Prestatiebeoordeling van vloeistof–vloeistof dispersieprocessen via druppelkarakterisering en evaluatie van vermogensdichtheid

This dissertation investigates the performance of multiple reactors for generating liquid-liquid dispersions. It examines how an oscillatory baffled reactor, a pulsed column, and a mixed vessel compare in terms of space-time yield and energy efficiency. To support this comparison, a methodology is developed to determine droplet size and the specific power input per volume introduced through fluid oscillations in a baffled reactor. Finally, a case study is presented in which both the oscillatory baffled reactor and the pulsed column are applied to a hydrometallurgical separation process of cobalt and nickel. Firstly, an image-based protocol for droplet size characterization is developed. The method combines sampling, stabilization, and dilution with off-line image acquisition using a microscope. This approach offers several advantages, the most significant being the clarity of the images, which enables automated analysis via a custom Python-based script. Automation eliminates the human bias inherent in manual analysis and allows for the detection of a significantly larger number of droplets, resulting in more statistically robust outcomes. The algorithm is capable of detecting up to 8000 droplets per minute and determining the Sauter mean diameter with an accuracy of 96%. Moreover, due to the dilution step, the protocol is applicable even at high dispersed phase volume fractions, such as 50 vol%. Another key aspect is the power density required to create a liquid-liquid dispersion. A novel method for calculating power density specifically tailored to oscillatory flow is developed. In contrast to literature, the method assumes that energy cannot be recovered during periods of negative instantaneous power density. It also incorporates experimentally measured dynamic pressure drop profiles. Additionally, the phase difference between pressure drop and fluid velocity is investigated. Results show that increasing the amplitude reduces the phase difference, whereas frequency has little to no effect. This behavior is hypothesized to originate from the different relationship that amplitude and frequency have with fluid velocity and acceleration. The newly developed method is further applied to determine the power density in both an oscillatory baffled reactor and a pulsed column. For various power densities, the approximate droplet size range is estimated, enabling a comparative ranking of the reactors based on increasing power density required to achieve a similar dispersion: pulsed column, mixed vessel, and oscillatory baffled reactor. However, the oscillatory baffled reactor can operate at significantly shorter residence times compared to the other two technologies, thereby potentially reducing the energy consumption per unit volume of processed liquid. The hydrometallurgical separation of cobalt and nickel is investigated as a case study of liquid-liquid dispersion processes. A custom 3D-printed oscillatory baffled reactor with an internal diameter of 9 mm is employed to investigate the effect of oscillation conditions and net flow rate on cobalt extraction efficiency. Residence times between 9 and 73 seconds are explored, resulting in a space-time yield ranging from 1180 to 9443 L/Lday. A percentage extraction of 80% is achieved within 9 seconds, increasing to 97% at a residence time of 36 seconds. Based on this performance, a productivity of 116.5 g Co^{2+}/day is attained with a single reactor. In comparison, a pulsed column operated in counter-current mode demonstrates a cobalt extraction efficiency of 99.7%, surpassing both the single-stage equilibrium value of 99.4% and the performance achieved with the oscillatory baffled reactor. However, flooding conditions constrain the operating window of the pulsed column, resulting in a space-time yield of 240 L/Lday.