Honeycomb Monolith Photoreactor

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

This paper presents a detailed simulation of a honeycomb monolith photoreactor using a Computational Fluid Dynamics (CFD) approach. The monoliths were treated as porous zones and photocatalytic oxidation reactions were assumed to follow Langmuir-Hinshelwood kinetics. The Discrete Ordinates (DO) model was used to simulate the light intensity. The coupling between the light intensity and kinetics were achieved using a user-defined subroutine. Photocatalytic oxidations of formaldehyde and toluene in the photoreactor were simulated and compared with published experimental data. The three-dimensional simulation results were successfully validated without any adjustable parameters, indicating that the porous media modeling using DO model could be employed to describe the photocatalytic degradation process. Using the modeling approach, parametric studies were carried out to investigate the impacts of some key factors on the degradation of the volatile organic compounds. With further validation, the model could be used to design, optimize, as well as to control a photoreactor.

Keywords:CFD, porous, photocatalytic oxidation, honeycomb monolith, discrete ordinates.


characteristic pore diameter [m]

diffusion coefficient for species [m2/s]

radiation intensity [W/m2]

spectral radiation intensity [W/m2]

species mass diffusion

refractive index

pressure [Pa]

contaminant sink term [kg/m3.s]

temperature [K]

contaminant concentration [ppmv]

water vapor concentration [ppmv]

species mass fraction

gravity vector

unit vector

velocity vector [m/s]

momentum sink term [N/m3]

density [kg/m3]

dynamic viscosity [kg/m.s]

wavelength [m]



stress tensor

scattering coefficient [1/m]

spectral scattering coefficient [1/m]

absorption coefficient [1/m]

spectral absorption coefficient [1/m]

solid angle

phase function


‘Sick Building Syndrome' is caused by long-term exposure to air toxics consisting mostly volatile organic compounds (VOCs) from indoor air which has been known to be carcinogenic to human health (Lomborj, 2001). The buildings which contain air toxics coming from the paintings, carpeting, building materials, office equipments et cetera, are known as ‘sick buildings'. Sick-building occupants will likely feel dizziness, headaches, poor concentration, fatigue, increased appearance of asthma attacks, and many more. Most of the syndromes are caused by poor air quality, which is mostly due to inadequacy in heating-ventilation-air-conditioning (HVAC) system in the buildings (CaluTech Inc., 2005). In order to prevent this, improving in indoor air quality becomes a crucial issue. In the past, there have been a number of conventional methods on air purification, such as thermal and catalytic oxidation, adsorption, and condensation (Ruddy and Carroll, 1993). However, these methods require rather high operating costs as they need high operating temperature and pressure, catalyst regeneration or waste handling (Castrillon and de Lasa, 2007). Heterogeneous photocatalytic oxidation technology, in another way, appears to be a more practical and promising alternative to the conventional methods for its ability to operate under ambient conditions. In a photoreactor, volatile organic compounds are adsorbed to the surface of the photocatalyst, and oxidized by the OH radicals formed by the photocatalyst due to irradiation and produce carbon dioxide and water. The photocatalyst used is commonly titanium dioxide due to low cost, high refractive index, and strong oxidative potential (CaluTech Inc., 2005).

Up to date, several photocatalytic reactor configurations have been investigated, some representatives are annular packed bed (Raupp et al., 1997), honeycomb monolith photoreactors (Hossain and Raupp, 1998, 1999), optical fiber photoreactors (Choi et al., 2001, Wang and Ku, 2002, 2003, Danion et al., 2004), and novel photoreactor-Photo-CREC-air (Castrillon et al., 2006). Several approaches have been adapted for modeling of these photoreactors, such as the Monte-Carlo method (Spadoni et al., 1978; Alexiadis, 2006), pseudohomogeneous model with two-flux incident submodel (Akehata et al., 1976), rigorous three-dimensional diffusion-convection-reaction model using Gauss-Legendre quadrature (Raupp et al., 1999), and two-dimensional heterogeneous convection-reaction model (Changrani and Raupp, 2000). Most recently, Computational Fluid Dynamics (CFD) studies have arouse a growing attention in modeling photocatalytic reactors. Mohseni and Taghipour investigated the flow field and photocatalytic reaction of air-borne vinyl chloride oxidation in an annular photocatalytic reactor using CFD Fluent® (Mohseni and Taghipour, 2004). Pareek et al. used the CFD Fluent® to simulate the behavior of an annular photocatalytic reactor for the degradation of spent Bayer liquor (Pareek et al., 2003) using Discrete Ordinates (DO) radiation model. S-Estivill et al. used the CFD Fluent® to integrate the radiation field with photocatalytic reaction kinetics to yield a rigorous model of a flat-plate, single-pass, flow-through photocatalytic reactor for air purification (S-Estivill et al., 2007). In recent times, Castrillon et al. have been working on the modification of a Photo-CREC-air proposed by de Lasa and Ibrahim (2004) as model unit, and CFX-5.7.1 was used to model and simulate the Photo-CREC-air for air treatment (Castrillon et al., 2006). Furthermore, CFD modeling was also applied to catalytic combustion in monoliths (Canu and Vecchi, 2002, Mazumber and Sengupta, 2002). In the studies of Mazumber and Sengupta, a rigorous sub-grid scale modeling was employed to treat heterogeneous catalytic reactions occurring in porous or honeycomb monoliths (Mazumber and Sengupta, 2002).

A major unresolved issue in commercialization of the photocatalytic oxidation technology is that simultaneous modeling, control and optimal design still exist as a challenge for photocatalytic reactors. It has been well known that honeycomb monolith reactors offer low pressure drop and high surface area (Sauer and Ollis, 1994). In this paper, Computational fluid dynamics (CFD) codes-Gambit® and Fluent® were utilized to simulate a honeycomb monolith photoreactor. Gambit was used to construct and mesh the reactor geometry, and the grid was then employed in Fluent® for further simulations. Discrete Ordinates (DO) radiation model was chosen for the radiation field, while Fluent® Species Transport model and Langmuir-Hinshelwood reaction kinetics were adopted for the photocatalytic oxidation reaction modeling. There were a number of studies on modeling of a single channel of a photoreactor (Hossain et al., 1999; Hossain and Raupp, 1998, 1999; Alexiadis, 2006), however, in order to assist in the design of an optimal reactor, modeling of the entire photoreactor plays a leading role. Therefore, in this paper, a novel method using porous media formulations with rather simpler and low computational cost was adapted. The monolith channels were treated as porous zones, and rigorous reaction modeling using local radiation intensities was included. Using the modeling approach, the photocatalytic degradation of formaldehyde and toluene were evaluated and compared with the experimental results from the studies of Hossain et al. (Hossain et al., 1999). In addition, parametric studies were carried out, to investigate the impacts of some key factors, such as light intensity, adsorption and reactor wall reflectivity.


The experimental data were taken from the studies of Hossain et al. (Hossain et al., 1999). In the study, a pilot scale demonstration unit was assembled to examine the experimental conversion of formaldehyde and toluene. The pilot unit was built into 8 ft. long, 1 ft. x 1 ft. cross-section duct, and can accommodate up to 6 lamp banks (consisting 4 UV lamps) and 5 individual square honeycomb monoliths. Compressed fiber glass was used as the reactor wall, and the reactor inner duct walls were coated with aluminium foil to reflect UV light. The reactor elements were alumina ceramic foams (HiTech Corp.) (Hossain et al., 1999), 10 pores per in. and wash-coated with Degussa P25 titania to yield a thickness of 15 to 70 µm. Unfiltered building air was blown into the reactor by a blower to provide the desired flow. The UV lamps used were of the type of low-pressure mercury lamps (Voltrac G10T5L-S400), with 98% of UV emitted at 254 nm. A UVC power meter (Oriel UVC Goldilux) was used to measure the light intensities on the monolith faces. The contaminants, formaldehyde and toluene were entered the reactor by introducing through a diffusion-controlled vaporizer and a compressed gas cylinder supply respectively. A Bruel & Kjaer 1302 multigas monitor and a gas chromatograph (IBM-9630) were used to measure the concentrations and the estimated absolute uncertainties were 0.020 ppmv, and 0.010 ppmv for formaldehyde and toluene respectively. (Note that the above experimental details were extracted from Hossain et al., 1999, “Three-dimensional developing flow model for photocatalytic monolith reactors").


Fluent®, which is a commercial CFD package, was used in the simulations, taking into consideration the coupling of the radiation field with porous media formulations. The regularly shaped monoliths were treated as porous zones, and several general assumptions were made: (1) the system was in steady state, isothermal conditions; (2) the flow was laminar and incompressible; (3) there were negligible absorption, scattering or emission of radiation by the gaseous medium; (4) the monolith channels were adiabatic and non-conducting; (5) A constant value of dilute-approximation was used for the diffusion coefficient; (6) no homogeneous chemical reaction occurred; and (7) there were sufficiently thick and uniform coatings so that no light can transmit through.

Physical System

The experiments carried out by Hossain et al. involved 1 to 5 honeycomb monoliths, flow rate ranged from 55 to 410 cubic feed per minute (CFM) (i.e. approximately 0.28 to 2.09 m/s), and the UV intensities measured on the monolith surfaces were 6.0 to 6.5 mW/cm2 (Hossain et al., 1999). Figure 1 shows the photoreactor consisting of 5 monoliths (10 surfaces) of 64 cells per square inch (CPSI) and aspect ratio 8. There are 6 lamp banks placed parallel to the monoliths, each consists of 4 individual lamps.

Air In


UV Lamps

Air Out

Figure 1: Monolith photoreactor.

Boundary Conditions

The following boundary conditions were employed: (1) uniform inlet air velocity was used and the velocity and concentration of the contaminants were set according to the experimental tests; (2) the outlet boundary condition was pressure-specified, with gauge pressure at 0. (3) the inlets and outlets of the monolith channels were modeled as 'Porous Jump'; (4) the UV lamps were modeled as semi-transparent walls; (5) the reactor wall was covered with aluminum foil, to reflect the UV light.

Radiation Field Model

The Discrete Ordinates (DO) model in Fluent® was used for the radiation field and the fineness of the angular discretization was set to 6 x 6. The radiative transfer equation (RTE) in the direction as a field function for a finite number of discrete solid angles is expressed as:


where the first term on the right hand side symbolizes the loss of photons due to absorption; the second term symbolizes the loss of photons due to out-scattering; and the last term symbolizes the in-scattering of photons.

Optical Properties of the medium

In this research, the reactive medium is air and the photocatalyst is titanium dioxide. It was assumed that the absorption and scattering coefficients for air are negligible, and for the Degussa P25 titania particles, the following relations were used (Romero et al., 1997):



where is the catalyst loading in g/m3. Moreover, for this optically thin medium, isotropic scattering was used and thus the phase function parameter in Equation (1) bears the value of unity.

Flow Model, Transport Equation and Kinetics

It is more practical to model the individual monolith channels by using porous media formulations instead of introducing computational grids, which are often in the range of sub-microns compared with the overall photoreactor (Mazumder and Sengupta, 2002). Fluent® solves Navier-Stokes equations based on the assumptions of conservation of mass and momentum in a moving fluid. The conservation of mass and momentum are described in Equations (4) and (5) respectively. For flows involving reactions or species transport, an additional species conservation equation, Equation (6) is solved:




where is the porosity which bears the value of 1 in bulk flow and 0.87 in monoliths. is the momentum sink term (N/m3) which is to include the pressure drop due to the photocatalyst support and the monolith channels. For laminar flow through the monolith channels,


where the first term on the right hand side accounts for viscous loss (Darcy's) and the second term accounts for inertia loss. and are user inputs, where




In Equation (6) is the mass diffusion in laminar flows,


where is the diffusion coefficient for species in the mixture. A constant dilute approximation of m2/s was used in the simulation to account for the diffusivity of the dilute concentration of VOCs in air.

in Equation (6) is the contaminant sink term, which is the kinetic model of the rate of consumption of the VOCs. In this work, a bimolecular form of the Langmuir-Hinshelwood (L-H) reaction rate was adopted to model the reaction rate of formaldehyde and toluene conversion at constant humidity. The correlation coefficients were extracted from Obee's work on a glass-plate photocatalytic reactor apparatus (Obee, 1996). According to the study, with the assumptions that the contaminant and water vapor are the only important factors in the reaction rates, the L-H rate equation was simplified to


where and are respectively the concentrations of the contaminant and water in ppmv. and are Langmuir adsorption constants in ppmv-1, (µmol.cm-2.h-1) is the rate constant at local UV flux incident on the monolith walls and is evaluated by:


where is the UV flux to the glass-plate reactor for which the kinetic constants were evaluated (Obee, 1996, Hossain et al., 1999). A rate exponent of 0.5 is chosen for both formaldehyde and toluene degradations. In Fluent® simulation, Equations (12) and (13) are compiled using user-defined function (UDF) to describe the degradation processes. The corresponding reaction constants are given in Table 1.














Table 1: L-H Correlation coefficients (Obee, 1996).


The results are summarized subsequently in the following order: Model Validation, Simulation Results: Pressure Drop, Fluid Flow and VOCs Removal, Effect of Light Intensity, Effect of Adsorption and Effect of Reactor Wall Reflectivity.

Model Validation

Equations (1) to (13) account for full modeling of the photoreactor. The simulation results using Fluent® were validated against the experimental results from the pilot-scale demonstration unit of Hossain et al.'s work. The tests were simulated according to different photoreactor layout and operating conditions of the experiments. A total of 12 cases were compared (6 cases each for toluene and formaldehyde conversions). Table 2 shows the experimental details and model results for different tests. The flow rates ranged from 55 to 410 ft3/min, and the mean measured UV flux on the 12" x 12" monolith surfaces was kept at 65 W/m2. It is important to note that all the flow rates were chosen such that the photoreactor was kinetically controlled (Hossain et al., 1999). Take test 1 as an assessment case, inlet air consisting of 0.3 ppmv toluene and 3500 ppmv water vapor was introduced at flow rate 63 ft3/min to the monolith photoreactor which contains 24 lamp surfaces (6 lamp banks) and 10 monolith surfaces (5 monoliths of length 1 in). The measured conversion of toluene was 68%, whereas the model conversion was 67.23%, with error percentage 1.13%.

Figures 2 and 3 show the graphical comparison of the model and experimental results for toluene and formaldehyde degradations respectively. Y-axis is the model conversion and X-axis is the measured/experimental conversion. The location of the data points determines the accuracy of the model data. Ideal case happens when all the data points fall on the diagonal of the plot. As seen in Figure 2, for toluene degradation, the slope of the best-fit line for the data points is 0.98, with R2 of 0.97. The model-predicted results successfully matched with the experimental results without any adjustable parameters.

Figure 2: Comparison between model and experimental data for toluene degradation.

Figure 3: Comparison between model and experimental data for formaldehyde degradation.

For formaldehyde conversions, as shown in Figure 3, the slope of the best-fit line is 0.90, with R2 of 0.86. The slope and R2 value are rather low compared to those of toluene conversions. Some possible reasons for the slight mismatch may be attributed to uncertainties in the experiments and the simplified reaction mechanism employed in the simulation. Since carbon monoxide is found to be one of the by-products from the photocatalytic oxidation of formaldehyde (Liu et al., 2005), the production of carbon monoxide could be one of the influencing factors. Nevertheless, the good agreement between the model and experimental results for both cases reveals that the novel approach presented in this paper is a good 'vehicle' to be used to describe the photocatalytic oxidation of VOCs in a honeycomb monolith photoreactor.

Simulation Results: Pressure Drop, Fluid Flow and VOCs Removal

Pressure Drop

The pressure drop profile across the photoreactor at the middle plane, (x, 0, 0), from the inlet to the outlet in the assessment case—test 1, is shown in Figure 5. The total area-weighted average pressure drop from the inlet to the outlet was found to be 1.98 Pa. This small pressure drop reveals the main benefit of a monolith photoreactor. As seen in Figure 5, due to the momentum sink term introduced by Equation (7), the pressure decreases linearly when the air flows through the monolith channels (there are 5 monoliths in this case) which are modeled as porous zones.

Figure 5: Simulated axial pressure drop across the middle plane of the photoreactor (, 0, 0).

Fluid Flow

The fluid flow in a monolith photoreactor can be observed through the velocity contour of the middle plane as shown in Figure 6. It can be seen that air enters with uniform velocity (green color), and when the flow reaches the lamps, it slips through the lamp surface, causing the highest velocity (indicated by red color) to happen in the spaces between lamps and the spaces between lamps and reactor wall.

Figure 6: Enlarged velocity contour around lamp area.



Flow Rate


UV Flux (W/m2)


Conc. (ppmv)

Water Conc.



Surface (Figure)

Monolith Length (in.)

Measured Conversion


Predicted Conversion (%)

Error Percent. (%)

(a) Toluene Degradation






24/10 (Figure 4a)










24/10 (Figure 4a)










24/10 (Figure 4a)










24/10 (Figure 4a)










24/10 (Figure 4a)










24/10 (Figure 4a)





(b) Formaldehyde Degradation






4/2 (Figure 4b)










4/2 (Figure 4b)










8/4 (Figure 4c)










8/4 (Figure 4c)










16/8 (Figure 4d)










24/10 (Figure 4a)





(c) Photoreactor Layout for Different Tests

Figure 4a: Lamp/surface: 24/10.

Figure 4b: Lamp/surface: 4/2.

Figure 4c: Lamp/surface: 8/4.

Figure 4d: Lamp/surface: 16/8.

Table 2: Operating conditions of different tests for (a) toluene degradation; (b) formaldehyde degradation (Hossain et al. 1999), corresponding to (c) Figure 4a, Figure 4b, Figure 4c and Figure 4d.

VOCs Removal

Again, taking test 1 as an assessment case, Figure 7(a) shows the mass fraction contour, showing the decrease in toluene mass fraction across the middle plane of the photoreactor (x, 0, 0), from the inlet to the outlet. Figure 7(b) further shows the decrease in toluene amount when the flow encounters the monoliths. It was determined that the amount of toluene was decreased to 50% of its initial amount when it came to about 15% of the reactor length (i.e. the second monolith).



Figure 7: (a) Mass fraction contour of toluene; (b) Mass fraction profile across the middle plane of the photoreactor (x, 0, 0).

Effect of Light Intensity

According to Hossain and Raupp, the UV light distribution profile is determined by TiO2 thin-film absorptivity and monolith channel aspect ratios (Hossain and Raupp, 1998). Here, a constant aspect ratio of 8 is used and the dependence of VOCs concentration on light intensity is plotted in Figure 8. In the studies from Obee, the oxidation rate dependence for toluene and formaldehyde degradation on light intensity follows a power law of exponents 0.64 and 0.6 respectively, with uncertainty 0.05 (Obee, 1996). The exponent used in this study is 0.5 for both cases. It is shown in Figure 8 that the amount of degradation was first-order at lower intensities and followed a power law of exponent 0.5 at higher intensities. This result conforms to the study of Yu et al, 2007.

Figure 8: Conversion dependence on UV intensity.

Adsorption Effect

By changing the Langmuir-Hinshelwood adsorption constants in Equation (12), the effect of adsorption on the degradation amount can be evaluated. As shown in Figure 9, when the adsorption is increased, the degradation process is enhanced. However, once a critical adsorption capacity is reached (for instance 14 ppmv-1 in test 10, and 6 ppmv-1 in test 11), the degradation process would not be improved any further, due to saturation of the photocatalyst with the contaminant molecules.

Figure 9: Conversion dependence on the adsorption constant.

Effect of Reactor Wall Reflectivity

Since the absorption, scattering and emission of radiation of the medium in this system are neglected, and the catalyst loading is of small amount, the medium is considered as an optically thin system. As shown in Figure 10, taking two examples, tests 10 and 12, when the reactor wall reflectivity increases, the conversion increases. Therefore, it can be concluded that for this system, the reactor wall reflectivity plays an important role in light distribution, which in turn, affects the effectiveness of the photodegradation of VOCs.

Figure 10: The effect of reflectivity on conversion.


A full simulation of the photocatalytic degradation of toluene and formaldehyde in a honeycomb monolith photoreactor using CFD—Fluent® was developed. The monoliths were treated as porous zones and Fluent® Discrete Ordinates (DO) radiation model, fluid flow and transport model were employed. In addition, user-defined function (UDF) was used to account for rigorous modeling of the reactions using Langmuir-Hinshelwood kinetics equation which involves local radiation intensities. The modeling data matched the experimental results remarkably well, by this means, signifying the reliability of the novel approach. Finally, parametric studies on the effects of light intensity, adsorption and reactor wall reflectivity were carried out. With further validation, this model could be used to assist in the design of an optimum reactor as well as the control configuration.



AKEHATA T., ITO K., and INOKAWA A., (1976), “Light intensity profiles in bubble-dispersed systems. An approach to analysis of heterogeneous photoreactors”, Kagaku Kogaku Ronbunshu, 2, 583.

ALEXIADIS Alessio, (2006), “2-D radiation field in photocatalytic channels of square, rectangular, equilateral triangular and isosceles triangular sections”, Chem. Eng. Sci., 61, 516-525.

CANU P., and VECCHI S., (2002), “CFD Simulation of reactive flows: Catalytic combustion in a monolith”, AlChE J., 48, 2921-2935.

CASTRILLON S. R-V., IBRAHIM H., de LASA H., (2006), “Flow field investigation in a photocatalytic reactor for air treatment (Photo-CREC-air)”, Chem. Eng. Sci., 61, 3343-3361.

CASTRILLON S. R-V. and de LASA H., (2007), “Performance evaluation of photocatalytic reactors for air purification using computational fluid dynamics (CFD)”, Ind. Eng. Chem. Res., 46, 5967-5880.

CHANGRANI R.G., RAUPP G.B., (2000). “Two-dimensional heterogeneous model for a reticulated-foam photocatalytic reactor.” AlChE J., 46, 829-842.

CHOI W., KO J.Y., PARK H., CHUNG J.S., (2001), “Investigation on TiO2-coated optical fibers for gas-phase photocatalytic oxidation of acetone”, Applied Cat. B: Envi, 31, 209-220.

DANION A., DISDIER J., GUILLARD C., ABDELMALEK F., JAFFREZIC-R. N., (2004), “Characterizeation and study of a single-TiO2-coated optical fiber reactor”, Applied Cat. B: Envi, 52, 213-223.

HOSSAIN M.M., GREGORY B., RAUPP G.B., (1998), “Radiation field modeling in a photocatalytic monolith reactor”, Chem. Eng. Sci., 53, 3771-3780.

HOSSAIN M.M., GREGORY B., RAUPP G.B., (1999), “Polychromatic radiation field model for a honeycomb monolith photocatalytic reactor”, Chem. Eng. Sci., 54, 3027-3034.

HOSSAIN M.M., RAUPP G.B., HAY S.O., Obee T.N., (1999), “Three-dimensional developing flow model for photocatalytic monolith reactors”, AlChE J., 45, 1309-1321.

LOMBORJ B., (2001), The sceptical Environmentalist, Cambridge University Press: Cambridge, U.K.

LIU H., LIAN Z., YE X. and SHANGGUAN W., (2005), “Kinetic analysis of photocatalytic oxidation of gas-phase formaldehyde over titanium dioxide”, Elsevier Sci. Inc., Chemosphere, 60, 630-635.

MAZUMDER S. and SENGUPTA D., (2002), “Sub-grid scale modeling of heterogeneous chemical reactions and transport in full-scale catalytic converters”, Elsevier Sci. Inc., Combustion and Flame, 131, 85-97.

MOHSENI M., TAGHIPOUR F., (2004), “Experimental and CFD analysis of photocatalytic gas phase vinyl chloride (VC) oxidation”, Chem. Eng. Sci., 59, 1601-1609.

OBEE T.N., (1996), “Phooxidation of subs-parts-per-million toluene and formaldehyde levels on titania using a glass-plate reactor”, Environ. Sci. Technol., 30, 3578-3584.

PAREEK V.K., BRUNGS M.P., ADESINA A.A. (2001). “CFD simulation of photoreactors”, 6th World Congress of Chem. Eng. Melbourne, Australia.

PAREEK V.K., and ADESINA A.A. (2004), “Light intensity distribution in a photocatalytic reactor using finite volume”, AlChE J., 50, 1273-1288.

PAREEK V.K., (2005), “Light intensity distribution in a dual-lamp photoreactor”, International J. of Chem. Reactor Engineering, 3, Article A56.

RAUPP G.B., NICO J.A., ANNANGI S., CHANGRANI R., and ANNAPRAGADA R., (1997), “Two-flux radiation-field model for an annular packed-bed photocatalytic oxidation reactor”, AlChE J., 43, 792-901.

ROMERO R.L., ALFANO O.M., CASSANO A.E., (1997), “Cylindrical photocatalytic reactors. Radiation absorption and scattering effects produced by suspended fine particles in an annular space”, Ind. Eng. Chem. Res., 36, 3097.

RUDDY, E. and CARROLL L., (1993), “Select the best VOC control strategy”, Chem. Eng. Prog., 89 (July), 28.

SALVADÓ-ESTIVILL I., M.HARGREAVES D., LI PUMA G., (2007), “Evaluation of the intrinsic photocatalytic oxidation kinetics of indoor air pollutants”, Environ. Sci. Technol., 41, 2028-2035.

SAUER M.L., OLLIS D.F., (1994), “Acetone oxidation in a photocatalytic monolith reactor”, J. Catal., 149, 81-91.

“Sick Building Syndrome and what you can do about it,” CaluTech, Inc. (2005)

SPADONI G., BANDINI E., and SANTARELLI F., (1978), “Scattering effects in photosensitized reactions”, Chem. Eng. Sci., 33, 517.

TAGHIPOUR F., MOHSENI M., (2005), “CFD simulation of UV photocatalytic reactors for air treatment”, AlChE J., 51, 3039-3047.

WANG W., KU Y., (2002), “The light transmission and distribution in an optical fiber coated with TiO2 particles”, Chemosphere, 50, 999-1006.

WANG W., KU Y., (2003), “The light transmission and distribution in an optical fiber coated with TiO2-particles”, Chemosphere, 50, 999-1006.

YU H.L., ZHAN K.L., ROSSI C., (2007), “Theoretical study on photocatalytic oxidation of VOCs using nano-TiO2 photocatalyst,” J. of Photochem. and Photobio A: Chem., 188, 65-73.