Energy & Fuels 1995,9, 406-412
406
Hydrogenation of Model Compounds in Syngas-D20 Systems Yuan C. Fu,* Katsuya Ishikuro, Tomoya Fueta, and Makoto Akiyoshi Department of Applied Chemistry, Muroran Institute of Technology, Muroran, Japan 050 Received September 29, 1994@
Coal model compounds were hydrogenated and desulfurized in the presence of solvent and catalyst by using syngas with steam in place of hydrogen. Nickel molybdate catalyst exhibited good activities for hydrogenation and desulfurization of model compounds at high pressure and elevated temperature with the use of syngas and steam. Extensive water-gas shift conversion also took place simultaneously. Hydrogenation of phenanthrene and dibenzothiophene was carried out in syngas-water systems, and the results were compared with those obtained in Hz system. The extent of hydrogenation increased with the increase of HdCO mole ratio in syngas. Hydrogen consumed during the hydrogenation was replenished by the accompanying water-gas shift reaction in syngas-HzO systems. Hydrodesulfurization of dibenzothiophene carried out in Hz-CO-D~O and D2-CO-H20 systems yielded deuterated products of cyclohexylbenzene and biphenyl in both systems, indicating that both active deuterium produced via water-gas shift reaction of CO with D20 as well as deuterium in gas phase participated in the hydrodesulfurization along with the hydrogenldeuterium exchange reaction with substrates.
Introduction In coal liquefaction and coprocessing of coal and petroleum residues to produce liquid fuels, the process economics could be improved significantly by using syngas (Hz CO) and steam instead of hydrogen as feed gas to the reactor. Cost analysis of a typical coal liquefaction process by Kattell reveals that as much as 30% of the overall cost is related to hydrogen production. Batchelder and Fu2 made a comparative process evaluation for a catalytic liquefaction of coal using H2 and syngas and reported that the use of syngas reduces the capital and operating costs by eliminating shift convertors and gas purifying systems and thus could reduce the process cost of the H-coal process by about 14%.The concept of using syngas is to bypass the expensive gas purification and shift conversion steps and send the “raw” syngas from coal gasifier directly t o the liquefaction reactor along with the recycle gas and added steam. The CO in the feed gas reacts with steam to form H2 in the liquefaction reactor rather than separately in the water-gas shift system. It has been demonstrated that l i ~ i t eand ~ , bituminous ~ coal@ are liquefied and coal l i q ~ i dis~hydrogenated ,~ by the use of carbon monoxide
+
Abstract published in Advance A C S Abstracts, April 15, 1995. (1)Kattel, S.“SYNTHOIL Process-Liquid Fuel from Coal Plant”, Report No. 75-20, Bureau of Mines, Morgantown, WV, J a n 1975. (2) Batchelder, R. F.; Fu, Y. C. Ind. Eng. Chem. Process Des. Deu. 1979,18,594-599. (3) Appell, H. R.; Wender, I.; Miller, R. D. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1969,31 (41, 39. (4) Appell, H. R.; Miller, R. D.; Wender, I. Presented a t the Division of Fuel Chemistry, 163rd ACS National Meeting, Boston, April 1014 - ., 1972 - - . -.
(5) Fu, Y. C.; Illig, E. G. Ind. Eng. Chem. Process Des. Deu. 15, 392-396.
1976,
(6) Del Bianco; Girardi, E. Proc. Int. Conf Coal Sci. 1987, 355358. (7) Vernon, L. W.; Pennington, R. E. U.S. Patent 3719588, 1973. (8) Stephens, H. P.; Kottenstette, R. J. Presented a t the Division of Fuel Chemistry, 194th ACS National Meeting, New Orleans, August 30-September 4, 1987.
or carbon monoxide containing gas with water. We have also reported previously that coal could be coprocessed with petroleum solvents (mixtures of n-paraffin and cycloalkane) and that model compounds could be hydrogenated and desulfurized in the presence of petroleum solvents and catalyst using syngas with steam.9 It was considered that the CO in the syngas reacts with steam to form active hydrogen via water-gas shift reaction and that the hydrogen formed could participate in the hydrogenation of coal, but the fate of hydrogen formed from the water-gas shift reaction was not known. Reactions of coal-DzO mixtures in microwave discharge were studied earlier.lOJ1 More recently, deuterium was used as a tracer to study the mechanism of coal hydrogenation.12-16 It is known that deuterium is incorporated into reaction products during coal liquefaction under deuterium gas14-16 and that hydrogen/ deuterium transfer occurs among solvent, reaction products, and gas when deuterium-labeled solvents are used in coal liquefaction experiments.13 Kabe et al.17 recently studied the reactions of coals and model compounds with tritiated water and reported that the reaction mechanism with water was quite different from that with molecular hydrogen and that the hydrogens attached to aromatic carbons in coal and model com(9) Fu, Y. C.; Akiyoshi, M.; Tanaka, F.; Fujiya, K. Prepr. Pap.-Am. Chem. SOC., Diu. Fuel Chem. 1991,36(41,1887-1891. (10)Fu, Y. C.; Blaustein, B. D. Chem. Ind. 1967,1257-1258. (11)Kessler, T.; Sharkey, A. G. Spectrosc. Lett. 1968,1 , 177-180. (12) Schweighardt, F. K.; Bockrath, B. C.; Friedel, R. A,; Retcofsky, H. L. Anal. Chem. 1976,48,1254. (13)Cronauer, D. C.; McNeil, R. I.; Young, D. C.; Ruberto, R. G. Fuel 1982,61, 610-619. (14) Wilson, M. A,; Vassalo, A. M.; Collin, P. J. Fuel Process. Technol. 1984,8,213. (15) Skowronski, R. P.; Ratto, J. J.; Goldberg, I. B.; Heredy, L. A. Fuel 1984,63, 440-448. (16) Dabbagh, H. A,; Shi, B.; Davis, B. H.; Hughs, C. G. Energy Fuels 1994,8,219-226. (17) Ishihara. A,: Takaoka, H.: Nakaiima, E.: Imai, Y.: Kabe. T. Energy Fuels 1993,7,362-336.
0887-0624/95/2509-0406$09.00/0 0 1995 American Chemical Society
Syngas-DaO Systems
pounds were able to exchange with water easily. These studies used deuterium or tritium tracer to investigate the mechanism of liquefaction. Deuterium sources could be molecular deuterium, deuterium oxide, and deuterated donor solvents. The main difference between coal liquefaction using syngas-HzO system and using hydrogen system is that there is a hydrogen source of HzO in addition to molecular hydrogen in the syngas-HzO system. Thus, it would be useful to use deuterium oxide and molecular deuterium as deuterium sources to study the fate of hydrogen formed from the water-gas shift reaction taking place in syngas-HzO system. This study focused on the effects of hydrotreating model compounds in syngas-Hz0 systems using phenanthrene and anthracene as representative three-ring aromatic hydrocarbons, and benzothiophene and dibenzothiophene as sulfur-containing compounds. Model compounds were hydrotreated in syngas-water systems in the presence of solvent and catalyst, and the results of hydrogenation and desulfurization were compared with those obtained from hydrogenation in hydrogen system. In order to understand the roles of molecular hydrogen and water in the syngas-HzO system during the hydrodesulfurization, experiments were also carried out to hydrotreat dibenzothiophene in the presence of solvent and catalyst under pressure using H2-CO-DzO and Dz-CO-H20 systems. Deuterium gas and DzO were used in an effort to study the incorporation of hydrogen from the gas and water by tracing the migration of deuterium through the system. Admittedly, we have chosen a relatively simple system for our study. Coal liquefaction is much more complicated and hydrogen transfer from a hydrogen donor solvent is also an important process in the cracking of large coal molecule~.~~
Experimental Section Reactants. For aromatic model compounds, chemically pure grade anthracene and phenanthrene (Tokyo Kasei) were used. Sulfur-containing compounds of benzothiophene and dibenzothiophene (98% pure) were obtained from Aldrich. Tetralin, decalin, (98% pure, Kanto Chemical), or n-dodecane (Tokyo Kasei) was used as the solvent in most cases. Reactions and Analyses. The hydrogenation of model compounds was conducted in a shaking 25-mL microreactor in the presence of solvent and a commercial NiMo/AlzOa catalyst (2.2 wt % Ni, 8.0 wt % Mo; Nippon Mining Co.) using syngas-H2O systems at an initial pressure of about 70 kg/ cm2 in most cases. Hydrotreating conditions of model compounds were varied, but the typical reaction temperature and time were 400 "C and 45 min, respectively. The reactor was quickly heated up in a fluidizing sand bath and maintained a t the reaction temperature for 45 min and then rapidly quenched in a cold water bath. I t usually takes less than 2 min to reach the desired temperature and 5-10 s to cool the reactor contents to below 50 "C. Various H2:CO:HzO mole ratios with HdCO mole ratios varying from 1 t o 3 and HzO/ CO mole ratios varying from 0.3 t o 1.0 were used. For the experiments, a mixture of 30 parts by weight of model compounds and 70 parts of solvent was placed in the microreactor, 10 wt % (based on the mixture) of ground presulfided catalyst was charged, and a calculated amount of water was added before the reactor was pressured with syngas. The amount of the mixture added was typically 2.0 g. A couple of small stainless-steel balls were placed in the reactor to achieve better mixing. After the reaction, the liquid products
Energy & Fuels, Vol. 9, No. 3, 1995 407 Table 1. Hydrogenation of Model Compounds in Hz and Syngas-HZO Systemsa
H2
gas system model compounds anthracene conv, % phenanthrene conv, % benzothiophene conv, % ethylbenzene formed, % decalin remained, % trandcis ratio tetralin remaining, % naphthalene formed,b% H2 consumpn: % CO conv, %
A, B
P, B
-
92.6 99.9 72.7 129.0 3.4 55.6 5.4 8.0
99.9
99.9 80.5 116.4 3.3 65.6 6.8 7.0
-
-
-
H2-CO-H20 (1:1:0.5) A, B
p, B -
-
74.7 99.9 84.2 99.8 2.8 73.8 14.2 1.8 37.2
99.3
99.9 88.4 97.7 2.8 75.3 15.5 1.8 37.4
Solvent: decalinltetralinln-decane. Catalyst: NiMo/AlzOs Initial pressure: 70 kg/cm2. Temperature 400 "C, time 45 min. Equal w t % of each component. A, anthracene; B, benzothiophene; P, phenanthrene. * Based on tetralin feed. Wt % of model compounds. were washed with tetrahydrofuran and analyzed by a Shimadzu GC-14A gas chromatograph using OV-1 fused silica capillary column (60 m x 0.25 mm o.d.), and gaseous products were analyzed by a Shimadzu GC-14A gas chromatograph. Material balances based on moles of reactant disappearance and solvent recovery were usually better than 94%. Durene or dibenzyl was used as an internal standard to analyze the samples, depending on the retention times of different product species involved. Reaction products were identified by comparing retention times with pure compounds and by gas chromatography-mass spectrometry using a Shimadzu QP1000. In experiments using D2 gas and DzO, an initial pressure of about 40 kg/cm2was used because of low pressure available for the D2 bomb. The liquid products were analyzed by GCMS. Spectral analysis was performed using the known split patterns of various undeuterated compounds. For these runs, samples of the gaseous products were obtained by passing through a dry ice and acetone trap to remove moisture prior to analysis using gas chromatograph and an ULVAC MSQ150A quadrupole mass analyzer. For hydrogen analysis, relative intensities of peaks of molecular ions were used t o estimate the relative portions of H2, HD, and D2.
Results and Discussion Hydroprocessing of Model Compounds. During coal liquefaction, large coal molecules fragment thermally and are hydrogenated via hydrogen transfer from a donor solvent. Coprocessing of coal with petroleum solvents using either hydrogen or syngas-HzO system was also aided by the presence of tetralinugAnthracene, phenanthrene, and benzothiophene were initially hydrotreated in the presence of a mixture of petroleum solvents containing decalin, tetralin, and n-decane to study the hydrogenation and desulfurization in a syngas-H2O system. It is seen in Table 1 that both anthracene and benzothiophene were hydrogenated or hydrodesulfurized easily in either Hz or syngas-HzO system in the presence of NiMo/AlzO3 catalyst. The conversion of phenanthrene was somewhat lower in the syngas-H20 system. In the syngas-HzO system, the extent of water-gas shift reaction was extensive and hydrogen consumption was lower, suggesting that Hz consumed in the hydrogenation reactions was replenished by the Hz formed from the CO-shift conversion. The CO conversions shown in the table were estimated on the basis of mole percent of CO converted t o CO2. The products formed from hydrotreating each reactant
408 Energy & Fuels, Vol. 9, No. 3, 1995
El
0 Phenanthrene
0 2HP
A IHP
A 8HP .PW.biph
OBW!
Reaction Temperature ("C)
Figure 1. Hydrogenation of phenanthrene at various temperatures (initial Hzpressure 30 kg/cm2;time 30 min).
will be discussed separately in the later sections. It was noted that hydrogenation of tetralin to form decalin was favored in the H2 system, whereas dehydrogenation of tetralin to form naphthalene was favored in the syngasH2O system. Under the hydrotreating condition, the petroleum solvents could also be significantlyhydrotreated, which could impact the results. It is also of interest to note that the trans f o d c i s form ratio of the remaining decalin increased t o 2.8-3.4 from the original trans/ cis ratio of 1.3 after the hydrotreating. The occurrence of isomerization from cis-decalin to trans-decalin was also observed by Clarke et al. during extraction of coal by decalin.l8 Hydrogenation of Phenanthrene. Shabtai et al.I9 studied hydrogenation of phenanthrene at 200-380 "C and reported that the first principal product was 1,2,3,4tetrahydrophenanthrene (4HP),which was converted to 1,2,3,4,5,6,7,8-octahydrophenanthrene (8HP) and perhydrophenanthrene (PHP). The yield of 9,lO-dihydrophenanthrene (2HP) was lower. Recent study by Lee and Satterfield,2O however, showed that 2HP was the primary product at 360 "C. Their kinetic model indicated that 4HP, the secondary product, formed 8HP with the progress of the reaction. Our objective was to evaluate the effects of hydrotreating phenanthrene in syngas-H2O systems a t reaction temperature and residence time comparable t o those under coal liquefaction conditions. Figure 1shows a plot of mole percent product distribution as a function of temperature, when phenanthrene was hydrogenated in n-dodecane solvent at initial hydrogen pressure of 30 kg/cm2 and reaction time of 30 min. At the reaction time of 30 min, the initial primary products at lower temperatures were 9,lO-dihydrophenanthrene(2HP), 1,2,3,4,5,6,7,8-octahydrophenanthrene (8HP),and 1,2,3,4tetrahydrophenanthrene (4HP). At 370 "C, 2HP decreased, but 4HP and 8HP became the main products. In addition to 8HP, cis- and trans-isomers of 1,2,3,4,4a,9,10,10a-octahydrophenanthrene(8HPI) were de(18) Clarke, J. w.;Rantell, T. D.; Snape, C. E. Fuel 1984,63,14761479. (19) Shabtai, J.; Veluswamy, L.; Oblad, A. G. Prepr. Pup.-Am. Chem. SOC.,Diu. Fuel Chem. 1978,23,107-113. (20) Lee, C. M.; Satterfield, C. N. Energy Fuels 1993,7 , 978-980.
F u et al.
tected as doublet peaks. Small amounts of perhydrophenanthrene (PHP) and hydrocracking products, biphenyl and unknowns, were also observed and lumped together in the figure. Figure 1 also shows that phenanthrene conversion decreased at 370 "C, probably due to the equilibrium shift toward the aromatic species at low H2 pressure and higher temperat~re.'~ Table 2 shows the results of hydrotreating phenanthrene at 370 "C and 70 kg/cm2 initial pressures in different gas systems using either decalin or tetralin as the solvent. Under N2 pressure, tetralin showed good activity as an effective donor solvent with higher yield of naphthalene, though the phenanthrene conversion was low. In Hz and syngas-H20 systems, however, no significant beneficial effects on the hydrogenation could be observed with the use of tetralin donor solvent except for some increase in the 2HP yield. The conversion of phenanthrene and the yields of 8HP and PHP were somewhat lower in the syngas-H20 system than in the H2 system. The extent of hydrogenation was higher with higher HdCO mole ratio of syngas. This seems reasonable because there is more hydrogen available at the start of the reaction and it is replenished as the reaction continues. Tetralin underwent disproportionation as was observed by Satterfield20and Curtis et a1.21 during hydrogenation of phenanthrene and pyrene, respectively. In the hydrogenation of phenanthrene at 370 "C for 45 min with H2 and NiMo/AlzOs catalyst, tetralin showed somewhat higher activity, with 11.3% hydrogenated to decalin and 2.0% dehydrogenated to naphthalene. In syngas-H2O systems, the dehydrogenation of tetralin was slightly favored. During the phenanthrene hydrogenation, isomerization of cis-decalin did not occur when decalin was used as the solvent, though the formation of trans-decalin rather than that of cis-decalin from tetralin was favored (trans/cis ratio = 2.3-2.7). We noticed that the isomerization of cisdecalin to trans-decalin was particularly pronounced when decalin was used as the solvent in the hydrogenation of sulfur-containing compounds (see later section also). To observe the progress of phenanthrene hydrogenation in syngas-H2O systems, the product distribution was plotted as a function of time during the hydrotreating in an H2-CO-H20 (1:1:0.5) system at 370 "C (see Figure 2). Similar to that observed in Figure 1, the initial primary products were 2HP and 4HP, and 8HP and 4HP became the major products as the phenanthrene conversion progressed. The latter observation agreed well with Satterfield's modellg that 8HP was formed from 4HP. Water-Gas Shift Reaction. During the phenanthrene hydrogenation in syngas-H20 system, watergas shift reaction and methane formation also occurred. Figure 3 shows the progress of CO-shift conversion and CHI formation with time. The CO-shift conversion and CH4 formation observed during a blank test at 400 "C with syngas, water, and NiMo/AlzOs catalyst but no added phenanthrene o r solvent were also included for comparison. It was noted that the CH4 formation during the blank run was rather substantial and reached about 14 mol % after 45 min. But the CH4 formation decreased markedly when phenanthrene was (21) Ting, P. S.; Curtis, C. W.; Cronauer, D. C. Energy Fuels 1992, 6 , 511-518.
Syngas-DZO Systems
Energy & Fuels, Vol. 9, No. 3, 1995 409 Table 2. Hydrogenation of Phenanthrene over NiMo/&Os Catalysta H2
decalin
tetralin
Hz-CO-H20 (2:1:0.5)
Hz-CO-HzO (1:1:0.5)
decalin
decalin
tetralin
tetralin
Nz decalin
tetralin ~~
product distribution, % phenanthrene 2HP 4HP 8HP 8HPIb other productsC solvent distribution, % decalin trandcis ratio tetralin naphthalene C O conversion, % H2 consumpn, wt % of phenanthrene
3.8 3.6 7.7 54.4 10.8 19.7
9.1 6.4 12.0 51.5 15.3 5.7
18.7 7.7 18.5 40.1 11.1 3.9
25.2 10.1 18.4 34.1 8.9 3.3
26.8 9.1 19.3 32.8 8.6 3.4
34.0 12.8 19.3 23.9 6.3 3.7
92.0 3.0 2.3 0.8 1.9
75.4 6.9 12.6 2.5 0.7 1.9
99.6 1.3 0.4
11.3 2.7 86.7 2.0 4.8
99.7 1.3 0.3 28.0 1.9
4.7 2.4 90.4 4.9 29.3 2.3
99.6 1.3 0.4 25.6 -0.1
3.0 2.3 92.7 4.3 27.1 -0.7
94.3 1.3 2.0 3.7 -
0.8 1.5 58.4 40.8 -
-
4.5
I
a Initial pressure 70 kg/cm2;temperature 370 "C; time 45 min. 2HP,9,lO-dihydrophenanthrene; 4HP,1,2,3,4-tetrahydrophenanthrene; 8HP,1,2,3,4,5,6,7,8-octahydrophenanthrene; 8HP1,1,2,3,4,4a,9,9a,lO-octahydrophenanthrene. Include cis- and trans- forms. Include perhydrophenanthrene, biphenyl, and unknowns.
OPhenanthrcne A LHP
0 2HP A BHP
OBHPl
PHP.blph
Reaction Time (min)
Figure 2. Hydrotreating phenanthrene in Hz-CO-HzO (1: 1:0.5)system (initial syngas pressure 70 kg/cm2;temperature 370 "C).
-"I 0
E
L
x f
t
z
C
c C
c
OOG
-
15
1
I
1
30
45
60
Reaction Time (min)
Figure 3. Progress of water-gas shift reaction and methane formation with time in syngas-HzO systems. Open symbols represent blank run without phenanthrene a t 400 "C; filled symbols represent run during phenanthrene hydrotreating at 370 "C.
present in the system as shown in the plot, suggesting that H2 concentration in the gas phase might be decreasing rapidly because of hydrogen uptake by phenanthrene and that the CO-shift conversion was undergoing actively to replenish the H2 consumed under this condition. When dibenzothiophene was hydrotreated in syngas-H2O systems at 400 "C for 45 min, the
CO conversion to CH4 was also found to be in the low range of 3-4 mol %. Hydrodesulfurization of Dibenzothiophene. Dibenzothiophene was hydrotreated in hydrogen and syngas-H2O systems in the presence of either decalin or tetralin solvent at 400 "C for 45 min. The results are shown in Table 3. Dibenzothiophene was easily hydrogenated and desulfurized by hydrogen to form mainly biphenyl, cyclohexylbenzene, bicyclohexyl, and decomposed products. Tetralin donated some hydrogen to form naphthalene, but was also hydrogenated to form decalin. Decalin converted to small amounts of tetralin and naphthalene. It was noted that the transhis ratio of decalin increased substantially t o greater than 4 in most runs. Although we do not know its significance, it cannot be explained why the extent of decalin isomerization is greater in hydrodesulfurization than in hydrogenation. Hydrodesulfurization with H2 resulted in high dibenzothiopheneconversions in both good and not so good donor solvents, indicating that catalytic hydrogenation by hydrogen gas plays the dominant role with respect to the hydrodesulfurization. When dibenzothiophene was treated in syngas-H20 systems, the hydrogenation and desulfurization activities were lower. The yields of hydrogenated products, cyclohexylbenzeneand bicyclohexyl, were lower. These results indicate that the performance of syngas in hydrodesulfurization did not come up to the high level that was observed in coal liquefaction. In coal liquefaction, the performances with syngas and H2 compare rather closely except that asphalten levels of the syngas products were somewhat higher.5 Table 3 also shows the results of hydrodesulfurizing dibenzothiopheneusing decalin solvent and syngas with steam at various H2:CO:H20 ratios. The dibenzothiophene conversion increased with HdCO mole ratio, while the H20/CO mole ratio was maintained at 0.6.The data indicate that the H20 level at the H20/CO mole ratio of 0.3-0.6 is appropriate to give good activities for water-gas shift conversion. Insufficient amounts of water may give low CO conversion and yield insufficient amounts of H2 while excessive amounts of water may leave unreacted water in the system. The extent of water-gas shift reaction was rather extensive as can be seen from high CO conversions obtained in all runs. When H20/CO mole ratio was 0.3,the moles of CO converted was greater than the moles of H20 added,
410 Energy & Fuels, Vol. 9, No. 3, 1995
Fu et al.
Table 3. Hydrodesulfurization of Dibenzothiophene at Various H2/CO/H20 Mole Ratios= Hz-CO-HzO Hz-CO-H~O Hz-CO-H~O H2-CO-HzO 1:1:1 1:1:0.3 1:1:0.6 2:1:0.6 H2 decalin tetralin decalin tetralin decalin decalin deca 1in product distribution, % dibenzothiophene 7.2 6.5 26.4 23.0 31.5 34.1 23.5 bicyclohexyl 8.8 4.2 1.7 1.9 1.6 1.4 2.7 cyclohexylbenzene 29.3 36.7 22.3 25.4 19.9 20.1 26.6 biphenyl 39.8 44.5 43.9 43.4 41.7 38.5 40.5 unidentified products 14.9 8.1 5.7 6.3 5.3 5.9 6.7 solvent distribution, % decalin 96.5 18.5 94.9 5.0 95.7 96.4 96.1 translcis ratio 4.8 4.0 4.0 4.0 4.0 3.3 4.3 tetralin 3.3 73.8 3.9 76.7 3.6 3.1 3.4 naphthalene 0.2 7.7 1.2 18.3 0.7 0.5 0.5 CO conversion, % 32.5 39.3 34.7 37.0 25.8 H2 consumpn, wt % of 4.2 6.0 2.0 1.8 0.9 -1.0 1.6 dibenzothiophene HdCO ratio after reaction 1.5 1.7 1.6 1.4 2.7 a
Hz-CO-H~O 3:1:0.6 decalin 20.8 3.0 28.8 40.5 6.9 96.5 4.3 3.1 0.4 35.3 2.7 5.2
Catalyst: NiMoiAl203. Solvent: decalin. Initial pressure 70 kg/cm2; temperature 400 "C; time 45 min.
Table 4. Hydrodesulfurization Using Syngas-DzO or Deuterium Containinn Gasesa Hz-CO-D~O D2-CO-H20 NZ-D~O (1:l:l) (1:l:l) (2:l) product distribution, % dibenzothiophene 29.1 34.0 68.7 bicyclohexyl 0.7 0.3 cyclohexylbenzene 16.1 12.7 0.6 biphenyl 50.1 50.2 28.8 unidentified products 4.0 2.8 1.9 solvent distribution, % deca1in 91.3 91.2 88.4 trans/cis ratio 3.8 4.0 2.2 tetralin 6.0 5.7 4.0 naphthalene 2.7 3.1 7.6 CO conversion, 5% 27.2 26.5
0 :I
e : DBT
50
-? -o
Catalyst: NiMo/AlzOs. Solvent: decalin. Initial pressure 40 kg/cm2;temperature 400 "C; time 45 min.
suggesting that water formed from methane formation was also participating in the shift conversion. The CO conversions were calculated on the basis of the amounts of CO2 formed. As was observed in coal liquefaction using ~ y n g a sH2 , ~ formation via water-gas shift reaction resulted in reduction of hydrogen consumption and increase of HdCO mole ratio after the reaction. The HdCO mole ratio of exit gas increased, because more CO than H2 was consumed. This is desirable because this H2-rich gas (after removal of C02) can be recycled and mixed with CO-rich "raw" syngas to increase the HJCO mole ratio of the feed gas to the liquefaction reactor. HydrodesulfurizationUsing Syngas Containing D20 and D2. Experiments with syngas containing D20 and D2 were carried out to observe how the hydrogentransfer reactions occurred from HzO and gas-phase hydrogen. Dibenzothiophene was hydrodesulfurized in the presence of decalin solvent using H2-CO-D20 (1: 1:l) and D2-CO-H20 (1:l:l) gas systems as shown in Table 4. Another experiment using N2-D20 (2:l) system was also carried out to determine the occurrence of WD exchange between dibenzothiophene and D2O in the absence of water-gas shift reaction. Initial pressures of syngas and N2 used were lower (about 40 kgl cm2)in this series of experiments because only a limited pressure of D2 gas was available. Similar to the results shown in Table 3, both the hydrodesulfurization and water-gas shift reaction progressed moderately. In the experiment using the N2-D20 system, the diben-
- W n
A : 9lph.nyl
0 :cHB
I7 \
y-40syobm
0 di
Figure 4. Distribution of deuterium in solvent and products from hydrodesulfurization of dibenzothiophene in different gas systems (CHB, cyclohexylbenzene; DBT, dibenzothiophene).
zothiophene conversion decreased markedly and some dibenzothiophene decomposed to form biphenyl. The formation of naphthalene from decalin increased a little in this case. The reaction products and the remaining dibenzothiophene and solvent were analyzed using GC-MS, and deuterium distributions are shown in Figure 4. In the H2-CO-DzO system, the cyclohexylbenzeneproduct was deuterated extensively to form dz, d3, dq, d g , and d6 species, but no do or dl species was present. Deuterium was also distributed widely to biphenyl and unreacted dibenzothiophene, and all species from do to de were present. Large parts of cis-decalin were isomerized to the trans form (transhis ratio was 1.3 before the reaction), buth both forms similarly contained about 30% of decalin-dl. Only trans-decalin is shown in Figure 4. The WD exchange of D20 with the decalin solvent was not as extensive as that with the products. In the D2-CO-H20 system, it was noted that deuterium was incorporated moderately into cyclohexylbenzene and biphenyl, both forming dl, da, and d3 species.
Syngas-DZO Systems
Energy &Fuels, Vol. 9, No. 3, 1995 411
Table 5. GC and MS Analysis of Gaseous Products H2-CO-D20 (1:l:l)
Dz-CO-Hz0 (1:l:l)
coz Nz hydrogen (Hz) (HD) (D2)
Table 6. Hydrogenation of Anthracene Using Syngas-D20- or Deuterium-ContainingGasesa Hz
gaseous products: mol %
co
N2-D20 (2:l)
32.8 13.6 46.1 (63) (31) (6)
37.2 14.5 42.5 (13) (11)
(76)
-
94.9 3.6 (100)
0 0
The remainders are CH4, CzHs, and HzS.
Only about 20% of d l species was present in either trans- or cis-decalin. It is of interest to observe that, under N2 pressure, H/D exchange occurred extensively between D2O and dibenzothiophene. The yield of biphenyl was lower, but deuterium was incorporated into the product extensively. Deuterium was also widely distributed t o the unreacted dibenzothiophene. The formation of deuterated cyclohexylbenzene, however, was almost negligible. The extensive H/D exchange observed between dibenzothiophene and DzO under Nz pressure is consistent with the results reported by Kabe et al.17 dealing with hydrogen exchange of HzO with coal or phenolic model compounds. In the case dealing with hydrogen exchange between syngas-H2O and dibenzothiophene, the situation is not the same. Water participates in the water-gas shift reaction as evidenced by the formation of C02. It is expected that the D2O in the syngas-DzO system reacts with CO to form active deuterium which leads to the formation of D2, and in the process, some active deuterium may be incorporated into dibenzothiophene and the substrates. In the N2-D20 system, as will be described later, the deuterated species were formed by exchange reactions rather than by hydrogenation reaction, because no deuterium exists in the gas phase. To observe the extent of D atoms from D20 transfer back to form dihydrogen such as HD and D2 in gaseous products, the gases were analyzed by gas chromatograph and quadrupole mass spectrometer. The analytical results are given in Table 5. It is noted that some deuterium in D20 is present in HD and D2 in the syngas-D20 system, while no deuterium is present in the “hydrogen gases” of the gaseous products in the N2D2O system. A small amount of H2 formed in the N2D2O system may have come from the dehydrogenation or decomposition of the solvent and reactant. In contrast t o the experiment in the Hz-CO-D~O system where HD and D2 were formed, the experiment in the D2-CO-H20 system yielded gaseous products containing HD and H2 in the “hydrogen gases”. These results indicate that, unlike D20 in the N2-D20 system, D2O in the syngas-DzO system or H2O in the D2-CO-H20 system forms active deuterium or hydrogen, respectively, which in turn is incorporated into the reactant and substrates or leads to the formation of HD and D2 or HD and H2, respectively. In the N2-D20 system, most of the deuteriums are probably incorporated into the reactant by exchange reactions with D2O. It may then be concluded that both gas-phase hydrogen and hydrogen formed via the water-gas shift reaction in the syngas-H2O system contribute to hydrodesulfurization and that the accompanying water-gas shift reaction is
product distribution, % anthracene 2HA 4HA 8HA 8HAIb PHA other products solvent distribution, % decalin trandcis ratio tetralin naphthalene
H2-CO-Dz0 (1:l:l)
Dz-CO-H~O (1:l:l)
0.4 1.5 18.5 45.1 18.3 5.2 11.0
7.1 13.4 59.4 11.6 3.7 0.0 4.8
4.7 9.0 63.5 13.7 4.4 0.3 4.4
98.7 1.3 1.3 -
93.3 1.3 3.4 3.3
97.2 1.3 2.1 0.7
Initial pressure 40 kg/cm2. Solvent: decalin. Temperature 400 “C;time 45 min. 2HA, 9,lO-dihydroanthracene; 4HA, 1,2,3,4tetrahydroanthracene;8HA, 1,2,3,4,5,6,7,8-0ctahydroanthracene; 8 W , 1,2,3,4,4a,9,9a,lO-octahydroanthracene; PHA, perhydroanthracene. Include cis and trans forms.
Table 7. Deuterium Distribution in Products from Hydrotreatment of Anthracene at 400 ‘Ca deuterium content (%) H2-CO-D20 system 2HA - 13 10 24 4HA 1 2 4 10 anthraceneremaining 6 12 19 22 Dz-CO-HzO system 2HA 22 19 31 19 4HA 8 23 29 25 anthraceneremaining 19 30 29 16
22 21 10 - 20 22 21 13 7 20 13 6 2 9 11
6
4 -
-
- - - -
a 2HA, 9,lO-dihydroanthracene; 4HA, 1,2,3,4-tetrahydroanthracene.
beneficial in reducing Hz consumption, from gaseous Hz that was contained in syngas charged to the reactor. Hydrogenation of Anthracene Using Syngas Containing Dz0 and Dz. Anthracene was also hydrotreated in H2, syngas-D20, and D2-CO-H20 systems at 400 “C for 45 min as shown in Table 6. The conversion of anthracene was high and the hydrogenated anthracenes detected were 9,lO-dihydroanthracene (2HA), 1,2,3,4-tetrahydroanthracene(4HA), 1,2,3,4,5,6,7,8-octahydroanthracene (8HA),two isomers of 1,2,3,4,4a,9,10,lOa-octahydroanthracene (8HAI),and perhydroanthracene (PHA). The progress of hydrogenation was greater with H2, and more than 80% of the products were 8HA, 8HAI, and PHA. In syngas-water systems, whether containing D20 or D2, the yield of 2HA increased and 4HA became the dominant product. These trends are somehow similar to those observed in the phenanthrene hydrotreatment. Table 7 shows deuterium distributions in 2HA, 4HA, and the remaining anthracene obtained from hydrotreatment in H2-CO-DzO and D2-CO-H20 systems. Both products and the remaining anthracene were deuterated extensively. H/D exchange occurred more extensively in the H2-CO-D20 system than in the Dz-CO-H~O system. In general, these results were in accord with those observed in the hydrotreatment of dibenzothiophene in the H2-CO-DzO and D2-CO-HzO systems, where HD and Dz were formed via water-gas shift conversion, and in the process some deuterium was incorporated into the substrates. It was also observed that, unlike the case of hydrotreating dibenzothiophene,
412 Energy & Fuels, Vol. 9, No. 3, 1995
the trandcis ratio of decalin solvent did not change at all after the hydrotreating of anthracene.
Fu et al.
reaction with substrates. Thus the accompanying watergas shift reaction during hydrotreatment of coal model compounds in syngas-H20 system would be beneficial in replenishing hydrogen consumed. The HdCO mole Conclusion ratio of the exit gas increased after the reaction, because Coal model compounds could be hydrogenated and of the greater consumption of CO rather than H2. hydrodesulfurized in syngas-H20 systems in the presThe results obtained with the use of model compounds ence of NiMo/AlzOs catalyst. During hydrotreating of indicate that coal liquefaction activity in syngas-HzO model compounds in syngas-H20 systems, the CO system could approach to the level comparable to that conversion to form CH4 was depressed and the CO-shift with H2. The off-gas, with increased HdCO mole conversion t o form Ha underwent actively. The extent ratio, can be recycled and mixed with the “raw” synof phenanthrene hydrogenation in syngas-H20 systems gas from coal gasifier and fed to the liquefaction rewas somewhat lower than that in H2 system, but the actor. The water-gas shift reaction would take place reaction path was similar in that the initial primary simultaneously during coal liquefaction. The obvious products were 9,lO-dihydrophenanthreneand 1,2,3,4- advantage of using syngas approach is the elimination tetrahydrophenanthrene, and that 1,2,3,4,5,6,7,8-0~of expensive purification and shift conversion steps for tahydrophenanthrene and 1,2,3,4-tetrahydrophenan- HZproduction from the “raw”syngas. It is believed that threne became the major products as the phenanthrene the use of syngas in the first stage of a two-stage process conversion progressed. instead of pure H2 would improve net hydrogen utilizaHydrotreating of dibenzothiophene carried out in an tion. H2-CO-D20 system yielded deuterated hydrogenation products and gaseous products of HD and Dz, indicating Acknowledgment. We thank M. Yamamoto and T. that active deuterium produced via water-gas shift of Hokkaido National Industrial Research Kotanigawa reaction of CO with D2O as well as hydrogen in gas for their assistance in GCNS analysis. Institute phase could both participate in the hydrogenation . EF9401841 reaction along withthe hydrogen/deuterium exchange
