Composites: Part A 35 (2004) 1267–1276

www.elsevier.com/locate/compositesa

Comparison of water absorption in natural cellulosic fibres from

wood and one-year crops in polypropylene composites
and its influence on their mechanical properties
Ana Espert, Francisco Vilaplana, Sigbritt Karlsson*
Department of Fibre and Polymer Technology, Royal Institute of Technology, SE-10044 Stockholm, Sweden
Received 19 September 2003; revised 29 March 2004; accepted 7 April 2004

Abstract
Environmentally beneficial composites can be made by replacing glass fibres with various types of cellulose fibres. Fibres from pine or
eucalyptus wood and also one-year crops such as coir, sisal, etc. are all good candidates. The poor resistance towards water absorption is one
of the drawbacks of natural fibres/polypropylene composites. New natural fibres/polypropylene composites were made and the water
absorption in them was studied by immersion of the composites in water at three different temperatures, 23, 50 and 70 8C. The process of
absorption of water was found to follow the kinetics and mechanisms described by Fick’s theory. In addition, the diffusivity coefficient was
dependent on the temperature as estimated by means of Arrhenius law. A decrease in tensile properties of the composites was demonstrated,
showing a great loss in mechanical properties of the water-saturated samples compared to the dry samples. The morphology change was
monitored by scanning electron microscopy studies of the samples before and after exposure to water and the devastating effect of water on
the fibre structure was shown.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: A. Polymer–matrix composites (PMCs); B. Mechanical properties; Water absorption

1. Introduction the natural fibres present lower density, less abrasiveness,

lower cost and they are renewable and biodegradable [2,3].
In the past decades, growing environmental awareness Moreover, their mechanical performance can be compared
has resulted in a renewed interest in the use of natural to that of synthetic fibres nowadays used and they present
materials for different applications [1]. New and harder good thermal properties [4]. The fact that they are fibres
environmental policies have forced industries like the obtained directly from natural resources makes them even
automotive, packaging and construction industry to search more attractive in terms of sustainability and environmental
for new materials that can substitute the traditional awareness. Unfortunately, several disadvantages of natural
composite materials consisting of a plastic matrix and fibres such as thermal and mechanical degradation during
inorganic filler as reinforcement. The most used composites processing can make them undesirable for certain
nowadays are the glass fibre filled thermoplastics, but also applications.
different inorganic fillers like aramid or carbon fibres are However, the main disadvantage of these natural
widely used. However, inorganic fibres present several fibre/polymer composites seems to be the incompatibility
disadvantages, like for example their non-biodegradability, between the hydrophilic natural fibres and the hydrophobic
the abrasion that they produce in the processing equipment thermoplastic matrix that makes necessary the use of
and the health problems that they cause to workers due to compatibilizers or coupling agents in order to improve the
the skin irritations they cause during processing and adhesion fibre/matrix [5 –8]. For the purpose of making
handling. Compared to the traditional synthetic fillers, engineering parts with this type of composites, a wide study
on the effect of different coupling agents, such as silanes,
* Corresponding author. Tel.: þ 46-8-790-8581; fax: þ46-8-100-775. maleic anhydride grafted polypropylene (PPgMA) or
E-mail address: sigbritt@polymer.kth.se (S. Karlsson). modification by acetylation, has been reported by these
1359-835X/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compositesa.2004.04.004
1268 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276

authors [9]. The most favourable properties were obtained constants. The value of coefficient n shows different
by modification with PPgMA, which was demonstrated to behaviour between cases; for Fickian diffusion n ¼ 1=2;
be the most suitable compatibilization method for this type while for Case II n ¼ 1 (and for Super Case II n . 1). For
of composites. anomalous diffusion, n shows an intermediate value ð1=2 ,
However, several considerations have to be taken into n , 1Þ: Moisture absorption in natural fibre reinforced
account in the design of natural fibres/PP composites. One plastics usually follows Case I Fickian behaviour, so, further
of the most important issues is the degradation behaviour of attention will be focused on its study.
the composites exposed to environmental conditions such as As mentioned before, apart from diffusion, two other
humidity, sunlight or microorganisms. The poor resistance minor mechanisms are active in moisture exposure of
of the fibres to water absorption can have undesirable effects composite materials. The capillary mechanism involves the
on the mechanical properties and the dimensional stability flow of water molecules into the interface between fibres
of the composites [10 – 15]. Therefore, it is important to and matrix. It is particularly important when the interfacial
study in detail the water absorption behaviour in order to adhesion is weak and when the debonding of the fibres and
estimate not only the consequences that the water absorbed the matrix has started. On the other hand, transport by
may have, but also how this water uptake can be minimized microcracks includes the flow and storage of water in the
in some way. cracks, pores or small channels in the composite structure.
Moisture penetration into composite materials is con- These imperfections can be originated during the processing
ducted by three different mechanisms. The main process of the material or due to environmental and service effects.
consists of diffusion of water molecules inside the The objective of this work was to relate kinetics and
microgaps between polymer chains. The other common characteristics of the water absorption in natural fibres/poly-
mechanisms are capillary transport into the gaps and flaws propylene composites. Water uptake leads to the degra-
at the interfaces between fibres and polymer, because of dation of the fibres and the fibre – matrix interface resulting
incomplete wettability and impregnation; and transport in a loss of mechanical performance. The differences in
by microcracks in the matrix, formed during the comp- mechanical properties due to type of fibre, type of matrix
ounding process [11,16]. In spite of the fact that all three and fibre content are reported [11,15].
mechanisms are active jointly in case of moisture exposure
of the composite materials, the overall effect can be
modelled conveniently considering only the diffusional 2. Experimental
mechanism.
In general, diffusion behaviour in glassy polymers can be For the preparation of composites, two different kinds of
classified according to the relative mobility of the penetrant polypropylene were used: a polypropylene homopolymer
and of the polymer segments. With this, there are three HH 315MO supplied by Borealis, Sweden and a post-
different categories of diffusion behaviour [16]. industrial polypropylene containing around 3 wt% of
Case I, or Fickian diffusion, in which the rate of diffusion ethylene vinyl acetate copolymer (EVA), PP/EVA supplied
is much less than that of the polymer segment mobility. The by Polykemi, Sweden. Four different fibres were used as
equilibrium inside the polymer is rapidly reached and it is fillers for the composites: cellulose fibres from the Kraft
maintained with independence of time. pulping process supplied by Komotini, Greece; sisal fibres,
Case II (and Super Case II), in which penetrant mobility coir fibres and Luffa sponge fibres supplied by the
is much greater than other relaxation processes. This Department of Mechanical Engineering of the University
diffusion is characterized by the development of a boundary of Dar es Salaam, Tanzania. In order to improve the
between the swollen outer part and the inner glassy core of compatibilization between the fibres and the matrix,
the polymer. The boundary advances at a constant velocity polypropylene –maleic anhydride copolymer, PPgMA Epo-
and the core diminishes in size until an equilibrium lenew supplied by Eastman, USA was employed.
penetrant concentration is reached in the whole polymer. The fibres were physically and chemically characterized.
Non-Fickian or anomalous diffusion occurs when the The physical characterization was realized using a Fiber-
penetrant mobility and the polymer segment relaxation are master equipment getting detailed information about
comparable. It is then, an intermediate behaviour between diameter, length and ratio aspect of the fibres. The chemical
Case I and Case II diffusion. characterization consisted of two methods: the SCAN-
These three cases of diffusion can be distinguished CM49 for determination of the acetone extract and the
theoretically by the shape of the sorption curve repre- AH23-18 for determination of the carbohydrates, the Klason
sented by lignin and the acid soluble lignin.
Mt Blends of the different polymeric matrix and different
¼ ktn ð1Þ fibre matrix varying in contents from 10 up to 30 wt% of
M1
fibre were prepared in a Brabender mixer at 190 8C and
where Mt is the moisture content at time t; M1 is the 70 rpm at approximately 10 min. The resulted blends were
moisture content at the equilibrium; and k and n are chopped and then films were prepared by compression
 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276 1269

Table 1
Physical characteristics and chemical composition of the fibres

Fibre Average length (mm) Average diameter (mm) Average shape factor (%) Cellulose Hemicelluloses Lignin Others

Cellulose from pulp 0.624 29.6 90.01 71.71 13.80 3.92 10.57
Sisal 0.615 25.9 94.36 56.52 16.49 10.62 16.36
Coir 0.838 38.4 93.36 33.00 17.91 36.14 12.95
Luffa sponge 0.500 34.5 92.59 50.15 15.57 14.41 19.87

moulding at 190 8C during 3 min. Compression moulded universal electromechanical testing machine using a gauge
was chosen since only small quantities of material were distance of 32 mm and a load of 0.1 kN at a strain speed of
available. The conditions in which compression moulding 5 mm/min. The specimens used were rectangular, 70 mm
was performed were chosen in order to reduce the formation long £ 5 mm wide and a thickness of 0.5 mm.
of voids and to improve the fibre distribution. The fibres in In order to understand better the effect of water on the
the films were randomly oriented. The films were cut in composites structure, scanning electron microscopy (SEM)
specimens 70 mm £ 5 mm £ 0.5 mm in order to perform experiments were also performed.
the water absorption experiments and the posterior tensile
tests.
Water absorption studies were performed following the 3. Results and discussion
ASTM D 570-98 standard. Six specimens of every sample
were submerged in distilled water at different temperatures, 3.1. Characterization of fibres
23, 50 and 70 8C. The specimens were removed from the
water at certain periods of time and weighted in a high By physical characterization, lengths and shape of the
precision balance and then submerged again in water. The fibres are given. These parameters may influence the
content of water was calculated by the weight difference. resulting mechanical properties of composites. The results
When the content of water remained invariable in the from the physical characterization of the fibres are presented
specimens, they were removed and tensile tests were in Table 1. The shape factor describes the degree of
performed in order to determine the mechanical properties. extension of the fibre. This measure goes from 0 to 100%,
Tensile tests were performed both in the dry and the where 100% means a completely straight fibre. In this case,
water-saturated samples in order to determine the influence the fibres analysed present very high values of average
of the water content in the mechanical properties following shape factor, which could be in some way expected since the
the ASTM D-882 standard. Tensile tests were performed at fibres are not so long, usually a longer fibre has a lower
23 8C and 40% relative humidity in an Instron 5566 shape factor.

Fig. 1. Absorption curve for PPEVA/cellulose composites immersed in water at 50 8C.

1270 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276

Fig. 2. Water absorption curves for 30sisalPP composite at different temperatures.

The chemical characterization revealed the proportion of the cellulose that presents the lowest. If we compare them,
each component in the fibres. The results are presented in coir is taken from the coconut fruit, which presents high
Table 1. Important differences were found depending on the rigidity and hardness, indicating its high content of lignin.
different types and precedence of the fibres, especially in On the other hand, cellulose fibres have already been treated
what refers to cellulose and lignin contents. Coir is the fibre to eliminate the lignin and only have the lignin remaining
that presents the highest content of lignin in contrast with after the pulping process.

Fig. 3. Diffusion case fitting plots for sisal/PP composites at 23 8C.

 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276 1271

3.2. Kinetics of the water absorption Table 2

Diffusion case selection parameters
The amount of water absorbed in the composites was Sample Temperature (8C)
calculated by the weight difference between the samples
exposed to water and the dried samples. Fig. 1 shows the 23 50 70
percentage of water absorbed plotted against time for all the
n k (h2) n k (h2) n k (h2)
samples. The same behaviour was observed, the samples
absorbed water very rapidly during the first stages (0 – 20 h) 10cell-PP 0.4145 0.1278 0.3008 0.1018 0.3691 0.3950
reaching a certain value, the saturation point, where no more 20cell-PP 0.4603 0.2034 0.3026 0.1710 0.5142 0.6616
water was absorbed and the content of water in the 30cell-PP 0.5016 0.2503 0.3831 0.2588 0.4739 0.7563
composites remained the same. The hydrophilic character 10sisal-PP 0.4993 0.7849 0.4219 0.1070 0.3906 0.4419
20sisal-PP 0.5414 0.1007 0.4225 0.1162 0.6825 0.4036
of natural fibres is responsible for the water absorption in the 30sisal-PP 0.5806 0.1044 0.6531 0.1386 0.4984 0.3468
composites, and therefore a higher content on fibres leads to 10coir–PP 0.5545 0.1495 0.3938 0.0824 0.5369 0.5103
a higher amount of water absorbed. The matrix had little 20coir–PP 0.6397 0.1397 0.3858 0.1022 0.4473 0.7019
effect on the amount of water absorbed as pure PP 30coir–PP 0.5890 0.1831 0.3450 0.1550 0.4275 0.8004
composites and recycled PP composites demonstrated 10luffa-PP 0.5790 0.1582 0.4399 0.1573 0.4929 0.3998
20luffa-PP 0.5908 0.0937 0.6386 0.1299 0.4834 0.5120
around the same water content, between 5.8 and 6.3% 30luffa-PP 0.6067 0.2007 0.4783 0.1928 0.5149 0.8200
after 100 h of exposure.
The temperature of the absorption process also has an 10cell-PP/EVA 0.4563 0.0783 0.5364 0.1018 0.4847 0.1900
20cell-PP/EVA 0.4977 0.1091 0.4871 0.1710 0.3789 0.3254
influence on the water absorption curves. Fig. 2 shows the 30cell-PP/EVA 0.5227 0.1208 0.4390 0.2588 0.3632 0.4704
absorption curves for 30sisalPP samples at different 10sisal-PP/EVA 0.6202 0.0415 0.5053 0.1070 0.4636 0.1848
temperatures. It can be concluded that the temperature 20sisal-PP/EVA 0.5432 0.0527 0.5633 0.1162 0.5094 0.2466
increases the water uptake of the composite materials. When 30sisal-PP/EVA 0.5461 0.0730 0.5922 0.1386 0.5115 0.3779
the fibre content and the temperature of immersion are 10coir–PP/EVA 0.5957 0.0523 0.5643 0.0824 0.4691 0.2008
20coir–PP/EVA 0.4750 0.0817 0.5237 0.1022 0.4775 0.3016
increased, the saturation time is shortened. For example, for 30coir–PP/EVA 0.5308 0.0731 0.5420 0.1550 0.4977 0.2999
the PPEVA/cellulose composites, the saturation time is 10luffa-PP/EVA 0.5767 0.1501 0.5022 0.1573 0.3639 0.4116
shortened by 168 h as the fibre content increases from 10 to 20luffa-PP/EVA 0.5055 0.0586 0.5849 0.1299 0.6152 0.2688
30%. When the temperature is increased from 23 to 70 8C, 30luffa-PP/EVA 0.6002 0.0925 0.5295 0.1928 0.6377 0.2986
the saturation time is reduced by 62 h for the composite with
30% sisal and PP, as it can be seen in Fig. 2. Recycled PP
composites showed longer saturation times than their the composite structure. For small times ðMt =M1 # 0:5Þ;
corresponding composites with pure PP as matrix. This the following equation can be used [17]
may be a sign about an improved compatibility of fibre

Mt 4 D 0:5 0:5
matrix interface due to the presence of EVA. EVA presents ¼ t ð3Þ
M1 L p
on one hand hydrophilic acetate groups that can bind to the
cellulosic fibres and on the other hand a polymeric chain where L is the thickness of the sample.
that can interact with the polymeric matrix, acting this way, By Eq. (3), the moisture diffusion coefficient is obtained
EVA as a coupling agent between fibres and matrix. from the slope of the linear part of the plot of Mt =M1 vs.
The analysis of the diffusion mechanism and kinetics was ðtimeÞ0:5 b21 ; where b represents the dimensions for the
performed based on the Fick’s theory and adjusting the thickness of the sample. Table 3 shows the values of
experimental values to the equation the moisture diffusion coefficient obtained for the fitting of

 the linear part to Eq. (3). The results for the diffusion
Mt
log ¼ logðkÞ þ n logðtÞ ð2Þ coefficient are in agreement with the results reported by
M1
other authors [15]. The increase in the fibre loading led to
which is derived from Eq. (1). higher values of the diffusion coefficients for all the different
Fig. 3 shows an example of the fitting of the experimental blends of composites. Due to the hydrophilic character of
data to Eq. (2). In Table 2, the values of the parameters n and natural fibres, the inclusion of water molecules inside the
k resulting from the fitting are shown for all the samples. composite material is favoured as demonstrated by the
The absorption of water in natural fibres/polypropylene kinetics of the diffusion processes. Also, the expanded
composites approaches towards the Fickian diffusion case, interfacial region, that is a result of the fibre loading,
as the values of n are very similar for all the samples and benefits the transport of the diffusant. Pure PP composites
they are very close to the value of n ¼ 0:5: present higher values of diffusion coefficient than their
Next step was the performance of an analysis of the respective composites with recycled PP as matrix. This
parameters of this theoretical model. The diffusion coeffi- observation can be an effect of the EVA, present in recycled
cient is the most important parameter of the Fick’s model, as PP acting as a coupling agent. With better adhesion between
this shows ability of solvent molecules to penetrate inside matrix and fibres, the velocity of the diffusional processes
1272 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276

Table 3 processing that could accelerate the diffusion in this case.

Diffusion coefficients for the composite materials at different temperatures From the results of the diffusion coefficient, any clear
Sample Diffusion coefficient Sample Diffusion coefficient conclusion about the influence of the type of fibre can be
(m2/s) £ 1013 (m2/s) £ 1013 extracted. It seems to be that what really influences the
moisture absorption behaviour regarding the nature of
23 8C 50 8C 70 8C 23 8C 50 8C 70 8C fibres, is the compatibilization between fibres and matrix.
With this, it appears that sisal fibre composites present
10cell-PP 1.67 7.81 16.5 10cell-PP/EVA 0.64 1.71 5.11
20cell-PP 5.07 20.6 71.4 20cell-PP/EVA 1.33 5.26 12.8
special good behaviour against water absorption, since their
30cell-PP 8.84 27.9 103.1 30cell-PP/EVA 3.41 8.34 29.9 diffusion coefficient values are generally the lowest.
10sisal-PP 0.83 6.01 27.1 10sisal-PP/EVA 0.54 1.67 4.69 The experimental diffusion results were compared with
20sisal-PP 2.30 10.3 45.3 20sisal-PP/EVA 0.74 3.88 10.3 the theoretical mathematical expression of the Fick’s law
30sisal-PP 3.84 13.3 69.5 30sisal-PP/EVA 1.85 7.27 31.2 for a film geometry, which is given by the equation:
10coir– PP 4.34 10.4 33.9 10coir–PP/EVA 0.86 1.86 6.93
!
20coir– PP 6.11 22.4 66.6 20coir–PP/EVA 1.07 2.59 15.2
Mt 8 X1
1 ð2n þ 1Þ2 p2
30coir– PP 10.9 39.3 111.8 30coir–PP/EVA 1.26 4.92 21.0 ¼12 2 exp 2 Dt ð4Þ
10luffa-PP 5.30 12.7 23.6 10luffa-PP/EVA 5.02 4.20 25.7 M1 p n¼0 ð2n þ 1Þ2 L2
20luffa-PP 3.08 19.3 53.4 20luffa-PP/EVA 0.66 5.68 29.8
30luffa-PP 18.30 45.1 171.5 30luffa-PP/EVA 5.07 9.54 40.0 Fig. 4 shows this comparison for PPEVA/cellulose
composites, for different fibre content and different
temperatures. Therefore, it can be concluded that the
decreases since there are less gaps in the interfacial region experimental data fit reasonably well the Fickian mode of
and also more hydrophilic groups as hydroxyls are blocked diffusion, especially for composites with low fibre content at
by the coupling effect. Theoretically, composites with fibres the initial stage of diffusion.
containing higher lignin content as fillers should present The diffusion processes are activated by an increase in
lower values of water uptake, since lignin is a hydrophobic temperature, as may be expected for a process that
compound that acts against hydrothermal degradation in the involves mobility of molecules. In case of Fickian
fibres. Therefore, coir composites should possess the lowest diffusion, the diffusion coefficient is supposed to follow
values of moisture diffusion coefficient, because of their an exponential Arrhenius-type relation with temperature as
high lignin content. Nevertheless, this behaviour is not expressed by
reproduced on the experimental values for the diffusion

coefficients shown in Table 3. This could be possible due to E
D ¼ D0 exp 2 a ð5Þ
a higher porosity or to the existence of voids formed during RT

Fig. 4. Comparison of the experimental data and the theoretical curves for PPEVA/cellulose composites with different fibre content at 50 8C. The points
represent the experimental data and the lines are the theoretical curves.
 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276 1273

Fig. 5. Arrhenius dependence of the diffusion coefficients for coir–PP composites.

where D0 is the permeability index, Ea is the activation due to the effect of the water molecules, which change
energy of the diffusion process and R is the universal gas the structure and properties of the fibres, matrix and the
constant. interface between them. Once the moisture penetrates
Fig. 5 is an example of the Arrhenius dependence of the inside the composite materials, the fibres tend to swell.
diffusion coefficient with temperature for different cellu- The matrix structure can also be affected by the water
lose-PP/EVA composites. Table 4 presents the results of the uptake by processes such as chain reorientation and
Arrhenius parameters, activation energy of the diffusion shrinkage. Hydrothermal ageing may also lead to the
process Ea and permeability index D0 : degradation of natural fibres by a hydrolysis mechanism.
Water absorption and their resulting effects contribute to
3.3. Effects of water absorption on the mechanical the loss of compatibilization between fibres and matrix,
properties which results in debonding and weakening of the
interface adhesion.
Mechanical tests were performed in all the samples Concerning the stress at maximum load, their values are
before and after water absorption. Figs. 6 –8 present somewhat reduced after water absorption for all the
these results. In general, the mechanical properties of composite blends. This effect is particularly evident for the
these materials are decreased after the moisture uptake, composites with higher fibre content, in which the maximum

Table 4
Arrhenius parameters for the dependence of the diffusion coefficient with the temperature

Sample Arrhenius parameters Sample Arrhenius parameters

Ea (kJ/mol) D0 (m2/s) Ea (kJ/mol) D0 (m2/s)

10cell-PP 4.99 3.67 £ 1026 10cell-PP/EVA 4.39 1.66 £ 1027

20cell-PP 5.65 9.28 £ 1025 20cell-PP/EVA 4.89 1.98 £ 1026
30cell-PP 5.20 3.39 £ 1025 30cell-PP/EVA 4.56 1.48 £ 1026
10sisal-PP 7.50 7.97 £ 1023 10sisal-PP/EVA 4.61 2.95 £ 1027
20sisal-PP 6.34 4.23 £ 1024 20sisal-PP/EVA 5.72 1.84 £ 1025
30sisal-PP 6.09 2.91 £ 1024 30sisal-PP/EVA 5.99 1.05 £ 1024
10coir–PP 4.32 8.64 £ 1027 10coir–PP/EVA 4.35 2.95 £ 1027
20coir–PP 5.10 1.74 £ 1025 20coir–PP/EVA 5.50 1.05 £ 1025
30coir–PP 4.99 2.20 £ 1025 30coir–PP/EVA 5.97 6.54 £ 1025
10luffa-PP 3.22 2.73 £ 1028 10luffa-PP/EVA – –
20luffa-PP 6.19 3.84 £ 1024 20luffa-PP/EVA 8.19 6.47 £ 1022
30luffa-PP 4.68 1.21 £ 1025 30luffa-PP/EVA 4.27 7.84 £ 1027
1274 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276

Fig. 6. Effect of the water absorption on the Young’s modulus for cellulose/PP composites.

stress gets drastically reduced. High temperatures of the the composite material, which should lead to an increase
moisture absorption also favour the decrease in the stress at of the maximum strain for the composites after water
maximum load. absorption [15]. Such an increase is shown for compo-
The changes in the strain at maximum load with sites with a low fibre content immersed in water at low
hydrothermal ageing are more complex. It has been temperatures, and in particular for those that employed
reported that water molecules act as a plasticizer agent in recycled PP as matrix. However, the maximum strain is

Fig. 7. Effect of the water absorption on the stress for cellulose/PP composites.
 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276 1275

Fig. 8. Effect of the water absorption on the strain for cellulose/PP composites.

instead reduced after hydrothermal ageing of composites 4. Conclusions

with higher fibre content and higher immersion tempera-
ture. The explanation is that the plasticizer effect of Water absorption on natural fibres/PP composites was
water only occurs at low water content, thus preferen- proved to follow the kinetics of a Fickian diffusion process,
tially for composites with low fibre loading, which use where the kinetics parameters are influenced by the content
recycled PP as matrix and which are immersed at low of fibre, the type of matrix and mainly, the temperature by
temperatures. For the other cases, the degradation effect means of the Arrhenius law.
of water in the structure of the material is superior to Mechanical properties are dramatically affected by the
the plasticizer effect, so that the maximum strain is water absorption. Water-saturated samples present poor
reduced. mechanical properties such as lower values of Young’s
SEM photographs clearly show the degradation of the modulus and stress at maximum load. For samples with low
fibres by the water absorption. Fig. 9 shows the composite fibre content, water acts in some way as a plasticizer leading
structure before and after water absorption, where the loss of to slightly higher values of strain, although this effect
adhesion between fibre and matrix, characterized by the disappears with the increase in temperature or in fibre
apparition of voids can be noticed. The fibres appear content.
seriously degraded, with complete lack of pectin layer and The use of a post-industrial polypropylene with a low
the microfibrils can be clearly observed. percentage of EVA in its composition in some cases leads to

Fig. 9. Scanning electron micrographs for 20sisal-PP/EVA composite before (a) and after (b) hydrothermal ageing at 50 8C.
1276 A. Espert et al. / Composites: Part A 35 (2004) 1267–1276

improved resistance to water absorption. It is believed then [6] Gassan J. A study of fiber and interface parameters affecting the
that EVA can further improve the compatibility between fatigue behaviour of natural fiber composites. Compos, Part A: Appl
Sci Manufact 2002;33A(3):369–74.
fibres and matrix. EVA is a copolymer of ethylene, which
[7] Gauthier R, et al. Interfaces in polyolefin/cellulosic fiber composites:
permits the adhesion to the PP matrix, and vinyl acetate, chemical coupling, morphology, correlation with adhesion and aging
which could bond due to its acetate groups to the hydroxyl in moisture. Polym Compos 1998;19(3):287 –300.
groups on the fibres. [8] George J, Sreekala MS, Thomas S. A review on interface modification
and characterization of natural fiber reinforced plastic composites.
Polym Eng Sci 2001;41(9):1471 –85.
[9] Espert A, Camacho W, Karlson S. Thermal and thermomechanical
Acknowledgements properties of biocomposites made from modified recycled
cellulose and recycled polypropylene. J Appl Polym Sci 2003;89(9):
The authors would like to thank the European Commu- 2353–60.
nity through the research project ECOSITES G5RD-CT- [10] Kajaks JA, et al. Effect of water on the physicomechanical properties
2000-00337 for the financial support. of composites containing low-density polyethylene and linen yarn
production waste. Mech Compos Mater 2001;37(2):167–70.
[11] Lin Q, Zhou X, Dai G. Effect of hydrothermal environment on
moisture absorption and mechanical properties of wood flour-
References filled polypropylene composites. J Appl Polym Sci 2002;85(14):
2824–32.
[1] Mohanty AK, Misra M, Drzal LT. Sustainable bio-composites from [12] Stamboulis A, et al. Environmental durability of flax fibres and their
renewable resources: opportunities and challenges in the green composites based on polypropylene matrix. Appl Compos Mater
materials world. J Polym Environ 2002;10(1/2):19 –26. 2000;7(5–6):273–94.
[2] Bledzki AK, Reihmane S, Gassan J. Thermoplastics reinforced with [13] Tajvidi M, Ebrahimi G. Water uptake and mechanical characteristics
wood fillers: a literature review. Polym Plast Technol Eng 1998;37(4): of natural filler-polypropylene composites. J Appl Polym Sci 2003;
451–68. 88(4):941–6.
[3] Bledzki AK, Gassan J. Composites reinforced with cellulose based [14] Ishak ZAM, et al. Hygrothermal aging and tensile behaviour of
fibres. Prog Polym Sci 1999;24(2):221– 74. injection-molded rice husk-filled polypropylene composites. J Appl
[4] Amash A, Zugenmaier P. Morphology and properties of isotropic and Polym Sci 2001;81(3):742–53.
oriented samples of cellulose fiber – polypropylene composites. [15] Joseph PV, et al. Environmental effects on the degradation behaviour
Polymer 1999;41(4):1589– 96. of sisal fibre reinforced polypropylene composites. Compos Sci
[5] Gassan J, Bledzki AK. Possibilities to improve the properties of Technol 2002;62(10 –11):1357–72.
natural fiber reinforced plastics by fiber modification—jute poly- [16] Comyn J, editor. Polymer permeability. 1985. 383 pp.
propylene composites. Appl Compos Mater 2000;7(5–6):373–85. [17] Crank J. The mathematics of diffusion; 1956, 347 pp.