Optimization of microfluidic synthesis of silver nanoparticles: A generic approach using machine learning

The properties of silver nanoparticles (AgNPs) are affected by various parameters, making optimisation their synthesis a laborious task. This is facilitated in this work concurrent use T-junction microfluidic system and machine learning approach. AgNPs synthesized reducing nitrate with tannic acid the presence trisodium citrate, which has dual role reaction as stabilizing agent. study uses decision tree-guided design experiment method for size AgNPs. developed approach kinetic nucleation growth constants derived from an independent set experiments to account chemistry synthesis, Reynolds number ratio Dean reveal effect hydrodynamics mixing within device storage temperature particle stability after collection. obtained model was used define parameter space additional carried out improve further. numerical results illustrate that well-designed can contribute more effectively development different models (decision tree, random forest XGBoost) opposed randomly added experiments.