< Back to previous page

Project

Characterization of Microstructured Reactors for Photochemical Transformations

Photochemical reactions are found in a large variety of applications such as synthesis of pharmaceuticals, polymers, and water treatment. In the last two decades, photochemistry was coupled successfully with micro‑scale reactors accelerating the research of light‑driven organic synthesis. Nowadays, the photochemical reactions are carried out in both customized and commercial flow reactors which are predominantly illuminated by Light-Emitting Diodes (LEDs).

The current challenge is the lack of information about the properties of light source, reactor and reaction medium and how they influence the reported conversions and yields. This results in reproducibility problems and limited understanding which leads further to extensive experimental efforts required for the optimization and scale‑up of photochemical processes. While this information can be obtained by characterization, the characterization studies are rare. The characterization methods are either missing or not convenient for flow reactors analysis, especially when they are illuminated by visible light sources. This thesis focusses on developing new accessible tools to characterize flow reactors and their application for understanding the performance of visible-light flow reactors in conditions that are mostly encountered in organic synthesis.

Firstly, a new visible-light chemical actinometer was developed for quantifying the number of photons with wavelengths ranging between 480 and 620 nm reaching the channels of flow reactors. An experimental methodology was established to determine the photon flux and optical pathlength with high reproducibility, using commercially available compounds, a simple experimental procedure, and on‑line absorbance detection. Secondly, a characterization tool complementary to actinometry was developed to quantify the irradiance and light distribution on the reactor surface when the LED light source was placed at distances smaller than 2 cm. The existing far-field model was improved using angular irradiance distributions extracted from near-field goniophotometer measurements using a commercial ray tracing software.

The irradiance model together with the chemical actinometer was applied to quantify the uniformity and the energy efficiency of different LEDs arrays, designed as light sources for a microstructured glass reactor. The LED arrays differ by layout and LED number. During the light source design, the electrical and thermal properties of LEDs were also addressed as they affect the optical output, stability and lifetime of the designed light sources. From the energy efficiency analysis, it was found that 1% of the electrical energy reached the reactor channel in the form of photons. The low energy efficiency was caused mainly by the high power consumption of the driving board and the small surface fraction of the glass plate occupied by the reactor channels.

Next, the chemical actinometer was combined with image analysis and residence time distribution experiments to investigate the photon transport and hydrodynamics in liquid and gas‑liquid flows in the same microreactor. The microreactor promoted a Taylor flow at all studied gas fractions. The obtained gas bubble and slug lengths were measured by image processing. The RTD measurements showed that the liquid residence time in gas‑liquid flows was similar to the value found in single-phase flow for the same overall flow rate, apart from the highest gas fraction. The photon flux per liquid volume increased exponentially with the gas content up to double the value measured in single‑phase flow. This observation was correlated to the volume of the liquid present in the film, region around the bubble caps and slug. A three-zone model was developed to predict the photon flux obtained from chemical actinometry. The new model indicated that the photon flux per liquid volume in the film and the region around the bubble caps were three times larger than the one in liquid slug. Furthermore, the optical pathlength decreased with the gas fraction. This variation can be predicted by a correlation that includes single-phase optical pathlength, gas fraction and an empirical factor.

A similar study was performed in a commercial milli‑scale reactor, Corning® G1 Advanced-Flow™ Reactor. This reactor promoted bubbly flow at all investigated gas fractions. From RTD experiments, it was found that the liquid residence time in all gas‑liquid flows was higher in comparison with the value found in single‑phase flow measured at the same total flow rate. In contrast with the increasing photon flux per liquid volume observed in Taylor flow, only a moderate rise was found in bubbly flow. In the absence of a thin liquid film, the photoreactor performance was independent of the gas fraction. Furthermore, the optical pathlength also decreased with the gas fraction. As in the case of Taylor flow, the optical pathlength in bubbly flow was a function of the single-phase optical pathlength, gas fraction and an empirical factor.

This thesis illustrates that the characterization of flow reactors provides a wide range of quantitative information regarding the light source, photon transport, and hydrodynamics. This information could be used as input in models to study photon transport in new flow patterns or reactor geometries. Moreover, the presented findings contributed to gain new insights regarding the parameters important for photoreactor performance. The acquired knowledge could support a rational photoreactor and process design.

 

Date:20 May 2015 →  31 Dec 2019
Keywords:Photo microreactors
Disciplines:General chemical and biochemical engineering, Process engineering, Transport phenomena
Project type:PhD project