Mechanically Robust, Self-Healable, and Highly Stretchable “Living” Crosslinked Polyurethane Based on a Reversible C?C Bond

Article abstract of DOI:10.1002/adfm.201706050

Stimuli-responsive polymers built by reversible covalent bonds used to possess unbalanced mechanical properties. Here, a crosslinked polyurethane containing aromatic pinacol as a novel reversible C?C bond provider is synthesized, whose tensile strength an

Full text of DOI:10.1002/adfm.201706050

FULL PAPER
Stimuli-Responsive Materials
www.afm-journal.de
Mechanically Robust, Self-Healable, and Highly Stretchable
“Living” Crosslinked Polyurethane Based on a Reversible

C C Bond
Ze Ping Zhang, Min Zhi Rong,* and Ming Qiu Zhang*
Here, we precisely design and synthesize
a crosslinked polyurethane containing a
Stimuli-responsive polymers built by reversible covalent bonds used to
possess unbalanced mechanical properties. Here, a crosslinked polyurethane

novel reversible C C bond, whose ten-

containing aromatic pinacol as a novel reversible C C bond provider is syn-
thesized, whose tensile strength and failure strain are tunable from 27.3 MPa
to as high as 115.2 MPa and from 324% to 1501%, respectively, owing to the
sile strength and failure strain are tunable
from 27.3 MPa to 115.2 MPa and from
324% to 1501%, respectively (Figure 1a).
The strategy offers a solution to the above
dilemma. In addition, the polyurethane
exhibits unusual “living” ability for initi-
ating polymerization as well as self-heala-
bility and reprocessability.

relatively high bond energy of the C C bond of pinacol as well as the
hydrogen bond between hard segments and semicrystalline soft segments.
Moreover, the dynamic equilibrium of pinacol enables self-healing and recy-
cling of the polymer. Interestingly, the dynamic exchange among macromol-
ecules, for the first time, successfully cooperates with solid-state drawing that
applies to thermoplastics, realizing strengthening of thermoset. Meanwhile,
the radicals derived from homolysis of pinacol can repeatedly initiate polym-
erization of vinyl monomers. The fruitful outcomes of this work may create a
series of promising new techniques.

C C bond has high bond energy
(e.g., 377.4 kJ mol−1 for ethane[2a]) and has
been used to form the basic frameworks
of polymers. Moreover, its bond energy
decreases with increasing bond length
due to steric and electronic effects, leading
to diverse properties.[3] It would thus be
a reasonable starting point to construct
reversibly crosslinked polymer with balanced mechanical per-
1. Introduction

formance by means of C C bond, so long as the latter becomes
Recently, quite a few reports prove that polymer networks con-
structed by reversible covalent bonds are able to be self-healed,
remolded, and recycled through reshuffling and rearrangement
of crosslinkages.[1] The behaviors not only differ from those of
traditional irreversible covalently bonded thermosets, but also
somewhat resemble thermoplastics. A series of new materials
with smart functionalities might thus be developed into next-
generation technology.
reversible under certain stimuli. Imato et al.,[4] for example,
prepared crosslinked polypropylene glycol-based polyurethane
gel with diarylbibenzofuranone as crosslinker. Intrinsic self-
healing of the gel material was enabled by dynamic reversible

equilibrium between C C linkage and carbon radicals. Because
of the nature of gel material, however, the resultant polymer
possesses low strength (0.1–0.8 MPa).[4]
Keeping this in mind, we turn to the derivatives of

With respect to the polymers networked by reversible cova-
lent bonds for structural applications, simultaneous integra-
tion of high strength and toughness is necessary. The materials
publicized so far, however, have either higher strength but low
extensionality or higher elongation-at-break but low strength
(Figure 1a), due to the relatively low bond energies of the
reversible bonds and/or fragile macromolecular structures.
1,2-disubstituted tetraphenylethanes (with central C C
bond energy of 247.3 kJ mol−1[2b]), and dihydroxyltetraphe-
nylethane (also called as aromatic pinacol) is chosen as it can
produce carbon radicals upon moderate heating/UV radia-
tion. Aromatic pinacol used to serve as initiator of free radical
polymerization,[5,6] but its dynamic reversible behavior has not
yet been dealt with. In addition to bond energy of reversible
bonds, molecular structure is another key factor that deter-
mines polymer properties. Comparatively, polyurethane is
more promising as a result of its unique soft/hard microphase
separation and strong intermolecular hydrogen bonding.
Dr. Z. P. Zhang, Prof. M. Z. Rong, Prof. M. Q. Zhang
Key Laboratory for Polymeric Composite and
Functional Materials of Ministry of Education
GD HPPC Lab
School of Chemistry
Sun Yat-sen University
Figure 2a shows the chemical structure of the target polymer
of this work (P1), which is synthesized by introducing aromatic
pinacol unit (Figure 2b) into the main chain of polyurethane
following the above consideration (Figure 2c). The core steps
lie in the synthesis of hydroxyl-functionalized aromatic pinacol
unit, and incorporating it into crosslinked polyurethane. Espe-
cially, hydroxyethoxys are respectively attached to the two phenyl
groups of aromatic pinacol unit, generating a new monomer
Guangzhou 510275, China
E-mail: cesrmz@mail.sysu.edu.cn; ceszmq@mail.sysu.edu.cn
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adfm.201706050.
DOI: 10.1002/adfm.201706050
©
Adv. Funct. Mater. 2018, 1706050
1706050 (1 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
aromatic pinacol perform so fast that the carbon radicals

derived from fission of C C bonds cannot be acquired. As a
result, a certain amount of 4-hydroxy-2,2,6,6-tetramethylpi-
peridin-1-oxyl (4-OH-TEMPO) is added to capture the radicals.
It is found that the onset homolysis temperature of the poly-
urethane containing DHETPED is roughly 80 °C, similar to
the value of DHETPED (Figure 3a–c). The thermally induced
dissociation of the pinacol units and recombination of carbon
radicals in polymer becomes more prominent at elevated tem-
perature (>80 °C). Meanwhile, it is found that the homolysis
reaction follows the first order kinetics, and the activation
energy of dissociation, Ed, is ≈112 kJ mol−1 (Figure S22, Sup-
porting Information). Theoretically, the bond dissociation
energy (BDE) of 1,2-disubstituted tetraphenylethane derivative
with similar molecular structure is 167.0 kJ mol−1 (Table S1,
Supporting Information), which is greatly affected by elec-
tron-withdrawing effect. Aromatic pinacol has a modest BDE
favoring (i) properties improvement of the polymer, and
(ii) reversible reactions. By taking advantage of the attractive
dissociation/recombination of carbon radicals, reshuffling of
the polyurethane chains containing aromatic pinacol structure
is enabled as demonstrated by the molecular weight change
of another linear control of P1, P4 (Figure S21, Supporting
Information), after heat treatment at 100 °C (Table S2 and
Figure S23, Supporting Information).
The dynamic reversible characteristics of P1, which is inher-
ited from aromatic pinacol moieties, can also be reflected by
stress relaxation tests. As shown in Figure 3d, there is a com-
plete stress relaxation for P1 from 80 to 120 °C. Obviously,
accelerated homolysis at elevated temperatures leads to faster
relaxation. In contrast, the control P2, which has nearly the
same compositions as P1 but does not contain any reversible
bonds (Figure S19, Supporting Information), can only reach
a steady state like conventional irreversibly crosslinked mate-
rials (Figure S24a, Supporting Information). It suggests that
the networked chains of P1 are allowed to be rearranged to a
less stretched state owing to bond fission/recombination of
the aromatic pinacol units. Similar results were found in other
polymers carrying dynamic reversible bonds.[7]
Since the relaxation is mainly controlled by the dynamic
reversible reaction, temperature dependence of the charac-
teristic relaxation time, τ* (determined at 1/e of the normal-
ized relaxation stress, Figure S24b, Supporting Information),
should be described by the Arrhenius law.[1u] Figure S24c in
the Supporting Information exhibits that it is the case, and
an activation energy, Ea, of 103 kJ mol−1 is estimated from
the slope, which is at the medium level of the reported values
(55–150 kJ mol−1[1o,p,t–v,8]) of dynamic exchange reactions. The
result agrees with the original design of P1. That is, the Ea should
not be too low to result in continuous structural instability at
room temperature. Besides, topology-freezing transition tem-
perature (at which reversible transition from viscoelastic solid to
liquid occurs and the viscosity reaches 1012 Pa∙s[9]), Tv, of P1 can
be estimated to be around 22 °C by extrapolation.[10] The value
is higher than Tg of P1 (−51.6 °C, refer to Table S3, Supporting
Information). It is seen from Figure S24d in the Supporting
Information that the P1 networks display like a strong glass
former similar to silica and other vitrimers,[1o,p,t–v,8,9] which is
different from conventional linear polymers such as polystyrene.
Figure 1. Mechanical properties. a) Dependences of tensile strength
on elongation-at-break of polymers containing different reversible cova-
lent bonds. PB, poly(butadiene); PU, polyurethane; PDS, polydisulfide;
PDMS, poly(dimethyl siloxane); MMA, methyl methacrylate; PUU,
poly(urea-urethane); PCO, polycyclooctene; PLA, poly(lactic acid); PHU,
poly(hydroxyurethane); VU, vinylogous urethane; PTIL, poly(1,2,3-triazo-
lium ionic liquid). b) Tensile strengths and elongation-at-break of solid-
state drawn P1 along and perpendicular to the drawing direction as a
function of drawing ratio.
(1,2-bis(4-(2-hydroxyethoxy)phenyl)-1,2-diphenylethane-1,2-diol,
DHETPED), which first reacts with the prepolymer of hexa-
methylene diisocyanate (HDI) and polytetramethylene ether
glycol (PTMEG), and then the polymer is crosslinked by polyols
crosslinker (Trisdiol) (Figure S17, Supporting Information).
Hereinafter, structure–properties relationship of P1 is care-
fully characterized and discussed, in hopes of exploring the

opportunities offered by homolysis reaction of C C bond
(Figure 2b) for developing novel smart materials.
2. Results and Discussion
2.1. Dynamic Reversibility
Although aromatic pinacol is a traditional radical initiator, it is

unclear whether its C C bond belongs to reversible bonds. To
reveal whether the pinacol units in P1 exist in a state of equilib-
rium of cleavage and recombination, electron spin resonance
(ESR) spectroscopy measurements are conducted. The results
indicate that the disconnection and recombination of the
©
Adv. Funct. Mater. 2018, 1706050
1706050 (2 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
Figure 2. a) Chemical structure of P1. b) Dynamic reversible reaction of aromatic pinacol. c) Design of the reversible polyurethane.
Therefore, no precise temperature control is needed for the
reversible reaction-induced processing of P1.
contrast to the case of P2, where no chemical interaction is
allowed. Moreover, the overlapped AFM pull-off curves of P1
collected during the repeated tests (Figure 4a) are indicative of
reversibility of the self-bonding ability in the same place. On the
other hand, visual inspection shows that the penetrating crack
on a film sample of P1 fades away after treatment at 100 °C
(Figure 4b–d), but the similar crack on P2 sample remains after
the same treatment (Figure S28, Supporting Information).
It is known that the capability of self-healing can also be
assessed by reprocessability.[11] As shown in Figure 5, the
specimen of P1 is first pulverized, and then the powers are
compression molded at a temperature near or higher than the
onset homolysis temperature (i.e., 80 °C, Figure 3c). Owing to
2.2. Autonomous Healing and Reprocessing Capabilities
The reshuffling and rearrangement of networks of P1 revealed
hereinbefore must benefit crack healing in principle. The
deduction is verified by both micro- and macroscopic experi-
ments as follows. The autohesion of P1 measured by atomic
force microscopic (AFM) pull-off curves (Figure 4a) indicates
that the adhesive energy of P1 at 100 °C is 6.4 mJ m−2, about
eight times higher than that at 25 °C (i.e., 0.78 mJ m−2). In
the meantime, the adhesive energy of control P2 at 100 °C is
2.8 mJ m−2, only four times higher than that at 25 °C
(i.e., 0.65 mJ m−2) (Figure S27, Supporting Information). Evi-

the disconnection and reconnection of the C C bonds, cova-
lent bonding can be re-established across the interfaces of the
fragmented P1. As a result, a compact bulk specimen is pro-
duced once more. By using dynamic microhardness tests,
elastic moduli ratio of the recycled and virgin specimens gives

dently, rapid exchange reaction of dynamic C C bonds could
take effect between P1 chains themselves. It forms striking
©
Adv. Funct. Mater. 2018, 1706050
1706050 (3 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
a measure of recovery efficiency. It is seen from Figure 4e,f that
almost complete recovery (93.6%) is achieved by hot pressing
under 3 MPa at 80 °C for 7 h, while 119.1% recovery is detected
at 100 °C for 3 h. Compared with the as-prepared sample
obtained from solution casting, hot pressing improves tap den-
sity of the molded specimen, so that elastic modulus of the recy-
cled material is slightly higher than the original value. On the
whole, the recovery efficiency increases with hot pressing time
at constant temperature, or it increases with molding tempera-
ture at constant pressing time. Clearly, the effect of recycling
a)
1
1: 30 oC
2: 60 oC
3: 70 oC
4: 80 oC
5: 90 oC
6: 100 oC
7: 110 oC
2
4
3
5
6
7

is controlled by the kinetics of C C bonds fission/recombina-
tion. The conclusion is supported by the reference experiment,
which shows when the powered P1 is compressed at
60 °C for as long as 48 h, no compact recycled specimen can be
obtained due to the very low homolysis rate of aromatic pinacol
units (Figure S22d, Supporting Information).
3330
3360
3390
3420
Magnetic field (G)
b)
1
1: 30 oC
2: 60 oC
3: 90 oC
4: 110 oC

It is worth noting that the dynamic reversible C C bonds
2
can do much more than self-healing and solid-state repro-
cessing. For example, the insoluble crosslinked P1 can be trans-
formed into solution. Figure 6 shows that the specimen of P1
is only swollen in chloroform at room temperature like conven-
tional thermosets, but it became completely dissolved within
2 h by heating to the onset homolysis temperature of aromatic
pinacol units (i.e., 80 °C). A transparent film is thus obtained
by casting. The result suggests a possible way for recycling
in solution by making use of thermally dissociable aromatic
pinacol units.
3
4
3315
3330
3345
3360
3375
3390
3405
Magnetic field (G)
2.3. Tuning Mechanical Properties by Solid-State Drawing
c)
The dynamic reversibility of P1 actually implies another
important possibility—properties’ regulation of crosslinked
polymers after curing, which is impossible when the
crosslinked network is full of irreversible bonds. That is,
solid-state drawing, which used to be applied to thermoplas-
tics for introducing ordered orientation and strengthening
the polymers,[12] may be employed to increase strength of
the reversibly crosslinked polyurethane. In general, the tech-
nique is not suitable for crosslinked networks because elon-
gation of the macromolecules is greatly restricted. In some
cases, thermoplastics are first extended at rubbery-like state
to improve their strength, and then crosslinked to fix the
structure.[13] To the best of our knowledge, however, there has
not been any attempt to bring in orientation in crosslinked
DHETPED / 4-OH-TEMPO
P3 / 4-OH-TEMPO
20
40
60
80
100
120
Temperature (oC)
4.0
d)
70 oC
80 oC
90 oC
100 oC
120 oC

polymer. Considering the fact that the C C bonds in aro-
matic pinacol units would be disconnected and reconnected
3.2
2.4
1.6
0.8
0.0
above the homolysis temperature, we believe that when P1 is
Figure 3. Characterization of dynamic reversibility. a) Typical ESR spectra
of DHETPED/4-OH-TEMPO measured at various temperatures. b) Typ-
ical ESR spectra of P3 (refer to Figure S20, Supporting Information)
mixed with 4-OH-TEMPO measured at various temperatures. c) Tem-
perature dependences of relative ESR signal intensities calculated from
integral areas of the normalized absorption curves of Figure S22a,b in
the Supporting Information. P3: a linear control of P1 for being able to
be dissolved in DMF as required by ESR test. 4-OH-TEMPO is added as
a radical spin-trapping reagent. d) Stress relaxation of P1 measured at
various temperatures.
2.0x103
4.0x103
6.0x103
8.0x103
1.0x104
0.0
Time (s)
©
Adv. Funct. Mater. 2018, 1706050
1706050 (4 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
macromolecules imposed by the reversible crosslinks is

released due to fission of C C bonds, the relevant chains
must be able to be stretched, which has already been proved
by the stress relaxation test (Figure 3d).
Figure 1b shows that the strength parallel to the drawing
direction, σ//, of P1 specimen indeed increases with increasing
the drawing ratio at 80 °C, and approaches to 115.2 MPa at
drawing ratio of 19. Interestingly, because the crosslinks in
the direction vertical to drawing can be synchronously re-
established, the strength in this direction, σ⊥, remains nearly
unchanged. The phenomenon differs from thermoplastics.
σ// of the solid-state drawn PP,[14] for instance, is 385 MPa
(≈9.6 times higher than the strength of undrawn PP (≈40 MPa))
at drawing ratio of 17, but its σ⊥ is only 4.5 MPa (≈11% of the
strength of undrawn PP). As for failure strain, its value along
the drawing direction, ε//, decreases with drawing ratio, which
is typical for orientation. Nevertheless, the failure strain in the
vertical direction, ε⊥, is also nearly independent on drawing
ratio, so that it is even higher than ε// with increasing drawing
ratio. Such an anisotropy of strength and failure strain has not
been reported before. Also, the above enhancement due to ori-
entation would not decay with increasing temperature so long
as the homolysis of aromatic pinacol is not activated. Figure S29
in the Supporting Information shows the stress–strain curves
of P1 that had been stretched at 80 °C under different drawing
ratios, which reveal more details of the effect of solid-state
drawing on the polymer's mechanical properties.
2.4. “Living” Polymer Network
In fact, carbon radicals derived from aromatic pinacol units
have long been utilized to initiate polymerization of vinyl
monomers. The diarylhydroxymethyl-based radicals do not
directly react with vinyl monomer through addition, but
conduct a secondary reaction by hydrogen transfer forming
Figure 4. Self-healing performance. a) Typical AFM pull-off curves
recorded during repeated approach and retraction of probes. The probes
are functionalized by P1, which is the same as the substrate material to
be tested. b–d) Photos showing thermal repair of P1 film: b) an artificial
crack is created; c,d) effect of treatment at 100 °C for c) 30 min and
d) 60 min. e,f) Load–unload test results of P1 and its recycled version
measured by dynamic microhardness tester. Loading force = 1.5 mN.
Compression pressure for recycling = 3 MPa.
drawn under the reversible circumstances, equilibrium of the

dynamic reversible reaction of C C bonds would be gradually
rebuilt up throughout the entire specimen in case the time is
long enough. In this context, as soon as the restriction on any
Figure 5. a) Original specimen of P1. b) Powdered P1 obtained by frozen
pulverization. c) Mold for reprocessing. d) Remolded specimen of P1.
Conditions of reprocessing: 100 °C, 2 h, and 3 MPa.
©
Adv. Funct. Mater. 2018, 1706050
1706050 (5 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
stretched above homolysis temperature to
largely improve the strength in the direction
of drawing. Moreover, the reversible cova-
lent chemistry imparts unusual properties
to the polyurethane. Tensile strength and
failure strain in the direction perpendicular
to drawing, for example, are almost the same
as the undrawn one. Meanwhile, the bulk
polymer itself is allowed to repeatedly initiate
polymerization of vinyl monomers by using
the carbon radicals from fission of the revers-
Figure 6. Optical images of P1 a) immersed in chloroform at beginning, b) swollen by chloro-
form at room temperature after 48 h, c) dissolved in chloroform at 80 °C for 2 h, and d) a fresh
transparent film obtained by casting of panel (c).

ible C C bonds in the absence of catalyst,
monomer radicals that initiate the polymerization (Figure S30,
Supporting Information).[15]
producing semi-interpenetrating polymer networks and grafting
polymer. The outcomes of this study reveal quite a few attractive
prospects of polymer engineering based on reversible covalent
chemistry, which goes beyond the scope of classic polymer engi-
neering and enriches the measures of material diversification.
According to this habit, P1 is dissolved in chloroform at
80 °C, and then different amounts of methyl methacrylate or
styrene monomers are added for polymerization at 100 °C for
24 h. The preliminary study evidences that the polymerization
can proceed as expected (Figures S31 and S32, and Table S4,
Supporting Information) and the polymerized product can
be repeatedly used as macromolecular initiator (Figure S32c,
Supporting Information). Although further investigation is
required for deeper understanding of the structure and proper-
ties of the polymerized products, it can be reasonably deduced
that they are semi-interpenetrating networks consisting of vinyl
monomer-grafted P1 and homopolymer of vinyl monomers
(Figure S30, Supporting Information). Since not all the aro-
matic pinacol units are exhausted and the grafted copolymer
also has the initiating ability, the polymerization of vinyl
monomers initiated by P1 exhibits living radical polymerization
feature to some extent.
Compared with existing control/living radical polymeriza-
tion method represented by atom transfer radical polymeriza-
tion (ATRP),[16] the polymerization initiated by P1 can be con-
ducted in air and does not require any catalyst. The detrimental
effects of ATRP including reduction in ageing-resistance of the
resultant polymer and toxicity ascribed to Cu ions no longer
exist. The new finding is not just about a new way of poly-
merization initiation. It probably results in a few novel tech-
niques. For example, by combining microencapsulated vinyl
monomers with the bulk P1 that becomes “living” above 80 °C,
one may develop intrinsic self-healing coupled with extrinsic
self-healing, micro-/nanopatterned surface, finely selective
functionalization, and gecko-like autonomous regrowing
mechanism.
4. Experimental Section
Materials and Reagents: 4-Hydroxybenzophenone (HBP, 98%),
1,3,5-triacryloylhexahydro-1,3,5-triazine (98%), 2-bromoethanol (95%),
4,4′-dihydroxytetraphenylmethane (DHTPM, 98%), 1-thioglycerol (98%),
and 1-hexylamine (99%) were purchased from Tokyo Chemical Industry
Co., Ltd. HDI (99%), PTMEG (Mn = 1000 and 2000 g mol−1), potassium
carbonate (99%), isopropyl alcohol (99.5%), acetic acid (99.5%), and
dibutyltindilaurate (DBTDL, 95%) were supplied by Aldrich. PTMEGs
were dehydrated at 100 °C under vacuum for more than 24 h prior to
use. N′N-dimethylformamide (DMF, 99.9%) was dried by activated
molecular sieves (4 Å) before usage. All the other chemicals and
solvents were used as received.
Synthesis of 4-(2-Hydroxyethoxy) Benzophenone (HEBP): A typical process
of synthesis of HEBP is described as follows (Figure S1, Supporting
Information). HBP (7.929 g, 40 mmol) was completely dissolved in
20 mL DMF and heated up to 90 °C. Then, 2-bromoethanol (7.497
g, 60 mmol) was slowly dropped into the solution from a dropping
funnel, followed by adding anhydrous potassium carbonate (11.057 g,
80 mmol) as the catalyst and stirring in argon for 24 h. Afterward,
the yellow mixture was cooled down and filtered for removal of
the solid catalyst. The filtrate was precipitated in deionized water,
followed by vacuum suction filtration and freeze-drying to obtain
white crystal with
a yield of 50%. Fourier transform infrared
spectroscopy (FTIR) (KBr, Figure S2, Supporting Information):

ν = 3291 (ν(OH)), 3051 (ν( CH)), 2949 (νas(CH2)), 2877 (νs(CH2)),


1640 (ν(C O)), 1603–1443 (ν(C C)), 1417 (δ(CH2)), 1378 (δ(OH)),





1254 (ν(C O)), 1094 (νs(C O C)), 1052 (ν(C O)), 847 (γ( CH)),
693 cm−1 (γ(OH)). Proton nuclear magnetic resonance spectroscopy
(1H NMR) (400 MHz, dimethyl sulfoxide-d6 (DMSO-d6), δ, Figure S3a,
Supporting Information): 3.754 (q, J = 5.2 Hz, 2H; CH2), 4.098 (t, J = 1.3 Hz,
2H; CH2), 4.946 (t, J = 5.6 Hz, 1H; OH), and 7.394–7.763 (m, 9H; CH of
the benzene ring). Carbon-13 nuclear magnetic resonance spectroscopy
(13C NMR) (400 MHz, DMSO-d6, δ, Figure S3b, Supporting Information):
59.639, 70.093, 114.713, 129.014, 129.696, 132.509, 132.654, 137.957,
162.956, and 194.942. Mass spectrum (ESI, Figure S4, Supporting
Information): m/z: [M + H]+ calculated for C15H14O3 243.2; found, 475.2593.
Synthesis of 1,2-Bis(4-(2-Hydroxyethoxy)Phenyl)-1,2-Diphenylethane-
1,2-Diol (DHETPED): Typically, DHETPED was synthesized as follows
(Figure S5, Supporting Information) by referring to literature.[17] HEBP
(4.854 g, 20 mmol) was completely dissolved in 20 mL DMF in a
50 mL quartz tube, and then isopropanol (30 mL) and two drops of
acetic acid were added. After homogeneous mixing, the solution was
exposed to ultraviolet light (350 nm, 3–4 mW cm−3) for 4–6 d and poured
into deionized water to obtain crude product. It was further purified by
column chromatography (silica gel) eluting with 1:1 petroleum ether-ethyl
3. Conclusions
Aromatic pinacol, a traditional radical initiator, is found to be

qualified as a novel reversible unit possessing adequate C C
bond energy. It can be cleft and rebonded under moderate
conditions. Accordingly, a simple and effective toolbox for pre-
paring thermo-triggered dynamic reversible crosslinked polyu-
rethane with excellent strength and toughness over the existing
reversible covalent systems is proposed by introducing aromatic
pinacol units with giant steric-hindrance effect into the main
chains. The crosslinked polyurethane cannot only be self-healed
and reprocessed through reshuffling of networks, but also be
©
Adv. Funct. Mater. 2018, 1706050
1706050 (6 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
acetate (v/v), achieving a yield of 31.9%. FTIR (KBr, Figure S6, Supporting
Information): ν = 3571 (ν(OH)), 3556 (ν(OH)), 3513 (ν(OH)), 3051
followed by stirring using a mechanical stirrer in argon for 6 h. Afterward,
the prepolymer solution was cooled down to 45 °C and reacted with
DHETPED (1.216 g, 2.5 mmol) and Trisdiol (0.478 g, 0.8 mmol) for
additional 18 h. At last, the mixture was poured into a silicone mold,
and then vacuum dried to obtain the material for the subsequent
characterization. FTIR (ATR, Figure S18a, Supporting Information):


(ν( CH)), 2926 (νas(CH2)), 2870 (νs(CH2)), 1605–1443 (ν(C C)),

 
O C)),
1418 (δ(CH2)), 1374 (δ(OH)), 1248 (ν(C O)), 1074 (νs(C
1035 (ν(C O)), 825 (γ( CH)), and 707 cm (γ(OH)). 1H NMR
(400 MHz, DMSO-d6, δ, Figure S7a,b, Supporting Information): 4.806
(s, 2H; OH) and 5.745 (t, J = 5.6 Hz, 2H; OH), 3.659 (q, J = 5.2 Hz,
4H; CH2), 3.865 (t, J = 5.2 Hz, 4H; CH2), and 6.605–7.282 (m, 18H;
CH of the benzene ring). 13C NMR (400 MHz, DMSO-d6, δ, Figure S7c,
Supporting Information): 59.867, 69.626, 82.936, 112.304, 125.998,
126.467, 129.542, 130.901, 138.913, 147.349, and 157.040. Mass
spectrum (API-ES, Figure S8, Supporting Information): m/z: [M + Na]+
calculated for C30H30O6 509.5; found, 509.0.
−1


ν = 3348 (ν(NH)), 2934 (νas(CH2)), 2854 (νs(CH2)), 2797 (ν(CH2)),
−1



1721 (ν(C O)), 1682–1445 (ν(C C)), 1248 (ν(C N)), and 1103 cm


(νs(C O C)). The control material that excludes pinacol moieties, HDI/
PTMEG2000/DHETPM/Trisdiol (P2), was prepared following similar
procedures by replacing DHETPED with DHETPM (Figure S19, Supporting
Information). The chemical structure of the product was also proved by
FTIR spectrum (Figure S18b, Supporting Information). When PTMEG2000
diol in P1 was replaced by PTMEG1000 diol, linear polyurethane without
the crosslinker (Trisdiol), HDI/PTMEG1000/DHETPED (P3), can be
prepared by the similar method (Figures S18c and S20, Supporting
Information). Moreover, the linear polyurethane containing higher
proportion of aromatic pinacol units and much better solubility in DMF,
HDI/DHETPED (P4) and its references HDI/DHETPM (P5) and HDI/
PEG400 (P6) were also synthesized (Figure S21, Supporting Information).
Characterization: 1H NMR and 13C NMR spectra were measured by
an AVANCE III 400 MHz with DMSO-d6 as solvent. Mass spectra were
obtained by Thermofisher LTQ Orbitrap Elite spectroscopy. Molecular
weights were determined at ambient temperature using Waters Breeze
gel permeation chromatography (GPC) system with DMF as eluent and
polystyrene standards for calibration. Differential scanning calorimetry
(DSC) measurements were performed on a TA Instruments DSCQ10
using a nitrogen purge and an empty aluminum pan as reference.
Dynamic mechanical analysis (DMA) tests were conducted on a METRA-
VIBDMA25 using tension mode under 1 Hz at a heating rate of 3 °C
min−1 in nitrogen. FTIR spectra were recorded by a Nicolet NEXUS 670.
ESR Experiments: ESR spectroscopy study was carried out on a
Bruker A300-10-12 spectrometer equipped with nitrogen heating
setup operating at 8.85 Hz. Modulation frequency and amplitude were
100 kHz and 0.1 mT, respectively. In the case of pinacol monomer
DHETPED, 0.243 g DHETPED (0.1 mol L−1) and 0.009 g 4-OH-TEMPO
(0.01 mol L−1) were completely dissolved in 5 mL DMF and measured in
a sealed quantitative capillary tube. The decomposition ratio of pinacol
monomer was calculated from Equation (1)
Synthesis of 4,4′-Dihydroxyethyltetraphenylmethane (DHETPM):
DHETPM, a chain extender excluding reversible bond, was synthesized
by utilizing substitution reaction between DHTPM and 2-bromoethanol,
as shown in Figure S9 in the Supporting Information. DHTPM (3.524 g,
10 mmol) and 30 mL DMF were added to a 100 mL round bottom
flask equipped with a magnetic stir bar. Then, 2-bromoethanol (3.749 g,
30 mmol) was slowly incorporated, followed by adding excessive catalyst
anhydrous potassium carbonate (5.528 g, 40 mmol). The solution was
allowed to be stirred at 90 °C in argon for 24 h. After separating the
solid catalyst from the system by filtration, a pale yellow filtrate was
left and poured into deionized water. The precipitate was collected and
recrystallized in methanol, producing white powders (3.709 g, 84%).
FTIR (KBr, Figure S10, Supporting Information): ν = 3329 (ν(OH)), 3029


(ν( CH)), 3056 (ν( CH)), 2931 (νas(CH2)), 2871 (νs(CH2)), 1608–1445


 
(ν(C C)), 1419 (δ(CH2)), 1247 (ν(C O)), 1080 (νs(C O C)), 1052
−1
1


(ν(C O)), 827 (γ( CH)), and 704 cm (γ(OH)). H NMR (400 MHz,
DMSO-d6, δ, Figure S11a,b, Supporting Information): 4.832 (t, J = 4.4 Hz,
2H; OH), 3.663 (q, J = 4.0 Hz, 4H; CH2), 3.931 (t, J = 4.0 Hz, 4H;
CH2), and 6.633–7.298 (m, 18H; CH of the benzene ring). 13C NMR
(400 MHz, DMSO-d6, δ, Figure S11c, Supporting Information): 60.118,
63.533, 69.857, 113.680, 126.325, 127.803, 130.733, 132.188, 138.996,
147.281, and 157.019. Mass spectrum (ESI, Figure S12, Supporting
Information): m/z: [M + Na]+ calculated for C29H28O4 463.5; found, 463.3.
Synthesis of 1,1′ ,1″-(1,3,5-Triazinane-1,3,5-Triyl)Tris(3-(2,3-Dihydroxy-
propylthio)Propan-1-One) (Trisdiol): Trisdiol was synthesized following
[18]
ref.
(Figure S13, Supporting Information). 1,3,5-triacryloylhexahydro-
1,3,5-triazine (1.246 g, 5 mmol) and excessive 1-thioglycerol (1.947 g,
18 mmol) (nCC/nSH = 0.83) were added to a 50 mL round flask
equipped with a magnetic stirrer. Then, methanol (8 mL) was added,
and the turbid mixture was stirred at 30 °C for 15 min. By adding
a drop of 1-hexylamine as the accelerator for thiol–ene reaction,
the turbid mixture was quickly converted into clear solution within
seconds, which was then stirred at 30 °C for 24 h. Afterward, the
reaction mixture was filtered and vacuum distillated to obtain pale
yellow oily liquid, which was further purified by recrystallization from
methanol for five times, achieving white crystalline compound with
a yield of 79.7%. FTIR (KBr, Figure S14, Supporting Information):
n1 I −I
2n2I1
(
)
=
I1 −I2
20I1
1
2
(1)
X =
where I1 and I2 are calculated from integral areas of the normalized
absorption curves of 4-OH-TEMPO reference solution without
pinacol and DHETPED/4-OH-TEMPO mixed solution under different
temperatures, respectively. n1 and n2 denote the molar quantities of
4-OH-TEMPO and DHETPED, respectively. Similarly, in the case of linear
polyurethane P3, 0.1 g P3 (containing ≈5.4 × 10−5 mol L−1 pinacol units)
and 0.001 g 4-OH-TEMPO (1 × 10−6 mol L−1) were completely dissolved
in 50 mL DMF and then tested.
Stress Relaxation Studies: Stress relaxation behavior was recorded by
METRA-VIBDMA25 (DMA) using tension mode. The test was conducted
at a strain of 10% that was applied at t = 0. Activation energy of the
exchange reaction, Ea, was estimated from Equation (2)

ν = 3380 (ν(OH)), 2919 (νas(CH2)), 2868 (νs(CH2)), 1657 (ν(C O)), 1427
−1


(δ(CH2)), 1373 (δ(OH)), 1251 (ν(C N)), 1180 (ν(C O)), and 677 cm
(γ(OH)). 1H NMR (400 MHz, DMSO-d6, δ, Figure S15a,b, Supporting
Information): 2.418–2.833 (m, 18H; CH2), 3.169 (d, J = 5.2 Hz,
6H; CH2), 3.565 (sext, J = 5.2 Hz, 3H; CH), 4.556(t, J = 5.6 Hz, 3H; OH),
4.751 (d, J = 4.8 Hz, 3H; OH), and 5.257 (s, 6H; CH2). 13C NMR
(400 MHz, DMSO-d6, δ, Figure S15c, Supporting Information): 27.686,
33.376, 35.848, 55.929, 64.819, 71.832, and 171.031. Mass spectrum
(ESI, Figure S16, Supporting Information): m/z: [M + Na]+ calculated for
C21H39O9N3S3 596.7; found, 596.2. Anal. calculated for C21H39O9N3S3:
C 43.9, H 6.8, N 7.3, S 16.7; found: C 42.7, H 7.0, N 7.1, S 16.4.
(2)
lnτ *(T) = lnτ 0 +Ea/RT
where τ* is the characteristic relaxation time defined as the time
required for the stress relaxation stress to reach 37% (1/e) of its initial
value, τ0 is the characteristic relaxation time at infinite temperature, R is
the universal gas constant, T is the absolute temperature at which the
stress relaxation experiment was performed.
Synthesis of Polyurethanes: Preparation of pinacol-containing
polyurethane P1 (Figure S17, Supporting Information) was conducted
through solution polymerization in
a four-neck flask. PTMEG2000
(10.000 g, 5 mmol) was completely melted at 60 °C and mixed with
Crack Healing Process: Qualitative evaluation of crack healing was
carried out by visual inspection. The specimens of P1 and P2 were
quenched in liquid nitrogen, followed by striking to create penetrating
cracks. Then, thedamagedspecimens were healed at 100 °C for30–60min.
100mLDMFcontainingonedropofDBTDL. Then, HDI(1.766g, 10.5mmol,
n
NCO/nOH
= 1.05, nHDI/nPTMEG2000/nDHETPED/nTrisdiol = 2.1/1/1/0.66)
was slowly dropped into the mixed solution from a dropping funnel,
©
Adv. Funct. Mater. 2018, 1706050
1706050 (7 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
Photos that record changes of the crack were taken by a camera attached
to a KEYENCE VHX1000C 3D optical microscope.
5632; c) A. Rekondo, R. Martin, A. R. de Luzuriaga, G. Cabañero,
H. J. Grandea, I. Odriozola, Mater. Horiz. 2014, 1, 237;
d)W.M.Xu,M.Z.Rong,M.Q.Zhang,J.Mater.Chem.A2016,4,10683;
e) B. T. Michal, C. A. Jaye, E. J. Spencer, S. J. Rowan, ACS Macro
Lett. 2013, 2, 694; f) G. Deng, F. Li, H. Yu, F. Liu, C. Liu, W. Sun,
H. Jiang, Y. Chen, ACS Macro Lett. 2012, 1, 275; g) N. Roy, E. Buhler,
J. M. Lehn, Polym. Int. 2014, 63, 1400; h) Z. P. Zhang, Y. Lu,
M. Z. Rong, M. Q. Zhang, RSC Adv. 2016, 6, 6350; i) C. Yuan,
M. Z. Rong, M. Q. Zhang, Polymer 2014, 55, 1782; j) S. Ji, W. Cao,
Y. Yu, H. Xu, Adv. Mater. 2015, 27, 7740; k) Y. Zhang, H. Ying,
K. R. Hart, Y. Wu, A. J. Hsu, A. M. Coppola, T. A. Kim, K. Yang,
N. R. Sottos, S. R. White, J. Cheng, Adv. Mater. 2016, 28, 7646;
l) Y. Amamoto, H. Otsuka, A. Takahara, K. Matyjaszewski,
Adv. Mater. 2012, 24, 3975; m) J. C. Lai, J. F. Mei, X. Y. Jia, C. H. Li,
X. Z. You, Z. Bao, Adv. Mater. 2016, 28, 8277; n) O. R. Cromwell,
J. Chung, Z. Guan, J. Am. Chem. Soc. 2015, 137, 6492;
o) J. P. Brutman, P. A. Delgado, M. A. Hillmyer, ACS Macro Lett.
2014, 3, 607; p) D. Montarnal, M. Capelot, F. Tournilhac, L. Leibler,
Science 2011, 334, 965; q) Q. Tian, Y. C. Yuan, M. Z. Rong,
M. Q. Zhang, J. Mater. Chem. 2009, 19, 1289; r) S. Yu, R. Zhang,
Q. Wu, T. Chen, P. Sun, Adv. Mater. 2013, 25, 4912; s) W. X. Liu,
C. Zhang, H. Zhang, N. Zhao, Z. X. Yu, J. Xu, J. Am. Chem. Soc.
2017, 139, 8678; t) D. J. Fortman, J. P. Brutman, C. J. Cramer,
M. A. Hillmyer, W. R. Dichtel, J. Am. Chem. Soc. 2015, 137, 14019;
u) W. Denissen, G. Rivero, R. Nicolaÿ, L. Leibler, J. M. Winne,
F. E. Du Prez, Adv. Funct. Mater. 2015, 25, 2451; v) M. M. Obadia,
B. P. Mudraboyina, A. Serghei, D. Montarnal, E. Drockenmuller,
J. Am. Chem. Soc. 2015, 137, 6078.
AFM Tests: AFM pull-off force curves were recorded by using a Bruker
MultiMode 8 at 25 and 100 °C, respectively. First, a glass microsphere
(diameter ≈ 50 µm) was coated by the polymer to be tested. Then, the
functionalized glass microsphere was glued to the tipless AFM cantilever
(model: TESP, spring constant: 20–30 N m−1) with UV curing adhesive.
Finally, the functionalized glass microsphere probe was brought into
contact with the same polymer substrate (thickness ≈ 1 mm) for 500 ms
and retracted. The adhesive energy was calculated by Johnson–Kendall–
Roberts theory[19]
Wadh = −2Fa /3πr
(3)
where Wadh is the adhesion energy on the contact interface, Fa is the
adhesion force (maximum pull-off force), and r is the curvature radius of
the probe tip (i.e., radius of the glass microsphere).
Recycling Experiments: Recycling of the target polymer was conducted
by pulverizing the specimen into powders with the aid of liquid nitrogen.
Then, the powers were compression molded under 3 MPa at 80–100 °C
for different times. Finally, the recycled specimens were tested by a
Shimadzu DUH-W201S dynamic ultra-micro hardness tester via load–
unload mode (loading force = 1.5 mN, loading speed = 0.0711 mN s−1,
loading time = 5 s) at 25 °C. The recovery efficiency is defined as the
ratio of elastic modulus of recycled and virgin specimens.
Solid-State Drawing Tests: Solid-state drawing was conducted by
a Hounsfield T10K-S universal tester equipped with a temperature-
controlled cabinet. The dumbbell-shaped specimens (50 × 4 × 0.2 mm3,
grip to grip separation: 15 mm) were heated to 80 °C and then stretched
to different draw ratios at a crosshead speed of 10 mm min−1. Finally, the
specimens were cooled down to 25 °C. Tensile properties of the drawn
specimens were measured by the same instrument at room temperature
under a crosshead speed of 50 mm min−1.
[2] a) M. Brouard, P. D. Lightfoot, M. J. Pilling, J. Phys. Chem. 1986,
90, 445; b) H. D. Beckhaus, B. Dogan, J. Schaetzer, S. Hellmann,
C. Rüchardt, Eur. J. Inorg. Chem. 1990, 123, 137.
[3] C. Rüchardt, H. D. Beckhaus, Angew. Chem., Int. Ed. 1980, 19, 429.
[4] K. Imato, M. Nishihara, T. Kanehara, Y. Amamoto, A. Takahara,
H. Otsukaet, Angew. Chem., Int. Ed. 2012, 51, 1138.
Supporting Information
[5] D. Braun, Macromol. Symp. 1996, 111, 63.
[6] T. Otsu, A. Matsumoto, Adv. Polym. Sci. 1998, 136, 75.
[7] a) Z. Q. Lei, H. P. Xiang, Y. J. Yuan, M. Z. Rong, M. Q. Zhang,
Chem. Mater. 2014, 26, 2038; b) M. Pepels, I. Filot, B. Klumperman,
H. Goossens, Polym. Chem. 2013, 4, 4955.
Supporting Information is available from the Wiley Online Library or
from the author.
[8] a) A. R. de Luzuriaga, R. Martin, N. Markaide, A. Rekondo,
G. Cabañero, J. Rodríguez, I. Odriozola, Mater. Horiz. 2016, 3, 241;
b) N. Zheng, Z. Fang, W. Zou, Q. Zhao, T. Xie, Angew. Chem., Int.
Ed. 2016, 55, 11421.
[9] M. Capelot, M. M. Unterlass, F. Tournilhac, L. Leibler, ACS Macro
Lett. 2012, 1, 789.
Acknowledgements
The authors thank the support of the Natural Science Foundation of
China (Grant Nos.: 51673219 and 51333008).
[10] W. Denissen, J. M. Winne, F. E. du Prez, Chem. Sci. 2016, 7, 30.
[11] Y. Zhang, A. A. Broekhuis, F. Picchioni, Macromolecules 2009, 42,
1906.
[12] a) I. M. Ward, Makromol. Chem., Macromol. Symp. 1988, 22, 59;
b) I. M. Ward, Polymer 1974, 15, 379; c) P. R. Pinnock, I. M. Ward,
Br. J. Appl. Phys. 1964, 15, 1559.
Conflict of Interest
The authors declare no conflict of interest.
[13] a) D. Milicˇevic´, D. S. Trifunovic´, M. Popovic´, T. V. Milic´,
E. Suljovrujic´, Nucl. Instrum. Methods Phys. Res., Sect. B 2007, 260,
603; b) E. Suljovrujic, Eur. Polym. J. 2009, 45, 2068.
[14] B. Alcock, N. O. Cabrera, N. M. Barkoula, J. Loos, T. Peijs, Compos-
ites, Part A 2006, 37, 716.
[15] D. Braun, K. H. Becker, Ind. Eng. Chem. Prod. Res. Dev. 1971, 10, 386.
[16] K. Matyjaszewski, J. Xia, Chem. Rev. 2001, 101, 2921.
[17] J. T. Li, J. H. Yang, J. F. Han, T. S. Li, Green Chem. 2003, 5, 433.
[18] C. Rim, L. J. Lahey, V. G. Patel, H. Zhang, D. Y. Son, Tetrahedron
Lett. 2009, 50, 745.
Keywords

aromatic pinacol,
self-healing
C
C
bonds, mechanical properties, polymers,
Received: October 18, 2017
Revised: October 27, 2017
Published online:
[1] a) Y. X. Lu, Z. Guan, J. Am. Chem. Soc. 2012, 134, 14226;
b) K. Imato, A. Takahara, H. Otsuka, Macromolecules 2015, 48,
[19] K. L. Johnson, K. Kendall, A. D. Roberts, Proc. R. Soc. London, Ser. A
1971, 324, 301.
©
Adv. Funct. Mater. 2018, 1706050
1706050 (8 of 8)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim