Radiation Absorption and Optimization of Solar Photocatalytic Reactors for Environmental Applications JOSE COLINA-M ´ ARQUEZ, † FIDERMAN MACHUCA-MART ´ INEZ, ‡ AND GIANLUCA LI PUMA* ,§ Chemical Engineering Department, Universidad de Cartagena, Cartagena, Colombia, Chemical Engineering School, Universidad del Valle, GAOX Group, Cali, Colombia, and Photocatalysis & Photoreaction Engineering, Department of Chemical and Environmental Engineering, The University of Nottingham, Nottingham NG7 2RD, United Kingdom Received January 13, 2010. Revised manuscript received May 20, 2010. Accepted May 25, 2010. This study provides a systematic and quantitative approach to the analysis and optimization of solar photocatalytic reactors utilized in environmental applications such as pollutant remediation and conversion of biomass (waste) to hydrogen. Ray tracing technique was coupled with the six-ﬂux absorption scattering model (SFM) to analyze the complex radiation ﬁeld in solar compound parabolic collectors (CPC) and tubular photoreactors. The absorption of solar radiation represented by the spatial distribution of the local volumetric rate of photon absorption (LVRPA) depends strongly on catalyst loading and geometry. The total radiation absorbed in the reactors, the volumetric rate of absorption (VRPA), was analyzed as a function of the optical properties (scattering albedo) of the photocatalyst. The VRPA reached maxima at speciﬁc catalyst concentrations in close agreement with literature experimental studies. The CPC has on average 70% higher photon absorption efﬁciency than a tubular reactor and requires 39% less catalyst to operate under optimum conditions. The “apparent optical thickness” is proposed as a new dimensionless parameter for optimization of CPC and tubular reactors. It removes the dependence of the optimum catalyst concentration on tube diameter and photocatalyst scattering albedo. For titanium dioxide (TiO 2 ) Degussa P25, maximum photon absorption occurs at apparent optical thicknesses of 7.78 for CPC and 12.97 for tubular reactors. Introduction Heterogeneous photocatalysis based on TiO 2 and modiﬁed semiconductor photocatalysts has received a great deal of attention in the literature as an environmentally friendly method for the treatment and puriﬁcation of lightly con- taminated water and air (1, 2), self-cleaning surfaces (2, 3), and as a route for sustainable energy production (4, 5). Irradiation of TiO 2 with radiation of energy higher than the band gap of the semiconductor (e.g., λ < 384 nm for TiO 2 anatase) produces electron-hole charges on the semicon- ductor which can initiate reduction and oxidation reactions. Oxidation of TiO 2 adsorbed water, by surface holes, produces hydroxyl and other radical species which are responsible for the titania’s wide-ranging activity toward large classes of contaminants (e.g., aromatics, halogenated hydrocarbons, pesticides, endocrine disrupting chemicals, inorganics, and others) (1-3, 6) and for the inactivation of microorganisms and toxins (e.g., coliforms, viruses, microcystins) (2, 3, 7, 8). Reduction of water by the photogenerated electrons and simultaneous oxidation of biomass (waste) and/or organic pollutants, in the absence of oxygen, has been proposed as a method for hydrogen production from renewable sources (9). The photogenerated electrons can also reduce harmful heavy metals present in the environment (1). The expansion and the technological application of heterogeneous photocatalysis in the above environmentally relevant processes require the development of efﬁcient photoreactors. Since sunlight is the only free source of photons, solar photoreactors are desirable. Photons are necessary reactants or initiators; therefore, the knowledge of the radiation ﬁeld in a photoreactor is crucial to maximize their efﬁcient use, products yield, and selectivity (10). However, due to lack of widely accessible engineering knowledge current photoreactors for environ- mental applications and sustainable energy production are designed by experience rather than from rigorous scientiﬁc principles. Furthermore, intrinsic kinetic parameters of photocatalytic reactions can be determined only with the knowledge of the radiation ﬁeld in the photoreactor (11-14). In general, optimal reactor geometry, photon distribution, and catalyst loading are interrelated concepts which require accurate modeling of the radiation ﬁeld, in the photoreactor. The spatial distribution of photons, i.e., the local volumetric rate of absorption (LVRPA) depends on the photon source, the optical properties of the system, the distribution of the catalyst and the reactor geometry. Several approaches have been proposed to calculate the LVRPA in photocatalytic reaction systems in which absorption and scattering of photons occur. The more rigorous studies have developed a numerical solution of the radiative transfer equation (RTE) (10, 11). This procedure implies a thorough mathematical work for describing the radiant ﬁeld for each given reactor geometry and the emission model of the photon source. Other studies have used a semiempirical approach (15) that calculates the LVRPA by ﬁtting model parameters to experi- mental data, in a attempt to avoid solving the RTE. Some authors (16) proposed a model without adjustable parameters considering all scattering directions to describe the photo- degradation of the herbicide carbaryl. Others (17) modeled the radiation ﬁeld in a parabolic trough solar photocatalytic reactor. The P1 approach was proposed for solving the RTE in solar tubular photoreactors (18). Other studies have considered Monte Carlo and the ﬁnite volume methods (19, 20) to solve the RTE. For practical purposes in photocatalytic processes, it is necessary to use simple models that describe the radiation ﬁeld and photon absorption in most reactor geometries and that can determine the essential parameters needed for design and optimization. The Six-Flux absorption-scattering model (SFM) has been proposed to estimate the LVRPA in hetero- geneous photocatalytic systems offering simplicity and accuracy. This model is based on establishing probabilities for six different Cartesian scattering directions. The SFM was applied to estimate the LVRPA in a ﬂat plate (21), in annular * Corresponding author phone: +44 (0)115 951 4170; fax: +44 (0)115 951 4115; e-mail: gianluca.li.puma@nottingham.ac.uk. † Universidad de Cartagena. ‡ Universidad del Valle. § The University of Nottingham. Environ. Sci. Technol. 2010, 44, 5112–5120 5112 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010 10.1021/es100130h  2010 American Chemical Society Published on Web 06/09/2010