Thermoset Phenolic Matrices Reinforced with Unmodified and … · 2016. 6. 29. · an interaction of physical, chemical, or physical-chemical nature between the two components.15

Thermoset Phenolic Matrices Reinforced with Unmodified andSurface-Grafted Furfuryl Alcohol Sugar Cane Bagasse and

Curaua Fibers: Properties of Fibers and Composites

W. G. Trindade,†,‡ W. Hoareau,‡ J. D Megiatto,† I. A. T. Razera,† A. Castellan,*,‡ andE. Frollini*,†

Instituto de Quımica de Sao Carlos, Universidade de Sao Paulo, USP, C.P. 780, CEP 13560-970,Sao Carlos, SP, Brazil, and Universite Bordeaux 1, Laboratoire de Chimie des Substances Vegetales

EA-494, F-33405 Talence Cedex, France

Received February 24, 2005; Revised Manuscript Received June 12, 2005

Composites based on phenolic matrices and unmodified and chemically modified sugar cane bagasse andcuraua fibers were prepared. The fibers were oxidized by chlorine dioxide, mainly phenolic syringyl andguaiacyl units of the lignin polymer, followed by grafting furfuryl alcohol (FA), which is a chemical obtainedfrom a renewable source. The fibers were widely characterized by chemical composition analysis, crystallinity,UV-vis diffuse reflectance spectroscopy, SEM, DSC, TG, tensile strength, and13C CP-MAS NMR. Thecomposites were analyzed by SEM, impact strength, and DMA. The SEM images and DMA results showedthat the oxidation of sugar cane bagasse fibers followed by reaction with FA favored the fiber/matrixinteraction at the interface. The same chemical modification was less effective for curaua fibers, probablydue to its lower lignin content, since the reaction considered touches mainly the lignin moiety. The tensilestrength results obtained showed that the fibers were partially degraded by the chemical treatment, decreasingthen the impact strength of the composites reinforced with them. In the continuity of the present project,efforts has been addressed to the optimization of fiber surface modification, looking for reagents preferablyobtained from renewable resources and for chemical modifications that intensify the fiber/matrix interactionwithout loss of mechanical properties.

Introduction

In the 20th century, the extraordinary growth of theapplication of synthetic plastics limited the application ofvegetal fibers, because of the advantages of syntheticpolymers concerning dimensional stability and plasticity,among other factors. The large availability of vegetal fibersall over the world and their low cost, along with the intrinsicproperties of these materials, have led to the search foralternative applications for these fibers besides the traditionaluses such as textile, paper production, and fuel.1,2

The present and urgent need to develop and commercializecomposite materials based on constituents derived fromrenewable sources has had a great impact on the drive toreduce the dependence on nonrenewable materials derivedfrom fossil sources, both from environmental and economicviewpoints.3,4 CO2 emission associated with anthropogenicactivities has reached levels that, among other things, havemade the interest in natural fibers, which had been consideredin the first decades of the 20th century, reappear.5-9

Several enterprises have started to use composites rein-forced with vegetal fibers, as, for example, those in theautomotive industry. Natural fibers are very efficient in sound

absorption, and in comparison to fiberglass, they are resistantto breakage into shards, have a lower cost, are lighter andbiodegradable, and can be obtained using 80% less energy.4

One of the aspects that must be considered in thediscussion of composites reinforced with vegetal fibersconcerns fiber supply regularity, as in many countries, woodis the main source for fiber supply. Annual plants have thedisadvantage of seasonal cropping over wood as fibersuppliers, which raises the need for subsequent cleaning,drying, storage, and other processes.10 In contrast, while atree takes years to grow, annual plants have a full cycle in12 months. The use of fibers for applications more sophis-ticated than the conventional ones requires process system-atization, from the plantation to the storage of the fibersextracted. Highly demanding markets, such as the automotiveindustry, have led the required modifications. Probably, withthe market growth, vegetal fibers with specific applicationswill be grown in areas distinct from those aimed at commonapplications,11 which in turn can decrease or even eliminatethe disadvantage previously mentioned.

With this in mind, fibers with short development cyclessuch as sisal,2 jute,12 sugar cane bagasse,13,14 and curaua2

have been considered as composite reinforcements. Thepresent work deals with the two last fibers.

In the mixture of two components of diverse chemicalnatures of any dimension or shape, the larger the contactarea (interfaces) among them, the greater the possibility of

* Authors for correspondence. E-mail: [email protected](A.C.); [email protected] (E.F.).

† Universidade de Saõ Paulo.‡ UniversiteBordeaux 1.

2485Biomacromolecules 2005,6, 2485-2496

10.1021/bm058006+ CCC: $30.25 © 2005 American Chemical SocietyPublished on Web 07/30/2005

an interaction of physical, chemical, or physical-chemicalnature between the two components.15 In all cases, theinteraction between the dispersed phase and the matrix phasedepends not only on the extension of the contact area, butalso on the affinity between the components. The affinitycan be intensified, for instance, either by physical or chemicaltreatments applied to the surface of the fibers.2

An important point to consider is that the reagents usedin chemical modifications cannot be too expensive, andideally, the modifications must involve a minimal numberof compounds obtained from nonrenewable sources. Re-cently, a new selective chemical modification of the surfaceof lignocellulosic natural fibers has been considered. Thismodification is based on the selective oxidation of guaiacyland syringyl units of lignin, generatingortho- and para-quinones able to react by Diels-Alder reaction with furfurylalcohol,14 that is commercially prepared by reduction offurfural, which in turn is obtained from agricultural residues.

To better understand this process at the molecular level, asimilar chemical modification was carried out with the ligninobtained by acidolysis from sugar cane bagasse and curaua,16

which were oxidized by ClO2 and reacted with furfurylalcohol. Similar reactions on sugar cane bagasse and curaualignins are described in the following sections. In the fieldof composites reinforced with natural fibers, it can beconsidered that most works deal with the preparation andcharacterization of thermoplastic matrix composites. Com-paratively few works have been made about thermosetmatrices, despite their large application, and among those,even fewer deal with phenolic thermoset matrices. The useof vegetal fibers as reinforcing agents of phenolic thermosetsreduces production costs, as the resin is substituted by fibers,and improves the properties of the thermoset, such as impactstrength. In the present work, phenolic thermoset matrices,reinforced with unmodified and modified sugar cane bagasseand curaua fibers, were prepared and characterized.

Experimental Section

Fiber Characterization. Humidity content was deter-mined according to ABNT NBR9656 (Associaçao Brasileirade Normas Tećnicas, the Brazilian association for technicalstandards), which consists of determining the percentagedifference between the initial weight of the sample (1.00 g)and after 4 h drying at 105°C.

Ash content was measured by considering the percentdifference between the initial weight of dried fiber of thesample and that after calcination for 4 h at 800°C.

Klason lignin content was evaluated following the TAPPIT13M-54 method, which is based on isolation of lignin afterpolysaccharide hydrolysis by concentrated sulfuric acid(72%). The samples (1.0 mg, previously dried) were macer-ated using 72% H2SO4 (15 mL) at room temperature andkept under these conditions during 24 h. After that, thematerial impregnated with sulfuric acid was transferred toan appropriate vessel, and 560 mL of distilled water wasadded. The system was then kept under reflux for 4 h. Theremaining solid (insoluble lignin and ash) was filtered(sintered glass funnel, no. 4, previously weighted), dried (105

°C, 24 h), and then weighed. The acid-insoluble lignincontent was calculated by considering the weight differencebetween the initial sample (fibers) and that of the filteredsolid and subtracting the ash content from the weight of thislast sample. The Klason-soluble lignin was determined byultraviolet spectroscopy, considering both the solution ob-tained after filtering off the insoluble lignin and a referenceconsisting of a sulfuric acid solution with the same acid finalconcentration of the soluble lignin solution (TAPPI T13M-54). Absorbance values were measured at 280 (A280) and 215(A215) nm. The concentration (C, g/L) of Klason-solublelignin was determined using the following equation, basedon the Lambert Beer law:

The holocellulose content was determined according toTAPPI T19m-54, which consists of a selective degradationof the lignin polymer. Sodium hypochlorite (2.5 g) andglacial acetic acid (1 mL) were added to an aqueoussuspension (120 mL put in an Erlenmeyer flask) of previ-ously dried and milled vegetal fibers (3 g). The system wascovered with an inverted Erlenmeyer flask and kept at 70°C, under magnetic stirring for 1 h. This sequence wasrepeated twice, that is, sodium hypochlorite (2× 2.5 g) andglacial acetic acid (2× 1 mL) were again added and thesystem kept under magnetic stirring for 1 h (2× 1 h). Then,after 3 h (3 × 1 h), the system was cooled to nearly 5°C,and the hollocellulose was filtered and washed exhaustivelywith water and methanol and then dried at 40°C, in avacuum oven, until constant weight.

For R-cellulose content determination, sodium hydroxidesolution (10 mL, 17.5%) was added to cellulose (1.0 g) atroom temperature. Then, the mixture was triturated for 8 min,and sodium hydroxide (10 mL, 17.5%) was added to themixture, which remained at rest for 20 min. Then, water (40mL) was added, and the solid residue was filtered and washedexhaustively with aqueous acetic acid and water. Thisremaining solid, consideredR-cellulose, was dried at 40°C,in a vacuum oven, until constant weight.17 If the alkalinesolution is kept at room temperature longer than 20 min, asusually described for wood, cellulose can also degrade,leading to erroneous results for hemicellulose content.

An average on three samples was done for all mentionedanalyses. The hemicellulose content was obtained by sub-tracting theR-cellulose part from the holocellulose content.

The crystallinity index, Ic, was determined by X-raydiffraction using a RIGAKU Rotaflex model RU-200Bdiffractometer operating at 40 kV, 20 mA, andλ (Cu KR)) 1540 Å. The crystallinity index was calculated using theBuschle-Diller and Zeronian equation: Ic) 1 - I1/I2,18

whereI1 is the intensity at the minimum of the crystallinepeak (18° < 2θ < 19°) andI2 is the intensity at its maximum(22° < 2θ < 23°).18

Scanning electron microscopy (SEM) was carried out ina Zeiss-Leica apparatus model 440, electron acceleration 20kV. The samples analyzed were covered with a thin layerof gold in a sputter-coat system.

C (g/L) )4.53× A215 - A260

300

2486 Biomacromolecules, Vol. 6, No. 5, 2005 Trindade et al.

Tensile strength was analyzed by considering fiber bundles15 mm long and nearly 0.5 mm diameter. A dynamicmechanical analyzer (DMA) model 2980, from TA Instru-ments, was used under the following conditions: 25°C, 1N/min to 18 N. A minimum of twenty samples was testedfor each material, and average values are reported in the nextsection.

Differential scanning calorimetry (DSC) analyses werecarried out using a Shimadzu DSC model 50. Samples ofapproximately 6.5-7.6 mg were placed in appropriate sealedpans and heated from 20 to 600°C, at 20°C/min, under N2

atmosphere (10 mL/min).Thermogravimetric (TG) analyses were carried out using

a Shimadzu model TGA-50TA. Samples of approximately7.5 mg were placed in appropriate pans and heated from 20to 800°C at 10°C/min, under N2 atmosphere (10 mL/min).

The UV-vis diffuse reflectance spectra were recorded ona Perkin-Elmer Lambda 18 spectrometer equipped with anintegrating sphere (Labsphere, 150 mm).

Solid-state 13C CP-MAS (cross polymerization-magicangle spinning) NMR spectra of unmodified and modifiedsugar cane bagasse were performed at room temperature ona Bruker DPX-400 NMR spectrometer (Bruker), using MASrates of 4 and 8 kHz, at a frequency of 100.61 MHz. Sampleswere packed in MAS 4 mm diameter zircon rotors. Chemicalshifts were relative to tetramethylsilane (TMS) used as anexternal standard. The acquisition time for all spectra wasset at 16 h (30 000 scans).

Oxidation of Bagasse and Curaua Fibers. 1. Oxidationof Bagasse Fibers. The sugar cane fibers (kindly providedby Mr. Bernard Siegmund, CIRAD, Reunion Island, France)and curaua (kindly given by Prof. Alcides Leaõ, DeptoCiencias Ambientais, UNESP, Botucatu, SP, Brazil) wereextracted (Soxhlet) with cyclohexane (Labsynth, Brazil, 99%)//ethanol (Labsynth, Brazil, 99%), 1:1 v/v, for 48 h and thenwith water for 24 h. The fibers were dried in an air-circulatedoven (60°C) until constant weight. The fiber extraction withcyclohexane/ethanol allows removal of waxes from the fibersurface, preventing weakening of the interactions at theinterface fiber matrix of these compounds. Oxidation of fibers(2 g) was performed with a ClO2-water solution (18 mL,1.88 mmol), prepared according to a procedure describedby Mark et al.,19 and acetic acid (0.5 mL) at 55°C. Afterreaction, the fibers, which turned yellow-red, were washedwith water until neutrality.

2. Reaction of Oxidized Bagasse and Curaua withFurfuryl Alcohol. The oxidized fibers (2 g), impregnatedwith furfuryl alcohol (Sigma-Aldrich, PA) (FA, 11.35 g),were heated at 100°C for 4 h. The excess of FA wasremoved by Soxhlet extraction using ethanol for 16 h. Then,the fibers were dried 24 h at 45°C, and weight gains due toreaction were determined on the basis of original and finaloven-dried fiber weights. They are reported as weight percentgain (WPG).

Prepolymer Synthesis.Phenolic prepolymer was synthe-sized by mixing phenol (Labsynth, Brazil, 99%), formalde-hyde (Labsynth, Brazil, 37%), and potassium hydroxide(Labsynth, Brazil, 85%), 1.38:1.00:0.06%, respectively, undermechanical stirring, at 70°C, for 1 h. Then, the solution

was cooled to room temperature, water was eliminated underreduced pressure, and the mixture was neutralized withconcentrated hydrochloric acid (Labsynth, Brazil, 36.5%).

Cure Reaction and Composite Preparation.Thermosetmaterials were obtained by mixing the prepolymer withresorcinol (Labsynth, Brazil), the curing accelerator (10:1,w/w) through mechanical stirring at 50°C for 30 min. Thecompression molding was carried out in a mold (220× 99.5× 5 mm3) under pressure. The molding cure cycle (75°C/1h/2.5 ton, 85°C/2 h/5.0 ton, 95°C/30 min/7.5 ton, 105°C/30 min/7.5 ton, 115°C/1 h/10.0 ton, 125°C/1.5 h/10.0 ton)was previously determined by DSC measurements.20

Composites reinforced with bagasse or curaua fibers(chemically modified or unmodified) were obtained byadding the fibers (18 g) to the prepolymers (102 g), themixture being submitted to mechanical stirring (30 min, 50°C) in order to get an optimum impregnation of thelignocellulosic materials in the prepolymer. The curingprocedure was analogous to the one described for thermosetpreparation (vide infra). Composites were prepared withrandomly oriented fibers (near 15 mm length).

Composite Characterization.For the Izod impact test,10 unnotched samples were cut from each plate and shapedaccording to ASTM D256 (63.5× 12.7× 4.0 mm3). Impactstrength was assessed using an Izod impact testor (CEASTResil 25). Impact tests were carried out at room temperaturewith an impact speed of 4 m/s and incident energy of 2.75J. In each experiment, at least 6 measurements were usedfor average calculation.

SEM, using the same conditions described for fibers, wasused to characterize the fractured samples.

Dynamic mechanical thermoanalysis (DMA or DMTA)was carried out using a thermal analyzer, TA Instrumentsmodel 2980, operating in the cantilevered horizontal measur-ing system. It was used in flexural mode to evaluate thespecimens. The experimental conditions used to analyze allsamples were as follows: oscillation amplitude of 10µm, 1Hz frequency, heating rate of 3°C/min, and temperaturerange 30-225 °C. Before starting every experiment, theequipment was stabilized at 30°C for 5 min.

Results and Discussion

Table 1 shows fiber composition data, which are inagreement with other data for these lignocellulosic materi-als.2,13

To further understand the chemical modification at themolecular level of the lignins present in the inner part of

Table 1. Properties of Sugar Cane Bagasse and Curaua Fibers

component (%)sugar cane

bagasse curaua

humidity 9.5 7.9ash 1.1 0.9holocellulose 72.1 83.5cellulose 55.2 73.6hemicelluloses 16.8 9.9klason lignin 25.3 7.5crystallinity 47 67

Composites of Phenolic Resin and Sugar Cane or Curaua Biomacromolecules, Vol. 6, No. 5, 2005 2487

the sugar cane and curaua fibers, lignin was extracted fromthe two vegetal sources considered in this work and submittedto the same modifications afterward. The lignins were reactedwith chlorine dioxide to oxidize the phenolic units of thismacromolecule and then treated with furfuryl alcohol. Weightpercent gains of 14% and 10% were obtained for sugar canebagasse and curaua, respectively.16 1H NMR of oxidizedlignins (figures not shown) indicated the diminution of thearomatic and methoxy regions after ClO2 oxidation, due tothe partial degradation of the macromolecule.16 Probably, thiskind of degradation of lignin was one of the factors thatreduced some mechanical properties of fibers, when theywere submitted to the same reaction, and thus of compositesreinforced with the chemically modified lignocellulosicfibers, as will be discussed later.

To quantify the different hydroxyl groups present in thelignin samples, curaua and bagasse, a31P NMR study of thephosphitylated polymers was carried out.16 Quantification ofthe different hydroxyl groups is presented in Table 2.

The total phenol and aliphatic hydroxyl groups for bagasseand curaua lignins after oxidation reaction are lower. ClO2

is known to react by one-electron oxidation with phenol togive phenoxy radicals, followed by chlorine dioxide additionto the latter.21 This process conduces to quinones andmuconic derivatives, mainly esters, in accordance with thelower content of phenol and acid groups observed (Table2). The low content of aliphatic hydroxyl groups might bedue to degradation of the propane chain, after chain displace-ment from carbon 1 of the aromatic ring in the radicalprocess. After modification with furfuryl alcohol, the valuesare also lower. This is indicative of the involvement of thehydroxyl group of furfuryl alcohol in the reaction. The

decrease of hydroxyl groups in the lignins after chemicalreactions (oxidation and grafting furfuryl alcohol) is tenta-tively confirmed by acetylation of the polymers. After theacetylation procedure, oxidized and chemically modifiedlignins were found not soluble in tetrahydrofuran, becauseof a lack of acetoxy groups necessary to render the polymerssoluble in this specific solvent.16

Once the reaction with ClO2 and FA was studied for sugarcane and curaua lignins, the related fibers were submittedto the same treatment. The WPG of sugar cane bagasse was18%, and the WPG of curaua was 7%. Sugar cane bagassehas a higher amount of lignin in its composition comparedwith curaua, and this reflects on the WPG, because thechemical modification of lignocellulosic fiber considered inthe present work mainly involves this macromolecule.

Observation of the red color displayed by the fibers ofsugar cane bagasse and curaua fibers after oxidation withClO2 is in accordance with the expected formation ofortho-quinones. Although UV-vis diffuse reflectance spectra ofthe fibers measured before and after oxidation (Figure 1)are not quantitative, they show a maximum near 455 nm;this value and the shape of the curve are indicative of theformation of guaiacyl and syringylortho-quinones.14 Also,it is likely thatp-quinones, absorbing in the near-UV region,are produced from thep-hydroxybenzyl alcohol structuralelements of lignin. The UV-vis diffuse reflectance spectraof sugar cane fibers (Figure 1, left) and curaua (Figure 1,right), ClO2 oxidized and treated with FA, show intenseabsorption in the visible region. This is in accordance withthe formation of complex quinonoid structures, as expectedfrom Diels-Alder reactions betweenortho-quinones andconjugated dienes.22

The content of lignin in sugar cane bagasse fiber changedfrom 25.3% (untreated fiber, Table 1) to 40.9% for the fiberoxidized and then treated with FA. Concerning curaua fibers,the content of lignins changed from 7.5% (untreated fiber,Table 1) to 11.7% after FA treatment. Probably for bothfibers, besides lignin, the FA polymer formed at the surfacealso remained insoluble during the analysis applied toquantify lignin, with the weight of FA polymer then addedto that of lignin.

The crystallinity of sugar cane bagasse changed from 47%to 52% after oxidation, probably because of the partialextraction of lignin, present in the amorphous region. The

Figure 1. UV-vis diffuse reflectance spectra of fibers expressed as log(1/R). (-) Unmodified fibers; (-) ClO2 oxidized fibers; (- - -) ClO2 oxidizedfibers and treated with FA modified fibers. Left: sugar cane bagasse fibers. Right: curaua fibers.

Table 2. Quantification of Several Hydroxyl Groups (mmol/glignin) in Curaua (CLig) and Bagasse (BLig) Isolated Lignins from31P NMR Analysis of Their Phosphitylated Derivatives (see ref 16)

CLig CLigoxa CLigox-FA

b BLig BLigoxa BLigox-FA

a

aliphatic 1.47 0.99 1.33 2.35 1.18 0.57S-OH 0.38 0.09 0.04 0.06 0.05 0.02G-OH 0.18 0.07 0.07 0.32 0.13 0.07H-OH 0.18 0.05 0.07 0.48 0.06 0.045-condensed 0.08 0.02 0.01 0.02 0.02 0.01total phenol 0.82 0.23 0.20 0.88 0.26 0.14acids 0.13 0.19 0.01 0.06 0.16 0.07

a Lignin oxidized. b Lignin oxidized and modified with FA.


reaction of oxidized fiber with FA decreased the crystallinityto 48%, which can indicate that the cellulose domain wasaffected in the considered reaction conditions. A similarbehavior was observed for curaua fibers, although withchanges occurring in a smaller scale than those in sugarcane: The crystallinity changed from 67% (untreated fiber,Table 1) to 70% (oxidized fiber) and 64% (oxidized and thenreacted with FA).

To confirm the occurrence of FA polymerization on thesurface of fibers, sugar cane bagasse was analyzed by solid-state13C CP-MAS NMR. Only sugar cane was consideredfor analysis because of its higher content of FA polymer onthe surface, in order to obtain spectra with adequate intensitypeaks.

Part of the structures of lignin, cellulose, and poly(furfurylalcohol), with the numbers and symbols assigned to theatoms, are showed in Figure 2.

The spectrum presented in Figure 3(A) shows character-istic signals of lignocellulosic fibers23 and more specificallyof sugar cane bagasse fibers.24,25Main resonance assignmentsare given in Table 3. The spectrum displays a signal at 22ppm, assigned to the methyl carbon of the acetyl group inhemicelluloses.

The region between 60 and 110 ppm is dominated bystrong signals, which are assigned mostly to the differentcarbons of cellulose, namely C1 (105 ppm), C4 crystalline(88 ppm), C4 amorphous (83 ppm), C2, 3, and 5 (75-73ppm), and C6 (64 ppm). Hemicelluloses also give signals in

this region. These signals overlap those due to differentaliphatic carbons of lignin. In this region, the only signalthat can be assigned to lignin is resonance peak 13 at 56ppm due to methoxy groups of aromatic moieties (Table 3).

The region between 110 and 160 ppm is specific to thearomatic carbon of lignin. At 173 ppm appears a signal dueto carbonyls of acetoxy and other ester groups in hemicel-luloses. The peak at 168 ppm originates from esterifiedp-coumaric acid or from the Cγ acidic carbon in etherifiedferulic acids. Thep-hydroxyphenyl (H) residues show aresonance signal at 128 ppm due to C2 and C6 carbons(Figure 3A and Table 3)

The chemical modification of bagasse fibers with chlorinedioxide and FA affected mainly the lignin polymer compo-nent, as is show in spectrum (B) (Figure 3). The carbohydratesignals remain unaffected, whereas signal 13 at 55.8 ppm,assigned to lignin methoxy groups, disappeared because ofthe formation of quinones after the oxidation reaction withClO2. Also, the lignin region between 110 and 160 ppm ischanged to a large extent. The difference spectra, (C)) (B)- (A), Figure 3, shows new resonance peaks at 153, 108,and 30 ppm due to the grafting of FA on the lignin polymer.The chemical shifts of these peaks are in accordance withthose found in linear poly(furfuryl alcohol) and might beassigned as follows: 153 ppm (C2), 108 ppm (C4 and C5),and 30 ppm (CR).26-28

These observations are in accordance with covalentgrafting of the furfuryl molecular framework on the ligninpolymer, via the oxidation of the phenolic lignin groups intoquinones and/or muconic acids.

The tensile strength and elongation of lignocellulosic fiberswere evaluated, and the results are shown in Table 4. Themaximum difference between each measurement and theaverage tensile strength and elongation (Table 4) was nearly25%, corresponding to the value of sugar cane bagasseoxidized with ClO2 fiber.

The data of Table 4 show that the oxidation of fibers withClO2 strongly reduces the tensile strength of both curaua andsugar cane bagasse. On the other hand, the reaction with

Figure 2. Schematic representation of the structures of lignin,cellulose, and poly(furfuryl alcohol).

Figure 3. 100-MHz CP-MAS 13 solid-state NMR spectra of bagassefibers (A), chemically modified bagasse fibers (oxidized by ClO2 andtreated with FA) (B), difference spectra (C) ) (B) - (A).

Table 3. Resonance Assignments of 13C CPMAS Spectrum ofSugar Cane Bagasse Fibers (refs 23-25)

peaknumber

chemical shift(ppm) assigmentsa

1 173 hemicelluloses: -COOsR and CH3COO-2 168 lignin: Cγ in ArsCHdCHCOOR

and LigOsArCHdCHCOOH3 153 lignin: S3(e), S5(e)4 150-142 lignin: S3(ne), S5(ne), G3, G4

5 140-122 lignin: S4, S1, G1, H2, H6

6 120-100 lignin: S2, S6, G2, G6

7 105 carbohydrates: C1


9 83 lignin: Câ. carbohydrates: C4

10 75 lignin: CR. carbohydrates: C2, C3, C5

11 73 carbohydrates: C2, C3, C5


13 56 lignin: OCH3

14 22 hemicelluloses: CH3COO-

a Abbreviations: S, syringyl; G, guaiacyl; H, p-hydroxyphenyl; ne,nonetherified; e, etherified.


FA seems to protect the fiber, probably because of thepolymeric layer introduced on the surface of fiber, as canbe inferred by the increase in tensile strength when theoxidized fibers react with FA (Table 4).

Figure 4 shows SEM images of unmodified and modifiedfibers

By comparing the SEM images shown in Figure 4, it ispossible to see that the chemical modification introduced acoating at the surface of the lignocellulosic fiber, which ismore evident for sugar cane bagasse because of the highercontent of FA polymer introduced, as has already beenmentioned.

The unmodified and modified fibers were characterizedin relation also to thermal stability (TG and DSC analysis,Figures 5 and 6). Studies show that the thermal decomposi-tion of lignocellulosic fibers is not necessarily an additivefunction resulting from the contribution of each fraction ofits components, that is, cellulose, hemicellulose, and lignin,because of the interactions between these fractions.29,30

In general, the thermolysis reaction of polysaccharides(cellulose and hemicellulose) occurs by the cleavage ofglycoside bonds, C-H, C-O, and C-C bonds, as well asby dehydration, decarboxylation, and decarbonylation. Con-

sidering the mixture arising from the degradation of cellulose,levoglucosan, produced by transglycosylation intramolecularreactions,31 is the most abundant product.32 Around 600°C,there can be carbonization of levoglucosan with the releaseof water.30,31

One of the degradation mechanisms of lignin to beconsidered occurs through dehydration, yielding derivativeswith lateral unsaturated chains and the release of water. Yet,carbon monoxide, carbon dioxide, and methane are alsoformed.32 The decomposition of aromatic rings occurs above400 °C.33 Continued burning leads to the saturation of thearomatic rings, the rupture of C-C bonds present in lignin,the release of water, CO2, and CO, and structural rearrange-ments.29 Considering the three main components of thelignocellulosic material, lignin is the component that presentsthermal degradation at a larger temperature range.30,34

The DSC and first-derivative TG curves (Figures 5 and6) obtained show peaks related to the decompositionprocesses mentioned already; however, the TG curves showin a more clear way the decomposition of polysaccharides(near 300°C) and lignin (near 480°C), the main componentsof sugar cane bagasse fibers. As the content of polysaccha-rides is considerably higher than that of lignin, the first-derivative TG peak related to the decomposition of polysac-charides is more intense than that of lignin. It must be pointedout that when end use of some material is considered, thetemperature of the beginning of the degradation is thelimiting factor concerning thermal stability. In the case ofthe fibers used in the present work, that temperature is near280 °C (Figure 5).

The profiles of the TG curves are similar for samplesoxidized only and for those oxidized and then reacted withFA. The DSC curve of this last one probably incorporatesseveral events in the large exothermic peak that appears from

Figure 4. SEM images of (a) unmodified and (b) modified curaua and (c) unmodified and (d) modified sugar cane bagasse fibers.

Table 4. Tensile Strength and Elongation (%) for Unmodified andModified Sugar Cane and Curaua Fibers

fiber

tensilestrength(mpa)

elongationε (%)

unmodified sugar cane bagasse 222 1.1oxidized sugar cane bagasse 126 0.7oxidized sugar cane bagasse reacted with FA 238 0.9unmodified curaua 636 0.8oxidized curaua 218 0.5oxidized curaua reacted with FA 565 1.2


300°C. It must be pointed out that the chemical modificationoccurs mainly at the surface, and in the thermal analyses,the material as a whole is analyzed. The content of FAgrafted on the surface is low, and besides, its thermaldecomposition probably occurred in the same temperaturerange of the components of fibers, because exothermic peaksnear 325°C are found for poly(furfuryl alcohol).28 Curauafibers presented similar TG and DSC curves (figures notshown).

Table 5 gives the impact strength results obtained forcomposites reinforced with sugar cane bagasse and curauafibers, both unmodified and chemically modified.

Cellulose content and fiber crystallinity are importantfactors for the mechanical properties of fibers. The largercellulose content and the larger proportion of crystallineregions of the curaua fiber in relation to those of the sugarcane bagasse fiber (Table 1) have an effect on the impactstrength of composites reinforced with curaua fibers, as their

phenolic composites are considerably more resistant thanthose reinforced with sugar cane bagasse (Table 5).

The data given in Table 5 confirm the data on tensilestrength; that is, in the conditions of the present work, ClO2

Figure 5. (a) TG curves of sugar cane bagasse, oxidized with ClO2 and modified with furfuryl alcohol. (b) TG curves of sugar cane bagasseoxidized with ClO2, N2 atmosphere, 10 mL/min, 20 °C/min.

Figure 6. (a) DSC curve of sugar cane bagasse oxidized with ClO2 and modified with furfuryl alcohol, (b) DSC curve of sugar cane bagasseoxidized with ClO2, (c) DSC curve of curaua oxidized with ClO2 and modified with furfuryl alcohol. N2 atmosphere, 10 mL/min, 20 °C/min.

Table 5. Izod (unnotched) Impact Strength of PhenolicComposites Reinforced with Unmodified and Chemically ModifiedSugar Cane Bagasse and Curaua Fibers and Their RespectiveStandard Deviations (%)

fiber composite impact strength (J/m)

Unmodifiedsugarcane 28 ( 7curaua 88 ( 19

Oxidized ClO2

sugarcane 15 ( 2curaua 71 ( 7

After Reaction with FAsugarcane 17 ( 2curaua 39 ( 1


oxidation led to fiber degradation, which in the case ofinfluence on impact strength of curaua composites, wasintensified by the later reaction with furfuryl alcohol, as theproperties of curaua fibers decreased noticeably after thesemodifications. In contrast, electronic scanning (SEM) imagesof the composites reinforced with sugar cane bagasse (seeFigure 7 and compare a with b and d) show that adhesionwas significantly intensified in the fiber/matrix interface bythese treatments. With relation to curaua, the SEM images

of composites reinforced with modified fibers also showedregions with good adhesion at the interface, although it isalso possible to detect regions in which the pullout mecha-nism occurred, as a consequence of poor adhesion (Figure7e).

Probably, the chemical modification the surface sufferedfacilitated the interaction with the phenolic prepolymer byincreasing its interdiffusion in the lignocellulosic network,leading to a more intense fiber/matrix interaction during the

Figure 7. SEM images of the impact fracture surface of sugar cane and curaua reinforced phenolic composites. (a) Bagasse unmodified(×5000); (b) bagasse oxidized with ClO2 (×5000); (c) bagasse oxidized with ClO2 and modified with FA (×5000); (d) bagasse oxidized withClO2 and modified with FA (×300); (e) curaua oxidized with ClO2 and modified with FA (×500; pull-out mechanism occurrence indicated byarrows).


fiber cure step, and consequently also facilitating theestablishment of physical interactions and/or possibly offiber/matrix chemical bonds.

Although the SEM images showed, at least for sugar canebagasse fibers, that the fiber/matrix interactions were im-proved because of the modifications of fiber surfaces, theimpact strengths of the composites reinforced with bothmodified curaua and sugar cane bagasse are lower than thoseof composites reinforced with unmodified fibers (Table 5).These results indicate that the intensification of interactionsat the interface was overshadowed by the worsening of themechanical properties of the fibers caused by the chemicaltreatment applied on their surfaces.

The composites were also characterized by DMA, atechnique not extensively used for vegetal fiber-reinforcedcomposites, in comparison to synthetic-fiber reinforcedcomposites.35 However, studies on the dynamical mechanicalproperties of these vegetal fiber-reinforced composites areof great importance, considering that these materials cansuffer dynamical stressing during service.36

Considering the three parameters obtained from thisanalysis, that is, the storage and loss moduli (E′ and E′′,respectively) and tanδ, it is possible to get information onmolecular mobility.37 When the polymer under considerationcorresponds to a thermoset, the mobility is related mainlyto the segments located among the cross-linkages. Besides,when the thermoset is part of a composite material, other

features must be considered. The dynamic mechanicalproperties of a composite are determined by the propertiesof its components, the system morphology and nature of theinterface between the two components. The layer of thematrix that circles the fibers, that is, the layer immediatelyposterior to the interface, can have different properties fromthose of the remaining material.

Figure 8 shows that the storage modulus (E′) of sugar canebagasse reinforced composites is lower than that of thethermoset and decreases when the fibers are modified, inthe whole temperature interval considered. On the contrary,composites reinforced with curaua have higher (modifiedfibers) or nearly the same (unmodified fiber)E′ as thermoset(Figure 8)

Relating to the lowerE′ values of composites reinforcedwith sugar cane bagasse when compared with phenolicthermoset (Figure 8a), this is probably mainly caused by thepoor mechanical properties of fiber, which lead to a less rigidmaterial compared to the thermoset. Besides, differencesbetween the coefficients of thermal expansion of the fiberand of the polymer can be another factor that can explainthis decrease inE′. Because of this difference, stress in thematrix can appear, and as a consequence, the modulusbecomes lower than the value obtained in the absence ofstresses. The same behavior was observed in a previous workfor composites reinforced with unmodified sugar canebagasse.13

Figure 8. Storage modulus (E′) for phenolic thermoset (PT) and composites (PC) reinforced with (a) sugar cane bagasse and (b) curaua fibers;(c) PT submitted to post cure (175 °C, 1 h).


Curaua is stiffer than sugar cane fiber, mainly because ofboth its higher cellulose content and crystallinity than bagasse(Table 1). As a consequence, the storage modulus of curaua-reinforced composites is higher than that of thermoset,because of the incorporation of a high-modulus fiber intothe polymer, which contributes to the overall properties.(Figure 8b).

The DMA study for both composites revealed still otherinteresting features. The tanδ and E′′ values of thecomposites reinforced with modified sugar cane fibers werelower than those of the composites reinforced with unmodi-fied fibers (Figure 9).

The tan δ value can be taken as an indication of theintensity of interactions at the fiber/matrix interface, becausethe interactions are stronger and the mobility of polymersegments is lower at the interface, and then, the damping isreduced. Yet, an improved composite interface leads to moreefficient transference of stress between the fiber and thematrix and lower energy dissipation and thusE′′.38 In thisway, both values, tanδ and E′′, seem to indicate that thechemical modification of sugar cane bagasse increased theinterface interaction. On the other hand,E′ is lower for thecomposite reinforced with modified fiber than that for thecomposite reinforced with unmodified sugar cane bagasse,

Figure 9. Loss modulus (E′′) and tan δ for phenolic thermoset (PT) and sugar cane bagasse reinforced phenolic composites (PC).

Figure 10. Loss modulus (E′′) and tan δ for phenolic thermoset (PT) and curaua fiber-reinforced phenolic composites (PC).

Figure 11. Loss modulus for sugar cane-reinforced composite: (a) unmodified, (b) oxidized and modified with FA.


as was already mentioned. The storage modulus (E′) reflectsmainly the stiffness of the material, and as the treatmentweakened the fiber, the material as a whole became less rigid.

Figure 10 shows that tanδ andE′′ values of compositesreinforced with chemically modified curaua fibers are higherthan those of composites reinforced with unmodified fibers.Considering the previous discussion for sugar cane bagassecomposite, it seems that the treatment of curaua fiber leadsto a less intense interaction with the matrix, thus increasingtanδ andE′′ values. As was already mentioned, the sites ofthe fiber that reacted with ClO2 and FA are mainly locatedin the lignin portion, and bagasse has a higher content oflignin than curaua (Table 1). Besides, curaua has a morphol-ogy different from that of bagasse (see Figure 7), and thechemical modification can change the surface in a way thatdecreases the interactions with the matrix at the interface.

The E′′ curves for both composites (Figures 9 and 10)show two peaks, near 125 and 200°C, the first one probablyrelated to the movement of the segments located among thecross linkages, that is related to the glass transition (Tg). Theinteractions at the interface seem to be different for eachfiber (curaua and bagasse) and even for unmodified andmodified fibers, as was mentioned, and at first, this can leadto a displacement inTg peaks for higher temperatures as theinteractions increase. However, in the present work, the shiftsin Tg peaks were not significant (Figures 9 and 10). The peakat 200°C is probably related to a residual cure that occursduring the scanning, because in some storage modulus curves(Figure 8a,b), theE′ value increases between 150 and 200°C, which is an indications of postcure, which in turn leadsto a more rigid material and then to a second transition as isobserved in loss modulus curves near 200°C. Figure 8cshows both theE′curves for phenolic thermoset before andafter postcure. It can be observed that the increase inE′values between 150 and 200°C practically disappeared afterthe sample was subjected to 175°C for 1 h, indicating thatthe main process leading to the observed increase in storagemodulus was the residual cure. Figure 8c shows also thatthe postcure leads to a stiffer material, with higher storagemodulus than the previous one, in all the temperatureintervals considered. It can be pointed out that not necessarilya fully cured matrix must be obtained, because it can leadto a more brittle material, as was already observed in previousworks.2,20As the postcure occurs at higher temperatures thanthose normally related to the end use of this kind of material,this process does not limit the application of them. The peakaround 50°C that appears in some curves probably is relatedto the mobility of some more free segments.

The E′′ peaks near 125 and 200°C are enlarged as aconsequence of fiber treatment, as can be observed in Figure10 for composites reinforced with curaua and in Figure 11,where the curves for composites reinforced with sugar caneboth unmodified and oxidized and modified with FA appearin a scale more suitable than in Figure 9 to observe thedifferences mentioned. The treatment introduces differentsites at the fiber surface and then diversifies the kind ofinteractions that can occur at the fiber/matrix interface. Theenlargement ofE′′ curves then reflects the increase inheterogeneity of segments, as they are involved in fiber

interactions with still more varied intensity compared withthe unmodified fibers.

Conclusions

A new process was set for chemically modifying sugarcane bagasse and curaua fibers. A quite specific oxidationby chlorine dioxide of syringyl and guaiacyl phenols of thelignin polymer creating quinones was obtained. The latterwere reacted with furfuryl alcohol, creating a coating aroundthe fiber more compatible to phenolic resins, to preparecomposites. This modification favored the fiber/matrixinteraction at the interface for sugar cane and curaua fiber-reinforced composites, but caused some fiber degradationthat affected their mechanical properties and decreased themechanical resistance of the composites reinforced withthem. It must be pointed out that the chemical modificationprocess of the fibers was conducted in water with mattercoming from natural resources such as nonwood fibers andfurfuryl alcohol. On a molecular basis, UV-vis diffusereflectance spectroscopy has pointed out the formation ofortho-quinones, whereas CP-MAS13C NMR have showndisappearance of methoxy groups in lignin phenyl units withstrong chemical modification of the aromatic ring. Concomi-tantly, peaks of furfuryl alcohol polymer were observed. Asthis is an ongoing project, it has been sought to optimizefiber surface modification in order to intensify fiber/matrixinteraction without the loss of fiber mechanical properties.

Curaua and sugar cane bagasse fibers presented differentbehaviors in relation to similar treatment and worked withdifferent intensities as reinforcing agents of the same matrix.These results indicated that is important to find optimalconditions for each lignocellulosic fiber individually, withrespect to its action as a reinforcing agent. Although thecomposites reinforced with curaua fibers showed higherimpact strength than those reinforced with sugar canebagasse, at first both of them can be used in somenonstructural applications, as, for instance, inner trim partsin automotive applications.

Acknowledgment. The authors are grateful to CAPESfor research fellowship for W.G.T. and I.A.T.R. and toCAPES/COFECUB (project 422/03) for traveling missionsbetween Brazil and France. E.F. is grateful to CNPq(National Research Council, Brazil) for research productivityfellowships and to FAPESP (The State of Saõ Paulo ResearchFoundation, Brazil) for a research fellowship for J.D.M. andfor financial support. W.H. is very grateful to ConseilRegional de la Reúnion (France) and to Fond SocialEuropeen (FSE) for a doctoral grant and A.C. is veryindebted to B. Siegmund (CIRAD) for his participation inthe sugar cane bagasse project.

References and Notes

(1) Young, R. A. Utilization of Natural Fibers: Characterization,Modification and Applications. InLignocellulosic-Plastics Compos-ites; Leao, A. L, Frollini, E., Carvalho, F. X., Eds.; UNESP-USP,Sao Carlos, SP, Brazil, 1997; p 1.

(2) Frollini, E.; Paiva, J. M. F.; Trindade, W. G.; Razera, I. A. T.; Tita,S. P. Lignophenolic and Phenolic Resins as matrix in vegetal fibersreinforced composites. InNatural Fibers, Polymers and Composites- RecentAdvances; Wallenberger, F., Weston, N., Eds.; KluwerAcademic Publishers: New York, 2004; p 193.


(3) Joshi, S. V.; Drzal, L. T.; Mohanty, A. K.; Arora, S.Composites,Part A 2004, 35, 371.

(4) Mohanty, A. K.; Wibowo, A.; Misra, M.; Drzal, L. T.Composites,Part A 2004, 35, 363.

(5) Wambua, P.; Ivens, J.; Verpoest, I.Compos. Sci. Technol.2003, 63,1259.

(6) Pervaiz, M.; Sain, M. M.Resour. ConserV. Recycl. 2003, 39, 325.(7) Keener, T. J.; Stuart, R. K.; Brown, T. K.Composites, Part A2004,

35, 357.(8) Maffezzoli, A.; Calo, E.; Zurlo, S.; Mele, G.; Tarzia, A.; Stifani, C.

Compos. Sci. Technol. 2004, 64, 839.(9) Thielemans, W.; Wool, R. P.Composites, Part A2004, 35, 327.

(10) Rowell, R. M.; Sanadi, A. R.; Caulfield, D. F.; Jacobson, R. E.Utilization of Natural Fibers in Plastic Composites: Problems andOpportunities. InLignocellulosics-Plastics Composites; Leao, A. L.,Carvalho, F. X., Frollini, E., Eds.; UNESP-USP, Saõ Carlos, SP,Brazil, 1997; p 23.

(11) Benatti, R., Jr. Natural fibers in Brazil-Utilization, Perspective andProcessing for Sisal, Ramie, Jute, Kenaf and Allied Fibers forComposites Production. InLignocellulosics-Plastics Composites;Leao, A. L., Carvalho, F. X., Frollini, E., Eds.; UNESP-USP, SaõCarlos, SP, Brazil, 1997; p 343.

(12) Razera, I. A. T.; Frollini, E.J. Appl. Polym. Sci.2004, 91, 1077.(13) Paiva, J. M. F.; Frollini, E.J. Appl. Polym. Sci. 2002, 83, 880.(14) Trindade, W. G.; Hoareau, W.; Razera, I. A. T.; Ruggiero, R.; Frollini,

E.; Castellan, A.Macromol. Mater. Eng.2004, 289, 728.(15) Yosomiya, R. InAdhesion and Bonding in Composites; Marcel

Dekker, Inc.: New York, 1990; p 1.(16) Hoareau, W.; Trindade, W. G.; Siegmund, B.; Castellan, A.; Frollini,

E. Polym. Degrad. Stab.2004, 86, 567.(17) Browning, B. L. InMethods of Wood Chemistry; Wiley: New York,

1967; Vol. 2, p 499.(18) Buschle-Diller, G.; Zeronian, S. H.J. Appl. Polym Sci.1992, 45,

967.(19) Mark, H. F.; Othmer, D. F.; Overberger, C. G.; Seaborg, G. T. Castor

oil to chlorosulfuric acid. InKirk-Othmer Encyclopedia of ChemicalTechnology, 4th ed.; John Wiley and Sons: New York, 1993; Vol.5, p 981.

(20) Paiva, J. M. F.; Trindade, W. G.; Frollini, E.; Pardini, L. C. InNaturalPolymers and Composites IV; Frollini, E., Leao, A. L., Mattoso, L.H. C., Eds.; USP-UNESP-EMBRAPA, Saõ Carlos, Brazil, 2002; p528.

(21) Ni, Y.; Van Heiningen, A. R. P.J. Wood Chem. Technol. 1994, 14,243.

(22) Hergert, H. L. InLignins, Occurrence, Formation, Structure, andReactions; Sarkanen, K. V., Ludvig, C. H., Eds.; Wiley-Inter-science: New York, 1971; p 165.

(23) Bardet, M.; Foray, M. F.; Tran, Q. K.Anal. Chem. 2002, 74, 4386.(24) Sun, J. X.; Sun, X. F.; Sun, R. C.; Fowler, P.; Baird, M. S.J. Agric.

Food Chem. 2003, 51, 6719.(25) Sun, X. F.; Sun, R. C.; Sun, J. X.Bioresour. Technol. 2004, 95,

343.(26) Choura, M.; Belgacem, N. M.; Gandini, A.Macromolecules1996,

29, 2839.(27) Principe, M.; Ortiz, P.; Martinez, R.Polym. Int. 1999, 48, 637.(28) Gonzales, R.; Figueroa, J. M.; Gonzales, H.Eur. Polym. J. 2002,

38, 287.(29) Perez, M. G.; Chaala, A.; Yang, J.; Roy, C.Fuel 2001, 80, 1245.(30) Paiva, J. M. F.; Trindade, W. G.; Frollini, E.; Pardini, L. C.Polym.

Plast. Technol. Eng. 2004, 43, 1179.(31) Scheirs, J.; Camino, G.; Tumiatti, W.Eur. Polym. J. 2001, 37, 933.(32) Meier, D.; Faix, O.Bioresour. Technol. 1999, 68: 71.(33) Rohella, R. S.; Sahoo, N.; Paul, S. C.; Choudhury, S.; Chakravortty,

V. Thermochim. Acta1996, 287, 131.(34) Orfao, J. J. M.; Antunes, F. J. A.; Figueiredo, J. L.Fuel 1999, 78,

349.(35) Ray, D.; Sarkar, B. K.; Das, S.; Rana, A. K.Compos. Sci. Technol,

2002, 62, 2002, 911.(36) Joseph, P. V.; Mathew, G.; Joseph, K.; Groeninckx, G.; Thomas, S.

Composites, Part A2003, 34, 275.(37) Pothan, L. A.; Thomas, S.Compos. Sci. Technol. 2003, 63, 2003,

1231.(38) Keusch, S.; Haessler, R.Composites, Part A1999, 30, 997.

BM058006+


Thermoset Phenolic Matrices Reinforced with Unmodified and … · 2016. 6. 29. · an interaction of physical, chemical, or physical-chemical nature between the two components.15

Documents