2344
Energy & Fuels 2006, 20, 2344-2349
Effect of Aromatics on Deep Hydrodesulfurization of Dibenzothiophene and 4,6-Dimethyldibenzothiophene over NiMo/ Al2O3 Catalyst Tao Song,† Zisheng Zhang,*,‡ Jinwen Chen,§ Zbigniew Ring,§ Hong Yang,§ and Ying Zheng† Department of Chemical Engineering, UniVersity of New Brunswick, Fredericton, Canada E3B 5A3, Department of Chemical Engineering, UniVersity of Ottawa, Ottawa, Canada K1N 6N5, and National Centre for Upgrading Technology, 1 Oil Patch DriVe, DeVon, Canada T9G 1A8 ReceiVed May 2, 2006. ReVised Manuscript ReceiVed July 18, 2006
The influence of different aromatics on the deep hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) over a commercial NiMo/γ-Al2O3 hydrotreating catalyst was investigated in a fixed-bed multiphase microreactor under designed conditions. Both the sulfur-containing compounds and the aromatic compounds in feeds and hydrotreated products were identified and quantified by GC-AED. Catalyst deactivation was observed, and a simple model for it was established. The kinetic behavior of these model compounds was studied with an assumption of pseudo-first-order reaction kinetics. The rate constants and the corresponding activation energies were determined, and the heat of adsorption for each compound on the catalyst surface was computed by density functional theory using the Material Studio software. The HDS rate for 4,6-DMDBT was much lower than that for DBT, which is attributed to the steric hindrance of the methyl groups at the 4 and 6 positions. The apparent activation energy of 4,6-DMDBT was higher than that of DBT under the same conditions. TThe HDS reaction rate significantly decreased with an increase of the content of aromatics in the feed. Aromatics with 2 or more rings were found to have stronger retardant effect on HDS than monoaromatics. This adverse effect was more pronounced for 4,6-DMDBT than for DBT. The competitive adsorption between the sulfur compounds and aromatics on the catalyst surface was the main reason for the decreased HDS efficiency as qualitatively verified by the heat of adsorption data.
1. Introduction Deep hydrodesulfurization (HDS) of diesel fuel has received great attention in recent years because of the worldwide increasingly stringent regulations on diesel fuel. Since 1997, the maximum sulfur content of diesel fuel has been limited to 500 wppm in most developed countries. New specifications have been issued or proposed by the U.S. Environmental Protection Agency (EPA) and European Commission (EC) to further reduce the sulfur content to the so-called ultralow-sulfur diesel (ULSD) level of 10 to 15 wppm and is expected to take effect by 20052006.1,2 To meet these more stringent specifications, highly efficient removal of sulfur compounds in diesel fuel is therefore of vital importance and necessity. It is well demonstrated that the sulfur compounds remaining in diesel fuel at a sulfur level lower than 500 ppm are predominantly alkyl-substituted dibenzothiophenes, in particular β-substituted alkyldibenzothiophenes: 4-methyldibenzothiophene * To whom correspondence should be addressed. Phone: 1-613-5625800 ext. 2047. E-mail:
[email protected]. † University of New Brunswick. ‡ University of Ottawa. § National Centre for Upgrading Technology. (1) Costa, P. D.; Potvin, C; Manoli, J.-M.; Lemberton, G.; Pe´rot, G.; Dje´ga-Mariadassou, G. New catalysts for deep hydrotreatment of diesel fuel: kinetics of 4,6-dimethyldibenzothiophene hydrodesulfurization over alumina-supported molybdenum carbide. J. Mol. Catal. A: Chem. 2002, 184, 323-333. (2) Song, C.; Ma, X. New design approaches to ultra-clean diesel fuels by deep desulfurization and deep dearomatization. Appl. Catal., B 2003, 41, 207-238.
(4-MDBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT). The latter is the most nonreactive compound toward HDS, which is the key point to achieve deep desulfurization.3-6 These species are termed refractory sulfur compounds. Both steric hindrance and electronic factors have been proposed to be responsible for the observed low-HDS reactivity.7,8 Increasing the severity of processing conditions (e.g., higher hydrogen pressure, lower space velocity, or higher reaction temperature) could be of some help for the removal of these refractory sulfurs. However, there are economic and thermodynamic limitations to the application of these options. As known widely, diesel fuel and its blend-stocks, termed as middle distillates, contain various components, such as nitrogen (3) Gates, B. C.; Topsoe, H. Reactivities in deep catalytic hydrodesulfurization: challenges, opportunities, and the importance of 4-methyldibenzothiophene and 4,6-dimethyldibenzothiophene. Polyhedron 1997, 16, 3213-3217. (4) Kabe, T.; Ishihara, A.; Tajima, H. Hydrodesulfurization of sulfurcontaining polyaromatic compounds in light cycle oil. Ind. Eng. Chem. Res. 1992, 31, 1577-1580. (5) Ma, X.; Sakanishi, K.; Mochida, I. Hydrodesulfurization reactivities of various sulfur compounds in diesel fuel. Ind. Eng. Chem. Res. 1994, 33, 218-222. (6) Ma, X.; Sakanishi, K.; Mochida, I. Hydrodesulfurization reactivities of various sulfur compounds in a vacuum gas oil. Ind. Eng. Chem. Res. 1996, 35, 2478-2497. (7) Daage, M.; Chianelli, R. R. Structure-function relations in molybdenum sulfide catalysts: The rim-edge model. J. Catal. 1994, 149, 414427. (8) Ma, X.; Sakanishi, K.; Isoda, T.; Mochida, I. Quantum chemical calculation on the desulfurization reactivities of heterocyclic sulfur compounds. Energy Fuels 1995, 9, 33-37.
10.1021/ef060199m CCC: $33.50 © 2006 American Chemical Society Published on Web 08/30/2006
Deep HDS of DBT and 4,6-DMDBT
compounds and aromatic species, which may affect the rate, selectivity, or both of the HDS of refractory sulfur compounds. The influence of these components on HDS catalyst activity becomes more significant under deep hydrodesulfurization conditions. For the aromatic-rich diesels, the inhibitory effect of those coexisting aromatics on deep HDS could be considerably large. Therefore, it is of great interest to have a better knowledge of the interactions between refractory sulfur compounds and different aromatic species during the deep HDS process and to elucidate the retarding mechanisms of aromatics on deep HDS. Moreover, a good understanding of these interactions and mechanisms will be one of the foundations for the development of high-efficiency HDS catalysts and processes. So far, many HDS studies were carried out with individual refractory sulfur species in inert solvents; the detailed catalytic mechanism and reaction networks of some representative sulfur compounds, such as benzothiophene, dibenzothiophene (DBT), and 4,6-DMDBT, have been reported by several research groups.9-14 There are also some reports on the solvent effect on the HDS of benzothiophene and dibenzothiophene, but at the deep HDS regime, the key sulfur compounds are the most refractory β-substituted alkyldibenzothiophenes, not these easyto-remove ones. Less information has been published on the inhibiting effect of aromatics on HDS of refractory organosulfur compounds. This led us to devote some efforts to the investigation of the effects of aromatics on deep HDS. In this work, simulated feeds, prepared with solvent and selected model sulfur and aromatic compounds at different aromatic concentrations, were hydrotreated through a microreactor under designed conditions. The adsorption heat of the model compounds and possible hydrotreated intermediates on the surface of catalyst were computed using the Dmol3 package of the Material Studio software (Accelrys Inc., San Diego, CA). The HDS kinetics of DBT and 4,6-DMDBT and the retarding effects of the aromatics on HDS were also obtained.
Energy & Fuels, Vol. 20, No. 6, 2006 2345 Table 1. Aromatic Mixture Composition Matrix aromatics
content (wt %)
mesitylene tetralin 2-methylnaphthalene fluorene phenanthrene pyrene
17.32 10.80 47.03 7.36 13.35 4.14
Table 2. Model Feeds Composition feed
n-hexadecane (%)
aromatics (wt %)
SDBT (wppm)
SDMDBT (wppm)
total sulfur (wppm)
base FA FB FC FD
99.84 94.82 89.84 80.02 59.83
0.00 5.02 10.00 19.82 40.01
165.5 166.3 166.2 167.8 169.1
88.1 89.1 88.6 92.5 90.6
253.7 255.4 254.8 260.3 259.7
Table 3. Experimental Conditions
pressure (bar) temp (°C) gas/oil ratio (NL/L) LHSV (h-1)
testing conditions
base condition
70 180-280 1000 1.11-3.79
70 200 1000 1.58
different aromatic concentrations were prepared as listed in Table 2. The gas used in this work was hydrogen with 99+% purity. The catalyst used was a crashed, equilibrated NiMo/γ-Al2O3 commercial hydrotreating catalyst (2.6 wt % Ni, 14.3 wt % Mo) with particle size in the range of 0.15-0.20 mm. 2.2. Apparatus and Procedure. The HDS experiments were carried out in a pilot-scale fixed-bed reactor system as shown in Figure 1. The reactor was a stainless steel tubular one (30.5 × 0.635
2. Experimental Section 2.1. Materials. The compositions of middle distillates and light cycle oils were used as the references in selecting the model compounds in this work. The relative levels of the mono-, di-, tri-, and tetraaromatics in the simulated feeds were kept the same as those in a LCO sample, determined in a GC/MS analysis. The specific model sulfur and aromatic compounds are DBT (99%), 4,6-DMDBT (97%), mesitylene (98%), tetralin (99%), 2-methylnaphthalene (97%), fluorene (98%), phenanthrene (98%), and pyrene (98%). They were all purchased from Sigma-Aldrich and used as supplied. N-hexadecane (99+ %) was obtained from Alfa Aesar and used as solvent in this work. The composition matrix of the aromatic mixture is given in Table 1. The simulated feeds with (9) Bataille, F.; Lemberton, J.-L.; Michaud, P.; Pe´rot, G.; Vrinat, M.; Lemaire, M.; Schulz, E.; Breysse, M.; Kaszetelan,S. Alkyldibenzothiophenes hydrodesulfurization-promoter effect, reactivity, and reaction mechanism. J. Catal. 2000, 191, 409-422. (10) Broderick, D. H.; Gates, B. C. Hydrogenolysis and hydrogenation of dibenzothiophene catalyzed by sulfided CoO-MoO3/γ-Al2O3: The reaction network. AIChE J. 1981, 27, 663-673. (11) Edvinsson, R.; Irandoust, S. Hydrodesulfurization of dibenzothiophene in a monolithic catalyst reactor. Ind. Eng. Chem. Res. 1993, 32, 391-395. (12) Ishihara, A.; Itoh, T.; Hino,T.; Nomura, M.; Qi, P. Y.; Kabe, T. Effects of solvents on deep hydrodesulfurization of benzothiophene and dibenzothiophene. J. Catalysis 1993, 140, 184-189. (13) Kabe, T.; Ishihara, A.; Zhang, Q. Deep desulfurization of light oil. Part 2: Hydrodesulfurization of dibenzothiophene, 4-methyldibenzothiophene and 4,6-dimethyldibenzothiophene. Appl. Catal., A 1993, 97, L1-L9. (14) Kabe, T.; Ishihara, A.; Nomura, M.; Itoh, T.; Qi, P. Effects of solvents in deep desulfurization of benzothiophene and dibenzothiophene, Chem. Lett. 1991, 2233-2236.
Figure 1. Flow scheme of pilot plant.
cm), operated in the continuous up-flow mode and heated by a three-zone electric furnace. The up-flow mode was adopted for catalyst wetting and liquid axial dispersion considerations. The reactor was equipped with an axial thermowell which houses one moveable thermocouple. The temperature differences in the packed bed were controlled to be within 2 °C. The reaction conditions were, therefore, considered isothermal. Catalyst was loaded into the reactor without dilution, giving a catalyst bed 0.20 m in length and 6 mL in volume. The preheating and postheating zones in the reactor were filled with 20-48 mesh glass beads. The reaction temperature was obtained by reading of the thermocouple placed in the middle of the catalyst bed. The catalyst was sulfided in situ with 3% H2S in H2 to allow complete reduction. After reduction, the catalyst was cooled to 200 °C, and the liquid and gaseous feeds were then introduced into the system. The experimental conditions employed in this work are summarized in Table 3. To investigate the deactivation of the catalyst, check-back runs were conducted regularly using the feed “base” as listed in Table 2 under a standard operating condition which is specified in Table 3. For each experiment, an on-stream time of at
2346 Energy & Fuels, Vol. 20, No. 6, 2006
Song et al.
Figure 2. Conversion of sulfur compounds in base feed under base conditions.
Figure 3. Declining trend of rate constants in check-back experiments: DBT ([), 4,6-DMDBT (9), base conditions.
least 20 h was allowed to reach a steady state, and then samples were taken from the product receiver for subsequent analysis. An on-line gas chromatography (GC) system was used to monitor the outlet gas composition for use in mass balance calculations. 2.3. Analysis. Product samples were collected after each experimental run had reached steady-state conditions. A HP6890 series gas chromatograph, equipped with an atomic emission detector (AED, sulfur channel 181 nm) and a Rtx-50 (60 m) from RESTEK, with an i.d. of 0.25 mm, capillary column (the stationary phase of this column was cross-linked 50% phenyl-50% methyl polysiloxane coated to 0.25 µm in thickness), was used to analyze individual sulfur species in the feed and hydrotreated products. The temperature for the GC analysis was initially set at 40 °C and programmed to increase at a rate of 2 °C/min to 285 °C, where the temperature was held for 20 min. The total analysis time was about 142 min for each sample. Partial identification was achieved by comparing the retention times of the sulfur species with those of standard compounds collected in a data bank and by using the GCMS spectra (HP5890-series II with 5972-series MS detector) obtained separately. The quantification of the sulfur in the samples was performed by the external-standard method in which a standard solution was injected under the same conditions as those used for the unknown sample. A response factor (RF) was obtained by dividing the sulfur content of the standard solution with the peak area (RF ) sulfur concentration/peak area). The standard solutions were a series of benzothiophene (BT) solutions with sulfur concentrations in the range of 30-200 ppm.
Obviously, the NiMo/Al2O3 catalyst employed here deactivated sharply during the first 1000 h of use: its activity became relatively stable afterward. Compared to the conversion of DBT, the impact of catalyst deactivation is much greater for 4,6DMDBT. The different deactivation behaviors for DBT and 4,6DMDBT revealed that the main HDS routes of these two compounds on this catalyst were different. To eliminate the effect of catalyst deactivation so that the effect of different aromatics (i.e., different feeds in Table 2) on HDS can be examined under the operating conditions of interests (as listed in Table 3), a simple catalyst deactivation model was established, and all experiment results were normalized according to this model. The check-back experiment at 455 h was set as the standard point, and pseudo-first-order reaction rate constants of DBT and 4,6-DMDBT of this point, kst, were calculated and considered as 1. The rate constants of other experiments, k, were calculated and compared to kst. Figure 3 illustrates the declining trend of the pseudo-firstorder rate constants of DBT and 4,6-DMDBT before the activity reached a steady state. The simple deactivation model based on DBT and 4,6-DMDBT and their correlation coefficients is given below
yDBT ) 22.269t-0.5083 R2 ) 0.9946
(1)
y4,6-DMDBT ) 429.03t-0.9811 R2 ) 0.9762
(2)
3. Results and Discussions 3.1. Catalyst Deactivation. During hydrotreatment, catalyst deactivation with time is invariably observed. There are generally several possible reasons for catalyst deactivation, including, loss of active phase dispersion because of sintering under high temperatures, poisoning of active sites by certain reactants and products, and pore-blocking by certain reaction products such as metal, coke, and polymers.15 In this work, because of the relatively low reaction temperatures used and the shortage of impurities that may result in catalyst poisoning in the feeds, coke formation was probably the main reason for the deactivation, although no further examination of the spent catalyst was performed. To account for the effect of catalyst deactivation on the experimental results with different feeds and under different operating conditions, check-back runs were conducted regularly using the “base” feed as listed in Table 2 under the standard operating condition as specified in Table 3. Figure 2 shows the conversion of DBT and 4,6-DMDBT in these experiments. (15) Topsøe, H.; Clausen, B. S.; Massoth, F. E. Hydrotreating Catalysiss science and Technology; Springer-Verlag: Berlin, 1996.
where y is the relative reactivity (y ) k/kst) and t is the onstream time. The data reported thereafter are all calculated and normalized using eqs 1 and 2. 3.2. Kinetics of Deep HDS. It is widely agreed that the HDS of dibenzothiophens over catalysts goes through two parallel reaction routes, namely, (1) direct HDS by hydrogenolysis of the reactants (DDS), yielding biphenyl compounds, and (2) indirect HDS by hydrogenation of one of the two aromatic rings, followed by hydrogenolysis of the hydrogenated intermediate products (HYD), first yielding tetrahydrodibenzothiophenes and then cyclohexybenzenes.16 Detailed reaction pathways and networks have been extensively reviewed by Whitehurst et al.17 For the HDS of 4,6-DMDBT, the HYD pathway is more favored because of the steric hindrance of substituted methyl groups. (16) Michaud, P.; Lemberton, J. L.; Pe´rot, G. Hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene: Effect of an acid component on the activity of a sulfided NiMo on alumina catalyst. Appl. Catal., A 1998, 169, 343-353. (17) Whitehurst, D. D.; Farag, H.; Nagamatsu, T.; Sakanishi, K.; Mochida, I. Assessment of limitations and potentials for improvement in deep desulfurization through detailed kinetic analysis of mechanistic pathways. Catal. Today 1998, 45, 299-305.
Deep HDS of DBT and 4,6-DMDBT
Energy & Fuels, Vol. 20, No. 6, 2006 2347 Table 5. Apparent Activation Energies [kJ/mol] of HDS of DBT and 4,6-DMDBTa source catalyst DBT 4,6-DMDBT a
Figure 4. ln(C0/C) vs 1/LHSV plots for DBT and 4,6-DMDBT: feed FB, temperature 240 °C, pressure 70 bar, H2/oil ) 1000 NL/L. Table 4. Pseudo-first-order Reaction Rate Constants of DBT and 4,6-DMDBTa
a
temp (K)
kDBT (h-1)
453 473 493 513 553
0.2782 0.8605 3.3108
kDMDBT (h-1) 0.0806 0.2879 1.0572 9.0843
Feed FB, pressure 70 bar, H2/oil ) 1000NL/L.
Figure 5. Arrhenius plots of DBT and 4,6-DMDBT: feed FB, pressure 70 bar, H2/oil ) 1000 NL/L, LHSV ) 1.58 h-1.
Plots of ln(C0/C) versus 1/LHSV for DBT and 4,6-DMDBT in feed “FB” at 240 °C are shown in Figure 4. C0 and C are the concentrations of sulfur compounds in the feed and product, respectively. Obviously, the disappearance of DBT and 4,6DMDBT follows a first-order kinetics with correlation coefficients (R2) higher than 0.95 for both. Therefore, the following kinetic equation could represent the HDS behavior of DBT and 4,6-DMDBT in this study
ri ) -kiCi
(3)
where i ) DBT or 4,6-DMDBT. The pseudo-first-order HDS rate constants of sulfur compounds, k, were determined from the slope of the straight lines in Figure 4. The values of k under different temperatures for feed FB were calculated, and the results are listed in Table 4. The rate of disappearance of DBT is almost 10-fold higher than that of 4,6-DMDBT under same temperatures, indicating that DBT is much easier to desulfurize than 4,6-DMDBT. Arrhenius plots, based on these pseudo-first-order rate constants, for DBT and 4,6-DMDBT are presented in Figure 5. The apparent activation energies calculated from the slopes of the straight
current work NiMo 115 129
ref 13 NiMo 100 167
ref 18 CoMo 108 130
Feed FB, pressure 70 bar, H2/oil ) 1000 NL/L, LHSV ) 1.58 h-1.
lines are shown in Table 5, where a comparison of our results with data from other research is presented. There are some discrepancies in the absolute values, but the trends in the apparent activation energies of the two sulfur compounds are fairly close to those reported in the literature. The differences in the apparent activation energies might be linked with the different catalysts and operating conditions employed. Model compound 4,6-DMDBT shows higher apparent reaction energies than DBT, which supports the observations made above. 3.3. Hydrogenation of Aromatics. During the hydrotreating processes in this work, the hydrogenation of the aromatic compounds present in the feeds would happen simultaneously with HDS and would unavoidably affect the HDS of the two model sulfur compounds. The concentrations of aromatics in each experiment were monitored as well to investigate their effects on HDS. For condensed aromatics, hydrogenation happens sequentially from one ring to the next. Usually, the higher the number of fused rings in aromatics, the more reactive the molecule is in the hydrogenation. The higher reactivity of these polyaromatics is speculated to be a consequence of the greater resonance stabilization of polyaromatic surface species formed by π-bonding of the polyaromatics moieties to active sites on the surface of catalyst. For example, naphthalene and substituted naphthalenes are an order of magnitude more active than benzene and substituted benzenes.19 The hydrogenations of aromatic hydrocarbons are reversible, and the conversion can therefore never reach 100%. However, because the reaction temperatures used in this study are relatively low, the reverse reaction, dehydrogenation, could be neglected, and the whole hydrogenation process can be considered as reactions in one direction. Meanwhile, as the concentration of each aromatic species is much higher than that of sulfur in feeds, the hydrogenation of aromatics can therefore be studied independently. Most reactions in the hydrogenation of aromatics are first-order. In this work, excellent fits of pseudofirst-order reaction kinetics were observed for the disappearance of all the model aromatic hydrocarbons except for mesitylene, which is so inactive that its concentration change was insignificant under the operating conditions examined. Figure 6 gives the plots of ln(C0/C) versus 1/LHSV for the rest of the 5 aromatic compounds in feed FB at 240 °C. The experimental data for the aromatic compounds were processed using a procedure similar to that of the organo-sulfur compounds. The ranges of the apparent activation energies of the six model aromatic hydrocarbons, determined from experimental data, are shown in Table 6. It can be seen that the differences of the apparent activation energies for different model aromatic species is not significant except for that of fluorene, which appears to be slightly higher. (18) Kabe, T.; Akamatsu, K.; Ishihara, A.; Otsuki, S.; Godo, M.; Zhang, Q.; Qian, W. Deep hydrodesulfurization of light gas oil. 1. kinetics and mechanisms of dibenzothiophene hydrodesulfurization, Ind. Eng. Chem. Res. 1997, 36, 5146-5152. (19) Sapre, A. V.; Gates, B. C. Hydrodeuslfurization of aromatic hydrocarbons catalyzed by sulfided CoO-MoO3/γ-Al2O3: The reaction networks, Ind. Eng. Chem. Prod. Res. DeV. 1981, 20, 68-73.
2348 Energy & Fuels, Vol. 20, No. 6, 2006
Figure 6. ln(C0/C) vs 1/LHSV plots for aromatic hydrocarbons: feed FB, 240 °C, 70 bar, H2/oil ) 1000 NL/L. Table 6. Apparent Activation Energies of Hydrogenation of Aromatic Hydrocarbonsa
a
aromatic
apparent activation energy (kJ/mol)
mesitylene tetralin 2-methylnaphthalene fluorene phenanthrene pyrene
73.4-80.9 93.4-113.6 81.5-97.7 118.3-141.6 90.4-108.0 91.6-107.6
70 bar, H2/oil ) 1000 NL/L, LHSV ) 1.58 h-1.
Figure 7. Assumed NiMo catalyst structure.
3.4. Adsorption Heats Calculation. With the development of modern computational chemistry, it is possible to use new computational approaches to aid in understanding and correlating the activities of sulfur compounds with their intrinsic properties and other components’ inhibition effects on HDS. In this study, the adsorption heat of all model compounds and possible hydrotreated intermediates on the surface of catalyst were computed using the Dmol3 package of the Material Studio software. There are many models proposed for the structure of promoted CoMo and NiMo catalysts.15 Among them, the “NiMoS-phase” model developed by Topsøe et al. has gained the greatest recognition, and it was therefore selected as the structure of catalyst in our computation. This model suggests that in alumina-supported Ni-Mo catalysts, the NiMoS phase is present as a single S-Mo-S sheet, with Ni present at the molybdenum sites. Ni-Mo-S structures are believed to be mainly responsible for the activity. A single-layer NixMo10-xS18 (for simplification, x ) 1 was chosen) nanocluster with the coordinatively unsaturated sites on the Ni-Mo edge was constructed as the catalyst model, as shown in Figure 7. All the calculations were performed with the DMol3 simulation module in the software. Individual model compounds, NixMo10-xS18 nanoclusters, and molecule-catalyst complexes were subject to geometry optimization before the adsorption
Song et al.
Figure 8. Conversions of sulfur compounds vs concentration of aromatics: DBT (×), 4,6-DMDBT (b), 240 °C, 70 bar, H2/oil ) 1000 NL/L. Table 7. Calculated Adsorption Heats compound
adsorption heat (kJ/mol)
mesitylene tetralin 2-methylnaphthalene fluorene phenanthrene pyrene DBT 4,6-DMDBT
-228.54 -208.00 -249.56 -278.29 -313.35 -313.09 -307.10 -279.84
heat calculation. The surface atoms of catalyst cluster and the adsorbed molecules were allowed to relax during the calculation, while the rest of the slab was held at the fixed bulk geometry. The total energy of the system was determined by a selfconsistent density functional theory calculation, and the exchangecorrelation energy was approximated by nonlocal generalized gradient corrections of PW91. All the electrons were included in the calculations with a double-numeric quality basis set. The real space cutoff of the atomic orbital was set at 4.0 Å, and a smearing of 0.005 Ha was used to count the orbital occupancy. The adsorption heat of a molecule on the catalyst cluster, ∆H, was given by the following equation:
∆H ) Em-cat - (Em + Ecat)
(4)
where Em-cat is the total energy of the molecule-catalyst complex, Em is the total energy of molecule, and Ecat is total energy of the catalyst cluster. The values of calculated adsorption heat are summarized in Table 7. It can be seen that the adsorption heat of aromatic compounds decreases in the following order: four ring > three ring > two ring > one ring. In particular, the adsorption heat of DBT is higher than that of 4,6-DMDBT. 3.5. Effects of aromatics on deep HDS. It is generally accepted that both the hydrogenation and sulfur extraction rates in HDS processes are reduced by the presence of other components in the feed or products. In this work, the effect of aromatics on deep HDS was the main objective. 3.5.1. Effect of Aromatic Concentration. To investigate the effect of aromatic concentration on deep HDS, the HDS reaction was carried out in the presence of different concentrations of total aromatic hydrocarbons. The conversions of two sulfur compounds with different aromatic concentrations at 240 °C were illustrated in Figure 8. It can be seen that the conversions of both DBT and 4,6-DMDBT decreased with the increase of total aromatic concentration. When the conversions at the two extreme conditions (e.g., no aromatic and 40% aromatic in feed) are compared, the conversion of DBT decreased from 100 to
Deep HDS of DBT and 4,6-DMDBT
Energy & Fuels, Vol. 20, No. 6, 2006 2349
Table 8. Rate Constants (k) of DBT and 4,6-DMDBTa
a
aromatic content (%)
DBT
4,6-DMDBT
0 5 10 20 40
3.5625 1.6801 0.8605 0.2768 0.1839
0.6226 0.1685 0.0806 0.0326 0.0222
200 °C, 70 bar, LHSV ) 1.58 h-1, H2/oil ) 1000 NL/L.
74%, while that of 4,6-DMDBT dropped sharply from 91 to 9%. Obviously, the retarding effect of aromatics on HDS of 4,6-DMDBT was more significant than that on HDS of DBT. Even at relatively low temperatures, the inhibition effect of aromatics on deep HDS can be observed. The pseudo-first-order rate constants of the two sulfur compounds were calculated using the procedure described in section 3.2, and some of the results are listed in Table 8. The ratios of the rate constants of DBT and 4,6-DMDBT with 40% aromatic concentration in the feed and without aromatics in the feed are 5.16 × 10-2 and 3.56 × 10-2, respectively. The reactivity decreased almost two orders of magnitude for both model compounds. 3.5.2. Effect of Different Aromatics. As mentioned above, the apparent reaction activation energies for model aromatic hydrocarbons are very close except that for fluorene in this study. However, their conversions during hydroprocessing showed greater differences. What caused this difference, and how do they inhibit the transforming of refractory sulfur species in HDS process? The higher conversion of the di- and polyaromatics than that of the monoaromatic indicates that condensed aromatic compounds are more active and easier hydrogenate, while the conversion of refractory sulfur compounds, as mentioned above, also goes mainly through the hydrogenation path. Thus, competitive adsorption between refractory organo-sulfur compounds and aromatics at the active catalytic sites is unavoidable. During the competitive adsorption, compounds with higher adsorption heats are easier adsorb and are likely to occupy more active sites. Calculation results show that adsorption heats of the two monoaromatic compounds are all lower than those of other condensed aromatics and sulfur compounds, while the adsorption heats of the condensed aromatic are higher or close to those of DBT and 4,6-DMDBT. Therefore, it is clear that the competitive adsorption between sulfur compounds and aromatics on the catalyst surface retards the HDS reaction, and aromatics with 2 or more rings have a stronger retardant effect on deep HDS than monoaromatics because of their higher adsorption capabilities. As shown in Table 7, the adsorption heat of DBT is higher than that of 4,6-DMDBT. The methyl substitutes in the 4 and
6 positions make the adsorption of 4,6-DMDBT on the catalyst surface more difficult than that of DBT, resulting in the lower adsorption heat. In competitive adsorption with aromatics, the sulfur compound with a lower heat of adsorption (i.e., the lower adsorption ability) is greatly inhibited. Thus, the retardant effect of aromatics on 4,6-DMDBT HDS is more significant than that on DBT in this study, which accounts for the sharp decrease of the conversion of 4,6-DMDBT. 4. Conclusions The aromatic compounds present in feeds do inhibit the conversion of sulfur species, DBT and 4,6-DMDBT, in deep HDS process. This adverse effect was more pronounced for 4,6DMDBT than for DBT, which might be attributed to steric hindrance from substituted groups on the 4 and 6 positions. The aromatic effect on HDS became significant with the increase of total aromatic concentration in feed, and condensed aromatic compounds have a stronger inhibitory effect on deep HDS than monoaromatics because they have higher adsorption abilities than other species; this was verified qualitatively by the adsorption heat data computed through a density functional theory program. Acknowledgment. The authors are grateful to the following individuals for their help and support: Mr. Dennis Carson and Mr. Tom Crothers for assistance in experiments and Ms. Yevgenia Briker for aromatic and sulfur analyses.
Nomenclature C0 and C ) component concentration in the feed and product, respectively Ecat ) total energy of the catalyst cluster Em ) total energy of molecule Em-cat ) total energy of the molecule-catalyst complex ∆H ) adsorption heat k ) pseudo-first-order reaction rate constant kst ) pseudo-first-order reaction rate constant for check-back run at 455 h R2 ) correlation coefficients 4-MDBT ) 4-methyldibenzothiophene 4,6-DMDBT ) 4,6-dimethyldibenzothiophene DBT ) dibenzothiophene EC ) European Commission EPA ) U.S. Environmental Protection Agency GC-AED ) gas chromatography with an atomic emission detector HDS ) hydrodesulfurization LHSV ) liquid hour space velocity NCUT ) National Centre for Upgrading Technology ULSD ) ultralow-sulfur diesel EF060199M