ИСТИНА |
Войти в систему Регистрация |
|
ИПМех РАН |
||
Both qualitative (identification of emission lines) and quantitative analysis of low- and high-alloy steels by means of Laser-Induced Breakdown Spectroscopy (LIBS) were considered in this research. We developed the special package of DLLs to use the ICCD camera “Nanogate-2V” (Nanoscan Ltd., Russia) in LabVIEW® environment for spectroscopic measurements and to perform a pretreatment of LIBS data1. The main technical characteristics (signal-to-noise ratio, dynamic range, instrumental function) of the laboratory setup built with a Czerny-Turner 32-cm spectrometer and the camera were carefully measured. We corrected the variations in multi-channel plate (MCP) amplification throughout its surface. Calibration was performed by direct uniform illumination of the detector surface by stable light source – lead hollow-cathode lamp. The spectral sensitivity curve of our system within the range 205-930 nm was determined via calibrated deuterium and tungsten-halogen lamps (StellarNet, Inc., USA). We applied our approach2 based on correlation of model spectra with experimental one for lines identification in LIBS spectra of steels. This procedure improves the accuracy of spectra identification and reduces its time costs. Since 2013 we sufficiently improved the procedure of synthetic spectra generation. Relative intensities were calculated by Saha-Boltzmann equations (i.e. plasma is under LTE), and afterwards spectral lines were convoluted by Voigt profile, including Doppler, collisional and Stark widths, and by instrumental profile. At the moment MySQL database is used for all data storage, including transitions fundamental parameters. The data of Stark broadening of emission lines are poor, but supplying from literature as possible. Finally, we tested several approaches to improve the accuracy of determination of both metallic (Ni, Cr, Mn, Al, V, Ti) and non-metallic (Si, C) components in steels. We applied multivariate principal component regression in the case of spectral interference of analytical lines with iron ones for high-alloy steels3. But the best results can be obtained with isolated analytical lines with appropriate internal standard line of iron. The calibration curves with coefficient of determination R2>0.996 were obtained for metals detemination with the use of common lamp-pumped Nd:YAG laser (Lotis Tii, Belarus) operated at 532 nm with the energy ~60 mJ/pulse with ring mode beam structure. We also demonstrated, that a portable DPSS Nd:YVO4 laser (Laser-export Co. Ltd., Russia, 527 nm, ~400 μJ/pulse, Gaussian beam profile) is suitable for steel analysis and, moreover, provided the significant reduction of matrix effect with absence of outlier samples in the calibration curves.