Biomacromolecules 2009, 10, 425–432 425

Cellulose Whiskers versus Microfibrils: Influence of the Nature

of the Nanoparticle and its Surface Functionalization on the
Thermal and Mechanical Properties of Nanocomposites
Gilberto Siqueira, Julien Bras, and Alain Dufresne*
Grenoble Institute of Technology (INP), The International School of Paper, Print Media and Biomaterials
(PAGORA), 461 rue de la Papeterie, BP 65 - F-38402 Saint Martin d′Hères Cedex, France
Received October 21, 2008; Revised Manuscript Received November 28, 2008

In the present work, nanowhiskers and microfibrillated cellulose (MFC) both extracted from sisal were used to
reinforce polycaprolactone (PCL). We report the influence of the nanoparticle’s nature on the mechanical and
thermal properties of the ensuing nanocomposites. The surface of both the nanoparticles was chemically modified
to improve their compatibilization with the polymeric matrix. N-Octadecyl isocyanate (C18H37NCO) was used as
the grafting agent. PCL nanocomposite films reinforced with sisal whiskers or MFC (raw or chemically modified)
were prepared by film casting. The thermal behavior (Tg, Tm, Tc, and degree of crystallinity) and the mechanical
properties of the nanocomposites in both the linear and the nonlinear range were determined using differential
scanning calorimetry (DSC), dynamical mechanical analysis (DMA), and tensile tests, respectively. Significant
differences were reported according to the nature of the nanoparticle and amount of nanofillers used as
reinforcement. It was also proved that the chemical treatment clearly improves the ultimate properties of the
nanocomposites.

Introduction physical properties and commercial availability make it very

Nowadays there is a simultaneous and growing interest in
developing biobased products and innovative process technolo-
gies that can reduce the dependence on fossil fuel and move to
a sustainable materials basis.1,2
In parallel, researchers have focused their works on the
processing of nanocomposites (materials with nanosized rein-
forcement) to enhance mechanical properties.3,4 Similarly to
traditional microcomposites, nanocomposites use a matrix where
the nanosized reinforcement elements are dispersed. The
reinforcement is currently considered as a nanoparticle when
at least one of its dimensions is lower than 100 nm. This
particular feature provides nanocomposites unique and outstand-
ing properties never found in conventional composites.4,5 There
are several possibilities to obtain nanoelements from renewable
resources depending on their biological origin (e.g., cellulose,
starch, chitin). We have focused our work on cellulose
nanoparticles.
Besides the low cost of the raw material, the use of cellulose
nanocrystals as a reinforcing phase in nanocomposites has
numerous well-known advantages.2,5-10 However, considering
application as reinforcement, cellulose nanoparticles can present
some disadvantages, for example, high moisture absorption, poor
wetability, incompatibility with most of polymeric matrices, and
limitation of processing temperature. Indeed, lignocellulosic
materials start to degrade near 220 °C, restricting the type of
matrix that can be used in association with natural fillers.5,11
Biobased nanocomposites are the next generation of materials
for the future.1,12 Polycaprolactone (PCL) is a striking candidate
as a matrix due to its biodegradability properties. PCL is a
semicrystalline polymer with a glass transition temperature
around -60 °C and a melting temperature of 59-60 °C. Its
* To whom correspondence should be addressed. E-mail:
alain.dufresne@efpg.inpg.fr.
10.1021/bm801193d CCC: $40.75
Published on Web 12/29/2008
attractive as a substitute material for nondegradable polymers.
Cellulose nanoparticles can be extracted from lignocellulosic
fibers and are available all around the World, with some of them
being abundant in tropical countries.13 Sisal fibers are obtained
from the leaves of the AgaVe sisalana14 being one of the most
widely used natural fibers and easily cultivated. It can be easily
found in India, Brazil, and Tanzania, the two formers being
considered as the main producing countries.13
Most studies found in the literature about biocomposites use
cellulose as simple “filler”, which in some cases contributes to
enhance the rigidity, but mostly embrittles the polymer. In the
present work, cellulose was used as a nanostructured high-
performance constituent, in the form of nanofibers (whiskers
or microfibrilated cellulose-MFC). Thus, the potential and
hierarchical structure of cellulose is totally exploited. The study
of cellulosic nanoparticles as a reinforcing phase in nanocom-
posite films started 20 years ago,15 and since this time, a huge
amount of literature has been devoted to nanocellulose and is
becoming a topical subject, as revealed from the abundant
literature. Different descriptors of these nanoparticles are used,
including whiskers, monocrystals, and nanocrystals. These
crystallites have also often been referred to in literature as
microfibrils, microcrystals, or microcrystallites, despite their
nanoscale dimensions. The term “whiskers” is used to designate
elongated crystalline rod-like nanoparticles, whereas the des-
ignation “microfibrils” should be used to designate long flexible
nanoparticles consisting of alternating crystalline and amorphous
strings. The obtaining of the former involves a specific step for
the digestion of amorphous cellulosic domains, generally acid
hydrolysis, whereas the latter is obtained from a mechanical
treatment. Although abundantly studied separately, no systematic
comparison between these two nanoparticles has been reported,
except a short communication in 2004.16
The aim of the present study was to evaluate the thermome-
chanical behavior of nanocomposite films obtained from PCL
 2009 American Chemical Society
426 Biomacromolecules, Vol. 10, No. 2, 2009
and cellulosic nanoparticles from sisal focusing on the differ-
ences induced by the nature of the nanoparticle (whiskers vs
microfibril) and its surface chemical modification.
Experimental Section
Materials. Native sisal fibers (AgaVe sisalana), originating from
northeast Brazil, were purchased in Mariana (Minas Gerais, Brazil).
Poly(caprolactone) (Mn ) 42500 g · mol-1, Mw ) 65000 g · mol-1),
sulphuric acid (g95 wt %), and N-octadecyl isocyanate were obtained
from Aldrich. Ethanol, acetone, chloroform, toluene, and dichlo-
romethane were purchased from Chimie-Plus.
Cellulose Whiskers. Sisal fibers were cut with a FRITSCH Pulveri-
sette mill, until fine particulate fibers were obtained. Then the fibers
were treated with a 4 wt % NaOH solution at 80 °C for 2 h under
mechanical stirring. This treatment was done three times to purify
cellulose by removing other constituents present in the fibers. After
each treatment, fibers were filtered and washed with distilled water
until the alkali was completely eliminated. A subsequent bleaching
treatment was carried out to bleach the fibers. The solution used in
this treatment consisted of equal parts of acetate buffer, aqueous chlorite
(1.7 wt % in water), and distilled water. The bleaching treatment was
performed at 80 °C for 4 h under mechanical stirring and was repeated
four times. After each treatment, the fibers were filtered and washed
with distilled water. Acid hydrolysis was achieved at 50 °C with 65
wt % sulfuric acid (preheated), for about 40 min, under mechanical
stirring. The fiber content during all these chemical treatments was in
the range 4-6 wt %. The suspension was diluted with ice cubs to stop
the reaction and washed until neutrality by successive centrifugations
at 10000 rpm at 10 °C for 10 min each step and dialyzed against
distillated water, in the sequence. Afterward the sisal whiskers
suspension was homogenized by using an Ultra Turax T25 homogenizer
for 5 min and filtered using glass filter No. 1. Some drops of chloroform
were added to the whiskers suspension which was stored at 4 °C.
Microfibrillated Cellulose (MFC). A 2.0% (wt/v) suspension of
bleached sisal MFC was pumped through a microfluidizer processor
(Model M-110 EH-30). The slurry was passed through the valves that
applied a high pressure. Size reduction of products occurs into
Interaction Chamber (IXC) using cellules of different sizes (400 and
200 µm). Pumping cycles were varied to optimize the fibrillation
process.
Surface Chemical Modification. The surface chemical modification
of cellulosic nanoparticles was performed in toluene. To avoid the
drying of the nanoparticles that undoubtedly should lead to a strong
aggregation process we used never dried cellulose whiskers or MFC.
A solvent exchange procedure from water to toluene was used. For
that, an aqueous suspension with the desired amount of cellulose
nanoparticles (1 wt %) was solvent exchanged to acetone and then to
dry toluene by several successive centrifugations and redispersion
operations. Sonication was performed after each solvent exchange step
to avoid aggregation.
In a three-necked round-bottomed flask, equipped with a reflux
condenser, 3 g of a whisker or MFC suspension in toluene and 100
mL of toluene were added. The system was kept in a nitrogen
atmosphere. An excess of n-octadecyl isocyanate (16.9 g) was added
drop by drop when the temperature of the system reached 90 °C. The
temperature was then increased up to 110 °C and it was kept in this
condition for 30 min. The modified whiskers were filtered and washed
with ethanol to remove amines formed during the reaction and the
isocyanates that did not react. Afterward, the modified materials were
washed with ethanol and centrifuged four times at 10000 rpm and 10
°C for 15 min each step. The final step consisted in changing the solvent
of the modified nanofibers (whiskers or MFC) from ethanol to
dichloromethane, which was the solvent used for the film preparation.
Details of the grafting procedure and characterization of the grafted
nanoparticles are reported elsewhere.17 The values of the degree of
substitution have been estimated from elemental analysis for both
whiskers and MFC.17 A degree of substitution of DS ) 0.07 and DS
Siqueira et al.
) 0.09 were found for whiskers and MFC, respectively. This result
takes into account the whole hydroxyl groups, whereas the experimental
conditions allow grafting only the surface OH groups. The grafting
was sufficient for this application. This grafting is important to better
understand the changes in compatibility and dispersion of these
nanoparticles.
Nanocomposite Films Preparation. The polycaprolactone (PCL
42500 g · mol-1) was first dissolved in dichloromethane at room
temperature for 20 h (0.036 g · L-1). Different amounts of whiskers or
MFC were used to prepare the nanocomposite films, namely, 0, 3, 6,
9, and 12 wt % on the dry basis. The maximum of 12 wt % is generally
enough to level off the properties of the materials. The corresponding
amounts of nanoparticles in suspension in dichloromethane were mixed
and magnetically stirred for 6 h with the PCL solution. The suspensions
were sonicated for 2 min before being cast in Teflon molds, where the
films were obtained by solvent evaporation at room temperature.
Characterizations. Samples for transmission electron microscopy
(TEM) were observed with a Philips CM200 transmission electron
microscope using an acceleration voltage of 80 kV. A drop of diluted
suspension of sisal whiskers was deposited on a carbon-coated grid.
The samples were stained with a 2 wt % solution of uranyl acetate. A
field emission scanning electron microscope (SEM), model Quanta 200
FEI, with accelerating voltage of 12.5 kV was used to study sisal MFC
surfaces’ topography. The samples were mounted onto a substrate with
carbon tape and coated with a thin layer of gold. Optical microscopy
observations were performed using an optical microscope Olympus
BH-2 in transmission mode.
Differential scanning calorimetry (DSC) experiments were carried
out with a DSC Q100 differential calorimeter (TA Instruments) fitted
with a manual liquid nitrogen cooling system. The samples were placed
in hermetically closed DSC devices. The heating and cooling rates were
10 °C · min-1 from -100 to 100 °C and from 100 to -100 °C,
respectively, in a N2 atmosphere. Sample weights were between 6 mg
and 8.5 mg.
Tensile tests were carried out with a RSA3 (TA Instruments, U.S.A.)
equipment with a 100 N load cell. Measurements were performed with
a cross head speed of 10 mm · min-1 at 25 °C. The samples were
prepared by cutting strips of the films 20 mm long and the distance
between jaws was 10 mm, whereas the width and the thickness of the
samples were measured before each measurement. The initial strain
rate was therefore ε̇ ) 1.67 × 10-2 s-1. Five samples were used to
characterize each nanocomposite.
Dynamical mechanical analysis (DMA) of the nanocomposite films
was carried out using a RSA3 (TA Instruments, USA) equipment
working in tensile mode. The measurements were performed at a
constant frequency of 1 Hz, strain amplitude of 0.05%, in the
temperature range from -100 to 100 °C, a heating rate of 5 °C · min-1
and a distance between jaws of 10 mm. The width of the samples varied
from 3 to 5 mm, which were measured before each analysis. Two
samples were used to characterize each nanocomposite.
Results and Discussion
Morphological Analyses. TEM micrographs of sisal whiskers
reported in Figure 1 show the homogeneity and nanometric
dimensions of sisal whiskers. The length and diameter of sisal
nanocrystals were determined by using digital image analysis
(ImageJ). The geometric average length and diameter were
around 215 nm ( 67 nm and 5 nm ( 1.5 nm, respectively. A
minimum of 421 and 205 measurements were used to determine
both the length and the diameter, respectively, of sisal whiskers.
These dimensions are in agreement with the results found by
Garcia de Rodriguez et al.18 However, the length of our whiskers
is lower compared to the one reported by Garcia de Rodriguez18
(250 nm), resulting in a lower aspect ratio (L/d ) 43 compared
to 6018). It is worth noting that sisal plants were not ground at
the same place and under the same conditions (Brazil vs India).
Cellulose Whiskers versus Microfibrils
Figure 1. Transmission electron micrographs of sisal whiskers.
Figure 2. SEM image sisal MFC (A) and optical microscopy image of sisal MFC (B).
Figure 3. Visual examination of sisal whiskers (A) and microfibrillated cellulose (B) in different liquids: (i) water, (ii) acetone, and (iii)
dichloromethane.
Figure 2 shows SEM image (A) and optical microscopy image
(B) of sisal MFC. The diameter of microfibrilated cellulose from
sisal were determined by digital image analysis (ImageJ) of SEM
micrographs. The average diameter was about 52 nm ( 15 nm
showing that microfibrils bundles were obtained. A minimum
of 50 measurements were performed for its determination.
Nonetheless it was not possible to determine the avegare length
of sisal MFC by SEM microscopy analysis. Indeed, when a drop
of MFC suspension was dried on the substrate prepared for SEM
analysis, a film formed which hindered the observation of
MFC’s length.
Films Preparation. PCL-based films reinforced with various
fractions of either raw or chemically modified sisal nanoparticles
have been prepared. Figure 3A shows the dispersion state of
unmodified whiskers in different organic solvents, namely, water
Biomacromolecules, Vol. 10, No. 2, 2009 427
(i), acetone (ii), and dichloromethane (iii). The dispersion state
of unmodified MFC in various liquids can be observed in Figure
3B. We clearly observe that MFC’s suspension remains
homogeneous when substituting water for acetone (Figure 3B,
(i) and (ii), respectively). However, because the polarity of the
solvent decreased, the solvent exchange step became more
difficult. So, MFC were not homogeneously dispersed in
dichloromethane (Figure 3C, (iii)). Therefore, the preparation
of unmodified MFC filled PCL films by casting from a
dichloromethane suspension was not possible.
The difference in the dispersibility between whiskers and
MFC could be related to their only differences, that is, the
possibility of entanglement and the presence of residual pectins
at the surface of MFC.19 The existence and the characterization
of residual hemicelluloses at the surface of MFC were reported
428 Biomacromolecules, Vol. 10, No. 2, 2009 Siqueira et al.

Figure 4. Typical stress-strain curves obtained from tensile tests for PCL-based nanocomposite films reinforced with 3 wt % (A) and 12 wt %
(B) sisal whiskers: unmodified (O) and chemically modified (∆). The behavior of the neat PCL matrix (9) is added as reference. The inset is an
expanded view of the low strain region.

elsewhere.20 It was found that around 2% of hemicelluloses were

present at the surface.
Tensile Tests. The nonlinear mechanical behavior of PCL-
based nanocomposite films was characterized by tensile tests
performed at room temperature. Typical stress-strain curves
obtained from tensile tests for PCL-based nanocomposite films
are shown in Figure 4. These figures clearly show the influence
of the grafting of cellulosic whiskers on the mechanical behavior
of the nanocomposite films. Young’s modulus values were
analyzed from the initial slope of the tensile curves, as detailed
in the expanded views. Figure 4A shows that the film reinforced
with 3 wt % of modified sisal whiskers displays higher tensile
modulus, strength, and strain at break, compared to its unmodi-
fied filler counterpart. It clearly shows the positive impact of
the surface chemical modification of the whiskers on the
mechanical behavior of the nanocomposite films. Moreover,
compared to the neat matrix, the addition of 3 wt % of modified
nanoparticles allows enhancing the tensile modulus without
detrimental effect on the ultimate mechanical properties (strength
and strain at break) contrarily to unmodified nanoparticles. At
this loading level, the nanocomposite films display a ductile
and rubber-like behavior similar to the PCL matrix and typical
of a semicrystalline thermoplastic material tested above its glass
transition temperature. On the contrary, highly filled composites
(12 wt %, Figure 4B) show a completely different behavior.
No plastic deformation was observed and their extreme brittle-
ness is clearly observed. The strain at break is extremely low
whether the surface of the nanoparticle was chemically modified
or not. A similar behavior has been reported for MFC reinforced
poly(styrene-co-butyl acrylate) by Dalmas et al.21
The influence of the morphology of the cellulosic nanoparticle
(whiskers vs MFC) on the tensile mechanical behavior of the
nanocomposite film is shown in Figure 5. The addition of 3 wt
% of chemically modified cellulosic nanoparticles induces an
increase of the tensile modulus regardless their nature, as can
be seen in Figure 6A. However, the ultimate mechanical
properties differ depending on the nature of the filler. Modified
sisal whiskers allow keeping a higher level of elongation at break
for the PCL film, compared to modified MFC. However, the
value of the strength is higher for modified MFC nanocompos-
ites than for modified whiskers nanocomposites.
Figure 6 (A) shows the evolution of the Young’s modulus
determined from the initial slope of the stress-strain curves as
a function of the nanoparticle content. Nominal data were used
Figure 5. Typical stress-strain curves obtained from tensile tests
for PCL-based nanocomposite films reinforced with 3 wt % modified
sisal whiskers (O) and 3 wt % modified MFC (∆). The behavior of
the neat PCL matrix (9) is added as reference.
for the stress and strain values. Results were reported for the
neat matrix and nanocomposite films reinforced with unmodified
and chemically modified whiskers, as well as chemically
modified MFC extracted from sisal.
The addition of raw sisal whiskers in PCL results in a global
decrease of the tensile modulus of the material. It could be
ascribed to the poor interfacial adhesion between the cellulosic
nanoparticles (hydrophilic) and the PCL matrix (hydrophobic).
However, previous studies have shown that not only the filler-
matrix adhesion but also the filler-filler interactions are important
when considering the reinforcing capability of cellulose whis-
kers.5 Indeed, it was shown that above the percolation threshold,
the cellulosic nanocrystals form a stiff percolating network
through hydrogen bonding that is responsible for the reinforcing
effect when the nanocomposite films were obtained by a casting/
evaporation technique. For rod-like nanoparticles with an aspect
ratio of 43, the percolation threshold is around 1.6 vol %, that
is, around 2 wt %.5 However, in our case the liquid medium
was dichloromethane and not water. This should most probably
affect the dispersion of the nanoparticles in the liquid and then
in the solid nanocomposite film. Therefore, the decrease of the
tensile modulus is most probably ascribed to the poor dispers-
ibility of the highly hydrophilic nanoparticles within the
Cellulose Whiskers versus Microfibrils
Figure 6. Evolution of the Young’s modulus (A), strength (B), and
strain at break (C) for PCL-based nanocomposites vs filler (whisker
or MFC) content: unmodified (9) and chemically modified whiskers
(O) and MFC (∆).
hydrophobic matrix rather than to the poor adhesion at the filler-
matrix adhesion as already suggested elsewhere.22 The crystal-
lization of the polymeric matrix could also affect the formation
of the percolating cellulosic network.
When the surface of the nanoparticles is chemically modified,
the tensile modulus tends to increase. It most probably results
from the compatibilizing effect of the grafted molecules that
induces a better dispersion of the reinforcing phase within the
matrix. Compared to the neat PCL matrix, the addition of only
3 wt % of modified sisal whiskers or modified MFC increases
the tensile modulus by a factor 15 or 60%, respectively. With
Biomacromolecules, Vol. 10, No. 2, 2009 429
12 wt % nanoparticles, the modulus is increased by 45 and
137%, respectively.
A simple visual inspection of the nanocomposite films allows
estimating the level of dispersion of the nanoparticles within
the polymeric matrix. Indeed, compared to the neat matrix, the
nanocomposite films reinforced with unmodified sisal whiskers
becomes dotted with black. When the cellulose whiskers were
chemically modified the occurrence of these aggregates vanish
and the appearance of the nanocomposite film becomes similar
to the one of the unfilled film evidencing the higher dispersion
level of the filler.
Several authors21,23,24 have also reported an increase in the
stiffness of polymeric films by increasing the filler content.
However, none of them have compared whiskers and MFC.
With these innovative results we prove that MFC has a better
impact on the stiffness increase and we confirm that the
reinforcing effect depends on the morphology of the filler.16
According to Dufresne et al.,19 pectins act as a binder between
cellulose microfibrils, improving the mechanism of load transfer
to the filler. This binding mechanism is controlled by hydrogen
bonding and/or covalent connections between pectins, hemi-
cellulose, and cellulose microfibrils. Therefore, the higher
Young’s modulus observed for modified MFC-based nanocom-
posites could be explained, at least partially, by the presence
of pectins at the surface of the nanoparticles. Pectins and
hemicelluloses are not present at the surface of the whiskers
because they are eliminated during the hydrolysis treatment.
The evolution of the strength and strain at break for PCL-
based films as a function of the filler content, presented in
Figures 6B and C, confirms the differences between the two
cellulose nanoparticles. Figure 6B shows that the strength is
drastically reduced upon loading PCL with unmodified sisal
whiskers, while both modified sisal whiskers and MFC are able
to maintain the strength of PCL at a relatively higher level,
especially for low contents of modified nanofillers (e.g., 3 wt
%). Low difference is reported between the two kinds of
modified nanoparticles, although the strength is systematically
higher for MFC. However, this difference is still significant in
regard with the error bars. As expected, the strain at break
(Figure 6C) decreases upon filler addition. However, for low
filler content (3 and 6 wt %), a clear difference is observed
depending on the nature of the nanoparticle. The most brittle
material is the nanocomposite film reinforced with unmodified
sisal whiskers, followed by modified MFC and then modified
whiskers. For higher filler content, the strain at break remains
similar regardless the nature of the nanoparticle.
The surface chemical modification of the nanoparticles most
probably affects their dispersibility as already suggested. This
poor dispersibility and the lack of intimate adhesion between
the filler and the matrix lead to numerous irregularly shaped
microvoids or microflaws in the composite structure. Because
of these microflaws the stress transfer from the matrix to the
filler is poor and the mechanical properties of the nanoparticles
are not fully utilized. The brittleness of the material is
accentuated by the probable aggregation of the unmodified
nanoparticles that leads to the formation of weak points. The
lowest elongation at break observed for modified MFC com-
pared to modified whiskers is most probably ascribed to the
possibility of entanglements of the former. Filler entanglements
are not likely to occur with whiskers that occur as straight rod-
like nanoparticles and then the behavior at break of the ensuing
nanocomposites is mainly governed by the matrix.
These results fully agree with the observations reported by
Azizi Samir et al.16 in their study of unmodified sugar beet
430 Biomacromolecules, Vol. 10, No. 2, 2009
Figure 7. Evolution of the logarithm of the storage tensile modulus
(log E′) vs temperature at 1 Hz for PCL-based nanocomposites films
reinforced with modified sisal whiskers (A) or modified MFC (B): 0 wt
% (9), 3 wt % (O), 6 wt % (+), 9 wt % (×), and 12 wt % (∆).
cellulose microfibrils reinforced poly(styrene-co-butyl acrylate).
They submitted cellulose microfibrils prepared from sugar beet
to different hydrolysis conditions in order to obtain nanofillers
with various entanglements characteristics. They observed a
decrease of both the strength and tensile modulus, and an
increase of the strain at break when the hydrolysis strength
increases, that is, when decreasing the possibility of filler
entanglement.
To conclude on the nonlinear mechanical behavior of sisal
nanoparticles reinforced PCL, the surface chemical modification
has a positive impact on both the stiffness and ductility of the
films. Moreover, the use of MFC instead of whiskers allows
obtaining stiffer but more brittle nanocomposite films.
Dynamic Mechanical Behavior. Figure 7 shows the evolu-
tion of the logarithm of the storage tensile modulus as a function
of temperature for modified sisal whiskers (panel A) and MFC
(panel B) reinforced PCL films. Modulus values have been
normalized at low temperature. Between the main relaxation
around -60 °C associated with the glass transition and the
irreversible polymeric chain flow around 50 °C due to the
melting of the PCL matrix, the storage modulus is increased
upon filler addition. This effect is more significant with MFC
in agreement with tensile tests.
Also, it is observed that the irreversible flow of the polymeric
matrix upon heating occurs at lower temperature for MFC
reinforced PCL films compared to cellulose whiskers reinforced
Siqueira et al.
films. This effect is ascribed to differences in the melting
behavior of the two sets of nanocomposites, as for instance the
melting temperature and degree of crystallinity, as it will be
outlined in the next section.
Thermal Characterization. The thermal characterization of
PCL-based nanocomposite films was carried out using DSC
measurements. From the analysis of DSC traces, the glass-rubber
transition temperature (Tg), the melting temperature (Tm),
associated heat of fusion (∆Hm), degree of crystallinity (χc),
and crystallization temperature (Tc) were obtained for the
unfilled PCL film, and nanocomposite materials reinforced with
either unmodified and modified whiskers, and modified MFC.
The resulting experimental data are listed in Table 1.
For all nanocomposites, the crystallization temperature is
significantly increased (around 10-12 °C) compared to the neat
matrix. For unmodified cellulosic whiskers, no evolution upon
filler loading is observed, whereas it is found to globally slightly
increase for modified nanoparticles. Moreover, the Tc increase
is more pronounced for a given whiskers content. The filler most
probably acts as a nucleating agent for the crystallization of
PCL. Panaitescu et al.25 also reported an increase of 10 °C of
the crystallization temperature for maleic anhydride modified
polypropylene (MAPP) composites reinforced with 20 wt % of
MFC or for polypropylene (PP) filled with 30 wt % of
microcrystalline cellulose. They also observed that the increase
in the crystallization temperature was more important in the
presence of filler-matrix compatibilizer. A similar result was
reported for MAPP-compatibilized palm tree fibers reinforced
PP.26
The degree of crystallinity of PCL is almost not influenced
by the presence of chemically modified sisal MFC. It remains
around 50%. Abdelmouleh et al.27 did not observed any
significant effect of lignocellulosic fibers on the degree of
crystallinity of the low density polyethylene (LDPE) matrix,
even after fiber modification with different silane agents. Yao
et al.28 analyzed the effect of the type of fiber and fiber content
on the degree of crystallinity of high density polyethylene
(HDPE)-based composites and no significant effect was reported
even for fiber content as high as 50 wt %.
On the contrary, the presence of sisal whiskers to the PCL
matrix seems to significantly increase the degree of crystallinity
of the host matrix, regardless their modification state. It is
increased from 51 wt % for the neat matrix to values close to
60 wt % for nanocomposite films regardless the filler content,
except for the composite reinforced with 3 wt % of unmodified
whiskers. This enhancement of the crystallinity of the PCL
matrix probably results, at least partially, in the improvement
of the stiffness for these nanocomposites reported previously.
The origin of this difference between cellulose whiskers and
MFC on the crystallization behavior of PCL is under investiga-
tion and results will be published shortly. An increase of the
crystallinity was also reported for residual lignocellulosic flour
from spruce and ground olive stone reinforced poly(hydroxy-
butyrate-co-valerate) (PHBV).29 Reinsch and Kelley30 reported
that short wood fibers acted as nucleating sites for the crystal-
lization of PHBV and enhanced its crystallization rate. It was
suggested by Luo and Netravali31 that Reinsch and Kelley’s
results may have been influenced by the lignin of wood fiber.
Then, despite the surface chemical modification of MFC,
residual hemicellulose and pectin at the surface may affect the
crystallization capability of PCL.
A significant increase in crystallinity of sorbitol plasticized
starch32 was reported when increasing cellulose whiskers
content. This phenomenon was ascribed to an anchoring effect
Cellulose Whiskers versus Microfibrils Biomacromolecules, Vol. 10, No. 2, 2009 431

Table 1. Thermal Characteristics of PCL-Based Nanocomposites Obtained from DSC Analysisa

filler sample Tg (°C) Tm(°C) ∆Hm (J · g-1) χcb Tc (°C)
PCL -62.0 63.4 80.7 0.51 22.8
unmodified whiskers WU 3% -53.0 65.4 80.1 0.53 32.7
WU 6% -55.8 65.9 86.7 0.59 33.7
WU 9% -56.0 65.7 90.0 0.63 33.2
WU 12% -59.6 64.2 85.5 0.62 32.6
modified whiskers WM 3% -58.5 65.7 94.1 0.62 33.8
WM 6% -57.6 64.2 91.2 0.62 33.5
WM 9% -57.9 64.6 88.9 0.62 33.9
WM 12% -57.2 64.1 86.7 0.63 34.9
modified MFC MFCM 3% -56.9 61.9 76.0 0.50 33.4
MFCM 6% -56.2 61.8 72.2 0.49 33.9
MFCM 9% -54.4 61.9 68.4 0.48 34.2
MFCM 12% -54.8 60.8 69.6 0.50 34.8
a
Glass transition temperature (Tg), melting temperature (Tm), enthalpy of fusion (∆Hm), degree of crystallinity (χc), and temperature of crystallization
(Tc). b χc ) ∆Hm/w∆Hm°, where ∆Hm° ) 157 J/g (heat of fusion for 100% crystalline PCL) and w is the weight fraction of polymeric matrix in the composite.

of the cellulosic filler, probably acting as a nucleating agent.
For POE based composites the degree of crystallinity of the
matrix was found to be roughly constant up to 10 wt % tunicin
whiskers33,34 and to decrease for higher loading level.35 Yao et
al.28 reported an increase of the degree of crystallization of
HDPE, from 16 to 19%, when 30 and 50 wt % of wood fibers
were added, respectively. It is worth noting that in our case the
addition of only 3 wt % of modified whiskers induce an increase
of the degree of crystallinity around 10%. Jimenez et al.36
studied the crystallization of PLC/clay composites and reported
that the crystallization was governed by two terms, the diffusion
and the nucleation. In addition, they observed that only a small
amount of clay was useful to serve as nucleation agent, while
a large amount seemed to hold back the transportation of
polymer segments. Moreover, the dependence of the clay content
on the occurrence of these effects was pointed out. Transcrys-
tallization of PP at cellulose nanocrystal surfaces was recently
evidenced and it was found to result from enhanced nucleation
due to some form of epitaxy.37
It seems that the nucleating effect of cellulosic nanocrystals
is mainly governed by surface chemical considerations. A
decrease in the degree of crystallinity of PCL was reported when
adding Riftia tubes chitin whiskers.38 It was suggested that,
during crystallization, the rod-like nanoparticles are most
probably first ejected and then occluded in intercrystalline
domains, hindering the crystallization of the polymer.
Concerning the melting point of the PCL matrix, it slightly
increases, when adding sisal whiskers, from about 63 °C to
64-66 °C (Table 1). This temperature is directly related to the
size of the crystalline domains. It means that the presence of
the cellulose whiskers does not interfere significantly with the
crystal growth, regardless their surface modification state. On
the contrary, the melting point tends to decrease upon modified
MFC addition. It means that the presence of MFC induces steric
hindrance effects restricting the growth of crystalline PCL
regions, probably resulting from entanglements between mi-
crofibrils. It results in both a lower melting point and lower
degree of crystallinity.
In the literature, melting temperature values were reported
to be nearly independent of the filler content in plasticized
starch32,39 and in POE-based materials33,34 filled with tunicin
whiskers. The same observation was reported for PCL reinforced
with Riftia tubes chitin whiskers38 and CAB reinforced with
native bacterial cellulose whiskers.40 However, for the latter
system, Tm values were found to increase when the amount of
trimethylsilylated whiskers increased. The authors ascribed this
difference to the stronger filler-matrix interaction in the case of
chemically modified whiskers. It might be also the case with
our results.
The glass transition temperature (Tg) of nanocomposites
reinforced with either unmodified or modified cellulosic nano-
particles is systematically higher than the one of the neat matrix
(Table 1). This effect is more significant with unmodified sisal
whiskers. It means that the molecular mobility of amorphous
PCL chains is restricted by the presence of the filler. The filler/
matrix compatibilization by surface grafting limits this phe-
nomenon. However, it is worth noting that the value of Tg can
also be affected by the degree of crystallinity of the matrix.
In most studies, no modifications of Tg values were reported
when increasing the amount of whiskers, regardless the nature
of the polymeric matrix. This result appeared to be surprising
because of the high specific area of these nanoparticles, that is,
around 170 m2 · g-1, for instance, for tunicin whiskers.39
In glycerol plasticized starch based composites, peculiar
effects of tunicin whiskers on the Tg of the starch-rich fraction
were reported depending on moisture conditions.39 For low
loading level (up to 3.2 wt %), a classical plasticization effect
of water was reported. However, an antiplasticization phenom-
enon was observed for higher whiskers content (6.2 wt % and
up). These observations were discussed according to the possible
interactions between hydroxyl groups on the cellulosic surface
and starch, the selective partitioning of glycerol and water in
the bulk starch matrix or at whiskers surface, and the restriction
of amorphous starch chain mobility in the vicinity of the starch
crystallite coated filler surface. For glycerol plasticized starch
reinforced with cellulose crystallites prepared from cottonseed
linter,42 an increase of Tg with filler content was reported and
attributed to cellulose/starch interactions. For tunicin whiskers/
sorbitol plasticized starch,32 Tgs were found to increase slightly
up to about 15 wt % whiskers and to decrease for higher
whiskers loading. Crystallization of amylopectin chains upon
whiskers addition and migration of sorbitol molecules to the
amorphous domains were proposed to explain the observed
modifications.
For waxy maize starch nanocrystals reinforced natural rubber,
a decrease of the onset glass transition temperature with the
increase of the nanoparticles content was reported.43 When a
glycerol plasticized starch matrix was used, an increased
temperature was reported for the main relaxation process
associated with the glass-rubber transition of amylopectin-rich
domains when increasing the starch nanocrystals content.44 The
reduction in the molecular mobility of matrix amylopectin chains
432 Biomacromolecules, Vol. 10, No. 2, 2009
for filled materials was explained by the establishment of
hydrogen bonding forces between both components. A similar
observation was reported for polyvinyl acetate (PVA) reinforced
with sisal cellulose whiskers18 and carboxymethyl cellulose
(CMC) reinforced with cotton cellulose whiskers.45
Conclusions
Cellulosic nanoparticles were prepared from sisal and used
to reinforce polycaprolactone (PCL). Both rod-like whiskers and
long flexible “hairy” microfibrils or microfibrillated cellulose
(MFC) were prepared. The surface of the nanoparticles was also
chemically modified by grafting N-octadecyl isocyanate
(C18H37NCO). The chemical grafting was found to improve the
dispersion of nanofillers in organic solvents. The effects of the
filler content, chemical grafting, and nature of the reinforcing
phase were investigated. Sisal whiskers were found to induce a
limited reinforcing effect because of the aggregation of the filler
within the nanocomposites film prepared from dichloromethane.
However, the presence of the filler increases the glass transition,
crystallization, and melting temperatures as well as the degree
of crystallinity of the PCL matrix, the cellulosic whiskers
probably acting as nucleation sites. It was not possible to process
PCL nanocomposite films reinforced with unmodified MFC
because of the poor dispersibility of the latter in dichlo-
romethane. The chemical grafting of cellulose whiskers improve
their compatibility with the polymeric matrix and dispersity.
The mechanical properties of ensuing nanocomposite films were
really improved in terms of both stiffness and ductility.
Characteristic temperatures and degree of crystallinity were also
found to increase with the filler content. When comparing sisal
whiskers and MFC (both chemically modified), it was found
that the modulus was higher for MFC-reinforced composites,
whereas the elongation at break was lower for a given loading
level. The difference was most probably ascribed to the
possibility of entanglements of MFC contrarily to rod-like
nanoparticles. The introduction of MFC within the PCL matrix
did not induce an increase of the degree of crystallinity of the
matrix and its melting point tends to decrease contrarily to what
was observed for sisal whiskers-reinforced films. Again, this
difference was ascribed to the possibility of entanglement of
MFC that tends to confine the polymeric matrix and restrict its
crystallization. Although the crystallinity of MFC-based nano-
composite films was lower compared to whiskers-based com-
posites, the stiffness of the former was higher, showing the high
impact of the entanglements on the mechanical properties.
Acknowledgment. The authors gratefully acknowledge AL-
BAN Program for the financial support (Ph.D. fellowship of
G.S.), and Y. Habibi and FCBA for the support in MFC
production.
References and Notes
(1) Pandey, J. K.; Kumar, A. P.; Misra, M.; Mohanty, A. K.; Drzal, L. T.;
Singh, R. P. J. Nanosci. Nanotechnol. 2005, 5, 497–526.
(2) Espert, A.; Vilaplana, F.; Karlsson, S. Composites, Part A 2004, 35,
1267–1276.
(3) Dufresne, A. J. Nanosci. Nanotechnol. 2006, 6, 322–330.
(4) Bondeson, D.; Mathew, A.; Oksman, K. Cellulose 2006, 13, 171–
180.
Siqueira et al.
(5) Azizi Samir, M. A. S.; Alloin, F.; Dufresne, A. Biomacromolecules
2005, 6, 612–626.
(6) Tserki, V.; Zafeiropoulos, N. E.; Simon, F.; Panayiotou, C. Composites,
Part A 2005, 36, 1110–1118.
(7) Pasquini, D.; Belgacem, M. N.; Gandini, A.; Curvelo, A. A. D. J.
Colloid Interface Sci. 2006, 295, 79–83.
(8) Bledzki, A. K.; Gassan, J. Prog. Polym. Sci. 1999, 24, 221–274.
(9) Gandini, A.; Belgacem, M. N. The State of the Art. Monomers,
polymers and composites from renewable resources, 1st ed.; Elsevier:
Great Britain, 2008; pp 1-16.
(10) Gandini, A.; Belgacem, M. N. Chemical Modification of Wood.
Monomers, polymers and composites from renewable resources, 1st
ed.; Elsevier: Great Britain, 2008; pp 419-432.
(11) Wambua, P.; Ivens, J.; Verpoest, I. Compos. Sci. Technol. 2003, 63,
1259–1264.
(12) John, M. J.; Thomas, S. Carbohydr. Polym. 2008, 71, 343–364.
(13) Sreekumar, P. A.; Joseph, K.; Unnikrishnan, G.; Thomas, S. Compos.
Sci. Technol. 2007, 67, 453–461.
(14) Li, Y.; Mai, Y. W.; Ye, L. Compos. Sci. Technol. 2000, 60, 2037–
2055.
(15) Chanzy, H.; Rotzinger, B.; Smith, P. Patent WO8703288, 1987.
(16) Azizi Samir, M. A. S.; Alloin, F.; Paillet, M.; Dufresne, A. Macro-
molecules 2004, 37, 4313–4316.
(17) Siqueira, G.; Bras, J.; Dufresne, A. 235th ACS National Meeting, New
Orleans, LA, U.S.A., April 6-10, 2008 Langmuir, to be submitted.
(18) Garcia de Rodriguez, N. L.; Thielemans, W.; Dufresne, A. Cellulose
2006, 13, 261–270.
(19) Dufresne, A.; Cavaille, J. Y.; Vignon, M. R. J. Appl. Polym. Sci. 1997,
64, 1185–1194.
(20) Dinand, E.; Vignon, M. R. Carbohydr. Res. 2001, 330, 285–288.
(21) Dalmas, F.; Chazeau, L.; Gauthier, C.; Cavaille, J. Y.; Dendievel, R.
Polymer 2006, 47, 2802–2812.
(22) Habibi, Y.; Dufresne, A. Biomacromolecules 2008, 9, 1974–1980.
(23) Kristo, E.; Biliaderis, C. G. Carbohydr. Polym. 2007, 68, 146–158.
(24) Georgopoulos, S. T.; Tarantili, P. A.; Avgerinos, E.; Andreopoulos,
A. G.; Koukios, E. G. Polym. Degrad. Stab. 2005, 90, 303–312.
(25) Panaitescu, D. M.; Donescu, D.; Bercu, C.; Vuluga, D. M.; Iorga,
M.; Ghiurea, M. Polym. Eng. Sci. 2007, 47, 1228–1234.
(26) Bendahou, A.; Kaddami, H.; Sautereau, H.; Raihane, M.; Erchiqui,
F.; Dufresne, A. Macromol. Mater. Eng. 2008, 293, 140–148.
(27) Abdelmouleh, M.; Boufi, S.; Belgacem, M. N.; Dufresne, A. Compos.
Sci. Technol. 2007, 67, 1627–1639.
(28) Yao, F.; Wu, Q. L.; Lei, Y.; Xu, Y. J. Ind. Crops Prod. 2008, 28,
63–72.
(29) Dufresne, A.; Dupeyre, D.; Paillet, M. J. Appl. Polym. Sci. 2003, 87,
1302–1315.
(30) Reinsch, V. E.; Kelley, S. S. J. Appl. Polym. Sci. 1997, 64, 1785–
1796.
(31) Luo, S.; Netravali, A. N. Polym. Compos. 1999, 20, 367–378.
(32) Mathew, A. P.; Dufresne, A. Biomacromolecules 2002, 3, 609–617.
(33) Marcovich, N. E.; Auad, M. L.; Bellesi, N. E.; Nutt, S. R.; Aranguren,
M. I. J. Mater. Res. 2006, 21, 870–881.
(34) Azizi Samir, M. A. S.; Alloin, F.; Gorecki, W.; Sanchez, J. Y.;
Dufresne, A. J. Phys. Chem. B 2004, 108, 10845–10852.
(35) Azizi Samir, M. A. S.; Alloin, F.; Sanchez, J. Y.; Dufresne, A. Polymer
2004, 45, 4149–4157.
(36) Jimenez, G.; Ogata, N.; Kawai, H.; Ogihara, T. J. Appl. Polym. Sci.
1997, 64, 2211–2220.
(37) Gray, D. G. Cellulose 2008, 15, 297–301.
(38) Morin, A.; Dufresne, A. Macromolecules 2002, 35, 2190–2199.
(39) Angles, M. N.; Dufresne, A. Macromolecules 2000, 33, 8344–8353.
(40) Grunert, M.; Winter, W. T. J. Polym. EnViron. 2002, 10, 27–30.
(41) Dufresne, A. Compos. Interfaces 2000, 7, 53–67.
(42) Lu, Y. S.; Weng, L. H.; Cao, X. D. Macromol. Biosci. 2005, 5, 1101–
1107.
(43) Angellier, N.; Molina-Boisseau, S.; Lebrun, L.; Dufresne, A. Macro-
molecules 2005, 38, 3783–3792.
(44) Angellier, H.; Molina-Boisseau, S.; Dole, P.; Dufresne, A. Biomac-
romolecules 2006, 7, 531–539.
(45) Choi, Y. J.; Simonsen, J. J. Nanosci. Nanotechnol. 2006, 6, 633–639.
BM801193D