RESUMO
Laser-induced breakdown spectroscopy (LIBS) associated with machine learning algorithms (ML) was used to evaluate the Brachiaria seed physiological quality by discriminating the high and low vigor seeds. A 23 factorial design was used to optimize the LIBS experimental parameters for spectral analysis. A total of 120 samples from two distinct cultivars of Brachiaria brizantha seeds exhibiting high vigor (HV) and low vigor (LV) in standard tests were studied. The raw LIBS spectra were normalized and submitted to outlier verification, previously to the reduction data dimensionality from principal component analysis. Supervised machine learning algorithm parameters were chosen by leave-one-out cross-validation in the test samples, and it was tested by external validation using a new set of data. The overall accuracy in external validation achieved 100% for HV and LV discrimination, regardless of the cultivar or the classification algorithm.
Assuntos
Brachiaria , Lasers , Aprendizado de Máquina , Sementes , Análise Espectral/métodosRESUMO
This study is dedicated to the direct determination of base (B and Fe) and some rare earth elements (REE; Dy, Gd, Nd, Pr, Sm and Tb) in hard disk magnets. Five calibration strategies were tested and compared. Two of them are related to multivariate calibration: multiple linear regression (MLR) and partial least squares (PLS). Both presented adequate trueness values within a range of 80-120% for almost all analytes. The only exception was Tb, which was probably due to matrix effects. The use of MLR and PLS permits the testing of calibration models in the presence of interference, but matrix effects are not corrected. Because of this, three other univariate calibration methods were also tested and compared: multi-energy calibration (MEC), one-point gravimetric standard addition (OP GSA) and two-point calibration transfer (TP CT). These three calibration approaches permit matrix effects corrections, but an appropriate selection of the blank and standard is mandatory. The standard error obtained ranged from 0.01 to 6% using these univariate calibration methods.
RESUMO
This study proposes an analytical procedure for microwave-assisted sample preparation of dietary supplements for athletes using dilute nitric acid solution followed by determination of elemental impurities (As, Cd, Hg and Pb) by inductively coupled plasma mass spectrometry (ICP-MS) according to the United States Pharmacopeia Chapters 2232 and 233. Calibration strategies as internal standardization (IS), multi-isotope calibration (MICal), and one-point standard addition (OP SA) were applied for correction of matrix effects. The optimization of the sample preparation procedure was performed using Doehlert experimental design based on overall desirability results (residual acidity, dissolved organic carbon and recoveries reached for certified reference material of Typical Diet) for each calibration method evaluated. Accuracy was also evaluated by recovery experiments according to the permissible daily exposure specific for each element and samples were spiked with element concentrations of 0.5J and 1.5J in order to check accuracies for As, Cd, Hg and Pb. Recoveries ranged from 82 to 120% using IS, 90 to 125% using MICal, 88 to 120% using OP SA and the repeatability was demonstrated by a precision lower than 10% RSD. Ten samples of dietary sport supplements were analyzed using the three calibration methods evaluated and the concentrations of As, Cd and Pb determined in eight samples were lower than the limits established by the Chapter 2232.
Assuntos
Arsênio/análise , Cádmio/análise , Suplementos Nutricionais/análise , Chumbo/análise , Mercúrio/análise , Micro-Ondas , Ácido Nítrico/química , Calibragem , Limite de Detecção , Espectrometria de Massas , Padrões de Referência , Reprodutibilidade dos Testes , Oligoelementos/análiseRESUMO
This study aims to develop methods for determination of Ca, K, Mg and Na by laser-induced breakdown spectroscopy (LIBS) and Ca, K, Mg, Na, P, S, Fe and Zn by wavelength dispersive X-ray fluorescence (WDXRF) in pressed pellets bivalve mollusks. LIBS and WDXRF calibration models were built with references values determined by inductively coupled plasma optical emission spectrometry (ICP OES) after acid digestion. The calibration models for LIBS and WDXRF were obtained from 28 samples (14 for calibration and 14 for validation). It was possible to implement a validation between LIBS and WDXRF methods for elements Ca, K, Mg and Na. The proposed calibration model obtained using LIBS and WDXRF data presented a good correlation with reference values obtained by ICP OES.
Assuntos
Bivalves/química , Metais/análise , Frutos do Mar/análise , Análise Espectral/métodos , Animais , Cálcio/análise , Calibragem , Análise de Alimentos/métodos , Lasers , Espectrometria de Fluorescência/métodos , Raios XRESUMO
A new method is proposed for the simultaneous determination of Mo and Ni in plant materials by high-resolution continuum source graphite furnace atomic absorption spectrometry (HR-CS GFAAS), employing direct solid sample analysis (DSS) and internal standardization (IS). Cobalt was used as internal standard to minimize matrix effects during Ni determinations, enabling the use of aqueous standards for calibration. Correlation coefficients for the calibration curves were typically better than 0.9937. The performance of the method was checked by analysis of six plant certified reference materials, and the results for Mo and Ni were in agreement with the certified values (95% confidence level, t-test). Analysis was made of different types of plant materials used as renewable sources of energy, including sugarcane leaves, banana tree fiber, soybean straw, coffee pods, orange bagasse, peanut hulls, and sugarcane bagasse. The concentrations found for Mo and Ni ranged from 0.08 to 0.63 ng mg(-1) and from 0.41 to 6.92 ng mg(-1), respectively. Precision (RSD) varied from 2.1% to 11% for Mo and from 3.7% to 10% for Ni. Limits of quantification of 0.055 and 0.074 ng were obtained for Mo and Ni, respectively.