where k is the rate constant of the rate-determining step and KT, KH2. KH2S are the adsorption constants of thiophene, H2 and H2S, respectively. Figure 1 ...
INTRINSIC KINETICS OF THIOPHENE HDS OVER A NiMo/SiO2 MODEL CATALYST A. Borgna, E.J.M. Hensen, J.A.R. van Veen and J.W. Niemantsverdriet Schuit Institute of Catalysis, Eindhoven University of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands Introduction Supported mixed transition metal sulfide catalysts play a pivotal role in refineries for the production of clean motor fuels. They are employed not only to hydrotreat the final products like gasoline and diesel, but also to pretreat fluid catalytic cracking or reformer feed [1]. Furthermore, they provide the hydrogenation functionality in most hydrocracking catalysts that upgrade vacuum residue to more valuable products. Two major drivers for the development of more active hydrotreating catalysts are (i) dwindling oil supplies forcing refiners to use heavier feedstock and (ii) ever-tightening motor fuel specifications (for instance, the EU Auto Oil programme II [2]). Noteworthy is that mostly the use of an improved catalyst is more economic than modifications to the process. While research on hydrotreating catalysts has been extensive over the last 50 years [3-5], it has proven very difficult to relate catalyst structure to catalyst activity and selectivity on the molecular level. In recent years, we have employed the model catalyst approach to heterogeneous catalysis. The essential point is to prepare planar model catalysts by spin-coating which mimics the industrial pore volume impregnation method. In this way, realistic models can be obtained which are easily amenable to high-resolution spectroscopic studies. We have shown before [6,7] that the strong tandem-mix of characterization and reactivity evaluation enables detailed insight into the influence of active phase formation on the activity. Here, we will show how these well-defined models can be used to obtain intrinsic kinetics of thiophene hydrodesulfurization. Experimental Model catalyst preparation NiMo/SiO2 planar model catalysts consisting of a Si(100) wafer covered by a thin layer of SiO2 as support were prepared by spin-coating as extensively described before [6,7]. Planar SiO2 model supports were prepared by oxidizing a Si(100) single-crystal wafer in air at 1023 K for 24h. After calcination the wafers were cleaned in a H2O2/NH4OH solution at 338 K. Subsequently, the surface was hydroxylated in boiling water for 20-30 min. As a typical example, we describe here the preparation of NiMo/SiO2. A model silica support was spin-coated in nitrogen atmosphere at 2800 rpm with an aqueous solution containing Ni(NO3)2.6H2O (Merck, p.a.) and (NH4)6Mo7O24.4H2O (Merck, purity >97%) with an atomic ratio of 1:3, respectively. The concentration of the precursor solution was adjusted to result in the required metal loadings (2 Ni at/nm2 and 6 Mo at/nm2). Prior to characterization or catalytic testing, a model catalyst was sulfided at a rate of 5 K/min to 673 K + 1 h in a mixture of 10 vol.% H2S in H2. Reactivity evaluation. Atmospheric thiophene HDS reactions were carried out in a batch-type reactor. After sulfidation, the reactor was flushed with the reactant mixture for 5 min at the desired reaction temperature. The reaction was then carried out in batch mode by closing the reactor inlet and outlet. This was marked as zero reaction time. After a reaction time of 1 h a gas-phase sample was taken from the reactor using a precision sampling gas syringe which was injected on a DB-1 column to analyze the main products: 1butene, n-butane, trans-2-butene, cis-2-butene, thiophene and small
amounts of C1-C3 hydrocarbons. The HDS activities are expressed as thiophene conversion after 1 h of reaction time and per 5 cm2 of model catalyst. These values were corrected for blank thiophene conversion measured using an empty reactor. Between two experiments the catalyst was resulfided at 673 K for 1 h. Results and Discussion Because diffusion limitations are absent, the use of planar model catalysts offers great advantages in kinetic studies. Catalytic data were obtained at different temperatures varying the partial pressures of hydrogen, thiophene and H2S, respectively. The kinetic data were analyzed considering both power-law and Langmuir-Hinshelwood (L-H) kinetics Between 573 and 673 K, positive orders in thiophene and hydrogen and moderately negative orders in H2S were obtained. Moreover, while the reaction order in thiophene significantly decreased as the reaction temperature decreases, the order in hydrogen was nearly constant and close to 1. The lower reaction order of thiophene as compared to the reaction order in hydrogen clearly indicates that thiophene adsorbs much stronger than hydrogen. The negative order in H2S indicates that the HDS reaction is inhibited by H2S, probably due to competition with thiophene for sulfur-defect sites. The trends in the reaction orders are in agreement with the literature [8,9]. Heats of adsorption and activation energies can be derived once a kinetic model has been adopted. Several mechanisms and associated kinetic models for the hydrodesulfurization of thiophene have been proposed in the literature [10-14]. Although the precise mechanism of thiophene HDS reaction is still under debate, we have chosen an often-used simplified reaction network consisting of elementary steps, in which thiophene exclusively adsorbs on sulfur vacancies and H2 adsorbs dissociatively on different sites to form butadiene (B) and H2S. Assuming that the surface reaction between adsorbed thiophene and H2 is the ratelimiting step and the H2S adsorption takes place on sulfur vacancies exclusively, in competition with thiophene adsorption, the following Langmuir-Hinshelwood rate expression can be obtained: k KT K H 2 pT pH 2 r = (1 + KT pT + K H S pH 2 S / pH 2 ) (1 + K 1H/ 2 p1H/ 2 ) 2 2
2
2
where k is the rate constant of the rate-determining step and KT, KH2 KH2S are the adsorption constants of thiophene, H2 and H2S, respectively. Figure 1 shows that the experimental data measured at constant thiophene and H2S partial pressures and varying the H2 partial pressure can be satisfactorily fitted using the L-H equation, giving an estimation of the adsorption equilibrium constants of hydrogen at different temperatures. The optimized values indicate that the adsorption equilibrium constant is statistically non-different from zero under our experimental conditions. Therefore, the hydrogen adsorption term in the denominator can be ignored. The optimized values for the adsorption equilibrium constants at 673 K of thiophene and H2S are 3.5 ± 0.7 bar-1 and 28.9 ± 1.8, respectively. The intrinsic kinetic parameters, the activation energy of the rate determining step and the heat of adsorptions, can be derived by fitting the kinetic data measured at different temperatures using non-linear multivariable fittings. Figure 2 shows the quality of this fitting and the intrinsic parameters are summarized in Table 1. The estimated value for the heat of adsorption of thiophene and H2S are –57.8 ± 10.9 kJ/mol and –117.1 ± 11.3 kJ/mol, respectively. A broad range of values for the heats of adsorption of both thiophene and H2S have been reported in the literature [11,12]. Theoretical values vary between -62 and -137 kJ/mol, whereas the values for the heat of adsorption of H2S vary between –75 and –93 kJ/mol [15].
Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 2003, 48(2), 603
Reaction Rate (mol
thioph.
Ni
/mol *S)
Table 1: Intrinsic kinetic parameters obtained from non-linear multivariable fittings on a sulfided NiMo/SiO2 model catalyst.
0 .0 8 673 K
0 .0 6
623 K
0 .0 4
0 .0 2
r = k ' .P H 2 / ( 1 + K H 2
0 .6
0 .5
.P H 2
0 .8
0 .5
)
2
rds Eact (kJ/mol)
83.5
0 Kthioph (bar-1)
8.7
H S
− ∆H ads2
0 .1 0
673 K
138.9 -117.1
623 K
598 K
0 .0 0 0 .0 0 0 .0 2 0 .0 4 0 .0 6 0 .0 8
p th io p h ( % ) Figure 2. Dependence of the rate of thiophene HDS on the thiophene partial pressure along with a non-linear multi-variable fitting according to the L-H model. PH2= 90 vol% and PH2S= 0 vol%. To summarize, the values of the heat of adsorption reported in this contribution indicate a relatively strong adsorption of both thiophene and H2S. The activation energy is about 85 kJ/mol, in line with theoretical calculations [15], which predict overall reaction enthalpies for C-S scission and sulfur removal of 70 kJ/mol and 73 kJ/mol, respectively. It is important to mention that much lower activation energies, usually obtained from Arrhenius plots, have been reported in the literature [16-18]. However, the activation energy derived from Arrhenius plots represents the apparent activation energy and not the real activation energy of the rate-determining step. Moreover, a decrease in apparent activation energy with increasing temperature has been frequently observed [16-18]. Such a decrease of the apparent activation energy has been often attributed to a decrease of the surface coverage of the reactants as temperature increases [17, 18]. The relation between apparent and real activation energy is: H
H
app rds T Eact = Eact + ∆H ads + ∆H ads2 Therefore, a negative apparent activation energy may be obtained when the adsorption energies of thiophene and hydrogen are large enough. This then explains the Volcano-type curve recently reported by Borgna et al. [18].
Conclusion Planar model catalysts can be applied to obtain intrinsic kinetic parameters for the relatively simple thiophene hydrodesulfurization reaction, offering an excellent opportunity to study reaction mechanism and kinetics without under diffusion-free conditions.
648 K
0 .0 4 0 .0 2
(kJ/mol)
-57.8
At high temperatures, the coverages become lower and in the limit of an empty surface, we find
0 .0 8 0 .0 6
(kJ/mol)
1 .0
Figure 1. Dependence of the rate of thiophene HDS on the partial pressure of H2 at different temperatures, along with data fit according to the L-H kinetic model. Pthioph.= 4 vol% and PH2S= 0 vol%.
Ni
0.0967
0 KH (bar-1) 2S
2
/mol *S)
k 0 (molthioph/molNi*S)
thioph
P H (% )
Thioph
Value
− ∆H ads
0 .4
Reaction Rate (mol
Parameter
H S
app rds T Eact = Eact + (1 − ΘT# ) ⋅ ∆H ads + (1 − Θ*H ) ⋅ ∆H ads2 − Θ #H S .∆H ads2 2
References 1. Gosselink, W., CatTech, 1998, 4, 127. 2. Auto-Oil II Programme, Com. EU, Com 626, 2000. 3. Prins, R., De Beer, V.H.J., Somorjai, G.A., Catal. Rev. Sci. Eng., 1989, 31, 1. 4. Topsøe, H., Clausen, B.S. Massoth, F.E., Hydrotreating catalysis, Springer, Berlin, 1996. 5. Eijsbouts, S., Appl. Catal. A, 1997, 158, 53. 6. Kishan, G, Coulier, L., de Beer, V.H.J, van Veen, J.A.R., Niemantsverdriet, J.W, J. Catal., 2000, 196, 180. 7. Coulier, L., de Beer, V.H.J, van Veen, J.A.R., Niemantsverdriet, J.W, J. Catal., 2001, 197, 26. 8. Hensen, E.J.M., Brans, H.J.A., Lardinois, G.M.H.J., de Beer, V.H.J., van Veen J.A.R., van Santen, R.A., J. Catal., 2000, 192, 98. 9. Leliveld, R.G., van Dillen, A.J., Geus J.W, Koningsberger, D.C., J. Catal., 1998, 175, 108. 10. Massoth F.E., Chung, K.S., Proc. 7th ICC, Tokyo, 1980, p. 629. 11. Satterfield, C.N., Roberts, G.W., AIChE Journal, 1968, 14, 159. 12. H.C. Lee, Butt, J.B., J. Catal., 1977, 49, 320. 13. Vrinat, M.L., Appl. Catal., 1983, 6, 137. 14. Girgis, M.J., Gates, B.C., Ind. Eng. Chem. Res., 1991, 30, 2021. 15. Neurock, M., Van Santen, R.A., J. Am. Chem. Soc., 1994, 116, 4427. 16. Leglise, J., van Gestel J., Duchet, J.C., Symp. Adv. Hydrotreating Catalysts, ACS, Washington, 533, 1994. 17. Hensen, E.J.M., Vissenberg, M.J., de Beer, V.H.J., van Veen J.A.R., Van Santen, R.A., J. Catal., 1996, 163, 429. 18. Borgna, A., Hensen, E.J.M., Coulier, L., de Croon, M.H.J.M., Schouten, J.C., van Veen, J.A.R. Niemantsverdriet, J.W., Submitted to Cat. Lett. (2003).
Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 2003, 48(2), 604
