My PhD dissertation.
1*#align(center)[#smallcaps[@thermosetting-materials-and-their-limitations[Chapter]]]*
2= #smallcaps[Thermosetting materials and their limitations]
3<thermosetting-materials-and-their-limitations>
4
5== Introduction
6<introduction>
7The crosslinking of polymers to create thermosets was a crucial
8innovation in human history. The first polymer scientists, the ancient
9Olmec, were known to extract latex from _Castilla elastica_ trees and mix
10it with juice from the _Ipomoea alba_ vine to convert the latex into
11rubber as early as 1600 B.C.E. In fact, the name Olmec means \"rubber
12people\" in Nahuatl. Olmec derives from the Nahuatl Ōlmēcatl (singular)
13or Ōlmēcah (plural), which in turn is derived from ōlli, meaning
14\"natural rubber\", and mēcatl, meaning \"people.\"@coe2017
15Archaeologists applied this name to the ancient culture before it was
16understood that the rubber people the Nahuas referred to were their
17contemporary neighbors in the Gulf Lowlands since the time of the Aztecs
18(more properly the Mēxihcah), not the ancient Olmecs of 2000 years
19prior. The name stuck, and it wonderfully illustrates the ingenuity of the
20Olmecs.@diehl2004
21
22#figure(image("Images/RubberBall.png", width: 80.0%),
23 placement: auto,
24 caption:
25 [Mesoamerican people were the first polymer scientists. An image of a
26 rubber ball preserved in soil, from El Manatí, 1600 BC.@perez2024]
27)
28<fig:rubberball>
29
30The Olmec used this rubber to fashion elastic balls as heavy as 9 lbs
31for use in the mesoamerican ballgame. By varying the ratio of latex to
32juice, rubbers of varying elasticity could be produced suited to
33different needs. In addition to making game balls (@fig:rubberball), they would also make rubber soled sandals, watertight
34containers, and waterproof fabrics by impregnating fabric with the
35latex/juice mixture.@tarkanian2011@hosler1999@tarkanian2003@ricarte2024
36This technology would spread to later mesoamericans, and eventually be
37noted by colonists.
38
39While this rubber was initially viewed as a curiosity by Europeans, some
40three and a half millennia later, this class of materials and its
41characteristic liquid to solid transition would come to underpin many
42developments in polymer science. Charles de la Condamine sent a sample
43of rubber to the Académie Royale des Sciences from Ecuador in 1736. La
44Condamine described rubber as originating from the milk, or as he called
45it \"latex,\" a term still in use today, of Hévé trees. Joseph Priestley
46coined the term \"indiarubber\" in 1770 after coming across a sample in
47an artist supply shop being sold to rub off pencil markings, eventually
48being shortened to just rubber.@wake1983 In 1844 Charles Goodyear
49rediscovered and received a patent for the process of vulcanization of
50natural rubber.@fisher1939@guise-richardson2010 In 1907 Leo Baekeland
51introduced the first commercial synthetic plastic, phenol-formaldehyde
52resin Bakelite, \"the material with a thousand uses.\"@baekeland1909 The
53molecular underpinnings of these materials was put forward by the
54\"father of polymer chemistry\" Hermann Staudinger in his groundbreaking
55paper \"On polymerization\" in 1920.@frey2020 Staundinger himself
56unintentionally supported his macromolecular hypothesis with
57cyclopentadiene crosslinked by Diels-Alder reactions, which would not be
58understood until 1928.@bruson1926@veldman2009@diels1928 The concept of
59gelation, the point at which polymers are crosslinked into a single
60molecule that spans a given volume element, was then first
61quantitatively described by Paul Flory in 1941,@flory1941 as well as a
62statistical mechanical treatment of crosslinked polymers in
631943.@flory1943@flory1943a
64
65The years since have seen momentous developments in the field of polymer
66chemistry enabling the synthesis of polymers, and by extension
67thermosets, by radical polymerizations with predictable molecular
68weights, low dispersity, and diverse functionality. Prior, free-radical
69polymerizations would produce \"dead\" polymers, those that cannot
70participate in further monomer addition, with heterogeneous degrees of
71polymerization. In 1956, Szwarc introduced the concept of a \"living\"
72polymerization,@szwarc1956 in which the propagating carbanion remains
73active even after all monomer is consumed. However, carbanions are
74easily destroyed by impurities and cannot be regenerated after their
75destruction. In 1993, Georges _et al._ demonstrated a \"living\"
76free-radical polymerization with a narrow dispersity using
77nitroxide-mediated polymerization (NMP).@georges1993 In 1995, Wang &
78Matyjaszewski and Kato _et al._ independently introduced atom transfer
79radical polymerization (ATRP).@kato1995@wang1995 In 1997, the third of
80the three major reversible deactivation radical polymerizations (RDRP),
81reversible addition-fragmentation chain transfer (RAFT) polymerization,
82was introduced by Chiefari _et al._@chiefari1998 All three methods produce
83polymers with stable, dormant end groups that can subsequently be
84reinitiated. Suddenly chemists and engineers had several techniques at
85their disposal for easily preparing well defined polymers with low
86dispersity, a variety of functional groups, and with controlled
87architectures, enabling the production of ever more sophisticated
88thermosets.
89
90Polymer thermosets have since come a long way, rising to dominate much
91of our world. By adjusting the crosslink density, as the ancient Olmecs
92discovered, and chemical structure the properties of the final material
93can be tuned to suit a whole host of applications. Additionally, the
94crosslinked nature of thermosetting polymers impart solvent resistance,
95thermo-mechanical resistance, and chemical, wear, and creep
96resistance.@alabiso2020 The combination of these properties, as well as
97their light weight, have made them irreplaceable in high performance
98environments such as aerospace@ma2023 and renewable energy.@wang2024
99With applications further ranging the gamut from soft materials like
100touch sensors,@reynolds2020 tissue mimicking hydrogels for cell
101culture,@jansen2022 re-processable pressure sensitive
102adhesives,@arrington2018 to high-modulus materials like printed circuit
103boards,@lee2016 protective coatings,@kathalewar2014 structural composite
104materials,@li2022 and many more too numerous to list, crosslinked
105materials have become completely irreplaceable.
106
107== Mechanoresponsive Thermosetting Materials
108<mechanoresponsive-thermosetting-materials>
109=== Introduction
110<introduction-1>
111Until recently, virtually all thermosetting materials have been
112crosslinked, a.k.a. cured, by either temperature, radiation, or simply
113spontaneous reactions upon mixing. These techniques generally induce the
114formation of covalent bonds between polymer chains or the polymerization
115of small multifunctional molecules to form a three-dimensional network
116structure. Each approach to crosslinking has distinct advantages and
117disadvantages. Thermal curing involves using high temperatures to drive
118forward crosslinking reactions that either occur very slowly at room
119temperature or not at all.@engels2004 Thermal curing is well suited for
120large samples, but is very energy intensive and requires an
121understanding of the thermal properties of the substrate to drive full
122conversion.@abliz2013 Spontaneously thermosetting materials are cured by
123mixing, often by step polymerization, however this relinquishes
124spatiotemporal control over gelation and can make applying the mixture
125difficult. Radiation-based curing, often achieved through free radical
126polymerization (FRP) with photosensitive radical initiators, can be very
127fast and energy efficient, but the penetration depth of light or
128electrons is limiting and the byproducts of radical initiators can be
129toxic. Reaching full cure is often not possible, and further thermal
130curing is required in addition to radiation exposure.@raghavan2000
131Additionally, radiation curing cannot be conducted through opaque
132materials, limiting their application space. As such, there is a need to
133develop new methods of curing thermosets that can be conducted through
134opaque materials, with lower energy cost, and with spatiotemporal
135control.
136
137Nature has had eons to develop and refine exquisite molecular machinery
138that puts even the very best of modern chemistry and engineering to
139shame. Looking across the natural world, one will find it replete with
140stimuli responsiveness. Retinal based photosynthesis captures and stores
141the energy of light through cis-trans isomerization coupled to ion
142pumps,@ernst2014 and chlorophyll based photosynthesis stores the energy
143of light through electron transfer.@vishniac1958 It's worth noting that
144chlorophylls have been harnessed by polymer chemists to induce
145polymerization with light.@shanmugam2015 Methods of sensing oxygen
146content, temperature, and pH just to name a few more, are crucial for
147survival.@kumar2007@damaghi2013@tan2016 More rare in synthetic
148materials, but common in nature, is constructive responsive to
149mechanical forces.@eyckmans2011
150
151Natural materials can respond to mechanical force constructively by
152polymerizing under deformation, as is the case for
153fibronectin.@gee2008@gee2013@klotzsch2009@mao2005 Under stain, the
154tertiary structure of fibronectin partially unravels, exposing cryptic
155binding sites that then participate in mechanically-induced
156polymerization and fibril formation. Mechanical force can also change
157gene expression. Extracellular mechanical forces can be propagated from
158focal adhesions through the cytoskeleton and LINC complex directly to
159chromatin, causing chromatin stretching and mechanosensitive gene
160expression.@dupont2022 Taking mechanically-induced unfolding of proteins
161and protein complexes as inspiration, our lab and others have developed
162gels that stiffen in response to applied cyclic compression by forming
163interchain disulfide bonds and thioethers.@lee2016a@tran2017@sonu2023
164However, these gels are not currently well suited for adhesives or
165coatings due to the difficulties of applying solid materials to
166substrates. The development of liquid pre-cursor solutions that can
167crosslink via mechanical perturbation for easy solution spread across a
168chosen substrate would transform a range of industries where reliance on
169thermal, UV, and electron beam curing makes on-demand curing of
170adhesives and post-crosslinking strengthening of coatings difficult.
171
172=== Sonication as a tool for applying targeted force
173<sonication-as-a-tool-for-applying-targeted-force>
174Imparting force onto polymer chains is commonly achieved when the
175polymer is in a fluid state, either as a melt, a glassy liquid, or a
176solution. When studying polymers, shear is most often applied, and shear
177is applied through the oscillations of parallel plates, extrusion, or
178ultrasonic waves. In solution, linear polymers exist in a solvent
179dependent coiled state (a Gaussian coil) in which the end-to-end
180distance is much smaller than the contour length of the polymer (the
181length of the polymer at maximum extension).@rubinstein2003 When polymer
182solutions flow rapidly near a surface (as in the case of parallel plate
183rheology or extrusion through a narrow opening)@graham2011@boger1987 or
184are sonicated,@akbulatov2017 the shear forces cause this coiled
185structure to be disrupted. Commonly, and perhaps intuitively, it is
186thought that tension along the polymer backbone reaches a maximum at the
187chain center, at which point enough force is accumulated along the
188backbone such that mechanochemistry can occur. Models have been proposed
189based on this, one suggesting polymers fully stretch, with the
190end-to-end distance reaching the contour length (overstretched chain),
191or there is only partial unfolding of the polymer coil based around the
192chain center (overstretched segment,
193@fig:dynamicchain).@diesendruck2023 More recent work has suggested
194instead that polymer unfolding in response to shear force may instead be
195a dynamic process in which chain unfolding begins at the polymer ends,
196with the overstretched segment propagating and growing along the
197backbone until it is sufficiently strained to react
198mechanochemically.@oneill2023
199
200#figure(image("Images/DynamicChain.png", width: 80.0%),
201 placement: auto,
202 caption: [
203 Dynamic overstretched chain model. Overstretched segments of polymer
204 (red) propagate and grow along the backbone in response to
205 cavitations caused by ultrasound.
206 ]
207)
208<fig:dynamicchain>
209
210In the case of polymers flowing near a surface, the shear forces are
211driven by physically pushing a solution through an opening or moving a
212surface rapidly parallel to the liquid. In the case of sonication, force
213is generated through the application of ultrasound, which has been known
214to degrade polymer chains since 1939.@schmid1939 Ultrasound consists of
215high frequency sound waves, with the 20 kHz to 2 MHz regime being
216\"destructive\" ultrasound necessary for for sonochemical and
217mechanochemical transformations. Sound waves compress and expand liquids
218as they are ultrasonicated, in places causing distances that are large
219enough to \"break\" the liquid and form cavitation bubbles. These
220bubbles expand over several cycles of compression and rarefaction until
221they grow large enough to become unstable, at which point it collapses.
222The collapse generates strong pressure gradients, causing a shear force
223as the surrounding area is pulled towards the center of the implosion,
224affecting an area on a micrometer scale.@paulusse2006 These pressure
225gradients and resultant shear forces are commonly cited as the source of
226mechanochemical transformations of polymers in response to sonication.
227Sonication as a means to drive chemical changes offers advantages over
228radiation and thermal curing. For one, sonication can be focused through
229opaque materials.@kuang2023 Additionally, sonication consumes up to
23025,000 times less energy than thermal curing processes.@biswal2023
231However, sonication also generates heat, and the strong shear forces
232generated can cause chain scission events.
233
234=== Mechanophores
235<mechanophores>
236The field of polymer mechanochemistry has its origins in papers
237published by none other than Staudinger himself in the 1930s describing
238the reduction in molecular weight of polystyrene samples after ball
239milling.@staudinger1930a@staudinger1930@staudinger1934 The idea that
240this decrease in molecular weight was a direct result of shear forces
241causing homolytic C--C bond cleavage was put forward by Kauzmann and
242Eyring in 1940,@kauzmann1940 and then confirmed by Tabata _et al._ by
243studying ultrasonicated polymers with electron paramagnetic resonance in
2441980.@tabata1980 That same year, Encina _et al._ demonstrated that
245polymers containing randomly distributed peroxide linkages
246preferentially undergo scission at O--O bonds over C--C
247bonds.@encina1980 In 2005, Berkowski _et al._ took the field a step
248further from random cleavages to targeted cleavages by demonstrating the
249preferential cleavage of chain-centered azo linkages.@berkowski2005
250Around this time at an Army Research Office workshop on polymer
251mechanochemistry, Caster suggested the name \"mechanophore\" for
252functional groups that undergo chemical transformations in response to
253mechanical force.@li2015
254
255The field grew after Hickenboth _et al._ published on biasing reaction
256pathways with mechanical force,@hickenboth2007 and while Woodward and
257Hoffman may have been certain there would never and could never be
258exceptions to their pericyclic selection rules,@woodward1969 exceptions
259would indeed arise. In this case, Hickenboth et al. conclusively showed
260that both cis and trans isomers of polymer chain centered
2611,2-diacetoxybenzocyclobutene undergo a disrotary and conrotary,
262respectively, electrocyclic ring opening to form the E,E isomer of
263ortho-quinodimethide in response to mechanical force, in direct
264violation of the Woodward-Hoffman rules (@fig:HickenbothFig).
265This of course caused an explosion of interest in polymer
266mechanochemistry, and mechanophores have since been developed for
267several applications including strain sensing,@raisch2018
268catalysis,@groote2013 and of course,
269crosslinking.@wang2015@groote2014@ramirez2013 Whereas previously
270mechanochemical reactions resulted in the breaking and weakening of
271polymer chains, now there are ways to link chains together to form
272crosslinked solids through mechanical force.@willis-fox2018@ghanem2021
273
274#figure(image("Images/HickenbothFig.png", width: 100.0%),
275 placement: auto,
276 caption: [
277 Cartoon depiction of the predicted reaction pathways of
278 1,2-diacetoxybenzocyclobutene. Reproduced from Hickenboth et al.
279 @hickenboth2007
280 ]
281)
282<fig:HickenbothFig>
283
284=== Limitations
285<limitations>
286A popular route to achieve mechanochemical transformation of polymeric
287materials is to employ mechanophores. These motifs are specially
288tailored weak bonds that can be incorporated into polymer chains that
289transform into predictable products upon application of
290force.@brantley2013@kean2013@kean2015@jung2021 However,
291mechanophore-mediated strengthening of polymeric materials remains a
292challenging approach for mechanical force curing of polymers for several
293reasons. First, these moieties typically are restrained by limited
294incorporation in polymer chains resulting in low crosslinking densities
295upon activation. Further, conversion of mechanophores is often low,
296resulting in poor activation even with good incorporation.@lloyd2023 The
297chemistry necessary to implement mechanophores is complex in its design
298and execution, making the approach inaccessible. These moieties are not
299commercially available. Typical means of mechanochemical activation tend
300to introduce significant and destructive bond scission, which one must
301be careful to avoid by selecting appropriate force
302regimes.@grandbois1999@garnier2000 Chapter 2 describes my approach using
303off the shelf chemistries to produce materials that are constructively
304sensitive to mechanical force with high conversions.
305
306== Polymer architecture and bulk properties
307<polymer-architecture-and-bulk-properties>
308=== Introduction
309<introduction-2>
310Likely the first report of controlled polymer architecture and the first report of comb polymers arrived in the 1944 when Rehberg and Fisher prepared high $n$-alkyl polyacrylates.@rehberg1944 This was followed by Kaufman _et al_. 1948 who observed that higher $n$-alkyl polyacrylates show sidechain crystallization.@kaufman1948 These comb-like polymers were defined as polymers where each monomer bears its own polymer sidechain@plate1974, though in modern polymer science they can be defined as polymers were only some monomers bear their own polymer sidechains.
311
312Another early example of controlled polymer architecture, block polymers,
313arrived in 1951 when Vaughn et al. published a study of new nonionic
314surfactants, including block copolymers of poly(ethylene oxide) (a.k.a.
315poly(ethylene glycol) (PEG)) or poly(propylene
316oxide).@vaughn1951@mankowich1954 Known commercially as poloxamers, this
317particular class of polymers remains highly studied and useful examples
318of molecular design to this day.@surve2024 Living polymerizations would
319be described just a few years later in 1956,@szwarc1956 which would
320subsequently enable the facile preparation of whole new classes of block
321polymers as well as far more exotic examples of molecular architecture.
322
323Living polymerizations would enable far more exotic examples of polymer
324architecture compared to linear polymers.@hadjichristidis2006 In the
325absence of impurities, the carbanion active site remains intact even at
326full monomer conversion, meaning a wealth of facile chain end
327modifications were now within reach. In 1984, Rempp et al. published the
328first synthesis of macromonomers by modification of the
329carbanion.@rempp1984@rempp1985 This would soon be followed up by the
330first publications on what are now called bottlebrush polymers by
331Tsukahara et al. in 1989@tsukahara1989 through a combination of living
332anionic polymerization and FRP, and then a more detailed investigation
333into their physical properties in 1994.@tsukahara1994 At this point
334however, the preparation of bottlebrush polymers was still quite
335difficult. The free radical homopolymerization of $omega$-methacryloyl
336functionalized polymers has several limitations, chiefly among them the
337dilute nature of the methacrylate functional group.@muehlebach2003 It
338would take further progress in polymer synthesis to unlock new routes to
339bottlebrush polymers, with ring opening metathesis polymerization and
340RDRPs in particular enabling the routine synthesis of cylindrical
341molecular brushes with a huge array of monomers.
342
343=== Comb polymers
344<comb-polymers>
345Polymer thermosets have several advantages over their small molecule
346counterparts, including shortened gelation time due to less reactions
347needing to take place to achieve a volume spanning element, tunable
348architecture and phase behavior inherited from the polymer architecture,
349and the formation of stable one-pot latexes. Properties of the final
350crosslinked material are often inherited from the properties of the
351constituent linked polymers. For example, the fibrillar structure of
352collagen I results in strain stiffening behavior,@motte2013 and
353synthetic block copolymers exhibit microphase separation affecting
354vitrimer processing,@lessard2020 allowing for additional handles to
355control the properties of these crosslinked networks.
356
357Comb polymers are no exception to this, and the vast design space of these branched polymers allows for the preparation of polymers with interesting phase behavior and conformation. For example, synthesizing reactive comb polymers with easily crystallizable sidechains allows for the preparation of crosslinked thermosets containing crystalline domains in the cured product,@claesson2004@lorenzana2024 increasing it's toughness and modulus. Their high molecular weight allows for quick
358gelation times,@reynolds2020 consequently requiring less energy to cure. The low viscosity of comb polymer solutions and melts further reduces the energy needed to process polymer combs,@kong2021 and can be used to modify the viscosity of polymer blends.@2009 Moreover, by adjusting the composition and stiffness of the backbone, combs can be used as next-generation compatibilizers between immiscible polymer
359phases@dong2015 as well as tough networks and adhesives on their own.@zhang2016a@ohnsorg2024
360
361=== Bottlebrush polymers
362<bottlebrush-polymers>
363Formally defined as a subset of comb polymers, bottlebrush polymers are a class of polymers characterized by the
364presence of densely grafted side chains along a main backbone. These
365polymers can be described using three variables: the degree of
366polymerization between each graft $(n_g)$, the degree of
367polymerization of the main backbone $(n_(b b))$, and the degree of
368polymerization of the side chains $(n_(s c))$. The crowding parameter
369$(ϕ)$ is defined by $n_g$ and $n_(s c)$, and provides a mathematical
370definition for the transition from loosely grafted comb to densely
371grafted bottlebrush polymer (@fig:BottlebrushPhases). Above a threshold $ϕ$, the intense sterics of
372the densely grafted side chains increase the rigidity of and creates
373tension along@panyukov2009 the main backbone, causing the brush to adopt
374a wormlike morphology, and preventing polymer chain
375entanglements.@reynolds2020@sheiko2019@xie2019 This lack of polymer
376entanglements, evidenced by the zero shear viscosity deviating from
377$eta_0 tilde.op N^3.4$,@dalsin2014 is most often exploited to create
378super soft additive free elastomers.
379
380#figure(image("Images/BottlebrushPhase.svg", width: 100.0%),
381 placement: auto,
382 caption: [
383 Phase diagram for methacrylic bottlebrush polymers. Depending on $n_(s c)$, $n_(b b)$, and $n_g$ polymers with grafted sidechains can be categorized either as combs, bottlebrushes with rigid sidechains (RSC), stretched sidechains (SSC), or stretched backbones (SBB).
384 ]
385)
386<fig:BottlebrushPhases>
387
388Perhaps owing to their recent discovery, bottlebrush polymers as a class
389of materials are relatively unexplored, and as a consequence poorly
390understood. Models of describing the conformation of individual
391bottlebrushes@liang2018 and their melt behavior@liang2017 have been put
392forward, but these almost always describe bottlebrush polymers with
393homopolymer side chains. Only recently has the self assembly of
394bottlebrush block polymers in melt@dalsin2015 and solution@pan2021 been
395described. Bottlebrush polymers with more complicated architecture, such
396as block polymer sidechains (core-shell architecture) are poorly
397described. Relatively well understood parameters of linear polymers such
398as the Flory-Huggins polymer-solvent interaction parameter, $χ$, are far
399less well modeled for bottlebrush architectures, and only grow more
400difficult to understand when secondary interactions like hydrogen
401bonding are taken into account.@bates2017 Predicting the Kuhn lengths
402and persistence lengths of bottlebrush polymers as well remains a
403challenge, with experimental and theoretical studies offering
404contradicting
405conclusions.@dutta2019@lecommandoux2002@feuz2005@yethiraj2006@clarke2024
406Studies of bottlebrush polymers composed of reactive monomers capable of
407participating in crosslinking are almost completely absent, and only
408studied as additives to other thermosetting resins.@moon2021 Some
409samples of bottlebrush mechanochemistry have been published, though
410these have very low functionality, only bearing mechanophores at the
411backbone-arm junction.@noh2021@peterson2021 Given the unique morphology,
412lack of entanglements, large size, and potential for phase separation
413and microdomain formation, bottlebrush polymers present an opportunity
414to design high performance thermosetting materials. Already they have
415found applications in fields such as touch sensors,@reynolds2020
416pressure sensitive adhesives,@arrington2018@maw2023 solvent free
417elastomers,@daniel2016 nanocapsules for drug delivery,@huang2011 and
418antifouling coatings,@yoshikawa2022 but highly crosslinked bottlebrushes
419have not been studied to date.
420
421=== Highly crosslinked combs
422<highly-crosslinked-combs>
423Chapter 3 details my work preparing robust, highly crosslinked
424comb polymers. When designing these materials, I imagined a
425polymer of extremely high molecular weight with a high density of
426crosslinkable sites. The final crosslinked material should be a high
427modulus, insoluble solid. To achieve this, I selected GMA as the
428reactive monomer, which bears a pendent epoxide ring capable of
429participating in nucleophilic additions to create intermolecular
430crosslinks. To investigate the spatial effects of crosslinks on
431toughness of the final thermoset, I identified BMA as an attractive
432copolymer for GMA to act as a small spacer and internal plasticizer. My
433approach to identifying a highly crosslinkable, tough combs was to
434first synthesize a suite of polymers of different architectures: linear
435and comb, homopolymer and copolymer, and random and block. Then,
436I investigated crosslinking these polymers with succinic acid, a small
437di-carboxylic acid, to create a tightly crosslinked network. Finally, we
438investigated the toughness of the resultant materials. Comb polymers
439with random sidechains produced crosslinked materials with higher
440modulus and toughness than polymers with block or homopolymer
441sidechains.
442
443== Hydrogel preparation
444<hydrogel-preparation>
445=== Introduction
446<introduction-3>
447Whereas the previously discussed thermosets are designed to be dispersed
448in organic solvents or used neat, polymers can also be designed to be
449hydrophilic. When crosslinked, hydrophilic polymers form hydrogels that swell in the presence of water but do not
450dissolve. Mammalian, plant, and bacterial cells are known to synthesize, secrete, and assemble a mix of proteins, proteoglycans, and sugars to
451create extracellular matrices (ECM).@flemming2010@mouw2014@seifert2010
452The ECM in turn provides both chemical and mechanical cues to
453cells,@ringer2017 directing cell fate, morphology, and
454phenotype.@lantoine2016@wei2022@peyton2008@yeung2005 Hydrogels have for
455decades attracted researchers due to their 3D network structure closely
456mimicking that of natural ECM, with the first cell encapsulation being
457demonstrated in 1980.@lim1980 Synthetic analogues in particular have
458been an area of intense focus as means to more precisely control the
459encapsulated cell environment compared to natural products, which often
460show large lot-to-lot variability.@kozlowski2021
461
462The first report of synthetic hydrogels as we currently know them
463appeared in 1958 in a publication by Danno in which poly(vinyl alcohol)
464in aqueous solution was crosslinked under gamma irradiation to form an
465insoluble gel.@danno1958 This was soon followed up by Wichterle and Lím
466when they polymerized 2-hydroxyethyl methacrylate in the presence of
467ethylene glycol dimethacrylate to create water swollen gels for use as
468contact lenses,@wichterle1960 which is still the basis for many contact
469lens formulations to this day.@teichroeb2008@nicolson2001 In 1970 PEG
470hydrogels were prepared by irradiation with gamma and electron
471radiation,@stafford1970@king1970 and PEG gels started to attract
472attention due to their favorable biocompatibility and non-fouling
473properties.@zhu2010 The chain end alcohols of PEG have since been
474functionalized with methacrylates for the preparation of gels by
475FRP,@tanaka1977 isocyanates@graham1984 and succinimidyl
476esters,@sakai2008 and a variety of chemistries capable of participating
477in Michael-type additions.@jansen2018@buwalda2014@metters2005
478
479=== Mixing and light
480<mixing-and-light>
481Cell encapsulation techniques within hydrogels must be both rapid enough
482to prevent cell settling and cytocompatible.@caliari2016 Free radicals
483and high energy UV light are well known to cause cellular
484damage,@chen1999 though UV photoinitiators and vinyl monomers are used
485to encapsulate cells. Michael-type reactions are commonly used as
486bio-orthoganol methods of crosslinking hydrogels, and are useful for
487their ability to incorporate bioactive peptides by exploiting Michael
488donors present in amino acids, such as the thiol in
489cysteine.@peyton2006@galarza2020 Additionally, the A--B type
490crosslinking by Michael-type reactions often produces highly homogenous
491gels.@sakai2008 However, the reaction kinetics vary amongst chemistries
492used, and tuning the kinetics within the framework of physiological
493conditions can be a challenge, not to mention relative difficulties of
494biology labs synthesizing functional PEG not commercially
495available.@rizzo2023@paez2020 Moreover, Michael-type reactions most
496often do not allow for spatiotemporal control, meaning the homogeneity
497of the network is dependent on the reaction kinetics.@darling2016
498
499=== Reversible deactivation radical polymerizations
500<reversible-deactivation-radical-polymerizations>
501RDRPs, by contrast, are rarely used in biomedical applications. The
502reasons for this vary, though in general, RDRPs are slower than FRPs,
503are sensitive to trace oxygen and are traditionally thermally
504initiated. These aspects make them unsuitable for polymerization in the presence of living cells.
505ATRP most commonly requires cytotoxic metal halide salts among several
506other reactants, although there are efforts to replace them with
507powerful organic reducing agents capable of homolytically cleaving
508carbon-halogen bonds.@corbin2022 NMP is traditionally extremely
509slow@grubbs2011 and has serious challenges with monomer
510compatibility.@guegain2015 RAFT requires tuning of the chain transfer
511agent to match the desired monomers@keddie2012 and is expensive to
512source.
513
514Despite all of this, RDRPs offer significant benefits over FRPs. FRPs
515produce gels with heterogeneous networks due to unavoidable termination,
516slow initiation relative to propagation, and slow segmental relaxation
517relative to chain growth in the gel phase.@lovestead2003@gu2020 FRPs
518commonly overcome inhibition from oxygen by overwhelming it with excess
519radical species, to the detriment of cytocompatibility. In contrast,
520RDRPs are able to produce highly homogenous gel networks due to their
521high initiation relative to propagation rates and low radical
522concentrations which minimize irreversible termination events as well as
523minimizing cellular damage. Furthermore, the reversible nature of RDRPs
524enable living gels@peyton2023 by postpolymerization modifications of
525gels.@bagheri2021 Chapter 4 of this dissertation will harness RDRPs
526using the photoiniferter process to rapidly polymerize and crosslink
527biocompatible hydrophilic monomers.
528
529== Hypothesis
530<hypothesis>
531Current methods of curing thermosets are limited in their applications
532and cannot be applied in all circumstances. This is particularly true of
533biomedical applications, where conditions must be kept within the strict
534confines of physiological conditions. I hypothesize that grafting
535pendent PEG chains along the main polymer backbone will provide a facile
536way to install mechanosensitivity by sterically shielding reactive
537sites, enabling on demand gelation. Furthermore, I hypothesize densely
538grafted bottlebrush polymers with reactive side chains can be prepared
539as high-modulus, tough thermosets by adjusting the sidechain architecture to
540favor intermolecular crosslinking. Finally, I hypothesize that rapid
541fragmentation of xanthogen disulfides in response to visible light will
542enable rapid preparation and crosslinking of biocompatible hydrogels. I
543predict that that this work will enable new methods of bottom up design
544of thermosetting materials.
545
546== Objectives
547<objectives>
548The following were the objectives for this dissertation:
549
550+ Optimize a library of sparsely grafted copolymers carrying reactive
551 pendent groups and methoxy poly(ethylene glycol) (PEG) grafts for use
552 as mechanically activated crosslinkable materials;
553
554+ Synthesize and characterize novel mechanically responsive core-shell
555 bottlebrush polymers furnished with a high density of crosslinkable
556 pendant groups;
557
558+ Develop a method to rapidly crosslink hydrogels in response to light
559 in the visible range.
560
561== Significance
562<significance>
563I have identified a new path for easy installation of
564mechano-responsiveness into synthetic polymers using off-the-shelf
565chemistries and materials. Further, I have developed strategies to
566combat intramolecular crosslinking in densely grafted reactive polymers.
567Finally, I have designed a method of rapidly synthesizing and
568crosslinking biocompatible hydrogels. Overall, this dissertation
569presents easy and accessible approaches to designing strain sensitive
570polymers as well as preparing hydrogels by reversible deactivation
571radical polymerization that will enable more rapid translation of
572mechanosensitive polymers and tailor-made hydrogels. Additionally, this
573dissertation provides new strategies for preparing tough thermosets from
574high molecular weight resins.