by DE Resasco · 2000 · Cited by 47 — The analysis of the nature and energetics of adsorption of certain adsorbates may help determine whether electronic effects have important influences in catalysis.

18 KB – 52 Pages

PAGE – 2 ============
11. INTRODUCTION Catalytic dehydrogenation of alkanes is an endothermic reaction, which occurs with an increase in the number of moles and can be represented by the expression Alkane ! Olefin + Hydrogen This reaction cannot be carried out thermally because it is highly unfavorable compared to the cracking of the hydrocarbon, since the C-C bond strength (about 246 kJ/mol) is much lower than that of the C-H bond (about 363 kJ/mol). However, in the presence of a suitable catalyst, dehydrogenation can be carried out with minimal C-C bond rupture. The strong C-H bond is a closed-shell ! orbital that can be activated by oxide or metal catalysts. Oxides can activate the C-H bond via hydrogen abstraction because they can form O-H bonds, which can have strengths comparable to that of the C- H bond. By contrast, metals cannot accomplish the hydrogen abstraction because the M- H bonds are much weaker than the C-H bond. However, the sum of the M-H and M-C bond strengths can exceed the C-H bond strength, making the process thermodynamically possible. In this case, the reaction is thought to proceed via a three centered transition state, which can be described as a metal atom inserting into the C-H bond. The C-H bond bridges across the metal atom until it breaks, followed by the formation of the corresponding M-H and M-C bonds. 1 Therefore, dehydrogenation of alkanes can be carried out on oxides as well as on metal catalysts. In fact, both types of dehydrogenation catalysts are typically found in a number of important industrial applications. Catalytic dehydrogenation is employed in the production of propylene and isobutylene from propane and isobutane, in the production of C 6 to C 19 mono-olefins from the corresponding normal alkanes, and of styrene from ethylbenzene. In this

PAGE – 3 ============
2contribution, the chemistry involved in these reactions will be discussed, with particular emphasis on the effects of catalyst structure and composition on the catalytic properties. 2. THERMODYNAMIC EQUILIBRIUM LIMITATIONS The dehydrogenation reactions are thermodynamically favored at high temperatures and low pressures. The equilibrium conversion can be readily calculated by the expression: K p = x ( h + x ) P (1) (1 Œ x ) (1+ i + h + x ) where K p is the equilibrium constant, x the equilibrium conversion, i and h the number of moles of inerts and hydrogen, per mole of alkane in the feed, respectively. The equilibrium conversion for the dehydrogenation of propane as a function of temperature is illustrated in Fig. 1 for typical hydrogen/alkane ( h) and inert/alkane ( i) feed ratios. In some industrial operations, part of the hydrogen produced is recycled to inhibit the formation of coke on the catalyst. However, at a given temperature, the equilibrium conversion decreases with the hydrogen/alkane feed ratio. For example, under typical dehydrogenation conditions (550ºC, 1 atm) the equ ilibrium conversion decreases from about 44 % to 25 % when the H 2/alkane ratio increases from 0 to 3. Therefore, the hydrogen/alkane feed ratio is a compromise between coke suppression and conversion. Contrarily, when inerts, such as steam or nitrogen, are added to the feed the equilibrium is benefited. For example, if the inert/alkane ratio is increased from 0 to 3 at 550ºC and 1 atm, the equ ilibrium conversion increases from 44 % to 64 %.

PAGE – 4 ============
33. CATALYTIC DEHYDROGENATION OF ISOBUTANE AND PROPANE The catalytic dehydrogenation of isobutane and propane are important reactions used commercially for the production of isobutylene and propylene, respectively. One of the major applications of isobutylene is as a feedstock in the manufacture of methyl tertiary butyl ether (MTBE). In the United States, MTBE has been used in relatively low concentrations as an octane booster in gasoline for more than 25 years. However, since 1992, it has been used in high concentrations in a large number of cities, to meet requirements of the Clean Air Act in the oxygenated and reformulated gasolines. In those cases, up to about 15 % MTBE is added to gasoline to allow more complete combustion and reduce emissions of carbon monoxide (CO) and volatile organic compounds (VOC). In recent years, objections to potential health effects of MTBE caused by water contamination have arisen, which may have an impact on the MTBE demand. 2 However, demand for polyisobutylene, and polybutenes in general, is in continuous growth. Thus, the needs for C 4 olefin production will probably remain high. C4 olefins are mainly produced from FCC (about 50 %) and steam cracking (21 %). However, in some industrial operations, it is necessary to have increased flexibility in the olefin supply and so, production of a single specific alkene is sometimes required. In those cases, the direct catalytic dehydrogenation is the ideal solution. Propylene is an important basic chemical building block for plastics and resins. Its worldwide demand has steadily grown for the last 15 years and it is projected that, in the coming years, demand growth for propylene will be equal to or even higher than that for ethylene. 3 Similar to isobutylene, propylene can be produced as a by-product from FCC and steam

PAGE – 5 ============
4cracking operations. However, in response to the growing demand and for the greater flexibility in the olefin pool that a dedicated process has in comparison to cracking units, several propane dehydrogenation plants are now being operated in the world. The dehydrogenation of lower alkanes is typically carried out on two different types of catalysts: a) Pt-based catalysts and b) chromia-based catalysts. 4-6 The main characteristics of these two types of catalysts will be discussed here, together with some reference to other less common materials. Several commercial processes for dehydrogenation of lower alkanes are currently available. The severely deactivating conditions imposed by the dehydrogenation process have challenged the process designers to develop efficient reactors for this difficult task. Several options have been tested and several have found successful applications. Among the various commercial processes available for dehydrogenation of propane and isobutane, one can find fixed bed reactors, operated in isothermal or adiabatic form. Some include cyclic operations, others include continuous catalyst regeneration with moving beds, or with fluidized beds. 3.1. Pt-based dehydrogenation catalysts: A side reaction that frequently competes with dehydrogenation is hydrogenolysis. Platinum is a primary component in many dehydrogenation catalysts due to its high activity for activating C-H bonds, coupled with an inferior activity for the rupture of C-C bonds, resulting in intrinsically high selectivities toward dehydrogenation. On a Pt surface, only low-coordination number sites (steps, kinks) are able to catalyze the C-C bond breaking, while essentially all Pt sites catalyze the rupture of the C-H bonds.

PAGE – 6 ============
5Another undesired side reaction that competes with dehydrogenation is coke formation. Since both, hydrogenolysis and coke formation, are more sensitive to the support structure than dehydrogenation, any impurity or inactive species on the surface may act as a site diluent and should increase the selectivity toward dehydrogenation. This effect is well documented and has been observed in a number of bimetallic systems. 7, 8 In addition to the dilution of sites, other more subtle factors may play a role in altering the activity of the Pt surface. For example, significant differences in the magnitude of these effects have been detected in a series of Pt catalysts promoted by the addition of different metals, which are themselves inactive, such as Sn, In, Pb, and Cu, but can alter the properties of Pt. 9 It was found that the promoting effect among these metals was highest from the addition of Sn and lowest from Cu. In fact, Sn has been the preferred promoter for most Pt dehydrogenation catalysts. 10, 11 In section 3.1.2, we discuss the effects of Sn in greater detail. 3.1.1. Reaction Mechanism and Kinetics on Pt Catalysts The isobutane dehydrogenation on Pt-based catalysts has generally been described by the following set of reaction steps: (A) C 4H10 + 2 * ” C 4H9* + H * (B) C 4H9* + 2 * ” C 4H8** + H * (C) C 4H8** ” C 4H8 + 2 * (D) H 2 + 2 * ” 2 H* Cortright et al. 12 have conducted deuterium tracing studies under reaction conditions, which showed a much larger extent of deuterium in the isobutylene product

PAGE – 8 ============
7temperatures, concentrations, and catalysts. 14 All the fitting parameters thus obtained had physical meaning and were used to compare thermodynamic functions on different catalysts. This kinetic expression clearly explains the reaction orders obtained on most previous studies, i.e., first order dependence with respect to the alkane and negative half to zeroth order dependence in hydrogen. 9, 13 3.1.2. Addition of Sn to Pt as a promoter for activity, selectivity, and catalyst life The addition of Sn has important beneficial effects on the catalytic properties of Pt for the dehydrogenation of lower alkanes. First of all, and as mentioned above, it increases the selectivity towards dehydrogenation by inhibiting hydrogenolysis. Similarly, the addition of Sn has a profound effect on the catalyst life. The Pt-Sn catalyst retains a much higher activity than the pure Pt catalyst. The enhancement in selectivity and stability, clearly illustrated in Fig. 2a and 2b, can be explained in geometric terms by the dilution of Pt ensembles by Sn. As described above, this dilution greatly reduces the activity towards reactions that require a large ensemble of Pt atoms to constitute the active sites, such as hydrogenolysis and coking. Also, the increase in selectivity of the pure Pt sample with time on stream shown in Fig. 2b can be explained by the same geometric arguments. The carbon deposited during the reaction plays the role of an inactive species that inhibits undesired reactions, including coking. This explanation accounts for two observations on the pure Pt catalyst. Both, the selectivity and the stability of the catalyst improve as a function of time on stream as the number of large ensembles is reduced by the presence of carbon. On the pure Pt catalyst the initial deactivation is very fast. Thus, it is difficult to measure the true initial activity of unpromoted Pt and determine whether the initial dehydrogenation rate on Pt is higher or

PAGE – 9 ============
8lower than on the bimetallic catalyst. In fact, pure Pt may be initially more active, but after a short time it deactivates while the bimeta llic retains its activity. This phenomenon is particularly pronounced in the absence of added H 2, as in the experiments of Fig. 2. The higher initial activity of pure Pt catalysts can only be observed when high H2/hydrocarbon feed ratios are employed. 15 3.1.3. Preparation of Pt-Sn Dehydrogenation Catalysts and its effects on performance The beneficial role of adding Sn to Pt catalysts has been observed on different supports and on catalysts prepared by various methods. In some preparations the addition of Sn has been conducted by sequential impregnation, first of an aqueous solution containing the Sn precursor (typically SnCl 2) and then of another solution containing the Pt precursor (typically H 2PtCl 6). In these preparations, the degree of Pt-Sn interaction is relatively low because Sn tends to interact with the support becoming segregated from Pt. Therefore, to maximize the metal-metal interaction, other preparations have been used. For example, the use of an aqueous solution containing both Pt and Sn precursors and excess HCl results in the formation of a bimetallic PtCl 2(SnCl 3)22- complex. 16,17 In this preparation, both metals are deposited on the surface in the same compound and, after the thermal treatment, a high degree of alloyed metals can be obtained. Other methods involve the Pt-catalyzed surface reduction of an organometa llic tin compound, resulting in the selective deposition of Sn over the Pt surface. 18 For example, the catalyst can be prepared from a solution of Sn(C 4H9)4 in an organic solvent, such as n-hexane. This solution is added onto a pre-reduced Pt/support sample, without exposure to air. The H

PAGE – 10 ============
9that remains adsorbed on Pt is responsible for the reduction of the Sn(C 4H9)4 , causing the selective deposition over the metal. 19 The catalyst preparation may critically influence the Pt-Sn interaction and consequently the catalytic behavior. As illustrated in Fig. 3 for a series of Pt-Sn catalysts supported on s ilica, the promoting effect of Sn strongly depends on the method employed in the preparation. Catalysts prepared by co-impregnation, particularly when the solvent is an aqueous solution containing HCl, result in a relatively high extent of Pt-Sn interaction. By contrast, sequential impregnation results in a large fraction of unalloyed Pt. Although alumina and silica have been the two supports most widely studied for dehydrogenation reactions, other non-acidic supports have also shown promising properties. For example, KL zeolite has been proposed to be an effective support for Pt- Sn dehydrogenation catalysts. It has been reported 20 that these catalysts maintain high isobutane dehydrogenation activity and selectivity for extended reaction intervals. Characterization of these materials indicates that a fraction of the Sn is alloyed with Pt and the rest is in the form of Sn 2+ ions, exchanged with K + previously present in the zeolite. It was further hypothesized that the K displaced from the zeolite framework can interact with the Pt-Sn alloy particles, promoting the activity. The promoting effect of K was demonstrated by deliberate addition of excess K to the catalysts, which greatly enhanced the dehydrogenation activity and selectivity. Table 1 makes a comparison of the isobutane dehydrogenation conversion on several Pt and Pt-Sn catalysts. 20 In the table we compare the reaction rates per gram of Pt. It is generally accepted that the best form of comparing the activity of a series of

PAGE – 11 ============
10 catalysts is in terms of turnover frequency (TOF), based on the number of sites as measured by hydrogen chemisorption. However, when Pt-Sn alloys and unalloyed Pt phases co-exist, the TOF may be misleading. The chemisorptive capacity of Pt is greatly changed when it is alloyed with Sn. For example, it has been shown that the heat of CO adsorption drops from Pt to PtSn alloy by as much as 20 kJ/mol. 21 This energy difference causes significant changes in the adsorption equilibrium constants at 298 K and, consequently, in the observed adsorption capacity. In recent work, 22 it was observed that the CO adsorption capacity of Pt-Sn/SiO 2 dropped to zero after 1 hour at 500ºC on isobutane stream. However, the catalyst was almost as active for dehydrogenation as at the beginning of the reaction. The reason for this apparent discrepancy is that the unalloyed Pt is rapidly covered by coke, while the Pt-Sn alloys remain more or less free of coke. Since alloyed Pt does not adsorb significant amounts of CO the CO/Pt measured on the fresh catalyst was mainly due to the fraction of unalloyed Pt, which after a while contributes little to the activity. The situation is very similar for the chemisorption of hydrogen. Verbeek and Sachtler 23 have shown that Pt-Sn alloys adsorb very little hydrogen and have ascribed this decrease to a lowering of the heat of adsorption. Microcalorimetry studies 24 have shown that, even though the addition of Sn resulted in a large decrease in the saturation adsorption coverages for H or CO, the heats of adsorption at zero coverage on Pt:Sn samples were similar to those on pure Pt. However, it must be noted that these SiO 2-supported samples were prepared by sequential impregnation and, as shown below, this technique leads to a large fraction of unalloyed Pt, which may be responsible for the measured high heats of adsorption. For changes in the initial heats of adsorption to be seen, the majority of Pt needs to be alloyed with Sn. In fact, when

18 KB – 52 Pages

By redy