GSA 2020 Connects Online

Paper No. 1-6
Presentation Time: 2:45 PM

REDUCING LABORATORY COSTS BY ESTIMATING TOTAL RARE EARTH CONCENTRATIONS FROM SEVEN OR FEWER ANALYZED ELEMENTS IN WESTERN NORTH DAKOTA COALS


KRUGER, Ned W., North Dakota Geological Survey, 600 East Boulevard Avenue, Bismarck, ND 58505-0840

Obtaining total lanthanide, yttrium, and scandium (rare earths) concentrations requires ICP-MS analysis of sixteen elements which costs hundreds of dollars per sample. If accurate totals can be estimated from a smaller number of element analyses, the cost savings can be applied to gathering more data points to better characterize a study area. A dataset of 828 sample analyses from lignite and associated rock types collected in western North Dakota, each with a complete suite of lanthanides, yttrium, and often scandium, was examined for patterns of elemental distributions which could be used to estimate total rare earth concentrations from laboratory analyses of seven or fewer of the sixteen elements.

The seven elements (Ce, Er, Gd, La, Nd, Sc, Y) chosen as lab-derived concentrations were selected based on their overall prevalence, affiliations with other rare earths, and economic rank. The remaining nine estimated-element concentrations were derived from scatter plot trendlines of each “unknown” element compared to its closest affiliated element with a lab concentration. A total rare earth concentration estimate was then calculated for each sample. The process was repeated, removing one known lab concentration at a time, to evaluate accuracies based on as few as three elements.

The total concentrations derived from the seven-element analyses for all samples in the dataset were within a deviation range of -2.6 to 1.8%, with an average deviation from laboratory reported concentrations of ±0.37%. The accuracy of the six-element analyses fell within a deviation range of -4.1 to 5.5% with an average deviation of ±0.81%. The four-element analyses fell within a deviation range of -10.2 to 6.3% with an average deviation of ±1.55%. Deviations for estimated elements individually fell within broader ranges.

In later sampling, the seven-element estimates of 21 new samples deviated from the lab concentrations within a range -0.6 to 1.4%, which is inside the range observed in the original dataset. Included was a lab reported total concentration of 1,598 ppm, believed to be the highest whole coal analysis reported to date from a North American coal, which was just one ppm less than had been estimated through the seven-element calculations.