激光诱导击穿光谱
偏最小二乘回归
分析化学(期刊)
化学
分析物
化学计量学
光谱学
校准
多元统计
单变量
等离子体原子发射光谱
感应耦合等离子体
数学
色谱法
等离子体
统计
物理
量子力学
作者
João Victor Borges Assis,Dennis da Silva Ferreira,Daniela de Assis Bócoli,Carlos Henrique Hoff Brait,Edenir Rodrigues Pereira-Filho
标识
DOI:10.1177/00037028231217974
摘要
This study was dedicated to developing analytical methods for determining macronutrients (Ca, K, and Mg) in soy leaf samples with and without petioles. The study's primary purpose was to present Laser-induced breakdown spectroscopy (LIBS) as a viable alternative for directly analyzing leaf samples using chemometric tools to interpret the data obtained. The instrumental condition chosen for LIBS was 70 mJ of laser pulse energy, 1.0 µs of delay time, and 100 µm of spot size, which was applied to 896 samples: 305 of soy without petioles and 591 of soy with petioles. The reference values of the analytes for the proposition of calibration models were obtained using inductively coupled plasma optical emission spectroscopy (ICP-OES) technique. Twelve normalization modes and two calibration strategies were tested to minimize signal variations and sample matrix microheterogeneity. The following were studied: multivariate calibration using partial least squares and univariate calibration using the area and height of several selected emission lines. The notable normalization mode for most models was the Euclidean norm. No analyte showed promising results for univariate calibrations. Micronutrients, P and S, were also tested, and no multivariate models presented satisfactory results. The models obtained for Ca, K, and Mg showed good results. The standard error of calibration ranged from 2.3 g/kg for Ca in soy leaves without petioles with two latent variables to 5.0 g/kg for K in soy leaves with petioles with two latent variables.
科研通智能强力驱动
Strongly Powered by AbleSci AI