1748
Ind. Eng. Chem. Res. 1998, 37, 1748-1754
Hydrodesulfurization of Catalytic Cracked Gasoline. 3. Selective Catalytic Cracked Gasoline Hydrodesulfurization on the Co-Mo/ γ-Al2O3 Catalyst Modified by Coking Pretreatment S. Hatanaka*,† and M. Yamada Department of Applied Chemistry, Graduate School of Engineering, Tohoku University, Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan
O. Sadakane Petroleum Research Laboratory, Mitsubishi Oil Company Ltd., 4-1 Ohgimachi, Kawasaki-ku, Kawasaki 210-0867, Japan
The interaction of olefins with the hydrodesulfurization (HDS) catalyst surface was studied by the time course of catalytic cracked gasoline (CCG) HDS on the Co-Mo/γ-Al2O3 catalyst. Olefin hydrogenation activity was decreased drastically in the beginning 30 h. After 75 h of time on stream, to wash the catalyst surface, toluene was flown into the reactor at room temperature and this toluene wash recovered some olefin hydrogenation activity. The oligomers of CCG olefins were found in the effluent of toluene wash. It was suggested that oligomerized and polymerized olefin (coke) selectively deactivated the olefin hydrogenation active site. The possibility of improving the HDS selectivity of Co-Mo/γ-Al2O3 by carbonaceous deposit was investigated for HDS reactions of CCG and model compounds. The catalyst with coking pretreatment after sulfiding showed the higher CCG HDS selectivity than that of the catalyst without coking pretreatment. Thiophene HDS in the presence of diisobutylene or 1-octene was also carried out, and the relative deactivation by the coking pretreatment was in the following order, isoolefin hydrogenation > thiophene HDS > n-olefin hydrogenation. The improvement of CCG HDS selectivity by the coking pretreatment was considered to be caused by the selective deactivation of isoolefin hydrogenation active site. Introduction CCG (catalytic cracked gasoline) which is produced from heavy gas oil or atmospheric residue by fluid catalytic cracking units in refineries, is one of the major components of motor gasoline. As CCG contains a high level of sulfur, CCG HDS (hydrodesulfurization) is a prospective process from the environmental point of view. To keep a high octane value, high HDS selectivity (higher activity for HDS and lower activity for olefin hydrogenation) is expected, because CCG contains 2040 vol % olefins. A few studies of CCG HDS on the CoMo/γ-Al2O3 catalysts have been carried out to focus the attention on the HDS selectivity (Desai et al., 1994; Hatanaka et al., 1997). Concerning the HDS selectivity of Mo catalyst, the effects of some modifications and pretreatments on the HDS activities and hydrogenation activities have been reported on model compounds. For example, by the modification of MoO/carbon catalyst with phosphate, kthiophene-HDS/kbutene-HG decreased to 1/10 of the original catalyst (Bouwens et al., 1988). The effects of regeneration conditions (i.e. temperature and presence of steam) on the catalyst selectivity was studied on Co-Mo/γAl2O3 catalysts using thiophene/cyclohexene/cyclohexane as a feedstock (Arteaga et al., 1987). In this study, * Corresponding author. Telephone: +81-44-344-3128. Fax: +81-44-344-3645. † Present address: Petroleum Research Laboratory, Mitsubishi Oil Co. Ltd., 4-1 Ohgimachi, Kawasaki-ku, Kawasaki 210-0867, Japan.
kthiophene-HDS/kcyclohexene-HG was decreased by the regeneration at 700 °C with steam. The study of the order of sulfidation and reduction treatment (Silvy et al., 1989) were also carried out on Co-Mo/γ-Al2O3 catalysts with the feedstock of thiophene/cyclohexene/cyclohexane. In this study, HDS activity was varied together with HG activity and the HDS selectivity was not improved. These studies seem to give a pessimistic view for selective CCG HDS, because any modifications and pretreatments of the catalysts reported so far always have decreased HDS selectivity (kHDS/kHG). However, the olefins used in these studies were limited to n-olefins (mostly butenes) or cycloolefins. Therefore, it is still unclear that these results can be applied for HDS of CCG, in which iso-C6-C10 olefins are major olefin components. To make clear the possibility of improving the HDS selectivity, the present series has tried to investigate the relationship between HDS activity and isoolefin hydrogenation activity. In part 1 of the present series (Hatanaka et al., 1997a), it was found that C6-C10 olefins strongly inhibit thiophene HDS. In part 2 (Hatanaka et al., 1997b), the inhibiting effect of olefins was further studied and three different types of active sites (i.e. HDS, n-olefin hydrogenation, and isoolefin hydrogenation) were proposed to exist on Co-Mo/γAl2O3 catalyst. Therefore, if it is possible to deactivate the type 3 active site (i.e. isoolefin hydrogenation active site) without changing the activity of the type 1 active site (HDS active site); it is expected to make a highly
S0888-5885(97)00735-5 CCC: $15.00 © 1998 American Chemical Society Published on Web 05/04/1998
Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998 1749 Table 1. Properties of CCG composition,a vol % saturates aromatics olefins (olefin structure breakdownb) 1-olefins H2CdCR2 H2CdCHR internal olefins HRCdCR2 HRCdCHR isoolefin ratio in C7 olefin, % total sulfur,c ppm density, g/cm3 at 15 °C distillation temperature,d °C IBP 10% 30% 50% 70% 90% EP research octane value
CCG-A
CCG-C
41.9 33.6 30.4 2.3 5.0 9.8 13.3 78.9 229 0.778
30.6 32.8 36.6 2.6 5.8 12.1 16.1 80.8 157 0.788
48 88 110 136 165 201 231 87.0
113 123 131 138 148 162 181 87.1
a By ASTM D-1319. b Type ratio was calculated by 1H NMR data on the assumption that the R2CdCR2 type olefin was not included. The NMR peak assignment for the olefin types, ppm by TMS standard: H2CdCR2, 4.50-4.80; H2CdCHR, 4.80-5.10 and 5.606.00; HRCdCR2, 5.10-5.25; HRCdCHR, 5.25-5.60. c By ASTM D-3120. d By ASTM D-2887.
selective HDS catalyst, because isoolefins are the major olefin components in CCG. The present work tried to make clear the possibility that the type 3 active site is selectively controlled by some catalyst pretreatments. The effects of oligomer and coke on Co-Mo/γ-Al2O3 were examined by the study of CCG HDS. To understand the effects of pretreatments on three types of active sites, variously pretreated catalysts were also supplied for thiophene HDS in the presence of diisobutylene or 1-octene. The catalyst surface on which the coke was deposited after CCG HDS was studied by electron spectroscopy for chemical analysis (ESCA) and FT-IR/DRA (powder diffuse reflectance). Experimental Section 1. Feedstocks. (a) CCG. Two kinds of CCG (CCGA and CCG-C) produced from low sulfur atmospheric residue were used here. CCG-A was the same as that
used in part 1 of this series (Hatanaka et al., 1997a). The properties of CCG are summarized in Table 1. (b) Model Compounds. Commercial grade thiophene was dissolved in the mixture of toluene (80 mol %) and olefins (20 mol %) at 2.83 × 10-4 mol/mol. Olefins used here were commercial grade diisobutylene and 1-octene without further purification. 2. Analyses for Feedstock and Product. The same analytical methods as those from parts 1 and 2 were used. The isoolefin ratio in C7 olefins of CCG was measured by GC equipped with a 50 m PONA column. 3. Catalysts. (a) Catalyst 1. The same batch of the homemade catalyst, as used in part 1, was pretreated and supplied for thiophene HDS study. (b) Catalyst 2. A commercial catalyst was pretreated in several ways and was used for the HDS study of CCG-C. Catalyst 2 which had been used for diesel fuel HDS in the refinery was also tested with or without regeneration treatment (calcined at 550 °C for 1 h). Physical and chemical properties of these catalysts, which are fresh, pretreated, and spent on HDS reaction, are summarized in Table 2. 4. Catalyst Pretreatment and HDS Reaction Procedure. A 4 mL aliquot of the catalyst crushed to 0.6-1.0 mm particles was packed in the same fixed bed microflow reactor as those used in parts 1 and 2. (a) Sulfiding Pretreatment. All of the catalysts were sulfided in situ at 250 °C for 2 h and then 300 °C for 2 h in the stream of a dibutyl-disulfide/straight run naphtha (bp 80-160 °C)/hydrogen mixture (sulfur, 3 wt %). (b) Aging Pretreatment. Following the sulfiding, some of the catalysts were aged at 300 °C for 48 h in the stream of a straight run naphtha (total sulfur 250 wt ppm)/hydrogen mixture. (c) Coking Pretreatment. Before or after sulfiding, the catalyst was treated in the stream of 1-methylnaphthalene (13 mol %) + cyclohexene (26 mol %) + argon (61 mol %) at 350 °C for 0.5 h and 0.5 MPa. Immediately after these sequential catalyst pretreatments, the feedstock was flown in the reactor with hydrogen. The reaction conditions are shown in Table 3. 5. Elemental Analysis of the Catalyst. The pretreated catalysts and the used catalysts were taken out
Table 2. Physical and Chemical Properties of the Fresh, Pretreated, and Postreaction Catalystsa catalyst analyses run no.
catalyst no.b
catalyst pretreatmentc
1 2 3 4 5 6 7 8 9 10 11 12
1-0 2-0 1-1 1-2 1-3 1-2c 2-2c 2-3c 2-4c 2-5c 2-6c 1-1t 1-2t 1-3t
(fresh) (fresh) SF SF + AG SF + AG + CO SF + AG SF + AG SF + AG + CO CO + SF + AG refinery spent, SF refinery spent, RG + SF + AG SF SF + AG SF + AG + CO
feedstock
before reaction before reaction before reaction CCG-A CCG-C CCG-C CCG-C CCG-C CCG-C thiop + olef + tole thiop + olef + tol thiop + olef + tol
carbon,d wt %
surface area, m2/g
pore volume, cm3/g
1.0 2.4 3.7 7.0 5.7 9.3 8.9 8.8 5.5 2.9 2.9 7.9
200 224 151 159 168 164 176 182 167 138 155 141 159 167
0.61 0.54 0.46 0.47 0.46 0.45 0.46 0.43 0.38 0.36 0.38 0.43 0.47 0.44
NO adsorption, mL/g of catalyst
5.2 4.0
5.1 5.1 4.4
a Metal content of fresh catalyst: catalyst 1, 15.0 wt % MoO , 4.6 wt % CoO; catalyst 2, 17.0 wt % MoO , 4.5 wt % CoO. b The meaning 3 3 of the catalyst number (X-Yz): X, catalyst 1 or 2; Y, pretreatment (0, fresh; 1, SF; 2, SF + AG; 3, SF + AG + CO; 4, CO + SF + AG; 5, refinery spent, SF; 6, refinery spent, RG + SF + AG); z, feedstock (c, CCG; t, thiophene). c SF, sulfiding; AG, aging; CO, coking; RG, regeneration. d Fresh catalyst base. e Thiophene + olefin + toluene.
1750 Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998 Table 3. Reaction Conditions for HDS Reaction for given reaction CCG HDS (time course)
CCG HDS
catalyst feedstock
catalyst 1 CCG-A
catalyst 2 CCG-C
gas gas/feed ratio feed time, h temperature, °C pressure, MPa contact time
hydrogen 85 NL/L 130 300 0.4 LHSV, 5.0 h-1
hydrogen 85 NL/L 24 300 0.4 LHSV, 5.0 h-1
HDS (model compounds) catalyst 1 80 mol % toluene; 20 mol % olefina thiophene: 2.83 × 10-4 mol/mol hydrogen 1.6 mol/mol ca. 100 175 (diisobutylene), 190 (1-octene) 1.3 W/F, 0.17-0.79 (g of catalyst min)/mol
a Diisobutylene or 1-octene. Diisobutylene composition: 2,4,4-trimethyl-1-pentene, 74.9 mol %; 2,4,4-trimethyl-2-pentene, 20.6 mol %; others, 4.5 mol %.
from the reactor and washed by toluene in a Soxhlet extractor for 24 h. The elemental analysis of the catalyst surface was carried out by ESCA (Shimadzu ESCA 850 M). The amount of the carbon deposited on the catalyst was measured by CHN elemental analyzer. 6. NO Adsorption on the Catalyst. The pretreated catalysts and the used catalysts were taken out from the reactor and washed by toluene in a Soxhlet extractor for 24 h. The amount of NO adsorbed on the catalyst was measured by the same pulse technique reported by Kasahara et al. (1995). The catalyst was treated at 300 °C in the flow of 5% H2S/H2. The pulse of NO was introduced at room temperature, and effused NO was detected by TCD. FT-IR spectra of NO adsorbed on the catalyst were obtained by the DRA method proposed by Yamada and Obara (1990). Finely powdered catalyst was placed in the IR cell, and 5% H2S/H2 was flown at 300 °C. NO was adsorbed at room temperature and then purged by He. Spectra of the adsorbed NO were recorded by a JEOL JIR-100FT spectrometer. 7. Definition. Since thiol and sulfide are produced by the HDS reaction in the presence of olefin, two different HDS (%) are defined in the following equations respectively in the present paper. Olefin hydrogenation (%) is also defined as follows. (a) CCG HDS.
total HDS (%): (1 - product sulfur/feed sulfur) × 100 olefin hydrogenation (%): (1 - olefin content in product/ olefin content in feed) × 100 (b) Model Compounds HDS.
total HDS (%): (1 - product sulfur/feed sulfur) × 100 thiophene HDS (%): {1 - (thiophene and THT sulfur in product)/ feed sulfur} × 100 THT ) tetrahydrothiophene olefin hydrogenation (%): (C8 paraffin content in product/ olefin content in feed) × 100 Results and Discussion 1. Deactivation of the Catalyst Activities by the Time Course of CCG HDS. In part 1 (Hatanaka et al., 1997a), it was newly found that C6-C10 olefins in
CCG show the strong inhibiting effect on thiophene HDS. In part 2 (Hatanaka et al., 1997b), through further studies of inhibiting effects of olefins on HDS, it was found that the interaction between the reactants (thiophene, n-olefin, and isoolefins) and the active sites for HDS and olefin hydrogenation strongly depend on the kind of reactants. The effects of H2S and Co on HDS and olefin hydrogenation activities were also studied. And it was noted that there is a clear difference between the HDS active site and the isoolefin hydrogenation active site. That is, HDS was inhibited by H2S and promoted by Co, while isoolefin hydrogenation was promoted by H2S and a little inhibited by Co. By these results, it has been proposed that three types of active sites, for thiophene HDS (type 1), n-olefin hydrogenation (type 2), and isoolefin hydrogenation (type 3), exist. This difference suggests the possibility of the selective CCG HDS by deactivating the isoolefin active site without affecting the HDS activity, because isoolefins are the major olefin components in CCG. With respect to the selective deactivation of type 3 active site, we noticed with the findings of part 2 that the type 3 (isoolefin hydrogenation) active site strongly interacts with isoolefin, resulting in the formation of oligomers and polymers of isoolefin. These higher molecules may finally be irreversibly deposited on the active site as a coke. If it is possible that type 3 active site is selectively deactivated by the oligomers and coke, it leads to selective CCG HDS. If the deposition of oligomers and/or coke proceeds during CCG HDS reaction, the catalyst activities for HDS and olefin hydrogenation may change during the reaction time course. In other words, the change is expected to reflect the effect of the deposition of oligomers and coke on these active sites. To understand the effect of the deposited species on the HDS activity and olefin hydrogenation activity, the time course of these activities during CCG HDS was studied over catalyst 1-2 (sulfided and aged). Figure 1 shows the time course of total HDS percentage and olefin hydrogenation percentage during the HDS of CCG-A for 130 h. HDS activity decreases gradually, while olefin hydrogenation activity decreases drastically in the first 30 h and stabilizes in the latter 100 h. After 75 h of time on stream, the reactor was cooled to room temperature and toluene was flown into the reactor at a liquid hourly space velocity (LHSV) 5.0 h -1 for 24 h (toluene wash). After the toluene wash, CCG HDS reaction was restarted. It is noted that olefin hydrogenation activity recovers much more than the total HDS activity by the toluene wash. To understand the recovery of olefin hydrogenation activity, the effluent
Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998 1751
Figure 1. Time course of CCG HDS: Catalyst 1-2. Reaction conditions: temperature, 300 °C; pressure, 0.4 MPa; LHSV, 5 h-1; H2/feed ratio, 85 NL/L. Feedstock: CCG-A.
of the toluene wash was collected and condensed by evaporation. The residue obtained was analyzed by IR. Since the IR spectrum showed the presence of aliphatic hydrocarbons, rather than aromatics, this residue was considered to be oligomers of the olefins contained in CCG. The molecular weight of the residue was from 300 to 800 by GPC analysis. From these results, it is suggested that olefins contained in CCG are oligomerized and polymerized to selectively deactivate the olefin hydrogenation active sites. After the toluene wash, olefin hydrogenation activity has recovered to some extent but still remained to be the low level. This suggests that some of the oligomer has polymerized to coke which is not soluble in toluene, and the coke effectively and irreversibly deactivates olefin hydrogenation active site. Through the whole run of 130 h, total HDS activity decreases gradually. On the other hand, olefin hydrogenation activity decreases quickly, especially in the first 30 h, and stabilizes in the latter 100 h. From these results, it is presumed that there are two different active sites for olefin hydrogenation. That is, one is easily deactivated by deposited species (derived from oligomerized and polymerized olefin) and the other is hardly deactivated by deposited species. 2. Effects of Coking Pretreatment on CCG HDS Selectivity. By the time course of CCG HDS, it was pointed out the possibility that hydrogenation activity of Co-Mo/γ-Al2O3 is selectively deactivated by the deposit of some oligomers and/or coke on the catalyst surface. This possibility suggests the improvement of the HDS selectivity of the catalyst. To understand the deactivation by the deposited species in more detail, various catalyst pretreatments which may produce carbonaceous deposit were tried. Three different combinations of the pretreatments were carried out on the fresh catalyst 2 (commercial). They were sulfiding + aging (catalyst 2-2), sulfiding + aging + coking (catalyst 2-3), and coking + sulfiding + aging (catalyst 2-4). The coking pretreatment was done by using a mixture of cyclohexene + 1-methylnaphthalene (Arteaga et al., 1987). The catalyst properties are listed in Table 2. CCG-C was hydrotreated over these catalysts pretreated variously. The result of CCG HDS is shown in Figure 2. Catalyst 2-2 gives a higher total HDS percentage than the catalysts with coking (catalysts 2-3 and 2-4), and this may be brought by less coke deposi-
Figure 2. Effects of catalyst pretreatments on CCG HDS selectivity. Reaction conditions: temperature, 300 °C; pressure, 0.4 MPa; LHSV, 5 or 10 h-1 (b*: 8 h-1); H2/feed ratio, 85 NL/L. Feedstock: CCG-C (sulfur, 157 wt ppm; olefin, 36.6 vol %).
tion of catalyst 2-2 than that of catalyst 2-3 and catalyst 2-4. Concerning the HDS selectivity, it is noted that catalyst 2-3 shows higher HDS selectivity than catalyst 2-2. However, catalyst 2-4 shows the same HDS selectivity level as catalyst 2-2. These phenomena suggest that the coking pretreatment much more affects the olefin hydrogenation active site than the HDS active site in the case of catalyst 2-3. And it is also shown that the coking pretreatment before sulfiding does not improve the HDS selectivity. The effects of coke were further examined by using a refinery spent catalyst (catalyst 2-5), which was used for 1 year in the diesel fuel HDS process. This catalyst, having 8.8 wt % coke deposition, was supplied for the CCG HDS activity test, and the results are shown in Figure 2. This spent catalyst also shows high HDS selectivity. However, the high HDS selectivity is lost by a regeneration procedure (catalyst 2-6). This result suggests that coke deposit improves the HDS selectivity. This improvement may be brought by the selective inhibiting effect of the coking pretreatment on isoolefin hydrogenation activity, because the GC analysis for C7 olefins in CCG shows that 80 wt % olefins are isoolefins (Table 1). This figure is very close to the isoparaffin percentage (79.0%) in CCG (Desai et al., 1994). However, it is still unclear how the coke affects three active sites, that is, the HDS active site, the n-olefin hydrogenation site, and the isoolefin hydrogenation active site. 3. Mechanism of the HDS Selectivity Improvement by Coke. To understand the improvement of CCG HDS selectivity by the coking pretreatment, the effects of coking pretreatment on the three types of active sites (thiophene HDS, isoolefin hydrogenation, n-olefin hydrogenation) were studied by the model reaction of thiophene HDS in the presence of diisobutylene or 1-octene. Three different pretreatments were carried out on the catalyst 1 (homemade). These pretreatments were as follows, sulfiding (catalyst 1-1), sulfiding + aging (catalyst 1-2), and sulfiding + aging + coking (catalyst 1-3). A different amount of coke was deposited on the catalysts by these pretreatments, and the amounts of carbon originated by the coke were 1.0 wt % by sulfiding, 2.4 wt % by sulfiding + aging, and 3.7 wt % by sulfiding + aging + coking (Table 2). Thiophene HDS was carried out over catalyst 1-1, 1-2, or 1-3 in the presence of diisobutylene or 1-octene (Figure 3). Total HDS and thiophene HDS are little reduced by the aging pretreatment but are much
1752 Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998
Figure 3. Total HDS and thiophene HDS in the presence of olefin.
Figure 4. Effects of catalyst pretreatments on the HDS selectivity (thiophene HDS in the presence of diisobutylene). Reaction conditions: temperature, 175 °C; pressure, 1.3 MPa; H2/feed ratio, 1.6 mol/mol; catalyst/feed, 0.17-0.79 (g of catalyst min)/mol. Feedstock: thiophene concentration, 2.83 × 10-4 mol/mol; toluene, 80 mol %; diisobutylene, 20 mol %.
Figure 5. Effects of catalyst pretreatments on the HDS selectivity (thiophene HDS in the presence of 1-octene). Reaction conditions: temperature, 190 °C; pressure, 1.3 MPa; H2/feed ratio, 1.6 mol/ mol; catalyst/feed, 0.17-0.60 (g of catalyst min)/mol. Feedstock: thiophene concentration, 2.83 × 10-4 mol/mol; toluene, 80 mol %; 1-octene, 20 mol %.
reduced by the coking pretreatment. Thiol and sulfide yields are also reduced by the coking pretreatment. The diisobutylene hydrogenation percentage is plotted against thiophene HDS in Figure 4. Little difference is observed between the selectivity of catalyst 1-1 and catalyst 1-2. However, it is noted that diisobutylene hydrogenation of catalyst 1-3 is much lower than those of catalyst 1-1 and catalyst 1-2. The 1-octene hydrogenation percentage is plotted against thiophene HDS in Figure 5. The selectivities of catalyst 1-1 and catalyst 1-2 are also the same. However, 1-octene hydrogenation of catalyst 1-3 is higher than those of catalyst 1-1 and catalyst 1-2. The effects of coking pretreatment on the three
different types of active sites (thiophene HDS, type 1; n-olefin hydrogenation, type 2; and isoolefin hydrogenation, type 3) are as follows, type 3 > type 1 > type 2. This order also suggests that the coking pretreatment effectively reduces isoolefin hydrogenation activity but hardly reduces n-olefin hydrogenation activity. The improvement of HDS selectivity by the coking pretreatment can be explained by the selective deactivation of isoolefin hydrogenation activity by coke. Concerning the control of olefin hydrogenation activity, the present results are partly consistent and partly inconsistent with the previous reports. For example, the ratio of thiophene HDS activity to butene hydrogenation activity (kHDS/kHG) was reported to be decreased
Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998 1753 Table 4. ESCA Measurement of the Catalyst Surface before and after CCG HDS elemental compositiona catalyst
condition
Al (2p)
Mo (3d)
Co (2p)
C (1s)
S (2p)
1-2 1-2c
after pretreatment after CCG HDS
100 100
8.7 8.1
8.3 8.0
66.0 122.3
19.2 16.0
a
Calculated with Al (2p) as standard.
by the modification with phosphate as mentioned in the Introduction (Bouwens et al., 1988). In the present study, in the case of thiophene HDS in the presence of 1-octene, HDS selectivity decreased by the coking pretreatment. However, it is noted that the ratio of thiophene HDS activity to isoolefin hydrogenation activity (kHDS/kiso-HG) is increased by the coking pretreatment. After the reaction, the carbon content of the used catalysts were measured (Table 2). The carbon content of the catalyst 1-3 is 7.9 wt % and is much larger than that of the catalysts 1-1 and 1-2 (2.9 wt %). This may be explained as follows. Polymerized materials were produced by coking pretreatment, and they covered isoolefin hydrogenation active sites. As some of the polymerized materials were toluene soluble and they were washed away by toluene in a Soxhlet extractor, the detected amount of carbon on the catalyst 1-3 before the HDS reaction was 3.7 wt %. During HDS reaction, toluene soluble polymerized materials were further polymerized and changed into toluene insoluble polymers. The amount of NO uptake of the spent catalysts were measured (Table 2). The carbon content are 2.9 wt % on catalysts 1-1 and 1-2 and 7.9 wt % on catalyst 1-3. The amounts of NO uptake on the catalysts are 5.1 mL/g on catalysts 1-1 and 1-2 and 4.4 mL/g on catalyst 1-3. The ratio of NO uptake on catalyst 1-3 to that on catalyst 1-2 is 0.86, and this figure is close to the ratio of the 1-octene hydrogenation reaction rate constant on catalyst 1-3 to that on catalyst 1-2 (0.76: k of catalyst 1-3/k of catalyst 1-2). 4. Surface Characterization. (a) Elemental Analysis of the Catalyst Surface. The amount of carbon deposited on the catalyst by the pretreatment and the reaction was measured by CHN elemental analysis, and the results are shown in Table 2. A 2.4 wt % amount of carbon is on catalyst 1-2 (after the pretreatments), and 7.0 wt % of carbon is on catalyst 1-2c (after CCG HDS reaction). The elemental composition of the catalyst surface was measured by ESCA, and the amounts of Mo, Co, C, and S, calculated with Al (2p) as a standard, are shown in Table 4. The amount of carbon on the catalyst 1-2c surface is almost double that of catalyst 1-2, and this result of ESCA is in accordance with the results of CHN elemental analysis. The amount of exposed Mo (as Al intensity is 100) decreases from 8.7 to 8.1 by CCG HDS, and this decrease means the coverage of Mo by coke deposition. (The amount of Co also decreases from 8.3 to 8.0; however, the accuracy of the Co amount is poor because of the satellite peak of oxygen.) (b) NO Adsorption Study on the Catalyst. To examine the effect of coke in more detail, NO uptake on catalyst 1-2 and catalyst 1-2c was measured by a pulse technique. The results are shown in Table 2. Before the CCG HDS reaction, NO uptake is 5.2 mL/g of catalyst. After the reaction, it decreases to 4.0 mL/g of catalyst. This result suggests that about 20% of the
Figure 6. FT-IR spectra of adsorbed NO on the catalyst.
coordinatively unsaturated site is covered by coke during CCG HDS. To observe the coordinatively unsaturated site (active sites) of the catalyst subjected to CCG HDS in more detail, an FT-IR study of NO adsorbed on catalysts 1-2 and 1-2c was carried out. The spectra obtained are shown in Figure 6. Three bands appear in the spectrum at 1850, 1800, and 1700 cm-1. The band assignments have been reported as follows: 1850 cm-1 for NO on Co, 1700 cm-1 for NO on Mo, and 1800 cm-1 for both NO on Co and NO on Mo (Okamoto et al., 1981; Yamada and Obara, 1990). Comparing the relative band intensity before CCG HDS with that after CCG HDS, the intensity of the Co site band decreases relatively against the Mo site band by CCG HDS reaction. The ESCA result, mentioned above, only shows that the catalyst is covered by coke (Table 4), and the FT-IR study further makes clear that the Co site is more covered by coke than the Mo site. If the Co site is selectively deactivated by coke, it is expected that HDS activity is decreased and n-olefin hydrogenation is increased, because it was found that thiophene HDS is promoted by Co but n-olefin hydrogenation is inhibited by Co, as has been reported in part 2 (Hatanaka et al., 1997b). Therefore, the present results do not simply mean the selective deactivation of the Co site by the coking pretreatment. It is considered that some other effect is also brought about by the coking pretreatment. Conclusions The present work tried to make clear the possibility of improving HDS selectivity. In the time course of CCG HDS, olefin hydrogenation activity decreased more than HDS activity and recovered to some extent by use of a toluene wash. IR analysis of the effluent of the toluene wash suggested that the deactivation of the olefin hydrogenation active site is caused by the selective deposition of oligomers and polymers on the active site. NO adsorption and ESCA studies supported the selective coke deposition on the catalyst active sites. To understand the deactivation by the deposited species, various catalyst pretreatments which may produce carbonaceous deposit were applied to Co-Mo/ γ-Al2O3. And these catalysts were supplied for the HDS selectivity studies. In CCG HDS, HDS selectivity was improved by the coking pretreatment. In the thiophene HDS in the presence of model olefins, it was also found that the coking pretreatment selectively deactivates
1754 Ind. Eng. Chem. Res., Vol. 37, No. 5, 1998
isoolefin hydrogenation activity. It was considered that the isoolefin hydrogenation active site is selectively deactivated by the coking pretreatment, leading to an improvement of the HDS selectivity for CCG HDS, because isoolefins are the major olefin components in CCG. Literature Cited Arteaga, A.; Fierro, J. L. G.; Grange, P.; Delmon, B. CoMo HDS Catalyst: Simulated Deactivation and Regeneration, Role of Various Regeneration Parameters. In Catalyst Deactivation; Delmon B., Froment G. F., Eds.; Elsevier Science Publishers: Amsterdam, 1987; p 59. Bouwens, S. M. A. M.; Vissers, J. P. R.; Beer, V. H. J.; Prins, R. Phosphorus Poisoning of Molybdenum Sulfide Hydrodesulfurization Catalysts Supported on Carbon and Alumina. J. Catal. 1988, 112, 401. Desai, P. H.; Lee, S. I.; Jonker, R. J.; De Boer, M.; Vrieling, J.; Sarli, M. S. Reduce Sulfur in FCC Gasoline. Fuel Reformulation. 1994 (Nov/Dec), 43. Hatanaka, S.; Sadakane, O.; Yamada, M. Hydrodesulfurization of Catalytic Cracked Gasoline. 1. Inhibiting Effects of Olefins on HDS of Alkyl-(Benzo)-Thiophenes Contained in Catalytic Cracked Gasoline. Ind. Eng. Chem. Res. 1997a, 36, 1519.
Hatanaka, S.; Sadakane, O.; Yamada, M. Hydrodesulfurization of Catalytic Cracked Gasoline. 2. The Difference between HDS Active Site and Olefin Hydrogenation Active Site. Ind. Eng. Chem. Res. 1997b, 36, 5110. Kasahara, S.; Miyabe, S.; Shimizu, T.; Takase, H.; Yamada, M. Effect of Catalysts Pretreatments on Hydrodesulfurization Activity and Surface Structure of Co-Mo/Al2O3. Sekiyu Gakkaishi 1995, 38, 81. Okamato, Y.; Katoh, Y.; Mori, Y.; Imanaka, T.; Teranishi, S. No Adsorption Sites in Sulfided MoO3/Al2O3 Catalysts. J. Catal. 1981, 70, 445. Silvy, R. P.; Grange, P.; Delmon, B. Activation of CobaltMolybdenum Hydrodesulfulization Catalysts: Influence of the Sulfidation Procedure on the Physico-Chemical Properties and Catalytic Activity. In Catalysts in Petroleum Refining; Trimm, D. L., Ed.; Elsevier Science Publishers: Amsterdam, 1989; p 233. Yamada, M.; Obara, T. Hydrogenation by Co-Mo/Al2O3 Catalyst (Part 6). FT-IR/DRA Study of Nitric Oxide Adsorbed on Sulfided Catalysts. Sekiyu Gakkaishi 1990, 33, 221.
Received for review October 22, 1997 Revised manuscript received February 23, 1998 Accepted February 27, 1998 IE970735L