nd

2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

INTERFACIAL CHARACTERISATION OF FLAX FIBRE - THERMOPLASTIC

POLYMER COMPOSITES BY THE PULL-OUT TEST

A. Stamboulis1, C. Baillie1, E. Schulz2

1
Imperial College of Science, Technology and Medicine, Department of Materials, Prince Consort
Road, London SW7 2BP, UK
2
Federal Institute of Materials Research and Testing, Lab VI.21, Unter den Eichen 87, D-12200 Berlin,
FRG

ABSTRACT
The interface between flax fibres and thermoplastic polymer matrices has been investigated.
Two types of flax fibres have been used: Dew retted and upgraded Duralin fibres. The latter fibres
have been treated by a novel treatment process for improved moisture and rot sensitivity. The
apparent shear strength of Dew-retted and upgraded Duralin fibres embedded in high and low density
polyethylene, polypropylene and maleic anhydride polypropylene, respectively has been calculated
from the maximum Force measured from the pull-out test using the Kelly-Tyson Equation. Higher
apparent shear strength values have been measured in the case of high density polyethylene. An
improvement in the interfacial shear strength by the use of upgraded flax fibres has not been clearly
observed. The curves of pull-out force versus displacement for all samples are typical of a brittle
fracture mixed mode interface behaviour. The pull-out test of flax fibre-polymer matrices exhibit similar
characteristics with the pull-out test of synthetic fibre-polymer systems.

1. INTRODUCTION

The interface between fibre and polymer matrix has been always an important area of
investigation since the early days of composites. The most forward way at first sight, to investigate
interface itself is by using single fibre tests. Tests with fibre bundles produce results, that are affected
by matrix effects. Pull-out, fragmentation, microcompression and microtension tests are currently used
for single fibre specimens(1-10). They all aim to characterise basic failure behaviour of the interface
and to measure a micromechanical value, the apparent shear strength.
In short-fibre composites, reinforcement is assumed to take place through elastic actions
accross the interface. The failure process in these composites involves fracture failure along the fibre
surfaces, with pull-out for short fibres and fracture for longer fibres. This process can be best
simulated by the single fibre pull-out test. The advantage of this test is that the debonding force can be
plotted as a function of embedded length. Knowing the maximum pull-out force, a so called apparent
shear strength τ can be calculated according to the Kelly/Tyson equation (11):

τ= P / DπL (1)

where P is the maximum debonding stress, D the diameter of the fibre and L the embedded length.
The whole procedure involves the assumption, that the interfacial loading is a shear mode, the shear
stress along the fibre is homogeneous and the friction between debonded fibre and matrix is
negligible. This, however, can be true in the case of a ductile matrix, where an extensive
rearrangement of the stresses is occured as a consequence of the plastic yielding of the matrix (12).
In the case of an ideal elastic fibre and matrix properties and a brittle interface fracture, strong stress

10-1
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

field inhomogeneities are appeared at the fibre entry as well as at the fibre end. Their quantities
depend strongly on the embedded length of the fibre and influence both the crack initiation and
propagation at stress loading (7,13,14). Concequently, the apparent shear stress calculated using
Equation (1) decreases with increasing fibre length in the case of an elastic material behaviour,
whereas it is independent from the fibre embedded length in the case of a very ductile matrix.
Nevertheless the pull-out test is very useful to determine the predominant material behaviour during
the failure process (15). Depending on the E-modulus proportion of fibre and matrix as well as the
embedding length in single fibre pull-out specimens, the debonding process mainly starts at the fibre
entry point and gradually extends along the fibre until an arrest occurs while the end debonds(16-18).
The fracture process is a mixed mode, which complicates the determination of the true shear strength.
The natural fibres are considered as composite materials themselves. The single fibres
(elongated cells) consist of a primary wall and three other secondary walls. At the middle there is a
lumen. The primary cell wall consists of pectin, which can be stiffened and dehydrated by the addition
of lignin. It then becomes part of the adhesive system holding cells together. In the secondary cell
wall, the cellulose molecules are synthesised by enzymes, which float around in the cell membrane,
spinning off primary fibrils of about 5 nm in diameter, each containing about 40 molecules. These
primary fibrils are assembled into microfibrils about 20-40 nm in diameter, which have hemicelluloses
decorating the outside. These hemicelluloses act as the connection between the microfibrils, giving
the primary structural network. Lignin can be deposited within this network with two ways: either as
isolated lumps (when it acts to limit the movement of the microfibres, thus increasing stiffness by steric
hindrance) or (later) as a continuous matrix, which then supplements and presumably replaces the
hemicellulose in importance to link the cellulose microfibrils. The fibres are highly anisotropic
especially when the diameter is considered. The lack of homogeneity in these fibres makes the
interpretation of the pull-out test difficult.
This work aims to investigate the fracture behaviour of single flax fibre thermoplastic
composites using the pull-out test and attempts to determine the interfacial characteristics in these
systems using this micromechanical model method. It is suggested that the pull-out test of these fibres
embedded in a polymer matrix has similarities with other fibres such as glass or carbon fibres and that
these systems can be evaluated taking into consideration besides the experimental difficulties with
natural fibres, similar problems involved in the micromechanical analysis of conventional synthetic
fibre systems.

2. EXPERIMENTAL

The flax fibres used in this work have been supplied by CERES BV, The Netherlands and
were of two types: Dew-retted flax fibres and upgraded flax fibres so called Duralin. The new
upgraded Duralin flax fibres have been treated by a novel treatment process for improved moisture
and rot sensitivity (19). Environmental studies have showed that these upgraded Duralin flax fibres
absorb less moisture than untreated flax fibres, whereas the mechanical properties of the upgraded
fibres are retained, if not improved (20). Individual dew-retted and Duralin flax fibres were carefully
separated by using small and thin tweezers. Single fibre pull-out specimens were prepared using a
specially designed apparatus developped at the BAM Institute. The single fibre specimens were
prepared on little aluminium blocks, which can be heated via an electric furnace at a controlled
temperature. A small droplet of polymers [high density polyethylene (HDPE), low density polyethylene
(LDPE), polypropylene (PP) and maleic-anhydride (5%) modified polypropylene (MA-PP)) was melted
on the aluminium sample holder. Then a single fibre of Duralin or Dew-retted flax was vertically placed
within the polymer droplet at a certain embedded length. The pull-out test was performed with the help
of a stiff test machine. The sample holder was mounted on the pull-out test device and the free fibre
end was glued as shown in Figure 1 (21).

10-2
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

Figure 1: Schematic diagram of the pull-out test device (7)

3. RESULTS AND DISCUSSION

Figure 2 shows a SEM micrograph of dew-retted fibres. The single fibres are well bonded
together to form larger fibre bundles. Kinks or nodes can be clearly seen at regular intervals along the
fibre surface, which may be attributed to fibre cell tip overlap within the bundles or defects on the
fibres surface due compression. Figure 3 shows single separated and bonded upgraded Duralin
fibres. The surface of some single fibres is smooth and kink bands are visible along the fibre axis. On
the left side of the micrograph the bonded fibres are surrounded by a layer of matrix which mainly
consists of pectin and some lignin. The fibres are well connected by the organic matrix and form a
fibre bundle. In both micrographs a roughness in surface is observed due to the organic matrix which
surrounds the primary cell wall of the fibres, which possibly gives reason to mechanical interlocking
and friction forces during the pull-out test.
Figures 4 and 5 show the relation between the apparent shear strength and embedded length
in the cases of Dew-retted /HDPE and –LDPE, respectively. In the case of Dew-retted/HDPE is
obvious that there is a non constant function of τ vs.L. This has been observed for all the single fibre
specimens with HDPE, PP and MA-PP and can be interpreted as a dominant brittle interface fracture
behaviour. In the case of the single flax (Dew-retted and Duralin) fibres-LDPE (Figure 5) the relation
between the shear strength and the embedded length is more constant indicating a less brittle matrix.

Figure 2: SEM micrograph of dew-retted fibres. Figure 3: SEM micrograph of upgraded Duralin
fibres

10-3
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

Dew-retted/HDPE

Apparent Shear Strength

20

[MPa]
15

10

0
0 100 200 300
Embedded length [µm]

Figure 4: Apparent Shear strength versus embedded length for Dew-retted flax-HDPE.

Dew-retted/LDPE

8
Apparent Shear
Strength [MPa]

3
100 150 200 250 300 350 400
Embedded Length [µm]

Figure 5: Apparent Shear Strength versus embedded length for Dew-retted flax-LDPE.

SEM micrographs (Figure 6) of a polypropylene matrix after a dew-retted fibre has been pulled-
out showed that micrifibrils separated from the single fibre remained within the matrix indicating weaker
interfaces within the fibre.

10-4
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

Figure 6: SEM micrograph of polypropylene matrix after a single dew-retted fibre has been pulled –out.

Table 1 summarises the results of the pull-out tests for both Dew-retted and Duralin fibres with
different polymer matrices. There are two values of shear strength, one is calculated with the help of
Equation 1 and then extrapolation of the trendline apparent shear strength versus embedded length to
L=0 and the other one is taken from the graphs of apparent shear strength versus embedded length at a
length of 150 µm.

Table 1: Shear strength resulting from single fibre pull-out tests.

Samples Apparent Shear strength (MPa)

Eq.1 and extrapolated to L=0 Embedded length of 150 µm
Dew-retted/HDPE 18.0 9.1
Dew-retted/LDPE 5.6 5.5
Dew-retted/PP 17.3 10.6
Dew-retted/MA-PP 17.8 11.4
Duralin-HDPE 16.2 10.1
Duralin-LDPE 7.1 6.2

It is obvious that higher values for both fibres are observed when HDPE was used as a polymer
matrix compared to LDPE. HDPE having a more rigid structure (linear chain) allows crystallisation
effects (transcrystallinity) to influence the interfacial shear strength. The maleic anhydride modified
polypropylene was tested as well as it is believed that improves the adhesion between these fibres and
the polypropylene matrix [22]. However, there is no significant improvement in the shear strength in the
case of Dew-retted/PP and Dew-retted/MA-PP. The shear strength of the latter is comparable with the
shear strength of Dew-retted-HDPE. Comparing Dew-retted and Duralin fibres there is no observed
improvement of the interfacial strength in the case of HDPE. On the other hand a very small
improvement is observed in the case of LDPE. However, this seems to be not very clear. Comparing
these values with the values at an embedded length of 150 µm a small improvement of the shear
strength in the case of Duralin fibres can be observed. The values of shear strength at an embedded
length of 150 µm are lower as it is expected. It is suggested that at small embedded lengths the values
of shear strength are scattered, whereas at higher embedded lengths the shear strength values become
less scattered. Therefore an intermediate embedded length has to be chosen in order to evaluate the
shear strength of the samples. From the Table 1 it is obvious that the values of shear strength at an
embedded length of 150 µm follow the same trend compared with the above mentioned values with the
exception of Dew-retted/HDPE.
Figure 7 shows a typical pull-out test of Dew-retted/HDPE. It is worth to notice that the shape of
the curve is typical of a mixed mode fracture process observed also by Hampe and Marotzke for glass
fibre/Polystyrol (23). All the samples showed similar behaviour during the pull-out test. The debonding
(point A) starts at the fibre entry point and it is extended gradually along the fibre. At the point B the end

10-5
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

debonding has already occured whereas at the point C there is still fibre length bonded and because of
the high fibre Poisson‘s shrinkages the friction forces are very small. It is worth to notice that the
behaviour of flax fibres embedded in a polymer matrix in the pull-out test is similar with the behaviour of
synthetic fibres although there are some anisotropic parameters such as the diameter of the fibres
which make the pull-out test more complicated.

Figure 7: Typical pull-out test curve of flax fibre-polymer systems (Dew-retted/HDPE).

4. CONCLUSIONS

The interfacial shear strength of Dew-retted and Duralin fibres embedded in different
thermoplastic matrices has been measured with single fibre pull-out test. An improvement in the
interfacial shear strength by the use of upgraded flax fibres has not been clearly observed. Higher
interfacial shear strength has been observed in the case of HDPE matrix that may be also attributed to
transcrystalline effects. The curves of pull-out force versus displacement for all samples are typical of a
brittle fracture mixed mode interface behaviour. The pull-out test of flax fibre-polymer matrices exhibit
similar characteristics with the pull-out test of synthetic fibre-polymer systems.
REFERENCES

1. M.R. Piggott, Comp.Sci.Techn., 57 (1997) 965-974.

2. S.R. Dai, M.R. Piggott, Comp.Sci.Technol.., 49 (1993) 81-87.
3. A.T. Di Benedetto, Comp.Sci.Technol.,42 (1991) 103-124.
4. A.T. Dibenedetto, M.R. Gurvich, J.Mat.Sci., 33 (1998) 4239-4248.
5. M. Mak, A. Lowe, B. Jar, Z. Stachurski, J.Mat.Sci.Let., 17 (1998) 645-647.
6. E.J.H. Chen, J.C. Young, Compos.Sci.Technol., 42 (1991)189-206.
7. E.Schulz, G. Kalinka, W. Auersch, J. Macromol.Sci.-Phys., B35(3&4), 527-546 (1996).
8. F.P.Liu, M. P. Wolcott, D.J. Gardner, T.G. Rials, Comp.Interf., Vol.2, No 6 (1994) 419-432.
9. P. Hedenberg, P. Gatenholm, J.Appl.Pol.Sci., 60 (1996) 2377-2385.
10. M. Desaeger, I. Verpoest, Compos.Sci.Technol., 48 (1993) 215-226.
11. A. Kelly, W.R. Tyson, J.Mech.Phys.Solids, 12 (1965) 329.
12. A. Hampe, C. Marotzke, ‚‚Messung der Scherfestigkeit in der Grenzfläche zwischen Polymer und
Glas durch Einzelfaserexperimente‘‘; in „ Haftung bei Verbundwerkstoffen und

10-6
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

Werkstoffverbunden“,ed. W. Brockmann, DGM Infomationsgesellschaft .Verlag, Oberursel 1989, p.

33-44.
13. C. Marotzke, Compos.Interf. 1 (1993) 153.
14. C. Marotzke, Compos.Sci. Technol., 50 (1994) 393.
15. S. Meretz, W. Auersch, C. Marotzke, E. Schulz, A. Hampe, Compos.Sci.Techol., 48 (1993) 285-
290.
16. M.R. Piggott, YX. Xiong, Compos.Sci.Technol., 52 (1994) 535-540.
17. A. Hampe, C. Marotzke, Compos.Sci.Technol., 57 (1997) 845-851
18. C. Marotzke, L.Qiao, Compos.Sci.Technol.57 (1997)887-897.
19. Pott, G.T., Pilot, R.J. and van Hazendonk, J.M., in: Proc. of the 5th European Conference on
Advanced Materials and Processes and Applications (EUROMAT 97), Vol.2 Polymers and
Ceramics, Maastricht, 21-23 April 1997,107.
20. A. Stamboulis, C.A. Baillie, S.K. Garkhail, H.G.H. van Melick, T. Peijs, Appl.Comp.Mat., 1999, in
press.
21. A. Hampe, G. Kalinka, S. Meretz, E. Schulz, Comp. 26 (1995) 40
22. A.K. Bledzki, J. Gassan, Natural Fiber Reinforced Plastics, in Handbook of Engineering Polymeric
Materials, Ed. N.P. Cheremisinoff, 1998, pp.787-810.
23. Personal communication with Prof. A. Hampe, BAM Institute.

10-7
 nd
2 International Wood and Natural Fibre Composites Symposium June 28-29, 1999 in Kassel/Germany

10-8