European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

Contents lists available at SciVerse ScienceDirect

European Polymer Journal

journal homepage: www.elsevier .com/locate /europol j

Macromolecular Nanotechnology

Incorporation of poly(glycidylmethacrylate) grafted bacterialcellulose nanowhiskers in poly(lactic acid) nanocomposites:Improved barrier and mechanical properties

0014-3057/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.eurpolymj.2013.04.035

⇑ Corresponding author. Tel.: +1 5105595794; fax: +1 5105595818.E-mail address: syed.imam@ars.usda.gov (S.H. Imam).

Marta Martínez-Sanz a, Mohamed A. Abdelwahab b,c, Amparo Lopez-Rubio a, Jose M. Lagaron a,Emo Chiellini b, Tina G. Williams d, Delilah F. Wood d, William J. Orts d, Syed H. Imam d,⇑a Novel Materials and Nanotechnology Group, IATA, CSIC, Avda. Agustín Escardino, 7, 46980 Paterna, Valencia, Spainb Laboratory of Bioactive Polymeric Materials for Biomedical and Environmental Applications (BIOlab), UdR INSTM, Department of Chemistry &Industrial Chemistry, University of Pisa, Via Risorgimento 35, 56126 Pisa, Italyc Department of Chemistry, Tanta University, 31527 Tanta, Egyptd USDA, Agricultural Research Service, WRRC-BCE 800 Buchanan Street Albany, CA 94710, USA

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

Article history:Received 16 January 2013Received in revised form 21 April 2013Accepted 29 April 2013Available online 23 May 2013

Keywords:Poly(lactic acid)Cellulose nanowhiskersBacterial cellulosePoly(glycidyl methacrylate)Surface modification

Poly(glycidyl methacrylate) (PGMA) was grafted onto bacterial cellulose nanowhiskers(BCNW) by means of a redox-initiated free radical copolymerization reaction. The incorpo-ration of PGMA chains decreased the thermal stability and crystallinity of BCNW. The neatand the PGMA-grafted BCNW were subsequently incorporated as fillers into the PLAmatrix. PGMA grafting improved both matrix-filler adhesion and the dispersion of cellulosenanocrystals. However, the dispersion of the nanofiller was still not completely optimizedand loadings higher than 3 wt.% resulted in increased agglomeration. The incorporation ofboth neat and PGMA-grafted BCNW significantly reduced the oxygen permeability of PLAfor low relative humidity conditions. However, due to the moisture sensitivity of cellulose,smaller improvements were attained when increasing the water activity. The detrimentaleffect of moisture on the oxygen permeability of nanocomposites was limited by the graft-ing copolymerization since PGMA-grafted BCNW seemed to present a more hydrophobicbehavior. Furthermore, increased elastic modulus and tensile strength were observed forall the nanocomposites, especially when the concentration of nanocrystals was aroundthe percolation threshold, i.e. 3 wt.%, but only nanocomposites containing PGMA-graftedBCNW preserved the ductility of neat PLA.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Currently, considerable research efforts have been fo-cused on developing new plastic materials with low envi-ronmental impact. Biodegradable polymers and materialsfrom renewable resources have attracted increased atten-tion for sustainable development and environmental con-servation. In particular, poly(lactic acid) (PLA) andpolyhydroxyalkanoates (PHAs) are biodegradable and

bio-based materials with promising properties. PLA is analiphatic polyester synthesized through lactic acid fermen-tation of natural feedstock such as corn or sugarcane [1,2]and it is one of the most widely used and well-studied bio-based polymers. It is highly transparent and rigid, withthermal and mechanical properties comparable topoly(ethylene terephthalate) (PET) [3]. However, its usehas been limited because of its brittleness, low heat distor-tion temperature (softening above 60 �C) and relativelylow barrier properties as compared to other packagingmaterials like PET. The incorporation of nanofillers suchas nanoclays, carbon nanotubes and cellulose nanocrystals

M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072 2063

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

has been proven to be an efficient strategy for enhancingthe properties of PLA and PLA-based hybrid materials [4–10].

Particularly, cellulose nanowhiskers (CNW) have gener-ated a great deal of interest in the field of nanocompositesbecause they offer renewable and biodegradable character,high surface area, low density, and high mechanicalstrength [11]. Cellulose is mostly derived from plantsources, but it is also synthesized by some bacterial spe-cies, and bacterial cellulose nanowhiskers (BCNW) can beextracted by subjecting bacterial cellulose to an acidhydrolysis treatment [12]. These BCNW have a highly crys-talline structure [12] and offer high aspect ratio [13,14].The main drawback for CNW is their hydrophilic character,making them difficult to disperse in non-polar media. Thismay cause CNW to agglomerate in organic solvents and ex-hibit poor adhesion with hydrophobic matrices. Surfacechemical modification by reaction of the hydroxyl groupspresent on the surface of cellulose nanocrystals with org-anomodifying agents has been proposed as one approachto overcome this shortcoming. In this regard, CNW havebeen successfully modified by acetylation [15–17], oxida-tion [18,19], silanization [20–22] and polymer graftingreactions [15,23–26] and incorporation of modified CNWinto polymeric matrixes, such as PLA has also been re-ported [15,17,21,22,27].

Plant derived CNW were partially silanized and subse-quently incorporated into PLA by solution casting, obtain-ing nanocomposites with an optimized nanofillerdispersion [21]. While the tensile strength and Young’smodulus increased with the incorporation of silanizedCNW, the elongation at break was reduced, suggesting thatthe matrix-filler adhesion did not improve. Similar resultswere observed for PLA nanocomposites incorporating acet-ylated CNW [27]. In a more recent study, silanized CNWwere incorporated into a PLA matrix by melt compoundingexhibiting a high level of nanofiller dispersion [22]. How-ever, improvements on the tensile or barrier propertieswere not reported. In another study, PLA chains weregrafted onto the surface of CNW by ring-opening polymer-ization and the subsequent incorporation into PLA by meltblending [15] indicated enhanced compatibility betweenthe matrix and the nanofiller. No barrier or mechanicalproperties were reported for such material. Surface modifi-cation of CNW may be a viable option for enhancing theirdispersion in organic solvents and to improve their com-patibility with hydrophobic matrices, but further researchis warranted to assess their impact on the mechanicaland barrier properties of PLA.

In this study, the surface of BCNW was chemically mod-ified by graft copolymerization of glycidyl methacrylate(GMA) using a redox-initiated free-radical method andmodified BCNW were subsequently incorporated into PLAby solution casting. The epoxy group in the GMA chainswas expected to react with the carboxylic group of PLA,thus promoting interfacial adhesion between BCNW andthe polyester matrix. While cellulose modification by graftcopolymerization of (GMA) has been reported [25,26],information on the incorporation of the grafted filler intopolymeric matrixes is lacking. The morphology, thermalproperties, oxygen barrier, water affinity and mechanical

properties of BCNW and BCNW-PGMA loaded PLA nano-composites were investigated and compared to determinethe influence of surface modification on the nanocompos-ites performance.

2. Materials and methods

2.1. Materials

Poly(lactic acid) (PLA) was a Polylactide Resin 4042Dgrade (Natureworks, LLC, Minnetonka, MN) with a weightaverage molecular weight (Mw) of ca. 6.6 � 104 g/mol.Glycidyl methacrylate (GMA) and ammonium cerium (IV)nitrate (CAN) were purchased from Aldrich (St. Louis,MO, USA). Sulfuric acid (96%) was purchased from Panreac(Barcelona, Spain). Nitric acid, methanol and acetone weresupplied by Fisher Scientific (New Jersey, USA). Tetrahy-drofuran (THF) and chloroform were purchased from Sig-ma–Aldrich (St. Louis, MO, USA).

2.2. Preparation of bacterial cellulose mats

Bacterial cellulose was obtained according to a pub-lished protocol [28]. Briefly, the bacterial cells (Glu-conacetobacter xylinus 7351) were grown in a modifiedHestrin/Shramm medium at 30 �C. The cells were pre-cul-tured in a test tube containing 5 mL of media and latertransferred and scaled up to a larger reactor of 20 L at Bioi-nicia S.L., Valencia (Spain). The synthesized bacterial cellu-lose pellicles were sterilized and cleaned in boiling waterfollowed by washing in a 10% (v/v) NaOH aqueous solution.

2.3. Preparation of bacterial cellulose nanowhiskers (BCNW)

At a neutral pH, bacterial cellulose (BC) pellicles wereground in a blender and the gel-like material was thencompressed to remove most of the absorbed water. Thismaterial was then subjected to acid hydrolysis accordingto a previously optimized method [12]. BC was treatedwith a sulfuric acid solution (301 mL sulfuric acid/L water).About 7 g of BC was suspended in 1 L of sulfuric acid solu-tion and it was heated and maintained at 50 �C with stir-ring for three days until a homogeneous solution wasobtained. The cellulose nanowhiskers were recovered as awhite precipitate after four centrifugation (12,500 rpm,20 min) cycles in a Beckman Avanti J-25 centrifuge witha JA14 rotor. Recovered material was re-suspended indeionized water and neutralized with sodium hydroxide(7.0 pH) and centrifuged (12,500 rpm, 20 min) to obtainapartially hydrated precipitate which was stored at 4 �Cor lyophilized for subsequent use. A portion of the materialwas solvent exchanged into acetone and chloroform. Thewater of the partially hydrated precipitate was replacedwith acetone by applying several centrifugation cycles inwhich the supernatant was removed and replaced withacetone. After that, several centrifugation cycles were ap-plied and acetone was replaced by chloroform. By meansof this procedure BCNW were suspended into chloroform.The concentration of BCNW after chloroform evaporationwas 48 wt.%.

2064 M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

2.4. Graft copolymerization of BCNW (PGMA-grafted BCNW)

PGMA-grafted BCNW were prepared according to a pre-viously published study [25]. Partially hydrated BCNWwere diluted with water to obtain a 0.2 wt.% suspensionand the pH was adjusted to 1 with dilute nitric acid. Thissuspension was stirred for 30 min while nitrogen was bub-bled through. Ammonium cerium (IV) nitrate (CAN) wasadded (4 mmol/L) as the initiator and stirred (15 min,35 �C). The GMA monomer was incorporated graduallyover a period of 30 min at a concentration of 40 mmolper g of dry BCNW and the mixture was stirred for 60 min.

After the reaction, the product was centrifuged(12,000 rpm, 20 min) and the precipitate was repeatedlywashed with deionized water (to remove traces of the ini-tiator and the acid) and vacuum filtered. Subsequently, thematerial was washed with THF to separate the homopoly-mer from the graft copolymer. To do this, the material wassuspended in 40 mL of THF and added dropwise to 400 mLchilled methanol (ice bath) and continuously stirred over-night after which the material was vacuum filtered andwashed with THF again. The PGMA-grafted BCNW wasthen vacuum dried overnight to remove any traces ofmethanol.

2.5. Preparation of PLA-BCNW nanocomposites

PLA-BCNW nanocomposites were prepared by solutioncasting followed by compression molding. Chloroformsolutions containing 5% (w/v) of polymer (PLA) were stir-red at room temperature for 48 h. Nanowhiskers (neatand PGMA-grafted BCNW) were added to the solutionsand dispersed by means of intense homogenization (Ul-tra-turrax). The samples were placed in an ultrasound bathfor 1 h to remove bubbles. The solutions were poured intoglass Petri dishes and placed on a leveled platform. Chloro-form was allowed to evaporate over 3 days and residualsolvent was removed from films by drying at 35–40 �Cfor 24 h. In order to improve the reaction between PLAand the modified nanowhiskers, about 1 g of the solutioncast samples were further pressed in a Carver press to ob-tain films of around 300–500 lm thickness. For this, sam-ples were placed in a mold and heated to 185 �C and leftisothermally for 3 min at 28 MPa. Films of neat PLA andPLA containing 1, 3 and 5 wt.% (relative to PLA weight) ofboth neat and PGMA-grafted BCNW were produced.

2.6. Transmission FT-IR analysis

For FT-IR analysis, about 2 mg of each sample wereground and dispersed in 200 mg of spectroscopic gradeKBr. A pellet was then formed by compressing the sampleat 150 MPa. Transmission FT-IR measurements were car-ried out under controlled conditions (21 �C, 40%RH) usinga Bruker (Rheinstetten, Germany) FT-IR Tensor 37 equip-ment. The spectra were taken at 1 cm�1 resolution averag-ing a minimum of 10 scans. Analysis of the spectra wasperformed using Grams/AI 7.02 (Galactic Industries, Salem,NH, USA) software.

2.7. Scanning Electron Microscopy (SEM)

Nanocomposites were cryo-fractured in liquid nitrogento observe the interior of the unstressed composite. Thenanocomposite sample was dropped directly into liquidnitrogen and fractured with a pre-chilled razor blade heldin a vice-grip. The fractured pieces were picked out of theliquid nitrogen using pre-chilled forceps and placed in adesiccator to thaw and reduce the condensation on thesurface of the material. Each specimen was placed on analuminum stub using double-sided adhesive tape andcoated with gold–palladium for 45 s in a Denton Desk IIsputter coating unit (Denton Vacuum USA, Moorestown,NJ). Specimens were viewed in a Hitachi S4700 field emis-sion scanning electron microscope (Hitachi HTA, Japan) at2 kV.

2.8. Differential scanning calorimetry (DSC)

Differential scanning calorimetry (DSC) was performedon a Perkin–Elmer DSC 7 thermal analysis system. Typi-cally, 4 mg of dry material were analyzed at a scanning rateof 10 �C/min from room temperature to the melting pointusing N2 as the purging gas. The first and second meltingendotherms after controlled crystallization at 10 �C/minfrom the melt were analyzed. Before evaluation, similarruns of an empty pan were subtracted from the thermo-grams. The DSC equipment was calibrated using indiumas a standard.

The crystallinity (%) of the PLA was estimated from thecorrected enthalpy for biopolymer content in the hybrid fi-bers, using the ratio between the enthalpy of the studiedmaterial and the enthalpy of a perfect PLA crystal, i.e.XCð%Þ ¼

DHf�DHc

DH0f� 100, where DHf is the enthalpy of fusion

and DHC the enthalpy of cold crystallization of the studiedspecimen, and DH0

f is the enthalpy of fusion of a totallycrystalline material. The DH0

f used for this equation was93 J/g for PLA [29].

2.9. Thermogravimetric analysis (TGA)

Thermogravimetric analyses (TGA) were performed un-der nitrogen atmosphere using a TA Instruments (TGA-2950). The samples were heated from room temperatureto 500 �C at a heating rate of 10 �C/min and at a nitrogengas flow rate of 60 mL/min. The derivatives of TGA curves(DTG) were obtained using TA analysis software.

2.10. Oxygen permeability

The oxygen permeability (OP) coefficient was derivedfrom oxygen transmission rate (OTR) measurements re-corded using an Oxtran 100 test system (Modern ControlInc., Minneapolis, MN, US). Experiments were carried outat 24 �C under both 0%RH and 80%RH. The 80% relativehumidity was generated by a built-in gas bubbler andwas checked with a hygrometer placed at the exit of thedetector. The samples were purged with nitrogen for aminimum of 20 h prior to exposure to an oxygen flow(10 ml/min). A 5 cm2 sample area was measured byusing an in-house developed mask. For testing oxygen

Fig. 1. PGMA-BCNW graft copolymerization mechanism.

M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072 2065

TECH

NO

LOG

Y

permeation in highly permeable materials, the samplingarea was reduced thus improving the reproducibility ofthe measurements.

2.11. Tensile testing

Young’s modulus, stress at break, and strain at breakwere determined according to ASTM method D638M un-der ambient conditions, using an Instron 5500R UniversalTesting Machine (Instron Corp., Canton, MA). Prior to test-ing, samples were equilibrated for one day at 50% relativehumidity in a chamber containing saturated solutions ofcalcium nitrate. The testing conditions used were: crosshead speed of 5 mm/min and load cell of 0.1 kN. Dumb-bell-shaped samples (0.2–0.5 mm thick) were tested witha gauge length of 20 mm. The reported values were theaverage of at least 10 measurements.

2.12. Contact angle measurements

Measurements of contact angle were performed at24 �C and ambient relative humidity (60%RH) in a Video-Based Contact Angle Meter model OCA 20 (DataPhysicsInstruments GmbH, Filderstadt, Germany). Contact anglemeasurements were obtained by analyzing the shape of adistilled water drop after it had been placed over the filmfor 30 s. Image analyses were carried out by SCA20software.

MA

CRO

MO

LECU

LAR

NA

NO

3. Results and discussion

3.1. Characterization of neat BCNW and PGMA-grafted BCNW

BCNW were modified by grafting glycidyl methacrylate(GMA), according to the grafting mechanism which isshown in Fig. 1. Firstly, ceric ion forms a chelate complexwith the glucosidic ring of BCNW that decomposes to gen-erate free radicals on the cellulose backbone. GMA has tworeactive functional groups, an acrylic and an epoxy group.The acrylic group is expected to react with the free radicalsand lead to graft copolymerization onto the surface of thenanocrystals [30,31]. As a result, BCNW with an epoxyfunctionality, which is reactive with the carboxylic groupof PLA, were generated.

To characterize the chemical structure and to confirmthe grafting of PGMA, FT-IR spectra of the BCNW beforeand after the copolymerization reaction were analyzed.Both neat and PGMA-grafted BCNW displayed absorptionpeaks that were characteristic of cellulose (cf. Fig. 2a). Inaddition, the PGMA grafted sample showed a new, strongband at 1724 cm�1, characteristic of an ester carbonylgroup [25], as well as the peaks at 905 cm�1 and844 cm�1, which are characteristic of the terminal vinylgroup of the PGMA chains and the epoxy group [25]respectively, thus confirming the presence of graftedPGMA.

Previous works demonstrated that the band ratio A1375/A2902, may be used to estimate the cellulose total crystal-line index by FT-IR [32,33]. The spectrum of PGMA-graftedBCNW was corrected by subtracting the PGMA spectrum

(cf. Fig. 2b) and a baseline correction was performedaccording to a previous work [33]. After that, the A1375/A2902 band ratio was calculated, obtaining a value of 0.85for neat BCNW and 0.17 for PGMA-grafted BCNW. This re-sult suggests that the crystalline structure of the nativenanocrystals may have been distorted to a certain extentas a consequence of the grafting reaction. For the graftcopolymerization parameters applied in this work, thegraft yield was 434% and the weight fraction of polymerin the modified CNW was 81% which was quite similar tothe work reported previously. Also, the extent of graftingwith GMA was much higher when compared to othermonomers like methyl methacrylate (MMA), ethyl acrylate(EA), 2-hydroxyethyl methacrylate (HEMA) and butyl acry-late (BuA) [25]. Therefore, an extensive grafting may haveaffected not only the more accessible cellulose amorphousdomains but also the crystalline regions.

As BCNW were intended for use as nanofillers in con-ventional plastic matrices, it was important to determinethe degradation temperature of the material to define thetemperature range in which the nanofiller could be pro-cessed without suffering thermal degradation. Fig. 3 showsthe DTG curves of the nanowhiskers before and after PGMAgrafting. The surface modification clearly resulted in amaterial with decreased thermostability. While the neatBCNW sample presented a sharp degradation peak with amaximum at approximately 319 �C, the PGMA-graftedBCNW sample showed multiple degradation steps. This re-sult is in agreement with previous studies in which PGMAwas grafted onto cotton cellulose [26] and nanofibrillatedcellulose [25].

Fig. 2. FT-IR spectra of neat BCNW and PGMA-grafted BCNW (a). FT-IRspectra of neat BCNW, PGMA-grafted BCNW, PGMA and PGMA-graftedBCNW after subtraction of the PGMA spectrum. The arrows indicate thebands which were used to estimate the crystallinity ratio in the neat andPGMA-grafted BCNW (b).

Fig. 3. Derivative thermogravimetric curves of neat BCNW and PGMA-grafted BCNW.

2066 M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

As deduced from the TGA parameters presented in Ta-ble 1, the degradation of native cellulose took place be-tween ca. 250 �C and 350 �C, while for the PGMA-grafted

filler, degradation occurred over a wider temperaturerange. Specifically, two different degradation stages couldbe distinguished for the cellulose fraction, followed bytwo additional ones for the PGMA. Regarding cellulosedegradation, a first peak with a maximum at ca. 221 �Cwas observed. This peak could be ascribed to the cellulosegrafted fraction which, as a consequence of excessive graft-ing, could be losing its highly crystalline nature to a certainextent, as previously pointed out by FT-IR analyses. This isin agreement with previous works on cellulose surfacemodification which have demonstrated that excessive sub-stitution may be detrimental to the structure and proper-ties of cellulose nanowhiskers [34,35]. Anotherdegradation peak for cellulose was observed, with a maxi-mum approximately matching that of neat cellulose, whichprobably corresponded to the ungrafted cellulose. For thePGMA degradation profile, one major peak and one shoul-der peak at ca. 367 �C and 396 �C were observed, respec-tively. In this regard, it has been shown that thedegradation of PGMA is a two step process, where one stepcorresponds to the depolymerization reaction and theother to the decomposition of esters [36].

3.2. Characterization of PLA nanocomposites

The neat and PGMA-grafted BCNW were incorporatedin a PLA matrix and morphology, thermal, barrier andmechanical properties of the produced nanocompositeswere evaluated.

3.2.1. Morphological characterizationInitial examination of the morphology by optical

microscopy indicated higher extent of agglomeration inthe nanocomposites incorporating neat BCNW comparedto the grafted BCNW as shown in Fig. 4a–f, where cellulosenanocrystals appeared as bright areas. When incorporatedin the amount of 5 wt.%, both native and grafted BCNWproduced agglomeration (cf. Fig. 4e and f).

To study the dispersion of the nanofiller within the ma-trix, the cryo-fractured surfaces of nanocomposite filmswere examined by SEM (Fig. 5a–g). For the nanocompositefilms, the figures correspond to the areas were agglomer-ates were detected. Observation of the samples confirmedthat neat BCNW agglomerated to a large extent in thenanocomposites (cf. Fig. 5b, d and f). Even though the neatBCNW had been subjected to a solvent exchange process toimprove their dispersion in organic solvents, their hydro-philicity led to low compatibility with the non-polar sol-vent in which PLA was dissolved. Furthermore, phaseseparation occurred between the polymeric matrix andthe BCNW agglomerates. The agglomeration and lack ofadhesion was reduced when the PGMA-grafted BCNWwere used. The grafted PGMA chains showed better com-patibility with the PLA matrix and, thus, PGMA-graftedBCNW seemed to be better distributed in the film. Nophase separation between the PLA and the PGMA-graftedBCNW was apparent in samples (cf. Fig. 5c, e and g). Nev-ertheless, the nanocrystals were still aggregated to a cer-tain extent and it was not possible to obtainnanocomposites with individual cellulose nanowhiskersuniformly dispersed in the PLA.

Table 1TGA maximum of the weight loss first derivate (TD1, TD2, TD3 and TD4) and the corresponding peak onset values (Tonset) and the residue at 490 �C (R490) for neatBCNW and PGMA-grafted BCNW.

Tonset (�C) TD1 (�C) TD2 (�C) TD3 (�C) TD4 (�C) R490 (%)

BCNW 295.5 ± 2.2 – 318.0 ± 1.3 – – 5.5 ± 3.5PGMA-grafted BCNW 197.6 ± 4.8 220.8 ± 2.6 319.1 ± 6.5 366.7 ± 2.2 395.5 ± 0.6 13.2 ± 0.9

Fig. 4. Optical microscopy images taken under polarized light of PLA nanocomposite films loaded with: (a) 1 wt.% neat BCNW; (b) 1 wt.% PGMA-graftedBCNW; (c) 3 wt.% neat BCNW; (d) 3 wt.% PGMA-grafted BCNW; (e) 5 wt.% neat BCNW and (f) 5 wt.% PGMA-grafted BCNW. Scale markers correspond to50 lm.

M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072 2067

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

Thus, morphological evaluation by optical microscopyand SEM revealed an improved dispersion of the PGMA-grafted BCNW, which showed increased compatibility withthe PLA phase and led to good matrix-filler adhesion. Aloading of 3 wt.% was found to be optimal to avoid largeagglomeration for both neat and PGMA-grafted BCNW.

3.2.2. Thermal properties and thermal stabilityAddition of both native and modified cellulose nano-

crystals in the PLA matrix caused a slight increase in themelting temperature which was accompanied by a de-crease in the cold crystallization temperature and a signif-icant increase in the crystallinity index (cf. Table 2).

Addition of neat BCNW or PGMA-grafted BCNW had lit-tle impact on the melting temperature of the PLA compos-ites. The lower cold crystallization temperatures observedin PLA composites containing neat or PGMA-grafted BCNWindicated that cellulose nanocrystals promoted the crystal-lization of PLA upon heating. The same effect has been pre-viously reported for PLA nanocomposites processed bymelt compounding in which highly dispersed BCNW actedas nucleating agents, promoting a faster crystallization ofPLA during heating [7]. This effect was more apparent at

lower nanofiller loadings and, therefore, in relation withthe morphological characterization, it appears that thisnucleating effect was favored by high nanofiller dispersion.A significant increase in the crystallinity index of nano-composites, compared to neat PLA, further supported thenucleating effect of cellulose nanocrystals. A similar in-crease in crystallinity was observed in PLA composites con-taining plant cellulose nanowhiskers produced by solutioncasting, suggesting that highly dispersed CNW acted asefficient nucleating agents [37]. Notably, the maximumcrystallinity increase was seen in samples loaded with1 wt.% nanofiller. Comparatively, samples with higherloadings of PGMA-grafted BCNW had slightly higher crys-tallinity than those loaded with neat BCNW. The glass tran-sition temperature was not greatly impacted by thepresence of cellulose nanowhiskers. These results werenot surprising as previous studies have shown that thepresence of cellulose nanowhiskers in a PLA matrix had lit-tle effect on the Tg of nanocomposites, even with a highnanofiller dispersion [37].

Thermal stability of PLA nanocomposites was studiedby TGA and Table 3 displays the parameters estimatedfrom thermogravimetric curves. Differences were observed

Fig. 5. SEM micrographs from the cryo-fractured surface of nanocomposite films: (a) Neat PLA; (b) PLA loaded with 1 wt.% neat BCNW; (c) PLA loaded with1 wt.% PGMA-grafted BCNW; (d) PLA loaded with 3 wt.% neat BCNW; (e) PLA loaded with 3 wt.% PGMA-grafted BCNW; (f) PLA loaded with 5 wt.% neatBCNW and (g) PLA loaded with 5 wt.% PGMA-grafted BCNW.

Table 2DSC maximum of melting (Tm), melting enthalpy (DHm), cold crystallization temperature (TCC), cold crystallization enthalpy (DHCC) and PLA crystallinity (Xc) ofPLA and its nanocomposite films with native and PGMA-grafted BCNW, obtained during the first heating run and glass transition temperature (Tg2) obtainedduring the second heating run.

Tm (�C) DHm (J/g) TCC (�C) DHCC (J/g) Xc (%) Tg2 (�C)

PLA 167.0 ± 0.1 29.8 ± 14.3 106.9 ± 1.4 21.7 ± 11.6 8.7 ± 2.9 61.0 ± 0.01%BCNW 167.6 ± 0.4 42.5 ± 0.5 100.8 ± 0.4 27.4 ± 1.0 16.2 ± 0.5 59.7 ± 0.43%BCNW 167.8 ± 0.4 41.1 ± 0.4 103.2 ± 0.7 30.1 ± 0.0 11.8 ± 0.4 60.0 ± 0.15%BCNW 168.0 ± 0.9 40.6 ± 0.2 107.8 ± 0.6 33.5 ± 0.2 7.7 ± 0.0 61.7 ± 0.71%PGMA-grafted BCNW 167.6 ± 0.4 40.9 ± 1.3 100.4 ± 0.0 26.6 ± 1.9 15.5 ± 0.6 59.9 ± 0.13% PGMA-grafted BCNW 167.4 ± 0.2 41.1 ± 1.6 102.9 ± 0.2 29.0 ± 0.7 12.9 ± 2.5 61.3 ± 0.15% PGMA-grafted BCNW 167.7 ± 0.0 39.9 ± 0.1 101.7 ± 0.9 30.3 ± 1.2 10.2 ± 1.4 60.9 ± 0.0

2068 M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

in the thermal stability of the nanocomposites loaded withneat or PGMA-grafted BCNW (Table 3). Low concentrationsof neat BCNW slightly increased the thermal stability ofthe nanocomposites, which was maximum for 3 wt.%BCNW loading. Other research works have shown thatthermal degradation was limited by the incorporation ofCNW since a strong nanowhiskers network, which re-

stricted the mobility of the polymer chains, was created[7,38]. In contrast, the incorporation of PGMA-graftedBCNW in PLA yielded composites with significantly de-creased thermal stability compared to that of neat PLA.As noted earlier, the grafting of PGMA onto the cellulosechains decreased the thermal stability of the nanofillerand, any addition of such fillers in a matrix was then

Table 3TGA maximum of the weight loss first derivate (TD) and the correspondingpeak onset values (Tonset) and the residue at 490 �C (R490) for neat PLA andPLA nanocomposites with native and PGMA-grafted BCNW.

Tonset (�C) TD (�C) R490 (%)

PLA 328.9 ± 3.5 362.1 ± 1.8 0.3 ± 0.21%BCNW 330.9 ± 6.3 362.7 ± 1.5 0.2 ± 0.13%BCNW 335.5 ± 1.6 364.1 ± 0.1 0.1 ± 0.05%BCNW 329.3 ± 0.7 356.9 ± 0.5 0.4 ± 0.51% PGMA-grafted BCNW 314.3 ± 0.1 348.5 ± 1.0 0.6 ± 0.03% PGMA-grafted BCNW 316.2 ± 3.7 349.0 ± 3.0 0.5 ± 0.15% PGMA-grafted BCNW 310.5 ± 0.3 346.6 ± 3.9 0.5 ± 0.1

M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072 2069

expected to have a detrimental effect on the thermal sta-bility of the nanocomposites. Additionally, the amount ofhydroxyl bonds available to establish interactions betweencellulose chains may also be reduced by PGMA grafting,diminishing any favorable effect of the nanocrystalsnetwork.

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

3.2.3. Oxygen permeability and surface water affinityOxygen permeability (OP) was measured both at 0%RH

and 80%RH and results are presented in Table 4. The per-meability value for neat PLA was similar to that previouslyreported for melt compounded PLA films [39,40]. Consis-tent with other studies, PLA displayed increased oxygenbarrier at higher relative humidity. This effect is attributedto a combination of increased oxygen diffusion and de-creased oxygen solubility at a higher relative humidity [3].

At 0%RH, the incorporation of native or PGMA-graftedBCNW into the PLA matrix resulted in significant reduc-tions in OP values. However, different trends were ob-served for the OP values when increasing the neat andPGMA-grafted BCNW loadings. While higher loadings ofPGMA-grafted BCNW resulted in an OP drop, a slight in-crease in the OP was observed when increasing the con-centration of neat BCNW. This observation points outthat the improved dispersion and matrix-filler adhesionachieved for the PGMA-grafted BCNW were responsiblefor the improved oxygen barrier attained when increasingtheir loading. A maximum OP drop of ca. 26% and ca. 32%was attained in samples with 1 wt.% BCNW and 5 wt.%BCNW-PGMA, respectively. Such degree of reduction inthe OP highlights the oxygen barrier properties of BCNWin PLA composites. It is also important to mention that,as previously suggested by FT-IR analyses, PGMA-graftedBCNW possessed lower crystallinity than neat BCNW andtherefore, they should present, in principle, a less effectivebarrier impact. This may explain why the OP was slightlylower for neat BCNW than for PGMA-grafted BCNW for aconcentration of 1 wt.%.

Table 4Oxygen permeability measured at 0%RH and 80%RH for PLA and nanocomposites

P 0%RH (m3 m/

PLA 2.15 ± 0.42e�18

1%BCNW 1.60 ± 0.19e�18

3%BCNW 1.65 ± 0.32e�18

5%BCNW 1.68 ± 0.54e�18

1% PGMA-grafted BCNW 1.74 ± 0.10e�18

3% PGMA-grafted BCNW 1.62 ± 0.23e�18

5% PGMA-grafted BCNW 1.46 ± 0.10e�18

Generally, under high relative humidity conditions(80%RH) addition of cellulose nanocrystals still reducedthe OP values. The only exception was the composite with5 wt.-% BCNW loading. It may also be observed that a trendof increased OP was produced when increasing the nano-filler loading, especially for the neat BCNW. A maximumreduction of ca. 17% and ca. 13% was observed in compos-ites with 1 wt.% neat BCNW and 1 wt.% PGMA-graftedBCNW, respectively. It is important to note that the perme-ability drop observed for neat PLA at higher water activitywas reduced or even reversed upon addition of neat BCNWinto the matrix. Similar effects were previously reportedfor PLA-BCNW nanocomposites, which presented optimaloxygen barrier properties for reduced BCNW loadings [7].These results could be explained by the fact that the addi-tion of BCNW further increased the oxygen diffusion coef-ficient of PLA at high relative humidity due to a strongerplasticization effect caused by the nanowhiskers.

In one study, up to 90% reduction in the OP at 80%RHhas been reported for PLA nanocomposites (solvent cast)loaded with cellulose nanowhiskers [37]. The lowest OPvalue reported by that study was 1.4 � 10�18 m3 m/m2 s Pa, which is quite similar to the range of OP reportedhere.

The fact that the addition cellulose nanocrystals limitedthe drop in the OP values of the PLA matrix at high RH isrelated to the hydrophilic character of cellulose, which al-lows water molecules to interact with hydroxyl groupsfrom BCNW distorting the hydrogen bonds networkestablished between cellulose chains and/or the polymericmatrix. Thus, water diffusion through the films was pro-moted at high RH.

The incorporation of PGMA-grafted BCNW providedmore evident OP reductions at high relative humidity thanthe addition of neat BCNW, suggesting that PGMA-graftedBCNW presented a less hydrophilic character. In order toinvestigate the different water affinity of PLA nanocompos-ites, contact angle measurements were carried out (resultslisted in Table 5). Data indicated that the contact angle wasnot significantly altered in the nanocomposites as com-pared to neat PLA. Despite BCNW being highly hydrophilic,at low nanofiller loadings all the nanocomposites had con-tact angles higher than 70�, which are typical of hydropho-bic surfaces [41]. With neat BCNW there was a trend ofdecreasing contact angle, i.e. more hydrophilic surface,when increasing the amount of nanocrystals in the com-posite. This trend was not observed with PGMA-graftedBCNW. In addition, in the composites with the highestPGMA-grafted BCNW concentration (5 wt.-%), a signifi-cantly higher contact angle (i.e. more hydrophobic surface)

containing neat BCNW and PGMA-grafted BCNW.

m2 s Pa) P 80%RH (m3 m/m2 s Pa)

1.63 ± 0.12e�18

1.36 ± 0.21e�18

1.60 ± 0.20e�18

1.86 ± 0.25e�18

1.42 ± 0.15e�18

1.51 ± 0.15e�18

1.55 ± 0.12e�18

Table 5Contact angles measured for PLA and its nanocomposites with neat andPGMA-grafted BCNW.

Contact angle (�)

PLA 76.3 ± 1.4ab

1%BCNW 76.2 ± 1.4ab

3%BCNW 75.2 ± 1.5ab

5%BCNW 74.2 ± 1.4a

1% PGMA-grafted BCNW 76.6 ± 1.7b

3% PGMA-grafted BCNW 75.6 ± 0.7ab

5% PGMA-grafted BCNW 77.0 ± 0.9b

The a and b letters correspond to the ANOVA statistical analysis andTukey test of the data indicating that with a 95% confidence level, thevalues are significantly different.

2070 M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

was observed compared to the composites with neatBCNW. This confirmed that the grafting process replacedhydroxyl groups on the surface with non-polar GMA chainsand conferred a more hydrophobic character to the PGMA-grafted BCNW. This also explains why the OP reductions at80%RH were greater in the PGMA-grafted BCNW comparedto neat BCNW.

3.2.4. Mechanical propertiesDespite showing some relatively large variability in

measurements, addition of both neat and PGMA-graftedBCNW resulted in increased Young’s modulus and tensilestrength (cf. Table 6). Notably, samples containing PGMA-grafted BCNW were slightly stiffer, with higher Young’smodulus and tensile strength, than those with neat BCNW.This probably resulted from better adhesion between thePGMA-grafted BCNW and the matrix and from the higherlevel of nanocrystals dispersion.

The elongation at break decreased slightly in compos-ites with neat BCNW compared to the neat PLA matrix.Interestingly, this decrease in elongation was not seen incomposites with PGMA-grafted BCNW. In general, theaddition of reinforcing agents in a polymeric matrix resultsin lower elongation at break values due to the stress con-tributed by their presence. However, previous studies haveshown that the development of some strong interactions,such as hydrogen bonding between the matrix and fillermay help reduce the stress concentration effect to a certainextent [42]. Results presented here suggest that the surfacemodification of BCNW promoted the interactions betweenthe PLA matrix and the nanocrystals and at the same timeimproved their dispersion, resulting in improved mechan-ical properties.

Interestingly, nanocomposites containing both neatBCNW and PGMA-grafted BCNW had optimal mechanical

Table 6Young’s modulus (E), tensile strength (rB) and elongation at break (eB) forPLA and nanocomposites containing neat BCNW and PGMA-grafted BCNW.

E (GPa) rB (MPa) eB (%)

PLA 2.41 ± 0.50 45.28 ± 8.68 4.12 ± 1.671%BCNW 2.50 ± 0.18 49.49 ± 2.11 3.68 ± 0.653%BCNW 2.57 ± 0.13 50.64 ± 7.59 3.53 ± 1.015%BCNW 2.42 ± 0.36 47.95 ± 6.67 3.90 ± 0.141% PGMA-grafted BCNW 2.53 ± 0.14 45.19 ± 3.04 5.53 ± 1.983% PGMA-grafted BCNW 2.79 ± 0.08 55.17 ± 0.37 4.09 ± 0.325% PGMA-grafted BCNW 2.61 ± 0.19 49.34 ± 5.81 4.02 ± 0.49

properties at a loading of 3 wt.%. This may be related tothe presence of a strong network of nanocrystals createdat this particular loading. The percolation threshold, wherethe nanowhiskers are interconnected by a 3D network, is acrucial condition to obtain strong mechanical reinforce-ment. By using the following equation, the percolationthreshold (tRc) can be estimated as a function of the aspectratio [43–45]:

tRc ¼0:7L=d

where L/d is the aspect ratio. As previously reported, BCNWhad an aspect ratio of approximately 30 [12]. Therefore,the percolation threshold should be reached with 2 v.-%of BCNW. This is indeed in agreement with the optimummechanical properties achieved in composites with3 wt.% nanofiller loading (which corresponds to 2.3 v.-%).The same effect was observed for both EVOH [45] andPLA [7] melt processed nanocomposites loaded withBCNW, which displayed optimal mechanical performancefor loadings between 2 wt.% and 3 wt.%.

Thus, the nanocomposite with 3 wt.% PGMA-graftedBCNW appears to be optimum in terms of mechanicalproperties, with an increase of ca. 16% and 22% in theYoung’s modulus and tensile strength, respectively, with-out impairing the ductility of the material. Similarimprovements have been reported for PLA nanocompositesloaded with BCNW prepared by melt compounding,although in that case the nanocomposites were much stif-fer due to the higher crystallinity resulting from the meltprocessing [7]. In contrast, a plasticization effect was re-ported for PLA nanocomposites loaded with plant-derivedCNW prepared by solution casting, presenting not onlyhigher elongation at break but also lower modulus [37].It seems that due to the lower aspect ratio of vegetalCNW, higher loadings were required in order to achieveimprovements in the mechanical properties. The data pre-sented here, show an efficient reinforcement effect pro-vided by the high aspect ratio of the BCNW fillers andthe effect of PGMA grafting onto the surface of BNCW toimprove their compatibility with the PLA matrix.

4. Conclusions

In this work, epoxy functionality was successfully intro-duced onto the BCNW surface by using cerium (IV), as free-radicals promoter, followed by grafting of glycidyl methac-rylate (GMA). The chemical modification of the BC nano-crystals surface was confirmed by FT-IR and TGA. It wasobserved that the thermal stability of BCNW was slightlyreduced as a result of grafting, most likely due to the dis-tortion of their highly crystalline structure to a certainextent.

Subsequently, neat BCNW and PGMA-grafted BNCWwere incorporated in a PLA matrix to prepare nanocompos-ites by solution casting followed by compression molding.

PGMA grafting resulted in enhanced matrix-filler adhe-sion and in a higher level of nanofiller dispersion, as re-vealed by morphological characterization. However,

M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072 2071

NO

LOG

Y

aggregation of BCNW took place to a certain extent, espe-cially when the loading was increased above 3 wt.%.

Incorporation of cellulose nanowhiskers significantlyimproved the oxygen barrier property of PLA at 0%RH evenin cases where BCNW were agglomerated. However, as aconsequence of the highly hydrophilic character of cellu-lose, when increasing the relative humidity up to 80%, low-er reductions in the oxygen permeability were attained,with optimum values corresponding to low nanofiller load-ings. Although the incorporation of BCNW led to nanocom-posites with slightly more hydrophilic surfaces ascompared to PLA, PGMA grafting did appear to limit thedetrimental effect of neat BCNW on the oxygen permeabil-ity at 80%RH.

The incorporation of both modified and unmodifiedBCNW resulted in stiffer materials. With the addition of3 wt.% PGMA-grafted BCNW, corresponding to the percola-tion threshold, the Young’s modulus and tensile strengthincreased by ca. 16% and 22%, respectively. Moreover, asa result of the improved matrix-filler adhesion, the ductil-ity of nanocomposites with PGMA-grafted BCNW was notreduced.

Thus, PGMA-grafting of BNCW improved their adhesionand dispersion in a hydrophobic PLA matrix, leading to im-proved barrier and mechanical properties compared to PLAnanocomposites loaded with neat BCNW.

ROM

OLE

CULA

RN

AN

OTE

CH

Acknowledgements

M. Martínez-Sanz would like to thank the Spanish Min-istry of Education for the FPU grant. The authors acknowl-edge financial support from MICINN (MAT2009-14533-C02-01 and MAT2012-38947-C02-01 projects) and the EUFP7 ECOBIOCAP project. Collaboration with researchers atthe Bioproduct Chemistry & Engineering Research Labora-tory of the WRRC-USDA in Albany, CA and the University ofPisa, Pisa, Italy is gratefully acknowledged. Part of the workwas conducted at the USDA laboratory in Albany, CA. A. Lo-pez-Rubio is the recipient of a ‘‘Ramon y Cajal’’ contractfrom the Spanish Ministry of Science and Innovation.

MA

C

References

[1] Auras R, Harte B, Selke S. An overview of polylactides as packagingmaterials. Macromol Biosci 2004;4(9):835–64.

[2] Drumright RE, Gruber PR, Henton DE. Polylactic acid technology. AdvMater 2000;12(23):1841–6.

[3] Auras R, Harte B, Selke S. Effect of water on the oxygen barrierproperties of poly(ethylene terephthalate) and polylactide films. JAppl Polym Sci 2004;92(3):1790–803.

[4] Bondeson D, Oksman K. Polylactic acid/cellulose whiskernanocomposites modified by polyvinyl alcohol. Compos Part A-Appl S 2007;38(12):2486–92.

[5] Chang JH, An YU, Sur GS. Poly(lactic acid) nanocomposites withvarious organoclays. I. Thermomechanical properties, morphologyand gas permeability. J Polym Sci Pol Phys 2003;41:94–103.

[6] Di Y, Iannace S, Maio ED, Nicolais L. Poly(lactic acid)/organoclaynanocomposites: thermal, rheological properties and foamprocessing. J Polym Sci Pol Phys 2005;43(6):689–98.

[7] Martínez-Sanz M, Lopez-Rubio A, Lagaron JM. On the optimization ofthe dispersion of unmodified bacterial cellulose nanowhiskersinto polylactide via melt compounding to significantly enhancebarrier and mechanical properties. Biomacromolecules 2012;13(11):3887–99.

[8] Oksman K, Mathew AP, Bondeson D, Kvien I. Manufacturing processof cellulose whiskers/polylactic acid nanocomposites. Compos SciTechnol 2006;66:2776.

[9] Paul MA, Alexandre M, Degee P, Henrist C, Rulmont A, Dubois P. Newnanocomposites materials based on plasticized poly(L-lactide) andorgano-modified montmorillonites: thermal and morphologicalstudy. Polymer 2003;44:443–50.

[10] Sanchez-Garcia MD, Gimenez E, Lagaron JM. On the use of plantcellulose nanowhiskers to enhance the barrier properties ofpolylactic acid. Cellulose 2010;17:987–1004.

[11] Habibi Y, Lucia LA, Rojas OJ. Cellulose nanocrystals: chemistry, self-assembly, and applications. Chem Rev 2010;110(6):3479–500.

[12] Martínez-Sanz M, Lopez-Rubio A, Lagaron JM. Optimization of thenanofabrication by acid hydrolysis of bacterial cellulosenanowhiskers. Carbohyd Polym 2011;85(1):228–36.

[13] De Souza Lima MM, Borsali R. Rodlike cellulose microcrystals:structure, properties, and applications. Macromol Rapid Comm2004;25(7):771–87.

[14] Hirai A, Inui O, Horii F, Tsuji M. Phase separation behavior inaqueous suspensions of bacterial cellulose nanocrystals prepared bysulfuric acid treatment. Langmuir 2009;25(1):497–502.

[15] Goffin AL, Raquez JM, Duquesne E, Siqueira G, Habibi Y, Dufresne A,et al. From interfacial ring-opening polymerization to meltprocessing of cellulose nanowhisker-filled polylactide-basednanocomposites. Biomacromolecules 2011;12(7):2456–65.

[16] Sassi JF, Chanzy H. Ultrastructural aspects of the acetylation ofcellulose. Cellulose 1995;2(2):111–27.

[17] Tomé LC, Pinto RJB, Trovatti E, Freire CSR, Silvestre AJD, Neto CP,et al. Transparent bionanocomposites with improved propertiesprepared from acetylated bacterial cellulose and poly(lactic acid)through a simple approach. Green Chem 2011;13(2):419–27.

[18] Cao X, Ding B, Yu J, Al-Deyab SS. Cellulose nanowhiskers extractedfrom TEMPO-oxidized jute fibers. Carbohyd Polym 2012;90(2):1075–80.

[19] Montanari S, Roumani M, Heux L, Vignon MR. Topochemistry ofcarboxylated cellulose nanocrystals resulting from TEMPO-mediatedoxidation. Macromolecules 2005;38(5):1665–71.

[20] Goussé, Chanzy H, Excoffier G, Soubeyrand L, Fleury E. Stablesuspensions of partially silylated cellulose whiskers dispersed inorganic solvents. Polymer 2002;43(9):2645–51.

[21] Pei A, Zhou Q, Berglund LA. Functionalized cellulose nanocrystals asbiobased nucleation agents in poly(l-lactide) (PLLA) – crystallizationand mechanical property effects. Compos Sci Technol 2010;70(5):815–21.

[22] Raquez JM, Murena Y, Goffin AL, Habibi Y, Ruelle B, DeBuyl F, et al.Surface-modification of cellulose nanowhiskers and their use asnanoreinforcers into polylactide: a sustainably-integrated approach.Compos Sci Technol 2012;72(5):544–9.

[23] Cao X, Habibi Y, Lucia LA. One-pot polymerization, surface grafting,and processing of waterborne polyurethane-cellulose nanocrystalnanocomposites. J Mater Chem 2009;19(38):7137–45.

[24] Goffin AL, Raquez JM, Duquesne E, Siqueira G, Habibi Y, Dufresne A,et al. Poly(e-caprolactone) based nanocomposites reinforced bysurface-grafted cellulose nanowhiskers via extrusion processing:morphology, rheology, and thermo-mechanical properties. Polymer2011;52(7):1532–8.

[25] Littunen K, Hippi U, Johansson L-S, Österberg M, Tammelin T, Laine J,et al. Free radical graft copolymerization of nanofibrillated cellulosewith acrylic monomers. Carbohyd Polym 2011;84(3):1039–47.

[26] Shukla SR, Athalye AR. Mechanical and thermal properties ofglycidyl methacrylate grafted cotton cellulose. J Appl Polym Sci1995;57(8):983–8.

[27] Lin N, Huang J, Chang PR, Feng J, Yu J. Surface acetylation of cellulosenanocrystal and its reinforcing function in poly(lactic acid).Carbohyd Polym 2011;83(4):1834–42.

[28] Martínez-Sanz M, Olsson RT, Lopez-Rubio A, Lagaron JM.Development of electrospun EVOH fibres reinforced with bacterialcellulose nanowhiskers. Part I: characterization and methodoptimization. Cellulose 2011;18(2):335–47.

[29] Liu X, Dever M, Fair N, Benson RS. Thermal and mechanicalproperties of poly(lactic acid) and poly(ethylene/butylenesuccinate) blends. J Environ Polym Degr 1997;5(4):225–35.

[30] Mishra A, Srinivasan R, Gupta R. Ppsyllium-g-polyacrylonitrile:synthesis and characterization. Colloid Polym Sci 2003;281(2):187–9.

[31] Stenstad P, Andresen M, Tanem B, Stenius P. Chemical surfacemodifications of microfibrillated cellulose. Cellulose 2008;15(1):35–45.

2072 M. Martínez-Sanz et al. / European Polymer Journal 49 (2013) 2062–2072

MA

CRO

MO

LECU

LAR

NA

NO

TECH

NO

LOG

Y

[32] Carrillo F, Colom X, Suñol JJ, Saurina J. Structural FTIR analysis andthermal characterisation of lyocell and viscose-type fibres. EurPolym J 2004;40(9):2229–34.

[33] Nelson ML, O’Connor RT. Relation of certain infrared bands tocellulose crystallinity and crystal lattice type. Part I. Spectra of latticetypes I, II, III and amorphous cellulose. J Appl Polym Sci 1964;8:1311–24.

[34] Andresen M, Johansson LS, Tanem BS, Stenius P. Properties andcharacterization of hydrophobized microfibrillated cellulose.Cellulose 2006;13(6):665–77.

[35] Ifuku S, Nogi M, Abe K, Handa K, Nakatsubo F, Yano H. Surfacemodification of bacterial cellulose nanofibers for propertyenhancement of optically transparent composites: dependence onacetyl-group DS. Biomacromolecules 2007;8(6):1973–8.

[36] Zulfiqar S, Zulfiqar M, Nawaz M, McNeill IC, Gorman JG. Thermaldegradation of poly(glycidyl methacrylate). Polym Degrad Stab1990;30(2):195–203.

[37] Sanchez-Garcia MD, Lagaron JM. On the use of plant cellulosenanowhiskers to enhance the barrier properties of polylactic acid.Cellulose 2010;17(5):987–1004.

[38] Yu L, Dean K, Li L. Polymer blends and composites from renewableresources. Progr Polym Sci 2006;31(6):576–602.

[39] Cava D, Giménez E, Gavara R, Lagaron JM. Comparative performanceand barrier properties of biodegradable thermoplastics and

nanobiocomposites versus PET for food packaging applications. JPlast Film Sheet 2006;22(4):265–74.

[40] Sanchez-Garcia MD, Nordqvist D, Hedenqvist M, Lagaron JM.Incorporating amylopectin in poly(lactic acid) by melt blendingusing poly(ethylene-co-vinyl alcohol) as a thermoplastic carrier. II.Physical properties. J Appl Polym Sci 2011;119(6):3708–16.

[41] Gilliland JW, Yokoyama K, Yip WT. Comparative study of guestcharge–charge interactions within silica sol–gel. J Phys Chem B2005;109(11):4816–23.

[42] George J, Ramana KV, Bawa AS, Siddaramaiah. Bacterial cellulosenanocrystals exhibiting high thermal stability and their polymernanocomposites. Int J Biol Macromol 2011;48(1):50–7.

[43] Favier V. PhD thesis, Etude de nouveaux matériaux compositesobtenus à partir de latex filmogènes et de whiskers de cellulose:effet de percolation mécanique. Joseph Fourier University, Grenoble,France; 1995.

[44] Sánchez-García MD, Hilliou L, Lagarón JM. Morphology and waterbarrier properties of nanobiocomposites of j/l-hybrid carrageenanand cellulose nanowhiskers. J Agr Food Chem 2010;58(24):12847–57.

[45] Martínez-Sanz M, Lopez-Rubio A, Lagaron JM. Nanocomposites ofethylene vinyl alcohol copolymer with thermally resistant cellulosenanowhiskers by melt compounding (II): water barrier andmechanical properties. J Appl Polym Sci 2013;128(3):2197–207.