Detecting Heavy Metals In Leafy Vegetables And Method Of Signal Enhancement Based On Laser Induced Breakdown Spectroscopy

For the purpose of quantitative analysis and real-time online detection of heavy metal elements in leaf vegetables by laser induced breakdown spectroscopy(LIBS), characteristic lines of PbⅠ405.78 nm, CrⅠ425.435 nm, CrⅠ427.480 nm, and CrⅠ428.972 nm,CuⅠ324.7nm and CuⅠ327.4nm were selected as the analytical spectral lines of lead, chromium, copper respectively. The characteristics of laser plasma of leaf vegetables heavy metal elements were studied by the experiments. Then different methods were applied to study the enhancement of laser plasma signal. Different data processing methods were applied to the LIBS spectrum data, and the affect of the accuracy of the models with different data processing methods were also discussed.The experimental results show that the optimum parameters configuration of detecting heavy metals in leafy vegetables by LIBS were setting the delay time as 1.7 μs, integration time as 0.06 ms, the laser energy as 103.8 mJ, which case the spectral intensity and SNR could achieve the better results. Adopt cylindrical bore constraint, net intensity of Pb was above average 1.35 times than unconstrained, the net intensity of Cu was above average 2 times than unconstrained. Adopt magnetic constraint, the net intensity of Pb, Cu were above average 2.2 and 1.5 times than unconstrained respectively. And with double pulse, the enhanced effects were more than five times. The results show that in the application of LIBS in leaf vegetables, enhancement of double pulse is the best, followed by magnetic confinement, constraint of the hole again. In this study, the sensitivity of LIBS detection has been improved effectively, a good foundation was provided for the detection of heavy metal elements in leaf vegetables by LIBS.Compared with intensity calibration model, the calibration model of partial least squares witf different variable selection methods has higher correlation coefficient. The correlation coefficient increased by 23.9% average. Comparison is made between the different partial least squares models, for LIBS testing data of Pb and Cr, siPLS model has the best quality, correlation coefficient of correction model, correlation coefficient of prediction model, RMSECV, RMSEP were 0.9967, 0.988, 0.737, 1.35 and 0.9709, 0.9713, 1.26, 1.43 respective. For LIBS testing data of Cu, has the quality of iPLS model is best, the four parameters of the model were 0.9951, 0.9955, 0.648 and 1.33 respective. In summary, PLS models are better than the intensity calibration model, the quality of iPLS, siPLS model are better than the PLS. For the spectral data of detecting heavy metals in leaf vegetables by LIBS, iPLS and siPLS methods can both obtain good modeling effect, the specific needs should be selected according to the different experimental conditions.