Análise multivariada de dados aplicada a pesticidas

Abstract

With the increase in population, the world demand for food grows vertiginously. To supply this need, human beings develop and sell products, known as pesticides, capable of inhibiting pests that devastate plantations. However, with the increasing production of these compounds, there is a need to monitor their physicochemical properties to delineate their reflexes with the environment and with the organisms present in it. Thus, this work aims to analyze the physical-chemical behavior of monitored pesticides in products for human consumption. For this, 246 pesticides were selected based on the Codex Alimentarius and Ordinance GM/MS Nº 888 and the influences caused by properties such as molar mass, water solubility, octanol-water partition coefficient, vapor pressure, polar surface area, molecular surface area and polarizability were tested by Principal Components Analysis (PCA) and Hierarchical Cluster Analysis (HCA). With HCA, the formation of six clusters can be observed, in which three clusters were formed by only one pesticide with atypical behavior (guazatine, cabosulfan, sulfuryl fluoride) and the remainder gathered pesticides with similar chemical behavior, even though they belong to different chemicals classes, such as the existing grouping among pesticides fenbutatin oxide, spinetoram, fenpicoxamide, afidopiropen and flubediamide. With PCA, the pesticides etofenprox, DDT and carbosulfan were grouped by their high octanol-water partition coefficient values, the pesticides fenbutatin oxide, spinetoram, fenpicoxamid, afidopyropen and flubendiamide were grouped by their high values of molar mass, molecular surface area and polarizability, the pesticides sulfuryl fluoride and bromomethane were grouped by high values of vapor pressure and the pesticide guazatine evidenced by its high value of solubility in water and polar surface area. With this, it was possible to estimate their chemical behavior and their dynamics with the environment.