GARNET XENOCRYSTS FROM THE DIAVIK MINE, NWT, CANADA: COMPOSITION, COLOR, AND PARAGENESIS

Recently sampled mantle-derived xenoliths from the A154–S kimberlite pipe, at the Diavik diamond mine, Northwest Territories, Canada, are dominantly eclogitic in paragenesis, suggesting that either the abundance of eclogite at depth exceeds that of peridotite, or that eclogite is sampled or preserved preferentially. To determine the relative abundance of eclogite and peridotite sampled by the ascending kimberlite more accurately and to evaluate the utility of color alone as a guide to composition, we studied xenocrysts of garnet recovered from coarse heavy-mineral concentrates. Over 10,000 garnet crystals were divided into 15 groups according to their color. From these groups, 174 crystals were analyzed for major and minor elements using an electron microprobe (WDS). Concentrations of trace elements of 55 garnet xenocrysts were measured by LA–ICP–MS. Orange xenocrysts have 2 O 3 , and are classified as having an eclogitic derivation according to two sets of criteria. Pink, red, and purple xenocrysts contain >2 wt.% Cr 2 O 3 , and are inferred to have a peridotitic derivation. The range of Cr content of purple xenocrysts extends to higher values (~14 wt.% Cr 2 O 3 ) than do those of pink (~10 wt.%) or red crystals (~9 wt.%). Orange-red and red-orange xenocrysts have Cr concentrations overlapping those of the orange and red groups, and may be either eclogitic or peridotitic. Chondrite-normalized REE (REE N ) patterns are either “normal”, with low LREE N , increasing concentrations into the MREE N , and relatively flat HREE at ~10–30× chondrite, or sinusoidal, with normalized concentrations increasing from La to Nd, decreasing from Nd to Ho, and then increasing again to Lu. “Normal” patterns are typical of orange and orange-red xenocrysts, although some show a LREE enrichment. The sinusoidal patterns are characteristic of purple and some pink and red xenocrysts. Garnet xenocrysts of pyroxenitic (including eclogitic and websteritic) and peridotitic affinity are sufficiently distinct in color to permit assignment of the >10,000 garnet crystals in our sample to one of the two suites based solely on color. Color alone, however, could not differentiate unambiguously between different peridotitic parageneses. In contrast to the xenolith data, we infer from the relative abundances of the xenocrystic garnet that the A154–S mantle sample is ~96% peridotite.