Feasibility of Surface-Enhanced Raman Spectroscopy for Rapid Detection of Aflatoxins in Maize

Rapid and sensitive surface-enhanced Raman spectroscopy (SEAS) for aflatoxin detection was employed for development of the models to classify and quantify aflatoxin levels in maize at concentrations of 0 to 1,206 mu g/kg. Highly effective SERS substrate (Ag nanosphere) was prepared and mixed with a sample extract for SERS measurement. Strong Raman bands associated with aflatoxins and changes in maize kernels induced by aflatoxin contamination were observed in different SERS spectroscopic regions. The k-nearest neighbors (KNN) classification model yielded high classification accuracy and lower prediction error with no misclassification of contaminated samples as aflatoxin negative. The multiple linear regression (MLR) models showed a higher predictive accuracy with stronger correlation coefficients (r = 0.939-0.967) and a higher sensitivity with lower limits of detection (13-36 mu g/kg) and quantitation (44-121 mu g/kg) over other quantification models. Paired sample t test exhibited no statistically significant difference between the reference values and the predicted values of SEAS in most chemometric models. The proposed SERS method would be a more effective and efficient analytical tool with a higher accuracy and lower constraints for aflatoxin analysis in maize compared to other existing spectroscopic methods and conventional Raman spectroscopy.