Chemical data and statistical interpretations for rocks and ores from the Ranger uranium mine, Northern Territory, Australia

Interlaboratory comparison of analyses of magnesium-rich metamorphic rocks and uranium ores by X-ray fluorescence, induction coupled plasma spectroscopy, and instrumental neutron activation

“…The ICP values do not compare well to the Survey INAA numbers (correlation coefficient of 0.34, average percent deviation of 195.6, table 2) and the numbers that disagree most widely (the "fliers") are all from samples that have a high uranium content. The Oak Ridge INAA, on the other hand, agrees very well with the Survey INAA and the DN values (correlations of 0.9769 and 0.9899, respectively, table 1) and thus the "Th N" values reported in Nash and Frishman (1983) are the more correct determinations. As noted before, the "U N" and "Th" values reported previously did not correlate well, and therefore we now feel that the postulated U-Th correlation at Ranger does not exist.…”

“…Uranium and Rare Earth Elements (REE) Nash and Frishman (1983) refer to a possible U-REE correlation at Ranger, and other investigators have also suggested that such an association may exist in the uranium deposits of the Pine Creek Geosyncline. McLennan and Taylor (1980) used spark-source mass spectrometry as an analytical tool on five mineralized samples from the Cahill Formation, one of which was from Ranger, and concluded that there was indeed a correlation between uranium and the REE.…”

“…It is obvious from table 1 that the Gd analyses listed in Nash and Frishman (1983) are worthless and that the Tb determinations by ICP are subject to very large errors (Gd deviations average nearly 82,000 percent, Tb determinations by ICP deviate by from 1000 to 5000 percent). We still wonder, however, if the overall correlation between U and the REE that appeared from our statistical manipulation of the entire 370 sample data set is real.…”

“…This method also suffers from an oversimplification because total error is no doubt a function of concentration in some cases, but our goal here is to assess the impact of analytical error on the geochemical insights gained from the statistical manipulations presented in Nash and Frishman (1983) and our simplifying assumption should not detract from that end.…”

“…The Oak Ridge analyses for the reference samples (standards BCR-1 and G-2) are published in Appendix II of Nash and Frishman (1983) and that paper also contains estimates of precision for the Oak Ridge analyses based on the 30 replicates. We did not submit blind replicates to the Geological Survey laboratories, but ratios of accepted values to determined values for a number of reference standards run along with the Survey INAA and ICP analyses are included here in Appendices II and III as tables A-l and A-2, respectively.…”