EDXRF Least Squares Full Spectral Fitting Algorithm Without Standard Samples

The main purpose of the energy dispersive X fluorescence analysis is to analyze the composition and content of the unknown sample.After calibrate the X ray fluorescence spectrometer accurately,the composition of the unknown sample is determined by the characteristic X ray of the target element,while the element content is determined by the characteristic peak area of the element.After obtaining the net peak area of the target element,the relationship between element net peak area and element content can be obtained by corresponding fluorescence analysis methods,and the element content of unknown samples can be obtained.Because the real sample is composed of many elements,which will produce matrix effect and affect the accuracy of fluorescence analysis.In addition,before the X fluorescence analysis,we need to calibrate the detection efficiency of X fluorescence spectrometer,and the accurate detection efficiency calibration is also the foundation of fluorescence measurement.In the energy dispersive X fluorescence analysis,the passive calibration and matrix correction of the instrument is always a difficult problem in this field.At present,the"empirical influence coefficient method"is commonly used in the X fluorescence analysis.This method needs specific types of samples,Firstly,the relationship between the element content and the net peak area of elements is established by using a large number of specific national standard samples.The unknown sample will be measured after the fitting curve of the element content is obtained,and the unknown sample will be obtained after the net peak area is substituted into the fitting curve to get the content of the target element of the unknown sample,so the analysis process is cumbersome and relies on a large number of national standard samples.There is a great difference between the national standard samples and the unknown samples,and some are even difficult to define.Therefore,a new method is needed to analyze the unknown samples,and reduce the dependence between the samples and the standard samples.Finally,realize the energy dispersive X-ray analysis without standard samples.Based on the above problems in energy dispersive X-ray fluorescence,this paper studies the algorithm of the Energy dispersive?EDXRF?least squares full spectrum fitting without standard sample analysis,relying on the Sichuan Science and Technology Support Program?2015GZ0272?and the Sichuan Education Department?17ZA0033?.The main contents and achievements of the paper are as follows:?1?Theoretical analysis of energy dispersive least squares method without standard samples.The theoretical analysis of the energy dispersive least square method is studied,and the feasibility of applying the least square method to the energy dispersive X fluorescence analysis is discussed,and how to realize the analysis without standard samples is also discussed.And through the Java programming language,using the Android Studio development tool to develop the EDXRF least squares full spectrum fitting without standard sample algorithm program.?2?Obtain the basic parameters of fluorescence analysisThe X-ray spectral distribution??E?of Rh target was simulated by Monte Carlo MCNP program,which was divided into 102 channels.The basic parameters of fluorescence analysis such as absorption transition factor,spectral line fraction,fluorescence yield and mass absorption coefficient are obtained by using Java programming algorithm.?3?Making K,Ca,Ti,Fe,Ni,Cu,Zn,Sr,Sn and Pb single element standard samples to determine the sample matrix.After making the sample,the ratio R of the coherent and incoherent scattering peak intensity of the K_?peak in the Rh target spectrum was obtained by measuring the single element sample of K,Ca,Ti,Fe,Ni,Cu,Zn,Sr,Sn and Pb.Because the ratio of single element standard sample content is known,it can accurately calculate the effective atomic number Zeff of single element standard sample,and establish the relation between R and the atomic number Zeff of the single element standard sample,and get the fitting curveZeff-R.To solve the unknown sample,the unknown sample is measured,and get the R of the unknown sample,and the effective atomic number of unknown sample is obtained by substituting R into the fitting curveZeff-R.And the effective atomic number is decimal,therefore,the effective atomic number is used for this paper,the EDXRF of the least-squares full spectrum fitting-free-sample algorithm requires interpolation to more accurately compute the mass absorption coefficients of the sample to the incident ray and the sample to the characteristic ray of the target element.?4?Obtaining the _iG parameters of the instrument scale through the algorithmThe single element standard sample of K,Ca,Ti,Fe,Ni,Cu,Zn,Sr,Sn,Pb is obtained.The net peak area of the single element is obtained.The _iG parameter is obtained by the first fluorescence formula.Then the parameter is fitted to the characteristicK_?energy of the element,and the fitting functionG_i-E_K?of the parameter with the characteristic energy change is obtained.It should be used in the later iteration for the EDXRF least-squares method for the full spectral fitting of non-standard samples.?5?Obtain the standard deviation parameters?_i of Gaussian broadening distribution of theoretical spectral lines.The standard deviation parameters?_i for Gaussian broadening of element characteristicK_?peaks are obtained by Gauss fitting the element characteristic peaks of the actual measured spectral lines with OriginPro9 software by using the standard samples of K,Ca,Ti,Fe,Ni,Cu,Zn,Sr,Sn and Pb in the measurement laboratory.Then the characteristic K_?ray energy of each element is fitted with the standard deviation parameter?_i,the fitting function?_i-E_K?of Gaussian distribution standard deviation parameter with the element characteristic energy is obtained,and the standard deviation parameters of other elements can be obtained according to the fitting function when calculating the Gaussian broadening of other elements.?6?Theoretical spectral line solution and real unknown sample analysis.Application of Java programming language to achieve theoretical spectral line solution,and 14 national standards of soil samples for practical analysis.The results show that the accuracy of the EDXRF least squares full spectrum fitting method is mainly influenced by the element content in the sample.The greater the element content,the more accurate the element characteristic peak area obtained by the instrument measurement spectral line,the more accurate the content of the algorithm fitting,the smaller the relative error.When the element content in the sample is low,the accuracy of the element characteristic peak area obtained by the instrument measurement spectral line is lower,and the error of the analytic content is greater.The mean error of Ti and Fe elements is less than that of 15%,Cu element average error of24.97%,Zn element is 48.71%,and the content of Ni element with only dozens of is too low,and the content of Ni element is more error due to the factors such as scattering background and other factors.This article has the following innovative points:?1?Through theoretical research,determine the feasibility of the least squares method applied to energy dispersive X-ray fluorescence analysis,and compile the least-squares full spectrum fitting model-free sample analysis program,and apply it to the actual unknown sample analysis.?2?To prepare single element standard sample,to determine the sample matrix,to obtain the standard deviation parameters?_i of Gaussian distribution broadening,instrument calibration parametersG_i.Based on the basic parameter method of energy dispersive x fluorescence Analysis,the method of least squares is applied to the analysis of energy dispersive X-ray fluorescence without standard samples,through the analysis of the actual national standard soil samples,it is shown that the algorithm can accurately determine the sample matrix and accurately calibrate the instrument efficiency,and achieve the expected target...