Al2O3

Journal of the Japan Petroleum Institute, 53, (5), 292-302 (2010)

292

[Regular Paper]

Hydrodesulfurization Activity of Co–Mo/Al2O3 Catalysts Prepared with Citric Acid: Post-treatment of Calcined Catalysts with High Mo Loading Nino RINALDI, Masahiro YOSHIOKA, Takeshi KUBOTA, and Yasuaki OKAMOTO＊ Dept. of Material Science, Shimane University, Matsue 690-8504, JAPAN (Received January 27, 2010)

Co_Mo/Al2O3 catalysts with high loading of Mo (20 wt% Mo) were prepared by the addition of citric acid (CA) using a simultaneous impregnation method or a post-treatment method. The catalyst activity was investigated for the hydrodesulfurization (HDS) of dibenzothiophene. TEM, Co K-edge XANES, NO adsorption capacity, and surface area measurements were conducted to characterize the catalysts. The catalysts were also investigated by the chemical vapor deposition technique using Co(CO)3NO to prepare Co/MoS2 and Co/Co_MoS2 catalysts. Addition of citric acid by the post-treatment method to calcined (Co)_Mo/Al2O3 oxide catalysts significantly enhanced the edge dispersion of MoS2 particles and the catalytic activity of Co/Mo/Al2O3 and Co_Mo/Al2O3, whereas citric acid addition by the simultaneous impregnation method hardly affected the edge dispersion. The highest HDS activity was obtained with a CA/Mo mole ratio of 1.5-2 for the post-treatment catalysts. The posttreatment method may be applicable to the regeneration of deactivated HDS catalysts. Keywords Catalyst preparation, Citric acid, Hydrodesulfurization, Cobalt molybdenum catalyst, Dibenzothiophene

Introduction Increasingly stricter environmental legislation controlling transportation fuel quality is extremely important for the protection of the environment. In particular, the allowable sulfur content in transportation fuels will be reduced to near zero in the near future. Consequently, further improvements in the activity, selectivity, and/or stability of hydrodesulfurization (HDS) catalysts are still needed. Supported nickel or cobalt_molybdenum sulfide catalysts are widely used in industrial HDS processes1)～3). The active sites of these catalysts are the so-called Co(Ni)_Mo_S phases, in which Co(Ni) occupies the edge sites of highly dispersed MoS2 particles2). Therefore, the formation of a greater amount of the Co(Ni)_Mo_S phase is essential to increase the activity of HDS catalysts1)～5). The addition of citric acid to HDS catalyst preparations as a chelating agent is a new method for preparing highly dispersed metal sulfide catalysts to improve the performance of HDS catalysts. The use of citric acid in the preparation of Co_Mo and Ni_Mo/Al2O3 catalysts with calcination of the dried samples at 773 K resulted in Co_Mo catalyst with superior HDS activity due to increased dispersion of MoS2 particles, but the resultant Ni _ Mo catalyst had inferior hydrogenation and ＊＊

To whom correspondence should be addressed. E-mail: [email protected]

hydrodenitrogenation activity, both compared to the catalysts prepared using ammonia6). The addition of citric acid is also effective, even after calcination, for the preparation of Ni_W catalysts for the hydrogenation of 1-methylnaphthalene7). Citric acid may increase the dispersion of active metals because of the increase in the viscosity of the impregnation solution8). Moreover, the presence of citric acid in the impregnation solution prevented the agglomeration of Mo, so Mo dispersion remained high throughout the preparation process, resulting in increased HDS activity9). Preparation of ultra-deep Co_Mo HDS catalysts using citric acid as a chelating agent combined with phosphoric acid without calcination (just after drying) found greatly improved HDS activity due to increased Co coverage on the edges of MoS2 particles, because the sulfidation temperature of Co was increased by the formation of a Co citrate complex10)～13). Addition of citric acid increased the activity of TiO2_ZrO2-supported Ni_Mo catalysts for the HDS of dibenzothiophene14). Addition of citric acid to a simultaneous impregnation solution also increased the HDS activity of Co _ Mo/Al2O3 and Co_Mo/B2O3/Al2O3 catalysts, possibly by increasing the dispersion of MoS2 particles as well as the increased Co coverage of the MoS2 edges15) . X-ray photoelectron spectroscopy (XPS) showed that the sulfidation of Mo is improved by the addition of citric acid11),16). X-ray absorption fine structure (XAFS) showed that the formation of surface Mo_citrate complexes results in

J. Jpn. Petrol. Inst., Vol. 53,

No. 5, 2010

293

increased sulfidation of Mo at a high Mo content for unpromoted Mo catalysts17). Our previous study found that the activity of Co_Mo/ B2O3/Al2O3 calcined catalysts with 8.7 wt% Mo for the HDS of thiophene was greatly improved by treatment with citric acid (post-treatment method)18). These catalysts were even more active for HDS than similar catalysts prepared by a simultaneous impregnation method. The hydrogenation activity of calcined Co_Mo catalysts was improved by treatment with ethylenediaminetetraacetate19). More recently, we showed that welldispersed Mo_CA complexes are formed on the support at high Mo content during the preparation of Mo/Al2O3 catalysts by a post-treatment method with citric acid17). Therefore, addition of citric acid by the post-treatment method improves the edge dispersion of MoS2 particles on the catalyst surface, resulting in significant increases in the catalytic activity for the HDS of thiophene and the sulfidation of Mo for Mo/Al2O3 sulfide catalysts at high Mo content17). In the present study, we prepared highly active Co_ Mo HDS catalysts with high Mo loading (20 wt% Mo) by the post-treatment method for comparison with Co_ Mo catalysts prepared by a conventional calcination technique or by a simultaneous impregnation method with citric acid. To further investigate the effect of citric acid addition on the characteristics of HDS catalysts, we used a less reactive sulfur compound, dibenzothiophene (DBT), as a reactant. The prepared catalysts were characterized by various physicochemical techniques, including chemical vapor deposition (CVD) using Co(CO)3NO as a probe molecule20)～23). Full Co occupancy of the MoS2 edge sites of supported MoS2 catalysts was established with the CVD technique to measure the maximum potential activity of Co_Mo catalysts. 1.

Experimental

1. 1. Catalyst Preparation A series of Mo/Al2O3 catalysts with various citric acid contents was prepared by an impregnation technique. γ-Al2O3 (JRC-ALO-7: 180 m2・g -1) was impregnated with an (NH4)6Mo7O24・4H2O (AHM) aqueous solution, followed by calcination at 773 K for 5 h. The Mo content of the calcined Mo/Al2O3 oxide catalyst was 20 wt% Mo. The calcined Mo/Al2O3 catalyst was impregnated with various amounts of aqueous solutions of citric acid (CA) (CA/Mo mole ratio＝0-3.5), followed by drying in an electric oven under static air at 383 K for 16 h (without calcination). The pH of the impregnation solution was not adjusted. This catalyst preparation technique is denoted here as the posttreatment method17),18). The catalysts are designated here as CA/Mo/Al. Another series of unpromoted Mo catalysts (20 wt% Mo) with various citric acid contents

was prepared by simultaneous impregnation of AHM and CA, followed by drying in an oven under static air at 383 K for 16 h (without calcination). The prepared catalysts are designated here as Mo_CA/Al. A series of conventional Co _ Mo/Al2O3 catalysts (0-8 wt% Co, 20 wt% Mo) with various Co contents was prepared by a double impregnation method (Mo first and then Co) using AHM and Co(NO3)2・6H2O (cobalt nitrate) aqueous solutions. An aliquot of the catalyst was calcined at 773 K for 5 h after each step. Then, the calcined Co_Mo/Al2O3 oxide catalysts were impregnated with citric acid (CA/Mo mole ratio＝2.0) and oven-dried under static air at 383 K for 16 h (without calcination) to prepare CA/CoMo/Al. For comparison, Co_Mo/Al2O3 catalysts were also prepared by a simultaneous impregnation method. γ-Al2O3 was simultaneously impregnated with AHM, cobalt nitrate, and CA to prepare CoMo_CA/Al, which was then dried at 383 K for 16 h (without calcination). The Mo, Co and CA contents of the simultaneous impregnation catalysts were the same as those of the post-treatment catalysts. The prepared catalysts were sulfided in a 10% H2S/ H2 flow at 673 K for 2 h. The sulfidation procedures were described previously17),18). A CVD technique using Co(CO)3NO as a precursor was applied to introduce Co into the sulfided catalyst17),18),22). Briefly, the sulfided catalyst was first evacuated at 673 K for 1 h and subsequently exposed for 5 min at room temperature to Co(CO)3NO vapor at 273 K. After 10 min evacuation at room temperature, the sample was sulfided again at 673 K for 1.5 h. The prepared catalyst is designated as, for example, Co/Mo_CA/Al, for the Mo _ CA/Al catalyst subjected to the CVD technique. The abbreviations for the preparation methods and catalysts are summarized in Table 1 for clarity. 1. 2. Reaction Procedure Prior to the HDS reaction, all catalysts were sulfided ex-situ at 673 K. The catalysts were evaluated for the HDS of DBT using a 200 ml autoclave. For each run, the autoclave was charged with 0.1 g of the catalyst and a solution containing 1% DBT (Aldrich, 98%) dissolved in 50 ml decalin, using a glove bag filled with N2 to avoid contact with air. After loading the reactant solution and the catalyst into the autoclave vessel, the vessel was placed in the autoclave in air. The air in the vessel was replaced with N2 and then H2 five times each at room temperature. The H2 pressure was increased and then the temperature was raised to 603 K (12 K/min) with stirring (300 rpm) of the reaction mixture. The reaction was carried out for 3-8 h at 603 K and initial H2 pressure of 1.43 MPa. The reaction products were analyzed by off-line gas chromatography with a hydrogen-flame ionization detector (FID, Shimadzu GC-14B). The HDS product was predominantly biphenyl with a small amount of cyclo-


No. 5, 2010

294 Table 1

List of Abbreviations for the Preparation Methods and Catalysts

Abbreviation

Preparation method and catalyst

CA Mo/Al (calc) Co/Mo/Al (calc) Mo_CA/Al Co/Mo_CA/Al CA/Mo/Al Co/CA/Mo/Al CoMo/Al (dried) Co/CoMo/Al (dried) CoMo/Al (calc) Co/CoMo/Al (calc) CoMo_CA/Al Co/CoMo_CA/Al CA/CoMo/Al Co/CA/CoMo/Al

Citric acid Mo/Al2O3 sulfide catalysts prepared from calcined Mo/Al2O3 oxide catalysts Co_Mo/Al2O3 sulfide catalysts prepared from Mo/Al (calc) by the CVD technique Mo/Al2O3 sulfide catalysts prepared from the simultaneous impregnation catalysts with CA Co_Mo/Al2O3 sulfide catalysts prepared from Mo_CA/Al by the CVD technique Mo/Al2O3 sulfide catalysts prepared from the post-treatment catalysts with CA Co_Mo/Al2O3 sulfide catalysts prepared from CA/Mo/Al by the CVD technique Co_Mo/Al2O3 sulfide catalysts prepared from dried Co_Mo co-impregnation catalysts without CA, CoMo_CA/Al without CA Co_Mo/Al2O3 sulfide catalysts prepared from CoMo/Al (dried) by the CVD technique Co_Mo/Al2O3 sulfide catalysts prepared from calcined Co_Mo/Al2O3 oxide catalysts, CA/CoMo/Al without CA Co_Mo/Al2O3 sulfide catalysts prepared from CoMo/Al (calc) by the CVD technique Co_Mo/Al2O3 sulfide catalysts prepared from the simultaneous impregnation catalysts with CA Co_Mo/Al2O3 sulfide catalysts prepared from CoMo_CA/Al by the CVD technique Co_Mo/Al2O3 sulfide catalysts prepared from the post-treatment catalysts with CA Co_Mo/Al2O3 sulfide catalysts prepared from CA/CoMo/Al by the CVD technique

hexylbenzene (1-2%). Therefore, only the HDS reaction rate was evaluated. Preliminary experiments showed that the reaction was first order with respect to the concentration of DBT. The reproducibility of the HDS activity as expressed by the rate constant was usually better than ±5%. 1. 3. Characterizations The physical properties of sulfided Mo/Al2O3 (specific surface area and pore volume) prepared by citric acid addition were evaluated using a BELSORP-18 PLUS (BEL Japan) nitrogen physisorption apparatus. Prior to the measurements, the sample was degassed at 673 K for 5 h under vacuum. Surface areas were determined by physical adsorption of N2 at liquid nitrogen temperature, using the BET equation. Pore volumes were obtained from the desorption curve of the isotherm using the procedure developed by Dollimore and Heal24). NO adsorption on sulfided CA/Mo/Al and Mo_CA/ Al was measured at room temperature by a pulse technique after the samples had been cooled in a 10% H2S/ H2 stream17),25),26). The sulfided catalyst was flushed with a high purity He stream before periodic introductions of 10% NO/He pulses. The reproducibility was usually better than ±5% of the total amount of NO adsorbed. Transmission electron microscopy (TEM) images of Mo/Al2O3 (20 wt% Mo) with citric acid addition were taken with an electron microscope (JEM-2010, JEOL) using an acceleration voltage of 200 keV26). The catalyst sample was evacuated and sealed in a glass tube without exposure to air. The presulfided sample was suspended in acetone and placed on a specimen grid in a glove bag filled with N2 gas to avoid as much contact with air as possible. The sample was then transferred in a stream of N2 to a sample holder attached to the microscope. The distributions of MoS2 particle size and stacking number were calculated over ca. 700 particles in five arbitrarily chosen areas. Co K-edge XAFS spectra were measured for sulfided

The activities of (△) Mo_CA/Al and (▲) CA/Mo/Al are also shown for comparison. Fig. 1●Catalytic Activities of (○) Co/Mo_CA/Al and (●) Co/CA/ Mo/Al as a Function of CA/Mo Mole Ratio for the HDS of DBT

Co_Mo/Al2O3 with or without citric acid addition and for reference compounds in the fluorescence or transmission mode, depending on the Co content, at room temperature at BL-9C in the Photon Factory of the Institute of Material Structure Science, High Energy Accelerator Research Organization (KEK-IMSS-PFAR) with 2.5 GeV ring energy and 380-350 mA stored current (proposal No. 2009G026, 2008G200). The synchrotron radiation was monochromatized by a Si (111) double crystal monochromator. 2.

Results and Discussion

2. 1.

Effect of Citric Acid Addition on the HDS Activity of Co/Mo/Al2O3 Figure 1 depicts the activities of Co/Mo/Al2O3 catalysts prepared by citric acid addition for the HDS of DBT as a function of CA/Mo mole ratio. Cobalt was


No. 5, 2010

295

Fig. 2●Amount of Co Deposited by CVD on (○) Co/Mo_CA/Al and (●) Co/CA/Mo/Al as a Function of CA/Mo Mole Ratio

added by the CVD technique to the Mo/Al2O3 catalysts to obtain the maximum potential activity22),23), which is attained by a full occupancy of MoS2 edges with Co_ Mo_S without blocking of active sites by catalytically inactive Co sulfides. The addition of Co by the CVD technique greatly enhanced the rate constant of the HDS reaction compared to that for Mo/Al2O3. With the simultaneous impregnation method, the rate constant of Co/Mo_CA/Al was only slightly increased by citric acid addition up to CA/Mo＝2, and then decreased by further addition of citric acid. On the other hand, the rate constant of the post-treatment catalyst (Co/CA/Mo/Al) was significantly increased by the addition of citric acid up to a CA/Mo mole ratio of 1.5 and leveled off around CA/Mo＝1.5-2, followed by a considerable decrease with further addition of citric acid. The rate constant of Co/CA/Mo/Al was increased by about 2.5 times by the post-treatment with citric acid (CA/Mo mole ratio＝2). The addition of citric acid by the post-treatment method clearly increased the catalytic activity for the HDS of DBT much more than addition by the simultaneous impregnation method at a high Mo content (20 wt% Mo). The amounts of Co anchored by the CVD process were evaluated for Co/Mo_CA/Al and Co/CA/Mo/Al to obtain information on the dispersion of MoS2 particles. The amounts of Co are depicted in Fig. 2 as a function of CA/Mo mole ratio. With the post-treatment method, the amount of Co on Co/CA/Mo/Al was significantly increased up to a CA/Mo mole ratio around 1-1.5 and decreased with further addition of citric acid. On the other hand, the amount of Co was decreased by the addition of citric acid using the simultaneous impregnation method. The amount of Co on Co/CA/Mo/Al (CA/Mo＝1.5) was about 1.5 times greater than that on Co/Mo_CA/Al, showing that the dispersion of MoS2 particles was considerably increased by the post-

Fig. 3●Amounts of NO Adsorption on (○) Mo _ CA/Al and (●) CA/Mo/Al as a Function of CA/Mo Mole Ratio

treatment with citric acid. With the unpromoted Mo catalyst, the rate constant for the HDS of DBT was only slightly increased but by about 1.5 times by the addition of citric acid (CA/Mo＝ 2) using the post-treatment method (Fig. 1). In agreement with the HDS of thiophene, the activity of CA/ Mo/Al was about twice as high as that of Mo _CA/Al for the HDS of DBT17). With the promotional effect of Co, the activity for the HDS of DBT was increased much more by the addition of Co with respect to Mo_ CA/Al or CA/Mo/Al, compared to the HDS of thiophene. The different promotional effects of Co between the HDS reactions of DBT and of thiophene were probably caused by the great difference in the hydrogen pressure used, 1.4 MPa for DBT and 20 kPa for thiophene. The maximum potential activity of the post-treatment catalyst (Co/CA/Mo/Al) was considerably higher than that of the simultaneous impregnation catalyst (Co/Mo _ CA/Al) for HDS reactions of both DBT and thiophene. 2. 2. Characterization of Mo/Al with Citric Acid NO adsorption measurements were conducted to evaluate the dispersion of MoS2 particles. The amounts of NO adsorbed on Mo_CA/Al and CA/Mo/Al are depicted in Fig. 3 as a function of CA/Mo mole ratio. The NO adsorption capacity was almost unchanged by the addition of citric acid by the simultaneous impregnation method, indicating that the edge d i s p e r s i o n o f M o S 2 p a r t i c l e s i s n o t m o d i fi e d . However, the NO adsorption capacity on CA/Mo/Al was drastically increased after addition of citric acid by the post-treatment method. The amount of NO adsorption capacity of CA/Mo/Al was almost twice as large as that of Mo/Al (calc) (post-treatment catalyst at CA/Mo＝0) or Mo_CA/Al. The NO adsorption and the Co content in the CVD catalyst were not proportional, in particular for the simultaneous impregnation catalysts, which may suggest that carbonaceous deposits


No. 5, 2010

296

Fig. 5●Distributions of Stacking Number of MoS2 Particles on (white bar) Mo_CA/Al and (gray bar) CA/Mo/Al (CA/Mo mole ratio＝1.5) as Observed by TEM

Fig. 4●Representative TEM Images of (a) Mo_CA/Al and (b) CA/ Mo/Al with CA/Mo Mole Ratio＝1.5

formed by the decomposition of citric acid during the sulfidation differently affect the adsorption sites of NO and Co(CO)3NO molecules on the edge sites of MoS2 particles. Adsorption of the latter molecules is likely to be more strongly hindered by carbonaceous deposits because of the larger molecular size. The amount of NO adsorption of CA/Mo/Al was about doubled by the addition of citric acid. In approximate agreement with the NO adsorption capacity, the HDS activity of the unpromoted CA/Mo/Al catalyst was increased about 1.5 times, compared to Mo/Al (calc). Figure 4 shows representative TEM images for CA/ Mo/Al and Mo_CA/Al with citric acid (CA/Mo＝1.5). The MoS2 particles were well dispersed by the posttreatment method compared with the simultaneous impregnation method. TEM images for both catalysts were analyzed to obtain more detailed information about the morphology of MoS2 particles. The distributions of stacking number and slab length of MoS2

Fig. 6●Distributions of Slab Length of MoS2 Particles on (white bar) Mo _CA/Al and (gray bar) CA/Mo/Al (CA/Mo mole ratio＝1.5) as Observed by TEM

particles are presented in Figs. 5 and 6, respectively. The average stacking number and slab length are summarized in Table 2. The average stacking number of MoS2 slabs was not affected by the addition method of citric acid. Almost 40% of MoS2 particles were still present as single slabs in both catalysts, as shown in Fig. 5. The most abundant slab length was in the range of 2-4 nm for the Mo_CA/Al catalyst. Table 2


No. 5, 2010

297 Table 2●Average Stacking Number and Slab Length of MoS2 Particles on Mo/Al2O3 (20 wt% Mo) Prepared by the Addition of Citric Acid (CA/Mo＝1.5) as Observed by TEM Catalyst

Slab length [nm]

Stacking number

Mo_CA/Al CA/Mo/Al

3.0 2.3

1.9 1.8

Table 3●Physical Properties of Sulfided Mo/Al2O3 Catalysts (20 wt% Mo) Prepared by the Addition of Citric Acid by the Simultaneous Impregnation Method and the Posttreatment Method (CA/Mo mole ratio＝1.5) Catalyst

Textural property Surface area [m ・g-1]

Pore volume [cm3・g-1]

180 116 205 251

1.26 0.72 0.51 0.47

2

γ-Al2O3 Mo/Al (calc) Mo_CA/Al CA/Mo/Al

shows that the average size of MoS2 particles on the post-treatment catalysts was considerably smaller than that on the simultaneous impregnation catalyst, in agreement with the NO adsorption results shown in Fig. 3. The higher dispersion of MoS2 particles by the posttreatment method thus explains the significantly higher HDS activity of Co/CA/Mo/Al compared to Co/Mo _ CA/Al (Fig. 1). Figures 1-3 show that post-treatment of the calcined Mo/Al2O3 oxide catalysts is a very effective preparation method for highly active HDS catalysts. Mo K-edge extended X-ray absorption fine structure (EXAFS) has shown that Mo_CA complexes are formed after citric acid is added in both preparation methods for Co_Mo/B2O3/Al2O3 with 8.7 wt% Mo18). We suggest that the unchanged dispersion of Mo in the simultaneous impregnation catalysts (Fig. 3) is due to the formation of large agglomerated Mo_CA complexes outside the pores of Al2O3 during the drying process, whereas with the post-treatment method, citric acid redisperses the Mo species by reaction with Mo oxide particles dispersed in the pores of the support to form in-situ well dispersed Mo _CA complexes inside of the pores17). This surface model is consistent with the insitu XAFS observations for Co_Mo/B2O3/Al2O3 with 8.7 wt% Mo, showing that the sulfidation of Mo_CA in CA/Mo/Al takes place at a lower temperature than in Mo_CA/Al27). The surface area and pore volume of the sulfided Mo/ Al2O3 catalysts prepared by the addition of citric acid are summarized in Table 3. In agreement with previous28) and our former observations17),18), the addition of citric acid greatly increased the specific surface area, in particular by post-treatment. In addition, the pore volume was decreased after the addition of citric acid, indicating modification of the physical properties of the catalysts. The increase in surface area and decrease in pore volume were greater for CA/Mo/Al than Mo_CA/

(a) Mo/Al (dried), (b) Mo_CA/Al, and (c) CA/Mo/Al. Adsorption (open symbols) and desorption (closed symbols). Fig. 7●N2 Adsorption-desorption Isotherms of 20 wt% Mo/Al2O3 Catalysts Prepared by Citric Acid Addition (CA/Mo mole ratio＝1.5)

Al. Figure 7 shows typical N2 adsorption-desorption isotherms for sulfided Mo/Al2O3 prepared with citric acid. A wide range hysteresis loop was observed at P/ P0 values around 0.4-0.9 for sulfided Mo_CA/Al or CA/ Mo/Al, suggesting the formation of ink-bottle type mesopores24),28). The ink-bottle type structure was more dominant for the simultaneous impregnation catalyst (Mo_CA/Al) compared to the post-treatment catalyst (CA/Mo/Al). We previously showed that the pore size diameter of sulfided 8.7 wt% Mo/Al2O3 catalysts is specifically modified to form ca. 4 nm pores by the presence of citric acid17),18). The adsorption isotherm of Mo/Al (dried) (Mo _CA/Al with CA/Mo＝0) did not show any large change in the pore structure of the sup-


No. 5, 2010

298

(△) CoMo/Al (dried), (▲) Co/CoMo/Al (dried), (○) CoMo_CA/Al, and (●) Co/CoMo_CA/Al.

(△) CoMo/Al (calc), (▲) Co/CoMo/Al (calc), (○) CA/CoMo/Al, and (●) Co/CA/CoMo/Al.

Fig. 8●HDS Activity of Co _ Mo/Al2O3 Catalysts (20 wt% Mo) Prepared by the Simultaneous Impregnation Method with Citric Acid (CA/Mo mole ratio＝2)

Fig. 9●HDS Activity of Co _ Mo/Al2O3 Catalysts (20 wt% Mo) Prepared by the Post-treatment Method with Citric Acid (CA/Mo mole ratio＝2)

port, whereas 10% citric acid/Al2O317) showed the formation of a specific pore structure. This suggests that carbonaceous deposits produced by the decomposition of citric acid during the sulfidation process forms a specific structure with 4-nm diameter pores. The carbonaceous deposits formed may inhibit participation of some MoS2 edge sites in the adsorption of Co(CO)3NO (Fig. 2), as suggested above. A greater decrease in the pore volume for CA/Mo/Al is probably due to the better dispersion of Mo _ CA complexes inside the pores27). 2. 3. HDS Activity of Co–Mo/Al2O3 with Citric Acid The effect of citric acid addition on the HDS of DBT was also studied for Co_Mo/Al2O3 catalysts (20 wt% Mo). The activities of the Co_Mo/Al2O3 catalysts with citric acid, prepared by the simultaneous impregnation method (CoMo _ CA/Al) and by the post-treatment method (CA/CoMo/Al), are shown in Figs. 8 and 9, respectively, as a function of Co content. With the CoMo/Al (dried) catalysts (CoMo_CA/Al with CA/Mo＝ 0) in Fig. 8, the rate constant increased with increasing Co content up to about 6 wt% Co and then leveled off. Addition of citric acid by the simultaneous impregnation method caused the rate constant of CoMo_CA/Al to considerably increase up to 4-6 wt% Co and then decrease at higher Co contents. The rate constant of CoMo_CA/Al was 1.5 times higher than that of CoMo/ Al (dried) (at 5.5 wt% Co). The CoMo/Al (dried) and CoMo _CA/Al catalysts were investigated by CVD for characterization of the surface structure of the catalysts. The rate constants of the catalysts are also depicted in Fig. 8. With CoMo/Al (dried), the rate constant of Co/CoMo/Al (dried) was considerably increased at 0 wt% Co (unpro-

moted Mo/Al (dried)), followed by a sharp decrease with increasing Co content, approaching that of CoMo/ Al (dried) at ＞ 6 wt% Co. The significant increase of the rate constant at 0 wt% Co by the addition of Co using the CVD method is due to the full occupancy by Co of the edge sites of MoS2 particles without blocking of the active sites by catalytically inactive Co sulfide clusters (maximum potential activity21)～23)). Moreover, the sharp activity decrease of Co/CoMo/Al (dried) with increasing Co content is ascribed to blocking of the MoS2 sites by catalytically inactive Co sulfide clusters during the sulfidation process22). The rate constant of CoMo_CA/Al was also considerably increased by the addition of Co using CVD at a low loading of Co (up to 6 wt% Co), suggesting that the edges of MoS2 particles in CoMo_CA/Al are only partially occupied by Co at ＜ 6 wt% Co22). No activity increase after CVD was observed at a higher Co loading, possibly because the MoS2 edge sites are already fully covered by Co at ＞ 6 wt% Co22) , although the MoS2 edges and thus Co_Mo_S are slightly blocked by catalytically inactive Co sulfides. On the other hand, the HDS activity of Co/CoMo _ CA/Al was constant below 6 wt% Co, indicating no notable active site blocking in CoMo_CA/Al at ＜ 6 wt% Co. Figure 8 also shows that the rate constants of CoMo/Al (dried) and CoMo _ CA/Al at 0 wt% Co were very similar after CVD, suggesting that the edge dispersion of MoS2 particles was not modified by the addition of citric acid by the simultaneous impregnation method, in agreement with the amount of NO adsorption in Fig. 3. We previously found that citric acid not only reacts with Mo to form Mo_CA complexes, but also with Co to form Co_ CA complexes, resulting in increased fraction of Co as


No. 5, 2010

299

Co _ Mo _ S after sulfidation15),18). Therefore, the increase in the rate constant of CoMo_CA/Al at a low Co content caused by the addition of citric acid is due to improved dispersion of Co, resulting in enhanced formation of Co_Mo_S and reduced formation of inactive Co sulfide clusters at ＜ 6 wt% Co, although the edge dispersion of MoS2 particles is not modified by the simultaneous impregnation method. The HDS activity of CoMo/Al (calc) (CA/CoMo/Al with CA/Mo＝0) in Fig. 9 increased up to 6 wt% Co and decreased at ＞ 6 wt% Co. The activity of CoMo/ Al (calc) was increased by CVD ( ＜ 4 wt% Co). However, the activity of Co/CoMo/Al (calc) sharply decreased below 4 wt% Co, suggesting that the active sites are significantly masked by the catalytically inactive Co sulfide clusters22) . The rate constant of CoMo/Al (calc) was slightly higher than that of the corresponding CoMo/Al (dried) catalysts (Fig. 8). Addition of citric acid by the post-treatment method (Fig. 9) caused the activity of CoMo/Al (calc) to drastically increase over the whole loading range of Co. The maximum activity was attained around 6 wt% Co, followed by decreased activity with further addition of Co. The maximum rate constant of the post-treatment catalysts (CA/CoMo/Al) was three times that of CoMo/ Al (calc) (5.5 wt% Co). Moreover, comparing with the activity shown in Fig. 8, the rate constant of CA/ CoMo/Al was considerably higher even than that of the CoMo_CA/Al catalysts prepared by the simultaneous impregnation method, as the HDS activity of CA/ CoMo/Al was 70% higher than of the CoMo_CA/Al catalyst at 6 wt% Co. Clearly the addition of citric acid to conventional Co_Mo/Al2O3 calcined catalysts by the post-treatment method is a promising method to increase the catalytic activity for the HDS of DBT. The post-treatment catalysts (CA/CoMo/Al) were also investigated by CVD22). The rate constants of the resultant CVD catalysts are shown in Fig. 9. Similar to Co/CoMo/Al (calc), the rate constant was greatly increased by CVD, but decreased with increasing Co content (＜ 6 wt% Co) and equalled the rate constant of CA/CoMo/Al at ＞ 6 wt% Co. The sharp activity increase of Co/CA/CoMo/Al at a low Co content (＜ 6 wt% Co) is due to the presence of vacant MoS2 edges for full promotion, whereas the decrease in the activity is caused by active site blocking by catalytically inactive Co sulfide clusters, in contrast to the CA/ CoMo/B2O3_Al2O3 catalysts (4 wt% Co and 8.7 wt% Mo) prepared by the post-treatment method, in which no such blocking was observed by the characterization with CVD18). Comparing the rate constants of Co/CoMo/Al (calc) and Co/CA/CoMo/Al at 0 wt% Co showed that the rate constant of the latter catalyst was notably higher than that of the former. Taking into consideration the amount of NO adsorption in Fig. 3 and the enhanced

Co contents are 1.5 and 7.5 wt% Co. The spectrum of Co9S8 is also shown as a reference. Fig. 10●Co K-edge XANES Spectra for CA/CoMo/Al (CA/Mo mole ratio＝2) and CoMo/Al (calc) (CA/Mo＝0)

HDS activity of Co/CA/CoMo/Al, the edge dispersion of MoS2 particles was probably increased by addition of citric acid by the post-treatment method. Moreover, Fig. 9 shows that the maximum potential activity of Co/CA/CoMo/Al with 0 wt% Co or Co/CA/Mo/Al was not only more than double that of Co/CoMo/Al (calc) (0 wt% CO), but also more than twice higher than that of Co/CoMo_CA/Al with 0 wt% Co (Fig. 8). Therefore, addition of citric acid by the post-treatment method is extremely effective to prepare highly active HDS catalysts with a high loading of Mo for the HDS of DBT. Co K-edge XANES measurements were conducted to study the chemical state of Co in the sulfided catalysts. Figures 10 and 11 show the Co K-edge X-ray absorption near edge structure (XANES) spectra for sulfided CA/CoMo/Al prepared by the post-treatment method and the sulfided original CoMo/Al (calc) catalysts together with a reference compound, Co9S8. At low Co content (1.5 wt% Co), a very weak 1s-3d preedge peak was observed for CoMo/Al (calc) (Fig. 11-b) compared with that for Co9S8, implying that Co exists mainly in a higher symmetry than tetrahedral coordination, as observed for Co_Mo_S. An even weaker preedge peak was observed for CA/CoMo/Al (1.5 wt% Co) (Fig. 11-a), showing that the fraction of Co9S8 was marginal. Clearly Co was better-dispersed on the edge sites of MoS2 particles at a Co content of 1.5 wt% for CA/CoMo/Al, which contributes to the larger increase in HDS activity caused by the addition of citric acid using the post-treatment method (Fig. 9).


No. 5, 2010

300

(a) CA/CoMo/Al (1.5 wt% Co), (b) CoMo/Al (calc) (1.5 wt% Co), (c) CA/CoMo/Al (7.5 wt% Co), (d) CoMo/Al (calc) (7.5 wt% Co), and (e) Co9S8 as a reference. Fig. 11●Co K-edge XANES Spectra of the 1s-3d Pre-edge Region for CA/CoMo/Al (CA/Mo mole ratio＝2) and CoMo/Al (calc) (CA/Mo＝0)

In contrast, the intensity of the 1s-3d pre-edge peak was considerably increased with higher Co content for CoMo/Al (calc) (7.5 wt% Co), as shown in Fig. 11-d. The intensity of the pre-edge peak was considerably decreased by the post-treatment with citric acid (Fig. 11c), indicating decreased fraction of Co sulfide clusters in CA/CoMo/Al, with respect to CoMo/Al (calc), consistent with the higher activity of CA/CoMo/Al than that of CoMo/Al (calc) (Fig. 9). The fraction of Co sulfide clusters was greater for the 7.5 wt% Co catalysts than that for 1.5 wt% Co catalysts, explaining the extensive active site blocking in both CA/CoMo/Al and CoMo/Al (calc) catalyst systems. As shown previously by in-situ XAFS for Co _ Mo/B2O3/Al2O3 catalysts (8.7 wt% Mo) with citric acid, the formation of Co-CA complexes increases the sulfidation temperature of Co, resulting in improved formation of Co _ Mo-S27) , as found previously10)～13). 2. 4. Improvement of the HDS Activity of Co–Mo/ Al2O3 by the Post-treatment Method Post-treatment with citric acid of calcined Mo/Al2O3 and calcined conventional Co _Mo/Al2O3 oxide catalysts, containing a high loading of Mo (20 wt% Mo), was much more effective for the preparation of highly active catalysts for the HDS of DBT compared to the simultaneous impregnation method. Figure 3 clearly shows that the dispersion of MoS2 particles in Mo/Al (calc) was greatly increased by the post-treatment with citric acid. X-ray diffraction (XRD) and Mo K-edge EXAFS showed that citric acid consumes crystalline MoO3 and CoMoO4 particles and, probably, coexisting well-dispersed Mo oxide clusters to form well-dispersed Mo_CA surface complexes17),18). In contrast, no increase in the dispersion of MoS2 particles was observed

with the simultaneous impregnation catalyst, as shown in Fig. 3. Although Mo _CA surface complexes are also formed after citric acid is added by the simultaneous impregnation method17), the dispersion of MoS2 particles was lower than or similar to that of the calcined catalyst (Fig. 3). According to the TEM observations (Table 2), the different catalytic behaviors of Mo/Al2O3 prepared by the post-treatment and simultaneous impregnation methods could be explained in terms of the dispersion of MoS2 particles after sulfidation. If AHM and citric acid are simultaneously present in the impregnation solution, citric acid reacts with AHM to form Mo_CA complexes, and these complexes are precipitated and agglomerated on the outer surface of pores during the drying process because of low solubility, leading to relatively large MoS2 particles on the support surface after sulfidation. On the other hand, if citric acid is added by the post-treatment method, citric acid combines with Mo oxide species including crystalline MoO3 particles on the calcined Mo/Al2O3 catalyst to form Mo_CA complexes in-situ17),18). Most of the Mo oxide species are already well dispersed in the pores by the wetting process during calcination5),29),30). Therefore, the resultant Mo_CA surface complexes are expected to be better dispersed inside the pores, forming small MoS2 particles after sulfidation, and leading to great improvement in the catalytic activity of the calcined catalyst, in accordance with the greatly increased NO adsorption capacity in CA/Mo/Al (Fig. 3) by the post-treatment method. Besides the increase in the edge dispersion of MoS2 particles, the addition of citric acid may also improve the dispersion of Co. Citric acid also interacts with Co2＋ in the impregnation solution and CoMoO4 particles to form Co _CA surface complexes11),15),18). These Co citrate complexes increase the decomposition temperature and accordingly the sulfidation temperature of Co during sulfidation11),27), resulting in the formation of a larger amount of Co_Mo_S, in particular by the post-treatment method18),27). The formation of Co_CA complexes also improves the dispersion of Co by the post-treatment method13), as indicated by the Co K-edge XANES results in Figs. 10 and 11. Therefore, the addition of citric acid to the calcined Co_Mo oxide catalysts by the post-treatment method is very effective to improve the performance of the catalysts. The other favorable effects of citric acid addition on the Mo/Al2O3 catalysts are probably also caused by weakened interactions between Mo and the support surface due to the formation of Mo_CA surface complexes as well as the increased dispersion of Mo precursors at the expense of crystalline MoO3 particles, as observed in our previous study17). Mo becomes easy to sulfide due to the weakened interactions between Mo and the support surface induced by the addition of citric acid. Mo K-edge EXAFS spectra17) show that the sulfidation


No. 5, 2010

301

of Mo on the Mo/Al catalysts is significantly increased by citric acid addition regardless of addition by the post-treatment method or the simultaneous impregnation method, as suggested by previous observations11),16). 3.

Conclusion

The present study investigated the effects of citric acid addition on the catalytic activity of Mo/Al2O3 and Co _ Mo/Al2O3 catalysts with a high loading of Mo (20 wt% Mo) for the HDS of DBT. Significant improvements in the HDS activity were obtained by the post-treatment method. The post-treatment of calcined Mo/Al2O3 and Co_Mo/Al2O3 oxide catalysts with citric acid transformed crystalline MoO3 and CoMoO4 particles into well-dispersed Mo_CA surface complexes, improving the edge dispersion of MoS2 particles on the catalyst surface after sulfidation. Therefore, the addition of citric acid by the post-treatment method is very effective to increase the edge dispersion of MoS2 particles on catalysts with high loading of Mo, and to significantly improve the HDS activity of the resulting Co/CA/Mo/Al and CA/CoMo/Al catalysts. Posttreatment of calcined Co_Mo/Al2O3 oxide catalysts is a promising method for the preparation of highly active HDS catalysts. Finally, we suggest that the posttreatment method with citric acid of calcined Co_Mo catalysts can be applied to the regeneration of deactivated HDS catalysts. Acknowledgment We are grateful to Dr. Takashi Fujikawa and Mr. Takeshi Ebihara (Research and Development Center, Cosmo Oil Co., Ltd.) for the TEM observations of the catalysts. References 1) Song, C., Catal. Today, 86, 211 (2003). 2) Topsøe, H., Clausen, B. S., Massoth, F. E., “Hydrotreating Catalysis,” Springer, Berlin (1996). 3) Whitehurst, D. D., Isoda, T., Mochida, I., Adv. Catal., 42, 345 (1998). 4) Kabe, T., Ishihara, A., Qian, W., “Hydrodesulfurization and Hydrodenitrogenation,” Kodansha, Tokyo (1999).

5) Eijsbouts, S., Appl. Catal. A: General, 158, 53 (1997). 6) Yoshimura, Y., Matsubayashi, N., Sato, T., Shimada, H., Nishijima, A., Appl. Catal. A: General, 79, 145 (1991). 7) Yoshimura, Y., Sato, T., Shimada, H., Matsubayashi, N., Imamura, M., Nishijima, A., Higo, M., Yoshitomi, S., Catal. Today, 29, 221 (1996). 8) van Dillen, A. J., Terörde, R. J. A. M., Lelisveld, D. J., Geus, J. W., de Jong, K. P., J. Catal., 216, 257 (2003). 9) Bergwerff, J. A., Jansen, M., Leliveld, B. R. G., Visser, T., de Jong, K. P., Weckhuysen, B. M., J. Catal., 243, 292 (2006). 10) Fujikawa, T., Kato, M., Kimura, H., Kiriyama, K., Hashimoto, M., Nakajima, N., J. Jpn. Petrol. Inst., 48, (2), 106 (2005). 11) Fujikawa, T., Kato, M., Ebihara, T., Hagiwara, K., Kubota, T., Okamoto, Y., J. Jpn. Petrol. Inst., 48, (2), 114 (2005). 12) Fujikawa, T., Catal. Surv. Asia, 10, 89 (2006). 13) Fujikawa, T., Kimura, H., Kiriyama, K., Hagiwara, K., Catal. Today, 111, 188 (2006). 14) Escobar, J., Barrera, M. C., de los Reyes, J. A., Toiedo, J. A., Santes, V., Colin, J. A., J. Molec. Catal. A: Chem., 287, 33 (2008). 15) Rinaldi, N., Usman, Al-Dalama, K., Kubota, T., Okamoto, Y., Appl. Catal. A: General, 360, 130 (2009). 16) Funamoto, T., Segawa, K., 11th Korea-Japan Symp. Catal., Seoul, 2007, Preprint, GO-37. 17) Rinaldi, N., Kubota, T., Okamoto, Y., Appl. Catal. A: General, 374, 228 (2010). 18) Rinaldi, N., Kubota, T., Okamoto, Y., Ind. Eng. Chem. Res., 48, 10414 (2009). 19) Mazoyer, P., Geantet, C., Diehl, F., Loridant, S., Lacroix, M., Catal. Today, 130, 75 (2008). 20) Okamoto, Y., Ochiai, K., Kawano, M., Kobayashi, K., Kubota, T., Appl. Catal. A: General, 226, 115 (2002). 21) Okamoto, Y., Ishihara, S., Kawano, M., Satoh, M., Kubota, T., J. Catal., 217, 12 (2003). 22) Okamoto, Y., Ochiai, K., Kawano, M., Kubota, T., J. Catal., 222, 143 (2004). 23) Okamoto, Y., Catal. Today, 132, 9 (2008). 24) Gregg, S. J., Sing, K. S. W., “Adsorption, Surface Area and Porosity,” 2nd ed. Academic Press, London (1988). 25) Okamoto, Y., Kawano, M., Kawabata, T., Kubota, T., Hiromitsu, I., J. Phys. Chem. B, 109, 288 (2005). 26) Kubota, T., Sato, K., Kato, A., Usman, Ebihara, T., Fujikawa, T., Araki, Y., Ishida, K., Okamoto, Y., Appl. Catal. A: General, 290, 17 (2005). 27) Kubota, T., Rinaldi, N., Okumura, K., Honma, T., Hirayama, S., Okamoto, Y., Appl. Catal. A: General, 373, 214 (2010). 28) Cheng, R., Shu, Y., Li, L., Zheng, M., Wang, X., Wang, A., Zhang, T., Appl. Catal. A: General, 316, 160 (2007). 29) Günther, S., Gregoratti, L., Kiskinova, M., Taglauer, E., Grotz, P., Schubert, U. A., Knözinger, H., J. Chem. Phys., 112, 5440 (2000). 30) Korányi, T. I., Paál, Z., Leyrer, J., Knözinger, H., Appl. Catal., 64, L5 (1990).


No. 5, 2010

302

要旨クエン酸を用いて調製した Co–Mo/Al2O3 触媒の水素化脱硫活性̶高モリブデン担持量焼成触媒のポスト処理̶ Nino RINALDI，吉岡政裕，久保田岳志，岡本康昭島根大学総合理工学部物質科学科，690-8504 松江市西川津町1060 クエン酸を同時含浸法あるいはポスト処理法により添加することにより，高モリブデン担持量

（20 2O3 触媒

Co_Mo/Al

wt%Mo）

媒を調製した。ポスト処理法でクエン酸を（Co _）Mo/Al2O3 焼成触媒に添加することにより，MoS2 粒子の分散性，および Co/

を調製した。触媒活性は，ジベンゾチオフェンの水素化脱硫反

Mo/Al2O3 および Co _ Mo/Al2O3 触媒の活性は著しく増加した。

応（HDS）に対して評価した。触媒のキャラクタリゼーション

一方，同時含浸法によるクエン酸添加では分散性の変化は見ら

は TEM，Co K-吸収端 XANES，NO 吸着量，表面積測定を用

れなかった。ポスト処理触媒では，最大活性はクエン酸╱ Mo

いて行った。また，触媒のキャラクタリゼーションのため，

モル比 1.5 ～ 2 で得られた。ポスト処理法は劣化触媒の再生に

Co（CO）3NO を用いる CVD 法で Co/MoS2 および Co/Co_MoS2 触

利用できることを示唆した。


No. 5, 2010