Effect of processing route on the composition and properties of hemp fibre

Fibers and Polymers 2008, Vol.9, No.5, 593-603

593

Effect of Processing Route on the Composition and Properties of Hemp Fibre

Sandra Korte* and Mark P. StaigerMaterials Engineering Group, Department of Mechanical Engineering, University of Canterbury,

Private Bag 4800, Christchurch, 8020, New Zealand(Received May 23, 2007; Revised June 22, 2008; Accepted August 25, 2008)

Abstract: There is great interest in the plant Cannabis sativa (hemp) as a source of technical fibres for the reinforcement ofpolymers in composite materials due to its high mechanical properties. In this work, the effect of enzymatic, hydrothermaland alkaline treatments on the composition and mechanical properties of hemp fibre are compared. The influence of enzymeconcentration and treatment time was examined (2.5-80 % Pectinex® Ultra SP-L, 6-48 hrs). Additionally, hydrothermal(170 oC, 10 bars) and alkaline treatments (18 wt. % NaOH, 40 oC) were used as pre-treatments to observe their effect on sub-sequent enzymatic treatment. The composition of hemp fibre was analysed by wet chemistry and Fourier transform infraredspectroscopy, while microstructure and mechanical properties were examined by scanning electron microscopy and tensiletesting, respectively. Enzymatic treatment resulted in extensive fibrillation and removal of non-cellulosic components, espe-cially when combined with hydrothermal treatment. However, a lengthy enzymatic treatment or combinative enzymatic-alka-line treatment led to extensive fibre breakdown that was accompanied by a pronounced reduction in the mechanicalproperties. Enzymatic treatment decreased Young’s modulus and tensile strength by 77 and 73 % respectively, and alkalinetreatment by 83 and 36 %. The hydrothermal treatment resulted in only minor changes in these properties.Keywords: Natural fibre, Hemp, Enzyme, Hydrothermal treatment, Alkaline treatment

Introduction

Mounting concerns for the environment have sparkedrenewed interest in the development of biodegradable,mechanically sound alternatives to synthetic fibres throughthe use of natural fibres found in the leaf or stalk (bast) ofplants such as hemp, flax, ramie and sisal [1]. Natural fibresthat are high in cellulose can be embedded in a polymermatrix to serve as an inexpensive, biodegradable and renewablealternative to glass fibres [2]. Firstly, natural fibres must beextracted from the leaf or stem and then broken down(fibrillated) further to produce suitable reinforcement forpolymer composites. It has been reported that natural fibre-reinforced polymer composites (also referred to as eco-composites or biocomposites) have specific strengths andstiffnesses comparable with their glass fibre counterparts[1,3-6]. Biocomposites have so far been selected for anumber of low load bearing industrial applications,especially in the automotive sector [7].

Natural fibres generally consist of bundles of individualcells (or primary fibres). The primary fibres transport waterand nutrients through the plant, facilitated by a hollowcentral canal (or lumen). There are four distinct regionssurrounding the lumen: the primary, secondary and tertiarycell walls, and the middle lamella, each varying incomposition, morphology and purpose [4]. The middlelamella at the exterior of the cell wall is comprisedpredominantly of lignified pectins that act to cement primaryfibres into bundles. Adjacent to the middle lamella is theprimary cell wall that consists of a disorganised arrangementof cellulose fibrils embedded in a matrix of pectin,

hemicellulose, lignin and proteins. Further in, the secondarycell wall, with the largest proportion of cellulose within thecell wall, consists of three layers of cellulose fibrils withvarying axial orientation that are bound by a matrix of ligninand hemicellulose. The basis for fibre processing is theremoval of lignified pectins from the middle lamella,thereby allowing the separation of the primary fibres fromtheir bundles. For composite applications, fibre treatmentsaim to improve the mechanical properties of the reinforcingfibre and enhance the adhesion between fibre and matrix,thus improving the interfacial strength and overall compositeperformance [8,9]. In particular, fibre treatments are able toincrease the surface area of the reinforcement available forchemical and/or mechanical bonding to a polymer matrix,thereby improving stress transfer within the final compositematerial [1,10-12].

Greater acceptance of biocomposites will hinge on thedevelopment of fibre treatment processes that areeconomically viable, environmentally sustainable and do notseverely degrade mechanical properties [13]. The life-cycleanalysis of biocomposites has shown that traditional fibreprocessing steps such as mercerization, water-retting andbleaching have negative environmental impacts due to thecopious consumption of water and polluting by-products[14,13,6]. The recyclability and degradability of enzymeshas stimulated research into their application to natural fibretreatment [15-23,11]. Enzymatic treatment has also beenshown to be a more reproducible processing route comparedwith the highly variable bacterial processes that occur duringfield retting [15,14].

Treatments in caustic soda have been adapted frommercerization of cotton and applied to natural fibres withdiffering process parameters, typically sodium hydroxide*Corresponding author: [email protected]

594 Fibers and Polymers 2008, Vol.9, No.5 Sandra Korte and Mark P. Staiger

(NaOH) concentration, treatment duration and temperatureand allowed fibre shrinkage [24,25,8,26-31]. Mwaikambo etal. [32] reported the strength and stiffness of hemp afteralkaline treatment to almost double with low NaOHconcentrations (0.16-0.24 %), while a decrease in theseproperties is typical with higher sodium hydroxide concen-trations (2-20 wt.%). After treatment with 20 % NaOH,Ouajai et al. [29] and Gassan et al. [25] observed a decreasein stiffness for hemp and jute fibre to 13 and 16 %,respectively, while the initial strength of the fibres decreasedby 30 and 70 %, respectively. The untreated fibre exhibitedelongation before break of ~2-2.5 %, while treatment withincreasing NaOH concentrations has reported increases inelongation of up to 7.5 % [33,34]. Degradation of the cellwall components and removal of amorphous non-cellulosicsare proposed to be possible causes for the observed decreasein stiffness and strength [8,29]. This is in contrast to Gassanet al. where stiffness and strength of jute was observed toincrease by ~250 and 300 %, respectively, after a 20 minutesisometric alkaline treatment at 25 wt.% [25]. It is proposed thatthe removal of amorphous components reduces the density ofthe interfibrillar volume, allowing alignment of the cellulosemicrofibrils in the direction of the applied tensile load (i.e.decreasing the microfibril angle).

Heat, steam (or hydrothermal) and steam explosiontreatments have also been found to be useful in fibrillatingnatural fibres. Rong et al. [35] and Yang et al. [34] observedthe strength and elongation at break of sisal to increase by10-30 % and 10-40 %, respectively, while the fibre stiffnessremained constant after a heat treatment at 150 oC in air.This heat treatment proved to be superior with regard tostrength and stiffness in comparison with a 2 % NaOHtreatment. The application of steam explosion has also beenshown to be highly effective [14,36-38]. For example,Kessler et al. [36] reported a 10 % increase in strength andelongation at break compared with the mechanicalseparation of flax fibre.

Various parameters of enzymatic treatment have beeninvestigated. The most influential parameters are known tobe the enzyme activity, enzyme concentration, treatmentduration and use of additives such as chelating agents (e.g.EDTA) [9,11,15-18,20,21,39,40-43]. Dreyer et al. [40]compared the mechanical properties of hemp after treatmentwith three different enzyme solutions with an alkalinetreatment in 0.4 % sodium carbonate at 100 oC. Theenzymatic treatments led to large variations in strength from<0.25 % to 150 % of the alkaline-treated fibre [40,20,29,9].Stuart et al. [9] observed a 30 % loss in strength andmarginal decrease in stiffness for flax fibre after treatmentwith a commercial enzyme solution. Shanks et al. [11]reported a distinct change in mechanical properties after 6hrs of enzymatic treatment with a reduction in strength andstiffness by 60 % while the elongation at break remainedconstant. The effect of enzyme “cocktails” on flax fibre can

decrease the strength by 30-50 % and reduce the elongationat break by up to 1.1 % depending on the treatment time andpre-soak in water [16,20].

A review of the literature on treatments of hemp fibresobviously does not permit a direct comparison of the appliedfibre treatments. Fibre material and methods differ substantiallybetween publications and mostly, if a comparison is made,only two treatment types are compared. The prospective ofthis work is therefore the characterization of fibres from asingle batch subjected to hydrothermal, alkaline andenzymatic treatments to accomplish a direct comparison oftheir respective effects. The treatments were applied inisolation and in combination with a special focus on theenzymatic processing route, where enzyme concentrationand treatment time were varied, in order to identify the mostinfluential treatment parameters on the composition andtensile properties of hemp fibre. The properties of the treatedfibres were evaluated using wet chemistry, Fourier-transform infrared (FTIR) spectroscopy, scanning electronmicroscopy (SEM) and tensile testing.

Experimental Methods

MaterialsHemp fibre (USO-31) was supplied by EcoFibre Industries

Pty. Ltd., Australia. Residual shivers were removed from theraw hemp using a laboratory scale carding machine. Acommercial enzyme solution (Pectinex® Ultra SP-L) wasobtained from Novozymes Australia Pty. Ltd. Pectinex®Ultra SP-L consists of a pectinase-carbohydrase complexmainly exhibiting pectolytic activity but with hemicellulosicand cellulosic side activities [44]. Pectinex® Ultra SP-L istypically used for the processing of fruit and vegetablemashes and maceration of plant tissues [44]. The enzyme isprepared from Aspergillus aculeatus with a standard activityof 26000 PG/ml measured by viscosity reduction at a pH of3.5 [45].

Fibre TreatmentsHydrothermal treatments were carried out in a rocking

autoclave in 50 g batches of fibre at 170 oC and 10 bars. Thefibres were thoroughly rinsed with water after hydrothermaltreatment. For the alkaline treatments, the fibres wereimmersed for 30 minutes in an 18 wt. % NaOH solution at40 oC and ultrasonically stirred to promote a homogeneoustreatment. After the alkaline treatment the fibres were rinsedwith 1 L of water per gram of fibre over a suction filter.Finally, the fibres were re-immersed in water, resulting in aslightly alkaline solution which was then neutralized withdiluted hydrochloric acid.

Enzymatic treatments were carried out under optimalconditions of pH 4.5 and 36 oC for the selected enzymes[45]. A solution to fibre ratio of 20:1 (wt/wt) was maintainedfor all experiments. The enzymatic treatments were carried out

Composition and Properties of Hemp Fibre Fibers and Polymers 2008, Vol.9, No.5 595

for 6, 12, 24 or 48 hrs. The enzymatic solution concentrationwas varied from 2.5 to 80 %. Finally, the resulting fibreswere rinsed in 1 L of water per gram of fibre.

Isolated hydrothermal, isolated alkaline and combinativehydrothermal-alkaline treatments were also carried out aspre-treatments to the above enzymatic treatments. A controlsample was prepared by enzymatic treatment for 48 hrswithout any pre-treatment. All of the resulting fibres afterthese treatments were then dried overnight at 50 oC.

In the following the treated fibres will be referred to by theabbreviations given in Table 1.

Fibre CharacterizationFTIR spectroscopy was carried out with a Shimadzu

FTIR-8201PC in transmission mode. Sample preparationinvolved firstly grinding the fibre and passing it through a 45μm mesh. Discs were then pressed from 2 mg of fibre mixedwith 200 mg of KBr, dried overnight at 50 oC in a vacuumoven and then analyzed immediately. A resolution of 4 cm−1

and total of 16 scans per sample were used to obtain spectrain the mid-infrared region from 400-4000 cm−1. A Happ-type apodization function was applied to the spectra tominimise the noise to signal ratio. The spectra were thensubjected to a linear baseline correction between the minimanear 1900 and 800 cm−1 and normalized in the range of 400-1900 cm−1.

The cellulose, hemicellulose and lignin contents of the as-received and treated fibres were determined by wetchemistry according to AOAC Method 973.18 using ~5 g offibre per sample. The remaining amorphous components(i.e. waxes, pectin and water-extractives) were groupedtogether as they could not be easily separated by theanalysis. This group of components was referred to as the‘residuals’ in this work.

A Perkin Elmer Diamond Dynamic Mechanical Analyser(DMA) was used in static mode to measure the tensileproperties of the hemp fibres based on ASTM standardD3822-01 [46]. The fibres were glued onto strips of paper ateither end of a 5 mm hole using an epoxy resin of adequateviscosity to ensure a smooth load transfer at the edgeswithout glue being pulled into the tested fibre lengthfollowing capillary forces. Great care was taken not to touchthe fibres over the tested length and to attach them straightand without slack exactly at the edges of the paper in orderto guarantee a constant gauge length of 5 mm. Thespecimens were then clamped into the DMA and the sides of

the hole cut carefully such that no load was placed onto thefibre. All tests were performed at a constant rate of 50micron/min. The fibre diameter was measured using areflected light microscope and the area calculated assuminga circular cross-section for single fibrils or a rectangularcross-section for fibre bundles. Imaging of each specimenfrom several orientations was used to confirm that singlefibrils had a circular cross-section obtain measurementsfrom two perpendicular directions, which were averaged togive the cross-sectional area. Fibre bundles usually consistedof fibrils interconnected within one plane, giving them arectangular cross-section with the thickness correspondingto the average fibre diameter. A minimum of 5-10 tensiletests were performed for each type of fibre treatment andspecimens that fractured at the possible stress concentrationnear the glue were discarded. Due to the difference in fibrediameter as a result of the treatments, the diameters availablefor testing were not uniform. For each sample the full rangeof diameters present after the treatments was tested wherefibres of sufficient length (~9 mm) were available. Theimpact of a change in diameter on fibre properties isdiscussed in 3.4. Enzyme treatment, especially after alkalinetreatment, led to a significant reduction in fibre length, sothat the thinnest fibrils were not available at a sufficientlength. A reduction in gauge length to the necessary lengthof ~1 mm to test these fibres was deemed impractical as anyerrors associated with the positioning and fixation of thefibres would be become significant relative to the testedlength. These fibres could therefore not be characterizedindividually with the available equipment.

Scanning electron microscopy (JOEL JSM-6100) wasused to examine the surface topology of the fibres aftertreatment with an accelerating voltage of 4 keV. Hemp fibrewas gold-coated for observations by SEM.

Results and Discussion

Effect of Enzyme Concentration on Fibre Mass LossThe influence of enzyme concentration (diluted by 2.5 to

80 %) on fibre mass loss was first assessed by monitoringthe weight loss after a treatment time of 24 hrs. A controlsample was placed in distilled water to identify any weightlosses attributable to dissolution in water or processing stepssuch as filtering or rinsing.

Enzyme kinetics under conditions of enzyme excess havebeen modelled by modifying the Michaelis-Menten equationfor substrate excess [47,48]:

where V is the velocity of the catalyzed reaction, E theenzyme concentration, KE the half-saturation constant andVmax the maximum velocity. Calafell and Garriga havesuggested that V can be related to the weight loss [39]. Thus,

V E( ) VmaxE

KE E+---------------=

Table 1. Abbreviations used to identify the treatments applied todifferent fibre samples

A Alkaline treatmentTh Hydrothermal treatment

6, 12, 24, 48 Enzymatic treatments, 6, 12, 24, 48 hrse.g. ATh24 Alkaline, hydrothermal and 24 hrs enzymatic treatment


the specific weight loss after 24 hrs treatment was measuredto determine the enzyme concentration that yields amaximum in the reaction rate (Figure 1). An enzymeconcentration of 20 % was chosen as a compromise for bothmaximising the reaction rates while minimising the usage ofenzymes. Careful collection of the enzymes after treatmentallowed recycling of the enzyme solution.

Effect of Treatment on Hemp Fibre MorphologyThe raw hemp exhibited a light brown colour with fibre

bundles measuring ~20-300 μm in diameter and severalcentimetres in length (Figure 2 a). The hydrothermaltreatment yielded darker brown fibres, exhibiting only slightfibrillation. The alkaline treatment resulted in yellowishcoloured fibres that exhibited some fibre curling (Figure 3).The alkaline treatment did not result in appreciable changesto the fibre dimensions according to a visual inspection. Incontrast, enzymatic treatment led to extensive fibrillationcompared with the other treatments. However, the enzymatictreatments did not achieve homogeneous fibrillation withineach batch. Prolonged enzymatic treatments led to theextensive breakdown of hemp fibre into ~1 mm lengths(Figure 2 b). Fibrillation and fibre shortening due to anenzymatic treatment was more pronounced when an alkalinepre-treatment was used (Figure 2 c). A hydrothermal pre-treatment followed by a short enzymatic treatment (6 hrs)yielded optimal results in terms of the fibres’ aspect ratio.

Scanning electron microscopy enabled a more detailedanalysis of the fibre morphologies. The surfaces of untreatedfibre bundles were found to be coated with smooth layersand agglomerates (Figure 4 a) that were partially removedby the hydrothermal treatment (Figure 4 b). An alkalinetreatment also resulted in degradation of these surface

Figure 1. Specific weight loss of hemp fibre as a function ofenzyme concentration.

Figure 2. Photographs of (a) untreated hemp fibre, (b) after 48 hrs enzymatic treatment, and (c) after the combined alkaline pre-treatment and24 hrs enzymatic treatment.


Figure 3. Scanning electron micrographs of fibres damaged as a result of an aggressive alkaline treatment at (a) low and (b) highmagnification.

Figure 4. Scanning electron micrographs of fractured fibres from tensile testing of (a) untreated fibres and fibres prepared by (b)hydrothermal treatment, (c) alkaline treatment, (d) 48 hrs enzymatic treatment, (e) combinative hydrothermal-alkaline treatment, and (f)combinative hydrothermal-alkaline-24 hrs enzymatic treatment.


layers; although, in contrast to the hydrothermally-treatedfibres, separation of the fibre bundles was increasinglyobserved (Figure 4 c). The fibre curling observed after alkalinetreatment is thought to be due to the shrinkage of cellulosefibrils that occurs on transformation from Type I to Type IIcellulose (Figure 3) [18,3,26]. Jahn [26] and Mwaikambo[27] have observed similar morphology changes after thealkaline treatment of different natural fibres. The enzymatictreatments also resulted in degradation of the non-fibroussurface layers, which resulted in partial flaking at the fibresurface (Figure 4 d).

Effect of Treatment on Fibre CompositionThe wet chemistry analysis revealed that the lignin content

remained at similar levels for most of the treatments, withenzymatic treatments resulting in the lowest values of theisolated treatments (Figure 5). The hydrothermal treatmentwas the most effective in removing hemicellulose. Theenzymatic treatment yielded the next lowest hemicellulosecontent at ~8 %. The alkaline treatment left the highestresidual hemicellulose content of ~10 %. Combinativealkaline-enzymatic or combinative hydrothermal-enzymatictreatments resulted in effective removal of hemicellulose,with a minimum of 4.35 % observed for the combinativehydrothermal-enzymatic treatment.

The combinative hydrothermal-enzymatic and hydrothermal-alkaline treatments exhibited the highest reduction of non-cellulosic components. The combinative hydrothermal-alkaline treatment led to the most extensive reduction of theresidual components while combinative treatments involvingan enzymatic treatment had a larger impact on thehemicellulose and lignin fractions. None of the treatmentswere able to reduce the content of non-cellulosic material tothis extent when used in isolation, suggesting that thecombinative treatments are synergistic in nature. Theenzymatic treatment achieved the most extensive removal ofresidual components (24 % down to ~10 %) of the threeisolated treatments.

Interestingly, the enzymatic treatment was less effective inremoving the residual components if applied after analkaline pre-treatment. The removal of hemicellulose duringthis combinative treatment does, however, indicate enzymaticactivity. Thus, it appears that the alkaline pre-treatmentaffects the structure of the residual components such that theenzymes cannot attack these components due to their highsubstrate-specificity. FTIR spectroscopy suggests that de-esterification of pectin is a possible cause of this observation.

FTIR spectra of untreated and treated hemp fibre areshown in Figure 6. Overlapping peaks from all of the majorcomponents are expected in the spectral window of 1000-1200 cm−1. However, cellulose is the main contributor due toits high content. Different changes to the spectrum withinthis region were observed depending on the fibre treatmentused. The hydrothermal and enzymatic treatments resultedin distinct peaks at 1090 and 1115 cm−1, while the alkalinetreatment resulted in a broadened spectrum with undefinedpeaks. The formation of slightly stronger local extremacould be observed with longer enzymatic treatments.According to Windeisen et al. [49], the observed peak

Figure 5. Composition of fibre samples as a function of the treatment (as determined by wet chemistry).

Figure 6. FTIR spectra of (a) untreated fibres, after (b) enzymatic,(c) hydrothermal, and (d) alkaline treatment.


sharpening can be explained by increasing depolymerisationof the polysaccharides that enables easier stimulation ofabsorption bands. A decrease in the degree of polymerizationof cellulose as a result of enzymatic treatment has also beenreported by Cao et al. [50].

Another noticeable change in the spectra is a change incrystallinity. The lateral order index (LOI) was calculated asthe ratio between the heights of the baseline-correctedabsorption peaks at 1430 and 897 cm−1 - peaks that correspondto sensitive absorption frequencies of crystalline andamorphous cellulose structures, respectively [51,28,11]. LOIdescribes the order of crystallinity rather than the amount ofcrystalline cellulose in relation to amorphous components.Table 1 shows the LOI calculated for hydrothermal, alkalineand enzymatic treatment with a conversion into crystallinityas determined by x-ray diffraction according to Wakelin [52]and Nelson [53]. However, the LOI is only of limitedsignificance for alkaline-treated fibres, due to the highspectral similarity of amorphous and Type II cellulose [51].Similar changes in the LOI after enzymatic treatments havebeen observed by Shanks et al. [11].

For mixed cellulose I and II lattice types Nelson andO’Connor [53] suggested the crystallinity index, CI, as theratio of the absorbances at 1372 cm−1 and 2900 cm−1 todetermine the level of crystallinity. The calculated values forthis ratio are also shown in Table 2. The raw fibre material

possesses the smallest relative amount of crystalline celluloseas revealed by the CI. However, as no treatments wereemployed that specifically change the crystalline order, e.g.axial obstruction of shrinkage, the raw fibre exhibited theundisturbed crystalline cellulose fraction with the highestobserved LOI. Both CI and LOI show similar values for thehydrothermal and enzymatic-treated fibres with enzymatictreatment resulting in slightly higher corresponding x-raycrystallinity. The large discrepancy between CI and LOI forthe alkaline-treated fibre was due to the reasons given aboveand since only the CI is applicable to mixed lattice types [53].In accordance with the CI the observed peak shift from 1426to 1430 cm−1 suggests preferential degradation of amorphouscellulose during enzymatic and hydrothermal treatment,which should lead to a higher crystallinity of the remainingfibre [54,11].

The broad absorbance band near 1740 cm−1 (Figure 6) isassigned to C=O carbonyl stretching in carboxylic groups[55] that mostly occur in the branched chain hemicelluloses[56] and esterified and carboxylic groups in pectin [11].Thus, the decrease of absorbance near 1740 cm−1 impliesthat de-esterification of the pectin is occurring during thealkaline treatment [57]. There is an increase in thecorresponding absorbance of anti-symmetric COO−-stretching around 1640 cm−1 [11] after the alkaline treatmentin accordance with the wet chemistry results (see Figure 5),

Table 2. Crystallinity index (CI) and lateral order index (LOI) of the untreated and treated fibre samples and the correlating x-ray crystallinityindexes determined according to Wakelin and Nelson [53,52]

LOI (a1429 cm−1/a897 cm−1) % CI (a1372 cm−1/a2900 cm−1) %Raw 4.02 90 0.51 30Hydrothermal (Th) 3.03 65 0.63 55Hydrothermal + 6 hrs enz. (Th6) 2.76 55 0.68 65Hydrothermal + 24 hrs enz. (Th24) 3.09 70 0.75 8048 hr Enzymatic (48) 3.39 75 0.73 80Hydroth + alk + 48 hrs enz (ThA48) 1.14 10 0.65 65Hydrothermal + alkaline (ThA) 0.93 5 0.66 60Alkaline (A) 0.99 5 0.57 40

LOI a1429 cm−1/a897 cm−1 CI a1372 cm−1/a2900 cm−1


indicating a higher remaining pectin content in the residualcomponent after this treatment.

Effect of Treatment on Tensile PropertiesFor all samples, the stiffness and strength were found to

increase with decreasing fibre diameter as also found byBaley [58]. This dependence of tensile strength on diameteris shown for alkaline treated fibres in Figure 7. The alkalinetreatment was found to be very aggressive and defibrillationand ultimate breakdown of the fibres made it impossible tomechanically test equivalent diameters for all samples.Therefore the maximum diameter tested after alkaline andhydrothermal/alkaline treatment was 10 μm while thediameters of all other samples ranged between 20 and 70 μm.Linear fits to the plots of tensile strength versus diametergave a very similar dependence on diameter for all sampleswith an increase of approximately 90 MPa per 10 μmreduction in diameter. Extrapolation of the tensile strengthtowards the smaller diameters yields the average tensilestrength found for the thinner fibres. However, the gradientshown in Figure 7 is steeper and therefore the size effect ismore pronounced. Whether this is a genuine effect caused bythe treatment conditions or associated with the increasedrelative uncertainty of the fibre diameter, decreaseduniformity as the size of fibrils is reduced, or the relative sizeof the lumen will be an aspect of future work. In spite of thisdifference in fibre diameter, a qualitative comparison andcorrelation of mechanical properties and treatment conditionsis feasible. In addition, the recorded data will be used as areference of fibre properties for future investigations onshort fibre composites. These can take into account the fullrange of fibres in each sample as the fibre length can be keptuniform and adjusted to the length of the shortest fibresavailable after treatment. However, although there is, in

contrast to single fibre tensile tests, no need to exclude theshort fibres produced by the enzymatic treatments in suchcomposite measurements, it is vital to initially obtain purefibre properties in order to assess additional factors such asinterfacial adhesion.

The absolute and relative values of Young’s modulus,tensile strength and elongation at failure along with therespective averaged diameters, as investigated by tensiletesting and optical microscopy, are displayed in Figure 8 andTable 3.

An isolated hydrothermal treatment changed the fibreproperties only marginally in terms of the tensile strength,stiffness and elongation at break in accordance with the onlyslight increase found by Rong et al. [35] and Kessler et al.[36]. The alkaline treatment was found to significantlyweaken hemp fibre. The tensile strength and Young’smodulus were reduced by ~36 % and 73 %, respectively,while the elongation at failure was observed to increase by220 %. Similar results have been reported for alkalization byGassan [3,25] and others [33,29,34].

Enzymatic treatment also had a negative influence on themechanical properties of the tested hemp fibres as has beenreported for hemp [40,11] and flax in the literature [16,20,9].A 48 hrs enzymatic treatment reduced the Young’s modulusand tensile strength by ~77 % and 83 %, respectively, Thefibres prepared by the combinative hydrothermal-enzymatictreatment exhibited a significant drop in mechanicalproperties after the enzymatic treatment, especially in thefirst few hours of treatment. Similarly, Shanks et al. haveobserved a pronounced initial drop in strength and stiffnesswith enzymatic treatment (Scourzyme® L) [11].

A superposition of the influences was observed in samplesexposed to the combinative alkaline-enzymatic treatment,leading to considerable reduction of mechanical properties.As was the case for isolated hydrothermal treatment, acombination with alkaline or enzymatic treatments did notsignificantly alter the mechanical properties in comparison tothe isolated application of alkaline or enzymatic treatments.

Figure 7. Tensile strength and Young’s modulus of the alkalinetreated fibres as a function of the fibre diameter. A similardistribution was found for all treatments.

Table 3. Summary of mechanical properties of hemp fibre aftervarious treatments (standard deviations are shown).

Fibretreatment

Young’smodulus

(GPa)

Tensile strength(MPa)

Elongationat break

(%)As-received 51.4±21.6 890±440 1.9±0.6Hydrothermal (Th) 54.7±27.7 904±374 2.1±1.048 hrs enzymatic 11.8±6.4 147±100 2.4±1.7Alkaline (A) 14.0±7.1 566±314 6.1±2.9Th6 24.9±6.6 244±83 1.3±0.2Th24 21.±9.5 93±8 0.7±0.3ThA 15.5±9.1 430±231 4.2±1.7ThA48 8.9±6.7 63±28 1.9±0.9


As-received hemp fibre and fibres exposed to hydro-thermal and enzymatic treatments generally exhibited brittlefailure, with fracture usually occurring along a single planeperpendicular to the fibre axis. The samples exposed to the

enzymatic treatments showed a tendency to form flakes atthe bundle surface and in between the elementary fibres (seeFigure 4 d). Similar observations after enzymatic treatmentof hemp fibres have been made by Buschle-Diller et al. [18].

Figure 8. Effect of fibre treatment on averaged mechanical properties. All values are normalized.

Table 4. Overview of obtained properties

Hydrothermal Alkaline Enzymatic Hydrothermal-enzymatic Alkaline-enzymaticTensile strength 0 - -- -- --Young’s modulus 0 -- − -/-- --Elongation 0 ++ 0 − 0Cellulose content + − + ++ −*0: little or no change, +: increase, and −: decrease.


Alkaline treatment resulted in an increased fragmentation offibre bundles and spatially more separated failure of theelementary fibres lengthwise along the main fibre axis(Figure 4 c, e, and f).

Although a general improvement of fibre strength andmodulus is expected with increasing crystalline cellulosecontent, in the present work no direct connection could bemade between cellulose content, its crystallinity andmechanical properties as can be seen in a comparison of thediagrams shown in Figure 8 and Table 2. However, theinfluence of the amount of crystalline cellulose might be ofonly inferior importance with regard to single fibre tenacityin comparison to the overall compositional changes anddisruption of the fibre’s natural composite structure.Separate work with a special focus on crystallinity would benecessary to identify its influence in correlation with eachprocessing step.

Conclusions

Hemp fibre was subjected to hydrothermal, alkaline andenzymatic treatments both in isolation and as combinativetreatments. The treated fibres were then analysed withrespect to their composition, morphology and mechanicalproperties. The main findings can be summarized as follows(see also Table 4):

• The employed hydrothermal treatment resulted in areduction of non-cellulosic material from 40 to 20 wt. %without negatively affecting the fibres’ mechanicalproperties;

• Treatment in sodium hydroxide led to dissolution of non-cellulosic components accompanied by transformationof the cellulose lattice type and de-esterification of pectinas observed by FTIR spectroscopy. The treatment did,however, weaken the fibre structure, resulting in areduction of mechanical properties;

• Enzymatic treatments with Pectinex® Ultra SP-Lproduced very effective fibrillation compared to thesteam and alkaline treatments. However, the fibrillationwas accompanied by a loss of mechanical strength andstiffness with longer treatment times;

• While combinative treatments generally exhibitedcomplementary mechanisms of action, negative effectswere observed for the combinative alkaline-enzymatictreatment, resulting in decreased release of non-cellulosic components, amplified fibre breakdown intovery short chips and consequently decreased mechanicalproperties;

The relative effects of the isolated and combinativeenzymatic treatments on the hemp fibre properties aresummarized in Table 4 (0 = little or no change, − decrease, +increase).

Acknowledgments

This work was financially supported by the Brian MasonScientific & Technical Trust and University of CanterburyGrant U6513. One of the authors (SK) would like to thank Dr.-Ing. Dieter Veit (Institut für Textiltechnik, RWTH Aachen)for his support and The German Academic ExchangeService for granting a scholarship. The authors would like tothank Mr. P. Warner for his support and the N. Veltre ofEcoFibre Industries Limited for supplying raw hemp fibre,and Dr. S. Jackson, Mr. D. Brown, Mr. T. Berry, Mr. R.Thompson, Mr. K. Stobbs and Mr. M. Flaws for technicalassistance.

References

1. A. K. Bledzki and J. Gassan, Prog. Polym. Sci., 24, 221(1999).

2. S. J. Eichhorn, C. A. Baillie, N. E. Zafeiropoulos, L. Y.Mwaikambo, M. P. Ansell, A. Dufresne, K. M. Entwistle,P. J. Herrera-Franco, G. C. Escamilla, L. Groom, M.Hughes, C. Hill, T. G. Rials, and P. M. Wild, J. Mater. Sci.,36, 2107 (2001).

3. J. Gassan, Ph. D. Dissertation, Universität-GH, Kassel,1997

4. P.-A. Koch, Faserstofftabellen nach, 1994.5. T. Peijs and C. A. Baillie, Compos. Sci. Technol., 63,

(2003).6. J. Sponner, L. Toth, S. Cziger, and R. R. Franck in “Bast

and Other Plant Fibres” (R. R. Franck Ed.), p.198, 2005.7. Nova-Institut für Ökologie und Innovation, Nachwachsende

Rohstoffe - Industrial Crops, 2003.8. J. Gassan and A. K. Bledzki, Compos. Sci. Technol., 59,

1303 (1999).9. T. Stuart, Q. Liu, M. Hughes, R. D. McCall, H. S. S.

Sharma, and A. Norton, Composites Part A, 37, 393 (2006).10. H.-R. Kymalainen, M. Koivula, R. Kuisma, A.-M. Sjoberg,

and A. Pehkonen, Bioresour. Technol., 94, 57 (2004).11. R. A. Shanks and S. Ouajai, Macromol. Biosci., 5, 124

(2005).12. B. Wielage, T. Lampke, and A. Bismarck, Verbundwerkstoffe

und Werkstoffverbunde, (2001).13. B. Lipp-Symonowicz, B. Tanska, and A. Sapieja, Fibres &

Textiles in Eastern Europe, 12, 63 (2004).14. C. Garcia-Jaldon, D. Dupeyre, and M. R. Vignon, Biomass

Bioenergy, 14, 251 (1998).15. D. E. Akin, J. A. Foulk, R. B. Dodd, and D. D. McAlister,

J. Biotechnol., 89, 193 (2001).16. D. E. Akin, I. Morrison, W. Herbert, L. L. Rigsby, J. D.

Evans, and J. A. Foulk, Industrial Crops and Products, 17,149 (2003).

17. F. Bruhlmann, M. Leupin, K. H. Erismann, and A. Fiechter,J. Biotechnol., 76, 43 (2000).


18. G. Buschle-Diller, C. Fanter, and F. Loth, Text. Res. J., 69,244 (1999).

19. Y. Cao and H. Tan, J. Mol. Struct., 705, 189 (2004).20. J. D. Evans, D. E. Akin, and J. A. Foulk, J. Biotechnol.,

97, 223 (2002).21. J. A. Foulk, D. E. Akin, and R. B. Dodd, Industrial Crops

and Products, 13, 239 (2001).22. G. Henriksson, D. E. Akin, D. Slomczynski, and K. -E. L.

Eriksson, J. Biotechnol., 68, 115 (1999).23. S. D. Mansfield, E. D. Jong, R. S. Stephens, and J. N.

Saddler, J. Biotechnol., 57, 205 (1997).24. S. H. Aziz and M. P. Ansell, Compos. Sci. Technol., 64,

1219 (2004).25. J. Gassan and A. K. Bledzki, J. Appl. Polym. Sci., 71, 623

(1999).26. A. Jahn, M. W. Schroder, M. Futing, K. Schenzel, and W.

Diepenbrock, Spectrochim. Acta, Part A, 58, 2271 (2002).27. L. J. Mwaikambo and M. P. Ansell, J. Appl. Polym. Sci.,

84, 2222 (2002).28. S. Y. Oh, D. I. Yoo, Y. Shin, and G. Seo, Carbohydr. Res.,

340, 417 (2005).29. S. Ouajai, A. Hodzic, and R. A. Shanks, J. Appl. Polym.

Sci., 94, 2456 (2004).30. S. Ouajai and R. A. Shanks, Polym. Degrad. Stab., 89, 327

(2005).31. I. Van de Weyenberg, J. Ivens, A. De Coster, B. Kino, E.

Baetens, and I. Verpoest, Compos. Sci. Technol., 63, 1241(2003).

32. L. Y. Mwaikambo, J. Mat. Sci., 41, 2483 (2006).33. Y. Li, Y.-W. Mai, and L. Ye, Compos. Sci. Technol., 60,

2037 (2000).34. G. C. Yang, H. M. Zeng, J. J. Li, N. B. Jian, and W. B. Zhang,

Acta Scientiarum Naturalium Universitatis Sunyatseni, 35,53 (1996).

35. M. Z. Rong, M. Q. Zhang, Y. Liu, G. C. Yang, and H. M.Zeng, Compos. Sci. Technol., 61, 1437 (2001).

36. R. W. Kessler, U. Becker, R. Kohler, and B. Goth, BiomassBioenergy, 14, 237 (1998).

37. L. P. Ramos, Quim. Nova, 26, 863 (2003).

38. M. R. Vignon, D. Dupeyre, and C. Garcia-Jaldon, Bioresour.Technol., 58, 203 (1996).

39. M. Calafell and P. Garriga, Enzyme Microb. Technol., 34,326 (2004).

40. J. Dreyer, J. Müssig, N. Koschke, W.-D. Ibenthal, and H.Harig, J. Industrial Hemp, 7, 43 (2002).

41. B. Lipp-Symonowicz, B. Tanska, A. Wolukanis, and H.Wrzosek, Fibres & Textiles in Eastern Europe, 12, 61(2004).

42. M. Ossola and Y. M. Galante, Enzyme Microb. Technol.,34, 177 (2004).

43. J. Zhang, G. Henriksson, and G. Johannson, J. Biotechnol.,81, 85 (2000).

44. Novozymes Switzerland AG, Corporate Homepage, 2005.45. Novozymes Switzerland AG, Product Sheet Pectinex®

Ultra SP-L, 2005.46. ASTM International, D 3822-01, Standard Test Method

for Tensile Properties of Single Textile Fibers, 2001.47. C. J. Bailey, Biochem. J., 262, 1001 (1989).48. B. Nidetzky, W. Steiner, M. Hayn, and M. Claeyssens,

Biochem. J., 298, 705 (1994).49. E. Windeisen, C. Strobel, and G. Wegener, Holz als Roh-

und Werkstoff, 61, 471 (2003).50. Y. Cao and H. Tan, Carbohydr. Res., 337, 1291 (2002).51. M. L. Nelson and R. T. O'Connor, J. Appl. Polym. Sci., 8,

1311 (1964).52. J. H. Wakelin, H. S. Virgin, and E. Crystal J. Appl. Phys.,

30, 1654 (1959).53. M. L. Nelson and R. T. O'Connor, J. Appl. Polym. Sci., 8,

1325 (1964).54. C. Y. Liang and R. H. Marchessault, J. Polym. Sci., 39, 269

(1959).55. A. Synytsya, J. Copikova, P. Matejka, and V. Machovic,

Carbohydr. Polym., 54, 97 (2003).56. K. K. Pandey, J. Appl. Polym. Sci, 71, 1969 (1999).57. R. Sun, J. M. Fang, L. Mott, and J. Bolton, Holzforschung,

53, 253 (1999).58. C. Baley, Composites Part A, 33, 939 (2002).

Effect of processing route on the composition and properties of hemp fibre

Documents