Influence of Pressure during the Alkylation of Toluene with Ethane

Institute of Chemical Technology, UniVersity of Stuttgart, 70550 Stuttgart, Germany. The direct alkylation of toluene with ethane to the isomeric ethy...
0 downloads 0 Views 111KB Size
Ind. Eng. Chem. Res. 2007, 46, 395-399

395

Influence of Pressure during the Alkylation of Toluene with Ethane Dirk Singer, Seyed Alireza Sadat Rezai, Sarah Sealy, and Yvonne Traa* Institute of Chemical Technology, UniVersity of Stuttgart, 70550 Stuttgart, Germany

The direct alkylation of toluene with ethane to the isomeric ethyltoluenes and hydrogen using a bifunctional Pd/H-ZSM-5 catalyst was carried out in a fixed-bed, flow-type reactor. The effect of pressure was investigated between 1 and 100 bar and at moderate temperatures of 300 and 350 °C. From stoichiometry, no influence of pressure is expected for ideal gases. However, thermodynamic equilibrium calculations for a network of reactions, taking into account nonideal gas behaVior, do predict an influence of pressure. Toluene conversions of 5-8% observed experimentally at 300 °C over the entire pressure range are in line with these calculations. In contrast, the toluene conversion observed at 350 °C and 50 bar is at 27% well above the equilibrium value calculated using a limited reaction network. It is proposed that the formation of hydrogen-rich light alkanes is favored at high pressures, creating a hydrogen sink. Hence, the equilibrium is shifted toward the formation of ethyltoluenes. The maximum experimental yield of ethyltoluenes achieved so far is approximately 14 wt %. 1. Introduction With the recent progress in chemical reaction engineering and highly active, multifunctional catalysts, the activation of relatively inert alkanes may be achieved. The direct use of shortchain alkanes, especially methane and ethane, as a basic chemical feedstock represents an attractive alternative to alkenes. Significant reserves of alkanes are found in nature in the form of wet natural gas, whereas alkenes have to be synthesized via other chemical reactions. The activation of alkanes can be achieved by oxidative and nonoxidative methods. Reactions of oxidative activation are often thermodynamically favored, e.g., due to formation of the stable but economically unattractive byproducts H2O and/or CO2. In contrast, reactions of nonoxidative activation may produce H2 as a valuable byproduct, but the equilibrium conversions of such reactions are generally low, and thus, measures have been taken to improve this situation, e.g., shifting equilibrium by removing a product.1,2 The alkylation of aromatics with alkanes, in particular, represents a potential new route for the synthesis of important feedstock chemicals such as n-propylbenzene1 and cumene,1,3 ethyltoluenes,4,5 and ethylbenzene.6 The use of propane as an alkylating agent, for the synthesis of cumene and n-propylbenzene from benzene, has received the most attention with studies on different types of catalysts.1,3 Using H-ZSM-5 catalysts, at low temperatures and propane conversions, cumene and npropylbenzene were the major observed products. However, with increasing temperature and conversion, ethylbenzene, toluene, ethane, and methane were the main products as a result of secondary cracking. The selectivity toward the desired products was improved by the incorporation of Pt, a dehydrogenating component, into the catalyst.1 Use of the less reactive ethane as an alkylating agent for benzene on bifunctional zeolite catalysts results in lower conversions than with propane; however, a high selectivity to the primary alkylation product ethylbenzene can be achieved.6 In the direct alkylation of toluene with ethane, Sealy et al.5 achieved an order-of-magnitude increase in conversion and a 30% increase in selectivity to ethyltoluenes with the addition of 0.9 wt % Pd to an H-ZSM-5 * Corresponding author: General Tel.: +49 711 685 64061. Fax: +49 711 685 64065. E-mail: [email protected].

catalyst. Doubling of this Pd content further increased conversion; however, the selectivity to ethyltoluenes was reduced with a simultaneous formation of methane and propane, possibly as a result of secondary reactions. In order to address the low conversion, due to equilibrium limitations, Smirnov et al.1 introduced intermetallic Zr2Fe, a hydrogen scavenger, to shift equilibrium. Thus, a 2-8 times increase in conversion could be achieved; however, the high conversions could not be maintained, either because of saturation of the hydrogen scavenger or fast catalyst deactivation at high conversions.1 In the present study, the direct gas-phase alkylation of toluene with ethane to form ethyltoluenes has been investigated in a plug-flow reactor. A bifunctional Pd/H-ZSM-5 catalyst with a high activity and selectivity for this reaction was used.4,5 The study focused on the effects of total pressure. From stoichiometry, no influence of the pressure is expected on the equilibrium conversion, as long as the reaction partners are believed to behave as ideal gases. However, experimentally much higher conversions and yields are observed at the higher total pressures applied in the present study as compared to our previous works carried out under atmospheric pressure.4,5 Thermodynamic calculations taking into consideration nonideal gas behavior at high pressure were carried out. 2. Experimental Section Catalyst Preparation. Zeolite ZSM-5 was synthesized without organic template from Kieselsol (30 wt % SiO2 in water, type VP-AC 4038, Bayer AG), sodium aluminate (54 wt % Al2O3, 41 wt % Na2O, Riedel-de Hae¨n), NaOH (>99 wt %, Merck), and demineralized water, producing a synthesis gel with the molar composition 60SiO2/Al2O3/9Na2O/2400H2O. Crystallization occurred within 5 days in a stainless-steel autoclave at 160 °C under stirring.7 The resulting zeolite was subjected to two consecutive ion-exchange steps with an aqueous solution (1 mol‚dm-3) of NH4NO3 (Fluka) at 80 °C, with each step lasting for 4 h. Palladium ion exchange was carried out by adding an aqueous solution of Pd(NH3)4Cl2 (Fluka) dropwise to a suspension of the ammonium ion-exchanged zeolite, under vigorous stirring, at room temperature. The zeolite powder was pressed without a binder and sieved to a particle size between 0.2 and 0.3 mm. The catalyst was

10.1021/ie060407y CCC: $37.00 © 2007 American Chemical Society Published on Web 12/14/2006

396

Ind. Eng. Chem. Res., Vol. 46, No. 2, 2007

Scheme 1. Reaction Network Taken into Account for Modeling Thermodynamic Equilibrium. From Top to Bottom: Alkylation of Toluene, Disproportionation of Toluene, and Alkylation of Benzene

Figure 1. Conversion of toluene on 1.0Pd/H-ZSM-5 (nSi/nAl ) 20) at 350 °C and various pressures as a function of time on stream.

activated in situ, prior to starting the experiment. To achieve a high dispersion of the noble metal, the catalyst was first heated in flowing air (50 cm3‚min-1) at a rate of 2 °C‚min-1 to a final temperature of 300 °C and then held at this temperature for 22 h. Next, the catalyst was purged with nitrogen for 1 h. During the subsequent reduction phase, the catalyst was heated in flowing hydrogen (50 cm3‚min-1) at 2 °C‚min-1 to 350 °C and held at this temperature for 22 h. The nSi/nAl ratio and the content of palladium were 20 and 1.0 wt % (referenced to the mass of the dry catalyst), respectively. The elemental composition of the catalyst was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES, Perkin-Elmer Plasma 400). Catalytic Experiments. Experiments were performed in a flow-type apparatus using a stainless-steel fixed-bed reactor at total pressures of 1-95 bar and reaction temperatures of 300 and 350 °C. Ethane (99.95 vol %, Westfalen AG) and nitrogen (99.999 vol %, Westfalen AG) were fed through a toluene (>99.9%, Merck) saturator containing Chromosorb P-NAW (Macherey-Nagel). Nitrogen was used as an internal standard, but also in order to avoid condensation of ethane and toluene above their critical pressures. The n˘ nitrogen/ n˘ ethane ratio was ∼5. Typically, the mass of the dry catalyst, the n˘ ethane/n˘ toluene ratio, and the WHSV (weight hourly space velocity of toluene and ethane) amounted to 0.5 g, ∼6, and 3.7 h-1, respectively. Product analysis was achieved using an on-line sampling system, a capillary gas chromatograph, and a CP-PoraPLOT Q column (length ) 30 m, inner diameter ) 0.32 mm, film thickness ) 20 µm, Chrompack). Two detectors in series were employed, namely, a thermal conductivity detector for the analysis of hydrogen followed by a flame ionization detector for the analysis of hydrocarbon products. Correction factors for the two detectors were determined separately. With nitrogen as internal standard, the results from both detectors were combined. From the mass or molar flows, the selectivities of all products were calculated in wt % (for the industrially relevant major products as displayed in Figure 3) or % (for the light products as displayed in Figure 8). The yields in % were determined from the selectivities and the toluene conversion. Model Development. Thermodynamic equilibrium was calculated for gas-phase total pressures of 1-100 bar, reaction temperatures of 300 and 350 °C, and an n˘ ethane/n˘ toluene ratio of 6. Separate calculations were performed for the alkylation reaction of toluene with ethane yielding m-ethyltoluene and hydrogen and for the disproportionation reaction of toluene

yielding m-xylene and benzene. In this article, these will be referred to as “single”-reaction calculations. Since the inclusion of nitrogen resulted only in small deviations in single-reaction calculations, nitrogen was excluded in all calculations presented here. In addition, calculations were also carried out for a reaction network (Scheme 1) accounting for the main products observed experimentally: ethyltoluene isomers and hydrogen from toluene alkylation with ethane, xylene isomers and benzene from toluene disproportionation, and ethylbenzene and hydrogen from benzene alkylation with ethane. Reactions within this network will be named as “network” in order to differentiate them from the single-reaction calculations. Chemical equilibrium equations were set up using eq 1 (see the Nomenclature section for details). Equilibrium constants were calculated, with eq 2, from standard Gibbs free energies of reaction at different temperatures.8 The Peng-Robinson equation of state was applied to determine fugacity coefficients of species and, hence, to take into account gas nonideality at high pressure (eq 3).9 The set of nonlinear equations was solved using a Powell hybrid method. Mass balance constraints were set in order to find the correct equilibrium concentrations of all species. l

Ka,j )

∏ i)1

aVi ij ) Kλ,j ‚

(

Ka,j ) exp l

Kλ,j )

∏ i)1

l

yVi ∏ i)1

)

∆Grxn j RT

() hf i

yiP

ij

(1)

(2)

νij

(3)

3. Experimental Results Time-on-Stream Behavior. Figure 1 shows the time-onstream (TOS) behavior of the conversion (X) of toluene over 1.0Pd/H-ZSM-5 at 350 °C. At a total pressure of 1 bar, the conversions are relatively high at the beginning of the experiment, gradually decreasing to the conversion ratio Xtoluene,TOS)1222min/ Xtoluene,TOS)43min ) 0.4. However, with increasing pressure, the catalyst stability improves. At a total pressure of 10 bar, the decrease in conversion is not as pronounced, with the conversion ratio Xtoluene,TOS)1200min/Xtoluene,TOS)33min ) 0.7. At and above 30 bar, conversion is constant during the entire run. Possible explanations for this effect could be that catalyst deactivation is reduced at higher pressures because of hydrogenation or extraction of coke precursors or by desorption of coke precursors

Ind. Eng. Chem. Res., Vol. 46, No. 2, 2007 397

Figure 2. Conversions of toluene and ethane on 1.0Pd/H-ZSM-5 (nSi/nAl ) 20) at 300 and 350 °C as a function of total pressure (the numbers are average values for the entire experiment).

due to stronger adsorption of nitrogen at higher pressures. It should be noted that further results are, therefore, presented as average conversions for the entire experiment. Effect of Pressure. The effect of pressure, from 1 to 95 bar, on the conversion of toluene and ethane at 300 and 350 °C is shown in Figure 2. As expected for equilibrium-limited endothermic reactions, conversion increases with increasing temperature from 300 to 350 °C. From stoichiometry, it is not expected that pressure should affect the static equilibrium conversion. However, at 300 °C, there is a small increase in conversion. At 1 bar, the average toluene conversion is ∼5% and increases to ∼8% at 30 bar; thereafter at higher pressures, an apparent plateau is observed. In contrast, at 350 °C and 1 bar, the toluene conversion is ∼6%. With increasing pressure, a maximum toluene conversion of 27% is observed at a total pressure of 50 bar. Further increasing pressure results in a decrease in conversion; at 95 bar, the toluene conversion is 19%. Precise analysis of ethane conversion as a function of pressure is made difficult because of the high n˘ ethane/n˘ toluene ratio, which was chosen to suppress side reactions of toluene. At 300 and 350 °C, ethane conversion is ∼1 and 2%, respectively. At 50 bar and 350 °C, a small maximum is observed in ethane conversion, corresponding to the maximum toluene conversion. The reproducibility of the results was demonstrated by repeating selected experiments with freshly prepared catalyst samples. For the experiment at 350 °C and 68 bar, for example, the relative error of the toluene conversion was 1.1% and that of the ethane conversion was 16%. Figure 3 shows the selectivity (S) toward the major aromatic products, namely, ethyltoluenes, xylenes, and benzene, at 300 and 350 °C versus pressure. At 1 bar, 300 and 350 °C, selectivity toward the desired ethyltoluenes is 94 and 83 wt %, respectively. The disproportionation reaction is preferred at higher temperatures, maybe because of reduced steric hindrance to the bulky diphenylalkane intermediates within the ZSM-5 pores (the effective pore dimensions increase with increasing temperature)10 and the high activation energy.11 The selectivity toward the disproportionation reaction also increases with increasing pressure: the selectivity ratio Sethyltoluenes,P)95bar/Sethyltoluenes,P)1bar is ∼0.8 at 300 °C, whereas at 350 °C, the same ratio is ∼0.5. 4. Thermodynamics Figure 4 presents the results of thermodynamic equilibrium calculations as a function of pressure at 300 and 350 °C for the independently calculated, single alkylation or disproportionation reaction of toluene with ethane to m-ethyltoluene and hydrogen

Figure 3. Selectivities of major products during the conversion of toluene and ethane on 1.0Pd/H-ZSM-5 (nSi/nAl ) 20) at 300 and 350 °C as a function of pressure (the numbers are average values for the entire experiment).

Figure 4. Equilibrium conversion of toluene for independent single alkylation and disproportionation reactions as a function of pressure (n˘ ethane/ n˘ toluene ) 6).

or to m-xylene and benzene. Thermodynamically, the disproportionation reaction is very much favored under the pressure range of interest. At 1 bar and 300 °C, the equilibrium conversions of the alkylation and disproportionation reactions are 4 and 45%, respectively. An initial increase in conversion is predicted for both reactions; however, the disproportionation reaction shows a maximum at 30 bar and 48% conversion. The maximum equilibrium conversion for the disproportionation reaction shifts from 30 bar at 300 °C to 50 bar at 350 °C. The maximum equilibrium conversion for the disproportionation reaction is due to a minimum in the product of fugacity coefficients (cf. Figure 5). For the alkylation reaction, a plateau of 5% equilibrium conversion above 30 bar is calculated at 300 °C (cf. Figure 4). The plateau shifts to 50 bar and 7% at 350 °C. At 300 and 350 °C, the increase in conversion for the alkylation reaction is 30 and 25%, respectively. These effects are due to real gases resembling ideal gases more closely at the higher temperature. Results of equilibrium calculations carried out for the network described in Scheme 1 are illustrated in Figure 6. Comparison of Figures 4 and 6 shows that there is no significant difference in equilibrium conversion trends for the single reactions versus the network. At 1 bar, thermodynamic conversion for the disproportionation reactions in the network has increased to 53% from 45% for the single reaction, yielding m-xylene and benzene. In contrast, the equilibrium conversion for the alkylation reactions in the network decreases to 3% at 1 bar and 300 °C, compared to 4% equilibrium conversion for the independently calculated, single alkylation reaction forming

398

Ind. Eng. Chem. Res., Vol. 46, No. 2, 2007

Figure 5. Product of fugacity coefficients for the alkylation and disproportionation reactions at 300 and 350 °C as a function of pressure.

Figure 6. Equilibrium conversion of toluene for a reaction network including alkylation and disproportionation of toluene and alkylation of benzene as a function of pressure (n˘ ethane/n˘ toluene ) 6). Table 1. Experimental Conversion X and Yield Y during the Alkylation of Toluene with Ethane on 1.0Pd/H-ZSM-5 (nSi/nAl ) 20) Varying the Total Pressure at 350 °C (The Numbers Are Average Values for the Entire Experiment) pressure/bar

1

10

30

50

68

95

Xethane/% Xtoluene/% Yhydrogen/% Ymethane/% Ypropane/% Ybenzene/% Yxylenes/% Yethylbenzene/% Yethyltoluenes/%

1.2 6.2 2.6 0.2 d.l. 0.4 0.3 0.1 3.4

1.1 9.6 2.2 1.4 0.3 1.0 0.8 0.3 4.0

1.3 16.2 3.3 1.7 0.6 2.1 2.0 0.4 4.9

2.3 27.1 2.5 6.7 0.8 4.2 4.0 0.9 7.8

1.6 20.5 2.4 3.9 0.8 3.5 3.4 0.7 5.2

2.1 19.2 2.5 4.8 0.7 2.8 3.1 0.6 4.4

m-ethylbenzene. This is as a result of competition in the network with the thermodynamically more favored disproportionation reaction. 5. Discussion Table 1 shows experimental conversions and product yields in mol % at 350 °C and different pressures. The yield of the desired products, the isomeric ethyltoluenes, increases with increasing toluene conversion. At 1 bar, the yield of ethyltoluene isomers is 3.4%, increasing to give a maximum of 7.8% at 50 bar. The yield of hydrogen is more constant, between 2 and 3% over the entire pressure range, whereas it would be expected that hydrogen should follow the same trend as the ethyltoluenes since hydrogen and ethyltoluenes are formed in stoichiometric amounts. Furthermore, although the methane yield is very low

Figure 7. Experimental conversion of toluene (the numbers are average values for the entire experiment) on 1.0 Pd/H-ZSM-5 (nSi/nAl ) 20) in comparison to the toluene conversion from the model using the equilibrium calculations for the reaction network.

at 1 bar, it increases to 6.7% at 50 bar, a significant yield as methane is formed in amounts similar to ethyltoluenes at high pressure. The low hydrogen yields can be explained by its consumption in secondary reactions, i.e., hydrogenation of coke precursors or carbonaceous deposits or toluene demethylation with hydrogen to form benzene and methane. Other light alkanes such as propane are also produced, however, only to a lesser extent, with selectivities below 1%. Ethane may also be produced, although this cannot be observed as it is a reactant. Yields of benzene and xylenes, formed predominantly by the disproportionation of toluene, are ∼4% at 50 bar. However, the amount of benzene is consistently higher than that of xylenes. This means that benzene is not only consumed in the alkylation with ethane forming ethylbenzene, but must also be produced, e.g., by toluene demethylation. The yield of ethylbenzene is low, always below 1%. Other reaction pathways may also be possible, e.g., xylenes demethylation with hydrogen to toluene, but also ethyltoluenes demethylation with hydrogen, forming ethylbenzene. Secondary cracking pathways have been previously observed at high temperature and conversion.1,5,6 The effects may also be more pronounced at the high pressures applied in this work. At the milder temperature of 300 °C, the effects are less pronounced (not shown). Much less methane is produced, and propane is below the detection limit. Ethyltoluene isomer yields are slightly higher than hydrogen yields. However, it should also be kept in mind that the hydrogen yields are determined with the thermal conductivity detector. Hence, hydrogen yields are less accurate than hydrocarbon yields. Figure 7 shows the experimental conversion of toluene at 300 and 350 °C as well as the toluene conversion from a model using the equilibrium calculations for the reaction network. The final model brings together the equilibrium conversions of the toluene alkylation and disproportionation reactions to yield a combined toluene conversion, which can be compared with the experimental results. Pore restrictions within zeolite ZSM-5 channels and the high n˘ ethane/n˘ toluene ratio suppress the toluene disproportionation reaction requiring two toluene molecules as opposed to the alkylation reaction requiring only one. Therefore, the model assumes that only 1 and 5%, respectively, of the disproportionation reaction can take place at 300 and 350 °C. Whereas at 300 °C the experimental results are in line with the model calculations, at 350 °C there is no correlation between experimental results and the calculated model. In order to

Ind. Eng. Chem. Res., Vol. 46, No. 2, 2007 399

Acknowledgment The authors thank Prof. Dr.-Ing. Jens Weitkamp for his support and encouragement and Prof. Klaus Mo¨ller for assistance with thermodynamic calculations. Financial support from Fonds der Chemischen Industrie is gratefully acknowledged. Nomenclature

Figure 8. Toluene conversion and selectivity of light products at 350 °C on 1.0 Pd/H-ZSM-5 (nSi/nAl ) 20) as a function of pressure (the numbers are average values for the entire experiment).

explain the large deviations at 350 °C, the experimentally observed selectivity to light products is given as a function of pressure in Figure 8. As has already been discussed, hydrogen yields are much lower than expected in comparison to ethyltoluene isomer yields. Furthermore, at 1 bar, hydrogen and methane selectivity are ∼40 and