Elsevier

Lithos

Volumes 318–319, October 2018, Pages 478-493
Lithos

New geochemical constraints on the origins of MARID and PIC rocks: Implications for mantle metasomatism and mantle-derived potassic magmatism

https://doi.org/10.1016/j.lithos.2018.08.036Get rights and content

Highlights

  • PIC rock genesis by kimberlite melt metasomatism of peridotite

  • No conclusive evidence to confirm MARID metasomatic or magmatic genetic models

  • MARID-derived melt compositions match those of alkaline magmas (e.g., orangeites)

Abstract

MARID (Mica-Amphibole-Rutile-Ilmenite-Diopside) and PIC (Phlogopite-Ilmenite-Clinopyroxene) rocks are unusual mantle samples entrained by kimberlites and other alkaline volcanic rocks. The formation of MARID rocks remains hotly debated. Although the incompatible element (for example, large ion lithophile element) enrichment in these rocks suggests that they formed by mantle metasomatism, the layered textures of some MARID samples (and MARID veins in composite xenoliths) are more indicative of formation by magmatic processes. MARID lithologies have also been implicated as an important source component in the genesis of intraplate ultramafic potassic magmas (e.g., lamproites, orangeites, ultramafic lamprophyres), due to similarities in their geochemical and isotopic signatures. To determine the origins of MARID and PIC xenoliths and to understand how they relate to alkaline magmatism, this study presents new mineral major and trace element data and bulk-rock reconstructions for 26 MARID and PIC samples from the Kimberley-Barkly West area in South Africa. Similarities between compositions of PIC minerals and corresponding phases in metasomatised mantle peridotites are indicative of PIC formation by pervasive metasomatic alteration of peridotites. MARID genesis remains a complicated issue, with no definitive evidence precluding either the magmatic or metasomatic model. MARID minerals exhibit broad ranges in Mg# (e.g., clinopyroxene Mg# from 82 to 91), which may be indicative of fractionation processes occurring in the MARID-forming fluid/melt. Finally, two quantitative modelling approaches were used to determine the compositions of theoretical melts in equilibrium with MARID rocks. Both models indicate that MARID-derived melts have trace element patterns resembling mantle-derived potassic magma compositions (e.g., lamproites, orangeites, ultramafic lamprophyres), supporting inferences that these magmas may originate from MARID-rich mantle sources.

Introduction

MARID (Mica-Amphibole-Rutile-Ilmenite-Diopside) rocks are coarse-grained, ultramafic, and ultrapotassic (4.0–9.5 wt% K2O) in composition (Dawson and Smith, 1977; Waters, 1987a). They occur as discrete xenoliths, or vein assemblages in composite xenoliths, entrained by archetypal (Group I) kimberlites and orangeites (formerly known as Group II or micaceous kimberlites; e.g., Smith, 1983). Bulk-rock enrichments in large ion lithophile elements (LILE) and light rare earth elements (LREE) have led to suggestions that MARID rocks represent a possible end-member product of progressive mantle metasomatism (Grégoire et al., 2002; Waters, 1987a, Waters, 1987b; Waters and Erlank, 1988). MARID rocks have also been proposed as a major source of mantle-derived ultramafic potassic magmas including lamproites (Matchan et al., 2009), ultramafic lamprophyres (Tappe et al., 2008, Tappe et al., 2017), kamafugites (Rosenthal et al., 2009), and orangeites (Giuliani et al., 2015). MARID rocks may also act as contaminants to magmas derived from the deeper, convective upper mantle (e.g., kimberlites), thereby causing significant shifts in their isotopic compositions (e.g., Tappe et al., 2011).

Grégoire et al. (2002) used mineral geochemical criteria to differentiate MARID rocks from a distinct type of strongly metasomatised, phlogopite-dominated mantle xenolith, which they termed ‘PIC’ (for its dominant mineralogy of Phlogopite-Ilmenite-Clinopyroxene). The majority of known MARID and PIC xenoliths originate from the Kimberley kimberlites (South Africa), although MARID samples have also been found in other southern African kimberlites (e.g., Letlhakane in the Magondi fold belt, and the Prieska region at the SW margin of the Kaapvaal craton; Field et al., 2008; Skinner et al., 1994; Stiefenhofer et al., 1997; Fig. 1), as well as in orangeites in the Barkly West area (e.g., Newlands and Roberts Victor: Dawson and Smith, 1977; Waters, 1987a, Waters, 1987b). The relative abundance of archetypal MARID rocks in the Kimberley kimberlites compared to other localities remains unexplained. Mantle metasomatism generating mica-amphibole-clinopyroxene-rich mantle lithologies (i.e. broadly mineralogically similar to MARID) occurs in many other locations (e.g., Germany and Uganda: Lloyd and Bailey, 1975; Canada: Peterson and le Cheminant, 1993; Tappe et al., 2006; Morocco: Wagner et al., 1996), but these rocks cannot be classified as MARID rocks because they do not display all of the following distinctive MARID characteristics:

  • 1)

    The presence of K-richterite (as opposed to Ca-rich amphiboles);

  • 2)

    Al-Cr-depletion of all silicate phases compared to peridotitic mineral compositions. This is most notable in phlogopite, wherein a significant amount of Fe substitutes for Al and Si into the tetrahedral site (e.g., Dawson and Smith, 1977); and

  • 3)

    The absence of olivine coexisting with MARID phases.

The distinguishing features of MARID rocks can also be used to differentiate between MARID and PIC rocks. For example, PIC rocks do not contain K-richterite, and PIC minerals generally contain less FeOT (i.e., total Fe expressed as FeO) than MARID minerals (i.e. PIC minerals have higher Mg#; Grégoire et al., 2002).

Despite several studies of the geochemistry of MARID xenoliths from South African localities, the genesis of these rocks remains unclear. MARID rocks may represent magmatic veins in the lithospheric mantle (Dawson and Smith, 1977; Jones et al., 1982; Waters, 1987a), which may be related to the crystallisation of orangeite magmas, based on trace element and Sr-Nd isotope evidence (e.g., Grégoire et al., 2002). This model may also require the fractionation of both olivine- and carbonate-rich components from the orangeite melt (Sweeney et al., 1993). The crystallisation of MARID rocks from a melt in the lithosphere may have led to the interaction of surrounding wall-rock with residual incompatible element-rich fluids, causing the adjacent peridotites to become metasomatised (e.g., Waters et al., 1989). Alternatively, MARID rocks may be formed by progressive metasomatic alteration of mantle peridotites (e.g., Sweeney et al., 1993). Erlank et al. (1987) described a metasomatic continuum from pristine garnet peridotite (GP), to garnet phlogopite peridotite (GPP) and phlogopite peridotite (PP), and finally to phlogopite K-richterite peridotite (PKP) based on petrography and mineral chemistry. According to Erlank et al. (1987), the development of phlogopite-K-richterite-rich rocks (including MARID) is the result of the most extensive metasomatism. However, owing to differences in mineral major element compositions in MARID and PKP samples, Erlank et al. (1987) proposed that the two could not be directly genetically related. Rocks belonging to the more recently defined PIC assemblage are suggested to form by extensive metasomatic alteration of peridotites by kimberlite melts, based on considerations of their trace element and radiogenic isotope compositions (Grégoire et al., 2002).

To provide new constraints on the genesis of MARID and PIC rocks, the compositions of their parental metasomatic fluids/melts, and the links between MARID and PIC rocks and ultramafic potassic magmas, we have examined the petrography and mineral major and trace element chemistry of 26 MARID and PIC xenoliths from southern African kimberlites and orangeites in the Kimberley and Barkly West areas (Fig. 1). “Reconstructed” bulk-rock compositions (calculated from modal and compositional data for primary minerals, thereby excluding secondary material in late stage veins and cracks) are presented and used to perform melting models of the unaltered MARID composition.

Section snippets

Geological setting

The xenolith samples used in the current study were collected from the Kimberley, Bultfontein, Wesselton, De Beers, and Kamfersdam kimberlites, and the Newlands orangeite, which are all located in or close to the town of Kimberley, in the Western terrane of the Kaapvaal craton (Field et al., 2008; Fig. 1). The Kimberley, Bultfontein, Wesselton, De Beers, and Kamfersdam pipes host archetypal kimberlite (Shee, 1985; Smith, 1983), whereas Newlands is an orangeite (e.g., Smith, 1983). The Kimberley

Samples and analytical methods

The 26 xenolith samples examined in this study were selected from a larger group of 40 xenoliths obtained from the collections housed in the John J. Gurney Upper Mantle Research Collection at the University of Cape Town, and the De Beers Consolidated Mines rock store. Additional samples were collected in 2015 and 2016 from the Boshof Road dumps, which host historical waste material from mining of the Bultfontein kimberlite. The samples vary from 5 to 15 cm in length. Samples for this study were

Petrography

Several samples examined in this study display massive textures (Fig. 2a), whereas others appear foliated due to the preferential orientation of mica grains. A few samples display modal layering of the main silicate minerals (typically of phlogopite and K-richterite or clinopyroxene) into discrete bands (Fig. 2b). Such textures are common features of MARID rocks (Dawson and Smith, 1977; Waters, 1987a, Waters, 1987b). One sample (AJE-319) is a dunite that hosts discontinuous veins of phlogopite

Mineral chemistry

To obtain representative major and trace element compositions of minerals from each sample, five to ten grains of each MARID or PIC mineral were analysed in thin section. Mineral compositions are generally homogeneous within a single sample, and the ranges of values are due to variations between samples. All chemical analyses acquired in this study are presented in Appendix A. Considerably larger compositional variability is associated with rims of MARID phases that are in contact with

Bulk-rock reconstructions

Bulk-rock compositions of MARID and PIC samples have been reported in several studies (e.g., Grégoire et al., 2002; Waters, 1987a, Waters, 1987b). However, XRF-based and solution-mode determinations of bulk-rock compositions fail to account for the ubiquitous late-stage modification of mantle samples, including MARID rocks, entrained by kimberlites (e.g., carbonate veins, mineral zonation; cf. Dawson and Smith, 1977; Fitzpayne et al., 2018; Richardson et al., 1985; Simon et al., 2007).

Discussion

Since the original classification of MARID rocks (Dawson and Smith, 1977), several studies have attempted to unravel MARID genesis by investigating MARID geochemistry and mineralogy (e.g., Grégoire et al., 2002; Konzett et al., 2014; Waters, 1987a). In contrast, PIC rocks were first described as a distinct lithology by Grégoire et al. (2002), owing to differences in mineral major and trace element compositions, as well as radiogenic isotope compositions. The data presented in this study broadly

Conclusions

This study presents a large suite of mineral major and trace element compositions, as well as reconstructed whole-rock major and trace element compositions, for MARID and PIC rocks from the Kimberley and Barkly West areas of South Africa. The new mineral chemical data support previous suggestions that PIC rocks are formed by intense peridotite metasomatism due to kimberlite melts. The MARID mineral major element compositions presented here exhibit wider ranges than previously reported. Although

Acknowledgements

We acknowledge the support of Graham Hutchinson during SEM and EPMA sessions, and Alan Greig for his help with LA-ICP-MS analyses. We also thank De Beers Consolidated Mines, the University of Cape Town, John Gurney and Simon Shee for access to the samples examined in this study, and Jock Robey for invaluable support during field work in the Kimberley area. We thank Andrew Kerr for his efficient editorial handling of this manuscript. Finally, we would like to thank Michel Grégoire, Sebastian

References (85)

  • A. Giuliani et al.

    Mantle oddities: a sulphate fluid preserved in a MARID xenolith from the Bultfontein kimberlite (Kimberley, South Africa)

    Earth Planet. Sci. Lett.

    (2013)
  • A. Giuliani et al.

    Constraints on kimberlite ascent mechanisms revealed by phlogopite compositions in kimberlites and mantle xenoliths

    Lithos

    (2016)
  • M. Grégoire et al.

    Spinel lherzolite xenoliths from the Premier kimberlite (Kaapvaal craton, South Africa): nature and evolution of the shallow upper mantle beneath the Bushveld complex

    Lithos

    (2005)
  • R. Hin et al.

    Formation and temporal evolution of the Kalahari sub-cratonic lithospheric mantle: Constraints from Venetia xenoliths, South Africa

    Lithos

    (2009)
  • D.E. Jacob

    Nature and origin of eclogite xenoliths from kimberlites

    Lithos

    (2004)
  • I. Katayama et al.

    Mineral chemistry and P–T condition of granular and sheared peridotite xenoliths from Kimberley, South Africa: Origin of the textural variation in the cratonic mantle

    Lithos

    (2009)
  • J. Konzett et al.

    The timing of MARID metasomatism in the Kaapvaal mantle: an ion probe study of zircons from MARID xenoliths

    Earth Planet. Sci. Lett.

    (1998)
  • J. Konzett et al.

    Two episodes of fluid migration in the Kaapvaal Craton lithospheric mantle associated with cretaceous kimberlite activity: evidence from a harzburgite containing a unique assemblage of metasomatic zirconium-phases

    Lithos

    (2013)
  • J. Konzett et al.

    The formation of saline mantle fluids by open-system crystallization of hydrous silicate-rich vein assemblages – evidence from fluid inclusions and their host phases in MARID xenoliths from the central Kaapvaal Craton, South Africa

    Geochim. Cosmochim. Acta

    (2014)
  • M. Lazarov et al.

    Thermal state and redox conditions of the Kaapvaal mantle: a study of xenoliths from the Finsch mine, South Africa

    Lithos

    (2009)
  • F.E. Lloyd et al.

    Light element metasomatism of the continental mantle: the evidence and the consequences

    Phys. Chem. Earth

    (1975)
  • E. Matchan et al.

    The geochemistry, petrogenesis and age of an unusual alkaline intrusion in the western Pilbara craton, Western Australia

    Lithos

    (2009)
  • W.F. McDonough et al.

    The composition of the Earth

    Chem. Geol.

    (1995)
  • S. Mollo et al.

    A new test for equilibrium based on clinopyroxene-melt pairs: clues on the solidification temperatures of Etnean alkaline melts at post-eruptive conditions

    Chem. Geol.

    (2013)
  • D.G. Pearson et al.

    Re-Os, Sm-Nd, and Rb-Sr isotope evidence for thick Archaean lithospheric mantle beneath the Siberian craton modified by multistage metasomatism

    Geochim. Cosmochim. Acta

    (1995)
  • T. Rehfeldt et al.

    Contrasting types of metasomatism in dunite, wehrlite and websterite xenoliths from Kimberley, South Africa

    Geochim. Cosmochim. Acta

    (2008)
  • S.H. Richardson et al.

    Kimberlite-borne garnet peridotite xenoliths from old enriched subcontinental lithosphere

    Earth Planet. Sci. Lett.

    (1985)
  • A. Rosenthal et al.

    Petrogenesis of strongly alkaline primitive volcanic rocks at the propagating tip of the western branch of the East African Rift

    Earth Planet. Sci. Lett.

    (2009)
  • R.L. Rudnick et al.

    Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and geochemical characteristics

    Earth Planet. Sci. Lett.

    (1993)
  • K. Schmidt et al.

    Trace element partitioning between phlogopite, clinopyroxene and leucite lamproite melt

    Earth Planet. Sci. Lett.

    (1999)
  • N.S.C. Simon et al.

    The origin of garnet and clinopyroxene in “depleted” Kaapvaal peridotites

    Lithos

    (2003)
  • A. Soltys et al.

    A new approach to reconstructing the composition and evolution of kimberlite melts: a case study of the archetypal Bultfontein kimberlite (Kimberley, South Africa)

    Lithos

    (2018)
  • S. Tappe et al.

    Between carbonatite and lamproite—Diamondiferous Torngat ultramafic lamprophyres formed by carbonate-fluxed melting of cratonic MARID-type metasomes

    Geochim. Cosmochim. Acta

    (2008)
  • S. Tappe et al.

    A fresh isotopic look at Greenland kimberlites: cratonic mantle lithosphere imprint on deep source signal

    Earth Planet. Sci. Lett.

    (2011)
  • S. Tappe et al.

    Sources and mobility of carbonate melts beneath cratons, with implications for deep carbon cycling, metasomatism and rift initiation

    Earth Planet. Sci. Lett.

    (2017)
  • F. Viljoen et al.

    Geochemical processes in peridotite xenoliths from the Premier diamond mine, South Africa: evidence for the depletion and refertilisation of subcratonic lithosphere

    Lithos

    (2009)
  • L. Wasch et al.

    An alternative model for silica enrichment in the Kaapvaal subcontinental lithospheric mantle

    Geochim. Cosmochim. Acta

    (2009)
  • P.A. Winterburn et al.

    Peridotite xenolith from the Jagersfontein kimberlite pipe: I. Primary and primary-metasomatic mineralogy

    Geochim. Cosmochim. Acta

    (1990)
  • G. Witt-Eickschen et al.

    The effect of temperature on the equilibrium distribution of trace elements between clinopyroxene, orthopyroxene, olivine and spinel in upper mantle peridotite

    Chem. Geol.

    (2005)
  • J. Adam et al.

    Trace element partitioning between mica-and amphibole-bearing garnet lherzolite and hydrous basanitic melt: 2. Tasmanian Cainozoic basalts and the origins of intraplate basaltic magmas

    Contrib. Mineral. Petrol.

    (2011)
  • S. Aulbach et al.

    Subcontinental lithospheric mantle origin of high niobium/tantalum ratios in eclogites

    Nature

    (2008)
  • S. Aulbach et al.

    Constraints from eclogite and MARID xenoliths on origins of mantle Zr/Hf-Nb/Ta variability

    Contrib. Mineral. Petrol.

    (2011)
  • Cited by (53)

    View all citing articles on Scopus
    View full text