Citation
Structure-property relationships in thiol-acrylate based main-chain liquid-crystalline elastomers

Material Information

Title:
Structure-property relationships in thiol-acrylate based main-chain liquid-crystalline elastomers
Creator:
Saed, Mohand Osman ( author )
Language:
English
Physical Description:
1 electronic file (132 pages) : ;

Subjects

Subjects / Keywords:
Elastomers ( lcsh )
Polymer liquid crystals ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
In this research, we used a profoundly new approach to synthesize liquid-crystalline elastomers (LCEs) based on using a thiol-acrylate “click” reaction and two-stage thiol-acrylate Michael addition- photopolymerization (TAMAP) reaction, both of which have not previously been investigated for LCE synthesis. The thiol-acrylate reaction was used initially to synthesize polydomain LCEs and then to examine the influence of crosslinking and spacer length. First, the influence of crosslinking on the thermomechanical behavior of LCEs was investigated. The isotropic rubbery modulus, glass transition temperature, and strain-to-failure showed strong dependence on crosslinker amount and ranged from 0.9 MPa, 3°C, and 105% to 3.2 MPa, 25°C, and 853%, respectively. The isotropic transition temperature (Ti) was shown to be influenced by the functionality of the crosslinker, while the crosslinker concentration had no effect. The magnitude of actuation can be tailored by controlling the amount of crosslinker and applied stress. Actuation increased with increasing the applied stress and decreased with greater amounts of crosslinking. Second, we hypothesized that tuning the LC phases in main-chain LCE systems can be achieved by varying the spacer length while maintaining the same mesogen (RM257). By increasing the length of spacers from two to eleven carbons along the spacer backbone (C2 to C11), we can modulate the mesophase from nematic to smectic, tailor the nematic to isotropic transition temperature between 90 and 140°C, and increase the average work capacity from 128 to 262 kJ/m3. Phase segregation and the smectic C phase is achieved at room temperature for the C6, C9, and C11 spacers. Upon heating, these samples transition into the nematic and later, the isotropic phase. Furthermore, this segregation occurs along with polymer chain crystallinity, which increasing the modulus of the networks by an order of magnitude; however, the crystallization rate is highly time dependent on the spacer length and can vary between 5 minutes for the C11 spacer and 24 hours for shorter spacers. A novel TAMAP methodology was implemented to synthesize monodomain LCEs using commercially available starting monomers. A wide range of thermomechanical properties was tailored by adjusting the amount of crosslinker, while the actuation performance was dependent on the amount of applied strain during programming.
Thesis:
Thesis (Ph.D.)--University of Colorado Denver
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Mohand Osman Saed.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
on10088 ( NOTIS )
1008887429 ( OCLC )
on1008887429

Downloads

This item has the following downloads:


Full Text
STRUCTURE-PROPERTY RELATIONSHIPS IN THIOL-ACRYLATE BASED
MAIN-CHAIN LIQUID-CRYSTALLINE ELASTOMERS
by
MOHAND OSMAN SAED
B.S., University of Gezira, 2008
M.S., University of Colorado Denver, 2014
A thesis submitted to Faculty of the Graduate School of the University of Colorado in partial fulfillments of the requirements of the degree of Doctor of Philosophy Mechanical Engineering Program
2017


2017
MOHAND OSMAN SAED
ALL RIGHTS RESERVED


The Thesis For Doctor Of Philosophy Degree by
Mohand Osman Saed has been approved for the Mechanical Engineering Program by
Dana R. Carpenter, Chair Christopher M. Yakacki, Advisor Ronald Rorrer Kai Yu Carl P. Frick
Christopher N. Bowman


Saed, Mohand Osman (Ph.D., Mechanical Engineering Program)
Structure-Property Relationships in Thiol-Acrylate Based Main-Chain Liquid-Crystalline Elastomers Thesis directed by Professor Christopher M. Yakacki
ABSTRACT
In this research, we used a profoundly new approach to synthesize liquid-crystalline elastomers (LCEs) based on using a thiol-acrylate click reaction and two-stage thiol-acrylate Michael addition-photopolymerization (TAMAP) reaction, both of which have not previously been investigated for LCE synthesis. The thiol-acrylate reaction was used initially to synthesize polydomain LCEs and then to examine the influence of crosslinking and spacer length. First, the influence of crosslinking on the thermomechanical behavior of LCEs was investigated. The isotropic rubbery modulus, glass transition temperature, and strain-to-failure showed strong dependence on crosslinker amount and ranged from 0.9 MPa, 3C, and 105% to 3.2 MPa, 25C, and 853%, respectively. The isotropic transition temperature (T;) was shown to be influenced by the functionality of the crosslinker, while the crosslinker concentration had no effect. The magnitude of actuation can be tailored by controlling the amount of crosslinker and applied stress. Actuation increased with increasing the applied stress and decreased with greater amounts of crosslinking. Second, we hypothesized that tuning the LC phases in main-chain LCE systems can be achieved by varying the spacer length while maintaining the same mesogen (RM257). By increasing the length of spacers from two to eleven carbons along the spacer backbone (C2 to Cl 1), we can modulate the mesophase from nematic to smectic, tailor the nematic to isotropic transition temperature between 90 and 140C, and increase the average work capacity from 128 to 262 kJ/m3. Phase segregation and the smectic C phase is achieved at room temperature for the C6, C9, and Cll spacers. Upon heating, these samples transition into the nematic and later, the isotropic phase. Furthermore, this segregation occurs along with polymer chain crystallinity, which increasing the modulus of the networks by an order of magnitude; however, the crystallization rate is highly time dependent on the spacer length and can vary between 5 minutes for
IV


the Cl 1 spacer and 24 hours for shorter spacers. A novel TAMAP methodology was implemented to synthesize monodomain LCEs using commercially available starting monomers. A wide range of thermomechanical properties was tailored by adjusting the amount of crosslinker, while the actuation performance was dependent on the amount of applied strain during programming.
The form and content of this abstract are approved. I recommend its publication.
Approved: Christopher M. Yakacki
v


To my family.
vi


ACKNOWLEDGEMENTS
I would like to thank my advisor, Prof. Christopher M. Yakacki for his guidance, encouragement, and continuous support over the past 5 years. Your passion for research, good work ethic, limitless ideas, and creativity has inspired me greatly. Working in his laboratory has thought me so many skills that will benefit me throughout my career.
I also would like to thank and acknowledge my thesis committee members, Prof. Dana Carpenter, Prof. Ron Rorrer, Prof. Kai Yu, Prof. Carl. Frick, and Prof. Christopher Bowman, who agreed to serve on my PhD committee despite their tense schedules and for their valuable feedback, which has helped me to further understand my research. Special gratitude extents to Prof. Frick, and Prof. Bowman for their willing to come from Laramie, WY, and Boulder, CO.
The interdisciplinary nature of my projects has taught me to collaborate extensively with many groups around the country. Explicitly, I would like to thank our collaborators at University Wyoming (Prof. Carl Frick and Dan Markel), University of Colorado Boulder (Rayshan Visvanathan, Prof.
Noel Clark, Matt McBride, Abeer Alzahrani, and Prof. Chris Bowman) and John Hopkins University (Aurelie Azoug and Vicky Nguyen).
I would like to thank the Smart Materials and Biomechanics Lab (SMAB) members for their support and encouragement. I would like to thank Dr. Amir Torbati, Ravi Patel, Ross Volpe, Nick Traugutt, Sam Mills, Michael Bollinger, Lillian Chatham, and Ryan Anderson for useful discussions and the wonderful time I spent working with them. I would also like to thank my undergraduate students Brandon Mang, Ellana Taylor, Chelsea Starr, and Kristen Bonifield.
Finally, I would like to thank my family for their love, support, and encouragement throughout this work. My wife, Omnia, has been a constant source of love and joy. I could not have accomplished my PhD without her support
Vll


TABLE OF CONTENTS
TABLE OF CONTENTS.................................................................xiii
LIST OF TABLE.......................................................................xi
LIST OF FIGURE.....................................................................xii
CHAPTER ..............................................................................
I. INTRODUCTION AND BACKGROUND.....................................................1
1.1 Liquid Crystals (LC).........................................................1
1.2 Liquid-crystalline Elastomers (LCEs).........................................2
1.2.1 Classification............................................................3
1.2.2 Preparations..............................................................6
1.2.3 Crosslinking History......................................................7
1.2.4 Stress-Strain Behavior....................................................8
1.2.5 Actuation.................................................................9
II. RESEARCH MOTIVATION AND GOALS..................................................11
III. THIOL-ACRYLATE MAIN-CHAIN LIQUID-CRYSTALLINE ELASTOMERS WITH
TUNABLE THERMOMECHANICAL PROPERTIES AND ACTUATION STRAIN............................15
3.1. Abstract.....................................................................15
3.2 Introduction.................................................................16
3.3. Experimental.................................................................19
3.3.1. Materials...............................................................19
3.3.2. Synthesis of Liquid-Crystalline Elastomers..............................20
3.3.3. Gel Fraction Tests......................................................21
3.3.4. X-Ray Scattering........................................................21
3.3.5. Differential Scanning Calorimetry (DSC).................................22
viii


3.3.6. Dynamic Mechanical Analysis (DMA).......................................23
3.3.7. Strain-to-Failure Tests.................................................23
3.3.8. Strain-Actuation Characterization.......................................23
3.4. Results......................................................................24
3.5. Discussion...................................................................32
3.6. Conclusions..................................................................38
3.6. Acknowledgements.............................................................38
IV. MODULATED MESOPHASE LIQUID CRYSTAL ELASTOMERS................................39
4.1. Abstract.....................................................................39
4.2. Introduction.................................................................39
4.3. Results and Discussion.......................................................42
4.4. Conclusions..................................................................54
4.5. Experimental Section.........................................................55
4.6. Acknowledgements.............................................................58
V. TAILORABLE AND PROGRAMMABLE LIQUID-CRYSTALLINE ELASTOMERS
USING A TWO-STAGE THIOL-ACRYLATE REACTION..........................................59
5.1. Main.........................................................................59
5.2. Conclusions..................................................................66
5.3. Acknowledgments..............................................................67
VI. SYNTHESIS OF PROGRAMMABLE MAIN-CHAIN LIQUID-CRYSTALLINE
ELASTOMERS USING A TWO-STAGE THIOL-ACRYLATE REACTION...............................68
6.1. Abstract.....................................................................68
6.2. Introduction.................................................................69
6.3. Protocol.....................................................................71
6.3.1. Preparation of Liquid Crystalline Elastomers LCEs.......................71
ix


6.3.2. Kinetics Study of Two-stage Reaction with Real-time Fourier Transform Infrared.73
6.3.3. Dynamic Mechanical Analysis (DMA)..............................................74
6.3.4. Strain-to-failure Tests........................................................75
6.3.5. Shape Fixity and Actuation Tests...............................................76
6.5. Discussion.........................................................................86
6.6. Acknowledgments....................................................................89
6.7. Materials........................................................................90
VII. CONCLUSIONS AND FUTURE WORK..........................................................91
7.1 Conclusions.........................................................................91
7.2. Recommendations for the Further Work...............................................93
BIBLIOGRAPHY...............................................................................95
APPRNDIX..................................................................................103
A- Wide-Angle X-Ray Scattering Characterizations.......................................103
B- Differential Scanning Calorimetry (DSC).............................................108
C- Dynamic Mechanical Analysis (DMA)...................................................Ill
x


LIST OF TABLE
Table 3-1. Summary of thermal analysis and thermomechanical properties of LCE...........xii
Table 4-1. Summary of DSC and WAXS data for 5 LCE systems tested. Each data point represents n=3.All of the samples contained equal amount of crosslinker. Tc was measured during the 1st heating scan, where as Tsmc, TNi, and AHfwere measured during the 2nd heating scan. The d-spacing valves were calculated from the ID plots see the supporting
information for more details............................................................47
Table 4-2. Dynamic Mechanical Analysis (DMA) behavior for the first and second temperature sweep; the glass transition temperature (Tg) was measured at the peak of tan 5;
(En) is the storage modulus measured at 25 C; where the rubbery modulus (Er) was measured the isotropic temperature TNi + 30 C. The first temperature sweep was performed after being stored for at least 24 hours at room temperature; whereas the second temperature
sweep was performed 5 minutes after the first sweep was completed.......................50
Table 6-1. Chemical Formulations for LCE Systems: Four different LCE systems used in this study. The naming convention is based on the molar ratio of thiol functional groups between PETMP and EDDET. All systems have an excess of 15 mol% acrylate functional groups. It should be noted, FTIR studies tested HHMP as well as DMPA as photoinitiators and reduced the amount DPA catalyst by half to help with the kinetic characterization.
*DPA is diluted in toluene at a ratio of 1:50...........................................85
Table 6-2. Summary of Thermomechanical Properties of LCE Systems: Dynamic Mechanical Analysis (DMA) test shows the thermomechanical properties of the initial polydomain LCE networks formed via the first-stage Michael-addition reaction. Both Ti
and E'r were measured at the lowest point of the storage modulus vs. temperature curve...86
Table 6-3. Materials used in the study...................................................90
xi


LIST OF FIGURE
Figure 1.1.. Polymeric materials displaying liquid crystallinity, (a) Liquid crystal polymer (LCP), (b) liquid crystal polymer networks (LCNs) are heavily crosslinked materials, (c) Liquid
crystal elastomers (LCEs) are lightly crosslinked materials.....................................1
Figure 1.2. Different attachment geometries for the synthesis of LCEs: side chain elastomers with end- on (a) or side-on (b) attached mesogenic side chains and main chain elastomers with
mesogenic units incorporated end-on (c) or side-on (d) into the polymer main chain..............3
Figure 1.3. Three important LC phases. In the nematic phase, the mesogens posses a short-range order and are aligned parallel in a uniform direction, defined by the director. Smectic A phases exhibit a layered structure with the mesogens parallel to the layer normal. In smectic C phases,
the mesogens are additionally tilted towards the layer normal...................................4
Figure 1.4. Photographs of oriented and unoriented nematic elastomers (a) and corresponding X-
ray patterns of the monodomain (b) and the polydomain (c) sample................................5
Figure 1.5., a)The first LCEs systems developed by Finkelmann et al, b) the most recent
chemistry used by White et al, 2015.............................................................7
Figure 1.6. Schematic of stress-strain curve for poly domain LCEs; shown the three regions......8
Figure 1.7., In the LC phase, the polymer backbones experience an anisotropic environment, which leads to an extended chain conformation. At the phase transition to the isotropic phase, the
polymer regains its coiled conformation, giving rise to a macroscopic shape change..............9
Figure 3.1. Schematic of LCE synthesis via thiol-acrylate Michael addition click reaction. The thiol monomers and the mesogens were selected as commercially available monomers. Stoichiometric mixtures of thiol to acrylate functional groups were used to create well-defined LCE networks....................................................................................19
xii


Figure 3.2. Gel fractions analysis for eight LCE networks with varying the concentration and
functionality of the crosslinker. Five samples (n=5) were tested at each composition.............22
Figure 3.3. Polydomain LCEs (a) can be aligned to monodomain (d) using mechanical strain.
The ID and 2D WAXS patterns for an example of 10 tri-thiol crosslinked networks. A representative of 2D WAXS pattern for 0% strain (c) and 80% strain (d). The intensity versus
azimuthal angle were measured at fixed strains of 0 (e) and 80% (f)..............................24
Figure 3.4. Polydomain nematic-to-isotropic transition associated with optical changes, (a) Nematic polydomain LCEs are optically opaque whereas isotropic LCEs are optically transparence, (b) Representative endotherms for eight LCE networks with varying crosslinker amount and functionality, (c) Average TM values along with standard deviations for LCE networks, (d) Average AHf values along with standard deviations for LCE networks. *Represents significant difference with respect to networks using the same functionality crosslinker (p-value < 0.05). +Represents significant difference with respect to equal crosslinker
counterpart (p-value < 0.05). Five samples (n=5) were tested at each composition.................26
Figure 3.5. a) storage modulus (E') and loss tangent (tan 5) traces for an example of 10 tri-thiol crosslinked networks measured at 3 C/min heating rate and 1 Hz frequency in tension mode, the second temperature sweep plot for a) tri-thiol LCE networks and b) tetra-thiol LCE networks.
The glass transition temperature (Tg) was measured at the peak of tan 5. b) Tg as a function of
the cross-linker content for tri- and tetra-thiol- acrylate LCE networks.........................27
Figure 3.6. The effect of the thiol cross-linker content on stress-strain curve measured at 0.2 mm/s at the glass transition temperature of each LCE composition using a) tri- and tetra-functional crosslinkers. The failure strain was defined by the fracture of the LCE sample, b) Failure strain as a function of the cross-linker content for tri- and tetra-thiol-acrylate LCE networks. +Represents significant difference with respect to equal crosslinker counterpart (p-
xin
value < 0.05.
29


Figure 3.7. The effect of the thiol cross-linker content on strain actuation. Samples were relaxed in a stress-free at 120 C for 10 min prior to testing at 5 C/min cooling rate and varying the applied stress in each cycle. A representative plot strain actuation as a function of the thiol crosslinker; a) tri-thiol crosslinker; b) tetra-thiol crosslinker. The magnitude of strain actuation was measured between strain at -20 C and strain at 120 C. The magnitude of strain actuation was plotted as function of applied stress for eight thiol-acrylate LCE networks by arraying thiol crosslinker content; c) tri-thiol cross-liker; d) tetra-thiol crosslinker. A minimum of 3 samples
(n=3) were tested at each condition..................................................................30
Figure 3.8. Work capacity for LCE networks as a function of crosslinking concentration. Work capacity was measured under a constant bias stress of 100 kPa while cooling from the isotropic
state................................................................................................32
Figure 4.1. Schematic of main-chain LCE synthesis via a thiol-acrylate Michael addition reaction. During synthesis, mesogens are in the isotropic (Iso) phase in a present of solvent (toluene) at 60C. The toluene is removed after the reaction is completed to form polydomain samples. The formation of nematic (N) and smectic C (SmC) domains form via phase separation
of mesogen and thiol spacer..........................................................................43
Figure 4.2. (a) Heat flows of five LCE networks with increasing spacer length from C2 to Cll.
Heat flows are shown on the second heating to reset the thermal history of the networks, (b) Comparison of first and second heating cycles of the C9 network. The second heating cycle shows an exotherm once heated above the glass transition temperature due to polymer chain crystallization, (c) Optical images showing isotropic (transparent) to polydomain (opaque)
transition of the five networks as a function of cooling............................................44
Figure 4.3. 2-D WAXS patterns for five LCE networks at room temperature. Diffraction was measured in (a) an unaligned poly domain state and (b) an aligned monodomain state. Alignment was achieved by stretching the samples to 100% engineering strain before analysis...................45
xiv


Figure 4.4. Temperature-controlled WAXS analysis of the LCE system using the C9 spacer. Diffraction patterns reveal the transition from a smectic C to nematic orientation when heated above 80C, while a nematic to isotropic transition occurs when heated above 100C. All images
were taken under 100% engineering strain...........................................................46
Figure 4.5. Storage modulus (E') and loss tangent (tan delta) traces for LCE networks with spacer lengths of C6, C9, and Cl 1. Samples were measured at 3C/min heating rate and 1 Hz frequency in tension. All samples were annealed above TNi and allowed to cool at room temperature for 24 hours before the first temperature sweep to allow the semi-crystallinity to fully form. Samples were tested four times and allowed to set isothermally at 25 C between each sweep for 5, 60, and 120 minutes to show the evolution of the mechanical properties due to polymer chain crystallization. The behavior of the C2 and C3 networks closely resembled that of
the C6 and are thus are only shown in appendix III.................................................49
Figure 4.6. (a) Selected actuation plots of five LCE networks with increasing spacer length from C2 to Cl 1 under a 50 kPa bias stress. Samples were equilibrated above Tv, and cooled at a rate of 5C/min. (b) Average work capacity for each network (n=3). Work capacity was
calculated by multiplying the bias stress by the actuation strain..................................52
Figure 4.7. Photo sequences highlighting multiple functionalities capable within these semicrystalline LCE networks. A C9 stent was synthesize with 15 mol% excess acrylate groups. The 9 mm stent was expanded to 15 mm and photo-crosslinked to lock in mesogen orientation, (a-b)
The LCE stent is capable of demonstrating a 1-way shape-memory effect when heated above its glass transition, (b-c) The LCE stent is also capable of reversible 2-way actuation when heated and cooled around its TM. (d) If the expanded stent is allowed time to develop polymer
crystallinity, it is capable of supporting a 100 g weight, compared to (e) an uncrystallized stent.54
Figure 5.1. (a) A diacrylate mesogen (RM257), dithiol flexible spacer (2,2-(ethylenedioxy) diethanethiol EDDET), and tetra-functional thiol crosslinker (pentaerythritol tetrakis (3-
xv


mercaptopropionate) PETMP) were selected as commercially available monomers. Non-equimolar monomer solutions were prepared with an excess of 15% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2-hydroxyethoxy)-2-methylpropiophenone (HHMP) were added as the respective catalyst and photo-initiator to the solutions, (b) Representative polydomain structure and physical samples demonstrating ability to mould different geometries, (c) A mechanical stress is applied to the polydomain samples to align the mesogens into a temporary monodomain, (d) A photopolymerization reaction is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample. Photo image compares sample before and after stretching and photo-curing, (e) WAXS pattern of aligned sample confirming nematic structure.
(f) POM image of unaligned sample at 20x magnification. Toluene was used as an optional
component to the system to reduce solution viscosity........................................60
Figure 5.2. Polydomain and monodomain LCE samples were subjected to 0 and 100 kPa bias stresses and cooled from 120 to -20C at a rate of 5C/min. Monodomain samples exhibited 45% actuation under zero stress. The monodomain samples in this experiment were programmed by
stretching a poly domain sample to 100% strain and photo-crosslinking for 10 minutes........62
Figure 5.3. (a) Alternating regions in a polydomain LCE are photo-crosslinked, which become resistant to transparent, monodomain alignment when stretched, (b) An unaligned LCE is heated to the isotropic state and crosslinked with a photo-mask. Upon cooling, photo-crosslinked areas
remain isotropic to revealan image..........................................................63
Figure 5.4. cytocompatibility of the TAMAP synthesized LCE was confirmed after both the first and second stages of the reaction using both elution and direct-contact test by an independent laboratory (WuXi AppTec, St. Paul, MN, USA). Cellular response to both (a) direct contact and
(b) elution tests are shown.................................................................66
Figure 6.1. Schematic of Monodomain Programing via a Two-Stage Thiol-Acrylate Reaction.
(a) A diacrylate mesogen (l,4-bis-[4-(3- acryloyloxypropyloxy)benzoyloxy]-2-methylbenzene
xvi


RM 257), dithiol flexible spacer (2,20-(ethylenedioxy) diethanethiol EDDET), and tetra-functional thiol crosslinker (pentaerythritol tetrakis(3-mercaptopropionate) PETMP) were selected as commercially available monomers. Non- equimolar monomer solutions were prepared with an excess of 15 mol% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2-hydroxyethoxy)-2-methylpropiophenone (HHMP) were added as the respective catalyst and photo- initiator to the solutions, (b) Representative polydomain structure forms via Michael addition (first stage) with a uniform cross-link density and latent excess acrylate functional groups, (c) A mechanical stress is applied to the polydomain samples to orient the mesogens into a temporary monodomain, (d)
A photopolymerization reaction (second stage) is used to establish crosslinks between the excess
acrylate groups, stabilizing the monodomain of the sample.........................................81
Figure 6.2. Kinetics Study of Michael Addition Reaction with Real-Time FTIR. (a) Representative two-stage thiol-acrylate reaction kinetics showing conversion as a function of time using DMPA photoinitiator. At the end of first stage, the thiol groups reached near 100% conversion while 22% of acrylate groups were unreacted. At the end of the second stage, unreacted acrylates reached 100% conversion, (b) FTIR absorbance spectra showing the thiol and acrylate conversion before curing at time 0, upon completion of the first stage at 300 minute,
and upon completion of the second stage at 320 minute.............................................82
Figure 6.3. Thermomechanics of TAMAP LCE Systems, (a) Representative strain-to-failure curves of four LCE systems with 15 mol% excess acrylate and varying amount of PETMP crosslinker, (b) Failure strain as a function of PETMP crosslinker, (c) The influence of temperature on failure strain for an LCE system with 15 mol% PETMP. The failure strain is compared alongside the tan d function of the material measured by DMA. (d) Representative glass transition behavior of four LCE systems tested, (e) Image of a stretched LCE specimen with 15 mol% PETMP compared to an untested specimen. Error bars in (b) and (c) represent
standard deviation.
83


Figure 6.4. Shape-Switching Pathways in an LCE. This schematic represents several different pathways available to achieve shape switching in LCEs. A custom dog-bone sample of 15 mol% PETMP is used in this demonstration with an initial shape of (a). Reversible stress-driven actuation is realized between (b-c) by adjusting the temperature about TNI while under a constant bias force (60.6 mN); the shape-memory effect is achieved by following the programming and recovery cycle of (a-b-d-e); and stress-free actuation can be activated thermally between (g-h) after a permanent monodomain has been programmed into the sample in step (f). The legend illustrates mesogen orientation in polydomain, monodomain, and isotropic states. T < TNI and T
>Tni images were taken at 22 and 90 C, respectively................................84
Figure 6.5. Thermomechanical Response in Programmed-Monodomain LCE Systems: (a) Shape fixity represents the efficiency of permanently aligning monodomain and all of samples show fixity above 90%. The magnitude of actuation measured between 22 and 90 C on a hot plate. Error bars represent standard deviation, (b) The magnitude of actuation measured on DMA from
-25 to 120 C, the actuation increase with increasing of applied programming strain.85
Figure 6.5. Thermomechanical Response in Programmed-Monodomain LCE Systems: (a) Shape fixity represents the efficiency of permanently aligning monodomain and all of samples show fixity above 90%. The magnitude of actuation measured between 22 and 90 C on a hot plate. Error bars represent standard deviation, (b) The magnitude of actuation measured on DMA from
-25 to 120 C, the actuation increase with increasing of applied programming strain.....85
Figure A.l. X-ray Scattering profile intensity as founction of azimuthal angle in WAXS..103
Figure A.2. .Numerical claculaion by fitting expermenatal data..........................104
Figure A.3. The ID WAXS patterns for eight LCE networks with different croslinking density.... 105 Figure A.4. The 1 nad 2D WAXS patternts for five LCE networks with different spacer length.... 106
Figure A.5. Temperature-controlled 2D WAXS patterns for C6..................................107
Figure A.6. Temperature-controlled 2D WAXS patterns for Cl..................................107
xvm


Figure B.l. The DSC for C2, 3, 6, and 11..................................................108
Figure B.2. LC phase transtion temperatures as a founction of spacer legth.................85
Figure B.3. DSC for polydomain sample tested before and after stage 2....................110
Figure C.l. Represenative of DMA plots for eight LCE network with different crosslink density.. Ill
Figure C.2. .Represenative of DMA plots for C2 and C3....................................112
Figure C.3. DMA plot for poly domain sample before and after stage 2......................113
xix


CHAPTER I
INTRODUCTION AND BACKGROUND
1.1 Uiquid Crystals (UC)
In 1888, Friedrich Reinitzer identified liquid crystals (LCs) as a state of matter, when he observed cholesterol melt at 144.5 C to a honey-like opaque fluid. Further heating up to 178.5 C, the liquid became clear and transparent.(1) The term liquid crystal is symbolized to the materials that demonstrate order like crystal while maintaining the behavior of a liquid. The molecules that give rise
Liquid crystal main-chain polymers (LCPs) Rigid main chain
Flexible main chain
Tm ->300 CC f->100GPa AS-0%
b
Glassy liquid crystal polymer networks (LCNs)
c
rg-40-120 c E -0.8-2 GPa AS-5%
H3C-Si(CH^O
Cross linker
0
1
Figure 1.1. Polymeric materials displaying liquid crystallinity, a; Liquid crystal polymer (LCP), b; liquid crystal polymer networks (LCNs) are heavily crosslinked materials, c; Liquid crystal elastomers (LCEs) are lightly crosslinked materials.
to this behavior are referred to as a mesogens. Mesogens are rigid molecules made of two to three
linearly connected aromatic rings, with anisotropic architecture. The molecules can be classified
based on shape to rod-like (calamitic) or disk-like (discotic). Mesogens order at the molecular level,
due to the maximization of the interaction energy and minimization the excluded volume (because of
1


their anisotropic shape architecture).(2) LC materials change their phase due to external stimuli and are classified into the subcategories of thermotropic (order depends on temperature), and lyotropic (order depends on the concentration of material in solvent). The interest in these materials has grown rapidly due to their commercial application values especially in the displays industry.(2) LC devices are dominating the market of displays for computers and telecommunication devices, which is now a hundreds of billion dollars market. (3) Recently, LC materials have been used as active components in other applications such as solar-energy,(4) optics and photonics,(5) mechanics, and biomedicine.(6) (7)Polymeric materials exhibiting liquid crystallinity are merging the liquid-crystalline order in the mesogens with elasticity in the polymer, can be classified based on crosslinking to thermoplastics or thermosets. Liquid-crystal polymers (LCPs) are thermoplastics uncrosslinked macromolecules such as Vectran. These materials are typically linear polymers, with melting temperatures (T ) above 300
C and moduli (E) that can exceed 100 GPa. Liquid-crystalline polymer networks (LCNs) and Liquid-crystalline elastomers (LCEs) are considered thermosets due to the existing of a crosslinking agent in their networks (Figure l.l).(S) Crosslinked Liquid-crystalline polymeric materials can also be further classified based on their glass transition temperature (Tg). LCNs are highly crosslinked materials with Tg above room temperature and moduli of approximately 1-2 GPa, whereas LCEs have a Tg below room temperature. (8) The scope of this dissertation work will only be focused on LCEs.
1.2 Liquid-crystalline Elastomers (LCEs)
LCEs are a remarkable class of materials that encompass the properties of both lightly-crosslinked polymer networks (rubber elasticity) and liquid-crystalline order (self-organization).(6, 9, 10) The concept of combining the properties of two subsystems was first proposed by de Gennes et al. in 1975 and was experimentally accomplished by Finkelmann et al. in 1981.(77, 72) The complex structure of LCEs enable several unique mechanical and optical properties; their most fascinating property is the ability to change their shape reversibly in response to external stimuli such as heat(9, 13) or light.(/4-
2


16) Applications for such materials range from sensors and actuators for biological applications such as artificial muscles,(17-19) biomimetic iris lenses,(20) tunable optical gathering devices,(21) cell scaffolds,(22, 23) micro-grippers for robotics,(24) microvalves for microfluidic systems,(25) and organic solar cells.(2d)
1.2.1 Classification
LCEs are highly diverse class of materials and can be classified into many categories. Herein, we will
Figure 1.2. Different attachment geometries for the synthesis of LCEs: side chain elastomers with end- on (a) or side-on (b) attached mesogenic side chains and main chain elastomers with mesogenic units incorporated end-on (c) or side-on (d) into the polymer main chain.
cover the main classes of LCEs. They can be categorized based on the attachment of mesogens to the polymer backbone (side-chains or main-chains), LC phase structure (smectic or nematic), and the domain type (polydomain or monodomain).
3


1.2.1.1 Classification Based on Attachment of the Mesogens to the Polymer Backbone
The structural components that give a rise to the ordered LC phases are called mesogens. They are the foundation of the LC domains and typically composed of two to three linearly connected aromatic rings (rod-like) with flexible ends. These moieties can be directly placed within the polymer backbone to create main-chain LCEs or as a side group (i.e. side-on or end-on LCEs) (Figure 1.2).(6) Before the discovery of main-chain in 1997, early work on LCEs was focused on side-chain LCEs on both nematic and smectic systems.(10) However, main-chain LCEs have attracted more attention due to the direct coupling between mesogenic order and polymer backbone conformations.(27-32) This direct coupling allows main-chain LCEs to exhibit higher degrees of mesogen orientation, mechanical anisotropy, and thus strain actuation, compared to side chain LCEs.(31) This dissertation work will only be dedicated to main-chain LCEs.
Smectic C Smectic A Nematic
Sc SA n
Figure 1.3. Three important LC phases. In the nematic phase, the mesogens possess a short-range order and are aligned parallel in a uniform direction, defined by the director. Smectic A phases exhibit a layered structure with the mesogens parallel to the layer normal. In smectic C phases, the mesogens are additionally tilted towards the layer normal.
1.2.1.2 Classification Based on LC Phase Type
The ordered LC phase state exists in a temperature range in between the solid crystalline and the
4


isotropic disordered liquid state. The three common types of LC phases are (Figure 1.3)(33): Nematic (N) mesogens are oriented in a uniform direction along a director (orientational order); smectic A (SmA) a layered structure with orientational and positional order; and smectic C (SmC) similar to SmA but with the mesogens titled with respect to the director.(34) In general, smectic LCEs have larger actuation, lower failure strain, higher modulus, and greater enthalpy compared to nematic LCEs. This is due to smectic LCEs typically having higher order parameters compared to nematic LCEs.(35) A general belief is that the formation of LC phases in elastomers is dictated by the mesogen structure: the mesogen core and flexible tails.(6, 34, 36, 37) Nematic mesogens should yield nematic LCE systems, and smectic mesogens should correspond to smectic LCEs.(AS) Krause et al. prepared numerous smectic C and nematic LCE systems with a variety of thermomechanical properties by modifying the mesogen structure.
Classification Based on Domain Type
b
1.2.1.3
Figure 1.4. Photographs of oriented and unoriented nematic elastomers (a) and corresponding X-ray patterns of the monodomain (b) and the poly domain (c) sample.
Poly domain LCEs are created in the absence of external fields, where monodomain must be created in a present of external fields. For many applications, in particular actuators, it is essential to prepare monodomain LCEs (also known as liquid single crystalline elastomers (LSCEs)), where the mesogens orient along a reference direction called a "director. A monodomain can be formed temporarily by applying an external stress (/. e. hanging a weight) to a sample, which will align the polymer chains and orient the mesogens in the direction of the stress. Permanent programming of the
5


monodomain can be achieved via a multi-step process, which involves producing a lightly crosslinked gel followed by immediate application of mechanical stress to induce orientation of the mesogens. Once aligned, the reaction is continued to form covalent crosslinks and stabilize the monodomain. (39) Other "one pot" alignment techniques can be performed in the presence of electric fields or by surface alignment (i.e. rubbing polyimide on a glass slide) during polymerization; however, these methods are generally limited to thin film samples.(6) LCEs prepared without an aligning method do not form monodomain. Rather, they tend to form a disordered arrangement of micro-domains, called polydomains, where each individual domain is defined as a region of uniform orientation. In other words, the mesogens are oriented along the director locally but lack orientation globally (Figure 1.4).(6) The polydomain state arises from quenched disorder, which comes from defects during synthesis caused by chain entanglements and crosslinking.(70, 41) Polydomain structures can be viewed under the polarizing optical microscope as a texture, called a Schlieren texture.(40, 42) Polydomain samples are optically opaque because LC domains strongly scatter light depending on their mesogen orientation, while monodomain samples are optically transparent because LC domains are oriented and as a result do not scatter light.(73) Polydomain and monodomain morphologies exist below the isotropic transition temperature (TO and disappear upon heating above T*.
1.2.2 Preparations
The first LCE network was synthesized by Finkelmann et al. utilizing a hydrosilylation reaction and was a polydomain nematic side-chain LCE (Figure 5a).{12) Since then, this reaction has been applied widely to synthesize side and main-chain LCEs.(44-47) The chemistry of main-chain LCEs has since evolved resulting in numerous synthetic approaches that have been used to create and modify materials to enhance the quality, stability, tailorability, reproducibility, accessibility, and to enable responsiveness to a variety of stimuli such as heat or light. Ortiz et al. have used epoxy resins to synthesize main-chain LCEs in a poly domain state. (48) Recently, Pei et al. have utilized exchangeable covalent bonds to repeatedly program the monodomain in epoxy-based main-chain
6


Figure 1.5. a)The first LCEs systems developed by Finkelmann et al, b) the most recent chemistry used by White et al, 2015.
LCEs. (79) It should be noted that epoxy-based LCEs typically have higher Tg values compared to hydrosilylation-based LCEs. Several photo-crosslinking reactions have also been proposed for the formation of main-chain LCE networks using functionalized pre-polymer chains such as polyesters or thiol-ene.(30, 32, 50) Thiol-ene/yne systems have received considerable attention in the field of LCE due to fast polymerization, low volume shrinkage and shrinkage stress, the formation of homogeneous networks, and minimal sensitivity to oxygen inhibition.(5/-J7) Recently, our group introduced a two-stage thiol-acrylate Michael addition-photopolymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic monodomain main-chain LCEs.(55, 56) Several recent examples of using TAMAP methodology to prepare main-chain LCEs have been performed. (5 7-59) The most recent chemistry develops came from White et al. this system utilizes a two-step method of aza-Michael addition and photopolymerization reaction.(99)
1.2.3 Crosslinking History
LCEs can be obtained via two different routes based on their synthesis and crosslinking history. LCEs that are polymerized in the isotropic state at high temperatures or in the presence of a solvent are known as isotropic polydomain nematic elastomers (i-PNEs). Otherwise, samples that are polymerized at temperatures below isotropic transition temperature (Tj) are known as nematic
7


polydomain nematic elastomers (n-PNEs). The resulting two types of polydomain LCEs share some thermomechanical properties with delicate differences in dynamic features.(61) For example, they have the same rubbery moduli and differ in stress-strain, dynamic, and actuation behavior (in preparation by Traugutt et al, 2017). i-PNE samples store more strain compared n-PNEs. More research must be done to fully investigate their differences.
1.2.4 Stress-Strain Behavior
Figure 1.6. Schematic of stress-strain curve for poly domain LCEs. Shown in three regions.
LCEs exhibit unique stress-stain behavior and can be depended mainly on the sample's domain type (monodomain or polydomain). The stress-stain behavior also can be influenced by temperature, LC phase, crosslink density and history, and the direction of the applied stress.(38, 62, 63) For example, polydomain LCEs undergo stress-strain behavior consisting of three regions. Region One is a linear elastic deformation of the polydomain, which occurs at relatively low strain values, and the slope of the stress-strain curve represents the modulus of the materials. Region Two undergoes the polydomain-monodomain transition, which takes place at an intermediate strain where the specimens change from opaque to optically transparent. Ideally, this transition leads to a plateau in the stress-strain curve with a large increase in strain a constant stress (soft elasticity). Region Three is continued
8


deformation of the monodomain. This occurs once the polydomains have all been oriented and the modulus increases more sharply as the polymer chains become aligned. The strain values differentiating these three regions highly depend on temperature, LC phase and the concentration of the crosslinker (Figure 7). (48) The stress-strain behavior of monodomain LCE (highly anisotropic material) is extremely dependent on the direction of the applied stress Monodomain LCEs stretched parallel to the applied stress have a linear elastic response strain. Where, monodomain LCEs stretched perpendicular to the applied stress display similar to polydomain LCEs stress-strain behavior with distinct soft-elasticity plateau. (64)
1.2.5 Actuation
Thermal actuation of LCEs relies on a reversible anisotropic-isotropic transition (TO associated with LC order.(33, 65) To program an LCE for actuation, the mesogens must first be oriented along a director to form a monodomain (i.e. anisotropic mesophase). The polymer chains elongate when the mesogens orient in the nematic or smectic phase, whereas in the isotropic phase they recover. When LCE is cooled from the isotropic to the nematic or smectic phase, the anisotropy of the polymer
Isotropic
T>T,
Figure 1.7. In the LC phase, the polymer backbones experience an anisotropic environment, which leads to an extended chain conformation. At the phase transition to the isotropic phase, the polymer regains its coiled conformation, giving rise to a macroscopic shape change.
chains in the nematic or smectic phase causes extension of the sample along the long axis of the
9


ellipse, which is the axis of the director for main chain materials, during cooling into the ordered phase.(8) After the initial elongation, main-chain systems may continue to elongate when cooled further below the isotropic transition due to the decrease in hairpin defects (polymer chain folding) with decreasing temperature (when the temperature increases the hairpin increases). This process is entirely reversible.(66)
10


CHAPTER II
RESEARCH MOTIVATION AND GOALS
Liquid-crystalline elastomers (LCEs) are smart materials that are known for their ability to undergo reversible thermal actuation due to the change in their liquid-crystalline phase from an anisotropic (nematic or smectic) to isotropic state (Figure 1.7).(33) LCEs can be activated by heat or light and have demonstrated reversible strains up to 400%.(67) As a result, LCEs have been proposed for many sensor and actuator applications, and most particularly as potential artificial muscles.(/(S', 68) Although, liquid crystal technology has experienced an impressive commercial success in display industry, which is multi-billion dollar business.(3) However, LCEs have not yet had the same success, due to many research challenges or unclear principles in our understanding to these fascinating materials. First, research challenges such as complex chemistry, synthesis, and programming conditions, which have all limited manufacturability and scalability of the LCEs for many researchers without extensive chemistry backgrounds. Therefore, the current LCE synthesis methods are not practical for large-scale manufacturing and are only used to produce small-scale samples such as thin fdms,(69, 70) fibers,(50, 71, 72) micro-pillars,(73) or micro-beads.(74, 75) Second, the unclear principles in our understanding to some structure-to-property relationships such as the influence of the crosslinker content and functionality on the LC phase behavior (i.e. phase type and transition temperature) and other thermomechanical properties. Moreover, there is little known about what dictate the LC phase formation in the elastomers. Last, one of the large problems associated with LCEs that prepared during the last three decades, is the lack of the trend in the thermomechanical properties with structure of the materials. Because most of LCEs have been obtained from differently prepared samples that accordingly had different qualities of monodomain, therefore is nearly impossible to compare the properties/ behaviors in these systems.(33) A comprehensive study to develop a new approach to prepare LCEs and carefully study their thermomechanical properties are therefore necessary to enhance manufacturability, understand material behavior, and assist the design
11


in new engineering applications.
In this research we used a profoundly new approach to develop LCEs based on thiol-acrylate click reaction and two-stage Michael addition-photopolymerization (TAMAP) reaction, both of which have not previously been investigated for LCE synthesis. First, the thiol-acrylate reaction is a powerful tool for polymer synthesis. It has been widely used recently, due to fast reaction rate with high yield, no byproduct, with or without solvent, in ambient temperature, with or without catalyst present in the system, and minimal oxygen inhibition.(52) All of these great qualities make this reaction suitable to use as a tool for LCE synthesis with solely commercial available starting LC materials and monomers, this will open the doors researchers to explore these materials in a very facile manner. Second, the TAMAP approach was implemented to create monodomain LCEs. This method is a new paradigm for LCE manufacturing at both large and small size scales by offering a high degree of tailorability not achievable by current methods via clear control of initial and final crosslinking density of the network and initial stretching conditions. For the first time it will be possible to investigate how these crosslinking densities, along with mechanical stretching conditions, influence the shape-fixity and actuation behavior of LCEs. Furthermore, this approach will help overcome the synthesis barrier to allow investigators with diverse backgrounds easier access to LCE research.
The ultimate goal of this work is to establish structure-property relationships in thiol-acrylate based main chain liquid crystalline elastomers. The main hypothesis of this work is that both thiol-acylate click reaction and two-stage thiol-acrylate reaction can be used to increase LCE tailorability and manufacturability in an effort to better understand the associated structure-property relationships. Specific aims to meet our four research objectives are listed below:
Specific aims #1: Investigate the impact of crosslinking (i.e. varying crosslinking concentration and functionality) on the LC phase transition and thermomechanical properties.
The purpose of this aim is to systematically investigate the impact of varying crosslinking density and
12


changing crosslinker functionality on the properties of thiol-acrylate LCE systems. We hypothesize that using a click reaction to produce more uniform networks may reveal clear structure-property relationships in LCEs as well as demonstrate improving in the actuation performance compare to other main-chain LCEs prepared using different synthetic techniques. The evolution in the network architecture influences the coupling between the liquid-crystalline behavior and polymer chains, which has a pronounced effect on the thermomechanical behaviors, including isotropic transition temperature (TO glass transition temperature (Tg), rubbery modulus, failure strain (ef), and the actuation performance of LCEs. This aim will utilize the high efficiency and orthogonality of the thiol-acrylate Michael addition reaction to synthesize well-defined, uniform networks. Such a reaction offers a facile way to tailor the network structure. This aim will highlight and quantify the effect of the crosslinker on influencing a broad range of thermomechanical properties. This will aid to demonstrate how simple modifications to composition can be used to optimize the physical and mechanical properties of these networks for a wide range of potential applications.
Specific aims #2: Examine what dictates the LC phase formation in LCEs? And modulate LC phase structure via varying the spacers length.
The purpose of this aim is to modulate LC phase structure using a single nematic mesogen, RM257, by controlling the thiol spacer length. Herein, we will use series of main-chain LCE systems that are capable of multiple LC phases at room temperature. We hypothesize that two or more LC phases will be realized, smectic C, and nematic. Smectic C phases should be observed when using longer thiol spacers, as these spacers drive a nano-scale segregation of ternary incompatible layers of bulk alkyl thiol functionalized spacers, flexible propylene oxide acrylic terminal chains, and mesogen cores. Segregation of these distinct segments is the main contribution to the formation of the smectic phase.(76) Shorter thiol spacers should not segregate; therefore, they must be engendered a nematic phase. Furthermore, the nematic to isotropic transition temperature (TM) should dependent on the spacer length (mesogen concentration). This approach should allow us to compare the
13


thermomechanical properties of smectic C and nematic LCE actuators with near identical chemical compositions, which has traditionally been extremely difficult to achieve due to vastly different reactions and compositions used in preparation.
Specific aims #3: Explore TAMAP reaction to create tailored bulk monodomian LCE samples capable of hands-free actuation well as offer spatio-temporal control over liquid-crystalline behavior.
As a simple, readily accessible, powerful methodology, we will introduce a previously unexplored approach to synthesize and program main-chain LCEs using TAMAP reaction. Initial polydomain LCE samples can be formed using a thiol-acrylate click reaction with the facile ability to tailor the crosslinking density and polymer structure. If an excess of acrylate groups exists, a second independent photopolymerization reaction can be used to align LCE into monodomain. This approach will offer neat and scalable synthesis of LCEs as well as offers exceptional spatio-temporal control of the second-stage photopolymerization reaction to influence liquid-crystalline behavior.
Specific aims #4: Investigate how the programming of the monodomain is influenced by the initial crosslinking density, initial programed strain, and temperature.
The purpose of this aim is to explore and demonstrate the robust nature of the TAMAP reaction to prepare main-chain LCEs by investigating the influence of crosslinking density and programed strain conditions, and temperature on the thermomechanics of the LCE systems. We will demonstrate a wide range of thermomechanical properties and actuation performance that are achievable using this reaction.
14


CHAPTER III
THIOL-ACRYLATE MAIN-CHAIN LIQUID-CRYSTALLINE ELASTOMERS WITH TUNABLE THERMOMECHANICAL PROPERTIES AND ACTUATION STRAIN
3.1. Abstract
The purpose of this study was to investigate the influence of crosslinking on the thermomechanical behavior of liquid-crystalline elastomers (LCEs). Main-chain LCE networks were synthesized via a thiol-acrylate Michael addition reaction. The robust nature of this reaction allowed for tailoring of the behavior of the LCEs by varying the concentration and functionality of the crosslinker. The isotropic rubbery modulus, glass transition temperature, and strain-to-failure showed strong dependence on crosslinker concentration and ranged from 0.9 MPa, 3C, and 105% to 3.2 MPa, 25C, and 853%, respectively. The isotropic transition temperature (/',) was shown to be influenced by the functionality of the crosslinker, ranging from 70C to 80C for tri- and fefra-functional crosslinkers. The magnitude of actuation can be tailored by controlling the amount of crosslinker and applied stress. Actuation increased with increased applied stress and decreased with greater amounts of crosslinking. The maximum strain actuation achieved was 296% under 100 kPa of bias stress, which resulted in work capacity of 296 kJ/m3 for the lowest crosslinked networks. Overall, the experimental results provide a fundamental insight linking thermomechanical properties and actuation to a homogenous poly domain nematic LCE networks with order parameters of 0.80 when stretched.
15


3.2 Introduction
Liquid-crystalline elastomers (LCEs) are a remarkable class of materials that encompass the properties of both lightly-crosslinked polymer networks (rubber elasticity) and liquid-crystalline order (self-organization). (6, 9, 10) The complex structure of LCEs enable several unique mechanical and optical properties; their most fascinating property is the ability to change their shape reversibly in response to external stimuli such as heat(9, 13) or light.(14-16) Applications for such materials range from sensors and actuators for biological applications such as artificial muscles,(17-19) biomimetic iris lenses,(20) tunable optical gathering devices,(21) cell scaffolds,(22, 23) micro-grippers for robotics,(24) microvalves for microfluidic systems,(25) and organic solar cells. (26)
For many applications, in particular actuators, it is essential to prepare monodomain LCEs, where the mesogens orient along a reference direction called a director. In practice, this can be achieved by applying a mechanical load or an aligning method, such as the use of rubbed polyimide or magnetic fields while crosslinking. (6) LCEs prepared without an aligning method do not form monodomains. Rather, they tend to form a disordered arrangement of micro-domains, called polydomains, where each individual domain is defined as a region of uniform orientation.
In other words, the mesogens are oriented along the director locally but lack orientation globally. The polydomain state arises from quenched disorder, which comes from defects during synthesis caused by chain entanglements and crosslinking.(40, 41) Polydomain structures can be viewed under the polarizing optical microscope as a texture, called a Schlieren texture.(40, 42) Polydomain samples are optically opaque because LC domains strongly scatter light depending on their mesogen orientation, while monodomain samples are optically transparent because LC domains are oriented and as a result do not scatter light.(43) Polydomain and monodomain morphologies exist below the isotropic transition temperature (/',) and disappear upon heating above /',.
16


The first polydomain LCE was synthesized by Finkelmann et al. utilizing a hydrosilylation reaction and was a nematic side-chain LCE.(12) Since then, this reaction has been applied to synthesize main-chain LCEs .(44-47) The chemistry of main-chain LCEs has since evolved resulting in various synthetic approaches that have been used to create and modify materials to enhance the quality, stability, tailorability, reproducibility, accessibility, and to enable responsiveness to a variety of stimuli such as heat or light. Ortiz et al. have used epoxy resins to synthesize main-chain LCEs in a poly domain state. (48) Recently, Pei et al. have utilized exchangeable covalent bonds to repeatedly program the monodomain in epoxy-based main-chain LCEs. (49) It should be noted that epoxy-based LCEs typically have higher Tg values compared to hydrosilylation-based LCEs. Several photo-crosslinking reactions have also been proposed for the formation of main-chain LCE networks using functionalized pre-polymer chains such as polyesters or thiol-ene.(30, 32, 50) Thiol-ene/yne systems have received considerable attention in the field of LCE due to fast polymerization, low volume shrinkage and shrinkage stress, the formation of homogeneous networks, and minimal sensitivity to oxygen inhibition.(5/-5-/) Recently, our group introduced a two-stage thiol-acrylate Michael addition-photopolymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic monodomain main-chain LCEs. (55) Several recent examples of using TAMAP methodology to prepare main-chain LCEs have been performed. (5 7, 59, 77, 78)
Many functional properties of LCEs rely on the reversible anisotropic-isotropic transition associated with LC order. Numerous studies have shown that /', can be correlated to the crosslink density in hydrosilylation-based LCE systems.(79-81) High crosslink densities of the network can disrupt the heat fusion and lower Tu leading to a less stable liquid-crystalline phase, where lightly crosslinking hardly affects 7-,.(79) Tsuchitanti, et al. has shown that the effects of crosslinker geometries have a pronounced effect on the orientation of the mesogens and 7',. The study suggests an increase in functionality of the crosslinker leads to an increase in 1). This is due to the
17


localization of mesogenic monomers in a heterogeneous network structure with a non-uniform distribution crosslinker.(82) Other studies have looked at the influence of crosslinking density on the mechanical properties such as the glass transition temperature (/'). failure strain, the breadth of the soft elasticity plateau and modulus.(35, 83, 84) These studies have found that Tg and modulus increase with an increase in the crosslinking density where the failure strain soft elasticity plateau decreases with the increasing crosslinking density. The synthetic method used in those studies relied on a hydrosilylation reaction. This method leads to random crosslinking. Therefore, the co-relationship between the structure, properties, and actuation performance may be hard to realize.
The purpose of this study is to systematically investigate the impact of varying crosslinking density and changing crosslinker functionality on the properties of thiol-acrylate LCE systems. We hypothesize that using a click reaction to produce more uniform networks may reveal clearer structure-property relationships in LCEs as well as demonstrate improved actuation performance compared to other main-chain LCEs prepared using different synthetic techniques. The evolution in the network architecture influences the coupling between the liquid-crystalline behavior and polymer chains, which has a pronounced effect on the thermomechanical behaviors, including 7i, Tg, rubbery modulus, failure strain (e/), and the actuation performance of LCEs. This study utilized the high efficiency and orthogonality of the thiol-acrylate Michael addition reaction to synthesize well-defined, uniform networks. Such a reaction offers a facile way to tailor the network structure. This work will highlight and quantify the effect of the crosslinker to influence a broad range of thermomechanical properties. This will aid to demonstrate how simple modifications to composition can be used to optimize the physical and mechanical properties of these networks for a wide range of potential applications.
18


3.3. Experimental
3.3.1. Materials
Tri-thiol LCE Networks
Tetra-thiol LCE Networks
Figure 3.1. Schematic of LCE synthesis via thiol-acrylate Michael addition click reaction. The thiol monomers and the mesogens were selected as commercially available monomers. Stoichiometric mixtures of thiol to acrylate functional groups were used to create well-defined LCE networks.
19


Pentaerythritol tetrakis(3-mercaptopropionate) (PETMP), trimethylopropane tris(3-mercaptopropionate) (TMPMP), 2,2-(ethylenedioxy) diethanethiol (EDDET), dipropylamine (DPA), and toluene were purchased from Sigma-Aldrich. 4-bis-[4-(3-
acryloyloxypropypropyloxy) benzoyloxy]-2-methylbenzene (RM257) was obtained from Wilshire Technologies, Inc. (Princeton, NJ, USA). The chemical structures of the monomers and catalyst are shown in Figure 1. All materials were used in their as-received condition without further purification.
3.3.2. Synthesis of Liquid-Crystalline Elastomers
LCE samples were synthesized via a thiol-acrylate Michael addition reaction. LCE networks were prepared starting with two thiol monomers. The thiol monomers were selected for their use as a multi-functional crosslinking monomer and ^/'-functional flexible spacer between mesogens. The flexible spacer (EDDET) was mixed with only one crosslinking monomer at a time, either the tri-functional TMPMP or tetra-functional PETMP. The ratio of thiol crosslinker to flexible spacer was systematically varied using 10, 20, 40, and 80 mol% of functional groups belonging to the crosslinker. Thiol solutions were added to the diacrylate mesogen, RM257, in a stoichiometric balance, which was dissolved in 50 wt% of toluene at 80C for 5 minutes prior to the addition of the thiol solution. Once the solution returned to room temperature, 1 mol% of DPA was added to catalyze the reaction. The solution was mixed vigorously using a Vortex mixer (No: 94540, Toronto, ON, Canada). Air bubbles were removed from the solution under a 500 mm-Hg vacuum. The solution was then injected between two glass slides separated with 1 mm spacers and left to cure overnight. After the polymerization was completed, the samples were placed in an oven for 24 hours at 80 C under a 500 mm-Hg vacuum to remove the solvent.
20


3.3.3. Gel Fraction Tests
LCEs were extracted in toluene for 1 week to determine the swelling ratio and gel fraction, GF, of the networks. LCE films were cut into rectangular samples measuring approximately 22 x 5 x 1 mm3. Each sample was then placed in a vial of 25 mL of toluene for the experiments. After 1 week, samples were removed from the swelling medium, dabbed dry with a paper towel, and dried for 48 hours in vacuum oven at 80 C. The gel fraction, was calculated by:
GF = 100 (3.1)
wt v 7
where Wt is the initial dry weight of the sample and Wf is final weight of the sample. Five samples (n = 5) were tested for each composition.
3.3.4. X-Ray Scattering
In order to investigate the nanostructure of liquid crystal in the network, X-ray analysis was performed using Forvis Technologies wide-angle X-ray scattering (WAXS) 30W Xenocs Genix 3D X-ray source (Cu anode, wavelength = 1.54 A) and Dectris Eiger R 1M detector. The beam size was 0.8 mm X 0.8 mm, and the data was collected at a sample-to-detector distance of 113 mm. The sample was exposed to the X-ray for 30 min. The flux was 4xl07X-rays/s. The scattering patterns were analyzed and plotted using intensity versus azimuthal angle by Rigaku SAXSgui and Igor Pro software to determine the d-spacing of LCEs using the Braggs equation below:
nA= 2d sin 9 (3.2)
where A is the X-ray radiation wavelength (1.5405 A), d is the spacing between long-range ordering of mesogens in LCE network, and 0 is the scattering angle. Data was gathered for sample
21


105
Tri-Thiol Crosslinker n=5
Tetra-Thiol Crosslinker
80l_l--u----i--U---iI-U----l_i_l-L---U
10% 20% 40% 80%
Crosslinker Content (mol)
Figure 3.2. Gel fractions analysis for eight LCE networks with varying the concentration and functionality of the crosslinker. Five samples (n=5) were tested at each composition.
stretch at 0 and 80% strain to identify the crystal structure for both polydomain and monodomain, respectively.
3.3.5. Differential Scanning Calorimetry (DSC)
DSC was performed using a TA Instruments Q2000 machine (New Castle, DE, USA). Samples with a mass of approximately 10 mg were loaded into a standard aluminum DSC pan. The samples were heated rapidly to 120 C at 10 C /min, held isothermally for 10 min, and cooled slowly to -50 C at a rate of 2 C/min to reset any thermal history within the sample. Samples were then heated to 120 C at a rate of 20 C/min. The isotropic transition temperature (T,) was defined as the minimum value of the endothermic peak. The reported enthalpy (AH?) change is measured by integrating the endothermic energy well corresponding to the transition from the nematic polydomain to isotropic state. Statistical analysis was performed on measurements of T; and AHf. ANOVA analysis followed by a Tukey's t-test was first performed to identify any significant differences between values within networks with either tri- or tetra-thiol crosslinkers. A Student's t-test was then performed to identify any differences between samples with equal crosslinker concentrations but different crosslinker functionality. A significant difference was identified when the p-valuc was less than 0.05.
100
95
90
85
22


3.3.6. Dynamic Mechanical Analysis (DMA)
DMA was performed using a TA Instruments Q800 machine (New Castle, DE, USA). Rectangular samples measuring approximately 20 x 10 x 0.8 mm3 were tested in tensile mode, with the active length measuring approximately 10 mm. Samples were cycled at 0.2% strain at 1 Hz and heated from -50 to 120 C at a rate of 3 C/min. T was defined as the temperature corresponding the peak of tan 5 curve. Nematic modulus (E and isotropic modulus (E',) were measured using the storage modulus values at 25 C and 115 C, respectively.
3.3.7. Strain-to-Failure Tests
Strain-to-failure tests were performed using an MTS Insight 30 (Eden Prairie, MN, USA) equipped with an LX-500 laser extensometer, thermal chamber, and 500 N load cell. For these experiments, samples were molded in an HDPE mold according to ASTM Type V dog-bone dimensions at a depth of 1 mm during synthesis. The gage cross-sectional area measured 3 mm x 1 mm. The samples were deformed at Tg at a rate of 0.2 mm/s until failure, defined by sample fracture.
3.3.8. Strain-Actuation Characterization
Strain actuation was measured using the Q800 machine. Sample ends were wrapped with aluminum foil and loaded in the DMA machine in tensile mode with an active length equal to 5 mm. The cross-sectional areas of the samples measured 1x5 mm2. Samples were equilibrated at 120 C. A constant bias stress (obias) was then applied to the samples, while the samples were heated and cooled between 120 and -50 C at 5 C/minute. The stress values investigated were 1, 10, 50, and 100 kPa. The maximum of actuation (ea) was defined by measuring the different between minimum and maximum engineering strain values measured at 120 and -50 C, respectively. The estimated volumetric work capacity of the networks was measured by multiplying the actuation strain by the applied bias stress (Eq. 3).
23


w
Work Capacity = =
^bias^L
LWT
Fbias _
WT L ~ abias£a
(3.3)
The maximum bias stress was selected to be 100 kPa to suit all the tested samples. Stresses greater than 100 kPa frequently caused fracture at the sample-grip interface at elevated temperatures for low-crosslinked samples.
3.4. Results
Figure 3.3. Polydomain LCEs (a) can be aligned to monodomain (d) using mechanical strain. The ID and 2D WAXS patterns for an example of 10 tri-thiol crosslinked networks. A representative of 2D WAXS pattern for 0% strain (c) and 80% strain (d). The intensity versus azimuthal angle were measured at fixed strains of 0 (e) and 80% (f).
Gel fraction analysis was performed as a measure of network formation and conversion in the LCE systems (Figure 2). Eight different compositions were synthesized and tested. LCE compositions are identified by mol% of functional groups belonging to the thiol crosslinker
24


throughout the study. The thiol-acrylate reaction yielded a high gel fraction >90% for all of LCEs networks. Both sets of networks showed an increase in gel fraction from 10 to 80 mol% crosslinker, converging at 99% gel fraction at high crosslink densities. At lower crosslink densities, networks formed with fefra-functional crosslinkers had higher gel fractions than their tri-functional counterparts.
WAXS analysis was performed in order to investigate the short-range and long-range order of the LCEs. The ID and 2D WAXS patterns of the LCE networks were generated when the samples were strained at 0 and 80% to measure the samples in unoriented (polydomain) (Figure 3.3a) and oriented (monodomain) states (Figure 3.3b), respectively; however, only the 10 wt% tri-thiol network is presented as a representative sample. All other samples are shown in the appendix. For
25


A f
w 1
15 mm t < t T > T.
c) 100 o
I 90 -
CD
g>
Q.
E £
c 80 -
o 70 -
Q-
O
60
11 i 11 11 n | 11 1111 m 11111 1111111 11111 m 111 111 11
o Tri-Thiol Crosslinker Tetra-Thiol Crosslinker
S-,
h
1111111111111111111111111
i I i i i i I i i i iti i i i I i M i
20 40 60 80
Crosslinker Content (mol%)
100
d) 1.2 1.0
0.8 .O)
^ 0.6 CO -C
5 0.4
0.2
0.0
111111111111 i 11111111111111 i 111111111 11111111 i 111
O Tri-Thiol Crosslinker _ Tetra-Thiol Crossliker'
7
I I 1 1 1 I I 1 1 I I I 1 I I I I 1 1 I I I I I I I L 1 1 I I I 1 1 I I 1 1 I I 1 I I L 1 I 1 I I
20 40 60 80
Crosslinker Content (mol%)
100
Figure 3.4. Polydomain nematic-to-isotropic transition associated with optical changes, (a) Nematic polydomain LCEs are optically opaque whereas isotropic LCEs are optically transparence, (b) Representative endotherms for eight LCE networks with varying crosslinker amount and functionality, (c) Average T; values along with standard deviations for LCE networks, (d) Average AHf values along with standard deviations for LCE networks. *Represents significant difference with respect to networks using the same functionality crosslinker (p-value < 0.05). +Represents significant difference with respect to equal crosslinker counterpart (p-value < 0.05). Five samples (n=5) were tested at each composition
unoriented samples at 0% strain, no clearly defined peaks were observed in the ID plot of intensity versus azimuthal angle for tri-thiol (Figure 3.3e) conversely, oriented samples at 80% strain all revealed periodic peaks separated by 180 (Figure 3.3f), which is indicative of nematic order a monodomain structure. All unoriented samples showed a diffuse ring in their 2D WAXS patterns, while oriented samples revealed two bright spots separated by 180. Representative 2D patterns are shown in Figures 3.3c and 3d for LCE samples strained at 0 and 80%, respectively. All LCEs exhibited similar scattering patterns with d-spacing values of 3.16 A, indicating that the
26


30
25
20
15
10
-o Tri-Thiol Crosslinker
Tetra-Thiol Crosslinker
A
1111111111111111111111111111111111 111111111
n = 5 -
20 40 60 80
Crosslinker Content (mol%)
100
Figure 4.5. a) storage modulus (E') and loss tangent (tan 5) traces for an example of 10 trithiol crosslinked networks measured at 3 C/min heating rate and 1 Hz frequency in tension mode, the second temperature sweep plot for a) tri-thiol LCE networks and b) tetra-thiol LCE networks. The glass transition temperature (Tg) was measured at the peak of tan 5. b) Tg as a function of the cross-linker content for tri- and tetra-thiol- acrylate LCE networks
microstructure of the LC domains was not affected by their composition. Small-angle X-ray
diffraction (SAXS) was also performed to verily no smectic phases were present (not shown).
Due to the presence of toluene during synthesis and crosslinking, these LCE samples are classified as isotropic polydomain-we/wnhc-elastomers (i-PNEs). The orientation parameter for this system is found to be ~ 0.80.
Thermal analysis was used to investigate the influence of the crosslinker concentration and functionality on LCE networks (Figure 3.4). Representative DSC traces showing heat flow as a function of temperature is shown in Figure 3.4b. All LCE networks show a stepwise decrease in the heat flow signals around temperatures attributed to Tgin the vicinity of 0 to 20 C; however, DSC was primarily used to characterize the isotropic transition. These networks demonstrated endothermic wells, shown as valleys, in the heat flow signals around 70 C for //7-thiol networks and 80 C for tetra-thiol networks. The minimum of the energy well marks the T of the nematic LCE. The //7-thiol networks had an average T, of 70 C with no statistical differences across compositions (Figure 3.4c). Conversely, tetra-thiol networks had significantly higher T; values than their //7-thiol counterparts, with the exception of the 80 mol% tetra-thiol network. The tetra-thiol networks with lower crosslinking amounts (10, 20, and 40 mol%) had an average T, of 80
27


C, while the 80 mol% network had a significantly lower average T, of 72 C. The influence of network structure on AHf was also investigated (Figure 3.4d). There were no significant differences in AHf across all of the networks tested with the exception of the 10 mol% tri-thiol network. The 10 mol% fn'-functional network has an average AHf of 0.955 J/g, which is 80% higher than its fefra-thiol counterpart. Although not statistically significant, a general trend appears to suggest AHf decreased with increasing crosslinker content. Data for each network is presented in Table 3.1. The thermomechanical response of the LCE networks was next investigated using DMA (Figure 3.5). A representative plot for the 10 mol% tri-thiol network (Figure 3.5a) compares the storage modulus (E ) and loss tangent (tan S) as a function of temperature. The sample demonstrated a glassy plateau below 0 C, followed by a stepwise decrease in E that corresponds with the onset of the glass transition (T0nset) for the network. In this study, Tg was measured at the maximum of the tan S curve and ranged from 3 to 25 C. There was a noticeable change in slope and concavity in the E for the LCE networks at Tg.
Furthermore, all LCE networks demonstrated similar DMA behavior (See Figure A3.1) and a dramatic decrease in E at Tb a phenomenon often termed dynamic soft elasticity The samples recover to a rubbery plateau as they are heated into the isotropic phase. It is important to notice that tan S of the networks remain elevated within the nematic region (i.e. between Tg and T,). While some of the samples exhibited one or two secondary peaks in tan S in the elevated region, this behavior was not consistent across all samples tested; however, all samples demonstrated tan S to decrease to a near-zero value once heated above T;. The dependence of Tg on crosslinker amount and functionality are presented in Figure 3.5b. The Tg of the LCE networks increased with increasing amounts of the crosslinker in a near-linear manner. At low crosslinker amounts, the Tg s of the networks are near equal; however, at higher crosslinking densities, the fefra-thiol crosslinked networks have higher Tg values compared to their //7-thiol counterparts. For all networks, the rubbery modulus values measured within the nematic and isotropic states increased
28


Table 1-1. Summary of thermal analysis and thermomechanical properties of LCE.
Crosslinker Crosslinker T onset Tg Ti En Ei AH,
(functionality) (mol%) (C) (C) (C) (MPa) (MPa) (J/g)
10 -1 2 52 68 6 2.5 0.5 0.9 0.2 0.955 0.1
Tri-Thiol 20 02 7 1 72 4 2.5 0.3 0.9 0.1 0.483 0.2
40 22 8 2 74 6 3.8 0.7 1.8 0.1 0.456 0.3
80 9 1 17 1 69 8 10.8 2.8 2.1 0.3 0.155 0.1
10 -4 1 3 1 79 1 1.4 0.1 1.4 0.1 0.534 0.3
Tetra-Thiol 20 -2 2 83 80 3 4.1 1.0 1.6 0.1 0.478 0.3
40 4 1 13 2 86 3 6.4 1.0 1.8 0.4 0.471 0.3
80 11 2 25 1 72 10 24.4 5.0 3.2 0.3 0.142 0.3
crosslinking density. A listing of thermomechanical values from DMA and DSC of the eight LCE
networks is shown in Table 3.1.
The influence of crosslinker concentration and functionality on the stress-strain behavior is shown Figure 3.6. Representative stress-strain curves for //7-thiol and /t'/ra-thiol networks can be seen in Figure 3.6a. All of the networks exhibited stress-strain behavior consisting of three regions. Region One is a linear elastic deformation of the polydomain, which occurs at relatively low strain values, and the slope of the stress-strain curve represents the modulus of the materials.
Figure 3.6. The effect of the thiol cross-linker content on stress-strain curve measured at 0.2 mm/s at the glass transition temperature of each LCE composition using a) tri- and tetra-functional crosslinkers. The failure strain was defined by the fracture of the LCE sample, b) Failure strain as a function of the cross-linker content for tri- and tetra-thiol-acrylate LCE networks. +Represents significant difference with respect to equal crosslinker counterpart (p-value < 0.05.
29


Region Two undergoes the polydomain-monodomain transition, which takes place at an intermediate strain where the specimens change from opaque to optically transparent. Ideally, this transition leads to a plateau in the stress-strain curve with a large increase in strain a constant stress (soft-elasticity); however, several networks only demonstrated a near-plateau in which the slope of the stress curve decreased within this transition (semi-soft elasticity). Region Three is continued deformation of the monodomain. This occurs once the polydomains have all been oriented and the modulus increases more sharply as the polymer chains become aligned. The
0 20 40 60 80 100 0 20 40 60 80 100
Applied Stress (kPa) Applied Stress (kPa)
Figure 3.7. The effect of the thiol cross-linker content on strain actuation. Samples were relaxed in a stress-free at 120 C for 10 min prior to testing at 5 C/min cooling rate and varying the applied stress in each cycle. A representative plot strain actuation as a function of the thiol crosslinker; a) tri-thiol crosslinker; b) tetra-thiol crosslinker. The magnitude of strain actuation was measured between strain at -20 C and strain at 120 C. The magnitude of strain actuation was plotted as function of applied stress for eight thiol-acrylate LCE networks by arraying thiol crosslinker content; c) tri-thiol cross-liker; d) tetra-thiol cross-linker. A minimum of 3 samples (n=3) were tested at each condition.
30


strain values differentiating these three regions highly depend on the concentration of the crosslinker. The strain in each region is shown to decrease with increasing concentration of crosslinker. The failure strain of the networks decreased with increasing concentration of the crosslinker, while the functionality of the crosslinker only had a significant difference at 10 mol% crosslinker (Figure 3.6b). As a result, the 10% tri-thiol network had the highest mean failure strain. Polydomain LCE samples need an applied bias stress in order to orient the mesogens within the network and exhibit thermo-reversible actuation. The magnitude of actuation was measured with respect to crosslinking, crosslinker functionality, and applied stress (Figure 3.7). Representative plots of //7-thiol networks (Figure 3.7a) and fefra-thiol networks (Figure 3.7b) were used to illustrate the strain actuation as samples were cooled from the isotropic state under a constant applied stress of 100 kPa. In both networks, the strain actuation is slowly increased while cooling in the isotropic region. After passing T, the actuation rate sharply increased until the forthcoming /'. below 7). the actuation strain plateaued as the polymer chains were frozen into place due to the vitrification of the network and reduction in chain mobility. The actuation strain increased non-linearly with the decreasing the amount of the crosslinker. The dependence of the actuation strain to the applied stress and the influence of the crosslinking density and functionality of the crosslinker is described in Figures 3.7c and 7d. In general, strain actuation was shown to increase with increased applied stress, while the type of the crosslinker had no apparent effect on the strain actuation. More specifically, the magnitude of strain actuation at a given bias stress decreased with increased crosslinking. The work capacity of the LCE networks actuating under a 100 kPa bias stress is compared in Figure 3.8. Work capacity decreased from average values of 296 to 72 kJ/m3 with increased crosslinker concentration from 10 to 80 mol%. These values were determined by multiplying the bias stress by strain actuation values. Because the bias stress is constant, the work capacity of the LCE networks is influenced by crosslinker concentration and function in the same manner as strain actuation.
31


3.5. Discussion
The purpose of this study was to systematically investigate the impact of (i) varying crosslinking density and (ii) varying crosslinker functionality in thiol-acrylate LCE systems. The network architecture influences the coupling between the liquid-crystalline behavior and polymer chains, which has a pronounced effect on the thermomechanics and the actuation performance of LCEs. This study utilized the high efficiency and orthogonality of the thiol-acrylate Michael addition reaction to synthesize well-defined, uniform networks. Such a reaction offers a facile way to tailor networks with control over thermomechanical properties and actuation performance.
Crosslinker Content (mol%)
Figure 3.8. Work capacity for LCE networks as a function of crosslinking concentration. Work capacity was measured under a constant bias stress of 100 kPa while cooling from the isotropic state.
Polydomain LCEs can be obtained via two different routes based on their synthesis and crosslinking history. Polydomain LCEs that are polymerized in the isotropic state at high temperatures or in the presence of a solvent are known as isotropic polydomain nematic elastomers (i-PNEs). Alternatively, samples that are polymerized at temperatures below T; are known as nematic polydomain nematic elastomers (n-PNEs).(Jf) The resulting two types of poly domain LCEs will display fairly similar thermomechanical properties with subtle differences in dynamic features. (46, 61, 85) Overall, this study focused on the relatively new method of thiol-
32


acrylate Michael-addition chemistry that utilizes a solvent for ease of manufacturing, and thus focused on i-PNEs.
Due to the nature of this reaction, near-complete conversion of thiol-acrylate monomers occurs.(54) Therefore, very high GF was accomplished over wide range of crosslink densities (Figure 3.2). There was a slight decrease in GF in samples with decreasing crosslinker content in both tri- and tetra-thiol LCE networks, which is to be expected due to the decrease in crosslinking density. It is important to note that this was achieved with commercially available materials without purification. All networks had GF values over 90%, which is excellent considering the monomers used in this study had purity values of 95%.
WAXS analysis is a standard procedure of determining LC structure and orientation order in the LCE systems. When the average orientation of mesogenic rods has a preferred direction of alignment, the intensity of Bragg scattering is azimuthally biased (peaks) around the ring (Figure 3.3d). The high-intensity regions in the wide-angle scattering are in the plane parallel to the alignment direction, 180 apart. Azoug et al. showed that for similar thiol-acrylate LCE systems, only 80% strain is needed for a strain-induced monodomain over wide range of temperature regardless of the strain rate. (55) Therefore, all of the samples underwent WAXS analysis at 80% strain to induce an oriented monodomain. WAXS profdes revealed all of the networks to have a nematic structure. The mesogenic monomer, RM257, has also previously been shown to create nematic LCEs when used in free-radical reactions.(51, 86) The orientation parameter for this system is found to be ~ 0.80 (see appendix A). This value might be considered high for a nematic phase, but still lower than the value obtained by Pei, et al. for nematic phase (0.86). It is important to note that all LCEs tested exhibited similar scattering patterns, indicating that the microstructure of the LC domains was not affected by crosslinker concentration or functionality in this system (Figure A.3).
33


Both thermal (DSC) and thermo-mechanical (DMA) analysis was used to characterize the LCE networks. The networks displayed Tg values between 3 and 25 C, which primarily depended on the crosslinker concentration. Previous studies using hydrosilylation reactions produced LCE networks with Tg values ranging from -28 to -12 C.(83) while epoxy based reactions produced high Tg materials that ranged from 46 to 65 C respectively.(84) Conversely, the T, values were on average 70 C for the tri-thiol networks and 80 C for the tetra-thiol networks. T, increased with increasing functionality of the crosslinker, but there was no appreciable effect of the amount of the crosslinker on T;. This latter result is consistent with a previous study by Burke et al. that showed T; was not influenced by increasing the crosslink density in smectic-C main-chain LCEs.(SO) The 80 mol% tetra-thiol composition had a significantly lower T, compared to less crosslinked networks; however, it should be noted that the measurement of T; became more difficult with smaller enthalpic wells at high crosslink densities. The significant effect of the crosslinker functionality on T, can be attributed to the localization of mesogenic monomers around the crosslinks.(82) With lower functionality crosslinkers (f=3), fewer mesogens are present around the network points. On the other hand, tetra-thiol networks have a higher number of mesogens around the network points, which may help stabilize the LC phase and result in a higher temperature required to transition between the nematic and isotropic phase. Nevertheless, increasing the amount of the crosslinker leads to broadening nematic-to-isotropic transition and reducing the magnitude of the enthalpic wells (Figure 3.4b). Networks with low crosslinking density and functionality should be less restrictive and help promote mesogen self-organization.
In this study, the 10 mol% tri-thiol network had the significantly highest AHf values, which is a measure of overall mesogenic order.
The dynamic mechanical response of LCEs has been a subject to increasing investigation due to its unique viscoelastic behavior.(42, 44, 87, 88) All of the reported LCE networks showed a distinctive drop in E as temperature approaches /',. This can be attributed to the dynamic soft
34


elasticity induced by mesogen instability around TV Landau theory states that mesogens can have two energy minimums preferring both order and disorder at Ti.(89) When dynamically cycled, mesogens can rotate between either minimum to lower the energy of the system and result in a drop in E Previous studies have shown that the dynamic stress must be in a direction to allow mesogen rotation, elsewise dynamic soft elasticity is not observed. (90) In this study, all samples tested were in the polydomain state and allowed mesogen rotation when cycled in uniaxial tension. Above Tthe E gradually recovers from the drop and became comparable to the isotropic modulus of conventional amorphous networks. Once cleared into the isotropic region, the value of the E is primarily dictated by the crosslink density of the networks. All networks tested in this study showed distinctive behavior in the tan S loss function, described by an initial peak at Tg followed by elevated values up until T,. The initial peak is attributed to a maximum in damping cause by the viscoelasticity of the polymer chains during the glass transition, while the viscosity and rotation of the liquid crystals causes the elevated values up to T;. Overall, both crosslinker density and functionality did not have a dramatic effect on the loss tangent behavior from our LCE systems, which supports Hotta and Terentjev proposing the use of LCEs as efficient damping materials.(87)
Crosslinking had a pronounced influence on the non-linear stress-strain behavior of the LCE networks. All networks exhibited soft or semi-soft elasticity when strained, which is caused by mesogens orienting themselves along the stretching direction and gradually transforming into a monodomain. This deformation costs almost zero energy and the stress remains nearly constant while the strain increases significantly.(10) The stress-strain behavior of the networks, including the modulus, breadth of soft-elasticity plateau, and failure strain were highly dependent on the crosslinking density. As crosslinking was decreased, the modulus decreased, while the soft-elasticity plateau and failure strain increased. The trends in failure strain as a function of crosslinker concentration (Figure 3.6b) follows the inherent inverse relationship between failure
35


strain as shown in other amorphous networks.(91) Safranski et al. show that the type of crosslinker used only had a significant influence on failure strain at low crosslinking densities for amorphous networks. The results of this study would suggest that while the overall stress-strain behavior of LCEs are markedly different than amorphous networks, the failure strain of LCEs are influenced by the same mechanisms. For example, at low crosslinking density, //7-thiol crosslinked networks had a significantly larger failure strain compared to fefra-thiol crosslinked networks. Increasing the amount of crosslinker not only affects the mechanical properties but also the local and global orientation. Eventually, with an increase in crosslinker, the modulus and strength of the networks will increase at the expense of failure strain and potential risk of preventing the formation of LC order.
The examined LCE networks displayed actuation upon heating and cooling. The actuation relies on the reversible phase transition of the LC domains from an isotropic phase to a globally anisotropic phase (i.e. an aligned monodomain). Hence, reversible thermal actuation in polydomain LCEs is highly dependent on the stress condition (i.e. the applied load) to orient the mesogens along the direction of stress.(SO, 92) It was shown that actuation strain increases with increasing applied stress in a non-linear fashion and approaches a limiting value. The magnitude of actuation diminished from 296 to 72% with increasing crosslink density in both sets of LCE networks. The results would suggest that less entropically restrictive networks (i.e. lower crosslink density) allow for higher magnitudes of actuation; however, the actuation behavior was not influenced significantly by crosslinker functionality. This seems in contrast to the AHf measurements, in which AHf increased as crosslinking was reduced but showed a significant difference in AHf between tri- and tetra-thiol crosslinkers at 10 mol%. Actuation is driven by the energetic contributions of the liquid crystals (enthalpy) and polymer chains (entropy elasticity). Given the 10 mol% crosslinked networks had similar network architecture, they should
36


demonstrate the same entropic contributions during actuation. Since the 10 mol% tri-thiol network had a significantly higher AHh we originally assumed this network would yield higher magnitudes of actuation; although, the 10 mol% tri-thiol networks showed the same actuation as 10 mol% tetra-thiol networks despite having a significantly higher AHf. We are currently investigating the actuation behavior of additional LCE networks with a higher and broader range of enthalpy values to further test our original assumption.
The stress-driven actuation and work capacity of our thiol-acrylate LCEs can be compared to other studies of smectic main-chain LCEs. Li et al. and Burke et al. measured actuation for an epoxy-based and hydrosilylation-based LCEs as a function of bias stress. (SO, 84) With a bias stress of 100 kPa, the actuation values of the two types of LCEs ranged between 63.2-78.7% and 18-21%, respectively. It should be noted that these values were for LCE networks with the lowest crosslink densities in these studies. By comparison, the low crosslinked LCEs in this study showed a mean actuation of approximately 296% and work capacity of 296 kJ/m3 at 100 kPa. Similar reversible actuation under uniaxial load showed work capacity about 96.9 kJ/m3 utilizing a two-step method of aza-Michael addition and photo-polymerization reaction.(93) Ware et al. measured work capacity for a +1 defect to be 3.6 kJ/m3 using similar reaction.(60) Mammalian skeletal muscle ranges from 8-40 kJ/nr'.(9-/)whilc other types of actuators under isotonic loading such as piezoelectric ceramics and electro-resistive polymers can have work capacities of 640 and 1250 kJ/m3 respectively.(95)
One of the most interesting behaviors of LCEs is their ability to actuate. The temperature range for actuation in LCEs is bounded between Tg and T,. The desirable range of values is application dependent. For example, low Tg and high T, will be ideal for applications such as soft robots and sensors, where low Tg and low T, will be ideal for biomedical applications such as a soft-expanding stents.(96) This study suggests that the thiol-acrylate Michael addition reaction may
37


serve as a good platform to tailor main-chain LCEs with enhanced actuation behavior.
3.6. Conclusions
Well-defined nematic, polydomain, main-chain LCE networks were synthesized using a thiol-acrylate Michael addition reaction using both tri-thiol and tetra-thiol crosslinkers. All networks had gel fraction values greater than 90% using commercially available starting materials. The 1-D WAXS patterns were not affected by crosslinking concentration or functionality and showed a nematic structure when strain to 80%. Crosslinker functionality showed a significant influence on the T; of the networks. The average T, for the tri-thiol networks was 70 C, while the average T, for the tetra-thiol networks was 80 C. Increased crosslinking was shown to reduce AHf of the polydomain to isotropic transition. An increase in crosslinking density was shown to increase Tg from 5 to 17 C and 3 to 25 C in tri-thiol and tetra-thiol crosslinked networks, respectively. Crosslinker density had an inverse relationship with failure strain, while crosslinker functionality only had a significant influence at the lowest degree of crosslinking. The average failure strain increased from 542 to 853% from tetra-thiol to tri-thiol networks at 10 mol%, respectively. Crosslinker functionality did not influence thermal actuation behavior, whereas an increase in crosslinking reduced the magnitude of actuation. Networks showed a decrease in work capacity from 296 to 72 kJ/m3 from 10 to 80 mol%.
3.6. Acknowledgements
NSF CAREER Award CMMI-1350436 and the Soft Materials Research Center under NSF MRSEC Grant DMR-1420736 supported this work
38


CHAPTER IV
MODULATED MESOPHASE LIQUID CRYSTAL ELASTOMERS
4.1. Abstract
Control of the mesophase in liquid crystalline elastomers (LCEs) is a critical aspect in harnessing their unique stimuli-responsive properties. Few studies have compared nematic and smectic main-chain LCEs in a direct way. Traditionally, it is believed that the mesogen core and synthetic route determines the phase behavior. In this study, we hypothesized that tuning the LC phases in main-chain LCE systems can be achieved by varying the spacer length while maintaining the same mesogen (RM257). By increasing the length of dithiol alkyl spacers containing two to eleven carbons along the spacer backbone (C2 to Cl 1), we can modulate the mesophase from nematic to smectic, tailor the nematic to isotropic transition temperature between 90 and 140C, and increase the average work capacity from 128 to 262 kJ/m3. Phase segregation and the smectic C phase is achieved at room temperature for the C6, C9, and Cl 1 spacers. Upon heating, these samples transition into the nematic and later, the isotropic phase. Furthermore, this segregation occurs along with polymer chain crystallinity, which increasing the modulus of the networks by an order of magnitude; however, the crystallization rate is highly time dependent on the spacer length and can vary between 5 minutes for the Cl 1 spacer and 24 hours for shorter spacers. This study illuminates several possibilities of the TAMAP reaction in modulation of the thermomechanical and liquid-crystalline properties of LCEs and discusses their potential use for biomedical applications.
4.2. Introduction
The extensive use of liquid-crystalline (LC) materials in modem technologies strongly relies on the modulation of their thermotropic behavior such as mesophase structure, phase transition
39


temperatures, and mesophase stability. In many applications, these materials are used as active mechanical or optical components.(8, 10, 33) The external stimuli-induced mechanical responsiveness phenomenon has attracted a lot of research attention and has been one of the most prolific frontier research areas in materials science recently.(97-99) Among them, liquid crystal elastomers (LCEs) are becoming an increasingly strong competitor in the development of a new generation of actuators because they provide advantages such as flexibility, large actuation strain, light weight, and tailorability.i/00) Such properties make them suitable for many for potential technological applications such as artificial muscles, sensors, and soft robotics.(33, 101) LCEs have already been demonstrated in many applications such as micro-grippers for robotics,(102) micro-electromechanical systems (MEMS) optical grating devices,(103, 104) tunable apertures,(105)and microfluidic systems.(106)
Actuation in LCEs is based on the unique combination of LC order, network elasticity, and chain mobility.(62, 107) The mobility and elasticity enable these supra-molecular systems to respond to different types of external stimuli such as heat or light. The order in these systems can be obtained by maximizing the interactions and minimizing the excluded volume between the LC mesogens, which gives rise to the mechanical anisotropy.(62) The ordered LC phase state exists in a temperature range in between the solid crystalline and the isotropic disordered liquid state. Three common three types of LC phases are: Nematic (N) mesogens are oriented in a uniform direction along a director (orientational order); Smectic A (SmA) a layered structure with orientational and positional order; and Smectic C (SmC) similar to SmA but with the mesogens titled with respect to the dircctor.(5-/) In general, smectic LCEs have larger actuation, lower failure strain, higher modulus, and greater enthalpy compared to nematic LCEs. This is due to smectic LCEs typically having higher order parameters compared to nematic LCEs. (35)
A general belief is that the formation of LC phases in elastomers is dictated by the mesogen structure: the mesogen core and flexible tails.(6, 34, 36, 37) Nematic mesogens should yield
40


nematic LCE systems, and smectic mesogens should correspond to smectic LCEs.(AS) Krause et al. prepared numerous smectic C and nematic LCE systems with a variety of thermomechanical properties by modifying the mesogen structure ;(3 7) however, this traditional method of tuning LC phase is very challenging and not practical for many research groups without extensive chemistry backgrounds. For example, acrylate-functionalized mesogens such as RM82, RM257, and 60BA have been used to prepare smectic A and C (60BA) and nematic LCEs (RM82, RM257).(7&S) This difference in LC phase is typically attributed to having different mesogen core structures (60BA vs. RM82, RM257). Interestingly, we have observed in the literature that smectic and nematic main-chain LCEs can be synthesized from the same mesogen cores. For example, mesogens with the same core structure such as 5Me, RM82, and RM257 have been used to prepare smectic and nematic main-chain LCEs. The Mather research group utilized 5Me as a mesogen to prepare smectic C LCEs,(SO, 109) while mesogens such as RM82 and RM257 have been employed by others to prepare nematic LCEs.(55, 60, 78) 5Me was prepared using a hydrosilylation reaction compared to RM82 and RM257, potentially suggesting that the type of reaction is dictating the LC phase. Looking at Torbati et al. and Saed et al., both used the same mesogen cores, same crosslinkers, and similar thiol-ene/acrylate reactions schemes to produce smectic C and nematic LCEs, respectively.(109, 110) The notable difference was the longer spacer employed by Torbati. This suggests that the formation of the LC phase is highly dictated by not only mesogenic structure but also the spacing between mesogens. Thereby, we hypothesize that modulating the LC phases in main-chain LCE systems could be achieved by varying the spacer length while maintaining the same mesogen. This suggests that a variety of LC phases could be made from similar mesogen structures. Having the ability to tune the LC phase by simply adjusting the spacer opens the door for the first time to further tune the LCE systems with exotic physical properties in a facile manner.
The purpose of this study is to modulate LC phase structure using a single nematic mesogen,
41


RM257, by controlling the thiol spacer length. Herein, we report series of main-chain LCE systems that are capable of multiple LC phases at room temperature. Two desired LC phases were realized, smectic C and nematic. Smectic C phases were observed when using longer thiol spacers, as these spacers drive a nano-scale segregation of ternary incompatible layers of bulk alkyl thiol functionalized spacers, flexible propylene oxide acrylic terminal chains, and mesogen cores. Segregation of these distinct segments is the main contribution to the formation of the smectic phase.(76) Shorter thiol spacers did not segregate; therefore, they engendered a nematic phase. Furthermore, the nematic to isotropic transition temperature (TNi) was highly dependent on the spacer length. As an unexpected result, the use of alkyl spacers resulted in semi-crystallinity within the LCE networks. This semi-crystallinity improved the mechanical properties of the elastomers; however, the rate of crystallization was highly dependent on the length of the spacers. This approach allows us to compare the thermomechanical properties of smectic C and nematic LCE actuators with near identical chemical compositions, which has traditionally been extremely difficult to achieve due to vastly different reactions and compositions used in preparation.
4.3. Results and Discussion
Main-chain LCE networks were synthesized via a thiol-acrylate Michael addition click reaction using a di-acrylate nematic mesogen (RM257), thiol-functionalized alkyl spacers (Cn), and a tetra-thiol crosslinker (PETMP). The reaction was catalyzed using a nucleophilic catalyst (DPA) at 60C in the presence of toluene; therefore, the crosslink history of the networks was established in the isotropic state. The polydomain state formed after the removal of toluene. The thiol-acrylate chemistry was chosen due to the wide availability of thiol and acrylate monomers and mesogens as well as the ability to react using a Michael addition reactions.(53) This synthesis methodology has previously been reported. (55, 111) In a study on the influence of the crosslinker, it was found that work capacity for similar networks increased with decreasing crosslinking density;(110) therefore, 2.5 mol% crosslinker (or 10 mol% of the thiol functional groups) was used for the
42


networks presented herein. The nature of the "click reaction provides a facile manner to tailor the network structure by simply adjusting the spacer length. Five compositions were synthesized with the total number of carbons in the spacer backbone ranging from 2 to 11 (labeled as C2, C3, C6, C9, and C11). The overall reaction scheme and study hypothesis are illustrated in Figure 4.1.
'ioluene
%r.
Iso
Short-range
Order
Figure 4.1. Schematic of main-chain LCE synthesis via a thiol-acrylate Michael addition reaction. During synthesis, mesogens are in the isotropic (Iso) phase in a present of solvent (toluene) at 60C. The toluene is removed after the reaction is completed to form polydomain samples. The formation of nematic (N) and smectic C (SmC) domains form via phase separation of mesogen and thiol spacer.
Differential scanning calorimetry (DSC) was used to initially characterize the phase transitions in
the LCE networks (Figure 4.2a). Spacer length had a significant influence on the isotropic
transition. TM decreased from 140C to 90C by increasing the spacer length from C2 to C11.
Furthermore, the heat of fusion (AHf) associated with this transition increased with spacer length. Two interesting phenomenon were observed as the spacer length increased beyond C6. First, an additional endothermic well appeared in the heat flows. This suggested that a second smectic phase could be present in the materials. Second, the presence of polymer chain crystallinity was observed. Physically, these samples demonstrated a substantial increase in modulus over the course of 24 hours. Comparing the heat flows of the first and second heat cycles showed the
43


samples could be thermally reset (Figure 4.2b). The presence of a crystallization exotherm can be seen followed by two endothermic wells for the C9 sample. Additional DSC experiments can be found in appendix B. The five LCE networks were heated to the isotropic state and allowed to cool slowly to demonstrate TM can be highly tailored using a single mesogen (Figure 4.2c).
Figure 4.2. (a) Heat flows of five LCE networks with increasing spacer length from C2 to Cl 1. Heat flows are shown on the second heating to reset the thermal history of the networks, (b) Comparison of first and second heating cycles of the C9 network. The second heating cycle shows an exotherm once heated above the glass transition temperature due to polymer chain crystallization, (c) Optical images showing isotropic (transparent) to polydomain (opaque) transition of the five networks as a function of cooling.
Wide-angle X-ray scattering (WAXS) was used next to investigate the influence of spacer length on the mesophase of the materials (Figure 4.3). Unstretched, polydomain samples all exhibited diffuse halos at approximately ~1.8 A"1, which is characteristic of unaligned LC domains. Both C9 and C11 samples also clearly exhibited inner halos at -0.24 A"1, which is indicative of smaller-scale order that can be associated with a smectic C phase. Next, an aligned monodomain
44


was created by applying a 100% engineering strain to the samples. The X-ray diffraction patterns evolved to show two bright spots separated by 180 and transverse to the direction of stretching. The C2 and C3 networks demonstrated characteristics of a nematic monodomain, while the C6 to C11 networks revealed four inner bright spots each separated by 90 that is characteristic of a smectic C monodomain. ID diffraction plots for each network can be found in Appendix 1.
C3-M C6-M
1 1 1 Cl
1 I
Figure 4.4. 2-D WAXS patterns for five LCE networks at room temperature. Diffraction was measured in (a) an unaligned polydomain state and (b) an aligned monodomain state. Alignment was achieved by stretching the samples to 100% engineering strain before analysis.
To further explore the mesophase behavior of these systems, the diffraction patterns of the C9
system were measured as a function of temperature (Figure 4.4). It is important to note, these
images were taken on a stretched sample that had 24 hrs since its last thermal cycle to equilibrate
before heating and analysis began. The 2-D WAXS showed a SmC-to-N transition completed by
85C and aN-to-iso transition completed by 120C. The DSC heat flows corroborate these
findings by exhibiting endothermic wells at 75C and 99C. Both C6 and C11 systems show the
same trend in mesophase behavior as a function of temperature and are shown in the supplement.
45


qy[A 1 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.
GJ O o o SmC
f :: a
2.0 1.5 1.0 qy[A l 0.5 0.0 -0.5 -1.0 -1.5 -2.
50 C SmC
0 a
OylA ]
j 70 C SmC
: t
80 C SmC
: 1 ^
Figure 3.4. Temperature-controlled WAXS analysis of the LCE system using the C9 spacer. Diffraction patterns reveal the transition from a smectic C to nematic orientation when heated above 80C, while a nematic to isotropic transition occurs when heated above 100C. All images were taken under 100% engineering strain.
This behavior can be attributed to the molecular structure containing flexible alkyl spacers, propylene oxide acrylic terminal chains, and rod-like mesogenic cores, which can be classified as a rod-coil tri-block system. With increased spacer length, this system can lead to phase separation of the incompatible segments and form self-assembled layered structures similar to ternary amphiphiles systems in ABC tri-block copolymers.(76, 112) The ordered periodic structures construct due to the mutual repulsion interaction forces of the dissimilar chemicals components of those segments and their packing constraints. (112) The LC assemblies of rod-coil tri-block systems may provide a facile means to modulate and influence the LC phase morphology. In the case of using shorter thiol spacers (C2 and C3), the resulting elastomers only exhibited a nematic phase. This can be attributed to the fact that short spacers coupled with the flexible propylene oxide acrylic terminal chains and form miscible new chains that are not amphiphilic, which can prevent phase separation and lead to a single N phase. The smectic C phase is a highly ordered LC phase; therefore, any sample that displayed the smectic C phase at room temperature (C6-C11) was expected to transition through the nematic phase before reaching an isotropic state. In contrast, C2 and C3 are not expected to show a SmC-N phase transition as a function of heating.
46


It is important to note the influence of time dependence in these samples when analyzing both DSC and WAXS data. For WAXS analysis, the nematic phase was not immediately detected in the C9 system if the sample was cooled from the isotropic state; therefore, our analysis was performed on a stretched sample held at room temperature for 24 hours. These results agree with our DSC analysis. During the cooling scan, the SmC-N transition was not detected and only showed up during heating scans. On the other hand, TNi was spotted in both heating and cooling scans with slight differences in temperature, which is due to instrument hysteresis and differences in the heating and cooling rates. During cooling, the smectic C phase is strongly depressed and did not appear on the heat flow traces; however, polymer chain crystallization and smectic C transitions were seen on the second heating. This behavior is due to a kinetic dependence of the phase separation. These LCE networks undergo polymer chain crystallization transition similar to Gelebart et al. 's systems. (78) The onset of the crystallization melting temperature can be revealed in the first heating scan and vanished during the cooling scan and the second heating scan for C9 sample.
Table 4-1. Summary of DSC and WAXS data for 5 LCE systems tested. Each data point represents n=3.All of the samples contained equal amount of crosslinker. Tc was measured during the 1st heating scan, where as TSmC, TM, and AHfwere measured during the 2nd heating scan. The d-spacing valves were calculated from the ID plots see the supporting information for more details.
Spacer Crosslinker (Func.mol%) Tc (C) T smC (C) T Ni (C) AH (J/g) d-spacing (A)

C2 10 92 8 140 5 0.7 0.5 -

C3 10 60 1 133 1 1.7 0.4 -
+ + +
C6 10 66 7 28 2 111 1 1.6 0.5 22.88

C9 10 50 3 75 4 99 1 10.4 1 32.25

Cll 10 53 1 81 1 90 1 13.3 1 33.54
The evolution of thermo-mechanical properties as a function of spacer length and polymer chain crystallization was studied using dynamic mechanical analysis (DMA). Storage modulus (£") and
47


loss tangent (tan 5) traces for the C6, C9, and Cl 1 are shown in Figure 4.5. Overall, the samples displayed distinct aspects of polydomain LCE thermo-mechanical behavior, such as having a broad elevated tan S curve between the glass transition temperature (Tg) and TNi as well as a decrease in modulus at TNi described as dynamic soft elasticity.(do, 57) More interestingly, all of the samples exhibited atypical behavior during the first heating scans, which was characterized by maintaining a high modulus (~100 MPa) when heated above Tg. This is in contrast to a sharp drop in modulus at the onset of the glass transition typically shown previously.(110) For shorter spacer lengths (C2 to C6), the second heating demonstrated this more typical response with a sharp drop in modulus (Figure 4.5a). C2 to C6 samples did not show any significant changes in thermomechanical behavior when allowing the sample to anneal for up to 120 min, indicating a slow crystallization rate (i.e. over 24 hours). The C9 spacer demonstrated the most unique behavior, as the annealed samples underwent recrystallization during testing. This is illustrated by an increase in modulus as the samples were heated above Tg. Each testing condition converged to the same values by 70C, which is just below the SmC-N transition for this system. For the longest spacer, C11 samples did not exhibit any differences in behavior between tests and annealing times (Figure 4.5c). This suggests that 5 minutes is enough to fully recrystallize the sample due to fast crystallization kinetics.
48


4
Figure 4.5. Storage modulus (E') and loss tangent (tan delta) traces for LCE networks with spacer lengths of C6, C9, and C11. Samples were measured at 3C/min heating rate and 1 Hz frequency in tension. All samples were annealed above TNi and allowed to cool at room temperature for 24 hours before the first temperature sweep to allow the semi-crystallinity to fully form. Samples were tested four times and allowed to set isothermally at 25 C between each sweep for 5, 60, and 120 minutes to show the evolution of the mechanical properties due to polymer chain crystallization. The behavior of the C2 and C3 networks closely resembled that of the C6 and are thus are only shown in appendix 3.
A summary and comparison of thermo-mechanical parameters between the first and second heating after 5 minutes is shown in Table 4.2. These data help define the difference between semi-crystallinity within the elastomers. C2 to C9 systems all showed an order of magnitude drop in nematic modulus En from the first to second heating. En of C11 systems did not change during the two tests due to its fast crystallization kinetics. Increasing the spacer length reduced Tg in the networks due to increasing the polymer chain mobility in longer spacers. The C2 system was the only system to have a Tg greater than ambient temperatures, and therefore may be considered a liquid-crystalline glassy network instead of elastomer. Spacer length showed no appreciable
49


Table 4-2. Dynamic Mechanical Analysis (DMA) behavior for the first and second temperature sweep; the glass transition temperature (Tg) was measured at the peak of tan 5; (En) is the storage modulus measured at 25 C; where the rubbery modulus (Er) was measured the isotropic temperature Ti + 30 C. The first temperature sweep was performed after being stored for at least 24 hours at room temperature; whereas the second temperature sweep was performed 5 minutes after the first sweep was completed
Spacer Crosslinker
T,1
E
E
r2
(Func.mol%) _________________^^______________________________________________________________________

C2 10 37 3 28 3 109 24 6 1 1.0 0.1 1.2 0.1

C3 10 21 1 21 1 76 16 4 1 0.7 0.1 1.0 0.4

C6 10 15 1 15 1 150 6 4 0 0.9 0.1 1.0 0.1

C9 10 15 3 11 1 146 28 4 0 0.7 0.1 0.8 0.2

Cll 10 21 1 20 3 140 22 128 177 0.7 0.2 1.0 0.3
differences in the isotropic modulus values, which was to be expected when using a relatively low
and constant amount of crosslinking between systems.
These systems offer a unique opportunity to control the thermo-mechanical properties of LCE networks, especially with respect to the trade-offs in actuator materials. Most materials that show two-way actuation either provide high modulus values with small strain actuation (i.e. such NiTi alloys or piezo-ceramics) or provide larger strain actuation with very low modulus values (i.e. LCEs or dielectrics).(33, 113-115) When it comes to the interplay between actuators and mechanical properties, "the stiffer the better is a traditional premise of a good design, as stiffness improves the precision, stability, and bandwidth of position-control of the device. (116) The use of LCE actuators for applications at room temperature, which typically near or above their Tg, is of particular importance for the sustained interest in their field of research. Traditional LCE materials undergo high strain actuation (around 400%) but the modulus in its deployed state typically in the range of 0.1-10 MPa.(8, 110) This behavior implies that a limitation of LCE actuators is their lack of mechanical strength and modulus after being deployed. In contrast to LCEs, the modulus of NiTi alloy actuators is typically around 60 GPa, but this high modulus is at the expense of the strain actuation (around 8%).{117, 118) Here, we present semi-crystalline LCE systems that exhibit an initial set of distinct mechanical properties suitable for deployment of
50


actuator devices and a second set of properties with an order of magnitude increase in modulus via self-stiffening. This is achieved by the formation of a semi-crystalline structure with variable re-crystallization rates. Other researchers have studied the increased potential for LCEs using dynamic self-stiffening, achieved by cycling the sample to align the mcsogcns.i /19) Enhancing the stiffness without compromising the strain actuation in inherently weak materials may lead to a new paradigm of designing the next generation of LCE actuators, which is possible by engineering the microstructure of the LCE networks.
Tni increased with decreased spacer length due to an intense localization of mesogenic monomers in the networks with short spacers. Hence, a higher temperature (more energy) was required to enable the phase change. This agrees with the results of our previous study that showed an increase in crosslinking functionality, which localized more mesogens around crosslinking netpoints, increased the TM of the networks. (110) On the other hand, the smectic-to-nematic transition temperature (TSN) showed an opposite trend as it decreased with decreasing the spacer length until completely disappearing at shorter spacers (C2 and 3). As stated earlier, the formation of the smectic C phase in this system is due to formation of layers structure and the stability of the layers and SmC phase is proportional to the length of the spacer. Thus, longer spacer required higher temperature (more energy) to disrupt the layers and enable the phase transition. To the best of our knowledge, there is no other similar studies demonstrating these relationships between LC phase transition temperatures, LC phase structure, and spacer length using a single type of reaction and mesogen.
51


The actuation performance of the five LCE networks was then measured as a function of temperature (Figure 4.6a). Samples were equilibrated in the isotropic state before testing began, which served to erase any thermal history and effects of polymer crystallinity at the beginning of the test. The magnitude of actuation strain increased with spacer length. Furthermore, the sharpness and rate of the actuation increased with spacer length. This is due to the range of temperatures decreases for longer spacers, as the Tg values relatively remained the same for all five networks while TM decreased with increased spacer length. The average actuation strain increased from 255 to 525%, which corresponding to volumetric work capacity under 50 kPa bias stress to increase from 128 to 262 kJ/nT (Figure 4.6
Figure 4.6. (a) Selected actuation plots of five LCE networks with increasing spacer length from C2 to C11 under a 50 kPa bias stress. Samples were equilibrated above T,, and cooled at a rate of 5C/min. (b) Average work capacity for each network (n=3). Work capacity was calculated by multiplying the bias stress by the actuation strain.
A stent model was selected to demonstrate the multi-functional potential of these semi-crystalline main-chain LCE samples. It should be noted a C9 LCE network was synthesized with 15 mol% excess acrylate groups to allow a second stage photo-polymerization reaction to program a monodomain in the stent following a previous method.(55, 111) This was to avoid the need of a bias stress during the demonstration and allow hands-free actuation. This study showed that the Tg of the networks could be tailored above room temperature and below body temperature (36C) by
52


utilizing shorter spacer lengths. Furthermore, our previous study showed Tg can be further refined by the amount of crosslinking within the networks.(110) This suggests that these materials can utilize the shape-memory effect around the glass transition to demonstrate a 1-way shape change. Figure 4.7a shows a stent deploy from a packaged shape similar to a previous study on shape-memory polymer (SMP) stents.(/20) The Mather research group has also investigated the shape-memory effect in LCEs to highlight the relevance of combining both phenomenon. Once deployed, the LCE stent can show reversible 2-way actuation by heating and cooling around TM (Figure 4.7b). During this process, the diameter of the stent reduces by approximately 40% when heated and returns to its original shape when cooled. Beilin et al. originally showed a tube-like stent that was capable of first expanding then contracting using a two-step programming process of SMPs:( 121) however, this method was not reversible and relied on two transition temperatures instead of a single TM. The ability to repeatedly and reversible switch the diameter of a stent could potentially solve several challenges in cardiovascular intervention, such as the need to reposition a device, remove a device, or adjust the device as the patient grows during adolescence. Lastly, shape-changing polymers, and especially elastomers, are inherently more compliant than their metallic counterparts. Depending on the application, they can lack the necessary rigidity to provide structural support, such as a stent resisting restenosis after balloon angioplasty. This limitation in shape-memory biomedical applications was previously discussed by Nair et al.(722) Figure 4.7de demonstrates how the LCE networks with semi-crystallinity can provide increased mechanical support. Our data shows LCE networks can undergo self-stiffening and the rate of chain crystallization can be controlled from between 5 minutes to 24 hours.
53


Figure 4.7. Photo sequences highlighting multiple functionalities capable within these semicrystalline LCE networks. A C9 stent was synthesize with 15 mol% excess acrylate groups. The 9 mm stent was expanded to 15 mm and photo-crosslinked to lock in mesogen orientation, (a-b) The LCE stent is capable of demonstrating a 1-way shape-memory effect when heated above its glass transition, (b-c) The LCE stent is also capable of reversible 2-way actuation when heated and cooled around its TM. (d) If the expanded stent is allowed time to develop polymer crystallinity, it is capable of supporting a 100 g weight, compared to (e) an uncrystallized stent.
4.4. Conclusions
A series of main-chain LCE networks with tunable LC mesophases were synthesized using a thiol-acrylate Michael addition reaction using a single nematic mesogen, RM257. The LCE networks exhibited different mesophases by controlling the length of thiol-functionalized alkyl spacers (C2, C3, C6, C9, and C11). Longer spacers (C6, C9, and C11) drive nano-scale segregation resulting in smectic C phases, whereas shorter spacers (C2 and C3) resulted in nematic phases only as confirmed by the 2D WAXS patterns. The length of the spacers showed a significant influence on the thermomechanical properties of the networks such as TM, AHf, and Tg. Tm decreased from 140 to 90C and AHf increased significantly with increasing spacer length.
The networks exhibited semi-crystallinity the rate of crystallization significantly influenced by the spacer length. Shorter spacers (C2, C3, and C6) displayed a slow recrystallization rate after being annealed, whereas C9 recrystallized within 2 to 3 hours and C11 samples only needed 5 minutes to fully recrystallize due to increased spacer flexibility. Smectic C networks demonstrated larger
54


magnitudes of actuation compared to nematic networks, with the C11 network showing an average actuation of 525% 57%. Networks showed an increase in work capacity from 128 to 262 kJ/m3 C2 to C11.
4.5. Experimental Section
4.5.1. Materials: Pentaerythritol tetrakis(3-mercaptopropionate) (PETMP), 1,2-Ethanedithiol (C2), 1,3-Propanedithiol (C3), 1,6-Hexanedithiol (C6), 1,9-Nonanedithiol (C9), 1,11-Undecanedithiol (Cl 1), dipropylamine (DPA), and toluene were purchased from Sigma-Aldrich. 4-bis-[4-(3- acryloyloxypropypropyloxy) benzoyloxy]-2-methylbenzene (RM257) was obtained from Wilshire Technologies, Inc. (Princeton, NJ, USA). The chemical structures of the monomers and catalyst are shown in Figure 4.1. All materials were used in their as-received condition without further purification.
4.5.1. Synthesis of Liquid-Crystalline Elastomers: LCE samples were synthesized via a thiol-acrylate Michael addition reaction. LCE networks were prepared starting with two thiol monomers. The thiol monomers were selected for their use as a tetra-functional cross-linking monomer and di-functional spacer between mesogens. The crosslinker (PETMP) was mixed with only one spacer monomer at a time, C2, C3, C6, C9, or C11. The ratio of thiol crosslinker to spacer was kept constant in all of samples, with 10 mol% of functional groups belonging to the cross-linker and 90 mol% belonging to the spacer. Thiol solutions were added to the diacrylate mesogen, RM257, in a stoichiometric balance unless otherwise stated, which was dissolved in 40 wt% of toluene at 80C for 5 min prior to the addition of the thiol solution. Once the solution returned to room temperature, 1 mol % of DPA was added to catalyze the reaction. The solution was mixed vigorously using a Vortex mixer (No: 94540, Toronto, ON, Canada). Air bubbles were removed from the solution under a 500 mm-Hg vacuum. The solution was then injected between two glass slides separated with 1 mm spacers and left to cure at 60C overnight. After the
55


polymerization was completed, the samples were placed in an oven for 24 h at 80C under a 500 mm-Hg vacuum to remove the solvent.
4.5.2. X-Ray Diffraction: In order to investigate the LC mesophases in the networks, X-ray diffraction was performed using Forvis Technologies wide-angle X-ray scattering (WAXS) 30W Xenocs Genix 3D X-ray source (Cu anode, wavelength = 1.54 A) and Dectris Eiger R 1M detector. The beam size was 0.8 mm x 0.8 mm, and the data was collected at a sample-to-detector distance of 197 mm. the sample was expose to the X-ray for 15 min. The flux was 4xl07X-rays/s. The scattering patterns were analyzed and plotted using intensity versus azimuthal angle by Rigaku SAXSgui and Igor Pro software to determine the d-spacing of LCEs using the Braggs equation below:
nA = 2d sin 9 (1)
where A is the X-ray radiation wavelength (1.5405 A), d is the spacing between long-range ordering of mesogens in LCE network, and 0 is the scattering angle. All of samples were annealed above their isotropic transition temperature and allowed to cool at room temperature for 24 hr prior testing. Data was gathered for all of samples (C2, 3, 6, 9, and Cl 1) at room temperature, stretch at 0 and 100% strain to identify the crystal structure for both poly domain and monodomain, respectively. Further investigation of the influence of the temperature on the nanostructure was done on the samples that exhibited smectic C phase at room temperature (C6, C9 and Cl 1). Samples were tested, while heating at 30, 50, 80, 85, 90, 100, and 120 C.
4.5.3. Differential Scanning Calorimetry (DSC): DSC was performed using a TA Instruments Q2000 machine (New Castle, DE, USA). Samples with a mass of approximately 10 mg were loaded into a standard aluminum DSC pan. All of samples were annealed above their isotropic transition temperature and allowed to cool at room temperature for 24 hr prior testing. The samples were equilibrated at -50 C and heated rapidly to the isotropic state (TNl + 30 C) at 10C/min, to measure the melting transition temperature (Tm) and cooled slowly to -50C at a rate
56


of 2C/min to allow LC self-assembly. Samples were then heated more rapidly to the isotropic state (Tm + 30 C) at a rate of 20C/min. The LC phase transition temperatures (TSNand TNi) were measured at the second heating scan, and defined as the minimum value of the first and the second endothermic peak, respectively. The reported enthalpy (AHf) change is measured by integrating the endothermic energy well corresponding to the transition from the SmC polydomain to N state and enthalpy and from the N polydomain to isotropic state.
4.5.4. Dynamic Mechanical Analysis (DMA): DMA was performed using a TA Instruments Q800 machine (New Castle, DE, USA). Rectangular samples measuring approximately 20 x 5 the isotropic state (TM + 30 C) x 0.8 mm3 were tested in tensile mode, with the active length measuring approximately 6 to 8 mm. Samples were cycled at 0.2% strain at 1 Hz and heated from -50 to the isotropic state (TM + 30 C) at a rate of 3C/min. All of samples were annealed above their isotropic transition temperature and allowed to cool at room temperature for 24 hr prior testing. Samples were temperature swept four times and allowed to set isothermally at 25C between each sweep for 5, 60, and 120 minutes to show elevation of the mechanical properties. Tg was measured at the second temperature sweep and defined as the temperature corresponding the peak of tan 5 curve. The LC modulus (E and isotropic rubbery modulus (E r) were measured using the storage modulus values at 10C and isotropic state (Tm + 30 C), respectively.
4.5.5. Strain-Actuation Characterization: Strain actuation was measured using the Q800 machine. Sample ends were wrapped with aluminum foil and loaded in the DMA machine in tensile mode with an active length equal to 4.2 mm. The cross-sectional areas of the samples measured 1x5 mm2. Samples were equilibrated at the isotropic state (TM + 30 C). A constant bias stress (obias = 50 kPa) was then applied to the samples, while the samples were heated and cooled between the isotropic state and -50C at 5C/minute. The maximum of actuation (ea) was defined by measuring the different between minimum and maximum engineering strain values measured at TNi+ 30C and -50C, respectively. Although, C9 and Cll samples were maximized
57


the DMAs length limit (25 mm) using a gage length of 4 mm. Therefore, only 2mm was allowed as an active gage length between the clamps by coating the rest of the sample with super glue and covering with Teflon tape such that the part was coated with super glue would not actuate. The estimated volumetric work capacity of the networks was measured by multiplying the actuation strain by the applied bias stress (Eq. 2).
W
Work Capacity =
^bias^L
LWT
Fbias _
WT L ~ abias£a
(2)
The maximum bias stress was selected to be 50 kPa to suit all the tested samples. Stresses greater than 50 kPa frequently caused greater actuation but fracture at the sample-grip interface at elevated temperatures for samples with longer spacer length (C9 and Cl 1).
4.5.6 Stent Fabrication: A stent was synthesized and programed using a two-stage thiol-acrylate Michael addition and photopolymerization (TAMAP) reaction described in our previous report.(55) Immediately after adding catalyst, he reaction was added to a two-piece cylindrical glass mold consisting of an inner core (diameter = 8mm) and an outer shell (inner diameter = 12mm) with a height of 6cm. After letting polymerize for 24 hours at 60 C, toluene was evaporated from the sample at 60 C overnight. Then, the sample was stretched radially to a diameter of 16mm and photopolymerized for about 30 minutes, rotating the light source 90 degrees every 7 minutes to evenly cure the sample.
4.6. Acknowledgements
NSF CAREER Award CMMI-1350436 and the Soft Materials Research Center under NSF MRSEC Grant DMR-1420736 supported this work.
58


CHAPTER V
TAILORABLE AND PROGRAMMABLE LIQUID-CRYSTALLINE ELASTOMERS USING A TWO-STAGE THIOL-ACRYLATE REACTION
5.1. Main
This study introduces an unexplored method to synthesize and program liquid-crystalline elastomers (LCEs) based on a two-stage thiol-acrylate Michael addition and photopolymerization (TAMAP) reaction. This methodology can be used to program permanently aligned monodomain samples capable of hands-free shape switching as well as offer spatio-temporal control over liquid-crystalline behaviour. LCE networks were shown to have a cytocompatible response at both stages of the reaction.
Liquid-crystalline elastomers (LCEs) are a class of smart materials that can exhibit reversible mechanical and optical functionalities. These materials incorporate self-organizing mesogenic structures into an elastomeric network to combine the properties of entropy elasticity and liquid-crystalline behaviour. Researchers have proposed LCEs for mechanical actuators,(33) artificial muscles,(18, 68) and switchable surfaces;(723) however, to enable actuation within the material, the mesogens must first be aligned uniformly, creating a liquid-crystalline monodomain (often referred to as a liquid single-crystal elastomer).(6)
The vast majority of main-chain LCEs are synthesized via hydrosilylation reactions, based on a method established by Bergmann and Finkelmann.(724) A multi-step approach is often used to achieve a monodomain in main-chain LCEs: the reaction is allowed to proceed to gelation, a sample is mechanically stretched to align the mesogens, and the reaction to proceeds to crosslink and stabilize the monodomain.(125) Other methods to produce a stabilized monodomain include using surface alignment techniques or magnetic fields to keep
59


Figure 5.1. (a) A diacrylate mesogen (RM257), dithiol flexible spacer (2,2'-(ethylenedioxy) diethanethiol EDDET), and tetra-functional thiol crosslinker (pentaerythritol tetrakis (3-mercaptopropionate) PETMP) were selected as commercially available monomers. Non-equimolar monomer solutions were prepared with an excess of 15% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2-hydroxyethoxy)-2-methylpropiophenone (HHMP) were added as the respective catalyst and photo-initiator to the solutions, (b) Representative polydomain structure and physical samples demonstrating ability to mould different geometries, (c) A mechanical stress is applied to the polydomain samples to align the mesogens into a temporary monodomain, (d) A photopolymerization reaction is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample. Photo image compares sample before and after stretching and photo-curing, (e) WAXS pattern of aligned sample confirming nematic structure, (f) POM image of unaligned sample at 20x magnification. *Toluene was used as an optional component to the system to reduce solution viscosity
mesogens aligned during synthesis via free-radical polymerizations of acrylate- or thiol-ene-
functionalized mesogens;(126, 127) however, these techniques have been limited to thin films or micro-geometries.(123)
As a simple, readily accessible, powerful methodology, we introduce a previously unexplored approach to synthesize and program main-chain LCEs using a two-stage thiol-acrylate Michael addition and photopolymerization (TAMAP) reaction. Initial poly domain LCE samples can be formed using a thiol-acrylate "click reaction with the facile ability to tailor the crosslinking density and polymer structure. If an excess of acrylate groups exists, a
60


second independent photopolymerization reaction can be used to further tailor the properties of the poly domain or stabilize an aligned monodomain. This approach offers elegant and scalable synthesis of LCEs as well as offers unprecedented spatio-temporal control over liquid-crystalline behavior.
A schematic of the two-stage TAMAP reaction is presented in Figure 5.1. For this study, commercially available starting materials with no additional purification were chosen to demonstrate the efficacy of the approach. RM 257 was selected for its use as a well-known diacrylate mcsogcn.(V59, 128) while a c/z-functional and a tetra-functional thiol monomer were selected for use as a flexible spacer and crosslinker, respectively. Non-equimolar solutions were simply mixed in a vial, poured into moulds, and allowed to cure in open air (detailed synthesis procedures and additional experimental results are provided in the Chapter 6). The first stage reaction is used to create a poly domain LCE via the thiol-Michael addition reaction, a click reaction between a thiol group and an electron deficient vinyl group (i.e. an acrylate), which is not limited in its scale. Previous work by Hoyle has demonstrated that nearly 100% conversion of the thiol groups can be attained and controlled over a timescale of approximately a few seconds to one day.(129) Several polydomain samples ranging from a thin film (~200 pm thick) to bulk samples (4 mm thick) are shown in Figure 5.1b to demonstrate the manufacturability of the thiol-acrylate reaction. Ultimately, this reaction will self-limit when the thiol groups have all reacted.
An independent, second-stage polymerization reaction between excess acrylate groups can then be photo-triggered. This second reaction is used to further tailor the properties of the LCE as well as permanently program an aligned monodomain sample via the establishment of new crosslinks. A demonstration of permanent monodomain alignment can also be seen in
61


Figure 5. led, shown by a rectangular sample being stretched, photo-crosslinked, and released. The opaque poly domain sample becomes transparent when stretched, visually indicating the formation of a monodomain. This method has provided an added degree of accessibility for our LCE collaborators, as we have successfully sent polydomain samples to separate laboratories to program a stable monodomain using the second photopolymerization reachon.For this communication, an LCE system with 13% of the thiol functional groups belonging to the crosslinker and a non-equimolar excess of 15% acrylate groups was used to
Temperature (C)
Figure 5.2. Polydomain and monodomain LCE samples were subjected to 0 and 100 kPa bias stresses and cooled from 120 to -20C at a rate of 5C/min. Monodomain samples exhibited 45% actuation under zero stress. The monodomain samples in this experiment were programmed by stretching a polydomain sample to 100% strain and photo-crosslinking for 10 minutes.
62


Figure 5.3. (a) Alternating regions in a polydomain LCE are photo-crosslinked, which become resistant to transparent, monodomain alignment when stretched.
(b) An unaligned LCE is heated to the isotropic state and crosslinked with a photo-mask. Upon cooling, photo-crosslinked areas remain isotropic to revealan image.
highlight the potential of this methodology. The presence of a liquid-crystalline state was verified with polarized optical microscopy, which showed birefringence that disappeared upon heating above 7). Single- and wide-angle x-ray analysis revealed the presence of a nematic structure at room temperature. Dynamic mechanical analysis revealed the glass transition temperature (Tg) increased from 15 to 19C from the first to second stage reaction,
while differential scanning calorimetry (DSC) revealed the nematic to isotropic transition temperature (7y to be 80C after the first stage reaction (Figure A2.3); however, 7) could not
63


be identified using DSC after the second stage reaction. The thermal-actuation behaviour of both polydomain and programmed monodomain samples can be seen in Figure 5.2.
Poly domain samples only exhibit a shape-switching response when under the presence of a bias stress. This stress drives the formation of a monodomain when cooling below 7',. Conversely, programmed monodomain samples show autonomous, hands-free actuation. Both polydomain and monodomain samples experienced increasing amounts of actuation strain with increased bias stress. These results indicate that mechanically useful LCE samples can be produced after both the first and second stages of this reaction. This behaviour cannot be accomplished using current hydrosilylation reactions, as the multi-step process does not consist of two independent reactions. Rather, the hydrosilylation method involves slowing the reaction at a critical point during the gelation process to align the monodomain, which can be difficult to replicate. It should be noted that recent studies have proposed unique methods to create more robust two-step techniques to program monodomain samples by introducing photo-sensitive crosslinking side groups along the main chain) J2) or using exchangeable crosslinks at high temperatures:! /JO) however, these methods do not offer facile control over polymer structure as the proposed TAMAP reaction.
The overall purpose of this work is to introduce a new approach to controlling both LCE structure and liquid-crystalline behaviour. The presented TAMAP reaction provides unique spatio-temporal control over the material to influence both mechanical and optical properties (Figure 5.3a). In this example, the second stage photo-polymerization reaction was used to increase crosslinking at specific alternating regions within a poly domain sample. Upon stretching, these regions have increased crosslinking and resist chain alignment and the formation of a transparent monodomain. Eventually, the process reveals a sample with alternating optical and mechanical properties. Another application of this approach is the control over the formation of liquid-crystalline domains (Figure 5.3b). In this example, the
64


second-stage reaction is used to reveal the formation of an image using opaque polydomains and the transparency of the the state. The addition of photo-crosslinks served to restrict the formation of the poly domain when cooled below Tt. Previous studies have not demonstrated this amount of precision and control over both LCE structure and liquid-crystalline behaviour. Furthermore, these results suggest that the second-stage photopolymerization reaction can be used to control the phase transitions between the poly domain, mondomain, and isotropic phases in response to a stimulus in specific locations.
LCEs are a class of active polymers that are capable of mechanical actuation in response to a stimulus, commonly heat or light.{131, 132) Unfortunately, LCEs have not experienced the same level of widespread research attention similar to other classes of actively moving polymers, such as shape-memory polymers (SMPs), though both systems are generally known for their ability to mechanically respond to a change in temperature. In addition, both systems require proper programming of the polymer to an aligned state before shape change can occur. The key difference is that SMPs exhibit a one-time shape-recovery event when heated above a thermal transition (Tg or and are driven by entropy elasticity,(133) while LCEs repeatedly undergo a shape-switching phenomenon driven by a reversible anisotropic-isotropic transition associated with liquid-crystalline order.(134) As a result, LCEs have an added degree of functionality capable of creating devices that repeatedly actuate over the lifetime of the device, such as in artificial muscles:!/#, 68) nevertheless, SMPs have received a higher profile of interest for proposed applications, especially biomcdically-rclatcd.i/J5-/J 7) It is of interest to note that recently researchers have proposed incorporating the shape-memory effect within LCE systems to take advantage of both mechanisms.(138)
65


The difficulty of synthesis and programing of monodomain LCE samples has been a longtime challenge.(77) The proposed TAMAP methodology may overcome traditional barriers to access these exquisite materials. Furthermore, it may provide an easily accessible platform to manufacture and tailor LCE-based biomedical devices. The composition presented in this study demonstrated non-cytotoxic responses after both first and second stage reactions (Figure 5.4). The proposed TAMAP approach provides the ability to explore potential biomedical applications of LCE materials with enhanced functionality and control. For
Figure 5.4. Cytocompatibility of the TAMAP synthesized LCE was confirmed after both the first and second stages of the reaction using both elution and direct-contact test by an independent laboratory (WnXi AppTec, St. Paul, MN. USA). Cellular response to both (a) direct contact and (b) elution tests are shown.
example, these data suggest the second stage reaction may be utilized to tailor the LCE
properties in vivo due to the non-cytotoxic response at both stages of the reaction. While
there have been a handful of toxicity studies performed on liquid crystal based materials and
sensors,{139, 140) biocompatibility data for LCEs remain largely unreported. Future studies
are needed to fully evaluate the biocompatible nature of these materials.
5.2. Conclusions
This study presented an unexplored two-stage TAMAP reaction. Mechanically robust polydomain samples were synthesized using the self-limiting Michael-addition reaction,
66


and demonstrated strain actuation under a bias stress in response to temperature. The second stage photopolymerization reaction was used to permanently program a monodomain within the samples, which demonstrated hands-free actuation without the need for a bias stress. Furthermore, this second reaction was used to tailor the mechanical properties and liquid-crystalline behaviour with spatio-temporal control. The composition investigated within this study elicited a cytocompatible response at each stage of the TAMAP reaction.
5.3. Acknowledgments
The National Science Foundation CAREER Award 1350436 supported this work. The authors would like to thank Ellana L. Taylor and Brandon Mang for their help in experimental testing as well as Amir Torbati, Jaimee Robertson, and Patrick T. Mather for their help in SAXS and WAXS characterization.
67


CHAPTER VI
SYNTHESIS OF PROGRAMMABLE MAIN-CHAIN LIQUID-CRYSTALLINE ELASTOMERS USING A TWO-STAGE THIOL-ACRYLATE REACTION
6.1. Abstract
This study presents a novel two-stage thiol-acrylate Michael addition-photopolymerization (TAMAP) reaction to prepare main-chain liquid- crystalline elastomers (LCEs) with facile control over network structure and programming of an aligned monodomain. Tailored LCE networks were synthesized using routine mixing of commercially available starting materials and pouring monomer solutions into molds to cure. An initial polydomain LCE network is formed via a self-limiting thiol-acrylate Michael-addition reaction. Strain-to-failure and glass transition behavior were investigated as a function of crosslinking monomer, pentaerythritol tetrakis(3-mercaptopropionate) (PETMP). An example non-stoichiometric system of 15 mol% PETMP thiol groups and an excess of 15 mol% acrylate groups was used to demonstrate the robust nature of the material. The LCE formed an aligned and transparent monodomain when stretched, with a maximum failure strain over 600%. Stretched LCE samples were able to demonstrate both stress-driven thermal actuation when held under a constant bias stress or the shape-memory effect when stretched and unloaded. A permanently programmed monodomain was achieved via a second-stage photopolymerization reaction of the excess acrylate groups when the sample was in the stretched state. LCE samples were photo-cured and programmed at 100%, 200%, 300%, and 400% strain, with all samples demonstrating over 90% shape fixity when unloaded. The magnitude of total stress-free actuation increased from 35% to 115% with increased programming strain. Overall, the two-stage TAMAP methodology is presented as a powerful tool to prepare main-chain LCE systems and explore structure-property-performance relationships in these fascinating stimuli-sensitive materials.
68


6.2. Introduction
LCEs are a class of stimuli-responsive polymers that are capable of exhibiting mechanical and optical functionalities due the combination of liquid-crystalline (LC) order and rubber elasticity. These materials can demonstrate extraordinary changes in shape, soft-elasticity behavior, and tunable optical properties in response to a stimulus such as heat or light,(6, 33, 141) which makes them suitable for many for potential technological applications such as artificial muscles,(27, 33, 142) sensors, and actuators.(17, 33)LCEs have already been demonstrated in many applications such as micro- grippers for robotics,(102) micro-electromechanical systems (MEMS).(/43) optical grating devices,(104, 143) tunable apertures,(105) and microfluidic systems.(106)
The structural components that give rise to the ordered LC phases are called mesogens. Mesogens are the basis of the LC domains and are typically composed of two or three linearly connected aromatic rings with flexible ends. These moieties can be directly placed within the polymer backbone to create main-chain LCEs or as a side group (i.e. side-on or end-on LCEs).(6, 80) Main-chain LCEs have generated a lot of interest due to their direct coupling between mesogenic order and polymer backbone conformations.(30, 31, 36, 46) This direct coupling allows main-chain LCEs to exhibit higher degrees of mesogen orientation, mechanical anisotropy, and strain actuation. (31)
Thermal actuation of LCEs relies on a reversible anisotropic-isotropic transition associated with LC order.(33) To program an LCE for actuation, the mesogens must first be oriented along a director to form a monodomain (i.e. anisotropic mesophase) and is often referred to as a liquid single-crystal elastomer. Actuation occurs as an aligned LCE is heated above an isotropic clearing temperature (Tj), which disrupts the order of the mesogens into an isotropic state and drives shape change. A monodomain can be formed temporarily by applying an external stress (i.e. hanging a weight) to a sample, which will align the polymer chains and orient the mesogens in the direction of the stress. Permanent programming of the monodomain can be achieved via a multi-step
69


process, which involves producing a lightly cross-linked gel followed by immediate application of mechanical stress to induce orientation of the mesogens. Once aligned, the reaction is continued to established covalent crosslinks and stabilize the monodomain.(39) Other "one pot" alignment techniques can be performed in the presence of electric fields or by surface alignment (i.e. rubbing polyimide on a glass slide) during polymerization; however, these methods are generally limited to thin film samples.(6, 30)
Finkelmann and Bergmann introduced the first synthetic route for the preparation main-chain LCEs using one-step platinum-catalyzed hydrosilylation reaction of a divinyl mesogen and a tetra-functional siloxane crosslinker.(46)This method has been widely adapted by many research groups to synthesize main-chain LCEs.(31, 138, 144) Polyesterification and epoxy-based reactions have also been used to make main-chain LCEs. (32) All of these methods require high purity starting materials and careful experimental conditions to prevent side reactions.(6) Furthermore, these methods rely on random cross-linking of the monomers, resulting in poorly defined network structure. Therefore, it is more difficult to correlate the structure to the properties of LCEs. Recent studies have used click chemistry as a tool to prepare more uniform LCE networks; however, these reactions require custom-synthesized starting mesogenic and thiol monomers, which can be challenging to produce, and have been limited to prepare micron-sized actuators rather than bulk samples.(145-147).
Current challenges in the LCEs focus on how to develop synthetic methods that are facile, reproducible, and scalable to design tailored LCE networks with programmable monodomains. Recently, our group introduced a two-stage thiol-acrylate Michael addition-photopolymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic main-chain LCEs. (55) Two-stage TAMAP reactions form dual- cure polymer networks, where the staging of the polymerization process allows modification of the polymer structure at two distinct time
70


points. This strategy has been adapted in the past few years to design and fabricate other advanced materials, other than mesomorphic systems, such as micro-actuators,(148) shape-memory polymers,(149, 150) and surface wrinkles.(151, 152) The TAMAP methodology utilizes a non-stoichiometric composition with an excess of acrylate functional groups. The first stage reaction is used to create a polydomain LCEs via the thiol Michael- addition reaction, which is self-limited by the thiol groups. This is an intermediate LCE network that would be capable of mesogenic domain orientation by applying mechanical stress. The polydomain resulting from the first-stage Michael-addition reaction is indefinitely stable and the alignment of the monodomain does not need to occur immediately after the reaction has completed. The second-stage photopolymerization reaction between excess acrylate groups is used to permanently fix an aligned monodomain and program the LCE for reversible and stress- free (i.e. "hands free") actuation. The purpose of this study is to explore and demonstrate the robust nature of the TAMAP reaction to prepare main-chain LCEs by investigating the influence of crosslinking density and programed strain on the thermomechanics of the LCE systems. We demonstrate a wide range of thermomechanical properties and actuation performance that are achievable using this reaction.
6.3. Protocol
6.3.1. Preparation of Liquid Crystalline Elastomers LCEs
1. Add 4 g of 4-bis-[4-(3-acryloyloxypropypropyloxy) benzoyloxy]-2-methylbenzene (RM257) into a 30 ml vial. RM257 is a di-acrylate mesogen and is received as a powder. Dissolve RM257 by adding 40 wt% (i.e. 1.6 g) of toluene and heat to 80 C on a hot plate. This process typically takes less than 5 min to dissolve the RM257 into a solution. Note: Other solvents can be used to dissolve the RM257, such as dichloromethane (DCM), chloroform, and dimethylformamide; however, toluene was chosen because it allows the monomers to cure at room temperature
71


without having the solvent evaporate quickly during reaction, while DCM and chloroform could evaporate quickly at room temperature before the Michael-addition reaction is completed. Dimethylformamide can dissolve RM257 immediately without heating, but requires very high temperatures to remove the solvent (~ 150 C). Kamal and Park used a combination of DCM and a liquid crystal, CB5, to dissolve RM257.(753)
2. Cool the solution to room temperature. Add 0.217 g of pentaerythritol tetrakis(3-mercaptopropionate) (PETMP), a tetra-functional thiol crosslinking monomer, and 0.9157 g of 2,2-(ethylenedioxy) diethanethiol (EDDET), a di-thiol monomer. The molar ratio of thiol functional groups between PETMP and EDDET is 15:85. This ratio will be referred to as 15 mol% PETMP throughout the study. Note: If the RM257 recrystallizes during this process, temporarily place the vial back onto the 80 C hot plate until the monomer returns to solution. Cool the solution to room temperature before proceeding to the next steps.
3. Dissolve 0.0257 g of (2-hydroxyethoxy)-2-methylpropiophenone (HHMP) into the solution. HHMP is a photoinitiator used to enable the second-stage photopolymerization reaction. This step can be skipped if the second-stage reaction will not be utilized.
4. Prepare a separate solution of a catalyst by diluting dipropylamine (DPA) with toluene at a ratio of 1:50. Add 0.568 g of diluted catalyst solution to the monomer solution and mix vigorously on a Vortex mixer. This corresponds to 1 mol% of catalyst with respect to the thiol functional groups. Note: Adding undiluted catalyst, such as DPA, to the solution will likely result in extremely rapid localized polymerization and will prevent manipulation of the polymer solution into the desired mold detailed in the next steps.
5. Place the monomer solution in a vacuum chamber for 1 min at 508 mmHg to remove any air bubbles caused by mixing. Perform this step immediately after mixing.
72


6. Immediately transfer the solution into the desired mold or inject the solution between two glass slides. Molds should be manufactured from HDPE. The molds do not need to be covered, as the Michael-addition reaction is relatively insensitive to oxygen inhibition.
7. Allow the reaction to proceed for at least 12 hr at room temperature. The solution will begin to gel within the first 30 min.
8. Place samples in a vacuum chamber at 80 C and 508 mmHg for 24 hr to evaporate the toluene. Once completed, the samples should have a glossy white and opaque appearance at room temperature.
9. Repeat the procedure to tailor the ratio of tetra-functional to di-functional thiol monomers in step 1.2 with ratios of 25:75 50:50, and 100:0, respectively. A detailed table of the chemical formulations used for this study is shown in Table 6.1.
6.3.2. Kinetics Study of Two-stage Reaction with Real-time Fourier Transform Infrared
1. Equip a spectrometer with a MCT/B detector and XT-KBr beam splitter.
2. Prepare a mixture using the protocol outlined above in the Preparation of LCE section using 0.5 mol% of catalyst with respect to thiol functional groups and 0.5 wt% of photoinitiator. Two initiators were tested separately, 2-2-dimethoxy-2-phenylacetophenone (DMPA),and HHMP. DMPA is a more commonly used initiator, while HHMP is more stable at elevated temperatures.
3. Place one drop of LCE mixture between NaCl crystals immediately after mixing using a glass pipette.
4. Record spectra at a 2.92 sec sampling interval rate.
5. Monitor the conversion of the thiol groups using a peak height profde with the S-H absorption
73


peak at 2,571 cm'1 with a baseline of 2,614 2.527cm'1.
6. Monitor the conversion of the acrylate groups using a peak height profile with the C=C absorption peak at 810 cm"^ with a baseline of 829 781 cm'l
7. Allow the reaction to proceed under FTIR at room temperature until the thiol peak height plateaus, showing 100% conversion of thiol groups.
8. Upon complete conversion of thiol group, turn on a 365-nm light source equipped with a light
9
guide for 10 min to complete the polymerization of excess acrylates at 350 mW/cm intensity, which can be measured by a radiometer photometer. 9. Monitor the conversion of the acrylate groups as described in 2.6.
6.3.3. Dynamic Mechanical Analysis (DMA)
1. Prepare two glass slides by spraying the surfaces of the slides with a hydrophobic surface agent and rubbing the surfaces with a paper towel until dry.
2. Stack slides together such that they are separated with a 1 mm spacer. Spacers can be cut by scoring and breaking a separate glass slide to measure approximately 25.4 mm x 5 mm x 1 mm. Clamp slides together using a binder clip at each end.
3. Inject monomer solution between the slides using a glass pipette. This requires approximately 1.5 g of the prepared monomer solution.
4. Allow the sample to cure for at least 12 hr according to step 1.7. Separate the glass slides and dry the sample according to Step 1.8.
5. Using a razor blade or scissors, cut a rectangular test specimens with dimensions of 30 x 10 x 1
3
mm .
74


6. Load the sample properly into a DMA machine. Test the sample in tensile mode, with active length measuring 10 to 15 mm. Take care not to over-tighten the grips on the test sample, as 0.1 N-m is often too much torque when tightening the grips.
7. Cycle the sample at 0.2% strain at 1 Hz from -50 to 120 C at a heating rate of 3 C/min. Set the force track to 125%.
8. Measure the glass transition temperature (Tg) at the peak of the tan d curve.
9. Measure the isotropic transition temperature (Tj) and the lowest point of the storage modulus curve.
10. Measure the rubbery modulus, E'r, at 27+30 C.
6.3.4. Strain-to-failure Tests
1. Prepare an HDPE mold by milling ASTM Type V dog-bone cavities at a depth of 1 mm.
2. Using a glass pipette, fill each dog bone cavity until the monomer solution is flush with the top of the mold. Allow the samples to cure and dry according to Steps 1.7 and 1.8.
3. Prepare 5 tensile specimens from LCE samples formulated with varying PETMP crosslinker concentrations of 15, 25, 50, and 100 mol%.
4. Set two pieces of reflective laser tape 5 to 7 mm apart within the gage length of the specimen.
5. Load the specimen into a mechanical tester equipped with a laser extensometer, thermal chamber, and 500 N load cell. Use self-tightening grips to secure the specimens, as specimens will dislodge from wedge grips at high strain values. Align the laser extensometer properly to track the accurate change in length as a function of applied strain.
75


6. Strain the specimens at room temperature with a displacement rate of 0.2 mm/sec until failure. Define failure by the fracture of the specimen.
7. Test additional specimens of 15 mol% PETMP crosslinking agent for strain-to-failure testing as a function of temperature. Test specimens at -40, -30, -20, -10, 0, 10, 22, 40, 60, and 80 C. Hold all specimens isothermally at the desired test temperature for 10 min prior to testing.
6.3.5. Shape Fixity and Actuation Tests
1. Prepare an HDPE custom dog-bone mold with gage length of 25 mm and cross-sectional area of 1 mm x 5 mm.
2. Prepare a 15 mol% PETMP monomer solution according to Steps 1.1 to 1.5.
3. Using a glass pipette, fill each mold cavity until the monomer solution is flush with the top of the mold.
4. Allow the samples to cure and dry according to Steps 1.7 and 1.8.
5. Set two pieces of reflective laser tape 5-7 mm apart within the gage length of the specimen. Load the sample according step 4.5. Using a permanent marker, mark a dot in the other side of each piece of reflective tape. Record the length between the dots.
6. Strain the specimens at room temperature with a displacement rate of 0.2 mm/sec to 100%, 200%, 300%, or 400% strain.
7. While maintaining the desired strain level, expose the sample to a 365 nm UV light source at an
9
intensity of ~10 mW/cm for 10 min by holding a UV Lamp approximately 150 mm from the sample.
8. Unload the sample and then heat it above 7/ to induce actuation. Allow the sample to cool back
76


to room temperature and record the length between the dots.
9. Calculate fixity using the following equation: where Sapplied is programming strain before photo-crosslinking (measured by the laser extensometer) and sfixed is the amount of permanent strain after photo-crosslinking (measured by the change in dot displacement).
10. Cut a 30 mm length sample from the center portion of the programmed specimen.
11. Load the sample properly into a DMA tester. Test the sample in tensile mode, with active length measuring 13 to 15 mm. Make sure not to over-tighten the grips on the test coupon.
12. Equilibrate the sample at 120 C under a preload of 0 N. Cool the sample from 120 to -25 C at a rate of 3 C/min. Maintain the pre-force at 0 N for the entire test.
Fixity (%) = x 100 (Eq. 1)
£applied
6.4. Representative Results
In this study, the two-stage TAMAP reaction cure kinetics were investigated using real-time FTIR. An FTIR series study on the conversion of the thiol and acrylate groups as a function of time to capture both the stages of the reaction was implemented and the normalized results are shown in Figure 2a. The first-stage thiol-acrylate Michael-addition reaction was initiated via base catalysis using DPA as the catalyst and results in the formation of a crosslinked polymer network. At the end of this initial reaction, the thiol functional groups achieve close to 100% conversion within 5 hours of under ambient conditions (~22 C), while the acrylate groups attained between 70% to 78% conversion under the same conditions. The thiol-acrylate Michael addition 'click' reaction is self-limiting in nature and can generate a step-growth, crosslinked, stable network in a facile manner based on the relative ratios of functional groups present. Subsequently, the second-
77


stage photopolymerization reaction was initiated via exposure to UV irradiation and the remaining unreacted acrylate groups present within the network were further crosslinked to achieve a final acrylate functional group conversion near 100%. Two photoinitiators, HHMP and DMPA and their reaction kinetics were studied within the polymer networks and both were seen to efficiently create crosslinked acrylate networks at the end of second stage polymerization. The conversion of the acrylate groups as a function of the intensity of exposure was also studied and seen to correlate. Overall, it was observed that though a number of variables such as photoinitiators and exposure times could be varied, it was possible to efficiently attain high final conversion of the acrylates at the end of the second stage within 10 minutes even with relatively low levels of UV
2 2
intensity (-10-25 mW/cm compared to 350 mW/cm ). Figure 6.2b shows the FTIR absorbance spectra of the two-stage reaction at 3 different time points, 0, 300, and 320 minutes. At time 0, the initial spectra captures the presence of both thiol and acrylate functional groups in their unreacted state. At the 300 minute time point, by the end of the first-stage thiol-Michael addition reaction, the thiol and acrylate peak heights are seen to reduce considerably, thereby implying the reaction between the thiol and acrylate functional groups has progressed to completion. The thiol peak is measured to be close to 100% conversion at this point, whereas the acrylates are seen to be consumed up to 78%. The complete disappearance of the thiol peak is not observed, most likely as the presence of the thiol-Michael adduct from the first-stage reaction is seen to appear and
overlap with the thiol peak at 2,571 cm'l At the end of the second-stage photopolymerization reaction initiated via UV exposure, at the 320 minutes point, the acrylate conversion is seen to proceed to completion, implying 100% conversion of remaining acrylic double bonds within the network.(75-/)
The two-stage TAMAP methodology provides facile control to explore structure-property relationships in LCEs. The influence of crosslinking density on stress-strain behavior is shown in Figure 6.3a. Modulus and fracture stress were shown to increase with increasing PETMP content,
78


while failure strain increased with decreasing PETMP content (Figure 6.3b). LCE samples with 50 and 100 mol% PETMP demonstrated initial elastic loading followed by a stress plateau and sharp increase in stress due to chain alignment. In comparison, samples with 15 and 25 mol% PETMP appeared to demonstrate more traditional elastomeric loading followed by an increase of stress due to chain alignment. All specimens tested showed a transition from white opacity to clear transparency when stretched (Figure 6.3e). It should be noted that all specimens maintained a large degree of permanent strain after fracture and did not recover to their original shape at room temperature; however, all specimens visually recovered to their original shape upon heating above Tj. The influence of temperature on failure strain was then investigated for the 15 mol% PETMP composition (Figure 6.3c). In the glassy state, LCE specimens exhibited brittle failure with no appreciable deformation. At the onset of the glass transition, the failure strain increased significantly and followed the general shape of the tan 5 function measure by DMA. The failure strain reached a maximum of 650% strain at 10 C. Representative glass transition behavior for the four LCE network systems is shown in Figure 6.3d. All of the LCE networks displayed non-traditional behavior in both the storage modulus and tan 5 curves. The storage modulus of all LCE networks displayed a distinct minimum that was roughly associated with Tj. The tan 5 functions were represented by an initial peak followed by an elevated region that diminished as the sample was heated into the isotropic state (a representative curve can be seen in Figure 6.3c). For the four LCE systems tested, both Tg and rubbery modulus increased with increasing crosslinking density. A summary of thermo-mechanical properties of the four LCE systems can be seen in Table 6.2.
LCEs offer the ability to demonstrate both the shape-memory effect and reversible actuation (Figure 6.4). An unaligned poly domain specimen of 15 mol% PETMP was used to illustrate the different shape-switching pathways that can be programmed into the material (Figure 6.4a).
79


Reversible stress-driven actuation is demonstrated by the pathway in Figure 6.4a-b-c. The polydomain specimen is stretched by hanging a 60.6 mN weight to apply a constant stress. This bias stress mechanically orients the mesogens into a transparent monodomain. The specimen contracts when heated to the isotropic state and elongates when cooled below Tj. This process can be repeated indefinitely. The shape-memory effect was exhibited when the bias stress was removed from the specimen when cooled below Tj to 22 C, which is still 18 C above Tg. While some elastic recoil was observed, a majority of the strain remained programmed into the material. It should be noted that the mesogens remained in a stable monodomain orientation, and there is a noticeable difference in optical properties within the free end of the sample where the clamp was attached (i.e. the gripped portion remained glossy white). Heating the sample above Tj activated full shape recovery, indicating the shape- memory cycle follows the pathway of Figure 6.4abde. The second-stage photopolymerization reaction can be used to achieve stress-free actuation without the need for a constant bias-stress or programming step between cycles. The temporarily
80


(a)
(b)
DPA
HHMP
Toluene
Michael
Addition
(Stage 1)
Alignment Monodomain
Figure 6.1. Schematic of Monodomain Programing via a Two-Stage Thiol-Acrylate Reaction, (a) A diacrylate mesogen (l,4-bis-[4-(3- acryloyloxypropyloxy)benzoyloxy]-2-methylbenzene RM 257), dithiol flexible spacer (2,20-(ethylenedioxy) diethanethiol EDDET), and tetra-functional thiol crosslinker (pentaerythritol tetrakis(3-mercaptopropionate) PETMP) were selected as commercially available monomers. Non- equimolar monomer solutions were prepared with an excess of 15 mol% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2-hydroxyethoxy)-2-methylpropiophenone (HHMP) were added as the respective catalyst and photo- initiator to the solutions, (b) Representative polydomain structure forms via Michael addition (first stage) with a uniform cross-link density and latent excess acrylate functional groups, (c) A mechanical stress is applied to the polydomain samples to orient the mesogens into a temporary monodomain, (d) A photopolymerization reaction (second stage) is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample.
9
aligned specimen was photo- cured using 365 nm light at ~10 mW/cm for 10 minutes (Figure 6.4f). The sample experienced minimal elastic recoil when unloaded due to the establishment of covalent crosslinks between the excess of unreacted acrylate groups (Figure 6.4g). Stress-free actuation was then activated by controlling the temperature about Tj using the reversible pathway in Figure 6.4g-h; however, it should be noted that the sample does not experience full recovery back to the initial shape of the specimen. The influence of applied programming strain (i.e. strain during photopolymerization) as function of fixity and actuation for the 15 mol% PETMP system is shown in Figure 6.5a. All specimens demonstrated fixity values higher than 90%. The amount of programming strain did not noticeably influence the fixity values for the strain range tested in this study. Conversely, actuation strain increased linearly with the amount of programming strain. On average, the actuation strain corresponded to approximately 30% of the programming strain value. Representative curves showing actuation as a function of temperature can be seen in Figure 6.5b. It should be noted that the actuation strain values in Figure 6.5a correspond to
81


Full Text

PAGE 1

STRUCTURE PROPERTY RELATIONSHIPS IN THIOL ACRYLATE BASED MAIN CHAIN LIQUID CRYSTALLINE ELASTOMERS by MOHAND OSMAN SAED B.S., University of Gezira, 2008 M.S., University of Colorado Denver, 2014 A thesis submitted to Faculty of the Graduate School of the University of Colorado in partial fulfillments of the requirements of the degree of Doctor of Philosophy Mechanical Engineering Program 2017

PAGE 2

! "" 2017 MOHAND OSMAN SAED ALL RIGHTS RESERVED

PAGE 3

! """ The Thesis For Doctor Of Philosophy Degree b y Mohand Osman Saed h as been approved for the Mechanical Engineering Program b y Dana R. Carpenter, Chair Christopher M. Yakacki, Advisor Ronald Rorrer Kai Yu Carl P. Frick Christopher N. Bowman Date: May 1 3 201 7

PAGE 4

iv Saed, Mohand Osman (Ph.D., Mechanical Engineering Program) Structure Property Relationships in Thiol Acrylate Based Main Chain Liquid Crystalline Elastomers Thesis directed by Professor Christopher M. Yakacki ABSTRACT In t his research, we used a profoundly new approach to synthesize l iquid crystalline e lastomers ( LCEs ) based on using a thiol acrylate "click" reaction and two stage thiol acrylate Micha el addition photopolymer ization (TAMAP) reaction, both of which have not previously been investigated for LCE synthesis. The thiol acrylate reaction was used initially to synthesize polydomain LCEs and then to examine the influence of crosslinking and spacer length. Fi r st, the influence of crosslinking on the thermomechanical behavior of LCEs was investigated. The isotropic rubbery modulus, glass transition temperature, and strain to failure showed strong dependence on cross linker amount and ranged from 0.9 MPa, 3¡C, and 105% to 3.2 MPa, 25¡C, and 853%, respectively. The isotropic transition temperature (T i ) was shown to be influenced by the functionality of the crosslinker, while the crosslinker concentration had no effect The magnitude of actuation can be tailored by c ontrolling the amount of crosslinker and applied stress. Ac tuation increased with increasing the applied stress and decreased with greater amounts of cross l inking. Second we hypothesized that tuning the LC phases in main chain LCE systems can be achieved by varying the spacer length while maintaining the same mesogen (RM257). By increasing the length of spacers from two to eleven carbons along the spacer backbone (C2 to C11), we can modulate the mesophase from nematic to smectic, tailor the nematic to isot ropic transition temperature between 90 and 140 ¡ C, and increase the average work capacity from 128 to 262 kJ/m 3 Phase segregation and the smectic C phase is achieved at room temperature for the C6, C9, and C11 spacers. Upon heating, these samples transiti on into the nematic and later, the isotropic phase. Furthermore, this segregation occurs along with polymer chain crystallinity, which increasing the modulus of the networks by an order of magnitude; however, the crystallization rate is highly time depende nt on the spacer length and can vary between 5 minutes for

PAGE 5

v the C11 spacer and 24 hours for shorter spacers. A novel TAMAP methodology was implemented to synthesize monodomain LCEs using commercially available starting monomers. A wide range of thermomechan ical properties was tailored by adjusting the amount of crosslinker, while the actuation performance was dependent on the amount of applied strain during programming. The form and content of this abstract are approved. I recommend its publication. Approved: Christopher M. Yakacki

PAGE 6

vi To my family.

PAGE 7

vii ACKNOWLEDGEMENTS I would like to thank my advisor, Prof. Christopher M. Yakacki for his guidance, encouragement, and continuous support over the past 5 years. You r passion for research, good work ethic, limitless ideas, and creativity has inspired me greatly. Working in his laboratory has thought me so many skills that will b enefit me throughout my career. I also would like to thank and acknowledge my thesis committee members, Prof. Dana Carpenter, Prof. Ron Rorrer, Prof. Ka i Yu, Prof. Carl Frick and Prof. Christopher Bowman, who agreed to serve on my PhD committee despite their tense schedules and for the ir valuable feedback, which has helped me to further understand my research. Special gratitude extents to Prof. Frick, and Prof. Bowman for their willing to come from Laramie, WY and Boulder, CO. The interdisciplinary nature of my projects has taught me to collaborate extensively with many groups around the country. Explicitly, I would like to thank our collaborators at University Wyoming (Prof. Carl Frick and Dan Markel), University of Colorado Boulder (Rayshan Visvanathan, Prof. Noel Clark, Matt McBrid e Abeer Alzahrani and Prof. Chris Bowman) and John Hopkins University (Aurelie Azoug and Vicky Nguyen). I would like to thank the Smart Materials and Biomechanics Lab (SMAB) members for their support and encouragement. I would like to thank Dr. Amir Tor bati, Ravi Patel, Ross Volpe, Nick Traugutt, Sam Mills, Michael Bollinger, Lillian Chatham, and Ryan Anderson for useful discussions and the wonderful time I spent working with them. I would also like to thank my undergraduate students Brandon Mang, Ellan a Taylor, Chelsea S tarr, and Kristen B o nifield. Finally, I would like to thank my family for their love, support, and encouragement throughout this work. My wife, Omnia has been a constant source of love and joy. I could not have accomplished my PhD with out her support

PAGE 8

viii TABLE OF CONTENTS TABLE OF CONTENTS ................................ ................................ ................................ .................... xii i LIST OF TABLE ................................ ................................ ................................ ................................ ... xi LIST OF FIGURE ................................ ................................ ................................ ................................ xii CHAPTER ................................ ................................ ................................ ........................... ..... .. I. INTRODUCTION AND BACKGROUND ................................ ................................ .................... 1 1.1 Liquid Crystals (LC) ................................ ................................ ................................ ................. 1 1.2 Liquid crystalline Elastomers (LCEs) ................................ ................................ ...................... 2 1.2.1 Classification ................................ ................................ ................................ ..................... 3 1.2.2 Preparations ................................ ................................ ................................ ....................... 6 1.2.3 Crosslinking History ................................ ................................ ................................ .......... 7 1.2.4 Stress Strain Behavior ................................ ................................ ................................ ....... 8 1.2.5 Actuation ................................ ................................ ................................ ............................ 9 II. RESEARCH MOTIVATION AND GOALS ................................ ................................ ............... 11 III. THIOL ACRYLATE MAIN CHAIN LIQUID CRYSTALLINE ELASTOMERS WITH TUNABLE THERMOMECHANICAL PROPERTIES AND ACTUATION STRAIN ..................... 15 3.1. Abstract ................................ ................................ ................................ ................................ ...... 15 3.2 Introduction ................................ ................................ ................................ ................................ 16 3.3. Experimental ................................ ................................ ................................ .............................. 19 3.3.1. Materials ................................ ................................ ................................ ............................. 19 3.3.2. Synthesis of Liquid Crystalline Elastomers ................................ ................................ ....... 20 3.3.3. Gel Fraction Tests ................................ ................................ ................................ ............... 21 3.3.4. X Ray Scattering ................................ ................................ ................................ ................ 21 3.3.5. Differential Scanning Calorimetry (DSC) ................................ ................................ .......... 22

PAGE 9

ix 3.3.6. Dynamic Mechanical Analysis (DMA) ................................ ................................ .............. 23 3.3.7. Strain to Failure Tests ................................ ................................ ................................ ........ 23 3.3.8. Strain Actuation Charact erization ................................ ................................ ...................... 23 3.4. Results ................................ ................................ ................................ ................................ ....... 24 3.5. Discussion ................................ ................................ ................................ ................................ .. 32 3.6. Conclusions ................................ ................................ ................................ ............................... 38 3.6. Acknowledgements ................................ ................................ ................................ ................... 38 IV. MODULATED MESOPHASE LIQUID CRYSTAL ELASTOMERS ................................ ..... 39 4.1. Abstract ................................ ................................ ................................ ................................ ...... 39 4.2. Introduction ................................ ................................ ................................ ............................... 39 4.3. Results and Discussion ................................ ................................ ................................ .............. 42 4.4. Conclusions ................................ ................................ ................................ ............................... 54 4.5. Experimental Section ................................ ................................ ................................ ................. 55 4.6. Acknowledgements ................................ ................................ ................................ ................... 58 V. TAILORABLE AND PROGRAMMABLE LIQUID CRYSTALLINE ELASTOMERS USING A TWO STAGE THIOL ACRYLATE REACTION ................................ ............................. 59 5.1. Main ................................ ................................ ................................ ................................ ........... 59 5.2. Conclusions ................................ ................................ ................................ ............................... 66 5.3. Acknowledgments ................................ ................................ ................................ ..................... 67 VI. SYNTHESIS OF PROGRAMMABLE MAIN CHAIN LIQUID CRYSTALLINE ELASTOMERS USING A TWO STAGE THIOL ACRYLATE REACTION ................................ .. 68 6.1. Abstract ................................ ................................ ................................ ................................ ...... 68 6.2. Introduction ................................ ................................ ................................ ............................... 69 6.3. Protocol ................................ ................................ ................................ ................................ ...... 71 6.3.1. Preparation of Liquid Crystalline Elastomers LCEs ................................ .......................... 71

PAGE 10

x 6.3.2. Kinetics Study of Two stage Reaction with Real time Fourier Transform Infrared .......... 73 6.3.3. Dynamic Mechanical Analysis (DMA) ................................ ................................ .............. 74 6.3.4. Strain to failure Tests ................................ ................................ ................................ ......... 75 6.3.5. Shape Fixity and Actuation Tests ................................ ................................ ....................... 76 6.5. Discussion ................................ ................................ ................................ ................................ .. 86 6.6. Acknowledgments ................................ ................................ ................................ ..................... 89 6.7. Materials ................................ ................................ ................................ ................................ 90 VII. CONCLUSIONS AND FUTURE WORK ................................ ................................ ................ 91 7.1 Conclusions ................................ ................................ ................................ ................................ 91 7.2. Recommendations for the Further Work ................................ ................................ ................... 93 BIBLIOGRAPHY ................................ ................................ ................................ ................................ 95 APPRNDIX ................................ ................................ ................................ ................................ ........ 103 A Wide Angle X Ray Scattering Characterizations ................................ ................................ .. 103 B Differential Scanning Calorimetry (DSC) ................................ ................................ .............. 108 C Dynamic Mechanical Analysis (DMA) ................................ ................................ .................. 111

PAGE 11

xi LIST OF TABL E Table 3 1 Summary of thermal analysis and thermomechanical properties of LCE. ................. xii Table 4 1 Summary of DSC and WAXS data for 5 LCE systems tested. Each data point represents n=3.All of the samples contained equal amount of crosslinker. T c was measured during the 1 st heati ng scan, where as T SmC T NI and H f were measured during the 2 nd heating scan. The d spacing valves were calculated from the 1 D plots see the supporting information for more details. ................................ ................................ ................................ ........ 47 Table 4 2. Dynamic Mechanical Analysis (DMA) behavior for the first and second temperature sweep; the glass transition temperature (T g ) was measured at the peak of tan !; (E n ) is the storage modulus measured at 25 ¡C; where the rub bery modulus (E r ) was measu red the isotropic temperature T NI + 30 ¡C. The first temperature sweep was performed after being stored for at least 24 hours at room temperature; whereas the second temperature sweep was performed 5 minutes after the first sweep was completed ................................ ......... 50 Table 6 1 Chemical Formulations for LCE Systems: Four different LCE systems used in this study. The naming convention is based on the molar ratio of thiol functional groups between PETMP and EDDET. All sy stems have an excess of 15 mol% acrylate functional groups. It should be noted, FTIR studies tested HHMP as well as DMPA as photoinitiators and reduced the amount DPA catalyst by half to help with the kinetic characterization. *DPA is diluted in toluene at a ratio of 1:50. ................................ ................................ ............... 85 Table 6 2. Summary of Thermomechanical Properties of LCE Systems: Dynamic Mechanical Analysis (DMA) test shows the thermomechanical propertie s of the initial polydomain LCE networks formed via the first stage Michael addition reaction. Both T i and E' r were measured at the lowest point of the storage modulus vs. temperature curve. ......... 86 Table 6 3. Materials used in the study. ................................ ................................ ........................ 90

PAGE 12

xii LIST OF FIGURE Figure 1.1 Polymeric materials displaying liquid crystallinity, ( a ) Liquid crystal polymer (LCP), ( b ) liquid crystal polymer networks (LCNs) are heavily crosslinked materials, ( c ) Liquid crystal elastomers (LCEs) are lightly crosslinked materials. ................................ ................................ 1 Figure 1.2 Different attachment geometries for the synthesis of LCEs: side chain elastomers with end on (a) or side on (b) attached mesogenic side chains and main chain elastomers with mesogenic units incorporated end on (c) or side on (d) into the polymer main chain. .......................... 3 Figure 1.3. Three important LC phases. In the nematic phase, the mesogens posses a short range order and a re aligned parallel in a uniform direction, defined by the director. Smectic A phases exhibit a layered structure with the mesogens parallel to the layer normal. In smectic C phases, the mesogens are additionally tilted towards the layer normal. ................................ ............................. 4 Figure 1.4. Photographs of oriented and unoriented nematic elastomers (a) and correspo nding X ray patterns of the monodomain (b) and the polydomain (c) sample. ................................ .................... 5 Fi gure 1.5 a)The first LCEs systems developed by Finkelmann et al, b) the most recent chemistry used by White et al, 2015. ................................ ................................ ................................ ..... 7 Figure 1.6 Schematic of stress strain curve for polydomain LCEs; shown the three regions. ............. 8 Figure 1.7 In the LC phase, the polymer backbones experience an anisotropic environment, which leads to an extended chain conformation. At the phase transition to the isotropic phase, the polymer regains its coiled conformation, giving rise to a macroscopic shape change. .......................... 9 Figure 3.1. Schematic of LCE synthesis via thiol acrylate Michael addition "click" reaction. The t hiol mon omers and the mesogens were selected as commercially available monomers. Stoichiometric mixtures of thiol to acrylate functional groups were used to create well defined LCE networks. ................................ ................................ ................................ ................................ ...... 19

PAGE 13

xiii Figure 3.2. Gel fractions analysis for eight LCE networks with varying the concentration and functionality of the crosslinker. Five samples (n=5) were tested at each composition ........................ 22 Figure 3.3 Polydomain LCEs (a) can be ali gned to monodomain (d) using mechanical strain. The 1D and 2D WAXS patterns for an example of 10 tri thiol crosslinked networks. A representative of 2D WAXS pattern for 0% strain (c) and 80% strain (d). The intensity versus azimuthal angle were measured a t fixed strains of 0 (e) and 80% (f). ................................ ................. 24 Figure 3.4. Polydomain nematic to is otropic transition associated with optical changes. (a) Nematic polydomain LCEs are optically opaque whereas isotropic LCEs are optically transparence. (b) Representative endotherms for eight LCE networks with varying crosslinker amount and functionality. (c) Average T NI values along with standard deviations for LCE networks. (d) Average "H f values along with standard deviations for LCE networks. *Represents significant difference with respect to networks using the same functionality crosslinker (p value < 0.05). Represents significant difference with respect to equal crosslinker counterpart (p value < 0.05). Five samples (n=5) were tested at each composition ............................ 26 Figure 3.5. a) storage modulus (E#) and loss tangent (tan !) traces for an example of 10 tri thiol crossli nked networks measured at 3 ¡C/min heating rate and 1 Hz frequency in tension mode, the second temperature sweep plot for a) tri thiol LCE networks and b) tetra thiol LCE networks. The glass transition temperature (T g ) was measured at the peak of tan !. b) T g as a function of the cross linker content for tri and tetra thiol acrylate LCE networks ................................ .............. 27 Figure 3.6. The effect of the thiol cross linker content on stress strain curve measured at 0.2 mm/s at the glass transition temperature of each LCE composition using a) tri and tetra functional crosslinkers. The failure strain was defined by the fracture of the LCE sample. b) Failure strain as a function of the cross linker content for tri and tetra thiol acrylate LCE networks. Represents significant difference with respect to equal crosslinker counterpart (p value < 0.05. ................................ ................................ ................................ ................................ ......... 29

PAGE 14

xiv Figure 3.7. The effect of the thiol cross linker content on strain actuation. Samples were relaxed in a stress free at 120 ¡C for 10 min prior to testing at 5 ¡C/min cooling rate and varying the applied stress in each cycle. A representative plot strain actuation as a function of the thiol crosslinker; a) tri thiol crosslinker; b) tetra thiol crosslinker. The magnitude of strain actuation was measured between strain at 20 ¡C and strain at 120 ¡C. The magnitude of strain actuation was plotted as function of applied stress for eight thiol acrylate LCE networks by arraying thiol crosslinker content; c) tri thiol cross lik er; d) tetra thiol crosslinker. A minimum of 3 samples (n=3) were tested at each condition. ................................ ................................ ................................ ..... 30 Figure 3.8. Work capacity for LCE networks as a function of crosslinking concentration. Work capacity was measured under a constant bias stress of 100 kPa while cooling from the isotropic state. ................................ ................................ ................................ ................................ ...................... 32 Figure 4.1. Schematic of main chain LCE synthesis via a thiol acrylate Michael addition reaction. During synthesis, mesogens are in the isotropic (Iso) phase in a present of solvent (toluene) at 60 ¡ C. The toluene is removed after the reaction is completed to form pol ydomain samples. The formation of nematic (N) and smectic C (SmC) domains form via phase separation of mesogen and thiol spacer. ................................ ................................ ................................ ................ 43 Figure 4.2. (a) Heat flows of five LCE networks with increasing spacer length from C2 to C11. Heat flows are shown on the second heating to reset the thermal history of the networks. (b) Comparison of f irst and second heating cycles of the C9 network. The second heating cycle shows an exotherm once heated above the glass transition temperature due to polymer chain crystallization. (c) Optical images showing isotropic (transparent) to polydomain (opaque) transition of the five networks as a function of cooling. ................................ ................................ ...... 44 Figure 4.3. 2 D WAXS patterns for five LCE networks at room temperature. Diffraction was measured in (a) an unaligned polydomain state and (b) an aligned monodomain state. Alignment was achieved by stretching the samples to 100% engineering strain before analysis. ......................... 45

PAGE 15

xv Figure 4.4. Temperature controlled WAXS analysis of the LCE system using the C9 spacer. Diffraction patterns reveal the transiti on from a smectic C to nematic orientation when heated above 80 ¡ C, while a nematic to isotropic transition occurs when heated above 100 ¡ C. All images were taken under 100% engineering strain. ................................ ................................ ......................... 46 Figure 4.5 Storage modulus (E#) and loss tangent (tan delta) traces for LCE networks with spacer lengths of C6, C9, and C11. Samples were measured at 3¡C/min heating rate and 1 Hz frequency in tension. All samples were annealed above T NI and allowed to cool at room temperature for 24 hours before the first temperature sweep to allow the semi crystallinity to fully form. Samples were tested four times and allowed to set isothermally at 25 ¡C between each sweep for 5, 60, and 120 minutes to show the evolution of the mechanical properties due to polymer chain crystallization. The behavior of the C2 and C3 networks closely resembled that of the C6 and are t hus are only shown in appendix III ................................ ................................ ............ 49 Figure 4.6 (a) Selected actuation plots of five LCE networks with increasing spacer length from C2 to C11 under a 50 kPa bias stress. Samples were equilibrated above T NI and co oled at a rate of 5 ¡ C/min. (b) Average work capacity for each network (n=3). Work capacity was calculated by multiplying the bias stress by the actuation strain. ................................ ......................... 52 Figure 4.7. Photo sequences highlighting multiple functionalities capable within these semi crystalline LCE networks. A C9 stent was synthesize with 15 mol% excess acrylate groups. The 9 mm stent was expanded to 15 mm and photo crosslinked to lock in mesogen orientat ion. (a b) The LCE stent is capable of demonstrating a 1 way shape memory effect when heated above its glass transition. (b c) The LCE stent is also capable of reversible 2 way actuation when heated and cooled around its T NI (d) If the expanded stent is a llowed time to develop polymer crystallinity, it is capable of supporting a 100 g weight, compared to (e) an uncrystallized stent. ...... 54 Figure 5.1 (a) A diacrylate mesogen (RM257), dithiol flexible spacer (2,2' (ethylenedioxy) diethanethiol EDDET), and tetra functional thiol crosslinker (pentaerythritol t etrakis (3

PAGE 16

xvi mercaptopropionate) PETMP) were selected as commercially available monomers. Non equimolar monomer solutions were prepared with an excess of 15% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (D PA) and (2 hydroxyethoxy) 2 methylpropiophenone (HHMP) were added as the respective catalyst and photo initiator to the solutions. (b) Representative polydomain structure and physical samples demonstrating ability to mould different geometries. (c) A mecha nical stress is applied to the polydomain samples to align the mesogens into a temporary monodomain. (d) A photopolymerization reaction is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample. Photo imag e compares sample before and after stretching and photo curing. (e) WAXS pattern of aligned sample confirming nematic structure. (f) POM image of unaligned sample at 20x magnification. *Toluene was used as an optional component to the system to reduce solu tion viscosity ................................ ................................ ......... 60 Figure 5.2. Polydomain and monodomain LCE samples were subjected to 0 and 100 kPa bias stresses and cooled from 120 to 20 ¡ C at a rate of 5 ¡ C/min. Monodomain samples exhibited 45% actuation under zero stress. The monodomain samples in this experiment were programmed by stretching a polydomain sample to 100% strain and photo crosslinking for 10 minutes. .................... 62 Figure 5.3 (a) Alternating regions in a polydomain LCE are photo crosslinked, which become resistant to transparent, monodomain alignment when stretched. (b) An unaligned LCE is heated to the isotropic state and crosslinked with a photo mask Upon cooling, photo crosslinked areas remain isotropic to revealan image. ................................ ................................ ................................ ...... 63 Figure 5.4. cytocompatibility of the TAMAP synthesized LCE was confirmed after both the first and second stages of the reaction using both elution and direct contact test by an independent laboratory (WuXi AppTec, St. Paul, MN, USA). Cellular response to both (a) direct contact and (b) elution tests are shown. ................................ ................................ ................................ ................... 66 Figure 6.1 Schematic of Monodomain Programing via a Two Stage Thiol Acrylate Reaction. (a) A diacrylate mesogen (1,4 bis [4 (3 acryloyloxypropyloxy)benzoyloxy] 2 methylbenzene

PAGE 17

xvii RM 257), dit hiol flexible spacer (2,20 (ethylenedioxy) diethanethiol EDDET), and tetra functional thiol crosslinker (pentaerythritol tetrakis(3 mercaptopropionate) PETMP) were selected as commercially available monomers. Non equimolar monomer solutions were prep ared with an excess of 15 mol% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2 hydroxyethoxy) 2 methylpropiophenone (HHMP) were added as the respective catalyst and photo initiator to the soluti ons. (b) Representative polydomain structure forms via Michael addition (first stage) with a uniform cross link density and latent excess acrylate functional groups. (c) A mechanical stress is applied to the polydomain samples to orient the mesogens into a temporary monodomain. (d) A photopolymerization reaction (second stage) is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample. ................................ ............................... 81 Figure 6.2. Kinetics Study of Michael Addition Reaction with Real Time FTIR. ( a ) Representative two stage thiol acrylate reaction kinetics showing conversion as a function of time using DMPA photoinitiator. At the end of first stage, the thiol groups reached near 100% conversion while 22% of acrylate groups were unreacted. At the end of the second stage, unreacted acrylates reached 100% conversion. ( b ) FTIR absorbance spectra showing the thiol and acrylate conversion before curing at time 0, upon completion of the first stage at 300 minute, and upon completion of the second stage at 320 minute. ................................ ................................ ..... 82 Figure 6.3 Thermomechanics of TAMAP LCE Systems. ( a ) Representative strain to failure curves of four LCE systems with 15 mol% excess acrylate and varying amount of PETMP crosslinker. ( b ) Failure strain as a function of PETMP crosslinker. ( c ) The influence of temperature on failure strain for an LCE system with 15 mol% PETMP. The failure strain is compared alongside the tan $ function of the material measured by DMA. ( d ) Representative glass transition behavior of four LCE systems tested. ( e ) Image of a stretched LCE specimen with 15 mol% PETMP compared to an untested specimen. Error bars in (b) and (c) represent standard deviation. ................................ ................................ ................................ ................................ 83

PAGE 18

xviii Figure 6.4 Shape Switching Pathways in an LCE. This schematic represents several different pathways available to achieve shape switching in LCEs. A custom dog bone sample of 15 mol% PETMP is used in this demonstration with an initial shape of (a). Reversible stress driven actuation is realized between (b c) by adjusting the temperature about T NI while under a constant bias force (60.6 mN); the shape memory effect is achieved by follo wing the programming and recovery cycle of (a b d e); and stress free actuation can be activated thermally between (g h) after a permanent monodomain has been programmed into the sample in step (f). The legend illustrates mesogen orientation in polydomain, monodomain, and isotropic states. T < T NI and T > T NI images were taken at 22 and 90 ¡C, respectively. ................................ ................................ ....... 84 Figure 6.5. Thermomech anical Response in Programmed Monodomain LCE Systems: (a) Shape fixity represents the efficiency of permanently aligning monodomain and all of samples show fixity above 90%. The magnitude of actuation measured between 22 and 90 ¡C on a hot plate. Error bar s represent standard deviation. (b) The magnitude of actuation measured on DMA from 25 to 120 ¡C, the actuation increase with increasing of applied programming strain. ....................... 85 Figure 6.5. Thermomechanical Response in Programmed Monodomain LCE Systems: (a) Shape fixity represents the efficiency of permanently aligning monodomain and all of samples show fixity above 90%. The magnitude of actuation measured between 22 and 90 ¡C on a hot plate Error bars represent standard deviation. (b) The magnitude of actuation measured on DMA from 25 to 120 ¡C, the actuation increase with increasing of applied programming strain. ....................... 85 Figure A.1 X ray Scattering profile intensity as founction of azimuthal angle in WAXS ............... 10 3 Figure A.2 Numerical claculaion by fitting expermenatal data. ................................ ...................... 10 4 Figure A .3 The 1D WAXS p atterns for eight LCE networks with different croslinking density ... 1 05 Figure A.4 The 1 nad 2D WAXS patternts for five LCE networks with different spacer length. ... 1 06 Figure A.5 Temperature controlled 2D WAXS patterns for C6 ................................ .................... 1 07 Figure A.6 Temperature controlled 2 D WAXS p atterns for C1. ................................ ..................... 1 07

PAGE 19

xix Figure B .1 The DSC for C2, 3, 6, and 11. ................................ ................................ ........................ 1 08 Figure B .2 LC phase transtion temperatures as a founction of spacer legth ................................ ..... 85 Figure B.3 DSC for polydomain sample tested before and after stage 2. ................................ ........ 1 1 0 Figure C .1 Represenative of DMA plots for eight L CE network with different crosslink density .. 1 1 1 Figure C.2 Represenative of DMA plots for C2 and C3.. ................................ ............................... 112 Figure C.3 DMA plot for polydomain sample before and after stage 2. ................................ .......... 113

PAGE 20

1 CHAPTER I INTRODUCTION AND BACKGROUND 1.1 Liquid Crystals ( LC ) In 1888, Friedrich Reinitzer identified liquid crystals (LCs) a s a state of matter, when he observed cholesterol melt at 144.5 C to a honey like opaque fluid. Further heating up to 178.5 ¡ C the liquid became clear and transparent. ( 1 ) The term liquid crystal is symbolized to the materials that demonstrate order like crystal while maintaining the behavior of a liquid. The molecules that give rise to this behavior are referred to as a mesogens. Mesogens are rigid molecules made of two to three linearly connected aromatic rings with anisotropic architecture. The molecules can be classified based on shape to rod like (calamitic) or disk like (discotic). Mesogens order at the molecular level, due to the maximization of the interaction energy and minimization the excluded volume (because of Figure 1 1 Polymeric materials displaying liquid crystallinity, a; Liquid crystal polymer (LCP), b; liquid crystal polymer networks (LCNs) are heavily crosslinked materials, c; Liquid crystal elastomers (LCEs) are lightly crosslinked materials.

PAGE 21

2 their anisotropic sha pe architecture). ( 2 ) LC materials change their phase due to external stimuli and are classified into the subcategories of thermotropic (order depends on temperature), and lyotropic (order depends on the conce ntr ation of material in solvent). The interest in these materials has grown rapidly due to their commercial application values especially in the displays industry. ( 2 ) LC devices are dominating the market of d isplays for computers and telecommunication devices, which is now a hundreds of billion dollars market. ( 3 ) Recently, LC materials have been used as active components in other app lications such as solar energy, ( 4 ) optics and photonic s ( 5 ) mechanics and biomedicine. ( 6 ) ( 7 ) P olymeric materials exhibiting liquid crystallinity are merging the liquid crystalline order in the mesogens with elasticity in the polymer, can be classified based on crosslinking to thermoplastics or thermosets. Liquid crystal polymers (LCPs) are thermoplastics uncrosslinked macromolecules such as Vectran. These materials are typically linear polymers, with melting temperatures ( T m ) above 30 0 ¡C and moduli ( E ) that can exceed 100 GPa. Liquid crystalline polymer networks (LCNs) and Liquid crystalline elastomers (LCEs) are considered thermosets due to the existing of a crosslinking agent in their networks (Figure 1. 1). ( 8 ) Cross linked Liquid crystalline polymeric materials can also be further classified based on their glass transition temperature ( T g ) LCNs are highly cross l inked materials with T g above room temperature and moduli of approximately 1 2 GPa, whereas LCEs have a T g below room temperature ( 8 ) The scope of t his dissertation work will only be focused on LCEs 1.2 Liquid crystalline Elastomers (LCEs) LCEs are a remarkable class of materials that encompass the properties of both lightly crosslinked polymer networks (rubber elasticity) and liquid crystalline order (self organization). ( 6 9 10 ) The concept of combining the properties of two subsystems was first proposed by de Gennes et al. in 1975 and was experimentally accomplished by Finkelmann et al. in 1981. ( 11 12 ) The complex structure of LCEs enable several unique mechanical and optical properties; their most fascinating property is the ability to change their shape reversibly in response to external stimuli such as heat ( 9 13 ) or light. ( 14

PAGE 22

3 16 ) Applications for such materials range from sensors and actuators for biological applications such as artificial muscles, ( 17 19 ) biomimetic iris lenses, ( 20 ) tunable optical gathering devices, ( 21 ) cell scaffolds, ( 22 23 ) micro grippers for robotics, ( 24 ) microvalves for microfluidic systems, ( 25 ) and organic solar cells. ( 26 ) 1.2.1 Classification LCE s are highly diverse class of materials and can be classified in to many categories Herein, we will cover the main classes of LCEs They can be categorized based on the attachment of mesogens to the polymer backbone (side chains or main chains), LC phase structure (smectic or nematic), and the domain t ype (polydomain or monodomain). Figure 1 2 Different attachment geometries for the synthesis of LCEs: side chain elastomers with end on (a) or side on (b) attached mesogenic side chains and main chain elastomers with mesogenic units incorporated end on (c) or side on (d) into the polymer main cha in.

PAGE 23

4 1.2.1.1 Classification Based on Attachment of the Mesogens to the Polymer Backbone The structural components that give a rise to the ordered LC phases are called mesogens. The y are the foundation of the LC domains and typically composed of two to three linearly connected aromatic rings (rod like) with flexible ends. These moieties can be directly placed within the polymer backbone to create main chain LCEs or as a side group ( i .e. side on or end on LCEs) (Figure 1. 2). ( 6 ) Before the discovery of main chain in 1997, early work on LCEs was focused on side chain LCEs on both nematic and smectic systems. ( 10 ) However, main chain LCEs have attracted more attention due to the direct coupling between mesogenic order and pol ymer backbone conformations. ( 27 32 ) This direct coupling allows main chain LCEs to exhibit higher degrees of mesogen orient ation, mechanical anisotropy, and thus strain actuation, compared to side chain LCEs. ( 31 ) This dissertation work will only be dedicated to main chain LCEs 1.2.1.2 Classification Based on LC Phase Type The ordered LC phase state exists in a temperature range in between the solid crystalline and the Figure 1 3 Three important LC phases. In the nematic phase, the mesogens possess a short range order and are aligned parallel in a uniform direction, defined by the director. Smectic A phases exhibit a layered struc ture with the mesogens parallel to the layer normal. In smectic C phases, the mesogens are additionally tilted towards the layer normal.

PAGE 24

5 isotropic disordered liquid state. T he t hree common types of LC phases are ( Figure 1.3 ) ( 33 ) : Nematic (N) mesogens are oriented in a uniform direction along a director (orientationa l order); s mectic A (SmA) a layered structure with orientati onal and positional order; and s mectic C (SmC) similar to SmA but with the mesogens titled with respect to the director. ( 34 ) In general, smectic LCEs have larger actuation, lower failure strain, higher modulus, and greater enthalpy compared to nema tic LCEs. This is due to smectic LCEs typically having higher order parameters compared to nematic LCEs. ( 35 ) A general belief is that the formation of LC phases in elastomers is dictated by the mesogen structure: the mesogen core and flexible tails. ( 6 34 36 37 ) Nematic m esogens should yield nematic LCE systems, and smectic mesogens should correspond to smectic LCEs. ( 38 ) Krause et al. prepared numerous smectic C and nematic LCE systems with a variety of thermomechanical properties by modifying the mesogen structure. 1.2.1.3 Classification Based on Domain T ype Polydomain LCEs are created in the absence of external fields, where monodomain must be created in a present of external fi elds. For many applications, in particular actuators, it is essential to prepare monodomain LCEs (also know n as liquid single crystalline elastomers (LSCE s )), where the mesogens orient along a reference direction called a "director". A monodomain can be fo rmed temporarily by applying an external stress ( i.e. hanging a weight) to a sample, which will align the polymer chains and orient the mesogens in the direction of the stress. Permanent programming of the Figure 1 4 Photographs of oriented and unoriented nematic elastomers (a) and corresponding X ray patterns of the monodomain (b) and the polydomain (c) sample.

PAGE 25

6 monodomain can be achieved via a multi step proces s, which involves producing a lightly cross linked gel followed by immediate application of mechanical stress to induce orientation of the mesogens. Once aligned, the reac tion is continued to form covalent crosslinks and stabilize the monodomain. ( 39 ) Other "one pot" alignment techniques can be performed in the presence of electric fields or by surface alignment ( i.e. rubb ing polyimide on a glass slide) during polymerization; however, these methods are generally limited to thin film samples. ( 6 ) LCEs prepared without an aligning method do not form monodomain. Rather, they tend to form a disordered arrangement of micro domains, called polydomains, where each individual domain is defined as a region of uniform orientation. In other words, the mesog ens are oriented along the director locally but lack orientation globally ( Figure 1. 4 ). ( 6 ) The polydomain state arises from quenched disorder, which comes from defects during synthesis caused by chain entanglements and cross linking. ( 40 41 ) Polydomain structures can be viewed under the polarizing optical microscope as a texture, called a Schlieren tex ture. ( 40 42 ) Polydomain samples are optically opaque because LC domains strongly scatter light depending on their mesogen orient ation, while monodomain samples are optically transparent because LC domains are oriented and as a result do not scatter light. ( 43 ) Polydomain and monodomain morphologies exist below the isotropic transition temperature (T i ) and disappear upon heating above T i 1.2.2 Preparations The first LCE network was synthesized by Finkelmann et al. utilizing a hydrosilylation reaction and was a polydomain nematic side chain LCE ( Figure 5a). ( 12 ) Since then, this reaction has been applied widely to synthesize side and main chain LCEs. ( 44 4 7 ) The chemistry of main chain LCEs has since evolved resu lting in numerous synthetic approaches that have been used to create and modify materials to enhance the quality, stability, tailorability, reproducibility, accessibility, and to enable responsiveness to a variety of stimuli such as heat or light. Ortiz et al. have used epoxy resins to synthesize main chain LCEs in a polydomain state. ( 48 ) Recently, Pei et al. have utilized exchangeable covalent bonds to repeatedly program the monodomain in epoxy based main chain

PAGE 26

7 LCEs. ( 49 ) It should be noted that epoxy based LCEs typically have higher T g values compared to hydrosilylation based LCEs. Several photo crosslinking reactions have also been propose d for the formation of main chain LCE networks using functionalized pre polymer chains such as polyesters or thiol ene. ( 30 32 50 ) Thiol ene/yne systems have received considerable attention in the field of LCE due to fast polymerization, low volume shrinkage and shrinkage stress, the formation of homogeneous networks, and minimal sensitivity to oxygen inhibition. ( 51 54 ) Recently, our group introduced a two stage thiol acrylate Michael addition photopolymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic monodomain main chain LCEs. ( 55 56 ) Several recent examples of using TAMAP methodology to prepare main chain LCEs have been performed. ( 57 59 ) The most recent chemistry develops came f rom White et al. this system utilizes a two step method of az a Michael addition and photopolymerization reaction. ( 60 ) 1.2.3 Crosslinking History LCEs can be obtained via two different routes bas ed on their synthesis and cross linking history. LCEs that are polymerized in the isotropic state at high temperatures or in the presence of a solvent are known as isotropic polydomain nematic elastomers (i PNEs). Otherwise samples that are polymerized at temperatures below isotropic tr ansition temperature ( T i ) are known as nematic Figure 1 5 a)The first LCEs systems developed by Finkelmann et al b) the most recent chemistry used by White et al, 2015.

PAGE 27

8 polydomain nematic elastomers (n PNEs). The resulting two types of polydomain LCEs share some thermomechanical properties with deli c ate differences in dynamic features. ( 61 ) For example, they have the same rubbery moduli and differ in stress strain dynamic and actuation behavior ( in preparation by Traugutt et al, 2017 ) i PNE samples store more strain compared n PNEs. More research must be done to fully investigate their differences. 1.2.4 Stress Strain B ehavior LCEs exhibit un ique stress stain behavior and can be depended mainly on the sample's domain type (monodomain or polydomain). The stress stain behavior also can be influenced by temperature, LC phase, crosslink density and history and the direction of the applied stress. ( 38 62 63 ) For example, p olydomain LCEs undergo stress strain behavior consisting of three regions. Region One is a linear elastic deformation of the polydomain, which occurs at relatively low strain values, and the slope of the st ress strain curve represents the modulus of the materials. Region Two undergoes the polydomain monodomain transition, which takes place at an intermediate strain where the specimens change from opaque to optically transparent. Ideally, this transition lead s to a plateau in the stress strain curve with a large increase in strain a constant stress (soft elasticity). Region Three is continued Figure 1 6 Schematic of stress strain curve for polydomain LCEs S hown in three regions.

PAGE 28

9 deformation of the monodomain. This occurs once the polydomains have all been oriented and the modulus increases more s harply as the polymer chains become aligned. The strain values differentiating these three regions highly depend on temperature, LC phase and the concentration of the crosslinker ( Figure 7 ). ( 48 ) The stress str ain behavior of monodomain LCE (highly anisotropic material) i s extremely dependent on the direction of the applied stress Monodomain LCEs stretched parallel to the applied stress have a linear elastic response strain. Where, monodomain LCEs stretched perpendicular to the applied stress display similar to polydomain LCEs stress strain behavior with distinct soft elasticity plateau. ( 64 ) 1.2.5 Actuation Thermal actuation of LCEs relies on a reversible anisotropic isotropic transition (T i ) associated with LC order. ( 33 65 ) To program an LCE for actuation, the mesogens must first be oriented along a director to form a monodomain ( i.e. anisotropic mesophase). The polymer chains elongate when the mesog ens orient in the nematic or smectic phase, whereas in the isotropic phase they recover. When LCE is cooled from the isotropic to the nematic or smectic phase, the anisotropy of the polymer chains in the nematic or smectic phase causes extension of the sam ple along the long axis of the Figure 1 7 In the LC phase, the polymer backbones experience an anisotropic environment, which leads to an extended chain con formation. At the phase transition to the isotropic phase, the polymer regains its coiled conformation, giving rise to a macroscopic shape change.

PAGE 29

10 ellipse, which is the axis of the director for main chain materials, during cooling into the ordered phase. ( 8 ) After the initial elongation, main chain systems may continue to elongate when cooled further below the isotropic transition due to the decrease in hairpin defects (polymer chain folding) with decreasing temperature (when the temperature increases the hairpin increases). This process is entirely reversible. ( 66 )

PAGE 30

11 CHAPTER II RESEARCH MOTIVATION AND GOALS Liquid crystalline elastomers (LCEs) are smart materials that are known for their ability to undergo reversibl e thermal actuation due to the change in their liquid crystalline phase from an anisotropic ( nematic o r smectic ) to isotropic state (Figure 1.7). ( 33 ) LCEs can be activated by heat or li ght and have demonstrated reversibl e strains up to 400%. ( 67 ) As a result, LCEs have been prop osed for many sensor and actuator applications, and most particularly as potential artificial muscles. ( 18 68 ) Although, liquid crystal technology has experienced an impressive commercial success in display industry which is multi billion dollar business ( 3 ) However, LCEs have not yet had the same success, due to many research challenges or unclear principles in our understanding to these fascinatin g materials. First, research challenges such as complex chemistry, synthesis, and programming conditions, which have all limited manufacturability and scalability of the LCEs for many researchers without extensive chemistry backgrounds. Therefore, the curr ent LCE synthesis methods are not practical for large scale manufacturing and are only used to produce small scale samples such as thin films, ( 69 70 ) fibers, ( 50 71 72 ) micro pil lars, ( 73 ) or micro beads. ( 74 75 ) Second the unclear principles in our understanding to some structure to property relationships suc h as the influence of the cross linker content and functionality on the LC phase behavior (i.e. phase type and transitio n temperature) and other thermo mechanical properties. Moreover, there is little known about what dictate the LC phase formation in the elastomers. Last, one of the large problems associated with LCEs that prepared during the last three decades, is the lack of the trend in the thermo mechanical properties with structure of the materials. Because most of LCEs have been obtained from differently prepared samples that accordingly had different qualities of monodomain, therefore is nearly impossible to c ompare the properties/ behaviors in these systems. ( 33 ) A comprehensive study to develop a new approach to prepare LCEs and carefully study their thermomechanical properties are therefore necessary to enhance manufacturability, understand material behavior, and assist the design

PAGE 31

12 in new engineering applications In t his research we used a profoundly new approach to develop LCEs based on thiol acrylate click reaction and two stage Micha el addition photopolymerization (TAMAP) reaction, both of which have not previously been investigated for LCE synthesis. First, the thiol acry late reaction is a powerful tool for polymer synthesis. It has been widely used recently, due to fast reaction rate with high yield, no byproduct, with or without solvent, in ambient temperature, with or without catalyst present in the syste m, and minimal oxygen inhibition. ( 53 ) All of these great qualities make this reaction suitable to use as a tool for LCE synthesis with solely commerc ial available starting LC materials and monomers, this will open the doors researchers to explore these materials in a very facile manner. Second, the TAMAP approach was implemented to create monodomain LCEs. This method is a new paradigm for LCE manufactu ring at both large and small size scales by offering a high degree of tailorability not achievable by current methods via clear control of initial and final crosslinking density of the network and initial stretching conditions For the first time it will b e possible to investigate how these crosslinking densities, along with mechanical stretching conditions, influence the shape fixity and actuation behavior of LCEs. Furthermore, this approach will help overcome the "synthesis barrier" to allow investigators with diverse backgrounds easier access to LCE research. The ultimate goal of this work is to establish structure property relationships in thiol acrylate based main chain liquid crystalline elastomers. The main hypothesis of this work is that both thiol a cylate click reaction and two stage thiol acrylate reaction can be used to increase LCE tailorability and manufacturability in an effort to better understand the associated structure property relationships. Specific aims to meet our four research objectives are listed below: Specific aims #1: Investigate the impact of cross linking (i.e. varying cross linking concentration and functionality) on the LC phase transition and thermomechanical properties. The purpose of this aim is to systematically inve stigate the impact of varying cros s linking density and

PAGE 32

13 changing cross linker functionality on the properties of thiol acrylate LCE systems. We hypothesize that using a click reaction to produce more uniform networks may reveal clear structure property rel ationships in LCEs as well as demonstrate improving in the actuation performance compare to other main chain LCEs prepared using different synthetic techniques. The evolution in the network architecture influences the coupling between the liquid crystallin e behavior and polymer chains, which has a pronounced effect on the thermomechanical behaviors, including isotropic transition temperature (T i ) glass transition temperature ( T g ), rubbery modulus, failure strain ( f ), and the actuation performance of LCEs. This aim will utilize the high efficiency and orthogonality of the thiol acrylate Michael addition reaction to synthesize well defined, uniform networks. Such a reaction offers a facile way to tailor the network structure. This aim wi ll highlight and quantify the effect of the cr oss linker on influencing a broad range of thermomechanical properties. This will aid to demonstrate how simple modifications to composition can be used to optimize the physical and mechanical properties of thes e networks for a wide range of potential applications. Specific aims #2: Examine what dictate s the LC phase formation in LCEs? And modulate LC phase structure via varying the spacer's length. The purpose of this aim is to modulate LC phase structure usin g a single nematic mesogen, RM257, by controlling the thiol spacer length. Herein, we will use series of main chain LCE systems that are capable of multiple LC phases at room temperature. We hypothesize that t wo or more LC phases will be realized, smectic C and nematic. Smectic C phases should be observed when using longer thiol spacers, as these spacers drive a n ano scale segregation of ternary incompatible layers of bulk alkyl thiol functionalized spacers, flexible propylene oxide acrylic terminal chains and mesogen cores. Segregation of these distinct segments is the main contribution to the formation of the smectic phase. ( 76 ) Shorter thiol spacers should not segregate; therefore, they must be engendered a nematic phase. Furthermore, the nematic to isotropic transition temperature (T NI ) should dependent on the spacer length (mesog e n concentration) Thi s approach should allow us to compare the

PAGE 33

14 thermomechanical properties of smectic C and nematic LCE actuators with near identical chemical compositions, which has traditionally been extremely difficult to achieve due to vastly different reactions and compos itions used in preparation. Specific aims #3: Explore TAMAP reaction to create tailored bulk monodomian LCE samples capable of "hands free" actuation well as offer spatio temporal control over liquid crystalline behavior. As a simple, readily accessible, powerful methodology, we will introduce a previously unexplored approach to synthesize and program main chain LCEs using TAMAP reaction. Initial polydomain LCE samples can be formed using a thiol acrylate "click" reaction w ith the facile ability to tailor the crosslinking density and polymer structure. If an excess of acrylate groups exists, a second independent photopolymerization reaction can be used to align LCE into monodomain. This approach will offer neat and scalable synthesis of LCEs as well as offers exceptional spatio temporal control of the second stage photopolymerization reaction to influence liquid crystalline behavior. Specific aims #4: Investigate how the programming of the monodomain is influenced by the in itial crosslinking density initial programed strain, and temperature. The purpose of this aim is to explore and demonstrate the robust nature of the TAMAP reaction to prepare main chain LCEs by investigating the influence of crosslinking density and progr amed strain conditions, and temperature on the thermomechanics of the LCE systems. We will demonstrate a wide range of thermomechanical properties and actuation performance that are achievable using this reaction.

PAGE 34

15 CHAPTER III THIOL ACRYLATE MAIN CHAIN LIQUID CRYST ALLINE ELASTOMERS WITH TUNABLE THERMOMECHANICAL PROPERTIES AND ACTUATION STRAIN 3.1 Abstract The purpose of this study was to investigate the influence of crosslinking on the thermomechanical behavior of liquid crystalline elastomers (LCEs). Main chain LCE networks were synthesized via a thiol acrylate Michael addition reaction. The robust nature of this reaction allowed for tailoring of the behavior of the LCEs by varying the concentration and functionality of the crosslinker. The iso tropic rubbery modulus, glass transition temperature, and strain to failure showed strong dependence on crosslinker concentration and ranged from 0.9 MPa, 3¡C, and 105% to 3.2 MPa, 25¡C, and 853%, respectively. The isotropic transition temperature ( T i ) was shown to be influenced by the functionality of the crosslinker, ranging from 70¡C to 80¡C for tri and tetra functional crosslinkers. The magnitude of actuation can be tailored by controlling the amount of crosslinker and applied stress. Actuation increas ed with increased applied stress and decreased with greater amounts of crosslinking. The maximum strain actuation achieved was 296% under 100 kPa of bias stress, which resulted in work capacity of 296 kJ/m 3 for the lowest crosslinked networks Overall, the experimental results provide a fundamental insight linking thermomechanical properties and actuation to a homogenous polydomain nematic LCE networks with order parameters of 0.80 when stretched.

PAGE 35

! 16 3.2 Introduction Liquid crystalline elastomers (LCEs) are a remarkable class of materials that encompass the properties of both lightly crosslinked polymer networks (rubber elasticity) and liquid crystalline order (self organization). ( 6 9 10 ) The complex structure of LCEs enable several unique mechanical and optical properties; their most fascinating property is the ability to change their shape reversibly in response to external stimuli such as heat ( 9 13 ) or light. ( 14 16 ) Applications for such materials range from sensors and actuators for biological applications such as artificia l muscles, ( 17 19 ) biomimetic iris lenses, ( 20 ) tunable opt ical gathering devices, ( 21 ) cell scaffolds, ( 22 23 ) micro grippers for robotics, ( 24 ) microvalves for microfluidic systems, ( 25 ) and organ ic solar cells. ( 26 ) For many applications, in particular actuators, it is essential to prepare monodomain LCEs, where the mesogens orient along a reference direction called a "director". In practice, this can be achieved by applying a mechanical load or an aligning method, such as the use of rubbed polyimide or magnetic fields while crosslinking. ( 6 ) LCEs prepared without an aligning method do not form monodomains. Rather, the y tend to form a disordered arrangement of micro domains, called polydomains, where each individual domain is defined as a region of uniform orientation. In other words, the mesogens are oriented along the director locally but lack orientation globally. Th e polydomain state arises from quenched disorder, which comes from defects during synthesis caused by chain entanglements and crosslinking. ( 40 41 ) Polydomain structures can be viewed under the polarizing optical microscope as a texture, called a Schlieren texture. ( 40 42 ) Polydomain samples are optically opaque because LC domains strongly scatter light depending on their mesogen orientation, while monodomain samples are optically transparent because LC domains are oriented and as a result do not scatter ligh t. ( 43 ) Polydomain and monodomain morphologies exist below the isotropic transition temperature ( T i ) and disappear upon heating above T i

PAGE 36

! 17 The first polydomain LCE was synthesized by Finkelmann et al. utilizing a hydrosilylation reaction and was a nematic side ch ain LCE. ( 12 ) Since then, this reaction has been applied to synthesize main chain LCEs. ( 44 47 ) The chemistry of main chain LCEs has since evolved resulting in various synthetic approaches that hav e been used to create and modify materials to enhance the quality, stability, tailorability, reproducibility, accessibility, and to enable responsiveness to a variety of stimuli such as heat or light. Ortiz et al. have used epoxy resins to synthesize main chain LCEs in a polydomain state. ( 48 ) Recently, Pei et al. have utilized exchangeable covalent bon ds to repeatedly program the monodomain in epoxy based main chain LCEs. ( 49 ) It should be noted that epoxy based LCEs typically have higher T g values compared to hydrosilylation based LCEs. Several photo crosslinking reactions have also been proposed for the formation of main chain LCE networks usi ng functionalized pre polymer chains such as polyesters or thiol ene. ( 30 32 50 ) Thiol ene/yne systems have received considerable attention in the f ield of LCE due to fast polymerization, low volume shrinkage and shrinkage stress, the formation of homogeneous networks, and minimal sensitivity to oxygen inhibition. ( 51 54 ) Recently, our group introduced a two stage thiol acrylate Michael addition photo polymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic monodomain main chain LCEs. ( 55 ) Several recent examples of using TAMAP methodology to prepare main chain LCEs have been performed. ( 57 59 77 78 ) Many functional properties of LCEs rely on the reversible anisotropic isotropic transition associated with LC order. Numerous studies have shown that T i can be correlated to the crosslink density in hydrosilylation based LCE systems. ( 79 81 ) High crosslink densities of the network can disrupt the heat fusion and lower T i leading to a less stable liquid crystalline phase, where lightly crosslinking hardly affects T i ( 79 ) Tsuchitanti, et al. has shown that the effects of crosslinker geometries have a pronounced effect on the orientation of the mesogen s and T i The study suggests an increase in functionality of the crosslinker leads to an increase in T i This is due to the

PAGE 37

! 18 localization of mesogenic monomers in a heterogeneous network structure with a non uniform distribution crosslinker. ( 82 ) Other studies have looked at the influence of crosslinking density o n the mechanical properties such as the glass transition temperature ( T g ), failure strain, the breadth of the soft elasticity plateau and modulus. ( 35 83 84 ) These studies have found that T g and modulus increase with an increase in the crosslinking density where the failure strain soft elasticity plateau decreases with the increasing crosslinking de nsity. The synthetic method used in those studies relied on a hydrosilylation reaction. This method leads to random crosslinking. Therefore, the co relationship between the structure, properties, and actuation perfo rmance may be hard to realize. The purpo se of this study is to systematically investigate the impact of varying crosslinking density and changing crosslinker functionality on the properties of thiol acrylate LCE systems. We hypothesize that using a "click" reaction to produce more uniform networ ks may reveal clearer structure property relationships in LCEs as well as demonstrate improved actuation performance compared to other main chain LCEs prepared using different synthetic techniques. The evolution in the network architecture influences the c oupling between the liquid crystalline behavior and polymer chains, which has a pronounced effect on the thermomechanical behaviors, including T i T g rubbery modulus failure strain ( f ), and the actuation performance of LCEs. This study utilized the high efficiency and orthogonality of the thiol acrylate Michael addition reaction to synthesize well defined, uniform networks. Such a reaction offers a facile way to tailor the network stru cture. This work will highlight and quantify the effect of the crosslinker to influence a broad range of thermomechanical properties. This will aid to demonstrate how simple modifications to composition can be used to optimize the physical and mechanical p roperties of these networks for a wide range of potential applications.

PAGE 38

! 19 3.3. Experimental 3.3.1. Materials F igure 3 1 Schematic of LCE synthesis via thiol acrylate Michael addition "click" reaction. The t hiol monomers and the mesogens were selected as commercially available monomers. S toichiometric mixtures of thiol to acrylate functional groups were used to create well defined LCE networks.

PAGE 39

! 20 Pentaerythritol tetrakis(3 mercaptopropionate) (PETMP), trimethylopropane tris(3 mercaptopropionate) (TMPMP), 2,2 (ethylenedioxy) diethanethiol (EDDET), dipropylamine (DPA), and toluene were purchased from Sigma Aldrich. 4 bis [4 (3 acryloyloxypropypropyloxy) benzoyloxy] 2 methylbenzene (RM257) was obtained from Wilshire Technologies, Inc. (Princeton, NJ, USA). T he chemical structures of the monomers and catalyst are shown in Figure 1. All materials were used in their as received condition without further purification. 3.3.2. Synthesis of Liquid Crystalline Elastomers LCE samples were synthesized via a thiol acryl ate Michael addition reaction. LCE networks were prepared starting with two thiol monomers. The thiol monomers were selected for their use as a multi functional crosslinking monomer and di functional flexible spacer between mesogens. The flexible spacer (E DDET) was mixed with only one crosslinking monomer at a time, either the tri functional TMPMP or tetra functional PETMP. The ratio of thiol crosslinker to flexible spacer was systematically varied using 10, 20, 40, and 80 mol% of functional groups belongin g to the crosslinker. Thiol solutions were added to the diacrylate mesogen, RM257, in a stoichiometric balance, which was dissolved in 50 wt% of toluene at 80C for 5 minutes prior to the addition of the thiol solution. Once the solution returned to room temperature, 1 mol% of DPA was added to catalyze the reaction. The solution was mixed vigorously using a Vortex mixer (No: 94540, Toronto, ON, Canada). Air bubbles were removed from the solution under a 500 mm # Hg vacuum. The solution was then injected betw een two glass slides separated with 1 mm spacers and left to cure overnight After the polymerization was completed, the samples were placed in an oven for 24 hours at 80 C under a 500 mm # Hg vacuum to remove the solvent.

PAGE 40

! 21 3.3.3. Gel Fraction Tests LCEs were extracted in toluene for 1 week to determine the swelling ratio and gel fraction, GF, of the networks. LCE films were cut into rectangular samples measuring approximately 22 x 5 x 1 mm 3 Each sample was then placed in a vial of 25 mL of toluene for th e experiments. After 1 week, samples were removed from the swelling medium, dabbed dry with a paper towel, and dried for 48 hours in vacuum oven at 80 C. The gel fraction, was calculated by: GF = ! !"" ( 3. 1) where W i is the initial dry weight of the sample and W f is final weight of the sample. Five samples (n = 5) were tested for each composition. 3.3.4. X Ray Scattering In order to investigate the nanostructure of liquid crystal in the network, X ray analysis was per formed using Forvis Technologies wide angle X ray scattering (WAXS) 30W Xenocs Genix 3D X ray source (Cu anode, wavelength = 1.54 ) and Dectris Eiger R 1M detector. The beam size was 0.8 mm X 0.8 mm, and the data was collected at a sample to detector dist ance of 113 mm. The sample was exposed to the X ray for 30 min. The flux was 4x10 7 X rays/s. The scattering patterns were analyzed and plotted using intensity versus azimuthal angle by Rigaku SAXSgui and Igor Pro software to determine the d spacing of LCE s using the Bragg's equation below: ! !"# ( 3. 2) where % is the X ray radiation wavelength (1.5405 ), d is the spacing between long range ordering of mesogens in LCE network, and & is the scattering angle. Data was gathered for sample

PAGE 41

! 22 s tretch at 0 and 80% strain to identify the crystal structure for both polydomain and monodomain, respectively. 3.3.5. Differential Scanning Calorimetry (DSC) DSC was performed using a TA Instruments Q2000 machine (New Castle, DE, USA). Samples with a mass of approximately 10 mg were loaded into a standard aluminum DSC pan. The samples were heated rapidly to 120 ¡C at 10 ¡C /min, held isothermally for 10 min, and cooled slowly to 50 ¡C at a rate of 2 ¡C/min to reset any thermal history within the sample. Samples were then heated to 120 ¡C at a rate of 20 ¡C/min. The isotropic tr ansition temperature ( T i ) was defined as the minimum value of the endothermic peak. The reported enthalpy ( H f ) change is measured by integrating the endothermic energy well corresponding to the transition from the nematic polydomain to isotropic state. St atistical analysis was performed on measurements of T i and H f ANOVA analysis followed by a Tukey's t test was first performed to identify any significant differences between values within networks with either tri or tetra thiol crosslinkers. A Student's t test was then performed to identify any differences between samples with equal crosslinker concentrations but different crosslinker functionality. A significant difference was identified when the p value was less than 0.0 5. Figure 3 2 Gel fractions analysis for eight LCE networks with varying the concentration and functionality of the crosslinker. Five samples (n =5) were tested at each composition

PAGE 42

! 23 3.3.6. Dynamic Mechanical Analysis (DMA) DMA was performed using a TA Instruments Q800 machine (New Castle, DE, USA). Rectangular samples measuring approximately 20 x 10 x 0.8 mm 3 were tested in tensile mode, with the active length measuring approximate ly 10 mm. Samples were cycled at 0.2% strain at 1 Hz and heated from 50 to 120 ¡C at a rate of 3 ¡C/min. T g was defined as the temperature corresponding the peak of tan curve. Nematic modulus ( E' n ) and isotropic modulus ( E' i ) were measured using the sto rage modulus values at 25 ¡C and 115 ¡C, respectively. 3.3.7. Strain to Failure Tests Strain to failure tests were performed using an MTS Insight 30 (Eden Prairie, MN, USA) equipped with an LX 500 laser extensometer, thermal chamber, and 500 N load cell. F or these experiments, samples were molded in an HDPE mold according to ASTM Type V dog bone dimensions at a depth of 1 mm during synthesis. The gage cross sectional area measured 3 mm x 1 mm. The samples were deformed at T g at a rate of 0.2 mm/s until failure, defined by sample fracture 3.3.8. Strain Actuation Characterization Strain actuation was measured using the Q800 machine. Sample ends were wrapped with aluminum foil and loaded in the DMA machine in tensile mode with an active length equal to 5 mm. The cross sectional areas of the samples measured 1 x 5 mm 2 Samples were equilibrated at 120 ¡ C. A constant b ias stress ( bias ) was then applied to the samples, while the samples were heated and cooled between 120 and 50 ¡ C at 5 ¡ C/minute. The stress values investigated were 1, 10, 50, and 100 kPa. The maximum of actuation ( a ) was defined by measuring the diffe rent between minimum and maximum engineering strain values measured at 120 and 50 ¡ C, respectively. The estimated volumetric work capacity of the networks was measured by multiplying the actuation strain by the applied bias stress (Eq. 3).

PAGE 43

! 24 !"#$ !" !"#$%& ! !"#$ !"# !"#$ !" ! !"#$ (3. 3 ) The maximum bias stress was selected to be 100 kPa to suit all the tested samples. Stresses greater than 100 kPa frequently caused fracture at the sample grip interface at el evated temperatures for low crosslinked samples. 3. 4. R esults Gel fraction analysis was performed as a measure of network formation and conversion in the LCE systems ( Figu re 2 ). Eight different compositions were synthesized and tested. LCE compositions are identified by mol% of functional groups belonging to the thiol crosslinker Figure 3 3 Polydomain LCEs (a) can be aligned to monodomain (d) using mechanical strain. The 1D and 2D WAXS patterns for an example of 10 tri thiol crosslinked networks. A representative o f 2D WAXS pattern for 0% strain (c) and 80% strain (d). The intensity versus azimuthal angle were measured at fixed strains of 0 (e) and 80% (f).

PAGE 44

! 25 throughout the study. The thiol acrylate reaction yielded a high gel fraction >90% for all of L CEs networks. Both sets of networks showed an increase in gel fraction from 10 to 80 mol% crosslinker, converging at 99% gel fraction at high crosslink densities. At lower crosslink densities, networks formed with tetra functional crosslinkers had higher gel fractions than their tri functional counter parts. WAXS analysis was performed in order to investigate the short range and long range order of the LCEs. The 1D and 2D WAXS patterns of the LCE networks were generated wh en the samples were strained at 0 and 80% to measure the samples in unoriented (polydomain) ( Figure 3. 3a ) and oriented (monodomain) states ( Figure 3. 3b ), respectively; however, only the 10 wt% tri thiol network is presented as a representative sample. All other samples are shown in the a p pendix For

PAGE 45

! 26 unoriented sampl es at 0% strain, no clearly defined peaks were observed in the 1D plot of intensity versus azimuthal angle for tri thiol ( Figure 3. 3e ) conversely, oriented samples at 80% strain all revealed periodic peaks separated by 180 ¡ ( Figure 3. 3f ), which is indicative of nematic order a monodomain structure. All unoriented samples showed a diffuse ring in their 2D WAXS patterns, while oriented samples revealed two bright spots separated by 180 ¡ Representative 2D patterns are shown in Figures 3. 3c and 3d for LCE samples strained at 0 and 80%, respectively. All LCEs exhibited similar scattering patterns with d spacing values of 3.16 , indicating that the Figure 3 4 Polydomain nematic to isotropic transition associated with optical changes. (a) Nematic polydomain LCEs are optically opaque whereas isotropic LCEs are optically transparence. (b) Representative e ndotherms for eight LCE networks with varying crosslinker amount and functionality. (c) Average T i values along with standard deviations for LCE networks. (d) Average "H f values along with standard deviations for LCE networks. *Represents significant diffe rence with respect to networks using the same functionality crosslinker (p value < 0.05). Represents significant difference with respect to equal crosslinker counterpart (p value < 0.05). Five samples (n=5) were tested at each composition

PAGE 46

! 27 microstructure of the LC domains was not affected by their composition. Small angle X ray diffraction (SAXS) was also performed to verify no smectic phases were present (not shown). Due to the presence of toluene during synthesis and crosslinking, these LCE samples are classified as isotropic polydomain nematic elastomers (i PNEs). The orientation parameter for this system is found to be ~ 0.80. Ther mal analysis was used to investigate the influence of the crosslinker concentration and functionality on LCE networks ( Figure 3. 4 ). Representative DSC traces showing heat flow as a function of temperature is shown in Figure 3. 4b All LCE networks show a s tepwise decrease in the heat flow signals around temperatures attributed to T g in the vicinity of 0 to 20 ¡C; however, DSC was primarily used to characterize the isotropic transition. These networks demonstrated endothermic wells, shown as valleys, in the heat flow signals around 70 ¡C for tri thiol networks and 80 ¡C for tetra thiol networks. The minimum of the energy well marks the T i of the nematic LCE. The tri thiol networks had an average T i of 70 ¡C with no statistical differences across compositions ( Figure 3. 4c ). Conversely, tetra thiol networks had significantly higher T i values than their tri thiol counterparts, with the exception of the 80 mol% tetra thiol network. The tetra thiol networks with lower crosslinking amounts (10, 20, and 40 mol%) had an average T i of 80 Figure 4 5 a) storage modulus (E# ) and loss tangent (tan !) traces for an example of 10 tri thiol crosslinked networks measured at 3 ¡C/min heating rate and 1 Hz frequency in tension mode, the second temperature sweep plot for a) tri thiol LCE networks and b) tetra thiol LCE networks. T he glass transition temperature (T g ) was measured at the pea k of tan !. b) T g as a function of the cross linker content for tri and tetra thiol acrylate LCE networks

PAGE 47

! 28 ¡C, while the 80 mol% network had a significantly lower average T i of 72 ¡C. The influence of network structure on H f was also investigated ( Figure 3. 4d ). There were no significant differences in H f across all of the networks tested with the exception of the 10 mol% tri thiol network. The 10 mol% tri functional network has an average H f of 0.955 J/g, which is 80% higher than its tetra thiol counterpart. Althou gh not statistically significant, a general trend appears to suggest H f decreased with increasing crosslinker content. Data for each network is presented in Table 3. 1 The thermomechanical response of the LCE networks was next investigated using DMA ( Figu re 3. 5 ). A representative plot for the 10 mol% tri thiol network ( Figure 3. 5a ) compares the storage modulus ( E' ) and loss tangent ( tan ) as a function of temperature. The sample demonstrated a glassy plateau below 0 ¡C, followed by a stepwise decrease in E' that corresponds with the onset of the glass transition ( T onset ) for the network. In this study, T g was measured at the maximum of the tan curve and ranged from 3 to 25 ¡C. There was a noticeable change in slope and concavity in the E' for the LCE net works at T g Furthermore, all LCE networks demonstrated similar DMA behavior (See Figure A3.1 ) and a dramatic decrease in E' at T i a phenomenon often termed "dynamic soft elasticity ". The samples recover to a rubbery plateau as they are heated into the is otropic phase. It is important to notice that tan of the networks remain elevated within the nematic region (i.e. between T g and T i ). While some of the samples exhibited one or two secondary peaks in tan in the elevated region, this behavior was not consistent across all samples tested; however, all samples demonstrated tan to decrease to a near zero value once heated above T i The dependence of T g on crosslinker amount and functionality are presented in Figure 3. 5b The T g of the LCE networks increased with increasing amounts of the crosslinker in a near linear manner. At low crosslinker amounts, the T g 's of the networks are near equal; however, at higher crosslinking densities, the tetra thiol crosslink ed networks have higher T g values compared to their tri thiol counterparts. For all networks, the rubbery modulus values measured within the nematic and isotropic states increased

PAGE 48

! 29 crosslinking density. A listing of thermomechanical values from DMA and DSC of the eight LCE networks is shown in Table 3. 1. The influence of crosslinker concentrati on and functionality on the stress strain behavior is shown Figure 3. 6. Representative stress strain curves for tri thiol and tetra thiol networks can be seen in Figure 3. 6a All of the networks exhibited stress strain behavior consisting of three region s. Region One is a linear elastic deformation of the polydomain, which occurs at relatively low strain values, and the slope of the stress strain curve represents the modulus of the materials. Table 1 1. Summary of thermal analysis and thermomechanical properties of LCE. Figure 3 6 The effect of the thiol cross linker content on stress strain curve measured at 0.2 mm/s at the glass transition temperature of each LCE composition using a) tri and tetra functional crosslinkers. The failure strain was defined by the fracture of the LCE sample. b) Failure strain as a function of the cross linker content for tri and tetra thiol acrylate LCE networks. Represents significant difference with respect to equal crosslinker counterpart (p value < 0.05.

PAGE 49

! 30 Region Two undergoes the polydomain monodomain transition, which takes place at an intermediate strain where the specimens change from opaque to optically transparent. Ideally, this transition leads to a plateau in the stress strain curve with a large increase in strain a constant stress (soft elasticity) ; however, several networks only demonstrated a near plateau in which the slope of the stress curve decreased within this transition (semi soft elasticity). Region Three is continued deformation of the monodomain. This occurs once the polydomains have all been oriented and the modulus increases more sharply as the polymer chains become aligned. The Figure 3 7 The effect of the thiol cross linker content on strain actuation. Samples were relaxed in a stress free at 120 ¡C for 10 min prior to testing at 5 ¡C/min cooling rate and varying the app lied stress in each cycle. A representative plot strain actuation as a function of the thiol crosslinker; a) tri thiol crosslinker; b) tetra thiol crosslinker. The magnitude of strain actuation was measured between strain at 20 ¡C and strain at 120 ¡C. T he magnitude of strain actuation was plotted as function of applied stress for eight thiol acrylate LCE networks by arraying thiol crosslinker content; c) tri thiol cross liker; d) tetra thiol cross linker. A minimum of 3 samples (n=3) were tested at each condition.

PAGE 50

! 31 strain values differentiating these three regions highly depend on the concentration of the crosslinker. The strain in each region is shown to decrease with incr easing concentration of crosslinker. The failure strain of the networks decreased with increasing concentration of the crosslinker, while the functionality of the crosslinker only had a significant difference at 10 mol% crosslinker ( Figure 3. 6b ). As a res ult, the 10% tri thiol network had the highest mean failure strain. Polydomain LCE samples need an applied bias stress in order to orient the mesogens within the network and exhibit thermo reversible actuation. The magnitude of actuation was measured with respect to crosslinking, crosslinker functionality, and applied stress ( Figure 3. 7 ). Representative plots of tri thiol networks ( Figure 3. 7a ) and tetra thiol networks ( Figure 3. 7b ) were used to illustrate the strain actuation as samples were cooled from th e isotropic state under a constant applied stress of 100 kPa. In both networks, the strain actuation is slowly increased while cooling in the isotropic region. After passing T i, the actuation rate sharply increased until the forthcoming T g below T g the actuation strain plateaued as the polymer chains were "frozen" into place due to the vitrification of the network and reduction in chain mobility. The actuation strain increased non linearly with the decreasing the amount of the crosslinker. The depend ence of the actuation strain to the applied stress and the influence of the crosslinking density and functionality of the crosslinker is described in Figures 3. 7c and 7d. In general, strain actuation was shown to increase with increased applied stress, whi le the type of the crosslinker had no apparent effect on the strain actuation. More specifically, the magnitude of strain actuation at a given bias stress decreased with increased crosslinking. The work capacity of the LCE networks actuating under a 100 kP a bias stress is compared in Figure 3. 8 Work capacity decreased from average values of 296 to 72 kJ/m 3 with increased crosslinker concentration from 10 to 80 mol%. These values were determined by multiplying the bias stress by strain actuation values. Bec ause the bias stress is constant, the work capacity of the LCE networks is influenced by crosslinker concentration and function in the same manner as strain actuation.

PAGE 51

! 32 3.5. D iscussion The purpose of this study was to systematically investigate the impact of (i) varying crosslinking density and (ii) varying crosslinker functionality in thiol acrylate LCE systems. The network architecture influences the coupling between the liquid crystal line behavior and polymer chains, which has a pronounced effect on the thermomechanics and the actuation performance of LCEs. This study utilized the high efficiency and orthogonality of the thiol acrylate Michael addition reaction to synthesize well defin ed, uniform networks. Such a reaction offers a facile way to tailor networks with control over thermomechanical properties and actuation performance. Polydomain LCEs can be obtained via two different routes based on their synthesis and crosslinking histor y. Polydomain LCEs that are polymerized in the isotropic state at high temperatures or in the presence of a solvent are known as isotropic polydomain nematic elastomers (i PNEs). Alternatively, samples that are polymerized at temperatures below T i are kn own as nematic polydomain nematic elastomers (n PNEs). ( 51 ) The resulting two types of polydomain LCEs will display fairly similar thermomechanical properties with subtle differences in dynamic features. ( 46 61 85 ) Overall, this study focused on the relatively new method of thiol Figure 3 8 Work capacity for LCE networks as a function of crosslink ing concentration. Work capacity was measured under a constant bias stress of 100 kPa while cooling from the isotropic state.

PAGE 52

! 33 acrylate Michael addition chemistry that utilizes a solvent for ease of manufacturing, and thus focused on i PNEs. Due to the nature of this reaction, near complete conversion of thiol a crylate monomers occurs ( 54 ) Therefore, very high GF was accomplished over wide range of crosslink densities ( Figure 3. 2 ). There was a slight decrease in GF in samples with decreasing crosslinker content in both tri and tetra thiol LCE networks, which is to be expected due to the decrease in crosslinking density. It is important to note that this was achieved with commercially available materials without purification. All networks had GF values over 90%, which is excellent considering the monomers used in this s tudy had purity values of 95%. WAXS analysis is a standard procedure of determining LC structure and orientation order in the LCE systems. When the average orientation of mesogenic rods has a prefe rred direction of alignment, the intensity of Bragg scattering is azimuthally biased (peaks) around the ring ( Figure 3. 3d ). The high intensity regions in the wide angle scattering are in the plane parallel to the alignment direction, 180 ¡ apart. Azoug et a l. showed that for similar thiol acrylate LCE systems, only 80% strain is needed for a strain induced monodomain over wide range of temperature regardless of the strain rate. ( 55 ) Therefore, all of the s amples underwent WAXS analysis at 80% strain to induce an oriented monodomain. WAXS profiles revealed all of the networks to have a nematic structure. The mesogenic monomer, RM257, has also previously been shown to create nematic LCEs when used in free rad ical reactions. ( 51 86 ) The orientation parameter for this system is found to be ~ 0.80 ( see appendix A ). This value might be considered high for a nematic phase, but still lower than the value obtained by Pei, et al. for nema tic phase (0.86). It is important to note that all LCEs tested exhibited similar scattering patterns, indicating that the microstructure of the LC domains was not affected by crosslinker concentration or functionality in this system ( Figure A. 3 ).

PAGE 53

! 34 Both ther mal (DSC) and thermo mechanical (DMA) analysis was used to characterize the LCE networks. The networks displayed T g values between 3 and 25 ¡ C, which primarily depended on the crosslinker concentration. Previous studies using hydrosilylation reactions prod uced LCE networks with T g values ranging from 28 to 12 ¡ C, ( 83 ) while epoxy based reactions produced high T g materials that ranged from 46 to 65 ¡ C respectively. ( 84 ) Conversely, the T i values were on average 70 ¡ C for the tri thiol netw orks and 80 ¡ C for the tetra thiol networks. T i increased with increasing functionality of the crosslinker, but there was no appreciable effect of the amount of the crosslinker on T i This latter result is consistent with a previous study by Burke et al. t hat showed T i was not influenced by increasing the crosslink density in smectic C main chain LCEs. ( 80 ) The 80 mol% tetra thiol composition had a significantly lower T i compared to less crosslinked networks; however, it should be noted that the measurement of T i became more difficult with smal ler enthalpic wells at high crosslink densities. The significant effect of the crosslinker functionality on T i can be attributed to the localization of mesogenic monomers around the crosslinks. ( 82 ) With lower functionality crosslinkers (f=3), fewer mesogens are present around the network points. On the other ha nd, tetra thiol networks have a higher number of mesogens around the network points, which may help stabilize the LC phase and result in a higher temperature required to transition between the nematic and isotropic phase. Nevertheless, increasing the amoun t of the crosslinker leads to broadening nematic to isotropic transition and reducing the magnitude of the enthalpic wells ( Figure 3. 4b ). Networks with low crosslinking density and functionality should be less restrictive and help promote mesogen self orga nization. In this study, the 10 mol% tri thiol network had the significantly highest # H f values, which is a measur e of overall mesogenic order. The dynamic mechanical response of LCEs has been a subject to increasing investigation due to its unique viscoel astic behavior. ( 42 44 87 88 ) All of the reported LCE networks showed a distinctive drop in E' as temperature approaches T i This can be attributed to the dynamic soft

PAGE 54

! 35 elasticity induced by mesogen instability' around T i Landau theory states that mesogens can have two ene rgy minimums preferring both order and disorder at T i ( 89 ) When dynamically cycled, mesogens can rotate between either minimum to lower the energy of the system and result in a drop in E' Previous studies have shown that the dynamic stress must be in a direction to allow mesogen rotation, elsewise dynamic soft elasticity is not observed. ( 90 ) In this study, all samples tested were in the polydomain state and allowed mesogen rotation when cycled in uniaxial tension. Above T i the E' gradually recovers from the drop and became comparable to the isotropic modulus of conventional amorphous networks. Once cleared into the isotropic region, the value of the E' i s primarily dictated by the crosslink density of the networks. All networks tested in this study showed distinctive behavior in the tan loss function, described by an initial peak at T g followed by elevated values up until T i The initial peak is attrib uted to a maximum in damping cause by the viscoelasticity of the polymer chains during the glass transition, while the viscosity and rotation of the liquid crystals causes the elevated values up to T i Overall, both crosslinker density and functionality di d not have a dramatic effect on the loss tangent behavior from our LCE systems, which supports Hotta and Terentjev proposing the use of LCEs as efficient damping materials. ( 87 ) Crosslinking had a pronounced influence on the non linear stress strain behavior of the LCE networks. All networks exhibited soft or semi soft elasticity when strained, which is caused by mesogens orienting themselves along the str etching direction and gradually transforming into a monodomain. This deformation costs almost zero energy and the stress remains nearly constant while the strain increases significantly. ( 10 ) The stress strain behavior of the networks, including the modulus, breadth of soft elasticity plateau, and failure strain wer e highly dependent on the crosslinking density. As crosslinking was decreased, the modulus decreased, while the soft elasticity plateau and failure strain increased. The trends in failure strain as a function of crosslinker concentration ( Figure 3. 6b ) follows the inherent inverse relationship between failure

PAGE 55

! 36 strain as shown in other amorphous networks. ( 91 ) Safranski et al. show that the type of crosslinker used only had a significant influence on failure strain at low crosslinking densities for amorphous networks. The results of this study would suggest that while the overall stress strain behavior of LCEs are markedly different than amorphous networks, the failure strain of LCEs are influenced by the same mechanisms. F or example, at low crosslinking density, tri thiol crosslinked networks had a significantly larger failure strain compared to tetra thiol crosslinked networks. Increasing the amount of crosslinker not only affects the mechanical properties but also the loc al and global orientation. Eventually, with an increase in crosslinker, the modulus and strength of the networks will increase at the expense of failure strain and potential risk of preventing the formation of LC order. The examined LCE networks displayed actuation upon heating and cooling. The actuation relies on the reversible phase transition of the LC domains from an isotropic phase to a globally anisotropic phase (i.e. an aligned monodomain). Hence, reversible thermal actuation in polydomain LCEs is hi ghly dependent on the stress condition (i.e. the applied load) to orient the mesogens along the direction of stress. ( 80 92 ) It was shown that act uation strain increases with increasing applied stress in a non linear fashion and approaches a limiting value. The magnitude of actuation diminished from 296 to 72% with increasing crosslink density in both sets of LCE networks. The results would suggest that less entropically restrictive networks (i.e. lower crosslink density) allow for higher magnitudes of actuation; however, the actuation behavior was not influenced significantly by crosslinker functionality. This seems in contrast to the # H f measuremen ts, in which # H f increased as crosslinking was reduced but showed a significant difference in # H f between tri and tetra thiol crosslinkers at 10 mol%. Actuation is driven by the energetic contributions of the liquid crystals (enthalpy) and polymer chains (entropy elasticity). Given the 10 mol% crosslinked networks had similar network architecture, they should

PAGE 56

! 37 demonstrate the same entropic contributions during actuation. Since the 10 mol% tri thiol network had a significantly higher # H f we originally assum ed this network would yield higher magnitudes of actuation; although, the 10 mol% tri thiol networks showed the same actuation as 10 mol% tetra thiol networks despite having a significantly higher # H f We are currently investigating the actuation behavior of additional LCE networks with a higher and broader range of enthalpy values to further test our o riginal assumption. The stress driven actuation and work capacity of our thiol acrylate LCEs can be compared to other studies of smectic main chain LCEs. Li et al. and Burke et al. measured actuation for an epoxy based and hydrosilylation based LCEs as a function of bias stress. ( 80 84 ) With a bias stress of 100 kPa, the actuation values of the two types of LCEs ranged between 63.2 78.7% and 18 21%, respectively. It should be noted that these values were for LCE networks with the lowest crosslink densities in these studies. By comparison, the low crosslinked LCEs in this study showed a mean actuation of approximately 296% and work capacity of 296 kJ/m 3 at 100 kPa. Similar reversible actuation under uniaxial load showed work capacity abou t 96.9 kJ/m 3 utilizing a two step method of aza Michael addition and photo polymerization reaction. ( 93 ) Ware et al. measured work capacity for a +1 defect to be 3.6 kJ/m 3 using similar reaction. ( 60 ) Mammalian skeletal muscle ranges from 8 40 kJ/m 3 ( 94 ) while other types of actuators under isotonic loading such as piezoelectric ceramics and electro resistive polymers can have work capacities of 640 and 1250 kJ/m 3 respectively. ( 95 ) One of the most interesting behaviors of LCEs is their ability to actuate. The temperature range for actuation in LCEs is bounded between T g and T i The desirable range of values is application dependent. For example, low T g and high T i will be ideal for applications such as soft robots and sensors, where low T g and low T i will be ideal for biomedical applications such as a soft expanding stents. ( 96 ) This study suggests that the thiol acrylate Michael addition reaction may

PAGE 57

! 38 serve as a good platform to tailor main chain LCEs with enhanced actuation behavior. 3.6. C onclusions Well defined nematic, polydomain main chain LCE networks were synthesized u sing a thiol acrylate Michael addition reaction using both tri thiol and tetra thiol crosslinkers. All networks had gel fraction values greater than 90% using commercially available starting materials. The 1 D WAXS patterns were not affected by crosslinkin g concentration or functionality and showed a nematic structure when strain to 80%. Crosslinker functionality showed a significant influence on the T i of the networks. The average T i for the tri thiol networks was 70 ¡ C, while the average T i for the tetra thiol networks was 80 ¡ C. Increased crosslinking was shown to reduce # H f of the polydomain to isotropic transition. An increase in crosslinking density was shown to increase T g from 5 to 17 ¡ C and 3 to 25 ¡ C in tri thiol and tetra thiol cross linked networks, respectively. Crosslinker density had an inverse relationship with failure strain, while crosslinker functionality only had a significant influence at the lowest degree of crosslinking. The average failure strain increased from 542 to 853% from tetra thiol to tri thiol networks at 10 mol%, respectively. Crosslinker functionality did not influence thermal actuation behavior, whereas an increase in crosslinking reduced the magnitude of actuation. Networks showed a decrease in work capacity fr om 296 to 72 kJ/m 3 from 10 to 80 mol%. 3.6. A cknowledgements NSF CAREER Award CMMI 1350436 and the Soft Materials Research Center under NSF MRSEC Grant DMR 1420736 supported this work

PAGE 58

! 39 CHAPTER IV MODULATED MESOPHASE LIQUID CRYSTAL ELASTOMERS 4.1. Abstract Control of the mesophase in liquid crystalline elastomers (LCEs) is a critical aspect in harnessing their unique stimuli responsive properties. Few studies have compared nematic and smectic main chain LCEs in a direct way. Traditionally, it is believed tha t the mesogen core and synthetic route determines the phase behavior. In this study, we hypothesized that tuning the LC phases in main chain LCE systems can be achieved by varying the spacer length while maintaining the same mesogen (RM257). By increasing the length of dithiol alkyl spacers containing two to eleven carbons along the spacer backbone (C2 to C11), we can modulate the mesophase from nematic to smectic, tailor the nematic to isotropic transition temperature between 90 and 140 ¡ C, and increase the average work capacity from 128 to 262 kJ/m 3 Phase segregation and the smectic C phase is achieved at room temperature for the C6, C9, and C11 spacers. Upon heating, these samples transition into the nematic and later, the isotropic phase. Furthermore, th is segregation occurs along with polymer chain crystallinity, which increasing the modulus of the networks by an order of magnitude; however, the crystallization rate is highly time dependent on the spacer length and can vary between 5 minutes for the C11 spacer and 24 hours for shorter spacers. This study illuminates several possibilities of the TAMAP reaction in modulation of the thermomechanical and liquid crystalline properties of LCEs and discusses their potential u se for biomedical applications. 4.2 Introduction The extensive use of liquid crystalline (LC) materials in modern technologies strongly relies on the modulation of their thermotropic behavior such as mesophase structure, phase transition

PAGE 59

! 40 temperatures, and mesophase stability. In many applica tions, these materials are used as active mechanical or optical components. ( 8 10 33 ) The external stimuli induced mechanical responsiveness phenomenon has attracted a lot of research attention and has been one of the most prolific frontier research areas in materials science recently. ( 97 99 ) Among them, liquid crystal elastomers (LCEs) are becoming an increasingly strong competitor in the development of a new generation of actuators because they provide advan tages such as flexibility, large actuation strain, light weight, and tailorability. ( 100 ) Such properties make them suitable for many for potential technological applications such as artificial muscles, sensors, and soft robotics. ( 33 10 1 ) LCEs have already been demonstrated in many applications such as micro grippers for robotics, ( 102 ) micro electromechani cal systems (MEMS) optical grating devices, ( 103 104 ) tunable apertures, ( 105 ) and microfluidic systems. ( 106 ) Actuation in LCEs is based on the unique combination of LC order, network elasticity, and chain mobility. ( 62 107 ) The mobil ity and elasticity enable these supra molecular systems to respond to different types of external stimuli such as heat or light. The order in these systems can be obtained by maximizing the interactions and minimizing the excluded volume between the LC mes ogens, which gives rise to the mechanical anisotropy. ( 62 ) The ordered LC phase state exists in a temperature range in between the solid crystalline and the isotropic disordered liqui d state. Three common three types of LC phases are: Nematic (N) mesogens are oriented in a uniform direction along a director (orientational order); Smectic A (SmA) a layered structure with orientational and positional order; and Smectic C (SmC) simi lar to SmA but with the mesogens titled with respect to the director. ( 34 ) In gen eral, smectic LCEs have larger actuation, lower failure strain, higher modulus, and greater enthalpy compared to nematic LCEs. This is due to smectic LCEs typically having higher order parameters compared to nematic LCEs. ( 35 ) A general belief is that the formation of LC phases in elastomers is dictated by the mesogen structure: the mesogen core and flexible tails. ( 6 34 36 37 ) Nem atic mesogens should yield

PAGE 60

! 41 nematic LCE systems, and smectic mesogens should correspond to smectic LCEs. ( 38 ) Krause et al. prepared numerous smectic C and nematic LCE systems with a variety of thermomechanical properties by modifying the mesogen structure; ( 37 ) however, this traditional method of tuning LC phase is very challenging and not practi cal for many research groups without extensive chemistry backgrounds. For example, acrylate functionalized mesogens such as RM82, RM257, and 6OBA have been used to prepare smectic A and C (6OBA) and nematic LCEs (RM82, RM257). ( 108 ) This difference in LC phase is typically attributed to having different mesogen core structures (6OBA vs. RM82, RM257). Interestingly, we have observed in the literature that smectic and nematic main chain LCEs can be synthesized from the same mesogen cores. For example, mesogens with the same core structure such as 5Me, RM82, and RM257 have been used to prepare smectic and nematic main chain LCEs. The Mather research group utilized 5Me as a mesogen to prepare smectic C LCEs, ( 80 109 ) while mesogens such as RM82 and RM257 have been employed by others to prepare nematic LCEs. ( 55 60 78 ) 5Me was prepared using a hydrosilylation reaction compared to RM82 and RM257, potentially suggesting that the type of reaction is dictating the LC phase. Looking at Torbati et al. and Sae d et al. both used the same mesogen cores, same crosslinkers, and similar thiol ene/acrylate reactions schemes to produce smectic C and nematic LCEs, respectively. ( 109 110 ) The notable difference was the longer spacer employed by Torbati. This suggests that the formation of the LC phase is highly dictated by not only mesogenic structure but also the spacing between mesogens. Thereby, we hypothesize that modulating the LC ph ases in main chain LCE systems could be achieved by varying the spacer length while maintaining the same mesogen. This suggests that a variety of LC phases could be made from similar mesogen structures. Having the ability to tune the LC phase by simply adj usting the spacer opens the door for the first time to further tune the LCE systems with exotic physical properties in a facile manner. The purpose of this study is to modulate LC phase structure using a single nematic mesogen,

PAGE 61

! 42 RM257, by controlling the thiol spacer length. Herein, we report series of main chain LCE systems that are capable of multiple LC phases at room temperature. Two desired LC phases were realized, smectic C and nematic. Smectic C phases were observed when using longer thiol spacers, as these spacers drive a nano scale segregation of ternary incompatible layers of bulk alkyl thiol functionalized spacers, flexible propylene oxide acrylic terminal chains, and mesogen cores. Segregation of these distinct segments is the main contribution to the formation of the smectic phase. ( 76 ) Shorter thiol spacers did not segregate; therefore, they e ngendered a nematic phase. Furthermore, the nematic to isotropic transition temperature (T NI ) was highly dependent on the spacer length. As an unexpected result, the use of alkyl spacers resulted in semi crystallinity within the LCE networks. This semi cry stallinity improved the mechanical properties of the elastomers; however, the rate of crystallization was highly dependent on the length of the spacers. This approach allows us to compare the thermomechanical properties of smectic C and nematic LCE actuato rs with near identical chemical compositions, which has traditionally been extremely difficult to achieve due to vastly different reactions and compositions used in preparation. 4.3. Results and Discussion Main chain LCE networks were synthesized via a thiol acrylate Michael addition "click" reaction using a di acrylate nematic mesogen (RM257), thiol functionalized alkyl spacers (C n ), and a tetra thiol crosslinker (PETMP). The reaction was catalyzed using a nucleophilic catalyst (DPA) at 60 ¡ C in the pres ence of toluene; therefore, the crosslink history of the networks was established in the isotropic state. The polydomain state formed after the removal of toluene. The thiol acrylate chemistry was chosen due to the wide availability of thiol and acrylate m onomers and mesogens as well as the ability to react using a Michael addition reactions. ( 53 ) This synthesis methodology has previously been reported. ( 55 111 ) In a study on the influence of the crosslinker, it was found that work capacity for similar networks increased with decreasing crosslinking density; ( 110 ) therefore, 2.5 mol% crosslinker (or 10 mol% of the thiol functional groups) was used for the

PAGE 62

! 43 networks presented herein. The nature of the "click" reaction provides a facile manner to tailor the network structure by simply adjusting the spacer length. Five compositions were synthesized with the total number of carbons in the spacer backbone ranging from 2 to 11 (labeled as C2, C 3, C6, C9, and C11). The overall reaction scheme and study hypothesis are illustrated in Figure 4. 1. Figure 4 1 Schematic of main chain LCE synthesis via a thiol acrylate Michael addition reaction. During synthesis, mesogens are in the isotropic (Iso) phase in a present of solvent (toluene) at 60 ¡ C. The toluene is removed after the reaction is completed to form polydomain samples. The formation of nematic (N) and smectic C (SmC) domains form via pha se separation of mesogen and thiol spacer. Differential scanning calorimetry (DSC) was used to initially characterize the phase transitions in the LCE networks ( Figure 4. 2a ). Spacer length had a significant influence on the isotropic transition. T NI decre ased from 140 ¡ C to 90 ¡ C by increasing the spacer length from C2 to C11. Furthermore, the heat of fusion ( H f ) associated with this transition increased with spacer length. Two interesting phenomenon were observed as the spacer length increased beyond C6. F irst, an additional endothermic well appeared in the heat flows. This suggested that a second smectic phase could be present in the materials. Second, the presence of polymer chain crystallinity was observed. Physically, these samples demonstrated a substa ntial increase in modulus over the course of 24 hours. Comparing the heat flows of the first and second heat cycles showed the

PAGE 63

! 44 samples could be thermally reset ( Figure 4. 2b ). The presence of a crystallization exotherm can be seen followed by two endothermi c wells for the C9 sample. Additional DSC experi ments can be found in appendix B The five LCE networks were heated to the isotropic state and allowed to cool slowly to demonstrate T NI can be highly tailored using a single mesogen ( Figure 4. 2c ). Wide a ngle X ray scattering (WAXS) was used next to investigate the influence of spacer length on the mesophase of the materials ( Figure 4. 3 ). Unstretched, polydomain samples all exhibited diffuse halos at approximately ~1.8 A 1 which is characteristic of unali gned LC domains. Both C9 and C11 samples also clearly exhibited inner halos at ~0.24 A 1 which is indicative of smaller scale order that can be associated with a smectic C phase. Next, an aligned monodomain Figure 4 2 (a) Heat flows of five LCE networks with increasing spacer length from C2 to C11. Heat flows are shown on the second heating to reset the thermal history of the networks. (b) Comparison of first and second heating cycles of the C9 network. The second heat ing cycle shows an exotherm once heated above the glass transition temperature due to polymer chain crystallization. (c) Optical images showing isotropic (transparent) to polydomain (opaque) transition of the five networks as a function of cooling

PAGE 64

! 45 was created by applying a 100% engineering strai n to the samples. The X ray diffraction patterns evolved to show two bright spots separated by 180 ¡ and transverse to the direction of stretching. The C2 and C3 networks demonstrated characteristics of a nematic monodomain, while the C6 to C11 networks rev ealed four inner bright spots each separated by 90 ¡ that is characteristic of a smectic C monodomain. 1D diffraction plots for each network can be found in A ppendix 1. Figure 4 4 2 D WAXS patterns for five LCE networks at room temperature. Diffraction was measured in (a) an unaligned polydomain state and (b) an aligned monodomain state. Alignment was achieved by stretching the samples to 100% engineering strain before analysis. To further explore the mesophase behavior of these systems, the diffraction patterns of the C9 system were measured as a function of temperature ( Figure 4. 4 ). It is important to note, these images were taken on a stretched sample that had 24 hrs since its last thermal cycle to equilibrate before heating and analysis began. The 2 D WAXS showed a SmC to N transition completed by 85 ¡ C and a N to iso transition completed by 120 ¡ C. The DSC heat flows corroborate these findings by exhibiting endothermic wells at 75 ¡ C and 99 ¡ C. Both C6 and C11 systems show the same trend in mesophase behavior as a function of temperature and are shown in the supplement.

PAGE 65

! 46 Figure 3 4 Temperature controlled WAXS analysis of the LCE system using the C9 spacer. Diffraction patterns reveal the transition from a smectic C to nematic orientation when heated above 80 ¡ C, while a nematic to isotropic transition occurs when heated above 100 ¡ C. All images were taken under 100% engineering strain. This behavior can be attributed to the molecular structure containing flexible alkyl spacers, propylene oxide acrylic terminal chains, and rod like mesogenic cores, which can be classified as a rod coil tri block system. With increased spacer length, this system can lead to phase separation of the incompatible segments and form self assembled layered structures similar to ternary amphiphiles systems in ABC tri block copolymers. ( 76 112 ) The ordered periodic structures construct due to the mutual repulsion interaction forces of the dissimilar chemicals components o f those segments and their packing constraints. ( 112 ) The LC assemblies of rod coil tri block systems may provide a facil e means to modulate and influence the LC phase morphology. In the case of using shorter thiol spacers (C2 and C3), the resulting elastomers only exhibited a nematic phase. This can be attributed to the fact that short spacers coupled with the flexible prop ylene oxide acrylic terminal chains and form miscible new chains that are not amphiphilic, which can prevent phase separation and lead to a single N phase. The smectic C phase is a highly ordered LC phase; therefore, any sample that displayed the smectic C phase at room temperature (C6 C11) was expected to transition through the nematic phase before reaching an isotropic state. In contrast, C2 and C3 are not expected to show a SmC N phase transition as a function of heating.

PAGE 66

! 47 It is important to note the influence of time dependence in these samples when analyzing both DSC and WAXS data. For WAXS analysis, the nematic phase was not immediately detected in the C9 system if the sample was cooled from the isotropic state; therefore, our analysis was performed on a stretched sample held at room temperature for 24 hours. These results agree with our DSC analysis. During the cooling scan, the SmC N transition was not detected and only showed up during heating scans. On the other hand, T NI was spotted in both heat ing and cooling scans with slight differences in temperature, which is due to instrument hysteresis and differences in the heating and cooling rates. During cooling, the smectic C phase is strongly depressed and did not appear on the heat flow traces; howe ver, polymer chain crystallization and smectic C transitions were seen on the second heating. This behavior is due to a kinetic dependence of the phase separation. These LCE networks undergo polymer chain crystallization transition similar to Gelebart et a l. 's systems. ( 78 ) The onset of the cr ystallization melting temperature can be revealed in the first heating scan and vanished during the cooling scan and the sec ond heating scan for C9 sample. Table 4 1 Summary of DSC and WAXS data for 5 LCE systems tested. Each data point represents n=3.All of the samples contained equal amount of crosslinker. T c was measured during the 1 st heating scan, where as T SmC T NI and H f were measured during the 2 nd heating scan. The d spacing valves were calculated from the ID plots see the supporting information for more details. The evolution of thermo mechanical properties as a function of spacer length and polymer chain crystallization was studied using dynami c mechanical analysis (DMA). Storage modulus ( E # ) and

PAGE 67

! 48 loss tangent ( tan ) traces for the C6, C9, and C11 are shown in Figure 4. 5. Overall, the samples displayed distinct aspects of polydomain LCE thermo mechanical behavior, such as having a broad elevated tan curve between the glass transition temperature (T g ) and T NI as well as a decrease in modulus at T NI described as dynamic soft elasticity. ( 55 57 ) More interestingly, all of the samples exhibited atypical behavior during the first heating scans, which was characterized by maintaining a high modulus (~100 MPa) when heated ab ove T g This is in contrast to a sharp drop in modulus at the onset of the glass transition typically shown previously. ( 110 ) For shorter spacer lengths (C2 to C6), the second heating demonstrated this more typical response with a sharp drop in modulus ( Figure 4. 5a ). C2 to C6 samples did not show any significant changes in thermo mechanical behavior when allowi ng the sample to anneal for up to 120 min, indicating a slow crystallization rate (i.e. over 24 hours). The C9 spacer demonstrated the most unique behavior, as the annealed samples underwent recrystallization during testing. This is illustrated by an incre ase in modulus as the samples were heated above T g Each testing condition converged to the same values by 70 ¡ C, which is just below the SmC N transition for this system. For the longest spacer, C11 samples did not exhibit any differences in behavior betwe en tests and annealing times ( Figure 4. 5c ). This suggests that 5 minutes is enough to fully recrystallize the sample due to fast crystallization kinetics.

PAGE 68

! 49 A summary and comparison of thermo mechanical parameters between the first and s econd heating after 5 minutes is shown in Table 4. 2 These data help define the difference between semi crystallinity within the elastomers. C2 to C9 systems all showed an order of magnitude drop in nematic modulus E n from the first to second heating. E n of C11 systems did not change during the two tests due to its fast crystallization kinetics. Increasing the spacer length reduced T g in the networks due to increasing the polymer chain mobility in longer spacers. The C2 system was the only system to have a T g greater than ambient temperatures, and therefore may be considered a liquid crystalline glassy network instead of elastomer. Spacer length showed no appreciable Figure 4 5 Storage modulus (E#) and loss tangent (tan delta) traces for LCE networks with spacer lengths of C6, C9, and C11. Samples were measured at 3¡C/min heating rate and 1 Hz frequency in tension. All samples were annealed above T NI and allowed to cool at room temperature for 24 hours before the first temperature sweep to allow the semi crystallinity to fully form. Samples were tested four times and allowed to set isothermally at 25 ¡C between each sweep for 5, 60, and 120 minutes to show the evolution of the me chanical properties due to polymer chain crystallization. The behavior of the C2 and C3 networks closely resembled that of the C6 and are thus are only shown in appendix 3

PAGE 69

! 50 differences in the isotropic modulus values, which was to be expected when using a relativ ely low and constant amount of crosslinking between systems. These systems offer a unique opportunity to control the thermo mechanical properties of LCE networks, especially with respect to the trade offs in actuator materials. Most materials that show two way ac tuation either provide high modulus values with small strain actuation (i.e. such NiTi alloys or piezo ceramics) or provide larger strain actuation with very low modulus values (i.e. LCEs or dielectrics). ( 33 113 115 ) When it comes to the interplay between actuators and mechanical properties, the stiffer the better is a traditional premise of a good design, as stiffness improves the precision, stability, and bandwidth of position control of the device. ( 116 ) The use of LCE actuators for applications at room temperature, which typically near or above their T g is of particular importa nce for the sustained interest in their field of research. Traditional LCE materials undergo high strain actuation (around 400%) but the modulus in its deployed state typically in the range of 0.1 10 MPa. ( 8 110 ) This behavior implies that a limitation of LCE actuators is their lack of mechanical strength and modulus after being deployed. In contrast to LCEs, the modulus of NiTi alloy actuators is typically around 60 GPa, but this high modulus is at the expense of the strain actuation (aroun d 8%). ( 117 118 ) Here, we present semi crystalline LCE systems that exhibit an initial set of distinct mechanical properties suitable for deployment of Table 4 2. Dynamic Mechanical Analysis (DMA) behavior for the first and second temperature sweep; the glass transition temperature (T g ) was measured at the peak of tan !; (E n ) is the storage modulus measured at 25 ¡C; where the rubbery modulus (E r ) was measured the isotropic temperature Ti + 30 ¡C. The first temperature sweep was performed after being stored for at least 24 hours at room temperature; whereas the second temperature sweep was pe rformed 5 minutes after the first sweep was completed

PAGE 70

! 51 actuator devices and a second set of properties with an order of magnitude increase in modulus via self stiffening. This is achieved by the formation of a semi crystalline structure with variable re crystallization rates. Other researchers have studied the increased potential for LCEs using dynamic self stiffening, achieved by cycling the sample to align the mesogens. ( 119 ) Enhancing the stiffness without compromising the strain actuation in inherently weak materials may lead to a new paradigm of designing the next g eneration of LCE actuators, which is possible by engineering the microstructure of the LCE networks T NI increased with decreased spacer length due to an intense localization of mesogenic monomers in the networks with short spacers. Hence, a higher tempera ture (more energy) was required to enable the phase change. This agrees with the results of our previous study that showed an increase in crosslinking functionality, which localized more mesogens around crosslinking netpoints, increased the T NI of the netw orks. ( 110 ) On the other hand, the smectic to nematic transition temperature (T SN ) showed an opposite trend as it decreased with decreasing the spacer length until completely disappearing at shorter spacers (C2 and 3). As stated earlier, the formation of the smectic C phase in this system is due to formation of layers structure and the stability of the layers a nd SmC phase is proportional to the length of the spacer. Thus, longer spacer required higher temperature (more energy) to disrupt the layers and enable the phase transition. To the best of our knowledge, there is no other similar studies demonstrating these relationships between LC phase transition temperatures, LC phase structure, and spacer length using a single type of reaction and mesogen.

PAGE 71

! 52 The actuation performance of the five LCE networks was then measured as a function of temperature ( Figure 4. 6a ). Samples were equilibrated in the isotropic state before testing began, which served to erase any thermal history and effects of polymer crystallinity at the beginning of the test. The magnitude of actuation strain increased with spacer length. Furthermore, the sharpness and rate of the actuation increased with spacer length. This is due to the range of temperatures decreases for longer spacers, as the T g values relatively remained the same for all five networks while T NI decreased with increased spacer length. The average actuation strain increased from 255 to 525%, which corresponding to volumetric work capacity under 50 kPa bias stress to increase from 128 to 262 kJ/m 3 ( Figure 4. 6 Figure 4 6 (a) Selected actuation plots of five LCE networks with increasing spacer length from C2 to C11 under a 50 kPa bias stress. Samples were equilibrated above T NI and cooled at a rate of 5 ¡ C/min. (b) Average work capacity for each network (n=3). Work capacity was calculated by multiplying the bias stress by the actuation strain A stent model was selected to demonstrate the multi functional potential of these semi crystalline main chain LCE samples. It should be noted a C9 LCE network was synthesized with 15 m ol% excess acrylate groups to allow a second stage photo polymerization reaction to program a monodomain in the stent following a previous method. ( 55 111 ) This was to avoid the need of a bias stress during the demonstration and allow hands free actuation. This study showed that the T g of the networks could be tailored above r oom temperature and below body temperature (36 ¡ C) by

PAGE 72

! 53 utilizing shorter spacer lengths. Furthermore, our previous study showed T g can be further refined by the amount of crosslinking within the networks. ( 110 ) This suggests that these materials can utilize the shape memory effect around the glass transition to demonstrate a 1 way shape change. Figure 4. 7a shows a stent deploy from a packaged shape similar to a previous study on shape memory polymer (SMP) stents. ( 120 ) The Mather research group has also investigated the shape memory effect in LCEs to highlight the relevance of combining both phenomenon. Once deployed, the LCE stent can show reversible 2 way actuation by heating and cooling around T NI ( Figure 4. 7b ). During this process, the diameter of the stent reduces by approximately 40% whe n heated and returns to its original shape when cooled. Bellin et al. originally showed a tube like stent that was capable of first expanding then contracting using a two step programming process of SMPs; ( 121 ) however, this method was not reversible and relied on two transition temperatures instead of a single T NI The ability to repeatedly and reversible switch the diam eter of a stent could potentially solve several challenges in cardiovascular intervention, such as the need to reposition a device, remove a device, or adjust the device as the patient grows during adolescence. Lastly, shape changing polymers, and especial ly elastomers, are inherently more compliant than their metallic counterparts. Depending on the application, they can lack the necessary rigidity to provide structural support, such as a stent resisting restenosis after balloon angioplasty. This limitation in shape memory biomedical applications was previously discussed by Nair et al. ( 122 ) Figure 4. 7de demonstrates how the LCE networks with semi crystallinity can provide increased mechanical support. Our data shows LCE networks can undergo self stiffening and the rate of chain crystallization can be controlled from between 5 minutes to 24 hours.

PAGE 73

! 54 Figure 4 7 Photo sequences highlighting multiple functionalities capable within these semi crystalline LCE networks. A C9 stent was synthesize with 15 mol% excess acrylate groups. The 9 mm stent was expanded to 15 mm and photo crosslinked to lock in mesogen orientation. (a b) The LCE stent is capable of demonstrating a 1 way shape memory effect when heated above its glass transition. (b c) The LCE stent is also capable of reversible 2 way actuation when heated and cooled around its T NI (d) If the expanded stent is allowed time to develop polymer crystallinity, it is capable of supporting a 100 g weight, compared to (e) an uncrystallized stent. 4.4. Conclusions A series of main chain LCE networks with tunable LC mesophases were synthesized using a thiol acrylate Michael addition reaction using a s ingle nematic mesogen, RM257. The LCE networks exhibited different mesophases by controlling the length of thiol functionalized alkyl spacers (C2, C3, C6, C9, and C11). Longer spacers (C6, C9, and C11) drive nano scale segregation resulting in smectic C ph ases, whereas shorter spacers (C2 and C3) resulted in nematic phases only as confirmed by the 2D WAXS patterns. The length of the spacers showed a significant influence on the thermomechanical properties of the networks such as T NI H f and T g T NI decrea sed from 140 to 90 ¡ C and H f increased significantly with increasing spacer length. The networks exhibited semi crystallinity the rate of crystallization significantly influenced by the spacer length. Shorter spacers (C2, C3, and C6) displayed a slow recry stallization rate after being annealed, whereas C9 recrystallized within 2 to 3 hours and C11 samples only needed 5 minutes to fully recrystallize due to increased spacer flexibility. Smectic C networks demonstrated larger

PAGE 74

! 55 magnitudes of actuation compared to nematic networks, with the C11 network showing an average actuation of 525% 57%. Networks showed an increase in work capacity from 128 to 262 kJ/m 3 C2 to C11. 4.5. Experimental Section 4.5.1 Materials : Pentaerythritol tetrakis( 3 mercaptopropionate) (PETMP), 1,2 Ethanedithiol (C2), 1,3 Propanedithiol (C3), 1,6 Hexanedithiol (C6), 1,9 Nonanedithiol (C9), 1,11 Undecanedithiol (C11) dipropylamine (DPA), and toluene were purchased from Sigma Aldrich. 4 bis [4 (3 acryloyloxypropypro pyloxy) benzoyloxy] 2 methylbenzene (RM257) was obtained from Wilshire Technologies, Inc. (Princeton, NJ, USA). The chemical structures of the monomers and catalyst are shown in Figure 4. 1 All materials were used in their as received condition without fur ther purification. 4.5.1. Synthesis of Liquid Crystalline Elastomers : LCE samples were synthesized via a thiol acrylate Michael addition reaction. LCE networks were prepared starting with two thiol monomers. The thiol monomers were selected for their use as a tetra functional cross linking monomer and di functional spacer between mesogens. The crosslinker (PETMP) was mixed with only one spacer monomer at a time, C2, C3, C6, C9, or C11. The ratio of thiol crosslinker to spacer was kept constant in all of sa mples, with 10 mol% of functional groups belonging to the cross linker and 90 mol% belonging to the spacer. Thiol solutions were added to the diacrylate mesogen, RM257, in a stoichiometric balance unless otherwise stated, which was dissolved in 40 wt% of t oluene at 80¡C for 5 min prior to the addition of the thiol solution. Once the solution returned to room temperature, 1 mol % of DPA was added to catalyze the reaction. The solution was mixed vigorously using a Vortex mixer (No: 94540, Toronto, ON, Canada) Air bubbles were removed from the solution under a 500 mm # Hg vacuum. The solution was then injected between two glass slides separated with 1 mm spacers and left to cure at 60¡C overnight. After the

PAGE 75

! 56 polymerization was completed, the samples were placed i n an oven for 24 h at 80¡C under a 500 mm # Hg vacuum to remove the solvent. 4.5.2. X Ray Diffraction : In order to investigate the LC mesophases in the networks, X ray diffraction was performed using Forvis Technologies wide angle X ray scattering (WAXS) 30W Xenocs Genix 3D X ray source (Cu anode, wavelength = 1.54 ) and Dectris Eiger R 1M detector. The beam size was 0.8 mm x 0.8 mm, and the data was collected at a sample to detector distance of 197 mm. the sample was expose to the X ray for 15 min. The flu x was 4x10 7 X rays/s. The scattering patterns were analyzed and plotted using intensity versus azimuthal angle by Rigaku SAXSgui and Igor Pro software to determine the d spacing of LCEs using the Bragg's equation below: ! !"# (1) where % is the X ray radiation wavelength (1.5405 ), d is the spacing between long range ordering of mesogens in LCE network, and & is the scattering angle. All of samples were annealed above their isotropic transition temperature and allowed to cool at room te mperature for 24 hr prior testing. Data was gathered for all of samples (C2, 3, 6, 9, and C11) at room temperature, stretch at 0 and 100% strain to identify the crystal structure for both polydomain and monodomain, respectively. Further investigation of th e influence of the temperature on the nanostructure was done on the samples that exhibited smectic C phase at room temperature (C6, C9 and C11). Samples were tested, while heating at 30, 50, 80, 85, 90, 100, and 120 ¡C. 4.5.3. Differential Scanning Calorim etry (DSC): DSC was performed using a TA Instruments Q2000 machine (New Castle, DE, USA). Samples with a mass of approximately 10 mg were loaded into a standard aluminum DSC pan. All of samples were annealed above their isotropic transition temperature and allowed to cool at room temperature for 24 hr prior testing. The samples were equilibrated at 50 ¡C and heated rapidly to the isotropic state (T Ni + 30 ¡C) at 10¡C/min, to measure the melting transition temperature (T m ) and cooled slowly to 50¡C at a r ate

PAGE 76

! 57 of 2¡C/min to allow LC self assembly. Samples were then heated more rapidly to the isotropic state (T Ni + 30 ¡C) at a rate of 20¡C/min. The LC phase transition temperatures (T SN and T NI ) were measured at the second heating scan, and defined as the minimum value of the first and the second endothermic peak, respectively The reported enthalpy ( H f ) change is measured by integrating the endothermic energy well corresponding to the transition from the SmC polydomain to N state and enthalpy and from the N polydomain to isotropic state. 4.5.4. Dynamic Mechanical Analysis (DMA): DMA was performed using a TA Instruments Q800 machine (New Castle, DE, USA). Rectangular samples measuring approximately 20 x 5 the isotropic state (T NI + 30 ¡C) x 0.8 mm 3 were te sted in tensile mode, with the active length measuring approximately 6 to 8 mm. Samples were cycled at 0.2% strain at 1 Hz and heated from 50 to the isotropic state (T NI + 30 ¡C) at a rate of 3¡C/min. All of samples were annealed above their isotropic tra nsition temperature and allowed to cool at room temperature for 24 hr prior testing Samples were temperature swept four times and allowed to set isothermally at 25¡C between each sweep for 5, 60, and 120 minutes to show elevation of the mechanical propert ies. T g was measured at the second temperature sweep and defined as the temperature corresponding the peak of tan curve. The LC modulus ( E' n ) and isotropic rubbery modulus ( E' r ) were measured using the storage modulus values at 10 ¡C and isotropic state (T NI + 30 ¡C) respectively. 4.5.5 Strain Actuation Characterization: Strain actuation was measured using the Q800 machine. Sample ends were wrapped with aluminum foil and loaded in the DMA machine in tensile mode with an active leng th equal to 4.2 mm. The cross sectional areas of the samples measured 1 x 5 mm 2 Samples were equilibrated at the isotropic state (T NI + 30 ¡C). A constant bias stress ( bias = 50 kPa ) was then applied to the samples, while the samples were heated and cooled between the isotropic state and 50 ¡ C at 5 ¡ C/minute. The maximum of actuation ( a ) was defined by measuring the different between minimum and maximum engineering strain values measured at T NI + 30¡C and 50 ¡ C, respectively. Although, C9 and C11 samp les were maximized

PAGE 77

! 58 the DMA's length limit (25 mm) using a gage length of 4 mm. Therefore, only 2mm was allowed as an active gage length between the clamps by coating the rest of the sample with super glue and covering with Teflon tape such that the part wa s coated with super glue would not actuate. The estimated volumetric work capacity of the networks was measured by multiplying the actuation strain by the applied bias stress (Eq. 2). !"#$ !"#"$%&' ! !"#$ !"# !"#$ !" ! !"#$ ( 2) The maximum bias stress was selected to be 50 kPa to suit all the tested samples. Stresses greater than 50 kPa frequently caused greater actuation but fracture at the sample grip interface at elevated temperatures for samples with lon ger spacer length (C9 and C11). 4.5.6 Stent Fabrication: A stent was synthesized and programed using a two stage thiol acrylate Michael addition and photopolymerization (TAMAP) reaction described in our previous report. ( 55 ) Immediately after adding catalyst, he reaction was added to a two piece cylindrical glass mold consisting of an inner core (diameter = 8mm) and an outer shell (inner diameter = 12mm) with a height of 6cm. After letting polymerize for 24 hours at 60 o C, to luene was evaporated from the sample at 60 o C overnight. Then, the sample was stretched radially to a diameter of 16mm and photopolymerized for about 30 minutes, rotating the light source 90 degrees every 7 minutes to evenly cure the sample. 4.6. Acknowled gements NSF CAREER Award CMMI 1350436 and the Soft Materials Research Center under NSF MRSEC Grant DMR 1420736 supported this work

PAGE 78

! 59 CHAPTER V TAILORABLE AND PROGRAMMABLE LIQUID CRYSTALLINE ELASTOMERS USING A TWO STAGE THIOL ACRYLATE REACTION 5.1. Main This study introduces an unexplored method to synthesize and program liquid crystalline elastomers (LCEs) based on a two stage thiol acrylate Michael addition and photo polymerization (TAMAP) reaction. This methodology can be used to program permanently aligned monodomain samples capable of "hands free" shape switching as well as offer spatio temporal control over liquid crystalline behaviour. LCE networks were shown to have a cytocompatible response at both stages of the reaction. Liquid crystalline elastomers ( LCEs) are a class of smart materials that can exhibit reversible mechanical and optical functionalities. These materials incorporate self organizing mesogenic structures into an elastomeric network to combine the properties of entropy elasticity and liquid crystalline behaviour. Researchers have proposed LCEs for mechanical actuators, ( 33 ) artificial muscles, ( 18 68 ) and switchable surfaces; ( 123 ) however, to enable actuation within the material, the mesogens must first be aligned uniformly, creating a liquid crystalline monodomain (often referred to as a liquid single crystal elastomer). ( 6 ) The vast majority of m ain chain LCEs are synthesized via hydrosilylation reactions, based on a method established by Bergmann and Finkelmann. ( 124 ) A multi step approach is often used to achieve a monodomain in main chain LCEs: the reaction is allowed to proceed to gelation, a sample is me chanically stretched to align the mesogens, and the reaction to proceeds to crosslink and stabilize the monodomain. ( 125 ) Other methods to produce a stabilized monodomain include using surface alignment techniques or magneti c fields to keep

PAGE 79

! 60 mesogens aligned during synthesis via free radical polymerizations of acrylate or thiol ene functionalized mesogens; ( 126 127 ) however, these techniques have been limited to thin films or micro geometri es. ( 123 ) As a simple, readily accessible, powerful methodology, we introduce a previously unexplored approach to synthesize and program main chain LCEs using a two stage thiol acrylate Michael addition and photopolymerization (TAMAP) reaction. Initial polydomain LC E samples can be formed using a thiol acrylate "click" reaction with the facile ability to tailor the crosslinking density and polymer structure. If an excess of acrylate groups exists, a Figure 5.1. (a) A diacrylate mesogen (RM257), dithiol flexible spacer (2,2' (ethylenedioxy) diethanethiol EDDET), and tetra functional thiol crosslinker (pentaerythritol tetrakis (3 mercaptopropionat e) PETMP) were selected as commercially available monomers. Non equimolar monomer solutions were prepared with an excess of 15% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2 hydroxyethoxy) 2 methylpropiophenone (HHMP) were added as the respective catalyst and photo initiator to the solutions. (b) Representative polydomain structure and physical samples demonstrating ability to mould different geometries. (c) A mechanical stress is applied to t he polydomain samples to align the mesogens into a temporary monodomain. (d) A photopolymerization reaction is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample. Photo image compares sample before and after stretching and photo curing. (e) WAXS pattern of aligned sample confirming nematic structure. (f) POM image of unaligned sample at 20x magnification. *Toluene was used as an optional component to the system to reduce solution viscosity

PAGE 80

! 61 second independent photopolymerization reaction can be used to furth er tailor the properties of the polydomain or stabilize an aligned monodomain. This approach offers elegant and scalable synthesis of LCEs as well as offers unprecedented spatio temporal control over liquid crystalline behavior. A schematic of the two stag e TAMAP reaction is presented in Figure 5. 1. For this study, commercially available starting materials with no additional purification were chosen to demonstrate the efficacy of the approach. RM 257 was selected for its use as a well known diacrylate mesog en, ( 69 128 ) while a di functional and a tetra functional thiol monomer were selected for use as a flexible spacer and crosslinker, respectively. Non equimolar solutions were simply mixed in a vial, poured into moulds, and allowed to cure in open air (detailed synthesis procedures and additional experimental results are provided in the Chapter 6). The first stage reaction is used to create a polydomain LCE via the thiol Michael addition reaction, a click reaction between a thiol group and an electron deficient vinyl group (i.e. an acrylate), which is not limited in its scale. Previous work by Hoyle has demonstrated that nearly 100% conversion of the thiol groups can be attained and controlled over a timescale of approximately a few seconds to one day. ( 1 29 ) Several polydomain samples ranging from a thin film (~200 'm thick) to bulk samples (4 mm thick) are shown in Figure 5. 1b to demonstrate the manufacturability of the thiol acrylate reaction. Ultimately, this reaction will self limit when the thiol g roups have all reacted. An independent, second stage polymerization reaction between excess acrylate groups can then be photo triggered. This second reaction is used to further tailor the properties of the LCE as well as permanently program an aligned mon odomain sample via the establishment of new crosslinks. A demonstration of permanent monodomain alignment can also be seen in

PAGE 81

! 62 Figure 5. 1cd, shown by a rectangular sample being stretched, photo crosslinked, and released. The opaque polydomain sample becomes transparent when stretched, visually indicating the formation of a monodomain. This method has provided an added degree of accessibility for our LCE collaborators, as we have successfully sent polydomain samples to separate laboratories to program a stable monodomain using the secon d photopolymerization reaction. For this communication, an LCE system with 13% of the thiol functional groups belo nging to the crosslinker and a non equimolar excess of 15% acrylate groups was used to Figure 5.2. Polydomain and monodomain LCE samples were subjected to 0 and 100 kPa bias stresses and cooled from 120 to 20 ¡ C at a rate of 5 ¡ C/min. Monodomain samples exhibited 45% actuation under zero stress. The monodomain samples in this experiment were programmed by stretching a polydomain sample to 100% strain and photo crosslinking for 10 minutes.

PAGE 82

! 63 highlight the potential of this methodology. The presence of a liquid crystalline state was verified with polarized optical microscopy, which showed birefringence that d isappeared upon heating above T i Single and wide angle x ray analysis revealed the presence of a nematic structure at room temperature. Dynamic mechanical analysis revealed the glass transition temperature ( T g ) increased from 15 to 19 ¡ C from the first to second stage reaction, while differential scanning calorimetry (DSC) revealed the nematic to isotropic transition temperature ( T i ) to be 80 ¡ C after the first stage reaction (Figure A2.3); however, T i could not Figure 5 3 (a) Alternating regions in a polydomain LCE are photo crosslinked, which become resistant to transparent, monodomain alignment when stretched. (b) An unaligned LCE is heated to the isotropic state and crosslinked with a photo mask. Upon cooling, photo crosslinked areas remain isotropic to revealan image.

PAGE 83

! 64 be identified using DSC after the second stage reaction. The thermal actuation behaviour of both polydomain and programmed monodomain samples can be seen in Figure 5. 2. Polydom ain samples only exhibit a shape switching response when under the presence of a bias stress. This stress drives the formation of a monodomain when cooling below T i Conversely, programmed monodomain samples show autonomous, "hands free" actuation. Both po lydomain and monodomain samples experienced increasing amounts of actuation strain with increased bias stress. These results indicate that mechanically useful LCE samples can be produced after both the first and second stages of this reaction. This behavio ur cannot be accomplished using current hydrosilylation reactions, as the multi step process does not consist of two independent reactions. Rather, the hydrosilylation method involves slowing the reaction at a critical point during the gelation process to align the monodomain, which can be difficult to replicate. It should be noted that recent studies have proposed unique methods to create more robust two step techniques to program monodomain samples by introducing photo sensitive crosslinking side groups a long the main chain ( 32 ) or using exchangeable crosslinks at high temperatures; ( 130 ) however, these methods do not offer facile control over polymer structure as the proposed TAMAP reaction. The overall purpose of this work is to introduce a new a pproach to controlling both LCE structure and liquid crystalline behaviour. The presented TAMAP reaction provides unique spatio temporal control over the material to influence both mechanical and optical properties (Figure 5. 3a). In this example, the secon d stage photo polymerization reaction was used to increase crosslinking at specific alternating regions within a polydomain sample. Upon stretching, these regions have increased crosslinking and resist chain alignment and the formation of a transparent mon odomain. Eventually, the process reveals a sample with alternating optical and mechanical properties. Another application of this approach is the control over the formation of liquid crystalline domains (Figure 5. 3b). In this example, the

PAGE 84

! 65 second stage reac tion is used to reveal the formation of an image using opaque polydomains and the transparency of the the state The addition of photo crosslinks served to restrict the formation of the polydomain when cooled below T i Previous studies have not demonstrate d this amount of precision and control over both LCE structure and liquid crystalline behaviour. Furthermore, these results suggest that the second stage photopolymerization reaction can be used to control the phase transitions between the polydomain, mond omain, and isotropic phases in response to a stimulus in specific locations. LCEs are a class of active polymers that are capable of mechanical actuation in response to a stimulus, commonly heat or light. ( 131 132 ) Unfortunately, LCEs have not experi enced the same level of widespread research attention similar to other classes of actively moving polymers, such as shape memory polymers (SMPs), though both systems are generally known for their ability to mechanically respond to a change in temperature. In addition, both systems require proper "programming" of the polymer to an aligned state before shape change can occur. The key difference is that SMPs exhibit a one time shape recovery event when heated above a thermal transition ( T g or T m ) and are drive n by entropy elasticity, ( 133 ) while LCEs repeatedly undergo a shape switching phenomenon driven by a reversible anisotropic isotropic transition associated with liquid crystalline order ( 134 ) As a result, LCEs have an added degree of functionality capable of creating devices that repeatedly actuate over the lifetime of the device, such as in artificial muscles; ( 18 68 ) nevertheless, SMPs have received a higher profile of interest for proposed applications, especially biomedically related. ( 135 137 ) It is of interest to note that recently researchers have proposed incorporating the shape memory effect within LCE systems to take advantage of both mechanisms. ( 138 )

PAGE 85

! 66 The difficulty of synthesis and programing of monodomain LCE samples has been a long time challenge. ( 71 ) The proposed TAMAP methodology may overcome traditional barriers to access these exquisite materials. Furthermore, it m ay provide an easily accessible platform to manufacture and tailor LCE based biomedical devices. The composition presented in this study demonstrated non cytotoxic responses after both first and second stage reactions (Figure 5. 4). The proposed TAMAP appro ach provides the ability to explore potential biomedical applications of LCE materials with enhanced functionality and control. For example, these data suggest the second stage reaction may be utilized to tailor the LCE properties in vivo due to the non cy totoxic response at both stages of the reaction. While there have been a handful of toxicity studies performed on liquid crystal based materials and sensors, ( 139 140 ) biocompatibility data for LCEs remain largely unreported. Future studies are needed to fully ev aluate the biocompatible nature of these materials. 5.2. Conclusions This study presented an unexplored two stage TAMAP reaction. Mechanically robust polydomain samples were synthesized using the self limiting Michael addition reaction, Figure 5.4. C ytocompatibility of the TAMAP synthesized LCE was confirmed after both the first and second stages of the reaction using both elution and direct contact test by an independent laboratory (WuXi AppTec, St. Paul, MN, USA). Cellular response to both (a) direct contact and (b) elution tests are shown.

PAGE 86

! 67 and demonstrated st rain actuation under a bias stress in response to temperature. The second stage photopolymerization reaction was used to permanently program a monodomain within the samples, which demonstrated "hands free" actuation without the need for a bias stress. Furt hermore, this second reaction was used to tailor the mechanical properties and liquid crystalline behaviour with spatio temporal control. The composition investigated within this study elicited a cytocompatible response at each stage of the TAMAP reaction. 5.3. Acknowledgments The National Science Foundation CAREER Award 1350436 supported this work. The authors would like to thank Ellana L. Taylor and Brandon Mang for their help in experimental testing as well as Amir Torbati, Jaimee Robertson, and Patrick T. Mather for their help in SAXS and WAXS characterization.

PAGE 87

! 6 8 CHAPTER VI SYNTHESIS OF PROGRAMMABLE MAIN CHAIN LIQUID CRYSTALLINE ELASTOMERS USING A TWO STAGE THIOL ACRYLATE REACTION 6.1. Abstract This study presents a novel two stage thiol acrylate Michael addition photopolymerization (TAMAP) reaction to prepare main chain liquid crystalline elastomers (LCEs) with facile control over network structure and programming of an aligned monodomain. Tail ored LCE networks were synthesized using routine mixing of commercially available starting materials and pouring monomer solutions into molds to cure. An initial polydomain LCE network is formed via a self limiting thiol acrylate Michael addition reaction. Strain to failure and glass transition behavior were investigated as a function of crosslinking monomer, pentaerythritol tetrakis(3 mercaptopropionate) (PETMP). An example non stoichiometric system of 15 mol% PETMP thiol groups and an excess of 15 mol% ac rylate groups was used to demonstrate the robust nature of the material. The LCE formed an aligned and transparent monodomain when stretched, with a maximum failure strain over 600%. Stretched LCE samples were able to demonstrate both stress driven thermal actuation when held under a constant bias stress or the shape memory effect when stretched and unloaded. A permanently programmed monodomain was achieved via a second stage photopolymerization reaction of the excess acrylate groups when the sample was in the stretched state. LCE samples were photo cured and programmed at 100%, 200%, 300%, and 400% strain, with all samples demonstrating over 90% shape fixity when unloaded. The magnitude of total stress free actuation increased from 35% to 115% with increase d programming strain. Overall, the two stage TAMAP methodology is presented as a powerful tool to prepare main chain LCE systems and explore structure property performance relationships in these fascinating stimuli sensitive materials.

PAGE 88

! 69 6.2. I ntroduction LCEs are a class of stimuli responsive polymers that are capable of exhibiting mechanical and optical functionalities due the combination of liquid crystalline (LC) order and rubber elasticity. These materials can demonstrate extraordinary changes in shape soft elasticity behavior, and tunable optical properties in response to a stimulus such as heat or light, ( 6 33 141 ) which makes them suitable for many for potential technological applications such as artificial muscles, ( 27 33 142 ) sensors, and actuators. ( 17 33 ) LCEs have already b een demonstrated in many applications such as micro grippers for robotics, ( 102 ) micro electromechanical systems (MEMS), ( 143 ) optical grating devices, ( 104 143 ) tunable apertures, ( 105 ) and microfluidic systems ( 106 ) The structural components that give rise to the ordered LC phases are called mesogens. Mesogens are the basis of the LC domains and are typically composed of two or three linearly connected aromatic rings with flexible ends. These moieties can be directly placed within t he polymer backbone to create main chain LCEs or as a side group ( i.e. side on or end on LCEs). ( 6 80 ) Main chain LCEs have generated a lot of interest due to their direct coupling between mesogenic order and polymer backbone conformations. ( 30 31 36 46 ) This d irect coupling allows main chain LCEs to exhibit higher degrees of mesogen orientation, mechanical anisotropy, and strain actuation. ( 31 ) Thermal actuation of LCEs rel ies on a reversible anisotropic isotropic transition associated with LC order. ( 33 ) To program an LCE for actuation, the mesogens must first be oriented along a director to form a monodomain ( i.e. anisotropic mesophase) and is often referred to as a liquid single crystal elastomer. Actuation occurs as an aligned LCE is heated above an isotropic clearing temperature ( T i ), which disrupts th e order of the mesogens into an isotropic state and drives shape change. A monodomain can be formed temporarily by applying an external stress ( i.e. hanging a weight) to a sample, which will align the polymer chains and orient the mesogens in the direction of the stress. Permanent programming of the monodomain can be achieved via a multi step

PAGE 89

! 70 process, which involves producing a lightly cross linked gel followed by immediate application of mechanical stress to induce orientation of the mesogens. Once aligned the reaction is continued to established covalent crosslinks and stabilize the monodomain. ( 39 ) Other "one pot" alig nment techniques can be performed in the presence of electric fields or by surface alignment ( i.e. rubbing polyimide on a glass slide) during polymerization; however, these methods are generally limited to thin film samples. ( 6 30 ) Finkelmann and Bergmann introduced the first synthetic route for the preparation main chain LCEs using one step platinum catalyzed hydrosilylation reaction of a divinyl mesogen and a tetra functional siloxane crosslinker. ( 46 ) This method has been widely adapted by many research groups to synthesize main chain LCEs. ( 31 138 144 ) Polyesterification and epoxy based reactions have also been used to make main chain LCEs. ( 32 ) All of these methods require high purity starting materials and careful experimental conditions to prevent side reactions. ( 6 ) Furthermore, these methods rely on random cross linking of the monomers, resulting in poorly defined network structure. Therefore, it is more difficult to correlate the structure to the properties of LCEs. Recent studies have used click chemistry as a tool to prepare more uniform LCE networks; ho wever, these reactions require custom synthesized starting mesogenic and thiol monomers, which can be challenging to produce, and have been limited to prepare micron sized actuators rather than bulk samples. ( 145 147 ) Current challenges in the LCEs focus on how to develop synthetic methods that are facile, reproducible, and scalable to design tailored LCE networks with programmable monodomains. Recently, our group introduced a tw o stage thiol acrylate Michael addition photopolymerization (TAMAP) methodology for the first time in mesomorphic systems to prepare nematic main chain LCEs. ( 55 ) Two stage TAMAP reactions form dual cure polymer networks, where the staging of the polymerization process allows modification of the polymer structure at two distinct time

PAGE 90

! 71 points. This strategy has been adapted in the past few years to design and fabricate other advanced materials, other than mesomorphic systems, such as micro actuators, ( 148 ) shape memory polymers, ( 149 150 ) and surface wrinkles. ( 151 152 ) The TAMAP methodology utilizes a non stoichiometric composition with an excess of acrylate functional groups. The first stage reaction is used to create a polydomain LCEs via the thiol Michael addition reaction, which is self limited by the thiol groups. This is an intermediate LCE network that would be capable of mesogenic domain orientation by applying mechanical stress. The polydomain resulting from the first stage Michael addition reaction is indefinitely stable and the alignment of the monodomain does not need to occur immediately after the reaction has completed. The second stage photopolymerization reaction between excess acrylate groups is used to pe rmanently fix an aligned monodomain and program the LCE for reversible and stress free ( i.e. "hands free") actuation. The purpose of this study is to explore and demonstrate the robust nature of the TAMAP reaction to prepare main chain LCEs by investigati ng the influence of crosslinking density and programed strain on the thermomechanics of the LCE systems. We demonstrate a wide range of thermomechanical properties and actuation performance that are achievable using this reaction. 6.3. P rotocol 6.3.1. Pre paration of Liquid Crystalline Elastomers LCEs 1. Add 4 g of 4 bis [4 (3 acryloyloxypropypropyloxy) benzoyloxy] 2 methylbenzene (RM257) into a 30 ml vial. RM257 is a di acrylate mesogen and is received as a powder. Dissolve RM257 by adding 40 wt% ( i.e. 1.6 g ) of toluene and heat to 80 ¡C on a hot plate. This process typically takes less than 5 min to dissolve the RM257 into a solution. Note: Other solvents can be used to dissolve the RM257, such as dichloromethane (DCM), chloroform, and dimethylformamide; how ever, toluene was chosen because it allows the monomers to cure at room temperature

PAGE 91

! 72 without having the solvent evaporate quickly during reaction, while DCM and chloroform could evaporate quickly at room temperature before the Michael addition reaction is c ompleted. Dimethylformamide can dissolve RM257 immediately without heating, but requires very high temperatures to remove the solvent (~ 150 ¡C). Kamal and Park used a combination of DCM and a liquid crystal, CB5, to dissolve RM257. ( 153 ) 2. Cool the solution to room temperature. Add 0.217 g of pentaerythritol tetrakis(3 mercaptopropionate) (PETMP), a tetra f unctional thiol crosslinking monomer, and 0.9157 g of 2,2 (ethylenedioxy) diethanethiol (EDDET), a di thiol monomer. The molar ratio of thiol functional groups between PETMP and EDDET is 15:85. This ratio will be referred to as 15 mol% PETMP throughout the study. Note: If the RM257 recrystallizes during this process, temporarily place the vial back onto the 80 ¡C hot plate until the monomer returns to solution. Cool the solution to room temperature before proceeding to the next steps. 3. Dissolve 0.0257 g of (2 hydroxyethoxy) 2 methylpropiophenone (HHMP) into the solution. HHMP is a photoinitiator used to enable the second stage photopolymerization reaction. This step can be skipped if the second stage reaction will not be utilized. 4. Prepare a separate solut ion of a catalyst by diluting dipropylamine (DPA) with toluene at a ratio of 1:50. Add 0.568 g of diluted catalyst solution to the monomer solution and mix vigorously on a Vortex mixer. This corresponds to 1 mol% of catalyst with respect to the thiol funct ional groups. Note: Adding undiluted catalyst, such as DPA, to the solution will likely result in extremely rapid localized polymerization and will prevent manipulation of the polymer solution into the desired mold detailed in the next steps. 5. Place the m onomer solution in a vacuum chamber for 1 min at 508 mmHg to remove any air bubbles caused by mixing. Perform this step immediately after mixing.

PAGE 92

! 73 6. Immediately transfer the solution into the desired mold or inject the solution between two glass slides. Mol ds should be manufactured from HDPE. The molds do not need to be covered, as the Michael addition reaction is relatively insensitive to oxygen inhibition. 7. Allow the reaction to proceed for at least 12 hr at room temperature. The solution will begin to ge l within the first 30 min. 8. Place samples in a vacuum chamber at 80 ¡C and 508 mmHg for 24 hr to evaporate the toluene. Once completed, the samples should have a glossy white and opaque appearance at room temperature. 9. Repeat the procedure to tailor the ratio of tetra functional to di functional thiol monomers in step 1.2 with ratios of 25:75 50:50, and 100:0, respectively. A detailed table of the chemical formulations used for this study is shown in Table 6. 1 6.3. 2. Kinetics Study of Two stage Reaction with Real time Fourier Transform Infrared 1. Equip a spectrometer with a MCT/B detector and XT KBr beam splitter. 2. Prepare a mixture using the protocol outlined above in the Preparation of LCE section using 0.5 mol% of catalyst with respect to thiol functional groups and 0.5 wt% of photoinitiator. Two initiators were tested separately, 2 2 dimethoxy 2 phenylacetophenone (DMPA),and HHMP. DMPA is a more commonly used initiator, while HHMP is more stable at elevated temperatures. 3. Place one drop o f LCE mixture between NaCl crystals immediately after mixing using a glass pipette. 4. Record spectra at a 2.92 sec sampling interval rate. 5. Monitor the conversion of the thiol groups using a peak height profile with the S H absorption

PAGE 93

! 74 peak at 2,571 cm 1 w ith a baseline of 2,614 2.527cm 1 6. Monitor the conversion of the acrylate groups using a peak height profile with the C=C absorption peak at 810 cm 1 with a baseline of 829 781 cm 1 7. Allow the reaction to proceed under FTIR at room temperature until the thiol peak height plateaus, showing 100% conversion of thiol groups. 8. Upon complete conversion of thiol group, turn on a 365 nm light source equipped with a light guide for 10 min to complete the polymerization of excess acrylates at 350 mW/cm 2 intens ity, which can be measured by a radiometer photometer. 9. Monitor the conversion of the acrylate groups as described in 2.6. 6.3. 3. Dynamic Mechanical Analysis (DMA) 1. Prepare two glass slides by spraying the surfaces of the slides with a hydrophobic surface agent and rubbing the surfaces with a paper towel until dry. 2. Stack slides together such that they are separated with a 1 mm spacer. Spacers can be cut by scoring and breaking a separate glass slide to measure approximately 25.4 mm x 5 mm x 1 mm. Clamp slides together using a binder clip at each end. 3. Inject monomer solution between the slides using a glass pipette. This requires approximately 1.5 g of the prepared monomer solution. 4. Allow the sample to cure for at least 12 hr according to step 1 .7. Separate the glass slides and dry the sample according to Step 1.8. 5. Using a razor blade or scissors, cut a rectangular test specimens with dimensions of 30 x 10 x 1 mm 3

PAGE 94

! 75 6. Load the sample properly into a DMA machine. Test the sample in tensile mode, with active length measuring 10 to 15 mm. Take care not to over tighten the grips on the test sample, as 0.1 N m is often too much torque when tightening the grips. 7. Cycle the sample at 0.2% strain at 1 Hz from 50 to 120 ¡C at a heating rate of 3 ¡C/min Set the force track to 125%. 8. Measure the glass transition temperature ( T g ) at the peak of the tan $ curve. 9. Measure the isotropic transition temperature ( T i ) and the lowest point of the storage modulus curve. 10. Measure the rubbery modulus, E' r at T i +30 ¡C. 6.3. 4. Strain to failure Tests 1. Prepare an HDPE mold by milling ASTM Type V dog bone cavities at a depth of 1 mm. 2. Using a glass pipette, fill each dog bone cavity until the monomer solution is flush with the top of the mold. Allow the samples to cure and dry according to Steps 1.7 and 1.8. 3. Prepare 5 tensile specimens from LCE samples formulated with varying PETMP crosslinker concentrations of 15, 25, 50, and 100 mol%. 4. Set two pieces of reflective laser tape 5 to 7 mm apart within the gage len gth of the specimen. 5. Load the specimen into a mechanical tester equipped with a laser extensometer, thermal chamber, and 500 N load cell. Use self tightening grips to secure the specimens, as specimens will dislodge from wedge grips at high strain value s. Align the laser extensometer properly to track the accurate change in length as a function of applied strain.

PAGE 95

! 76 6. Strain the specimens at room temperature with a displacement rate of 0.2 mm/sec until failure. Define failure by the fracture of the specime n. 7. Test additional specimens of 15 mol% PETMP crosslinking agent for strain to failure testing as a function of temperature. Test specimens at 40, 30, 20, 10, 0, 10, 22, 40, 60, and 80 ¡C. Hold all specimens isothermally at the desired test temperat ure for 10 min prior to testing. 6.3. 5. Shape Fixity and Actuation Tests 1. Prepare an HDPE custom dog bone mold with gage length of 25 mm and cross sectional area of 1 mm x 5 mm. 2. Prepare a 15 mol% PETMP monomer solution according to Steps 1.1 to 1.5. 3. Using a glass pipette, fill each mold cavity until the monomer solution is flush with the top of the mold. 4. Allow the samples to cure and dry according to Steps 1.7 and 1.8. 5. Set two pieces of reflective laser tape 5 7 mm apart within the gage length of the specimen. Load the sample according step 4.5. Using a permanent marker, mark a dot in the other side of each piece of reflective tape. Record the length between the dots. 6. Strain the specimens at room temperature with a displacement rate of 0.2 mm/sec to 100%, 200%, 300%, or 400% strain. 7. While maintaining the desired strain level, expose the sample to a 365 nm UV light source at an intensity of ~10 mW/cm 2 for 10 min by holding a UV Lamp approximately 150 mm from the sample. 8. Unload the sample and t hen heat it above T i to induce actuation. Allow the sample to cool back

PAGE 96

! 77 to room temperature and record the length between the dots. 9. Calculate fixity using the following equation: where ( applied is programming strain before photo crosslinking (measured by the laser extensometer) and ( fixed is the amount of permanent strain after photo crosslinking (measured by the change in dot displacement). 10. Cut a 30 mm length sample from the center portion of the programmed specimen. 11. Load the sample properly into a DMA tester. Test the sample in tensile mode, with active length measuring 13 to 15 mm. Make sure not to over tighten the grips on the test coupon. 12. Equilibrate the sample at 120 ¡C under a preload of 0 N. Cool the sample from 120 to 25 ¡C at a rate of 3 ¡C/min. Maintain the pre force at 0 N for the entire test. 6.4. Representative Results In this study, the two stage TAMAP reaction cure kinetics were investigated using real time FTIR. An FTIR series study on the conversion of the thiol and acryla te groups as a function of time to capture both the stages of the reaction was implemented and the normalized results are shown in Figure 2a The first stage thiol acrylate Michael addition reaction was initiated via base catalysis using DPA as the catalys t and results in the formation of a crosslinked polymer network. At the end of this initial reaction, the thiol functional groups achieve close to 100% conversion within 5 hours of under ambient conditions (~22 ¡C), while the acrylate groups attained betwe en 70% to 78% conversion under the same conditions. The thiol acrylate Michael addition 'click' reaction is self limiting in nature and can generate a step growth, crosslinked, stable network in a facile manner based on the relative ratios of functional gr oups present. Subsequently, the second

PAGE 97

! 78 stage photopolymerization reaction was initiated via exposure to UV irradiation and the remaining unreacted acrylate groups present within the network were further crosslinked to achieve a final acrylate functional gr oup conversion near 100%. Two photoinitiators, HHMP and DMPA and their reaction kinetics were studied within the polymer networks and both were seen to efficiently create crosslinked acrylate networks at the end of second stage polymerization. The conversi on of the acrylate groups as a function of the intensity of exposure was also studied and seen to correlate. Overall, it was observed that though a number of variables such as photoinitiators and exposure times could be varied, it was possible to efficient ly attain high final conversion of the acrylates at the end of the second stage within 10 minutes even with relatively low levels of UV intensity (~10 25 mW/cm 2 compared to 350 mW/cm 2 ). Figure 6. 2b shows the FTIR absorbance spectra of the two stage reactio n at 3 different time points, 0, 300, and 320 minutes. At time 0, the initial spectra captures the presence of both thiol and acrylate functional groups in their unreacted state. At the 300 minute time point, by the end of the first stage thiol Michael add ition reaction, the thiol and acrylate peak heights are seen to reduce considerably, thereby implying the reaction between the thiol and acrylate functional groups has progressed to completion. The thiol peak is measured to be close to 100% conversion at t his point, whereas the acrylates are seen to be consumed up to 78%. The complete disappearance of the thiol peak is not observed, most likely as the presence of the thiol Michael adduct from the first stage reaction is seen to appear and overlap with the t hiol peak at 2,571 cm 1 At the end of the second stage photopolymerization reaction initiated via UV exposure, at the 320 minutes point, the acrylate conversion is seen to proceed to completion, implying 100% conversion of remaining acrylic double bonds w ithin the network. ( 154 ) The two stage TAMAP methodology provides facile control to explore structure property relationships in LCEs. The influence of crosslinking density on stress strain behavior is shown in Figure 6. 3a Modulus and fracture stress were shown to increase with in creasing PETMP content,

PAGE 98

! 79 while failure strain increased with decreasing PETMP content ( Figure 6. 3b ). LCE samples with 50 and 100 mol% PETMP demonstrated initial elastic loading followed by a stress plateau and sharp increase in stress due to chain alignment In comparison, samples with 15 and 25 mol% PETMP appeared to demonstrate more traditional elastomeric loading followed by an increase of stress due to chain alignment. All specimens tested showed a transition from white opacity to clear transparency when stretched ( Figure 6. 3e ). It should be noted that all specimens maintained a large degree of permanent strain after fracture and did not recover to their original shape at room temperature; however, all specimens visually recovered to their original shape upon heating above T i The influence of temperature on failure strain was then investigated for the 15 mol% PETMP composition ( Figure 6. 3c ). In the glassy state, LCE specimens exhibited brittle failure with no appreciable deformation. At the onset of the g lass transition, the failure strain increased significantly and followed the general shape of the tan function measure by DMA. The failure strain reached a maximum of 650% s train at 10 ¡C. Representative g lass transition behavior for the four LCE network systems is shown in Figure 6. 3d All of the LCE networks displayed non traditional behavior in both the storage modulus and tan curves. The storage modulus of all LCE networks displayed a distinct minimum that was roughly associated with T i The tan f unctions were represented by an initial peak followed by an elevated region that diminished as the sample was heated into the isotropic state (a representative curve can be seen in Figure 6. 3c ). For the four LCE systems tested, both T g and rubbery modulus increased with increasing crosslinking density. A summary of thermo mechanical properties of the four LCE systems can be seen in Table 6. 2 LCEs offer the ability to demonstrate both the shape memory effect and reversible actuation ( Figure 6. 4 ). An unalign ed polydomain specimen of 15 mol% PETMP was used to illustrate the different shape switching pathways that can be programmed into the material ( Figure 6. 4a ).

PAGE 99

! 80 Reversible stress driven actuation is demonstrated by the pathway in Figure 6. 4a b c The polydoma in specimen is stretched by hanging a 60.6 mN weight to apply a constant stress. This bias stress mechanically orients the mesogens into a transparent monodomain. The specimen contracts when heated to the isotropic state and elongates when cooled below T i This process can be repeated indefinitely. The shape memory effect was exhibited when the bias stress was removed from the specimen when cooled below T i to 22 ¡C, which is still 18 ¡C above T g While some elastic recoil was observed, a majority of the st rain remained programmed into the material. It should be noted that the mesogens remained in a stable monodomain orientation, and there is a noticeable difference in optical properties within the free end of the sample where the clamp was attached ( i.e. th e gripped portion remained glossy white). Heating the sample above T i activated full shape recovery, indicating the shape memory cycle follows the pathway of Figure 6. 4abde The second stage photopolymerization reaction can be used to achieve stress free actuation without the need for a constant bias stress or programming step between cycles. The temporarily

PAGE 100

! 81 aligned specimen was photo cured using 365 nm light at ~10 mW/cm 2 for 10 minutes ( Figure 6. 4f ). The sample experienced minimal elastic recoil when un loaded due to the establishment of covalent crosslinks between the excess of unreacted acrylate groups ( Figure 6. 4g ). Stress free actuation was then activated by controlling the temperature about T i using the reversible pathway in Figure 6. 4g h ; however, i t should be noted that the sample does not experience full recovery back to the initia l shape of the specimen. The influence of applied programming strain ( i.e. strain during photopolymerization) as function of fixity and actuation for the 15 mol% PETMP sy stem is shown in Figure 6. 5a All specimens demonstrated fixity values higher than 90%. The amount of programming strain did not no ticeably influence the fixity values for the strain range tested in this study. Conversely, actuation strain increased linearly with the amount of programming strain. On average, the actuation strain corresponded to approximately 30% of the programming str ain value. Representative curves showing actuation as a function of temperature can be seen in Figure 6. 5b It should be noted that the actuation strain values in Figure 6. 5a correspond to Figure 6.1. Schematic of Monodomain Programing via a Two Stage Thiol Acrylate Reaction. ( a ) A diacrylate mesogen (1,4 bis [4 (3 acryloyloxypropyloxy)benzoyloxy] 2 methylbenzene RM 257), dithiol flexible spacer (2,20 (ethylenedioxy) diethanethiol EDDET), and tetra functional thiol crosslinker (pentaerythritol tetrakis(3 mercaptopropionate) PETMP) were selected as commercially available monomers. Non equimolar monomer solutions were prepared with an e xcess of 15 mol% acrylate functional groups and allowed to react via a Michael addition reaction. Dipropyl amine (DPA) and (2 hydroxyethoxy) 2 methylpropiophenone (HHMP) were added as the respective catalyst and photo initiator to the solutions. ( b ) Repre sentative polydomain structure forms via Michael addition (first stage) with a uniform cross link density and latent excess acrylate functional groups. ( c ) A mechanical stress is applied to the polydomain samples to orient the mesogens into a temporary mon odomain. ( d ) A photopolymerization reaction (second stage) is used to establish crosslinks between the excess acrylate groups, stabilizing the monodomain of the sample.

PAGE 101

! 82 measurements between room temperature, 22 ¡C, and 90 ¡C, while the behavior shown in Figure 6. 4b was monitored between 25 and 120 ¡C. This expanded temperature range caused additional actuation strain to be realized: 80%, 102%, 125%, and 207% actuation strain for samples programmed at 100%, 200%, 300%, and 400% strain, r espectively. Figure 6 2 Kinetics Study of Michael Addition Reaction with Real Time FTIR. ( a ) Representative two stage thiol acrylate reaction kinetics showing conversion as a function of time using DMPA photoinitiator. At the end of first stage, the thiol groups reached near 100% conversion while 22% of acrylate groups were unreacted. At the end of the second stage, unreacted acrylates reached 100% conversion. ( b ) FTIR absorbance spectra showing the thiol and acrylate conversion before curing at time 0, upon completion of the first stage at 300 minute, and upon completion of the second stage at 320 minute.

PAGE 102

! 83 Figure 6 3 Thermomechanics of TAMAP LCE Systems. ( a ) Representative strain to failure curves of four LCE systems with 15 mol% excess acrylate and varying amount of PETMP crosslinker. ( b ) Failure strain as a function of PETMP crosslinker. ( c ) The influence of temperature on failure strain for an LCE system with 15 mol% PETMP. The failure strain is compared alongside the tan $ function of the material measured by DMA. ( d ) Representative glass transition behavior of four LCE systems tested. ( e ) Image of a stretched LCE specimen with 15 mol% PETMP compared to an untested specimen. Error bars in (b) and (c) represent standard deviation.

PAGE 103

! 84 F igure 6 4 Shape Switching Pathways in an LCE. This schematic represents several different pathways available to achieve shape switching in LCEs. A custom dog bone sample of 15 mol% PETMP is used in this demonstration with an initial shape of (a). Reversib le stress driven actuation is realized between (b c) by adjusting the temperature about T i while under a constant bias force (60.6 mN); the shape memory effect is achieved by following the programming and recovery cycle of (a b d e); and stress free actuat ion can be activated thermally between (g h) after a permanent monodomain has been programmed into the sample in step (f). The legend illustrates mesogen orientation in polydomain, monodomain, and isotropic states. T < T i and T > T i images were taken at 22 and 90 ¡C, respectively.

PAGE 104

! 85 Figure 6 5 Thermomechanical Response in Programmed Monodomain LCE Systems: ( a ) Shape fixity represents the efficiency of permanently aligning monodomain and all of samples show fixity above 90%. The magnitude of actuation measured between 22 and 90 ¡C on a hot plate. Error bars represent standard deviation. ( b ) The magnitude of actuation measured on DMA from 25 to 120 ¡C, the actuation increase with increasing of applied programming strain Table 6 1 Chemical Formulation s for LCE Systems: Four different LCE systems used in this study. The naming convention is based on the molar ratio of thiol functional groups between PETMP and EDDET. All systems have an excess of 15 mol% acrylate functional groups. It should be noted, FTIR studies tested HHMP as well as DMPA as photoinitiators and reduced the amount DPA catalyst by half to help with the kinetic characterization. *DPA is diluted in toluene at a ratio of 1:50.

PAGE 105

! 86 6.5. D iscussion Main chain LCEs have been investigated for numerous potential applications ranging from actuators and sensors to artificial muscles. Unfortunately, synthesis and monodomain alignment remain significant challenges that prevent many of these applications from being fully realized. ( 153 ) Recent work has explored new methods to help overcome these challenges, such as using exchangeable crosslinks to be able to re program an aligned monodomain multiple times. ( 49 ) The purpose of this study was to present a relatively unexplored approach to LCE synthesis and monodomain programming using a two stage TAMAP reaction. The first st age reaction is a "click" reaction based on a thiol acrylate Michael addition using an amine catalyst. Due to the nature of this reaction, full conversion of thiol acrylate Michael addition reaction was accomplished within 5 hours at room temperature using DPA as a catalyst ( Figure 2 ). It is important to note that this was achieved with commercially available materials without purification and using a relatively simple "mix and pour" method. 0.5 mol% of DPA with respect to thiol functional groups was chosen in this study for the control it gave over polymerization rate, allowing transferring the monomers solution into the mold. It is very important to note that the polymerization rate of Michael addition is merely dedicated by the catalyst concentration. Hig h Table 6 2. Summary of Thermomechanical Properties of LCE Systems: Dynamic Mechanical Analysis (DMA) test shows the thermomechanical properties of the initial polydomain LCE networks formed via the first stage Michael addition reaction. Both T i and E' r were measured at the lowest point of the storage modulus vs. temperature curve.

PAGE 106

! 87 catalyst concentration results an immediate gelation with high monomers conversions where is too low catalyst concentration allows slow conversions and often times high conversions cannot be achieved even as a function of time. Ultimately, the polymeriza tion rate can be tuned by the catalyst concentration. ( 129 ) One of the advantages this methodology offers is that the resulting intermediate polydomain LCE network is uniform and stable, such that the second stage reaction can be delayed indefinitely. This can enable the s ynthesis and programming steps to be performed in separate laboratories. Furthermore, the second stage reaction can be coupled with standard photolithography techniques to provide spatio temporal control over photo crosslinking. ( 55 ) For preparation of our experimental samples, HHMP was used as a photoinitiator because of its stability in present of visible light and at elevated temperatures, allowing for samples to be thermally cycled for stress driven actuation or the shape memory effect without triggering the initiator. A separate photo initiator was used for the FTIR portion of this study, helping illustrate that this methodology has the potential to be used with a variety of free radical initiators to dr ive the second stage reaction. The presented TAMAP methodology offers facile control over the structure of the initial polydomain LCE network. The four LCE networks synthesized demonstrate a wide range of achievable thermomechanical properties by varying the ratio between the PETMP crosslinker and EDDET spacer. Failure strain decreased with increasing in the co ncentration of PETMP, while T g and rubbery modulus (E r ) increased with increasing in PETMP concentration. This behavior is explained as an increase in PETMP concentration increases the crosslinking density of the networks and restricts chain mobility withi n the network. The strain to failure behavior follows the inherent inverse relationship between rubbery modulus and failure strain as shown in other amorphous shape memory polymer (SMP) networks. ( 155 ) LCE systems with high failure strains are generally more desirable, as they allow for increased alignment of the monodomain with larger programming strains. The failure strain of our 15 mol% PETMP system was maximized

PAGE 107

! 88 when strained near T g as measured by the peak of tan $. This is also in good agreement with previous studies that demonstrated the maximum strain in amorphous SMP networks o ccurred between the onset of the glass transition and T g ; ( 155 156 ) however, the LCE samples did not experience a rapid decrease in failure strain when heated above T g as shown in most elastomers. ( 157 ) This can be attributed to the elevated tan $ region that exists between T g and T i (i.e. the nematic phase) Previous groups have investigated and verified the unique tan $ loss behavior in nematic LCE networks. ( 158 159 ) This loss behavior is attributed to the soft elasticity in the nematic phase, such that the anisotropic shape of the mesogens can accommodate strains by rotation without experiencing an increase in stress. LCEs have generated a lot of scientific interest due to their stimuli responsive shape cha nging capabilities. ( 132 ) Polydomain LCE samples can be programmed to demonstrate reversible stress driven thermal actuation if stretched under a constant stress to program a temporary monodomain ( Figure 6. 4b c ); however, the shape memory ef fect can also be realized in LCE networks. ( 47 138 ) In this study, the TAMAP synthesized LCE samples could be programmed for shape memory at room temperature, in which a significant amount of strain remained stored in the sample even though the samp le was above T g To enable stress free or "hands free" actuation, the second stage photopolymerization reaction can be used to program a permanently aligned monodomain in stretched LCE samples. The efficiency of the second stage reaction can be examined by measuring fixity as a function of increasing stretch. It should be noted that fixity is a common metric used to evaluate the programming of SMP networks. ( 160 ) In this study, samples were programed at different levels of strain ( i.e. 100%, 200%, 300%, and 400%) and showed excellence fixity over 90%. Our results demonstrated that the magn itude of thermal actuation scaled with programming strain are in good agreement with previous results that link an increase in order parameter to increased mechanical actuation. ( 161 ) For example, LCE samples that were

PAGE 108

! 89 photo cured at 400% strain demonstrated on average 115% actuation when heated and cooled between 22 and 90 ¡C and 207% actuat ion when heated and cooled between 25 and 120 ¡C. Compared to other LCE studies, Ahir et al. ( 67 ) reported 400% actuation in LC polymer fibers and Yang et al. ( 145 ) reported 300% to 400% actuation for micro LCE pillars. It is important to note that the present study measures actuation differently from much of the LCE literature, which often calculates strain based on the length of the sample in the isotropic state. In this study, actuation strain is al ways based on the original length of the synthesized polydomain sample. This is more appropriate for the TAMAP methodology as it provides a more effective measure of the efficiency of the programming strain and photo crosslinking on both strain fixity and recovery. Regardless, our reported actuation strains are still lower than 400% as reported in other studies. However, this TAMAP reaction is still relatively unexplored and the influence of photo crosslinking has yet to fully be uncovered. While photo cros slinking is necessary to fix a permanent monodomain, too much photo crosslinking will prevent actuation from occurring. Theoretically, there should exist an optimal amount of photo crosslinking to both stabilize the monodomain and allow for maximum actuati on. Overall, the TAMAP methodology provides a powerful tool to synthesize LCE systems, tailor their structure, program permanent monodomain alignment, and ultimately explore this fascinating class of materials. 6.6. Acknowledgments This work was supporte d by NSF CAREER Award CMMI 1350436 as well as the University of Colorado Denver Center for Faculty Development. The authors would like to acknowledge Jac Corless, Eric Losty, and Richard Wojcik for their help in developing fixtures and molds for the synthe sis and characterization of these materials. The authors would also like to thank Brandon Mang and Ellana Taylor for their preliminary characterization of the materials.

PAGE 109

! 90 6.7. Materials Table 6 3 Materials used in the study.

PAGE 110

! 91 CHAPTER VII CONCLUSIONS AND FUTURE WORK 7.1 Conclusions The overall objective of this work was to establish structure property relationships in thiol acrylate based main chain liquid crystalline elastomers. In t his research work we used a solely new approach to prepare LCEs base d on using a thiol acrylate click reaction for polydomain networks and TAMAP reaction for monodomain networks, both of which have not previously been investigated for LCE synthesis. Three majors advances have been accomplished related to structure proper ty relationships of LCE for the first time, which a re the impact of the cross linking, tuning the LC phase via v arying the spacer length, and devolving TAMAP reaction with tuna ble (initial and finally) cross linking and programed strain conditions. The impact of cross l inking was examined in C hapter III Well defined nematic polydomain main chain LCE networks were synthesized using a thiol acrylate Michael addition reaction using both tri thiol and tetra thiol cross linkers. All networks had gel fractio n values greater than 90% using commercially available starting materials. The 1D WAXS patterns were not affected by crosslinking concentration or functionality and showed a nematic structure when strain to 80%. Crosslinker functionality showed a significa nt influence on the T i of the networks. The average T i for the tri thiol networks was 70 ¡ C, while the average T i for the tetra thiol networks was 80 ¡ C. Increased crosslinking was shown to reduce H f of the polydomain to isotropic transition. An increase in crosslinking density was shown to increase T g from 5 to 17 ¡ C and 3 to 25 ¡ C in tri thiol and tetra thiol crosslinked networks, respectively. Crosslinker density had an inverse relationship with fai lure strain, while cross linker functionality only had a significant influence at the lowest degree of crosslinking. The average failure strain increased from 542 to 853% from tetra thiol to tri thiol networks at 10 mol%, respectively. Crosslinker function ality did not

PAGE 111

! 92 influence thermal actuation behavior, whereas an increase in crosslinking reduced the magnitude of actuation. Networks showed a decrease in work capacity from 296 to 72 kJ/m 3 from 10 to 80 mol%. Another new technique was introduced to tune t he LC phase structure in C hapter IV A series of main chain LCE networks with tunable LC mesophases were synthesized using a thiol acrylate Michael addition reaction using a single nematic mesogen, RM257. The LCE networks exhibited different mesophases by controlling the length of thiol functionalized alkyl spacers (C2, C3, C6, C9, and C11). Longer spacers (C6, C9, and C11) drive a nano scale segregation resulting in smectic C phases, whereas shorter spacers (C2 and C3) resulted in nematic phases only as co nfirmed by the 2D WAXS patterns. The length of the spacers showed a significant influence on the thermomechanical properties of the networks such as T NI H f and T g T NI decreased from 140 to 90 ¡ C and H f increased significantly with increasing spacer length. The networks exhibited semi crystallinity the rate of crystallization significantly influenced by the spacer length. Shorter spacers (C2, C3, and C6) displayed a slow recrystallization rate after bein g annealed, whereas C9 recrystallized within 2 to 3 hours and C11 samples only needed 5 minutes to fully recrystallize due to increased spacer flexibility. Smectic C networks demonstrated larger magnitudes of actuation compared to nematic networks, with th e C11 network showing an average actuation of 525% 57%. Networks showed an increase in work capacity from 128 to 262 kJ/m 3 C2 to C11. In C hapter VI a new method was used to program monodomain LCE. In the study we presented an unexplored two stage TAMAP reaction. Mechanically robust polydomain samples were synthesized using the self limiting Michael addition reaction, and demonstrated strain actuation under a bias stress in response to temperature. The second stage photopolymerization reaction was used t o permanently program a monodomain within the samples, which demonstrated "hands free" actuation without the need for a bias stress. Furthermore, this second reaction was used to tailor the mechanical properties and liquid

PAGE 112

! 93 crystalline behavior with spatio temporal control. The composition investigated within this study elicited a cytocompatible response at each stage of the TAMAP reaction. Additional studies of TAMAP reaction me thod were presented in chapter VI both the initial (thiol acrylate) and the sec ond reaction (acrylate photopolymerization) had nearly 100% reaction conversion. Moreover, A wide range of thermomechanical properties was tailored by adjusting the amount of crosslinker, while the actuation performance was dependent on the amount of appli ed strain during programming. 7.2. Recommendations for the Further Work Thiol acrylate reaction is among the most widely implemented click reactions in polymerization and polymer syntheses; ( 53 ) however, implication of this powerful reaction in liquid crystalline materials is still in its beginning. In fact, this is the first attempt to apply this reaction in a bulk polymerization of liquid crystal line elastomers. While we had a lot of success on utilizing this reaction to tune the LC phase behavior and thermomechanic al properties of LCEs ( C hapter III and IV ), but a lot more need to be done. For example, the LC phase behavior in nematic LCE was foun d to be unaffected by the amount of the crosslinking, because nematic LCEs do not have layer structures; however, that may not be the case in smectic LCE due to their layered structure which may well be affected or changed to nematic b y changing the crossl inking. In C hapter IV we have found that the spacing between the mesogens is linked to controlling the LC phase behavior. Only one chemistry type was used as spacer between the mesogen with various lengths, which allowed getting nematic and smectic phases in the elastomers with distinctive transition temperatures. We recommend using different spacer chemistry or combine two different spacer chemistries in the future. That could result in new LC phases and more importantly bringing the isotropic transition temperature down spastically when long and more flexible spacers such as h exa(ethylene glycol) dithiol is used Low transition temperature actuators have the potential to be used in many engineering applications such as biomedical

PAGE 113

! 94 applications. Moreover, It will be highly endorsed to understand the LC structure of the LCEs prepare with combining two different spacer lengths with same or different chemistries. In this work we have developed a system that allowed us to prepare smectic and nematic LCEs using nearly identical chemistry. A comprehensive mechanical testing (i.e. stress strain) needs to be conducted to compare some unique LCE behaviors such as high damping, soft elasticity, and s e mi c rystallinity and polymer chain conformation. This will help bui ld a trend and relationship between the properties LCE and LC phase structure. Such a relationships will help aid design new en gineering applications. The configuration of liquid crystalline elastomer networks fundamentally dictated by the ordered mesogeni c properties (LC phase) and polymer chain conformations. ( 162 ) Very few studies tried to isolate, "which d ictates which". First, De Gennes predicted that LC phase effect the polymer chain conformation and later realized by Tokida et al. ( 162 163 ) As mesogens kee p the order polymer chain compensate the anisotropic by change their conformation into chain folds such as hairpin to reduce the entropy. Second, in this work we showed polymer chains also influence the LC phase formation. Global study is needed to uncove r the relationship between the ordered LC phase and polymer chain conformations. Such a study will help better understand many LCEs properties and behaviors such as actuation mechanism, the dissipation mechanism, and soft elasticity.

PAGE 114

! 95 BIBLIOGRAPHY 1. F. Reinitzer, BeitrŠge zur kenntniss des cholesterins. Monatshefte fŸr Chemie/Chemical Monthly 9 421 (1888). 2. J. W. Goodby et al., Liquid crystals with restricted molecular topologies: supermolecules and supramolecular assemblies. Chemical Communications, 2057 (1998). 3. D. Pauluth, K. Tarumi, Advanced liquid crystals for television. Journal of Materials Chemistry 14 1219 (2004). 4. L. Schmidt Mende et al., Self o rganized discotic liquid crystals for high efficiency organic photovoltaics. Science 293 1119 (2001). 5. P. P. Verbunt et al., Increased efficiency of luminescent solar concentrators after application of organic wavelength selective mirrors. Optics expres s 20 A655 (2012). 6. F. Bršmmel, D. Kramer, H. Finkelmann, in Liquid Crystal Elastomers: Materials and Applications. (Springer, 2012), pp. 1 48. 7. S. J. Woltman, G. D. Jay, G. P. Crawford, Liquid crystal materials find a new order in biomedical applicat ions. Nature materials 6 929 (2007). 8. T. J. White, D. J. Broer, Programmable and adaptive mechanics with liquid crystal polymer networks and elastomers. Nature materials 14 1087 (2015). 9. C. Ohm, M. Brehmer, R. Zentel, Applications of Liquid Crystalli ne Elastomers. 250 49 (2012). 10. M. Warner, E. M. Terentjev, Liquid crystal elastomers. (OUP Oxford, 2003), vol. 120. 11. P. De Gennes, RŽflexions sur un type de polymres nŽmatiques. CR Acad. Sci. Ser. B 281 2 (1975). 12. H. Finkelmann, H. J. Kock, G. Rehage, Investigations on liquid crystalline polysiloxanes 3. Liquid crystalline elastomers a new type of liquid crystalline material. Die Makromolekulare Chemie, Rapid Communications 2 317 (1981). 13. M. Bispo, D. Guillon, B. Donnio, H. Finkelmann, Main chain liquid crystalline elastomers: monomer and cross linker molecular control of the thermotropic and elastic properties. Macromolecules 41 3098 (2008). 14. D. Corbett, M. Warner, Changing liquid crystal elastomer ordering with light a route to opto me chanically responsive materials. Liquid Crystals 36 1263 (2009). 15. D. Krishnan, H. T. Johnson, Light induced deformation in a liquid crystal elastomer photonic crystal. Journal of the Mechanics and Physics of Solids 62 48 (2014). 16. K. M. Lee, H. Koer ner, R. A. Vaia, T. J. Bunning, T. J. White, Light activated shape memory of glassy, azobenzene liquid crystalline polymer networks. Soft Matter 7 4318 (2011). 17. S. Petsch et al., Smart artificial muscle actuators: Liquid crystal elastomers with integra ted temperature feedback. Sensors and Actuators A: Physical 231 44 (2015). 18. D. L. Thomsen et al., Liquid Crystal Elastomers with Mechanical Properties of a Muscle. Macromolecules 34 5868 (2001/08/01, 2001). 19. H. Wermter, H. Finkelmann, Liquid crystalline elastomers as artificial muscles. e Polymers 1 111 (2001). 20. S. Schuhladen et al., Iris like tunable aperture employing liquid crystal elastomers. Adv Mater 26 7247 (Nov 12, 2014). 21. E. Sungur et al., Tem perature tunable optical gratings in nematic elastomer. Applied Physics A 98 119 (2009). 22. T. Bera et al., Liquid Crystal Elastomer Microspheres as Three Dimensional Cell Scaffolds Supporting the Attachment and Proliferation of Myoblasts. ACS Applied Ma terials & Interfaces 7 14528 (2015/07/08, 2015). 23. Y. Gao et al., Biocompatible 3D Liquid Crystal Elastomer Cell Scaffolds and Foams with Primary and Secondary Porous Architecture. ACS Macro Letters 5 4 (2015).

PAGE 115

! 96 24. A. S‡nchez Ferrer et al., Photo Crosslinked Side Chain Liquid Crystalline Elastomers for Microsystems. Macromolecular Chemistry and Physics 210 1671 (2009). 25. A. Sanchez Ferrer et al., Liquid crystalline elastomer microvalve for microfluidics. Adv Mater 23 4526 (Oct 18, 2011). 26. D. Wei, New liquid crystal molecule advances organic solar cells. Liquid Crystals Today 24 99 (2015). 27. P. G. de Gennes, Un muscle artificiel semi rapide. Comptes Rendus de l'AcadŽmie des Sciences Series IIB Mechanics Physi cs Chemistry Astronomy 324 343 (1997). 28. B. Donnio, H. Wermter, H. Finkelmann, Simple and versatile synthetic route for the preparation of main chain, liquid crystalline elastomers. Macromolecules 33 7724 (Oct 17, 2000). 29. G. H. F. Bergmann, H. Finke lmann, V. Percec, M. Zhao, Liquid crystalline main chain elastomers. Macromol Rapid Comm 18 353 (1997). 30. H. Yang et al., Synthesis and physical properties of a main chain chiral smectic thiol ene oligomer. Liquid Crystals 37 325 (2010). 31. A. R. Garc ia Marquez, B. Heinrich, N. Beyer, D. Guillon, B. Donnio, Mesomorphism and Shape Memory Behavior of Main Chain Liquid Crystalline Co Elastomers: Modulation by the Chemical Composition. Macromolecules 47 5198 (2014). 32. P. Beyer, E. M. Terentjev, R. Zente l, Monodomain liquid crystal main chain elastomers by photocrosslinking. Macromolecular Rapid Communications 28 1485 (Jul 17, 2007). 33. C. Ohm, M. Brehmer, R. Zentel, Liquid Crystalline Elastomers as Actuators and Sensors. Advanced Materials 22 3366 (Au g 17, 2010). 34. J. Goodby. (Wiley VCH, New York, 1998). 35. A. S‡nchez Ferrer, H. Finkelmann, Thermal and mechanical properties of new main chain liquid crystalline elastomers. Solid State Sciences 12 1849 (2010). 36. B. Donnio, H. Wermter, H. Finkelmann A simple and versatile synthetic route for the preparation of main chain, liquid crystalline elastomers. Macromolecules 33 7724 (2000). 37. S. Krause et al., Nematic main chain elastomers: Coupling and orientational behavior. Comptes Rendus Chimie 12 8 5 (2009). 38. A. S‡nchez Ferrer, H. Finkelmann, Thermal and mechanical properties of new main chain liquid crystalline elastomers. Molecular Crystals and Liquid Crystals 508 348/[710] (2009). 39. J. KŸpfer, H. Finkelmann, Nematic liquid single crystal ela stomers. Die Makromolekulare Chemie, Rapid Communications 12 717 (1991). 40. S. Fridrikh, E. Terentjev, Polydomain monodomain transition in nematic elastomers. Physical Review E 60 1847 (1999). 41. S. Clarke, E. Terentjev, I. Kundler, H. Finkelmann, Text ure evolution during the polydomain monodomain transition in nematic elastomers. Macromolecules 31 4862 (1998). 42. C. Ortiz, R. Kim, E. Rodighiero, C. Ober, E. Kramer, Deformation of a polydomain, liquid crystalline epoxy based thermoset. Macromolecules 31 4074 (1998). 43. S. Clarke, E. Nishikawa, H. Finkelmann, E. Terentjev, Light scattering study of random disorder in liquid crystalline elastomers. Macromolecular Chemistry and Physics 198 3485 (1997). 44. K. A. Burke, P. T. Mather, Soft shape memory i n main chain liquid crystalline elastomers. Journal of Materials Chemistry 20 3449 (2010). 45. K. H. Hanus, W. Pechhold, F. Soergel, B. Stoll, R. Zentel, Phase behavior and elastic properties of a slightly crosslinked liquid crystalline main chain polymer Colloid and Polymer Science 268 222 (1990). 46. G. H. Bergmann, H. Finkelmann, V. Percec, M. Zhao, Liquid crystalline main chain elastomers. Macromolecular rapid communications 18 353 (1997). 47. I. A. Rousseau, P. T. Mather, Shape memory effect exhibited by smectic C liquid crystalline elastomers. J. Am. Chem. Soc. 125 15300 (2003).

PAGE 116

! 97 48. C. Ortiz, M. Wagner, N. Bhargava, C. Ober, E. Kramer, Deformation of a polydomain, smectic liquid crystalline e lastomer. Macromolecules 31 8531 (1998). 49. Z. Pei et al., Mouldable liquid crystalline elastomer actuators with exchangeable covalent bonds. Nature materials 13 36 (2014). 50. H. Yang et al., Novel liquid crystalline mesogens and main chain chiral smec tic thiol ene polymers based on trifluoromethylphenyl moieties. Journal of Materials Chemistry 19 7208 (2009). 51. T. H. Ware, Z. P. Perry, C. M. Middleton, S. T. Iacono, T. J. White, Programmable Liquid Crystal Elastomers Prepared by Thiol Ene Photopolym erization. ACS Macro Letters 4 942 (2015). 52. D. Martella, C. Parmeggiani, D. S. Wiersma, M. Pi–ol, L. Oriol, The first thiol yne click chemistry approach for the preparation of liquid crystalline elastomers. Journal of Materials Chemistry C 3 9003 (201 5). 53. C. E. Hoyle, C. N. Bowman, Thiol ene click chemistry. Angewandte Chemie International Edition 49 1540 (2010). 54. A. H. Torbati, P. T. Mather, A hydrogel forming liquid crystalline elastomer exhibiting soft shape memory. Journal of Polymer Science Part B: Polymer Physics 54 38 (2016). 55. C. Yakacki et al., Tailorable and programmable liquid crystalline elastomers using a two stage thiol acrylate reaction. RSC Advances 5 18997 (2015). 56. C. M. Yakacki, M. Saed, D. P. Nair, T. Gong, C. Bowman. (U S Patent 20,160,009,863, 2016). 57. M. O. Saed, A. H. Torbati, D. P. Nair, C. M. Yakacki, Synthesis of Programmable Main chain Liquid crystalline Elastomers Using a Two stage Thiol acrylate Reaction. Journal of visualized experiments: JoVE, (2016). 58. V. V. A. Azoug, J. Doolinga, M.O. Saed, C.M. Yakackib, T.D. Nguyena, The viscoelasticity of the polydomain monodomain transition in main chain liquid crystal elastomers. Submitted (2016, 2016). 59. C. Ahn, X. Liang, S. Cai, Inhomogeneous stretch induced pat terning of molecular orientation in liquid crystal elastomers. Extreme Mechanics Letters 5 30 (2015). 60. T. H. Ware, M. E. McConney, J. J. Wie, V. P. Tondiglia, T. J. White, Voxelated liquid crystal elastomers. Science 347 982 (2015). 61. K. Urayama, Se lected issues in liquid crystal elastomers and gels. Macromolecules 40 2277 (2007). 62. E. M. Terentjev, Liquid crystalline elastomers. Journal of Physics: Condensed Matter 11 R239 (1999). 63. A. S‡nchez Ferrer, H. Finkelmann, Polydomain Monodomain Orien tational Process in Smectic C Main Chain Liquid Crystalline Elastomers. Macromolecular rapid communications 32 309 (2011). 64. T. H. Ware, J. S. Biggins, A. F. Shick, M. Warner, T. J. White, Localized soft elasticity in liquid crystal elastomers. Nature communications 7 (2016). 65. C. Ohm, M. Brehmer, R. Zentel, in Liquid Crystal Elastomers: Materials and Applications. (Springer 2012), pp. 49 93. 66. M. Tokita, H. Tagawa, R. Ishige, K. Osada, J. Watanabe, Thermally reversible distortion observed for monodomain nematic elastomer of cross linked main chain polyester. Molecular Crystals and Liquid Crystals 465 193 (2007). 67. S. V. Ahir, A. R. Tajbakhsh, E. M. Terentjev, Self Assembled Shape Memory Fibers of Triblock Liquid Crystal Polymers. Advanced Functional Materials 16 556 (2006). 68. H. Wermter, H. Finkelmann, Liquid crystalline elastomers as artificial muscles. e Polymers 13 (2001). 69. K. M. Lee, T. J. Bunning, T. J. White, Autonomous, Hands Free Shape Memory in Glassy, Liquid Crystalline Polymer Networks. Advanced Materials 24 2839 (Jun 5, 2012).

PAGE 117

! 98 70. S. Krause, R. Dersch, J. H. Wendorff, H. Finkelmann, Photocrosslinkab le Liquid Crystal Main Chain Polymers: Thin Films and Electrospinning. Macromolecular Rapid Communications 28 2062 (2007). 71. S. V. Ahir, A. R. Tajbakhsh, E. M. Terentjev, Self Assembled Shape Memory Fibers of Triblock Liquid Crystal Polymers. Advanced F unctional Materials 16 556 (2006). 72. J. Naciri et al., Nematic Elastomer Fiber Actuator. Macromolecules 36 8499 (2003/11/01, 2003). 73. A. Buguin, M. H. Li, P. Silberzan, B. Ladoux, P. Keller, Micro actuators: When artificial muscles made of nematic liquid crystal elastomers meet soft lithography. Journal of the American Chemical Society 128 1088 (Feb 1, 2006). 74. C. Ohm, E. K. Fleischma nn, I. Kraus, C. Serra, R. Zentel, Control of the Properties of Micrometer Sized Actuators from Liquid Crystalline Elastomers Prepared in a Microfluidic Setup. Advanced Functional Materials 20 4314 (Dec 21, 2010). 75. C. Ohm, C. Serra, R. Zentel, A Contin uous Flow Synthesis of Micrometer Sized Actuators from Liquid Crystalline Elastomers. Advanced Materials 21 4859 (Dec 18, 2009). 76. C. Tschierske, Liquid crystal engineering new complex mesophase structures and their relations to polymer morphologies, na noscale patterning and crystal engineering. Chemical Society Reviews 36 1930 (2007). 77. A. Azoug et al., Viscoelasticity of the polydomain monodomain transition in main chain liquid crystal elastomers. Polymer, (2016). 78. A. H. Gelebart, M. Mc Bride, A P. Schenning, C. N. Bowman, D. J. Broer, Photoresponsive Fiber Array: Toward Mimicking the Collective Motion of Cilia for Transport Applications. Advanced Functional Materials, (2016). 79. B. Y. Zhang, L. M. Liu, Z. L. Feng, Y. G. Jia, Mesomorphic prope rties of liquid crystalline polysiloxanes and elastomers based on a tri vinyl crosslinker. Liquid crystals 30 1015 (2003). 80. K. A. Burke, I. A. Rousseau, P. T. Mather, Reversible actuation in main chain liquid crystalline elastomers with varying crossli nk densities. Polymer 55 5897 (2014). 81. Y. G. Jia, B. Y. Zhang, M. Tian, W. Pan, Synthesis and structure of polysiloxane liquid crystalline elastomers with a mesogenic crosslinking agent. Journal of applied polymer science 93 1736 (2004). 82. A. Tsuchi tani, H. Ashida, K. Urayama, Pronounced effects of cross linker geometries on the orientation coupling between dangling mesogens and network backbones in side chain type liquid crystal elastomers. Polymer 61 29 (2015). 83. W. Ren, P. J. McMullan, A. C. Griffin, Stress strain behavior in main chain liquid crystalline elastomers: effect of crosslinking density and transverse rod incorporation on "Poisson's ratio". physica status solidi (b) 246 2124 (2009). 84. Y. Li et al., C ontrolled shape memory behavior of a smectic main chain liquid crystalline elastomer. Macromolecules 48 2864 (2015). 85. J. KŸpfer, H. Finkelmann, Liquid crystal elastomers: influence of the orientational distribution of the crosslinks on the phase behavi our and reorientation processes. Macromolecular chemistry and physics 195 1353 (1994). 86. K. M. Lee, T. J. Bunning, T. J. White, Autonomous, Hands Free Shape Memory in Glassy, Liquid Crystalline Polymer Networks. Advanced Materials 24 2839 (2012). 87. A Hotta, E. Terentjev, Dynamic soft elasticity in monodomain nematic elastomers. The European Physical Journal E 10 291 (2003). 88. H. Qin, P. T. Mather, Combined one way and two way shape memory in a glass forming nematic network. Macromolecules 42 273 (2008). 89. D. Andrienko, Introduction to liquid crystals. IMPRS school, Bad Marienberg, (2006). 90. S. Clarke, A. Tajbakhsh, E. Terentjev, M. Warner, Anomalous viscoelastic response of nematic elastomers. Physical review letters 86 4044 (2001).

PAGE 118

! 99 91. D. L Safranski, K. Gall, Effect of chemical structure and crosslinking density on the thermo mechanical properties and toughness of (meth)acrylate shape memory polymer networks. Polymer 49 4446 (9/23/, 2008). 92. M. Mohiuddin, T. T. Tung, Nanomaterials Embed ded Liquid Crystal Elastomers in Electronics Devices Application. Flexible and Stretchable Electronic Composites, 365 (2015). 93. T. Ware, T. White, Programmed liquid crystal elastomers with tunable actuation strain. Polymer Chemistry 6 4835 (2015). 94. I W. Hunter, S. Lafontaine, in Solid State Sensor and Actuator Workshop, 1992. 5th Technical Digest., IEEE. (IEEE, 1992), pp. 178 185. 95. G. M. Spinks, V. T. Truong, Work per cycle analysis for electromechanical actuators. Sensors and Actuators A: Physic al 119 455 (2005). 96. C. M. Yakacki et al., Unconstrained recovery characterization of shape memory polymer networks for cardiovascular applications. Biomaterials 28 2255 (2007). 97. A. Lendlein, S. Kelch, Shape memory polymers. Angewandte Chemie Intern ational Edition 41 2034 (2002). 98. M. D. Lima et al., Electrically, chemically, and photonically powered torsional and tensile actuation of hybrid carbon nanotube yarn muscles. Science 338 928 (2012). 99. K. Otsuka, C. M. Wayman, Shape memory materials. (Cambridge university press, 1999). 100. Y. Yu, T. Ikeda, Soft Actuators Based on Liquid Crystalline Elastomers. Angewandte Chemie International Edition 45 5416 (2006). 101. C. Liu, H. Qin, P. Mather, Review of progress in shape memory polymers. Journal of Materials Chemistry 17 1543 (2007). 102. A. S‡nchez Ferrer et al., Photo Crosslinked Side Chain Liquid Crystalline Elastomers for Microsystems. Macromol. Chem. Phys. 210 1671 (2009). 103. H. Yang, G. Ye, X. Wang, P. Keller, Micron sized liquid crystalline elastomer actuators. Soft Matter 7 815 (2011). 104. E. Sungur et al., Temperature tunable optical gratings in nematic elastomer. Applied Physics A: Materials Science & P rocessing 98 119 (2010). 105. S. Schuhladen et al., Iris Like Tunable Aperture Employing Liquid Crystal Elastomers. Advanced Materials 26 7247 (2014). 106. A. S‡nchez Ferrer et al., Liquid Crystalline Elastomer Microvalve for Microfluidics. Advanced Mate rials 23 4526 (2011). 107. (!!! INVALID CITATION !!! ). 108. C. L. Gonzalez et al., Nanoporous Membranes of Hydrogen bridged Smectic Networks with Nanometer Transverse Pore Dimensions. Advanced Materials 20 1246 (2008). 109. A. H. Torbati, P. T. Mathe r, A hydrogel forming liquid crystalline elastomer exhibiting soft shape memory. Journal of Polymer Science Part B: Polymer Physics 54 38 (2016). 110. M. O. Saed et al., Thiol acrylate main chain liquid crystalline elastomers with tunable thermomechanical properties and actuation strain. Journal of Polymer Science Part B: Polymer Physics 55 157 (2017). 111. M. O. Saed, A. H. Torbati, D. P. Nair, C. M. Yakacki, Synthesis of Programmable Main chain Liquid crystalline Elastomers Using a Two stage Thiol acryl ate Reaction. e53546 (2016/01/19/, 2016). 112. M. Lee, Y. S. Yoo, Supramolecular organization of block oligomers based on rod shaped mesogen into liquid crystalline assembly. Journal of Materials Chemistry 12 2161 (2002). 113. R. H. Baughman et al., Carbo n nanotube actuators. Science 284 1340 (1999). 114. E. F. Crawley, J. De Luis, Use of piezoelectric actuators as elements of intelligent structures. AIAA journal 25 1373 (1987).

PAGE 119

! 100 115. A. Nespoli, S. Besseghini, S. Pittaccio, E. Villa, S. Viscuso, The high potential of shape memory alloys in developing miniature mechanical devices: A review on shape memory alloy mini actuators. Sensors and Actuators A: Physical 158 149 (2010). 116. G. A. Pratt, M. M. Williamson, in Intelligent Robots and Systems 95.'Human Robot Interaction and Cooperative Robots', Proceedings. 1995 IEEE/RSJ International Conference on. (IEEE, 1995), vol. 1, pp. 399 406. 117. S. Seelecke, I. Muller, Shape memory alloy actuators in smart structures: Modeling and simulation. Applied Mechanics Reviews 57 23 (2004). 118. A. Bezrouk, J. Hanus, J. Z‡hora, Temperature characteristics of nitinol spiral stents. Scr. Med.(Brno) 78 219 (2005). 119. A. Agrawal et al., Dynamic self stiffening in liquid crystal elastomers. Nature communications 4 1739 (2013). 120. C. M. Yakacki et al., Unconstrained recovery characterization of shape memory polymer networks for cardiovascular applications. Biomaterials 28 2255 (May, 2007). 121. I. Bellin, S. Kelch, R. Langer, A. Lendlein, Polymeric triple shape materia ls. Proc Natl Acad Sci U S A 103 18043 (Nov 28, 2006). 122. D. P. Nair et al., Two Stage Reactive Polymer Network Forming Systems. Advanced Functional Materials 22 1502 (2012). 123. H. Yang et al., Micron Sized Main Chain Liquid Crystalline Elastomer Act uators with Ultralarge Amplitude Contractions. Journal of the American Chemical Society 131 15000 (Oct 21, 2009). 124. G. H. F. Bergmann, H. Finkelmann, V. Percec, M. Y. Zhao, Liquid crystalline main chain elastomers. Macromolecular Rapid Communications 1 8 353 (May, 1997). 125. J. KŸpfer, H. Finkelmann, Nematic Liquid Single Crystal Elastomers. Makromol Chem Rapid 12 717 (Dec, 1991). 126. K. M. Lee, H. Koerner, R. A. Vaia, T. J. Bunning, T. J. White, Light activated shape memory of glassy, azobenzene liq uid crystalline polymer networks. Soft Matter 7 4318 (2011). 127. H. Yang et al., Synthesis and physical properties of a main chain chiral smectic thiol ene oligomer. Liq Cryst 37 325 (2010). 128. K. M. Lee, H. Koerner, R. A. Vaia, T. J. Bunning, T. J. White, Relationship between the Photomechanical Response and the Thermomechanical Properties of Azobenzene Liquid Crystalline Polymer Networks. Macromolecules 43 8185 (Oct 12, 2010). 129. J. W. Chan, C. E. Hoyle, A. B. Lowe, M. Bowman, Nucleophile initiated thiol michael reactions: effect of organocatalyst, thiol, and ene. Macromolecules 43 6381 (2010). 130. Z. Q. Pei et al., Mouldable liquid crystalline elastomer actuators with exchangeable covalent bonds. Nature Materials 13 36 (Jan, 2014). 131. D. Iqbal, M. H. Samiullah, Photo Responsive Shape Memory and Shape Changing Liquid Crystal Polymer Networks. Materials 6 116 (Jan, 2013). 132. M. Behl, A. Lendlein, Actively moving polymers. Soft Matter 3 58 (2007). 133. C. Liu, H. Qin, P. T. Mather, Review of progress in shape memory polymers. Journal of Materials Chemistry 17 1543 (2007). 134. C. Ohm, M. Brehmer, R. Zentel, in Advances in Polymer Science. (2012), vol. Liquid Crystal Elastomers: Material s and Applications. 135. C. M. Yakacki, K. Gall, in Advances in Polymer Science. (2010), vol. Shape Memory Polymers, pp. 147 175. 136. A. Lendlein, S. Kelch, Shape memory polymers as stimuli sensitive implant materials. Clinical Hemorheology and Microcircu lation 32 105 (2005). 137. W. Sokolowski, A. Metcalfe, S. Hayashi, L. Yahia, J. Raymond, Medical applications of shape memory polymers. Biomed. Mater. 2 S23 (Mar, 2007). 138. K. A. Burke, P. T. Mather, Soft shape memory in main chain liquid crystalline e lastomers. Journal of Materials Chemistry 20 3449 (2010).

PAGE 120

! 101 139. S. J. Woltman, G. D. Jay, G. P. Crawford, Liquid crystal materials find a new order in biomedical applications. Nat Mater 6 929 (12//print, 2007). 140. A. Sharma et al., Biocompatible, Biodeg radable and Porous Liquid Crystal Elastomer Scaffolds for Spatial Cell Cultures. Macromolecular Bioscience, n/a (2014). 141. J. Garcia Amor—s, M. Mart’nez, H. Finkelmann, D. Velasco, Photoactuation and thermal isomerisation mechanism of cyanoazobenzene bas ed liquid crystal elastomers. Phys. Chem. Chem. Phys. 16 8448 (2014). 142. H. Finkelmann, H. Wermter, in ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY. (AMER CHEMICAL SOC 1155 16TH ST, NW, WASHINGTON, DC 20036 USA, 2000), vol. 219, pp. U493 U493. 143. R. L. Selinger, B. L. Mbanga, J. V. Selinger, in Integrated Optoelectronic Devices 2008. (International Society for Optics and Photonics, 2008), pp. 69110A 69110A 5. 144. A. Agrawal et al., Surface wrinkling in liquid crystal elastomers. Soft Matter 8 7138 (2012). 145. H. Yang et al., Micron sized main chain liquid crystalline elastomer actuators with ultralarge amplitude contractions. J. Am. Chem. Soc. 131 15000 (2009). 146. H. Yang et al., Novel liquid crystalline mesogens and main chain chiral sm ectic thiol ene polymers based on trifluoromethylphenyl moieties. Journal of Materials Chemistry 19 7208 (2009). 147. Y. Xia, R. Verduzco, R. H. Grubbs, J. A. Kornfield, Well defined liquid crystal gels from telechelic polymers. J. Am. Chem. Soc. 130 173 5 (2008). 148. Y. Meng, J. Jiang, M. Anthamatten, Shape actuation via internal stress induced crystallization of dual cure networks. ACS Macro Letters 4 115 (2015). 149. H. Peng et al., High Performance Graded Rainbow Holograms via Two Stage Sequential Or thogonal Thiol Click Chemistry. Macromolecules 47 2306 (2014). 150. D. P. Nair et al., Two Stage Reactive Polymer Network Forming Systems. Advanced functional materials 22 1502 (2012). 151. A. A. Alzahrani et al., Photo CuAAC Induced Wrinkle Formation in a Thiol Acrylate Elastomer via Sequential Click Reactions. Chemistry of Materials 26 5303 (2014). 152. S. J. Ma, S. J. Mannino, N. J. Wagner, C. J. Kloxin, Photodirected Formation and Control of Wrinkles on a Thiol ene Elastomer. ACS Macro Letters 2 474 (2013). 153. T. Kamal, S. y. Park, Shape Responsive Actuator from a Single Layer of a Liquid Crystal Polymer. ACS applied materials & interfaces 6 18048 (2014). 154. D. P. Nair et al., The thiol Michael addition click reaction: a powerful and widely used tool in materials chemistry. Chemistry of Materials 26 724 (2013). 155. D. L. Safranski, K. Gall, Effect of chemical structure and crosslinking density on the thermo mechanical properties and toughness of (meth) acrylate shape memory polymer networks. Polymer 49 4446 (2008). 156. C. M. Yakacki, S. Willis, C. Luders, K. Gall, Deformation Limits in Shape Memory Polymers. Advanced Engineering Materials 10 112 (2008). 15 7. T. L. Smith, Ultimate tensile properties of elastomers. I. Characterization by a time and temperature independent failure envelope. Journal of Polymer Science Part A: General Papers 1 3597 (1963). 158. S. Clarke, A. Hotta, A. Tajbakhsh, E. Terentjev, E ffect of cross linker geometry on dynamic mechanical properties of nematic elastomers. Physical Review E 65 021804 (2002). 159. P. Martinoty, P. Stein, H. Finkelmann, H. Pleiner, H. R. Brand, Mechanical properties of mono domain side chain nematic elastom ers. The European Physical Journal E: Soft Matter and Biological Physics 14 311 (2004). 160. T. Sauter, M. Heuchel, K. Kratz, A. Lendlein, Quantifying the shape memory effect of polymers by cyclic thermomechanical tests. Polymer Reviews 53 6 (2013).

PAGE 121

! 102 161. S. Clarke, A. Hotta, A. Tajbakhsh, E. Terentjev, Effect of crosslinker geometry on equilibrium thermal and mechanical properties of nematic elastomers. Physical Review E 64 061702 (2001). 162. P. L. Crystals. (Academic Press, New York, 1982). 163. M. Tok ita, H. Tagawa, H. Niwano, K. Osada, J. Watanabe, Temperature induced reversible distortion along director axis observed for monodomain nematic elastomer of cross linked main chain polyester. Japanese journal of applied physics 45 1729 (2006). 164. A. Lea dbetter, E. Norris, Distribution functions in three liquid crystals from X ray diffraction measurements. Molecular Physics 38 669 (1979). 165. P. Davidson, D. Petermann, A. Levelut, The measurement of the nematic order parameter by x ray scattering recons idered. Journal de Physique II 5 113 (1995). 166. J. P. Lagerwall, F. Giesselmann, M. D. Radcliffe, Optical and x ray evidence of the "de Vries" Sm) A* Sm) C* transition in a non layer shrinkage ferroelectric liquid crystal with very weak interlayer tilt correlation. Physical Review E 66 031703 (2002).

PAGE 122

! 103 APPRNDIX A Wide Angle X Ray Scattering Characterizations Summary X ray diffraction is a powerful technique to identify microscopic structures and measure the orientation order parameter in LCEs. The method used to calculate the orientation order para meter of selected lightly cross linked sample is presented in thi s Appendix. Additional selected 1 and 2 D patterns from studies of the influence of the cross linking (chapter III) and spacer (chapter IV ) on the LC phase structure are also presented in this Appendix. Orientational Order: Numerical analysis of the scattering profile I( $) is done to calculate the orientational order parameter. Leadbetter et al. showed the relation between scattering profile I( $) and orientational distribution function f(%) in their famous paper. ( 164 ) ! !"# !"# ! !"# !"# ! ! !" (A.1) Figure A.1 Scattering profile I(* ) as a function of azimuthal angle in the wide angle regime

PAGE 123

! 104 Davidson et al. showed that the scattering profile I( $) can also be written as a series of cos 2n functions. ( 165 ) ! ! ! ! !"# ! (A.2) w here the fitting parameter f 2n is determined numerically by fitting Eq. 2 to the experimentally determined scattering profile I( $) Figure. S6 shows the best fit of Eq. (A2) to the data. The fitting parameter f 2n is then used to calculate the orientational distribution function f(%) given by the equation below: ! !"# ! ! (A3) The expectation value of the inclination angle % and orientational order parameter S 2 is calculated using the following equation [A 3]: ( 166 ) Figure A2 ! !"#$ is numerically calculated by fitting Eq. S2 to the experimental data. The average inclination angle % and the orientational order parameter S 2 are provided in the inset.

PAGE 124

! 105 ! ! !"#$ !" ! ! !"#$ !" ! (A4) where the average inclination angle and orientational ord er parameter are obtained by substituting X in Eq. (4) with % and !"# ! 2, respectively. By using numerical analysis, the average inclination angle % is 19.51, and the orientational order parameter S 2 ~ 0.80 The influence of cross linking on the LC phase s tructure WAXS analysis was performed in order to investigate the short range and long range order of the LCEs. The 1D WAXS patterns of the LCEs were generated when the samples were strained at 0 and 80% to measure the samples in unoriented (polydomain) an d oriented (monodomain) states, respectively. For unoriented samples at 0% strain, no clearly defined peaks were observed in the 1D plot of intensity versus azimuth for tri thiol (Figure A .3.a) and tetra thiol LCEs (Figure A .3b); conversely, oriented samples at 80% strain all revealed periodic peaks separated by 180 ¡ which is indicative of nematic order a monodomain structure. It is important to note that all LCEs tested Figure A.3 The 1D WAXS patterns for eight LCE networks with varying the concentration and functionality of the crosslinker. The intensity versus azimuth for (a) tri thiol and (b) tetra thiol LCEs were measured at fixed strains of 0 and 80%.

PAGE 125

! 106 exhibited similar scattering patterns, indicating that the microstructure of the LC domains was not affected by crosslinker concentration or functionality i n this system. Figure A.4. 1 and 2 D WAXS patterns for five LCE networks at room temperature. Diffraction was measured in an aligned monodomain state. Alignment was achieved by stretching the samples to 100% engineering strain before analysis.

PAGE 126

! 107 The influence of sp acer on the L C phase structure Wide angle X ray scattering (WAXS) was used next to investigate the influence of spacer length on the mesophase of the materials ( Figure A .4 ). C6, C9, and C11 revealed a peak at the smectic region (~0.24 A 1 ). The intensity of the peak increased with increasing the length of the spacer. Interestingly, the spacing between the layers enlarged linearly with increasing of the length of the spacers. Both C2 and C3 did not appearance any peak at the smectic region. The evolution of the mesopha se of C6 and C11 was studied as a function of temperature ( Figure A .5 and 6). C6 bared a SmC to N transition c above 50 ¡ C, where C11 did demonstration a SmC to N transition within the tested temperature rang. Figure A.5. Temperature controlled WAXS analysis of the LCE system using the C6 spacer. Diffraction patterns reveal the present of the smectic c phase over broad range of temperature. All images were taken under 100% engineeri ng strain. Figure A.6 Temperature controlled WAXS analysis of the LCE system using the C11 spacer. Diffraction patterns reveal the transition from a smectic C to nematic orientation when heated above 50 ¡ C. All images were taken under 100% engin eering strain.

PAGE 127

! 108 B Differential Scanning Calorimetry (DSC ) S ummary DSC is a very useful technique to characterize the LC and polymer phase transition temperatures. Additional selected studies of 1s t and 2nd heating from chapter IV the influence spacer on the LC phase structure and transition temperatures are pr esented in this Appendix. Monodonain LCEs were prepared via two stage thiol acrylate Michael addition reaction (Stage 1) and photo polymerization reaction (Stage 2). The DSC trance of LCE before and after the 2 nd stage reaction is shown in in this Appendix. Figure B .1. Heat flow from standard DSC of the C2, C3, C6, and C11 In all plots, the networks were first heated to the isotropic state (T NI + 30¡C) at 10¡K/min and then cooled to 80 ¡K at 2 ¡K/min. Then heated again at 20 ¡C/min. The mel ting transition temperatures were measured at the first heating scan where the LC transition temperatures were measured at the sound heating scan.

PAGE 128

! 109 The influence of spacer on the LC phase structure DSC was used to initially characterize the phase transitions in the LCE networks ( Figure B .1 ). The first heating scan displayed two or more endothermic wells corresponded to melting temperature and one or two LC transition temperatures. The first (from the left) endothermic well for the all networks is thought to b e link to melting temperature of the polymer chain semi crystallinity. C2 and C6 had an extra endothermic well related to melting temperature. While we are not sure why did these materials have two melting transitions but this behavior can be explained ba sed on the self assembly theory of the nano scale segregation of ternary incompatible layered. The molecular structure of the mesogen (RM257) is containing mesogen core and propylene oxide acrylic terminal chains. While propylene oxide is miscible C3 spacer and not miscible with C2 and C6 spacers Therefore, the resulting LCE of C3 had one melting transition well where C2 and C6 had two melting transitions well. The second heating scan exhibited only LC transitions without melting transition due to slow recrystallization. C2 and C3 revea led one endothermic well, which corresponded to the nematic to isotropic transition. C6 and C11 had an extra endothermic well. This advised that a second smectic phase is present in the materials. The trend of the transition temperature as a function of th e spacer length is shown in Figure B .2. The nematic transition (T NI ) increased with decreased spacer length, Figure B.2 LC transition temperatures from the DSC traces of five LCE networks with difference spacer length. Smectic to nematic (T SN ) and nematic to isotropic (T NI ) transition temperatures were measured at the second heating scan at 20 K/min

PAGE 129

! 110 where the smectic to nematic transition temperature (T SN ) showed an opposite trend as it decreased with decreasing the spacer length until complete ly disappearing at shorter spacers (C2 and C3). DSC trances of Monodomain LCE before and after the 2 nd reaction DSC was also performed using a TA Instruments Q2000 machine. Samples with a mass of approximately 10 mg were loaded into a standard aluminum DSC pan. The samples were heated rapidly to 125¡C at 10¡C /min, held isothermally for 3 min, and cooled slowly to 80¡C at rate of 2¡C /min to reset any thermal history within the sample. Samples were then heated to 125¡ C at a rate of 10 at 10¡C/min. T I w as defined as the minimum value of the endothermic peak. All samples were run in triplicate (n=3) and representative curves can be seen in Figure B .3 The T I was difficult to detect using DSC for the Stage 2 samples. A broad and weak transition appears to occur at approximately 70C and is best illustrated in Figure A3c. Figure B. 3. Polydomain LCE samples are tested after completion of Michael addition reaction (Stage 1) and photo polymerization reaction (Stage 2). (b) Heat flow is plotted as a function of temperature to identify isotropic transition. (c) An expanded view of heat flow as a function of temperature for the LCE samples shown in (b).

PAGE 130

! 111 Dynamic Mecha nical Analysis (DMA ) : D MA is a powerful technique to study the evolution of thermo mechanical properties in LCEs. Additional selected studies of the influence of the cross linking (chapter III) and spacer (chapter IV ) on LCE mechanical behavior are shown in this Appendix. Monodonain LCEs were prepared via two stage thiol acrylate Michael addition reaction (Stage 1) and photo polymerization reaction (Stage 2). The DMA trance of LCE before and after the 2 nd stage react ion is shown in in this Appendix The influence of crosslinker concentration and functionality on dynamic behavior is shown in (Figure C .1). Representative plots for tri thiol networks ( Figure C.1.a ) and tetra thiol networks (Figure C.1. b) compare storage modulus ( E' ) and loss tangent ( tan ) of LCEs as a function of temperature. All LCE networks demonstrate a glassy plateau below 0¡C followed by a stepwise decrease in E' that corresponds with the onset of the glass transition ( T onset ) for each network. In this study, T g was measured at the maximum of the tan curve and ranged from 3 to 25¡C. There was a noticeable change in slope and concavity in the E' for the LCE networks at Tg. Furthermore, all LCE networks demonstrated a dramatic decrease in E' at Ti, a phenomenon often termed "dynamic soft elasticity". The samples recover to a rubbery plateau as they are heated into the isotropic phase. It is important to notice that tan of the networks remain elevated within the nematic region (i.e. between Tg and Ti). While some of the representative curves may Figure C. 1. Storage modulus (E#) and loss tangent (tan !) traces for eight thiol acrylate LCE networks measured at 3 ¡C/min heating rate and 1 Hz frequency in tension mode, the second temperature sweep plot for a) tri thiol LCE networks and b) tetra thiol LCE networks. The glass transition temperature (T g ) was measured at the peak of tan

PAGE 131

! 112 show one or two secondary peaks in tan !, this behavior was not consistent across all samples tested; however, all samples demonstrated tan to decrease to a near zero value once heated above Ti. The advancement of thermo mechanical properties as a function of spacer length and polymer chain crystallization was studied using DMA Storage modulus (E# ) and loss tangent (tan !) traces for the C2 and C3 are shown in Figure C .2. A distinct thermo mechanical behavior for the first heating compared to the rest of the heating scans due to the melting of the semi crystalline structure during the first heat ing scans. Monodonain LCEs were prepared via two stage thiol acrylate Michael addition reaction (Stage 1) and photo polymerization reaction (Stage 2). All samples were run in triplicate (n=3) and representative curves of DMA before and after the 2nd stage can be seen in Figure C .3. It should be noted that the tan $ functions of the samples were much broader compared to traditional elastomeric materials, which typically display a relatively symmetric single peak. Furthermore, both Stage 1 and S tage 2 LCE samples exhibited a distinct dip (i.e. minimum) in the rubbery modulus regime, which was appears to be associated with Ti. This behavior has previously been reported by Warner and Terentjev when testing LCEs under relatively low frequencies. ( 10 ) Figure C.2 Storage modulus (E#) and loss tangent (tan delta) traces for LCE networks with spacer lengths of C2 and C3. Samples were measured at 3¡C/min heating rate and 1 Hz frequency in tension. All samples were annealed above T NI and allowed to cool at room temperature for 24 hours before the first temperature sweep to allow the semi crystallinity to fully form. Samples were tested four times and allowed to set isothermally at 25 ¡C between each sweep for 5, 60, and 120 minutes to show the evolution of the mechanical properties due to polymer chain crystallization.

PAGE 132

! 113 Figure C.3. Polydomain LCE samples are tested after completion of Michael addition reaction (Stage 1) and photo polymerization reaction (Stage 2). (a) Storage modulus and tan $ are plotted as a function of temperature to characterize glass transition