Hyperlink 4. The Perils of Sampling Particulate Materials and Effects on Analytical Results

Concentration can be very high while the actual amount of the contaminant is low. The reported concentration depends on the mass of the sample that is analyzed and how it was prepared prior to analysis.

Figure H4-1 illustrates adsorption of contaminant to mineral grains Since iron (Fe) hydroxide is in a particulate form within the soil matrix, the contaminants sorbed to the iron hydroxide particle also behave as particulates. The light-colored material covering certain grains is arsenic. The concentration of a contaminant such as arsenic depends on how many particles of arsenic-laden iron complex are present within a “cleaner” silicate matrix. For the soil particle in Figure H4-1, the total mass of arsenic on the iron particles is miniscule (on the order of 0.005 µg), but the amount of total, mostly nonarsenic bearing, soil matrix is also very small (approximately 1 µg). Therefore, on a concentration basis (mass of contaminant per mass of soil) expressed in typical units, an analytical result would come out very high (approximately 5000 mg/kg). Figure H4-1 illustrates the important point that concentration is not the actual measure of exposure. The mass, the actual amount of contaminant that is present, is the measure of exposure. If a receptor ingested the 1 µg soil particle in Figure H4-1, the actual mass of arsenic ingested is only 0.005 µg. The concentration value for the particle, 5000 mg/kg, however, makes the exposure appear much worse than it really is.

Figure H4-1. Arsenic-laden iron hydroxide particles within a “clean” silicate matrix.

AS = 5,000++ mg/kg

Figure H4-1. Arsenic-laden iron hydroxide particles within a “clean” silicate matrix. The total masses of the soil particle and arsenic are 1 µg and 0.005 µg, respectively.
Photograph courtesy of Roger Brewer, HDOH.

Typically, sands bind less contaminant mass than other soil types. Their larger grain size limits their surface area, and the major mineral component (silica) in sand does not have the sorption ability of more clayey soils. Both organic and inorganic contaminants that do adhere to sand particles are relatively easily removed.

Figure H4-2. Illustration of the difficulties of collecting representative samples from particulate parent material.
Figure H4-2. Illustration of the difficulties of collecting representative samples from particulate parent material.

Organic carbon is another important component of soils that plays a large role in giving soil structure and in binding contaminants. A mature compost pile is an excellent example of soil material that is very high in organic carbon. Organic carbon aggregates absorb and concentrate hydrophobic organic contaminants into their matrix. Organic carbon is the primary food for soil microorganisms, and as it is cycled through the microbial community it ages until it turns into humic and fulvic acids. These organic acids are very persistent in soil and have a large surface area. Their molecules include many oxygen atoms carrying partial negative charges that attract contaminant molecules or atoms with an overall positive charge or with positively charged atomic groups. Organic carbon can be a very important soil component that sequesters contaminants.

The bottom line is that the measured concentration of a soil sample greatly depends on whether many or few contaminant-laden particles are present in the analytical subsample. A representative soil sample will have the same proportion of these particles as the targeted soil population. The smaller the sample or subsample, the more difficult it is to get the same proportion as in the parent material. As shown in Figure H4-2, a small sample (Sample A) taken from particulate material can miss concentrated particles that are scattered throughout the original matrix. This scenario leads to an analytical result that is less than the concentration of the original. On the flip side, there is also the chance that a small sample could capture contaminated particles (Sample C), but it is likely that the proportion of “hot” to “clean” particles will be different from the original. Because the ratio for Sample C is skewed in favor of the “hot” particles, Sample C would have an analytical result that is higher than the true concentration of the parent material. A larger sample (Sample B) has a much better chance of accurately representing the original.