Evaluation of a portable spectrometer for detection of pesticide residue on cabbage

• Main Author: Che Mohammad Ishkandar El-Rahimin, Che Dini Maryani
• Format: Thesis
• Language: English
• Published: 2018
• Subjects: Cabbages - Diseases and pests - Control of - Malaysia; Pesticides - Application techniques
• Online Access: http://psasir.upm.edu.my/id/eprint/77616/1/FK%202019%2018%20IR.pdf; http://psasir.upm.edu.my/id/eprint/77616/
• Institution: Universiti Putra Malaysia

Cabbage is one of the most popular vegetables grown in Malaysia. However, pesticides have been widely used to control the disease-born pests on the vegetable, which can lead to residue presence in agricultural products. Most of the existing technologies for detecting pesticide residues requires a destructive method for sample preparations, which is time-consuming, costly and complex process. Thus, this study investigated the potential application of non-destructive spectroscopic method by using visible shortwave near infrared (Vis-SWNIR) spectrometer to detect the presence of pesticide residues on cabbage samples. A total of 105 organic cabbages was used in the experiment. There were two types of pesticides used for this research namely deltamethrin and malathion. The samples were divided into seven batches, where the first three batches were treated with deltamethrin pesticide at different concentrations namely low (0.08% v/v), medium (0.11% v/v) and high (0.14% v/v). Another three batches were treated with malathion pesticide at different concentrations namely low (0.18% v/v), medium (0.25% v/v) and high (0.32% v/v). The seventh batch was not treated with any pesticide and used as a control sample for the experiment. The maximum residue limit (MRL) (mg/kg) obtained from gas chromatography (GC) analysis showed that the cabbage sprayed with a high concentration of pesticide gave high MRL value, whereas the low concentration of pesticide sprayed gave the low value of MRL. Based on the t-test, it shows the MRL data for both pesticides are significantly different (p<0.05). The spectral data focused on wavelength between 650 and 700 nm; and 740 and 940 nm to observe the trend of the spectral curve. The spectral curve of cabbage with pesticide and without pesticide shows there is a difference between the sample with and without pesticide residue. The spectral data were correlated with MRL data. The coefficient of determination (R2) for MRL and spectral value were obtained using partial least square (PLS) analysis. The R2 value for the calibration model of deltamethrin and malathion was 0.982 and 0.987 respectively and the R2 value ii for the prediction model was 0.948 and 0.989 respectively. ANN classifier was done using spectral data of deltamethrin and malathion pesticide with classification accuracies 66.70% and 73.30% respectively. The overall accuracy of data was increased to 93.33% and 100.00% for deltamethrin and malathion pesticide, respectively by applying the PCA. These results indicated that the portable spectroscopy was able to differentiate between cabbage samples with and without pesticide residues at different levels of pesticide concentration non-destructively. In comparison, the other method such as gas chromatography, high-performance liquid chromatography and biosensor was destructive, time consuming and high cost. In conclusion, the Vis-SWNIR spectrometer has the potential to be used for the detection of pesticide residues on cabbage samples.