Metal-catalyzed transesterification for healing and assembling of thermosets

Article abstract of DOI:10.1021/ja302894k

Catalytic control of bond exchange reactions enables healing of cross-linked polymer materials under a wide range of conditions. The healing capability at high temperatures is demonstrated for epoxy-acid and epoxy-anhydride thermoset networks in the presence of transesterification catalysts. At lower temperatures, the exchange reactions are very sluggish, and the materials have properties of classical epoxy thermosets. Studies of model molecules confirmed that the healing kinetics is controlled by the transesterification reaction rate. The possibility of varying the catalyst concentration brings control and flexibility of welding and assembling of epoxy thermosets that do not exist for thermoplastics.

Full text of DOI:10.1021/ja302894k

Communication
pubs.acs.org/JACS
Metal-Catalyzed Transesterification for Healing and Assembling of
Thermosets
Mathieu Capelot, Damien Montarnal, Franco̧ is Tournilhac, and Ludwik Leibler*
Matier
̀
e Molle et Chimie (UMR 7167 ESPCI-CNRS), Ecole Super
́
ieure de Physique et Chimie Industrielles de la Ville de Paris
ESPCI ParisTech, 10 rue Vauquelin, 75005 Paris, France
S
* Supporting Information
range must be well-controlled as in the case of thermoplastics.
Other reversible reactions such as amino/carbonyl reactions13
or siloxane equilibration14 have also been studied. Another
method consists of using radicals to initiate exchange
reactions15 and plasticity in thermosets.16 Materials cross-
linked with such radical-exchangeable reactions exhibit good
mending properties either thermally,17 photochemically,18 or
even trigger-free.19 Unfortunately, these methods require
purpose-made monomers, which limits immediate widespread
utilization.
Here we show that classical and largely available epoxy
thermosets can be welded efficiently by inducing and
controlling transesterification reactions. Such epoxy resins are
currently widely employed in applications such as coatings,
electronics, adhesives, and light-emitting diodes or as structural
matrices in composites.20 To allow for transesterification, the
hardeners must be chosen from among carboxylic acids and
acid anhydrides. Both react with epoxy rings to yield ester
links.20,21
We recently demonstrated that with proper catalysis, the
topologies of epoxy−acid and epoxy−anhydride networks can
be rearranged by transesterification exchange reactions without
modification of the numbers of links and average function-
ality.22 At high temperatures, the networks can flow by
topology rearrangements and behave like viscoelastic liquids.
When the temperature is decreased, the transesterification
exchanges slow down, and the topology of the network freezes.
Below the topology freezing transition, the materials have
properties of permanently cross-linked thermosets. The
reversible transition from the viscoelastic liquid state to the
elastic solid has the attributes of a glass transition. The viscosity
gradually decreases when the temperature is increased,
following an Arrhenius law as for the glass transition in silica
and a few other inorganic compounds, the so-called strong glass
formers.23 We call this new class of organic materials exhibiting
the network topology freezing glass transition “vitrimers”. Like
classical thermosets, vitrimers are insoluble irrespective of
temperature. The purpose of this paper is to prove that
autogenous welding of two pieces of vitrimer can be done over
a very broad range of temperatures. The principle of welding by
exchange reactions is presented in Figure 1a.
ABSTRACT: Catalytic control of bond exchange reac-
tions enables healing of cross-linked polymer materials
under a wide range of conditions. The healing capability at
high temperatures is demonstrated for epoxy−acid and
epoxy−anhydride thermoset networks in the presence of
transesterification catalysts. At lower temperatures, the
exchange reactions are very sluggish, and the materials
have properties of classical epoxy thermosets. Studies of
model molecules confirmed that the healing kinetics is
controlled by the transesterification reaction rate. The
possibility of varying the catalyst concentration brings
control and flexibility of welding and assembling of epoxy
thermosets that do not exist for thermoplastics.
he ability to be welded by simple heating is an important
feature of metals and silica glasses. Indeed, welding allows
T
manufacturing of complex objects. For thermoplastic polymers,
welding induced by the diffusion of polymer chains through the
interface via a reptation process have been observed above1 and
slightly below2 the glass transition temperature. However, the
temperature must be carefully controlled to avoid material flow
because the viscosity drops abruptly in a very narrow
temperature range near the glass transition.3 Chemically
cross-linked polymers, also called thermosets, have outstanding
thermal and mechanical properties and are irreplaceable in
demanding applications, such as in structural parts needed by
the aircraft and automotive industries. However, thermosets are
not easily welded, as any kind of reprocessing is impossible
once the curing reaction is complete. Various methods of
assembly by adhesion have been envisaged, such as rough
surface interlocking4 and interdiffusion of network defects5 or
added linear thermoplastic chains6 at the interface. For more
robust welding, it is possible in some cases to stop the curing
reaction before it goes to completion. Welding can then be
induced by restarting the reaction once the desired parts are in
contact.7
Recently, elegant dynamic covalent chemistries8 have been
introduced in thermosetting polymers to yield self-mending or
stress-relaxation properties,9 as predicted from theoretical
works.10 Wudl and co-workers have developed a thermosetting
material based on the thermoreversible Diels−Alder reaction.11
When heated, the sample can be mended, and the initial
properties are recovered after cooling. However, depolymeriza-
tion causes an abrupt viscosity drop during heating.12
Therefore, if molds are not used, the assembly temperature
The epoxy−acid networks used in this study were
synthesized from the diglycidyl ether of bisphenol A
(DGEBA) and a mixture of dicarboxylic and tricarboxylic
Received: March 30, 2012
Published: April 26, 2012
© 2012 American Chemical Society
7664
dx.doi.org/10.1021/ja302894k | J. Am. Chem. Soc. 2012, 134, 7664−7667

Journal of the American Chemical Society
Communication
Figure 1. (a) Scheme illustrating the welding of two pieces of epoxy resin by transesterification exchange reactions. (b) Structure of the epoxy−acid
network used in the study. The epoxy resin DGEBA is linked through β-hydroxy esters to dicarboxylic and tricarboxylic acids. (c) Model molecules
E18 and E19 investigated to mimic β-hydroxy ester links. Numerous transesterification products appear when the mixture is heated, such as diester
DE28 and β-hydroxy esters E17 and E20.
fatty acids (Figure 1b). The reaction of epoxy rings with
carboxylic acids is rather complex, because several backside
reactions can affect the network structure and properties. Three
main reactions have been reported: (a) the addition of
carboxylic acid groups on epoxy rings, yielding β-hydroxy
esters; (b) the addition of hydroxyl groups on epoxy rings,
yielding ethers; and (c) Fischer esterification of the hydroxyl
groups by carboxylic acids [see the Supporting Information
(SI)].24 Several catalysts, such as 2-methylimidazole (2-MI),
have been used over the years to promote reaction (a) over (b)
and (c).25 Under these conditions, the gelation may be
predicted from the stoichiometry and average functionality
using Monte Carlo simulations.26 Also, for a stoichiometry of
one acid to one epoxy ring, all of the ester and hydroxyl groups
are obtained in equal amounts and are more prone to
Figure 2. (a) Normalized GC chromatograms of an E18 + E19
mixture with 6 mol % 2-MI + 5 mol % Zn(acac)2 at 150 °C after 5
min, 45 min, and 2 h. A quantitative kinetics study was performed by
transesterification reactions.27
To quantify the kinetics of transesterification and control the
dynamics of exchanges by metal catalysis, model compounds
were studied. We synthesized two different β-hydroxy esters,
E18 and E19 (where Ex is the β-hydroxy ester with x carbon
atoms; Figure 1c), that mimic the ester links formed between a
carboxylic acid and an epoxy unit. The exchange reaction was
studied by GC−MS. Indeed, by design the exchange products
are distinguished by the number of carbon atoms, and thus,
they could be well-separated on a GC chromatogram and easily
identified by mass spectrometry. After the mixture of model
compounds E18 and E19 was heated, transesterification
products appeared. Thus, at slightly shorter and longer
retention times than the initial esters (Figure 2a), we observed
the formation of the monoesters E17 and E20, respectively. At
long retention times, the diester products corresponding to all
six transesterification combinations among E17, E18, E19, and
E20 were also present (an example is DE28 in Figure 1b; see
the SI). As the four monoesters have very similar structures, a
quantitative kinetics study could be performed by following the
area ratio r = (E17 + E20)/(E18 + E19) as a function of time.
following the appearance of E17 and E20 with time. (b) Trans-
esterification kinetics results for the model molecules. With no metal
catalyst added or the addition of only 1 mol % Zn(OAc)2, equilibrium
□
was attained after 15 h. With 5 mol % Zn(OAc)2 ( ) or Zn(acac)2
■
( ), this time was reduced to 2 h.
Without the addition of any metal catalyst, the trans-
esterification reaction took ∼15 h at 150 °C (Figure 2b). The
method allowed us to screen different catalysts selected from
the literature.28 By adding metal salts such as zinc acetate or
zinc acetylacetonate at 5 mol %, we were able to reduce the
exchange time to only 2 h. Interestingly, zinc acetate is also an
efficient esterification catalyst25 and can thus be used alone in
epoxy−acid networks. The metal concentration plays an
important role, as shown in Figure 2b.
Epoxy−acid networks were synthesized in the presence of
zinc acetate. The synthesis produced a soft epoxy network
exhibiting a Tg of ∼15 °C, a storage modulus of 4 MPa, and
good elastomeric properties at room temperature.
7665
dx.doi.org/10.1021/ja302894k | J. Am. Chem. Soc. 2012, 134, 7664−7667

Journal of the American Chemical Society
Communication
To investigate the welding properties of these networks, we
used the lap-shear test (Figure 3a). This test is intrinsically
several cycles can be done consecutively on the same samples
(see the SI).
To extend this welding process to technologically relevant
hard epoxy resins, networks were synthesized from DGEBA
and glutaric anhydride. When stoichiometric amounts of epoxy
rings and acid anhydrides were used, the addition−esterification
reaction yielded only diesters and no hydroxyl groups.
However, studies with model molecules showed that the
presence of hydroxyl groups is crucial for transesterification
reactions (Figure 4a). Under the same conditions of catalysis
Figure 3. (a) Photograph of a sample during the lap-shear test. (b)
Stress−strain curves of samples loaded with different catalyst
concentrations and welded for 1 h at 150 °C. (c) Stress−strain curves
of samples loaded with 5 mol % Zn(OAc)2 and welded under different
conditions.
difficult to pass for material assemblies, since a fracture is easily
initiated at the interface.29 In our experiments, two rectangular
samples of material (25 mm × 5 mm × 1.4 mm) were
superimposed on a 15 mm length and held together under
pressure for welding times ranging from 15 min to 15 h at a
controlled temperature (room temperature or 100, 125, or 150
°C). A good contact was ensured by applying a 25%
compression during the treatment. The welding efficiency
was then evaluated by carrying out tensile tests at room
temperature with a cross-head speed of 10 mm/min on the
assembly and comparing the forces at break.
With 1 mol % Zn(OAc)2, after 1 h at 150 °C, the assembly
broke quickly, with a force at break of ∼15 N. Welded under
the same conditions, a sample catalyzed at 5 mol % broke at 27
N (Figure 3b). For comparison, a simple rectangular ribbon
with the same total length (35 mm × 5 mm × 1.4 mm) broke
at 35 N. Welding for a longer time led to full recovery of the
mechanical properties: after several hours at 150 °C, the
rupture became cohesive and occurred in the bulk material far
from the welded interface. For an easier comparison with
results from model molecules, we limit the discussion here to
temperatures up to 150 °C. Faster weldings were achieved at
higher temperatures, however.
Figure 3c shows stress−strain curves of assemblies of
material catalyzed at 5 mol % that were welded for various
times at different temperatures. Just above the glass transition
(room temperature), only adhesion due to interdiffusion of
polymer chains (network defects) across the interface
contributed to the separation force, which was ∼6 N after
welding for 1 h. Longer times (3 or 15 h) did not improve the
adhesion very much. In contrast, at 150 °C, longer welding
times resulted in better assembly properties (see the SI). It is
noteworthy that the same welding properties could be obtained
at two different temperatures by adjusting the treatment time:
identical behavior was indeed observed with welding for 1 h at
125 °C or 30 min at 150 °C. This confirms that in contrast to
thermoplastics, efficient welding can be achieved over a broad
span of temperature. Thus, it should be possible to adjust more
easily the welding temperature and welding time for any object
dimensions. Indeed, by applying a contact pressure that is small
compared to the elastic modulus of the sample, the interfaces
can be healed without modification of the dimensions of the
welded parts (see the SI). Because chemical moieties are only
exchanged and not consumed during the welding process,
Figure 4. (a) Model molecule studies confirm the importance of
hydroxyl groups in the exchange reactions. (b) The stoichiometry
between epoxy rings and anhydride groups plays an important role in
the welding properties, as it influences the concentration of hydroxyl
groups in the network. The cross-sectional area was 1.4 mm × 5 mm.
Each value is an average of five independent tests.
and temperature, full transesterification exchange between two
hydroxy esters occurred within 2 h, only 50% exchange was
obtained between β-hydroxy ester E18 and diester DE29 after
30 h, and no exchange at all could be observed between the two
diesters DE26 and DE29 even after 24 h. Hydroxyl groups are
thus needed in the exchange process, and their concentration
controls the welding kinetics.
Experimentally, hydroxyl groups are observed in epoxy−
anhydride networks when an excess of epoxy is introduced.30
The epoxy−anhydride reaction has been extensively studied in
the literature,21 and the commonly accepted mechanism
suggests that the reaction is an alternate ring-opening
polymerization of anhydrides and epoxides.31 In a first step,
an anhydride is opened by a hydroxyl group or a catalyst,
yielding a monoester and a carboxylic acid. In a second step,
this acid opens an epoxy ring to form a diester and regenerate a
hydroxyl group (see the SI). With an excess of epoxy,
homopolymerization initiated by hydroxyl groups also occurs
at the end of the process.32 Since both reactions regenerate
hydroxyl groups, all of the hydroxyl groups present in the cured
material should have been present from the beginning. In the
present case, hydroxyl groups could be the result of epoxy-
opening reactions by water molecules, which are naturally
present in the metal catalyst. When the relative ratio between
the anhydride and epoxy was varied from 1:1 to 0.5:1, IR
studies showed that the quantity of hydroxyl groups in the
networks increased (see the SI).
Swelling experiments confirmed that the materials are well
above the gel point and are insoluble. Dynamic mechanical
experiments and differential scanning calorimetry showed that
these networks have a storage modulus ranging from 60 to 100
MPa, and a Tg above 60−70 °C. The materials are therefore
hard resins, comparable to those used in composites. The
addition of a transesterification catalyst in such formulations
enables efficient welding, as confirmed by lap-shear tests at
7666
dx.doi.org/10.1021/ja302894k | J. Am. Chem. Soc. 2012, 134, 7664−7667

Journal of the American Chemical Society
Communication
(c) Leibler, L.; Rubinstein, M.; Colby, R. H. J. Phys. II 1993, 3, 1581.
(d) Leibler, L; Rubinstein, M.; Colby, R. H. Macromolecules 1991, 24,
4701.
(11) (a) Chen, X.; Matheus, A. D.; Kanji, O.; Ajit, M.; Hongbin, S.;
Nutt, S. R.; Sheran, K.; Wudl, F. Science 2002, 295, 1698. (b) Murphy,
E. B.; Bolanos, E.; Schaffner-Hamann, C.; Wudl, F.; Nutt, S. R.; Auad,
M. L. Macromolecules 2008, 41, 5203.
(12) (a) Adzima, B. J.; Aguirre, H. A.; Kloxin, C. J.; Scott, T. F.;
Bowman, C. N. Macromolecules 2008, 41, 9112. (b) Zhang, Y.;
Broekhuis, A. A.; Picchioni, F. Macromolecules 2009, 42, 1906.
(13) (a) Skene, W. G.; Lehn, J.-M. P. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 8270. (b) Lehn, J. M. Prog. Polym. Sci. 2005, 30, 814.
(c) Deng, G.; Tang, C.; Li, F.; Jiang, H.; Chen, Y. Macromolecules
2010, 43, 1191.
(14) Zheng, P.; McCarthy, T. J. J. Am. Chem. Soc. 2012, 134, 2024.
(15) Otsuka, H.; Aotani, K.; Higaki, Y.; Takahara, A. J. Am. Chem.
Soc. 2003, 125, 4064.
(16) (a) Scott, T. F.; Schneider, A. D.; Cook, W. D.; Bowman, C. N.
Science 2005, 308, 1615. (b) Park, H. Y.; Kloxin, C. J.; Scott, T. F.;
Bowman, C. N. Macromolecules 2010, 43, 10188.
(17) (a) Nicolay, R.; Kamada, J.; Wassen, A. V.; Matyjaszewski, K.
Macromolecules 2010, 43, 4355. (b) Canadell, J.; Goossens, H.;
Klumperman, B. Macromolecules 2011, 44, 2536.
(18) (a) Ghosh, B.; Urban, R. W. Science 2009, 323, 1458.
(b) Amamoto, Y.; Kamada, J.; Otsuka, H.; Takahara, A.;
Matyjaszewski, K. Angew. Chem., Int. Ed. 2011, 50, 1660.
(19) Imato, K.; Nishihara, M.; Kanehara, T.; Amamoto, Y.; Takahara,
A.; Otsuka, H. Angew. Chem., Int. Ed. 2012, 51, 1138.
(20) Pascault, J. P.; Williams, R. J. J. Epoxy Polymers: New Materials
and Innovations; Wiley-VCH: Weinheim, Germany, 2010.
(21) May, C. A. Epoxy Resins: Chemistry and Technology, 2nd ed.;
Marcel Dekker: New York, 1988.
(22) Montarnal, D.; Capelot, M.; Tournilhac, F.; Leibler, L. Science
2011, 334, 965.
(23) Angell, C. A. Science 1995, 267, 1924.
(24) (a) Dusek, K.; Matejka, L. Adv. Chem. Ser. 1984, 208, 15.
(b) Hoppe, C. E.; Galante, M. J.; Oyanguren, P. A.; Williams, R. J. J.
Macromol. Mater. Eng. 2005, 290, 456.
(25) (a) Blank, W. J.; He, Z. A.; Picci, M. J. Coat. Technol. 2002, 74,
33. (b) Merfeld, G.; Molaison, C.; Koeniger, R.; Acar, A. E.;
Mordhorst, S.; Suriano, J.; Irwin, P.; Warner, R. S.; Gray, K.; Smith,
M.; Kovaleski, K.; Garrett, G.; Finley, S.; Meredith, D.; Spicer, M.;
Naguy, T. Prog. Org. Coat. 2005, 52, 98. (c) Pire, M.; Norvez, S.;
Iliopoulos, I.; Le Rossignol, B.; Leibler, L. Polymer 2011, 52, 5243.
(d) Pire, M.; Lorthioir, C.; Oikonomou, E. K.; Norvez, S.; Iliopoulos,
I.; Le Rossignol, B.; Leibler, L. Polym. Chem. 2012, 3, 946.
(26) Montarnal, D.; Tournilhac, F.; Hidalgo, M.; Leibler, L. J. Polym.
Sci., Polym. Chem. 2010, 48, 1133.
room temperature at 0.1 mm/min. Figure 4b shows that the
welding is stronger at lower anhydride/epoxy ratios.
These experiments demonstrate that metal-catalyzed trans-
esterification permits the controlled establishment of chemical
links at the interface between two epoxy networks. The key
parameters controlling the kinetics and final strength of the
welding are the concentration and the nature of the
transesterification catalyst and the concentration of hydroxyl
groups present in the networks. As for silica glass, the welding
process is robust and can be achieved in different time−
temperature windows without any changes in the material
dimensions. Catalysis brings additional control and flexibility of
process that do not exist for thermoplastics. The concept,
already illustrated here for both soft and hard epoxy networks,
can be readily extended to other polyester networks33 and other
systems that exhibit exchange reactions.
ASSOCIATED CONTENT
* Supporting Information
■
S
Curing main reactions; synthesis of networks; synthesis and
NMR and GC studies of model molecules; lap-shear tests; and
FTIR spectra of hard networks. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
ludwik.leibler@espci.fr
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge funding from ESPCI, CNRS, and Arkema. We
are grateful to Croda for providing Pripol1040. We are
indebted to Laura Breucker for help in epoxy−anhydride
network experiments.
REFERENCES
■
(1) (a) Wu, S. Polymer Interfaces and Adhesion; Marcel Dekker: New
York, 1982. (b) Wool, R. P. Polymer Interfaces: Structure and Strength;
Hanser: Munich, 1994.
́
(2) (a) Guerin, G.; Mauger, F.; Prud’homme, R. E. Polymer 2003, 44,
7477. (b) Akabori, K.-I.; Baba, D.; Koguchi, K.; Tanaka, K.; Nagamura,
T. J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 3598.
(3) Ferry, J. Viscoelastic Properties of Polymers; Wiley: New York,
1980.
(4) (a) Outwater, J. O.; Gerry, D. J. J. Adhes. 1969, 1, 290.
(b) Rahmathullah, M. A. M.; Palmese, G. R. J. Appl. Polym. Sci. 2009,
113, 2191.
(5) Yamaguchi, M.; Ono, S.; Terano, M. Mater. Lett. 2007, 61, 1396.
(6) (a) Hayes, S. A.; Jones, F. R.; Marshiya, K.; Zhang, W. Composites,
Part A 2007, 38, 1116. (b) Luo, X. F.; Ou, R. Q.; Eberly, D. E.;
Singhal, A.; Viratyaporn, W.; Mather, P. T. ACS Appl. Mater. Interfaces
2009, 1, 612. (c) Peterson, A. M.; Kotthapalli, H.; Rahmathullah, M. A.
M.; Palmese, G. R. Compos. Sci. Technol. 2012, 72, 330.
(7) Troughton, M. J. Handbook of Plastics Joining: A Practical Guide,
2nd ed.; William Andrew: Norwich, NY, 2008.
(8) (a) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K.
M.; Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 898. (b) Maeda, T.;
Otsuka, H.; Takahara, A. Prog. Polym. Sci. 2009, 34, 581.
(9) (a) Wojtecki, R. J.; Meador, M. A.; Rowan, S. J. Nat. Mater. 2011,
10, 14. (b) Bergman, S. D.; Wudl, F. J. Mater. Chem. 2008, 18, 41.
(c) Kloxin, C. J.; Scott, T. F.; Adzima, B. J.; Bowman, C. N.
Macromolecules 2010, 43, 2643.
(27) (a) Craun, G. P. J. Coat. Technol. 1995, 67, 841. (b) Matejka, L.;
Pokorny, S.; Dusek, K. Polym. Bull. 1982, 7, 123.
(28) (a) Otera, J. Chem. Rev. 1993, 93, 1449. (b) Craun, G. P.; Kuo,
C. Y.; Neag, C. M. Prog. Org. Coat. 1996, 29, 55.
(29) (a) Awaja, F.; Gilbert, M.; Kelly, G.; Fox, B.; Pigram, P. J. Prog.
Polym. Sci. 2009, 34, 948. (b) Maes, F.; Montarnal, D.; Cantournet, S.;
́
Tournilhac, F.; Corte, L.; Leibler, L. Soft Matter 2012, 8, 1681.
(30) Stevens, G. C. J. Appl. Polym. Sci. 1981, 26, 4259.
(31) (a) Matejka, L.; Lovy, J.; Pokorny, S.; Bouchal, K.; Dusek, K. J.
Polym. Sci., Polym. Chem. 1983, 21, 2873. (b) Dusek, K.; Lunak, S.;
Matejka, L. Polym. Bull. 1982, 7, 145. (c) Mauri, A. N.; Galego, N.;
Riccardi, C. C.; Williams, R. J. J. Macromolecules 1997, 30, 1616.
(d) Chian, W.; Timm, D. C. Macromolecules 2004, 37, 8098.
(32) (a) Mauri, A. N.; Riccardi, C. C. J. Appl. Polym. Sci. 2002, 85,
2342. (b) Corcuera, M. A.; Mondragon, I.; Riccardi, C. C.; Williams,
R. J. J. J. Appl. Polym. Sci. 1997, 64, 157.
(33) (a) Frich, D.; Goranov, K.; Scheggenburger, L. A.; Economy, J.
Macromolecules 1996, 29, 7734. (b) Frich, D.; Economy, J.; Goranov,
K. Polym. Eng. Sci. 1997, 37, 541. (c) Frich, D.; Hall, A.; Economy, J.
Macromol. Chem. Phys. 1998, 199, 913.
(10) (a) Green, M. S.; Tobolsky, A. V. J. Chem. Phys. 1946, 14, 80.
(b) Rubinstein, M.; Semenov, A. N. Macromolecules 1998, 31, 1386.
7667
dx.doi.org/10.1021/ja302894k | J. Am. Chem. Soc. 2012, 134, 7664−7667