Ind. Eng. Chem. Res. 2003, 42, 675-679 675 APPLIED CHEMISTRY Conversion of a Used Poly(ethylene terephthalate) Bottle into Oxalic Acid and Terephthalic Acid by Oxygen Oxidation in Alkaline Solutions at Elevated Temperatures Toshiaki Yoshioka,* Masaki Ota, and Akitsugu Okuwaki Department of Applied Chemistry, Graduate School of Engineering, Tohoku University, Aramaki Aza Aoba 07, Aoba-ku, Sendai 980-8579, Japan Chemical recycling of poly(ethylene terephthalate) (PET) flakes to terephthalic acid (TPA) and oxalic acid was investigated by simultaneous hydrolysis and oxygen oxidation in concentrated NaOH. PET flakes were hydrolyzed to sodium terephthalate and ethylene glycol (EG) in NaOH solutions before oxygen introduction. Because sparingly soluble sodium terephthalate in concentrated NaOH solutions was stable to the oxidation, the TPA yield was approximately 100 mol % under all conditions. In contrast, EG was oxidized to oxalate and CO 2, and the maximum oxalic acid yields were 60.7 mol % using flakes from the bodies of transparent bottles and 65.9 and 71.4 mol % using commercial transparent flakes and a mixture of transparent and green flakes from PET, respectively. If the gate fee (83.3 /kg of PET) can be obtained as a waste PET recycling subsidy from the Japanese government, this process will generate a profit of $1,134.20 for the treatment of 1000 kg of PET as a net profit for 70 mol % of oxalic acid yield. 1. Introduction Recently, disposal of waste materials has come into focus as an environmental problem that affects everyone. Setting up intermediate treatment plants, such as incinerators, or obtaining the land for reclamation is difficult. Poly(ethylene terephthalate) (PET) resin is inexpensive, light, transparent, tough, and easy to process and, as a result, is used not only in several industries but also in daily life. In particular, PET is used for beverage bottles, the demand for which is increasing year by year. In 2000, domestic production of PET bottles reached 401 000 tons in Japan. Currently used PET is recycled chemically by solvolysis such as methanolysis 1-3 and glycolysis. 4-8 In these methods, PET is converted to dimethyl terephthalate (DMT) and ethylene glycol (EG) and/or a monomer such as dihydroxyethyl terephthalate (DHET) and a lowmolecular-weight oligomer. Terephthalic acid (TPA) and EG can be obtained by hydrolysis 9-14 in acidic and basic conditions. The hydrolysis rate of PET in water was measured from 60 to 175 C by Golike and Lasoski 15 and from 260 to 285 C by Campanelli et al. 16 Zimmerman and Kim 17 investigated the effect of metal catalysts, such as Sb, Zn, Mn/ Sb/Pb, Zn/Sb, and Ca/Sb, on the hydrolysis from 280 to 300 C. Sanders and Zeronian, 18 Schlts and Lhymn, 19 and Dave et al. 20 studied the alkaline hydrolysis of PET fiber in a PET/glass composite with respect to the stabilization of PET. A number of researchers have studied the chemical recycling process of PET by alkalis in anhydrous conditions in the presence of alcohol or glycol and aprotic solvent to recover the monomeric components and sodium terephthalate. 21-23 Paszun and Spychaj have presented a review of investigations of PET chemical recycling. 24 However, the development of a new process to promote the recycling of waste PET by recovering more valuable chemicals is necessary. One strategy is the conversion of TPA and EG units in PET into more expensive chemicals during the recycling process. Oxalic acid, which not only is used as a semifine raw chemical but also is thought to be a raw chemical of oxamide ((CONH 2 ) 2, which is a slow-releasing nitrogenous fertilizer 25 ), is produced from EG. In the present study, used PET flakes were oxidized by oxygen in NaOH solutions, PET was hydrolyzed to TPA and EG, and only EG was oxidized simultaneously to oxalic acid. The optimum conditions under which to obtain TPA and oxalic acid from used PET flakes are reported. 2. Experimental Section 2.1. Materials. Additive-free PET flakes (0.4 5 5 mm) were prepared from the bodies of used transparent PET bottles and did not contain crystallized cap or bottom parts. The average molecular weight was approximately 30 000. In the experiment of section 3.3, PET flakes were prepared from both the bodies and crystallized cap parts of used transparent bottles and green bottles. The flakes were washed in water under ultrasonic irradiation for 10 min. All chemicals used were reagent grade. 2.2. Oxygen Oxidation in Alkaline Solutions. A mixture of 0.67 g of PET flakes, 20 cm 3 of water, and a 10.1021/ie010563z CCC: $25.00 2003 American Chemical Society Published on Web 01/28/2003
676 Ind. Eng. Chem. Res., Vol. 42, No. 4, 2003 prescribed amount of NaOH was placed in a poly- (tetrafluoroethylene) (PTFE) beaker (wall thickness of 2 mm, o.d. of 30 mm, and depth of 120 mm). The PTFE beaker was placed into a 316 stainless steel autoclave of 100 cm 3 in volume which was fitted with a magnetdriven stirrer. The air was then replaced with N 2, and the beaker was heated to 160-250 C at 3 C/min in an electric furnace. After reaching the prescribed temperature, the oxygen oxidation was started by pressurized oxygen at 1-10 MPa, and this time was defined as the reaction start. The time at which the autoclave was removed from the electric furnace was defined as the reaction end time, after which air cooling of the autoclave to room temperature was performed compulsorily. The product in the beaker was washed out with water, filtrated using a 1G4 glass filter, diluted to 500 cm 3, and then stocked as a reaction solution for analysis. 2.3. Analysis and Definition. Sodium terephthalate in the reaction solution was precipitated as TPA by the addition of 2MH 2 SO 4, filtered with a glass filter, and washed with cold water. The TPA on the glass filter was dried at 105 C for 2hinanoven. After removal of sodium ions in the reaction solution using a cation-exchange column, the amount of oxalate ion was determined by ion chromatography (Dionex Quic; columns AG4 and AS4) and that of EG and other C 2 substrates was determined by high-performance liquid chromatography (column, Shodex Ionpac 810P- KC 811; detector, RI; 30 C) and total organic carbon analysis (Shimadzu TOC 5000). The amount of CO 2 was determined by gas chromatography (Hitachi 164; column, stainless steel-silica; detector, TCD; 40 C), after evolution of CO 2 by adding 6MH 2 SO 4. The degree of weight loss and yields of terephthalate, EG, oxalate, and CO 2 are defined as follows: degree of weight loss (wt %) ) {(W 0 - W t )/W 0 } 100 (1) TPA yield (mol %) ) (m TPA /m PET ) 100 (2) Figure 1. Effect of the reaction time on the TPA (0), EG (4), oxalate (O), and CO 2 (b) yields in 27.5 M NaOH at 250 C and 5 MPa P O2. Figure 2. Effect of the NaOH concentration on the TPA (0), EG (4), oxalic acid (O), and CO 2 (b) yields at 250 C and 5 MPa P O2 for5h. 1 h. The CO 2 yield increased little after 1 h, indicating that oxalic acid is formed from released EG and primarily that CO 2 may be produced in the very beginning of the oxidation of EG. EG yield (mol %) ) (m EG /m PET ) 100 (3) oxalic acid yield (mol %) ) (m ox /m PET ) 100 (4) CO 2 yield (mol %) ) (m CO2 /2m PET ) 100 (5) where W 0 is the initial weight of PET, W t is the weight of the PET residue, m PET is the initial number of moles of PET-based monomeric units, and m TPA, m EG, m ox, and m CO2 are moles of TPA, EG, oxalic acid, and CO 2, respectively. 3. Results and Discussion 3.1. Effects of Reaction Conditions on the Yields of TPA, Oxalic Acid, and CO 2. 3.1.1. Oxidation Curve. A typical oxidation curve of PET flakes is shown in Figure 1. Both TPA and EG yields were quantitative at 0 h, which indicates that the hydrolysis of PET flakes progressed satisfactorily. The TPA yield remained quantitative, indicating that PET was hydrolyzed easily to sodium terephthalate and EG before oxygen introduction (eq 6). In contrast, EG was oxidized rapidly to oxalic acid and CO 2 until 1 h after the oxygen introduction. The oxalic acid yield increased linearly, having a small slope corresponding to a decrease of the EG yield after Some chemicals, such as glycolic aldehyde, glyoxal, glycolic acid, and glyoxylic acid, are produced until the formation of oxalic acid in the oxidation of EG. For example, when each substrate was oxidized as a starting material, oxalic acid was quantitatively formed from glyoxal, glycolic acid, and glyoxylic acid without the formation of CO 2. In contrast, CO 2 and oxalic acid were formed from EG and glycolic aldehyde. Therefore, in the oxidation method of EG to oxalic acid, the majority of CO 2 may be produced in the beginning of the oxidation as shown in Figure 2. 3.1.2. NaOH Concentration. The effect of NaOH concentration on TPA, EG, oxalic acid, and CO 2 yields is shown in Figure 3. EG was not measured until reaching 27.5 mol/kg H2O NaOH due to the complete oxidation and the EG yield increased beyond 27.5 mol/
Ind. Eng. Chem. Res., Vol. 42, No. 4, 2003 677 Figure 3. Effect of the NaOH concentration on the TPA (0), EG (4), oxalic acid (O), and CO 2 (b) yields at 250 C and 5 MPa P O2 for5h. Figure 5. Effect of the reaction temperature on the TPA (0), EG (4), oxalic acid (O), and CO 2 (b) yields in 27.5 M NaOH at 5 MPa P O2 for5h. Table 1. Oxalic Acid Yield from Various PET Flakes (NaOH, 27.5 M; P O2, 5 MPa; 250 C; 5 h; 1000 rpm) type yield (mol %) PET flakes (clear type, body) 60.7 PET flakes (clear type, cap and body) 65.9 PET flakes (clear and green types, cap and body) 71.4 Figure 4. Effect of the oxygen partial pressure on the TPA (0), EG (4), oxalic acid (O), and CO 2 (b) yields in 27.5 M NaOH at 250 C for 5 h. kg H2O. Oxalic acid was formed just below 5 mol/kg H2O NaOH, indicating that the base-catalyzed oxygen oxidation does not work because free alkali was absent because of the neutralization of NaOH using TPA and CO 2. The oxalic acid yield increased linearly from 3.5 mol % at 5 mol/kg H2O NaOH to 60.7 mol % at 27.5 mol/ kg H2O NaOH. The sharp linear increase in the oxalic acid yield indicates that oxalic acid is formed by the basecatalyzed oxygen oxidation of EG because the oxidation rate is proportional to the first-order function of the OH - concentration. 26 Over 27.5 mol/kg H2O NaOH, the oxalic acid yield decreased whereas the EG yield increased remarkably. This is caused by lowering the rate of oxidation of EG because of the decrease in the solubility of oxygen in concentrated NaOH solutions, and some mass transfer, such as the transfer of oxygen, controlled the basecatalyzed oxygen oxidation. 3.1.3. Oxygen Partial Pressure. The effect of the oxygen partial pressure on TPA, EG, oxalic acid, and CO 2 yields in 27.5 mol/kg H2O NaOH is shown in Figure 4. The TPA yield was also quantitative and was not affected by the oxygen partial pressure. The EG yield was quantitative at 0 MPa P O2 but decreased rapidly by 0 mol % at 5 MPa with increasing oxygen partial pressure. In contrast, the oxalic acid yield increased with increasing oxygen partial pressure up to 5 MPa but decreased at 10 MPa, indicating the possible production of less oxalic acid because of a stronger oxidation of EG. 3.1.4. Temperature. The effect of the reaction temperature on TPA, EG, oxalic acid, and CO 2 yields is shown in Figure 5. TPA was always produced quantitatively. The EG yield decreased, but oxalic acid and CO 2 yields increased linearly with increasing temperature. The slope of the formation of CO 2 was smaller than that of oxalic acid, which means that the selectivity of oxalic acid formation is advantageous at higher temperatures ranging from 160 to 250 C. However, because the temperature dependence is low, the basecatalyzed oxygen oxidation appears to be controlled by mass transport processes, as shown those in section 3.1.2. Thus, detailed kinetics of the oxidation of EG will be required from this point on. 3.1.5. Effect of Copper(II) Oxide. The oxygen oxidation of acetate in a NaOH solution was reported to be accelerated by transition-metal oxide catalysts such as CuO and NiO, 27 and the yield of oxalic acid is increased. Similarly, the effect of CuO can be held responsible for the improvement in the yield of oxalic acid in this reaction. The catalytic effect of CuO was observed to some extent and the yield of oxalic acid increased from 60.7 to 65.3 mol % in the presence of 5 wt % CuO powder on PET flakes under 250 C, 27.5 mol/kg H2O NaOH, 5 MPa P O2,and5h. 3.2. Effect of the Type of PET Flakes Obtained from Postconsumer PET Bottles. Commercial PET flakes from used bottles contain the hard parts of the cap and bottom, the crystallinity of which is higher than that of the body. Two types of flakes containing the hard parts were used as samples. One sample was produced from clear-type bottles, and the other was produced from a mixture of clear and green bottles. These samples were oxidized in 27.5 M NaOH at 250 C and an oxygen partial pressure of 5 MPa for 5 h. The resulting product solutions were colorless even for the mixture type, indicating that the green dyestuff in PET bottles was sufficiently oxidized. The results are shown in Table 1. The oxalic acid yield was 65.9 mol % for the clear-type sample and 71.4 mol % for the mixture-type sample. These yields were higher than 60.7 mol % for the PET flakes made of the body alone. The reasons for this
678 Ind. Eng. Chem. Res., Vol. 42, No. 4, 2003 Table 2. Total Balance of the Alkaline Hydrolysis-Oxidation Process of Waste PET Bottles (Oxalic Acid Yield ) 70 mol %) raw materials product byproduct PET NaOH H 2SO 4 O 2 TPA oxalic acid Na 2SO 4 molar ratio 1 4 2 4 1 0.7 2 weight ratio 192.18 160.00 196.18 64.00 166.14 63.03 284.10 weight (kg) 1,000 832 1,021 333 865 328 1478 unit price ( /kg) a -83.3 37.5 19.1 50.1 104.2 345.8 50.1 sum ($) -833.0 312.0 196.2 166.8 901.3 1134.2 740.5 total ($) -158.9 1134.2 a Trust money for recycling of waste PET bottles: 83.3 /kg (1$ ) 120, 2000 in Japan). government. Therefore, this process will be able to generate a profit of $1,293.10 for the treatment of 1000 kg of PET as a net profit for 70 mol % of oxalic acid yield. Moreover, the market for TPA, oxalic acid, and Na 2 SO 4 with be served. Acknowledgment The authors thank Ueyama Works, for supplying the PET flakes. Literature Cited Figure 6. Alkali hydrolysis-oxidation process flow of waste PET bottles. remain unclear, but the selectivity is thought to have produced oxalic acid according to the effect of basecatalyzed oxygen-oxidation increases at the time of EG release because the rate of hydrolysis of the cap and bottom parts of high crystallinity is too slow. The increase in the oxalic acid yield for green bottles is due to the pigment working as a catalyst. This result indicates that pretreatments such as a water washing and separation between transparent and green bottles are not necessary. 3.3. Chemical Recycling Process for Waste PET. Figure 6 is a flow sheet of a feedstock recycling process for waste PET using simultaneous hydrolysis and oxygen oxidation in alkaline solution, and Table 2 is a balance sheet based only on raw material costs. In this process, the main products are TPA, oxalic acid, and sodium carbonate, with sodium sulfate as a byproduct. Conversion into TPA and oxalic acid as a free acid adding sulfuric acid is necessary because these products are obtained as sodium salts. In contrast, CO 2 was precipitated as Na 2 CO 3 using sodium terephthalate and sodium oxalate. However, this can be released as gaseous CO 2 by adding sulfuric acid, and consequently Na 2 SO 4 is precipitated. Na 2 SO 4 is thus produced as a byproduct. We can obtain the gate fee (83.3 /kg of PET) as a waste PET recycling subsidy from the Japanese (1) Vereinigte Glanzstoff Fabriken A.G. Process for the Conversion of Polyethylene Terephthalate into Dimethyl Terephthalate. Br. Patent 755,071, 1956. (2) E. I. du Pont de Nemours and Co. Improvements in the Preparation of High Quality Dimethyl Terephthalate. Br. Patent 784,248, 1957. (3) Grushke, H.; Hammerschick, W.; Nauchem, B. Process for Depolymerizing Polyethylene Terephthalate to Terephthalic Acid Dimethyl Ester. U.S. Patent 3,403,115, 1968. (4) MacDowell, J. T.; Klusion, N. C. Reclaiming Linear Terephthalate Polyesters. U.S. Patent 3,222,999, 1965. (5) Vaidya, U. R.; Nadlkarni, V. M. Unsaturated Polyesters from PET Waste: Kinetics of Polycondensation. J. Appl. Polym. Sci. 1987, 34, 235-245. (6) Vaidya, U. R.; Nadlkarni, V. M. Polyester Polyols for Polyurethanes from PET Waste: Kinetics of Polycondensation. J. Appl. Polym. Sci. 1988, 35, 775-785. (7) Trowell, J. T. Polyols from Scrap Polyethylene Terephthalate and Dimethyl Terephthalate Process Residue. U.S. Patent 4,- 720,571, 1988. (8) Gintis, D. Glycolytic Recycle of Poly(ethylene terephthalate) (PET). Macromol. Chem., Macromol. Symp. 1992, 57, 185-190. (9) Brown, G. E., Jr.; O Brien, R. C. Method for Recovery of Trephthalic Acid and Ethylene Glycol from Polyester Materials. U.S. Patent 3,952,053, 1976. (10) Pusztaseri, S. F. Method for Recovery of Trephthalic Acid from Polyester Scrap. U.S. Patent 4,355,175, 1987. (11) Yoshioka, T.; Sato, T.; Okuwaki, A. Hydrolysis of Waste PET by Sulfuric Acid at 150 C for a Chemical Recycling. J. Appl. Polym. Sci. 1994, 52, 1353-1355. (12) Yoshioka, T.; Okayama, N.; Okuwaki, A. Kinetics of Hydrolysis of PET powder in Nitric Acid by a Modified Shrinking- Core Model. Ind. Eng. Chem. Res. 1998, 37, 336-340. (13) Yoshioka, T.; Motoki, T.; Okuwaki, A. Kinetics of Hydrolysis of Poly(ethylene terephthalate) powder in Sulfuric Acid by a Modifified Shrinking-Core Model. Ind. Eng. Chem. Res. 2001, 40, 75-79. (14) Mandoki, J. W. Depolymerization of Condensation Polymer. U.S. Patent 4,605,762, 1986. (15) Golike, R. C.; Lasoski, S. W., Jr. Kinetics of Hydrolysis of Polyethylene Terephthalate Films. J. Phys. Chem. 1960, 64, 895-898. (16) Campanelli, J. R.; Kamal, M. R.; Cooper, D. G. A Kinetics Study of the Hydrolytic Degradation of Polyethylene Terephthalate at High Temperatures. J. Appl. Polym. Sci. 1993, 48, 443-451. (17) Zimmerman, H.; Kim, N. T. Investigation on Thermal and Hydrolytic Degradation of Poly(ethylene terephthalate). Polym. Eng. Sci. 1980, 20, 680-683.
Ind. Eng. Chem. Res., Vol. 42, No. 4, 2003 679 (18) Sanders, E. M.; Zeronian, S. H. An Analysis of the Moisture-Related Properties of Hydrolyzed Polyester. J. Appl. Polym. Sci. 1982, 27, 4477-4481. (19) Schlts, J. M.; Lhymn, C. Stress-Rupture Lifetime of a Poly- (ethylene terephthalate)/glass Composite Under an Alkaline Solution Environment. Polym. Compos. 1984, 7, 208-214. (20) Dave, J.; Kumar, R.; Stivastave, H. C. Studies on Modification of Polyester Fabrics I: Alkaline Hydrolysis. J. Appl. Polym. Sci. 1987, 33, 455-477. (21) Tindall, G. W.; Perry, R. L. Esters Hydrolysis and Depolymerization of Polyesters and Polycarbonate Polymer. U.S. Patent 5,045,122, 1991. (22) Tindall, G. W.; Perry, R. L.; Spaugh, S., Jr. Depolymerization of Substantially Amorphous Polyesters. U.S. Patent 5,328,982, 1994. (23) Oku, A.; Hu, L. C.; Yamada, E. Alkali Decomposition of Poly(ethylene terephthalate) with Sodium Hydroxide in Nonaqueous Ethylene Glycol: A Study on Recycling of Terephthalic Acid and Ethylene Glycol. J. Appl. Polym. Sci. 1997, 63, 595-598. (24) Paszun, D.; Spychaj, T. Chemical Recycling of Poly- (ethylene terephthalate). Ind. Eng. Chem. Res. 1997, 36, 1373-1383. (25) Okuwaki, A.; Okabe, T. Development of a New Route to Oxamide from Coal and Ammonia. Trends Inorg. Chem. 1991, 2, 145-158. (26) Wakabayashi, T.; Okuwaki, A. Oxidation of Coals in Liquid Phases. Kinetics of the Base-Catalyzed Oxidation of Acetate Ion by Oxygen at Elevated Temperatures. Bull. Chem. Soc. Jpn. 1988, 61, 4329-4334. (27) Ichinose, S.; Okuwaki, A. Oxidation of Coals in Liquid- Phase X. Mechanism of the Cleavage of Benzenecarboxylic Acid to Oxalic Acid and Carbon Dioxide by the Base-Catalyzed Oxygen- Oxidation. Bull. Chem. Soc. Jpn. 1990, 63, 159-165. Received for review June 26, 2001 Revised manuscript received December 16, 2002 Accepted December 21, 2002 IE010563Z