Scaling-Up of Photoreactors: Applications to Advanced Oxidation Processes

Abstract

A general procedure to perform the scaling-up of photochemical reactors from first principles is presented, working with three applications to Advanced Oxidation Technologies. This work is not a review paper on the subject; it is an abridged description of the analysis and design methods developed with our coworkers during the last fifteen years. The objective of our commitment was to scaling-up photoreactors of any type, starting from small-scale laboratory reactors results, with the additional imposed restraint of never making use of empirically adjusted parameters. All our proposals are based on the principles of chemical reactor engineering science and the rigorous application of radiation field theory. The description is made starting from one process where the design makes use of a radiation model that only requires two optical properties: the radiation absorption and reflection coefficients of the employed catalyst (TiO2). The procedure is illustrated with catalytic wall reactors in the gas phase. The kinetics was obtained in a small flat plate laboratory reactor (Total reacting surface area=81 cm2) and extrapolated to a much larger reactor made of three concentric annular cylindrical tubes with all their walls coated with the catalyst (Total reacting surface area=5,209 cm2). Both the laboratory and the pilot size reactor had to be modeled. The pollutant degraded was perchloroethylene (PCE) in air. A complete reaction sequence was employed. In both cases, Philips lamps of different sizes but of the same type TL/08-F4T5/BLB were used. Results were quite satisfactory. The next case concerns a simple homogeneous reaction studying the degradation of formic acid employing H2O2 and UVC radiation (Germicidal). The kinetic model and the reaction sequence were obtained in a small laboratory batch reactor (V=70 cm3) and extrapolated to a continuous tubular reactor of annular cross section (A=65 cm2) having L=2m of reaction length. In this case, the required optical property is the absorption coefficient of hydrogen peroxide to apply the radiative transfer equation (RTE) without scattering. Similarly as in previous case, both the laboratory reactor and the larger continuous reactor had to be mathematically described. In both cases, the same type of lamp was employed: Philips TUV lamp. Predictions of the reactor performance were quite acceptable. The third illustration, considers a TiO2, slurry photocatalytic reactor for degrading 4- chlorophenol. In this case, the complete RTE is required. Thus, all the optical properties are needed: the absorption and scattering coefficients as well as the phase function for scattering. The kinetic model and the reaction sequence were obtained in a small laboratory, batch reactor (of variable size from 29 to 290.4 cm3) and extrapolated to a larger reactor (V=734.4 cm3) having completely different irradiation conditions and configuration. Once more, it was necessary to model both the laboratory reactor and the larger scale one. In both cases, lamps with very similar spectral output power distribution were employed (Actinic type: Philips TL/08 and TLK/09N). Very satisfactory results were obtained. There are several key points to note in a procedure that permits to move from laboratory reactors to significant scale-ups without experimentally adjustable parameters. The approach is based on four necessary conditions: (i) to have a validated kinetic scheme (a detailed mechanism or a precise empirical representation), (ii) to have a validated, intrinsic reaction kinetic expression as a function of position and time [R(x,t)], (iii) to use in both reactors the same spectral radiation output power distribution [l for monochromatic radiation and f(l) for the polychromatic cases], and (iv) to apply and correctly solve a rigorous mathematical model to both the laboratory and the large-scale reactor.