206
Energy & Fuels 1990,4, 206-214
perature. Without catalyst, the hydrogen addition from the gas phase to the coal was very small and liquefaction mainly proceeded with the hydrogen donation from the solvent while the hydrogen exchange between the gas phase and the coal proceeded rapidly even at the initial stage of the reaction. Although the Ni-Mo-A1203 catalyst
was not effective for the conversion of coal, it promoted the cracking of BIS-THFS to HS as shown in Figure 4. This catalyst also promoted the hydrogen-exchange reaction between the gas phase and the coal. Registry No. Ni, 7440-02-0;Mo, 7439-98-7;tetralin, 119-64-2.
Reaction Pathways of Model Coprocessing Systems Using Molybdenum Naphthenate and Excess Sulfur Honggon Kim and Christine W. Curtis* Chemical Engineering Department, Auburn University, Auburn, Alabama 36849 Received July 11, 1989. Revised Manuscript Received January 2, 1990
The activity and selectivity of catalysts generated in situ from molybdenum naphthenate and excess sulfur for reactions occurring in coal-oil reactions have been examined. The reaction pathways for species representative of those present in coal and petroleum residuum are presented. The MoSz catalyst generated in situ promotes partial saturation of the multiring aromatic to hydroaromatic species but does not promote further saturation of the hydroaromatic or of single-ringalkyl-substituted aromatics. Heteroatom removal with this in situ generated MoS2 catalyst follows two simultaneous reaction pathways: one for heteroatom removal prior to saturation of the aromatic rings producing alkyl-substituted single-ring aromatics and the other for heteroatom removal after saturation of the aromatic rings producing primarily alkyl-substituted cyclic hydrocarbons.
Introduction The coprocessing of coal with petroleum residuum has been shown to be a feasible process for simultaneously converting coal and upgrading coal liquids and petroleum residuum into higher quality Previous research has shown that molybdenum naphthenate (MoNaph) is an effective catalyst for c o p r o c e ~ s i n g . ~Mo~ lybdenum naphthenate has also been shown to be effective for upgrading end products from coal 1iquids'O and petroleum crudes and Molybdenum naphthenate is a metal salt of a mixture of organic acids and serves as a catalyst precursor. The active catalytic species, M0S2,I4is formed after thermal decomposition of the molybdenum naphthenate and reaction with sulfur. The activity and selectivity of the catalyst generated from molybdenum naphthenate and excess sulfur for specific types of reactions which occur in coprocessing have only been surveyed.6 The purpose of this work is to provide an in-depth study of the reaction pathways occurring for model hydrocarbon and heteroatomic species that are representative of some functionalities present in coal and residuum in the presence of molybdenum naphthenate and excess sulfur. The model systems selected were naphthalene representing aromatics in coal; indan and indene representing partially saturated cyclic hydrocarbons present in coal and petroleum residuum; benzothiophene, an organic sulfur compound representing sulfur species present in both coal and petroleum residuum; quinoline and indole, organic nitrogen compounds representing those present in coal and petroleum residuum; o-cresol, representing phenolic compounds present in coal; and benzofuran, an etheric oxygen com* Author
t o whom correspondence should be addressed.
0887-0624/90/2504-0206$02.50/0
pound representing heterocyclic oxygen compounds present in both coal and petroleum residuum. Two salient findings resulted from this work. First, MoNaph with excess sulfur showed selectivity for producing hydroaromatics as the primary products from the hydrogenation of aromatics not containing heteroatomic species. The increased amount of hydroaromatics in the reaction system with MoNaph and excess sulfur may promote coal conversion resulting in the increased observed activity of MoNaph in coprocessing systems? Second, the degree of ring saturation of products from the hydrogenation of different heteroatomic species was dependent upon the heteroatomic species present. The presence of nitrogen appeared to activate the adjacent aromatic ring compared to either oxygen or sulfur. The types and quantities of products obtained from hydrogenation reactions of aromatic and heteroatomic compounds are presented in the (1) Moschopedis, S. E.; Hawkins, R. W.; Fryer, J. F.; Speight, J. G. Fuel 1980,59, 647-653. (2) Monnier, J. Review of the Coprocessing of Coals and Heavy Oils of Petroleum Origin; CANMET Report 84-53, March 1984. (3) Mochida, I.; Iwamoto, K.; Tahara, T.; Korai, Y.; Fujitau, H.; Takeshita, K. Fuel 1982, 61,603-609. (4) Yan, T. Y.; Espenscheid, W. F. Fuel Process. Technol. 1983, 7 , 121-133. (5) Audeh, C. A. U.S. Patent 4,390,409, 1983. (6) Pellegrino,J. L. Master's Thesis, Auburn University, Auburn, AL, 1987. (7) Curtis, C. W.; Tsai, K. J.; Guin, J. A. Ind. Eng. Chem. Res. 1987, 26, 12-18. (8) Curtis, C. W.; Cassell, F. N. Energy Fuels 1988,2, 1. (9) Aldridge, C. L.; Bearden Jr., R. U.S. Patent 4,298,454, 1981. (10) Tsai, K. J. Ph.D. Dissertation, Auburn University, Auburn, AL, 1985. (11) Bearden Jr., R.; Aldridge, C. L. U S . Patent 4,134,825, 1979. (12) Herbstman, S. U S . Patent 4,125,455, 1978. (13) Gleim, W. K. T.;Lake, I.; Gatsis,J. G. U.S. Patent 3,165,463,1965. (14) Kim, H.; Curtis, C. W.; Cronauer, D. C.;Sajkowski, D. J. P r e p . Pap.-Am. Chem. SOC.,Diu.Fuel Chem. 1989, 34, 1431-1438.
0 1990 American Chemical Society
Energy & Fuels, Vol. 4, No. 2, 1990 207
Reaction Pathways of Model Coprocessing Systems
individual pathways given in this paper.
Experimental Section The model reactants, naphthalene (99%),indan (97%), indene (99%), benzothiophene (97%), quinoline (99%), indole (99%), o-cresol(99%), and benzofuran (99.5%), were selected as representative species of some functionalities in both coal and petroleum residuum. n-Hexadecane (99%) was used as the solvent. All of the model reactants, their hydrogenation products, and the solvent were obtained from Aldrich Chemical Company. Two molybdenum-containingoil-soluble catalysts,molybdenum naphthenate (MoNaph) and molybdenum octoate (MoOct), were obtained from Shepherd Chemical Company. The Mo content of the original catalyst precursors was 6 wt % Mo in MoNaph and 8 wt % Mo in MoOct. Shell 324 NiMo/A1203powder used in some reactions was presulfided prior to use. NiMo/A1203 extrudates (1/16 in.) were presulfided in a tube furnace under a flow of 10% H g in H2 at 20 mL/min for 5 h. The presulfiding temperature was raised from 204 to 371 "C through several step increases during the heating period.1° The presulfided catalyst was allowed to cool in an inert atmosphere, ground to -150-mesh powder and then stored in a vacuum desiccator. Conditions and Procedures. Hydrogenation reactions were conducted in 20-cm3stainless steel batch microtubing bomb reactors which were oriented horizontally. Hydrogen was introduced to the reactor at 1250 psig at ambient temperature prior to the reaction. The reactions were performed at 380 "C for 30 min at an agitation rate of 550 cpm. The pressure of the system at reaction temperature was estimated to be ca. 2700 psig. Four grams of reactant solution was charged for each reaction and included 2 wt % naphthalene or 1wt % of the other reactants in the solvent hexadecane. Reactions were performed both thermally and catalytically and were at least duplicated. The results are presented as average values with standard deviations in the form of X f ern. For the catalytic reactions, MoNaph or MoOct was premixed a t a level between 2850 or 2950 ppm Mo in the reactant solution, always keeping a constant ratio of Mo to reactant, and -0.024 g of elemental sulfur was added to generate the active catalytic species. The atomic ratio of S to Mo in the reactor was 6, providing 3 times the stoichiometric amount of sulfur required to produce MoS2. T w o types of thermal reactions were performed the first with no excess sulfur and the second with -0.016 g of elemental sulfur which was the approximate amount of excess sulfur that would be present in the catalytic reaction after reaction with MoNaph. The thermal reaction with sulfur was performed to evaluate the effect of excess sulfur on the reaction pathway. The liquid products were analyzed by gas chromatography using a 30-m fused silica capillary DB-5column of 0.32-mm inner diameter from J&W Scientific and FID detection. p-Xylene was used as the internal standard. The products were identified by using a VG 7OEHF mass spectrometer coupled with a Varian 3700 gas chromatograph. Model Reactants and Catalyst.
Results and Discussion Generation of the Active Catalytic Species from Different Molybdenum Precursors. Two different molybdenum precursors, MoNaph and MoOct, were examined in this work. The active catalytic species generated in situ from the oil-soluble molybdenum salts of organic acids in the presence of sulfur is believed to be a molybdenum sulfide whose stoichiometry is close to M o S ~ The MoS2 produced was a black powderlike solid with a particle size range of 500-2500 A in diameter;I4 this small particle size eliminated the limitation of pore diffusion which has been shown to affect adversely the activity of catalyst extrudates in copro~essing.~"~' (15)Curtis, C. W.; Guin, J. A.; Kamajian, B. L.; Moody, T. E. Fuel Process. Technol. 1986,12,111-125. (16)Ho,P. N.;Weller, S. W. Fuel Process. Technol. 1981,4,21-29. (17)Curtis, C. W.; Tsai, K. J.; Guin, J. A. Ind. Eng. Chem. Process Des. Deu. 1986,24,1259-1266.
Table I. Comparison of the Catalytic Activity of Different Mo Precursors reaction svstem" MoNaDh MoOct individual naphthalene reaction % HYD of naphthalene 42.7 f 0.3 42.0 combined naphthalene/benzothiophene reaction % HYD of naphthalene 43.7 f 0.2 43.9 % HYD of benzothiophene 41.1 f 1.1 39.5 % HDS of benzothiophene 100.0 f 0.0 100.0 a
% HYD (hydrogenation) and % HDS (hydrodesulfurization)
are expressed as the averaged values and standard deviations of X f u,.
- Ln2 NAPH
T
cis- & t r a n s - D
CYC4"' BB
Figure 1. Reaction pathway for naphthalene hydrogenation catalyzed by molybdenum naphthenate (broken arrow: possible reaction pathway).
In order to compare the activity and selectivity of the two molybdenum oil-soluble catalysts, catalytic reactions were performed with MoNaph and MoOct using the model systems, naphthalene and naphthalene with benzothiophene. A comparison of the activities of the two catalysts at 380 "C with 3000 ppm Mo and excess sulfur is given in Table I. Two terms are used to define the degree of reactivity of the species: percent hydrogenation and percent hydrodesulfurization. Percent hydrogenation is defined as the number of moles of hydrogen required to achieve the final liquid product distribution as a percentage of moles of hydrogen required to achieve the most hydrogenated product; the most hydrogenated liquid products defined were decalin for naphthalene and cyclohexane for benzothiophene. Percent hydrodesulfurization is the summation of the mole percents of products not containing sulfur. At this reaction condition, the two different precursors exhibited nearly the same activity as indicated by equivalent amounts of hydrogenation and hydrodesulfurization for both naphthalene and benzothiophene. MoNaph was chosen as the catalyst for all subsequent catalytic reactions. Reaction Pathways. The reaction pathways observed for catalytic reactions using MoNaph and excess sulfur have been examined on the basis of the analysis of liquid products recovered. The products produced and the pathways described were compared to those that appear in the literature for supported Mo catalyst6 and bimetallic hydrotreating catalysts like NiMo/Alz03, CoMo/Alz03, and NiW/Alz03.1924 The reaction pathways obtained here ~ J ~were similar to those with presulfided Mo/A1203but dif(18)Hudlicky, M. Reductions in Organic Chemistry; Wiley: New York, 1984. (19)Steigel, G. J.; Shah, Y. T.; Krishnamurthy, S.; Pauvelker, S. V. Reaction Engineering in Direct Coal Liquefaction; Shah, Y. T., Ed.; Addison-Wesley: London, 1981;Chapter 6. (20)Badilla-Ohlbaum, R.;Pratt, K. C.; Trimm, D. L. Fuel 1979,58, 309-314. (21)Stern, E. W. J. Catal. 1979,57,390-396. (22)Rollmann, L. D.J. Catal. 1977,46,243-252. Ollis, D. F. J. Catal. 1983,BO, 76-89. (23)Odebunmi, E. 0.; (24)Schulz, H.; Schon, M.; *an, N. M. Catalytic Hydrogenation; Cervery, L., Ed.; Elsevier Science: New York, 1986; Chapter 6.
Kim and Curtis
208 Energy &Fuels, Vol. 4, No. 2, 1990 Table 11. Comparison of Naphthalene Hydrogenation Using MoNaph with Sulfur and NiMo/Al,O$ presulfided NiMo/ products, thermal MoNaph A1203' mol 70 thermal with S with Sb naphthalene 0.0 f 0.0 99.1 90.2 f 2.8 2.0 f 0.4 tetralin 92.5 f 0.5 9.5 f 0.6 0.9 9.8 f 2.8 decalin 0.0 0.0 f 0.0 5.5 f 0.7 86.6 f 0.3 2.2 f 0.1 n-butylbenzene 0.0 0.0 f 0.0 trace 1.7 f 0.1 othersd 0.0 0.0 f 0.0 0.0 f 0.0 2.9 f 0.0 t-D/c-D' NAf NAf 1.4 f 0.1 "he product distribution is expressed as mole percentage using the averaged values and standard deviations of X f 6,. Mo = ca. 3000 ppm. 150 mesh powder catalyst of Shell 324 NiMo/Al,O,; Ni and Mo = ca. 3000 ppm. Other products of molecular weight 138. e Weight ratio of trans-decalin to cis-decalin. 'Not available.
*
ferent from those with the above bimetallic catalysts. Mo/A1203 showed a selectivity for producing partially saturated aromatics and nearly no activity for saturating alkylbenzenes.6 By contrast, the bimetallic supported catalysts containing Ni or Co showed a selectivity for saturating single-ring alkylaromatics and aromatics to alicyclics but showed a relatively low selectivity for producing hydroaromatics which are known as excellent solvents for coal l i q u e f a c t i ~ n . ~ ~ , ~ ~ Naphthalene Hydrogenation. The reaction pathway for naphthalene hydrogenation catalyzed by MoNaph and excess sulfur is given in Figure 1. Under thermal hydrogenation conditions, less than 1%of the naphthalene was converted to tetralin as shown in Table 11. With the addition of excess sulfur, 10% conversion of naphthalene was observed. The catalytic reaction produced nearly complete naphthalene conversion and yielded tetralin as the primary product at 92% and small amounts of cis- and trans-decalin. Both tetralin and decalin were used as reactants at 380 O C in the presence of MoNaph. Only -5% of tetralin was hydrogenated to decalin, but the reverse dehydrogenation to naphthalene was not detected at this temperature. No dehydrogenation of decalin was observed. As the temperature was increased from 380 to 425 "C or the time was increased from 30 min to 1 h, a trace amount of tetralin was hydrogenolyzed to n-butylbenzene. At 425 "C,naphthalene reacted to form 12% decalin and 5% n-butylbenzene; however, slightly more naphthalene, -4%, remained unconverted,indicating that dehydrogenation of tetralin to naphthalene possibly occurred at the higher temperature. This dehydrogenation of tetralin to naphthalene at higher temperatures than 400 O C was also found with sulfided NiW/A1,0, catalyst and explained by the effect of thermodynamic equilibrium on aromatic conversion: the chemical equilibrium between aromatics and naphthenes shifts in favor of aromatics with increasing temperature.,' A comparison of the products achieved with MoNaph and excess sulfur and presulfided NiMo/Al,O, is given in Table 11. Under equivalent conditions including the same metal content, MoNaph and excess sulfur promoted the hydrogenation of naphthalene to the hydroaromatic, tetralin, while the introduction of NiMo/Al,03 resulted in completely saturating the aromatic rings and producing decalin as the primary product. Indan and Indene Hydrogenation. The reaction pathways for indan and indene hydrogenation are given (25) Chiba, K.; Tagaya, H.; Saito, N. Energy Fuels 1987,1, 338. (26) Curtis, C. W.; Guin,J. A.; Hale, M. A.; Smith, N. Fuel 1985, 64, 461-469. (27) Wilson, M.F.; Fisher, I. P.; Kriz,J. F.J.Catal. 1985,95,155-166.
INE
-*I
"2
mLaJ IN
c i s - & trans-HHIN
H21
".1
1 qC3"' ] [ CJHC3"' ] "11 " 2 1 -'2"6
-'2%
UCH3 0"' TOL
MCH
Figure 2. Reaction pathway for indene and indan hydrogenation catalyzed by molybdenum naphthenate.
in Figure 2. Indene readily hydrogenated to indan under both thermal and catalytic conditions as shown in Table 111. Indan itself was stable thermally, but catalytically 10% of indan was converted to cis- and trans-hexahydroindan and in trace amounts to toluene and methylcyclohexane. The cis isomer of hexahydroindan was predominant. No propylbenzene was detected in the products; however, trace amount of toluene observed suggested that a reaction pathway passing through propylbenzene was plausible. Benzothiophene Hydrodesulfurization. Benzothiophene was thermally much more stable than indene; less than 8 and 26% of benzothiophene was converted thermally and thermally with sulfur, respectively (Table IV). With sulfur present, more hydrodesulfurized product, ethylbenzene, was thermally produced than without sulfur. In the catalytic reaction with MoNaph and excess sulfur, complete desulfurization occurred, yielding ethylbenzene as the major product at 90% and ethylcyclohexaneas the minor product at 6% (Table IV). In order to evaluate the reaction pathway leading to the production of ethylcyclohexane as shown in Figure 3, ethylbenzene was hydrogenated under the same conditions. Less than 2% ethylbenzene was hydrogenated to ethylcyclohexane and methylcyclohexane (Table V). However, 6% ethylcyclohexane was produced from the hydrogenation of benzothiophene. The additional ethylcyclohexane produced from benzothiophene must be obtained through another reactive intermediate such as octahydrobenzothiophene or 2-ethylcyclohexanethiol,neither of which was found in the product analysis. Hydrogenation of the observed intermediate, o-ethylthiophenol, produced slightly less ethylcyclohexane than benzothiophene hydrogenation, suggesting a possible alternative reaction pathway from benzothiophenethrough octahydrobenzothiophene. Thus, in order to account for the production and quantity of ethylcyclohexane, two simultaneous reaction pathways are proposed for benzothiophene hydrodesulfurization. The primary pathway of reaction was through hydrodesulfurization of benzothiophene by saturation of the five-membered ring, followed by hydrogenolysis and desulfurization, thereby forming the major product, ethyl-
Reaction Pathways of Model Coprocessing Systems
Energy & Fuels, Vol. 4, No. 2, 1990 209
Table 111. Thermal and Catalytic Hydrogenation of Indan and Indene Using MoNaph and S u l f u P
products, mol % indene indan cis-hexahydroindan trans-hexahydroindan toluene methylcyclohexane
indan thermal with S b MoNaph with Sc 0.0 1.4 f 0.0 100.0 88.4 f 1.1 0.0 5.8 0.7 0.0 2.4 f 0.0 0.0 0.7 f 0.7 0.0 1.3 f 0.3
thermal 0.0
100.0
*
0.0 0.0 0.0 0.0
indene thermal with S b MoNaph with Sc 10.4 f 7.3 0.0 f 0.0 89.6 f 7.3 86.9 f 0.6 0.0 f 0.0 5.8 f 0.4 0.0 f 0.0 3.0 f 0.6 0.0 f 0.0 2.8 f 0.4 0.0 f 0.0 1.5 f 0.0
thermal 48.5 51.5 0.0 0.0 0.0 0.0
OThe product distribution is expressed as mole percentage using the averaged values and standard deviations of X f a,. b S = 0.016 g. 'Mo = 2850-2950 ppm; Mo/indan = 0.30; S = 0.024 g. Table IV. Thermal and Catalytic Hydrodesulfurization of Benzothiouhene and Ethylthiophenol with MoNauh and S u l f u P
products, mol % benzothiophene dihydrobenzothiophene o-ethylthiophenol ethylbenzene ethylcyclohexane othersb
thermal without S 92.4 f 0.2 4.1 f 0.7 0.0 f 0.0 3.5 f 1.0 0.0 f 0.0 0.0 f 0.0
benzothiophene thermal with S 74.1 f 3.9 6.9 f 1.9
o-ethylthiophenol MoNaph with S
MoNaph with S 0.0 f 0.0 0.0 f 0.0
0.0 f 0.0
0.0 f 0.0
19.0 f 3.3 0.0 f 0.0 0.0 f 0.0
90.4 f 2.0 5.9 f 1.9 3.7 f 2.0
0.0 f 0.0 95.3 f 0.2 3.0 f 0.1 1.7 f 0.1
'The product distribution is expressed as mole percentage using the averaged values and standard deviations of X f a,. hexane and toluene.
Methylcyclo-
Table V. Comparison of Propylbenzene, Ethylbenzene, and Toluene Hydrogenation Using MoNaph and Sulfur and NiMo/AllOf BZT
catalyst (metal content, ppm) presulfided NiMo/A1203 MoNaph with (3000 Ni and products, mol % S (3000 Mo) Mo) Propylbenzene Hydrogenation propylbenzene 99.5 f 0.6 19.8 f 3.9 propylcyclohexane 0.5 f 0.6 78.0 f 3.6 0.0 f 0.0 1.5 f 0.3 butylcyclopentane methylcyclohexane 0.0 f 0.0 0.7 f 0.1
H21
m DHBZT
".1
Ethylbenzene Hydrogenation 17.9 f 1.7 ethylbenzene 96.9 f 0.5 ethylcyclohexane 1.9 f 0.6 81.8 f 1.5 0.0 f 0.0 toluene 0.9 f 0.1 methylcyclohexane 0.3 f 0.1 0.3 f 0.3 Toluene Hydrogenation toluene 98.0 f 0.2 methylcyclohexane 1.8 f 0.0 ethylcyclopentane 0.2 f 0.2
3H2
H2C J
OETP
20.3 f 0.1 78.6 f 0.1
The product distribution is expressed as mole percentage using the averaged values and standard deviations of X f u,,.
1
ECHT
CY2"'
CYCZH'
1.1 f 0.1
benzene. The minor reaction pathway was through saturation of the aromatic ring prior to desulfurization,yielding the minor product, ethylcyclohexane. o -Cresol Hydrodeoxygenation. o-Cresol was thermally stable at 380 "C but reacted catalytically with MoNaph and excess sulfur. Toluene was the major product and was produced in more than 55% yield as shown in
I
ECH
EB
"p.
H21-CH,
U C H 3
(ICH3 HCH
TOL F i g u r e 3.
Reaction pathway for benzothiophene hydrodesulfurization catalyzed by molybdenum naphthenate.
Table VI. Thermal and Catalytic Hydrodeoxygenation of o-Cresol and Its Products'
products, mol % o-cresol 2-methylcyclohexanol unknownb to1uene ethylcyclopentane methylcyclohexane
thermal 100.0 0.0 0.0 0.0 0.0 0.0
o-cresol MoNaphwith S thermal with S (3000 ppm Mo) 94.4 13.3 f 0.6 0.0 0.0
2.4 0.0
3.2
0.0 0.0 0.0 f 0.0 55.2 f 1.0 6.8 f 0.3 24.7 f 0.1
_
_
MoNaph with S (600 ppm Mo) 76.7 f 2.2 0.0 f 0.0 2.0 f 0.2 14.8 f 1.6 2.3 f 0.3
4.2
f
0.2
~
2-methylcyclohexanol MoNaph with S (3000 PPm Mo) 0.0 f 0.0 0.0 f 0.0 0.0 f 0.0
4.7 f 0.2 16.9 f 0.4 78.4 f 0.5
'The product distribution is expressed as mole percentage using the averaged values and standard deviations of X dimethylcyclohexane as a possible structure.
f u,,.
bUnknown:
Kim and Curtis
210 Energy & Fuels, Vol. 4, No. 2, 1990 Table VII. Thermal and Catalytic Hydrodeoxygenation of Benzofuran and Its Products"
benzofuran products, mol % benzofuran dihydrobenzofuran o-ethylphenol 2-ethylcyclohexanol
ethylbenzene
ethylcyclohexenes ethylcyclohexane othersb
thermal thermal with S 96.5 1.6 1.9 0.0 0.0 0.0 0.0 0.0
78.3 4.0 14.5 0.0 1.1 0.0 2.2 0.0
MoNaph with S (3000ppm) 0.0 f 0.0 0.0 f 0.0 20.7 f 2.1 0.0 f 0.0 48.5 f 1.4 0.2 f 0.3 29.1 f 0.3 1.5 f 0.6
MoNaph with S (600ppm) 0.9 f 0.6 2.3 f 0.5 74.3 1.9 0.0 f 0.0 14.0 f 1.3 1.3 f 0.1 5.9 f 0.3 1.4 f 1.0
*
ethylphenol MoNaph with S (3000ppm Mol
2-ethylcyclohexanol MoNaph with S (3000ppm Mo)
26.2 f 0.2 0.0 f 0.0 43.4 f 0.1 0.7 f 0.0 28.3 f 0.2 1.4 f 0.1
0.0 f 0.0 7.1 f 0.7 2.0 f 0.1 88.7 f 0.7 2.2 f 0.1
The product distribution is expressed as mole percentage using the averaged values and standard deviations of X f u,. hexane, methylcyclohexenes, and toluene.
Methylcyclo-
BZP
H21 ......... 0% ....... 2% 4.. - 2H,
TOL
Hzl
OCH3 + GCH3 CXJ ".1 *I DHBZF
MCHE ' S H2
OCH3 O/C2H5 +
MCH
ECP
Figure 4. Reaction pathway for o-cresol hydrodeoxygenation catalyzed by molybdenum naphthenate (broken arrow: possible reaction pathway).
Table VI. Substantial amounts of methylcyclohexane, -24%, and ethylcyclopentane, -7%, were also produced. The reaction pathway for o-cresol using MoNaph and excess sulfur is proposed in Figure 4. The production of both alkyl-substituted alicyclics, ethylcyclopentane and methylcyclohexane, and alkyl-substituted aromatics, toluene, was observed. However, as has been shown in Table V, the direct hydrogenation of toluene using MoNaph produced less than 2% methylcyclohexane and no ethylcyclopentane. In order to obtain nearly 30% methylcyclohexane and ethylcyclopentane from 0-cresol, an alternative route through a reactive intermediate that allows for the formation of alkyl-substituted saturated cyclic compounds must be considered. cis- and trans-methylcyclohexanol is a plausible intermediate although it was not observed as a product. A mixture of cis- and trans2-methylcyclohexanol was hydrogenated with MoNaph and excess sulfur and yielded 78% methylcyclohexane and 17% ethylcyclopentane as the primary and secondary products, respectively. A small amount of toluene (5%) was also produced. Because dehydrogenation of methylcyclohexane did not occur under the conditions of these reactions, this toluene was most likely produced from the dehydrogenation of methylcyclohexene which is proposed as an unstable intermediate from methylcyclohexanol. Thus, two reaction pathways occurring simultaneously for o-cresol hydrodeoxygenation are proposed on the basis of the products observed from o-cresol and its hydrogenated intermediates in the presence of MoNaph and excess sulfur. The two pathways consisted of deoxygenation prior to aromatic ring saturation producing toluene and aromatic ring saturation prior to deoxygenation producing methylcyclohexane and ethylcyclopentane.
Q(;z5
OEP Hz1-H20
OCZH5 EB
H 2 1 -CH,
D C H 3
TOL Figure 5. Reaction pathway for benzofuran hydrcdeoxygenation
catalyzed by molybdenum naphthenate.
Benzofuran Hydrodeoxygenation. The reaction pathway for benzofuran hydrodeoxygenation is given in Figure 5. Benzofuran was thermally stable but was partially hydrodeoxygenated with excess sulfur as shown in Table VII. The thermal products observed were dihydrobenzofuran and o-ethylphenol. Catalytic hydrogenation of benzofuran with MoNaph and excess sulfur resulted in ethylbenzene (48%) as the primary product and ethylcyclohexane (29%) as the secondary product. As shown in Table V, ethylbenzene under these reaction conditions can only be hydrogenated to a small extent to ethylcyclohexane. In order to account for the production of substantial amounts of ethylcyclohexane from benzofuran, the reactive intermediates, octahydrobenzofuran or 2-ethylcyclohexanol, are proposed. When 2-ethylcyclohexanol was hydrogenated, the primary product was ethylcyclohexane a t 89% and the secondary product was ethylbenzene a t 7%. Thus, two simultaneous reaction pathways are proposed for benzofuran hydrodeoxygenation with MoNaph and excess sulfur. The primary pathway removed the oxygen prior to aromatic ring saturation by
Energy & Fuels, Vol. 4, No. 2, 1990 211
Reaction Pathways of Model Coprocessing Systems
Table VIII. Thermal and Catalytic Hydrodenitrogenation of Quinoline and Its Productsn quinoline MoNaph MoNaph tetrahydroquinoline decahydroquinoline o-propylaniline thermal with S with S MoNaph with S MoNaph with S MoNaph with S products, mol % (1500 ppm Mo) (1500 ppm Mo) (1500 ppm Mo) thermal with S (3000 ppm) (1500 ppm) quinoline 8f3 4f3 Of0 0 Of0 Of0 92f3 91f3 Of0 37 32 f 3 3f2 tetrahydroquinoline Of0 Of0 Of0 21 16 f 5 28 f 20 decahydroquinoline Of0 5 f l 10f2 7 7f0 Of0 8 o-propylaniline 010 Of0 5*1 0 2fO 1f2 11 propylbenzene Of0 Of0 10f2 21 24 f 2 41 f 10 8 propylcyclohexenes Of0 Of0 75f3 14 19 f 1 27 f 13 73 propylcyclohexane and butylcyclopentane OThe product distribution is expressed as mole percentage using the averaged values and standard deviations of X f un.
passing from o-ethylphenol to ethylbenzene. The secondary pathway proposed consisted of saturation of the aromatic ring prior to oxygen removal and utilized 2ethylcyclohexanol as a reactive intermediate. Quinoline Hydrodenitrogenation. Quinoline was thermally reactive; the nitrogen-containing aromatic ring was hydrogenated, producing substantial amounts of 1,2,3,4-tetrahydroquinoline(THQ) as shown in Table VIII. Both the hydrogenation and the hydrogenolysis of quinoline were enhanced with the addition of excess sulfur to the thermal reaction. Only a trace amount of 5,6,7,8tetrahydroquinoline (THQ) was detected, but hydrodenitrogenation of quinoline was not achieved in the thermal reaction at 380 "C. The catalytic reaction of quinoline with MoNaph and excess sulfur produced n-propylcyclohexaneas the primary product with a yield of 68%. An appreciable amount, -5%, of n-propylbenzenewas also produced. By contrast, in reactions using presulfided NiMo/A1203as the catalyst, 5,6,7,8-THQ was produced at 350-390 0C28 and decahydroquinoline was a major product at 350 OC.B However, with the 3000 ppm Mo catalyst and reaction conditions used in this study, neither 5,6,7,8-THQ nor decahydroquinoline was detected. To test whether 5,6,7,8-THQ or decahydroquinolinewas a reactive intermediate, quinoline hydrogenation was performed with 1500 ppm Mo. A substantial amount of decahydroquinoline was produced, but 5,6,7,8-THQ was not detected (Table VIII). On the basis of the product slate observed, the reaction pathway for quinoline in the presence of MoNaph and excess sulfur is proposed in Figure 6. The individual intermediates were also hydrogenated at the reaction conditions of this study. The hydrogenation of 1,2,3,4-THQ at 1500 ppm Mo produced decahydroquinoline but not quinoline. The hydrogenation of decahydroquinoline at 1500 ppm Mo produced a small amount of 1,2,3,4-THQ as well as substantial quantities of npropylcyclohexenes, n-propylcyclohexane, and n-butylcyclopentane. The major products from the hydrogenation of o-propylaniline at 1500 ppm Mo were n-propylcyclohexane and n-butylcyclopentane. However, when npropylbenzene and n-propylcyclohexane were used as reactants, neither reactant reacted at 3000 ppm Mo (Table V). Less than 1% n-propylbenzene was hydrogenated to n-propylcyclohexane, and n-propylcyclohexane was not dehydrogenated to any other compounds. Thus, npropylbenzene could not be a major precursor for npropylcyclohexane and its isomers. From these results, 2-propylcyclohexylamineis proposed as a reactive intermediate through which saturated al(28)Satterfield,C.N.;Yang, S. H.I d . Eng. Chem. Process Des. Deu. 1984,23,11-19. (29)Chung, W.J. Master's Thesis,Auburn University, Auburn, AL, 1987.
QN 2H21
sa - 3H2
mQ
DHQ
".1
H21
W"' NH2
OPA
3H2
i
1
PCHA
-WH7 0"'"' a"'"' +
-2H2
PB
PCHE'S
".1
UC3"'+ (\Ic4H9 PCH
BCP
Figure 6. Reaction pathway for quinoline hydrodenitrogenation catalyzed by molybdenum naphthenate.
kyl-substituted cyclic hydrocarbons were produced from both o-propylaniline and decahydroquinoline. Although 2-propylcyclohexylamine was not detected in the reaction products, its use as a reactive intermediate is in accordance with Satterfield's findings in quinoline hydrogenation.28p30 On the basis of the observation of a very small amount ( m 1%)of n-propylbenzene and substantial quantities of n-propylcyclohexenes (-41 9%) from the hydrodenitrogenation reaction of decahydroquinoline, it is proposed that the n-propylbenzenewas produced from the dehydrogenation of n-propylcyclohexenes. When opropylaniline was reacted, it produced m 11% n-propylbenzene. Therefore, it is likely that n-propylbenzene was produced primarily through the direct denitrogenation of o-propylaniline and only to a small degree through the dehydrogenation of 1-and 3-propylcyclohexenes. Thus, the hydrodenitrogenation of quinoline occurred according to two different pathways. The primary path(30) Satterfield, C. N.; Smith, C. M. Ind. Eng. Chem. Process Des. Deu. 1986,25,942-949.,
Kim and Curtis
212 Energy & Fuels, Vol. 4, No. 2, 1990
Table IX. Thermal and Catalytic Hydrodenitrogenation of Indole and Its Products" indole o-ethylaniline MoNaph with S MoNaph with S MoNaph with S Droducts, mol % thermal thermal with S (3000 DDm Mo) (600 Dum Mo) (3000 Dum Mo) indole 29.5 16.7 0.0 f 0.0 16.6 71.7 0.0 f 0.0 indoline 70.5 27.6 0.0 11.6 4.6 f 4.6 o-ethylaniline 0.0 f 0.0 35.0 0.0 0.0 8.3 f 0.6 13.8 f 0.1 ethylbenzene 2.9 0.0 0.8 f 0.5 0.0 f 0.0 0.0 ethylcyclohexene 6.4 0.0 0.0 82.0 f 5.2 ethylcyclohexane 11.5 84.6 f 0.2 0.0 0.0 3.6 f 0.2 0.0 methylcyclohexane trace 0.0 0.0 0.7 f 0.5 othersb 0.0 1.6 f 0.2 'The product distribution is expressed as mole percentage using the averaged values and standard deviations of X f u,. *Cyclohexane and benzene.
way was denitrogenation after both rings had been fully saturated that mainly yielded saturated end products like n-propylcyclohexane and n-butylcyclopentane. The secIND ondary pathway was hydrogenolysis followed by denitrogenation without saturation of the remaining single aromatic ring that produced n-propylbenzene as the end product. Satterfield et al. proposed a similar reaction pathway 3% for quinoline reaction system using N ~ M O / A ~ ~They O~.~*~ suggested that n-propylcyclohexane and its isomers could INDN PHIND be produced only through propylcyclohexenes. However, when NiMo/Alz03was used as catalyst a t the conditions of this study (Table V), 78% n-propylcyclohexane was .21 produced from n-propylbenzene. Stohl also showed a 3H2 hydrogenation pathway of n-propylbenzene to n-propylNHZ cyclohexane using N ~ M O / A ~ ~ O ~At: ~least * under the conditions of this study, the reaction pathway for quinoline OEA ECHA in the presence of NiMo/Al,O, seems to involve the pathway from o-propylaniline to n-propylbenzene to n' 2 1 -NH3 propylcyclohexane. Indole Hydrodenitrogenation. Indole was readily hydrogenated to indoline under thermal conditions as shown in Table IX. Indole was further hydrogenated and EB ECH hydrogenolyzed to o-ethylaniline with the addition of excess sulfur to the thermal reaction, but no nitrogen removal occurred. The reaction pathway for the catalytic hydrodenitrogenation of indole with MoNaph and excess sulfur U C H 3 is given in Figure 7 . The primary product was ethylcyclohexane a t 79%, and the secondary product was MCH TOL ethylbenzene at 8%. At a lesser catalyst loading of 600 Figure 7. Reaction pathway for indole hydrodenitrogenation ppm Mo, indoline was observed in the catalytic products. catalyzed by molybdenum naphthenate. As has been mentioned previously and shown in Table V, ethylbenzene cannot be readily hydrogenated to ethylprimarily producing alkylcyclohexane. cyclohexane under these catalytic reaction conditions. Product Distribution in Heteroatom Removal. The Therefore, 2-ethylcyclohexylamine is proposed as a likely product distributions obtained from different heteroatomic reactive intermediate between o-ethylaniline and the alspecies showed a different ratio of ethylbenzene to kyl-substituted denitrogenated cyclic products. Ethylethylcyclohexane for a catalyst loading of 3000 ppm Mo. cyclohexane could then be obtained from the denitrogenFor benzothiophene, a 90 to 6 ratio of ethylbenzene to ation of 2-ethylcyclohexylamine to ethylcyclohexene folethylcyclohexane was obtained; for benzofuran, a 49 to 29 lowed by further hydrogenation. When o-ethylaniline was ratio; and for indole and o-ethylaniline, an 8 to 82 ratio reacted, ethylcyclohexane was produced as the major and a 14 to 85 ratio, respectively. A low ratio of propylproduct (84%) and ethylbenzene as the secondary product benzene to propylcyclohexane and butylcyclopentane, 5 (14%). The products were observed in similar proportions to 75, was also obtained from quinoline hydrogenation. as in indole hydrogenation. These ratios indicated that sulfur removal occurred priTwo reaction pathways were observed with indole as was marily without saturation of the adjacent aromatic ring, observed with the other hydrocarbon and heteroatomic nitrogen removal occurred primarily after saturation of the species with the same molecular skeletal structure. The adjacent aromatic ring, and oxygen removal followed both two pathways followed were nitrogen removal from the pathways with oxygen removal prior to saturation of the aromatic ring prior to ring saturation producing alkylaromatic ring being predominant. benzene and ring saturation prior to denitrogenation Some parallelism can be drawn with other hydrotreating catalysts such as N ~ M o / A and~ C~ O~ M~ O ~ /~ A ~ ~ ~ ~ . ~ These hydrotreating catalysts as well as MoNaph with (31) Sbhl, F.V.Prepr. Pap.-Am. Chem. Soc., Diu.Fuel Chem. 1987, 32,325-331. excess sulfur showed the same order of hydrogen usage for
"d
H21
aCzH5 a"'"'OC2"'
OCH3
Reaction Pathways of Model Coprocessing Systems catalytic heteroatom removal: desulfurization < deoxygenation < denitrogenation. Reactivity of the Heteroatomic Species. When the reactivities of the different model hydrocarbon and heteroatomic systems having the same molecular skeletal structure are compared at equivalent conditions, they can be ranked in terms of reactivity for both thermal and catalytic reactions. The four compounds compared are indene, indole, benzothiophene, and benzofuran. Thermal hydrogenations occurring in the five-membered ring first ranked the model systems in the reactivity order of indene > indole > benzothiophene > benzofuran Thermal hydrogenolyses which occurred in the presence of excess sulfur ranked the model systems in the order of benzothiophene benzofuran > indole >> indene
-
The reactivity of the heteroatomic species for catalytic heteroatom removal with MoNaph and excess sulfur was as follows: benzothiophene > indole > benzofuran Indole was the most thermally reactive heteroatomic species in saturating the five-membered ring. However, when excess sulfur was present in the thermal reactions, a change in the reactivity of the dihydro forms was observed. Both dihydrobenzothiophene and dihydrobenzofuran were more reactive than indoline. More carbonheteroatom bond scission occurred with dihydrobenzothiophene and dihydrobenzofuran than with indoline, indicating that the saturated five-membered nitrogen-containing ring was more stable than those containing oxygen and sulfur. The ease of heteroatom removal occurred thermally and thermally with sulfur in the order of desulfurization > deoxygenation > denitrogenation. With MoNaph and excess sulfur, the heteroatom sulfur was readily removed from the hydrogenolyzed five-membered ring. Nitrogen was more easily removed from the hydrogenolyzed species, o-ethylaniline, than oxygen was removed from o-ethylphenol. In fact, the catalytic heteroatom removal from benzothiophene,indole, benzofuran or from o-ethylthiophenol, o-ethylaniline and o-ethylphenol with MoNaph and excess sulfur was ranked according to increasing order of bond strengths of C-S, C-N, and C-0.32 This same dependence for the ease of heteroatom removal on the carbon and heteroatom bond strengths was observed with other hydrotreating catalysts, such as NiMo/Al2O3l0and C O M O / A ~ ~ ~ ~ . ~ ~ I ~ ~ As mentioned earlier, the pathway for nitrogen removal required saturating the aromatic ring while that for sulfur removal did not. Oxygen removal could proceed by either pathway. Hence, once the aromatic ring possessing an adjacent heteroatom became saturated with the aid of catalyst, the order of heteroatom removal followed the ranking of carbon and heteroatom bond strengths irrespective of the types of Mo-containing catalysts.
Summary The pathways for hydrogenation reactions catalyzed by MoNaph and excess sulfur were given for hydrocarbon and for heteroatomic species on the basis of the products produced and the reactions observed for the intermediate products. MoNaph with excess sulfur showed different selectivity in the hydrogenation of aromatic and heteroatomic species compared to that of the commercial hy(32) Furimsky, E.AZChE J. 1979,25, 306-311.
Energy & Fuels, Vol. 4, No. 2, 1990 213 drotreating bimetallic catalysts like NiMo/A1203,NiW/ A1,0,, and CoMo/Alz03. MoNaph promoted partial saturation of the aromatic ring by producing the hydroaromatic tetralin from naphthalene and indan from indene as the primary products. Saturation of alkyl-substituted aromatics, such as toluene, ethylbenzene, or propylbenzene, was not promoted. This selectivity of Mo catalyst from MoNaph and excess sulfur for producing hydroaromatics, which donate hydrogen during coal liquefaction, may be one of the important promoting influences of MoNaph in actual coprocessing systems and partially responsible for its high activity.8 The reaction pathways observed for the heteroatomic species containing oxygen, sulfur, and nitrogen followed two paths: one pathway for heteroatom removal prior to saturation of the aromatic ring producing alkyl aromatics and one for heteroatom removal following saturation of the aromatic ring producing alkyl alicyclics. All of the heteroatomic species followed both pathways; however, each followed these to different degrees. The primary pathway for benzothiophene WBS direct hydrogenolysis and desulfurization to ethylbenzene; the minor pathway was the hydrogenation to 2-ethylcyclohexanethiol followed by desulfurization to ethylcyclohexane. The major pathway for o-cresol was deoxygenation to toluene without ring saturation; the secondary pathway occurred when o-cresol was hydrogenated to the postulated intermediate, 2methylcyclohexanol, that was later deoxygenated to methylcyclohexene, methylcyclohexane, and ethylcyclopentane. Benzofuran followed the same primary and secondary paths as 0-cresol, yielding ethylbenzene and ethylcyclohexane as the major end products. For quinoline, the major pathway was through saturation of the aromatic rings followed by hydrogenolysis, thereby yielding the reactive intermediate, 2-propylcyclohexylamine,that was later denitrogenated to 1-and 3-propylcyclohexenes, n-propylcyclohexane,and n-butylcyclopentane. The secondary pathway for quinoline was hydrogenation to 0propylaniline, which was in part directly denitrogenated to n-propylbenzene. Indole showed the same major and minor pathways as quinoline and yielded ethylcyclohexane as the major product and ethylbenzene as the minor product. Therefore, the extent of hydrogenation of the aromatic ring adjacent to the heteroatom-containing ring increased from the lowest level with sulfur, then with oxygen, and then highest with nitrogen. The extent of hydrogen usage in heteroatom removal followed the decreasing order of denitrogenation, deoxygenation, and desulfurization with the greatest usage in denitrogenation and the least in desulfurization. The lower the carbon and heteroatom bond strength, the greater was the ease of heteroatom removal.
Acknowledgment. We gratefully acknowledge the support of this work by the Department of Energy under Grant DE-FG2285PC80502. The provision of catalyst precursors by Shepherd Chemical is also gratefully acknowledged. The GC/MS analysis by Dr. George W. Goodloe is especially appreciated. BB BCP BZF BZT CR D DHBZF DHBZT DHQ
Nomenclature n-butylbenzene n-butylcyclopentane benzofuran benzothiophene o-cresol
trans- and cis-decalin dihydrobenzofuran dihydrobenzothiophene decahydroquinoline