Patterning Elastomer, Thermoplastics and Shape Memory

Home , Soft lithography

PATTERNING ELASTOMER, THERMOPLASTICS AND SHAPE MEMORY

MATERIAL BY UVO LITHOGRAPHY AND SOFT LITHOGRAPHY

A Dissertation

Presented to

The Graduate Faculty of The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Ying Chen

January, 2017

PATTERNING ELASTOMER, THERMOPLASTICS AND SHAPE MEMORY

MATERIAL BY UVO LITHOGRAPHY AND SOFT LITHOGRAPHY

Ying Chen

Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Alamgir Karim Dr. Sadhan C. Jana

______Committee Member Dean of the College Dr. Abraham Joy Dr. Eric J. Amis

______Committee Member Dean of the Graduate School Dr. Kevin A. Cavicchi Dr. Chand Midha

______Committee Member Date Dr. Xiong Gong

______Committee Member Dr. Jutta Luettmer-Strathmann

ABSTRACT

Micro/nano patterning and structures, especially hierarchical structures have attracted great research interest in the past few decades, because these structures exhibit unique properties like tunable adhesion and wetting. Especially, micro/nano patterning is of great interest for semiconductor, micro/nano fluidics, optical and photonic devices applications. Traditional methods used for the fabrication of hierarchical structures typically involve the formation of complex patterned features through multistep lithography processes. These methods usually require expensive equipments and specialty reagents, so that the low-cost fabrication of well-controlled micro-nano patterns for diverse potential applications remains a challenge.

We demonstrated a versatile and inexpensive method for controlling the surface relief structure of polymer films over large areas through a two-step imprinting process.

First, nanoscale patterns were formed by nanoimprinting elastomer (PDMS) films with a pattern on a DVD disk. Micron-scale patterns were then superimposed on the nanoimprinted PDMS films by exposing them to ultraviolet radiation in oxygen (UVO) through a photomask. UVO exposure leads to a conversion and densification of PDMS to

SiOx, leading to micron height relief features that follow a linear scaling relation with pattern dimension. Further, the pattern scopes are shown to collapse into a master curve by

iii

normalized feature values. Interestingly, these relief structures preserve the nanoscale features. In this dissertation, the influence of the self-limiting PDMS densification, walls stress at the boundary of micro depression, PDMS thickness, modulus and UVO exposure energy is studied in control of the micro depression scale. The method fidelity was evaluated in coarse-grained molecular dynamics simulations and confirmed experimentally.

In the second part of this study, this simple two-step imprinting process involving both nanoimprinting and UV radiation, is studied for pattern transfer demonstration of the dimension adjustable micro-nano hierarchically structures not only on elastomer films, self-assembled monolayer and nanoparticles but also by imprinting onto thermoplastic polymer films. The patterning of thermoplastic polymer films is achieved through capillary force lithography (CFL). Fundamental study of CFL is conducted to understand the influence of film thickness and annealing process. Another two forms of soft lithography examined here include replica molding, and microcontact printing. Thus, these generated patterns are successfully extended to self-assembled monolayer and nano particles, which enlightens the further potential applications of the proposed patterning method.

Finally, Shape memory compounds based on mixtures of an ionomer with a FA are used to develop shape memory or shape morphing surfaces with micro- or nano-scale features. Three different FAs, zinc stearate (ZnSt), stearic acid (SA) or lauric acid (LA) are mixed with Zn salt of sulfonated EPDM, respectively, to act as the temporary networks.

As with the bulk shape memory compounds based on the ionomer/FA design, the switching temperature for micro- and nano-scale surface pattern recovery can be easily tuned by simply changing the FA used in the composition. The shape memory recovery efficiency iv

of the micro and nano scaled surface topography is compared with bulk materials. The recovery behavior for the surface nanopattern, however, had lower efficiency than micropattern scale and bulk shape memory of the same material, which may be due to the effects of the excess surface energy on the dynamics of the surface patterns or creep of the temporary or permanent networks due to the high stress used to deform the nano-scale grating pattern and produce the temporary crosshatched pattern.

ACKNOWLEDGEMENTS

For the past five years of studying at The University of Akron, I have received generous help from a lot of people, and this dissertation benefits so much from their kind help. I am grateful to meeting so many nice persons and worked with them, they have very big influences not only on my career but how I think about life and myself.

With this project, my advisor Dr. Karim was a tremendous source of insight and practicality. I would like to express my special appreciation to him: Dr. Karim. I am also grateful to my committee members, Dr. Xiong Gong, Dr. Kevin Cavicchi, Dr. Abraham

Joy, Dr. Jutta Luettmer-Strathmann for their critical review and helpful suggestions.

Special thanks to all current group members and alumni of the Karim Research

Group, especially for Dr. Manish Kulkarni and Dr. Diya Bandyopadhyay, due to their kind help for all of suggestions in starting and developing projects. I am extremely grateful for that.

Outside the committee, my sincere thanks go to Dr. Jack Douglas, from NISR, Dr.

Robert Weiss, Dr. Matthew L. Becker and Dr. Andrey V. Dobrynin, from Polymer

Department for their collaborations, discussion, advices, and help during my research. It has been delighted to collaborate with them during the past five years.

I would also like to thank Dr. Larry Rhodes at Schneller, Inc. I spent great four months doing an internship under his management. He provided me with the great opportunity to work in an industrial environment and guided me to the industry of semiconductor.

I also want to thank students in the Department of Polymer Science and Engineering,

Zhiyang Zhao, Zilu Wang and Changhuai Ye, for collaborations and helps throughout my research. They have been really helpful for past years, and I am very grateful for their kindness and friendship.

The most special thanks go to my best family, my husband, mom, dad and brother.

Your love and unconditional support is the most important thing in my life.

vii

CONTENTS

1 Introduction ...... 1

2 Background ...... 5

2.1 UV Lithography ...... 5

2.2 Soft Lithography ...... 10

Replica Molding ...... 12

Microcontact Printing ...... 16

Capillary Force Lithography ...... 20

2.3 Shape memory behavior...... 25

3 Methodology of a novel UVo Lithography for a dimension controlled micro- nano hierarchical patterning...... 37

3.1 Materials ...... 39

3.2 Experimental methods ...... 39

viii

3.3 Micro-nano hierarchical patterns and parameters influencing the

efficiency of UVO Lithography ...... 42

Chemistry of PDMS under UVO exposure ...... 42

Micro-nano hierarchical pattern ...... 45

Parameters influencing the process of UVO induced PDMS

depression ...... 49

Simulation of the nano patterning process ...... 55

3.4 Summary ...... 61

4 Investigating the application of the prepared micron-nano hierarchical mold for Soft Lithography ...... 62

4.1 Materials ...... 63

4.2 Experimental methods ...... 63

Pattern transfer – Replica Molding ...... 63

Pattern transfer – Microcontact Printing ...... 64

Pattern transfer - Capillary Force Lithography ...... 65

4.3 Micro-nano patterning through replica molding ...... 66

4.4 Micro-nano patterning microcontact imprinting ...... 69

4.5 Defects analysis and elimination via a process optimized Capillary Force

Lithography ...... 72

4.6 Summary ...... 83

5 shape memory behavior of Micro and nano scale patterned

FA/elastomer composites ...... 85

5.1 Materials ...... 86

5.2 Experimental methods ...... 86

5.3 Shape memory behavior of bulk composites ...... 91

5.4 Surface micropattern memory behavior ...... 97

5.5 Shape memory behavior of nano patterned film ...... 100

5.6 Summary ...... 103

6 Conclusions ...... 105

7 References ...... 109

LIST OF TABLES

Table Page

2.1 Summary of resist challenges and imaging challenges...... 10

3.1 Young’s modulus and nano-pattern height for the negative mold with different PDMS base cross-linker weight ratio where initial pattern height H0=155 nm. Data are obtained from pure nanopattern samples (cured at 120 °C for 2 h)...... 58

5.1 Shape fixing and recovery efficiencies for SMP-ZnSt for five consecutive shape memory cycles...... 96

LIST OF FIGURES

Figure Page

2.1 Process flow of photolithography using negative photoresist (not to scale)...... 7

2.2 Process of photolithography by using negative and positive photoresists...... 7

2.3 (a) Visible light absorption of the open and closed forms of the photochromic material used for absorbance modulation photolithography. (b) Figure of the creation of a controlled, reprogrammable, near-field aperture using NSOM...... 8

2.4 SEM images of voxels created by MAP...... 9

2.5 Micro mold collapse modes. (a) Sidewall collapse. (b) Roof collapse. (c) Buckling. (d)

Lateral collapse...... 12

2.6 Procedure of Replica Molding...... 13

2.7 SEM images of nanopillars and nano holes at different processing steps of replica molding. First row: nanopillars on silicon masters. Second row: nano holes at PDMS molds replicated from corresponding silicon masters. Third row: nanopillars at epoxy molded from the corresponding PDMS molds in second row...... 15

xii

2.8 (left) SEM image of the generated micro-nano patterns composed of microspheres and

[(PAH/PAA)(PAH/SiO2)3]n films. (right) Profiles of a water droplet on the generated film, and sliding on the surface...... 15

2.9 Procedure for Microcontact Printing...... 18

2.10 (a) Subtractive procedure used to print a SAM on a metallic film. Pattern is defined by a subsequent etching process. (b) Additive process used to print seeding solutions or catalysts that forms nanowires using electroless deposition (ELD). (c–f) Atomic Force

Microscope (AFM) images of gold, silver, copper, and palladium nanowires formed by subtractive procedure as shown in a...... 19

2.11 Process of Micro molding in Capillaries(MIMIC)...... 21

2.12 Process of Capillary Force Lithography...... 22

2.13 Figure diagram of capillary force lithography (CFL): a) when film is relatively thick with respect to the mold pattern depth (excessive material for mold filling) and b) when polymer film is thin...... 24

2.14 Molecular mechanism of thermal SMPs...... 26

2.15 Illustration of the molecular mechanisms of the thermal induced shape-memory function for a) a multiblock copolymer, where Tc=Tm, b) a covalently cross-linked polymer, where Tc=Tm, and c) a polymer network, where Tc=Tg...... 27

xiii

2.16 Fabrication and deformation/recovery of a 2D SMP membrane. (a) Fabrication of a mold master by holographic lithography. (b) Fabrication of the SMP membrane with periodic holes by replica molding. (c) micro-holes shape memory demonstration...... 31

2.17 Pattern flat out and recovery in a SMP membrane. Optical images of the: (a) original;

(b), (c) partially deformed with ε ~ (13 ± 2)% and (20 ± 2)%, respectively; (d) completely deformed with ε ~ (13 ± 2)%; and (e) recovered SMP membranes. (f–j) Corresponding

SEM images of the SMP membranes shown in (a–e). (k–o) Higher magnification SEM images of (f–j). (p) Transparency comparison between the original and the deformed SMP film. (q) Display of ‘‘Penn’’ logos underneath the SMP films. (r) UV-Vis spectra of the

SMP films at different steps corresponding to a, b, d and e...... 32

2.18 Top: Figure of the programming and recovery steps. (a)–(k) digital images, AFM images and AFM line profiles of the patterns at the permanent, temporary and recovered stages. (j) and (k) the plots of remaining pattern heights after programming (Hprogrammed) and the recovered pattern height (Hrecovered) as a function of the permanent pattern height

(Hpermanent), respectively...... 34

2.19 Top: Schematic of the programming and recovery steps with a second grating pattern.

(a)–(c) AFM height images and line profiles of the patterns at the permanent, temporary and recovered stages...... 35

2.20 Figure illustration of the permanent and temporary network at a Zinc oleate/Zn-

SEPDM system...... 36

xiv

3.1 models and dimension information of the TEM grids used in experiments...... 39

3.2 Schematic illustration of nanopatterning ...... 41

3.3 AFM images of channel patterns at the original (A) and PDMS replica(B) of CD and

DVD...... 41

3.4 Experimental assembly for micro patterning...... 42

3.5 (A) XPS color map of C 1s element at 40 min UVO exposed PDMS with 100 mesh

TEM grid as a photomask, (B) XPS spectra of Si 2p level at UVO exposed and blocked

PDMS surface...... 44

3.6 AFM images of hierarchical PDMS pattern combining DVD with TEM grid patterns

(300 hexagonal grid mesh, 1000 grid mesh and 2000 grid mesh) after a 4 h (85.3 J/cm2)

UVO exposure. Insets are the structures with high resolution, with dimension of 20 µm x

20 µm, 10 µm x 10 µm and 20 µm x 20 µm, from top to bottom. The bottom shows the

Fast Fourier Transform (FFT) of the AFM images with hierarchical pattern combining

2000 grid mesh (side width: 7.5 µm) and DVD channel. Line profile of two micro repeat units corresponding to the 2000 grid mesh imprinting case is shown...... 48

3.7 AFM images of hierarchical pattern combining CD with TEM grid pattern (A1) 300 hexagon, (B1) 1000 mesh square, (C1) 2000 mesh square, with 4 h UVO exposure. .... 48

3.8 Digital pictures of hierarchically patterned PDMS combining DVD pattern and square patterns with side lengths of (a) 500 μm and (b) 250 μm...... 49

3.9 (A) AFM images of flat PDMS irradiated under UVO for 4 h through a TEM grid with different patterns: 300 HEX mesh (57 µm side), 1000 square mesh (19 µm side), 2000 square mesh (7.5 µm side). (B) UVO unnormalized micropattern depth of the UVO micro patterning as a function of in-plane dimension. (C) normalized micropattern depth of the UVO micro patterning (from top surface to center of pattern)...... 51

3.10 (A) Depth/aspect ratio changes with UVO exposure time/cumulative radiation energy in the case of a rough TEM grid (1000 mesh, 19 μm). (B) Plot of the depth/ aspect ratio changes with UVO exposure time/cumulative radiation energy. The black and black dashed curves are measured on PDMS samples, half-blocked by a flat silicon wafer, for PDMS of thickness (1.22 ± 0.031) mm and (3.89 ± 0.116) mm, respectively. The blue curve is the depth measured from flat PDMS covered by 19 μm mesh photomask. The red curve is measured from DVD patterned PDMS covered by 19

μm photomask. The graph shows a non-monotonic PDMS micropattern depth change with increasing UVO exposure time: 0.5 h (UVO energy: 10.7 J/cm2), 1 h

(21.4 J/cm2), 2 h (42.7 J/cm2), 3 h (64.0 J/cm2), 5 h (106.6 J/cm2), 7 h (149.2 J/cm2) and 9 h (191.8 J/cm2) corresponding to data points in curves. Error bars in (A) and

(B) are estimated from the standard deviation of repeated measurements. (C) PDMS

Depth changes with different UVO exposure at 0.5 h (red columns) and 5 h (black columns) upon PDMS with Young’s modulus of 0.57 MPa...... 54

3.11 Evolution of the replica shape produced via cross-linking (blue color) compared with the corresponding master mold shape (yellow color) as a function of master mold size H0 and shear modulus GN of the negative mold network...... 58 xvi

3.12 Dependence of the deformation ratio ΔH/H0 on the elastocapillary number for different systems: the negative mold (black solid square), melt with negative mold of system 1 (red empty square), melt with negative mold of system 2 (green half solid square).

Experimental data from Table 3.1 are shown by pink stars. Inset shows definitions of pattern deformation in studied systems...... 60

4.1 Procedure of replica molding...... 64

4.2 Experimental process for pattern transfer from PDMS to thermoplastic material (PS in this study) via Capillary Force Lithography(CFL)...... 65

4.3 Left column shows AFM images of hierarchical PDMS pattern combining DVD with

TEM grid patterns after a 4 h (85.3 J/cm2) UVO exposure. The right column shows negative imprint of left column PDMS patterns. Insets are the structures with high resolution, with dimension of 20 µm x 20 µm, 10 µm x 10 µm and 20 µm x 20 µm, from top to bottom.

The G (left) and H (left) show the Fast Fourier Transform (FFT) of the AFM images with hierarchical pattern combining 2000 grid mesh (side width: 7.5 µm) and DVD channel patterns at both of direct imprint and replica molding conditions, respectively. The G (right) and H (right) show the line profiles of two micro repeat units corresponding to the 2000 grid mesh imprinting case at both of direct imprint and replica conditions. The I, J and H, three plots show the height changes of micropattern, UVO blocked and unblocked nano- patterns at mold and replica (a, b, c, d, e and f, marked in AFM images accordingly). The data is calculated from AFM images of 300 hexagonal mesh grid/DVD hierarchical

xvii

patterns. Uncertainties in plots indicate the standard deviations in repeated measurements.

...... 68

4.4 AFM images of SAM transferred onto Si from the PDMS mold with original and negative patterns. The structures in red and blue squares show the zoom in features of the ordered nano channels...... 70

4.5 SEM images of silver nanoparticles replicating the mold micro-nano hierarchical structures...... 72

4.6 Microscopic and AFM images of PS patterning with different Mw and film thickness.

...... 73

4.7 Line profiles of micro units at PS films with thickness of 87 nm, 390 nm and 626 nm.

...... 74

4.8 Analysis based on the AFM pictures. a, notation description of H, hm, hb. b, the plot of the height of micro protrusion (hm), and nano pattern at boundary of 2 micro units (hb1) and

4 micro units (hb2) as a function of as cast PS film thickness. c, height/depth comparison of PDMS mold depth and PS micro pattern protrusion...... 75

4.9 Mechanism discussion for the thermal expansion induced thin film ceiling touching phenomenon...... 77

4.10 AFM line profiles of the surface topography at PDMS micro molds before and after thermal expansion...... 77

xviii

4.11 Schematic illustration of Lithographically Induced Self-construction (LISC)...... 78

4.12 (a) AFM height images of PS with film thicknesses of 87 nm, 240 nm and 532 nm.

The red and blue regions show the topography of channels on top of the micro units and at the surrounding of the micro units. (b) Channel features on top of micro units after CFL with gradual annealing. (c) Channel features on top of micro units by using a neat PDMS mold (generated by dual replication of the original PDMS/SiOx hybrid mold). (d) Line profiles of the channels on top micro units, with original PDMS/SiOx hybrid mold, by gradual oven annealing, and by using a neat PDMS mold (dual replication), respectively.

...... 80

4.13 Schematic of the cold/dynamic zone annealing system...... 82

4.14 AFM images of nano channel at micro protrusion of PS (Mw = 161,200 g/mol) and

PS-PMMA (Mw = 57,000 – 25, 000 g/mol) in oven annealing and DZA conditions. .... 82

5.1 Synthesis of Zn-SEPDM...... 88

5.2 (a) Chemical composition of Zn-SEPDM, (b) Microstructure of SMP-FA. The permanent network forms from the aggregation of the zinc sulfonate ionic groups (red circles), which physically crosslink the ionomer and the temporary network consists of microcrystalline FA which also serves as supramolecular crosslink junctions due to strong dipole-dipole or ion-dipole interactions between the ionomer and FA (see arrow). This figure is not shown in scale...... 88

xix

5.3 (top) Schematic of the fabrication of the PDMS nano-mold and AFM images of DVD master and nanopatterned PDMS. (bottom) Schematic of the surface shape memory experiment: (a) nano-channel pattern transfer from PDMS to SMP-ZnSt using solution casting; (b) melt fabrication of temporary, crosshatched nanopattern at Tc = 100 ºC; (c) recovery of permanent pattern by reheating unstressed temporary nanopattern to Tc = 100

ºC...... 91

5.4 Thermal behavior of SMP-x compounds shown by DSC (Perkin Elmer DSC 8500) heating thermograms with a temperature gradient of 10 ºC/min...... 92

5.5 Tensile storage modulus measured at f = 1 Hz by DMA (TA Instruments Q800). ... 93

5.6 Shape memory cycle for a SMP-ZnSt film. The numbers indicate the distinct steps of the cycle as described in the text and the strains correspond to the values used in equations

(1) and (2). Path 1: The sample was heated to 100 ºC with an applied tensile stress of 1.5

Pa to prevent the sample from sagging. Path 2: The applied stress was increased to 0.2 MPa and the sample was allowed to equilibrate. Path 3: The deformed sample was cooled to 20

ºC under a constant stress of 0.2 MPa resulting in a strain 휀푚. Path 4: The applied stress was removed to fix the free-standing temporary shape with a strain 휀푢. The sample was then reheated without any applied stress to 100 ºC to remove the temporary network (path

5) and cooled to 20 ºC to recover the permanent shape, strain 휀푢...... 94

5.7 Five consecutive shape memory cycles for SMP-ZnSt; The function of temperature

(black curve), stress (red curve) and strain (blue curve) with time...... 96

5.8 Micropattern shape memory demonstration for SMP-ZnSt: (a) original surface imprinted micropattern (permanent pattern), (b) temporary pattern, (c) recovered pattern.

Top photos are optical micrograph and the bottom photos and graphs are the AFM...... 99

5.9 Line profiles of the AFM images for the original, temporary and recovered micropatterns for SMP-LA, SMP-SA and SMP-ZnSt. The original pattern is the same as shown in Figure ...... 100

5.10 Nanopattern shape memory results for SMP-ZnSt. (a, b, c) optical micrograph and

AFM image of (a) the imprinted grating pattern (p-direction), (b) crosshatched temporary pattern and (c) recovered pattern; (d) line profiles corresponding to the solid or dashed lines on the AFM image for the permanent, temporary and recovered patterns; (e) stress distribution of temporary shape calculated by finite element analysis (ANSYS)...... 103

xxi

CHAPTER I

1 INTRODUCTION

Hierarchical structures are ubiquitous in nature, ranging from the surface structure of taro and other plants to the feathers of birds and the foot-hairs of geckos.1-3 These highly organized materials often have specific functional properties that are derived from their multiscale structure, e.g., superhydrophobicity, self-cleaning, water resistance, transport for membranes and the capacity to sustain high loading forces.4-6 Traditional methods used for the fabrication of hierarchical structures typically involve the formation of complex patterned features through multistep lithography processes7-10, that includes photolithography, electron-beam lithography, soft lithography, and ultraviolet (UV) photolithography.11,12 These methods usually require expensive equipments and specialty reagents, so that the low cost fabrication of well-controlled micro-nano patterns for diverse potential applications remains a challenge.

This dissertation is aimed at the technology development, fundamental study and application exploration of micro-nano hierarchically pattern polymer surface. The objective is to propose a method to pattern surface hierarchically by a cheap and easy- processing technique and understand the physical mechanism and factors influencing 1

during the processing. Furthermore, the generated pattern is used to investigate the potential applications on soft lithography. In addition to general thermoplastic materials and “ink” materials, the pattern is applied onto shape memory materials to understand the different shape memory effect on micro and nano scale patterns comparing to the bulk system.

In this dissertation, Chapter 2 is presented to discuss the background information of the scientific topics relates to my work. It’s presented in 3 categories: UVO lithography,

Soft lithography and shape memory behavior. UVO lithography part provides the information of processing procedures and different working mechanisms involved in using different photosensitive materials. The vast applications of UV Lithography in semiconductor industry are introduced. This chapter also reviews the trends and recent progress of UV lithography in critical dimension shrinking. In addition, the exist challenges relating to the material selection and processing modification are particularly discussed.

Soft lithography part is firstly introduced by comparing the pros and cons by using soft molds. Subsequently, it’s explained by focusing on the 3 most actively studied forms: replica molding, microcontact printing and capillary force lithography. Processing procedures, potential applications and focus of research are discussed for each type of soft lithography. The discussion on the part of shape memory behavior is started from the concepts and basic working mechanism of shape memory effects. The 4 aims of the current study on shape memory composites are fully reviewed. Particularly, I reviewed the recent research works about the shape memory function on micro and nano scale topographies.

Chapter 3 presents the research work on the methodology of the novel dual- lithography for a dimension controlled micro-nano hierarchical patterning. In 2011, Xue et al.13 developed a method for direct patterning of PDMS films by simply exposing the films to UV radiation in air (UVO). The present work builds on this promising work by combining this approach with nanoimprinting. 13 Specifically, we modify the method by combining imprint lithography with a UVO exposure step that enables the fabrication of hierarchical film patterns with a large depth relief over large areas. The factors influencing the efficiency of patterning, including the photomask dimension, UV exposure energy,

PDMS thickness, surface energy and modulus are investigated to provide a full map of the controlling parameters for this patterning process.

Chapter 4 extends the research by applying the generated micro-nano hierarchical pattern for different forms of soft lithography applications. Here, the applications on replica molding, microcontact printing and capillary force lithography are explored. The special chemistry of the PDMS/SiOx hybrid brings pros and cons during the Soft Lithography applications. For demonstration of replica molding, PDMS prepolymer is used as model material. The “ink” materials used for microcontact printing are self-assembled monolayer and silver particles. The favorable patterning results in these two lithographies may due to the soft PDMS remaining hard shell at UVO exposed regions. In the process of capillary force lithography, defects are observed, which may result from the air trapping during the patterning process. However, the defects failure is fully investigated and eliminated by using a dynamic zone annealing (DZA) technique.

Chapter 5 investigates the micro and nano scale shape memory behavior based on a family of novel shape memory composite materials. Three different FAs, zinc stearate

(ZnSt), stearic acid (SA) or lauric acid (LA) are mixed with Zn salt of sulfonated EPDM, respectively. Here, the FAs function as temporary networks, and Zn-SEPDM acts as a permanent network. A shape memory system with different transition temperatures is obtained due to the different melting temperature of different FAs. The shape memory recovery efficiency of the micro and nano scaled surface topography is compared with bulk materials. A relative low recovery efficiency is observed due to that the stored elasticity is consumed by the dominated surface energy increase at micro/nano recovery.

Finally, chapter 6 summarizes the research findings presented in chapters 3, 4 and

5. The importance of this systematic work is emphasized.

CHAPTER II

2 BACKGROUND

2.1 UV Lithography

UV lithography is an actively studied patterning method used in microfabrication to create surface micro or nano structures on polymer thin films or bulk substrates.

Generally, UV lithography can also be termed photolithography, as UV wavelength is the most typical radiation used to sculpt the patterns. In terms of standard procedures, UV lithography involves several typical procedures: substrate cleaning, substrate surface preparation, photoresist application, UV exposure, developing, etching, and photoresist removal (Figure 2.1). UV Lithography patterns polymer films by taking advantage of photosensitive polymer material, on which nanopattern is created by transferring the micro or nano pattern on a photomask to the photosensitive polymer by irradiating UV light.

Photoresist polymer is crosslinked/polymerized at the UV exposed regions, while the UV shielded part of polymer film is developed by using a corresponding photoresist developer.

This kind of UV curing polymer is termed negative tone photoresist (such as SU-8, and epoxy), which is quite widely used in semiconductor pattern generation, because it’s cost saving, wet chemical resistant and silicon-adhesive. A major shortcoming of negative

photoresist is that the exposed regions swell easily as the counterpart is dissolved by the photoresist developer, which compromises the resolution of the process slightly.14 Positive photoresist (such as PMMA) works reversely, as the UV exposed area is to be removed.

For positive photoresist, UV exposure changes the chemical structure of the resist so that it becomes more soluble in its corresponding developer. In a lot of cases, especially in antenna and RF circuits Industry, current practice in the photolithography mainly relies on positive resists since they present higher resolution than negative resists.15 The figure illustration of the working mechanisms of positive and negative photolithography is shown in Figure 2.2 in detail. Chemical etching often accompanies photolithography as the process of fabricating silicon or metallic patterns using photoresists and etchants etch out selected areas corrosively. It has gained a wide popularity since it could produce highly complex patterns with high resolution 14, 15.

Figure 2.1 Process flow of photolithography using negative photoresist (not to scale). 16

(Reprinted with permission from Reference 16)

Figure 2.2 Process of photolithography by using negative and positive photoresists.15

(Reprinted with permission from Reference 15)

UV Lithography is one of the most successful technologies in microfabrication,12 as it has boosted the semiconductor industry in past decades to meet the Moore’s law. Until now, nearly all integrated circuits are made by this technology. The features size of the IC patterns has shrunken to ~ 10 nm by utilizing UV light with smaller wavelengths: from 436 nm, 365nm, to excimer laser 248 nm, 193 nm, 157 nm, 126 nm, up to the most recent extreme UV lithography with wavelength of 13.5 nm.12

To improve the resolution of the generated patterns, industrial approaches are mainly explored on using radiation or charged particles17 of shorter wavelengths. However, a lot of the recently developed techniques are proposed to generate nanoscale patterns by using visible light to cut the expense18. For example, Near-field scanning optical microscopy (NSOM) takes advantages of a contact mask incorporating with photochromic molecules to enhance the resolution by overcoming the light diffraction limits.19-22 The working mechanism is shown in Figure 2.3. Photochromic exists two different conformations that have different light absorptions (Figure 2.3 a). In addition, the red- absorption form could be converted to the UV absorption form under red light exposure.

During operation, 2 visible lights with 2 different wavelengths λ1 and λ2 are exposed onto a photochromic film upon a layer of photoresist film, as shown in Figure 2.3 b. As a consequence, only the area at the peak of λ1 allows the transmission of λ1 wavelength and further exposing and curing the photoresist underneath. Thus, the reduced nano-sized features could be generated.

Figure 2.3 (a) Visible light absorption of the open and closed forms of the photochromic material used for absorbance modulation photolithography. (b) Figure of the creation of a 8

controlled, reprogrammable, near-field aperture using NSOM. 21 (Reprinted with permission from Reference 21)

In addition to NSOM, multiphoton absorption polymerization (MAP)23-25 is another mature approach to super-resolved photolithography. MAP takes advantages of nonlinear interactions with light by absorbing two or more photons, even with none of which have sufficient energy to excite the photo-initiator by its own. As a result of this kind of nonlinear interaction, the final excited regions could be significantly smaller than the wavelength of light employed. For example, Li et al.26 successfully achieved λ/20 resolution by irradiating a 800 nm light, as shown by Figure 2.4.

Figure 2.4 SEM images of voxels created by MAP.26 (Reprinted with permission from

Reference 26)

However, with the requirements of dimensional feature size reaching a few nanometers, photolithography remains a challenging process to not only safe the prominent

role of etching, but also pattern surface with significant height (sometimes over tens of microns). The details of the challenges and major issues photolithography involves are presented in Table 2.1.27

Table 2.1 Summary of resist challenges and imaging challenges. 27

Resist Challenges Imaging Challenges Step Coverage Depth of Focus Topography Resist integrity over Quality images over significant steps significant steps Thickness Uniformity CD Control Thickness ctontrol on Linewidth cotrol across varying topography significant tipography Pattern Transfer Etch Resistance Image Quality Resist integrity under Quality images through thick aggressive etch conditions resist Adhesion Resolution in Thick Resist Resist adhesion to prevent Step edge profiles for fine undercutting geometries in thick resist

2.2 Soft Lithography

Soft lithography is a family of techniques for fabricating micro or nano structures using elastomeric stamps, molds, or conformable photomasks.28 Poly (dimethyl siloxane),

(PDMS) is one of most typical mold materials.12 Soft Lithography is termed mainly because soft mold is used as photomask during the applications of pattern replication or creation. Due to the soft nature of mold, soft lithography shows a lot of strengths including29: (1) Cost saving. Soft mold is cheap and it could be easily replicated from one expensive master, which significantly reduces the cost of the master template fabrication.

(2) Conformal contact. With conformal contact, external pressure is not mandatary anymore even when printing with an undulated substrate. (3) Insensitive to particle contaminants. Particles trapped between mold and substrate is less problematic as the soft mold could locally deform around a particle avoiding damage to the mold or substrate which leads to an improved yield of the process for the extended applications. (4) Avoiding anti-adhesive layer. Non-stick layer is not necessary during Soft Lithography application due to the relative low surface energy of flexible mold materials. For example, the surface tension of PDMS is around 20 mJ/m2.30 (5) Decreased imprinting and demolding force. (6)

Gradually and sequential micro-contact and “peel-off” separation during processing.

However, soft Lithography also has some intrinsic drawbacks.12, 29, 31, 32 (1)

Deformation and distortion of soft molds. Due to the relatively low Young’s modulus, soft molds show significant deformation under pressure, which remains a major issue limiting the resolution, uniformity and reproducibility of the imprinted patterns. High-aspect-ratio structures tend to collapse even much severely. The most typical defects modes are shown in Figure 2.5: sidewall collapse, roof collapse, buckling and lateral collapse.33 Different forms of mold failure not only result from the different dimension or aspect ratio of soft mold, but also the improper application pressures. Thus, mold deformation remains a challenge for further potential applications. (2) Poor dimensional stability of the patterned film. Due to the poor solvent resistance as well as the deformation induced by pressure and thermal expansion, the dimensional stability of printed patterns is degraded. (3) Mold aging.

Since the hardness and resistance to solvent are poor, soft molds have a relatively short lifetime. These limitations must be solved and overcame for extensive applications.

（a）（b）（c）（d）

Figure 2.5 Micro mold collapse modes. (a) Sidewall collapse. (b) Roof collapse. (c)

Buckling. (d) Lateral collapse. 33 (Reprinted with permission from Reference 33)

Generally, there are different types of soft lithography when using different experimental assembling geometries. In the following sections, the most typical techniques of soft lithography focusing on the basic mechanism, typical flow of processing and the range of applications will be discussed.

Replica Molding

Replica molding is an efficient technique for duplicating the shape and size of the pattern features on a soft mold/master, which provides patterning with a wide range of polymer materials. Contrary to photolithography, replica molding replicates 3-dimensional structures in one single step. Figure 2.6 outlines the procedures of replica molding. The

PDMS molds are prepared by casting against rigid masters using a standard PDMS thermal curing procedure. The pattern morphology on PDMS mold can, in turn, be replicated with relative high resolution by using this patterned PDMS as a mold to form structures in a second UV-curable (or thermally curable) prepolymer, even with a shrinkage of < 3% after

thermal curing or UV curing. In the process, a prepolymer is firstly deposited on the PDMS mold by casting or spin coating. Following that procedure, the prepolymer is thermally cured or UV cured, and it’s subjected to separation from the master by peeling them apart.

The relief structures on the replica are complementary/negative to those on the mold and very similar in size to those on the original master. 12, 34

Figure 2.6 Procedure of Replica Molding.34 (Reprinted with permission from Reference

34)

In addition, the shapes and sizes of the surface features at a PDMS mold could be manipulated in a well-controlled way by deforming and distorting the mold through 13

mechanical compression, bending, stretching, or a combination of these deformations35. In these approaches, the relief features on the surface of PDMS molds are deformed mechanically and then replicated using cast molding. If desired, multiple shapes and shrunken sizes could be successfully obtained by using a single PDMS soft mold.

Currently, the studies on replica molding are mainly focused on the generation of pattern with high aspect ratio/resolution36-38 and exploration of potential applications by combining with other techniques39-41. For example, Zhang et al.36 investigated the influence of material stiffness on pattern collapse as the aspect ratio of pattern is above 6. When a stiffer material, such as polyurethane and epoxy is used, a pattern with higher aspect ratio could still be generated with no lateral collapse observed. The SEM images of Si master,

PDMS mold and the replicated epoxy nanopillars are shown in Figure 2.7.

Figure 2.7 SEM images of nanopillars and nano holes at different processing steps of replica molding. First row: nanopillars on silicon masters. Second row: nano holes at

PDMS molds replicated from corresponding silicon masters. Third row: nanopillars at epoxy molded from the corresponding PDMS molds in second row. 36 (Reprinted with permission from Reference 36)

Zhao et al.39 proposed a method to generate micro-nano hierarchically patterned superhydrophobic surface with polyimide by combining replica molding and layer by layer assembly. They successfully obtained a surface with low surface energy and bio-mimic hierarchical structure as shown in the Figure 2.8 left. Figure 2.8 right shows the resulting static water contact angle reaching 160º and sliding angle of less than 10°.

Figure 2.8 (left) SEM image of the generated micro-nano patterns composed of microspheres and [(PAH/PAA)(PAH/SiO2)3]n films. (right) Profiles of a water droplet on the generated film, and sliding on the surface. 39 (Reprinted with permission from

Reference 39)

Microcontact Printing

Microcontact printing, another form of soft lithography, works straightforwardly.

It transfers relief patterns from the surface of a patterned PDMS stamp to the surface of a substrate, which enables the transfer of various ‘ink’ materials, including self-assembled monolayers (SAMs) 42, 43, and small molecules, such as quantum dots44-46, nanoparticles47-

49 and even biomolecules, proteins, polyelectrolytes and suspensions of cells34, 50-52. In microcontact printing, only the contact region between the PDMS stamp and a substrate transfers the ‘ink’ from the protruded surface of the relief features of the stamp to substrate, and generates patterns with feature sizes as small as 100 nm over areas as large as ~1 m2.34

Figure 2.9 shows the procedure for the microcontact printing of “ink”. The soft nature and elastic property of the PDMS stamp enable conformal contact between the stamp and the substrate, which assists the pattern transfer even with non-planar surfaces (including rough and curved surfaces). Self-assembled monolayers (SAMs) pattern transfer is a large category in microcontact printing, which is also one of the most intensively studied materials for microcontact printing. Generally, SAMs are formed when PDMS stamps

‘inked’ with functionalized end group, such as –OH, –NH2, –COOH, or –SH groups, which further varies the wetting and interfacial properties of the target substrates. SAMs are organic molecules produced by assemblies formed spontaneously on surfaces by chemisorption or self-organization and are organized into ordered domains. SAMs could also provide surface functions in demand by controlling the end functional groups. For example, Alkanethiols (SH-(CH2)n-X), with long alkyl chains (n = 16 or 18) form hydrophobic monolayers, whereas those with different terminal functional groups (X) can

form hydrophilic, hydrophobic or charged SAMs on surfaces of gold, silver, palladium, platinum or other metals34, 53. Many synthetic methods are available for attaching ligands to functional groups (X) on SAMs to further adjust the surface chemical properties of the patterns on objective substrates.53 In addition, microcontact printing can also be used to pattern surfaces with multiple SAMs, especially for generating islands of SAMs that adsorb proteins and cells, and which are surrounded by SAMs that resist the adsorption of biomaterial. 54, 55 PDMS stamps have also been used to create patterns of various “inks” onto glass, polystyrene or silicon for different applications51, 56, 57. SAMs could be used to tune the wetting behavior of surfaces, create the secondary structures at another material,58 act as an etch resist,59 or mimic the natural extracellular matrix of cells.60

“Ink”

PDMS stamp

Invert stamp; bring into contact with substrate

Substrate

Peel away PDMS stamp

‘Ink’

Figure 2.9 Procedure for Microcontact Printing. 34 (Reprinted with permission from

Reference 34)

Microcontact printing is a high throughput technique (especially when using a rolling stamp, or cylindrical61) and it’s applicable to many different types of materials. For example, Geissler et al.59 proposed two different methods to generate metallic nanowires by microcontact printing: subtractive processing and additive processing, as shown in

Figure 2.10 . Figure 2.10a illustrates the process of subtractive etching the unprinted regions to generate patterns. Figure 2.10b illustrates the printing of seeding solutions or catalysts for electroless deposition from solution. Figure 2.10c-f show nanowires of gold, silver, copper, and palladium prepared by subtractive processing.

Figure 2.10 (a) Subtractive procedure used to print a SAM on a metallic film. Pattern is defined by a subsequent etching process. (b) Additive process used to print seeding solutions or catalysts that forms nanowires using electroless deposition (ELD). (c–f)

Atomic Force Microscope (AFM) images of gold, silver, copper, and palladium nanowires formed by subtractive procedure as shown in a.59 (Reprinted with permission from

Reference 59)

Capillary Force Lithography

Capillary Force Lithography is a special type of lithography for micro molding, which only relies on capillary force to drive material flow. Before Capillary Force

Lithography is introduced, Micromolding in Capillaries(MIMIC) is proposed in the early

1990s.62, which was considered as a complementary to photolithography as MIMIC enables the fabrication of microstructures with various materials, including organic polymers, inorganic and organic salts, ceramics, metals and crystalline microparticles.62

MIMIC relies on the flow of a dilute polymer solution inside micro cavities due to the conformal contact between a stamp and a flat bare substrate, as shown in Figure 2.11. In contrast to MIMIC, Capillary Force Lithography (CFL) is a relatively new method proposed in 2001 by Suh et al.63, which requires a previously coated polymer film on the substrate. CFL created patterns by using a typographically patterned layer of PDMS as a mold. The patterned PDMS mold is brought into conformal contact with polymer thin film and then it’s subjected to an elevated temperature (T>Tg) to enable the mobility of polymer to fill in the PDMS mold. Here, capillary force is the only driving force for the mold filling.

After mold filling, the polymer film is suddenly cooled down to fix the patterns. By peeling the mold away carefully, a pattern of micro or nano patterned polymer is left on the substrate surface.34 The process of Capillary Force Lithography is shown in Figure 2.12.

Substrate

Add prepolymer and cure

Prepolymer

Remove PDMS channels

Polymer microstructures

Figure 2.11 Process of Micro molding in Capillaries(MIMIC). 34 (Reprinted with permission from Reference 34)

Figure 2.12 Process of Capillary Force Lithography.

CFL has been actively studied during the past decade because it combines the essential feature of imprint lithography (molding a polymer melt) and soft lithography

(using an elastomeric mold), making it an alternative choice to lithographies for large-area patterning64. At the temperature above the glass-transition temperature (Tg) of the polymer, capillary force drives polymer fluid to fill in the elastomer mold. The initial polymer film thickness and the feature dimensions of the mold determine the type of the resulting

polymer patterns. When the polymer film is thick enough to completely fill in the cavity of the mold, a residual polymer layer remains on the substrate. However, if the polymer film is thin, the polymer meniscus shape is observed after patterning due to the insufficient polymer filling in the voids, which eliminates the need for etching steps to remove residual layer63, 65 (Figure 2.13). The advantages of retaining incomparable pattern fidelity from imprint lithography, and non-pressure imprinting, makes CFL an attractive technique for fabricating patterns in one step.63, 64, 66, 67 However, the thermally expandable nature of elastomer (such as PDMS) prevents the CFL from being an attractive technique for fabricating highly accurate patterns, especially in a large area.12 Besides, the sagging of

PDMS as PDMS mold recession width is much bigger than the pattern depth, 푑 ≫ ℎ, caused by compressive forces between the stamp and the substrate contributes to the defects in the pattern.12 The defects molds have been discussed in detail in Figure 2.5.

Figure 2.13 Figure diagram of capillary force lithography (CFL): a) when film is relatively thick with respect to the mold pattern depth (excessive material for mold filling) and b) when polymer film is thin. 63 (Reprinted with permission from Reference 63)

2.3 Shape memory behavior

Shape memory polymers (SMP) are smart materials that can be programmed into a temporary shape, but can subsequently recover their original shapes when exposed to an external stimulus, such as heat, light, moisture, pH or an electrical or magnetic field.68, 69

Shape memory requires two distinct networks. One network provides a permanent shape, that is crosslinked with either covalent bonds or supramolecular bonds that have a relaxation time much longer than the characteristic time of the shape memory behavior.

The second network must be reversible upon application of a suitable external stimulus, such as thermal energy, that functions as the temporary network that supports a temporary shape. For efficient shape memory effect, the temporary physical network must be susceptible to an external stimulus that disrupts the supramolecular bonds and activates the molecular mobility, yet sufficiently robust that it suppresses permanent deformation, i.e., creep.70

Among these stimuluses, thermal triggered SMP is a typical one, which has been actively studied in academic and used in industry. Figure 2.14 shows the molecular mechanism of thermal sensitive SMP. Thermal sensitive SMPs usually have a physical cross-linking structure (crystalline/amorphous hard phase) or chemical cross-linking structure as permanent networks, and crystalline, amorphous or liquid-crystal phase acts as switching temporary networks. Generally, SMPs are firstly processed or thermally set to have an original shape, with zero internal stress unreleased. If the SMP is deformed, large internal stress can be stored in the permanent cross-linking structure by cooling the polymer

to the temperature below its switch temperature (Tc). With some interfacial coherence between temporary (hard solid at the temperature below Tc) and permanent network, the creeping of permanent work is largely suppressed. By heating the polymer above the switch transition temperature (Tc), the SMP recovers its permanent shape as a result of releasing internal stress stored in the cross-linking structure.71 Figure 2.15 shows the detail of the process and molecular mechanism of 3 different thermal sensitive shape memory effect.68

Comparing to shape memory ceramics and shape memory metallic alloys (SAMs), SMPs have the advantages of lightweight, low-priced72, highly processable73, switch temperature adjustable73-75, high shape deformability and recoverability71, 76, 77. Those characteristics enable the potential applications in biomedical devices78, 79 electronics80, self-healing properties81-83, smart adhesives84, deployable structures85 and functional textiles86.

Figure 2.14 Molecular mechanism of thermal SMPs. 71 (Reprinted with permission from

Reference 71)

Figure 2.15 Illustration of the molecular mechanisms of the thermal induced shape- memory function for a) a multiblock copolymer, where Tc=Tm, b) a covalently crosslinked polymer, where Tc=Tm, and c) a polymer network, where Tc=Tg. 68 (Reprinted with permission from Reference 68)

Current studies on SMPs composite/blends are mainly focused for 4 aims71:

(1) to improve shape recovery stress and mechanical properties; Comparing to the high mechanical strength of shape memory alloy (SMAs) (700 MPa-2000 MPa)87-89, the mechanical strength of SMPs (in the range of 5 MPa-100 MPa)90 need to be largely 27

strengthened. The recovery stress of SMPs is relatively low comparing to SMAs as well.87,

91, 92 Generally, the typical methods to reinforce SMPs is mix them with high modulus organic or inorganic fillers, such as microfibers93, 94 carbon nanotube (CNT)95, 96, nanoclay97, nano SiC, graphene98, 99, carbon black (CB) 100-102, and other organic fillers, such as polyhedral oligomeric silsesquioxanes (POSS)103-105, and cellulose106, 107.

(2) to decrease shape recovery induction time by increasing thermal conductivity;

Due to the organic nature of SMPs, SMPs are thermally insulated with a thermal conductivity usually below 0.30 W/mK108. Rapid heating and cooling of SMPs are required to deform or fix SMP to largely suppress the creep of the permanent network, which remains a big challenge to SMPs. Fillers such as metal particles and inorganic fillers including alumina109, 110, fused silica111, SiC112, 113, boron nitride (BN)85, 114, carbon nanotube115, 116, and glass fiber117, with heat transfer properties, are investigated to increase the thermal conductivity of SMPs.71

(3) to create new polymer/polymer blends with shape-memory effect (SME); In the past decades, shape memory polymer/ polymer blends have been actively studied to generate novel shape memory systems with improved properties comparing to the available

SMPs. The mixing of a SMP with a miscible polymer could be easily used to adjust the switching temperature and mechanical properties of the SMP. For example, Jeong et al.118 proposed that by blending the poly(caprolactone) (PCL) segmented thermoplastic polyurethane (TPU) (acting as the soft segment), and hexamethylene diisocyanate

(HDI)/1,4-butane diol (BD)/4,4′-dihydroxy biphenyl (DHBP) based hard segment with phenoxy resin, shape memory effect could be achieved. In addition to this new generation 28

of SMPs in polymer blend system, the tunable transition temperature and strengthened mechanical strength could also be achieved by polymer blending. For example, Jeong et al.119 continued their study by mixing PVC with PCL segmented thermoplastic polyurethanes (TPUs) and they proposed that the transition temperature (Tc) of this miscible amorphous domain could be varied by tuning the blending composition.

(4) to generate shape memory materials sensitive to electricity114, 120-123, magnetic124-126, light127-130, moisture71, 131, 132, and other stimuluses71. Not only the thermal sensitive SMPs, other SMPs triggered by different stimulus is another very important development trends of SMP composite due to the vast applications in sensors and devices.

In addition to those 4 aims mentioned above, other novel functions such as multiple shape memory effect (SME) and two-way SME are also actively studied as it enlightens the potential applications for muscles and actuators.133-139. Even though these aims have been widely studied, more studies are needed for the relationship between the structure of

SMP composite and blend material, properties of the material, and requirements of specific applications.71

Nevertheless, the majority of work in this field has concerned bulk or macroscopic scale shape memory features, microscopic and nanoscopic scale pattern memory has recently been reported and has attracted great attention,140-144 such as 2D voids145 and 3D microwrinkles146 and microprotrusions141, 143, 147, 148 by making use of the efficient and large modulus change at the shape memory transition temperature Tc. The unusual properties of patterned or textured films, e.g., surface wetting149-151, optical properties141, surface

roughness152, friction147, and adhesion84, 153, have generated considerable interest in switchable micro- and nano-scale patterns and textures. However, the nano pattern recovery and micro pattern recovery are challenging. Because during the shape recovery process, the entropic/elastic energy stored in the deformed state is released, while the permanent micro/nano scale shape could not be easily recovered until the stored energy could fully compensate the increasingly dominant surface energy (induced by the increased surface area as the deformation size is decreased to micro and nano scales). So, until now, a small number of reports have been published showing the full recovery to the original shape from the micro patterns and nano patterns.140, 148

For example, Li et al.148, achieved colored-transparent-colored switchable film with a micro scale patterned shape memory polymer (a mixture with molar ratio of 5:1:3 of the melted diglycidyl ether of bisphenol A epoxy (EPON 826), poly- (propylene glycol)bis(2- aminopropyl)ether (Jeffamine D-230) and decylamine (DA)), for the potential applications of displays, privacy window and camouflages.154 Figure 2.16 shows the processing steps of pattern master fabrication, SMP pattern fabrication and SME demonstration. Figure

2.17 a, f, k shows that the pores patterned polymer is colorful due to the diffraction grating effect. After strain application, circular pores are deformed to elliptical slits. The holes are further closed and the color is further dismissed due to the strain is further applied up to

30 %. When the deformed film is reheated to 90 °C, it recovers to the original shape with

100 % of the original pitch and 97.6% of original hole size, as shown in Figure 2.17 j, o.

Figure 2.17 e shows that after pattern recovery, the color is recovered as well. Figure 2.17 p and q are another evidence showing the different optical diffraction from the surface of

the films with “Penn” logos placed underneath. The UV-Vis spectra shown in Figure 2.17 r indicates the transparency change at the different stages of a, b, d, and e, which confirms the visual appearance shown at Figure 2.17 q. Most importantly, the obtained color spectrum by simply tuning the mechanical deformation, could even probably be used to design a full-color display with variable structure parameters.148

Figure 2.16 Fabrication and deformation/recovery of a 2D SMP membrane. (a) Fabrication of a mold master by holographic lithography. (b) Fabrication of the SMP membrane with periodic holes by replica molding. (c) micro-holes shape memory demonstration. 148

(Reprinted with permission from Reference 148)

Figure 2.17 Pattern flat out and recovery in a SMP membrane. Optical images of the: (a) original; (b), (c) partially deformed with ε ~ (13 ± 2)% and (20 ± 2)%, respectively; (d) completely deformed with ε ~ (13 ± 2)%; and (e) recovered SMP membranes. (f–j)

Corresponding SEM images of the SMP membranes shown in (a–e). (k–o) Higher magnification SEM images of (f–j). (p) Transparency comparison between the original and the deformed SMP film. (q) Display of ‘‘Penn’’ logos underneath the SMP films. (r) UV-

Vis spectra of the SMP films at different steps corresponding to a, b, d and e. 148 (Reprinted with permission from Reference 148)

In another example, Wang et al.140 demonstrated programming and fully recovery of nano patterns on a range of nanoimprint lithographically created SMPs. They utilized acrylate based SMPs to generate line-and-space gratings by thermal embossing

Nanoimprint Lithography (TE-NIL). The whole process of the pattern generation, flatten out and recovery is shown in Figure 2.18. Figure 2.18 also shows the optical images, AFM images and profiles of the surface morphology at permanent shape (a, d, g), temporary shape (b, e, h) and recovered shape (c, f, i). Iridescent color dismiss is observed at the temporary shape due to that the majority of light is transmitted through the smooth film.

The iridescent color is caused by the similar phenomenon as discussed in the last example

(study of Li et al.148). Interestingly, regardless of programmed height and original permanent height, the final pattern height is recovered to the original height, as evidenced by Figure 2.18 j and k. Unlike bulk system, the degree of pattern recovery (푅푟) could be quantified by the equation: 푅푟 = 퐻푟푒푐표푣푒푟푒푑⁄퐻푝푒푟푚푎푛푒푛푡 , where 퐻푟푒푐표푣푒푟푒푑 is the recovered pattern height, and 퐻푝푒푟푚푎푛푒푛푡 is the height of the permanent pattern set at the beginning. From the fit line shown in Figure 2.18 k, 푅푟 = 100%, full recovery is obtained at varied permanent heights of channel patterns. Not only for one tier of pattern, dual- pattern or multi-pattern could also be obtained with nearly 100% recovery, as shown in

Figure 2.19.

Figure 2.18 Top: Figure of the programming and recovery steps. (a)–(k) digital images,

AFM images and AFM line profiles of the patterns at the permanent, temporary and recovered stages. (j) and (k) the plots of remaining pattern heights after programming

(Hprogrammed) and the recovered pattern height (Hrecovered) as a function of the permanent

pattern height (Hpermanent), respectively. (Reprinted with permission from Reference 140)

Figure 2.19 Top: Schematic of the programming and recovery steps with a second grating pattern. (a)–(c) AFM height images and line profiles of the patterns at the permanent, temporary and recovered stages. (Reprinted with permission from Reference 140)

Our collaborator, Dr. Robert Weiss started working on shape memory polymers about 10 years ago74 and he previously reported the development of shape memory function based on compounds of the zinc salt of a sulfonated poly ethylene-co-propylene-co-(5- ethylidene-2-norbornene) ionomer, Zn-SEPDM, and a fatty acid or fatty acid salt, FA74, 155,

156. In those materials, an ionic nanostructure provides a permanent network due to the long relaxation times of the ionic interactions within the ionomer phase and the FA formed a microcrystalline phase that provided a supramolecular, temporary network. Figure 2.20155 shows the molecular mechanism of SME comprising the temporary and permanent network.

The size of the microdomain at shape memory systems is a key factor as it may influence the surface morphology change especially when micro and nano scale patterns relate during the shape memory process. However, the details of the size of the microcrystalline FA phase are unclear, despite several attempts to characterize mixtures of Zn-SEPDM

ionomers and ZnSt using X-ray and electron microscopy157-159. The difficulty has to do with the small size of the crystallites, the presence of an ionic domain nanophase and artifacts such as crystals of excess neutralizing agent. The general opinion is that the crystalline ZnSt phase is poorly organized and the size is less than ~0.5 µm. Handlin159 concluded that the ZnSt crystals were thin lamellar flakes with thickness ~5 nm, but larger lateral dimensions. Duvdevani et al.157 attributed ~0.1 nm sized features in scanning/transmission electron micrographs to the ZnSt crystals, though they also observed

ZnSt crystals with sizes ranging from ~0.1 – 1 µm. Based on those studies, the relative clarity of the samples and the large melting point depression of the FAs in the compounds, we believe that the FA crystals in the compositions considered in this SMP system are predominantly less than ~0.5 µm.

Figure 2.20 Figure illustration of the permanent and temporary network at a Zinc oleate/Zn-SEPDM system. 155 (Reprinted with permission from Reference 155)

CHAPTER III

3 METHODOLOGY OF A NOVEL UVO LITHOGRAPHY

FOR A DIMENSION CONTROLLED MICRO- NANO

HIERARCHICAL PATTERNING.

Hierarchical structures are ubiquitous in nature, ranging from the surface structure of taro and other plants to the feathers of birds and the foot-hairs of geckos. 1-3 These highly organized materials often have highly specific functional properties that derive from their multiscale structure, e.g., superhydrophobicity, self-cleaning, water resistance, transport for membranes and the capacity to sustain high loading forces. 4-6 Traditional methods used to produce hierarchical structures typically involve the formation of larger features and smaller roughness in steps by lithography. 7-10 It includes photolithography, electron-beam lithography, soft lithography, and ultraviolet (UV) photolithography for making complex patterns. 11,12 These methods often involve expensive equipments and specialty reagents, so that the low-cost fabrication of well-controlled micro-nano patterns for diverse potential applications remains a challenge. In this study, PDMS is used for patterning process, largely due to the soft nature of PDMS which provides conformal contact and tolerance to

particle contaminants during patterning. It’s also a typical material used for soft lithography, which further expands the applications of surface patterns.

In our approach to the fabrication of hierarchical film patterns with a large depth relief over large areas, we have combined imprint lithography12 with a UVO exposure step13. In this method, nano-sized DVD patterns are initially transferred onto an elastomeric substrate (PDMS) through direct nanoimprinting, while in-plane and out-of- plane micro-sized structuring of the films are created by simply UVO irradiating the film through TEM grids with different hole configurations. The mechanism of how the wall stress built up at the boundary region of micro recession is systematically studied. In-depth understanding of the process of micro recession was attained by elucidating the influence of the frontal development (top-down reduced PDMS densification) of PDMS under UVO irradiation. To probe the fidelity and controllability of the nano-pattern preparation and imprinting, we performed a coarse grained molecular dynamics simulation, and established the universality of the pattern deformation as a function of the material property.

This methodology is proposed to understand:

➢ how UV light could be manipulated to control micro-nano 3D topography,

including pattern shape and dimension.

➢ what’s the role of PDMS elastics and surface energy plays during pattern

replication.

3.1 Materials

Poly (dimethyl siloxane), PDMS, Sylgard 182 is bought from Dow Corning, USA.

Digital Video Disk (DVD)s are products from Sony Corporation, Japan. CDs are bought from Magnavox, USA. TEM grids are bought from Ted Pella. USA, with models of G300HEX, G1000HS and G2000HS. The dimensional information for these models is shown in Figure 3.1.

Figure 3.1 models and dimension information of the TEM grids used in experiments.

3.2 Experimental methods

DVDs have channel pattern at the interface between polycarbonate and intermediate aluminum layer. The channel patterns have 120 nm height and 750 nm wavelength. To make DVD patterned PDMS, the surface polycarbonate layer and metallic reflecting layer in DVD were removed and then the residue patterned polycarbonate layer is rinsed with methanol. PDMS was prepared by hand mixing base and curing agent with a weight ratio of 10:1. Air bubbles trapped in the mixture were removed by exhausting in vacuum desiccators for 15 min. Then the PDMS liquid mix was poured onto the DVD

master, followed by curing at 120 oC for 2 h in an oven. The patterned PDMS was obtained by peeling off from the DVD master. (Figure 3.2) Elastomeric CD pattern was produced in the same way. The AFM height images for both of original and PDMS replica of CD and DVD are shown in Figure 3.3. The profile information from AFM shows that both of

CD and DVD replicas keep the pattern shapes and dimensions with very high fidelity. The

PDMS replica of the DVD template has an average height of 119.2 nm, with a standard deviation of 7.4 nm, which is nearly identical to the DVD template (wavelength = 750 nm,

height = 120 nm). The standard deviations in this dissertation are all calculated from repeated measurements. Evidently, we obtained a faithful pattern reproduction.

Figure 3.2 Schematic illustration of nanopatterning

Figure 3.3 AFM images of channel patterns at the original (A) and PDMS replica(B) of

CD and DVD.

Micro patterns were produced on DVD nano channel patterned PDMS, as well as

on smooth version of PDMS pads. This was accomplished by placing TEM grids on

PDMS surfaces. Subsequently, these were placed in a UVO chamber (PSD series,

Novascan Technologies, Inc.) and exposed to the simultaneous dual wavelengths of 185 nm and 254 nm. Micro-patterns were generated via consequent “sinking” of the regions

exposing to UVO. The experimental assembly is shown in Figure 3.4. The average thickness of PDMS used in this study is 1.22 mm, with a standard deviation of 0.031 mm.

Figure 3.4 Experimental assembly for micro patterning.

3.3 Micro-nano hierarchical patterns and parameters influencing the efficiency

of UVO Lithography

Chemistry of PDMS under UVO exposure

PDMS can be oxidized under UVO radiation with two dominant wavelengths: 185 nm and 254 nm. Mechanically, oxygen in the atmosphere is activated by the 185 nm UV

13 light producing ozone and the CH3 groups in the PDMS side chain are eliminated by the reaction with ozone under 254 nm UV.160, 161 In particular, the operative reactions and rationale for this reaction are discussed below:

In our experiments, flat PDMS was exposed to UVO radiation with an uncoated

TEM copper grid photomask upon it. Only the regions not shielded by the wire and grid were exposed to radiation. To understand the different chemistry of the photo-masked

PDMS at UVO exposed and blocked parts, XPS was performed. VersaProbe II Scanning

XPS Microprobe from Physical Electronics (PHI), under ultra-high vacuum conditions

with a pressure of 2 × 10-6 Pa was used for XPS spectra. Automated dual beam charge neutralization was used during the analysis of the samples to provide accurate data. The energy passed analyzer was 117.4 eV for the survey spectra and 23.5 eV for the high- resolution Si 2p scans. Each spectrum was collected using a monochromatic (Al Kα) X- ray beam (E = 1486.6 eV) over a 200 µm diameter probing area with a beam power of 50

W. Figure 3.5 shows the color image of C 1s intensity indicating the distribution of carbon ratio at 40 min UVO exposed PDMS surface (~10 nm depth) with a 100 mesh TEM grid

(230 µm side width) as a photomask. The higher carbon density is observed at UVO blocked regions with the white translucent lines as auxiliary lines. The Si 2p peaks in Figure

3.5 display a binding energy shift from UVO blocked region to UVO exposed region. The

Si 2p peak shift, from 102.27 eV to 102.84 eV, confirms the transition tendency, as the

162, 163 literature value of SiO2 is 103.3 – 103.7 eV . The details of atomic ratios at UVO exposed and blocked PDMS surface are presented in Table 1. After 40 min UVO exposure, the carbon ratio decreased from the original 43.1 %, which is close to the theoretical value

of 50 %, to 19.1 %, due to the chemistry of PDMS under UVO inducing evaporable carbon dioxide. Additionally, the ratio of Si to O increased to 1:2 approximately, which

corresponds to the expectation of the transformation from PDMS to SiO2 as discussed above.

Figure 3.5 (A) XPS color map of C 1s element at 40 min UVO exposed PDMS with 100 mesh TEM grid as a photomask, (B) XPS spectra of Si 2p level at UVO exposed and blocked PDMS surface.

Table 1 Atomic concentration (C 1s, O 1s, and Si 2p) determined with XPS of untreated

PDMS (theoretical and measured) and 40 min UVO exposed PDMS.

C (%) O (%) Si (%)

PDMS (theoretical) 50.0 25.0 25.0

PDMS (measured) 43.1 31.9 25.0

PDMS (exposed for 40 min) 19.1 54.3 26.1

In addition to the PDMS chemistry change before and after UVO exposure, the

3 3 transition from O-Si-C (density is 0.965 g/cm ) to SiOx (density is 2.64 g/cm for x = 2) results in a large density increase160, contributing to the depth increase of PDMS under

UVO exposure.160 Thus, PDMS behaves like a traditional dry positive photoresist in this

UVO densification method.160,164 The change of density of the polymer upon radiation is not the only factor relevant to creating surface topography under illumination. Changes in the local chemical composition associated with non-uniform UVO exposure will cause the change of surface energy of the film and we must then anticipate variations in surface 44

topography.161, 165-168 Previous studies show that the migration of low molecular weight species of PDMS from bulk to the surface causing the recovery of surface energy and surface hydrophobicity.169 In our experiments, no obvious surface depth recovery is observed, as determined by measurements of the structure after 3 weeks (unchanged). A combination of densification and interfacial energy variation is then responsible for the

UVO induced surface modification. We next quantify this patterning process to exert dimensional control over the surface pattern process.

Micro-nano hierarchical pattern

Figure 3.6 shows the results of the combined micro-nano hierarchical patterning in PDMS (cured directly on DVD pattern and then UVO exposed under a TEM grid). It shows the hierarchical PDMS patterns obtained after 4 h (accumulated energy: 85.3

J/cm2) of UVO exposure through the three TEM grids (as shown in Figure 3.1) after a common DVD nano-patterning step. Figure 3.6 (bottom) shows the Fast Fourier

Transform (FFT) of the AFM image of the hierarchical pattern from G2000 TEM grid

(side width: 7.5 µm). The planar orderings of the hierarchical patterns are clearly reflected in these FFT spectra of the AFM images. The 5 obvious high-order peaks indicate the pronounced channel orientation. The in-plane average channel distance of

757 nm (very close to channel wavelength obtained from AFM: 750 nm) was derived from the FFT, which also reveals the long-range uniformity of channel exceed by far the investigated area. The macro square patterns are arranged with high periodicity in both horizontal and vertical directions, according to the parallel dot matrixes. According to the line profile of the hierarchical pattern units generated from the G2000 TEM grid, we 45

observe that the UVO exposed regions become depressed by several hundred nanometers, yet quite surprisingly, the nanostructures created through the initial DVD patterning survive the deep UVO exposure densification process. Comparing to the original nano channel height (120 nm), shrinkages are observed at the UV exposed region, which is a consequence of the same UVO depression phenomenon. The AFM images of hierarchical pattern with CD channel as template are shown as well. (Figure 3.7)

Figure 3.6 AFM images of hierarchical PDMS pattern combining DVD with TEM grid patterns (300 hexagonal grid mesh, 1000 grid mesh and 2000 grid mesh) after a 4 h (85.3

J/cm2) UVO exposure. Insets are the structures with high resolution, with dimension of 20

µm x 20 µm, 10 µm x 10 µm and 20 µm x 20 µm, from top to bottom. The bottom shows the Fast Fourier Transform (FFT) of the AFM images with hierarchical pattern combining

2000 grid mesh (side width: 7.5 µm) and DVD channel. Line profile of two micro repeat units corresponding to the 2000 grid mesh imprinting case is shown.

Figure 3.7 AFM images of hierarchical pattern combining CD with TEM grid pattern (A1)

300 hexagon, (B1) 1000 mesh square, (C1) 2000 mesh square, with 4 h UVO exposure.

In practice, we even generated patterns with larger mesh sizes with this method, so there is no issue with the in-plane scalability of dimensions. This is illustrated in Figure

3.8, which shows the digital photographs of the hierarchical patterns with the combination of DVD nano-pattern and micro square grid patterns with side lengths of 500 µm and 250

µm, respectively (USA standard testing sieve from VWR, USA) after 4 h UVO exposure.

AFM scans of these larger patterns revealed that due to the “weaving” of the wire mesh at

the intersections of the sieve, the boundaries of square patterns are not sharp. The multi- color surface is due to the nature of light interference from DVD patterns.

Figure 3.8 Digital pictures of hierarchically patterned PDMS combining DVD pattern and square patterns with side lengths of (a) 500 μm and (b) 250 μm.

Parameters influencing the process of UVO induced PDMS depression

Figure 3.9A shows a range of PDMS topographies created by simply varying the

UVO photomask pattern size with the constant of 4 h UVO exposure (accumulated energy: 85.3 J/cm2). Three kinds of TEM grids: G300HEX, G1000, and G2000, with different shape and sizes were utilized. The detailed shape and dimension information is shown in Figure 3.1. The resulted PDMS have similar widths of (56.1 ± 2.4) µm, (18.5 ±

1.9) µm, and (7.5 ± 0.1) µm, respectively. The average pattern depths (D) of these three micro patterns are 1165 nm, 600 nm and 355 nm, respectively. Figure 3.9B shows, interestingly, a linear scaling relationship exists between the PDMS depression depth (D) and pattern width (W), D = (15.8 W + 244) nm, where the depth (D) is calculated from the differential of the deepest part to the top surface of UVO blocked part, and the width

(W) is micro depression width in μm. Error bars are the standard deviation of repeated 49

measurements. Although the depth of the depression created by UVO patterning is nearly linear over the range of our observations, the Y-intercept (244 nm) of this extrapolated line would not be zero, implying that this micro-pattern scaling must break down at submicron pattern length scales, as shown by black dashed line extrapolated to zero in

Figure 3.9B, left side. However, detailed analysis of that regime is out of the scope of our present study. At the other extreme of “non-confinement”, we placed a smooth silicon wafer on a smooth PDMS pad to create a semi-infinite boundary. The edge height across this single step-boundary gave a UVO exposure saturation depth of 1150 nm. We then extrapolate pattern width of Figure 3.9B to this maximum saturation depth, and determine that it would correspond to a micro-pattern width of 57.3 µm. It means that PDMS sculpted through micro-pattern mask widths above 57.3 µm (about G300HEX mesh size) should then be free of finite size effects, i.e. the linear scaling relation will no longer apply.

An examination of the shape of these depression offers further evidence relating to the physical origin of this pattern formation: a balance of elastic wall restoring forces and the pined densified glassy base layer. If the shapes of the depressions of the patterns having a given symmetry are rather independent of TEM grid pattern dimensions, it would suggest a dominant role of a balance of elastic energy and pinning force between the UVO densified base layer and non-UVO exposed wall regions. We quantify this superficial impression in normalized plots to compare PDMS trench depth profile shapes. As a function of reduced dimensions, we scale our AFM observations of the pattern depth (Z) normalized by the maximum depth of the pattern (Zmax), and the in-plane coordinate of the pattern width W

normalized by grid pattern width (Wpattern). Figure 3.9C shows that the shapes of all these cavities are indeed rather similar, except near the edge of these patterns. We see that the edge ‘irregularities’ in the UVO micro depression patterns extend beyond the edge of the

TEM mask and they become progressively smaller on a relative height scale as the mask patterns become larger. This secondary patterning at the edge of the pattern probably reflects the backflow of material associated with the formation of the micro depression pattern upon UV exposure, but this effect must be convoluted with the scattering of UV light under the mask near the edges of the mask holes.

Figure 3.9 (A) AFM images of flat PDMS irradiated under UVO for 4 h through a TEM grid with different patterns: 300 HEX mesh (57 µm side), 1000 square mesh (19 µm side),

2000 square mesh (7.5 µm side). (B) UVO unnormalized micropattern depth of the

UVO micro patterning as a function of in-plane dimension. (C) normalized micropattern depth of the UVO micro patterning (from top surface to center of pattern).

Next, the influence of cumulative UVO exposure on DVD nano-patterned PDMS is considered to further explored the extent to which the micro topography can be controlled. The red and black curves in Figure 3.10A show the depths of the UVO exposed

PDMS region with G1000 TEM grid as photomasks, as a function of UVO energy. The main feature of Figure 3.10A is the non-monotonic change of micro-pattern depth with

UVO exposure. We conjecture that the depth increases initially due to that a dense silica layer forms at PDMS surface after a long-term UVO exposure that reduces the diffusion of oxygen through the PDMS layer, and block the direct UV exposure preventing the further in-depth densification. In parallel, but on a much slower time scale, there is a densification at the direct UV blocked walls of PDMS because of the scattered UV irradiation and ozone diffusion below the rough TEM grid surface. Another reason of this progressively reduced wall height is probably the insufficient vertical wall stress against the dragging-down force from densified glass layer as the PDMS surface declined to some extent. Figure 3.10A also plots the depth of the micro-pattern replica. Its height is slightly higher than the mold. The reason for this phenomenon is discussed in detail in Chapter 4.2.1. In addition, because, the TEM grids used are not symmetric in surface roughness, i.e. one side is smoother (shiny surface) than the other side (matte surface). Figure 3.5, 3.6, 3.9 were collected with the rougher “matte” side of the TEM grid in contact with the nano-patterned PDMS. In contrast,

the UVO imprinting was conducted by applying the smooth side of TEM grid for micro pattern generation (red and blue curves in Figure 3.10B). We observe that the stable equilibrium micro-pattern heights are obtained even with a long term of UVO exposure, which indicates that the effect of wall stress is largely prevented. Here, the blue and red curves show the results from a flat PDMS film covered by the smooth side of 19 μm square

TEM grid (G1000), and a DVD nano-patterned PDMS also covered by the smooth side of

G1000 TEM grid, respectively. For both DVD nano-patterned and smooth surface, a limiting micro-pattern saturation depth of nearly 500-700 nm is attained. While, the increased rough contact between TEM grid and DVD patterned PDMS accelerates the

PDMS surface densification at TEM grid covered parts. As a control, we placed a smooth silicon surface cleaved along a 100 plane, in contact with a smooth PDMS pad and exposed it to UVO. The depth evolution in the PDMS UVO exposed region (measured at the boundary as a step height), is plotted in the black solid line in Figure 3.10. It increases gradually and saturates at a depth of 1150 nm, nearly twice the depth attained under the smooth contacting surface of a G1000 TEM grid. We attribute this depth difference to the finite size effects of the micro square pattern (discussed in Figure 3.9). The final depth of

1150 nm represents a balance between the retracting UVO densified region stresses and restoring wall stresses. We conducted further studies to determine the effect of PDMS thickness on surface densification. The black dashed line in Figure 3.10B is measured on a PDMS layer of thickness (3.89 ± 0.116) mm, nearly 3 times as the previous PDMS thickness used in this study. Comparing to the black solid line in Figure 3.10B on PDMS of thickness (1.22 ± 0.031) mm, the surface depression of the thicker PDMS is about 200 nm deeper at large UVO dose, but much less than 3 times of the depth of the 1.22 mm 53

PDMS. This non-linear dependence on PDMS thickness phenomena is presumably due to a gradual diminishment of PDMS densification from the surface to bulk interior as the surface glassy layer works as a natural barrier to UV and ozone permeation. Thus, the surface indent depth of PDMS after UVO exposure is not in direct proportion to its bulk thickness.

The Young’s modulus measured at a ratio of 20:1 is about 0.6 MPa, which is very close to the results reported in other literatures33-35. The 18 h (Young’s modulus: 0.57 MPa) cured PDMS with 20:1 mixing ratio was utilized to study the effect of PDMS thickness on the UVO densification efficiency. Figure 3.10C shows the surface height reduction of

PDMS, which were measured from PDMS half-blocked by silicon wafer under 0.5 h (10.7

J/cm2) and 5 h (106.6 J/cm2) UVO exposure. The red and black columns show the PDMS surface decrease at UVO exposures of 0.5 h and 5 h, respectively. We observe the same trends between surface depth and PDMS thickness: larger the PDMS thickness, larger the surface depth under both 0.5 h and 5 h exposure conditions. The effect of PDMS thickness shown in Figure 3.10C lead to the same conclusion as discussed in Figure 3.10B.

Figure 3.10 (A) Depth/aspect ratio changes with UVO exposure time/cumulative radiation energy in the case of a rough TEM grid (1000 mesh, 19 μm). (B) Plot of the depth/ aspect ratio changes with UVO exposure time/cumulative radiation energy. The black and black 54

dashed curves are measured on PDMS samples, half-blocked by a flat silicon wafer, for PDMS of thickness (1.22 ± 0.031) mm and (3.89 ± 0.116) mm, respectively. The blue curve is the depth measured from flat PDMS covered by 19 μm mesh photomask. The red curve is measured from DVD patterned PDMS covered by 19

μm photomask. The graph shows a non-monotonic PDMS micropattern depth change with increasing UVO exposure time: 0.5 h (UVO energy: 10.7 J/cm2), 1 h

(21.4 J/cm2), 2 h (42.7 J/cm2), 3 h (64.0 J/cm2), 5 h (106.6 J/cm2), 7 h (149.2 J/cm2) and 9 h (191.8 J/cm2) corresponding to data points in curves. Error bars in (A) and

(B) are estimated from the standard deviation of repeated measurements. (C) PDMS

Depth changes with different UVO exposure at 0.5 h (red columns) and 5 h (black columns) upon PDMS with Young’s modulus of 0.57 MPa.

Regardless of these complications relating to the surface roughness of the photomask, we are able to generate hierarchical patterns of tunable depth within the 1 m range or larger by varying the mask pattern dimension, UVO exposure time (dose), and

PDMS thickness.

Simulation of the nano patterning process

In order to elucidate how surface tension and elastic properties of polymeric materials used in nano-pattern fabrication process influence deformation of the pattern structure, we perform coarse-grained molecular dynamics simulations of the fabrication process following main steps outlined in experimental methods. In our simulations, we used bead-spring representation of the precursor polymers and cross-linked polymeric

films. The surface of the rigid mold structure (DVD pattern) is modelled by the fixed beads.

The pattern has a periodic structure of bars with square cross-section of side length H0 and distance between bars Lx. The interactions between beads are described by the Lennard-

Jones (LJ) potential:

12 6 휎 휎 휎 12 휎 6 4휀 [( ) − ( ) − ( ) − ( ) ] 푟 ≤ 푟 퐿퐽 푟 푟 푟 푟 푖푗 푐푢푡 푈퐿퐽(푟푖푗) = { 푖푗 푖푗 푐푢푡 푐푢푡 (S.1) 0 푟푖푗 > 푟푐푢푡

where rij is the distance between ith and jth beads with diameter assumed to be the same for all types of beads in a system, εLJ is the LJ interaction parameter and rcut is the cutoff distance. By varying the value of the LJ interaction parameter, we can control affinity between different species. Connectivity of the beads into polymer chains and crosslinking bonds are modelled by the FENE bonds.170 The attractive part of the bond potential is described by the finite extension nonlinear elastic (FENE) potential:

2 1 2 푟 푈퐹퐸푁퐸(푟) = − 퐾푅0푙푛 (1 − 2) 2 푅0

2 with the spring constant K=30kBT/σ and the maximum bond length R0=1.5σ. The repulsive part of the bond potential is described by the pure repulsive LJ potential with

LJ=1.0 kBT (where kB is the Boltzmann constant and T is the absolute temperature) and

1/6 rcut= 2 σ.

The system temperature is controlled by coupling system to the Langevin thermostat. The equation of motion of the ith bead coupled with a thermostat has the following form: 56

푑푣⃑⃑⃑ (푡) 푚 푖 = 퐹 (푡) − 휉 ⃑푣⃑⃑ (푡) + 퐹 푅(푡) 푑푡 푖 퐿 푖 푖

where 푣⃑⃑⃑ 푖(푡) is the velocity of the ith bead, 퐹 푖(푡) is the net deterministic force acting

1/2 on the ith bead, and L=0.1 m/τLJ is the friction coefficient where τLJ = σ(m/εLJ) is the

푅 standard LJ time. In our simulations, all beads have the same mass m equal to unity. 퐹푖 (푡)

푅 is the stochastic force with zero average value 〈퐹푖 (푡)〉 = 0 and 훿-functional correlations

푅 푅 ′ 〈퐹푖 (푡)퐹푖 (푡′)〉 = 6휉퐿푘퐵푇훿(푡 − 푡 ). The velocity-Verlet algorithm with a time step 0.005

휏퐿퐽 is used for integration of equations of motion. All simulations are performed using

LAMMPS.171

In our simulations, all beads have diameter σ. The transferred pattern into polymeric film is fixed either by cross-linking polymeric chains making up the film (analog of the PDMS replica) or by a temperature quench (analog of the PS replica). To establish relationship between parameters of the interaction potentials and macroscopic system properties such as surface tension and shear modulus we perform a set of additional molecular dynamics simulations in which we calculate the shear modulus GN of polymeric network and its surface tension as a function of the cross-linking density and strength of the LJ interaction parameters describing affinity between different species making up the system.

Figure 3.11 shows the simulation results for deformation of the transferred replica fixed by cross-linking polymeric film for patterns with different cross section sizes H0 and different shear modulus GN of the negative mold. It follows from this figure

that as the shear modulus of the negative mold increasing, the deformation decreases

monotonically for all sizes. It is interesting to point out that, compared with the square

shape of the master mold, smaller pattern size shows a larger relative deformation.

Figure 3.11 Evolution of the replica shape produced via cross-linking (blue color)

compared with the corresponding master mold shape (yellow color) as a function of master

mold size H0 and shear modulus GN of the negative mold network.

Table 3.1 Young’s modulus and nano-pattern height for the negative mold with different

PDMS base cross-linker weight ratio where initial pattern height H0=155 nm. Data are

obtained from pure nanopattern samples (cured at 120 °C for 2 h).

PDMS Base: Cross-linker 10:1 15:1 20:1

Young’s Modulus (MPa) 1.74±0.03 0.70±0.05 0.29±0.04

Pattern Height (nm) 137±3 119±1 87±7

The observed deformation of the transferred pattern can be shown to be a manifestation of the balance elastic and capillary forces172-193. The elastic force per unit area generated in the crosslinked polymeric film due to pattern height displacement by

ΔH can be estimated as 푮푵∆푯. This force is balanced by the surface tension ϒ of the network-air interface. This results in the following scaling relation of the pattern height deformation and ratio of the elastocapillary number 휸/푮푵푯ퟎ to pattern dimensions H0:

횫푯/푯ퟎ ∝ 휸/푮푵푯ퟎ. In Figure 3.12, we combined all our simulation data for pattern deformation as a function of the dimensionless parameter 휸/푮푵푯ퟎ. All our data sets have collapsed into one universal plot. In the regime of the small values of the parameter

휸/푮푵푯ퟎ < ퟏ when the network elasticity provides a dominant contribution stabilizing the pattern deformation, the pattern deformation increases linear with increasing the value of the dimensionless parameter 휸/푮푵푯ퟎ. However, in the opposite limit of large values of the parameter 휸/푮푵푯ퟎ > ퟏ the structure of the transferred pattern is completely destroyed. This pattern smoothing is a result of capillary forces that minimize the surface area of the network-air interface by flattening it.

Figure 3.12 Dependence of the deformation ratio ΔH/H0 on the elastocapillary number for different systems: the negative mold (black solid square), melt with negative mold of system 1 (red empty square), melt with negative mold of system 2 (green half solid square).

Experimental data from Table 3.1 are shown by pink stars. Inset shows definitions of pattern deformation in studied systems.

To compare our simulation data with experiments, this figure also shows the dependence of pattern deformation of the negative mold at different cross-linker density

(see the experimental measurement in Table 3.1). By fitting the experimental results to the universal plot, we obtained the surface tension of the PDMS negative mold 휸 ≈

ퟐퟓ 풎푵/풎. This value is close to the report from Dow Corning Inc. (22.7 mN/m).194

Note that capillary forces also play an important role in controlling the fidelity of the pattern fixed by temperature quench. In this case, the deformation of the interface

between the soft mold image and precursor polymeric film controls the quality of the transferred pattern. By quenching the system, we are fixing the location of the interface and pattern deformation.

3.4 Summary

In conclusion, we have developed a simple and versatile technique for fabrication of dual scale micro-nano hierarchical interfacial patterns in the elastomer by combining nano-imprinting and UVO microlithography. Using this approach, tunable and scalable hierarchically patterned structures in PDMS could be created by controlling nano-pattern masters, UVO irradiation, and micro-patterned photomasks. The presented systematic study of the effects of the wall stress, frontal development (gradual diminishment of PDMS densification from the surface to bulk interior under UVO exposure), UVO exposure energy and PDMS modulus provided a valuable information for optimization of the pattern transfer technique. Comparison of computer simulations and experimental results for pattern deformation pinpoint the elastocapillarity as a driving force behind pattern deformation at different stages of the pattern transfer process. The future development of this technique could become a valuable tool for other lithography applications, including microcontact printing with potential applications in the fabrication of microfludic devices, biologycally active substrates and electronic systems.

CHAPTER IV

4 INVESTIGATING THE APPLICATION OF THE

PREPARED MICRON-NANO HIERARCHICAL MOLD

FOR SOFT LITHOGRAPHY

Soft lithography is a family of techniques for fabricating micro or nano structures using elastomeric stamps, molds, and conformable photomasks.28 Poly (dimethyl siloxane),

(PDMS) is one of the most typical mold material.12 In this study, the micro-nano hierarchically patterned PDMS/SiOx hybrid mold is used to study the advantages of hard roofs at the applications of the different forms of soft lithography.

This study is conducted to understand:

➢ whether the PDMS/SiOx micro-nano mold could transfer pattern successfully

without soft mold deformation defects.

➢ comparing to conventional neat soft mold, what’s the difficulties the hybrid

mold brings at the applications of soft lithography.

➢ what could be done to eliminate the defects hybrid mold brings.

4.1 Materials

Poly (dimethyl siloxane), PDMS, Sylgard 182 is bought from Dow Corning, USA.

The solution of silver nanoparticles in ethylene glycol is bought from Sigma-

Aldrich.

Dopamine hydrochloride is obtained from Sigma-Aldrich.

Trichloro (1H, 1H, 2H, 2H-perfluorooctyl) silane (FOTS) is obtained from Sigma-

Aldrich.

Polystyrene (PS) (Mw = 3,000 g/mol, Mw = 37,800 g/mol, and Mw = 161,200 g/mol) are bought from Polymer Source.

4.2 Experimental methods

Pattern transfer – Replica Molding

The negative imprint of the hierarchical pattern or micro pattern is processed by pouring uncross-linked PDMS (PDMS oligomer) onto the originally patterned PDMS mold.

After 2 h of heating at 120 °C, PDMS oligomer is cross-linked. After peeling off, the negative imprinted pattern is prepared. (Figure 4.1)

Figure 4.1 Procedure of replica molding.

Pattern transfer – Microcontact Printing

Microcontact printing of FOTS

First of all, the DVD patterned PDMS was exposed to UVO for 2 h with 2000 mesh

TEM grid placed on it. Then, FOTS was vapor deposited onto the patterned PDMS in a desiccator with 2 h incubation time. After UVO exposure for 20 min, hydroxyl groups form on silicon wafer, which could be used for contact imprinting adhesion. Then, the

FOTS deposited PDMS is gently applied onto UVO exposed silicon wafer to enable the

FOTS transfer. After several seconds, PDMS is peeled off, leaving the FOTS printed onto the silicon wafer.

Microcontact printing of silver nanoparticles

A random smooth substrate, here a silicon wafer was taken as an example, was immersed into a solution of 3-Hydroxytyramine hydrochloride in 10 mM Tris butter (pH

= 8.5) to 2 mg/ml. After 1 h, polydopamine (PDA) was formed and this surface modified substrate was ready to use. Silver particles (diameter: 50 nm) were deposited onto micro- nano hierarchically patterned PDMS mold by Dr. Blade with a silver solution of ethylene 64

glycol with fine control of sliding speed and blade gap. By contacting the silver coated

PDMS mold to PDA treated substrate, the silver particles were transferred consequently, due to the mussels’ adhesive mechanism of PDA.195, 196

Pattern transfer - Capillary Force Lithography

Hierarchical patterns of glassy polystyrene (PS) was generated by placing the hierarchically patterned PDMS face down onto PS films on smooth polished silicon wafer.

For this, PS (Mw = 3,000 g/mol, Mw = 37,800, and Mw = 161,200) films with thicknesses of 87 nm, 240 nm, 390 nm, 532 nm and 620 nm were flow coated onto silicon wafers. The pattern replication occurs when the entire “PDMS-PS-Silicon wafer” sandwich structure is annealed in a vacuum oven at 180 °C for 2 h. The whole process of pattern fabrication is shown in Figure 4.2.

Figure 4.2 Experimental process for pattern transfer from PDMS to thermoplastic material

(PS in this study) via Capillary Force Lithography(CFL).

4.3 Micro-nano patterning through Replica Molding

Figure 4.3 shows the results of the combined micro-nano hierarchical patterning process in PDMS (cured directly on DVD pattern and then UVO exposed under TEM grid). The left column of Figure 4.3 shows hierarchical PDMS patterns obtained after 4 h

(accumulated energy: 85.3 J/cm2) of UVO exposure through the same three TEM grids

(as shown in Figure 3.1) after a common DVD nano-patterning step. The right column is the negative imprint of the pattern on the left, obtained by direct molding. Figure 4.3 G

(left) shows the Fast Fourier Transform (FFT) of the AFM image of the hierarchical pattern from G2000 TEM grid (side width: 7.5 µm). Figure 4.3H (left) is the FFT from the negative replica of Figure 4.3G (left). The planar orderings of the hierarchical patterns are clearly reflected in these FFT spectra of the AFM images. The 5 obvious high-order peaks indicate the pronounced channel orientation. The in-plane average channel distance of 757 nm (very close to channel wavelength obtained from AFM: 750 nm) was derived from the FFT in Figure 4.3G (right), which also revealed the long-range uniformity of channel exceed by far the investigated area. The macro square patterns are arranged with high periodicity in both horizontal and vertical directions, according to the parallel dot matrixes. According to Figure 4.3G (right: line profile of the hierarchical pattern units generated from the G2000 TEM grid) and 3H (right: line profile from the negative imprinted pattern of 3G (right)), we observe that the UVO exposed regions become depressed by several hundred nanometers, yet quite surprisingly, the nanostructures created through the initial DVD patterning survive the deep UVO exposure densification process. Comparing to the original nano channel height (120 nm),

shrinkages are observed, as illustrated in curves c (showing the nano height in open region) and d (showing the nano height at replica of c) in Figure 4.3J, where we see that the PDMS nanoscale channel height is smaller in open regions (c) than the covered part

(e in Figure 4.3K) of the TEM grid. The replica of hierarchically patterned PDMS has a higher micro-pattern height but a smaller nano-patterned height as compared to the original hierarchically patterned PDMS as shown in Figure 4.3I and Figure 4.3J respectively. Likewise, the UVO blocked nano-pattern region in the original hierarchical pattern is higher compared to the hierarchical replica in Figure 4.3K.

I believe the reason for the height of the micro-pattern at the replica of the hierarchical pattern being slightly larger than the original micro depression, is that the vertical thermal expansion of the non-densified UVO blocked walls of the mold pattern, at the elevated temperature during the curing of the replica PDMS fluid. This decreased relief pattern height trend in nano channel pattern replication is opposite to the increased micro pattern amplitude replication, both outside and inside UVO exposed regions, simply due to PDMS contraction associated with cooling.

Figure 4.3 Left column shows AFM images of hierarchical PDMS pattern combining DVD with TEM grid patterns after a 4 h (85.3 J/cm2) UVO exposure. The right column shows negative imprint of left column PDMS patterns. Insets are the structures with high resolution, with dimension of 20 µm x 20 µm, 10 µm x 10 µm and 20 µm x 20 µm, from top to bottom.

mesh imprinting case at both of direct imprint and replica conditions. The I, J and H, three plots show the height changes of micropattern, UVO blocked and unblocked nano-patterns at mold and replica (a, b, c, d, e and f, marked in AFM images accordingly). The data is calculated from AFM images of 300 hexagonal mesh grid/DVD hierarchical patterns.

Uncertainties in plots indicate the standard deviations in repeated measurements.

4.4 Micro-nano patterning through Microcontact Imprinting

Generally, micro-contact printing is widely studied for generating patterned SAMs

of alkanethiolates on gold, but it’s not actively studied on Si/SiO2 and glass. The SAMs of

alkytrichlorosilanes on the hydroxyl-terminated surfaces are less ordered than those of

alkanethiolates on gold, and they form more slowly.197, 198 To validate the SAM transfer

from the micro-nano hierarchically patterned PDMS to Si wafer, trichloro (1H, 1H, 2H,

2H-perfluorooctyl)silane (FOTS) was utilized to conduct the experiments. Firstly, the

DVD patterned PDMS was exposed to UVO for 2 h with 2000 mesh TEM grid placed on

it. Then, FOTS was vaper deposited onto the patterned PDMS in a desiccator for 2 h. After

UVO exposure for 20 min, silicon wafer with hydroxyl group exposed was used for contact

imprinting. During contact imprinting, Si-O-Si crosslinking of covalent bonds forms by

the reaction of silane group and hydroxyl group at the contact regions199. The transferred

FOTS on silicon wafer from the original and negative molds were shown in Figure 4.4.

The full profiles and the zoom-in AFM images show FOTS aligned quite well on silicon with both of the micro squares and nano channel shapes. No mold deformation is observed, which may be result from the hard shell at the top of micro recession. As in neat soft mold cases, the ultra-high aspect ratio of micro mold induces the roof collapse easily, as discussed in chapter 2. The patterned SAMs comprising alkylsiloxanes on Si/SiO2 and glass provides an alternative to the traditional UV photolithography.

Figure 4.4 AFM images of SAM transferred onto Si from the PDMS mold with original and negative patterns. The structures in red and blue squares show the zoom in features of the ordered nano channels.

To further explore the application of the contact imprinting of the introduced hierarchical patterns and expand this method to other different substrates, rather than the general gold and silicon substrates. A test on silver particle transfer via microcontact printing was conducted, with polydopamine (PDA), a mussel-inspired surface modification, developed upon objective substrates. A random smooth substrate, here a silicon wafer was taken as an example, was immersed into a solution of 3-Hydroxytyramine hydrochloride in 10 mM Tris butter (PH = 8.5) to 2 mg/ml. After 1 h, PDA was formed and this surface modified substrate was ready to use. Silver particles (diameter: 50 nm) were deposited on micro-nano hierarchically patterned PDMS mold by Dr. Blade with a silver solution of ethylene glycol with fine control of sliding speed and blade gap. By contacting the silver coated PDMS mold to PDA treated substrate, the silver particles were transferred consequently, due to the mussels’ adhesive mechanism of PDA. 195,196 The SEM images of the resulted silver patterns are shown in Figure 4.5. These 3 images show the assembly of micro squares, a single micro unit, and the zoom-in nano channel structures, respectively.

The perfectly ordered nanoparticles prove the success of microcontact printing and implies the further potential applications of combining the developed micro-nano hierarchical patterning and contact imprinting, such as microfluidic, cell growth studies, optical gratings200 and nanowire electronics.

Figure 4.5 SEM images of silver nanoparticles replicating the mold micro-nano hierarchical structures.

4.5 Defects analysis and elimination via a process optimized Capillary Force

Lithography

Micro-nano hierarchically patterned PDMS could be created by the method presented in chapter 3. For Capillary Force Lithography, mold with hierarchical pattern was generated via densification and consequent “sinking” of the regions exposing to UVO irradiation for 4 h (85.3 J/cm2) through the TEM grid pattern in direct contact geometry with a nano channel patterned PDMS. To transfer the fabricated micro-nano hierarchical pattern to Polystyrene (PS), patterned PDMS was placed face down onto the spin casted

PS (Mw = 3,000 g/mol, 37,800 g/mol and Mw = 161,200 g/mol) films with thicknesses (Ho) of 87 nm, 240 nm, 390 nm, 532 nm and 626 nm. The pattern replication occurs when the entire sandwich structure of hierarchical patterned PDMS placed upon the PS film is

annealed in a vacuum oven at 180 °C for 2 h. Figure 4.6a shows the optical microscopic images of PS with different film thickness (Ho) and molecular weight (Mw). The microscope images show that well defined micro structures at varied Mw and Ho are obtained in large areas. Figure 4.6b provides the whole AFM images of the single micro- protrusion units at the patterned PS film (Mw: 161,200 g/mol) with varying film thickness.

The conversion of the line shapes, from meniscus to near rectangle, indicates the transition from non-efficient filling to full filling of the micro protrusion with increasing film thickness, as shown in Figure 4.7. A similar phenomenon was reported in other research works63, 64, 67, 201, which were considered as a signature mark of capillary rise in CFL, as discussed in chapter 2.2.3.

Figure 4.6 Microscopic and AFM images of PS patterning with different Mw and film thickness.

Figure 4.7 Line profiles of micro units at PS films with thickness of 87 nm, 390 nm and

626 nm.

To obtain more information about the CFL in micro-nano hierarchical patterning, quantitative analysis was conducted with the AFM images in Figure 4.6. Figure 4.8a illustrates the notations used in the discussion. Here, H, hm, and hb refer to the PDMS cavity depth at room temperature, the absolute height of PS micro protrusion, and height of PS base, after imprinting, respectively. Furthermore, hb1 and hb2 denote heights of patterned

PS base at the boundary between two micro protrusions and at the crossing of four micro protrusions, respectively. Figure 4.8b illustrates the measured hm, hb1, and hb2 with varying

PS film thickness. The larger PS base height at the crossing region (hb2) comparing to the boundary region (hb1) is largely due to the lower PDMS micro mold caused by the enhanced pinning stress at the boundary of the densified glassy region. (The detailed information is available in previous work shown in chapter 3) The grey shadow region (hm - hb1) indicates the increased relative height of micro protrusion with the increasing PS film thickness,

which is also exhibited by the red curve in Figure 4.8c. The green dash line in Figure 4.8c shows the average depth of micro depression in PDMS mold depth. In thin films, increased pattern height is observed with the increasing as cast film thickness, however does not reach the depth of the micro pattern at the mold (H). As a consequence of continuous increasing of the as cast film thickness, the final relative height of the micro pattern ceases to change by reaching a steady plateau. Thus, we successfully obtained the micro-nano hierarchically patterned PS film with micro height adjustable by using one constant PDMS mold. In other words, thin PS film (thickness: 87 nm – 390 nm) replicated the far away

PDMS nanochannel structure (mold depth: 410 nm) effectively. It’s expected to be induced from the capillary undulation at the elevated temperature as the PS material is not enough in volume for mold filling.

Figure 4.8 Analysis based on the AFM pictures. a, notation description of H, hm, hb. b, the plot of the height of micro protrusion (hm), and nano pattern at boundary of 2 micro units

(hb1) and 4 micro units (hb2) as a function of as cast PS film thickness. c, height/depth comparison of PDMS mold depth and PS micro pattern protrusion.

To illustrate the mechanism of the thin film replicating far away structures, the schematic of the CFL micro-nano imprinting process is presented in Figure 4.9. Before annealing, the bulk PDMS has the micro recession with the depth of H, and PS has the as cast film thickness, H0. During thermal annealing, the “irregular” replicating behavior may be induced by the thermal expansion of PDMS bulk (Coefficient of Thermal Expansion,

CTE = 3.0 *10-4 /K202), which suppresses the driving upward capillary force to get a counterweight balanced state, inducing the lower PS-PDMS meeting level. In the condition of non-efficient mold filling material (PS), the hydrostatic pressure(downward) and the pressure drop across the meniscus are balanced by the PDMS thermal expansion.

훻퐿 = 퐿 ∙ 훼푃퐷푀푆 ∙ 푑푡

Because of thermal expansion, the PS (Coefficient of Thermal Expansion, CTE, 8.6

* 10-5/K 161) shrink furthermore after cooling down to room temperature. These could be the reasons for the hierarchically patterned PS thin film shows short micro protrusion while still replicating the nano channels.

To understand if thermal expansion is the factor inducing the special phenomon of thin film replicating top structures at the mold, AFM is used to analyse the topography after thermal expansion. Figure 4.10 shows the line profiles of PDMS micro patterns before and after thermal expansion. It shows that after thermal expansion (at 180 °C) the depth of

PDMS is enlarged to 911 nm from the original 812 nm. It proves that during CFL, the top ceiling of micro mold raises up rather than sinks. This fact suggests that the hypothessis of

thermal expansion and capillary force blanced meeting assumption should not be the explaination for the observed phenomenon.

Figure 4.9 Mechanism discussion for the thermal expansion induced thin film ceiling touching phenomenon.

Figure 4.10 AFM line profiles of the surface topography at PDMS micro molds before and after thermal expansion.

Actually, this interesting thin film replicating far away features is considered as a similar phenomenon as reported by Dr. Stephane Chou. Lithographically induced self- construction(LISC)203, an electrohydrodynamic patterning method, has been shown to 77

produce periodic patterns by placing a mask upon polymer films with no contact. The assembling strategy is shown in Figure 4.11. According to a lot of studies204-207, the electrostatic force (either internally generated by the contact potentials or externally applied) drives the instability and the surface tension stabilized the system. According to

Dr. Stephen Chou’s study208, with specific distance of air gap, any defect dot/plain plates in the topology will result in charge accumulation in certain regions, especially at defect regions, thus giving a non-uniform charge distribution, as shown in Figure 4.11. Therefore, the electrostatic force between the polymer and the mask is stronger in some places than in others. The final generated pattern shows the height of protrusions is much larger than the original film thickness.203, 206 In our study, by taking advantages of this lithographically induced self-construction, we successfully generated height adjustable micro-nano hierarchical patterns by simply adjusting film thickness even with a fixed mold.

Figure 4.11 Schematic illustration of Lithographically Induced Self-construction (LISC).

208 (Reprinted with permission from Reference 208)

In addition to the nano channel structures at micro protrusions, “bamboo” structures are observed on these channel features unexpectedly. The channel structures, inside and outside of micro protrusion, for PS films with thicknesses of 87 nm, 240 nm and 532 nm are shown in Figure 4.12a. Actually, according to AFM analysis, similar topography 78

applies to the whole set of film thicknesses. Comparing to the neat channel structures at the surface of PS base, the widely-distributed bamboo features should not from the non- sufficient materials as it shows everywhere on the channels at micro protrusion on PS with film thickness range from 87 nm to 626 nm. In nanoimprint lithography, bubble defects were reported eliminable by an appropriate selection of film thickness and imprint conditions, like temperature and pressure. While in CFL, quite few work has been done to study the formation and elimination of bubble defects. Gradual oven annealing is not an efficient way to expel the air trapped between PS and PDMS/SiOx hybrid mold, which is evidenced by the remaining bamboo structure after gradual annealing, as shown in Figure

4.12b. Here, the gradual oven annealing is conducted by slowly increase the temperature for CFL by reaching 180°C from room temperature in 45 min. However, this bamboo defects can be successfully removed by utilizing the neat PDMS mold by dual replicating the original PDMS/SiOx hybrid mold. Dual replication is conducted by two times of replica molding by using PDMS as an intermediate material. Because the original PDMS mold in use consists a hard shell at the top of micro recession, which is a SiOx dominated glassy layer. This glassy layer is air trappable. After dual replication, the soft PDMS mold not only retains the air-permeable property12, but also shows the exact same micro-nano hierarchical structures. This neat PDMS mold gives the excellent bubble eliminating properties, which could be confirmed with high resolution AFM line profiles of the nanochannels, as shown in Figure 4.12d.

Figure 4.12 (a) AFM height images of PS with film thicknesses of 87 nm, 240 nm and 532 nm. The red and blue regions show the topography of channels on top of the micro units and at the surrounding of the micro units. (b) Channel features on top of micro units after

CFL with gradual annealing. (c) Channel features on top of micro units by using a neat

PDMS mold (generated by dual replication of the original PDMS/SiOx hybrid mold). (d)

Line profiles of the channels on top micro units, with original PDMS/SiOx hybrid mold, by gradual oven annealing, and by using a neat PDMS mold (dual replication), respectively.

After understanding the origin of the occurrence of hole defects at the micro protrusion, I want to explore potential methods to expire hole defects to increase the fidelity of hierarchical patterning via CFL by using the UVO depressed PDMS mold. Dynamic

Cold Zone Annealing (DZA), a technique widely used to enhance ordering kinetics of phase separation of block copolymer (BCP)209-211 is introduced here to study the potential benefit from localized annealing on bubble elimination to tell whether temperature gradient

helps mold filling process. In DZA, the temperature gradient is generated by using two cold blocks on two side of a hot block. The temperatures of these blocks are well controlled.

As the sample is dragged through the temperature gradient with a motor, the temperature profile is generated on each point of the sample, as shown in Figure 4.13. In this study, PS

(Mw = 161,200 g/mol) and PS - PMMA (Mw = 57k-25k g/mol) thin films with a film thickness of 370 nm were used for DZA by firstly spin coated on UVO cleaned quartzes.

The PDMS mold was fabricated firstly by DVD replica molding and then UVO exposure of 85.3 J/cm2 with a photomask of 300 mesh square TEM grid. With PS or PS-PMMA film confined by hierarchically patterned PDMS mold, the whole sandwich set undergone DZA with cold-hot-cold regions at the speed of 20 µm/sec. Figure 4.14 shows the AFM images of channel structure at micro protrusion of PS film through oven annealing, and DZA with nano channel parallel and perpendicular to DZA direction. Bubble defects are successfully eliminated by DZA due to the in-situ thermal annealing expelling air bubbles efficiently.

A controlled experiment was conducted to increase the temperature in oven annealing slowly to compare with the influence of the localized temperature gradient in DZA. Results show that bubble defects remain in oven annealing with even slow temperature increase

(by spending 1 h to increase the temperature from room temperature to 180 °C), indicating the great advantage of DZA in suppressing undesirable patterns with enhanced in-situ patterning kinetics created through sharp temperature history.

Figure 4.13 Schematic of the cold/dynamic zone annealing system. 211 (Reprinted with permission from Reference 211)

Figure 4.14 AFM images of nano channel at micro protrusion of PS (Mw = 161,200 g/mol) and PS-PMMA (Mw = 57,000 – 25, 000 g/mol) in oven annealing and DZA conditions.

4.6 Summary

In this chapter, we demonstrated the capability of soft lithography by using the micro-nano hierarchical patterned PDMS/SiOx mold. These micro-nano patterned molds are generated by combining UVO lithography and general replica molding (as discussed in chapter 3). The three forms of soft lithography examined here includes replica molding, microcontact printing and capillary force lithography.

During replica molding, both of micro and nano patterns are replicated from crosslinked PDMS to another layer of PDMS without using any releasing agent. However, opposite behaviors of micro patterning and nano patterning are studied. This is considered due to the different thermal expansion and surface energy compensation at different pattern scales.

For microcontact printing, FOTS, a fluorinated small volatile molecular, and silver nanoparticle are used as “ink” material to explore the stability of the shell hardened PDMS mold. FOTS could be chemically bonded to substrate due to the chemistry between chlorine group and hydroxyl group. In addition, dopamine is used as a surface functionalizer to generate surface adhesion between silver particle and substrate. This part of work demonstrates the potential application of the micro-nano hierarchically pattern

PDMS for surface modification.

In the Capillary Force Lithography part, capillary force acts as the only driving force to drive the mold filling of thermoplastic material by using the same PDMS hybrid

mold. We obtained a method to replicate the structures “far away” at the top of PDMS mold even with a relatively thin polymer film. The mechanism of the this “irregular” phenomenon is fully investigated. This phenomenon contributes to be an effective means to get a micro pattern with controlled height without sacrificing the fidelity of nano channel pattern by simply controlling the film thickness. In addition, we investigated the causes of bubble defects. As a consequence of understanding the causes, bubbles defects were finally eliminated by replacing with a neat soft PDMS mold or by applying DZA for CFL.

Furthermore, DZA-CFL is a very efficient technique as it’s scalable via Roll to Roll (R2R) processing, providing a potential chance for large area and high fidelity hierarchical patterning.

CHAPTER V

5 SHAPE MEMORY BEHAVIOR OF MICRO AND NANO

SCALE PATTERNED FA/ELASTOMER COMPOSITES

This chapter describes the shape memory behavior of micro- and nano-scale patterns imprinted on the surface of a shape memory polymer composed of mixtures of Zn-

SEPDM with three different FAs, zinc stearate (ZnSt), stearic acid (SA) or lauric acid (LA).

The advantage of an ionomer/FA compound as a shape memory polymer is that it is derived from two easily obtained or synthesize materials and can easily be scaled to greater than pound quantities. Another advantage of this type of SMP design is that the switching temperature, which is the melting point of the FA, can be easily changed by using different

FAs. Also, since supramolecular bonds were used for the permanent and temporary networks in this shape memory polymer design, the materials are thermoplastic – i.e., they can be molded and remolded into any shape using conventional thermoplastic processes.

This study is conducted to understand:

➢ whether the Zn-SEPDM/FA SMP, for which bulk shape memory has been

previously demonstrated74, 155, 156, can be used for creating switchable micro-

and nano-scale patterns imprinted on its surface.

➢ whether simply changing the FA used also allows one to design a family of

surface shape memory materials with varying switching temperature.

5.1 Materials

A statistical terpolymer, poly[ethylene-co-propylene-co-(5-ethylidene-2- norbornene)], EPDM, with a backbone composition of 49.0% ethylene, 46.8% propylene, and 4.2% 5-ethylidene-2-norbornene is obtained from Exxon Chemical Co.

Poly (dimethyl siloxane), PDMS, Sylgard 182 is bought from Dow Corning,

USA.

Zinc stearate (ZnSt), stearic acid (SA) and lauric acid (LA) are obtained from

Sigma-Aldrich.

5.2 Experimental methods

Material synthesis and formulation

Sulfonated EPDM was synthesized by sulfonating the EPDM with acetyl sulfate at ~25ºC in a hexane solution following the procedure Makowski et al. proposed212. As shown in Figure 5.1, the resulting sulfonated polymer was neutralized with excess zinc acetate dehydrate to form the zinc sulfonate derivative, Zn-SEPDM, which had a 86

sulfonation level of 10.1 meq/100g as measured by elemental sulfur analysis. Zn-SEPDM is a thermoplastic elastomer that can be melt-processed at elevated temperatures, T > 150

ºC. Three different FAs, zinc stearate (ZnSt), stearic acid (SA) or lauric acid (LA) were obtained from Sigma-Aldrich and used as received. Compositions were prepared by solution blending the ionomer with 20 wt% of fatty acid/salt in a 95/5 (v/v) mixture of toluene and methanol and then evaporating the solvent. The same Zn-SEPDM ionomer was used for each compound, and the sample notation used herein is SMP-x, where x =

ZnSt, LA or SA. Previous work has shown that the fatty acids are homogeneously distributed in the blend, but phase separates as microscopic crystals are stabilized by interaction with the salt groups of the ionomer.158 When the compounds are heated to above the melting point of the fatty acid, the molten fatty acid acts as a plasticizer to improve melt processing, but the strong intermolecular interactions between the carboxylic acid or metal carboxylate groups of the fatty acid and the zinc sulfonate groups of the ionomer prevent the fatty acid from migrating to the surface of the processing equipment during melt mixing or diffusing to the surface of the solidified molded compound.212 (Figure 5.2)

Figure 5.1 Synthesis of Zn-SEPDM.

Figure 5.2 (a) Chemical composition of Zn-SEPDM, (b) Microstructure of SMP-FA. The permanent network forms from the aggregation of the zinc sulfonate ionic groups (red circles), which physically crosslink the ionomer and the temporary network consists of

microcrystalline FA which also serves as supramolecular crosslink junctions due to strong dipole-dipole or ion-dipole interactions between the ionomer and FA (see arrow). This figure is not shown in scale.

Thick film (bulk) specimen preparation

Thick film with thickness varied from 1 µm to 20 µm was prepared on polyethylene terephthalate (PET) film by using Dr. Blade and followed by heating at the temperature of

60 ºC. Film was ready to use after vacuum dry in oven overnight at 60 ºC. SMP-X solution was mixed in a solvent mixture of 95/5 (v/v) toluene/methanol.

Shape memory of micropatterned SMP-X film

A copper TEM grid with 58 μm wide hexagonal holes separated by 25 μm was pressed into a 50 µm thick film with a pressure of 0.29 MPa at 190 ºC for 15 min to imprint a permanent micropattern on the surface of the film. The temporary shape was generated by pressing a smooth PDMS onto micropatterned sample at Tc with a compressive stress of 0.30 MPa for 3 min, followed by cooling under stress to 10 ºC to fix the temporary pattern. The original surface micropattern was recovered by reheating the unstressed temporary pattern to Tc for 2 min.

Shape memory of nanopatterned SMP-ZnSt film

A shape memory nanopattern was prepared by imprinting a channel (grating) pattern from a PDMS nano-mold onto a SMP-ZnSt film, Figure 5.3 (top). The PDMS channel nano-mold was prepared by using direct pattern transfer by pressing a 89

polycarbonate digital video disc (DVD) master (channel height = 120 nm, wavelength =

750 nm) into a liquid PDMS film (Sylgard 182, Dow Corning) on a glass plate. The master was obtained by separating the polycarbonate layer from the metallic reflection layer of the

DVD, and the polycarbonate lithographic master was washed with methanol before imprinting. Air bubbles were removed from the PDMS liquid between the glass plate and the DVD master by placing the system in a vacuum desiccator for 15 min, and then the

PDMS was cured at 120 ºC for 2 h. The nanopatterned PDMS was peeled off the DVD master and used to prepare the permanent nanopattern on a SMP-ZnSt surface by casting a solution of SMP-ZnSt in a mixed solvent of 95/5 (v/v) toluene/methanol onto the patterned PDMS film and evaporating the solvent overnight at 60 ºC, Figure 5.3a.

The nanopatterned SMP sample (permanent pattern) was programmed into a crosshatched pattern by forming a temporary grating pattern orthogonal to the permanent grating pattern with the PDMS nano-mold using thermal nanoimprint lithography at 100

ºC with a compressive stress of 1.6 MPa, Figure 5.3b. Note that the stress was calculated using the area of the pattern protrusions, which occupied only about one-fourth the planar area of the film surface. The temporary crosshatched pattern was fixed by quenching the

SMP-ZnSt film with the embedded PDMS nano-mold in liquid nitrogen without releasing the imprinting compressive stress. The recovery step was achieved by reheating the film

with the temporary crosshatched nanopattern to 100 ºC for 5 min with no applied stress,

Figure 5.3c.

Figure 5.3 (top) Schematic of the fabrication of the PDMS nano-mold and AFM images of DVD master and nanopatterned PDMS. (bottom) Schematic of the surface shape memory experiment: (a) nano-channel pattern transfer from PDMS to SMP-ZnSt using solution casting; (b) melt fabrication of temporary, crosshatched nanopattern at Tc = 100

ºC; (c) recovery of permanent pattern by reheating unstressed temporary nanopattern to Tc

= 100 ºC.

5.3 Shape memory behavior of bulk composites

Tg of the three SMP-x compounds was the same, -56 ºC, but the melting point (Tm) of the FA within the compound, defined as the peak in the DSC melting endotherm, varied from 39 ºC for SMP-LA to 63 ºC for SMP-SA to 119 ºC for SMP-ZnSt. (Figure 5.4). The temperature interval over which melting occurred was much broader for the FA in the shape memory compound than for the pure FA, and in general, Tm of the FA crystals in the compound was depressed 6 – 10 ºC from that of the pure FA due to the strong interactions between the ionomer and the FA, which produced microscopic FA crystals. For these

shape memory compounds, the shape memory switching temperature, Tc ~ Tm, so the DSC data in Figure 5.4 demonstrate the tunability of Tc by simply changing the FA used in the compound.

Figure 5.4 Thermal behavior of SMP-x compounds shown by DSC (Perkin Elmer DSC

8500) heating thermograms with a temperature gradient of 10 ºC/min.

A typical shape memory cycle of a SMP-ZnSt film (~ 50 × 20 × 0.05 mm) measured by DMA (TA Instruments Model Q-800) is shown in Figure 5.5. The film was heated to

Tc (path 1) to partially melt the FA crystalline domains. The tensile stress was then increased to 0.2 MPa and the sample was allowed to reach an equilibrium strain, (path

2). Note that only partial melting was used for SMP-ZnSt, so that sufficient FA was melted to allow the sample to form and retain a temporary shape, but to also allow the residual

ZnSt crystals to prevent creep or flow of the permanent shape. The sample was then cooled to 20 ºC at 20 ºC/min with a constant applied stress (0.2 MPa) to obtain a temporary shape with a strain 휀푚 (path 3). The applied load was removed (path 4) to fix the temporary shape, strain 휀푢. Note that an ideal shape memory polymer should behave as an elastic solid with 휀푚 = 휀푢. The sample was then held at 20 ºC with no applied stress for 3 min, during which time the sample length changed by < 1%, and then the temperature was raised to

100 ºC again to recover the permanent shape (path 5). The sample was held at 100 ºC for

10 min without stress to allow the recovery process to reach equilibrium, and finally the sample was cooled to 20 ºC (path 6). (Figure 5.6)

Figure 5.5 Tensile storage modulus measured at f = 1 Hz by DMA (TA Instruments Q800).

The shape fixing, F, and recovery, R, efficiencies were calculated from the DMA data using the following equations213,

휀 퐹 = 푢 ×100% (1) 휀푚

휀 −휀 푅 = 푚 푝 ×100% (2) 휀푚−휀푖

where 휀푖, 휀푢, 휀푚 and 휀푝 are the strains of the original, permanent shape, the fixed temporary shape, the stretched shape prior to cooling/unloading and the recovered shape, respectively.

Figure 5.6 Shape memory cycle for a SMP-ZnSt film. The numbers indicate the distinct steps of the cycle as described in the text and the strains correspond to the values used in equations (1) and (2). Path 1: The sample was heated to 100 ºC with an applied tensile stress of 1.5 Pa to prevent the sample from sagging. Path 2: The applied stress was increased to 0.2 MPa and the sample was allowed to equilibrate. Path 3: The deformed sample was cooled to 20 ºC under a constant stress of 0.2 MPa resulting in a strain 휀푚.

Path 4: The applied stress was removed to fix the free-standing temporary shape with a strain 휀푢. The sample was then reheated without any applied stress to 100 ºC to remove the temporary network (path 5) and cooled to 20 ºC to recover the permanent shape, strain 휀푢.

From DMA data, the fixing efficiency for SMP-ZnSt (weight percentage: 20%) calculated from Equation (1) is 80%. The contraction shown during path 4 in Figure 5.6 is due to creep of the sample due to the fact that it is actually a viscoelastic solid. In this case, the creep actually occurred in the permanent network155, 156, which in this system is a supramolecular network formed by the formation of ionic nanodomains ionomer74, 214. The onset of shape recovery occurred at around ∼80 ºC, though the major shape recovery occurred between 90 and 100 ºC, which was a consequence of the relatively broad melting transition of the ZnSt in the compound.

Five consecutive shape memory cycles for the ZnSt/Zn-SEPDM using the stress and temperature program described in the previous paragraph for each cycle are shown in

Figure 5.7, and the fixing and recovery efficiencies are summarized in Table 5.1. These results demonstrate the reproducibility of the shape memory performance of the SMP-ZnSt.

The shape fixing and recovery behavior were repeatable for at least four shape memory cycles following the first cycle (note the first cycle had abnormally low F and/or R due to relaxation of thermal stresses generated during sample preparation).

Shape memory cycles of ZnSt/ZnSEPDM SMP

120 0.25 30

25 100 0.20

80 0.15

C o

60 0.10 Strain % Strain

10 (Mpa) Stress

Temperature 40 0.05 5

20 0.00 0

0 0 100 200 300 400 Time (min)

Figure 5.7 Five consecutive shape memory cycles for SMP-ZnSt; The function of temperature (black curve), stress (red curve) and strain (blue curve) with time.

Table 5.1 Shape fixing and recovery efficiencies for SMP-ZnSt for five consecutive shape memory cycles. (see Figure 5.7)

Cycle 1 2 3 4 5

F 0.81 0.80 0.79 0.80 0.80

R 0.69 0.93 0.95 0.95 0.96

The shape memory cycles are also conducted on other FA mixed shape memory materials. For SMP-ZnSt (Tc = 100 ºC), F = 80 ± 0.93 % and R = 95 ± 1.0 %; for SMP-LA

(Tc = 60 ºC), F = 94 ± 1.3 %; R = 89 ± 3.2 % and for SMP-SA (Tc = 38 ºC). F = 88 ± 2.8 %;

R = 100 ± 1.0 %.

5.4 Surface micropattern memory behavior

A demonstration of microscale memorization of the SMP-ZnSt is shown in Figure

5.8. A copper TEM grid with 58 μm wide hexagonal holes separated by 25 μm was pressed into a 50 µm thick film with a pressure of 0.29 MPa at 190 ºC for 15 min to imprint a permanent micropattern on the surface of the film, Figure 5.8a. The temporary shape was fabricated by pressing a smooth PDMS onto micropatterned sample at 100 ºC with a compressive stress of 0.30 MPa for 3 min, followed by cooling under stress to 10 ºC to fix the temporary pattern, Figure 5.8b. The original surface micropattern was recovered by reheating the unstressed temporary pattern to 100 ºC for 2 min. Figure 5.8c.

The optical micrographs and AFM strips in Figure 5.8 show that although the original hexagonal pattern, Figure 5.8a, was compressed during the formation of the temporary pattern, Figure 5.8b, the flattening of the pattern was not complete. That was probably a consequence of using a soft elastomer pad, PDMS, to deform the hexagonal micropattern into a flat surface during the temporary pattern fabrication. That procedure only compressed the hexagonal pattern by ~37%, see the AFM line profile in Figure 5.8b.

The failure to completely flatten the surface in this experiment is not believed to be a limitation of the pattern memory behavior of these compounds, but rather a consequence

of using the PDMS to transmit the pressure to the film in the attempt to flatten the hexagonal micropattern. However, in a subsequent experiment, when a stiffer substrate was used to fabricate the temporary shape, the hexagonal pattern was successfully compressed into a flat surface.

The recovery of the permanent pattern, Figure 5.8c was efficient as demonstrated by the excellent fidelity of the recovered pattern with the permanent pattern shown in

Figure 5.8a. Equations (1) and (2) could not be used to quantify the pattern memory behaviors, because it was not possible to measure the pattern depth profile (i.e., 휀푚) prior to unloading the stress during the fixing step. Instead, Equation (3) was used to characterize the recovery efficiency of the pattern memory, R,

R  Hr/Hp (3)

where Hr and Hp are the depths of the recovered and original micropatterns as measured by AFM. In this case, R ~ 100 % for the micropattern shape memory.

One of the main advantages of the ionomer/FA shape memory polymer design is that the switching temperature, Tc, used for programing a temporary shape and recovering the permanent shape can be tuned by simply varying the FA used74. Families of SMPs with varying Tc can easily be developed for a single host polymer by compounding a thermoplastic ionomer with different FAs using conventional polymer processing techniques, e.g., extrusion, and complicated shapes or textures can be manufactured by injection molding, embossing or stamping. Where a common solvent is available, solution processing can also be used to prepare shapes or surface textures. For example, SMPs with

Tc ranging from 10 ºC to 120 ºC may be prepared from compounds of the Zn-SEPDM ionomer with various FAs74 and bulk objects may be manufactured by extrusion, injection molding and solution-casting.

Figure 5.8 Micropattern shape memory demonstration for SMP-ZnSt: (a) original surface imprinted micropattern (permanent pattern), (b) temporary pattern, (c) recovered pattern.

Top photos are optical micrograph and the bottom photos and graphs are the AFM.

Micropatterns similar to that shown in Figure 5.8 for SMP-ZnSt were also prepared from SMP-LA and SMP-SA, where Tc = 50 ºC and 80 ºC, respectively. The AFM line profiles for the hexagonal texture for the original, temporary and recovered shapes for the three SMP-x films are shown in Figure 5.9, along with the values for R – all of which were greater than 90 %.

Figure 5.9 Line profiles of the AFM images for the original, temporary and recovered micropatterns for SMP-LA, SMP-SA and SMP-ZnSt. The original pattern is the same as shown in Figure 5.8a and the same pattern-memory procedure that was described for

Figure 5.8 was used for all the shape memory compounds.

5.5 Shape memory behavior of nano patterned film

A shape memory nanopattern was prepared by imprinting a channel (grating) pattern from a PDMS nano-mold onto a SMP-ZnSt film, Figure 5.3. The fidelity of the imprinted nanopattern was characterized by a fill factor defined as the ratio of the cross- sectional areas of the imprinted channels on the SMP-ZnSt surface and on the PDMS nano-

100

mold. That calculation indicated that the pattern was reproduced to an accuracy of 91 %.

The loss of fidelity was most likely due to difficulties in flowing of the polymer into the small grooves and/or relaxation of the polymer after the mold was removed.

Figure 5.10 shows the surface morphology, as measured by optical microscopy and

AFM, of the original imprinted (permanent), temporary and recovered surface textures of

SMP-ZnSt for the shape memory experiment described in Figure 5.3. The directions p and t shown in the insert to Figure 5.10 correspond to parallel and orthogonal to the direction of the channels in the original imprinted film (i.e., the permanent pattern). The imprinted nanopattern, Figure 5.10a, reproduced well the nanopattern of the PDMS nano- mold, though the channel height was only ~114 nm, i.e., ~95 % of the height in the lithographic master, see the yellow shaded region in Figure 5.10d. The temporary pattern, where a second grating pattern was applied orthogonally (t-direction in Figure 5.10) to the permanent grating pattern (p-direction), was a crosshatched pattern, Figure 5.10b, with the heights of the channels in both directions were about 60% of the original grating pattern, see red and black dashed lines in Figure 5.10d. The recovered pattern, Figure 5.10c retained some of the temporary pattern with the channels in the t-direction having a height of ~18 nm and the recovery, R, of the permanent shape was only 73%. Note also, that the wavelengths of the two crossed grating patterns in the temporary shape were broader than that of the nano-mold, while for the recovered shape, the wavelength of the permanent grating pattern (p-direction) decreased and the temporary grating pattern (t-direction) remained pretty much the same as that in the temporary shape, Figure 5.10d.

101

The lower than expected value of R may be a consequence of impediment of the pattern relaxation due to the surface energy when features are at the nanometer length scale148, 213, or it may be due to the relatively high pressure on the grating protrusions during formation of the temporary shape that may have produced excessive creep of the permanent or temporary networks. Figure 5.10e shows the stress distribution in the crosshatched, temporary pattern, as calculated by finite element analysis (ANSYS software). The maximum stress occurs at the edge of channels.

102

Figure 5.10 Nanopattern shape memory results for SMP-ZnSt. (a, b, c) optical micrograph and AFM image of (a) the imprinted grating pattern (p-direction), (b) crosshatched temporary pattern and (c) recovered pattern; (d) line profiles corresponding to the solid or dashed lines on the AFM image for the permanent, temporary and recovered patterns; (e) stress distribution of temporary shape calculated by finite element analysis (ANSYS).

5.6 Summary

Shape memory compounds based on mixtures of an ionomer with a FA can be used to develop shape memory or shape morphing surfaces with micro- or nano-scale features.

The recovery behavior for the surface nanopattern, however, had lower efficiency than micropattern scale and bulk shape memory of the same material, which may be due to the effects of the excess surface energy on the dynamics of the surface patterns or creep of the temporary or permanent networks156 due to the high stress used to deform the nano-scale grating pattern and produce the temporary crosshatched pattern. Unfortunately, the source of the inefficiency cannot be easily confirmed without measuring the local strain of the micro- or nanoscale texture features during the deformation step used in forming the temporary pattern.

As with the bulk shape memory compounds based on the ionomer/FA design, the switching temperature for micro- and nano-scale surface pattern recovery can be easily tuned by simply changing the specific FA used in the composition. Another feature of the

Zn-SEPDM system that was not used in this study is the ability to crosslink the ionomer

103

using electron beam radiation156. That should improve the creep-resistant behavior of the permanent supramolecular network as was reported in a previous publication156.

104

CHAPTER VI

6 CONCLUSIONS

Surface morphology, especially hierarchical structures have attracted great research interest in the past few decades, both for industrial applications and scientific understanding. There are a lot of nature features which enlightens potential structure- property relationships, ranging from the surface structure of taro and other plants to the feathers of birds and the foot-hairs of geckos.1-3 These highly-organized materials often have specific functional properties that are derived from their multiscale structure, e.g., superhydrophobicity, self-cleaning, water resistance, transport for membranes and the capacity to sustain high loading forces. Traditional methods used for the fabrication of hierarchical structures typically involve the formation of complex patterned features through multistep lithography processes,7-10 that includes photolithography, electron-beam lithography, soft lithography, and ultraviolet (UV) photolithography11, 12. These methods usually require expensive equipments and specialty reagents. Therefore, a low-cost fabrication of well-controlled micro-nano patterns for diverse potential applications remains a challenge. The present research was motivated to design a new technique to 105

generate a dimension adjustable patterns. The pattern structures are further extended to small molecules, nanoparticles, thermoplastics and shape memory supramolecule.

The objectives of this research are: 1) to study the methodology of a new patterning method, which is easy-processing and dimension controllable; 2) to explore the applications, understand the working mechanism and scientific issues of the generated patterns for soft lithography applications, including replica molding, microcontact printing and capillary force lithography; 3) to provide an understanding of the micro and nano pattern memory of a FA/elastomer composites system. For these objectives, 3 projects are developed based on 3 different categories of main materials. For the methodology, PDMS elastomer is used as a “to-be-etched” material for UVO lithography. For soft lithography demonstration, PDMS, small molecules (FOTS), nanoparticles (silver), and thermoplastics

(polystyrene) are selected. In the part of shape memory study on surface pattern, fatty acids and zinc salt of sulfonated EPDM is utilized. Here, to provide different transition temperature, the selected fatty acids include zinc stearate (ZnSt), stearic acid (SA) and lauric acid (LA). Each part has promising scientific and industry applications, and thus show great research significance.

In the first part of this study, we developed a simple and versatile technique for fabrication of dual scale micro-nano hierarchical interfacial patterns in elastomer films by combining nano-imprinting and UVO microlithography. Using this approach, tunable and scalable hierarchically patterned structures in elastomeric polymers can be generated with different nano-pattern masters, UVO irradiation, PDMS thickness, modulus, and micropatterned photomasks. The presented systematic study of the effects of the wall stress, 106

frontal development (gradual diminishment of PDMS densification from surface to bulk interior under UVO exposure), UVO exposure energy and PDMS modulus provided a valuable information for optimization of the pattern transfer technique. Comparison of computer simulations and experimental results for pattern deformation pinpoint the elastocapillarity as a driving force behind pattern deformation at different stages of the pattern transfer process. The future development of this technique could become a valuable tool for micro contact imprinting with potential applications for microfludic devices, biologycally active substrates and electronic systems.

The second part of this study further extends the research by applying the generated micro-nano hierarchical pattern for different forms of soft lithography. These micro-nano patterned molds were prepared by combining UVO lithography and general replica molding (as discussed in the first part of this study). The three forms of soft lithography examined here includes replica molding, microcontact printing and capillary force lithography. For a demonstration of replica molding, PDMS prepolymer is used a model material. The “ink” materials used for microcontact printing are self-assembled monolayer and silver particles. The favorable patterning results in these two lithographies may due to the soft PDMS remaining hard shell at UVO exposed regions. In the process of capillary force lithography (CFL), we investigated the phenomenon of replicating the structures “far away” at the top of PDMS mold even with a relatively thin polymer film. This phenomenon contributes to be an effective means to get a micro pattern with controlled height without sacrificing the fidelity of nano channel pattern by simply controlling the film thickness.

During this approach, bubble defects are observed, however, which were finally eliminated

107

by replacing with a neat soft PDMS mold or by applying Dynamic Zone Annealing (DZA) for capillary force lithography. Furthermore, DZA-CFL functions as a very efficient technique as it’s scalable via Roll to Roll (R2R) processing, providing a potential chance for large area and high fidelity hierarchical patterning.

In the third part, I investigated the shape memory effect of micro and nano patterned surfaces, based on a family of novel shape memory composite materials. In this study, three different FAs, zinc stearate (ZnSt), stearic acid (SA) and lauric acid (LA) were mixed with

Zn salt of sulfonated EPDM, respectively, to investigate the shape memory effect of bulk material and film surface patterns. As with the bulk shape memory compounds based on the ionomer/FA design, the switching temperature for micro- and nano-scale surface pattern recovery can be easily tuned by simply changing the specific FA used in the composition. The shape memory recovery efficiency of the micro and nano scaled surface topography is compared with bulk materials. The recovery behavior for the surface nanopattern, however, has lower efficiency than micropattern scale and bulk shape memory of the same material, which may be due to the effects of the excess surface energy on the dynamics of the surface patterns or creep of the temporary or permanent networks156 due to the high stress used to deform the nano-scale grating pattern and produce the temporary crosshatched pattern.

108

CHAPTER VII

7 REFERENCES

1 K. Autumn, Y. A. Liang, S. T. Hsieh, W. Zesch, W. P. Chan, T. W. Kenny, R.

Fearing and R. J. Full, Nature, 2000, 405, 681-685.

2 M. Sitti and R. S. Fearing, J. Adhes. Sci. Technol., 2003, 17, 1055-1073.

3 A. Geim, S. Dubonos, I. Grigorieva, K. Novoselov, A. Zhukov and S. Y.

Shapoval, Nat. Mater., 2003, 2, 461-463.

4 X. Gao and L. Jiang, Nature, 2004, 432, 36-36.

5 C. Neinhuis and W. Barthlott, Ann. Bot., 1997, 79, 667-677.

6 A. Nakajima, K. Hashimoto, T. Watanabe, K. Takai, G. Yamauchi and A.

Fujishima, Langmuir, 2000, 16, 7044-7047.

7 G. Li, R. Zhu and Y. Yang, Nat. Photonics 2012, 6, 153-161.

8 L. Zhai, F. Ç. Cebeci, R. E. Cohen and M. F. Rubner, Nano Lett., 2004, 4, 1349-

1353.

109

9 D. Xia and S. Brueck, Nano Lett., 2004, 4, 1295-1299.

10 H.-L. Gou, J.-J. Xu, X.-H. Xia and H.-Y. Chen, ACS Appl. Mat. Interfaces 2010,

2, 1324-1330.

11 A. d. Campo and C. Greiner, J. Micromech. Microeng., 2007, 17, R81.

12 Y. Xia and G. M. Whitesides, Annu. Rev. Mater. Sci., 1998, 28, 153-184.

13 C.-Y. Xue, W. Zhang, W. H. S. Choo and K.-L. Yang, Langmuir, 2011, 27,

13410-13414.

14 L. Raymond, L. Nelson, D. Hamilton and W. Kerwin, IEEE T. Compon. Hybr.,

1979, 2, 395-398.

15 H. R. Khaleel, A. I. Abbosh and H. M. Al-Rizzo, Design, fabrication, and testing of flexible antennas, INTECH Open Access Publisher, 2013.

16 https://en.wikipedia.org/wiki/Photolithography.

17 I. Utke, P. Hoffmann and J. Melngailis, J. Vac. Sci. Technol. B, 2008, 26, 1197-

1276.

18 J. T. Fourkas, J. Phys. Chem. Lett, 2010, 1, 1221-1227.

19 E. Betzig, J. Trautman, T. Harris, J. Weiner and R. Kostelak, Science, 1991, 251,

1468-1470.

110

20 S. Sun, K. S. Chong and G. J. Leggett, J. Am. Chem. Soc., 2002, 124, 2414-2415.

21 T. L. Andrew, H.-Y. Tsai and R. Menon, Science, 2009, 324, 917-921.

22 H.-Y. Tsai, H. I. Smith and R. Menon, Opt. Lett., 2008, 33, 2916-2918.

23 C. N. LaFratta, J. T. Fourkas, T. Baldacchini and R. A. Farrer, Angew. Chem. Int.

Ed., 2007, 46, 6238-6258.

24 S. Maruo and J. T. Fourkas, Laser Photonics Rev., 2008, 2, 100-111.

25 D. Yang, S. J. Jhaveri and C. K. Ober, MRS Bull., 2005, 30, 976-982.

26 L. Li, R. R. Gattass, E. Gershgoren, H. Hwang and J. T. Fourkas, Science, 2009,

324, 910-913.

27 D. Craven, Photolithography challenges for the micromachining industry, 1996.

28 Y. Xia, J. A. Rogers, K. E. Paul and G. M. Whitesides, Chem. Rev., 1999, 99,

1823-1848.

29 H. Lan, Updates in Advanced Lithography, 2013.

30 J. Brandup, E. Immergut and E. Grulke, Polymer handbook, 1989.

31 H. Lan, Y. Ding and H. Liu, in Nanoimprint lithography: Principles, processes and materials, Nova Science Publishers, Incorporated, 2011, pp. 169-195.

111

32 L. J. Guo, Adv. Mater., 2007, 19, 495-513.

33 J. E. Petrzelka and D. E. Hardt, J. Micromech. Microeng., 2012, 22, 075015.

34 D. B. Weibel, W. R. DiLuzio and G. M. Whitesides, Nat. Rev. Microbiol., 2007,

5, 209-218.

35 Y. Xia, E. Kim, X.-M. Zhao and J. A. Rogers, Science, 1996, 273, 347.

36 Y. Zhang, C.-W. Lo, J. A. Taylor and S. Yang, Langmuir, 2006, 22, 8595-8601.

37 D. Chandra, J. A. Taylor and S. Yang, Soft Matter, 2008, 4, 979-984.

38 K. A. Heyries and C. L. Hansen, Lab Chip, 2011, 11, 4122-4125.

39 Y. Zhao, M. Li, Q. Lu and Z. Shi, Langmuir, 2008, 24, 12651-12657.

40 D. J. Lipomi, M. A. Kats, P. Kim, S. H. Kang, J. Aizenberg, F. Capasso and G.

M. Whitesides, ACS nano, 2010, 4, 4017-4026.

41 L. Song, M. Zhu, Y. Chen and K. Haraguchi, Polym. J., 2008, 40, 800-805.

42 J. L. Wilbur, A. Kumar, H. A. Biebuyck, E. Kim and G. M. Whitesides,

Nanotechnology, 1996, 7, 452.

43 J. L. Wilbur, A. Kumar, E. Kim and G. M. Whitesides, Adv. Mater., 1994, 6, 600-

604.

112

44 A. Rizzo, M. Mazzeo, M. Palumbo, G. Lerario, S. D'Amone, R. Cingolani and G.

Gigli, Adv. Mater., 2008, 20, 1886-1891.

45 A. Rizzo, M. Mazzeo, M. Biasiucci, R. Cingolani and G. Gigli, Small, 2008, 4,

2143-2147.

46 L. Kim, P. O. Anikeeva, S. A. Coe-Sullivan, J. S. Steckel, M. G. Bawendi and V.

Bulovic, Nano Lett., 2008, 8, 4513-4517.

47 V. Santhanam and R. P. Andres, Nano Lett., 2004, 4, 41-44.

48 L. A. Porter, H. C. Choi, J. Schmeltzer, A. E. Ribbe, L. C. Elliott and J. M.

Buriak, Nano Lett., 2002, 2, 1369-1372.

49 S. T. Han, Y. Zhou, Z. X. Xu, L. B. Huang, X. B. Yang and V. Roy, Adv. Mater.,

2012, 24, 3556-3561.

50 A. P. Quist, E. Pavlovic and S. Oscarsson, Anal. Bioanal. Chem., 2005, 381, 591-

600.

51 E. Delamarche, in Protein Sci., 2004, pp. 31-52.

52 D. B. Weibel, A. Lee, M. Mayer, S. F. Brady, D. Bruzewicz, J. Yang, W. R.

DiLuzio, J. Clardy and G. M. Whitesides, Langmuir, 2005, 21, 6436-6442.

53 J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo and G. M. Whitesides, Chem.

Rev., 2005, 105, 1103-1170.

113

54 S. A. Ruiz and C. S. Chen, Soft Matter, 2007, 3, 168-177.

55 L. Yan, X.-M. Zhao and G. M. Whitesides, J. Am. Chem. Soc., 1998, 120, 6179-

6180.

56 C. J. Campbell, S. K. Smoukov, K. J. Bishop and B. A. Grzybowski, Langmuir,

2005, 21, 2637-2640.

57 M. Mayer, J. Yang, I. Gitlin, D. H. Gracias and G. M. Whitesides, Proteomics,

2004, 4, 2366-2376.

58 A. L. Briseno, S. C. Mannsfeld, M. M. Ling, S. Liu, R. J. Tseng, C. Reese, M. E.

Roberts, Y. Yang, F. Wudl and Z. Bao, Nature, 2006, 444, 913-917.

59 M. Geissler, H. Wolf, R. Stutz, E. Delamarche, U.-W. Grummt, B. Michel and A.

Bietsch, Langmuir, 2003, 19, 6301-6311.

60 M. Mrksich, C. S. Chen, Y. Xia, L. E. Dike, D. E. Ingber and G. M. Whitesides,

P. Natl. A. Sci., 1996, 93, 10775-10778.

61 Y. Xia, D. Qin and G. M. Whitesides, Adv. Mater., 1996, 8, 1015-1017.

62 E. Kim, Y. Xia and G. M. Whitesides, J. Am. Chem. Soc., 1996, 118, 5722-5731.

63 K. Y. Suh, Y. S. Kim and H. H. Lee, Adv. Mater., 2001, 13, 1386-1389.

64 K. Y. Suh and H. H. Lee, Adv. Funct. Mater., 2002, 12, 405-413.

114

65 C. M. Bruinink, M. Péter, P. A. Maury, M. De Boer, L. Kuipers, J. Huskens and

D. N. Reinhoudt, Adv. Funct. Mater., 2006, 16, 1555-1565.

66 D. Ho, J. Zou, B. Zdyrko, K. S. Iyer and I. Luzinov, Nanoscale, 2015, 7, 401-414.

67 K.-Y. Suh, M. C. Park and P. Kim, Adv. Funct. Mater., 2009, 19, 2699-2712.

68 A. Lendlein and S. Kelch, Angew. Chem. Int. Ed., 2002, 41, 2034-2057.

69 P. T. Mather, X. Luo and I. A. Rousseau, Annu. Rev. Mater. Res., 2009, 39, 445-

471.

70 T. Xie, Polymer, 2011, 52, 4985-5000.

71 Q. Meng and J. Hu, Compos. Part A: Appli. S., 2009, 40, 1661-1672.

72 J. Lin and L. Chen, J. Appl. Polym. Sci., 1998, 69, 1563-1574.

73 J. Lin and L. Chen, J. Appl. Polym. Sci., 1998, 69, 1575-1586.

74 R. Weiss, E. Izzo and S. Mandelbaum, Macromolecules, 2008, 41, 2978-2980.

75 T. Xie and I. A. Rousseau, Polymer, 2009, 50, 1852-1856.

76 H. Tobushi, T. Hashimoto, N. Ito, S. Hayashi and E. Yamada, J. Intell. Mater.

Syst. Struct., 1998, 9, 127-136.

115

77 J. H. Yang, B. C. Chun, Y.-C. Chung and J. H. Cho, Polymer, 2003, 44, 3251-

3258.

78 A. Lendlein, M. Behl, B. Hiebl and C. Wischke, Expert Rev. Med. Devices, 2010,

7, 357-379.

79 K. Kratz, U. Voigt and A. Lendlein, Adv. Funct. Mater., 2012, 22, 3057-3065.

80 Z. Yu, Q. Zhang, L. Li, Q. Chen, X. Niu, J. Liu and Q. Pei, Adv. Mater., 2011, 23,

664-668.

81 J. Nji and G. Li, Polymer, 2010, 51, 6021-6029.

82 R. R. Kohlmeyer, M. Lor and J. Chen, Nano Lett., 2012, 12, 2757-2762.

83 H. Meng, P. Xiao, J. Gu, X. Wen, J. Xu, C. Zhao, J. Zhang and T. Chen, Chem.

Commun., 2014, 50, 12277-12280.

84 T. Xie and X. Xiao, Chem. Mater., 2008, 20, 2866-2868.

85 X. Lan, Y. Liu, H. Lv, X. Wang, J. Leng and S. Du, Smart Mater. Struct., 2009,

18, 024002.

86 J. Hu, H. Meng, G. Li and S. I. Ibekwe, Smart Mater. Struct., 2012, 21, 053001.

87 J. Van Humbeeck, Adv. Eng. Mater., 2001, 3, 837-850.

88 Y. Bellouard, Mater. Sci. Eng., A, 2008, 481, 582-589.

116

89 C. L. Dimitris, Shape memory alloys: Modeling and engineering applications,

Springer, 2008.

90 T. Ohki, Q.-Q. Ni, N. Ohsako and M. Iwamoto, Compos. Part A: Appli. S., 2004,

35, 1065-1073.

91 R. Kainuma, Y. Imano, W. Ito, Y. Sutou, H. Morito, S. Okamoto, O. Kitakami, K.

Oikawa, A. Fujita and T. Kanomata, Nature, 2006, 439, 957-960.

92 Y. Zhuohong, H. Jinlian, L. Yeqiu and Y. Lapyan, Mater. Chem. Phys., 2006, 98,

368-372.

93 Y. Wang, G. Zhu, J. Xie, Q. Men, T. Liu and F. Ren, J. Polym. Res., 2014, 21, 1-

94 Z. Tang, D. Sun, D. Yang, B. Guo, L. Zhang and D. Jia, Compos. Sci. Technol.,

2013, 75, 15-21.

95 M. Raja, S. H. Ryu and A. Shanmugharaj, Eur. Polym. J., 2013, 49, 3492-3500.

96 H. Lu and W. M. Huang, Appl. Phys. Lett., 2013, 102, 231910.

97 F. Quadrini, L. Santo and E. A. Squeo, Polym. Plast. Technol. Eng., 2012, 51,

560-567.

98 H. J. Yoo, S. S. Mahapatra and J. W. Cho, J. Phys. Chem. C, 2014, 118, 10408-

10415.

117

99 M. Yoonessi, Y. Shi, D. A. Scheiman, M. Lebron-Colon, D. M. Tigelaar, R.

Weiss and M. A. Meador, ACS Nano, 2012, 6, 7644-7655.

100 K. Wang, X.-g. Yan, F. Ren and X.-p. Cui, Chin. J. Polym. Sci., 2016, 34, 466-

474.

101 X. Wang, J. Zhao, M. Chen, L. Ma, X. Zhao, Z.-M. Dang and Z. Wang, J. Phys.

Chem. B, 2013, 117, 1467-1474.

102 Y. Wang, G. Zhu, X. Cui, T. Liu, Z. Liu and K. Wang, Colloid. Polym. Sci., 2014,

292, 2311-2317.

103 B. Alvarado-Tenorio, A. Romo-Uribe and P. T. Mather, Macromolecules, 2015,

48, 5770-5779.

104 B. Alvarado-Tenorio, A. Romo-Uribe and P. T. Mather, MRS Proceedings, 2013,

1453, mrss12-1453-gg1416-1402.

105 B. Alvarado-Tenorio, A. Romo-Uribe and P. T. Mather, MRS Proceedings, 2012,

1450, mrss12-1450-cc1403-1423.

106 Y. Zhu, J. Hu, H. Luo, R. J. Young, L. Deng, S. Zhang, Y. Fan and G. Ye, Soft

Matter, 2012, 8, 2509-2517.

107 Y. Bai, C. Jiang, Q. Wang and T. Wang, Carbohydr. Polym., 2013, 96, 522-527.

108 M. Y. Razzaq and L. Frormann, Polym. Compos., 2007, 28, 287-293.

118

109 S. Hasan, R. Thompson, H. Emery, A. Nathan, A. Weems, F. Zhou, M. Monroe and D. Maitland, RSC Adv., 2016, 6, 918-927.

110 G. Ji and G. Li, Mater. Des., 2013, 51, 79-87.

111 B. D. Beake and J. F. Smith, Philos. Mag. A, 2002, 82, 2179-2186.

112 C. Likitaporn and S. Rimdusit, Key Eng. Mater., 2015, 659, 373-377.

113 H. Du, Z. Song, J. Wang, Z. Liang, Y. Shen and F. You, Sens. Actuators, A 2015,

228, 1-8.

114 H. Lu, W. M. Huang and J. Leng, Compos. Part B: Eng., 2014, 62, 1-4.

115 H. Lu, Y. Yao, W. M. Huang, J. Leng and D. Hui, Compos. Part B: Eng., 2014,

62, 256-261.

116 H. Lu, J. Yin, B. Xu, J. Gou, D. Hui and Y. Fu, Compos. Part B: Eng., 2016, 100,

146-151.

117 K. Wei, G. Zhu, Y. Tang, X. Li, T. Liu and L. Niu, Compos. Part B: Eng., 2013,

51, 169-174.

118 H. M. Jeong, B. K. Ahn and B. K. Kim, Eur. Polym. J., 2001, 37, 2245-2252.

119 H. M. Jeong, J. H. Song, S. Y. Lee and B. K. Kim, J. Mater. Sci., 2001, 36, 5457-

5463.

119

120 H. Lu and J. Gou, Polym. Adv. Technol., 2012, 23, 1529-1535.

121 H. Lu, W. M. Huang, F. Liang and K. Yu, Materials, 2013, 6, 3742-3754.

122 J. H. Kang, E. J. Siochi, R. K. Penner and T. L. Turner, presented in part at the

242nd American Chemical Society National Meeting and Exposition, Denver, CO;

United States, 2011.

123 G. Fei, G. Li, L. Wu and H. Xia, Soft Matter, 2012, 8, 5123-5126.

124 Y. Cai, J. S. Jiang, B. Zheng and M. R. Xie, J. Appl. Polym. Sci., 2013, 127, 49-

56.

125 Y. Liu, K. Xu, Q. Chang, M. A. Darabi, B. Lin, W. Zhong and M. Xing, Adv.

Mater., 2016, 28, 7758-7767.

126 M. Y. Razzaq, M. Behl, K. Kratz and A. Lendlein, Adv. Mater., 2013, 25, 5730-

5733.

127 D. Habault, H. Zhang and Y. Zhao, Chem. Soc. Rev., 2013, 42, 7244-7256.

128 H. Zhang, J. Zhang, X. Tong, D. Ma and Y. Zhao, Macromol. Rapid Commun.,

2013, 34, 1575-1579.

129 K. Yu, Y. Liu and J. Leng, RSC Adv., 2014, 4, 2961-2968.

130 D. Iqbal and M. H. Samiullah, Materials, 2013, 6, 116-142.

120

131 S. Chen, F. Mo, F. J. Stadler, S. Chen, Z. Ge and H. Zhuo, J. Mater. Chem. B,

2015, 3, 6645-6655.

132 S. Chen, J. Hu and S. Chen, Polym. Int., 2012, 61, 314-320.

133 H. Qin and P. T. Mather, Macromolecules, 2008, 42, 273-280.

134 I. Bellin, S. Kelch and A. Lendlein, J. Mater. Chem., 2007, 17, 2885-2891.

135 T. Chung, A. Romo-Uribe and P. T. Mather, Macromolecules, 2008, 41, 184-192.

136 S. V. Ahir, A. R. Tajbakhsh and E. M. Terentjev, Adv. Funct. Mater., 2006, 16,

556-560.

137 M. Shahinpoor, Y. Bar-Cohen, J. Simpson and J. Smith, Smart Mater. Struct.,

1998, 7, R15.

138 D. K. Shenoy, D. L. Thomsen III, A. Srinivasan, P. Keller and B. R. Ratna, Sens.

Actuators, A 2002, 96, 184-188.

139 D. L. Thomsen, P. Keller, J. Naciri, R. Pink, H. Jeon, D. Shenoy and B. R. Ratna,

Macromolecules, 2001, 34, 5868-5875.

140 Z. Wang, C. Hansen, Q. Ge, S. H. Maruf, D. U. Ahn, H. J. Qi and Y. Ding, Adv.

Mater., 2011, 23, 3669-3673.

121

141 H. Xu, C. Yu, S. Wang, V. Malyarchuk, T. Xie and J. A. Rogers, Adv. Funct.

Mater., 2013, 23, 3299-3306.

142 M. Ebara, K. Uto, N. Idota, J. M. Hoffman and T. Aoyagi, Adv. Mater., 2012, 24,

273-278.

143 M. Ebara, K. Uto, N. Idota, J. M. Hoffman and T. Aoyagi, Int. J. Nanomed., 2014,

9, 117.

144 J. Li, Y. An, R. Huang, H. Jiang and T. Xie, ACS Appl. Mat. Interfaces 2012, 4,

598-603.

145 J.-H. Jang, C. Y. Koh, K. Bertoldi, M. C. Boyce and E. L. Thomas, Nano Lett.,

2009, 9, 2113-2119.

146 Y. Zhao, W. Huang and Y. Fu, J. Micromech. Microeng., 2011, 21, 067007.

147 W. Huang, N. Liu, X. Lan, J. Lin, J. Pan, J. S. Leng, S. Phee, H. Fan, Y. Liu and

T. Tong, Formation of protrusive micro/nano patterns atop shape memory polymers,

2009.

148 J. Li, J. Shim, J. Deng, J. T. Overvelde, X. Zhu, K. Bertoldi and S. Yang, Soft

Matter, 2012, 8, 10322-10328.

149 C. M. Chen and S. Yang, Adv. Mater., 2014, 26, 1283-1288.

150 T. Lv, Z. Cheng, E. Zhang, H. Kang, Y. Liu and L. Jiang, Small, 2016, n/a-n/a.

122

151 N. García-Huete, J. M. Cuevas, J. M. Laza, J. L. Vilas and L. M. León, Polymers,

2015, 7, 1674-1688.

152 Y. Fang, Y. Ni, S.-Y. Leo, C. Taylor, V. Basile and P. Jiang, Nat. Commun.,

2015, 6.

153 S. Reddy, E. Arzt and A. del Campo, Adv. Mater., 2007, 19, 3833-3837.

154 M. Izumi, A. M. Sweeney, D. DeMartini, J. C. Weaver, M. L. Powers, A. Tao, T.

V. Silvas, R. M. Kramer, W. J. Crookes-Goodson and L. M. Mäthger, J. Roy. Soc.

Interface, 2009, rsif20090299.

155 J. Dong and R. Weiss, Macromolecules, 2011, 44, 8871-8879.

156 J. Dong and R. Weiss, Macromol. Chem. Phys., 2013, 214, 1238-1246.

157 I. Duvdevani, R. Lundberg, C. Wood-Cordova and G. Wilkes, Modification of ionic associations by crystalline polar additives, 1986.

158 D. Jackson, J. Koberstein and R. Weiss, J. Polym. Sci., Part B: Polym. Phys.,

1999, 37, 3141-3150.

159 D. L. Handlin Jr and E. L. Thomas, Macromolecules, 1983, 16, 1514-1525.

160 F. Egitto and L. Matienzo, J. Mater. Sci., 2006, 41, 6362-6373.

161 S. Yu, P. Hing and X. Hu, J. Phys. D: Appl. Phys., 2000, 33, 1606.

123

162 J. F. Moulder, J. Chastain and R. C. King, Handbook of X-Ray Photoelectron

Spectroscopy: a Reference Book of Standard Spectra for Identification and Interpretation of XPS Data, Perkin-Elmer Eden Prairie, MN, 1992.

163 B. Schnyder, T. Lippert, R. Kötz, A. Wokaun, V.-M. Graubner and O. Nuyken,

Surf. Sci., 2003, 532, 1067-1071.

164 F. H. Dill, IEEE T. Electron Dev., 1975, 22, 440-444.

165 K. Efimenko, W. E. Wallace and J. Genzer, J. Colloid Interface Sci., 2002, 254,

306-315.

166 J. M. Katzenstein, D. W. Janes, J. D. Cushen, N. B. Hira, D. L. McGuffin, N. A.

Prisco and C. J. Ellison, ACS Macro Letters, 2012, 1, 1150-1154.

167 D. W. Janes, J. M. Katzenstein, K. Shanmuganathan and C. J. Ellison, J. Polym.

Sci., Part B: Polym. Phys., 2013, 51, 535-545.

168 I. C. Sanchez, J. Chem. Phys., 1983, 79, 405-415.

169 K. Ma, J. Rivera, G. J. Hirasaki and S. L. Biswal, J. Colloid Interface Sci., 2011,

363, 371-378.

170 K. Kremer and G. S. Grest, J. Chem. Phys., 1990, 92, 5057-5086.

171 S. Plimpton, J. Comput. Phys., 1995, 117, 1-19.

124

172 B. Andreotti, O. Baumchen, F. Boulogne, K. E. Daniels, E. R. Dufresne, H.

Perrin, T. Salez, J. H. Snoeijer and R. W. Style, Soft Matter, 2016, 12, 2993-2996.

173 B. Andreotti and J. H. Snoeijer, EPL, 2016, 113, 66001.

174 Z. Cao and A. V. Dobrynin, Macromolecules, 2016, 49, 3586-3592.

175 D. P. Holmes, P. T. Brun, A. Pandey and S. Protiere, Soft Matter, 2016, 12, 4886-

4890.

176 H. Liang, Z. Cao and A. V. Dobrynin, Macromolecules, 2016.

177 S. H. Tawfick, J. Bico and S. Barcelo, MRS Bull., 2016, 41, 108-114.

178 Z. Cao and A. V. Dobrynin, Macromolecules, 2015, 48, 443-451.

179 Z. Cao, M. J. Stevens and A. V. Dobrynin, Macromolecules, 2014, 47, 6515-

6521.

180 R. W. Style, J. S. Wettlaufer and E. R. Dufresne, Soft Matter, 2014, 11, 672-679.

181 X. Xu, A. Jagota and C.-Y. Hui, Soft Matter, 2014, 10, 4625-4632.

182 M. De Volder and A. J. Hart, Angew. Chem. Int. Ed., 2013, 52, 2412-2425.

183 N. Nadermann, C.-Y. Hui and A. Jagota, Proc. Natl. Acad. Sci. U. S. A., 2013,

110, 10541-10545.

125

184 J. H. Weijs, B. Andreotti and J. H. Snoeijer, Soft Matter, 2013, 9, 8494-8503.

185 M. Ebara, Sci. Technol. Adv. Mater., 2016.

186 J.-M. Y. Carrillo and A. V. Dobrynin, J. Chem. Phys., 2012, 137, 214902.

187 J.-M. Y. Carrillo and A. V. Dobrynin, Langmuir, 2012, 28, 10881-10890.

188 C. Duprat, S. Protiere, A. Y. Beebe and H. A. Stone, Nature, 2012, 482, 510-513.

189 R. W. Style and E. R. Dufresne, Soft Matter, 2012, 8, 7177-7184.

190 B. Roman and J. Bico, J. Phys.: Condens. Matter, 2010, 22, 493101.

191 B. Pokroy, S. H. Kang, L. Mahadevan and J. Aizenberg, Science, 2009, 323, 237-

240.

192 J. Huang, M. Juszkiewicz, W. H. de Jeu, E. Cerda, T. Emrick, N. Menon and T. P.

Russell, Science, 2007, 317, 650-653.

193 C. Y. Hui, A. Jagota, Y. Y. Lin and E. J. Kramer, Langmuir, 2002, 18, 1394-

1407.

194 M. Owen, Chim. Nouv., 2004, 27-33.

195 H. Lee, S. M. Dellatore, W. M. Miller and P. B. Messersmith, science, 2007, 318,

426-430.

126

196 H.-W. Chien, W.-H. Kuo, M.-J. Wang, S.-W. Tsai and W.-B. Tsai, Langmuir,

2012, 28, 5775-5782.

197 M. E. McGovern, K. M. Kallury and M. Thompson, Langmuir, 1994, 10, 3607-

3614.

198 Y. Xia, M. Mrksich, E. Kim and G. M. Whitesides, J. Am. Chem. Soc., 1995, 117,

9576-9577.

199 J. Sagiv, J. Am. Chem. Soc., 1980, 102, 92-98.

200 M. Wang, J. E. Comrie, Y. Bai, X. He, S. Guo and W. T. Huck, Adv. Funct.

Mater., 2009, 19, 2236-2243.

201 H. Li, W. Yu, J. Xu, C. Yang, Y. Wang and H. Bu, RSC Adv., 2014, 4, 39684-

39690.

202 N. Bowden, W. T. Huck, K. E. Paul and G. M. Whitesides, Appl. Phys. Lett.,

1999, 75, 2557-2559.

203 S. Y. Chou, L. Zhuang and L. Guo, Appl. Phys. Lett., 1999, 75, 1004-1006.

204 E. Schäffer, T. Thurn-Albrecht, T. P. Russell and U. Steiner, EPL (Europhysics

Letters), 2001, 53, 518.

205 L. F. Pease and W. B. Russel, Langmuir, 2004, 20, 795-804.

127

206 L. Wu and S. Y. Chou, Appl. Phys. Lett., 2003, 82, 3200-3202.

207 P. Deshpande, L. F. Pease III, L. Chen, S. Y. Chou and W. B. Russel, Phys. Rev.

E, 2004, 70, 041601.

208 L. Chen, L. Zhuang, P. Deshpande and S. Chou, Langmuir, 2005, 21, 818-821.

209 G. Singh, K. G. Yager, D.-M. Smilgies, M. M. Kulkarni, D. G. Bucknall and A.

Karim, Macromolecules, 2012, 45, 7107-7117.

210 K. G. Yager, N. J. Fredin, X. Zhang, B. C. Berry, A. Karim and R. L. Jones, Soft

Matter, 2010, 6, 92-99.

211 X. Zhang, K. G. Yager, J. F. Douglas and A. Karim, Soft matter, 2014, 10, 3656-

3666.

212 H. Makowski, R. Lundberg, L. Westerman, J. Bock and A. Eisenberg, Adv. Chem.

Ser., 1980, 187, 3.

213 W. L. Lee and H. Y. Low, Sci. Rep., 2016, 6.

214 M. Behl and A. Lendlein, Mater. Today, 2007, 10, 20-28.

128