Mercury Distribution around Taconite Concentrators

Abstract

Sources of mercury in the environment have become an area of concern for the various regulatory agencies. Taconite production has been identified as one of the sources. Previous programs have indicated that the majority of the mercury in the ore is rejected to the tailings basins, but it appeared that some plants rejected a greater portion than other plants. The purpose of this study was to determine if the differences in mercury rejection were ore related or flowsheet related. Samples were taken around the various concentrator unit operations from !spat-Inland, National Steel Pellet Company, Evtac, Hibtac, and Minntac. The sampling program was not designed to provide definitive mercury balances around the concentrator, but rather to provide a "snapshot" of what was happening with each unit operation. Each unit operation was sampled so that the feed was taken first, followed quickly by simultaneous sampling of the concentrate and tails. This sampling should provide a good balance around each unit operation. The sampling protocol, potential error sources, and measures to minimize the various potential errors are discussed. All of the samples were analyzed for total iron, Satmagan iron, silica, alumina, calcium, magnesium, manganese and sulfur at the Coleraine Minerals Research Laboratory (CMRL). A split of each sample was sent to Frontier Geosciences for total mercury analyses and sequential extraction mercury analyses on the feed, coarse tails, fine tails and concentrate samples from each plant. The sequential extraction is supposed to give an indication of the mercury compounds present. It was anticipated that the sequential extraction method could be used to determine if different plants had different forms of mercury in their ore and if different mercury compounds behaved differently during concentration. The first stage of extraction uses DI water; the second stage uses a pH 2 solution; the third stage uses lN KOH; the fourth stage uses 12N nitric acid; and the fifth and final stage uses aqua regia. The mercury content in the plant crude ore feeds ranged from a low of 9.44 ppb to 27.90 ppb. The plant concentrates ranged from a low of 6.19 ppb to a high of 16.10 ppb. There appeared to be no relationship between the mercury in the feed and the mercury in the concentrate. The standard Excel correlation program was run to relate the various elements to the mercury analyses. For all plants the correlation coefficients for sulfur and mercury content were greater than 0.8 and were greater than 0.9 for four of the five plants. As would be expected, the mercury content correlated positively with the gangue elements and negatively with the iron and magnetic iron analyses. This indicates a selective rejection of mercury to the tails and suggests that the mercury is not associated with the magnetite. The sequential extractions indicated no significant leaching of mercury in either the water or pH 2 solution. This indicates that there is no easily solubilized mercury present in the ore and that the mercury rejected to the tailings should not leach into the water. The sequential extractions indicated no obvious differences in the feed to the five plants. There is some question as to whether the sequential extraction is reflecting different mercury forms or is just indicating the difficulty in extracting mercury from the mineral lattices. The sequential extractions indicated no differences in plant performances.