Water Quality Detection By Laser Induced Breakdown Spectroscopy

In recent years,the excess emission of industrial wastewater and domestic sewage has become more and more serious to water pollution due to the rapid development of industrialization and urbanization.Various salt substances in sewage enter the soil through a series of chemical reactions such as ion exchange and transfer,which promotes the transfer of calcium-soluble salts in the soil to the water,causing an increase in hardness in the water.The traditional water hardness testing method needs long measurement time,which can't achieve on-line monitoring,introduce secondary pollution,complicated operation,narrow detection range,etc.For the current problem,laser induced breakdown spectroscopy?LIBS?was used to study the high-precision detection method of quantitative analysis Ca element in Single solution and mixed solution.LIBS was widely used in various elemental analysis without the need for pretreatment of samples,short analysis time,simultaneous multi-element analysis,no secondary contamination of samples,and unlimited sample morphology.In this paper,the LIBS online detection platform based on liquid jet injection mode was built,and the data analysis algorithm was combined to optimize the quantitative analysis results.The main contents include as flow:?1?For the problem that the continuous background interference was strong in the experiment and the accuracy of the quantitative analysis is low.The baseline calibration data processing method was used to calibrate the collected spectra and then perform internal standard quantitative analysis.For the calibration curve Ca II 393.366 nm and Ca II 396.847 nm,the R² phase was significantly improved compared with baseline calibration from 0.9498 and 0.9156 to 0.9752 and 0.9667,respectively.At the same time,the limited of detection was also increased from 30.68ppm and 32.05ppm to 27.99ppm and 28.10ppm.?2?Since the detection accuracy of the internal standard method after calibration was not high,the quantitative analysis was carried out by partial least squares method firstly,and the fitting coefficient R² of Ca element was 0.9842.At this time,the fitting coefficient was still not high.Based on this,the least squares support vector machine was used for quantitative analysis.The final fitting coefficient R² increased from 0.9752 to 0.9947,RMSEP decreased from 1.3031wt.%to 0.2358wt.%,and ARE decreased from 8.6929%to1.7487%.?3?The water quality analysis environment is complex.In order to simulate the water quality analysis environment,the experiment was configured with mixing solutions for CaCl₂ and Mg?NO₃?₂,CaCl₂ and NaCl,CaCl₂ and Cu?NO₃?₂,CaCl₂ and Mg?NO₃?₂,NaCl and Cu?NO₃?₂.The four mixed solutions were quantitatively analyzed by the calibration model of the least squares support vector machine.The mixed solution of CaCl₂ and NaCl obtained the highest fitting coefficient of 0.9932.The RMSEC,RMSEP,and ARE errors were 0.5823wt.%,0.3347 wt.%,and 2.1326%.Based on the least squares support vector machine algorithm?LSSVM?,the accuracy of quantitative analysis of Ca ions in water was improved.The feasibility of LIBS for water quality detection and the accuracy of least squares support vector machine algorithm were verified by quantitative analysis results in single solution and mixed solution,which providing an important basis for the application of LIBS technology water quality testing.