The spinning disc reactor (SDR) uses surface rotation to produce thin film flow with improved mixing and reduced residence times in chemical processing applications. Solvent-antisolvent precipitation is one such process that can benefit from these properties. This study investigates the film hydrodynamics and precipitation of starch nanoparticles by contacting starch dissolved in sodium hydroxide with ethanol as the antisolvent. One objective of this study is to understand how interactions of the disc surface topography (grooved and smooth) with other parameters such as liquid flowrate, antisolvent to solvent flow ratio and disc speed impact the mixing and precipitation processes. Results indicate that an increase in flow rate and rotational speed leads to smaller nano-particles and narrower size distributions, which is attributed to increased shear and instabilities within the liquid film. It was also observed that an increased antisolvent to solvent ratio caused a reduction in particle size, as increased antisolvent generated higher supersaturation. Results showed that although particle size was not significantly influenced by the disc texture, the size distribution was narrower and higher yields were obtained with the grooved disc surface. The grooved disc therefore offers the opportunity for higher throughput in the solvent-antisolvent precipitation of starch particles with better product quality.

Empirical correlations have been developed to relate experimentally determined starch nanoparticle size obtained in a solvent–antisolvent precipitation process with key hydrodynamic parameters of a spinning disc reactor (SDR). Three different combinations of dimensionless groups including a conventional Reynolds number (Re), rotational Reynolds number (Re_ω) and Rossby number (Ro) have been applied in individual models for two disc surfaces (smooth and grooved) to represent operating variables affecting film flow such as liquid flowrate and disc rotational speed, whilst initial supersaturation (S) has been included to represent varying antisolvent concentrations. Model 1 featuring a combination of Re, Re_ω and S shows good agreement with the experimental data for both the grooved and smooth discs. For the grooved disc, Re has a greater impact on particle size, whereas Re_ω is more influential on the smooth disc surface, the difference likely being due to the passive mixing induced by the grooves irrespective of the magnitude of the disc speed. Supersaturation has little impact on particle size within the limited initial supersaturation range studied. Model 2 which characterises both flow rate and disc rotational speed through Ro alone and combined with Re was less accurate in predicting particle size due to several inherent limitations.