PhD-Dissertation / Chapter2.typ

*#align(center)[#smallcaps[@molecularly-shielded-on-demand-ultrasound-cured-polymer-networks[Chapter]]]*
= #smallcaps[Molecularly shielded, on-demand, ultrasound-cured polymer networks]
<molecularly-shielded-on-demand-ultrasound-cured-polymer-networks>
Reprinted (adapted) with permission from Lorenzana, A. A. _et al._
Molecularly Shielded, On-Demand, Ultrasound-Cured Polymer Networks.
#emph[Macromolecules] #strong[57], 6465--6473 (2024). Copyright 2024
American Chemical Society.

== Abstract
<abstract>
Networks formed from polymers can range from soft hydrogels to ultrahard
protective coatings, making them useful for a wide range of applications
from cell culture to highly bonded adhesives. Polymer networks are
commonly crosslinked _via_ heat or high energy light, and recently
mechanical force has also been used to induce the formation of
crosslinks in pre-existing networks. Here, I demonstrate a new strategy
to use mechanical deformation and ultrasound to induce liquid-to-solid
crosslinking. I synthesized graft copolymers with large poly(ethylene
glycol) (PEG) side-chains acting as molecular shielding groups to
protect otherwise highly reactive epoxide group. Solutions of highly
shielded polymers could remain as a liquid solution when left
undisturbed, and I could initiate gelation of these solutions with
ultrasound in 20 seconds. These ultrasound-sensitive polymers are
particularly useful in light and heat sensitive applications, and where
precise control over the gelation time is required.

== Introduction
<introduction>
Polymer networks can be crosslinked _via_ permanent covalent bonds.
Polymer networks can include super-soft hydrogels that mimic human
tissue,@rimmer2007 protective ultra-hard coatings,@villani2015 and
highly bonded adhesives.@wang2018 Highly crosslinked lightweight
networks, such as those formed with epoxide, are crucial in industrial
applications like transportation, where reducing vehicle weight improves
passenger safety and reduces harmful greenhouse gas emissions. The
process of crosslinking or "curing" polymers is typically accomplished
via a) mixing, b) heat, c) high energy light, or d) electron
beams.@frounchi2006 Heat and light are popular routes, as they
facilitate curing on-demand, allowing liquid application to a substrate.
However, light and heat are not always feasible, as light cannot pass
through opaque materials, and heat can damage delicate or flammable
substrates. Electron beams are also popular industrially due to their
high energy efficiency and excellent uniformity, however transmittance
through metals can be challenging.

Alternatively, natural polymers (e.g. peptides, saccharides, nucleic
acids), can form networks in response to temperature, light, and
solvents by partially unfolding, thus exposing previously buried, or
"cryptic", binding sites. Of particular interest to us, these cryptic
binding sites can also be revealed in response to a mechanical
stimulus.@brown2009@bu2012 For example, fibronectin will dynamically
unfold and polymerize into fibrils in response to cell-generated
forces.@gee2008@smith2007@oberhauser2002 In contrast, synthetic polymers
commonly weaken or even rupture under force.@kuijpers2004

Inspired by the unfolding triggered crosslinking of proteins like
fibronectin, I sought to develop a new method of installing
mechanosensitivity within synthetic polymer networks. Recently, we
developed organogels@tran2017 and hydrogels@sonu2023 with
mechano-responsive properties, both based on preformed diacrylate
crosslinks with reactive pendent thiols for post-polymerization
crosslinking. Both systems begin as a crosslinked network and respond to
compression, strengthening several hundreds of kPa in elastic modulus
over repeated cycles. The mechanosensitivity results from long PEG
molecular shielding groups grafted to the polymer backbone, which
prevent the reactive thiol groups from crosslinking until compression
brings them together.

To date, the most successful method of creating synthetic
mechanosensitive polymers that undergo liquid-to-solid transition is by
inserting weak bonds, "mechanophores," within polymer chains that are
converted to an active intermediate in response to force, capable of
strengthening the material.@ramirez2013@lenhardt2015@hickenboth2007
Still other approaches to designing force-sensitive materials involve
the design of small molecules with several ways of participating in
intermolecular interactions such as hydrogen bonding, π--π stacking, and
van der Waals forces. Peptide-based isomers functionalized with
cholesterol and napthalic groups have been shown to create micellar
assemblies that undergo a gel-gel transition with the application of
ultrasound.@yu2010 This work, in contrast, uses the shielding group
concept, starting with uncrosslinked, shielded polymers that can undergo
a rapid liquid-to-solid transition upon application of force. To
accomplish this, graft polymers bearing reactive epoxide@vidil2016
groups are mixed with small molecule amine/thiol crosslinkers.
Ultrasonic irradiation is used to apply high strain rates to the
shielded polymers. Straining of the graft polymers overcomes their
steric barrier to interaction with the small molecule crosslinkers,
facilitating a reaction, that rapidly strengthens the material. The
resultant materials achieve elastic modulus values comparable to ultra
hard commercial epoxy coatings. I anticipate that these shielded
polymers will be useful as extremely hard and solvent-resistant coatings
and as adhesives that can be cured by focusing ultrasound through the
surfaces the adhesive is bound to.

== Materials and methods
<materials-and-methods>
=== Chemical and polymer sourcing
<chemical-and-polymer-sourcing>
Materials were purchased from Sigma-Aldrich unless otherwise mentioned.
(500 g/mol and 950 g/mol, PEGMA500 and respectively), glycidyl
methacrylate (97%, GMA), and 2-methoxyethyl methacrylate (99%, MEMA)
were passed through a column of neutral alumina to remove inhibitors
before use. 2,2′-(Ethylenedioxy)diethanethiol (95%, EDT), ethylene
diamine (99%, EDA), (99%, PPB), (98%, CPA), and
2-(azo(1-cyano-1-methylethyl))-2-methylpropane nitrile (98%, AIBN),
1-butanol (99.9%, BuOH), 1,4-dioxane (99%, dioxane),
N,N-dimethylformamide (99.8%, DMF) were used as received. Diethyl ether
(99%, ether), lithium hydroxide monohydrate (98.5%, LiOH), and
acetonitrile (99%, MeCN) were purchased from Fisher Chemical and used as
received. Basic alumina 60-325 mesh was purchased from Fisher Scientific
and used as received.

=== Representative polymer synthesis
<representative-polymer-synthesis>
Poly(GMA_-co-_PEGMA) and poly(GMA_-co-_MEMA) of all molar ratios and degree
of polymerization (DP) were synthesized by reversible
addition-fragmentation chain transfer (RAFT) polymerization. The
targeted monomer ratios and DP are described in @shieldedpolymerstab. Each reaction was fed 0.01 moles of monomer total.
For example, 0.71 g (0.005 mol) GMA, 0.72 g (0.005 mol) MEMA, 0.0559 g
CPA (0.2 mmol), 6.6 mg AIBN (0.04 mmol) (\[50\]:\[1\]:\[0.2\]
\[M\]:\[CTA\]:\[I\], where \[M\]:\[CTA\] defines the DP), 4 mL of
1,4-dioxane, and a stir bar were added to a 20 mL scintillation vial.
The vial was sealed with a rubber septum and the solution was purged
with N2 (g) for $tilde.op$20-30 min in an ice bath to prevent solvent
and monomer evaporation (PEGMA solutions were bubbled in cool water to
prevent PEG crystallization). Subsequently, the vial was placed in a
thermostated aluminum reaction block at 60 °C on top of a magnetic
stir/hot plate. The reaction was left to stir overnight, yielding a
viscous liquid. The solution was removed from heat and exposed to air to
terminate the polymerization. The solution was precipitated into cold
(-20 °C) ether, the solid washed twice more with cold ether, and dried
at 0.01 mbar overnight.

=== Polymer characterization
<polymer-characterization>
Polymer DP and the comonomer incorporation ratio were determined through
#super[1]H NMR on a Bruker Avance 500 at 500 MHz in CDCl#sub[3] (@fig:MEMAGMANMR - @fig:GMAMEMA7030).@izunobi2011 The ratio of monomers was
determined by integration of #super[1]H spectral resonances of the PEGMA/MEMA
methoxy protons and the methanetriyl proton of the GMA glycidyl ring,
normalized to the aromatic proton peak at the para position of the CPA
phenyl ring, assuming there is one Z group@keddie2012 on every polymer
chain.

#figure(image("Images/C2S3.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_MEMA) targeting 50 DP and 1:1
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:MEMAGMANMR>

#figure(image("Images/C2S4.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA500) targeting 50 DP and 1:1
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA500NMR>

#figure(image("Images/C2S5.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA950) targeting 50 DP and 1:1
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA950NMR>

#figure(image("Images/C2S6.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA950) targeting 50 DP and 60:40
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA9506040>

#figure(image("Images/C2S7.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA950) targeting 50 DP and 70:30
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA9507030>

#figure(image("Images/C2S8.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA500) targeting 50 DP and 60:40
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA5006040>

#figure(image("Images/C2S9.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA500) targeting 50 DP and 70:30
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA5007030>

#figure(image("Images/C2S10.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_PEGMA500) targeting 50 DP and 30:70
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAPEGMA5003070>

#figure(image("Images/C2S11.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_MEMA) targeting 50 DP and 60:40
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAMEMA6040>

#figure(image("Images/C2S12.png", width: 100.0%),
  placement: bottom,
  caption: [
    #super[1]H NMR spectra of poly(GMA_-co-_MEMA) targeting 50 DP and 70:30
    comonomer ratio. Spectra recorded at 500 MHz in CDCl#sub[3].
  ]
)
<fig:GMAMEMA7030>

=== Copolymer solution preparation
<copolymer-solution-preparation>
Polymer solutions were initially prepared to be 50 wt% polymer. For
example, 0.3 g of polymer was dissolved in 0.3 g of solvent, and
crosslinker was added such that the nucleophilic functional group was
equimolar with the total epoxide concentration. To control for the
concentration of crosslinking points in solution, polymers were
subsequently formulated to be 1 M of epoxide in solution. For copolymers
containing PEGMA2000, solutions were formulated in MeCN at 0.5 M of
epoxide functional units due to the large pendant chains dominating the
overall mass of the sample and crosslinked with EDT catalyzed by LiOH.
Each sample was vortexed for 5 sec to ensure complete mixing before
proceeding with rheometry or sonication.

For all experiments crosslinked with amines, reactions were conducted in
a solvent system of 1:1 BuOH:DMF. Alcohols are known to catalyze the
reaction between amines and epoxides through the formation of a
trimolecular complex.@ehlers2007 Thiol-crosslinked reactions were
conducted in MeCN with 10 µL of 2 M LiOH as a catalyst, necessary to
deprotonate the thiols in order to perform a nucleophilic attack on the
epoxide ring.@gadwal2015

=== Parallel plate rheology
<parallel-plate-rheology>
Gelation times and storage moduli ($G'$), and $t a n delta$ of polymer
solutions/gels were determined on a Kinexus Pro parallel plate rheometer
(Netzsch, Selb, Bayern, Germany). Measurements were run on a 20 mm plate
with a 1 mm gap at 1% strain and 1 -- 100 rad s#super[-1] frequency
sweep. Each frequency sweep lasted approximately 5 min, and the entire
measurement lasted approximately 15 hr. The gel point was defined using
the Winter-Chambon criterion, for which the time of gelation is defined
as the point at which $t a n delta$ becomes frequency independent at
small frequencies.@winter1986@chambon1985@chambon1987 For samples with
very high modulus, the elastic modulus was determined using compressive
rheology by taking the slope of the stress strain curve of cured gels
with a 4 mm diameter. Rheological experiments were analyzed using IRIS
Rheo-Hub (IRIS Development, Amherst, MA).@poh2022

=== Sonication-induced gelation of shielded polymers
<sonication-induced-gelation-of-shielded-polymers>
Polymer solutions were sonicated using a QSonica Q500 with a microtip
attachment. The microtip QSonica probe was immersed in a polymer
solution in MeCN. Water was flowed across the outer surface of the tube
using a custom-made jacketed beaker to control bulk temperature
(University of Massachusetts Amherst Scientific Glassblowing Laboratory,
Amherst, MA). Temperature was monitored with an IRT205 IR thermometer
(General Tools, Secaucus, NJ) and confirmed with a mercury thermometer.
This cooling setup was not sufficient to control temperature after 2 min
and 40 s of sonication. Samples were sonicated at 10% amplitude and 20
kHz for 5 sec at a time, with 10 sec breaks in between pulses to avoid
probe overheating. Gelation was determined by the point at which the
power output would drop to $tilde.op$0 W and noise from vibrations would
cease when the polymer had formed a solid gel. Samples were then
immediately moved to the adjacent needle induced cavitation (NIC) setup
to determine the elastic modulus immediately post sonication.

=== Differential scanning calorimetry
<differential-scanning-calorimetry>
Differential scanning calorimetry (DSC, Q200, TA Instruments) was used
for crystallization characterization. A sample of (3-5 mg) was sealed in
a standard aluminum hermetic pan using TZERO press (TA Instruments)
before being added to the calorimeter with an identical empty reference
pan. The equipment was lowered to -90 °C and heated to 100 °C at a rate
of 5 °C/min to remove the thermal history of the sample. The equipment
was then lowered to -90 °C again and heated to 100 °C at the same rate,
where enthalpy of melting (\$\\Delta H\\textsubscript{m}\$) was obtained
from the area of the melting curve divided by the sample
weight.@kong2002 Thermogravimetric analysis (TGA, Q50, TA Instruments)
was used to determine the degradation of the samples before running DSC
to meet the criteria of a maximum 1.5 wt% loss.

=== Needle induced cavitation
<needle-induced-cavitation>
Characterization of elastic modulus of sonicated gels was done with
needle induced cavitation (NIC) using a custom-made setup with water as
the fluid, pressurized with a NE 1000 syringe pump (New Era,
Farmingdale, NY), contained in a 6 mL disposable syringe with a 27-gauge
stainless steel disposable needle, microstand, and Px409-015 GUSBH
pressure gauge (Omega, Norwalk, CT). Data collected from NIC was
recorded on a Surface Mini using a custom LabView program to interface
with the pressure sensor and record the pressure values (Crosby Lab,
University of Massachusetts Amherst, Amherst, MA). When calculating the
elastic modulus of gels, the effects of surface tension were ignored and
values were computed using Equation 1.@dougan2022@barney2020 Each NIC
experiment lasted on average from 30-90 sec. $ E = frac(6 P_c, 5) $

== Results and discussion
<results-and-discussion>
=== Gelation kinetics of polymers under static conditions
<gelation-kinetics-of-polymers-under-static-conditions>
My goal was to create a polymer network that was shelf-stable and would
gel in response to force. First, I created a suite of polymers with
varying crosslinker to comonomer ratios. Shielded and control copolymers
were synthesized using RAFT polymerization of PEGMA (molecular shielder,
grafting-through process@li2021) or MEMA (control) with GMA monomers.
Poly(GMA-co­-PEGMA) and poly(GMA_-co-_MEMA) were synthesized with varying
monomer ratios (30:70 GMA:PEGMA/MEMA to 70:30 GMA:PEGMA/MEMA) and shield
lengths (1, $tilde.op$10, and $tilde.op$20 PEG repeats for MEMA,
PEGMA500, and PEGMA950, respectively) to determine their effect on
gelation. Additionally, DP for each composition was varied to determine
the effect of polymer length on force sensitivity.

When developing these materials, I imagined a polymer system that would
be easily spreadable onto a substrate as a liquid that would then
transition to a solid state after the introduction of mechanical
stimuli. The final solid material should be bonded together permanently
with covalent crosslinks. To achieve this goal, I selected the monomer
GMA for its robust epoxide reactive group. Epoxides are known to undergo
a ring-opening reaction in the presence of nucleophiles like amines and
thiols. To introduce mechano-sensitivity, I sought to copolymerize my
epoxide functional monomers with monomers functionalized with groups
that could provide steric hindrance. Towards this goal, GMA was
co-polymerized with PEGMA of varied molecular weights from 140 to 950
g/mol that I hypothesized could provide a steric hindrance to
crosslinking _via_ their ether side-chains.

Synthesis of this suite of polymers proceeded as expected, with final
DPs and incorporation ratios closely matching the targeted DP and feed
ratio when conducted in dioxane (@shieldedpolymerstab). Successful
incorporation and molar ratio of constituent monomers was confirmed
using #super[1]H NMR spectroscopy (@fig:MEMAGMANMR - @fig:GMAMEMA7030). DP and incorporation ratios
of poly(GMA_-co-_PEGMA) samples were less consistent compared to their
MEMA counterparts, attributed to the inherent dispersity of PEGMA
macromonomers skewing the actual molar amount added to reactions. After
successfully synthesizing the desired copolymers, I moved on to assess
their gelation kinetics.


#figure(
  placement: auto,
  align(center)[#table(
    columns: 5,
    align: (center,right,right,right,right,right,right,),
    table.header([Name], [Target DP], [Feed ratio], [Actual DP], [Actual
      ratio],),
    table.hline(),
    [50:50 GMA:MEMA], [50], [1:1], [73], [55:45],
    [50:50 GMA:PEGMA500], [50], [1:1], [122], [51:49],
    [50:50 GMA:PEGMA950], [50], [1:1], [91], [56:44],
    [60:40 GMA:PEGMA950], [50], [60:40], [102], [62:38],
    [70:30 GMA:PEGMA950], [50], [70:30], [87], [72:28],
    [30:70 GMA:PEGMA500], [50], [30:70], [76], [30:70],
    [60:40 GMA:PEGMA950], [50], [60:40], [119], [60:40],
    [70:30 GMA:PEGMA500], [50], [70:30], [95], [68:32],
    [60:40 GMA:MEMA], [50], [60:40], [134], [60:40],
    [70:30 GMA:MEMA], [50], [70:30], [140], [70:30],
    [50:50 GMA:PEGMA950 25DP], [25], [1:1], [34], [1:1],
    [50:50 GMA:PEGMA950 50DP], [50], [1:1], [61], [54:46],
    [50:50 GMA:PEGMA950 100DP], [100], [1:1], [93], [1:1],
    [50:50 GMA:PEGMA950 150DP], [150], [1:1], [130], [52:48],
    [50:50 GMA:PEGMA950 200DP], [200], [1:1], [191], [57:43],
  )]
  , caption: [Polymers used in each experiment, their target DP,
  comonomer feed ratio, actual DP, and actual comonomer ratio as
  determined by #super[1]H NMR.]
  , kind: table
  )
<shieldedpolymerstab>
For the crosslinkers, I chose EDT due to its non-volatile nature and
reasonable stability in air, and EDA as it is commonly used to cure
epoxy resins. Amines and thiols were chosen as two candidates both
because they are frequently used in commercial epoxy formulations and to
compare the effects of different reaction kinetics on the shielded
copolymer system. I sought to determine a molecular weight of shielding
groups that would facilitate delayed crosslinking of the epoxide groups
in the presence of a bifunctional nucleophile without preventing it
entirely. In my experiments, I tested a range of effects including
varying the DP of grafted chains from 1 to 20, varying the DP of the
polymer backbone from 25 to 670, adjusting nucleophilic attack kinetics,
and varying the ratio of comonomers from 30 to 70% GMA concentration.
The monomers used to form the copolymers and the different crosslinkers
in these experiments are represented in @fig:shieldinglength a.

#figure(image("Images/C2F1.png", width: 100.0%),
  placement: auto,
  caption: [
    Large molecular shields inhibit or delay crosslinking. (a)
    Illustrations of polymer components used throughout the paper. (b)
    Effect of shielding group functionality on storage modulus ($G'$) over
    time with amine crosslinks at constant 50 wt % polymer, reacting
    with EDA. Inset depicts high density poly(GMA_-co-_MEMA) with many
    epoxy groups. (c) Effect of shielding group functionality on storage
    modulus over time with amine crosslinks at constant 1 M
    concentration epoxy, reacting with EDA. Inset depicts low density
    poly(GMA_-co-_MEMA) with a fixed amount of epoxy groups. (d) Effect of
    shielding group functionality on storage modulus over time with
    thiol crosslinks at constant 1 M concentration epoxy, reacting with
    EDT. Inset depicts poly(GMA_-co-_PEGMA950) with a fixed amount of
    epoxy groups and large shielding groups preventing crosslinking. For
    all experiments, a 1:1 ratio of GMA:MEMA, PEGMA500, or PEGMA950 was
    used. Error bars show the standard deviation of $G'$ at each timepoint
    (n = 3). For all conditions, including the enhanced kinetics
    provided by the thiol-epoxy reaction, a latency period before
    gelation at static conditions is present.
  ]
)
<fig:shieldinglength>

First, the effect of pendent shield size on crosslinking was assessed at
constant weight percent and static conditions (@fig:shieldinglength b). When solutions are formulated at 50 wt% of
polymer with EDA, poly(GMA_-co-_MEMA) crosslinks very quickly (1 h) and
reaches a final $G'$ on the order of 106 Pa. Conversely,
poly(GMA_-co-_PEGMA500) crosslinks more slowly (8 h) and reaches a final
$G'$ on the order of 104 Pa, and poly(GMA_-co-_PEGMA950) shows no change
in modulus indicating no crosslinking occurred. At constant 50 wt% of
polymer in solution, the concentration of epoxide for unshielded samples
(MEMA) is very high compared to the shielded polymers (PEGMA). At this
fixed concentration, the overall mass for the shielded polymer solutions
is dominated by the presence of ether in the PEGMA side chains, skewing
the sample in favor of unreactive ether and decreasing the number of
possible crosslinks. The lack of increase in modulus with PEGMA950 may
be due to this ether dominance preventing the formation of a
volume-spanning network. Additionally, the relatively large mass of the
ether side-chains decreases the amount of reactive epoxy in solution.

To control for the effect of variable epoxide concentration, samples
were next formulated at a constant epoxide molar concentration (@fig:shieldinglength c). Epoxide concentration was set to 1 M, resulting
in variable weight percent polymer in solution: control polymer samples
with low (25%) and shielded samples with high (61%) weight percent. At
25 wt%, poly(GMA_-co-_MEMA) crosslinks more slowly (2 h) than at 50 wt%
and reaches a lower final $G'$ on the order of 105 Pa. For
poly(GMA_-co-_PEGMA950), wt% changes from 50 to 61 and expectedly shows
only a small increase in $G'$ of 35 Pa. For poly(GMA_-co-_PEGMA500
samples, 1 M epoxide concentration is equal to 50 wt% of polymer. Trends
in the effect of shielding groups are the same at constant wt% polymer
or mol% epoxides: as the shielding group MW increases, the time to
gelation and the final modulus both decrease.

Finally, the effect of more reactive nucleophiles on crosslinking were
investigated by replacing EDA with EDT and keeping the mol% epoxide
constant (@fig:shieldinglength d). Thiols are known to be stronger
nucleophiles than primary amines, and the ring opening reaction between
thiols and epoxides proceeds orders of magnitude faster than between
amines and [e-mail]2013 At a constant 1 M epoxide
concentration, poly(GMA_-co-_MEMA) with EDT crosslinked more rapidly (30
min) than the amine condition and attained a similar final $G'$.
Poly(GMA_-co-_PEGMA500) samples crosslinked rapidly (42 min) with EDT, but
more slowly than the MEMA copolymer and attained a final modulus on the
order of 104 Pa. Poly(GMA_-co-_PEGMA950) samples still did not show any
signs of gelation, increasing only to a final modulus of 10 Pa. Even
with faster reaction kinetics, the PEGMA950 shielding groups suppress
gelation.

For permanently crosslinked polymer networks, the equilibrium modulus of
the cured material can be predicted by Flory's theory of rubber
elasticity@flory1941@flory1953 and is proportional to the number of
elastically effective chains in the network.@ferry1980@kulicke1989 As
the number of elastically effective chains increases, so does the
equilibrium modulus; therefore, a low equilibrium modulus implies the
presence of unreacted crosslinks. With the same number of crosslinks
possible in MEMA, PEGMA500, and PEGMA950 samples, and taking the
equilibrium modulus of the MEMA polymer in @fig:shieldinglength c,
PEGMA500 and PEGMA950 can be inferred to be have a lower crosslinking
desntiy due to the protective effects of the polyether chains. This led
us to believe that 950 g/mol shielding groups are most effective at
creating a steric barrier to reaction, preventing crosslinking between
adjacent polymers and resulting in lower final $G'$ values.

=== Controlling gel time through shield graft density
<controlling-gel-time-through-shield-graft-density>
We next aimed to determine the minimum molar ratio of shielding groups
necessary to prevent spontaneous crosslinking by varying the ratio of
GMA:PEGMA (@fig:shieldingratio a). I expected that high contents
of shielding monomer would entirely inhibit gelation over the
measurement time, eventually prohibiting crosslinking even under force.
To assess the minimum molar ratio necessary for preventing gelation
without applied mechanical stimulus, the mole percent of PEGMA950
($tilde.op$20 repeat units) and PEGMA500 ($tilde.op$10 repeat units)
shielding monomers within each polymer chain was varied from 30 to 50
mol%. Variations in mole percent of MEMA copolymers was assessed as a
negative control. The total concentration of epoxides in solution
remained constant at 1 M.

#figure(image("Images/C2F2.png", width: 100.0%),
  placement: auto,
  caption: [
    Ratio of pendent shields to reactive groups controls gelation time.
    (a) Illustrations of polymers at different GMA:PEGMA molar ratios,
    showing the change in backbone flexibility and exposed reactive
    sites. b-e. Storage modulus evolution over time for: (b-c) varying
    mole percentage of PEGMA950 with a diamine (b) or dithiol (c)
    crosslinker; (d) varying mole percentage of PEGMA500 and (e) MEMA
    with a diamine crosslinker. Arrows represent trends in shielding
    resulting from increased ratio of shielding monomer.
  ]
)
<fig:shieldingratio>

In the presence of EDA or EDT with shielding group concentrations at mol
50% (PEGMA950), a negligible increase in $G'$ was seen; at 40%, a very
slow increase in $G'$ with a final value on the order of 103 Pa was
demonstrated; and at 30%, a rapid increase in $G'$ with a final $G'$ of
104 Pa (@fig:shieldingratio b-c). At higher shielding monomer
percentages, gelation was entirely inhibited over the measurement time,
even with the quick crosslinking EDT. In both the thiol and amine cases,
the trend toward decreasing gel time with increasing PEGMA950 content is
the same.

Next, polymers with PEGMA500 shielding units ($tilde.op$10 repeat units)
were varied from 30 to 70 mol% shielding monomer content while keeping
the total epoxide group concentration in solution constant at 1 M
(@fig:shieldingratio d) with EDA. When the shielding group
concentration was 30 and 40 mol%, the material crosslinks rapidly and
reaches final $G'$ values on the order of 105 Pa. At 50% concentration
of shielding groups, the material reaches a lower final modulus on the
order of 104 Pa. At the maximum tested 70% molar ratio of shielding
group to reactive group, the shielded polymers still form a gel, but do
not attain an equilibrium modulus during the experimental timeframe. The
PEGMA500 shielding groups do not provide a sufficient steric barrier to
reaction but do provide some hinderance to reaction evidenced by the
decreased final modulus values compared to control samples.

Finally, polymers with one repeat unit pendent chains (MEMA) were varied
between 30 and 50 mol% control monomer and reacted in the presence of
EDA. Increasing the control monomer ratio from 30 to 50% slightly
decreased to rate at which the material crosslinked and the final
modulus, from 105 Pa at 30 and 40 mol% control monomer to just above 104
Pa at 50 mol% control monomer (@fig:shieldingratio e). As
expected, the small size of the MEMA comonomer did not contribute
significantly to suppressing the crosslinking kinetics of the
crosslinking polymers.

At low ratios of shielding monomer to reactive monomer, there are
statistically likely to be more stretches of reactive monomer with no
steric effects to prevent them from crosslinking, as well as increased
backbone flexibility. At high ratios of shielding monomer to reactive
monomer, there are far fewer reactive monomer sequences as well as a
straighter backbone due to pendent chains preventing backbone flexing.
In summary, only the compositions achieved this with a high degree of
shielding. Gelation was completely inhibited at a 1:1 ratio of reactive
to shielding groups. This composition was selected as the most promising
candidate for force-activated gelation.

=== Force-induced gelation of shielded polymers
<force-induced-gelation-of-shielded-polymers>
We hypothesized sonication would be a facile method to mechanically
induce gelation of shielded polymer. Sonication can achieve enormous
strain rates approaching 10#super[8] s#super[-1].@hennrich2007 This
enormous strain rate arises from cavitations introduced during
ultrasonic irradiation, nearly instantaneously creating and destroying
microscopic bubbles that in turn create pressure gradients able to apply
force through fast solvent flows to polymers of sufficient size. The
force accumulated along the polymer backbone result in overstretched
regions, which is what is generally accepted to drive conventional
mechanochemical reactions.@oneill2023

#figure(image("Images/C2F3.png", width: 100.0%),
  placement: auto,
  caption: [
    Sonication induces shielded polymer crosslinking. (a) Gel time of
    poly(GMA_-co-_PEGMA950) under static and sonicated conditions at
    varying DP with 1:1 molar ratio. Samples at 25 and 50 DP did not
    form a gel. Insets show a liquid polymer solution during a bubble
    test and a polymer cured through sonication, still attached to the
    sonicator probe. (b) Elastic modulus of poly(GMA_-co-_PEGMA950) cured
    with sonication as measured _via_ NIC. Samples were crosslinked with a
    1:1 molar ratio of thiol to epoxy and at a DP of 100, 150, or 200
    and measured 60 s post sonication and after two weeks.
  ]
)
<fig:sonication>

Crosslinking of shielded polymers induced #emph[via] sonication was
assessed at DP of 25 to 200 monomer units per chain (@fig:sonication a). Each polymer sample was prepared at 1 M epoxide group
concentration and reacted with EDT catalyzed by LiOH. Utilizing an
ultrasonic probe immersed in polymer solutions, samples were subjected
to ultrasonic waves for 5 s at a time, with 10 s of pause in between to
prevent probe overheating. All conditions have delayed gelation at
static conditions, allowing for the characterization of faster
crosslinking with induced strain. At DP equal or greater to 100, samples
gelled within 60 s of sonication time. At 100 DP, I observed a two
order of magnitude decrease in gelation time when comparing unperturbed
samples with sonicated samples. Samples of DP 150 and 200 gelled more
rapidly, within 30 and 20 seconds of sonication time, respectively.
Poly(GMA_-co-_PEGMA950) of lower DP (25 and 50) did not show any strain
responsiveness, and the solution boiled before any gelation or viscosity
change was observed due to the heat generated by the ultrasonic probe,
reaching a temperature of 56 °C measured through an IR thermometer, at
which point the solution began to boil while sonication was being
applied. Counterintuitively, the heat generated by sonication is
counterproductive to gelation of this system, possibly due to changes in
the conformation of PEGMA shielding groups at higher temperatures
(@fig:shieldedtemp). It is well understood that PEGMA copolymers
have a lower critical solution temperature in water that is dependent on
the polyether length and the ionic strength of the environment,@lutz2006
but it is not clear that this behavior extends into aprotic organic
solvents. Gelation time under static conditions decreased as a function
of DP like sonicated samples but showed a leveling off after 150 DP
unlike the sonicated samples. This decrease in gel time is likely due to
the longer backbone lengths of the polymers beginning closer to the
percolation threshold for gelation, resulting in fewer epoxide-thiol
reactions needing to take place to form a volume spanning elastic path
and a shorter time to the critical gel.@daoud2000@winter2016

#figure(image("Images/C2S1.png", width: 100.0%),
  placement: auto,
  caption: [
    Evolution of $G'$ for poly(GMA_-co-_PEGMA950) at a 1:1 molar ratio of
    comonomers and 25 DP during parallel plate rheology. Frequency
    sweeps were run from 1 to 100 rad/s over 17.5 hrs using a 20 mm top
    plate. Data plotted at 1 Hz and 1% strain. At 25 °C the sample
    increases in modulus rapidly after a 2.5 hr latency period. At 40 °C
    the same shows a small uptick in modulus after 12.5 hrs and never
    fully gels during the measurement period.
  ]
)
<fig:shieldedtemp>

It has been shown that polymers of sufficient molecular weight are
sensitive to shear forces. The large size of polymers results in
restriction of bond angle conformers available due to chain and bond
torsional strain, meaning polymers can accumulate force along their
backbone as entropic potential
energy.@wang2013@hermes2011@shi2006@cui2009 High molecular weight
polymers undergo chain scission in response to strong shear forces
generating two distinct carbon-centered radicals.@caruso2009@beyer2005
These sufficiently strong shear forces result in overstreched segments
of polymer adjacent to the chain center, generating a tensile force that
drives mechanochemical reactions.@oneill2023 The chain scission rate
increases with molecular weight.@madras2000 This molecular weight
dependence is more accurately described as a polymer length
dependence.@may2016 It follows that shielded poly(GMA_-co-_PEGMA950) of
sufficient DP is more easily influenced by shear forces in solution if
the chain length is long enough, surpassing at least 100 units in
length. The increased DP of the polymer also increases the viscosity of
the sample. Prior literature has shown that highly viscous media
decreases the effectiveness of ultrasonic micromixing,@monnier1999
making it less likely that the dependence of gel time on DP is a result
of mixing phenomena. This study does not elucidate the mechanism for
this system's strain sensitivity. It is not clear what aspect of
crosslinking is sped up by the application of ultrasound, the addition
of EDT to polymer or the addition of polymer+EDT to another polymer.
Future studies using mono-thiols functionalized with UV tags would shed
light on the precise molecular mechanism of strain-sensitive
crosslinking.

Cavitation rheology was used to assess post-gelation elastic moduli of
gels formed _via_ sonication (@fig:sonication b). NIC has previously
been shown to be effective at extracting elastic modulus information
from soft materials.@dougan2022 Sonicated samples were measured to have
an elastic modulus near 1 kPa for samples starting at 100 DP, and 20 kPa
for samples between 150 and 200 DP as measured by NIC. After a week of
resting in a sealed tube to allow for residual epoxides to be consumed
by thiols, the modulus of each sample increased to an average of 20 kPa
for samples starting at 100 DP and 60 kPa for samples starting at 150 to
200 DP. The final modulus for 150 and 200 DP polymers had a wide range,
varying from 30 to 170 kPa. This variance is likely error from
cavitation rheology, which tends to have higher variance for samples
with higher elastic moduli.@zimberlin2007@barney2019 The modulus derived
from NIC shows polymers shielded with PEGMA950 cure into relatively weak
materials.

=== Ultrahard materials from shielded copolymers
<ultrahard-materials-from-shielded-copolymers>
Conventional epoxy resins and composites can attain $G'$ values
approaching and surpassing 10#super[9] Pa.@baral2008@maka2015 Choosing
this value as a benchmark for comparison, I formulated
poly(GMA_-co-_PEGMA2000) copolymers at a 1:1 monomer ratio and 670 DP. The
extremely long shielding group and long DP were chosen to provide a
material that had both maximum latency and sensitivity to ultrasound.
After sonicating these samples and leaving them to cure for 48 hr, the
polymer crosslinked into an opaque white solid. Samples were prepared as
5x4 mm cylinders, and their moduli were assessed on a rheometer via
compression with a 4 mm diameter plate. An elastic modulus value of 62
MPa was extracted from the resultant stress-strain curve (@fig:ultrahardshield a), approaching that of conventional epoxy
materials. Immersing gels of this copolymer into acetone and ethanol
showed no visible change in the material, but in MeCN, DCM, and water
the gels crumbled into insoluble chunks (@fig:ultrahardshield b),
leading me to conclude that the material's strength comes from a
combination of epoxide-thiol covalent crosslinks and PEG side chain
crystallization. It is well known that graft copolymers with
crystallizable side chains will form crystal
domains.@takeshita2010@inomata2005 Using DSC I was able to measure a
melting temperature for a cured GMA:PEGMA2000 sample. confirming the
material is partially crystallized (@fig:dsc). Using a steric
shielding approach, I created an ultrahard material through an
unexpected combination of crystallinity and covalent bonding.

#figure(image("Images/C2F4.png", width: 100.0%),
  placement: auto,
  caption: [
    Shielded copolymers create ultrahard and durable materials. (a)
    Compression modulus of a fully cured poly(GMA_-co-_PEGMA2000) with 1:1
    molar ratio of monomers. Elastic modulus is calculated by taking the
    slope during the linear portion of the stress-strain curve. Red line
    shows the linear best fit through four points. (b) Fully cured
    poly(GMA_-co-_PEGMA2000) gels immersed into acetone, ethanol, water,
    acetonitrile, and dichloromethane.
  ]
)
<fig:ultrahardshield>

#figure(image("Images/C2S2.png", width: 100.0%),
  placement: auto,
  caption: [
    Differential scanning calorimetry thermogram of 1:1 GMA:PEGMA2000
    crosslinked with EDT with all acetonitrile solvent evaporated off.
    $Δ$H#sub[m] of the sample was calculated to be 97.9 J/g with T#sub[m] at 49.29 °C.
  ]
)
<fig:dsc>

== Conclusion
<conclusion>
We synthesized novel strain-sensitive shielded polymers containing both
reactive epoxides and molecular shields. These shielding PEG chains
provide a steric barrier to an otherwise powerful and efficient
crosslinking reaction between amines or thiols and epoxides. This
approach to creating strain sensitive materials provides a facile route
to creating strain responsive coatings and adhesives, using well-known
and commercially available monomers. Through this I demonstrated, for
the first time, a liquid-to-solid transition accelerated under force
using shielded reactive polymers. I showed that force stimulated
gelation could be achieved with ultrasound. I further showed that
steric shielding can create ultrahard materials. Suppressed gelation
without force, combined with ultrasound sensitivity, make this polymer
an ideal candidate for an adhesive in a heat or light sensitive
application.

#emph[The work in this chapter represents a collaboration with Jichao
Song and Professor Jessica Schiffman from the Chemical Engineering
Department at University of Massachusetts Amherst. DSC measurements in
@fig:dsc were conducted and data analyzed by Jichao Song. NIC
data was collected and analyzed by Hsu Shwe Yee Naing.]