Xylene isomerization model reaction

The effect of crystal size of these zeolites on the resulted toluene conversion can be ruled out as the crystal sizes are rather comparable, which is particularly valid for ZSM-5 vs. SSZ-35 and Beta vs. SSZ-33. The concentrations of aluminum in the framework of ZSM-5 and SSZ-35 are comparable, Si/Al = 37.5 and 39, respectively. However, the differences in toluene conversion after 15 min of time-on-stream (T-O-S) are considerable being 25 and 48.5 %, respectively. On the other hand, SSZ-35 exhibits a substantially higher concentration of strong Lewis acid sites, which can promote a higher rate of the disproportionation reaction. Two mechanisms of xylene isomerization were proposed on the literature [8] and especially the bimolecular one involving the formation of biphenyl methane intermediate was considered to operate in ZSM-5 zeolites. Molecular modeling provided the evidence that the bimolecular transition state of toluene disproportionation reaction fits in the channel intersections of ZSM-5. With respect to that formation of this transition state should be severely limited in one-dimensional (1-D) channel system of medium pore zeolites. This is in contrast to the results obtained as SSZ-35 with 1-D channels system exhibits a substantially higher... 

As a result of steric constraints imposed by the channel structure of ZSM-5, new or improved aromatics conversion processes have emerged. They show greater product selectivities and reaction paths that are shifted significantly from those obtained with constraint-free catalysts. In xylene isomerization, a high selectivity for isomerization versus disproportionation is shown to be related to zeolite structure rather than composition. The disproportionation of toluene to benzene and xylene can be directed to produce para-xylene in high selectivity by proper catalyst modification. The para-xylene selectivity can be quantitatively described in terms of three key catalyst properties, i.e., activity, crystal size, and diffusivity, supporting the diffusion model of para-selectivity. 

A quantitative model requires knowledge of the diffusivity under reaction conditions and of the intrinsic activities for toluene disproportionation and xylene isomerization. While these are not easily obtained, the methodology has been worked out for the case of paraffin and olefin cracking (5). So far, we have obtained an approximate value for the diffusivity, D, of o-xylene at operation conditions from the rate of sorptive o-xylene uptake at lower temperature and extrapolation to 482°C (Table V). 

Rates of model reactions are more commonly used to determine relative rather than absolute surface acidities and a variety of acid-catalyzed reactions have been used for this purpose (1-3). Xylene isomerization is a particularly well-substantiated model reaction, thanks to work by Ward and Hansford (43). They demonstrated that the conversion of o-xylene to p- and /n-xylenes over a series of synthetic silica-alumina catalysts increases as the alumina content is increased from 1 to 7%. The number of strong Brdnsted acids in each member of the catalyst series was measured by means of infrared spectroscopy. Since conversion of o-xylene was found to be a straight-line function of the number of Br0nsted acids (see Fig. 9), rate of xylene isomerization appears to be a valid index of the amount of surface acidity for this catalyst series. This correlation also indicates that the acid strengths of these silica-alumina preparations are roughly equivalent. 

Impregnation of ZSM-5 with different compounds produce small changes in the zeolite pore dimensions which increases the selectivity to para-xylene (168,169) as can be seen from the results in Table 12 taken from (169). Iliese results can be explained on the bases of a model which considers the interplay between the relative rates of the toluene disproportionation, product diffusion, and xylene isomerization (170), as it is schematized in Figure 32, taken from (170), in where D is the diffusion coefficient, K is the rate constant, and r is the crystal radius. The reactions considered are Isomerization "(I) and Disproportionation (D), while the reaction products are benzene (Bz), toluene (I and xylenes (P,M,0). Initial products (i), primary products (o), and secondary products (P,M,0) arc considered in this scheme. 

A further improvement of the approach of Wei [107] was reported in 1989 by Hashimoto et al. [42], which considered not only adsorption effects, but also the nonselective reactions occurring at the outer surface of the crystallites. The nonselective influence of these reactions has also been recognized by Fraenkel [35] in 1990, who studied the formation of xylene from toluene on a HZSM-5 catalyst. Fraenkel assumed that inside the crystallite only />-xylene is formed, whereas the ortho and meta isomers are sterically inhibited there. Hence, he concluded that the amount of o- and m-xylene observed during his experiments must be due to the isomerization of p-xylene at the outer surface of the crystallites. This two-step mechanism was first suggested in 1987 by Paparetto et al. [82] for the ethylation of toluene. It may also be worth noting that Fraenkel s model took into account not only the isomerization but also the nonselective alkylation at the outer crystallite surface. 

We will use now the same method and Mordenite zeolite model as in the previous part, and investigate the isomerization of xylene isomers. As described in the previous part, this reaction can proceed via two alternative routes, viz. a methyl shift isomerization, and disproportionation reactions. Moreover, we observed than in the case of toluene isomerization, the location of toluene with respect to the Br0nsted acidic site for the shift isomerization was of no consequence for the activation energy barrier. We will check these mechanisms for the three xylenes. 

Well designed distillation columns are close to electronic filters and while complex their performance can be well predicted from thermodynamic data. The complexity of chemical reactors varies from case to case. The performance of a well designed methanol or ammonia reactor can be predicted as well as that of a distillation column, whereas in more complex systems the risks of predictions is greater. In some sense identification of chemical reaction models has some features of a purely statistical correlation. Even in a simple well defined system such as isomerization of xylene where we can measure all reaction... 

---