Multicomponent prediction of Sudan dye adulteration in crude palm oil using SERS – Based bimetallic nanoflower combined with genetic algorithm

Abstract

Crude palm oil (CPO) is an important edible vegetable oil used globally, recently subjected to Sudan dye adulteration. This study explored the feasibility of a novel SERS-based Au@Ag substrate to detect four Sudan dyes (I - IV) in CPO. When mixed with spiked CPO, the SERS substrate produced strong signals that increased with increasing concentrations from 0.001 to 4.0 ppm. The genetic algorithm partial least square (GA-PLS) model outperformed the partial least square (PLS) and the ant colony optimization – PLS (ACO-PLS) models with Rc values of 0.9844, 0.9865, 0.9884, and 0.9888 0.9846. The calculated LOD were 0.00088 ppm, 0.00092 ppm, 0.00095 ppm, and 0.00097 and real sample recovery rates of 88.0–113.0%, 92.0–103.0%, 91.2–98.5% and 97.0–109.5% for Sudan I, II, III, and IV, respectively. The findings affirmed the SERS sensor’s considerable potential for rapid and selective detection of Sudan dyes in CPO when combined with chemometrics.