激光诱导击穿光谱技术在沉积物元素定量分析中的应用  

Application of Laser-Induced Breakdown Spectroscopy in Quantitative Analysis of Sediment Elements

作　　者：付晓芬 宋友桂[1,2] 张明玉 冯中琦 张大成 刘慧芳 FU Xiao-fen;SONG You-gui;ZHANG Ming-yu;FENG Zhong-qi;ZHANG Da-cheng;LIU Hui-fang(State Key Laboratory of Loess and Quaternary Geology,Institute of Earth Environment,Chinese Academy of Sciences,Xi'an 710061,China;CAS Center for Excellence in Quaternary Science and Global Change,Xi'an 710061,China;Xi'an Institute for Innovative Earth Environment Research,Xi'an 710061,China;School of Physics and Optoelectronic Engineering,Xidian University,Xi'an 710071,China)

机构地区：[1]中国科学院地球环境研究所黄土与第四纪地质国家重点实验室,陕西西安710061 [2]中国科学院第四纪科学与全球变化卓越创新中心,陕西西安710061 [3]西安地球环境创新研究院,陕西西安710061 [4]西安电子科技大学物理与光电工程学院,陕西西安710071

出　　处：《光谱学与光谱分析》2024年第3期641-648,共8页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(41977385,42002205);第二次青藏高原综合考察研究项目(2019QZKK0101)资助。

摘　　要：激光诱导击穿光谱(LIBS)可以快速测量样品中各种元素的含量或组成,被广泛应用于环境样品的测试分析,但对第四纪沉积物的分析还鲜有报道。该实验是以青海湖岩芯第四纪湖泊沉积物和国家标准土壤样品为研究对象,利用Savitzky-Golay卷积平滑和标准正态变量变换(SNV)联用对原始光谱进行预处理,然后结合单变量定标分析、偏最小二乘回归(PLSR)对青海湖沉积物样品中Na、Ca、Mg、Si、Al、Fe、Mn、Sr和Ba 9种元素进行定量分析。以交叉验证的结果作为PLSR模型参数寻优的标准,分别以预测决定系数(R2)、交叉验证均方根误差(RMSECV)、预测均方根误差(RMSEP)和相对分析误差(RDP)作为评估PLSR模型的定量精度和稳定性。结果表明,PLSR算法显著改善了传统单变量分析的定量效果,预测决定系数分别为0.94、0.94、0.98、0.94、0.97、0.84、0.89、0.98和0.76,相对分析误差分别为2.74、2.35、3.27、2.97、3.56、1.68、1.54、4.18和0.75,结合交叉验证均方根误差及预测均方根误差结果可知,LIBS技术结合PLSR算法对Na、Ca、Mg、Si、Al和Sr元素的预测准确度较高,而Fe、Mn和Ba元素的定量效果不是很理想,说明PLSR算法在预测精度和适用性方面存在一定的局限性。为进一步探究LIBS技术应用于元素地球化学指标测试的可行性,利用LIBS预测含量比值与参考含量比值进行对比,二者曲线变化趋势基本一致,验证了LIBS技术应用于沉积物元素地球化学的可行性和有效性。为第四纪沉积物样品中元素定量分析提供了可靠方法,也为古气候古环境重建提供新的技术手段和研究思路。Laser-induced breakdown spectroscopy can quickly measure the content or composition of various elements in samples and is widely used in the testing and analysis of environmental samples.However,its application to analysis of multiple elements in geological samples is rarely reported.This study took the drill-core Quaternary Lake sediments of Qinghai Lake and national standard soil samples as the research objects.The original spectra were preprocessed by Savitzky-Golay convolution smoothing and standard normal variable transformation,and through univariate calibration analysis as well as partial least squares regression algorithm to quantitatively analyze nine elements of Na,Ca,Mg,Si,Al,Fe,Mn,Sr and Ba in Qinghai Lake sediment samples.The results of cross-validation were used as the criteria for optimizing the parameters of the PLSR model,and the quantitative accuracy and stability of the PLSR models were evaluated by the coefficient of prediction determination(R 2),root mean square error of cross-validation(RMSECV),root mean square error of prediction(RMSEP)and residual predictive deviation,respectively.The results show that the PLSR algorithm significantly improves the quantitative effect of traditional univariate analysis;the coefficients of determination for prediction are 0.94,0.94,0.98,0.94,0.97,0.84,0.89,0.98 and 0.76,and the relative analysis errors are 2.74,2.35,3.27,2.97,3.56,1.68,1.54,4.18 and 0.75.Combined with the results of cross-validation root mean square error and prediction root mean square error,it can be seen that LIBS technology combined with the PLSR algorithm has high prediction accuracy for Na,Ca,Mg,Si,Al and Sr elements.However,the quantitative effects of Fe,Mn and Ba elements are not very satisfactory,indicating that the PLSR algorithm has certain limitations and applicability in the prediction accuracy.In order to further explore the reliability of the LIBS technique is applied to index test of geochemical elements,this paper compared the predicted content ratio of LIBS with the referen

关 键 词：激光诱导击穿光谱 偏最小二乘回归 定量分析 第四纪沉积物 古环境 

分 类 号：O433.4[机械工程—光学工程]