Design and Evaluation of Translucent Structures for Photocatalytic Reactors

Photocatalysis is a technology which opens up new processing windows for organic synthesis or organic transformation reactions. A common application is the photocatalytic degradation of organic pollutants in water. Degradation reactions are basically oxidation reactions, where the desired end product is usually carbon dioxide. The intermediate products formed during the mineralization process are potentially interesting to valorize meaning that photocatalysis can be used to perform selective oxidation reactions using molecular oxygen. The catalyst used in a photocatalytic process can be present in the same phase as the reacting liquid, which is called homogeneous photocatalysis, or in a different phase than the reacting liquid, which is called heterogeneous photocatalysis. In this work, the focus will be on heterogeneous photocatalysis.

Performing a heterogeneous photocatalytic reaction requires a light source, a photocatalyst and a reactor geometry. The reactor needs to be designed in such a way, that it synergizes photon transport and mass transport with a high catalyst loading. Two main types of reactors exist. In the first reactor type, the slurry reactor, catalyst particles are suspended in the reacting liquid resulting in a low controllability over the light distribution and requiring a catalyst separation step after the reaction is completed. A slurry configuration is often used in combination with a batch reactor. In the second category, the catalyst layer is immobilized on a support. This approach is more promising for synergizing photocatalysis with flow chemistry. The catalyst loading can be precisely controlled by the catalyst layer thickness, and there is no need for a catalyst separation step.

Despite the interesting possibilities of photocatalysis, no industrial application exists as of yet. One of the reasons can be attributed to the reactor design. The problem regarding reactor design can be made clear using a simple flat-plate reactor as a reference system. In a flat-plate reactor, the reacting fluid is flowing between two plates where both plates can be coated with a photocatalyst which is irradiated perpendicular to the flow direction. The catalyst layer is chosen sufficiently thick to absorb all the incoming light energy. Since photocatalysts are mostly highly absorbing materials, the light transmission is almost equal to 0 after travelling approximately 10 µm in the catalyst layer. In this type of reactor, two problems occur. Firstly, because of the laminar flow profile between the two plates, mass transport from the bulk liquid towards the catalyst layer is dominated by diffusion, which limits the supply of reagents to the catalyst surface. Secondly, when the catalyst layer is too thick, mass transfer limitations inside the layer occur. A potential solution is to decrease the volume of the reactor to microreactor scale in order to decrease the diffusion length and to increase the surface-to-volume area meaning thinner layers can be applied for the same catalyst loading with respect to the liquid volume. However, microreactors are not scalable, not easy to illuminate efficiently due to the small reactor window, and have a low total surface area. As a result, the issue of internal diffusion limitations in the catalyst layer is not solved because the catalyst layer thickness has to be increased to increase the catalyst loading. Also, applying thinner catalyst layers on a microreactor surface leads to light energy losses at the backside of the reactor. The solution to solve all these issues simultaneously is to incorporate multiple microchannels in a larger translucent structure. A translucent structure can be illuminated easily, provides a large surface-to-volume area as well as a large total surface area and has an improved bulk liquid mass transfer. These properties allow for easy light source design and flexibility in choosing the correct catalyst layer thickness avoiding diffusion limitations using the geometry of the structure as a parameter to tweak total catalyst loading and surface area. In this work, the used translucent structure is a packed bed of equally sized borosilicate glass spheres.

In a proof-of-concept study, it was proven that the productivity and energy efficiency of the packed bed reactor was higher than the existing designs from the literature. For the photocatalytic degradation of methylene blue, the reactor had a productivity similar to that of a microreactor, but a photocatalytic space-time yield, which is a measure for productivity versus energy efficiency, of 4 to 5 orders of magnitude higher. It was shown that the improved surface-to-volume area leads to high catalyst loadings up to 2.9 g L-1 while maintaining layer thicknesses below 300 nm. In the second part of the work, the effect of the bead size and catalyst layer thickness on external mass transfer and energy absorption was assessed. The bead size of the packed bed reactor can be used to alter the external mass transfer properties of the photocatalytic bed and to tweak the light absorption and surface area in the bed. When in a fixed reactor volume the bead size is decreased while maintaining the same catalyst loading, the number of scattering boundaries increases meaning that less light is being absorbed. In the third part of the work, the optimal catalyst configuration and number of structural layers in a translucent structure were determined. For a light intensity of 200 W m-2 the optimal layer thickness is more or less equal to 1.5 µm in a translucent structure containing a minimum of 5 structural layers. These values vary as a function of the light intensity, since mass transfer limitations are dependent on the catalyst layer reactivity. Finally, translucent packed bed structures were tested for gas-liquid oxidation reactions and it was proven that also for gas-liquid applications translucent structures can improve the productivity of a photocatalytic reaction.