Application of NIRs coupled with PLS and ANN modelling to predict average droplet size in oil-in-water emulsions prepared with different microfluidic devices

In this study, the potential of microfluidic systems with different microchannel geometries (microchannel with teardrop micromixers and microchannel with swirl micromixers) for the preparation of oil-inwater (O/W) emulsions using two different emulsifiers (2 % and 4 % Tween 20 and 2% and 4 % PEG 2000) at total flow rates of 20-280 mu L/min was investigated. The results showed that droplets with a smaller average Feret diameter were obtained when a microfluidic device with tear drop micromixers was used. To predict the average Feret diameter of O/W emulsion droplets, near-infrared (NIR) spectra of all prepared emulsions were collected and coupled with partial least squares (PLS) regression and artificial neural network modelling (ANN). The results showed that PLS models based on NIR spectra can ensure acceptable qualitative prediction, while highly non-linear ANN models are more suitable for predicting the average Feret diameter of O/W droplets. High R-2 values (R-2 validation greater than 0.8) confirm that ANNs can be used to monitor the emulsification process. (C) 2022 Elsevier B.V. All rights reserved.

Automated summary

This study investigates the potential of microfluidic systems with different microchannel geometries for preparing oil-in-water emulsions. The results show that a microfluidic device with tear drop micromixers produces smaller droplets. Near-infrared spectra combined with partial least squares regression and artificial neural network modeling are used to predict the average diameter of emulsion droplets. The results demonstrate that artificial neural network models are more suitable for predicting droplet diameter.