364
Energy & Fuels 1999, 13, 364-368
Effect of Solvents on the Liquid-Phase Cracking of Thermosetting Resins Yoshiki Sato,* Yoichi Kodera, and Tohru Kamo Hydrocarbon Research Laboratory, National Institute for Resources and Environment, 16-3 Onogawa, Tsukuba-shi, Ibaraki-ken, 305-8569 Japan Received March 16, 1998
The effect of solvents on the liquid-phase cracking of thermosetting resins was examined in the presence of tetralin (hydrogen-donor solvent), decalin (nondonor solvent), or petroleum heavy oil with a reaction time of 60 min at 430-450 °C in a 200-mL autoclave under 2 MPa of initial nitrogen atmosphere. Epoxy and novolak-type phenol resins, as typical thermosetting resins, showed almost 100 wt % conversion and about an 85 wt % yield of distillable oil when an iron oxide-sulfur catalyst and tetralin solvent were used. With decalin, the oil yield from epoxy resin decreased to 75 wt %; however, the phenol resin gave an extremely low conversion of 30 wt % with an oil yield of 25 wt %. Cracking of a resol-type phenol resin gave only 9 wt % conversion with decalin but increased to 99 wt % with tetralin. ABS and urethane resin also showed high conversions of more than 90 wt % and oil yields of 60-90 wt % with tetralin. In all experimental runs, gas yields were very low at 1-3 wt %, except for the case of urethane foam in which about 20 wt % of carbon dioxide was produced. In the oil produced from epoxy and phenol resins, 4074 wt % of phenol compounds were detected. This indicates that liquid-phase cracking proceeds through simple bond dissociation followed by quick hydrogen supply from tetralin without any condensation.
1. Introduction Since the first commercial production of polyvinyl chloride in 1945, the production of plastics in Japan has increased remarkably from 0.7 million tons in 1961 to 12.3 million tons in 1993. The generation of waste plastic amounted to 7.6 million tons, almost one-half of the production. Three million tons (40%) was disposed of directly in landfills; 4 million tons (51%) was incinerated both directly (35%) and with energy recovery (16%). Only 690 000 tons (9%) was recycled. Chemical recycling of waste plastics is an important and urgent problem to be solved from the standpoints of environmental protection and effective reuse of organic resources. Pyrolysis of waste plastics has been actively studied using bench-scale units to produce fuel oil. However, because of the lack of hydrogen sources, the oil contains olefinic hydrocarbons, relatively heavy boiling-point fractions, and high concentrations of environmentally hazardous elements such as sulfur, chlorine, etc. About 1.6 million tons of thermosetting resins, such as phenol and epoxy resins, are produced annually for use mainly in electrical appliances and automobiles as well as heat-proof tableware and adhesive agents. However, it is well-known that liquefaction of these resins by pyrolysis is impossible because of the hardening when heated. Until now there have been no experimental investigations reported on chemical recycling of thermosetting resins. We have been studying chemical recycling of polyolefinic waste plastics1,2 and automobile tires3 by liquid-
phase cracking at 400-450 °C. Organic components in tires were quantitatively cracked to low-sulfur hydrocarbon oil in the presence of a hydrogen-donor solvent. In this reaction, yields of gaseous products were very low, 2-5 wt %, and the recovered carbon black could be reused for rubber manufacturing because of its fine particle size and physical properties.4 In the present study, we applied this liquid-phase cracking using hydrogen-donor or nondonor solvents to chemical recycling of thermosetting resins and some other plastics (ABS, urethane foam and resin) that are also very difficult to liquefy under pyrolytic conditions. The effects of reaction temperature, catalyst, and nitrogen or hydrogen atmosphere on the composition of the produced oil are discussed. 2. Experimental Section Pellet-form pure plastics with no contamination were used as the reactant plastics. The novolak and resol types of phenol resins used were PSM-4389, with weight-average molecular weight (Mw) of 4500-5000, and N-411M, with weight-average molecular weight of 2000-2500, produced by Gun Ei Chemical Industry Co., Ltd. The epoxy and thermoresistant ABS resins (1) Wann, J.-P.; Kamo, T.; Yamaguchi, H.; Sato, Y. 212th ACS National Meeting, Preprints 41(4), 384-386 1996. (2) Wann, J.-P.; Kamo, T.; Yamaguchi, H.; Sato, Y. 214th ACS National Meeting, Preprints 42(4), 972-977 1997. (3) Sato, Y.; Kamo, T. Globec ′96, 9th Global Environment Technology Congress, Proceeding 13.1.3-13.1.8 1996. (4) Sato, Y.; Kurahashi, S.; J. Jpn. Inst. Energy 1995, 74(2), 91989.
10.1021/ef9800516 CCC: $18.00 © 1999 American Chemical Society Published on Web 01/15/1999
Liquid-Phase Cracking of Thermosetting Resins
Figure 1. Liquid-phase cracking of phenol resins (reaction temperature: 440 °C. Reaction time: 60 min. Initial N2 pressure: 2 MPa). used were EPOMIC, with number-average molecular weight (Mn) of 1600, produced by Mitsui Chemical, Inc., and MTH-2, with Mn of 46 000, produced by Sumika A & L, Inc. The polyurethane resin used was a standard sample supplied by the Japan Urethane Industries Institute (Mw ) 3000). Experimental runs were carried out in a 200-mL stainless steel magnedrive autoclave at 430, 440, and 450 °C under 2.0 MPa of initial nitrogen pressure and a reaction time of 60 min in the presence of either tetralin or decalin. Tetralin was used as a model for a typical hydrogen-donor solvent, and decalin was used as a typical nondonor solvent. For the purpose of comparison, we used hydrogen instead of nitrogen; in some experiments we used an iron oxide-sulfur catalyst, waste engine oil or LCO (light cycle oil) from an FCC unit in the petroleum refinery process as an inexpensive, conventional solvent. In a typical run, the autoclave was charged with 23 g of plastic, 57 g of solvent, and 0.7 g of iron oxide with 0.3 g of elemental sulfur. In all runs, the autoclave was heated at a rate of 15 °C/min and was air-cooled after the reaction time. Gaseous products were analyzed with a gas-chromatograph equipped with a TCD detector using Porapak N, molecular sieves 5A and 13X, Gasukuropack 54 and VZ-7 packed columns. Liquid products were separated into solids and a liquid by filtration. Solid products were washed with THF and dried under a vacuum at 110 °C. Liquid products were vacuum-distilled at 330 °C under 3 Torr and separated into the distillable oil and vacuum residue. Conversion and yields were gravimetrically calculated from the amounts of solid residue, distillable oil, vacuum residue, and gas. Elemental analyses of solid and liquid products were performed with a Carlo Erba CHN-O elemental analyzer. Microanalyses of sulfur and nitrogen were also carried out using an ANTEK 700 elemental microprocessor. The distilled oils were subjected to gas chromatography for compositional analysis and to mass spectrometry for identification purposes.
3. Results and Discussion Phenol Resins. The conversion and product distribution from the liquid-phase cracking of novolak and resol types of phenol resins at 440 °C and 60 min under nitrogen are shown in Figure 1 and Table 1. The gaseous products consisted of mainly methane and hydrogen with traces of C2 and C3 hydrocarbon gases. The novolak type of phenol resin showed almost 100% conversion when tetralin was used as a solvent. Very high conversion and oil yields with low gas production were observed in both the thermal and catalytic reactions. When using an iron catalyst, the distillable oil yielded
Energy & Fuels, Vol. 13, No. 2, 1999 365
by vacuum distillation was fairly high at 85 wt % and, in contrast, the yields of gas and vacuum residue were very low at 1.2 and 7.2 wt %, respectively. The conversion and the oil yield were also high at 100 and 70 wt % under thermal cracking conditions. The difference in the distillable oil yield between the catalytic and thermal reactions was explained by the hydrogentransfer activity of the iron catalyst. There was no significant difference in the conversion or the oil yield between reaction temperatures of 430 and 450 °C (see Table 1). The consumption and product distribution of solvents is also given in Table 1. In these reactions, tetralin plays the role not only of the solvent to solubilize the resin and the product, but also as a hydrogen-donating agent. About 19 wt % of the naphthalene in the distillable oil was produced from the tetralin by catalytic cracking and 10 wt % by thermal cracking. Hydrogen production from the reactant resin was calculated to be 1.02 wt % by catalytic cracking and 0.49 wt % by thermal cracking. The generation of hydrogen in the gaseous product was only 0.26 and 0.07 wt %, so consequently, 0.76 wt % of the hydrogen was effectively used in the catalytic conversion and 0.42 wt % in the thermal conversion of the novolak type of phenol resin to an oil product. This means that only 33.7 and 16.2 wt % of the tetralin provided was needed to supply hydrogen for liquidphase cracking of phenol resin (novolak) equivalent in conversion and quantity of distillable oil, respectively, to those obtained by thermal cracking. Experimental results using engine oil showed a low conversion, 47 wt %, and low oil yield, 29 wt % (Table 1). This is predominantly due to the poor hydrogen-donor ability of the paraffinic oil. However, the results obtained with the use of LCO as a solvent and by hydrogen gas with decalin showed higher conversions and oil yields. It is very clear that a certain amount of hydrogen gas or a hydrogen-donor solvent such as tetralin or alkyltetralin is required for the cracking of phenol resin. In contrast, very different results were obtained with the use of decalin as a solvent. More than 70% of the resin could not be converted either with or without the iron catalyst. Only 20 wt % of oil yield was observed, but a mixture of solid products stuck to the stirrer and the walls of the autoclave. Our experimental findings indicate the difficulties with continuous operation of pyrolysis in a large-scale facility. There are almost no reports on the pyrolysis of thermosetting resins, but it is generally known that cracking or pyrolysis of thermosetting resins is impossible due to their characteristic properties. The results observed for the use of decalin showed a reaction behavior similar to pyrolysis without any solvent. The liquid products, as analyzed by gas chromatography, consisted of large amounts of phenols and cresols in the distillable oil, 26 and 38 wt %, respectively, for catalytic cracking and 22 and 24 wt % for thermal cracking (see Table 2). This indicates that more than 75 wt % of the alkylphenols in the distillable oil were recovered in the catalytic reaction. The resol type phenol resin, which has a threedimensional cross-linkage structure by a methylol group, again showed a high conversion level of more than 99
366 Energy & Fuels, Vol. 13, No. 2, 1999
Sato et al.
Table 1. Liquid-Phase Cracking of Phenol Resin at 440 °C and a Reaction Time of 60 min in the Presence of a 2 MPa Nitrogen Atmosphere resin type novolak solvent solvent/resin wt ratio catalyst conversion (wt %) product yield (wt %) gas oil VR water conversion of tetralin to naphthalene (wt %) a
resol
tetralin 2.54 Fe2O3 96.8
tetralin 2.53 non 100.0
tetralina 2.53 Fe2O3 98.4
tetralinb 2.53 Fe2O3 98.3
tetralin 2.53 Ni-Mo/Al2O3 99.7
t-decalinc 2.53 Fe2O3 98.5
LCO 2.53 Fe2O3 89.6
engine oil 2.53 Fe2O3 46.9
tetralin 2.53 Fe2O3 99.1
decalin 2.53 non 9.2
1.3 84.6 7.2 3.7 25.8
0.1 68.0 29.1 2.8 12.3
0.9 84.0 9.9 3.6 23.6
2.6 85.5 6.4 3.8 28.2
1.0 92.4 2.3 4.0 37.7
0.7 64.8 23.4 9.6
4.6 46.6 36.3 2.1
7.5 29.3 9.7 0.4
1.1 50.6 44.7 2.7 25.7
0.3 5.6 2.4 0.9
Reaction temperature: 430 °C. b Reaction temperature: 450 °C. c 5 MPa of hydrogen.
Table 2. Liquid-Phase Cracking of Epoxy and Phenol Resins in Tetralin at 440 °C phenol (novolak) catalyst Fe2O3 conv (wt %) 96.8 yield of product (wt %) gas 1.3 oil 84.6 product distribution (wt %) phenol 25.7 cresol 38.1 isopropylphenol dimethylphenol 11.8 total 75.6 VR 7.2 water 3.7
epoxy
phenol (resol)
non Fe2O3 non Fe2O3 100.0 98.1 99.9 99.1 0.1 68.0
2.7 88.1
9.9 1.1 74.1 50.6
22.2 23.7
16.7 1.2 24.3 0.6 42.8 7.3 0.0
4.7 1.9 5.0 2.8 14.4 15.9 0.0
7.5 53.4 29.1 2.8
9.8 26.8 20.4 57.0 44.7 2.7
wt % in the presence of tetralin. There was not a high oil yield (about 50 wt %), but again, about 60 wt % of phenol compounds were found in the oil products. With decalin, the conversion was again less than 10 wt %. The differences in cracking reactivity between the novolak and resol types of phenol resins are not considered to be due to bond dissociation energies but rather to the difficulties of hydrogen transfer needed to stabilize the radicals produced by C-C and/or C-O cleavage. From a typical network structure model of phenol resin proposed by Zinke et al.5 and Hochtlen et al.,6 cracking starts with C-C and/or C-O bond cleavage to produce radicals. However, because of the extremely rigid molecular structure (Figure 2), radicals in the complex three-dimensional network cannot effectively contact the solvent or any other hydrogen source to stabilize. Thus, stabilization is controlled by the solubility or mobility of the hydrogen-donating solvent. Because of the highly complex structure of the resol type resins, it was obviously more difficult for the radicals to contact the solvent than in the novolak type resins. In cracking without any supply of hydrogen, such as in pyrolysis, radicals produced by bond cleavage finally react with another neighboring part in the same molecule, as shown also in Figure 2, and form a very rigid structure leading to coking. Epoxy, ABS, Urethane Resin, and Urethane Foam. Epoxy resin, a very popular adhesive agent, is thought to be difficult to liquefy by pyrolysis. ABS, urethane resin, and urethane foam are also known to show poor conversion to oil by pyrolysis. Until now, there has been no scientific data reported on the (5) Zinke, A.; Ziegler, E. Wien. Chem. Ztg. 1944. (6) Hochtlen, A. Kunststoffe 44, 533 1954.
recycling or pyrolysis of epoxy, urethane resin, or urethane foam. The results of the liquid-phase cracking of epoxy resin, ABS, urethane resin, and urethane foam at 440 °C and 60 min under nitrogen atmosphere are shown in Figure 3. In the case of epoxy resin, more than 98 wt % conversion was obtained in the presence of iron catalyst-tetralin and no catalyst-decalin systems. Solvent or catalyst made no significant difference in conversion. The yield of distillable oil obtained from the catalytic reaction with tetralin, 88 wt %, was higher than the 74 wt % from the thermal reaction. Under the former reaction conditions, about 20 wt % of naphthalene was produced from tetralin. Production of hydrogen gas from the resin was only 0.1 wt %, and therefore, 0.5 wt % of hydrogen in the tetralin was transferred to stabilize the phenol and isopropylphenol. The concentrations of monomer compounds, such as phenol and isopropylphenol, in the oil were also very high at 42.8 wt % in the catalytic reaction with tetralin (Table 2). It is well-known that epoxy resin is the copolymer of phenol and isopropylphenol. The production ratio of phenol/isopropylphenol from the catalytic cracking with tetralin was 0.69 wt/wt, (0.99 mol/mol). This indicates that the cracking of epoxy resin proceeded stoichiometrically by the cracking and the hydrogen-transfer activities of the iron catalyst. On the other hand, with the thermal reaction in decalin, the yield of monomer compounds, phenol and isopropylphenol, was only 14.4 wt %. The production of tetralin and naphthalene was only 1.4 wt %, and the radicals produced by cracking of epoxy resin were thought to be stabilized by hydrogen transfer from other parts of the resin producing olefinic bonds. As shown in Figure 4, the structure of epoxy resin contains an isopropyl functional group that would be a source of hydrogen for stabilization of cracked radicals. This indicates that the use of a hydrogen-donor solvent has the advantage of producing higher yields of monomers and saturated compounds with high stability. The liquid-phase cracking of ABS, urethane resin, and urethane foam in the presence of tetralin also represented a conversion of more than 90 wt % with a high yield of oil (Figure 3). Notwithstanding the fact that t-decalin was used as a solvent instead of tetralin, it was possible for these resins, including epoxy resin, to be converted into an oil fraction whose reaction behavior was quite different from that under pyrolysis. This indicates that the solvent plays the role of not only the hydrogen donor, but also the medium for effective mixing and contact of the reactant with the catalyst,
Liquid-Phase Cracking of Thermosetting Resins
Energy & Fuels, Vol. 13, No. 2, 1999 367
Figure 2. Mechanism of pyrolysis and liquid-phase cracking of phenol resin.
Figure 3. Liquid-phase cracking of epoxy, ABS, and urethane resins (Reaction temperature: 440 °C. Reaction time: 60 min. Initial N2 pressure: 2 MPa).
preventing point heating and promoting a stable reaction in the reactor. Conclusion Phenol resins showed very different reaction behaviors in the liquid-phase cracking depending on the hydrogen-donor characteristics of the solvent. Both novolak and resol types of phenol resin gave almost 100 wt % conversion with the use of tetralin and high yields of monomer-size compounds. On the other hand, cracking of these resins could not proceed at all with decalin as the solvent, and the solid residue produced by the reaction in decalin could not be degraded in a reaction
Figure 4. Structure of phenol and epoxy resins.
using tetralin at 440 °C. It is obvious that thermal condensation changed the structure due to the lack of a hydrogen source for the stabilization of intermediate radicals. Conversion in the liquid-phase cracking of epoxy, ABS, urethane resin, and urethane foam was very high in the case of both tetralin and decalin. High yields of stable monomer compounds such as phenol and isopropylphenol were produced by the hydrogen supplied from tetralin. However, tetralin and naphthalene, which were the dehydrogenation products from decalin, were not detected at cracking conditions of 440 °C. The reason these resins showed high conversion in decalin is
368 Energy & Fuels, Vol. 13, No. 2, 1999
attributable to the isopropyl functional group and styrene structure in these resins, which can supply hydrogen to the intermediate radicals and stabilize them to form the corresponding olefinic compounds. In contrast, due to their rigid three-dimensional structure, phenol resins do not have a structure that can supply hydrogen. Therefore, they are very difficult to crack without a hydrogen supply either from the solvent or from gaseous hydrogen.
Sato et al.
Acknowledgment. The authors are grateful for frequent discussions with Mr. K. Tatsumoto, Mr. M. Haneda, and Mr. M. Kameyama of Mitsui SRC Co., Ltd. We also thank Gun Ei Chemical Industry Co., Ltd., Mitsui Chemical, Inc., Sumika A & L, Inc., and Japan Urethane Industries Institute for providing plastic materials. EF9800516
