Research On Rapid Detection Quantitative Analysis Method Based On DP-LIBS

The development of efficient and rapid analysis techniques for the elements in crops,agricultural products,and agricultural production materials can provide important technical support for precision agriculture and have important significance for agricultural development.Laser induced breakdown spectroscopy(LIBS)allows quantitative and qualitative analysis of several elements at the same time,and has the following advantages:rapid detection,simple sample preparation,and environmental friendliness.This study is based on a collinear double-pulse laser-induced breakdown spectroscopy detection platform.Nitrogen,phosphorus,and potassium in compound fertilizers were treated as research objects.The effects of system parameters and quantitative analysis algorithms on simultaneous detection of multiple elements were explored.The impact of modeling methods on the simultaneous detection of multiple elements,and the relationship between the accuracy of single element detection and the detection efficiency of multiple elements were also analyzed.Taking heavy metal elements in rice plants as the research object,a quantitative analysis model of heavy metal elements was established.The main results are showed as follows:(1)Research on LIBS detection system:System parameter optimization method and quantitative model establishment of LIBS for nitrogen,phosphorus,and potassium detection in compound fertilizers were studied.In order to detect three elements at the same time,the influences of single and double pulses,laser energy,wavelength,pulse time interval,and excitation timing on the spectral signal were explored.According to the relationship between the spectral signal intensity and the laser energy,40 mJ energy was selected as the experimental parameter.Single and double pulse experiments were conducted,to optimize the time interval of the double pulse laser system according to the line intensity,and explore the influence of single pulse and double wavelength excitation timing of different wavelengths on the quantitative model.In the N,P,and K multi-elements analysis,the spectral intensity value collected under the 532 nm laser first excitation condition was large,and the time interval between the two laser pulses was optimized to 0.4 ?s;considering the accuracy,under the four experimental conditions,partial least square method was used to quantitatively analyze and model the nitrogen,phosphorus,and potassium in the compound fertilizer.The single-pulse phosphorus element at 532 nm had the best modeling effect,with the correlation coefficient of the PLS model as 0.90 and the correlation coefficient as 0.82.For the nitrogen and potassium elements,the best modeling effect was found under the DP-LIBS condition of 532 nm first excitation,with the correlation coefficients of the PLS model as 0.91 and 0.89,and the correlation coefficients of the verification results as 0.87 and 0.86,respectively.2)Establishment of LIBS quantitative model:Quantitative analysis methods and models of trace heavy metal elements Cu,Cd,and Pb in rice plants were established.Based on LIBS,the heavy metals Cu,Cd,and Pb in rice stems,leaves,and grains were quantitatively analyzed,and a univariate linear regression model of characteristic spectral lines and a partial least squares regression(PLS)model of characteristic bands were established.? In the linear regression model based on spectral line Cu I:324.7 nm,the correlation coefficients for stem,leaf,and grain were 0.79,0.81,and 0.80,respectively;in the partial least square regression(PLS)model based on the characteristic band,the corresponding correlation coefficients were 0.96,0.99,and 0.97,respectively,and the validation sets were 0.98,0.98,and 0.76,respectively.?In the linear regression model based on spectral line Cd I:328.8 nm,the correlation coefficients for the stem,leaf,and grain were 0.62,0.65,and 0.72;the corresponding correlation coefficients in the least squares regression(PLS)model were 0.98,0.96,and 0.82;while the correlation coefficients of the validation set were 0.96,0.98,and 0.75,respectively.? In the linear regression model based on spectral line Pb I:405.7 nm,the correlation coefficients of the stem,leaf,and grain were 0.76,0.93,and 0.66,respectively;in the partial least square regression(PLS)model,they were 0.98,0.99,and 0.92,while for the validation set they were 0.92,0.88,and 0.76,respectively.Overall,our findings showed that:Quantitative analysis of trace heavy metals in rice and macro nutrients in compound fertilizers can be achieved,after the LIBS system parameter optimization and model establishment.The technique has the advantages of rapid multi-element analysis as well as accurate single-element analysis.The findings can provide theoretical basis and technical support for LIBS instrument systemization and integration in the future.