Measurements of cadmium in soil extracts using multi-variate data analysis and electrochemical sensors

Abstract

There is a growing awareness of potential health risks due to exposure to heavy metals. One source of uptake is via agriculture, when heavy metals in the soil are taken up by the crop. The metal cadmium holds a special position, since it is considered to be a health risk, even at the low concentrations observed in our food supply, furthermore, it is ranked as eight on the top 20 hazardous substances list. Two measurement systems are described based on stripping voltammetry for analysis of cadmium. One is based on a three metal direct probe system (TMDPS) with three working electrodes (platinum, gold and rhodium), combined with a polishing unit, the other is an automatic flow through system, using one working electrode of gold, also equipped with a polishing unit. A number of different soils were extracted with an ammonium-lactate solution and analyzed with the systems, and the data obtained were subjected to multi-variate data analysis (MVDA). Using modeling based on partial least square (PLS), concentrations of cadmium in the soil extracts could be predicted for the TMDPS in the concentration area 0.5–10 μg/l with a root mean square error of prediction (RMSEP) of 0.8 μg/l and a relative predicted deviation (RPV) of 2.0. One sample could be analyzed in 4 min. It was also shown that by using different PLS models, the concentration of the elements copper, aluminum, lead and iron could be predicted. The possibilities of using the technique for field use were also evaluated by studies of mixtures of different soils in 0.1 M HNO₃ solution, the time for an analysis was, however, rather large, around 20 min.