Research On Heavy Metal Copper And Lead In Plants Based On Near-infrared And Raman Spectroscopy

Heavy metals in plants are directly related to the growth and development status ofplants, and also indirectly affecting human health through the food chain transmission.Therefore, detecting the content of heavy metals in plants not only helps to choose the highquality plants and improve the crop quality and yield, but also can reduce the harm caused byeating the crops, which containing excessive heavy metals. The combination of digestion andchemical analysis instrument is mainly used for the detection of the heavy metal content inplants, which is time-consuming and plant-damage. Therefore, seeking a kind of rapid andnondestructive method for detecting heavy metal content in plants has the important andpractical significance.Heavy metals in plants usually have complex chelation with the organic moleculargroups that have spectral information. Therefore heavy metals in plants can be indirectlydetected by using spectral technique basing on the chelation. In this paper, near infraredspectroscopy and Raman spectroscopy were carried out in detecting the content of copper andlead in Ludwigia prostrate leaves and Vetiver root, stem and leaf. The quantitative analysismodels of detecting heavy metal content in plants were established. The main researchcontent and results were shown as follows:(1) Study the feasibility of detecting heavy metal copper content in Ludwigia prostrateleaves using near infrared spectroscopy. Combined with different spectral preprocessingmethods, partial least squares (PLS) was used to establish a near infrared quantitative analysismathematical model of detecting heavy metal content in Ludwigia prostrate leaves. The fullcross validation correlation coefficient and root mean square error of this model were0.95and5.99mg/kg, respectively.24unknown samples were used to test the quality of this model,the correlation coefficient and the root mean square error of prediction were0.923and7.38mg/kg, respectively. The result shows that the near infrared spectroscopy is suitable for thequantitative detection of the copper content in Ludwigia prostrate leaves, it provides thereference for the detection of heavy metal content in Vetiver grass.(2) Study the quantitative detection methods of heavy metal content of copper and leadin Vetiver grass roots, stem, leaf tissues based on near infrared spectroscopy. It shows that theVetiver grass leaf is the main part in enrichment and storage of copper and lead, through thecontrast analysis of copper and lead content in Vetiver grass roots, stem, leaf tissues.Spectrum was optimized by different spectral preprocessing methods, combined with geneticalgorithm (GA) for variable selection, PLS modeling method is adopted to establish the near infrared quantitative analysis mathematics models for detecting the content of copper and leadin Vetiver grass roots, stems and leaves,20unknown samples were used to test the quality ofthese models, respectively. In the copper experimental group, the prediction correlationcoefficient of Vetiver grass root, stem and leaf tissue were0.932,0.835,0.876, respectively,the prediction root mean square error were0.069g/kg,0.024g/kg,0.106g/kg, respectively;In the lead experimental group, the prediction correlation coefficient of Vetiver grass root,stem and leaf tissue were0.636,0.603,0.832, respectively, the prediction root mean squareerror were0.039g/kg,0.028g/kg,0.034g/kg, respectively.(3) Explore the method of detecting content of copper and lead in Vetiver grass rootbased on Raman spectroscopy. Different pretreatment methods were used to optimize theRaman spectra of Vetiver grass roots, continuous projection algorithm (SPA) was applied tofilter the bands of Raman spectrum, PLS modeling method was employed, Ramanspectroscopy quantitative analysis models of copper and lead content in Vetiver grass rootwere set up,20unknown samples were used to test the quality of these models, respectively.The prediction correlation coefficient and the root mean square error were0.561,0.607and0.148g/kg,0.040g/kg, respectively.The above results verify the potential to detect the contents of heavy metals in plants bynear infrared and Raman spectroscopy. Through the contrast analysis, near infraredspectroscopy has higher precision and accuracy, and Raman spectroscopy possesses theadvantages of rapid and field testing.