Electronic Thesis and Dissertation Repository

Degree

Doctor of Philosophy

Program

Chemical and Biochemical Engineering

Supervisor

de Lasa, Hugo

Abstract

Scaling up a photoreactor requires both knowledge of optical properties of the slurry medium and an established kinetic model. Measuring the scattering and absorption coefficients of particles suspended in water involves the use of specialized optical equipment, as well as the partial solution of the radiative transfer equation (RTE). In addition, modeling of the radiation field in photoreactors with complicated geometries offers special challenges.

On the other hand, most of the kinetic models (KM) for phenol photodegradation reported in the literature were obtained for a single organic chemical species only. In fact, neglecting all the intermediate species generated during the photoreaction, is a common oversimplification that limits the KM application. As a result, once the radiation and kinetic models fully established, energy efficiencies can be obtained.

In this PhD dissertation, the photocatalytic degradation of phenol over four different TiO2 catalysts is studied. It is proven that phenol yields hydroquinone, catechol, benzoquinone, and acetic and formic acids as main intermediate species.

The radiation field inside photocatalytic reactors is predicted by solving the RTE. From the solution of the RTE, the local volumetric rate of energy absorption (LVREA) is also calculated. The radiation field inside an annular photoreactor is simulated using the Monte Carlo (MC) method for different TiO2 suspensions in water. All simulations are performed by using both the spectral distribution, and the wavelength-averaged scattering and absorption coefficients.

The Henyey-Greenstein phase function is adopted to represent forward, isotropic and backward scattering modes. It is assumed that the UV lamp reflects the back-scattered photons by the slurried medium. It is proven, photo-absorption rates, using MC simulations and spectral distribution of the optical coefficients, agree closely with experimental observations from a macroscopic balance (MB). It is also found that the scattering mode of the probability density function, is not a critical factor for a consistent representation of the radiation field.

When solving the RTE, two optical parameters are needed: (1) the absorption and scattering coefficients, and (2) the phase function. In this research work, the MC method, along with an optimization technique, is shown to be effective in predicting the wavelength-averaged absorption and scattering coefficients for different TiO2 powders. To accomplish this, the LVREA and the transmitted radiation (Pt) in the photoreactor are determined by using a MB. The optimized coefficients are calculated ensuring that they comply with a number of physical constrains, falling in between bounds established via independent criteria.

The optimization technique is demonstrated by finding the absorption and scattering coefficients for different semiconductors that best fit the experimental values from the MB. The objective function in this optimization is given by the least-squared error for the LVREA. A photocatalyst is synthesized and its optical properties determined by the proposed method. This approach is a general and promising one; not being restricted to reactors of concentric geometry, specific semiconductors and/or particular photocatalytic reactor unit scale.

Based on the proposed intermediate reactions, a phenomenological based unified kinetic model is proposed for describing the obtained experimental observations in phenol photodegradation. This Langmuir-Hinshelwood (L-H) kinetic model is based on a “Series-Parallel” reaction network. This reaction model is found to be applicable to the various TiO2 photocatalyst in the present study.

This unified kinetic network is based on the identified and quantified chemical species in the photoconversion of phenol and its intermediates. In order to minimize the number of optimized parameters, the adsorption constants of the different intermediate species on the different catalysts configuration, are obtained experimentally. It is shown that the unified kinetic model requires a number of significant assumptions to be effective; avoiding overparametization. As a result, the unified kinetic model is adapted for each specific TiO2 photocatalyst under study.These different models adequately describe the experimental results. It is shown that this approach results in good and objective parameter estimates in the L-H kinetic model, which is typically applied to photocatalytic reactors.

Finally, two efficiency factors, the quantum yield and the photochemical and thermodynamic efficiency factor, are obtained, in this PhD dissertation. These factors are based on the kinetic model proposed and the radiation being absorbed by the photocatalyst particles. The efficiency calculations consider stoichiometric relationships involving observable chemical species and OH· groups. The obtained efficiency factors point toward a high degree of photon utilization and, as a result, the value of photocatalysis and Photo-CREC-Water reactors for the conversion of organic pollutants in water is confirmed.

Share

COinS