Determinação direta de fósforo disponível em solos utilizando laser-induced breakdown spectroscopy (LIBS)

Abstract

Phosphorus is a vital element for sustaining life as we know it, being a constituent of nucleic acids, ATP, and cell membranes. Its proper availability in the soil is essential to achieving optimal crop yields, establishing it as a highly relevant macronutrient. Considering the advances in LIBS in recent decades, especially in soil analysis, this study aimed to explore methodologies to enable the quantification of this analyte in soil matrices. This work investigated the application of laser-induced breakdown spectroscopy (LIBS) as an alternative to conventional chemical analysis for determining plant-available phosphorus in agricultural soils. Regression and classification models were developed using the AGLIBS 2.0 instrumentation from Agrorobótica to quantify and classify extractable phosphorus content directly in solid soil samples. A total of 393 samples were collected from the states of Mato Grosso, Mato Grosso do Sul, and Goiás, prepared as pressed pellets for spectral acquisition. The LIBS data were correlated with results from the standard Mehlich-1 extraction method using ICP OES, which served as reference for model calibration and validation. Multivariate calibrations (MLR, PLS, PCR) outperformed univariate models, with validation standard errors around 19 mg/dm³. Classification approaches (KNN, SIMCA, and PLS-DA) achieved over 70% accuracy in discriminating agronomic phosphorus classes. Despite instrumental limitations and challenges in correlating LIBS signals to phosphorus availability, results suggest the technique’s potential as a complementary or alternative analytical tool for agronomic analysis. It presents advantages such as reduced chemical waste and faster analytical throughput.