Moisture Absorption Characteristics ofNatural Fibre Composites

J. GtRIDHAR, KISHORE AND R. M. V. G. K. R-\O*

Depar-trTu:nlof MetallurgyIndian Insrirute of ScienceBangalore-560 0/2, India

(Rccc:i\-cdJune 9. ,1985)

ABSTRACT

A comparative study was made between the moisture absorption bd13.vioursof sisal andjute fibre composites in an epoxy matrix under immersion conditions. Sisal fibres. in spiteof possessing more compact structure than jute fibres, exhibited high~r moisrure absorp-tion levels in' theiF'Compositefoffi1; contrary to'expectations,-This tendency v,"Q.Sanriburedto the high cellulose coment and a possible interfacial effect in the rormer.

INTRODUCTION't '

O F THE VARlOUS NATURAL FIBRES. JUTE AND 'SISAL FIBRES ARE OF I~lEREST

to researchers of composite'materi'}ls. These fibres are very abundant inIndia and often inexpensive as compared to other relatively adyanced man-madefibres, like glass. However, the former category are not only inferior in strength,but also suffer from a very high moistUre absP'f.ption tendency. a property thatcan be directly attribut~ to their chemical ri~fure and physical structure.

" It ,is evident from Table I that sisal fibres h~~e a higher cellulose content thanjute fibres. Sisal fibres find traditional, age-old applications in the form of highstrength ropes and floor mats. These materials have not been really examinedfrom a composite angle.

Glass and graphite fibres are noted to be impermeable not only to moisture,but also to the matrix that surrounds them in a composite. These impermeablefibre composites' obey the Fickian diffusion process [1-3]. In recent im'estiga-tions Rao et. al. [4] reponed the \alidity of the Fickian diffusion model forcarefully fabricated jute-epoxy composites, the fibres being permeable both tomoisture and the matrix resin. They showed that jute fibres absorb as much as88 % by weight of the matrix resin during the composite preparation. Theequilibrium moisture levels in this fibre (140%) were reduced to as 10\\' as 9.5 %

*Materials Science Division, National AeroIl.1utlcal Laboratory. Bangalore-560 017. India,

Joumal of REINFORCED PLASTICSAND CO\lPOSITES, Vol. 5-Apri/ 1986

0!:'1-68~/S6 02 0141-1054.S0/0

141

142 1. GIRIDHAR, KJSHORE AND R. M. V. G. K, RAo

..y

Table 1. Chemical composition of some natural fibres.

in a composite at a fibre volume fraction of 0.70.'-These significant findings'prompted a similar stUdy of another composite based again on natUral fibres.

In this paper, the authors present experimental data on sisal-epoxy compositessubjected to water immersion, with a variation in the fibre volume fraction. Theycompare the absorption trends in sisal and jute fibres before.and after 'resin im-pregnation and bring- out the disparities eXhibitedby the sisat:Jibres. They discussfurther the drawbacks of sisal fibres in forming good com,ppsites and highlightthe ''excellant wettability of jute fibres in a resin matrix. "

"... " .. ,- ... ','" THEORETICAL ASPECTS ."

(aY;Diffusion Process

Diffusion in polymer composites is by and large reported to take place ,as perthe Fickian rule. For a one dimensioI1al diffusion process through either side ofa bone dry solid of finite thickness (h), Shen & Springer [1] give the followingexpressIOn,

Mr 8 co [ - (2j + 1)2r (D,;lh2))

Mm = 1 '- r j I; 0 exp (2j + 1)2(1)

As reported by Rao et aI, for permeable fibre composites, Equation (1) can berewritten in a simplified version as

F. = 1 - :' [

1

f!l(D:r Ih2) 7r2

expexp

](2)

-«Y

so that an experimental plot of Fs versus D;r Ih2 on an analytical line representingEquation (2) reveals the extent of correlation between the experimental data andthe theory.

Lignin&

Fibre Cellulose Moisture Ash Pectins Extractives

% % % % % %

Jute 63.24 9.93 0.68 24.41 1.42

Sisal 77.20 6.20 1.0 14.50 1.1

/Ii

l/oisrure Absorprion Characrerisrics of Naruml Fibre Cmnposires1/

Ii

li The composite diffusion coefficient can be calculated from the moisture ab-,F sorption plots (Me versus Ji) sinceI

143

D'= 71'"

(

h

)

2

(

M2 - M1

)

1

c 4Mm .J[; - K

b. Nature of Natural Fibres

(3)

Considering the physical structure proper, commercial grade jute fibres aremade of fine fibrils twisted together to form continuous rovings useful for pack-ing and fastening applications. These fibres have further been proved to have a

i good potential for large scale use in the composite industry, provided certainj precautions are taken. However, lack of sufficient strength on the part of theI basic fibrils hinders their being twisted to a high degree of compaction, which

.1 imparts to them a loose and open structure in the rovings form. It is interesting'j to note, however, that the open structUre becomes almost impregnated with thet I matrix resin during the composite making process, as reported earlier by Raojl et al. This is the reason why the moisture absorption level of a virgin jute fibre; J (140% by ~eight) is drastically reduced when the jute fibre is impregnated with

""i+..,an epoxy~-n~sin~'<'~_P'-"""d- ''''''-.'""..., _.

!i Sisal fibres, on the other hand, have a more compact physical structure than: ; jute fibres. It may therefore be expected that virgin jute fibres would exhibit: ! greater resin and moisture absorption levels than sisal fibres. However, upon

i I resin impregnation, these fibres will ~xhibit m~is~re absorption p~operties .that! : very rn~%.2~}iependon the level of rysln abso~tlOn lnt~ the fi~re,:I?nor to cunng,! i the cel:f.tl.I9.~econtent of the fibre and the resm-fibre mterface.! .; . ...~.~ .

i. i. ,

; i,EXPERIMENTAL DETAILS )'

iii

: i

Bonef:Cdry test specimens 25mm X 25mm X 3mm wer~:"prepared from.unidirectional composite laminates made of commerical grade jute and sisalfibres in a hot curing epoxy matrix (LY 556, HT 972 system supplied by Mis.Ciba Geigy (I) Ltd.). These specimens were then immersed in distilled watermaintained at 298oK. The specimens were periodically withdrawn from thewater bath, wiped dry and weighed. The percentage weight increase wascalculated as a function of time till the specimens reached equilibrium absorptionlimits. These measurements were also made on sisal composites of varying fibrevolume fractions.

RESULTS AND DISCUSSION

a. ~1oisturf Absorption Trends

Figure 1 shows that the sisal fibres absorbed much less resin than the jutefibres prior to curing. Referring to Figure 2 it is seen that. being compact, thesisal fibres absorbed less moisture (about 70 % by weight) than the jute fibres(140%). Figure 3 shows that the moisture absorption level increases as the

'Y

100

~ 80a.;:)

.:xu

'0..60c...,eI

0::1.0

0

-<:-

2J

0 1 2) - '{, 3 i.

:.rf" ."{:f"; (~mih)~'"" .

Figure 1. % resin absorption in si'IJn:YI1)and jute (0) fibres.

.,

)""

200

160

80

~ 120.:L

1.0

v 0 5 10 15 20 25 30

it .(hr)'11

Figure 2. % moisture absorption in sisal (11)and (0) jute fibres.

144

-------

j

/iscure Absorprion.

.

~/ell

Ii '6//

i(1~

~

CJwmclcrislics of Narural Fihre C011l/Josiles J4S

v/"ii;

IIIiI'iI!!".'

ilLii:1

Ii!!;;

IIII!'::ii"i!.i

L.

2

0 10 20 30 ~ 50 60 70.

~.( hr)1J

Figure 3. % moisture absorption injute corrposite at different volume fractions 69-0.9; 'S-O. 7;A~Q' ,

volume fraction of jut~fit)re in the cbmposite is increased. The absorption treridin the sisal composite}~tth regard to the fibre volume fraction remains sim~lar,'

as evident from F~gu~e\4.~How~ve~. it.may be noted tha~ at a vol.ume fraqtionof 0.7 the absorptlOnfieVel of vIrgm sIsal fibre (70%, FIgure 2) IS reducep toonly 40% upon resin ~}\-npregnation,as clearly demonstrated in this figure.~:Forthe same volume fraction the absorption level of a virgin jute fibre (140 %) isbrought down to as low as 9.5% (Figure 3).

This striking disparity between jute and sisal composites can be attributed onthe one hand. to the basic compactness of sisal fibres (and hence the low moistureabsorption of virgin sisal fibres as compared to that of virgin jute fibres), andto the high cellulose content of sisal fibres (and hence the lower reduction inmoisture absorption level of virgin sisal as compared to jute upon impregnationwith the resin).

Thus, while the basic compactness in the physical structure accounts for theabsorption level in an unimpregnated (virgin) fibre, the chemical nature(cellulose content) appears to govern largely the absorption behaviour of an im-pregnated fibre (the composite).

Now looking at the virgin fibres, the diffusivity of a sisal composite is a func-tion of the diffusion coefficients of the fibre (D!) and the resin (Dr). For the jutefibre, however, the composite diffusi\'ity is controlled mostly by the resin im-pregnated jute fibre (D;).

,,,

'2

10

.

-e: 8-

6

- -146 1. GIRIDHAR. KISHORE AND R. M. V. G. K. RAo

56

14

~\L .

48

1.0

-:- 32-:L

16

..,':s

J {WI'

," ('-.r""'-~.

JV,i [1

a 10 L'uJ\',1l20 30 I 1.0 50 60 70;,'.1' Jt.(hr)tJ

Figure 4. % moisture absorption'in sisal composite at cifferentvoume fractions 0-0.8; tl-O.7.

I

During experiments, it was seen that sisal fibres seldom wetted satisfactoril)in the resin matrix, while the wetting of jure fibres was very good. This alscleads to the conclusion that sisal fibres donor lend themselves as attractive rein.forcements in polymer matrices unless otherwise surface treated. It was in facinoticed that the fibres withdrew easily from the laminates when pulled from oneend, with cured resin sticking all along the fibres. This then lends some credenceto a possible interfacial factor also contributing to the high absorption levels irthe sisal composites.

b. Fickian Correlation in Sisal-Epoxy Composites

Figure 5 shows the superimposition of the experimental data for the sisal.composite on the analytical curve representing Equation (2). A good correlatiorcan be noticed for the composite at a fibre volume fraction of 0.7. Deviations~however, appear at the higher fibre fraction (0.8). Jute fibre composites, howeve[exhibited very good correlation at all volume fractions considered, as reportedby Rao et al.

This can be attributed largely to the non-wening surface characteristics of thesisal fibres, as observed in these experiments. In a sisal composite, the fibre ap'peared as if encapsulated in a resin casting, while in the ,case of the jute com.posite, the fibres have lost their identity.

-- -147

1-0

I/

lMoiscure Absorplion CJw.racteristics of Natural Fibre Composites;/iF"

;lf

(:1

.1

II

~i

0-8 Analytical OJrv

E::f-:.. (}.6::f"lr'

0.

0-2

I

I

I

I

'I

;1

i''jj

'jii

-6 -5 -I. -3 -2 -1 0 1 2In(Oc t/h~)

Figure 5. % comparison of analytical (Eqlladon2) and measured "F:' va.Uesfor the sisal com-posite. (298K, water immersion) E)-0.8; h.-D. Z

c. Composite Diffusivities of Natural Fibre Composites-,' ". . -' .. ' . . , .. .

The diffusion coefficients of the jute and sisal composites as calculated fromEquation (3) are presented in Figure 6. The overall diffusion coefficient increaseswith the respective fibre volume fractions for both types of composites as ex-pected. The variation of the diffusivity index (D:/Dr) as given in Figure 7,however, indicates that the diffusivity'of the sisal epoxy composite is more sen-sitive to the fibre volume fractidm:than that of the jute composite. This may berelated to the surface characteristics of these fibres.

Table 2 shows the overall diffusion characteristics of neat-epoxy, jute-epoxy,and sisal-epoxy composite systems);"'; , .

.At this juncture, it may be worthwhile to note that even the advanced aramidfibre composite absorbs moisture [5] to levels as high as 6% under immersionconditions and that the absorption increases with an increase in the fibre volumefraction [6]. This, however, doesn't seem to hinder the application of these com-posites in advanced areas. Nevertheless, it does attract the attention of moistureabsorption analysts, and it is very appropriate to establish to what extent the highlevels of moisture noticed in the aramid composites affect their mechanicalproperties.

CONCLUSIONS

Sisal fibres are physically more compact and hence absorb less moisture thanjute fibres. Their composites, however, exhibit much higher absorption levelsthan jute composites. This can be attributed largely to a high cellulose contentin sisal fibres and to some extent to the poor fibre-resin interface noticed in thesecomposites.

Sisal fibres appear much less promising than jute fibres, in an untreated condi-tion, in the fabrication of composites. The cost of sisal fibres is also high as com-

-110

-810

,.U0.1~

..Eu ~u 10

0

J-II

6x~.: on,. :: '0', ,. _O:CilL

.5x "G::0.1 0,0 0,7 0-9 '-0

Vf ,'It'.'

Figure 6. Comparison of diffusion coefficient of jute a.IK:J sisal composites at ditter;?.T!JJ.jfitxe frac-Dons.~sisaJ-epoxy. mm -jute-epoxy. Da -epoxy-resin. .JCJX'/

-I: 40

. i

3Z

8

24

H0

16

a 0.2 0., 0,6 (}aI

Vf' Vf

Figure 7. Variationof ditfusivity index (Dr)with fibre fracDons. 0-sisa/-epoxy. 0-jute-epoxy.

--

foisture Ab.\cr[Jtioll Characteristics of Nalllral FilJre Composites 149

Table 2. Diffusion properties of overall jute-epoxy andsisal-epoxy composites.

Jute Epoxy Composite Sisal Epoxy Composite,

Diffusion Property

Equilibrium Moisture'Content (Mm(%)

Diffusion Coefficient D~or Dc(cm2sec.')

Applicability of FickianDiffusion Model

Effect of Fibre VolumeFraction

3.2 8.5

v,' = 0.9

11.2

v, = 0.741.6

VI = 0.8Epoxy Resin V( = 0.7

49.0

8.3 X 10"0 4.5 x 10" 2.66 x 10" 1.66 x 10" 2.92 x 10"

Applicable Applicable Applicable

Increase Diffusivity Increase Diffusivity

DiffusivityIndex 0, 1 5.42 32.1 20.1 35.1

pared to jute, which places another limitation on their use in low cost compositedevelopment.

NOl\ffi'i CLATlJJ1E

.IV;-""""F"..., -

s

..,' .?., h-"":"

,,,,i'i'}vI, -

," '. M",,-

'.'/'~1' {-

Diffusion coefficient of permeable fibre compositeFractional moisture absorption. ",!,Specimen thickness .

Moisture content at any ti~eEquilibrium moisture contentExposure time

-',' !;'.'..,

. Cm2 See'(M,! Moo)Cm%%hr

""':":-<'."

- ,:~:'{rACKNO\VLEDGEl\1ENTS

The authors are grateful to the authorities of IISc and NAL for the facilitiesprovided at the Department of Metallurgy and the Materials Science Division,respectively. They would like to thank Dr. A. K. Singh, Head, Materials ScienceDivision for his help and interest in the work.

REFERENCES

1. Shen, Chi-Hung and Springer, George S.. ~Moisture Absorption and Desorption of CompositeMaterials,~ J. Composite Materials. Vo\. 10. pp. 2 - 20 (1976).

2. Springer, George S., "Moisture Contents of Composites under Transient Conditions," J. Com-posite Materials, Vo\. 11, pp. 107-122 (1977).

3. Romanenkov, I. G. and Machavariani, Z. P., "Water Absorption of GRPS," Sovier Plasrics, pp.49-51 (June 1966).

4. Rao,R. M.V. G. K., Chanda,Manasand Balasubramanian, N., "A Fickian Diffusion Model forPermeable Fibre Polymer Composites.- J. Reinforced Plastics and Composires, Vol. 2 (October1983).

150 1. GIRlDHAR, KlSHORE AND R. M. Y. G. K. RAo

5. Lo, S. Y. and Hahn, H. T., -Swelling of Kevlar-49/Epoxy and S2 Glass/Epoxy Composites,~ Pro-gress in Science and Engineering of Composires (1982).

6. Gopalan, R., Rao, R. M. V. G. K., Murthy, M. V. V. and Dattaguru, B., uDitfusion Srudies onHigh Perfonnance Hybrid ComfX)sites,"Communicated to the J. Reinforced Plasrics and Com-posires, for publication.

',..

0<",,,,,~,~~-,'"'''' ,'w'<~";:';~&"~",,,",=.= ~ _......