Fast visualization of distribution of chromium in rice leaves by re-heating dual-pulse laser-induced breakdown spectroscopy and chemometric methods

Knowledge of distribution of toxic metal in crop is essential for studying toxic metal uptake, transportation and bioaccumulation, and it is important for environmental pollution monitoring. In this study, the macro spatial distribution of chromium in rice leaves was visualized by re-heating dual-pulse laser-induced breakdown spectroscopy (DPLIBS) and chemometric methods. After the optimization of two important parameters (delay time and energy ratio) in DPLIBS, chromium prediction model was established based on global spectra. The global model achieved acceptable performance while slight overfitting for model was found because of numerous irrelevant variables. Feature variables including emissions from chromium and other elements were successfully selected by the values of regression coefficient in partial least square regression model. Best performance was achieved by using the feature variables and support vector machine, with correlation coefficient of prediction of 0.959, root mean square error of prediction of 13.4 mg/kg and residual predictive deviation of 3.6. Finally, the distribution of chromium in rice leaves was visualized with the best prediction model. The distribution image showed that chromium distributed approximately symmetrically along the vein and was likely to be accumulated in leaf apex. The preliminary results provide an approach for investigating the macro spatial distribution of elements in crops, which is important for environmental protection and food safety.