wood pulp composites

Composites: Part A 46 (2013) 45–52

Contents lists available at SciVerse ScienceDirect

Composites: Part A journal homepage: www.elsevier.com/locate/compositesa

The influence of fibre length and damage on the mechanical performance of polypropylene/wood pulp composites Armin Thumm ⇑, Alan R. Dickson Scion, Private Bag 3020, Rotorua 3010, New Zealand

a r t i c l e

i n f o

Article history: Received 8 June 2012 Received in revised form 13 September 2012 Accepted 15 October 2012 Available online 2 November 2012 Keywords: A. Fibres A. Wood A. Thermoplastic resin B. Mechanical properties

a b s t r a c t Compression moulded wood fibre-reinforced thermoplastic composites were formed with radiata pine pulp fibres having four distinct fibre length distributions and three levels of fibre damage. The influence of these two parameters in regard to mechanical performance was investigated. Fibre length only had an influence when it dropped below a critical length. This critical length was found to be consistent with model calculations which put the critical length for radiata pine fibres at 0.8 lm. Fibre damage, defined in this study as dislocation and nodes, only had a minor influence on panel properties. Injection moulded reference samples showed trends for the influence of fibre length and fibre damage which were consistent with compression moulded samples. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Increasingly, wood pulp fibres are finding applications in fibre-reinforced thermoplastic composites. For example; NCell™ (GreenCore Composites Inc. Canada); Woodforce (Sonae Industria, Portugal); CreaMix (CreaFill Fibers Corp. USA); KarelineÒ (Kareline natural composites, Finland); UPM ForMi (UPM, Finland) are all products that incorporate wood pulp, rather than wood flour, for the reinforcement of plastics. There are a number of different pulp fibre types depending on wood source (relatively short but narrow hardwoods versus long but wide softwoods) or pulping method. Pulps where the fibres are liberated from each other by mechanical means at high temperature (e.g. thermo-mechanical pulps (TMPs) or medium density fibre board pulps (MDFs)) have higher yield but contain more fibre damage, such as fibre breakage and cell wall disruptions [1] than pulps produced by chemical processes that separate fibres by dissolving out wood components (e.g. kraft pulps). Additionally, pulps may be further processed (refined) by a mechanical process designed, largely, to increase fibre flexibility and generate fibrillar material. For paper products, the effect of refining is to increase the density and bonded area of the sheet, increasing the network strength. The differences in fibre dimensions, chemical composition, flexibility, amount of fines material and other pulp properties is tailored for the type and use of the paper product. There is also growing interest in optimising wood pulp properties for composites. Fibre strength and length are critical factors for

composite reinforcement performance. Investigating flax fibres between 3 and 25 mm length in polypropylene (PP) Garkhail et al. [2] found no improvement in composite strength for fibres longer than 6 mm. Laws and McLaughlin [3] conclude that the moduli of composites reinforced with short (glass) fibres behave as if they were reinforced with fibres of infinite length once the aspect ratio of the fibres is >100. The minimum length of a reinforcement fibre, i.e. the critical length, can be described as the length required to ensure sufficient adhesion between fibre and matrix so that the stress in the fibre can reach a fibre fracture stress. In its simplest form the critical length has been defined as [4]

Critical fibre length ðlc Þ ¼

1359-835X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compositesa.2012.10.009

ð1Þ

In order to calculate the critical length, the tensile strength of the fibre (rt) and the interfacial shear strength (s) need to be known. The fibre diameter is d and if the equation is divided by d it can be expressed as the ‘‘critical’’ aspect ratio being a ratio between fibre strength and interfacial shear strength. These factors have been measured for larger, agricultural fibres (e.g. van den Oever and Bos [5]) but are difficult to measure for small wood fibres. An alternative approach was taken by Rodriguez et al. [6] who calculated the fibre strength and interfacial shear strength through modelling of the mechanical properties of composites according to the method of Bowyer and Bader [4]. Using the deformation of the composite the critical fibre length in this case is defined as

⇑ Corresponding author. Tel.: +64 7 343 5899; fax: +64 7 343 5507. E-mail addresses: [email protected] (A. Thumm), alan.dickson @scionresearch.com (A.R. Dickson).

d rt 2s

lc ðect Þ ¼

Eft ect d 2s

ð2Þ

46

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

This equation requires knowledge of the modulus of the fibre in tension (Eft ). Based on the approach described by Rodriguez, Eft and the interfacial shear strength (s) can be calculated and consequently the critical fibre length determined (ect is the deformation of the composite which can be obtained from the stress/strain curve). In the case of the corn stalk fibre/PP composite used by Rodriguez critical fibre length can be calculated to be around 0.5 mm when appropriately compatibilised with maleated polypropylene (MAPP). The average diameter of the cornstalk fibres is given as 16.1 lm which would result in a ‘‘critical aspect ratio’’ of about 30. Fibre strength is also important and is influenced by fibre damage and defects. The presence, terminology and effect of fibre defects in wood fibre have been reviewed a number of times in relation to wood or paper properties [7–10] and more recently in relation to composite properties [11]. For example, the number of microcompressions [8] has been shown to correlate with a reduction in fibre stiffness [12] and the presence of defects such as microcompressions and nodes [8] is known to lead to a reduction in fibre strength [13]. Such defects can be formed during pulp processing, such as mixing [14,15]. The differences in the scale and types of processing and their effects on fibres explains why laboratory produced pulps are generally stronger than those produced in mills [16]. Studies on the effects of wood pulp fibre length and damage on composite properties have been contradictory. Long fibres have been shown to be beneficial for composites made from birch chemi-thermomechanical pulp and polyethylene [17], and thermomechanical pulp (TMP) and polylactic acid [18]. On the other hand, little difference in strength performance was found for bleached kraft/PP composites [19]. However, in these studies fibre length fractions were obtained by refining [17] and screening [18,19] and it is known that under such conditions length is not independent of other factors [20]. Beg and Pickering [21] investigating kraft pulp reinforced PP, attributed reductions in composite tensile strength and stiffness to fibre damage and length reduction brought about by increased levels of pulp refining. There have been a number of studies on a range of fibre types in a range of matrices that show that processing operations such as compounding and injection moulding have a major influence on fibre length [22– 27] and width [22,24,26] and consequently aspect ratio. The magnitude of the reductions in width and length are largely attributed to the magnitude of the shear forces [22] although the initial fibre dimensions can also have an influence on the dimensions after processing [24]. To our knowledge the mechanisms of fibre shortening for wood fibres in composites has not been examined in detail and will also not be examined in detail here. However, fibre breakage in flax fibres during processing [26] and in composite failure [27] has been shown to be related to the presence of prominent dislocations or nodes (generally called kink bands in non-wood fibres). Where bundles, rather than individual fibres are processed and show reductions predominantly in width but also in length, this is likely due at least in part to the separation into individual fibres. For wood fibres the effect of shear and temperature at the fibre–fibre interface is well understood and forms the basis of the different forms of mechanical pulping for the pulp and paper industries [1]. In this study we investigate, in a controlled manner, the related effects of fibre length and damage on fibre-reinforced thermoplastic composite properties independently. The influence of fibre length was investigated with a method in which an initially long fibre fraction was screened from a TMP. The fraction was then reduced in fibre length in a controlled and predictable manner by manual cutting. To investigate the influence of fibre damage a kraft pulp was submitted to extended periods of blending in a mixer (disintegrator). A wet forming technique and compression moulding were then used for composite production to minimise any

further damage and length reduction that is known to occur during compounding [6,24,25,28,29]. The investigation also examined the influence of injection moulding on these materials. This was done for two reasons: firstly, to investigate whether trends observed in compression moulding would also be present in injection moulded samples and secondly, to investigate what effect injection moulding has on fibre length. The material for injection moulding was created by manually cutting compression moulded samples into pellets and feeding them directly into an injection moulder. 2. Material and methods 2.1. Materials The TMP for fibre length investigation as well as the kraft pulp for fibre damage investigation was produced at Scion, New Zealand from Pinus radiata chips. Polypropylene (PP) fibres (Atofina) were provided by FiberVisions (Covington, Georgia, USA). The PP fibres were 23 lm wide and 5 mm long with a melting point (by DSC) of 165–166 °C. The fibres include an undisclosed coupling agent. PP pellets (Seetech M1600, LG Daesan Petrochem, Korea) were injection moulded with a bleached kraft pulp (Carter Holt Harvey, NZ) to produce wood fibre tensile specimen required for modelling purposes. 2.2. Production of pulps with different fibre length A 2-stage radiata pine slabwood TMP (120 Canadian standard freeness) was produced using 125 °C preheating and a residence time of three minutes. The pulp was pre-screened at the production site to obtain a fraction with increased fibre length. This fraction was used as a starting point for the experiments below. To generate pulps with different fibre length independently of other fibre variables a modified version of that employed by He et al. [30] was used. The TMP was first passed through a 10 mesh screen (approximately 2 mm opening width) and the material on the screen collected to obtain a pulp with an average length weighted fibre length (LWL) of 3.0 mm.

LWL ¼

X

2

ni li

 X 1  ni li

ð3Þ

This long fibre fraction was then reduced by making handsheets with random fibre orientation and cutting them with a guillotine. A FiberScan instrument (Andritz Sprout-Bauer) was used to measure fibre length distributions of the resulting pulps. The measurement principle of this instrument is based on fibres passing through a narrow passage between a light source and detector and the resulting light fluctuations are measured and converted to fibre length. The instrument is well suited to handle long stiff fibres such as TMP and MDF but has limited sensitivity to very small fibres such as fines. Four different fibre lengths were achieved by different cutting approaches as detailed in Table 1. 2.3. Production of pulps with different levels of fibre damage A kraft pulp with a kappa number of 16 was prepared from P. radiata slabwood chips under the following conditions: 16% effective alkali charge; 30% sulphidity; 4:1 liquor:wood ratio; 90 min to 170 °C; 90 min at 170 °C. The general principles and procedures of kraft pulping can be found in [31]. The cooked chips were defibrated at 1% consistency in a Crompton Parkinson laboratory stirrer at 1400 rpm for 10 min and then screened through a 0.25 mm flatbed screen and collected on a 200 mesh screen (74 lm opening). Damage was inflicted on the pulp by disintegration at 1.2% consistency following AS/NZS 1301.203s:2007. The disintegrator is a laboratory mixer as specified by AS/NZS 1301.214s:2007, with a

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

47

Table 1 Cutting procedures to achieve different fibre lengths. Fibre length (LWL)

Cutting procedure

3.0 mm 2.1 mm 1.3 mm 0.8 mm

– Handsheets are cut into 2 mm squares Handsheets are cut into 1 mm squares Handsheets are cut into 2 mm squares then re-pulped and formed into new handsheets. This process is repeated 3 times

3-bladed propeller rotating at 2940 rpm. To determine the effect of disintegration time on fibre length, disintegration was performed for 6 h with aliquots being removed periodically for the measurement of length and fines contents. Based on the 6 h experiment samples with disintegration times of 2, 30 and 150 min were selected to provide an appropriate range of fibre damage. Length and fines contents were measured using a Metso FiberLab which uses capillary flow and 2-CCD cameras. For the 6 h experiment single FiberLab measurements were made (approximately 4000–5000 fibres). For the selected treatment times, 12 replicate measurements of length and fibres contents were performed (approximately 50,000–66,000 fibres). Disintegration is required to deflocculate fibres prior to processing, therefore, there was no ‘‘control’’ (0 min) without any disintegration. Relative fibre damage was assessed by mounting the wet fibres in glycerol on glass slides and observing them using cross-polarised light on a Leica DMRB microscope fitted with a 40  objective and a Leica EC3 digital camera. Individual fibres were randomly selected and the camera rotated so the long axis of the fibre corresponded to the long axis of the photo frame (246 lm when using a 40  objective). Damage was quantified based on the presence of dislocations and nodes in the fibre walls. Dislocations are single slip planes in fibre walls visible under cross-polarised light. Nodes are regions of bending or compressive failure [8]. Examples of dislocations and nodes are shown in Fig. 1. The number of nodes and dislocations present on one side of the fibre were counted. Additionally, fibres were subjectively assessed for their relative severity of damage. Two criteria were used: (1) the visible delamination of the fibre wall, usually in association with microcompressions and (2) the visible presence of a fibre fracture or indication of a disruption of the fibre wall more severe than just the presence of microcompressions. The slides were randomised and measurements were performed by an unbiased observer because, as evident in Fig. 1, the distinction between dislocations and nodes and their extent can be subjective. At least 150 fibres per treatment were measured or assessed.

2.4. Sample production 2.4.1. Compression moulding Polypropylene (PP) fibres and the appropriate wood pulp were weighed out to give a 40% wood fibre loading (wt%) and the PP fibres and the pulp fibres were disintegrated separately in water. For PP and TMP the disintegrations were performed for 10 min each. The kraft pulps did not receive any disintegration beyond that used to inflict the original levels of fibre damage (2, 30, 150 min). The PP and wood pulps were then combined and mixed using an overhead stirrer for several minutes. The combined fibres are formed into 17 cm square pads using a forming box. After evacuation of the water the pads were air dried for storage and then dried overnight at 60 °C before compression moulding. Pads were compression moulded at 180 °C for three minutes using a 3 mm spacer to define the thickness of the panel. A force of 200 kN (7 MPa) was sufficient to achieve this. The resulting density was 1000 kg m3. Three panels were made for each treatment in the fibre length investigation and five panels for each

Fig. 1. Damage in pulp fibres as seen under cross-polarised light. (a) Dislocations are present in the fibre walls (arrows) as single white or black lines. Fibres have had 2 min of disintegration. (b) Dislocations again present as single lines in fibre walls. In places several dislocations appear close together ({) and form thicker white or black regions. Such features are classed as nodes. In this case dislocations (arrows) can be seen to be continuous around at least part of the perimeter of the fibre joining a similar node on the opposing wall. Fibres have had 2 min of disintegration. (c) Large prominent nodes on both sides on the fibre. Individual dislocations can also be seen. Fibres have had 150 min of disintegration.

treatment for the fibre damage investigation. From each panel six flexural test samples of dimensions 13  130 mm and seven impact testing samples of dimensions 13  65 mm were cut. 2.4.2. Injection moulding The tensile specimens required for modelling purposes were produced by injection moulding. 40% (w/w) pelletised kraft pulp, 56.5% (w/w) PP pellets and 3.5% (w/w) MAPP were pre-blended in a mixing vessel. The blend was extruded using a LabTech Engineering LTE26-40 twin screw extruder with 26 mm screws and a length to diameter ratio of 40:1. The extruded strand was cooled in a waterbath and pelletised. Pellets were dried in a DRI-AIR

48

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

desiccant dryer for a minimum of 3 h at 60 °C before being injection moulded using a BOY 35 M moulding machine. Tensile specimens were injection moulded into a dog bone shape, based on ASTM D 638-01, with dimensions of 3.2  165 mm. Compression moulded samples were also used as the starting material for injection moulding. For this approach flexural test pieces were cut into pellet-sized pieces of about 12  5 mm and fed directly into the injection moulder to produce flexural test specimen with dimensions of 13  135 mm. The pellets were intentionally kept relatively large to minimise any shortening of fibres. The fibre length distributions (FibreScan) and damage assessments of fibres in injection moulded samples were determined after extraction with Xylene. The extraction procedure used 1 g of composite sample which was boiled in a paper thimble for 4 h in 200 ml of xylene. The thimble was then transferred to a soxhlet extractor and washed with xylene under reflux for a further 4 h. 2.4.3. Statistics All error bars shown are based on ±95% confidence intervals. The results of an analysis of variance (Anova) is shown in graphs using lower case letters. Different letters mean a significant difference at a = 0.05.

Fig. 2. Fibre length distributions of a long fibre fraction and the same fraction cut to three shorter lengths. The original pulp is included for reference. Fibre length was measured in 28 lm intervals.

Table 2 Fibre length data for the pulps used in the fibre length study. Pulp

Average length (mm)

Average length weighted length (mm)

3.0 mm 2.1 mm 1.3 mm 0.8 mm Original

1.31 1.17 0.83 0.47 1.18

3.00 2.09 1.33 0.77 2.24

2.5. Mechanical testing Flexural properties were measured using an Instron 5566 machine according to ASTM D 790-96a. The crosshead speed was 1.3 mm/min. The support span to depth ratio (L/d) was 16 to 1 giving a span of 50 mm. Specimens were tested to failure to obtain the Young’s modulus and the ultimate strength. Modulus and strength were corrected for minor variations in density by dividing values through density. Tensile properties were measured with the same instrument fitted with a 10 kN load cell and an external extensometer, according to ASTM D638-01. The crosshead speed was 5 mm/min. The gauge length was 115 mm and the extensometer length was set to 25 mm. Specimens were tested to failure to obtain the Young’s modulus, maximum tensile strength and in some cases elongation at Break. Tensile properties were only able to be measured for injection moulded samples. Samples for Izod notched impact strength were of dimensions 12.6 mm  63.5 mm  thickness in accordance with ASTM D 6110-06. Using a Ceast Notchvis, a 45° notch was machined into each sample to give a width of 10.16 mm between the tip of the notch and the edge of the material. Impact testing was undertaken using a Ceast Resil impact-testing machine. Izod samples were tested according to ASTM D 6110-06, at a velocity of 3.46 m/s, a 150° angle and a 0.5 J hammer.

Fig. 3. Strength and stiffness of a long fibre fraction vs the same fraction cut to three shorter length. (Different letters denote significantly different treatments within each series, a = 0.05).

3. Results and discussion 3.1. Fibre length The fibre length analysis shows distinct distributions for the long fibre fraction and the three pulps of shorter lengths produced from it (Fig. 2 and Table 2). No significant difference in composite strength can be observed for the pulps ranging from 3 mm to 1.3 mm in average length (LWL). A significant drop in performance does however occur if the fibre length of the pulp is reduced from 1.3 mm to 0.8 mm (Fig. 3). A similar pattern can be observed for composite stiffness (Fig. 3). The impact strength, as measured by the Izod notched test, shows a substantially increased performance only for the longest,

Fig. 4. Impact strength of composite panels made from different fibre length pulps. (Different letters denote significantly different treatments, a = 0.05).

3 mm, fibre fraction (Fig. 4). The pulps with reduced fibre length have similar performance with a slight drop for the pulp of shortest fibre length. None of the fibre-filled samples does however improve the impact performance beyond the unfilled PP.

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

Fig. 5. Strength and stiffness of injection moulded specimens based on pelletised compression moulded samples. (Different letters denote significantly different treatments, a = 0.05). Length numbers at the top are from fibres extracted from the injection moulded samples.

Fig. 6. Impact strength of injection moulded specimens based on pelletised compression moulded samples. (Different letters denote significantly different treatments, a = 0.05).

The compression moulded samples from the fibre length investigation were cut into pellet-sized pieces and used for the production of injection moulded samples. The injection moulded test pieces showed a very similar strength performance to the compression moulded samples in relative and absolute terms (Fig. 5). A comparison of fibre length after injection moulding with the original fibre length (Fig. 5) reveals some fibre length retention related to the original fibre length. The slight increase in strength for the 1.3 mm fibre might be due to a better processability of material with shorter fibres requiring substantially lower injection pressure compared to the long fibre pulps. Unlike the compression moulding samples the impact strengths of the injection moulded samples were similar for all the pulps (Fig. 6). It has to be kept in mind however, that injection moulding

49

leads to a shortening of fibres causing fibres of various initial length to be cut to similar final length in the composite, possibly explaining the similar impact performance of the different fibre length pulps. The slight increase for the 1.3 mm fibres is statistically significant but in practical terms not relevant. The shortening caused by injection moulding is likely to take fibre length substantially below a critical fibre length for impact resistance. The critical fibre length for radiata pine fibres was estimated following the method of Rodriguez given in the introduction using a bleached kraft pulp. The method requires tensile samples which were produced by injection moulding. The average stress–strain curve for these samples and the fibre length distribution of extracted fibres (Fig. 7) was used to calculate the interfacial shear strength (s) which was found to be 12 MPa. This is based on a measured fibre diameter of 23 lm (FibreLab) of extracted fibre rather than the literature value for P. radiata which is closer to 32 lm. It is known that fibre dimensions can decrease dramatically with drying [32,33]. The Young’s modulus of the fibre was calculated as 18.2 GPa using the Hirsch model as applied by Rodriguez. This is similar to values which have been found for kraft pulp of Norway spruce (16.1 GPa, [34]) and Scots pine (18.0 GPa, [35]). The critical fibre length was then calculated according to Eq. (2) as 0.8 mm. Naturally, the calculated numbers can only be taken as an estimate. They are based on the specific system of an injection moulded bleached kraft/PP composite. However, all pulps used in this study are from P. radiata so fibre dimension are comparable and critical fibre length of radiata pine fibres may in general be around 0.8 mm with some influence of fibre strength depending on the type of pulp. Assuming an average fibre diameter in the composite of 23 lm results in a ‘‘critical aspect ratio’’ of 35. The calculation also considers the influence of compatibilisation between the fibres and matrix, as a lack of compatibilisation leads to a lower interfacial shear strength (s) and results in a longer critical fibre length. Fibre alignment will also have an influence on the critical fibre length. The methodology by Bowyer and Bader [4]uses an orientation factor which is 1 for fibres aligned to the testing direction. It reduces to a third for fibres randomly disposed in a plane and reduces further to one sixth if fibres are randomly disposed in three dimensions. The orientation factor for the samples used in the calculations above was found to be 0.47. This indicates some alignment of fibres which may be caused by the flowing action of the matrix during injection moulding. A compression moulded panel with totally random alignment in x and y direction would be expected to have a lower orientation factor and as a consequence a slightly longer critical fibre length compared to a more aligned injection moulded sample.

Fig. 7. Stress/strain curve for an injection moulded wood fibre tensile reference sample made from kraft pulp and compatibilised PP and the fibre length distribution of fibres extracted from it. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

50

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

The drop in performance for the shortest pulp in the compression moulded fibre length series is likely due to a majority of fibres in this pulp dropping below a critical length which lies between 1.3 and 0.8 mm. The critical length was also indicated by the injection moulded samples. All pulps which did contain substantial amounts of fibres >1.0 mm show similar performance. When average fibre length drops below this value a substantial drop in performance results. These results are consistent with the calculated critical fibre length of 0.8 mm. It has however to be kept in mind that the calculations were based on injection moulded tensile samples which is a requirement of this theoretical approach. An improvement in impact strength was only found for the longest fibre pulp. This may indicate that a threshold has been reached at this fibre length where physical entanglement of fibres starts to have an influence. It can be hypothesised that a network of fibres bonded by friction and OH-bonding would interfere with crack propagation and consequently more energy would be required to promote a crack.

Fig. 10. Influence of fibre damage on flexural strength and stiffness for compression moulded samples. Different small letters denote significantly different treatments (a = 0.05).

3.2. Fibre damage The intention of this work was to examine the effect of fibre damage independently of fibre length. However the disintegration process resulted in a slight reduction in mean fibre length (2.14, 2.09 and 2.09 mm for the 2, 30 and 150 min disintegrations respectively). The drop from 2.14 to 2.09 mm was significant at the 95% confidence interval. An extensive analysis of the length distribution data revealed no individual fibre length class with a significant difference between treatments (data not shown). The analysis that illustrated the greatest visual differences between treatments was the cumulative frequency distribution (Fig. 8) where a slight separation between the lines can be seen between 0 and 2 mm. We largely attribute the differences in mean length to the generation of fines (lengths less than 0.02 mm) where fragments of fibre wall material have become separated from the fibre wall, right skewing (positive skew) the length distribution (data not shown) and reducing the mean length. We consider this length difference to be negligible and, additionally, there was no significant difference in length distribution between the 30 and 150 min treatments. Consequently the pulps were treated as if of equal length and

Fig. 8. Influence of disintegration time on fibre length.

Fig. 11. Impact strength of composites made from pulps with increasing levels of fibre damage (compression moulded samples). Different small letters denote significantly different treatments (a = 0.05).

Fig. 12. Injection moulded samples of composites made from pulps with increasing levels of fibre damage. Different small letters denote significantly different treatments (a = 0.05). Length numbers at the top are from fibres extracted from the injection moulded samples.

Fig. 9. The level of fibre damage caused by disintegration illustrated by (a) the number of microcompressions and nodes and (b) delaminations and fractures or disruptions.

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

51

Fig. 13. The presence of microcompressions and wall fractures/disruptions before and after injection moulding.

examined by microscopy to determine the level of fibre damage (Fig. 9). The longest disintegration treatment caused an approximately 45% increase in microcompressions, an 84% increase in nodes and a 323% increase in the presence of visible cell wall delaminations compared to the minimum 2 min disintegration treatment. There was no significant change in the presence of wall fractures and disruptions. A qualitative visual assessment showed few signs of cell wall fibrillation. These pulps where then used for panel manufacture using compression moulding to avoid further damage to the fibres. Flexural testing showed all pulps had a substantial increase in panel strength and stiffness over the unfilled polypropylene (Fig. 10). There was a trend to decreasing performance with increased fibre damage. The decline is however relatively minor especially for panel strength. No significant difference was found for impact strength of the different treatments (Fig. 11). Samples were pelletised by manual cutting and used to produce injection moulded pieces in the same manner as for the fibre length investigation (Fig. 12). Injection moulded samples with increasing levels of fibre damage show a similar behaviour to compression moulded samples with only a slight deterioration of panel performance caused by fibre damage. The extracted fibres showed a high degree of twisting [36] resulting from the injection moulding and/or the extraction process. The fibres also appeared to generally have smaller diameters and cell wall delaminations were rarely present, suggesting cell wall thickness reduction with drying. Due to the hydrophilic nature of cellulose, dried fibres generally at least partially rehydrate in the glycerol mounting medium. This was not evident with the extracted fibres, possibly due to residual PP or xylene making the fibres more hydrophobic, although this was not investigated. The extracted fibres were assessed for the presence of microcompressions and fractures and disruptions (Fig. 13). The fibres showed similar trends in microcompressions after injection moulding as the original fibres, albeit at a higher level. There was also a significant increase in the presence of fractures and disruptions after injection moulding. The fibres disintegrated for 30 min showed higher levels of fractures and disruptions than the other samples. The increased fibre damage together with a reduced fibre length (Fig. 12) is unexplained but may have been due unnoticed issues during injection moulding. The disintegration process used here to introduce fibre damage was used to minimise length reduction, fines generation and external fibre wall fibrillation. Microcompressions, nodes and cell wall delaminations follow similar trends as both delaminations and nodes are a consequence of microcompressions [8]. The study successfully minimised fibre shortening and consequently the factors which led to fibre shortening, fibre fractures and disruptions, do not change significantly across the three treatments. The observation that a large increase in microcompressions, nodes and wall

delaminations resulted in only a relatively minor effect on composite properties is supported by work by Eder et al. [37]. In their study, damage was approximately comparable to the current study and was induced in fibres by compression of wooden blocks. Mechanically isolated fibres containing different levels of microcompressions were subjected to tensile tests, resulting in more modest decreases in ultimate stress than anticipated. They concluded that the local mechanical distortion of cellulose fibrils may play only a minor role in the strength reduction known to occur in mill produced pulps [16]. Similarly, Thuvander et al. [38], measuring individual fibres, saw only very slight drops in the mean and variability of strain to failure data for refined kraft pulp fibres (which would be expected to show higher levels of fibre damage than the current study). They also noted there was no change in the intrinsic cellulose viscosity, and therefore no evidence for cellulose chain scission. Similarly, there is little evidence for significant changes in cellulose structure or crystallinity as a result of refining [39,40] or changes in pulp viscosity with the increased incidence of microcompressions [14]. Lennholm and Iversen [40] suggest that the small changes in NMR spectra they saw as a result of refining were due to changes in hydrogen bonding in response to the formation of microcompressions. According to Thuvander et al. cellulose chain scission is not required for the decrease of fibre strength. Instead they suggest that the minor changes they saw in fibre strength was due to interfibrillar debonding resulting in reduced interfibrillar stress transfer capability leading to failure characterised by interfibrillar shear damage. It is possible the increase of microcompressions after injection moulding (Fig. 13) was a result of the injection moulding process incrementally damaging fibres in a consistent manner. However, this is contradicted by the treatment disintegrated for 30 min which had significantly larger increases in fractures and disruptions (Fig. 13) and a corresponding reduction in fibre length (Fig. 12) probably caused by operating conditions of the injection moulder at the time. If the severity of the injection moulding process had an effect on microcompressions it would be expected that the 30 min sample would show an increased incidence relative to the other samples. As this did not occur it seems unlikely that injection moulding had a significant influence on the presence of the microcompressions. Instead, the higher incidence of microcompressions after injection moulding may be because they were more visible, as the dry and contracted fibre wall was, apparently, not able to be rehydrated.

4. Conclusion In general the effects of fibre length and fibre damage on mechanical performance where not as significant as was expected. Fibre length, if independent of any other fibre factors, indeed only has a substantial influence on performance when it drops below a critical length. This is consistent with composites theory which

52

A. Thumm, A.R. Dickson / Composites: Part A 46 (2013) 45–52

states that once a reinforcement fibre is over the critical length it can be considered as infinite in regard to composite strength. The ‘‘critical’’ fibre length which was estimated by modelling to be around 0.8 mm for P. radiata is difficult to achieve with compounding by an extruder which tends to cut fibres substantially shorter. Given the drop in performance that occurs when reinforcement fibres drop below this ‘‘critical’’ fibre length, gentle compounding methods should show significant benefits when wood pulp fibres are used for reinforcement. Fibre length only increased impact resistance for the longest fibre length pulp. Creating and maintaining an average fibre length of about 3 mm in a composite using pine fibres is, however, a challenging task, especially when fibres and matrix are compounded in an extruder. It is likely that fibre length alone will not be a significant factor in the improvement of impact strength compared to other opportunities such as fibre or matrix modification. The controlled application of fibre damage had little effect on composite performance. Processing that induces fibre wall delamination and microcompressions without significant scission of cellulose fibrils seems to have only a minor affect on fibre strength due to reduced interfibrillar stress transfer capability. Where fibre damage does have a major affect on composite strength it is most likely due to the formation of fractures and disruptions in the fibre wall that will also bring about a reduction in fibre length during processing. Acknowledgments The TMP pulp was supplied by Karl Murton. The kraft pulp was supplied by Steven Chapman and John Lloyd. Most of the laboratory work including pulp processing, pulp pad making and measurement and analysis was performed by Donna Smith and Maxine Smith. The injection moulding was performed by Daniel Parker. This work was supported through Scion’s core funding. References [1] Corson SR. Aspects of mechanical-pulp fiber separation and development in a disk refiner. Pap Puu-Pap Tim 1989;71(7):801–14. [2] Garkhail SK, Heijenrath RWH, Peijs T. Mechanical properties of natural-fibre-mat-reinforced thermoplastics based on flax fibres and polypropylene. Appl Compos Mater 2000;7(5–6):351–72. [3] Laws N, McLaughlin R. The effect of fibre length on the overall moduli of composite materials. J Mech Phys Solids 1979;27(1):1–13. [4] Bowyer WH, Bader MG. On the re-inforcement of thermoplastics by imperfectly aligned discontinous fibres. J Mater Sci 1972;7:1315–21. [5] Van Den Oever MJA, Bos HL. Critical fibre length and apparent interfacial shear strength of single flax fibre polypropylene composites. Adv Compos Lett 1998;7(3):81–5. [6] Rodriguez M, Rodriguez A, Bayer RJ, Vilaseca F, Girones J, Mutje P. Determination of corn stalk fibers’ strength through modeling of the mechanical properties of its composites. Bioresources 2010;5(4):2535–46. [7] Keith CT, Côté WA. Microscopic characterization of slip lines and compression failures in wood cell walls. Forest Prod J 1968;18(3):67–74. [8] Page DH, Seth RS, Jordan BD, Barbe MC. Curl, crimps, kinks and microcompressions in pulp fibres – their origin, measurement and significance. Fundamentals of papermaking: transactions of the eigth fundamental research symposium held at Oxford; 1985, p. 183–227. [9] Hartler N. Aspects on curled and microcompressed fibers. Nord Pulp Pap Res J 1995;10(1):4–7. [10] Nyholm K, Ander P, Bardage S, Daniel G. Dislocations in pulp fibres – their origin, characteristics and importance – a review. Nord Pulp Pap Res J 2001;16(4):376–84. [11] Hughes M. Defects in natural fibres: their origin, characteristics and implications for natural fibre-reinforced composites. J Mater Sci 2012;47(2): 599–609.

[12] Hartler N, Nyrén J. Misaligned zones in cellulosic fibres. Part 3: their influence on the fibre stiffness. Svensk Papperstidning 1968;71(21):788–9. [13] Page DH, El-Hosseiny F. The mechanical properties of single wood pulp fibres. Part 4. The infulence of defects. Svensk Papperstidning 1976;79(14):471–4. [14] Ellis MJ, Allison RW, Kibblewhite RP. The effect of laboratory mixing at medium pulp concentration on radiata pine kraft pulp. Appita J 1995;48(5): 358–62. [15] Ellis MJ, Duffy GG, Allison RW, Kibblewhite RP. Fibre deformation during medium consistency mixing: role of residence time and impeller geometry. Appita J 1998;51(1):29–34. [16] MacLeod JM. New improved kraft pulp quality. Pap Puu-Pap Tim 1990;72(8): 780–7. [17] Migneault S, Koubaa A, Erchiqui F, Chaala A, Englund K, Krause C, et al. Effect of fiber length on processing and properties of extruded wood-fiber/HDPE composites. J Appl Polym Sci 2008;110(2):1085–92. [18] Peltola H, Laatikainen E, Jetsu P. Effects of physical treatment of wood fibres on fibre morphology and biocomposite properties. Plast Rubber Compos 2011; 40(2):86–92. [19] Farshid B, Safdari V, Nourbakhsh A, Pilla S. The effects of fiber length and fiber loading on the mechanical properties of wood-plastic (polypropylene) composites. Turk J Agric For 2010;34(3):191–6. [20] Kang T, Dickson A, Paulapuro H. The response of the various TMP fractions to refining. Nord Pulp Pap Res J 2008;23(2):224–30. [21] Beg MDH, Pickering KL. Mechanical performance of Kraft fibre reinforced polypropylene composites: Influence of fibre length, fibre beating and hygrothermal ageing. Compos Part A-Appl S 2008;39(11):1748–55. [22] Iannace S, Ali R, Nicolais L. Effect of processing conditions on dimensions of sisal fibers in thermoplastic biodegradable composites. J Appl Polym Sci 2001;79(6):1084–91. [23] Bouafif H, Koubaa A, Perre P, Cloutier A. Effects of composite processing methods on wood particle development and length distribution: consequences on mechanical properties of wood-thermoplastic composites. Wood Fiber Sci 2010;42(1):62–70. [24] Quijano-Solis C, Yan N, Zhang SY. Effect of mixing conditions and initial fiber morphology on fiber dimensions after processing. Compos Part A-Appl S 2009;40(4):351–8. [25] Guo Q, Cheng B, Kortschot M, Sain M, Knudson R, Deng J, et al. Performance of long Canadian natural fibers as reinforcements in polymers. J Reinf Plast Compos 2010;29(21):3197–207. [26] Duc AL, Vergnes B, Budtova T. Polypropylene/natural fibres composites: analysis of fibre dimensions after compounding and observations of fibre rupture by rheo-optics. Compos Part A-Appl S 2011;42(11):1727–37. [27] Rask M, Madsen B, Sørensen BF, Fife JL, Martyniuk K, Lauridsen EM. In situ observations of microscale damage evolution in unidirectional natural fibre composites. Compos Part A-Appl S 2012;43(10):1639–49. [28] Barkoula NM, Garkhail SK, Peijs T. Effect of compounding and injection molding on the mechanical properties of flax fiber polypropylene composites. J Reinf Plast Compos 2010;29(9):1366–85. [29] Bos HL, Müssig J, van den Oever MJA. Mechanical properties of short-flax-fibre reinforced compounds. Compos Part A-Appl S 2006;37(10):1591–604. [30] He J, Batchelor WJ, Johnston RE. A microscopic study of fibre–fibre contacts. Appita J 2004;57(4):292–8. [31] Kocurek MJ, Malcolm EW, Grace TM. Alkaline pulping. In: Kocurek MJ, editor. Pulp and paper manufacture, 3rd ed., vol. 5. The Joint Textbook Committee of the Paper Industry; 1989. p. 637. [32] Kibblewhite RP. Effects of pulp drying and refining on softwood fibre wall structural organisations. In: Baker C, Punton VW, editors. Fundamentals of papermaking. London: Mechanical Engineering Publications Ltd.; 1989. p. 121–52. [33] Weise U, Paulapuro H. Changes of fibre dimensions during drying. International paper physics conference; 1995, p. 121–4. [34] Leopold B. Effect of pulp processing on individual fiber strength. Tappi J 1966;49(7):315–8. [35] Duncker B, Nordman L. Determination of the strength of single fibres. Pap PuuPap Tim 1965;10:539–52. [36] Mark RE, Thorpe JL, Angello AJ, Perkins RW, Gillis PP. Twisting energy of holocellulose fibers. J Polym Sci Pol Sym 1971;36(1):177–95. [37] Eder M, Terziev N, Daniel G, Burgert I. The effect of (induced) dislocations on the tensile properties of individual Norway spruce fibres. Holzforschung 2008;62(1):77–81. [38] Thuvander F, Gamstedt EK, Ahlgren P. Distribution of strain to failure of single wood pulp fibres. Nord Pulp Pap Res J 2001;16(1):46–56. [39] Liitiä T, Maunu SL, Hortling B. Solid state NMR studies on cellulose crystallinity in fines and bulk fibres separated from refined kraft pulp. Holzforschung 2000;54(6):618–24. [40] Lennholm H, Iversen T. The effects of laboratory beating on cellulose structure. Nord Pulp Pap Res J 1995;10(2):104–9.