Hydrodesulfurization Activity of CoMo Catalysts Supported on ...

26 downloads 4235 Views 113KB Size Report
Ce in the support are 2 and 10 wt %, respectively. ... sense, support plays an important role.4 .... tograms of the cation-free titania support after various.
Energy & Fuels 2003, 17, 81-86

81

Hydrodesulfurization Activity of CoMo Catalysts Supported on Stabilized TiO2 P. Schacht,† G. Herna´ndez,† L. Ceden˜o,‡ J. H. Mendoza,† S. Ramı´rez,† L. Garcı´a,† and J. Ancheyta*,† Instituto Mexicano del Petro´ leo, Eje Central La´ zaro Ca´ rdenas No 152, Me´ xico D.F. 07730, Me´ xico, and UNICAT, Departamento de Ingenierı´a Quı´mica, UNAM, Me´ xico, D.F. 04510 Me´ xico Received June 24, 2002

In this work we report the results of characterization and catalytic behavior of different CoMobased catalysts supported on titania modified with lanthanum and cerium. Co and Mo contents of catalysts were between 25 and 100% of the corresponding amount of metals commonly found in commercial catalysts (Co + Mo ) 12.85 wt %). For all catalyst a constant Co/Mo weight ratio was kept at about 0.22. Catalytic activities were evaluated using dibenzothiophene hydrodesulfurization as model reaction. On the bases of specific surface area results and XRD measurements analyses after hydrothermal deactivation, it was found that suitable contents of La and Ce in the support are 2 and 10 wt %, respectively. DBT conversion of Ti-La catalysts was slightly higher than Ti-Ce catalysts, and both showed maximum values of conversion at 50 and 75% of the amount of metals in commercial catalysts.

1. Introduction As emission regulations continue to tighten in markets around the world, the need for ultralow sulfur diesel (ULSD) fuels will continue to increase. The exact definition of ULSD has not been fully clarified yet, but some environmental authorities are considering a range between 10 and 30 wppm sulfur. There are various process options for producing ULSD in commercial hydrotreating (HDT) plants. Most of them involve revamping of existing units, increase of reaction severity or improvement of HDT feed quality, which will depend on the original design of the hydrotreating unit.1,2 Another way for achieving ULSD specifications is by means of increasing activity of the HDT catalyst. However, to meet increasingly stringent environmental regulations, catalysts having activity more than four to five times of the present level are required.3 In this sense, support plays an important role.4 Most of the HDT catalysts used in commercial applications are molybdenum based promoted with cobalt or nickel and supported on gamma-alumina. Less attention has been given toward commercial application of other supports. †

Instituto Mexicano del Petro´leo. Departemento de Ingenierı´a Quı´mica. (1) Marroquı´n, G.; Ancheyta, J. Appl. Catal. A 2001, 207, 407420. (2) Marroquı´n, G.; Ancheyta, J.; Ramı´rez, A.; Farfa´n, E. Energy Fuels 2001, 15, 1213-1219. (3) Delmon, B.; Froment, G. F.; Grange, P. Hydrotreatment and Hydrocracking of Oil Fractions 1999, 153, 397, 401. (4) Maity, S. K.; Rana, M. S.; Bej, S. K.; Ancheyta, J.; Murali Dhar, G.; Prasada Rao, T. S. R. Appl. Catal. A 2000, 205, 215-225. ‡

Mo supported on titania is 4.4 times more active in the hydrodesulfurization (HDS) of tiophene, compared to Mo supported on alumina.5 This behavior has been explained on the electronic promotion of Mo phases by titania, similar in nature to the one of the Co.6 It has also been shown that Mo catalysts on titania have higher activities in dibenzotiophene HDS than alumina supported catalysts.7,8 Hence, TiO2 seems to be promising HDT catalyst support, especially to be used for obtaining ULSD. However, one of the disadvantages in using titania is its low specific surface area as well as its low thermal stability at high temperatures.9 Some literature reports indicate that sol-gel is a good method for improving the specific surface area of titania and allowing the addition of doping agents, to move to higher temperatures the transition anatase phase to rutile and to increase the thermal stability of this material. The doping agents commonly used are Cerium (Ce) and Lantanum (La).10,11 To evaluate the possible application of La and Ce doped titania as supports for hydrotreating catalysts in the present work Co-Mo catalysts have been prepared, characterized and evaluated in the HDS of dibenzothiophene (DBT). (5) Ramı´rez, J.; Fuentes, S.; Dı´az, G.; Vrinat, M.; Breyesse, M.; Lacroix, M. Appl. Catal. A 1989, 52, 211. (6) Ramı´rez, J.; Ceden˜o, L.; Busca, G. J. Catal. 1999, 59. (7) Ozaki, H. Catalysis Surv. Jpn. 1997, 1, 143. (8) Whitehurst, D. D.; Isoda, T.; Mochida, I. Adv. Catal. 1998, 42, 345. (9) Ramamoorthy, R. D.; King- Smith, Vanderbilt, D. Phys. Rev. B 1994, 49, 15. (10) Vargas, S.; Arroyo, R.; Haro, E.; Rodrı´guez, R. J. Mater. Res. 1999, 14, 3932. (11) Koebrugge, G. W.; Winnubst, L.; Burgraaf, A. J. J. Mater. Chem. 1993, 3, 1095.

10.1021/ef020144u CCC: $25.00 © 2003 American Chemical Society Published on Web 12/05/2002

82

Energy & Fuels, Vol. 17, No. 1, 2003

2. Experimental Section 2.1. Preparation of Support and Catalyst. Titania support was doped with different amounts of La and Ce in order to define suitable contents for preparing catalysts. Lanthanum- and Cerium-doped titania supports were prepared by sol-gel method. Titanium butoxide (Aldrich, 97%) dissolved in 2-propanol (J. T. Baker, 99.8%) was used as titanium source. Nitric acid (Aldrich, 71%) was employed as catalyst for hydrolysis reaction. From the total amount of solvent, 80% was mixed with titanium butoxide, and the remaining 20% was mixed with water and nitric acid. Later, this last mixture was added to the first, drop to drop, shaking at room temperature during 6 h. The molar ratios of acid/alcoxide, solvent/alcoxide, and water/alcoxide were 0.2, 75, and 10, respectively. The peptization of the resulting hydrolyzate with nitric acid was then left during 60 h at room temperature and then it was dried at 120 °C. Calcination took place at 500 °C during 3 h. The lanthanum and cerium nitrate doping agents were dissolved totally in water for hydrolysis, and added later to the solvent. The final solid was made with the stabilized titania and boehmite as binder. Catalysts doped with Ce and La were prepared with different active metal contents but with a constant Co/Mo weight ratio of about 0.22. A load of 2.8 atoms of Mo/nm2 was considered as optimal according to the literature.12,13 The catalysts were prepared by the method of coimpregnation of Co(NO3)‚2.6H2O and (NH4)6Mo7O24 on doped titania. The solution with Co and Mo was added to the support drop to drop. After, the sample was aged around 2 h, dried during 24 h at 100°C, and finally calcined at 500 °C for 2 h. The supports were named as Ti (%), Ti-La (%), and Ti-Ce (%), which correspond to CoMo/TiO2, CoMo/La-TiO2, and CoMo/Ce-TiO2 catalysts, respectively. The value in parentheses (%) indicates the percentage of the total amount of metals commonly found in commercial catalysts. For instance, the Ti (100%) catalyst corresponds to the CoMo/TiO2 sample with the same amount of metals than commercial catalysts. We have taken the following metal contents found in typical commercially available catalysts: 2.35 wt % Co and 10.5 wt % Mo. 2.2. Activity Test. Activity tests were carried out in a Parr batch reactor. Two hundred forty milligrams of catalyst were loaded to the reactor for each test. Before HDS reaction, the catalysts were sulfided with a mixture of H2/H2S (90/10 vol %) at 400 °C during 4 h. DBT hydrodesulfurization was chosen as catalytic activity test. The reaction was carried out at the following operating conditions: 300 °C reaction temperature, 400 psig pressure, 6.3 mol/mol hydrogen-to-oil ratio, 6 mg/mL catalyst/feed ratio, and 6 h time-on-stream. The reaction mixture consisted of DBT and hexadecane as solvent (0.5 wt % sulfur as DBT). Product analyses were performed on-line in a PERKIN ELMER chromatograph equipped with a 50 m HP Ultra 2 capillary column with flame ionization detector (FID). Conversion in each test was defined as 100 minus unconverted DBT in weight percent, and product selectivity was calculated by dividing product yield between conversion. 2.3. Catalyst Characterization. The elemental analysis was performed by quantitative atomic absorption spectrometry in a Varian Spectra-10 plus spectrometer. Specific surface area was determined by means of BET method, using an ASAP 2000 sorptometer. (12) Ramı´rez, J.; Harle, V.; Ruiz, L.; Ceden˜o, L.; Vrinat, M.; Breysse, M. In Proceedings of the XIIIth Iberoamerican Simposio of Catalysis I; Bermudez, O., Ed.; Guanajuato, Mexico, 1992; p 151. (13) Vissenberg, M. J. Preparative Aspects of Supported Metal Sulfide Hydrotreating Catalysts; Eindhoven University: Eindhoven, The Netherlands, 1999; p 97.

Schacht et al. Table 1. Results of Hydrothermal Deactivation of Laand Ce-Doped-Titania Supports area retention (%) support

doping agent

wt % of doping agent

2 h at 550 °C

2 h at 700 °C

TiO2 Ti-La Ti-La Ti-La Ti-Ce Ti-Ce Ti-Ce

La La La Ce Ce Ce

0 2 5 10 2 5 10

67 72 64 67 72 70 73

35 49 46 44 35 37 39

X-ray diffraction (XRD) spectra were recorded using a Siemens D-5000 instrument fitted with a monochromator for CuKR radiation operating valve 40 kV/40 mA.

3. Results and Discussion 3.1. Definition of La and Ce Contents. Before preparing catalysts, it was necessary to define suitable contents of doping agents in the support. For this reason, TiO2 supports with various La and Ce amounts (2, 5, and 10 wt %) were prepared. These supports were deactivated hydrothermally at 550 and 700 °C during 2 h. Before hydrothermal deactivation supports were calcined at 500 °C during 6 h. The changes in specific surface area before and after deactivation test were taken as a measure of stability of supports. The results are given in Table 1. It is clear from this table that contents of 2 wt % for La and 10 wt % for Ce have higher retention of specific surface area than other contents at both temperatures. Retention of specific surface area for these doping agent contents is also higher than pure titania support. For this reason, the catalysts were prepared with these amounts of La and Ce. 3.2. X-ray Diffraction of Supports. X-ray diffractograms of the cation-free titania support after various heat treatments are shown in Figure 1. Calcination was done during 6 h at temperatures in the range of 300700°C. XRD of TiO2 support without treatment is also included in this figure. XRD patterns of calcined samples at 300 °C show the typical diffraction peak of anatase phase. At higher temperature (i.e., 400 °C) diffraction peaks became more intense and narrow due to particle growth. After calcination at 500 °C new peaks appeared and the anatase phase peaks became less intense, which means that anatase was transformed to a more stable phase (rutile). At 700 °C, peaks of anatasa phase almost disappeared completely. Total transformation of anatasa phase has been reported in the literature at 800 °C.11,14 The same XRD measurements were carried out for La and Ce-doped-titania supports, which are shown in Figures 2 and 3, respectively. It is observed from both spectra that anatasa is the predominant phase in the range of temperatures of 300-700 °C. Small peaks of rutile phase appeared at 600 °C. Then, if XRD spectra of La and Ce-doped-titania supports (Figures 2 and 3) are compared to cation-free titania support (Figure 1) we can see that doping agents (14) Zhaobin, W.; Qin, X.; Xiexian, G. Appl. Catal. A 1990, 63, 305317.

Hydrodesulfurization Activity of CoMo Catalysts

Energy & Fuels, Vol. 17, No. 1, 2003 83

Figure 1. X-ray powder diffraction patterns of TiO2 support.

Figure 2. X-ray powder diffraction patterns of La-doped-TiO2 support.

inhibited the formation of rutile phase. These results confirm the stabilizing effect of the promoters added in the gelation stage of the synthesis. 3.3. Chemical Analysis of Catalysts. CoMo/TiO2, CoMo/La-TiO2, and CoMo/Ce-TiO2 catalysts were prepared with various amounts of Co and Mo. This amount of metals was considered as a percentage of that found in commercial catalysts. Theoretical contents (wt %) of Co and Mo for each catalyst were (25) Co ) 0.588, Mo ) 2.625, Co+Mo ) 3.213, (50) Co ) 1.176, Mo ) 5.25, Co+Mo ) 6.426, (75) Co ) 1.764, Mo ) 7.875, Co+Mo ) 9.639, (100) Co ) 2.35, Mo ) 10.5, Co+Mo ) 12.85. For all catalysts the Co/Mo ratio was of 0.224. Table 2 shows the results of the chemical analyses of the three series of catalysts. It can be seen that Mo and

Co contents as well as Co/Mo ratio are in very good agreement with theoretical values. Hereafter, we will refer the results as theoretical values. 3.3. Specific Surface Area of Catalysts. The BET areas of the catalysts prepared with various amounts of metals are shown in Figure 4. For all catalysts the specific surface areas gradually decreased with increasing content of metals, and dropped from 84 to 91 to 55-60 m2/g when Co+Mo loading changed from 3.2 to 12.8 wt % (25 and 100% of the amount of metals in commercial catalysts, respectively). It can also be observed that specific surface areas of samples doped with La and Ce do not reduce drastically compared to the corresponding samples without cations. The maximum reduction was found with Ce-doped catalyst (5-7 m2/g). This slight decrease means that

84

Energy & Fuels, Vol. 17, No. 1, 2003

Schacht et al.

Figure 3. X-ray powder diffraction patterns of Ce-doped-TiO2 support.

Figure 4. BET areas of catalysts. (O) CoMo/TiO2, (0) CoMo/ La-TiO2, (4) CoMo/Ce-TiO2. Table 2. Chemical Analysis of the Catalysts (wt %) catalyst

Mo

Co

Co+Mo

Co/Mo ratio

Ti (25) Ti (50) Ti (75) Ti (100)

2.60 5.25 7.80 10.50

CoMo/TiO2 0.60 1.17 1.70 2.35

3.20 6.42 9.50 12.85

0.231 0.223 0.218 0.224

Ti-La (25) Ti-La (50) Ti-La (75) Ti-La (100)

2.40 5.20 7.70 10.50

CoMo/La-TiO2 0.50 2.90 1.15 6.35 1.70 9.40 2.20 12.70

0.208 0.221 0.221 0.210

Ti-Ce (25) Ti-Ce (50) Ti-Ce (75) Ti-Ce (100)

CoMo/Ce-TiO2 2.80 0.60 3.40 5.30 1.18 6.48 7.50 1.73 9.23 10.370 2.40 13.10

0.214 0.223 0.231 0.224

incorporation of Ce or La does not affect texturally titania. 3.4. Catalyst Activities. Before discussing results of activity tests, it should be mentioned that kinetic

analysis was not the purpose of the present study but the comparison of catalyst activity. For this reason only one condition was evaluated for all catalysts and the difference in activity of catalysts was then analyzed by the conversion and product selectivities in the reaction of hydrodesulfurization of DBT. As was mentioned in introduction section, one of the major purposes of the present contribution is to prepare catalysts for obtaining ULSD. As well-known, the major sulfur compound existing in diesel fuel with a S level less than 500 wppm are the alkyl DBT’s, the so-called refractory sulfur compounds. Almost no DBT exists in the diesel fuel with the S level less than 300 wppm. Hence, more refractory compounds than DBT such as 4,6-DMDBT should be employed for catalyst evaluation for deep HDS as the HDS reactivity and reactions mechanism of the refractory sulfur compounds are very different from DBT. However, as this first work was an exploratory-type research, we have decided to use DBT as model compound only to have a preliminary idea about the advantages of application of La and Ce dopped titania as supports for hydrotreating catalysts. Results of experiments with other sulfur compounds will be reported in next contributions. The reaction was carried out at 300°C during 6 h using sulfided CoMo/La-TiO2 and CoMo/Ce-TiO2 catalysts. According to the reaction network (Figure 5) proposed by Houalla et al.,15 the products of the DBT HDS reaction are unconverted dibenzothiophene (DBT), hydrogenated DBT (TH-DBT, tetrahydrodibenzothiophene, and HH-DBT, hexahydrodibenzothiophene), biphenyl (BP), cyclohexylbenzene (CHB), and bicyclohexyl (BCH). The products identified by GC-MS in this work correspond to those indicated in Figure 5. Additionally, benzene (BZ), cyclohexane (CH), and some other cracked products (Lights) were also detected, which agrees with (15) Houalla, M.; Nag, N. K.; Sapre, A. V.; Broderick, D. H.; Gates, B. C. AIChE J. 1978, 24, 1015-1021.

Hydrodesulfurization Activity of CoMo Catalysts

Energy & Fuels, Vol. 17, No. 1, 2003 85

Table 3. Results of Catalytic Activities in the Hydrodesulfurization of DBT Using CoMo/La-TiO2 and CoMo/Ce-TiO2 Catalysts CoMo/Ce-TiO2

CoMo/La-TiO2

wt %

Ti-Ce (25)

Ti-Ce (50)

Ti-Ce (75)

Ti-Ce (100)

Ti-La (25)

Ti-La (50)

Ti-La (75)

Ti-La (100)

BZ CH BCH BP CHB HH-DBT TH-DBT DBT Lights DBT conversion

5.04 2.91 2.00 10.31 6.44 1.93 3.55 59.11 8.71 40.89

7.33 3.64 2.97 11.66 9.56 2.09 6.15 41.97 14.63 58.03

7.95 5.90 2.64 13.11 10.39 2.85 7.40 40.11 9.65 59.89

7.51 5.17 1.61 12.43 7.11 0.80 3.35 51.20 10.82 48.80

2.02 0.96 1.53 10.39 5.40 1.35 6.95 58.60 12.80 41.40

6.41 3.60 2.91 12.66 9.52 2.00 7.11 39.55 16.24 60.45

7.43 4.91 2.55 13.51 10.31 2.05 5.42 40.31 13.51 59.69

8.55 5.17 1.59 14.51 7.19 0.80 3.96 50.19 8.04 49.81

Figure 5. Reaction scheme of hydrodesulfurization of DBT.

other reports in the literature.16 This means that HDS of DBT undergoes the two well-known path reactions: (1) hydrodesulfurization via C-S bond hydrogenolysis to form BP, and (2) hydrogenation to form hydrogenated DBT (TH-DBT and HH-DBT). All these products then desulfurize to form CHB and finally BCH by hydrogenation. The results of catalytic activity are presented in Table 3. It can be observed that Ti-La catalysts yield slightly more DBT conversions than Ti-Ce catalysts. For both series of catalysts the maximum value of conversion was obtained between 50 and 75 wt % of the Co+Mo content reported in commercial catalysts (Figure 6). Products distribution is also shown in Table 3 and main products selectivities (CHB, BP, BCH, and H-DBTd TH-DBT+HH-DBT) are plotted in Figure 7 for both TiLa and Ti-Ce catalysts. It is observed that La and Ce affect in different way products selectivities. In the case of BCH and CHB, yields and selectivities are higher when supported byTi-La, and for both series of catalysts maximum values were obtained at 50 and 75 wt % of the amount of metals in commercial catalysts, respectively. The inverse behavior was found for BP selectivity (Figure 7). All these results suggest that CoMo/La-TiO2 and CoMo/Ce-TiO2 may be suitable catalysts to be used in hydrotreating process for producing ULSD. Metals (Co and Mo) contents would be lesser than those reported in commercial catalysts. Of course more experimental work with real feeds and typical HDT operating conditions is necessary. (16) Olgin, E.; Vrinat, M.; Ceden˜o, L.; Ramı´rez, J.; Borque, M. Appl. Catal. A 1997, 165, 1.

Figure 6. Conversion of DBT. (0) CoMo/La-TiO2, (4) CoMo/ Ce-TiO2.

Figure 7. Product selectivity of DBT hydrodesulfurization.

3.5. Future Work. It should be emphasized that this work was exploratory in nature, and it should be considered as an initial screening of potential applicability of promoted titania supports for HDS catalysts. Hence, there are some points that need to be verified and performed according to conventional approaches, such as the following: • Activity Test. Experiments were carried out with DBT and for developing catalysts for ULSD production, more refractory compounds, i.e., 4,6-DMDBT, are needed for catalyst activity evaluation.

86

Energy & Fuels, Vol. 17, No. 1, 2003

• Sulfiding Stream. In the present work we used a mixture of H2S/H2 and during commercial operation this stage is carried out with the same feed with or without sulfiding agent, such as DMDS. Then, a real feed, i.e., straight-run gas oil (SRGO), is necessary for sulfiding. • Sulfiding Procedure. Sulfiding was done at 400 °C, and better catalyst behavior has been reported with sulfiding in various steps, i.e., drying, soaking, and activation at two different temperatures. • Reaction Conditions. Experiments were conducted at atmospheric pressure in a glass reactor and real conditions are necessary, which include evaluation of catalyst with real feed such as SRGO alone or blended with light cycle oil. Time-on-stream was only 6 h and for evaluation of catalyst stability it is necessary to perform experiments during more time, i.e., 60 h. The effect of important operating variables, such as reaction temperature and hydrogen partial pressure, is also needed. • Co/Mo Ratio of Catalysts. We have prepared catalysts with constant Co/Mo weight ratio of about 0.22, which was taken as reference from typical commercially available catalysts. The effect of this ratio on catalyst behavior should be also studied. • Comparison with Commercial Catalyst. Relative activity for HDS in comparison with a conventional CoMo/alumina catalyst is necessary. This comparison should be done in terms of activity and stability on the same feed, reaction system, and operating condition basis.

Schacht et al.

4. Conclusions On the basis of our results of supports and catalysts characterizations as well as activity test of catalysts, the following conclusions can be raised: • Titania supports with 2 wt % La and 10 wt % Ce were selected for catalyst preparation due to their higher retention of specific surface area than other samples during hydrothermal deactivation. • XRD measurements showed the stabilizing effect of La and Ce which inhibited formation of rutile phase compared to cation-free titania support. • Specific surface areas of La- and Ce-doped-titania supports do not reduce drastically compared to the corresponding samples without cations which means that incorporation of these agents does not affect texturally titania. • CoMo/Ti-La catalyst showed slightly more DBT conversions than with CoMo/Ti-Ce catalyst, and maximum values of conversion were achieved with 50 and 75 wt % of the amount of metals in commercial catalysts. Acknowledgment. The authors thank Instituto Mexicano del Petro´leo for its financial support (97-10III FIES project). EF020144U