matrix adhesion of cellulose fibres in PLA, PP and MAPP: A critical review of pull-out test, microbond test and single fibre fragmentation test results

Accepted Manuscript Fibre/matrix adhesion of cellulose fibres in PLA, PP and MAPP: A critical review of pull-out test, microbond test and single fibre fragmentation test results Nina Graupner, Joraine Rö ßler, Gerhard Ziegmann, Jörg Müssig PII: DOI: Reference:

S1359-835X(14)00113-4 http://dx.doi.org/10.1016/j.compositesa.2014.04.011 JCOMA 3602

To appear in:

Composites: Part A

Received Date: Revised Date: Accepted Date:

10 June 2013 11 February 2014 17 April 2014

Please cite this article as: Graupner, N., Rö ßler, J., Ziegmann, G., Müssig, J., Fibre/matrix adhesion of cellulose fibres in PLA, PP and MAPP: A critical review of pull-out test, microbond test and single fibre fragmentation test results, Composites: Part A (2014), doi: http://dx.doi.org/10.1016/j.compositesa.2014.04.011

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Fibre/matrix adhesion of cellulose fibres in PLA, PP and MAPP: A critical review of pull-out test, microbond test and single fibre fragmentation test results Nina Graupner1*, Joraine Rößler1, Gerhard Ziegmann2, Jörg Müssig1 1

University of Applied Sciences Bremen, Dept. Biomimetics – The Biological Materials Group, Neustadtswall

30, D-28199 Bremen, Germany 2

Institute for Polymer Materials and Plastics Processing, Clausthal University of Technology, Agricolastr. 6, D-

38678 Clausthal-Zellerfeld, Germany

* corresponding author: [email protected]

Abstract The present work deals with the practical fibre/matrix adhesion of regenerated cellulose fibres (lyocell) and bast fibre bundles (flax, kenaf) in different matrices (polylactide-PLA, polypropylene-PP, maleic-anhydride-grafted polypropylene-MAPP). The influence of different testing procedures (pullout-test, microbond-test, fragmentation-test) on the fibre/matrix characteristics is discussed. The results of the different tests showed the same trends, but the absolute values differ. Clearly higher interfacial shear strength (IFSS) for cellulose fibres was found in PLA and MAPP in comparison to PP due to higher polarity. In addition, bast fibres displayed higher apparent IFSS values compared to lyocell because of their rougher surface and their chemical composition. The apparent IFSS of the pull-out-test resulted in higher values compared to results obtained from the fragmentation-test. This phenomenon is explained by different stress distributions due to variable specimen geometry, different behaviour of failure and the friction which occurs between fibre and matrix during fibre pull-out in the pull-out-test.

Keywords: A. Fibres; A. Thermoplastic resin; B. Fibre/matrix bond; B. Fragmentation

1

Introduction Besides the properties of the reinforcing fibre and the polymer matrix, the fibre/matrix interaction has a decisive influence on the properties of a composite [1] [2]. The chemical composition of the fibre and the constitution of the fibre surface play an important role [3] [4] [5]. Failure of a composite is dependent on different failure mechanisms. If the force exceeds the strength of the interface, debonding and fibre pull-out occur. Via frictional forces, tension may still be transferred to a fibre along the interface, whereby energy is absorbed [6]. This kind of fibre/matrix adhesion is called practical fibre/matrix adhesion or apparent interfacial shear strength, respecively. If the forces are higher than the local fibre strength, the fibre fails due to tension. A distinction is drawn between the theoretical and the practical fiber/matrix adhesion. In contrast to the practical adhesion the theoretical adhesion depends only indirectly on the work of fracture (fibre pull-out). The theoretical adhesion is difficult to measure because it is influenced by factors such as thermal stress, stress gradients, plastic deformation, bridging molecule chains, etc., which have little to do with the theoretical adhesion [7]. Due to this fact the measured fibre/matrix adhesion in this work is described as practical adhesion. In general a good adhesion between a stiff and strong fibre and the matrix results in a stiff composite with high tensile strength, whereby often brittle impact behaviour is caused, because energy absorbing fibre pull-outs are prevented. However, weaker but still good fibre/matrix interactions usually lead to good impact strength due to the good energy absorption by the pull-outs. Depending on the reinforcing fibre used, the strength of a composite can also be reduced by weak fibre/matrix adhesion. A weak fibre/matrix interaction, in which force transfer from the matrix to the fibre is not possible, lead to a reduction in the tensile properties, and may also result in a decrease in the impact strength due to the absence of friction during the fibre pull-outs. Hence, for the interpretation of the mechanical properties of a composite, the characterisation of fibre/matrix adhesion is of special importance. For the analysis of conventional composites, such as glass or carbon fibre reinforced composites, some tests are already available [4] [8] [9]. Examples are the pull-out test, the microbond test, the push-in test, the three-fibre test or the fragmentation test. Brief descriptions of these tests with the advantages and disadvantages can be found in [4] [10]. Clyne

2

and Jones [11] showed test procedures for different material combinations and list a variety of testing methods, which are used by different authors. In the present work, the fragmentation test [12] [13], the microbond test [14] and the pull-out test [15] [16] are used. For the fragmentation test, the fibre is completely embedded in the matrix and the sample is loaded axially to the fibre direction in a tensile test up to a certain strain. During this procedure, the samples are not to be allowed to break. Depending on the quality of the adhesion, more or fewer fragments are formed. From the lengths of these fragments, the critical fragment length, and an assessment of the fibre tensile strength the interfacial shear strength (IFSS) can be calculated. For the pull-out and microbond tests, the fibre is embedded only partially, to a certain length, into the matrix, and is then pulled out axially from the matrix. From the measured forces, the IFSS, and if the fibre tensile strength is known, the critical fibre length can be calculated. The described methods are sometimes used for the characterisation of the fibre/matrix interactions of cellulose fibre-reinforced polymers. The preparation for the measurements is often labour intensive, especially for very fine and/or brittle fibres which often result in fibre breakage. The fragmentation test is well-suited for brittle fibres in a matrix, which has at least a three times higher elongation than the fibre [2], like flax fibres in a PP matrix [17] [18] [19] [20]. For material combinations such as lyocell fibres in a PLA matrix, this test is not applicable due to the high elongation of the fibres compared to the matrix. The matrix would break before fragmentation can start. Another test which is used for cellulose fibres in polymeric matrices is the pull-out test, e.g. for flax in PE or PP [21] as well as in PLA or PHB [22], sisal in PP or PLA [23] [24] or coir in PET [25]. The microbond test differs from the pull-out test substantially in the shape of the test specimen. It can be used for the investigation of ramie and lyocell in PP [3] or pineapple leaf fibre bundles in PHBV [5]. Preparation, in particular of fine and brittle fibres with low tensile strength, is difficult and limited to fibres having a minimum diameter and a minimum strength as described for the pull-out test [1]. The pull-out and microbond test can be primarily used for fibres with high tensile strength and/or weak fibre/matrix interactions [4]. The embedding length of the fibre is limited; otherwise, the fibre would break before it could be pulled out of the matrix. Due to the possible formation of a meniscus of the matrix to the

3

fibre during sample preparation, the exact embedding length cannot always be determined correctly, and there is the possibility that values are distorted [2]. Due to the high deviations of the IFSS reported in the literature for similar material pairings, a comparative study was carried out of results obtained with the single fibre fragmentation test, the pullout test and the microbond test. In this framework, the following hypotheses were formulated and are subjected to critical scrutiny:

•

Due to the influence of different test procedures, the results (IFSS and critical fibre or fragment length), which are determined with the pull-out test and the fragmentation test do not match.

•

Bast fibre bundles, like flax or kenaf, have a higher surface roughness and lead to higher apparent IFSS values compared to lyocell fibres with a smooth surface.

•

The results of the pull-out and microbond test will not differ significantly due to the similar testing procedure.

•

IFSS and the critical fibre or fragment length of different bast fibre bundles (flax and kenaf) will not differ significantly due to the similar mechanical properties.

•

The large variations of the IFSS values reported in the literature for similar material pairings are dependent on the preparation and testing parameters. If constant production and testing parameters are used, a higher reproducibility of the results is possible.

Materials & methods Fibres •

Lyocell fibres with a fibre fineness of 1.5 tex, a staple fibre length of 90 mm, a fibre diameter of 35.7 µm and a density of 1.5 g/cm³ were supplied by Lenzing AG (Lenzing, Austria) in 2009,

•

well dew retted and scutched kenaf fibre bundles (Hibiscus cannabinus L.) from Bangladesh for textile processing with a manually measured cut length of 30 – 40 mm and a fibre width of

4

46,8 ± 24,9 µm (determined by Fibreshape [29]) were supplied by Holstein Flachs (Mielsdorf, Germany) in 2004 and •

field retted flax fibre bundles (Linum usitatissimum L.; HO-0401b) with a manually measured cut length of 150 – 230 mm and a fibre width of 38,7 ± 32,3 µm (determined with Fibreshape) were supplied by Holstein Flachs (Mielsdorf, Germany) in 2004.

For the evaluation of the amount of individual fibre cells within the flax and kenaf fibre bundles the fibre bundles of the supplied samples were prepared in a hot shrink tube (Conrad Electronic SE, Hirschau, Germany) and embedded in polyethylene glycol (type 1000, Rotipuran® Ph. Eur., Carl Roth GmbH + Co KG, Karlsruhe, Germany). Cross-sections were prepared with a microtome and 40 fibre bundles of flax and kenaf were examined with a light microscope. The fibre cross-sectional area, the fibre width and fibre thickness were measured with the open source software ImageJ (Wayne Rasband, National Institutes of Health, USA) and the number of cells was counted in each fiber bundle. The results of the measurement have shown a dependence of the cross-sectional area of the fibre bundle on the amount of individual fibre cells with a coefficient of determination of 0.68 for flax and 0.66 for kenaf. The dependence of the fibre width on the number of fibre cells shows similar coefficients of determination with 0.68 for flax and 0.70 for kenaf. Since the fibre width is measured for the evaluation of the values determined with the pull-out test and the single fibre fragmentation test, from the linear equation of the regression line, the amount of individual fibres within the flax and kenaf fibre bundles was calculated. As an example for flax a fibre bundle width of 100 µm corresponds an amount of 19 individual fibres. A number of 21 individual fibres was calculated for kenaf. Matrices •

PLA Ingeo fibres type SLN 2660 D (Eastern Textile Ltd., Taipei, Taiwan) with a fibre fineness of 0.67 tex and a length of 64 mm produced from a NatureWorks™ PLA with a density of 1.24 g/cm³, a melting temperature of 160 - 170 °C and a glass transition temperature of 60 – 65 °C,

•

polypropylene granules (PP) (SABIC PP 575P) homopolymer with a density of 0.905 g/cm³, a melting temperature of 173 °C and a melt flow rate (MFR230; 2,16) of 10.5 g/10 min and

5

•

polypropylene granules (SABIC PP 755P) grafted with 2% maleic anhydride (Licomont AR504, acid number = 41 mg KOH/g, density = 0.91 g/cm³ and a melting temperature of 156 °C) as coupling agent (MAPP).

Fibre tensile test The lyocell fibres and the bast fibre bundles were conditioned for at least 18 h at 20 °C and 65% relative humidity prior tensile testing. A Fafegraph M testing machine (Textechno, Mönchengladbach, Germany) working with a load cell of 100 cN for lyocell fibres and 10000 cN for bast fibre bundles was used for the tensile tests. The machine was equipped with pneumatic clamps covered with hard PVC. According to a cotton standard (ISO 3060) the gauge length was set to 3.2 mm at a testing speed of 2 mm/min. A clamp with a mass of 200 mg was used for pre-tensioning the fibres. Due to the big standard deviation of the tensile characteristics of cellulose fibres at least 70 fibres were tested.

Single fibre pull-out test For the single fibre pull-out test test specimens were prepared from single lyocell fibres, flax fibre bundles and kenaf fibre bundles in a PLA, PP and MAPP matrix. The kenaf fibre bundles used for the pull-out test with a PLA matrix have an amount of 3 to 24 (median value 11) single fibres in the bundle with a cross-sectional area ranging between 0.0006 and 0.0036 mm² (median value 0.0013). Flax fibre bundles used, contain 9 to 51 individual fibre cells with a surface area between 0.0008 and 0.0128 mm². Test specimens were prepared with a PLA, MAPP and PP matrix. The median value of the amount of individual flax fibre cells within the bundle was determined with 25, 23 and 23, respectively. The cross-sectional area was determined with 0.0035, 0.0042 and 0.0039 mm² for the preparation with the PLA, MAPP and PP matrix, respectively. For the preparation of the pull-out test specimens polymer-films with a thickness of 100 – 250 µm were produced. The polymer-films were manufactured in a Joos hot press type HP-S10 (Joos, Pfalzengrafenweiler, Germany) at 180 °C for 3 min under a pressure of approx. 20 N/cm² between two Teflon-foils. Foils were cooled in a cold press (Joos, Pfalzengrafenweiler, Germany, type HP S-60) at 25°C for 2 min at a pressure of 20 N/cm². The polymer-film was cut into sheets of approx. 8 x 8 mm²

6

which were equipped with a little gap made by a sewing needle. A Teflon foil type 0903 AS (Böhme Kunststofftechnik GmbH + Co. KG, Schwarzenbek, Germany) with dimensions of 15 x 15 mm² was furnished with a little gap, too. The fibre was first pushed through the hole of the Teflon foil and was fixed on the one side of an aluminium spacer (thickness 2 mm). The other end of the fibre was pushed through the polymer-film and was fixed in a metal frame to adjust the fibre in the vertical direction as shown in Figure 1 A. In order to prevent a deformation of the molten samples by air current, the metal frame was wrapped with aluminum foil. After this, the samples were stored in a forced air oven type BW91270 (Zwick/Roell GmbH, Ulm, Germany) for 5 min at 185 °C. The polymer melted and surrounded the fibre. After cooling at room temperature and demoulding the specimens were conditioned for at least 18 h at 23 °C and 50% relative humidity. The IFSS was measured by pulling out the fibre of the matrix by using a Zwick/Roell universal testing machine type Z020 (Zwick/Roell GmbH, Ulm, Germany). To fix the specimen in the testing machine, a rectangular frame was fabricated. The frame had a slot with a width of approx. 700 µm and was fixed with the upper clamps of the testing machine. The test specimen was inserted into the metal frame and the fibre was pushed through the slot and was fixed with the lower clamps as shown in Figure 1B. The testing machine operated with a load cell of 500 N, the free gauge length was set to 5 mm and the testing speed was 1 mm/min. The amount of tested samples is shown in Table 2. For the evaluation only test specimens with a complete fibre pull-out were analysed. This was controlled with an optical light microscope after testing. Valid results were used to calculate the apparent or apparent IFSS (τ) in N/mm² according to Kelly and Tyson (1965) [30] (Equation 1)

τ=

Fmax d f ⋅ π ⋅ l eF

(1)

with Fmax = the maximum load in N, df = the diameter of the fibre or fibre bundle in mm and leF = embedding length of the fibre in the matrix in mm. The critical fibre length Lc in mm was calculated and according to Kelly Tyson [30] (Equation 2)

Lc =

σ f ⋅df 2 ⋅τ

(2)

with σf = tensile strength of the fibre in N/mm².

7

The calculation according to Kelly Tyson [30] assumes a uniform stress distribution, a uniform fibre diameter and a homogeneous wetting of the fibre with the matrix. If these conditions are given, the debonding force would increase linearly with increasing embedding surface. This behavior could be determined only in trend for the investigated composite systems with low coefficients of determination (~ 0.2). The low linear correlation confirms that at least one of the requirements has not been met. Since it is very improbable that the conditions are met for any investigated sample, the calculation of the IFSS and the critical fibre length represents an approximation which can be expressed as the apparent IFSS and the apparent critical fibre length, respectively.

Microbond test For the preparation of test specimens for the microbond test single lyocell fibres were fixed in a metal frame at both ends with adhesive tape (Tesa SE, Hamburg, Germany). Two PLA fibres having a fineness of 0.67 tex were knotted around the fibre. The protruded ends of the PLA-fibres were removed with a razor blade (see Figure 2 A). Thereafter, the samples were heated for 5 min at 185 °C and cooled at room temperature as described for the pull-out test. Result is a PLA-droplet around the fibre with a diameter ranging between 100 and 200 µm. The testing procedure and the evaluation of the results of 18 microbond samples was carried out with the parameters described for the pull-out test. The fixation of the microbond test specimens is shown in Figure 2 B.

Single fibre fragmentation test (SFFT) For the investigation of the practical fibre/matrix adhesion of cellulose fibres in PP- and MAPPmatrices, a single fibre fragmentation test (SFFT) was additionally chosen. As fibres single lyocell fibres as well as flax and kenaf fibre bundles were used. The median value of the amount of single fibre cells within the bundle ranged between 6 and 8. In contrast to the pull-out test smaller fibre bundles were selected because the crack can run through single fiber cells and not through the entire fibre bundle when larger fibre bundles are used. If the crack does not run through the entire fibre

8

bundle, no meaningful analysis of the fragment length can be done. The probability for this behavior increases with a raising amount of fibre cells within the bundle. As part of sample preparation, PP- and MAPP-sheets with a thickness between 180 and 250 µm were produced as described for the pull-out test section. Single fibres and fibre bundles were oriented, placed on one sheet and fixed with adhesive tape on both ends. The sheet was covered with a second sheet and pressed again to stack the films together (hot pressing under a pressure of 20 N/cm² at 180 °C for 5 min followed by cold pressing at 25 °C for 2 min at a pressure of 20 N/cm²). To determine the critical fragment length and IFSS, the sheets were cut into dump bell shaped specimens with a width of approx. 3 mm and a length of 36 mm. Testing was carried out in a Zwick/Roell universal testing machine (Zwick/Roell, Ulm, Germany) type Z 020 operating with a load cell of 500 N and a testing speed of 0.2 mm/min. The gauge length was set to 15.5 mm. Before the specimen is broken the test was finished. Prior to testing the test specimens were conditioned for at least 18 h at 23 °C and 50% relative humidity. The fragment length of the individual samples was measured under an optical light microscope Axiostar Plus (Zeiss, Jena, Germany) using an object micrometer. For the analysis of the fragment length only samples were taken into account where the crack went completely through the fibre (compare Figure 3). The amount of tested samples is shown in Table 2. The measured results allow the calculation of the critical fragment length Lfc in mm according to Feih et al. [12] (Equation 3)

L fc =

4 − 4 n li ⋅l = ∑ 3 3 i =0 n

(3)

with Lfc = critical fragment length in mm, Ī = average fragment length, n = number of measurements and li = single fragment length in mm and the IFSS τ in N/mm² according to Kelly and Tyson [30]. The apparent IFSS was calculated according to Equation 4 [30]

τ=

σ f ⋅df 2 ⋅ L fc

(4)

with σf = tensile strength of the fibre in N/mm² and df = diameter of the fibre or fibre bundle in mm.

9

Similar to the pull-out test the calculation of the critical fragment length and the IFSS assume a uniform stress distribution, a uniform fibre diameter and a homogeneous wetting of the fiber with the matrix.

Scanning electron microscopy (SEM) SEM-investigations of the fibre surface and the interface of pull-out test specimens were carried out with a SEM type JSM 6510 (Jeol, Eching, Germany). Prior to the investigations the samples were prepared on sample holders equipped with conductive carbon adhesive tape and sputtered for 90 s and a voltage of 56 mA with a BAL-TEC SCD-005 sputter coater (Bal-Tec, Liechtenstein) with gold. Micrographs were taken with an accelerating voltage of 5 kV with secondary electrons.

Atomic force microscopy (AFM) AFM investigations were carried out with a NanoWizard AFM (company JPK, Berlin, Germany). Surfaces of flax and lyocell fibres were scanned in a contact mode. The cantilevers (type Arrow, company NanoWorld, Neuchâtel, Switzerland) have a spring constant of 0.2 N/m and a resonance frequency of 14 kHz. The scanning speed was set to 1.5 Hz and the set point was configured to < 1 V.

Statistical evaluation of the test results The statistical evaluation of all results was carried out with the JMP 8.0 software (SAS Institute, Cary, USA) with a Shapiro-Wilk test (a = 0.05) regarding a normal distribution. To proof if there are significant differences between the data of variable samples, for normally distributed data the TukeyKramer HSD test (a = 0.05) and for data which are not distributed normally the Wilcoxon test (a = 0.05) were chosen. Results are shown as Box-Whisker plots. Beside the Box-Whisker plots mean values and standard deviations are presented in the figures. Significant differences are shown with different letters and a “*” shows results which are not distributed normally.

10

Results and discussion Tensile properties of fibres The tensile characteristics of the cellulose fibres and fibre bundles are shown in Table 1. It is apparent that the bast fibre bundles display a clearly higher tensile strength and stiffness in comparison to the lyocell fibres, but the elongation at break of lyocell is clearly higher than that of the bast fibre bundles. The tensile strength results were used for the calculation of the apparent IFSS and the critical fibre length.

Fibre/matrix-adhesion of lyocell, flax and kenaf in different matrices Interfacial shear strength (IFSS) Influence of the testing system: SFFT vs. pull-out test Figure 4 A shows the results of the apparent IFSS of lyocell/PLA, lyocell/MAPP and lyocell/PP measured with the pull-out test. In comparison to lyocell/PP with an apparent IFSS of 6.3 MPa, a significantly higher value was measured for lyocell/PLA at 10.3 MPa. Compared to PP, MAPP leads to a clear increase, up to a value of 8.8 MPa (compare Table 2). Similar values for lyocell/PP were reported by Adusumalli et al. [10] measured with the microbond test. The authors determined an apparent IFSS of 5.3 MPa. The treatment of the fibres with a 0.2% MAPP solution led to an increase in the apparent IFSS up to 8.2 MPa. The higher apparent IFSS is caused by the improved practical adhesion between the cellulose fibres and the matrix. The ester linkage between the hydroxyl groups and the hydride carbonyl groups of the maleic anhydride results in improved adhesion [31] [32]. A literature study by Mechraoui et al. [33] showed an optimum improvement of the practical fibre/matrix adhesion with a MAPP concentration ranging between 1.5 and 2.5 mass-%. The improvement in fibre/matrix adhesion by the presence of MAPP in PP matrices is effected by an improved load transmission from the matrix to the fibre, which leads to an increase in the tensile strength of the composites; this has been shown in other studies, too [34] [35] [36] [37]. Unlike PP, PLA shrinks slightly during processing and has a higher hydrophilicity and polarity [38] than PP, which has a more unipolar and hydrophobic character [39]. Khoshkava and Kamal [41] investigated nano-crystalline cellulose reinforced PLA and PP matrices and pointed out that the work

11

of adhesion depends on the polarity of the polymer matrix. The authors investigated the dispersion, the polar and the total surface energy of PLA, PP and nano-crystalline cellulose at 25 °C. They found out that nano-crystalline cellulose displays the highest polar energy (28 mJ/m²) and total surface energy (69 mJ/m²). The lowest polarity with 0.5 mJ/m² and total surface energy (32 mJ/m²) was measured for PP. PLA displays a value of 8 mJ/m² for the polarity and 41 mJ/m² for the total surface energy. Due to this, PLA exhibit a higher work of adhesion value than PP. In addition, the authors applied the Lifshitz−van der Waals/acid−base approach for the calculation of work of adhesion and stated that this model is more reasonable for PLA because it considers hydrogen bonding interactions between nanocrystalline cellulose and PLA [41]. All the ascribed properties are responsible for a better wetting of the cellulose fibres with the PLA matrix and for a worse wetting of the cellulose fibres with the PP matrix which results in gaps in the interface as shown in Figure 13 C. As measured with the pull-out test, a similar trend was observed for the samples examined with the fragmentation test. A higher apparent IFSS was determined for lyocell/MAPP (5.0 MPa) compared to lyocell/PP (3.8 MPa) (Figure 4 B). As a trend, the values which were determined by means of the pullout test were 1.7 to 1.8 times higher compared to the results of the fragmentation test. The results of the apparent IFSS of flax in PLA, MAPP and PP show a similar trend as those measured for lyocell in different matrices (compare Figure 5). The highest apparent IFSS was determined with the pull-out test with a value of 28.3 MPa for flax/PLA, followed by flax/MAPP with 24.3 MPa and flax/PP with 17.9 MPa (Figure 5 A). The results of the pull-out tests showed 1.5 to 1.8 times higher values than those obtained with the fragmentation test. Flax/MAPP resulted in apparent IFSS of 15.8 MPa and flax/PP of 9.8 MPa, measured with the fragmentation test (Figure 5 B). Similar trends have been described by different authors. Morlin and Czigáni [44] reported an IFSS of 3.4 MPa for flax fibre bundle reinforced PP and 5.1 MPa for hemp fibre bundle reinforced PP measured with the microbond test. In a PLA matrix, a value of 9.0 MPa was determined for flax fibre bundles and of 11.3 MPa for hemp fibre bundles. Taha [24] investigated sisal fibre bundle reinforced PLA and sisal fibre bundle reinforced PP. The higher apparent IFSS (17.1 MPa) was measured for sisal/PLA in comparison to sisal/PP (6.1 MPa). An improvement by the use of MA as an adhesion promoter was also confirmed by literature values. With the use of 5% MA, Stamboulis et al. [21]

12

determined an increase of the apparent IFSS of single flax fibres in a PP-matrix (measured with the pull-out test) from 10.6 MPa to 11.4 MPa. The apparent IFSS of different bast fibre bundles show no significant differences for flax/PLA and kenaf/PLA (Figure 6 A) investigated with the pull-out test as well as for flax/MAPP and kenaf/MAPP (Figure 6 B) measured with the fragmentation test (Table 2).

Influence of fibre geometry Due to the fact that lyocell fibres were prepared as single fibres and bast fibres were prepared as fibre bundles, the results are not directly comparable. As shown in Figure 7 typical pull-out force elongation characteristics of bast fibre bundles and single lyocell fibres show differences. While for bast fibre bundles a decrease in friction after debonding was observed, similar to results ascribed by Yang and Thomason [40] for single glass fibres in a PP matrix with a low IFSS, a relative constant friction was obtained for the single lyocell fibre during pull-out. Bast fibre bundles display a complex structure and consist of some individual fibre cells with a certain adhesion between the single fibres and the multilayered cell wall. Le Duigou et al. [45] described that the failure during a pull-out test can even be initiated in the primary fibre cell. The primary cell wall can be debonded from the secondary fibre cell wall. The authors observed a peeling of the primary cell wall after debonding and argued that this behaviour lead to a decrease of the frictional force during fibre pull-out [45]. Microscopic investigations were used for the identification of specimens that have shown a peeling of individual fibre cells or the cell wall resulting in adherent fibre material on the matrix. The peeling effect leads to a decline of the apparent IFSS. Specimens showing this behaviour were not considered for the evaluation.

Influence of interfacial friction and surface roughness For single lyocell fibres and flax fibre bundles in a PLA-matrix sometimes a sudden release of the stored energy resulted in a complete loss of the measured force by a sudden pull-out of a large portion of the embedded fibre. This behavior is described also by Yang and Thomason [40] for glass fibres in a PP matrix at high pull-out forces.

13

Beside the apparent IFSS the interfacial friction was determined for all pull-out samples, showing friction after debonding according to [45]. The results are summarised in Table 3. It is obvious that the standard deviation is very large due to the high variability of the frictional forces observed. In trend the higher interfacial friction was determined for the bast fibre bundles in comparison to single lyocell fibres. While the interfacial friction of bast fibre bundles corresponds to 51 – 61% of the apparent IFSS, the values of lyocell range between 38 and 71 % for the different matrices. It is hypothesised that the differences are caused by the variable complexity of the fibre structure. As explained above bast fibre bundles consist of some individual fibre cells. Single fibres can be splitted from the bundle as shown in Figure 8 A-C. The splitted single fibres can be embedded separately into the matrix and can bear a part of the applied load. When the main fibre bundle is debonded from the matrix the splitted single fibres could be pulled out of the matrix and lead to additional frictional forces. In dependence on the different matrices and the strength of the adhesion between splitted single fibres, fibrils and the matrix the frictional force can vary. The debonding of splitted single fibres and fibrils can lead to a decrease of the friction force after debonding due to a peeling effect. In contrast to bast fibre bundles lyocell fibres display a smooth surface (compare Figure 8 D). Due to the smooth surface the friction force dropped down to a similar value for the different matrices after debonding (Table 3). As shown in AFM micrographs in Figure 9 a higher average roughness could be confirmed for bast fibre bundles with a mean value of 124 nm in comparison to lyocell with a mean average roughness of 34 nm. As presented in Table 3 the interfacial friction of the bast fibre bundles is very high which confirms a high surface roughness, too. This interfacial friction has nothing to do with the theoretical adhesion. Due to this fact the interfacial friction was subtracted from apparent IFSS resulting in lower values, especially for bast fibre bundles. Nevertheless, the results show the same trend and the values of bast fibre bundles are still higher than those of lyocell (Table 3). The corrected pull-out values display in trend lower values than the results determined with the single fibre fragmentation test. Compared to the pull-out test the friction is of less importance for the evaluation of test results obtained from the fragmentation test because the measured fragment length is used for the calculation of the IFSS and not the measured force.

14

Influence of chemical composition Beside friction, another aspect for the better apparent adhesion of bast fibre bundles may be due to the chemical composition. While bast fibres contain 2 – 19% lignin [43] this component is missing in lyocell fibres. A preliminary study has shown that lignin can lead to an improvement in fibre/matrix adhesion, resulting in higher tensile strength of the composite [46]. The reason for the better adhesion is based on the fact that lignin contains polar, polarisable and nonpolar groups while pure cellulose contains only polar groups. The nonpolar groups of lignin can interact with PLA and PP by van-derWaals forces and the polar groups with cellulose. Unlike PP, PLA might also interact with cellulose by its ester groups. MAPP can interact with cellulose via its anhydride [42].

Critical fibre and fragment length The critical fibre length, which is determined with the pull-out test, as well as the critical fragment length, which was measured with the fragmentation test, showed a consistent trend with the calculated apparent IFSS. Good fibre/matrix adhesion leads to short fibre or fragment lengths, whereas poor fibre/matrix adhesion leads to longer critical fibre or fragment length. Thus, for the pull-out test, the shortest critical fibre length was measured for lyocell/PLA with 0.75 mm, followed by lyocell/MAPP with 0.94 mm and lyocell/PP with 1.31 mm (compare Figure 10 A). Adusumalli et al. [10] described a comparable critical fibre length of 0.7 mm for lyocell/MAPP and 1.2 mm for lyocell/PP (determined with the microbond technique). Analogous to the lower IFSS determined with the fragmentation test, a longer critical fragment length was measured in comparison to the pull-out test. Lyocell/MAPP resulted in a value of 1.46 mm and lyocell/PP of 2.50 mm (Figure 10 B). Similar trends as described for lyocell were found for flax. From the pull-out test, a critical fibre length of 1.95 mm was measured for flax/PLA, 1.87 mm for flax/MAPP and 3.42 mm for flax/PP (Figure 11 A). The critical fragment length determined with the fragmentation test showed values of 2.14 mm for flax/MAPP and 2.98 mm for flax/PP (Figure 11 B). In contrast to the IFSS, significant differences were identified between flax and kenaf with respect to the critical fragment lengths (Figure 12). The critical fragment length of flax/PLA was 1.95 mm and that of kenaf/PLA was determined with 0.73 mm (Figure 12 A). Similar trends for flax/MAPP and

15

kenaf/MAPP are shown in Figure 12 B. For flax/MAPP, a value of 2.14 mm and for kenaf/MAPP, a value of 0.6 mm was determined. These differences are based on the variable cross-sectional area of the fibre bundles. The examined kenaf fibre bundles which were investigated with the pull-out test displayed a median fibre bundle width of 27.7 µm and the fibre bundles examined with the fragmentation test had a fibre bundle width of 35.0 µm. Investigated flax fibre bundles were much coarser (77.4 µm for the pull-out test and 53.3 µm for fragmentation test). While the measured force for the calculation of the IFSS is related to the outer surface of the embedded fibre section, the critical fibre length largely depends on the aspect ratio. This means for the same apparent IFSS, a larger fibre length is required for the coarser fibres.

Fragmentation test versus pull-out test The hypothesis stating that the results from the fragmentation test and pull-out test show similar values could not be confirmed. The apparent IFSS values measured with the pull-out test showed higher values by factors between 1.5 and 1.8 than those which were found with the fragmentation test. Similarly, the measured critical fragment lengths were 1.2 to 1.9 times higher than the critical fibre length. The measured differences of the apparent IFSS and critical fibre lengths by different testing methods have been confirmed by some authors [1] [47] [48] [49]. However, in contrast to the results presented in this study, in most studies, higher apparent IFSS values were determined with the fragmentation test compared to the pull-out test [1] [47] [48] [50]. To evaluate this effect, the following results will be presented regarding the differences of the results measured with the pull-out test and the fragmentation test for different material pairings. From these statements, explanation attempts are derived for the described phenomenon. Herrera-Franco et al. [48] measured for henequen fibre bundles in a HDPE matrix an apparent IFSS value of 5.4 MPa and 2.5 MPa with the fragmentation test and the pull-out test, respectively. ValadezGonzalez et al. [47] reported an apparent IFSS value of 4.2 MPa determined with the fragmentation test and of 2.4 MPa determined with the pull-out test for henequen fibre bundle reinforced HDPE. The authors explained the higher values with better wetting of the fibres with the matrix, due to the pressure which is applied during the preparation of fragmentation test. However, Wagner et al. [50]

16

justified the higher apparent IFSS of single Kevlar 49 fibre reinforced epoxy with the fragmentation test compared to the microbond test with the methods of evaluation used. The authors described that for the evaluation of the results of the fragmentation test, only those samples are considered which were fragmented, while the results of the specimens which were measured with the microbond test or the pull-out test are only valid if the fibre is pulled out of the matrix. Thus, values measured with the fragmentation test tend to be overestimated, whereas the values of the microbond test are underestimated. Since cellulose fibres have a significantly lower tensile strength than glass or Kevlar fibres, fragments can be found reliably in fragmentation test samples. This means that fewer samples in which the fibre has not failed are rejected from the evaluation. However, Piggott and Dai [49] explained a higher apparent IFSS of single Kevlar fibre reinforced PE investigated with the pull-out test compared to the fragmentation test with the fibre/matrix friction during fibre pull-out. It is assumed that this effect is valid for the investigated samples, especially for the bast fibres. Moreover, cellulose fibres have a rougher surface, and cause higher friction compared to glass or Kevlar fibres. A higher surface roughness of bast fibre bundles in comparison to lyocell fibres could be confirmed with SEM and AFM studies (compare Figure 8 and Figure 9). If the determined interfacial friction is subtracted from the apparent IFSS, the values are clearly reduced, especially for the bast fibre bundle reinforced samples which display the highest interfacial friction. The correction of the pull-out test values with the interfacial friction results in trend to higher apparent IFSS values measured with the fragmentation test. This trend is described by different authors for uncorrected apparent IFSS values [1] [42] [43] [45]. In comparison to results from the pull-out test the interfacial friction has less influence on the results determined with the fragmentation test due to the measurement of the critical fragment length. The measured force is not used for the calculation of the apparent IFSS.

Pull-out test versus microbond test To investigate the influence of two similar centrosymmetric test methods on the apparent IFSS and the critical fibre length, experiments were carried out with the pull-out test and microbond test for lyocell/PLA. According to the hypothesis, it was expected that the results of both methods should not

17

differ significantly from each other. In essence, the two methods differ in terms of the fixation in the experimental procedure and the embedding of the fibre into the polymeric matrix. The specimen geometry differed considerably (see Figure 14), resulting in a different stress distributions during testing. The results of this investigation is summarised in Figure 15. Figure 15 A shows the apparent IFSS in dependence of the testing method. The determined apparent IFSS measured with the pull-out test is significantly higher (10.3 ± 4.9 MPa) as compared to the apparent IFSS determined with the microbond test (6.9 ± 3.0 MPa). Hence, a significantly greater critical fibre length was measured for the microbond test specimens with 1.02 ± 0.53 mm. With the pull-out test, a critical fibre length of 0.75 ± 0.51 mm was calculated (Figure 15 B). The differences in the results are based on the different specimen geometry and the resulting different stress states [9]. Yang and Thomason [6] confirmed the higher apparent IFSS for the pull-out test in comparison to the microbond test for glass fibres in a PP matrix. With the microbond test, a value of 2.3 MPa was measured and a value of 3.3 MPa was determined using the pull-out test. The interfacial friction was not affected whether the pull-out test or the microbond test was used. It resulted in 4.4 MPa and 4.1 MPa for the pull-out test and the microbond test, respectively.

Reproducibility of the results Due to the variation of the characteristics within the test series as well as the various results measured with the fragmentation test, the pull-out test and the microbond test, it is necessary to clarify the reproducibility of the results measured with the described methods. It is evident that the scattering of the results within the test series is very large. In the fragmentation test, this effect is mainly caused by few fragments, which were determined with a very small length leading to a very high apparent IFSS and low critical fragment length. These readings mean that statistical analysis often shows no normal distribution of the results. The large scatter of the pull-out test values are the result of the small embedding length (around 100 µm). It is hypothesized that small voids can exist in the interface leading to a considerable reduction in the apparent IFSS and thus to an overestimation of the critical fibre length.

18

Data in a literature review on IFSS and critical fibre or fragment lengths determined with fragmentation, pull-out and microbond tests confirmed the large variability of data within trials as well as by different test parameters and the evaluation of the fibre geometry and embedding length. A summary of apparent IFSS and critical fibre length investigated with the fragmentation, pull-out and microbond tests is given in Table 5 for different fibre/matrix combinations without optimised fibre/matrix adhesion. It becomes clear that the results may vary significantly from study to study. Figure 16 shows the apparent IFSS in dependence on the critical fibre length for different fibre/matrix

compositions without optimized fibre/matrix adhesion. The variation is particularly visible for bast/PP and bast/MAPP. This effect can be based on the different cross-sectional areas, the different amount of single fibres within the fibre bundle and a different degree of retting of the fibres which can influence the surface characteristics (roughness, chemical composition) of the bast fibre bundles as discussed by Le Duigou et al. [63]. However, similar problems have been described for single and homogeneous carbon fibre reinforced epoxy composites. Ji et al. [9] examined, in the context of the carbon fibre Round Robin test, the reproducibility of the fragmentation, pull-out, push-in and microbond test results in 12 different laboratories. The results show considerable scatter. Due to the constant and reproducible fibre diameter of the single carbon fibres, the determination of the fibre cross-sectional area has a smaller influence in comparison to natural fibres / fibre bundles. The test parameters were prescribed as part of the Round Robin test. It can be concluded that the differences of the test results are based on the sample preparation as well as the evaluation of the embedding length. Even for single glass fibre/epoxy, Piggott [51] determined for the listed methods (fragmentation, pull-out, push-in and microbond test) clear deviations of the results. The author explained this scattering with unknown effects in different produced samples like the pressure across the interface, the coefficient of friction, a mixed fracture failure and the fact that it is not clear how a work of fracture can be related to the actual properties affected by the interface, such as compressive strength, and shear strength. Kobayashi et al. [52] pointed out that the specimen manufacturing parameters have a clear influence on the apparent IFSS of jute yarn/PLA composites measured with the bundle pull-out test. The authors investigated the influence of moulding pressure, temperature and press-time on the apparent IFSS. It was found that the jute fibre bundles in the jute yarn are damaged at a pressure above 0.5 MPa, and the

19

IFSS decreased with increasing pressure. The pressing temperature was examined in the range between 170 and 220 °C. The apparent IFSS increased significantly at a change of temperature from 170 °C to 180 °C. At 170 °C, the PLA is very viscous and a bad wetting of the fibres with the matrix occured. At 185°C, the highest apparent IFSS was determined and, at temperatures above 200 °C, there was a significant reduction in the IFSS due to fibre degradation. At a pressing temperature of 190 °C and a pressure of 1.33 MPa, the authors reported a decrease in the apparent IFSS with an extended press time between 2 and 20 minutes [52]. The ascribed studies pointed out that the production parameters, the testing procedure and especially the evaluation of fibre geometry and embedding length is of special importance to generate reproducible results. The reproducibility of the results obtained with the pull-out test and the fragmentation test is described in the following sections.

Pull-out test Some authors have reported a function of the apparent pull-out IFSS in dependence on the embedding length. Islam et al. [53] and Pickering et al. [54] showed a decrease in the apparent IFSS with an increasing embedding length of hemp fibre bundle/PLA. As shown in Figure 17 this trend could not be confirmed for lyocell. The influence of the embedding length described by Islam et al. [53] and Pickering et al. [54] can be attributed to a greater scattering of the embedded fibre length. In their studies the embedding lengths of hemp fibre bundles in PLA ranged between 400 and 2000 µm. The higher embedding length of hemp fibre bundles in comparison to lyocell can be realised by the larger cross-sectional area of the hemp fibre bundles. Thus, greater forces can be applied for the pull-out of the fibre from the matrix, without fibre failure. Due to the smaller cross-sectional area of lyocell, the maximum embedding length is limited to approx. 250 µm. Hence, the embedding length of lyocell ranged only between 80 and 250 µm. It is assumed that higher variability in the embedding length would have a greater influence on the apparent IFSS values. Regarding the comparability of results, it is of special importance to pay attention to the embedding length. Beckert and Lauke [15] described the difficulties that arise in single-fibre interface tests, particularly for the pull-out test. Due to inhomogeneities in fibre surfaces, different stress distributions occur at the

20

interface and especially at the embedded ends of the fibres, which lead to stress concentrations. It is difficult to reproduce the same stress distribution of one sample. The stress distribution even changes with small inhomogeneities on the fibre surface or tiny voids. The stress concentrations have particularly significant influences on short embedding lengths [15]. Piggott [51] pointed out that the shrinkage pressure and the roughness of the fibre surface has no influence on the theoretical adhesion between fibre and matrix. Nevertheless, the shrinkage pressure and the surface roughness influence the apparent IFSS. That is the reason why a direct comparison between fibres with a rougher surface (eg. bast fibre bundles) and fibres with a smoother surface (eg. lyocell) is difficult. For this reason in the present study, the approach was chosen to subtract friction energy from the apparent IFSS to reduce the influence of a rougher fibre surface according to [45]. For lyocell fibres sometimes a double peak is observed within the force elongation curves as shown in Figure 18. Usually, lyocell fibres do not display a uniform fibre diameter over the fibre length. After the first debonding of the fibre (first peak in Figure 18) a sharp drop of force occurred and a second peak can be seen after a certain elongation. A similar behavior was observed by Yang and Thomason [40] for glass/PP. The authors found out that this phenomenon is based on a complete loss of the measured force by the pull-out of a large part of the embedded fibre. It is assumed that this behaviour produces a fibre slack, which leads to a displacement at no registered force until the fibre is restrained again [40]. In the case of an increasing fibre diameter in the fibre direction we assume that the fibres can additionally act as a wedge during pull-out and cause friction leading to a higher second peak. This behavior has nothing to do with adhesion. If this phenomenon occurs during the debonding of the fibre from the matrix and leads to an increase of the pull-out force the calculated IFSS is overestimated. The probability for this phenomenon increases with increasing embedding length. Results of specimens which showed this behaviour were excluded from the evaluation. In the overall consideration the IFSS is dependent on the debonding force, the fibre cross-sectional area and the embedding length. The determination of the fibre cross-sectional area plays a major role and has a significant influence on the calculated apparent IFSS. Especially, with non-round fibre cross-sections (elliptical or polygonal shape), a common source of error is the measurement of the fibre and fibre bundle cross-sectional area. As an examination of Valadez-Gonzales et al. [47] shows,

21

the calculated apparent IFSS values vary depending on the kind of the determination of the crosssectional area. The authors determined the cross-sectional area of henequen fibre bundles on the one hand on the basis of the fibre diameter and on the other hand from the perimeter of the same fibre bundle. The henequen fibre bundles were processed with a PLA matrix to pull-out test specimens. An apparent IFSS of 3.3 MPa was achieved when the diameter was used for the calculation of the crosssectional area and an IFSS of 2.2 MPa was calculated when the perimeter was used. This example shows that an absolute comparison of results is only possible, in addition to other factors, when an exact measurement of the embedding length and a uniform determination of the cross-sectional area of the fibre is carried out. The measurement and evaluation of the cross-sectional area is especially difficult for natural bast fibre bundles which display a polygonal shape.

Fragmentation test Fragmentation tests carried out by Huber and Müssig [20] and Awal et al. [56] exactly match the experimental parameters which are used in the present work. The same type of flax fibre and an identical MAPP matrix was used. The calculated apparent IFSS of flax/MAPP in this study is 15.8 MPa and a critical fragment length of 2.8 mm was determined. Huber and Müssig [20] measured a value of 12.0 MPa for the apparent IFSS and 3.2 mm for the critical fragment length. The data of Awal et al. [56] provided 13.9 MPa for the IFSS and 2.4 mm for the critical fibre length. The slightly higher apparent IFSS in the present work can be explained by the correction of the fibre bundle diameter. While Huber and Müssig and Awal et al. adopted the fibre width as fibre diameter, the fibre crosssectional area was corrected in the present study because of the polygonal shape of the fibre bundles. For this purpose, fibre width and fibre thickness were measured from optical light micrographs by using the ImageJ software 1.3.7v (Wayne Rasband, National Institute of Health, USA). The calculated fibre cross-sectional area is based on an ellipse. Since the fibre thickness is usually clearly smaller than the fibre width, a smaller cross-sectional area is calculated, resulting in a higher apparent IFSS. Without this correction, the apparent IFSS is on the level of the values determined by Huber and Müssig and Awal et al. The results show that the fragmentation test is reproducible and independent of the examiner under constant test conditions. The investigations of Awal et al. [56] also showed that the

22

test speed has a low impact on the apparent IFSS of the samples. The results of flax/MAPP ranged from 11.2 MPa at a test speed of 2 mm/min and 12.3 MPa at a test speed of 0.1 mm/min, but a larger clamping length (30 mm) resulted in a clearly lower apparent IFSS of 9.1 MPa. This result shows the clear dependence of the characteristics on the experimental setup and the experimental procedure. As the discussion has shown for the apparent IFSS, the data are highly dependent on the fibre crosssectional area and only comparable with different experiments when the same measurement of the cross-sectional area occurred. Valadez-Gonzales [47] showed this effect for henequen in a HDPE matrix. The apparent IFSS was determined with the fragmentation test. The authors calculated the cross-sectional area of the fibre diameter on the one hand from the fibre diameter, resulting in an IFSS of 6.0 MPa, while the calculation from the perimeter resulted in a value of 4.4 MPa. A study by Bos [18] showed the influence of the size-effect regarding the cross-sectional area of single flax fibres and fibre bundles in a PP matrix. On the one hand, flax fibre bundles with a diameter of 30 - 120 µm, and on the other hand single flax fibres with a diameter of 10 - 15 µm were investigated. The results clearly differed. Whereas for the flax fibre bundles an IFSS of 8.0 MPa was calculated, the IFSS of the examined single fibres was significant higher and resulted in a value of 13.0 MPa. This also had an effect on the critical fragment length, which was determined for the flax fibre bundles with 3.80 mm and for the single fibres with 0.98 mm. As discussed for the pull-out test, the transmission of the characteristic values to whole composite structures is difficult. Not only fibre/matrix interactions, but also fibre/fibre interactions due to fibre agglomeration of several fibres take place. Practical tests show that the IFSS depends significantly on the fibre content in the composite. Thomason [6] established for long fibre-reinforced PP a decrease in the IFSS with increasing fibre mass content. This case cannot be represented by single fibre test methods.

Summary & conclusions The measurement of the fibre/matrix interaction has shown that established measurement methods for the determination of the fibre/matrix adhesion of cellulose fibres in a PLA, MAPP or PP matrix are

23

suitable. As described by the hypothesis, significantly higher values were achieved for bast fibre bundles as compared to regenerated cellulose fibres. The investigated flax and kenaf fibre bundles showed the same values for the apparent IFSS; however, differences in the critical fibre length were seen, which can be explained by different fibre bundle fineness. Thus, the hypothesis that states that different bast fibre bundles have different apparent IFSS values could be discarded in the case of kenaf versus flax. Despite of the adhesion promoter, the differences between lyocell/MAPP and lyocell/PP were not significant, but higher values were demonstrated for the MAPP samples compared to the PP samples. Even small improvements can have a considerable influence on the mechanical properties of the composite, because the fibres are embedded in a very large amount. The apparent IFSS of cellulose/PLA and cellulose/MAPP showed, from a statistical perspective, the same results. The hypothesis that the fragmentation test and the pull-out test give different results cannot be refuted. For cellulose fibres, higher apparent IFSS was measured with the pull-out test compared to the fragmentation test due to different sample preparation methods and stress distributions in the sample. If the interfacial friction which has nothing to do with the theoretical adhesion is subtracted from the apparent IFSS clearly lower values are obtained with the pull-out test, especially for bast fibre bundles which displayed a higher friction due to the higher surface roughness. The hypothesis that the pull-out test and microbond test provide similar values must be discarded. Significant differences were determined. The discussion shows that the measurement of practical fibre/matrix-adhesion is more complex for fibre bundles than for single fibres and the determination of the embedding length and the crosssectional area is of great importance. In the overall consideration, the reproducibility of the test methods is limited as shown by the extreme variation of literature values. In addition to sample preparation and fibre fineness, the test procedure (test speed, clamping length, embedding length, etc.) and the stress distribution, which is influenced by small defects, is usually not reproducible; these parameters have a significant influence on the results. Nevertheless, single-fibre test methods are a valuable tool to study the adhesion of fibres and fibre bundles in different matrices, and to compare them. The fibre/matrix interactions of small amounts of a sample can be investigated, e.g. to study the influence of different fibre treatments.

24

References

[1] Rao, V.; Franco, P. H.; Ozzello, A. D. & Drzal, L. T.: A direct comparison of the fragmentation test and the microbond pull-out test for determining the interfacial IFSS. The Journal of Adhesion, 1991, 34, 65-77 [2] Herrera-Franco, P. & Drzal, L.: Comparison of methods for the measurement of fibre/matrix adhesion in composites. Composites, 1992, 23, 1-27 [3] Adusumali, R.-B.; Reifferscheid, M.; Weber, H.; Roeder, T.; Sixta, H. & Gindl, W.: Mechanical properties of regenerated cellulose fibres for composites. Macromolecular Symposia, 2006, 244, 119125 [4] Narkis, M. & Chen, E.: Review of methods for characterisation of interfacial fibre-matrix interactions. Polymer Composites, 1988, 9, 245-251 [5] Luo, S. & Netravali, A.: Interfacial and mechanical properties of environment-friendly "green" composites made from pineapple fibres and poly(hydroxybutyrate-co-valerate) resin. Journal of Materials Science, 1999, 34, 3709-3719 [6] Thomason, J. L. & Vlug, M. A.: Influence of fibre length and concentration on the properties of glass fibre-reinforced polypropylene. 4. Impact properties. Composites: Part A: Applied Science and Manufacturing, 1997, 28, 277-288 [7] Dutschk, V.: Oberflächenkräfte und ihr Beitrag zu Adhäsion und Haftung in glasfaserverstärkten Thermoplasten. PhD thesis. Fakultät für Maschinenwesen der Technischen Universität Dresden, 2000. – in German [8] Zhandarov, S. & Mäder, E.: Characterisation of fibre/matrix interface strength: Applicability of different tests, approaches and parameters. Composites Science and Technology, 2005, 65, 149-160 [9] Ji, X.; Dai, Y.; Zheng, B. L.; Ye, L. & Mai, Y.-W.: Interface end theory and re-evaluation in interfacial strength test methods. Composite Interfaces, 2003, 10, 567-580 [10] Adusumalli, R.-B.; Reifferscheid, M.; Weber, H. K.; Roeder, T.; Sixta, H. & Gindl, W.: IFSS of the lyocell fibre/polymer matrix interface evaluated with the microbond technique. Journal of Composite Materials, 2010, 46, 359-367 [11] Clyne, T. & Jones, F.: In: Mortensen, A. (ed.): Encyclopaedia of materials: science and technology. Chapter 3.7.17: Composites: Interfaces, Elsevier, Oxford, UK, 2001, 3-13 [12] Feih, S.; Wonsyld, K.; Minzari, D.; Westermann, P. & Lilholt, H.: Testing procedure for the single fibre fragmentation test. Risø National Laboratory, Roskilde, Denmark, 2004

25

[13] Ramirez, F.; Carlsson, L. & Acha, B.: A method to measure fracture toughness of the fibre/matrix interface using the single-fibre fragmentation test. Composites: Part A: Applied Science and Manufacturing, 2009, 40, 679-686 [14] Guar, U. & Miller, B.: Microbond method for determination of the IFSS of a fibre/resin interface: Evaluation of experimental parameters. Composites Sciences and Technology, 1989, 34, 35-51 [15] Beckert, W. & Lauke, B.: Critical discussion of the single-fibre pull-out test: does it measure adhesion? Composites Science and Technology, 1997, 57, 1689-1706 [16] Wang, C.: Fracture mechanics of single-fibre pull-out test. Journal of Materials Science, 1997, 32, 483-490 [17] van den Oever, M. J. A. & Bos, H. L.: Critical fibre length and apparent interfacial IFSS of single flax fibre polypropylene compsites. Advanced Composite Materials Letters, 1998, 7, 81-85 [18] Bos, H. L.; Molenveld, K.; Teunissen, W.; van Wingerde, A. M. & van Delft, D. R. V.: Compressive behaviour of unidirectional flax fibre reinforced composites. Journal of Materials Science, 2004, 39, 2159-2168 [19] Park, J.-M.; Quang, S. T.; Hwang, B.-S. & DeVries, K. L.: Interfacial evaluation of modified jute and hemp fibres/polypropylene (PP)-maleic anhydride polypropylene copolymers (PP-MAPP) composites using micromechanical technique and nondestructive acoustic emission. Composites Science and Technology, 2006, 66, 2686-2699 [20] Huber, T. & Müssig, J.: Fibre matrix adhesion of natural fibres cotton, flax and hemp in polymeric matrices analyzed with the single fibre fragmentation test. Composite Interfaces, 2008, 15, 335-349 [21] Stamboulis, A.; Baillie, C. & Schulz, E.: Interfacial characterisation of flax fibre-thermoplastic polymer composites by the pull-out test. Angewandte Makromolekulare Chemie, 1999, 272, 117-120 [22] Wong, S.; Shanks, R. & Hodzic, A.: Effect of Additives on the Interfacial Strength of Poly(Llactic acid) and Poly(3-hydroxy butyric acid)-Flax Fibre Composites. Composites Science and Technology, 2007, 67, 2478-2484 [23] Yuan, X. W.; Jayaraman, K. & Bhattacharyya, D. Fakirov, S. & Bhattacharyya, D. (ed.): Mechanical performance of plasma-treated natural fibre-polypropylene composites. Handbook of engineering biopolymers - homopolymers, blends and composites, Carl Hanser Verlag, Munich, Germany, 2007, 329-412 [24] Taha, I. M. A.: Processing and characterisation of selected discontinous natural fibres and their polymer composites. Institut für Polymerwerkstoffe und Kunststofftechnik, TU Clausthal, Fakultät für Materialwissenschaften, Clausthal-Zellerfeld, Germany, 2007

26

[25] Prasad, S.; Pavithran, C. & Rohatgi, P.: Alkali treatment of coir fibres for coir-polyester composites. Jounal of Materials Science, 1983, 18, 1443-1454 [26] Mottershead, B. & Eichhorn, S.: Deformation micromechanics of model regenerated cellulose fibre-epoxy/polyester composites. Composites Science and Technology, 2007, 67, 2150-2159 [27] Angelini, L. G.; Lazzeri, A.; Levita, G.; Fontanelli, D. & Bozzi, C.: Ramie (Boehmeria nivea (L.) Gaud.) and spanish broom (Spartium Junceum L.) fibres for composite materials: Agronomical aspects, morphology and mechanical properties. Industrial Crops and Products, 2000, 11, 145-161 [28] Anuar, H.; Zuraida, A.; Morlin, B. & Kovács, J. G.: Micromechanical property investigations of poly(lactic acid)-kenaf fibre biocomposites. Journal of Natural Fibres, 2011, 8, 14-26 [29] Schmid, G. H.; Dvorak, M.; Müller, J. & Müssig, J.: Characterizing flock fibers using quantitative image analysis. Flock - English issue, 2004, 30, 6-12 [30] Kelly, A. & Tyson, W. R.: Tensile properties of fibre-reinforced metals: Copper/tungsten and copper/molybdenum. Journal of the Mechanics and Physics of Solids, 1965, 13, 329-338 [31] Felix, J. & Gatenholm, P.: The nature of adhesion in composites of modified cellulose fibres and polypropylene. Journal of Applied Polymer Science, 1991, 42, 609-620 [32] Joseph, P.; Rabello, M.; Mattoso, L.; Jospeh, K. & Sabu, T.: Environmental effects on the degradation behaviour of sisal fibre reinforced polypropylene composites. Composite Science and Technology, 2002, 62, 1357-1372 [33] Mechraoui, A.; Riedl, B. & Rodrigue, D.: The effect of fibre and coupling agent content on the mechanical properties of hemp/polypropylene composites. Composite Interfaces, 2007, 14, 837-848 [34] Arbelaiz, A.; Fernandez, B.; Cantero, G.; Llano-Ponte, R.; Valea, A. & Mondragon, I.: Mechanical properties of flax fibre/polypropylene composites. Influence of fibre/matrix modification and glass fibre hybridization. Composites: Part A: Applied Science and Manufacturing, 2005, 36, 1637-1644 [35] Beckermann, G.; Pickering, K. & Foreman, N.: Evaluation of the mechanical properties of injection moulded hemp fibre reinforced polypropylene composites. Advanced Materials Research, 2007, 29-30, 303-306 [36] El-Sabbagh, A.; Steuernagel, L. & Ziegmann, G.: Mechanical properties modelling of natural fibre thermoplastic composites. In: M-Base Engineering & Software GmbH, Aachen, D. (ed.) NFibreBase Workshop, Bremen, Germany, 21.-22.9.2009 2009 [37] Mirza, F. A.; Rasel, S. M.; Kim, M. S.; Afsar, A. M.; Kim, B. S. & Song, J. I.: Lyocell fibre reinforced polypropylene composites: Effect of matrix modification. Advanced Materials Research, 2010, 123-125, 1159-1162

27

[38] Endres, H.-J. & Siebert-Raths, A. (ed.): Engineering Biopolymers: Markets, Manufacturing, Properties and Applications. Carl Hanser Verlag, Munich, Germany, 2011 [39] Reddy, N.; Nama, D. & Yang, Y.: Polylactic acid/polypropylene polyblend fibres for better resistance to degradation. Polymer Degradation and Stability, 2008, 93, 233 – 241 [40] Yang, L. & Thomason, J.: Development and application of micromechanical techniques for characterising interfacial shear strength in fibre-thermoplastic composites. Polymer Testing , 2012, 31, 895 - 903 [41] Khoshkava, V. & Kamal, M. R.: Effect of Surface Energy on Dispersion and Mechanical Properties of Polymer/Nanocrystalline Cellulose Nanocomposites. Biomacromolecules, 2013, 14, 3155-3163 [42] Graupner, N.; Fischer, H.; Ziegmann, G. & Müssig, J.: Improvement of fibre/matrix adhesion of regenerated cellulose fibre reinforced PP-, MAPP- and PLA-composites by the use of lignin. In: Scarponi, C. (Ed.): Proceedings of the Fourth International Conference on Innovative Natural Fibre Composites for Industrial Applications, Rome, Italy, 2013 [43] Müssig, J.; Fischer, H.; Graupner, N. & Drieling, A.: Testing methods for measuring physical and mechanical fibre properties. Chapter 13. In: Müssig, J. (Ed.): Industrial applications of natural fibres: structure, properties and technical applications, John Wiley & Sons, Ltd., 2010, pp. 269-311 [44] Morlin, B. & Czigány, T.: Investigation of the surface adhesion of natural fibre reinforced polymer composites with acoustic emission technique. Proceeding of the 8th polymers for advanced technologies international symposium, Budapest, Hungary, 2005, 13-16 [45] Le Duigou, A.; Bourmaud, A.; Balnois, E.; Davies, P. & Baley, C. Improving the interfacial properties between flax fibres and PLLA by a water fibre treatment and drying cycle. Industrial Crops and Products , 2012, 39, 31 - 39 [46] Graupner, N.: Application of lignin as natural adhesion promoter in cotton fibre-reinforced poly(lactic acid) (PLA) composites. Journal of Materials Science, 2008, 43, 5222-5229 [47] Valadez-González, A.; Cervantes-Uc, J. M.; Olayo, R. & Herrera-Franco, P. J.: Effect of fibre surface treatment on the fibre-matrix bond strength of natural fibre reinforced composites. Composites: Part B: Engineering, 1999, 30, 309-320 [48] Herrera-Franco, P. & Valadez-González, A.: A study of the mechanical properties of short natural-fibre reinforced composites. Composites: Part B: Engineering: Engineering, 2005, 36, 597-608 [49] Piggott, M. R. & Dai, S. R.: Fibre pull-out experiments with thermoplastics. Polymer Engineering and Science, 1991, 31, 1246-1249

28

[50] Wagner, H. D.; Gallis, H. E. & Wiesel, E.: Study of the interface in kevlar 49-epoxy composites by means of microbond and fragmentation tests: Effects of materials and testing variables. Journal of Materials Science, 1993, 28, 2238-2244 [51] Piggott, M. R.: Why interface testing by single-fibre methods can be misleading. Composites Science and Technology, 1997, 51, 965-974 [52] Kobayashi, S.; Yamamoto, T. & A. Nakai, A.: Interfacial shear strentgh evaluation of jute/poly(lactic acid). Journal of Solid Mechanics and Materials Engineering, 2009, 3, 1063-1070 [53] Islam, M.; Pickering, K. & Foreman, N.: Influence of alkali treatment on the interfacial and physico-mechanical properties of industrial hemp fibre reinforced polylactic acid composites. Composites: Part A: Applied Science and Manufacturing, 2010, 41, 596-603 [54] Pickering, K. L.; Sawpan, M. A.; Jayaraman, J. & Fernyhough, A.: Influence of loading rate, alkali fibre treatment and crystallinity on fracture toughness of random short hemp fibre reinforced polylactide bio-composites. Composites: Part A: Applied Science and Manufacturing, 2011, 42, 11481156 [55] Burgstaller, C.: Comparison of interfacial IFSS from fibre pull-out and mechanical testing in polypropylene sisal composites. International Journal of Materials and Product Technology, 2009, 36, 11-19 [56] Awal, A.; Cescutti, G.; Ghosh, S. & Müssig, J.: Interfacial studies of natural fibre/polypropylene composites using single fibre fragmentation test (SFFT). Composites: Part A: Applied Science and Manufacturing, 2011, 42, 50-56 [57] Cho, D.; Seo, J. M.; Lee, H. S.; Cho, C. W.; Han, S. O. & Park, W. H.: Property improvement of natural fibre-reinforced green composites by water treatment. Advanced Composite Materials, 2007, 16, 299-314 [58] Islam, M. S.: The influence of fibre processing and treatments on hemp fibre/epoxy and hemp fibre/PLA composites. University of Waikato, Faculty of Materials and Process Engineering, Hamilton, New Zealand, 2008 [59] Ji, S.; Cho, D.; Park, W. & Lee, B.: Electron beam effect on the tensile properties and topology of jute fibres and the interfacial strength of jute-PLA green composites. Macromolecular Research, The Polymer Society of Korea, co-published with Springer, 2010, 18, 919-922 [60] Karlsson, J. O.; Blachot, J.-F.; Peguy, A. & Gatenholm, P.: Improvement of adhesion between polyethylene and regenerated cellulose fibres by surface fibrillation. Polymer Composites, 1996, 17, 300-304 [61] Bos, H.-L.: The potential of flax fibres as reinforcement for composite materials. Doctoral thesis. Department of Chemical Engineering, Technische Universiteit Eindhoven, 2004

29

[62] Lee, S. G.; Choi, S.-S.; Park, W. H. & Cho, D.: Characterisation of surface modified flax fibres and their biocomposites with PHB. Macromolecular Symposia, 2003, 197, 88-89 [63] Le Duigou, A; Davies, P. & Baley, C.: Improving the interfacial bonding of flax fibre/poly(Llactide) bio-composites. Composites Science and Technology, 2010, 70, 231-239 [64] Sawpan, M. A.: Mechanical performance of industrial hemp fibre reinforced polylactide and unsaturated polyester composites. The University of Waikato, Faculty of Materials and Process Engineering, Hamilton, New Zealand, 2010 [65] Taha, I. M. & Ziegmann, G.: Potential of sisal reinforced biodegradable polylactic acid and polyvinyl alcohol composites. Key Engineering Materials, 2010, 425, 167-178 [66] Torres, F. & Cubillas, M.: Study of the interfacial properties of natural fibre reinforced polyethylene. Polymer Testing, 2005, 24, 694-698 [67] Yang, L. & Thomason, J. L.: Interface strength in glass fibre-polypropylene measured using the fibre pull-out and microbond methods. Composites: Part A: Applied Science and Manufacturing, 2010, 41, 1077-1083

30

Figures

Figure 1: Preparation of a pull-out test specimen (A) and testing procedure of the pull-out test (B).

Figure 2: Preparation of a test specimen for the microbond test (A) and testing procedure of the microbond test (B). The metal frame is clamped in the upper clamps and the fibre in the lower clamps.

31

Figure 3: Typical fragments of a lyocell fibre (left) and a flax fibre bundle (right) in a MAPP matrix after the fragmentation test.

Figure 4: Box-Whisker-plots with mean values and standard deviation (grey) of the interfacial IFSS (IFSS) for lyocell/PLA, lyocell/MAPP and lyocell/PP measured with the pull-out test (A) and lyocell/MAPP and lyocell/PP measured with the single fibre fragmentation test (B) (* shows results which are not distributed normally; different letters show significant differences).

Figure 5: Box-Whisker-plots with mean values and standard deviation (grey) of the IFSS for flax/PLA, flax/MAPP and flax/PP, measured with the pull-out test (A) and the single fibre fragmentation test (B) (* shows results which are not distributed normally, different letters show significant differences).

32

Figure 6: Box-Whisker-plots with mean values and standard deviation of the IFSS for flax/PLA and kenaf/PLA measured with the pull-out test (A) and flax/MAPP and kenaf/MAPP measured with the single fibre fragmentation test (B) (* shows results which are not distributed normally; different letters show significant differences).

Figure 7: Typical force-elongation-curves of a bast fibre bundle and a single lyocell fibre in a PLA, MAPP or PP matrix.

33

Figure 8: SEM micrographs of kenaf (A & B) flax (C) and lyocell (D).

Figure 9: Atomic force micrograph with surface topography and average roughness values measured at the vertical and horizontal grey lines (grey numbers) of lyocell (A) and flax (B).

34

Figure 10: Box-Whisker-plots with mean values and standard deviation (grey) of the critical fibre length for lyocell/PLA, lyocell/MAPP and lyocell/PP measured with the pull-out test (A) and lyocell/MAPP and lyocell/PP measured with the fragmentation test (B) (* shows results which are not distributetd normally; different letters show significant differences).

Figure 11: Box-Whisker-plots with mean values and standard deviation (grey) of the critical fibre length for flax/PLA, flax/MAPP and flax/PP measured with the pull-out test (A) and flax/MAPP and flax/PP measured with the fragmentation test (B) (* shows results which are not distributetd normally; different letters show significant differences).

Figure 12: Box-Whisker-plots with mean values and standard deviation (grey) of the critical fibre length of flax/PLA und kenaf/PLA, measured with the pull-out test (A) and flax/MAPP and kenaf/MAPP, measured with the fragmentation test (B) (* shows results which are not distributetd normally; different letters show significant differences).

35

Figure 13: SEM-micrographs of the interface of misleaded pull-out test specimens after testing of lyocell/PLA (A), lyocell/MAPP (B) and lyocell/PP (C).

Figure 14: Macroscopic view on a pull-out test sample (A) and micrograph of a microbond test sample (B); both prepared with a lyocell fibre.

Figure 15: Box-Whisker plots with mean values and standard deviation (grey) of the IFSS (A) and Lc (B) of lyocell/PLA measured with the pull-out test and the microbond test (* shows results which are not distributetd normally; different letters show significant differences).

36

Figure 16: IFSS in dependence on the critical fibre length of different mating materials measured with pull-out tests, fragmentation tests and microbond tests. Flax, hemp, ramie and kenaf were pooled to the group of “bast”. Results of pull-out tests and microbond tests were additionally pooled. Own data and literature data listed in Table 5 were used for this presentation. Only fibres without pre-treatment were considered.

37

Figure 17: IFSS in dependence on the embedding length of lyocell/PLA (A), lyocell/MAPP (B) and lyocell/PP (C).

Figure 18: Force-elongation-curve of a single lyocell fibre in a PLA, MAPP or PP matrix.

38

Tables

Table 1: Tensile properties of the cellulose fibres. * denotes results which are not distributed normally. Tensile strength

Young´s modulus

Elongation at break

Type of fibre

n

in MPa

in GPa

in %

Lyocell

78

332 (± 81)*

6.86 (± 2.62)*

9.1 (± 2.2)*

Kenaf

72

653 (± 364)*

26.9 (± 11.2)*

2.5 (± 0.9)*

Flax

80

945 (± 776)*

22.9 (± 15.2)*

4.6 (± 2.1)*

Table 2: Interfacial IFSS in MPa of lyocell, flax and kenaf in different matrices (PLA, MAPP, PP) measured with the pull-out test and single fibre fragmentation test (* shows results which are not distributed normally). Pull-out test Material

Fragmentation test

Mean

Ratio

Mean

pairing

n

value

Sd.

Median

Lyocell/PLA

63

10.3*

4.9

9.8

Flax/PLA

22

28.3

10.9

27.7

Kenaf/PLA

33

25.7

10.6

25.6

Lyocell/PP

32

6.3*

4

Flax/PP

24

17.9

Lyocell/MAPP

34

Flax/MAPP

21

Pull-out

n

value

Sd.

Median

Fragmentation

5

64

3.8*

2.5

3.2

1.7

10.5

16.4

17

9.8*

6.8

7.5

1.8

8.8*

5.7

6.7

178 5.0*

4.3

3.9

1.8

24.3

11.1

23

36

15.8*

14.5 11.8

149 17.2*

37.8 10.9

Kenaf/MAPP

/

1.5

Table 3: IFSS versus interfacial friction (IF) of lyocell, flax and kenaf in different matrices (PLA, MAPP, PP) measured with the pull-out test (* shows results which are not distributed normally). Material pairing

IFSS in MPa

Sd.

IF MPa

Sd.

IFSS – IF in MPa

Lyocell/PLA

10.3*

4.9

4.4

2.3

5.9

Flax/PLA

28.3

10.9

17.4

10.0

10.9

Kenaf/PLA

25.7

10.6

15.9

8.0

9.8

Lyocell/PP

6.3*

4.0

4.5

0.8

1.8

Flax/PP

17.9

10.5

9.1

4.7

8.8

Lyocell/MAPP

8.8*

5.7

3.4

2.6

5.4

Flax/MAPP

24.3

11.1

14.4

6.3

9.9

39

Table 4: Critical fibre length of lyocell, flax and kenaf embedded in different matrices (PLA, MAPP, PP) determined with the fragmentation and pull-out test (* shows results which are not distributed normally). Critical fibre length

/ Critical fragment length

Pull-out-Test

Ratio

Fragmentationstest

Composite

Fragmentation n

Mean

Sd.

Median

n

Mean

S d.

Median

Pull-out

Lyocell/PLA

63

0.748*

0.513

0.583

Flachs/PLA

22

1.945*

2.521

1.204

Kenaf/PLA

33

0.727*

0.544

0.587

Lyocell/PP*

32

1.307*

0.884

1.153

64

2.500*

2.470

1.813

1.9

Flachs/PP

24

3.415*

2.988

2.784

17

3.993*

2.978

3.267

1.2

Lyocell/MAPP*

34

0.944*

0.557

0.853

180

1.457*

1.616

0.947

1.7

Flachs/MAPP

21

1.872*

1.191

1.270

36

2.804*

2.142

2.133

1.5

149

1.117*

0.601

0.987

Kenaf/MAPP

Table 5: IFSS and critical fibre / fragment length of different untreated fibre/matrix combinations investigated with pull-out, microbond and fragmentation tests (* single fibre, ** fibre bundle). Testing

IFSS

Critical fibre / fragment length in mm

Fibre

Matrix

procedure

in MPa

Flax*

HDPE

Pull-out

9.1

Flax*

HDPE

Pull-out

10.1

Henequen**

HDPE

Fragmentation

5.4

Henequen**

HDPE

Fragmentation

4.2

Henequen**

HDPE

Pull-out

2.5

Henequen**

HDPE

Pull-out

2.4

Flax*

LDPE

Pull-out

5.5

Flax*

LDPE

Pull-out

6.2

Lyocell*

LDPE

Fragmentation

4.8

Lyocell*

LDPE

Microbond

6.0

Micromodal*

LDPE

Microbond

5.0

Ramie**

LDPE

Microbond

10.0

Flax**

PP

Fragmentation

4.5

4.86

Flax*

PP

Fragmentation

13.0

0.98

Flax**

PP

Fragmentation

8.0

3.80

[21] [21] [48] [47] [48] [47] [21] [21] [60] [10] [10] [10] [56] [61] [61]

Flax**

PP

Fragmentation

9.8

3.99

this study

Flax**

PP

Microbond

3.4

Flax*

PP

Pull-out

10.6

0.76

Reference

[44] [21]

40

/

Flax**

PP

Pull-out

17.9

Hemp**

PP

Microbond

5.10

Kenaf**

PP

Fragmentation

7.4

5.05

this study

Lyocell*

PP

Fragmentation

3.8

2.50

this study

Lyocell*

PP

Microbond

4.2 1.20

[3] [10] [10]

1.31

this study

Lyocell*

PP

Microbond

5.3

Lyocell*

PP

Microbond

5.0

Lyocell*

PP

Pull-out

6.3

Micromodal*

PP

Microbond

3.0

Ramie**

PP

Microbond

4.9

Ramie**

PP

Microbond

5.9

3.45

this study

[44]

Ramie**

PP

Microbond

6.0

Sisal**

PP

Microbond

4.6

Sisal**

PP

Pull-out

5.0

Sisal**

PP

Pull-out

6.1

Cotton*

MAPP

Fragmentation

0.7

5.03

Flax**

MAPP

Fragmentation

13.9

2.42

Flax**

MAPP

Fragmentation

12.3

2.60

Flax**

MAPP

Fragmentation

12.8

2.45

Flax**

MAPP

Fragmentation

12.7

2.51

Flax**

MAPP

Fragmentation

11.2

2.60

Flax**

MAPP

Fragmentation

9.1

3.00

Flax**

MAPP

Fragmentation

12.0

3.20

[10] [3] [10] [10] [44] [55] [24] [65] [20] [56] [56] [56] [56] [56] [56] [20]

Flax**

MAPP

Fragmentation

15.8

2.80

this study

Flax*

MAPP

Pull-out

11.4

Flax**

MAPP

Pull-out

24.3

1.87

this study

Hemp**

MAPP

Fragmentation

14.3

3.16

Henequen*

MAPP

Microbond

4.1

[20] [57]

Kenaf**

MAPP

Fragmentation

17.2

1.12

this study

Lyocell*

MAPP

Fragmentation

5.0

1.46

this study

Lyocell*

MAPP

Pull-out

8.8

0.94

this study

Ramie**

MAPP

Fragmentation

24.9

0.98

Sisal**

PE

Fragmentation

2.16

Sisal**

PVA

Pull-out

17.3

Flax*

PLA

Microbond

15.3

Flax*

PLA

Microbond

18.2

Flax*

PLA

Microbond

22.2

Flax*

PLA

Microbond

9.9

Flax*

PLA

Microbond

9.0

[56] [66] [24] [65] [63] [63] [63] [63] [63] [22]

1.95

this study

Flax**

PLA

Pull-out

12.5

Flax**

PLA

Pull-out

28.3

Hemp**

PLA

Microbond

11.3

Hemp**

PLA

Pull-out

1.6

1.10

[21]

[44] [53]

41

Hemp**

PLA

Pull-out

1.8

Hemp**

PLA

Pull-out

5.6

Jute*

PLA

Microbond

5.5

Jute**

PLA

Microbond

4.6

Kenaf**

PLA

Microbond

5.4

Kenaf*

PLA

Microbond

10.7

Kenaf**

PLA

Pull-out

25.8

0.73

this study

Lyocell*

PLA

Pull-out

10.3

0.75

this study

Sisal**

PLA

Microbond

14.3

Sisal**

PLA

Pull-out

17.1

Flax**

PHB

Microbond

8.8

Kenaf**

PHB

Pull-out

13.2

1.95

this study

Lyocell*

PHB

Pull-out

7.1

0.92

this study

Ananas**

PHBV

Microbond

8.2

Flax**

Mater-Bi

Microbond

4.2

Hemp**

Mater-Bi

Microbond

3.0

Sisal**

Mater-Bi

Microbond

3.2

1.63

[58] [64] [57] [59] [28] [57]

[44] [24] [65] [62]

[5] [44] [44] [44]

42