Analyses by atomic absorption for detrital gold in more than 2,000 beach, offshore, marine-terrace, and alluvial sands from southern Oregon have shown that the values determined from raw or unconcentrated sediment containing small amounts of gold are neither reproducible nor representative of the initial sample. This difficulty results from a 'particle sparsity effect', whereby the analysis for gold in a given sample depends more upon the occurrence of random flakes of gold in the analyzed portion than upon the actual gold content of the sample.
The particle sparsity effect can largely be eliminated by preparing a gold concentrate prior to analysis. A combination of sieve, gravimetric, and magnetic separation produces a satisfactory concentrate that yields accurate and reproducible analyses. In concentrates of nearly every marine and beach sand studied, the gold occurs in the nonmagnetic fraction smaller than 0.124 mm and with a specific gravity greater than 3.3. The grain size of gold in stream sediments is somewhat more variable.
Analysis of concentrates provides a means of greatly increasing the sensitivity of the analytical technique in relation to the initial sample. Gold rarely exceeds 1 part per million in even the richest black sand analyzed; to establish the distribution of gold (and
platinum) in marine sediments and its relationship to source and environmental factors, one commonly needs to know their content to the part per billion range. Analysis of a concentrate and recalculation to the value in the initial sample permits this degree of sensitivity.
Additional publication details
USGS Numbered Series
Marine sediment sample preparation for analysis for low concentrations of fine detrital gold