You are on page 1of 8

Current Opinion in Colloid & Interface Science 29 (2017) 1–8

Contents lists available at ScienceDirect

Current Opinion in Colloid & Interface Science

journal homepage: www.elsevier.com/locate/cocis

Cellulose nanomaterial reinforced polymer nanocomposites


Alain Dufresne
Univ. Grenoble Alpes, CNRS, Grenoble INP, LGP2, F-38000, Grenoble, France

a r t i c l e i n f o a b s t r a c t

Article history: Several forms of cellulose nanomaterials, notably cellulose nanocrystals and cellulose nanofibrils, exhibit attrac-
Received 30 September 2016 tive properties and are potentially useful for a large number of industrial applications. These include the paper
Received in revised form 24 January 2017 and cardboard industry, use as reinforcing filler in polymer nanocomposites, basis for low-density foams, additive
Accepted 31 January 2017
in adhesives and paints, as well as a wide variety of filtration, electronic, food, hygiene, cosmetic, and medical
Available online 8 February 2017
products. This entry focuses on cellulose materials as filler in polymer nanocomposites. The ensuing mechanical
Keywords:
properties obviously depend on the type of nanomaterial used, but the crucial point is the processing technique.
Cellulose nanomaterial The emphasis is on the melt processing of such nanocomposite materials that has not yet been properly resolved
Cellulose nanocrystal and remains a challenge.
Cellulose nanofibrils © 2017 Elsevier Ltd. All rights reserved.
Nanocomposites
Processing
Mechanical properties

1. Introduction the last 20 years [5,13,15,27]. Indeed, nanocomposites show unique


properties, because of the nanometric size effect, compared to conven-
There is an explosion of interest in the use of biomass as a source tional composite even at low filler content. Nanofillers have strong rein-
of renewable energy and materials. A promising source of biomass is forcing effects and studies have also shown their positive impact
cellulose, the most abundant polymer on Earth, which has been used in barrier packaging. However, the processing of the material plays a
for centuries in highly diverse applications. A recent focus of this activity crucial role since it determines its morphology (spatial arrangement of
has followed from the recognition that, by suitable chemical and me- the phases).
chanical treatments, it is possible to produce fibrous materials with
one or two dimensions in the nanometer range from any naturally 2. Cellulose nanomaterials
occurring sources of cellulose. The terms “nanocellulose” or “cellulose
nanomaterial” are now used to cover the range of materials derived The possibility for preparing nanomaterials or nanoparticles from
from cellulose with at least one dimension in the nanometer range. cellulose or other polysaccharides such as chitin or starch results
This material has been described as a new ageless bionanomaterial from a common feature, viz. 1) they are semicrystalline polymers, and
[14], as it exists in nature since the dawn of time but was only extracted 2) they display a hierarchical multiscale nanostructure. After purifica-
rather recently. Isolation of crystalline cellulosic regions, in the form tion, nanoparticles can be extracted from these naturally occuring poly-
of monocrystals, by an acid hydrolysis process was first reported in mers mainly using a top-down mechanically- or chemically-induced
1947 [30]. The first report on the mechanical destructuration of cellu- deconstructing strategy. Even if considered as cellulose nanomaterial,
lose fibers was published latter in 1983 in two companion papers bacterial cellulose results from a different strategy since it involves a
[20,33]. Nanocellulose-based materials have a low carbon footprint, bottom-up procedure, and will consequently not be addressed in this
are sustainable, renewable, recyclable, and nontoxic; they thus have entry.
the potential to be truly green nanomaterials with many useful and un-
expected properties. How not to love them? 2.1. Cellulose nanofibrils
With a Young's modulus in the range 100–130 GPa and a surface
area of several hundred m2.g− 1, new promising properties can be The mechanically-induced destructuration strategy consist in ap-
considered for cellulose. The potential of this nanomaterial has been plying severe multiple mechanical shearing actions to a cellulosic fiber
proved for special functional nanomaterials [21••] but it is as a biobased slurry to release more or less individually the constitutive microfibrils.
reinforcing nanofiller that it has attracted significant interest during Different shearing equipments such as homogenizer, microfluidizer or
ultra-fine friction grinder are generally used. This material is usually
called microfibrillated cellulose (MFC), nanofibrillated cellulose (NFC)
E-mail address: alain.dufresne@pagora.grenoble-inp.fr. or cellulose nanofibrils (CNF) and is obtained as an aqueous suspension.

http://dx.doi.org/10.1016/j.cocis.2017.01.004
1359-0294/© 2017 Elsevier Ltd. All rights reserved.
2 A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8

Recently, TAPPI (Technical Association of the Pulp and Paper Industry) atom skeleton, but differ in their hydrogen bonding patterns. The Iα
has proposed to standardize the terminology (Standard Terms and form represents a triclinic phase with one-chain-per-unit cell, while
Their Definition for Cellulose Nanomaterial WI 3021) and below a the Iβ form represents a monoclinic phase with two-chains-per-unit
given degree of destructuration, the term CNF should be used. cell. The ratio of the two allomorphs Iα and Iβ differs greatly depending
This production route is normally associated to high energy con- on the species. The Iα phase is mainly found in celluloses produced by
sumptions for fiber delamination and different pretreatments have primitive organisms such as algae or bacteria while cellulose Iβ lies
been therefore proposed to facilitate this process, e.g. mechanical cutting, mainly in the cellulose produced by higher plants (cotton, wood,...)
acid hydrolysis, enzymatic pretreatment, and introduction of charged and animals such as in the envelope of marine animals.
groups through carboxymethylation or 2,2,6,6-tetramethylpiperidine- The elastic modulus EL of the crystalline regions of cellulose I in the
1-oxyl (TEMPO)-mediated oxidation. CNF appears as flexible filaments direction parallel to the chain axis was measured both experimentally
consisting of both individual and aggregated nanofibrils made of alter- and through atomistic molecular dynamics simulations. Some values
nating crystalline and amorphous cellulose domains (Fig. 1a). The are reported in Fig. 2 and Table 1. These values were determined at
width is generally in the range 3–100 nm depending on the source of room temperature and considering intramolecular hydrogen bonding.
cellulose, defibrillation process and pretreatment and the length is Much lower values were reported when intramolecular hydrogen
considered to be higher than 1 μm. It is worth noting that hydrophobic bonding was not taken into account, evidencing their important role
compounds are still present at the surface of CNF and that the surface in the determination of the crystallite modulus and chain deformation
charge is fixed by the pretreatment step. mechanism. Nevertheless, a broad range of values is observed and the
average experimental and predicted tensile modulus of cellulose I crys-
2.2. Cellulose nanocrystals tal depicted by horizontal lines in Fig. 2 is 124 GPa and 134 GPa, respec-
tively. It is therefore reasonable to consider that the tensile modulus
The chemically-induced destructuration strategy consists generally of crystalline cellulose I is around 130 GPa. The modulus of cellulose
in applying a controlled strong acid hydrolysis treatment to cellulosic fi- microfibrils is expected to result from a mixing rule of the modulus of
bers allowing dissolution of amorphous domains and therefore longitu- the crystal, the amorphous regions of cellulose and defects/air in the
dinal cutting of the microfibrils. The ensuing nanoparticles are generally sample. Its average value is around 100 GPa [13••].
called cellulose nanocrystals (CNCs) according to TAPPI (Standard Cellulose nanomaterials should therefore act as excellent reinforcing
Terms and Their Definition for Cellulose Nanomaterial WI 3021) and nanomaterial in a polymeric matrix which modulus is around few GPa
are obtained as an aqueous suspension. These nanoparticles occur as in the best conditions, i.e. in its glassy state. However, homogeneously
high aspect ratio rod-like nanocrystals, or whiskers (Fig. 1b). Their geo- blending this nanomaterial with a polymeric matrix is challenging.
metrical dimensions depend on the origin of the cellulose substrate and
hydrolysis conditions. Sulfuric acid is classically used for the preparation 4. Processing of polymer nanocomposites
of CNC and this process induces the formation of negatively charged sul-
fate groups at it surface. The average length is generally of the order of Both CNC and CNF are generally obtained as diluted aqueous (or
a few hundred nanometers and the width is of the order of a few nano- at least polar liquid) dispersions. The stability of the colloidal suspen-
meters. An important parameter for CNCs is the aspect ratio, which is sion results from the presence of residual hemicelluloses for CNF and
defined as the ratio of the length to the width. charged sulfate groups in the case of sulfuric acid-hydrolyzed CNC.
Several applications have been envisaged for these nanoparticles. The pioneering works on the preparation of nanocellulose reinforced
However, owing to the structural function of cellulose in nature, the polymer nanocomposites consequently involved an aqueous medium
application of cellulose nanomaterials as a reinforcing phase for nano- to keep this natural dispersion of the nanoparticles. This mode of pro-
composite applications is the most prominent and broadly investigated cessing allows preserving the individualization state of the nanoparti-
end use. cles resulting from their colloidal dispersion in water. Water-soluble
polymers or polymer aqueous dispersions (latex) are therefore and
3. Stiffness of the cellulose nanomaterial continue to be favorable systems for such application.
Processing of nanocomposite materials from cellulosic nanoparticles
Cellulose is a ubiquitous structural semicrystalline polymer and and polymer latex was historically the first one reported in the literature
cellulose crystals give trees and plants their strength and light weight. [16••]. The nanocomposite system reported in this study consisted of
In nature, cellulose is found in the crystalline form of cellulose I (native CNC extracted from tunicate and the polymeric matrix was obtained
cellulose) which is a mixture of two polymorphs, cellulose Iα and Iβ. by the statistical copolymerization of styrene and butyl acrylate
These two crystalline forms have the same conformation of the heavy (poly(S-co-BuA)). Solid nanocomposite films were obtained by mixing

Fig. 1. Transmission electron micrographs from a dilute suspension of (a) cellulose nanofibrils prepared from Opuntia ficus-indica fibers [23], and (b) cellulose nanocrystals extracted from
ramie fibers [18].
A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8 3

nanoparticles. This system can be considered as a model system since


the polymeric matrix was fully amorphous and the ratio S/BuA was cho-
sen so that its glass transition temperature was around 0 °C, enabling the
processing at room temperature. It has inspired a large number of subse-
quent works.
A similar strategy can be considered with water-soluble polymers
except that a polymer solution is used. Such a strategy was used in the
first report dealing with CNF reinforced polymer nanocomposites. It
involved a glycerol-plasticized starch matrix and CNF extracted from
sugar beet pulp [12••].
This wet strategy avoiding drying of the nanoparticles and ensuing
irreversible aggregation may also be envisaged with a broad range of or-
ganic liquid media. Solvent exchange can be conducted to disperse the
nanoparticles in the suitable solvent. Even if the suspension is less stable
than in water, it can be sufficient for processing. The polymeric matrix
can subsequently be dissolved in this solvent and homogeneously
mixed with the nanomaterial before solvent removal. Coating with a
surfactant or surface chemical modification of the nanoparticles can
Fig. 2. Tensile modulus values for crystalline native cellulose (cellulose I) at room also be considered to promote their dispersion. The global objective is
temperature: experimental (●) and calculated data (○). The solid and dashed horizontal to reduce their surface energy in order to improve their dispersibility/
lines correspond to the average value for experimental and calculated data, respectively.
compatibility with non-polar media. However, as we will see later, it
is generally preferable to keep available the surface-OH groups even if
and casting the two aqueous suspensions followed by water evaporation their modification promotes the dispersion of the nanoparticles.
performed above the glass transition temperature of the polymer. The surface chemical modification of cellulose nanomaterials to pro-
During water evaporation, the solid content in the medium continuous- mote their compatibility with media of low polarity has been widely in-
ly increases and the latex particles get closer and closer. The polymeric vestigated. Covalent functionalization is obviously the favorite strategy
particles act as impenetrable domains to nanocrystals during evapora- that allows chemists to freely express their expertise and fantasy be-
tion due to their high viscosity. When getting in touch each other cause of the reactivity of cellulose. It generally involves reactive hydroxyl
these soft polymeric particles deform and adopt a polyhedral form, groups from the surface. The common surface chemical modifications of
trapping the cellulosic nanoparticles in their dispersion state. Boundary CNC are shown in Fig. 3. They can be categorized into three distinct
between the former latex particles disappears by chain diffusion leading groups: (1) substitution of surface hydroxyl groups with small
to a continuous polymer film containing the dispersed cellulosic molecules (labeled with red arrows in Fig. 3); (2) polymer surface

Table 1
Longitudinal (EL) modulus of crystalline cellulose. The first column corresponds to data points reported on the x-axis in Fig. 1.

Experimental determination

Data Material Method EL (GPa) Reference

1 Ramie fibers (cellulose I) X-ray diffraction 134 Sakurada et al. (1962). J. Polym. Sci. 57, 651
2 122–135 Matsuo et al. (1990). Macromolecules 23, 3266
3 138 Nishino et al. (1995). J. Polym. Sci. Part B: Polym. Phys. 33, 1647
4 Cellulose Iβ Raman 143 Šturcova et al. (2005). Biomacromolecules 6, 1055
5 Cellulose I 57–105 Rusli and Eichhorn (2008). Appl. Phys. Lett. 93, 033111
6 Ramie fibers (cellulose I) Inelastic X-ray scattering 220 Diddens et al. (2008). Macromolecules 41, 9755
7 Cellulose Iβ AFM 145–150 Iwamoto et al. (2009). Biomacromolecules 10, 2571–2576

Theoretical determination

Data Material EL (GPa) Reference

1 Cellulose I 77–121 Meyer and Lotmar (1936). Helv. Chim. Acta 19, 68
2 56 Treloar (1960). Polymer 1, 290
3 172.9 Tashiro and Kobayashi (1985). Polym. Bull. 14, 213
4 76 Jaswon et al. (1968). Proc. Roy. Soc. 306, 389
5 Cellobiose (two hydrogen 136 Kroon-batenburg et al. (1986). Polym. Commun. 27, 290
bonds - cellulose I)
6 Cellulose I 167.5 Tashiro and Kobayashi (1991). Polymer 32, 1516
7 Cellulose I 134–135 Reiling and Brickmann (1995). Macromol. Theory Simul. 4, 725
8 Cellulose Iα 127.8 Neyertz et al. (2000). J. Appl. Polym. Sci. 78, 1939
9 Cellulose Iβ 115.2
10 Cellulose Iβ 124–155 Tanaka and Iwata (2006). Cellulose 13, 509
11 Cellulose Iα 136–155 Eichhorn and Davies (2006). Cellulose 13, 291
12 Cellulose Iβ 116–149
13 Cellulose Iβ 156 Bergenstråhle et al. (2007). J. Phys. Chem. B 111, 9138
14 Disaccharide cellulose Iβ 85.2 Cintrón et al. (2011). Cellulose 18, 505
Disaccharide cellulose Iβ 99.7
Extended cellulose Iβ chains (10–40 glucoses) 126.0
15 Cellulose Iβ 139.5 Wu et al. (2013). Cellulose 20, 43
16 Cellulose Iβ 206 Dri et al. (2013). Cellulose 20, 2703
17 Cellulose Iβ 107.8 Wu et al. (2014). Cellulose 21, 2233
18 Cellulose Iβ 196 Dri et al. (2014). Modelling Simul. Mater. Sci. Eng. 22, 085012
19 Cellulose Iβ 138 Djahedi et al. (2015). Carbohydr. Polym. 130, 175
4 A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8

Fig. 3. Common chemical modifications of cellulose nanocrystal (CNC) [21••]. [PEG: poly(ethylene glycol); PEO: poly(ethylene oxide); PLA: poly(lactic acid); PAA: poly(acrylic acid);
PNiPAAm: poly(N-isopropylacrylamide); PDMAEMA: poly(N,N-dimethylaminoethyl methacrylate)].

modification based on the “grafting onto” strategy with different cou- and grafted by an esterification reaction. The homogeneity and thermal
pling agents (labeled with blue arrows in Fig. 3); and (3) polymer sur- resistance of the ensuing nanocomposite was found to increase with the
face modification based on the “grafting from” strategy with the radical length of the grafted chains (Fig. 4a). From this Figure, it can be seen that
polymerization of ring opening polymerization (ROP), atom transfer neat LDPE is obviously translucent as any low thickness polymeric film
radical polymerization (ATRP) and single-electron transfer living radi- with a relatively low degree of crystallinity. When adding 10 wt.%
cal polymerization (SET-LP) (labeled with yellow arrows in Fig. 3). of CNC, the film becomes dotted with black. These heterogeneities
Experimental conditions should avoid swelling media and peeling reveal the poor and inhomogeneous dispersion of the filler within
effect of surface-grafted chains inducing their dissolution in the reac- the polymeric matrix as well as thermal degradation of the cellulosic
tion medium. The chemical grafting process has therefore to be mild nanomaterial. When using chemically modified CNC, the occurrence of
in order to preserve the integrity of the nanoparticle. However, cova- these aggregates progressively vanishes and the appearance of the com-
lent functionalization of the nanofiller normally involves complicate posite film becomes similar to the one of the unfilled film. However,
and expensive steps which can be prohibitive for most industrial uses. the question arises again regarding the process that is generally lengthy
The most convenient, as well as industrially and economically viable and cumbersome.
processing method for cellulose nanomaterial reinforced thermoplastic A possible way for overcoming self-aggregation of the cellulose
polymer nanocomposites is most probably melt processing such as nanomaterial during melt extrusion consists in using never-dried nano-
extrusion or injection-molding. This is well supported by the fact that particles. Liquid feeding of CNF dispersed in a mixture of water, acetone
the commercial exploitation of cellulose nanomaterials has already and glycerol triacetate and subsequent extrusion with polylactic acid
commenced. Moreover, it can be considered as a green process since (PLA) was reported [19]. However, effective venting and vacuum ports
no solvent is involved. However, the inherent incompatibility between are needed. Moreover, only low filler content composites (1 wt.%) can
hydrophilic cellulose and generally hydrophobic polymer matrix, be prepared because of the high viscosity of the CNF dispersion.
as well as thermal stability issues need to be addressed. This last Therefore, many researchers have quite recently directed their
point is particularly sensitive for sulfuric acid-hydrolyzed CNC because research to a simplest physical functionalization of the cellulose
the acid hydrolysis process introduces less thermally stable sulfate nanomaterial and preferably favoring water as dispersing medium.
groups on the surface. Nevertheless, these issues can be quite easily The guideline consists of adding to the nanoparticle aqueous dispersion
tackled, e.g. by chemically modifying the surface of the nanoparticles a water-soluble molecule that can homogeneously diffuse and disperse
as detailed above. Generally, good results can be obtained as shown in in the liquid medium. At least part of this molecule needs to present
Fig. 4a. In this study, organic acid chlorides-grafted CNCs extracted favorable interactions with the cellulosic surface to coat it. Upon drying
from ramie fibers were extruded with low density polyethylene the irreversible aggregation of the nanoparticles through hydrogen
(LDPE) [11]. Aliphatic chains presenting different lengths were used bonding is avoided because of this coating. The homogeneous
A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8 5

(a)

(b)

LDPE LDPE+CNC LDPE+M-CNC

(c)

3% 6% 3% 6%

LDPE

1%M-CNC 3%M-CNC 6%M-CNC 10%M-CNC


(d)

5%CNC

Fig. 4. Photographs of extruded CNC reinforced nanocomposite films: (a) neat LDPE film and extruded nanocomposite films reinforced with 10 wt.% of unmodified and C18 acid chloride-
grafted CNC [11]; (b) PP reinforced with various contents of quaternary ammonium salt-coated CNC (M-CNC) or 3 wt.% neat CNC, and neat PP (from left to right) [28]; (c) neat LDPE,
and LDPE reinforced with 3 or 6 wt.% of neat CNC or PEO-coated CNC (M-CNC) (from left to right) [31•]; (d) neat LDPE, and LDPE reinforced 5 wt.% neat CNC or various contents of
(PEO–PPO–PEO)-coated CNC (M-CNC) [29].

dispersion of this dried material within a polymeric matrix during melt would probably avoid this phenomenon. This strategy was explored
processing can therefore be simplified especially if the coating material using polyoxyethylene (PEO) which is a water-soluble and polar poly-
can present favorable interactions with the dispersing medium. mer [6]. The favorable interactions between PEO chains and cellulosic
Such kind of molecule can be a surfactant. Indeed, surfactants were evidenced from rheological measurements. A progressive de-
are usually organic compounds that are amphiphilic, i.e. containing crease in the viscosity was observed when adding CNC to an aqueous
both hydrophobic groups (the tail) and hydrophilic groups (the head). PEO solution up to a critical nanocrystal content after which the vis-
Since sulfuric acid-hydrolyzed CNC are negatively charged, the strength cosity started to increase. This phenomenon evidenced strong interac-
of the interactions between the cellulosic surface and the surfactant tions between PEO chains and nanoparticle surface. The effect observed
molecule can be enhanced using cationic surfactants such as quaternary can also be understood as the effect of depletion forces. The critical CNC
ammonium salts. Such a strategy was investigated and the negatively content corresponding to the minimum viscosity value decreased as
charged sulfate groups borne by CNC were profitably used to ionically the molecular weight of PEO decreased because of stronger affinity of
adsorb quaternary ammonium salt by a simple aqueous method [28]. hydroxyl end groups of PEO with cellulose than its ether oxygen groups
Improved thermal stability of surfactant-coated CNC was observed com- [3,31]. PEO-coated CNC obtained after freeze-drying the aqueous mixture
pared to neat CNC. After removal of unadsorbed quaternary ammonium was used to prepare nanocomposites with LDPE [6,31], polystyrene (PS)
salt molecules by dialysis, the mixture was freeze-dried and the modi- [22] or PLA [2] matrix by melt extrusion. Improved dispersibility and re-
fied CNC was extruded with polypropylene at 190 °C. Hydrophobized duced thermal degradation of the nanoparticles were observed showing
CNC dispersed well in different nonpolar solvents and polypropylene, the compatibilizing action of PEO (Fig. 4c). It was also shown that phys-
which was impossible for neat CNC as shown in Fig. 4b. The neat PP ical interaction of the polymer was enhanced when chemically grafting
film is translucent as any low thickness semicrystalline polymeric film. the same polymer on the surface of the nanoparticles [22].
When adding neat CNC, the film became homogeneously darker as a re- Other polymers can be used as cellulose nanomaterial carrier, but
sult of the thermal degradation of the cellulosic filler. The appearance improved compatibilization could be obtained with block copolymers
of the nanocomposite films reinforced with up to 10 wt.% quaternary consisting of hydrophilic and hydrophobic sequences. A triblock
ammonium salt-coated CNC was similar to that of neat PP showing (PEO–PPO–PEO) copolymer having two hydrophilic ends with high af-
that the coating was able to both protect the surface sulfate groups and finity with cellulosic surface attached to hydrophobic polypropylene
improve the dispersion of the nanoparticles. oxide (PPO) displaying higher affinity with hydrophobic polymers was
Melt processing techniques such as extrusion or injection-molding used to coat CNC [29]. The aspect of the nanocomposite films obtained
involve high shear. Under these conditions small adsorbed molecules from these coated CNC and LDPE is shown in Fig. 4d. Again, limited ther-
(surfactants) could be removed from the nanoparticle surface and mal degradation and improved dispersibility were observed for coated
blended with the matrix. Wrapping the cellulosic nanomaterial with CNC.
longer molecules, i.e. macromolecules, bearing moieties susceptible to Another option consists in coating the cellulosic nanomaterial using
interact physically with the cellulosic surface and with an apolar matrix the same polymer as for the matrix using a dissolution/precipitation
6 A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8

process. The ensuing concentrated masterbatch can then be diluted by


extrusion. This strategy was explored for processing CNC reinforced
polyamide 6 (PA6) [10] or polycarbonate (PC) [24] nanocomposites.
The first step consists in dispersing the cellulose nanomaterial in a sol-
vent of the polymer (formic acid for PA6 and pyridine for PC) and dis-
solving the polymer in this medium. The mixture is then precipitated
in water and coated CNC are recovered. Improved dispersion was ob-
tained from this masterbatch process compared to composite materials
obtained by direct extrusion with neat CNC.

5. Mechanical properties of polymer nanocomposites

As we have seen previously, the cellulose crystal, and to less extent


the cellulose microfibril, displays high stiffness and raised attention as
a potential engineering nanomaterial for nanocomposite applications.
Therefore, if homogeneously dispersed within a polymeric matrix, im-
proved mechanical performance is expected and should be observed. Fig. 5. Evolution of the tensile modulus for NR films reinforced with CNC and CNF
Moreover, as lightweight material (the density of crystalline cellulose extracted from the rachis of date palm tree as a function of the filler content (data
is around 1.5–1.6 g.cm−3) it has all the characteristics to be an efficient extracted from [7]).

reinforcing agent for nanocomposite applications. Indeed, the specific


Young's modulus, which is the ratio between the Young's modulus
and the density, of cellulose nanomaterials is around 65 J.g−1 for CNF two well-defined peaks and the relative magnitude of the high tempera-
and 85 J.g−1 for CNC whereas it is around 25 J.g−1 for steel. ture peak was found to increase when increasing the CNF content. This
This presumption was well verified in the pioneering work reported splitting of the relaxation process could be ascribed to strong interactions
for tunicin CNC reinforced poly(S-co-BuA) [16]. An outstanding im- between CNF and the NR matrix because of the presence of hydrophobic
provement on the rubbery modulus and its thermal stability were compounds at the surface of the filler. These interactions could lead to
observed even at low CNC content. In the rubbery state of the thermo- the formation of an interfacial layer surrounding the filler and which
plastic matrix, the modulus of the composite with a loading level as mobility is restricted compared to the bulk matrix. This phenomenon
low as 6 wt.% (i.e. 4 vol%) was more than two order of magnitude higher is obviously emphasized because of the nanometric scale of the filler
than the one of the unfilled matrix. Moreover, the introduction of 3 wt.% and omnipresence of the surface. Lower water uptake for CNF reinforced
or more CNC provides an outstanding thermal stability of the matrix NR compared to CNC composites [7], as well as higher contact angle for
modulus up to the temperature at which cellulose starts to degrade pure CNF film compared to CNC film [32] were additional evidences.
(500 K). This behavior was well described using a mechanical percola- For the second assumption (impact of the aspect ratio that could
tion approach. It was shown that in suitable conditions, a mechanically govern the critical nanoparticle content to reach percolation), one can
percolating stiff network of nanoparticles can form within the polymer refer to the study reported by [8]. In this report, CNC extracted from
matrix that supports the mechanical solicitation. The largest part of the the same source (Posidonia oceanica) was prepared from the leaves
observed reinforcing effect is attributed to this network and the intrinsic (POL) and from the balls (POB). CNC with different average width
stiffness of the nanoparticle plays only a minor role. The formation of (length) around 7 nm (520 nm) for POL and 8 nm (283 nm) for POB
this network is conditioned by the homogeneous dispersion of the filler, were obtained. It corresponded to an average aspect ratio around 75
the percolation threshold that depends on the aspect ratio of the nano- and 35 for CNC from POL and POB, respectively. Nanocomposite mate-
particles, and the strength of the filler/filler interactions. The stiffness of rials were prepared from these CNCs and poly(styrene-co-butyl acrylate)
the percolating CNC network was found to increase with the aspect ratio by casting/evaporation. A much higher reinforcing effect was observed
of the nanocrystal [9]. It therefore means that the use of higher aspect for POL compared to POL. From the comparison between experimental
ratio CNC is more interesting from a mechanical point of view because and predicted data using the percolation approach it was concluded
it first induces a decrease of the critical percolation threshold and also that the higher aspect ratio of CNC obtained from POL results in both a
stiffens the formed continuous network. lower percolation threshold and a stiffer percolating CNC network.
Another evidence of this fact lies in the highest reinforcing effect ob- Regarding the third assumption (effect of entanglements between
served for CNF than for CNC, despite the lower modulus of the former. flexible CNF inducing stronger connection between the nanoparticles),
This is well exemplified in Fig. 5 where the tensile modulus of nanocom- a good evidence was provided for nanocomposites prepared from
posites obtained by casting/evaporation from a mixture of natural poly(S-co-BuA) and cellulose nanomaterials obtained from sugar beet
rubber (NR) latex and CNF/CNC prepared from the rachis of date palm pulp [4]. CNF was used as the starting material and used as such or
tree is plotted as a function of the filler content [7]. Similar results acid hydrolyzed at different extends to have a range of morphology be-
were observed for CNC/CNF prepared from sisal using polycaprolactone tween CNF and CNC. All nanocomposite films contained 6 wt.% of cellu-
as matrix [32]. It was attributed to (i) the presence of residual hydro- lose nanomaterial. It was observed that both the tensile modulus and
phobic compounds such as lignin, extractive substances and fatty the tensile strength increased when adding the cellulosic filler, but
acids, at the surface of CNF that could play the role of compatibiliser; the reinforcing effect depended on the morphology of the filler. The
(ii) the higher aspect ratio for CNF that could result in the connection highest reinforcing effect was reported for the unhydrolyzed CNF-
of the nanoparticles for lower contents; but also (iii) the possibility of based composite. As the hydrolysis strength increased (going towards
entanglements for flexible CNF inducing stronger connections between CNC-like nanoparticles), both the modulus and the strength decreased,
the nanoparticles. showing the strong influence of entanglements on the rigidity of the
The first assumption (presence of residual hydrophobic compounds material. Moreover, it was shown that the elongation at break increased
at the surface of CNF) was well evidenced from the shape of the loss as the strength of the hydrolysis increased. This was an indication
angle peak observed using dynamic mechanical analysis (DMA) per- that, after destruction of the cellulosic network, the behavior of the
formed on CNC/CNF reinforced NR as shown in Fig. 6 [7]. Whereas for composite was mainly governed by the matrix. For the long entangled
CNC reinforced NR a single peak was observed for the main relaxation CNF-based composite the entangled cellulosic network governed the
process associated to the glass transition of the NR matrix, it splits into ultimate properties. As the result, the elongation at break decreased.
A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8 7

Fig. 6. Evolution of the loss angle tangent versus temperature at 1 Hz for NR films reinforced with (a) CNC and (b) CNF extracted from the rachis of date palm tree. Filler content: 0 (●),
1 (○), 2.5 (▲), 5 (Δ), 7 5 (×), 10 (■), and 15 wt.% (□) [7].

All these studies used a wet processing method, i.e. the nanocom- Apart these two phenomena which are inherent to the melt pro-
posites were prepared from never-dried CNC, allowing keeping the cessing technique, the homogeneous dispersion of hydrophilic cellu-
good dispersion state of the cellulosic nanomaterial in water and en- losic nanomaterials within a generally hydrophobic polymer matrix
suing homogeneous dispersion in the polymeric matrix after water requires their coating to hinder the formation of inter-particle hydrogen
evaporation. We have previously seen that different strategies can bonding. The reactivity of surface hydroxyl groups can be turned off
be used to avoid the irreversible aggregation of the nanoparticles by surface chemical grafting or physical coating of the nanoparticles as
upon drying and promote their homogeneous dispersion with limited previously explained. However, this strategy introduces defects and de-
thermal degradation within a hydrophobic matrix by melt extrusion. teriorates the cellulose nanoparticle percolation, and the full potential of
However, disappointing mechanical properties are reported for these the nanofiller is lost leading to mechanical properties which are far from
systems, far from the expectations observed for wet-processed nano- the expectations.
composites, even if improvement is generally observed compared to This effect was well exemplified for chitin nanocrystal (ChNC)
direct extrusion of unmodified nanoparticles. In addition, it is worth reinforced NR [17]. Neat ChNC as well as chemically modified ChNC
noting than in most cases a semicrystalline polymeric matrix was used using different coupling agents, namely phenyl isocyanate (PI), alkenyl
and it was clearly shown that cellulose nanomaterials generally act as succinic anhydride (ASA), and 3-isopropenyl-α,α’-dimethylbenzyl
nucleating agent inducing an increase of the degree of crystallinity of isocyanate (TMI) were prepared. All nanocomposite materials were
the matrix. This enhanced crystallinity of the polymeric matrix obvious- obtained by casting-evaporation thus avoiding mechanical degradation
ly participates to the improvement in the stiffness of the material, which and orientation of the nanorods. The liquid processing medium for
is difficult to dissociate from the real direct reinforcing effect of the hydrophilic neat ChNC was water (NR in latex form) whereas it was
nanoparticle. However, a recent publication showed that the dimen- toluene for hydrophobic modified ChNC (a solvent for NR). Even if im-
sions of CNC have a stronger impact on the crystallinity of the matrix proved filler/matrix adhesion was evidenced for the various chemical
and mechanical stiffness of the nanocomposite than its volume fraction, treatments, the mechanical performances of the composites strongly
and that the mechanical reinforcing effect is mostly caused by the CNC- decrease after the chemical modification. This loss of performance was
induced crystallinity of the matrix [25]. attributed to the partial or total destruction of the three-dimensional
A probable cause of this unexpected poor mechanical reinforcement ChNC network assumed to be present in the unmodified composites.
behavior results from the high shear rates involved in the extrusion
process. Indeed, a mechanical degradation (length decrease) of CNC
upon melt extrusion was reported [1]. Moreover, it can obviously 6. Conclusions
induce orientation of the elongated nanoparticles in the extrusion direc-
tion, limiting the formation of a percolating network, which is the basis Cellulose nanomaterials can be extracted from any cellulosic source
of the reinforcing effect of cellulose nanomaterials. This preferential ori- in the form of cellulose nanofibrils (CNF) or cellulose nanocrystals
entation phenomenon was evidenced using small angle X-ray scattering (CNC) using a mechanical or acid hydrolysis treatment, respectively.
(SAXS) experiments for CNC reinforced LDPE prepared by extrusion and Both nanomaterials display high stiffness and are lightweight making
was found to be amplified for increasing CNC contents [29]. It was also them ideal candidates for the preparation of polymer nanocomposites.
observed using 2D–small amplitude oscillary shear (2D–SAOS) experi- Casting-evaporation from an aqueous (or at least polar) medium is
ments for injection-molded CNC reinforced polybutyrate adipate tere- the most suited processing method, since it preserves the dispersion
phthalate (PBAT) [26]. The effect of a thermal annealing treatment state of the nanoparticles in this medium as well as the surface hydroxyl
of CNC reinforced injection-molded nanocomposites on the possible groups leading to unexpected mechanical properties. This slow wet
auto-reorganization of the nanofiller as a first step to create an isotropic process gives the highest mechanical performance materials compared
material, the initial condition to induce 3D–network formation under to other processing techniques. Indeed, during liquid evaporation
mild conditions, was investigated [26]. Even if the high viscosity of the strong interactions between nanoparticles can settle and promote the
polymer melt limited the movement of the nanoparticles and hindered formation of a strong percolating network through H-bonding. This
the formation of the desired percolating network, a spatial reorganiza- network ensures the mechanical stiffness of the material, the intrinsic
tion of the nanorods was observed after short conditioning time. It sug- stiffness of the nanoparticles surprisingly playing only a minor role.
gested that it should be possible to partially alter the organization of the However, it occurs only above their percolation threshold although
particle within the polymeric matrix imposed by the injection-molding this continuous network is overexploited and often mentioned in
process. However, adequate lower shear rates melt processing should the literature even if the filler content is lower than the percolation
limit both the breakage of CNC and its orientation. threshold. This strategy is probably well adapted for niche applications.
8 A. Dufresne / Current Opinion in Colloid & Interface Science 29 (2017) 1–8

However, if industrial and large scale manufacturing of cellulose [6] Ben Azouz K, Ramires EC, Van den Fonteyne W, El Kissi N, Dufresne A. ACS Macro
Lett 2012;1:236–40.
nanomaterial reinforced polymer nanocomposites is the final target, [7] Bendahou A, Kaddami H, Dufresne A. Eur Polym J 2010;46:609–20.
melt processing is definitively the most interesting technique, since [8] Bettaieb F, Khiari R, Dufresne A, Mhenni MF, Belgacem N. Carbohydr Polym 2015;
the final product can be easily shaped by extrusion, injection-molding, 123:99–104.
[9] Bras J, Viet D, Bruzzese C, Dufresne A. Carbohydr Polym 2011;84:211–5.
blow-molding or compression-molding, but also the most challenging. [10] Corrêa AC, Teixeira EM, Carmona VB, Teodoro KBR, Ribeiro C, Mattoso LHC, et al.
Issues of self-aggregation and thermal degradation have been clearly Cellulose 2014;21:311–22.
identified as the main challenges to overcome. Nevertheless, different [11] de Menezes AJ, Siqueira G, Curvelo AAS, Dufresne A. Polymer 2009;50:4552–63.
[12] Dufresne A, Vignon MR. Macromolecules 1998;31:2693–6.
strategies presented in this entry allow easily overcoming these chal- ••
[13] Dufresne A. Nanocellulose: From Nature to High Performance Tailored Materials.
••
lenges. The fact remains that the mechanical properties of ensuing Berlin/Boston: Walter de Gruyter GmbH & Co. KG; 2012.
materials are always disappointing and far from the expectation. It is [14] Dufresne A. Mater Today 2013;16:220–7.
[15] Eichhorn SJ, Dufresne A, Aranguren M, Marcovich NE, Capadona JR, Rowan SR, et al.
mainly attributed to the inevitable coating of the nanoparticles (to
J Mater Sci 2010;45:1–33.
avoid self-aggregation upon drying) and inherent hiding of surface [16] Favier V, Canova GR, Cavaillé JY, Chanzy H, Dufresne A, Gauthier C. Polym Adv
••
hydroxyl groups which prevent their mechanical percolation though Technol 1995;6:351–5.
H-bonding. The solution might be to adapt this coating in order to retain [17] Gopalan Nair K, Dufresne A, Gandini A, Belgacem MN. Biomacromolecules 2003;4:
1835–42.
some free OH groups for further nanoparticle interaction. Even more [18] Habibi Y, Goffin AL, Schiltz N, Duquesne E, Dubois P, Dufresne A. J Mater Chem 2008;
ideally, it might be to hide surface OH groups during processing with a 18:5002–10.
compound that could be eliminated by a specific treatment (thermal, [19] Herrera N, Mathew AP, Oksman K. Compos Sci Technol 2015;106:149–55.
[20] Herrick FW, Casebier RL, Hamilton JK, Sandberg KR. J Appl Polym Sci Appl Polym
radiative,…) after shaping the product. There are probably significant Symp 1983;37:797–813.
scientific and technological challenges to take up! [21]
••
Lin N, Huang J, Dufresne A. Nanoscale 2012;4:3274–94.
[22] Lin N, Dufresne A. Macromolecules 2013;46:5570–83.
[23] Malainine ME, Dufresne A, Dupeyre D, Mahrouz M, Vignon MR. Carbohydr Polym
Acknowledgments 2003;51:77–83.
[24] Mariano M, El Kissi N, Dufresne A. Eur Polym J 2015;69:208–23.
LGP2 is part of the LabEx Tec 21 (Investissementsd'Avenir — grant [25] Mariano M, Chirat C, El Kissi N, Dufresne A. J Polym Sci B 2016;54:2284–97.
[26] Mariano M, El Kissi N, Dufresne A. Langmuir 2016;32:10093–103.
agreement no. ANR-11-LABX-0030) and of the PolyNat Carnot Institut [27] Moon RJ, Martini A, Nairn J, Simonsen J, Youngblood J. Chem Soc Rev 2011;40:
(Investissementsd'Avenir — grant agreement no.ANR-11-CARN-030-01). 3941–94.
[28] Nagalakshmaiah M, El Kissi N, Dufresne A. ACS Appl Mater Interfaces 2016;8:
References and recommended reading•,•• 8755–64.
[29] Nagalakshmaiah M, Pignon F, El Kissi N, Dufresne A. RSC Adv 2016;6:66224–32.
[30] Nickerson RF, Habrle JA. Ind Eng Chem 1947;39:1507–12.
[1] Alloin F, D'Aprea A, Dufresne A, El Kissi N, Bossard F. Cellulose 2011;18:957–73.
[31]

Pereda M, El Kissi N, Dufresne A. ACS Appl Mater Interfaces 2014;6:9365–75.
[2] Arias A, Heuzey M-C, Huneault MA, Aussias G, Bendahou A. Cellulose 2015;22:
[32] Siqueira G, Bras J, Dufresne A. Biomacromolecules 2009;10:425–32.
483–98.
[33] Turbak AF, Snyder FW, Sandberg KR. J Appl Polym Sci Appl Polym Symp 1983;37:
[3] Azizi Samir MAS, Alloin F, Sanchez J-Y, Dufresne A. Polymer 2004;45:4033–41.
815–27.
[4] Azizi Samir MAS, Alloin F, Paillet M, Dufresne A. Macromolecules 2004;37:4313–6.
[5] Azizi Samir MAS, Alloin F, Dufresne A. Biomacromolecules 2005;6:612–26.


of special interest.
••
of outstanding interest.

You might also like