Nucleation and solidification on the microscale

The development of new processes for synthesising nano-particles and hydrogel microparticles is mostly based on a trial and error approach. This results in high consumption of chemicals and energy, which adds a negative impact on the environment. Using microfluidic technology reduces the amount of reactants necessary for technology development and optimisation. In addition, predictive numerical tools further reduce the experimental part of research and development and save resources, energy and time. In this case study, we aim to develop multi-fidelity models to predict an output of microfluidic reactions based on and validated by tailored experimental studies. Also, it will enable the automated choice of reactants and process parameters to synthesise the final product (silver nano-particles or alginate hydrogel microparticles) with required properties. We will collect a wide range of experimental data, develop high-fidelity physical models based on first principles and parameters derived from experiments, and use data analysis and machine learning approaches to improve the fidelity of physical models and develop simple low fidelity models. The last will allow a fast estimation of parametric space, providing desired experimental output.

The results of this project will be useful for developing new synthesis processes on silver nano-particles and alginate hydrogel microparticles. The results will also provide a methodology for developing predictive tools of variable fidelity levels for other chemical synthesis and formulation engineering areas.