Reproduced in Part with Permission from Badge, I

Reproduced in part with permission from

Badge, I., Sethi, S., Dhinojwala, A. Langmuir 2011, 27, 14726–14731

TUNING WETTABILITY AND ADHESION OF STRUCTURED SURFACES

A Dissertation

Presented to

The Graduate Faculty of The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Ila Badge

May, 2014

TUNING WETTABILITY AND ADHESION OF STRUCTURED SURFACES

Ila Badge

Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Ali Dhinojwala Dr. Coleen Pugh

______Committee Member Dean of the College Dr. Mesfin Tsige Dr. Stephen Z.D. Cheng

______Committee Member Dean of the Graduate School Dr. Li Jia Dr. George R. Newkome

______Committee Member Date Dr. Darrell Reneker

______Committee Member Dr. Robert Weiss

ABSTRACT

Structured surfaces with feature size ranging from a few micrometers down to nanometers are of great interest in the applications such as design of anti-wetting surfaces, tissue engineering, microfluidics, filtration, microelectronic devices, anti- reflective coatings and reversible adhesives. A specific surface property demands particular roughness geometry along with suitable surface chemistry. Plasma Enhanced

Chemical Vapor Deposition (PECVD) is a technique that offers control over surface chemistry without significantly affecting the roughness and thus, provides a flexibility to alter surface chemistry selectively for a given structured surface. In this study, we have used PECVD to fine tune wetting and adhesion properties. The research presented focuses on material design aspects as well as the fundamental understanding of wetting and adhesion phenomena of structured surfaces.

In order to study the effect of surface roughness and surface chemistry on the surface wettability independently, we developed a model surface by combination of colloidal lithography and PECVD. A systematically controlled hierarchical roughness using spherical colloidal particles and surface chemistry allowed for quantitative prediction of contact angles corresponding to metastable and stable wetting states. A well-defined roughness and chemical composition of the surface enabled establishing a correlation between theory predictions and experimental measurements.

iii

We developed an extremely robust superhydrophobic surface based on Carbon-

Nanotubes (CNT) mats. The surface of CNTs forming a nano-porous mesh was modified using PECVD to deposit a layer of hydrophobic coating (PCNT). The PCNT surface thus formed is superhydrophobic with almost zero contact angle hysteresis. We demonstrated that the PCNT surface is not wetted under steam condensation even after prolonged exposure and also continues to retain its superhydrophobicity after multiple frosting- defrosting cycles. The anti-wetting behavior of PCNT surface is consistent with our model predictions, derived based on thermodynamic theory of wetting.

The surface of gecko feet is a very unique natural structured surface. The hierarchical surface structure of a Gecko toe pad is responsible for its reversible adhesive properties and superhydrophobicity. van der Waals interactions is known to be the key mechanism behind Gecko adhesion. However, we found that the wettability, thus the surface chemistry plays a significant role in Gecko adhesion mechanism, especially in the case of underwater adhesion. We used PECVD process to deposit a layer of coating with known chemistry on the surface of sheds of gecko toes to study the effect that wettability of the toe surface has on its adhesion.

In summary, we demonstrated that PECVD can be effectively used as means of surface chemistry control for tunable structure-property relationship of three types of structured surfaces; each having unique surface features.

DEDICATION

I would like to dedicate this dissertation to my parents, Ravindra and Vasanti

Badge, for their unconditional love, support and words of encouragement throughout my

Ph.D.

ACKNOWLEDGEMENTS

I would like to take this opportunity to thank several individuals who have contributed to my success in the graduate school and helped make the Ph.D. experience one of the most valuable ones in my life.

First and foremost, I would like to thank my advisor, Dr. Ali Dhinojwala, for his valuable guidance and encouragement at every step of graduate school. None of the work presented here would have been possible without his support. Under his guidance, I have learnt to be more disciplined and organized, to thoroughly and critically analyze the problems and learn from failures. Having him as my mentor has made me not only a better scientist but also a better person in the past few years.

I would like to acknowledge and express my gratitude towards my committee members, Dr. Mesfin Tsige, Dr. Li Jia, Dr. Darrell Reneker and Dr. Robert Weiss, for agreeing to be on my dissertation committee and their valuable advice and feedback from time to time. I would like to thank Dr. Li Jia and his former graduate student Dr. Sarang

Bhawalkar for collaborating in the project of study of wettability using patterns formed with colloidal lithography. I would also like to thank Dr. Peter Niewiarowski and Alyssa

Stark for collaboration in the work of gecko adhesion and also for numerous valuable discussions.

I would like to thank Dr. Bojie Wang for training me to use SEM and TEM techniques, which I heavily relied upon in my research. I would also like to thank Ed

Laughlin for his excellent machine work and Jack Gillespie for his help with glass related work. I would like to specially mention my gratitude towards both of them for their valuable ideas and suggestions in part designs. I would like to thank Goodyear Tire and

Rubber Company for the funding and giving me an opportunity to work as an industrial intern.

I would like to thank all my group members, past and present, for their help in my research. The communications on a daily basis and exchange of ideas in small chats have always helped me solve problems related to the research. I would like to specifically mention my gratitude to Dr. Sunny Sethi for teaching me carbon nanotube synthesis technique and mentoring me in the initial years of my Ph.D. I would like to thank Dr.

Frederic Siffer for training me to use the PECVD set up. I would specially like to thank

Dr. Vasav Sahni, Dr. Michael Heiber, Mena Klittich and Alyssa Stark for proof reading manuscripts and abstracts and their valuable feedback for the same. I would like to thank

Mena Klittich for her help in proof reading my dissertation as well.

Last but not the least I would like to thank my friends and family, for their unconditional love and support, for sharing the moments of happiness and for helping me get through tough times. I would like to thank my parents and my younger brother,

Mayur Badge for believing in me and their constant love and support. I would specially like to thank friends in Akron who have been a family away from home. I thank Kushal

Bahl, Saurabh Batra, Sarang Bhawalkar, Diya Bandyopadhyay and Poonam Songar for all their support and being just wonderful friends.

vii

TABLE OF CONTENTS Page

LIST OF TABLES ...... xii

LIST OF FIGURES ...... xiv

CHAPTERS

I. INTRODUCTION ...... 1

II. BACKGOUND ...... 5

2.1 Wetting ...... 5

2.1.1 Wetting by water ...... 6

2.2 Contact Angle ...... 7

2.2.1 Background ...... 7

2.2.2 Young’s Equation ...... 8

2.2.3 Static and dynamic contact angles ...... 9

2.2.4 Surface energy versus surface roughness ...... 12

2.2.5 Wenzel model ...... 15

2.2.6 Cassie-Baxter model ...... 17

2.3 Designing surface with desired wettability ...... 19

2.3.1 Superhydrophobic surfaces ...... 20

viii

2.3.2 Superwetting surfaces ...... 33

2.3.3 Oleophobic and omniphobic surfaces ...... 34

2.4 Thermodynamic stability of wetting states ...... 38

2.5 Loss of superhydrophobicity ...... 41

2.6 Adhesion of structured surfaces...... 45

2.6.1 Mechanical adhesion ...... 46

2.6.2 Chemical adhesion ...... 46

2.6.3 Work of adhesion ...... 47

2.6.4 van der Waals interactions ...... 48

2.6.5 Pressure sensitive adhesives (PSA) ...... 49

2.7 Gecko adhesion ...... 50

2.7.1 Effect of environmental factors ...... 53

2.7.2 Self cleaning of gecko toe pads ...... 55

2.8 Plasma Enhanced Chemical Vapor Deposition (PECVD) or Plasma polymerization ...... 56

2.8.1 PECVD nomenclature ...... 57

2.8.2 PECVD of organic precursors ...... 59

2.8.2.1 Kinetics of the PECVD reaction (mechanism of plasma polymerization) ...... 61

2.8.2.2 PECVD of fluorinated molecules ...... 64

2.8.2.3 PECVD of organosilicon compounds ...... 68

2.8.2.4 PECVD of maleic anhydride ...... 70

III. EXPERIMENTAL SECTION ...... 73

3.1 Fabrication of single and dual layer particle patterns ...... 73

3.1.1 HNCP patterning using colloidal self-assembly of spherical particles 73

3.1.2 Morphology characterization of HNCP patterns ...... 74

3.1.3 PECVD coatings of HNCP pattern ...... 75

3.1.4 AFM measurements to measure R2 of HNCP patterns ...... 77

3.1.5 XPS analysis for chemical composition of the PECVD coating ...... 78

3.1.6 Wettability studies of PECVD coated HNCP patterns ...... 78

3.1.7 Wettability model calculations for HNCP patterns ...... 79

3.2 Carbon nanotube (CNT) based coatings ...... 80

3.2.1 Synthesis of Carbon nanotubes ...... 80

3.2.2 PECVD coating on the surface of CNT mats (formation of PCNT) ...81

3.2.3 Characterization of CNT and PCNT structures ...... 82

3.2.4 Wetting studies of CNT and PCNT ...... 83

3.2.5 Wetting model calculations for CNT and PCNT surfaces ...... 85

3.3 Gecko adhesion : The effect of surface wettability ...... 86

3.3.1 Gecko shed sample preparation ...... 86

3.3.2 PECVD coating on sheds ...... 86

3.3.3 Surface characterization of PECVD coated sheds ...... 87

3.3.4 Wetting and adhesion tests...... 87

3.3.5 Statistical analysis ...... 90

IV. TUNING SURFACE WETTABILITY USING ORDERED ARRAYS OF PARTICLES ...... 91

4.1 Motivation...... 91

4.2 Results and discussion ...... 94

4.2.1 Single Layer Roughness ...... 94

4.2.2 XPS analysis for surface chemical composition of the PECVD coated HNCP patterns ...... 105

4.2.3 Dual Roughness ...... 110

4.3 Conclusion ...... 121

V. CARBON NANOTUBES (CNT) BASED ROBUST SUPERHYDROPHOBIC SURFACES ...... 123

5.1 Motivation...... 123

5.2 Results and discussion ...... 125

5.2.1 Morphology and chemical composition of CNT and PCNT ...... 125

5.2.2 Wettability of CNT versus PCNT ...... 132

5.2.2.1 Steamphobicity ...... 132

5.2.2.2 Low temperature condensation ...... 138

5.2.2.3 Anti-frost properties ...... 141

5.3 Conclusion ...... 142

VI. ADHESION AND WETTING OF GECKO TOE PADS : THE ROLE OF SURAFCE CHEMISTRY ...... 143

6.1 Motivation...... 143

6.2 Results and discussion ...... 148

6.2.1 Effect of contact surface wettability (surface chemistry) ...... 148

6.2.1.1 Whole-animal Adhesion ...... 148

6.2.1.2 Smooth Surface Adhesion Model ...... 150

6.2.1.3 Patterned Surface Adhesion Model ...... 152

6.2.2 Effect of gecko hair wettability (surface chemistry) ...... 160

6.3 Conclusion ...... 176

VII. SUMMARY AND CONCLUSIONS ...... 179

BIBLIOGRAPHY ...... 183

APPENDIX ...... 200

xii

LIST OF TABLES Table Page

4.1. Diameter, height and calculated α for the spheres coated using PECVD………….98

4.2. Peak position and assignments for high resolution XPS C1s spectra of PECVD layer of 1H,1H,2H-perfluoro-1-dodecene on the surface of flat Si wafer and HNCP array of spherical particles……………………………………………………………………….106

4.3. Peak position and assignments for high resolution XPS C1s spectra of PECVD layer of HMDSO on the surface of flat Si wafer and HNCP array of spherical particles...... 108

5.1. Chemical quantification of the plasma deposited coating using XPS analysis: C1s peak assignments, functional group retention and atomic concentrations……………...130

6.1. Contact angles of water and n-hexadecane on all four surfaces used in whole- animal experiments and modeling. Errors are means  1 SEM………………………..149

6.2. Ratios of wet to dry adhesion on all four surfaces used in whole-animal experiments and modeling, the ratios calculated based on experimental results are included for comparison as well……………………………………………………………………...153

6.3. Summary of atomic composition and atomic ratios of B-S, M-S and F-S surfaces calculated using survey XPS spectra…………………………………………………...163

6.4. Wwet:Wdry calculated from experimental data for adhesion measurements on glass and OTS-SAM surface……………………………………………………………………...167

6.5. Summary of water contact angles, The intrinsic contact angles (θY) are measured on flat, control samples and apparent contact angles (θapp) are measured on blank and PECVD coated shed samples. The inset figures show water in contact with different surafces, it beads up to form Cassie-Baxter wetting state on the surafces of B-S and F-S and spreads completely on the surface of M-S (scale bar : 500 µm)…………………...169

A1. Table of ratios for the non-tetrad patterned surface in each of the four wetting cases and on each test surface………………………………………………….215

A2. The parameters used to calculate Hamaker constants for the absorption frequency (νe) of 3 x 1015 s-1 …………………………………………………………………………...216

xiii

A3. Hamaker constant values calculated for different contact surfaces; the subscripts 1, 2 and 3 correspond to “gecko hair-like” n-hexadecane, substrate (glass, plexiglass and PTFE) and air (or water) respectively…………………………………………………..216

xiv

LIST OF FIGURES Figure Page

2.1. Schematic representation of equilibrium contact angle of a liquid droplet on a solid surface……………………………………………………………………………………..7

2.2. Schematic representation of (a) advancing contact angle ( ) and (b) receding contact angle ( ) using volume addition/removal method, (c) Schematic representation of tilt plate method for measuring CAH…………………………………………………10

2.3. (a) Image of a water droplet beaded up on the surface of lotus leaf, (b) SEM image of hierarchical structure of lotus leaf surface consisting of waxy bumps, every bump is covered with finer structures called wax tubules and (c) high resolution SEM image of waxy tubules……………………………………………………………………………..13

2.4. A schematic representation of Wenzel state of a liquid droplet on rough solid surface, is the apparent contact angle measured on the surface called as “Wenzel angle”…...16

2.5. Schematic representation of Cassie-Baxter state of a superhydrophobic surface…...17

2.6. Theoretical curves representing the apparent contact angle (θR) as a function of equilibrium contact angle (θ), calculated using Wenzel and Cassie-Baxter models…….18

2.7. Examples of natural superhydrophobic surfaces. Different types of structural patterns observed on (a) lotus leaf surface, (b) taro leaf surface, (c) rice leaf, (d) surface of butterfly wing, (e) stenocara beetle’s back shell surface and (f) mosquito eye structure..20

2.8. Examples of single level surface roughness. The ordered patterns and systematically controlled roughness was created using different lithographic techniques. (a) photolithographic towers and (b) indented square posts[1], (c) diced silicon wafer[2], (d) photolithographic towers[3], (e) silicon nano-towers[4] (f) laser-modified SU8 surface[5] (g) SU8 towers [6], (h) silicon islands and (i) silicon nanowires grown on those silicon islands[7]…………………………………………………………………………………23

2.9. Examples of random roughness created using a number of different techniques. (a) poly(perfluoroalkyl ethyl methacrylate) coated electrospun fibers [8], (b) SEM image of MgAl2O4 monolith through a novel single-source inorganic precursor route, and after chemical modification with n-octadecanoic acid, the surface shows superhydrophobic (Inset) [9], (c) SEM image of nanorod film of Cu-ferrite by sol–gel process [10], (d)

xv porous copper films created by electrochemical deposition at a 0.8 A cm_2 cathodic current density in 0.5 M H2SO4 and 0.1 M CuSO4 for 45 s [11], (e) copper plate immersed in an aqueous solution of 2.0 M NaOH and 0.1 M K2S2O8 for 60 min, showing good superhydrophobic property after dodecanoic acid modification (inset) [12], (f) porous membrane produced by solvent casting of 17.9 mg ml−1 polypropylene solution using methyl ethylketone as the nonsolvent [13], (g) multifilament woven fabric [14], (h) micro-bead connected fibres formed by elecrospinning [15] and (i) cobalt hydroxide crystalline nano-pins (brucite-type) with diameter of 6.5 nm [16]………………………27

2.10. Examples of superhydrophobic surfaces with hierarchical dual roughness, (a) Two- tier textures: micropillars are etched in silicon, and CNT nanopillars are subsequently deposited [17], (b) CNT-coated polystyrene-sphere array [18] and (c) raspberry-like particulate film fabricated by assembling one layer of 35 nm silica particles on large silica particulate film prepared using Langmuir–Boldgett (LB) deposition [19]. The superhydrophobic surfaces with hierarchical roughness have been shown to show better stability of Cassie-Baxter state of a water droplet compared to single layer roughness…34

2.11. (a) and (b) correspond to schematic diagrams illustrating possible liquid-vapor interfaces on two different surfaces having the same solid surface energy and the same equilibrium contact angle (θ), but different geometric angles (ψ)[20] Figure (b)represents a re-entrant curvature with a net surface force balance directing upward implying that the liquid is supported on the surface rather than getting imbibed inside the pores……………………………………………………………………………………...35

2.12. Exampels of oleophobic surfaces, (a) electrospun surface containing 44.4 wt% fluorodecyl POSS and possessing the beads-on-strings morphology. The inset shows the molecular structure of fluorodecyl POSS molecules, (b) and (c) are reentrant curvature surfaces, called as ‘micro-hoodoo’ surfaces with flat cap and square top, respectively. (d) and (e) represent different liquids forming a very high contact angles on oleophobic surfaces formed with electrospun fibrous coating on a flat surface (d) and duck feather surface (e). Without fiber coating coating, oil completely wets these surfaces………….36

2.13. A schematic representation of SLIPS surface (a) [21] and carbon-soot based oleophobic surface (b) [22]……………………………………………………………....38

2.14. An example of free energy phase diagram constructed for pillar shaped geometry. The Wenzel and Cassie-Baxter wetting state regions can be identified and the difference between the two can be calculated.[23]………………………………………………….40

2.15. Schematic representation of some common types of interactions resulting in adhesion between the two surfaces, (a) Capillary forces, (b) suction, (c) mechanical interlocking (Velcro) and (d) electrostatic interactions………………………………….46

2.16. Schemtic representation of work of adhesion [24]………………………………...47

xvi

2.17. The hierarchical structure of gecko foot is shown. Each toe is made up of lamellar strips (a), every lamellar strip is made of an array of microscopic hairs called setae (b and c) and one seta branches out into hundreds of siner structures called as spatula (d). [25, 26]………………………………………………………………………………………..51

2.18. A schematic representation of different reactions taking place in a typical PECVD of organic molecules [27]………………………………………………………………..61

2.19. Empirical energy diagram for different types of reactions taking place in a typical PECVD process.[27]……………………………………………………………………..63

2.20. Chemical structure of maleic anhydrie molecule, showing two surface functionalities…………………………………………………………………………….70

3.1. Schematic representation of PECVD setup operated under vacuum...... 75

3.2. Vacuum operated PECVD set-up. The cyan glow of 1H,1H,2H-perfluoro-1-dodecne (a) can be differentiated from a pink glow of HMDSO (b)...... 77

3.3. A schematic of unit cell representing HNCP structure of spherical particles is shown. In the case of dual roughness, the second layer of smaller particles also form the same pattern, thus, the same unit cell can be used in the model for predicting the wettability ...... 79

3.4. A schematic representation of a furnace used for Chemical Vapor Deposition (CVD) process used for CNT growth...... 81

3.5. A schematic representation of steam condensation experiments……………………84

3.6. The CNT and PCNT surface can be visualized as a stack of multiple layers of nano- meshes (a). The smallest unit of the mesh is a rectangle bound by four tubes such that the width and height of the rectangle are “t” and “h” respectively; it is the unit cell geometry used in the model calculations (b)……………………………………………………….85

3.7. A schematic representation of shear adhesion measurement set up used to measure adhesive force between a shed sample and a substrate (glass or OTS-SAM coated glass)……………………………………………………………………………………..88

4.1 : SEM image of single layer HNCP array of 250 nm silica particles (a) and 500 nm silica particle (b)...... 94

4.2 : Examples of single layer roughness; a HNCP arrays formed with (a) 100 nm particles (scale bar : 1μm) and (b) 2700 nm particles (scale bar : 10 μm); (c) A schematic representation of a hexagonal unit cell. One unit cell contains one particle. L is the inter- particle distance measured between the centres of two adjacent particles and D is the

xvii particle diameter. A schematic representation of (d) Wenzel-state and (e) Cassie-Baxter state for single layer roughness formed by HNCP arrays of spherical particles………...95

4.3. Plot of contact angle vs r1 for different sized silica particles and the calculated Wenzel model……………………………………………………………………………97

4.4 . The angle ‘α’ and height ‘H’ defined for particles coated using PECVD………….98

4.5 . (a) AFM 2D height image of 1H,1H,2H-perfluoro-1-dodecene deposited by PECVD on the surface of clean Silicon wafer, scan size : 1μm, (b) AFM 3D height image of 1H,1H,2H-perfluoro-1-dodecene deposited by PECVD on the surface of clean Silicon wafer, scan size : 1μm and (c) R’ calculated using AFM image analysis for a PECVD coating of 1H,1H,2H-perfluoro-1-dodecene as a function of scan size. The extrapolation to zero scan size gives the value of R2…………………………………………………...99

4.6. (a) A comparison of experimental contact angles with Wenzel (W) and Cassie-Baxter (CB) model curves for a single layer of HNCP array of 250 nm particles, plotted as a function of r1. The dotted line represents the meta-stable state, whereas the solid line corresponds to the stable wetting state. The transition point is at r1~1.5. (b) A plot of contact angles as a function of effective roughness (R) of a single layer of HNCP patterns for all particle sizes used. The experimental data points for all particle sizes follow the Wenzel model behaviour. The dotted lines represent the upper and lower limits of the Wenzel model region. The behaviour is consistent with the plot in figure (a)…………103

4.7. Plot of free energy G* values vs r1 for Wenzel and Cassie-Baxter states for single layer HNCP array of 250 nm particles………………………………………………….104

4.8. A survey spectrum of PECVD coating of 1H,1H,2H-perfluoro-1-docence on the surface of HNCP pattern………………………………………………………………..104

4.9. C1s spectra of PECVD coating of 1H,1H,2H-perfluoro-1-dodecene on the surface of cleaned Silicon wafer (a) and HNCP pattern (b)……………………………………….105

4.10. A survey spectrum of PECVD coating of HMDSO on the surface of HNCP pattern…………………………………………………………………………………..107

4.11. C1s spectra of PECVD coating of HMDSO on the surface of cleaned Silicon wafer (a) and HNCP pattern (b)……………………………………………………………….108

4.12. Examples of surfaces with dual hierarchical roughness used for experimental studies. SEM images represent samples with (a) r1 = 3.5, r2 = 1.5, (b) r1 = 3.0, r2 = 1.5, (c) r1 = 2.5, r2 = 1.5, (d) r1 = 2.0, r2 = 1.5, (e) r1 = 2.0, r2 = 2.0. Scale Bar: 1 µm for all the images. Decreasing r1 from 3.5 to 2 can clearly be seen in the images (a) to (e)………110

4.13. R’ calculated using AFM image analysis for a PECVD coating of HMDSO as a function of scan size…………………………………………………………………….111

xviii

4.14. A schematic representation of (a) Wenzel state, (b) Cassie-Baxter state and (c) Penetrating Cassie-Baxter state for hierarchical dual roughness. Wenzel state corresponds to the penetration of water all the way inside the surface roughness features at both levels of roughness hierarchy. The Cassie-Baxter state corresponds to α1CB which defines the penetration of water between the bigger particles. The depth of penetration of water between the smaller particles in Cassie-Baxter and Penetrating-Cassie-Baxter states is characterized by α2CB and αPC, respectively...... 113

4.15. (a) Contact angles predicted for dual roughness unit cell model using Penetrating- Cassie-Baxter wetting state equations and compared to corresponding Cassie-Baxter state, θY = 110º, r2= 0.9 and αPC=70º, (b) Contact angles predicted for dual roughness unit cell model using Penetrating-Cassie-Baxter wetting state equations and compared to corresponding Wenzel and Cassie-Baxter states, θY = 95º, r2= 1 and αPC=80°……...... 114

4.16. The comparison of model curves and experimental data for contact angles as a function of r1 for hierarchical dual roughness; (a) Contact angles for HNCP dual roughness arrays with PECVD layer of 1H,1H,2H-perfluro-1dodecne deposited on the surface plotted as a function of r1, r2 = 1.5 is maintained constant. The Cassie-Baxter wetting state which is the only model that gives real contact angle values is denoted as CB. The model lines are generated using α1CB=30º and α2CB=70º. (b) Contact angle data for HNCP dual roughness patterns with a PECVD layer of HMDSO deposited on their surface are shown in with r2 = 1.5 and r2= 2. The possible wetting states are denoted as W (Wenzel) and CB (Cassie-Baxter).Contact angles measured by droplet deposition on the surface as well as by condensing water on the surface in both the cases are compared. Free energy calculations show that Wenzel is the most stable wetting state. The experimental data are consistent with the model predictions irrespective of the way in which the droplet is formed on the surface…………………………………………...... 116

4.17. (a) Comparison of G* values vs r1 for Wenzel(W) and Cassie-Baxter (CB) states for HMDSO coated dual roughness surface, CB-1.5 and W-1.5correspond to r2= 1.5; CB- 2 and W-2 correspond to r2=2, (b) Comparison of G* profiles as a function of r1 for Cassie-Baxter (CB) and Penetrating-Cassie-Baxter (PCB) states for 1H.1H,2H-perfluoro- 1-dodecene PECVD coated dual roughness surface, θY =110º, r2=0.9, αPC=70º, α1CB=30º and α2CB=70º, (c) Comparison of G* profiles as a function of r1 for Wenzel (W), Cassie- Baxter (CB) and Penetrating-Cassie-Baxter (PCB) states for HMDSO PECVD coated dual roughness surface, θY =95º,α1=42.3, r2=1, αPC=80º, α1CB=45º and α2CB=85º……..118

5.1. SEM images of CNTs(a) and plasma coated CNTs (PCNT)(b)…………………..124

5.2. TEM images of (a) uncoated CNTs and (b) plasma coated CNTs (PCNT), (c) Diameter distribution of uncoated (CNT) and plasma coated CNTs (PCNT)………….125

5.3. TGA analysis of (a) CNT and (b) PCNT…………………………………………..126

5.4. IR spectra of monomer and the coating formed by PECVD, (b) XPS C1s spectra of uncoated (CNT), plasma coated (PCNT) CNTs and plasma coated Al and Si…………127

xix

5.5. XPS survey spectra of (a) CNT (uncoated carbon nano-tubes) and (b) PCNT (Plasma coated Carbon nanotubes), (c) XPS F1s spectrum…………………………….128

5.6. XPS C1s spectra of uncoated (CNT), plasma coated (PCNT) CNTs and plasma coated Al and Si………………………………………………………………………...129

5.7. (a) Condensed steam droplets on PCNT surface after steam exposure of 1 hour; Contact angle measured on (b) CNT (uncoated) and (d) PCNT (plasma coated) surfaces; following the steam condensation water contact angle measured on (c) CNT and (e) PCNT surfaces; (f) Water contact angle measured on PCNTs as a function of steam exposure time…………………………………………………………………………...132

5.8. (a) A schematic showing CNT mesh structure. It can be visualized as multiple meshes stacked on top of each other, (b) A schematic of rectangular unit cell model. It has inner side dimension “(t x h)” and outer dimension "(t+d) x (h+d)", where t and h are the average length and width respectively of a rectangular cavity between nanotubes in a mesh structure and d is the average diameter of CNT, (c) ΔG* calculated as a function of r for θY = 80° (CNT) and θY = 110° (PCNT); 1< r <2.9……………………………….133

5.9. Water condensation on the surface of CNT (scale bar: 50µm)…………………….138

5.10. Water condensation on the surface of PCNT (scale bar: 50µm)…………………139

5.11. Frost formation on the surface of CNT (a) and PCNT (b)………………………..140

6.1 . Whole-animal shear adhesion data from Tokay geckos (Gekko gecko) tested on four surfaces in dry or wet contact…………………………………………………………..148

6.2. Unit cell representing the tetrad pattern of the gecko toe surface………………….151

6.3. Images of a gecko foot (A) in dry contact with glass, (B) in wet contact with a hydrophobic surface (OTS-SAM coated glass) and (C) in wet contact with a hydrophilic surface (glass)…………………………………………………………………………..154

6.4. The SEM images of B-S (a,b), M-S (c,d) and F-S (e,f). Comparison of lower magnification images (a,c and e) indicate that the tetrad pattern of the sheds was retained after PECVD layer deposition. The higher magnification images (b,d and f) show the finer structures present on the sheds (spatulae)………………………………………...160

6.5. (a) XPS survey spectra of B-S, M-S and F-S. The peaks assignable to constituent elements are labelled. The survey spectra were used for calculating surface atomic composition. The high resolution C1s XPS spectra of (b)B-S, (c) M-S and (d) F-S are compared………………………………………………………………………………..161

6.6. Average force per unit area of tokay gecko (Gekko gecko) skin sheds either the blank (B), M-S (M) or F-S (F). Samples were tested on a clean hydrophilic glass slide (a) or on

xx a glass slide coated with hydrophobic OTS-SAM (b). Ten samples in each coating and surface group were tested in air (black bars) and ten were tested in water (grey bars). Error is reported as 1 s.e.m. and significant differences are indicated with a *………..164

6.7. A schematic representation of a section of unit cell corresponding to gecko toe morphology is shown. A pillar shaped seta has a tier of smaller pillars on top of it. Unit cell consists of four such setae...... 169

6.8. ΔG* plotted as a function of R for different θY. The inset of the figure shows magnified plots for θY of 100°, 110°and 120°...... 170

A1. A schematic representation of the non-tetrad patterned unit cell morphology. One unit cell is boxed in red. Columns are 60m tall and 4m wide. Each column is separated by 1m………………………………………………………………………………….215

xxi

CHAPTER I

INTRODUCTION

Surfaces with patterns made up of features ranging from a few millimeters to a few nanometers are called structured surfaces. The patterns could either be regular or random. The structure-property relationship of the structured surfaces has been of great interest for various fields of materials technology and research developments. Over the last couple of decades, creation of such patterned surfaces with precise control over feature sizes, even as small as a few nanometers, has been achieved with highly sophisticated lithography and surface etching techniques (carried out with ion and electron beams). The developments of these techniques surely mark some of the milestones in the fields of science and technology. The inspiration for synthetic structured surfaces, however, lies in nature, where a variety of unique surface structures have played major roles in the evolution and biological functioning for survival of a number of biological systems. The intricate designs of natural patterns and their correlation to the surface properties still continue to amaze mankind. There is a whole range of structured surfaces where surface structure is at the heart of the surface property, including natural surfaces such as superhydrophobic plant leaves, dry adhesive gcko feet, structural colors on bird and butterfly feathers, the anti-reflective moth eye surface and synthetic surfaces such as water repellent fabrics, anti-glaring coatings, supercapacitors,

1 data storage devices with the capability to store up to a few terabytes of data, transparent and flexible electronics, efficient fuel and solar cells, to list a few.

Wetting and adhesion are the surface properties which depend on the structure of a surface, or more commonly referred to as surface roughness. The two important consequences of surface roughness are, first, that the surface roughness results in higher surface area at the contact interface between the two surfaces, compared to the corresponding two-dimensional area. Second, the surface roughness is also responsible for creating a heterogeneous contact interface in contact with a liquid, due to the presence of air pockets that are created as a result of the structure. The two factors form the basis of the effect of surface roughness on wettability and adhesion of structured surfaces.

Even though the geometric effects are significant, the intrinsic surface energy plays an equally important role in deciding the wettability or adhesion of a given surface.

Thus, tuning of wettability and adhesion not only requires control over the roughness but also over the surface chemistry. The optimum combination of the two is required for achieving desired wettability or adhesion.

Creating a surface with suitable wettability or adhesion is generally a two step process; one step is required to create surface structure (using etching, lithography, electrospinning and many more) and the other to achieve desired surface chemistry

(deposition, solution casting or molecular assembly). It is a very rare scenario to obtain both the parameters in a single step, i.e. to form a structured surface using a material with suitable surface chemistry. Thus, even though there is abundance of techniques to make structured surfaces and to modify surface chemistry, the need for a suitable combination of the two, when targeting a specific wetting or adhesion property of the surface, narrows

2 the choices down to a very few permutations. Experimental challenges and cost associated with the fabrication of a structured surface impose further limitations. Thus, there is a still need to develop a strategy that is versatile, easy, and cost effective for fine- tuning the wetting and adhesion of structured surfaces.

Wettability of surfaces, in addition to surface roughness and inherent chemistry, can also be affected by the external stimuli or environments in which they operate. For example, the wettability of surfaces exposed to varying temperatures, humid environments, vapor treatments, corrosive atmospheres or mechanical impacts could be different than the wettability of the native surface. Similarly, the external factors such as surface contamination and presence of water have unfavorable effects on adhesion of surfaces. Thus, wetting and adhesion tunability demands taking into account the external factors as well, in order to impart robustness to the given structured surface.

Wetting and adhesion are not completely isolated phenomena, especially for structured surfaces. For example, the cell adhesion to structured surface is often related to water wettability of the surface; the self cleaning ability of the fibrillar adhesive is commonly linked to its superhydrophobic behavior. The unique fibrillar adhesive system that geckos use in their habitat is one such system where wetting and adhesion depend strongly on hierarchical structure of the surface. It is amazing how the natural system has evolved to retain its adhesion even in humid and dirty environments, both of which tend to result in adhesive failure for any typical adhesive. The gecko adhesive is a complex biomaterial that is only little understood with respect to its surface chemistry and bulk material composition. The limited knowledge restricts both understanding of the adhesive

3 mechanism and, effectively, the ability to tune the adhesive properties of gecko inspired mimetic structures.

In this study, we have studied the wettability of a structured surface as a fundamental phenomenon using model surfaces that we fabricated. We used this knowledge further to design surfaces with fine tuned wettability for improving their performance in specific applications. We also have studied the effect the surface wettability has on adhesion using an example of gecko adhesion. We used a plasma enhanced chemical vapor deposition (PECVD) process for chemical modification of surfaces, as it offers advantages of being a vapor based dry process, and can easily be used for a wide range of organic precursors. It is also simple and easily scalable, since it can be operated at ambient pressure as well. We have explored the PECVD chemistry to form functional films and coatings on the structured surfaces.

CHAPTER II

BACKGROUND

2.1 Wetting

Wetting, in simple words, can be defined as an interaction between a solid and a liquid surface which decides the extent to which a given solid surface would wet when contacted with a liquid. The wetting for a given pair of solid and liquid, thus, indicates if liquid would spread on solid surface or would bead up to form well defined spherical cap shaped droplets with a finite contact angle. The devices and processes used in a number of industrial applications such as filtration membranes, waterproofing, paint and adhesive formulations, nanotechnology, catalysis, corrosion, marine applications, biological systems etc. rely on a fundamental understanding of surface wetting phenomenon. In majority of these applications, the organic liquids form a major component and are commonly used as solvents, dispersion agents, or viscosity modifiers, from a laboratory scale (utilizing a few milliliters or liters of solvents) all the way up to industrial scale.

The efficiency of performance of these liquids is strongly dependent on their wettability with respect to the other constituents in the formulations.

2.1.1 Wetting by water

Water acts as the main component in the aqueous systems. Additionally, water is not only a vital part of life and biological systems but it is also a vital part of industrial processes such as steam generation, hydroelectric power generation and heat exchangers.

Thus, water and its interactions with different types of surfaces has always been a topic of great interest to the researchers. Moreover, unlike other liquids, other physical forms of water, i.e. solid (ice) and vapor (steam), are also commonly encountered in our day to day life. The comprehension of the wettability of water and the effect its solid or vapor form might have on the wettability, thus, is an important aspect of surface and interface properties of water. There is, therefore, an obvious reason for continued interest and intense focus to understand the fundamentals of wetting of water (ice or steam) and to apply them to technological developments.

The chemistry of water molecules is very unique as well, which further stirs an interest in its surface and interface interactions with different materials. Water is a triatomic molecule with the molecular formula H2O. It is a highly polar molecule, and exhibits a log range of intermolecular hydrogen bonding in all its physical forms, which is one of the main reasons for some of its very special surface properties. Water has its highest density at about 4°C. The density of water reduces at it is cooled down and transformed into solid ice, which is contrary to the phase transition behavior shown by any other liquid. The phenomenon is most commonly known as the “anomalous behavior of water”. The hydrogen bonding of water is a key component of surface forces which decides the wettability of a given solid surface with respect to water.

2.2 Contact Angle

Contact angle is one of the most commonly used physical parameters which define the wettability of surfaces.

2.2.1 Background

The wettability or wetting property of a given pair of solid and liquid is measured in terms of the contact angle of a droplet of liquid deposited on the given solid surface.

The liquid droplet, upon contacting a solid surface, equilibrates to form a symmetrical spherical cap shape, as a result of surface force balance of adhesive and cohesive interactions of the two surfaces brought in contact. The equilibrium shape of the droplet and the contact angle acquired by it with respect to the three phase contact line (solid- liquid-vapor, saturated vapor in equilibrium with the liquid phase), is characteristic of a given solid-liquid pair and is termed as ‘equilibrium contact angle’.

Figure 2.1. Schematic representation of equilibrium contact angle of a liquid droplet on a solid surface

2.2.2 Young’s Equation

The schematic in figure 2.1 shows the equilibrium spherical cap shape of a liquid droplet in contact with a solid surface. The surface forces are shown with the arrows indicating the directions in which they act. A simple surface force balance at a three phase contact line on the surface can be written as follows:

……………………………………………………………… (2.1)

………………………………………………………...... (2.2)

Equation 2.2 is called Young’s equation, which defines the equilibrium contact angle ( ) for the given solid-liquid pair. The important assumptions in the derivation of

Young’s equation are: (1) A solid surface is chemically homogenous i.e. the chemical composition of the surface (defined by the presence of functional group) over the entire contact area at the interface is constant and, (2) the solid surface is perfectly smooth at the contact interface. Any real surface however, is neither perfectly homogeneous chemically nor perfectly smooth. Also, the definition of surface roughness is only relative. The frame of reference for the smallest length scale of roughness below which surface could be assumed to be flat is still a topic of debate, since some of the theoretical studies have recently shown that even the atomic scale roughness can be expected to have an effect on surface wettability. Experimentally, usually atomically smooth surfaces

(silica, glass, and mica), thin films (polymer films and films deposited with vapor deposition techniques) and molecular monolayers or any surfaces with absence of obvious and intentional surface features present on it, are assumed to be a frame of reference for flat surface for estimating using Young’s equation. The effect that

8 surface roughness has on its wettability will be discussed in great detail later. = 90° is used as a distinction reference for differentiating liquid liking surfaces from liquid repellent surfaces, e.g. in the case of water, the surfaces with < 90° are referred to as

‘hydrophilic’ surfaces and the ones with > 90° are called as ‘hydrophobic’ surfaces.

The contact angles and other wettability parameters, although applicable to any solid- liquid pair, are most commonly used and defined for water. Hence, some of the following concepts will be discussed for water-solid wettability. It should be noted, however, that these general concepts are applicable to all liquids in contact with given solids.

2.2.3 Static and dynamic contact angles

The wettability of the surface can be expressed in terms of different types of contact angles. The static contact angle is measured with the surface held perfectly flat horizontally such that a liquid droplet forms a symmetrical shape on the surface (as shown in figure 2.1 above).

There are two types of dynamic contact angles, namely advancing contact angle

( ) and receding contact angle ( ). The advancing contact angle is the maximum angle possible on the surface whereas, the receding contact angle is defined as the minimum possible contact angle on the given surface. The difference between the two is called as contact angle hysteresis (CAH). The CAH is used to define the entire range of contact angles possible on a surface and is commonly used to describe the extent to which the surface has an affinity towards water. The equilibrium contact angle value typically lies within this range i.e. < < . CAH is a consequence of physical defects and chemical heterogeneities on the surface and results in the pinning of a droplet to the

9 surface. Typically there are two ways to measure advancing and receding contact angles on the surface.

As shown in figure 2.2, advancing and receding angles can be measured on the surface by gradually increasing and reducing the droplet volume with the help of the syringe, respectively till the angle is changed on the surface. Other way to measure advancing and receding contact angle is a tilting plate method. In this method, the angles are measured on the surface by tilting the base of surface to which it is fixed with the droplet of liquid sitting on top of it, as shown in figure 2.2 c. As the tilt angle increases, the angle at the advancing edge increases and that at the receding edge reduces. The tilting is continued till the droplet rolls off the surface. The maximum and minimum angles at advancing and receding edges before the droplet just rolled off the surface are measured as advancing and receding contact angles respectively. The CAH is dependent more on chemical heterogeneity than surface roughness, as the measurement using the tilt angle method suggested.[29]

Figure 2.2. Schematic representation of (a) advancing contact angle( ) and (b) receding

Figure 2.2 (continued). contact angle ( ) using volume addition/removal method, (c)

Schematic representation of tilt plate method for measuring CAH[28]

The origin of CAH is still ambiguous. The two contradicting hypotheses for explaining CAH as a physical parameter have been debated against each other over the past few years in the wetting literature. One of the proposed hypotheses, which has been backed up by experimental demonstration, showed that the wettability of a surface is a completely surface tension derived phenomenon, which mathematically is defined as the force per unit length along the surface and acts perpendicular to the line of contact.[28]

Based on this argument, it was suggested that the wettability measured in terms of contact angle is independent of contact area at the interface, and solely depends on how the contact line force moves along a surface. Thus, it was proposed that CAH can occur only in the scenarios where the contact line experiences the surface heterogeneity, either chemical or structural. On the other hand, thermodynamic arguments based on free energy derivations propose that CAH is, in fact, dependent on the area of contact interface and its chemical composition.[30] The metastability in contact angles observed on such heterogeneous smooth surfaces resulting in CAH was also shown to be dependent on droplet volume, based on thermodynamic analysis.[31] The hypothesis has been shown to be in good agreement with experimental demonstrations in this case as well. In spite of all the controversy, CAH has been used as a very important quantification parameter for differentiating “sticky” versus “slippery” surfaces[32] and is an indicator of the Cassie-Baxter to Wenzel wetting transition as demonstrated by

Nosonovsky, M. et. al. In their study, they also showed that CAH is a consequence of kinetic motion of the triple line of contact at the solid-liquid interface and the adhesion

11 hysteresis of the surface.[23] CAH as a function of surface feature geometry was systematically studied by Extrand, C. W., and a criterion was developed to predict the

CAH associated with a structured surface as a function of contact line density and surface force balance.[33] It is difficult, however, to precisely predict the CAH for a typical structured surface, firstly because the roughness shape is not precisely controlled and well defined, and secondly, because of the thermodynamic wetting transitions taking place on the surface where each wetting state is associated with different CAH.

In addition to static and dynamic contact angles, roll-off angle and sliding angle are defined occasionally for a surface to explain its wettability. The roll-off angle is defined as the tilt angle at which, when the surface with a liquid droplet on it is tilted, the droplet rolls off of the surface. Similarly, as the name suggests the sliding angle is defined as the tilt angle at which the liquid droplet starts sliding off the surface with the same angle acquired by a spherical liquid cap at both the ends with zero CAH. Both roll- off and tilt angles are related to CAH; the lower the roll-off angle, the lower the CAH.

The same relationship holds good for tilt angle and CAH as well. In fact tilt angle is a tool to measure CAH, as explained above.

2.2.4 Surface energy versus surface roughness

As discussed above in the derivation of Young’s equation, the contact angle of a liquid measured on a given surface is a function of interfacial energy of solid-liquid, liquid-vapor, and solid-vapor interfaces i.e. the surface energy of a solid and liquid in contact. The wettability of a liquid can be defined in terms of the minimum and maximum contact angles possible for a given liquid on solid surfaces with different surface energies. For example, for the wettability of a surface with respect to water, it has

12 been shown that the minimum contact angle on the surface could be as low as 0°, implying complete spreading of water droplet on the surfaces such as surfaces of clean glass, silica, and mica. On the other hand, the maximum contact angle that water could acquire on a surface is about 120°, on the surface on polytetrafluoroethylene (PTFE, most commonly known as Teflon). The lowest possible surface energy of a solid surface corresponding to the maximum limit of a contact angle is attributed to the -CF3 groups oriented at the air interface. The vapor deposited n-perfluoroeicosane on the surface of glass has been reported to have the contact angle similar to that of PTFE and was measured to have the surface energy of about 6.7 mJ/m2; the lowest possible for a solid surface. The low surface energy in this case was also attributed to -CF3 groups oriented at the air interface.[34]

Figure 2.3. (a) Image of a water droplet beaded up on the surface of lotus leaf, (b) SEM image of hierarchical structure of lotus leaf surface consisting of waxy bumps, every

Figure 2.3. (continued) bump is covered with finer structures called wax tubules and (c) high resolution SEM image of waxy tubules.[35]

Because no other functional group has surface energy lower than the CF3, it could be expected that water would not have contact angle more than 120°. However, it is a common observation that water in contact with the natural surfaces such as plant leaves

(most of them) bead up to form spherical droplets and rolls off quickly. It has also been identified as a surface cleaning mechanism for the leaf’s surface. A spherical droplet beaded up on a leaf surface is shown in figure 2.3a above. The water contact angle measured on the surface of lotus leaf was about 160°, much higher than the maximum limit established based solely on the surface energy component. The surface of the lotus leaf thus was called a ‘superhydrophobic surface’. Thus, a superhydrophobic surface is typically associated with a water contact angle of about 150° and low contact angle hysteresis, as was observed in the case of the lotus leaf surface. Although the surface of lotus leaf is made up of a waxy material (moderately hydrophobic with contact angle ~

75°, measured on the layer of smooth carnauba wax, excreted from the wax palm leaves and have similar chemical signatures as that of lotus leaf waxy cuticles)[36], with an intrinsic wettability lower than that of PTFE , the surface still acquired a static contact angle much higher than that of PTFE. Based on this observation, it was further concluded that the surface topography of the lotus leaf had an equal contribution towards its wettability. The hierarchical structured surface of lotus leaf is shown in figure 2.3b and

2.3c. As can be seen, the surface of a lotus leaf is made up of micron scale bumps of a waxy material. These features on its surface are further covered with tiny nano-scale tubular features. The unique topography and structural hierarchy, along with the

14 moderate hydrophobicity of intrinsic materials, was proposed to be responsible for the superhydrophobicity of a lotus leaf. Thus, it was demonstrated with the example of a natural leaf surface that the surface roughness, along with the surface energy, is a critical factor for deciding surface wettability of solids. In fact, it was identified as the most important parameter in designing the specific wettability of surfaces, such as oleophobicity and superabsorbency, as will be discussed later.

Even though it was only over the last decade or two that the synthetic mimics of superhydrophobic surfaces were developed using numerous synthetic methods and surface morphologies, the models to account for the effect of surface roughness on surface wettability date back to 1940’s. Two models, namely the Wenzel model, developed in 1936, and the Cassie-Baxter model, developed in 1944, are used still today to explain the effect of surface roughness on its wetting properties, both quantitatively and qualitatively.

2.2.5 Wenzel model

The contribution of surface roughness to surface wettability was first thoroughly studied by Robert N. Wenzel, while studying the water repellency of textile materials. He suggested that it is the physical state of the surface, in addition to the surface tensions of the solid and liquid, that decides the wettability of solid surfaces. He defined a dimensionless and unitless parameter, r, called roughness factor. It is defined as the ratio of actual area at the contact interface to the geometric surface area, i.e. the projected contact area. Thus, as per the definition, r is always greater than 1 since the actual surface area is always more than the projected flat surface area.

Figure 2.4. A schematic representation of Wenzel state of a liquid droplet on rough solid surface, is the apparent contact angle measured on the surface called as “Wenzel angle”.

A schematic shown in figure 2.4 represents a Wenzel state. As shown in the figure, for a typical rough surface, a water droplet on coming in contact with the surface wets the surface area completely, such that the asperities on the surface also are wetted by the liquid. The apparent contact angle of a droplet resting on the surface in this case ( ) is given by the following equation:[37]

……………………………………………………………………(2.3)

From the equation 2.3 above, it can be concluded that the inherent wettability of the surface is amplified by a factor ‘r’, representing the physical state of the roughness associated with the surface, i.e. the Wenzel equation predicts that the hydrophilic surface will become more hydrophilic and a hydrophobic surface will become more hydrophobic as a result of surface roughness.

2.2.6 Cassie-Baxter model

The Wenzel model for wettability of a rough surface is applicable to homogeneous wetting of a surface, as discussed in the previous section. The Cassie-

Baxter model, on the other hand, was developed for a heterogeneous interface formed between a liquid and a solid surface with chemical heterogeneity present on the surface.

For example, in the case of superhydrophobic surfaces with low CAH, a Cassie-Baxter state is acquired by a water droplet deposited on a surface with roughness asperities present. In such a case, water droplet is partially in contact with the solid surface and is partially supported on the air entrapped between the roughness asperities, thus creating a chemical heterogeneity at the contact interface, since the interfacial energy of the solid- water and the air-water interfaces are different from each other.[38]

Figure 2.5. Schematic representation of Cassie-Baxter state of a superhydrophobic surface.

As shown in the schematic representation of Cassie-Baxter state in figure 2.5 above, the liquid prefers to rest on top of the rough features leading to the presence of air

17 pockets in the gaps between the roughness features. The apparent contact angle in this case ( ) can be estimated using following equation:

……………………………………………………………....(2.4)

and correspond to the area fractions of solid-liquid and vapor-liquid interface respectively.

Figure 2.6 represents the theoretical curves defining the Wenzel and Cassie-

Baxter regimes. A plot of θR as a function of intrinsic contact angle (θ) is in fact a very convenient way to quickly identify two distinctly different wetting regions for a given surface roughness, when done in the cosine domain of contact angle values, as shown in the figure for a fractal surface.[39]

Figure 2.6. Theoretical curves representing the apparent contact angle (θR) as a function of equilibrium contact angle (θ), calculated using Wenzel and Cassie-Baxter models. [39]

The important assumption underlying the application of the Cassie-Baxter equation for predicting the apparent equilibrium contact angle on a surface with heterogeneous surface composition is that the area fractions of the two phases contacting a liquid droplet are assumed to be constant. The Cassie-Baxter equation thus, cannot be used for calculating advancing and receding apparent contact angles, and thus, CAH values associated with the liquid contacting the surface, since the area fractions vary as the contact line either advances or recedes. In order to address this issue, a modified

Cassie-Baxter equation was proposed. The modified equation is applicable for calculating advancing and receding contact angles for surfaces with feature sizes on the order of few micrometers. It accounts for a distorted contact line upon contacting the surface structures, as opposed to an almost flat contact line in the case of smaller roughness features. A surface texture parameter, called as , was introduced by Choi, W. et. al. and the modified Cassie-Baxter equation was formulated as follows [40]:

……………………………………(2.5)

Where, , and correspond to advancing contact angle and equilibrium contact angles of surfaces 1 and 2, respectively, in contact with the liquid droplet.

2.3 Designing surface with desired wettability

As discussed in the section above, we know that the surface chemistry and roughness both are equally important in deciding the wetting state of a given surface. The synergistic effect needs to be accounted for not only to create optimum roughness geometry but also for the choice of material with appropriate surface chemistry to make a patterned roughness. For example, it is theoretically possible to create a

19 superhydrophobic surface using a material that is hydrophilic by choosing the right surface roughness. However, it may not be the most favored wetting state thermodynamically and may not be observed under experimental conditions. [41]

2.3.1 Superhydrophobic surfaces

As discussed in the previous section briefly, the unique water repellent property of a lotus leaf surface, which is attributed to its structured surface, has been an inspiration for surface scientists to design a synthetic mimic. The lotus leaf, however, is not the only surface that exhibits superhydrophobicity. There are numerous natural superhydrophobic surfaces; examples of some of them are shown in figure 2.7. The common feature of all of the surfaces is a roughness patterns on their surfaces, as can be seen in the figure.

Figure 2.7. Examples of natural superhydrophobic surfaces. Different types of structural patterns observed on (a) lotus leaf surface, (b) taro leaf surface, (c) rice leaf, (d) surface

Figure 2.7. (continued) of butterfly wing, (e) stenocara beetle’s back shell surface and (f) mosquito eye structure.[42, 43]

The study of the lotus leaf’s surface using microscopic and spectroscopic tools revealed the secret behind its very special surface property, i.e. water repellency. The surface that appears flat macroscopically is made up of microscopic tiny bumps consisting of a hydrophobic waxy material which also are further covered with wax crystalloids, both of which are further covered with nanoscopic hairy structures (also called as wax tubules). Typically, the bumps are about 3-10 µm in dimension, and the wax crystalloids are about 70-100 nm. The nano-hairs, the smallest unit of the structural hierarchy, are about 50 nm in diameter. A combination of structural hierarchy and low surface energy wax layer makes the surface of lotus leaf superhydrophobic. Similar to the lotus leaf surface, the surfaces of the rice leaf and the taro leaf are also made up of structural hierarchy and are also superhydrophobic.[42] There are other natural surfaces as well which show strong water repellency and have inspired a number of synthetic mimics. Some of such natural surfaces include gecko lizard’s toe pads, bird feathers, butterfly wings, water striders’ legs, mosquito eyes, cicada wings, rose petals and the

Stenocara beetle’s back.[43] In addition to the superhydrophobic property, the distinct roughness geometry also contributes to some unique surface properties of these surfaces, such as reversible adhesion and structural colors.

The surface is called superhydrophobic when the water contact angle on the surface is 150° or more.[44] The superhydrophobic state is thus, with the right choice of surface roughness and chemistry, possible in both Wenzel and Cassie-Baxter states of a water droplet. However, the droplets in Wenzel state are sticky and are associated with

21 very high CAH. It is a matter of debate to define superhydrophobicity of a surface; whether to account for both high contact angle and low CAH, or if high contact angle is enough to characterize the surface as superhydrophobic is an ambiguous argument. For example, there are numerous examples of textured surfaces, both natural and artificial, on which water forms almost perfectly spherical droplet (contact angle of about 150° or more) and also rolls easily off the surface. Some of the examples include lotus leaf surface, silicon posts with a layer of fluorinated silanes, and electrospun micro and nano fiber mats.[45] On the other hand, some surfaces result in a high water contact angle but also exhibit very high CAH. An example is of a surface of etched polystyrene which forms a static water contact angle of about the same (~150°) but the droplet is sticky and does not leave the surface even when the surface is flipped upside down i.e. tilted by

180°.[46] In this case, in spite of the fact that hysteresis is very high, the surface was still referred to as being ‘superhydrophobic’. It is therefore, a common practice to specify a

CAH value along with the contact angle measured on a superhyrophobic surface in order to distinguish a surface forming a non-stick water droplet versus the one that forms a sticky droplet. It has been shown that hierarchical roughness of the surface helps in reducing the hysteresis since the surface pinning is reduced due to high roughness.[47]

Synthetic routes to superhydrophobic surfaces:

The extensive study of water repellency of natural surfaces showed that surface roughness is a critical parameter to create a superhydrophobic surface synthetically.

Having understood this, a number of models and techniques have been developed to achieve the desired roughness. The roughness could either be a well defined geometry with regularly shaped and spaced features or random surface roughness. Some of the

22 most common technologies used for synthetic superhydrophobic surfaces are discussed below. The most common strategy to make a surface superhydrophobic is to create roughness on the surface, using one of the techniques discussed below, followed by hydrophobizing the thus formed structured surface with functionality having low surface energy, such as the functional groups which are either silane based or which contain fluorinated chemistry. Some of the examples of surface structures which impart superhydrophobicity to the surface are shown in figures 2.8, 2.9 (ordered roughness and random roughness respectively) and 2.10 (dual surface roughness).

Figure 2.8. Examples of single level surface roughness. The ordered patterns and systematically controlled roughness were created using different lithographic techniques.

Figure 2.8. (continued) (a) photolithographic towers and (b) indented square posts[1], (c) diced silicon wafer[2], (d) photolithographic towers[3], (e) silicon nano-towers[4] (f) laser-modified SU8 surface[5] (g) SU8 towers [6], (h) silicon islands and (i) silicon nanowires grown on those silicon islands[7]

(1) Wet chemical reactions

Wet chemical reactions refer to the ones taking place in a solution. The use of wet chemical reactions for creating superhydrophobic surfaces has been most commonly used for surface etching of metals or for growth and deposition of metal oxide nano-particles from solution in order to form a textured surface. Because a metal surface inherently has a high surface energy, the creation of a structured surface needs to be followed by a deposition of a low surface energy coating to make the surface superhydrophobic. Some of the examples where wet chemistry was used to form superhydrophobic surfaces include a two step process used to form ZnO microstructures from solution, followed by its modification by organic self assembled monolayer (SAM) [48], ZnO nanorods formed on a cotton fabric surface using a suitable zinc solution[49], surface etching of polished copper substrate by immersion in oxalic acid solution to produce roughness followed by hydrophobic layer deposition[50] and flowery structures formed on nickel surface using monoalkyl phosphonic acid solution[51]. Thus, wet chemistry offers a facile method to create a variety of surface textures of metals.

(2) Electrochemical deposition

Electrochemical deposition using electrolyte solutions is another versatile and novel method to create structured surfaces. Multilayer polyelectrolyte films formed by the electrochemical deposition offered a preformed matrix for deposition of gold

24 nanoparticles in the dendritic patterns. The deposition of a hydrophobic layer on the thus formed structured surface resulted in a superhydrophobic surface with very low

CAH.[52] Thin structured films of ZnO formed by cathodic electrochemical deposition, upon modification with a layer of perfluorosilane based coating, were shown to form a superhydrophobic surface.[53] A multilayer matrix formed by layer-by-layer (LbL) assembly of polyelectrolytes was found to be suitable for the deposition of silver (Ag) aggregates by the method of electrodeposition. The morphology of Ag aggregates, and effectively the surface roughness, could easily be adjusted by manipulating the time and potential used in the deposition process.[54] A similar approach was also used to form gold (Au) aggregates.[55] A combination of LbL and electrodeposition serves as a nanofabrication tool to create very unique structures which resemble the water striders’s leg structures, one of the natural superhydrophobic surfaces.

(3) Molecular self assembly

The formation of self assembled monolayers (SAM) using vapor phase deposition is one of the most widely used methods to make the surface hydrophobic by molecular modification. As discussed previously, a majority of structured surfaces rely on this process as a second step in creating a superhydrophobic surface. Formation of a hydrophobic SAM, followed by roughening a surface, has also been explored as a way to create superhydrophobic surfaces. A self-assembly of an optimum mixture of mono, di-, tri- and tetra- chloro silanes was demonstrated by Gao, L. et. al. to react with the surface to initiate vertical polymerization and resulted in the formation of structured network called as “Lichao’s surface”. These surfaces are reported to have the highest water contact angle (175°-178°), with almost no CAH. They were referred to as “perfect

25 superhydrophobic surfaces”.[56] The mechanical manipulation of contact points on the elastomeric surface on which molecular self assembly is taking place was developed as a surface modification method to create structured surfaces. The assembled layers are called as “mechanically assembled monolayers” (MAMs). The structured MAMs surfaces formed with semi-fluorinated molecules were shown to be superhydrophobic.[57] Some other examples of hydrophobic SAMs include films formed using methyloctyldimethoxysilane and fluorooctylmethyldimethoxysilane on the silicon wafer surface. The optimum roughening of the SAM surface was used to make it superhydrophobic with very low CAH.[58]

(4) Layer by layer deposition

Layer by layer (LbL) deposition from polyelectrolyte solutions is often combined with electrochemical deposition to form superhydrophobic surfaces, as previously discussed. This technique has also been explored to create wettability patterns on a surface. LbL deposition of hydrophilic layers on a superhydrophobic surface, carried out in a controlled manner by Zhai, L. et. al., resulted in the formation of a patterned surface which has wetting properties similar to that of a desert beetle. Creating such surfaces with wettability patterns could find applications in the fields of controlled drug delivery and water harvesting.[59]

(5) Surface lithography

Typical lithography techniques used for creating superhydrophobic surfaces include soft lithography [60, 61], nanoimprint lithography [62, 63], electron and X-ray beam lithography [64], photocatalytic lithography [65, 66] and colloidal lithography [67,

68]. The surface patterns are created using high energy beams or assembly of colloidal particles. Soft lithographic techniques most commonly use a photosensitive mask, an elastomeric stamp with patterns on its surface. Precise control of surface patterning offered by photolithography was used by Spori, D. M. et. al. for creating a surface pattern gradient by varying the density of pillars and holes on the surface. This unique design approach enabled the control over surface roughness to obtain the Cassie-Baxter wetting state of a water droplet along the entire length of the gradient and to study the effect of gradient geometry on the CAH and roll off angles associated with a Cassie-Baxter wetting state. A relatively high hysteresis in the Cassie-Baxter wetting regime was attributed to the pinning of the contact line at the pillar tips, whereas a lower hysteresis was caused in the regions with higher hole density.[69]

Figure 2.9. Examples of random roughness created using a number of different techniques. (a) poly(perfluoroalkyl ethyl methacrylate) coated electrospun fibers [8], (b)

Figure 2.9. (continued) SEM image of MgAl2O4 monolith through a novel single-source inorganic precursor route, and after chemical modification with n-octadecanoic acid, the surface shows superhydrophobic (Inset) [9], (c) SEM image of nanorod film of Cu-ferrite by sol–gel process [10], (d) porous copper films created by electrochemical deposition at

2 a 0.8 A cm cathodic current density in 0.5 MH2SO4 and 0.1 MCuSO4 for 45 s [11], (e) copper plate immersed in an aqueous solution of 2.0 M NaOH and 0.1 M K2S2O8 for 60 min, showing good superhydrophobic property after dodecanoic acid modification (inset)

[12], (f) porous membrane produced by solvent casting of 17.9 mg ml−1 polypropylene solution using methyl ethylketone as the nonsolvent [13], (g) multifilament woven fabric

[14], (h) micro-bead connected fibers formed by elecrospinning [15] and (i) cobalt hydroxide crystalline nano-pins (brucite-type) with diameter of 6.5 nm [16]

(6) Chemical vapor deposition

A chemical vapor deposition (CVD) process has been explored by a number of different researchers to create structured surfaces, as well as thin film modification for making a structured surface hydrophobic. A hot filament chemical vapor deposition

(HFCVD) process was used by Lau, K.K.S. et. al. to deposit a thin layer of polytetrafluoroethylene (PTFE) on the surface of vertically aligned carbon nanotube

(CNT) forests. The CNT surface thus formed was not only superhydrophobic but also retained the vertical alignment and water repellency under water condensation.[70] CVD was used by Sethi, S. et. al. to synthesize mesh-like Carbon Nanotubes (CNT) patterns on the surface of stainless steel. A layer of CNT deposited acts as a superhydrophobic coating for stainless steel surface. The CNT based superhydrophobic coating was

28 demonstrated to be conductive and retained the water repellent properties under extreme thermal stresses.[71]

(7) Plasma enhanced chemical vapor deposition (PECVD)

The use of highly energetic organic plasma source for deposition or surface etching is one of the most versatile tools to create superhydrophobic surfaces, since both the desired surface roughness and chemistry can be achieved in a single step. PECVD was used to create a superhydrophobic carbon nanotube nanoforest by Lau, K. et. al.[70]

Fluorination using PECVD process was identified as a way to make the surface of cellulose superhydrophobic, while retaining its biodegradability and mechanical properties.[72] PECVD as a process and its wetting and adhesion related applications are discussed further in greater details in section 2.9.

(8) Sol-gel processes

A sol-gel process is most commonly used to create patterns based on organometallic materials, typically silica, alumina, and titanium based. A flowerlike superhydrophobic-superhydrophilic pattern was created using a sol-gel method by

Tadanaga, K. et. al. A three layered patterned surface was formed with Al2O3 (first layer),

TiO2 (second layer), and fluoroalkyl chains as a third layer.[73] A similar process was used to create optically transparent Al2O3 based superhydrophobic coating.[74] An intrinsically hydrophobic organosilica based foam was formed using a sol-gel process. It is a single step process for creating superhydrophobic structured surface since the foam material is inherently hydrophobic.[75] A dual functional surface, which is both anti- reflective and superhydrophobic was created using a sol-gel process. The surface

29 roughness consisted of two layer coating where trimethylsiloxane functionalized silica particles were partially embedded into the organosilicon matrix binder layer.[76]

(9) Surface etching

Surface etching has been demonstrated as a useful technique for creating surface patterns using hydrophilic materials such as silicon. Silicon ‘nanograss’ was fabricated using a multistep unisotropic etching process. As a result of thin layer of poly(heptadecafluorodecylacrylate) photochemically attached to the surface, the resultant superwetting nanograss surface was converted into a superhydrophobic surface by

Dorrer, C. et. al. The silicon nanograss based superhydrophobic surface continued to be water repellent even under water condensation. The condensed water droplets were shown to move off the surface after coalescing with each other, leaving the surface completely dry.[77] A chemical etching of metal surfaces (aluminum, copper and zinc) followed by hydrophobizing the surface with fluoroalkysilane was established as a method to create superhydrophobic surfaces using such polycrystalline metals.[78]

(10) Hydrothermal etching

Hydrothermal etching of glass surface to form a rose flower like structure [79] and hierarchical structure[80] has been developed as a single step process for creating surface roughness. Deposition of hydrophobic monolayers on these structures makes the surfaces superhydrophobic with very low CAH.

(11) Templating and microprinting

Templating and microprinting are seldom used methods for creating patterns for superhydrophobicity. A femto-second microprinting was used by Li, B. et. al. to create titanium pillar based superhydrophobic surfaces which exhibit CAH as low as 5°.[81]

(12) Phase separation

The phase separation of a polymer solution by manipulation of solvent-nonsolvent composition or temperature is one of the easiest methods to create superhydrophobic surfaces. A simple, and probably the most inexpensive method, for making a polypropylene based gel-like superhydrophobic surface was developed by taking an advantage of its different solubility in different solvents and solution temperatures.[82]

The enhanced phase separation using a non-solvent (ethanol) was used to create structured surfaces from polyvinyl chloride (PVC) by drop casting its solution in the mixture of solvent (tetrahydrofuran) and a non-solvent (ethanol). The surface morphology of the dried porous PVC films was controlled by controlling the concentration of ethanol. The superhydrophobic surface thus formed was demonstrated to be robust enough to retain its water repellent properties upon outdoor exposure and immersion in water for prolonged period of time.[83]The principle of phase separation of polymers has been developed as a technique to create surface roughness required for superhydrophobic surfaces. Micelles formed by vapor induced phase separation of copolymer of styrene and dimethylsiloxane [PS-b-PDMS] were demonstrated as a convenient phase separation approach for making a superhydrophobic surface.[54]

(13) Electrospinning

Electrospinning is a tool to form mesh-like fibrous mats from dilute solutions.

The diameter of the fibers can be controlled from a few micrometers to nanometers (of the order of 100 nm), thus creating micro scale or nano scale porosity. Because the majority of the polymers are low surface energy materials (since they are hydrocarbon based materials), most commonly, the superhydrophobic surfaces created using electrospinning eliminate the need for surface modification (hydrophobic modification).

Elctroelectrospun mats formed using poly(acrylonitrile-b-a,a-dimethyl meta- isopropenylbenzylisocyanate) [poly(AN-b-TMI)] were shown to have excellent Cassie-

Baxter wetting state stability under uniaxial loading of about 400 Pa.[84] The robustness of the mats also facilitates further modification, as was demonstrated by Ma, M. et. al. In their work, electrospun fibers were further decorated with nanoparticles to produce a hierarchical roughness; the surface thus formed was superhydrophobic with very low

CAH.[8] The electrospun fabric material based on poly(caprolactone) (made using electrospinning, followed by hydrophobic surface modification using vapor deposition) was shown to be not only superhydrophobic (contact angle of about 175°) but also exhibited good oleophobicity.[85] Electrospinning of a block copolymer of styrene and dimethylsiloxane (PS-b-PDMS) (inherently a hydrophobic material) was demonstrated as a simple and facile method to make a superhydrophobic surface with very low CAH.

Inherent low surface energy enabled the process to be single step.[86] In addition to polymeric materials, electrospinning can also be used to create superhydrophobic fibrous nanostructures using inorganic materials, some of which also possess unique ability to undergo wetting transition when triggered by an external stimulus.[87]

2.3.2 Superwetting surfaces

Using an optimum combination of surface roughness and surface chemistry, a surface can be designed such that the water contact angle on the surface is zero. The surfaces are called superhydrophilic surfaces, the surface property exactly opposite to that of superhydrophobicity. As the name suggests, such surfaces show great affinity to water.

Water in contact with such a surface spreads or gets absorbed almost instantaneously, i.e. there is no spherical droplet of water observed on a superhydrophilic surface. The surface roughness allows for absorption of water, which is not possible on a flat hydrophilic surface. The micropatterning and nanopatterning techniques discussed above, which are used to create superhydrophobic surfaces, can be used for superhydrophilic surfaces as well, except for the hydrophobization step.[88] Some of the techniques for creating superhydrophilic surfaces include the assembly of particles using LbL technique[89], using polymer-SiO2 composite[90], and using inorganic oxide coatings[91]. Superwetting surfaces are important in designing superabsorbent fabrics and filtration membranes for water-oil separation. The surfaces with reversible switching of wettability from superwetting to superhydrophobic were also made possible by choosing the right design parameters. The use of external stimuli such as pH of the solution [92] and temperature

[93] were shown to make the switching possible.

Figure 2.10. Examples of superhydrophobic surfaces with hierarchical dual roughness,

(a) Two-tier textures: micropillars are etched in silicon, and CNT nanopillars are subsequently deposited [17], (b) CNT-coated polystyrene-sphere array [18] and (c) raspberry-like particulate film, fabricated by assembling one layer of 35 nm silica particles on large silica particulate film prepared using Langmuir–Boldgett (LB) deposition [19]. The superhydrophobic surfaces with hierarchical roughness have been shown to show better stability of Cassie-Baxter state of a water droplet compared to a single layer roughness.

2.3.3 Oleophobic and omniphobic surfaces

Unlike superhydrophobic surfaces, there is no naturally occurring surface that is oleophobic, i.e. which shows oil repellency. The biggest challenge in creating an oleophobic surface is the low surface tension of oils or liquid hydrocarbons. Thus, even though there are thousands of different ways to create superhydrophobic surfaces, there have been only a few successful attempts of creating oleophobic surfaces. The thermodynamic modeling of surface roughness showed that the roughness geometry is the most critical parameter for making a surface oleophobic, in addition to a low surface energy (typically obtained with siloxane or fluorinated functionality on the surface). This 34 roughness geometry is called as re-entrant curvature, as shown in the schematic in figure

2.11 below.

Figure 2.11. (a) and (b) correspond to schematic diagrams illustrating possible liquid- vapor interfaces on two different surfaces having the same solid surface energy and the same equilibrium contact angle (θ), but different geometric angles (ψ)[20] Figure

(b)represents a re-entrant curvature with a net surface force balance directing upward implying that the liquid is supported on the surface rather than getting imbibed inside the pores.

A similar theoretical criterion for creating liquid (oil or water) repellent surfaces, based on the direction of net surface force and contact line shape, was developed by

Extrand, C.W. to demonstrate the possibility of oil repellency theoretically.[94] Tuteja,

A. et. al. designed reentrant curvature geometry of the surface such that when an oil droplet with very low surface tension (typically about 10-15 mJ/m2) comes in contact with the surface, the net surface force balance is directed upwards, and results in the formation of a superoleophobic surface, since oil droplet is supported on the air pockets rather than getting imbibed into the pores (which would be the case with net force acting downwards).

Figure 2.12. Examples of oleophobic surfaces, (a) electrospun surface containing 44.4 wt% fluorodecyl POSS and possessing the beads-on-strings morphology. The inset shows the molecular structure of fluorodecyl POSS molecules, (b) and (c) are re-entrant curvature surfaces, called ‘micro-hoodoo’ surfaces with flat caps and square tops, respectively. (d) and (e) represent different liquids forming very high contact angles on oleophobic surfaces formed with an electrospun fibrous coating on a flat surface (d) and a duck feather surface (e). Without a coating of fibers, oil completely wets these surfaces.

[20, 95]

By systematically changing the dimensions of the re-entrant curvature on the surface, the critical point beyond which the net force balance would be directed downwards and oil would be imbibed inside the porous surface, was calculated and demonstrated experimentally.[95] The thermodynamic analysis of oil repellent state of the surface also shows that even though oleophobicity is possible on such surfaces with

36 reentrant curvature of some sort, it is not the most stable wetting state, and is limited by the hydrostatic pressure and mechanical impact that it can withstand.[20]

The oleophobic surfaces are also superhydrophobic, since the surface tension of water is much higher than that of typical oils. Thus, all the superoleophobic surfaces are superhydrophobic as well. It should be realized that there are very few materials that can be used for making the oleophobic surfaces with high oil contact angle (~150°). Another approach that was recently demonstrated was to design surfaces with almost zero CAH, irrespective of what the contact angle of oil on the surface was. Wong, T.-S. et. al. designed and demonstrated zero CAH of oil droplets on the surface of SLIPS (Slippery

Liquid Infused Porous Surfaces).[21] These surfaces are made up of porous polymer network and instead of air entrapped between the surface asperities, the pores are filled with a lubricating fluid, which has low surface energy and is chemically inert. This approach and design of a surface was inspired by the Nepenthes pitcher plant, a carnivorous plant which uses a similar mechanism for prey-trapping. As the oil droplet is brought in contact with such a surface, it slips off the surface, instead of getting imbibed in the pores. The main requirement of this design is the immiscibility and chemical passivity of the two liquids coming in contact with each other at the contact interface (a schematic representation is shown in figure 2.13 a). A simpler design approach was developed by Deng, X. et. al. in developing an “amphiphobic” transparent coating based on candle soot template. The nano-meter scale structure was obtained by the deposition of soot on a glass by holding it over the candle flame, depositing a silica shell followed by calcination to make the surface transparent. Silanization of the structures was used to lower the surface energy of the coating (a schematic representation of the surface is

37 shown in figure 2.13 b below). The coating was also shown to be mechanically robust since it retained its superoleophobic and superhydrophobic properties even after the abrasion caused by sand blasting.[22] The design of oil repellent surfaces is important for applications involving oil-water separation and oil fractionation processes.

Figure 2.13. A schematic representation of SLIPS surface (a) [21]and carbon-soot based oleophobic surface (b) [22].

2.4 Thermodynamic stability of wetting states

For structured superhydrophobic surfaces, it has been shown that Wenzel, Cassie-

Baxter, and a number of intermediate wetting states are possible depending upon the penetration depth of water droplet inside the roughness features. Thermodynamic free energy barriers separate all the possible wetting states. The possibility of transition from one state to the other depends on if the barrier can be overcome. The water repellent property of a lotus leaf surface, which is an inspiration for synthetic superhydrophobic surfaces, is, in fact, a metastable wetting state. The surface is designed in such a way,

38 however, that it retains its anti-wetting properties under the environmental conditions it is commonly exposed to.[96] The model studies showed that the quantification in terms of contact angles for wetting states can be used to distinguish one from the other. Also, difference between the free energy corresponding to every wetting state determines the barrier that is required to overcome in transitioning between the wetting states. A thorough analysis of the equilibrium wetting of a liquid droplet on a structured surface was carried out by Marmur, A. He estimated the boundary conditions for the equilibrium wetting regime based on thermodynamic analysis of the competing wetting states,

Wenzel and Cassie-Baxter.[97] The transition between the wetting states can be brought about in a number of different ways. The multiple wetting states possible on a structured surface were first demonstrated by Lafuma et. al. In their study, the transition was induced by compressing a droplet of water between two identical superhydrophobic surfaces, which initially exhibit a Cassie-Baxter wetting state, with the droplet supported on air pockets between the roughness features. The induced transition resulted in a

Wenzel state, where the air pockets were observed to be filled with water.[98] Other methods to induce such a transition include the condensation of water on the surface, evaporation, vibration induced transition, and mechanical impact. The capillary pressure was also used as a way to quantify the transition pressure between the wetting states. The transition between the wetting states may or may not be reversible depending upon if the roughness or surface chemistry is altered upon contact of water with the surface. A theoretical study carried out by Patankar, N. et. al. with pillar shaped roughness features pointed out the important aspect of such a transition. The quantitative analysis of free energies showed that when the Wenzel state is thermodynamically more favorable, the

39 transition of a liquid droplet in the Cassie-Baxter state on a given surface to Wenzel state can only take place if the free energy barrier could be overcome by external stimulus.

Thus, this study emphasized the importance of quantifying free energy associated with wetting states.[99]

The calculation of free energy of wetting for different wetting states on the surface can be used to construct a phase diagram of surface wettability. An example of a phase diagram is shown in figure 2.14 below. As can be seen in the figure, free energy for the Cassie-Baxter and Wenzel states were calculated for pillar geometry and plotted as a function of contact angle. With the help of a phase diagram, the range of contact angles which correspond to Cassie-Baxter state could be predicted. Also, the relative comparison of Wenzel state and corresponding Cassie-Baxter state (for same contact angle on the plots) allows to predict which is the most favorable state and the magnitude of difference between the two, in order to bring about the transition. A phase diagram can be used to identify wetting regime for the chosen surface roughness and chemistry and to design the surface accordingly for the desired wettability.

Figure 2.14. An example of free energy phase diagram constructed for pillar shaped geometry. The Wenzel and Cassie-Baxter wetting state regions can be identified and the difference between the two can be calculated.[23] 40

Thorough thermodynamic analyses with different model surface morphologies have been carried out in a similar manner, in order to demonstrate the presence of different wetting and transitions between them. These include surfaces with single layer roughness and surfaces with hierarchical roughness. The most common roughness geometry used for theoretical predictions is the pillar shaped surface feature. A theoretical study by Liu, T. et. al. highlights the importance of hierarchical structure in designing an anti-condensation surface, i.e. a surface which is not wetted under water condensation. It was shown that, for a surface made up of hierarchical patterns of pillar shaped features, a transition from Wenzel to Cassi-Baxter of a condensed water droplet is thermodynamically more favorable and would take place almost spontaneously. [100]

2.5 Self cleaning property of superhydrophobic surfaces

Self cleaning with water, also referred to as wet self cleaning, is a very important consequence of superhydrophobicity of surfaces.[101] Due to low contact angle hysteresis of water on these surfaces, water easily rolls off of the surface and takes the dirt particles away as it is rolling off. It is the self-cleaning ability of the lotus leaf for which it is called a symbol of purity. It is required for a droplet to be in a Cassie-Baxter state in order to effectively show self cleaning behavior. Wenzel state of a water droplet usually is associated with its loss of self cleaning as well.

2.6 Loss of superhydrophobicity

As a consequence of the multiple wetting states possible for a superhydrophobic surface, and the transitions between the states, the superhydrophobic property of the surface could be lost under different conditions, such as in the case where the Wenzel

41 state is the most thermodynamically favored. This is the biggest challenge that superhydrophobic surfaces most commonly face, and is the biggest limitation that is imposed on using these surfaces in applications where their water repellency is crucial.

The loss of superhydrophobicity may or may not be reversible. As discussed in the previous section, it is the transition from Cassie-Baxter to Wenzel state that is associated with the wetting of a surface or the loss of superhydrophobicity. In an attempt to overcome this challenge, there are typically two approaches adapted by the researchers:

(1) creating a surface which shows Cassie-Baxter as the most stable wetting state and (2) manipulate surface wettability such that the transition from Wenzel to Cassie-Baxter could be made possible to reverse the ‘wetting’ of a surface.

In an attempt to design a surface with thermodynamically favored Cassie-Baxter state under condensation, a number of studies took an advantage of surface tension gradient. A Marangoni flow resulting in tear drops of wine is a common example of kinetic motion induced due to surface tension gradient.[102, 103] Based on the same principle, a continuous radial surface tension gradient was created by radial diffusion controlled silanization of the monolayer formed on the surface. When the resultant wettability gradient surface was brought in contact with condensing steam, it was observed that the condensing droplets coalesce and move from hydrophobic center to hydrophilic edges (radially outwards), with velocities two to three orders of magnitude higher than typical Marangoni flow; the effect is attributed solely to the surface tension gradient.[104] A similar vertical surface energy gradient has been proposed to be responsible for the retention of superhydrophobicity of the natural lotus leaf surface under water condensation or dew formation. The composition of the wax like material

42 and the hair density variation are two components which could result in a surface energy gradient on a lotus leaf surface.

Similar to Cassie-Baxter to Wenzel transition, a reversible Wenzel to Cassie-

Baxter transition could be brought about by a number of different external stimuli. A surface of lotus leaf held fixed on a cold plate was observed to result in wetting as water condensed on the surface. A vibration induced reverse transition was shown to restore its superhydrophobicity effectively.[105] The coalescence of condensed droplets has been shown to play a crucial role to drive the Wenzel to Cassie-Baxter transition. . The conversion of interfacial energy of the droplets in Wenzel state into kinetic energy is the transition mechanism which results in either jumping of droplets and resting on the surface in the Cassie-Baxter state, or in droplets getting rolled off the surface without leaving any residue behind.[17, 106, 107] This is purely an interfacial energy driven transition and does not require any external stimulus. The presence of nano-scale roughness features or hierarchical structures with nano-scale features as a part of the structural hierarchy have been shown to facilitate Wenzel to Cassie-Baxter transition. The exact mechanism behind the ability of nano-roughness to aid in transition is, however, not very well understood.

Metastability of the superhydrophobic property of a surface underwater also restricts its use in underwater applications. The plastron layer, i.e. the layer of air entrapped between a surface and water which is responsible for superhydrophobicity of a surface, decays rapidly as the superhydophobic surface is immersed in water. The rate of decay, and thus, the wetting transition depends on the depth of penetration and, effectively, the hydrostatic pressure of the water column. The thermodynamic instability

43 of a plastron layer was demonstrated by Poetes, R. et. al. in the case of a natural superhydrophobic lotus leaf surface, along with a synthetic mimic fabricated using PTFE.

For a superhydrophobic lotus leaf, the plastron life time was estimated to be about 1 hour at a penetration depth of 55cm.[108] The plastron decay life can be expected to be differ for different superhydrophobic samples and depths of underwater penetration. The lifetime, however, can be expected to be of the order of few hours for a typical superhydrophobic surface, making it impossible to use in applications involving continuous exposure to water. This may not hold well for a surface with a thermodynamically favorable Cassie-Baxter state. A theoretical possibility of underwater superhydrophobicity was analyzed by Marmur A. It was shown that such a superhydrophobicity could be possible for the surface with very high roughness ratio. In addition to surface forces, the hydrodynamic forces underwater, which are not accounted for in wetting predictions, could also result in the failure of superhydrophobicity. [109]

It is important to note that it is not always the thermodynamic instability of the

Cassie-Baxter wetting state of a superhydrophobic surface that is responsible for its loss of superhydrophobicity. As discussed in the previous sections, the surface is made up of a number of extremely fine structures. The structures are typically made up from a hydrophilic material and are coated with a thin layer of hydrophobic material, in order to make it superhydrophobic. The hydrophobic coating has a very poor wear and abrasion resistance and tends to come off easily due to its poor adhesion to the surface, in the case of mechanical contact. The similar limitation is observed for the hierarchical structures of the surface roughness patterns, where multiple tiers of roughness are not very well bound to each other. In the case of surfaces with a very high aspect ratio (such as fiber networks

44 and nano-tube structures), capillary collapse of the structures is also responsible for loss of superhydrophobicity.[110] The mechanical wear of the surface is surely an irreversible damage and in such a case no restoration of superhydrophobicity is possible. Thus, in addition to accounting for surface energy and optimum roughness, superhydrophobic surfaces should also be tested and tuned for mechanical and abrasion resistance.

2.7 Adhesion of structured surfaces

Adhesion is defined as the tendency of two dissimilar surfaces to interact at their contact interface in order to form a bond, also referred to as “adhesive interactions” between the surfaces. Adhesion is an attractive interaction between two materials. In the most commonly used conventional sense, adhesion is defined for a pair of “adhesive” and

“adherent”; an adhesive being an intermediate layer sandwiched between two similar or dissimilar surfaces in order to hold them together and adherent is the surfaces which are being adhered using an adhesive. The strength of adhesion depends on the interactions between the adhesive and adherent materials. Mechanical interactions and chemical interactions (at molecular scale) are the two main categories of interfacial interaction resulting in adhesion between two surfaces. Some examples of different types of interactions resulting in adhesion between the two surfaces are shown in figure 2.15.

Figure 2.15. Schematic representation of some common types of interactions resulting in adhesion between the two surfaces, (a) capillary forces, (b) suction, (c) mechanical interlocking (Velcro) and (d) electrostatic interactions

2.7.1 Mechanical adhesion

Mechanical adhesion is one of the key contributors to the adhesion between two surfaces. The strength of mechanical adhesion solely depends on the surface roughness; the higher the surface roughness, the higher the area of contact, and thus, the higher the number of contact points at the interface, and effectively, higher adhesion. Velcro is a very common example of an adhesive which is strongly dependent on mechanical interlocking mechanism.

2.7.2 Chemical adhesion

The chemical nature of the two surfaces coming in contact plays an equally important role in adhesion between the two, in addition to mechanical interlocking. The polar versus non-polar nature of the functional groups, the surface composition at the

46 molecular scale and the charged versus neutral surface are some of the very important factors which decide different types of chemical interactions possible between the two surfaces. Also, the type of chemical interactions also defines the strength of adhesion and its possibility of being reversible. The types of chemical interactions include capillary forces, electrostatic interactions, covalent bonding, hydrogen bonding and van der Waals interactions (some of these are schematically shown in figure 2.15 above).

2.7.3 Work of adhesion

The surface energy and the surface tension are physical parameters that are used to account for all the possible interactions between two surfaces coming in contact. The surface energy is defined as the energy spent in creating unit surface area of a given material. Mathematically, both the physical quantities have the same dimensions. Most commonly, however, surface energy is used for solids, whereas surface tension is used for liquids. The work of adhesion is defined as the energy required to separate unit area of contact between the two surfaces, 1 and 2, as shown schematically in the figure 2.16.

Figure 2.16. Schemtic representation of work of adhesion [24]

Work of adhesion (W12) can be estimated using following equation:

W12 = γ1+γ2-γ12………………………………………………………………………(2.6)

Where, γ1, γ2 and γ12 are the surface energies of component 1, 2 and the interfacial energy at contact interface of 1 and 2, respectively. In the case of work of adhesion between a solid and a liquid surface, the equation 2.6 in combination with Young’s equation for equilibrium contact angle for a solid-liquid pair (as discussed in section 2.2.2) gets reduced as follows:

WSL = γL-V(1 + cosθY)…………………………………………………………………(2.7)

2.7.4 van der Waals interactions

Amongst different types of surface forces, van der Waals (vdW) forces are the weakest intermolecular forces. However, they are always present between two molecules, irrespective of the nature of molecules. It is, thus, not straight forward to understand the nature of these forces. The vdW forces are long range and are effective over the distance as large as 10 nm down to 0.2 nm between two interacting entities.[111, 112] The vdW energy of interaction (W) is directly proportional to Hamaker constant (A) and intermolecular distance (r) as follows:

………………………………………………………………………….(2.8)

A is calculated based on Lifshitz theory, using following equation:

……………………………………………………………………...(2.9)

…..(2.10)

Here, 1 and 2 correspond to two phases interacting across medium 3, is a dielectric constant, is refractive index and is the main electronic absorption frequency.

2.7.5 Pressure sensitive adhesives (PSA)

For any adhesive system to work efficiently, it needs to have liquid like wetting properties to be able to spread over the large area of contact of the surface easily. After spreading, solidification follows with the help of external stimuli such as heat, light or some chemical reaction. The solid like properties are required to impart the required mechanical strength to the adhesive. A single component pressure sensitive adhesives

(PSAs) were developed as an alternative to conventional adhesive systems. PSAs are made up of viscoelastic materials which possess properties of both solid and liquid. The use of PSAs also eliminated the use of any external stimuli for their applications. PSAs could also be used for reversible bonding and de-bonding. However, since they are made up of viscoelastic materials, there are some intrinsic shortcomings. Viscoelastic PSAs usually have much low moduli which limit the maximum stress that these adhesives can bear. The tacky surface of these adhesives makes them highly susceptible to contamination.[113, 114, 115] Thus, an ideal PSA system which can be bonded and debonded effectively must be made up of an elastic material. However, elastic materials are not compliant enough to generate an intimate contact. The answer to this problem lies in one of the natural adhesive systems, gecko feet structures. The gecko adhesive system has been a topic of interest for scientific communities over years now. It is a unique adhesive system that is dry, can be bonded and de-bonded reversibly and still shows very high adhesive strength (unlike viscoelastic reversible PSAs).

2.8 Gecko adhesion

Geckos are lizards that are known for their unique adhesive ability. They have adapted to stick to almost all surfaces, interestingly without the use of any glue. Another interesting fact about gecko adhesion is that their toe pads allow them not only to stick firmly to any surface as they hold onto it, but also to walk on a vertical wall or even a ceiling with their body upside down. The gecko adhesion is, thus, known as reversible adhesion. The conventional “glue” based adhesive that possesses a similar reversible reusability is the one that is most commonly used in sticky notes. The adhesive strength of the latter however, is nowhere comparable to gecko adhesive system. The gravity defying ability and reversibility of gecko adhesive system have triggered interest in the adhesion, material science, and biomimicry fields. The focus is to study the properties of the gecko toe pad and mimic them into a synthetic adhesive which may find numerous applications in robotics, electronics, space engineering, and biomedical fields. “The

Spiderman” could actually be a reality in the near future!

The reversible adhesive property of geckos is attributed to the hairy structure on their toe pads. Three tier structural hierarchies are observed on the surface of gecko feet, as shown in the images in figure 2.17.

(a) (b)

Figure 2.17. The hierarchical structure of gecko foot is shown. Each toe is made up of lamellar strips (a), every lamellar strip is made of an array of microscopic hairs called setae (b and c) and one seta branches out into hundreds of siner structures called as spatula (d). [25, 26]

Each foot of gecko is made up of about 7-8 lamellar strips of hairs, which are spaced and oriented in an orderly manner. A lamella essentially is made up of thousands of microscopic hairs called “setae”. Typically, a seta is about 60 µm long and about 10

µm wide. A seta consists of numbers of hairy strands intertwined to form a cord like structure. The tip of every seta branches out into finer hairy structures, called “spatulae”.

The origin of this nomenclature lies in the fact that the tips of these finest hairs are spatula shaped (as shown in figure 2.17d). The spatula is about 200 nm wide. The number

51 density of spatulae could vary from anywhere between 100-1000 per seta, depending upon the type of a gecko species.[25, 26] Although the structure and morphology of gecko toe pad is well understood, the composition of the inherent material, which the hairs are made up of, is not yet completely known. It has been showed that one of the major constituents of the gecko’s hairy structures is β-keratin-like protein.[116] It was also demonstrated recently by Hsu, et. al. that geckos leave footprints behind as they walk. The prints correspond to phospholipid molecules.[117] Thus, the gecko hairs can be understood as a composite system made up of a protein and hydrocarbon-like phospholipid material. However, what the exact bulk and surface composition of the hairs are still remains unanswered.

The adhesion mechanism of gecko toe pads is a two step process. As the animal takes every step, a foot is brought normally in contact with the surface followed by shearing action in the opposite direction of motion. The gecko adhesive toe pad is thus also classified as a ‘pressure sensitive adhesive’. The hairs on the gecko toe pads are flexible, deforming easily under the application of load. Thus, upon shearing the hairs can easily be bent to create higher area of contact between the toe pads and the surface underneath, compared to the contact area formed when the two come together in the normal direction. The bending of the hairs is completely reversible. In order to take the next step with the same foot, the animal peels the foot off of the surface using hyperextension muscular motion. The most efficient peeling angle has been identified to be close to 30°.[118]

The nature of the adhesive forces responsible for gecko adhesion has been a topic of interest and debate over the years. It was proposed and demonstrated experimentally

52 that the gecko adhesion is achieved purely based on van der Waals interactions between the toe pads and a contact surface.[119] This implied that the strength of adhesion is dependent only on how much contact interface is formed and independent of surface chemistry of contact surface. Later, it was pointed out by Huber et. al. that the capillary forces are responsible for higher adhesive forces produced by spatulae at ambient humidity conditions. [120] Thus, it was expected to get better adhesion from a hairy surface made up of a hydrophilic material since it is the tip surface that comes in intimate contact with the contact surface. Although this may hold good at the spatula level, the contribution of capillary forces was not reported at the setal or animal scale, thus, the effect could be minimal compared to the overall adhesion of the system. The presence of water and high humidity, however, has significant impact on material properties and, effectively, adhesion, as will be discussed in the following sections.

2.8.1 Effect of environmental factors

The animals encounter a number of environmental fluctuations, such as rain water and temperature and humidity variations. They also need to be able to walk on different types of natural surfaces, such as tree barks, plant leaves, and rock surfaces. In all of the scenarios, they need to have appropriate functioning of their adhesive system, failure of which could be fatal for their survival. The gecko toe adhesive is, thus, yet another example of nature’s magnificent engineering. Even though it is expected that the gecko’s clinging abilities be independent of temperature and humidity fluctuations, because the interfacial forces are of van der Waals in nature, when studied for the two factors it was observed that the whole animal scale adhesion is sensitive to temperature; it is higher at lower temperatures (about 12°C) than higher temperatures (about 35°C). Humidity

53 variations were reported to have a very intriguing effect on overall animal adhesion. At low temperatures, high humidity was observed to have a synergistic effect, and increased adhesion. The adhesion trends at higher temperature tested, however, were almost insensitive to humidity fluctuations.[121]

The effect of water on gecko adhesion is another interesting aspect of the gecko adhesive system. It was not until recently that the role of water on the adhesion mechanism was studied extensively and its important technological applications were realized. The surface of gecko toe pads is superhydrophobic. Water, in contact with the surface, beads up to form a spherical droplet with a contact angle of about 150°, and rolls easily off the surface- the characteristics typical of a superhydrophobic surface. The superhydrophobicity of the gecko toes was used to show the very low contact fraction of hairs in their non-loaded, non-sheared state. The surface wetting property was used to emphasize the occurrence of appropriate loading and shearing of the toe surface by

Autumn, K. et. al.[122] The water repellency of the gecko toe surface was reported to be lost over time. It was demonstrated that the non-sticky water droplet in contact with the setal patch surface became sticky when maintained in contact for about 20 minutes, and showed necking behavior. The change in wettability was attributed to possible conformational changes of the surface functional groups, and expected to lower the adhesion, based on van der Waals interactions predictions. Another interesting aspect of gecko toe wettability was reported by Hsu et. al., in which, with the help of surface sensitive spectroscopic technique, it was shown that a thin layer of water (typically, a monolayer) present on the surface does not interfere with the contact between the surface and the gecko toe, i.e. a dry contact can be easily made with the surface, as the toe expels

54 the water layer away.[117] This could explain the retention of adhesive strength of toe pads in highly humid conditions. When tested for their adhesion on wet surface having a thick layer of water (~ 0.5 cm deep), it was observed that the geckos tend to slip on the surface, even though their toes remain dry.[123] This pointed towards a critical limit of superhydrophobicity of the toe surface, which can repel water column of certain height

(less than 0.5 cm) to establish a contact with the surface underneath. It has been observed that gecko toes do not wet easily even when held underwater. The shiny silver appearance of the toes corresponding to the air plastron layer between the hairs is a common indication of dry toes underwater. Even though it is a default state of the toes, it has been reported that the toes can be wetted for forcing a wetting transition by rubbing their toes underwater, which forces water to replace air entrapped between the structures.

The appearance of the toes on wetting changes to grey color which is distinctly different compared to the dry toe appearance underwater. Although reversible, wetting of the toes is unfavorable for the geckos as they cannot stick to any surface with their wet toes. Also, they fail to stick to wet hydrophilic surfaces even when their toes remain dry. Both of these observations represent adhesion failure modes for the animals.[123]

2.8.2 Self cleaning of gecko toe pads

In addition to reversible adhesion and superhydrophobicity, another unique feature of gecko toe pads is their self cleaning ability. Geckos contact numerous dirty surfaces and are still able to retain adhesion without any grooming. Hansen et. al. demonstrated the self cleaning ability of the toe surface based on the mechanical contact model. It was shown that the energy equilibrium mismatch at the dirt-surface interface

55 versus seta-dirt interface drives the self cleaning of the surface. It typically takes the animal about 4-5 steps to restore its adhesion almost completely.[25]

2.9 Plasma Enhanced Chemical Vapor Deposition (PECVD) or Plasma polymerization

Plasma is an ionized gas, and is also referred to as the fourth fundamental state of matter, along with solid, liquid and gases. The electrical gas discharge glow is a typical characteristic of the plasma state. Lightening in the sky or electric sparks are some of the common examples of plasmas. Plasma can be created by heating the gas up or by the electric breakdown caused by application of high voltage across a rarefied gas. Although plasma based processes such as metal welding and corona discharge have been used for years in industrial applications, the ability of plasma to deposit thin organic films of highly branched or cross-linked polymer-like materials was discovered relatively recently. Plasma polymerization is also referred to as Plasma Enhanced Chemical Vapor

Deposition (PECVD) of organic molecules, which is, in fact, a more appropriate nomenclature for the process. PECVD is a very different category of chemical reactions compared to the conventional polymerization since, even though there are initiation, propagation, and termination steps taking place in the PECVD process, which resemble those occurring in the case of conventional polymerization; the PECVD material is insoluble and infusible film unlike the forms in which polymers are formed. This is also the reason the PECVD process of organic molecules has never drawn enough attention in the field of polymer science. The process however, offers a number of advantages and potential uses in the fields of surface functionalization, modification, and coatings, as will be discussed in later sections.

2.9.1 PECVD nomenclature

Precursor and ionization gas:

PECVD chemistry involves gas phase chemical reactions. The reaction mixture in the plasma process is made up of two gases, precursor gas and ionization gas. The precursor gas is typically an organic monomer chosen depending upon the surface functionality that is to be imparted. The precursor does not necessarily have to be a gas, it can be a solid or a liquid as well; in which case it needs to be either sublimed or vaporized to get into a gas phase before plasma can be generated. The ionization gas is typically an inert gas such as Argon, Helium, Nitrogen, Oxygen, or Air. The function of ionization gas is to generate reactive species which consist of ions and free radicals. The reactive species then collide with precursor gas molecules to generate precursor reactive species. A number of different reactions such as recombination, linear addition or cross- linking of precursor molecules result in the formation of three dimensional network of monomer which gets deposited on the solid surface. One of the biggest advantages of

PECVD of organic molecules is the fact that there is no surface specificity of the deposited layer, i.e. irrespective of what the chemical nature of the substrate is, a film with more or less uniform thickness can be deposited.

Carrier gas:

A carrier gas is an essential component in atmospheric pressure operated PECVD unit. The purpose of the carrier gas is to generate high enough flow rates so that the reactive species are carried to the designated plasma length and the glow is sustained. It may or may not be used in vacuum operated plasma set-up depending upon the range of

57 vapor pressure of operation. In many cases, the ionization gas also serves the purpose of carrier gas. In the case where two different gases are chosen for two different purposes, a carrier gas needs to be chosen which is relatively passive compared to the ionization gas in the inlet gas stream.

Plasma power:

The power specified for a given PECVD experiment corresponds to the peak power of the AC or DC signal used to generate plasma.

Duty cycle:

In the case of AC source used for the excitation of the plasma, duty cycle refers to the fraction of one cycle for which the plasma is in its “on” state. It is usually expressed as a percentage. The rate of deposition of the films and the structural retention of the organic functionalities depend strongly on the duty cycle used in the PECVD process.

Continuous versus pulsed plasma:

In the pulsed plasma, flashes of plasma are created which correlate to the duty cycle used. On the contrary, continuous plasma refers to constant glow of the plasma throughout the deposition process.

Vacuum based PECVD system:

The vaporization of plasma precursor is most commonly carried out under vacuum, most commonly of the order of few mTorr. The reactor types, classified mainly based on the type of electric field and coupling used, include internal and external electrode based PECVD and electrode-less PECVD reactor set-up. Most commonly AC

58 voltages are used for plasma excitation. The excitation frequency can be anywhere from a few kHz to radio frequency (MHz) to microwave frequency (GHz). The lower frequency glow discharges usually show DC glow discharge characteristics. The electrodes alternatively act as anode and cathode, typically when source frequency lower than 1 kHz is used. As the frequency is increased, positive ions in the vapor phase tend to immobilize and retain partial charge from previous cycle. Thus, when using AC frequencies in the range of kHz, there is always a chance of the plasma getting extinguished due to the loss of reactive species by charge neutralization or molecular recombination taking place in the gaseous volume. In the case of frequencies MHz or higher, these phenomena do not usually take place due to high frequency of the field used. Thus, a continuous plasma glow is easily possible.

2.9.2 PECVD of organic precursors

The use of PECVD for deposition of organic films, more commonly known as plasma polymerization, has gained enormous attention in a number of applications, which include the fields of protective coatings [124], biomedical applications [125, 126], water repellent surfaces (fabric coatings), semiconductor and electronics applications, fluid filtration and separation membranes[127], linear and non-linear optics, so on and so forth.

The biggest advantage that the process of PECVD offers is its ability to form thin films from almost any organic material in its vapor phase. Additionally, some of the key features of organic PECVD process are as follows [128, 129]:

(1) The types of monomers/precursors that are used in the PECVD process also

include the ones which could not be polymerized by means of conventional

polymerization.

(2) The films deposited using PECVD do not show any surface selectivity, i.e. the

films are bonded equally well to different types of surfaces such as glasses, metals

and plastics.

(3) It is a completely dry, solvent-less process. Thus, it has a potential to serve as a

“green” substitute process to a number of wet processes which tend to generate a

lot more volatile organic content (VOC).

(4) The thickness of the film could vary from a few nanometers to micrometers.

The usually non-reactive monomers, such as saturated hydrocarbons (alkanes), could also be polymerized using this technique. Even though such a polymerization is possible, it was shown that the rate of PECVD deposition was faster when the precursor had higher unsaturation, i.e. the rate of deposition was the fastest for acetylene, slower for ethylene, and the slowest for ethane. The presence of unsaturation in the monomer structure also resulted in the higher unsaturation and higher level of oxidation in the plasma deposited layer. The physical form of the PECVD layer also depends on the rate of deposition and thus, effectively the precursor unsaturation. It has been observed that high deposition rates result in a powdered form, an intermediate rate forms either powder or film, and a lower rate produces only films. In spite of limited control over quality and composition of the PECVD films formed with non-reactive monomers, the fact that functional films could actually be formed with the help of PECVD is surely very unique feature of this process.[130]

2.9.2.1 Kinetics of the PECVD reaction (mechanism of plasma polymerization)

Figure 2.18. A schematic representation of different reactions taking place in a typical

PECVD of organic molecules [27]

There have been a number of models developed to describe the kinetics of

PECVD process. Mostly all of these models rely on a ‘free radical polymerization’ mechanism. The schematic representation of the PECVD mechanism is shown in Figure

2.18 above. The two dominant mechanisms in the gas phase reaction include “plasma

61 induced polymerization” and “plasma polymerization” as defined by Yasuda, H. et. al.

The case of plasma induced polymerization can be understood to take place similar to that of a chain addition polymerization. On the other hand, in the case of “plasma polymerization”, stable intermediate gaseous species are formed, which could follow one of the two routes: first, the stepwise activation followed by addition could lead to the

PECVD deposition and second, remaining in the gaseous state as byproducts. Another unique characteristic of PECVD is the process of ablation, i.e. surface etching by the reactive gaseous byproducts formed (most prominently H2, O2, or F2 gases). It is analogous to the process of “de-polymerization”. Thus, it mostly is detrimental to the deposited layer. All of the three mechanisms occurring simultaneously in a typical

PECVD process were studied using model hydrocarbon and fluorocarbon precursors

(C2H4 and C2F4). The discrete energy levels defined in terms of W/Fm (the ratio of plasma power to the flow rate) corresponding to the three different types of reactions were empirically estimated and plotted as a function of Fm, as shown in the Figure 2.19 below.

The energy level diagram, thus, helps in determining the choice of reaction parameters depending upon the regime in which the deposited layer is required to fall. It has also been shown that the overall process kinetics are affected by monomer flow rate. If the flow rate is too small, there is not enough monomer concentration to generate reactive species and propagate reaction continuously, resulting in a low deposition rate. On the other hand, in the case of very high flow rates of monomer, it was observed that the residence time of not only monomers, but also of reactive species, was too short to reach the substrate, resulting in the decreased deposition rate. Thus, optimization of the flow

62 rate to create balance between the two competing processes is a crucial factor for PECVD process.

Figure 2.19. Empirical energy diagram for the different types of reactions taking place in a typical PECVD process.[27]

In the PECVD process, because the substrate surface on which the deposition takes place is subjected to the environment of highly energetic species, the surface may not always remain inert. In addition to the organic film deposition, the other processes, including surface cleaning, surface ablation, cross-linking of near surface molecules, and

63 modification of surface composition, can affect not only the surface but also the film properties.

2.9.2.2 PECVD of fluorinated molecules

Perfluorocarbons and fluorinated hydrocarbon based polymers or coatings are very interesting due to the electrical and chemical inertness of the basic C-F bond dominantly constituting the molecules. Additionally, the fact that the functional groups

CF2 and CF3 are among those with the lowest possible surface energy makes them useful in the non-stick and water repellent applications. The chemical structure of the fluorinated precursors (analogous to monomers used in polymerization) was thoroughly investigated by Yasuda, H. et. al. It was observed that the rate of plasma activation and the deposition of thin films using PECVD is slower for fluorine containing monomers compared to corresponding hydrocarbon monomers, an effect that was attributed to stronger C-F bond, which disabled the detachment of fluorine atoms. The presence of unsaturation or cyclic structure in the precursor structure was shown to accelerate the rate of PECVD. [131] The effect of process parameters on the quality and the chemistry of fluorocarbon containing polymers deposited using PECVD was investigated further by

Yasuda, H. et. al. which revealed a dependence on the factor W/Fm, where W was the discharge power and Fm was the monomer flow rate. The correlation between the two showed that the higher the flow rate of monomer used, the higher the power needed to be to sustain the plasma glow. It also was demonstrated that beyond a critical value of W/Fm, monomer decomposition is a dominant mechanism which resulted in the irregular surface composition of the film.[132, 133] It was demonstrated by Masuoka, T. et. al. that addition of hydrogen gas to the reaction mixture of perfluorocarbon precursor

(hexafluoroethane, HFE particularly) accelerated the PECVD reaction rate provided the molar ratio of H2:HFE was less than 1. Interestingly for ratios exceeding 1, the H2 gas acted as a scavenger for reactive fluoride species, resulting in not only retarded rate of reaction but also in the ablation of reaction chamber surface.[134]

The most commonly used precursor for PECVD process of fluorocarbons is tetrafluoroethylene (TFE, C2F4) in order to mimic the properties such as low surface energy, low friction coefficient and low dielectric constant of polytetrafluroethylene

(PTFE). The choice of pulsed versus continuous plasma, plasma power, and duty cycle have all been shown to affect the surface composition i.e. relative percentages of -CF2-, -

CF3, and >CF functional groups on the surface. A single parameter, or a combination of two or more parameters, could be varied systematically to control the surface composition.[135] In addition to TFE, a number of other fluorocarbon molecules have also been explored to form PECVD films with fluorinated functional group moieties.

Unlike conventional monomers, which need to have reactive sites (unsatuartion or reactive functional groups) in their molecular structure in order to polymerize, the plasma enhanced process has been shown to efficiently deposit cross-linked polymer-like structures which act as functional films. The choice of fluorocarbon precursor depends on the desired end-use properties of the films. Some examples of unique fluorocarbon precursor systems, their properties, applications, and advantages that PECVD offers specifically are discussed below.

The PECVD carried out using tertafluoromethane (CF4) precursor was observed to form better quality pinhole free thin films, as compared to similar thin films that could be produced by surface etching of PTFE. The films formed were optically transparent

65 with a refractive index of about 1.35.[136] Fluorinated and perfluorinated alkanes and alkenes are among the commonly studied category of organic plasma precursors. The plasma created using a mixture of methane (CH4) and octaflurocyclobutane (C4F8) in a capacitively coupled plasma set-up formed thin films which showed excellent thermal stability up to 350°C and very low surface energy (the water contact angle measured was

100°). The same reaction mixture also was used to create a fractal surface which possessed superhydrophobic properties.[137] PECVD films of C4F8 deposited using an inductively coupled plasma set up on the surface of an anodic Al2O3 membrane were able to switch the wettability of the nano-porous membranes from superhydrophilic to superhydrophobic.[138] More recently, PECVD films of C4F8 were also demonstrated to function as hydrophobic valves in microfluidic biochemical applications.[139] A mixture of hexafluoroethane (C2F6) with either acetylene or butadiene offered a very well controlled surface energy variation, depending upon the relative compositions in the reactant mixture stream.[140] Perfluoropropylene oxide (C3F6O) and a mixture propylene and nitrogen trifluoride (NF3) on plasma depositions also allowed a similar control over surface composition. In the latter case, in fact a gradient of fluorinated molecules could be created by continuously changing the feed composition as plasma reactions were taking place.[141, 142] The PECVD films of perfluorocyclohexane (C6F12) were shown to have a very good structural retention with the use of pulsed PECVD[143], whereas perfluorooctane (C8F18) PECVD films exhibited excellent dielectric and insulation properties.[144] The PECVD of hexfluoropropylene (C3F6), in addition to the film deposition, was observed to form submicron particles of the precursor, which are integrated into the film[145] Perfluorohexenes (trimers of C9F18 compounds), with well

66 controlled plasma parameters were able to deposit films with –CF3 being a dominant functional group on the surface, a chemical moiety which has been shown to have the lowest surface energy among organic functional groups.[146] A PECVD deposition using a series of fluorocarbon precursors, ranging from saturated fluorocarbons (CnFn+2, n=1-4 and 6), unsaturated fluorocarbons (C2F4, C3F6) and a gaseous mixture of CF4 and H2 was demonstrated to be effective to impart water repellent and moisture resistant properties to hygroscopic nylon fabrics. This approach was particularly significant for selectively modifying the surface properties of fiber materials while maintaining their bulk mechanical properties.[147] The PECVD films deposited using pentafluoroethane

(C2HF5) and octafluorocyclobutane (C4F8) on the surface of paper and regenerated cellulose were observed to improve the hydrophobicity and barrier properties of cellulosic materials.[148] The fact that PECVD involves highly reactive energetic species which are capable etching even the metal surface, was taken advantage of in creating organometallic films. A CF4 precursor used to create plasma in a RF planar magnetron with an Al target resulted in the Al particles being incorporated in the deposited film.[149] The ultraviolet emission spectrum of CF2 radicals, which are the dominant reactive species in the PECVD of fluorocarbons, is well characterized, and has been used to study in-situ composition of reactive species.[150] This offers an advantage of estimating the number density of fluorinated species when the reaction is in progress, and to correlate this to the rate of deposition and the surface composition of the film that is deposited after reaction completion. It is particularly important for the PECVD process, since the mechanism and kinetics of the reactions are still not very well understood.

Some of the higher chain length fluorocarbon molecules, such as 1H,1H,2H-perfluoro-1-

67 docence and perfluoroacrylates, have been successfully used in the PECVD process. The films formed using these precursors have been reported to have possibly the lowest surface energy and find applications in creating oil repellent surfaces; not many synthetic or natural surfaces can show stable and prolonged oil repellency due to the very low surface tension of oils.[151, 152, 153]

2.9.2.3 PECVD of organosilicon compounds

Similar to fluorocarbon precursors, organosilicon compounds is yet another class of molecules which has been studied extensively as PECVD precursor for various applications, ranging from biomedical uses, protective coatings, optical lenses, anti- reflective and insulating coatings and dry lithography.[154]

The thin film formation using the PECVD process and the control of the surface composition based on hetero-atomic precursors with the general formula (CH3)3-Si-X-Si-

(CH3)3 was demonstrated by Inagaki et. al. (X = none, CH2, NH, O, S). Thus, PECVD can be identified as among the most versatile surface functionalization tools which is applicable for a wide range of monomer structures.[155] Some of the commonly used organosilicon PECVD precursors include hexamethyldisiloxane [HMDSO, (CH3)3-Si-O-

Si-(CH3)3] [156, 157], hexamethlydisilazane [HMDSN, (CH3)3-Si-NH-Si-(CH3)3] and tertraethoxysilane [TEOS, SiC8H20O4].[158] Films formed using vinyltrimethylsilane

[CH2=CH-Si-(CH3)3] and HMDSO [(CH3)3-Si-O-Si-(CH3)3 in a capacitively coupled RF glow discharge setup were shown to trap free radicals, which later could be reacted with oxygen and nitric oxide. The reactions such as bond reorganization and chain fragmentation were observed to be dominant in this case of PECVD and were expected to affect the final structure of the deposited film.[159] A film of HMDSO deposited on the

68 surface of titanium was also identified as a potential surface for fibronectin adsorption, mostly used in dental implant applications.[160] A thin layer of HMDSO and HMDSN on the surface of biodegradable polymer foils was successfully deposited using PECVD in order to make the surface hydrophobic, to avoid direct interaction of water with the bulk of the material. The PECVD coating, thus, acted as a water repellent protective layer while the biodegradability of the materials remained unaffected in spite of the presence of the plasma layer.[161] A film deposited on the metal surfaces (about 2 µm in thickness) using organosilicon precursors, such as HMDSO, HMDSN, and hexamethylcyclotrisilazane (HMCTSN, C6H21N3Si3) were observed to provide excellent corrosion protection in a highly corrosive simulated marine environment and higher thermal stability compared to conventional protective layers, an effect attributed to the highly crosslinked nature of the films.[162] The choice of a suitable organosilicon precursor depends on the efficiency of the structure to undergo plasma excitation and the following reactions. A comparison of hexamethyldisilane [(CH3)3SiSi(CH3)3] and tetramethylsilane [(CH3)4Si] showed that hexamethyldisilane more readily underwent the

PECVD reactions; the behavior was attributed to the presence of chromophoric Si-Si bond in the structure. [163] The atmospheric pressure operated plasma systems were also used to deposit PECVD films of HMDSO and HMDSN. In spite of higher oxidation, compared to vacuum processes, the plasma-enhanced films that were formed retained their transparency and low surface energy. Thus, it provided a proof of concept of using

PECVD based processes more conveniently at atmospheric pressure, compared to vacuum based glow discharge reactors which could also offer an advantage of forming films over larger surface areas.[164] A PECVD carried out for methylsiloxane and

69 methylsilazane precursors allowed the analysis of intermediate oligomeric species, which were identified to be mostly dimeric. The analysis was used to concretely develop reaction kinetics models and to identify the conditions under which ionic mechanism dominates over the free radical.[165]

2.9.2.4 PECVD of maleic anhydride

Maleic anhydride is an interesting organic reagent, since it has two different functionalities: a carbon double bond and an anhydride functionality (as shown in the structure of molecule in Figure 2.20).

Figure 2.20. Chemical structure of maleic anhydride molecule, showing two surface functionalities

Due to two types of functionalities being present in the same molecule, maleic anhydride can be used in various synthetic chemical reactions, which include Diels-Alder reactions, photochemical reactions, homopolymerization, copolymerization and surface grafting. The films deposited of maleic anhydride using PECVD are of particular interest, since they can impart a surface with two different functionalities which are reactive sites, and could be taken advantage of to act as a tie layer between the two surfaces to improve the interfacial adhesion and to create surface graft in the form of desired functionality.

The high surface energy of the films is also responsible for easy spreading and high wettability by liquids (aqueous and organic based systems). Thus, the films act as efficient adhesive tie-layers. Deposition from vapor based plasma reaction offers an advantage of being a single-step solvent-less process.

The systematic variation of reaction parameters demonstrated that the highest retention of anhydride functionality was possible when the mildest plasma conditions were used, i.e. pulsed mode of plasma and low duty cycles were identified to be the most favorable for anhydride functionality.[166, 167] In a study to determine the surface reactivity of the maleic anhydride films deposited using PECVD, the deposited films were derivatized using either decylamine or benzylamine. A uniform coverage of amine moieties correlated to an abundance of anhydride functionality of PECVD deposition. A

PECVD layer deposited on a surface already having a self assembled monolayer on it was further seen to improve the adhesion of maleic anhydride films to the substrate for the applications involving prolonged underwater exposure. The reactivity of maleic anhydride, thus, offers a flexibility of forming multilayer laminates of thin films, depending upon the target application.[168] The anhydride functionality of the PECVD films can be hydrated underwater to form a carboxylic acid functionalized (-COOH) surface. The acid functionalized film was shown to effectively support a negatively charged DMPG lipid bilayer using Ca2+ as a chelating agent.[169] PECVD functionalization thus serves as a very easy process for protein adsorption on any surface of interest. The anhydride groups of maleic anhydride were also shown to irreversibly bind to bovine serum albumin (BSA).[170] An ability of the functional films to bind to different surfaces, even in aqueous media and ability to retain anhydride linkages upon

PECVD deposition which show high reactivity towards amine functional moieties, make the maleic anhydride films a potential tunable surface for bioengineering. The ease of film deposition and the well controlled film properties using PECVD process are added advantages for biosurface and interface applications.

CHAPTER III

EXPERIMENTAL SECTION

3.1 Fabrication of single and dual layer particle patterns

Fabricating PECVD coated Hexagonal non-contiguously closed packed (HNCP) arrays of spherical particles is a two-step process. In the first step, the single or dual patterns were formed with suitable particle size chosen and predetermined inter-particle distance. In the second step, the surface of the particles was coated with a layer of suitable PECVD coating to impart desired surface functionality. The contact angles were measured on these surfaces either by depositing a droplet or by condensing a droplet of water on surfaces. The model calculations using hexagonal unit cell account for inherent wettability of PECVD coating and long-range order of a hexagonally patterned surface.

3.1.1 HNCP patterning using colloidal self-assembly of spherical particles

To form HNCP arrays of particles, silica particles (diameters from 100 nm to 750 nm,

Fiber Optic Center Inc.) were first functionalized using a mixture of p-(t-

Butyl)phenethyltrichlorosilane (BPTCS, Gelest) and 2-(4-chlorosulfonylphenyl)ethyl trichlorosilane (SPTCS, 50% v/v in methylene chloride, Gelest) followed by suspension in isopropanol (Reagent grade, Aldrich). Polystyrene particles (2700 nm diameter,

Interfacial Dynamics) with sulfonate functional groups were suspended in isopropanol after purification using four centrifugation-re-dispersion cycles. The particle suspension

73 was spread at the air-water interface in a Langmuir trough (NIMA). The film thus formed at the air-water interface was then compressed using the movable barriers at a controlled rate to a specified final area. The suspension concentration and the final area to which the particle film was compressed were predetermined with calculations for the desired center-to-center particle distance. A silicon wafer coated with a thin adhesive layer

[poly(n-butyl acrylate-ran-N,N-diethylamino ethyl acrylate)] was brought in contact with the particle film and the particle film was then transferred to the substrate. In the case of two-tiered roughness, the first layer of the particle film (2700 nm polystyrene particles) was coated with the same adhesive and was used to transfer a second layer (250 nm silica particles) of HNCP array of particles from the air-water interface. Additional details of the experimental procedure are described elsewhere.[171]

3.1.2 Morphology characterization of HNCP patterns

A JEOL JSM-7401F field emission scanning electron microscope was used for

SEM imaging of the patterns formed with colloidal lithography. The patterned surfaces were sputter coated before the imaging with a thin layer of silver using magnetron sputter coater. The SEM images were analysed using ImageJ software to calculate particle diameters (D) as well as to measure center-to-center particle distance (L). The Fast

Fourier Transform function in ImageJ was used to measure L by measuring the d- spacings in the crystal lines in FFT.

AFM was used to measure the height (H) of the particles after coating them using

PECVD. The AFM height image scanning was always started from a flat region on the sample of Si wafer where there were no particles. The difference between the flat line and the maximum height (measured as the scanning tip encountered the particle) is measured

74 as H for the sample. The AFM height images for atleast 25 particles from each sample were acquired and used to calculate the average H.

3.1.3 PECVD coatings of HNCP patterns

The rf-PECVD (radio frequency PECVD) process was used to deposit two different coatings on the surface of HNCP patterns formed by colloidal lithography. The precursors chosen for PECVD coating have distinctly different intrinsic wettability;

PECVD coating of 1H, 1H, 2H-perfluoro-1-dodecne (C10F21-CH=CH2 , 97% Matrix

Scientific) gave intrinsic contact angle ( ) of 110° ± 2° and that of hexamethyl disiloxane (O[Si(CH3)3]2) gave intrinsic contact angle of 95° ± 3°. was measured on the surface of flat silicon wafer with a layer of the same PECVD coating deposited. The

PECVD experiments were carried out in an electrode-less, inductively coupled vacuum chamber, as shown schematically in figure 3.1.

Figure 3.1. Schematic representation of PECVD setup operated under vacuum

As can be seen in the figure above, the metal coil wound around the chamber is connected to the 13.56 MHz radio frequency generator through an impedance matching network. The bell jar shaped glass reactor (vacuum chamber) is connected to the liquid nitrogen cold trap and a rough pump. A pressure gauge is connected at the outlet of the reactor and is used to monitor and regulate the vapour pressure.

Both the precursors chosen in this case were liquids. They were contained in a glass delivery tube with a control valve. The tube was connected to the inlet of the reactor. The PECVD for 1H,1H,2H-perfluoro-1-dodecne was carried out at 200 mTorr.

The vacuum chamber was run in a continuous mode at 35 W input power. PECVD was a three-step process. It consisted of precursor vaporization, plasma ignition followed by precursor vaporization as the last step. A typical cycle for PECVD of 1H,1H,2H- perfluoro-1-dodecne consisted of total 15 min, 5 min for each of the three steps described above.[172] The PECVD for hexamethyldisiloxane (HMDSO) followed the same sequence of steps. In the case of HMDSO, however, the precursor vaporization was carried out at about 250-300 mTorr to generate enough volume of vapour. The vapour pressure for resonance at which the plasma ignites in this case was however about 70-80 mTorr. Thus, after the first precursor vaporization step, the vapour pressure was reduced to about 80 mTorr and plasma was ignited without any delay. The rest of the steps were followed similar to that of the 1H,1H,2H-perfluoro-1-dodecne, making sure that the appropriate vapour pressure was maintained at each step. Clean silicon wafers were used as a flat reference surfaces since the patterned surfaces were also formed on silicon wafer surfaces.

The color of the plasma glow in the PECVD process can be used as an indicator of the molecule excitation since plasmas of some species have a distinct characteristic color. In this case, 1H,1H,2H-perfluoro-1-dodecne produces a cyan colored glow, whereas HMDSO plasma is pink/violet in color (shown in figure 3.2 below). It is difficult to distinguish HMDSO plasma from air/O2 plasma however, since both produce similar colored glow.

a b

Figure 3.2. Vacuum operated PECVD set-up. The cyan glow of 1H,1H,2H-perfluoro-1- dodecne (a) can be differentiated from a pink glow of HMDSO (b).

3.1.4 AFM measurements to measure R2 of HNCP patterns

The roughness of the PECVD coating layer (R2) needs to be accounted for in order to estimate overall roughness (R). We used Atomic Force Microscopy (AFM) for measurement of R2. The height images were acquired on a silicon wafer with a layer of

PECVD of 1H,1H,2H-perfluoro-1-dodecene, deposited under same conditions as that of the HNCP pattern of particles. We used Nanaoscope IIIa, DI multimode SPM (Digital

Instruments) for acquiring the AFM images. AFM tip with a resonant frequency between

300-400 kHz was used. The images were taken in a Tapping mode of the tip. For larger scan sizes (5μm - 20μm), scan frequency of 0.5 Hz was used, whereas for smaller scans

(500 nm - 2µm) scan frequency of 0.25 Hz was used.

3.1.5 XPS analysis for chemical composition of the PECVD coating

We used XPS to determine the chemical composition of the coating formed using

PECVD on the surface of HNCP patterns. We acquired XPS spectra using PHI Quantum

2000 XPS microprobe that uses Al Kα radiation. The data for survey spectra were obtained over 0eV-1100eV range of binding energy. We used 117.5 eV pass energy and

0.5 eV step size for survey spectra. These spectra were used to determine the atomic composition of elements on the surface. For further quantification, the high resolution spectra over narrower range of binding energies were acquired and analyzed.

3.1.6 Wettability studies of PECVD coated HNCP patterns

The contact angles for all samples were measured experimentally with a 10-12 μL droplet of de-ionized water. The measurements were carried out using Ramé-Hart

Instruments Advanced Goniometer 500 F1 with Drop Image Advanced software. The contact angle measurements were carried out at ambient temperature and pressure and as quickly as possible to avoid any loss due to water evaporation. At least three points were measured on each sample. For every point a total of 5 measurements were taken at an interval of 2 s each. Each measurement gave contact angle at the left and the right edge of the three phase contact line. The average of each measurement and finally that of total five measurements resulted in the average contact angle value and corresponding error in the measurement for one data point. The rest of the data points were acquired following the same method and final average contact angle value was reported for one sample. The advancing and receding contact angles were measured by tilting the base of the goniometer, while the sample was held fixed to the base. The base was tilted until the droplet rolled off, with advancing and receding contact angles recorded at every degree of

78 tilt. The difference between the advancing and receding contact angle just before the droplet rolled off was reported as the contact angle hysteresis.

The contact angles of condensed droplets were also measured on these surfaces.

For condensation experiments, the sample was introduced in a humidity room maintained at 90% ± 3% humidity and at a temperature of 30°C ± 1°C. The sample was allowed to equilibrate by letting it stand in the chamber for about 5 minutes. This ensured a thin layer of water formed on the surface. Maintaining high humidity for longer times so that large enough water droplets were condensed on the surface to measure the contact angles was challenging. Thus, after 5 min, an external droplet was deposited on the surface with a layer of water condensed on it already. The contact angle of the droplet deposited was measured using a goniometer and compared to the contact angle measurements carried out at room temperature and humidity.

3.1.7 Wettability model calculations for HNCP patterns

Figure 3.3. A schematic of unit cell representing HNCP structure of spherical particles is shown. In the case of dual roughness, the second layer of smaller particles also forms the same pattern, thus, the same unit cell can be used in the model for predicting the wettability.

To estimate surface roughness, the interfacial contact areas and the wetting state for the given design parameters, we proposed a hexagonal unit cell model that represents

HNCP pattern of spherical particles at both levels of roughness hierarchy. A model hexagonal unit cell is shown in Figure 3.3. As seen in the figure, each unit cell has six adjacent hexagonal unit cells. The dimension of the side of a hexagon is equal to L , where L is the center-to-center distance between the particles. Each unit cell contains one particle. Since one unit cell is representative of the whole surface, it is sufficient to calculate the surface roughness for one cell. D in the figure denotes the diameter of the spherical particle.

3.2 Carbon nanotube (CNT) based coatings

The synthesis and characterization methods of CNT based coatings are described in the following sections.

3.2.1 Synthesis of Carbon nanotubes

Mesh like CNTs are grown on acid treated stainless steel substrate using floating catalyst Chemical Vapor Deposition (CVD) process. The CVD process is carried out in a tubular reactor, shown schematically in figure 3.4 below. The furnace used has either two or three zones.

The stainless steel (SS 304) plates are acid treated by etching in 9 M H2SO4 for 10 minutes followed by washing with de-ionized water. CVD process used was floating catalyst method. As the name of the process suggests, the catalyst was not deposited on the surface; it is in fact a part of inlet gaseous mixture stream. Xylene was used as the carbon source and Ferrocene [dicyclyopentadienyl iron, Fe(C5H5)2] as the catalyst. The

80 acid treated stainless steel plates are heated to 750° C inside the furnace under Argon-

Hydrogen (85:15 v/v) flow atmosphere. The vapors of the solution of 1g of Ferrocene in

100 mL of Xylene are sublimed at 190° C and introduced in the furnace at a rate of 0.11 mL/min. The reaction was continued for 1 hour (6.6 mL of catalyst solution continuously injected in the flow), while maintaining the rest of the parameters constant (flow rates and temperatures). After the reaction was complete, H2 was turned off and the flow of Ar was reduced. The temperature of the furnace was reduced gradually in steps of 100°-

200° C. Argon flow was maintained till the furnace cooled off completely.

Figure 3.4. A schematic representation of a furnace used for Chemical Vapor Deposition

(CVD) process used for CNT growth

3.2.2 PECVD coating on the surface of CNT mats (formation of PCNT)

The mesh like structure is coated using rf-PECVD (radio frequency-Plasma

Enhanced Chemical Vapor Deposition). The precursor chosen for PECVD is 1H,1H,2H- perfluoro-1-dodecene (C10F21-CH=CH2) [97% pure, purchased from Matrix Scientific] 81 which is a representative of the class of long chain perfluoro alkyl monomers. The

PECVD experiments are carried out in a vacuum plasma chamber, similar to the process discussed in section 3.1.3. CNT mats formed on SS plate by CVD acts as a structured surface, whereas, clean aluminum plates and silicon wafers are used as flat reference substrates.

3.2.3 Characterization of CNT and PCNT structures

(1) Morphology

Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy

(TEM) are used to characterize the CNT mesh. JEOL JSM-7401F Field Emission

Scanning Electron Microscope was used for acquiring SEM images. Images were acquired under vacuum of 10-6 torrs and a potential difference of 5 kV. No sputter coating was required for the CNT and PCNT samples. TEM images were acquired using JEOL

JEM 1230 Electron Microscope and at potential difference of 120 kV. The diameter distribution of uncoated and plasma coated CNTs are obtained and compared by calculating diameters of about 130 tubes in each case. The diameters were calculated from TEM images using ImageJ software.

(2) Chemical composition

The IR spectroscopy and XPS are used for chemical composition analysis. FTS

3000 Excalibur Series DIGILAB IR spectrometer is used in the transmission mode.

Standard IR liquid cell is used to obtain IR spectrum of the liquid precursor. The plasma- enhanced films are deposited on the compression molded KBr discs formed with IR grade KBr powder (purchased from Sigma Aldrich) and the spectra are acquired in the

82 transmission mode. 64 scans are collected for IR spectra of both the background and the sample.

The XPS spectra are collected using PHI Quantum 2000 Scanning XPS

Microprobe instrument that uses Al Kα radiation. The survey spectra are acquired to detect the presence of elements on the surface. The survey scans are obtained over the entire range of binding energies i.e. 0eV-1000 eV with pass energy of 117.5 eV and step size of 0.5 eV. The survey scans are run for 5 minutes. The high-resolution spectra for

C1s and F1s regions are obtained over 20 eV binding energy width i.e. 280eV-300eV for

C1s and 680eV - 700 eV for F1s region. The high-resolution spectra in C1s region are acquired at 11.75 eV pass energy and 0.05 eV step size. The F1s region spectra are obtained at 58.7 eV pass energy and 0.125 eV step size. The high resolution scans in the

C1s and F1s regions are run for 10 minutes. These spectra are also used for the chemical quantification. TGA data for CNT and PCNT samples were acquired under N2 atmosphere and following the same procedure. We used TA Instruments TGA Q500 for the TGA experiment and analysis.

TGA analysis of CNT and PCNT samples was carried out using TA Instruments

TGA Q500TGA. The data for both the samples were acquired under N2 atmosphere.

3.2.4 Wetting studies of CNT and PCNT

Wetting behavior of plasma coated (PCNT) and uncoated (CNT) mesh is studied by contact angle analysis. CNT and PCNT surfaces are exposed to steam and analyzed visually. Immediately after exposure to steam, contact angles are measured on the surfaces at room temperature. The contact angle measurements are carried out using

Ramé-hart Instruments Advanced Goniometer 500 F1 with in-built software Drop Image

Advanced, as described in section 3.1.6.

The steam condensation experiments are carried out by exposing the sample to the continuous stream of condensing steam. The schematic representation of the experimental is shown in figure 3.5.

Figure 3.5. A schematic representation of steam condensation experiments

Steam is generated in a steam chamber composed of stainless steel and heated using resistive heating. The temperature of the chamber is monitored using a J- thermocouple and kept constant for all the measurements. The chamber is fed with water at a continuous rate. As the water enters the heated chamber it boils. This generated steam generated is carried through a stainless steel pipe to the sample block in which the sample plates are held horizontally. The entire assembly is thermally insulated to minimize the loss of heat to the atmosphere. The sample holder block is provided with a thermocouple and the temperature inside the holder is monitored continuously. The samples are exposed to condensing steam at atmospheric pressure and the temperature of

96° ± 2° throughout the experiment. Immediately after the steam exposure, the water contact angle is measured and plotted as a function of steam exposure time.

3.2.5 Wetting model calculations for CNT and PCNT surfaces

The CNT mats grown on stainless steel plates show non-aligned, mesh like morphology. In three-dimensional space, the morphology of the layer of CNTs formed can be visualized as a stack of number of such entangled meshes as shown schematically

in Figure 3.6a. Calculations for f s used in Cassie-Baxter model are based on a rectangular unit cell shown in Figure 3.6b.

Figure 3.6. The CNT and PCNT surface can be visualized as a stack of multiple layers of nano-meshes (a). The smallest unit of the mesh is a rectangle bound by four tubes such that the width and height of the rectangle are “t” and “h” respectively; it is the unit cell geometry used in the model calculations (b).

Figure 3.6b represents a single, topmost layer of a mesh like structure. In the

Cassie-Baxter state, the water droplet is assumed to be in contact with the projection of this single layer and not the entire thickness of the CNT mesh. The average length “t” and average width “h” of a rectangular cavity enclosed by four nanotubes are determined from SEM images using ImageJ software. In each SEM image, about 20 rectangular unit cells are chosen at different positions and an image is rotated as required to be able to draw a rectangle that fits the given unit cell and the values of “t” and “h” are calculated.

The same procedure is repeated for about 8-10 SEM images to calculate average values and standard deviations.

3.3 Gecko adhesion : The effect of surface wettability

The sample preparation and test methods for measuring the adhesive strength of gecko feet at whole animal scale and using sheds of toe skin are carried out as follows.

3.3.1 Gecko shed sample preparation

Samples were prepared from skin sheds that were collected from ten naturally molting Tokay geckos (Gekko gecko). Sheds were stored in a -20°C freezer until sample preparation. This method preserved both the adhesive setal fields and multiple lamellae of each toe that contact the surface during adhesion, similar to the configuration we expect for whole-animal studies. Previous reports suggest there is no significant change in the coefficient of friction in aged samples (~2.4 years old) when compared to those freshly removed from the animal[173] however as a precaution we did not use samples that were older than 1.5 years from the original collection date. Samples were made by cutting away the setal toe pad region from the rest of the foot shed (non-adhesive region) and mounting the toe shed on adhesive tape (Scotch tape) with the adhesive hairs facing up. Sheds were attached by applying pressure to the inter-lamallae region (hairless region) using forceps. This was done along the toe shed, taking care to not disturb the adhesive hairs. Excess tape was cut away from the sample so that none of the synthetic adhesive could interfere with adhesion trials.

3.3.2 PECVD coating on sheds

The vacuum operated PECVD set up, as discussed in details in section 3.1.3, was used for the PECVD coating of the sheds. The PECVD of Maleic anhydride was carried

86 out at about 100 mTorr vapor pressure and that of 1H,1H,2H-perfluoro-1-dodecene was carried out at about 200 mTorr.[172, 174] The PECVD process consisted of standard three steps, including precursor vaporization, plasma ignition, followed by lastly precursor vaporization, as discussed before.

3.3.3 Surface characterization of PECVD coated sheds

The blank sheds and shed samples with PECVD layer deposited were characterized using SEM and XPS. A JEOL JSM-7401F field emission scanning electron microscope was used for SEM imaging of the shed samples. The samples were sputter coated prior to imaging. The samples are biological and tend to get charged significantly in SEM imaging. To be able to view the surface features better the imaging was always carried out in LEI mode with very low accelerating voltage, typically 1-2 kV and emission current of 20 µA.

Chemical composition analysis was carried out using XPS. The spectra were acquired using PHI Versaprobe II that uses Al Kα radiation. The data for survey spectra were obtained over 0eV-1100eV range of binding energy using pass energy of 117.5 eV and step size of 0.5 eV. These spectra were used to determine the atomic composition of elements on the surface. The high resolution spectra were acquired over narrower range of binding energies (typically about 20 eV) using 11.75 eV pass energy and 0.1 eV step size.

3.3.4 Wetting and adhesion tests

The contact angle measurements were carried out using 10-15 µL droplet of de- ionized water, as discussed in details in section 3.1.6. At least two measurements were done for each sample and the average of the measurements was reported as the final

87 value. The same PECVD coating deposited on the flat Si wafer surface were used as flat

(non-structured) control samples.

The adhesion tests were performed at room temperature and 30-40% relative humidity. Samples were attached to a horizontally mounted force apparatus [121] using double sided copper tape, again cutting away any excess tape. A schematic of adhesion measurement setup is shown in figure 3.7 below.

Controller Glass plate

Motorized force Shed sample Nylon string gauge Hollow Stainless Steel box

Figure 3.7. A schematic representation of shear adhesion measurement set up used to measure adhesive force between a shed sample and a substrate (glass or OTS-SAM coated glass)

A nylon thread was fitted between two notches at the top of a clean glass slide and hooked to a motorized force reader which pulled at a controlled rate. Glass slides were cleaned in a base bath, blow dried with compressed N2 followed by oven drying at 120° prior to use. A second glass slide was coated with a self assembled monolayer of

Octadecyltrichlorosilane (OTS-SAM) to test the effect of surface hydrophobicity (the water contact angle on the surface of OTS-SAM was 95° ± 2°).[175] To form the OTS-

SAM, a glass plate was rinsed with de-ionized water to remove water soluble contaminants and dried with N2. Following this, the plate was rinsed with isopropyl

88 alcohol (IPA), dried with N2 and put in a base bath for about 3 hours. After being removed from the base bath, the plate was rinsed thoroughly with de-ionized water and blow dried with N2. The cleaned glass plate was dried further at 120°C for about 3 hours to ensure complete drying. After drying the glass plate was immersed in a freshly prepared OTS solution (1mM solution of OTS in toluene) and allowed to stay immersed for 30 minutes. The glass container containing the solution was sealed to minimize contact with atmospheric air and avoid possible degradation of OTS. After 30 minutes, the plate was taken out of the solution and was rinsed successively with toluene, acetone, chloroform and IPA, blow drying with N2 between. The SAM formed on the surface of the glass plate was annealed in a vacuum oven at approximately 150° C overnight. The glass plates used in these experiments weighed about 46 g and were laid over the sample prior to sliding, thus a uniform pre-load of 46 g was applied to each sample. The weighted slide was then pulled along the sample in the direction of adhesion. Slides were cleaned with ethanol after each sample. We tested each coating group (B, M and F) on the hydrophilic glass slide and the hydrophobic OTS-SAM coated slide in both air and water. Water trials were done the same as air trials, except that after the sample was positioned inside a small open chamber, water was poured in so that the shed was completely submerged and covered with at thin layer of water. We then placed the weighed slides on the sample, as was done for air, and performed the adhesion test. Only the maximum force reading during the sliding was used for analysis. Occasionally, we used the maximum force prior to sample failure (ripped or damaged) for analysis. For each sample sliding behavior was also noted to assess whether stiction or adhesion occurred. Total two-dimensional area of the samples was measured after the experiment

89 using a dissecting microscope (x, city) and ImageJ software (National Institutes of

Health, Bethesda, MD, USA).

3.3.5 Statistical analysis

Ten individuals contributed a total of 12 toe sheds, one to each of the 12 experimental groups. By using one sample per individual in each group we controlled for individual variation because each individual contributed equally to all treatments. In some cases multiple shed samples from the same individual were used to make the 12 total samples for testing. Samples were tested in random order by group. ANOVA was used to test for the effect of coating, treatment and substrate and all interactions. Area was used as a covariate. Force values were log transformed to normalize the residuals.

Individual contrasts were performed between groups of interest, specifically contrasting maximum force in air and maximum force in water for each surface coating tested on each substrate. To account for multiple hypothesis testing we performed a sequential bonferroni correction. A Pearson's chi square test was used to test if stiction or friction occurred more often in air or water in each coating treatment and both substrates.

CHAPTER IV

TUNING SURFACE WETTABILITY USING ORDERED ARRAYS OF PARTICLES

4.1 Motivation

As it has been discussed in the previous sections, surface wettability control is achieved by an optimum combination of surface chemistry and surface roughness.

Typically, the wetting property of a structured surface is described using either a Wenzel model or a Cassie-Baxter model. Although both of these models are used commonly to qualitatively describe the wetting properties of a structured surface, a quantitative correlation between the models and experimental data is always challenging mainly due to lack of control over surface roughness. Even though there exist a number of methods for making surfaces rough, the scope of designing a well-controlled surface roughness with dual or more levels of hierarchy for obtaining a desired wetting state is restricted mostly to some theoretical models [176, 177] and a few experimental demonstrations[178]. From fabrication view-point, as the length scale of surface feature size decreases, the precise control over surface roughness not only becomes more difficult but also relies mainly on extremely expensive lithographic technique. [179, 180,

181, 182, 183] Moreover, rectangular pillar geometry is the most commonly studied roughness geometry due to its relative ease of fabrication and modelling.[3, 184, 185]

Roughness features with curvatures such as cylindrical or spherical shaped features are

91 less explored and impose even more challenges for precisely controlling roughness parameters for hierarchical structures. Some of the studies reported on wettability of spherical particles give a quantitative picture of wetting transitions of regular arrays of spheres in terms of critical capillary pressure and show its dependence on the intrinsic wettability of the surface.[186, 187, 188, 189] However, they do not show any correlation between the wetting transitions and apparent contact angles. Also, the reported experiments deal with particles with dimensions comparable to the typical water droplet sizes used for contact angle measurements (few mm in diameter), in which case contact angle measurements with even a single particle is possible. Precisely engineering the patterned roughness using particles with dimensions of the order of a few nano-meters such that the wettability can be quantified and correlated in terms of contact angles is still challenging, firstly due to experimental challenges in fabricating such surfaces and secondly due to the fact that contact angle can have a real value only between 0º and

180º. The surface roughness can easily have values for which the wetting models fail to calculate real values of contact angles.

In the case where multiple wetting states can exist for a structured surface, the thermodynamic stability of every wetting state and its probable transition to one or more possible wetting states is crucial in almost all applications of structured surfaces.

Typically used process parameters such as dynamic fluid flow, hydrostatic pressure, impact and compressive forces, used in the applications of these surfaces could easily overcome the energy barrier between the wetting states, resulting in transition. This transition may be irreversible and may result in the failure of the anti-wetting

92 characteristic of the surface.[36, 98, 185, 190] It is thus important, not only to precisely tailor the roughness features with suitable materials but also to estimate the stability limits of all the associated wetting states.

In an attempt to develop a systematic approach for studying the wetting properties of structured surfaces, we studied a model surface system of hexagonally arranged spherical particles with a well-defined surface geometry. The regular shapes of the roughness features enabled us to model all the possible wetting states and to predict their thermodynamic stability using classical thermodynamic theory of wetting.

We controlled surface roughness using patterns of spherical particles formed by colloidal lithography. It is a simple, less expensive technique as compared to other analogous techniques, such as photolithography or surface etching. It also offers the advantage of easily patterning non-planar curved surfaces, which is particularly important for non-pillar geometries. In this study, hexagonal non-contiguously close packed

(HNCP) arrays of silica and polystyrene spherical particles are formed with a single and double layer patterns. A single layer of particles (the first layer in the case of dual, hierarchical patterns) consists of spheres with diameters ranging between 100 nm to 2700 nm. The two-tiered roughness is formed with the first layer of HNCP array of polystyrene spheres (2700 nm in diameter) followed by the HNCP array of smaller spherical silica particles (250 nm in diameter) on top of it. The particle size and inter-particle distance for both large and small particles was varied to alter the surface roughness. We used plasma enhanced chemical vapour deposition (PECVD) technique for surface chemistry and thus, the intrinsic wettability control of a surface. The choice of PECVD in this case is

93 particularly advantageous because it is a dry process. Therefore, unlike wet processes where the solvents used may result in the disruption of the pattern due to capillary evaporation or disturb the surface charge that holds the arrays together, PECVD allows retention of the surface pattern. Also, PECVD is not restricted to any specific precursor.

Thus, a range of surface functionalities could be incorporated using this technique.

Colloidal lithography combined with PECVD is unique approach for tuning surface wettability.

4.2 Results and discussion

The characterization of surface microstructure and chemistry for predicting and measuring surface wettability was carried out as discussed in the following sections.

4.2.1 Single Layer Roughness

The single layer of HNCP pattern was formed with four different particle sizes

(100 nm, 250 nm, 500 nm and 2700 nm). For each particle size, the ratio of inter-particle distance (measured between the centres of any two adjacent particles) to the particle diameter ( = L1/D1) was varied in the range of 1 to 3.

Figure 4.1. SEM image of single layer HNCP array of 250 nm silica particles (a) and 500 nm silica particle (b)

Examples of typical single layer of HNCP arrays of spherical particles with 250 nm and 500 nm silica particles are shown in Figure 4.1a and 4.1b respectively. The

HNCP patterns with 100 nm silica particles and 2700 nm polystyrene particles were formed in a similar way (SEM pictures are shown in figures 4.2a and 4.2b respectively).

The value of was determined for every sample based on the values of L1 and D1

(corresponding to unit cell shown in figure 4.2c, also shown in figure 3.3 in the previous section). The surface chemistry for all single layer HNCP patterns was maintained constant. A hydrophobic coating formed by PECVD of 1H,1H,2H-perfluoro-1-dodecne

( of 110° ± 2°) was deposited on the surface of patterned particle arrays. The angle was determined by the water contact angle measurement on a flat surface coated with the given PECVD coating, as discussed in the experimental section above.

(a) (b)

Figure 4.2. Examples of single layer roughness; a HNCP arrays formed with (a) 100 nm particles (scale bar : 1μm) and (b) 2700 nm particles (scale bar : 10 μm); (c) A schematic contd. Figure 4.2. representation of a hexagonal unit cell. One unit cell contains one

Figure 4.2. (continued) particle. L is the inter-particle distance measured between the centres of two adjacent particles and D is the particle diameter. A schematic representation of (d) Wenzel-state and (e) Cassie-Baxter state for single layer roughness formed by HNCP arrays of spherical particles.

For a single layer of HNCP pattern, there are two possible wetting states, either the Wenzel-state or the Cassie-Baxter state, as shown schematically in Figures 4.2d and

4.2e respectively.[37] [38] The apparent contact angle for the Wenzel-state ( ) is defined by the following equation [37]

...... (4.1)

R ( ) is the roughness factor and is the intrinsic contact angle. Based on the

Wenzel model equation, we can conclude that the inherent wetting characteristic of the surface is amplified with the roughness.

For the Wenzel-state of the unit cell (Figure 4.2d), the ratio of actual wet surface area to the projected wet area is defined here as R1 and can be calculated as follows

(derivation in the appendix).

...... (4.2)

180

) ° 160 Wenzel Model 140 250 nm 120 500 nm

Contact Angle Angle Contact( 100 2.7 microns 1 2 3 100 nm r1

Figure 4.3. Plot of contact angle vs r1 for different sized silica particles and the calculated

Wenzel model 96

Using equation 4.2, the apparent contact angles calculated as a function of r1 are shown in figure 4.3; the model predictions are compared to the experimentally measured values for patterns formed with different particle sizes. The equation 4.2 for assumes a perfectly spherical shape of the particle in the unit cell with a single point contact at the bottom of the surface. However, the contact between the sphere and the bottom surface is not perfect since the polystyrene spherical particles tend to flatten, resulting in slight deformation. With a layer of PECVD coating deposited on the surface of particles, the shape of the particle is a spherical cap rather than a perfect sphere. The spherical cap shaped coated particle is characterized by height H and angle (as shown in Figure 4.4).

This renders the following relationship:

………………………………………………………………(4.3)

The change in geometry thus, introduces a correction factor in equation 4.2 that needs to be accounted for to predict the Wenzel angle accurately. The value of R1 of the PECVD film coated unit cell is given as follows:

...... (4.4)

Figure 4.4 : The angle ‘α’ and height ‘H’ defined for particles coated using PECVD

The roughness factor is now dependent on and r1 as opposed to just r1 before.

SEM was used to measure the diameters of the features (D) and AFM was used to measure the height (H), as discussed above. and are thus calculated from these values. (Table 4.1)

Table 4.1. Diameter, height and calculated α for the spheres coated using PECVD

Particle Diameter (nm) Height (nm) Height/Diameter α

100 nm 222 ± 11 131 ± 5 0.59 79.6

250 nm 319 ± 21 241 ±14 0.76 58.7

500 nm 495 ± 28 414 ± 23 0.83 48.7

2700 nm 2850 ± 73 2500 ± 65 0.87 42.3

It is important to note that the layer of coating deposited using PECVD on the surface of the particles is not perfectly smooth as well. The roughness of the coating surface, even though very small compared to the particle dimensions, can add to the overall roughness. We used AFM measurements to determine the roughness of the layer of PECVD coating. (Figures 4.5a and 4.5b) We used the triangulation analysis of the

AFM height images to calculate the ratio of actual surface area to the projected area of the coating layer (R2). The data was fit to the exponential function (Figure 4.5c) and the extrapolation of R2 value to zero scan size gives the roughness factor, R2, for PECVD coatings.[191, 192] The effective roughness factor (R) for PECVD coated HNCP layer can be estimated as follows:

...... (4.5)

Figure 4.5. (a) AFM 2D height image of 1H,1H,2H-perfluoro-1-dodecene deposited by

PECVD on the surface of clean Silicon wafer, scan size : 1μm, (b) AFM 3D height image of 1H,1H,2H-perfluoro-1-dodecene deposited by PECVD on the surface of clean Silicon contd. Figure 4.5. wafer, scan size : 1μm and (c) R2 calculated using AFM image analysis for a PECVD coating of 1H,1H,2H-perfluoro-1-dodecene as a function of scan size. The extrapolation to zero scan size gives the value of R2.

The value of is estimated to be about 1.19 ± 0.05 for a PECVD coating of

1H,1H,2H-perfluoro-1-dodecene. Substituting the value of R in equation 4.1 gives the value of expected from the Wenzel model.

The second possible wetting state is the Cassie-Baxter state as shown schematically in Figure 4.2e. In this case, it is only a fraction of particle surface that is in

99 contact with water; f1 defines solid-liquid interfacial area fraction and f2 defines liquid- vapour interfacial area fraction. The calculations are based on two main assumptions: first, the air-water interface at the three-phase contact line is flat, which is reasonable for the typical water droplet sizes used in the experiments. The second assumption is that the value of f1 depends on the depth of penetration of water, which is characterized by an angle . It is not possible to directly measure . Because the contact angles depend on , the optimum value of was required to estimate an accurate correlation between model calculations and experimental data. The contact angle for Cassie-Baxter state of unit cell is given by the following equation. (Please see derivation in the appendix)

...... (4.6)

For a single layer of spherical cap shaped particles, the fractions can be estimated as follows:

and ...... (4.7)

As seen in equation 4.7, f1 and f2 are both functions of and . is determined experimentally, as explained earlier. is varied between 0° and 90°. The depth of penetration of water inside the gap is assumed to have an upper limit of 90°.

A thermodynamic barrier separates the Wenzel and Cassie-Baxter states. A more thermodynamically stable state is preferred. The possibility to overcome this barrier and transition from one state to the other is a function of surface roughness and its inherent contact angle. To be able to predict if the experimentally measured contact angle on the surface of HNCP patterned arrays correspond to the most stable wetting state or represent

100 the meta-stable wetting state, a thermodynamic analysis of all possible wetting states is required. The free energy difference (in its dimensionless form) as the water droplet is deposited on the surface is written as follows:[172, 96, 97]

...... (4.8)

Where, is the volume of the water droplet, and is the surface tension of water, θ is the apparent contact angle, f1 and f2 are solid-liquid and liquid-vapour interface area to projected area ratios as defined before. is the intrinsic contact angle. In the case of

Wenzel state, f2=0 and f1=R. In the case of Cassie-Baxter state, f1 and f2 are calculated based on AS, AL and Af which correspond to solid-liquid interface area, liquid-vapour interface area and projected area respectively (please see derivations in appendix for details).

Figure 4.6a shows contact angles (θ) measured experimentally as a function of .

The figure also shows a comparison between experimental values and θ calculated using

Wenzel and Cassie-Baxter models for the surface with HNCP array of 250 nm particles.

The corresponding free energy profiles look qualitatively identical and show a similar trend (Figure 4.7). The dotted line in the graph indicates relatively unstable wetting state

(figure 4.6a), whereas a solid line corresponds to a relatively stable wetting state at a given . Based on our model calculations, we can expect that below = 1.5, Cassie-

Baxter (CB) is more favored wetting state and above =1.5 Wenzel (W) is more favourable state. We found that the experimentally measured contact angles are consistent with the model predictions. Even though the values calculated using the model are not exactly the same as the ones obtained experimentally, they match within a range

101 of experimental errors. Qualitatively, both the model and experimental values show an identical trend for >1.5. The points very close to or lower than =1.5 are observed to follow CB line. However, with limited number of variations possible in the region

, a clear trend could not be obtained for experimental data points in the region r1 < 1.5. The errors in the experimental values arise mainly from surface defects and variations in particle size and size distribution, which are also evident from the SEM images of the patterns shown in figures 4.1 and 4.2.

A similar trend is also observed for samples made with different particle sizes.

Thus, it can be concluded based on the model calculations that for a single layer roughness formed with HNCP pattern with a hydrophobic coating on the surface

(intrinsic contact angle of about 110°), Wenzel state is the most stable wetting state for the range of measured in this study. A very narrow window of ( shows a Cassie-Baxter wetting state to be the most stable state. The experimental data are also seen to be in good agreement with this trend at a given point in the range of variation used.

Figure 4.6b shows a master curve in which the experimentally measured apparent contact angles for different particle sizes (100 nm, 250 nm, 500 nm and 2700nm) are plotted as a function of R. These experimental results are compared to the predictions of the Wenzel model calculated using equation 1 and the values of R calculated from the experimental measurements. A reasonably good fit is observed between the predictions of the Wenzel model, without an adjustable parameter and the experimental measurements of the contact angles for values of R between 1 and 3.

102

Figure 4.6. (a) A comparison of experimental contact angles with Wenzel (W) and

Cassie-Baxter (CB) model curves for a single layer of HNCP array of 250 nm particles, plotted as a function of r1. The dotted line represents the meta-stable state, whereas the contd. Figure 4.6. solid line corresponds to the stable wetting state. The transition point is at r1~1.5. (b) A plot of contact angles as a function of effective roughness (R) of a single layer of HNCP patterns for all particle sizes used. The experimental data points for all particle sizes follow the Wenzel model behaviour. The dotted lines represent the upper and lower limits of the Wenzel model region. The behaviour is consistent with the plot in figure (a).

103

Figure 4.7. Plot of free energy G* values vs r1 for Wenzel and Cassie-Baxter states for single layer HNCP array of 250 nm particles

4.2.2 XPS analysis for surface chemical composition of the PECVD coated HNCP

patterns F1s

80x103

cps FKLL 40

C1s F2s 0 0 200 400 600 800 1000 B.E. (eV)

Figure 4.8. A survey spectrum of PECVD coating of 1H,1H,2H-perfluoro-1-docence on the surface of HNCP pattern.

104

The survey scans were used to estimate the atomic compositions of the PECVD coating on flat Si wafer and patterned surfaces. On comparing the results, we see that both the spectra show identical elemental peaks. The presence of identical peaks in survey spectra of the two also implies that the coating thickness is more than the typical analysis depth of 10 nm of XPS scans, which is consistent with the PECVD coated particle thickness measurements we carried out using AFM. (Table 4.1)

(a) 1000 perF-Si

800

600 cps

400

200

280 285 290 295 300 B.E. (eV) (b)

700 perF- HNCP pattern 600

500

400 cps 300

200

100

280 285 290 295 300 B.E. (eV) Figure 4.9. C1s spectra of PECVD coating of 1H,1H,2H-perfluoro-1-dodecene on the surface of cleaned Silicon wafer (a) and HNCP pattern (b).

The spectra (a) and (b) in figure 4.9 look very similar implying that the surface functionality on flat Si wafer and the patterned HNCP array are the same. The peak 105 assignments confirm that all the peak positions correspond to different C atoms in the precursor molecule structure. The results are tabulated below (table 4.2).

Table 4.2. Peak position and assignments for high resolution XPS C1s spectra of PECVD layer of 1H,1H,2H-perfluoro-1-dodecene on the surface of flat Si wafer and HNCP array of spherical particles

Peak Positions and Assignments

Peak ID Binding Energy (eV)

Flat Si Structured HNCP Peak Assignments

pattern

1 286.21 286.47 C-CFn

2 288.47 288.66 CF-CFn

3 291 290.98 CF2

4 292.94 293 CF3

The similar characteristics are seen in the case of PECVD HMDSO coated HNCP arrays when compared to the HMDSO coating on the surface of flat Si. The survey scan of HMDSO coated HNCP patterns are shown in figure 4.10, whereas the high resolution

C1s scans of HMDSO coating on a flat control and patterned surface are shown in figures

4.11a and 4.11b respectively. The peak assignments for the same are listed in table 4.3 below.

106

20x103

O1s C1s 15

cps

OKLL Si2p

5 Si2s O2s 0 0 200 400 600 800 1000 B.E. (eV) Figure 4.10. A survey spectrum of PECVD coating of HMDSO on the surface of HNCP pattern.

107

(a)

1400 HMDSO-Si

1200

1000

800 cps

600

400

200

280 285 290 295 300 B.E. (eV) (b)

1000 HMDSO-HNCP pattern

800

cps 600

400

200

280 285 290 295 300 B.E. (eV) Figure 4.11. C1s spectra of PECVD coating of HMDSO on the surface of cleaned Silicon wafer (a) and HNCP pattern (b).

Table 4.3. Peak position and assignments for high resolution XPS C1s spectra of PECVD layer of HMDSO on the surface of flat Si wafer and HNCP array of spherical particles

Peak Positions Peak Assignments

Flat Si HNCP pattern

1 283.36 283.34 Si-CH3 or Si-CH2-CH2-

2 284.3 284.27 Si-CH3 or Si-CH2-CH2-

3 286.5 286.14 -C-O-

108

4.2.3 Dual Roughness

For the hierarchically rough structures, the surface roughness is controlled at two different length scales and is characterized by parameters (bigger particles) and

(smaller particles), defined as and , respectively. L and D here are inter- particle distance and particle diameter respectively, as defined before for single roughness. For hierarchical HNCP arrays of particles, the value of is maintained constant at either 1.5 ± 0.17 or 2 ± 0.2. Dual roughness is formed with two levels of

HNCP patterns. For these patterns, we used D1 = 2700 nm and D2 = 250 nm. The variation in and is achieved by varying inter-particle distances L1 and L2, respectively. Similar to single roughness, particle concentration and transfer rate were chosen depending upon the value of r2 chosen for a given sample.

The examples of hierarchical HNCP arrays are shown in SEM images in figure

4.12. The images in Figures 4.12a-4.12d correspond to a series of samples where is decreased from 3.5 to 2 at an interval of 0.5. In all these cases, was kept constant at

1.5. The decreasing inter-particle distance for the first layer of HNCP arrays of bigger particles can be observed clearly in figures. Figure 4.12e corresponds to the sample with

= 2 and = 2. The variation in for the second layer of HNCP arrays of particles is not easily distinguishable at the magnification used for the images shown in Figure 4.12.

The precisely controlled values of r1 and r2 at both levels provide wide enough range of surface roughness to test the existence of different wetting states as a function of surface roughness.

109

Figure 4.12. Examples of surfaces with dual hierarchical roughness used for experimental studies. SEM images represent samples with (a) r1 = 3.5, r2 = 1.5, (b) r1 = 3.0, r2 = 1.5, (c) r1 = 2.5, r2 = 1.5, (d) r1 = 2.0, r2 = 1.5, (e) r1 = 2.0, r2 = 2.0. Scale Bar: 1 µm for all the images. Decreasing r1 from 3.5 to 2 can clearly be seen in the images a to e.

The surface chemistry control in this case was achieved by using two different surface functionalities. PECVD coating of 1H, 1H, 2H-perfluoro-1-dodecne with intrinsic contact angle ( ) of 110°± 2º and hexamethyldisiloxane (HMDSO) with of about

110

95°± 3º was deposited on the surface and the wettability was determined with contact angle measurements, as explained in the experimental section. The coating roughness in this case is referred to as R3. R3 was determined in a similar way using AFM measurements, as discussed above for single scale roughness. Figure 4.13 below represents a curve fitted to the data points generated using AFM measurements.

Figure 4.13. R’ calculated using AFM image analysis for a PECVD coating of HMDSO as a function of scan size.

A hexagonal model unit cell approach similar to that of a single layer roughness was used to theoretically predict the surface wettability and the thermodynamic stability of different wetting states. In addition to the Wenzel and the Cassie-Baxter states, another wetting state is also possible due to two levels of surface roughness.[176] The intermediate wetting state is called the Penetrating-Cassie-Baxter state. The schematic representation of all three wetting states is shown in figure 4.14. It can be seen that the

Wenzel-state corresponds to the complete penetration of water between the small as well

111 as the large particles as defined by the Wenzel model (figure 4.14a). The Cassie-Baxter state for this type of surface is defined by partial penetration of water between small and large particles. The angles (for large particles) and (for small particle layer on top of a big particle) correspond to the depth of penetration of water layer forming a solid-water-air composite interface at both levels of structural hierarchy (figure 4.14b).

The Penetrating-Cassie-Baxter state is defined as the wetting state in which the water penetrates completely between the large particles. However, water is only partially penetrated between the small particles. The depth of penetration is defined using a parameter (figure 4.14c). As explained in the case of single layered HNCP patterns, the bigger particles are spherical cap shaped instead of perfect spheres. The spherical cap shape of the particle is accounted for by defining the parameter (figures 4.14a and

4.14c).

The apparent contact angle in the Wenzel state ( ) is calculated using equation

4.1. The effective surface roughness of unit cell, R, in this case is defined as follows:

...... (4.9)

R1 and R2 here correspond to the roughness ratios of the first layer of bigger particles and second layer of smaller particles, respectively. R3 accounts for the surface roughness of the PECVD coating layer, as explained before for a single layer. It is estimated to be 1.19 and 1.21 for PECVD layer of 1H,1H,2H-perfluoro-1-dodecene and HMDSO respectively. In a more general form, for a system with hierarchical pattern of particle structures with ‘n’ layers of roughness, the total surface roughness for Wenzel wetting state can be defined as follows:

112

...... (4.10)

Figure 4.14. A schematic representation of (a) Wenzel state, (b) Cassie-Baxter state and

(c) Penetrating Cassie-Baxter state for hierarchical dual roughness. Wenzel state corresponds to the penetration of water all the way inside the surface roughness features at both levels of roughness hierarchy. The Cassie-Baxter state corresponds to α1CB which defines the penetration of water between the bigger particles. The depth of penetration of water between the smaller particles in Cassie-Baxter and Penetrating-Cassie-Baxter states is characterized by α2CB and αPC, respectively.

113

The apparent contact angle for the Cassie-Baxter state for hierarchical surface roughness is calculated using equation 4.6, similar to the single layer roughness. The fractions f1 and f2 in this case, are estimated using following equations:

...... (4.11)

……………………………...... (4.12)

It can be seen from equations 4.11 and 4.12 that both f1 and f2 depend on , , r1 and r2.

Figure 4.15. (a) Contact angles predicted for dual roughness unit cell model using

Penetrating-Cassie-Baxter wetting state equations and compared to corresponding Cassie-

Baxter state, θY = 110º, r2= 0.9 and αPC=70º, (b) Contact angles predicted for dual roughness unit cell model using Penetrating-Cassie-Baxter wetting state equations and compared to corresponding Wenzel and Cassie-Baxter states, θY = 95º, r2= 1 and αPC=80°

For the Penetrating-Cassie-Baxter state, the apparent contact angle can be estimated using equation 4.6 as well, since it shows a partial Cassie-Baxter wetting character. Because water has penetrated between the large particles, the equations for f1

114 and f2 are significantly different for the Penetrating-Cassie-Baxter state in comparison to the Cassie-Baxter state. (Please see the appendix for the derivations.)

and ...... (4.13)

...... (4.14)

It can be seen that both f1 and f2 for Penetrating-Cassie-Baxter state depend on r1, r2 and αPC. For the range of r2 we used in our experiments for varying the surface roughness, no real values of contact angle (θPCB) could be predicted and compared with the experimental data. The real contact angles value prediction is possible however for lower values of r2. The corresponding data for Penetrating-Cassie-Baxter state is shown and compared with Wenzel and Cassie-Baxter states in figure 4.15. Controlling r2 precisely in a lower range is challenging experimentally.

Figure 4.16a compares the theoretical and experimental values of contact angle as a function of for the dual roughness surfaces coated with a PECVD layer of 1H, 1H,

2H-perfluoro-1-dodecne. r2 in this case is maintained constant at 1.5. It can be seen that the range of contact angles measured experimentally in this case is higher compared to the single roughness for the same surface chemistry. The figure also shows a range of contact angles predicted using Cassie-Baxter model. For the range of and that we used no real values for and can be calculated using HNCP model. Thus, the model curves for those two cases are not shown in Figure 4.16a for comparison. The calculations for Cassie-Baxter model assume values for and, since actual estimation of these variables is not possible experimentally for single particle unit cell. In

order to generate a model curve region, and was used.

115

Figure 4.16. The comparison of model curves and experimental data for contact angles as a function of r1 for hierarchical dual roughness; (a) Contact angles for HNCP dual roughness arrays with PECVD layer of 1H,1H,2H-perfluro-1dodecne deposited on the surface plotted as a function of r1, r2 = 1.5 is maintained constant. The Cassie-Baxter wetting state which is the only model that gives real contact angle values is denoted as

CB. The model lines are generated using α1CB=30º and α2CB=70º. (b) Contact angle data for HNCP dual roughness patterns with a PECVD layer of HMDSO deposited on their surface are shown in with r2 = 1.5 and r2= 2. The possible wetting states are denoted as W

116

Figure 4.16. (continued) (Wenzel) and CB (Cassie-Baxter).Contact angles measured by droplet deposition on the surface as well as by condensing water on the surface in both the cases are compared. Free energy calculations show that Wenzel is the most stable wetting state. The experimental data are consistent with the model predictions irrespective of the way in which the droplet is formed on the surface.

Although our initial assumption was to treat both and as fitting parameters, it is only this combination of the selected penetration depth angles that can give real contact angle values for the entire range of r1 we used. Therefore, uncertainty about the exact values of and and limited range of variation possible, the model does not exactly predict the experimental trend we observed. Even though the contact angle values do not match exactly, the experimental measurements show contact angles more than 150º and very low contact angle hysteresis (~ 5º), which are typical characteristics of the Cassie-Baxter state. Also, as can be seen from the experimental data, the measured contact angle values increase as r1 increases; which is a trend consistent with the Cassie-Baxter model prediction. Thus, we can conclude that both the experimental data points and model curves lie in the Cassie-Baxter region. and are the only unknown variables and since there are only limited number of combinations possible for the two, we could not define the upper and lower limits of the Cassie-Baxter region in this case.

Figure 4.16b shows the wettability data with contact angles plotted as a function of (r1 varied between 1 and 4) for the HMDSO coated dual roughness surface samples with = 1.5 ± 0.17 and = 2 ± 0.2, respectively. The purpose of choosing HMDSO for

PECVD coating is to change the inherent wettability ( of 95°) for the same roughness

117 geometry. Figure 4.16b compares the Wenzel and the Cassie-Baxter model predictions with the experimentally measured contact angles. The corresponding free energy curves are shown in the supporting information. Experimentally measured contact angles follow

Wenzel line for both r2 = 1.5 and r2 = 2. Because the dual roughness surfaces are not defect free, similar to the single roughness surfaces, the errors associated with r2 are accounted for when generating the model lines. Even though the experimental contact angle data shown in figure 4.16b clearly demonstrates the Wenzel wetting state for both r2 , there is a possibility of existence of Cassie-Baxter state for the parameters we used in the experiments.

Figure 4.17. (a) Comparison of G* values vs r1 for Wenzel(W) and Cassie-Baxter (CB) states for HMDSO coated dual roughness surface, CB-1.5 and W-1.5correspond to r2=

1.5; CB-2 and W-2 correspond to r2=2, (b) Comparison of G* profiles as a function of r1

118

Figure 4.17. (continued) for Cassie-Baxter (CB) and Penetrating-Cassie-Baxter (PCB) states for 1H.1H,2H-perfluoro-1-dodecene PECVD coated dual roughness surface, θY

=110º, r2=0.9, αPC=70º, α1CB=30º and α2CB=70º, (c) Comparison of G* profiles as a function of r1 for Wenzel (W), Cassie-Baxter (CB) and Penetrating-Cassie-Baxter (PCB) states for HMDSO PECVD coated dual roughness surface, θY =95º,α1=42.3, r2=1,

αPC=80º, α1CB=45º and α2CB=85º.

From our free energy analysis we also know that Wenzel is the most favorable wetting state, in the case where both Cassie-Baxter and Wenzel wetting states are possible (figures 4.17a). In the case where all three states are possible, the model predictions show that penetrating Cassie-Baxter would in fact be the most stable wetting state (figures 4.17b and 4.17c). We further confirmed this by measuring a static contact angle of a condensed droplet on these surfaces. Condensation is the easiest way to induce transition to the most stable Wenzel state, in case metastable Cassie-Baxter wetting state existed on the surfaces. The contact angles measured with condensed water droplet also fall on the Wenzel lines (figure 4.16b). Thus, irrespective of the mechanism of contact of a water droplet with the surface (deposition versus condensation), the most stable Wenzel state is acquired by the water droplet for given parameters of the dual roughness surface.

In the case of dual roughness, a wide range of variation was possible by controlling known variables r1, r2 and θY. However, because predicted contact angles can have values only between 0º and 180º, choosing a right combination of the variables which give the real values of contact angles to compare with the experimental data was the most crucial step in this case. and although are unknown variables, did not seem to have a significant effect on model predictions. With all the limitations in model

119 calculations identified, we were not only be able to design the surfaces with suitable combination of variables but also demonstrated a reasonable correlation between experimental data and model predictions for two different wetting states (Cassie-Baxter in the case of 1H,1H,2H-perfluoro-1dodecene and Wenzel in the case HMDSO PECVD coating.) We did not show existence of Penetrating-Cassie-Baxter state experimentally.

However, we identified a range of r2 in which Penetrating-Cassie-Baxter state is possible.

Thorough thermodynamic analysis of the three possible wetting states in this case serves as a design approach for fine tuning the surface wettability .

4.3 Conclusion

In this study, we demonstrate that the combination of colloidal lithography and

PECVD allows for precise control over surface roughness as well as surface chemistry.

The study of hexagonal unit cell of HNCP patterns to model the colloidal assembly of particles enabled us to test the validity of thermodynamic wetting models and also made it possible to predict the stability of a given wetting state. The experiments designed allowed qualitative as well as quantitative comparison of the Wenzel and the Cassie-

Baxter models with the experimental data.

Based on our experimental findings and model calculations, it can be concluded that , and are the three variables for single layer and dual layer of HNCP patterns of spherical particles. The regular hexagonal arrangement of patterns and precise control over the variables enabled the predictions of different wetting states possible with single and dual roughness geometry. We also demonstrated a reasonably good correlation between the model predictions and experimentally measured contact angles for Wenzel and Cassie-Baxter states. Additionally, the unit cell model allowed us to predict an

120 existence of Penetrating-Cassie-Baxter wetting state for dual roughness and identify the range of r2 corresponding to the same. A very unique geometry of curved spherical particles, their well-controlled arrangement at both levels of roughness hierarchy and flexibility to precisely alter the surface chemistry, surely makes a combination of colloidal lithography and PECVD an efficient tool to fine tune the surface wettability.

Additionally, it also allows for modelling of the surface to predict not only the apparent contact angles but also the metastability of the wetting states. This approach allows to fabricate HNCP patterns, which resemble lotus leaf surface and moth-eye structure; with tunable wettability and can be used for biomimetic applications such as designing superhydrophobic, self-cleaning surfaces and anti-reflective coatings.

121

CHAPTER V

CARBON NANOTUBES BASED ROBUST SUPERHYDROPHOBIC SURFACES

5.1 Motivation

Surfaces that remain dry under a flux of steam are desired for many applications, such as blades of steam turbines and surfaces of heat exchangers. Different techniques are employed to create water repellent surfaces, commonly referred to as superhydrophobic surfaces. They are defined as surfaces on which water forms a contact angle of greater than 150° and are formed by suitable combination of surface chemistry and roughness.[44, 45, 193] However, the ability of a surface to remain dry by letting the water droplets roll-off depends on contact angle hysteresis (CAH) defined as the difference between advancing and receding water contact angle on a surface, as discussed in the previous sections. The major factor influencing the CAH on a superhydrophobic surface is the ‘superhydrophobic state’ of the water droplets.[32, 98] A droplet of water on such a surface may be in a wetting state (Wenzel state) with water completely filling the roughness asperities on the surface[37] or a non-wetting state (Cassie-Baxter state) with the water droplet supported on the air-solid composite surface.[38] In the Cassie-

Baxter state, water droplets show lower CAH and easily roll off the surface as compared to the ones in the Wenzel state.

Compared to water repellency of a superhydrophobic surface where water droplet is deposited on the surface, in the case of condensation of water vapor on such a surface,

122 the formation of water droplet follows a nucleation and growth mechanism where nucleation may initiate within the pores. This would result in formation of water droplets in the Wenzel state.[194] Therefore, a superhydrophobic surface which shows low contact angle hysteresis with liquid water may not show similar behavior in case of steam condensation. For the surface to remain dry under condensation, it is required that the water droplets undergo a transition from Wenzel to Cassie-Baxter regime. Depending on if this transition is favorable thermodynamically, a typical superhydrophobic surface may or may not act as anti-condensation surface. Few approaches proposed and tested in the past to create surfaces that can undergo such a transition include the design of surfaces with differential intrinsic wettability either in the form of a continuous gradient[104] or patches of hydrophilic-hydrophobic regions on the surface.[195] A similar vertical surface energy gradient was proposed to explain the ability of a lotus leaf to remain dry under dew formation.[196] A systematic design of roughness features on the surface such that nano-scale roughness is present either as a part of the structural hierarchy or the surface being nano-porous by itself have also been shown as design approaches to make a surface remain dry under condensation. [17, 70, 77] However, the exact mechanism behind the wetting behavior of a structured surface under condensation is still not understood completely. All the condensation studies reported pertain to low temperature and pressure of condensing water vapor, carried out typically in the range of 0°-5°C. In the case of steam condensation, the mechanism of droplet formation is similar to low temperature condensation. However, the environment of high temperature and pressure of condensing steam to which the surface is exposed may exceed the mechanical and hydrostatic stability limits of the non-wetting state of the surface. Thus, the stability

123 limitations make it more difficult to design a steamphobic surface even though control over surface energy and geometry is easily achieved with a wide range of materials.

The current work creates unique coatings for stainless steel that remain dry under prolonged exposure to steam and maintain their superhydrophobicity. The coatings are formed using mesh-like carbon nanotube (CNT) structure. The structure provides desired surface roughness and porosity. The chemistry of CNT structures is modified using

Plasma-Enhanced Chemical Vapor Deposition (PECVD), more commonly known as plasma polymerization process.

5.2 Results and discussion

The characterization of CNT surfaces and its correlation to the wetting property is discussed below.

5.2.1 Morphology and chemical composition of CNT and PCNT

Figure 5.1. SEM images of CNTs(a) and plasma coated CNTs (PCNT)(b)

The SEM images in figure 5.1 show the structure of as grown CNTs on stainless steel (SS) plates and PECVD coated CNTs (PCNT). The morphology of CNT mats can be described as an entangled, nano-porous structure. This morphology is retained even

124 after the deposition of plasma coating. The PCNT structures are not easily distinguishable from CNT structures, as can be seen by comparing the SEM images.

(a) (b)

100 nm 100 nm

Figure 5.2. TEM images of (a) uncoated CNTs and (b) plasma coated CNTs (PCNT), (c)

Diameter distribution of uncoated (CNT) and plasma coated CNTs (PCNT).

Thickness of the plasma coated films on CNT is estimated by analysis of distribution of diameters of CNTs. TEM is used to image the CNT structure (as shown in figures 5.2a and 5.2b, corresponding to CNT and PCNT respectively) and Image-J

125 software is used to analyze the tube diameters. Figure 5.2c compares the overall size distribution of CNTs and PCNTs. It can be seen that the minimum diameter range for

CNTs is 5-10 nm, whereas for PCNTs it is 10-15 nm. The peak in the diameter distribution of CNTs lies in the 15-20 nm range. In the case of PCNTs, it is seen at 20-25 nm diameter range. From these observations, the thickness of the plasma deposited coating can be approximated to be about 5-10 nm, which is not easily detectable with the resolution limits of the imaging instruments used and given the non-uniformity of the tube diameters. Thus, the size distribution analysis gives a good estimate of the coating thickness and the thin layer of the plasma coating serves the purpose of surface chemistry modification without significantly affecting the nano-structure of the CNTs. A thin layer of organic coating is also desirable to prevent any significant effects on heat transfer coefficients.

(a) (b)

Figure 5.3. TGA analysis of (a) CNT and (b) PCNT

Comparison of TGA profiles of CNT and PCNT shows that the weight loss of about 3% at around 455° C can be attributed to the plasma deposited coating (as seen in figure 5.3b). No such weight loss is observed in the case of CNT sample in the same

126 temperature range implying the absence of coating. The weight loss following the plateau at 572°C in figure 5.3b corresponds to Carbon nano-tubes. The similar profile of weight loss can be seen in figure 5.3a starting at 535°C.

1214

1150

658 725 888 1017 Monomer 1151 1209

1077 Plasma coating

Absorbance Absorbance [a.u.] 737 1518

400 800 1200 1600 2000 -1 Wavenumber (cm )

Figure 5.4. IR spectra of monomer and the coating formed by PECVD.

To further confirm the presence of plasma coated films on nano-structures, IR spectra of 1H,1H,2H-perfluoro-1-dodecene monomer and the coating formed by plasma enhanced deposition are compared (figure 5.4a). The IR spectrum of plasma deposited coating shows that there is a loss of characteristic alkene C-H bending vibrations in the region 1000-650 cm-1. This implies an activation of double bond of the monomer in the plasma state. The characteristic perfluoroalkyl chain stretching bands are observed in the

-1 -1 region 1110-1280 cm (CF2) and 1110-1340 cm (CF3) in both, monomer and plasma deposited coating. The broadening of the peaks in this region for the plasma deposited coating can be attributed to lower structural retention of the monomer arising due to

127 fragmentation, rearrangement and cross-linking reactions that take place in the plasma enhanced deposition process.[153, 151, 197]

(a) (b) 3 50x10

3 F1s

C1s 100x10

40 80

60 cps cps 20 40 FKLL

10 20

C1s F2s 0 0 0 200 400 600 800 1000 0 200 400 600 800 1000 B.E. (eV) B.E. (eV)

3 (c) 60x10 50

30 CPS 20

680 685 690 695 700 B.E.(eV)

Figure 5.5. XPS survey spectra of (a) CNT (uncoated carbon nano-tubes) and (b) PCNT (

Plasma coated Carbon nanotubes), (c) XPS F1s spectrum

The quantitative chemical analysis of the plasma deposited coating is carried out with XPS. Figures 5.5a and 5.5b show survey spectra for CNT and PCNT respectively.

As can be seen in figure 5.5a (CNT), the presence of only C is detected. In figure 5.5b

(PCNT), C and F are the elements detected on the surface.

Figure 5.6 shows C1s spectra of uncoated CNT mats and 1H,1H,2H-perfluoro-1- dodecene plasma coated CNT mats. The C1s spectra of the same plasma coating on the aluminum and silicon surfaces which are used as reference flat substrates are also shown in the figure. The spectra are fitted to a Gauss-Lorentz function. The main peak in the

128 case of uncoated CNTs, which shows up at around 284 eV, can be assigned to graphitic

C-C bond in a carbon nanotube. The spectra for coated aluminum and silicon surfaces look almost identical and do not show any peak at 284 eV. All the four peaks observed for the coating on aluminum and silicon can be assigned to the characteristic binding energy range of C-Fn bonds.[151, 152, 153]

C-C (CNT)

CNT

CF2

C-CFn CF-CFn CF3

Coated Al Intensity [a.u.] Intensity Coated Si

PCNT

280 285 290 295 300 B.E. (eV)

Figure 5.6. XPS C1s spectra of uncoated (CNT), plasma coated (PCNT) CNTs and plasma coated Al and Si.

The percent retention of CF2 (291 eV) and CF3 (293 eV) at the surfaces can be estimated by integrating the area under the curve. This is particularly significant because the presence of these functional groups on the surface is responsible for making the CNT

129 surface hydrophobic i.e. modifying its intrinsic wettability.[198] In the case of PCNTs, an additional peak at around 284 eV is assigned to C-C bond of CNT. The presence of this peak along with other C-Fn peaks implies that the thickness of the coating on CNT should be less than typical analysis depth of about 10 nm for the XPS. Therefore, the thickness of the coating estimated is consistent with the results obtained with the size distribution analysis of the tubes. The comparable values obtained from the functionality retention and F:C atomic ratios of plasma coated aluminum, silicon and CNTs are consistent with the fact that the PECVD was carried out under identical conditions for the deposition of the coating on all the surfaces.

The F1s spectra obtained for all the samples show a single peak at about 688 eV, which can be assigned to the C-F bond (figure 5.5c). Similar F1s XPS spectra are observed for coatings on Al, Si and CNT surfaces. XPS analysis thus confirms the presence of a thin layer of plasma deposited coating on the carbon nanotube surface. The atomic compositions calculated using survey scans and peak assignments are tabulated below.

Table 5.1.Chemical quantification of the plasma deposited coating using XPS analysis:

C1s peak assignments, functional group retention and atomic concentrations

Peak Positions and Assignements: Binding Energy (eV)

Peak Peak ID Assignments

Uncoated Plasma Plasma Plasma CNTs coated Al coated coated Si CNTs

130

Table 5.1. (continued) Chemical quantification of the plasma deposited coating using XPS analysis: C1s peak assignments, functional group retention and atomic concentrations

1 284.14 284.09 C-C 2 284.35 286.32 286.22 286.13 C-CFn 3 283.61 288.48 288.47 CF-CFn 4 291.12 290.99 290.81 CF2 5 293.06 292.93 292.76 CF3 Functional Groups Retention: % CF2 46.26 41.67 36.64 % CF3 14.54 16.04 18.73 Atomic Concentration: F:C 1.79 1.95 1.69

5.2.2 Wettability of CNT versus PCNT

The wettability was measured in terms of contact angle, anti-steam condensation and anti-freezing behavior of surfaces, as discussed below.

5.2.2.1 Steamphobicity

CNT mats show a water contact angle of 151°± 3° and CAH of 5°. Water contact angle on PCNT surface is 156° ± 2° and CAH is 2°. Thus, there is no significant difference in the wetting behavior of CNT and PCNT in the water contact angle test and both the surfaces can be characterized as superhydrophobic surfaces. Figures 5.7b and

5.7d show optical images of water droplets on CNT and PCNT surfaces.

131

Figure 5.7. (a) Condensed steam droplets on PCNT surface after steam exposure of 1 hour; Contact angle measured on (b) CNT (uncoated) and (d) PCNT (plasma coated) surfaces; following the steam condensation water contact angle measured on (c) CNT and

(e) PCNT surfaces; (f) Water contact angle measured on PCNTs as a function of steam exposure time.

Under steam condensation, however, it is observed that the two surfaces have a drastic difference in their wetting properties. The experiments consist of exposing the

CNT and PCNT surfaces (held horizontally) to the condensing steam flux and following the steam exposure measuring the water contact angle by depositing a water droplet on the surface. On exposure to steam, it is seen that the condensate forms as small spherical droplets on PCNT surface, as shown in Figure 5.7a. The non-uniform coverage of these droplets on the surface may imply removal of bigger droplets by coalescence. [77, 106]

The droplet size continues to shrink due to evaporation and eventually no droplets are seen on the surface. When PCNT surface is measured for its water contact angle, it is seen that the surface is superhydrophobic even after it is exposed to steam (Figure 5.7e).

132

However, in the case of uncoated CNT mat, within the time scale of minutes, a thin film of condensed water appears to be formed with steam exposure. Even though the film is not clearly visible, the complete wetting of the deposited droplet following steam exposure on this surface implies the presence of thin water film. Thus, the uncoated CNT surface loses its superhydrophobicity on exposure to steam (Figure 5.7c). We also observe that the superhydrophobicity of the CNT surface is restored after it is dried off completely. Figure 5.7f shows the contact angle measured on PCNT mats after prolonged exposure to steam as a function of exposure time. Interestingly, it is observed that the superhydrophobicity of the PCNT surface is retained even after steam exposure of as long as 10 hours, whereas the CNT surface loses its superhydrophobicity when exposed to steam only for a few minutes.

Figure 5.8. (a) A schematic showing CNT mesh structure. It can be visualized as multiple meshes stacked on top of each other, (b) A schematic of rectangular unit cell model. It has inner side dimension “(t x h)” and outer dimension "(t+d) x (h+d)", where t and h are

133

Figure 5.8. (continued) the average length and width respectively of a rectangular cavity between nanotubes in a mesh structure and d is the average diameter of CNT, (c) G* calculated as a function of r for θ Y = 80° (CNT) and θ Y = 110° (PCNT); 1< r <2.9.

The calculation of ( G* CB - G* W ) depends on the value of fs. We calculated the value of fs to be 0.204±0.05 using the rectangular cell model. At every point, G* for corresponding fs value was calculated and difference between G* and G* was

plotted as a function of r. The range of f s thus, defines the lower and upper limit for in the case of both CNT and PCNT as shown by highlighted regions in the graph.

Therefore, for any given “r”, the range of can be estimated from the model calculations here.

Differences in the wetting behavior of PCNT and CNT mats under steam condensation can be explained based on the modification of surface chemistry and, thus, the intrinsic wettability (θ ) of CNTs. The intrinsic contact angle of the surface of a single CNT has been measured to be about 80°.[199] After plasma coating, we estimate the intrinsic contact angle of PCNT surface to be 110° from the contact angles measured for plasma-coated flat plates. We propose here a model to predict whether the Cassie-

Baxter or the Wenzel states are thermodynamically favored for CNT or PCNT surfaces.

The free energy of the given wetting state as a function of corresponding apparent contact angle (θ) and intrinsic wettability (θ ) can be expressed in its dimensionless expression form.[97] For a Wenzel state, the free energy can be written as:

-2/3 2 G* = [F (θ W )] [2 – 2 cosθ – r sin θ cos θ ]…………………………………(5.1)

134

Where, θ W is the Wenzel contact angle defined as follows:[37]

cos θ = r cos θ Y …………………………………………………………………(5.2)

Here, “r” is the surface roughness defined as the ratio of actual surface area to the projected surface area (r >1). F(θ) is defined as:

F(θ) = (2 – 3 cos θ + cos3θ)……………………………………………………………(5.3)

Similarly, the expression for free energy of the Cassie-Baxter state can be written as:

-2/3 2 G* CB = [F (θ CB )] [2 – 2 cos θ CB – sin θ (–1 + f s + f s cos θ )]…………………..(5.4)

Here, θ is the Cassie-Baxter contact angle defined as:[38] cos θ = – 1 + f (1+ cos θ )………………………………………………………(5.5)

f is the fraction of solid in contact with water when the droplet is supported on air-solid composite surface. The free energy difference between the Cassie-Baxter and

Wenzel state can be defined as G*= G* – G* W . We use a model rectangular unit cell as shown in Figure 5.8b to estimate f . Using SEM and image analysis we calculate fs =

0.204±0.05. Based on the values of fs, the model predicts θ value to be 140° ± 5° for

CNT surface and 150° ± 4° for PCNT surface. It can be seen that the θ values calculated using the Cassie-Baxter model are close to the static contact angle measured experimentally. Since, 0° < θ < 180°, G* and thus, can only be defined for 1 < r

< 2.92. The upper limit on “r” is imposed by θ = 110° for PCNT. In the range of “r” where real θ is defined, it can be seen (figure 5.8c) that for CNT surface is an increasing function of “r” implying that Wenzel state becomes more favorable as surface roughness increases for θ = 80°. In the case of PCNT surface with θ = 110°,

135 decreases as “r” increases and crosses the G*= 0 line for higher values of “r”. The negative values of imply that the Cassie-Baxter state becomes thermodynamically more favorable state as the surface roughness is increased. For these CNT and PCNT surfaces we expect “r” to be high and most likely higher than 3. Thus, for nano-porous mesh like PCNT surface the Cassie-Baxter state with high contact angle is more favored whereas for CNT surface, Wenzel state with low contact angle is more favored thermodynamically. This implies that even though the Cassie-Baxter wetting state is possible for both CNT and PCNT surfaces (based on values of fs estimated), thermodynamically, it represents only a metastable state for CNT surface. After steam condensation, when a droplet in the more stable Wenzel state is formed on this surface, its failure to transition to Cassie-Baxter state results in the loss of superhydrophobicity.

On the other hand, even if condensate forms a Wenzel state droplet on exposure to steam on PCNT surface, the transition from Wenzel to Cassie-Baxter state is favored in this case. This explains the retention of superhydrophobicity of a PCNT surface even after prolonged exposure to steam.

The thermodynamic model here thus helps predicting the stable wetting state of the water droplet as it is formed. Unlike water droplet deposited on the CNT and PCNT surfaces in water contact angle tests, in the case steam condensation, the water vapor condenses on the mat surface to form a liquid droplet. The wetting state in which the condensed water droplet is formed and its relative stability is the deciding factor for steamphobicity of a surface. For a surface to be steamphobic i.e. to stay dry under steam condensation, the water droplet needs to be stable in the Cassie-Baxter state associated with low contact angle hysteresis, irrespective of the mechanism by which it forms on the

136 given surface. The model here clearly implies that formation of a stable Cassie-Baxter state compared to Wenzel state is possible only in the case of PCNTs and thus, explains steamphobicity of the PCNT surface. CNT surface on the other hand fails to act steamphobic due to stable Wenzel state formed of the condensed water droplet. From the classical theory of nucleation and growth of a condensate, it is known that it is dependent

on intrinsic wettability (θ Y ) of the surface. It is estimated that the free energy barrier that is required to overcome for the condensation of water vapor is higher and condensation rates are lower on a hydrophobic surface compared to a hydrophilic surface. [104, 195]

Since, CNT and PCNT surfaces used here have similar structures and differ only in intrinsic wettability, the condensation kinetics can be expected to be different and may also have contributions is defining steamphobicity of a surface.

5.2.2.2 Low temperature condensation

The steam condensation experiments clearly demonstrated the difference in the wettability of CNT and PCNT surface. We further examined the behavior of these surfaces under low temperature condensation (dewing) to find out if the wettability contrast could still be observed.

137

Figure 5.9. Water condensation on the surface of CNT (scale bar: 50µm)

The CNT sample held on cold stage, when cooled down at a constant rate (40-

45% humidity), it was observed that water condensation started at 3° C (figure 5.9). At

1°C, the droplet can be seen to have grown big enough. The observation of the shape of the droplet can be seen to be distorted (not perfectly) spherical, which implies wetting of the surafce. However, the condensed droplet did not form a film, unlike steam condensation, which implies that the wetting state observed here could possibly an intermediate wetting state of the CNT surface i.e. between Wenzel (complete wetting of the surface) and Cassie-Baxter (Non stick droplet).

138

Figure 5.10. Water condensation on the surface of PCNT (scale bar: 50µm)

The condensation on the surface of PCNT carried out in a similar way shows that condensation is retarded slightly (starts at about 1°C). As the water condensation continued, it was observed that the surface of PCNT ha to be cooled down to almost -

11°C for the droplet to grow as big as the one on the CNT surface at 1°C, which suggests slower rate of condensation on the PCNT surface. Also, the shape of the droplet remained perfectly spherical even when the droplet froze at -12°C. As the stage was heated and frozen droplet melted, it still was observed to continue be a perfectly spherical shape

(Cassie-Baxter state) and rolled off the surface upon coalescence leaving the surface perfectly dry, which suggests the retention of superhydrophobicity on PCNT surface under low temperatures condensation as well. The anti-wetting properties were observed to be unaffected over multiple cooling and heating cycles. The difference in wettability could be attributed to mainly the difference in inherent wettability of CNT versus PCNT

139 since the thermal conductivity could be expected to remain more or less the same due to almost constant surface roughness.

5.2.2.3 Anti frost properties

a b

Figure 5.11. Frost formation on the surface of CNT (a) and PCNT (b)

The frost formation on the surface of CNT and PCNT was carried out by rapidly cooling the surfaces down in contact with dry ice (~ -75°C). The optical images of frost formed are shown in figure 5.11 above. There was no difference observed between the morphology of the frost formed (as can be seen in images 5.11a and 5.11b). Also, the rate of the frost formation did not appear to differ too much. The expected differences in the frost formation on the CNT and PCNT surfaces are probably too subtle to be observed under the experimental conditions in which they were carried out. Nonetheless, the wettability differences were evident in the defrosting step. Defrosting on the surface of

CNT resulted in the water droplets left behind on the surface. On the other hand, defrosting on the surface of PCNT surface results in the formation of water droplet which either bead up on the surface with almost perfectly spherical shape (Cassie-Baxter state of the droplet) or roll off the surface without a need of tilting the surface leaving the surface dry perfectly. The behavior is attributed to thermodynamically more favorable

140

Cassie-Baxter state of water droplet on the surface of PCNT. Over the multiple frosting- defrosting cycles, the superhydrophobicity of the surface was observed to remain unaffected.

5.3 Conclusion

It can thus be concluded here that two surfaces which do not differ much structurally and are characterized as superhydrophobic surfaces exhibit extreme steamphobic behavior. The combination of nano-scale roughness of the CNT mats along with the surface energy contrast achieved by deposition of low surface energy plasma deposited coating results in the formation of a highly robust steamphobic surface. The biggest advantage of PECVD process used here for chemical modification is that it is a completely dry process. Unlike wet processes, in which the capillary forces tend to collapse the nano-porosity, the PECVD process enables deposition of a very thin layer of a film on such a surface without disturbing its structural integrity. Use of stainless steel to form the steamphobic coatings is an added advantage since it is a widely used industrial material. The robust steamphobic surfaces are of particular interest in designing heat exchangers. They are ideal for dropwise steam condensation and can be expected to remarkably improve the exchanger efficiency due to fast droplet removal and low adhesion of condensed steam to the surfaces. Robustness of the structure can also be taken advantage of in anti fog, anti ice and anti dew coatings.

141

CHAPTER VI

ADHESION AND WETTING OF GECKO TOE PADS: THE ROLE OF SURAFCE

CHEMISTRY

6.1 Motivation

The gecko adhesive system has been a topic of interest for researchers for centuries, due to its unique ability to stick reversibly with the help of dry adhesive toe pads. Over the last decade, researchers have made extraordinary progress in understanding how the gecko adhesive system works. Indeed, many laboratories have tested hundreds of synthetic mimics for potential use in robotics, medicine, space and everyday life. The intense focus to understand the principle behind the working of adhesive toe pads and mimic their ability has led to the understanding that van der Waals interaction is the primary mechanism for adhesion. Thus, the studies are typically focused on the nano-, micro-, and whole-animal mechanics of gecko adhesion on clean, dry substrates, there is relatively little known about the effect of surface chemistry on gecko adhesion.

While the range and performance of synthetic “gecko-tapes” is impressive, there remain important gaps in our knowledge of the system and its capabilities in natural environments. Geckos are extremely diverse, constituting over 1,400 species world- wide.[200, 201] However, knowledge of the natural substrates and conditions geckos use is very limited. For example, it is likely that many species move across leaves and other

142 plant structures which are not perfectly smooth and have variable surface chemistries.[202, 203] In principle, the interaction of gecko feet with such surfaces can have a significant effect on adhesion, yet gecko research has only just begun to tackle such questions. [204, 205, 206] Additionally, natural surfaces are also likely to become wet (especially in the tropics) and dirty, potentially reducing adhesion. Although research on the ability of geckos to remove dirt from their toes has received some attention [25,

207], studies on wetting and the effect of water are limited, despite the well known anti- wetting properties of the toes which are both superhydrophobic and have a low contact angle hysteresis. [122, 208]

The gecko foot is surely an intriguing “smart” surface that has evolved to not only act as a reversible, dry pressure sensitive adhesive but also exhibit unique surface properties such as superhydrophobicity[122, 208] and self cleaning[25]. Geckos' incredible adhesive abilities to climb up vertical surfaces or hang from a ceiling upside down are attributed to the hierarchical, structured surface of their toe pads.[26, 209, 210]

It has been demonstrated with the help of systematic experimental measurements that the van der Waals interactions is the key mechanism behind gecko-like adhesion.[119, 211]

Therefore, it is intuitive that the geckos acquire their adhesion using a split contact mechanism of their hairy structures. The adhesion is dependent upon the morphology and mechanical properties of the hairy structures and independent of inherent surface chemistry of the material the hairs are made of. Although, this principle holds true for adhesion of geckos studied on a variety of surfaces, the intrinsic wettability of gecko hairs can be expected to have an effect on superhydrophobicity and self cleaning property and possibly on adhesion, especially in the presence of high humidity, frequent rainfall or

143 while walking on dirty surfaces. It is not very uncommon for the geckos to encounter water or particulate dirt as they walk on different natural surfaces. Under these circumstances, the geckos need to have developed a mechanism to retain their characteristic surface wetting properties along with the adhesion to be able to survive.

Thus, superhydrophobicity, reversible adhesion and self cleaning ability are not independent properties, understanding the correlation between them is important for studying the adhesion mechanics of gecko toes and engineering behind a natural surface that is optimized for its three distinct properties.

A few recent studies have reported that the effect of water on gecko adhesion is very intriguing. When tested for their adhesion on the surfaces with different wettability, it was observed that on a hydrophilic glass surface geckos can stick strongly with dry toe pads. However, they tend to slip and adhere very poorly with their toes soaked wet with water.[123] Also, a very thin layer of water can be easily expelled by their toes to make a dry contact with a hydrophilic surface, however, no such contact is possible with a thicker water layer (~ 0.5 cm deep).[117, 123] We recently also observed that as the hydrophobicity of the surface that geckos are contacting increases, the underwater adhesion is stronger compared to corresponding dry adhesion.[212] Based on these studies, it is implied that the hydrophobicity of contact surface along with the superhydrophobicity of the gecko toe surface possibly have a synergistic effect on adhesion in the presence of water. The inherent surface chemistry and a very unique surface roughness made up of hierarchical hairy structures are responsible for the superhydrophobic property of gecko toe pads. For a typical superhydrophobic surface, these two factors are accounted for in the Wenzel[37] (homogeneous solid-liquid

144 interface) and the Cassie-Baxter[38] (heterogeneous solid-liquid and air-liquid interface) wetting theories. Interestingly, a water droplet in contact with a gecko toe pad forms a

Cassie-Baxter state, with a contact angle of about 150° and a very low contact angle hysteresis (about 2-3°).[122] On the other hand, a Wenzel state of a water droplet corresponding to complete wetting of the toe is also possible which results in the failure of geckos to adhere.[123] The transition between the states is suggested to be not spontaneous and expected to take place gradually over a period of time (of the order of few minutes).[213] Thus, the metastability of superhydrophobic state and estimation of transition barrier between the superhydrophobic and wetted toe state are important factors for determining wettability of gecko toe pads since it is not favorable for geckos to have wet toes. The biggest challenge in studying the wettability of gecko toe pads is the limited knowledge of the material that the hairy pads are made up of i.e. the surface chemistry of hairy structures. One of the main constituents of gecko toe hairs is β-keratin like protein.[26, 116, 211, 214] However, it was demonstrated by Hsu et. al. that the geckos leave footprints behind as they walk. The analysis of the footprints showed that they correspond to phospholipid molecules, an oil like material which is relatively more hydrophobic than β-keratin.[117] Therefore, the hairy pads of gecko toes can be visualized more like a composite surface made up of β-keratin and some lipid like material. However, since it is a natural surface we have no control over the surface or bulk composition of such a complex composite system, both the components of which have distinctly different wettability. Further, experimental and sample preparation limits impose challenges in analyzing the surface chemistry accurately. Thus, even though the surface morphology of the gecko toe pad is well characterized, the lack of knowledge of

145 inherent surface chemistry limits the fundamental understanding of its wettability. It is expected that the surface chemistry directly affects the surface wettability however, if the surface wettability (effectively, surface chemistry) of toe pads is really a critical factor for wet adhesion still remains unanswered. In this study we have come up with a strategy that allowed us to study the effect of wettability (surface chemistry) of contact surface and isolate the effect of surface chemistry from surface morphology of gecko toe pads in an attempt to understand the effect of surface chemistry on the effective functioning of the gecko toe surface.

We tested the effect of water on the gecko adhesive system using a range of surfaces with different wettability defined by water contact angle ( . Since geckos probably encounter both hydrophilic (such as some flowers and roots) as well as hydrophobic (most plant leaves) surfaces [202, 215], we tested surfaces that are hydrophilic ( = 49.6°  1.4°), intermediately wetting ( = 85.2°  0.5°) and hydrophobic ( = 94.0°  0.5°). We also tested the effect of water on adhesion to polytetrafluoroethylene (PTFE), a synthetic substrate geckos cannot adhere to in dry conditions. [216]

We adopted a novel approach of selectively changing the surface chemistry by depositing a thin coating of known functionality on setae surface, without affecting the surface roughness significantly. We have used Plasma Enhanced Chemical Vapor

Deposition (PECVD) process for this purpose. The parameters for PECVD can be chosen to deposit thin layer so that the surface roughness is affected only minimally. Further, this dry vapor process is particularly advantageous in comparison to a wet process. A wet process results in collapsing or disruption of hairs due to capillary forces during drying.

146

We chose two PECVD precursors with different intrinsic wettabilities (quantified in terms of equilibrium water contact angle, θY), a hydrophilic precursor (maleic anhydride,

θY ~ 48°) and a hydrophobic precursor (1H,1H,2H-perfluoro-1-dodecene, θY ~ 110°)for the PECVD process. The deposition using PECVD was carried out on the toe skin sheds molted naturally by geckos. The wetting and adhesion results of PECVD coated sheds with known surface chemistry were compared to the results obtained using non-coated, blank sheds keeping the surface roughness relatively unchanged.

6.2 Results and discussion

The experimental measurements and theoretical calculations carried out to study the effect of surface chemistry on gecko adhesion are discussed in the following sections.

6.2.1 Effect of contact surface wettability (surface chemistry)

The surface chemistry of both contact surfaces i.e. the gecko toe and the surface it is adhering to can be expected to play role in interfacial adhesion. In order to isolate the effect, we first studied the effect of surface wettability while keeping the shed surface chemistry constant.

6.2.1.1 Whole-animal Adhesion

Gecko adhesion in the shear direction (frictional adhesion) was tested on four surfaces with different wetting properties: glass, plexiglass, OTS-SAM coated glass and

PTFE. Water contact angles ranged from approximately 50° to 100° (Table 6.1). Whole- animal shear adhesion was measured on all four surfaces when submerged under water and when dry to test for a difference in adhesion when either air or water was the medium of contact. We found a significant interaction between surface wettability and medium of

147 contact. The gecko adhesion on the glass surface was significantly higher when the surface was dry (17.13  3.93N) than when the glass substrate was submerged under water (5.40  1.33N). On the plexiglass and the OTS-SAM coated surfaces a significant difference in wet versus dry adhesion was not observed (23.98  3.92N plexiglass; 17.63

 2.22N OTS-SAM in wet conditions and 26.43  1.94N plexiglass; 19.95  1.71N OTS-

SAM in dry conditions). Conversely, and quite surprisingly when tested on PTFE geckos produced a significantly higher adhesive force underwater (8.04  1.09N) than on dry

PTFE (1.56  0.66N), unlike the results on all other surfaces (Figure 6.1).

Figure 6.1. Whole-animal shear adhesion data from Tokay geckos (Gekko gecko) tested on four surfaces in dry or wet contact.

148

Table 6.1. Contact angles of water and n-hexadecane on all four surfaces used in whole-animal experiments and modeling. Errors are means  1 SEM.

Water (°) n-Hexadecane (°) Glass 49.6 ± 1.4 12.8 ± 2.0 Plexiglass 85.2 ± 0.5 8.9 ± 2.3 OTS-SAM 94.0 ± 0.5 28.5 ± 2.9 PTFE 97.4 ± 0.3 30.2 ± 1.6

6.2.1.2 Smooth Surface Adhesion Model

We tested the hypothesis that the ratio of shear adhesion values from the whole- animal experiments on wet and dry surfaces could be predicted from a classical thermodynamic model, based on the work of normal adhesion (W) between two flat surfaces in air and in water. The Dupré equation for calculating the work of adhesion between two surfaces, 1 and 2, in its general form is written as follows:

………………………………………………………….(6.1)

Here, is the work of adhesion between the surfaces 1 and 2, AC is the area of contact, , , and are the surface energies of components 1 and 2 and interfacial energy of the contact between 1 and 2 respectively.

We used equation 6.1 to calculate the work of adhesion between each of our four surfaces (glass, plexiglass, OTS-SAM coated glass and PTFE) and the “gecko hair-like” n-hexadecane surface, assuming that the contact interface formed as a result of contact between the two is flat (Table 6.2). The Young-Dupré equation for the dry contact between the two surfaces when air is the medium of contact can be written as follows (1):

149

……………………………………………….(6.2)

Where, is the interfacial energy at the contact interface between the “gecko hair- like” surface and the contact surface (glass, plexiglass, OTS-SAM coated glass or PTFE),

is the surface energy of “gecko hair-like” n-hexadecane surface and is the surface energy of the contact surface. Young’s equation for the contact angle ( ) that n- hexadecane makes on a given contact surface is:

…………………………………………………….....(6.3)

Substituting equation 6.3 in equation 6.2 for we get:

…………………………………………………….(6.4)

We measured the contact angle of n-hexadecane on all four surfaces that we used for the gecko trials to obtain the value of (see second column of Table 6.1). The value

2 of is known to be 25 mJ/m . Substituting all the known values in equation 6.4 gives the work of dry adhesion ( . Similarly, the equation for wet adhesion (Wwet) can be derived as follows (derivation in appendix):

………………………..(6.5)

Using equations 6.4 and 6.5, we calculated the ratio for our model and compared this to the ratio of wet to dry shear adhesion values obtained experimentally from whole-animal adhesion measurements on each of the four surfaces

(Table 6.2). The model calculations show that relative normal adhesion on any given test surface when tested in air and under water is a function of surface wettability. As the hydrophobicity of the surface increases (water contact angle increases), the value of

150

increases. The experimental values from whole-animal adhesion experiments also show a similar trend (Table 6.2), supporting our hypothesis that surface hydrophobicity can have a significant thermodynamic impact on gecko adhesion when tested underwater.

6.2.1.3 Patterned Surface Adhesion Model

To consider the effect of patterning of the setal mats we took into account the gecko toe surface morphology by modeling a tetrad-patterned “gecko hair-like” surface, as shown schematically in figure 6.2. The tetrad unit cell was made up of four square pillars each with a square side width of 4 μm and height of 60 m. The distance between any two adjacent pillars in one tetrad was 1 μm and the separation between two adjacent tetrads was 2 μm (Figure 6.2).

Figure 6.2. Unit cell representing the tetrad pattern of the gecko toe surface

When the tetrad structure of “gecko hair-like” n-hexadecane is in contact with either of the four flat surfaces (glass, plexiglass, OTS-SAM coated glass or PTFE), is estimated using equation 6.4. For of a patterned surface there are four different

151 cases possible of “separated” and “in contact” states that are associated with air pockets between the square columns before or after contact with the flat test surface (see Table

6.2 for schematics of each case). In all cases the water surrounding the unit cell is also accounted for in , whereas the adhesive interface is always assumed to be dry. In case 1, both, the “in contact” and the “separated” states maintain air pockets in the space separating the square columns. for case 1 can be calculated using the following equation (equation 6.6) where A1 and A2 correspond to total surface areas of the “gecko- hair-like” n-hexadecane tetrad patterned unit cell and the surface it is in contact with

(glass, plexiglass, OTS-SAM coated glass or PTFE), respectively.

………………………..(6.6)

In the “separated” state of case 2, water penetrates completely inside the tetrad asperities in such a way that the entire surface area of the tetrad unit cell is in contact with water. However, the “in contact” state expels all the water out and the asperities occupied with water initially are replaced by air pockets. The equation to calculate for case 2 is:

……(6.7)

Cases 3 and 4 represent the possibility that the gaps between the pillars in a tetrad are completely filled with water in the “in contact” state, however, in the “separated” state the gaps are either air pockets or filled with water for case 3 and case 4 respectively.

Equation 6.8 is used to calculate for case 3 and equation 6.5 calculates for case 4 (since it is only AC that is different for the patterned surface compared to the flat surface).

152

………………………………………………………………………………(6.8)

Derivations for each case are included in the appendix as are derivations for a non-tetrad

patterned unit cell model.

Table 6.2. Ratios of wet to dry adhesion on all four surfaces used in whole-animal

experiments and modeling, the ratios calculated based on experimental results are

included for comparison as well.

Surface Smooth Gecko Surface

Case 1 Case 2 Case 3 Case 4

Glass 0.56 -0.28 30.54 -30.27 0.56 0.32

Plexiglass 1.38 1.27 31.9 -29.25 1.38 0.91

OTS-SAM 1.64 1.73 34.14 -30.76 1.64 0.89

PTFE 1.74 1.91 34.57 -30.92 1.74 5.19

Using our calculations of the ratios of each surface in various

conditions, we qualitatively compared our model with experimental results. The

ratios for smooth surfaces and ratios for case 4 of the patterned surface are the

same (Table 6.2). This is due to the fact that the common multiplication factor

corresponding to the total surface area for and cancels out as the ratio of the

two is taken. Similar to the smooth surface model, hydrophobicity of the surface has a

significant impact on the ratio however each case presents different results

153 for a given surface, suggesting that the presence of water between or around the hairs also plays an important role in adhesion underwater. Based on our results we can clearly rule out case 2 and case 3 as possible conditions of pre- and post contact between the patterned “gecko hair-like” surface and the flat test surfaces, as ratio values are either extremely high, favoring wet adhesion over dry (case 2) or extremely low, favoring dry adhesion over wet (case 3), neither of which occur in the whole-animal system.

ratios for case 1 and case 4 are most similar to those from whole-animal experiments, where glass is the least favorable in wet conditions and ratios on the hydrophobic surfaces are near one, supporting equal adhesion in wet and dry conditions

(Table 2). Modeled ratios involving PTFE do not explain our experimental results (the experimental ratio is much higher than all modeled ratios) and will be discussed later.

To further investigate case 1 and case 4 we imaged the gecko foot contacting glass and the OTS-SAM coated surfaces under water after the gecko took four steps, similar to experimental trials (Figure 6.3).

Figure 6.3. Images of a gecko foot (A) in dry contact with glass, (B) in wet contact with a hydrophobic surface (OTS-SAM coated glass) and (C) in wet contact with a hydrophilic surface (glass).

154

The hydrophobic OTS-SAM coated surface (Figure 6.3b) shows typical characteristics of a dry contact (Figure 6.3a) such as a white rather than a gray appearance of the adhesive toe pads and their tendency to remain dry when taken out of the water. This suggests that our results are most similar to case 1, a constantly dry contact under water with air pockets between the surface asperities in both the

“separated” and “in contact” states. On the hydrophilic glass surface however, we see one of three scenarios taking place typically, either a consistently dry contact (case 1), an entrapped air bubble between the toe and the glass surface (shown as a silvery layer on the toe; Figure 6.3c) or one where water wets the toes (case 3) and the wet toes appear gray when removed from the test surface.

The shear adhesion to the hydrophilic glass surface was much lower in water than in air experimentally (Figure 6.1) and was consistently lower in our model calculations when compared to the other three surfaces (Table 6.2). The anomalous behavior in the case of the glass surface is not surprising given that the surface is hydrophilic and thus shows higher affinity to water compared to other test surfaces. For example, three scenarios may occur when two surfaces contact underwater.[217] First, a layer of water between the surfaces acts as a barrier to the establishment of contact between the two.

Second, roughness of the contact surface may change the contact area and may also leave small pools of water behind. Thirdly, water may be fully excluded from the two surfaces and completely dry contact occurs. Depending on the gecko’s foot placement, we found that either a dry contact formed by squeezing water out completely (case 1) or an air bubble formed (Figure 6.3c) which acted like a lubricating layer causing the gecko to slip. There is also another scenario that may occur in whole-animal gecko adhesion on

155 wet surfaces where a wetting transition occurs and the toe wets completely, causing a substantial drop in adhesion experimentally[123] and theoretically (case 3). These observations suggest that natural placement of the foot on the wet surface is important and adhesion can be highly variable when the surface is hydrophilic, potentially limiting a gecko’s movement on wet hydrophilic surfaces. Interestingly not all feet, and possibly toes, show the same scenario in any single trial; rather it is always a combination of two or more behaviors happening at the same time which complicates the use of models, like ours, that assume homologous contact.

Our results on plexiglass and OTS-SAM coated glass are consistent with our hypothesis that adhesion will not be affected by water on surfaces that are more hydrophobic in nature than glass. This suggests that gecko adhesion is not impaired by surface wetting in their natural environments, assuming their native substrates are at least moderately hydrophobic (perhaps  85°). One of the main surfaces we expect geckos to be walking on is plant surfaces, such as leaves, which are primarily hydrophobic due to their waxy cuticle.[202] By imaging the foot in contact with the hydrophobic OTS-SAM coated glass it is clear that water is excluded from the majority of the adhesive toe pad and the contact made underwater is dry (Figure 6.3b). Our experimental results are also supported by our model and ratios of wet to dry adhesion are near one (wet adhesion is not different from dry adhesion) in case 1 where the “gecko hair-like” surface remains dry prior to and during contact. The maintenance of dry contact is interesting because when two surfaces become separated water is likely to penetrate the separation crack forming between the two surfaces and even propagate further growth.[218] This may be what occurs when testing normal and frictional (shear) adhesion of a patch of setae,

156 where normal adhesion in water is much lower, due to water penetration, than frictional adhesion in water under high loads which presumably keep water from separating the two surfaces.[213]

As discussed before, the dominant mechanism behind the gecko’s ability to stick is van der Waals forces which in dry contact should be relatively insensitive to surface chemistry. Of the three surfaces with similar Hamaker constants (see derivations in appendix): glass (6.5 x 10-20J), plexiglass (6.2 x 10-20J) and a glass plate coated with

OTS-SAM (6.5 x 10-20J, assuming the OTS coating has no effect on the dielectric properties of glass surface on which it is formed), dry adhesion on the plexiglass surface was almost significantly higher than both the dry glass (t = 2.33, df = 5, p = 0.0672) and the dry glass coated with OTS-SAM (t = -2.52, df = 5, p = 0.0535). The plexiglass surface had a contact angle of 85.2°  0.5° and a slightly lower Hamaker constant.

Although it is not entirely clear why dry plexiglass was marginally better than the other two surfaces in our whole-animal experiments, it is interesting to consider an optimal surface for the gecko adhesive system and how such a surface may correlate to the natural surfaces of their environment.

In contrast, the Hamaker constants calculated in water are lower than in air for all the substrates studied here (calculations in appendix). This is expected based on van der

Waals interactions and may indicate that the adhesion forces to separate non-polar surfaces in water should be lower in water. However, this is misleading because Hamaker constants do not accurately predict the interfacial energies of non-polar materials in water and requires the addition of hydrogen bonding.[24] A simpler explanation for higher

157 shear adhesion forces for non-polar surfaces in water is the higher interfacial energies of non-polar materials in water than in air as calculated by our model.[219]

The final surface, PTFE, provided surprising results. We quantitatively confirmed previous observations[216] that geckos do not stick well to dry PTFE (1.56  0.66N).

When tested on PTFE submerged under water however, the geckos clung to the fully wetted surface significantly better than the dry (8.04  1.09N), contrary to our findings on all other tested surfaces. Unlike our previous results on each of the other surfaces, our model does not predict the whole-animal experimental results. Our experimental ratio is fivefold higher than our theoretical ratio and this discrepancy could be due to the low adhesion of geckos to dry PTFE or the comparably high adhesion geckos have on wet

PTFE. The experimentally measured shear adhesion to dry PTFE is 10-15 times lower than shear adhesion to a surface with similar water contact angle (OTS-SAM coated glass). This low adhesion on PTFE cannot be explained by the small difference in the

Hamaker constants (4.6 x 10-20J for PTFE compared to 6.5 x 10-20J for OTS-SAM). We also do not believe that static charging is playing a role here because the adhesion values for PTFE are lower, rather than higher, than those predicted by contact angles or

Hamaker constants. In addition, surface charges will be neutralized under water and cannot influence the shear adhesion values measured in water. One reason the experimental values of dry shear adhesion on PTFE are lower than the expected values from our theoretical models, or compared to a surface with similar water contact angle, may be related to the abnormally low coefficient of friction of PTFE. Interestingly, underwater shear adhesion values for PTFE are vastly improved and closer to our expected values for hydrophobic surfaces. We believe that the adhesion in air is

158 anomalous for PTFE, resulting in much larger ratios for wet versus dry shear adhesion forces. Additionally, we hypothesize that the roughness of PTFE may also play an important role. Although the roughness of the PTFE surface does not change when wet or dry, when PTFE is submerged under water the roughness may be less important to adhesion because water is able penetrate between the rough surface asperities. However when dry, the roughness of PTFE may cause air gaps and reduced contact area, lowering adhesion values. It is clear that further work is necessary to clarify the effect of roughness on adhesion to wet and dry surfaces. Interestingly, synthetic gecko-like PTFE pillars tested underwater with a silica probe were also successful in achieving adhesion; however adhesion values underwater were not five times higher than dry as measured in our experiments.[220]

6.2.2 Effect of gecko hair wettability

As discussed in the motivation, the ambiguity of the surface chemistry of gecko hairs is the biggest challenge for studying the effect hair wettability might have on dry and wet adhesion. Based on the results of the effect of surface wettability, as discussed in the previous section, we expect the surface wettability of the gecko toe to have some contribution towards its adhesion as well. In order to study the effect of hair wettability, we adopted an approach to modify the surface of hairs with known chemistry, making sure that the roughness is affected only minimally. This was done by deposition of

PECVD coating on the surface hairs. We used two different types of precursors, maleic anhydride (M-S) and 1H,1H,2H-perfluror-1-dodecene (F-S) for plasma deposition. We chose shed samples for the coating. The results are discussed in the following section and are compared to the wetting and adhesion data of the blank sheds (B-S).

159

Figure 6.4. The SEM images of B-S (a,b), M-S (c,d) and F-S (e,f). Comparison of lower magnification images (a,c and e) indicate that the tetrad pattern of the sheds was retained after PECVD layer deposition. The higher magnification images (b,d and f) show the finer structures present on the sheds (spatulae)

We used SEM imaging to study the morphology of uncoated and PECVD coated sheds. Figure 6.4 shows SEM images of blank sheds (B-S), the sheds coated with a

PECVD layer of Maleic Anhydride (M-S) and 1H,1H,2H-perfluoro-1-dodecene (F-S). A tetrad pattern of micron scale setae and every seta branching out into hundreds of finer spatulae were the two typical surface features of the toe pad sheds, as observed in images of B-S in Figures 6.4a and 6.4b. Figures 6.4c and 6.4d correspond to M-S and Figures

6.4e and 6.4f correspond to F-S. The typical characteristics such as tetrad pattern and hierarchical structure of the sheds were retained even after the deposition of a PECVD layer, as can be seen from these images. Based on this observation it can be assumed that

160 the thickness of the coating was only a few nanometers and therefore, did not affect the overall roughness of the surface significantly.

Figure 6.5. (a) XPS survey spectra of B-S, M-S and F-S. The peaks assignable to constituent elements are labelled. The survey spectra were used for calculating surface atomic composition. The high resolution C1s XPS spectra of (b)B-S, (c) M-S and (d) F-S are compared

The chemical composition of blank and PECVD coated sheds was studied using surface sensitive XPS measurements. Figure 6.5 compares the survey and C1s high resolution XPS spectra of B-S, M-S and F-S. The survey spectra shown in Figure 6.5a are used for calculating relative atomic compositions of a given surface. The survey spectrum of the B-S sample showed the presence of C1s, N1s, O1s and S2p peaks, implying that they were the constituent elements of gecko toe pad surface. The survey scan of M-S

161 showed the presence of C1s and O1s peaks, whereas that of F-S showed C1s and F1s.

The elements detected in both the spectra were consistent with the chemical structure of the precursors we chose for PECVD and atomic compositions were comparable with control measurements. Interestingly, no peaks exclusively corresponding to B-S (N1s and

S2p) were seen in the survey scans of either M-S or F-S. Thus, we estimated the PECVD layer thickness on the surface of M-S and F-S to be roughly about 10-15 nm, comparable to the typical analysis depth of XPS. The atomic compositions and ratios are listed in

Table 6.3. To further analyze the composition of the surface functional groups, we carried out high resolution scans in the narrower scan width. Figures 6.5b, 6.5c and 6.5d show high resolution C1s spectra of B-S, M-S and F-S, respectively. The C1s envelope obtained for M-S (Figure 6.5c) was de-convoluted into three Gaussian peaks, 284.6 eV

(CxHy), 285.6 eV (different types of bond between C and O as listed in supporting information) and 288.2 eV (anhydride linkage), all the peaks were consistent with

PECVD layer of maleic anhydride (data for control samples are provided in supporting ino).[166] [167] Similarly, all the four peaks in the C1s spectrum of F-S were assigned to different types of C-C and C-F bonds in the structure of 1H,1H,2H-perfluoro-1-dodecene

(Figure 6.5d). The peak at 288.8 eV was the most dominant one and attributed to the surface CF2 [172, 151]. The analysis of C1s spectrum of B-S was complicated since for the first time we have used XPS analysis to characterize the gecko toe pad surface. Even though the exact chemical composition of the B-S surface is not known, β-keratin [26] and phospholipid [117] have been identified as the components of the B-S surface. Based on this knowledge, the possible chemical bonds contributing to the C1s spectrum of B-S surface could be identified. The peak positions of the spectrum were consistent with C-N,

162

C-S and C-O bond regions. However, there is a possibility of an overlap of multiple bonds for a single peak since the exact molecular structure is not known. With the help of

XPS analysis thus, we confirmed that a thin layer of PECVD coating was deposited on the structured surface of toe sheds such that all three surfaces of B-S, M-S and F-S have distinctly different surface chemistry and thus, intrinsic wettability.

Table 6.3. Summary of atomic composition and atomic ratios of B-S, M-S and F-S surfaces calculated using survey XPS spectra

In order to study the effect that surface chemistry (or the surface wettability) has on adhesion, we tested the adhesion of blank and coated sheds, i.e. B-S, M-S and F-S samples on a hydrophilic glass surface and a hydrophobic OTS-SAM surface in air (Wdry) and underwater (Wwet). The results for dry versus wet adhesion measurements are shown in Figure 6.6.

The whole model for the effect of coating (B-S, M-S, F-S), test substrate

(hydrophilic glass or hydrophobic OTS-SAM) and treatment testing conditions (air or water) was highly significant (df = 12, F = 19.7571, p <0.0001*) as were many of the interactions, including the three-way interaction of coating*surface* treatment. The shed

163 area (a two-dimensional area of contact) was included as a co-variate however none of the interactions with area were significant and these interactions were subsequently removed from analysis.

Figure 6.6. Average force per unit area of tokay gecko (Gekko gecko) skin sheds either the blank (B), M-S (M) or F-S (F). Samples were tested on a clean hydrophilic glass slide

(a) or on a glass slide coated with hydrophobic OTS-SAM (b). Ten samples in each coating and surface group were tested in air (black bars) and ten were tested in water

(grey bars). Error is reported as 1 s.e.m. and significant differences are indicated with a *.

Our results showed that the effect of coating varies as a function of test substrate and treatment condition. Specifically, there was a significant drop in force when samples were tested underwater on a hydrophilic glass substrate compared to tests performed in air (Figure 6.6a). However, this did not occur on the hydrophobic OTS-SAM coated substrate (Figure 6.6b). Overall force values from samples tested on the OTS-SAM substrate in air were lower than those tested on the glass substrate. Using multiple contrasts, we found that uncoated, blank samples (B-S) tested in air and water on the glass substrate were significantly different (F = 49.9798, p <0.0001*), as were those of

164

M-S (F = 54.9360, p <0.0001*) and F-S (F = 13.2583, p = 0.0004*). None of the comparisons in air and water on the OTS-SAM coated substrate were significantly different (F = 0.5782, p = 0.4487 for B-S, F = 0.7179, p = 0.3987 for M-S and F =

0.8312, p = 0.3640 for F-S). We observed that a layer of water was always present at the glass-shed interface, even though the B-S and F-S samples continued to stay dry after underwater adhesion measurements. The M-S surface however, wetted as soon as it came in contact with water. The comparison of wet versus dry adhesion measured on hydrophilic glass and hydrophobic OTS-SAM surface is consistent with the similar measurements done at whole animal scale i.e. the underwater adhesion stays either unchanged or increases on a hydrophobic surface, as long as there is no intervening layer of water.[212] Even though the trends expected on glass and OTS-SAM match qualitatively, the discrepancy between the absolute values of F/A measured was striking.

We attributed these differences to two different mechanisms by which the samples adhered to a given surface, under constant normal preload and shear rate. Examination of the behavior of the samples during testing revealed whether they held firmly to the contact surface and released at a maximum force value just before the sliding started

(stiction) or if a maximum force value was reached during sliding of a surface over the samples (friction). As a general trend, the force values measured in the case of "stiction" were higher compared to "friction". Three different scenarios varied in their propensity for stiction and friction behavior. When tested on hydrophobic OTS-SAM we found that friction predominated in air and stiction in water (χ2= 10.7692, p = 0.0010*) for M-S samples. We also found that B-S samples tested on the hydrophobic OTS-SAM substrate nearly reached maximum force values by friction in air (χ2 = 3.8095, p = 0.0510).

165

Conversely, we found that B-S tested on the hydrophilic glass surface reached maximum force values by stiction in air and friction in water (χ2 = 9.8990, p = 0.0017*). The values of dry and wet adhesion (forces measured experimentally) thus, could not be compared directly for glass and OTS-SAM since the contribution of either friction or stiction taking place during sliding experiments was not the same for each surface. The ratio Wwet:Wdry was in fact more appropriate comparison for B-S, M-S and F-S measurements on glass and OTS-SAM surface (Table 6.4). Irrespective of the surface chemistry of the shed sample, the wet adhesion on a hydrophilic glass surface was reduced significantly compared to dry adhesion (Table 6.4). On the other hand, the ratio Wwet:Wdry being close to 1 on hydrophobic OTS-SAM implied that the wet adhesion was almost as good as dry adhesion (the ratios are consistent with the multiple contrast analysis discussed before).

Thus, our results showed that sheds can stick only to hydrophobic surfaces in the presence of water and fail to adhere to the wet hydrophilic surface. The wet adhesion on hydrophobic surface however, showed no dependence on intrinsic wettability of gecko toe hairs, i.e. irrespective of if they are hydrophilic or hydrophobic (resulting in the surface to be either wet or dry underwater) the Wwet:Wdry ratio was consistent. This is surprising since it has been reported before that the geckos fail to adhere when their toes are wet.[123] It is contradictory to the wet adhesion measurements of hydrophilic M-S surface which wets almost instantaneously in contact with water and was still able to retain the adhesion by possibly creating a dry contact interface in contact with OTS-SAM underwater. It is thus, implied that unlike hydrophilic M-S surface, the ability of B-S surface to stay dry underwater is in fact critical for being able to stick. It also suggests that the interactions between B-S surface and water are much more complex than what

166 could possibly be explained based on hydrophilicity or hydrophobiicty of surfaces; additional mechanisms such as reversible conformational changes of functional groups at the surface of B-S probably also take place in the presence of water. Accounting for all the different types of interactions possible with water is however, beyond the scope of this study. Nonetheless, understanding how easily B-S surface could be wetted, if it could restore its superhydrophobicity on drying and be able to stick as well as its dry state, demands systematic study of wettability of B-S surface to establish boundary conditions for a dry versus wet state of the toe surface. Again, the lack of knowledge of inherent surface chemistry of B-S limits studying its wettability. Therefore, we used M-S and F-S surfaces, with known surface chemistries, as reference surfaces to study the surface wetting properties of B-S which is important to help understand the consequences of surface wettability of gecko toe pads on adhesion and self cleaning better.

Table 6.4. Wwet:Wdry calculated from experimental data for adhesion measurements on glass and OTS-SAM surface

167

Table 6.5. Summary of water contact angles, The intrinsic contact angles (θY) are measured on flat, control samples and apparent contact angles (θapp) are measured on blank and PECVD coated shed samples. The inset figures show water in contact with different surafces, it beads up to form Cassie-Baxter wetting state on the surafces of B-S and F-S and spreads completely on the surface of M-S (scale bar : 500 µm)

To study the wettability of gecko sheds, water contact angles were measured on

B-S, M-S and F-S surfaces (Table 6.5). The surfaces of both B-S and F-S show characteristics of a typical superhydrophobic surface i.e. contact angle of 150° (inset of

Table 6.5 shows images of water droplet in contact with shed surfaces) and a very low contact angle hysteresis (about 2-3°). The water droplet in contact with both of these surfaces corresponds to a Cassie-Baxter state. On the contrary, the water droplet placed on the surface of M-S almost instantaneously spread on the surface resulting in complete wetting. The color of the hairs on M-S surface was also observed to change from shiny white to grey in contact with water, which is a common indication of surface wetting of gecko toes.[123] The wetting of the shed in this case corresponds to Wenzel state. With contact angle measurements, thus, we concluded that for the gecko toe surface morphology, the wettability can vary from superwetting to superhydrphobic, depending upon what the surface chemistry of the structures is. The contrast in the surface wettability can be attributed to surface chemistry differences since surface roughness can 168 be assumed to be the same, compared to the water droplet size used (typically 10-15 µm) to measure contact angle.

Figure 6.7. A schematic representation of a section of unit cell corresponding to gecko toe morphology is shown. A pillar shaped seta has a tier of smaller pillars on top of it.

Unit cell consists of four such setae.

Contact angle measurements imply that the surface to gecko setal hairs is most likely hydrophobic (similar to the surface of F-S). However, because superhydrophobicity of the surface is the result of a combination of surface roughness and surface wettability, there are more than one intrinsic contact angles of material possible which may result in superhydrophobicity of a structured surface. We expected the same to hold true for the gecko toe morphology. The easiest approach to estimate the range of surface intrinsic wettability which can result in superhydrophobic characteristics for the gecko toe surface roughness was to carry out apparent contact angle calculations for model surface, which could enable the predictions of apparent contact angle as a function of intrinsic contact angle, keeping the surface morphology constant. The model

169 calculations were also used for thermodynamic stability analysis of wetting states.[97]

[174, 172] A schematic representation of a section of unit cell is shown in Figure 6.7. The dimensions of the unit cell were calculated using SEM images. Unit cell consists of four setae (tetrad pattern), each with a square cross section. The pillar is 60 µm tall and 4µm wide. The top face of every seta consists of a number of cubes, 0.2 µm in size.

Figure 6.8. ΔG* plotted as a function of R for different θY. The inset of the figure shows magnified plots for θY of 100°, 110°and 120°.

The roughness parameters, which depend on surface morphology, were calculated for the unit cell and incorporated in the Cassie-Baxter and Wenzel equations to predict the apparent contact angles (θCB and θW) respectively (see appendix for calculations).

These contact angle values were used to calculate G*CB and G*W, the thermodynamic free energy corresponding to Cassie-Baxter and Wenzel states respectively. The

170 difference between the two was calculated as ΔG*. The range 70°≤ θY ≤ 120° resulted in

θCB of about 145°-160° that matched the range of the contact angle values we measured experimentally on the surfaces of B-S and F-S. Also, the upper limit for θY without the effect of surface roughness is 120°[34] and thus, any θY above 120°could not be used for model predictions. Figure 6.8 shows ΔG* calculated as a function of roughness, R, for different intrinsic wettability θY. Although the range of R where contact angles could be predicted was very small compared to the actual roughness of model surface as well as actual gecko toe surface, it qualitatively predicted the wettability as a function of θY and meta-stability of Cassie-Baxter state. In the case of this model surface, for θY < 90°, ΔG* was an increasing function of R implying that Wenzel state is thermodynamically more favorable. On the other hand for θY > 90°, ΔG* was a decreasing function of R and becomes negative at a critical R, which is different for different θY (inset of Figure 6.8), implying that Cassie-Baxter is thermodynamically more favorable state for R values more than the critical R. The model predictions were consistent with our experimental observations that the surface of M-S wetted completely in contact with water resulting in the Wenzel wetting state, whereas that of F-S continued to remain dry and retain its superhydrophobicity.

We demonstrated with static contact angle measurement that the water droplet on the surface of B-S exists in the Cassie-Baxter state. In order study if this is in fact a metastable state as predicted by model calculations, we measured the wetting of B-S, M-

S and F-S under water condensation. In these experiments, the samples were kept under

100% humidity. Following the humidity exposure, they were tested for contact angle by depositing a water droplet externally. We observed that the surface of M-S wetted

171 completely under condensation as well. The surface of F-S retained its superhydrophobic characteristics and continued to stay dry. The droplet of water on the surface of B-S however, started penetrating inside the roughness; it was deformed and resulted in wetting of a surface. Thus, water droplet in this case acquired the Wenzel wetting state.

The exposure of a B-S surface to high humidity thus, resulted in its transition from

Cassie-Baxter superhydrophobic state to Wenzel wetting state. Comparing the wetting properties of B-S with that of M-S and F-S, both of which have known inherent wettability, the wettability of B-S surface can be explained on the relative scale of inherent wettability. The B-S surface thus, is inherently hydrophobic enough to form a stable superhydrophobic Cassie-Baxter state, unlike that of M-S which wets instantaneously. However, it is only a metastable state. The transition to the Wenzel state was brought about under water condensation. It is important to note here that the wetting transition is completely reversible; the toe sheds regain their superhydrophobicity on drying. Also, the typical time frame for inducing this transition is of a few days (the sample exposed to high humidity undisturbed for 3-4 days) and there is also a possibility of a reversible changes in surface functional group conformations along with wetting transition. The θY for B-S surface is not known, however, based on the model predictions and the experimental observations, we can predict the range for θY for B-S to be between

70°-90°. This is the range of θY for which the Cassie-Baxter wetting state is possible but it is not thermodynamically the most favorable state of gecko toe model surface.

Humidity experiments showed the possibility of transition from a Cassie-Baxter to Wenzel state. We quantified the transition barrier that separates the two states for all the surfaces in terms of hydrostatic pressure of water column. These measurements were

172 done by introducing the sample in a water column contained in a transparent glass tube, allowing it to equilibrate at a given height for about 5 minutes and observing for its wet versus dry appearance. The M-S sample wets as soon as it is introduced in the water column. On the other hand, in the case of both B-S and F-S we observed that when the sample was introduced in the water column, it appeared shiny silver implying the presence of air plastron layer.[108] Both the surfaces continued to retain the plastron layer as the immersion depth was increased to the maximum column height of 4'. The plastron layer did not disappear even when the sample was held at 4' depth for 7-8 hours.

Thus, the transition barrier between the Cassie-Baxter and Wenzel states of B-S and F-S surfaces can be estimated to be greater than 11.95 kPa, almost three orders of magnitude higher than that of M-S. Moreover, we anticipate it to be higher for F-S surface compared to B-S surface based on the model predictions and condensation experiment results, where B-S surface was observed to transition to Wenzel and no such transition was seen for F-S surface. The differences in the magnitude of transition barriers between Cassie-

Baxter and Wenzel wetting states for all the three surfaces can be attributed to the differences in inherent wettability.

In summary, we used gecko toe shed as our model surface to test the effect of surface chemistry on wettability and wet adhesion of gecko toe pads. We observed that for wet adhesion the most important parameter is chemistry of the contacting surface

(hydrophobicity of OTS-SAM surface resulted in wet adhesion comparable or better than its dry adhesion) rather than the surface chemistry of the hairy shed surface. The underwater adhesion of B-S, M-S and F-S in contact with hydrophilic glass dramatically reduced compared to their dry adhesion. Interestingly, in contact with the OTS-SAM

173 surface underwater we observed that there was no significant decrease in wet adhesion compared to dry adhesion for all three chemistries of hairy shed surfaces, M-S, B-S and

F-S implying a formation of dry contact interfaces underwater.[217, 221] Based on our results it appears that similar surface chemistry of a flat surface versus that of a rough surface do not have the same effect on wet adhesion. The trends we observed for wet versus dry adhesion of shed samples correlate with the similar experiments carried out using whole animal. The case of M-S and OTS-SAM interface was particularly interesting since the surface of M-S wetted as soon as it came in contact with water. We observed that the surfaces of B-S and F-S however, remained dry when subjected to underwater adhesion test. This implied an obvious dependence of surface chemistry on wettability of gecko toe pads however, no effect on the wet adhesion if the toe surface was wet (M-S) or dry (B-S and F-S) underwater in contact with a hydrophobic OTS-

SAM surface.

In addition to reversible adhesion, self cleaning ability is another important function of gecko hairs. Even though we observed that wetting of toe surface is not detrimental for the static adhesion of gecko toe sheds, the self cleaning ability could be affected with the wet toes since it is required for a self-cleaning surface to stay dry in contact with water. The lotus leaf surface is a classic example of self cleaning surface that has its self cleaning ability dependent on its superhydrophobicity.[101, 222] Retention of self cleaning ability in the presence of water would be important for locomotion of animals and would depend on retention of superhydrophobicity of gecko toes. It was thus, required to systematically study the wettability of B-S surface, using M-S and F-S surfaces with known surface chemistry as reference points. We carried out extensive

174 wettability studies using contact angle measurements, wetting properties studied under condensation and hydrostatic transition pressure measurements of B-S and compared to the measurements carried out with M-S and F-S surfaces. Our experiments revealed that the surface of gecko toe hairs is made up of a material with intrinsic contact angle (θY) in the range of 70°-90° which allows it to exhibit stable superhydrophobic state. It is not the most stable state however and the transition to wetted Wenzel state could be brought about. Although, wetting of the B-S surface is possible, the transition pressure needed for it is of the orders of 10 kPa, 3 orders of magnitude higher than M-S (θY ~ 48°) and comparable to F-S (θY ~ 110°). In addition to metastable superhydrophobicity, other phenomena such as surface group rearrangement, changes in mechanical properties of hairs in the presence of water are possibly coupled together and could result in not only wetting of B-S surface but also its failure to adhere to surfaces, as reported before.[123,

213] Further investigation of these factors is required to understand the underwater adhesion of gecko toe pads better. The shed surface can be used as a model surface for studying the gecko adhesion, which definitely is a better control surface for experimental measurements and characterization than testing the whole animal since we demonstrated that the adhesion the results obtained using shed samples correlate with the whole animal studies.

6.3 Conclusion

In this study, we studied the effect of surface chemistry on adhesion and wetting of gecko toe pads. We used the toe skin sheds as the structured surface which has the same hierarchical surface morphology as that of the gecko toes. We used a PECVD process to deposit a thin layer of known surface functionality on the shed surface. A

175 hydrophilic coated shed (M-S) and a hydrophobic coated shed (F-S) were studied for their surface wettability and compared with blank, uncoated shed (B-S). The adhesion of

B-S, M-S and F-S measured on a hydrophilic glass surface and a hydrophobic OTS-SAM surface revealed that the wettability of shed doesn't affect the wet adhesion, the hydrophobicity of the surface it is contacting is in fact more important. Based on wettability model predictions and experimental measurements, we could predict the intrinsic wettability of B-S surface to be in the range of 70°-90°, hydrophobic enough to show a stable superhydrophobic Cassie-Baxter state. The wettability results are also consistent with the hypothesis that the lipids are present on the surface of gecko toe pads.

We estimated the hydrostatic pressure to bring about transition from dry Cassie-Baxter superhydrophobic state to Wenzel wetting state to be more than 11.95 kPa. These findings shed light on some of the very important consequences of intrinsic material properties of hairy structures the gecko toe pads are made up of. The hairy toe pads don't get wet easily which is not particularly important for the animals to stick to surfaces but could otherwise result in the failure of the toe surface to possess self cleaning in contact with water. The adhesive forces generated in humid or water environment may not be the highest that geckos can produce (sheds in contact with a hydrophobic OTS-SAM surface) but are possibly just enough to help them hold on to surfaces and not fail (sheds in contact with hydrophilic glass surface). Surface hydrophobicity of hairs could also be expected to aid in their particle self cleaning ability; a hydrophilic material would not help getting rid of the contaminants easily and could be extremely unfavorable for the animals in their natural habitat. For synthetic mimics of gecko inspired surfaces, the surface roughness is probably the most important design aspect for gecko inspired

176 adhesives even the ones targeted for the use underwater. However, for designing gecko inspired superhydrophobic or self cleaning surfaces, the material chosen should be moderately hydrophobic to not easily result in wetting of the surface which could be a failure mode of application of a given functional surface.

177

CHAPTER VII

SUMMARY AND CONCLUSIONS

In the work presented here, we used plasma enhanced chemical vapor deposition

(PECVD) process for chemical modifications of structured surfaces. PECVD being a versatile vapor phase deposition process allowed to deposit thin coating layers (as small as 5 nm) thus, minimally affecting the surface area and overall roughness of the structured surfaces (even with smallest feature sizes of the order of couple of hundreds of nanometers) before and after PECVD deposition.

We demonstrated that the combination of colloidal lithography and PECVD allows for systematic control over surface roughness as well as surface chemistry enabled us to test the validity of thermodynamic wetting models (both qualitatively and quantitatively) and also made it possible to predict the stability of a given wetting state.

Based on our experimental findings and model calculations, it can be concluded that ,

and are the three variables for single layer and dual layer of HNCP patterns of spherical particles. The regular hexagonal arrangement of patterns and precise control over the variables enabled the predictions of different wetting states possible with single and dual roughness geometry. We also demonstrated a reasonably good correlation between the model predictions and experimentally measured contact angles for Wenzel and Cassie-Baxter states. Additionally, the unit cell model allowed us to predict an

178 existence of Penetrating-Cassie-Baxter wetting state for dual roughness and identify the range of r2 corresponding to the same. A very unique geometry of curved spherical particles, their well-controlled arrangement at both levels of roughness hierarchy and flexibility to precisely alter the surface chemistry, surely makes a combination of colloidal lithography and PECVD an efficient tool to fine tune the surface wettability.

Additionally, it also allows for modelling of the surface to predict not only the apparent contact angles but also the metastability of the wetting states.

The combination of PECVD and nano-porosity of CNT mats was also taken advantage of in creating steamphobic surfaces. We demonstrated that two nano- structured surfaces (CNT and PCNT) which do not differ much structurally and are characterized as superhydrophobic surfaces exhibit extreme steamphobic behavior. The combination of nano-scale roughness of the CNT mats along with the surface energy contrast achieved by deposition of low surface energy plasma deposited coating results in the formation of a highly robust steamphobic surface. The biggest advantage of PECVD process used here for chemical modification is that it is a completely dry process. Unlike wet processes, in which the capillary forces tend to collapse the nano-porosity, the

PECVD process enables deposition of a very thin layer of a film on such a surface without disturbing its structural integrity. Use of stainless steel to form the steamphobic coatings is an added advantage since it is a widely used industrial material. The robust steamphobic surfaces are of particular interest in designing heat exchangers. They are ideal for dropwise steam condensation and can be expected to remarkably improve the exchanger efficiency due to fast droplet removal and low adhesion of condensed steam to

179 the surfaces. Robustness of the structure can also be taken advantage of in anti fog, anti ice and anti dew coatings.

Fibrillar adhesives, categorized as dry and reversible pressure sensitive adhesives, offer numerous advantages over conventional wet or viscoelastic adhesive systems. They are inspired from the adhesion mechanism that geckos used for their locomotion. Gecko inspired adhesion is purely van der Waals based and is expected to be insensitive to surface chemistry. However, this does not hold well for the adhesion taking place in the presence of water. The fact that geckos can retain their adhesion in the presence of water

(humid weather and tropical rainforests) which most of the adhesives fail to do, raises a need to understand the effect of wettability on the adhesion ability of the animals. In order to systematically understand the wet adhesion mechanism, we first tested the effect that contact surface wettability has on the adhesion. We observed that as the hydrophobicity of the contact surface increases, the wet adhesion is significantly higher than the corresponding dry adhesion. These tests were performed at whole animal scale.

The experimental results were consistent with the predicted adhesion model calculations.

In order to study the effect of shed wettability, the biggest challenge was the ambiguity of surface chemistry of native gecko toe hairs. We took an advantage of PECVD process that is helps in retaining the surface structure as closely as that of the native surface, to deposit thin layer of coatings with known surface chemistry. A hydrophilic coating

(maleic anhydride) and a hydrophobic coating (1H,1H,2H-perfluoro-1-dodecene) were deposited on the gecko shed surface (sheds naturally molted by geckos) and the wetting and adhesion properties were tested and compared to the uncoated shed results. Firstly, the wetting and adhesion properties of blank, uncoated sheds were observed to be

180 consistent with whole animal data. The wettability of PECVD coated sheds and that of blank sheds was distinctly different, where hydrophilic coated sheds result in spontaneous wetting upon contact with water, whereas hydrophobic coated sheds and blank sheds are superhydrophobic surfaces. In spite of clear differences in surface wettability, the wet adhesion did not get affected significantly for any surface chemistry of sheds as long as the contact surface was hydrophobic (OTS-SAM coated glass plate). This was surprising, yet points to a very important criterion i.e. surface structure is more important for wet adhesion of the structured surface as opposed to the surface chemistry. The flexibility that

PECVD offers can be paired with different techniques used to create structured surfaces, for mimicking fibrillar adhesives. We showed that the surface chemistry of such an adhesive system can be easily tuned using PECVD, depending upon if it is required for a structure to wet or remain dry in the applications involving prolonged exposures to water.

Thus, with the help of examples of structured surfaces (colloidal lithography patterns, CNT mats and gecko shed surface) we demonstrated that the tunability of their wetting and adhesion properties could be achieved using PECVD process. PECVD, thus is surely one of the promising tool that can be used for achieving surface chemistry control for optimizing not only wetting and adhesion but also other functional properties which structured surfaces offer.

181

BIBLIOGRAPHY

[1] D. Öner and T. J. McCarthy, “Ultrahydrophobic surfaces: effects of topography length scales on wettability,” Langmuir, vol. 16, no. 20, pp. 7777–7782, 2000.;.

[2] Z. Yoshimitsu, A. Nakajima, T. Watanabe, and K. Hashimoto, “Effects of surface structure on the hydrophobicity and sliding behavior of water droplets,” Langmuir, vol. 18, no. 15, pp. 5818–5822, 2002.;.

[3] L. Zhu, Y. Feng, X. Ye, and Z. Zhou, “Tuning wettability and getting superhydrophobic surface by controlling surface roughness with well-designed microstructures,” Sensor. Actuat. A: Phys., vol. 130, pp. 595–600, 2006.;.

[4] C.-H. Choi and C.-J. Kim, “Fabrication of a dense array of tall nanostructures over a large sample area with sidewall profile and tip sharpness control,” Nanotechnology, vol. 17, no. 21, p. 5326, 2006.;.

[5] R. M. Wagterveld, C. W. Berendsen, S. Bouaidat, and J. Jonsmann, “Ultralow hysteresis superhydrophobic surfaces by excimer laser modification of SU-8,” Langmuir, vol. 22, no. 26, pp. 10904–10908, 2006.;.

[6] N. J. Shirtcliffe, S. Aqil, C. Evans, G. McHale, M. I. Newton, C. C. Perry, and P. Roach, “The use of high aspect ratio photoresist (SU-8) for super-hydrophobic pattern prototyping,” J. Micromech. Microeng., vol. 14, no. 10, p. 1384, 2004.;

[7] L. Cao, H.-H. Hu, and D. Gao, “Design and fabrication of micro-textures for inducing a superhydrophobic behavior on hydrophilic materials,” Langmuir, vol. 23, no. 8, pp. 4310–4314, 2007.;.

[8] M. Ma, M. Gupta, Z. Li, L. Zhai, K. K. Gleason, R. E. Cohen, M. F. Rubner, and G. C. Rutledge, “Decorated electrospun fibers exhibiting superhydrophobicity,” Adv. Mater., vol. 19, no. 2, pp. 255–259, 2007.;.

[9] L. Zou, X. Xiang, J. Fan, and F. Li, “Single-source precursor to complex metal oxide monoliths with tunable microstructures and properties: the case of Mg-containing materials,” Chem. Mater., vol. 19, no. 26, pp. 6518–6527, 2007.;.

182

[10] Z. Huang, Y. Zhu, J. Zhang, and G. Yin, “Stable biomimetic superhydrophobicity and magnetization film with Cu-ferrite nanorods,” J. Phys. Chem. C, vol. 111, no. 18, pp. 6821–6825, 2007.;.

[11] Y. Li, W.-Z. Jia, Y.-Y. Song, and X.-H. Xia, “Superhydrophobicity of 3D porous copper films prepared using the hydrogen bubble dynamic template,” Chem. Mater., vol. 19, no. 23, pp. 5758–5764, 2007.;.

[12] Q. Pan, H. Jin, and H. Wang, “Fabrication of superhydrophobic surfaces on interconnected Cu(OH)2 nanowires via solution-immersion,” Nanotechnology, vol. 18, no. 35, p. 355605, 2007.;.

[13] J. A. Franco, S. E. Kentish, J. M. Perera, and G. W. Stevens, “Fabrication of a superhydrophobic polypropylene membrane by deposition of a porous crystalline polypropylene coating,” J. Membr. Sci., vol. 318, no. 1, pp. 107–113, 2008.;.

[14] S. Michielsen and H. J. Lee, “Design of a superhydrophobic surface using woven structures,” Langmuir, vol. 23, no. 11, pp. 6004–6010, 2007.;.

[15] J. Zheng, A. He, J. Li, J. Xu, and C. C. Han, “Studies on the controlled morphology and wettability of polystyrene surfaces by electrospinning or electrospraying,” Polymer, vol. 47, no. 20, pp. 7095–7102, 2006.;.

[16] E. Hosono, S. Fujihara, I. Honma, and H. Zhou, “Superhydrophobic perpendicular nanopin film by the bottom-up process,” J. Am. Chem. Soc., vol. 127, no. 39, pp. 13458– 13459, 2005.;.

[17] C.-H. Chen, Q. Cai, C. Tsai, C.-L. Chen, G. Xiong, Y. Yu, and Z. Ren, “Dropwise condensation on superhydrophobic surfaces with two-tier roughness,” Appl. Phys. Lett., vol. 90, no. 17, pp. 173108–173108, 2007.;.

[18] Y. Li, X. J. Huang, S. H. Heo, C. C. Li, Y. K. Choi, W. P. Cai, and S. O. Cho, “Superhydrophobic bionic surfaces with hierarchical microsphere/SWCNT composite arrays,” Langmuir, vol. 23, no. 4, pp. 2169–2174, 2007.;.

[19] H.-J. Tsai and Y.-L. Lee, “Facile method to fabricate raspberry-like particulate films for superhydrophobic surfaces,” Langmuir, vol. 23, no. 25, pp. 12687–12692, 2007.;.

[20] A. Tuteja, W. Choi, J. M. Mabry, G. H. McKinley, and R. E. Cohen, “Robust omniphobic surfaces,” PNAS, vol. 105, no. 47, pp. 18200–18205, 2008.;.

[21] T.-S. Wong, S. H. Kang, S. K. Tang, E. J. Smythe, B. D. Hatton, A. Grinthal, and J. Aizenberg, “Bioinspired self-repairing slippery surfaces with pressure-stable omniphobicity,” Nature, vol. 477, no. 7365, pp. 443–447, 2011.;.

183

[22] X. Deng, L. Mammen, H.-J. Butt, and D. Vollmer, “Candle soot as a template for a transparent robust superamphiphobic coating,” Science, vol. 335, no. 6064, pp. 67–70, 2012.;.

[23] M. Nosonovsky and B. Bhushan, “Biomimetic superhydrophobic surfaces: multiscale approach,” Nano Lett., vol. 7, no. 9, pp. 2633–2637, 2007.;.

[24] J. N. Israelachvili, Intermolecular and surface forces: revised third edition. Academic press, 2011.;.

[25] W. Hansen and K. Autumn, “Evidence for self-cleaning in gecko setae,” PNAS, vol. 102, no. 2, pp. 385–389, 2005.;.

[26] N. Rizzo, K. Gardner, D. Walls, N. Keiper-Hrynko, T. Ganzke, and D. Hallahan, “Characterization of the structure and composition of gecko adhesive setae,” J. R. Soc. Interface, vol. 3, no. 8, pp. 441–451, 2006.;.

[27] H. Yasuda and T. Hsu, “Plasma polymerization investigated by the comparison of hydrocarbons and perfluorocarbons,” Surf. Sci., vol. 76, no. 1, pp. 232–241, 1978.;.

[28] L. Gao and T. J. McCarthy, “Contact angle hysteresis explained,” Langmuir, vol. 22, no. 14, pp. 6234–6237, 2006.;.

[29] C. Extrand and Y. Kumagai, “An experimental study of contact angle hysteresis,” J. Colloid Interf. Sci., vol. 191, no. 2, pp. 378–383, 1997.;.

[30] R. E. Johnson Jr and R. H. Dettre, “Contact angle hysteresis. iii. study of an idealized heterogeneous surface,” J. Phys. Chem., vol. 68, no. 7, pp. 1744–1750, 1964.;.

[31] A. Marmur, “Contact angle hysteresis on heterogeneous smooth surfaces,” J. Colloid Interf. Sci., vol. 168, no. 1, pp. 40–46, 1994.;.

[32] G. McHale, N. Shirtcliffe, and M. Newton, “Contact-angle hysteresis on superhydrophobic surfaces,” Langmuir, vol. 20, no. 23, pp. 10146–10149, 2004.;.

[33] C. Extrand, “Model for contact angles and hysteresis on rough and ultraphobic surfaces,” Langmuir, vol. 18, no. 21, pp. 7991–7999, 2002.;.

[34] T. Nishino, M. Meguro, K. Nakamae, M. Matsushita, and Y. Ueda, “The lowest surface free energy based on-CF3 alignment,” Langmuir, vol. 15, no. 13, pp. 4321–4323, 1999.;.

[35] H. J. Ensikat, P. Ditsche-Kuru, C. Neinhuis, and W. Barthlott, “Superhydrophobicity in perfection: the outstanding properties of the lotus leaf,” Beilstein J. Nanotechnol., vol. 2, no. 1, pp. 152–161, 2011.;.

184

[36] Y.-T. Cheng and D. E. Rodak, “Is the lotus leaf superhydrophobic?” Appl. Phys. Lett., vol. 86, no. 14, pp. 144101–144101, 2005.;.

[37] R. N. Wenzel, “Resistance of solid surfaces to wetting by water,” Ind. Eng. Chem., vol. 28, no. 8, pp. 988–994, 1936.;.

[38] A. Cassie and S. Baxter, “Wettability of porous surfaces,” Trans. Faraday Soc., vol. 40, pp. 546–551, 1944.;.

[39] K. Kurogi, H. Yan, and K. Tsujii, “Importance of pinning effect of wetting in super water-repellent surfaces,” Colloids and Surfaces A: Physicochem. Eng. Aspects, vol. 317, no. 1, pp. 592–597, 2008.;.

[40] W. Choi, A. Tuteja, J. M. Mabry, R. E. Cohen, G. H. McKinley et al., “A modified cassie-baxter relationship to explain contact angle hysteresis and anisotropy on non-wetting textured surfaces,” J. Colloid Interf. Sci., vol. 339, no. 1, pp. 208–216, 2009.;.

[41] A. Marmur, “From hygrophilic to superhygrophobic: theoretical conditions for making high-contact-angle surfaces from low-contact-angle materials,” Langmuir, vol. 24, no. 14, pp. 7573–7579, 2008.;.

[42] Z. Guo, W. Liu, B.-L. Su et al., “Superhydrophobic surfaces: from natural to biomimetic to functional.” J. Colloid Interf. Sci., vol. 353, no. 2, p. 335, 2011.;.

[43] K. Liu, X. Yao, and L. Jiang, “Recent developments in bio-inspired special wettability,” Chemical Society Reviews, vol. 39, no. 8, pp. 3240–3255, 2010.;.

[44] G. McHale, N. Shirtcliffe, and M. Newton, “Super-hydrophobic and superwetting surfaces: Analytical potential?” Analyst, vol. 129, no. 4, pp. 284–287, 2004.;.

[45] D. Quéré, “Non-sticking drops,” Rep. Prog. Phys., vol. 68, no. 11, p. 2495, 2005.;.

[46] M. Jin, X. Feng, L. Feng, T. Sun, J. Zhai, T. Li, and L. Jiang, “Superhydrophobic aligned polystyrene nanotube films with high adhesive force,” Adv. Mater., vol. 17, no. 16, pp. 1977–1981, 2005.;.

[47] L. Gao and T. J. McCarthy, “The “lotus effect” explained: two reasons why two length scales of topography are important,” Langmuir, vol. 22, no. 7, pp. 2966–2967, 2006.;.

[48] X. Wu, L. Zheng, and D. Wu, “Fabrication of superhydrophobic surfaces from microstructured ZnO-based surfaces via a wet-chemical route,” Langmuir, vol. 21, no. 7, pp. 2665–2667, 2005.;.

185

[49] B. Xu and Z. Cai, “Fabrication of a superhydrophobic ZnO nanorod array film on cotton fabrics via a wet chemical route and hydrophobic modification,” Appl. Surf. Sci., vol. 254, no. 18, pp. 5899–5904, 2008.;.

[50] Z. Guo, J. Fang, L. Wang, and W. Liu, “Fabrication of superhydrophobic copper by wet chemical reaction,” Thin Solid Films, vol. 515, no. 18, pp. 7190–7194, 2007.;.

[51] M. Li, J. Xu, and Q. Lu, “Creating superhydrophobic surfaces with flowery structures on nickel substrates through a wet-chemical-process,” J. Mater. Chem., vol. 17, no. 45, pp. 4772–4776, 2007.;.

[52] X. Zhang, F. Shi, X. Yu, H. Liu, Y. Fu, Z. Wang, L. Jiang, and X. Li, “Polyelectrolyte multilayer as matrix for electrochemical deposition of gold clusters: toward super-hydrophobic surface,” J. Am. Chem. Soc., vol. 126, no. 10, pp. 3064–3065, 2004.;.

[53] M. Li, J. Zhai, H. Liu, Y. Song, L. Jiang, and D. Zhu, “Electrochemical deposition of conductive superhydrophobic zinc oxide thin films,” J. Phys. Chem. B, vol. 107, no. 37, pp. 9954–9957, 2003.;.

[54] N. Zhao, F. Shi, Z. Wang, and X. Zhang, “Combining layer-by-layer assembly with electrodeposition of silver aggregates for fabricating superhydrophobic surfaces,” Langmuir, vol. 21, no. 10, pp. 4713–4716, 2005.;.

[55] F. Shi, Z. Wang, and X. Zhang, “Combining a layer-by-layer assembling technique with electrochemical deposition of gold aggregates to mimic the legs of water striders,” Adv. Mater., vol. 17, no. 8, pp. 1005–1009, 2005.;.

[56] L. Gao and T. J. McCarthy, “A perfectly hydrophobic surface (θa/θr= 180°/180°),” J. Am. Chem. Soc., vol. 128, no. 28, pp. 9052–9053, 2006.;.

[57] J. Genzer and K. Efimenko, “Creating long-lived superhydrophobic polymer surfaces through mechanically assembled monolayers,” Science, vol. 290, no. 5499, pp. 2130–2133, 2000.;.

[58] X. Song, J. Zhai, Y. Wang, and L. Jiang, “Fabrication of superhydrophobic surfaces by self-assembly and their water-adhesion properties,” J. Phys. Chem. B, vol. 109, no. 9, pp. 4048–4052, 2005.;.

[59] L. Zhai, M. C. Berg, F. C. Cebeci, Y. Kim, J. M. Milwid, M. F. Rubner, and R. E. Cohen, “Patterned superhydrophobic surfaces: toward a synthetic mimic of the namib desert beetle,” Nano Lett., vol. 6, no. 6, pp. 1213–1217, 2006.;.

[60] X. Gao, X. Yan, X. Yao, L. Xu, K. Zhang, J. Zhang, B. Yang, and L. Jiang, “The dry-style antifogging properties of mosquito compound eyes and artificial analogues prepared by soft lithography,” Adv. Mater., vol. 19, no. 17, pp. 2213–2217, 2007.;.

186

[61] B. Liu, Y. He, Y. Fan, and X. Wang, “Fabricating super-hydrophobic lotus-leaf- like surfaces through soft-lithographic imprinting,” Macromol. Rapid Commun., vol. 27, no. 21, pp. 1859–1864, 2006.;.

[62] A. Pozzato, S. D. Zilio, G. Fois, D. Vendramin, G. Mistura, M. Belotti, Y. Chen, and M. Natali, “Superhydrophobic surfaces fabricated by nanoimprint lithography,” Microelectron. Eng., vol. 83, no. 4, pp. 884–888, 2006.;.

[63] G. Zhang, J. Zhang, G. Xie, Z. Liu, and H. Shao, “Cicada wings: a stamp from nature for nanoimprint lithography,” Small, vol. 2, no. 12, pp. 1440–1443, 2006.;.

[64] X.-M. Li, D. Reinhoudt, and M. Crego-Calama, “What do we need for a superhydrophobic surface? a review on the recent progress in the preparation of superhydrophobic surfaces,” Chem. Soc. Rev., vol. 36, no. 8, pp. 1350–1368, 2007.;.

[65] H. Notsu, W. Kubo, I. Shitanda, and T. Tatsuma, “Super-hydrophobic/superhydrophilic patterning of gold surfaces by photocatalytic lithography,” J. Mater. Chem., vol. 15, no. 15, pp. 1523–1527, 2005.;.

[66] Y. Lai, C. Lin, H. Wang, J. Huang, H. Zhuang, and L. Sun, “Superhydrophilic– superhydrophobic micropattern on TiO2 nanotube films by photocatalytic lithography,” Electrochem. Commun., vol. 10, no. 3, pp. 387–391, 2008.;.

[67] G. Zhang, D. Wang, Z.-Z. Gu, and H. Möhwald, “Fabrication of superhydrophobic surfaces from binary colloidal assembly,” Langmuir, vol. 21, no. 20, pp. 9143–9148, 2005.;.

[68] Y. Li, J. Zhang, S. Zhu, H. Dong, F. Jia, Z. Wang, Y. Tang, L. Zhang, S. Zhang, and B. Yang, “Bioinspired silica surfaces with near-infrared improved transmittance and superhydrophobicity by colloidal lithography,” Langmuir, vol. 26, no. 12, pp. 9842– 9847, 2010.;.

[69] D. Spori, T. Drobek, S. Zurcher, and N. Spencer, “Cassie-state wetting investigated by means of a hole-to-pillar density gradient,” Langmuir, vol. 26, no. 12, pp. 9465–9473, 2010.;.

[70] K. K. Lau, J. Bico, K. B. Teo, M. Chhowalla, G. A. Amaratunga, W. I. Milne, G. H. McKinley, and K. K. Gleason, “Superhydrophobic carbon nanotube forests,” Nano Lett., vol. 3, no. 12, pp. 1701–1705, 2003.;.

[71] S. Sethi and A. Dhinojwala, “Superhydrophobic conductive carbon nanotube coatings for steel,” Langmuir, vol. 25, no. 8, pp. 4311–4313, 2009.;.

[72] B. Balu, V. Breedveld, and D. W. Hess, “Fabrication of “roll-off” and “sticky” superhydrophobic cellulose surfaces via plasma processing,” Langmuir, vol. 24, no. 9, pp. 4785–4790, 2008.;.

187

[73] K. Tadanaga, J. Morinaga, A. Matsuda, and T. Minami, “Superhydrophobic- superhydrophilic micropatterning on flowerlike alumina coating film by the sol-gel method,” Chem. Mater., vol. 12, no. 3, pp. 590–592, 2000.;.

[74] K. Tadanaga, K. Kitamuro, A. Matsuda, and T. Minami, “Formation of superhydrophobic alumina coating films with high transparency on polymer substrates by the sol-gel method,” J. Sol-Gel Sci. Technol., vol. 26, no. 1-3, pp. 705–708, 2003.;.

[75] N. J. Shirtcliffe, G. McHale, M. I. Newton, and C. C. Perry, “Intrinsically superhydrophobic organosilica sol-gel foams,” Langmuir, vol. 19, no. 14, pp. 5626–5631, 2003.;.

[76] M. Manca, A. Cannavale, L. De Marco, A. S. Arico, R. Cingolani, and G. Gigli, “Durable superhydrophobic and antireflective surfaces by trimethylsilanized silica nanoparticles-based sol- gel processing,” Langmuir, vol. 25, no. 11, pp. 6357–6362, 2009.;.

[77] C. Dorrer and J. Rühe, “Wetting of silicon nanograss: from superhydrophilic to superhydrophobic surfaces,” Adv. Mater., vol. 20, no. 1, pp. 159–163, 2008.;.

[78] B. Qian and Z. Shen, “Fabrication of superhydrophobic surfaces by dislocation- selective chemical etching on aluminum, copper, and zinc substrates,” Langmuir, vol. 21, no. 20, pp. 9007–9009, 2005.;.

[79] F. Shi, X. Chen, L. Wang, J. Niu, J. Yu, Z. Wang, and X. Zhang, “Roselike microstructures formed by direct in situ hydrothermal synthesis: from superhydrophilicity to superhydrophobicity,” Chem. Mater., vol. 17, no. 24, pp. 6177–6180, 2005.;.

[80] X. Liu and J. He, “One-step hydrothermal creation of hierarchical microstructures toward superhydrophilic and superhydrophobic surfaces,” Langmuir, vol. 25, no. 19, pp. 11822–11826, 2009.;.

[81] B. Li, M. Zhou, R. Yuan, and L. Cai, “Fabrication of titanium-based microstructured surfaces and study on their superhydrophobic stability,” J. Mater. Res, vol. 23, no. 2, pp. 491–2, 2008.;.

[82] H. Y. Erbil, A. L. Demirel, Y. Avc, and O. Mert, “Transformation of a simple plastic into a superhydrophobic surface,” Science, vol. 299, no. 5611, pp. 1377–1380, 2003.;.

[83] H. Chen, Z. Yuan, J. Zhang, Y. Liu, K. Li, D. Zhao, S. Li, P. Shi, and J. Tang, “Preparation, characterization and wettability of porous superhydrophobic poly (vinyl chloride) surface,” J. Porous Mater., vol. 16, no. 4, pp. 447–451, 2009.;.

188

[84] K. Acatay, E. Simsek, C. Ow-Yang, and Y. Z. Menceloglu, “Tunable, superhydrophobically stable polymeric surfaces by electrospinning,” Angew. Chem. Int. Ed., vol. 43, no. 39, pp. 5210–5213, 2004.;.

[85] M. Ma, Y. Mao, M. Gupta, K. K. Gleason, and G. C. Rutledge, “Superhydrophobic fabrics produced by electrospinning and chemical vapor deposition,” Macromolecules, vol. 38, no. 23, pp. 9742–9748, 2005.;.

[86] M. Ma, R. M. Hill, J. L. Lowery, S. V. Fridrikh, and G. C. Rutledge, “Electrospun poly (styrene-block-dimethylsiloxane) block copolymer fibers exhibiting superhydrophobicity,” Langmuir, vol. 21, no. 12, pp. 5549–5554, 2005.;.

[87] X. Wang, B. Ding, J. Yu, and M. Wang, “Engineering biomimetic superhydrophobic surfaces of electrospun nanomaterials,” Nano Today, vol. 6, no. 5, pp. 510–530, 2011.;.

[88] J. Drelich, E. Chibowski, D. D. Meng, and K. Terpilowski, “Hydrophilic and superhydrophilic surfaces and materials,” Soft Matter, vol. 7, no. 21, pp. 9804–9828, 2011.;.

[89] X. Liu and J. He, “Hierarchically structured superhydrophilic coatings fabricated by self-assembling raspberry-like silica nanospheres,” J. Colloid Interf. Sci., vol. 314, no. 1, pp. 341–345, 2007.;.

[90] H. Dong, P. Ye, M. Zhong, J. Pietrasik, R. Drumright, and K. Matyjaszewski, “Superhydrophilic surfaces via polymer- SiO2 nanocomposites,” Langmuir, vol. 26, no. 19, pp. 15567–15573, 2010.;.

[91] X. Zhou, X. Guo, W. Ding, and Y. Chen, “Superhydrophobic or superhydrophilic surfaces regulated by micro-nano structured zno powders,” Appl. Surf. Sci., vol. 255, no. 5, pp. 3371–3374, 2008.;.

[92] Y. Jiang, Z. Wang, X. Yu, F. Shi, H. Xu, X. Zhang, M. Smet, and W. Dehaen, “Self-assembled monolayers of dendron thiols for electrodeposition of gold nanostructures: toward fabrication of superhydrophobic/superhydrophilic surfaces and ph-responsive surfaces,” Langmuir, vol. 21, no. 5, pp. 1986–1990, 2005.;.

[93] T. Sun, G. Wang, L. Feng, B. Liu, Y. Ma, L. Jiang, and D. Zhu, “Reversible switching between superhydrophilicity and superhydrophobicity,” Angew. Chem. Int. Ed., vol. 43, no. 3, pp. 357–360, 2004.;.

[94] C. Extrand, “Criteria for ultralyophobic surfaces,” Langmuir, vol. 20, no. 12, pp. 5013–5018, 2004.;.

189

[95] A. Tuteja, W. Choi, M. Ma, J. M. Mabry, S. A. Mazzella, G. C. Rutledge, G. H. McKinley, and R. E. Cohen, “Designing superoleophobic surfaces,” Science, vol. 318, no. 5856, pp. 1618–1622, 2007.;.

[96] A. Marmur, “The lotus effect: superhydrophobicity and metastability,” Langmuir, vol. 20, no. 9, pp. 3517–3519, 2004.;.

[97] A. Marmur, “Wetting on hydrophobic rough surfaces: to be heterogeneous or not to be?” Langmuir, vol. 19, no. 20, pp. 8343–8348, 2003.;.

[98] A. Lafuma and D. Quéré, “Superhydrophobic states,” Nature Materials, vol. 2, no. 7, pp. 457–460, 2003.;.

[99] N. A. Patankar, “Transition between superhydrophobic states on rough surfaces,” Langmuir, vol. 20, no. 17, pp. 7097–7102, 2004.;.

[100] T. Liu, W. Sun, X. Sun, and H. Ai, “Thermodynamic analysis of the effect of the hierarchical architecture of a superhydrophobic surface on a condensed drop state,” Langmuir, vol. 26, no. 18, pp. 14835–14841, 2010.;.

[101] R. Fürstner, W. Barthlott, C. Neinhuis, and P. Walzel, “Wetting and self-cleaning properties of artificial superhydrophobic surfaces,” Langmuir, vol. 21, no. 3, pp. 956– 961, 2005.;.

[102] M. K. Chaudhury and G. M. Whitesides, “How to make water run uphill,” DTIC Document, Tech. Rep., 1992.;.

[103] A. D. Nikolov, D. T. Wasan, A. Chengara, K. Koczo, G. A. Policello, and I. Kolossvary, “Superspreading driven by marangoni flow,” Advances in colloid and interface science, vol. 96, no. 1, pp. 325–338, 2002.;.

[104] S. Daniel, M. K. Chaudhury, and J. C. Chen, “Fast drop movements resulting from the phase change on a gradient surface,” Science, vol. 291, no. 5504, pp. 633–636, 2001.;.

[105] J. B. Boreyko and C.-H. Chen, “Restoring superhydrophobicity of lotus leaves with vibration-induced dewetting,” Phys. Rev. Lett., vol. 103, no. 17, p. 174502, 2009.;.

[106] C. Dietz, K. Rykaczewski, A. Fedorov, and Y. Joshi, “Visualization of droplet departure on a superhydrophobic surface and implications to heat transfer enhancement during dropwise condensation,” Appl. Phys. Lett., vol. 97, no. 3, pp. 033104–033104, 2010.;.

[107] J. B. Boreyko and C.-H. Chen, “Self-propelled dropwise condensate on superhydrophobic surfaces,” Phys. Rev. Lett., vol. 103, no. 18, p. 184501, 2009.;.

190

[108] R. Poetes, K. Holtzmann, K. Franze, and U. Steiner, “Metastable underwater superhydrophobicity,” Phys. Rev. Lett., vol. 105, no. 16, p. 166104, 2010.;.

[109] A. Marmur, “Underwater superhydrophobicity: theoretical feasibility,” Langmuir, vol. 22, no. 4, pp. 1400–1402, 2006.;.

[110] T. Verho, C. Bower, P. Andrew, S. Franssila, O. Ikkala, and R. H. Ras, “Mechanically durable superhydrophobic surfaces,” Adv. Mater., vol. 23, no. 5, pp. 673– 678, 2011.;.

[111] H. Margenau, “Van der waals forces,” Rev. Mod. Phys., vol. 11, pp. 1–35, Jan 1939. [Online]. Available: http://link.aps.org/doi/10.1103/RevModPhys.11.1 ;

[112] H. Hamaker, “The london—van der waals attraction between spherical particles,” Physica, vol. 4, no. 10, pp. 1058–1072, 1937.;.

[113] A. Paiva, N. Sheller, M. D. Foster, A. J. Crosby, and K. R. Shull, “Study of the surface adhesion of pressure-sensitive adhesives by atomic force microscopy and spherical indenter tests,” Macromolecules, vol. 33, no. 5, pp. 1878–1881, 2000.;.

[114] H.-J. Kim and H. Mizumachi, “Miscibility and peel strength of acrylic pressure- sensitive adhesives: Acrylic copolymer–tackifier resin systems,” J. Appl. Polym. Sci., vol. 56, no. 2, pp. 201–209, 1995.;.

[115] A. Gent and R. Petrich, “Adhesion of viscoelastic materials to rigid substrates,” Proc. R. Soc. London, Ser. A, vol. 310, no. 1502, pp. 433–448, 1969.;.

[116] L. Alibardi, “Ultrastructural autoradiographic and immunocytochemical analysis of setae formation and keratinization in the digital pads of the gecko Hemidactylus turcicus (Gekkonidae, reptilia),” Tissue & Cell, vol. 35, no. 4, pp. 288–296, 2003.;.

[117] P. Y. Hsu, L. Ge, X. Li, A. Y. Stark, C. Wesdemiotis, P. H. Niewiarowski, and A. Dhinojwala, “Direct evidence of phospholipids in gecko footprints and spatula– substrate contact interface detected using surface-sensitive spectroscopy,” J. R. Soc. Interface, vol. 9, no. 69, pp. 657–664, 2012.;.

[118] K. Autumn and A. M. Peattie, “Mechanisms of adhesion in geckos,” Integr. Comp. Biol., vol. 42, no. 6, pp. 1081–1090, 2002.;.

[119] K. Autumn, M. Sitti, Y. A. Liang, A. M. Peattie, W. R. Hansen, S. Sponberg, T. W. Kenny, R. Fearing, J. N. Israelachvili, and R. J. Full, “Evidence for van der waals adhesion in gecko setae,” PNAS, vol. 99, no. 19, pp. 12252–12256, 2002.;.

[120] G. Huber, H. Mantz, R. Spolenak, K. Mecke, K. Jacobs, S. N. Gorb, and E. Arzt, “Evidence for capillarity contributions to gecko adhesion from single spatula nanomechanical measurements,” PNAS, vol. 102, no. 45, pp. 16293–16296, 2005.;.

191

[121] P. H. Niewiarowski, S. Lopez, L. Ge, E. Hagan, and A. Dhinojwala, “Sticky gecko feet: the role of temperature and humidity,” PLoS One, vol. 3, no. 5, p. e2192, 2008.;.

[122] K. Autumn and W. Hansen, “Ultrahydrophobicity indicates a non-adhesive default state in gecko setae,” J. Comp. Physiol., A, vol. 192, no. 11, pp. 1205–1212, 2006.;.

[123] A. Y. Stark, T. W. Sullivan, and P. H. Niewiarowski, “The effect of surface water and wetting on gecko adhesion,” J. Exp. Biol., vol. 215, no. 17, pp. 3080–3086, 2012.;.

[124] H. Yasuda, “Plasma polymerization for protective coatings and composite membranes,” J. Membrane Sci., vol. 18, pp. 273–284, 1984.;.

[125] H. Yasuda, M. Bumgarner, H. Marsh, B. Yamanashi, D. Devito, M. Wolbarsht, J. Reed, M. Bessler, M. Landers, D. Hercules et al., “Ultrathin coating by plasma polymerization applied to corneal contact lens,” J. Biomed. Mater. Res., vol. 9, no. 6, pp. 629–643, 1975.;.

[126] H. Yasuda and M. Gazicki, “Biomedical applications of plasma polymerization and plasma treatment of polymer surfaces,” Biomaterials, vol. 3, no. 2, pp. 68–77, 1982.;.

[127] H. Yasuda and C. Lamaze, “Preparation of reverse osmosis membranes by plasma polymerization of organic compounds,” J. Appl. Polym. Sci., vol. 17, no. 1, pp. 201–222, 1973.;.

[128] H. Yasuda, Plasma polymerization. Access Online via Elsevier, 1985.;.

[129] Y. Osada, Plasma polymerization processes. Elsevier Science Ltd, 1992, vol. 3.;.

[130] H. Kobayashi, A. Bell, and M. Shen, “Plasma polymerization of saturated and unsaturated hydrocarbons,” Macromolecules, vol. 7, no. 3, pp. 277–283, 1974.;.

[131] H. Yasuda and T. Hsu, “Some aspects of plasma polymerization of fluorine- containing organic compounds,” J. Polym. Sci., Part A: Polym. Chem., vol. 15, no. 10, pp. 2411–2425, 1977.;.

[132] H. Yasuda, T. Hsu, E. Brandt, and C. Reilley, “Some aspects of plasma polymerization of fluorine-containing organic compounds. ii. comparison of ethylene and tetrafluoroethylene,” J. Polym. Sci., Part A: Polym. Chem., vol. 16, no. 2, pp. 415–425, 1978.;.

[133] H. Yasuda and T. Hirotsu, “Critical evaluation of conditions of plasma polymerization,” J. Polym. Sci., Part A: Polym. Chem., vol. 16, no. 4, pp. 743–759, 1978.;.

192

[134] T. Masuoka and H. Yasuda, “Plasma polymerization of hexafluoroethane,” J. Polym. Sci., Part A: Polym. Chem., vol. 20, no. 9, pp. 2633–2642, 1982.;.

[135] K. Nakajima, A. Bell, M. Shen, and M. Millard, “Plasma polymerization of tetrafluoroethylene,” J. Appl. Polym. Sci., vol. 23, no. 9, pp. 2627–2637, 1979.;.

[136] H. Biederman, S. Ojha, and L. Holland, “The properties of fluorocarbon films prepared by rf sputtering and plasma polymerization in inert and active gas,” Thin Solid Films, vol. 41, no. 3, pp. 329–339, 1977.;.

[137] Y. Matsumoto and M. Ishida, “The property of plasma-polymerized fluorocarbon film in relation to CH4/C4F8 ratio and substrate temperature,” Sensor. Actuat. A - Phys., vol. 83, no. 1, pp. 179–185, 2000.;.

[138] D. A. Brevnov, M. J. Barela, M. J. Brooks, G. P. López, and P. B. Atanassov, “Fabrication of anisotropic super hydrophobic/hydrophilic nanoporous membranes by plasma polymerization of C4F8 on anodic aluminum oxide,” J. Electrochem. Soc., vol. 151, no. 8, pp. B484–B489, 2004.;.

[139] H. Andersson, W. van der Wijngaart, P. Griss, F. Niklaus, and G. Stemme, “Hydrophobic valves of plasma deposited octafluorocyclobutane in drie channels,” Sensor. Actuat. B, Chem., vol. 75, no. 1, pp. 136–141, 2001.;.

[140] S. A. Visser, C. E. Hewitt, J. Fornalik, G. Braunstein, C. Srividya, and S. Babu, “Surface and bulk compositional characterization of plasma-polymerized fluorocarbons prepared from hexafluoroethane and acetylene or butadiene reactant gases,” J. Appl. Polym. Sci., vol. 66, no. 3, pp. 409–421, 1997.;.

[141] C. R. Savage, R. B. Timmons, and J. W. Lin, “Molecular control of surface film compositions via pulsed radio-frequency plasma deposition of perfluoropropylene oxide,” Chem. Mater., vol. 3, no. 4, pp. 575–577, 1991.;.

[142] A. Tasaka, A. Komura, Y. Uchimoto, M. Inaba, and Z. Ogumi, “Preparation of functionally gradient fluorocarbon polymer films by plasma polymerization of NF3 and propylene,” J. Polym. Sci., Part A: Polym. Chem., vol. 34, no. 2, pp. 193–198, 1996.;.

[143] A. Hynes, M. Shenton, and J. Badyal, “Pulsed plasma polymerization of perfluorocyclohexane,” Macromolecules, vol. 29, no. 12, pp. 4220–4225, 1996.;.

[144] C. Biloiu, I. A. Biloiu, Y. Sakai, H. Sugawara, and A. Ohta, “Amorphous fluorocarbon polymer (a-C:F) films obtained by plasma enhanced chemical vapor deposition from perfluoro-octane (C8F18) vapor. ii. dielectric and insulating properties,” J. Vac. Sci. Technol., A, vol. 22, p. 1158, 2004.;.

193

[145] R. Chen, V. Gorelik, and M. Silverstein, “Plasma polymerization of hexafluoropropylene: film deposition and structure,” J. Appl. Polym. Sci., vol. 56, no. 5, pp. 615–623, 1995.;.

[146] J.-H. Wang, J.-J. Chen, and R. B. Timmons, “Plasma synthesis of a novel CF3- dominated fluorocarbon film,” Chem. Mater., vol. 8, no. 9, pp. 2212–2214, 1996.;.

[147] Y. Iriyama, T. Yasuda, D. Cho, and H. Yasuda, “Plasma surface treatment on nylon fabrics by fluorocarbon compounds,” J. Appl. Polym. Sci., vol. 39, no. 2, pp. 249– 264, 1990.;.

[148] S. Vaswani, J. Koskinen, and D. W. Hess, “Surface modification of paper and cellulose by plasma-assisted deposition of fluorocarbon films,” Surf. Coat. Technol., vol. 195, no. 2, pp. 121–129, 2005.;.

[149] H. Biederman and L. Holland, “Metal doped fluorocarbon polymer films prepared by plasma polymerization using an RF planar magnetron target,” Nucl. Instrum. Methods Phys. Res., vol. 212, no. 1, pp. 497–503, 1983.;.

[150] M. Millard and E. Kay, “Difluorocarbene emission spectra from fluorocarbon plasmas and its relationship to fluorocarbon polymer formation,” J. Electrochem. Soc., vol. 129, no. 1, pp. 160–165, 1982.;.

[151] S. Coulson, I. Woodward, J. Badyal, S. Brewer, and C. Willis, “Plasmachemical functionalization of solid surfaces with low surface energy perfluorocarbon chains,” Langmuir, vol. 16, no. 15, pp. 6287–6293, 2000.;.

[152] S. R. Coulson, I. S. Woodward, J.P.S. Badyal, S.A. Brewer, and C. Willis, “Ultralow surface energy plasma polymer films,” Chem. Mater., vol. 12, no. 7, pp. 2031– 2038, 2000.;.

[153] L. Laguardia, D. Ricci, E. Vassallo, A. Cremona, E. Mesto, F. Grezzi, and F. Dellera, “Deposition of super-hydrophobic and oleophobic fluorocarbon films in radio frequency glow discharges,” in Macromol. Symp., vol. 247, no. 1. Wiley Online Library, 2007, pp. 295–302.;.

[154] F. Hu and J. C. Tou, “Method of forming a plasma polymerized film,” Feb. 27 1996, uS Patent 5,494,712.;.

[155] N. Inagaki, S. Kondo, M. Hirata, and H. Urushibata, “Plasma polymerization of organosilicon compounds,” J. Appl. Polym. Sci., vol. 30, no. 8, pp. 3385–3395, 1985.;.

[156] S. H. Lee and D. C. Lee, “Preparation and characterization of thin films by plasma polymerization of hexamethyldisiloxane,” Thin Solid Films, vol. 325, no. 1, pp. 83–86, 1998.;.

194

[157] K. G. Sachdev and H. S. Sachdev, “Characterization of plasma-deposited organosilicon thin films,” Thin Solid Films, vol. 107, no. 3, pp. 245–250, 1983.;.

[158] F. Fracassi, R. d’Agostino, and P. Favia, “Plasma-enhanced chemical vapor deposition of organosilicon thin films from tetraethoxysilane-oxygen feeds,” J. Electrochem. Soc., vol. 139, no. 9, pp. 2636–2644, 1992.;.

[159] M. Vasile and G. Smolinsky, “Organosilicon films formed by an RF plasma polymerization process,” J. Electrochem. Soc., vol. 119, no. 4, pp. 451–455, 1972.;.

[160] T. Hayakawa, M. Yoshinari, and K. Nemoto, “Characterization and protein- adsorption behavior of deposited organic thin film onto titanium by plasma polymerization with hexamethyldisiloxane,” Biomaterials, vol. 25, no. 1, pp. 119–127, 2004.;.

[161] J. Behnisch, J. Tyczkowski, M. Gazicki, I. Pela, A. Holländer, and R. Ledzion, “Formation of hydrophobic layers on biologically degradable polymeric foils by plasma polymerization,” Surf. Coat. Technol., vol. 98, no. 1, pp. 872–874, 1998.;.

[162] H. Schreiber, M. Wertheimer, and A. Wrobel, “Corrosion protection by plasma- polymerized coatings,” Thin Solid Films, vol. 72, no. 3, pp. 487–494, 1980.;.

[163] J. Fonseca, S. Tasker, D. Apperley, and J. Badyal, “Plasma-enhanced chemical vapor deposition of organosilicon materials: a comparison of hexamethyldisilane and tetramethylsilane precursors,” Macromolecules, vol. 29, no. 5, pp. 1705–1710, 1996.;.

[164] D. Trunec, Z. Navrátil, P. Stahel, L. Zajcková, V. Buršková, and J. Cech, “Deposition of thin organosilicon polymer films in atmospheric pressure glow discharge,” J. Phys. D : Appl. Phys., vol. 37, no. 15, p. 2112, 2004.;.

[165] A. Wrobel, M. Kryszewski, and M. Gazicki, “Oligomeric products in plasma- polymerized organosilicones,” J. Macromol. Sci., Chem., vol. 20, no. 5-6, pp. 583–618, 1983.;.

[166] M. Ryan, A. Hynes, and J. Badyal, “Pulsed plasma polymerization of maleic anhydride,” Chem. Mater., vol. 8, no. 1, pp. 37–42, 1996.;.

[167] F. Siffer, A. Ponche, P. Fioux, J. Schultz, and V. Roucoules, “A chemometric investigation of the effect of the process parameters during maleic anhydride pulsed plasma polymerization,” Anal. Chim. Acta, vol. 539, no. 1, pp. 289–299, 2005.;.

[168] S. Schiller, J. Hu, A. Jenkins, R. Timmons, F. Sanchez-Estrada, W. Knoll, and R. Förch, “Chemical structure and properties of plasma-polymerized maleic anhydride films,” Chem. Mater., vol. 14, no. 1, pp. 235–242, 2002.;.

195

[169] A. Jenkins, J. Hu, Y. Wang, S. Schiller, R. Foerch, and W. Knoll, “Pulsed plasma deposited maleic anhydride thin films as supports for lipid bilayers,” Langmuir, vol. 16, no. 16, pp. 6381–6384, 2000.;.

[170] S. Liu, M. M. Vareiro, S. Fraser, and A. T. A. Jenkins, “Control of attachment of bovine serum albumin to pulse plasma-polymerized maleic anhydride by variation of pulse conditions,” Langmuir, vol. 21, no. 19, pp. 8572–8575, 2005.;.

[171] S. P. Bhawalkar, J. Qian, M. C. Heiber, and L. Jia, “Development of a colloidal lithography method for patterning nonplanar surfaces,” Langmuir, vol. 26, no. 22, pp. 16662–16666, 2010.;.

[172] I. Badge, S. Sethi, and A. Dhinojwala, “Carbon nanotube-based robust steamphobic surfaces,” Langmuir, vol. 27, no. 24, pp. 14726–14731, 2011.;.

[173] K. Autumn, C. Majidi, R. Groff, A. Dittmore, and R. Fearing, “Effective elastic modulus of isolated gecko setal arrays,” J. Exp. Biol., vol. 209, no. 18, pp. 3558–3568, 2006.;.

[174] I. Badge, S. P. Bhawalkar, L. Jia, and A. Dhinojwala, “Tuning surface wettability using single layered and hierarchically ordered arrays of spherical colloidal particles,” Soft Matter, 2013.;.

[175] Y. Yi, H. Robinson, S. Knappe, J. Maclennan, C. Jones, C. Zhu, N. Clark, and J. Kitching, “Method for characterizing self-assembled monolayers as antirelaxation wall coatings for alkali vapor cells,” J. Appl. Phys., vol. 104, no. 2, pp. 023534–023534, 2008.;.

[176] S. H. Sajadinia and F. Sharif, “Thermodynamic analysis of the wetting behavior of dual scale patterned hydrophobic surfaces,” J. Colloid Interf. Sci., vol. 344, no. 2, pp. 575–583, 2010.;.

[177] N. A. Patankar, “Mimicking the lotus effect: influence of double roughness structures and slender pillars,” Langmuir, vol. 20, no. 19, pp. 8209–8213, 2004.;.

[178] Y. Yan, N. Gao, and W. Barthlott, “Mimicking natural superhydrophobic surfaces and grasping the wetting process: A review on recent progress in preparing superhydrophobic surfaces,” Adv. Colloid Interfac., vol. 169, no. 2, pp. 80–105, 2011.;.

[179] X. Feng and L. Jiang, “Design and creation of superwetting/antiwetting surfaces,” Adv. Mater., vol. 18, no. 23, pp. 3063–3078, 2006.;.

[180] C. Ran, G. Ding, W. Liu, Y. Deng, and W. Hou, “Wetting on nanoporous alumina surface: transition between wenzel and cassie states controlled by surface structure,” Langmuir, vol. 24, no. 18, pp. 9952–9955, 2008.;.

196

[181] P.-C. Lin and S. Yang, “Mechanically switchable wetting on wrinkled elastomers with dual-scale roughness,” Soft Matter, vol. 5, no. 5, pp. 1011–1018, 2009.;.

[182] V. Zorba, E. Stratakis, M. Barberoglou, E. Spanakis, P. Tzanetakis, and C. Fotakis, “Tailoring the wetting response of silicon surfaces via fs laser structuring,” Appl. Phys. A, vol. 93, no. 4, pp. 819–825, 2008.;.

[183] P. Roach, N. J. Shirtcliffe, and M. I. Newton, “Progess in superhydrophobic surface development,” Soft Matter, vol. 4, no. 2, pp. 224–240, 2008.;.

[184] G. Caputo, B. Cortese, C. Nobile, M. Salerno, R. Cingolani, G. Gigli, P. D. Cozzoli, and A. Athanassiou, “Reversibly light-switchable wettability of hybrid organic/inorganic surfaces with dual micro-/nanoscale roughness,” Adv. Funct. Mater., vol. 19, no. 8, pp. 1149–1157, 2009.;.

[185] T. Koishi, K. Yasuoka, S. Fujikawa, T. Ebisuzaki, and X. C. Zeng, “Coexistence and transition between cassie and wenzel state on pillared hydrophobic surface,” PNAS, vol. 106, no. 21, pp. 8435–8440, 2009.;.

[186] C. Extrand and S. I. Moon, “Intrusion pressure to initiate flow through pores between spheres,” Langmuir, vol. 28, no. 7, pp. 3503–3509, 2012.;.

[187] C. Extrand, “Repellency of the lotus leaf: Resistance to water intrusion under hydrostatic pressure,” Langmuir, vol. 27, no. 11, pp. 6920–6925, 2011.;.

[188] S. Bán, E. Wolfram, and S. Rohrsetzer, “The condition of starting of liquid imbibition in powders,” Colloids Surf., vol. 22, no. 2, pp. 291–300, 1987.;.

[189] N. J. Shirtcliffe, G. McHale, M. I. Newton, F. B. Pyatt, and S. H. Doerr, “Critical conditions for the wetting of soils,” Appl. Phys. Lett., vol. 89, no. 9, pp. 094101–094101, 2006.;.

[190] Q.-S. Zheng, Y. Yu, and Z.-H. Zhao, “Effects of hydraulic pressure on the stability and transition of wetting modes of superhydrophobic surfaces,” Langmuir, vol. 21, no. 26, pp. 12207–12212, 2005.;.

[191] H. Kamusewitz and W. Possart, “Wetting and scanning force microscopy on rough polymer surfaces: Wenzel’s roughness factor and the thermodynamic contact angle,” Appl. Phys. A., vol. 76, no. 6, pp. 899–902, 2003.;.

[192] H. Rangwalla, A. D. Schwab, B. Yurdumakan, D. G. Yablon, M. S. Yeganeh, and A. Dhinojwala, “Molecular structure of an alkyl-side-chain polymer-water interface: origins of contact angle hysteresis,” Langmuir, vol. 20, no. 20, pp. 8625–8633, 2004.;.

[193] S. Wang and L. Jiang, “Definition of superhydrophobic states,” Adv. Mater., vol. 19, no. 21, pp. 3423–3424, 2007.;.

197

[194] R. Narhe and D. Beysens, “Nucleation and growth on a superhydrophobic grooved surface,” Phys. Rev. Lett., vol. 93, no. 7, p. 076103, 2004.;.

[195] K. K. Varanasi, M. Hsu, N. Bhate, W. Yang, and T. Deng, “Spatial control in the heterogeneous nucleation of water,” Appl. Phys. Lett., vol. 95, no. 9, pp. 094101–094101, 2009.;.

[196] Y. Zheng, D. Han, J. Zhai, and L. Jiang, “In situ investigation on dynamic suspending of microdroplet on lotus leaf and gradient of wettable micro-and nanostructure from water condensation,” Appl. Phys. Lett., vol. 92, no. 8, pp. 084106– 084106, 2008.;.

[197] P. He, D. Shi, J. Lian, L. Wang, R. C. Ewing, W. van Ooij, W. Li, and Z. Ren, “Plasma deposition of thin carbonfluorine films on aligned carbon nanotube,” Appl. Phys. Lett., vol. 86, no. 4, pp. 043107–043107, 2005.;.

[198] E. Hare, E. Shafrin, and W. Zisman, “Properties of films of adsorbed fluorinated acids,” J. Phys. Chem., vol. 58, no. 3, pp. 236–239, 1954.;.

[199] A. H. Barber, S. R. Cohen, and H. D. Wagner, “Static and dynamic wetting measurements of single carbon nanotubes,” Phys. Rev. Lett., vol. 92, no. 18, p. 186103, 2004.;.

[200] D. Han, K. Zhou, and A. M. Bauer, “Phylogenetic relationships among gekkotan lizards inferred from C-mos nuclear DNA sequences and a new classification of the Gekkota,” Biol. J. Linn. Soc., vol. 83, no. 3, pp. 353–368, 2004.;.

[201] T. Gamble, E. Greenbaum, T. R. Jackman, A. P. Russell, and A. M. Bauer, “Repeated origin and loss of adhesive toepads in geckos,” Plos one, vol. 7, no. 6, p. e39429, 2012.;.

[202] K. Koch, B. Bhushan, and W. Barthlott, “Diversity of structure, morphology and wetting of plant surfaces,” Soft Matter, vol. 4, no. 10, pp. 1943–1963, 2008.;.

[203] R. Jetter and S. Schäffer, “Chemical composition of the Prunus laurocerasus leaf surface. dynamic changes of the epicuticular wax film during leaf development,” Plant Physiol., vol. 126, no. 4, pp. 1725–1737, 2001.;.

[204] P. H. Niewiarowski, A. Stark, B. McClung, B. Chambers, and T. Sullivan, “Faster but not stickier: Invasive house geckos can out-sprint resident mournful geckos in moorea, french polynesia,” J. Herpetol., vol. 46, no. 2, pp. 194–197, 2012.;.

[205] A. Russell and M. Johnson, “Real-world challenges to, and capabilities of, the gekkotan adhesive system: contrasting the rough and the smooth,” Can. J. Zool., vol. 85, no. 12, pp. 1228–1238, 2007.;.

198

[206] G. Huber, S. N. Gorb, N. Hosoda, R. Spolenak, and E. Arzt, “Influence of surface roughness on gecko adhesion,” Acta Biomater., vol. 3, no. 4, pp. 607–610, 2007.;.

[207] S. Hu, S. Lopez, P. H. Niewiarowski, and Z. Xia, “Dynamic self-cleaning in gecko setae via digital hyperextension,” J. R. Soc. Interface, vol. 9, no. 76, pp. 2781– 2790, 2012.;.

[208] K. Liu, J. Du, J. Wu, and L. Jiang, “Superhydrophobic gecko feet with high adhesive forces towards water and their bio-inspired materials,” Nanoscale, vol. 4, no. 3, pp. 768–772, 2012.;.

[209] K. Autumn, “Gecko adhesion: structure, function, and applications,” MRS Bull., vol. 32, no. 06, pp. 473–478, 2007.;.

[210] K. Autumn and N. Gravish, “Gecko adhesion: evolutionary nanotechnology,” Philos. Trans. R. Soc. London, Ser. A, vol. 366, no. 1870, pp. 1575–1590, 2008.;.

[211] K. Autumn, Y. A. Liang, S. T. Hsieh, W. Zesch, W. P. Chan, T. W. Kenny, R. Fearing, and R. J. Full, “Adhesive force of a single gecko foot-hair,” Nature, vol. 405, no. 6787, pp. 681–685, 2000.;.

[212] A. Y. Stark, I. Badge, N. A. Wucinich, T. W. Sullivan, P. H. Niewiarowski, and A. Dhinojwala, “Surface wettability plays a significant role in gecko adhesion underwater,” PNAS, 2013.;.

[213] N. S. Pesika, H. Zeng, K. Kristiansen, B. Zhao, Y. Tian, K. Autumn, and J. Israelachvili, “Gecko adhesion pad: a smart surface?” J. Phys.:Condens. Matter, vol. 21, no. 46, p. 464132, 2009.;.

[214] A. M. Peattie, C. Majidi, A. Corder, and R. J. Full, “Ancestrally high elastic modulus of gecko setal β-keratin,” J. R. Soc. Interface, vol. 4, no. 17, pp. 1071–1076, 2007.;.

[215] L. Feng, Y. Zhang, Y. Cao, X. Ye, and L. Jiang, “The effect of surface microstructures and surface compositions on the wettabilities of flower petals,” Soft Matter, vol. 7, no. 6, pp. 2977–2980, 2011.;.

[216] U. Hiller, “Untersuchungen zum Feinbau und zur Funktion der Haftborsten von Reptilien,” Zoomorphology, vol. 62, no. 4, pp. 307–362, 1968.;.

[217] M. K. Chaudhury and G. M. Whitesides, “Direct measurement of interfacial interactions between semispherical lenses and flat sheets of poly (dimethylsiloxane) and their chemical derivatives,” Langmuir, vol. 7, no. 5, pp. 1013–1025, 1991.;.

199

[218] H. Haidara, M. Chaudhury, and M. Owen, “A direct method of studying adsorption of a surfactant at solid-liquid interfaces,” J. Phys. Chem., vol. 99, no. 21, pp. 8681–8683, 1995.;.

[219] M. K. Chaudhury and J. Y. Chung, “Studying friction and shear fracture in thin confined films using a rotational shear apparatus,” Langmuir, vol. 23, no. 15, pp. 8061– 8066, 2007.;.

[220] H. Izadi, B. Zhao, Y. Han, N. McManus, and A. Penlidis, “Teflon hierarchical nanopillars with dry and wet adhesive properties,” J. Polym. Sci. B Polym. Phys, vol. 50, no. 12, pp. 846–851, 2012.;.

[221] M. K. Chaudhury, “Interfacial interaction between low-energy surfaces,” Mater. Sci. Eng., R, vol. 16, no. 3, pp. 97–159, 1996.;.

[222] R. Blossey, “Self-cleaning surfaces—virtual realities,” Nat. Mater., vol. 2, no. 5, pp. 301–306, 2003.;.

200

APPENDIX

A1. Derivation of Wenzel equation

Let us assume a droplet of water with volume V which when placed on such a rough surface takes up a spherical cap shape with a contact angle of . Let rb be the radius of the spherical cap. The droplet exists in the Wenzel wetting state, as shown in the schematic representation above. The change in the surface free energy of the system (ΔG) can be estimated as follows.

(1.1)

Here, is the curved surface area of a spherical cap, is the area of contact at the solid-liquid interface and S is the surface area of the water droplet before it contacts the

solid surface ( ), which is constant for a constant volume of the droplet. , and are the surface tensions at the liquid-vapor, solid-liquid and solid-vapor interfaces respectively. Let be the projected area of spherical cap, in which case according to

Wenzel model, . Also, from Young’s equation we know that

. Appropriate substitutions in equation 1.1 reduce it to the following equation:

(1.2)

201

Taking derivative on both sides and substituting and , we

get the following equation for Wenzel state:

(1.3) A2. Derivation of Cassie-Baxter equation

Let us assume a droplet of water with volume V which when placed on the rough surface takes up a spherical cap shape with a contact angle of . Let rb be the radius of the spherical cap. In the case of a Cassie-Baxter wetting state of the droplet, the change in free energy of the system (ΔG) is given by the following equation (similar to the derivation of Wenzel equation).

(2.1)

Here, is the curved surface area of a spherical cap, is the area of contact at the liquid-vapor interface at the base of the cap, is the area of contact at the solid-liquid interface and S is the surface area of the water droplet before it contacts the solid surface, which is constant for a constant volume of the droplet. , and are the surface tensions at the liquid-vapor, solid-liquid and solid-vapor interfaces respectively. Let be the projected area of spherical cap and let and . Also, from

Young’s equation we know that . Appropriate substitutions in equation 2.1 reduce it to the following equation:

(2.2)

Taking derivative on both sides and substituting and , we

get the following equation for Cassie-Baxter state:

202

(2.3)

A3. Single Layer Roughness : Model calculations

Hexagonal Unit cell calculations:

(3.1)

The total solid surface area is the area of the hexagon in addition to the surface area of the particle (Assuming single point contact between the particle and the flat surface).

(3.2) where is the particle diameter.

Wenzel state calculations:

Roughness factor ( defined as the ratio of total surface area to the projected area is calculated using following equation.

(3.3)

The roughness of the PECVD layer deposited on the surface (R2) also adds to the overall surface roughness. The effective roughness of the surface (R) is thus, defined using following equation:

(3.4)

Wenzel state contact angle ( thus, can be calculated using following equation:

(3.5)

Here, is the equilibrium contact angle defined for a homogenous surface.

203

Correction on R1:

The surface area of the part of the sphere and the roughness factor are subsequently different from before due to spherical cap shape of the particle, as opposed to perfectly sphere shape.

(3.6)

(3.7)

(3.8)

Cassie-Baxter state of single layer roughness:

Let A S , A L and A f be the area of solid-liquid interface, liquid-vapor interface and projected surface areas of a unit cell respectively and let ‘n’ be the total number of unit cells constituting the surface.

(3.9)

(3.10)

Unit Cell Calculations: Cassie-Baxter state

For unit cell with single layer roughness, as shown in figure 1c, the area fractions and thus, the contact angles are calculated as follows:

(3.11)

(3.12)

204

(3.13)

With appropriate substitutions, we get:

and (3.14)

Substituting f1 and f2 gives the equation for Cassie-Baxter angle as follows:

(3.15)

A4. Dual Roughness: Model calculations

In the case of dual roughness, the parameters L1, D1, r1 correspond to the layer of bigger particles; L2, D2 and r2 correspond to that of small particles. α1 is the parameter that defines the flattened base of the spherical particles.

(1) Wenzel State:

(4.1)

205

(4.2)

Let (4.3)

and (4.4)

From equations 4.3 and 4.4, we know that R1 and R2 exclusively correspond to big particle layer and small particle layer respectively. Let R3 be the roughness of the

PECVD coating layer in this case. Thus, the effective roughness (R) for the Wenzel state of dual roughness can be defined as follows:

(4.5)

(2) Cassie-Baxter state: The fractions for dual roughness can be calculated in a way similar to that of Cassie-

Baxter state of single roughness.

(4.6)

L-V interface here implies liquid-vapor interface.

(4.7)

(4.8)

It can be assumed that (4.9)

206

With appropriate substitutions we get equations for calculating f1, f2 and thus, the contact angle for Cassie-Baxter state.

(4.10)

(4.11)

(4.12)

(3) Penetrating Cassie-Baxter state:

A schematic of Penetrating Cassie-Baxter state is shown in figure 4c. Since, this state exhibits partial Cassie-Baxter character equation 1.12 can be used to calculate the respective contact angles. The fractions for this type of wetting state can be calculated using following equations.

(4.13)

(4.14)

(4.15)

Equations 4.13 and 4.14 can be simplified by defining a common factor K.

(4.16)

207

With appropriate substitutions we get the equation to calculate f1, f2 and thus, contact angle for penetrating Cassie-Baxter state (θPCB).

and (4.17)

(4.18)

For the r2 values that we used in our experiments (1.5 and 2), no real values of θPC could be calculated using any value of fitting parameter αPC. However, for lower values of r2

(0.9 for 1H,1H,of 2H-perfluoro-1-dodecene and 1 for HMDSO) real values of θPC could be calculated and compared with θW and θCB.

A5. Free energy calculations for different wetting states

In order to be able to predict the most stable wetting state and to calculate the energy barrier between different possible wetting states, we calculated free energy for each wetting state. The free energy equations are derived as follows.

Similar to the derivation of Cassie-Baxter equation, let a droplet of water with volume V be deposited on a rough surface. The droplet takes up a spherical cap shape with radius rb and forms a contact angle of θ. The change in free energy for this case is calculated using equation 2.1.

S in equation 2.1 has a constant value for a given volume of a droplet and thus, it results in a constant relative change in all the respective free energies of wetting states.

The constant value contributed to the overall free energy change is thus, not accounted for in the calculations here.

Now, for a spherical cap with contact angle θ,

208

(5.1)

Let Af be the projected area of the spherical cap (as defined before in the case of Cassie-

Baxter equation derivation).

(5.2)

Substituting equations 5.1 and 5.2 in equation 2.1 we get,

(5.3)

Now, (5.4)

(5.5)

2 Substituting for rb (equation 5.5) in equation 5.3, we get an equation of free energy in terms of apparent contact angle as follows:

(5.6)

In the case of Wenzel state, f2 = 0, f1 =R and θ = θW

Equation 5.6 thus reduces to:

(5.7)

In the case of Cassie-Baxter and Penetrating Cassie-Baxter states, equation 5.6 could be used to calculate G*, with appropriate substitutions made for apparent contact angle

(either predicted using model or measured experimentally), f1 and f2.

209

A6. Adhesion model calculations

Dry contact of smooth surfaces:

The Dupré equation for calculating the work of adhesion between two surfaces, 1 and 2, in its general form is written as follows:

(6.1)

We used equation 1 to calculate the work of adhesion between each of our four surfaces (glass, plexiglass, OTS-SAM coated glass and PTFE) and the “gecko hair-like” n-hexadecane surface, assuming that the contact interface formed as a result of contact between the two is flat (Table 2). The Young-Dupré equation for the dry contact between the two surfaces when air is the medium of contact can be written as follows:

(6.2)

Where, is the interfacial energy at the contact interface between the “gecko hair- like” surface and the contact surface (glass, plexiglass, OTS-SAM coated glass or PTFE),

is the surface energy of “gecko hair-like” n-hexadecane surface and is the surface energy of the contact surface.

Young’s equation for the contact angle ( ) that n-hexadecane makes on a given contact surface is:

(6.3) 210

Substituting equation 3 in equation 2 for we get:

(6.4)

We measured the contact angle of n-hexadecane on all four surfaces that we used for the gecko trials to obtain the value of (see second column of Table 1). The value of is known to be 25 mJ/m2. Substituting all the known values in equation 4 gives the work of dry adhesion ( .

Wet contact of Smooth Surfaces:

In the case of wet adhesion i.e. the case where water is the medium of contact, the work of adhesion ( is calculated using the following equation:

(6.5)

Here, is the work of adhesion between two surfaces contacting under water. Similar to dry contact, denotes the interfacial energy at the contact surface-water interface (contact surface is glass, plexiglass, OTS-SAM coated glass or PTFE), is the n-hexadecane-water interfacial energy and is the interfacial energy at the surface-n-hexadecane contact interface.

Similar to the contact angle of n-hexadecane , the contact angle of water ) was also measured on all four surfaces (first column of Table 1). It gives the following relationship:

(6.6)

Where is the surface tension of water. Substituting equations 3 and 6 in 5 for

and gives the following equation: 211

(6.7)

2 2 The values of and are known to be 50 mJ/m and 25 mJ/m respectively.

and were determined experimentally as discussed in Text S3. Table 1 summarizes the contact angles of water and n-hexadecane on different test surfaces. Substituting all the values in equation 7 gives the value of wet adhesion. Thus, knowing all the parameters, we can estimate using equation 8 below (derived from equations 4 and 7):

(6.8)

Patterned surface calculations : Tetrad pattern of setae

In the case of contact between the tetrad-patterned “gecko hair-like” surface and the contacting surfaces (glass, plexiglass, OTS-SAM coated glass or PTFE), there are four possible cases as shown schematically in Table 2. The ratio for all the cases can be estimated as follows (final ratios are reported in Table 2).

Case 1:

(6.9)

A1 and A2 in the case of patterned surfaces correspond to total surface areas of the

“gecko-hair-like” n-hexadecane tetrad patterned unit cell and the surface it is in contact with (glass, plexiglass, OTS-SAM coated glass or PTFE), respectively. Further simplification of equation 9 and appropriate substitutions give an equation to calculate

for case 1 (below).

(6.10) 212

is thus calculated using equations 4 and 10:

(6.11)

Case 2:

Similar to case 1, the equation for for case 2 is derived as:

(6.12)

Further simplification gives equation 12 in terms of parameters measurable experimentally.

(6.13)

for this case is derived using equations 4 and 13:

(6.14)

Case 3:

for case 3 is derived in a similar way as cases 1 and 2 above.

(6.15)

Simplification and substitution reduces equation 15 in the following form:

(6.16)

Using equations 4 and 16, the ratio can be calculated as shown below.

213

(6.17)

Case 4:

The equation for in case 4 is calculated as follows:

(6.18)

Simplification on substitution gives equation 18 in terms of θ1and θ2.

(6.19)

The equation for the ratio is derived below:

(6.20)

A1, A2 and AC here represent total surface area of tetrad unit cell, total surface area of the test surface and the contact area respectively. We calculated the areas using the dimensions of a unit cell we estimated from SEM images. For the tetrad-patterned unit

2 2 2 cell, A1 = 3961μm , A2 = 121μm and AC = 64μm . The respective ratios of areas for

were calculated based on these values and substituted.

Non-tetrad Patterned Surface in Wet and Dry Contact

A schematic of a non-tetrad patterned surface is shown below (Figure A1). One unit cell is boxed. Similar to a tetrad-patterned unit cell, dimensions of this type of gecko toe morphology were estimated using SEM imaging. For this type of unit cell, A1 = 996

2 2 2 μm , A2 = 36 μm and AC = 16 μm . ratios were thus calculated using these values and equations derived above for four different cases of wet contact. The results are tabulated below (Table A1)

214

Figure A1. A schematic representation of the non-tetrad patterned unit cell morphology.

One unit cell is boxed in red. Columns are 60m tall and 4m wide. Each column is separated by 1m.

Table A1. Table of ratios for the non-tetrad patterned surface in each of the four wetting cases and on each test surface.

Surface Case 1 Case 2 Case 3 Case 4

Glass -0.62 30.39 -30.45 0.56

Plexiglass 1.23 32.04 -29.43 1.38

OTS-SAM 1.78 34.37 -30.95 1.64

PTFE 1.98 34.83 -31.12 1.74

215

Hamaker Constant Calculations

The parameters used for Hamaker constant calculations and the values obtained for different substrates are tabulated in Table A2 and Table A3 respectively.

Table A2. The parameters used to calculate Hamaker constants for the absorption

15 -1 frequency (νe) of 3 x 10 s

Dielectric Constant Refractive Index (ε) (n) Water 80 1.33 n-hexadecane 2.05 1.42 Glass 3.7 1.54 Plexiglass 2.6 1.5 PTFE 2.1 1.36

Table A3. Hamaker constant values calculated for different contact surfaces; the subscripts 1, 2 and 3 correspond to “gecko hair-like” n-hexadecane, substrate (glass, plexiglass and PTFE) and air (or water) respectively.

Surface Hamaker constant (J) -20 -20 A132(air) (×10 J) A132(water) (×10 J) Glass 6.53 0.75 Plexiglass 6.16 0.68 PTFE 4.59 0.34

216