Prediction Methods And Longitudinal Distribution Patterns Of Soil Heavy Metals Based On Colinear Double Pulse LIBS

Soil is the habitat on which human beings depend for survival.With the accelerated industrialization,cadmium（Cd）,as a harmful heavy metal element,will accumulate in soil and affect the number and species of soil microorganisms,destroying the stability of soil ecosystem.Meanwhile,Cdin the soil will migrate longitudinally with time and moisture,and in the process of migration Cdis easily absorbed by plant roots,thus inhibiting the normal growth of plants,and also endangering human health by polluting groundwater and food chain.Therefore,it is necessary to detect the soil Cdcontent and reveal its longitudinal migration and distribution pattern in soil.Laser-induced breakdown spectroscopy（LIBS）,as a new spectroscopic technique,has the advantages of simple sample pretreatment,rapid and micro-damage,and is gradually becoming one of the effective tools for soil heavy metal detection.In this study,the Cdcontent in brown soil of Shaanxi Province was quantitatively analyzed by using double pulse（DP）-LIBS with machine learning methods,and the longitudinal migration of Cdwas simulated at different vertical depths from 0 to 50 cm to predict the Cdcontent in Shaanxi soil and explore its longitudinal distribution.The main findings and conclusions of this paper are as follows:（1）Soil heavy metal Cdelemental spectral signal enhancement and optimization of key parameters.In this study,60 pancake shaped soil samples were prepared by adding 15gradients of cadmium chloride（CdCl₂）aqueous solution to simulate soils contaminated with Cdat different levels;the Cdspectral signal enhancement was achieved by using a colinear double-pulse experimental setup with a colinear optical path and lasers of 532 nm and 1064nm wavelengths,and the results showed that the Cdspectral intensity of DP-LIBS was increased by 1.16 times and 5.00 times,and the signal-to-back ratio was increased by 1.24times and 1.60 times,respectively,compared with the single-pulse LIBS with the wavelengths of 532 nm and 1064 nm.This result is a prerequisite for accurate analysis of the LIBS data at a later stage.（2）Quantitative analysis of soil heavy metal Cdbased on DP-LIBS technology.The characteristic spectral lines of Cdwas selected through the National Institute of Standards and Technology（NIST）atomic spectral database and the spectral lines were 357.5-362.5 nm.The spectral data are preprocessed by median absolute deviation（MAD）,multiple scattering correction（MSC）,wavelet transform denoising,and data normalization and normalization..Partial Least Squares Regression（PLSR）,Particle Swarm Algorithm（PSA）optimized Least Squares-Support Vector Machine（LS-SVM）,and Genetic Algorithm（GA）optimized Back Propagation Artificial Neural Network（BP-ANN）was applied to quantitative analysis.The Coefficient of Determination（R²）and Root Mean Square Error（RMSE）evaluated the effectiveness of the models.The results showed that the PSA-optimized LS-SVM predicted the best results with the test set R² of 0.999 and the mean RMSEP of 0.359 mg/kg.This result indicated that DP-LIBS is an effective tool for soil heavy metal Cddetection.（3）Research on the longitudinal distribution pattern of soil heavy metal Cdelements in brown soils of Shaanxi province.In this study,CdCl₂ aqueous solution at the allowed concentration under the US Environmental Protection Agency’s general soil screening criteria was added to the soil column to simulate Cd-contaminated soil,and distilled water was added to simulate natural irrigation and precipitation.The LS-SVM model with PSA optimization was used to invert the time-varying soil heavy metal Cdcontent at different vertical depths using the colinear DP-LIBS.The results showed that the soil heavy metal Cdcontent was significantly negatively correlated with the depth of the soil column,and was mainly concentrated at 0-10 cm,and less distributed below 40 cm,and the Cdcontent slowly migrated downward with time and water.The results provide a theoretical basis for the development of remediation and prevention technologies of soil heavy metal Cdpollution,and also protect the green and safe agricultural production.