self-healing and vitrimeric polymers based on dynamic-covalent boronic esters · 2017. 9. 21. ·...
TRANSCRIPT
SELF-HEALING AND VITRIMERIC POLYMERS BASED ON DYNAMIC-COVALENT BORONIC ESTERS
By
JESSICA J. CASH
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2017
© 2017 Jessica J. Cash
To my family, who have provided constant support
4
ACKNOWLEDGMENTS
Many thanks to my friends and fellow graduate students, past and present, who
have been there to encourage when experiments failed and celebrate when they
succeeded. I would especially like to thank Soma Mukherjee, Mayra Rostagno,
Yuqiuong Dai (Daily), Sandhya Guntaka, and Erica Amato.
This material is based upon work supported by the National Science Foundation
(DMR-1410223 and DMR-1606410). J.J.C. thanks the DOD for a Science Mathematics
and Research for Transformation Fellowship.
5
TABLE OF CONTENTS page
ACKNOWLEDGMENTS .................................................................................................. 4
LIST OF TABLES ............................................................................................................ 7
LIST OF FIGURES .......................................................................................................... 8
LIST OF SCHEMES ...................................................................................................... 10
LIST OF ABBREVIATIONS ........................................................................................... 11
ABSTRACT ................................................................................................................... 13
CHAPTER
1 INTRODUCTION .................................................................................................... 15
1.1 Diels-Alder Linkages ......................................................................................... 16 1.2 Carboxylate Transesterification ........................................................................ 20 1.3 Sulfur-based Functional Groups ....................................................................... 24
1.4 Nitrogen-based Functional Groups ................................................................... 30 1.5 Other Dynamic-Covalent Linkages ................................................................... 36
1.6 Boron-based Functional Groups ....................................................................... 39
2 RESEARCH OBJECTIVE ....................................................................................... 43
3 ROOM-TEMPERATURE SELF-HEALING POLYMERS BASED ON DYNAMIC-COVALENT BORONIC ESTERS ........................................................................... 46
3.1 Overview ........................................................................................................... 46
3.2 Experimental Section ........................................................................................ 50 3.2.1 Materials .................................................................................................. 50
3.2.2 Instrumentation and Analysis .................................................................. 51 3.2.3 Synthesis and Experimental Procedures ................................................. 52
3.3 Results and Discussion ..................................................................................... 56
3.3.1 Monomer Synthesis and Reversibility...................................................... 56 3.3.2 Network Formation .................................................................................. 57
3.3.3 Network Characterization ........................................................................ 59 3.3.4 Self-healing ............................................................................................. 61
3.4 Conclusions ...................................................................................................... 65
4 BALANCING STATIC AND DYNAMIC BONDS IN SELF-HEALING BORONIC ESTER NETWORKS .............................................................................................. 67
4.1 Overview ........................................................................................................... 67
6
4.2 Experimental Section ........................................................................................ 71
4.2.1 Materials .................................................................................................. 71 4.2.2 Instrumentation and Analysis .................................................................. 71
4.2.3 Synthesis and Experimental Procedures ................................................. 72 4.3 Results and Discussion ..................................................................................... 74
4.3.1 Networks with Boronic Esters and Free Diols .......................................... 74 4.3.2 Networks with Permanent Crosslinks and Boronic Esters ....................... 80 4.3.3 Combined Free Diol and Permanent Crosslinker Networks .................... 84
4.4 Conclusions ...................................................................................................... 87
5 BORONIC ESTER VITRIMERS .............................................................................. 89
5.1 Overview ........................................................................................................... 89 5.2 Experimental Section ........................................................................................ 91
5.2.1 Materials. ................................................................................................. 91 5.2.2 Instrumentation and Analysis. ................................................................. 92
5.2.3 Synthesis and Experimental Procedures ................................................. 92 5.3 Results and Discussion ..................................................................................... 99
5.3.1 Synthesis and Comparison of Relative Hydrolysis and Transesterification of Model Boronic Esters .................................................. 99
5.3.2 Preparation and Characterization of Dithiol Oligomer and Networks ..... 103
5.3.3 Dynamics of the Networks ..................................................................... 104 5.4 Conclusions .................................................................................................... 108
6 CONCLUSIONS AND FUTURE DIRECTIONS .................................................... 109
APPENDIX
A: DSC RESULTS OF INITIAL SELF-HEALING NETWORK COMPOSITIONS ......... 111
B: RELAXATION TIMES OF FREE DIOL, PERMANENT CROSSLINKED, AND COMBINED NETWORKS ..................................................................................... 112
LIST OF REFERENCES ............................................................................................. 113
BIOGRAPHICAL SKETCH .......................................................................................... 124
7
LIST OF TABLES
Table page 4-1 DSC results of low glass transition temperature polymers showing values
below room temperature. .................................................................................... 75
5-1 Hydrolysis constants measured by 1H NMR ....................................................... 95
5-2 Tg values measured by DMA from the tan delta peak ...................................... 104
5-3 Summary of the relaxation times of 5-BE networks .......................................... 105
5-4 Summary of the relaxation times of 6-BE networks .......................................... 107
A-1 DSC results of low glass transition temperature polymers showing Tg values below room temperature. .................................................................................. 111
B-1 Relaxation times from tensile stress-relaxation measurements ........................ 112
8
LIST OF FIGURES
Figure page 1-1 Diels-Alder molecules based on cyclopentadiene. ............................................. 19
1-2 Healing of photo-locking Diels-Alder polymers. .................................................. 20
1-3 Changes in topology during transesterification bond exchange.......................... 22
1-4 The affect of zinc catalyst on transesterification. ................................................ 24
1-5 Thiol-disulfide exchange ..................................................................................... 26
1-6 Pre-orientation of the disulfide bond easing exchange ....................................... 28
1-7 Hindered ureas made more easily reversible by a bulky substituent .................. 34
1-8 Noncatalytic Michael addition ............................................................................. 35
1-9 Crosslink conservation through Ru-catalyzed olefin metathesis. ........................ 37
1-10 Diarylbibenzo-furanone dissociation ................................................................... 39
1-11 Conductive composite made from Ag-nanowires in a boroxine polymer network ............................................................................................................... 41
3-1 Diene boronic ester ............................................................................................ 57
3-2 Synthesis of boronic ester network materials via photoinitiated thiol-ene curing. ................................................................................................................. 58
3-3 FTIR spectra of the boronic ester diene, tetrathiol (PTMP), the solution prior to crosslinking, and the final crosslinked network. .............................................. 58
3-4 Water absorption of disk shaped samples of boronic ester-crosslinked network materials . . completely submerged in water as a function of time. ....... 59
3-5 Water contact angles as a function of time for a boronic ester-crosslinked network . . . equilibrated at 85, 75, or 23% humidity measured at a series of times after removal from respective humidity chambers. .................................... 60
3-6 Self-healing of boronic ester network proposed mechanism of healing. ............. 62
3-7 Self-healing of boronic ester network materials as evaluated by tensile testing. ................................................................................................................ 64
3-8 Self-healing of boronic ester network materials after multiple cycles of damage and repair as evaluated by tensile testing. ............................................ 65
9
4-1 An illustration of the three prepared systems ..................................................... 70
4-2 Percent mass change of the network that contained 5% free diol while immersed in water. ............................................................................................. 76
4-3 Stress relaxation of 0, 1, 3, and 5% free diol networks after exposure to humidity. ............................................................................................................. 78
4-4 Evaluation of healing by tensile testing after exposure to humid environments. ..................................................................................................... 79
4-5 The effect of water on permanently crosslinked samples ................................... 81
4-6 Stress relaxation data for 80, 85, and 90% permanent crosslinker networks ..... 83
4-7 Self-healing of boronic ester networks as evaluated by tensile testing. .............. 84
4-8 Water absorption of samples containing both 5% free diol and 80% permanent crosslinker. ....................................................................................... 85
4-9 Stress relaxation curves run in duplicate of samples containing 5% free diol, 80% permanent crosslinker, and both ................................................................ 86
4-10 Healing measured after 3 days by tensile testing for samples containing 5% free diol, 80% permanent crosslinker, or both .................................................... 86
5-1 ln(x) ln(x0) vstime for the hydrolysis of 0.1 M SRBE with 10 M D2O in d-DMSO, at 25 ˚C. 96
5-2 Reagents for subsequent studies ....................................................................... 99
5-3 The overall hydrolysis. ...................................................................................... 101
5-4 NMR transesterification with inverted and referenced protons labeled. ............ 102
5-5 Stress-relaxation of 5-BE networks .................................................................. 105
5-6 Stress relaxation of 6-BE networks ................................................................... 107
10
LIST OF SCHEMES
Schemes page 1-1 Examples of Diels-Alder linkages used in bulk dynamic-covalent polymeric
materials. ............................................................................................................ 17
1-2 Two types of carboxylate-based transesterification used in vitrimers, a subset of dynamic-covalent polymeric materials. ........................................................... 21
1-3 Types of sulfur-based exchange reactions used in bulk dynamic-covalent polymeric materials. ............................................................................................ 25
1-4 Types of nitrogen-based exchange reactions used in bulk dynamic-covalent polymeric materials with the exception of oximes, which were determined to be irreversible in the absence of solvent. ........................................................... 31
1-5 Other types of exchange reactions used in bulk dynamic-covalent polymeric materials. ............................................................................................................ 36
1-6 Types of boron-based exchange reactions used in bulk dynamic-covalent polymeric materials. ............................................................................................ 40
11
LIST OF ABBREVIATIONS
APD 3-Allyloxy-1,2-propanediol
ASTM American Society for Testing and Materials
ATR Attenuated Total Reflectance
BE Boronic Ester
CAN Covalently Adaptable Network
DSC Differential Scanning Calorimeter
DABBF Diarylbibenzofuranone
DART Direct Analysis in Real Time
DCM Dichloromethane
D2O Deuterium oxide
DODT 3,6-Dioxa-1,8-octanedithiol
DMA Dynamic Mechanical Analyzer
DMPA 2,2-Dimethoxy-2-phenylacetophenone
DMSO
EPR
FRBE
FTIR
HRMS
IPDI
LBADSA
Mn
NMR
PTMP
SEC
Dimethylsulfoxide
Electron Paramagnetic Resonance
Five-membered Ring Boronic Ester
Fourier Transform Infrared
High-Resolution Mass Spectrometry
Isophorone diisocyanate
Low-Bond Axi-symmetric Drop Shape Analysis
Number Average Molecular Weight
Nuclear Magnetic Resonance
Pentaerythritol tetrakis(3-mercaptopropionate)
Size Exclusion Chromatography
12
SEM
SRBE
TEA
TEGDVE
TEMPO
TEMPS
Tg
THF
TMS
TOF-MS
Tv
UV
Vis
VPBA
VPBE
Scanning Electron Microscope
Six-membered Ring Boronic Ester
Triethylamine
Tri(ethylene glycol) divinyl ether
2,2,6,6-Tetramethylpiperidine-1-oxyl
Tetramethylpiperidine-1-sulfanyl
Glass transition temperature
Tetrahydrofuran
Tetramethylsilane
Time-of-Flight Mass Spectrometry
Topology freezing temperature
Ultraviolet
Visible
4-Vinylphenylboronic acid
4-((Allyloxy)methyl)-2-(4-vinylphenyl)-1,3,2-dioxaborolane
13
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
SELF-HEALING AND VITRIMERIC POLYMERS BASED ON DYNAMIC-COVALENT
BORONIC ESTERS
By
Jessica J. Cash
May 2017
Chair: Brent Sumerlin Major: Chemistry
With the growing interest in self-healing, recyclable, and malleable materials,
much effort has been devoted to exploring dynamic-covalent networks. Unlike other
dynamic chemistries, boronic esters have been utilized in polymers for drug delivery,
sensors, and responsive hydrogels, but not in the absence of solvent immersion. We
prepared thiol-ene networks with boronic esters. The reversibility of boronic esters in the
presence of gaseous and liquid water allowed hydrolysis to be used as an associative
mechanism to produce self-healing in bulk materials with good efficiency at room
temperature.
Given the propensity for boronic ester networks to gradually creep on exposure
to humidity in ambient conditions and the necessity for shape stability in many
applications, we proceeded to investigate ways to minimize creep and stress relaxation
by preparing boronic ester networks with various quantities of free diol, to promote
associative exchange, and permanent crosslinker, for an irreversible framework.
Network stability was probed with prolonged water immersion, stress-relaxation
experiments, and tensile testing after healing under humid and dry conditions. Samples
with a combination of free diol and permanent crosslinker had improved stability while
14
retaining better healing efficiency compared to samples with only free diol or permanent
crosslinker. These results could be used to design more stable polymer networks with
other reversible chemistries.
Vitrimers, strong organic glasses made from dynamic bonds with an associative
mechanism of exchange, have recently emerged as category of interest. We
synthesized the first boronic ester vitrimers. Using a stiffer dithiol oligomer to make thiol-
ene networks, the glass transition temperature of the boronic ester samples was raised
above room temperature producing slower molecular motion within the polymer. Two
different boronic esters were investigated for hydrolytic stability and rates of
transesterification by NMR spectroscopy with model compounds and through stress
relaxation experiments of corresponding bulk networks. Both types of boronic ester
networks fitted well to the Arrhenius equation, suggesting temperature solely affecting
exchange rate and not the equilibrium of the reversible linkages.
The research described here demonstrates the benefits of having reversible
bonds in the bulk polymers, and similar strategies can be used to design dynamic
materials tuned for other applications.
15
CHAPTER 1 INTRODUCTION
In recent years, bulk dynamic-covalent polymers are being recognized more and
more for their broad potential to address drawbacks of current polymeric materials and
for their capabilities to function in novel ways beyond the scope of static molecular
structures. As an example of the first point, they are being used to make materials that
can be healed and be recycled. Regarding the second point, shape memory,1
composite fabrication after the polymer network synthesis,2 nano-imprinting of already
formed polymer networks,3 and humidity-driven actuation4 are being explored.
The optimal dynamic chemistry for such a material changes depending on the
demands of the specific application. For example, self-healing autonomously at room
temperature requires a chemical linkage that either exchanges at room temperature or
that dissociates with mechanical force and reforms at room temperature. In contrast, the
goal with many recyclable systems is to create a material with the same properties as
irreversibly crosslinked networks in typical-use conditions and only have exchange and
reprocessing upon the application of a stimuli. Likewise, hydrolytic degradation,
although frequently thought of as a disadvantage, can be an advantage when
judiciously used to design environmentally friendly adhesives2 and reprocessable
materials.5
Broadly, the ideal mechanism of dynamicity also varies depending on the desired
application. For instance, phase change materials need a dissociative mechanism6
while vitrimers, strong organic glass-formers, rely on an associative mechanism.7
Neither would be possible with the mechanism of the other.
16
Furthermore, the polymer topology, also chosen for specific applications, can
influence how the dynamic chemistry behaves and thus, the selection of the dynamic
chemistry. For example, polymers for structural materials are highly crosslinked to
ensure strength, durability, and solvent resistance8 and often have higher glass
transition temperatures. In contrast, elastomers have low crosslink density and lower
glass transition temperatures. However, it’s important to consider that the structure of
the main polymer chain can affect the dynamic chemistry. Otsuka and coworkers
recently showed that the architecture around the dynamic bond can change the
activation of the dynamic bond.9 They found that higher molecular weight and more
connections near the site of a dynamic bond made it more susceptible to homolytic
cleavage, all conditions being equal. Mechanochemical affects also need to be
considered in bulk conditions for exchange reactions that would not typically be thought
of as being mechanically affected. In another earlier work, free furans and maleimides
were known to be present at the surface of cracks, well below the normal retro Diels-
Alder reaction temperature,10 likewise, for anthracene maleimide Diels-Alder in polymer
networks.11
This chapter presents a review of the dynamic-covalent chemistries used in bulk
polymeric materials focusing on the conditions for their dynamic behavior.
1.1 Diels-Alder Linkages
Diels-Alder chemistry was one of the first dynamic covalent reactions used to
design polymers with responsive behavior in bulk.12,13 Since the first system, materials
with a variety of dienes and dienophiles have been explored in the absence of solvent
(Scheme 1-1).8,11,12,14-16 General characteristics of the Diels-Alder chemistries in all
these systems include atom conservation for the forward and reverse reactions, a
17
dissociative mechanism, catalyst-free reversibility, tolerance to oxygen, and with a few
exceptions, due to the side reactions of the dissociated molecular units,17 hydrolytic
stability. Depending on the desired application for the material, the temperature of the
retro reaction can be tailored by the selection of the dienes and dienophiles. This
temperature can range from room temperature in bulk14 to 200 ˚C in solution,18 which is
beyond the degradation temperature of many polymers.
Scheme 1-1. Examples of Diels-Alder linkages used in bulk dynamic-covalent polymeric materials.
The earliest reversible Diels-Alder chemistry used for bulk polymers was the
furan maleimide chemistry published by Wudl and coworkers.12 A tetra-functional furan
was crosslinked with a tri-functional maleimide. The crosslinked network was examined
for reaction completion by UV-vis spectroscopy, solid state reversibility by DSC and
solid state 13C NMR, healing capability by tensile testing and SEM imaging, and
oxidative degradation by heating in air followed by compression testing. The efficiency
of the forward reaction varied from 60% to 70% at 24 ˚C after 5 days up to 95% ± 5%
18
at 75 ˚C after 3 hours. For the reverse reaction, 25% of the Diels-Alder adducts were
calculated to have reversed after cycling to 150 ˚C and none when cycled below 120 ˚C.
An average healing efficiency of approximately 50% was obtained after healing for 2 h
at 150 ˚C. This was a significantly better healing than previous thermoplastic efforts.19
Numerous systems since have used similar furan maleimide dynamic chemistries.
Following this work, a unique single component cyclopentadiene Diels-Alder
system was subsequently published by Wudl and other coworkersy8 based on a
previous observation of insolubility by Stille,20 which turned out to be due to crosslinking
at the double bond of norbornene. The use of a single component system automatically
addressed concerns about sample uniformity. Two dicyclopentadiene monomers were
presented. Depending on the monomer, polymerization was done by heating to either
150 ˚C or 120 ˚C for 10 h to open the preformed cyclic dicyclopentadiene monomers
and then slowly cooling to room temperature to form a crosslinked network (Fig. 1-1).
Reversibility was stated to occur at 120 ˚C, and healing experiments were performed at
this temperature.
Another notable type of Diels-Alder linkages encountered in bulk polymers are
anthracene maleimide.11 The forward reaction of anthracene maleimide Diels-Alder
reaches completion of greater than 95%, as measured by UV-vis spectroscopy, after 3
days, 2 h, or 1 h, respectively, at room temperature, 60 ˚C, or 100 ˚C. Anthracene
maleimide Diels-Alder reversibility, in solution, requires temperatures greater than 200
˚C.18 Many polymers degrade at these temperatures.11 Anthracene maleimide linkages
in polymer networks showed reversibility at lower temperatures when exposed to forces
19
Figure 1-1. Diels-Alder molecules based on cyclopentadiene. A) The scheme of linear
polymer formation. B) One of the cyclopentadiene monomers and a unit of the molecular X-ray crystal structure. C) A trimer representative of the crosslinks in the polymer network and a unit of the molecular X-ray crystal structure. Adapted with permission from Murphy, E. B.; Bolanos, E.; Schaffner-Hamann, C.; Wudl, F.; Nutt, S. R.; Auad, M. L. Macromolecules 2008, 41, 5203-5209. Copyright 2008 American Chemical Society.
causing damage. After damage, by finger tapping at room temperature, free anthracene
was observed with UV-vis spectroscopy at room temperature. If cut surfaces were
immediately placed back in contact with each other, the best recovery at room
temperature was 14% of tensile strength and 34% of elongation at break. However,
better healing efficiencies of up to 55% in tensile strength and 90% in elongation at
break could be obtained after healing at 100 ˚C for 3 days.
Various types of hetero Diels-Alder linkages have also been explored.
Cyclopentadiene dithioester Diels-Alder networks were formed from protected
monomers by heating to 120 ˚C for the deprotection retro Diels-Alder reaction to take
place.14 Complete reversibility for this system is measured at just above 170 ˚C.
Sufficient linkages were opened at 120 ˚C after 10 min with 1 kN of pressure applied
that upon cooling healing efficiencies of 106% and 96% for the first and second healing
20
cycles, respectively, were produced. Another type of hetero Diels-Alder linkages was
noted for reversibility at room temperature.15
Further advances in Diels-Alder chemistry for bulk polymers continue to be
investigated. Recent work showcased a design for turning on and off furan maleimide
chemistry through the incorporation of a molecular photo-switch (Fig. 1-2).16
Figure 1-2. Healing of photo-locking Diels-Alder polymers. A) The scheme of the photo-
switching thermo-responsive Diels-Alder polymers healing. B) Pictures of the films healing when switched on (top) or not healing when switched off (bottom). Adapted from Fuhrmann, A.; Gostl, R.; Wendt, R.; Kotteritzsch, J.; Hager, M. D.; Schubert, U. S.; Brademann-Jock, K.; Thunemann, A. F.; Nochel, U. N.; Behl, M.; Hecht, S. Nat. Commun. 2016, 7, 1-7. Copyright 2016 Rights Managed by Nature Publishing Group.
1.2 Carboxylate Transesterification
Although also an old chemistry, in direct contrast to Diels-Alder chemistry,
carboxylate transesterification has been highlighted in recent work, especially in vitrimer
applications, because of its associative mechanism, which allows crosslink conservation
21
(Fig. 1-3). Temperature only affects rate of exchange, in this case, and not the number
of dynamic linkages. The constant crosslink density makes these materials insoluble
regardless of temperature. Carboxylate transesterification is tunable by catalyst
incorporation. The two major types of carboxylate transesterefications studied in
vitrimers are epoxy-derived and urethanes (Scheme 1-2). Both are applicable
commercially.
Scheme 1-2. Two types of carboxylate-based transesterification used in vitrimers, a
subset of dynamic-covalent polymeric materials.
Epoxy-derived networks in bulk with carboxylate transesterification have been
studied generally in context of healing21,22 and reprocessability.5,23 Unique to the
associative mechanism, a slow change in viscosity with temperature allows welding
without molds, unlike thermoplastics.21,22 By studying model reactions, Leibler and
coworkers found that transesterification achieved equilibrium after ~15 h at 150 ˚C.21
Metal salts lowered the time to 2 h. Increased quantity of catalyst improved welding of
the bulk material in the same time. Above the glass transition temperature, the same
recovery of tensile strength could be achieved at a lower temperature, if the time was
adjusted to be longer. The free hydroxyl groups were necessary for exchange to occur.
22
A similar system was further explored for multiple cycles of grinding and reprocessing
(Fig. 1-3).22 The initial samples with this system showed the dynamic behavior of stress
relaxation at 180 ˚C, but behaved like permanently crosslinked elastomers at 80 ˚C.
Thus, significant exchange was not occurring at 80 ˚C. After 30 min at 180 ˚C with a
small amount of pressure to ensure particle contact, a previously pulverized sample
again appears to be a transparent undamaged sample. If the pressure is 90 kPa, the
reprocessed sample had mechanical properties similar to the initial sample. While some
property loss from damage to irreversible bonds took place, a 4th cycle sample was able
to achieve ultimate stretch comparable to a fresh sample.
Figure 1-3. Changes in topology during transesterification bond exchange: left (before),
middle (intermediate during exchange), right (after). Adapted from Yu, K.; Taynton, P.; Zhang, W.; Dunn, M. L.; Qi, H. J. RSC Adv. 2014, 4, 10108-10110. Copyright 2014 The Royal Society of Chemistry.
23
A slightly different bio-derived epoxy transesterification material was made by
Altuna and coworkers.22 The exchange rate in this system without catalyst appeared to
be comparable to the previous metal catalyzed system. The faster rate could be
attributed to the presence of more free hydroxyls and lower crosslink density. The
healed samples from this bio-derived epoxy attained the same stress and strain levels
as the original samples after healing after 2 h at 160 ˚C. While this system is important,
the ease of adding metal catalysts to current commercial systems remains an area of
interest to the scientific community.
Tournilhac and coworkers recently more closely examined the role of Zn2+ as a
catalyst in the mechanism of transesterification (Fig. 1-4).23 As can be seen in Figure 1-
4B, the general mechanism for transesterification in basic conditions is a nucleophilic
attack at the carbonyl. Model compounds studied by GC-MS confirmed the catalysis is a
function of concentration (Fig. 1-4A). From X-ray spectroscopy, the zinc was determined
to be tetracoordinated in the vitrimers, and from IR spectroscopy, the presence of zinc
affected the C-O and C=O vibrations. Tentatively, this interaction was represented as in
Figure 1-4C. Zn2+ acts as a catalyst in three ways according to this work. These are first
by activating the carbonyl of the ester (Fig. 1-4C), second by stabilizing the carboxylate
and shifting the equilibrium towards it, and third by bringing reactive sites closer
together.
As an alternative to epoxy based vitrimers, Hillmyer and coworkers explored
urethane vitrimers with a much higher concentration of esters, one per three back bone
atoms.24 Due to the synthetic route, it was possible to incorporate various
24
concentrations of free hydroxyl as well. All sample compositions had fast relaxation
times below 50 s at 140 ˚C.
Figure 1-4. The affect of zinc catalyst on transesterification. A) Transesterification
kinetics of model compounds with various loadings of zinc catalyst at 150 ˚C. B) The mechanism of base catalyzed transesterification without metal catalyst. C) Zinc cations coordinating with oxygen accelerating transesterification kinetics by two methods. Adapted from Demongeot, A.; Mougnier, S. J.; Okada, S.; Soulié-Ziakovic, C.; Tournilhac, F. Polym. Chem. 2016, 7, 4486-4493. Copyright 2016 The Royal Society of Chemistry.
1.3 Sulfur-based Functional Groups
Bulk dynamic covalent materials based on sulfur chemistries are very diverse
(Scheme 1-3). The exchange mechanisms for many of these chemistries rely on sulfur-
sulfur bond breakage and formation although a few use carbon-sulfur bonds. Some of
the mechanisms rely on radical sulfur while others operate through sulfur anions. In
some cases, the same functionality can exhibit different mechanisms depending on the
environmental conditions. Sometimes the mechanism of exchange in a material
changes as one of the functionalities turns into another because of the environment.
These chemistries are thermo-responsive, photo-responsive, or both.
25
Scheme 1-3. Types of sulfur-based exchange reactions used in bulk dynamic-covalent
polymeric materials.
An early type of sulfur-based exchange chemistry purposefully examined in
context of bulk polymers was thiol disulfide chemistry.25-27 Initially this example showing
healing was thought to be caused by disulfide-disulfide exchange,25 a closer look at the
kinetics in subsequent work suggested observed healing could likely be attributed to
thiol-disulfide exchange.26 The samples in this first work showed full recovery of
elongation at break after 1 h at 60 ˚C up to the third healing cycle. Better healing was
obtained with a higher concentration of exchangeable groups. When model compounds
were studied by SEC in the follow-up work, the presence of 1% of a small molecule
amine was necessary for nearly full exchange at 60 ˚C for 1 h. The first system had a
26
small amount of amine present. The rate of thiol-disulfide exchange by a nucleophilic
mechanism (Fig. 1-5A) was found to be much faster than disulfide-disulfide exchange
by a radical mechanism. The rate did depend on basicity of the environment and the
specific thiol and disulfide. The system with the highest exchange rate has a pH
corresponds to the pKa of the thiol. In Fig. 1-5B, the two regions seen in the stress
relaxation experiments, particularly for samples on day 0, suggested two mechanisms
of relaxation. The decreased stress relaxation of older samples was explained by
oxidation of thiols to form disulfides.
Figure 1-5. Thiol-disulfide exchange. A) Proposed mechanism of exchange. B) Stress
relaxation experiments after air exposure for 0-2 days. Adapted from Pepels, M.; Filot, I. A. W.; Klumperman, B.; Goossens, H. Polym. Chem. 2013, 4, 4955-4965. Copyright 2012 The Royal Society of Chemistry.
Another early thiol-disulfide material was investigated by Kowalewski and
coworkers.27 This particular system healed scratches at room temperature. The polymer
network was formed from the oxidation of thiol functionalized star polymers, which are
known to have low intrinsic viscosity. This ease of movement could contribute to the
exchange at a lower temperature. Also noteworthy from this work is the concept of
27
“zipping up” to heal damage where the areas nearest to each other and undamaged
sites heal first.
Disulfide-disulfide exchange chemistry has also been purposefully used to make
bulk dynamic covalent materials.28-31 Disulfide-disulfide exchange is desirable compared
to thiol-disulfide exchange because the presence of oxygen is no longer a concern.
Rowan and coworkers prepared photo-healable networks through complete oxidation of
thiol precursors.28 Full recovery of mechanical properties occurred after 5 min under UV
light. The mechanism of healing would proceed via radicals formed with the aid of the
UV light. This particular system was also heated, but only to melt the crystalline regions
of the network. The heating alone did not cause healing. Zhang and coworkers with
another system showed that self-healing without UV light or additional heat could be
achieved at room temperature in disulfide containing materials if a phosphine catalyst
was present.29 The phosphine catalyst, in a basic environment, was proposed to
stabilize the intermediate sulfur ion, facilitating metathesis of disulfide bonds (Fig. 1-6A).
A healing efficiency of 91% was obtained after 24 h at room temperature. Another
method used to lower the temperature for disulfide-disulfide exchange was presented
by Odriozola and coworkers.30,31 Poly(urea-urethanes) having aromatic disulfides with
adjacent hydrogen bonding sites on either side were measured to have healing
efficiencies of 97% after 24 h at room temperature (Fig. 1-6B). The hydrogen bonds
assured the disulfides would be in a desirable proximity and orientation to each other for
easy metathesis. The quadruple hydrogen bonding also contributed to healing recovery
by as much as 51%. Somewhat unexpectedly, minimal stress relaxation was observed
at 25 ˚C (Fig. 1-6C), even though healing and disulfide metathesis occur at this
28
Figure 1-6. Pre-orientation of the disulfide bond easing exchange. A) Proposed
mechanism with a phosphate catalyst. Adapted with permission from Lei, Z. Q.; Xiang, H. P.; Yuan, Y. J.; Rong, M. Z.; Zhang, M. Q. Chem. Mater. 2014, 26, 2038–2046. Copyright 2014 American Chemical Society. B) Fixing orientation of the disulfide bond with quadruple hydrogen bonds. C) Stress relaxation of aromatic disulfides with H-bonding. Adapted from Martin, R.; Rekondo, A.; Ruiz de Luzuriaga, A.; Cabañero, G.; Grande, H. J.; Odriozola, I. J. Mater. Chem. A 2014, 2, 5710-5716. Copyright 2014 The Royal Society of Chemistry.
29
temperature. From IR spectroscopy, the hydrogen bonding of the urethane carbonyl is
disrupted before 90 ˚C, and that of the urea carbonyl is disrupted at ~110 ˚C. The
increase in stress relaxation at 90 ˚C and above seen in Fig. 1-6C was explained by the
decrease in hydrogen bonding.
Very recently, another disulfide linkage producing catalyst-free healing was used
in bulk polymer networks by Otsuka and coworkers.32 Tetramethylpiperidine-1-sulfanyl
(TEMPS) is a sulfur analogue of the common molecular unit 2,2,6,6-tetramethylpiperidine-
1-oxyl, TEMPO. TEMPS unlike TEMPO usually exists as a dimer. The S-S dissociation
energy in TEMPS is about half that of alkyl disulfides. TEMPS was stable even after
heating at 100 ˚C for 24 h in solution while exposed to air. Recovery of 93% of
elongation at break was achieved after healing for 24 h at 100 ˚C.
Beyond the two sulfur atoms in disulfides, copolysulfides, with many sequential
sulfurs, have been recently prepared and studied by Pyun and coworkers.33-40 Heating
for 72 h at 100 ˚C provided sufficient molecular exchange to recover imaging quality for
copolysulfide lenses in both the visible and IR regions after they had been previously
damaged with sandpaper.34
Two other types of photo-responsive sulfur-based linkages have also been used
for dynamic covalent bulk polymer materials, trithiocarbonate and thiuram disulfide.41,42
Trithiocarbonate functionalities exchanged with UV light under nitrogen in bulk polymer
samples leading to visual healing after 48 h even when subjected to a bending force
with tweezers.41 Thiuram disulfide materials healed to 90% healing efficiency with
visible light under air after 12 h.42
30
Konkolewicz and coworkers used thiol-Michael linkages from a pre-formed thiol-
acrylate crosslinker to make polymer networks with dynamic behavior likely from retro-
Michael reactions.43 The best healing sample had ~85% recovery of strain at break after
healing for 16 h at 90 ˚C. Hydrogen bonding accounted for ~25% of this recovery.
1.4 Nitrogen-based Functional Groups
Like materials based on sulfur functional groups, bulk dynamic covalent materials
based on nitrogen chemistries are very diverse (Scheme 1-4). Some of them are stable
in the presence of oxygen while others are not. Hydrolytic stability also varies with some
functionalities being susceptible to degradation in the presence of water. Like previously
noted sulfur-based functionalities, the exchange mechanism, and mechanistic class, of
some nitrogen-based functionalities also varies depending on environmental conditions.
With the exception of alkoxyamines, all of these exchange chemistries rely on breaking
and reforming a C-N bond.
Vitrimers have been formed from two of the more hydrolytically stable nitrogen-
based functional groups used in dynamic covalent bulk polymers.44,45 The first of these,
vinylogous urethanes with vinylogous amides, bears a resemblance to the previously
discussed vitrimers with carboxylate transesterfication.44 Vinylogous amides are
hydrolytically stable and exchange by catalyst-free transamination. Ground and
reformed samples recovered their mechanical properties after 30 min at 150 ˚C.
Samples had relaxation times of 85 s at 170 ˚C. The second type of nitrogen-based
functionality used to make vitrimers is transalkylation of triazolium salts.45 The relaxation
times ranged from a few seconds at 200 ˚C to 30 min at 130 ˚C. Stress relaxation is
also influenced by counter-ion choice. These materials are particularly interesting as
they are conductive.
31
Scheme 1-4. Types of nitrogen-based exchange reactions used in bulk dynamic-
covalent polymeric materials with the exception of oximes, which were determined to be irreversible in the absence of solvent.
The next category of interest, imines, hydrazones, and oximes, bear a molecular
similarity to each other. In this order, they have increasing stability and resistance to
hydrolysis.46 A couple types of bulk polyimines have been studied by different
groups.5,47 Zhang and coworkers prepared polyimines and showed them to be dynamic
in dry conditions with heat and wet conditions at room temperature.5 In a dry
environment, 90% stress relaxation was achieved after 30 min at 80 ˚C, while it would
take ~480 days at room temperature. Stress relaxation while the sample was immersed
in water was comparable to the dry sample at 127.5 ˚C. Samples were able to be
32
reprocessed from a powder through either wet pressing or heating at 80 ˚C and
pressing. The degree of hydrolysis measured by NMR spectroscopy was minimal even
with water saturated samples. The authors suspect that the mechanism of exchange in
the wet samples is imine-amine exchange. In other work, Zhang and Barboiu using
polyimines to make asymmetric membranes for water transport.47 Hydrazones while
more hydrolytically stable than imines seem to be less common in bulk polymers. Lehn
and coworkers used the thermally responsive hydrazone exchange of two layered
polymer films to create color and florescent images.48 One film was of a polymer with a
thiophene moiety between two hydrazones while the other had a phenyl groups
attached to the hydrazones. Upon exchange, the new conjugation from the phenyls
through the thiophene caused the color change and florescence. A self-healing material
was also created with bis-acylhydrazones, which could have dynamic behavior from
hydrogen bonding and hydrazone exchange from the same molecular unit.49 The
healing of 90% recovery of strain after 4 h at room temperature measured in this work
was ascribed mainly to hydrogen bonding exchange. In other work, polymers with
acylhydrazones were shown to heal through hydrazone exchange, but the healing
conditions were heating at 100 ˚C for 64 h.50 Bulk polymers with oximes, the most
hydrolytically stable functionality of this group, have been studied quite recently.46 No
healing was observed despite glass transition temperatures below -20 ˚C, heating for 5
days at 100 ˚C, heating at 150 ˚C, exposing to acid while heating at 100 ˚C, including
acid in the polymer network, having excess aldehyde or ketone monomer, and using
less stable aliphatic oximes.
33
Similar in some ways to the vinylogous urethanes, hindered ureas have been
used in polymer networks to make self-healing materials.51 Hindered ureas, however,
rely on a dissociative mechanism. The chemistry itself is believed to be dynamic for two
reasons. The weakening of the C-N bond is expected to disrupt the orbital co-planarity
of the carbonyl and the substituted nitrogen, and unlike a disrupted amide functionality
producing the unstable ketene, the byproduct of the disrupted hindered urea is the more
stable isocyanate (Fig. 1-7). Ideally, as shown by the values listed in Fig. 1-7, the
equilibrium constant for the dynamic reaction sufficiently favors bond formation so that
the material is a crosslinked network, while the forward and reverse rates are fast
enough for exchange to produce healing on a reasonable time scale. Hindered ureas
are a particularly attractive dynamic chemistry because conventional polyureas and
poly(urethan-urea)s have readily available inexpensive starting materials that can be
made dynamic simply by using bulky amines. Samples healed for 12 h at 37 ˚C
recovered to give 87% of the original strain at break. Full recovery was prevented by the
susceptibility to hydrolysis of the isocyanate intermediate. Despite this drawback, this
example is one of the few examples of catalyst-free low-temperature healing.
Literature contains many examples of dynamic covalent bulk polymers based on
alkoxyamines.52-59 As mentioned in the introduction to this section, this is the only
nitrogen-based chemistry used in bulk dynamic covalent polymers that relies on
breaking and reforming a C-O bond instead of a C-N bond. Additionally, its exchange
mechanism is unique because unlike the other reactions it relies on the creation of a
stable radical. One concern with alkoxyamines is irreversible combination of carbon
34
Figure 1-7. Hindered ureas made more easily reversible by a bulky substituent. Adapted
from Ying, H.; Zhang, Y.; Cheng, J. Nat. Commun. 2014, 5, 3218-3226. Copyright 2014 Macmillan Publishers Limited. All rights reserved.
radicals, but most studies show that multiple cycles of healing have only a small
decrease in healing efficiency.52-57,59 In addition to being thermal responsive,
alkoxyamines can also be photo-responsive with UV light.59 Early thermo-responsive
designs required 2.5 h at 130 ˚C to get 75.9% healing and were air sensitive.56
Subsequent work increased the stability of the carbon radical by placing a nitrile
substituent on it.55 This also decreased the sensitivity to oxygen, and lowered the
healing temperature to 15 ˚C. After healing for 48 h at this temperature, a healing
efficiency of 94.6% was achieved. Very recently, Zhang and coworkers suggested that
35
the alkoxyamine dissociation temperature could be selectively designed by synthesizing
the alkoxyamine from the many commercially available azo-initiators.52 Also unique to
this work, an amide group next to the radical carbon was used to increase the radical
stability and promote oxygen insensitivity. Healing this particular system for 4 h at 80 ˚C
in air produced 100% healing efficiency.
Figure 1-8. Noncatalytic Michael addition. A) Thiol Michael addition enhanced by
electron withdrawing groups in a small molecule study. Adapted with permission from Krishnan, S.; Miller, R. M.; Tian, B.; Mullins, R. D.; Jacobson, M. P.; Taunton, J. J. Am. Chem. Soc. 2014, 136, 12624-12630. Copyright 2014 American Chemical Society. B) Aza-Michael addition with an activated double bond in a bulk network. Adapted with permission from Baruah, R.; Kumar, A.; Ujjwal, R. R.; Kedia, S.; Ranjan, A.; Ojha, U. Macromolecules 2016, 49, 7814-7824. Copyright 2016 American Chemical Society.
36
The last nitrogen-based linkage to be addressed in this section is the reversible
aza-Michael reaction (Fig. 1-8B). The reversible aza-Michael addition was recently
studied in combination with a pH sensitive amide linkage by Ojha and coworkers.60
They were inspired by previous studies examining thiol Michael reversibility. Thiol
Michael reversibility had been found to be possible when the double bond was more
electron deficient due to conjugation with a carboxylate or cyano group (Fig. 1-8A).61,62
In bulk polymer materials with reversible aza-Michael reactions, healing of a scratch
was possible after 30 min at 50 ˚C.60
1.5 Other Dynamic-Covalent Linkages
Scheme 1-5. Other types of exchange reactions used in bulk dynamic-covalent
polymeric materials.
37
The remaining dynamic covalent linkages seen in scheme 1-5, with the exception
of the siloxane silanol, rely on dynamic C-C bond dissociation and reformation.
Additionally, the middle three are photo-responsive. Aside from these similarities, they
are all quite different from each other.
The first reversible reaction, olefin metathesis, requires the presence of a
catalyst.63,64 The exchange mechanism proceeds through a metallocyclobutane
intermediate (Fig. 1-9). Increasing the amount of the second-generation Grubbs’ Ru
metathesis catalyst led to faster rates of exchange and lower crosslink density.63 The
faster exchange caused faster stress relaxation63 and healing.64 Samples fully healed
in 3 h at room temperature and 30 min at 30 ˚C. A catalyst-containing sample welded to
a catalyst-free sample to give ~80% of the maximum strain after 3 h at room
temperature.
Figure 1-9. Crosslink conservation through Ru-catalyzed olefin metathesis. Adapted
with permission from Lu, Y.-X.; Tournilhac, F.; Leibler, L.; Guan, Z. J. Am. Chem. Soc. 2012, 134, 8424-8427. Copyright 2012 American Chemical Society.
38
The next three linkages to be discussed are photo-responsive. The first two of
these proceed through the formation of a cyclic structure. Anthracene, in addition to
being used in anthracene-maleimide Diels-Alder linkages, can also [4+4] photodimerize
with UV light with wavelengths longer than 300 nm and then dissociate either at UV
wavelengths shorter than 280 nm or thermally at 200 ˚C.65 Likewise, cinnamoyl groups,
photodimerize with [2+2] cycloaddition.66 UV light above 280 nm crosslinks the
cinnamoyl groups and UV light below 280 nm dissociates the groups. Healing took
place after 10 min under UV light greater than 280 nm. Better healing took place when
samples were heated at 100 ˚C while irradiated. The last photo-responsive linkage to be
discussed, diarylbibenzofuranone (DABBF), has been used to study mechanochemical
activation.9,67 The dissociated DABBF radical has a noticeable blue color so that when
incorporated into a polymer it acts a visual sensor (Fig. 1-10B). A thermoplastic
urethane with DABBF in the backbone showed repeatable color change in response to
stress.67 The DABBF units had the potential to act both as damage detectors and the
healing enabler. DABBF has also been used to examine the effect of molecular weight
and polymer architecture on mechanical activation in bulk polymers.9 Polymers
increasing molecular weight or greater branching were synthesized with DABF in the
center. As measured by EPR spectroscopy (Fig. 1-10A), higher molecular weight or
higher branching led to more activation.
The last exchange chemistry for bulk polymers in this section, siloxane
exchange, is quite old. Polydimethylsiloxane has been studied for dynamic behavior in
solid state since the 1950s.68 Samples with the “living” reactive anionic species can heal
after 24 h at 90 ˚C.69 The anionic tetramethylammonium dimethylsilonate end groups
39
catalyze the siloxane exchange. This anionic end group is air and water stable at lower
temperatures and can be readily removed at the “decatalyzation” temperature of 150 ˚C.
In addition to healing, the materials were also remoldable.
Figure 1-10. Diarylbibenzo-furanone dissociation. A) Dissociation quantification
measured by EPR for polymers with different molecular weights and architectures. B) Images of the color change in the bulk polymer before and after the application of mechanical stress. Adapted with permission from Oka, H.; Imato, K.; Sato, T.; Ohishi, T.; Goseki, R.; Otsuka, H. ACS Macro Lett. 2016, 5, 1124-1127. Copyright 2016 American Chemical Society.
1.6 Boron-based Functional Groups
Boron-based functional groups in bulk dynamic polymer materials have just
begun to be explored. Unlike many of the previous sections were the cited examples
were pertinent, often the first, studies from the many available in literature, to the best of
40
my knowledge, this section is a comprehensive overview of boron-based dynamic-
covalent bulk polymers. The three that have been studied to date are presented in
scheme 1-6. Like many of the previously mentioned chemical functional groups, such as
esters, imines, and hydrazones, boron-based functional groups can also be hydrolyzed.
This hydrolysis can be purposely used as a reversibility mechanism. The first two
mechanisms discussed in this section will address the use of hydrolysis for reversibility.
The last mechanism to be discussed relies on transesterification reminiscent of the
carboxylate-based transesterification of esters.
Scheme 1-6. Types of boron-based exchange reactions used in bulk dynamic-covalent
polymeric materials.
Boroxine formation, unlike all the previously discussed reversible functional
groups, requires three functionalities to form instead of two. The very high modulus of
PDMS materials containing boroxines, two times higher than previously reported self-
healing PDMS systems, was attributed to the uniqueness of boroxines.2 In addition to
increased stiffness, the boroxine units also caused an increased material strength. With
41
the application of a small amount of water, damaged interfaces healed with efficiencies
of 95% after 5 h at 70 ˚C. Adhesive properties of this material, after curing for 24 h at 70
˚C, were similar to those of cyanoacrylate and epoxy commercial glues. Conductive
semi-transparent self-healing composites were made with silver nanowires simply by
pre-wetting the surface of the boroxine PDMS materials and applying the silver
nanowires (Fig. 1-11). These composites had improved adhesion between the silver
nanowires and the polymer matrix compared with previous self-healing silver nanowire
composites. When mechanically tested at various humidities, the materials remained
relatively stable despite being susceptible to hydrolysis.
Figure 1-11. Conductive composite made from Ag-nanowires in a boroxine polymer
network. A) Image of semi-transparency of the composite. B) UV spectra of the polymer network with and without Ag-nanowire. C) SEM image of the composite after multiple adhesion-peeling tests, red denotes the Ag-nanowires. D) Composite surface damaged with a crack. E) Composite surface after healing. Adapted from Lai, J.; Mei, J.; Jia, X.; Li, C.; You, X.; Bao, Z. Adv. Mater. 2016, 28, 8277-8282. Copyright 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
42
The remaining two exchange types rely on two different mechanisms of
exchange for boronic esters. The first, like boroxines, is the dissociative hydrolysis and
reformation of the boronic ester functionality.70,71 The first bulk dynamic polymer using
boronic ester hydrolysis as the exchange mechanism is presented in Chapter 3 of this
dissertation. Briefly, with the application of a small amount of water at a cut interface,
recovery of maximum stress and strain at break was observed for samples after healing
at room temperature for 3 min followed by a drying process.70 On the basis of this work,
Zuo, Gou, Zhang, and Feng prepared similar polysiloxane materials that also showed
self-healing within 30 min at room temperature with a small amount of water at the
damaged site.71 Chapter 4 will look at materials capable of both a hydrolysis exchange
mechanism and a boronic ester transesterification mechanism. The latter mechanism of
exchange, boronic ester transesterification, was used to cause malleability and self-
healing in polymer networks prepared by Guan and coworkers.72 1,2-diol-containing
polycyclooctene was crosslinked with one of two diboronic esters leaving free diol for
transesterification. The boronic ester with the faster exchange rate was stabilized
through coordination with an attached amine. Networks with the faster boronic ester
healed after 16 h at 50 ˚C. These samples could also be reprocessed from small pieces
with a melt press at 80 ˚C. Reformed samples approached the original maximum stress
and strain at break values. Stress relaxation was measured for solvent swollen
samples. Chapter 5 will examine using boronic ester transesterification to make a
vitrimer with stress relaxation times at increasing temperatures following Arrhenius
behavior.
43
CHAPTER 2 RESEARCH OBJECTIVE
The purpose of the present research was to synthesize and investigate bulk
polymer materials with dynamic-covalent boronic ester linkages to create new self-
healing materials and strong organic glasses. Much effort in recent years has gone
towards optimizing the selection of the exchange mechanism of reversible linkages for
specific applications and developing an understanding of the best polymer architectures
in dynamic materials to produce various mechanical properties. These investigations
are ongoing through the efforts of researchers globally. Several types of reversible
linkages have been employed for dynamic bulk materials, but boronic esters have
received less attention. We investigated the reversibility of boronic esters in polymer
networks, selectively promoting mechanisms of hydrolysis and transesterification. Over
the course of our work, networks were prepared with boronic esters between all
crosslinking points, with the addition of free diol, and with irreversible linkages
substituted for some of the boronic ester. Networks with five and six-membered boronic
esters rings were also evaluated to determine how their different rates of boronic ester
exchange and the positions of their equilibrium altered the mechanical properties of the
bulk polymers. Boronic ester reversibility was also studied in networks with glass
transition temperatures below room temperature and above room temperature.
Specific objectives of each research project and a brief summary of our results
are given below.
The goal of the research described in Chapter 3 was to synthesize cross-linked
polymers constructed with dynamic-covalent boronic esters via photo-initiated radical
thiol−ene click chemistry. Because the reversibility of the boronic ester cross-links was
44
readily accessible, the resulting materials were capable of undergoing bond
exchange to covalently mend after failure. The reversible bonds of the boronic esters
were shown to shift their exchange equilibrium at room temperature when exposed to
water. Nevertheless, the materials were observed to be stable and hydrophobic and
absorbed only minor amounts of water over extended periods of time when
submerged in water or exposed to humid environments. The facile reversibility of the
networks allowed intrinsic self-healing under ambient conditions. Highly efficient self-
healing of these bulk materials was confirmed by mechanical testing, even after
subjecting a single site to multiple cut−repair cycles. Several variables were considered
for their effect on materials properties and healing, including cross-link density,
humidity, and healing time (Chapter 3).
The objective of the research described in Chapter 4 studied the effect of
combining varying ratios of dynamic boronic ester crosslinks and static (i.e., irreversible)
crosslinks in bulk polymeric materials. Networks that contain boronic ester crosslinks
have been shown to undergo dynamic bond exchange that enables self-healing
behavior and reprocessability. However, networks crosslinked exclusively by rapidly
exchanging bonds are also susceptible to creep and stress relaxation, which limits
many potential materials applications. Different mechanisms of bond exchange were
also considered by also preparing networks that contained free diols to enable crosslink
exchange of boronic esters by transesterification. The networks were evaluated in terms
of responsiveness to moisture, proclivity towards deformation, and ability to self-heal.
The networks containing free diols could be remolded and healed upon heating. By
controlling the humidity and temperature of the environment, the dominant boronic ester
45
exchange mechanism could be shifted from hydrolysis/esterification to
transesterification. Incorporating a fraction of permanent crosslinks yielded networks
that maintained their structural integrity yet still underwent good healing and
reproducible repair after multiple cut/heal cycles. When both free diols and irreversible
crosslinks were incorporated into a single network, shape stability was enhanced, and
improved healing was observed when compared to networks that contained either free
diol or permanent crosslinks independently (Chapter 4).
The goal of the research described in Chapter 5 was to broaden the limited
scope of vitrimer exchange reactions. Boronic ester transesterification was explored as
an alternative to previous chemistries. Vitrimers being strong organic glass formers
create a class of materials with the lightness and insolubility of crosslinked thermosets
and rubbers while possessing the ability to be reshaped. First, kinetic and
thermodynamic studies of model boronic ester compounds revealed the improved
hydrolytic stability of the six-membered ring in comparison to the five-membered ring
and the relative transesterification rates. Next, dynamic boronic ester networks of both
ring types were prepared by photo-initiated thiol-ene polymerization using a difunctional
spacer to produce crosslinked networks with Tg values above room temperature. The
materials were characterized by stress relaxation experiments, which confirmed they
were vitrimeric in nature. Our report demonstrates how a fundamental understanding of
small molecule boronic esters can be implemented to enhance stability and to optimize
the potential of boronic ester polymers for new applications such as vitrimers (Chapter
5).
46
CHAPTER 31 ROOM-TEMPERATURE SELF-HEALING POLYMERS BASED ON DYNAMIC-
COVALENT BORONIC ESTERS
3.1 Overview
Polymeric materials constructed via dynamic-covalent bonds with sufficient chain
mobility have the capacity to demonstrate reversible equilibria73,74 in the solid state,
which may have broad implications for the design and development of smart
materials.75 Many recent reports in the literature detail the exploitation of dynamic-
covalent bonds to effect self-healing behavior.76-81 Progress in this arena fundamentally
expands the macromolecular hypothesis as stated by Flory82 and originally published by
Staudinger.83 Indeed, as the field of dynamic polymers expands, the covalent structure
of polymers is no longer solely responsible for the unique characteristics and properties
of novel polymers. Rather, the additive effects of reversible bonds can lead to new
properties and structurally dynamic polymers,75 which are defined by Rowan and
coworkers to be macromolecules with macroscopic responses to changes at a
molecular level due to reversible chemistry.75 With these new capabilities, novel
functions and applications, such as rewritable surfaces, robust recyclable materials, and
self-healing coatings are attainable.84
Self-healing based on reversible bonds can occur via two pathways, depending
on the end condition and timescale of the damage. Macroscopic failure can often be
prevented by the preemptive healing of microscopic damage as it forms,85-87 On the
other hand, healing of macroscopic damage can be either externally triggered to heal or
internally induced via shape memory,28,88,89 by placing the edges of the damaged
Reproduced with permission from Macromolecules 2015, 48, 2098-2106. Copyright 2015 American
Chemical Society.
47
interfaces in close proximity to allow bond exchange. Both routes can prolong material
lifetime and may allow repeated healing.
Although there have been recent reports of gels that undergo healing by
exchange of dynamic-covalent bonds,90-94 there are relatively few examples of dynamic-
covalent chemistry being applied toward materials that can self-heal in the bulk. Healing
of macromolecular materials in the bulk state is arguably the most important area of
self-healing, given that many polymers are typically utilized in a solventless
environment. Given the diversity of dynamic-covalent bonds and the wide variety of
conditions under which their exchange can be triggered, there remain many
opportunities for improvement and expansion in this area. Most previous examples of
bulk self-healing have relied on bond exchange externally triggered by heat or light. The
earliest thermally responsive systems, prepared by Wudl and coworkers, were based on
furan-maleimide Diels-Alder chemistry with temperatures greater than 115 ˚C needed
for healing.12,13 However, exposure to high temperatures is often not feasible or
desirable for many applications, which has lead to other approaches being explored to
induce healing. For example, Klumperman and coworkers developed a healing process
based on a thiol-disulfide exchange mechanism at 60 ˚C.25,26 Although this approach
proved highly successful for the preparation of healable materials, the concentration of
thiols was observed to decrease over time under an ambient atmosphere because of
continuous oxidization of thiols to disulfides. Alkoxyamines have also been considered
as thermoreversible dynamic-covalent bonds that can bring about self-healing behavior.
However, the radical products of alkoxyamine dissociation are also sensitive to oxygen
and high temperatures (90-130 ˚C), which are typically required to induce dissociation of
48
the labile O−C bond.57,95 There have been only limited reports of healing via photo-
irradiation of alkoxyamines at room temperature.59 Like thiol-disulfide exchange, the
efficiency of alkoxyamine healing is expected to decrease with time, in this case, due to
the inevitable irreversible combination of carbon-centered radicals. Alternatively, in one
of the earliest reports that relied on a light trigger, Chung and coworkers achieved
covalent healing by the photo-induced [2+2] cycloaddition of cinnamoyl groups to
reversibly form a cyclobutane derivative.66 More recently, trithiocarbonate moieties were
employed by Matyjaszewski and coworkers in the first example of macroscopic fusion
from UV-induced healing.41
While photo- and heat-induced dynamic-covalent chemistries have proven
valuable in many self-healing systems, autonomous healing with no external trigger is
often most desirable. Fewer examples exist of bulk systems being healed with no
significant outside stimulus being necessary. In most of these cases, such healing
typically occurs as a result of the stimuli being present under ambient conditions (i.e.,
ambient light or heat). For example, the relatively stable radicals from thiuram disulfide
have been employed for visible-light self-healing over a 24 h period, which succeeded
efficiently in air as long the damaged pieces had not been separated for an extended
time.96 Ghosh and Urban developed a UV self-repairing polyurethane based on oxetane
rings capable of scratch healing within an hour, only needing power densities similar to
sunlight.97,98 Other previously reported materials rely on the heat present at room
temperature to induce healing. However, some of these systems require the presence
of catalysts in the bulk matrix. For example, Lehn and coworkers designed a double
dynamic bis-imino carbohydrazide polymer infused with acid catalyst that healed at
49
room temperature in bulk.49 Similarly, disulfide metathesis can also cause healing at
room temperature with the aid of an aliphatic phosphine catalyst.29 Catalyst-free
approaches comprise an even smaller subset of strategies to achieve bulk healing at
room temperature. Preliminary investigations qualitatively suggest room-temperature
self-healing can occur through tailor-made Diels-Alder moities.15 Furthermore, Odriozola
and coworkers have demonstrated a compelling example of room-temperature intrinsic
self-healing in the solid state due to aromatic disulfide metathesis;30 however, this
material relies largely on reversible supramolecular hydrogen bonds for healing instead
of reversible covalent bonds. We are interested in exploiting new dynamic-covalent
chemistries for intrinsic self-healing in bulk under ambient conditions.
For this purpose, we were interested in boronic acids, which are known to form a
variety of dynamic-covalent bonds.99-101 For example, the dehydration of boronic acids
to form boroxines is readily reversible by hydrolysis. Boroxine formation has been
employed to prepare a number of dynamic-covalent assemblies.102-104 The direction of
the boroxine/boronic acid equilibrium can be readily controlled by temperature, the
addition of Lewis bases, or the addition of water. Boronic acids are also capable of
forming dynamic-covalent bonds by reacting with diols, typically either in basic aqueous
media or in anhydrous organic solutions to form boronate esters or boronic esters,
respectively. Our group has demonstrated that boronic ester-based macromolecular
stars can be rendered dynamic in organic solutions.105 Boronate esters have also been
employed to prepare self-healing hydrogels, wherein covalent healing can be effected
by formation of new boronate ester bonds along the interface of damage.90,106-111
Esterification of boronic acids can also be exploited to bring about mending in the
50
absence of water. Lavigne and coworkers have prepared dynamic-covalent linear
polymer chains by polymerization of low molecular weight bis-diols with diboronic
acids.100 The resulting polymers were hydrolyzed in organic solution, isolated by drying,
and restored back to the original molecular weight under vacuum.
As compared to these previous reports, we were interested in using boronic
esters for self-healing of networks in the bulk, reasoning that these linkages may be
ideal for self-healing because they can be rendered dynamic at room temperature under
ambient conditions. We reasoned that hydrolysis of surface-exposed boronic esters in a
bulk material could occur by intentionally wetting the surface at the site of damage (or
from water present in the atmosphere under ambient humidity) to induce exchange of
boronic esters to heal the material by covalent bridge formation across the damage
interface. Accordingly, a styrenic boronic ester was synthesized and incorporated into a
network by a radical-based thiol-ene process. The bulk behavior of these networks was
investigated for their self-healing properties that arise from the dynamic-covalent nature
of their boronic ester crosslinks. The polymeric networks were capable of bulk-state
healing at room temperature, suggesting they may hold promise for various
applications, including being used as coatings, composites, and biological materials.
3.2 Experimental Section
3.2.1 Materials
Divinylbenzene (Sigma-Aldrich, 80%) was passed through a column of basic
alumina. Dimethylsulfoxide-d6 (d-DMSO, Cambridge Isotope, 99.9% D) was dried
overnight over 4A molecular sieves. Dichloromethane (DCM, Sigma-Aldrich) was dried
using an anhydrous solvent system (Innovative technologies). 4-Vinylphenylboronic acid
(VPBA, Combi-blocks, 98%), 3-allyloxy-1,2-propanediol (Acros Organics, 98%),
51
pentaerythritol tetrakis(3-mercaptopropionate) (PTMP, Sigma-Aldrich, 95%), 3,6-dioxa-
1,8-octanedithiol (DODT, TCI America, 95%), 2,2-dimethoxy-2-phenylacetophenone
(DMPA, Sigma-Aldrich, 99%), potassium chloride (BDH, 99%), sodium chloride (Fisher,
99%), potassium acetate (Macron Chemicals, 99%), deuterium oxide (D2O, Cambridge
Isotope, 99.9% D), and molecular sieves (4 A, Mallinckrodt) were used as received.
3.2.2 Instrumentation and Analysis
1H NMR (500 MHz), 13C NMR (125 MHz), and 11B NMR (160 MHz) spectra were
recorded using an Inova 500 spectrometer. For 11B NMR spectroscopy, 5-mm thin-
walled quartz NMR tubes were used. Chemical shifts are reported in parts per million
(ppm) downfield relative to tetramethylsilane (TMS, 0.0 ppm). Multiplicities are reported
using the following abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet;
br, broad. High-Resolution Mass Spectrometry (HRMS) was conducted with an Agilent
6220 TOF-MS mass spectrometer in the Direct Analysis in Real Time (DART) mode
with the IonSense DART source. Infrared spectra were collected on a Thermo Nicolet
5700 FTIR spectrometer equipped with a single bounce diamond stage attenuated total
reflectance (ATR) accessory. Differential scanning calorimetry (DSC) measurements
were performed on a TA Instruments Q1000 equipped with a liquid nitrogen cooling
accessory and calibrated using sapphire and high purity indium metal. All samples were
prepared in hermetically sealed pans (4−7 mg/sample) and were referenced to an
empty pan. A scan rate of 10 °C per minute was used. Glass transition temperatures
were evaluated as the midpoint of a step change in heat capacity. Thermal experiments
were conducted as follows: samples were heated through 220 °C, followed by cooling at
10 °C per minute to −80 °C, and then heated through 220 °C at 10 °C per minute. Data
reported reflects the average of the second and third heating scans. Stress/strain
52
properties of all network compositions were measured on a standard Instron testing
machine (No. 4204) using test specimens in the form of dogbones according to ASTM
standard and procedure (D 638). The gauge length was 50.0 mm, and the crosshead
speed was 10 mm/min at 25°C and 50% humidity. The data reported are the averages
of five measurements. The UV lamp used for photocuring was a UVP Blak-Ray Model
B100AP at 365 nm with 8.9 mW/cm2.
3.2.3 Synthesis and Experimental Procedures
Synthesis of 4-((allyloxy)methyl)-2-(4-vinylphenyl)-1,3,2-dioxaborolane
(VPBE). 4-Vinylphenylboronic acid (24.4 g, 165 mmol) and 3-allyloxy-1,2-propanediol
(19.8 g, 150 mmol) were stirred in dry DCM (200 mL) with molecular sieves (4 Å, ca. 20
g) at room temperature. The progress of the reaction was monitored by 1H NMR
spectroscopy, and additional molecular sieves were added as needed to drive the
reaction to completion. After confirming reaction completion, VPBE was purified by
filtering, centrifuging, filtering again, and concentrating to give the final colorless-to-pale
yellow liquid (30.4 g, 125 mmol, 83%). The product was characterized by 1H NMR and
13C NMR spectroscopy. 1H NMR (500 MHz, d-DMSO): δ (ppm) 7.68 (d, 2H), 7.50 (d,
2H), 6.76 (dd, 1H), 5.92 (d, 1H), 5.87 (m, 1H), 5.34 (d, 1H), 5.24 (d, 1H), 5.13 (d, 1H),
4.73 (m, 1H), 4.37 (dd, 1H), 4.08 (t, 1H), 4.01 (m, 2H), 3.56 (m, 2H). 13C NMR (125
MHz, d-DMSO): δ (ppm) 140.0, 136.5, 135.0, 134.8, 125.7, 116.5, 115.7, 75.9, 71.4,
71.4, and 67.5. 11B NMR (160 MHz, d-DMSO): δ (ppm) 35.2. ESI-HRMS: Calcd. for
[M+Na+]: 266.1199. Found: 266.1187. Elemental analysis: Calculated for C14H17BO3: C,
68.98%; H, 7.02%. Found: C, 68.82%; H, 7.08%.
Model Degradation of VPBE. VPBE (40.0 mg, 0.164 mmol) was dissolved in
dry d-DMSO (0.7 mL). After the 1H and 13C NMR spectra were obtained, D2O (0.050
53
mL, 2.5 mmol) was added to the NMR tube, and 1H and 13C NMR spectra were
recorded again after the solution was mixed well.
Synthesis of Thiol-ene Networks. Ratios of VPBE, DODT, PTMP, and DMPA
at targeted ratios were mixed and then sonicated for several minutes until a
homogenous solution was formed. Reaction size scale was typically 1-5 g total with
vinyl to thiol functionalities being maintained in a 1:1 ratio. Networks with different ratios
of DODT and PTMP were prepared. The samples had either 25:75, 50:50, or 75:25 of
the thiol equivalence coming from DODT:PTMP. All reactions used 1 wt% DMPA. The
same procedure was followed for the reference material made with divinylbenzene in
the place of VPBA. The solution was then transferred to preformed molds and cured for
30 min while being irradiated at 365 nm, rotating periodically to ensure uniform curing
from the top and bottom of the samples. This same procedure was followed for all
networks. ATR-FTIR spectroscopy was used to monitor thiol conversion using the S−H
absorption peak at 2590 cm-1. Thermal characterization of the final materials was
performed by DSC.
Water Absorption of Networks. Three samples with compositions of 75:25
DODT:PTMP were weighed and placed in vials of water at room temperature. At 24 h
intervals, samples were removed, patted dry, and weighed.
Humidity Chamber Preparation. Constant humidity environments were
prepared as previously reported.112 Briefly, 23, 75, and 85% humidity atmospheres were
made by filling the bottom of sealed containers with saturated salt solutions with
potassium acetate, sodium chloride, and potassium chloride, respectively. These salts
54
were selected to minimize humidity variation within our typical laboratory temperature
ranges, thereby optimizing the environments for use in controlled healing experiments.
Water Contact Angle Measurements. Three samples with compositions of
75:25 DODT:PTMP were kept for 21 days in 23, 75, and 85% humidities. The samples
were then removed, and the water contact angle was immediately measured with a
goniometer using high-resolution video. A minimum of 5 drops was recorded per
composition. Still images were isolated from the video at 0, 10, and 20 s. Mathematical
fitting software (Low-Bond Axi-symmetric Drop Shape Analysis (LBADSA) Plugin for
ImageJ) was used to determine the water contact angles, which are reported as
averages of 5-7 measurements.
Network Healing. To evaluate healing, cured samples of all compositions were
healed for varying times under ambient conditions after dabbing the cut interfaces with
water and reconnecting the two individual pieces along their freshly exposed interface.
Likewise, divinylbenzene control networks were cut, dabbed or immersed in water, and
reconnected for more than 3 days. Healing was also attempted for all boronic ester
samples over several days in ambient conditions with the absence of liquid water. The
75:25 DODT:PTMP composition was additionally tested for healing with heat in the
absence of water at 50 ˚C for 30 min. To assess healing quantitatively, dogbone shaped
samples were cured in silicone molds, cut, dabbed with water, reconnected, returned to
the mold, and placed in 85% humidity chambers for up to 3 days. After removal from the
chambers, the samples were vacuum dried for over 18 h and with consistent intermittent
storage in a desiccator until their final weights were within 0.3 wt% of their weights
before humidity exposure. For repeat damage and healing studies, samples were cut,
55
healed, and dried for approximately 6 h before being cut and healed in the same place
for several cycles. All samples were characterized by tensile testing.
56
3.3 Results and Discussion
3.3.1 Monomer Synthesis and Reversibility
To demonstrate boronic esters can be used for macroscopic healing in the bulk
state and to examine simple variations in the network composition, the polymer
formation reaction needed to possess a few key characteristics, such as tailorability, an
absence of side products, and easy processability to facilitate sample preparation. All
these features can be found in reactions categorized as click chemistry.113-115
Specifically, we chose radical photo-initiated thiol-ene chemistry because of the readily
available monomers, the potential to conduct the reaction without solvent, and mild
reaction conditions that allow for ambient temperatures and the tolerance of oxygen.116
To enable subsequent thiol-ene chemistry, we used commercially available thiols and a
novel boronic ester diene.
The boronic ester diene was formed in good yields from the corresponding
boronic acid and diol in dry organic media. The resulting boronic ester diene was a
liquid at room temperature and miscible with common thiol-ene reagents, DODT and
PTMP, and the photoinitiator. Furthermore, the diene was stable at room temperature
when stored in a dry environment. Because the healing behavior of the polymeric
networks was being designed to exploit the reversibility of boronic esters, the monomer
was evaluated in a model experiment by 1H NMR spectroscopy under both dry and wet
conditions to show the extreme shifts in equilibrium. The dry sample demonstrated clear
downfield shifts of the protons near the oxygen atoms of the boronic ester and a
narrowing of the distance between the two aromatic peaks relative to the starting
materials. After adding a small amount of water and waiting for 30 s, 97% of the boronic
ester was observed to have hydrolyzed to its constituent boronic acid and diol
57
components (Fig. 3-1), clearly demonstrating the hydrolytic reversibility of the boronic
ester bonds.
Figure 3-1. Diene boronic ester. A) Dynamic equilibrium of the boronic ester diene B)
1H NMR spectrum of a solution of the boronic ester in d-DMSO before and after the addition of D2O.
3.3.2 Network Formation
The diene boronic ester was reacted with readily available di- (DODT) and tetra-
(PTMP) thiols and a photoinitiator to produce networks with boronic ester linkages
between all crosslinking points (Fig. 3-2). The crosslink density was varied by altering
the ratio of the di- to tetrathiol while maintaining a 1:1 ratio of thiol to –ene. The Tg of all
the samples was below room temperature, suggesting the samples would have
sufficient chain mobility under ambient conditions to allow efficient healing (Table A-1).
Complete consumption of thiols was verified for all monomer compositions by
ATR-FTIR spectroscopy.25 Confirmation of the absence of excess thiols was essential
58
Figure 3-2. Synthesis of boronic ester network materials via photoinitiated thiol-ene
curing.
to verify that healing of the resulting materials was due to boronic ester exchange and
not the result of thiol-disulfide or disulfide-disulfide exchange reactions (vide infra),
which have previously been shown to allow healing.25,26 Fig. 3-3 shows the FTIR
spectrum for the boronic ester, the tetrathiol, the pre-crosslinked solution, and the
network after curing. The S−H absorbance peak completely disappeared after curing,
suggesting the thiol-ene crosslinking reaction was essentially quantitative and that the
network-forming monomers completely reacted to give a highly crosslinked product.
Figure 3-3. FTIR spectra of the boronic ester diene, tetrathiol (PTMP), the solution prior
to crosslinking, and the final crosslinked network.
59
3.3.3 Network Characterization
The cured samples were insoluble in dry DMSO, THF, and acetone, all solvents
that were previously demonstrated to be good for the precursor monomer components.
This transition in solubility offered further evidence of the crosslinked network structure.
Because the reversibility of the boronic ester functionality is hydrolytically driven, we
examined the network stability and interaction with water. Despite the lability of the
boronic ester crosslinks, samples completely immersed in water were stable and did not
degrade over a period of 60 days, though prolonged immersion did result in more
significant creep. The samples also absorbed a small amount (<10%) of water during
this time (Fig. 3-4), and while this amount was greater than would be theoretically
necessary to hydrolyze all of the boronic esters, the persistent stability suggests the
relative hydrophobicity of the networks was sufficient to retard infiltration of water to
preserve the boronic ester crosslinks and prevent overall material degradation.
Figure 3-4. Water absorption of disk shaped samples of boronic ester-crosslinked
network materials (DODT:PTMP = 75:25) completely submerged in water as a function of time.
To more thoroughly assess the response of the material to aqueous
environments, a series of water contact angle measurements was performed on
60
samples equilibrated for 21 days at three different humidities (Fig. 3-5). These samples
were then removed to ambient conditions and immediately tested. The initial contact
angles (95-103˚) revealed the relative hydrophobicity of the networks. Interestingly, the
measured contact angles decreased with time, a behavior often associated with surface
rearrangement. However, given the expected uniform distribution of polar groups within
the network and the results of the water absorption experiments, we attribute the
reduction in contact angle to be at least partly the result of droplet absorption during the
measurement. This hypothesis was supported by the fact that the samples aged in the
most humid environment (85% humidity) prior to analysis demonstrated the least
change in contact angle over time, as the outer surface of the material already had
sufficient opportunity to absorb water from the humid atmosphere
Figure 3-5. Water contact angles as a function of time for a boronic ester-crosslinked
network (DODT:PTMP = 75:25) equilibrated at 85, 75, or 23% humidity measured at a series of times after removal from respective humidity chambers.
61
in which it was previously stored. On the other hand, the sample aged at 23% humidity
demonstrated the most significant change in contact angle, which is consistent with it
being initially drier. The drier samples also demonstrated lower initial contact angles,
which we attribute to water absorption occurring immediately upon contact with water
thereby complicating the absolute measurement of contact angle. If the change in
contact angle was the result of increased polarity due to surface rearrangement, the
samples aged at higher humidity would be expected to be the most hydrophilic initially.
3.3.4 Self-healing
Given the reversibility and exchange of boronic esters in the presence of water,
we reasoned the networks may be capable of self healing when damaged. In this case,
the envisioned healing behavior would arise from boronic esters present at the interface
of a cut exchanging with those on the adjacent surface to span the divide at the site of
damage. Healing was first explored qualitatively (Fig. 3-6). Disk-like samples were cut,
separated, and their newly revealed cut surfaces were dabbed with 2-3 drops of water
before being placed back in contact with one another. We expected the addition of
water to the freshly cleaved surfaces could shift the equilibrium of surface-exposed
boronic ester groups toward the disassociated state so that the resulting free boronic
acid and diol groups on each surface could more readily form bridges via bond
formation across the fracture interface.117 Additionally, because of the increased
propensity for creeping when wet, moistening the surfaces prior to healing should also
facilitate more intimate contact and increased chain mobility along the cut surfaces,
which may help healing to occur through a “zipping-up” process starting from areas in
contact.86 Interestingly, the samples were not particularly tacky when dry, so self-
healing behavior would be distinct from self-adhesion in which two surfaces with their
62
reversible bonds at equilibrium are brought in contact. In this case, the bridges would
only be allowed to grow slowly because of low concentrations of reactive groups
present at their equilibrated surfaces.117 It was envisioned that self-healing of these
boronic ester materials would be achieved by cutting a sample and wetting its surface to
create a situation far from the bulk equilibrium, with many pairs of boronic acids and
diols being available to induce healing by bridge formation. Also, given that there could
be preexisting finite concentration of free diol groups within the matrix and at the cut
interface, we cannot entirely exclude the possibility of healing via transesterification.
Figure 3-6. Self-healing of boronic ester network materials. A) Proposed mechanism of
healing. (Note that while the middle image suggests complete hydrolysis of the boronic esters at the damage interface, there may also be intact boronic esters that could participate in the healing process by transesterification.) B) Photos of the healing process for a boronic ester sample with 75:25 DODT:PTMP. C) Control experiment demonstrating attempted healing of a network with 75:25 DODT:PTMP crosslinked via divinylbenzene, an irreversible diene.
63
Adhesion of the two separate pieces was noticeable within several minutes, and
after 3.5 days, the sample had healed to such an extent that the original scar had nearly
disappeared. At the end of 4 days, the materials could be manually stretched to more
than twice the original length without fracturing. The relatively long time required for
healing is consistent with the theory proposed by Rubinstein and coworkers in that
recovery of bonds across the fracture is the result of exchange of reactive components
(i.e., boronic acids and diols) between different bonded partners.117 The materials were
also evaluated for their ability to heal without the direct addition of water to the cut
surfaces under ambient conditions and in high humidity environments (85%). While the
networks healed in both cases, healing appeared to be faster and more efficient with the
minor addition of water. As previously reported, careful realignment when contacting the
cleaved pieces along the damaged interface proved to be critical.25
The increased healing efficiency observed when the two freshly cleaved surfaces
were wet prior to healing seemed to support the important role of bridge formation via
boronic ester exchange along the surface. Additional evidence was obtained to support
the proposed mechanism of healing. First of all, the absence of thiols after curing, as
confirmed by FTIR spectroscopy, would seem to exclude healing by thiol disulfide
exchange reactions, as has been described in previous publications.25,26 Additionally,
when the networks were cut, adjoined without water, and heated in a dry environment,
no healing was observed, further suggesting the mending process was likely not due to
exchange of a small amount of disulfides or transesterification of boronic esters.
Furthermore, control samples were prepared with the same thiol components but with
divinylbenzene, instead of the boronic ester, as the –ene component (Fig. 3-6c). In this
64
case, no healing was observed with or without the addition of water, even after 4 days.
As compared to the networks formed with the boronic ester crosslinker, there was no
stickiness along the failure interface when the two cut pieces were placed in contact,
implying that chain entanglement likely has very little role in the healing process.
Tensile testing experiments were conducted to quantify the efficiency of
healing.25 Three different compositions were considered (75:25, 50:50, and 25:75
DODT:PTMP), with particular attention being paid to the effect of healing time on the
recovery of tensile strength and elongation at break (Fig. 3-7). The tensile properties of
the healed samples were compared to those of the original uncut material. After healing
for 3 days at 85% humidity, good recovery of peak stress and strain at break
Figure 3-7. Self-healing of boronic ester network materials as evaluated by tensile
testing. A) Representative stress-strain plot for a network with 25:75 DODT:PTMP healed for 3 days at 85% humidity. B) Maximum stress and C) elongation at break as a function of the relative amount of dithiol (i.e., DODT:PTMP) and healing time.
65
were observed for all three compositions. When samples were healed for only 3 min,
efficient healing was observed only for the 25:75 samples, while the 50:50 and 75:25
samples showed only partial recovery.
The samples were also investigated for their ability to heal after multiple cycles of
damage and repair at the same site. These experiments were particularly important, as
the ability of repeated recovery is a primary advantage of healing through a mechanism
of reversible bond exchange. The material with a composition of 75:25 DODT:PTMP
composition demonstrated excellent healing in up to three cycles, with nearly full
recovery of tensile strength and elongation at break being observed after each cycle of
repair (Fig. 3-8).
Figure 3-8. Self-healing of boronic ester network materials after multiple cycles of
damage and repair as evaluated by tensile testing. A) Maximum stress and B) elongation at break of boronic ester network materials with 75:25 DODT:PTMP (Repair conditions: 3 days at 85% humidity).
3.4 Conclusions
These results indicate it is possible to achieve bulk-state self-healing via boronic
ester exchange. Repair was demonstrated to take place under ambient conditions, in
the presence of air, at room temperature, and in the absence of solvent, although
application of a thin layer of water to the freshly cleaved surfaces facilitated bond
exchange and increased the efficiency of healing. Despite being composed of
66
hydrolytically labile boronic esters, the materials were stable in humid environments and
even when completely submerged in water. The dynamic-covalent nature of the bonds
allowed healing to occur over multiple cycles, though the reversibility of all crosslinks did
lead to creep. Future studies will involve the incorporation of both reversible and
permanent crosslinks to investigate the effect on creep and healing efficiency. Beyond
this work, boronic ester incorporation could be expanded to networks made by a variety
of chemistries that go beyond the thiol-ene approach employed here. The reversibility of
the boronic ester linkages in the bulk state also suggests this functionality may have
applicability beyond self-healing, in other types of materials that rely on structurally
dynamic polymers.
67
CHAPTER 4 BALANCING STATIC AND DYNAMIC BONDS IN SELF-HEALING BORONIC ESTER
NETWORKS
4.1 Overview
When designing polymeric materials capable of both long-term durability and
end-of-life recycling, having the capacity to tune macroscopic structure as well as
mechanical properties is critical. The stability of most material’s initial macroscopic
shape is largely determined by its ability to relieve internal stresses developed during
synthesis.118 However, stability in materials crosslinked via dynamic bonds can also be
established post-synthesis, provided sufficient time for crosslink exchange. Materials
made from networks with dynamic bonds can function as static or dynamic networks,
depending on the rate of crosslink exchange. As a result, selection of the type of
dynamic linkage used for crosslinking greatly influences the overall behavior of the
resulting material.119,120 However, it is important to consider that networks containing
dynamic bonds do not consist solely of dynamic bonds.121 The majority of the bonds
present in these networks are static, not dynamic. Nevertheless, these static bonds
affect the dynamics of the network by restricting local mobility and dictating the
positioning of the reversibly reactive moieties after crosslink dissociation.122
A tremendous of interest has been paid to materials that contain dynamic
crosslinks. The reversible nature of the crosslinks allows, in many cases, self-healing
and reprocessability. However, dynamic exchange of crosslinks can also lead to creep
and stress relaxation, which prevents the materials from being useful for many
applications. Whether the shape of a particular material is stable or malleable is largely
determined by the selection, placement, and ratio of static and dynamic bonds present
in the network.
68
Control over crosslink density can be maintained in a dynamic network through
careful selection of crosslink exchange mechanism. Covalent adaptable networks
(CANs), defined as reversibly crosslinked polymers with sufficient reversible bonds to
allow the network to respond chemically to an external stimulus,73 can be divided into
two types based on the mechanism of exchange.7 In dissociative CANs, bonds are
continuously broken and re-formed. In this case, crosslink density depends on the
equilibrium between the associated and dissociated bonds. In associative CANS, bond
cleavage is accompanied by simultaneous formation of a new crosslink. As a result,
crosslink density remains constant. The dissociative mechanism can produce non-
equilibrium conditions at damaged interfaces, facilitating bridge formation, an important
step in the self-healing process.117 The associative mechanism, in contrast, affords
retention of crosslink density and mechanical properties.123 This latter mechanism has
been most often demonstrated in networks that undergo crosslink exchange via
traditional thermally-responsive transesterification reactions,21,124,125 though several
other chemistries have been considered recently.7,24,26,30,44,45,63,126-128 Despite the recent
advances in this area, there still exists a need for increased chemical diversity in the
field of associative CANs.
While boronic ester and diol transesterification reactions have been investigated
and characterized in small molecule contexts,129,130 this type of chemistry has been less
widely applied to polymeric systems. In a recent example of boronic ester
transesterification in a polymer network, a diol-functionalized polymer was crosslinked
with two types of difunctional boronic esters.72 Network healing could occur by a
combination of two mechanisms: hydrolysis/re-esterification (dissociative exchange)
69
and transesterification (associative exchange). Given the relatively low hydrolytic
stability of most boronic esters, a completely dry environment would be needed to
isolate transesterification as the only operative mechanism of bond exchange.129
Manipulating the contributions of hydrolysis and transesterification could lead to
materials that not only exhibit a rapid response to external stimuli but also a high
crosslink retention, enabling improved conservation of material properties. The
combination of two dynamic exchange mechanisms operating on different timescales
for self-healing material applications has previously been reported by other
authors.131,132 Boronic esters can potentially achieve similar properties with a single type
of crosslink.
Another way molecular structure can be harnessed to control the macroscopic
behavior of polymeric networks is through the selective placement of static bonds
between crosslinking points to complement and support connections with exchangeable
bonds. Permanent crosslinking enables retention of mechanical properties and reduces
the ability of the material to relax under an applied stress,133-135 as has been
demonstrated by swelling experiments on gels with various concentrations of
permanent crosslinker.136 By varying the quantity of permanent crosslinks, networks can
be designed so that an irreversible framework allows for better macroscopic structural
integrity, while a shifting equilibrium exchange in dynamic bonds provides a separate
healing behavior.
We are interested in directing macromolecular behavior through the design of
networks that marry dynamic and static chemistries. We sought to improve the
structural properties of self-healing materials through mechanistic control of the
70
exchange reaction and by the inclusion of permanent crosslinks within otherwise
reversible networks, expanding upon our previous investigations of reversibly
crosslinked polymer architectures in solution, such as stars,102,105,137-139 micelles,140-143
and gels,90,144 to explore dynamic networks in the bulk state.70
We determined that the boronic ester functionality would be well suited for
exploring the effect of architecture on shape stability and pliability in self-healing
materials through simultaneous stimuli-responsive associative and dissociative
exchange and the incorporation of irreversible crosslinks. We prepared three systems:
Figure 4-1. An illustration of the three prepared systems. A) Network with excess diol, B) network with irreversible crosslinks, and C) combination network with excess diol and irreversible crosslinks.
(i) boronic ester networks with various concentrations of excess diol for concurrent bond
exchange via dissociative hydrolysis/re-esterification and associative transesterification,
(ii) boronic ester networks that were augmented by the addition of irreversible
71
crosslinks, and (iii) boronic networks that contained both free diols and irreversible
crosslinks (Fig. 4-1). For each system, water absorption, stress-relaxation, and healing
behaviors were characterized to elucidate the relative roles of static and dynamic
crosslinks. While the initial two systems excelled at rearrangement and shape retention,
respectively, the third system yielded a material that favorably balanced both behaviors.
4.2 Experimental Section
4.2.1 Materials
4-((Allyloxy)methyl)-2-(4-vinylphenyl)-1,3,2-dioxaborolane (VPBE) was prepared
as previously reported,70 except the purification was modified to eliminate the need for
centrifugation. Dimethylsulfoxide-d6 (d-DMSO, Cambridge Isotopes, 99.9% D) was dried
overnight over 4 A molecular sieves. Dichloromethane (DCM, Sigma-Aldrich) was dried
using an anhydrous solvent system (Innovative technologies). 4-Vinylphenylboronic acid
(VPBA, Combi-blocks, 98%), 3-allyloxy-1,2-propanediol (APD, Acros Organics, 98%),
pentaerythritol tetrakis(3-mercaptopropionate) (PTMP, Sigma-Aldrich, 95%), 3,6-dioxa-
1,8-octanedithiol (DODT, TCI America, 95%), 2,2-dimethoxy-2-phenylacetophenone
(DMPA, Sigma-Aldrich, 99%), potassium chloride (BDH, 99%), potassium acetate
(Macron Chemicals, 99%), tri(ethylene glycol) divinyl ether (TEGDVE, Sigma-Aldrich,
98%), and molecular sieves (4 A, Mallinckrodt) were used as received.
4.2.2 Instrumentation and Analysis
Infrared spectra were collected on a Perkin Elmer Spectrum One FTIR
spectrometer equipped with a ZnSe crystal attenuated total reflectance (ATR)
accessory. Differential scanning calorimetry (DSC) measurements were performed on a
TA Instruments Q1000 equipped with a liquid nitrogen cooling accessory and calibrated
using sapphire and high purity indium metal. All samples were prepared in hermetically
72
sealed pans (4−7 mg/sample) and were referenced to an empty pan. A scan rate of 10
°C per minute and helium purge gas were used. Glass transition temperatures were
evaluated as the midpoint of a step change in heat capacity. Thermal experiments were
conducted as follows: samples were heated to 50 °C, followed by cooling at 10 °C per
minute to −80 °C, and then heated to 50 °C at 10 °C per minute. Data reported reflects
the average of the second and third heating scans. Tensile tests and stress relaxation
tests were performed on a TA.XTPlus texture analyzer from Texture Technologies with
a 5 kg load cell using test specimens in the form of dogbones that were sized according
to ASTM standard D1708 with a 15 mm grip-to-grip separation at 25°C and 50%
humidity. The tensile data reported are the average of five measurements collected with
a rate of 10 mm/min. The stress relaxation data were collected in duplicate with
samples placed at 10% strain at a rate of 120 mm/min and held under constant strain
for 5 min. The light used for photocuring was a UV nail gel curing lamp with four 9 W
bulbs (available online from ad hoc suppliers) with a peak emission at 360 nm and
intensity of 7.0 mW/cm2.
4.2.3 Synthesis and Experimental Procedures
Synthesis of thiol-ene networks. The synthetic procedure was the same as
previously described,70 with samples mixed, sonicated, and cured with 360 nm UV light
for 30 min using 1 wt% DMPA. All samples had a 75:25 ratio of DODT:PTMP, with the
specific composition of the vinyl component being varied to substitute either free diol
(1%, 3%, or 5% molar equivalent of vinyl) or irreversible crosslinker (80%, 85%, 90%,
97%, 98%, or 99% molar equivalent of vinyl) for boronic ester diene while still
maintaining a one-to-one ratio of thiol to vinyl groups. The same procedure was
followed for the reference material made with TEGDVE entirely in the place of VPBA.
73
ATR-FTIR spectroscopy was used to monitor thiol conversion using the S−H absorption
peak at 2590 cm-1. Thermal characterization of the final materials was performed by
DSC.
Water absorption studies. Rectangular samples with either 5% free diol, 80%
TEGDVE, 85% TEGDVE, 90% TEGDVE, or combined samples with 5% free diol and
80% TEGDVE were weighed and then submerged in water for 14 days. Samples of
each composition in duplicate were removed every 24 h, patted dry, and weighed
before re-immersing.
Humidity chamber preparation. Constant humidity environments were
prepared as previously reported.70,112 Briefly, 23% and 85% humidity atmospheres were
made by filling the bottom of sealed containers with saturated solutions of potassium
acetate or potassium chloride, respectively. These salts were selected to minimize
humidity variation within our typical laboratory temperature ranges, thereby optimizing
the environments for use in controlled healing experiments.
Network relaxation. For the stress relaxation experiments, all network
compositions were equilibrated for a minimum of 24 h in either a desiccator, to
approximate 0% humidity, or the desired salt humidity chamber. Samples were
transported to the testing instrument in the selected equilibration environments. Upon
removal from these chambers, the dogbones were immediately tested to minimize
exposure to the lab environment.
Network healing. Three representatives of each network type along with the
combination network, (0, 1, 3, 5% free diol, 80, 85, 90% TEGDVE, and combined 5%
free diol and 80% TEGDVE), were healed for 3 days under ambient conditions after
74
dabbing the completely separated cut interfaces with water and reconnecting the two
individual pieces along the freshly exposed interface. Samples with 5% free diol were
also investigated for healing with heat (1.5 and 3 h at 70 ˚C), omitting the addition of
water to the cut interface. For the samples containing irreversible TEGDVE crosslinks, a
longer healing time (7 days) and up to 3 cycles of healing were also examined. Multiple
healing cycles were carried out by repeatedly damaging and healing the material at the
same location, with a 6 h drying time between cycles. After removal from the chambers,
all samples exposed to humidity were vacuum dried for over 18 h with intermittent
storage in a desiccator until their final weights were within 0.3% of their weights before
exposure to humidity. All samples were characterized by tensile testing.
4.3 Results and Discussion
4.3.1 Networks with Boronic Esters and Free Diols
We previously demonstrated that bulk macroscopic healing could be achieved in
networks capable of boronic ester exchange.70 While efficient healing was possible,
rapid exchange of the dynamic crosslinks also led to creep over extended times. To
provide insight into the the relationship between relaxation behavior and healing
efficiency, we decided to consider two variables that would allow tuning of the dynamic
behavior such that rapid and efficient healing could be achieved in a system that
demonstrates minimal creep. By incorporating free diol functionality within a network
crosslinked via boronic esters, crosslink exchange can be achieved by
transesterification, thereby providing an alternative mechanism for dynamic behavior.
On the other hand, to minimize the potentially detrimental effects of stress relaxation
and creep, we also investigated the effect of incorporating irreversible crosslinks within
boronic ester-containing materials (Fig. 4-1).
75
Boronic ester-crosslinked networks with varying concentrations of free diol were
made through thiol-ene chemistry.70 These networks were fully cured as determined by
ATR-FTIR spectroscopy via the absence of the peak that arises from the thiol at 2590
cm-1. Varying amounts of 1%, 3%, and 5% were used to prepare networks with different
ratios of boronic esters to free diols. All compositions resulted in glass transition
temperatures below room temperature, which led to sufficient molecular mobility to heal
at ambient temperature (Table 4-1).
Table 4-1. DSC results of low glass transition temperature polymers showing values below room temperature.
Free Diol (%) Permanent Crosslinker (%)
Tg (˚C)
5 -25 3 -20 1 -20 80 -55 85 -53 90 -51 5 80 -52
Since the presence of water has a significant influence on the rate of hydrolytic
self-healing as well as the predominant healing mechanism when hydrolysis and
transesterification are occurring simultaneously, the dynamic behavior of these
materials was evaluated as a function of the humidity of the surrounding environment.
However, beforehand, samples were also completely immersed in water to investigate
their structural integrity. Even samples with the highest concentration of free diols
present in the network were stable, with only a slight increase in weight (~5%) being
observed over the course of 13 days (Fig. 4-2). All networks, regardless of free diol
content, became opaque over the course of the first day while submerged in water,
suggesting that their hydrophobicity led to a collapsed skin layer forming on the surface
76
the of the samples, which likely limited infiltration of water (and the hydrolysis that would
have resulted). However, the presence of free diol did seem to enhance chain mobility
and resulted in samples that underwent a macroscopic shape change from a thin
rectangular shape to a more spherical form when submerged in water. This shape
change occurred relatively quickly (12 h) after submersion and was not observed for the
samples that contained no free diol.
Figure 4-2. Percent mass change of the network that contained 5% free diol while immersed in water.
The observed shape change during water exposure, as well as observations of
creep even while stored in a dry environment, suggested the importance of more
thoroughly investigating the effect of free diol content and the role of hydrolysis on the
dynamic nature of these materials. Stress relaxation experiments were conducted to
provide insight into the extent of crosslink exchange and the propensity for creep.
Samples were conditioned in chambers of varying humidity before being immediately
characterized for stress relaxation under tension. Stress relaxation times () were
determined by measuring the time needed for the stress to drop to 1/e of its original
value.126 Generally speaking, there appeared to be a significant increase in network
77
mobility (i.e., lower ) for samples exposed to higher humidity environments or
containing increasing concentrations of free diols. For example, when stress relaxation
was investigated after conditioning the networks in dry atmospheres, where boronic
ester hydrolysis should be minimal and the crosslink density of all samples could be
assumed to be similar (Fig. 4-3A), increasing the free diol content resulted in more rapid
network relaxation. For example, at 0% humidity the relaxation time dropped from near
100 s down to 25 s for samples with 3 and 5% free diols, respectively. These results
suggest that more free diol leads to faster diol-boronic ester exchange (Table B-1),
which is consistent with free diols leading to accelerated bond exchange due to
transesterification. At 23% humidity (Fig. 4-3B), increasing free diol content (0, 1, 3, and
5%) even more significantly reduced relaxation times (107, 92, 41, and 7.3 s,
respectively). At 85% humidity (Fig. 4-3C), hydrolysis appears to be the dominant,
mechanism of crosslink exchange, with all compositions demonstrating relaxation times
of approximately 1-2 s. These results suggest that both increased concentration of free
diol and increased humidty lead to accelerated crosslink exchange in boronic ester-
containing materials.
At relatively low humidities, the rate of exchange is strongly correlated with the
concentration of free diols, as crosslink cleavage occurs primarily via transesterification.
On the other hand, at higher humidities crosslink cleavage occurs mainly via hydrolysis,
leading to a minimal effect of free diol concentration on the rate of stress relaxation.
Given this, we demonstrated that networks containing 5% free diol content could be
reshaped after prolonged exposure to water (i.e., after immersion or exposure to high
78
humidity environments) (Fig. 4-3D), suggesting these materials may have utility for
recycling applications.5
Figure 4-3. Stress relaxation of 0, 1, 3, and 5% free diol networks after exposure to humidity. A) 0%, B) 23%, and C) 85% humidity. D) Qualitative demonstration of 5% free diol network remolding after immersion in water and exposure to 85% humidity.
Given the apparent exchange reactions occurring in both humid and dry
environments during the stress relaxation experiments, healing in both environments
was expected for networks containing free diol, unlike networks comprised purely of
boronic esters.70 To examine this, samples with and without free diol were exposed to
humidity over time to determine whether they were capable of hydrolytic self-healing
(i.e., hydrolysis and re-esterification). The materials were then subjected to tensile
testing (Fig. 4-4A and Fig. 4-4B). Exposure of damaged free diol networks to high
humidity resulted in healing similar to that observed with 100% boronic ester networks,
with high healing efficiencies of 95% from peak stress. Healing was also investigated
79
after the samples were exposed to dry environments. While networks with no free diol
content demonstrated no healing,70 which is to be expected given that bond hydrolysis
and re-esterification are expected to be slow under these conditions,70
Figure 4-4. Evaluation of healing by tensile testing after exposure to humid environments (85% humidity for 3 days) for A) peak stress B) strain at break. C) After exposure to dry conditions (under vacuum and heating at 70 °C) for peak stress and D) strain at break.
which is to be expected given that bond hydrolysis and re-esterification are expected to
be slow under these conditions, networks containing 5% free diol samples did
demonstrate a modest degree of healing, with efficiencies of 45% from peak stress and
18% from strain at break being observed after 3 h at 70 °C. These results suggest that
transesterification between free diols and boronic esters leads to crosslink exchange
and healing even in the absence of water, a result that is consistent with the stress
relaxation results observed for these networks at 0% humidity (Fig. 4-4C and Fig. 4-4D).
80
The relatively low healing efficiency likely results from the slow rate of exchange
between diols and boronic esters.72
4.3.2 Networks with Permanent Crosslinks and Boronic Esters
While networks crosslinked via dynamic bonds can be healable, reprocessable,
and respond to certain environmental cues, a potential complication arises on a
sufficiently long time scale when the material is under stress by tension, compression,
bending, or gravity. In these cases, dynamic bond exchange can also lead to frequency-
dependent rheological behavior, such as creep or plastic deformation. One solution to
minimize the effects of creep or stress relaxation is to imbue the networks with a small
fraction of irreversible crosslinks that impart structural integrity. We decided to evaluate
the effect of including permanent crosslinks within our boronic ester materials, with
particular attention being paid to the amount of irreversible crosslinks that could be
included without significantly compromising the benefits of the dynamic organoboron
bonds.
TEGDVE was included during curing of dithiol, tetrathiol, and boronic ester diene
to create networks that included both dynamic boronic esters and irreversible crosslinks.
Quantitative curing was observed by ATR-FTIR, and all networks had glass transition
temperatures below room temperature (Table 4-1). Because free diols were not
intentionally incorporated in these networks, it was envisioned that healing would rely on
hydrolysis and re-esterification of boronic esters. As compared to the materials
previously considered, complete immersion of the samples in water led to an initial
increase in mass due to water absorption followed by a gradual weight loss over time
(Fig. 4-5A). The reduction in mass after prolonged exposure to water is likely the result
of the increased hydrophilicity of the samples that resulted from incorporation of the
81
relatively polar permanent crosslinker. This idea is further supported by the fact that the
networks containing increasing amounts of TEGDVE became less opaque when
submerged in water.
Figure 4-5. The effect of water on permanently crosslinked samples. A) Change in mass over time for networks containing 80, 85, and 90% TEGDVE (i.e., permanent crosslinks) submerged in water and B) images of the samples after exposure in water for 2 days.
Stress relaxation experiments were carried out for samples with various
concentrations of permanent crosslinks that were equilibrated beforehand in three
different humidities: 0, 23, and 85%. As shown in Figure 6a, the networks aged in the
lowest humidity environments and that contained the higher incorporation of TEGDVE
demonstrated behavior that was most similar to covalently crosslinked networks. As the
concentration of boronic ester was increased and the amount of permanent crosslinks
was decreased, the maximum stress at the start of the experiment increased and a
gradual decrease in the slope was apparent with time. This increase in stress is to be
82
expected and has been previously observed in dynamic covalent networks and has
been attributed to “catching” that occurs as the dynamic bonds reform, increasing the
maximum force needed for strain.145 By the same reasoning, this increased dynamic
behavior correlating with the increasing concentration of boronic ester, also leads to a
greater magnitude of relaxation. This is also true for samples equilibrated in the
absence of moisture, because even atmospheric water present during testing could
cause slight shifts in equilibrium, leading to a gradual relaxation and a lower predicted
plateau region that corresponds to a decrease in network crosslink density.
All samples pre-equilibrated in a dry environment had relaxation times greater
than 100 s (Table B-1), suggesting very slow bond exchange consistent with minimal
effects of hydrolysis and re-esterification in the absence of water. Networks exposed to
23% humidity (Fig. 4-6B) relaxed faster, as expected, with the plateau level of the
curves decreasing with increasing concentration of the hydrolytically-labile boronic ester
bonds. Following the same trend of increased humidity leading to faster bond exchange,
networks exposed to the highest humidity of 85% (Fig. 4-6C) demonstrated the fastest
relaxation times. For all networks exposed to humid environments, an increased content
of permanent crosslinks led to reduced and slower stress relaxation.
Tensile testing was used to elucidate the effect of incorporating permanent
crosslinks on healing efficiency (Fig. 4-7A and Fig. 4-7B). Healed samples
demonstrated maximum stress (86%) and strain at break (90%) values that were within
error of the virgin samples for the networks in which 80% of the crosslinks were
permanent (i.e., 20% boronic ester crosslinks).
83
Figure 4-6. Stress relaxation data for 80, 85, and 90% permanent crosslinker networks. A) With exposure to 0% humidity. B) With exposure to 23% humidity. C) With exposure to 85% humidity.
The incorporation of permanent crosslinks in dynamic-covalent networks leads to an
increased potential for irreversible bond cleavage during damage and thus a lower
possibility of full recovery during healing. Therefore, repeat healing experiments were
undertaken in which samples were repeatedly cut and healed at the same location. The
samples with 80% of the crosslinks being permanent were selected for testing due to
their excellent healing and moderate stress relaxation at low humidity. After three
healing cycles, both peak stress and strain at break were within error of the values
obtained for the virgin materials prior to damage (Fig. 4-7C and Fig. 4-7D). These
results suggest that healing can be efficient even when the majority of crosslinks
present in the network are non-dynamic.
84
Figure 4-7. Self-healing of boronic ester networks as evaluated by tensile testing. A) Maximum stress and B) elongation at break as a function of the relative amount of permanent crosslinker (TEGDVE) and healing time. C) Maximum stress and D) elongation at break of boronic ester network materials with 80% TEGDVE after multiple cycles of damage and repair.
4.3.3 Combined Free Diol and Permanent Crosslinker Networks
Given the separate benefits of incorporating free diols for faster healing and
permanent crosslinks for reduced stress relaxation and creep, we prepared networks
that contained both components APD and TEGDVE. By combining both types of
crosslinker in a single material, it was anticipated that the less desirable aspects of each
system could be mitigated (i.e., creep in the free diol networks and low healing
efficiency in the permanently crosslinked networks. To test this hypothesis, a network
that contained 5% free diol and 80% irreversible diene was selected, as these
concentrations, when incorporated individually, showed optimal behaviors. These
85
networks were observed to undergo a mass loss of approximately 6% over nine days
(Fig. 4-8). The presence of free diol and TEGDVE led to a network that was more
hydrophilic than those considered earlier, which may have facilitated hydrolysis on the
exterior of the material. However, no further mass loss was observed after 9 days, the
networks became opaque over time, and their shapes remained stable, suggesting that
the materials are relatively stable to hydrolysis over extended periods.
Figure 4-8. Water absorption of samples containing both 5% free diol and 80% permanent crosslinker.
The stress relaxation behavior for these materials showed that the mechanical
behavior most closely resembled that of the networks that contained permanent
crosslinks alone (i.e., without free diol) (Fig. 4-9, Table B-1). As the substitution of 80%
irreversible diene for boronic ester diene affects a greater fraction of the network than
incorporation of the relatively small amount of 5% free diol, this trend observed in
relaxation behavior is expected. These results indicate that the propensity of these
hybrid materials to undergo creep is relatively minimal.
Finally, while the shape and structural integrity in response to external forces was
much improved, an evaluation of the effect of combining both approaches on healing
efficiency was carried out. Healing efficiency was determined by tensile experiments,
86
Figure 4-9. Stress relaxation curves run in duplicate of samples containing 5% free diol, 80% permanent crosslinker, and both. A) After exposure to 0% humidity. B) After exposure to 23% humidity. C) After exposure to 85% humidity.
Figure 4-10. Healing measured after 3 days by tensile testing for samples containing 5% free diol, 80% permanent crosslinker, or both. A) Peak stress and B) strain at break recovery.
87
comparing the peak stress and strain at break of the virgin and healed materials. The
extent of healing was measured after three days, as samples with only free diol showed
good healing efficiency in this timeframe while the samples with permanent crosslinker
required longer for similar healing. The hybrid networks demonstrated healing behavior
that was intermediate to those of the networks with only free diol or permanent
crosslinks alone (Fig. 4-10).
4.4 Conclusions
These experiments demonstrate that polymeric networks crosslinked via
dynamic-covalent boronic esters can be modified to improve healing efficiency and to
reduce stress relaxation and creep by the incorporation of free diols or permanent
crosslinks, respectively. While materials crosslinked exclusively via boronic esters
undergo exchange by hydrolysis and re-esterification or, potentially, by boronic ester
metathesis, including free diols within the networks allows an additional method of bond
exchange to occur by transesterification. Including this associative mechanism of
exchange allows for samples leading to higher healing efficiencies at the expense of
faster stress relaxation and creep. On the other hand, stress relaxation and creep can
be moderated by the incorporation of permanent crosslinks that lend increased
structural integrity, albeit with a loss of healing efficiency. In all cases, the mechanical
and healing behavior of the materials was dependent on the humidity of the
environment to which the networks were exposed and the healing time. Arguably, the
materials that demonstrated the best balance of rapid bond exchange to promote
healing and static crosslinks to provide structural integrity were those that contained
both free diols and permanent crosslinks. These results may provide insight into the
88
design of dynamic-covalent materials that rely on other examples of reversible covalent
bonds.
89
CHAPTER 5 BORONIC ESTER VITRIMERS
5.1 Overview
In 2011, Leibler and coworkers introduced a new category of materials, organic
strong glass formers known as vitrimers, possessing the structural integrity of
thermosets at typical-use temperatures and the processing properties of thermoplastics
at higher temperatures.125 This unique performance was a consequence of reduced
viscosity caused by thermally-responsive transient molecular networks. Specifically,
networks are made of linkages undergoing dynamic isodesmic reactions so that a
constant number of crosslinks is present at all times. The conservation of crosslinks
causes the molecular framework to endure even at high temperatures and upon
exposure to solvent.124 Thus, this type of material remains insoluble under all conditions
like permanent networks. However, unlike permanent networks, dynamic networks are
capable of dissipating stress by thermally-induced bond rearrangement.146-156 In
contrast to typical polymers with William-Landers-Ferry behavior above their transition
temperature,157-159 vitrimers show a unique and slower Arrhenius-dependent decrease
in viscosity with increasing temperature. The freezing transition temperature defined
when η = 1012 Pa·s is commonly used to distinguish the solid-to-liquid transition
viscosity124,125,157,159,160 so that different vitrimers can be compared, and a correlation
between exchange rate, relaxation time, and vitrimer attributes can be made. Early
vitrimers were based on catalyzed transesterification reactions.21,22,24,124,161,162 More
recent work has expanded dynamic chemistry selection to include transamination of
vinylogous urethanes44 and transalkylation of triazolium salts.45 Vitrimeric materials
90
allow new opportunities in recycling, adhesion/welding, and post-polymerization nano-
imprinting applications.163
Vitrimers, with fast exchange chemistries, ideally have glass temperatures above
the freezing transition temperature.7 In this case, above the transition temperature the
material behavior is controlled by the exchange rate and not diffusion. A general way to
raise the glass transition temperature and stiffness is through more rigid chemistry.
Specifically, for thiol-ene networks, thiol reagents with ring structures can be pre-formed
with thiourethane chemistry using commercially available isocyanates.164-167
Using thiol-ene chemistry is desirable for a number of reasons.
Photopolymerization to form vitrimer networks allows unique control over the time and
location of curing, opening the door to more applications, particularly patterning
applications. Radical thiol-ene “click” chemistry takes place under mild reaction
conditions.113,116 The step-growth mechanism of thiol-ene chemistry puts dynamic
bonds in the network backbone in contrast to a chain-grown polymer with pendent
dynamic groups.163
Boronic esters are used as the dynamic group in large molecules for a number of
reasons. Previously, boronic esters have been noted for high thermal stability in organic
frameworks168 and high thermodynamic stability in context of amorphous polymer
networks.72 The B−O bond has a bond disassociation energy of 124 kcal/mol,169 which
is higher than C−C bond disassociation energy of ~80 kcal/mol. The large kinetic
tunability of boronic ester transesterification72 allows for application-specific tailoring of
polymer systems.
91
The susceptibility of boronic esters to hydrolysis170 remains a significant
challenge to employing boronic esters as the dynamic chemistry in vitrimer applications.
Two approaches exist for addressing this issue. The network chemistry can be selected
to minimize water infiltration,70 and the boronic ester chemistry can be optimized to be
more resistant to hydrolysis.171,172 One way to modify the boronic ester is by changing
the diol precursor. Selecting a 1,3-diol instead of a 1,2-diol leads to a six-membered ring
boronic ester, which is expected to be more stable than a five-membered ring boronic
ester due to the reduction in ring strain.171 Previous bulk polymer networks with boronic
esters have all used five-membered rings.70-72
Here the preparation of new boronic ester vitrimers with five and six-membered
ring boronic esters is reported. Model studies confirmed that the six-membered ring
boronic ester favored the boronic ester in equilibrium and hydrolyzed slower than the
five-membered ring boronic ester. When heated, the thiol-ene networks showed stress-
relaxation on two different time scales corresponding to the five and six membered ring
boronic ester exchange chemistries. Both materials demonstrated Arrhenius behavior.
5.2 Experimental Section
5.2.1 Materials
Dimethylsulfoxide-d6 (d-DMSO, Cambridge Isotope, 99.9% D) and
Dichloromethane (DCM, Sigma-Aldrich) were dried overnight over 4A molecular sieves.
4-Vinylphenylboronic acid (VPBA, Combi-blocks, 98%), 3-allyloxy-1,2-propanediol
(APD, Acros Organics, 98%), pentaerythritol tetrakis(3-mercaptopropionate) (PTMP,
Sigma-Aldrich, 95%), 3,6-dioxa-1,8-octanedithiol (DODT, TCI America, 95%), 2,2-
dimethoxy-2-phenylacetophenone (DMPA, Sigma-Aldrich, 99%), tri(ethylene glycol)
divinyl ether (TEGDVE, Sigma-Aldrich, 98%), triethylamine (TEA, Sigma-Aldrich,
92
99.5%), isophorone diisocyanate (IPDI, Sigma-Aldrich, 98%), tetrahydrofuran (THF,
EMD, 99.5%), acetone (Fisher Chemical, 99.5%), and molecular sieves (4 A,
Mallinckrodt) were used as received.
5.2.2 Instrumentation and Analysis
1H NMR (500 MHz), 13C NMR (125 MHz), and 11B NMR (160 MHz) spectra were
recorded using an Inova 500 spectrometer. For 11B NMR spectroscopy, 5-mm thin-
walled quartz NMR tubes were used. Chemical shifts are reported in parts per million
(ppm) downfield relative to tetramethylsilane (TMS, 0.0 ppm). Multiplicities are reported
using the following abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet;
br, broad. High-Resolution Infrared spectra were collected on a Thermo Nicolet 5700
FTIR spectrometer equipped with a single bounce diamond stage attenuated total
reflectance (ATR) accessory. Molecular weights and molecular weight dispersities were
determined by size exclusion chromatography (SEC). The UV lamp used for
photocuring was a UV nail gel curing lamp (available online from ad hoc suppliers) with
four 9 W bulbs and peak emission near 360 nm with 7.0 mW/cm2.
5.2.3 Synthesis and Experimental Procedures
Synthesis of 4-((allyloxy)methyl)-2-(4-vinylphenyl)-1,3,2-dioxaborolane
(VPBE). 4-Vinylphenylboronic acid (2.00 g, 13.5 mmol) and 3-allyloxy-1,2-propanediol
(1.62 g, 12.3 mmol) were stirred in dry DCM (20 mL) with molecular sieves (4 A, ca. 2 g)
at room temperature. The progress of the reaction was monitored by 1H NMR
spectroscopy, and additional molecular sieves were added as needed to drive the
reaction to completion. After confirming reaction completion, VPBE was purified by
filtering and concentrating to give the final colorless-to-pale yellow liquid (1.87 g, 7.64
mmol, 62%). The product was characterized by 1H NMR and 13C NMR spectroscopy.
93
Synthesis of 6-membered ring boronic ester (6MRBE). 4-Vinylphenylboronic
acid (2.00 g, 13.5 mmol) and 2-(allyloxymethyl)-2-ethyl-1,3-propanediol (2.15 g, 12.3
mmol) were stirred in dry DCM (20 mL) with molecular sieves (4 A, ca. 2 g) at room
temperature. The progress of the reaction was monitored by 1H NMR spectroscopy, and
additional molecular sieves were added as needed to drive the reaction to completion.
After confirming reaction completion, VPBE was purified by filtering and concentrating to
give the final colorless-to-pale yellow liquid (2.39 g, 8.34 mmol, 68%). The product was
characterized by 1H NMR and 13C NMR spectroscopy. 1H NMR (300 MHz, d-DMSO):
delta (ppm) 7.63 (d, 2H), 7.44 (d, 2H), 6.70 (dd, 1H), 5.87 (d, 1H), 5.83 (m, 1H), 5.24 (d,
1H), 5.17 (d, 1H), 5.11 (d, 1H), 3.97 (d, 1H), 3.92 (d, 2H), 3.83 (d, 1H), 3.42 (s, 2H),
1.37 (q, 2H), 0.83 (t, 3H). The boron-bound carbon was not observed due to
quadrupolar relaxation. 13C NMR (75 MHz, d-DMSO): delta (ppm) 139.4, 136.8, 135.1,
134.0, 125.5, 116.6, 115.2, 71.1, 68.8, 66.5, 38.3, 23.3, and 7.33. Elemental analysis:
Calcd for C17H23BO3: C, 71.35%; H, 8.10%. Found: C, 71.24%; H, 8.07%.
Synthesis of 5-membered ring reference boronic ester (FRBE) (1).
Phenylboronic acid (2.01 g, 16.5 mmol) and 3-allyloxy-1,2-propanediol (1.97 g, 14.9
mmol) were stirred in dry DCM (20 mL) with molecular sieves (4 A, ca. 2 g) at room
temperature. The progress of the reaction was monitored by 1H NMR spectroscopy, and
additional molecular sieves were added as needed to drive the reaction to completion.
After confirming reaction completion, SRBE was purified by filtering and concentrating
to give the final colorless-to-pale yellow liquid/white solid mixture at room temperature
and completely liquid at ~50 °C (2.35 g, 9.95 mmol, 67%). The product was
characterized by 1H NMR and 13C NMR spectroscopy. 1H NMR (300 MHz, d-DMSO):
94
delta (ppm) 7.69 (d, 2H), 7.49 (t, 2H), 7.40 (t,1H), 5.88 (m, 1H), 5.25 (d, 1H), 5.13 (d,
1H), 4.74 (m, 1H), 4.38 (dd, 1H), 4.09 (t, 1H), 4.02 (m, 2H), 3.57 (m, 2H). The boron-
bound carbon was not observed due to quadrupolar relaxation. 13C NMR (75 MHz, d-
DMSO): delta (ppm) 134.9, 134.4, 134.0, 131.5, 127.9, 127.3, 116.5, 75.9, 71.4, and
67.5. Elemental analysis: Calcd for C12H15BO3: C, 66.10%; H, 6.93%. Found: C,
65.48%; H, 6.72%.
Synthesis of 6-membered ring reference boronic ester (SRBE) (2).
Phenylboronic acid (2.01 g, 16.5 mmol) and 2-(allyloxymethyl)-2-ethyl-1,3-propanediol
(2.60 g, 14.9 mmol) were stirred in dry DCM (20 mL) with molecular sieves (4 A, ca. 2 g)
at room temperature. The progress of the reaction was monitored by 1H NMR
spectroscopy, and additional molecular sieves were added as needed to drive the
reaction to completion. After confirming reaction completion, SRBE was purified by
filtering and concentrating to give the final colorless-to-pale yellow liquid/white solid
mixture at room temperature and completely liquid at ~50 °C (3.18 g, 11.4 mmol, 77%).
The product was characterized by 1H NMR and 13C NMR spectroscopy. 1H NMR (300
MHz, d-DMSO): delta (ppm) 7.69 (d, 2H), 7.44 (t, 1H), 7.34 (t, 2H), 5.86 (m, 1H), 5.25
(d, 1H), 5.13 (d, 1H), 3.97 (d, 1H), 3.94 (d, 2H), 3.86 (d, 1H), 3.35 (s, 2H), 1.39 (q, 2H),
0.85 (t, 3H). The boron-bound carbon was not observed due to quadrupolar relaxation.
13C NMR (75 MHz, d-DMSO): delta (ppm) 134.9, 133.5, 130.7, 116.4, 71.4, 68.7, 66.4,
38.2, 23.1, and 7.2. Elemental analysis: Calcd for C15H21BO3: C, 69.26%; H, 8.14%.
Found: C, 69.14%; H, 8.13%.
Synthesis of Bisthiol oligomer. Triethylamine (0.103 g, 1.01 mmol) was mixed
with 3,6-dioxa-1,8-octanedithiol (5.47 g, 30.0 mmol) and a portion of the acetone (100
95
mL). Isophorone diisocyanate (4.43 g, 20.0 mmol) was then added to this mixture while
stirring. The solution was allowed to react for 7 h at room temperature under nitrogen.
Reaction completion was verified by the disappearance of the isocyanate peak at 2260
cm-1 as shown in ATR-FTIR spectra. Acetone was evaporated at 50 °C under vacuum.
The Mn as measured by SEC with polystyrene standards was 1150 g/mol.
Synthesis of Thiol-ene Networks. The boronic ester networks with bisthiol
oligomer were synthesized as was previously described,70 except with the addition of
dry THF in a ratio of 2/3 of the total thiol-ene reagent mass prior to cure and a THF
removal step post-cure. Samples with bisthiol oligomer required THF to become fully
miscible and be uniformly mixed. The removal of THF was done by heating at 80 ˚C
while pulling vacuum for a week.
Model Hydrolysis and Transesterification of FRBE and SRBE. The
equilibrium was studied on samples ca. 0.1 M, prepared by dissolving a weighed
amount of boronic ester in 0.5 mL of d-DMSO. The hydrolysis constants, in mol-1-L, are
given below (Table 5-1).
Table 5-1. Hydrolysis constants measured by 1H NMR
Sample mester (g) xacid xdiol xH2O KHydrolysis
FRBE(1) 0.0245 0.68 0.63 4.745 0.2435 FRBE(2) 0.0271 0.82 0.75 7.485 0.1879 SRBE(1) 0.0152 0.18 0.15 4.97 0.0110 SRBE(2) 0.0152 0.315 0.235 9.68 0.0088
The error in the measurements, ca. 20%, comes mostly from weighting, and it
could be improved by weighting larger amounts. As the ratio KFRBE/KSRBE was of
interest, this ratio was additionally measured precisely in a mixture of FRBE and SRBE,
96
where KFRBE/KSRBE = ([1’]/[1])/([2’]/[2]) (Fig. 5-3). A quantitative proton spectrum of the
mixture was taken with a relaxation delay of 60 s, a 5 s acquisition time and a 45 pulse.
The hydrolysis was a two step process, and a reaction order or a kinetic
equation could not be assigned. The apparent hydrolysis rate could not be determined
by NMR easily, since a sample of 0.1 M ester and 10 M D2O, which gave good
quantitative data in 25 s, and satisfied the condition of negligible variation in the
concentration of D2O during the reaction, was past the initial rate regime (the initial
linear part (Fig. 5-1)) in a couple of minutes, even for the slower hydrolyzing SRBE.
Figure 5-1. ln(x) ln(x0) vs. time for the hydrolysis of 0.1 M SRBE with 10 M D2O in d-
DMSO, at 25 ˚C.
An attempt was made at measuring the k’1a, k’1b, and k’2 in an inversion/recovery
difference experiment using the NOESY1D pulse sequence. Selective inversion of the
proton at 4.74 ppm in 1 revealed population transfer at 3.78 and 4.53 ppm, which were
assigned to 1a and 1b, respectively, based on their chemical shifts. Similarly, selective
97
inversion at 3.86 ppm in 2, revealed the signal at 3.36 ppm in 2a. In an array of mixing
time = 50, 100, 150, 200 and 250 ms, the intensity of these signals stayed the same.
To measure the transesterification rate between the diol and ester, a signal of
one of them, was selectively inverted, i.e. its magnetization is changed from 1 to 1. In
time, the exchange process transferred this negative magnetization to the isochronous
signal in the diol, decreasing its intensity. In a competing process, both these signals
relaxed to the equilibrium, with the rates the inverse of their T1. The evolution of the
intensity of both signals in exchange was monitored by acquiring spectra after a delay
which was arrayed. Two sets of data were collected for each sample, one in which the
first pulse was a selective inversion one, and one in which it was a hard 180. The
exchange rate was fitted in the first data set, and the two T1s in the second, using the
CIFIT program by Alex Bain,173 iterating until the values converge.
For the experiment to work, a significant exchange ca. 10% was needed, in a
time comparable to T1. At 25 C, this was the case for the transesterification of 1, but
not 2, therefore the measurements were made at 75 C. Ideally both isochronous
signals considered in the experiment should not overlap other signals, but this is not the
case for 1, therefore the whole region 3.24 3.44 ppm was selectively inverted. Only
protons from diol 1’ (H1a, H1b, H3a and H3b in 1 allylglycerol) are in this region, so
interfering NOES in 1 are not a problem, and we had to assume that all T1s of these
protons are equal, which was not the case, but had little effect on the results. The signal
monitored in the ester was one of H3a and H3b, at 4.38 ppm. In the case of 2 and 2’,
98
the methylene protons in the ethyl groups are free of overlap, and the ones in the ester
were selectively inverted.
Glass Transition Temperature Determination. Dynamic mechanical analysis
was performed on a Q800 DMA (TA Instruments), and the glass transition temperature
(Tg) was defined as the peak of the tan delta curve. For glassy films containing FRBE
and SRBE, the tan delta was monitored from 20 to 100 °C, using a ramp rate of 5
°C/min, a frequency of 1 Hz, and a fixed oscillatory strain of 0.025%. To determine the
ultimate Tg of the polymer films, three scans were performed and the average of the tan
delta peak of the second and third scans was taken as the Tg value.
Stress Relaxation. Stress relaxation experiments were performed on a Q800
DMA. The sample was equilibrated at the selected temperature for 3 min, followed by
an immediate application of strain and monitoring of stress decrease over approximately
10 min for the five-membered ring boronic ester network and 30 min for the six-
membered ring boronic ester network. For temperature step experiments, where
multiple temperatures were tested, a small strain of 1% was used to avoid excessive
sample deformation over multiple cycles.
99
Figure 5-2. Reagents for subsequent studies. A) Model compounds for NMR spectroscopy hydrolysis and transesterification experiments B) Reference two boronic ester dienes and –ene functional diols C) Di- and tetra-functional thiols for network formation.
5.3 Results and Discussion
The design goals for the polymer networks in this study were to synthesize a
more hydrolytically stable boronic ester network so that temperature change would only
affect the reaction rate not the equilibrium crosslink density, raise the Tg compared to
previous boronic ester networks, ideally achieving a Tg above room temperature, and
reduce the total amount of boronic ester needed while retaining boronic esters between
all crosslinking points.
5.3.1 Synthesis and Comparison of Relative Hydrolysis and Transesterification of Model Boronic Esters
As mentioned above, six-membered ring boronic esters are known to be more
hydrolytically stable compared to five-membered ring boronic esters as they are more
thermodynamically stable.171 Because this could not be examined by looking the
chemical structure in the bulk polymer samples, small molecule model compounds in
solution were studied (Fig. 5-2B). As it had previously been reported that the five-
membered ring boronic ester hydrolyzed within minutes at room temperature,70 the
hydrolysis was initially investigated by 1H NMR spectroscopy at -23 ˚C. The five-
membered ring boronic ester appeared already hydrolyzed on the first scan, and the
six-membered ring boronic ester was already partially hydrolyzed within the first few
scans. A higher temperature was then investigated, and the six-membered ring boronic
ester, even in the presence of 40 times molar excess of water at 65 ˚C still appeared to
be in equilibrium with the unhydrolyzed boronic ester. Further investigations were
performed at room temperature. Equilibrium was measured for samples with only the
100
five-membered ring boronic ester (1) or only the six-membered ring boronic ester (2)
and also for a 1:1 mixture of 1 and 2 in a single sample (Fig. 5-3). As this was a model
experiment meant to compare the two boronic esters, the ratio of the equilibriums was
understood to be the important result. Since the environment of the bulk polymer would
be drastically different from the solution, the numerical values were expected to change,
while the general relationship, in bulk only measurable by mechanical properties,
remained the same. The solution measurements from the individual and mixed
component experiments were in agreement. The measurements from the individual
experiments gave a ratio of K1/K2 = 22 and the measurement from the mixed sample
gave a ratio of K1/K2 = 23.
Next, the rate of hydrolysis was more closely examined. The hydrolysis was a
two-step process, and a reaction order or a kinetic equation could not be assigned. The
apparent hydrolysis rate could not be determined by NMR easily, since a sample of 0.1
M ester and 10 M D2O, which gave good quantitative data in 25 s, and satisfied the
condition of negligible variation in the concentration of D2O during the reaction, was
past the initial rate regime in a couple of minutes, even for the slower hydrolyzing 2.
From 1H NMR spectroscopy inversion/recovery difference experiments (NOESY1D), it
was confirmed that the rate determining step was the hydrolysis of the monoester. In
the shortest measurable mixing time, 50 ms, the boronic ester was already in
equilibrium with the monoester, but no alcohol had formed. The rate determining step,
the hydrolysis of the monoester, was taken to be similar for 1 and 2, as ring strain no
longer influenced their behavior and they were structurally alike. Given this
approximation, the actual overall rate was taken to be proportional to the concentration
101
of monoester (1a, 1b, and 2a). From analyzing relative equilibriums, the concentration
of the monoesters of 1 (1a and 1b) are approximately 20 times more than 2a, so the
five-membered ring boronic ester (1) was understood to hydrolyze at a rate
approximately 20 times faster than the six-membered ring boronic ester (2).
Figure 5-3. The overall hydrolysis constant is K1 = K’1a K”1a = K’1b K”1b = ([1’][3]) / ([1][H2O]2), and the rate constants are k’1a/k’-1a = K’1a, etc. A) Hydrolysis of the five-membered boronic ester model compound and B) hydrolysis of the six-membered boronic ester model compound.
102
Like the hydrolysis, the information from the transesterification model
experiments was useful in terms of comparison (Fig. 5-4). The bulk polymer samples
were expected to follow a similar trend to the solution models. The six-membered ring
boronic ester, as would be anticipated, transesterified slower than the five-membered
ring boronic ester, for the same reasons it hydrolyzed slower. It might be assumed that
this automatically makes the five-membered ring boronic ester the more ideal. However,
note should be taken that the stability of the six-membered ring boronic ester to
hydrolysis means that it could potentially behave dynamically and Arrheniusly through a
transesterification mechanism under wider range of water exposure conditions than are
possible for the five-membered ring boronic ester given its susceptibility to hydrolysis.
Figure 5-4. NMR transesterification with inverted and referenced protons labeled. A) Transesterification of the five-membered boronic ester model compound B) Transesterification of the six-membered boronic ester model compound.
103
5.3.2 Preparation and Characterization of Dithiol Oligomer and Networks
To raise the Tg at least 50 ˚C above similar previous boronic ester networks
made by radical photo-thiol-ene for self-healing, a dithiol oligomer was prepared using
thiol-isocyanate reactions with a slight molar excess of thiol. Thiol-ene networks with
thiol-urethane linkages have been previously shown to have glass transition
temperatures significantly higher than typical thiol-ene networks.164,166 Additionally, the
isocyanate used to make the oligomer, isophorone diisocyanate, was selected so that
the overall network would be more hydrophobic. The complete consumption of the
isocyanate group was confirmed by ATR-FTIR spectroscopy showing the
disappearance of the isocyanate peak at 2551 cm-1. ATR-FTIR spectroscopy also
confirmed the continued presence of the thiol functionality at 2590 cm-1. The oligomer
was also characterized by 1H NMR. Based on the ratio of thiol to isocyanate, the
number average degree of polymerization from the Carothers equation was calculated
to be 5. The Mn was then predicted to be 989 g/mol. The Mn as measured by GPC and
derived from polystyrene standards was 1150 g/mol, which corresponded well with the
calculated value.
The thiol-ene networks were prepared as mostly usual with DMPA as the photo-
initiator.70 The ratio of thiol to ene was kept 1:1, and the only dienes were the boronic
ester dienes, with either a five or six membered ring, so that all crosslinks were
exchangeable as required by the vitrimer definition. In contrast to prior boronic ester
thiol-ene networks, with these precursors, the networks could not be made in the
absence of solvent. The boronic esters, PTMP, and the dithiol oligomer were immiscible
in each other even though each of them was a liquid. A small amount of dry THF was
added to form a uniform solution prior to curing by UV light. Subsequently the THF was
104
removed from the networks by heating at 80 ˚C while pulling vacuum for a week. The
dried samples were glassy at room temperature. The Tg values of the boronic ester
networks were measured as the average of the tan delta peak from the second and
third heating ramps of temperature ramp by DMA. For the five and six membered ring
boronic ester networks, the values are 49 ˚C and 52 ˚C, respectively, well above room
temperature (Table 5-2).
Table 5-2. Tg values measured by DMA from the tan delta peak
Network 4-SH (rel. mol%)
Diothiol Oligomer (rel. mol %)
Free Diol (rel. mol %)
Tg (˚C)
BE(5) 25 75 3 49 BE(6) 25 75 3 52
5.3.3 Dynamics of the Networks
Stress relaxation experiments were used to measure the time scale of dynamic
behavior at a variety of temperatures. The starting temperature for the stress-relaxation
experiments was selected from the end of the tan delta peak because a non-Arrhenius
transition region exists near the Tg. The five-membered boronic ester network was
tested from 90 ˚C to 125 ˚C in 5 ˚C increments allowing 10 min for relaxation per
temperature. The stress relaxation at each temperature was fitted to a single element
Maxwell model to determine relaxation time. The model gave a good fit for this system
at 105 ˚C. The relaxation times (Table 5-3) were calculated from the stress relaxation
traces from 105 ˚C through 125 ˚C (Fig. 5-5A). The natural log of the relaxation times
plotted against inverse temperature showed a linear fit (R2 = 0.996) consistent with
Arrhenius behavior (Fig. 5-5B).
105
The characteristic relaxation times for this system are much faster than some
examples in in literature. Zn(II)-catalyzed transesterification of polyester epoxies had a
relaxation time of 1 h at 150 ˚C.124 Soybean oil-based B-hydroxyesters epoxies had a
catalyst-free relaxation time of 5.5 h at 150 ˚C.22 However, the relaxation times of the
five-membered boronic ester network are also still longer than other examples in
literature. Sn(II)-catalyzed polylactide vitrimers relaxed below 50 s at 140 ˚C.24 The
relaxation times of the network with the five-membered boronic ester are clearly well
within the range of standard examples.
Figure 5-5. Stress-relaxation of 5-BE networks. A) E/E0 vs time. B) Arrhenius fit of relaxation times.
Table 5-3. Summary of the relaxation times of 5-BE networks
T (˚C) * (min)
105 3.7 110 3.3 115 3.0 120 2.7 125 2.3
The slope of the line from the Arrhenius plot (Fig. 5-5B) gave the activation
energy, Ea, divided by the gas constant, R. The apparent activation energy for the
transesterification of the five-membered ring boronic ester was ~29 kJ/mol. Additionally,
the topology freezing temperature, Tv, was calculated using the same method as
106
Capelot and coworkers124 from the linear fit to give a value of -94 ˚C. The Tv in this case
is well below the Tg of 49 ˚C, so that when the network is heated above the Tg, the
stress-relaxation would be expected to correspond to the behavior of an ideal vitrimer.
As with the networks of the five-membered ring boronic ester, for the six-
membered ring boronic ester network, the starting temperature for the stress-relaxation
experiments was selected from the end of the tan delta peak. Although the top of the
tan delta peak, the Tg, was very similar for both boronic ester networks, the end of the
peak for the network with the six-membered ring boronic ester appeared to be at 100
˚C, 10 ˚C higher than the 90 ˚C of the previous network. Thus, the six-membered
boronic ester network was tested from 100 ˚C to 165 ˚C in 5 ˚C increments allowing 30
min for relaxation per temperature. The relaxation time allowed per temperature was
extended to 30 min due to the slower rate of exchange of the network with the six-
membered ring boronic ester. The stress-relaxation at each temperature was fitted to a
single element Maxwell model to determine relaxation time. The model gave a
reasonable fit of the stress-relaxation trace starting at 150 ˚C. This could be partially
due to lower temperature requiring more than 30 min relaxation to give a good fit. The
relaxation times (Table 5-4) were calculated from the stress relaxation traces from 150
˚C through 165 ˚C (Fig. 5-6A). The natural log of the relaxation times plotted against
inverse temperature showed a linear fit (R2 = 0.997) consistent with Arrhenius behavior
(Fig. 5-6B).
Compared to the previous five-membered ring boronic ester network, the six-
membered ring boronic ester network has a relaxation behavior approximately 30 times
slower. While this is significantly slower, the change in rate was expected from the
107
model compounds. Additionally, the relaxation times remain in the range of those
published in literature for vitrimers.
Figure 5-6. Stress relaxation of 6-BE networks. A) E/E0 vs time. B) Arrhenius fit of relaxation times.
Table 5-4. Summary of the relaxation times of 6-BE networks
T (˚C) * (min)
150 20 155 17 160 14 165 12
The activation energy for the six-membered ring boronic ester network was also
calculated from the slope of the line derived from the Arrhenius plot (Fig. 5-6B). The
apparent activation energy for the transesterification of the six-membered ring boronic
ester network was ~54 kJ/mol, almost twice the activation energy of the five-membered
ring boronic ester network. Both activation energies of the boronic ester networks are in
the range observed for carboxy ester transesterification in soybean oil174 but quite lower
than the ~90 kJ/mol seen by Liebler and coworkers for vitrimers with zinc-catalyzed
transesterification.124 The topology freezing temperature, Tv, was also calculated for the
six-membered ring boronic ester network by the same method to give a value of 5 ˚C.
As before, the Tv in this case is well below the Tg of 52 ˚C, and likewise, when the
108
network is heated above the Tg, the stress-relaxation would be expected to correspond
to the behavior of an ideal vitrimer. It should be remembered that the value of Tv is its
usefulness as a conventional reference point to compare between systems. Historically,
Tv has been used for silicates and other inorganic glasses.160,175 Practically, a viscosity
of 1012 Pa-s is very high. For the two boronic ester networks studied in this work, the
relaxation times at theTv, as derived from the linear fit, would be 65 days and 42 days
for the five and six-membered ring boronic ester networks respectively. At room
temperature, the relaxation times would be ~1 h and ~12.5 days for the five and six-
membered ring boronic ester networks respectively.
5.4 Conclusions
In this work, boronic ester transesterification was used to make new vitrimers.
These materials possessed relaxation times that changed with temperature according to
Arrhenius behavior. To ensure typical irreversible polymeric behavior at room
temperature, photo-curable thiol-ene networks were made with higher glass transition
temperatures from a synthesized more rigid oligomeric dithiol,. Additionally, model
studies demonstrated that the newly prepared six-membered ring boronic ester was
more stable to hydrolysis compared with the five-membered ring boronic ester, even
when exposed to high concentrations of liquid water, broadening the scope of possible
applications for these materials. These new boronic ester vitrimers expand the library of
vitrimer chemistries and create unique opportunities for vitrimers in areas where boron-
containing materials are of interest, such as adhesives, sensors, and flame retardants.
109
CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS
The research presented in this dissertation is meant to highlight the benefits of
incorporating reversible linkages in bulk polymers. Responsive behavior due to dynamic
bonds has broadened the original definition of the word polymer to include not only the
additive effects of covalent bonds, but also the additive effects of many reversible
covalent bonds. These effects expand the fundamental capabilities of polymeric
materials to include ability to undergo dramatic changes in properties with the
application of a specific stimulus or the presence of a specific stimulus in the
environment of the polymer. The present work demonstrates the potential for boronic
ester linkages in bulk polymers. The work on boronic ester self-healing systems
highlights the capabilities of self-healing materials with dynamic covalent linkages to
achieve repeated healing while retaining the strength covalent bonds provide to the
network. These studies also emphasizes the capability of dynamic linkages to be
combined cooperatively with permanent linkages within polymer networks to produce
materials that retain good healing efficiency while gaining resistance to creep and stress
relaxation. Moreover, the strong organic glasses, vitrimers, formed with boronic esters
show that with thoughtful modification of the boronic ester chemistry, transesterification,
an associative mechanism, can be promoted as a primary mechanism, even though
boronic esters are capable of two mechanisms of reversibility.
While the research descibed here focused exclusively on simple boronic esters in
bulk dynamic polymers, similar strategies can be expanded to prepare bulk materials
based on more complex boronic esters as well as reversible linkages with other
chemistries. The boronic ester chemistry could be optimized for use in materials
110
designed for applications such as adhesives, reversibly imprinted surfaces, and shape
memory materials.
111
APPENDIX A DSC RESULTS OF INITIAL SELF-HEALING NETWORK COMPOSITIONS
Table A-1. DSC results of low glass transition temperature polymers showing Tg values below room temperature.
Reactants Composition Tg (˚C)
1 : 2 25 : 75 -16 1 : 2 50 : 50 -3.5 1 : 2 75 : 25 -9.5 1 : 2 (DVB control) 50 : 50 -3.9
112
APPENDIX B RELAXATION TIMES OF FREE DIOL, PERMANENT CROSSLINKED, AND
COMBINED NETWORKS
Table B-1. Relaxation times from tensile stress-relaxation measurements Free Diol (%) Permanent Crosslinker (%) Humidity (%) τ (s)
5 - 0 25
3 - 0 100
1 - 0 110
0 0 0 93
- 80 0 116
- 85 0 123
- 90 0 127
- 97 0 109
- 98 0 114
- 99 0 104
5 80 0 146
5 - 23 7.3
3 - 23 41
1 - 23 92
0 0 23 107
- 80 23 62
- 85 23 78
- 90 23 83
- 97 23 85
- 98 23 105
- 99 23 82
5 80 23 69
5 - 85 1.8
3 - 85 1.3
1 - 85 0.7
0 0 85 1.9
- 80 85 2.7
- 85 85 3.2
- 90 85 3.4
- 97 85 6.1
- 98 85 6.9
- 99 85 7.4
5 80 85 5.0
113
LIST OF REFERENCES
(1) Heo, Y.; Sodano, H. A. Adv. Funct. Mater. 2014, 24, 5260-5260. (2) Lai, J.; Mei, J.; Jia, X.; Li, C.; You, X.; Bao, Z. Adv. Mater. 2016, 28, 8277-8282. (3) Cox, L. M.; Li, Z.; Sowan, N.; Nair, D.; Xiao, J.; Bowman, C. N.; Ding, Y. Polymer
2014, 55, 5933-5937. (4) Wang, D. H.; McKenzie, R. N.; Buskohl, P. R.; Vaia, R. A.; Tan, L.-S.
Macromolecules 2016, 49, 3286-3299. (5) Taynton, P.; Yu, K.; Shoemaker, R.; Jin, Y.; Qi, H. J.; Zhang, W. Adv. Mater.
2014, 26, 3938-3942. (6) Swanson, J. P.; Rozvadovsky, S.; Seppala, J. E.; Mackay, M. E.; Jensen, R. E.;
Costanzo, P. J. Macromolecules 2010, 43, 6135–6141. (7) Denissen, W.; Winne, J. M.; Du Prez, F. E. Chem. Sci. 2016, 7, 30-38. (8) Murphy, E. B.; Bolanos, E.; Schaffner-Hamann, C.; Wudl, F.; Nutt, S. R.; Auad,
M. L. Macromolecules 2008, 41, 5203-5209. (9) Oka, H.; Imato, K.; Sato, T.; Ohishi, T.; Goseki, R.; Otsuka, H. ACS Macro Lett.
2016, 5, 1124-1127. (10) Yoshie, N.; Watanabe, M.; Araki, H.; Ishida, K. Polym. Degrad. Stab. 2010, 95,
826-829. (11) Yoshie, N.; Saito, S.; Oya, N. Polymer 2011, 52, 6074-6079. (12) Chen, X.; Dam, M. A.; Ono, K.; Mal, A.; Shen, H.; Nutt, S. R.; Sheran, K.; Wudl,
F. Science 2002, 295, 1698-1702. (13) Chen, X.; Wudl, F.; Mal, A. K.; Shen, H.; Nutt, S. R. Macromolecules 2003, 36,
1802-1807. (14) Oehlenschlaeger, K. K.; Mueller, J. O.; Brandt, J.; Hilf, S.; Lederer, A.; Wilhelm,
M.; Graf, R.; Coote, M. L.; Schmidt, F. G.; Barner-Kowollik, C. Adv. Mater. 2014, 26, 3561-3566.
(15) Reutenauer, P.; Buhler, E.; Boul, P. J.; Candau, S. J.; Lehn, J. M. Chem. Eur. J.
2009, 15, 1893-1900. (16) Fuhrmann, A.; Gostl, R.; Wendt, R.; Kotteritzsch, J.; Hager, M. D.; Schubert, U.
S.; Brademann-Jock, K.; Thunemann, A. F.; Nochel, U. N.; Behl, M.; Hecht, S. Nat. Commun. 2016, 7, 1-7.
114
(17) Roy, N.; Lehn, J.-M. Chem. Asian J. 2011, 6, 2419-2425. (18) Syrett, J. A.; Mantovani, G.; Barton, W. R. S.; Price, D.; Haddleton, D. M. Polym.
Chem. 2010, 1, 102-106. (19) Raghavan, J.; Wool, R. P. J. Appl. Polym. Sci. 1999, 71, 775-785. (20) Stille, J. K.; Plummer, L. J. Org. Chem. 1961, 26, 4026–4029. (21) Capelot, M.; Montarnal, D.; Tournilhac, F.; Leibler, L. J. Am. Chem. Soc. 2012,
134, 7664-7667. (22) Altuna, F. I.; Pettarin, V.; Williams, R. J. J. Green Chem. 2013, 15, 3360-3367. (23) Demongeot, A.; Mougnier, S. J.; Okada, S.; Soulié-Ziakovic, C.; Tournilhac, F.
Polym. Chem. 2016, 7, 4486-4493. (24) Brutman, J. P.; Delgado, P. A.; Hillmyer, M. A. ACS Macro Lett. 2014, 3, 607-
610. (25) Canadell, J.; Goossens, H.; Klumperman, B. Macromolecules 2011, 44, 2536-
2541. (26) Pepels, M.; Filot, I. A. W.; Klumperman, B.; Goossens, H. Polym. Chem. 2013, 4,
4955-4965. (27) Yoon, J. A.; Kamada, J.; Koynov, K.; Mohin, J.; Nicolaÿ, R.; Zhang, Y.; Balazs, A.
C.; Kowalewski, T.; Matyjaszewski, K. Macromolecules 2012, 45, 142-149. (28) Michal, B. T.; Jaye, C. A.; Spencer, E. J.; Rowan, S. J. ACS Macro Lett. 2013, 2,
694-699. (29) Lei, Z. Q.; Xiang, H. P.; Yuan, Y. J.; Rong, M. Z.; Zhang, M. Q. Chem. Mater.
2014, 26, 2038–2046. (30) Rekondo, A.; Martin, R.; Ruiz de Luzuriaga, A.; Cabañero, G.; Grande, H. J.;
Odriozola, I. Mater. Horiz. 2014, 1, 237-240. (31) Martin, R.; Rekondo, A.; Ruiz de Luzuriaga, A.; Cabañero, G.; Grande, H. J.;
Odriozola, I. J. Mater. Chem. A 2014, 2, 5710-5716. (32) Takahashi, A.; Goseki, R.; Otsuka, H. Angew. Chem. Int. Ed. 2017, 56, 2016-
2021.
115
(33) Kleine, T. S.; Nguyen, N. A.; Anderson, L. E.; Namnabat, S.; LaVilla, E. A.; Showghi, S. A.; Dirlam, P. T.; Arrington, C. B.; Manchester, M. S.; Schwiegerling, J.; Glass, R. S.; Char, K.; Norwood, R. A.; Mackay, M. E.; Pyun, J. ACS Macro Lett. 2016, 5, 1152-1156.
(34) Griebel, J. J.; Nguyen, N. A.; Namnabat, S.; Anderson, L. E.; Glass, R. S.;
Norwood, R. A.; Mackay, M. E.; Char, K.; Pyun, J. ACS Macro Lett. 2015, 4, 862-866.
(35) Dirlam, P. T.; Simmonds, A. G.; Shallcross, R. C.; Arrington, K. J.; Chung, W. J.;
Griebel, J. J.; Hill, L. J.; Glass, R. S.; Char, K.; Pyun, J. ACS Macro Lett. 2015, 4, 111-114.
(36) Simmonds, A. G.; Griebel, J. J.; Park, J.; Kim, K. R.; Chung, W. J.; Oleshko, V.
P.; Kim, J.; Kim, E. T.; Glass, R. S.; Soles, C. L.; Sung, Y.-E.; Char, K.; Pyun, J. ACS Macro Lett. 2014, 3, 229-232.
(37) Griebel, J. J.; Namnabat, S.; Kim, E. T.; Himmelhuber, R.; Moronta, D. H.;
Chung, W. J.; Simmonds, A. G.; Kim, K.-J.; van der Laan, J.; Nguyen, N. A.; Dereniak, E. L.; Mackay, M. E.; Char, K.; Glass, R. S.; Norwood, R. A.; Pyun, J. Adv. Mater. 2014, 26, 3014-3018.
(38) Griebel, J. J.; Li, G.; Glass, R. S.; Char, K.; Pyun, J. J. Polym. Sci. A Polym.
Chem. 2014, 173-177. (39) Griebel, J. J.; Nguyen, N. A.; Astashkin, A. V.; Glass, R. S.; Mackay, M. E.; Char,
K.; Pyun, J. ACS Macro Lett. 2014, 3, 1258-1261. (40) Namnabat, S.; Norwood, R.; Dereniak, E.; Griebel, J.; Pyun, J. SPIE 2014. (41) Amamoto, Y.; Kamada, J.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. Angew.
Chem. Int. Ed. 2011, 50, 1660-1663. (42) Amamoto, Y.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. Adv. Mater. 2012, 24,
3975-3980. (43) Zhang, B.; Digby, Z. A.; Flum, J. A.; Chakma, P.; Saul, J. M.; Sparks, J. L.;
Konkolewicz, D. Macromolecules 2016, 49, 6871-6878. (44) Denissen, W.; Rivero, G.; Nicolaÿ, R.; Leibler, L.; Winne, J. M.; Du Prez, F. E.
Adv. Funct. Mater. 2015, 25, 2451-2457. (45) Obadia, M. M.; Mudraboyina, B. P.; Serghei, A.; Montarnal, D.; Drockenmuller, E.
J. Am. Chem. Soc. 2015, 137, 6078-6083.
116
(46) Kuhl, N.; Abend, M.; Bode, S.; Schubert, U. S.; Hager, M. D. J. Appl. Polym. Sci. 2016, 133, 1-9.
(47) Zhang, Y.; Barboiu, M. Chem. Commun. 2015, 51, 15925-15927. (48) Ono, T.; Fujii, S.; Nobori, T.; Lehn, J.-M. Chem. Commun. 2007, 4360-4363. (49) Roy, N.; Buhler, E.; Lehn, J.-M. Polym. Int. 2013, 63, 1400-1405. (50) Kuhl, N.; Bode, S.; Bose, R. K.; Vitz, J.; Seifert, A.; Hoeppener, S.; Garcia, S. J.;
Spange, S.; van der Zwaag, S.; Hager, M. D.; Schubert, U. S. Adv. Funct. Mater. 2015, 25, 3295-3301.
(51) Ying, H.; Zhang, Y.; Cheng, J. Nat. Commun. 2014, 5, 3218-3226. (52) Zhang, Z. P.; Lu, Y.; Rong, M. Z.; Zhang, M. Q. RSC Adv. 2016, 6, 6350-6357. (53) Zhang, Z. P.; Rong, M. Z.; Zhang, M. Q. Polymer 2014, 55, 3936-3943. (54) Yuan, C. e.; Rong, M. Z.; Zhang, M. Q. Polymer 2014, 55, 1782-1791. (55) Zhang, Z. P.; Rong, M. Z.; Zhang, M. Q.; Yuan, C. e. Polym. Chem. 2013, 4,
4648-4654. (56) Yuan, C.; Rong, M. Z.; Zhang, M. Q.; Zhang, Z. P.; Yuan, Y. C. Chem. Mater.
2011, 23, 5076-5081. (57) Yuan, C.; Zhang, M. Q.; Rong, M. Z. J. Mater. Chem. A 2014, 2, 6558-6566. (58) Aiba, M.; Higashihara, T.; Ashizawa, M.; Otsuka, H.; Matsumoto, H.
Macromolecules 2016, 49, 2153-2161. (59) Telitel, S.; Amamoto, Y.; Poly, J.; Morlet-Savary, F.; Soppera, O.; Lalevée, J.;
Matyjaszewski, K. Polym. Chem. 2013, 5, 921-930. (60) Baruah, R.; Kumar, A.; Ujjwal, R. R.; Kedia, S.; Ranjan, A.; Ojha, U.
Macromolecules 2016, 49, 7814-7824. (61) Serafimova, I. M.; Pufall, M. A.; Krishnan, S.; Duda, K.; Cohen, M. S.; Maglathlin,
R. L.; McFarland, J. M.; Miller, R. M.; Frodin, M.; Taunton, J. Nat. Chem. Biol. 2012, 8, 471-476.
(62) Krishnan, S.; Miller, R. M.; Tian, B.; Mullins, R. D.; Jacobson, M. P.; Taunton, J.
J. Am. Chem. Soc. 2014, 136, 12624-12630.
117
(63) Lu, Y.-X.; Tournilhac, F.; Leibler, L.; Guan, Z. J. Am. Chem. Soc. 2012, 134, 8424-8427.
(64) Lu, Y. X.; Guan, Z. J. Am. Chem. Soc. 2012, 134, 14226-14231. (65) Yokoe, M.; Yamauchi, K.; Long, T. E. J. Polym. Sci., Part A: Polym. Chem. 2016,
54, 2302-2311. (66) Chung, C.; Roh, Y.; Cho, S.; Kim, J. Chem. Mater. 2004, 16, 3982-3984. (67) Imato, K.; Kanehara, T.; Nojima, S.; Ohishi, T.; Higaki, Y.; Takahara, A.; Otsuka,
H. Chem. Commun. 2016, 52, 10482-10485. (68) Osthoff, R. C.; Bueche, A. M.; Grubb, W. T. J. Am. Chem. Soc. 1954, 76, 4659-
4663. (69) Zheng, P.; McCarthy, T. J. J. Am. Chem. Soc. 2012, 134, 2024-2027. (70) Cash, J. J.; Kubo, T.; Bapat, A. P.; Sumerlin, B. S. Macromolecules 2015, 48,
2098-2106. (71) Zuo, Y.; Gou, Z.; Zhang, C.; Feng, S. Macromol. Rapid Commun. 2016, 1-8. (72) Cromwell, O. R.; Chung, J.; Guan, Z. J. Am. Chem. Soc. 2015, 137, 6492-6495. (73) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart, J. F.
Angew. Chem. Int. Ed. 2002, 41, 898-952. (74) Jin, Y.; Yu, C.; Denman, R. J.; Zhang, W. Chem. Soc. Rev. 2013, 42, 6634-6654. (75) Wojtecki, R. J.; Meador, M. A.; Rowan, S. J. Nat. Mater. 2010, 10, 14-27. (76) Burattini, S.; Greenland, B. W.; Chappell, D.; Colquhoun, H. M.; Hayes, W.
Chem. Soc. Rev. 2010, 39, 1973-1985. (77) Wu, D. Y.; Meure, S.; Solomon, D. Prog. Polym. Sci. 2008, 33, 479-522. (78) Wool, R. P. Soft Matter 2008, 4, 400-418. (79) Fiore, G. L.; Rowan, S. J.; Weder, C. Chem. Soc. Rev. 2013, 42, 7278-7288. (80) Murphy, E. B.; Wudl, F. Prog. Polym. Sci. 2010, 35, 223-251. (81) Billiet, S.; Hillewaere, X. K. D.; Teixeira, R. F. A.; Du Prez, F. E. Macromol. Rapid
Commun. 2012, 34, 290-309.
118
(82) Flory, P. J. In Principles of Polymer Chemistry; Cornell University Press: New York, 1953, p 672.
(83) Mülhaupt, R. Angew. Chem. Int. Ed. 2004, 43, 1054-1063. (84) Watanabe, M.; Yoshie, N. Polymer 2006, 47, 4946-4952. (85) Bergman, S. D.; Wudl, F. J. Mater. Chem. 2008, 18, 41-62. (86) Yoon, J. A.; Kamada, J.; Koynov, K.; Mohin, J.; Nicolaÿ, R.; Zhang, Y.; Balazs, A.
C.; Kowalewski, T.; Matyjaszewski, K. Macromolecules 2012, 45, 142-149. (87) Ghosh, B.; Chellappan, K. V.; Urban, M. W. J. Mater. Chem. 2012, 22, 16104-
16113. (88) Heo, Y.; Sodano, H. A. Adv. Funct. Mater. 2014, 24, 5261-5268. (89) Yang, Y.; Urban, M. W. Chem. Soc. Rev. 2013, 42, 7446-7467. (90) Deng, C. C.; Brooks, W. L. A.; Abboud, K. A.; Sumerlin, B. S. ACS Macro Lett.
2015, 220-224. (91) Wei, Z.; Yang, J. H.; Zhou, J.; Xu, F.; Zrinyi, M.; Dussault, P. H.; Osada, Y.;
Chen, Y. M. Chem. Soc. Rev. 2014, 43, 8114-8131. (92) Phadke, A.; Zhang, C.; Arman, B.; Hsu, C.-C.; Mashelkar, R. A.; Lele, A. K.;
Tauber, M. J.; Arya, G.; Varghese, S. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 4383-4388.
(93) Fairbanks, B. D.; Singh, S. P.; Bowman, C. N.; Anseth, K. S. Macromolecules
2011, 44, 2444-2450. (94) Urban, M. W. Nat. Chem. 2012, 4, 80-82. (95) Yuan, C.; Rong, M. Z.; Zhang, M. Q.; Zhang, Z. P.; Yuan, Y. C. Chem. Mater.
2011, 23, 5076-5081. (96) Amamoto, Y.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. Adv. Mater. 2012, 24,
3975–3980. (97) Ghosh, B.; Urban, M. W. Science 2009, 323, 1458-1460. (98) Ghosh, B.; Chellappan, K. V.; Urban, M. W. J. Mater. Chem. 2011, 21, 14473-
14486. (99) Niu, W.; Smith, M. D.; Lavigne, J. J. J. Am. Chem. Soc. 2006, 128, 16466-16467.
119
(100) Niu, W.; O'Sullivan, C.; Rambo, B. M.; Smith, M. D.; Lavigne, J. J. Chem.
Commun. 2005, 4342-4344. (101) Cheng, F.; Jäkle, F. Polym. Chem. 2011, 2, 2122-2132. (102) De, P.; Gondi, S. R.; Roy, D.; Sumerlin, B. S. Macromolecules 2009, 42, 5614-
5621. (103) Qin, Y.; Cui, C.; Jäkle, F. Macromolecules 2007, 40, 1413-1420. (104) Korich, A. L.; Walker, A. R.; Hincke, C.; Stevens, C.; Iovine, P. M. J. Polym. Sci.,
Part A: Polym. Chem. 2010, 48, 5767-5774. (105) Bapat, A. P.; Roy, D.; Ray, J. G.; Savin, D. A.; Sumerlin, B. S. J. Am. Chem. Soc.
2011, 133, 19832-19838. (106) Xu, J.; Yang, D. G.; Li, W. J.; Gao, Y.; Chen, H. B.; Li, H. M. Polymer 2011, 52,
4268-4276. (107) Piest, M.; Zhang, X.; Trinidad, J.; Engbersen, J. F. J. Soft Matter 2011, 7, 11111-
11118. (108) Tarus, D.; Hachet, E.; Messager, L.; Catargi, B.; Ravaine, V.; Auzély-Velty, R.
Macromol. Rapid Commun. 2014, 35, 2089-2095. (109) He, L.; Fullenkamp, D. E.; Rivera, J. G.; Messersmith, P. B. Chem. Commun.
2011, 47, 7497-7499. (110) Roberts, M. C.; Hanson, M. C.; Massey, A. P.; Karren, E. A.; Kiser, P. F. Adv.
Mater. 2007, 19, 2503-2507. (111) Kotsuchibashi, Y.; Agustin, R. V. C.; Lu, J.-Y.; Hall, D. G.; Narain, R. ACS Macro
Lett. 2013, 2, 260-264. (112) Greenspan, L. J. Res. NBS A Phys. Chem. 1977, 81A, 89. (113) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem. Int. Ed. 2001, 40, 2004-
2021. (114) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem. 2001, 113, 2056-2075. (115) Sumerlin, B. S.; Vogt, A. P. Macromolecules 2010, 43, 1-13. (116) Hoyle, C. E.; Bowman, C. N. Angew. Chem. Int. Ed. 2010, 49, 1540-1573.
120
(117) Stukalin, E. B.; Cai, L.-H.; Kumar, N. A.; Leibler, L.; Rubinstein, M. Macromolecules 2013, 46, 7525-7541.
(118) Scott, T. F.; Schneider, A. D.; Cook, W. D.; Bowman, C. N. Science 2005, 308,
1615-1617. (119) Roy, N.; Bruchmann, B.; Lehn, J.-M. Chem. Soc. Rev. 2015, 44, 3786-3807. (120) Amamoto, Y.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. Adv. Mater. 2012,
3975-3980. (121) Otsuka, H. Polym. J. 2013, 45, 879-891. (122) Garcia, S. J. Eur. Polym. J. 2014, 53, 118-125. (123) Yang, H.; Yu, K.; Mu, X.; Shi, X.; Wei, Y.; Guo, Y.; Qi, H. J. Soft Matter 2015, 11,
6305-6317. (124) Capelot, M.; Unterlass, M. M.; Tournilhac, F.; Leibler, L. ACS Macro Lett. 2012,
1, 789-792. (125) Montarnal, D.; Capelot, M.; Tournilhac, F.; Leibler, L. Science 2011, 334, 965-
968. (126) Fortman, D. J.; Brutman, J. P.; Cramer, C. J.; Hillmyer, M. A.; Dichtel, W. R. J.
Am. Chem. Soc. 2015, 137, 14019-14022. (127) Romano, F.; Sciortino, F. Phys. Rev. Lett. 2015, 114, 078104-078105. (128) Imbernon, L.; Oikonomou, E. K.; Norvez, S.; Leibler, L. Polym. Chem. 2015, 6,
4271-4278. (129) Hall, D. G. In Boronic acids: preparation and applications in organic synthesis
and medicine; Wiley-VCH: Weinheim, 2005, p ACS Applied Materials & Interfaces.
(130) Lennox, A. J. J.; Lloyd-Jones, G. C. Chem. Soc. Rev. 2014, 43, 412-443. (131) Zhang, B.; Digby, Z. A.; Flum, J. A.; Foster, E. M.; Sparks, J. L.; Konkolewicz, D.
Polym. Chem. 2015, 6, 7368-7372. (132) Fox, C. H.; ter Hurrne, G. M.; Wojtecki, R. J.; Jones, G. O.; Horn, H. W.; Meijer,
E. W.; Frank, C. W.; Hedrick, J. L.; a, J. M. G. i. Nat. Commun. 2015, 6, 1-8. (133) Iyer, B. V. S.; Salib, I. G.; Yashin, V. V.; Kowalewski, T.; Matyjaszewski, K.;
Balazs, A. C. Soft Matter 2013, 9, 109-121.
121
(134) Sordo, F.; Mougnier, S.-J.; Loureiro, N.; Tournilhac, F.; Michaud, V.
Macromolecules 2015, 48, 4394-4402. (135) Döhler, D.; Peterlik, H.; Binder, W. H. Polymer 2015, 69, 264-273. (136) Nicolaÿ, R.; Kamada, J.; Van Wassen, A.; Matyjaszewski, K. Macromolecules
2010, 43, 4355-4361. (137) Mukherjee, S.; Bapat, A. P.; Hill, M. R.; Sumerlin, B. S. Polym. Chem. 2014, 5,
6923-6931. (138) Bapat, A. P.; Ray, J. G.; Savin, D. A.; Sumerlin, B. S. Macromolecules 2013, 46,
2188-2198. (139) Bapat, A. P.; Ray, J. G.; Savin, D. A.; Hoff, E. A.; Patton, D. L.; Sumerlin, B. S.
Polym. Chem. 2012, 3, 3112-3120. (140) Figg, C. A.; Simula, A.; Gebre, K. A.; Tucker, B. S.; Haddleton, D. M.; Sumerlin,
B. S. 2015, 6, 1230-1236. (141) Roy, D.; Sumerlin, B. S. Macromol. Rapid Commun. 2014, 35, 174-179. (142) Cambre, J. N.; Roy, D.; Sumerlin, B. S. J. Polym. Sci., Part A: Polym. Chem.
2012, 50, 3373-3382. (143) Roy, D.; Cambre, J. N.; Sumerlin, B. S. Chem. Commun. 2009, 2106-2108. (144) Mukherjee, S.; Brooks, W. L. A.; Dai, Y.; Sumerlin, B. S. Polym. Chem. 2016, 7,
1971-1978. (145) Salib, I. G.; Kolmakov, G. V.; Gnegy, C. N.; Matyjaszewski, K.; Balazs, A. C.
Langmuir 2011, 27, 3991-4003. (146) Leibler, L.; Rubinstein, M.; Colby, R. H. J. Phys. II 1993, 3, 1581−1590. (147) Osthoff, R. C.; Bueche, A. M.; Grubb, W. T. J. Am. Chem. Soc. 1954, 76,
4659−4663. (148) Zheng, P.; McCarthy, T. J. J. Am. Chem. Soc. 2012, 134, 2024−2027. (149) Scott, T. F.; Schneider, A. D.; Cook, W. D.; Bowman, C. N. Science 2005, 308,
1615−1617. (150) Kloxin, C. J.; Scott, T. F.; Park, H. Y.; Bowman, C. N. Adv. Mater. 2011, 23,
1977−1981.
122
(151) Nicolay, R.; Kamada, J.; Wassen, A. V.; Matyjaszewski, K. Macromolecules
2010, 43, 4355−4361. (152) Amamoto, Y.; Kamada, J.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. Angew.
Chem., Int. Ed. Engl. 2011, 50, 1660−1663. (153) Amamoto, Y.; Otsuka, H.; Takahara, A.; Matyjaszewski, K. ACS Macro Lett.
2012, 1, 478−481. (154) Lu, Y.-X.; Tournilhac, F.; Leibler, L.; Guan, Z. J. Am. Chem. Soc. 2012, 134,
8424−8427. (155) Deng, G.; Tang, C.; Li, F.; Jiang, H.; Chen, Y. Macromolecules 2010, 43,
1191−1194. (156) Deng, G.; Li, F.; Yu, H.; Liu, F.; Liu, C.; Sun, W.; Jiang, H.; Chen, Y. ACS Macro
Lett. 2012, 1, 275−279. (157) Ediger, M. D.; Angell, C. A.; Nagel, S. R. J. Phys. Chem. 1996, 100,
13200−13212. (158) Williams, M. L.; Landel, R. F.; Ferry, J. D. J. Am. Chem. Soc. 1955, 77,
3701−3707. (159) Dyre, J. C. Rev. Mod. Phys. 2006, 78, 953−971. (160) Angell, C. A. Science 1995, 267, 1924−1935. (161) Yang, Y.; Pei, Z.; Zhang, X.; Tao, L.; Wei, Y.; Ji, Y. Chem. Sci. 2014, 5, 3486-
3492. (162) Pei, Z.; Yang, Y.; Chen, Q.; Terentjev, E. M.; Wei, Y.; Ji, Y. Nat. Mater. 2014, 13,
36-41. (163) Lyon, G. B.; Cox, L. M.; Goodrich, J. T.; Baranek, A. D.; Ding, Y.; Bowman, C. N.
Macromolecules 2016, 8905–8913. (164) Li, Q.; Zhou, H.; Wicks, D. A.; Hoyle, C. E. J. Polym. Sci., Part A: Polym. Chem.
2007, 45, 5103-5111. (165) Eglin, D.; Griffon, S.; Alini, M. J. Biomater. Sci. Polym. Ed. 2010, 21, 477-491. (166) McNair, O. D.; Brent, D. P.; Sparks, B. J.; Patton, D. L.; Savin, D. A. ACS Appl.
Mater. Interfaces 2014, 6, 6088-6097.
123
(167) Nguyen, L.-T. T.; Truong, T. T.; Nguyen, H. T.; Le, L.; Nguyen, V. Q.; Le, T. V.; Luu, A. T. Polym. Chem. 2015, 6, 3143-3154.
(168) Tilford, R. W.; Mugavero III, S. J.; Pellechia, P. J.; Lavigne, J. J. Adv. Mater.
2008, 20, 2741-2746. (169) Cao, J.; Liu, S.; Chen, Y.; Shi, L.; Zhang, Z. Polym. Chem. 2014, 5, 5029-5036. (170) Xu, J.; Duran, D.; Mao, B. J. Liq. Chromatogr. Relat. Technol. 2006, 29, 2795-
2806. (171) Roy, C. D.; Brown, H. C. J. Organomet. Chem. 2007, 692, 784-790. (172) Achilli, C.; Ciana, A.; Fagnoni, M.; Balduini, C.; Minetti, G. Cent. Eur. J. Chem.
2013, 11, 137-139. (173) Bain, A. D.; Cramer, J. A. J. Magn. Reson. A 1996, 118, 21-27. (174) Freedman, B.; Butterfield, R. O.; Pryde, E. H. J. Am. Oil Chem. Soc. 1986, 63,
1375− 1380. (175) Dyre, J. C. Rev. Mod. Phys. 2006, 78, 953−972.
124
BIOGRAPHICAL SKETCH
Jessica Cash grew up in the Mojave Desert by the Eastern Sierra Nevada
mountains in California and received her B.S. in Chemistry from the University of
California, Santa Barbara in 2011. She worked in Dr. Craig Hawker’s group synthesizing
and functionalizing magnetic nanoparticles with RAFT polymers and incorporating them
into thiol-ene matrices and performing materials characterization. She also worked as a
student intern at China Lake Naval Reseach Lab studying film formation and properties
for optical waveguide devices, high-temperature composite resins, and ionic liquids.
While at the University, she was named a UCSB Regents Scholar and a member of
USA Today’s 2nd All-California College Academic Team. She enrolled into the
University of Florida in fall 2012 as a graduate student in organic chemistry under the
direction of Dr. Brent Sumerlin, where she studied polymer chemistry. In summer 2014,
she was awarded the Science, Mathematics And Research for Transformation
Scholarship sponsored by the Department of Defense. She received her PhD in
chemistry in the spring 2017.