Ind. Eng. Chem. Process Des. Dev. lQ01, 20, 68-73
60
Overall hydrodenitrogenaton rate is the very strong adsorption Onto the catalyst of secondary minesformed as reaction intermediates. Acknowledgment
Literature Cited Beuther. H.; Larson, 0. A. Ind. Eng. Chem. Process Des. D e v . 1985, 4 , 177.
Cocchetto, J. F. Ph.D. Thesis, M.I.T., Cambridge, M ~ S S . ,1979. Goudriaan, F.; " a n . H.; Vlugter, J. C. J. Inst. Pet. 1973, 59, 40. Goudriaan, F. Thesis, Twente University of Technology. Enschede, The Netherlands. 1974. Gultekin,-S. Ph.D. Thesis, M.I.T., Cambridge, Mass., 1980. Hk"lblau. P. M.: Jones. C. R.: Bischoff, K. B. I d . €nu. Chem. Fundam.
Received for review April 14,1980 Accepted August 4,1980
Hydrogenation of Aromatic Hydrocarbons Catalyzed by Sutfided Co0-MOO&-AI,O,. Reactivities and Reaction Networks Ajit V. Sapre and Bruce C. Gates' Center for Catalytic Science and Technology, Department of Chemical Engineerlng, University of Delaware, Newark, Delaware 19717
Operation of a reactor for hydroprocessing of coalderived liquids to be recycled to a coal liquefaction reactor requires optimum conversion of aromatics to hydroaromatics. In this work hydrogenation of aromatic compounds (benzene, biphenyl, naphthalene, and 2-phenylnaphthalene) was studied with a batch reactor at 75 atm and 325 OC in the presence of particles of sulfided Co0-MoO3/y-AI2O3 catalyst. Each reaction network involved reversible hydrogenation of the aromatic hydrocarbon (e.g., naphthalene)to give a hydroarometic hydrocarbon (e.g., tetralin), which experienced further, slow hydrogenation (e.g., to decalin). Slow isomerization reactions of biphenyl to give methylcyclopentylbenzeneswere also observed. The quantitative kinetics data are summarized in networks with each reaction being represented as pseudo first order in the hydrocarbon reactant.
Recent forecasts of energy supply and demand show a projected shortfall in liquid hydrocarbons which will become severe by the end of this century (Ford Foundation, 1979). There has consequently been a surge of development work on processes for the conversion of coal, oil shale, tar sands, and petroleum residua into liquid fuels. Upgrading of these heavy fossil fuels, which is essential for the production of commercially useful liquid products, is normally effected by hydrotreating in the presence of catalysts like "cobalt molybdate" or "nickel molybdate" at 50-250 atm and 300-450 "C (Gates et al., 1979). In these applications, the principal catalytic reactions are hydrodesulfurization, hydrodenitrogenation, hydrocracking, and hydrogenation of aromatics. Among the slowest of these reactions are the aromatic hydrogenations. These take on great importance because nitrogen removal from polycyclic aromatics does not take place until ring saturation has occurred (Shih et al., 1977; Bhinde, 1979), and hydrodenitrogenation is required to prevent poisoning of downstream catalysts (e.g., cracking catalysts) by nitrogen-containing compounds. Hydrogenation of aromatics is also important in the coal liquefaction technology now under development, as exemplified by the donor solvent process (Neavel, 1976) and some forms of the solvent refined coal process (National Academy of Sciences, 1977). In these processes, the liquefied coal undergoes catalytic hydrogenation and is then recycled to the liquefaction reactor, where it transfers hydrogen to the liquefying coal (Furlong et al., 1976).
The composition of the recycle solvent is determined by the nature of the liquefying coal. In bituminous and subbituminous coals, aromatic lamallae are believed to contain predominantly one, two, three, and four-ring components (Wiser, 1975). The prevalence of small molecular units in the coal structure is substantiated by analyses of liquefaction products (Ruberto et al., 1977). A large part of coal resembles either naphthalene or biphenyl (Harrison, 1966; Davies and Lawson, 1967). The average structures of the liquefaction products in the heavier oils and lighter asphaltenes (having average molecular weights of about 400) (Farcasiu, 1977) imply that, upon hydroprocessing, large quantities of two-ring aromatics like naphthalene and biphenyl would be formed. Direct evidence of one- and two-ring aromatics in the oils formed in coal liquefaction is provided by NMR, infrared, and high-pressure liquid chromatography analyses (Jones et al., 1980). All of these results indicate the importance of aromatic hydrogenation of one- and two-ring aromatics in the hydroprocessing of coal-derived liquids. It therefore appears that with the best available catalysts, hydrogenation of aromatics is the class of hydroprocessing reactions most deserving of study in regard to synthetic fuels conversion. The literature provides only fragmentary and qualitative information about the reaction networks and reactivities of aromatic hydrocarbons, most of the work having been done with unpromoted catalysts under conditions different from those of industrial interest (Weisser and Landa, 1973). The reported kinetics data
0196-4305/81/1120-0068$01.00/00 1980 American Chemical Society
Ind. Eng. Chem. Process Des. Dev., Vol. 20, No. 1, 1981
for hydrogenation of aromatic hydrocarbons in the presence of promoted sulfided catalysts like "cobalt molybdate" are limited to benzene, toluene, naphthalene, and phenanthrene (Voorhoeve and Stuiver, 1971; Ahuja et al., 1970; Huang, 1976; Rollmann, 1977; de Beer, et al., 1976; Hagenbach et al., 1971). The experiments reported here were carried out to establish the reactivities and reaction networks for hydrogenation of benzene, biphenyl, naphthalene, and 2phenylnaphthalene in the presence of a commercial catalyst (sulfided "cobalt molybdate," Co0-MoO3/~-Al2O3)at a pressure (75 atm), and temperature (325 "C) representative of industrial conditions. Biphenyl and 2phenylnaphthalene were chosen as model reactants in part because they are the principal aromatic products formed in the hydrodesulfurization of dibenzothiophene and of benzo[b]naphtho[2,3-d]thiophene,respectively (Houalla et al., 1978; Sapre et al., 1980). The latter two compounds are typical of the least reactive sulfur-containing compounds found in the heavier petroleum residua and coal-derived liquids. Naphthalene was chosen because, upon hydrogenation, it forms tetralin, the model compound most commonly used as a hydrogen donor in coal liquefaction experiments. Experimental Section Equipment and Procedures. The reaction experiments were carried out in a commercial 1-L batch reactor (Autoclave Engineers) equipped with a magnetically driven turbine stirrer, thermowell, furnace, cooling coil, gas inlet, and sampling lines. The design of the equipment allowed charging of the catalyst after the reactants had been brought to reaction temperature, thereby virtually eliminating any heat-up period which would prevent elucidation of isothermal kinetics from the conversion data (Shih et al., 1977). The catalyst was particles ( 150 pm) of COO-Moo3/ 7-Alz03 (American Cyanamid HDS 16A),2 g of which were externally sulfided at 400 "C with 10 vol % HzS in Hz at atmospheric pressure for 2 h prior to being charged to the reactor. The catalyst composition and physical properties are given elsewhere (Houalla et al., 1978). The reactant solution (375 mL) contained roughly 0.75 mol % of the hydrocarbon reactant in n-hexadecane (Humphrey Chemical Co., redistilled). The solution was saturated with hydrogen at the reaction temperature, providing a large stoichiometric excess of this reactant. The agitator speed was about 1500 rpm. Calculations and comparisons with data obtained in complementary experiments showed that the interphase and intraparticle mass transfer resistances were negligible. The reactor was operated at 75 f 2 atm and 325 f 1"C in all the experiments. Carbon disulfide, 0.1 mol ?G , was added to the reaction mixture, providing a source of sulfur that was essential to stabilize the catalyst by maintaining it in the sulfided form (Sapre, 1980). The CSzwas almost instantaneously converted into HzS under the reaction conditions, the rate of the hydrogenolysis reaction CSz + 4Hz 2HzS + CHI being at least 8000 times greater than that of the aromatic hydrogenation reaction. Materials. The reagents naphthalene (>99%) and benzene (>99 % ) were supplied by Fisher; bicyclohexyl (>98%), cyclohexylbenzene (>98%), and tetralin (>99%) by Aldrich; biphenyl (>98%) by Eastman, and 2phenylnaphthalene (>98%) by ICN Pharmaceuticals. All these reagents were used without further purification. Product Analysis. Product samples were drawn periodically from the reactor and analyzed with a PerkinElmer 3920B gas chromatograph (GLC) equipped with a N
-
69
flame ionization detector and a 0.8 mm 0.d. X 0.025 mm i.d. X 100 m long wall-coated open tubular column (the liquid phase was OV-101, Perkin-Elmer) with an all-glass splitter injector. The column temperature was varied depending on the reactant compound, and temperature programming was used to improve peak resolution (Sapre, 1980). Principal reaction products formed from benzene, biphenyl, and naphthalene were identified by comparing GLC retention times of components of the product mixtures with those of known compounds. For identification of the primary products of 2-phenylnaphthalene hydrogenation, tricarbonylchromium complexes of the products were synthesized using chromium hexacarbonyl (Nicholls and Whiting, 1959). The tricarbonylchromium complexes which were formed were separated by column chromatography using alumina (6% water) as the solid phase. The hydrocarbon solvent and uncomplexed compounds were eluted with n-pentane. The separation of the complexes on the column was achieved with different proportions of n-pentane and methylene chloride as the eluting solvent. The structures of the recrystallized complexes were determined with proton NMR spectroscopy; details of this technique are given elsewhere (Sapre et al., 1980). Results Reaction Networks. Blank experiments performed with hydrogen and biphenyl or with hydrogen and cyclohexylbenzene in the batch reactor confirmed that there was no observable conversion of these reactants even under conditions more severe (400 "C and 100 atm) than those used in the catalysis experiments. These results confirm that all the observed reactions were catalytic, being consistent with thermochemical calculations (Ross and Blessing, 1979) showing the stability of the C-C bonds in these molecules toward homolysis and free radical formation under the conditions studied here. The thermal dissociation of tetralin in the presence of hydrogen is observable at temperatures exceeding 350 "C and becomes significant only at temperatures exceeding 400 "C (Hooper et al., 1979). In all the catalysis experiments, mass balances obtained from analyses of the unconverted reactant and the products were good, typically 96%. The naphthalene products gave especially good mass balances, >98% in all experiments. We suggest that errors in the product analysis reflect the occurrence of undetected reactions giving products which (1) were lost in sampling due to their volatility or (2) were masked in the GLC analysis. Regarding the second possibility, it has been shown that hydroaromatic structures form stable adducts with the acceptor molecules (e.g., free radicals) formed during coal liquefaction (Cronauer et al., 1979). When experiments were carried out for relatively long times or at high temperatures (400 "C), n-hexadecane gave low-molecularweight products indicative of its cracking. The cracking reactions might proceed via free-radical intermediates, and these might have formed high-molecular-weight adducts with the hydroaromatics or aromatics, which could have been masked by the solvent peak. In the experiments carried out with benzene as the reactant, cyclohexane was the only primary product observed. The concentration-time profiles for the benzene reaction are shown in Figure 1. The reaction network inferred from the data is shown in Figure 2; the unidentified hydrocarbon indicated in the network was not separated in the GLC analysis; it is speculated to be methylcyclopentane. Since the reversibility of the benzene-
70
Ind. Eng. Chem. Process Des. Dev., Vol. 20, No. 1, 1981
f
2 5 k
8
V
0
0
5
15
IO TiME * r
20
Figure 1. Conversion of benzene and Hzin the presence of sulfided Co0-MoO3/r-Al2O3 in a batch reactor at 325 "C and 75 atm. The curves are the predictions of the reaction network of Figure 2.
0 0
-0
BENZENE
HYDROCARQON
m5py-m CYCLOHEXYLBENZENE
BlCYCiOHEXYL
OECALNS NAPHTHALENE
TETRALIN
&5ms /
2-PHENYLNAPHTHALENE
2-DHENYLTETRALIN
c)
a 0
0
5
IO
= 15
= 20
--9-HYDROCARE 25
0 NS
30
0
35
TIME, hr
Figure 3. Conversion of biphenyl and H2in the presence of sutfided CoO-MoO,/y-Al,O, in a batch reactor at 325 "C and 75 atm. The curves are the predictions of the reaction network of Figure 2. Table I. Results of Isothermal Gas Chromatographic Analyses with OV 101 as the Stationary Phasea
CYCLOHEXAVF
BIPHENYL
5
E-PHENYLTETRALIN
HYDROCARBONS
Figure 2. Reaction networks for hydrogenation of benzene, biphenyl, naphthalene, and 2-phenylnaphthalene in the presence of sulfided CoO-MoO3/r-AlzO3at 325 "C and 75 atm. Each reaction is approximated as first order in the organic reactant; the numbers next to the arrows are the pseudo-first-order rate constants in cubic meters per kilogram of catalyst per second.
cyclohexane reaction is so well documented (Weisser and Landa, 1973), no experiments with cyclohexane as a reactant were carried out. The most thorough set of catalytic reaction experiments was performed with biphenyl, which reacted with hydrogen to give cyclohexylbenzene as a primary product. Cyclohexylbenzene was subsequently converted into bicyclohexyl. The concentration-time plots for the biphenyl reaction network are shown in Figure 3; the data show clearly that bicyclohexyl is a secondary product, since the slope of the curve is zero at zero time. Three unidentified hydrocarbon products having nearly equal retention times in the GLC column were also observed in low concentrations (in total, about 13% of the product at the maximum conversion). Hydrogenation experiments carried out with cyclohexylbenzene and with bicyclohexyl as the reactants showed that these hydrocarbons were formed as primary products from cyclohexylbenzene, but not from bicyclohexyl; there was too little sample to allow identification of these products by combination of GLC/mass spectrometry, and so retention times were used for the identification. The retention times of a number of compounds considered to be plausible reaction products were represented using the Kovats retention index system (Kovats, 1965; Haken, 1976), and the results are collected in Table I. [In calculating these
component
Kovats retention index
cyclohexane benzene toluene ethylbenzene n-propylbenzene tert-butylbenzene sec-butylbenzene n-butylbenzene n-pentylbenzene n-hexylbenzene indan indene cis-decalin trans-decalin tetralin naphthalene bicy clohexyl cyclohexylbenzene biphenyl 1-methylnaphthalene 2-methylnaphthalene 1,2-dimethylnaphthalene 1,4-dimethylnaphthalene
644 650 756 848 935 97 5 993 1038 1130 1220 1005 1017 1108 1085 1158 1183 1295 1300 1360 1300 1285 1450 1430
In isothermal gas chromatography the retention index I of compound X is found by logarithmic interpolation between the values for two normal paraffins, as follows (Kovats, 1965)
where Vg(P,) < V g ( X ) < Vg(P,+,) and V g is the specific retention volume (determined by the GLC retention time), and P, is a normal paraffin C,H,,,,.
indices, the measured retention times were used in combination with literature data (Svob and Deur-Siftar, 1974; Sawatzky et al., 19761.1 To transform the indices from one stationary phase to another, the Kovats rules were used (Kovats, 1965). Two of the rules pertinent for our results are (1) the retention indices of nonpolar compounds are almost constant for any kind of stationary phase, and (2) the retention index of any compound is about the same for all nonpolar stationary phases. The retention indices for the three unknown products were 1260,1268, and 1277. Comparing these values with the data of Table I, we infer that the unknowns were neither products of ring opening (which would give substituted benzene or substituted cyclohexane) nor substituted indane, tetralin, or naphthalene (which indicates that they had not been formed by cracking or C-C bond
Ind. Eng. Chem. Process Des. Dev., Vol. 20, No. 1, 1981 71
3 0 0 -@ D
0
2
4
6
8
IO
TIME, hr
Figure 4. Conversion of naphthalene and H2 in the presence of sulfided Co0-Mo03/y-A1203in a batch reactor at 325 OC and 75 atm. The curves are the predictions of the reaction network of Figure 2.
cleavage followed by ring closure). A basis for identification of the products is provided by dibenzothiophene hydrodesulfurization experiments,giving 3-methylcyclopentylcyclohexane and 3,3’-dimethyldicyclopentyl as products with a catalyst and reaction conditions similar to those used here (Urimoto and Sakikawa, 1972). The observation of these products indicates the occurrence of isomerization reactions. The literature reports isomerization of cyclohexane to give methylcyclopentane, of tetralin to give methylindane, and of octahydrophenanthrene to give several methyl-substituted isomers during hydroprocessing (Hagenbach et al., 1971; Weisser and Landa, 1973; de Beer et al., 1976; Cronauer et al., 1978; Cronauer et al., 1979). These results lead to the suggestion that the three unknown products are isomerized hydrocarbons formed from cyclohexylbenzene, namely, cyclopentylbenzene with methyl substituents on the five-membered ring.
Further support for this suggestion is given by the distribution of these products; they appeared roughly in 3:31 molar proportions. Cyclohexylbenzenewould give only the three isomers shown above, and the formation of the third is expected to be slower as a result of steric hindrance. These three products together are represented simply as “hydrocarbons” in the reaction network shown in Figure 2. This network shows the reversibility of the initial hydrogenation reaction (l),and experiments with hydrogen
and cyclohexylbenzene as reactants confirmed that this reaction id reversible. When the reactants were naphthalene and hydrogen, 1,2,3,4-tetrahydronaphthalene(tetralin) was obtained as a primary product. Tetralin was further hydrogenated to give cis- and trans-decahydronaphthalene(decalin) (Figure 4). trans-Decalin was the predominant isomer of decalin. No other hydrocarbon products were detected. The data show that equilibrium between tetralin and naphthalene was closely approached. In experiments with hydrogen and tetralin as reactants, naphthalene was formed, confirming the reversibility of the primary hydrogenation reaction. Hydrogenation of 2-phenylnaphthalene gave two primary products, 2-phenyltetralin and 6-phenyltetralin, each of which experienced further hydrogenation (Figure 5). The secondary products, consisting chiefly of phenylsubstituted decalins, are represented together as “hydro-
TIME, hr
Figure 5. Conversion of 2-phenylnaphthaleneand H2 in the presence of sulfded CoO-Mo03/yA1203in a batch reactor at 325 ‘C and 75 atm. The curves are the predictions of the reaction network of Figure 2. Table 11. Pseudo-First Order Rate Constants for Hydrogenation Catalyzed by Sulfided CoO-MoO,/r-Al,O, at 325 “C and 7 5 atm
lo6 x
reactant benzene biphenyl naphthalene Z-phenylnaphthalene
concn, mol % 0.65 0.80 0.85 0.75
pseudo-first-order rate constant (m3/kgof cat. s)= 2.8 i 3.0 i 58.9 t 61.4 f
0.1 0.2 3.6 5.3
a Data are reported with 95% confidence limits and were determined by extrapolation to initial concentrations of reactants.
carbons” in this figure. Since 2-phenyltetralin and 6phenyltetralin were not commercially available, experiments were not performed with them as reactants to confirm the reversibility of the primary hydrogenation reactions. The experience with biphenyl and naphthalene, however, provides a firm basis for assuming that the primary hydrogenation reactions are reversible. The data of Figure 5 show that equilibrium between 2-phenylnaphthalene and its two primary products (2-phenyltetralin and 6-phenyltetralin) was closely approached. Kinetics. The reactions in the networks of Figure 2 are represented as pseudo first order in the hydrocarbon reactants. The pseudo-first-order rate constants for the overall disappearance of each component, determined from the low-conversion data, are summarized in Table 11. Since the primary hydrogenation reactions are reversible, a deviation from the pseudo-first-order behavior was observed as equilibrium was approached. Therefore, for the determination of rate constants to provide the best representation of the reaction networks over the whole ranges of conversion data, all the data were used, and it was assumed that each reaction in each network was pseudo first order in the organic reactant. The pseudo-first-order rate constants were estimated as follows. (1) Raw concentration vs. time data for benzene and biphenyl hydrogenation, but not the others, were smoothed by fitting spline functions using the IMSL subroutine ICSMOU (this program performs one-dimensional data smoothing by error detection and is highly effective in smoothing a data set only mildly contaminated with isolated errors). (2) The data (now smoothed for two of the networks) were used with equal weighting in the Carlton 2 program (Himmelblau et al., 1967) to determine best values of the individual rate constants. The values of the rate constants determined for the individual reactions in these networks
72
Ind. Eng. Chem. Process Des. Dev., Vol. 20, No. 1, 1981
Table 111. Equilibrium Constants for Aromatic Hydrogenation Evaluated from Kinetics Data reaction benzene + 3H, f cyclohexane biphenyl + 3H, Z cyclohexy lbenzene naphthalene + 2H, 2 tetralin 2-phenylnaphthalene + 2H, 2 2-phenyltetralin 2-phenylnaphthalene + 2H, t 6-phenyltetralin
KCa
340
+
180 L3/g-mo13
365
z
110 L3/g-mo13
145 i 80 L2/g-mo12 20 i 6 L2/g-mo12 400
t
1 2 0 L2/g-mo12
a For the reaction HC, + nH, 2 HC,. the concentration equilibrium constant is defined as K , =*'[CHC,/ CHC,'CH 1 equfi.
are shown in Figure 2, and the lines representing the predictions from these networks and the estimated pseudo-first-order rate constants are shown in Figures 1,3,4, and 5. It is clear from these figures that the data are well represented by these approximations. Equilibria. The maximum observed conversions of naphthalene and 2-phenylnaphthalene were about 99%, whereas the maximum observed conversions of biphenyl and benzene were only about 75%. The equilibrium constants for the five hydrogenation reactions at 325 "C are collected in Table 111. The values for biphenyl and naphthalene hydrogenation, K , = (CCHB/ CBpH-CH,3) = 365 f 110; K , = (CT/CN.CH:)esuil, = 145 f 80, agree airly well with results of gas-phase equilibrium experiments (Frye, 1962; Wilson et al., 1958) (modified to conform to the above definitions of the equilibrium constants), which establish a value of 485 for biphenyl and 90 for naphthalene. The large uncertainty in the K , values shown in Table I n reflects the large uncertainty in estimating the hydrogen concentration in the solution at high temperature and pressure. Low-temperature solubility data for hydrogen in n-hexadecane are given by Cukor and Prausnitz (1972); solubility is represented by a Henry's law constant at infinite dilution for a range of temperatures. To extend the solubilities to higher temperatures, the Jonah-King equation for the temperature coefficient of solubility was used (King et al., 1979). To extend the solubilities to high pressure, the Krichevsky-Kasamovsky equation was used (Prausnitz, 1969). Partial molar volumes needed in the above calculations were estimated from a correlation given by Lyckman et al. (1965). Fugacities for hydrogen were calculated from the virial equation of state (American Institute of Physics Handbook, 1963). The density of the reaction mixture at 325 OC and 75 atm was estimated by the Lee-Kesler correlation (Lee and Kesler, 1975), which required use of asymmetric mixing rules given by Plocker et al. (1978). There was one adjustable parameter, the acentric factor for n-hexadecane, which was adjusted such that low-temperature pure component densities of nhexadecane are predicted accurately by the above procedure. Discussion The kinetics results (Table I1 and Figure 2) show that benzene and biphenyl have nearly the same reactivities, and naphthalene and 2-phenylnaphthalene have nearly the same reactivities, one order of magnitude greater than that of benzene. 2-Phenylnaphthalene reacts almost exclusively in the naphthalene part of the molecule, as expected from this pattern. These results agree well with the results of high-pressure, high-temperature studies of the hydrogenation of benzene and naphthalene in the presence of
"F"'
unsupported MoS2and WS2catalysts, showing a 19-fold higher reactivity of naphthalene than of benzene (Weisser and Landa, 1973). We infer that supporting the sulfide catalyst on alumina results primarily in an increase in the number of catalytic sites but not in a significant change in the nature of the catalytic activity for aromatic hydrogenation reactions. The catalytic sites are widely believed to be surface anion vacancies, exposing Mo cations which can bond to the aromatic nucleus by interaction with its a electrons (Volter, 1964; Volter et al., 1968; Singhal and Espino, 1978 Maggiore et al., 1979; Kwart et al., 1980). The a-bonded intermediate may rearrange to give a a-bonded species, causing activation of the ring for hydrogen addition in a process resembling electrophilic aromatic substitution. These ideas provide a rough explanation of the differences in reactivity between benzene and naphthalene. The surface a complex involving naphthalene experiences greater resonance stabilization than that involving benzene. If this structure resembles the transition state, then the activation energy would be expected to be less for naphthalene than for benzene hydrogenation (Hogg, 1974). The results of Figure 2 show that phenyl substituents on benzene and naphthalene have negligible effects on the reactivities for hydrogenation. We speculate that small increases in reactivity resulting from the electron-donating influence of the aryl substituents might be roughly compensated by increased steric restriction of adsorption. The simplicity and apparent generality of the reaction networks elucidated in this work provide us with a basis for evaluating critically some recent work carried out with 1-methylnaphthalene as a reactant to model coal-liquid hydrogenation in a trickle-bed reactor (Patzer et al., 1979). Comparing ten different hydroprocessing catalysts, the authors observed that the ratio of concentrations of tetralin and naphthalene in the product was nearly constant (0.49 f 0.031, being unaffected by changes in catalyst composition, feed flow rate, and reaction temperature in the range 316-399 "C. Patzer et al. attributed this result to an equilibrium limitation and inferred the reaction network in eq 2. Their
interpretation of an equilibration of the products is surprising in view of the lack of dependence of the ratio of tetralin to naphthalene on temperature. Therefore, we are led to the following reinterpretation of their results. Since Patzer et al. presented too few data to allow meaningful estimation of all four rate constants, we adopt the one value they provided (for k,) and estimate the other parameters from our data obtained for naphthalene hydrogenation at approximately the same pressure (75 vs. 70 atm) and temperature (325 vs. 316 "C)and with a similar catalyst. The values (adjusted to the units used by Patzer et al.) are as follows: k1 = 1.6, k2 = 1.2, k2= 0.05, and k3 = 0.08 h-'. Figure 6 shows the relative concentration vs. (LHSV)-' profiles calculated for these values and for reaction network (2). For comparison, we have also plotted data from Table I1 of Patzer et al. Agreement between their data and the predicted curves is good. In the experiments reported by Patzer et al., (LHSV1-l was typically varied only from 0.3 to 0.6. The curves in Figure 6 show that, over a small range of (LHSV)-l less than 0.8, both naphthalene and tetralin concentration are almost linearly dependent on (LHSV)-', and the naph-
Ind. Eng. Chem. Process Des. Dev., Vol. 20, No. 1, 1981 73
activity for formation of constituents like tetralin. But it should not have excessive activity to over-hydrogenate and form poor hydrogen donors like decalin, the formation of which might be a further detriment because of the excessive consumption of costly hydrogen. Literature Cited
' LHSV ) - I ,
hr
Figure 6. Prediction o f relative concentration profiles f o r catalytic hydrogenation o f 1-methylnaphthalene a t 70 a t m and 316 O C . T h e curves are predicted f r o m t h e d a t a o f Patzer e t al. (1979) and t h e results for naphthalene hydrogenation (Figure 2).
thalene to tetralin ratio is consequently almost constant, giving the misimpression of an equilibrium distribution. Reported elsewhere (Sapre, 1980) is a sensitivity analysis concerning a wide range of cracking, hydrogenation, and dehydrogenation rate constants and showing that equilibrium between naphthalene and tetralin could not have been reached under the conditions employed by Patzer et al. We also infer that the ten catalysts used must have been about the same in their hydrogenation activities and that the activation energies for naphthalene hydrogenation and tetralin hydrogenation are approximately equal. These conclusions about product distributions in hydrogenation of aromatics are important for coal liquefaction processes, especially if the product liquid is recycled to the liquefaction reactor to provide a source of transferrable hydrogen. The reaction networks in Figure 2 show that the primary products of aromatic hydrogenation are hydroaromatics, which are much more effective hydrogen donors than aromatics or saturates (Orchin and Storch, 1948; Doyle, 1975). Liquefaction of coal has been shown to occur almost instantaneously upon its reaching a high temperature in the presence of a hydrogen-donor solvent (Neavel, 1976; Whitehurst and Mitchell, 1976). Thus the slow reaction in the hydrogenation/liquefaction process may be the hydrogenation of the recycle solvent to replenish the hydrogen-donor species (Guin et al., 1979). One of the objectives of the solvent hydrogenation is to optimize the yields of hydroaromatics. The reactions leading to consumption of hydroaromatics (e.g., further hydrogenation of tetralin to give decalins or isomerization of hydroaromatics) are undesirable, since they result in a solvent of reduced hydrogen-donor capacity. From Figure 2 it is evident that these undesirable side reactions are relatively slow in comparison with the primary hydrogenation reactions. Under the conditions of this study, the hydrogenationdehydrogenation equilibrium favored high yields of hydroaromatics. The data of Figures 4 and 5 show that within about 2 h,, the concentrations of tetralin, 2phenyltetralin, and 6-phenyltetralin passed through maxima, with yields of hydroaromatics exceeding 90 % . The results of kinetics studies of biphenyl hydrogenation (Sapre, 1980) indicate that at higher temperatures, dehydrogenation reactions become more important; the isomerization of hydroaromatics is also significant at higher temperatures (Hooper et al., 1979). These results suggest a strategy of maximizing yields of hydroaromatics by applying lower temperatures and high hydrogen partial pressures. the catalyst would also play an important role; the ideal catalyst would have moderate hydrogenation
"American Institute of Physics Handbook", Gay, D. E., Ed.; McGraw-Hili: New York, 1963. Ahuja, S. P.; Derrlen, M. L.; LePage. J. F. Ind. Eng. Chem. Prod. Res. Dev. 1970, 9 , 272. Bhinde, M. V. Ph.D. Thesis, University of Delaware, 1979. Cronauer, D. C.; Jeweii, D. M.; Shah, Y. T.; Keuser, K. A. Ind. Eng. Chem. Fundam. 1978, 17, 291. Cronauer, D. C.; Jeweii, D. M.; Shah, Y. T.; Modi, R. J.; Seshadri, K. S. Ind. Eng. Chem. Fundam. 1979, 18, 368. Cukor, P. M.; Prausnitz, J. M. J. Phys. Chem. 1972, 76, 598. Davies, L.; Lawson, G. J. Fuel, 1987, 48, 95. de Beer, V. H. J.; Dahimans, J. G. S.; Smeets, J. G. M. J. Cafal. 1978, 42, 467. Doyle, G. Prepr. Dlv. Pet. Chem. Am. Chem. SOC. 1975. 21, 165. Farcasiu, M. Fuel 1977, 56, 10. Ford Foundation Report, "Energy, The Next Twenty Years", Bailing Publishing Co., 1979. Frye, C. G. J. Chem. Eng. Data 1982, 7 , 592. Furlong, L. E.; Effron, E.; Vernon, L. W.; Wilson, E. L. Chem. Eng. Prcg. 1978, 72(8). 69. Gates, B. C.; Katzer, J. R.; Schuit, G. C. A. "Chemistry of Catalytic Processes", McGraw-Hili: New York, 1979; Chapter 5. Guin, J. A.; Tarrer, A. R.; Lee, J. M.; Lo, L.; Curtis, C. W. Ind. fng. Chem. Process Des. Dev. 1979, 18, 371. Hagenbach, G.; Courty, P. L.; Deimon, 8. J. Cafal. 1971, 23, 295. Haken, J. K. Adv. Chromatogr. 1978, 14. Harrison, E. K. Fuel 1988, 45,397. Himmeiblau, D. M.; Jones, C. R.: Bischoff, K. B. Ind. fna. Chem. Fundam. 1987, 8 , 539. Hogg, The English University .. D. R. "Electronic Effects arid their Applications", .. Press, London, 1974. Hooper, R. J.; Battaerd, H. A. J.; Evans, D. 0. Fuel.1979, 58, 132. Houalla, M.; Nag, N. K.; Sapre, A. V.; Broderick, D. H.; Gates, B. C. AIChE J. 1978. 24. 1015. Huang. &S.' M.S. Thesis, University of Mississippi, 1976. ICSMOU, computer subroutine, International Mathematical and Statistical Library, 1972. Jones, D. G.; Rottendorf, H.; Wilson, M. A,; Collin, P. J. Fuel 1980, 59, 19. King, M. B.; Ai-Nazzar, H.; Ail, J. K. Chem. Eng. Sci. 1979, 34, 1080. Kovats, E. Adv. Chromatcgr. 1985, 1 . Kwart, H.; Schuit, G. C. A.; Gates, B. C. J. Catal. 1980, 67, 128. Lee, B. 1.; Kesler, M. 0. AIChf J. 1975. 21, 510. Lyckman, E. W.; Eckert. C. A.; Prausnttz, J. M. Chem. Eng. Sci. 1985, 20, 685. Maggiore, R.; Glordano, N.; Crisafuiii, C.; Casteiii, F.; Solarino, L.; Bart, J. C. J. J. Catal. 1979. 50. 193. National Academy of Scknces, Assessment of Technology for Liquefaction of Coal, Washington, D.C., 1977. Neavel, R. C. Fuel 1978, 55,237. Nichoiis, B.; Whiting, M. J. Chem. SOC.1959, 551. Orchin, M.; Storch, H. H. Ind. Eng. Chem. 1948, 40(8). 1385. Patzer, J. F., 11; Ferrauto, R. J.; Montagna, A. A. Ind. Eng. Chem. Process Des. Dev. 1979, 18, 625. Piocker, U.; Knapp, H.; Prausnitz, J. M. Ind. Eng. Chem. Process Des. Dev. 1978, 17, 324. Prausnitz, J. M. "Molecular Thermodynamics of Fluid-Phase Equilibria". h e n tice-Hail: Engiewood Cliffs. N.J., 1969. Roiimann, L. D. J. Catal. 1977, 46, 243. Ross, D. S.; Blessing, J. E. Prepr. Am. Chem. SOC.Div. FuelChem. 1979, 24. 129. Ruberto, R. G.; Cronauer, D. C.; Jeweii, D. M.; Seshadri, K. S. Fuel1977, 56, 18. Sapre, A. V.; Broderick, D. H.; Fraenkei. D.; Gates, 6. C.; Nag, N. K. AIChE J., 1980, 24, 690. Sapre, A. V. Ph.D. Thesis, University of Delaware, 1980. Sawatzky, H.; Albert, E. G.; Smiley, G. T.; Montgomery, D. S. Fuel1978. 55, 329. Shih, S. S.; Katzer, J. R.; Kwart, H.; Stiles, A. B. Prepr. ACS Dlv. Pet. Chem. Am. Chem. SOC. 1977, 22, 919. Singhai, G. H., Espino, R. L. Prepr. ACS Div. Pet. Chem. Am. Chem. Soc. 1978, 23(1). 36. Svob, V.; Deur-Siftar, D. J. Chromatogr. 1974, 91. 677. Urimoto, H.; Sakikawa, N. Sekiyu Gakkaishi 1972, 15, 926. Voiter, J. J. Catal. 1984, 3 , 297. Voher, J.; Hermann, M.; Heise. K. J. Catal. 1888, 12, 307. Voorhoeve, R. J. H.; Stuiver, J. C. M. J. Catal. 1971, 23, 228. Weisser, 0.;Landa, S., "Sulphide Catalysts, Their Properties and Applications", Pergamon: New York, 1973. Whtehurst, D. D.; Mitchell, T. 0. Prepr. Am. Chem. SOC. Dlv. Fuel Chem. 1978, 21(5), 127. Wilson, T. P.; Caflisch. E. G.; Hurley, G. F. J . Phys. Chem. 1958, 62, 1059. Wiser, W. H. Prepr. Am. Chem. SOC. Dlv. Fuel Chem. 1975 20(2), 122. Receiued for r e v i e w M a r c h 6, 1980 A c c e p t e d August 4, 1980
