Catal Lett (2009) 130:291–295 DOI 10.1007/s10562-009-9919-9
Highly HYD Selective Pd–Pt/support Hydrotreating Catalysts for the High Pressure Desulfurization of DBT Type Molecules V. G. Baldovino-Medrano Æ Sonia A. Giraldo Æ Aristo´bulo Centeno
Received: 27 November 2008 / Accepted: 24 February 2009 / Published online: 25 March 2009 Ó Springer Science+Business Media, LLC 2009
Abstract A silica–alumina supported Pd–Pt bimetallic catalyst was found to be highly active in dibenzothiophene hydrodesulfurization with preferential selectivity to the hydrogenation (HYD) route of desulfurization at the expense of the direct removal of sulfur. In addition, it was demonstrated that for Pd–Pt catalysts selectivity to HYD can be drastically changed by a ‘‘soft’’ change in the nature of the carrier. Keywords Pd–Pt SiO2–Al2O3 DBT HDS HYD route
1 Introduction The development of the hydrogenation (HYD) route of desulfurization of dibenzothiophene (DBT) type molecules is urged in order to accomplish deep hydrodesulfurization (HDS) of heavy stock oils [1–3]. The main advantage of HYD is that it makes more flexible the aromatic structure of DBT [3], which in turn reduces the steric hindrance of highly refractory b-alkyl-DBTs (i.e. alkyl-dibenzothiophenes substituted in 4- and 6-positions such as 4,6-dimethyldibenzothiophene (4,6-DMDBT)) [1–3]. In spite of the large amount of reports claiming new more active CoMo and NiMo sulfided catalysts, none of them have proven preferential conversion of DBT via HYD so far [1–6].
V. G. Baldovino-Medrano S. A. Giraldo A. Centeno (&) Centro de Investigaciones en Cata´lisis (CICAT), Escuela de Ingenierı´a Quı´mica, Universidad Industrial de Santander (UIS), Cra. 27 Calle 9, Bucaramanga, Colombia e-mail:
[email protected]
This is due to the fact that the role of Co and Ni as promoters is mainly to enhance the direct C–S bond scission function of the MoS2 active phase [4]. Also, to increase the selectivity to HYD over the direct desulfurization route (DDS), the active phase of the catalyst should be able to selectively adsorb the DBT molecule in a flat p-mode through one of its aromatic rings and perform its HYD rather than binding and breaking the S heteroatom in a ‘‘single’’ step [7]. Furthermore, DDS is the favored thermodynamic pathway under standard reaction conditions [8]. Consequently, comparing 4,6-DMDBT and DBT, only the latter allows a direct measurement of the intrinsic catalytic selectivity of the tested material either to HYD or to DDS. It is important to recall that both DBT and 4,6-DMDBT share a common mechanistic dynamics comprising parallel HYD and/or DDS reaction routes when their HDS is performed over ‘‘non-strongly’’ acidic supported catalysts [1–4]. This is why in this work DBT was chosen as a more suitable model molecule than 4,6-DMDBT to develop highly HYD selective catalysts. Recent studies show that Pd–Pt supported catalysts possess high activity in the HDS of DBT [9–13] and, in addition, they exhibit better selectivity to HYD than conventional sulfided CoMo and NiMo catalysts [10, 11]. Indeed, Kabe et al. [10] studied the HDS of DBT over highly loaded Pd–Pt/c-Al2O3 catalysts which displayed high HDS activity and almost complete conversion via HYD. Niquille-Ro¨thlisberger and Prins [11] tested Pd–Pt/cAl2O3 in the HDS of DBT and reported 30% selectivity to HYD. Many other studies have shown that the hydrogenating capacity and thiotolerance of Pd–Pt catalysts allow them to perform the HDS of 4,6-dimethyl-DBT (4,6-DMDBT) via HYD [9–12]. In spite of this, even over Pd–Pt neither preferential DBT desulfurization via HYD has been achieved nor has HYD been effectively controlled only by ‘‘soft’’ changes in the nature of the carrier. This has
123
292
been achieved in this work, where highly HYD selective Pd–Pt catalysts were developed for the high pressure HDS of DBT as a function of slight changes in the nature of the support. As discussed before, such materials are very promising candidates for the efficient HDS of b-alkyl-DBT because such molecules react almost exclusively via HYD.
2 Experimental 2.1 Materials Preparation Four bimetallic Pd–Pt catalysts, supported in different carriers, with a total 2 wt% (Pd ? Pt) nominal metal content and a Pd/Pt = 4 molar ratio were prepared by incipient wetness co-impregnation. Supports used comprised: home-made SiO2 (Si), Al2O3 (Al) and SiO2–Al2O3 with Si/Al = 0.1 molar ratio (Al–Si), and Procatalyse c-Al2O3 (Al-P). Home-made carriers were prepared adapting the method presented by Cai et al. [14] for MCM. In a typical synthesis of 10 g of support, c.a. 430 mL of deionized water and 27 mL of NH4OH (Merck, reagent grade) were heated at 353 K under vigorous stirring. Then, approximately 8 g of the cationic surfactant, hexadecyltrimethylammonium bromide (CTAB) (Merck, 98%), was added to the mixture until complete dilution. Afterwards, appropriate amounts of silica (tetraethyl orthosilicate, Merck) and/or aluminum trisecbutoxide (Merck) were added slowly to this solution. A 0.125(CTAB):1(silica and/ or aluminum precursors) molar ratio was used in all preparations. All throughout the preparation process, the pH was allowed to remain at its natural value (pH = 13–14). The resulting white slurry was filtered and dried at ambient conditions. To eliminate the surfactant a Soxhlet extraction was carried out for 24 h using an ethanol (250 mL)–HNO3 (5 mL) solution. Solids were then pressed into pellets, calcined at 773 K for 2 h in air flow (100 mL/min), and grounded to Dp = 0.18–0.6 mm. All supports were calcined for 4 h in air flow at 773 K before impregnation of the metals. Palladium and Platinum precursors used were PdCl2, Pd(II) acetylacetonate (Pd(acac)2), Pt(II) acetylacetonate (Pt(acac)2) and H2PtCl6 4H2O (all provided by Sigma–Aldrich), respectively. A mixed solution of chloride (PdCl2 or chloroplatinate) or organometallic precursors was prepared and impregnated onto the support. Amounts of Pt and Pd were adjusted to obtain the desired metallic contents. In the case of chloride precursors, both precursors were diluted in a water solution containing a small excess of HCl. The solution was heated under stirring until complete dissolution. The organometallic precursors were simultaneously diluted in hot toluene. Impregnated solids were aged at room temperature for 1 day and then dried at 393 K
123
V. G. Baldovino-Medrano et al.
for 12 h and calcined in air flow at 773 K for 4 h (5 K/min). The catalysts prepared with organometallic Pd and Pt precursors were designated as PdPt-Ac/Si, PdPt-Ac/Al, PdPtAc/Al-P and PdPt-Ac/Al–Si, respectively. The catalyst prepared with chloride precursors was supported on Procatalyse c-Al2O3 being labeled as PdPt-Cl/Al-P. In addition, for comparison purposes, two conventional CoMo (2 wt% CoO and 10 wt% MoO3) and NiMo (4.5 wt% NiO and 9 wt% de MoO3) catalysts, supported on Procatalyse c-Al2O3, were prepared using a procedure presented elsewhere [15]. 2.2 Catalytic Tests Before the catalytic tests, samples of 0.5 g of the calcined Pd–Pt catalysts were in situ dried at 393 K under N2 flow for 1 h, and subsequently, they were reduced in hydrogen flow (100 ml/min) for 3 h at 673 K. In the case of conventional CoMo and NiMo sulfidation with a 15vol% H2S in H2 mixture was performed under the same conditions as mentioned above. Reaction conditions were chosen to recreate the conditions of temperature and pressure of a typical HDT reactor [1–3]. Catalysts were tested in a continuous flow high pressure fixed-bed reactor at 5 MPa and 583 K for Pd–Pt catalysts. In the case of CoMo and NiMo, reaction temperature was adjusted to obtain a DBT conversion around 75%, specifically, T = 573 K for CoMo and 563 K for NiMo, respectively, to allow a direct comparison in HYD selectivity between NiMo, CoMo, PdPt-Ac/Al–Si and PdPt-Cl/Al–P. The composition of the model charge (liquid charge) was 2 wt% DBT (Aldrich) dissolved in cyclohexane (commercial grade). Hexadecane (2 wt%) (Aldrich) was used as an internal standard for the GC analysis. A liquid flow rate of 30 ml/h and a H2/liquid charge volume ratio of 500 LN/L were used. For all tests, the catalyst was diluted in glass spheres (1 mm in diameter) and placed between two glass wool plugs. The absence of any diffusion limitations was previously verified. Liquid products were analyzed using a HP 6890 GC equipped with an FID, an HP-1 capillary column (100 m 9 0.25 mm 9 0.5 lm) and a split injector. Catalytic tests were conducted until reaching the steady state attained after 4 h. Reaction products detected were: biphenyl (BP), cyclohexylbenzene (CHB) and partially hydrogenated intermediates: tetrahydro-dibenzothiophene (THDBT) and hexahydro-dibenzothiophene (HHDBT). Activity was expressed as the conversion of DBT (%CDBT), products distribution as the yield (%y) to each reaction product. Selectivity to HYD (SHYD) was defined as the (100 - %yBP)/%yBP ratio.
Highly HYD Selective Pd–Pt/support Hydrotreating Catalysts Fig. 1 Steady state average HYD selectivity for the prepared catalysts. Ac = organometallic Pd(acac)2 and Pt(acac)2 precursors; –Cl=PdCl2 and H2PtCl6 4H2O precursors; -P = Procatalyse c-Al2O3. Reaction conditions: T = 583 K; P = 5 MPa; liquid flow rate = 30 ml/h; and, H2/liquid charge volume ratio = 500 LN/L. *NiMo catalyst tested at 563 K. **CoMo catalyst tested at 573 K
293
2.0
Selectivity (S HYD)
1.5
1.0
0.5
0.0
PdPt-Ac/Si
PdPt-Ac/Al PdPt-Ac/Al-P PdPt-Cl/Al-P PdPt-Ac/Al-Si
NiMo*
CoMo**
Catalyst †
3 Results and Discussion Figure 1 shows the average steady state HYD selectivity SHYD in the HDS of dibenzothiophene for the prepared Pd–Pt catalysts and the two NiMo and CoMo conventional catalysts. Both MoS2-based catalysts display a very low SHYD in agreement with previous literature reports [1–6]. For the Pd–Pt system, SHYD drastically varies as a function of the carrier used. Thus, SHYD follows the order: PdPt-Ac/ Al–Si [ PdPt-Cl/Al-P [ PdPt-Ac/Si [ PdPt-Ac/Al-P [ PdPt-Ac/Al. The fact that a very sharp enhancement of SHYD is obtained over Pd–Pt with only ‘‘soft’’ changes in the nature of the carrier employed is rather un-precedent in HDS literature ([1–6, 16–18] and references therein). In Table 1 the total DBT conversion, %CDBT, and the products distribution for every catalyst is presented. The results presented in Table 1 show that for PdPt-Ac/Al–Si the HYD Table 1 Average steady state conversion and products distribution in DBT HDS for the different catalysts prepared Catalyst
%CDBT
%yCHB
%yBP
%yHHDBT
%yTHDBT
PdPt-Ac/Si
40
29.3
54.6
2.9
13.2
PdPt-Ac/Al
40
13.3
82.0
0.6
4.1
PdPt-Ac/Al-P
44
22.1
68.5
1.6
7.9
PdPt-Cl/Al-P
70
44.6
52.4
0.4
2.6
PdPt-Ac/Al–Si
76
61.0
35.0
0.7
3.3
NiMo
78
9.1
90.9
0.0
0.0
CoMo
75
8.0
92.0
0.0
0.0
%yCHB = %yield to cyclohexylbenzene; %yBP = %yield to biphenyl; %yTHDBT = %yield to tetrahydro-DBT; %yHHDBT = %yield to hexahydro-DBT
route of desulfurization is completely developed as it is evidenced by the fact that CHB is the main reaction product. Moreover, for this catalyst %CDBT is at the same level than for conventional CoMo and NiMo, as it was intended to be able to compare HYD selectivity at isoconversion. As previously mentioned, HYD is a thermodynamically un-favored reaction route [8]. Consequently, only those catalysts in which DBT desulfurization is kinetically controlled, for example un-promoted MoS2 and Pd, HYD reaction rates have been reported to be comparable to DDS ones [4, 19]. On the other hand, it must be remarked that the total nominal metallic content of all Pd–Pt catalysts is 2 wt% which is far from those contents reported by Kabe et al. [10]. These authors [10] observed almost complete conversion of DBT to CHB over a 2 wt% Pt-10 wt% Pd catalyst. Theirs results indicated that it is possible to develop the HYD route of desulfurization of the DBT molecule over Pd–Pt alloys, but it was necessary to design a Pd–Pt catalytic system with a similar good performance but employing lower noble metal contents. Such high Pt and Pd loads seem prohibitive to the commercial HDS processes. On the other hand, Niquille-Ro¨thlisberger and Prins [11] have prepared high HDS performance low content noble metal catalysts and developed the HYD route of desulfurization for 4,6-DMDBT. However, for this model molecule DDS is practically unfeasible [7]. From this standpoint the HDS performance of PdPt-Ac/Al–Si is quite extraordinary and relevant to the future design of highly HYD selective catalytic materials aimed to be employed in deep HDS processes. Other important differences in SHYD and products distribution as a function of the support must be commented.
123
294
In first place, partially hydrogenated DBT intermediates in high amounts are registered for the Pd–Pt catalysts and not for CoMo and NiMo. It may be noted that significant amounts of such intermediates (Table 1) were detected for the silica supported Pd–Pt catalyst (PdPt-Ac/Si). This indicates that PdPt-Ac/Si is able to develop HYD but it seems to experience certain level of limitation regarding the ability to perform the final C–S bond scission step in HYD as compared to the other Pd–Pt catalysts which yielded lower amounts of such intermediates. On the other hand, if one compares the performance of the catalysts supported on alumina (PdPt-Ac/Al-P and PdPt-Ac/Al) but prepared from the same precursor, it is important to remark how HYD over Pd–Pt is highly susceptible to a probable change in the nature of the alumina support. It is known that the surface chemistry of every alumina is rather complex [20] and it can influence the catalytic activity in HDS [21], but not usually HYD selectivity in DBT desulfurization. Our results for PdPt-Ac/Al and PdPt-Ac/ Al-P catalysts are similar to those presented by NiquilleRo¨thlisberger and Prins [11], who reported that PdPt/cAl2O3 catalysts, prepared from nitrate precursors, have a better selectivity to DDS over HYD, and also displayed low amounts of non-desulfurized DBT intermediates. This can be very advantageous when performing the HDS of nonsterically hindered DBT molecules which in turn require less hydrogen consumption in the removing of sulfur [2]. In general, the alumina supported Pd–Pt catalysts prepared from organometallic precursors possess a higher DDS selectivity (%yBP) compared to PdPt-Ac/Si. Moreover, as mentioned before, the C–S–C bond breaking capacity of alumina supported Pd–Pt catalysts is higher than that of PdPt-Ac/Si. Therefore, the differences in the nature of the carriers are clearly modulating the selectivity in HDS over Pd–Pt. Although extensive information exists concerning the effect of support on HDS reactions over MoS2 based catalysts [1–3, 5, 6, 15–17], regarding Pd–Pt catalysts, to our knowledge, there are no works in open literature analyzing the effect of the support on the selectivity of this catalytic system in the HDS of DBT. An in-depth analysis of this effect is still required. On the other hand, comparing the behavior of the alumina supported catalysts prepared from chloride precursors (PdPt-Cl/Al-P) with those prepared from the organometallic precursors it is clear that the nature of the precursors of the active elements can really manipulate both HDS activity and HYD over Pd–Pt. While %yCHB for PdPt-Ac/Al-P is 22% this value is doubled when employing chloride precursors PdPt-Cl/Al-P. In addition, the %CDBT of the latter catalyst (PdPt-Cl/Al-P) is c.a. 70%, while in the former (PdPt-Ac/Al-P) it reaches 44%. A comparison of the products distribution shows that the use of chloride precursors mostly favors conversion of DBT to CHB, as the yield to BP and DBT partially hydrogenated
123
V. G. Baldovino-Medrano et al.
intermediates decreases. Therefore, even the possible presence of Cl ions remaining attached to the alumina surface [22] can have a deep impact in the development of HYD over Pd–Pt catalysts. Such Cl ions are known to increase alumina acidity of reforming noble metal catalysts thus favoring their HYD function [23, 24]. This points out to the importance of carrier acidity, either intrinsic or due to doping substances like Cl. Acidity can enhance HDS rates as well as thiotolerance of noble metal based catalysts [13, 24, 25]. Galindo and de los Reyes [25] found a correlation between thiophene HDS, thiotolerance and acidity for alumina–titania Pd–Pt supported catalysts. Barrio et al. [13] ascribed the high activity of Pd–Pt/ASA catalysts in the HDS of DBT and HYD of naphthalene partly to strong Brønsted acidic sites. However, they did not present results on HYD selectivity, but reported that compared to monometallic Pd/ASA a Pd–Pt/ASA catalyst displayed a much higher C–S hydrogenolysis capacity. For conventional MoS2 based catalysts it has been claimed that an increase in support’s acidity promotes HYD pathway [5, 6, 17, 26, 27], but only minor increases are reported. This effect is also promoted by dopants such as F [27]. It can be speculated that a similar phenomena is responsible of enhancing the HDS performance of the Pd–Pt alloy by increasing conversion of dibenzothiophene via HYD. Of course, at the current stage of this research no rigorous conclusions can be drawn concerning a relationship between acidity and the observed catalytic behavior registered, for the acidic properties of the presented materials are currently being measured. What is important is to notice that in order to observe non-trivial changes in the HYD/DDS selectivity of DBT over conventional Co(Ni)Mo sulfided catalysts drastic changes in the nature of the support must be made. Large amounts of literature on the subject indicate so. All the above presented evidence points out to the high sensitivity of the Pd–Pt systems to ‘‘soft’’ changes in the nature of the carrier for the development of HYD during the HDS of DBT. In the same sense, it must be noticed that the Al2O3–SiO2 support prepared has a very low Si/Al ratio (Si/Al = 0.1) which allows to think that it would not possess very strong acidic sites. This aspect is very important because highly acidic supports tend to favor undesirable side reactions during HDS [28] and are additionally more susceptible to poisoning from hydrodenitrogenation [29]. Finally, the data confirm the fact that Pd–Pt catalysts exhibit a balance of catalytic functions different from that displayed by conventional NiMo or CoMo catalysts.
4 Conclusions It has been found that the HDS of DBT can be selectively oriented to the HYD reaction pathway over a silica–
Highly HYD Selective Pd–Pt/support Hydrotreating Catalysts
alumina supported Pd–Pt catalyst of very low Si/Al ratio. These results, in particular the role of chlorine, confirm the fact that the reaction pathway in DBT hydrotreating over Pd–Pt strongly depends on the nature of the carrier. They also confirm that the catalytic behavior in HDS of these catalysts notably differs from that of conventional sulfided CoMo and NiMo catalysts. Finally, the developed catalysts are very promising for the HDS of sterically hindered b-alkyl-DBTs. Acknowledgments This work was possible due to the financial support of COLCIENCIAS, a government institution that promotes science and technology in Colombia, in the frame of the project 110206-17636. V. G. Baldovino-Medrano thanks COLCIENCIAS and Universidad Industrial de Santander for their financial support. Authors gratefully thank Prof. B. Delmon for fruitful discussions on the manuscript.
References 1. 2. 3. 4.
Song C, Ma X (2004) Intern J Green Energy 1:167 Whitehurst DD, Isoda T, Mochida I (1998) Adv Catal 42:345 Bej SK, Maity SK, Turaga UT (2004) Energy Fuels 18:1227 Mijoin J, Pe´rot G, Bataille F, Lemberton JL, Breysse M, Kasztelan S (2001) Catal Lett 71:139 5. Pawelec B, Halachev T, Olivas A, Zepeda TA (2008) Appl Catal A Gen 348:30 6. Zepeda TA, Pawelec B, Fierro JLG, Olivas A, Fuentes S, Halachev T (2008) Microp Mesop Mat 111:157 7. Yang H, Fairbridge C, Ring Z (2003) Energy Fuels 17:387
295 8. Vrinat ML (1983) Appl Catal 6:137 9. Reinhoudt HR, Troost R, van Langeveld AD, Sie ST, van Veen JAR, Moulijn JA (1999) Fuel Proc Tech 61:133 10. Kabe T, Qian W, Hirai Y, Li L, Ishihara A (2000) J Catal 190:191 11. Niquille-Ro¨thlisberger A, Prins R (2006) J Catal 242:207 12. Niquille-Ro¨thlisberger A, Prins R (2007) Catal Today 123:198 13. Barrio VL, Arias PL, Cambra JF, Gu¨emez MB, Pawelec B, Fierro JLG (2003) Fuel 82:501 14. Cai Q, Luo Z-S, Pang W-Q, Fan Y-W, Chen X-H, Cui F-Z (2001) Chem Mater 13:258 15. Giraldo SA, Centeno A (2008) Catal Today 133–135:255 16. Effects of support in hydrotreating catalysis for ultra-clean fuels (2003) Catal Today 86:1–288 17. Pe´rot G (2003) Catal Today 86:111 18. Ancheyta J, . Froment GF (Eds) International symposium on advances in hydroprocessing of oils fractions (ISAHOF 2007) Catal Today 130 (2008) 1–264 19. Baldovino-Medrano VG, Giraldo SA, Centeno A (2008) J Mol Cat A Chem 87:1917. doi:10.1016/j.molcata.2008.11.021 20. Kelber JA (2007) Surf Sci Rep 62:271 21. Adachi M, Contescu C, Schwarz JA (1996) J Catal 162:66 22. Jackson SD, Willis J, Mclellan GD, Webb G, Keegan MBT, Moyes RB, Simpson S, Wells PB, Whyman R (1993) J Catal 139:191 23. Łomot D, Juszczyk W, Karpin´ski Z (1997) Appl Catal A Gen 155:99 24. Stanislaus A, Cooper BH (1994) Catal Rev-Sci Eng 36:75 25. Galindo IR, de los Reyes JA (2007) Fuel Proc Tech 88:859 26. Pophal C, Kameda F, Hoshino K, Yoshinaka S, Segawa K (1997) Catal Today 39:21 27. Kim H, Lee JJ, Moon SH (2003) Appl Catal B Env 44:287 28. Guo H, Sun Y, Prins R (2008) Catal Today 130:249 29. Furimsky E, Massoth FE (2005) Catal Rev 47:297
123
