8/20/2019 1-s2.0-S0141391014001955-main

1/11

PLA-PHB/cellulose based  lms: Mechanical, barrier and disintegrationproperties

M.P. Arrieta  a , e, E. Fortunati  b ,  *, F. Dominici  b , E. Rayon  c, J. L opez  a , J.M. Kenny  b , d

a Instituto de Tecnolog í a de Materiales, Universitat Polit ecnica de Valencia, 03801 Alcoy, Alicante, Spainb Materials Engineering Centre, UdR INSTM, NIPLAB, University of Perugia, 05100 Terni, Italyc Instituto de Tecnolog í a de Materiales, Universitat Polit ecnica de Valencia, E-46022 Valencia, Spaind Institute of Polymer Science and Technology, CSIC, Juan de la Cierva 3, Madrid 28006, Spaine  Analytical Chemistry, Nutrition and Food Sciences Department, University of Alicante, P.O. Box 99, E-03080 Alicante, Spain

a r t i c l e i n f o

 Article history:

Received 11 February 2014Received in revised form29 April 2014Accepted 2 May 2014Available online 29 May 2014

Keywords:

Poly(lactic acid)Poly(hydroxybutyrate)Modied cellulose nanocrystalsNanocompositesBiodegradationBarrier properties

a b s t r a c t

Nanocomposite   lms based on poly(lactic acid)-poly(hydroxybutyrate) (PLA-PHB) blends and synthe-sized cellulose nanocrystals (CNC) or surfactant modied cellulose nanocrystals (CNCs), as bio-basedreinforcement, were prepared by melt extrusion followed by   lm forming. The obtained nano-composites are intended for short-term food packaging. Thus, the mechanical, optical, barrier andwettability properties were studied. Functionalized CNCs contribute to enhance the interfacial adhesionbetween PLA and PHB, leading to improved mechanical stiffness and increased  lm stretchability. Thesynergic effects of the PHB and CNCs on the PLA barrier properties were conrmed by increases inoxygen barrier properties and reductions in surface wettability of the nanocomposites. In addition, themeasurements of the viscosity molecular weight for ternary systems showed practically no degradationof PLA and smaller degradation of PHB during processing due to nanocrystal presence. The disintegrationprocess in composting conditions of PLA was delayed by the addition of PHB, while CNC speeded it up.PLA-PHB-CNCs formulations showed enhanced mechanical performance, improved water resistance,

reduced oxygen and UV-light transmission, as well as appropriate disintegration in compost suggestingpossible applications as packaging materials.© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Many positive characteristics of PLA have situated it as the mostused biopolyester for biodegradable food packaging industry.Among other properties, PLA compared to other biopolymersshows easy processability [1], superior transparency, availability inthe market [2], excellent printability [3], high rate of disintegrationin compost [4]. Conversely, the use of PLA  lms for food packaging

has been strongly limited because of their poor mechanical andbarrier properties [5]. Moreover, for large-scale industrial produc-tion PLA must guarantee adequate thermal stability or low thermaldegradation during processing and use [6].

Initially, the materials for food applications were semi-rigid orexible monolayer systems. Afterward, to improve the barrierproperties they were replaced by more complex multilayer sys-tems, which are still used in the market. However, its dif culty of 

recycling and the worldwide trend to reduce the polymer con-sumption per package unit, demand the development of simplestpackaging formulations particularly focused on blending strategies[7]. Thus, melt blending PLA with another biopolymer can lead tosignicant improvement of the   nal properties through a costeffective, easy and readily available processing technology [8].

Food packaging, besides containment and information, shouldprovide foodstuff protection against water, light or oxidative pro-

cess [9]. It is known that the crystalline phase has an importantimpact on mechanical and permeation properties; as a result,considerable academic and industrial research efforts have beenfocused to increase PLA crystallinity. In this sense, the addition of poly(hydroxybutyrate) (PHB), a highly crystalline biopolymer, tothe PLA matrix by melt blending has been considered as an easyway to increase PLA crystallinity and regulate its properties  [10].PHB, the most common representative of poly(hydroxyalkanoates)(PHA), with a high degree of crystallinity, has been also proposedfor short-term food packaging applications [11]. PHB has a similarmelting temperature to PLA, allowing blending both polymers inthe melt state.In a previouswork, PLA was melt blended with 25 wt

*  Corresponding author. Tel.:  þ39 0744492921; fax:  þ39 0744492950.E-mail address: elena.fortunati@unipg.it (E. Fortunati).

Contents lists available at ScienceDirect

Polymer Degradation and Stability

j o u r n a l h o m e p a g e :   w w w . e l s e v i e r . c o m/ l o c a t e / p o l y d e g s ta b

http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010

0141-3910/©

 2014 Elsevier Ltd. All rights reserved.

Polymer Degradation and Stability 107 (2014) 139e149

mailto:elena.fortunati@unipg.ithttp://www.sciencedirect.com/science/journal/01413910http://www.elsevier.com/locate/polydegstabhttp://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://dx.doi.org/10.1016/j.polymdegradstab.2014.05.010http://www.elsevier.com/locate/polydegstabhttp://www.sciencedirect.com/science/journal/01413910http://crossmark.crossref.org/dialog/?doi=10.1016/j.polymdegradstab.2014.05.010&domain=pdfmailto:elena.fortunati@unipg.it

8/20/2019 1-s2.0-S0141391014001955-main

2/11

% of PHB showing an improvement in oxygen barrier and waterresistant, whilst reducing the inherent high transparency of PLA[8,12]. Transparency is an essential issue to be considered in thedevelopment of materials intended for food packaging since seeingthrough the packaging is one of the most important requirementsfor the consumers [2]. Moreover, to preserve food products untilthey reach the consumer, packaging sometimes require protectingfood products from ultraviolet light   [13]. Thus, the design of transparent films with enhanced UV protection is particularlyrelevant in many packaging applications [14]. It has been reportedthat PHB acts as better light barrier in the visible and ultravioletlight regions [11] than PLA.

On the other hand, novel and ef cient polymer materials basedon nanotechnology can provide innovative solutions to increase thepolymers performance for food packaging. The incorporation of nanoparticles into polymers matrices has shown to improve me-chanical properties and thermal stability, as well as it offers manyadditionally advantages such as reduction in raw materials andelimination of expensive lamination secondaryprocesses, while themodication of the thermal properties of polymers leads to areduction in machine cycle time and temperature [15].

The use of nanocomposites to improve the inherent shortcom-

ings of PLA based packaging materials has proven to be a promisingtechnology. The ideal nanoparticle should be biobased and biode-gradable. In this sense, bioresources obtained from agricultural-related industries have received signicant attention, particularlyfocused on cellulosic materials and especially to its specic form of cellulose nanocrystals (CNC), which have been revealed to be aninteresting model   ller   [16]   for various biopolymer matricesincluding PLA [3,17,18]  and PHB   [19], beside others biopolymerssuch as poly(vinyl alcohol) (PVA) [16] and poly(hydroxybutyrate-co-hydroxyvalerate) (PHBV)   [20]. Cellulose nanocrystals haveshown better mechanical properties than a majority of thecommonly used reinforcing materials and offer additional excep-tional advantages such as biodegradability, high stiffness and lowdensity [17] abundance in nature and low cost [21].

Although nanocelluloses have a great potential as mentionedabove, the high amount of   eOH on the surface of the crystals in-duces high attraction between them [22]. Thus, the high polarity of cellulose surface and the resultant low interfacial compatibilitywith hydrophobic polymer matrices make dif cult the homoge-nous dispersion of nanocellulose in polymers [23]. For that reason,a surface modication of CNC by a surfactant (CNCs) has beenproposed and successful dispersion of CNCs in the PLA matrix wasachieved [3,17]. This specic type of modication enhances theinterfacial adhesion polymer/nanoller and thus improves somenal properties of the   nal nanocomposites such as mechanicalperformance [17], oxygen barrier and water resistance [24], whichare particularly interesting for materials intended for foodpackaging.

In a previous work, the processing performance of PLA-PHBwith CNC or CNCs was optimized and it was veried that thefunctionalization of CNCs favours the dispersion into PLA-PHBblend matrix enhancing the interfacial adhesion by meansincreasing the thermal stability [25]. In the current work, in orderto address the positive effect of cellulose nanocrystals on thethermal stability of PLA and PHB, a more detailed study concerningthe potential reduction of their molecular weight due thermalprocessing is reported. Moreover, since the main objective of thisresearch is to propose this high performance nanocomposite  lmsfor biodegradable food packaging industry a complete character-ization related with thiseld of application was conducted. For thispurpose, the combination of ternary system based in PLA, PHB andcellulose nanocrystals blend was developed to enhanced PLA bar-

rier and mechanical properties. Cellulose nanocrystals (CNC) were

synthesized from microcrystalline cellulose (MCC) as well asfurther modied using a surfactant (CNCs) to improve the disper-sion in the biopolymer matrix. Then nanocrystals were melt-blended with a previous prepared PLA-PHB masterbatch andnally processed into  lms. The processing of these systems andtheir crystalline and thermal stability properties were reportedpreviously   [25]. The mechanical, optical and barrier propertieswere tested with the aim to evaluate their suitability for the foodpackaging sector and are reported here. Additionally, the dis-integrability under composting of the multifunctional materialswas evaluated to get information about their post-use.

2. Experimental

 2.1. Materials

Poly(lactic acid) (PLA 2002D,  M n ¼ 98,000 g mol1, 4 wt%  D-iso-

mer) was supplied by NatureWorks (USA). Poly(hydroxybutyrate)(PHB, under the trade name PHI002) was acquiredfrom NaturePlast(France) and microcrystalline cellulose (MCC, dimensions of 10e15 mm) was purchased from SigmaeAldrich.

 2.2. Nanocrystal synthesis and modi cation

Acid hydrolysis of microcrystalline cellulose (MCC) was carriedout by using sulphuric acid 64% (wt/wt) at 45   C for 30 min withcontinuous stirring [17]. The obtained cellulose nanocrystals (CNC)in an acid solution were washed with ultrapure water (1:200),centrifugated and dialyzed until neutral pH. An ion exchange resinwas added to the cellulose suspension for 24 h and then wasremoved by   ltration in order to ensure that all ionic materialswere removed except the Hþ counter ions associated with thesulphate groups on the CNC surfaces. After that, nanocrystal sus-pensions were ultrasonicated (Vibracell 75043, 750 W, BioblockScien-tic) for 2 min in an ice bath. Surface modied cellulosenanocrystals (CNCs) were also prepared by adding a surfactant

(STEFAC TM 8170, Stepan Company Northeld) in 1/1 (wt/wt).Finally, cellulose nanocrystals in powder were obtained by a freeze-drying process of previously neutralized solutions (1.0% (wt/wt) of 0.25 mol l1 NaOH).

 2.3. PLA-PHB-nanocomposite preparation

PLA (75 wt%) was blended with 25 wt% of PHB and then rein-forced with 5wt% of pristine (CNC) or surfactant modied (CNCs)cellulose nanocrystals. Masterbatches were prepared by using atwin-screw microextruder (DSM explorer 5&15 CC MicroCompounder) by following the same processing conditions asdescribed in a previous work   [25]. Briey, using a temperatureprole of mixing process with a maximum temperature of 200   C

with three-step temperature procedure of 180e

190e

200 

C and ascrew speed of 150 rpm for 2 min. Masterbatches were pelletizedand mixed for 1 min and directly processed in  lms with a headforce of 3000N. Then, a  lm procedure was conducted to obtain sixformulations, including the neat PLA and PLA-PHB blend with athickness ranged from 10 to 30 mm. The obtained formulations andthe proportion of each component are summarized in Table 1.

 2.4. Characterization techniques

The capillary viscosity was measured at roomtemperature usinga Ubbelohde viscometer (type 1C) according to ISO 1628 [26] for alllm sample diluted in chloroform (SigmaeAldrich 99% purity) andat least three concentrations were used. PLA and PHB pellet were

also measured as control. The concentration of PLA and/or PHB in

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149140

8/20/2019 1-s2.0-S0141391014001955-main

3/11

the  nal  lm formulations were expressed taken into account theproportion reported in Table 1. The intrinsic viscosity [h] of sampleswas determined to estimate the viscosity molecular weight bymeans the MarkeHouwink relation:

½h ¼ K   M aV    (1)

were K  and a are 1.53  102 and 0.759 for PLA, respectively [27] aswell as K  and a for PHB are 1.18  102 and 0.780 in that order [28].

The mechanical behaviour was investigated by tensile test in a

digital Lloyd instrument LR 30K, performed on rectangular probes(100 mm  10 mm) at room temperature by following the UNE-ENISO 527-3 standard [29] with a crosshead speed of 5 mm/min, aload cell of 500N and an initial gauge length of 50 mm. Averagetensile strength (TS ), percentage elongation at break (εb   %) andYoung's modulus (E ) were calculated from the resulting stresse-strain curves as the average of   ve measurements of eachcomposition.

The absorption spectra of nanocomposites, obtained in the700e250 nm region, were investigated by a PerkineElmer (Lambda35, USA) UV eVIS spectrophotometer. Nanocomposite   lm colourproperties were evaluated in the CIELAB colour space by using aKONICA CM-3600d COLORFLEX-DIFF2, HunterLab, Hunter Associ-ates Laboratory, Inc, (Reston, Virginia, USA). The instrument was

calibrated with a white standard tile. Yellowness index (YI ) andcolour coordinates,   L  (lightness), a* (red-green) and b* (yellow-blue) were measured at random positions over the   lm surface.Average values of  ve measurements were calculated. Total colourdifference (DE ) was calculated with respect to the control pure PLAlm or PLA-PHB  lm as:

DE  ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Da*2 þ Db*2 þDL*2

p   (2)

The oxygen transmission rate (OTR) was measured to studythe oxygen permeability of the nanocomposites by using a Sys-tech Instruments 8500 oxygen permeation analyzer (MetrotecS.A, Spain) at room temperature and 2.5 atm. 14 cm diametercircle   lms were compressed between the upper and lower

diffusion chamber. Pure oxygen (99.9% purity) was introducedinto the upper half of the sample chamber while nitrogen wasinjected into the lower half. To prepare the appropriate samplesfor OTR measurements, masterbatch pellets were set in discs byusing a DSM Xplore 10-ml injection moulding machine at 175, 180and 190   C with a pressure prole in three steps: 6 bar for 4 min,8 bar for 5 min and 8 bar for 3 min. Discs were then processedinto   lms by compression moulding process at 180   C in a hotpress (Mini C 3850, Caver, Inc., Wabash, IN, USA) with a pressurecycle of 3 MPa for 1 min, 5 MPa for 1 min, and 10 MPa for 2 min.Nanocomposite   lms were then quenched to room temperatureat atmospheric pressure. Their average thickness was between180 and 250  mm.

Surface wettability of   lms was studied through static water

contact angle measurements with a standard goniometer

(EasyDrop-FM140, KRÜSS GmbH, Hamburg, Germany) equippedwith a camera and Drop Shape Analysis SW21; DSA1 software wasused to test the water contact angle (q) at room temperature. Thecontact angle was determined by randomly putting 5 drops of distilled water (z2  mL) with a syringe onto the  lm surfaces and,after 30 s, the average values of ten measurements for each dropwere used. The maximum standard deviation in the water contactangle measurements did not exceed  ±3% [30].

 2.5. Disintegrability under composting conditions

The disintegration under composting conditions of PLA andPLA-PHB nanocomposites was investigated on the basis of the ISO20200 standard  [31]. A solid synthetic waste was prepared bymixing 10% of compost supplied by Gesenu S.p.a. (Perugia, Italy),with 30% rabbit food, 10% starch, 5% sugar, 1% urea, 4% corn oil and40% sawdust.The watercontent of the substrate wasaround 50 wt%and the aerobic conditions were guaranteed by mixing it softly [4].Nanocomposite   lms (cut in 15    15 mm2) were weighed andburied at 4e6 cm depth in perforated plastic boxes, containing theprepared mix, and incubatedat 58 C. Each nanocompositelmwasrecovered at 1, 2, 3, 7, 10, 14 and 21 days of disintegration, cleanedwith distilled water, dried in an oven at 37   C during 24 h andreweighed. The disintegration degree was calculated by normal-

izing the sample weight, at different days of incubation, to theinitial weight. In order to determine the time at which 50% of eachlm was degraded, disintegrability degree values were then  ttedusing the Boltzmann equation (OriginPro 8.1.software) as follows:

m ¼  ðmi  m∞Þ

1 þ eð1ðt 50=dt ÞÞ  (3)

where  mi  and  m∞ are the initial and   nal mass values measuredrespectivelyat the beginning of the exposition to compost and afterthe  nal asymptotes of the disintegrability test, and  t 50 is the timeat which materials disintegrability reaches the average value be-tween mi and  m∞, known as the half-maximal degradation,  dt is aparameter that describes the shape of the curve between the upper

and lower asymptotes [32].Photographs of recovered samples were taken for visual com-parison. Surface microstructure of PLA and PLA-PHB nano-composites before and after 3 days of incubation in compostingwere studied by optical microscopy using a LV-100 Nikon Eclipseequipped with a Nikon sight camera at 20  magnications usingthe extended depth of  eld (EDF-z) imaging technique to obtain atridimensional vision of   lms surfaces. This technique uses amotorized  z -axis (height of focus) to take images at different heightplanes. Subsequently, by means of a dedicated algorithm installedin the NIS-Elements software a 3D image is reconstructed followingthe original texture of sample.

The relationship between  meso-lactide and   L ,D-lactide form inthe polymer after compost incubation was also studied by

Pyrolysis-Gas Chromatography/Mass Spectrometry (Py-GC/MS) by

 Table 1

PLA and PLA-PHB nanocomposite  lm formulations and their viscosity molecular weight (M v).

Film formulations Materials (wt %) PLA M v (g mol1) PHB  M v  (g mol

1)

PLA PHB CNC CNCs

PLA 100   e e e   90,600  ±  10,700   ePLA-CNC 95   e   5   e   86,200  ±  18,800   ePLA-CNCs 95   e e   5 81,100  ±  14,400   ePLA-PHB 75 25   e e   91,900  ±  19,000 224,300  ±  45,150PLA-PHB-CNC 71.25 23.75 5   e   95,000  ±  19,500 231,700  ±  46,400PLA-PHB-CNCs 71.25 23.75   e   5 95,600  ±  18,800 233,000  ±  44,700

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149   141

8/20/2019 1-s2.0-S0141391014001955-main

4/11

means of a Pyroprobe 1000 pyrolyzer (CDS Analytical, Oxford,Pennsylvania, USA) at 1000   C for 0.5 s, coupled with a gas chro-matograph (6890N, Agilent Technologies) and a mass selectivedetector (Agilent 5973N) on the basis of a previous developedmethod [32].

Fourier infrared spectra of the samples in the 400e4000 cm1

range were recorded by a Jasco FTIR 615 spectrometer, in trans-mission mode.

 2.6. Statistical analysis

Signicance in the mechanical, wettability and colour parame-ters differences were statistically by one-way analysis of variance(ANOVA) using OriginPro 8 software. To identify which groups weresignicantly different from other groups, means comparison weredone employing a Tukey's test with a 95% condence level.

3. Results and discussion

 3.1. Viscosity molecular weight 

The estimated viscosity molecular weight (M v) of PLA and PHBpellets were 95,800 ± 4400 g mol1and 255,300 ± 39,000 g mol1,respectively. It is know that polymers can undergo thermaldegradation during processing and diminution of  M v values for alllm formulations with respect of PLA pellet and/or PHB pellet weredetected (Table 1). PLAprocessed into lm resulted in a reduction of the M v of PLA around 5%. Further decrease occurred in binary PLAbased nanocomposites. While, a reduction of PLA  M v value of 10%was detected for PLA-CNC, PLA-CNCs showed a higher reduction of 15%. This results is in accordance with previous work where weshowed that the thermal stability of PLA was reduced with theaddition of cellulose nanocrystals, particularly for PLA-CNCs  [25].

The addition of PHB also produced a reduction on the M v of PLA of about 4% and a higher reduction (11%) was showed, as expected, inthe case of the  M v  value of PHB. This higher reduction in the  M vvalue of PHB than in the   M v   value of PLA can be explained bysomewhat PHB thermal degradation that might take place at theprocessing temperature used here, in the range of 180e200 C [33].The viscosity molecular weight of PLA in ternary PLA-PHB basednanocomposites was almost constant. Meanwhile, the viscositymolecular weight of PHBwas reduced by approximately 9% in theseformulations (PLA-PHB-CNC and PLA-PHB-CNCs). The addition of cellulose nanocrystals into PLA-PHB blends lead to a enhancementthe interface interaction between PLA and PHB leading to animprovement of the thermal stability of both polymers, particularlyfor PHB that usually shows a small processing window [25]. Thussmaller reductions on PHB Mv  values were observed after pro-cessing due the positive effect of CNC and CNCs presence.

 3.2. Mechanical properties

As it can be seen from tensile curves (Fig. 1-a), the neat PLAlm showed a characteristic plastic deformation that it was

reduced with both, PHB and CNC incorporation. CNC and PHBproved to be effective to increase PLA modulus (Fig. 1-b), but nosignicant differences were observed between the Young'smodulus of PLA-CNCs and PLA  lms. While CNCs or PHB produceda decrease on the tensile strength (TS ) of PLA, the combination of PHB and CNCs produce a nanocomposite (PLA-PHB-CNCs) withcomparable TS  with respect to PLA. This behaviour can be relatedwith the more ef cient dispersion of functionalized cellulosenanocrystals (CNCs)   [3]   resulting in an enhancement in theinterfacial adhesion and therefore in a better interaction betweenPLA and PHB [25]. Moreover, the PLA-PHB-CNCs  lm revealed thehighest deformation at break, showing an increase of 175% with

Fig. 1.   Tensile test results of PLA, PLA-PHB and nanocomposite  lms: a) Stress-strain curves, b) Young 's Modulus (E ), c) Tensile strength (TS ) and d) Elongation and break (εB).aed

Different letters on the bars within the same image indicate signi

cant differences between formulations ( p <

 0.05).

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149142

8/20/2019 1-s2.0-S0141391014001955-main

5/11

respect to the neat PLA  lm, while signicant lower elongations atbreak respect to the PLA  lm were measured for the other nano-composite formulations. Films for food packaging are required tomaintain their integrity in order to withstand the stress that oc-curs during shipping, handling and storage  [34]   and PLA-PHB-CNCs is therefore far stretchable and stiff with comparablestrength than PLA and could be dened as the best formulation forfood packaging applications.

 3.3. Optical and colourimetric properties

The absorption spectra of PLA, PLA-PHB and nanocomposites are

shown in Fig. 2(a) while the visual appearances of the 

lms aredisplayed in Fig. 2(b). Neat PLA  lm proved to be the most trans-parent showing the highest transmission in the visible region of the

spectra (400e700 nm). No signicant changes were observed dueto the presence of CNC or CNCs in the visible region of the spectra,thus PLA-CNC and PLA-CNCs resulted in highly transparent  lms,referred to the light transmission in the range of 540 e560 nm[13,35], as it can be observed in Fig. 2(b). The good transparency of PLA-CNC  lms has been related with the good dispersion of cellu-lose nanocrystals into PLA matrix  [17,35]. On the other hand, a25 wt% of PHB, provokes a reduction of the light transmission of thelms. Both, PHB and cellulose nanocrystals show a blocking effecton the virtually transparent PLA matrix at the UV spectra region(250e400 nm). Cellulose nanocrystals reduced the UV light trans-mission with a maximum centered at 275 nm which corresponds tothe UV-C region (280e100 nm), generally created from articiallight sources [13]. This behaviour was more evident with surfacemodied nanocrystals (CNCs). The PLA-PHB-CNCs   lm showed ablocking effect in the UV light spectra region with the lowest UV-Clight transmission, while maintaining the high transparency in thevisible spectra region.

Table 2  summarizes the colour parameters obtained for PLA,PLA-PHB and cellulose nanocrystal based nanocomposites. PLAshowed the highest   L   value conrming it characteristic highbrightness. L  is signicantly affected by CNC, CNCs and PHB pres-

ence, although all   lm samples present still higher  L  values thancommercial low density polyethylene (LDPE) and poly(ethyleneterephthalate) (PET) lms [13]. Negative values of the a* coordinatereveal a deviation towards green, while positive values for  b* areindicative of a deviation towards yellow. As a consequence of thePHB presence, the highest deviations towards green and yellowcolours were observed in the PLA-PHB blend, in accordance toprevious studies [36]. As a result, the yellowness index (YI ) showedthe maximum value for the PLA-PHB  lm, followed by PLA-PHB-CNCs and PLA-PHB-CNC. The  YI  is used to describe the change incolour of a sample from clear toward yellow. It must be noticed,that the obtained YI  values are signicant lower than those previ-ously reported for PLA [12,13] and PLA-PHB (75:25) blends [12,36].The main reason for this important reduction in  YI   is due to the

lower lms thickness obtained by the processing lm methodologyused in the present work. Despite the total colour differences ob-tainedwith respect to the neat PLAlm were signicant different inall cases, the total colour differences were in general smaller than2.0, being this value the threshold of perceptible colour differencefor the human eye [37], with the exception of the PLA-PHB  lm ascan be conrmed in Fig. 2b.

 3.4. Oxygen transmission rate and wettability

In a previous reported work the incorporation of CNC andfunctionalized CNCs had shown reductions in OTR values of neatPLA   lm (30.5 cm3 * mm * m2 * day1) of about 43% and 48%,

respectively   [24]. In this case, the addition of PHB reduced theoxygen permeation of PLA to 13.3 cm3 * m m * m2 * day1

(reduction of 56%) due to the increased crystallinity in the system

 Table 2

Colour parameters from CIELab space and  YI  of PLA, PLA-PHB and nanocomposite  lms.

Samples   L   a* b*   DE    YI

PLA 94.64  ±  0.01a 0.98 ±  0.01a 0.76 ±  0.02a e   0.70  ±  0.03a

PLA-CNC 93.57  ±  0.01b 1.08 ±  0.02b 0.62 ±  0.02b 1.08  ±  0.01 0.52  ±  0.02b

PLA-CNCs 94.28  ±  0.01c 1.01 ±  0.01c 0.83 ±  0.01c 0.37  ±  0.01 0.82  ±  0.02c

PLA-PHB 93.79  ±  0.01d 1.07 ±  0.01b 1.54 ±  0.01d 1.15  ±  0.01 2.14  ±  0.02d

PLA-PHB-CNC 93.55  ±  0.01e 0.91 ±  0.01d 1.03 ±  0.01e 1.12  ±  0.01 1.29  ±  0.01e

PLA-PHB-CNCs 93.69  ±  0.01f  1.04 ±  0.01c 1.48 ±  0.01f  1.19  ±  0.01 2.07  ±  0.03f 

DE  Calculated by using PLA  lm colour coordinates as reference.aef 

Different superscripts within the same column indicate signi

cant differences between formulations ( p <

 0.05).

Fig. 2.  PLA, PLA-PHB and nanocomposite  lms: a) UV eVis spectra, b) Visual appear-

ance, c) contact angle measurements.  aedDifferent letters on the bars within the same

image indicate signicant differences between formulations ( p  <  0.05).

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149   143

8/20/2019 1-s2.0-S0141391014001955-main

6/11

[12,25]. Nevertheless, from the OTR values obtained in ternarysystems (15.3 cm3 * m m * m2 *day1 for PLA-PHB-CNC and 13 cm3

* mm * m2 * day1 for PLA-PHB-CNCs), it can be noticed that theincorporation of CNC or CNCs to PLA-PHB blend did not provokemajor changes in OTR values. The low OTR values obtained for thePLA-PHB blend and ternary nanocomposites highlight the advan-tage of blending PLA with crystalline PHB. Theselms are thereforeattractive for food packaging applications were barrier to oxygen iscritical to avoid or reduce oxidative processes.

Additionally,   lms for food packaging are required to protectfoodstuff from humidity during transport, handling and storage.Thus, water contact angle measurements were carried out toevaluate the hydrophilic/hydrophobic character of   lms and theresults are shown in  Fig. 2(c)   [37]. It should be noticed that allformulations showed values higher than 65, being materialsacceptable for the intended end-use applications. PHB has a hy-drophobic character dueto the poor af nity of the water tothe non-polar polymer surface [38]. In this way, the PLA-PHB blend showedsignicant increased water resistance in comparison with neat PLA,in good accordance with a previous reported work [12]. The pres-ence of CNC in PLA and PLA-PHB caused an increase in wettability,while functionalized CNCs did not signicantly change PLA or PLA-

PHB wettability. The positive effect of cellulose nonocrystalchemical modication in the wettability of PLA and PLA-PHB  lmsis mainly due to the presence of sulphate groups with low polarityon the surface that increase the surface hydrophobicity of the  nalmaterial.

 3.5. Disintegration under composting 

Fig. 3 (a)   shows the visual appearance of PLA, PLA-PHB andcellulose nanocrystal based nanocomposites after different time of disintegration in composting conditions where it is possible toconrm the biodegradable character of all the formulations stud-ied. After only 1 day of incubation,  lms become smaller, with theexception of PLA-PHB blend, which started the  lm size reductionon the second day of incubation. After 7 days of incubation binaryand ternary formulation  lms became breakable and small piecesof  lms were recovered. It also could be noticed that they changedtheir colour and became more opaque after 7 days. When thedegradation process of the polymer matrices started, a change inthe refraction index of the materials was observed as a result of water absorption and/or presence of products formed by the hy-drolytic process  [39]. Additionally, the   lms disintegrability wasevaluated in terms of mass loss as a function of incubation time(Fig. 3 (b)), in which the line at 90% of disintegration represents thegoal of disintegrability test [4]. Unmodied cellulose nanocrystals(CNC) speed up the disintegration of PLA and PLA-PHB blend from14 days to 21 days, respectively, to 10 days. Comparable   ndingswere previously reported for PLA nano-biocomposite   lms with

functionalized cellulose nanocrystals and silver nanoparticles [40].Accordingly, after 10 days CNC incorporated   lms were visiblydisintegrated (Fig. 3 (a)), while CNCs incorporated counterpartsreached between 50 and 60% of disintegrability and need 14 days toreach the goal of the disintegrability test (Fig. 3 (b)). It isknown that

Fig. 3.  a) Visual appearance of   lm samples before and after different incubation days under composting conditions. b) Degree of disintegration of   lms under composting

conditions as a function of time.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149144

8/20/2019 1-s2.0-S0141391014001955-main

7/11

the PLA disintegration in compost starts by a hydrolysis process[41], thus this different behaviour observed between the pristineand modied cellulose nanocrystals could be ascribed to the morehydrophobic character of functionalized cellulose nanocrystals thatprotect the polymer matrix from the water attack. In brief, PLA-CNCand PLA-PHB-CNC lost more than 90% of the initial matter in 10days; PLA, PLA-PHB-CNC and PLA-PHB-CNCin 14 days and PLA-PHBin 21 days. These short degradation times have been related to thelow thickness of tested samples  [4]. Some changes in compostcolour were observed (Fig. 3 (a)) due to the aerobic fermentationthat results in dark humus soil.

However,all formulationsshowed different rate of disintegrationand thusthe Boltzmannfunction was used to correlate the sigmoidalbehaviour of the mass loss during the disintegrability in the com-posting process (Fig. 4). The estimated regression parameters of thetted results of the non-linear model and t 50 were calculated, while

mi and m∞ values were assigned as 0% and 100% of disintegrability,respectively. The correlation coef cients between theoretical andexperimental data(R2) werehigher than 0.990 in allcases,indicatingthat only minor differences were observedin thetting of themodelto experimental values. The rate of disintegration under compostingconditions was longer for PLA-PHBwith a half-maximal degradation(t 50) at about 14 days, with respect to neat PLA that showed  t 50 atabout 10 days, due to the fact that the polymer disintegrability incomposting starts in the amorphous phase of the polymers [42] andthe increasing crystallinity in PLA-PHB blend due to the PHB pres-ence delays the PLA degradation rate [12,36].

Both cellulose nanocrystals speed up the rate of disintegration of PLA and PLA-PHB shifting half-maximal degradation to lowervalues. The t 50 of PLA was shifted from 10 days to 7 days in PLA-CNCand remains practically constant in PLA-CNCs, while the t 50 of PLA-PHB was shifted from 14 days to 7 days and to 10 days in PLA-PHB-

Fig. 4.  Disintegrability of PLA, PLA-PHB and nanocomposite 

lms as a function of time.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149   145

8/20/2019 1-s2.0-S0141391014001955-main

8/11

CNC and PLA-PHB-CNCs, respectively. As a result, PLA-PHB-CNCshowed higher rate of disintegration than PLA-CNCs, even whenPLA-PHB-CNC (cc   ¼   18.5%) is more crystalline than PLA-CNCs(cc  ¼  11.0%) [25]. This unexpected result could be explained withthe fact that during disintegration in compost, PLA surface is  rstlyattacked by water where polymer chain are hydrolysed [4] and as aresult the smaller molecules become susceptible for enzymaticdegradation mediated by microorganisms   [41]. Meanwhile, PHBdisintegration is rstly caused by polymer surface erosion mediatedby microorganisms which then are able to spread gradually insidethe polymer matrix   [43]. PLA-PHB-CNC showed higher surfacepolarity than PLA-CNCs. Thus, the water attack starts on the moresusceptible component, CNC, with hydroxyl groups available on thesurface, allowing the hydrolysis in PLA-PHB-CNC, which is followedby the microorganisms attack, while CNCs is protecting PLA in PLA-CNCs. In the meantime, available   eOH are now able to attack hecarbon of the ester group and produce intramolecular degradation,followed by the hydrolysis of the ester link [44]. As a consequence,the higher surface polarity of PLA-PHB-CNC and the eOH presencethat catalyse the hydrolysis process, leading to a higher disinte-gration rate for PLA-PHB-CNC than PLA-CNCs.

Micrograph observations of lm surfaces before and after 3 days

in composting, and their proles measured by the EDF-z technique,are shown in   Fig. 5. PLA and PLA nanocomposite   lms beforecomposting showed smoother prole than PLA-PHB counterparts.Similar behaviourwas observed in a previous work where neat PLA,neat PHB  lms roughness were investigated by means of confocalmicroscopy. The PLA   lm showed a smoother roughness prolethan PHB [12]. In general, after 3 days in composting all formula-tions showed a more irregular EDF-z prole with respect to thesame formulation before composting.

The chemical changes of nanocompositelmsbefore and after 1,2 and 3 days in composting were followed by FTIR analysis. At

higher time of disintegration,  lm samples could not be studied inthe FTIR spectrometer due the small portions of   lms recovered.The main differences found were in the 2000e1200 cm1 region of the FTIR spectra as shown in   Fig. 6. The typical asymmetricstretching of the carbonyl group (eC]O) of PLA centered at1760 cm1 become broader during composting due to an increasein the numberof carboxylic endgroups in the polymer chain duringthe hydrolytic degradation [39]. Moreover, a band at 1722 cm1 wasalso apparent (shownby grey arrows) that has been associated withcrystalline C]O stretching vibration of PHB [45]. In some samples,this band appears as a shoulder due to its low intensity and thestrong stretching vibration of carbonyl group. A small band at1687 cm1 was apparent (shown by black arrows) in PLA-PHB andPLA-PHB-CNCs after three days of composting. This band has beenreported to be a crystalline band, although its spectral origin is notyet assigned [45,46]. The clear appearance of this band in PLA-PHBand PLA-PHB-CNCs is supported by the early degradation of theamorphous phase of the polymer blend while the crystalline PHBremains in the polymer matrix. Similar results indicating that PHBslow down the disintegration rate of PLA in PLA-PHB blends havebeen previously reported  [36]. In the region between 1550 and1650 cm1 the appearance of a broad band was observed for all

formulations. The appearance of this band has been previouslyobserved during the degradation of PLA-MCC based compositesand wasrelated to the presence of carboxylate ions in degraded PLAcomposites [4].

Fig. 7(a)  shows the typical Py-GC/MS chromatogram of PLA-PHB-CNCs obtained by pyrolysing the   lm at 1000   C for 0.5 s.The pyrolysis of all PLA based lms is characterized by the presenceof two peaks with very similar mass spectra (m/ z  ¼  32, 43, 45 and56) in which the peak at 17 min corresponds to  meso-lactide andthe peak at 18 min with the highest signal intensity in all samples to(L) and/or (D)-lactide [32]. Films with PHB showed the broad peak

Fig. 5.  Optical micrographs (20

) of PLA, PLA-PHB and nanocomposite 

lms and their EDF-z pro

les.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149146

8/20/2019 1-s2.0-S0141391014001955-main

9/11

of crotonic acid (6.7 min, m/ z  ¼ 39, 41, 68, 69, and 86) [12,36], whilenanocomposte lms showed a peak at 22.5 min assigned tothermaldegradation products of the cellulose structure (m/ z  ¼  55, 69, 87and 103)  [47]. The groups of small peaks appearing at retentiontimes between 19 min and 22 min were assigned to the thermaldegradation products of PLAwith the characteristic series of signalsat m/z  ¼ 56 þ (n  72) attributed to PLA degradation products suchas dimers (n ¼ 2) and trimers (n ¼  3) [32]. In general, the intensityof peaks decreased with composting time. However, the   meso-lactide intensity showed a lower decrease with respect to the (D,L)-lactide equivalent. The ratio meso-lactide:lactide has been used as a

semi-quantitative sign of the degradation of PLA [12,32,36,48,49].

Fig 7(b) shows the reduction of (D,L)-lactide with respect to meso-lactide after the pyrolysis of the recovered samples. No signicantdifferences were observed between PLA and PLA-PHB blend until 7days in composting, but after 10 days, the PLA relationship [lactide/meso-lactidet ¼10 days/lactide/meso-lactidet ¼0 days] highly decreased.Nanocomposites showed similar reduction in 3 days of composting,but higher times revealed higher reduction for unfunctionalizednanocomposites (PLA-CNC and PLA-PHB-CNC). The estimatedreduction of the (D,L)-lactide form with respect to the meso-lactidefollowed a similar tendency that the disintegrability test. In thissense, PLA-CNC and PLA-PHB-CNC showed the highest degradation

rate suggestive of the polymer shortening by the hydrolysis

Fig. 6.   Infrared spectra (2000e1200 cm1) of PLA, PLA-PHB and nanocomposite  lms before and after different time of incubation under composting conditions.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149   147

8/20/2019 1-s2.0-S0141391014001955-main

10/11

resulted in a higher amount of lactic acid. It is known that micro-organisms prefer the L-lactide form of lactic acid rather than the D-form, thus there was a higher enzymatic degradation of L-lactideinuencing the formation of a higher amount of  meso-lactide form

during the pyrolysis test [48].

4. Conclusions

PLA-PHB based nanocomposite   lms reinforced with synthe-sized CNC and functionalized CNCs intended for food packagingwere developed and characterized. A reduction of the viscositymolecular weight of PLA, by approximately 5%, occurred due tothermal processing reaching higher reduction with the presence of CNC (10%) and CNCs (15%). Higher detriment of the viscosity mo-lecular weight was observed for PHB after processing in PLA-PHB.Conversely, in the case of PLA-PHB-CNC and PLA-PHB-CNCs, theaddition of nanocellulose improve the thermal stability leading to alesser reduction of PHB viscosity molecular weight and practically

unaffected the viscosity molecular weight of PLA. The combinationof PHB and the better dispersed CNCs demonstrated the reinforcingeffect increase simultaneously the Young modulus and elongationat break,with comparable tensile strength to those of neat PLA. PHBand functionalized CNCs showed a slight UV blocking effect on thevirtually transparent PLA matrix. Although the addition of PHB ledto a decrease in PLA high transparency, it did not compromise theultimate optical properties due to the low  lm thickness achieved.The presence of PHB increased the crystallinity of PLA and itsnucleation effect reduced the polymer chains mobility enhancingthe oxygen barrier performance of  nal PLA-PHB blend  lms whilethe wettability was reduced. Moreover, functionalized CNCs, whichincreases the polymer-nanoparticle interfacial adhesion, alsoreduced the oxygen transmission at the same time as it decreased

the surface adhesive forces improving the water resistance. Finally,CNC based nanocomposites showed the highest rate of disinte-gration in compost, while the surface hydrolysis of functionalizedCNCs nanocomposite   lms started somewhat later and the pres-ence of crystalline PHB delayed the disintegration process.

The results of this research suggest that the novel combinationof PLA-PHB blends and functionalized CNCs opens a new perspec-tive for their industrial application as short-term food packaging.

 Acknowledgements

This research was supported by the Ministry of Science andInnovation of Spain (MAT2011-28468-C02-01 and MAT2011-28468-C02-02). M.P. Arrieta thanks Generalitat Valenciana (Spain)

for Santiago Grisolí 

a Fellowship (GRISOLIA/2011/007) and

Universitat Politecnica de Valencia for the Development SupportProgramme PAID-00-12 (SP20120120). The Authors acknowledgeGesenu S.p.a. for compost supply. Authors gratefully thank Prof.Alfonso Jimenez (University of Alicante, Spain) and Prof. Marí a

Dolores Salvador Moya (Universitat Polit

ecnica de Val

encia) fortheir assistance with OTR measurements and optical microscope-EDF measurements, respectively.

References

[1]  Fortunati E, Armentano I, Iannoni A, Kenny JM. Development and thermalbehaviour of ternary PLA matrix composites. Polym Degrad Stab 2010;95:2200e6.

[2]  Arrieta MP, L opez J, Ferrandiz S, Peltzer MA. Characterization of PLA-limoneneblends for food packaging applications. Polym Test 2013;32:760 e8.

[3]   Fortunati E, Armentano I, Zhou Q, Puglia D, Terenzi A, Berglund LA, et al.Microstructure and nonisothermal cold crystallization of PLA compositesbased on silver nanoparticles and nanocrystalline cellulose. Polym DegradStab 2012;97:2027e36.

[4]   Fortunati E, Armentano I, Iannoni A, Barbale M, Zaccheo S, Scavone M, et al.

New multifunctional poly(lactide acid) composites: mechanical, antibacterial,and degradation properties. J Appl Polym Sci 2012;124:87 e98.

[5]   Mattioli S, Peltzer M, Fortunati E, Armentano I, Jimenez A, Kenny JM. Struc-ture, gas-barrier properties and overall migration of poly(lactic acid)   lmscoated with hydrogenated amorphous carbon layers. Carbon 2013;63:274e82.

[6]  Carrasco F, Gamez-Perez J, Santana OO, Maspoch ML. Processing of poly(lacticacid)/organomontmorillonite nanocomposites: microstructure, thermal sta-bility and kinetics of the thermal decomposition. Chem Eng J 2011;178:451e60.

[7]   Lagaron JM, Lopez-Rubio A. Nanotechnology for bioplastics: opportunities,challenges and strategies. Trends Food Sci Technol 2011;22:611e7.

[8] Arrieta MP, Samper MD, L ope J, Jimenez A. Combined effect of poly(-hydroxybutyrate) and plasticizers on polylactic acid properties for   lmintended for food packaging. J Polym Environ; 2014.   http://dx.doi.org/10.1007/s10924-014-0654-y.

[9]   Castro L opez MdM, L opez de Dicastillo C, L opez Vilari~no JM, GonzalezRodrí guez MV. Improving the capacity of polypropylene to be used in anti-oxidant active   lms: incorporation of plasticizer and natural antioxidants.

 J Agric Food Chem 2013;61:8462e70.[10]   Zhang M, Thomas NL. Blending polylactic acid with polyhydroxybutyrate: the

effect on thermal, mechanical, and biodegradation properties. Adv PolymTechnol 2011;30:67e79.

[11]   Bucci DZ, Tavares LBB, Sell I. Biodegradation and physical evaluation of PHBpackaging. Polym Test 2007;26:908e15.

[12]   Arrieta MP, L opez J, Hernandez A, Rayon E. Ternary PLAePHBeLimoneneblends intended for biodegradable food packaging applications. Eur Polym J2014;50:255e70.

[13]   Auras R, Harte B, Selke S. An overview of polylactides as packaging materials.Macromol Biosci 2004;4:835e64.

[14]   Lagaron JM, Nú~nez E. Nanocomposites of moisture-sensitive polymers andbiopolymers with enhanced performance for  exible packaging applications.

 J P lastic Film Sheeting 2012;28:79 e89.[15]  Silvestre C, Duraccio D, Cimmino S. Food packaging based on polymer nano-

materials. Prog Polym Sci 2011;36:1766e82.[16]   Fortunati E, Luzi F, Puglia D, Terenzi A, Vercellino M, Visai L, et al. Ternary PVA

nanocomposites containing cellulose nanocrystals from different sources and

silver particles: part II. Carbohydr Polym 2013;97:837e

48.

Fig. 7.  a) Py-GC/MS chromatogram of PLA-PHB-CNCs nanocomposite  lm, b) mesolactide:lactide ratio loss of PLA, PLA-PHB and nanocomposite  lms.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149148

http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://dx.doi.org/10.1007/s10924-014-0654-yhttp://dx.doi.org/10.1007/s10924-014-0654-yhttp://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref16http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref15http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref14http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref13http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref12http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref11http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref10http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://refhub.elsevier.com/S0141-3910(14)00195-5/sref9http://dx.doi.org/10.1007/s10924-014-0654-yhttp://dx.doi.org/10.1007/s10924-014-0654-yhttp://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref7http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref6http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref5http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref4http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref3http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref2http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1http://refhub.elsevier.com/S0141-3910(14)00195-5/sref1

8/20/2019 1-s2.0-S0141391014001955-main

11/11

[17]   Fortunati E, Armentano I, Zhou Q, Iannoni A, Saino E, Visai L, et al. Multi-functional bionanocomposite  lms of poly(lactic acid), cellulose nanocrystalsand silver nanoparticles. Carbohydr Polym 2012;87:1596 e605.

[18]   Martí nez-Sanz M, Lopez-Rubio A, Lagaron JM. High-barrier coated bacterialcellulose nanowhiskers   lms with reduced moisture sensitivity. CarbohydrPolym 2013;98:1072e82.

[19]   Patrí cio PSDO, Pereira FV, Dos Santos MC, De Souza PP, Roa JPB, Orece RL.Increasing the elongation at break of polyhydroxybutyrate biopolymer: effectof cellulose nanowhiskers on mechanical and thermal properties. J ApplPolym Sci 2013;127:3613e21.

[20]   Martí nez-Sanz M, Villano M, Oliveira C, Albuquerque MGE, Majone M, Reis M,et al. Characterization of polyhydroxyalkanoates synthesized from microbialmixed cultures and of their nanobiocomposites with bacterial cellulosenanowhiskers. New Biotechnol 2013;31:364e76.

[21]   Fortunati E, Peltzer M, Armentano I, Jimenez A, Kenny JM. Combined effects of cellulose nanocrystals and silver nanoparticles on the barrier and migrationproperties of PLA nano-biocomposites. J Food Eng 2013;118:117e24.

[22]   Aranguren MI, Marcovich NE, Salgueiro W, Somoza A. Effect of the nano-cellulose content on the properties of reinforced polyurethanes. A study us-ing mechanical tests and positron anihilation spectroscopy. Polym Test2013;32:115e22.

[23]   Jonoobi M, Mathew AP, Abdi MM, Makinejad MD, Oksman K. A comparison of modied and unmodied cellulose nanober reinforced polylactic acid (PLA)prepared by twin screw extrusion. J Polym Environ 2012;20:991 e7.

[24]  Fortunati E, Peltzer M, Armentano I, Torre L, Jim enez A, Kenny JM. Effects of modied cellulose nanocrystals on the barrier and migration properties of PLAnano-biocomposites. Carbohydr Polym 2012;90:948e56.

[25]   Arrieta MP, Fortunati E, Dominici F, Rayon E, L opez J, Kenny JM. Multifunc-tional PLA-PHB/cellulose nanocrystal lms: processing, structural and thermalproperties. Carbohydr Polym 2014;107:16e24.

[26]  UNE EN ISO 1628e1. Determination of the viscosity of polymers in dilutesolution using capillary viscometers. Part 1: general principles; 2009.

[27]  Van Den Oever MJA, Beck B, Müssig J. Agrobre reinforced poly(lactic acid)composites: effect of moisture on degradation and mechanical properties.Compos Part Appl Sci Manuf 2010;41:1628e35.

[28]   Spitalský Z, Lací k I, Lathova E, Janigova I, Chodak I. Controlled degradation of polyhydroxybutyrate via alcoholysis with ethylene glycol or glycerol. PolymDegrad Stab 2006;91:856e61.

[29]   UNE-EN ISO 527e3. Plastics: determination of tensile properties. Part 3: testconditions for  lms and sheets; 1995.

[30]   Fombuena V, Balart J, Boronat T, Sanchez-Nacher L, Garcia-Sanoguera D.Improving mechanical performance of thermoplastic adhesion joints by at-mospheric plasma. Mater Des 2013;47:49e56.

[31]   UNE EN ISO 20200. Determination of the degree of disintegration of plasticmaterials under simulated composting conditions in a laboratory-scale test;2006.

[32]   Arrieta MP, Parres F, L opez J, Jimenez A. Development of a novel pyrolysis-gas

chromatography/mass spectrometry method for the analysis of poly(lacticacid) thermal degradation products. J Anal Appl Pyrol 2013;101:150 e5.

[33]  Erceg M, Kovacic T, Klaric I. Thermal degradation of poly(3-hydroxybutyrate)plasticized with acetyl tributyl citrate. Polym Degrad Stab 2005;90:313 e8.

[34]  Bonilla J, Fortunati E, Atares L, Chiralt A, Kenny JM. Physical, structural andantimicrobial properties of poly vinyl alcohol-chitosan biodegradable   lms.Food Hydrocoll 2013;35:463e70.

[35]  Armentano I, Bitinis N, Fortunati E, Mattioli S, Rescignano N, Verdejo R, et al.Multifunctional nanostructured PLA materials for packaging and tissue engi-neering. Prog Polym Sci 2013;38:1720e47.

[36] Arrieta MP, L opez J, Rayon E, Jimenez A. Disintegrability under compostingconditions of plasticized PLA-PHB blends. Polym Degrad Stab; 2014. http://

dx.doi.org/10.1016/j.polymdegradstab.2014.01.034.[37]   Arrieta MP, Peltzer MA, L opez J, Garrigos MDC, Valente AJM, Jimenez A.

Functional properties of sodium and calcium caseinate antimicrobial activelms containing carvacrol. J Food Eng 2014;121:94e101.

[38]  Puglia D, Fortunati E, D'Amico DA, Manfredi LB, Cyras VP, Kenny JM. Inuenceof organically modied clays on the properties and disintegrability in compostof solution cast poly(3-hydroxybutyrate)   lms. Polym Degrad Stab 2014;99:127e35.

[39]   Fortunati E, Puglia D, Santulli C, Sarasini F, Kenny JM. Biodegradation of Phor-mium tenax/poly(lacticacid) composites. J Appl Polym Sci 2012;125:E562e72.

[40]   Fortunati E, Rinaldi S, Peltzer M, Bloise N, Visai L, Armentano I, et al. Nano-biocomposite   lms with modied cellulose nanocrystals and synthesizedsilver nanoparticles. Carbohydr Polym 2014;101:1122e33.

[41]  Kale G, Auras R, Singh SP. Comparison of the degradability of poly(lactide)packages in composting and ambient exposure conditions. Packag Technol Sci2007;20:49e70.

[42]   Calv~ao PS, Chenal JM, Gauthier C, Demarquette NR, Bogner A, Cavaille JY.Understanding the mechanical and biodegradation behaviour of poly(-hydroxybutyrate)/rubber blends in relation to their morphology. Polym Int2012;61:434e41.

[43]  Weng YX, Wang L, Zhang M, Wang XL, Wang YZ. Biodegradation behavior of P(3HB,4HB)/PLA blends in real soil environments. Polym Test 2013;32:60e70.

[44]  Lucas N, Bienaime C, Belloy C, Queneudec M, Silvestre F, Nava-Saucedo JE.Polymer biodegradation: mechanisms and estimation techniques   e a review.Chemosphere 2008;73:429e42.

[45]  Guo L, Sato H, Hashimoto T, Ozaki Y. FTIR study on hydrogen-bonding in-teractions in biodegradable polymer blends of poly(3-hydroxybutyrate) andpoly(4-vinylphenol). Macromolecules 2010;43:3897e902.

[46]   Mousavioun P, George GA, Doherty WOS. Environmental degradation of lignin/poly(hydroxybutyrate) blends. Polym Degrad Stab 2012;97:1114e22.

[47]   Jimenez A, Ruseckaite RA. Binary mixtures based on polycaprolactone andcellulose derivatives: thermal degradation and pyrolysis. J Therm Anal Calo-rim 2007;88:851e6.

[48]   Khabbaz F, Karlsson S, Albertsson AC. Py-GC/MS an effective technique tocharacterizing of degradation mechanism of poly (L-lactide) in the differentenvironment. J Appl Polym Sci 2000;78:2369e78.

[49]   Westphal C, Perrot C, Karlsson S. Py-GC/MS as a means to predict degree of 

degradation by giving microstructural changes modelled on LDPE and PLA.Polym Degrad Stab 2001;73:281e7.

M.P. Arrieta et al. / Polymer Degradation and Stability 107 (2014) 139e149   149

http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://dx.doi.org/10.1016/j.polymdegradstab.2014.01.034http://dx.doi.org/10.1016/j.polymdegradstab.2014.01.034http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref49http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref48http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref47http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref46http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref45http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref44http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref43http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref42http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref41http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref40http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref39http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref38http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://refhub.elsevier.com/S0141-3910(14)00195-5/sref37http://dx.doi.org/10.1016/j.polymdegradstab.2014.01.034http://dx.doi.org/10.1016/j.polymdegradstab.2014.01.034http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref35http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref34http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref33http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref32http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref31http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref30http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref29http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref28http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref27http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref26http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref25http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref24http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref23http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref22http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref21http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref20http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref19http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref18http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17http://refhub.elsevier.com/S0141-3910(14)00195-5/sref17