Rare earth element analysis by isotope dilution

At Stony Brook we have been analyzing igneous rocks, granites through gabbros, basalts, and ultramafic rocks for rare earth elements (REE) as well as the other major, minor, and trace elements in order to determine the origin of these rocks utilizing quantitative trace element models. For rare earths we are using the analytical procedure of J. G. Arth, USGS, Reston, Virginia, which utilizes two Dowex 50W by 8 cation columns, one for nitric acid to get a group separation of the REE and the other for hydrochloric acid to separate the specific REE. The rare earth elements La, Ce, Nd, Sm, Eu, Gd, Dy, Er, Yb, and Lu can usually be analyzed in three to four mass spectrometric runs with a total time for analysis of about 8 hours. The number of REE analyzed should not be reduced or the great advantage of the REE, the regular but not consistent variation from element to element, cannot be fully utilized in an interpretation. Precision is about 1 percent for all elements except for Lu and Gd. The uncertainty in the accuracy is somewhat greater. Much of the REE data in the literature has uncertainties of about 5 percent of greater, and for gross comparison of rock types this data is adequate. REE data however, with uncertainties much greater than 10 percent are no longer acceptable, as they are not useful even for gross comparisons of rock types. In a detailed study of a magmatic sequence in order to distinguish differentiated fractions versus different original parents and to determine which residual or fractionating minerals are involved, then high precision on the order of 1 percent is absolutely necessary. In such a sequence, since it is only necessary to compare samples with each other, high precision is adequate as one can normalize all samples to a sample in the sequence. In order to compare data with other laboratories, however, high accuracy is also necessary. The uncertainty in the accuracy is mainly in not having adequate rare earth standards