RECYCLING OF END-OF-LIFE THERMOPLASTIC COMPOSITE BOATS

M.E. Otheguya*, A.G. Gibsona, M. Robinsona, E. Findonb, B. Crippsb, A. Ochoa Mendozac, M.T. Aguinaco Castroc

a NewRail Research Centre, School of Mechanical and Systems Engineering, Newcastle University, England, NE1 7RU, UK

b Small Boat Centre of Excellence, BVT Surface Fleet Portsmouth Ltd, 3/236 PP112, Military Road, HM Naval Base, Portsmouth, England, PO1 3NH, UK

c E.U.I.T.Industriales, UPM, Polytechnic University of Madrid, Ronda de Valencia, 3, 28012 Madrid, Spain

* Corresponding author, m.e.otheguy@ncl.ac.uk SUMMARY This paper discusses the recycling of thermoplastic composite materials in the context of boatbuilding. Work was carried out on the recycling of an experimental thermoplastic composite rigid inflatable boat, originally built by BVT Surface Fleet and tested in service by the RNLI. It was found that a range of useful injection moulding materials could be prepared from the hull material of the craft, demonstrating that structural thermoplastic composites are recyclable in practice as well as in principle, and confirming that they are a sustainable alternative to thermosets.

1. INTRODUCTION This paper relates to recycling issues associated with a 8.5m experimental rigid inflatable boat (RIB), manufactured in thermoplastic (polypropylene-glass) composite to demonstrate the viability of thermoplastics technology in marine composites. An experimental Atlantic 85 RIB built in thermoplastic was manufactured by BVT Surface Fleet (VT Halmatic) in 2004 for the Royal National Lifeboat Institution (RNLI) and had been extensively service tested. BVT Surface Fleet has been at the forefront of composite boat building in the UK since 1952. Although BVT provides a new build capability, it is also committed to the provision of through-life support and disposal of naval ships and small boats. Successfully managing the life cycle of BVT products is of great importance: the recycling and disposal of composite structures and small craft is central to this philosophy. Composite materials have become dominant over five decades of evolution in military and leisure boatbuilding mainly due to their aesthetic characteristics, low weight and good mechanical properties. The market for boat hull glass fibres is more than 130,000 tpa and there is expanding use of carbon and aramid fibre-reinforced materials [1,2]. The boatbuilding field, however, is predominantly based on thermosetting resins, albeit with a movement away from open mould processes towards closed mould ones. The marine composites industry has recently attracted attention in terms of VOC emissions, carbon footprint and recyclability [3]. As a result, in the EU, it has become desirable to develop end-of-life strategies that avoid landfill (2004). This follows the example set in the automotive industry by the End of Life Vehicles Directive (ELVD, 2003), which requires 95% of the mass of each vehicle manufactured after January 2015 to be reused or recovered. These new rules have led many car manufacturers to redesign their vehicles to facilitate disassembly, as well as to careful materials selection to promote recycling and reuse.

2. COMPOSITE MATERIALS RECYCLING The existing technically viable recycling techniques for thermosetting composites comprise mechanical, thermal and solvent routes (Figure 1).

Figure 1. Recycling processes for thermoplastic and thermosetting composites. Mechanical recycling has been explored and applied in the recent past, by the ERCOM project and by Phoenix Fibreglass [4]. Laminates are ground, milled, then used as fillers in new composite products. Neither of these early initiatives achieved economic viability, due to the lack of suitable markets for product containing recycled material. There was also the economic problem that the cost of the recycled filler exceeded that of virgin fillers, such as calcium carbonate. Thermal processes aim at fibre, material and energy recovery through controlled combustion of the organic portion of the composite scrap. Combustion with energy recovery makes use of the calorific value of the resin for heating or energy applications. The key problem is that the glass fibres, being heated along with the resin, dilute the useful heat that can be extracted. For this reason, processes, such as cement manufacture have been investigated in which the hot ashes have some use in the product. Handling of hazardous combustion products is a significant problem. Fluidised bed processes for composite scrap operate at 400-700ºC and involve blowing air through a silica sand bed. The fibres are recovered by filtering the hot gas stream emerging from the bed. The gas can then be combusted in an afterburner to recover heat. Any metallic inserts in the composite can be recovered from the sand. The glass fibres recovered with this technique are short and show a strength loss of about 25%. Finally, composites can be heated in the absence of air, again resulting in short fibres, along with and gaseous and liquid decomposition products from the resin. Unfortunately, the current glass fibre price makes it unprofitable to recover fibres using these methods. By contrast, carbon fibre is claimed to be profitably recycled by companies such as Adherent Technologies Inc. (US) [5, 6] and Recycled Carbon Fibre Ltd. (former Milled Carbon, UK) with proprietary catalysed processes that use lower temperatures. Other recycling methods include chemical decomposition with solvents or supercritical fluids [7]. However, the chemicals used can constitute an environmental hazard and the resulting chemical mixtures require complex processing to recover useable products.

Thermoplastic resins offer the possibility of recyclability [8-10]. There has been considerable interest in their use in boat structures, primarily for reasons of toughness, as well as recyclability. With real end-of-life structures, such as the hull of a rigid inflatable boat there are several “contaminant” materials in addition to the composite material. For instance the hull laminate is painted with a primer-topcoat system, and the internal structure comprises a sandwich configuration including a balsa core. One of the aims of the present study was to examine the effect of these contaminants on the properties of the recycled composite. In the present study it was decided to manufacture mouldable thermoplastic granules by granulating the hull material and diluting it with additional thermoplastic resin to aid processability and bring the resin content in line with that of conventional injection moulding compounds that could be used to produce new parts. 3. EXPERIMENTAL 3.1 MATERIALS The materials employed in this study were laminates taken from the thermoplastic version of the Atlantic 85 RIB, Figure 2, and an additional set of purpose-made glass/PP laminates. The main structural material used in the boat was Twintex T PP 60 1485 woven polypropylene-glass commingled fabric, containing 60 wt.% glass in the form of 18µm diameter fibres. The hull surface was painted with an epoxy-based primer, PPG Industries NEXA Autocolor P572-212, followed by AWLGRIP, an acetate-based G-line Standard Marine topcoat. Treatment with a primer is usually necessary when coating thermoplastics because of their non-polar nature that inhibits adhesion.

Figure 2. RNLI rigid inflatable boat, Atlantic 85, in service. The boat internal structure was a sandwich construction of the Twintex fabric and DIAB Pro-balsa Standard core material, with a specific gravity of 0.155. The sandwich laminates employed various thicknesses and numbers of plies, depending on function and position. In addition, a set of

laminates was manufactured for recycling, containing exactly the same products (Batches 1-3). A further set of laminates (Batch 4) was manufactured with plain Twintex PP 60 to enable an assessment of the influence of dry paint (Batch 5) on the recycled hull material properties to be undertaken. It was desirable to decrease the fibre fraction to facilitate processing into injection moulding material, as manufacturers would not normally employ material containing more than about 40wt.% glass. A high melt flow index PP homopolymer would normally be chosen for impregnated wood and glass fibre products. In the present case, the material used was Sasol HTV145, which had a melt flow rate (MFR) value of 45g/10min. It should be noted that, although virgin PP was used here, appropriate PP recyclate is available. For wood and/or glass fibre impregnation the addition of a coupling agent to the thermoplastic melt is known to improve the fibre-resin adhesion and thus the final mechanical properties. In this case 2wt.% by of Polybond 3200 maleic anhydride modified PP was added. 3.2 RECYCLING ROUTE AND EQUIPMENT Panels were removed from the boat with a pneumatic saw and subsequently cut into 15 mm (approx.) squares, which were fed to, a Homa moulding Granulator Type 01, 4.1 kW. The resulting granulate was dry blended with virgin PP pellets, and fed into a single screw extruder, for compounding. There were 5 temperature-controlled zones, set at (from feed to die): 200ºC, 225ºC, 250ºC, 235ºC and 220ºC. The solidified extrudate was again passed through a granulator. The resulting granules were injection moulded into tensile bars, using a Sandretto HP 40 (40 ton) machine. The following optimum conditions were used: injection pressure: 300 bar; barrel temperatures: 220ºC, 230ºC, 240ºC and mould temperature: 65ºC. The resulting dog-bone samples complied with ISO 527-1:1993. Five batches of tensile samples were produced, the overall compositions and final specific gravity (SG), after moulding, being shown in Table 1. Batches 1-3 were all based on the granulated hull sandwich laminate. Batches 1 and 2 were diluted with virgin PP and Batch 3 contained laminate only. The main processing difficulty occurred due to the presence of the fluffy lightweight balsa component, which resulted in some feeding problems, which probably could have been overcome by using a twin-screw extruder with a specialised augur feed. There were also some initial wet-out problems with the balsa, which again could probably have been overcome by using more specialised compounding equipment. Batch 3, with the highest balsa and fibre content, was the most difficult to process. Despite the compounding difficulties the process resulted, in Batches 1-3, in granules that were acceptably injection mouldable.

Table 1. Recycled material batches, compositions and specific gravities.

Batch Composition, % in weight S.G. 1 87.6% PP, 9.4% glass, 3.0% balsa wood 0.970 2 53.0% PP, 35.5% glass, 11.4% balsa wood 1.201 3 33.2% PP, 50.5% glass, 16.2% balsa wood 1.390 4 60.3% PP, 39.7% glass 1.204 5 As 4, with 0.7% paint added 1.201

Batch 4, which contained no balsa, was granulated from pure laminated Twintex material (60wt.% glass) and designed to indicate the maximum achievable properties in recycling. Again, the glass content was reduced by adding virgin PP at the compounding stage, to achieve a final glass

content of 40wt.% in the moulding material. Batch 5 was the same with 0.03% of paint residue added, to ascertain whether this would have a deleterious effect. Both these batches processed well. 3.3 RESULTS AND DISCUSSION The specific gravities of Batches 1-3 after moulding were compared with the calculated values based on the initial SGs of the components, as shown in Table 2. The calculated value was made assuming an SG of 0.155 for the balsa, as given by the supplier. It can be seen that the actual SG values are all significantly higher than those calculated, the most probable reason being that the compounding and injection moulding processes had removed a significant proportion of the air initially present in the balsa. To test this hypothesis, the SGs were re-calculated, replacing the balsa SG with a higher value of 1.5, corresponding to the value for wood fibre [11]. This second estimate would correspond to the situation where most of the air in the balsa was removed during processing. These values can be seen to agree much more closely with the measured SGs. Comparison of the values suggests that approximately 65% of the air enclosed in the structure of balsa wood was probably eliminated in the compounding and moulding processes. Remaining porosity can be observed by Electron Microscopy, Figure 10.

Table 2. Calculated and measured specific gravity of Batches 1-3.

Batch S.G. (No air removed)

S.G. (Measured)

S.G. (All air removed)

1 0.834 0.970 0.977 2 0.684 1.201 1.249 3 0.620 1.390 1.485

The tensile strength of all the batches of recycled mouldings was measured and compared to the values expected for injection moulded samples of similar glass content [12, 13]. A number of interesting results were observed. Batches 4 and 5, based on pure Twintex diluted with PP, show the highest tensile strength values, albeit with a small reduction in the case of the paint-contaminated batch (Figure 3).

0

20

40

60

80

100

120

Tens

ile s

treng

th, M

Pa

12345

0%

1%

2%

3%

4%

Elon

gatio

n at

bre

ak, %

12345

Figure 3. Tensile strength of recycled injection mouldings

(bars show 95% confidence limits). Material batch numbers correspond to those in Table 1.

Figure 4. Elongation-at-break of recycled injection mouldings (bars show 95% confidence limits). Material

batch numbers correspond to those in Table 1.

Both these batches show properties comparable with those expected of conventional PP compounds of similar glass content. The reduction in the case of the paint contamination may be due to sharp edged, low strength paint particles. These particles (Figure 11) are probably not well-bonded to the polypropylene, acting as stress raisers.

0

2

4

6

8

10

12

14M

odul

us o

f ela

stic

ity, G

Pa

12345

0

20

40

60

80

100

120

140

160

0% 10% 20% 30% 40% 50% 60%

Glass content (% weight)

Tens

ile s

tren

gth,

MPa

Recycled, glass onlyRecycled, glass and paintRecycled, glass and woodVirgin, ComAlloy® 121,135,143,150Virgin, glass only (Thomason), LFVirgin, glass only (Thomason), SF

Figure 5. Modulus of elasticity of recycled injection mouldings (bars show 95% confidence limits). Material batch numbers correspond to Table 1

Figure 6. Tensile strength values for recycled laminates compared to virgin materials (Thomason from [13],

ComAlloy® from [12]). The batches containing the balsa show reduced, though still acceptable, properties. Although the glass content of Batch 4 (39.7wt.%) is lower than that of Batch 3 (50.5wt.%) the strength is greater, underlining the negative effect of balsa. Again this could be due to insufficient bonding with the PP matrix, or to degradation of some of the balsa components. As reported by Jakab et al., wood decomposition commences at around 200°C, with decomposition of lignin and hemicellulose, accompanied by the release of water and formaldehyde [14]. Thus, partly degraded wood particles could act as stress raisers, initiating failure. Most of the samples containing wood showed wood particles on fracture surface (Figures 9, 10). The tensile strengths were compared with values for standard glass/PP materials [12, 13]. Figure 6 shows that the strength values of the recycled samples are lower than those of ‘long fibre’ moulding materials. This is not surprising, since the latter materials, which are manufactured by a different method, have a fibre length of the order of 1mm, which is significantly larger than in the present case. However, the strengths of the recyclates can be seen to be comparable with those of the more widely used ‘short fibre’ materials, which are also produced by extrusion compounding, and which have an average fibre length of around 200 microns. These results place the properties of these mixed recycled materials well within the range that would be acceptable in the marketplace for short fibre moulding compounds.

0%

1%

2%

3%

4%

0% 10% 20% 30% 40% 50% 60%

Glass content (% weight)

Elon

gatio

n at

bre

ak, %

Recycled, glass onlyRecycled, glass and paintRecycled, glass and woodVirgin, ComAlloy® 143,150Virgin, glass only (Thomason), LFVirgin, glass only (Thomason), SF

0

2

4

6

8

10

12

14

0% 10% 20% 30% 40% 50% 60%

Glass content (% weight)

Mod

ulus

of E

last

icity

, GPa

Recycled, glass onlyRecycled, glass and paintRecycled, glass and woodVirgin, Borealis® 205,402,306Virgin, glass only (Thomason), LFVirgin, glass only (Thomason), SF

Figure 7. Elongation-to-break values for recycled laminates compared to virgin materials (Thomason

from [13], ComAlloy® from [12]).

Figure 8. Young’s Modulus values for recycled laminates compared to virgin materials (Thomason

from [13], Borealis® materials from [12]).

The values of elongation-to-break in Figure 4 show the expected decrease with increasing fibre content. Again the deleterious effect of the balsa can be seen. Nevertheless the elongation-to-break values are still in the region of commercial acceptability, as can be seen from Figure 7. Figures 5 and 8 show that, for all compounds, the Young’s modulus increases with glass content. Comparing the result for Batch 2 (~35wt.% glass, 11wt.% wood ) to that for Batch 4 (40% glass) it can be inferred wood particles do have a positive influence on modulus (Figure 5). By contrast, paint particles seem to have a negative effect, probably because their modulus is similar to that of PP.

Figure 9. SEM image of a Batch 3 sample fracture surface showing an approx. 2mm wide wood chip.

Figure 10. Wood chip detail, Batch 3 sample (see Figure 9).

Figure 11. Detail of a paint particle on fracture

surface, Batch 5 sample.

6. CONCLUSIONS The present research has demonstrated the recyclability of thermoplastic-based composites. In particular we have shown that the hull of a rigid inflatable boat, composed of glass/polypropylene laminate along with balsa core material and paint, can be recycled by melt processing into injection mouldable granules which have acceptable properties when processed. Although both balsa and paint have a deleterious effect on moulded strength and elongation-to-break the properties achievable in the compounded granules are well-within the region of commercial interest for reinforced polypropylene moulding materials.

These materials could be used either in non-appearance automotive applications, where talc and glass reinforced PP are currently used. Alternatively, they could be used in decking and wood imitation applications where wood-reinforced composites are currently being considered. The Young’s modulus of the recycled materials is comparable to that of conventional short fibre polypropylene moulding materials. The presence of balsa does have a small positive effect on modulus. 7. ACKNOWLEDGEMENTS We would like to acknowledge the support of the European Commission under the MOMENTUM Research Training Network, Contract No. MRTN-CT-2005-019198. 8. REFERENCES 1. MOURITZ, A.P., GELLERT, E., BURCHILL, P. and CHALLIS, K., 'Review of advanced

composite structures for naval ships and submarines', Composite Structures, 53, 1, 21-42, 2001

2. MCCONNELL, V.P., 'Composites in North America', Reinforced Plastics, 49, 11, 26-33, 2005 3. GRAMMAN, H. and KRAPP, R.V.B., 'Disposal and recycling of HSC materials', 6th

International Conference on High-Performance Marine Vehicles, Naples, 271 - 280, 2008 4. PICKERING, S.J., 'Recycling technologies for thermoset composite materials--current status',

Composites Part A: Applied Science and Manufacturing, 37, 8, 1206-1215, 2006 5. ALLRED, R.E. and BUSSELLE, L.D., 'Tertiary recycling of automotive plastics and

composites', J. of Thermoplastic Composite Materials, 13, 2, 92-101, March 1, 2000. 6. GIULVEZAN, G. and CARBERRY, B., 'Composite recycling and disposal, an environmental

R&D issue', Boeing Environmental Technotes, 8, 4, 2003. 7. POULAKIS, J.G. and PAPASPYRIDES, C.D., 'Recycling of polypropylene by the

dissolution/reprecipitation technique: I. A model study', Resources, Conservation and Recycling, 20, 1, 31-41, 1997

8. GIBSON, A.G., 'Continuous molding of thermoplastic composites', Comprehensive Composite Materials, Pergamon, Ed. A.Kelly and C. Zweben, Oxford, 979-998, 2000

9. GIBSON, A.G. and MÅNSON, J.A.E., 'Impregnation technology for thermoplastic matrix composites', Composites Manufacturing, 3, 4, 223-233, 1992

10. IJAZ, M., ROBINSON, M. and GIBSON, A.G., 'Cooling and crystallisation behaviour during vacuum-consolidation of commingled thermoplastic composites', Composites Part A: Applied Science and Manufacturing, 38, 3, 828-842, 2007

11. BLEDZKI, A.K., S. REIHMANE, S. and GASSAN, J., 'Thermoplastics reinforced with wood fillers: a literature review', Polymer-Plastics Technology and Engineering, 37, 4, 451 - 468, 1998

12. AUTOMATION CREATIONS I.N.C., 1996-2009, MatWeb: Material Property Data, http://www.matweb.com (accessed on April 2009).

13. THOMASON, J.L., 'The influence of fibre length and concentration on the properties of glass fibre reinforced polypropylene: 5. Injection moulded long and short fibre PP', Composites Part A: Applied Science and Manufacturing, 33, 12, 1641-1652, 2002

14. JAKAB, E., VÁRHEGYI, G. and FAIX, O., 'Thermal decomposition of polypropylene in the presence of wood-derived materials', Journal of Analytical and Applied Pyrolysis, 56, 2, 273-285, 2000