Oxidation of ortho-xylene

Abstract

The thesis embodies the work on the vapour?phase oxidation of o-xylene over sintered and compacted vanadium pentoxide catalyst in a fixed?bed reactor (FBR) and a continuous stirred?tank catalytic reactor (CSTCR).

A survey of literature reveals that although considerable amount of work has been reported on o-xylene oxidation over various unpromoted and promoted catalysts, no kinetic study has been made over sintered and compacted vanadia catalyst.

Thermodynamic analysis of the oxidation of o-xylene to phthalic anhydride, maleic anhydride and carbon dioxide reveals that while the individual reactions are not hindered by equilibrium limitations, as the magnitude of the equilibrium constants are very high, the occurrence of parallel reactions is influenced by the air?to?o-xylene mole ratio.

From a preliminary study on the oxidation of o-xylene over unpromoted and promoted (molybdenum and tungsten oxides) vanadia impregnated on celite, pumice, silicon carbide and quartz, sintered and compacted vanadia catalyst was selected for detailed kinetic study. Experimental data were collected in the region of constant catalytic activity.

The products of oxidation were found to be phthalic and maleic anhydrides and carbon dioxide. It was experimentally established that there was no reaction in homogeneous phase.

In a fixed?bed reactor, kinetic and mechanism data were collected in the kinetic regime and interphase and intraphase diffusional resistances were experimentally found to be negligible. Theoretical computation revealed that the mass and heat?transfer gradients were not significant. The range of process variables used with the catalyst of average particle diameter of 0.0356 cm were: (i) Concentration of o-xylene: 0.18–0.90% (below lower limit of inflammability) (ii) Time factor: 16.67–41.67 g cat/(gm mole o-xylene/hr) (iii) Temperature: 450–517°C (iv) Partial pressure of o-xylene: 0.1236–0.6180 cm Hg (v) Partial pressure of oxygen: 5.7019–14.2548 cm Hg

The data representing the effect of time factor on the degree of conversion at different temperature levels were analysed by assuming the reactions follow first?order kinetics, which was confirmed by fitting the rate expression derived for pseudo?first?order kinetic reaction. The general reaction scheme proposed for the oxidation of o-xylene to phthalic anhydride, maleic anhydride and carbon dioxide narrowed down to a scheme in which o-xylene oxidation proceeds via parallel routes. This was based on quantitative values of the rate constants and also the pattern of product distribution.

Experimental data collected at different partial pressures of o-xylene and oxygen were used to explain the kinetic data over the catalyst surface. Even though there is overwhelming support for the redox mechanism to explain the oxidation of o-xylene over vanadia catalyst, besides the above mechanism, the other mechanisms tested were: two?stage redox mechanism with desorption of oxygen not negligible, three?stage redox mechanism, Rideal mechanism, Langmuir–Hinshelwood mechanism and power?law model. The models were tested based on the initial rates. Screening of these models by both classical and statistical (non?intrinsic parameter) methods indicated that the two?stage redox model with an order combination (1.0, 1.0 with respect to o-xylene and oxygen) best represented the data with an average absolute deviation of 3.95%.

The rate constants of the two?stage redox model were estimated by linear and non?linear regression analysis. The activation energies and pre?exponential factors of the Arrhenius equation for pseudo?homogeneous and heterogeneous reactions were evaluated by linear regression analysis.

In the second part of the work, kinetics of the oxidation of o-xylene over sintered and compacted vanadia catalyst was studied in a continuous stirred?tank catalytic reactor. A CSTCR provided with a four?bladed paddle catalyst basket was fabricated for this purpose.

The reactor (CSTCR) behaviour in respect of mass?transfer and mixing characteristics was determined by: (i) Following the course of the reaction at different basket?stirring speeds at two flow rates and two temperature levels (ii) Mass?transfer correlations developed from naphthalene evaporation experiments (iii) Ford–Perlmutter method (iv) Residence?time?distribution studies

Kinetic measurements were then made under conditions where mass?transfer effects were negligible and perfect mixing existed. The temperature gradient across the film surrounding the catalyst particle was computed and found to be small. The range of variables used with the catalyst of average particle diameter 0.325 cm were: (i) Time factor: 14.47–46.30 g cat/(gm mole o-xylene/hr) (ii) Temperature: 450–517°C (iii) Partial pressure of o-xylene: 0.1932–0.6180 cm Hg (iv) Partial pressure of oxygen: 5.7019–14.2548 cm Hg

The CSTCR data were analysed using the findings from FBR experiments. The rate data based on the parallel?reaction scheme were fitted to both a pseudo?homogeneous model and a heterogeneous (two?stage redox) model. Rate constants were evaluated by linear regression; for the heterogeneous model, non?linear regression was also used. Activation energies and pre?exponential factors were computed.

A comparison between PBR and CSTCR was made.

Summary of conclusions: (i) Sintered and compacted vanadium pentoxide is an excellent high?temperature catalyst for vapour?phase oxidation of o-xylene (450–517°C). (ii) Phthalic anhydride, maleic anhydride and carbon dioxide are formed via parallel routes. (iii) Kinetics are well represented by a two?stage redox model with reaction orders 1.0 (o-xylene) and 1.0 (oxygen). (iv) Kinetic parameters differ between PBR and CSTCR; disagreement is smaller when pseudo?homogeneous model is used. (v) CSTCR is preferable to FBR for intrinsic?kinetic studies of highly exothermic reactions such as o-xylene oxidation; improvements in CSTCR design are suggested.