[Regular Paper] 1. Introduction

Transcription

1 [Regular Paper] Methods of Activating Catalysts for Hydrodesulfurization of Light Gas Oil (Part 2) Catalytic Properties of CoMo/Al2O3 Presulfided by Polysulfides for Deep and Ultra-deep Hydrodesulfurization of Light Gas Oil (Received October 27, 2000) Hydrodesulfurization (HDS) of a straight run light gas oil (SRLGO, Sulfur: 1.50wt%) and ultra-deep HDS of a hydrotreated light gas oil (HDLGO, Sulfur: 0.045wt%) were carried out over a commercial Co-Mo/Al2O3 catalyst presulfided with a novel sulfiding agent, bis-(1,1,3,3-tetramethylbutyl)-polysulfide (CS-40). The HDS reac- Nm3/m3. The catalyst presulfided in situ with CS-40 showed comparable activity to the corresponding catalyst presulfided with dimethyldisulfide (DMDS), and higher activity than that presulfided with H2S in HDS of SRLGO. Catalyst presulfided ex situ with CS-40 plus sulfur showed significantly enhanced activity for HDS of alkyl-substituted dibenzothiophenes (DBTs). No significant changes in apparent activation energies for HDS of DBT were observed. In contrast, the catalyst ex situ sulfided with CS-40 plus sulfur had smaller activation energies for HDS of 4,6-dimethyldibenzothiophene (4,6-DMDBT). The ex situ sulfiding method using CS-40 plus sulfur may enhance the formation of active sites for hydrogenation and results in increased activity for indirect desulfurization of alkyl-substituted DBTs via hydrogenation of the aromatic ring of DBTs to form cyclohexylbenzenes. 1. Introduction More severe environmental regulations will require that the content of sulfur in diesel-engine fuel be lowered to less than 500ppm in Japan, the US and Europe1),2). Since further proposals suggest that the content of sulfur in diesel-engine fuel be lowered to less than 50ppm within the next five years2), deep or ultra-deep hydrodesulfurization (HDS) of light gas oil (LGO) has become an urgent priority. Methyl groups at the 4- and 6-positions of dibenzothiophene (DBT) greatly reduce the rate of HDS, and 4,6-dimethyldibenzothiophene (4,6-DMDBT) is very difficult to desulfurize even under deep desulfurization conditions1),3). Therefore, successful deep or ultra-deep desulfurization of LGO will depend on removal of alkyl-substituted DBTs such as 4,6-DMDBT from the light gas oil. Removal of alkyl-substituted DBTs is easier after the aromatic ring is hydrogenated4). Therefore, there are two main methods to achieve deep and ultra-deep desulfurization of LGO: Development of a catalyst with higher activity for both HDS and hydrogenation (HYD), and more effective activation of present catalysts by developing a new activation methods. * To whom correspondence should be addressed. The catalytic activity for HYD depends greatly on the activation procedure of the catalyst and the novel sulfidingagent-bis-(1,1,3,3-tetramethylbutyl)-polysulfide (CS-40), which has higher boiling point and lower toxicity than current agents, can be used to activate the catalyst for deep HDS of LGO5). The present study investigated the feasibility of CS-40 as a presulfiding agent for catalysts used in the HDS of light gas oil by evaluating conversions of alkyl-substituent DBTs present in LGO. The effects of the in situ and ex situ presulfiding procedure, and H2S, DMDS, and CS-40 sulfiding agent were investigated on the catalytic activity of Co-Mo/Al2O3. In addition, catalyst presulfided with CS-40 was evaluated for the HDS of an original hydrotreated light gas oil (sulfur content less than 500 ppm) to investigate the potential for ultra-deep HDS of light gas oil. 2. Experimental 2.1. Materials The feedstocks were a straight run light gas oil (SRLGO), obtained by direct distillation of Arabian light crude oil and an original hydro-desulrided light gas oil (HDLGO), containing 1.5wt% and 0.045wt% of total sulfur, respectively, with the properties shown

2 Table 1 Properties of Feedstocks Fig. 1 Temperature Program for Presulfiding in Table 1. Hydrogen and hydrogen sulfide in hydrogen (H2S 4.95vol%) were obtained from Tohei Kagaku Co. Dimethyldisulfide (DMDS, Kishida Chemicals Co., Ltd.) and the organic polysulfide, bis-(1,1,3,3- tetramethylbutyl)-polysulfide (C8H17SnC8H17, mean sul- Inc.), were used as sulfiding agents. The properties of flow rate of 0.01m3/h according to the temperature the sulfiding agents were described previously5). The program shown in Fig. 1. catalyst was commercially available Co-Mo/Al2O3 catalyst as described previously5), and supplied as 1/32- Total amounts of sulfur incorporated into the catalyst in the various procedures are summarized in Table 2. inch extrudates, which were crushed and screened to Catalytic Activity Test to 35 mesh granules. The pretreated catalyst was used in the same reactor 2.2. Presulfiding Procedure of Catalyst for hydrotreatment of a straight run light gas oil Based on the results obtained in the previous study5), (SRLGO) or an original hydrotreated light gas oil two types of sulfiding procedures, in situ and ex situ, (HDLGO). The HDS reactions were conducted under were used, as shown in Fig. 1. PSI-1: Presulfiding in situ with a mixture of 4.95 pressure, 3.0MPa; LHSV, h-1; Gas/Oil ratio, vol% H2S in a stream of H2 under atmospheric pressure. 125Nm3/m3. Samples were collected from a gas-liq- The fresh catalyst (2.54g) was packed into a uid separator. The reaction temperature was then bench-scale reactor, with a diameter of 20mm and containing changed and after 4-5h, further sampling was carried a thermocouple tube with a diameter of 2mm in out. Products were analyzed by gas chromatography the longitudinal direction, as described elsewhere3). with an FPD detector (GC-14A, Shimadzu Corp.) and First, the mixed gas of hydrogen sulfide and hydrogen AED detector (HP-5921A). The identification of sul- compounds was carried out by comparing the chro- fur m3/h. Then, the temperature was increased from room matogram of GC-FPD and GC-AED, as described previously3). Further, it was assumed that the HDS reac- PSI-2 and PSI-3: Presulfiding in situ with the light gas oil spiked with 2.5wt% of sulfur in the form of dimethyldisulfide (DMDS) or CS-40. First, the fresh catalyst packed in the reactor was heated in the hydrogen gas stream under 3.0MPa from room temperature was introduced under the conditions: H2/Oil, 125Nm3/ m3; LHSV, 4,0h-1. The temperature was raised to in Fig. 1. PSE-1 to 3: Presulfiding ex situ with CS-40, or CS- 40 plus elemental sulfur. The fresh catalyst (1g) was immersed in an acetone solution (2ml) of CS-40 (0.5 g) or in a CS2 solution of CS-40 ( g) and elemental sulfur ( g) at room temperature. Then the solvent was removed from the mixture in a draft tion. The catalyst was packed into the reactor and tion of DBTs could be treated as a first-order reaction and the pseudo-first-order reaction rate constant (khds) for HDS of DBTs was determined according to the previous method6),7).

3 Table 2 Total Sulfur Incorporated into the Catalyst (Stotal) in Various Sulfiding Procedures Stoichiometric sulfur amount in the catalyst is mol/g.cat when all active metals are present as Co9S8 and MoS2. a) Originating from DMDS or CS-40. b) Originating from SRLGO. Fig. 2 GC-FPD Chromatogram of Polyaromatic Sulfur-containing Compounds in SRLGO Fig. 3 Removal Ratio of Total Sulfur in SRLGO at Several Temperatures 3. Results 3.1. HDS of SRLGO on in situ Sulfided Catalyst Figure 2 shows the GC-FPD chromatogram of SRLGO and indicates the sulfur compounds. Dibenzothiophenes (DBTs) are the major sulfur compounds refractory to HDS in light gas oil3),6). DBTs decrease in reactivity to HDS as follows: nonsubstituted DBT and alkyl-substituted DBTs with no substituents at the 4- and 6-positions, alkyl-substituted DBTs with a substituent at the 4-position, and alkylsubstituted DBTs with substituents at both 4- and 6- positions3),6). Figure 3 shows the removal of total sulfur in SRLGO with HDS reaction temperature over Co-Mo catalysts presulfided with H2S (PSI-1), DMDS (PSI-2) and CS-40 (PSI-3). The activity of the catalyst presulfided with CS-40 was higher than that presulfided with H2S or DMDS. This result is slightly different from the previous results obtained5), in which the catalyst presulfided by H2S showed the highest activity. To interpret the result, the effects of sulfiding agents were compared on the conversions of three sulfur compounds resistant to HDS, DBT, 4-methyldibenzothiophene (4-MDBT) and 4,6-DMDBT. Figure 4 shows the Arrhenius plots of pseudo-first-order rate constants (khds) for HDS of DBT, 4-MDBT and 4,6- Numbers in parentheses refer to peaks in Fig. 2. Fig. 4 Arrhenius Plots of khds for HDS of DBTs in SRLGO on the Catalyst Presulfided in situ by H2S, DMDS and CS-40 DMDBT. As expected, the values of khds for DBT decreased slightly in the order of PSI-1 (H2S)>PSI-2 contrast, the values of khds for 4-MDBT and 4,6- DMDBT over catalyst using DMDS (PSI-2) and CS-40

4 Table 3 Activation Energies for HDS of DBTs in SRLGO on in situ Presulfided Catalyst [kcal/mol] (PSI-3) were greater than those over catalyst using H2S (PSI-1). These results indicate that CS-40 is an effective presulfiding agent for the catalytic HDS of LGO, as for HDS of DBT5). The activation energies for HDS of DBT, 4-MDBT and 4,6-DMDBT are summarized in Table 3. These values for PSI-1 agree with previously reported values6). Moreover, the sulfiding agent did not significantly influence the activation energies for HDS of DBT, 4-MDBT and 4,6-DMDBT using catalyst presulfided in situ HDS of SRLGO on ex situ Sulfided Catalyst Arrhenius plots of pseudo-first-order rate constants for HDS of DBT, 4-MDBT and 4,6-DMDBT over catalysts presulfided ex situ with CS-40 or CS-40 plus sulfur are shown in Fig. 5 a)-c). The results using catalyst sulfided in situ with CS-40 (P51-3) are also shown for reference. Almost no difference in HDS rate constants were found for catalyst prepared by PSI-3 and PSE-1, although the amount of used CS-40 in the former was about 7 times more than that in the latter. This indicates that the ex situ sulfiding method with CS-40 is an efficient method for activating the catalyst. The catalytic activities for HDS of 4-MDBT and 4,6- DMDBT of catalysts using CS-40 plus sulfur (PSE-2 and PSE-3) were remarkably higher than those using only CS-40 (PSE-1 and PSI-3), especially at lower temperatures. These results suggest that using the mixture of CS-40 plus sulfur is superior to using only CS- 40 for activating the catalyst. This also implies that the ex situ sulfiding method is more favorable to the removal of alkyl-substituted DBTs, and thus for the ultra-deep desulfurization of LGO. The apparent activation energies for HDS of DBT, 4-MDBT and 4,6- DMDBT are summarized in Table 4. There was no significant difference in the activation energies for HDS of DBT in all cases. In contrast, comparing PSE-2 or PSE-3 with PSE-1 or PSI-3 the activation energies for HDS of 4-MDBT and 4,6-DMDBT Fig. 5 Arrhenius Plots of khds for HDS of DBTs in SRLGO on ex situ and in situ Presulfided Catalysts Table 4 Activation Energies for HDS of DBTs in SRLGO on ex situ Presulfided Catalyst [kcal/mol] 3.3. Ultra-deep HDS of HDLGO on ex situ Sulfided Catalyst HDS of an original hydrotreated light gas oil (HDLGO, total sulfur: 0.045%) was carried out over catalysts pretreated according to the procedures PSE-1, PSE-2 and PSE-3, respectively, to assess the potential of the ex situ sulfiding method for ultra-deep desulfurization of LGO. Figure 6 shows the GC-FPD chromatogram for sulfur compounds in HDLGO. After deep HDS, 4,6-DMDBT, several trimethyl-substituted dibenzothiophenes (C3-DBT) and a few other sulfur compounds still remain in the oil. Figure 7 a)- d) show the conversions of the four main sulfur compounds at several reaction temperatures, with the results over the catalyst sulfided in situ with H2S (PSI- 1) for reference. The catalytic activities were higher for catalysts prepared by ex situ than by in situ presulfidation. Moreover, the catalyst pretreated by PSE-3 showed almost the same activity as that by PSE-2, and much greater than those by PSE-1 and PSI-1. The Arrhenius plots of khds for HDS of 4,6-DMDBT in HDLGO are shown in Fig. 8. The values of khds

5 Fig. 6 GC-FPD Chromatogram of Polyaromatic Sulfur-containing Compounds in HDLGO Table 5 Activation Energies for HDS of 4,6-DMDBT in HDLGO [kcal/mol] decreased in the order of PSE-2=PSE-3>PSE-1> PSI-1. The apparent activation energies for HDS of 4,6-DMDBT are summarized in Table 5. Catalyst presulfded by PSE-2 and PSE-3 had smaller activation 2kcal/mol). These results indicates that the ex situ sulfiding method using a mixture of CS-40 plus sulfur is a very promising procedure for preparing CoMo catalyst for the ultra-deep desulfurization of light oil. 4. Discussion Numbers in parentheses refer to peaks in Fig. 6. No significant difference in apparent activation energies for HDS of DBT was observed as shown in Tables 3 and 4. This indicates that the sulfiding agent did not affect the HDS mechanism of DBT. Both DMDS and polysulfides were easily decomposed with hydrogen under atmospheric pressure on hydrotreatment catalysts, primarily forming methanethiol (CH3SH) at lower temperatures, and subsequently producing H2S9),10). The decomposition of polysulfides starts at able to consider that, in a sulfiding process using CS- 40, CS-40 is primarily transformed to H2S in an atmosphere of H2, and the catalyst is sulfided by the H2S produced. Thus, if an adequate amount of sulfur is incorporated into the catalyst, the structure of the sulfided catalyst is little influenced by the sulfiding agent, and the activity for HDS of DBT depends upon the number of active sites produced. This could explain the previous results indicating that the catalytic activity for HDS of DBT increased with the amount of labile sulfur Fig. 7 Conversions of 4,6-DMDBT (48) and C3-DBTs (56, 59 and 62) in HDLGO at Several Temperatures Fig. 8 Arrhenius Plots of khds for HDS of 4,6-DMDBT in HDLGO

6 formed in the sulfiding process5). On the other hand, the ex situ sulfiding method with CS-40 plus sulfur resulted in a significant increase in catalytic activity for the conversion of alkyl-substituted DBTs such as 4,6- DMDBT, as shown in Figs. 5 and 7. The apparent activation energies for HDS of 4,6-DMDBT over catalysts prepared by PSE-2 and PSE-3 were less than those by PSI-3 and PSE-1, as shown in Tables 4 and 5. There are two routes for HDS of DBTs: Direct desulfurization producing biphenyls (BPs), and indirect desulfurization producing cyclohexylbenzenes (CHBs)1),4),12). Further, the removal of alkyl-substituted DBTs is easier after the aromatic ring is hydrogenated4),12),13). Kinetic analysis of the HDS of DBTs has suggested that the activation energy for the formation of CHBs is smaller than that for the formation of BPs1). Thus, the ex situ sulfiding method using CS-40 plus sulfur may have significantly enhanced the hydrogenation capacity of the catalyst, i.e., the activity via the indirect desulfurization route. The indirect desulfurization route becomes more important for catalysts prepared by PSE-2 or PSE-3, which lower the apparent activation energy for HDS of 4,6-DMDBT, close to the apparent activation energy for HDS of 4,6-DMDBT to form 3,3'-DMCHB. The catalytically active sites involved in hydrotreatment reactions are generally viewed as sulfur vacancies (or anion vacancies), but whether both hydrogenation and hydrogenolysis occur on the same type of site or on different types of sites remains unclear14). A dual-site mechanism involving Mo atoms at different oxidation levels, the higher oxidation state for the hydrogenation site and the lower one for the hydrogenolysis site, may be involved15),16) The "Rim-Edge" model for MoS2 was proposed to interpret the hydrogenation and hydrogenolysis sites, in which the catalyst particle can be described as a stack of several discs17). The top and bottom discs are associated with the rim sites. The discs "sandwiched" between the top and bottom discs are associated with the edge sites. Sulfur hydrogenolysis occurs on both the rim and edge sites, whereas hydrogenation occurs exclusively on the rim sites. Therefore, the number of stacking layers becomes greater when the sulfiding temperature is higher, resulting in lower hydrogenation activity. This could explain our previous results showing that lower final sulfiding temperature resulted in higher HYD activity5). Thus, when CS-40 plus sulfur (PSE-2 and PSE-3) were used, CS-40 was decomposed into several sulfur species at lower temperatures, and was used to sulfide the catalyst. On the other hand, the impregnated elemental sulfur cannot be transformed to H2S at lower temperatures even in a H2 atmosphere. Part of the elemental sulfur is transformed to H2S at higher temperatures, so the sulfiding process can continue, resulting in higher activity of the catalyst than when using only CS- 40. Further, other sulfur species together with H2S may be formed when using CS-40, and will participate in the formation of active sites, especially for HYD. A polysulfide (di-tert-nonylpentasulfide, TNPS) acti- average number of layers of 1.6. These values are significantly lower than those determined for the H2S-H2 ber of active sites at the top and bottom of the MoS2 discs, i, e., the rim sites, increases and results in higher hydrogenation activity. An increase in hydrogenation can be expected for the case of CS-40, as in the case of TNPS. 5. Conclusions Catalyst presulfided in situ with CS-40 showed comparable activity to the corresponding catalyst presulfided with DMDS, and a higher activity than that presulfided with H2S, for the HDS of light gas oil. The ex situ sulfiding method using CS-40 plus sulfur (PSE- 2 and PSE-3) significantly enhanced the catalytic activity for HDS of alkyl-substituted DBTs. No significant changes in apparent activation energies for HDS of the sulfiding method decreased the apparent activation energies for HDS of 4,6-DMDBT in the order of PSI-1 CS-40 plus sulfur enhances the formation of active sites for hydrogenation and results in increased indirect desulfurization of alkyl-substituted DBTs via hydrogenation and subsequent desulfurization of the aromatic ring of DBTs to form cyclohexylbenzenes. References 1) Kabe, T., Ishihara, A., Qian, W., "Hydrodesulfurization and Hydrodenitrogenation," Kodansha and Wiley-VCH, Tokyo (1999). 2) Octane Week, 8, (15), Apr. 14 (1997). 3) Kabe, T., Ishihara, A., Tajima, H., Ind. Eng. Chem. Res., 31, (6), 1577 (1992). 4) Zhang, Q., Ishihara, A., Kabe, T., Sekiyu Gakkaishi, 39, (6), 410 (1996). 5) Yamada, Y., Qian, W., Wang, G., Yoda, Y., Ishihara, A., Kabe, T., Sekiyu Gakkaishi, 44, (4), 217 (2001). 6) Kabe, T., Akamatsu, K., Ishihara, A., Otsuki, S., Godo, M., Qian, W., Zhang, Q., Yamada, S., Sekiyu Gakkaishi, 42, (3), 150 (1999). 7) Kabe, T., Qian, W., Funato, A., Okoshi, Y., Ishihara, A., Phys. Chem. Chem. Phys., 1, 921 (1999). 8) Kabe, T., Akamatsu, K., Otsuki, S., Godo, M., Zhang, Q., Qian, W., Ind. Eng. Chem. Res., 36, 5146 (1997). 9) van Gestel, J., Leglise, J., Duchet, J.-C., J. Catal., 145, 429 (1994). 10) Prada Silvy, R., Grange, P., Delannay, F., Delmon, B., Appl. Catal., 46, 113 (1989). 11) Berrebi, G., US Pat

7 12) Amorelli, A., Amos, Y. D., Halsig, C. P., Kosman, J. J., Jonker, R. J., Wind, M. de, Vrieling, J., Hydrocarbon Processing, 1992, June, ) Landau, M. V., Berger, D., Herskowitz, M., J. Catal., 158, 236 (1996). 14) Stanislaus, A., Cooper, B., Catal. Rev.-Sci. Eng., 36, (1), 75 (1994). 15) Moreau, C., Aubert, C., Durand, R., Zminita, N., Geneste, P., Catal. Today, 4, 117 (1988). 16) Moreauand, C., Geneste, P., "Theoretical Aspects of Heterogeneous Catalysis," ed. by Moffat, J. B., Van Nostrand Reinhold, New York (1990), p ) Daage, M., Chianelli, R. R., 149, 414 (1994). 18) Labruyere, F., Dufresne, P., Lacroix, M., Breysse, M., Catalysis Today, 43, 111 (1998). 19) Dufresne, P., Brahma, N., Labruyere, F., Breysse, M., Catalysis Today, 29, 251 (1996). Keywords Presulfidation, Polysulfide, Cobalt molybdenum catalyst, Light gas oil, Ultra-deep hydrodesulfurization