typical atomic force microscope is shown in Figure

2.1. A cantilever with a sharp tip is moved across a surface by applying a voltage to the piezoelectric scanner. A laser is reflected off the end of the cantilever to a four quadrant photodetector and the signal is used to monitor the tip position. The tip can be raised or lowered in the Z direction to follow the sample topography by applying another voltage to the piezoelectric scanner.

Sample

Figure 2.1. Schematic diagram of an atomic force microscope.

2.2. 1 . 1 Contact Mode AFM

In contact mode, the cantilever is brought into direct contact with the sample surface. The cantilever bends to a certain predetermined deflection that the machine

tries to keep constant. As the tip is scanned across the surface parallel to the cantilever axis, the computer monitors the deflection via the laser position on the photodetector. If

52 the deflection changes, i.e., the tip reaches a peak or valley on the surface, a voltage is immediately applied to the piezoelectric scanner to cause a change in the Z height that returns the deflection back to the setpoint value. The computer then converts the applied voltage to a height change based on an internal calibration curve. In this way a

topographic image of the sample is obtained.

The probe can also be scanned perpendicular to the cantilever axis. In lateral force mode (LFM), friction between the tip and the sample causes a lateral deflection or

torsion of the cantilever. The image that is obtained shows areas where the friction is different. If the detector sensitivity and cantilever torsion force constant are known, the

friction force between the surface and the tip can be determined. The friction force is related to the tip-sample interactions and can be changed by chemical modification of

the sample and tip.

Adhesive forces between the tip and sample can be measured using a force

curve as shown in Figure 2.2. The probe is not moved in the xy plane, but simply

brought down into contact with the surface and retracted. The cantilever deflection is

monitored versus the tip-sample separation distance. There is no deflection as the tip

moves toward the surface ( 1 ). At a very small separation, long range forces may deflect

the tip down toward the surface (2). Once in contact, the tip deflects to a certain

predetermined amount and then begins to retract (3). If there is an adhesive force

between the tip and the sample, the cantilever may deflect in the opposite direction as it

sticks to the surface (4). Finally, the tip snaps free and the deflection returns to zero (5).

The cantilever displacement can then be converted to a force if the spring constant of

53 the cantilever is known. In this way, tip-sample adhesive forces can be measured and compared.

Tip-sample separation (nm)

Figure 2.2. An example of a force-distance curve used to measure tip-sample adhesive forces.

2.2.1.2 Tapping Mode AFM

Tapping mode AFM uses an oscillating cantilever to intermittently contact the

surface, which applies less shear force to the sample than contact mode and is especially

useful for soft samples. The oscillation amplitude of the cantilever is held constant and monitored as the tip is scanned across the surface. Any change in the amplitude results in an instant and appropriate change in the piezoelectric scanner height to return the amplitude to the setpoint value. This produces a topographic image of the sample. The phase of the oscillation is also monitored during scanning and the phase lag between the

drive signal and the cantilever response is recorded. An image of the relative phase

change at each xy position on the sample surface can be plotted. This is extremely

useful when the topography is minimal, as sample moiphology can still be determined.

54 Phase changes are caused by tip-sample interactions that can be attributed to

viscoelasticity, hydrogen bonding. Van der Waals force, etc.'^"^

2.2.2 AFM Probe Surface Chemistry

Tips for tapping mode AFM are generally made from silicon, which is the

material of choice because of its resonance frequency and stiffness. The probes are manufactured using typical semiconductor processes, such as lithography and etching.

Silicon probes have a single beam cantilever geometry with a tip having a radius of ~5-

50 nm. An oxide layer of SiO: covers the surface and is terminated in silanol groups at the air interface.

a b

Figure 2.3. (a) Tapping mode cantilever and tip (b) Contact mode cantilever and tip.

Tips for contact mode AFM are made from silicon nitride. This material is

harder and less brittle than silicon and is better able to withstand the constant friction with the surface without becoming dull or breaking. The triangular, double cantilever design allows one to tailor the spring constant by changing the dimensions.

55 Silicon nitride ceramics can be prepared by a variety of methods including direct nitration of silicon and vapor phase reactions of silanes with ammonia.^ Silicon nitride

7 8 AFM probes are prepared via a low-pressure chemical vapor deposition process. '

Dichlorosilane and ammonia react to form silicon nitride with the approximate chemical formula of Si3N4, which depends on the exact ratio of starting materials. Like many other ceramics, Si3N4 oxidizes at ambient conditions.^' A considerable amount of

11-17 research has been done to determine the structure of the oxide. ' The consensus is that a fully oxidized surface has a Si02 layer at the solid-air interface with a SiNxOy gradient down to the bulk Si3N4. XPS, FT-IR, and electrochemical studies have shown that secondary silylamines and silanols are the main functional groups on a partially oxidized surface. However, the exact ratio of functional groups can vary greatly and depends on both the deposition conditions and subsequent cleaning procedures.

2.2.3 AFM Probe Surface Modification

Thiol self-assembly on gold is one of the most often used methods to modify

AFM probes. ' It is well known that thiols with sufficiently long alkyl chains self assemble on gold and other metal surfaces to form densely packed monolayers.""^"^^ The alkyl chain can be terminated with a variety of functional groups so the resulting surface

chemistry can be tailored. This technique is relatively simple, but the AFM probes must

first be coated with a chromium or titanium adhesive layer followed by gold. These

coatings increase the radius of the tip, which results in a loss of resolution. Also, the difference in the thermal expansion coefficient between the gold and the silicon or silicon nitride can cause the cantilever to become warped.

56 ^'"^^'^^ Silane chemistry is also applicable to AFM probe modification. Silanol

groups on the surface of silicon or silicon nitride can be modified with mono, di, or tri

'^^""^' functional alkoxy or chlorosilanes.^^ Figure 2.4 illustrates the possible reactions that can occur.

OH OH OH OH OH OH QH" OH | Y Q" 9" I I I I

vertical polymerization self-assembly covalent attachment

Figure 2.4. Possible reactions of trichloro- and monochlorosilanes with surface silanols.

The reaction begins with the hydrolysis of the chloro or alkoxy groups to form silanols, which can then react with surface silanols to form Si-O-Si linkages with loss of water. Trifiinctional silanes are most often used for AFM tip surface modification. In

the best case scenario a complete, single monolayer is formed, which has been shown to consist of laterally polymerized silane chains that are intermittently covalently attached

^'^'^"^ to the surface. However, vertical polymerization is also possible which results in

multilayer, irregularly formed coatings. Special attention to reaction conditions is needed to form a monolayer. Difunctional silanes also permit vertical polymerization

and irregular surfaces and reproducibility is often a problem if the conditions are not

57 carefully controlled. Monoflinctional silanes can only covalently bond once to the

surface and cannot form multilayers. However, since the reaction occurs by random

attachment of the molecules to the surface, a complete monolayer is never formed as

molecules that are already attached to the surface sterically screen some surface

silanols.^^'^"* The advantage of silane chemistry versus thiols is that no metallization is

needed so the tips stay sharp. This work will focus on the use of monochlorosilanes to

modify AFM probes in order to avoid multilayer formation that is a common problem with trichlorosilanes.

2.2.4 Surface Chemical Analysis Using AFM

Atomic force microscopy was first used to look at sample topography at the nanometer level. Now various techniques allow researchers to study the chemical make-up of a surface while taking advantage of the outstanding resolution of AFM.

Force curve, lateral force, and phase imaging modes will be discussed. There are several reviews covering the wide literature of AFM for chemical analysis so only a few

^"^"^"'^ representative examples of each technique will be given.

2.2.4.1 Force Curves

As previously discussed, tip-sample adhesive forces can be measured using force curves. A force curve is generated by bringing the AFM tip into contact with a

surface and then retracting it. Adhesion between the tip and sample causes a deflection

of the cantilever that is recorded versus the tip sample separation distance. The

58 —

adhesive force can be calculated if the cantilever spring constant is used to convert the deflection to a force.

Several experimental parameters are important. The tip radius can affect the adhesive force, as a larger tip will have more functional group interactions with the

sample and hence a higher adhesive (or repulsive) force. The adhesive force is often normalized by dividing by the tip radius.^*^ Adhesive force measurements in air can be complicated by adsorbed water and contaminants on the sample. At small tip-sample separations the water can exert a capillary force that may be greater than the adhesive

"^^'^^ forces one is attempting to measure. The measurements can be done in a liquid environment to avoid such interference."*''^' Single adhesive force measurements can vary widely so most researchers take many force curves and find the mean of the resulting histogram.^"^^

> 1 1 — 1 I r 0 5 10 15 2« 5E5 Z displacBment (nrrij

Figure 2.5. Force-distance curves for COOH-CH3, CH3-CH3, and COOH-COOH tip- samples interactions in ethanol. Reprinted with permission from reference 45. Copyright 1995 American Chemical Society.

59 Numerous groups have studied tip-sample adhesive forces/^"^^'^^'^^''"'^^''*^ Figure

2.5 shows an example from a paper by Noy, et al. that reveals the different adhesive forces between thiol modified AFM tips and surfaces.^^ Their results showed that, in ethanol, an AFM tip modified with a carboxyhc acid-terminated thiol has the highest adhesive force with a similarly modified surface. The lowest adhesion was for a carboxylic acid tip with a methyl surface.

2.2.4.2 Lateral/Friction Force Images

Lateral force AFM is done by scanning the tip perpendicular to the cantilever axis. Images are obtained when the surface has areas with different friction coefficients.

The greater the difference in the friction coefficients between the tip and the surface, the better the contrast and resolution.

Figure 2.6. AFM images of a COOH-terminated thiol square in a CH3 matrix, (a) height image, (b) lateral force image with a COOH-temiinated AFM tip, (c) lateral force image with a CHr tenninated tip. Reprinted with permission from reference 45. Copyright 1995 American Chemical Society.

Unmodified silicon nitride tips can be used to image surfaces in friction mode

because, even with no modification, there are silylamines and silanols on the tips.'*''"**^

However, chemically modifying the AFM tips allows the user to tailor the friction force

''^'^'^ contrast for various surfaces.'''^"'^^ Friction force images of a carboxylic acid-

60 terminated thiol square in an alkanethiol matrix are shown in Figure 2.6a-c. ' The

topographic image in Figure 2.6a shows nothing because there is little height difference between the two thiols. The image in Figure 2.6b was taken with a carboxyHc acid

modified tip and Figure 2.6c with an alkanethiol tip. The friction force contrast is reversed when the tip chemistry changes.

As with force curve measurements, various experimental parameters are important to obtain reproducible friction force images. The tip radius must be determined because larger tips will have more functional group interactions with the surface and hence higher friction forces. At ambient conditions adsorbed water can introduce a capillary force between the tip and sample that can affect the image contrast.

Friction force images can be done in liquid media to avoid such complications.^^

Instrument parameters such as scan speed and the applied normal force can also change the observed images.^

2.2.4.3 Tapping Mode Phase Images

The phase of oscillation of an AFM tip in tapping mode can change as the result

of tip-sample interactions. Phase lag is dependent on sample elasticity, viscoelasticity, chemistry, adsorbed water, tip quality, and instrument parameters. If the interactions are strong, either attractive or repulsive, the phase lag between the applied and recorded oscillation will be large. In this way, contrast in phase images can be enhanced by chemically modifying the tip to interact more strongly with one area of the

"^^ sample than another."^'^' Figure 2.7 shows two phase contrast images of a carboxylic acid thiol-modified hollow square pattern in an alkanethiol matrix."^' The first image

61 was taken with a COOH-modified tip and the second with a CHs-modified tip. The

phase contrast was inverted when the tip was changed as a result of different tip-sample

adhesion forces that affected the phase lag.

Figure 2.7. Tapping mode AFM phase images of a COOH-terminated thiol open square in a CH3-terminated thiol matrix with a (c) COOH-modified AFM tip (d) CHs-modified tip. (25 ]im X 25 fim image) Reprinted with permission from reference 51. Copyright 1998 American Chemical Society.

a b c

Figure 2.8. Tapping AFM phase images of a 20:80 blend of PMMA/polybutadiene changing only the tapping force: (a) rsp=0.84 (b) rsp=0.73 (c) rsp=0.58. Reprinted with permission from reference 54. Copyright 2000 American Chemical Society.

Experimental parameters are also important with this imaging technique. As previously mentioned, the effects of tip radius and capillaiy force distortion must be taken into account when doing AFM. The oscillation amplitude of the cantilever plays an important role in tapping mode.^^'^^ The cantilever has a certain free air amplitude.

62 Ao, which is dampened to a constantly maintained setpoint amplitude. Asp, when it

begins scanning the surface. The value of Ao and the setpoint amplitude ratio,

rsp=Asp/Ao, determine the tapping force, with a higher value of Ao and a lower value of

rsp resulting in a larger force applied to the surface. Changing those values can affect

the phase image and even invert the contrast. Figure 2.8 shows how changing only the

value of rsp can result in different phase images of the same sample. Generally,

smaller amplitudes favor chemical contrast while at higher amplitudes contrast is

affected by sample elasticity.""^

2.3 Experimental

2.3.1 Materials

Atomic force microscope probes were purchased from Nanodevices, Inc.

Silicon Tap300 tapping mode probes had a resonance frequency of -300 kHz. Silicon nitride 0TR8 contact mode probes had a nominal spring constant of 0.15 N/m for the

200 \im long cantilever. A Nioprobe tipcheck sample was purchased from Aurora

Nanodevices, Inc. 3-Aminopropyldimethylethoxysilane (APDMES), 3,3,3- trifluoropropyldimethylchlorosilane (F3), n-propyldimethylchlorosilane (PDCS), tris(trimethylsiloxy)chlorosilane (Tris(TMS)Cl), and tridecafluoro-1,1,2,2- tetrahydrooctyldimethylchlorosilane (FDCS) were purchased from Gelest and used as received. After purchase from Aldrich, octadecylmercaptan (OCT) was vacuum

1 distilled under reduced pressure and 1 -mercaptoundecanioc acid (MDA) was recrystallized from water. Xylene, toluene, ethanol, and silicon tetrachloride were purchased from Aldrich and used as received. Dow Coming Sylgard 184 silicone

63 elastomer base and curing agent were used as received. Silicon wafers were purchased from International Wafer Service (<100> orientation, P/B doped, resistivity from 20.0 to 40 Q. cm). House-purified water (reverse osmosis) was further purified using a

Millipore Milli-Q system (18.2 MQ cm).

2.3.2 Experimental Procedures

2.3.2.1 Silane Modification

AFM probes and silicon wafers were modified with silanes via a vapor phase reaction. Oxygen plasma was used to remove contaminants from the surfaces: 2 minutes for the probes and 5 minutes for the silicon wafers. The wafers, probes, or both were then suspended in a Schlenk tube above 0.5 mL of the selected silane. The vessel was sealed and heated to 65-70 °C in an oil bath for a predetemiined amount of time.

After the reaction the samples were removed and carefiilly dipped sequentially into toluene, ethanol, and water to remove physisorbed silane.

2.3.2.2 Silicon Dioxide Growth and Modification

Silicon dioxide was grown from Tris(TMS) monolayers by a chemical vapor

^^''^'^ deposition process that was refined by Jia and coworkers. Silicon wafers were first modified with Tris(TMS)Cl in a vapor phase silane reaction as described above. The residual silanols on the surface that were not modified by the Tris(TMS) could react with another silane. The samples were placed into one chamber of a vessel separated by a valve from another chamber containing SiCU. The SiCU was cooled to -23 ''C with a

64 slush of liquid nitrogen and ortho-xylene. Both chambers were separately purged with nitrogen and pumped down to 50 mTorr with the help of a Mano-Watch (Model MW-

1000, Instruments for Research and Industry, I"R, Inc.) device attached to the vacuum

system. The valve between the two chambers was opened for 1 minute to allow the

SiCU vapor to diffuse to the sample chamber and react with the surface silanols. The valve was closed and the sample chamber purged with nitrogen and pumped down to at least 20 mTorr to remove the residual SiCU. The samples were then exposed to air for 5 minutes to allow water vapor to adsorb to the freshly grown silica. The samples were again purged with nitrogen, pumped down to 20 mTorr, equilibrated to 50 mTorr, and

exposed to SiCU for 1 minute. This two step cycle, exposure to SiCU followed by exposure to air, was repeated to get the desired silicon dioxide thickness. The silicon dioxide was then modified with another silane to produce a surface with two chemical functionalities.

2.3.2.3 Microcontact Printing

Gold surfaces were prepared by thermally evaporating 1 0 nm of chromium onto a silicon wafer followed by 50 nm of gold. The gold surfaces were oxygen plasma cleaned for 5 minutes.

PDMS stamps with various features were prepared. Silicon wafers lithographically patterned with post features' were modified with FDCS to make them

hydrophobic. Ten parts by weight of PDMS resin was mixed with 1 part curing agent.

The mixture was stirred for 2 minutes and then degassed under vacuum until no bubbles were present. The mixture was poured over the hydrophobic silicon posts and placed

65 under vacuum to ensure complete filling of the features. The PDMS was cured in an oven at 60 overnight and carefully removed from the silicon mold.

Solutions of 1 mM OCT and 1 mM MDA were prepared in ethanol. Single flinctionality surfaces were made by dipping a freshly cleaned gold substrate in a

solution for 1 hour. Binary surfaces were made using the PDMS stamps. A few drops of thiol solution was applied to the stamp and allowed to sit for 10 seconds before being blown dry with nitrogen for 20 seconds. The stamp was placed on a clean gold substrate for 30 seconds with initial light finger pressure to ensure contact. After

stamping, the sample was placed in the other thiol solution for 1 hour to modify the unstamped regions.

2.3.3 Characterization

AFM measurements were performed on a Dimension 3 1 00 Scanning Probe

Microscope (Veeco, Inc.) equipped with a Nanoscope Ilia controller. The scanner head was calibrated against a standard sample to ensure that accurate dimensions were recorded. Measurements in liquid were made using the glass tip holder accessory. All

AFM images were optimized for viewing using a third order flattening algorithm to remove bowing and bending distortions. X-ray photoelectron spectra were obtained on a Physical Electronics Quantum 2000 Scanning ESCA Microprobe. Dynamic contact angle measurements were performed using a Rame-Hart telescopic goniometer and a

Gilmont syringe with a 24-gauge flat tipped needle. Advancing (0a) and receding (Or) contact angles were measured while the probe fluid was added to and withdrawn from the drop, respectively. A JEOL JSM 6320F was used to obtain field emission scanning

66 electron micrographs. Monolayer thicknesses were determined using a Rudolph

Research AutoEL-11 ellipsometer.

2.3.3.1 Chemical Force Microscopy

Adhesion forces were determined from tip deflection versus tip-sample separation plots recorded as the AFM tip was brought down into contact with and retracted from the surface. The tip deflection was converted to a force using the nominal spring constant of the cantilever, 0.15 N/m, as stated by the manufacturer. The average adhesion force value was determined from 15-20 different spots on the sample.

Tapping mode images were obtained for a variety of surfaces with patterned chemistry. The instrument parameters were kept the same as possible for each image.

The phase image was refined with a lowpass filter which replaces each data point in the image with the weighted average of the 3x3 cell of points surrounding and including the point. Most features of interest were larger than 3 pixels by 3 pixels so they were not affected, but the random background noise was greatly reduced. The roughness analysis feature of the AFM software was then used to identify the highest phase peaks by using the absolute highest peak as the reference point and adjusting the threshold setpoint. The peaks were averaged to get the phase difference between the two areas of the surface with different chemistries.

2.3.3.2 X-ray Photoelectron Spectroscopy

X-ray photoelectron spectroscopy (XPS) is a nondestructive surface analysis technique that probes the elemental composition of the outer tens of angstroms of a

67 material. X-rays from a monochromatic source bombard the sample." The X-rays interact with and transfer their energy to the inner shell electrons of the atoms in the sample, ejecting a photoelectron. The kinetic energy and number of electrons per unit

time is recorded by a detector. Subtraction of the electron's kinetic energy, Ek, from the

initial energy of the X-ray, hv, gives the binding energy with which the photoelectron was held in the atomic orbital, Eb.

E,=hv-E^ (2.1)

A typical XPS spectrum shows the binding energy versus the photoelectron counts per second. Electrons in each orbital of the various elements have a different binding energy. Reference tables of binding energy allow the determination of all the elements in the sample. Additionally, the individual peaks can be integrated to determine the atomic concentration of each element.

The binding energy of an electron is influenced by its chemical environment.

Neighboring functional groups that add or subtract electron density can shift the binding energy. These chemical shifts can be used to identify different functional groups in the

sample. For example, an electron in the carbon 1 s orbital in a -CF3 group has a higher

binding energy than the carbon 1 s electron in a -CH3 group because fluorine is more electronegative than hydrogen. Accurate reference tables of chemical shifts due to a

functional group are needed for correct identification. Also, not all functional groups can be detemiined because chemical shift differences are often small and peaks overlap.

Photoelectrons that arc ejected from the various atoms in the sample experience elastic and inelastic collisions on the way to the detector. The intensity of the ejected

68 photoelectron reaching the detector without a loss in kinetic energy decays exponentially with sample depth:

f l(=)^rexp (2.2) J

where V is the initial photoelectron intensity at a point in the solid, ?La(Ei<) is the

attenuation length of the photoelectron with kinetic energy Ek, 0 is the angle between

the surface and the photoelectrons' path to the detector, and z is the depth below the

surface the photoelectron starts from. The deeper inside the sample it starts, the more

chance it has to collide with other atoms before reaching the surface. Therefore, only the photoelectrons from the outer tens of angstroms are detected.

A few terms need to be defined before the discussion continues. Escape depth,

is the distance normal to the surface at which the probability of an electron escaping

without significant energy loss due to inelastic scattering drops to e ' (36.8%) of its

original value. Attenuation length, X^, is the average distance that an electron having a given energy travels between successive inelastic collisions assuming the elastic

collisions are insignificant. Sampling depth, S, is the distance normal to the surface from which a specific percentage of the detected electrons originate.

Escape depth is related to attenuation length as shown in Equation 2.3,

4r = /i sin 6^ (2.3)

where 9 is the angle between the path of the photoelectron to the detector and the surface.

The sampling depth is typically taken as 3(!^ or 95% of the observed signal. Therefore,

the maximum sampling depth occurs when 6=90 and is a minimum when 0 approaches

69 0. A series of spectra at different angles allows nondestmctive chemical depth profiling of the sample surface. The exact sampling depth depends on the attenuation length, which varies with the material and the X-ray energy. As an example, using a value of

20 A for the attenuation length of the carbon Is electron, the sampling depths at 15" and

75° are 16 A and 58 A respectively.

2.4 Results and Discussion

2.4.1 Silane Modification of Probes and Surfaces

The silanes used in this project were chosen to provide a range of functionality

(Figure 2.9). All the silanes contained only a single hydrolyzable group permitting one bond with the substrate so no multilayers could form. Short chain silanes were used to allow comparison of the chemical effects, not the chain length. The amine-

functionalized silane was chosen because it can participate in hydrogen bonding interactions. The short chain alkyl and fluoro silane have only van der Waals interactions. Finally, a longer chain fluorosilane was used to provide a very hydrophobic surface.

CF3(CF2)5CH2CH2Si(CH3)2CI

Figure 2.9. Silanes used for chemical modification of AFM probes; (a) 3- aminopropyldimcthylcthoxysilanc, (b) 3,3,3-trifluoropropyldimethylchlorosilane. (c) n-

propyldimcthylchlorosilane, (d) tridccatluoro- 1 , 1 ,2,2- tetrahydrooctyldimethylchlorosilanc.

70 The AFM tip plays a critical role in obtaining accurate images of a sample surface. A sharp, symmetric tip with a radius smaller than the features of interest provides the best resolution. The contact area of the tip on the sample also influences

the adhesive forces that are measured. Therefore it is important to monitor the tip condition. One option is to take an SEM image of the tip and measure the radius.

Figure 2. 10 shows SEM images of a silicon tapping mode probe before and after it was

oxygen plasma cleaned for 2 minutes. There is no significant difference in the tip radius. A similar check was performed to ensure that plasma cleaning did not affect the silicon nitride probes.

a b

Figure 2.10. SEM images of a silicon nitride AFM tip (a) before and (b) after a 2 minute oxygen plasma cleaning.

SEM imaging was too time consuming for routine tip evaluation so an alternate method was used. A niobium film was purchased which had sharp surface features of

~5 nm which are smaller than the radius of the typical AFM tip used in this study, -20-

40 nm. Figure 2. 11 a shows an SEM image of the standard. When it was imaged by

AFM, the features were slightly distorted as shown in Figure 2.11b. A cross-section

71 1

analysis of the image revealed features that were a mirror image of the tip. Equation 2. was used to calculate the tip radius from the width and height of the features.

a b

Figure 2.11. (a) SEM image of a commercial niobium film with sharp features that are smaller than an AFM tip,^*^ (b) Contact mode AFM height image of a similar film.

Vapor phase reactions were used to silanize the silicon nitride and silicon AFM probes along with silicon wafers in the same vessel. The silicon wafers allowed dynamic contact angle and thickness to be measured, since the AFM probes were too small for such techniques. Table 2.1 shows the XPS atomic concentrations for silicon

AFM probes modified with the specified silane for 1 day. The spectra were taken with

a 1 00 |im spot size so the tip and part of the cantilever were included in the measurement. Two minutes of oxygen plasma cleaning removed most of the carbon, leaving only 3.2%. Reaction with F3 was revealed by the presence of 4.8% fluorine as

72 well as a hydrophobic contact angle, 92/75, as measured on a similarly modified silicon wafer. The surface silanized with PDCS also had a hydrophobic contact angle, 91/79, and XPS showed an increase in the carbon concentration. The polar amine group gave the APDMES surface a hydrophilic contact angle of 78/39 and the carbon concentration

increased to 14.5%. Nitrogen was not detected by XPS for two reasons: ( 1 ) nitrogen

has a small X-ray cross section and is difficult to detect at small concentrations even on the best samples and (2) the angle of incidence was 45" which means that a lot of the signal was coming from the substrate and not just the monolayer. (A 15° take-off angle, which would look at a thinner slice of the surface, could not be used because the signal to noise ratio was poor.)

Table 2.1. XPS analysis of silane-modified silicon tapping mode AFM probes.

Atomic composition Water AFM tip surface (45° take-off angle data) contact chemistry angle'* Si 0 C F N Au (degrees)

Plasma cleaned 29.8 46.9 6.2 16.6 0.6 spreads

-CFs 29 33.5 16.2 4.8 15.6 1 92/75

-CH3 28.6 45.1 9.9 15.8 0.6 91/79

-NH2 20.5 35.4 32.3 16.4 0.6 78/39

^Contact angles were recorded on flat silicon wafers modified in the same vessel.

The XPS data for the modified silicon nitride AFM probes is found in Table 2.2.

The high concentration of oxygen points to the presence of an oxide layer on the surface.

The Si2p signal in the XPS actually appears as two overlapping peaks that are just

73 barely resolved at -101 .5 eV and 100.5 eV, corresponding to the Si-0 bond in silicon

dioxide and the Si-N bond in silicon nitride respectively. The opposite side of the

cantilever is coated with gold to improve the reflection of the laser beam and during the

evaporation process a small amount of gold, < 1% by XPS, made its way to the tip side.

Fluorine is present on the surface after the reaction with F3 indicating that the

silanization was successful. Reaction with PDCS increased the carbon concentration and a distinct Cls peak was seen in the XPS at 283.3 eV. An increase in nitrogen

concentration was not observed after reaction with APDMES because it was obscured by the high concentration of nitrogen in the silicon nitride substrate. However, close inspection of the Nls region shows a small but distinct peak at -399 eV corresponding to APDMES at the base of the more intense peak at 396 eV corresponding to the silicon nitride. Similar reactions were not run on flat silicon nitride wafers because the reaction conditions used to prepare silicon nitride affect the final composition. The substrate

would be different from the AFM probes and it was decided that little additional information would be gained.

Table 2.2. XPS analysis of silane-modified silicon nitride contact mode AFM probes.

Atomic composition AFM tip surface (45° take-off angle data) chemistry Si 0 C F N Au

Plasma cleaned 29.8 46.9 6.2 16.6 0.6

-CF3 29 33.5 16.2 4.8 15.6 1

-CH3 28.6 45.1 9.9 15.8 0.6

-NH2 20.5 35.4 32.3 16.4 0.6

74 2.4.2 Adhesion Force Measurements

The adhesion force between a chemically modified AFM tip and a

flinctionalized surface were measured. A typical force vs. displacement curve was previously shown in Figure 2.2. The radius of the tip affects the measured adhesion

force because a greater area of contact means more tip-surface interactions. The forces were divided by the radius of each tip to allow a more direct comparison between the different functionalized AFM probes. The radii of the probes were: 64 nm for -NH2, 25 nm for -CF3, 155 nm for -CH3 and 165 nm for the unmodified silicon nitride probe which has a surface layer of-Si-OH. All of these values were measured after the adhesion experiments.

Table 2.3. Normalized adhesion force (N/m x 10'^) between alkylsilane-modified silicon nitride contact mode AFM probes and similarly modified silicon wafers, measured in air.

Surface AFM tip chemistry chemistry -NH2 -CF3 -CH3 -Si-OH

-NH2 13. 1 (±0.6) 26{±3) 3(±i) 20

-CF3 53(±3) 20(±4) 3(±i) 15 (±3)

-CH3 46(±6) 20(±3) 3(±i) 16(±3)

Table 2.3 shows the experimentally determined forces for each tip and surface

combination. Looking down each column, there is not a large difference between the interactions of each tip with the various surfaces. Comparing the different tip chemistries across rows reveals that the -CH3 modified tip has the lowest adhesion

force on all the surfaces. That is reasonable because the only attractive forces are due to

75 van der Waals interactions. What is a little surprising is that the -CF3 modified tip had much stronger adhesive forces than the -CH3 tip. The surface energy of the -CF3 group

is the lowest of any of the tip functionalities and it would seem intuitive that the adhesion force would be lowest. However, a possible explanation is that the -CF3 group has a larger permanent dipole due to the electron withdrawing effect of the

fluorine atoms. This additional interaction is why the adhesion force is greater for -CF3 than -CH3. A similar observation has been noted by others.^'

There is known to be capillary condensation of water between the AFM tip and sample under ambient conditions and the capillaiy force can be greater than the

adhesion forces being investigated.'^°'"*'^'^^ The capillary force can be eliminated by doing the measurements in liquid. The adhesion forces were detennined again, but with the tip and sample submerged in water. Table 2.4 shows the adhesion forces normalized by the tip radius. The adhesion forces for all tip-sample combinations were

less than in air, showing that capillary force was indeed affecting the results under

ambient conditions. The largest decrease was seen for the amine modified tip. which

being hydrophilic, probably had the most water vapor condensed on it from the air.

Looking down each column, again there was little difference between the modified tip and each surface.

Comparing the different tips, the -CF3 functionality showed the highest

adhesion force with all surfaces followed by -NH2 and finally -CH3. Hydrophobic interactions could explain the high adhesion force between the -CF3 modified tip and the -CF3 and -CH3 surfaces. However, hydrophobic interactions should not be as strong with the more hydrophilic -NH2, but the adhesion force data shows otherwise. A

76 reason for this discrepancy is not clear. The -CH3 tip has the least adhesion force with

all the surfaces. Apparently the hydrophobic interactions are not as strong and only van

der Waals forces contribute to the measured force. Although, it should be noted again

that comparisons between different tips can be problematic even after trying to

normalize the data using the tip radius.

Table 2.4. Normalized adhesion force (N/m x 10") between alkylsilane-modified silicon nitride AFM contact mode probes and similarly modified silicon wafers, measured in water.

Surface AFM tip chemistry chemistry -NH2 -CF3 -CH3

-NH2 4.9(±o.5) 11(±2) 1 .4(±U.5)

-CF3 3.5(±o.3) 11 (±2) 1.4(±0.5)

-CH3 3.8(±o.5) 9(±2) 1.5(±0.5)

The adhesion measurements with functionalized tips did not show very good contrast between the various surfaces. The problem may have been that the adhesion

forces were similar and the uncertainty in the measurement did not allow a distinction to be revealed. Each force value was an average of 15-20 measurements on the same

surface. However, it has been shown that the mean value of hundreds of individual

measurements provides more precise data.^"^"^"'^ The research focus turned to distinguishing chemical differences on surfaces with two different chemistries.

77 2.4.3 Tris(TMS)/Si07/Silane Surfaces

2.4.3. 1 Silicon Dioxide Growth and Modification

Surfaces with areas of different chemical functionality would be useful to

determine if the chemically modified AFM probes could distinguish one area from

another. A method to prepare such a surface was developed in our labs by a former

student.^^ Figure 2.12 shows a schematic representation of the process. A silicon wafer

was modified with Tris(TMS)Cl, which is a bulky silane that reacts randomly and inefficiently on the surface leaving areas of exposed silanols that are too sterically

hindered to allow another Tris(TMS) molecule to fit. Those silanols can be modified by

a smaller silane*''* or used as a reactive anchor to tether single polymer molecules to the surface. In this particular process, silicon dioxide was grown from the interstitial silanols by alternately exposing the surface to SiCU and water vapor. The silicon tetrachloride was hydrolyzed to Si(0H)4 by the adsorbed water on the hydrophilic silanols and then reacted by a condensation mechanism to form Si02. Multiple cycles were used to grow thicker silicon dioxide. Finally, the silicon dioxide was modified with another silane to produce islands of one chemical functionality surrounded by a sea of another. (A similar binary chemical surface could have been prepared by modifying

the interstitial silanols with a second silane without first growing silicon dioxide.

However, the average surface area of each silanol patch was estimated to only be -0.5 nm" on a fully covered Tris(TMS) surface. The second silane would be difficult to detect at the resolution of the AFM used in this study.)

78 Table 2.5. Kinetics of the vapor phase reaction of Tris(TMS)Cl with a silicon wafer.

Time (hrs) Contact angle (degrees) Tris(TMS) coverage (%)

1 74/60 67 4 82/67 77 24 82/70 77 48 92/74 87 72 95/78 90 96 99/84 93

79 The kinetics of the Tris(TMS) reaction are shown in Table 2.5. The advancing contact angle data reveals that the initial reaction is fast, but the reaction continues over a period of days as evidenced by the gradual increase in the receding angle. The surface coverage was calculated using the Israelachvili-Gee equation,

)' (l = (l cos (l cos + cos ef ./I + 6', + /, + 0, Y

(2.5)

where f| and 0| are the surface coverage and contact angle for the first component, f2

and O2 are the same for the second component, and 9 is the measured contact angle of the composite surface. A contact angle of 108° was used for a complete Tris(TMS) surface and a value of 0° was used for a complete silanol surface.

Different thicknesses of silicon dioxide were grown on surfaces with varying

Tris(TMS) coverage. The goal was to find the best combination of coverage and cycles to prepare isolated islands of Si02 that could be further modified. A three day vapor phase reaction in Tris(TMS) followed by 2 cycles of silicon dioxide growth gave the most reproducible results. AFM images of the surface after each stage of the reaction are shown in Figure 2.13.

The images qualitatively show the change in surface topography after each step and Table 2.6 reveals the quantitative differences. The average and root-mean-square roughness increase after each step, but that does not give a complete picture. Skewness

is a measure of the symmetry of the data around the mean, with values greater than or

less than zero indicating more peaks or valleys on the surface respectively. Kurlosis is a

80 measure of the sharpness of the height distribution ftinction, which for a normal

Gaussian distribution would be 3. Values less than or greater than three indicate few or many features in the extreme tails of the distribution. The silicon wafer and Tris(TMS) surface both have a slightly positive value of skewness, and kurtosis just above three indicating a nearly Guassian distribution of peak heights. However, the silicon dioxide surface has large positive skewness, and kurtosis much higher than three, indicating many peaks, that are much taller than the mean height. This provides a better description of the surface than just average roughness.

Figure 2.13. AFM tapping mode height images of (a) plasma cleaned silicon wafer, (b) silicon wafer modified with Tris(TMS)Cl for 3 days, (c) Si02 posts grown in the incomplete Tris(TMS) monolayers (0.5 x 0.5 |j,m scan size, 2 nm z scale).

81 Table 2.6. Roughness parameters for Tris(TMS)Cl-modified surfaces determined by AFM.

Surface Average RMS Skewness Kurtosis roughness, Ra roughness, 5

Si wafer^ 0.082 0.109 0.119 3.133 Tris(TMS)^ 0.143 0.186 0.203 3.422 Tris/SiO:' 0.502 0.809 2.828 14.862

^ ^ O2 plasma cleaned 5 minutes, vapor phase, 3 days, ' 1 cycle SiCU

Table 2.7. XPS analysis of chemically modified silica posts grown from an incomplete Tris(TMS) monolayer.

Atomic composition Water contact (75° take-off angle data) Surface chemistry^ angle (degrees) Si 0 C F N

Tris(TMS) 49.5 42.1 8.4 92/80

Tris/Si02/-CF3 51 40.4 7.2 1.4 97/84

Tris/Si02/-CH3 51.7 41.9 6.4 98/87

Tris/Si02/-NH2 50.7 40.8 7.8 0.6 92/74

^ Silicon wafer modified with Tris(TMS)Cl for 3 days, 2 cycles SiCU, and finally 1 day with the specified silane.

The silicon dioxide posts were then modified by silanes in a vapor phase

reaction for 1 day. Table 2.7 shows the XPS and contact angle analysis of the surfaces.

The spectra were taken at a 75 degree take-off angle to avoid the shadowing effect that the nanometer high silicon dioxide posts would have at a lower inclination.

Modification of the silicon dioxide with F3 was confirmed by the presence of fluorine, and nitrogen indicated a successful silanization with APDMES. The advancing and receding contact angles increased when the surfaces were modified with hydrophobic

82 PDCS and F3, but a decrease in the receding angle was seen upon reaction with

hydrophihc APDMES.

2.4.3.2 Phase Images

Tapping mode phase images are produced when the AFM measures the

difference between the phase of the oscillating cantilever and the phase of the applied

signal from the piezoelectric device. Interactions between the tip and sample change the phase, with greater interaction causing a greater phase lag. Usually the phase image

is just used qualitatively to view the sample morphology. In this study an attempt was made to quantitatively determine the phase difference between two areas with different chemistry using a new method described in the experimental section. Briefly, a lowpass

filter was used to reduce background noise as shown in Figure 2. 14. The main phase peaks were not affected, but the background noise was greatly reduced. Then the ~25 tallest peaks were averaged to find the phase contrast between the Tris(TMS) monolayer and the chemically modified silica posts.

As mentioned previously, the tip and sample interactions depend on the sample elasticity, viscoelasticity, chemistry, adsorbed water, tip quality, instrument settings and

possibly other parameters." In order to identify the chemical component it was important to keep the other factors constant. The different silane monolayers should not affect the modulus of the surface, which was dominated by the stiff silicon dioxide. A free air amplitude, Ao, setting of 100 mV was used for all samples. The damping ratio,

rsp, varied a bit, but was kept as high as possible, typically -0.8, while still maintaining contact with the surface. The tip quality always differed slightly between tips so the

83 same one was used on each set of surfaces. The radii of all the probes were between

19-22 nm.

a b

Figure 2.14. AFM tapping mode phase images of Si02 posts grown from Tris(TMS) monolayers (a) before and (b) after a lowpass filter algorithm was applied to reduce background noise.

Table 2.8 shows the average phase contrast as recorded using a variety of fiinctionalized AFM probes. Comparisons should only be made down columns, not across rows because the tip shape can affect the image. For each of the tip chemistries

there was little difference in the phase contrast on the various surfaces. The most

important information on the table is the control experiment in the last row. The silicon dioxide areas were modified with Tris(TMS) so the entire surface had the same chemistry and the only difference was the topography. All of the modified tips showed at least as much phase contrast on this chemically homogeneous surface as they did on the mixed surfaces. Topography also affected the phase contrast so the effects of surface chemistry could not be determined.

84 Table 2.8. Degrees of phase contrast between chemically modified AFM tips and similarly modified silica post surfaces.

Surface AFM tip chemistry chemistry'' -NH2 -CF3 -CH3 -Si-OH

-NH: 3.4(±o.6) 3(+i) 7( + 2) 20

-CF3 2.4(±o.2) 3(±i) 7(±2) 25

-CH3 2. 3 (±0.3) 3(±i) 1 1(±4) 23

-Si(CH3)3 14(±5) 3.0(+o.6) 6(±n 39(±3)

Silicon wafer modified with Tris(TMS)Cl for 3 days, 2 cycles SiCU, and finally 1 day with the specified silane.

2.4.4 Microcontact Printed Thiol Surfaces

2 AAA Preparation

Surfaces with patterned chemistry and no topography were prepared using microcontact printing of thiols on gold (Figure 2.15).*^^'^^ A patterned stamp of crosslinked PDMS was made by using a lithographically patterned silicon wafer as a

mold. A 1 mM solution of octadecylmercaptan (OCT) in ethanol was used to "ink" the stamp, which was then placed in contact with a gold surface. The thiol was transfen-ed from the PDMS stamp to the surface and replicated the pattern, producing modified

areas and bare gold. The samples were then placed in a 1 mM solution of 11- mercaptoundecanoic acid in ethanol to modify the remaining gold. In this way a binary, chemically patterned surface was prepared.

85 HS(CK.)rCH- PDMS —^— PDMS 9thanol 1 i L J L

lotarnp to

[transfer thiol

HS (CH2),oC02H AU Cr ^hariol oi

Figure 2.15. Schematic diagram of the process to microcontact print alkanethiols onto gold.

Table 2.9. XPS analysis of gold surfaces before and after modification with alkanethiols.

Atomic composition Water contact (45° take-off angle data) Gold surface angle (degrees) C Au 0 S

as deposited 39.9 53.7 6.5 89/41

plasma cleaned^ 23.2 70.9 5.8 34/8

OCT modified'' 65.7 33 1.3 0.1 110/98

MDA modified' 48.9 46.4 4.3 0.5 33/8

^5 ''24 ""^ minutes, O2 plasma; hr, 1 mM in ethanol; 24 hr, I mM in ethanol.

86 XPS was used to characterize the gold surface as well as the thiol monolayers

(Table 2.9). After evaporation, the gold surface had a substantial amount of carbon

contamination and the contact angle was accordingly high. Plasma cleaning removed

some of the carbon, but XPS still showed 23.2%, however the contact angle was much more hydrophilic. Thiols are known to displace carbon contamination on gold and preferentially adsorb so the fact that a small amount remained after cleaning was not a concern.^'' The gold surface had a roughness of -1 nm and a grain size of ~20 nm as determined by AFM. Sulphur was apparent in the XPS spectra of both the MDA and

OCT monolayers, and the carbon concentration increased when either was adsorbed.

10^/98° The water contact angles were the most telling, with OCT being hydrophobic 1 and MDA being hydrophilic 3378°.

The microcontact printed samples were characterized in a variety of ways. SEM was used to view the patterns (Figure 2. 16a). In the optical micrograph in Figure 2.16b,

the gold in the circles was not modified by OCT like the matrix and it preferentially

dissolved in a 1 mM NaCN, 1 M NaOH solution. This proved that the microcontact printing produced a complete monolayer. The pattern of a binary OCT/MDA surface

was easily viewed by cooling the sample and allowing water to condense onto it.

Figure 2.16c shows that the hydrophilic MDA patches were preferentially wetted.

87 Figure 2.16. (a) SEM image of OCT microcontact printed onto gold (circles are not covered by the thiol monolayer), (b) Optical micrograph of a similar sample after the exposed gold was etched 2 hours in a 1 mM NaCN, 1 M NaOH solution, (c) Optical micrograph of water condensation on the hydrophilic MDA features of a surface microcontact printed with 32x32 }im OCT and MDA squares in a checkerboard pattern.

88 Figure 2. 1 7. Tapping mode AFM height (left) and phase (right) images of microcontact printed thiols on gold consisting of MDA circles surrounded by OCT. The AFM tip chemistry was different for each image: (a) -CH3, (b) F13, (c) -NH2, (d) poly( vinyl alcohol).

89 2.4.4.2 AFM Characterization

A variety of size scale features were prepared, but most were too large to be easily viewed by AFM. A patterned surface with 2 micron MDA features surrounded by an OCT matrix proved best. Figure 2.1 7 shows tapping mode height and phase images of the patterned surfaces taken with a variety of tip chemistries. All images

were obtained using the same instrument parameters. It is interesting to note that the height images in all cases show that the MDA features are higher than the OCT matrix even though by ellipsometry the monolayers were nearly the same thickness. One

possibility is that a thin film of water vapor was wetting the hydrophilic MDA areas which increased the tip-sample interactions, causing a dampening of the oscillation amplitude. The AFM falsely interpreted that the dampening was caused by a height difference. This artifact has been reported by others.^^

However, it was the phase image that was of interest. There was very little phase contrast when the surface was imaged with the more hydrophobic tips, -CH3 and

F13, but the features began to appear when hydrophilic tips, -NH2 and poly(vinyl alcohol) were used. The hydrophilic tips may interact with the adsorbed water layer on the MDA regions, enhancing the phase contrast. The grain structure of the gold substrate added noise to the phase image making the edges of the features indistinct.

The images were analyzed in the same manner described previously in an attempt to quantitate the phase contrast. However, the same technique would not work because there were two levels of phase contrast, a large contrast between the chemically distinct areas and a smaller, overlapping contrast as a result of the roughness of the gold surface.

90 An attempt was made to obtain tapping mode images of the patterned surfaces

immersed in water to determine if a thin film of water was causing the distortion of the height image and affecting the phase contrast. Water dampens the oscillation of the cantilever and the silicon probes could not be used because the resonance frequency was not suitable. However, the silicon nitride probes normally used for contact mode had the right resonance frequency when submerged. The trick was to be able to manually find the resonance frequency of the cantilever. After many unsuccessful attempts on two different types of AFMs, research effort was directed toward other areas.

AFM lateral force images were also measured on the microcontact printed thiol surfaces. In this technique the cantilever was scanned perpendicular to the cantilever axis. Friction between the tip and sample surface caused torsion of the cantilever which was recored by the AFM. Different chemistries of the tip and sample caused a change in the friction force that was measured.

Figure 2.18a shows the height and friction images taken with an unmodified silicon nitride tip on a surface microcontact printed with OCT leaving circles of bare

gold. In the height image, the OCT monolayer is slightly higher than the bare gold,

which was expected. The friction force image indicates that there is more friction with

the OCT layer than the bare gold. Figure 2. 1 8b was a similar sample except that the circles were filled with MDA. The height image now reveals no features because the lengths of the thiol chains are nearly the same. The friction force image shows well defined features with a higher friction force observed on the OCT regions. Fricfion

91 force images were also measured using silane-modified probes. However, the images were always poor and even qualitative assessment was difficult.

Figure 2.18. AFM height (left) and lateral force (right) images of (a) OCT matrix microcontact printed onto gold leaving bare gold circles, (b) OCT matrix with MDA- modified circles.

2.5 Conclusions

Silicon and silicon nitride AFM probes were modified with a variety of monofunctional silanes to change the surface chemistry. XPS revealed an oxide layer on each of the materials that allowed the condensation reaction to occur. Amine, alkyl, and fluorine groups were attached to the surface and the various chemistries were

92 confirmed by XPS. Indirect analysis was done by measuring dynamic contact angles on similarly modified silicon wafers, which also indicated that silanization took place.

Adhesion forces were measured for all combinations of modified tips and similarly modified silicon wafers. Capillary forces seemed to dominate the chemical

interactions when the measurements were conducted in air. The forces measured for the hydrophilic amine and native silicon dioxide tips were an order of magnitude greater than when the hydrophobic alkyl and fluoroalkyl modified tips were used on the same surfaces. The same adhesion experiments were conducted in water to avoid the capillary force. The difference in the adhesion force did not vary significantly when a single modified tip was used to probe the various surfaces.

A new method was developed to quantitate the tapping mode phase contrast.

Surfaces with two different chemistries were prepared to test the technique. Silicon dioxide was grown from incomplete Tris(TMS) monolayers and then modified with various silanes. The average phase difference between the modified posts and the

Tris(TMS) background was determined. However, even surfaces with homogeneous chemistry showed phase contrast because of the topography.

Microcontact printing of thiols on gold was used to make chemically patterned surfaces with hydrophilic and hydrophobic areas but no topography. The samples were characterized by a number of techniques including SEM and XPS. AFM tapping mode

images were taken with chemically modified tips and it was found that hydrophilic tips gave the best contrast.

93 .

2.6 References

(1) Binnig, G.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930-933.

(2) Bar, G.; Brandsch, R.; Whangbo, M. H. Surf. Sci. 1998, 41 h L802-L809.

(3) Takano, H.; Kenseth, J. R.; Wong, S. S.; O'Brien, J. C; Porter, M. D. Chem. Rev. 1999, 99, 2845-2890.

(4) Finot, M. O.; McDermott, M. T. J. Am. Chem. Soc. 1997, 779, 8564-8565.

(5) Tamayo, J.; Garcia, R. Langmmr 1996, 72, 4430-4435.

(6) Petzow, G.; Herrmann, M. High Performance Non-Oxide Ceramics li 2002, 102, 47-167.

(7) Senden, T. J.; Drummond, C. J. Colloids and Surfaces a-Physicochemical and

Engineering Aspects 1995, 94, 29-5 1

(8) Lee, G. U.; Chrisey, L. A.; Oferrall, C. E.; Pilloff, D. E.; Turner, N. H.;

Colton, R. J. Israel Journal of Chemistiy 1996, 36, 81-87.

(9) Petzow, G.; Herrmann, M. In High Performance Non-Oxide Ceramics li; Springer-Verlag: Berlin, 2002; Vol. 102, pp 47-167.

( 1 0) Riley, F. L. Journal of the American Ceramic Society 2000, 83, 245-265.

(11) Raider, S. I.; Flitsch, R.; Aboaf, J. A.; Pliskin, W. A. Journal of the Electrochemical Society 1976, 123, 560-565.

(12) Brow, R. K.; Pantano, C. G. Journal of the American Ceramic Society 1986, 69, 314-316.

(13) Busca, G.; Lorenzelli, V.; Baraton, M. I.; Quintard, P.; Marchand, R. Jofz/vw/ ofMolecular Structure 1986, 143, 525-528.

94 (14) Ogbuji, L. U. T.; Jayne, D. T. Journal of the Electrochemical Society 1993, 140, 759-766.

(15) Ogbuji, L.; Bryan, S. R. Journal of the American Ceramic Society' 1995, 78, 12^72- 1278.

(16) Ogbuji, L. Journal ofthe American Ceramic Society 1995, 78, 1279-1284.

( 1 7) Okada, K.; Fukuyama, K.; Kameshima, Y. Journal of the American Ceramic Society 1995, 78, 2021-2026.

( 1 8) Bergstrom, L.; Bostedt, E. Colloids and Surfaces 1990, 49, 1 83-1 97.

(19) Bergstrom, L. Colloids and Surfaces 1992, 69, 53-64.

(20) Chovelon, J. M.; Jaffrezicrenault, N.; Clechet, P.; Cros, Y.; Fombon, J. J.;

Baraton, M. I.; Quintard, P. Sensors and Actuators B-Chemical 1991, 4, 385-389.

(21) Harame, D. L.; Bousse, L. J.; Shott, J. D.; Meindl, J. D. leee Transactions on Electron Devices 1987, 34, 1700-1707.

(22) Frisbie, C. D.; Rozsnyai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071-2074.

(23) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo,

R. G. J. Am. Chem. Soc. 1989, 7/7, 321-335.

(24) Ulman, A. Chemical Reviews 1996, 96, 1533-1554.

(25) Barrat, A.; Silberzan, P.; Bourdieu, L.; Chatenay, D. Europhys. Lett. 1992, 20, 633-638.

(26) Nakagawa, T.; Ogawa, K.; Kurumizawa, T.; Ozaki, S. Jpn. J. Appl. Phys. Part 2 - Lett. 1993, 32, L294-L296.

(27) Sagiv, J. Journal of the American Chemical Society 1980, 102, 92-98.

95 (28) Maoz, R.; Sagiv, J. Journal of Colloid and Interface Science 1984, 100, 465-496

(29) Wasserman, S. R.; Tao, Y. T.; Whitesides, G. M. Langmuir 1989, 5, 1074-1087

(30) Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 15, 3759-3766.

(31) Sung, M. M.; Kluth, G. J.; Maboudian, R. Journal of Vacuum Science & Technology a-Vacuum Surfaces and Films 1999, 17, 540-544.

(32) Tripp, C. P.; Hair, M. L. Langmuir 1995, 11, 1215-1219.

(33) Tripp, C. P.; Hair, M. L. Langmuir 1995, 11, 149-155,

(34) Rye, R. R.; Nelson, G. C; Dugger, M. T. Langmuir 1997, 13, 2965-2972.

(35) Butt, H. J.; Jaschke, M.; Ducker, W. Bioelectrochem. Bioenerg. 1995, 38,\9\- 201.

(36) Noy, A.; Vezenov, D. V.; Lieber, C. M. Anmi. Rev. Mater. Sci. 1997, 27, 381- 421.

(37) Colton, R. J. J. Vac. Sci. Technol. B 2004, 22, 1609-1635.

(38) Cappella, B.; Dietler, G. Surf Sci. Rep. 1999, 34, 1-104.

(39) Bumham, N. A.; Chen, X.; Hodges, C. S.; Matei, G. A.; Thoreson, E. J.;

Roberts, C. J.; Davies, M. C; Tendler, S. J. B. Nanotechnology 2003, 14, 1-6.

(40) Grigg, D. A.; Russell, P. E.; Griffith, J. E. J. Vac. Sci. Technol. A-Vac. Surf Films 1992, 10, 680-683.

(41) Weisenhorn, A. L.; Hansma, P. K.; Albrechl, T. R.; Quate, C. F. Appl. Phys. Lett. 1989,54, 2651-2653.

96 (42) Drake, B.; Prater. C. B.; Weisenhom, A. L.; Gould, S. A. C; Albrecht, T. R.; Quate, C. F.; Cannell, D. S.; Hansma, H. G.; Hansma, P. K. Science 1989, 243, 1586-1589.

(43) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 35i, 239-241.

(44) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831-1836.

(45) Noy, A.; Frisbie, C. D.; Rozsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 7/7, 7943-7951.

(46) Bumham, N. A.; Colton, R. J. J. Vac. Sci. Technol. A-Vac. Surf. Films 1989, 7, 2906-2913.

(47) Ovemey, R. M.; Meyer, E.; Frommer, J.; Brodbeck, D.; Luthi, R.; Howald,

L.; Guntherodt, H. J.; Fujihira, M.; Takano, H.; Gotoh, Y. Nature 1992, 359, 133-135.

(48) Ovemey, R. M.; Meyer, E.; Frommer, J.; Guntherodt, H. J.; Fujihira, M.; Takano, H.; Gotoh, Y. Langmuir 1994, JO, 1281-1286.

(49) Green, J. B. D.; McDemiott, M. T.; Porter, M. D.; Siperko, L. M. J. Phys. Chem. 1995, 99, 10960-10965.

(50) Bar, G.; Brandsch, R.; Whangbo, M. H. Langmuir 1998, 14, 7343-7348.

(51) Noy, A.; Sanders, C. H.; Vezenov, D. V.; Wong, S. S.; Lieber, C. M.

Langmuir 1998, 14, 1508-151 1.

(52) Sasaki, K.; Koike, Y.; Azehara, H.; Hokari, H.; Fujihira, M. Appl Phys. A- Mater. Sci. Process. 1998, 66, S1275-S1277.

(53) Basnar, B.; Friedbacher, G.; Brunner, H.; Valiant, T.; Mayer, U.; Hoffmann, H. Appl. Surf. Sci. 2001, 171, 213-225.

(54) Raghavan, D.; VanLandingham, M.; Gu, X.; Nguyen, T. Langmuir 2000, 16, 9448-9459.

97 (55) Brandsch, R.; Bar, G.; Whangbo, M. H. Langmuir 1997, 13, 6349-6353.

(56) Kim, H. C; Jia, X. Q.; Stafford, C. M.; Kim, D. H.; McCarthy, T. J.;

Tuominen, M.; Hawker, C. J.; Russell, T. P. Adv. Mater. 2001, 13, 795-797.

(57) Jia, X. Q.; McCarthy, T. J. Langmuir 2003, 19, 2449-2457.

(58) Oner, D.; McCarthy, T. J. Langmuir 2000, 16, 1111-11%!.

(59) Chan, C. M. Polymer Surface Modification and Characterization; Hansen New York, 1994.

(60) SEM courtesy of Aurora Nanodevices, Inc.

(61) Hayashi, K.; Sugimura, H.; Takai, O. Appl Surf Sci. 2002, 188, 513-51 8.

(62) Thundat, T.; Zheng, X. Y.; Chen, G. Y.; Sharp, S. L.; Warmack, R. J.;

Schowalter, L. J. Appl. Phys. Lett. 1993, 63, 2150-2152.

(63) Jia, X. Q.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society 2001, 222, U320-U320.

(64) Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 75, 7238-7243.

(65) Stafford, C. M.; Fadeev, A. Y.; Russell, T. P.; McCarthy, T. J. Langmuir 2001, 77,6547-6552.

(66) Kumar, A.; Biebuyck, H. A.; Abbott, N. L.; Whitesides, G. M. Journal of the American Chemical Society 1992, 114, 9188-9189.

(67) Kumar, A.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1994, 10, 1498-1511.

(68) Biebuyck, H. A.; Larsen, N. B.; Delamarche, E.; Michel, B. Ihm Journal of

Research and Development 1997, 41, 1 59- 1 70.

98 Love, J. C; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M.

Chemical Reviews 2005, 105. 1 103-1 169.

99 CHAPTER 3

TWO DIMENSIONAL DROPLET FLUIDICS BASED ON DIFFERENTIAL LYOPHOBICITY

3.1 Introduction

The motion of liquids on surfaces is important in a wide variety of applications

from car wax to microfluidic devices. As will be discussed, in order to induce motion

of a droplet sitting on a surface, an external force needs to be applied to overcome the

surface tension forces. A range of forces have been investigated, but what is not often

mentioned is the role and importance of surface chemistry. One can reduce the applied

force by carefully selecting the chemical functionality of the sample surface.

Surfaces with patterned chemistry were prepared by a combination of UV

lithography and vapor phase silanization. A hydrophobic silane with a low hysteresis

was used to prepare "roads" on which the liquid would move very easily. A second

silane with a higher advancing contact angle was used to make "curbs" that confined and directed the motion of the liquid. A variety of chemical patterns were prepared to emphasize the importance of both advancing contact angle difference and low hysteresis on the motion of liquids on a planar surface.

100 3.2 Background

3.2. 1 Contact Angle, Wettability, and Hysteresis

The shape and motion of a liquid drop on a surface has been the subject of

hundreds of years of research and thousands of publications. It continues to be an important area of research although, as will be discussed, many basic concepts have not yet been fully explained and others are often misunderstood.

Thomas Young is credited with the first thermodynamic explanation of the shape of a liquid drop on a surface.' He reasoned that the equilibrium shape was a result of the minimization of the surface free energy of the system at the three phase

contact line Figure 3.1. The equation relating the static contact angle, 0y, to the

interfacial free energies is

YlvCOsG = (Ysv-Ysi) (3.1)

where yiv, ysv, and ysi are the liquid/vapor, solid/vapor, and solid/liquid interfacial tensions respectively. The derivation assumes a smooth, homogeneous, isotropic, and non-deformable surface.

Figure 3.1 . Static contact angle of a drop of liquid on a surface.

101 The Young equation predicts one contact angle for a static liquid drop on an

ideal surface, however, it is observed that there can be many different static contact

angles on real surfaces.^ The obvious reason is that real surfaces do not conform to the

assumptions that the equation is based on, i.e., they are not smooth or completely

homogeneous. The observed contact angle varies between two extremes, the advancing

and the receding contact angle, which are measured while increasing and decreasing the

drop volume respectively. The advancing angle is always greater than or equal to the

receding. The difference between the two is called the hysteresis and will be discussed

in more detail later. The effects of surface defects on the static contact angle will be

considered first.

Cassie investigated contact angles on surfaces with chemical heterogeneity.'' He showed that the contact angle could be determined by summing the contributions from the various components. For a surface with n different chemical functionalities the contact angle could be calculated using Equation 3.2,

cos^.=Xy:cos^, X./:=l (3.2)

where fj is the fraction of the surface covered by functionality i, 0i is the contact angle

of functionality i on a flat surface, and 0c is the contact angle on the mixed surface. The

equation is best applied to patchy surfaces that have domains smaller than the size of the drop but larger than molecular dimensions.

102 More recently, Israelachvili and Gee modified the Cassie equation to take into account samples that were mixed on the molecular level:

(l + cos^J^=X./;(l + cos,)^ X/=l (3.3)

where the parameters are the same as defined in the Equation 3.2 above.

Surface roughness also changes the observed contact angle. Wenzel modified

the Young equation to include a roughness factor, r, defined as the actual surface area divided by the projected surface area."^ A flat surface having a contact angle of 0y

would have a contact angle of 0w when roughened by a factor of r.

cos = r cos 0^. (3.4)

This equation predicts that roughening a smooth surface with a contact angle less than

90'^ will lower the contact angle, but for a surface with a contact angle greater than 90° the contact angle will increase. The equation works well for low, uniform roughness, but fails on surfaces with high roughness ratios.

103 Figure 3.2. Schematic diagram and photograph of (a) Cassie-Baxter type wetting (b)

Wenzel type wetting. Reprinted with permission from reference 6. Copyright 2003 American Chemical Society.

At high roughness values, the liquid does not wet the entire surface but instead

sits atop surface asperities (Figure 3.2a).^ The surface can thus be considered a

composite with one component being air. The Cassie-Baxter equation describes such a

surface where fi and fo are the fraction of the solid surface and air under the drop respectively.^'^

cos^,, =./;cos^, yi+/2=l (3.5)

All the equations discussed thus far describe how surface chemical heterogeneity and roughness affect the equilibrium contact angle. However, as was

already pointed out, a single static contact angle is rarely observed on real surfaces.

The reason is that surface heterogeneity and roughness also hinder the motion of the three phase contact line. The contact line pins on surface defects and many metastable, local minima in free energy are accessible. ' The magnitude of the energy barrier between states, the vibrational energy of the droplet and how the drop was placed on the

surface affect the observed contact angle value. If the energy barrier is low and the

104 vibrational energy is high, the drop is more likely to find the global minimum, but if the

vibrational energy is low and the barrier high the drop shape will only correspond to a local minimum. Therefore, static contact angles give minimal information about the surface.

Dynamic contact angle measurement provides more reproducible values. The

advancing angle is measured as the volume of a liquid drop is increased and the

receding angle while the drop size is decreased. For an ideal, homogeneous surface, the advancing, receding and static angles would all be equal, but for a real surface 0a>

0static> 0R- The difference between the advancing and receding angles is called the contact angle hysteresis. Chemical heterogeneity, surface roughness, surface reorganization, and surface swelling by the probe fluid have been shown to cause hysteresis.

Wettability is often studied using liquid drops moving down an inclined surface.''" '^' The drop cannot move until the downhill edge reaches the advancing contact angle and the trailing edge reaches the receding angle. Various researchers have derived and revised an equation relating the surface tension forces holding the drop in

''^ place to the angle the surface must be inclined to induce drop motion: '^

- mg sin a = kwyi^. (cos 6,^ cos 0^ ) (3.6)

where m is the drop mass, g is the acceleration of gravity, a is the tilt angle, k is a

constant, w is the droplet contact diameter, and yiv is the liquid vapor surface tension.

105 To decrease the surface forces, one can either decrease the hquid/solid contact width or decrease the hysteresis.

While various researchers have studied drop motion on inclined surfaces, they

'^"'^'^"^'"^ are usually interested in the drop mobility' and shape on a homogeneous

surface. What is not often discussed is changing or confining the drop motion using chemically defined surface paths. In this project the focus was on influencing not only the drop mobility but also the direction by preparing samples with patterned surface chemistry.

3.2.2 Liquid Mobility in Micro and Droplet Fluidic Devices

The field of microfluidics has received significant attention in recent years and

is covered in several reviews." ' Typical microfluidics devices utilize millimeter to micron size capillaries or channels to confine fluid flow. The small-scale experiments that can be performed in these devices offer several advantages to the typical large-scale

techniques. The amount of reagents is reduced, multiple analyses can be perfonned on a single, portable device, and reaction kinetics often increase due to higher reagent concentrations.

Much of microfluidics research is focused on materials issues that arise upon scaling down the size of the features.^"* Surface tension forces become increasingly important and make fluid transport more demanding on the pumping system. The small capillaries can also become blocked by contaminants or air bubbles and can be hard to clean. The fabrication of complex three-dimensional structures can also be complicated.

106 Confinement of fluid flow based on surface chemistry differences has been demonstrated in microfluidic devices.""^ Channels can be made hydrophilic while hydrophobic patches act as valves, stopping the flow until a pressure increase

overcomes the surface tension forces. Zhao, et al. patterned the surface of microchannels with hydrophilic and hydrophobic paths."^^"^^ They showed that fluid flow could be confined to the hydrophilic pathways as long as a critical pressure was

not exceeded. Oh^° and Lam, et al."" removed the channels and simply patterned two surfaces with hydrophilic and hydrophobic stripes and stacked them together with a small spacer in between. The fluid only wet the hydrophilic regions and spanned the gap between the two surfaces because of capillary force.

In another method of fluid manipulation called drop fluidics the liquid is moved

^"'"^^ on an open surface. The two dimensional system eliminates many of the problems related to fluid confinement in microchannels or small gaps. Surface forces are less of a hindrance to fluid motion and the open arrangement allows access to reagents for easy manipulation and cleanup. Obviously evaporative losses need to be considered if long analyses are to be performed. The liquid can be moved across the surface by various

methods such as electrostatics^"^, electrowetting''^ thermal gradients^^""^'', chemical

gradients"^^"^^'^', and surface acoustic waves'''""^"'.

The chemistry of the surface plays an important role. The path of the drops can be defined using two different surface chemistries; one as the road and the other as the

curb. The drop follows the higher surface energy (lower contact angle) road and is

confined by the lower surface energy (higher contact angle) curbs. However, what is

rarely discussed is the importance of contact angle hysteresis. A drop of liquid on a

107 surface with no hysteresis would require only a very small force to induce motion so less work would have to be done by the system.

3.3 Experimental

3.3.1 Materials

Trimethylchlorosilane (TMCS), dimethyldichlorosilane (DMDCS), tridecafluoro-

1,1,2,2-tetrahydrooctyldimethylchlorosilane (FDCS) and 3- aminopropyldimethylchlorosilane were purchased from Gelest and used as received.

Shipley Megaposit SPR220-7 positive photoresist, Rohm and Haas Microposit MF-319, a 0.26 N tetramethylammonium hydroxide developing solution, and Shipley Microposit

EC- 1 1 ethyl lactate solvent were used as received. Silicon wafers were purchased from

International Wafer Service (<100> orientation, P/B doped, resistivity from 20.0 to 40

Q cm). House-purified water (reverse osmosis) was further purified using a Millipore

Milli-0 system { 1 8.2 MQ cm).

3.3.2 Experimental Procedures

3.3.2.1 UV Pattern Fonnation

Silicon wafers were modified with silancs via a vapor phase reaction. Five minute exposure to oxygen plasma (Harrick Scientific Corp. plasma cleaner model

PDC-001 ) was used to remove contaminants from the surfaces. The wafers were then suspended in a Schlenk tube above 0.5 mL of the selected silane. The vessel was sealed

108 and heated to 65-70 °C in an oil bath for 3 days. After the reaction the samples were

removed and carefully dipped sequentially into toluene, ethanol, and water to remove

physisorbed silane.

A positive photoresist was then spin coated onto the modified wafers at 6000

rpm for 60 seconds. The resulting thickness was ~4-5 ^im. The samples were heated

from room temperature to 125 "C over ~3 minutes and held at that temperature for

another 2 minutes to remove any residual solvent. The photoresist was exposed to 365 nm UV radiation at 350 W with an intensity of -10 MW/cm' through a mask having the desired features. The masks were made by printing the preferred pattern in black ink onto a plastic overhead transparency. Simple patterns were prepared using aluminum foil as a mask. The exposed regions were selectively dissolved by soaking in a 0.26 N

tetramethylammonium hydroxide solution for 1 minute. The sample was then plasma cleaned for 2 minutes to remove any remaining resist in the exposed areas and also to etch through the initial silane monolayer. A second silane was applied to the exposed regions using the same method described previously, except the time was usually

shortened to 1 day. Finally the remaining resist was removed using ethyl lactate as a solvent.

3. 3.2.2 Characterization

The two main characterization techniques were XPS and contact angle. A

Physical Electronics Quantum 2000 Scanning ESCA Microprobe was used to obtain

XPS data. Contact angle data was obtain using a goniometer and computer-controlled video capture system custom built by members of Professor Crosby's research group in

109 the Polymer Science and Engineering Department. All contact angle measurements are

averages of 3-5 spots per sample and the standard deviation was typically ~2°.

3.4 Results and Discussion

3.4.1 Patterned Silane Surfaces

The requirements for the two silanes were that one had a low hysteresis so that the liquid droplet could move easily and the other had a higher advancing contact angle

than the first. The silanes chosen were TMCS and DMDCS for the low hysteresis areas and FDCS for the barriers. Silicon wafers were modified with the three different silanes as previously described. All three of the surfaces were hydrophobic with the following

contact angles: 1 167101° for FDCS, 1047100° for DMDCS, and 92786° for TMCS.

TMCS and DMDCS had reasonably low hysteresis, with some DMDCS samples showing nearly no hysteresis. The thickness of the silane layers was slightly different

for each. TMCS had the thinnest layer, 3 A, which was expected because it can only form a monolayer and the initial silane length was the shortest. FDCS can also only

form a monolayer, but the chain length is slightly longer leading to a thicker silane layer

of ~10 A. DMDCS can form two Si-0 bonds so it can vertically polymerize from the surface and had the thickest silane layer, -30 A.

A variety of methods were explored to prepare surfaces with patterned silane

chemistry. An aqueous sodium hydroxide solution was used to etch through the first silane layer to re-expose the silica for fiirther modification with another silane.

110 .

However, it proved difficult to control the placement of the solution as it preferred to

remain spherical due to surface tension, but the interest lied in making lines.

Electron beam lithography was also attempted. Research has shown that silane

monolayers can be precisely removed using an electron beam.^^""*^' The exposed silica

can then be modified with another silane. The obvious drawbacks are the expense and

time needed to prepare each sample. FDCS samples were used as a test. The beam

current and area dosage were varied over a matrix of 1x1 |im and 30x30 \im squares to

determine the best conditions. The samples were then modified with 3-

aminopropyldimethylchlorosilane for one day. The amine groups of different samples

were then labeled in various ways in an attempt to visualize the pattern. A solution of

negatively charged gold was used to selectively bind to the amine groups and

fluorescein isothiocyanate was used to fluorescently label the amine regions. Neither

labeling method allowed visualization of the patterns. AFM showed no selective deposition of the gold particles and confocal microscopy did not reveal any regions that were selectively labeled.

A new technique was developed to prepare surfaces with patterned chemistry by combining UV lithography and vapor phase silanization. The various processes are shown in Figure 3.3 and were described in detail above. The first step was to modify the silicon wafer with one of the silanes, followed by spin coating with a positive

photoresist. It was found that the photoresist dewet FDCS so the first monolayer was either TMCS or DMDCS. XPS analysis of the resist layer can be found in Table 3.1

The most prevalent element is carbon from the polymer resin. No silicon from the

substrate can be seen because the resist is so thick, but a small amount of silicone

111 surfactant contributes to the small Si concentration. The oxygen is attributed to hydroxyl and carbonyl functionalities in the resin mixture. Sulphoxide groups are commonly found on components in the resin mixture.

hydrophob ic silicon silane

UV positive I photoresist t iiiyiiiiiii •mask

(1)develop resist

1 (2)plasma etch silane

(1) 2nd silane

(2) remove resist

Figure 3.3. Schematic diagram of the UV lithography process used to prepare surfaces with patterned silane chemistry.

Table 3.1. XPS analysis of an alkylsilane modified silicon wafer after being coated with a resist layer, exposed to UV radiation, developed, and plasma cleaned.

Atomic composition take-off angle data) surface

C Si 0 S

resist layer 81.8 0.3 17.3 0.7

developed'' 30.1 33.4 36.5

plasma cleaned^ 0.7 48.6 50.7

^ 1 min exposure to UV followed by 60 s rinse in developing solution; 2 min oxygen plasma.

112 The resist was then exposed to UV radiation through a mask with the desired

pattern. The exposure time was important as too little exposure caused incomplete

dissolution of the resist and too much exposure led to broadening of the interface and

poor definition of the features. After many attempts, 60 seconds was found to be the

best exposure time. The resist was developed for 60 seconds to remove the regions

exposed to UV. The XPS data in Table 3. 1 shows that most, if not all of the resist was

removed as indicated by the increase in silicon concentration from the substrate.

Oxygen plasma was used to remove the remaining resist and etch through the

first silane layer down to the silica. Time was important again, as too little oxidation

did not remove all the contamination and after too much time the edges of the resist

began to be etched away. Two minutes was found to be optimal. Table 3.1 shows that

nearly all the carbon was removed. The exposed silica was then modified with a second

silane, usually FDCS. Finally, the remaining resist was dissolved leaving a sample with

patterned silane chemistry. The resist was sometimes difficult to ftiUy remove after the

reaction as evidenced by a decrease in the receding contact angle and increased

hysteresis.

After this technique was developed, a closer inspection of the literature revealed procedures that were similar. Ondarcuhu and Veyssie made patterned silane surfaces using UV lithography."*^ They coated an alkyl silane modified silicon wafer with a

positive photoresist and exposed it to UV radiation through a mask. The sample was

developed in NaOH followed by UV/ozone treatment to oxidize the exposed silane.

They were interested in hydrophilic and hydrophobic areas and did not react the

exposed silanol groups with a second silane. Palladre, et al. used a PMMA resist layer.

113 e-beam lithography, and vapor phase silane modification to prepare patterned silane surfaces.""^ The most similar procedure was recently published by Harrant, et al"^^ and basically used nearly the exact same process that was developed independently in our laboratory.

3.4.2 Characterization by XPS and Contact Angle

The patterned surfaces were characterized by XPS and contact angle. Figure 3.4 shows an XPS line scan measuring fluorine concentration across a surface that was patterned with DMDCS on the left side and FDCS on the right. A 20 \im beam size was

used. At the boundary between the two sections there is an abrupt increase in the fluorine concentration showing that the interface is reasonably sharp. The DMDCS side had no fluorine and Figure 3.5a shows the Cls region as a single peak as expected.

Figure 3.5b shows the Cls region for FDCS which shows three distinct peaks from the three types of carbon moieties in the silane chain.

It is well known that liquids, when placed at a chemical interface, move spontaneously from areas of lower surface energy (higher contact angle) to higher surface energy (lower contact angle).^^'^*^ This phenomenon has been utilized to make

liquids move across samples with a surface energy gradient."^*^'^' Conversely, low

surface energy areas can be used to confine and direct the motion of a liquid. There is a barrier for a sliding drop to move on a tilted surface from a region of lower contact angle to one with a higher contact angle, even if both are hydrophobic. The ability of the line to act as a barrier depends on the relative difference in the advancing contact angle with the background. The optimum combination of chemistry would yield a low

1 14 contact angle, low hysteresis background to allow ease of drop motion and a high advancing contact angle barrier.

Figure 3.4. XPS line scan of fluorine concentration on a silicon wafer patterned on one side with DMDCS and on the other with FDCS.

3000 n 1600- 2500- 1400- / 1 ^

2000- / 1 -S >« >« 1200- +j 1 '« '<« / c 1500- / c 1000- 0) / 0)

/ \ — 1000- 800-

500- 600-

< 1 1 1 1 1 1 1 1 1 1 1 1 1 t 1 I 298 296 294 292 290 288 286 284 282 280 278 298 296 294 292 290 288 286 284 282 280 278

Binding Energy (eV) Binding Energy (eV)

a b

Figure 3.5. XPS multiplex spectra of the Cls region on the (a) DMDCS side and (b) FDCS side of a patterned sample similar to that shown in Figure 3.4.

115 Figure 3.6a shows a water droplet on a surface having DMDCS on one side and

FDCS on the other. The contact angles are very similar to those on homogeneous surfaces. Although the receding contact angle for DMDCS is a bit low, the hysteresis remained low. A 2 )iL droplet in Figure 3.6b is shown pinned at the interface between

the DMDCS surface that it is wetting and the FDCS side which has a higher contact angle. The droplet will remained pinned until its advancing angle reaches that of the

FDCS surface. However, since the surface is already at the steepest angle possible, the only way for the advancing contact angle to increase would be to increase the volume of

the drop. Figure 3.6c is the same DMDCS/FDCS sample. As the sample is tilted, the drop remains pinned to, but slides along the interface.

Two regions of the type shown in Figure 3.6 are not required to confine a droplet to a line; a more hydrophobic line of sufficient width on a less hydrophobic

background is adequate to direct droplet movement on a surface. The modified silicon

wafers still looked hke mirrors after the modification reaction (Figure 3.7a). An easy way to view the pattern was to dip the samples in ethanol or hexane which preferentially wetted the alkyl silane and not the FDCS (Figures 3.7b,c).

Figures 3.7b,c show examples of the patterned surfaces that were prepared. The background is TMCS and the lines are FDCS. Water droplets moved easily on the

TMCS when the sample was tilted until contact was made with a FDCS line. The droplet then stayed pinned until either the advancing contact angle of the fluorinated line was reached by tilting the surface to a higher angle or the sample was rotated causing the drop to slide along the line.

116 jt% ft. 103°/98° I 111160/101060/1010

Figure 3.6. Silicon wafer chemically patterned with sections of DMDCS and FDCS: (a) Water contact angle measurements on the DMDCS (left) and FDCS (right) sections, (b) Same surface held vertical with a 2 )iL water droplet on the DMDCS side pinned at the FDCS interface, (c) Water droplet sliding along the DMDCS/FDCS interface.

Figure 3.7. (a) Silicon wafer after chemical pattern fonnation, (b, c) TMCS modified silicon wafers with FDCS lines after immersion in ethanol.

117 The pattern in Figure 3.7b can be used to make a water droplet move uphill. A

water droplet placed near the outside of the spiral follows along the fluoroalkyl line as

the surface is held nearly vertically and rotated clockwise. The droplet moves vertically

from the border of the wafer to the center of the invisible spiral. A simple maze pattern

is shown in Figure 3.7c. A water drop can be used to solve this "invisible" puzzle.

Figure 3.8. DMDCS modified silicon wafer with FDCS lines. Water droplets placed at each invisible vertex are sequentially combined as the sample is slightly tilted and rotated clockwise.

Figure 3.8 shows a DMDCS surface patterned with V-shaped FDCS lines. A ~2

[iL water droplet was placed at three of the invisible vertices with one 20 \iL at the fourth. When the sample was tilted slightly and rotated clockwise, the larger droplet slid along the FDCS line and into the next vertex collecting the smaller water droplet. The

118 other droplets did not move because the force of gravity was not enough for them to overcome the slight hysteresis present in the sample as explained previously (Equation

3.6). Continued rotation allowed all the droplets to be combined in succession. This pattern permits confining the droplet to points for defined periods of time. Various reagents or indicators could be covalently attached to each vertex and a drop could be sequentially analyzed by each. Increasing the number of vertices would allow many more analyses of a single drop.

Gravity is not the only force that can be used to move liquids on a surface. The centrifugal force of a spinning object has been used in microfluidics to move liquids

without using a pumping system. The liquid is placed near the center of the device and,

as it is rotated, the fluid moves radially outward along the microfluidic channels. The progress of the liquid can be impeded using a passive valve. A hydrophobic barrier can

stop the motion of the liquid until enough force is applied to reach the advancing

'^'^ contact angle of the chemical impediment.^ ' Another system uses capillary pressure

^"'"^^ to hinder the liquid motion. The rotational speed can be increased, providing a high enough force for the liquid to overcome the passive barriers.

A two-dimensional version of a similar system was invesfigated. A DMDCS surface was patterned with FDCS lines as shown in Figure 3.9a. The sample pictured had four pairs of angled hydrophobic lines that directed a drop of water to the narrow opening between them. The gaps varied from 1-2 mm for the sample in Figure 3.9a and samples with gaps up to 4 mm were prepared. The sample was placed on a spin coater and a 10 )iL drop of water was placed in between each pair of FDCS lines. The rotations per minute were slowly increased until a drop slid off the surface. One would

119 expect that the water would pass more easily through the larger gaps at lower RPMs.

Qualitative determination showed that was generally the case. The first drop would slide off the surface at -3000 RPM, but the order was not always the same, probably due to slight imperfections on the surface. A quantitative determination of the effects of such parameters as drop size, gap opening, RPMs, and surface chemical difference was not attempted, but might prove to be an interesting investigation.

a b

Figure 3.9. (a) Pattern prepared with FDCS lines on a DMDCS background that directs water droplets outward through the gap under the influence of centrifugal force, (b) Pattern of hydrophobic lines that could be used to combine liquid droplets.

This technique is likely more general than the few experiments described here.

Patterned surfaces such as that shown in Figure 3.9b could be used to make multiple

droplets converge and mix by changing the surface tilt angle and the gap between the

fluorinated lines. Two converging lines pin a droplet by distorting it from a section of a

sphere. The tilt angle at which the droplet is released is controlled by the sizes of the droplet and opening and the angle between the lines. The motion of liquids other than water can also be controlled by tailoring the surface chemistry.

120 3.5 Conclusions

The ability to confine and direct the motion of liquid under the force of gravity using only patterned surface lyophobicity was demonstrated. A new method was developed to prepare such patterned surfaces using a combination of UV lithography and vapor phase silane reactions. Either DMDCS or TMCS was used to prepare a low hysteresis matrix on which liquid moved easily and FDCS lines, which had a higher advancing contact angle, were used to direct the motion of the droplets. The originality of this project was that both the matrix and lines were hydrophobic, but the low hysteresis of the matrix allowed the liquids to be confined by only a slight difference in contact angle.

A variety of patterned surfaces were prepared to demonstrate the applications of this technique. XPS line scans showed a reasonably sharp interface between the two chemistries and the Cls multiplex spectra revealed the presence of each silane in the proper region as defined by the lithographic process. Water droplets placed on the patterned surfaces moved easily on the low hysteresis matrix, but pinned at the FDCS lines. When the samples were rotated the droplets slid along the hydrophobic barriers.

It was shown that multiple droplets could be confined to certain areas of the same sample and then sequentially manipulated using only gravity. Centrifugal force was also used to induce drop motion. Many other patterns could be imagined, but the

important point is that a decrease in the hysteresis reduces the force needed to induce drop motion and also lowers the barriers that are needed to confine the droplets.

121 3.6 References

(1) Young, T. Philosophical Transactions of the Royal Society ofLondon 1805, 95, 65-87.

(2) Rayleigh, L. Philosophical Magazine 1890, 30, 386-400.

(3) Cassie, A. B. D. Discussions of the Faraday Society 1948, i, 11-16.

(4) Israelachvili, J. N.; Gee, M. L. Langmiiir 1989, 5, 288-289.

(5) Wenzel, R. N. Industrial and Engineering Chemistry 1936, 28, 988-994.

(6) He, B.; Patankar, N. A.; Lee, J. Langmuir 2003, 79, 4999-5003.

(7) Cassie, A. B. D.; Baxter, S. Transactions of the Faraday Society 1944, 40, 546- 551.

(8) Cassie, A. B. D.; Baxter, S. Nature 1945, 755, 21-22.

(9) Shuttleworth, R.; Bailey, G. L. J. Transactions of the Faraday Society 1948, 44, 16-22.

( 1 0) Bikerman, J. J. Journal ofPhysical and Colloid Chemistiy 1950, 54, 653-658.

(11) Good, R. J. J. Am. Chem. Soc. 1952, 74, 5041-5042.

(12) Johnson, R. E.; Dettre, R. H. J. Phys. Chem. 1964, 68, 1744-&.

(13) Johnson, R. E.; Dettre, R. H. In Surface and Colloid Science; Matijevic, E., Ed.; Wiley: New York, 1969; Vol. 2, pp 85-153.

(14) Garbassi, F.; Morra, M.; Occhiello, E. Polymer Surfaces: From Physics to Technology; Wiley: New York, 1994.

122 (15) Macdougall, G.; Ockxent, C. Proceedings of the Royal Society ofLondon. Series A, Mathematical and Physical Sciences 1942, 180, 151-173.

( 1 6) Kawasaki, K. Journal of Colloid Science 1960, 75, 402-407.

( 1 7) Furmidge, C. G. Journal of Colloid Science 1962, 1 7, 309-324.

( 1 8) Dussan, E. B.; Chow, R. T. P. Journal of Fluid Mechanics 1983, 137, 1 -29.

(19) ElSherbini, A. I.; Jacobi, A. M. Journal of Colloid and Interface Science 2004, 273, 556-565.

(20) Suzuki, S.; Nakajima, A.; Kameshima, Y.; Okada, K. Surface Science 2004, 557, L163-L168.

(21) Kim, H. Y.; Lee, H. J.; Kang, B. H. Journal of Colloid and Interface Science 2002, 247, 372-380.

(22) ElSherbini, A. I.; Jacobi, A. M. Journal of Colloid and Interface Science 2004, 273, 566-575.

(23) Milinazzo, P.; Shinbrot, M. Journal of Colloid and Interface Science 1988, 121, 254-264.

(24) Stone, H. A.; Kim, S. Aiche Journal 2m\, 47, 1250-1254.

(25) Reyes, D. R.; lossifidis, D.; Auroux, P. A.; Manz, A. Analytical Chemistry 2002, 74, 2623-2636.

(26) Stone, H. A.; Stroock, A. D.; Ajdari, A. Annual Review ofFluid Mechanics 2004, i6, 381-411.

(27) Zhao, B.; Moore, J. S.; Beebe, D. J. Science 2001, 291, 1023-1026.

(28) Zhao, B.; Moore, J. S.; Beebe, D. J. Analytical Chemistiy 2002, 74, 4259-4268.

123 .

(29) Zhao, B.; Viemes, N. O. L.; Moore, J. S.; Beebe, D. J. Journal of the American Chemical Society 2002, 124, 5284-5285.

(30) Oh, C: United States Patent 5,904,824, 1999.

(31) Lam, P.; Wynne, K. J.; Wnek,G. E.Ifl«^ww/r 2002, 7(^,948-951.

(32) Washizu, M. leee Transactions on Industry^ Applications 1998, 34, 1 32-131

(33) Lehto, A.; Kojola, H.; Lovgren, T.; Lonnbreg, H.: World Patent 99/54730, 1999.

(34) Torkkeli, A.; Ph.D. Dissertation, Helsinki University of Technology, 2003.

(35) Cho, S. K.; Moon, H. J.; Kim, C. J. Journal ofMicroelectromechanical Systems 2003, 72, 70-80.

(36) Brochard, F. Langmuir 1989, 5, 432-438.

(37) Darhuber, A. A.; Davis, J. M.; Troian, S. M.; Reisner, W. W. P/?v.s/c^ o/F/w/t/^ 2003, 75, 1295-1304.

(38) Darhuber, A. A.; Valentino, J. P.; Davis, J. M.; Troian, S. M.; Wagner, S. Applied Physics Letters 2003, 82, 657-659.

(39) Darhuber, A. A.; Valentino, J. P.; Troian, S. M.; Wagner, S. Journal of Microelectromechanical Systems 2003, 72, 873-879.

(40) Daniel, S.; Chaudhury, M. K.; Chen, J. C. Science 2001, 297, 633-636.

(4 1 ) Chaudhury, M. K.; Whitesides, G. M. Science 1992, 256, 1539-1541.

(42) Wixforth, A. Superlattices and Microstructures 2003, 33, 389-396.

(43) Wixforth, A.: United States Patent 6,777,245, 2004.

124 .

(44) Lercel, M. J.; Tiberio, R. C; Chapman, P. F.; Craighead, H. G.; Sheen, C. W.; Parikh, A. N.; AUara, D. L. Journal of Vacuum Science & Technology- B 1993, 11, 2823-2828.

(45) Carr, D. W.; Lercel, M. J.; Whelan, C. S.; Craighead, H. G.; Seshadri, K.; Allara, D. L. Journal of Vacuum Science & Technology a-Vacuum Surfaces and F/7m.v 1997, 75, 1446-1450.

(46) Krupke, R.; Malik, S.; Weber, H. B.; Hampe, O.; Kappes, M. M.; von

Lohneysen, H. Nano Letters 2002, 2, 1 161-1 164.

(47) Ondarcuhu, T. Journal De Physique li 1995, 5, 227-24 1

(48) Pallandre, A.; Glinel, K.; Jonas, A. M.; Nysten, B. Nano Letters 2004, 4, 365- 371.

(49) Harant, A. W.; Wliipple, S. G.; Douglas, K.; Bowman, C. N. Journal of Vacuum Science & Technology B 2005, 2i, 354-358.

(50) Ondarcuhu, T.; Veyssie, M. Journal De Physique li 1991, /, 75-85.

(51) Zoval, J. v.; Madou, M. J. Proceedings of the IEEE 2004, 92, 140- 1 53.

(52) Gyros Microlabs. Microfluidics at Work. http://www.gyros.com/technology/technology_microlab_microfluidics.html (accessed April 2006).

(53) Duffy, D. C; Gillis, H. L.; Lin, J.; Sheppard, N. F.; Kellogg, G. J. Analytical Chemistiy 1999, 77, 4669-4678.

(54) Lai, S.; Wang, S. N.; Luo, J.; Lee, L. J.; Yang, S. T.; Madou, M. J. Analytical Chemistiy 2004, 76, 1832-1837.

125 CHAPTER 4

CONDENSATION ON ULTRAHYDROPHOBIC SURFACES AND ITS EFFECT ON WATER DROPLET MOBILITY

4.1 Introduction

Many researchers have recently become interested in ultrahydrophobic surfaces

as a variety of experimental techniques have allowed the preparation of samples with

controlled roughness.'"' ' Possible applications include drag reduction in microfluidic

devices'"*"'^ and so called "self-cleaning" surfaces'''"'^' on which liquid droplets easily

roll and take dirt and contaminants off of the surfaces. For most of the applications of

these surfaces, the important issue is not the fact that the contact angle is high, but how

readily the liquid moves on the surface, a dynamic process. In this research study,

condensation and its effect on liquid droplet mobility were studied on several

ultrahydrophobic surfaces. We show that condensed water can dramatically reduce the

water repellency of ultrahydrophobic surfaces as evidenced by dynamic contact angle

measurements.

4.2 Background

4.2. 1 Superhvdrophobic Surfaces

As discussed in the previous chapter, surface roughness can enhance the

hydrophobicity of a surface. The highest advancing contact angle achievable on a flat,

hydrophobic surface is -120" with close packed -CF3 groups"", but roughness can

126 '^^ increase the value to nearly 1 SO".""' Surfaces with advancing contact angles above

-150" are typically termed superhydrophobic in the literature.

Superhydrophobic surfaces are of interest because liquids often move very

easily on them. However, as was pointed out by Chen et al., researchers frequently

report only static contact angles which give little information about the liquid mobility

on the surface."^^ They suggested that hydrophobicity should be determined with

regards to liquid mobility rather than just absolute contact angle values. The term

ultrahydrophobic was recommended to describe surfaces with both high contact angle

and low hysteresis on which water droplets move very easily. They gave examples of

surfaces with no hysteresis on which a drop of water would not come to rest. Even

surfaces with contact angle <90° could be considered water repellant because, with no

hysteresis, a drop would move at a small tilt angle. (Throughout this text, the term water repellant will be used to describe a surface on which liquid droplets move with

minimal applied force. In some field of study "water repellanf is reserved exclusively

for surfaces with contact angles >90".)

As previously discussed in Chapter 3 (Equation 3.6), in order for a liquid droplet to move on an inclined an external force must be applied to overcome the surface tension forces keeping the drop pinned.

- mg sin a - kn'/i, (cos 0^ cos 0^ ) (4.1)

Equation 4. 1 shows that the tilt angle needed to induce drop motion would be zero for a sample with no hysteresis. As the volume of a droplet increases, the pull of gravity

127 "

becomes greater and a lower tilt angle is needed to induce motion. The equation also

reveals that a small amount of hysteresis becomes less of a hindrance to drop motion as

the contact angle increases. This was illustrated nicely by Quere in a recent review.

Figure 4.1 shows a plot of the mean contact angle, 6 = (0a+Or)/2, versus the critical

contact angle hysteresis, AG, below which the droplet will move and above which the

droplet remains pinned. For a certain tilt angle and droplet size, in this case a = 90° and

w = 1 mm, the critical hysteresis value is low for hydrophilic surfaces. Only a few

degrees of hysteresis can be tolerated before the droplet mobility is lost. The critical

value grows slowly with the average contact angle until it exponentially increases at

~1 60". Double digit hysteresis does not hinder the motion of liquid droplets as much when the surface has very high contact angles. The reasons for this observation will be discussed below.

50 A0 C) 40-

30-1 dr

10 j

0 0 ~20 40 60 80 100 ^20 140 160 180 9

Figure 4.1. The critical contact angle hysteresis, AG, plotted versus the mean contact angle, G, for a water droplet with a 0.5 mm radius on a surface tilted 90". Below the critical line the drop can move but above it the drop remains pinned to the surface. Reprinted from reference 26. 4.2.2 Hysteresis on Rough Hydrophobic Surfaces

Johnson and Dettre performed several detailed studies of the wettability of

'"'" rough surfaces.^ They confirmed that, upon increasing the surface roughness, a drop of liquid transitioned from wetting the entire surface to sitting on top of the features as

predicted by Wenzel^' (Figure 4.2b) and Cassie-Baxter^'^"'^ ' (Figure 4.2a) respectively.

More importantly, they showed that the transition was accompanied by a substantial decrease in contact angle hysteresis. Figure 4.3 shows the change in dynamic contact

angle with increasing roughness for a hydrophobic surface that is initially flat with a static contact angle of 120°."^^ At small roughness ratios, in the Wenzel regime (B), the

liquid wets the entire surface and the hysteresis is high as shown by the difference in the advancing and receding contact angles, (dashed lines (D) and (C) respectively). At

roughness ratios above 2.0, in the Cassie-Baxter regime, the liquid drop is supported on

top of the surface asperities and the hysteresis is low. Liquid is expected to roll off the porous surfaces more easily, making them the preferred type for most water repellency applications.

Figure 4.2. Diagram of a water droplet on a rough, hydrophobic surface that is in either the (a) Cassie-Baxter or (b) Wenzel wetting regime.

129 40 - 30 L

1.0 2.0 3.0

Roughness E.atio, r

Figure 4.3. Contact angle as a function of roughness for an initially flat surface having a static contact angle of 120°. Reprinted with permission from reference 27. Copyright 1964 American Chemical Society.

4.2.3 Wetting Transitions

The discussion thus far would make it seem that a hydrophobic surface could only be in the Wenzel or Cassie-Baxter regime, but in fact both types of wetting can occur on the same surface. Starting from the Cassie-Baxter regime, water cannot wet in

between the surface features unless a pressure is applied to overcome the negative

capillary pressure of the hydrophobic posts. The capillary pressure, Pc, is defined as.

Lcp (4.2) A

where yiv is the liquid/vapor surface tension, 0o is the advancing contact angle of the material, Lcp is the total solid perimeter in a repeat unit of the surface configuration, and

130 *^'^ Acp is the area of air space in the repeat unit.''' This critical pressure can be achieved by increasing the size of the drop, the density of the liquid, or letting the drop fall from a certain height onto the surface. Therefore, as pointed out by Toridceli, "a minimum

criteria for an ultrahydrophobic surface is that the weight of the liquid drop not exceed the Laplace pressure".

Several researchers have recently studied the transition from the Cassie-Baxter to Wenzel wetting regime on a single surface." Quere and Laftima showed that a drop of water would transition from the Cassie-Baxter to Wenzel wetting regime by

placing it between two ultrahydrophobic surfaces and applying pressure.^ Patankar and co-authors published several papers discussing the wetting transition on an

37 38 ultrahydrophobic surface made of PDMS. They observed that a drop placed gently

on the surface sat on top of the surface features, but wet between the features if it fell

onto the surface from a certain height. Marmur''^'"*^, Carbone et al."^"^, and Dupuis et al."*^ have discussed the theory behind metastable states on superhydrophobic surfaces.

Another way to observe a wetting transition on an ultrahydrophobic surface is to

condense water on it. Water condenses on a surface that is cooled below the dewpoint.

The drops initially nucleate evenly over a rough surface including between the asperities. Quere and co-authors compared the dynamic contact angle values of a water drop placed gently on an ultrahydrophobic surface with the values obtained for a drop

7 35 that was grown by condensation on the same surface. ' The advancing contact angle only decreased from 164° on the dry surface to 141° on the wet surface, but the contact angle hysteresis increased from 5° to over 100°! This indicates that the water drop switched from sitting on top of the surface features to wetting entire surface. The huge

131 hysteresis observed on the wetted surface means that a drop would not move easily if

the sample was tilted and therefore its water repellency would be poor.

Quere and co-workers mentioned the condensation experiments only briefly in

their papers. In this project the dynamic contact angles were measured for a variety of

ultrahydrophobic surfaces with varying surface features. Another question that was

addressed was the morphology of the condensed water on the surface. When the

amount of condensed water on the surface increased, would the water drops remain in

between the asperities or would there be a critical point when the drops were forced up

onto the tops of the features?

After this research was initiated in our lab, Narhe and Beysens published a paper

about condensation on superhydrophobic surfaces. They were interested in the growth

dynamics of the droplets and showed time-lapsed optical micrographs of water

condensation on grooved silicon samples. Drops formed both in the channels and on the ridges. Eventually the entire surface was covered, confirming that Wenzel-type

wetting occurs during condensation. Very little was mentioned about contact angle angle or hysteresis.

4.3 Experimental

4.3.1 Materials

Dimethyldichlorosilane (DMDCS) was purchased from Gelest and used as received. Silicon wafers were purchased from International Wafer Service (<100>

132 orientation, P/B doped, resistivity from 20.0 to 40 Q cm). House-purified water

(reverse osmosis) was further purified using a Millipore Milli-Q system (18.2 MQ cm).

4.3.2 Experimental Procedures

4.3.2.1 Ultrahvdrophobic Surfaces

Rough silicon wafers with a variety of patterned, micron-sized features were

previously prepared by another student using lithography.^'^'' After being cleaned in a

7:3 mixture of sulfuric acid and hydrogen peroxide (30%), the samples were

hydrophobized using a vapor phase reaction. The wafers were suspended above 0.5 mL

of DMDCS in a sealed vessel for 3 days at ~65 °C. After the reaction they were rinsed

with toluene, ethanol, and water.

4.3.2.2 Condensation Experiments

Optical microscopy of condensation on various ultrahydrophobic surfaces was

performed on an Olympus BX5 1 research microscope equipped with a Qlmaging

Micropublisher CCD camera. A simple cooling stage was custom made using common

laboratory materials. A piece of hollow 1/4 copper tubing was bent at 90° and a flat copper sample stage was attached to one end. The other end was placed into liquid nitrogen and the sample was slowly cooled by conduction. If the dewpoint was especially low, drops of water were placed near the sample to create a more humid environment.

133 Another custom made cold stage was built using 1/8" copper tubing which was snaked back and forth underneath a 1" x 2" steel sample stage to provide efficient cooling. The ends of the tubing were attached to a circulating water bath filled with a

7:3 mixture of water and ethylene glycol antifreeze. A clear cover was made from glass microscope slides and polystyrene and placed on top of the sample stage to create an enclosure where the humidity could be adjusted to 100%. There was a small hole in the top so the needle could reach the sample for contact angle measurements.

Contact angle data was obtain using a goniometer and computer-controlled video capture system custom built by members of Professor Crosby's research group in the Polymer Science and Engineering Department. Advancing and receding contact angles were measured by adding or withdrawing water from a droplet respectively. The

droplet diameter was usually less than 2.7 mm, which is the capillary length for water,

48 i.e., the critical drop size above which gravity begins to affect the shape of the drop.

All contact angle measurements are averages of at least 2 spots per sample.

4.4 Results and Discussion

4.4.1 Preparation of Ultrahydrophobic Surfaces

A series of topography-based ultrahydrophobic surfaces containing hydrophobized silicon posts were used for the condensation experiments. The details of the photolithographic preparation and silane modification of the silicon post surfaces were described above and have been reported elsewhere.^ The various surfaces are designated ^X^ '^, where W is the post width, Y and Z are the dimensions of the unit

134 cell, and X is the post shape: SP for square post, StP for four arm star, IP for indented square post (Figure 4.4a,b). All of the posts are 40 ]xm in height. A representative scanning electron microscopy image of the lithographically prepared,

dimethyldichlorosilane hydrophobized silicon posts is shown in Figure Ic. Water droplets placed on such surfaces rest on top of the posts on a solid/air composite surface. We have previously shown that the surfaces are ultrahydrophobic and have reasonably low contact angle hysteresis."^ Water droplets move easily when the surfaces

are slightly tilted.

C

Figure 4.4. (a) Diagram showing the dimensions used to designate the various silicon post surfaces: W is the width of the post, Y and Z are the dimensions of the unit cell, (b) Illustrations of the top of an indented square post, IP, and a four armed star post, StP, (c) Representative SEM of lithographically prepared silicon posts.

135 Water droplets can either sit on top of the hydrophobic posts or wet between them as was discussed earher and shown schematically in Figure 4.2. A demonstration

^'^Mmgpi28,^56Mm^ of the transition between the two wetting regimes on such a surface, shown in Figure 4.5. (The feature sizes were so large that the contact line could be observed jumping from one post to the next.) The water droplet in Figure 4.5a was placed gently on the surface and had a static contact angle of -135°. The advancing and

receding contact angles were 155° and 115° respectively. It moved very easily when the surface was tilted to -20°. When the suspended droplet in Figure 4.5a touched the first droplet, the contact angle dramatically changed to -90° (Figure 4.5b), with dynamic

contact angle values of 135°/80°. The droplet no longer moves, even if the sample is held vertical. The feature size and spacing was such that the negative capillary pressure keeping the droplet suspended on top of the posts was low and was easily overcome by the force of the second droplet falling on the surface.

a b

Figure 4.5. Optical micrographs of two different static contact angles on the same rough, hydrophobic surface. Image (a) was taken after the droplet was carefully placed on the surface and (b) was taken after the two drops in image (a) merged.

136 ' Variations of this simple experiment have already been reported elsewhere ,

but it vividly illustrates the transition between the two wetting regimes. It also reveals a

limitation to keep in mind when choosing applications for surface. For example, if the

sample was meant to be water repellant, it would not work very well if raindrops fell on

it with enough force to overcome the negative capillary pressure.

4.4.2 Condensation on Ultrahydrophobic Surfaces

4.4.2.1 Optical Microscopy

The growth of condensed water was observed on a variety (see Table 1 ) of these surfaces by optical microscopy. A custom built cold stage was used to cool the samples below the dewpoint. Figure 4.6 shows a succession of optical micrographs of

-^""sp^'^Mm condensation on sample water droplets initially nucleate on the sides of the posts and grow to span the gap between neighboring features (Figure 4.6a,b). The droplets grow larger as more water condenses, but they are laterally confined by the hydrophobic posts. Surface tension forces keep the drop reasonably spherical so instead of growing around the posts, the droplets grow upward and protrude above the surface features as indicated by the white arrows in Figures 4.6c,d.

137 c d

Figure 4.6. Sequential optical micrographs of condensed water growth on (-Mmgp4,8(jm^ hydrophobized silicon posts: 40 ]im tall, 2 fim wide, 4x8 \im unit cell White arrows indicate drops that protrude above the tops of the posts.

Figure 4.7. Sequential optical micrographs of water evaporating from hydrophobized (^Mmsp^.^Mni-^ silicon posts: 40 fxm tall, 2 )im wide, 4x8 ^im unit cell White arrows indicate drops that transitioned from sitting on top of the posts to wetting between them.

138 Evaporation of the water droplets from the same surface is shown in Figure 4.7.

The droplets indicated by the white arrows initially stick out above the tops of the hydrophobic posts. Their volume shrinks as they evaporate and in Figure 4.7c the droplets have formed capillary bridges between 3 or 4 neighboring features. The droplets have transitioned from sitting on top of the posts to wetting between them, an exact reversal of what was seen during condensation.

The condensation experiments were reproduced using an environmental SEM

(FEI Quanta 200) equipped with a Peltier cooling stage. The ESEM allowed precise control over the temperature, pressure, and humidity so many condensation and evaporation cycles could be run reasonably quickly. A sample with similar, but slightly larger features than that shown in Figures 4.6 and 4.7 was placed in the ESEM. The chamber was pumped down to 5.44 Torr and the cooling stage set to 5 "C. Water vapor initially condensed on the sides of the posts (Figure 4.8a) and over time the droplets began to coalesce and span between the posts (Figure 4.8b). Controlled cooling showed

that the initial capillary bridges formed between two posts before wetting multiple

features. This is in slight contrast to optical microscopy where the first observable drops nearly always bridging between multiple posts. The transition from wetting between the features to sitting on top of them was not observed upon further cooling

because, even if it did occur, the contrast and depth of view would make it very difficult to distinguish.

139 a b

Figure 4.8. Sequential environmental SEM images of water condensation on hydrophobized silicon posts: 40 fim tall, 8 ]im wide, 16x32 \im unit cell (^^'"^sp'^-^-^'"^)

A direct comparison with optical microscopy is not completely appropriate because the ESEM measurements were done at reduced pressure. Nonetheless, a similar development of the condensation morphology was observed. Additional ESEM condensation experiments on other hydrophobic samples were not performed because

the sample stage was too small and the expensive post surfaces had to be broken to fit

on it. Also, the size of the features was such that optical microscopy was more than adequate to view the condensation, although with slightly less control of the experimental parameters.

Large water droplets with diameters of hundreds of microns formed when condensation was allowed to progress for a long time. Figure 4.9 shows several large water drops on the same surface as in Figures 4.6 and 4.7. The droplets function as lenses and reveal the area beneath them that consists of a mosaic of regions where either

water (darker areas) or air (lighter areas) is present between the posts. These droplets are in both the Wenzel and Cassie-Baxter wetting regimes, sitting on both a solid/air

140 and solid/liquid composite surface. The contact lines of the large droplets are contorted and not circular as pointed out by the arrows. Water between the posts has coalesced with the droplets and has pinned the contact line. As will be shown by contact angle

measurements, the mobility of the liquid droplets on the surface is decreased dramatically.

Figure 4.9. Optical micrograph of large condensed water droplets on a surface of (-Mmgp4,8j.m^ hydrophobized silicon posts: 40 |Lim tall, 2 |im wide, 4x8 fim unit cell Black arrows indicate distortions of the contact line. Darker areas, both underneath and between the drops, indicate where water penetrates between the posts.

4.4.2.2 Contact Angle Measurement

Dynamic contact angle measurements were made on the same surface as previously described both before and after condensation to determine what effect, if any, the condensed water had. The sample was cooled below the dewpoint using a custom made sample stage connected to a circulating water bath. Damp cloth was placed next to the sample and both were covered by a transparent enclosure to keep the humidity as

141 high as possible. Contact angle measurements began when condensation was visible on

the surface and multiple spots were used as condensation progressed.

Figure 4.10 shows snapshots of the advancing and receding contact angle before

and after condensation. The dry surface had advancing and receding contact angles of

168" and 130° respectively (Figure 4.10a). Cooling this sample below the dewpoint induced dramatic changes in wettability; as shown in Figure 4.10b, the dynamic contact

angles decreased substantially after water vapor condensed. The advancing angle still shows a hydrophobic value of 97°, but the receding angle decreases to 0° resulting in an

enormous increase in hysteresis. This increase in hysteresis upon condensation is

7 35 similar to that observed by others. '

b

Figure 4. 10. Advancing and receding contact angles on a surface with hydrophobized (2Mmsp4.8Mm>j. silicon posts: 40 |xm tall, 2 |im wide, 4x8 (im unit cell surface (b) surface with condensed water (needle diameter is 0.5 mm).

142 The difference between the advancing and receding contact angles is referred to

as the hysteresis and this value is sometimes used to contrast the wetting characteristics

of two surfaces. However, a better value to use for comparisons is the difference

between the cosines of the receding and advancing angle. This value is directly related

to the mobility of the liquid on a surface, as shown in Equation 3.6, with lower values

indicating that a liquid droplet moves with less applied force, assuming the other

parameters remain constant. The cosine of the receding contact angle minus the cosine

of the advancing contact angle will be ternied COS(HY) in the rest of the text..

The COS(HY) for the dry surface in Figure 4.10 was 0.34 and after

condensation it increased to 1.12. This indicates that a larger force must be applied to

overcome the surface tension forces and move the water droplet on the "wef surface.

The force in Equation 4.1 also depends on the contact width of the drop with the surface.

At an equal volume, the drop with the lowest contact angle, the least hydrophobic, will

have the largest contact width. This will further increase the necessary applied force to

cause drop motion.

A qualitative, but more direct measure of the water repellency of the surface is

shown in Figure 4. 11 . The dry sample was inclined nearly vertically and a drop of water was placed on the surface. The water droplet easily runs off and the surface

remains dry and water repellant (Figure 4. 1 la). The same sample is shown in Figure

4.11b after water vapor has condensed on it. Now the drop of water remains pinned to the surface.

143 Figure 4.12. (a-d) Sequential optical micrographs of condensed water growth on hydrophobized silicon posts: 40 iim tall, 8 fim wide, 32x80 unit cell (^^"'SP^--^^^"'').

(e,f) Water drops easily roll off the tilted surface when it is dry (e), but stick after condensation (f).

144 .

A similar series of condensation experiments was performed on the

^Mmgp3-,«0M"i ultrahydrophobic surface (pjgure 4.12). The posts were larger and further

apart than on the first surface, -^*™SP"^'^^'"\ Condensation began on the sides of the posts and the droplets grew until they coalesced and fonned capillary bridges between the features. The drops grew much larger before they protruded above the tops of the posts.

Water easily rolled off the diy surface when it was tilted (Figure 4.12e), but after condensation the added water drops remained pinned (Figure 4. 120- The dynamic

170''/ 145'' 13°/51° contact angle values before and after condensation were and 1

respectively (Table 4. 1 ). The COS(HY) of the dry surface was 0. 1 7 while that of the

wet surface was 1 .02, nearly six times as high.

The dynamic contact angles for a number of other surfaces based on

hydrophobized silicon pillars before and after condensation are shown in Table 4. 1

The dry surfaces are ultrahydrophobic with relatively low COS(HY) values varying

from 0. 1 7 to 0.45. The advancing and receding contact angles decrease in all cases after water vapor condensation with surfaces having COS(HY) values from 1.02 to 1.74.

While the advancing contact angles all remained hydrophobic (>90°), all the receding angles dropped below 90^*. The values of the receding contact angles varied from complete wetting, 0°, to 72°. (Additional optical micrographs of condensation on the other hydrophobic surfaces listed in the Table 4.1 can be found in Appendix A.)

145 1

Table 4.1. Dynamic contact angle measurements on various hydrophobic silicon post surfaces before (dry) and after (wet) condensation.

A 1 1 im'' ^ suriacc Wr (aeg) COSOr-COSOa /\ v|J.m )

S|.inic'n-'2,32nm 1'^ '>8 1 0 0 1 1 0 063 Ul y 1

Wcl 1 TO 1 41

8|.iniQr)32,80fjm vji y I 70 145 0 17 0 05 0 075

WCl 1 1 1 s 1 1 0?

2)iniQp4,8|.im or Hrv 1 V/ 0 1 30 0 34 0 25 0 50

t)7 1 1 WCl y / u 9

8|iniQpl6,32(ini Hrv 1 7^ 1 23 0 4S 0 95 0 1 3

W d 1 17 11 1 04 8|ainTnl^.32|am Ht*\/ "^8 li Lii y 11/171 1 ^0 0 37 0 1 Q 0 1 5

wet 128 0 1.62

8nnig^pl6,32|ani dry 171 140 0.22 0.13 0.14

wet 138 0 1.74

32|Lim square dry 152 127 0.28 0.13 0.12

hollow wells wet 140 39 1.54

fi is the fractional contact area between a liquid droplet and the post tops; A is the contact line density, i.e., the post top perimeter per unit area.^'^

Table 4. 1 also contains two parameters that describe the surface topography.

The column fi is the fractional contact area between a liquid droplet and the post tops or in other words, the fraction of the projected surface area that the post tops occupy. The

sum of the solid fraction, f|, and air fraction, f:, equals 1. When sun-ounded by air, a

water droplet adopts the shape of a sphere to minimize its surface area. Therefore it makes sense to maximize the air fraction of the surface and minimize the solid fraction to get the highest possible contact angle, keeping in mind that if the posts are too sparse they will not support the weight of the drop.

The values of fi and f: can be used in the Cassie-Baxter equation (Equation 3.5) to calculate the equilibrium contact angle on the surface. The theoretical contact angle

146 was determined for each surface and was very similar to the average of the experimentally measured advancing and receding contact angles (Table 4.2). (The

average of the advancing and receding contact angles is often used as an estimate of the equilibrium contact angle. The Cassie-Baxter equation does not take hysteresis into account.) No correlation between the fractional solid surface area and the change in contact angle or hysteresis after condensation could be found.

The second parameter that describes the various surfaces is A, which is the post

top perimeter per unit area.^*^ Our group has shown that increasing the contact line length often causes an increase in the receding contact angle due to a more tortuous and contorted path that the contact line must take on such surfaces."* The importance of the

interactions at the contact line to the observed contact angle is still unclear. The most widely used equations, such as those derived by Wenzel and Cassie-Baxter, emphasize that the contact area underneath the liquid droplet determines the contact angle.

However, some experimental evidence is at odds with that prediction (see Extrand"^° and references therein). For example, Extrand did a simple experiment on a surface patterned as shown in Figure 4.13. The sample had a hydrophobic spot surrounded by a

hydrophilic matrix. When a small drop was placed in the center it had the expected advancing contact angle of the hydrophobic material, 95°. However, when the volume was increased and the contact line crossed onto the hydrophilic area the contact angle was immediately that of the hydrophilic region, 6°. Recall Cassie's equation for a chemically heterogeneous surface,

COS 6^ = / COS ^1 + /, cos^. (4.3)

147 which predicts that the observed contact angle is the sum of the contact angle on each

different chemistry multiplied by the fractional area each chemistry occupies under the

drop. Using this equation, Extrand calculated the theoretical contact angle to be 33",

which was at odds with the experimental evidence.

V////A I

a b e

Figure 4.13. (a) Hydrophobic spot on a hydrophilic matrix, (a) Water contact angle on the hydrophobic spot (b) Contact angle on a mixed surface.

Even though that example was of a chemically heterogeneous surface, the basic premise can be extended to rough surfaces. We believe that the increase in the contact

line length is more important than the increase in solid/liquid contact area underneath the drop in regard to an increase in hysteresis. Figure 4.9 showed that large water droplets grown by condensation sat on a composite surface consisting of areas with

either water or air between the posts. A close look at the contact line showed that it was distorted by smaller droplets protruding from the tops of the hydrophobic posts. While

not conclusive proof, it strongly suggests that pinning at the three phase contact line affects the droplet mobility on surfaces wet with condensation. An attempt was made to discover a relationship between A and the increase in contact angle hysteresis or receding contact angle to strengthen the argument that the contact line was important.

However, no correlation could be found.

148 4.4.2.3 Capillary Pressure

There is still the question of why the water droplets initially condense between

the posts until a certain size and then protrude above the surface. Let us start with the

opposite question: why does an evaporating drop transition from sitting on top of a

rough surface to wetting in between the hydrophobic features? As mentioned in the

introduction there is a negative capillary force resisting penetration of the water droplets

in between the hydrophobic posts (Equation 4.2). It was already shown that the

resistive force could be overcome by pushing on the drop or increasing its mass. There

is an inherent Laplace pressure in a liquid droplet caused by surface tension, which increases as the radius of the droplet decreases."^' Equation 4.4 states

2y AP = (4.4) R

where AP is the difference in pressure between the droplet and the surrounding medium,

yiv is the liquid, vapor surface tension and R is the radius of the droplet. Some experimental evidence suggests that the inherent Laplace pressure of a small droplet can

overcome the negative capillary force. Palzer et al., prepared hydrophobic glass capillaries and placed water droplets of varying radii on the end.'^" They found that

when a droplet was small enough it was able to penetrate the hydrophobic capillary without the application of external pressure. Quere and coworkers have studied

^'^'"^'^ transitions between the Cassie-Baxter and Wenzel wetting regimes. They showed a picture of two water droplets on a superhydrophobic surface.^ ' The large drop was

sitting on top of the surface features and had a high contact angle, > 1 50", and a much

149 smaller drop had a contact angle <90° and was probably wetting between the surface

features.

This could explain the behavior that was obsei"ved by optical microscopy in this

study. As the droplets evaporate the radius becomes smaller and the Laplace pressure

of the droplet increases. At some critical radius it can overcome the negative capillary pressure exerted by the hydrophobic posts. The reverse would be true during condensation. The initial droplets were small, but as they grew larger, their internal

Laplace pressure decreased and they were pushed to the top of the posts by capillary force.

The negative capillary pressure exerted by the hydrophobic posts can be calculated using Equation 4.2. A more specific form of the equation was derived for a regular array of posts by Zheng and coworkers^' (A similar equation was developed by

Torkkeli.)"'* Equation 4.5 shows the expression for the negative capillary pressure, Pc,

YJ cose P = (4.5)

where f is fraction of the projected surface occupied by the post tops and X is the post

top surface area divided by its perimeter. Table 4.2 shows the calculated values of Pc for the hydrophobic post surfaces. Setting Equation 4.4 equal to Equation 4.5 we can determine critical droplet radius, R^, that produces a Laplace pressure equal to the negative capillary pressure. The last column of Table 4.2 gives the results.

150 Table 4.2. Experimental and calculated parameters describing the various hydrophobic post surfaces.

(eA+eR)/2 Pe X 10" surface Rc (l^m) (degrees) (degrees) (Pa)

8)im^jp32,32|im 154 154 2.0 1.08 135

X)amgp32.80).im 158 164 2.0 0.40 366

2|imgp4.8|am 149 145 0.5 10.1 14

8|jnigpl 6,32|am 148 143 2.0 2.5 58

8|imjpI6,32|jm 149 148 1.24 2.8 52

X|.inig^pl6,32f.im 156 154 0.89 2.4 60

0CB is the theoretical contact angle calculated using the Cassie-Baxter equation

(Equation 3.5), A. is the post top surface area divided by its perimeter, P^ is the negative capillary pressure as calculated using Equation 4.5, Rc is the droplet radius that produces a Laplace pressure (Equation 4.4) equal to Pc.

The calculated value of Rc can be compared to the critical radius observed by optical microscopy. Looking back at Figures 4.6 and 4.7, the radius of the droplets after extrusion from (condensation) or intrusion into (evaporation) the hydrophobic posts of

*^^'"^ surface -^'"^sp"' was on the order of ~5 |im compared to the calculated Rc of 14 jim.

*

|im and -15 \im respectively. The calculated value of Rc is the right order of magnitude,

but seems to an overestimate compared to the observed radius. Apparently there is another force on the drop that also opposes the negative capillary pressure. Gravity was

neglected, but it should not have much effect on such small droplets. Another reason for the discrepancy between the observed and calculated Rc has not been determined.

151 4.4.2.4 Nanoscale Roughness

All of the surfaces previously described had micron size features. Water vapor

was condensed on a surface with nanometer scale roughness to examine its effect on

water droplet mobility. The surface was prepared by another member of the research

group using a solution phase silane reaction to modify a silicon wafer (Figure 4.14a).

The details of the sample preparation will be published elsewhere. What is important is

that the dry sample was very hydrophobic with dynamic contact angle values of

17071 17". The contact angles measured after condensed water droplets formed on the

sample were 1 15752°. Figure 4.14b,c shows the receding contact angles before and

after condensation. The COS(HY) increased from 0.53 for the dry surface to 1 .04 for

the wet surface indicating a loss of liquid mobility. Environmental SEM was used in an

attempt to view the initial nucleation and growth of the water droplets but the contrast

was too poor at high magnification.

a b c

Figure 4.14. (a) A hydrophobic surface with nanoscale roughness, (b,c) Receding contact angle measurements on the surface before and after condensation respectively

(needle diameter is 0.5mm).

152 4.4.2.5 Low Hysteresis during Condensation

Condensation affects water droplet mobility on a variety of hydrophobic

surfaces with different size scale features so the obvious question to ask was: Can a

surface be prepared on which water droplets remain mobile even during condensation?

Surfaces with low hysteresis but minimal roughness, i.e., not ultrahydrophobic, might

be an option since penetration of the condensation between the features on the rough

surfaces appears to be the reason for the increase in hysteresis.

A polished silicon wafer was modified with dimethyldichlorosilane via the same

vapor phase reaction used to hydrophobize the silicon post surfaces. The modified

surface had a thickness of ~3 nm and an average roughness of only a few angstroms.

Dynamic contact angle measurements on the DMDCS surface before and after condensation were 101"/99'' and 95794" respectively. The hysteresis was very low both before and after condensation. (The slightly lower contact angles measured after condensation are due to either the lower temperature which caused a small change in surface tension or difficulty in accurately measuring the contact line tangent due to small water droplets on the surface obstructing the view of the three phase contact line.)

Water droplets moved easily when the dry or wet surface was slightly tilted confirming that liquid mobility was maintained after condensation.

153 4.5 Conclusions

Condensation on a variety of ultrahydrophobic surfaces was examined and it

was shown to cause an increase in the contact angle hysteresis and a resultant decrease

in macroscopic water drop mobility. Optical microscopy showed that the water

condensed between the posts and spanned between multiple features before being

pushed to the surface by the negative capillary pressure of the hydrophobic posts. A

macroscopic water drop coalesced with the condensed water and underwent a wetting

transition from sitting on top of the posts to wetting between them. The condensed

liquid pinned the contact line of the macroscopic droplet and dynamic contact angle

measurements revealed an increase in hysteresis which corresponds to a decrease in

liquid mobility. In applications where liquid mobility is important and condensation is likely to occur, an alternative to ultrahydrophobic surfaces would be a smooth surface

with little or no hysteresis.

154 4.6 References

(1) Onda, T.; Shibuichi, S.; Satoh, N.; Tsujii, K. Langmiiir 1996, 72, 2125-2127.

(2) Youngblood, J. P.; McCarthy, T. J. Macromolecules 1999, 32, 6800-6806.

(3) Bico, J.; Marzolin, C; Quere, D. Europhys. Lett. 1999, 47, 743-744.

(4) Oner, D.; McCarthy, T. J. Langmuir 2000, 16, 1111-11%!.

(5) Yoshimitsu, Z.; Nakajima, A.; Watanabe, T.; Hashimoto, K. Langmuir 2002, 75,5818-5822.

(6) Quere, D. Physica A 2002, 573, 32-46.

(7) Laftima, A.; Quere, D. Nat. Mater. 2003, 2, 457-460.

(8) Shiu, J. v.; Kuo, C. W.; Chen, P. L.; Mou, C. Y. Chem. Mat. 2004, 16, 561- 564.

(9) Ma, M.; Hill, R. M.; Lowery, J. L.; Fridrikh, S. V.; Rutledge, G. C. Langmuir 2005, 27, 5549-5554.

(10) Shirtchffe, N. J.; McHale, G.; Newton, M. I.; Perry, C. C. Langmuir 2005, 27, 937-943.

(11) Jisr, R. M.; Rmaile, H. H.; Schlenoff, J. B. Angew. Chem.-lnt. Edit. 2005, 44, 782-785.

(12) Yabu, H.; Takebayashi, M.; Tanaka, M.; Shimomura, M. Langmuir 2005, 27, 3235-3237.

(13) Shiu, J.-Y.; Kuo, C.-W.; Chen, P.; Mou, C.-Y. Chem. Mat. 2004, 16, 561-564.

(14) Torkkeli, A.; Ph.D. Dissertation, Helsinki University of Technology, 2003.

155 ( 15) Kim, J.; Kim, C.-J. In Proc. IEEE Int. Coiif. MEMS, 2002; pp 479-482.

(16) Ou, J.; Perot, B.; Rothstein, J. P. Phys. Fluids 2004, 16, 4635-4643.

(17) Barthlott, W.; Nemhuis, C. Planta 1997, 202, 1-8.

(18) Feng, L.; Li, S. H.; Li, Y. S.; Li, H. J.; Zhang, L. J.; Zhai, J.; Song, Y. L.; Liu, B. Q.; Jiang, L.; Zhu, D. B. Adv. Mater. 2002, 14, 1857-1860.

( 1 9) Blossey, R. Nat. Mater. 2003, 2, 30 1 -306.

(20) Furstner, R.; Neinhuis, C; Barthlott, W. Nachr. Chem. 2000, 48, 24-28.

P. 956- (2 1 ) Furstner, R.; Barthlott, W.; Neinhuis, C; Walzel, Langmiiir 2005, 21, 961.

(22) Nishino, T.; Meguro, M.; Nakamae, K.; Matsushita, M.; Ueda, Y. Langmiiir 1999, 75, 4321-4323.

(23) Bemett, M. K.; Zisman, W. A. J. Phys. Chem. 1962, 66, 1207-&.

(24) Bartell, F. E.; Shepard, J. W. J. Phys. Chem. 1953, 57, 211-215.

(25) Chen, W.; Fadeev, A. Y.; Hsieh, M. C; Oner, D.; Youngblood, J.; McCarthy,

T. J. Langmuir 1999, 75, 3395-3399.

(26) Quere, D. Rep. Prog. Phys. 2005, 68, 2495-2532.

(27) Johnson, R. E.; Dettre, R. H. In Advances in Chemistiy Series; Gould, R. F., Ed.; American Chemical Society: Washington, 1964; Vol. 43, pp 136-144.

(28) Dettre, R. H.; Johnson, R. E. In Advances in Chemistry Series; Gould, R. F., Ed.; American Chemical Society: Washington, 1964; Vol. 43, pp 136-144.

(29) Dettre, R. H.; Johnson, R. E. In Wetting; Society of Chemical Industry: London, 1967; pp 144-155.

156 (30) Johnson, R. E.; Dettre, R. H. In Surface and Colloid Science; Matijevic, E., Ed.; Wiley: New York, 1969; Vol. 2, pp 85-153.

(31) Wenzel, R. N. Industrial and Engineering Chemistiy 1936, 2<*?, 988-994.

(32) Cassie, A. B. D.; Baxter, S. Transactions of the Faraday Society 1944, 40. 546- 551.

(33) Cassie, A. B. D.; Baxter, S. Nature 1945, 755, 21-22.

(34) Bico, J.; Marzolin, C; Quere, D. Europhys. Lett. 1999, 47, 220-226.

(35) Quere, D.; Laflima, A.; Bico, J. Nanotechnology^ 2003, 74, 1 109-1 1 12.

(36) Patankar, N. A. Langmuir 2003, 79, 1249-1253.

(37) He, B.; Patankar, N. A.; Lee, J. Langmuir 2003, 79, 4999-5003.

(38) He, B.; Lee, J.; Patankar, N. A. Colloid Surf A-Physicochem. Eng. Asp. 2004, 248, 101-104.

(39) Patankar, N. A. Langmuir 2004, 20, 8209-8213.

(40) Patankar, N. A. Langmuir 2004, 20, 1091-7 1 02.

(41 ) Zheng, Q. S.; Yu, Y.; Zhao, Z. H. Langmuir 2005, 27, 12207-12212.

(42) Marmur, A. Langmuir 2003, 79, 8343-8348.

(43) Marmur, A. Langmuir 2004, 20, 35 1 7-3519.

(44) Carbone, G.; Mangialardi, L. European Physical Journal E 2005, 16, 67-76.

(45) Dupuis, A.; Yeomans, J. M. Langmuir 2005, 27, 2624-2629.

157 (46) Narhe, R. D.; Beysens, D. A. Phys. Rev. Lett. 2004, 93.

(47) Oner, D.; University of Massachusetts, 2001.

(48) Clanet, C; Beguin, C; Richard, D.; Quere, D. Journal ofFluid Mechanics

2004, 57 7, 199-208.

(49) Extrand, C. W. Langmuir 2004, 20, 501 3-50 1 8.

(50) Extrand, C. W. Langmuir 2003, 19, 3793-3796.

(51) Adamson, A. W. Physical Chemistiy ofSurfaces, 5th ed.; Wiley: London, 1990.

(52) Palzer, S.; Sommer, K.; Hiebl, C. Chem. Eng. Technol. 2003, 26, 962-966.

(53) CaUies, M.; Quere, D. Soft Matter 2005, 7, 55-61.

158 APPENDIX A

SUPPLEMENTAL DATA FOR CHAPTER 2 AND 4

A.l Chapter 2

A. LI Tris(TMS) Silylation in scCO^

The introduction to Chapter 1 explained that scCO: is a good solvent for a

variety of chemical reactions. A former student studied the reactions of silanes with

silicon wafers in carbon dioxide.' He found that both liquid and supercritical CO2 could be used as solvents for the reaction of (dimethylamino)trialkylsilanes and chlorosilanes.

Tris(trimethylsiloxy)chlorosilane (Tris(TMS)Cl) is a bulky silane that reacts randomly and inefficiently with a silica surface leaving areas of exposed silanols that

are too sterically hindered to allow another Tris(TMS) molecule to fit. The kinetics of the reaction, as monitored by contact angle, reveal that the initial Tris(TMS) molecules react with the surface within an hour.^ The contact angle then slowly increases over the course of 5 days as the bulky Tris(TMS) slowly finds and reacts with the residual silanols that become more sterically hindered as the reaction proceeds. Comparison of the vapor phase and solution phase reactions showed that more densely packed monolayers were formed using the vapor phase reaction."

In Chapter 2, Tris(TMS)Cl was used to prepare incomplete monolayers from which silicon dioxide was grown. Carbon dioxide was examined as a potential solvent for the reaction of Tris(TMS)Cl with silicon wafers. The thought was that perhaps the low surface tension and high diffusivity of scCOt would speed up the reaction kinetics.

159 Reactions were run in both liquid and supercritical CO2 and compared with vapor phase reaction conditions. The process was similar to that used by Cao.' Silicon wafers were cleaned by oxygen plasma and placed in an 8.5 mL stainless steel high- pressure vessel. The vessel was sealed and filled with CO2 at room temperature, -23 °C, and 950 psi. A second smaller (0.35 mL) high-pressure vessel containing 2 mmol of

Tris(TMS)Cl was attached inline with the first and filled with CO2 under the same conditions. This was to prevent any reaction from occurring before the CO2 was added.

The other end of the small vessel was attached to a syringe pump set to deliver carbon dioxide at 1200 psi and 23 °C. The valve connecting the small vessel to the large one was opened to create a pressure gradient that forced the silane into the vessel with silicon wafers along with the CO2. The reaction was allowed to proceed at room

temperature for the liquid CO2 experiments or it was heated to 70 °C (4800 psi) for the

SCCO2 reactions. After the reaction the wafers were rinsed with toluene, ethanol, and water.

Table A.l gives a comparison of the kinetics via contact angle measurement for liquid CO2, SCCO2, and vapor phase Tris(TMS)Cl reactions. The surface coverage was calculated using the Israelachvili-Gee equation (Equation 2.5). After 4 hours the advancing contact angles for both liquid CO2 and SCCO2 are higher than for the vapor

phase reaction and remain so until the reaction is complete. The surface coverage calculation indicates that Tris(TMS)Cl reacts faster in carbon dioxide than under

conventional vapor phase conditions. However, it is important to note that the receding contact angles of the vapor phase reaction are nearly the same as the values recorded for

the CO2 conditions at each time interval. The receding angle is more sensitive to

160 hydrophilic groups on the surface, such as the residual silanols that become more sterically hindered as the reaction proceeds. Overall both the liquid and SCCO2 reaction conditions are at least equal to the conventional vapor phase reaction.

Table A. 1 . Comparison of vapor phase Tris(TMS)Cl silylation of silicon wafers with the same reaction using liquid CO2 or scCO: as the solvent.

Liquid CO2 SCCO2 Vapor phase

Time (hr) Oa/OrC") Coverage" eA/eR(^^) Coverage Oa/OrD Coverage

4 98/62 92 94/68 89 82/67 77

24 96/62 91 98/69 92 82/70 77

48 106/78 99 105/74 98 92/74 87

96 103/73 96 108/87 -100 99/84 93

Surface coverage (%) was calculated using the Israelachvili-Gee equation.

A. 1 .2 Phase Contrast on Block Copolymers

Block copolymers came to mind when looking for a surface with patterned surface chemistry and limited topography to test the modified AFM probes. Block copolymers with the right molecular weight and block lengths phase separate into a

variety of morphologies. One pattern of interest is a cylindrical phase of one polymer oriented perpendicular to the substrate surface and surrounded by a matrix of the other

polymer. The size scale of the phase separation is typically in the nanometer range which makes AFM a useful characterization technique. The AFM height image does not usually provide much information, but the morphology can be determined using the

161 tapping mode phase image. However, sometimes the phase contrast is not very high.

Chemically modified AFM tips could be used to improve the phase contrast.

Samples of various block copolymers with cylinders aligned perpendicular to the surface were generously donated by students in the Russell research group.

Polystyrene-b-poly(methyl methaciylate), polystyrene-b-poly(ethylene oxide), and

polystyrene-b-poly(4-vinyl pyridine) were used. In all cases the PS formed the matrix and the other polymer formed the cylinders.

Table A. 2. Degrees of phase contrast between the different phases of block copolymer films as determined using chemically modified AFM tips.

Block AFM tip chemistry copolymer -NH2 F,3 -CH3 -Si-OH

PS/PMMA 0.5(±o.3) 0.7(±o.3) 2(±i) 0.8(±o.i)

PS/PEO 1.9(±0.3) 1.7(±0.8) 0.8(±o.2) 0.5(±o.2)

PS/PVP 2.7(±o.3) 2.19(±o.o3) 0.7(±o.2) 0.73(±o.o5)

Tapping mode images of each block copolymer surface were obtained with the chemically modified AFM tips. The average phase contrast between the matrix and the

cylinders was then determined as described in Chapter 2 (Table A. 2). The amine and fluorine modified AFM tips showed the highest phase contrast with PS/PVP, while the n-propyl modified and unmodified tips had the highest phase contrast with PS/PMMA.

As mentioned previously, it is difficult to compare the different tips directly because of differences in tip radius. The images were also taken on slightly different areas of the

surface for each sample. However, it does appear that for PS/PEO and PS/PVP the amine and fluorine modified tips provide better phase contrast than the unmodified tips

162 that are commonly used. This could be of practical use for those who study block polymer moi"phology.

A.2 Chapter 4

Condensation was observed on silicon posts hydrophobized with dimethydichlorosilane having a variety of spacings and shapes. The following images show condensation on the structures listed in Table 4.1.

Figure A. 1 . (a-d) Sequential optical micrographs of condensed water growth on ('^Mmgp3^.3^Mm^ hydrophobized silicon posts: 40 |im tall, 8 \im wide, 32x32 [xm unit cell

163 e f

Figure A. 2. (a-d) Sequential optical micrographs of condensed water growth on (^Mni3pi6,32Mm^ hydrophobized silicon posts: 40 tall, 8 fim wide, 16x32 |um unit cell

(e,f) Water drops easily roll off the tilted surface when it is dry (e), but stick after condensation (f).

164 a b

Figure A. 3. (a-d) Sequential optical micrographs of condensed water growth on (^MmjpiM2nm^ hydrophobized silicon posts: 40 |am tall, 8 )Lim wide, 16x32 |^m unit cell

C d

Figure A.4. (a-d) Sequential optical micrographs of condensed water growth on (^Mmg^pi6.3.Mm^ hydrophobized silicon posts: 40 |im tall, 8 jim wide, 16x32 fim unit cell

165 r 1*" I*' n Tr

» 4*. jf 4 ak. /Ifc. !f 1*' >f ir

U t\ 4i. Ak, di.- dk. Jk, Jk. Jik ir ^^ -^r ir If Y CI r J. A 4k .Ik..

C

Figure A.5. (a-d) Sequential optical micrographs of condensed water growth on hydrophobized silicon: 32 \im hollow wells. (e,f) Water drops easily roll off the tilted surface when it is dry (e), but stick after condensation (t).

166 b

Figure A. 6. Advancing and receding contact angles on a dimethyldichlorosilane modified silicon wafer: (a) dry surface (b) surface with condensed water. Hysteresis remains low even after condensation (needle diameter is 0.5 mm).

167 A.3 References

(1) Cao, C. T.; Fadeev, A. Y.; McCarthy, T. J. Langmii ir 200\, 17, 757-761.

(2) Fadeev, A. Y.; McCarthy, T. J. Langmiiir 1999, 75, 7238-7243.

(3) Stafford, C. M.; Fadeev, A. Y.; Russell, T. P.; McCarthy, T. J. Langmuiv 2001, 77, 6547-6552.

(4) Bates, F. S.; Fredrickson, G. H. Physics Today 1999, 52, 32-38.

168 APPENDIX B

EXPERIMENTS WITH PDMS

B.l Buckled PPXN on PDMS

Polysiloxanes are interesting polymers because they have an inorganic backbone and hydrocarbon side groups. Silicones, as they are more commonly called, have useful

properties over a wide range of temperatures from -100 ''C to 250 ''C' The Si-0 bond

in the polymer backbone gives silicones a low Tg (-127 ''C) and good low temperature flexibility. Silicones are used for a wide variety of applications from adhesives to health products.

Poly(dimethyl siloxane) (PDMS) is often used in research laboratories because

of its elastomeric properties which make it useful for such applications as a stamp for microcontact printing." The buckling of metal fdms deposited on poly( dimethyl

siloxane) was studied by Bowden et al. ' The metal was evaporated onto PDMS which was heated by the process. As the sample cooled, the PDMS shrank, placing a compressive strain on the rigid metal film which buckled with a wavelength of -20-50

fim. The waves were disordered when the metal was deposited on a flat, unconstrained piece of PDMS, but when the PDMS had raised surface features, the buckling was well ordered.

The buckling phenomenon is not limited to metal films, but can also be observed

for polymer films placed on PDMS. Stafford et al. developed a new technique to determine the elastic moduli of polymeric thin films."* A thin film of polystyrene (PS) was placed on a piece of PDMS which was then compressed causing the PS to buckle

169 normal to the applied force. The buckling wavelength, film thickness, and Poisson's

ratio were used to calculate the film thickness.

That paper gave us the idea to deposit PPXN on PDMS to prepare buckled films.

Recall fi-om Chapter 1 that PPXN is prepared by vapor deposition polymerization. The thought was to polymerize PPXN directly onto PDMS that was stretched and held in place. After the reaction the PDMS would be released and shrink causing a compressive strain in the PPXN film. The PPXN film would buckle in an ordered fashion. Since PPXN can be deposited with very precise control over thickness, the wavelength of the buckling could be tailored. Possible applications could include use as a deformable diffraction grating or a stamp for contact printing.

A series of screening experiments were perfomied. PDMS samples were

prepared using Sylgard 1 84 (Dow Coming). The resin and curing agent were mixed in a 10:1 ratio and cured at 60 °C overnight. PDMS samples were cut to the desired size,

stretched a certain amount, and clamped under tension (Figure B. 1). PPXN was then deposited onto the constrained PDMS at room temperature in a process described in

detail in Chapter 1 . After the reaction the tension was released and optical microscopy was used to observe the buckling of the PPXN film.

a b c Figure B.l. (a) PDMS film held in tension, (b) PPXN deposited on constrained PDMS, (c) Buckled PPXN film on unconstrained PDMS.

170 A piece of PDMS was cut so that the width varied from ~4 mm to 18mm (Figure

B.2a). The PDMS was stretched and the gradient width caused the elongation (tensile strain) to vary across the sample. (The total elongation for the sample could not be

accurately determined using the simple constraint system, but it was <10%.) PPXN (60 nm) was deposited onto the constrained PDMS at room temperature. The tension on the

PDMS film was released and the sample was examined by optical microscopy.

c d

Figure B.2. (a) Gradient width PDMS sample, with arrows indicating the direction the sample was stretched and constrained and letters showing the approximate location of the following optical micrographs, (b,c,d) Buckled PPXN film on PDMS after the tension was released.

The PPXN film has a higher elastic modulus, 2.4 GPa,^ than the PDMS, -1-2

MPa,"* so when the PDMS was unconstrained it placed a compressive strain on the

PPXN causing it to buckle. Figures B.2b,c,d show the buckled PPXN at points where

171 the PDMS sample had a successively smaller width respectively and hence the PPXN was under a progressively higher compressive force. The expected buckling pattern, based on the orientation of the images, was vertical waves with a gradually decreasing wavelength as the compressive strain increased. Figures B.2b,c,d do show vertical buckles with a decreasing wavelength, but they do not extend across the sample and are not well ahgned. AFM showed that the amplitude was ~2 fim. A closer inspection

indicates that the expected buckling is superimposed on another unexpected buckhng

pattern perpendicular to the first and in the same direction as the elongation.

The buckling parallel to the elongation direction was examined by doing optical microscopy of PPXN/PDMS sample still held under tension. A gradient thickness

PDMS sample was prepared by curing the silicone on a tilted surface. (Stretching the gradient thickness sample resuhed in a similar gradient elongation as in the variable width sample.) The PDMS sample was held in tension and PPXN (100 nm) was deposited at room temperature. Figure B.3 shows optical microscopy of the constrained

sample. The underlying PDMS thickness varies for each image so the tensile strain is

different at each point. The PPXN is buckled parallel to the applied elongation direction. The images also show that as the tensile strain increases (thickness decreases) the buckling becomes progressively more ordered. The amplitude was determined to be

1 .2 |im by AFM. When the tension was released, a pattern of overlapping perpendicular and parallel buckling was observed which was similar to that shown in

Figure B.2.

172 a b

Figure B.3. Buckled PPXN on a gradient thickness PDMS sample held under tension in the horizontal direction. The thickness of the underlying PDMS layer varies for each

image: (a) 2.3 mm, (b) 1.6 mm, (c) 1.3 mm, (d) 1 mm.

Figure B.4. Buckled PPXN on PDMS that was not constrained during PPXN deposition.

173 PPXN was deposited on an unconstrained sample of PDMS. Optical microscopy showed isotropic buckling of the PPXN film (Figure B.4). This indicates that the dimensions of the PDMS are different during and after the deposition process.

One possible explanation is that during PPXN deposition, the heat of polymerization causes the surface of the PDMS to expand. The substrate is cooled by water flowing around the glass deposition chamber. In other PPXN depositions the mass of the substrate easily dissipates the heat generated by PPXN. However, the PDMS samples are a few millimeters thick and are good insulators so the temperature at the top surface could be higher than room temperature. The PDMS would shrink upon cooling causing an isotropic compressive strain on the PPXN film resulting in buckling.

This helps explain the previous observations. In Figure B.3 the PDMS was constrained in the horizontal direction. Therefore, during PPXN deposition the PDMS surface could expand more easily in the vertical than horizontal direction. As the

PDMS cooled, a compressive strain was placed on the PPXN film in the vertical

direction which is the reason for the observed orientation of the buckling. There was also a tensile strain gradient in the horizontal direction due to the thickness gradient.

The PDMS could still expand a httle in the horizontal direction at low tensile strains so

upon cooling there was still a small compressive strain in that direction. That is why

the buckled PPXN in Figure B.3a (slight horizontal compressive strain) is less ordered

than in Figure B.3d (even less horizontal compressive strain). Huang et al. perfonned a simulation study of the buckling of a rigid film on a flexible substrate and predicted a similar trend.'' They showed that as the compressive strain became more isotropic, the

174 buckling pattern changed from one similar to Figure B.4 to one even more aligned than

Figure B.4d.

Another possible complication is that during the vapor deposition the monomer could diffuse into the PDMS film creating a gradient between the PPXN and PDMS. A few samples were cryomicrotomed at low temperature and examined by TEM. The

results were inconclusive as it was difficult to find the interface between the PPXN and

PDMS (whether it be gradient or not).

The bottom line is that the dimensional changes of the PDMS surface during

PPXN deposition and the applied tensile strain both contribute to the formation, orientation and order of the buckled PPXN. This complicates the preparation of a well

defined pattern. However, it might be interesting to do further experiments to see what other types of patterns could be prepared.

B.2 PDMS in Anodized Aluminum Membranes

There is a lot of interest in using anodized aluminum membranes ( AAMs) as templates to prepare small objects. For example, an article by Martin in 1994 reviewing

7 the use of various membranes as templates has been cited over 1200 times.' The AAMs are typically microns thick and have close packed, high aspect ratio pores with diameters from tens to hundreds of nanometers (Figure B.5). The diameter, spacing, and order of the pores can be controlled by the reaction conditions used to prepare them.^

Our research group and others have used the membranes to prepare polymer

' objects.*^ ' The polymer can be introduced into the pores in a variety of ways. The

175 membrane can be held in contact with a hot polymer melt whereupon the pores fill by capillary force. The polymer can also be introduced by filling the pores with a polymer solution and then evaporating the solvent. The membrane can be dissolved leaving a forest of polymer rods aligned perpendicular to the substrate.

Figure B.5. (a) Bottom surface and (b) cross section of a commercial AAM.

The ability of a gecko to adhere to flat surfaces and even crawl across the ceiling has captivated scientists.'^ A lot of research is being done to investigate the origins of

the adhesion and mimic it with man-made samples. Examination of the gecko's foot

shows that it has many features with a variety of length scales.'^ There are hundreds of thousands of hairs called setae that are tens of microns long and a few microns wide.

The tips of the setae are further divided into thin hairs called spatulae which are -200

nm in diameter. The superior adhesion of the gecko's foot is believed to be the result of the cumulative van der Waals forces of the millions of spatulae.'"*

Various methods have been tried to mimic the surface features of a gecko's foot.

Geim et al. prepared arrays of 0.2 to 4 |j,m diameter polyimide posts spaced 0.4 to 4.5

176 "^ |am apart. ' Their results showed that the density of the features was the most important

parameter. A 1 cm" array of posts 2 |am high, 0.5 )am in diameter with a periodicity of

1 .6 |im could support 3 N which is the same magnitude as the adhesive force of a gecko's foot, 10 N.

Their sample was prepared by electron beam lithography which is expensive and time consuming. A fast, inexpensive alternative would make applications more practical. The idea for the following experiments was to prepare an array of PDMS posts that could be used as an adhesive. Commercial AAMs have pore diameters of

-200 nm (Figure B.5) which is similar to the spatulae of a gecko's foot. The pores are also densely packed so there would be many van der Waals interactions between the

PDMS and the surface to be adhered to.

The PDMS resin and curing agent were mixed in a ratio of 10 to 1 and a small amount was deposited on a glass slide. A commercial AAM (Whatman Anodisc 13) with -200 nm pores was placed on the PDMS and was immediately wet by the silicone.

After curing for an hour or so at 1 00 °C, the sample was cooled placed in a 1:1 mixture of isopropanol and 50 wt% phosphoric acid in water to dissolve the AAM. (The basic sodium hydroxide solution normal used to dissolve the AAM was found to slowly degrade the silicone.) After a few days the membrane was mostly removed and SEM images were obtained.

177 a b

Figure B.6. SEM cross-section images of PDMS rods prepared in an AAM after the membrane was dissolved. Arrows indicate a film of PDMS connecting all the rods.

The SEM images in Figure B.6 show that a dense array of PDMS rods was prepared. However, the PDMS also coats the top surface of the membrane and all the

rods are connected by a ~4 fj,m thick layer of PDMS at the top surface as indicated by the arrows in Figure B.6b.

PDMS was mixed as before, but was then diluted with various portions of toluene. The mixture was spin coated onto silicon wafers at various speeds to get liquid

PDMS films with different thicknesses. An AAM was placed on each with the idea that

the thin film of PDMS would not have enough volume to fill the whole capillary. After

curing at 1 00 °C for an hour the membrane was dissolved over the course of a few days in phosphoric acid. SEM showed that no matter how thin the spin coated PDMS layer, there was almost always some PDMS on the surface of the top surface of the membrane.

A method was devised to cap one end of the AAM to prevent the PDMS from wetting the entire surface. Short polystyrene rods were grown in one side of the AAM by another member of the research group as describe elsewhere.*^ The open pore side of the membrane was placed on the liquid PDMS mixture, placed under vacuum, and

178 cured at room temperature. The sample was soaked in toluene after curing to remove the polystyrene and then placed in phosphoric acid to dissolve the AAM. SEM images were obtained looking down onto the tops of the rods. Figure B.7 shows that the PDMS rods were definitely separate, as a thick top layer similar to that seen in Figure B.6b

would totally obscure the view. It is also obvious that the rods have all fallen down and

are lying parallel to the substrate. It seems obvious in retrospect, but the modulus of the

PDMS is too low to prepare such high aspect ratio rods, -300: 1 height to diameter. A recent paper concluded that the maximum aspect ratio for a PDMS posts with a diameter of 0.36 |im was ~3.

a b

Figure B.7. Top view of PDMS rods prepared in an AAM, shown after the membrane was dissolved.

It should be possible to make smaller aspect ratio PDMS rods by first growing

polystyrene rods that nearly fill the entire AAM. However, that adds a few extra steps when the goal was to make a quick, easy, and inexpensive mimic of a gecko's foot.

179 B.3 References

(1) Odian, G. Principles ofPolymerization, 3rd ed.; John Wiley & Sons, Inc.: New York, 1991.

(2) Zhao, X. M.; Xia, Y. N.; Whitesides, G. M. Journal of Materials Chemistry 1997,7,1069-1074.

(3) Bowden, N.; Brittain, S.; Evans, A. G.; Hutchinson, J. W.; Whitesides, G. M. Nature 1998, 393, 146-149.

(4) Stafford, C. M.; Harrison, C.; Beers, K. L.; Karim, A.; Amis, E. J.; Vanlandingham, M. R.; Kim, H. C; Volksen, W.; Miller, R. D.; Simonyi, E. E. Nature Materials 2004, J, 545-550.

(5) Beach, W. F.; Lee, C; Bassett, D. R.; Austin, T. M.; Olson, R. In Encyclopedia ofPolymer Science and Engineering, 2nd ed.; Wiley: New York, 1989; Vol. 17, pp 990-1025.

(6) Huang, Z. Y.; Hong, W.; Suo, Z. Physical Review E 2004, 70.

(7) Martin, C. R. Science 1994, 266, 1961-1966.

(8) Chik, H.; Xu, J. M. Mater. Sci. Eng. R-Rep. 2004, 43, 103-138.

(9) Moon, S. I.; McCarthy, T. J. Macromolecules 2003, 36, 4253-4255.

(10) Cepak, V. M.; Martin, C. R. Chem. Mat. 1999, 11, 1363-1367.

(11) Hulteen, J. C; Martin, C. R. J. Mater. Chem. 1997, 7, 1075-1087.

(12) Autumn, K. Am. Scientist 2006, 94, 124-132.

(13) Williams, E. E.; Peterson, J. A. Science 1982, 215, 1509-151 1.

180 Autumn, K.; Sitti, M.; Liang, Y. C. A.; Peattie, A. M.; Hansen, W. R.;

Sponberg, S.; Kenny, T. W.; Fearing, R.; Israelachvili, J. N.; Full, R. J. Proc.

Natl. Acad. Sci. U. S. A. 2002, 99, 12252-12256.

Geim, A. K.; Dubonos, S. V.; Grigorieva, I. V.; Novoselov, K. S.; Zhukov, A.

A.; Shapoval, S. Y. Nat. Mater. 2003, 2, 461-463.

Roca-Cusachs, P.; Rico, F.; Martinez, E.; Toset, J.; Farre, R.; Navajas, D. Langmuiv 2005, 27, 5542-5548.

181 BIBLIOGRPAHY

Adamson, A. W. Physical Chemistiy ofSurfaces, 5th ed.; Wiley: London, 1990.

Andrade, C. K. Z.; Alves, L. M. Current Organic Chemistry 2005, 9, 195-218.

Auspos, L. A.; Bumam, C. W.; Hall, L. A. R.; Hubbard, J. K.; Kirk, W. M.;

Schaefgen, J. R.; Speck, S. B. Journal ofPolymer Science 1955, 75, 19-29.

Auspos, L. A.; Hall, L. A. R.; Hubbard, J. K.; Kirk, W. M.; Schaefgen, J. R.; Speck, S. B. Journal ofPolymer Science 1955, 75, 9-17.

Autumn, K. American Scientist 2006, 94, 124-132.

Autumn, K.; Sitti, M.; Liang, Y. C. A.; Peattie, A. M.; Hansen, W. R.; Sponberg, S.;

Kenny, T. W.; Fearing, R.; Israelachvili, J. N.; Full, R. J. Proceedings of the National Academy of Sciences of the United States ofAmerica 2002, 99, 12252- 12256.

Bam, CD.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G.

J. Am. Chem. Soc. 1989, 777, 321-335.

Bar, G.; Brandsch, R.; Whangbo, M. H. Surf Sci. 1998, 411, L802-L809.

Bar, G.; Brandsch, R.; Whangbo, M. H. Langmuir 1998, 14, 7343-7348.

Barrat, A.; Silberzan, P.; Bourdieu, L.; Chatenay, D. Europhys. Lett. 1992, 20, 633- 638.

Bartell, F. E.; Shepard, J. W. Journal ofPhysical Chemistry 1953, 57, 21 1-215.

Barthlott, W.; Neinhuis, C. Planta 1997, 202, 1-8.

Basnar, B.; Friedbacher, G.; Brunner, H.; Valiant, T.; Mayer, U.; Hoffmann, H. Appl. Surf Sci. 2001, 777,213-225.

Bates, F. S.; Fredrickson, G. H. Physics Today 1999, 52, 32-38.

182 Beach, W. F. Macromoleciiles 1978, 7/, 72-76.

Beach, W. F. In Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed.;

Kroschwitz, J. I.; Howe-Grant, M., Eds.; Wiley: New York, 1998; Vol. Supplement, pp 863-901.

Beach, W. F.; Lee, C; Bassett, D. R.; Austin, T. M.; Olson, R. In Encyclopedia of Polymer Science and Engineering, 2nd ed.; Wiley: New York, 1989; Vol. 17, pp 990-1025.

Beckman, E.; Porter, R. S. J. Polym. Sci. 1987, Part B: Polym. Phys. 25, 1511-1517.

Beecroft, L. L.; Ober, C. K. Chemistiy ofMaterials 1997, 9, 1302-1317.

Berens, A. R.; Huvard, G. S.; Korsmeyer, R. W.: United States Patent 4,820,752, 1989.

Berens, A. R.; Huvard, G. S.; Korsmeyer, R. W.; Kunig, F. W. Journal ofApplied Polymer Science 1992, 46, 231-242.

Bergstrom, L. Colloids and Surfaces 1992, (59, 53-64.

Bergstrom, L.; Bostedt, E. Colloids and Surfaces 1990, 49, 183-197.

Bemett, M. K.; Zisman, W. A. Journal ofPhysical Chemistiy 1962, 66, 1207-&.

Bico, J.; Marzolin, C; Quere, D. Europhysics Letters 1999, 47, 220-226.

Bico, J.; Marzolin, C; Quere, D. Europhysics Letters 1999, 47, 1A?>-14A.

Biebuyck, H. A.; Larsen, N. B.; Delamarche, E.; Michel, B. Ibm Journal ofResearch and Development 1997, 41, 159-170.

Bikerman, J. J. Journal ofPhysical and Colloid Chemistry 1950, 54, 653-658.

Binnig, G.; Quate, C. F.; Gerber, C. Phys. Rev. Lett. 1986, 56, 930-933.

183 Blossey, R. Nature Materials 2003, 2, 301-306.

Bochenkov, V. E.; Stephan, N.; Brehmer, L.; Zagorskii, V. V.; Sergeev, G. B. Colloids and Surfaces A: Physicochemical and Engineering Aspects 2002, 198- 200, 9\\-9\5.

Bowden, N.; Brittain, S.; Evans, A. G.; Hutchinson, J. W.; Whitesides, G. M. Nature 1998, i9i, 146-149.

Brandsch, R.; Bar, G.; Whangbo, M. H. Langmuir 1997, 73, 6349-6353.

Brochard, F. Langmuir 1989, 5, 432-438.

Broer, D. J.; Luijks, W. Journal ofApplied Polymer Science 1981, 26, 2415-2422.

Brow, R. K.; Pantano, C. G. Journal of the American Ceramic Society 1986, 69, 314- 316.

Brown, C. J.; Farthing, A. C. Journal of the Chemical Society 1953, 3270-&.

Brown, G. D.; Watkins, J. J. Abstracts ofPapers of the American Chemical Society 2001.227, U369-U369.

Bumham, N. A.; Chen, X.; Hodges, C. S.; Matei, G. A.; Thoreson, E. J.; Roberts, C.

J.; Davies, M. C; Tendler, S. J. B. Nanotechnology 2003, 14, 1-6.

Bumham, N. A.; Colton, R. J. J Vac. Sci. Technol. A-Vac. Surf Films 1989, 7, 2906- 2913.

Busca, G.; Lorenzelli, V.; Baraton, M. I.; Quintard, P.; Marchand, R. Journal of Molecular Structure 1986, 143, 525-528.

Butt, H. J.; Jaschke, M.; Ducker, W. Bioelectrochem. Bioenerg. 1995, 38, 191-201.

Callies, M.; Quere, D. Soft Matter 2005, 7, 55-61.

184 Cao, C. T.: Fadeev, A. Y.; McCarthy, T. J. Langmuir 2001, 77, 757-761.

Cao, C. T.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society 2000, 279, U440-U440.

Cao, C. T.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society- 2001, 227, U362-U362.

Cappella, B.; Dietler, G. Surf. Sci. Rep. 1999, 34. 1-104.

Carbone, G.; Mangialardi, L. European PhysicalJournal E 2005, 76, 61-16.

Carr, D. W.; Lercel, M. J.; Whelan, C. S.; Craighead, H. G.; Seshadri, K.; Allara, D. L. Journal of Vacuum Science & Technolog\- a- Vacuum Surfaces and Films 1997, 75, 1446-1450.

Cassie, A. B. D. Discussions of the Faraday Society 1948, i, 11-16.

Cassie, A. B. D.; Baxter, S. Transactions of the Faraday Society 1944, 40, 546-551.

Cassie, A. B. D.; Baxter, S. Nature 1945, 755, 21-22.

Cepak, V. M.; Martin. C. R. Chemistiy ofMaterials 1999, 77, 1363-1367.

Chan, C. M. Polymer Surface Modification and Characterization:, Hanser: New York, 1994.

Chan, C. M.; Ko, T. M.; Hiraoka, H. Surf Sci. Rep. 1996, 24, 3-54.

Chaudhury. M. K.; Whitesides, G. M. Science 1992, 256, 1539-1541.

Chen, W.; Fadeev, A. Y.; Hsieh, M. C; Oner, D.; Youngblood, J.; McCarthy, T. J. Langmuir 1999, 75, 3395-3399.

Chen, W.; McCarthy, T. J. Macromolecules 1997, 30, 78-86.

185 Chik, H.; Xu, J. M. Materials Science & Engineering R-Reports 2004, 43, 103-138.

Chiou, J. S.; Barlow, J. W.; Paul, D. R. Journal ofApplied Polymer Science 1985, 30, 2633-2642.

Chiou, J. S.; Barlow. J. W.; Paul, D. R. Journal ofApplied Polymer Science 1985, 30, 3911-3924.

Cho, S. K.; Moon, H. J.; Kim, C. J. Journal of Microelectromechanical Systems 2003, 72, 70-80.

Chovelon, J. M.; Jaffrezicrenault, N.; Clechet, P.; Cros, Y.; Fombon, J. J.; Baraton,

M. I.; Quintard, P. Sensors and Actuators B-Chemical 1991, 4, 385-389.

Chow, S. W.; Loeb, W. E.; White, C. E. Journal oj Applied Polymer Science 1969, 13, 2325-2332.

Clanet, C; Beguin, C; Richard, D.; Quere, D. Journal ofFluid Mechanics 2004, 57 7, 199-208.

Colton, R. J. J. Vac. Sci. Technol. B 2004, 22, 1609-1635.

Cooper, A. I. Journal ofMaterials Chemistiy 2000, 10, 207-234.

Corley, R. S.; Haas, H. C; Kane, M. W.; Livingston, D. I. Journal ofPolymer Science 1954,73,137-156.

Daniel, S.; Chaudhury, M. K.; Chen, J. C. Science 2001, 291, 633-636.

Darhuber, A. A.; Davis, J. M.; Troian, S. M.; Reisner, W. W. Physics ofFluids 2003, 15, 1295-1304.

Darhuber, A. A.; Valentino, J. P.; Davis, J. M.; Troian, S. M.; Wagner, S. Applied Physics Letters 2003, 82, 657-659.

Darhuber, A. A.; Valentino, J. P.; Troian, S. M.; Wagner, S. Journal of Microelectromechanical Systems 2003, 12, 873-879.

186 DeSimone, J. M.; Guan, Z.; Elsbemd, C. S. Science 1992, 257, 945-947.

Dettre, R. H.; Johnson, R. E. In Advances in Chemistiy Series; Gould, R. F., Ed.; American Chemical Society: Washington, 1964; Vol. 43, pp 136-144.

Dettre, R. H.; Johnson, R. E. In Wetting; Society of Chemical Industry: London, 1967; pp 144-155.

Drake. B.; Prater, C. B.; Weisenhom, A. L.; Gould, S. A. C; Albrecht, T. R.; Quate, C. P.; Cannell, D. S.; Hansma, H. G.; Hansma, P. K. Science 1989, 243, 1586- 1589.

Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 239-241.

Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmiiir 1992, 8, 1831-1836.

Duffy, D. C; Gillis, H. L.; Lin, J.; Sheppard, N. F.; Kellogg, G. J. Analytical Chemistiy 1999, 71, 4669-4678.

Dupuis, A.; Yeomans, J. M. Langmuir 2005, 21, 2624-2629.

Dussan, E. B.; Chow, R. T. P. Journal ofFluid Mechanics 1983, 137, 1-29.

El-Sayed, M. A. Accounts of Chemical Research 2001, 34, 257-264.

ElSherbini, A. I.; Jacobi, A. M. Journal of Colloid and Interface Science 2004, 273, 556-565.

ElSherbini, A. I.; Jacobi, A. M. Journal of Colloid and Interface Science 2004, 273, 566-575.

Errede, L. A.; Szwarc, M. Quarterly Reviews 1958, 72, 301-320.

Extrand, C. W. Langmuir 2003, 19, 3793-3796.

Extrand, C. W. Langmuir 2004, 20, 5013-5018.

187 Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 75, 7238-7243.

Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, /5, 3759-3766.

Feng, L.; Li, S. H.; Li, Y. S.; Li, H. J.; Zhang, L. J.; Zhai, J.; Song, Y. L.; Liu, B. Q.; Jiang, L.; Zhu, D. B. Advanced Materials 2002, 14, 1857-1860.

Finot, M. O.; McDermott, M. T. J. Am. Chem. Soc. 1997, 779, 8564-8565.

Fortin, J. B.; Lu, T. M. Chemistry ofMaterials 2002, 14, 1945-1949.

Frisbie, C. D.; Rozsnyai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071-2074.

Furmidge, C. G. Journal of Colloid Science 1962, 77, 309-324.

Furstner, R.; Barthlott, W.; Neinhuis, C; Walzel, P. Langmuir 2005, 27, 956-961.

Furstner, R.; Neinhuis, C; Barthlott, W. Nachrichten Aus Der Chemie 2000, 48, 24-28.

Garbassi, F.; Morra, M.; Occhiello, E. Polymer Surfaces: From Physics to Technology; Wiley: New York, 1994.

Gaynor, J. F.; Desu, S. B. J. Mater. Res. 1994, 9, 3125-3130.

Gaynor, J. F.; Desu, S. B. J. Mater. Res. 1996, 77, 236-242.

Gaynor, J. F.; Desu, S. B.; Senkevich, J. J. Macromolecules 1995, 28, 7343-7348.

Geim, A. K.; Dubonos, S. V.; Grigorieva, L V.; Novoselov, K. S.; Zhukov, A. A.; Shapoval, S. Y. Nature Materials 2003, 2, 461-463.

Gerasimov, G. N.; Nikolaeva, E. V.; Smimova, E. L; Sochilin, V. A.; Trakhtenberg, L. L Doklady Physical Chemistry 2001, 380, ll^-lld.

188 Gerasimov, G. N.; Popova, E. L.; Nikolaeva, E. V.; Chvalun, S. N.; Grigoriev, E. I.;

Trakhtenberg, L. I.; Rozenberg, V. I.; Hopf, H. Macromolecidar ChemistiT and

Physics 1 998, 799, 2 1 79-2 1 84.

Gerasimov, G. N.; Sochilin, V. A.; Chvalun, S. N.; Volkova, L. V.; Kardash, I. Y. Macromolecidar Chemistry and Physics 1996, 197, 1387-1393.

Godovski, D. Y. In Thermal and Electrical Conductivity ofPolymer Materials;

Springer-Verlag: Berlin, 1995; Vol. 1 19, pp 79-122.

Godovsky, D. Y. In Biopolymers/Pva Hydrogels/Anionic Polymerisation Nanocomposites; Springer-Verlag: Berlin, 2000; Vol. 153, pp 163-205.

Good, R. J. Journal of the American Chemical Society' 1952, 74, 5041-5042.

Gorham, W. F. Journal ofPolymer Science-Polymer Chemistry Part A- 1 1966, 4, 3027- 3039.

Gorham, W. F. Advances in Chemistiy Series 1969, 643-&.

Grace, J. M.; Gerenser, L. J. J. Dispersion Sci. Technol. 2003, 24, 305-341.

Green, J. B. D.; McDermott, M. T.; Porter, M. D.; Siperko, L. M. J. Phys. Chem. 1995, 99, 10960-10965.

Greiner, A.; Mang, S.; Schaefer, O.; Simon, P. Acta Polymerica 1997, 48, 1-15.

Greiner, A.; Mang, S.; Schafer, O.; Simon, P. Acta Polymerica 1997, 48, 1-15.

Greiner. A.; Simon, P.: United States Patent 5,310,858, 1994.

Grigg, D. A.; Russell, P. E.; Griffith, J. E. J. Vac. Sci. Technol. A-Vac. Surf Films 1992, JO, 680-683.

Grigor'ev, E. I.; Zav'yalov, S. A.; Chvalun, S. N. Vysokomolekularnye Soedineniya. Ser.A Ser.B Ser.C - Kratkie Soobshcheniya 2003, 45, 1308-1313.

189 Harame, D. L.; Bousse, L. J.; Shott, J. D.; Meindl, J. D. leee Transactions on Electron Devices 1987, 34, 1700-1707.

Harant, A. W.; Whipple, S. G.; Douglas, K.; Bowman, C. N. Journal of Vacuum Science & Technology^ B 2005, 23, 354-358.

Hayashi, K.; Sugimura, H.; Takai, O. Appl. Surf. Sci. 2002, 7

He, B.; Lee, J.; Patankar, N. A. Colloids and Surfaces a-Physicochemical and Engineering Aspects 2004, 248, 101-104.

He, B.; Patankar, N. A.; Lee, J. Langmuir 2003, 19, 4999-5003.

Herrera-Alonso, M.; McCarthy, T. J. Langmuir 2004, 20, 9184-9189.

Hsieh, M. C; Farris, R. J.; McCarthy, T. J. Macromolecules 1997, 30, 8453-8458.

Huang, Z. Y.; Hong, W.; Suo, Z. Physical Review E 2004, 70.

Hulteen, J. C; Martin, C. R. Journal ofMaterials Chemistiy 1997, 7, 1075-1087.

Hyatt, J. A. Journal of Organic Chemistry 1984, 49, 5097-5101.

Isoda, S.; Tsuji, M.; Ohara, M.; Kawaguchi, A.; Katayama, K.-i., 1983; Vol. 24, pp 1155-1161.

Israelachvili, J. N.; Gee, M. L. Langmuir 1989, 5, 288-289.

Itoh, T. Prog Polym. Sci. 2001, 26, 1019-1059.

Iwamoto, R.; Wunderlich, B. J Polym. Sci. Pt. B-Polym. Phys. 1973, 77, 2403-241 1.

Jia, X. Q.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society 2001, 222, U320-U320.

190 Jia, X. Q.; McCarthy, T. J. Langmuir 2003, 19, 2449-2457.

Jisr. R. M.; Rmaile, H. H.; Schlenoff, J. B. Angewandte Chemie-International Edition 2005, 44, 782-785.

Johnson, R. E.; Dettre, R. W. Journal ofPhysical Chemistiy 1964, 68, 1744-&.

Johnson, R. E.; Dettre, R. H. In Advances in Chemistiy Series; Gould, R. F., Ed.; American Chemical Society: Washington, 1964; Vol. 43, pp 136-144.

Johnson, R. E.; Dettre, R. H. In Surface and Colloid Science; Matijevic, E., Ed.; Wiley: New York, 1969; Vol. 2, pp 85-153.

Kaufman, M. H.; Mark, H. P.; Mesrobian, R. B. Journal ofPolymer Science 1954, 13, 3-20.

Kawasaki, K. Journal of Colloid Science 1960, 75, 402-407.

Kazarian, S. G.; Brantley, N. H.; Eckert, C. A. Chemtech 1999, 29, 36-41.

Kendall, J. L.; Canelas, D. A.; Young, J. L.; DeSimone, J. M. Chemical Reviews 1999, 99, 543-563.

Kim, H. C; Jia, X. Q.; Stafford, C. M.; Kim, D. H.; McCarthy, T. J.; Tuominen, M.;

Hawker, C. J.; Russell, T. P. Adv. Mater. 2001, 13, 795-797.

Kim, H. Y.; Lee, H. J.; Kang, B. H. Journal of Colloid and Interface Science 2002, 247, 372-380.

Kim, J.; Kim, C.-J. In Proc. IEEE Int. Conf MEMS, 2002; pp 479-482.

Kirby, C. F.; McHugh, M. A. Chemical Reviews 1999, 99, 565-602.

Kozlov, M.; McCarthy, T. J. Langmuir 2004, 20, 9170-9176.

191 Kozlov, M.; Quarmyne, M.; Chen, W.; McCarthy, T. J. Macromolecules 2003, 36, 6054-6059.

Krukonis, V. J. Polymer News 1985, 77, 7-16.

Krupke, R.; Malik, S.; Weber, H. B.; Hampe, O.; Kappes, M. M.; von Lohneysen, H.

- Nano Letters 2002, 2, 1 1 6 1 1 1 64.

Kubo, M.; Iwatsuki, S. In Polymeric Materials Encyclopedia; Salamone, J. C, Ed.; CRC Press: Boca Raton, 1996.

Kubo, M.; Toyota, K.; Itoh, T.; Iwatsuki, S. Polym. Bull. 1995, 35, 615-619.

Kubo, S.; Wunderlich, B., 1972; Vol. 10, pp 1949-1966.

Kumar, A.; Biebuyck, H. A.; Abbott, N. L.; Whitesides, G. M. Journal of the American Chemical Society 1992, 114, 9188-9189.

Kumar, A.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1994, 10, 1498-151 1.

Kung, E.; Lesser, A. J.; McCarthy, T. J. Macromolecules 1998, 31, 4160-4169.

Kung, E.; Lesser, A. J.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society \999, 2 18, mil-m\l.

Kung, E.; Lesser, A. J.; McCarthy, T. J. Macromolecules 2000, 33, 8192-8199.

Lafuma, A.; Quere, D. Nature Materials 2003, 2, 457-460.

Lahann, J.; Balcells, M.; Rodon, T.; Lee, J.; Choi, I. S.; Jensen, K. P.; Langer, R. Langmuir 2002, 18, 3632-3638.

Lahann, J.; Klee, D.; Hocker, H. Macromol. Rapid Commun. 1998, 19, 441-444.

Lahann, J.; Klee, D.; Thelen, H.; Bienert, H.; Vorwerk, D.; Hocker, H. J. Mater. Sci.- Mater. Med 1999, 10, 443-448.

192 .

Lahann, J.; Langer, R. Macromoleciiles 2002, 35, 4380-4386.

Lai, S.; Wang, S. N.; Luo, J.; Lee, L. J.; Yang, S. T.; Madou, M. J. Analytical Chemlstiy 76, 1832-1837.

Lam, P.; Wynne, K. J.; Wnek, G. E. Langmiiir 2002, J 8, 948-951.

Larena, A.; Pinto, G. Spectr. Lett. 1992, 25, 447-461

Lee, G. U.; Chrisey. L. A.; Oferrall, C. E.; Pilloff, D. E.; Turner, N. H.; Colton, R. J. Israel Journal ofChemistiy 1996, 36, 81-87.

Lehto, A.; Kojola, H.; Lovgren, T.; Lonnbreg, H.: World Patent 99/54730, 1999.

Lercel, M. J.; Tiberio. R. C; Chapman, P. P.; Craighead, H. G.; Sheen, C. W.; Parikh, A. N.; Allara, D. L. Journal of Vacuum Science & Technology B 1993, 77,2823-2828.

Lora, M.; McHugh, M. A. Fluid Phase Equilibria 1999, 757, 285-297.

Love, J. C; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chemical

Reviews 2005, 105, 1 103-1 169.

Lu, Z. X.; McCarthy, T. J. Abstracts ofPapers of the American Chemical Society 2004, 227, U379-U379.

Ma, M.; Hill, R. M.; Lowery, J. L.; Fridrikh, S. V.; Rutledge, G. C. Langmuir 2005, 27, 5549-5554.

Macdougall, G.; Ocki-ent, C. Proceedings of the Royal Society ofLondon. Series A, Mathematical and Physical Sciences 1942, J 80, 151-173.

Maoz, R.; Sagiv, J. Journal of Colloid and Interface Science 1984, 100, 465-496.

Marmur, A. Langmuir 2003, 79, 8343-8348.

193 Marmur, A. Langmuir 2004, 20, 35 1 7-3519.

Martin, C. R. Science 1994, 266, 1961-1966.

Martini, D.; Shepherd, K.; Sutcliffe, R.; Kelber, J.; Edwards, H.; San Martin, R. Applied Surface Science 1999, 141, 89-100.

McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction: Principles and Practice, 2nd ed.; Butterworth-Heinemann: Boston, 1994.

Milinazzo, F.; Shinbrot, M. Journal of Colloid and Interface Science 1988, 727, 254- 264.

Moon, S. I.; McCarthy, T. J. Macromolecules 2003, 36, 4253-4255.

Murthy, N. S.; Kim, H. G. Polymer 1984, 25, 1093-1096.

Nakagawa, T.; Ogawa, K.; Kurumizawa, T.; Ozaki, S. Jpn. J. Appl. Phys. Part 2 - Lett. 1993, 32, L294-L296.

Narhe, R. D.; Beysens, D. A. Physical Review Letters 2004, 93.

Nishino, T.; Meguro, M.; Nakamae, K.; Matsushita, M.; Ueda, Y. Langmuir 1999, 75, 4321-4323.

Nowlin, T. E.; Smith, D. F., h. Journal ofApplied Polymer Science 1980, 25, 1619- 1632.

Noy, A.; Frisbie, C. D.; Rozsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 777, 7943-7951.

Noy, A.; Sanders, C. H.; Vezenov, D. V.; Wong, S. S.; Lieber, C. M. Langmuir 1998, 14, 1508-1511.

Noy, A.; Vezenov, D. V.; Lieber, C. M. Annu. Rev. Mater. Sci. 1997, 27, 381-421.

194 Odian, G. Principles ofPolymerization, 3rd ed.; John Wiley & Sons, Inc.: New York, 1991.

Ogbuji, L. Journal of the American Ceramic Society 1995, 78, 1279-1284.

Ogbuji, L.; Bryan, S. K. Journal of the American Ceramic Society 1995, 78, 1272-1278.

Ogbuji, L. U. T.; Jayne, D. T. Journal of the Electrochemical Society 1993, 140, 759- 766.

Oh, C: United States Patent 5,904,824, 1999.

Okada, K.; Fukuyama, K.; Kameshima, Y. Journal of the American Ceramic Society 1995,75,2021-2026.

Onda, T.; Shibuichi, S.; Satoh, N.; Tsujii, K. Langmuir 1996, 12, 2125-2127.

Ondarcuhu, T. Journal De Physique li 1995, 5, 227-241.

Ondarcuhu, T.; Veyssie, M. Journal De Physique li 1991, 1, 75-85.

O'Neil, A.; Watkins, J. J. Green Chemistiy 2004, 6, 363-368.

Oner, D.: University of Massachusetts, 2001.

Oner, D.; McCarthy, T. J. Langmuir 2000, 16, 1111-11%!.

Ou, J.; Perot, B.; Rothstein, J. P. Physics ofFluids 2004, 16, 4635-4643.

Ovemey, R. M.; Meyer, E.; Frommer, J.; Brodbeck, D.; Luthi, R.; Howald, L.;

Guntherodt, H. J.; Fujihira, M.; Takano, H.; Gotoh, Y. Nature 1992, 359, 133- 135.

Ovemey, R. M.; Meyer, E.; Frommer, J.; Guntherodt, H. J.; Fujihira, M.; Takano, H.; Gotoh, Y. Langmuir 1994, 10, 1281-1286.

195 Pallandre, A.; Glinel, K.; Jonas, A. M.; Nysten, B. Nemo Letters 2004, 4, 365-371.

Palzer, S.; Sommer, K.; Hiebl, C. Chemical Engineering & Technology 2003, 26, 962- 966.

Pataiikar, N. A. Langmidr 2003, 79, 1249-1253.

Patankar, N. A. Langmitir 2004, 20, 8209-8213.

Patankar, N. A. Langmiiir 2004, 20, 7097-7102.

Perrut, M. Industrial & Engineering Chemistiy Research 2000, 39, 4531-4535.

Petzow, G.; Herrmann, M. High Performance Non-Oxide Ceramics li 2002, 102, 47- 167.

Petzow, G.; Herrmann, M. In High Performance Non-Oxide Ceramics li; Springer- Verlag: Berlin, 2002; Vol. 102, pp 47-167.

Phuvanartnuruks, V.; McCarthy, T. J. Macromolecules 1998, 31, 1906-1914.

Prausnitz, J. M.; Lichtenthaler, R. N.; Azevedo, E. G. Molecular Thermodynamics of Fluid-Phase Equilibria, 2nd ed.; Prentice-Hall: New Jersey, 1986.

Quere, D. Physica a-Statistical Mechanics and Its Applications 2002, 313, 32-46.

Quere, D. Reports on Progress in Physics 2005, 68, 2495-2532.

Quere, D.; Laftima, A.; Bico, J. Nanotechnolog)' 2003, 14, 1 109-1 1 12.

Raghavan, D.; VanLandingham, M.; Gu, X.; Nguyen, T. Langmiiir 2000, 16, 9448- 9459.

Raider, S. I.; Flitsch, R.; Aboaf, J. A.; Pliskin, W. A. Journal of the Electrochemical Society 1976, 123, 560-565.

196 Rayleigh. L. Philosophical Magazine 1890, 30, 386-400.

Reyes, D. R.; lossifidis, D.; Auroux, P. A.; Manz, A. Analytical ChemistiT 2002, 74, 2623-2636.

Riley, F. L. Journal of the American Ceramic Society 2000, 83, 245-265.

Roca-Cusachs, P.; Rico, F.; Martinez, E.; Toset, J.; Farre, R.; Navajas, D. Langmuir 2005, 27, 5542-5548.

Rogojevic, S.; Moore, J. A.; Gill, W. N. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1999, 77, 266-274.

Rye, R. R.; Nelson, G. C; Dugger, M. T. Langmuir 1997, 13, 2965-2972.

Sagiv, J. Journal of the American Chemical Society 1980, 102, 92-98.

Sand, M. L.: United States Patent 4,598,006, 1986.

Sasaki, K.; Koike, Y.; Azehara, H.; Hokari, H.; Fujihira, M. Appl. Phys. A-Mater. Sci. Process. 1998, 66, S1275-S1277.

Senden, T. J.; Drummond, C. J. Colloids and Surfaces a-Physicochemical and Engineering Aspects 1995, 94, 29-51.

Senkevich, J. J.; Desu, S. B. Chemistiy ofMaterials 1999, 77, 1814-1821.

Senkevich, J. J.; Yang, G. R.; Lu, T. M. Colloid Surf. A-Physicochem. Eng. Asp. 2003, 214, 119-126.

Senkevich, J. J.; Yang, G. R.; Lu, T. M. Colloid Surf. A-Physicochem. Eng. Asp. 2003, 216, 167-173.

Sergeev, G.; Zagorsky, V.; Petrukhina, M. Journal ofMaterials Chemistiy 1995, 5, 31- 34.

197 Sergeev, G.; Zagorsky, V.; Petrukhina, M.; Zavyalov, S.; Grigorev, E.; Trakhtenberg,

L. Analytical Communications 1997, 34, 1 13-1 14.

Shenhar, R.; Norsten, T. B.; Rotello, V. M. Advanced Materials 2005, /7, 657-669.

Shirtcliffe, N. J.; McHale, G.; Newton, M. I.; PeiTy, C. C. Langmnir 2005, 27, 937-943.

Shiu, J.-Y.; Kuo, C.-W.; Chen, P.; Mou, C.-Y. ChemisDy of Materials 2004, 76, 561- 564.

Shiu, J. Y.; Kuo, C. W.; Chen, P. L.; Mou, C. Y. Chemistiy of Materials 2004, 76, 561-564.

Shuttleworth, R.; Bailey, G. L. J. Transactions ofthe Faraday Society 1948, 44, 16-22.

Simon, P.; Greiner, A. Polymer Journal 1992, 24, 1317-1320.

Simon, P.; Mang, S.; Hasenliindl, A.; Gronski, W.; Greiner, A. Macromolecules 1998, 31, 8775-8780.

Sochilin, V. A.; Mailyan, K. A.; Alexandrova, L. N.; Nikolayev, A. A.; Pebalk, A.

v.; Kardash, I. Y. Doklady Akademii Nauk Sssr 1991, 319, 173-176.

Sochilin, V. A.; Pebalk, A. V.; Semenov, V. I.; Kardash, I. Y. Vysokomolekulyarnye Soedineniya Seriya A 1991, 3i, 1536-1542.

Stafford, C. M.; Fadeev, A. Y.; Russell, T. P.; McCarthy, T. J. Langmnir 2001, 77, 6547-6552.

Stafford, C. M.; Harrison, C; Beers, K. L.; Karim, A.; Amis, E. J.; Vanlandingham, M. R.; Kim, H. C; Volksen, W.; Miller, R. D.; Simonyi, E. E. Nature

Materials 2004, 3, 545-550.

Stewart, D. D.; Union Carbide Corporation: United States Patent 3,600,216, 1971.

Stone, H. A.; Kim, S. Aiche Journal 2m\ , 47, 1250-1254.

198 Stone, H. A.; Stroock, A. D.; Ajdari, A. Annual Review of Fluid Mechanics 2004, 36, 381-411.

Sung, M. M.; Kluth, G. J.; Maboudian, R. Journal of Vacuum Science & Technology a

Vacuum Surfaces and Films 1 999, 1 7, 540-544.

Suppes, G. J.; McHugh, M. A. Journal of Chemical and Engineering Data 1989, 34, 310-312.

Suzuki, S.; Nakajima, A.; Kameshima, Y.; Okada, K. Surface Science 2004, 557, L163-L168.

Szwarc, M. Nature 1947, 160, 403.

Szwarc, M. Discussions of the Faraday Societ}- 1947, 2, 46.

Szwarc, M. Journal ofPolymer Science 1951, 6, 319-329.

Takada, M.; Tanigaki, M.; Ohshima, M. Polymer Engineering and Science 2001, 41, 1938-1946.

Takano, H.; Kenseth, J. R.; Wong, S. S.; O'Brien, J. C; Porter, M. D. Chem. Rev. 1999, 99, 2845-2890.

Tamayo, J.; Garcia, R. Langmuir 1996, 12, 4430-4435.

Thundat, T.; Zheng, X. Y.; Chen, G. Y.; Sharp, S. L.; Warmack, R. J.; Schowalter, L ].Appl. Phys. Lett. 1993, 6i, 2150-2152.

Tolstopyatov, E. M. Journal ofPhysics D: Applied Physics 2002, 35, 1516-1525.

Tomasko, D. L.; Li, H. B.; Liu, D. H.; Han, X. M.; Wingert, M. J.; Lee, L. J.; Koelling, K. W. Industrial & Engineering Chemistiy Research 2003, 42, 6431- 6456.

Torkkeh, A.; Ph.D. Dissertation, Helsinki University of Technology. 2003.

199 Tripp, C. P.; Hair, M. L. Langmiiir 1995, 77, 1215-1219.

Tripp, C. P.; Hair, M. L. Langmiiir 1995, 77, 149-155.

Ulman, A. Chem. Rev. 1996, 96, 1533-1554.

Wang, W. C. v.; Kramer, E. J.; Sachse, W. H. Journal ofPolymer Science Part B-

Polymer Physics 1982, 20, 1 37 1 - 1 384.

Washizu, M. leee Transactions on Industiy Applications 1998, 54, 732-737.

Wasserman, S. R.; Tao, Y. T.; Whitesides, G. M. Langmuir 1989, 5, 1074-1087.

Watkins, J. J.; Ph.D. Dissertation, University of Massachusetts, 1997.

Watkins, J. J.; McCarthy, T. J. Macromolecules 1994, 27, 4845-4847.

Watkins, J. J.; McCarthy, T. J. Chemistiy ofMaterials 1995, 7, 1991-&.

Watkins, J. J.; McCarthy, T. J. Macromolecules 1995, 28, 4067-4074.

Weisenhom, A. L.; Hansma, P. K.; Albrecht, T. R.; Quate, C. F. AppL Phys. Lett. 1989, 54, 2651-2653.

Wells, S. L.; DeSimone, J. Angewandte Chemie-International Edition 2001, 40, 519- 527.

Wenzel, R. N. Industrial and Engineering Chemistiy 1936, 28, 988-994.

Williams, E. E.; Peterson, J. A. Science 1982, 275, 1509-151 1.

Wissinger, R. G.; Paulaitis, M. E. Journal ofPolymer Science Part B-Polymer Physics 1987,25,2497-2510.

Wixforth, A. Superlattices and Microstructures 2003, 33, 389-396.

200 Wixforth, A.: United States Patent 6,777,245, 2004.

Woods, H. M.; Silva, M.; Nouvel, C; Shakesheff, K. M.; Howdle. S. M. Journal of

Materials Chcmistiy 2004, 14, 1 663- 1 678.

Yabu, H.; Takebayashi, M.; Tanaka, M.; Shimomura, M. Langnniir 2005, 27, 3235- 3237.

Yilgor, I.; McGrath, J. E. Polymer Bulletin 1984, 12, 491-497.

Yilgor, I.; McGrath, J. E. Polymer Bulletin 1984, 72, 499-506.

Yoshimitsu, Z.; Nakajima, A.; Watanabe, T.; Hashimoto, K. Langmuir 2002, 7^, 5818-5822.

Young, J. L.; DeSimone, J. M. Pure and Applied Chemistry 2000, 72, 1357-1363.

Young, T. Philosophical Transactions of the Royal Society ofLondon 1805, 95, 65-87.

Youngblood, J. P.; McCarthy, T. J. Macromolecules 1999, 32, 6800-6806.

Yu, Q. S.; Deffeyes, J.; Yasuda, H. Prog. Org. Coat. 2001, 47, 247-253.

Zagorskii, V. V.; Ivashko, S. V.; Bochenkov, V. E.; Sergeev, G. B. Nanostructured Materials 1999, 72, 863-866.

Zhao, B.; Moore, J. S.; Beebe, D. J. Science 2001, 297, 1023-1026.

Zhao, B.; Moore, J. S.; Beebe, D. J. Analytical Chemistiy 2002, 74, 4259-4268.

Zhao, B.; Viemes, N. O. L.; Moore, J. S.; Beebe, D. J. Journal of the American Chemical Society 2002, 724, 5284-5285.

Zhao, X. M.; Xia, Y. N.; Whitesides, G. M. Journal ofMaterials Chemistiy 1997, 7, 1069-1074.

201 Zheng, Q. S.; Yu, Y.; Zhao, Z. H. Langmuir 2005, 27, 12207-122

Zoval, J. v.; Madou, M. J. Proceedings of the IEEE 2004, 92, 140

202