Garnet dissolution and the emplacement of kimberlites

https://doi.org/10.1016/S0012-821X(99)00019-9Get rights and content

Abstract

The dissolution of mantle-derived garnet in H2O-bearing kimberlite melt was investigated experimentally to provide some constraint on the survival of garnet xenocrysts commonly found in kimberlites. Garnet dissolution was determined by measuring the rate of radius change of garnet spheres in synthetic kimberlite melt above its liquidus as a function of both temperature and pressure. To apply the results, a population of 80 garnet xenocrysts in a hypabyssal facies, macrocrystic kimberlite in the Slave Province, Canada, were examined in detail petrographically. The time scale required for kimberlite ascent in order to preserve these garnets is dependent on the temperature path of the magma adopted during ascent. Assuming a minimum temperature of the kimberlite solidus, some estimate of the minimum ascent rate of kimberlites can be derived. Garnet xenocrysts would dissolve in one to ten hours at 1000–1200°C in kimberlite magmas rising from the mantle. The dissolution data also require that kimberlite magmas could only have existed near their liquidus temperature in the mantle for time periods of minutes. Newly-formed fritted dissolution rims on cracked garnet xenocrysts, possibly related to explosive emplacement, indicate that the very final stage of kimberlite ascent in the crust to the root zone of the diatreme possibly occurred on a time scale of only minutes to seconds.

Introduction

Little is known about the ascent path of kimberlites from their depth of origin to the surface where they eventually form diatremes during emplacement at shallow levels in the Earth's crust. The mechanism of ascent for kimberlite magmas must to some degree be dictated by their physical properties, but these have been difficult to constrain. At pressures less than 3 or 4 GPa, the solubility of CO2 in kimberlite magma changes drastically, causing exsolution of a fluid-phase [1], enhancing the buoyancy of the kimberlite magma, causing rapid rise through the crust, further exsolution and eventual emplacement, possibly explosively [2]. The result of this ascent path is intense disaggregation of mantle-derived xenoliths to form xenocryst-laden magma, sometimes intermixed with crustally-derived rock fragments 3, 4, 5. The characterization of this entire magmatic system (magma, fluid phase and abundant xenolithic material) is complex, and has obscured a clear understanding of the physical properties of kimberlites.

Some knowledge of the ascent mechanisms of kimberlites at shallow levels in the crust has come from the geology of diatremes, with evidence for hydrofracture and intense brecciation of country rock 3, 4, 5. Information on the deeper ascent path for these magmas, from the mantle to the base level (or root zone) of the diatreme, however, is not preserved in the diatreme. This part of the ascent path might be recorded in xenolithic material entrained within the kimberlite itself, such as compositional zoning of some minerals. For example, Franz et al. [6]measured Ca zoning profiles in olivine from mantle xenoliths in Gibeon kimberlites from Namibia, and suggested that these samples resided near the crust–mantle boundary for weeks in their host kimberlite magma. If this inference is correct, it would imply that kimberlite magma moves in pulses, with a hiatus in its rise from the mantle to the crust, perhaps near the rheological boundary between the crust and mantle. The length of this hiatus could control the preservation of diamond xenocrysts transported in kimberlites.

Mantle xenoliths disaggregate in kimberlites during decompression, releasing garnet xenocrysts into magma. One constraint bearing on the deep-seated movement of kimberlite may potentially be derived from this process. The ascent rate of the kimberlite must be fast enough to preserve garnet xenocrysts that are in a reaction relation with their host magma. The rate of dissolution of garnet in kimberlite magmas dictates the maximum time the garnet xenocrysts could have resided in the magma during ascent and emplacement before eventual formation of diatremes. The aim of this study is to provide some experimental and petrographical constraints on the emplacement processes in kimberlite by determining the dissolution rate of garnet in kimberlite magma near its liquidus. This experimental data is then applied to petrographic observations for 80 garnet xenocrysts hosted in a macrocrystic, hypabyssal facies phase of a kimberlite from the Archean Slave Province, Canada. The results of the study provide some constraints on the rise of kimberlite magmas through the mantle and crust, and give some insight into reaction kinetics in kimberlite magmas.

Section snippets

Methods

Spheres of natural pyrope-rich mantle garnet (Table 1) were made by abrading chips of garnet following the method of Bond [7]. After abrasion, the spheres were washed ultrasonically for 5 minutes in dilute HCl and examined optically. A subset of the most spherical garnets from this process were chosen for use in the experiments, and their diameters were measured several times with a micrometer, giving a maximum and minimum radius.

The garnet dissolution experiments were conducted using a

Garnet dissolution

All dissolution experiments in H2O-bearing kimberlite SK3 contained garnet, or its reaction products, set in a matrix of glass, long blades of olivine and bubbles with a seriate range of sizes. The larger bubbles were present during the experiment, as evidenced by their impingement by blades of olivine that formed during quenching (Fig. 2a). The presence of this texture demonstrates that the kimberlite melt was fluid-saturated at the PT conditions of these experiments, and that the large

Discussion

Because the experiments in this study were performed on garnet spheres, it was deemed appropriate to apply the results only to natural garnet xenocrysts that approximated a spherical shape. Thus, many of the 80 natural mantle garnets with oblong and strongly oval shapes were eliminated from application of the experimental data, although their sizes and features fall within the limits of the spherical garnet subset, and should not affect the conclusions developed below. A subset of 20 spherical

Acknowledgements

We especially thank D. Bateman, S. Modeland and J. Spence for their assistance with sample preparation and some of the measurements on the kimberlite garnets. We thank J. Ganguly, C. Shaw, D. Smith, and M. Kirkley for their helpful reviews. Access to the Grizzly drill core was generously provided by BHP Minerals (Kelowna). This research was supported by Research and Equipment grants by NSERC of Canada to DC. Sampling of the kimberlite drill core was assisted with a LITHOPROBE grant to DC. [CL]

References (26)

  • W. Bond

    Making small spheres

    Rev. Sci. Instrum.

    (1951)
  • K. Bose et al.

    Quartz–coesite revisited: reversed experimental determination at 500–1200°C and retrieved thermochemical properties

    Am. Mineral.

    (1995)
  • D.W. Williams et al.

    Melting curve of diopside to 50 kilobars

    J. Geophys. Res.

    (1969)
  • Cited by (0)

    View full text