polymers

Article
Recyclable and Mendable Cellulose-Reinforced
Composites Crosslinked with Diels–Alder Adducts
KeumHwan Park 1 , Cheolmin Shin 2 , Ye-Seul Song 1 , Hee-Jin Lee 1 , Chiho Shin 2 and
Youngmin Kim 1, *
1 Display Materials & Components Research Center, Korea Electronics Technology Institute, 25 Saenari-ro,
Bundang-gu, Seongnam-si, Gyeonggi-do 13509, Korea; khpark@keti.re.kr (K.P.); bp53224@daum.net (Y.-S.S.);
gmlwls5010@gmail.com (H.-J.L.)
2 Kwangsung Corporation Co., Ltd. 212-14, Neungan-gil, Songsan-myeon, Dangjin-Si,
Chungcheongnam-do 31711, Korea; cmshin@belite.co (C.S.); chshin@belite.co (C.S.)
* Correspondence: ymkim109f@gmail.com; Tel.: +82-31-789-7432




Received: 8 October 2018; Accepted: 6 January 2019; Published: 11 January 2019


Abstract: Owing to their natural abundance and exceptional mechanical properties, cellulose fibers
(CFs) have been used for reinforcing polymers. Despite these merits, dispersing hydrophilic CFs
in a hydrophobic polymer matrix is challenging. To address this, an amphiphilic ammonium
salt was employed as the dispersant for CFs in this study. The hydrophobic CFs were mixed
with a healable polymer to produce CF-reinforced composites. As the thermosetting polymer was
crosslinked with Diels–Alder (DA) adducts, it was mended and recycled via a retro DA reaction at
120 ◦ C. Interestingly, the CF-reinforced polymer composites were mended and recycled as well.
When 5 wt % of the hydrophobic CFs was added to the polymer, maximum tensile strength,
elongation at break, Young’s modulus, and toughness increased by 70%, 183%, 75%, and 420%,
respectively. After recycling, the CF-reinforced composites still featured better mechanical properties
than recycled polymer.

Keywords: cellulose; reinforced; composite; healing; recycling

1. Introduction
Cellulose fibers (CFs) have drawn much attention as composite materials due to their intrinsic
advantages, such as natural abundance, low cost, and excellent mechanical properties [1]. As celluloses
feature a linear chain in which glucose is linked through β-1,4-glucosidic bonds, they aggregate to form
fibers through hydrogen bonding [2]. This provides CFs with a high elastic modulus of 138 GPa [3].
Although this hydrogen bonding provides CFs with excellent mechanical properties, it hampers CF
dispersion in the hydrophobic polymer matrix in composites. To overcome this, celluloses have
been modified using NIO4 [4,5], acid anhydride [6], carbonyl chloride [7,8], TEMPO/NaClO [9,10],
deep-eutectic solvents [11,12], and silanes [13]. Although these methods are effective for the
modification of CFs, they require multiple steps and the use of organic solvents. To reduce the
cost of composite fabrication, a simple and cost-effective way to generate hydrophobic CFs is needed.
Polymers containing reversible bonds can mend damage when they are impacted,
which ultimately extends their lifetime [14,15]. Polymers featuring disulfide bonds [16,17],
dynamic urea bonds [18,19], π–π interaction [20], and hydrogen bonding [21–23] have been studied
for application as self-healable materials. As the flow of polymers is the key factor in healing
damage, a high temperature is required for healing to occur [24]. Although some polymers could
mend under mild conditions, their poor mechanical properties needed to be improved for their
practical application [16,17]. Recently, a polymer featuring dynamic urea bonds showed a high

Polymers 2019, 11, 117; doi:10.3390/polym11010117 www.mdpi.com/journal/polymers

Polymers 2019, 11, 117 2 of 10

E-modulus of 0.34 GPa and a healing efficiency of 50% at 100 ◦ C [25]. Additionally, Kim et al.
reported that a disulfide-containing polymer with the excellent mechanical properties was mended at
room temperature (RT) within 2 h [26]. Since Diels–Alder (DA) adducts were reversibly converted
to their dienes and dienophiles through a retro DA reaction [27], they have also been employed as
crosslinkers for self-healable polymers [28–30]. When furans and maleimides were employed as dienes
and dienophiles, respectively, DA reaction and retro DA reaction occurred at around 50 and 120 ◦ C,
respectively. The healing efficiency of this combination reached 83% [31], implying that polymers
with DA adducts are promising healable materials. To improve the mechanical properties of healable
polymers, we fabricated CF-reinforced composites using thermosetting polymers with DA adducts
as crosslinkers. The simple treatment of cationic surfactant made CFs hydrophobic, which enhanced
the dispersion of the CFs in the healable polymers. The effect of the CFs on the mechanical properties
of the composites was investigated. While mechanical properties of the composites with 1–5 wt %
CFs increased, those of the composite with 7 wt % CFs decreased due to agglomeration of the CFs
in the composite. Like unreinforced polymer, CF-reinforced polymer composites were healable via
a retro DA reaction. Besides this, a retro DA reaction made thermosetting polymers and composites
recyclable through dissolution processes.

2. Materials and Methods

2.1. Materials
Poly(propylene glycol) (Mn ~ 1000 g/mol), tetramethylammonium chloride (TM),
tetrabutylammonium chloride (TB), hexadecyltrimethylammonium chloride (TH), propylene glycol
monomethyl ether acetate (PGMEA), methyl ethyl ketone (MEK), dimethylformamide (DMF),
and dibutyltin dilaurate were purchased from Sigma-Aldrich Korea Ltd, Yongin, Korea.
Isophorone diisocyanate (IPDI), bis(3-ethyl-5-methyl-4-maleimidophenyl)methane (BMI),
and furfurylamine were purchased from SEJINCI, Seoul, Korea. Cellulose was purchased
from MOORIM P&P, Seoul, Korea. All chemicals were used as received, without purification. The
furan diol 1 was synthesized by following the known method [32]. The poly(propylene glycol) used
was designated as PPO-1000.

2.2. Preparation of the TH-Treated CFs

An aqueous suspension of CFs (166.6 g, 3.0 wt %) was diluted with deionized water (100 g)
to produce a CF suspension with a solid content of 1% by weight. This was mixed with
hexadecyltrimethylammonium chloride (5.0 g) and stirred at room temperature. After 30 min,
hydrophobic CFs were isolated by filtration using a Buchner funnel, and dried in an oven (KIPAE ENT
CO., LTD, Suwon, Korea).

2.3. Preparation of the F-PU Solution

The flask was charged with furan diol 1 (2.15 g, 10.0 mmol) and PPO-1000 (10.0 g, 10.0 mmol).
To this mixture were added IPDI (4.5 g, 20.2 mmol) and PGMEA (16.6 g). A catalytic amount of
dibutyltin dilaurate was added to the mixture at room temperature. The solution was heated at 60 ◦ C
for 3 h and cooled to room temperature to produce an F-PU solution with a solid content of 50%
by weight.

2.4. Synthesis of the DA-PU

To the as-prepared F-PU in PGMEA (5.0 g) was added BMI (0.33 g). This mixture was heated at
60 ◦ C.
After 20 min, the solution was poured into a PTFE mold, followed by being kept in an oven.
Oven temperature was increased from 40 to 80 ◦ C slowly to produce a DA-PU pad.
Polymers 2019, 11, 117 3 of 10

2.5. Synthesis of the CF-Reinforced Composites

To the as-prepared F-PU in PGMEA (5.0 g) were added the TH-treated CFs. Once the cellulose
fibers were dispersed in the F-PU solution, BMI (0.33 g) was added and stirred at 60 ◦ C for 20 min.
The solution was poured into a PTFE mold, followed by being kept in an oven. Oven temperature
was increased from 40 to 80 ◦ C slowly to produce CF-reinforced composites. The ratio of CF/DA-PU
ranged from 1% to 7% by weight.

2.6. Healing Tests

The as-prepared DA-PU solution was poured on a glass substrate with dimensions of 5 cm × 5 cm.
Then, it was spun at 500 rpm for 20 s to form a thin polymer layer on the glass substrate. This was kept
at 80 ◦ C overnight to produce a DA-PU layer on the glass substrate. The DA-PU layer was scratched
by a cutter knife, then it was kept at 120 ◦ C for 20 min for healing. The change in the scratch before
and after thermal treatment was monitored by optical microscopy. For the healing test of a composite,
a solution containing CF-reinforced composite with a CF content of 5 wt % was used.

2.7. Instrumentation
The proton NMR spectra were measured at 400 MHz on an NMR spectrometer equipped with
Bruker Top Spin 3.2 software (AscendTM 400, Bruker, Madison, WI, USA) using DMSO-d6 as a solvent
at room temperature. FTIR spectra in the range from 600 to 4000 cm−1 were measured on a Fourier
transform infrared spectrophotometer (IRAffinity-1S, SHIMADZU, Kyoto, Japan), by placing samples
on the surface of a ZnSe crystal using the attenuated total reflectance method. Surface profiles were
measured by laser confocal microscopy (VK-9710K, Keyence, Osaka, Japan). The surfaces of the
fractured samples were observed by scanning electron microscopy (SEM; JSM-6480, JEOL Ltd., Tokyo,
Japan). Thermal stability of polymers, TH-treated CFs, and CF-reinforced composites were evaluated
under N2 atmosphere by elevating the temperature from 23 to 400 ◦ C at the rate of 10 ◦ C/min using a
differential thermal thermogravimetry analyzer (DTG-60H, Shimadzu, Japan). The molecular weight
was obtained on an ACQUITY APC (Waters Corporation, Milford, MA, USA) equipped with a column
(ACQUITY APC XT 2002.5 µm) by flowing a THF solution of a sample at the rate of 0.7 mL/min.
After calibrating the column set using polystyrenes with molecular weights ranging from 1220 to
264,000 g/mol, the molecular weight of the sample was measured.

3. Results and Discussion

To make CFs hydrophobic, their aqueous solutions were treated with quaternary alkyl
ammoniums (QAAs). The absorption of the QAAs on the surface of the CFs would alter the
hydrophilicity of the CFs. We chose three types of QAAs (i.e., TM, TB, TH) to investigate the effect
of the alkyl groups on the hydrophobicity of the CFs. The changes in the hydrophobicity of the
QAA-treated CFs were readily observed by the extraction method using a water/toluene mixture.
In this, the aqueous suspension of CFs was diluted to 1% using de-ionized water and stirred for
30 min. The solution was stirred by a magnetic stirrer at RT for 30 min. Then, toluene was added
to the CF suspension to make the ratio of water/toluene one-to-one by weight. After the two layers
were mixed in a separatory funnel, the mixture was allowed to stand for 30 min to check whether
the CFs had been transferred to the organic layer (Figure 1). The CFs in the water phase were not
transferred to the toluene phase when TM and TB were used as the dispersants. However, the use
of TH made the CFs stay in the toluene phase instead of the water phase. The different behaviors of
the dispersant-treated CFs may be attributed to the hydrophobicity and the shape of the dispersants.
For example, the effect of cetyltrimethylammonium bromide on exfoliating graphite was reported
previously [33]. Although TB and TH possess comparable carbon numbers, TB did not extract the
CFs from the water phase to the toluene phase, unlike TH. This indicates that the bulky substituents
of the butyl groups hampered interactions between the TB cations and the hydroxyl groups on the
Polymers 2019, 11, 117 4 of 10
Polymers
Polymers2018,
2018,10,
10,xxFOR
FOR PEER
PEERREVIEW
REVIEW 44 of
of 10
10

celluloses.
celluloses. In
In the
the case
case of
of TM,
TM, the
the smaller
smaller size
size of
of the
the substituent
substituent could
could lead
lead to
to the
the interaction
interaction between
between
the dispersants
the dispersants
dispersantsand and the
andthe CFs,
theCFs,
CFs, but
but thethe
but the hydrophobicity
hydrophobicity
hydrophobicity of
of methyl
methyl
of methyl groups
groups
groups was
was not not
waslarge large
large enough
not enough enough to
to
to extract
extract
extract the
the CFs
CFs in
in water
water
the CFs in water by toluene. by
by toluene.
toluene.

Figure
Figure 1.
1. Cellulose
Cellulose fibers
fibers (CFs)
(CFs) in
in water/toluene
water/toluene solution
water/toluene solutionafter
solution aftermixing.
after mixing.The
mixing. TheCFs
The CFswere
CFs weretreated
were treated with
with the
the
following surfactants: (a) none, (b) tetramethylammonium chloride (TM), (c) tetrabutylammonium
following surfactants: (a) none, (b) tetramethylammonium chloride (TM), (c) tetrabutylammonium
tetrabutylammonium
chloride
chloride (TB),
(TB), and
and (d)
(d) hexadecyltrimethylammonium
hexadecyltrimethylammonium chloridechloride (TH).
(TH).

To fabricate
To
To fabricatehealable
fabricate healable
healable composites
composites
composites usingusingthe TH-treated
using the
the TH-treated CFs, weCFs,
TH-treated synthesized
CFs, we polymers crosslinked
we synthesized
synthesized polymers
polymers
by DA adducts.
crosslinked
crosslinked by
by DAFuran
DA diol 1 Furan
adducts.
adducts. was
Furan readily
diol 11synthesized
diol was readilyby
was readily reacting glycerol
synthesized
synthesized by
by reactingcarbonate
reacting glycerol
glyceroland carbonate
furfuryl
carbonate amine
and
and
at the
furfuryl ratio
furfuryl amine of
amine at 1:1 following
at the
the ratio
ratio ofthe reported
of 1:1
1:1 following method.
following the Then,
the reported a
reported method.mixture of
method. Then, furan diol
Then, aa mixture 1, polyol
mixture of (PPO-1000),
of furan
furan diol
diol 1,1,
and IPDI
polyol
polyol in PGMEA
(PPO-1000),
(PPO-1000), andwas
and IPDIheated
IPDI in at 60 ◦ Cwas
in PGMEA
PGMEA forheated
was 3 h to produce
heated at
at 6060 °C°C F-PU,
for 33 hhinto
for towhich
produce
produce theF-PU,
furan in
F-PU, groups
whichwere
in which the
the
grafted
furan
furan onto the
groups
groups were
werepolyurethane
grafted
grafted onto onto backbones
the (Scheme 1).
the polyurethane
polyurethane The completion
backbones
backbones (Schemeof
(Scheme 1).the
1). The
The reaction between
completion
completion of the
of the
the
polyol and the isocyanate was indicated by the disappearance of the IR absorption at 2260 cm −1 ,
reaction
reaction between
between the the polyol
polyol and and the the isocyanate
isocyanate was was indicated
indicated by by the the disappearance
disappearance of of the
the IRIR
corresponding
absorption
absorption at to NCO
at 2260
2260 cm−1,,stretching
cm −1 corresponding
corresponding (Figure to S1).
NCOThe
to NCO identity
stretching
stretching of F-PU
(Figure
(Figure S1).
S1). was Thecharacterized
The identity
identity of by NMR
of F-PU
F-PU was
was
spectroscopy. by
characterized
characterized The appearance
by NMR
NMR of the proton
spectroscopy.
spectroscopy. The resonance atof
The appearance
appearance of4.8–4.9
the ppm in
the proton
proton the 1 H-NMR
resonance
resonance at spectrum
at 4.8–4.9
4.8–4.9 ppm
ppm showed
in
in the
the
11that the hydroxyl group in the polyol had reacted within isocyanate to had
form a urethane
H-NMR
H-NMR spectrum
spectrum showed
showed that
that the
the hydroxyl
hydroxyl group
group in the
the polyol
polyol had reacted
reacted withbond.
with This was
isocyanate
isocyanate to
to
further
form
form confirmed
aa urethane
urethane bond. by 2DThis
bond. COSY-NMR
This was
was further (Figure
further S2), in which
confirmed
confirmed by
by 2D 2Dthere was a correlation
COSY-NMR
COSY-NMR (Figurebetween
(Figure S2),
S2), in two protons
in which
which there
there
at 4.86
was
was and 1.20 ppm.
aa correlation
correlation In addition,
between
between two
two protonsthe carbon
protons at
at 4.86 resonance
4.86 and 1.20atppm.
and 1.20 155.5In
ppm. ppm
In furtherthe
addition,
addition, confirmed
the carbon the presence
carbon resonance
resonance at
at
of theppm
urethane bond 13
in the C-NMR spectrum. Theurethane
weight average and
155.5
155.5 ppm further
further confirmed
confirmed the
the presence
presence of
of the
the urethane bond
bond in in thenumber
the 13C-NMR
13
C-NMR average molecular
spectrum.
spectrum. The
The
weights
weight of F-PUand
weight average
average were
and measured
number
number average
averageto be 40,302 and
molecular
molecular 34,270of
weights
weights ofg/mol,
F-PU
F-PU were respectively.
were measured
measured The to polydispersity
to be
be 40,302
40,302 and
and
(M
34,270
34,270/M
w g/mol, )
g/mol,
n was calculated
respectively.
respectively. Theto be 1.176.
The polydispersity
polydispersity (M (Mww/M/Mnn)) was
was calculated
calculated to to bebe 1.176.
1.176.

Scheme
Scheme 1.
1. Synthesis
Synthesis of
of F-PU
F-PU and
and DA-PU.
DA-PU.
Polymers 2019, 11, 117 5 of 10
Polymers 2018, 10, x FOR PEER REVIEW 5 of 10

To synthesize
To synthesizepolyurethane
polyurethanecrosslinked
crosslinked through
through DADA adducts,
adducts,bismaleimide
bismaleimide 2 was2 added
was added to F-PUto
in PGMEA, and the solution was spin-coated on a glass substrate followed by drying at 80 ◦ C overnight.
F-PU in PGMEA, and the solution was spin-coated on a glass substrate followed by drying at 80 °C
When the solvent
overnight. When the was slowly
solvent wasevaporated, DA adducts
slowly evaporated, formed to
DA adducts produce
formed a free-standing
to produce film of
a free-standing
DA-PU.
film Owing to
of DA-PU. the crosslinking
Owing through DA
to the crosslinking adducts,
through DA thermosetting DA-PU was
adducts, thermosetting not dissolved
DA-PU was not
in organicin
dissolved solvents
organicsuch as MEK
solvents suchand as DMF
MEK at andRT.DMFTo investigate
at RT. To ainvestigate
retro DA reaction,
a retro DA DA-PU was
reaction,
immersed ◦
DA-PU wasinimmersed
DMSO-d6inand kept at
DMSO-d 120 kept
6 and C for
at20120min. DA-PU
°C for 20 min.powder
DA-PU was dissolved
powder was in DMSO-d
dissolved in6
over time to make the solution yellow in color (Figure S3). The deuterated solution
DMSO-d6 over time to make the solution yellow in color (Figure S3). The deuterated solution was was characterized
by 1 H-NMR spectroscopy.
characterized The 1 H-NMR The
by 1H-NMR spectroscopy. spectrum
1H-NMR showed resonances
spectrum showed at 6.21, 6.35, and
resonances 7.53 ppm,
at 6.21, 6.35,
corresponding
and to protons of furan,
7.53 ppm, corresponding indicating
to protons of furan,thatindicating
the thermal thattreatment
the thermal ledtreatment
to a retroled
DAtoreaction
a retro
(Figure
DA S4). The
reaction proton
(Figure resonances
S4). The proton at 7.21 ppm for ethylenic
resonances at 7.21 ppm protons of maleimide
for ethylenic confirmed
protons the retro
of maleimide
DA reaction
confirmed asretro
the well.DAThereaction
de-crosslinking of a polymer
as well. The was crucial
de-crosslinking for its healing
of a polymer ability for
was crucial because a soft
its healing
polymer
ability flowedareadily
because and filled
soft polymer scratches.
flowed readily Weand
performed healing tests
filled scratches. We with this compound
performed (Figure
healing tests 2).
with
The surface of DA-PU was scratched by a cutter knife, then it was kept at 120 ◦ C. After 20 min, the mark
this compound (Figure 2). The surface of DA-PU was scratched by a cutter knife, then it was kept at
on the
120 °C.surface
After 20had completely
min, the markdisappeared
on the surface andhad
the completely
polymer became yellow, indicating
disappeared that a retro
and the polymer DA
became
reactionindicating
yellow, had occurred.
that a retro DA reaction had occurred.

Figure 2. Optical
Figure 2. Opticalmicroscopy
microscopyimages.
images. The
The surface
surface of DA-PU
of DA-PU (a) before
(a) before and (b)and (b)the
after after the healing
healing process.
process. Theof
The surface surface of the composite
the composite with 5 wtwith 5 wt
% CFs (c)%before
CFs (c) before
and and the
(d) after (d) after theprocess.
healing healing process.

Besides
Besides the theself-healing
self-healing ability, a retro
ability, DA reaction
a retro made the
DA reaction madethermosetting polymers recyclable
the thermosetting polymers
[34]. To this, the mixture of F-PU and bismaleimide 2 in PGMEA
recyclable [34]. To this, the mixture of F-PU and bismaleimide 2 in PGMEA was poured into was poured into a PTFE mold witha
the
PTFE size of 4 with
mold cm ×the2 cm,sizeand
of 4heated
cm × from
2 cm, 40andtoheated
80 °C slowly
from 40 totoproduce◦ rectangular
80 C slowly DA-PU
to produce samples.
rectangular
For
DA-PUrecycling of DA-PU,
samples. it wasof
For recycling dissolved
DA-PU, in PGMEA.
it was As the
dissolved pad was not
in PGMEA. dissolved
As the pad was innot
PGMEA at RT,
dissolved in
the
PGMEAsolution was
at RT, theheated
solution at was
120 °C for 1ath.120
heated ◦
When the
C for solution
1 h. When the wassolution
clear, itwas
wasclear,
poured into
it was the PTFE
poured into
mold
the PTFEandmold
maintained at temperatures
and maintained in the range
at temperatures of range
in the 40 to 80 °C to
of 40 to 80 ◦ C to produce
produce pads of pads
the same
of theshape.
same
Although there was
shape. Although therea was
slight difference
a slight in color
difference between
in color between thethe twotwopads,
pads,the
thethermosetting
thermosetting was
successfully recycled.The
successfully recycled. The thermal
thermal stability
stability of theofDA-PU
the DA-PU was evaluated
was evaluated by the
by increasing increasing
temperature the
up to 400 ◦ C up
temperature by to 400 °C by thermogravimetric
thermogravimetric analysis (TGA). analysis (TGA).showed
The results The resultsthat showed
DA-PU was that stable
DA-PU was
below
250 ◦ Cbelow
stable 250 °C
with little within
change little change
weight in weight
(Figure S5). As(Figure S5). As DA-PU
thermosetting thermosetting DA-PUand
was mended was mended
recycled at
and ◦
120 recycled at 120 °C, itswas
C, its decomposition decomposition
not a concern. wasThese
not aresults
concern. wereThese results
in good were in with
agreement goodthe agreement
thermal
with theofthermal
stability stabilitycrosslinked
other polymers of other polymers
by DA adducts crosslinked
[32]. The bymechanical
DA adducts [32]. The
properties mechanical
of DA-PU were
properties of DA-PU
measured through were test
a tensile measured
(Figurethrough a tensile
3). A polymer test with
sample (Figure 3). A polymer
dimensions of 4 cmsample
× 2 cmwithwas
dimensions
extended along of 4thecmz-axis
× 2 cmuntilwasitextended
was broken. alongThethe z-axis until
maximum it was
stress andbroken. The maximum
the elongation at break stress
of the
and the elongation
polymer were 2.83 at MPabreakandof53%,
the polymer wereRecycling
respectively. 2.83 MPa and 53%,the
through respectively.
dissolution Recycling through
process changed
the
thesedissolution
propertiesprocess changedsample
of the recycled these properties
to 2.44 MPaofand the 81%.
recycledThe sample to 2.44 MPa stemmed
change presumably and 81%.from The
change presumably
the different density stemmed from the of
in the crosslinking different
the twodensity
samples indue
the tocrosslinking of theoftwo
the reversibility thesamples due
DA adducts.
to the were
They reversibility
comparable of thetoDA theadducts.
maximum They were(2.4
stress comparable
MPa) andtothe theelongation
maximumatstressbreak (2.4 MPa)
(88%) ofand
the
the elongation polyurethane
silicone-based at break (88%)crosslinked
of the silicone-based polyurethane
with DA adducts crosslinked
in the previous with[34].
report DA adducts in the
previous report [34].
Polymers 2019, 11, 117 6 of 10
Polymers 2018, 10, x FOR PEER REVIEW 6 of 10

Figure
Figure3.3.Stress-strain
Stress-straincurves
curvesofofCF-reinforced
CF-reinforcedDA-PU
DA-PUcomposites
composites(a)
(a)before
beforeand
and(b)
(b)after
afterrecycling.
recycling.

Encouraged by
Encouraged bythis,
this,wewe
mixed TH-treated
mixed CFs and
TH-treated CFsDA-PU to improve
and DA-PU the mechanical
to improve properties
the mechanical
properties of DA-PU. To compare the dispersibility of TH-treated CFs with that of pristine CFs,mixed
of DA-PU. To compare the dispersibility of TH-treated CFs with that of pristine CFs, they were they
with amixed
were PGMEA withsolution of F-PU,
a PGMEA at the CF/F-PU
solution of F-PU, ratio of 100-to-5
at the CF/F-PUbyratioweight, and sonicated
of 100-to-5 for 30 min.
by weight, and
After the addition
sonicated of bismaleimde
for 30 min. 2 to both of
After the addition composite solutions,
bismaleimde 2 to they
bothwere pouredsolutions,
composite into PTFE they
moldswere
and
kept in an oven overnight. While the TH-treated CFs were well dispersed
poured into PTFE molds and kept in an oven overnight. While the TH-treated CFs were wellin the polymer, the pristine
CFs aggregated
dispersed in the matrix
in the polymer, (FigureCFs
the pristine S6).aggregated
This demonstrates that (Figure
in the matrix the hydrophobicity facilitated
S6). This demonstrates
dispersion
that of the TH-treated
the hydrophobicity CFs in the
facilitated hydrophobic
dispersion of thepolymer matrix.
TH-treated CFsDespite
in the this, aggregation
hydrophobic of the
polymer
TH-treated CFs was monitored when the CF-to-polymer ratio was 7% by weight.
matrix. Despite this, aggregation of the TH-treated CFs was monitored when the CF-to-polymer To investigate the
effectwas
ratio of the
7%TH-treated
by weight.CFs on the mechanical
To investigate properties,
the effect composites
of the TH-treated CFswith CF/DA-PU
on the ratios
mechanical ranging
properties,
from 1% to 7% by weight were prepared. The mechanical properties of the composites
composites with CF/DA-PU ratios ranging from 1% to 7% by weight were prepared. The mechanical are summarized
in Table 1. of the composites are summarized in Table 1.
properties
Table 1. Mechanical properties of polymer and composites.
Table 1. Mechanical properties of polymer and composites.
Tensile Strength Elongation at Young’s Toughness
CF Tensile Strength
CF Content Elongation at Young’s Modulus Toughness
(MPa) Break Modulus (MPa) (MJ/m3 )
Content (MPa) Break (MPa) (MJ/m3)
0 wt % 2.83 ± 0.06 53.55 ± 3.35 10.10 ± 0.32 1.09 ± 0.02
0 wt % 2.83 ± 0.06 53.55 ± 3.35 10.10 ± 0.32 1.09 ± 0.02
1 wt % 3.70 ± 0.24 73.33 ± 8.63 10.80 ± 0.87 1.96 ± 0.04
1 wt % 3 wt %3.70 ± 0.24
3.95 ± 0.31 73.3388.33
± 8.63
± 6.20 10.80
23.80 ±± 0.87
2.67 1.96 ± 0.04
2.62 ± 0.02
3 wt % 5 wt %3.95 ± 0.31
4.83 ± 0.82 88.33 ± 6.20
151.57 ± 28.01 23.80
17.74 ±± 2.67
2.29 2.62 ± 0.02
5.67 ± 0.06
5 wt % 1.38 ± 0.25 151.57
7 wt %4.83 ± 0.82 36.66 ± 10.42
± 28.01 7.21 ±± 1.60
17.74 2.29 0.35 ± 0.29
5.67 ± 0.06
7 wt % 1.38 ± 0.25 36.66 ± 10.42 7.21 ± 1.60 0.35 ± 0.29
The stress-strain curves show that mechanical properties were improved as the CF content in
The stress-strain
the thermosetting curves show
composites that mechanical
increased (Figure 3).properties
When thewere improved
CFs were mixed asinthethe
CFpolymer
content by
in
the
5 wtthermosetting
%, maximum composites increased
tensile strength, (Figure
elongation at 3). When
break, the CFs
Young’s were mixed
modulus, in the polymer
and toughness by 5 wt
increased by
%, maximum tensile strength, elongation at break, Young’s modulus, and toughness
70%, 183%, 75%, and 420%, respectively, in comparison to unreinforced DA-PU. This demonstrates increased by
70%, 183%,
that the 75%, and
addition 420%,
of CFs respectively,
made the compositein comparison
tough. Even tothough
unreinforced DA-PU.
the increase in This
both demonstrates
modulus and
that the addition
elongation of theofcomposites
CFs made thewascomposite tough.composites
unusual, some Even though thecarbon
with increase in both modulus
nanotubes [35–37] and
and
elongation of the composites was unusual, some composites with carbon nanotubes
graphene oxide [38] have shown similar results. The Young’s modulus was higher for the composite [35–37] and
graphene
with a CFoxide [38]ofhave
content 3 wtshown
% thansimilar results.
that with a CFThe Young’s
content of 5modulus was higher
wt % because for the composite
the increased modulus
with a CF content of 3 wt % than that with a CF content of 5 wt % because the
by CFs was compromised by TH, which acted as a plasticizer in composites [39]. Composites also increased modulus by
CFs was compromised by TH, which acted as a plasticizer in composites
mended scratches like DA-PU under the same conditions, indicating that the improved mechanical [39]. Composites also
mended
properties scratches
did not like DA-PUthe
deteriorate under the same
healing conditions,
efficiency indicating
of the polymer that the
(Figure 2). improved mechanical
Well-dispersed CFs in
properties did not deteriorate the healing efficiency of the polymer (Figure 2). Well-dispersed
the DA-PU polymer matrix were also observed by optical microscopy after the healing test. However, CFs in
the DA-PU polymer matrix were also observed by optical microscopy after
the mechanical properties of the composite with a CF content of 7 wt % were worse than those of the healing test.
However,
DA-PU due thetomechanical properties
poor dispersion of CFs
of the the composite
in the polymerwith a(Figure
CF content of 7 wt
3). Once the%CFs were worse than
aggregated to
those of DA-PU due to poor dispersion of the CFs in the polymer (Figure 3).
become bundles in the composites, they served as stress concentrators, which induced pull-out failure. Once the CFs
aggregated to become bundles in the composites, they served as stress concentrators, which induced
pull-out failure. The surfaces of the fractures in the composites after the tensile tests were analyzed
Unlike other composites, the toughness of the composite with a CF content of 7 wt % was improved
after recycling. This was probably due to the increased dispersion of CFs in the composite solution at
high temperature during recycling.

Table 2. Mechanical properties of polymer and composites after recycling.

Polymers 2019, 11, 117 7 of 10
CF Tensile Strength Elongation at Young’s Modulus Toughness
Content (MPa) Break (MPa) (MJ/m3)
The surfaces
0 wt % of the fractures
2.44 ± 0.09 in the composites after the
81.67 ± 9.36 tensile tests were analyzed
4.61 ± 0.25 by scanning electron
1.48 ± 0.03
microscopy
1 wt % (Figure 4). While
2.61 ± 0.20 the composites with
76.66 ±17.14 a CF content of 1–5
9.83 ± 1.50 wt % showed 1.63 ± 0.06holes
small
(ca. 10 %
3 wt µm), those with
2.96 ±a 0.37
CF content of 113.33
7 wt %±featured
12.73 large holes 9.14(>ca. 100 µm). As the
± 1.74 2.70diameter
± 0.04 of
the CFs
5 wt % was ca. 10 µm, the
3.01 ± 0.81large holes stemmed from
245.00 ± 44.33 the pullout of the
6.57 ± 1.36 CF aggregates. This implied
4.80 ± 0.08
that dispersion
7 wt % of the CFs plays
1.73 ± 0.44 a key role in enhancing
78.33 ± 42.30 the mechanical
4.00 ± 1.15properties of composites
0.98 ± 0.71 [35].

Figure
Figure 4.4.SEM images
SEM of theofsurfaces
images of fractures
the surfaces in the composites
of fractures after the tensile
in the composites after tests. Composites
the tensile tests.
with a CF content of (a) 1 wt %, (b) 3 wt %, (c) 5 wt %, and (d) 7 wt %. Red circles
Composites with a CF content of (a) 1 wt %, (b) 3 wt %, (c) 5 wt %, and (d) 7 wt %. Red circles indicate holes
generated by the
indicate holes pullout by
generated of the
the CFs.
pullout of the CFs.

We recycled the CF-reinforced thermosetting polymer composites by dissolving them in PGMEA

at 120 ◦ C, followed by pouring in a PTFE mold. The mechanical properties of the recycled ones
were measured and are summarized in Table 2. The stress-strain curves of recycled composites
showed that tensile strength and modulus were decreased, with a concomitant increase in elongation.
This indicated that crosslinking density of the composites was reduced after recycling because DA
adducts were reversibly converted to their dienes and dienophiles at high temperature. The toughness
of composites was rather similar before and after recycling. When the composite with a CF content
of 5 wt % was recycled, tensile strength, elongation at break, Young’s modulus, and toughness
increased by 23%, 200%, 42%, and 224% in comparison to the recycled DA-PU polymer. Unlike other
composites, the toughness of the composite with a CF content of 7 wt % was improved after recycling.
This was probably due to the increased dispersion of CFs in the composite solution at high temperature
during recycling.

Table 2. Mechanical properties of polymer and composites after recycling.

Tensile Strength Elongation at Young’s Toughness

CF Content
(MPa) Break Modulus (MPa) (MJ/m3 )
0 wt % 2.44 ± 0.09 81.67 ± 9.36 4.61 ± 0.25 1.48 ± 0.03
1 wt % 2.61 ± 0.20 76.66 ±17.14 9.83 ± 1.50 1.63 ± 0.06
3 wt % 2.96 ± 0.37 113.33 ± 12.73 9.14 ± 1.74 2.70 ± 0.04
5 wt % 3.01 ± 0.81 245.00 ± 44.33 6.57 ± 1.36 4.80 ± 0.08
7 wt % 1.73 ± 0.44 78.33 ± 42.30 4.00 ± 1.15 0.98 ± 0.71
Polymers 2019, 11, 117 8 of 10

4. Conclusions
Hydrophobic CFs were not only dispersed in hydrophobic polymers but also acted as both fillers
and plasticizers, which improved the mechanical properties of composites. The combined enhancement
in Young’s modulus and elongation realized that toughness of the composites increased in proportion
to the amount of the dispersed CFs compared to the unreinforced polymer. The incorporation
of the hydrophobic CFs in thermosetting, crosslinked by DA adducts, resulted in highly-durable
mendable composites without deteriorating healing ability. In addition, thermosetting composites were
recyclable via a retro DA reaction, which is not possible for conventional thermosetting crosslinked by
irreversible bonds.

Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4360/11/1/117/s1,

Figure S1. FTIR spectrum of F-PU. Figure S2. 2D COSY NMR of polymer F-PU. There was a correlation between
the proton (a) next to the urethane bond and the protons (b) on the methyl group. Figure S3. DA-PU in the NMR
tube (left) before and (right) after being heated at 120 ◦ C for 20 min. Figure S4. 1 H-NMR spectrum of DA-PU in
DMSO-d6 ranging from 6.0 ppm to 8.0 ppm after heat treatment. Figure S5. TGA graphs over temperature for
(a) TH-treated CFs, (b) DA-PU and (c) the CF-reinforced composite. Figure S6. CF-reinforced polymer composites
with a CF content of 5 wt %. The composites were prepared using (left) the TH-treated CFs and (right) the
pristine CFs.
Author Contributions: Supervision and writing—original draft, Y.K. and K.P.; conceptualization and
methodology, C.S. (Cheolmin Shin) and C.S. (Chiho Shin); formal analysis and investigation, Y.-S.S. and H.-J.L.
Funding: This work was supported by the Technology development Program (S2482255) funded by the Ministry
of SMEs and Startups (MSS, Korea).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Lee, K.-Y.; Aitomäki, Y.; Berglund, L.A.; Oksman, K.; Bismarck, A. On the use of nanocellulose as
reinforcement in polymer matrix composites. Compos. Sci. Technol. 2014, 105, 15–27. [CrossRef]
2. Cazón, P.; Vázquez, M.; Velazquez, G. Novel composite films based on cellulose reinforced with chitosan
and polyvinyl alcohol: Effect on mechanical properties and water vapour permeability. Polym. Test. 2018, 69,
536–544. [CrossRef]
3. Sato, A.; Kabusaki, D.; Okumura, H.; Nakatani, T.; Nakatsubo, F.; Yano, H. Surface modification of cellulose
nanofibers with alkenyl succinic anhydride for high-density polyethylene reinforcement. Compos. Part A
2016, 83, 72–79. [CrossRef]
4. Liimatainen, H.; Visanko, M.; Sirviö, J.A.; Hormi, O.E.O.; Niinimaki, J. Enhancement of the Nanofibrillation
of Wood Cellulose through Sequential Periodate–Chlorite Oxidation. Biomacromolecules 2012, 13, 1592–1597.
[CrossRef] [PubMed]
5. Sirvio, J.; Hyvakko, U.; Liimatainen, H.; Niinimaki, J.; Hormi, O. Periodate oxidation of cellulose at elevated
temperatures using metal salts as cellulose activators. Carbohydr. Polym. 2011, 83, 1293–1297. [CrossRef]
6. Pandey, J.K.; Chu, W.S.; Kim, C.S.; Lee, C.S.; Ahn, S.H. Bio-nano reinforcement of environmentally degradable
polymer matrix by cellulose whiskers from grass. Compos. Part B 2009, 40, 676–680. [CrossRef]
7. Freire, C.S.R.; Silvestre, A.J.D.; Neto, C.P.; Belgacem, M.N.; Gandini, A. Controlled heterogeneous
modification of cellulose fibers with fatty acids: Effect of reaction conditions on the extent of esterification
and fiber properties. J. Appl. Polym. Sci. 2006, 100, 1093–1102. [CrossRef]
8. Freire, C.S.R.; Silvestre, A.J.D.; Pascoal Neto, C.; Rocha, R.M.A. An Efficient Method for Determination of
the Degree of Substitution of Cellulose Esters of Long Chain Aliphatic Acids. Cellulose 2005, 12, 449–458.
[CrossRef]
9. Habibi, Y.; Chanzy, H.; Vignon, M.R. TEMPO-mediated surface oxidation of cellulose whiskers. Cellulose
2006, 13, 679–687. [CrossRef]
10. Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A. Homogeneous Suspensions of Individualized
Microfibrils from TEMPO-Catalyzed Oxidation of Native Cellulose. Biomacromolecules 2006, 7, 1687–1691.
[CrossRef]
Polymers 2019, 11, 117 9 of 10

11. Hosseinmardi, A.; Annamalai, P.K.; Wang, L.; Martin, D.; Amiralian, N. Reinforcement of natural rubber
latex using lignocellulosic nanofibers isolated from spinifex grass. Nanoscale 2017, 9, 9510–9519. [CrossRef]
[PubMed]
12. Sirviö, J.A.; Visanko, M.; Liimatainen, H. Deep eutectic solvent system based on choline chloride-urea as a
pre-treatment for nanofibrillation of wood cellulose. Green Chem. 2015, 17, 3401–3406. [CrossRef]
13. Fu, J.; Zhang, J.; Song, X.; Zarrin, H.; Tian, X.; Qiao, J.; Rasen, L.; Li, K.; Chen, Z. A flexible solid-state
electrolyte for wide-scale integration of rechargeable zinc–air batteries. Energy Environ. Sci. 2016, 9, 663–670.
[CrossRef]
14. Yang, Y.; Ding, X.; Urban, M.W. Chemical and physical aspects of self-healing materials. Prog. Polym. Sci.
2015, 49–50, 34–59. [CrossRef]
15. An, S.Y.; Arunbabu, D.; Noh, S.M.; Song, Y.K.; Oh, J.K. Recent strategies to develop self-healable crosslinked
polymeric networks. Chem. Commun. 2015, 51, 13058–13070. [CrossRef] [PubMed]
16. Rekondo, A.; Martin, R.; Ruiz de Luzuriaga, A.; Cabañero, G.; Grande, H.J.; Odriozola, I. Catalyst-free
room-temperature self-healing elastomers based on aromatic disulfide metathesis. Mater. Horiz. 2014, 1,
237–240. [CrossRef]
17. Canadell, J.; Goossens, H.; Klumperman, B. Self-Healing Materials Based on Disulfide Links. Macromolecules
2011, 44, 2536–2541. [CrossRef]
18. Ying, H.; Zhang, Y.; Cheng, J. Dynamic urea bond for the design of reversible and self-healing polymers.
Nat. Commun. 2014, 5, 3218. [CrossRef]
19. Park, J.I.; Choe, A.; Kim, M.P.; Ko, H.; Lee, T.H.; Noh, S.M.; Kim, J.C.; Cheong, I.W. Water-adaptive and
repeatable self-healing polymers bearing bulky urea bonds. Polym. Chem. 2018, 9, 11–19. [CrossRef]
20. Burattini, S.; Colquhoun, H.M.; Fox, J.D.; Friedmann, D.; Greenland, B.W.; Harris, P.J.F.; Hayes, W.;
Mackay, M.E.; Rowan, S.J. A self-repairing, supramolecular polymer system: Healability as a consequence of
donor–acceptor π–π stacking interactions. Chem. Commun. 2009, 6717–6719. [CrossRef]
21. Zhang, D.-D.; Ruan, Y.-B.; Zhang, B.-Q.; Qiao, X.; Deng, G.; Chen, Y.; Liu, C.-Y. A self-healing PDMS
elastomer based on acylhydrazone groups and the role of hydrogen bonds. Polymer 2017, 120, 189–196.
[CrossRef]
22. Montarnal, D.; Cordier, P.; Soulié-Ziakovic, C.; Tournilhac, F.; Leibler, L. Synthesis of self-healing
supramolecular rubbers from fatty acid derivatives, diethylene triamine, and urea. J. Polym. Sci. Part
A Polym. Chem. 2008, 46, 7925–7936. [CrossRef]
23. Folmer, B.J.B.; Sijbesma, R.P.; Versteegen, R.M.; van der Rijt, J.A.J.; Meijer, E.W. Supramolecular Polymer
Materials: Chain Extension of Telechelic Polymers Using a Reactive Hydrogen-Bonding Synthon. Adv. Mater.
2000, 12, 874–878. [CrossRef]
24. Garcia, S.J. Effect of polymer architecture on the intrinsic self-healing character of polymers. Eur. Polym. J.
2014, 53, 118–125. [CrossRef]
25. Zechel, S.; Geitner, R.; Abend, M.; Siegmann, M.; Enke, M.; Kuhl, N.; Klein, M.; Vitz, J.; Gräfe, S.; Dietzek, B.;
et al. Intrinsic self-healing polymers with a high E-modulus based on dynamic reversible urea bonds. NPG
Asia Mater. 2017, 9, e420. [CrossRef]
26. Kim, S.-M.; Jeon, H.; Shin, S.-H.; Park, S.-A.; Jegal, J.; Hwang, S.Y.; Oh, D.X.; Park, J. Superior Toughness
and Fast Self-Healing at Room Temperature Engineered by Transparent Elastomers. Adv. Mater. 2018, 30,
1705145. [CrossRef] [PubMed]
27. Liu, Y.-L.; Chuo, T.-W. Self-healing polymers based on thermally reversible Diels–Alder chemistry.
Polym. Chem. 2013, 4, 2194–2205. [CrossRef]
28. 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. [CrossRef]
29. Lee, D.H.; Heo, G.; Pyo, K.-H.; Kim, Y.; Kim, J.-W. Mechanically Robust and Healable Transparent Electrode
Fabricated via Vapor-Assisted Solution Process. ACS Appl. Mater. Interfaces 2016, 8, 8129–8136. [CrossRef]
30. Heo, G.; Pyo, K.-H.; Lee, D.H.; Kim, Y.; Kim, J.-W. Critical Role of Diels–Adler Adducts to Realise Stretchable
Transparent Electrodes Based on Silver Nanowires and Silicone Elastomer. Sci. Rep. 2016, 6, 25358. [CrossRef]
31. Du, P.; Wu, M.; Liu, X.; Zheng, Z.; Wang, X.; Sun, P.; Joncheray, T.; Zhang, Y. Synthesis of linear polyurethane
bearing pendant furan and cross-linked healable polyurethane containing Diels–Alder bonds. New J. Chem.
2014, 38, 770–776. [CrossRef]
Polymers 2019, 11, 117 10 of 10

32. Pyo, K.-H.; Lee, D.H.; Kim, Y.; Kim, J.-W. Extremely rapid and simple healing of a transparent conductor
based on Ag nanowires and polyurethane with a Diels–Alder network. J. Mater. Chem. C 2016, 4, 972–977.
[CrossRef]
33. Vadukumpully, S.; Paul, J.; Valiyaveettil, S. Cationic surfactant mediated exfoliation of graphite into graphene
flakes. Carbon 2009, 47, 3288–3294. [CrossRef]
34. Kim, J.-W.; Lee, D.H.; Jeon, H.-J.; Jang, S.I.; Cho, H.M.; Kim, Y. Recyclable thermosetting thermal pad
using silicone-based polyurethane crosslinked by Diels-Alder adduct. Appl. Surf. Sci. 2018, 429, 128–133.
[CrossRef]
35. Blond, D.; Barron, V.; Ruether, M.; Ryan, K.P.; Nicolosi, V.; Blau, W.J.; Coleman, J.N. Enhancement of
Modulus, Strength, and Toughness in Poly(methyl methacrylate)-Based Composites by the Incorporation of
Poly(methyl methacrylate)-Functionalized Nanotubes. Adv. Funct. Mater. 2006, 16, 1608–1614. [CrossRef]
36. Yang, B.-X.; Shi, J.-H.; Pramoda, K.P.; Goh, S.H. Enhancement of the mechanical properties of polypropylene
using polypropylene-grafted multiwalled carbon nanotubes. Compos. Sci. Technol. 2008, 68, 2490–2497.
[CrossRef]
37. Yang, B.-X.; Shi, J.-H.; Pramoda, K.P.; Goh, S.H. Enhancement of stiffness, strength, ductility and toughness
of poly(ethylene oxide) using phenoxy-grafted multiwalled carbon nanotubes. Nanotechnology 2007, 18,
125606. [CrossRef]
38. Zhao, Y.; Lu, J.; Liu, X.; Wang, Y.; Lin, J.; Peng, N.; Li, J.; Zhao, F. Performance enhancement of polyvinyl
chloride ultrafiltration membrane modified with graphene oxide. J. Colloid Interface Sci. 2016, 480, 1–8.
[CrossRef]
39. Zou, H.; Xu, W.; Zhang, Q.; Fu, Q. Effect of alkylammonium salt on the dispersion and properties of
Poly(p-phenylene sulfide)/clay nanocomposites via melt intercalation. J. Appl. Polym. Sci. 2006, 99,
1724–1731. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).