Improving the use of eclogitic garnet as a diamond indicator mineral, and constraining the origin of eclogites in the subcontinental lithospheric mantle

• Author / Creator

  Hardman, Matthew

• Diamond occurs in the subcontinental lithospheric mantle (SCLM) and is transported to the surface by kimberlite-lamproite volcanism and other deeply-derived volcanic rocks. In the SCLM diamond is often hosted by peridotitic or eclogitic substrates. Even when it occurs in economic abundances in a kimberlite deposit, diamond is extremely scarce, typically below the parts per million level. Therefore, diamond exploration practices often seek out “diamond indicator minerals,” such as garnet, that may have co-existed with diamond or equilibrated under conditions where diamond may have been stable. Some practices that employ low-Cr garnets to identify diamondiferous deposits hosted by eclogitic substrates, however, are susceptible to error: one of the causes of this is that there is significant compositional overlap between low-Cr garnets from mantle eclogites – which may be diamondiferous – and garnets from lower crustal granulites, which are barren.
  One existing methods for the discrimination of crustal- and mantle-derived garnets is the Mg# (Mg/[Mg+Fe]) versus Ca# (Ca/[Ca+Mg]) method of Schulze (2003). To determine if the Schulze (2003) method can successfully discriminate garnets derived from granulites from those derived from mantle eclogites, I determined the major- and trace-element compositions of garnets from 190 new lower crustal granulite and 529 new mantle eclogite xenoliths, from a variety of kimberlites globally. These data are combined with the major-element compositions of 2977 garnets from published literature. When this combined dataset is applied to the Schulze (2003) method, the full error rate is 17.1 ± 2.1 %. To improve the discrimination of garnets from crustal and mantle rocks, in Chapter 2 I derive new probabilistic single-grain discriminants for crustal and mantle garnets using major-element compositions. These discriminants are based on the multivariate statistical methods linear discriminant analysis and logistic regression. The cross-validated error rate of the logistic regression method is 7.5 ± 1.9 %, the lowest overall in published literature for the classification of low-Cr garnets derived from crustal and mantle low-Cr garnets.
  In Chapter 3 I assess the value of garnet trace-element data in improving classification error rates during diamond exploration. Using a combined dataset of garnet trace-element compositions from new xenoliths in this study and data in the published literature, I find that classification error rates for garnets from crustal granulites and mantle eclogites are improved by adding tracing element data as classifiers. I present a new trace-element classifier using the statistical method Classification and Regression Trees (CART). This CART classifier is additive to the outputs of the major-element method in Chapter 2, and adds garnet Eu-anomalies and Sr concentrations as variables. The combination of the trace-element CART method and the major-element logistic regression method results in an error rate as low as 4.7 % on calibration data.
  Finally, in Chapter 4 I undertake a study of eclogite xenoliths from the former Roberts Victor diamond mine, South Africa. I analysed a new suite of 65 eclogite xenoliths from Roberts Victor for their major- and trace-element compositions. In addition to a new dataset of 34 oxygen isotope analyses by SIMS, I report the first triple oxygen isotope data (δ17O, δ18O) for eight kimberlite-derived eclogites. Eight new samples in the dataset have sub-chondritic whole-rock LREE abundances and are low in Sr, HFSE, sodium-in-garnet, potassium-in-clinopyroxene, Zr/Hf, and δ18O (< 4.0 ‰). These samples classify as Group II eclogites, based on textural equilibrium exemplified by interlocking grains with straight grain boundaries. For the larger sample set of Group I eclogites from Roberts Victor, based on their major- and trace–element characteristics, I concur with previous authors that they are metamorphosed basaltic-picritic lavas or gabbroic cumulates from oceanic crust, crystallised from melts of depleted MORB mantle. For the Group II eclogites, however, I propose formation as cumulates in deep oceanic crust from melts that were chemically less-enriched than N-MORB due to derivation from a residual mantle source. Previous melting of this depleted mantle source at garnet- ± spinel-facies preferentially extracted incompatible elements and fractionated Zr-Hf in the residue. Cumulates precipitated from the second-stage melts inherited the residual chemical signature of their mantle sources. Coupling the low δ18O values of the Group II eclogites, which fall outside of the canonical mantle range, with their variable europium anomalies, indicates that they crystallised in plagioclase-facies oceanic crust.

• Subjects / Keywords

  • exploration
  • garnet
  • crust
  • eclogite
  • kimberlite
  • diamond
  • mantle

• Graduation date

  Spring 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-h68a-py97

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.