Interaction of turbulence with nanoparticle formation in reaction crystallisation

Abstract

Reaction crystallisation, or precipitation, is a process involving particle formation through a chemical reaction in the liquid phase. The resultant particle size distribution (PSD) is the most important process parameter and determines the product quality and suitability for particular applications. The formation of particles involves nucleation and growth of particles, both of which are fast processes that can have timescales comparable with mixing, a fact that promotes complex interactions between crystallisation and mixing. The objective of the present thesis is to develop a methodology for coupling direct numerical simulation with population balance modelling for particulate processes and to employ this methodology to investigate the interplay between turbulence, nucleation and growth in a reaction crystallisation process, as well as the effect of turbulence fluctuations on the PSD evolution.

The DNS-PBE methodology is first developed and validated with experimental measurements from the literature on BaSo4 nanoparticle precipitation in a T-mixer. This approach captures the smallest flow scales and does not require additional turbulence model or closures for population balance modelling. It is then employed for studying the mixing-precipitation interactions. Timescales and lengthscales of the processes involved are evaluated to show evidence of scale overlap between turbulence and precipitation. Local dominant mechanisms are identified and a map of them is constructed via the process timescales. It is found that the shape of the PSD is controlled by the extent of the nucleation zone, in which the number of formed seeds determines the subsequent growth consumption. As a means of controlling PSD evolution, mixing acts as an agent for changing the dominant zone distribution and topology.

The turbulent fluctuations at different scales are also studied in the present work. These fluctuations contribute to the evolution of the PSD in the form of unclosed terms in the averaged PBE. Contribution of these terms is examined and the consequences of neglecting them are outlined. Finally, a preliminary study of the effects of sub-Kolmogorov scale mixing is performed. While thinner reaction structure is obtained when the Batchelor scale at high Schmidt number is resolved, no considerable deviations are found on the PSD evolution and total consumption in low Schmidt number simulation due to the extent of reaction zones being balanced by the strength of reaction rates.