The Development Of Energy Dispersive X-ray Fluorescence Spectrometer

X-ray fluorescence analysis method has an advantage of nondestructive, multi element analysis,high accuracy, fast analysis, and environmental, which applicable for the production process detectionand control. It has a broad market prospect and important research significance. This paper aims at thedemand of the RoHS detection. With analyzing the theory foundation of the x-ray fluorescenceanalysis method, we define the working principle of the X-ray fluorescence spectrometer and themethod of spectral analysis software.To obtain the characteristic X-ray of the sample and achieve automation of the spectrometer, wedesign the whole hardware platform of the spectrometer, which include the optical system and controlsystem. The optical system take charge of exciting the sample and collecting the characteristic X-ray,which is composed of X-ray tube, filter, collimator and Amptek X-123detection systems. The controlsystem achieves the control to the optical system and the communication between each part of thesystem, which is based on the embedded processor STM32F103R8T6. With the requirement ofanalytical conditions, we design the embedded system hardware form the angle of functional integrity.We introduce the selection principles of filter and collimator. Besides, we design circuit to realizeautomatic replacement of filter and collimator based on the A3983chip and stepping motor. Withanalyzing the tube voltage and tube current which can affect the energy of X-Ray, we design D/Acircuit to control the energy of X-ray based on the LTC2704-16. Meanwhile, the hardware circuit ofUSB HUB is designed. And PCB layout of the circuit is designed in Altium Designer.After the design of hardware, this paper presents a design method of the fluorescence spectrumanalysis to get the category and content of elements in sample. The method includes qualitativeanalysis and quantitative analysis. Besides, we develop the analysis software with Delphi. Qualitativeanalysis is realized through Spectrum processing. First, we use Savitzky and Golay Polynomialsfilters to smooth the spectrum. Next, in order to get characteristic peaks, we remove backgroundbased on the straight-line method and the peak stripping method. Through using the second derivativemethod, we can obtain the position and height of spectral peaks. Also, we get the energycorresponding to the every peak by energy calibration. Meanwhile, we fit area of the characteristicpeak that is the intensity of characteristic X-ray by Gaussian function. We establish the mathematicalmodel of fundamental parameter method for the correction of the matrix effect in Quantitativeanalysis process.At last, this paper introduces the actual effect of analysis software. Also, Combined with hardware platform, the software is applied in testing. The results show that the spectrometer achievesthe expected performance index and meets practical demands. Also, it has high automation degree.Besides, the data processing of spectrum is visualized and easy-management.