A Preliminary Study of Aflatoxin B1 Detection in Peanut Oil by Terahertz Time-Domain Spectroscopy

Abstract

Abstract. Aflatoxins are fungal secondary toxic metabolites that can cause serious health problems in animals and humans. Terahertz (THz) spectroscopy is promising as a nondestructive analytical tool for chemical and biochemical detection. In this technical note, THz time-domain spectroscopy (THz-TDS) was used for the quantitative analysis of aflatoxin B1 (AFB1) in peanut oil. Chemometrics methods including partial least squares regression (PLSR) and stepwise multiple linear regression (SMLR) were used to build quantitative models. Their performance was compared using two spectral pretreatment methods (first and second derivative). The results show that the SMLR model was more stable than the PLSR model, although its prediction precision was slightly lower. When 1.11 and 0.72 THz frequencies were used, the root mean square errors of calibration (RMSEC) and cross-validation (RMSECV) were 18.2 and 27.7 ppm, respectively, and the corresponding correlation coefficients (rc and rcv) were 0.9639 and 0.9309, respectively. The overall results indicate that THz-TDS combined with frequency selection could be applied as a rapid and nondestructive tool for the determination of AFB1 content in peanut oil.