Assessing and mapping the elemental composition of coffee leaves and fruits using portable X-ray fluorescence spectrometry

Abstract

Foliar analysis is essential for monitoring the nutritional status of plants, ensuring high productivity and quality. Traditionally, this analysis is performed using wet digestion methods, which are time-consuming and require large amounts of chemical reagents. Portable X-ray fluorescence spectrometry (pXRF) has shown high performance in assessing various earth materials and holds the potential for characterizing plant leaves and fruits. This study hypothesizes that pXRF measurements on fresh and intact coffee leaves can contribute to assessing the nutritional status of plants compared to conventional laboratory methods. This experiment aimed to verify the sensitivity of pXRF in monitoring the nutritional status of coffee plants. The experiment was conducted in a 67-hectare coffee area in the Macaia district of Bom-Sucesso, Minas Gerais, where 35 georeferenced points were randomly established. At each point, eight diagnostic leaves were collected and directly analyzed on the adaxial surface using a pXRF Bruker S1 Titan model operating in Geoexploration mode (dwell time of 90 seconds). After in situ pXRF measurements, samples from each point were oven-dried at 60°C, ground, and re-analyzed using: i) pXRF Bruker S1 Titan in Geoexploration mode (90 seconds); ii) pXRF Bruker Tracer 5g calibrated for plant materials; iii) ICP-OES after acid digestion. Additionally, green and mature fruits from the same points were collected, oven- dried, ground, and analyzed via pXRF Bruker S1 Titan and ICP-OES after acid digestion. Results showed that pXRF measurements on oven-dried and ground samples using the Titan and Tracer models were similar to the ICP-OES method. For K, Mg, and S, in situ, pXRF measurements were comparable to other methods, although P and Ca concentrations in situ were underestimated. All methods were comparable for micronutrients except for Zinc (Zn). The pXRF sensor could differentiate green and mature fruits in principal component analyze (PCA) and showed a strong correlation between leaves and green fruits. Similarities were observed between pXRF and ICP-OES in spatial distribution maps of elemental composition.