Preparation and Viscoelastic Properties of Composite ...publications.lib.chalmers.se/records/fulltext/241483/local_241483.pdf · Preparation and Viscoelastic Properties of Composite

Research ArticlePreparation and Viscoelastic Properties of Composite FibresContaining Cellulose Nanofibrils Formation of a CoherentFibrillar Network

Tobias Moberg123 Hu Tang4 Qi Zhou234 and Mikael Rigdahl123

1Department of Materials and Manufacturing Technology Chalmers University of Technology 412 96 Gothenburg Sweden2Wallenberg Wood Science Center Chalmers University of Technology 412 96 Gothenburg Sweden3Royal Institute of Technology 100 44 Stockholm Sweden4Royal Institute of Technology School of Biotechnology 100 44 Stockholm Sweden

Correspondence should be addressed to Mikael Rigdahl mikaelrigdahlchalmersse

Received 15 March 2016 Revised 2 June 2016 Accepted 16 June 2016

Academic Editor Ilker S Bayer

Copyright copy 2016 Tobias Moberg et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Composite fibres with a matrix of poly(ethylene glycol) (PEG) and cellulose nanofibrils (CNF) as reinforcing elements wereproduced using a capillary viscometer Two types of CNFwere employed one based on carboxymethylated pulp fibres and the otheron TEMPO-oxidized pulp Part of the latter nanofibrils was also grafted with PEG in order to improve the compatibility betweenthe CNF and the PEGmatrixThe nominal CNF-content was kept at 10 or 30 weight-The composite fibres were characterized byoptical and scanning electron microscopy in addition to dynamic mechanical thermal analysis (DMTA) Evaluation of the storagemodulus indicated a clear reinforcing effect of the CNF more pronounced in the case of the grafted CNF and depending on theamount of CNF An interesting feature observed during the DMTA-measurements was that the fibrils within the composite fibresappeared to forma rather coherent and load-bearing networkwhichwas evident even after removing of the PEG-phase (bymelting)An analysis of the modulus of the composite fibres using a rather simple model indicated that the CNF were more efficient asreinforcing elements at lower concentrations which may be associated with a more pronounced aggregation as the volume fractionof CNF increased

1 Introduction

The interest in using polymers and composites based onrenewable resources has increased tremendously duringrecent years compare for example [1] This is coupled to theaim towards a more sustainable society and decoupling of thedependence of fossil-based raw materials The developmentand use of poly(lactic acid) which can be derived fromstarch or sugar can be used as an example here as well asthermoplastic starch see for example [2 3] Natural fibres asreinforcing elements in composite materials also contributeto the aim towards more sustainable materials An obviousexample of such elements is cellulosic fibres such as woodand plant fibres The idea of using cellulosic materials indifferent forms as reinforcements or fillers is certainly notnew [4] but modifications and refinements of the material

have during recent years opened up new possibilities withregard to different applications

Cellulose fibres can be said to consist of a number ofnanofibrils or microfibrils and fibril aggregates which areimportant building blocks for the fibres These fibrils can beseparated from the cellulose fibres by aqueous homogeniza-tion process at high pressurehigh shear rates often combinedwith a pretreatment of the fibres using for example enzymes[5] 2266-tetramethylpiperidine-1-oxyl (TEMPO) radical-mediated oxidation [6] or carboxymethylation [7] in orderto reduce the energy consumption The fibrils are denotedas cellulose nanofibrils (CNF) or nanofibrillated cellulose(NFC) and the term microfibrillated cellulose (MFC) issometimes also used CNF have been suggested for sev-eral different applications and products for example high-toughness paper [8ndash11] barriermaterial (lowering the oxygen

Hindawi Publishing CorporationJournal of NanomaterialsVolume 2016 Article ID 9569236 10 pageshttpdxdoiorg10115520169569236

2 Journal of Nanomaterials

and oil permeability) [12ndash14] aero- and hydrogels [15ndash18]and reinforcing elements in composite materials [15 19ndash24]In this context it is certainly of interest that as pointed outby Lee et al [19] the number of publications on cellulosenanocomposites has increased exponentially over the recentyears but already in the 1980s initial attempts were made touse nanocellulose as a reinforcing element in thermoplasticssee for example [24] Several types of matrix polymers havebeen combined with CNF (and similar elements) such aspoly(lactic acid) poly(vinyl alcohol) epoxy starch polyure-thane and unsaturated polyester just to mention a few

After the homogenization process the concentration ofthe CNF is usually quite low of the order of 2ndash5 weight-(wt-) or lower in order to minimize or avoid aggregationThe aggregation is associated with colloidal interactions andthe entanglement of the fibrils Drying of the suspensionscan lead to hornification of the material and redispersing thedry fibril aggregates is very difficult This behaviour preventsa straightforward incorporation of CNF into commoditythermoplasticmatrices which inmost cases are hydrophobicThis consequently makes manufacturing of nanocompositesusing conventional processing techniques such as extrusionand injection moulding challenging [25]

Closely connected to the processability of materials aretheir rheological properties Even at low concentrationsthe CNF suspensions exhibit quite a complex rheologicalbehaviour they are shear-thinning at least at not too highshear rates and elastic in nature compare for example[5 26ndash28] Both the shear viscosity and the viscoelasticparameters of the CNF suspensions are affected by additiveslike salts carboxymethyl cellulose cationic starch and poly-methacrylates [29 30] The CNF suspensions with a fibrilcontent exceeding the percolation threshold exhibit a gel-likebehaviour with a yield stress a storage modulus greater thanthe loss modulus and the moduli being rather insensitiveto changes in the measuring frequency for example [5 26]and to moderate changes in temperature up to 80∘C [5 31]When subjected to an extensional flow the CNF suspensionsexhibited quite a high extensional viscosity with a Troutonratio which significantly exceeded three [32]

The increased interest in CNF (cf [15]) is to some extentcoupled to their potentially good mechanical performanceTheir tensile modulus (including values of the cellulosecrystal) has been estimated to be in the range 100 to 160GPa[19 33] obviously making them interesting as reinforcingelements in polymermatrices For comparison typical valuesof the modulus in the longitudinal direction for glass fibresand aramid fibres are 70 and 125GPa respectively [34]

The tensile modulus of nanopapers and wet-spun-fibresproduced from CNF or similar elements has been reportedto be between about 15 and somewhat higher than 30GPa[10 35ndash37] In order to utilize the potentially high modulusof the CNF in a polymer composite the reinforcing elementsshould bewell dispersed and the adhesion between thematrixand the elements should be sufficient [34] Aggregation of theCNF leads to lower specific area a less efficient stress transferbetween thematrix and the fibrils and thus a less efficient useof modulus of fibrils compare [19] In the case of unsaturatedpolyester containing CNF Ansari and coworkers [23]

evaluated an effective tensile modulus of the CNF and foundit to decrease from 42GPa at a CNF-content of 16 wt- downto 25GPa at 45wt- CNF It was speculated that aggregationof the CNF was associated with the decrease in modulus withincreasing amounts of CNF thus deceasing the efficiency ofthe CNF Ansari et al [38] also noted a similar behaviourwith regard to the effective modulus in the case of CNF-containing epoxy In a recent work Tang et al [39] preparedoriented ribbon-shaped composites consisting of CNF withgrafted poly(ethylene glycol) The effective modulus of theCNF was in that case estimated to be of the order of 55GPaat a volume fraction CNF of 0585 in the polymer composite

The aim of the present study is to evaluate the possi-bility of producing composite fibrous materials containingCNF using a rather simple technique based on capillaryviscometry In principle the technique used for the man-ufacturing can after some modifications be scaled up ina straightforward manner in order to produce significantamounts of these composite macrofibres The viscoelasticproperties of the composite fibres in the solid state andthe reinforcing efficiency of the CNF were evaluated usingdynamic-mechanical thermal analysis (DMTA)The polymerused for thematrixmaterial was poly(ethylene glycol) (PEG)The polymer is dissolved in the aqueous CNF suspensionin order to retard the aggregation (and collapse) of thenanofibrils during drying of the system thus contributingto the homogeneity of the composite and enhancing theefficiency of the CNF with regard to the stiffness of the finalcomposite Two kinds of CNF were employed and in a seriesof experiments also CNF grafted with PEG according tothe procedure outlined in [39] were used The grafted PEGshould improve the compatibility with the PEG used for thematrix thus providing a more stable interphase region Inaddition to the stiffness of the final composite as determinedusing DMTA the rheological behaviour of the CNF-polymersuspensionswas assessed since it reflects possible interactionsbetween the fibrils and the polymer aswell as providing usefulinformation regarding the processability of the compositesystem The integrity of the formed network of CNF inthe composite fibres was assessed through the removal ofthe PEG-phase In principle composite fibres of the kindconsidered here can be used as reinforcing elements in othertypes of composite materials

2 Materials

Two different types of nanofibrillated (or microfibrillated)cellulose (CNF) were used in this work One was pre-pared by high pressure homogenization of aqueous woodfibre suspensions (softwood sulphite dissolving pulp fromDomsjo Fabriker Sweden) and kindly supplied by InnventiaAB Sweden The pulp was carboxymethylated before beingpassed through a high pressure homogenizer Due to thecarboxymethylation the resulting fibrils with a width in therange 5ndash10 nm were negatively surface charged correspond-ing to a degree of substitution close to 01 The preparationand the properties of the CNF are described in more detailby Wagberg et al [7] and this type was denoted as C-CNF The concentration of the cellulose in the as-received

Journal of Nanomaterials 3

suspension was 22 weight percent (wt-)The other suspen-sion denoted as T-CNF was prepared by TEMPO-mediatedoxidation of the pulp fibres (softwood sulphite pulp providedby Nordic Paper) a part of this suspension was graftedwith poly(ethylene glycol) onto the pulp fibres as describedin [39] The grafting was achieved through carbodiimide-mediated amidation with amino-terminated poly(ethyleneglycol) Aqueous suspensions of the chemically modifiedfibres were mechanically disintegrated in order to obtainsuspensions of nanofibrillated cellulose with a width ofthe order of 4 nm for the nongrafted fibrils The TEMPO-oxidized fibres had total carboxylate content of 15mmolgcellulose corresponding to surface degree of substitution of05 [39] The amount of grafted poly(ethylene glycol) was33wt- In the suspension this was found to correspond to awidth of about 5 nm in the case of the grafted fibrils [39]TheCNF concentration of these suspensions (from the TEMPO-oxidized pulp fibres) was 02 wt-

The polymer used as a matrix in the composite materialwas poly(ethylene glycol) PEG or poly(ethylene oxide)PEO from Clariant International Ltd Muttenz Switzerlandgrade 35000 SThis polymer iswater-soluble enabling a ratherstraightforward mixing into the CNF suspension and pro-viding a better compatibility with cellulose than commoditythermoplastics which normally are quite hydrophobic ThePEG-grade used in this work had according to the supplieran average molecular weight of 35000 gmole a density of120 gcm3 and a melting point of at least 57∘C

21 Preparation of Composite Fibres For producing thecomposite material a certain amount (depending on theaimed cellulose content) of PEG was dissolved in the CNFsuspension The polymer-containing suspension was thendried at 65∘C until the weight of the material was the same asthe calculated dry weight of the sample This composite wasthen used for producing the composite fibres Two nominalamounts of C-CNF in the PEGmatrix were used here 10 and30wt- (dry state) In the case of the TEMPO-oxidized CNFgrafted with PEG only 30wt- was added this means thatwith the grafted fibrils the cellulose content was only 20wt- since one-third of those fibres consisted of grafted PEGcompare [39]

The composite fibres were produced using a capillaryviscometer (Gottfert Rheograph 2002 Germany) The useof a capillary viscometer allowed for an estimation of theviscosity of the composite compound as the fibres wereformed The capillary used had a diameter of 2mm a lengthof 20mm and an entrance angle of 90∘ Smaller diametersof the capillary led to an excessive pressure build-up in thebarrel The temperature was set to 63∘C Lower temperaturesresulted in an incomplete melting of the polymer and whenincreasing the temperature above 66-67∘C the viscosity of themelt became too low to allow for a proper fibre formation

Seven different speeds of the piston pushing the materialthrough the capillary were used from 001 to 1mms corre-sponding to shear rates between 144 and 144 sminus1 in case ofthe used capillary The pressure was measured by a pressuretransducer in the barrel close to the capillary The pressurereadings and the corresponding piston speeds were used to

calculate the apparent viscosity in the usual manner Thepiston speed was increased when the pressure reading hadstabilizedThefibres obtained at the highest piston speedwereused for the dynamic-mechanical testing see below PEG-fibres without CNF to be used as a reference were producedin a corresponding manner

22 Dynamic-Mechanical Thermal Analysis (DMTA) Theviscoelastic properties given by the storage and loss mod-uli in tension of the fibres from the capillary viscometerwere determined by dynamic-mechanical thermal analysis(DMTA) at an applied frequency of 1Hz using a RheometricsRSA-2 equipment The specimens were subjected to twodifferent types ofmeasurements a strain sweep and a thermalsweep The strain sweep was performed at room temperatureand ambient pressure and the storage (1198661015840) and loss moduli(11986610158401015840) as well as the mechanical loss factor (tan 120575) were deter-mined as functions of an increasing applied strain amplitudeAt low strain amplitudes the moduli were independent ofthe deformation but when a ldquocriticalrdquo strain amplitude wasexceeded they decreased with increasing strain as observedearlier [26 27] Below the critical value the fibres thusexhibited a linear viscoelastic behaviour For each sample thestrain sweep was repeated three times (using three differentspecimens) and the variation in storage modulus betweenthe specimens was less than 15 (standard deviation) Thediameter of the fibre was measured using a digital calliper atthree different positions along its length and the variationwastypically 1

In the temperature sweep the maximum applied strainamplitude was kept within the linear viscoelastic region(given by the strain sweep) The starting temperature was25∘C (room temperature) and the viscoelastic parameterswere determined as a function of temperature at a heatingrate of 2∘Cmin up to 120∘C For each sample the temperaturesweep was in most cases repeated three times (using differentspecimens) The variation in storage modulus between thespecimens was 20 or less (standard deviation)

23 Scanning Electron Microscopy (SEM) The cross sectionsof fractured fibres as well as their surface structure werevisually examined in a scanning electron microscope (SEM)The samples were fractured after being submerged in liquidnitrogen and the fracture surfaces were then coated with anapproximately 5 nm thick gold layer using a Sputter CoaterS150B BOC Edwards UK The SEM used in this work wasa digital scanning electron microscope Carl Zeiss DSM 940Germany

24 The CNF-Content in the Extrudates In order to estimatethe actual amount of CNF in the composite fibres somepart of the fibres nominally containing 30wt- C-CNF wasimmersed in an excess amount of deionized water for at least24 hours This was done in order to dissolve the PEG Thesuspension was then filtered through a filter paper and boththe solid residue and the filtrate were collected The solidresidue was again dissolved in deionized water for at least 24hours and once again filtered and collectedThis solid residuewas then dried at low pressure and 65∘C until its weight was

4 Journal of Nanomaterials

Pure PEG10wt- C-CNF30wt- C-CNF

101

102

103

100

Shear rate (sminus1)

103

104

105

106

Visc

osity

(Pa s

)

Figure 1 The shear viscosity as a function of shear rate at 63∘C forunfilled and C-CNF-containing PEG

constant The weight of the residue corresponded to 28wt-of the initial weight of the sample which is quite close to thenominal weight fraction of the CNF in the fibresThus it maybe reasonable to conclude that the formation of the compositefibres in the capillary viscometer did not significantly changethe CNF-content

3 Results

31 Melt Viscosity of the Composite Melts Figure 1 shows theviscosity as a function of the shear rate at 63∘C for unfilledPEG and for the composite melts containing nominally 10and 30wt- carboxymethylated CNF (denoted as C-CNF)As expected the viscosity at a given shear rate increasedwith increasing C-CNF concentration compare for example[40 41] Except for the melt containing the highest amountof CNF there was a tendency for the melts to approach aNewtonian plateau at low shear rates At higher shear ratesall specimens exhibited a shear-thinning behaviour which isnot unexpected

With exception of the low shear rate region the resultsshown in Figure 1 can be fitted to a power-law type of relationthat is

120578 () = 119870 ()119899minus1

(1)

where 120578 is the shear viscosity the shear rate 119870 theconsistency and 119899 a flow index The latter two are materialparameters Table 1 gives 119870 and 119899 for the different materialsused here and it is obvious that the degree of shear-thinningincreased as the amount of C-CNF increased since the flowindex decreased This might be due to a more pronouncedorientation and disentanglement of the fibrils at higher con-centrations as the shear rate increases A similar behaviourcan be noted for other polymeric systems containing cellulosefibres compare [40]

Table 1 The power-law parameters for PEG-melts containing CNFat 63∘C

Material Consistency 119870 kPa s119899 Flow index 119899Unfilled PEG 212 05810wt- C-CNF 854 04130wt- C-CNF 2706 02830wt- nongrafted T-CNF 4396 03130wt- grafted T-CNF 1243 044

101

102

103

100

Shear rate (sminus1)

103

104

105

106

GraftedUngrafted

Visc

osity

(Pa s

)

Figure 2 The shear viscosity as a function of shear rate at 63∘C forthe composite melts containing TEMPO-oxidized CNF with andwithout grafted PEG The nominal CNF-content was 30wt-

The viscosity of the composite melt containing 30wt-TEMPO-oxidized CNF (denoted as T-CNF) was also mea-sured during the fibre formation in the capillary viscometerThe viscosity of thesemelts (with nongrafted aswell as graftedCNF) is shown as a function of the shear rate at 63∘C inFigure 2 see also Table 1

The viscosities of the melts containing 30wt- C-CNFand nongrafted T-CNF were not too different although theviscosity of the latter was somewhat higher at a given shearrate Whether the flexible and slightly more slender T-CNFgave a more flocculated structure thus enhancing the visco-sity may be speculated since it is known that aggregationnormally will lead to an increase in viscosity [41] The differ-ent surface charge character betweenC-CNF andT-CNFmayalso play a role here compare [41] Grafting PEG on the T-CNF reduced the viscosity significantly and also diminishedthe shear-thinning character of the composite melt Animproved compatibility between the grafted fibres and thepolymer matrix leading to less agglomerated fibril structuremay be one reason here [41 42] but reduction of the cellulosecontent from 30 to 20wt- can also provide an explanationat least partially

32 Microscopy Studies of the Composite Fibre StructureFigure 3 is a representative collection of optical micrographs

Journal of Nanomaterials 5

1mm

Figure 3 Optical micrographs of the extruded composite fibresFrom left to right unfilled PEG fibre with 10wt- C-CNF fibrewith 30wt- C-CNF fibre with 30wt- nongrafted T-CNF andfibre with 30wt- grafted T-CNF The diameter of the fibres was2 plusmn 002mm

of the different fibres produced here The diameter of thefibres was 2 plusmn 002mm and they were manufactured usingthe highest piston speed during the capillary extrusionTherewere clear signs of fibril agglomeration along the lengthof the composite fibres seen as somewhat whitish areas inthe micrographs that is the fibrils did not form a uniformnetwork in the radial direction of the composite fibresthroughout the fibre length and the fibrils were not unidi-rectional oriented although a more preferred orientation ofthe fibrils along the fibre axis could probably be expected Inmore detail the fibrils formed an interconnecting structure inthe axial direction of the fibres (as evident from the DMTA-results presented below) but the CNF distribution was nothomogeneous in the radial direction

The fibres containing T-CNF fibrils exhibited a yellow-ish tint compared to those with C-CNF This effect hasbeen observed by others in the case of TEMPO-oxidizedcellulose for example as described by Takaichi et al [43]They reported that oven-drying of the cellulose promotedthe formation of C6-aldehydes and C2C3 ketones whichcontributed to the yellowing

Scanning electron micrographs of typical fracture sur-faces (see Figure 4) indicated that the fibrils were quite welldistributed over the cross section at fibril contents of 10 and30wt- C-CNF in some sections of the fibres However asalready pointed to this varied along the length of the fibresin some sections there seemed to be less fibrils and thenthey were more concentrated towards certain parts of thecross section No observable difference between the fracturesurfaceswas notedwhenC-CNFwas exchanged for T-CNFAcloser examination of the fracture surfaces showed that therewere very few signs of fibril pull-out from the matrix whichindicates a good adhesion between the fibrils and the matrixcompare [40] This was the case even when the nongraftedfibrils were used as the reinforcing elements Another inter-esting observation was that the (outer) fibre surface becamesmootherwith increasing fibril content see Figure 4Thiswasalso evident when touching (feeling) the fibres The unfilledPEG-fibre exhibited the highest surface roughness Therecould be different reasons for this behaviour For example

the addition of the fibrils could stabilize the flow through thecapillary in itself It may also be that the surface irregularitiesappear as a result of the die-swell when the melt exits thecapillary compare [44] Adding fibrils to the melt can reducethe die-swell (cf [44]) and then reduce the surface roughness

33 Dynamic-Mechanical Thermal Analysis of the CompositeFibres As described above two different kinds of measure-ments were performed both using the same equipment Thedynamic strain sweep was performed at room temperatureand ambient pressure The reference sample that is thesample containing only PEG matrix material exhibited analmost constant storage modulus 1198641015840 in the linear viscoelasticregion The value of the storage modulus in this region wasabout 1 GPa (more precise 084GPawith a standard deviationof 007GPa) and the onset of the nonlinear behaviouroccurred at strains around 01 as shown in Figure 5 Strain-sweep experiments were also performed with the compositefibres containing 10 and 30wt- C-CNF Figure 5 includesan example of such a measurement in case of a fibre witha nominal C-CNF-content of 30wt- The storage modulusin the linear viscoelastic region was here about 2GPa (moreprecise 18 GPa with standard deviation of 036GPa) thatis the addition of the CNF increased the modulus It maybe remarked that in view of the inhomogeneous aggregatestructure visualised in Figure 3 the variation in storagemodulus of the composite fibres was quite low indicatingthat CNF reinforced the polymer in a rather coherentmanneralong the fibre length

The critical strain for the onset of nonlinear behaviour(corresponding in a sense to a disruption of the structureof the material) in case of the composite fibre shown inFigure 5 was about 005 Adding fibrils to the PEG matrixthus decreased the linear viscoelastic region for the materialwhich is not unusual when fillers are incorporated intopolymeric materials [45]

The onset of the nonlinear region denotes the maximumstrain that can be applied when performing the temperaturesweeps since such measurements should be performed atconditionswhere the viscoelastic parameters are independentof the applied strain Figure 6 shows 1198641015840 as a function of thetemperature for the reference PEG matrix material In thiscase the temperature region was cut short since the materialmelted around 60∘C and thus no signal was obtained attemperatures exceeding 65∘C The applied strain amplitudehere was 002 which is within the linear viscoelastic rangeaccording to the strain sweep At temperatures below themelting point themodulus decreased as expected somewhatwith increasing temperature

Figure 7 shows the temperature dependence of the stor-age modulus and of the mechanical loss factor tan 120575 at 1Hzfor the same type of composite fibre as in Figure 5The strainamplitude was set to 001 that is well below the critical strain

In contrast to the behaviour of the unfilled PEG matrixthe storage modulus of the composite fibre did not decreaseto a very low value at the melting point of the polymerInstead the modulus was reduced about a decade andattained a ldquopseudo-rdquo plateau level around 100MPa whichdecreased rather slowly with increasing temperature The

6 Journal of Nanomaterials

200120583m

(a)

200120583m

(b)

Figure 4 Scanning electron micrographs showing the fracture surface of a fibre containing 10wt- C-CNF (a) and one containing 30wt-C-CNF (b) The surface (outer) area can be seen in the bottom right corner (a) and bottom left corner (b) of the micrographs

PEG 30wt- CNFPure PEG

100

10minus2

10minus1

10minus3

10minus4

Strain

10minus3

10minus2

10minus1

100

101

Stor

age m

odul

us (G

Pa)

Figure 5The storage modulus 1198641015840 as a function of the applied strainamplitude at a frequency of 1Hz for the PEG matrix (containing noCNF) and for a composite fibre containing 30wt- C-CNF

plateau extended up to the highest temperature used here120∘C without exhibiting any major changes Since the PEGmatrix melted already around 60∘C it may be concludedthat the quite high storage modulus is associated with acoherent and interconnected fibrillar network of CNF in thecomposite This network was then formed when preparingthe composite material andor when the composite meltflowed through the capillary The mechanical loss factorpassed through a pronounced maximum in the meltingregion of the matrix polymer and then decreased ratherslowly with increasing temperature when the melting pointhad been exceeded It is not uncommon that tan 120575 exhibitsa high value when a polymer softens compare for example[46]

10minus5

10minus4

10minus3

10minus2

10minus1

100

101

Stor

age m

odul

us (G

Pa)

8050 60 704030

Temp (∘C)

Figure 6 The storage modulus 1198641015840 as a function of temperature forthe PEG matrix (containing no CNF)

As shown in Figure 8 the composite fibre containing10wt- C-CNF also exhibited a plateau at higher tempera-tures The modulus associated with the plateau was howeverlower approaching 10MPa than that of the fibre containing30wt- C-CNF which is in line with the lower fibril content

The composite fibres containing nominally 30wt-grafted and nongrafted T-CNF displayed a modulus-strainamplitude relation that was almost identical to that of thecorresponding fibre containing the same amount of C-CNFthe extent of the linear viscoelastic region was not markedlyaffected The strain amplitude used when performing thetemperature sweep was 001 Figure 9 shows the temperaturedependence of the storage modulus for the composite fibrescontaining the grafted as well as the nongrafted T-CNF

The temperature dependence of 1198641015840 resembled to a largeextent that of the composite fibre containing 30wt- C-CNF

Journal of Nanomaterials 7

00

01

02

03

04

Tan

delta

80 12040 60 10020

Temperature (∘C)

10minus2

10minus1

100

101

Stor

age m

odul

us (G

Pa)

Figure 7 The storage modulus 1198641015840 and the mechanical loss factortan 120575 at 1Hz as functions of temperature for a fibre containing 30wt- C-CNF

8040 50 60 7030 90 100 110 120 13020

Temperature (∘C)

10minus3

10minus2

10minus1

100

101

Stor

age m

odul

us (G

Pa)

Figure 8 The storage modulus 1198641015840 as a function of the temperaturefor a composite fibre containing 10wt- C-CNF

Again a plateau in the storage modulus was noted at temper-atures higher than the melting point of the matrix polymerfor both the T-CNF-containing fibres The correspondingvalue of the modulus was of the order of 100MPa beingsomewhat higher when the nongrafted T-CNF was used asthe reinforcing phase However it should be rememberedthat the composite containing nominally 30wt- graftedT-CNF actually contained only 20wt- cellulosic materialindicating that the grafting promoted a stiffer (and probablystronger) fibrillar network in the composite fibres This isquite plausible since the storage moduli of composite fibrescontaining nongrafted CNF apparently scaled with the CNF-content compare Figures 7 and 8

4 Discussion

The addition of the CNF to the polymer matrix obviouslyhad a strong effect on the modulus of the fibres thatis with about 20 volume- CNF the modulus at room

80 12040 60 10020

Temperature (∘C)

10minus2

10minus1

100

101

Stor

age m

odul

us (G

Pa)

GraftedUngrafted

Figure 9 The storage modulus 1198641015840 as a function of the temperaturefor composite fibres containing 30wt- grafted and nongrafted T-CNF

temperature was approximately doubled Here the weightfraction of the CNF (30wt-) has been recalculated into acorresponding volume content using a density of 1500 kgm3for the CNF giving about 20 volume- The nanofibrils hadalso the expected effect on themelt viscosity of the compositemelt that is it increased compare for example [40] andthe critical strain for onset of the structure deteriorationdecreased With regard to the stiffness enhancement noappreciable difference between C-CNF and nongrafted T-CNF was observed The grafting appeared to improve theefficiency of the reinforcing elements since about the samestorage modulus of the composite fibres was noted as in thecase of the nongrafted T-CNF although the cellulose contentwas about 30 lower This may be interpreted as a result ofan improved compatibility between the grafted PEG and thePEG matrix Although this is plausible the lack of extensivefibre pull-out in the fracture surfaces of all the compositefibres did indicate that the adhesion between theCNF and thematrix was in general quite goodThe enhanced compatibilitymay then result in an improved dispersion of the CNF in thematrix which would promote the mechanical properties

Both the mechanical performance and the micrographsshow that the manner in which the composite fibres wereproduced counteracted at least to some extent the collapse ofthe fibril network when dryingThus the associated decreasein surface area of the fibrillary network was counteracted andthe reinforcing effect of the fibrils was enhanced This wasalso one of the aims of the preparation technique Admittedlyand as pointed to earlier a perfect uniform distribution wasnot obtained and the micrographs Figure 3 revealed thataggregates were clearly formed to some extent in the fibresIn a series of experiments attempts were made to align thefibrillar network by orienting the composite fibres at elevatedtemperature andor in a moistened state These attemptswere however not successful since either the polymer matrixwas too brittle or its softening (melting) region was not

8 Journal of Nanomaterials

sufficiently broad both of which are likely to be associatedwith the rather low molecular weight of the PEG matrix(35 000 gmol) A higher degree of alignment would howeverbe desirable from the mechanical performance point of view

A striking feature of the temperature dependence of themodulus of the composite fibres is the second plateau at tem-peratures above the melting point of the polymer matrix seeFigures 7ndash9 The magnitude of the corresponding modulusscales with cellulose content (with exception of the compositecontaining the grafted T-CNF) An interpretation is that theCNF form a coherent network in the matrix (otherwise themodulus would decrease to zero) This network can resultfrom the preparation of the composite material andor beformed when the composite melt flows through the capillaryVisual examination of the fibres after the exposure to the hightemperatures revealed that they consistedmore or less only ofcellulose fibrils in a collapsed state

In this context it is interesting to estimate how well thereinforcing ability of the CNF with regard to the stiffness isutilized in the composite fibres The modulus of a compositecan be modelled in several ways compare for example[34] and the modelling involves a number of assumptionsand estimations The fibrils are assumed to be straight withcylindrical cross section and without any defects whichclearly is not the case here and it is assumed that the adhesionbetween the matrix and the reinforcement is sufficient Arather simple model that can be used for the intendedpurpose is that of Cox-Krenchel [19 34 47]

119864119888= 120578119889120578119897V119891119864119891+ (1 minus V

119891) 119864119898 (2)

where 119864119888is the elastic modulus of the composite fibres in

the fibre direction V119891the volume fraction of CNF 119864

119891the

(effective) modulus of the fibrils in the axial direction and119864119898the modulus of the matrix (here taken to be 09GPa from

the measurements of the storage modulus) The factor 120578119889

accounts for the orientation of the fibrils here it is assumedthat the fibrils are randomly oriented in three dimensions (asroughly estimated from the optical micrographs) and in sucha case 120578

119889= 02 [46] (which might be an underestimation)

The correction for a finite fibril length given by 120578119897stems from

the shear-lag theory [34] and is given by

120578119897= 1 minus

tanh (119898119886)119898119886 (3)

Here 119886 is the aspect ratio of the fibrils that is the ratiobetween the fibril length (119897) and the diameter (119889) From thedata on fibril dimensions reported by Wagberg et al [7] 119886 =100 appears to be a reasonable value the analysis performedhere is actually not very sensitive to values of 119886 between 50and 200 The factor119898 is obtained from

119898 = radic2119866119898

119864119891ln (2119877119889)

(4)

where 119866119898

is the shear modulus of the matrix and 2119877 isthe distance between the fibrils The shear modulus can beestimated from the storage modulus of PEG and assuming

that the fibrils are arranged in square array the ratio 119877119889 canbe related to the volume fraction of fibrils V

119891

With these assumptions and with a volume fraction offibrils of 02 an effective fibril modulus 119864

119891of about 34GPa

was obtained using the values of the storagemodulus at roomtemperature of the composite fibres containing C-CNF At avolume content of 016 corresponding to composite contain-ing grafted T-CNF (only accounting for the cellulosic mate-rial) the effectivemodulus was 44GPa and a further decreaseof the volume content to 008 corresponding to the fibrecontaining 10wt- C-CNF increased the evaluated 119864

119891-value

to more than 60GPa The calculated values of the effectivemodulus may be somewhat overestimated (mainly due to theassumption regarding the fibril distribution) but the trend isclear With decreasing fibril content the effective fibril mod-ulus increases Similar results were reported in [23 38] usingthe Halpin-Tsai model where CNF were incorporated intopolyester and epoxy matrices It was suggested that agglom-eration at higher fibril contents could lead to a less efficientload transfer between the fibrils and the matrix (and a loweravailable surface area of the fibrils) resulting in dependenceof the effective modulus on the fibril content which seemsquite plausible The calculated values of the fibril modulusare high but clearly below the possible limits [19 33] whichgives room for improvements with regard to the mechanicalperformance of the polymer-based composites in terms ofa more homogeneous fibril distribution and in the presentcase an enhanced alignment of theCNF in the fibre direction

5 Conclusions

The addition of the CNF to the polymer matrix clearlyenhanced the stiffness of the composite fibres and decreasedthe critical strain for onset of the structure deteriorationThe grafting of PEG on the T-CNF improved the efficiencyof the reinforcing elements since about the same storagemodulus (stiffness) of the composite fibres was noted as in thecase of the nongrafted T-CNF although the cellulose contentwas about 30 lower This can be interpreted in terms ofan improved compatibility between the grafted PEG and thePEGmatrix possibly associated with an enhanced dispersionof the CNF in the matrix

The DMTA-measurements clearly revealed a secondplateau in themodulus-temperature curves for the compositefibres at temperatures exceeding themelting point of the PEGmatrix It is suggested that this plateau can be associatedwith fibrillary network that is coherent and interconnectedin the axial direction of the composite fibres although thedistribution of the CNF was not homogeneous in the radialdirection The analysis of the experimental stiffness resultspointed to the fact that agglomeration at higher fibril contentscould lead to less efficient load transfer between the fibrils andthe matrix (and a lower available surface area of the fibrils)and thus a lower effective reinforcement provided by theCNF

Competing Interests

The authors declare that they have no competing interests

Journal of Nanomaterials 9

Acknowledgments

The authors acknowledge Wallenberg Wood Science Centerand Chalmers University of Technology for the financialsupport

References

[1] A J Ragauskas C K Williams B H Davison et al ldquoThe pathforward for biofuels and biomaterialsrdquo Science vol 311 no 5760pp 484ndash489 2006

[2] K Oksman Y Aitomaki A P Mathew et al ldquoReview of therecent developments in cellulose nanocomposite processingrdquoComposites A Applied Science and Manufacturing vol 83 pp2ndash18 2016

[3] MThunwall V Kuthanova A Boldizar andM Rigdahl ldquoFilmblowing of thermoplastic starchrdquo Carbohydrate Polymers vol71 no 4 pp 583ndash590 2008

[4] S J Eichhorn C A Baillie N Zafeiropoulos et al ldquoCurrentinternational research into cellulosic fibres and compositesrdquoJournal of Materials Science vol 36 no 9 pp 2107ndash2131 2001

[5] M PaakkoM Ankerfors H Kosonen et al ldquoEnzymatic hydro-lysis combined with mechanical shearing and high-pressurehomogenization for nanoscale cellulose fibrils and strong gelsrdquoBiomacromolecules vol 8 no 6 pp 1934ndash1941 2007

[6] A Isogai T Saito and H Fukuzumi ldquoTEMPO-oxidized cellu-lose nanofibersrdquo Nanoscale vol 3 no 1 pp 71ndash85 2011

[7] L Wagberg G Decher M Norgren T Lindstrom M Anker-fors and K Axnas ldquoThe build-up of polyelectrolyte multilay-ers of microfibrillated cellulose and cationic polyelectrolytesrdquoLangmuir vol 24 no 3 pp 784ndash795 2008

[8] Oslash Eriksen K Syverud andOslash Gregersen ldquoThe use ofmicrofib-rillated cellulose produced from kraft pulp as strength enhancerin TMP paperrdquoNordic Pulp and Paper Research Journal vol 23no 3 pp 299ndash304 2008

[9] T Taipale M Osterberg A Nykanen J Ruokolainen andJ Laine ldquoEffect of microfibrillated cellulose and fines on thedrainage of kraft pulp suspension and paper strengthrdquoCellulosevol 17 no 5 pp 1005ndash1020 2010

[10] M Henriksson L A Berglund P Isaksson T Lindstrom andT Nishino ldquoCellulose nanopaper structures of high toughnessrdquoBiomacromolecules vol 9 no 6 pp 1579ndash1585 2008

[11] M Ankerfors T Lindstrom and D Soderberg ldquoThe use ofmicrofibrillated cellulose in fine paper manufacturingmdashresultsfrom a pilot scale papermaking trialrdquo Nordic Pulp and PaperResearch Journal vol 29 no 3 pp 476ndash483 2014

[12] C Aulin M Gallstedt and T Lindstrom ldquoOxygen and oil bar-rier properties of microfibrillated cellulose films and coatingsrdquoCellulose vol 17 no 3 pp 559ndash574 2010

[13] H Fukuzumi T Saito T Iwata Y Kumamoto and A IsogaildquoTransparent and high gas barrier films of cellulose nanofibersprepared by TEMPO-mediated oxidationrdquo Biomacromoleculesvol 10 no 1 pp 162ndash165 2009

[14] I Siro D Plackett M Hedenqvist M Ankerfors and TLindstrom ldquoHighly transparent films from carboxymethylatedmicrofibrillated cellulose the effect ofmultiple homogenizationsteps on key propertiesrdquo Journal of Applied Polymer Science vol119 no 5 pp 2652ndash2660 2011

[15] D Klemm F Kramer S Moritz et al ldquoNanocelluloses a newfamily of nature-based materialsrdquo Angewandte ChemiemdashInter-national Edition vol 50 no 24 pp 5438ndash5466 2011

[16] C Chang and L Zhang ldquoCellulose-based hydrogels presentstatus and application prospectsrdquo Carbohydrate Polymers vol84 no 1 pp 40ndash53 2011

[17] H Jin M Kettunen A Laiho et al ldquoSuperhydrophobic andsuperoleophobic nanocellulose aerogel membranes as bioin-spired cargo carriers on water and oilrdquo Langmuir vol 27 no5 pp 1930ndash1934 2011

[18] H Sehaqui M Salajkova Q Zhou and L A Berglund ldquoMech-anical performance tailoring of tough ultra-high porosity foamsprepared from cellulose I nanofiber suspensionsrdquo Soft Mattervol 6 no 8 pp 1824ndash1832 2010

[19] K-Y Lee Y Aitomaki L A Berglund K Oksman and ABismarck ldquoOn the use of nanocellulose as reinforcement inpolymer matrix compositesrdquo Composites Science and Technol-ogy vol 105 pp 15ndash27 2014

[20] K Abe F Nakatsubo and H Yano ldquoHigh-strength nanocom-posite based on fibrillated chemi-thermomechanical pulprdquoComposites Science and Technology vol 69 no 14 pp 2434ndash2437 2009

[21] I Siro and D Plackett ldquoMicrofibrillated cellulose and newnanocomposite materials a reviewrdquo Cellulose vol 17 no 3 pp459ndash494 2010

[22] M Jonoobi A PMathewMM AbdiM DMakinejad and KOksman ldquoA comparison of modified and unmodified cellulosenanofiber reinforced polylactic acid (PLA) prepared by twinscrew extrusionrdquo Journal of Polymers and the Environment vol20 no 4 pp 991ndash997 2012

[23] F Ansari M Skrifvars and L Berglund ldquoNanostructured bio-composites based on unsaturated polyester resin and a cellulosenanofiber networkrdquoComposites Science and Technology vol 117pp 298ndash306 2015

[24] A Boldizar C Klason J Kubat P Naslund and P Saha ldquoPre-hydrolyzed cellulose as reinforcing filler for thermoplasticsrdquoInternational Journal of Polymeric Materials vol 11 no 4 pp229ndash262 1987

[25] J K Pandey A N Nakagaito and H Takagi ldquoFabrication andapplications of cellulose nanoparticle-based polymer compos-itesrdquo Polymer Engineering and Science vol 53 no 1 pp 1ndash82013

[26] T Moberg and M Rigdahl ldquoOn the viscoelastic properties ofmicrofibrillated cellulose (MFC) suspensionsrdquo Transactions ofthe Nordic Rheology Society vol 20 pp 123ndash130 2012

[27] A Naderi T Lindstrom and J Sundstrom ldquoCarboxymethy-lated nanofibrillated cellulose rheological studiesrdquo Cellulosevol 21 no 3 pp 1561ndash1571 2014

[28] L Jowkarderis and T G M van de Ven ldquoRheology of semi-dilute suspensions of carboxylated cellulose nanofibrilsrdquo Car-bohydrate Polymers vol 123 pp 416ndash423 2015

[29] A-H Vesterinen P Myllytie J Laine and J Seppala ldquoTheeffect of water-soluble polymers on rheology of microfibrillarcellulose suspension and dynamic mechanical properties ofpaper sheetrdquo Journal of Applied Polymer Science vol 116 no 5pp 2990ndash2997 2010

[30] A Karppinen A-H Vesterinen T Saarinen P Pietikainen andJ Seppala ldquoEffect of cationic polymethacrylates on the rheologyand flocculation of microfibrillated celluloserdquo Cellulose vol 18no 6 pp 1381ndash1390 2011

[31] M-P Lowys J Desbrieres andM Rinaudo ldquoRheological char-acterization of cellulosic microfibril suspensions Role of poly-meric additivesrdquo Food Hydrocolloids vol 15 no 1 pp 25ndash322001

10 Journal of Nanomaterials

[32] T Moberg M Rigdahl M Stading and E Levenstam BragdldquoExtensional viscosity ofmicrofibrillated cellulose suspensionsrdquoCarbohydrate Polymers vol 102 no 1 pp 409ndash412 2014

[33] I Sakurada Y Nukushina and T Ito ldquoExperimental determi-nation of the elastic modulus of crystalline regions in orientedpolymersrdquo Journal of Polymer Science vol 57 no 165 pp 651ndash660 1962

[34] N G McCrum C P Buckley and C B Bucknall Principles ofPolymer Engineering Oxford Science Publications Oxford UK1997

[35] S Iwamoto A Isogai and T Iwata ldquoStructure and mechanicalproperties of wet-spun fibers made from natural cellulosenanofibersrdquo Biomacromolecules vol 12 no 3 pp 831ndash836 2011

[36] J G Torres-Rendon F H Schacher S Ifuku and A WaltherldquoMechanical performance of macrofibers of cellulose and chitinnanofibrils aligned by wet-stretching a critical comparisonrdquoBiomacromolecules vol 15 no 7 pp 2709ndash2717 2014

[37] KMO Hakansson A B Fall F Lundell et al ldquoHydrodynamicalignment and assembly of nanofibrils resulting in strongcellulose filamentsrdquoNature Communications vol 5 article 40182014

[38] F Ansari S Galland M Johansson C J G Plummer and LA Berglund ldquoCellulose nanofiber network for moisture stablestrong and ductile biocomposites and increased epoxy curingraterdquo Composites Part A Applied Science and Manufacturingvol 63 pp 35ndash44 2014

[39] H Tang N Butchosa and Q Zhou ldquoA transparent hazy andstrong macroscopic ribbon of oriented cellulose nanofibrilsbearing poly(ethylene glycol)rdquo Advanced Materials vol 27 no12 pp 2070ndash2076 2015

[40] M Thunwall A Boldizar M Rigdahl et al ldquoProcessing andproperties of mineral-interfaced cellulose fibre compositesrdquoJournal of Applied Polymer Science vol 107 no 2 pp 918ndash9292008

[41] H A Barnes J F Hutton and K Walters An Introduction toRheology chapter 7 Elsevier Science Amsterdam The Nether-lands 1989

[42] P Gatenholm H Bertilsson and A Mathiasson ldquoEffect ofchemical composition of interphase on dispersion of cellulosefibers in polymers I PVC-coated cellulose in polystyrenerdquoJournal of Applied Polymer Science vol 49 no 2 pp 197ndash2081993

[43] S Takaichi T Saito R Tanaka and A Isogai ldquoImprovement ofnanodispersibility of oven-dried TEMPO-oxidized celluloses inwaterrdquo Cellulose vol 21 no 6 pp 4093ndash4103 2014

[44] J M Dealey and K F Wissbrun Melt Rheology and Its Role inPlastics Processing Van Nostrand Reihold New York NY USA1990

[45] H BertilssonOn the transition tomarked nonlinear viscoelastic-ity in solid polymer [PhD thesis] Royal Institute of TechnologyStockholm Sweden 1977

[46] H A BarnesAHandbook of Elementary Rheology vol 13 chap-ter 13 University of Wales Cambrian Printers AberystwythUK 2000

[47] F L Matthews and R D Rawlings Composite MaterialsEngineering and Science Chapman amp Hall London UK 1994