the values determined for injection-moulded test specimens with those exhib- ited by the original tank material and those specified by the quality order since the test specimens were only 3 mm thick, compared with the average thickness of a tank, which is between 5 and 8 mm. Degree of crystallinity The degree of crystallinity of both the fluorinated and unfluorinated materials, both after one regrinding and one remoulding, lay within the provisions of the quality order. Melting temperature The remelting temperature of fluorinated and unfluorinated materials, both after regrinding and after remoulding, lay below the temperature range stipu- lated by the quality order. Flow index No difference could be detected between fluorinated and unfluorinated sam- ples evaluated by a spiral flow mould test. The recycling of material from used fuel tanks Parallel investigations were carried out to establish how much fuel the tank material absorbs, and thetimeinvolved for the fuel to migrate fromthematerial when no longer exposed to fuel. At a room temperature, samples punched from the original (fluorinated) tank achieved constant weight, with weight increases in the range 6.6-7.3%. Since material recycling cannot take place without elimination of the fuel, drying tests were carried out under a range of conditions. It was established that tempera- tures exceeding 100 o C were necessary to eliminate all fuel. The material used for the investigation was obtained from Volkswagen Golf and Passat vehicles, which have employed plastic fuel tanks for some 10 years. The time required for tank removal, cleaning, and elimination of tank compo- nents, made either of metal or a plastic other than HDPE, averaged some 5 min 30 sec. Ata labour cost of DM 45/hthis implies some DM 4.2for recovery of a fuel tank yielding 3.8 kg of HDPE with, at a price of about DM 1.8/kg for secondary HDPE, a nominal value of DM 6.9. 148 The Use of Recyclable Plastics in Motor Vehicles Ballistic test All test samples showed a ductile fracture. It is, however, difficult to compare perature of 230-260 o C. Subsequent material reprocessing caused no problems other than the unpleasant odour, which was absent from samples made from virgin material. Since all finished fuel tanksareinfact fluorinated, the mechanical, rheological and thermal properties of the recovered material were compared only with the fluorinated samples. Torsion test The recovered material met the requirements of the quality order in respect of torsional properties. Tensile test The moduli of elasticity of samples made from recovered material were much higher than those made from unused tanks. Though perhaps partially attribut- able to cross-linking, it is more likely to reflect a change of specification of the Lupolen 4261A, whose modulus of elasticity was in 1978 given as 1200 N/mm 2 and in 1989as 850 N/mm 2 . The highervalue of the modulus of elasticitydoes not necessarily constitute a poorer quality of material so far as the requirements of plastic fuel tanks are concerned. Charpy impact test As before, none of the test specimens broke, so satisfying the criteria of the quality order. It was noted that the impact properties did not deteriorate on re- cycling, as usually occurs. Ballistic testing Since all specimens exhibited ductile fracture it was possible to compare used and unused polymers.It was noted that, althoughthe materials had been inser- vice for ten years before recycling, the low-temperature impact properties had not deteriorated. As before, it is difficult to compare the values determined for injection-moulded test specimens with those exhibited by the original tank ma- terial and those specified by the quality order since the test specimens were only 3 mm thick, compared with the average thickness of a tank, which is between 5 and 8 mm. M. E. Henstock and K. Seidl 149 Before the tank was reground and reprocessed the fuel was extracted for 24 h at 120 o C, to avoid the risk that reprocessing might reach the fuel ignition tem- Degree of crystallinity The results fell within the ± 10% tolerance band of the quality order. Melting temperature The melting temperature, evaluated by differential scanning calorimetry, was lower than that specified by the quality order. SUMMARY AND CONCLUSIONS The monetary value of recoverable materials in road vehicles has fallen as plastics have replaced steel. Increases have occurred in the cost of the labour for scrapyard dismantling and in the amount of unsaleable residue and, therefore, in the associated disposalcosts.Improvements in scrapyard economics may pos- sibly be achieved by the prior removal from vehicles of large polymeric compo- nents and their recycling as well-characterised plastics fractions. Some trials with material recovered from used plastic fuel tanks show promising results for the manufacture of new tanks. REFERENCES 1. M. E. Henstock, Design for Recyclability, London, The Institute of Metals, pp. 3-6 (1988). 2. K. C. Dean and J. W. Sterner, Dismantling a Typical Junk Automobile to Produce Quality Scrap, United States Bureau of Mines, RI 7350, Washington, (1969). 3. M. E. Henstock, Conservation and Recycling, 2 (1), 69 (1988). 4. R. Franklin, Recycling Cars, Materials Reclamation Weekly, 8 December, 21 (1990). 5. K. D. Marshall, The Economics of Automotive Weight Reduction, Soc. of Automot. Engrs, Paper No. 700 174, Automot. Engng. Congr., Detroit, 12 January, 1970. 6. M. C. Flemings, K. B. Higbie, and D. J. McPherson, Report of Conf.: Energy Conservation and Recycling in the Aluminum Industry, Massachusetts Institute of Technology, (co-sponsored by the Center for Materials Science, M.I.T. and the U.S. Bureau of Mines, with the cooperation of the aluminium Ass.), 18-20 June, 1974. 7. Anon, The Energy Content of Plastics Articles, Association of Plastics Manufacturers in Europe, Distributed in the United Kingdom by the British Plastics Federation, Publication No.309/1, (April, 1986). 8. K. Seidl, Development of a Total Energy Balance for the Manufacture of Fuel Tanks in Steel and Plastics, Unpublished calculations (1991). 9. K. Muller, The Increasing Use of Plastics and Its Impacts on the Recyclability of Automobiles and on Waste Disposal in West Germany, The United States and Japan, Vortrag zur Recyclingplas II, Washington, D.C., 18-19 June, 1987. 150 The Use of Recyclable Plastics in Motor Vehicles Automobile Recycling, Univ. of Wisconsin, 16 October, 1975. 11. The calculations are intended to illustrate only the absolute changes in quantities of recoverable metal with change in model year and to facilitate comparison by application of metal prices at arbitrary times, i.e. April, 1976 and December, 1986. No attempt has been made to determine the revenues obtainable by scrapping each model year after a predetermined period and then deflating prices to 1988 levels. 12. M. E. Henstock, Design for Recyclability, The Institute of Metals, London, p. 74 (1988). 13. K. E. Boeger and N. R. Braton, Resources and Conservation, 14, 133 (1987). 14. G. R. Daborn and M. Webb, Treatment of Fragmentizer Waste by Starved Air Incineration - a Brief Feasibility Study, Department of Industry, Warren Spring Laboratory, LR 465 (MR) M, October, 1983. M. E. Henstock and K. Seidl 151 10. R. W. Roig, M. Narkus-Kramer, and A. L. Watson, Impacts of Materials Substitution in Automobile Manufacture on Resource Recovery, Symp., The Technology of Ground Rubber Tire-Polymer Composites K. Oliphant, P. Rajalingam, and W. E. Baker Department of Chemistry, Queen’s University, Kingston, Ontario, Canada K7L 3N6 INTRODUCTION Discarded tires represent a significant component of the overall plastics recy- cling challenge. They are an easily segregated, large volume part of the waste stream and present their own, somewhat unique, waste utilization problems. The whole issue is complicated by many alternative proposals, varying govern- ment legislation and preferences, incomplete technical information, and eco- nomic uncertainties. 1 A number of reviews have already discussed this overall disposal problem and examined many of theproposed approaches. 1-5 The major problem liesin finding approaches that are both economically and environmentally sound. Some of the methods of utilizing scrap tires that have been investigated are: burning, pyrol- ysis, use in cleaning up oil spills, road surfaces, roofing materials, and play- ground surfaces (for details see above mentioned reviews). While some of these approaches have been put into practice, the scrap tire disposal problem is still clearly a case where supply far outstrips available uses, andnew methods of uti- lization (or technological advances to extend existing ones) areclearlyneeded. One area that has the potential to utilize large volumes of discarded tires is their use as a filler in polymercomposites.Itistheproblems associated with this approach, and the technological advances made in overcoming these problems, that are the focus of this review. Unfortunately, much of the work in this area has been undertaken byindustryand is not available in theliterature.The liter- ature which is available, however, is presented, along with an in-depth look at the work carried out in our laboratories. K. Oliphant, P. Rajalingam, and W. E. Baker 153 GROUND RUBBER TIRE COMPOSITE BEHAVIOR Although the use of ground rubber tire (GRT) as a filler in polymer blends is a potentially attractive approach, it is fraught with a number of difficulties. Gen- erally, when the large GRT particles are added to either thermoplastic or thermoset matrices there is a large drop in mechanical properties, even at rela- tively low filler loadings. 6,7 Given that the approach here is to use the GRT as a low cost additive, and that there are a number of other materials competing in this regard, overcoming this large drop in properties has to be accomplished with little added cost (both intermsofadditivesand additional processing). This has proven to be quite achallengingtask.Inthe following sections the major fac- tors influencing thermoplastic GRT composite properties are discussed along with approaches to improving these properties. Tire Grinding In order to be used as afillerinpolymercomposites,tires are first ground into a fine powder on the order of 100-400 µ m. This is accomplished typically through either cryogenic or ambient grinding. General reviews of the size-reduction pro- cess have been published. 8-10 A typical process 4 generally involves tire splitters to cut the tire initially, followed by a two-roll grooved-rubber mill or hammer 154 Ground Rubber Tire-Polymer Composites Figure 1. Ground rubber tire particles. (left) cryogenically ground, (right) ambiently ground. mill. The bead wire is removed by hand or with magnets and fiber is removed at intermediate operations with hammer mills, reel beaters, and air tables that blow a steady stream of air across the rubber, separating the fiber. Between ambient and cryogenic grinding there is a noticeabledifferenceinthe nature of the ground rubber tire (GRT) particles. As shown in Figure 1, the sur- face of the cryogenically ground rubber is smooth and regular (because the parti- cles are cooled below their glass transition temperatures before fracture) compared to the rough irregular surfaces of ambiently ground material. There has been no complete study (to our knowledge) on the advantages (or disadvan- tages) of cryogenically versus ambiently ground GRT particles in polymer com- posites. Such a study is complicated because of the influences of particle size, particle size distribution and contaminant levels on GRT-polymer composites, and the factthat these all vary from supplierto supplier. It has beenfound, how- ever, that significant differences in composite properties are found for GRT of similar mesh (particle) size sourced from different suppliers andfurtherstudyof particle characteristics is currently underway in our laboratories. Characteristics of Tire Particles It is the complex nature of the GRT particle that complicates its use as a filler in polymer composites. The most pertinent feature of GRT particles is that they are still highly cross-linked. This has two major consequences: • there is little breakdown of the particles under normal melt compounding conditions (see section on particle size) • there is a sharp interface resulting in poor adhesion between the GRT par- ticles and the matrix (see section on adhesion). GRT particles are also compositionally quite complex. Tires contain a number of different rubbers (SBR, butyl rubber, natural rubber, polybutadiene rubber etc.), carbon black filler, antioxidants, and additional additives, the exact com- position depending on the type of tire and the part of the tire (e.g. tread vs. side- wall, vs. liner). Elementally, a typical tire is comprised of carbon 83%, hydrogen 7%, ash 6%, oxygen 2.5%, sulfur 1.2%, and nitrogen 0.3%. 1 There is approxi- mately 45-55% rubber hydrocarbon, 10-15% acetone extractables, 20-30% car- bon black, and 6% ash. 11 K. Oliphant, P. Rajalingam, and W. E. Baker 155 Polymer Matrix As this paper focuses on the use of GRT as a filler in thermoplastic systems the literature pertaining to thermoset systems will only be briefly reviewed. In gen- eral, addition of GRT to rubber vulcanates reduces all physical properties, the extent of deterioration depending upon the amount and particle size of the GRT added. 12-14 The use of coupling agents 15 has been reported to improvetheproper- ties of these systems. GRT particles are also found to decrease the tensile, flex- ural and storage shear modulus in composites with an unsaturated polyester resin. 16 Very poor properties are again reported for a GRT-phenoliccompound. 15 Similar poor behavior for GRT-thermoplastic composites is often reported. Deanin and Hashemielya 17 report on GRT composites made with six different polymers (HIPS, PP, HDPE, LLDPE, LDPE, and ABS) and five different elasto- mers (SBS, SEBS, SIS, butyl, and EDPM). The addition of the GRT reduced the tensile strength noticeably, and is not reported to provide for any increase in im- pact strength. Poorer properties are reported for more brittle matrices. The GRT in this study was, however, broken down to some extent in a prior mastication step, reportedly enabling it to form into a “thin thermoplastic sheet”. Its exact nature, compared to the still highly cross-linked large GRT particles typically employed in GRT composites was not reported however. Tuchman and Rosen 15 examined composites of GRT and PP, ABS, PS, LDPE, and HDPE. Addition of GRT was reported to reduce all the mechanical properties of ABS, LDPE, and HDPE. The Izod impact strengths of PP and PS, however, are reportedly in- creased (by up to three times for PP at 40 wt% GRT) upon the addition of GRT. Phadke and De, 18 however, report that there is a decreaseintheimpactstrength of PP when GRT is added. Duhaime and Baker 6 report on the properties of LLDPE-GRT composites. The impact strength is seen to drop by 35% even at 10 wt% filler loading. The impact strength continues to drop until at 30 wt% filler it is 50% lower than that of the pure LLDPE. It remains approximately constant, however, from 30-60 wt% GRT. The impact behavior was characterized using a Rheometrics drop-weight instrumented impact tester,in which a piezoelectric loadcell is tip-mounted to a high velocity dart permitting the recording of the load-displacement curve for the entire impact event. For pure LLDPE, impact failure is seen to be a ductile yielding process in which the dart draws the material out as it passes through. 19 For the GRT-LLDPE composites the failure is similar except that ‘yielding‘ oc- curs at much lower forces and elongations, and the material is not drawn out to 156 Ground Rubber Tire-Polymer Composites as high elongations because of premature failure caused by the large, poorly bonded rubber particles. In tensile tests, the tensile strength and ultimate elon- gation are seen to decrease steadily with increasing GRT content. At 40 wt% GRT the tensile strength has decreased by 48% and the tensile elongation by 80%. Oliphant and Baker 7 report on blends of GRTwithLLDPEand HDPE. The ad- dition of GRT to LLDPE results in composites with properties similar to those described by Duhaime. 6 The deleterious effects of theGRTparticles on HDPE is, however, more pronounced than for blends of the GRT with LLDPE (70% de- crease in impact strength for HDPE compared to 50% for LLDPE at 40 wt% GRT), although the observed trends are similar. In contrast to the failure of LLDPE described previously, the failure of pure HDPE, although it involves some plastic deformation, is observed tooccurthroughcatastrophicpropagation of a crack through the impact zone. This type of failure is also observed in the GRT-HDPE composites (compared to the ductile tearing process observed for GRT-LLDPE composites). It is this difference in impact failure which is sug- gested to beresponsible for the poorer propertiesof GRT-HDPE composites. It is postulated that in HDPE-GRT composites the failure remains semi-brittle be- cause the particles aretoo large to induce abrittleto ductile transition (or ashift in the brittle to ductile transition temperature to below the test temperature). Failure then occurs largely through crack propagation and the large particles K. Oliphant, P. Rajalingam, and W. E. Baker 157 Table 1 Influence of melt flow index on the impact energy of LLDPE/GRT composites MFI Impact energy (J) Drop (%) Pure PE 40 wt% GRT 0.3* 15.8 7.6 52 1.0 21.0 13.9 34 5.0 14.9 7.9 47 12.0 12.9 7.4 43 20.0 11.4 6.8 40 *LDPE act as serious flaws, providing an easy path for the crack to follow. The addition of GRT to a semi-brittle matrix is therefore believed to require much higher lev- els of adhesion (to retardcrack growth at the particle/matrix interface), or much lower particle sizes (to lower the brittle-ductile transition temperature). This is borne out to some extent experimentally (see reference 7 and section on adhe- sion). Rajalingam and Baker 20 have studied GRT compositeswith a number of differ- ent LLDPE’s and a LDPE. The impact properties of the pure matrices and their corresponding 40 wt% GRT composites are given in Table 1. Although there is some variation in the percent drop in impact strength with MFI, the addition of GRT is shown in all cases to have a similar influence on mechanical properties. There are two exceptions of note: the higher molecular weight 1 MFI LLDPE ap- pears to produce composites with lower material property drop, and the reduc- tion in impact strength for the LDPE is slightly higher than for the LLDPE composites. Deanin and Hashemiolya 17 have also reported that LDPE produced poorer GRT composites than LLDPE. It is also interesting to note that the LLDPE of MFI = 1.0 produces a 40 wt% GRT composite with slightly higher im- pact strength thanfor pure LLDPE’s with MFIs of12 and 20 dg/min. Thus,if the higher viscosities of the composite can be tolerated in processing, these materi- als may prove to be useful composites. In general then, it is seen that simple addition of GRT to most polymers by melt blending results in a significant deterioration in mechanical properties. Particle Size The large rubber particle size used in GRT composites is reported to be one of the two major factors (the other being adhesion) contributing to the poor me- chanical properties generally observed for GRT-polymer composites. The impor- tance of particle size (and particle size distribution and shape) on mechanical properties of composites in general is well known in the literature. 21 For rubber toughening applications it is generally reported that there is an optimum parti- cle size (typically in the 0.1-5 µ m range) for toughening brittle polymers, 21 and a minimum particle size (or inter-particle distance) for toughening semi-ductile polymers (typically less than 1 µ m). 22 For hard particulate fillers, impact strength is generally observed to increase with a decrease in particle size. In general, for optimum composite properties, a low particle size is desired. In GRT-polymer composites, however, the particle size is (relatively) quite large. 158 Ground Rubber Tire-Polymer Composites [...]... modulus for the GTB particles than for the GRT particles, which could also lead to the mechanical property improvements However, no difference in modulus of the particles is suggested on comparing the relative stiffness of blends with GTB and GRT Scanning electron 162 Ground Rubber Tire-Polymer Composites micrographs (SEM) of the composites revealed little or no breakdown of the particles on blending... reveals an oxygen surface content of 5-15%, which may indicate the presence of -OH or -COOH functionalities There have been a number of reports of processes that claim to improve properties of GRT-polymer composites through enhancing adhesion Many of these are from industrial sources and include little specific information For thermoset systems, coating the surface of the GRT particles with an unsaturated... tests of LLDPE melt bonded to a tire compound which simulates the tire tread composition that there is negligible adhesion That this was largely an artifact of the cross-link density was shown in the significantly higher adhesion of LLDPE to a tire compound having a much lower concentration of curing agents and hence a lower degree of cross-linking In this sense then GRT particles are akin to hard particulate... Given the influence of particle size on GRT composites (and on the properties of particulate composites in general) there is some question as to what properties are ultimately obtainable with the large GRT particles even if good adhesion can be obtained In other words, is there any justification for focusing on improving adhesion while leaving particle size fixed? To this end, blends of ground tire bladder... 20 and Baker to cover particle sizes ranging from 28 mesh (600 µm) to 200 mesh (74 µm) These materials were obtained by sieving off different mesh sizes from a sample originally ground to 40 mesh (300 µm) Four wt% of a reactive coupling agent was added to the blends (see section on adhesion) Figure 2 shows the influence of particle size on impact energy and the MFI for blends of 40 wt% GRT in LLDPE... believed to be the major factor (in addition to the large particle size) leading to the large mechanical property decreases observed upon incorporation of GRT into most polymers K Oliphant, P Rajalingam, and W E Baker 161 This poor adhesion is, at least in part, due to a high degree of crosslinking in the GRT particles The highly crosslinked nature of the particles inhibits molecular diffusion across the... was shown to be, at least partially, due K Oliphant, P Rajalingam, and W E Baker 163 to cross-linking of the PE phase by residual initiator In fact, spray coating of the GRT surface with low concentrations of a peroxide initiator (= 0.1 wt%) before blending with LLDPE resulted in composites with impact strengths up to 80% of those of the pure LLDPE Although straight addition of the peroxide resulted... particles with an EAA copolymer overcame most of the deleterious effects of adding GRT to LLDPE (while still retaining composite processability) A blend of 40 wt% EAA coated GRT particles (4 wt% EAA) with LLDPE was shown to have impact and tensile strengths 90% of those for pure LLDPE, representing increases of 60% and 20% respectively, over blends with uncoated particles It was suggested that an interaction... crack formation on failure) 7 are less tolerant of the large GRT particles and, as described previously, consequently require either much higher levels of interaction (adhesion) or much smaller particle sizes Matrix Modification An alternative approach to improving upon GRT-thermoplastic composite properties is to add materials that will enhance the properties of the matrix, and thus the overall composite... the large particle size Little is reported in the open literature on the nature of the GRT particle surface or on its potential for chemical interactions that would increase adhesion Infra-red spectroscopy has been reported to show the presence 28 of unreacted double bonds but does not mention the presence (or absence) of other functional groups Electronic Spectra for Chemical Analysis (ESCA) of 20 GRT . cost of DM 45/hthis implies some DM 4.2for recovery of a fuel tank yielding 3.8 kg of HDPE with, at a price of about DM 1.8/kg for secondary HDPE, a nominal value of DM 6.9. 148 The Use of Recyclable. for GRT of similar mesh (particle) size sourced from different suppliers andfurtherstudyof particle characteristics is currently underway in our laboratories. Characteristics of Tire Particles It. found 7 that pre-coating of GRT particles with an EAA copolymer overcame most of the dele- terious effects of adding GRT to LLDPE (while still retaining composite processability). A blend of 40 wt% EAA coated GRT particles