Glass-transition temperature governs the thermal decrosslinking behavior of Diels–Alder crosslinked polymethacrylate networks

Article abstract of DOI:10.1002/pola.29524

A series of Diels–Alder (DA) crosslinked polymethacrylate networks covering a broad range of glass-transition temperatures (T_g) was prepared to establish the relationship between the T_g and the thermal decrosslinking behavior of these networks. A series of permanently crosslinked and uncrosslinked analogues were also prepared to better understand the thermoset-to-thermoplastic transition occurring in the DA networks at elevated temperatures. The network series were studied using dynamic mechanical analysis, which established an inverse relationship between T_g and decrosslinking ability. Differential scanning calorimetry confirmed the viability of the DA linkages in all formulations, and a trapping experiment with 9-anthracenemethanol demonstrated that even the least responsive network was capable of undergoing decrosslinking given appropriate thermal treatment. While polymer chain mobility has long been understood to be a critical factor in healable materials, this work verifies the importance of this parameter in the decrosslinking of DA networks.

Full text of DOI:10.1002/pola.29524

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ORIGINAL ARTICLE
Glass-transition temperature governs the thermal decrosslinking behavior
of Diels–Alder crosslinked polymethacrylate networks
Daniel J. Dobbins,1,2 Georg M. Scheutz,1 Hao Sun ,1 Christopher A. Crouse,2
1
Brent S. Sumerlin
1George and Josephine Butler Polymer Research Laboratory, Center for Macromolecular Science and Engineering, Department of
Chemistry, University of Florida, Gainesville, Florida 32611-7200
2Air Force Research Laboratory, Munitions Directorate, Eglin AFB, Florida 32542
Correspondence to: B. S. Sumerlin (E-mail: sumerlin@chem.ufl.edu)
Received 8 July 2019; Revised 24 September 2019; accepted 27 September 2019
DOI: 10.1002/pola.29524
ABSTRACT: A series of Diels–Alder (DA) crosslinked polymethacrylate
networks covering a broad range of glass-transition temperatures
(Tg) was prepared to establish the relationship between the Tg and
the thermal decrosslinking behavior of these networks. A series of
permanently crosslinked and uncrosslinked analogues were also pre-
pared to better understand the thermoset-to-thermoplastic transition
occurring in the DA networks at elevated temperatures. The network
series were studied using dynamic mechanical analysis, which
established an inverse relationship between Tg and decrosslinking
ability. Differential scanning calorimetry confirmed the viability of the
DA linkages in all formulations, and a trapping experiment with
9-anthracenemethanol demonstrated that even the least responsive
network was capable of undergoing decrosslinking given appropri-
ate thermal treatment. While polymer chain mobility has long been
understood to be a critical factor in healable materials, this work ver-
ifies the importance of this parameter in the decrosslinking of DA net-
works. © 2019 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym.
Chem. 2019
KEYWORDS: Diels-Alder polymers; dynamic mechanical proper-
ties; glass transition
formation.2 The application of a proper stimulus results in
linkage reversion, affording uncrosslinked polymers, oligo-
mers, or monomers depending on the location of the revers-
ible linkage.
INTRODUCTION Plastics have historically been categorized by
the effect of temperature on bulk mechanical properties.1
Thermoplastics soften and eventually flow when heated,
whereas thermosets retain mechanical integrity until thermal
degradation. This dissimilarity in mechanical response to ther-
mal stimulation is a result of differences in molecular struc-
ture. Whereas thermoplastics are comprised of uncrosslinked
chains, thermosets consist of an infinite covalent network.
Although thermoplastics do have an advantage over thermo-
sets in that they can be thermally remolded, the glass-
transition temperature (Tg) limits the service temperature of
the material, often barring thermoplastics from use in applica-
tions requiring sustained exposures to elevated temperatures.
Conversely, the infinite covalent network morphology renders
thermosets attractive for many situations requiring durability
under harsh conditions, such as the construction industry,
electrical applications, aerospace engineering, insulation mate-
rials, and specialty adhesives. Despite these desirable attri-
butes, this infinite covalent network also precludes end-of-life
reprocessing or recycling. Covalent adaptable networks
(CANs) serve as a bridge between thermoplastics and thermo-
sets, utilizing dynamic-covalent linkages for network
A wide variety of dynamic-covalent linkages3 have been harnessed
for the development of stimuli-responsive networks. Popular
examples include boronate esters,4–9 imines,10–12 oximes,13
enamines,14,15 hydrazones,16 and disulfides,17–22 to mention a
few.23 Diels–Alder (DA) adducts, in comparison, often feature
superior orthogonality,24 tunability,25 and cyclability.26 The DA
reaction product is an example of a thermally reversible dynamic-
covalent linkage: a bond capable of undergoing thermally induced
dissociation without undergoing decomposition or other irrevers-
ible side reactions.27 While numerous other thermally reversible
covalent chemistries have been applied to polymer substrates,28
the aforementioned qualities have ensured the privileged position
of the DA reaction in modern polymer science. First reported in
1928,29 the DA reaction is a popular choice for the construction of
thermally reversible networks.30–36 Adducts of furans and
maleimide derivatives are the most commonly encountered DA
products, largely due to the conveniently accessible temperatures
at which the forward and reverse reactions proceed (e.g., the
Additional supporting information may be found in the online version of this article.
© 2019 Wiley Periodicals, Inc.
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2019
1

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
JOURNAL OF
POLYMER SCIENCE
ORIGINAL ARTICLE
WWW.POLYMERCHEMISTRY.ORG
cycloaddition of underivatized furan and maleimide occurs at
60 ꢀC or below, and the cycloreversion takes place at temperatures
of 90 ꢀC or greater).37 While much of the literature focuses on
completely transforming dienes and dienophiles to cyclohexenes
and vice versa, the dynamic equilibrium between the two at inter-
mediate temperatures has also been considered.38 By crosslinking
a furfuryl-pendent oligomer with bismaleimide, Zhang et al. were
able to demonstrate network plasticity at temperatures between
60 and 100 ꢀC.
(DMSO-d6, 99.9% D + 0.05% V/V TMS) were purchased from
Cambridge Isotope Laboratories, Inc.
The photopolymerization cuvettes were prepared by cutting
12" × 12" sheets of 2 mm thick 50A (medium) high-
temperature silicone rubber (McMaster-Carr) with a razor
blade to form a rectangular gasket fitting flush between two
plain glass microscope slides (Fisherfinest™ Premium,
Fisher Scientific), all of which were rinsed with DCM and
acetone and were dried prior to assembly. After cleaning,
the glass slides were coated once with Chem-Trend Zyvax®
Sealer GP™ mold sealer and allowed to dry overnight. The
apparatus was held together with 0.500 binder clips (Office
Depot).
The effect of Tg on bond exchange reactions in dynamic-
covalent systems, particularly in healable materials, is a signif-
icant consideration.39–41 Specifically, the requirement for
healable systems to possess a sufficiently low Tg in order to
render the polymer chains adequately mobile for reactive
functionalities to undergo bond exchange is one of the most
important aspects of designing self-healing networks.42,43
Since reports concerning dynamic-covalent healable networks
possessing high Tg are rare, especially those utilizing DA moie-
ties, we were interested in probing the effect of Tg on network
decrosslinking induced by the retro Diels-Alder (rDA) reac-
tion. Toward this end, a DA crosslinker, consisting of a DA
linkage flanked by two polymerizable functionalities, was pre-
pared and characterized. This crosslinker was copolymerized
with various monomers to prepare a series of copolymer net-
works of varying Tg, and these materials were characterized
by dynamic mechanical analysis (DMA) to monitor the evolu-
tion of mechanical properties with temperature. In particular,
the reduction of mechanical properties at higher temperatures
due to the rDA reaction was correlated with Tg, and conclu-
sions were drawn regarding the backbone contribution to the
thermally induced decrosslinking of DA networks.
Methods
All polymerizations were performed with UV curers that con-
tained four 9 W fluorescent UV light tubes with a peak emis-
sion at 365 nm and a combined intensity between 3.1 and
4.9 mW/cm2 as determined by a General UV513AB Digital
UVA/B Light Meter calibrated at 365 nm. 1H and 13C NMR
spectra were recorded at room temperature in DMSO-d6 or
CDCl3 using an Inova 500 MHz spectrometer. Chemical shifts
(δ) are given in parts per million (ppm) relative to TMS and
referenced to residual protonated solvent; (DMSO-d6: δ1H
2.50 ppm; CDCl3: δ1H 7.26; δ13C 77.16 ppm). Abbreviations
used are s (singlet), d (doublet), t (triplet), q (quartet), b
(broad), and m (multiplet). DMA experiments were performed
with a TA Instruments Q800 DMA in the “DMA Multi-
Frequency - Strain” mode in air using a temperature ramp
experiment from 0 to 200 ꢀC at a rate of 3 ꢀC/min, a frequency
of 1 Hz, and the single cantilever clamp. Differential scanning
calorimetry (DSC) measurements were obtained on a TA
Instruments Q1000 DSC using hermetically sealed aluminum
pans and sample sizes between 6 and 10 mg. Temperature
ramps of 10 ꢀC/min from 0 to 200 ꢀC with 5 min isothermal
steps at the beginning and end of each ramp were performed
twice, with all DA data taken from the first heating cycle. Prior
to DSC, samples were characterized by thermogravimetric anal-
ysis (TGA) to ensure that the 5% mass loss temperature was
greater than 200 ꢀC. TGA data were acquired with a TA Instru-
ments Q5000 TGA using sample sizes between 5 and 13 mg
and temperature ramps of 10 or 20 ꢀC/min from room tempera-
ture to 600 ꢀC under nitrogen. Size-exclusion chromatography
(SEC) was performed in N,N-dimethylacetamide with 50 mM
LiCl at 50 ꢀC and a flow rate of 1.0 mL/min (Agilent isocratic
pump, degasser, and autosampler; columns: ViscoGel I-series
5 μm guard + two ViscoGel I-series G3078 mixed bed columns,
molecular weight range 0–20 × 103 and 0–100 × 104 g/mol).
Detection consisted of a Wyatt Optilab T-rEX refractive index
detector operating at 658 nm. Apparent molecular weights were
calculated using nine poly(methyl methacrylate) standards from
988,000 g/mol to 2200 g/mol. All thermal data (DMA, DSC, and
TGA) were processed and plotted using TA Universal Analysis
software, and all NMR data were processed and plotted using
MestReNova software. DMA, DSC, and SEC data were replotted
using Origin software.
EXPERIMENTAL
Materials
All chemicals were used as received unless otherwise noted.
Furan (≥ 99%), maleic anhydride (≥ 99.0%), toluene (anhydrous,
99.8%), methanol (HPLC grade), ethanolamine (≥ 98%), benzene
(≥ 99.0%),
4-(dimethylamino)pyridine
(DMAP,
≥ 99%)
methacrylic anhydride (94%), ethyl methacrylate (EMA, 99%),
butyl methacrylate (BMA, 99%), and hexyl methacrylate (HMA,
98%) were purchased from Sigma-Aldrich. Dichloromethane
(DCM), acetone, sodium bicarbonate, ammonium chloride, and
sodium chloride were purchased from Fisher Scientific. Dry DCM
was obtained with an MBraun MB-SPS solvent purification sys-
tem. Furfuryl alcohol (98%), triethylamine (TEA, 99%),
9-anthracenemethanol (AnthMeOH, 98%), and magnesium sul-
fate (anhydrous, ≥ 99.5%) were purchased from Alfa Aesar. Basic
alumina (Brockmann Activity I, 50–200 μm, 60 Å) and methyl
methacrylate (MMA, 99%) were purchased from Acros Organics.
2,2-Dimethoxy-2-phenylacetophenone (DMPA, > 98.0%) and tri-
ethylene glycol dimethacrylate (TEGMA, ≥ 95%) were purchased
from TCI America. Diethyl ether was purchased from EMD
Millipore (Burlington, Massachusetts). N,N-Dimethylformamide
(DMF, ≥ 99.8%) was purchased from VWR. Deuterated chloro-
form (CDCl3, 99.8% D + 0.05% V/V TMS) and dimethyl sulfoxide
2
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2019

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ORIGINAL ARTICLE
Synthesis of Furan-Maleic Anhydride Oxanorbornene
Furan (185 mL, 2.59 mol) and maleic anhydride (50 g,
0.51 mol) were suspended in toluene (500 mL) in a 1 L
round-bottom flask equipped with a stir bar, and the reaction
was stirred for 3 days at room temperature. The precipitated
product was washed with toluene and dried under vacuum
(82.4 g, 97%). 1H NMR (500 MHz, DMSO-d6, δ): 6.58 (s, 2H),
5.35 (s, 2H), 3.31 (s, 2H).
warm to room temperature overnight. The solution was then
diluted with DCM (40 mL) and washed with saturated sodium
bicarbonate (80 mL), ammonium chloride (80 mL), and brine
(80 mL). The organic layer was dried with magnesium sulfate
and was passed through two plugs of basic alumina. The sol-
vent was removed by rotary evaporation and reduced pres-
sure, followed by bubbling through with argon (2.31 g, 74%).
1H NMR (500 MHz, CDCl3, δ): 6.55 (d, J = 5.6 Hz, 1H), 6.44 (d,
J = 5.7 Hz, 1H), 6.12 (s, 1H), 6.07 (s, 1H), 5.59 (s, 1H), 5.56 (s,
1H), 5.26 (s, 1H), 4.99 (d, J = 12.9 Hz, 1H), 4.51 (d,
J = 12.8 Hz, 1H), 4.29 (t, J = 5.3 Hz, 2H), 3.81 (m, 2H), 3.00 (d,
J = 6.5 Hz, 1H), 2.93 (d, J = 6.5 Hz, 1H), 1.95 (s, 3H), 1.90 (s,
3H); 13C NMR (500 MHz, CDCl3, δ): 175.5, 174.0, 167.1, 166.9,
137.6, 137.3, 136.0, 135.9, 126.4, 126.3, 89.8, 81.2, 61.6, 61.0,
50.1, 48.5, 38.1, 18.4, 18.3.
Synthesis of Furan-Protected N-(2-Hydroxyethyl)
Maleimide
Furan-maleic anhydride oxanorbornene (81 g, 0.49 mol) was
suspended in methanol (180 mL) in a 500 mL round-bottom
flask equipped with a stir bar and a reflux condenser. The
reaction was cooled in an ice bath for 30 min. Ethanolamine
(45 mL, 0.75 mol) was added dropwise with stirring. The
resulting mixture was stirred in an ice bath for 1 h and was
warmed to room temperature. The reaction was refluxed
overnight. After cooling to room temperature, the product pre-
cipitated and was recovered by vacuum filtration. The filtrate
was concentrated and stored in a freezer overnight to collect
a second crop of product (72.1 g, 71%). 1H NMR (500 MHz,
DMSO-d6, δ): 6.55 (s, 2H), 5.13 (s, 2H), 4.78 (m, 1H), 3.42 (m,
4H), 2.93 (s, 2H).
Generalized Photopolymerization Procedure
Monomers were passed through plugs of basic alumina to
remove inhibitor prior to use. DMPA (0.5 mol% respective to
total monomer) and crosslinker (DA, TEGMA, or none) were
weighed into a vial. De-inhibited monomer (2 mL total vol-
ume) was added, and the solution was shaken until homoge-
nous. The solution was injected into the assembled cuvette,
and the cuvette was irradiated under the UV lamp for 8 h.
Synthesis of N-(2-Hydroxyethyl)Maleimide
RESULTS AND DISCUSSION
Furan-protected
N-(2-hydroxyethyl)maleimide
(72 g,
DA Crosslinker Design, Synthesis, and Characterization
The design of the DA crosslinker began with the selection of
the diene and dienophile. For the ease of synthesis and conve-
nient cycloaddition and cycloreversion temperatures,37
derivatized furans and maleimides were adopted.
Methacryloyl groups were chosen for the polymerizable func-
tionality to aid in copolymerization with methacrylic mono-
mers, enabling the Tg of the networks to be tuned both widely
due to the broad Tg range of polymethacrylates and in a
0.34 mol) was added to a 1 L round-bottom flask equipped
with a stir bar and a short reflux condenser. Toluene
(430 mL) was added, and the reaction was refluxed overnight.
After cooling to room temperature, the product precipitated
with further cooling in an ice bath and was recovered by vac-
uum filtration. The filtrate was concentrated and stored in a
freezer overnight to collect a second crop of product (36.7 g,
1
76%). H NMR (500 MHz, DMSO-d6, δ): 7.01 (s, 2H), 4.78 (m,
1H), 3.47 (m, 4H).
straightforward
manner
by
simply
changing
the
comonomer(s) used. The synthetic pathway is shown in
Scheme 1. After the forward DA reaction of furan and maleic
Synthesis of DA Diol
N-(2-hydroxyethyl)maleimide (7.0 g, 49.6 mmol) was dis-
solved in benzene (75 mL) in a 250 mL round-bottom flask
equipped with a stir bar and a reflux condenser. Freshly dis-
tilled furfuryl alcohol (4.86 g, 49.6 mmol) was added, and the
solution was refluxed overnight. The product, which precipi-
tated upon cooling, was washed with diethyl ether and dried
under reduced pressure (7.4 g, 62%). 1H NMR (500 MHz,
DMSO-d6, δ): 6.52 (m, 2H), 5.08 (m, 1H), 4.94 (dd, J = 6.1,
5.5 Hz, 1H), 4.77 (m, 1H), 4.04 (dd, J = 12.5, 6.1 Hz, 1H), 3.69
(dd, J = 12.5, 5.5 Hz, 1H), 3.41 (m, 4H), 3.04 (d, J = 6.4 Hz,
1H), 2.88 (d, J = 6.4 Hz, 1H).
Synthesis of DA Dimethacrylate Crosslinker
To a 100 mL round-bottom flask equipped with a magnetic
stir bar and an addition funnel was added DA diol (2.0 g,
8.4 mmol), TEA (7 mL), DMAP (407 mg), and dry DCM
(40 mL). The reaction was cooled in an ice bath. The addition
funnel was charged with methacrylic anhydride (5.0 mL,
33.5 mmol), which was added dropwise over a period of
30 min. The resulting mixture was stirred and allowed to
SCHEME 1 Synthesis of DA dimethacrylate crosslinker starting
from furan and maleic anhydride.
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2019
3

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
JOURNAL OF
POLYMER SCIENCE
ORIGINAL ARTICLE
WWW.POLYMERCHEMISTRY.ORG
anhydride, the resulting oxanorbornene was converted to the
imide using ethanolamine. The hydroxyethyl maleimide was
obtained after rDA deprotection, liberating furan. This
maleimide was then reacted with furfuryl alcohol to give the
DA diol, which was converted to the dimethacrylate with
methacryloyl anhydride.
employed in this study. Initially, methyl methacrylate (MMA,
Tg: 105 ꢀC49) and hexyl methacrylate (HMA, Tg: −5 ꢀC49) were
selected, but the low Tg of HMA made mechanical characteri-
zation of uncrosslinked poly(hexyl methacrylate) (PHMA) dif-
ficult. Therefore, butyl methacrylate (BMA, Tg: 20 ꢀC49) was
chosen as a replacement because it is the next lowest Tg com-
mercially available n-alkyl methacrylate.
The syntheses of the DA diol and intermediates were per-
formed according to a modified literature procedure,44 and
the conditions for the preparation of the dimethacrylate were
adapted from a published precedent.45 Along with two
patents,46,47 this is one of the first reported preparations and
uses of this dimethacrylate DA crosslinker. All syntheses were
straightforward, scalable, and afforded good yields after opti-
mization. The workups were facile, as the pure products pre-
cipitated from the crude reaction mixtures. The final step
required acidic and basic washes to remove excess
triethylamine (TEA), 4-(dimethylamino)pyridine (DMAP), and
methacrylic acid, respectively. The 1H NMR spectra for the
furan-maleic anhydride oxanorbornene, furan-protected
hydroxyethyl maleimide, deprotected hydroxyethyl maleimide,
DA diol, and DA dimethacrylate crosslinker can be found in
the Supporting Information as Figures S1–S5, respectively.
The 13C NMR spectrum of the DA dimethacrylate crosslinker
is shown in Figure S6.
For the polymerization, a conventional free-radical approach
was taken to make this chemistry readily applicable to larger
scale processes. A thick film photopolymerization strategy was
utilized to avoid potential risks associated with conventional
thermal initiation, namely premature crosslinker cleavage at
elevated cure temperatures and voids in the film due to the
nitrogen byproduct of azo initiator decomposition. Bulk condi-
tions were chosen to aid in film preparation by eliminating the
need for solvent removal. The films were prepared in cuvettes
consisting of glass microscope slides and silicone rubber sheets
cut to form rectangular gaskets. Solutions consisting of initiator
and crosslinker dissolved in monomer were injected into the
cuvettes through the spacer, and the cuvettes were placed
under an inexpensive commercial UV lamp to cure. This
approach allowed for the preparation of uniformly thick films,
and film recovery was facilitated by the ease of disassembly of
the cuvette. Representative photographs of the procedure are
displayed in the Supporting Information as Figure S7.
Network Design, Synthesis, and Characterization
Since DA networks gradually change their mechanical behav-
ior from thermosetting to thermoplastic with increasing tem-
perature, thermoset and thermoplastic controls were
prepared as model examples of the respective behaviors as
illustrated in Scheme 2. Triethylene glycol dimethacrylate
(TEGMA) was used as a permanent crosslinker for the ther-
moset control films, and crosslinker was omitted to fabricate
the thermoplastic controls. A series of 15 samples was pre-
pared consisting of five different monomer molar feed ratios
(100:0, 75:25, 50:50, 25:75, and 0:100 BMA:MMA), and each
feed ratio was either crosslinked with the DA crosslinker for
the thermally reversible films or TEGMA for the thermoset
controls, or kept uncrosslinked for the thermoplastic controls.
Unless otherwise specified, the amount of crosslinker used
(DA or TEGMA) was 1 mol% relative to total monomer.
To isolate the effect of Tg on the change in mechanical proper-
ties of DA crosslinked networks at elevated temperatures
brought about by the rDA reaction, a methodology for prepar-
ing a series of polymers of varying Tg without introducing
additional variables was necessary. We settled on the copoly-
merization of two monomers at different feed ratios that
when individually polymerized afforded polymers of contra-
sting Tg. Gradually varying the ratio of one monomer to the
other allowed for the network Tg to be tuned in a systematic
fashion. To prevent inconsistencies across the series that may
result from copolymerizing two monomers with drastically
different structures, n-alkyl methacrylates were selected, as
this allowed facile control over Tg via altering the length of
the alkyl chain. While longer lengths generally impart lower
Tg, polymethacrylates begin to exhibit side-chain crystallinity
with alkyl chain lengths between 10 and 12 carbon atoms.48
Therefore, only short-chain n-alkyl methacrylates were
Mechanical Characterization
Assessment of the mechanical properties of the films with
increasing temperature was performed by DMA. The storage
moduli of the TEGMA-crosslinked and uncrosslinked controls
as a function of temperature from 0 to 200 ꢀC are shown in
Figure 1. Both controls, uncrosslinked and permanently
crosslinked, display behavior typical of thermoplastics and
thermosets, respectively.50 All formulations exhibit relatively
high glassy moduli, which increase with increasing MMA con-
tent. As the temperature increases, the samples undergo the
glass transition, which shifts to higher temperatures at higher
MMA concentrations. After this, the TEGMA-crosslinked sam-
ples level off and enter the rubbery plateau regime. The stor-
age moduli then remain constant until the end of the run. The
uncrosslinked samples, on the other hand, yield shortly after
SCHEME
2 Overview of polymerization approach including
structures of crosslinkers used in this study.
4
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2019