Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

J Mater Sci (2015) 50:7660–7672

DOI 10.1007/s10853-015-9330-4

Thermo-compression forming of flax fibre-reinforced polyamide 6


composites: influence of the fibre thermal degradation
on mechanical properties
Shaoxiong Liang1 • Hedi Nouri1 • Eric Lafranche1

Received: 24 April 2015 / Accepted: 3 August 2015 / Published online: 12 August 2015
Ó Springer Science+Business Media New York 2015

Abstract The thermal degradation of plant-based fibres at [1–3]. Thanks to their inherent advantages such as low
high temperature is a main issue when manufacturing vegetal density and biodegradability, with specific mechanical
fibres/thermoplastic composites due to the high melting point properties comparable to those of glass fibres, natural fibres
of engineering polymers such as polyamide 6 (PA6). This such as flax, hemp, jute, sisal, pineapple and wheat are
paper aims at investigating the influence of the thermo-com- currently industrially developed for applications in auto-
pression forming process parameters, such as the temperature motive, tertiary, aeronautic or other cutting-edge tech-
and consolidation time, on the mechanical properties of con- nologies [2, 4]. Particularly, flax is a locally abundantly
tinuous flax fibre-reinforced PA6 composites. Woven fabric available and renewable resource in Europe. Flax fibre has
flax/PA6 composites were prepared by compression moulding better mechanical properties [2] and thermal resistance [5,
using a film-stacking process under different consolidation 6] than other natural fibres which makes it an ideal can-
conditions according to a two-factor Doehlert design of didate, among vegetal fibres, for the polymer composites
experiments. Tensile and flexural properties were assessed. A reinforcement.
second-order polynome was used to correlate the processing Most of the studies concerning vegetal fibre-reinforced
parameters and the mechanical properties (i.e. stiffness, polymer composites have been focused on thermosetting
strength and strain at break). Fracture mode changes with the matrices due, on the one hand to the low viscosity of the
consolidation conditions. The weakening of the fibres and uncured resin which results in good fabric impregnation,
eventually of the composite performances was ascribed to the and on the other hand to the low curing temperature which
increase in flax fibre thermal degradation. is beneficial to prevent fibre thermal degradation [7–13].
Nevertheless, thermoplastic matrices offer an attractive
alternative due to their high ductility, short processing
Introduction
cycle time even for complex parts and their recycling
ability. For instance, the curing process of flax/epoxy
The development of tailored high-performance natural
composite has a 2–8 h duration [7–9, 14] while the con-
fibres makes it possible to use plant-based fibre-reinforced
solidation of flax/poly(lactic acid) composite takes only
composites in structural or semi-structural applications
7 min [15]. Among those thermoplastic polymers,
polypropylene and polyamide are widely used in particular
& Shaoxiong Liang in the automotive industry. However, these polymers
Shaoxiong.liang@mines-douai.fr require a processing temperature over 170 °C, whereas the
Hedi Nouri mechanical properties of plant-based fibres are prone to be
hedi.nouri@mines-douai.fr affected above this temperature due to the presence of
Eric Lafranche thermally sensitive constituents such as cellulose
eric.lafranche@mines-douai.fr (60–70 wt%), hemicelluloses (10–20 wt%) and lignin
1 (2–10 wt%) [16]. Since vegetal fibres contain about
Department of Polymers and Composites Technology &
Mechanical Engineering, Mines Douai, 941 rue Charles 10 wt% water, even drying at high temperature can cause
Bourseul CS 10838, 59508 Douai, France irreversible change of the fibre structure (closure of small

123
J Mater Sci (2015) 50:7660–7672 7661

pores in the cellulosic structure) and decrease the fibre degradation and the mechanical properties of woven fabric
plasticity to make it more brittle [17]. This structural flax fibre/PA6 composite moulded by film stacking method.
change, associated with the cellulosic depolymerization
when exposed to high temperature, leads to mechanical
performance loss. Baley et al. [18] reported that drying at Experimental
elevated temperature leads to a significant decrease in flax
fibre tensile strength from 1500 to 870 MPa. Gourier et al. Materials
[19] studied the effect of thermal processing cycle of 8 min
at different temperatures on the flax fibre mechanical Flax woven fabric (3 9 1 twill) of 600 g m-2 (dry) was
properties, showing that the heating cycles affect the provided by Dehondt TechnologiesÒ (France) under the
mechanical behaviour by changing the shape of the stress– trade name NATTEX N/2D600 (Fig. 1).
strain curves and tensile properties. The mechanical prop- PA6 (AkulonÒ grade K222-D) was purchased from
erties are stable up to a processing temperature of 190 °C Royal DSM Company (The Netherlands). The PA6 used
and significantly drop (-30 to -66 %) at 250 °C. Besides, had a melting temperature of 220 °C and density of
Wielage et al. [20] measured a decrease of 17 and 41 % in 1130 kg m-3. The main mechanical properties of the
flax fibre tensile strength after heating cycles at 180 and polymer and flax fabrics are summarized in Table 1.
220 °C, respectively, for a period of 15 min. Thus, the
thermal degradation of the cellulosic fibres in natural fibre/ Composite materials processing
thermoplastic composites manufacturing is a main concern.
Thanks to their relatively low melting temperature Flax/PA6 composite plates of 270 9 340 9 2 mm3 were
associated with their good economical/performance bal- manufactured by compression moulding using the film
ance, natural fibre/polypropylene composites have been
widely studied especially in the form of chopped fibres
[21–29]. Regarding natural fibre/polyamide 6 (PA6) sys-
tems, a lack of studies is noticed probably due to the high
melting temperature (*220 °C) of this matrix. Ozen et al.
[30] investigated 20 wt% short flax fibre-reinforced PA6,
showing an increase in mechanical properties sufficient to
compete with the conventional glass fibre-reinforced
composites. In the case of chopped cellulose fibre/PA6 and
PA6.6 composites, Xu [31] reported a significant decrease
in strength due to the thermal degradation of the fibres. El-
Sabbagh et al. [6] prepared flax/PA6 composite by injec-
tion moulding. Tensile modulus of 50 wt% flax/PA6
composite increases up to 6 GPa compared to neat PA6
(*1.3 GPa). It is worth mentioning that the very limited
number of studies concerning flax/PA6 composites
addressed extrusion compounding followed by injection
moulding with chopped fibres. This processing method
involves short manufacturing times (‘‘compounding ? in-
jection’’ cycle) of about 8 min [19] and hence good
preservation of the fibre. However, better mechanical per-
formances can be achieved when using continuous fibres to
reinforce the composite materials. For example, a flax mat/
PLA composite fabricated by film stacking achieves a
tensile strength of 89 MPa while that of injection-moulded
short flax fibre/PLA is 46 MPa only [15].
To our better knowledge, there is nevertheless no com-
prehensive study reported in the literature detailing the
influence of processing parameters on thermal degradation
and resulting mechanical behaviour of continuous flax fibre/
PA6 composites. Thus, this paper aims investigating the effect
of process parameters (time and temperature) on the fibre Fig. 1 Nattex N/2D600 flax fibre woven fabric: a top side; b rear side

123
7662 J Mater Sci (2015) 50:7660–7672

stacking process. 0.15-mm-thick PA6 films were first Table 2 Compression moulding conditions
obtained by cast-extrusion process at a temperature of Processing parameter Set-up conditions
250 °C using a Collin single-screw extruder (Germany),
then stored at 23 °C and 50 % relative humidity (RH). Pre-heating timea Fixed 3 min
Before manufacturing, the PA6 films were dried under Pre-heating pressure Fixed 0.44 MPa
vacuum at 0.2 bar and 80 °C for 4 h. According to Santo Compression/impregnation temperature Variable 230–250 °C
et al. [33], fibre drying is not necessary for flax/PA6 pro- Compression/impregnation time Variable 2–8 min
cessing. In the present study, flax fabrics stored at 23 °C Compression pressure Fixed 2.17 MPa
and 50 % RH conditions were not dried to minimize fibre Cooling temperature Fixed 20 °C
damage due to the loss of moisture [18]. Dried PA6 films Cooling time Fixed 10 min
and two layers of flax fabric were then placed alternately in a
Pre-heating temperature is equal to that in the phase of compres-
an aluminium mould following a stacking sequence PA6/ sion/impregnation
Flax/PA6/Flax/PA6. The compression moulding was
achieved in a Dolouet press (France) according to the
processing conditions summarized in Table 2.
A second-ordered two-factor Doehlert design of exper-
iments (DOE) [34] was used to establish the relationship
between processing parameters and mechanical properties.
According to the literature, the temperature and time are
the main parameters controlling the fibre/matrix impreg-
nation quality and the fibre thermal degradation. Thus,
these parameters were chosen as processing variables in
this study. This DOE consists in distributing uniformly the
7 experimental points within the experimental domain
(Fig. 2). These points are placed in equal distance (coding
value) to each other.
This DOE uses commonly coded value (normalized
value) in order to compare the effect of the parameters
having different nature. The relation between coded value
ðXÞ and physical value ðxÞ is obtained through Eq. 1:
x  Center
X¼ ; ð1Þ
Step Fig. 2 Two-factor Doehlert DOE. The point number 8 is an extra
condition
where Step ¼ ðxmax  xmin Þ=2 and Center ¼ ðxmax þ
xmin Þ=2, where xmax and xmin are the ultimate values of
parameter x. This method allows to establish a second- Y ¼ a0 þ a1 X1 þ a2 X2 þ a11 X12 þ a22 X22 þ a12 X1 X2 ; ð2Þ
ordered polynomial relation (Eq. 2) between the response
ðYÞ (mechanical properties), and the experimental vari- where an are the regression coefficients to be determined,
ables (compression moulding temperature ðX1 Þ and com- and a0 is the average experimental response. Graphically,
pression time ðX2 Þ) Eq. 2 is a three-dimensional spherical surface.
According to the two-factor Doehlert DOE, the pro-
cessing conditions of different experimental points (N°1 to
Table 1 Mechanical properties of the dried and conditioned (23 °C, N°7) are presented in Table 3. The specimens are desig-
50 % RH) materials
nated according to their processing conditions (e.g.
PA 6 Flax fibre [32] 230d5mn corresponds to sample cut from a flax/PA6
Dry Conditioned composite plate moulded at 230 °C for 5 min).

Tensile modulus (GPa) 3.8 1.2 54 Mechanical testing


Yield tress (MPa) 95 55 –
Yield strain (%) 3.5 25 – Rectangular samples of 250 9 25 9 2 mm3 were cut for
Stress at break (MPa) 1339 tensile testing according to ISO 527-4 standard. No end-
Flexural modulus (GPa) 2.6 – taps were used. Tensile properties were measured using a
Flexural strength (MPa) 100 – 100 kN tensile machine (Instron 8501, UK) with a

123
J Mater Sci (2015) 50:7660–7672 7663

Table 3 Composite materials processing variables


Experiment number Composite reference Processing temperature Compression time
Physical value x1 (°C) Coded value X1 Physical value x2 (min) Coded value X2

1 230d5mn 230 -1 5 0
2 240d5mn 240 0 5 0
3 250d5mn 250 1 5 0
4 235d2mn 235 -0.5 2 -0.866
5 235d8mn 235 -0.5 8 0.866
6 245d2mn 245 0.5 2 -0.866
7 245d8mn 245 0.5 8 0.866
8a 230d2mn 230 -1 2 -0.866
a
Extra condition

crosshead speed of 2 mmmin-1 at 23 °C and 50 % RH. TGA/DSC1 device (USA) in a 10 mlmin-1 constant
Five specimens were tested for each processing condition. nitrogen flow. The heating rate was 10 °C min-1 over a
The average value and the standard deviations (SD) were temperature range from 25 to 700 °C. Isothermal gravi-
reported. For Young’s modulus measurement, a digital metric analysis was conducted for flax fibre and for PA6 at
image correlation (DIC) method was used in order to three constant temperature levels (230, 240 and 250 °C).
improve the precision of the strain measurement (cf The weight loss was measured during 12 min.
‘‘Young’s modulus’’ section). An Allied camera (Germany)
of 4 M pixel resolution associated with the Vic-2D 2009 Rheological characterization
software was used for the strain measurement.
Three-point bending tests were performed using an Capillary flow measurements were carried out on a capil-
Instron 1185 machine (UK) equipped with a 1 kN force lary rheometer (Model 75 Rheograph, Göttfert, Germany)
cell. The sample dimension was 40 9 15 9 2 mm3 and the at 230, 240 and 250 °C in the range of 10 to 104 s-1 with a
span/depth ratio 16/1 (32 mm in span), according to the 1-mm-diameter die. The length/diameter (L/D) ratios of the
ISO 14125 standard. Ten specimens were tested per con- capillary were 20, 40 and 60. The apparent viscosity
dition. The average values and the standard deviation were obtained was corrected for both the non-Newtonian beha-
reported. viour (Rabinowitsch correction) and the elastic effect
All (tensile and flexural) specimens were cut using a (Bagley correction). The curves were interpolated accord-
Charly-4U milling robot (France) with a 2-mm steel drill. ing to the Carreau Yasuda law (Fig. 3).
Samples were firstly conditioned under vacuum of 0.2 bar
and 50 °C for 72 h, then at 23 ± 3 °C and 50 ± 5 % RH
for 360 h before mechanical testing. Results and discussion

Microscopy observation Impregnation and porosity

A Leica DM 2500 M microscope and a Leica DFC420 Before mechanical analysis, optical microscopy observa-
camera (USA) were used for the observation of the com- tions were carried out in order to ensure the good fibre
posite impregnation quality. A Philips SEM 525 M scan- impregnation and an acceptable porosity level in the
ning electron microscope (SEM) (The Netherlands) was composites. Figure 4 shows a representative optical
used to observe fracture surfaces of flax/PA6 composites micrograph of the composites. The flax woven fabric used
sputtered with a thin layer of gold. was composed by a majority of technical fibres (fibre
bundles) rather than by elementary separated fibres (fila-
Thermal gravimetric analysis (TGA) ments). A very good fibre impregnation quality, even in
fibre rich area (Fig. 4a), is observed. Limited air pocket
Thermal gravimetric analysis (TGA) was performed to porosities are visible (Fig. 4b), representing less than 1 %
evaluate the flax fibre and PA6 thermal degradation. The of the area of a total of 75 microscope images. It is well
measurements were carried out using a Mettler Toledo known that the presence of porosity has a negative effect

123
7664 J Mater Sci (2015) 50:7660–7672

on the mechanical performance of composite materials. influence on the stiffness of the composite, which is small
Using the correction factor ð1  Vp Þ2 proposed by Madsen and can be neglected compared to the differences observed
et al. [35] for plant-based fibre composites, with Vp the between samples consolidated under different processing
porosity volume fraction, 1 % void fraction can have 2 % conditions (‘‘Mechanical behaviour’’ section). Therefore,
the influence of the porosity is ignored in the following
discussion.

Young’s modulus

The main difficulty to evaluate the elastic modulus of the


studied composite materials comes from the fabric struc-
ture itself (size and linear weight of the fibre strand)
inducing local inhomogeneity represented by the distribu-
tion of plastic zones (matrix accumulations) and elastic
zones (continuous fibre strands). Thus, a 12.5-mm strain
extensometer (Ext.) placed on the sample edge or strain
gauges (KFG-10-120-C1-11L1M2R unidirectional strain
gauge) attached on the composite surface (GC) can exhibit
a discrepancy in strains measurement (Fig. 5b).
Fig. 3 Capillary rheological behaviour of PA6 Figure 5a illustrates a typical longitudinal strain field
obtained by DIC. Because of the large size of the flax
strands (7–8 mm width) and the 3/1 twill woven architec-
ture of the fabric (Fig. 1), the strain field is particularly
heterogeneous. In order to increase the accuracy of the
strain measurement, longitudinal strains were calculated by
DIC. Three virtual strain gauges of 3 9 10 mm2 corre-
sponding to the above-mentioned unidirectional strain
gauge size were defined and placed at the centre (GC) and
at locations where the maximum (GMax) and minimum
(GMin) strains were measured. Also, a global strain (Glb)
was calculated corresponding to the average values of a
20 9 90 mm2 area located in the middle of the sample.
The tensile stress–strain curves recorded from different
measurement methods and locations were compared in
Fig. 5b. The strain measured by the gauge located at the
centre (GC) shows higher values than the global strain

Fig. 5 Tensile strain measurement with different methods: a longitu-


Fig. 4 Optical micrographs of the fibre/matrix impregnation quality: dinal tensile strain field, by DIC; b tensile stress–strain curves using
a impregnation in fibre rich area, b air pocket porosity different methods

123
J Mater Sci (2015) 50:7660–7672 7665

(Glb) due to a high strain zone nearby. The extensometer plotted in Fig. 7. The tensile stress–strain curves show a
records lower strain (Ext.) than the global strain (Glb) due knee point at about 0.1 % of strain. This change of slope is
to the influence of a low strain area. This discrepancy in commonly observed for continuous flax fibre-reinforced
strain measurement can highly affect the mechanical composites [7] and attributed to the flax fibre tensile
properties yielded from these curves in particular when the behaviour [18, 32].
Young’s modulus is considered. Due to the complex combination of normal and shear
Finally, the influence of gauge size on the strain mea- stresses in the sample and the smaller size of the flexural
surement by DIC was studied by varying the gauge size specimens compared to that of the tensile ones, the flexural
from 5 9 5 to 20 9 90 mm2. The gauge shape was also curves show larger scattering.
changed from square to rectangular for size bigger than For both tensile and flexural tests, a significant gap in
20 9 20 mm2 due to the specimen width limitation. Strains terms of strength and strain at break can be observed
have been measured in several random locations on the (Fig. 7) between the two sets of composites having the
surface of the same specimen according to different gauges highest and the lowest performances, respectively, refer-
sizes. The strain at break is plotted as function of the gauge enced as 230d2mn and 250d5mm. Whereas the elastic
size in Fig. 6. The scattering of the strain measurement moduli of both composites, calculated as the linear
decreases when the gauge size increases. A convergence regression slope of the stress–strain curves in the strain
towards the average value is noted from gauge sizes of range of 0.05–0.25 %, are very similar.
20 9 80 mm2. Logically, smaller gauge sizes provide more The tensile and flexural properties of the neat PA6 and
precise local information of the specimen behaviour. flax/PA6 composites consolidated in the experimental
However, the present paper focuses on the global behaviour domain are reported in Table 4. Lower values in tensile
of the flax/PA6 composite. Therefore, all strain data will be properties are noticed for neat PA6 compared to those of
calculated hereafter using the 20 9 90 mm2 gauge size. fully dried PA6 (Table 1) due to the drying conditions

Mechanical behaviour

The tensile and flexural stress–strain curves of composites


having the highest and lowest mechanical performance are

Fig. 6 Convergence of the strain measurement as function of the Fig. 7 a Tensile and b flexural stress–strain curves of 230d2mn and
gauge size 250d5mn composite samples

123
7666 J Mater Sci (2015) 50:7660–7672

used in this study (cf ‘‘Composite materials processing’’ Analysis of the DOE response surfaces
section).
The Young’s moduli of the composites are above The coefficients of the polynomial representation (Eq. 2),
10 GPa and the flexural moduli are higher than 8 GPa, identified for the DOE response surfaces of the mechanical
indicating a significant reinforcement effect brought by the properties of the flax/PA6 composites are reported in
flax fabric. Besides, their variations in the experimental Table 5. The statistical analysis confirms the significance
domain, from 10.4 to 12.7 GPa (22 %) for Young’s mod- of all coefficients, with the exception of a12 for tensile
ulus and from 8 to 9.1 GPa (14 %) for the flexural modulus modulus and a11 for flexural modulus, which are non-sig-
indicate a minor influence of the time/temperature param- nificant for a significance level (p value) of 5 %. The sig-
eters on these properties within the experimental domain. nificant coefficients of the non-zero second-ordered terms
The strain at break of 0.29–0.92 % (i.e. 217 % differ- (a11 and a22 ) indicate that the influence of time and tem-
ence) in tension and 1.01–2.63 % (i.e. 160 % difference) in perature on the mechanical properties is non-linear. The
bending shows a particular sensitivity on these two pro- interaction term (a12 ) reveals that the effects of compres-
cessing parameters [15, 33, 36]. Indeed, the temperature sion time and heating temperature on the mechanical
and time directly affect the mechanical properties of the properties are dependent on each other.
vegetal fibre reinforced composites. In general, lower The response surfaces as well as the experimental results
temperature and shorter processing time lead to higher (spherical points) are plotted in Figs. 8 and 9. All response
composite performance [15, 36]. The tensile strength of surfaces, with the exception of the flexural moduli, indicate
230d2mn composites shows a significant improvement better performances with a decrease in the moulding tem-
compared to that of neat PA6 (?30 %), whereas it is 53 % perature and compression time. Hence, the extra experi-
lower than neat PA6 when flax/PA6 was consolidated at mental sample 230d2mn has the highest strength and strain
250 °C, 5 min (250d5mn). In the same way, in bending, under both loading types (Table 4). The predicted results
the samples moulded with the lower temperature (230 °C) correlate well with the experimental ones (Table 4). The
and shorter consolidation time (2 min) display the highest experimental tensile modulus of 230d2mn sample is not
flexural strength (?60 % compared to neat PA6). higher than that of 230d5mn (longer processing time) or
Although the high viscosity of the PA6 matrix (Fig. 3) 235d2mn (higher temperature) ones. This can be explained
should induce an increase of either processing temperature by the fact that, for this property, the experimental results
or compression time, the preservation of flax fibre from between composites manufactured under different pro-
thermal degradation requires a temperature and time cessing conditions are quite close. In fact, both the response
reduction. The results of the extra experimental point surfaces of tensile and flexural moduli are relatively flat,
(230d2mn), which is consolidated at the lowest tempera- and they have small coefficients (a1 ; a2 ; a11 ; a22 and a12 )
ture and shortest time within the experimental domain, with respect to the average value a0 (Table 5). According to
confirm thus a significant increase in composite properties. the literature [36], the modulus is less influenced by the

Table 4 Mechanical properties (average value ± standard deviation) of the flax/PA6 composites, PA6 and corresponding predictions
Material Tensile properties Flexural properties
Modulus (GPa) Strength (MPa) Max strain (%) Modulus (GPa) Strength (MPa) Strain at max stress (%)

230d5mn 12.7 ± 1.2 68.7 ± 3.7 0.64 ± 0.08 9.13 ± 1.53 133 ± 22 2.41 ± 0.31
240d5mn 11.5 ± 0.9 42.2 ± 2.6 0.39 ± 0.03 8.40 ± 1.28 89 ± 13 1.37 ± 0.27
250d5mn 11.0 ± 0.5 32.2 ± 2.4 0.29 ± 0.03 8.00 ± 1.59 66 ± 13 1.01 ± 0.14
235d2mn 12.2 ± 0.3 76.7 ± 2.1 0.78 ± 0.08 8.37 ± 1.44 125 ± 20 2.13 ± 0.35
235d8mn 10.8 ± 0.5 40.2 ± 5.5 0.40 ± 0.07 9.68 ± 1.27 90 ± 18 1.22 ± 0.26
245d2mn 11.6 ± 0.4 54.4 ± 2.9 0.54 ± 0.04 8.55 ± 1.05 102 ± 13 1.65 ± 0.29
245d8mn 10.4 ± 1.5 33.9 ± 4.5 0.36 ± 0.02 8.81 ± 1.02 75 ± 8 1.16 ± 0.20
230d2mn 12.0 ± 1.0 90.2 ± 3.6 0.92 ± 0.02 8.34 ± 0.98 140 ± 16 2.63 ± 0.20
PA6 2.91 ± 0.2 68.4 ± 2.6 19.2 ± 0.5 2.10 ± 0.03 88 ± 1 6.79 ± 0.07
P230d2mna 13.0 96.7 0.99 8.54 154 2.92
a
Response surfaces of the DOE for 230d2mn sample

123
J Mater Sci (2015) 50:7660–7672 7667

Table 5 Identified coefficients of the polynomial response for different mechanical properties
Coefficient Tensile properties Flexural properties
Modulus (GPa) Strength (MPa) Max. strain (%) Modulus (GPa) Strength (MPa) Strain at max. stress (%)

a0 11.51 42.22 0.388 8.40 89.35 1.365


a1 -0.74 -16.91 -0.165 -0.49 -28.62 -0.551
a2 -0.76 -16.46 -0.162 0.45 -18.18 -0.404
a
a11 0.36 8.22 0.080 0.17 10.81 0.351
a22 -0.50 9.38 0.151 0.55 8.32 0.117
a12 0.16a 9.23 0.121 -0.61 4.23 0.247
a
Non-significant value (p value \ 5 %)

processing parameters than the strength and strain due to its Thermal analysis of flax fibre and PA6
lower sensitivity to the adhesion quality between fibres and
matrix. TGA of flax fibre and PA6 matrix is reported in Fig. 11.
From these response surfaces, it is clear that, mechanical According to the literature [23, 37–39], flax fibre is char-
properties decrease when moulding temperature and com- acterized by four weight loss stages. The first one corre-
pression time increase. Both parameters favour the thermal sponds to the water release (dehydration) corresponding to
degradation of the natural flax fibre [19, 20]. the moisture weight in the material, which is approximately
The best mechanical properties of flax/PA6 composite 5 % in our case. The second stage characterized by a
tensile strength (90.2 MPa) and flexural strength weight loss starting at around 200 °C with a plateau
(140 MPa) are significantly higher than those of neat PA6 (shoulder) at 260 °C, corresponds to the hemicelluloses
(68.4 MPa in tension and 88 MPa in bending). It is worth degradation. The major weight loss starts at 260 °C with a
noting that, this composite material was consolidated under maximum peak value at 350 °C and the end at 385 °C.
230 °C/2 min conditions; the total processing time of This stage is linked to the cellulose pyrolysis. Afterwards,
5 min (3 min for preheating and 2 min for consolidation) is the weight continues to decrease in a stable way until the
long compared to that of injection-moulded small plastic end of the scanning temperature which is 700 °C related to
parts (several seconds to tens of seconds) but in the same the non-cellulosic substances [37] (such as lignin and sugar
range (several minutes) as that of injection-moulded big Maillard derivatives).
automotive components. Nevertheless, the processing cycle The TGA leads to the conclusion that the lignin is more
is shorter than that of plant-based fibre/thermoplastic thermally stable than hemicelluloses and cellulose with a
composites (7–22 min [2, 15, 19]) generally found in the decomposition temperature range of 150–900 °C [40, 41].
literature. Using thermoplastic polymers compared to Besides, Yang et al. [40] have demonstrated that the ther-
thermosetting resins as composite matrix makes it possible mal degradation of these components depends on their
to reduce processing time (several hours [9, 11, 18]) and inherent structures and chemical nature. The hemicellulose
warrants recyclability of the material. is constituted of various random and amorphous polysac-
charides which can be easily removed at low temperature,
Fracture surfaces while the cellulose is constituted of highly ordered glucose,
a polymer without branches giving high thermal stability.
Figure 10 shows the SEM images of the tensile fractured The lignin formed of aromatic rings associated with vari-
surfaces of composites processed under different condi- ous branches shows a wide degradation temperature range
tions of the Doehlert DOE. The processing conditions have [40]. The residual weight of 20 % currently observed at the
a significant influence on the fracture behaviour of the end of the TGA test (700 °C) is consistent with the values
composites. With the increase of temperature (i.e. Fig. 10a reported in the literature [23].
to b, c to e, f to g) as well as the compression time (i.e. The TGA and DTGA curves of PA6 show a first mass
Fig. 10f to a, g to b), the length of the fibre pull out loss (2.1 %) corresponding to the presence of water in the
decreases. In the case of the most severe conditions matrix [42]. The beginning of decomposition is noticed at
(250 °C, 5 min), the flax fibres were broken just near the 370 °C and reaches a maximum rate at 455 °C much
transverse fracture plane (Fig. 10 e) indicating a resistance higher than those of flax fibre. Moreover, isothermal TGA
of the flax fibre lower than that of the fibre/matrix interface. curves (Fig. 12) of the PA6 indicate a minor weight loss of

123
7668 J Mater Sci (2015) 50:7660–7672

Fig. 8 Response surfaces of


tensile properties of flax/PA6
composites as a function of
compression time and heating
temperature: a Young’s
modulus, b tensile strength,
c maximum tensile strain. Each
point represents an experimental
result

about 0.2 % during the first 10 min at 250 °C. The thermal The isothermal gravimetric analysis (Fig. 12) of flax
degradation of the matrix does not seem to happen under fibre shows a non-linear mass loss, the thermal degrada-
the present composite processing conditions (230–250 °C tion increasing as a function of the processing time. The
for several min). mass loss caused by the fibre dehydration is not taken into

123
J Mater Sci (2015) 50:7660–7672 7669

Fig. 9 Response surfaces of


flexural properties of flax/PA6
composites as a function of
compression time and heating
temperature: a flexural modulus,
b flexural strength, c flexural
strain at maximum stress. Each
point represents an experimental
result

account here due to the preheating period of the apparatus 250 °C, indicating a decrease in the kinetic of
(3 min) where the water contained in the flax fibre is degradation.
vaporized. According to Fig. 12, the flax fibre mass loss is Regarding the experimental domain, the composite
higher between 230 and 240 °C than between 240 and plates were moulded between 230 and 250 °C for 2 to

123
7670 J Mater Sci (2015) 50:7660–7672

Fig. 10 SEM images of tensile


fractured surfaces for different
processing conditions of the
Doehlert DOE

Fig. 11 Thermal gravimetric analysis of the PA6 and flax fibre


Fig. 12 Isothermal gravimetric analysis of the PA6 and flax fibre

8 min. Thermal degradation should have occurred in flax Conclusion


fibres. In the elementary flax fibre, the cellulose forms
crystalline mesofibrils, structured in spirals within amor- The thermo-compression moulding of flax fibre fabric-re-
phous matrix made of hemicelluloses while the lignin acts inforced PA6 was experimentally investigated in order to
as an adhesive between them [32, 43]. The decomposition minimize the thermal degradation of the flax fibres during
of lignin and hemicelluloses destroys the fibrils network processing. The influence of the processing temperature
and hence results in the drop in the mechanical properties and time on the tensile and flexural mechanical behaviour
of the reinforcement fibre and consequently the mechanical was evaluated using a second-ordered two-factor Doelhert
properties of the composites. DOE.

123
J Mater Sci (2015) 50:7660–7672 7671

The DOE response surfaces of the mechanical properties 7. Liang S, Gning PB, Guillaumat L (2014) Quasi-static behaviour
of flax/PA6 composites have shown a decrease in perfor- and damage assessment of flax/epoxy composites. Mater Des
2015(67):344–353. doi:10.1016/j.matdes.11.048
mance when the consolidation temperature and time 8. Liang S, Gning PB, Guillaumat L (2012) A comparative study of
increase. The increase of the temperature from 230 to fatigue behaviour of flax/epoxy and glass/epoxy composites.
250 °C and of the time from 2 to 5 min results in a drop of Compos Sci Technol 72(5):535–543. doi:10.1016/j.compscitech.
tensile strength of the flax/PA6 composite from 90.2 to 2012.01.011
9. Gning PB, Liang S, Guillaumat L, Pui WJ (2011) Influence of
32.2 MPa (-64 %). The dependence of the mechanical process and test parameters on the mechanical properties of flax/
properties on the processing parameters is non-linear. epoxy composites using response surface methodology. J Mater
The mechanical strengths of flax/PA6 composite con- Sci 46:6801–6811. doi:10.1007/s10853-011-5639-9
solidated at 230 °C, 2 min are significantly higher than 10. Nguyen VH, Deleglise-Lagardere M, Park CH (2015) Modeling
of resin flow in natural fiber reinforcement for liquid composite
those of the neat PA6 (?32 % in tension and ?60 % in molding processes. Compos Sci Technol 113:38–45. doi:10.1016/
bending). The total processing time of 5 min (3 min for j.compscitech.2015.03.016
preheating and 2 min for consolidation) is shorter than 11. Li Y, Li Q, Ma H (2015) The voids formation mechanisms and
those usually used to prepare vegetal fibre/thermoplastic their effects on the mechanical properties of flax fiber reinforced
epoxy composites. Compos A Appl Sci Manuf 72:40–48. doi:10.
composite materials reported in the literature (7–22 min). 1016/j.compositesa.2015.01.029
The fracture mode of the composites (length of fibre 12. Zhang D, Milanovic NR, Zhang Y, Su F, Miao M (2013) Effects
pull-out) depends on the severity of the processing condi- of humidity conditions at fabrication on the interfacial shear
tions. The TGA has highlighted four decomposition stages strength of flax/unsaturated polyester composites. Compos B Eng
2014(60):186–192. doi:10.1016/j.compositesb.12.031
of the flax fibre: dehydration (25–150 °C), decomposition 13. Le Duigou A, Kervoelen A, Le Grand A, Nardin M, Baley C
of hemicelluloses (200–260 °C), cellulose (260–385 °C) (2014) Interfacial properties of flax fibre–epoxy resin systems:
and lignin (from 170 °C). Also, the time/temperature existence of a complex interphase. Compos Sci Technol
dependence in the flax fibre degradation process was con- 100:152–157. doi:10.1016/j.compscitech.2014.06.009
14. Nguyen VH, Lagardere M, Park CH, Panier S (2014) Perme-
firmed. A maximum loss of 3.5 % in mass is measured for ability of natural fibre reinforcement for liquid composite mold-
flax fibres exposed at 250 °C for 11 min. ing processes. J Mater Sci 49(18):6449–6458. doi:10.1007/
s10853-014-8374-1
Acknowledgements Thanks are due to Dehondt TechnologiesÒ 15. Le Duigou A, Davies P, Baley C (2009) Seawater ageing of flax/
(France) for kindly supplying flax fibre fabric (Nattex N/2D600) and poly(lactic acid) biocomposites. Polym Degrad Stab
to the Nord-Pas-de-Calais Region (France) for funding Hedi Nouri’s 94:1151–1162. doi:10.1016/j.polymdegradstab.2009.03.025
post-doctoral grant (Contract No. 14002212). The authors also 16. Gassan J, Bledzki AK (2001) Thermal degradation of flax and
acknowledge International Campus on Safety and Intermodality in jute fibers. J Appl Polym Sci 82:1417–1422
Transportation (CISIT), European Community (FEDER) and Nord- 17. Bos JH, Müssig J, van den Oever MJA (2006) Mechanical
Pas-de-Calais Region for funding the TGA equipment. properties of short-flax fibre reinforced compounds. Compos A
Appl Sci Manuf 37:1591–1604. doi:10.1016/j.compositesa.2005.
10.011
18. Baley C, Le Duigou A, Bourmaud A, Davies P (2012) Influence
of drying on the mechanical behaviour of flax fibres and their
References unidirectional composites. Compos A Appl Sci Manuf
43:1226–1233. doi:10.1016/j.compositesa.2012.03.005
1. NabiSaheb D, Jog JP (1999) Natural fibre polymer composites: a 19. Gourier C, Le Duigou A, Bourmaud A, Baley C (2014)
review. Adv Polym Technol 18(4):351–363. doi:10.1002/ Mechanical analysis of elementary flax fibre tensile properties
(SICI)1098-2329(199924)18:4\351:AID-ADV6[3.3.CO;2-O after different thermal cycles. Compos A Appl Sci Manuf
2. Wambua P, Ivens J, Verpoest I (2003) Natural fibres: can they 64:159–166. doi:10.1016/j.compositesa.2014.05.006
replace glass in fibre reinforced plastics? Compos Sci Technol 20. Wielage B, Lampke T, Marx G, Nestler N, Starke D (1999)
63:1259–1264. doi:10.1016/S0266-3538(03)00096-4 Thermogravimetric and differential scanning calorimetric analy-
3. Bos HL (2004) The potential of flax fibres as reinforcement for sis of natural fibres and polypropylene. ThermochimicaActa
composite materials. PhD thesis, Technische Universiteit Eind- 337:169–177. doi:10.1016/S0040-6031(99)00161-6
hoven, Eindhoven 21. Sobczak L, Lang RW, Haider A (2012) Polypropylene compos-
4. Müssig J (2010) Industrial applications of natural fibres: struc- ites with natural fibers and wood: general mechanical property
ture, properties and technical applications. Wiley, New York, profiles. Compos Sci Technol 72:550–557. doi:10.1016/j.comps
ISBN: 978-0-470-69508-1 citech.2011.12.013
5. Manfredi LB, Rodriguez ES, Wladyka-Przybylak M, Vazquez A 22. Lafranche E, Oliveira V, Martins CI, Krawczak P (2015) Pre-
(2006) Thermal degradation and fire resistance of unsaturated diction of injection-moulded flax fibre reinforced polypropylene
polyester, modified acrylic resins and their composites with nat- tensile properties through a micro-morphology analysis. J Com-
ural fibres. Polym Degrad Stab 91:255–261. doi:10.1016/j.poly pos Mater 49(1):113–128. doi:10.1177/0021998313514875
mdegradstab.2005.05.003 23. Arbelaiz A, Fernandez B, Ramos JA, Mondragon I (2006)
6. El-Sabbagh A, Steuernagel L, Ziegmann G, Meiners D, Toepfer Thermal and crystallization studies of short flax fibre reinforced
O (2014) Processing parameters and characterisation of flax fibre polypropylene matrix composites: effect of treatments. Ther-
reinforced engineering plastic composites with flame retardant mochim Acta 440:111–121. doi:10.1016/j.tca.2005.10.016
fillers. Compos B Eng 62:12–18. doi:10.1016/j.compositesb. 24. Arbelaiz A, Fernandez B, Cantero G, Llano-Ponte R, Valea A,
2014.02.009 Mondragon I (2005) Mechanical properties of flax fibre/

123
7672 J Mater Sci (2015) 50:7660–7672

polypropylene composites. Influence of fibre/matrix modification 34. Guillaumat L (2000) Reliability of composite structures- impact
and glass fibre hybridization. Compos A Appl Sci Manuf loading. Compos Struct 76:163–172. doi:10.1016/S0045-
36:1637–1644. doi:10.1016/j.composittesa.2005.03.021 7949(99)00166-2
25. Arbelaiz A, Fernandez B, Cantero G, Llano-Ponte R, Valea A, 35. Madsen B, Thygesen A, Lilholt H (2009) Plant fibre composites—
Mondragon I (2005) Mechanical properties of short flax fibre porosity and stiffness. Compos Sci Technol 2009(69):1057–1069.
bundle/polypropylene composites: influence of matrix/fibre doi:10.1016/j.compscitech.01.016
modification, fibre content, water uptake and recycling. Compos 36. Song YS, Lee JT, Ji DS, Kim MW, Lee SH, Youn JY (2012)
Sci Technol 65:1582–1592. doi:10.1016/jcompscitech.2005.01. Viscoelastic and thermal behavior of woven hemp fiber rein-
008 forced poly(lactic acid) composites. Compos B Eng 43:856–860.
26. Ku H, Wang H, Pattarachaiyakoop N, Trada M (2011) A review doi:10.1016/j.compositesb.2011.10.021
on tensile properties of natural fiber reinforced polymer com- 37. Bourmaud A, Morvan C, Baley C (2010) Importance of fibre
posites. Compos B Eng 42:856–873. doi:10.1016/j.compositesb. preparation to optimize the surface and mechanical properties of
2011.01.010 unitary flax fiber. Ind Crops Prod 32:662–667. doi:10.1016/j.
27. Garkhail S, Wieland B, George J, Soykeabkaew N, Peijs T (2009) indcrop.2010.08.002
Transcrystallization in PP/flax composites and its effect on 38. Van De Velde K, Kiekens P (2002) Thermal degradation of flax:
interfacial and mechanical properties. J Mater Sci 44(2):510–519. the determination of kinetic parameters with thermogravimetric
doi:10.1007/s10853-008-3089-9 analysis. J Appl Polym Sci 83(12):2634–2643. doi:10.1002/app.
28. Facca AG, Kortschot MT, Yan N (2006) Predicting the elastic 10229
modulus of natural fibre reinforced thermoplastics. Compos A 39. Qua EH, Hornsby PR, Sharma HSS, Lyons G (2011) Preparation
Appl Sci Manuf 37:1660–1671. doi:10.1016/j.compositesa.2005. and characterisation of cellulose nanofibres. J Mater Sci
10.006 46(18):6029–6045. doi:10.1007/s10853-011-5565-x
29. Facca AG, Kortschot MT, Yan N (2007) Predicting the tensile 40. Yang H, Yan R, Chen H, Lee DH, Zheng C (2007) Character-
strength of natural fibre reinforced thermoplastics. Compos Sci istics of hemicellulose, cellulose and lignin pyrolysis. Fuel
Technol 67:2454–2466. doi:10.1016/jcompscitech.2006.12.018 86:1781–1788. doi:10.1016/j.fuel.2006.12.013
30. Ozen E, Kiziltas A, Kiziltas EE, Gardner DJ (2013) Natural fiber 41. Ciobanu C, Ungureanu M, Ignat L, Ungureanu D, Popa VI (2004)
blend-nylon 6 composites. Polym Compos 34(4):544–553. Properties of lignin–polyurethane films prepared by casting
doi:10.1002/pc.22463 method. Ind Crops Prod 20:231–241. doi:10.1016/j.indcrop.2004.
31. Xu X (2008) Cellulose fiber reinforced nylon 6 or nylon 66 04.024
composites. PhD thesis, Georgia Institute of Technology 42. Araujo JR, Adamo CB, De Paoli MA (2011) Conductive com-
32. Baley C (2002) Analysis of the flax fibres tensile behaviour and posites of polyamide-6 with polyaniline coated vegetal fiber.
analysis of the tensile stiffness increase. Compos A Appl Sci Chem Eng J 174:425–431. doi:10.1016/j.cej.2011.08.050
Manuf 33:939–948. doi:10.1016/S1359-835X(02)00040-4 43. Gorshkova TA, Gurjanov OP, Mikshina PV, Ibragimova NN,
33. Santos PA, Spinace MAS, Fermoselli KKG, De Paoli MA (2007) Mokshina NE, Salnikov VV, Ageeva MV, Amenitskii SI, Cher-
Polyamide-6/vegetal fiber composite prepared by extrusion and nova TE, Chemikosova SB (2010) Specific type of secondary cell
injection molding. Compos A Appl Sci Manuf 38:2404–2411. wall formed by plant fibers. Russ J Plant Physiol 57(3):328–341.
doi:10.1016/j.compositesa.2007.08.011 doi:10.1134/S1021443710030040

123

You might also like