A comparison of two methods to assess diamond potential using major and trace element analysis of diamond heavy mineral concentrate (peridotitic and eclogitic garnets)

Thirty garnets from each of nine kimberlite pipes from southern Africa and Yakutia were analysed for major element as well as for trace element concentrations using electron and proton probes. The concentrates came from a mix of barren and diamondiferous kimberlites as well as from both on-craton and off-craton localities. The major element approach (using plots of CaO vs Cr\\(_2\\)O\\(_3\\)and Na\\(_2\\)O vs TiO\\(_2\\)) was used to predict the diamond potential of the source(s) from which the garnets were derived. Histograms of TiO\\(_2\\) were used to separate high and low temperature garnet populations. The major element method correctly predicted the diamond potential of the source of five out of nine (56%) of the concentrates from the geochemistry of the eclogitic garnets contained within them. The same garnets were analysed for trace element levels of nickel and the nickel concentration in each garnet grain was used to calculate its temperature of crystalisation by use of the garnet-nickel geothermometer. This temperature was then related to diamond potential by interpreting it in terms of the graphite and diamond stability fields. The two sets of predictions were then compared. The garnet-nickel geothermometer gave an accurate assessment of the diamond potential for six of eight pipes studied (75%). The two inaccurate predictions derived from the interpretative process not taking into consideration the shallowness of the lithosphere for one source (Nouzee) - giving a steeper than \ideal\" geotherm- and to the presence of a low temperature suite of concentrate garnets in the Roberts Victor sample. The analyses failed to confirm the presence of a single G10 garnet in one hundred and twenty seven peridotitic garnets examined from eight of the heavy mineral concentrates even though seven of these came from diamondiferous kimberlites five of which were of economic grade. This indicates that in those areas where soil sampling recovers quite small numbers of (peridotitic) garnets it may not be possible to adequately assess diamond potential using the presence or absence of G10 garnets."