Before recommending an herbicide, it is necessary to know its interactions with the soil attributes that will influence its sorption and half-life in the soil. This knowledge is an essential condition for minimizing any negative effects that may result from its application. However, due to the great diversity of soil and climate in Brazil, especially for products of recent use in the country, such as the herbicide tembotrione, this knowledge is often not available. This may be the main cause of occurrence of intoxication reports in crops carried out in succession to corn sprayed with this herbicide. In order to reduce possible impacts on successor crops and the contamination of surface and groundwater using the high-performance liquid chromatography, the tembotrione half-life was determined in soils with different attributes. The evaluated soils were a Red-Yellow Latosol with and without liming (Viçosa, MG), Red-Yellow Latosol (Rio Paranaíba, MG), Yellow Latosol (Sooretama, ES), and Red-Yellow Latosol with and without liming (Gurupi, TO). The results showed that liming might favor tembotrione degradation in the soil, as well as direct relationships between values of tembotrione half-life and contents of clay and soil organic matter. Tembotrione half-life in samples of Red-Yellow Latosol (without liming) collected in Viçosa and Rio Paranaíba, MG, was higher than 90 days, indicating carryover risks in successive crops to corn sprayed with tembotrione in these regions.

Studies on herbicide behavior in soils may be performed using biological and chemical methods. The efficiency of the biological method depends on the indicator species sensitivity to low herbicide concentrations in the soil solution. Among the herbicides commonly used in Brazil for corn cultivation, tembotrione stands out. In the last agricultural seasons, the intoxication of some crops when cultivated in succession to corn has been reported, which may be attributed to tembotrione recommendations without the knowledge of their interactions with colloids of tropical soils. In this research, we selected an indicator plant species of tembotrione residues in the soil after sorting 12 species. The sorption of this herbicide was estimated in two Oxisols (Latossolo Amarelo and Latossolo Vermelho-Amarelo, Brazilian Soil Classification) and in a Histosol (Organossolo, Brazilian Soil Classification) with different attributes. Initially, after sorting the 12 plant species, the two most efficient in detecting tembotrione in the soil were selected. In the second stage, the most sensitive species was selected. In the third stage, tembotrione sorption was assessed in the three soils by using the selected species. Among the 12 species, Beta vulgaris and Brassica oleracea var. capitata were the most efficient in detecting tembotrione at low concentrations in the soil, being B. vulgaris (beet) the most sensitive. Thus, due to its ease of cultivation, high sensitivity, and fast initial growth, beet can be used as an indicator species of tembotrione presence in the soil solution. The sorption of this herbicide was higher in the Histosol and it is directly related to the organic matter content.

As an alternative to control of resistant biotypes, farmers have resorted to the use of herbicides that have residual soil activity and good pre-emergence action. One of the most used mixtures in Brazil is that of sulfentrazone + glyphosate. When a herbicide has a pre-emergence action, small leaching of its molecules is necessary for its superficial incorporation into the soil, where most of the weed seeds are found. However, if excessive leaching occurs, the herbicide may have reduced efficiency in controlling weed, increasing the risks of groundwater contamination. In this study, sulfentrazone leaching was evaluated in columns with soil, in isolated applications and in mixture with formulations of glyphosate (Roundup Ready®, Roundup Ultra® and Zapp Qi®). For this, the columns were filled with samples of Red-Yellow Latosol, collected in the 0-20 cm depth layer. After moistening the columns to near field capacity, the herbicides were applied to the top and, 24 hours later, a 60 mm rainfall was simulated. After 72 hours, the columns were sectioned in 5 cm segments, collecting the soil from each segment. In these samples, indicator plants (Sorghum bicolor) were cultivated for the presence of the herbicide. It is concluded that, in general, glyphosate does not alter the leaching potential of sulfentrazone in the Red-Yellow Latosol. However, mixtures with Roundup Ready® and Zapp Qi® reduce the sorptive forces of sulfentrazone to soil colloids. Therefore, in some types of soil, herbicide desorption may occur with greater intensity, making its availability in the soil solution more difficult. This may result in better control of plants and/or intoxication of sensitive crops.