Error Associated with the Sampling of Gold Ores

"Sampling is an important step for ore body evaluation and daily plant material balances. The Gy’s equation is regularly used to assess the size of the samples to be collected and to estimate the reproducibility of the collected samples. The equation works well for iron and base metal ores but seems to produce pessimistic estimates of the gold content reproducibility for gold ores. A model is proposed to simulate the division of samples using a standard splitter and discrepancies are observed between the Gy’s equation predictions and simulated samples reproducibility for crushed gold ore. Results of the simulation show that sampling a pile of ore crushed to -2 mm should lead to more reproducible gold content than it is observed in practice and predicted by the Gy’s equation.INTRODUCTION With current gold prices above $1400 per troy ounce operators seek for new economical ore bodies. The metallurgical evaluation of an ore body is one of the numerous steps required to justify the investment for the development of a gold mine. Metallurgical testing requires to sample the ore body and to prepare the ore sample for the grinding and separation tests. The preparation implies crushing the ore usually to less than 2mm and its division into batches of a given weight for subsequent processing. For gold ore testing a batch of crushed ore is ground in a laboratory mill and leached using cyanide. Gold recovered in the solution is used as an indicator of the ore processing amenability. One difficulty associated with ore testing is due to the reproducibility of gold content measurements of the feed sample. According to the Gy’s equation (Gy, 1979; Morisson, 2009) the division of a 20 kg batch of -2 mm ore containing 5 g/t of gold into batches of 1 kg may lead to batches containing 5 ± 5 g/t of gold. For this reason the gold content of a feed sample is usually back calculated from the assays of the solution and solids residues of the leaching process. However, these two assays may contain measurement errors that can propagate through the calculation of the gold recovery. The use of the feed assays as a redundant information could add robustness in the evaluation of ore amenability to processing. This paper discusses the problem of sampling reproducibility for a gold ore. The analysis is carried out with reference to the Gy’s equation for the sampling of particular materials."