WO1997037936A1 - A photocatalytic reactor for water purification and use thereof - Google Patents

Abstract

The invention relates to a photocatalytic reactor suitable for water purification and comprising a housing having inlet means for feeding the reactor with polluted water, outlet means for removing the treated water from the reactor, at least one light conductor means and at least one UV-source capable of emitting UV-light into and through the light conducting means. Preferably, the light conducting means is a hollow tube made of quartz glass or pyrex glass. Most preferably, the photocatalytic reactor is provided with a UV-tube light which is coated on its surface with a photocatalyst. The photocatalytic reactor is suitable for the purification of waste water containing toxic organic pollutants.

Description

A PHOTOCATALYTIC REACTOR FOR WATER PURIFICATION AND USE THEREOF The invention relates to a photocatalytic reactor which is suitable for the purification of waste water containing organic pollutants, in particular toxic pollutants.

Background of the invention

Water contaminated with traces of toxic organic compounds is a common problem in Europe and throughout the world. In recent years, application of advanced oxidation technologies involving strongly oxidising hydroxyl radicals have gained increasing interest in the treatment of industrial wastewater, contaminated ground and drinking water [Legrini et al. , 1993; Hoffmann et al., 1995]. In particular, heterogeneous photocatalytic degradation in the presence of a semiconductor catalyst (i.e. a photocatalyst) has been shown to be a promising method for the destruction of toxic chemicals [Ollis et al., 1989; Fox and Dulay, 1993]. The appeal of this process technology is the prospect of partial or complete mineralization of pollutants to environmentally harmless compounds.

In recent years, interest has focused on the use of TiO₂ as a photocatalyst for the destruction of polluting materials [Hagfeldt and Gratzel, 1995]. Semiconductors are flexible reagents for waste treatment as the catalysts are not consumed during the reaction and can carry out oxidations and reductions simultaneously. Activation of the catalyst is achieved through the absorption of a photon of ultraviolet bandgap energy resulting in the formation of electron donor (reducing) sites and electron acceptor (oxidising) sites. The carbon containing pollutants can be oxidised either partially or completely to carbon dioxide and water while the other elements bonded to the organic compounds can be converted to anions such as nitrate, sulphate or chloride.

The semiconductor catalyst can be employed either in a colloidal form or as an immobilised film . Each of these two classes has advantages and disadvantages. In photoreactors operated with catalyst particles as a slurry [Matthews, 1992], the reaction rate is predominantly determined by the light intensity on the surface, the quantum efficiency of the catalyst and the adsorption properties of the reacting and non-reacting components in solution. However, the use of suspensions requires the separation and recycling of the ultrafine catalyst from the treated liquid and the depth of penetration of UV light is limited because of strong absorption by both catalyst particles and dissolved organic species. Above problems could be avoided in photoreactors in which catalyst particles are immobilised [Zeltner et al., 1993]. However, immobilisation of a semiconductor on a support generates a unique problem. The reaction occurs at the liquid-solid interface and mass transfer from the bulk of the liquid to the catalyst surface may now play an important role in the overall rate.

Research on photocatalytic processes to achieve mineralization of both organic and inorganic pollutants has been widely tested, but only at laboratory scale [Ollis et al., 1989; Legnni et al., 1993;. Mills et al., 1993]. However, to date no viable pilot plant has yet been developed successfully using this technology. There is much engineering and scale-up issues that must be addressed before commercial process units can be realised.

Problems in the development of an effective photocatalytic reactor.

The problem of scale-up of multiphase photocatalytic reactors is considerably more complex than that of conventional chemical reactors or homogeneous photoreactors. The need to use a solid catalyst makes the whole problem quite complicated as another phase is added to the system. Moreover, without photons of appropriate energy content the catalyst shows no activity. The demand for catalyst illumination is in fact an extra engineering factor in the reactor design besides conventional reactor scale-up complications such as mixing and mass transfer, reactant-catalyst contacting, flow patterns, reaction kinetics, catalyst installation, temperature control, etc. The illumination factor is of utmost importance since the amount of catalyst that can be activated determines the water treatment capacity by the reactor. The volume of photocatalytic reactor can be expressed as

where Q is the volumetric flow rate (m³/s), C_in is the inlet pollutant concentration (mol/m³), X is the fractional conversion desired, k is illuminated catalyst surface area in contact with reaction liquid inside the reactor volume (m²/m³) and R is the average mass destruction rate (mol/m²/s). Hence, smallest reactor volume will result when k and R are as large as possible for fixed desired values of capacity, Q, inlet concentration, C_in. and fractional conversion desired, X. R is a reaction specific parameter as it expresses the performance of a catalyst for the breakdown of a specific model component, while k is a reactor spec if ic parameter representing the amount of catalyst inside the reactor that is sufficiently illuminated so that it is active and is in contact with the reaction liquid.

One major barrier in the development of photocatalytic reactor is that the reaction rate is usually slow compared to conventional chemical conversion rates due to low concentration levels of the pollutants. Improvement of the breakdown rates would lead to the need of a less demanding amount of catalyst to be illuminated and therefore, a smaller reactor volume.

Another crucial hurdle in the development of effective photocatalytic reactor design is the need to provide large amounts of active catalyst inside the reactor. Even though the effective surface area of the porous anatase catalyst coating is high, there - in the case of an immobilized photocatalyst - can only be a thin coating (about 1 μm thick) applied to a surface. Thus, the amount of immobilized active catalyst in the reactor is limited and, even if individual degradation processes can be made relatively efficient, the overall conversion efficiency will still be low. This problem severely restricts the processing capacity of the reactor and the necessary time required to achieve high conversions are measured in hours, if not days.

In view of the above considerations, applicant is of opinion that new reactor configurations must address the two most important parameters, namely, light distribution inside the reactor through the absorbing and scattering liquid to the catalyst, and providing high surface areas for catalyst coating per unit volume of reactor. The new reactor design concepts according to the invention must provide a high ratio of activated catalyst to illuminated surface and also must have a high density of active catalyst in contact with liquid to be treated inside the reactor. For comparison of design efficiency of different photocatalytic reactors in terms of their efficacy to install as much as activated catalyst per unit volume of reaction liquid inside the reactor, applicant proposes a parameter k, namely illuminated specific surface area, representing the total illuminated surface area of catalyst within the reactor that is in contact with the reaction liquid.

A number of photocatalytic reactors have been patented in recent years but probably none has so far been developed to pilot scale level. Based on the arrangement of the light source and reactor vessel, all these reactor configurations fall under the categories of immersion type with lamp(s) immersed within the reactor, external type with lamps outside the reactor or distribut ive type with the light distributed from the source to the reactor by optical means such as reflectors or optical fibres. Majority of reactors patented are in fact variation of the classical annular reactor of immersion or external type in which catalyst is immobilised on reactor wall (Sato, 1992; Taoda, 1993); on pipes internally (Matthews, 1990); on ceramic membranes (Anderson et al., 1991). on glass wool matrix between plates (Cooper, 1989). on semipermeable membranes (Miano and Borgarello, 1991; Oonada, 1994). embedded in water permeable capsules (Hosokawa and Yukimitsu, 1991). on a mesh of fiberglass (Henderson and Robertson, 1989). on beads (Heller and Brock, 1993). on fused silica glass fibers (Hofstadler et al., 1994). on porous filter pipes (Haneda, 1992), on glass fiber cloth (Masuda et al., 1994). etc. The reactors are either helical (Ritchie, 1991). spiral (Matthews, 1988), shallow cross flow basins (Cooper and Ratcliff, 1991) or optical fiber (Wake and Matsunaga, 1992). However, all these reactor designs are limited solely to small scales by the low values of the key parameter, k. The only way to apply these systems for large-scale applications are by using multiple units of more than 10,000 for treatment capacity of 10 m³/hr.

Summary of the invention

In order to overcome the abovementioned deficiencies inherent to conventional photocatalytic reactors Applicant has designed a photocatalytic reactor comprising a housing having inlet means for feeding the reactor, outlet means for removing the treated water from the reactor, at least one light conductor means providing a large light-emitting surface illuminating the photocatalyst present in the reactor and at least one UV-source capable of emitting UV- light into and through the light conductor means.

Brief description of the Figures

The invention will be better understood by reference to the attached figures:

Fig. 1 : Top view of a distributive type photocatalytic reactor comprising glass slabs or rods as light conductors (1), lamp-reflectors (2) and wherein the direction of the liquid stream is indicated by arrows.

Fig. 2 : Distribution of light rays in a glass slab or rod [1:

glass, 2: air/water/catalyst, 3: air, n₁ = 1,5, n₂ = 1

(air), 1,33 (water), 2,8 (titania); n₃ = 1, θ_c = 41.8°

(glass-air), 62.5° (glass-water)

Fig. 3 : Distribution of light rays in a hollow tube [1: air, 2:

glass wall, 3: catalyst titania; n₃ = 1, n₂ = 1.5. n₃ =

2.8]

Fig. 4 : Schematic diagram of a Multiple Tube Reactor (11) (MTR) comprising a reflector (12), a UV light source (13), a lens (14), a number of TiO₂ coated hollow glass tubes (15), an inlet (16) and an outle (17).

Fig. 5 : Calculated light density diagram for a smooth (solid circles) and a roughened glass wall surface (squares).

Fig. 6 : Experimentally obtained light density diagram for a tube glass surface coated with aluminium [outside diameter of tube = 0.006 m; tube wall thickness = 0.001 m]. Fig. 7 : Radial Light Intensity Profile I_radial (X) for a glass tube, where the incoming light intensity I_o = 500 W/m²; angle of incidence 3 . outside tube diameter d_o : 0.006 m, tube wall thickness w = 0.001 m, surface roughness factor γ : 5-5 m^{- 1} and absorption coefficient for the medium B: 4.5 m^-1.

Fig. 8 : Effect of input light intensity on I_rad (X) profile.

[α = 3 deg; d_o = 6 mm; w = 1 mm; β = 3 n^-1; γ = 6 m^-1] Fig. 9 : Effect of angle of incidence α on I_rad (X) profile.

[I_o = 500 W/m²; d = 0.006 m; W = 0.001 m, γ = 5-5 m^-1, β = 4. 5 m^-1]

Fig. 10: Effect of tube diameter on I_rad (X) profile.

[I_o = 500 W/m²; α = 3 deg; W = 0.001 m, β = 4 5 m^-1 , Y = 5.5 n X

Fig. 11: Effect of degree of roughness γ on I_rad (X) profile

[I_o = 500 W/m²; α = 3 deg; d = 0.006 m; W = 0.001 m; β =

4.5 m^-1]

Fig. 12: Effect on pollutant concentration on degradation rate.

[I_o = 500 W/m²; α = 3 deg; d = 0.006 m; W = 0.001 m; β = 4.5 m^-1 , γ = 5.5 m^-1]

Fig. 13: Effect of the light intensity, angle of incidence, tube diameter and wall thickness on the overall performance of the tube reactor.

[I_o = 500 W/m²; α = 3 deg; d_o = 0.006 m, W = 0.001 to; β = 4.5 m^-1 Y= 5.5 m^-1 C_in = 10 ppm]

Fig. 14: Flow diagram of an experimental set-up comprising a reactor

(21), a cuvette (22), a monitor (23), a recorder (24), a bypass line (25), a 3-way valve (26), a Dye-reservoir (27), a water reservoir (28) and a pump (29).

Fig. 15: Photocatalytic degradation of SBB dye for various initial concentrations

Reactor specifications: volume of reactor = 1.23 10^-3 m³ catalyst surface area = 0.51 m²;

catalyst amount = 3-8 10^-3 kg/m²;

Experimental conditions: flow rate = 3.0 10^-5 m³/s volume of liquid treated = 7.5 10^-4 m³.

Fig. 16: Schematic diagram of Tube Light Reactor (TLR) . Fig. 17: Schematic view of the dip-coating apparatus.

Fig. 18: Catalyst layer thickness with number of times coated in the dip-coating apparatus.

Fig. 19: Graph showing adsorption of the dye in the system, dark reaction and photocatalysis.

Fig. 20: Influence of external mass transfer (flow rate) on the overall rate.

Fig. 21: Mass transfer rate, km, reaction rate, K_r. overall rate,

K_overall and fractional contribution of external mass transfer, φ , towards overall rate versus Reynolds number.

Fig. 22: Photocatalytic degradation of SBB dye for various initial concentrations

Reactor specifications: volume of reactor = 5.36 10^-4 m³ illuminated catalyst surface area = 0.15 m²;

number of lamps = 21; catalyst amount = 4.0 x 10^-3 kg/m ²

Experimental conditions: flow rate = 1.67 10^-5 m³/s; volume of treated liquid = 4.05 10^-4 m³.

Detailed description of the invention

An embodiment of a photocatalytic reactor according to the invention is shown in Figure 1. The reactor is a rectangular vessel in which light conductors as glass slabs (or rods) coated on its outside surface with catalysts are embedded vertically. The lamps together with reflectors are placed on two sides of the reactor while liquid enters and exits from the other two sides. Light rays enter the conductors through one end are repeatedly internally reflected down the length and at each reflection come in contact with the catalyst present around the outer surface of the conductors. Thus, the conducting materials might be considered as a means of light carrier to the catalyst. Since the ratio of the surface area on which catalyst is present to the light entering area could be as high as 500, evidently a very large catalyst area can be illuminated. Moreover, the fact that a large number of such light conducting material may be packed inside the reactor, the configuration provides a high total light transfer area and allows for a higher illuminated catalyst area per unit reactor volume. Densely packing of the reactor with light conducting objects not only increases surface to volume ratio but also may reduce the effective mass transfer diffusion length for the pollutant to catalyst surface.

The vital issue in the distributive type of reactor concept is how to introduce light from the external source efficiently into the light conductors, and likewise, how to get it out again at the proper location and in the appropriate amount. The predominant obstacle Applicant came across in the use of glass slabs (or rods) as the light conducting object is the occurrence of total internal reflection. It transpires when light travels from denser to rarer medium and is determined by the critical angle given by

where n₁ and n₂ are the refractive indices of the denser and rarer medium respectively. In the case of light travelling from air, to glass to air (or water) as depicted in figure 2 where medium 1, 2 and 3 are glass, air (or water) and air respectively, the angle θ will always exceed the critical angle, θ_c, for the interface between glass and air (or water) irrespective of angle of incidence, α (0 to 90°). In other words, all the light rays that are entering through the top surface will experience the phenomena of total internal reflection and will come out axially rather than emerging from the lateral surface. However, refractive index of TiO₂ (between 2.4 to 2.8) is higher than that of glass (about 1.5) in the wavelength range of 200 to 400 nm and it is likely that total internal reflection will not take place when the glass surface is coated with titania. However, if coating consists of small spheres of catalyst particles being in the range of 0.001-1.0 μm dispersed along the surface, the actual glass-titania interface will be small as most of the glass surface will still be in contact with water. For this reason it is best to avoid the occurrence of total internal reflection entirely.

One way of avoiding total internal reflection is by surface roughening. Moreover, surface roughening further assist - if necessary - in achieving a better catalyst adhesion to the substrate. Both are indeed found out to be the case experimentally. In fact, when lateral surface was roughened by sand blasting most of the light emerged within few centimetres and hardly any light remained thereafter in the axial direction. This is not only because roughenings desist total internal reflection phenomena but also UV- transparency of most light conducting material is very poor. Table 1 shows measured transmission intensity values for different qualities of glass material.

When the measured value for a path length of 0.05 m is extrapolated to a path length of 0.5 m assuming exponential decay, it appears that only 2% to 2.5% of light remains for Corning and Tempax type of glass whereas 15-52 remains for Quartz. Consequently, use of Quartz as light conductors will naturally help to overcome the abovementioned light transmission problem.

The total internal reflection problem may also be effectively avoided when the surface's light has to pass through conductors which are parallel instead of perpendicular. One such configuration is a hollow glass tube coated on its surface with semiconductor catalysts as shown in figure 3. Although, total internal reflection could be avoided completely in this configuration the angle of incidence of light will be a critical factor. When light falls on the glass surface a part of it is reflected and the rest is transmitted. The ratio between the reflection and transmission of light is a strong function of angle of incidence. When the light beam is nearly parallel with the surface (α close to Oº), most of the light is reflected and exits axially rather than laterally while for light rays with α close to 90° most of the light will emerge laterally within few centimetres and barely any light will remain thereafter as reflection is only 4% for a glass-air interface. Table

2 shows calculated values of percentage of light that remains from top for different values of α.

TABEL 2

The percentage of light that is still present at x m from top for propagation of a single light ray in a hollow space between glass walls.

Clearly, angle of incidence has an immense influence on the amount of light that will come out from the lateral and axial directions along with on the distance between successive contact with the glass surface. Therefore, it is apparent that in the design of reactor based on hollow tubes light must be guided into the conductors at a precise angle through a combination of optical lenses and reflectors.

A mathematical model to estimate light distribution in a single hollow tube is developed by Applicant and subsequently, the effect of various parameters on the radial light intensity profile, I_red(x). The overall reactor performance was then evaluated to determine the effect of specific radial light distribution on the breakdown performance for the model component. Special Brilliant Blue (SBB), a textile dye. Based on the model results, optimal parameters were evaluated to help guide reactor design and construction, and finally experiments were performed to find out the prospect of the reactor for further development. A model for evaluation of MTR system

Reactor description : A new concept using hollow tubes as light conductors satisfying most of the design criteria for the scale-up of photocatalytic reactor are considered. The new design allows for a large surface area of catalyst within a relatively small reactor volume compared to classical reactor consisting of several lamps externally surrounding a cylindrical vessel. A 70 to 100 fold increase in surface area per m³ reactor volume can be obtained over classical reactor design. The hollow tube might be considered as a pore carrying light to the catalyst. In this novel configuration, light rays entering through one end of the hollow tube are repeatedly internally reflected down the length of the tube and at each reflection came across the annular catalyst coating present around the outer surface of the tube. Since the ratio of the cylindrical surface area to the circular end surface (light entrance area) of the tube is of the order of 500, evidently a very large area can be illuminated. This new configuration according to the invention provides a high light transfer area and allows for a higher illuminated catalyst area per reactor volume. Another potential advantage of distributing light within hollow tubes is that light does not have to pass through the reactant and product phases in the reactor. This type of reactor is abbreviated as "MTR" for Multiple Tube Eeactor.

The basic set up for which the present model has been developed is a cylindrical vessel containing the reaction liquid within which an array of hollow tubes is placed (figure 4 ) . Light enters the tube bundle from one end with an input light intensity of I_o and an angle of incidence of α. The light bundle is considered homogeneous so that all light rays enter at one specific angle and have the same intensity. Under this condition the light distribution in one tube will be representative of the entire unit.

Formulation for the estimation of light distribution in a single tube

Light from top is entering a single tube through two regions.

One region is the hollow shaft and the other is through the annular glass of thickness w. The behaviour of the light rays is quite different in these two regions, the glass wall and the hollow shaft, of the tube and are considered separately.

(1) In the shaft : Light propagation is determined by the amount that reflects (or enter the glass wall through refraction) at every bounce of the light beam/ray with the glass wall. The distance between two successive bounces, D, can be calculated from the shaft diameter, d_i, and the angle with which light beam strikes the glass surface, θ.

At any axial position, x, the number of bounces that has already occurred can be calculated from

The portion that reflects at every bounce is a function of the incident angle, θ, and percentage reflected can be calculated from the following equations :

If it is assumed that the perpendicular and parallel plane polarised components of the light are present in equal proportions, then for angle of incidence of up to 20° the average reflectivity can be approximated by the following expression

R = exp[-0.0905θ] (6)

The amount of light remaining in the shaft at any axial position x is then given by

I_{ax, s} (X ) =I_DRⁿb^(x) (7)

(2) In the wall : If both surfaces of the tube wall are smooth, then all light rays entering from top into the annular region will be totally reflected as the glass is optically denser than either air or water. Hardly any light rays will be refracted and transmitted through the outside wall into the environment. But, when the outer surfaces of the hollow tubes are roughened or coated with TiO₂ catalyst, light transmission situation is quite different. Total internal reflection will break down and a portion of the light will be refracted and will come out from the tube surface in radial direction. In addition, the light rays entering this wall region from top will also be absorbed by the glass. In the case of light propagation in the wall, the effects of the outer wall condition and the absorption are likely to be much larger than angular influences, and hence for simplicity, the effect of entrance angle is neglected. Since the light energy losses due to absorption in the glass medium can be characterised by an absorption coefficient, β, it is mathematically convenient to define a similar wall effect coefficient, γ, for surface roughening, γ relates to the portion of light that is passing the outer wall of the tube instead of being reflected. Hence, γ = O, can be considered for the condition of total internal reflection (smooth uncoated glass surface), and high value of γ corresponds to high degree of roughness with most of the light energy leaving the tubes by immediately passing through the tube wall surface upon contacting with the glass wall. The light intensity that remains at any position x in this region can then be given by the following exponential expression l_ax,w(X) = I_o exp[-(β+γ)x] (8)

(3) Total axial light intensity profile in the tube : The total axial light intensity profile in the tube can be found by adding the respective components from the wall and the shaft region but must be weighted to their respective cross-sectional area. This is given by

where d_o and d_i are the outside and inside diameter of the tubes. (4) Light coming out of the tube wall surface onto the catalyst :

The outgoing light can be calculated from the light energy balance over a segment of dx.

In the above equation it is assumed that the light energy difference between x and x+dx is only due to light energy coming out radially and absorption by the glass medium. The absolute values of the radial intensity will be much lower than the axial intensity because of the large differences in surface areas of the tube cross section and its outer peripheral surface wall. The ratio is given by 4L/d_o. With tube diameters about few millimetres and tube length in the order of half a meter, this ratio is in the range of 100 and 1000. It means that in the MTR concept of photocatalytic reactor high intensities are needed at the top of the tubes.

(5) Test of the light distribution model with experimentally measured data : In figure 5, I_ax(x) obtained from the equation 9 is compared with experimentally observed data. In the experimental set- up an optical lens is used to focus the light rays into the tube precisely within a few degrees. In the figure, solid circles and squares represent experimental measured data for axial outcoming light for the smooth and roughened (sand blasted) glass wall surface respectively for 0.006 m diameter tube while solid line represents results from model calculation. It is apparent from the figure that the model fits the measured data reasonably well. Figure 6 shows experimentally measured I_ax(x) results when the outer glass surface is coated with highly reflecting aluminium. In this case the difference in measured light intensity value will be entirely due to absorption by the glass material. When the measured data was fitted with an exponential function the absorption coefficient, β, for the glass material obtained was 4.53 m^-1.

Results

(1) Typical I_rad(x) profile : The effect of design parameters on I_rβd(x) are studied by varying each parameter one at a time keeping the others fixed at a reference value. The reference set of values are taken as follows : the incoming light intensity, I₀, as 500 W/m² ; outside tube diameter, d_o, as 0.006 m ; tube wall thickness, w, as 0.001 m ; angle of incidence, α, as 3 degrees ; surface roughness factor, γ, as 5 m^-1 ; absorption coefficient for the medium, β, as 4 .5 m^-1 . Figure 7 shows the radial intensity profile for the above set of reference values. From the figure it can be seen that even though the input light intensity was 500 W/m², the radial outcoming light intensity is only between 3-7 W/m² (near the top) and 0.3 W/m² (near the bottom) for tubes of 0.5 m long. This low value of radial intensity can be recognised if we consider the two important factors. Firstly, most of the light is leaving axially (for example, about 50% in figure 5) from the end of the tube and is practically lost, and secondly, there is a large difference in surface areas of inlet and outlet of light energy. The second factor, the large value of the ratio kL/d_o, have been discussed already. The first factor can be improved if we coat the bottom of the tube with aluminium (or by placing a mirror) to reflect most of the light that is coming out axially back into the tube. In this way, the I_rad value at the end of the tube could be increased from the present value of 0.3 W/m².

(2) Effect of input light intensity : Figure 8 shows the effect of input light intensity, I_o, on the radially outcoming light intensity profile along the length of the tube. As expected I_o has a large impact on I_rad(x). However, use of a too high input light intensity is not very practical as only about 0.7% of input light intensity is coming out as I_rad. If we assume, the diameter of the reactor as 0.1 m, then light power on top of the reactor must be at least 4 W for a required 500 W/m² of I_o of λ_< 365 nm. If the lamps spectral distribution contains 152 of UV-A, then the required lamp power is 27 W. In other words, when a 27 W lamp is placed at the focal point inside a parabolic reflector that spans 0.1 m, the intensity of UV-A content of the light will be 500 W/m² on top of the reactor.

(3) Effect of angle of incidence : The effect of light incident angle is important since it sets the criteria for the arrangement of the tube bundle inside the reactor and/or the manner light source, lens and parabolic reflector are to be placed around the reactor. Figure 9 shows the effect of angle of incidence, α, on I_rad(x). When α is equal to 5°. more light is coming out radially within the first few centimetres and hardly any light is left within the shaft thereafter. Whereas, at very low angles (α equal to 1°) when the light rays enter the tube almost in parallel along the centreline most of the light is being reflected rather than refracted by the tube wall. Therefore, for angles lower or higher, either I_rad(x) is too high in the beginning (with very low values at the end of the tube) or I_rad(x) is too low all along the tube. Hence, there exists an optimal value for α for which the I_rad(x) profile is more evenly distributed and thereby will result in a more uniform catalyst activation. From the figure it appears that α around 3° may be the best for obtaining a most even distribution for I_rad(x). (4) Effect of tube diameter : Figure 10 shows the effect of tube diameter on I_rad(x). Increase of tube diameter increases the value of I_rad(x) profile. This result can be explained easily if one recognises that wider tubes (d_o equal to 0.01 m) receive relatively more total light from top due to larger surface area available for light inlet compared to smaller tube diameters (d_o equal to 0.003 m). High values of I_rad(x) for large values of d_o will activate the catalyst more and will give better performance for each tube for conversion of pollutants . However, this does not necessarily mean that we must select tubes with large d_o. High performance of a single tube may be well compensated by the fact that less tubes can be placed within a reactor unit and therefore the total amount of catalyst present within the reactor will also be limited. This has indeed found to be the case and is discussed below. (5) Effect of surface roughness : The effect of the surface roughening factor, γ, on the I_rad(x) profile is shown in figure 11. Tube surface wall roughening appeared to have a similar effect as varying the angle of incidence. When the wall surface is roughened more (γ=l6.5 m^-1), more light will come out within the first few centimetres of the tube length and hardly any light remains thereafter, whereas when the tube surface is too smooth (γ = 2.75 m^-1 ) hardly any light will emerge radially from the surface as most of the light will be piped down the tube and will come out axially at the end of the tube. Consequently, there may be an optimum value of Y for which I_rad(x) will as evenly be distributed as possible. In addition, surface roughening to some extent is advantageous for catalyst fixation and its durability. Estimation of overall reactor performance : The rate of degradation of SBB dye is given by {

where k_s = O .38 μmol/m²/s ; a = 0. 18 m²/W ;

b = 0.85 ; K = I8.5 m³/mol. The rate of degradation of the dye, R[I_rad(x),C], along the axial position of tube was calculated for a fixed value of pollutant concentration. Figure 12 shows rates as a function of axial position for different pollutant concentrations. The rate of degradation is reasonably high at the beginning and decreases gradually towards the e

nd of a 0.5m long tube. With the decrease of dye concentration from 50 ppm to 1 ppm, the profile of R[I_rad(x),C] decreases progressively. Consequently, it will take a long time to remove the last traces of pollutants.

For an estimation of the chemical activity that results from the light distribution in a single tube, an average reaction rate can be calculated from

by averaging the rate over the tube length. The average tube performance, ATP, the average breakdown performance of one tube, can be calculated from

ATP (μmol/s) = R_avg * A_t (13) where A_t is area of each tube. Since all tubes are considered equal within a reactor unit, the overall performance of the reactor unit, RPU, is simply ATP multiplied by N_t , the number of tubes in one unit. Results

To investigate the effect of different parameters on overall performance of the reactor, a diameter of 0.2 m is considered. Applicant further assumed that 60% of the reactor volume is occupied by tubes. In Figure 13, the effect of various parameters on the overall performance of the reactor is demonstrated through bar diagrams.

(a) Effect of input light intensity : I_o has a huge impact on RPU as expected. However, use of a too high input light intensity is not very practical as only a small fraction of input light intensity can be forced into a single tube and the increase of I_rad for a unit increase in I_o is not very substantial. Therefore, I_o must be selected as high as possible but practically reasonable.

(b) Effect of angle of incidence : The effect of light incident angle is important since this sets the criteria for the light source bundle and the (parabolic) reflector. The overall reactor performance concerning angle of incidence shows an optimum value for RPU. At very low angles (α = 1º) RPU has a low value, goes through a maximum for α value of about 3º and then tails off to a constant value for α greater than 5º. When α is high (greater than 5°), the first few centimetres receive much more light, thereby yielding a higher performance for the total system which in fact is not fully compensated by the better performance of the lower regions for a values between 3 to 5 degrees. At very low angles when light enters the tube almost in parallel along the centrelines, not enough light touches the wall of the shaft. (c) Effect of tube diameter : Figure 10 shows that the performance of one tube is better when we select a tube diameter as large as possible. This is because a wider tube catches relatively more total light on top so that I_rad(x) has higher values. However, the effect on overall reactor performance appears to be overshadowed as more smaller tubes can be fitted into a reactor unit thereby increasing k. For example, when 0.01 m diameter tubes are used, 240 of them can be placed in a reactor unit with a k value of 600 m²/m³. Whereas, 2700 tubes can be installed into the same reactor unit when 0.003 m diameter tubes are selected resulting in a k value of 2000 m²/m³. The two counteracting effects (inlet area and k) result in an optimum tube diameter for some particular set of reference values. Further there are two important advantages in the use of quartz tubes. First, transmission of quartz is high and therefore, less light will be absorbed by the glass wall. Secondly, strengths of quartz is superior to Pyrex and hence, a smaller wall thickness can be selected.

(d) Effect of wall thickness : With increasing wall thickness from 0.0005 m to 0.002 m, the overall performance deteriorates. This can be understood from the fact that when wall thickness is increased keeping d_o fixed, comparatively more light is lost due to absorption within the glass media. Hence, wall thickness of the tube should be selected as small as possible.

Minimum light intensity required for catalyst activation:

Without photons of appropriate energy content the catalyst shows no activity. Therefore, it is important to estimate the minimum amount of light each catalyst particle must receive for activation. One mole TiO₂ requires 3.2 ev (≡ 380 nm wavelength light) or 308378 J/mol. If the average degradation rate of Special Brilliant Blue dye is taken as 0.15 μmol/m²/s then assuming 5% quantum efficiency, minimum energy required is about 1.0 W/m². If we assume that about 3.0 x 10^-3 kg catalyst can be immobilised per m², then energy required is about 0.33 kW per kg of catalyst. Hence, it is evident that not much energy is required to activate TiO₂ catalyst. Selection of lamp and reactor dimension : Based on the above study, a 40 W lamp contained in a parabolic reflector spanning 0.056 m is selected. The input light intensity of λ < 365 nm is 2436 W/m². If 0.5 m long tubes are used then for a packing density of 60% , 54 tubes of diameter 0.006 m can be placed within the reactor. The volume of the reactor, total surface area of catalyst and K are 1.23 x 10^-3 m³, 0.51 m², 1087 m²/m³, respectively. The performance rate of the reactor is 0.015 μmol/s when 10 ppm textile dye is treated and the initial efficiency of the reactor expressed as conversion per unit time per unit volume of reactor per unit energy input is 0.305 umol/s/m³/W. A further embodiment of the invention relates to a tube light reactor (TLR). More in particular these tube lights present in the reactor are fluorescent tube lamps of low wattage emitting lights in the wavelength of interest (λ < 380 nm) and an extremely small diameter i.e. preferably in the lower part of the range of 0.0045 - 0.025 m. These tube lights of Philips, for instance Philips NDF-U2 49-6W lamps, address many of the solutions to the problems that have restricted the development of technical scale photocatalytic reactors for water purification. These lamps are available in various shapes and lengths, and can be placed inside a reactor to form a variety of different configurations. Development of a reactor using these lamps will provide all the advantages of the above discussed MTR plus the additional advantage that catalyst could be activated at its highest level. In a preferred reactor configuration, the catalyst was deposited on the outer surface of these low wattage lamps. Thus, the main problem met in the development of the reactor based on multiple hollow tubes could be avoided. In the MTR design it is not possible to obtain a uniform light distribution along the total length of the tubes and therefore, it will restrict the maximum length of tubes that can be used inside a reactor and thereby the overall performance of the reactor. The lamps of Philips will eliminate this drawback of the MTR reactor concept as the present new design is capable of uniform light distribution over entire tube lengths. Of course, this is possible with classical lamps too. However, the new lamps allow for a 70 to 100 times larger surface area for catalyst per unit reactor volume compared to a classical reactor design.

Further it is apparent that these abovementioned lamps of Philips create great opportunities to build a much more efficient photocatalytic reactor for water purification, in the case of both an immobilized photocatalyst and a photocatalyst present in slurry form.

One such reactor configuration in which improvement in reactor performance can be anticipated is the one in which these narrow diameter lamps are wrapped with glass wool fiber and catalyst is then deposited on the resultant enhanced surface area. In this way, amount of catalyst can be increased even further for the same reactor volume and light distribution within the reactor thereby increasing the efficiency of the reactor.

The catalyst coating on the lamp's surface was found out to be durable and the activity of the catalyst did not deteriorate during the course of the experiments. However, it is well known that the lifetime of lamps deteriorates when it is operated immersed in liquid instead of air. One way of taking advantage of the lamps while avoiding reactor operation in which the lamps are immersed is to place one such tubelight lamp inside each of the hollow tubes described in the multiple tube reactor (MTR) configuration. In this way all the advantages of the MTR concept can be utilized while eliminating the basic drawback of uniform light distribution dilemma. Moreover, this will further increase surface area for catalyst and thereby pollutant conversion for the same electrical power input. Of course, this will increase the reactor volume as well but the increase in the amount of catalyst and proper utilisation of electrical energy might increase the overall reactor efficiency.

With respect to the photocatalyst to be used in the reactor according to the invention it is remarked that any suitable photocatalyst known in prior art may be used. Examples of such photocatalysts are anatase TiO₂, Pt-TiO₂, ZnO, CdS. CdSe. SnO₂, SrTiO₃, W0₃, Fe₂O₃, ZrO₂, Ta₂O₅, GaP, SiC and SnO. However, preferably anatase TiO₂ is used.

The photocatalysts indicated above have generally a surface area of 20-200 m²/g and a particle size of 0.001-1.0 μm.

In the case of an immobilized photocatalyst the light conductor is coated with 0.5-5 g/m² of photocatalyst.

The UV-source which is applied in the reactor according to the invention emits light in the range of 200-440 nm, preferably 320-380 nm. Examples of such UV-sources are high pressure and low pressure Hg vapour lamps and Xe arc lamps as well as the above-indicated tube lights. A further aspect of the invention relates to a process for degrading organic substances and microorganisms present in a polluted aqueous solution or suspension by means of the photocatalytic reactor according to the invention. According to this process the polluted aqueous solution or suspension is fed to a reactor according to the invention, wherein the photocatalyst is irradiated with UV-light and after being subjected to the influence of the activated photocatalyst - present in either an immobilized form or in suspension - the treated solution or suspension is removed from the reactor. Further extra oxidants such as O₂, H₂O₂ and O₃ may be used in the purification process. Examples of polluted aqueous solutions or dispersions are domestic, industrial and agricultural waste waters comprising one or more saturated or unsaturated hydrocarbons optionally comprising one or more heterogeneous atoms selected from the group consisting of halogen, like chlorine, bromine and iodine, N, S, O and P. Representatives of such compounds are (halo)hydro-carbons like benzene, toluene, chlorophenols, alcohols, acids, esters, ketones, amines as well as pesticides, etc.

The invention will be illustrated by means of the following examples showing the use of light conductors coated on its outside surface with a photocatalyst, but the invention may not be restricted thereto.

Example I

The reactor : Figure 4 shows schematic drawing of the novel bench- scale reactor system based on hollow tubes. The reactor consists of a cylindrical vessel of diameter 0.056m within which 54 hollow Quartz glass tubes of diameter 0.006m coated on its peripheral surface with catalyst were placed. The tubes were held securely within the reactor by two teflon end plates on which 54 holes were drilled. The reactor resembles that of a shell- and tube- heat exchanger with reaction liquid flowing through the shell-side over the outside surfaces of the coated tubes while light travels through the inside of hollow tubes. The tubes were arranged in triangular pitch of 0.007m thereby achieving a very high surface area per unit volume. The feed was introduced through four equally distributed ports at one end of the vessel thereby minimizing formation of any dead zones. Similarly, the exit flow from the reactor was collected through four ports distributed at the other end of the reactor. One end of each tubes were closed to prevent any reaction liquid entering the inside of the tubes. The closed ends were also coated with aluminium for better utilisation of axially exiting light.

Experimental set-up: A gear pump (Verder model 2036) was used for pumping the reactant between the reactor and the reservoir via a flow-through cuvette placed inside a universal photometer for continuous on-line measurement of the model component (figure 14). Two three-way glass valves were used between the water and specially designed reactant reservoir for initial zero setting of the analytical instrument before the start of an experiment, introduction of the reactant into the system, elimination of bubbles formed during experiment, and final flushing of the entire system.

The light source (Philips GBF 6436, 12V, 40W) used was a low voltage halogen lamp optically positioned in a light weight highly glossy anodised aluminium reflector spanning 0.056m for a clearly defined beam spread. In addition, a condenser lens of focal length 0.04m were placed between the lamp and the reactor to obtain light beam at a half intensity beam angle between 2 to 4 degrees.

Catalyst: Degussa P25 grade TiO₂ was used as catalyst for all the experiments. The crystalline product is nonporous primarily in the anatase form (70:30 anatase to rutile) and is used without further treatment. It has a BET surface area of (5-5 ± 1.5) * 10^-1 m²/kg and crystallite sizes of 30 nm.

Catalyst immobilization: For better catalyst fixation and its durability, the glass surface of the tubes on which titania was deposited was roughened by sand blasting. This makes the catalyst surface uneven but increases the strength and amount of catalyst per unit area that could be deposited. The lamps surface were coated with catalyst in a dip-coating apparatus. The immobilisation method employed is described in Example II.

Model reactant component: The model reactant component used was a brightly colored water soluble acid dye, Special Brilliant Blue (SBB, MW 812), laboratory reagent grade (in 20% solution) of which was obtained from Bayer (catalogue number 42735). This reactive dye was found to be an excellent model component for characterization of a photo-catalytic reactor. Analysis: Changes in SBB dye concentration were measured on-line by flowing a bypass stream of the dye from reactor outlet continuously through a bottom loader flow-through cuvet (Hellma, path length 0.001m) placed inside a Colorimeter (Vitatron Universal Photometer 6000) and was recorded continuously by a Kipp & Zonen (Model BD80) recorder.

Experimental procedure: At the start of every experiment the reactor was rinsed with Milli-Q water (deionized water obtained by using a Milli-Q water purification- system from the Millipore Corporation) for several times before zero-setting the analytical instrument. The reactor and connecting lines were then filled with the dye solution with the help of the three-way valves and it was ensured that no air bubbles remained in the system. The change in the dye concentration was continuously analysed and recorded. New silicon connecting tubes and fresh catalyst were found to adsorb the dye for about an hour, but no noticeable adsorption by the entire system was observed afterwards. Light was turned on only when the colorimeter reading was stabilised. Results

The catalyst coating on the lamp's surface was found out to be durable and the activity of the catalyst did not deteriorate during the course of the experiments. Figure 15 shows experimental results in the multiple tube reactor for the photocatalytic destruction of the SBB dye for various starting concentrations. Experimental results show that photocatalytic destruction of the dye pollutant is possible in the present configuration and it also reveals that 90% of the pollutant was degraded in about 100 minutes. This is in spite of the fact that the reactor was far from optimum. In fact, performance of the reactor can be instantly improved by decreasing the length of the hollow tubes used. At present, 0.5m long tubes have been used and it is likely that the catalyst is almost inactive near the end of the tube away from the light source.

Example II

The reactor : A bench scale reactor system using an extremely narrow diameter U-shaped fluorescent lamp was constructed. The schematic diagram of the reactor is shown in figure 16. The reactor consists of two parts made of stainless steel. One part consists of a flat plate (0.132m x 0.0l6m) with 21 holes onto which another plate (0.248m x 0.132m ) was welded. Twenty-one U-shaped lamps were placed around the plate and its end extended through the holes for electrical connections. Electrical wires were connected to the novel lamps through copper holders that are screwed around the lamps end. This part acts as a clamp for the lamps and the unit was used as a holder for both in dip-coating lamp's surface with catalyst and photocatalytic degradation experiments. The second part of the reactor assembly was a rectangular vessel containing the reaction liquid into which the other part can be placed. The feed was introduced at the top of the vessel and was equally distributed over the width of the reactor through 5 inlet ports. Similarly, the exit flow from the reactor was collected through 5 ports distributed on the other side of the reactor. The effective illuminated surface areas of the catalyst and the volume of the reactor were 0.15 m² and 5.36 x 10^-4 m³, respectively. The parameter k, defined as total illuminated catalyst surface area that is in contact with reaction liquid per unit volume of liquid treated in the reactor volume, was equal to 6l8 m²/m³.

Lamps: The U-shaped lamps (Philips NDF-U2 49-6W) used had a length of 0.498m and an extremely narrow diameter of 0.0045m. The lamps operate at 1020 Volts and the power demand was 6 Watts of which 15% was in the UV-A region thereby the incident light intensity (of λ < 380 nm) on catalyst particles was 127.8 W/m². Experimental set-up: A gear pump (Verder model 2036 with a maximum flow rate of 3.0 x 10^-5 m³/s) was used for pumping the reactant between the reactor and the reservoir via a flow-through cuvette placed inside a universal photometer for continuous on-line measurement of the model component (figure 14). Two three-way glass valves were used between the water and the reactant reservoir for initial zero setting of the analytical instrument before the start of an experiment, introduction of the reactant into the system, elimination of bubbles formed during experiment, and final flushing of the entire system. The reactor assembly was placed inside a thermostat bath to prevent overheating of the reaction liquid.

Catalyst: Degussa P25 grade TiO₂ was used as a catalyst for all the experiments. The crystalline product is non-porous, primarily in the anatase form (70:30 anatase to rutile) and is used without further treatment. It has a BET surface area of (5.5 ± 1.5) x 10⁴ m²/kg and crystallite sizes of 30 nm in 0.1 μm diameter aggregates.

Catalyst immobilization: For catalyst fixation and its durability, the glass surface of the lamps on which titania was deposited was roughened by sand blasting. This makes the catalyst surface uneven but increases the strength and amount of catalyst per unit area that could be deposited. The lamp's surfaces were coated with catalyst in the dip-coating apparatus (designed and constructed by the Applicant) (Figure 17). This is a completely automated equipment capable of immobilizing catalyst onto a variety of different shaped and sized substrates to any desired thickness by way of successive dipping of the objects into a suspension at a controlled speed that can be varied between (0.4 - 4.0) x 10^-4 m/s. Four 250 W infrared lamps were attached to a clamp that can be moved both vertically and horizontally for instant drying of the coating. Figure 18 shows the amount of catalyst immobilised per unit area on the outside of a roughened glass tube surface with the number dipping. Model reactant component: The model reactant component used was a brightly colored water soluble acid dye, Special Brilliant Blue (SBB, MW 812), laboratory reagent grade (in 20% solution) of which was obtained from Bayer (catalogue number 42735). This is an excellent model component for characterization of photocatalytic reactor as the dye is reactive only in presence of both TiO₂ and UV light, biologically not degradable, and present in wastewater streams from textile industry. The complete oxidation reaction of SBB dye is as follows:

C₄₃H₇₁O₆N₃S₂Na + 64½ O₂ → 43 CO₂ + 32 H₂O + 2 H₂SO₄ + 3 HNO₃ (14) Analysis: Changes in SBB dye concentration were measured on-line by flowing a bypass stream of the dye from reactor outlet continuously through a bottom loader flow-through cuvet (Hellma, path length 0.001m) placed inside a Colorimeter (Vitatron Universal Photometer 6000) and was recorded continuously by a Kipp &. Zonen (Model BD8O) recorder. For concentrations up to 0.5 mol/m³ the calibration line obeyed the Beer-Lambert law with good precision and the absorptivity coefficient, € (at λ_max= 605 nm), was found to be 5000 m³/mol/m.

Experimental procedure: At the start of every experiment the reactor was rinsed with Milli-Q water for several times before zero-setting the analytical instrument. The reactor was then filled with the dye solution with the help of the three-way valves and it was ensured that no air bubbles remained in the system. The change in the dye concentration was continuously analysed and recorded. New silicon connecting tubes and fresh catalyst were found to adsorb the dye for about an hour, but no noticeable adsorption by the entire system was observed afterwards. Light was turned on only when the colorimeter reading was stabilised. The experimental result is shown in figure 19. The decrease in concentration in the first two parts can be attributed to the adsorption of the dye by the connecting tubing (no catalyst present) and fresh catalyst, respectively, while that of the last part was when light was turned on and actual photocatalysis occurred. Evidence of complete mineralisation : The degradation of the parent organic compound usually produces intermediates that may be stable, resistant to further oxidation and more toxic. Chemical oxygen demand (COD) experiments were performed to determine the degree of mineralisation and to ensure that in the present investigation the analytical measurement was not merely that of the decolorisation of the dye. COD measures the amount of oxygen required during the oxidation of a compound with hot acid dichromate and provides a measure of the oxidisable matter present. The COD for the SBB dye is given by

COD = y . [M_oxygen/M_{SBB dye}] . n = 2.542 y (15) where y is the concentration of the SBB dye in ppm, M_oxygen and M_{SBB dye} are the respective molecular weights and n for the dye is 64½. The COD was measured for various starting dye solutions, for liquid collected at the end of several experiments and for pure Milli-Q water. COD values obtained for the starting dye solution were always the equal to value calculated from equation (15) and that of the treated liquid and the value of the blank was zero. The COD value for the final treated liquid shows that complete mineralisation occurred and for the present model component colorimetπc analysis was justified as it not merely measured the decolorization of the dye, but the mineralisation of it.

Influence of external mass transfer It is evident that when catalyst is immobilized onto a surface the forced convective mass transport of reactants from a well-mixed liquid phase to the catalyst surface may play an important role. The overall measured rate can be dominated by surface reaction rate, mass transfer rate or by a combination of both. Indeed experiments with increase of the flow rate through the reactor caused an increase in the rate of disappearance of the dye (figure 20) showing influence of external mass transfer. Experimental results for photocatalytic degradation have been found in the literature to be described by typical Langmuir-Hinshelwood kinetics. However, recent kinetic studies by Applicant for the SBB dye showed that the rate is essentially first order for the concentration below 30 ppm. At steady state, various rates are the same and given by N = k_overall( C_o-O) = k_m(C_o-C_s) = k_r C_s (16)

where C_o, C_s are the concentrations in the bulk and very near the catalyst surface, respectively; k_m, k_r and k_overall are mass transfer rate, reaction rate and experimentally measured overall rate, respectively. Equation (16) simplifies to

The mass transfer rate can be calculated for a particular configuration and flow conditions of the system from known correlations involving the dimensionless Sherwood, Reynolds and Schmidt numbers. For the present problem, the mass transfer rate is calculated from the following correlation of flow across tube banks (Holman, 1981)

where Re = [Qd /A_minυ], Sc = υ / D, d is the diameter of the tubes,

D and υ are diffusion coefficient and kinematic viscosity respectively, Q is volumetric flow rate, A_min is the minimum flow area. The constants m, C and n for the present geometry and flow conditions are 0.77. 0.8 and 0.4, respectively. The above correlation is valid for 10 < Re < 10⁶. The various rates are plotted as a function of Reynolds number in figure 21. The figure indicates that mass transfer rate is of the same order of magnitude as that of the reaction rate and even if the reactor is operated at highest possible Reynolds number there will be considerable influence of mass transfer on the overall rate. In the figure, fractional contribution of external mass transfer towards the overall rate , φ , is also shown . Most of the experiments were performed at a flow rate of 1.67 x 10^-5 m³/s (≡ Re 167) .

Experimental results : Testing of the reactor containing the extremely narrow diameter lamps with catalyst coated on its outer peripheral surface was done to determine the performance of the lamps. Figure 22 shows experimental results for the photocatalytic destruction of the SBB dye for various starting concentrations. This figure reveals that 90% of the pollutant was degraded in about 30 minutes.

Comparative Example 1

In Table 3 reactor specifications and experimental conditions used for multiple tube reactors are compared with a classical annular reactor [Assink et al., 1993] and with a tube light reactor.

TABLE 3

Comparison of reactor specifications and experimental conditions applied in a classical annular reactor (CAR), tube light reactor (TLR) and multiple tube reactor (MTR)

In the classical reactor, the walls were coated with 8 layers of TiO₂ catalyst and were surrounded by 10 Philips TLK 40W/10R UV-A lamps. Whereas in the tube light reactor, 21 extremely narrow diameter (0.0045m) novel lamps of Philips coated with TiO₂ catalyst on its peripheral surface were used. From the table it can be seen that much higher value for the parameter k can be achieved for an MTR than for both a classical and a tube light reactor. It is expected that the performance of TLR will surpass that of MTR because of superior catalyst activation, but the overall reactor efficiency will be greatly suppressed due to the application of an excess of light energy than required for catalyst activation.

When comparing the efficiency of this test reactor, expressed in terms of mols converted per unit time per unit reactor volume per unit electrical power consumed for the same model component with that of a classical annular reactor, an increase of about 4502 was observed (Table 4).

This increase in efficiency is realized in spite of the fact that the design of this test reactor was far from optimum. In addition, the proposed novel reactor design has the capability of scale-up to any dimensions whereas the classical photocatalytic reactors are restricted to small reactor units only. It is apparent that both the TLR and the MTR design idea creates great opportunities for building much more efficient photocatalytic reactors.

Cited References

Anderson, M.A., Tunesi, S. and Xu, Q., 1991, Degradation of organic chemicals with titanium ceramic membranes , US 5035784 A 910730.

Assink, J. W., Koster, T.P.M. and Slaager J.M., 1993,

Fotokatalytische oxydatie voor afvalwaterbehandeling, Internal report reference no. 93-137. TNO - Milieu en Energie, Apeldoorn, The

Netherlands .

Cooper,G.A., 1989, Photocatalyst in a glass wool matrix between plates , US 4888101.

Cooper, G.A. and Ratcliff, M.A., 1991, System for and method for decontaminating a contaminated fluid by using photocatalytic partic les , W09108813 Al 910627.

Fox, M. A. and Dulay, M. T., 1993, Heterogeneous photocatalysts , Chem. Rev., 93, 34l.

Hagfeldt, A. and GrStzel, M., 1995, Light- induced redox reactions in nano-crystalline systems , Chem. Rev., 95, 49.

Haneda, K., 1992, Photocatalytic e lement pipe and photocatalytic chemical reactor , patent JP 04061933 A2920227.

Heller, A. and Brock, J.R., 1993. Materials and methods for enhanced photocatalysis of organic compounds in oi l spi l l treatment , patent

W09317971 Al 930916.

Henderson, R.B. and Robertson, M.K., 1989, Fluid purification by photodegradation of organic pollutants and microorganisms, patent EP 3063301 Al 890308.

Hoffmann, M.R., Martin, S.C., Choi, W. and Bahnemann, D.W., 1995,

Environmental applications of semiconductor photocatalysis , Chem.

Rev., 95, 69.

Hofstadler, K., Bauer. R., Novallc, S. and Helsler, G.. 1994, New reactor design for photocatalytic treatment with TiO2 immobi lized on fused-si lica glass fibers: photo-mineralization of 4-chlorophenol ,

Environmental Science and Technology, 28, 670.

Hosokawa, M. and Yukimitsu, K., 1988. Treatment of waste fluids with titania partic les , patent JP 63042793 A2 880223.

Legrini, O., Oliveros, E. and Braun, A. M., 1993. Photochemical processes for water treatment, Chem. Rev., 93, 671. Masuda, R., Kawashima, K., Takahashi, W., Murabayashi, M. and Ito,

K.. 1994, Photocatalysts for treatment of harmful substances and its apparatus, patent JP 06320010 A2941122.

Matthews, R.W., 1988, Semiconductor photocatalytic method and system for determining organic matter in an aqueous solution including an oxidizing agents, patent WO 8806730 Al 880907.

Matthews, R.W., 1990, Coating photoactive metal oxides onto substrates and their use in water purification, patent AU 600289 B2

900809.

Matthews, R.W., 1992, Photocatalytic oxidation of organic contaminants in water : An aid to environmental preservation, Pure and Applied Chemistry, 64 (9),1285.

Miano, F. and Borgarello, E., 1991. Method for the heterogeneous catalytic photo-degradation of pollutants, patent EP 417847 Al 910320.

Mills, A., Davies, R. H. and Worsley, D., 1993. Water purification by semiconductor photocatalysis, Chemical Society Reviews, December,

417.

Ollis, D. F., Pelizzetti. E. and Serpone, N., 1989, Photocatalysis : Fundamentals and Applications, Wiley, New York.

Ollis, D. F., Pelizzetti, E. and Serpone, N., 1991, Destruction of water contaminants, Environmental Science and Technology, 25 (9).

1523.

Oonada, J., 1994. Water purifying method, patent JP 06071256 A2 940315.

Ray, A.K. and Beenackers, A.A.CM., 1995a, A novel swirl-flow reactor for kinetic studies of semiconductor photocatalysis, submitted to xxx.

Ray, A.K. and Beenackers, A.A.CM., 1995b, A novel photocatalytic reactor for water purification, submitted to xxx.

Ritchie, D.G., 1991. Photocatalytic fluid purification apparatus having helical nontransparent substrate, patent US 5069885 A 911203.

Sato, K., 1992, Drinking water container capable of decomposing organic halogen compounds by light irradiation, patent JP 04114791 A2920415.

Taoda, H., 1993, Water treatment, patent JP 05076877 A2930330. Wake, H. and Matsunaga, T., 1994, Bedox reaction using photocata lysis of semiconductor, patent JP 06134476 A2 940517.

Zeltner, W. A., Hill, C. G. and Anderson, M. A., 1993. Supported ti tania for photodegradation, ChemTech, 21, May.

Claims

1. Photocatalytic reactor comprising a housing having inlet means for feeding the reactor with polluted water, outlet means for removing treated water from the reactor, a photocatalyst and a UV- source, characterized in that the reactor comprises at least one light conductor means and at least one UV-source capable of emitting UV-light into and through the light conductor means.

2. Reactor according to claim 1, characterized in that the light conductor means is a hollow tube, made of light conducting material.

3. Reactor according to claim 1 or 2, characterized in that the hollow tube has a diameter in the range of 0.0025-0.1 m.

4. Reactor according to any of claims 1-3, characterized in that the hollow tube is made of quartz glass or pyrex glass.

5. Reactor according to any of claims 1-4, characterized in that the photocatalyst is selected from the group consisting of anatase TiO₂, Pt-TiO₂, ZnO, CdS, CdSe. SnO₂. SrTiO₃, WO₃, Fe₂O₃, ZrO₂, Ta₂O₅, GaP, SiC and SnO, preferably anatase TiO₂.

6. Reactor according to any of claims 1-5, characterized in that the photocatalyst has a surface area of 20-200 m²/g and a particle size of 0.001-1.0 μm.

7. Reactor according to any of claims 1-6, characterized in that the UV-source emits light in the range of 200-440 nm, preferably 320-380 nm.

8. Reactor according to any of claims 1-7, characterized in that the UV-source is a high pressure or a low pressure Hg vapour lamp or an Xe arc lamp or a tube light.

9. Reactor according to any of claims 1-8, characterized in that the light conductor means is coated on its outside surface with a photo-catalyst.

10. Reactor according to claim 9. characterized in that the diameter of the tube light is from 0.0045 - 0.025 m.

11. Reactor according to claim 9 or 10, characterized in that the UV-source is a tube light coated on its outside surface with a photocatalyst.

12. Reactor according to any of the claims 9-11 , characterized in that the UV-source is a tube light, incorporated in a hollow tube, coated on its outside surface with a photocatalyst.

13. Reactor according to any of the claims 9-12, characterized in that the light conductor means is coated with 0.5 - 5-0 g/m² of photocatalyst.

14. Reactor according to any of the claims 9-13. characterized in that the light conductor means has firstly been coated with a glass wool fiber material on its outside surface.

15. A process of degrading organic substances and microorganisms present in a polluted aqueous solution or suspension by means of a photocatalytic reaction, characterized by feeding the polluted aqueous solution or suspension to a reactor according to any of the claims 1-8, irridiating the photocatalyst, present in the form of a suspension, with UV-light and removing the treated aqueous solution or dispersion from the reactor.

16. A process of degrading organic substances and microorganisms present in a polluted aqueous solution or suspension by means of a photocatalytic reaction, characterized by feeding the polluted aqueous solution or suspension to a reactor according to any of the claims 9-14, irradiating the photocatalyst with UV-light and removing the treated aqueous solution or dispersion from the reactor.

17. A process according to claim 15 or 16, characterized in that extra oxidants such as O₂, H₂O₂ and O₃ are used.

18. A process according to any of the claims 15 or 17. characterized in that the polluted aqueous solution or dispersion is a domestic, industrial or agricultural waste water comprising one or more saturated or unsaturated hydrocarbons, optionally comprising one or more heterogeneous atoms selected from the group consisting of halogen, N, S, O and P. *****