无标样EDXRF基体效应及含量校正的微分算子模型研究

Non-standard and auto analysis is the main development tendency of Energy Dispersive X-ray Fluorescence (EDXRF), while the matrix effect is an intrinsic problem to EDXRF. This program will study non-standard EDXRF technology and correction method, based on geochemistry, non-destructive nuclear analysis, applied and computational mathematics. Firstly, the differential operator series response spectrum will be established of element characteristic X-ray spectrum..The Monte Carlo simulation flux will be fusion broadened by combining with the detector response function (DRF) model, and form the digital simulation spectrum library. Secondly, study the matrix effect correction method based on differential operators, establish the function between element content and characteristic X-ray physical cross section. The third, study a kind of full spectrum quantitative analysis method for EDXRF technology. Establish the mathematical equation between elements content and X-ray spectra, and get the accurate content after corrected by iterative computations procedure based on weighted least-squares and differential operator correction model. Finally, we will develop the analysis programme to analyze inhomogeneous complex geological sample, and complete the cooperation and evaluation. The innovation point of this study is to achieve non-standard EDXRF technology and matrix effect correction new technology. The prospect result will help to promote the automatic level of EDXRF, and has very important scientific significance and application values.

无标样自动化是能量色散X荧光（EDXRF）技术的主要发展趋势，而基体效应校正是该技术面临的固有难题，本项目利用放射性物探方法及计算数学，开展无标样EDXRF中基体效应校正及含量校正方法研究。首先，利用探测器响应函数(DRF)对Monte Carlo模拟注量进行融合展宽，再利用微分谱进行泰勒级数二次展宽，获得与实验谱相吻合的数字化模拟谱；其次，建立元素含量与特征X射线物理截面的函数，进行微分算子基体效应校正方法研究，实现对元素含量的校正；再次，进行无标样EDXRF定量分析方法研究，建立元素含量与数字化模拟谱间的数理方程及权重残差微扰校正模型，利用权重最小二乘迭代计算出校正后的元素含量；最后，开发分析程序，对非均匀复杂地质样品分析，完成方法的评价。本项目以实现无标样EDXRF准确分析和基体效应校正新方法为创新点，预期成果有助于促进技术的自动化，在X射线分析技术领域有着较好的科学意义和应用价值。

本项目针对如何利用MonteCarlo模拟技术实现特征X射线荧光技术对元素的无标样分析，研究完成了宽能段X射线DRF模型和Monte Carlo 模拟注量融合展宽算法以及泰勒级数二次展宽算法研究。实验测量获得了Fe、Cu、Ni、Pb等十余种金属元素纯样（化学分析纯）及其若干个混合样品的X射线能谱，对各元素的特征X射线进行了能量刻度，保证了元素的定性分析准确度。根据各特征能谱分析，建立了包含高斯峰函数、阶梯型本底函数、Si逃逸峰函数和低能侧拖尾函数的探测器响应函数。利用Monte Carlo模拟，得到各元素的特征X射线注量谱，利用探测器响应函数，建立了元素谱线融合展宽算法，实现了模拟与实验的连接。为了获得更精确的能谱拟合，研究建立了 泰勒二次展宽算法，利用泰勒级数模型，进行了微分建模研究。最后，对研究建立的能谱分析算法，对一批混合样品进行了多元素无标样EDXRF定量计算，对Fe、Cu、Mn、Ni、Cr、Zn元素的分析误差最小为0.01%，最大为0.23%，与工业要求的1%的误差相比，达到应用要求。.受本项目资助的研究成果已发表3篇SCI论文。