Heterogeneities in samples of contaminated soil

Abstract: Many factors are important during sampling of contaminated soil. One of these is the sample sizes used. By estimation of the Heterogeneity Invariant (HI) it is possible to balance the sample size to one uncertainty contribution that is attached to sampling and sample handling. This is called the Fundamental error. Large overestimation of HI leads to larger samples and increased cost for sample handling. Large underestimation of HI leads to too small samples and to a larger uncertainty contribution, and possible skewed samples. Some methods to estimate HI is described and discussed from the perspective of contaminated soil sampling. Contaminated soil is a widely used concept, it is quite different challenges to sample if the contaminant is e.g. metallic particles, in-evaporated solutions or oils. None of the discussed methods is without any limitation. Field portable X-ray spectroscopy fluorescence is (XRF) is a popular instrument to analyse for metals in contaminated soils. It can provide quick and cheap analysis. The most prepared samples are dried, ground finely and put into Mylar thin film XRF sample cups. The sample volume is about 5-7 cm3, but the actually measured volume is only a fraction thereof. In the, likely, case of segregation in the sample volume this can lead to a biased measurement. Mechanisms for segregation in the XRF sample cup are discussed. Further suggestions are made to reduce the segregation. Despite precautions to reduce the effect of segregation it is difficult to eliminate it.