polymers

Article
Highly Toughened and Transparent Biobased Epoxy
Composites Reinforced with Cellulose Nanofibrils
Sandeep S. Nair 1 , Christopher Dartiailh 2 , David B. Levin 2 and Ning Yan 1,3, *
1 Faculty of Forestry, University of Toronto, 33 Willcocks Street, Toronto, ON M5S 3B3, Canada;
sandeep.sudhakarannair@utoronto.ca
2 Department of Biosystems Engineering, University of Manitoba, E2-376 Engineering and Information
Technology Complex (EITC), 75 Chancellors Circle, Winnipeg, MB R3T 5V6, Canada;
umdartia@myumanitoba.ca (C.D.); david.levin@umanitoba.ca (D.B.L.)
3 Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street,
Toronto, ON M5S 3E5, Canada
* Correspondence: ning.yan@utoronto.ca; Tel.: +1-416-946-8070; Fax: +1-416-978-3834




Received: 5 March 2019; Accepted: 2 April 2019; Published: 3 April 2019


Abstract: Biobased nanofillers, such as cellulose nanofibrils (CNFs), have been widely used as
reinforcing fillers for various polymers due to their high mechanical properties and potential for
sustainable production. In this study, CNF-based composites with a commercial biobased epoxy resin
were prepared and characterized to determine the morphology, mechanical, thermal, and barrier
properties. The addition of 18–23 wt % of CNFs to epoxy significantly increased the modulus, strength
and strain of the resulting composites. The addition of fibrils led to an overall increase in strain
energy density or modulus of toughness by almost 184 times for the composites compared to the
neat epoxy. The addition of CNFs did not affect the high thermal stability of epoxy. The presence of
nanofibrils had a strong reinforcing effect in both glassy and glass transition region of the composites.
A significant decrease in intensity in tan δ peak for the epoxy matrix occurred with the addition
of CNFs, indicating a high interaction between fibrils and epoxy during the phase transition. The
presence of highly crystalline and high aspect ratio CNFs (23 wt %) decreased the water vapour
permeability of the neat epoxy resin by more than 50%.

Keywords: biobased epoxy; cellulose nanofibrils; composite toughness

1. Introduction
Epoxy resins are a class of polymers that exhibit high strength, stiffness, and creep resistance,
as well as dimensional and thermal stability. They have been widely used in coatings, as well as in
composites, construction, adhesives, electronics, and electrical applications [1,2]. Most epoxy resins are
derived from petroleum-based resources. Almost 75% of global production of epoxy resins are derived
from the condensation reaction of bisphenol A (BPA) and epichlorohydrin, forming diglycidyl ether of
bisphenol A (DGEBA). More than 60% of the molecular weight of DGEBA is petroleum-derived [3,4].
Growing environmental concerns, increased legislated obligations to move towards a sustainable
economy, and volatile oil prices have all compelled both industry and academia to develop biobased
epoxy resins. Therefore, a significant measure of research is committed to developing biobased epoxy
resins or partially substituting the use of petroleum-based materials in synthesizing epoxy resin
monomers. Several studies have focused on the preparation of epoxy resin monomers from various
renewable materials such as tannins, cardanols, lignin, carbohydrates, vegetable oils, terpenes, and
resin acids [1,2].
Despite all the good properties of biobased epoxy systems, they are inherently brittle. Epoxy
resins fail under less vigorous conditions when compared to other engineering polymers due to its high

Polymers 2019, 11, 612; doi:10.3390/polym11040612 www.mdpi.com/journal/polymers

Polymers 2019, 11, 612 2 of 13

crosslink density [2,5]. As a result, considerable research has been focused on improving the toughness
of epoxy systems. Different ways have been used to improve the epoxy resin toughness. The major
ones include (a) elastomer or rubber modification [6], (b) thermoplastic modification [7], (c) using
inorganic rigid particles [8], and (d) miscellaneous methods, such as using dendritic molecules [9] or
block copolymers [10]. However, most of these materials have their own limitations in toughening
epoxy resins. Although, the addition of elastomers significantly increases toughness, it decreases the
strength and stiffness [11]. Additionally, it is difficult to control the particle size of elastomers in the
epoxy matrix [6]. The addition of thermoplastics increases ductility, including stiffness and strength.
However, thermoplastics with high compatibility can negatively impact toughness due to non-phase
separation [2,12]. Dendritic molecules or block copolymers are highly expensive [13]. In addition,
most of these toughening agents are petroleum based, and/or non-renewable [2]. Over the years,
scientists have developed different green approaches to overcome these difficulties. The major ones
include developing epoxidized vegetable oils [14] and use of bio-based nanofillers such as CNFs [15]
and cellulose nanocrystals (CNCs) [16] to toughen the epoxy.
Biobased nanofillers, such as cellulose nanofibrils (CNFs) have widths in nano-dimensions and
high aspect ratio (length-to-width ratio). In plant fibers, several cellulose molecules are bundled
together by strong hydrogen bonding to form elementary fibrils or microfibrils. These microfibrils
or aggregates of microfibrils with diameters less than 100 nm are called as CNFs [17,18]. Intensive
mechanical treatment is used to disintegrate or defibrillate plant fibers to nano-sized fibrils [19,20].
CNFs are abundant, biodegradable, renewable, and have unique physical properties. Modulus and
strength as high as 3 GPa and 135 MPa, respectively, have been reported for some CNFs obtained from
plant fibers [21,22]. The highly fibrillated CNFs have lesser defects than micro-sized cellulose fibers,
and together with high aspect ratio can effectively increase the strength and strain of the reinforcing
polymers, and thereby toughen the whole composite [18]. CNFs have been widely used as a reinforcing
agent for a variety of thermoset resins [23,24] and thermoplastic polymers [25,26]. CNFs have also
been studied as filler for various polymers for enhancing the water and oxygen barrier properties.
High crystallinity, smaller dimensions, and the ability to form dense networks within the polymer are
some of the significant barrier-enhancing properties of CNFs [17,27].
Several studies have reported the use of CNFs to enhance the mechanical properties of synthetic
or petroleum based thermoset resin [23,24,28,29]. However, to the best of our knowledge, the addition
of CNFs to enhance the mechanical properties of biobased epoxy resins has been very limited. In
this study, our aim is to incorporate CNFs in a commercial biobased epoxy resin, to investigate the
morphology and physical properties of the final composite material in terms of mechanical, thermal,
and barrier properties.

2. Materials and Methods

2.1. Materials
The epoxy investigated is a commercial biobased epoxy resin, BioPoxy 36, supplied by Ecopoxy Inc.
(Morris, MB, Canada). A medium hardener was the curing agent obtained from the same manufacturer.
Cellulose nanofibril suspensions were obtained from the Center for Biocomposites and Biomaterials
Processing, University of Toronto, Toronto, Canada. The cellulose content for the CNFs were 83%
(based on dry weight), and the average fibril diameter was approximately 10–25 nm. The degree of
polymerization and crystallinity for the CNFs were 1200 and 71%, respectively. All other chemicals
were obtained from Caledon labs (Georgetown, ON, Canada).

2.2. Cellulose Nanofibril Sheet Preparation

CNF suspension at 0.1 wt % solid consistency was vacuum filtered through a membrane (0.22 µm
PVDF, Millipore GVWP14250, MilliporeSigma, Bedford, MA, USA) in a Buchner funnel. The wet films
obtained after dewatering the suspensions along with the membranes were placed between blotting
Polymers 2019, 11, 612 3 of 13

papers and pressed at 345 kPa for 6 min each. Some of the films were removed and dried at 50 ◦ C
under a pressure of approximately 23 kg for two days. A separate batch of wet films was then solvent
exchanged to acetone by soaking in acetone for 48 h. The dried as well as solvent exchanged films
were cut into dog-bone shape tensile specimens according to ASTMD 638-Type V specifications.

2.3. Production of CNF/Epoxy Composites

The cellulose nanofibril network obtained by vacuum filtration was solvent exchanged with
acetone before impregnating with epoxy. The solvent exchange is done to prevent aggregation of fibrils.
The epoxy resin consisted of epoxy/curing agent mixed in a ratio of 4.29:1 by weight. To reduce the
viscosity of resin, the epoxy resin was diluted with acetone in three different ratios: higher viscosity
(epoxy and acetone at a weight ratio of 3:1); medium viscosity (1:1 ratio); and lower viscosity (1:3 ratio).
Solvent exchanged CNF sheet specimens were immersed in each of these epoxy resin solutions. The
impregnated specimens were removed from the solutions and suspended using fine metal wires in an
oven for two hours prior to the curing. In addition to impregnated CNF specimens, neat epoxy films
were solvent casted in aluminum pans. The specimens were then cured in the oven for 2 h at 120 ◦ C.
The cured samples were analyzed for various physical properties.

2.4. Mechanical Properties

The dried CNF films, neat epoxy films and composites were tested according to ASTM D 638-Type
V specifications. Tensile tests were performed using an Instron machine (Model 3367, Instron, Norwood,
MA, USA) equipped with a 2 kN load cell, in which a gauge length of 2.5 cm and a cross head speed
of 10 mm min−1 were used. Dynamic mechanical analysis (DMA) were carried out using a TA
Instruments Q800 analyzer (TA Instruments, New Castle, DE, USA) in tension mode. Rectangular test
specimens were 6–7 mm wide and the span was 15 mm. The frequency and amplitude of oscillation
were set at 1 Hz and 15 µm, respectively. The temperature varied from room temperature to 180 ◦ C at
a heating rate of 3 ◦ C/min.

2.5. Thermogravimetric Analysis (TGA)

The thermal stabilities of the CNF films, neat epoxy films, and composites were obtained using
a TGA Q500 (TA Instruments, New Castle, DE, USA) equipped with platinum crucibles. Samples
(8−10 mg) were heated at a constant rate of 10 ◦ C/min from room temperature to 800 ◦ C in nitrogen
atmosphere. The Tonset for the sample was defined as the temperature at which significant weight
loss begins and determined using the tangent method. The Tmax , defined as the temperature at which
the maximum weight loss rate occurred, corresponds to the maximum value of the derivative weight
curve (dm/dTmax ).

2.6. Water Vapour Transmission (WVT)

WVT was determined according to ASTM E-96-95 water method. The test conditions were set to
23 ◦ C and 50 % relative humidity (RH). The CNF films, neat epoxy and composite films were cut into
6.4 cm diameter specimens and conditioned to test conditions. The specimens were used to seal a test
dish containing 50 mL of distilled water. The weight change of the test dish was recorded periodically
and WVT was determined from the slope of weight change versus time, using the following equation:

WVT = ∆G/∆tA (1)

Here, ∆G is the weight change of dish, ∆t is the time during which the change occurred, and A is
the test area (dish mouth area).
Polymers 2019, 11, 612 4 of 13

2.7. Other Characterizations

Scanning electron microscopy (SEM) images of the cross-sectional fracture surfaces of the CNF,
epoxy and composite films were obtained using a JEOL 6610LV (Seal Laboratories, El Segundo,
CA, USA) operated at 15 kV. Samples were mounted on an aluminum stub using carbon tapes and
sputtered with gold to provide adequate conductivity. Attenuated total reflectance−Fourier transform
infrared spectroscopy (ATR-FTIR) was used to obtain spectra of CNF film, neat epoxy, and composites
((ATR-FTIR, Nicolet iS50 FT-IR, ThermoFisher Scientific, Waltham, MA, USA). The sample spectra
were collected from 600 to 4000 cm−1 at a resolution of 4 cm−1 after 64 scans. Transmittance spectra of
the CNF, epoxy and composite films were obtained in the visible region (300–800 nm) using UV–VIS
spectrophotometer (Lambda 2, PerkinElmer, Waltham, MA, USA).

2.8. Statistical Analysis

Twelve to fifteen specimens were tensile tested to obtain the average mechanical properties of
pure CNF, neat epoxy and composite films. A graphical representation of the change in mechanical
properties of composites with varying CNF weight percent will be obtained. Three specimens from
each group (CNF, neat epoxy and composites films) will be analyzed for TGA, DMA, WVT, FTIR, and
UV–VIS transmission spectra.

3. Results
A series of composites were prepared by impregnating the pure CNF films in different
epoxy/acetone solution ratios. The average thickness of CNF film ranged from 55–60 µm (Figure 1a).
The impregnation of CNF films in medium viscosity (epoxy and acetone at a ratio of 1:1) yielded the
best results. The composite cross-section showed well-dispersed fibrils or fibril aggregates within the
epoxy matrix (Figure 1c,d). The white spots on the matrix are very short pull out lengths of fibrils
or fibril aggregates indicating strong interfacial adhesion. The average thickness of the composites
impregnated in medium viscosity ranged from 180–200 µm. However, the impregnation of CNF films
in higher viscosity (epoxy and acetone at a ratio of 3:1) and lower viscosity (epoxy and acetone at a
ratio of 1:3) did not yield good results due to the poor distribution of epoxy in and around the CNF
fibrils. Figure 2a,b shows the cross section of composites obtained from impregnation in high viscosity
epoxy solution. There was clear phase separation between the epoxy matrix and CNF film. Huge
amount of epoxy formed the face layers, while the CNF film formed the core with dimensions similar
to pure CNF film (50–70 µm). The total average thickness of the composite ranged from 290–320 µm.
This implies that there was only limited penetration of epoxy into the CNF film. In addition, there
were voids within the sample due to high viscosity of epoxy. For films impregnated in low viscosity
solution (Figure 2c,d), the cross-sections showed non-uniform distribution of fibril within the matrix.
There were areas totally devoid of any fibrils, sparse distribution of fibrils, and areas without any resin.
The average thickness of composite ranged from 88–96 µm.
The medium viscosity led to good impregnation and resulted in composites with different CNF
contents ranging from 18 to 23 wt %. Figure 3 shows the change in mechanical properties with varying
CNF wt %. Table 1 shows the average mechanical properties of pure CNF film, neat cured epoxy resin,
and the composite. The representative stress-strain curves for the tensile tests are shown in Figure 4.
The strong reinforcement of CNF on the epoxy polymer matrix is quite evident from different loading
of fibrils (18–23 wt %). The strain energy density or modulus of toughness was calculated as the total
area under the stress strain curves. The modulus of the epoxy increased from an average value of
2.28 to 3.56 MPa, the tensile strength increased from 62.5 to 108.25 MPa, and the strain increased from
2.92% to 4.81%, which led to an overall increase in modulus of toughness by almost 184 times (0.98 to
2.79 MPa) for the composites compared to the neat epoxy.
Polymers 2019, 11, 612 5 of 13
Polymers
Polymers 2019,
2019, 11,
11, xx FOR
FOR PEER
PEER REVIEW
REVIEW 55 of
of 13
13

Figure
Figure 1. 1.
Figure SEM
1.SEM images
SEMimages
imagesofof fracture
offracture surfaces
surfaces of
fracture surfaces of (a)
of (a) Pure
(a) Pure CNF
Pure CNFfilm;
CNF film;(b)
film; (b)neat
(b) neatepoxy
neat epoxy
epoxy film; (c)
film;
film; 23
(c)(c) wt
23 23 %
wtwt
%%
CNF/epoxy
CNF/epoxy composite
CNF/epoxycomposite (low
composite(low resolution);
(lowresolution); and (d)
resolution); and (d) 23
(d) 23 wt
23 wt %
wt% CNF/epoxy
%CNF/epoxy
CNF/epoxy composite
composite
composite (high resolution).
(high
(high resolution).
resolution).

Figure
Figure
Figure2. 2. SEM
2.SEM images
SEMimages
images of of fracture
fracturesurfaces
fracture surfacesof
surfaces ofofcured
cured
cured composite
composite
composite made from
made
made impregnating
from
from impregnating
impregnating CNFs in
in (a)
CNFsCNFs (a) in
epoxy/acetone
(a)epoxy/acetone solution
epoxy/acetonesolution at a ratio
solutionatata aratio of 3:1
ratioofof3:1 (low
3:1(low resolution);
(lowresolution); (b) epoxy/acetone
resolution);(b)(b)epoxy/acetone
epoxy/acetonesolution at
solution
solution a ratio
at aatratio of
of 3:1
a ratio of 3:1
3:1
(high
(high
(high resolution);
resolution);
resolution); (c)(c)
(c) epoxy/acetone solution
solution at
epoxy/acetone solution
epoxy/acetone at aa ratio
ratio of
of 1:3
of 1:3 (low
1:3 (lowresolution);
(low resolution);and
resolution); and(d)
and (d)epoxy/acetone
(d) epoxy/acetone
epoxy/acetone
solution
solution
solution at
at at a ratio
a ratio
a ratio of
ofof1:31:3 (high
1:3(high resolution).
(highresolution).
resolution).
 epoxy
epoxyresin,
resin,andandthethecomposite.
composite.The The representative stress-straincurves
representative stress-strain curvesfor
for the
the tensile
tensile tests
tests areare shown
shown
in in
Figure
Figure 4. 4.The
Thestrong
strongreinforcement
reinforcement of of CNF
CNF on on the
the epoxy
epoxypolymer
polymermatrix
matrixis is quite
quite evident
evident fromfrom
different
different loading
loadingofoffibrils
fibrils(18–23
(18–23 wt
wt %).
%). The strain
strain energy
energydensity
densityorormodulus
modulus of of toughness
toughness waswas
calculated
calculated asasthethetotal
totalarea
areaunder
underthethe stress strain curves.
stress strain curves.The
Themodulus
modulusofofthethe epoxy
epoxy increased
increased fromfrom
anan average
average value
valueofof2.28
2.28toto3.56
3.56MPa,
MPa, the
the tensile strengthincreased
tensile strength increasedfrom
from62.5
62.5
toto 108.25
108.25 MPa,
MPa, andand
thethe
strain
strain
Polymers 2019, 11,increased
increased
612 from2.92
from 2.92%%to to4.81
4.81 %,
%, which
which led
led to
to an
anoverall
overallincrease
increaseininmodulus
modulus of of
toughness
toughness by by6 of 13
almost
almost 184 184 times(0.98
times (0.98toto2.79
2.79MPa)
MPa) for
for the
the composites
composites compared
comparedtotothe
theneat epoxy.
neat epoxy.

Figure 3. Mechanical properties of the CNF/epoxy composites.

Figure
Figure 3. Mechanical
3. Mechanical propertiesof
properties ofthe
the CNF/epoxy
CNF/epoxy composites.
composites.
Table 1. Mechanical properties of CNF film cured neat epoxy resin, and composites.
1. Mechanical
TableTable properties
1. Mechanical ofofCNF
properties CNFfilm
filmcured neat epoxy
cured neat epoxyresin,
resin,and
and composites.
composites.
Tensile Strength Tensile Modulus Tensile Strain Strain Energy
Sample
Tensile(MPa)
Strength
Tensile (GPa)
Strength Tensile
TensileModulus (%)
Modulus Tensile Strain Density
Strain (MPa)
StrainEnergy
Energy
Sample
Sample Tensile Strain (%)
Pure CNF film 220 ±
(MPa) 21.7
(MPa) 7.7 ± 0.85
(GPa)
(GPa) 5.0 ± 0.47
(%) 7.53 ± 0.4(MPa)
Density
Density (MPa)
Pure
Pure Epoxy
CNF
CNF film
film 62.5 ±
220 220 5.86
± 21.7
± 21.7 2.28 ±
7.77.7 0.41
± 0.85
± 0.85 2.92 ± 0.33± 0.47
5.0 ±5.00.47 0.98 ± 0.06
7.53
7.53 ±± 0.40.4
Composite (18–23 wt
Epoxy 62.5 ± 5.86 ± 0.41 ± 0.33 ± 0.06
Epoxy 62.5
108.5± ±5.86
8.43 2.282.28
3.56 ±±0.28
0.41 2.92±2.92
4.81 ±1.04
0.33 0.98
0.98
2.79 ± 0.06
± 0.49
Composite% CNF)wt % CNF)
(18–23 108.5 ± 8.43 3.56 ± 0.28 4.81 ± 1.04 2.79 ± 0.49
Composite (18–23 wt
108.5 ± 8.43 3.56 ± 0.28 4.81 ± 1.04 2.79 ± 0.49
% CNF)

Figure 4. Representative
Figure stress-strain
4. Representative curves
stress-strain curvesof
of the CNFfilm
the CNF filmcured
cured neat
neat epoxy
epoxy resin,resin, and composite.
and composite.

Furthermore, the experimental data were compared with the predicted values using the rules of
Figure 4. Representative stress-strain curves of the CNF film cured neat epoxy resin, and composite.
mixture (ROM). The predicted values were obtained using the following equations:

Ecomposite = Vf Ef + (1 − Vf ) Em (2)

σcomposite = Vf σf + (1 − Vf ) σm (3)

Here, Ecomposite and σcomposite are the predicted modulus and tensile strength for the composites,
respectively. Ef, Em, σf, σm is the tensile modulus of CNF, tensile modulus of matrix, tensile strength of
CNF, and tensile strength of matrix, respectively. Vf denotes the volume fraction of CNF. The tensile
property of pure CNF film (Ef, σf ), neat epoxy (Em, σm ) obtained from Table 1 were used to fit in the
predicted model. The predicted strength and modulus values for 18 wt % loading (Vf = 13.7%) were
84.08 MPa and 3.02 GPa, respectively. The predicted strength and modulus values for 23 wt % loading
(Vf = 17.77%) were 90.49 MPa and 3.24 GPa, respectively. The experimental values (Figure 3) for
18 wt % (Ecomposite = 3.31 GPa and σcomposite = 97.11) and 23 wt % (Ecomposite = 3.79 GPa and σcomposite =
118.7 MPa) were higher than the predicted values.
Polymers 2019, 11, 612 7 of 13

Cellulose molecules show high polarity and, therefore, need chemical functionalization of the
CNF to have a significant reinforcement effect on the epoxy matrix [30]. A good interaction between
the CNF and epoxy can create an effective interphase region around fibrils. The stress in the interphase
region will be higher than the neat epoxy, and could contribute to higher volume fraction of fibrils,
thereby contributing to the high mechanical properties [31]. The interphase or the increase in volume
fraction due to interphase is not taken into consideration for the ROM prediction model, which could
result in lower predicted values. The results imply that the CNF used in this study showed good
compatibility with the epoxy matrix, resulting in high mechanical properties. The composite tensile
properties obtained in this study were comparable to, or even better than, some of the petroleum-based
epoxy/CNF composites in the literature [28,29,32]. The reinforcement obtained by the addition CNF
Polymers 2019, 11, x FOR PEER REVIEW 7 of 13
were even better than some of the glass fiber reinforced epoxy composites [33,34].
FigureFurthermore,
5 shows the theinfrared
experimentalspectra
data of pure
were CNF film,
compared cured
with the neatvalues
predicted epoxy, andthe
using composite
rules of with
23 wt %mixture
CNF. (ROM).
CNF film showed major absorption peaks at 3330 cm −1 (O-H stretching), 1161 cm−1
The predicted values were obtained using the following equations:
(C-O-C asymmetric stretching), 990 cm−1 and 1056 cm−1 (C-O stretching), 1369 cm−1 (symmetric
Ecomposite = VfEf + (1 − Vf) Em (2)
C-H deformation), 1425 cm−1 (asymmetric C-H deformation) associated with various polysaccharides
(cellulose and hemicellulose) [35,36]. σcomposite = Vfσf +neat
The cured (1 − Vepoxy
f) σm showed major peaks at 3370 (3) cm−1 (O-H
stretching),Here,
1607Ecm − 1 − 1
(C=C stretching in phenyl groups), 1036 cm (C-O symmetric stretching in
composite and σcomposite are the predicted modulus and tensile strength for the composites,

ether groups), 1334 cm −1 (C-H deformation in epoxide groups), 1295 cm−1 (C-H vibration in epoxide
respectively. Ef, Em, σf, σm is the tensile modulus of CNF, tensile modulus of matrix, tensile strength of
groups),CNF,
andand cm−1strength
915tensile (in-plane of matrix,
asymmetricrespectively. Vf denotesofthe
deformation volume ring)
epoxide fraction of CNF. The
[35,37,38]. tensile
For composites,
property of pure CNF film (E f, σf), neat epoxy (Em, σm) obtained from Table 1 were used to fit in the
the infrared spectra were obtained on the polished surface. This was done mainly to remove the
predicted model. The predicted strength and modulus values for 18 wt % loading (Vf = 13.7%) were
capping of epoxy layers on the surface (Figure 1c). In addition to hydroxyl peaks, there was significant
84.08 MPa and 3.02 GPa, respectively. The predicted strength and modulus values for 23 wt % loading
weakening of peakwere
(Vf = 17.77%) intensities
90.49 MPa inandthe3.24
composite sample The
GPa, respectively. compared to epoxy,
experimental values indicative
(Figure 3) forof18possible
interaction between CNF and resin. The epoxide groups react with amines
wt % (Ecomposite = 3.31 GPa and σcomposite = 97.11) and 23 wt % (Ecomposite = 3.79 GPa and σcomposite as well as the
= 118.7 MPa) hydroxyl
groups were
in CNF
higher during
than the curing [23,35].
predicted values. Compared to cured epoxy, the composite samples showed a
reduction inCellulose
intensity molecules
or completeshow high polarity and, of
disappearance therefore,
absorptionneed peaks
chemical atfunctionalization
1295 and 915 cm of − 1 , related
the
CNF to have a significant reinforcement effect on the epoxy matrix [30]. A good interaction between
to the epoxide/oxirane rings in the resin. The CNF has high surface area with enormous amount of
the CNF and epoxy can create an effective interphase region around fibrils. The stress in the
hydroxyl groups, which reacts with epoxides forming stable ether linkages [39]. In addition, several
interphase region will be higher than the neat epoxy, and could contribute to higher volume fraction
studiesofhave highlighted
fibrils, the strong
thereby contributing to theaccelerating
high mechanical effect of celluloses
properties [31]. The or other hydroxyl
interphase groups on
or the increase
epoxy curing
in volume [37,40].
fraction Figure
due to 6interphase
shows the images
is not andconsideration
taken into the light transmittance measurements
for the ROM prediction model, versus
which could
the wavelength for result
pure in CNF,lower predicted
epoxy, values. The results
and composite films. At imply that the
600–800 nm CNF used in thisthe
wavelength, study
composite
samplesshowed good compatibility
transmitted 86–88% ofwith the the epoxy
light. Thematrix,
additionresulting in high mechanical
of uniform nano-sized properties.
cellulose Thefibrils in
composite tensile properties obtained in this study were comparable to, or even better than, some of
composites did not affect the light transmittance, when compared to neat epoxy films. The high light
the petroleum-based epoxy/CNF composites in the literature [28,29,32]. The reinforcement obtained
transmittance confirms the uniform dispersion of nanofibrils in the epoxy matrix [41].
by the addition CNF were even better than some of the glass fiber reinforced epoxy composites [33,34].

FigureFigure 5. FTIR
5. FTIR spectra
spectra representative of
representative of CNF,
CNF,neat
neatcured
curedepoxy resin,
epoxy and composites.
resin, and composites.

Figure 5 shows the infrared spectra of pure CNF film, cured neat epoxy, and composite with
23 wt % CNF. CNF film showed major absorption peaks at 3330 cm−1 (O-H stretching), 1161 cm−1 (C-
O-C asymmetric stretching), 990 cm−1 and 1056 cm−1 (C-O stretching), 1369 cm−1 (symmetric C-H
deformation), 1425 cm−1 (asymmetric C-H deformation) associated with various polysaccharides
(cellulose and hemicellulose) [35,36]. The cured neat epoxy showed major peaks at 3370 cm−1 (O-H
stretching), 1607 cm−1 (C=C stretching in phenyl groups), 1036 cm−1 (C-O symmetric stretching in ether
groups), 1334 cm−1 (C-H deformation in epoxide groups), 1295 cm−1 (C-H vibration in epoxide groups),
 studies have highlighted the strong accelerating effect of celluloses or other hydroxyl groups on
epoxy curing [37,40]. Figure 6 shows the images and the light transmittance measurements versus
the wavelength for pure CNF, epoxy, and composite films. At 600–800 nm wavelength, the composite
samples transmitted 86–88% of the light. The addition of uniform nano-sized cellulose fibrils in
composites
Polymers 2019, 11,did
612not affect the light transmittance, when compared to neat epoxy films. The high light
8 of 13
transmittance confirms the uniform dispersion of nanofibrils in the epoxy matrix [41].

Figure
Figure 6.6.(a)
(a)Images
Imagesofofpure
pureCNF
CNF(A),
(A), cured
cured epoxy resin
resin only
only(B),
(B),and
andthe
the2323wt
wt%%composite film
composite (C).
film (C).
(b) Light transmittance measurements versus the wavelength for pure CNF, the cured epoxy
(b) Light transmittance measurements versus the wavelength for pure CNF, the cured epoxy resin, and resin,
theand
23 the
wt % 23 composite.
wt % composite.

Figure
Figure 7 7shows
showsthe theTGA
TGAand andDTGDTG curves
curves for pure CNF, epoxy
epoxyand and2323wtwt%%composite.
composite.The Thepure
pure
CNF
CNF showed
showed aT a onset
TonsetandandTTmaxmax of
of286
286and
and321 ◦
321°C,C,respectively.
respectively.Despite
Despite being
beingabundant
abundant and sustainable,
and sustainable,
thethe CNFs
CNFs arestill
are stilllimited
limitedinincomposite
composite industry
industry mainly
mainly due
dueto toits
itslow
lowthermal
thermalstability [42].
stability However,
[42]. However,
epoxies are a class of thermosetting polymers with a high thermal stability. The epoxy resinused
epoxies are a class of thermosetting polymers with a high thermal stability. The epoxy resin used inin
our study showed a T onset and Tmax of 323 and 355 °C, respectively.
◦ The composites
our study showed a Tonset and Tmax of 323 and 355 C, respectively. The composites with 23 wt % with 23 wt % CNF
loading
CNF showed
loading showed a Tonset
a Tand Tmax of 308 and 345 °C, respectively.
◦ The decrease in Tonset and Tmax was
onset and Tmax of 308 and 345 C, respectively. The decrease in Tonset and
Tmax was only 5% and 3%, respectively, in the composite comparedthe
only 5% and 3%, respectively, in the composite compared to purepure
to the epoxy.
epoxy.TheTheresults areare
results
impressive, given the fact that we have added high amount of cellulose nanofibrils
impressive, given the fact that we have added high amount of cellulose nanofibrils in the composite. in the composite.
The high thermal stability for the composite can be attributed to effective interaction between fibrils
The high thermal stability for the composite can be attributed to effective interaction between fibrils
and epoxy [23]. Figure 8 shows the storage modulus as a function of temperature for pure CNF,
and epoxy [23]. Figure 8 shows the storage modulus as a function of temperature for pure CNF, epoxy,
epoxy, and 23 wt % composite. The presence of CNF had a strong reinforcing effect in both glassy
and 23 wt % composite. The presence of CNF had a strong reinforcing effect in both glassy and glass
and glass transition region. The storage modulus significantly increased with the addition of 23 wt %
transition region. The storage modulus significantly increased with the addition of 23 wt % CNFs.
CNFs. Figure 9 shows the tan δ as a function of temperature for epoxy and 23 wt % composite. The
Figure 9 shows the tan δ as a function of temperature for epoxy and 23 wt % composite. The glass
glass transition temperature (Tg◦) at 91 °C did not show any significant difference between the epoxy
transition temperature (Tg ) at 91a C
and the composite. However,
did not show any significant difference between the epoxy and the
significant decrease in intensity in tan δ peak for the epoxy matrix
composite. However, a significant
occurred with the addition of CNFs. The resultsdecrease in intensity in there
imply that tan δwaspeakhigh
for interaction
the epoxy matrix
betweenoccurred
fibrils
with the addition of CNFs. The results
and epoxy during the phase transition [25]. imply that there was high interaction between fibrils and epoxy
during the phase transition [25].
Figure 10 shows the water vapour transmission of the neat epoxy, 23 wt % composite and the
pure CNF film. The permeability of water vapour molecules was reduced by more than 60 % with the
addition of CNFs to epoxy. The numerical values for WVT are given in Table S1. The water vapour
permeability through polymer film depends on the dissolution of water molecules and its rate of
diffusion in the film [27,43]. The diffusion of molecules in epoxy film takes place mainly through the
interaction between the water molecules and epoxy matrix. Water molecules can hydrogen bond with
either the hydroxyl groups or the amine groups in the resin [44]. The presence of hydroxyl groups
in the cured epoxy was evident from the FTIR results. The water vapour barrier efficiency of pure
CNF film is low mainly due to high hydrophilicity of cellulose molecules. However, the addition of
highly crystalline, high aspect ratio CNFs can increase the tortuosity of permeating molecules within
the epoxy matrix [27]. Different studies have showed that the water vapour efficiency of epoxy can be
improved by using inorganic nanoparticles [45,46]. Water molecules can also diffuse through minute
nano or micro voids within the film. Therefore, the high-water vapour barrier efficiency of composites
also indicates the effective adhesion or interphase formation between the epoxy and CNFs. The WVT
values we obtained for the composites can be compared to some of the films made from commercial
petroleum-based polymers [27,47].
Polymers 2019, 11, 612 9 of 13
Polymers 2019, 11, x FOR PEER REVIEW 9 of 13

Polymers 2019, 11, x FOR PEER REVIEW 9 of 13

Figure
Figure 7. 7.
(a)(a)TGA
TGAand
and(b)
(b)DTG
DTG curves
curvesfor
forpure
pureCNF,
CNF,epoxy, and
epoxy, thethe
and 23 wt
23 %
wtcomposite.
% composite.
Figure 7. (a) TGA and (b) DTG curves for pure CNF, epoxy, and the 23 wt % composite.

Figure 8. 8.Temperature
Figure dependence
Temperature dependence of storage
of the the storage modulus
modulus E0 CNF,
E′ for pure for pure CNF,
epoxy, epoxy,
and the 23 wtand
% the
23 wt % composite.
composite.

Figure 8. Temperature dependence of the storage modulus E′ for pure CNF, epoxy, and the 23 wt %
composite.
Polymers 2019, 11, 612 10 of 13
Polymers 2019, 11, x FOR PEER REVIEW 10 of 13

Figure 9. Temperature dependence of the tan δ for pure CNF, epoxy, and the 23 wt % composite.

Figure 10 shows the water vapour transmission of the neat epoxy, 23 wt % composite and the
pure CNF film. The permeability of water vapour molecules was reduced by more than 60 % with
the addition of CNFs to epoxy. The numerical values for WVT are given in Table S1. The water vapour
permeability through polymer film depends on the dissolution of water molecules and its rate of
diffusion in the film [27,43]. The diffusion of molecules in epoxy film takes place mainly through the
interaction between the water molecules and epoxy matrix. Water molecules can hydrogen bond with
either the hydroxyl groups or the amine groups in the resin [44]. The presence of hydroxyl groups in
the cured epoxy was evident from the FTIR results. The water vapour barrier efficiency of pure CNF
film is low mainly due to high hydrophilicity of cellulose molecules. However, the addition of highly
crystalline, high aspect ratio CNFs can increase the tortuosity of permeating molecules within the
epoxy matrix [27]. Different studies have showed that the water vapour efficiency of epoxy can be
improved by using inorganic nanoparticles [45,46]. Water molecules can also diffuse through minute
nano or micro voids within the film. Therefore, the high-water vapour barrier efficiency of
composites also indicates the effective adhesion or interphase formation between the epoxy and
CNFs. The WVT values we obtained for the composites can be compared to some of the films made
from commercial petroleum-based polymers [27,47].
Figure 9. Temperature
Figure dependence
9. Temperature dependenceof ofthe
the tan for pure
tan δδ for pureCNF,
CNF,epoxy,
epoxy,and
andthethe
23 23
wt wt % composite.
% composite.

Figure 10 shows the water vapour transmission of the neat epoxy, 23 wt % composite and the
pure CNF film. The permeability of water vapour molecules was reduced by more than 60 % with
the addition of CNFs to epoxy. The numerical values for WVT are given in Table S1. The water vapour
permeability through polymer film depends on the dissolution of water molecules and its rate of
diffusion in the film [27,43]. The diffusion of molecules in epoxy film takes place mainly through the
interaction between the water molecules and epoxy matrix. Water molecules can hydrogen bond with
either the hydroxyl groups or the amine groups in the resin [44]. The presence of hydroxyl groups in
the cured epoxy was evident from the FTIR results. The water vapour barrier efficiency of pure CNF
film is low mainly due to high hydrophilicity of cellulose molecules. However, the addition of highly
crystalline, high aspect ratio CNFs can increase the tortuosity of permeating molecules within the
epoxy matrix [27]. Different studies have showed that the water vapour efficiency of epoxy can be
improved by using inorganic nanoparticles [45,46]. Water molecules can also diffuse through minute
nano or micro voids within the film. Therefore, the high-water vapour barrier efficiency of
composites also indicates the effective adhesion or interphase formation between the epoxy and
10. Water
FigureFigure vapour
10. Water transmission
vapour transmission of the
thepure
pureCNF,
CNF, epoxy,
epoxy, andand the
the 23 wt23 wt % composite.
CNFs. The WVT values we obtained for the composites can be compared to % composite.
some of the films made
from commercial
4. Conclusions petroleum-based polymers [27,47].

A strong reinforcement of CNFs on the epoxy polymer matrix was quite evident from different
loading of CNFs (18–23 wt %). The average tensile modulus increased from 2.28 to 3.56 MPa, the tensile
strength increased from 62.5 to 108.25 MPa, and the strain increased from 2.92% to 4.81%, which led
to an overall increase in modulus of toughness by 184 times for the composites compared to the neat
epoxy. The addition of fibrils also decreased the water vapor permeability of the neat epoxy, increasing
the water vapour barrier efficiency of resulting composites by more than fifty percent. All these results
indicate an effective interaction between the CNFs and epoxy used in this study. Also, the addition
of substantial amount of cellulose fibrils (23 wt %) in the resulting composites did not affect the high
thermal stability of the epoxy matrix. With improved mechanical properties, the commercial biobased
epoxy resin reinforced with CNFs obtained in our study has great potential for various structural
applications. The use of biobased epoxy and the addition of CNFs also highlight their environmental
suitability. The low water vapour permeability obtained by the addition of CNFs in the epoxy are
favorable attributes and can be extremely desirable for other applications such as coating or adhesives.

Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4360/11/4/612/s1,

Figure 10. Water vapour transmission of the pure CNF, epoxy, and the 23 wt % composite.
Table S1. Water vapour transmission of the epoxy, composite and the CNF film.
Author Contributions: Investigation: S.S.N.; resources, C.D., D.B.L., and N.Y.; writing—original draft preparation,
S.S.N.; writing—review and editing: C.D., D.B.L., and N.Y.; supervision: N.Y.; funding acquisition: D.B.L. and N.Y.
Polymers 2019, 11, 612 11 of 13

Funding: The authors would also like to thank Genome Canada for funding support through the Genomics
applied Partnership Program (GAPP) grant titled, “Fibre composite and biomatrix genomics (FiCoGen).
Acknowledgments: The authors would also like to thank Ecopoxy Inc. for providing the commercial biobased
epoxy samples. We would also like to thank Center for Biocomposites and Biomaterials Processing, University of
Toronto, for supplying the cellulose nanofibril samples.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Auvergne, R.; Caillol, S.; David, G.; Boutevin, B.; Pascault, J.P. Biobased thermosetting epoxy: Present and
future. Chem. Rev. 2014, 114, 1082–1115. [CrossRef] [PubMed]
2. Roudsari, G.M.; Mohanty, A.K.; Misra, M. Green Approaches to Engineer Tough Biobased Epoxies: A Review.
ACS Sustain. Chem. Eng. 2017, 5, 9528–9541. [CrossRef]
3. Shen, L.; Haufe, J.; Patel, M.K. Product Overview and Market Projection of Emerging Bio-Based Plastics
PRO-BIP 2009; Report for European Polysaccharide Network of Excellence (EPNOE) and European
Bioplastics; European Polysaccharide Network of Excellence (EPNOE) and European Bioplastics: Utrecht,
The Netherlands, 2009.
4. Celikbag, Y.; Meadows, S.; Barde, M.; Adhikari, S.; Buschle-Diller, G.; Auad, M.L.; Via, B.K. Synthesis and
characterization of bio-oil-based self-curing epoxy resin. Ind. Eng. Chem. Res. 2017, 57, 9389–9400. [CrossRef]
5. Unnikrishnan, K.P.; Thachil, E.T. Toughening of epoxy resins. Des. Monomers Polym. 2006, 9, 129–152.
[CrossRef]
6. Day, R.; Lovell, P.; Pierre, D. Toughening of epoxy resins using particles prepared by emulsion polymerization:
Effects of particle surface functionality, size and morphology on impact fracture properties. Polym. Int. 1997,
44, 288–299. [CrossRef]
7. Kishi, H.; Uesawa, K.; Matsuda, S.; Murakami, A. Adhesive strength and mechanisms of epoxy resins
toughened with pre-formed thermoplastic polymer particles. J. Adhes. Sci. Technol. 2005, 19, 1277–1290.
[CrossRef]
8. Spanoudakis, J.; Young, R.J. Crack propagation in a glass particle-filled epoxy resin. J. Mater. Sci. 1984, 19,
473–486. [CrossRef]
9. Boogh, L.; Pettersson, B.; Månson, J.A.E. Dendritic hyperbranched polymers as tougheners for epoxy resins.
Polymer 1999, 40, 2249–2261. [CrossRef]
10. Könczöl, L.; Döll, W.; Buchholz, U.; Mülhaupt, R. Ultimate properties of epoxy resins modified with a
polysiloxane–polycaprolactone block copolymer. J. Appl. Polym. Sci. 1994, 54, 815–826. [CrossRef]
11. Bagheri, R.; Marouf, B.T.; Pearson, R.A. Rubber-toughened epoxies: A critical review. Polym. Rev. 2009, 49,
201–225. [CrossRef]
12. Chen, M.C.; Hourston, D.J.; Sun, W.B. The morphology and fracture behaviour of a miscible epoxy
resin-polyetherimide blend. Eur. Polym. J. 1995, 31, 199–201. [CrossRef]
13. Mezzenga, R.; Boogh, L.; Månson, J.A.E. A review of dendritic hyperbranched polymer as modifiers in epoxy
composites. Compos. Sci. Technol. 2001, 61, 787–795. [CrossRef]
14. Miyagawa, H.; Misra, M.; Drzal, L.T.; Mohanty, A.K. Fracture toughness and impact strength of
anhydride-cured biobased epoxy. Polym. Eng. Sci. 2005, 45, 487–495. [CrossRef]
15. Kuo, P.Y.; de Assis Barrosb, L.; Yan, N.; Sain, M.; Qing, Y.; Wu, Y. Nanocellulose composites with enhanced
interfacial compatibility and mechanical properties using a hybrid-toughened epoxy matrix. Carbohydr.
Polym. 2017, 177, 249–257. [CrossRef] [PubMed]
16. Xu, S.; Girouard, N.; Schueneman, G.; Shofner, M.L.; Meredith, J.C. Mechanical and thermal properties of
waterborne epoxy composites containing cellulose nanocrystals. Polymer 2013, 54, 6589–6598. [CrossRef]
17. Lavoine, N.; Desloges, I.; Dufresne, A.; Bras, J. Microfibrillated cellulose—Its barrier properties and
applications in cellulosic materials: A review. Carbohydr. Polym. 2012, 90, 735–764. [CrossRef] [PubMed]
18. Khalil, H.P.S.A.; Bhat, A.H.; Yusra, A.F.I. Green composites from sustainable cellulose nanofibrils: A review.
Carbohydr. Polym. 2012, 87, 963–979. [CrossRef]
19. Spence, K.L.; Venditti, R.A.; Rojas, O.J.; Habibi, Y.; Pawlak, J.J. A comparative study of energy consumption
and physical properties of microfibrillated cellulose produced by different processing methods. Cellulose
2011, 18, 1097–1111. [CrossRef]
Polymers 2019, 11, 612 12 of 13

20. Nair, S.S.; Zhu, J.Y.; Deng, Y.; Ragauskas, A.J. Characterization of cellulose nanofibrillation by micro grinding.
J. Nanopart. Res. 2014, 16, 2349. [CrossRef]
21. Saito, T.; Kuramae, R.; Wohlert, J.; Berglund, L.A.; Isogai, A. An ultrastrong nanofibrillar biomaterial: The
strength of single cellulose nanofibrils revealed via sonication-induced fragmentation. Biomacromolecules
2013, 14, 248–253. [CrossRef] [PubMed]
22. Matsuo, M.; Sawatari, C.; Iwai, Y.; Ozaki, F. Effect of Orientation Distribution and Crystallinity on the
Measurement by x-ray Diffraction of the Crystal Lattice Moduli of Cellulose I and II. Macromolecules 1990, 23,
3266–3275. [CrossRef]
23. Nair, S.S.; Kuo, P.-Y.; Chen, H.; Yan, N. Investigating the effect of lignin on the mechanical, thermal, and
barrier properties of cellulose nanofibril reinforced epoxy composite. Ind. Crops Prod. 2017, 100, 208–217.
[CrossRef]
24. Chen, H.; Nair, S.S.; Chauhan, P.; Yan, N. Lignin containing cellulose nanofibril application in pMDI wood
adhesives for drastically improved gap-filling properties with robust bondline interfaces. Chem. Eng. J. 2019,
360, 393–401. [CrossRef]
25. Nair, S.S.; Chen, H.; Peng, Y.; Huang, Y.; Yan, N. Polylactic Acid Biocomposites Reinforced with Nanocellulose
Fibrils with High Lignin Content for Improved Mechanical, Thermal, and Barrier Properties. ACS Sustain.
Chem. Eng. 2018, 6, 10058–10068. [CrossRef]
26. Teixeira, E.d.M.; Pasquini, D.; Curvelo, A.A.S.; Corradini, E.; Belgacem, M.N.; Dufresne, A. Cassava bagasse
cellulose nanofibrils reinforced thermoplastic cassava starch. Carbohydr. Polym. 2009, 78, 422–431. [CrossRef]
27. Nair, S.S.; Zhu, J.; Deng, Y.; Ragauskas, A.J. High performance green barriers based on nanocellulose. Sustain.
Chem. Process. 2014, 2, 1–7. [CrossRef]
28. Ansari, F.; Galland, S.; Johansson, M.; Plummer, C.J.G.; Berglund, L.A. Cellulose nanofiber network for
moisture stable, strong and ductile biocomposites and increased epoxy curing rate. Compos. Part A Appl. Sci.
Manuf. 2014, 63, 35–44. [CrossRef]
29. Lee, K.Y.; Tammelin, T.; Schulfter, K.; Kiiskinen, H.; Samela, J.; Bismarck, A. High performance cellulose
nanocomposites: Comparing the reinforcing ability of bacterial cellulose and nanofibrillated cellulose. ACS
Appl. Mater. Interfaces 2012, 4, 4078–4086. [CrossRef]
30. Lu, J.; Askeland, P.; Drzal, L.T. Surface modification of microfibrillated cellulose for epoxy composite
applications. Polymer 2008, 49, 1285–1296. [CrossRef]
31. Nair, S.S.; Wang, S.; Hurley, D.C. Nanoscale characterization of natural fibers and their composites using
contact-resonance force microscopy. Compos. Part A Appl. Sci. Manuf. 2010, 41, 624–631. [CrossRef]
32. Bruce, D.M.; Hobson, R.N.; Farrent, J.W.; Hepworth, D.G. High-performance composites from low-cost plant
primary cell walls. Compos. Part A Appl. Sci. Manuf. 2005, 36, 1486–1493. [CrossRef]
33. Asadi, A.; Miller, M.; Moon, R.J.; Kalaitzidou, K. Improving the Interfacial and Mechanical Properties of
Short Glass fiber/Epoxy Composites by Coating the Glass Fibers with Cellulose Nanocrystals. Express Polym.
Lett. 2016, 10, 587–597. [CrossRef]
34. Miwa, M.; Ohsawa, T.; Tahara, K. Effects of Fiber Length on the Tensile Strength of Epoxy-Glass Fiber and
Polyester-Glass Fiber Composites. J. Appl. Polym. Sci. 1980, 25, 795–807. [CrossRef]
35. Karumuri, S.; Hiziroglu, S.; Kalkan, A.K. Thermoset-cross-linked lignocellulose: A moldable plant biomass.
ACS Appl. Mater. Interfaces 2015, 7, 6596–6604. [CrossRef]
36. Pandey, K.K. A study of chemical structure of soft and harwood and wood polymers by FTIR spectrscopy. J.
Appl. Polym. Sci. 1999, 71, 1969–1975. [CrossRef]
37. Varma, A.J.; Jamdade, Y.K.; Nadkarni, V.M. Curing Characteristics of Epoxy-Resins Filled with Cellulose
and Oxidized Cellulose. Angew. Makromol. Chem. 1984, 122, 211–218. [CrossRef]
38. Mertzel, E.; Koenig, J. Application of FT-IR and NMR to epoxy resins. In Epoxy Resins and Composites II;
Springer: Berlin, Germany, 1986; pp. 73–112.
39. Kuo, P.Y.; Yan, N.; Sain, M. Influence of cellulose nanofibers on the curing behavior of epoxy/amine systems.
Eur. Polym. J. 2013, 49, 3778–3787. [CrossRef]
40. Partansky, A.M. A study of accelerators for epoxy-amine condensation reaction. Adv. Chem. 1970, 92, 29–47.
41. Abe, K.; Iwamoto, S.; Yano, H. Obtaining cellulose nanofibers with a uniform width of 15 nm from wood.
Biomacromolecules 2007, 8, 3276–3278. [CrossRef]
42. Poletto, M.; Zattera, A.J.; Forte, M.M.C.; Santana, R.M.C. Thermal decomposition of wood: Influence of
wood components and cellulose crystallite size. Bioresour. Technol. 2012, 109, 148–153. [CrossRef]
Polymers 2019, 11, 612 13 of 13

43. Yasuda, H.; Stannett, V. Permeation, solution, and diffusion of water in some high polymers. J. Polym. Sci.
1962, 57, 907. [CrossRef]
44. Soles, C.L.; Yee, A.F. Discussion of the molecular mechanisms of moisture transport in epoxy resins. J. Polym.
Sci. Part B Polym. Phys. 2000, 38, 792–802. [CrossRef]
45. Guadagno, L.; Vertuccio, L.; Sorrentino, A.; Raimondo, M.; Naddeo, C.; Vittoria, V.; Iannuzzo, G.; Calvi, E.;
Russo, S. Mechanical and barrier properties of epoxy resin filled with multi-walled carbon nanotubes. Carbon
2009, 47, 2419–2430. [CrossRef]
46. Sangermano, M.; Periolatto, M.; Signore, V.; Russo Spena, P. Improvement of the water-vapor barrier
properties of an uv-cured epoxy coating containing graphite oxide nanoplatelets. Prog. Org. Coat. 2017, 103,
152–155. [CrossRef]
47. Steven, M.D.; Hotchkiss, J.H. Comparison of flat film to total package water vapour transmission rates for
several commercial food wraps. Packag. Technol. Sci. 2002, 15, 17–27. [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/).