Energy & Fuels 1996, 10, 587-590
587
Naphthene Upgrading with Pillared Synthetic Clay Catalysts Edwin S. Olson* and Ramesh K. Sharma University of North Dakota, Energy & Environmental Research Center, P.O. Box 9018, Grand Forks, North Dakota 58202 Received November 28, 1995. Revised Manuscript Received February 27, 1996X
Catalytic hydrotreatment of methylcyclohexane was investigated as a model for upgrading of coal-derived naphthenes to isoparaffinic structures with synthetic clay catalysts. Nickelsubstituted synthetic mica montmorillonite (NiSMM), alumina-pillared NiSMM, and zirconiapillared NiSMM were prepared and tested for hydrocracking and hydroisomerization of methylcyclohexane. Total acidity and surface area increased with pillaring of NiSMM with polyoxyaluminum and polyoxyzirconium cations. Catalytic reactions of the methylcyclohexane gave high conversions of the substrate to aliphatic hydrocarbons. Most of the products were branched alkanes (isoparaffins). These compositions are highly desirable for environmentally acceptable transportation fuels. Dehydrogenation was not a major pathway, as indicated by the minimal formation of aromatic hydrocarbons, coke, or other oligomeric materials. This paper describes the effect on the product distribution of various operating conditions, which included reaction temperature, contact time, hydrogen pressure, and catalyst.
Introduction A growing need exists for new and improved catalysts to improve the overall economics of coal liquefaction. Coal-derived liquids are heavy, viscous, aromatic, heterocompound-rich, and very difficult and expensive to refine into transportation fuels. Extensive hydrogenation with molybdenum sulfide catalysts can convert the coal liquid to substituted cycloalkanes (naphthenes). Further catalytic hydroisomerization and hydrocracking of naphthenes to isoparaffins without reversion to aromatics is required to meet current gasoline standards. Natural clays were used as catalysts in petroleum cracking until replaced by more active and selective zeolites.1 Synthetic clays (SMM) containing a hydrogenation component were more active for hydrocracking and hydroisomerization.2-4 Synthetic clays such as nickelsubstituted synthetic mica montmorillonite (NiSMM) were effective for cleaving carbon-carbon bonds and isomerizing the alkyl chains to give branched alkanes.5-8 These catalysts have also been found to be effective in hydrocracking and reforming of biomass materials into the isoparaffinic composition desired for transportation fuels.9,10 Unlike zeolites, very low concentrations of aromatic components were obtained using synthetic clays. Traditional platinum catalysts may be extensively poisoned by the sulfur compounds present in the Abstract published in Advance ACS Abstracts, April 1, 1996. (1) Gates, B. C.; Katzer, J. R.; Schuit, G. C. A. Chemistry of Catalytic Processes; McGraw-Hill: New York, 1979; p 1. (2) Jaffe, J.; Kittrell, J. R. U.S. Patent 3 535 229, 1970. (3) Mulaskey, B. F. U.S. Patent 3 682 811, 1972. (4) Csicsery, S. M.; Mulaskey, B. F. U.S. Patent 3 655 798, 1972. (5) Swift, H. E.; Black, E. R. Ind. Eng. Chem. Prod. Res. Dev. 1974, 13, 106. (6) Black, E. R.; Montagna, A. A.; Swift, H. E. U.S. Patent 4 022 684, 1977. (7) Black, E. R.; Montagna, A. A.; Swift, H. E. U.S. Patent 3 966 642, 1976. (8) Olson, E. S.; Sharma, R. K. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1994, 39, 706.
coal-derived feedstocks; however, NiSMM is not deactivated by sulfur, although nitrogen compounds may inhibit the acidic catalyst.11,12 In pillared clays, intercalation of hydroxylated or complexed metal cations maintains the clay layer structure after loss of water and generates larger pore sizes. These structures are stable to temperatures over 600 °C13 and even 800 °C when pillaring is performed with mixed colloidal clusters.14 Pillared natural clays impregnated with a hydrogenation component have been reported, but the potential for synthetic and pillared synthetic clays in naphthene reforming has not been thoroughly explored. The goal of this study was to develop an understanding of the conditions and catalysts appropriate for hydrocracking and hydroisomerization of coal-derived cycloalkanes (naphthenes). Catalytic hydrotreating reactions of methylcyclohexane (MeCH) were conducted to evaluate the potential of synthetic and pillared synthetic clay catalysts for the upgrading of coal-derived liquids to an isoparaffinic product. The reactions of MeCH were carried out under several conditions to determine the effects of temperature, reaction time, and hydrogen pressure on the hydrocracking and hydroisomerization activity of NiSMM, alumina-pillared NiSMM, and zirconia-pillared NiSMM. These initial studies did not investigate the effects of sulfur and nitrogen on the MeCH reaction.
X
0887-0624/96/2510-0587$12.00/0
(9) Olson, E. S.; Sharma, R. K. In Energy from Biomass and Wastes XVI; Klass, D. L., Ed.; Institute of Gas Technology: Chicago, IL, 1993; p 739. (10) Sharma, R. K.; Olson, E. S. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1994, 39, 1040. (11) Sharma, R. K.; Olson, E. S. Semiannual technical progress report for the period January-June 1991, EERC publication, 1991. (12) Olson, E. S.; Sharma, R. K. Semiannual technical progress report for the period July-August, 1991, EERC publication, 1991. (13) Occelli, M. L. In Keynotes in Energy-Related Catalysts; Kaliaguine, S., Ed.; Elsevier: Amsterdam, 1988; pp 101-137. (14) Occelli, M. L.; Peaden, P. A.; Ritz, G. P.; Iyer, P. S.; Yokoyama, M. Macroporous Mater. 1993, 1, 99.
© 1996 American Chemical Society
588
Energy & Fuels, Vol. 10, No. 3, 1996
Olson and Sharma
Experimental Section Catalyst Preparation and Characterization. NiSMM. NiSMM was prepared by the procedure of Heinerman15 and co-workers. Sodium silicate solution was exchanged with Amberlite IR 120-H cation-exchanged resin, and a solution of nickel acetate was added and stirred with aluminum isopropylate and ammonium fluoride. After evaporation, aqueous ammonia was added, and the slurry was heated in a pressure vessel at 300 °C for 16 h. The product was separated by centrifugation and washing with deionized water, dried at 120 °C, and calcined in air at 540 °C for 6 h. Alumina-Pillared NiSMM (AlPNiSMM). Pillaring of NiSMM was carried out according to the procedure described by Gaaf and co-workers.16 A slurry of NiSMM (5 g in 1000 mL) was stirred at 70 °C for 20 h with an aged (3 day) polyoxyaluminum solution prepared from aluminum chloride (30 mmol) and sodium hydroxide (60 mmol). The pillared clay was separated by centrifugation, washed with deionized water several times, and freeze-dried. The dried solid was calcined at 350 °C for 6 h. Zirconia-Pillared NiSMM (ZrPNiSMM). ZrPNiSMM was prepared by adding the zirconia pillaring solution prepared from zirconyl chloride (75 mmol) to 5 g of NiSMM in 1000 mL of water and heating at 70 °C for 20 h. Acidity Measurements. The acidity of the solid acid catalysts was determined by pyridine adsorption and desorption studies. A small amount of sample (100 mg) was placed in a glass chamber attached to a vacuum pump, a gas inlet, and a gas outlet. The chamber was evacuated, and argon saturated with pyridine was introduced into the chamber until the weight increase ceased. At this stage, the chamber was evacuated until the physisorbed pyridine was removed, as indicated by the constant weight of the base-absorbed sample. The amount of pyridine chemisorbed was used to determine total acidity of the catalyst. Surface Area Measurements. Surface area measurements were performed with a Micromeritics AccuSorb 2100E static unit for nitrogen physisorption at 77 K (BET method). Catalytic Reactions. In a typical experiment, 0.5 g of MeCH and 0.25 g of the catalyst were placed in a tubing bomb (15-mL microreactor). The microreactor was evacuated, pressurized with the desired amount of hydrogen, and placed in a rocking autoclave preheated to the desired temperature. At the end of the reaction period, the microreactor was cooled in a dry ice-acetone slurry, degassed, and opened. The desired amount of the internal standard was added to the product slurry; the product slurry was transferred into a centrifugation tube by washing with methylene chloride; and the solid catalyst was removed by centrifugation. Quantitative gas chromatography-flame ionization detection (GC-FID) analyses of the liquid samples were performed with a HewlettPackard 5890a gas chromatograph equipped with a Petrocol capillary column. The solid was dried in a vacuum at 110 °C for 3 h. Since they were not found in the reaction products, 2,2,4-trimethylpentane and n-octadecane were added as the internal standards. Alkane components were determined from the area percent ratio with respect to octadecane, assuming the same response factor.
Results and Discussion Catalyst Characterization. The total acidity of the NiSMM as measured by the amount of chemisorbed pyridine was found to be 0.84 mmol/g of catalyst, indicating the presence of acidic sites in the catalyst. These sites are Lewis as well as Brønsted acids. Pillaring the clay with polyoxyaluminum cations increased (15) Heinerman, J. J. L.; Freriks, I. L. C.; Gaaf, J.; Pott, G. T.; Coqlegem, J. G. F. J. Catal. 1983, 80, 145. (16) Gaaf, J.; van Santen, R.; Knoester, A.; van Wingerden, B. J. Chem. Soc., Chem. Commun. 1983, 655.
Figure 1. Capillary column gas chromatogram of the products from catalytic hydrotreatment of methylcyclohexane with AlPNiSMM at 400 °C for 90 min in 500 psi of hydrogen. Labeled peaks are identified in Table 5. Table 1. Synthetic Clay Characterization catalyst
total acidity, mmol/g
surface area, m2/g
NiSMM AlPNiSMM ZrPNiSMM
0.84 1.02 1.14
240 230 274
Table 2. Catalytic Reactions of Methylcyclohexanea catalyst NiSMMb NiSMM AlPNiSMMc AlPNiSMM AlPNiSMM AlPNiSMM AlPNiSMM AlPNiSMM AlPNiSMM ZrPNiSMMd ZrPNiSMM
time, temp, hydrogen, conversion, products, % min °C psig % liquid gas 90 90 30 30 30 90 90 90 180 90 90
400 400 350 400 400 400 400 400 400 400 400
500 1000 1000 1000 200 500 200 1000 1000 500 1000
97 99 29 93 37 90 36 99 100 98 99
72 41 23 72 28 68 25 42 21 68 54
25 58 6 21 9 22 11 57 79 30 45
a Substrate ) methylcyclohexane; catalyst:substrate ratio ) 0.5. Nickel-substituted synthetic mica montmorillonite. c Aluminapillared NiSMM. d Zirconia-pillared NiSMM.
b
the total acidity to 1.02 mmol/g of catalyst. The increase in acidity is due to the additional acidic sites created by the dehydroxylation of the hydroxylated aluminum cations. Similarly, total acidity of zirconia-pillared NiSMM increased to 1.14 mmol/g of catalyst (Table 1). The surface area of the NiSMM was found to be 240 m2/g. Pillaring the clay with polyoxyzirconium cations resulted in an increase in the surface area (274 m2/g). The increase in surface area may be attributed to an increase in the size of micropores created by the intercalated metal cation clusters (pillars). Although the interlayer spacing increases, the packing of the pillars is tight, so that a large increase in surface area is not observed. Alumina-pillared NiSMM had a little lower surface area (230 m2/g) as reported earlier. The repeat distance of the layers in the alumina-pillared NiSMM was previously reported to be 1.73 nm.16 Catalytic Reactions with MeCH. NiSMM synthetic clay and two pillared NiSMM catalysts were used in a series of microreactor tests carried out with MeCH under hydrogen pressures. The results from the catalytic hydrocracking/hydroisomerization reactions of MeCH are given in Tables 2 and 3. The reaction products were a complex mixture of hundreds of components formed as a result of the extensive rearrangements or isomerization of the primary cracked products.
Naphthene Upgrading with Synthetic Clay Catalysts
Energy & Fuels, Vol. 10, No. 3, 1996 589
Table 3. Products from Reactions of Methylcyclohexanea catalyst
hydrogen, psig
BAb/n-Ad
DMCP/BA1
OCAc/BA
i-C4/n-C4
i-C5/n-C5
i-C6/n-C6
i-C7/n-C7
NiSMM ZrPNiSMM AlPNiSMM AlPNiSMM AlPNiSMM
500 500 500 200 1000
3.2 2.7 1.9 4.9 3.3
0.2 0.1 0.6 6.7 0.6
0.2 0.1 0.2 2.3 0.3
2.2 2.1 2.6 2.3 2.5
5.1 4.3 6.6 7.9 6.6
5.7 5.2 7.5 10.6 8.4
3.1 4.2 0.7 12.4 2.6
a Reaction temperature ) 400 °C; reaction time ) 90 min; catalyst:substrate ratio ) 0.5. b Branched alkanes. c Cycloalkanes other than DMCP. d Normal alkanes.
Table 4. Major Components from Catalytic Hydrotreating of Methylcycohexanea
d
nA
BAb
CAc
DMCP
BTXd
propane butane pentane hexane heptane
isobutane isopentane 2,3-dimethylbutane 2-methylpentane 3-methylpentane 2,3-dimethylpentane 2,4-dimethylpentane 3,3-dimethylpentane 2-methylhexane 3-methylhexane
cyclopentane cyclohexane methylcyclopentane ethylcycopentane
1,1-dimethylcycopentane cis-1,2-dimethylcycopentane trans-1,2-dimethylcyclopentane
benzene, toluene
cis-1,3-dimethylcyclopentane trans-1,3-dimethylcyclopentane
a Catalyst:substrate ratio ) 0.5; H ) 500 psi; reaction temperature ) 400 °C; reaction time ) 90 min. b Branched alkanes. c Cycloalkanes. 2 Benzene, toluene, xylene.
Table 5. Yield Data for Hydrotreatment of Methylcyclohexanea retn time, min 5.012 5.287 5.525 6.403 6.885 7.687 8.756 8.815 8.964 9.538 10.337 11.815 12.074 13.470 14.102 14.905 15.025 15.234 15.671 16.216 16.480 16.745 16.935 18.162 20.276 21.730 24.580 a
compd propane isobutane butane isopentane pentane solvent cyclopentane 2,3-dimethylbutane 2-methylpentane 3-methylpentane hexane methylcyclopentane 2,4-dimethylpentane benzene cyclohexane 2-methylhexane 2,3-dimethylpentane 1,1-dimethylcyclopentane 3-methylhexane cis-1,3- dimethylcyclopentane trans-1,3-dimethylcyclopentane trans-1,2- dimethylcyclopentane 2,2,4-trimethylpentane (standard) heptane methylcyclohexane (substrate) ethylcyclopentane toluene
yield, wt % 1.7 8.8 4.2 9.7 2.3 0.5 0.8 3.0 2.2 1.2 1.9 0.1 0.2 0.3 0.6 0.3 0.6 0.8 0.7 0.7 1.1 0.4 1.5 0.6 0.9
wt % based on starting weight of MeCH.
The major products from higher-temperature reactions were the desired isoparaffins (mono-, di-, and trimethylalkanes) with lower amounts of n-alkanes. The amount of aromatic hydrocarbons was insignificant. Major products from the reaction are listed in Table 4. Complete quantitative yield data from the reaction of methylcyclohexane with alumina-pillared NiSMM at 400 °C for 90 min in 500 psi of H2 pressure are given in Table 5. No coke or insoluble materials were detected on the recovered catalyst. Temperature Effects. A reaction of MeCH was carried out with alumina-pillared NiSMM at 350 °C for 30 min in the presence of 1000 psi of initial hydrogen pressure. This reaction gave only 29% conversion of MeCH to products. The yields of liquid and gas prod-
ucts were 23% and 6%, respectively. A reaction at 400 °C gave considerably higher conversion (93%) of MeCH to liquid (72%) and gas (21%) products. This experiment established that 400 °C is a more appropriate temperature, and subsequent reactions were performed at 400 °C. Pressure Effects. Reactions of MeCH were conducted under different initial hydrogen pressures to determine the effect on product yields and distribution. The reaction of MeCH with NiSMM at 400 °C for 90 min in the presence of 1000 psi of initial hydrogen pressure gave almost complete conversion of the substrate to liquid (41%) and gas (58%) products. The reaction at 500 psi resulted in significant improvement in the yield of liquid products (72%), and gas yield was significantly lower (25%), owing to lower hydrocracking of MeCH and rearrangement products at the lower pressure. The pillared clay catalysts exhibited the same effect. The reaction of MeCH with alumina-pillared NiSMM (AlPNiSMM) at 400 °C for 90 min in the presence of 200 psi of hydrogen (initial) gas gave only 36% conversion of the substrate to liquid (25%) and gas (11%) products. Increasing the hydrogen pressure to 500 psi resulted in a substantial increase in the conversion of MeCH (90%) to give 68% yield liquid products and 22% gas products. The reaction with 1000 psi of initial hydrogen pressure gave significantly high conversion of MeCH (99%), but the yield of liquid products was significantly lower (42%) with the remaining gas. The reaction of ZrPNiSMM with MeCH at 400 °C for 90 min in 500 psi (initial) of hydrogen pressure gave 98% conversion to liquid product (68%) and gas product (30%). However, a similar reaction with 1000 psi of hydrogen pressure gave significantly lower amounts of liquid products and high gas yield (Table 2). The large pressure effect on conversion of MeCH and also on gas yield is attributed to a simple hydrogen concentration effect. Reaction Time. The effect of reaction time on the product yields and product distribution was investigated with AlPNiSMM for two hydrogen pressure conditions.
590
Energy & Fuels, Vol. 10, No. 3, 1996
Olson and Sharma
Table 6. Product Distribution from Catalytic Hydrotreating of Methylcyclohexanea catalyst
nA
BA
CA
DMCP
BTX
NiSMM A1PNiSMM ZrPNiSMM
16.5 18.5 20.1
53.2 35.2 55.1
9.2 6.1 5.3
9.5 22.1 6.4
2 2 2.1
a Catalyst:substrate ratio ) 0.5; H ) 500 psi; reaction tem2 perature ) 400 °C; reaction time ) 90min.
In an experiment with an initial hydrogen pressure of 1000 psi, a 30-min reaction gave 93% conversion of MeCH to products; the yields of liquid and gas products were 72% and 21%, respectively. At the same pressure, the reaction for 90 min resulted in almost complete conversion of MeCH, to give a liquid product yield of 42%. The remaining substrate was converted into gaseous products. After 180 min, MeCH was completely converted into liquid (21%) and gaseous (79%) products. Thus, a shorter reaction time is required for liquid isoparaffinic products. The effect of time was also studied by comparing the reactions of MeCH with AlPNiSMM for 30 and 90 min at 400 °C in the presence of 200 psi of initial hydrogen pressure. These reactions gave similar conversions of MeCH to volatile products (Table 2). The liquid product yield was somewhat higher for the 30-min reaction than for the 90-min reaction. Thus, even at the lower pressure, primary cracked products are further cracked to form undesirable gaseous components over a longer period of time. Selectivity. A comparison of the conversions obtained with the three catalysts at 500 and at 1000 psi at 400 °C (Table 2) indicates that there is no difference in activity of the catalysts. At 1000 psi, the ZrPNiSMM gave a significantly better liquid yield (54%) than the NiSMM (41%) and AlPNiSMM (42%). But at the moderate pressure (500 psi), there is no difference in the selectivity for total liquid product. As described above, 500 psi is the preferred pressure to perform the reaction. Further information on the selectivity for hydrocracking and hydroisomerization reaction of the MeCH can be gained from comparison of the cyclic and isoparaffinic components in the products (Tables 3 and 6). The reaction of MeCH with AlPNiSMM in 200 psi of hydrogen gave large amounts of dimethylcyclopentanes (DMCP), indicating that hydrocracking is incomplete at lower hydrogen concentrations. The formation of four DMCP isomers (1,1-DMCP, trans-1,2-DMCP, cis-1,3DMCP, and trans-1,3-DMCP) indicates that carbonium ions form initially on all carbons of the MeCH ring and rapidly rearrange via protonated cyclopropane (actually bicyclohexane) intermediates to the isomeric DMCPs. The products obtained at 500 psi with the three catalysts are mainly the isoparaffins. Thus, at higher pressures, cracking of the DMCP rearrangement products as well as the MeCH via carbonium ion intermediates increases, giving branched C7 isomers and subsequently other isoparaffins. The amounts of aromatic hydrocarbons were found to be minimal with all of the catalysts, so dehydrogenation is not a major reaction of the MeCH. The ZrPNiSMM gave the lowest ratio of DMCP to branched alkanes and therefore appears to be the most efficient catalyst for the ring hydrocracking reaction. This result is consistent with the higher acidity of this
pillared catalyst (Table 1). Previous work demonstrated that there is no difference in the three catalysts with respect to shape selectivity for linear substrates,9 but shape selectivity with respect to reaction rates for branched alkanes was not explored. It is more likely that the effects seen here result from acidity and small differences in the reaction pathways. The ZrPNiSMM gave low amounts of the n-alkanes and other cycloalkanes, but slightly more of these than did the other catalysts. This could mean that the reaction has progressed further toward an equilibrium concentration than the slightly less active catalysts. But, a higher i-C7 to n-C7 ratio was obtained, compared with the other catalysts. Thus, the preference for isomerization of (2methylcyclohexyl)carbonium ion to DMCP carbonium ion intermediates in the first step of the reaction of MeCH rather than isomerization to (2-ethylcyclopentyl)carbonium ion and then to n-C7 intermediates is higher for the ZrPNiSMM. The AlPNiSMM gives a quite different distribution of products from those of the other two catalysts. Since the n-heptane yield is high, the initial carbonium ion reaction may favor the rearrangement of (2-methylcyclohexyl)carbonium ion to (2-ethylcyclopentyl)carbonium ion, which cracks to n-heptane rather than branched heptanes. Ethylcyclopentane is one of the larger components in the products from this reaction. The lowered content of branched alkanes increased the DMCP to branched alkanes (BA) ratio. Thus the acidic AlPNiSMM is also efficient at hydrocracking, but different initial rearrangements lead to different products. Conclusions Methylcyclohexane was effectively isomerized and hydrocracked with the synthetic clay and pillared synthetic clay catalysts to give mainly isoparaffinic components. Dehydrogenation of the MeCH was not a major pathway, as indicated by the minimal formation of aromatic hydrocarbons (
