Disproportionationof Toluene*

by Takao Iwamura**,Seiya Olam*** and Masaki Sato** プ 解

Summary: Toray has developedthe "TATORAY" process which is applicable to produce benzeneand xylenefrom tolueneand C9 aromatics by disproportionationand transalkylationtech- nique. The first commercialplant was built in the Kawasaki plant of Toray in summer 1969 and the plant has been successfullyon-stream for over oneyear. Distinguishingfeatures of "TATORAY" process proved in commercialoperation are as follows: (1) Product molar ratio of benzeneand xylene is adjustable in accordancewith demand of nylon and polyesterfibers. (2) The operatingcondition is much milderthan a conventionalhydrodealkylation process. Over all aromatic yield can be maintained over 97% in "TATORAY". (3) For productionof p- and o-xylenes,low contentof ethylbenzene(normally less than 2mol%) and high content of p- and o-xylenes (more than 20mol%) are of great advantage becauseof higher separation efficiencyand yield. (4) Consumptionof hydrogenduring the reaction is very low since the "TATORAY" reac- tion is carried out without appreciableside reactions. (5) The catalyst is totally regenerableand a single on-streamperiod can be extendedover several months, as coke formation on the catalyst is substantiallyprevented. This makes a singlefixed bed reactor systempossible.

1 Introduction has been successfully on-stream since the begin- Recently the demand for aromatics, particu- ning of October, 1969 at the Kawasaki plant larly benzene and xylenes, has been considerably of Toray Industries, Inc.. increased in Japan. A large and growing amount 2 Disproportionation of Toluene of benzene is being used as a starting material 2.1 Explanation of the Reaction in manufacturing nylon, polystyrene and other Disproportionation of toluene is normally car- synthetic materials, while rapid expansion in the ried out in vapor phase with a solid acid catalyst. production of polyester fiber and film has en- Principal reactions: hanced the position of xylenes. The aromatics have mainly come from the catalytic reforming of naphtha, pyrolysis gasoline and gas oil. However, these processes also give less valuable aromatics, such as toluene and C9 aromatics, as well as valuable benzene and If trimethylbenzene is present in the feed, xylene. the following transalkylation reaction also takes There are several routes to treat toluene. For place. example, toluene can be converted to benzene, caprolactam, terephthalic acid and toluene di- isocyanate. Recently Toray has developed new catalyst and process to convert toluene not only Typical side reactions are disproportionation to benzene but also to xylene, and it was called of xylene and hydrodealkylation of toluene. "TATORAY" . It can more favourably treat C9 aromatics as a feed stock to produce more xylene. The first commercial plant, a unit of 70,000 metric tons of fresh toluene feed per year

* Received January 12, 1971. ** Chemicals Research & Development Laboratories, Other side reactions are aromatic ring rupture Toray Industries, Inc. (1,111, Tebiro, Kamakura, Kanagawa, Japan) and formation of heavier materials, such as fluo- *** Kawasaki Plant, Toray Industries, Inc. rene, naphthalene and its derivatives. To pre-

Bulletin of The Japan Petroleum Institute Iwamura, Otani and Sato: Disproportionation of Toluene 117

vent side reactions and coke formation on a cata- temperature was measured by thermocouples lyst, the presence of hydrogen is effective, through installed at several points in the reactor. For no hydrogen is required in the formula33),44). normal experiments, T-81 catalyst pellets having 2.2 Catalyst 5mm in diameter and 4mm in length were Much has been reported in previous publi- packed in the reactor. Gas chromatography was cations regarding disproportionation catalysts. used to analyse feeds and effluents. Basically, there are the following kinds of cata- 2.3.2 Chemical Equilibrium

lyst systems; Friedel-Crafts1)～10), silica-alu- Chemical equilibrium data on the toluene

mina11)～31) and zeolite type32)～65). disproportionation reaction was reported an For an efficient commercial catalyst, there are the system of the methylbenzenes71)～73). How- several requirements. One of the most impor- ever, the equilibrium with the feed which con- tant is to prevent fast fouling in catalytic activi- tains ethylbenzene and ethyltoluene has not been ty due to coke formation on catalyst surface. reported thus far. Hence the experimental runs If the amount of coke deposite on catalyst can were conducted using the apparatus shown in be decreased, regeneration of catalyst would not Fig. 1 and the results summarized in Table 1. be needed as frequently, and a single fixed-bed The results shows fairly good agreement with reactor can be achieved. The Toray effort had the equilibrium concentrations calculated by been devoted to develop such catalyst systems Egan. Only a slight different was observed in and over five hundred kinds of catalyst were ex- the C8 aromatic compositions, which was appa- plored. Finally an excellent new catalyst rently caused by disproportionation reactions of system (T-81) was developed, which shows high Table 1 Equilibrium Concentration Unit: mol selectivity, high conversion per pass and good stability over a period of several months without

any regeneration66)～70). As well as firm development of catalyst, cor- responding suitable operating conditions have been established. Final reactor zone consists of a single adiabatic fixed bed reactor. 2.3 Kinetics of the Reaction 2.3.1 Experimental Apparatus8) Fig. 1 shows an experimental apparatus for investigating the reaction at high pressure. The reactors inner diameter is 28mm and it is made of stainless steel (SUS 32). Heating of the reactor was done by surrounding electric heater and

Fig. 2 Equilibrium Composition of Methylbenzenes Fig. 1 Experimental Apparatus Calculated at 750°K by API Data84)

Volume 13, No, 1, May 1971 118 Iwamura, Otani and Sato: ethylbenzene and ethyltoluene. vity as the reaction proceeds. Fig. 2 shows equilibrium composition on me- However, T-81 catalyst is completely regene- thylbenzenes and it indicates that the maximum rable by combustion of coke on the catalyst. conversion level of toluene per pass would be The ignition point of coke on the T-81 catalyst restricted to less than 58 mol%. was measured by a differential thermal analyser

2.3.3 Kinetics of the Reaction (Shimazu DT-2A) and it was found to be 430℃. Since the catalytic reaction with T-81 catalyst Also, the authors measured the combustion rate was carried out in a quite stable manner, the of coke on the T-81 catalyst under atmospheric study on the details of the reaction kinetics was pressure. The results are expressed by the fol- possible by employing a long test period with lowing formula. the same catalyst. rc ∞exp(-13,300/RT)nc3Ρ02 First of all, the external film diffusion around where the pellet was examined. Conversion was mea- nc: amount of coke on catalyst<2.0×10-3 sured with different amount of packed catalyst (mol-C/g-cal) with the other conditions remaining the same. T: Temperature= 450～530℃ The result showed that the film diffusion resistance P02: partial pressure of oxygen (atm) is negligibly small. This is also supported by For regeneration, oxygen less than two mol% independent calculation on film resistance. in nitrogen gas can be used. For combustion As the next step, internal pore diffusion within of coke without any damage on the catalyst ac- the pellet was studied. Pellets of different size tivity. including catalyst powder were made and con- 3 "TATORAY" Process version was measured. The result indicated that pore diffusion resistance is again negligible 3.1 Outline of the Process under operating conditions and consequently The flowscheme of the "TATORAY" pro- effectiveness factor of the pellet is approximately cess is shown in Fig. 378). Feed toluene, carried unity and calculation using Thiele modulus74) with makeup and recycle hydrogen, is heated, agreed with the result, thus the intrinsic rate first in an exchanger, then in a furnace. The for the catalyst could be measured. mixture then passes to a single adiabatic fixed The study on the effect of total pressure on the bed reactor at a moderate temperature and rate indicated that the surface reaction appears pressure. The reactor effluent is sent to the to be the rate controlling step. The initial re- separator and separated gas is recycled to the lative rate curve, strongly effected by partial feed. The liquid is sent first to a stabilizer and pressure of toluene, indicated that it will be then to the clay treater and to benzene, toluene, suitable to summarize the results by the Lang- xylene and trimethylbenzenes recovery columns. muir-Hinshelwood equation17). Recovered toluene and trimethylbenzenes are re- Temperature dependency of initial relative cycled to the feed. Suitable operating ranges are: rate were correlated by Arrhenius plot. The Total pressure 10～50atm activation energy for the case is calculated to be Temperature 350～530℃ 22.6kcal/mol-toluene. As a summary of kine- Mole ratio of H2/feed 5～20mol/mol tics study, initial rate of toluene disproportio- H2 concentration in recycle gas over 70 nation on T-81 catalyst, can be described by mole% rc ∞ KPT2/(1+KTPT)2 Process performances such as conversion, aro- where matic ring loss, over-all yield and on-stream periods are enterdependent. Higher pressure or higher temperature increases conversion per pass, but give larger ring loss, resulting in decrease in over- all yield. Lower ratio of H2/feed gives more For scaling up of a reactor, those rate data coke formation on catalyst; it is desirable to obtained by experiments can be directly adopted. maintain this ratio above 5. All these con- For other catalysts, several papers reported on ditions were well optimized in the plants67). kinetic study of toluene disproportionation75)～77). 3.2 Feature of the Process 2.4 Regeneration of T-81 Catalyst 3.2.1 Comparison with a Hydrodealkylation Disproportionation catalyst is decrease in acti- Process

Bulletin of The Japan Petroleum Institute bisproportionation of Toluene 119

Fig. 3 Flowscheme of "TATORAY" Process

In toluene dealkylation the methane molecule which is split off has a relatively low value com- pared with either toluene or benzene. Dispro- portionation, on the other hand, preserves the value of the entire toluene molecule by producing one molecule of xylene and one of benzene from two molecules of toluene. The comparison of economics for both processes mainly depends on the price statue of benzene, toluene and xylene. If hydrogen is not sufficiently available, this will also be important. 3.2.2 Product Distribution "TATORAY" process can afford a broad pro - duct distribution, according to the demands for benzene and xylene, by changing the reaction con- dition and feed composition. This is very useful from the standpoint of commercial production. The authors conducted experiments on the product distribution using experimental apparatus. Fig. 4 Effect of Trimethylbenzene Concentration in As equilibrium exists among methylbenzenes, Feed trimethylbenzenes were used for changing the ratio of methyl groups to a benzene ring. The about 4% trimethylbenzenes. result is shown in Fig. 4. The ratio of xylene A high content of p- and 0-xylenes and low to benzene is distributed from 0.7 to over 100; content of ethylbenzene are advantageous in when the concentration of trimethylbenzenes higher efficiency production of p- and 0-xylenes. is feed is 50mol%, the maximum yield of xylene 3.2.4 Trimethylbenzene and Light Paraffins is obtained. When the trimethylbenzene con- The author's experiments indicated that the centration is about 4mol% in feed, the ratio isomeric composition of trimethylbenzenes pro- of xylene to benzene would be unity. duced in "TATORAY" was approximately con- 3.2.3 Quality of the Products stant as 1,3,5-:1,2,4-:1,2,3=29:57:14, al- The received benzene and xylene have high though the feed composition of trimethylbenzenes product quality. Typical data, which were was different. This is in a good agreement with obtained from the "TATORAY" operation in the equilibrium composition of trimethylben- the Kawasaki plant80) are shown in Table 2 zenes at 700ｰK predicted by Hastings and Ame- and 3. The feed contains 96% toluene and miya72),81).(1,3,5-:1,2,4-:1,2,3=23:67:10).

Volume 13, No. 1, May 1971 Iwamura, Otani and Sato: 120

The authors also studied the effect of ethyl- was affected by the ethyltoluene concentration. toluene content in the feed. In the experiments, When the total C9 fraction, which is by- the concentration of ethyltoluene was changed produced in disproportionation of toluene and from 0 to 0.4 in total C9 aromatics, keeping the amounts to about 4mol% of fresh toluene, is concentration of C9 aromatics in feed as 4mol%. recycled to the feed stream, some equilibrium The results are shown in Fig. 5, as well as the content of ethyltoluene and ethylbenzene may change of ethylbenzene concentration of the exist. These are found to be 15mol% of ethyl- effluent. It also indicated that ethylbenzene toluene in C9 aromatics and 2mol% ethyl- benzene in C8 aromatic hydrocarbons. Table 2 Typical Analysis of "TATORAY" Benzene To study the effect of paraffinic and naph- thenic impurities in the feed, experiments were carried out increasing the paraffin content in the feed. Table 4 showing the results of the experiment, indicates that considerable amount

Table 3 Typical Analysis of "TATORAY" Xylene

Fig. 5 Effect of Ethyltoluene Concentration in Feed

Table 4 Reaction with Feed Containing Paraffins

Bulletin of The Japan Petroleum Institute Disproportionation of Toluene 121

of light paraffins in the feed were decomposed makeup hydrogen has over 93mol% purity. during the reaction. Furthermore, by the ad- BTX fractionation columns are integrated with dition of paraffins to the feed, no affect was de- hydrodealkylation79) and xylene separation tected on benzene ring rupture, toluene con- processes. The excellent quality of the products version and product distribution. This is ad- confirmed by commercial operation was already vantageous for production of high purity prod- shown in Table 2 and 3. The summary of ucts. plant operation is shown in Fig. 6. It shows 3.2.5 Yield and Hydrogen Consumption high stable conversion and good aromatic yield "TATORAY" process shows a good aromatic of over 97%. Regeneration of catalyst was yield and low hydrogen consumption. The carried out only once, and that after ten months typical ring rupture was calculated as only 0.7 from the initial start-up. mol% of toluene per pass, by analysis of the purge By this regeneration, catalytic activity was gas and liquid effluent. Also chemical con- completely recovered. The catalyst life is expec- sumption of hydrogen was calculated as 0.7mol% ted to be greater than two years. of toluene per pass. 3.4 Process Economics 3.3 Result of Plant Operation Typical "TATORAY" economic evaluation "TATORAY" plant in Toray is using toluene data are shown in Table 5, all of which were and C9 aromatics by-product as feed stocks and confirmed in the pilot plant work. The case of X/B=1 was also well supported by the commer- cial plant operation for over one year. For the other case, the similar long test runs in the com- mercial plant are scheduled in near future.

4 Other Disproportionation Process

4.1 "Xylene-Plus" "Xylene -Plus" is the disproportionation or transalkylation of aromatics developed by Sin- clair Oil Corporation (USA) 25),16),82),83).It was reported that commercial plant was built and in operation in Houston Refinery of At- lantic Richfield Company. The details of the process was not disclosed but the differences from "TATORAY" process would be in the following points. Fig. 6 Summary of Plant Operation (1) No hydrogen is used.

Table 5 Process Economics of "TATORAY"

Volume 13, No. 1, May 1971 122 Iwamura, Otani and Sato: Disproportionation of Toluene

(2) Moving bed is adopted because of neces- [Zeolite Catalyst] sity of catalyst regeneration. 32) "Zeolite Symposium" Chemistry and Chemical Industry, 21, 1233 (1968). This appears to make over-all selectivity of 33) Hara, N. et al., ibid., 21, 1274 (1968). aromatics lower than 95%. 34) Morita, Y., J. Japan Petrol. Inst., 10, 429 (1967). 35) Yamamoto, N. et al., ibid., 9, 441 (1966). 4.2 Other Applications 36) Yoshitomi, S., ibid., 13, (2), 92 (1970). A few other transalkylations may be interesting 37) British Petrol. Co. Ltd., Brit., 1, 081, 373 (1966). for industrial application but little has been 38) Socony Mobil Oil, U. S., 3, 173,855 (1965). 39) Socony Mobil Oil, U. S., 3, 250, 728 (1966). reported. 40) Socony Mobil Oil, Japan 65-25372 (1965). (1) Production of durene via tetramethyl- 41) Socony Mobil Oil, Brit., 1,022, 687 (1966). 42) Universal Oil Products, U. S., 3, 437, 710 (1969). benzene by disproportionation of trime- 43) Shell Oil, Brit., 1, 039, 246 (1966). thylbenzene. 44) Shell Oil, U. S., 3, 281, 483 (1966). 45) Shell International Research Maatschappij N. V., (2) Production of ethylbenzene from ethyl- Japan 70-21930 (1970). toluene. 46) Mobil Oil, Japan, 68-25967 (1968). (3) Transalkylation of ethylbenzene to di- 47) Mobil Oil, Japan, 69-26365 (1969). ethylbenzene. 48) Mobil Oil, Brit., 1,059, 524 (1967). 49) Mobil Oil, U. S., 3, 506, 731 (1970). Acknowledgement 50) Mobil Oil, U. S., 3, 377, 400 (1968). 51) Texaco Inc., U. S., 3, 476, 821 (1969). Authors wish to thank all members of CPX 52) Texaco Development Corp., Brit., 1, 191,094 (1970). Project Team for their assistance in this work. 53) Esso Research & Engineering Co., Japan, 70-23828 (1970). The authors also with to express their sincere 54) Toray Industries, Neth., 68-17615 (1969). appreciation to Mr. C. V. Berger of UOP Process 55) Toray Industries, Neth., 68-17616 (1969). 56) Toray Industries, Belg., 716, 016 (1968). Division for the discussion and advice on develop- 57) Toray Industries, Japan, 67-19082 (1967). ment and design of "TATORAY" process. 58) Toray Industries, Neth., 69-08032 (1969). 59) Toray Industries, Neth., 69-05766 (1969). References 60) Toray Industries, Belg., 732, 127 (1969). 61) Toray Industries, Belg., 724, 845 (1969). [Friedel-Crafts Catalyst] 62) Yashima, T. et al., 22nd Am. Meeting of Chem. 1) Standard Oil Co., (Indiana), U. S., 3, 006, 977 (1961). Soc. Japan, No. 12219 (1969). 2) Standard Oil Co., (Indiana), U. S., 3, 009, 004 (1961). 63) Morita, Y. et al., J. Chem. Soc. Japan, Ind. Chem. 3) Esso Research & Engineering Co., U. S., 3, 068, 302 Sec., 70, 1363 (1967). (1962). 64) Matsumoto, H. et al., J. Japan Petrol. Inst., 10, (8), 4) Standard Oil Co., (Indiana), U. S., 2, 954, 414 (1960). 572 (1967). 5) Universal Oil Products, U. S., 2, 886, 535 (1959). 65) Yashima, T. et al., Bull. Japan Petrol. Inst., 12, 106 6) Schriesheim, A., J. Org. Chem., 26, 3530 (1961). (1970). 7) British Petrol. Co. Ltd., Brit., 1,044, 490 (1966). [Others] 8) Lien, A. P., J. Am. Chem. Soc., 75, 2407 (1953). 66) Otani, S. et al., Japan Chem. Quarterly, IV-IV, 17 9) Sun Oil, U. S., 2, 739, 993 (1956). (1968). 10) Council of Scientific & Industrial Research, F. R., 67) Otani, S., Spectral Lecture Presented at National 1,408,076 (1965). Meeting of AIChE (1969). [Silica-Alumina Catalyst] 68) Ogawa, D. et al., J. Chem. Soc. Japan, Ind. Chem. 11) Houdry Process Corp., U. S., 2, 795, 629 (1957). Sec., 72, 2165 (1969). 12) Sun Oil, Canada, 532, 860 (1958). 69) Otani, S. et al., J. Japan Petrol. Inst., 13, (4), 13) Monsanto Co., U. S., 3, 221, 072 (1965). 282 (1970). 14) Shimobayashi, M. et al., J. Japan Petrol. Inst., 7, 70) Otani, S., Chem. Eng., July 27, 118 (1970). (1), 25 (1964). 71) Egan, C. J., J. Chem. Eng. Data, 5, 298 (1960). 15) Mitsubishi Petrochem., Japan 66-10, 807 (1966). 72) Hastings, S. H., Nikolson, D. E., ibid., 6, 1 (1961). 16) Ai, M. et al., J. Catalysis, 10, (8), 527 (1967). 73) Pitzer, K. S. et al., J. Am. Chem. Soc., 65, 803 17) Echigoya, E. et al., Kagaku Kogaku (Japan), 31, 386 (1943). (1967). 74) Thiele, F. W., Ind. Eng. Chem., 31, 916 (1939). 18) Izumi, Y. et al., J. Chem. Soc. Japan, 84, 699 (1963). 75) Ai, M. et al., Bull. Japan Petrol. Inst., 7, 46 (1965). 19) Izumi, Y. et al., Bull. Chem. Soc. Japan, 37, 1797 76) Izumi, Y. et al., Bull. Chem. Soc. Japan, 37, 1797 (1964). (1964). 20) British Petrol. Co. Ltd., Belg., 626, 446 (1963). 77) Matsumoto, H. et al., Bull. Japan Petrol. Inst., 10, 21) British Petrol. Co. Ltd., Fr., 1, 467, 723 (1967). 8 (1968). 22) Calif Research, U. S., 3,182,095 (1965). 78) Toray Industries, Belg., 716, 016 (1968). 23) Esso Research & Engineering Co., U. S., 3, 126, 422 79) Toray Industries, Belg., 732, 127 (1969). (1964). 80) Toray Industries, Belg., 731, 485 (1969). 24) Mobil Oil, Brit., 1,143,913 (1969). 81) Amemiya, T. et al., J. Japan Petrol. Inst., 3, (10), 25) Sinclair Research, U. S., 3, 116,340 (1963). 813 (1960). 26) Sinclair Research, U. S., 3, 437, 709 (1969). 82) Verdel, J. A., Oil & Gas J., June 9, 63 (1969). 27) Texaco Inc., U. S., 3, 476, 821 (1969). 83) Sinclair Research, U. S., 3,350,469 (1967). 28) Toray Industries, U. S., 3, 413, 374 (1968). 84) API selected values of physical and thermodynamic 29) Toray Industries, Japan 67-19, 082 (1967). properties of hydrocarbons and related compound, 30) Universal Oil Products, U. S., 2, 966, 529 (1960). Carnegie Press (1953). 31) Universal Oil Products, U. S., 3, 417, 157 (1968).

Bulletin of The Japan Petroleum Institute