Recycled carbon fibre-reinforced polypropylene thermoplastic composites

Recycled carbon fibre-reinforced polypropylene thermoplastic composites

Composites: Part A 43 (2012) 79–86 Contents lists available at SciVerse ScienceDirect Composites: Part A journal homepage:

2MB Sizes 41 Downloads 471 Views

Composites: Part A 43 (2012) 79–86

Contents lists available at SciVerse ScienceDirect

Composites: Part A journal homepage:

Recycled carbon fibre-reinforced polypropylene thermoplastic composites M.H. Akonda a,⇑, C.A. Lawrence b, B.M. Weager c a

Tilsatec Advanced Materials, Tilsatec Ltd., Wakefield, UK Centre for Technical Textiles, University of Leeds, Leeds, UK c NetComposites Ltd., Broom Business Park, Chesterfield, UK b

a r t i c l e

i n f o

Article history: Received 15 January 2011 Received in revised form 7 September 2011 Accepted 10 September 2011 Available online 16 September 2011 Keywords: A. Carbon fibres A. Recycling A. Yarn B. Mechanical properties

a b s t r a c t Comingled carbon fibre (CF)/polypropylene (PP) yarns were produced from chopped recycled carbon fibres (reCF) (20 mm in length, 7–8 lm diameter) blended with matrix polypropylene staple fibres (60 mm in length, 28 lm diameter) using a modified carding and wrap spinning process. Microscopic analysis showed that more than 90% of the reCF were aligned along the yarn axis. Thermoplastic composite test specimens fabricated from the wrap-spun yarns had 15–27.7% reCF volume content. Similar to the yarn, greater than 90% of the reCF comprising each composite sample made, showed a parallel alignment with the axis of the test specimens. The average values obtained for tensile, and flexural strengths were 160 MPa and 154 MPa, respectively for composite specimens containing 27.7% reCF by volume. It was concluded that with such mechanical properties, thermoplastic composites made from recycled CF could be used as low cost materials for many non-structural applications. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Carbon fibre (CF) is one of the most widely used reinforcing fibres for thermoplastic and thermoset composite materials, owing to its specific strength, specific modulus and thermo-physical properties. Since its commercial introduction in late 1969 for aerospace applications, the volume and end use of this fibre have grown, and continue to grow year-on year [1]. For example, in 2004, an estimated 22,000 metric tons of CF were consumed worldwide, about 8400 tons (about 18.5 million pounds) in the United States alone [2]. The worldwide carbon fibre consumption almost doubled in 2010 to approximately 40,000 tons [3] and it is estimated that the total global demand will be approximately 65,000 tons by 2014 [4]. As the worldwide volume of CF composite usage grows, there is concern about the potential tonnage of waste from manufacturing processes and end-of-life products. The waste related to CF products will quickly reach a significant level to become an important environmental issue since such products are not biodegradable. Manufacturers will therefore need to identify ways to comply with legislation on sustainability [5]. In addition, the cost of virgin CF fibres is high (£20–40/kg, depending on grade), reflecting the energy consumption in its manufacture [6,7], which means recovery and recycling of waste offer sound economic benefits. Consequently, there is a strong interest in developing processes for recovering and recycling CF from waste materials. Several methods have been developed for CF recovery from CF reinforced plastic (CFRP) waste. ⇑ Corresponding author. Tel.: +44 (0) 1924 375742; fax: +44 (0) 1924 376204. E-mail address: [email protected] (M.H. Akonda). 1359-835X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.compositesa.2011.09.014

These methods include catalytic conversion [8] at low temperatures (i.e. 200 °C); a ‘‘continuous process [9–12] in which the scrapped materials are shredded and fed into a fluidised gas chamber at 550 °C wherein the matrix undergoes thermal oxidative decomposition releasing the CF to be collected; the use of microwave heating [13] and supercritical alcohols [14] have also been reported. Published results [12,15] indicate that some of these recovery processes leave the mechanical properties of the recovered CF relatively unaffected. Reportedly the thermal recovery processes did not significantly change the oxygen content but the functional groups [email protected] (7.82% for recovered CF, 6.24% for unsized virgin) and COOH (5.27% for recovered CF, 3.79% for unsized virgin) were higher [16]. Despite the increase in oxygenated functional groups, the interfacial bonding strength with epoxy was unchanged [16]. Therefore the recovered fibres were considered reusable for composites manufacturing mainly as short fibres utilised as milled fillers. Lately, attention has focused on the recovery and application of longer fibres. However, no studies have yet been reported on converting the longer recovered CFs into a useable roving or yarn structure for downstream manufacturing. The work reported here describes the processing of recovered CF blended with thermoplastic staple fibres into yarns that were then used in thermoplastic composite fabrication. 2. Materials 2.1. Recycled carbon fibres The recovered CF used was supplied by Recycled Carbon Fibre Ltd., UK. Long lengths of fibres, >500 mm (Fig. 1) were recovered


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

from scrap TR50S carbon fibre–epoxy composite which was heated treated at 500 °C for 10 min in air at atmospheric pressure to completely burn off the epoxy matrix. The fibres were then cut to 50– 55 mm to be processed. For the sake of convenience, the material will be referred to, from here onwards, as recycled carbon fibre (reCF). 2.2. Matrix fibres Polypropylene fibres (PP) were selected as the matrix to be processed with the reCF, because this polymer is widely used as a low cost matrix for thermoplastic composites. The PP matrix fibre was manufactured and supplied by Drake Extrusion (UK) in staple form (60 mm length; 3.3 dtex fineness). 3. Experimental work 3.1. Characterisation of reCF The recovery process in air at high temperatures (500 °C) may affect the fibre’s physical properties, especially the strength [13]. Therefore, single fibre tensile tests were carried out in accordance with BS ISO 11566 (1996). The gauge length was 25 mm, the crosshead speed 1 mm/min and the fibre was loaded until failure. The average value of 20 single fibres taken randomly from the chop-material provided was determined. A Scanning Electron Microscope (SEM) was used to observe the 20 fibre surfaces and geometries. The SEM micrographs obtained showed the fibres to

have cylindrical shapes and therefore the micrographs were used to determine the diameters of the fibres. The average cross sectional area and denier of the fibres were calculated using by the following equations: [17].

hX i 2 ðpd Þ=4 =20

ð1Þ 2

Denier ¼ q  A ¼ 1:75  ðpd =4Þ  9  103 g=9000



The density (q) of the CF fibre was 1.75 g/cm . This is the density value reported in the fibre’s technical data sheet. 3.2. Production of reCF/PP wrap-spun yarn blends The chopped reCF was blended with the PP fibres (see Fig. 2a) to 30:70 and 50:50 weight ratios by weight. Stretch-breaking has reportedly been used as a preparatory method [18,19] for producing carbon fibre/matrix fibre blends for wrap spun yarn production. However, to date this process requires continuous filament (virgin) carbon fibre tow. The blends of reCF/PP staple fibres were produced using a carding process [20] to obtain a continuous sliver (Fig. 2b), which was subsequently drafted [21] and wrap spun [22] to convert the sliver into a yarn (see Fig. 2c). The intermingling of the staple reCFs with the matrix fibres was achieved by modifying a conventional card fitted with stationary flats [20], in particular, components such as the feed roller, feed plate, and the cylinder, flats and doffer wire-clothing had to be modified for the reCF/PP blends to be suitably processed, details of which are published

Fig. 1. (a) Recycled carbon fibres and (b) individual recycled carbon fibres.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. (a) reCF/PP blends (42:58 by weight), (b) sliver and (c) yarn made from blended fibres. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

Melting Consolidation

Solidification 20

15 min

5 mm

Pressure, bar



Time, min



Fig. 3. (a) Yarn winding arrangement and (b) process cycle for composite moulding. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

elsewhere [23]. [The modified card was used to produce lightweight slivers of 3 g/m linear density (3 ktex) (CV% = 12) with an average breaking load of 1 N, tenacity of 0.83 N/ktex. During the process of intermingling (termed blending) the reCF with PP matrix fibres on the modified card, it was found that the crimped PP fibres acted as carriers for the un-crimped reCF. This provided the required fibre-to-fibre cohesion which helped individual separation of the reCFs (termed opening) with minimum fibre breakage and good blending. After carding, the blended reCF/PP sliver produced was drafted in order to parallelize the CF fibres and achieve fibre mass regularity along the sliver and subsequently the yarn length. The degree of intermingling of the fibres was assessed by an image analysis technique using the software Image Pro+[24]. This involved analysing photomicrographs of samples (80 mm  80 mm) of the carded web removed prior to consolidation into a sliver. Wrap-spun yarn samples Yarn 1 and Yarn 2, each of 1000 ± 100 tex linear density (g/1000 m yarn), were respectively produced from the 30:70 and 50:50 drawn-sliver blends using a wrapping levels of 80 turns per metre.

3.3. Composite fabrication and mechanical testing The reCF/PP yarns produced were wound onto a steel frame unidirectionally (Fig. 3a), and the layers were subsequently compressed and consolidated using hot compression moulding plates at 220 °C, under 20 bars pressure for 15 min (Fig. 3b). Composite specimens were also made from the card sliver. Lengths of slivers were placed into a 20 mm width channel and was pressed under the above conditions. At the end of the dwell time, the composite panels were cooled at a rate of 15 °C/min to room temperature (20 °C). The UD composite samples were cut into 250 mm  25 mm test pieces for tensile measurements (in accordance with BS EN ISO 527-1:1996) and 15 mm  85 mm (68 mm span length) for flexural properties (following the standard ISO 14125).

3.4. Fibre analysis of sliver yarn and composite sheets Representative specimens taken from the sliver, yarn and composite test pieces were heated to 500 °C in an oven under a nitrogen atmosphere for 30 min to burn off the PP matrix completely from the specimens in accordance with ASTM D2584-68. Once the resin was completely removed, the weight of the residue was measured and the mass loss calculated as the resin content of the specimens [25].

The fibre volume percentages were determined from the known densities for reCF and PP, 1.75 g/cm3 and 0.90 g/cm3 respectively, and the calculated % weight of the matrix fibres (resin matrix). The mean fibre lengths and length distributions were measured by using the Baer diagram/Fibrogram technique [20,21] for sample sizes of 200 fibres from the specimens. To determine fibre orientation, optical microscopic observations (reflective mode) were made of polished cross-sections taken at right angles to the axes of the sliver and yarn samples encapsulated in epoxy resin, and of the respective composite samples. The CF fibres that met the crosssectional plane have either a circular or elliptical foot print; the former if aligned with the considered axis, the latter if inclined to the axis. The orientation of each fibre to the considered axis was, therefore, calculate from its ellipticity [26,27].

4. Results and discussion Microscopic measurements showed the average fibre diameter for reCF was slightly (5.3%) higher than that of the virgin CF (see Table 1). The SEM images of the reCF surfaces showed the fibres to be free of carbonaceous deposits and smooth textured. No visible surface flaws were apparent on the fibres (Fig. 4). The load– extension curves obtained from the tensile testing of 20 single reCF and virgin fibres are presented in Fig. 5 and a summary of the test data listed in Table 1. The results showed that the reCF had greater than 85% of the virgin fibre strength and was only 7% lower in modulus. Other properties such as linear density, fibre diameter and breaking extension were also comparable to the virgin carbon fibre. Fig. 6a shows as an example the histogram obtained from the image analysis of the 50:50 (reCF/PP) blend sample. It can be seen that the maximum intensity of greyness was at approximately the 120 (pixel level) of the grey scale with a narrow peak width, indicating good blend uniformity. Absolute black and absolute white equate to 0 and 200 pixel levels respectively, while 130 represents

Table 1 Comparison study of tensile properties of single virgin and recycled carbon fibres (average values for 20 single fibres). Properties

Virgin fibre (TR50S)

Recycled fibres (TR50S)

Fibre diameter (lm) Linear density (dtex) Tensile strength (GPa) Tensile modulus (GPa) Elongation at break (%)

7.60 ± 0.01 0.75 ± 0.07 3.19 ± 0.56 242.0 ± 5.4 1.10 ± 0.22

8.0 ± 0.01 0.79 ± 0.07 2.86 ± 0.39 225.0 ± 7.5 1.09 ± 0.17


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

Fig. 4. SEM images of fibre surface (a) recycled and (b) virgin carbon fibre.

a uniform grey [28]). Wider distributions, with a peak positioned near to the zero pixel level, were obtained for poorly blended samples produced prior to the card modifications; these samples had many carbon fibre rich zones (Fig. 6b). The burn-off test for the slivers and yarns produced from the 30:70 and 50:50 blends resulted in a reCF content of 25% and 42% by weight. The reduction in reCF content may be attributed to the loss of the very short lengths of broken fibres which occurred during the carding process, resulting in a reduction in the average fibre length; from 55 mm to 23 ± 3 mm in the 30:70 and to 17 ± 3 mm in the 50:50 blended yarn samples, respectively (see Fig. 7). The larger decrease in the mean reCF length for the 50:50 blend, indicates that higher reCF content would result in greater shortening of the fibre during carding, owing to the need for more intensive opening action. The diameter (7–8 lm) and the linear density (0.7 dtex) of the reCF were nearly four times smaller than those of PP fibres (diameter 28 ± 0.5 lm, linear density 3.3 dtex). The total number of reCF in the 50/50 blend would therefore be four times higher than that of PP fibres. This means that there would have been insufficient PP fibres to reduce the impact of the carding action and thereby the breaking the reCF.

The slivers, 42% reCF and 58% PP by weight had a low breaking strength (breaking load 1 N) (see Fig. 7). Thin compress-moulded composite laminate sheets were made from this sliver (thickness 0.35 mm / width 16 mm) and Fig. 8 shows the tensile characteristics obtained from the sliver composites. The average breaking load and tensile strength of the sliver composite were determined from five test specimens and found to be 550 ± 10 N and 98 MPa, respectively. (Fig. 9). No further reduction in fibre length occurred during the making of the composite test specimens with the spun yarns. The measured reCF contents of these specimens were as follows:  Composite C1 (1.15 mm thick) made from Yarn 1 (30:70, reCF/ PP blends) contained 15% reCF by volume (25% by weight) (see Fig. 10).  Composite C2 (1.4 mm thick) made from Yarn 2 (50:50, reCF/PP blends), contained 27.7% reCF by volume (42% by weight) (see Fig. 10). The slightly lower % of reCF in both composite test pieces was attributed to process losses.

Force–extention curves of recycled carbon fibres

Force–extention curves of virgin carbon fibres



Fig. 5. Load–extension curves of single (a) recycled and (b) virgin carbon fibres.


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

Fig. 6. (a) Ideal blend of 42:58 (by weight) carbon / PP fibres and histogram obtained from ideal blend sample, (b) poor blends sample and histogram of the poor blended sample. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Load, N

%, relative frequency

700 600

Sample 1


Sample 7

Sample 4 Sample 8


Sample 10

300 200 100 0



Class length, mm








Extension, mm

Fig. 7. Histograms of reCF length distribution in 25:75 (30:70 actual blends) reCF:PP and 42:58 (50:50 actual blends) reCF:PP (by weight) yarn specimens.

Fig. 9. Load–extension curves obtained from individual reCF/PP (42/58 by weight) consolidated slivers.

50 By weight %


Carbon fibre, %


Load, N


By volume %

30 20 10

0.4 0 Yarn 1 0

Yarn 2

Composite C1

Composite C2

Fig. 10. % of reCF in yarns and composite specimens.







Extension, mm Fig. 8. Load–extension curves obtained from individual reCF/PP (42:58 by weight) sliver.

SEM observations of the composite panels made from the wrap yarns showed no trace of the wrap-yarn structure. Both the constituent PP staple fibres and the PP filament wrapper that held together the yarn structure, melted during the hot-pressing


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

majority of the reCF are parallel and aligned with the yarn axis and this is reflected in the related composite produced. Fig. 14 shows the results of measured relative alignment of the fibres in the yarn and depicts that 90% of fibres showed good alignment. This may be a possible reason for the higher tensile values obtained with the yarn-based composite. However, SEM studies of the fracture zones (Fig. 13b) of all the composites samples showed failure occurred largely by fibre pull-out from the PP matrix, indicating weak fibre/matrix interfacial bonding. This may due to the surface characteristic of the PP matrix which has non-polar chemical bonds and therefore would be non-reactive with the smooth surface of reCF [29].





Ten. modulus (GPa)

Ten. strength (MPa)

moulding, and the reCF had spread out uniformly in the molten polymer matrix. Owing to the high degree of blending of the fibre components, voids were not evident in the composites. The composite specimen C1 had tensile and flexural strengths (Fig. 11) of 105 MPa and 100 MPa, respectively, whereas, the C2 composite specimen had 50% higher values (tensile strength of 160 MPa and flexural strength of 154 MPa). The flexural modulus of the C2 composite was 31% higher than C1 (19 GPa for C1 and 25 GPa for C2). The higher value compared with C1 may be attributed to the 80% higher reCF volume fraction. Figs. 12 and 13a show SEM micrographs of the alignment of the reCF in the wrap-spun yarn and ultimately in the composite C2, and it would appear that the

105 70 35 0

21 14 7 0





Flex. strength (MPa)

Flex. modulus (GPa)





Fig. 11. (a) Tensile strength, (b) tensile modulus, (c) flexural strength and (d) flexural modulus of the unidirectional reCF-PP composites.

Fig. 12. SEM images of (a) reCF in yarn and (b) reCF in burnt off yarn sample.


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

Fig. 13. SEM images of (a) reCF alignment in composite and (b) reCF pulled out from PP matrix.


Fibre ratio, %





0 Machine direction 10 degrees from 20 degrees from (90 degrees) machine direction machine direction

30 degrees from machine rection

50 degrees from machine direction

Angle Fig. 14. reCF fibre orientation distribution in yarn (42:58 reCF:PP by weight) sample.

5. Conclusion A process for converting recycled carbon fibre into a staplespun yarn blend with polypropylene fibres has been described. These yarns were used to produce thermoplastic composite test specimens. The specimens showed acceptable mechanical properties attributed to good alignment of the carbon fibre with the precursor yarn axis. It may be concluded that with such mechanical properties, thermoplastic composites made from recycled CF could be used as low cost materials for many non-structural applications. Also, it would seem reasonable to assume that the conversion process could be used to intermingle recycled carbon fibre with other thermoplastic matrix fibres such as PET, PA, PPS and PEEK.

Acknowledgements The work presented in this paper was carried out at University Leeds. This research work was co-funded by Technology Strategy Board’s Collaborative Research and Development Programme (Fibre Cycle; Grant Number GR/R80605/01) and six industrial partners; Tilsatec Ltd., Advanced Composites Group Ltd. (UK), Recycled

Carbon Fibre Ltd., NetComposites, Exel Composites (UK) and Sigmatex (UK).

References [1] Hajduk F, Carbon fibres overview; global outlook for carbon fibres 2005. Intertech conferences. San Diego: CA; October 11–13, 2005. [2] Market report on world carbon fibre composite. Germany: Acmite Market Intelligence; July, 2010. [3] Masayoshi K. Toray carbon fibre composite materials businesses. Toray IR Seminar No. 7. June 6, 2005. [4] Roberts A. Carbon fibre & precursor outlook, 2008–2015. Carbon fibre 2007 conference. Washington: D.C.; December 4–7, 2007. [5] European directive 2000/53/EC. On end of life vehicles. [6] Pickering SJ, Kelly R, Kennerley JR, Rudd CD, Fenwick NJ. A fluidised–bed process for the recovery of glass fibres from scrap thermoset composite. Compos Sci Technol 2000;60(4):509–23. [7] Davidson J. Carbon fibre composites: advancements in reclamation processes and recycled material forms. In: Composites innovation 2007 conference proceedings. Spain: Barcelona; October 4–5, 2007. [8] Alldred RE, Coons AB, Simonson RJ. Properties of carbon reclaimed from composite manufacturing scrap by tertiary recycling. In: Int Proc SAMPE 28th Int Tech Conf Seattle, 1996, SAMPE: 1996. [9] Pickering SJ. Recycling of carbon fiber composites using a fluidised bed process. In: 8th International conference on fiber reinforced composites (FRC200). p 565–72.


M.H. Akonda et al. / Composites: Part A 43 (2012) 79–86

[10] Pickering SJ, Yip H, Kennerley JR, Kelly R, Rudd CD. The recycling of carbon fibre composites using a fluidised bed process. In: FRC 2000, Proceedings of the international conference on fibre reinforced composites. Newcastle upon Tyne: UK; 8th Sept 13–15, 2000. [11] Jiang G, Wong WH, Pickering SJ, Rudd CD, Walker GS. Study of a fluidised bed process for recycling carbon fibre from polymer composite. In: Proceedings of 7th world congress for chemical engineering. Glasgow: UK; 2005. [12] Yip H, Pickering SJ, Rudd CD. Characterization of carbon fibres recycled from scrap composites using fludised bed process. Plast Rubber Compos 2002;31(6):278–82. [13] Peng JCM, Donnet JP, Wang TK, editors. Surface treatment of carbon fibres, 3rd ed. In: Donnet J-P et al., editors. Carbon Fibres, New York: Marcel Dekker; 1998. p. 161–229. [14] Hyde, Jason R, et al. Supercritical propanol, a possible route to composite carbon fibre recovery: a viability study. Compos Part A: Appl Sci Manufac 2006;37(11):2171–5. [15] Connor M, Allen B, Heil J. Recycled carbon fiber analysis: mechanical properties. In: SAMPE Fall technical conference proceedings: global advances in materials and process engineering. Dallas: TX; November 6–9, 2006. [16] Jiang G, Pickering SJ, Walker GS, Wong KH, Rudd CD. Surface characterisation of carbon fibre recycled using fluidised bed. Appl Surf Sci 2007. [17] Allen BE. Characterization of reclaimed carbon fibres and their integration into new thermoset polymer matrixes via existing composite fabrication techniques. MSc Thesis, Materials Science and Engineering, USA: North Carolina State University; 2008. [18] US Patent No. 4856147. Composites of stretch broken aligned fibres of carbon and glass reinforced resin. 1989.

[19] Gregory PD, Daniel W. Comparison of continuous fibre and stretch broken carbon fibre (SBCF) materials in forming processes. In: The international technical conference proceedings of the society for the advancement of material and process engineering (SAMPE). USA: Long Beach; 2005. [20] Lawrence CA. Fundamentals of spun yarn technology. Florida: CRC press; 2003. [21] Lord PR. Handbook of yarn production: technology, science and economics. CRC press; 2003. [22] Palmer DW. Wrap Spinning. Carpet & Rug Industry. 15, Oct., 1983. [23] UK Patent Application. Application No. GB 1001868.7, UK: February, 2010. [24] Kim KW, Lee KH, Lee BS, Ho HS, Oh SJ, Kim HY. Effects of Blend Ratio and Heat Treatment on the Properties of the Electrospun Poly(ethylene terephthalate) Nonwovens. Fibres Polym 2005;6. [25] Abdalla FH, Megat MH, Sapuan MS, Sahari BB. Determination of volume fraction values of filament wound glass and carbon fibre reinforced composites. ARPN J Eng Appl Sci 2008;3(4). [26] Hine PJ, Davidson N, et al. Hydrostatically extruded glass-fiber reinforced polyoxymethylene. I: The development of fibre and matrix orientation. Polym Compos 1996;17:720–9. [27] Hine PJ, Duckett RA, Davidson N, et al. Modelling of the elastic properties of fibre reinforced composites.1. orientation measurement. Comp Sci Tech 1993;47:65–73. [28] Ng MCF, Zhou J. Innovative layered-combination mode for digital jacquard fabric design. Text Res J 2009;79:737. [29] Bunsel AR, Renard J. Fundamentals of fibre-reinforced composite materials. In: Cantor B, Goringe MJ, Ma E, editors. Series in Materials Science and Engineering, London: Institute of Physics publishing; 2005.