Pergamon ('hemwal En~li,leerin9 5cWnce. Vol 52, Nos 21 22. pp 4205-4213. 1997 1997 Elsevier Science Lid All rights reserved Prmted in Great Britain Pil: S0009-2509(97)00263-7 o0o9 2509 97 $17.00 + 0.00 Effect of catalyst concentration and simulation of precipitation processes on liquid-phase catalytic oxidation ofp-xylene to terephthalic acid A. Cincotti, R. Orrfi, A. Broi and G. Cao* Dipartimento di Ingegneria Chimica e Materiali, Universita'degli Studi di Cagliari. Piazza d'Armi, 09123 Cagliari, Italy (Accepted 7 July 1997) Abstract---The influence of catalyst concentration, i.e. cobalt naphthenate, on product distribu- tion and kinetic constants of the lumped kinetic scheme of liquid-phase p-xylene oxidation proposed in previous works (cf. Cao et al., 1994a, b) is investigated. The experiments involving various levels of catalyst concentrations (from 1.67 to 33.3 × 10-4 mol/kgi) are conducted in an isothermal semi-batch oxidation reactor where both the gas and the liquid phase are well mixed. The dependence of the kinetic constants of the lumped kinetic scheme on the catalyst concentration is examined. In addition, the interaction between the chemical reactions of the lumped kinetic scheme for p-xylene oxidation to terephthalic acid and the precipitation kinetics of both 4-carboxybenzaldehyde and terephthalic acid is analyzed theoretically. A semi-batch gas-liquid reactor model which incorporates the description of the above phenomena allows us to identify their interplay..(" 1997 Elsevier Science Ltd Keywords: p-xylene; oxidation; terephthalic acid; catalyst: precipitation. INTRODUCTION In the chemical industry, liquid-phase catalytic oxida- tion of organic compounds, which can be either homolytic or heterolytic depending upon the mecha- nism of oxygen activation (cf. Emanuel and Gal, 1986), constitutes a wide class of important processes. In particular, when addressing the study of homolytic oxidation processes, e.g. cyclohexanol from cyclo- hexane, terephthalic acid from p-xylene, isophthalic acid from m-xylene, emphasis was placed on the in- vestigation of the catalytic system, i.e. catalyst and promoter concentration, nature of solvent, reaction temperature, etc., and its influence on the oxidation rate (Sheldon and Kochi, 1981; Raghavendrachar and Ramachandran, 1992). This type of approach typi- cally accounts for the complex nature of the catalyst within the already complex chain elementary reaction schemes, which involve a very large number of rad- icals as well as molecular species (cf. Carr~ and San- tacesaria, 1980). However, when simulating the gas-liquid reactors, the formulation of detailed kinetic models of oxidation processes requires a large compu- tation effort for solving the diffusion-reaction equa- *Corresponding author. Tel.: 39/70/675-5058: fax: 39.'70/675-5067: e-mail: cao@visnu.dicm.unica.it. tions in the film at the gas-liquid interface, coupled with the continuity equations for both gaseous and liquid bulk phases. This is due to the intrinsic differ- ence between the space scale typical of molecular and radical species. In addition, the estimation of kinetic parameters of chain elementary reaction kinetic schemes may not be reliable due to the difficulty of monitoring all participating species, including highly reactive radicals, as a function of time in semi-batch reactors where specific experimental studies are typi- cally conducted. The most common approach is to lump the detailed mechanism into a set of global reactions which in- volves only molecular species, whose concentration can be, in principle, easily monitored as a function of time. Without involving formal procedure of general validity but simply including the minimum number of reactions to describe the behavior of all the species of interest, various lumped kinetic schemes for homolytic oxidation processes have been developed in the litera- ture, by Cavalieri d'Oro et al. (1980) for p-xylene oxidation, Chen et al. (1985) for o-xylene oxidation, Morbidelli et al. (1986) for ethyl-benzene autoxidation and Krzysztoforski et al. (1986) for cyclohexane oxi- dation. By accounting for the most important inter- mediates and final products of the process, i.e. p-tolualdehyde, p-tolualcohol, p-toluic acid, 4205