Selectivity Effect of Oxygenates in Hydrocracking of Fischer-Tropsch Waxes

Transcription

1 662 Energy & Fuels 2007, 21, Selectivity Effect of Oxygenates in Hydrocracking of Fischer-Tropsch Waxes Dieter Leckel* Fischer-Tropsch Refinery Catalysis, Sasol Technology Research and DeVelopment, P. O. Box 1, Sasolburg 1947, South Africa ReceiVed NoVember 29, ReVised Manuscript ReceiVed January 16, 2007 The potential effect of oxygenates on the hydrocracking of iron-based Fischer-Tropsch (FT) wax was investigated using bifunctional amorphous silica-alumina base metal and noble metal catalysts. Hydrogenation of the FT wax resulted in higher hydrocracking activity and increased conversion compared to unhydrogenated wax, but contrary to expectations, a higher diesel selectivity was obtained with the untreated wax. Experiments with a hydrogenated FT wax fraction whereby oxygenate model compounds such as tetradecanol and lauric acid were added to the wax showed that the oxygenates selectively affected the product spectrum by competitive adsorption with hydrocarbons on the active sites. Analyses of the product spectrum led to the conclusion that the oxygenates impacted the catalyst s metal-acid balance, which subsequently changed the bifunctional behavior of the original catalyst. The presence of the alcohol in the hydrogenated FT wax improved the hydrocracking selectivity to diesel. Addition of the acid led to more secondary cracking and as a result a higher selectivity to lower-boiling hydrocarbons. The hydrocracking product spectrum of a hydrogenated FT wax can consequently be successfully manipulated by the selective addition of oxygenates. Introduction Low-temperature Fischer-Tropsch (LTFT) synthesis based on indirect coal liquefaction produces predominantly straightchain hydrocarbons, resulting in waxes that are essentially free of aromatics and sulfur. 1 The FT synthesis products produced by cobalt-based or iron-based FT catalysts contain also olefins and oxygenates, with the iron-based catalyst producing a higher concentration of olefins, because iron is less hydrogenating than cobalt. The hydrocracking of LTFT wax using commercial bifunctional catalysts has been researched at Sasol since the 1970s. 2 Mild hydrogenation of LTFT synthesis products prior to hydrocracking improved the conversion significantly. 3 Oxygenates seemed to modify the catalyst properties by competitive adsorption on the active sites, affecting the balance of dehydrogenation/hydrogenation to acid sites. 4 In 1965, Union Oil Company of California was assigned a patent 5 in which water (as steam) and water precursors such as alcohols, ketones, aldehydes, and esters were added during the hydrocracking of crude oil fractions boiling in the gas oil range to improve the selectivity to gasoline. Water in concentrations below 2 wt % (based on the feed) was found to be beneficial. The intrinsic activity of the hydrocracking catalyst was thereby improved. This was not surprising, because the activity of amorphous silica-alumina catalysts is maximized at 1.4% * Tel.: ; fax: ; dieter.leckel@sasol.com. (1) Dry, M. E. Catal. ReV. Sci. Eng. 1981, 23 (1,2), (2) Dry, M. E. Catal. Today 2002, 71, (3) Leckel, D. Energy Fuels 2005, 19, (4) Leckel, D. Oxygenates in Fischer-Tropsch Waxes - A Threat or Opportunity in Fuels Hydrocracking. Proceedings of 7th European Congress on Catalysis, Sofia, Bulgaria, August 28-September 1, 2005; Book of Abstracts: Paper O5-05; p 164. (5) Peralta, B. U.S. Patent 3,173,853, 1965; assigned to Union Oil Company of California. water. 6 Another U.S. patent 7 reports the successful hydrocracking of higher-boiling hydrocarbon fractions ( C) to transportation fuels in the presence of water at low hydrogen partial pressures. The upgrading of FT syncrude using a hydroprocessing catalyst hydroconversion step is described by the Shell Middle Distillate Synthesis Process. 8 During hydroconversion, a fraction of the primary FT hydrocarbon product is hydrogenated without substantial hydroisomerization or hydrocracking converting the minor quantities of olefins and oxygenates. The resulting n-paraffins are finally hydroisomerized and hydrocracked to fuels. It was found that using the olefin and oxygen-free feed enhanced the selectivity of the hydrocracking catalyst to gas oil. 9 In addition, the catalyst lifetime and product properties improved. A process for preparing lubricant basestocks is provided by Cody et al., 10 where the dewaxing catalyst was selectivated by oxygenates. An increase in the degree of isomerization and a reduction in the cracking activity after the treatment of a dewaxing catalyst (ZSM-48 zeolite loaded with 0.6 wt % platinum) with a mixture containing carboxylic acids, alcohols, esters, aldehydes, ethers, and ketones were observed. The latter is a composition of oxygenates typically found in FT waxes. Abazaijan et al. 11 reported, on the other hand, that oxygenates in FT synthesis products, most predominantly the alcohols, have a deactivating effect on the hydroprocessing catalyst employed in FT syncrude workup units. Water and oxygen-containing compounds may affect the catalyst (6) Finch, J. N.; Clark, A. J. Phys. Chem. 1969, 73, (7) Egan, C. J. U.S. Patent 4,097,364, 1978; assigned to Chevron Research Company. (8) Eilers, J.; Posthuma, S. A.; Sie, S. T. Catal. Lett. 1990, 7, (9) Eilers, J.; Posthuma, S. A. EP 0, 583,836, A1, 1993; assigned to Shell Int. Res. Co. (10) Cody, I. A.; Murphy, W. J.; Hantzer, S. S. WO 2004/ A2, 2004; assigned to ExxonMobil Research and Engineering Co. (11) Abazaijan, A.; Tomlinson, H. L.; Havlik, P. Z.; Clingan, M. D. EP 1,449,906 A1, 2004; assigned to Syntroleum Corporation /ef060603h CCC: $ American Chemical Society Published on Web 02/23/2007

2 SelectiVity Effect of Oxygenates Energy & Fuels, Vol. 21, No. 2, Table 1. Physical Properties of the Hydrocracking Catalyst catalyst properties NiMoSiO 2-Al 2O 3 PtWSiO 2-Al 2O 3 Pt, mass % 0.3 NiO, mass % 3.2 MoO 3, mass % 15 WO 3, mass % 2.2 SiO 2/Al 2O 3 mass ratio BET surface area, m 2 /g 360 pore volume, cm 3 /g 0.30 total acidity, mmol NH 3/g catalyst 0.21 activity by competitive adsorption 12 on active sites, which is nondestructive with regard to the catalyst. Furthermore, oxygenates and water can deactivate a catalyst by acting on their structure or chemical composition, which can be seen as having a destructive impact on the catalyst. In general, the influence of oxygenates depends on the type of catalyst and operating conditions. In this work, unhydrogenated (UnH) and hydrogenated (H) FT waxes were hydrocracked, and the influence of the different feeds on product selectivities and yields was studied. Tetradecanol and lauric acid were chosen as representative FT wax oxygenates and added to a hydrogenated FT wax to investigate their effect on the bifunctional catalyst behavior during the hydrocracking reaction. Experimental Catalysts. A commercial sulfided NiMo on an amorphous silica-alumina hydrocracking catalyst and an in-house-prepared amorphous silica-alumina-supported noble metal catalyst modified by the addition of WO 3 was used for the hydrocracking tests. The Siral75 support material, which is noncommercial, was supplied by Sasol Germany GmbH (formerly Condea Chemie GmbH). The Siral material is manufactured according to procedures describe in the patent literature. 13 Aluminum alkoxides with a carbon distribution of C 2 to C 20+ were thereby hydrolyzed with deionized water and stirred for approximately 45 min. The aqueous alumina suspension separating from the alcohols was then mixed with orthosicilic acid deionized by means of an ion exchanger in a total quantity of 3% by weight. The solid suspension obtained was dried with a spray drier at C. The amorphous support material Siral75 contained 75 mass % silica and 25 mass % alumina. Some characteristic physical properties of the catalysts are listed in Table 1. The Brunauer-Emmett-Teller (BET) surface area (by nitrogen physisorption) and pore volume were measured using equipment supplied by Micromeritics. Temperature-programmed desorption (TPD) of ammonia was performed with a Micromeritics TPD 2900 NH 3 analyzer for the determination of total acidity. The metal loading was determined by inductively coupled plasma. The analyses of the commercial catalyst were restricted due to an agreement with the catalyst supplier. Reactor System. A bench-scale fixed-bed reactor (27.6 mm i.d. and 1.8 m length) operating in the downflow mode was used for the isothermal studies. The reactor system was described in a previous work. 3 Noble Metal Catalyst Impregnation Procedures and Catalyst Activation. 14 The catalysts were prepared by incipient wetness impregnation. Thus, the Siral support material (1.5 mm extrudates, 100 g) was calcined in air applying a heating rate from ambient to 500 C at2 C/min. The temperature was kept at 500 C for3h. (12) O Connor, C. T.; Langford, S. T.; Fletcher, J. C. Q. Proc. of 9th Int. Zeolite Conference, Montreal, 1992; von Bollmoos, R., Eds., et al., Butterworth-Heinemann: 1993; pp (13) Meyer, A.; Noweck, K.; Reichenauer, A.; Schimanski, J.; German Patent DE C1, Meyer, A.; Noweck, K.; Reichenauer, A.; Schimanski, J. U.S. Patent 5,045,519, 1991; assigned to Condea Chemie GmbH. The calcined material was then impregnated with a solution of ammonium metatungstate [(NH 4 ) 6 H 2 W 12 O 40 xh 2 O, Fluka, g85% WO 3 ] sufficient to load 3.0 mass % WO 3. The material was again calcined applying a heating rate of 2 C/min from ambient to 500 C and held at the latter temperature for 4 h. Thereafter, 100 g of the calcined material was subsequently impregnated with an aqueous solution of [Pt(NH 3 ) 4 ](NO 3 ) 2 (99.995% purity, Sigma Aldrich) containing the required amount to obtain 0.30 mass % of platinum in the final catalyst. The catalytic material was calcined once more at 500 C in air for 4 h. Following this, the catalyst was activated in situ by first drying the material at 120 C for2hinnitrogen (0.5 l/min) and subsequently conducting the reduction at 400 C for 3 h using a hydrogen flow of 1.7 l/min (heating rate of 25 C/h from 120 to 400 C). The NiMo catalyst was predried in nitrogen at 125 C and 0.1 MPa for 8 h prior to sulfidation. Then, nitrogen was replaced by hydrogen (100 l/h flow rate) and the reactor pressurized to 3.0 MPa. Dimethyldisulfide (2 mass %) in a C 9 -C 11 paraffin mixture was used as a sulfiding agent. The catalyst was first wetted thoroughly with the sulfiding mixture using a pump speed of 500 ml/h, whereafter the pump rate was decreased to 200 ml/h. The temperature was increased hourly from 125 C in steps of 25 to 250 C, where it was kept for 4 h. H 2 S breakthrough was monitored in the exit gas stream by means of Draeger tubes. A 2000 vppm H 2 S concentration was seen as sufficient to increase the temperature in steps of 25 C further to 350 C, where it was kept for another 2 h. Thereafter, the reactor temperature was decreased to 240 C, which completed the sulfiding. Then, the reactor pressure was raised and the desired hydrogen feed rate established. To ensure that the wax did not solidify in the lines during the test runs, all lines to and from the reactor were heated to above the melting point of the wax. Dimethyl disulfide was continuously added to the wax feed during hydrocracking. Experience at Sasol has proven that a value of 200 vppm in the exit gas is sufficient to keep the catalyst in its active state. Procedure. Catalyst activity was monitored by drawing product samples from the reactor after steady-state conditions were reached, typically 72 h after reaction conditions were changed. The following 8 h period was then used to collect a representative sample for product analysis. Mass balances of % were achieved with collection of the tail gas, condensed lighter hydrocarbons, and liquid products from the reactor stream. Condensed light hydrocarbons were refrigerated prior to analysis. Operating Conditions. Hydrocracking was performed in the temperature range of C. A pressure of 7.0 MPa was applied. The liquid hourly space velocity (LHSV) was varied between 0.50 and 1.0 h -1, and the hydrogen-to-feed ratio was adjusted between 1000 and 1500:1 m 3 /m 3 (gas volume expressed at 0 C, 101 kpa). Experiments were performed in the once-through mode. Feed. LTFT reactors produce gas, condensates, and reactor wax. The average liquid product yields from an iron-based slurry reactor are 38 wt % condensates and 62 wt % reactor wax. Linear paraffins are the major constituents in the products. The wax is essentially free of sulfur and aromatics. The iron-based slurry LTFT reactor wax consists of molecules with a carbon number between C 10 and C 120 and peaks at a carbon number of C 30. The wax consists of 92 mass % n-paraffins, 5.5 mass % isoparaffins, 1.5 mass % oxygenates, and about 1 mass % R-olefins (Figure 1). C80 wax is a commercially produced wax fraction at Sasol, distilled from reactor wax which is subsequently hydrogenated to remove all unsaturates and oxygenates. It consists of paraffin molecules between C 20 and C 60 and peaks at the carbon number C 38 (Figure 2). In house Sasol analytical procedures based on modified American Society for Testing and Materials (ASTM) methods are used for the wax feed analyses including the ASTM methods D-1386, (14) Liwanga-Ehumbu, A-M.; Visagie, J. L.; Leckel, D. O. GB , April 23, 2003; U.S. Patent 2003/ A1, Sept 18, 2003; WO 01/90280 A2, April 03, 2003; AU B2, Patent No , May 24, 2001; all assigned to Sasol Technology.

3 664 Energy & Fuels, Vol. 21, No. 2, 2007 Leckel Figure 1. Carbon number distribution of the Sasol iron-based low-temperature Fischer-Tropsch (LTFT) slurry bed reactor wax (figure includes all relevant components including oxygenates). Figure 2. Carbon number distribution of the hydrogenated iron-based C80 wax. D-1387, and D5622 for oxygenate determination. The carbon number distributions were determined with a high-temperature gas chromatograph (GC) on a 15 m Restek, 0.28 mm i.d., and 0.15 µm-phase thickness MXT-1 metal capillary column. The wax sample was spiked with a C 36 internal standard and dissolved in 25 ml of xylene. The instrumental conditions were as follows: initial column temperature of 60 C and initial column holding time of 5 min; column ramp rate of 10 C/min; final column temperature of 440 C with a final column holding time of 5 C; initial injector temperature of 40 C; initial injector holding time of 0 min; injector ramp rate of 70 C; final injector temperature of 420 C; final injector holding time of 48 min; hydrogen carrier gas; detector temperature of 450 C; sample volume injection of 1 µl. Product Analyses. The reactor tail gas was collected in a glass gas sample bomb that was connected to the reactor exit stream, and the gas was then analyzed by gas chromatography with flame ionization detection (GC-FID). All liquid products were collected and fractionated into a naphtha (C 5 -C 9 ) cut, diesel (C 10 -C 22, corresponding to C), and a residue product boiling above 370 C (i.e., C 23+ ). Simulated distillation (ASTM D-2887) was applied to obtain the boiling-point distribution of the liquid products. An Agilent 6890N GC-FID was used for the analyses to calculate conversion, selectivity, and yields. The products were separated on a 50 m HP PONA methyl siloxane column with 200 µm internal diameter and 0.5 µm film thickness. Terminologies. Conversion, selectivity, and yields in this work were calculated as follows. Conversion is here understood to be true conversion, since, according to the definitions used, naphtha and diesel already present in the feed are corrected for. The hydrocracking conversion ( true conversion) is defined as described in eq 1. %C 23 + Conversion ) ( wt% C 23+ in Feed - wt% C 23 + in Product wt% C 23 + in Feed ) 100 (1) The diesel (C 10 -C 22 ) selectivity was calculated according to eq 2. %(C 10 -C 22 ) Selectivity ) wt% (C 10 -C 22 ) in Product - wt% (C 10 -C 22 ) in Feed (wt% C 23+ in Feed - wt% C 23+ in Product)/wt% C 23+ in Feed) ( 100 (2) The naphtha (C 5 -C 9 ) and gas (C 1 -C 4 ) selectivity were calculated analogously. The yields were calculated by multiplying the fractional conversion by the selectivity. Results and Discussion Differences in conversion and selectivity were found during the hydrocracking of UnH and H iron-based LTFT waxes that were attributed to the effect of the oxygenate-containing heteroatoms in the FT wax. The presence of oxygenates

4 SelectiVity Effect of Oxygenates Energy & Fuels, Vol. 21, No. 2, Figure 3. Hydrocracking of unhydrogenated (UnH) and hydrogenated (H) iron-based FT wax using a sulfided NiMo/SiO 2-Al 2O 3 catalyst at 7.0 MPa, 0.55 h -1 LHSV, and an H 2-to-wax ratio of 1500:1 m 3 /m 3 (conversion plotted over operating temperature). Figure 4. Naphtha and diesel selectivity obtained during hydrocracking of unhydrogenated (UnH) and hydrogenated (H) iron-based FT slurry reactor wax using a sulfided NiMo/SiO 2-Al 2O 3 catalyst at 7.0 MPa, 0.55 h -1 LHSV, and an H 2-to-wax ratio of 1500:1 m 3 /m 3. inhibited the hydrocracking conversion and influenced the catalyst properties, resulting in an increase in the selectivity to diesel. This led to more fundamental studies on the effect of oxygenates on the product spectrum and the catalyst performance. A. Hydrocracking of Unhydrogenated and Hydrogenated LTFT Reactor Wax with a Commercial Sulfided NiMo Catalyst Supported on Amorphous Silica-Alumina. Hydrogenated and unhydrogenated LTFT slurry reactor wax was hydrocracked at a pressure of 7.0 MPa, a 0.55 h -1 LHSV, and ah 2 -to-wax ratio of 1500:1 m 3 /m 3. In Figure 3, the conversion obtained for the two wax feeds is plotted over the operating temperature. Compared to the hydrocracking of unhydrogenated FT wax, prehydrogenating of the wax resulted in C lower operating temperatures and significantly higher conversions. At an operating temperature of 360 C, a conversion level of only 25% was observed for the unhydrogenated wax, while a 70% conversion was already achieved for the hydrogenated FT wax, which presents an almost 3-fold higher hydrocracking conversion. Operating at higher temperatures markedly improved the hydrocracking activity for the unhydrogenated wax. In Figure 4, the product selectivities for the hydrocracking of unhydrogenated and hydrogenated FT waxes are plotted against the operating temperature, which is not the generally employed way of comparing selectivity because selectivity has to be compared at constant conversion. However, it best shows the different effects of the two wax feeds. It was observed that the diesel selectivity for the unhydrogenated wax remained almost unchanged (75%) up to temperatures of 375 C, whereafter it decreased markedly. The naphtha selectivity increased accordingly from 19 to 34%. The conversion of unhydrogenated wax in the temperature range of C generally resulted in a 10-30% higher diesel and 35-50% lower naphtha selectivity. Table 2 compares the selectivities and properties for the products obtained from FT wax hydrocracking at similar conversions, 70% and 99% and 97%. It was noticed that prehydrogenation most significantly affected the diesel-tonaphtha ratio, which was always found to be higher for hydrocracking of unhydrogenated FT wax. The difference in the diesel-to-naphtha ratio though seemed to vanish upon going to high conversion levels. Diesel selectivity was slightly worse for the hydrogenated feed. The oxygenates present in the LTFT wax are more polar than the alkenes formed from the dehydrogenation of the alkanes at the metal sites and adsorb competitively with the latter at the

5 666 Energy & Fuels, Vol. 21, No. 2, 2007 Leckel Table 2. Comparison of Product Selectivity Obtained for the Hydrocracking of Hydrogenated (H) and Unhydrogenated (UnH) FT Wax Using a Sulfided NiMo/SiO 2-Al 2O 3 Catalyst at 7.0 MPa, 0.55 h -1 LHSV, and an H 2-to-Wax Ratio of 1500:1 m 3 /m 3 H UnH H UnH conversion, mass% temperature, C Selectivity, % gas (C 1-C 4) naphtha (C C) diesel ( C) Yield, % gas (C 1-C 4) naphtha (C C) diesel ( C) diesel-to-naphtha ratio acid sites, specifically at the stronger acid sites. While the number of metal sites accessible to the alkanes/alkenes (N Me ) remained constant, the number of acid sites (N A ) available to the alkenes is decreased due to competitive adsorption of the oxygenates. This leads to a change in the acid-metal balance of the catalyst. Since the ratio of N Me /N A increased, the possibility that a carbenium ion or alkene intermediate is exposed to more than one acid site is reduced. Thus, the formation of multibranched isomers and secondary cracking is limited because the possibility that the intermediate alkene encounters a metal side is higher than encountering an acid site for further isomerization or cracking. As a consequence, the diesel selectivity is higher and the naphtha formations lower during the hydrocracking of unhydrogenated wax. Similar observations were reported by Alvarez et al., 15 who hydroisomerized and hydrocracked n-decane over a PtHY catalyst with different metal-to-acid ratios. They found low cracking conversions at high metal-to-acid site ratios, while the formation of light products and the conversion was favored at low ratios of N Me /N A. B. The Effect of Tetradecanol and Lauric Acid on the Hydrocracking of an Iron-Based Hydrogenated FT Wax (C 20 -C 60 ) Using an Amorphous Noble Metal Hydrocracking Catalyst. Since FT wax contains a variety of oxygenate compound classes, it is difficult to determine which oxygenate species is the most likely to interfere with the hydrocracking reaction. Therefore, we decided to investigate the effect of the most abundant oxygenate classes present in FT waxes, namely, alcohols and acids. A fully hydrogenated iron-based FT wax fraction (C80 wax presented in Figure 2) was used as an oxygenate-free wax basis to which 5 mass % of the relevant oxygenate, tetradecanol and lauric acid, respectively, was added. The waxes were then hydrocracked at 7.0 MPa, a temperature of 370 C, a 1.0 h -1 LHSV, and a H 2 -to-wax ratio of 1200:1 m 3 /m 3 using an amorphous silica-alumina catalyst, modified with W and Pt. It was found that the presence of the alcohol had a more detrimental effect than that of the acid on the wax conversion (Table 3). The lauric acid in the wax led to a 20% lower conversion, while the addition of tetradecanol decreased the conversion significantly by almost 40%. If both oxygenate types would affect the same active sites of the catalyst, then irrespective of the type of oxygenate added, the inhibiting effect would be analogous. However, as Table 3 shows, adding the acid increased the selectivity to low-boiling hydrocarbons such as naphtha and gas, while addition of the alcohol led to an (15) Alvarez, F.; Ribeiro, F. R.; Perot, G.; Thomazeau, C.; Guisnet, M. J. Catal. 1996, 162, Table 3. The Effect of Lauric Acid and Tetradecanol on the Conversion and Product Selectivity during Hydrocracking of Iron-Based C80 Wax at 370 C C80 C80 + C 11COOH C80 + C 14OH conversion, mass % selectivity, % gas (C 1-C 4) naphtha (C C) diesel ( C) diesel-to-naphtha ratio iso-to-n-paraffin ratio in diesel Table 4. The Effect of Tetradecanol Addition on the Product Selectivity and the iso-to-n-paraffin Ratio in the Diesel for Hydrocracking the Hydrogenated C80 Wax at Similar Conversions (Conversion and Selectivity in Mass %) selectivity temp., C conversion diesel naphtha naphtha ratio iso-to-n-paraffin ratio in diesel diesel-to- C C 80 + C 14OH C C 80 + C 14OH increase in the diesel fraction. Consequently, the alcohol must have affected different active sites. The addition of the acid compound led to a 9% decrease in the diesel and a 12% increase in naphtha selectivity, while the presence of the alcohol increased the selectivity to diesel by ca. 15% and decreased that of the naphtha by 30%. As a consequence, hydrocracking the alcohol-containing LTFT wax increased the diesel-to-naphtha ratio substantially (over 60%). The addition of lauric acid led to a 2% higher gas formation as compared to the clean wax. Specifically, a higher methane content was observed, which obviously resulted from the decarboxylation of the acid with subsequent hydrogenation to methane. Similar results were reported previously for the hydrocracking of n-hexadecane with added hexanoic acid. 4 Table 4 illustrates the effect of tetradecanol on the product selectivity and yield as well as the iso-to-n-paraffin ratio in the diesel for wax hydrocracking compared at similar conversions (57% and 95%). At the lower conversion (57%), the presence of tetradecanol in the FT wax resulted in a slight increase (4%) in diesel selectivity and a 10% decrease in naphtha selectivity, thus resulting in a ca. 20% higher diesel-to-naphtha ratio. At the high conversion level (95%), the selectivity to diesel increased by 8% while that to naphtha decreased by about 23%. The lower naphtha selectivity obtained can in first approximation be attributed to the lower isomerization activity of the catalyst, as reflected by the lower iso-to-n-paraffin ratio in the diesel inhibiting secondary cracking reactions. Oxygenates such as alcohols or carboxylic acids in FT feeds are known to influence the activity of acid catalysts, 16,17 and from our results, it seems apparent that the alcohol addition affected the wax hydrocracking the most. Alcohols are expected to adsorb on acidic catalysts in competition with the olefins generated during the hydrocracking reaction; however, they are easily dehydrated to produce an olefin and water. 18 The water generated could adsorb on the catalyst surface and affect the catalyst activity by converting Lewis acid sites into Brønsted acid sites, thus increasing the acidity of the catalyst. If it would be true that the presence of water would increase the Brønsted acidity of the catalyst, an increased formation of low-boiling (16) Smook, D.; de Klerk, A. Ind. Eng. Chem. Res. 2006, 45, (17) Cowley, M. Energy Fuels 2006, 20, (18) Berteau, P.; Delmon, B.; Dallons, J.-L.; van Gysel, A. Appl. Catal., A 1991, 70,

6 SelectiVity Effect of Oxygenates Energy & Fuels, Vol. 21, No. 2, material should be observed with the alcohol-containing C80 wax compared to the plain C80 wax at similar conversion (Table 4). However, this is not the case. Similar findings were reported by Corma et al., 19 who investigated the interaction of water with zeolite catalyst surfaces during catalytic cracking. It was found that under cracking reaction conditions neither an increase of Brønsted acidity nor a competing adsorption effect of water occurred. In addition, according to calculations for the vapor pressure of water using the UNIFAC method 20 and taking our reaction conditions applied into account, water should readily desorb from the catalyst. The addition of tetradecanol in our study changed on the one hand the acid-metal balance of the original catalyst and suppressed the isomerization by decreasing the number of acid sites available for carbenium ion reactions through competitive adsorption of the oxygenate, while the number of metal sites on the other hand remained constant. Thus, relatively more metal sites were available for the (de)hydrogenation reaction. As a consequence, the unsaturated alkene intermediates are hydrogenated rather than being subjected to further acid-catalyzed skeletal rearrangements or cracking reactions, which suppresses secondary cracking and the formation of low-boiling hydrocarbons. This in turn rationalizes the observed lower iso-to-nparaffin ratio in the diesel as well as the improved diesel and lower naphtha selectivity. This is characteristic of ideal hydrocracking. 21,22 The effect of the lauric acid addition on the product selectivity was not as expected. It was assumed that, as in the case of the alcohol addition, an increase in diesel selectivity would occur during hydrocracking of the acid-containing C80 wax. However, contrary to expectations, a decrease in diesel selectivity and an increase in naphtha selectivity were observed. We propose that, due to its acidic nature, instead of adsorbing on the acid sites, the lauric acid preferentially adsorbs on the metal site. The adsorption of carboxylic acids on metals and transition metal oxides with its subsequent decomposition was previously reported as a probe reaction for the catalytic properties of metals and metal oxides. 23,24 Also, the formation of surface carboxylate species has been reported for Pt before. 25,26 This adsorption (19) Corma, A.; Marie, O.; Ortega, F. J. J. Catal. 2004, 222, (20) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of Gases and Liquids, 4th ed.; McGraw-Hill: New York, (21) Debrabandere, B.; Froment, G. F. Stud. Surf. Sci. Catal. 1997, 106, 379. (22) Thybaut, J. W.; Narasimhan, C. S. L.; Denayer, J. F.; Baron, G. V.; Jacobs, P. A.; Martens, J. A.; Marin, G. B. Ind. Eng. Chem. Res. 2005, 44, decreases the metal-acid number ratio (N Me /N A ), resulting in nonideal hydrocracking behavior leading to a lower diesel-tonaphtha ratio. Coonradt and Garwood 27 also described the reactions and product distributions found for catalysts with different hydrogenation activities. Catalysts with a highly active hydrogenation component, in our case representative of the catalyst where the oxygenate (alcohol) adsorbed on the acid site, showed controlled primary hydrocracking, resulting in a product spectrum containing higher molecular weight cracked paraffins. With regards to our study, we observed the higher diesel-to-naphtha ratio. By contrast, a catalyst with a lower hydrogenation function, in our case representative of the catalyst where the oxygenate (lauric acid) adsorbed on the metal site, led to more secondary cracking. In our study, we observed a lower diesel-to-naphtha ratio, the consequence of the higher naphtha formation. Conclusion Hydrogenation of Fischer-Tropsch wax leads to higher conversion during hydrocracking; however, at similar conversion, a higher diesel selectivity was found for hydrocracking unhydrogenated wax, which was attributed to the oxygenates present in the wax. The addition of tetradecanol to a hydrogenated FT wax improved the hydrocracking selectivity to diesel, while the addition of lauric acid led to a decrease in diesel and an increase in naphtha as well as gas selectivity. The results suggest that the oxygenates change the balance of metal-to-acid sites of the original catalyst by competitive adsorption on the active sites. The alcohol adsorbs on the acid sites thereby increasing the metal-to-acid site number ratio, while the carboxylic acid preferentially adsorbs on the metal sites resulting in a decrease in the metal-to-acid site number ratio. Acknowledgment. The author gratefully acknowledges the technical contributions of M. A. Liwanga-Ehumbu, B. Mothebe, and G. G. Swiegers (302/97) and also appreciates the permission of Sasol Technology Research and Development to publish this work. EF060603H (23) Madix, R. J. Catal. ReV.-Sci. Eng. 1984, 26, 281. (24) Kim, K. S.; Barteau, M. A. Langmuir 1988, 4, (25) Avery, N. R. Vacuum Sci. Technol. 1982, 20, 592. (26) Avery, N. R. Appl. Surf. Sci. 1982, 11/12, 774. (27) Coonradt, H. L.; Garwood, W. E. Ind. Eng. Chem. Proc. Des. DeV. 1964, 3 (1), 38.