Some Observations On Mineral Properties And Analytical Reproducibility In Geochemical Samples

Geochemical laboratories are commonly criticized by geologists about poor analytical reproducibility and erratic anomaly patterns, especially when gold and trace metals from resistant minerals are reported. Geochemical analysis in mineral exploration is a compromise between high productivity on the one hand, imposed by large numbers of samples, and analytical precision and accuracy on the other. However, the physical properties of resistant minerals, such as cassiterite, gold, beryl, chromite, zircon and others, interfere with both of these requirements. Therefore, the degree of sample homogeneity that can reasonably be achieved in sampling and sample preparation must be considered. Subsequently, the understanding of its effects on analytical data quality and the consequences on data interpretation will provide a basis for understanding the problems common to exploration as an interdisciplinary science. Complaints about poor and inadequate analytical performance are not confined to exploration geochemistry. They are a common feature in mining and metallurgy and wherever sampling and analysis of "grains" are concerned - the "nugget" or "grain size effect." The "grain size" effect Poor analytical reproducibility is normal for gold and a well-defined group of other elements usually found in resistant placer minerals. Those who use and interpret geochemical data must realize and appreciate the chemical, the physicochemical, the physical and the mineralogical properties of the elements and their minerals. Especially for the "placer" minerals and their elements, such comprehensive interpretation is crucial. The most important factor to consider is the behavior of a mineral and its metal component(s) during weathering and their integration into the sampled medium, in exploration mostly sediments and soils. (Data interpretation for samples of water, gas, to a certain extent rock, follows different patterns.) A mineral may be chemically and physically stable or it may easily disintegrate physically or decompose chemically, or any combination of such processes, at varying degrees. The breakdown products, in turn, may or may not, interreact with the environment. If the nature of an element or a mineral predestines it for a heterogenous distribution in the sample medium, then nature, size and number of grains likely to occur should be considered in relation to the sample portion taken for analysis. This allows the estimation of their effect on analytical precision and accuracy. Elaborate and sophisticated statistical calculations exist on this subject. But these approaches do not cope with the complexity of the natural surface environment. The miner alogical, chemical and environmental behavior of elements and minerals can be estimated but not calculated. However, the mineral grain sizes and their influence on analytical precision can be precisely calculated if certain conditions, assumptions and idealizations are made. If the geochemical and mineralogical characteristics of minerals and elements are understood, such calculations demonstrate the grain (or nugget) effects that mineral properties and (geo)chemical behavior of minerals and elements cause on the precision and accuracy of geochemical analysis through their influence on sample homogeneity. Two other factors that influence the sample homogeneity and the nugget effect are the efficiency of sample preparation and the sample portion taken for analysis. In this way, certain element- or mineral-specific parameters can be established as a guide for the sampling program. The information, for example, may assist in determining sampling procedures in the field, especially the sample weight to be taken for "representative samples." Also, it may help assess whether analytical data, as provided by the laboratory, are acceptable. Finally, it may help determine the approach in data interpretation. However, all such simplified calculations are based on idealized, that is, unreal assumptions and conditions. As such, they represent one extreme end on the scale of probabilities. The reality is found somewhere away from this extreme, towards homogeneity. An example from a study of an eluvial gold prospect may be given for illustration: •Original sample weight: 20 kg (44 lbs) of rock gravel, crushed and ground to -0.18 mm (-80 mesh). •Au content: 20 grains of Au, average size of about 0.5 mm3 (0.03 cu in.) each = 7.5 mg each, making a total of about 150 mg Au in the sample = 7.5 ppm Au. Assumptions •Au occurs in the sample as free, discrete grains only. •Not more than one grain, if any, goes into each sample split (analyte) portion (20 grains of 7.5 mg Au each). •Analysis of original rock sample: 100 g sample for analysis, 20 kg/100 g = 200 samples 20 samples with 1 grain each: result, - 75 ppm Au. 180 samples with no grain: result, 0 ppm Au chances 1:9 •10 g sample for analysis: