Hyperlink 12. Within-Sample Heterogeneity

The nonuniform distribution of particles (within-sample heterogeneity) poses a challenge for obtaining a representative subsample for analysis. There are three things that can be done in the laboratory to address this: increase mass of subsample analyzed, reduce the size of particles in the sample (controlling fundamental error), or increase the number of increments that compose the analytical subsample (controlling grouping and segregation error) by using the techniques described in Section 6. Fundamental and grouping errors are discussed in Section 2.5.1.

Obtaining a representative subsample in the face of within-sample heterogeneity involves additional labor, time, equipment, and cost. But the alternative can be data that do not stand up to scrutiny, that cannot be defended as reliable, and that lead to erroneous decisions about risk and remedial design.

The exact procedures selected to prepare samples always depend on the numerous project variables related to soil type, contaminants of interest, staffing, budget, availability of equipment, desired workflow, number of samples, subsequent sample preservation or preparation steps, etc. Details of these procedures are provided in Sections 5 and 6.

Increasing the mass of soil that is digested or extracted manages within-sample heterogeneity by reducing FE. Increasing the mass increases the likelihood that the same ratio of particle types will be present in the analytical sample as in the original sample (see Figure H4-1). It also reduces the “nugget effect,” which produces very high concentration results when a “hot” particle is captured in a very small volume of “clean” soil particles. Because the soil mass used for typical metals analysis is very small (0.5–2 g, depending on the laboratory), the nugget effect can be particularly pronounced when samples are analyzed for metals. If it is not practical to increase the mass digested for metals analysis, then careful attention must be paid to sample preparation. Because analyses for organics typically involve 10–30 g of soil, the nugget effect is less of a problem than for metals; however, it can still occur. Therefore, thought needs to be given to what sample preparation is necessary to ensure an analytical sample that gives a concentration result representative of the matrix in the field.

The other way to address within-sample heterogeneity through reduction of FE is reduction of particle sizes through milling as discussed in Section 6.