TY - CONF A1 - Rühlmann, Madlen A1 - Büchele, Dominique A1 - Ostermann, Markus A1 - Schmid, Thomas A1 - Paul, Andrea T1 - Design of an online-analysis technique for the determination of major and minor nutrients in soils using DP-LIBS N2 - LIBS (laser-induced breakdown spectroscopy) is known as a fast and simultaneous multi-element analysis with little or no sample preparation. Because of that, over the last few years there has been a growing interest in applications of LIBS in the field of agriculture. As part of the National Research Strategy BioEconomy 2030 the German Federal Ministry of Education and Research (BMBF) started an innovation programme called BonaRes. BonaRes consists of ten interdisciplinary research project as-sociations which are dealing with soil as a sustainable resource for the bio-economy. One of these research projects is I4S (intelligence for soil) which has the goal to develop an integrated system for site-specific soil fertility management. This system includes a sensor platform, which contains different sensors, like XRF, VIS-NIR, Gamma and LIBS. The main task of LIBS measurements in this project is the real time determination of the elemental contents of major and minor nutrients in soils, like calcium, magnesium, potassium. For this purpose, a special setup has been designed. The sample uptake operates with the help of a rotatable sample plate, on which the loose soil sample can be placed in form of a track. The sample plate circulates with different velocities to simulate the application on the field. To provide a higher intensity and a better re-producibility of the obtained signal, a double-pulse Nd:YAG laser (1064 nm) was used. In order to minimize dust formation from the soil during the operation of the laser, a dust removal by suction has been integrated When using relative methods such as LIBS, a suitable calibration curve is needed for absolute quantification. Within the I4S project a large number of soil samples from different testing grounds with known data of chemical composition, texture etc. is available. For the first calibration curves seven soil samples from different grounds in Germany were prepared as reference materials and four certified reference materials from China and Canada were purchased. With the help of these reference materials, calibration curves for different elements were initially calculated based on internal standard addi-tion. Copper was used as internal standard because of its low concentration in soils. The complex matrix of soils, as well as the influence of moisture and grain size in soils makes the absolute quantification by LIBS challenging. To overcome these influences, a calibration curve based on multivariate analysis was generated. Therefore, baseline correction on the second derivative with a Savitzky-Golay filter was followed by a partial least squares regression (PLSR). Multivariate analysis leads to noise reduction and neglection of interfering signals. Therefore, a better correlation between signal intensity and nutrient concentration is observed and a robust calibration curve is obtained. T2 - 9. Euro-Mediterranes Symposium LIBS CY - Pisa, Italy DA - 11.06.2017 KW - LIBS KW - Soil KW - Matrix KW - Multi-element PY - 2017 AN - OPUS4-40807 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -