Analytical optimisation of eco-friendly soap production using hyperspectral imaging and chemometric modelling of physicochemical properties.

Abstract

The pressing environmental imperative to curb petrochemical detergent pollution has driven the development of circular approaches that valorise waste lipids. In this work, we establish an integrated chemometric–hyperspectral framework to optimise bar soap production from used frying oils (UFOs). A fractional Taguchi screening, followed by a central composite rotatable design (CCRD), systematically evaluated the effects of NaOH concentration (14–22 % w/v) and NaOH/UFO ratio (0.30–0.70) on soap pH and mechanical hardness. The optimal formulation (14.08 % NaOH; ratio 0.30) yielded bars with pH 10.31 ± 0.02 and hardness 359.6 ± 5.2 g, alongside superior textural resilience and cohesion. Near-infrared hyperspectral imaging (896–1704 nm) coupled with partial least squares regression (PLSR) enabled non-invasive, real-time pH prediction (R2 = 0.83; SEP = 0.18), while a simplified multiple linear regression (MLR) model refined alkalinity forecasts to R2 = 0.87. Hardness modelling (R2 < 0.60) highlighted the need for advanced variable-selection and nonlinear strategies to capture complex microstructural dynamics. By uniting NIR-HSI with data-driven calibration, our methodology delivers rapid quality control, reduces reliance on laborious assays and demonstrates a scalable, sustainable template for eco-innovative personal-care manufacturing.