Mechanistic Study of Stress Relaxation in Urethane-Containing Polymer Networks

Article abstract of DOI:10.1021/acs.jpcb.8b11489

Cross-linked polymers are used in many commercial products and are traditionally incapable of recycling via melt reprocessing. Recently, tough and reprocessable cross-linked polymers have been realized by incorporating cross-links that undergo associative exchange reactions, such as transesterification, at elevated temperatures. Here we investigate how cross-linked polymers containing urethane linkages relax stress under similar conditions, which enables their reprocessing. Materials based on hydroxyl-terminated star-shaped poly(ethylene oxide) and poly((±)-lactide) were cross-linked with methylene diphenyldiisocyanate in the presence of stannous octoate catalyst. Polymers with lower plateau moduli exhibit faster rates of relaxation. Reactions of model urethanes suggest that exchange occurs through the tin-mediated exchange of the urethanes that does not require free hydroxyl groups. Furthermore, samples were incapable of elevated-temperature dissolution in a low-polarity solvent (1,2,4-trichlorobenzene) but readily dissolved in a high-polarity aprotic solvent (DMSO, 24 to 48 h). These findings indicate that urethane linkages, which are straightforward to incorporate, impart dynamic character to polymer networks of diverse chemical composition, likely through a urethane reversion mechanism.

Full text of DOI:10.1021/acs.jpcb.8b11489

Article
pubs.acs.org/JPCB
Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX
Mechanistic Study of Stress Relaxation in Urethane-Containing
Polymer Networks
Jacob P. Brutman,† David J. Fortman,‡,§ Guilhem X. De Hoe,† William R. Dichtel,*,‡
and Marc A. Hillmyer*,†
†Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455-0431, United States
‡Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States
§Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University, Ithaca, New York 14853, United States
S
* Supporting Information
ABSTRACT: Cross-linked polymers are used in many
commercial products and are traditionally incapable of
recycling via melt reprocessing. Recently, tough and
reprocessable cross-linked polymers have been realized by
incorporating cross-links that undergo associative exchange
reactions, such as transesterification, at elevated temperatures.
Here we investigate how cross-linked polymers containing
urethane linkages relax stress under similar conditions, which
enables their reprocessing. Materials based on hydroxyl-
terminated star-shaped poly(ethylene oxide) and poly(( )-
lactide) were cross-linked with methylene diphenyldiisocya-
nate in the presence of stannous octoate catalyst. Polymers
with lower plateau moduli exhibit faster rates of relaxation. Reactions of model urethanes suggest that exchange occurs through
the tin-mediated exchange of the urethanes that does not require free hydroxyl groups. Furthermore, samples were incapable of
elevated-temperature dissolution in a low-polarity solvent (1,2,4-trichlorobenzene) but readily dissolved in a high-polarity
aprotic solvent (DMSO, 24 to 48 h). These findings indicate that urethane linkages, which are straightforward to incorporate,
impart dynamic character to polymer networks of diverse chemical composition, likely through a urethane reversion mechanism.
of vitrimers is gradual and follows an Arrhenius relationship,
which differs from the typical Williams−Landel−Ferry response
observed for thermoplastics. This strong glass-forming behavior
confers a topology freezing transition temperature (Tv) when it
occurs above or at the Tg. Below Tv, the vitrimer behaves as a
traditional thermoset, and above Tv, the cross-links undergo
dynamic exchange reactions that give rise to thermally activated
stress relaxation.
Vitrimers and vitrimer-like materials have been developed on
the basis of dynamic chemistry including alkene metathesis,15,16
hindered urea exchange,17,18 disulfide/polysulfide metathe-
sis,19−22 thiol-disulfide exchange,23 vinylogous urethane ex-
change,24,25 transcarbamoylation,26,27 siloxane equilibrium,28,29
and boronic ester exchange.30−32 We previously reported
reprocessable materials based on poly(( )-lactide) (PLA)
cross-linked with a diisocyanate in the presence of a stannous
octoate catalyst.33 We initially hypothesized that transester-
ification caused stress relaxation (Scheme 1A) and that the high
concentration of ester functionalities in the PLA backbone was
responsible for the rapid stress relaxation and efficient
INTRODUCTION
■
Cross-linked polymer networks are prevalent in adhesives,
composites, and other durable products. Unlike thermoplastics,
traditional cross-linked polymer networks are effectively
irreparable, cannot be reshaped, and are nonrecyclable through
traditional means. Many studies have explored their reprocess-
ing by incorporating dynamic covalent cross-links.1−6 Early
examples relied on thermally reversible moieties such as Diels−
Alder cycloadducts, which allowed the materials to depoly-
merize and be reformed in a new shape.7,8 Although this
approach represented a significant conceptual advance in the
reprocessing of cross-linked polymers, the materials often
exhibit poor thermal stability and solvent resistance, which
limit their potential applications.9
A new class of reprocessable polymer networks, termed
vitrimers,10−12 feature cross-links that typically undergo
exchange through associative rather than dissociative processes.
Vitrimers maintain their cross-linked nature at elevated
temperature and/or in the presence of solvents, as demonstrated
for polyester epoxy resins containing a Zn(II) transesterification
catalyst.10,13,14 The cured resins were capable of reprocessing
and injection molding at 280 °C; however, they only swell in hot
solvent (180 °C) rather than dissolving, as is typical for cross-
linked polymers. The temperature dependence of the viscosity
Received: November 28, 2018
Revised: January 15, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
with urethane reversion being the predominant mechanism of
stress relaxation in polymer networks cross-linked by N-aryl
urethanes.
Scheme 1. Proposed Possible Relaxation Mechanisms for
Urethane Cross-Linked PLA
a
EXPERIMENTAL SECTION
■
Materials. All reagents were purchased from Sigma-Aldrich
(Milwaukee, WI) and were used as received unless otherwise
stated. All glassware was held at 105 °C overnight prior to use
unless otherwise specified. ( )-Lactide was kindly provided by
Altasorb (Piedmont, SC) and used as received. Stannous octoate
[Sn(Oct)2] was purified by vacuum distillation (three times,
∼130−150 °C, ca. 30−50 mTorr argon). Dichloromethane
(DCM) and methanol were purchased from Fisher Scientific
(Hampton, NH). DCM was purified via a GC-SPS-4-CM glass
contour 800-L solvent purification system obtained from Pure
Process Technologies (Nashua, NH). Four-arm hydroxyl-
terminated PLA and urethane cross-linked PLA were synthe-
sized using a literature procedure; the number-average molar
a
(A) Transesterification, (B) transcarbamoylation, and (C) urethane
reversion.
reprocessability of these materials. However, the Arrhenius
activation energy (Ea) for stress relaxation (150 kJ mol−1) was
much higher than that determined for PLA transesterification
(80 kJ mol−1),34 which indicated that other stress relaxation
mechanisms were more likely. We later showed that cross-linked
polyhydroxyurethanes relax stress through hydroxyl−urethane
exchange (e.g., transcarbamoylation, Scheme 1B),26,27 suggest-
ing that the urethane linkages might also be responsible for stress
relaxation in the PLA networks. Furthermore, stress relaxation of
polyurethane networks lacking free hydroxyl groups, which are
required for transcarbamoylation, was reported by Tobolsky in
the 1950s and was hypothesized to occur via urethane reversion
(Scheme 1C).35,36 Exploiting the dynamic nature of urethane
bonds has recently become a promising approach to reshaping
polyurethane networks, although mechanistic aspects of these
processes are not entirely clear. For example, Zheng et al.
hypothesized that urethanes were responsible for stress
relaxation in polyurethane elastomers in the presence of
dibutyltin dilaurate.37 Yan et al. further studied the tin-mediated
relaxation of polyurethanes and showed preliminary results
indicating that this behavior can be correlated to the
reprocessability of cross-linked polyurethanes.38,39 Additional
work by Zheng et al. demonstrated that networks based on N-
aryl urethanes relax stress at elevated temperatures even in the
absence of an external catalyst,40 although no explicit evidence
for urethane reversion was presented in any of these systems.
Yang and Urban postulated that the repair of cross-linked
polyurethanes occurs through the generation of amines upon
mechanical failure, which react further with urethanes to form
ureas.41 Although modest changes in the Raman and IR spectra
were consistent with this hypothesis, more definitive character-
ization was not obtained. Considering these findings, a robust
understanding of the mechanisms of stress relaxation for
urethane-containing polymer networks has not yet been
established.
1
mass (Mn) of the prepolymer was determined by H NMR
spectroscopy.33 Urethane cross-linked PEO was synthesized
under the same conditions as for the PLA samples using
commercially available pentaerythritol ethoxylate (Mn ≈ 797 g
mol−1). PLA and PEO samples are respectively denoted as PLA-
X-Y and PEO-X-Y, where X is the Mn of the prepolymer (kg
mol−1) and Y is the cross-linker used: either methylene diphenyl
diisocyanate (MDI) or poly(methylene diphenyl diisocyanate)
(PMDI). In some cases, a sample was swollen in methanol
(details given in the Characterization Methods section) or
contained no catalyst, denoted by SM and NC at the end of the
sample name, respectively.
Synthesis of Epoxide Cross-Linked Poly(4-methylcap-
rolactone). The 4-methylcaprolactone monomer was synthe-
sized using a literature procedure.42 Sn(Oct)2 (4 mg, 0.025 mol
%) was dissolved in toluene (ca. 0.1 mL) and charged in a
pressure vessel, along with 4-methylcaprolactone (5 g, 39 mmol)
and pentaerythritol (0.073 g, 0.54 mmol). The reaction mixture
was held at 160 °C for 3 h, and then succinic anhydride (0.3 g, 3
mmol) was added under N2 and stirred for 1 h. The mixture was
cooled, dissolved using DCM, and then precipitated into
methanol (ca. 10 times the volume of the product solution). The
crude polymer was again dissolved in DCM and reprecipitated
in hexanes (ca. 10 times the volume of the product solution).
The resulting carboxylic acid-terminated poly(4-methylcapro-
lactone) (P4MCL) was dried under N2 for 24 h and then dried
under vacuum (20 mTorr) at 60 °C for 72 h; the isolated yield
was 76%. 1H NMR (500 MHz, CDCl3; 25 °C) δ 4.20−4.01 (m,
158 H), 2.64 (m, 16 H), 2.42−2.23 (m, 150 H), 1.74−1.39 (m,
390 H), 0.97−0.87 (m, 240 H). Mn = 10.1 kg mol−1. DSC: Tg =
−60 °C.
P4MCL (1.52 g, 1.0 equiv of COOH groups), triglycidyl
isocyanurate (64 mg, 1.0 equiv of epoxide groups), and
Sn(Oct)2 (6 mg, 2.5 mol % with respect to COOH groups)
were dissolved in DCM. The solution was poured into a
polypropylene container and allowed to sit for 24 h before being
heated under N2 at 120 °C for 3 h. The gel percent of the
resultant epoxide cross-linked P4MCL network was 95%.
Representative Synthesis of N−H Model Compounds.
To a flame-dried round-bottomed flask under a nitrogen
atmosphere was added alcohol (16.8 mmol) and anhydrous
tetrahydrofuran (20 mL). A solution of Sn(Oct)2 (130 mg, 0.34
mmol, 2 mol %) dissolved in anhydrous tetrahydrofuran (1 mL)
was added, followed by the addition of isocyanate (16.8 mmol)
using a syringe. The resulting solution was stirred at room
Here we investigate the thermally activated stress relaxation of
urethane-containing polymer networks as a function of temper-
ature, catalyst content, and polymer structure. The similarity of
Ea values for the stress relaxation of PLA- and poly(ethylene
oxide) (PEO)-based polymers cross-linked with aryl isocyanates
indicates that urethane bonds are the dynamic linkages in both
networks and further suggests that transesterification-based
relaxation is negligible in PLA-derived urethane networks.
Exchange studies of urethane model compounds at elevated
temperature provide further insight into the mechanism
responsible for stress relaxation. Variable-temperature NMR
spectroscopy and polymer swelling experiments are consistent
B
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
temperature for 24 h, and solvent was removed at reduced
pressure to yield a white solid. The crude solid was purified by
chromatography on silica gel using 20% ethyl acetate/hexanes to
yield the product.
g, 11.7 mmol), and triphenylmethane (27.0 mg, 0.11 mmol) as
an internal standard. The resulting mixture was heated and
stirred in a preheated oil bath. Aliquots were removed using a
syringe at various time points, diluted with DCM, and subjected
to GC−MS analysis. Concentrations of the transcarbamoylation
product, N-phenyl-O-decyl urethane, could not be determined
quantitatively because of the partial decomposition of the
urethanes on the GC column at temperatures required to
sufficiently volatilize them.
1
N-Phenyl-O-octyl Urethane. White solid, 82% yield. H
NMR (400 MHz, CDCl3) δ 7.38 (d, J = 8.0 Hz, 2H), 7.35−7.23
(m, 2H), 7.05 (tt, J = 7.1, 1.2 Hz, 1H), 6.56 (br s, 1H), 4.16 (t, J
= 6.7 Hz, 2H), 1.71−1.63 (m, 2H), 1.43−1.23 (m, 10H), 0.89
(t, J = 7.8 Hz, 3H). 13C NMR (100 MHz, CDCl3) δ 153.7, 138.0,
128.9, 123.2, 118.5, 65.4, 31.7, 29.19, 29.15, 28.9, 25.8, 22.6,
14.0. IR (neat, ATR) 3304, 2956, 2920, 2853, 1698, 1599, 1544,
1444, 1236, 1055, 747 cm−1.
Model Alcohol−Urethane Exchange Reaction Ana-
lyzed by NMR Spectroscopy. To a vial was added Sn(Oct)2
(11.6 mg, 0.03 mmol, 2.5 mol % with respect to urethane), N-
tolyl-O-(triethyleneglycol monomethyl ether) urethane (340
mg, 1.14 mmol), and 1-decanol (1.81 g, 11.4 mmol). The
resulting mixture was then heated and stirred in a preheated oil
bath. Aliquots of ca. 10 mg were removed using a syringe at
various time points, dissolved in CDCl3 (containing 10.0 mg/
mL tribromobenzene as an external standard) to a concentration
N-Tolyl-O-decyl Urethane. White solid, 85% yield. 1H NMR
(400 MHz, CDCl3) δ 7.28−7.22 (m, 2H), 7.10 (d, J = 8.3 Hz,
2H), 6.49 (br s, 1H), 4.14 (t, J = 6.7 Hz, 2H), 2.30 (s, 3H),
1.72−1.60 (m, 2H), 1.44−1.24 (m, 14H), 0.92−0.84 (m, 3H).
13C NMR (100 MHz, CDCl3) δ 153.8, 135.4, 132.8, 129.4,
118.7, 65.3, 31.9, 29.51, 29.50, 29.27, 29.25, 28.9, 25.8, 22.6,
20.7, 14.1. IR (neat, ATR) 3327, 2919, 2851, 1696, 1596, 1531,
1314, 1235, 1071, 814 cm−1.
1
of 50.0 mg/mL, and analyzed by H NMR spectroscopy (the
−CH2O− peaks of the starting material and transcarbamoyla-
tion product are distinct). Concentrations of the trans-
carbamoylation product, N-tolyl-O-decyl urethane, could not
be determined quantitatively because of the overlap of the
product resonance at ca. 4.1 ppm with a side product. (The
calculated amount of product formed is greater than the amount
of starting material lost. Running the reaction to high conversion
clearly shows an overlapping resonance convoluting the product
peak.)
Model Urethane−Urethane Exchange Reaction. To a
vial was added Sn(Oct)2 (23.9 mg, 0.059 mmol, 2.5 mol % with
respect to urethane), N-phenyl-O-octyl urethane (294 mg, 1.18
mmol), N-tolyl-O-decyl urethane (344 mg, 1.18 mmol), and
triphenylmethane (30.5 mg, 0.12 mmol) as an internal standard.
The resulting mixture was heated and stirred in an oil bath
preheated to the desired temperature. Aliquots were removed
using a syringe at various time points, diluted with DCM, and
subjected to GC−MS analysis. Concentrations of the urethane
exchange product, N-phenyl-O-decyl urethane, could not be
determined quantitatively because of the partial decomposition
of the urethanes on the GC column at temperatures required to
sufficiently volatilize them.
N-Tolyl-O-(triethyleneglycol monomethyl ether) Ure-
1
thane. Colorless oil, 57% yield. H NMR (400 MHz, CDCl3)
δ 7.25 (d, J = 8.6 Hz, 2H), 7.10 (d, J = 8.0 Hz, 2H), 6.73 (br s,
1H), 4.35−4.28 (m, 2H), 3.78−3.71 (m, 2H), 3.73−3.62 (m,
6H), 3.59−3.52 (m, 2H), 3.38 (s, 3H), 2.30 (s, 3H). 13C NMR
(100 MHz, CDCl3) δ 153.5, 135.3, 132.7, 129.3, 118.7, 71.8,
70.42, 70.41, 70.40, 69.3, 63.9, 58.8, 20.6. IR (neat, ATR) 3306,
2873, 1727/1709, 1599, 1530, 1315, 1222, 1207, 1102, 1069,
816 cm−1.
Representative Synthesis of N-CH3 Model Com-
pounds. To a flame-dried round-bottomed flask under a
nitrogen atmosphere was added urethane model compound (4
mmol) and anhydrous dimethylformamide (15 mL). The
mixture was cooled in an ice bath, and sodium hydride (192
mg, 8.0 mmol, 320 mg dispersion in mineral oil) was added,
resulting in gas evolution. The resulting mixture was stirred for
10 min, and then iodomethane (1.419 g, 10 mmol, 0.62 mL) was
added using a syringe. The resulting mixture was stirred at room
temperature for 18 h and then diluted with water (150 mL). This
solution was extracted with diethyl ether (200 mL), which was
washed with water (2 × 150 mL), dried over MgSO4, and
filtered. Solvent was removed at reduced pressure to yield a
colorless oil. The crude oil was purified via chromatography on
silica gel using 10% ethyl acetate/hexanes to yield the product.
N-Methyl-N-phenyl-O-octyl Urethane. Colorless oil, 65%
yield. 1H NMR (400 MHz, CDCl3) δ 7.34 (m, 2H), 7.27−7.15
(m, 3H), 4.09 (t, J = 6.7 Hz, 2H), 3.30 (s, 3H), 1.63−1.53 (m,
2H), 1.33−1.22 (m, 10H), 0.88 (t, J = 6.8 Hz, 3H). 13C NMR
(100 MHz, CDCl3) δ 155.7, 143.4, 128.7, 125.8, 125.6, 65.8,
37.5, 31.7, 29.11, 29.08, 28.8, 25.8, 22.6, 14.0. IR (neat, ATR)
2925, 2855, 1703, 1598, 1498, 1346, 1154, 695 cm−1.
Synthesis of the Diethyl Urethane Adduct of MDI. To a
round-bottomed flask was added MDI (10 g, 40 mmol), ethanol
(9.3 mL, 160 mmol), and DCM (10 mL) with stirring. Once the
solution was homogeneous, it was cooled to 0 °C and Sn(Oct)2
(0.81 g, 2 mmol, 2.5 mol % with respect to NCO groups) in
DCM (10 mL) was added. After 10 min at 0 °C, the solution was
left at room temperature for 20 h before concentrating by rotary
evaporation and drying under high vacuum (∼20 mTorr) for 48
h. The crude product was then dissolved in dimethylformamide
(ca. 20 mL) and precipitated into deionized water (ca. 200 mL),
yielding a yellow solid that was dried under vacuum (20 mTorr)
for 48 h (quantitative yield). 1H NMR (500 MHz, DMSO-d6) δ
9.52 (s, 2H), 7.38 (d, 4H), 7.10 (d, 4H), 4.12 (q, 4H), 3.79 (s,
2H), 1.24 (t, 6H). 13C NMR (125 MHz, DMSO-d6) δ 154, 138,
136, 129, 119, 60, 40, 15.
N-Methyl-N-tolyl-O-decyl Urethane. Colorless oil, 55%
1
yield. H NMR (400 MHz, CDCl3) δ 7.18−7.07 (m, 4H),
4.08 (t, J = 6.7 Hz, 2H), 3.27 (s, 3H), 2.34 (s, 3H), 1.64−1.54
(m, 2H), 1.35−1.23 (m, 14H), 0.89 (t, J = 6.9 Hz, 3H). 13C
NMR (100 MHz, CDCl3) δ 155.9, 140.8, 135.6, 129.3, 125.6,
65.8, 37.7, 31.9, 29.5, 29.4, 29.3, 29.1, 28.9, 25.8, 22.6, 20.9, 14.0.
IR (neat, ATR) 2923, 2854, 1703, 1515, 1343, 1155, 1111, 820,
768 cm−1.
Characterization Methods.26,33,43 NMR spectroscopy
was performed on a 500 MHz Bruker Avance III HD with a
SampleXpress spectrometer (Billerica, MA) or a 400 MHz
Agilent DD MR-400 spectrometer. Solutions were prepared in
99.8% CDCl3 (Cambridge Isotope Laboratories). All NMR
spectra were acquired at 20 °C with at least 16 scans and a 1 s
delay unless otherwise specified. Chemical shifts are reported in
Model Alcohol−Urethane Exchange Reaction Ana-
lyzed by GC−MS Analysis. To a vial was added Sn(Oct)2
(11.9 mg, 0.03 mmol, 2.5 mol % with respect to urethane), N-
phenyl-O-octyl urethane (293 mg, 1.17 mmol), 1-decanol (1.86
C
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
a
Scheme 2. Synthesis of Urethane Cross-Linked PEOs and PLAs
a
The bracketed representation for MDI and PMDI is used to show that PMDI is a mixture of regioisomers and oligomers with an average
functionality of 3.2 isocyanate groups and an average Mn ≈ 400 g/mol.
ppm with respect to residual CHCl3 (7.26 ppm). Variable-
temperature (VT) NMR was performed on a 500 MHz Bruker
III at 100 and 140 °C. DMSO-d6 (Cambridge Isotope
Laboratories, 99.9%) was purified by distillation over CaH2.
The solution for VT-NMR was prepared and sealed in a high-
pressure NMR tube under N2. The same solution was allowed to
equilibrate at the desired temperature for 10 min before the 1H
NMR and 13C NMR spectra were acquired.
FTIR was performed on a Bruker Alpha Platinum with a
single-reflection diamond ATR head or a Thermo Nicolet iS10
equipped with a ZnSe ATR attachment. Spectra were obtained
from 400 to 4000 cm−1 using a minimum of 16 scans. For each
spectrum, the transmission intensity data were normalized with
respect to the carbonyl stretch.
where G′ is the storage modulus under shear, R is the universal
gas constant, T is the absolute temperature in the rubbery region
(ca. 373 K), and ρ is the density of PLA (1.25 g cm−3) or PEO
(1.13 g cm−3).
The stress−relaxation analysis (SRA) experiments were
performed under strain control at a specified temperature
(110−170 °C depending on the sample). The samples were
allowed to equilibrate at a chosen temperature for approximately
10 min, after which the axial force was then adjusted to 0 N with
a sensitivity of 0.05 N. Each sample was then subjected to an
instantaneous 5% strain. The stress decay was monitored while
maintaining a constant strain (5%) until the stress relaxation
modulus had relaxed to 37% (1/e) of its initial value. The
characteristic relaxation time (τ*), or the time required for the
modulus to reach 37% (1/e) of its initial value, was measured
three times in succession for each sample at each temperature.
These points were then plotted versus 1/T and fit to the
Arrhenius relationship in eq 2.
Gas chromatography−electron impact mass spectrometry
(GC−MS) was performed on an Agilent 6890N Network GC
system with a JEOL JMS-GCmate II mass spectrometer
(magnetic sector). Triphenylmethane was used as an internal
standard.
Ea/RT
*
τ (T) = τ0e
(2)
Dynamic mechanical thermal analysis (DMTA) was
performed on a TA Instruments RSA-G2 analyzer (New Castle,
DE) using dog-bone-shaped films (ca. 0.5 mm (T) × 3 mm (W)
× 25 mm (L) and a gauge length of 14 mm). DMTA
experiments were conducted in tension film mode, where the
axial force was first adjusted to a tension of 0.2 N (sensitivity of
0.01 N) to ensure no buckling of the sample. The proportional
force mode was set to force tracking to ensure that the axial force
was at least 100% greater than the dynamic oscillatory force. The
strain adjust was then set to 30% with a minimum strain of
0.05%, a maximum strain of 5%, and a maximum force of 0.2 N
to prevent the sample from going out of the specified strain
range. A temperature ramp was then performed from −50 to 200
°C at a rate of 5 °C min−1, with an oscillating strain of 0.05% and
an angular frequency of 6.28 rad s−1. PLA samples required a
higher starting temperature (25 to 35 °C) because of transducer
overload in the glassy regime. The glass-transition temperature
(Tg) was determined from the maximum value of the loss
modulus. The cross-link density (νe) and the molar mass
between cross-links (Mx) were calculated using the storage
modulus (E′) at 100 °C with eq 1.
where τ0 is the characteristic relaxation time at infinite T, Ea is
the activation energy of stress relaxation (kJ mol−1), R is the
universal gas constant, and T is the temperature in K at which
SRA was performed.
Tv is defined as the point at which a material exhibits a
viscosity of 1012 Pa s, also known as the liquid-to-solid transition
viscosity (ηv).44−46 Using Maxwell’s relation (eq 3)47 and the E′
measured by DMTA at 100 °C, the τ* at Tv (τv*) was determined
for each sample. The Arrhenius fit for each sample was then
extrapolated to the corresponding τ*v to determine Tv.
1
3
*
ηv = E′τv
(3)
Differential scanning calorimetry (DSC) was conducted on a
TA Instruments Discovery DSC (New Castle, DE). The
instrument was calibrated using an indium standard. All samples
(ca. 3−7 mg) were analyzed using T-Zero hermetic pans under a
N2 purge of 50 mL min−1. The samples were initially cooled to
−80 °C and then heated to 150 °C at 10 °C min−1. After a 1 min
isotherm, the samples were cooled to −80 °C at 10 °C min−1 and
heated again to 150 °C at the same rate. Values for Tg were
acquired at the midpoint of each transition in the second heating
curve using Trios software. Thermogravimetric analysis (TGA)
was performed on a TA Instruments Q500 (New Castle, DE)
3ρRT
Mx
E′(T) = 3G′(T) = 3RTν =
e
(1)
D
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
under air at a heating rate of 10 °C/min from room temperature
to 550 °C. A typical sample size was between 8 and 15 mg.
Solvent extraction experiments were performed by placing a
small amount of cross-linked polymer (ca. 20−100 mg) into a 20
mL vial filled with DCM or MeOH. The vial was then closed and
stirred for 48 h before removing the solvent by gravity filtration.
The recovered sample was dried under reduced pressure for 48 h
at 20 mTorr, after which the sample was weighed and the gel
percent was determined. A high-temperature swell test was also
performed with PEO-0.8-MDI (with and without catalyst)
submerged in 1,2,4-trichlorobenzene (TCB) or anhydrous
dimethylsulfoxide (DMSO) at 140 °C for 7 days or until full
dissolution occurred.
Trace Sn analysis was performed by Chemical Solutions Ltd.
using inductively coupled plasma−mass spectrometry (ICP−
MS) after microwave digestion of the samples. Five samples
were subjected to trace Sn analysis: PEO-0.8-PMDI (0.44 wt %
Sn), PEO-0.8-MDI (0.54 wt % Sn), PEO-0.8-PMDI-SM (0.96
wt % Sn), PEO-0.8-MDI-SM (0.52 wt % Sn), and PEO-0.8-
MDI-NC (0.000053 wt % Sn).
Figure 1. DMTA of urethane cross-linked PEO and PLA samples. The
dashed lines indicate samples soaked in methanol (SM) and dried
under vacuum, and the dotted line indicates the control sample with no
catalyst (NC). The analysis was run at 1 Hz with an oscillation strain of
0.05%. PLA samples were tested below 150 °C to avoid thermal
decomposition and at higher initial temperatures to avoid transducer
overload in the glassy regime.
RESULTS AND DISCUSSION
temperature. Varying the Mn of the PLA prepolymer affected the
plateau modulus as expected: higher Mn prepolymers yielded
less densely cross-linked materials and thus lower plateau
moduli. We attributed the similar plateau moduli of PLA-3.8-
MDI and PLA-10-MDI to the presence of trapped entangle-
ments, which became more prevalent as the prepolymer Mn was
increased (Figure 1, Table 1). The reported molar mass between
entanglements (Me) for PLA is 4 kg mol−1, and the critical molar
mass (Mc) at which entanglements are experimentally observed
for PLA is 9 kg mol−1,48 which is consistent with the Mx of 8.9 kg
mol−1 determined for these samples.
SRA of the PEO-based materials revealed stress relaxation
similar to that of PLA-based materials (Figure 2A), suggesting
that urethanes, not esters, are the dominant functional groups
contributing to stress relaxation. These findings are also
consistent with the remarkably slow stress relaxation observed
for an aliphatic polyester network containing Sn(Oct)2 but no
urethanes (epoxide cross-linked poly(4-methylcaprolactone),
Figure S2). Furthermore, the lack of significant transester-
ification observed in poly(4-methylcaprolactone) and PLA-
based materials is in agreement with a recently reported model
compound study using Sn(Oct)2: transesterification between
benzyl alcohol and a β-lactone was rapid at 120 °C, but no
transesterification of the resultant unstrained products was
observed after 24 h.42 The characteristic relaxation times (τ*)
for PEO- and PLA-based materials at various temperatures were
fitted to an Arrhenius model (eq 2), from which an Ea of stress
relaxation was extracted (Figure 2B, Table 1). The Ea did not
vary significantly between the PEO- and PLA-based materials
containing Sn(Oct)2 (the range was 139−165 kJ mol−1,
consistent with our previous work), providing further evidence
that the urethane functionality dominates the stress relaxation
behavior in all cases. FTIR spectra of all materials before and
after SRA are indistinguishable, indicating that the functional
groups in the networks are stable at the elevated temperatures
employed in this study (Figures S3−S5). Explicitly, we see no
formation of common byproducts seen in polyurethanes, such as
allophanates, ureas, biurets, and isocyanurates.
■
In our previous study of urethane cross-linked PLA,33 the Ea of
stress relaxation (150 kJ mol−1) was far higher than that reported
for Sn(Oct)2-catalyzed transesterification in a PLA melt (80 kJ
mol−1).34 To further investigate, we substituted the PLA
component with PEO; if transesterification reactions are the
dominant mechanism of stress relaxation, then PEO-based
materials would exhibit distinctly different stress relaxation
behavior as a result of the lack of ester linkages in PEO. Urethane
cross-linked PEO materials were prepared using conditions
similar to those used for urethane cross-linked PLA (Scheme
2).33 The commercially available PEO prepolymer was
combined with MDI (NCO functionality of 2, 0.75 NCO
group per OH group) and Sn(Oct)2 (2.5 mol % relative to initial
OH groups) to afford cross-linked networks. The cross-link
density of the PEO-based materials was varied by using PMDI
(average NCO functionality of 3.2, 0.75 NCO group per OH
group) as an alternate cross-linker. The NCO/OH ratio was
chosen in order to maintain consistency between this study and
our previous study.33 To directly compare these materials with
PLA-based polymers, we prepared hydroxyl-terminated, star-
shaped PLA prepolymers with Mn = 1.0, 3.8, and 10 kg mol−1
and cross-linked them under the same conditions as those used
for the PEO prepolymers (Scheme 2). Some PEO-based
samples were swollen in methanol (MeOH) after curing to
study the effect of alcohol treatment on stress relaxation.
Similarly, a control PEO-based sample was prepared without
catalyst.
DMTA of the PEO-based materials exhibited a plateau
modulus of 3.7 MPa for the MDI-based materials and a modulus
of 6.6 MPa for those prepared with PMDI, indicating that the
higher isocyanate functionality of PMDI increased the cross-link
density as expected (Figure 1, Table 1). During the DMTA
experiment, an increase in the modulus of the materials with
catalyst was observed above 150 °C. After cooling the samples
back to room temperature, we subjected them to a second
DMTA experiment and found that the plateau moduli had
increased slightly (Figure S1). These increases in the moduli are
likely the outcome of further cross-linking at ≥150 °C. By
comparison with the DMTA data for materials without catalyst
and those swollen in MeOH, it is apparent that the presence of
catalyst was necessary to observe further cross-linking at high
While the Ea values were similar across all polymers studied,
the rates of relaxation differed significantly depending on the
structure of the networks. Relaxation rates increased dramati-
cally with decreasing plateau modulus (Figure 3), which is
E
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
Table 1. Mechanical and Thermal Data of Urethane Cross-Linked PEOs and PLAs
a
b
d
d
e
f
g
h
i
sample
gel %
E′ (MPa) υe × 104 (mol mL−1
)
Mx (kg mol−1
)
Tg,DMTA (°C)
Tg,DSC (°C)
Td,5% (°C)
Ea (kJ mol−1
)
Tv (°C)
PEO-0.8-PMDI
PEO-0.8-MDI
PLA-1-MDI
PLA-3.8-MDI
PLA-10-MDI
PEO-0.8-PMDI-SM
PEO-0.8-MDI-SM
PEO-0.8-MDI-NC
99
100
99
99
99
100
100
95
6.6
3.7
2.4
1.3
1.3
6.6
4.2
3.5
7.1
4.0
2.6
1.4
1.4
7.1
4.5
3.8
1.6
2.8
4.8
8.9
8.9
1.6
2.5
3
3
1
−3
55
55
53
−1
−4
−14
303
294
189
188
202
299
293
309
142 11
83
76
67
56
59
108
81
91
−1
63
54
53
0
0
−10
139
155
153
165
5
5
3
3
127 31
87 27
102 37
a
Samples are named according to the prepolymer type-molar mass (kg mol−1)-cross-linker used. SM denotes the samples swollen in methanol, and
NC denotes the sample made without catalyst. Determined as the ratio of the dry mass after a swell test to the mass before the swell test
b
c
d
e
multiplied by 100. Determined at 100 °C. Determined using eq 1. Defined as the temperature where the maximum in the loss modulus (E″)
f
occurs in DMTA. Measured after erasing the thermal history at 150 °C for 1 min, cooling to −80 °C, and heating back to 150 °C at a rate 10 °C
g
h
min−1. Determined by heating from 20 to 550 °C under air at 10 °C min−1. Determined from the Arrhenius fit from SRA (eq 2). The standard
i
error as determined by Origin software is shown with these values. Determined by extrapolating the Arrhenius fit from SRA to τ*v for each
individual sample, which is determined using eq 3.
Figure 2. (A) SRA for PEO- and PLA-based materials at 130 °C and 5% strain. The experiments were stopped once the sample reached 1/e (37%) of
the initial stress relaxation modulus (G0). (B) Arrhenius analyses of the characteristic relaxation times for each sample. Three repetitions at each
temperature are plotted. Dashed lines indicate samples that were swollen in MeOH, and the dotted line indicates the sample prepared with no catalyst.
Figure 3. Plots of the characteristic relaxation time (τ*) at 130 °C as a function of (A) the storage modulus (E′) at 130 °C and (B) [OH]res, which was
estimated from the cross-linking reaction stoichiometry. The values of [OH]res are also shown in A using the color that corresponds to each material.
consistent with previous studies performed on polyester
vitrimers.49 On the basis of stoichiometry, PEO-0.8-PMDI
and PEO-0.8-MDI have nearly identical residual hydroxyl group
concentrations ([OH]res); however, the slower stress relaxation
behavior of PEO-0.8-PMDI suggested that the modulus
influenced the relaxation rate more strongly than did [OH]res
(Figure 3). Likewise, PLA-3.8-MDI and PLA-10-MDI had
similar plateau moduli and vastly different [OH]res values, yet
their rates of stress relaxation were essentially identical (Figure
3). Although it is difficult to completely decouple the influence
of E′ and [OH]res on the stress relaxation rate, it is apparent that
across all samples, both in this work and in our previous study,
the modulus has a stronger influence on the relaxation rate
than the residual hydroxyl concentration.33 Samples with a lower
plateau modulus have a lower urethane concentration, which
suggests that the likely mechanism of relaxation is not
bimolecular. We speculate that relaxation is more rapid at
lower cross-link density because of the increased ability of
reactive groups to diffuse within the network. Furthermore, if the
mechanism of relaxation were dissociative, each urethane
F
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
reversion would have a proportionally greater effect on the
reduction of the overall cross-link density in more loosely cross-
linked networks.
In our previous studies with PLA-based materials, we found
that urethane-based stress relaxation occurred more rapidly in
the presence of Sn(Oct)2;33 however, we did not determine an
Ea for stress relaxation in the absence of Sn(Oct)2. Therefore, we
investigated the stress relaxation of PEO-based materials
prepared without a catalyst as well as catalyst-containing
materials that had been swollen in MeOH. The rate of relaxation
was approximately 30 times slower in the absence of Sn(Oct)2
(Figure 2B, dotted line), indicating that the presence of tin
greatly facilitates stress relaxation. We observed similar slow
relaxation behavior in the samples swollen in MeOH (PEO-0.8-
MDI-SM and PEO-0.8-PMDI-SM, Figure 2B, dashed lines).
Although we initially hypothesized that MeOH treatment would
completely remove the catalyst, the Sn contents of samples pre-
and postswelling were evaluated by ICP-MS and determined to
be between 0.4 and 1 wt %, which indicated that Sn was not
effectively removed by swelling the samples in MeOH. The Sn
content of a sample prepared without catalyst was negligible (<1
ppm). Therefore, the ICP-MS and SRA results suggest that
swelling in excess MeOH possibly deactivates the Sn(Oct)2
catalyst toward urethane exchange rather than removing it. For
the MeOH-treated and catalyst-free samples, the τ* values
increased significantly after each successive run at a given
temperature, resulting in large errors for the calculated activation
energies. There were no evident changes in the FTIR spectra
before and after SRA for these samples, and no significant mass
loss observed before 290 °C (Figures S5 and S15). However,
SRA resulted in discoloration of the samples, and we suspect that
the inconsistent stress relaxation behavior between successive
runs was at least partially due to the decomposition of the
reactive functionalities or the formation of additional non-
dynamic cross-links. The multiple processes contributing to
stress relaxation in the absence of active tin species convolute the
comparison of the measured activation energies between
MeOH-treated and catalyst-free samples and those with
Sn(Oct)2 present.
To further understand the exchange reactions of urethanes
and the effects of hydroxyl groups on these processes, we studied
the behavior of urethane-containing model compounds at
elevated temperature. In the presence of excess 1-decanol and
Sn(Oct)2 (2.5 mol % with respect to urethane) at 150 °C, N-
phenyl-O-octyl urethane reacted to yield the O-decyl urethane at
a much slower rate compared to the timescale of the stress
relaxation (Figure 4A). We therefore analyzed the reaction
between two discrete urethanes in the absence of exogenous
alcohol (Figure 4B). Under these conditions, the formation of
crossover products is observed more quickly and at timescales
more similar to stress relaxation, suggesting that urethane
reversion, not alcohol-induced associative exchange, is the
dominant exchange mechanism (Figure 4B). Furthermore,
when two discrete urethanes were heated in the presence of
exogenous alcohol, only exchange with the free alcohol was
observed (at a rate similar to urethane-hydroxyl exchange)
(Figure S6A). These results indicate that excess free alcohols
trap the isocyanate intermediates that would otherwise enable
urethane−urethane exchange. However, the free alcohols also
slow the rate of exchange by coordinating to the Sn atoms,
thereby attenuating their ability to catalyze urethane exchange.
Decreasing the concentration of free alcohol in this model
exchange reaction leads to a recovery of the reactivity, consistent
Figure 4. GC−MS of model compound studies. (A) Urethane−
hydroxyl exchange conducted at 150 °C for 4 h (10 OH groups per
urethane) and (B) urethane−urethane exchange performed at 150 °C
for 2 h (equimolar in both urethanes). Peak intensities are normalized
to triphenylmethane (retention time = 13.7 min). Relative area
percentages of each compound are shown, and although qualitatively
significant, compound degradation on the GC column prevented
quantitative determination (see Figure S6 for more information).
with this interpretation (Figure S7). These findings are
consistent with the slower stress relaxation observed for the
MeOH-treated samples because while Sn atoms are still present,
the rates of relaxation are similar to that of the catalyst-free
material.
A urethane−urethane crossover experiment was also
conducted on analogous N-methylated urethanes to explore
the possibility of a distinct metathesis mechanism (Figure S6B)
because N,N-disubstituted urethanes are incapable of rever-
sion.26,50 Neither exchange products nor byproducts were
observed, again supporting a dissociative urethane exchange
mechanism for the nonmethylated urethanes. Unfortunately, the
GC−MS results for all model studies were not amenable to
quantitative reaction rate determination due to partial
decomposition on the GC column (see Figure S6 for more
1
information). Attempts to use H NMR spectroscopy as an
alternative method to quantitatively determine the rate of
reaction of a free alcohol with a model urethane were
complicated by side reactions occurring at the higher temper-
atures required for hydroxyl-urethane exchange (see Figure S8
for more information).
G
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
Although it appeared that a urethane reversion pathway was
operative in these systems, this mechanism was inconsistent with
the insolubility of these networks in TCB at 140 °C.33 Other
reprocessable cross-linked polymers based on reversible
reactions, such as Diels−Alder cycloadditions, are capable of
full dissolution at elevated temperatures.9 Therefore, we posited
that the low polarity of TCB was not appropriate for dissolving
the polymers but that a polar aprotic solvent such as DMSO
could favor dissolution. Indeed, all three variations of PEO-0.8-
MDI as well as PLA-1.0-MDI were insoluble in TCB but
dissolved fully in anhydrous DMSO at elevated temperature
(Figures S9 and S10). Model urethane−urethane exchange
studies performed with the addition of DMSO and TCB show
very similar reactivity, suggesting that the samples dissolve in
DMSO due to the increased swellability or solubility of the
samples in DMSO rather than a significant difference in the
reactivity of the urethane functional groups (Figure S12). The
samples swollen in TCB were dried under vacuum and analyzed
using FTIR spectroscopy. While there were some minor
detectable differences, no significant change in the carbonyl
resonances was observed despite substantial discoloration
(Figure S11). In contrast to our expectations based on rates of
stress relaxation, the samples treated with MeOH dissolved in
DMSO significantly faster (24 h) than samples with (36 h) and
without catalyst (48 h to be mostly dissolved, 96 h for complete
dissolution). We hypothesize that the additional alcohols in the
MeOH-treated samples result in a more rapid net loss of cross-
links during the swelling experiment.
an upfield shift of the N−H peak due to the loss of hydrogen
bonding (Figure S13), indicating that the isocyanate concen-
tration at these temperatures is below the detection limit of
NMR spectroscopy. The fact that urethane is strongly favored at
equilibrium at these temperatures is consistent with the
Arrhenius-type dependence of stress relaxation because the
cross-links never dissociate to the degree required to cause a
rapid drop in viscosity.
We sought other evidence for a reversion mechanism by
indirectly detecting isocyanate-derived species upon removal of
alcohol. A mixture of the diethyl urethane adduct of MDI and
Sn(Oct)2 (2.5 mol % with respect to urethane) was heated in a
distillation apparatus to drive the formation of MDI by the
removal of ethanol. We found that no ethanol was recovered
after 24 h at 140 or 150 °C. At 160 °C, however, some ethanol
was recovered (18% of the theoretical amount), and we
observed an insoluble brown solid in the distillation pot. The
FTIR spectrum of the solid indicated the presence of
isocyanurate moieties (1509, 1411, and 1171 cm−1, Figure
S14),51 which is consistent with the formation of isocyanates
followed by trimerization to isocyanurates during alcohol
removal. This experiment confirmed that the isocyanate species
was transient at temperatures relevant to stress relaxation (140−
150 °C), but is consistent with a reversion-based mechanism.
On the basis of SRA, model reactions, and literature
precedent,52,53 we propose two potential mechanisms of stress
relaxation in these materials (Scheme 3). In the presence of
exogenous alcohol, the mechanism in Scheme 3A would
predominate. Low conversion observed in the hydroxyl−
urethane exchange model reactions (10 OH groups per
urethane) and low relaxation rates in the MeOH-treated
samples may arise from the coordination of the Sn(II) metal
center by alcohols, inhibiting its ability to catalyze urethane
reversion. However, rates would increase as [OH] decreases,
which is consistent with the fast urethane−urethane exchange
observed in the absence of exogenous alcohol. Furthermore,
because [OH]res in the cross-linked materials is low, inhibition of
the Sn(II) centers by excess alcohol is less likely to occur.
Meanwhile, in the absence of exogenous alcohol, urethanes can
more freely bind to the catalytic Sn(II) center, allowing for full
reversion and subsequently fast exchange (Scheme 3B). This
mechanism is consistent with previous observations of polyur-
ethane stress relaxation in both the presence37,38 and
absence35,36,40 of tin catalysts.
Because DMSO was capable of completely dissolving the
PEO-based samples, we sought to directly detect the formation
of the isocyanate intermediates. We acquired a 13C NMR
spectrum of the diethyl urethane adduct of MDI in DMSO-d6 at
25, 100, and 140 °C (Figure 5). No peaks were observed in the
13C NMR spectrum that corresponded to the isocyanate,
suggesting that the equilibrium significantly favored the ure-
thane at these temperatures. A 1H NMR spectrum showed only
CONCLUSIONS
■
We demonstrate that PEO- and PLA-based networks with
urethane cross-links are capable of stress relaxation at elevated
temperature, likely by urethane reversion. The modulus of the
materials is apparently the principal factor controlling the
relaxation time, as samples with lower moduli relax more
quickly. Although an increase in the concentration of residual
hydroxyl moieties slightly lowers the Ea of stress relaxation,
[OH]res in these materials does not affect the overall rate of
relaxation as significantly as the storage modulus, suggesting that
reversion is the primary mechanism for urethane exchange.
Urethane reversion is further supported by model compound
studies in which rapid urethane exchange is observed in the
absence of free hydroxyl groups and only when the urethane
contains N−H as opposed to N−Me. Furthermore, the alcohol-
based inhibition of the Sn(Oct)2 catalyst is apparent in samples
treated with MeOH as well as in model compound reactions
performed in a large excess of exogenous alcohol. Although we
Figure 5. 13C VT-NMR spectra (125 MHz, DMSO-d6) of MDI and the
diethyl urethane adduct of MDI; concentrations of approximately 100
mg mL−1 were used to achieve a better signal-to-noise ratio. The diethyl
urethane solution was under a N2 atmosphere and contained Sn(Oct)2
(2.5 mol % with respect to urethanes); the solution was allowed to
equilibrate at the desired temperature for 10 min before 128 scans were
collected.
H
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
Scheme 3. Proposed Stress Relaxation Mechanisms of (A) the Hydroxyl-Urethane Exchange (Urethane Reversion and Reaction
a
with a New Hydroxyl) and (B) the Urethane−Urethane Exchange (Urethane Reversion and Recombination)
a
Stress relaxation by mechanism A is much slower than by mechanism B due to the competitive inhibition from free hydroxyl groups coordinating
to the tin center.
were unable to directly observe the presence of isocyanates using
NMR spectroscopy at elevated temperatures, the generation of
isocyanurate moieties after the distillation of ethanol from the
diethyl urethane adduct of MDI provided further indirect evi-
dence of the formation of isocyanate intermediates. These
studies yield insight into the vitrimer-like behavior of polyur-
ethane networks at elevated temperatures and suggest that
further investigations of the reprocessability of commercially
ubiquitous cross-linked polyurethanes will be a beneficial
approach to their recycling.
supporting this work. We also thank Dr. Letitia Yao for
performing VT-NMR experiments.
REFERENCES
■
(1) Wu, D. Y.; Meure, S.; Solomon, D. Self-healing Polymeric
Materials: A Review of Recent Developments. Prog. Polym. Sci. 2008,
33, 479−522.
(2) Tyagi, P.; Deratani, A.; Quemener, D. Self-Healing Dynamic
Polymeric Systems. Isr. J. Chem. 2013, 53, 53−60.
(3) Kloxin, C. J.; Bowman, C. N. Covalent Adaptable Networks:
Smart, Reconfigurable and Responsive Network Systems. Chem. Soc.
Rev. 2013, 42, 7161−7173.
ASSOCIATED CONTENT
■
S
(4) Wool, R. P. Self-healing Materials: A Review. Soft Matter 2008, 4,
400−418.
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jpcb.8b11489.
(5) Wojtecki, R. J.; Meador, M. A.; Rowan, S. J. Using the Dynamic
Bond to Access Macroscopically Responsive Structurally Dynamic
Polymers. Nat. Mater. 2011, 10, 14−27.
1
SRA, FTIR, GC, TGA, DSC, and H NMR results and
pictures of the high-temperature swelling experiments
(PDF)
(6) Roy, N.; Bruchmann, B.; Lehn, J.-M. DYNAMERS: Dynamic
Polymers as Self-healing Materials. Chem. Soc. Rev. 2015, 44, 3786−
3807.
(7) Chen, X.; Dam, M. A.; Ono, K.; Mal, A.; Shen, H.; Nutt, S. R.;
Sheran, K.; Wudl, F. A Thermally Re-mendable Cross-Linked
Polymeric Material. Science 2002, 295, 1698−1702.
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: wdichtel@northwestern.edu.
*E-mail: hillmyer@umn.edu. Phone: 612-625-7834.
(8) Chen, X.; Wudl, F.; Mal, A. K.; Shen, H.; Nutt, S. R. New
Thermally Remendable Highly Cross-Linked Polymeric Materials.
Macromolecules 2003, 36, 1802−1807.
ORCID
(9) Zhang, Y.; Broekhuis, A. A.; Picchioni, F. Thermally Self-Healing
Polymeric Materials: The Next Step to Recycling Thermoset Polymers?
Macromolecules 2009, 42, 1906−1912.
(10) Montarnal, D.; Capelot, M.; Tournilhac, F.; Leibler, L. Silica-Like
Malleable Materials from Permanent Organic Networks. Science 2011,
334, 965−968.
David J. Fortman: 0000-0002-0422-3733
Guilhem X. De Hoe: 0000-0002-0996-7491
William R. Dichtel: 0000-0002-3635-6119
Marc A. Hillmyer: 0000-0001-8255-3853
Notes
(11) Imbernon, L.; Norvez, S. From Landfilling to Vitrimer Chemistry
in Rubber Life Cycle. Eur. Polym. J. 2016, 82, 347−376.
(12) Denissen, W.; Winne, J. M.; Du Prez, F. E. Vitrimers: Permanent
Organic Networks With Glass-like Fluidity. Chem. Sci. 2016, 7, 30−38.
(13) Capelot, M.; Montarnal, D.; Tournilhac, F.; Leibler, L. Metal-
Catalyzed Transesterification for Healing and Assembling of
Thermosets. J. Am. Chem. Soc. 2012, 134, 7664−7667.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors acknowledge the Center for Sustainable Polymers
at the University of Minnesota, a National Science Foundation-
supported center for chemical innovation (CHE-1413862), for
I
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B
Article
(14) Capelot, M.; Unterlass, M. M.; Tournilhac, F.; Leibler, L.
Catalytic Control of the Vitrimer Glass Transition. ACS Macro Lett.
2012, 1, 789−792.
(35) Colodny, P. C.; Tobolsky, A. V. Chemorheological Study of
Polyurethan Elastomers. J. Am. Chem. Soc. 1957, 79, 4320−4323.
(36) Offenbach, J. A.; Tobolsky, A. V. Chemical Relaxation of Stress in
Polyurethane Elastomers. J. Colloid Sci. 1956, 11, 39−47.
(37) Zheng, N.; Fang, Z.; Zou, W.; Zhao, Q.; Xie, T. Thermoset
Shape-Memory Polyurethane with Intrinsic Plasticity Enabled by
Transcarbamoylation. Angew. Chem., Int. Ed. 2016, 55, 11421−11425.
(38) Yan, P.; Zhao, W.; Fu, X.; Liu, Z.; Kong, W.; Zhou, C.; Lei, J.
Multifunctional Polyurethane-Vitrimers Completely Based on Trans-
carbamoylation of Carbamates: Thermally-Induced Shape Memory
Effect and Self-Welding. RSC Adv. 2017, 7, 26858−26866.
(39) Yan, P.; Zhao, W.; Wang, Y.; Jiang, Y.; Zhou, C.; Lei, J. Carbon
Nanotubes-Polyurethane Vitrimer Nanocomposites with the Ability of
Surface Welding Controlled by Heat and Near-Infrared Light.
Macromol. Chem. Phys. 2017, 218, 1700265.
(40) Zheng, N.; Hou, J.; Xu, Y.; Fang, Z.; Zou, W.; Zhao, Q.; Xie, T.
Catalyst-Free Thermoset Polyurethane with Permanent Shape
Reconfigurability and Highly Tunable Triple-Shape Memory Perform-
ance. ACS Macro Lett. 2017, 6, 326−330.
(41) Yang, Y.; Urban, M. W. Self-Healing of Glucose-Modified
Polyurethane Networks Facilitated by Damage-Induced Primary
Amines. Polym. Chem. 2017, 8, 303−309.
(42) De Hoe, G. X.; Zumstein, M. T.; Tiegs, B. J.; Brutman, J. P.;
McNeill, K.; Sander, M.; Coates, G. W.; Hillmyer, M. A. Sustainable
Polyester Elastomers from Lactones: Synthesis, Properties, and
Enzymatic Hydrolyzability. J. Am. Chem. Soc. 2018, 140, 963−973.
(43) Brutman, J. P.; De Hoe, G. X.; Schneiderman, D. K.; Le, T. N.;
Hillmyer, M. A. Renewable, Degradable, and Chemically Recyclable
Cross-Linked Elastomers. Ind. Eng. Chem. Res. 2016, 55, 11097−11106.
(44) Angell, C. A. Formation of Glasses from Liquids and
Biopolymers. Science 1995, 267, 1924−1935.
(45) Ediger, M. D.; Angell, C. A.; Nagel, S. R. Supercooled Liquids and
Glasses. J. Phys. Chem. 1996, 100, 13200−13212.
(46) Dyre, J. C. Colloquium: The Glass Transition and Elastic Models
of Glass-forming Liquids. Rev. Mod. Phys. 2006, 78, 953−972.
(47) Hiemenz, P. C.; Lodge, T. P. Polymer Chemistry, 2nd ed.; CRC
Press: Boca Raton, FL, 2007.
(48) Auras, R. A.; Lim, L. T.; Selke, S. E. M.; Tsuji, H. Poly(lactic acid):
Synthesis, Structures, Properties, Processing, and Applications; John Wiley
& Sons: Hoboken, NJ, 2010.
(49) Yu, K.; Taynton, P.; Zhang, W.; Dunn, M. L.; Qi, H. J. Influence
of Stoichiometry on the Glass Transition and Bond Exchange Reactions
in Epoxy Thermoset Polymers. RSC Adv. 2014, 4, 48682−48690.
(50) Gaylord, N. G.; Sroog, C. E. The Reactions of Carbamates with
Alcohols. J. Org. Chem. 1953, 18, 1632−1637.
(51) Pawar, G. M.; Buchmeiser, M. R. Polymer-Supported, Carbon
Dioxide-Protected N-Heterocyclic Carbenes: Synthesis and Applica-
tion in Organo- and Organometallic Catalysis. Adv. Synth. Catal. 2010,
352, 917−928.
(52) Bloodworth, A. J.; Davies, A. G. Organometallic Reactions. Part I.
The Addition of Tin Alkoxides to Isocyanates. J. Chem. Soc. 1965,
5238−5244.
(53) Delebecq, E.; Pascault, J.-P.; Boutevin, B.; Ganachaud, F. On the
Versatility of Urethane/Urea Bonds: Reversibility, Blocked Isocyanate,
and Non-isocyanate Polyurethane. Chem. Rev. 2013, 113, 80−118.
(15) Lu, Y.-X.; Tournilhac, F.; Leibler, L.; Guan, Z. Making Insoluble
Polymer Networks Malleable via Olefin Metathesis. J. Am. Chem. Soc.
2012, 134, 8424−8427.
(16) Lu, Y.-X.; Guan, Z. Olefin Metathesis for Effective Polymer
Healing via Dynamic Exchange of Strong Carbon−Carbon Double
Bonds. J. Am. Chem. Soc. 2012, 134, 14226−14231.
(17) Ying, H.; Zhang, Y.; Cheng, J. Dynamic Urea Bond for the Design
of Reversible and Self-Healing Polymers. Nat. Commun. 2014, 5, 3218−
3227.
(18) Zhang, Y.; Ying, H.; Hart, K. R.; Wu, Y.; Hsu, A. J.; Coppola, A.
M.; Kim, T. A.; Yang, K.; Sottos, N. R.; White, S. R.; Cheng, J. Malleable
and Recyclable Poly(urea-urethane) Thermosets Bearing Hindered
Urea Bonds. Adv. Mater. 2016, 28, 7646−7651.
(19) Martin, R.; Rekondo, A.; Ruiz de Luzuriaga, A.; Cabanero, G.;
Grande, H. J.; Odriozola, I. The Processability of a Poly(urea-urethane)
Elastomer Reversibly Crosslinked With Aromatic Disulfide Bridges. J.
Mater. Chem. A 2014, 2, 5710−5715.
(20) Rekondo, A.; Martin, R.; Ruiz de Luzuriaga, A.; Cabanero, G.;
Grande, H. J.; Odriozola, I. Catalyst-free Room-temperature Self-
healing Elastomers Based on Aromatic Disulfide Metathesis. Mater.
Horiz. 2014, 1, 237−240.
(21) Xiang, H. P.; Qian, H. J.; Lu, Z. Y.; Rong, M. Z.; Zhang, M. Q.
Crack Healing and Reclaiming of Vulcanized Rubber by Triggering the
Rearrangement of Inherent Sulfur Crosslinked Networks. Green Chem.
2015, 17, 4315−4325.
(22) Xiang, H. P.; Rong, M. Z.; Zhang, M. Q. Self-healing, Reshaping,
and Recycling of Vulcanized Chloroprene Rubber: A Case Study of
Multitask Cyclic Utilization of Cross-linked Polymer. ACS Sustainable
Chem. Eng. 2016, 4, 2715−2724.
(23) Pepels, M.; Filot, I.; Klumperman, B.; Goossens, H. Self-healing
Systems Based on Disulfide-thiol Exchange Reactions. Polym. Chem.
2013, 4, 4955−4965.
(24) Denissen, W.; Rivero, G.; Nicolay, R.; Leibler, L.; Winne, J. M.;
̈
Du Prez, F. E. Vinylogous Urethane Vitrimers. Adv. Funct. Mater. 2015,
25, 2451−2457.
(25) Denissen, W.; Droesbeke, M.; Renaud, N.; Leibler, L.; Winne, J.
M.; Du Prez, F. Chemical Control of the Viscoelastic Properties of
Vinylogous Urethane Vitrimers. Nat. Commun. 2017, 8, 14857.
(26) Fortman, D. J.; Brutman, J. P.; Cramer, C. J.; Hillmyer, M. A.;
Dichtel, W. R. Mechanically Activated, Catalyst-Free Polyhydroxyur-
ethane Vitrimers. J. Am. Chem. Soc. 2015, 137, 14019−14022.
(27) Fortman, D. J.; Brutman, J. P.; Hillmyer, M. A.; Dichtel, W. R.
Structural Effects on the Reprocessability and Stress relaxation of
Cross-linked Polyhydroxyurethanes. J. Appl. Polym. Sci. 2017, 134,
44984.
(28) Zheng, P.; McCarthy, T. J. A Surprise from 1954: Siloxane
Equilibration Is a Simple, Robust, and Obvious Polymer Self-Healing
Mechanism. J. Am. Chem. Soc. 2012, 134, 2024−2027.
(29) Schmolke, W.; Perner, N.; Seiffert, S. Dynamically Cross-Linked
Polydimethylsiloxane Networks with Ambient-Temperature Self-
Healing. Macromolecules 2015, 48, 8781−8788.
(30) Cash, J. J.; Kubo, T.; Bapat, A. P.; Sumerlin, B. S. Room-
Temperature Self-Healing Polymers Based on Dynamic-Covalent
Boronic Esters. Macromolecules 2015, 48, 2098−2106.
(31) Cromwell, O. R.; Chung, J.; Guan, Z. Malleable and Self-Healing
Covalent Polymer Networks through Tunable Dynamic Boronic Ester
Bonds. J. Am. Chem. Soc. 2015, 137, 6492−6495.
(32) Roettger, M.; Domenech, T.; van der Weegen, R.; Breuillac, A.;
Renaud, N.; Leibler, L. High-performance Vitrimers from Commodity
Thermoplastics through Dioxaborolane Metathesis. Science 2017, 356,
62−65.
(33) Brutman, J. P.; Delgado, P. A.; Hillmyer, M. A. Polylactide
Vitrimers. ACS Macro Lett. 2014, 3, 607−610.
(34) Yu, Y.; Storti, G.; Morbidelli, M. Kinetics of Ring-Opening
Polymerization of l,l-Lactide. Ind. Eng. Chem. Res. 2011, 50, 7927−
7940.
J
DOI: 10.1021/acs.jpcb.8b11489
J. Phys. Chem. B XXXX, XXX, XXX−XXX