Discrimination of untreated and sodium sulphite treated bean sprouts by Fourier transform infrared spectroscopy and chemometrics

Abstract

Sprouts of black beans (Phaseolus vulgaris L.), soybeans (Glycine max L.) and mung beans (Vigna radiata L.) are widely consumed foods containing abundant nutrients with biological activities. They are commonly treated with sulphites for the preservation and extension of shelf-life. However, our previous investigation found that immersing the bean sprouts in sulphite might convert the active components into sulphur-containing derivatives, which can affect both the quality and safety of the sprouts. This study explores the use of FTIR in conjunction with chemometric techniques to differentiate between non-immersed (NI) and sodium sulphite immersed (SI) black bean, soybean and mung bean sprouts. A total of 168 batches of raw spectra were obtained from NI and SI-bean sprouts using FTIR spectroscopy. Four pre-processing techniques, three modelling assessment techniques and four model evaluation indices were examined for differences in performance. The results show that the multiplicative scatter correction is the most effective pre-processing method. Among the models, the accuracy rate of the three models was as follows: radial basis function neural network (95%) > convolutional neural network (91%) > random forest (82%). The overall findings indicate that FTIR spectroscopy, in conjunction with appropriate chemometric approaches, has a high potential for rapidly determining the difference between NI and SI-bean sprouts.

Keywords:

• Bean sprouts
• FTIR spectroscopy
• sodium sulphite immersion
• chemometrics
• rapid discrimination

Disclosure statement

No potential conflict of interest was reported by the author(s).

Authors’ contributions

Yaxin Li: Investigation, formal analysis, writing; Baoguo Chen: Investigation, formal analysis; Shuhong Ye: Resources, funding acquisition; Qi Wu: Methodology; Lin Zhu: Resources, funding acquisition; Yan Ding: Resources, writing – review and editing, funding acquisition.

Additional information

Funding

This research was supported by the Natural Science Foundation of Liaoning Province (2021-MS-299), the National Natural Science Foundation of China (31770725) and the China Scholarship Council (CSC No. 202008210171).