sodium borohydride as a protective agent for the alkaline ... · pdf filetreatment of sisal...

Composite Interfaces 18 (2011) 407–418brill.nl/ci

Sodium Borohydride as a Protective Agent for the AlkalineTreatment of Sisal Fibers for Polymer Composites

A. G. O. Moraes a,∗, M.-R. Sierakowski b, T. M. Abreu a and S. C. Amico a

a Department of Materials Engineering, Federal University of Rio Grande do Sul,Porto Alegre, P.O. Box 15010, 91501-970, RS, Brazil

b BioPol, Department of Chemistry, Federal University of Paraná, Curitiba,P.O. Box 19081, 81531-990, PR, Brazil

Received 9 March 2011; accepted 7 July 2011

AbstractThe vegetable fibers used for polymer matrix composites are usually treated to improve their adhesion withthe matrix. The chemical treatment with sodium hydroxide (NaOH), although widely used, may damage thefiber surface structure, reducing its strength. The possibility of protecting vegetable fibers against alkalinechemical aggression by using hydride ions (H−) was investigated in this work. Sisal fibers were modifiedby immersion in a NaOH aqueous solution (2, 5 and 10% wt/vol), with or without the addition of sodiumborohydride (NaBH4) (1% wt/vol), under variable conditions (immersion time and temperature). The effectof using NaBH4 was investigated using fiber tensile and pull-out tests, critical length calculation, along witha Weibull statistical analysis. This agent was found to minimize sisal degradation under highly concentratedalkaline conditions in comparison with sisal treated with the pure NaOH solution. The results suggest the5% wt/vol treatment for 60 min under room temperature in the presence of the hydride ions as the mostsuitable for sisal. This result may be extended to other vegetable fibers of similar composition and maypromote their use in polymer composites.© Koninklijke Brill NV, Leiden, 2011

KeywordsVegetable fibers, alkaline treatment, polysaccharides degradation, borohydride ions, strength, adhesion

1. Introduction

Natural fibers can be of animal (e.g., silk, wool), mineral (e.g., asbestos) or veg-etable (e.g., wood, pineapple, sisal, curaua, henequen, hemp, jute, ramie, coir, bam-boo, sugar cane) origin. Vegetable fibers are essentially microcomposites consistingof cellulose fibers embedded in an amorphous lignin and hemicellulose matrix [1].A variety of uses has been found for these fibers [2], for instance, as reinforcementof polymers, where they represent an alternative for the partial replacement of glass

* To whom correspondence should be addressed. E-mail: [email protected]

© Koninklijke Brill NV, Leiden, 2011 DOI:10.1163/156855411X595825

408 A. G. O. Moraes et al. / Composite Interfaces 18 (2011) 407–418

fiber composites. Indeed, many industrial sectors already use composites reinforcedwith vegetable fibers, such as the automotive industry [3], in a variety of compo-nents (mainly internal ones), focusing on fuel consumption reduction among otherbenefits [4].

Vegetable fibers cannot directly bond to the hydrophobic polymer resin and,therefore, surface, mostly chemical, treatments are generally applied to promotegreater compatibility/affinity between these phases. Many papers already reportedon the effect of different chemical treatments on vegetable fibers investigating dis-tinct agents, concentrations and treatment conditions (time and temperature) [5–7].

Perhaps the most widely employed treatment for vegetable fibers is the alka-line treatment, which is usually performed using an aqueous sodium hydroxide(NaOH) solution [8], which promotes ionization of the hydroxyl (OH) groups ofthe fiber to alkoxide [5]. Although these treatments usually increase adhesion, theycan be costly, preventing their widespread use, and they may also diversely affectthe mechanical properties of the fibers, reducing fiber strength [5, 6, 9], which willultimately be detrimental to the composite produced with these fibers. As an ex-ample, in the work of the Mishra et al. [10], the polyester composite reinforcedwith sisal fibers treated with 5% NaOH showed higher tensile strength than thatcontaining sisal fibers treated with 10% NaOH.

Investigation of the degradation mechanism [11–14] and protection of polysac-charides (e.g., cellulose and xyloglucan), under alkaline aqueous conditions withthe use of protective agents based on borohydride ions (BH4

−) [12], has recentlybeen applied to vegetable fibers [15]. During alkaline treatment, chemical com-pounds containing hydride ions (H−), e.g., sodium borohydride (NaBH4), can actas a reducing agent in reducing end-groups (aldehyde) present in the C − 1 freepolysaccharides chain by a mechanism called end-wise degradation or β-elim-ination (peeling or unzipping) [14], therefore, minimizing fiber degradation.

This study aims to investigate the use of NaBH4 as a protective agent of sisalfibers subjected to chemical treatment with NaOH aqueous solution with respect tofiber strength and fiber adhesion to an unsaturated polyester matrix.

2. Experimental

2.1. Materials

The sisal fibers were supplied as yarns by Casa Gaúcha de Barbantes (Porto Ale-gre/RS) and the linear density of the wire and the fiber were 2588 and 31 tex,respectively. Glacial acetic acid 99.7 wt% (Quimex), sodium borohydride 97.0 wt%(Nuclear), sodium hydroxide 98 wt% (Vetec), orthophthalic unsaturated polyesterresin (Uceflex UC 5518, Elekeiroz) and ethylmethylketone peroxide (initiator:Butanox® M-50, Akzo Nobel) were used as received.

A. G. O. Moraes et al. / Composite Interfaces 18 (2011) 407–418 409

2.2. Fiber Treatment and Characterization

The sisal fibers were surface modified by immersion in a sodium hydroxide (NaOH)aqueous solution (2, 5 and 10% wt/vol), under mechanical stirring (70 rpm), with orwithout the addition of sodium borohydride (NaBH4; 1% wt/vol) in batches of 3.8 gof fiber per 500 ml of solution, for 15, 30 or 60 min, at room temperature (20◦C)or at 40◦C. The remaining solution in the fibers was neutralized with an acetic acidaqueous solution (HAc; 0.2% wt/vol), followed by washing with distilled water,drying in an oven with air circulation at 105 ± 1◦C for 1 h and conditioning. Thevarious fiber samples studied were referred to as shown in Table 1. The in naturafiber and the fibers which had only been immersed in distilled water for 60 min(named H2O_t60) were used as control groups.

Individual fiber filaments were randomly chosen from the yarn and used intensile testing. The conditioning of samples was carried out according to ASTMD1776. Tensile tests (ASTM D2256) were carried out in a universal testing ma-chine (Emic DL 10000) equipped with 50 N load cell. The speed of testing andthe gauge length were constant in all tests, 5 mm/min and 100 mm, respectively.A pre-load of about 0.5 N was applied prior to testing to reduce initial slippageof the fibers. All tests were conducted at room temperature, typically 20◦C. Datawere discarded if the fiber fractured at or near the clamps and a new fiber was used.Between 18 and 50 useful fibers were tested in each case. After a successful test,both fiber fracture extremes were measured with a Mitutoyo micrometer (0.01 mmresolution). The cross-section of the fibers was assumed circular for simplificationand the mean diameter of both fiber ends was used in the calculations.

To evaluate shear strength at the sisal/polyester interface, a single fiber (30 mmlength) was partially incorporated (1.5 mm depth) into an unsaturated orthophthalic

Table 1.Nomenclature used to identify the treatments applied to the sisal fibers

H2O_t60 Washing with water for 60 min

OH2_t60 Treatment with 2% (wt/vol) of NaOH for 60 min,OH2_BH_t60 without or with 1% (wt/vol) of NaBH4, respectively

OH5_t15, Treatment with 5% (wt/vol) of NaOH for 15, 30 or 60 minOH5_t30,OH5_t60

OH5_t60_T40 Treatment with 5% (wt/vol) of NaOH for 60 min at 40◦C,OH5_BH_t60_T40 without or with 1% (wt/vol) of NaBH4, respectively

OH5_BH_t15, Treatment with 5% (wt/vol) of NaOH and 1% (wt/vol) ofOH5_BH_t30, NaBH4 for 15, 30 or 60 min, respectivelyOH5_BH_t60

OH10_t60 Treatment with 10% (wt/vol) of NaOH for 60 min,OH10_BH_t60 without or with 1% (wt/vol) of NaBH4, respectively


polyester block (12 × 12 mm) cured using 1.0 wt% of ethylmethylketone perox-ide as initiator. The fiber was pulled along its axis orientation until full detachmentfrom the resin. Tests were carried out at 1 mm/min and at room temperature (typi-cally 20◦C) in the same testing machine detailed above. Data were discarded if thefiber fractured and between 12 and 23 useful specimens were used for each fibertreatment. From the mean fiber diameter and tensile strength values, and using thepolymer/matrix interfacial strength, the critical fiber length for each treatment canbe calculated as in [16].

The Weibull analysis was applied to all data. The α and β parameters of thisstatistical distribution [17, 18] determine the shape and scale of the distribution,respectively. The β parameter is called the ‘characteristic life’, i.e., the failure prob-ability is approximately equal to 63.2%. The α parameter is dimensionless, beinginversely proportional to the dispersion of the values of the property assessed. An-other important parameter is the T50, which corresponds to the 50th percentile (themedian) of the data. In order to estimate these parameters, two methods are usuallymentioned — the maximum likelihood technique and the use of a probability plot,which was employed in this work.

3. Results and Discussion

3.1. Effect of Treatment on Fiber Diameter and Strength

To check the hypothesis of a Weibull distribution for the average diameter, the an-alytical W -test was used. In six of fourteen sets of fibers, the W value did not liebetween W0.05 and W0.95 (not shown here). Figure 1 shows the Weibull probabilityplot of the diameter data for twenty seven OH2_BH_t60 fibers, and a low coeffi-cient of determination (R2 = 0.925) can be noticed, indicating poor agreement tothe Weibull distribution. This suggests that this sort of statistical analysis for thediameter values might not be necessary or even justifiable for vegetable fibers suchas sisal, even more due to the excellent agreement between T50 and the mean for allfibers, as shown in Table 2.

The mean fiber diameter values and their standard deviations for each chemicaltreatment are presented in Table 2 and high standard deviations are found for allfibers, as expected for vegetable fibers. Considering the scatter and the somewhatconflicting results, it is difficult to undoubtedly draw conclusions regarding thesedata. Bearing in mind that every sisal yarn consisted of around 84 filaments andless than half of them were measured for each treatment, the sampling may notbe representative. Nevertheless, shorter treatment times, as in OH5_t15, tends tomaintain the original fiber diameter, whereas for longer periods it tends to decrease.The NaBH4 was found to generally promote fiber swelling.

The hypothesis of Weibull distribution for tensile strength, elastic modulus andstrain at break was also checked. The analytical W -test was again used and, in thiscase, in only 2–4 sets of data did the W values not lie between W0.05 and W0.95values (not shown here). When these data followed a Weibull distribution, the R2


Figure 1. Weibull probalitity plot of the diameter values for twenty-seven OH2_BH_t60 fibers.

Table 2.Mean diameter and T50 values obtained for the in natura and treated SF

Treatment Average mean diameter (µm) T50 (µm)

In natura 169 (±39) 164H2O_t60 174 (±31) 170OH2_t60 154 (±28) 151OH2_BH_t60 172 (±29) 170OH5_t15 169 (±31) 165OH5_BH_t15 156 (±18) 155OH5_t30 153 (±32) 149OH5_BH_t30 168 (±28) 165OH5_t60 143 (±31) 140OH5_BH_t60 149 (±21) 147OH5_t60_T40 174 (±23) 172OH5_BH_t60_T40 181 (±32) 178OH10_t60 174 (±27) 171OH10_BH_t60 163 (±42) 158

in the Weibull probability plot showed a high value. To illustrate, the histogram andthe Weibull distribution parameters of the tensile strength data for the OH2_BH_t60fibers are shown in Fig. 2(a), whereas Fig. 2(b) shows the respective Weibull prob-ability plot.

In order to correctly estimate fiber tensile strength, the diameter of the fiber mustbe known and this fact alone justifies the measurement of each fiber being tested.


(a)

(b)

Figure 2. Histogram (a) and Weibull probalitity plot (b) of tensile strength data for the OH2_BH_t60fibers.

Even then, the standard deviations are always very high, sometimes higher than20% of the mean [19, 20]. In this work, the mean and standard deviation of thetensile strength data for each set of fibers can be seen in Table 3 along with theWeibull parameters. The tensile strength of the in natura SF is consistent with the


Table 3.Mean tensile strength and T50 values obtained for the in natura and treated SF

Treatment Mean tensile strength (MPa) T50 (MPa) β (MPa)

In natura 399 (±111) 383 437H2O_t60 418 (±94) 407 451OH2_t60 382 (±87) 372 413OH2_BH_t60 396 (±86) 386 428OH5_t15 341 (±87) 330 371OH5_BH_t15 409 (±87) 399 443OH5_t30 343 (±85) 332 373OH5_BH_t30 410 (±85) 401 446OH5_t60 398 (±63) 393 423OH5_BH_t60 399 (±71) 393 426OH5_t60_T40 320 (±54) 315 340OH5_BH_t60_T40 392 (±69) 386 420OH10_t60 332 (±79) 321 364OH10_BH_t60 337 (±83) 327 368

results obtained by Silva et al. [20], who found 400 ± 126 MPa (testing speed andgauge length of 0.1 mm/min and 40 mm, respectively).

The variations of β,T50 and the mean for the tensile strength data show that ten-sile strength decreases with the increase in NaOH concentration. However, due tothe high standard deviation of the results (around 25%), a statistical test is necessaryto verify if there is a significant decrease. To determine the confidence with whichone can infer differences between mean values, an approximate graphical calcula-tion [18] was carried out taking into account sampling size and slope of the Weibulldistribution. From the results obtained, it can be said, with more than 90% confi-dence, that only the OH5_t60_T40 and OH10_t60 treatments significantly decreasetensile strength in comparison with the in natura SF. Interestingly, when comparedto the OH10_BH_t60 treatment, the confidence decreases to 89%, i.e., the use ofNaBH4 turns the sample similar to the in natura fiber in terms of strength.

Table 4 shows the mean and standard deviation of the elastic modulus data foreach set of fibers along with the respective Weibull parameters. The in natura SFmodulus was somewhat similar to that reported by Silva et al. [20], who found avalue of 19 ± 7 GPa. The elastic modulus mostly decreased with stricter NaOHtreatments and a recovery in modulus was obtained when using NaBH4. The oppo-site behavior showed by the treatment with 10% NaOH suggests greater aggregationof fibrils during the chemical treatment, with less interaction with the solution [12].

The dependence of the mechanical properties (tensile strength and elastic modu-lus) with the average diameter of the in natura SF was also studied. Tensile strengthand elastic modulus appeared to increase when the average diameter decreased. Inaddition, modulus tended to be higher for stronger fibers.


Table 4.Mean elastic modulus and T50 values obtained for the in natura and treated SF

Treatment Mean elastic modulus (GPa) T50 (GPa) β (GPa)

In natura 17.8 (±4.8) 17.1 19.5H2O_t60 16.8 (±5.0) 16.0 18.4OH2_t60 18.6 (±3.9) 18.1 19.9OH2_BH_t60 18.8 (±4.4) 18.3 20.3OH5_t15 16.1 (±4.1) 15.5 17.6OH5_BH_t15 17.1 (±3.7) 16.7 18.6OH5_t30 16.3 (±4.5) 15.6 17.9OH5_BH_t30 17.5 (±3.0) 17.2 18.7OH5_t60 17.5 (±3.6) 17.1 18.9OH5_BH_t60 18.0 (±3.5) 17.6 19.5OH5_t60_T40 14.3 (±2.9) 14.0 15.5OH5_BH_t60_T40 17.2 (±2.9) 16.8 18.4OH10_t60 17.2 (±3.6) 16.8 18.5OH10_BH_t60 14.3 (±3.6) 13.8 15.5

Table 5.Mean strain at break and T50 values obtained for the in natura and treated SF

Treatment Mean strain at break (%) T50 (%) β (%)

In natura 2.3 (±0.5) 2.2 2.5H2O_t60 2.4 (±0.5) 2.3 2.5OH2_t60 2.0 (±0.4) 2.0 2.2OH2_BH_t60 2.1 (±0.4) 2.0 2.2OH5_t15 2.1 (±0.3) 2.0 2.1OH5_BH_t15 2.6 (±0.5) 2.5 2.7OH5_t30 2.2 (±0.5) 2.1 2.3OH5_BH_t30 2.5 (±0.4) 2.4 2.6OH5_t60 2.3 (±0.5) 2.2 2.5OH5_BH_t60 2.4 (±0.4) 2.3 2.5OH5_t60_T40 2.4 (±0.3) 2.4 2.5OH5_BH_t60_T40 2.5 (±0.4) 2.4 2.6OH10_t60 1.9 (±0.5) 1.8 2.1OH10_BH_t60 2.3 (±0.4) 2.2 2.4

Obs.: Up to about 0.5 N (pre-load), the fibers displayed around 1.7% strain.

The results of mean and standard deviation of the strain at break data for eachset of fibers along with the respective Weibull parameters are presented in Table 5.A significant reduction in strain at break was found only for the fiber treated with10% NaOH and again the use of NaBH4 yielded a recovery in this property.


3.2. Effect of Treatment on Adhesion

The results of the pull-out tests are shown in Fig. 3 for all treatments carried out for60 min. The in natura, H2O_t60, OH5_t60_T40, OH10_t60 and OH10_BH_t60fibers showed the poorest adhesion with polyester. The adhesion of the in naturafibers was consistent with the results obtained by Towo et al. [19], who reported5.4 ± 1.2 MPa.

The best adhesion performance was found for the OH5_BH_t60 and OH5_BH_t60_T40 treatments (7.5–7.7 MPa), both with NaBH4. It is interesting to stress thatthe recovery in adhesion brought about by the NaBH4 action in the OH5_t60_T40treatment was remarkable, whereas the OH10_BH_t60 treatment was again respon-sible for very poor results, even in the presence of the NaBH4 (6.0 and 5.1 MPa,respectively). Thus, considering adhesion performance, the NaBH4 action is moresignificant under lower NaOH concentrations, at room temperature or higher.

Regarding critical length, this parameter simulates an actual load transfer situ-ation, from the matrix to the fiber, in a composite material, and is influenced bya combination of diameter and tensile strength of the fiber and fiber/matrix adhe-sion. The critical length decreased from around 5.6–6.1 mm for the in natura andH2O treated fibers to around 3.8–4.9 mm for the 2 and 5% NaOH treated fibers(Fig. 4). The OH5_BH_t60 and OH5_t60 fibers showed the lowest values perhapssuggesting that these would be the most suitable chemical treatment. Just to il-

Figure 3. Adhesion values obtained from pull-out tests for the in natura and treated fibers.


Figure 4. Calculated critical length values for the in natura and treated fibers.

lustrate, critical length was reported as 11.3 and 10.2 mm for sisal/starch basedpolymer [21] and curaua/polyester pairs [16], respectively.

4. Conclusions

The diameter measurements showed great scatter, so a meaningful estimate of fibertensile strength requires the determination of the diameters of the particular fiberbeing tested. The tensile strength data (β,T50 and the mean) showed a decreasewith increasing NaOH concentration in the chemical treatment, and the addition ofNaBH4 during treatment resulted in strength recovery. The same was noticed forthe fiber elastic modulus and the strain at break, although this was not noticed forthe 10% NaOH treatment, for which the modulus decreased: this was attributed toa greater aggregation of fibrils during the chemical treatment, with less interactionwith the solution, which affected the effectiveness of the BH4

− ion.Fiber strength, fiber adhesion to polyester and critical length calculations sug-

gest that the 5% NaOH treatment with NaBH4 for 60 min at room temperatureyielded the best combination of properties and should be favored when manufac-turing sisal/polyester composites. In all, the combination of results indicates thatthe NaBH4 is effective in minimizing degradation of the sisal fiber, validating itsuse as a protective agent for vegetable fibers during alkaline chemical treatment.This result may be extended to other vegetable fibers of similar composition andmay help promoting their general use in polymer composites.


Acknowledgements

The authors would like to thank CNPq and CAPES for the financial support and theChemistry Institute (Prof. Dr. Marly Jacobi)/UFRGS for the mechanical tests.

References

1. M. J. John and R. D. Anandjiwala, Recent developments in chemical modification and character-ization of natural fiber-reinforced composites, Polym. Compos. 29, 187–207 (2008).

2. N. M. A. Cordeiro, C. J. R. Gouveia, A. G. O. Moraes and S. C. Amico, Natural fibers characteri-zation by inverse gas chromatography, Carbohyd. Polym. 84, 110–117 (2011).

3. R. Zah, R. Hischier, A. L. Leao and I. Braun, Curaua fibers in the automobile industry — a sus-tainability assessment, J. Clean. Prod. 15, 1032–1040 (2007).

4. W. G. Trindade, W. Hoareau, J. D. Megiatto, I. A. T. Razera, A. Castellan and E. Frollini, Ther-moset phenolic matrices reinforced with unmodified and surface-grafted furfuryl alcohol sugarcane bagasse and curaua fibers: properties of fibers and composites, Biomacromolecules 6, 2485–2496 (2005).

5. X. Li, L. G. Tabil and S. Panigrahi, Chemical treatments of natural fiber for use in natural fiber-reinforced composites: a review, J. Polym. Environ. 15, 25–33 (2007).

6. A. Valadez-Gonzalez, J. M. Cervantes-Uc, R. Olayo and P. J. Herrera-Franco, Effect of fiber treat-ment on the fiber–matrix bond strength of natural fiber reinforced composites, Composites, Part B:Engng 30, 309–320 (1999).

7. T. H. D. Sydenstricker, S. Mochnaz and S. C. Amico, Pull-out and other evaluations in sisal-reinforced polyester biocomposites, Polym. Test. 22, 375–380 (2003).

8. J. T. Kim and A. N. Netravali, Mercerization of sisal fibers: effect of tension on mechanical prop-erties of sisal fiber and fiber-reinforced composites, Composites, Part A: Appl. Sci. 41, 1245–1252(2010).

9. S. C. Amico, Vegetable fibers as multifunctional materials, Materia-Rio De Janeiro 15, 386–395(2010).

10. S. Mishra, A. K. Mohanty, L. T. Drzal, M. Misra, S. Parija, S. K. Nayak and S. S. Tripathy, Studieson mechanical performance of biofibre/glass reinforced polyester hybrid composites, Compos. Sci.Technol. 63, 1377–1385 (2003).

11. M. A. Glaus and L. R. Van Loon, Degradation of cellulose under alkaline conditions: new insightsfrom a 12 years degradation study, Environ. Sci. Technol. 42, 2906–2911 (2008).

12. M. R. Gouvêa, C. Ribeiro, C. F. de Souza, I. Marvila-Oliveira, N. Lucyszyn and M.-R. Sier-akowski, Rheological behavior of borate complex and polysaccharides, Mater. Sci. Engng C: Bio.Sci. 29, 607–612 (2009).

13. J. Sartori, A. Potthast, A. Ecker, H. Sixta, T. Rosenau and P. Kosma, Alkaline degradation kineticsand CE-separation of cello- and xylo-oligomers. Part I, Carbohydr. Res. 338, 1209–1216 (2003).

14. C. J. Knill and J. F. Kennedy, Degradation of cellulose under alkaline conditions, Carbohydr.Polym. 51, 281–300 (2003).

15. A. G. O. Moraes, M. R. Sierakowski, T. M. Abreu and S. C. Amico, Adhesion evaluation of treatedsisal fiber/polyester interfaces and determination of critical lenght, in: Proc. 7th Int. Symp. NaturalPolymers and Composites, pp. 1314–1317, Gramado (2010).

16. S. N. Monteiro, R. C. M. P. Aquino and F. P. D. Lopes, Performance of curaua fibers in pullouttests, J. Mater. Sci. 43, 489–493 (2008).


17. S. S. Shapiro, in: Handbook of Statistical Methods for Engineers and Scientists, 2nd edn,H. M. Wadsworth (Ed.), pp. 6.13–6.18. McGraw-Hill, New York, USA (1998).

18. W. Nelson, Applied Life Data Analysis. John Wiley and Sons, New York, USA (1982).19. A. N. Towo, M. P. Ansell, M. L. Pastor and D. E. Packham, Weibull analysis of microbond shear

strength at sisal fibre–polyester resin interfaces, Compos. Interf. 12, 77–93 (2005).20. F. D. Silva, N. Chawla and R. D. de Toledo, Tensile behavior of high performance natural (sisal)

fibers, Compos. Sci. Technol. 68, 3438–3443 (2008).21. V. A. Alvarez and A. Vazquez, Influence of fiber chemical modification procedure on the mechan-

ical properties and water absorption of MaterBi-Y/sisal fiber composites, Composites, Part A:Appl. Sci. 37, 1672–1680 (2006).

sodium borohydride as a protective agent for the alkaline ... · pdf filetreatment of sisal...

Documents