Response Surface Methodology for Optimizing LIBS Testing Parameters: A Case to Conduct the Elemental Contents Analysis in Soil
Keqiang Yu,Yanru Zhao,Yong He,Dongjian He
DOI: https://doi.org/10.1016/j.chemolab.2019.103891
IF: 4.175
2019-01-01
Chemometrics and Intelligent Laboratory Systems
Abstract:Optimization of testing parameters are the prerequisite for laser induced breakdown spectroscopy (LIBS) further data analysis, which can offer important reference value for the soil detection in the field. This work investigated the influence of the main testing parameters laser energy (LE), delay time (DT), and lens to sample distance (LTSD) of LIBS system. Based on the spectral characteristic of main elements in soils, the testing parameters of LIBS for soil detection were obtained and verified. The optimization analysis of three testing parameters LE (50-160 mJ), DT (0.5-4.5 mu s), and LTSD (94-102 mm) were conducted by response surface methodology (RSM). Central composite design (CCD) in RSM was introduced to carry out experimental runs. The combined signal-background-ratio (SBR) of characteristic spectral lines from main elements (Si, Fe, Mg, Ca, Al, Na, K, etc.) in soil were defined as the objective function (named Y-SBR). The interaction influences among three independent variables (LE, DT, and LTSD) on soil plasma characteristics were explored and the optimized testing parameters of LIBS were summarized. Results revealed as follows: the factor LE showed a remarkable linear effect to Y-SBR, and factors DT and LTSD exhibited opposite results. The interactive items of three factors displayed a non-significant relationship. Meanwhile, the quadratic items of LE2, DT2 and LTSD2 offered significant surface relationships. Through the RSM analysis, the optimized testing parameters for LIBS soil detection were LE: 103.09 mJ; DT: 2.92 mu s; LTSD: 97.69 mm; and a peak value Y-SBR of 198.60. After that, the LIBS data of 21 representative soil samples were collected under the optimized LIBS testing parameters. Partial least squares regression (PLSR) was introduced to predict the main elemental contents. Results indicated that PLSR models offered promising outputs for predicting the contents of Al, Ca, Fe, K, Mg, and Na in the sampled soil, which revealed that the testing parameters of LIBS optimized by RSM were available. This work provided a theoretical basis for the accurate LIBS data analysis and regarded as a technical support for the field soil JIBS testing parameters selection.